Datasets:

Saeed

/

TissueGPT_dataset

like 0

10. 7554/eLife. 93220

eLife

Transdifferentiation of fibroblasts into muscle cells to constitute cultured meat with tunable intramuscular fat deposition

Current studies on cultured meat mainly focus on the muscle tissue reconstruction in vitro, but lack the formation of intramuscular fat, which is a crucial factor in determining taste, texture, and nutritional contents. Therefore, incorporating fat into cultured meat is of superior value. In this study, we employed the myogenic/lipogenic transdifferentiation of chicken fibroblasts in 3D to produce muscle mass and deposit fat into the same cells without the co-culture or mixture of different cells or fat substances. The immortalized chicken embryonic fibroblasts were implanted into the hydrogel scaffold, and the cell proliferation and myogenic transdifferentiation were conducted in 3D to produce the whole-cut meat mimics. Compared to 2D, cells grown in 3D matrix showed elevated myogenesis and collagen production. We further induced fat deposition in the transdifferentiated muscle cells and the triglyceride content could be manipulated to match and exceed the levels of chicken meat. The gene expression analysis indicated that both lineage-specific and multifunctional signalings could contribute to the generation of muscle/fat matrix. Overall, we were able to precisely modulate muscle, fat, and extracellular matrix contents according to balanced or specialized meat preferences. These findings provide new avenues for customized cultured meat production with desired intramuscular fat contents that can be tailored to meet the diverse demands of consumers.

Introduction Cultured meat is an innovative and emerging technique that produces meat directly from cell cultures, potentially providing a high-quality, safe, and stable source of animal protein ( Chriki and Hocquette, 2020 ). In comparison to traditional livestock and poultry farming, cultured meat generates fewer greenhouse gases, utilizes less arable land, and causes less animal harm ( Mattick et al. , 2015 ). Proper selection of seed cells is one of the keys to the success of cultured meat production. The starting cells must be easily obtainable and be able to proliferate numerous times to enable mass production. While the stem cells like muscle stem cells and pluripotent stem cells have frequently been utilized as the cellular source for cultured meat, these stem cells are rare in the animal body and difficult to obtain and amplify on a large scale. In contrast, the somatic cells, which constitute the body, could be efficiently converted into muscle cells under certain conditions. The fibroblast is one of the most abundant and widely distributed cell types present in the body and could be easily collected via minimally invasive biopsy procedure without sacrificing the farm animals. The fibroblasts can replicate indefinitely in vitro and are amenable to myogenesis, adipogenesis, and chondrogenesis ( French et al. , 2004 ; Yin et al. , 2010 ), which produce muscle, fat, and extracellular matrix (ECM) proteins that constitute the meat and the associated texture and flavor. Recently, the fibroblast cells from farm animals have been utilized as the source cells for cultured meat production, with or without myogenesis ( Jeong et al. , 2022 ; Pasitka et al. , 2023 ), demonstrating the feasibility of somatic cell-derived seed cells as a sustainable and ethical option for cultured meat production. We have previously developed a protocol for the controlled transdifferentiation of chicken fibroblasts into myoblasts, which subsequently form multinucleated myotubes and express mature muscle proteins ( Ren et al. , 2022 ). Moreover, chicken fibroblasts could also be efficiently converted into fat-depositing lipocytes by treating them with chicken serum (CS) medium or other substances such as fatty acids ( Hausman, 2012 ; Kim et al. , 2020 ; Lee et al. , 2021 ). Therefore, the induced myogenic and adipogenic competency, along with the inherent fibrogenic collagen-producing ability of chicken fibroblasts, can enable us to simultaneously synthesize muscle, fat, and collagen during the cultured meat production. Nevertheless, further techno-functional research is necessary to precisely control the composition of the end product of fibroblast-derived cultured meat in order to achieve a more balanced and personalized nutritional profile and meet the specific consumer preferences. In this study, the chicken fibroblast cells were implanted into the hydrogel scaffold to analyze the 3D cellular dynamics involving cell proliferation and myogenic/adipogenic transdifferentiation. We also optimized the low-serum culture conditions of chicken fibroblasts to reduce the cost of mass production. The myogenic transdifferentiated cells were confirmed to be skeletal muscle lineage but not myofibroblasts. Importantly, the cells were subjected to myogenesis and adipogenesis sequentially in 3D hydrogel matrix to resemble the whole-cut meat with the controllable intramuscular fat and collagen content. By using transdifferentiation strategies, the depositing ratio of fat into the cultured meat could be manually and precisely adjusted, allowing the nutrients to be naturally synthesized from the organized muscle structure. This demonstrates the potential for manipulating cultured meat to meet consumer preferences for specific fat content and texture. Results Chicken fibroblasts proliferate stably in low-serum conditions The chicken fibroblast cells were chosen as the ideal cell source for cultured meat production because they can propagate indefinitely and undergo myogenesis whenever the induction is provided. These cells can be readily obtained from fertilized eggs without the need to harvest animals. We have previously constructed an inducible myogenic transdifferentiation system in chicken fibroblasts with the stable integration of Tet-On-MyoD cassette ( Ren et al. , 2022 ). The MyoD is the key myogenic transcription factor, and the chicken fibroblasts could be converted into striated and elongated myotubes (myofibers) upon forced expression of MyoD ( Ren et al. , 2022 ; Weintraub et al. , 1989 ). The Tet-On-MyoD cassette enables the inducible and reversible activation of chicken MyoD factor, and in the current fibroblast cells, the myogenic transdifferentiation is only switched on by adding doxycycline (DOX). Without the DOX treatment, the control chicken fibroblast cells (Tet-On-MyoD) do not differ from the wild-type cells in terms of morphology, proliferation rate, and gene expression ( Figure 1—figure supplement 1A–C ). Hence, as a proof-of-concept experiment, we utilized this inducible myogenic fibroblast cell line to develop protocols for cultured chicken meat production. To achieve sustainable and cost-effective cell production, it is essential to minimize the serum usage in the culture medium. Therefore, we implemented a progressive serum reduction approach to acclimate the cells in the prospect of obtaining a stable cell source that can be propagated in low serum concentrations. In the initial experiment, we used 12% fetal bovine serum (FBS) in 1640 basal medium, which served as the control group. The results showed that the cells were able to proliferate normally at 6, 3, and 1. 5% serum concentrations, but at decreasing rates as shown by the EdU assay ( Figure 1A and B ). We also tested the CS as an alternative to bovine serum in order to avoid cross-species contamination of animal derivatives. The results showed that chicken fibroblasts could proliferate stably in the 1% CS ( Figure 1A and B, Figure 1—figure supplement 1D ) and the cell populations multiplied three times in 3 d as demonstrated by the CCK-8 assay ( Figure 1C ). Thus, we conclude that the low-serum medium can effectively support the stable propagation of chicken fibroblasts. Figure 1. Chicken fibroblasts proliferate stably in low-serum conditions. ( A ) Cellular morphology and EdU staining of chicken fibroblasts under different low-serum conditions. FBS: fetal bovine serum; CS: chicken serum. Scale bars, 200 µm. ( B ) Quantification of the proportion of EdU-positive cells in (A). Error bars indicate s. e. m. n = 3. *p<0. 05, **p<0. 01, ***p<0. 001. Paired t -test. ( C ) The CCK-8 cell proliferation assay showed the proliferation of chicken fibroblasts in 1% CS. Error bars indicate s. e. m, n = 3. Figure 1—figure supplement 1. Experimental scheme of myogenic transdifferentiation. ( A ) Scheme of the MyoD-induced transdifferentiation. ( B ) Morphology and EdU staining of chicken fibroblasts under different conditions from (A). Scale bars, 200 µm. ( C ) Quantification of the proportion of EdU-positive cells in (B). Error bars indicate s. e. m, n = 3. ns: not significant. Paired t -test. ( D ) Cellular morphology of chicken fibroblasts under different low-serum conditions. Orange triangles mark the sharper and smoother morphology of cell edge contours. Please note that the cells showed abnormal morphology in the lowest serums of 1% FBS and 0. 5% CS. FBS: fetal bovine serum; CS: chicken serum. Scale bar, 100 µm. 3D culture of chicken fibroblasts in GelMA hydrogels The behavior of cells grown on the top of 2D flat surface may differ from that of cells in 3D space. To simulate the 3D natural growth environment of cells in vivo, we utilized the gelatin methacrylate (GelMA)-based hydrogels as scaffolds for chicken fibroblasts. GelMA hydrogel can form a stable and porous structure for cell implantation and is commonly used in tissue engineering because of its great biocompatibility and mechanical tenability ( Pepelanova et al. , 2018 ). We created hydrogels with varying concentrations at 3, 5, 7, and 9wt%, and then observed their surface characteristics when immersed in culture medium using an emission scanning electron microscope. The porosity was measured to be 83. 27, 65. 01, 62. 47, and 57. 96%, respectively. Thus, the higher the mass concentration of hydrogels, the tighter the 3D mesh structure formed and the smaller the porosity ( Figure 2A ). The pore diameters in the scaffolds were determined to range from 3 µm 2 to 100 µm 2, showing the biggest pore size in 3% hydrogel and the smallest in 9% hydrogel ( Figure 2B ). Therefore, the porosity and pore size of the scaffold could be adjusted to achieve optimal physical strength and nutrients delivery that are suited for long-term cell culturation. Figure 2. 3D culture of chicken fibroblasts in gelatin methacrylate (GelMA) hydrogels. ( A ) Microscopic images of GelMA hydrogels at different concentrations (3, 5, 7, and 9 wt%) taken by scanning electron microscopy (SEM) and their corresponding simplified maps of pore distributions. Scale bar, 10 µm. ( B ) Quantification of pore area in (A). Error bars indicate s. e. m, n = 3. ****p<0. 0001. Paired t -test. ( C ) Brightfield and red fluorescent images of cells in 3D culture after PKH26 staining at different times (1 hr, 1 d, 3 d, 5 d, and 9 d). Scale bars, 100 µm. ( D ) Relative area of PKH26-linked cells in (C). Error bars indicate s. e. m, n = 3. *p<0. 05, ***p<0. 001, ****p<0. 0001. ( E ) Representative EdU staining shows the proliferation of cells in 3D culture on 1 d, 3 d, and 5 d after cell implantation in hydrogel. Scale bars, 100 µm. ( F ) Quantification of the proportion of EdU-positive cells in (E). Error bars indicate s. e. m, n = 3. **p<0. 01. Paired t -test. Figure 2—figure supplement 1. Cellular morphology of chicken fibroblasts cultured in 3D. Morphological changes of cells implanted and cultured on four different concentrations of hydrogels at 3, 5, 7, and 9 wt% for different times, and the 3 wt% hydrogel collapsed after the second day of growth. Scale bars, 100 µm. Figure 2—figure supplement 2. The labeling of cells with PKH26 and comparisons of cell morphology and proliferation between 2D and 3D. ( A ) Brightfield and red fluorescent images of chicken fibroblasts in 2D culture after PKH26 labeling. Scale bars, 100 µm. ( B ) Morphology of cells in gelatin methacrylate (GelMA) hydrogels cultured in 3D before and after dissociation. Scale bars, 100 µm. ( C ) Morphology and EdU staining of chicken fibroblasts under different conditions. 3D→2D indicates that cells isolated from 3D (B) were re-cultured in 2D. Scale bars, 200 µm. ( D ) Quantification of the proportion of EdU-positive cells in (C). Error bars indicate s. e. m, n = 3. ns: not significant. Paired t -test. To determine the most ideal gel concentration and pore size suited for chicken fibroblast attachment and growth, we conducted an experiment where we implanted the same number of cells into gels with varied pore sizes and examined the cellular dynamics over a 9-day period ( Figure 2—figure supplement 1 ). The cells adhered to the gels immediately and started to exhibit typical extended and irregular fibroblast cell morphology after 1 d, indicating the hydrogel’s good biological compatibility with chicken fibroblast cells. However, the 3% hydrogel collapsed in the medium 2 d after cell implantation and thus did not support the long-term cell growth. We continued to monitor the 3D cell proliferation and found that the cells successfully propagated in 5, 7, and 9% hydrogels for the entire duration of the experiment. Among the tested hydrogels, the 5% concentration was found to be the best option for cell attachment and growth as shown by the densest cellular structure formed, and thus be utilized for the subsequent analysis. Due to the challenges of observing and distinguishing the cells and pores within the hydrogel scaffold using the brightfield of ordinary light microscopy, we then used the PKH26 fluorescent cellular dye to accurately observe the cellular morphology and quantify the cell numbers. The cells were firstly labeled with PKH26 in 2D culture ( Figure 2—figure supplement 2A ) and then transferred into the gel matrix. Upon implantation, the cells appeared mostly round on the first day, but gradually extended and expanded within the hydrogel scaffold. By the fifth day, most of the cells were elongated with irregular shape and short tentacles and packed tightly, and by the ninth day they multiplied more than 15 times and formed dense fibrous bundles ( Figure 2C and D ). We also examined the proliferation dynamics of the cells in 3D scaffold via the EdU assay. The proliferation efficiency of chicken fibroblast cells gradually increased over time in 3D culture and reached replication levels even higher than those in 2D culture ( Figures 1B and 2E, F ). This result indicates that 3D culture conditions provide a more favorable environment for cell growth. Moreover, when the 3D cultured cells were detached from the gel by collagenase dissociation ( Figure 2—figure supplement 2B ) and seeded back in 2D monolayer culture plates, the cells again exhibited similar morphology but slightly increased proliferation capacity compared to the original fibroblasts in 2D conditions ( Figure 2—figure supplement 2C and D ). Thus, the chicken fibroblast cells cultured in 3D maintain their normal physiological characteristics and we continue to stimulate the cellular myogenesis and lipogenesis to prepare them for cultured meat production. Transdifferentiation of chicken fibroblasts into muscle cells in 3D Despite the complexity of the composition of fresh meat and processed products, muscle cells are the major and indispensable component of meat foods. The muscle cell-derived myotubes (myofibers) provide a rich source of proteins and nutrients and constitute the meat texture. To this end, we aimed to transform the chicken fibroblast into muscle cells through the established MyoD overexpression protocol ( Ren et al. , 2022 ). We first tested and optimized the myogenic transdifferentiation parameters in 2D culture through the DOX-induced MyoD expression ( Figure 3—figure supplement 1 ). We observed elongated multinucleated myotubes and abundant expression of myosin heavy chain (MHC) after 3 d of myogenic induction ( Figure 3—figure supplement 2A and B ). The ‘serum deprivation’ protocol was the classical strategy to stimulate terminal myogenic differentiation in myoblasts of various species including chicken ( Nakashima et al. , 2005 ). Thus, we examined the myogenic transdifferentiation efficiency (myotube fusion index) in the chicken fibroblast cells overexpressing MyoD in combination with reduced concentrations of bovine or horse serums. However, we found that reducing serum concentrations did not increase the myotube formation but instead caused massive cell death ( Figure 3—figure supplement 2B ). Therefore, it seems that the ‘serum deprivation’ could not further improve the myogenic transdifferentiation in chicken fibroblast cells and a simple MyoD overexpression strategy is sufficient for efficient production of mature muscle cells. After demonstrating the feasibility of induced transdifferentiation of chicken fibroblasts into muscle cells in 2D culture, we continued myogenic transdifferentiation in 3D to simulate the construction of concrete whole-cut meat. The cells were inoculated into a hydrogel scaffold and allowed to proliferate for 7 d before inducing transdifferentiation with MyoD overexpression ( Figure 3A ). We used whole-amount MHC immunofluorescence staining to examine the myotube formation directly inside the gel. We identified abundant multinucleate MHC + myotubes at multiple focal planes within the gel ( Figure 3B, Figure 3—video 1 ), indicating successful myogenic transdifferentiation of 3D cultured chicken fibroblasts. In contrast, no MHC signal or autofluorescence was detected in the 3D cultured chicken fibroblasts ( Figure 3—figure supplement 2C ) and hydrogel without cells ( Figure 3—figure supplement 2D ). With the assistance of confocal Z-stack analysis, the stacked images showed densely packed MHC + myotubes from a piece of cellular hydrogel complex at a depth of 68 μm ( Figure 3D ). The separate XY axis views of the orthogonal projections at different depths ( Figure 3E ) and a multiangle video ( Figure 3—video 2 ) also showed that several myotubes were aligned together. Nevertheless, many myotubes were oriented in different directions, preventing the entire matrix from aligning in one direction. In conclusion, we successfully transformed the chicken fibroblast cells into mature muscle cells in 3D environment. Figure 3. Transdifferentiation of chicken fibroblasts into muscle cells in 3D. ( A ) Experimental design for fibroblast myogenic transdifferentiation in 3D culture. ( B ) Representative images of myosin heavy chain (MHC) staining showed the myogenic ability of chicken fibroblasts in 3D culture. Scale bars, 50 µm. ( C ) Comparison of the mean myogenic fusion index between 2D and 3D. Error bars indicate s. e. m, n = 3. *p<0. 05. Paired t -test. ( D ) 3D images of MHC staining of cells cultured in 3D. The right panel is depth-coded image, which indicate different depths from the deepest (cyan) to the surface (yellow). ( E ) Orthogonal projections of three sets of MHC staining of cells in 3D culture at different depths. Scale bars, 50 µm. ( F ) Expression of skeletal muscle-related genes was determined by RT-qPCR in 2D and 3D cells upon myogenic transdifferentiation and control 3D cells without stimulation. Note that the myogenic transdifferentiation driven by MyoD stimulates the expression of classical myogenic factors. Error bars indicate s. e. m, n = 3. *p<0. 05, **p<0. 01, ***p<0. 001, ****p<0. 0001. Paired t -test. ( G ) Macroscopic morphology of the empty hydrogel matrix (left) and cultured meat (right). The cultured meat is the product obtained after 3D cell culture and induction of myogenesis. Scale bars, 1 cm. Figure 3—figure supplement 1. Expression of MyoD upon doxycycline (DOX) treatment. ( A ) The MyoD-3xFlag was fused in frame and under the control of Tet-On system. Representative immunofluorescence staining of MyoD in fibroblast_MyoD and fibroblast_Con. Scale bars, 50 μm. ( B ) Representative immunofluorescence staining of Flag in fibroblast_MyoD and fibroblast_Con. The anti-Flag immunostaining indicate the exogenous MyoD transgene expression. Scale bars, 50 μm. Figure 3—figure supplement 2. Myogenic transdifferentiation in 2D. ( A ) Experimental design for fibroblast myogenic transdifferentiation in 2D culture. ( B ) Myosin heavy chain (MHC) staining demonstrates the myogenic capacity of chicken fibroblasts. Note that horse serum (HS) treatment causes massive loss of cells. Scale bars, 100 µm. ( C ) Immunofluorescence staining of MHC in 3D cultured chicken fibroblasts without activation of MyoD factor as a negative control. Scale bar, 100 µm. ( D ) Immunofluorescence staining MHC in cell-free gelatin methacrylate (GelMA) hydrogels as the control. Scale bar, 200 µm. Figure 3—figure supplement 3. Cell gelatin methacrylate (GelMA) 3D culture units and macroscopic morphology after 7 d of culture, with a white plastic frame as the fixation ring. Scale bars, 1 cm. Figure 3—video 1. MHC + myotubes in the 3D cultured fibroblast after myogenic transdifferentiation. Scale bar, 50 μm. Figure 3—video 2. Multiangle video showing myotubes were aligned together. Compared to 2D, the transdifferentiated myotubes induced in 3D were more organized and densely packed, resembling the native myofiber distribution in vivo ( Figure 3B ). The myotube formation efficiency (fusion index) in 3D reached 30. 49%, which was significantly higher than that of the 2D under the same transdifferentiation conditions ( Figure 3C ). We also evaluated the expression of several myogenic markers by RT-qPCR ( Figure 3F ). MyoG (myogenin) is a transcription factor regulating terminal skeletal muscle differentiation and could be induced by MyoD ( Cao et al. , 2006 ). MYH6 (myosin heavy chain 6) is the major protein comprising the muscle thick filament and functions in muscle contraction ( van Rooij et al. , 2009 ). MCK (muscle creatine kinase) is an enzyme that is primarily active in mature skeletal and heart muscle ( Vincent et al. , 1993 ). All myogenic factors, including MyoG, MYH6, and MCK, significantly increased upon transdifferentiation in both 2D and 3D. In addition, the transdifferentiated cells exhibited significantly higher expression of MyoG and MCK in 3D conditions compared to that in 2D, indicating more robust myogenic differentiation and maturation of cells in the 3D microenvironment. We speculate that the porous structure of the hydrogel matrix may support the cells to grow in all directions, similar to the environment in which the myofibers form in vivo. Macroscopically, the muscle filaments and the dense cellular network structures formed by myogenic transdifferentiation could make the hydrogel matrix more compact and solid ( Figure 3G ). Thus, compared to the empty transparent scaffolds without cells, the hydrogels cultivated with muscle cells more closely resemble whole-cut meat, similar to fresh animal meat. Myogenic transdifferentiation of fibroblast does not produce myofibroblasts Fibroblasts can be induced to differentiate into myofibroblast upon injury, leading to tissue fibrosis. Similar to the skeletal muscle cells, the myofibroblast is also contractile and expresses the myogenic factor MyoD as well as certain types of sarcomeric MHCs ( Hecker et al. , 2011 ). To confirm that the myogenic transdifferentiated cells were indeed skeletal muscle cells but not the myofibroblasts, we utilized a panel of cell lineage-specific markers to delineate the cell conversion progress and determine the cellular identity of fibroblasts, myoblasts, and myofibroblasts, respectively. The immunofluorescence staining of 3D cultured cells showed that the muscle-specific intermediate filament Desmin ( Paulin and Li, 2004 ) was expressed only in the MyoD-transdifferentiated cells, but not in the original fibroblasts or the embryonic skin-derived myofibroblasts ( Figure 4A ). In contrast, the classical myofibroblast marker alpha-smooth muscle actin (α-SMA) ( Hinz et al. , 2007 ) was expressed only in myofibroblasts, but not in the fibroblast or the transdifferentiated cells ( Figure 4 ). The fibroblast intermediate filament Vimentin ( Tarbit et al. , 2019 ) was abundantly expressed in the fibroblasts but reduced in the myogenic transdifferentiated cells ( Figure 4C ). The 2D and 3D cultured cells showed consistent pattern of marker protein expression, indicating that the different culture models and conditions do not affect the cell identity conversion ( Figure 4—figure supplement 1A and B ). These results confirmed that the MyoD indeed transdifferentiate the cells toward the skeletal muscle lineage but not the myofibroblast. Furthermore, the RT-qPCR showed that, after myogenic induction in 3D, the skeletal muscle-specific genes Desmin and Six1 ( Relaix et al. , 2013 ) were significantly elevated ( Figure 4D ), whereas the fibroblast gene Thy-1 was significantly reduced ( Figure 4E ). The transforming growth factor β (TGFβ) signaling is the most potent known inducer of myofibroblast differentiation ( Vaughan et al. , 2000 ). We found that the expression of core TGFβ signaling components, TGFβ-1, TGFβ-3, and Smad3, remained unchanged during the transdifferentiation process ( Figure 4F ), indicating that the classical myofibroblast lineage was not induced. Together, these data confirm that the myogenic transdifferentiation of fibroblast does not produce myofibroblasts. Figure 4. Myogenic transdifferentiation of fibroblasts does not produce myofibroblasts. ( A ) Immunofluorescence staining of 3D cultured cells showed that the skeletal muscle marker Desmin was expressed only in the transdifferentiated cells but not in fibroblasts or myofibroblasts. Scale bars, 50 µm. ( B ) Immunofluorescence staining of 3D cultured cells showed that the myofibroblast marker alpha-smooth muscle actin (α-SMA) was expressed only in the myofibroblasts but not in fibroblasts or transdifferentiated cells. Scale bars, 50 µm. ( C ) Immunofluorescence staining of 3D cultured cells showed that the fibroblast marker Vimentin was abundantly expressed in fibroblasts and myofibroblasts but greatly reduced in transdifferentiated cells. Scale bars, 50 µm. ( D ) RT-qPCR showed that the myogenic genes Desmin and Six1 were significantly increased upon myogenic transdifferentiation. ( E ) RT-qPCR showed the fibroblast marker gene Thy-1 was significantly reduced upon myogenic transdifferentiation. ( F ) The myofibroblast marker genes TGFβ-1, TGFβ-3, and Smad3 remain unchanged during myogenic transdifferentiation. Error bars indicate s. e. m, n = 4. *p<0. 05, **p<0. 01, ***p<0. 001, ****p<0. 0001. ns: not significant. Paired t -test. Figure 4—figure supplement 1. Myogenic transdifferentiation of fibroblasts does not produce myofibroblasts in 2D culture. ( A ) Immunofluorescence staining of 2D cultured cells showed that the skeletal muscle marker Desmin was expressed only in the transdifferentiated cells but not in fibroblasts or myofibroblasts, and the myofibroblast marker alpha-smooth muscle actin (α-SMA) was expressed only in the myofibroblasts, but not in fibroblasts or transdifferentiated cells. Scale bars, 50 µm. ( B ) Immunofluorescence staining of 2D cultured cells showed that the fibroblast marker Vimentin was abundantly expressed in fibroblasts but greatly reduced in MyoD-transdifferentiated cells. Scale bars, 50 µm. Stimulate the fat deposition in chicken fibroblasts in 3D The intramuscular fat is a crucial component of meat that can determine its quality attributes, such as taste and flavor. The chicken fibroblasts have been reported to be amenable to lipogenesis through various stimuli, including CS, insulin, fatty acids, and retinoic acids ( Kim et al. , 2021 ; Kim et al. , 2020 ; Lee et al. , 2021 ). We first attempted different lipogenic stimulations on 2D cultured cells and stained them for Oil Red O to visualize and quantify fat deposition in the fibroblasts. Very few scattered Oil Red O signals were found in the treatments consisting of only serums (12. 5% FBS, 1% CS, or 2% CS), indicating no lipogenesis during normal proliferation conditions. However, when we added insulin and fatty acids (oleic/linoleic acid) to the medium, lipid droplets in the cells dramatically increased as detected by the Oil Red O staining. After extensive optimization of the concentrations of supplements in the medium, we identified that the 8 µg/ml fatty acids plus 60 µg/ml insulin (abbreviated as FI, F: fatty acids, I: insulin) can induce lipogenesis most efficiently in 2D chicken fibroblasts ( Figure 5—figure supplement 1A–C ). Next, the same lipogenesis induction strategy was applied to the 3D cultured cells ( Figure 5A ). We found extensive Oil Red O signal inside the hydrogel matrix at different focal planes and the lipid droplets were clearly visible as beaded strings under magnification ( Figure 5B and C, Figure 5—video 1 ). The RT-qPCR analysis illustrated the expression of genes involved in lipogenesis and triglyceride synthesis was significantly higher in the cells with lipogenic stimulation compared to the control ( Figure 5D ). Notably, the genes encoding for PPARγ (peroxisome proliferator-activated receptor gamma), Gpd1 (glycerol-3-phosphate dehydrogenase), and FABP4 (fatty acid binding protein 4) all showed higher expression in 3D than that in 2D. Interestingly, the expression of Znf423 (zinc finger protein 423), which is a PPARγ transcriptional activator ( Addison et al. , 2014 ; Longo et al. , 2018 ), only increased upon lipogenic induction in 3D but not in 2D conditions. It seems that the cells grown in 3D hydrogel showed enhanced lipogenesis compared to the flat surface cultured cells, similar to the myogenic transdifferentiation process in 3D. In addition to the lipogenic induction with fatty acids and insulin (FI), we also validated the use of CS in promoting lipid accumulation as previously reported ( Kim et al. , 2021 ). Our results showed that the CS alone could also stimulate fat accumulation effectively ( Figure 5—figure supplement 2A–C ). We further measured the triglyceride content and found that the lipogenic induction increased the triglyceride significantly in the cell matrix ( Figure 5E ). In conclusion, the 3D cultured chicken fibroblast can efficiently deposit lipids by different stimulations. Figure 5. Stimulation of fat deposition in chicken fibroblasts in 3D. ( A ) Experimental design for fibroblast lipogenesis in 3D culture (‘F’ is for fatty acids and ‘I’ is for insulin). ( B ) Representative images showing the Oil Red O staining of lipid content accumulated in cells at different focal planes at the same position. The control group was normal medium without lipogenesis. Scale bars, 100 µm. ( C ) Relative area of lipid droplets in (B). Error bars indicate s. e. m, n = 3. ****p<0. 0001. Paired t -test. ( D ) Expression of lipid synthesis-related genes determined by RT-qPCR in 2D and 3D cells upon lipogenic induction and control 3D cells without stimulation. Error bars indicate s. e. m, n = 3. *p<0. 05, **p<0. 01, ***p<0. 001, ****p<0. 0001. Paired t -test. ( E ) Triglyceride content in the cultured meat upon different lipogenic inductions and control 3D cells without stimulation. Error bars indicate s. e. m, n = 3. **p<0. 01. Paired t -test. Figure 5—figure supplement 1. Efficient lipogenesis in 2D chicken fibroblasts. ( A ) Experimental design of fibroblast lipogenic differentiation in 2D culture. ( B ) Oil Red O staining of lipids in 2D under different conditions. Scale bar, 100 µm. ( C ) Relative area of lipid droplets in (B). Error bars indicate s. e. m, n = 3. ***p<0. 001, ****p<0. 0001. Paired t -test. Figure 5—figure supplement 2. Chicken serum (CS)-induced lipogenesis in 3D cultured chicken fibroblast. ( A ) Experimental design for fibroblast lipogenic differentiation in 3D culture induced by CS. ( B ) Oil Red O staining of 3D culture of cells after lipogenic induction (10% CS) and representative images were taken consecutively at different focal planes in the same position. ‘1’, ‘2’, ‘3’, are magnifications of the corresponding areas. Scale bars, 100 µm. ( C ) Relative area of lipid droplets in (B). Error bars indicate s. e. m, n = 3. ****p<0. 0001. Paired t -test. Figure 5—video 1. Lipid droplets were observed in 3D cultured fibroblasts upon lipogenic induction. Scale bar, 100 μm. Controlled fat deposition in the transdifferentiated muscle cells in 3D hydrogel The above-presented data shows that chicken fibroblast cells have a superior capacity for transforming into muscle and depositing fat when cultured in a 3D hydrogel matrix. Next, we tried to combine the myogenic and lipogenic stimuli together to modulate the fat deposition in the cultured meat to simulate the various intramuscular fat contents in the conventionally raised meat. Rather than converting fibroblasts into muscle cells and fat cells separately and mixing them later, we adopted a new strategy that can induce de novo lipid deposition in the muscle by first inducing myogenic transdifferentiation and then followed by lipid induction in the same cells ( Figure 6A, Figure 6—figure supplement 1A ). In 2D conditions, plenty of MHC + myotubes and Oil Red O-stained lipids were found to intermingle after the myogenic/lipogenic treatment ( Figure 6—figure supplement 1B ), and some of the red marked lipid droplets were located inside the myotubes, indicating that the transformed muscle cells indeed deposit fat autonomously to constitute intramyocellular lipids ( Figure 6—figure supplement 1C ). Next, we applied similar treatment to the cells cultured in 3D hydrogel and also identified Oil Red O-labeled lipid droplets mixed with the densely packed MHC + myotubes ( Figure 6B and C ). These findings suggest that the use of myogenic/lipogenic treatments can induce the formation of muscle cells that are capable of depositing fat in both 2D and 3D. We further examined the expression levels of both myogenic and lipogenic factors in the 3D cultured cells by RT-qPCR. Compared to the control 3D cells without myogenic/lipogenic stimulations, the induced cells showed significantly higher expression of genes involved in both myogenesis and lipogenesis ( Figure 6D ). Interestingly, the extent of gene upregulation upon the combined myogenic/adipogenic stimulations was comparable to that of myogenic or adipogenic induction alone. This finding suggests that sequential myogenic transdifferentiation and lipid deposition do not interfere with each other when conducted in the same cells. Figure 6. Controlled fat deposition in the transdifferentiated muscle cells in 3D hydrogel. ( A ) Experimental design for fibroblast myogenic/lipogenic differentiation in 3D culture. ( B ) Representative images of myosin heavy chain (MHC) and Oil Red O staining of cells upon myogenesis/lipogenesis in 3D culture. Scale bars, 50 µm. ( C ) Orthogonal projections of three sets of MHC and Oil Red O staining of cells in 3D culture at different depths. Scale bars, 50 µm. ( D ) Expression of muscle-related genes (top) and lipid-related genes (bottom) in the cells with myogenesis/lipogenesis induction and control 3D cells without any stimulation were determined by RT-qPCR. ( E ) Triglyceride content of cultured meat under different conditions and real meat compare to fibroblasts_control. ‘Meat_leg’ and ‘Meat_breast’ were taken from the leg and breast muscles of adult chickens. Error bars indicate s. e. m, n = 3. *p<0. 05, **p<0. 01, ***p<0. 001, ****p<0. 0001. Paired t -test. Figure 6—figure supplement 1. Myogenic/lipogenic stimulation in chicken fibroblasts. ( A ) Experimental process for sequential myogenic/lipogenic stimulation in 2D culture. ( B ) Myosin heavy chain (MHC) staining and Oil Red O staining of cells with 2D induction of myogenesis/lipogenesis, with the triangular arrow indicating that Oil Red O-labeled lipid droplets are shown under the red fluorescent channel. Scale bars, 100 µm. ( C ) Immunofluorescence staining and Oil Red O staining demonstrating lipid deposition inside the transdifferentiated muscle cells, with the triangular arrow indicating the location of the lipids within the muscle cell. Scale bars, 100 µm. The intramuscular fat is an integral component of both traditional animal meat and cultured meat, and it directly influences the meat flavor and texture ( Frank et al. , 2016 ). Hence, we compared the triglyceride levels in the 3D hydrogel cells with different types of lipogenic stimuli with the chicken breast and leg meat. The results showed that the lipogenic stimulation in the 3D hydrogel cells increased the triglyceride content in the cultured meat to the levels comparable to or even higher than real chicken meat ( Figure 6E ). In contrast, the control cells without any induction or with only myogenic stimulation do not show apparent triglyceride accumulation ( Figure 6E ). Therefore, the fat content in the cultured meat could be synthesized in a controlled manner, and then we tried to purposely manipulate the triglyceride contents in the meat matrix by adjusting the potency of adipogenic stimulation. By fine-tuning the concentrations of insulin and fatty acids during lipogenic induction, the triglyceride contents in the final product of cultured meat can precisely reach any customized levels across the range from 1. 5 mg/g to 7 mg/g ( Figure 6E ), which overlap and surpass the levels in the fresh chicken breast and leg meat. As a result, this strategy greatly expands the diversity and category of cultured meat products, allowing for precise control over intramuscular fat contents to meet consumer preferences. Therefore, guided and graded fat deposition in cultured meat allows for the creation of various meat products with controlled intramuscular fat contents. The collagen content and extracellular matrix components of cultured meat Fibroblasts are an essential source of extracellular matrix (ECM), including the collagen, which provides elasticity to the tissue in the body and enriches the texture of the cultured meat ( Ben-Arye and Levenberg, 2019 ). In theory, the fibroblast should generate abundant ECM to produce a more realistic meat product. We then examined the collagen content in the cultured meat and found that the total collagen protein gradually increased and reached the plateau at 1. 59 µg/mg in the final product ( Figure 7A ). This is mainly due to the increased cell numbers and the accumulation of secreted collagen in the hydrogel matrix during myogenic/adipogenic transdifferentiation. Nevertheless, the RT-qPCR showed that the genes encoding the major components of ECM exhibited various expression patterns with the extension of culture time. The expression of COL1A1 (collagen, type I, alpha 1) and COL1A2 (collagen, type I, alpha 2) gradually decreased, whereas the fibronectin increased during the time course of meat synthesis. The expression of elastin and laminin genes remained stable throughout the whole course of the experiment ( Figure 7B ). However, the laminin protein content was accumulated and increased steadily during 3D culturation ( Figure 7C ). Overall, the synthesis and accumulation of different types and amounts of ECM components during the myogenic/lipogenic stimulations can improve the texture of the cultured meat prepared from fibroblast cells. Figure 7. The collagen content and expression of extracellular matrix (ECM) components in cultured meat. ( A ) Total collagen content of cultured meat at different days of cultivation. Error bars indicate s. e. m, n = 3. *p<0. 05, ***p<0. 001. Paired t -test. ( B ) Expression of ECM-related genes determined by RT-qPCR of cultured meat. Error bars indicate s. e. m, n = 3. *p<0. 05, ***p<0. 001, ****p<0. 0001. Paired t -test. ( C ) Representative Laminin staining of cells in 3D culture on 1 d, 3 d, 5 d, and 7 d after cell implantation and transdifferentiation in hydrogel. Scale bars, 100 µm. The characterization of molecular changes during myogenic transdifferentiation and fat deposition in cultured meat To provide insights into the functional shifts during the transdifferentiation from fibroblasts toward muscle, fat, or muscle/fat cells in 3D culture, we further analyzed the transcriptomes from the different populations of cells including the ‘original fibroblasts’ (3D_fibroblast), ‘myogenic transdifferentiated cells’ (3D_MyoD), ‘adipogenic transdifferentiated cells’ (3D+FI), and ‘myogenic/adipogenic transdifferentiated cells’ (3D_MyoD+FI) ( Figure 8A ). To illustrate the relationship between these cell groups, we conducted an unsupervised hierarchical clustering analysis of the whole transcriptome. The findings revealed that the 3D+FI group clustered distinctly from the others, while the 3D_MyoD and 3D_MyoD + FI groups exhibited greater similarity. Moreover, the 3D_fibroblasts formed a distinct sub-cluster on their own ( Figure 8B ), suggesting that myogenic or adipogenic transdifferentiation drives these cells away from their original fibroblastic state. The principal component analysis (PCA) of the transcriptomes also showed that distinct trajectories of myogenic and adipogenic transdifferentiation routes were derived from the original fibroblasts and finally integrated together into the myogenic/adipogenic cells (3D) ( Figure 8C ). It indicates that the myogenic and adipogenic signalings could operate simultaneously and separately during the generation of the culture meat composed of muscle and fat. We also compared the differentially expressed genes (DEGs) from ‘3D_MyoD vs 3D_fibroblast’, ‘3D+FI vs 3D_fibroblast’, and ‘3D_MyoD + FI vs 3D_fibroblast’. The results showed that the majority (78%) of DEGs in the 3D_MyoD+FI are overlapped with 3D_MyoD and 3D+FI, indicating that sequential myogenic/adipogenic induction in 3D_MyoD+FI is consistent with myogenic or adipogenic function individually ( Figure 8D ). The heat map also highlighted the representative myogenic or adipogenic genes that were upregulated in the myogenic, adipogenic, or myogenic/adipogenic cells constitute the culture meat. In contrast, the fibroblast genes were diminished during the transdifferentiation, confirming the loss of fibroblast identity ( Figure 8E ). In addition, the Gene Ontology (GO) analysis of upregulated DEGs in 3D_MyoD+FI cells confirmed that myogenic-specific pathways such as ‘muscle organ development’ and adipogenic-specific pathways such as ‘PPAR signaling pathway’ were enriched. In addition, we also identified several multifunctional signaling pathways such as ‘JAK-STAT signaling pathway’, ‘NF-kappa B signaling pathway’, and ‘MAPK signaling pathway’ that were simultaneously activated during myogenic/adipogenic transdifferentiation, which should have profound effects on both myogenesis ( Bakkar and Guttridge, 2010 ; Jang and Baik, 2013 ; Keren et al. , 2006 ) and adipogenesis ( Batista et al. , 2012 ; Bost et al. , 2005 ; Richard and Stephens, 2011 ; Figure 8F ). The upregulated genes in the representative pathways (such as ‘ Stat1’ in JAK-STAT signaling pathway, ‘ Mapkapk3’ in MAPK signaling pathway) are shown in Figure 8G. In conclusion, the transcriptome analysis of the different types of transdifferentiated cells revealed important molecular mechanisms including not only the myogenic- and adipogenic-specific pathways driving the muscle formation and fat deposition respectively, but also several key multifunctional signaling pathways that can promote the cell fate transition and differentiation in different cellular contexts including the muscle and fat tissues. Figure 8. Gene expression profiles during transdifferentiation and fat deposition in 3D. ( A ) Scheme of the RNA-seq samples marked by different colors. ( B ) Hierarchical clustering analysis of whole transcriptomes of 3D_fibroblasts, 3D_MyoD, 3D+FI, and 3D_MyoD+FI using Euclidean distance with ward. D cluster method. ( C ) Principal component analysis (PCA) of transcriptome changes during myogenic transdifferentiation and fat deposition (n = 10, 247 genes). The ellipses group includes three biological replicates in each cell type. The arrows represent the reprogramming of gene expression under different conditions. The routes were derived from the original fibroblast toward two differentiation routes, namely ‘myogenic transdifferentiation’ and ‘adipogenic transdifferentiation’. ( D ) Venn diagram showing the overlap of differentially expressed genes (DEGs) from 3D_MyoD, 3D+FI, and 3D_MyoD+FI compared to the original 3D_fibroblasts. ( E ) Heat map showing the representative genes differentially expressed between 3D_MyoD+FI and 3D_fibroblast cells (n = 3 biologically independent samples). ( F ) Gene Ontology (GO) analysis of upregulated DEGs between 3D_MyoD+FI vs. 3D_fibroblast cells. ( G ) GOChord analysis of the upregulated genes within representative pathway between 3D_MyoD+FI and 3D_fibroblast cells. Discussion Mature muscle tissue primarily consists of myofibers (myotubes), which are long, multinucleated cells that contract to generate force and movement. In addition to myofibers, muscle tissue contains a variety of other cell types, including fibroblasts and adipocytes, all of which play important roles in the structure and function of the tissue. The ECM, which is mainly secreted by fibroblast cells, provides support and structural integrity to the muscle tissue. It is made up of a complex network of proteins and molecules, such as collagen, that provide a scaffold for the cells to attach to and interact with ( Franco-Barraza et al. , 2016 ). Together, these components contribute to the unique features of the skeletal muscle tissue and the fresh meat, including its strength, flexibility, and elasticity. It has been and still is challenging to recreate those characteristics in an in vitro condition of cultured meat production. One of the main obstacles in this process is the co-culturing of different cell types with distinct properties. In this study, we overcame this limitation of co-culturing different types of cells by utilizing a single-source cell to generate various meat components, including muscle, fat, and collagen. Precisely, we employed chicken fibroblasts to produce muscle, deposit fat, and synthesize collagen in a well-controlled and adjustable manner within a 3D setting to produce meat with desirable characteristics. Fibroblasts are one of the most common cell types in animals and could serve as the seed cells for cultured meat production due to their unique and versatile features. First, the fibroblast cells are widely available in the bodies of agricultural animals and could be easily isolated and cultured in vitro. Second, many groups have successfully transformed the chicken fibroblasts into immortalized cell lines ( Himly et al. , 1998 ; Pasitka et al. , 2023 ). These cell lines can provide an unlimited cellular resource for cultured meat production and eliminate the need for animal or embryo harvest. Third, fibroblasts are capable of adapting to low serum concentration medium or even serum-free medium, which greatly reduce the culturing cost and risk of serum-bound pathogens ( Genbacev et al. , 2005 ; Lohr et al. , 2009 ; Pasitka et al. , 2023 ). Lastly, the fibroblast cells can also undergo transformation to enable high-density propagation in suspension culture ( Bürgin et al. , 2020 ; Fluri et al. , 2012 ; Shittu et al. , 2016 ), which is a crucial step toward scaling up to mass production. The Food and Drug Administration (FDA) has already approved the use of chicken fibroblast cells for cultured meat production, and recently, Eat Just Inc successfully utilized chicken fibroblasts for the commercial production and sale of cultured meat in Singapore and the USA (FDA, 2023). One latest study has also demonstrated the feasibility and consumer acceptance of cultured meat derived solely from native chicken fibroblast cells, indicating a very promising development. However, the resulting meat products do not seem to contain any muscle components ( Pasitka et al. , 2023 ). We have previously established an effective strategy for myogenic transdifferentiation, allowing for the production of muscle cells from fibroblast cells of various species including chicken and pig in 2D culture ( Ren et al. , 2022 ). In the present study, we further enhanced the myogenic transdifferentiation process in 3D and simultaneously simulated the fat deposition to create cell-based meat that more closely resembles real meat. As a proof of concept, we utilized the transgene method to achieve maximum myogenic induction and the final products still retain the foreign transgene fragment in the cells’ genome. It is therefore posing a risk of genetic modified food that is not suitable for mass production. In the next step, other non-transgenic means such as non-integrating vectors, chemical reprogramming, modified RNAs, and recombinant transgene removal techniques will be explored to develop transgene-free end products. Another food safety concern in this study is the use of GelMA hydrogel for culture meat production. Due to its excellent biocompatibility and mechanical flexibility, GelMA-based hydrogel has demonstrated significant potential in scalable 3D cell culture for creating artificial tissue ranging in sizes from millimeters to centimeters. It is widely used in 3D cell culture and tissue engineering for regenerative medicine, but less common in food processing and agricultural applications. Due to its special photo-crosslinking properties, biocompatibility, and degradability, it allows this material to be shaped into complex tissue structures by 3D printing or modeling. Many researchers have also used GelMA hydrogel as a scaffold for culture meat production ( Jeong et al. , 2022 ; Zheng et al. , 2021 ). Later research will carefully consider hydrogel as well as other types of scaffold biomaterials for cost-effective and food-safety-compliant culture meat production ( Bomkamp et al. , 2022 ). Numerous studies have identified the crucial role of fat in the aroma, juiciness, and tenderness of meat. In general, a very low level of intramuscular fat results in dry meat with a plain taste, whereas the high intramuscular fat contents can improve the cooking flavor and greatly increase the value of meat products, such as the high marbling beef from Japanese Wagyu cattle ( Gotoh et al. , 2018 ; Motoyama et al. , 2016 ). The fat content of fresh meat mainly comes from lipids contained in the fat cells. Thus, closely mimicking the intramuscular fat properties in cultured meat would require co-culturing of muscle cells with fat cells. For example, co-culturing pre-adipocytes with myoblasts may increase the intramuscular fat content, tenderness, and taste intensity of cultured meat ( Lau et al. , 1996 ; Pandurangan and Kim, 2015 ; Zagury et al. , 2022 ). However, co-culture of different cell types is technically challenging since each cell type grows and differentiates in the specific optimized medium. When different cell types are cultured in the same medium, these culture conditions may be suboptimal for one cell type or the other and result in inefficient cellular growth ( Pallaoro et al. , 2023 ). Previous studies have underlined the influence that adipocytes growing near muscle cells can impair myogenesis ( Seo et al. , 2019 ; Takegahara et al. , 2014 ). Simultaneous or sequential induction of both myogenic and lipogenic differentiation in the same starting seed cells would resolve these co-culture conflicts, and the multi-lineage competent chicken fibroblast cells were chosen to explore this double transdifferentiation strategy. As a proof of concept, we successfully transformed chicken fibroblast cells into muscle cells and deposited fat into the same cells in 3D hydrogel matrix. Notably, the intramuscular fat content in the cultured meat could be tailored to any specific level within a certain range. From a nutritional point of view, a direct comparison with the traditional chicken meat was performed. The triglyceride content in the cultured meat is comparable to that of chicken meat, and, more importantly, the amount of fat could be easily manipulated in order to achieve a more attractive nutritional profile ( Figure 9 ). Figure 9. Model for myogenic and lipogenic transdifferentiation of chicken fibroblasts in 3D culture to produce meat with precisely controlled levels of intramuscular fat and extracellular matrix. In this study, the deposition of fat in the myotubes/myofibers facilitated the storage of significant lipid quantities in transdifferentiated muscle cells, known as intramyocellular lipids. Additionally, we observed Oil Red O staining in the remaining un-transdifferentiated fibroblasts, resembling cells of intramuscular adipocytes (extramyocellular lipids) found within muscle tissue. Hence, current adipogenic induction treatment caused lipogenesis in both the MyoD-transdifferentiated cells and un-transdifferentiated fibroblasts. At present, we do not know whether the fatty acids profile would be different between the two types of origins. Unsaturated fats are conventionally regarded as 'healthier' than saturated fats ( Berglund et al. , 2007 ; Hamley, 2017 ). One of the superior assets of cultured meat production is the ability to increase the unsaturated fat levels by adjusting the additive fatty acids, oleic/linoleic acids in this case, in the culture medium. In the future, further comprehensive analysis of the fatty acids profile in the cultured meat is required to elucidate the fat-associated attributes similar or distinct to real meat. The current transcriptome analysis during the cellular transdifferentiation also revealed that key regulatory signaling pathways control the formation of cultured meat derived from fibroblasts. Notably, we found that several multifunctional pathways such as JAK-STAT and MAPK signaling pathways were significantly enriched in the final cells that underwent double myogenic/adipogenic transdifferentiation. These signalings could drive the differentiation process of many different types of cells, including muscle and adipose tissues ( Bost et al. , 2005 ; Jang and Baik, 2013 ; Keren et al. , 2006 ; Richard and Stephens, 2011 ), and may be manipulated in more settings to further improve the generation of final cultured meat with tunable fat content. In conclusion, we have effectively utilized immortalized chicken fibroblasts in conjunction with classical myogenic/adipogenic transdifferentiation approaches within the 3D hydrogel to establish a cultured meat model. This model allows for the precise regulation of the synthesis of key components of conventional meat, including muscle, fat, and ECM. This approach can be readily extrapolated to other species such as pigs and cows, and presents promising avenues for the large-scale production of customized and versatile meat products that may cater to varying consumer preferences. Materials and methods Cell preparation and inducible myogenic transdifferentiation The cellular transdifferentiation was constructed as described previously ( Ren et al. , 2022 ; Ren et al. , 2023 ). Briefly, we cloned the chicken full MyoD coding sequence fused in-frame with 3xFlag into a DOX-inducible lentiviral system (Tet-On-MyoD). The wild-type chicken fibroblasts were infected with lentivirus and subjected to puromycin selection, and finally obtained the transdifferentiation cell lines. In addition, myofibroblasts were isolated from the skin of 10-day-old chicken embryos in the same way as in the previous study ( Kosla et al. , 2013 ). Cells were cultured in 1640 basal medium (Gibco, #C11875500BT) supplemented with 12% FBS (CELLiGENT, #CG0430A) and 1% penicillin-streptomycin (Gibco, #11140050) at 39°C under 5% CO 2 atmosphere and were given fresh medium every 2 d. When grown to approximately 80% confluence, the cells were trypsinized and passaged. The study was approved by the Animal Care and Use Committee of Shandong Agricultural University. Domestication of cells in low-concentration serum medium The chicken fibroblast cells were domesticated with the progressive reduced concentrations of serum. In general, cells were cultured with 12% FBS in 1640 basal medium, and when grown to about 80% confluence, the medium was replaced with a 6% FBS medium for further cell culture and passage. The medium was then changed to 3% FBS medium for further cell culture and passage depending on the cell status, and so on. This was started again and the above steps were repeated as soon as the cells grew badly or died. We directly changed the cell culture medium from 12% FBS to 2% CS then reduced it to 1% CS when using CS (Solarbio, #S9080) medium and the cells could adapt to the medium with low concentration of CS after 3–5 passages. Preparation of 3D scaffold and cell culture in 3D matrix GelMA hydrogels were purchased from Beijing ShangPu for this experiment. Appropriate amount of GelMA powder was weighed, dissolved in 1640 basal medium on a water bath at 70°C for about 30 min, and then filtered with 0. 22 μm sieve. Then, 1/8 volume of lithium acylphosphinate salt photoinitiator was added to the dissolution solution to obtain GelMA hydrogel solution, then stored at 37°C until usage but no more than 24 hr. Fibroblast cells were obtained by trypsin treatment and suspended with GelMA hydrogel solution and gently mixed, followed by treatment of 405 nm UV light for 10–20 s to get a 3D hydrogel scaffold. In addition, the hydrogel was secured on a fixation ring for better cell growth and easy movement ( Figure 3—figure supplement 3 ). The cell hydrogel complexes were placed in 24-well plate and culture with 1640 medium containing 12% FBS. Then, they were transferred to a new well after 12 hr, and fresh medium was added and changed every 24 hr. The cell hydrogel complexes were gently washed three times with PBS buffer and digested by collagenase II enzyme (Worthington, LS004177) in an incubator at 39°C for 6 min until the cells in the hydrogel package became round and shed. The digest was terminated by adding medium and then centrifuged at 1000 × g for 10 min at room temperature (25 °C). Cell Counting Kit-8 assay Cells cultured in 1% CS and 12% FBS were seeded into 96-well plates with 100 μl of medium per well and incubated for 0 hr, 24 hr, 48 hr, and 72 hr. 10 μl of the Cell Counting Kit-8 assay (CCK-8) solution (Solarbio, #CA1210, China) was then added to each well and incubated for 2 hr. The absorbance values at 450 nm were measured with an EnSpire multifunctional spectrophotometer (PerkinElmer, USA). EdU assay 2D cells were cultured in 6-well plates, and 1 ml of the growth medium was added to each well with 0. 25 μl of 1. 25 mg/ml 5-ethynyl-2'-deoxyuridine (EdU) (Beyotime, #ST067-1g). After 30 min of incubation, cells were fixed by 4% paraformaldehyde (PFA) for 30 min. The cells were stained with a prepared reaction solution consisting of 1 mmol/L CuSO 4, 100 mmol/L Tris–HCl, 100 mmol/L ascorbic acid, 1:1000 Alexa Fluor 555 Azide as previous described ( Zhang et al. , 2022 ). After 30 min, the cells were washed with PBS, and the nuclei were stained with 50 ng/ml DAPI for 10 min. 3D cultures of cells were cultured in 24-well plate by adding 1 ml of the medium with 0. 25 μl EdU, incubated for 1 hr, and then fixed with 4% PFA for 24 hr. Staining time was extended to 1 hr and nuclear staining to 20 min. Fluorescence images were collected using fluorescence microscopy. Cell differentiation For fibroblast induction into myoblasts, 50 ng/ml DOX (Sigma, #D3000000) in 1640 basal medium containing 12% FBS was added for 3 d, and the differentiation medium was replaced when the cells reached 80% confluence. In addition, in 3D culture, cells were proliferated for 7 d before changing the differentiation medium when a dense arrangement of cells can be observed under the microscope. For induction of lipogenic differentiation, the differentiation medium was changed when the fusion rate of cells reached 80% in 2D culture or proliferated for 5–7 d in 3D culture. The lipogenesis was induced for 48 hr with fresh medium changes every 24 hr. For the differentiation experiments, 1640 with 12% FBS was used as a control, and the lipogenic medium was consistent with a 1:100 fatty acid ('F' for short) composition of 1:1 oleic acid (Sigma, #O3008, 2 mol oleic acid/mole albumin) and linoleic acids (Sigma, #L9530, 2 mol linoleic acid/mole albumin; 100 mg/ml albumin), and insulin ('I' for short) (Sigma, #I0516) concentration of 60 μg/ml. Immunofluorescence staining Similar to our previous steps for immunofluorescence staining of cells in 2D culture ( Luo et al. , 2022 ), cells were fixed in 4% PFA for 30 min and washed three times with PBS buffer at room temperature, then permeabilized with 0. 5% Triton X-100 in PBS for 10 min and blocked with 10% goat serum at 0. 5% TritonX-100 for 1 hr. Primary antibodies were diluted with 10% goat serum in 0. 5% Triton X-100 at 4°C for 12 hr. The primary antibodies for MHC (DSHB, #AB2147781), Desmin (Sigma, #D8281) and laminin (Sigma, #F1804) were added at a dilution of 1:500, and the primary antibodies for Vimentin (DSHB, #AB528504) was added at a dilution of 1:200. This was followed by incubation with Alexa secondary antibody (Invitrogen, #A-21202, #A-21206, and #A-31570) at a 1:500 dilution for 2 hr at room temperature. The antibody for α-SMA (Sigma, C6198) was added at a dilution of 1:500 and incubated with 10% goat serum in 0. 5% Triton X-100 at room temperature for 2 hr. Nuclear staining was performed with 50 ng/ml DAPI (Sigma, #D8417) for 10 min. Cells in the 3D culture were fixed for 48 hr, the total time of permeabilizing and the blocking was no more than 24 hr, and the incubation time of the primary antibody was extended to 24 hr, the secondary antibody to 4 hr, and the nuclear staining to 20 min. For MHC detection, fluorescence images were collected using fluorescence microscopy, and the MHC + DAPI + /DAPI + differentiation index was calculated from three or more images using confocal microscopy (Zeiss LSM 800). Oil Red O staining For Oil Red O staining, cells were fixed with 4% PFA for 30 min or 24 hr, respectively, in 2D and 3D culture. After washing thrice with PBS at room temperature, the cells were soaked in 60% isopropanol for better coloration of Oil Red and then washed for 5 min or 10 min, respectively, in 2D and 3D culture. The cells were stained with Oil Red O (Sigma, #O0625) for 30 min or 60 min, respectively, in 2D and 3D. The cells were then washed with 60% isopropanol for 30 s or 1 min, respectively, in 2D and 3D to remove surface staining. Then, the cells were washed with distilled water three times and the stained lipid droplets were visualized using a microscope. mRNA extraction and RT-qPCR Cells were lysed in Trizol (Simgen, #5301100) and RNA was extracted following the manufacturer’s recommendations. RNA concentration was measured on NanoDrop2000 (Thermo Scientific, USA). 1 μg of RNA was reverse-transcribed using PrimeScript RT reagent kit (Takara, #RR047A). Real-time quantitative PCR was performed using SYBR Green Mix (Abclonal, #RK21203) following the manufacturer’s instructions. Expression was normalized to GAPDH using delta-delta-CT method. For comparisons of the expression, we used a one-tailed Student’s t -test. The error bars indicate the SEM. The RT-qPCR primers are described in Table 1. Table 1. List of primers of qPCR. Gene Forward primer Reverse primer Gapdh TCGGAGTCAACGGATTTGGC ATAGTGATGGCGTGCCCATT MyoD ACTACAGCGGGGAGTCAGAT GCTTCAGCTGGAGGCAGTAT MyoG AGCCTTCGAGGCTCTGAAAC AAACTCCAGCTGGGTGCTC Myh15 AGATAAAGGAACTACAGGCTCGT CGCCAGCTTCAGGAACTCA CKM ACCTGGACCCCAAATACGTG TCGAACAGGAAGTGGTCGTC Desmin GGAGATCGCCTTCCTCAAGA CAGGTCGGACACCTTGGATT Six1 ACTGCTTCAAGGAGAAGTCG TTCTCCGTGTTCTCCCTCTC Thy-1 TGTCATCCTGACAGTGCTGC GGTAGAGGCACACCAGGTTC TGFβ–1 GAGCTGTACCAGGGTTACG GAAGCCTTCGATGGAGATG TGFβ–3 CTCCCCGAGCACAATGAGT TATATGCTCATCTGGCCGCA Smad3 GCAAGATCCCACCAGGATG GAGGTGCAGCTCAATCCAG Pparg TGCCAAGCATTTGTAT TGCGAATTGCTACTTCTTTGTT Znf423 CCAGTGCCCACAGAAGTTCT CCACTGTGCCACCATCAAGT Fabp4 CAAGCTGGGTGAAGAGTTTGATG TCGTAAACTCTTTTGCTGGTAAC Gpd1 GGCTTTTGCCAAGACTGGGAA GGTTTGCCCTCATAGCAGATCTG Collagen I α1 GTCCTGCTGGATTTGCTGG GAAACCAGTAGCACCAGGG Collagen I α2 TGATCCATCTAAAGCGGCTG TTTGCCAGGGTGACCATCTT Laminin CGCGATTTCTGATTTTGCCG CATTGCAGTCACAAGGCAAG Fibronectin GTGCTACGACGATGGGAAAA GCAGTTGACGTTGGTGTTTG Elastin CTACTGGGACAGGTGTTGGA CACCATAGGCTCCTGCCTT Transcriptome analysis The RNA-seq library-preparation protocol was based on the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, #E7530L). Insert size was assessed using the Agilent Bioanalyzer 2100 system and qualified insert size was accurate quantification using StepOnePlus Real-Time PCR System (Library valid concentration＞10 nM) and then paired-end sequencing using an Illumina platform. RNA-seq was performed on triplicates for each sample. Sequencing data have been deposited in SRA database under accession code PRJNA1102033. RNA-seq data analysis The RNA-seq raw data were first trimmed adapters by trim_galore software ( Bolger et al. , 2014 ) and then the clean data of RNA-seq was mapped to the chicken genome (Ensembl, GRCg7b) using hisat2 ( Kim et al. , 2019 ) with default parameters. Because of using paired-end reads, the concordant unique mapping reads/pairs were kept based on the mapping flags. The duplications were removed based on the coordinates of the reads/pairs. The de-duplication unique mapping reads/pairs were used for further analysis in this study. The read counts for each sample were computed with the featureCounts ( Liao et al. , 2014 ) software, and the RefSeq gene annotation for chicken genome assembly is GRCg7b. The transcript per million (TPM) was normalized using the read counts. DEGs of RNA-seq data were analyzed using DESeq2 called by q<0. 01 and fold change >2 thresholds. The chicken genes were transformed to homolog human genes using the ensemble bioMart database, the GO in this study was conducted in Metascape, and the enriched top pathways are shown. All plots were generated with R (v4. 0. 3). Measurement of total collagen content in cultured meat Total collagen content was determined by the concentration of hydroxyproline. The medium was replaced with fresh medium in the incubator for 2 hr before the assay and washed three times with PBS. According to the instructions of the Hydroxyproline Assay Kit (Jiancheng, #A030-1-1, China), 1 volume of saline was added to dried cultured meat at a 1:1 ratio of weight and volume. After mechanical homogenization in an ice-water bath, the digestion solution was incubated at 37°C in a water bath for 4 hr. Hydroxyproline content was measured by the absorbance at 550 nm using an EnSpire multifunctional spectrophotometer (PerkinElmer). For each measurement of cultured meat, the hydroxyproline content of the corresponding blank cell-free hydrogel was subtracted. The amount of collagen was calculated from the hydroxyproline concentration with a conversion factor of 7. 25 in μg/mg wet tissue ( Vasanthi et al. , 2007 ; Zheng et al. , 2021 ). Measurement of triglyceride content in cultured meat The triglyceride content in cultured meat was measured using the kit (Solarbio, #BC0620, China). Before the assay, the culture system was washed three times with PBS and uniformly added to the same medium containing 12% FBS in an incubator for 2 hr. It was then removed and washed three more times with PBS. The cultured meat was taken out and churned, and the precipitate is obtained by centrifugation, which should be well air-dried. The precipitate was weighed and Trizol was added to lyse cell for 2 hr. A mixture of n-heptane and isopropanol at a ratio of 1:1 was then added, shaken, and mixed. Then, potassium hydroxide was added and fullly shaken to produce glycerol and fatty acids, and the other reagents were added in sequence according to the instructions. The final triglyceride content was measured by the specific light intensity at 420 nm, as described previously. Emission scanning electron microscope The cell hydrogels were washed three times with PBS and then fixed in 4% PFA for 1 hr. The samples were randomly clamped into spiking trays, rapidly frozen in liquid nitrogen, and then the images were randomly collected using the emission scanning electron microscope (Hitachi SU8010, Japan). Statistical analyses Statistical analyses were performed using the GraphPad Prism software. For normally distributed data sets with equal variances, a two-sample t -test was used. The significance of differences is provided in the figure legends.

10. 7555/JBR. 27. 20130001

Journal of Biomedical Research

Physical and degradation properties of PLGA scaffolds fabricated by salt fusion technique

Tissue engineering scaffolds require a controlled pore size and interconnected pore structures to support the host tissue growth. In the present study, three dimensional (3D) hybrid scaffolds of poly lactic acid (PLA) and poly glycolic acid (PGA) were fabricated using solvent casting/particulate leaching. In this case, partially fused NaCl particles were used as porogen (200-300µ) to improve the overall porosity (≥90%) and internal texture of scaffolds. Differential scanning calorimeter (DSC) analysis of these porous scaffolds revealed a gradual reduction in glass transition temperature (Tg) (from 48°C to 42. 5°C) with increase in hydrophilic PGA content. The potential applications of these scaffolds as implants were further tested for their biocompatibility and biodegradability in four simulated body fluid (SBF) types in vitro. Whereas, simulated body fluid (SBF) Type1 with the optimal amount of HCO 3 − ions was found to be more appropriate and sensible for testing the bioactivity of scaffolds. Among three combinations of polymer scaffolds, sample B with a ratio of 75:25 of PLA: PGA showed greater stability in body fluids (pH 7. 2) with an optimum degradation rate (9% to 12% approx). X-ray diffractogram also confirmed a thin layer of hydroxyapatite deposition over sample B with all SBF types in vitro.

INTRODUCTION The core idea of tissue engineering is to allow the cells to repair and regenerate damaged tissues and organs by promoting cell growth and differentiation over the scaffolds. These scaffolds are biodegradable matrices designed to support cell proliferation, which finally provides a functional tissue. Since scaffolds are temporary matrices, the degradation performance of the scaffolds must correspond to the regeneration rate of the affected tissues. In this regard, the selection of scaffold material is very important to facilitate the cells to behave in the desired manner to generate tissues or organs of our requirement. The materials used to synthesize biodegradable scaffolds for bone tissue engineering applications ranges between inorganic materials such as ceramics to synthetic polymers. Among them, synthetic polymers have the ability to tailor mechanical properties and degradation kinetics of scaffolds to suit various tissue engineering applications. The intrinsic properties of the polymer materials play a strategic role in the morphology, texture and performances of the scaffold. To be an artificial scaffold, the structure and the surface morphology of the scaffolds have to meet general requirements specific for the targeted tissue: i) three-dimensional architecture; ii) interconnected pores to ensure cell growth, diffusion of nutrients and metabolic waste; iii) suitable surface chemistry; iv) suitable mechanical properties; v) controllable biodegradation and bioresorbability. In the current study, three combinations of hybrid scaffolds were prepared by blending polylactic acid (PLA) and polyglycolic acid (PGA) at the ratios of 80:20, 75:25 and 70:30 and these polymers are known for their cell-based tissue engineering approaches. These polymers have been shown to be degraded mainly by hydrolysis of ester bonds into acidic monomers, which can be removed from the body by physiological metabolic pathways, ( Fig. 1 ). The process which we adopted for preparation of microporous biodegradable scaffolds was solvent casting/particle leaching, where we used non-dispersed sodium chloride (NaCl) as particulate porogen for improved pore interconnectivity. In this case, pores were interconnected via fused salt particles prior to the synthesis of three-dimensional (3D) polymer scaffolds. Thus, dissolution of this fused porogen matrix leaves a highly interconnected pore structure in the polymer scaffolds. Once the polymer scaffolds are made, there is an immediate need to test the scaffolds for both in vivo and in vitro in order to consider them for human applications. These studies include their physical, chemical and mechanical properties helpful for assessing their bioavailability. In case of in vitro studies, scaffolds were exposed to a group of model solutions simulating the inorganic portions of blood plasma to study their surface interaction and changes. The composition of the most used simulated body fluids differs from that of human blood plasma by high content of Cl- and lower content of HCO 3 − ions. Considering the composition of bone like apatite, which contains carbonate ions, the test results could be influenced by this difference. In this study, we prepared four different simulated body fluids with varied concentrations of the above said ions and we monitored the influence of these simulated fluids on the physico-mechanical properties of polymer scaffolds. Fig. 1 Structure of PLGA molecule. The figure shows hydrolytic degradation to PLA and PGA monomers in the presence of physiological fluids. PLGA: poly (lactic co glycolic acid). MATERIALS AND METHODS Scaffold preparation PLA (2. 9 kg/mol) and poly glycolic acid (PGA) (IV= 1. 2 dL/g) were procured from Sigma (St. Louis, MO, USA). Porous polymer scaffolds were prepared by solvent casting/particulate leaching where we used NaCl salt as particulate porogen (200-300 µm). PLA and PGA with higher molecular weight/inherent viscosity were used in our studies to ensure that the scaffolds would hold adequate mechanical integrity despite their relatively high porosity (≥90%). Briefly, NaCl matrices were prepared by subjecting NaCl particles to 95% humidity for 12 hours prior to solvent casting. PLA and PGA were blended at the ratios 70:30, 75:25 and 80:20 and dissolved in Hexafluoro-2-propanol (Sigma). This molten polymer blends were poured in to non-dispersed NaCl matrices (or) scaffold before solvent evaporation. Then, these scaffolds were vacuum dried for 48 hours before NaCl particles were further leached out by immersing scaffolds in de-ionized water. Scaffold characterization Electron microscopy The transverse sections of NaCl scaffolds were imaged using scanning electron microscope (SEM) prior to solvent casting to monitor NaCl crystal fusion. In addition, transverse sections of polymer scaffolds after salt leaching were also imaged using SEM (Zeiss EVO® MA15). Determination of glass transition temperature (Tg) Glass transition temperature (Tg) of all three scaffold types was determined as per ASTM D7426 standard by differential scanning calorimeter (DSC) equipped with liquid nitrogen cooling system (Auto Q20, TA instruments). Ten mg of polymer samples were quantitatively transferred to sealed aluminum pans and subjected to cooling and heating cycles from 0°C to +200°C with cooling and heating rates of 5°C/min. During experiment, DSC cell was purged with dry nitrogen at 40 mL/min. The baseline correction was performed by recording a run with empty pans. Preparations of simulated body fluids We tested the polymer degradation as per ASTM F1635-04a standard with four simulated body fluid types by varying Cl − and HCO 3 −. The composition of simulated body fluids is shown in Table 1. SBFs were prepared in polypropylene beakers by dissolving NaCl, NaHCO 3, KCl, K 2 HPO 4, MgCl 2, 1M HCl, CaCl 2, Na 2 SO 4, Tris HCl in double distilled water and pH was adjusted to 7. 4 and the fluids were further incubated at 37°C for 3 to 4 days and monitored for hydroxyapatite (HA) deposition. In vitro degradation of scaffolds by immersion method (ASTM F1635-04a) Polymer scaffolds were cut to uniform sizes and each sample was weighed before the immersion test. Then, scaffolds were placed in separate polypropylene beakers and fully immersed in simulated body fluids (0. 2 mL of SBF/mm 3 of scaffold) and incubated for 21 days at 37°C. These polymer scaffolds were taken out at preferred time intervals (at 14 and 21 days) and rinsed with distilled water and dried further for studying the morphological changes and weight loss; simultaneously, we checked for pH shift in simulated body fluids due to polymer degradation. Analysis of the scaffold surface by X-ray diffraction The interaction of polymer scaffolds with body fluids was evaluated by studying surface modifications over the scaffolds by X-ray diffraction (XRD) analysis. Precisely, polymer scaffolds immersed in four different simulated body fluids for 21 days were further vacuum dried in order to test HA deposition over the scaffold surface using XRD (Shimadzu) with a 2 theta (2θ) angle between 10 to 80 degrees at a scan speed of 5°/min. Table 1 The composition of inorganic components required for the preparation of simulated body fluids (mmol/L) SBF1 SBF2 SBF3 SBF4 Na + 142. 0 142. 0 142. 0 142. 0 K + 2. 0 2. 0 2. 0 2. 0 Ca 2+ 2. 5 2. 5 2. 5 2. 5 Mg 2+ 1. 0 1. 0 1. 0 1. 0 Cl − 116 121. 0 126. 0 131. 0 HCO 3 − 20. 0 15. 0 10. 0 5. 0 SO 4 2− 1. 0 1. 0 1. 0 1. 0 HPO 4 2− 1. 0 1. 0 1. 0 1. 0 RESULTS Scaffold preparation Highly porous (≥90%) hybrid poly (lactic co glycolic acid) (PLGA, scaffolds (3 mm thickness) were prepared by the solvent casting/particulate leaching method. These scaffolds were cut to uniform sizes for further characterization. For our convenience, scaffolds with 80:20, 75:25 and 70:30 of PLA: PGA are denoted as Sample A, Sample B and Sample C respectively. Scaffold characterization Electron microscopy Initial incubation of NaCl crystals in a humidifier (95%) resulted in fusion of salt crystals, creating interconnected matrices ( Fig. 2A ). This fused NaCl crystals prior to addition of molten polymer mixture increased pore interconnectivity, which improved the overall porosity of scaffolds (≥90%) ( Fig. 2B ). Determination of glass transition temperature (Tg) The DSC analysis of PLGA scaffolds revealed their amorphous nature identified by the presence of glass transition temperature (Tg) and by the absence of melting temperature (Tm). The PLGA scaffolds showed increased polymer degradation with increases in the PGA proportion in the scaffolds. We also noticed a gradual reduction in Tg of polymer scaffolds A, B and C to 48°C, 44. 5°C and 42. 5°C, respectively. Fig. 2 Scanning electronic microscopic images of fused salt crystals (A) and porous polymer scaffolds (B). In vitro degradation of scaffolds by the immersion method (ASTM F1635-04a) 1) Morphological variations with time After 14 days in simulated body fluids, the morphological changes were found to be irregular, with increase in pore size over the scaffold surface. After 21 days, regular and recognizable morphological changes were detected in PLGA scaffolds. We also compared the morphological variations with weight loss in PLGA scaffolds during degradation studies. 2) Weight loss Sample C had shown accelerated weight loss when compared to the other two samples; this may be due to higher PGA content in the scaffold, which is hydrophilic in nature. Though the percentage degradation was high in case of samples C, the degradation rate aws not even (approximately 11% to 22%). This uneven degradation property makes this combination inappropriate for in vivo application. Comparatively, sample B had a stable and optimum degradation rate (approximately 9% to 12%), whereas sample A exhibited a low degradation rate (approximately 4% to 8%) compared to the other two samples. These observations indicated that increase in the PLA content made the scaffolds more hydrophobic and denser, thus making them tougher to be degraded both in vivo and in vitro. 3) pH shift in the simulated body fluids over time By the end of 30 days of incubation, Fig. 3 shows the pH shift in the simulated body fluids over time at 37°C. Sample A showed a slight increase in pH from 7. 4 to 7. 6, which might be due to slower polymer degradation and continues release of (PO 4 ) 3- ions which acts as conjugate bases. Sample C, due to its accelerated degradation to lactic and glycolic acids; showed higher decline in pH from 7. 4 to 6. 7, making this sample vulnerable for in vivo applications whereas sample B was found to be stable with a shift in pH from 7. 4 to 7. 2. 4) Analysis of scaffold surface by XRD XRD analysis using the X-ray diffractometer revealed the presence of straight base line and semi-sharp peaks ( Fig. 4 ), suggesting the semi crystalline nature of our PLGA scaffolds. The XRD patterns also clearly indicated the deposition of HA traces over the scaffold surface. Compared to other SBF types, the XRD spectrum of SBF1 showed better HA deposition whose carbonate content and phosphate contents are more similar to human blood plasma compared to other SBF solutions. Fig. 3 Time dependence on pH of simulated body fluids. A: 80:20 PLGA; B:75:25 PLGA; C: 70:30 PLGA. And 2, 3 and 4 represents four SBF types Fig. 4 XRD analysis of polymer scaffolds for HA deposition. A: XRD spectrum of Pure HA. B: XRD spectrum of Sample A. C: XRD spectrum of Sample B. D: XRD spectrum of Sample C. DISCUSSION Porous hybrid polymer scaffolds (PLGA) were prepared by the solvent casting/particle leaching method, where we fused NaCl particles by prolonged exposure to rich moist environment (95% humidity), resuling in enhanced pore interconnectivity in the PLGA scaffolds ( Fig. 5A ). In this study, fused NaCl particles resulted in the creation of holes on the walls of the scaffolds, which increased the comprehensive modulus of the polymer scaffolds. Improved pore interconnectivity is also helpful in a variety of tissue engineering applications, particularly those requiring close cell to cell contact. Fig. 5B depicts the transverse section of PLGA scaffolds prepared using fused salt particles as porogen. Scanning electron micrographs illustrate the polymer scaffolds with highly porous and well interconnected network. The microstructures of the scaffold determine its interaction with the cells and molecular transport of nutrients and biological wastes from within the scaffold. Exclusively, the pore size of the scaffolds determines the cell seeding efficiency into the scaffold; small pores prevent the cells from piercing into the scaffold, while very large pores prevent cell adhesion due to reduced area to colonize cells. Therefore, scaffold with an open and interconnected pore network and high degree of porosity (≈90%) is described as a perfect model to integrate with the host tissue. PLGA scaffolds prepared by salt fusion also showed irregular pore sizes, ranging between few microns to 300 µm. This variation in porosity may be due to a phenomenon known as solid-liquid phase separation which is attributed to solvent crystallization. When the temperature of the polymer solution is lower than the solvent freezing point (crystallization temperature), solvent crystallizes and the polymer phase is expelled as impurity. A continuous polymer-rich phase is formed by the aggregation of polymer fractions excluded from solvent crystals. After solvent crystals have been sublimated, the scaffold is produced with a micro-porosity similar to the geometry of solvent crystals. In in vitro degradation studies, PLGA scaffolds were degraded by hydrolysis of their ester linkages. The presence of methyl side chain in PLA makes it more hydrophobic and denser than PGA and hence lactide-rich PLGA copolymers are less hydrophilic, absorb less water and are subsequently degraded at a lower rate. Additionally, reduced molecular weight with increased PGA content influences the reduction in “Tg” of the PLGA scaffolds (from 48°C to 42. 5°C, which was quite near the incubation temperature of 37°C). All the above discussed features made sample B (PLA: PGA (75:25)) as a favorite in comparison with other scaffolds. Also, controlled degradation (9-12%) of sample B might provide the room for tissue growth both in vivo and in vitro as biodegradable or restorable material. Fig. 5 Micro structure of highly porous PLGA scaffolds. Whereas A depicts fused NaCl particles with contact points resulted in salt bridges between the particles; B: SEM images of the polymer scaffolds exhibit optimum porosity (≥90%). X-ray diffraction spectra of polymer scaffolds after interaction with different model solutions are explained in Figure 5. After 21 days of immersion in simulated boy fluids, a thin layer of HA deposition was observed over the scaffold surface. Comparatively, SBF1 showed characteristic peaks for HA with all three scaffold types. This fact indicates that SBF1 with carbonate content similar to the human blood plasma could be more suitable and sensitive for in vitro testing of bioactivity and the diffusive character of observed peaks might be the result of poor crystallinity of the precipitated product due to the relative short time of exposure in the simulated body fluids and/or thinner precipitated layer. The slower apatite deposition with SBF1 in comparison with other body fluids could enable the more sensitive in vitro testing of bioactive materials. Moreover, during interaction with human blood plasma, the creation of carbonated hydroxyapatite could be awaited rather than pure HA precipitation. Therefore, the content of carbonate ions in the solutions can be important for the plausibility of in vitro test. In conclusion, in vitro degradation behaviors of PLGA scaffolds in three different formulations were tested systematically with four simulated body fluids for 30 days. Detailed quantitative studies on the physiological features of scaffolds in wet environment along with other material parameters were tested. During these studies, sample B was found to be more appropriate with better physiological characteristics for further in vivo studies. The composition of SBF1 also proved as a better source for further optimization studies.

10. 7555/JBR. 28. 20120119

Journal of Biomedical Research

Morphological MRI and T2 mapping of cartilage repair tissue after mosaicplasty with tissue-engineered cartilage in a pig model

Abstract The aim of this study was to evaluate the efficacy of mosaicplasty with tissue-engineered cartilage for the treatment of osteochondral defects in a pig model with advanced MR technique. Eight adolescent miniature pigs were used. The right knee underwent mosaicplasty with tissue-engineered cartilage for treatment of focal osteochondral defects, while the left knee was repaired via single mosaicplasty as controls. At 6, 12, 18 and 26 weeks after surgery, repair tissue was evaluated by magnetic resonance imaging (MRI) with the cartilage repair tissue (MOCART) scoring system and T2 mapping. Then, the results of MRI for 26 weeks were compared with findings of macroscopic and histologic studies. The MOCART scores showed that the repaired tissue of the tissue-engineered cartilage group was statistically better than that of controls ( P < 0. 001). A significant correlation was found between macroscopic and MOCART scores ( P < 0. 001). Comparable mean T2 values were found between adjacent cartilage and repair tissue in the experimental group ( P > 0. 05). For zonal T2 value evaluation, there were no significant zonal T2 differences for repair tissue in controls ( P > 0. 05). For the experimental group, zonal T2 variation was found in repair tissue ( P < 0. 05). MRI, macroscopy and histology showed better repair results and bony incorporation in mosaicplasty with the tissue-engineered cartilage group than those of the single mosaicplasty group. Mosaicplasty with the tissue-engineered cartilage is a promising approach to repair osteochodndral defects. Morphological MRI and T2 mapping provide a non-invasive method for monitoring the maturation and integration of cartilage repair tissue in vivo.

INTRODUCTION Articular cartilage injury is one of the most common injuries seen in orthopaedic surgery, especially in young athletes. In a retrospective review of 31, 510 knee arthroscopies, the incidence of chondral lesions was 63%. Full-thickness articular cartilage lesions with exposed bone were found in 20% of patients, with 5% in those younger than 40 years old. Deep chondral defects in a weight-bearing area are at high risk of progressing to osteoarthritis (OA), a process accelerated in young and active individuals. Treatment of focal chondral injuries of the knee remains challenging and controversial as the cartilage does not spontaneously repair. Various surgical options have been developed to restore articular cartilage and produce a durable repair ; however, these traditional techniques have had limited success. Most of these procedures lead to the formation of fibrous or fibrocartilage tissue with poor biomechanical and biochemical properties compared to hyaline cartilage. Recently, new techniques such as autologous osteochondral transplantation (AOT, mosaicplasty), autologous chondrocyte implantation (ACI), and tissue engineering models (matrix-associated autologous chondrocyte implantation-MACI) have been developed to treat full-thickness cartilage defects located in weight-bearing regions to create a hyaline or hyaline-like cartilage repair tissue -. Mosaicplasty has been used for more than a decade in clinical practice with successful results. Mosaicplasty involves transfer of cylindrical osteochondral plugs from a non-weight-bearing area of the knee to the defective site in the form of a mosaic. Advantages include the simplicity of one-stage surgical procedure, low morbidity and cost, and a better clinical outcome. However, mosaicplasty also has its disadvantages, such as limited donor cartilage, different thickness and mechanical properties of donor and recipient cartilage, subsidence of graft surface, and poor integration of remnant defect -. The resolution of these issues, especially integration of remnant defect between plugs, is crucial in advancing mosaicplasty as a cartilage repair technique. As tissue engineering is promising for the future of cartilage repair 7 -, mosaicplasty may improve osteochondral repair options. However, their outcomes remain unknown. As articular cartilage repair techniques are increasingly used in clinical practice, methods to assess repair tissue will become increasingly important. Arthroscopy is unsuitable for routine follow-up due to its invasiveness and potential morbidity. In contrast, magnetic resonance imaging (MRI) is a noninvasive method that can provide a more global assessment of the entire repair area, longitudinally at different time points. Additionally, the biochemical properties of the cartilage can be evaluated with newer MRI techniques -. T2 mapping, a new quantitative technique of MRI, provides information about the interaction of water molecules and collagen network within the articular cartilage. It is becoming more popular for the evaluation of cartilage repair. In both animal and human studies, T2 mapping has been shown to differentiate the hyaline cartilage from the fibrocartilage after cartilage repair. We designed a new technique, mosaicplasty with tissue-engineered cartilage, to repair full-thickness osteochondral defects in a pig model. By using a pig model of near equivalent size, anatomy, mechanical loading of the joint and the cartilage and composition as a human subject, our model can be easily transferred to clinical practice. MRI morphology data was obtained 26 weeks postoperatively and compared with macroscopic and histologic findings. We aimed to observe the maturation of cartilage repair tissue after our procedure and evaluate its efficacy by 3. 0T-MRI with morphological scoring and in vivo biochemical T2 mapping. MATERIALS AND METHODS Animals Eight adolescent miniature pigs (5 months old, male to female ratio of 1:1, weighing 25 kg) were utilized for this study. The study protocol was approved by the local institutional review board at the authors' affiliated institutions. Animal welfare and the experimental procedures were carried out strictly in accordance with the Guide for Care and Use of Laboratory Animals (National Research Council of USA, 1996). Formation of tissue engineered cartilage Both knees of each pig were performed by the same operating procedure. Small slices of articular cartilage were harvested from non-weight bearing sites. Chondrocytes were isolated and cultured. Scaffolds of polylactic-co-glycolic acid (PLGA; PLA:PGA, 9:1; AUVON, USA) were immersed in polylysine and sterilized with ethylene oxide. Chondrocytes of the second passage were seeded into the PLGA scaffolds by pipetting directly and repeatedly at a density of 2×10 6 cells/mL. Then, they were cultured in a bioreactor (HFB-40, Celdyne, Houston, TX, USA) containing DMEM/F12 (Gibco, Carlsbad, CA, USA) for 4 weeks to form tissue engineered cartilage for implantation. Surgical procedure After 8 weeks, a standardized full-thickness defect (diameter: 6 mm; depth: 6 mm) - was created in the weight-bearing area of the femoral medial condyle of both knees by using a hollowed drill ( Fig. 1A ). The defects were repaired firstly by autologous osteochondral mosaicplasty. Four autologous osteochondral plugs (diameter: 3 mm) harvested from the non-weight bearing area of the distal femur were implanted into the defect. The right knees were assigned to the experimental group while the left knees served as the control group (8 knees in each group). The dead spaces between plugs were treated as follows: in the control group, dead spaces were left empty ( Fig. 1B ) while in the experimental group, in vitro tissue-engineered cartilage was performed to fill the dead spaces ( Fig. 1C ). Animals began full weight bearing within a very short period of time postoperatively and then were allowed to return to outdoor farm when they regained their normal gait. Fig. 1 Implantation surgery. A: A full-thickness defect (diameter, 6 mm; depth, 6 mm) was created in the weight-bearing area of the medial femoral condyle of the femur of pigs. B: Control group: The defect was repaired by single mosaicplasty. C: Experimental group: The defect was repaired by mosaicplasty with tissue engineered cartilage. The dead space among plugs were filled with the tissue-engineered cartilage. MRI evaluation In vivo, repair tissues were longitudinally assessed 6, 12 and 18 weeks postoperatively with 3T-MRI. Pigs were anesthetized with intramuscular ketamine (12 mg/kg) and midazolam (1 mg/kg) before MRI. They were sacrificed 26 weeks postoperatively with a lethal dose of intravenous pentobarbitone. Knees were harvested en bloc with intact capsule and surrounding muscles for the final MRI, after which each specimen was sectioned into halves corresponding as closely as possible to the MRI sagittal plane. Specimens were then submitted for macroscopic and histologic assessment. Image acquisition MRI was performed on a 3T MR system (Magnetom Trio, A Tim system, Siemens, Erlangen, Germany) with a dedicated 8-channel knee coil. Each pig was placed feet-first in the prone position in the middle of the coil. T2 relaxation times were obtained from T2-maps reconstructed using a multiple spin echo technique with a repetition time (TR) of 1200 milliseconds. Six echo times (TE) were collected (13. 8, 27. 6, 41. 4, 55. 2 and 69 milliseconds). A 16 cm× 16 cm field of view (FOV), 2 mm slice thickness, distance factor of 10% and bandwidth of 228 Hz/pixel were used. Sixteen slices were measured for each pig knee. Images were acquired in the sagittal plane for the medial condyles. Scan time was 4 minutes 36 seconds. For morphologic evaluation, a standard knee protocol was used for all animals, consisting of a sagittal PD-weighted image with fat suppression (TR: 3600 ms; TE: 17 ms) and a coronary T1WI (TR: 765 ms; TE: 16 ms) as well as a T2WI with fat suppression sequence (TR: 4000 ms; TE: 48 ms) in the sagittal and coronary planes. All sequences were obtained with the following parameters, including FOV of 16 cm× 16 cm and section thickness of 2 mm, 16 slices, distance factor of 10% and bandwidth of 228 Hz/pixel. The total acquisition time for morphologic sequences was 5 minutes 17 seconds. Data analysis To evaluate the morphologic condition of cartilage repair tissue and surrounding cartilage by using routine sequences, we used the MR observation of cartilage repair tissue (MOCART) scoring system. The maximum score achievable in the evaluation of 9 variables (the degree of defect repair, the integration of cartilage repair tissue to the border zone, the surface and structure of cartilage repair tissue, signal intensity, the constitution of the subchondral lamina and bone, possible adhesions and effusion) was 100. Two experienced musculoskeletal radiologists assessed and reviewed the images. A representative score for each parameter was determined by averaging the scores of the two observers. The T2 maps were post-processed by using workstation software (Syngo acquisition workplace, Siemens Medical, Erlangen, Germany). Morphologic images provided by the conventional sequences in combination with the slices of cartilage repair tissue sites were selected on T2 images. In each slice, two regions of interest (ROIs) were defined. One ROI, encircling the repair site, was marked with black arrows, and one ROI placed on adjacent normal cartilage in the same slice was used as reference cartilage (white arrow) ( Fig. 2B, D, F, H, J, I, N and P ). ROIs were analyzed as mean ROIs (from the subchondral bone to the cartilage surface with identified cartilage repair sites) and zonal ROIs (dividing the global thickness of cartilage repair tissue and reference cartilage into equal-sized deep and superficial aspects). In all knees, cartilage repair sites were observed over 2 to 3 contiguous sagittal sections and ROIs were placed within repair tissue per section. Thus, a total of 4 to 6 ROIs were measured within each cartilage repair site. Control cartilage sites were also chosen from the same zone surface, but were required to be more than 2 cm away from cartilage repair tissue. T2 maps were then fused on the corresponding anatomic images utilizing workstation fusion software. Sample T2 maps with ROIs positioned are shown in Fig. 2B, D, F, H, J, I, N and P. Fig. 2 Sagittal T2WI-FS (A, C, E, G, I, K, M, O) and corresponding fusing T2 maps (B, D, F, H, J, L, N, P) appearance of the repaired tissue (between the black arrows) and the adjacent normal cartilage (white arrows) in the control group (A, B, E, F, I, J, M, N) and the experimental group (C, D, G, H, K, L, O, P) at 6 weeks (A-D), 12 weeks (E-H), 18 weeks (I-L) and 26 weeks (M-P) after surgery. Macroscopic and histologic analysis Harvested samples were initially examined macroscopically. The longitudinal cross-section of defects was studied, as well as the interface between repaired tissue and adjacent normal osteochondral tissue. The repaired cartilage was macroscopically scored by using a previously developed scale. After macroscopic analysis, samples were submitted for histologic and immunohistochemical staining. The cartilage tissue blocks were fixed with 10% formalin and further processed for histology. Specimens were decalcified for 1 month, dehydrated and embedded in paraffin. Five μm sections were cut transversely with a microtome and hydrated in graded ethanol series. Repair tissue and normal cartilage sections were stained with H&E to evaluate the structural features of repair tissue, and with immunohistochemical staining to study collagen content. Statistical analysis Continuous variables were presented as mean ± standard deviation. Differences among the mean T2 values and macroscopic grading in both groups were analyzed by Student's t -test. Considering the multiple measurements of each pig, a two-way ANOVA with post-hoc test was also performed. Bivariate Pearson correlation analysis was performed to determine the correlation between MOCART scores and gross scores. All analyses were performed using the SPSS software (version 16. 0, SPSS Inc. , Chicago, IL, USA). P -value of less than 0. 05 was considered statistically significant. RESULTS All 8 pigs survived during the study period. No postoperative wound inflammation, infection, swelling or other complications were observed. Morphologic results According to the MOCART scoring system, the experimental group scores were statistically higher than those of controls ( P < 0. 001; Fig. 2, Fig. 3 ). A significant increase in MOCART scores over time was also found in both groups, especially in the experimental group. By individually considering the 9 variables, significant differences were found from the integration of the border zone, surface and signal intensity of repair tissue at 26 weeks ( P < 0. 05). Defect repair in the experimental group was better than controls. In the experimental group, the integration of the border zone was nearly complete, repair area surfaces were smooth and continuous, and a gradual decrease of graft signal intensity was found over time ( Fig. 2O ). However, fluid signal clefts, irregular surface, inhomogeneous signal and cyst formation in repair tissue were found in the control group ( Fig. 2M ). Fig. 3 MOCART scores between the control group (CG) and the experimental group (EG) at different follow-up intervals. * P <0. 05, vs CG. Mean (full-thickness) T2 values T2 values (milliseconds) are shown in Table 1. The adjacent normal cartilage showed comparable T2 values in the two groups at different follow-up intervals ( P > 0. 05) ( Fig. 4 ). Concerning cartilage repair tissue, T2 values in controls were slightly, though not significantly, higher than those of the experimental group at 6 and 12 weeks postoperatively ( P = 0. 86, P = 0. 581, respectively; Fig. 2B, D, F and H ). However, at 18 and 26 weeks, the mean T2 values of repair tissue in the experimental group were significantly lower compared with controls ( P = 0. 005 and P = 0. 002, respectively; Fig. 2J, L, N and P ). The mean T2 values of repair tissue in both groups decreased over time, and were most obvious in the experimental group ( Fig. 4 ). When comparing adjacent normal cartilage and repair tissue in the experimental group, we found that mean T2 values for repair tissue were significantly higher than those of adjacent normal cartilage at 6 and 12 weeks postoperatively ( P < 0. 001). However, comparable results were found for both adjacent cartilage and repair tissue at 18 and 26 weeks ( P = 0. 124 and P = 0. 375, respectively). In the control group, mean T2 values for repair tissue were significantly higher than those of adjacent normal cartilage ( P < 0. 05). At the interface of plugs with native cartilage, there were regions of prolonged T2 relaxation time for the control group ( Fig. 2N ). Fig. 4 Mean (full-thickness) T2 values (milliseconds) of the repair tissue (RT) and adjacent normal cartilage (NC) for the control group (CG) and the experimental group (EG) at different follow-up intervals. NC showed comparable T2 values for two groups at any follow-up interval. Mean T2 values of RT for two groups showed a decreased trend over time. This decreased trend, however, was most obvious in EG. Comparable mean T2 values were found for NC and RT at 18 and 26 weeks in EG. But for CG, mean T2 values of RT were significantly higher than that of NC at either follow-up time point. Zonal T2 values The evaluation of zonal T2 value of adjacent normal cartilage sites showed a significant increase between the deep and superficial layers of articular cartilage at various postoperative follow-up intervals ( P < 0. 05; Fig. 5 ). By evaluating cartilage repair tissue, there were no significant differences for zonal T2 values within the control group at any time point ( P > 0. 05; Fig. 5A ). However, the repair tissues in the experimental group showed varying results with regard to zonal assessment at different periods after implantation. Zonal variation could not be found in repair tissue at 6 and 12 weeks ( P > 0. 05). However, at 18 and 26 weeks, repair tissues in the experimental group showed a significant zonal increase from deep to superficial layers ( P < 0. 05; Fig. 5B ). Table 1 illustrates the results for cartilage repair tissue in the two groups. Fig. 5 Zonal (deep and superficial) T2 values (milliseconds) of the repair tissue (RT) and adjacent normal cartilage (NC) for the control group (A) and the experimental group (B) at different follow-up intervals. * P <0. 05. Macroscopic findings After 26 weeks post-implantation, cartilage autografts in both groups survived in situ, without rising, descending or becoming detached. The integration of the subchondral bone and native bone was good. In the experimental group, gaps among the plugs were replaced by neo-cartilage hyaline cartilage, which appeared macroscopically similar to host cartilage. The neo-cartilage surface was as smooth and congruous as native articular cartilage. According to their cross-sectional appearance, the regenerated cartilage macroscopically appeared similar to host cartilage with good integration. No fissuring was noted ( Fig. 6C ). However, in the control group, the zone of dead spaces was covered by just a thin layer of fibrous tissue. According to cross-sectional appearance, the surface of fibrous tissue was lower than that of adjacent cartilage, making the entire reconstructed articular surface rough. Integration in interface was poor ( Fig. 6F ). The corresponding macroscopic grading of all knees in the both groups is shown in Table 2. Furthermore, there was a significant correlation between MOCART scores and macroscopic scores (r = 0. 970, P < 0. 001). Fig. 6 Histological and immunohistochemical analyses and gross cross-sectional appearance of the repaired cartilage at 26 weeks after surgery. A and D: H&E staining (×100). B and E: immunohistochemical staining for collagen II (×100). C and F: Gross cross-sectional appearance. A-C: In the experimental group, defects were repaired by hyaline or hyaline-like cartilage, which was almost completely integrated into surrounding normal cartilage. D-F: In the control group, the surface of repair tissue was not smooth. The zones of dead spaces had visible fissures and were filled with fibrous tissue which stained negatively for type II collagen. Table 1 Mean and zonal T2 values (milliseconds) for cartilage repair tissue (RT) and adjacent normal cartilage (NC) of two groups according to postoperative follow-up intervals Group Time point (weeks) Subgroup Mean T2 (milliseconds) Deep T2 (milliseconds) Superficial T2 (milliseconds) CG 6 RT 69. 13±4. 36 a 69. 06±4. 26 69. 23±4. 53 NC 50. 90±1. 96 a 49. 00±2. 03 52. 80±2. 00 12 RT 65. 30±4. 61 a 64. 89±4. 61 65. 71±4. 65 NC 52. 75±1. 84 a 51. 56±1. 92 53. 94± 1. 78 18 RT 60. 06±3. 06 b 59. 23±3. 05 60. 91±3. 09 NC 53. 90±1. 98 a 52. 49±2. 15 55. 31±1. 98 26 RT 57. 20±2. 67 b 56. 46±2. 74 58. 06±2. 60 NC 53. 78±1. 99 a 52. 35±2. 06 55. 25±2. 01 EG 6 RT 68. 75±4. 27 a 68. 61±4. 24 69. 26±4. 26 NC 50. 99±2. 35 a 49. 35±2. 70 52. 61±2. 08 12 RT 64. 18±3. 23 a 63. 53±3. 56 64. 94±3. 02 NC 52. 63±1. 45 a 51. 38±1. 30 53. 89±1. 63 18 RT 55. 58±2. 31 b 54. 16±2. 45 56. 98±2. 23 NC 53. 91±1. 70 a 52. 58±1. 80 55. 29±1. 76 26 RT 52. 34±2. 43 b 50. 80±2. 52 53. 83±2. 45 NC 53. 31±1. 77 a 51. 96±1. 70 54. 66±1. 92 Note—CG = the control group; EG = the experimental group; Each data represents the mean and standard deviation. a No significant difference in mean T2 values between NC and RT ( P >0. 05), b Significant difference in mean T2 values between NC and RT ( P <0. 05). Table 2 Results of grading scale for gross appearance Grading CG EG P Coverage 2. 63±0. 74 3. 63±0. 52 0. 008 Neocartilage color 2. 50±0. 76 3. 50±0. 53 0. 009 Defect margins 1. 75±0. 71 3. 25±0. 46 0. 000 Surface 2. 13±0. 83 3. 37±0. 52 0. 003 Total 9. 00±2. 00 13. 75±1. 75 0. 000 CG: the control group; EG: the experimental group. Each data represents the mean ± standard deviation. Histologic results After 26 weeks, PLGA scaffold was completely resorbed. The autologous cartilage from mosaicplasty survived and was observed in situ. The subchondral bone was healed with good integration. Type II collagen staining of the auto-transplanted cartilage was nearly normal in the two groups. In the experimental group, the zones of dead spaces in all defects were covered by hyaline-like cartilage. The reconstructed articular surface was smooth and uniform with better contiguity to native cartilage ( Fig. 6A ). Collagen II in the repaired tissue was evident ( Fig. 6B ). In the control group, dead spaces in all defects failed to regenerate hyaline cartilage. Only a small amount of fibrous tissue was observed ( Fig. 6D ). Type II collagen staining presented negative staining ( Fig. 6E ). The surface of fibrous tissue was lower than that of native cartilage, making the surface rough. Visible fissures were found in the cartilage layer. Sclerotic subchondral bone was also found. DISCUSSION The full-thickness defects of articular cartilage caused by trauma or joint diseases are common in orthopaedic practice. Regeneration cartilage is poor due to the absence of neurovascular supply; hence, the treatment of articular cartilage injury remains a challenge for varying outcomes. Focusing on the development of reliable approaches to restore joint congruency with hyaline cartilage and to integrate this neo-cartilage with surrounding host cartilage is a priority. Current approaches used in clinical practice are mosaicplasty, ACI and tissue engineering techniques such as MACI 5 -. Compared with other methods, mosaicplasty provides better outcomes. It can immediately fill focal osteochondral defects with mature intact hyaline cartilage. The major advantages of mosaicplasty are the convenience of one step procedure, relatively brief rehabilitation period and low cost. However, some problems associated with mosaicplasty include limited donor cartilage, dead spaces between cylindrical grafts and the lack of integration of donor and recipient hyaline cartilage, subsequent cyst formation and uneven surface. Some studies, have demonstrated that these implants would deteriorate over time as a result of the lack of integration and formation of fibrocartilage between host and donor cartilage. Persistent gaps might make pressure conduction abnormal, which causes cartilage degeneration over time. In our study, we also found that the congruency of joint surface could not be restored to its original status in single mosaicplasty due to the dead space between cylindrical grafts. Thus, we aimed to design a model that would improve single mosaicplasty procedure and outcome. Tissue engineering approaches that use 3D biomaterial scaffolds seeded with chondrocytes as carriers for cell growth are popular areas for research. One advantage may be more efficient redifferentiation of chondrocytes, and hence the formation of hyaline-like repair tissue. Tissue engineering is a promising strategy for cartilage repair and has been successful in preliminary studies 8 -. Tissue engineered cartilage has shown to be successful in repairing cartilage defect in a large animal model. In contrast to other repair methods, tissue engineered cartilage has demonstrated improved integration with surrounding cartilage in vitro. Sun et al. demonstrated that tissue engineering technique may solve poor concrescence of remnant defect and the integration of single mosaicplasty in a goat model. Thus, engineered cartilage has the potential to overcome the limitations of poor integration in single mosaicplasty. In our study, autologous chodrocytes were seeded on a 3D scaffold and cultured in vitro to form tissue engineered cartilage. This engineered cartilage was histologically demonstrated to be hyaline-like, forming a highly organized collagen structure similar to native cartilage. Monitoring graft characteristics is increasingly important to assure a good outcome, and to prevent cartilage degeneration that eventually leads to osteoarthritis. MRI permits noninvasive and direct assessment of cartilage defects and can dynamically monitor the progression of cartilage repair. In addition, newer MR sequences, such as T2 mapping, can show major the ultrastructural components of cartilage. The quantitative measures of T2 relaxation time of repair tissue may therefore be useful in the assessment and long-term follow-up of MACT or OAT repairs, with regards to gradual maturation, native differentiation and integration. In mosaicplasty, the integration with MR assessment has posed continuous challenges. Link et al. found that 15% of patients had an incongruity of cartilage-cartilage interface, although the results of the cartilage repair were satisfactory. In an indirect MR arthrography study after OAT, persistent fissure-like gaps between implanted cartilage and native cartilage were demonstrated. After single ostochondral autograft and allograft implantation for the treatment of isolated grade IV osteochondral defects in a canine model, Glenn et al. reported that a persistent cleft was noted at the interface of the articular surfaces of graft and host in 90% of all histologic specimens 6 months postoperatively. The remaining 10% of knees filled clefts with fibrous tissue (comparable to our results). We also found that the dead space between grafts after mosaicplasty did not become completely integrated. For the optimal use of MRI in the evaluation of tissue repair, a simple evaluation and point scoring system that allows efficient statistical data analysis is necessary. The MOCART score has been designed to systematically describe the constitution of the area of cartilage repair and surrounding tissues, and shown to be reliable and reproducible, and can be applied to different surgical cartilage repair techniques. The system is very helpful for longitudinal follow-up of cartilage repair. Based on the use of MOCART score after MACI, Trattnig et al. reported that in longitudinal follow-up after 12, 24 and 52 weeks, there was an increase in the overall MOCART score from 53 to 63, and to 73, after 12, 24 and 52 weeks. This increase in trend for the overall MOCART score was also found in our study. It may possibly reflect a gradually improved regeneration process for tissue repair over time. At 26 weeks, MOCART score in the experimental group (77. 50 ± 6. 55) was obviously higher than that of controls (59. 37 ± 6. 78). In the control group, the congruency of joint surface was unable to be restored to its original status, even up to 26 weeks. Fluid signal clefts at the interface between graft tissue and native cartilage indicated poor graft integration. Inhomogeneous signal and cyst formation in the repair tissue of controls could be seen on T2WI with fat suppression. Cyst formation was associated with edema-like signal intensity and also demonstrated a fibrocartilage or fibrous tissue appearance. The inhomogeneous signal within graft cartilage may have been due to volume averaging of developing fibrous tissue, or possibly fluid within the interstices between plugs. In the experimental group, the integration of border zone was nearly complete. Continuous smoothing of the graft surface was observed. A gradual decrease in graft signal intensity was found over time. These findings may be related to graft integration and remodeling. Our results demonstrated a strong correlation between MOCART scores and macroscopic scores. Therefore, we believe that mosaicplasty with tissue engineered cartilage may solve the problem of poor integration found with single mosaicplasty, and that morphological MR can assess the maturation and integration process of repair tissue. For the gaps between transplanted cartilage plugs and native cartilage, the length of postoperative period should be considered, as surface congruity will improve over time due to fibrocartilage tissue formation between osteochondral plugs. However, fibrocartilage repair tissue, or fibrous tissue, lacks the properties of hyaline cartilage for optimal joint function. It is critical to understand the biochemical constitution of repair tissue and this can be accomplished by using biochemical T2 mapping techniques. Several studies have demonstrated the validity of T2 mapping for the assessment of articular cartilage and cartilage repair tissue 7 -. T2 relaxation times allow for the evaluation of the major matrix components of cartilage such as water concentration, collagen architecture and orientation. An increased cartilage full-thickness T2 value is associated with early cartilage injuries. Furthermore, the spatial distribution of T2 values can be used for in vivo monitoring of the biomechanical properties and maturation changes of various cartilage layers. Many factors have an effect on T2 values, so that T2 values show an individual variability. In our study, we chose adjacent healthy cartilage as a comparison to repair tissue. In healthy control cartilage, T2 values between groups were comparable at each follow-up time point. T2 spatial variation of normal cartilage showed an increased trend from the deep to superficial zones, as previous studies have reported. This correlated with type II collagen fiber matrix organization in each zone of hyaline cartilage. Considering full-thickness T2 values in the two groups in follow-up period, repair tissue was characterized by high mean T2 values versus adjacent healthy cartilage in early follow-up period (6 weeks and 12 weeks postoperatively). These higher T2 values are probably inherent to chondral edema and a difference in matrix content versus adjacent healthy cartilage. T2 values of repair tissue in the single mosaicplasty group were still significantly higher than in the healthy control cartilage at 26 weeks. In addition, dead space regions at graft-host interface were commonly associated with prolonged T2 relaxation time. This phenomenon was also reported in a canine model study after OAT by Glenn et al. . In another study by Salzmann et al. , OAT plugs also demonstrated elevated T2 values compared to healthy control cartilage. Furthermore, significantly different T2 values were found in the immediate adjacent regions surrounding the cylinders after OAT. A possible explanation is that OAT plugs do not integrate at the chondral margin. However, T2 values of repair tissue after MACI were significantly lower compared with healthy control cartilage. In our study, we demonstrated that elevated T2 values in the repair tissue after OAT was associated with poor integration among plugs, leading to a higher water concentration and faster water mobility in those regions. However, in the experimental group at longer follow up intervals (18 weeks postoperatively), full-thickness T2 repair tissue values approached those of adjacent normal cartilage as previously reported. These results indicated that repair tissue collagen and water content were similar to the adjacent normal cartilage. The decrease in full-thickness T2 values over time reflects tissue maturation. Thus, full-thickness T2 values demonstrate that mosaicplasty with tissue engineered cartilage can restore full-thickness defects and achieve a good integrative condition. In our study, besides full-thickness evaluation, the zonal assessment also provided possible insight into the maturation of cartilage repair tissue. In the longitudinal follow-up by using T2 mapping, over 6 to 26 weeks, the trend of increasing T2 values from deep to superficial cartilage layers became apparent. A significant difference was first found at 18 weeks in the experimental group, indicating that remodeling can occur in repair tissue, resulting in the reorganization of implanted cartilage. However, zonal variation in the depth of repair tissue in controls was not detected at each follow-up time point. At 26 weeks, histologic outcomes demonstrated that defects in the experimental group were covered by hyaline-like cartilage, including the dead spaces filled by tissue engineered cartilage. In the control group, only a small amount of fibrous tissue was observed in dead space. Although cartilage autografts survived as hyaline cartilage, the poor integration of the dead space between grafts may lead to inferior properties of the whole repair tissue versus native cartilage in the control group. The fissure existence and fibrous materials filling in the defect may provide abnormal stress distribution with weight-bearing which produce change in autografts cartilage matrix. The different T2 spatial distribution between groups may be based on differences in composition of the repair tissue, maturation and integration of cartilage repair tissue. Previous histologically validated animal studies also revealed the increase in zonal T2 as an indicator of the formation of hyaline or hyaline-like cartilage structure of the cartilage repair tissue. The zonal variation of T2 values was observed in human cartilage repair tissue after MACI. Thus, full-thickness assessment may aid in the visualization of the maturation of cartilage repair tissue. In our study, T2 maps were used to characterize cartilage repair tissue as a virtual biopsy and provided initial information about the integration of repair tissue. This study had several limitations. Firstly, the sample size was small with a relatively short-term follow-up period. Additional studies with larger numbers of samples, long-term follow-up and more new repair procedures are necessary to validate the approach of T2 mapping as a measure of the quality of repair tissue. Secondly, there were no histologic and pathologic results compared to MR results before 26 weeks, due to the small number of specimens and good repair efficacy in the long-term follow-up. On the other hand, at 26 weeks, we did find clear histologic correlation with MR findings. In addition, partial volume effects have limited the results concerning zonal variation due to thinner cartilage diameter. Multiple ROI measurements may diminish this error. Finally, the imaging of the cartilage repair tissue is more challenging than the imaging of native articular cartilage in many ways, due to the different repair procedures applied. Standardized assessment protocols for cartilage repair would be ideal. In conclusion, we have demonstrated the superiority of mosaicplasty with tissue engineered cartilage over single mosaicplasty for the repair of articular defects in the knee. Tissue engineered cartilage can promote the integration of dead spaces in single mosaicplasty. MRI is a noninvasive method for monitoring the development and integration of cartilage repair tissue. Full-thickness and zonal T2 analysis may provide additional information concerning the maturation of cartilage repair tissue in vivo.

10. 7555/JBR. 29. 20150027

Journal of Biomedical Research

From nerve to blood vessel: a new role of Olfm2 in smooth muscle differentiation from human embryonic stem cell-derived mesenchymal cells

No abstract available

Vascular smooth muscle cell (SMC) differentiation is an important process in vasculogenesis and angiogenesis during embryonic development. The alterations in the differentiated state in SMCs contribute to a variety of major cardiovascular diseases such as atherosclerosis, hypertension, restenosis and vascular aneurysm [ 1, 2 ]. A better understanding of the cellular and molecular mechanisms that control SMC differentiation is essential to help develop new approaches to both prevent and treat these diseases. Therefore, development of reliable and reproducible in vitro cellular models in order to study the differentiation mechanisms is important although it has been challenging because of intrinsic peculiarities of SMC. SMCs originate from at least eight different progenitors during embryonic development including neural crest, proepicardium, mesothelium, splanchnic mesoderm, secondary heart field, mesoangioblasts, somites and various stem/progenitor cells [ 1 ]. SMC populations from different embryological origins are observed in different vessels as well as within the same vessel segments although showing sharp boundaries with no intermixing of cells from different lineages [ 3 ]. Importantly, SMCs from different origins are regulated differentially and can exhibit a wide range of different phenotypes. Even in adult organs, SMCs are not terminally differentiated because the cells may undergo phenotypic modulation in response to alterations in local environmental cues including growth factors/inhibitors, mechanical influences, inflammatory mediators, cell-cell and cell-matrix interactions [ 2 ]. SMC differentiation is a complex but poorly defined process although much progress has been made in identifying molecular mechanisms controlling the expression of SMC specific genes. Accumulating evidence has shown that a precisely coordinated molecular network orchestrates the SMC differentiation program involved in a range of signaling pathways including TGF-β, retinoid, extracellular matrix, Notch, reactive oxygen species, histone deacetylase and microRNA signaling [ 4 ]. Several in vitro model systems have been developed to mimic the SMC differentiation in vivo including using C3H10T1/2 cells, neural crest cells, A404, embryoid body and embryonic stem cells. Although these models have significantly contributed to our understanding of SMC differentiation, each of these models has its limitations. In addition, human embryonic stem cell can differentiate to both endothelial cell (EC) and SMC populations in the same differentiation conditions. Though the cells are excellent for in vivo neoangiogenesis and regeneration of blood vessels, they may not be ideal for precisely dissecting the molecular mechanism governing SMC differentiation because SMCs differentiated from embryonic stems cells are heterogenic and thus contain a mixed population. We recently developed a novel in vitro model for TGF-β-induced SMC differentiation from human embryonic stem cell-derived mesenchymal cells (hES-MCs). hES-MCs, derived from H9 human embryonic stem cells, are natural SMC progenitors for mesoderm-derived SMCs that account for most of the vascular SMCs [ 1 ]. hES-MCs have the capacity to produce the three lineages associated with mesenchymal stem cells including osteogenic, chondrogenic and SMC lineages. We found that hES-MCs can be robustly differentiated to SMC phenotype upon TGF-β stimulation and exhibit a morphology resembling functional SMCs. hES-MCs have the potential to be used for tissue engineering for regeneration of human SMCs due to their mesodermal origin. Interestingly, the nervous and the vascular systems share many common features including a similar and often overlapping anatomy characterized by highly branched and ramified layouts, and common signaling pathways. Many similarities can also be found at the cellular and even extend to the molecular levels. There is strong evidence for coordination between the two systems [ 5 ]. In some cases this coordination may be achieved by utilizing the same cues or signals, suggesting that common molecules may regulate the development of both nervous and vascular systems. Olfactomedin 2 (Olfm2), first found in the frog olfactory neuroepithelium, belongs to the family of Olfactomedin domain-containing proteins consisting of at least 13 members in mammals. Olfm2 expression is developmentally regulated. Blockade of Olfm2 reduces eye size, hinders optic nerve extension, and disrupts anterior central nerve system and head development including neural crest cell-derived cartilaginous structures of the pharyngeal arches in zebrafish [ 6 ]. In humans, a R144Q substitution in Olfm2 protein is thought to be the disease-causing mutation in Japanese patients with open-angle glaucoma. Effects of Olfm2 on eye development in developing zebrafish appear to be related to Pax6 signaling [ 6 ]. Pax6 is a master transcriptional factor for eye development and functions. Importantly, Pax6 has been shown to physically interact with TGF-β, which contributes to maintaining functional status of eyes. These results suggest a possible role of Olfm2 in the TGF-β signaling cascade during early eye development. Our recent study has shown that Olfm2 plays a role in vascular development, especially in TGF-β-induced SMC differentiation [ 7 ]. Olfm2 is dramatically upregulated during TGF-β-induced SMC differentiation of hES-MCs. Olfm2 knockdown suppresses TGF-β-induced expression of SMC markers while Olfm2 overexpression promotes the marker gene expression. Interestingly, TGF-β induces Olfm2 nuclear accumulation, consistent with our finding that Olfm2 is abundantly expressed in nuclei of SMC in normal human aorta. Olfm2 expression is Smad2/3-dependent. In addition, Olfm2 acts as a nuclear cofactor binding to serum response factor (SRF) to promote SRF/CArG box interaction, leading to an enhanced transcription and expression of SMC marker genes. Olfm2 promotes SRF binding to SMC marker promoters through inhibiting the expression of HERP1 (Hrt2, Hey2, Hesr2, and CHF1) and thus attenuating the SRF association with HERP1, which is a downstream target of Notch signaling and a transcriptional repressor involved in SMC differentiation [ 8, 9 ]. Our study indicates that the homeostatic balance between Olfm2 and HERP1 expression may be one of the factors that determine whether or not SMC marker genes can be effectively induced by TGF-β after the initial phase of SMC differentiation. Interestingly, in addition to SMC differentiation, both SRF and HERP1 are also essential for nerve cell differentiation [ 10 ]. Our study identifies Olfm2 as a novel contributor that can regulate both processes. Further in-depth analysis of the Olfm2-SRF-HERP1 axis may provide new insights into the molecular networks coordinately regulating the neural and vascular development during embryogenesis. Moreover, identification of additional new factors that regulate both nervous and vascular systems is likely to unravel additional common mechanisms underlying the unique interaction between nerve and blood vessel.

10. 7555/JBR. 32. 20160066

Journal of Biomedical Research

Bone regeneration with adipose derived stem cells in a rabbit model

It has been shown that stem cells are able to calcify both in vitro and in vivo once implanted under the skin, if conveniently differentiated. Nowadays, however, a study on their efficiency in osseous regeneration does not exist in scientific literature and this very task is the real aim of the present experimentation. Five different defects of 6 mm in diameter and 2 mm in depth were created in the calvaria of 8 white New Zealand rabbits. Four defects were regenerated using 2 different conveniently modified scaffolds (Bio-Oss® Block and Bio-Oss Collagen®, Geistlich), with and without the aid of stem cells. After the insertion, the part was covered with a collagen membrane fixed by 5 modified titan pins (Altapin®). The defect in the front was left empty on purpose as an internal control to each animal. Two animals were sacrificed respectively after 2, 4, 6, 10 weeks. The samples were evaluated with micro-CT and histological analysis. Micro-CT analysis revealed that the quantity of new bone for samples with Bio-Oss® Block and stem cells was higher than for samples with Bio-Oss® Block alone. Histological analysis showed that regeneration occurred in an optimal way in every sample treated with scaffolds. The findings indicated that the use of adult stem cells combined with scaffolds accelerated some steps in normal osseous regeneration.

Introduction The repair of bone defects continues to be a challenging part of many reconstructive procedures [ 1 – 2 ]. Although autogenous bone grafts remain the standard in the reconstruction of bone defects, they have disadvantages, including the limited amount of available bone and morbidity of the donor site [ 3 ]. A previous approach to this problem focused on the development of various artificial materials instead of autogenous bone. However, artificial bone substitutes may expose the patient to the risks of foreign body reactions and infections [ 4 ]. Recent advances in cell culture techniques may provide an elegant solution to these restrictions [ 5 ]. Several recent studies have reported the ubiquitous distribution of adult stem cells in various tissues and organs, including bone marrow, muscle, brain, skin, and more recently, subcutaneous fat. Stem cells represent the new frontier in the field of regenerative medicine and are seen as a promising and suitable means to overcome the mentioned drawbacks [ 6 ]. Several studies report that adult stem cells can be isolated from many organs and tissues [ 2 ]. In particular, adipose tissue contains cells that have the ability to proliferate and differentiate into multiple cell lines [ 1, 7 ]. These stem cells may have important applications in tissue engineering. As a matter of fact, adipose tissue-derived stem cells (ADSCs) have the potential to differentiate into bone, cartilage, fat, myocardium, skin, and neurons [ 8 – 9 ]. In current clinical practice, mesenchymal stem cells are commonly collected from the bone marrow. However, no significant differences between adipose-derived stem cells and bone marrow-derived mesenchymal stem cells from the same patient were observed with regard to the yield of adherent cells, their growth kinetics, cell senescence, differentiation capacity, and gene transduction efficiency [ 10 – 11 ]. Moreover, adipose tissue can be collected under local anesthesia more easily than bone marrow, making the procedure less invasive to the donor [ 12 – 13 ]. The aim of this study was to investigate the differences in vivo between traditional bone regeneration and the combination with tissue engineering in the animal model. Materials and methods Eight New Zealand rabbits weighing about 4. 5 kg, treated according to the "European conventions for the protection of vertebrate animals used for experimental and other scientific purposes" (1999/575/EC) and Italian regulations (DL 116. 1993), underwent the first surgery for the removal of adipose tissue. The animals were operated under anesthesia with Xylazine and Zoletil ®. After shaving and disinfecting the skin with Betadine ®, a flap was created for the removal of intrascapular adipose tissue. A single withdrawal in each animal was made from their adipose tissue and stem cells were subsequently isolated to avoid problems of rejection and the inconvenience of intervening in immunosuppressed animals. In this way any replanting will be possible with self cells taken directly from the animal. After collection, Vicryl ® sutures and an additional Betadine ® disinfection on the skin were performed. In the following days, the animals were given an antibiotic and anti-inflammatory analgesic therapy with enrofloxacin (Baytril ®) and carprofen (Rimadyl ®) to prevent complications. Isolation of adult mesenchymal stem cells The removed adipose tissue was transported in the laboratories of SISSA of Trieste, where the process of isolation of mesenchymal stem cells immediately began. The extracellular matrix was digested by a 0. 1% collagenase solution in a water bath at 37 °C for 60 minutes. Thereafter, the cells obtained were seeded in Dulbecco's modified Eagle's medium containing 10% bovine serum and antibiotics (control medium) and centrifuged for 3 minutes at 1, 500 r/minute. Then, yjey were filtered through a nylon membrane with a pore size of 100 µm and the cells were placed in control medium culture. The selection of cells that adhered to diskette and the gradual elimination of adipocytes were made following a well described protocol by Rietze et al. [ 14 ]. Preparation of scaffolds To support cell growth in the plant site we decided to use two different scaffolds produced by Geitslich ®: deproteinized bovine bone (Bio-Oss ® Block Geistlich) and bovine cancellous granular with the addition of a collagen matrix to 10% (Bio-Oss Collagen ® Gei-stlich). The Bio-Oss ® Block was processed under sterile conditions to obtain discs with a diameter of5 mm and a thickness of 2 mm so that it fits perfectly with the type of defect. The Bio-Oss Collagen ® were cut in two portions, since preliminary experiments suggested that this was the amount needed to adequately fill the defect. Osteogenic differentiation For osteogenic differentiation, it was decided to adopt a protocol developed by Kakudo, which was already well documented. Cells differentiated in this way are able to calcify very quickly if implanted subcutaneously, but there is no evidence in the literature on the present operation to repair a critical defect. A total of 1, 000, 000 cells were seeded in each scaffold in control medium and kept for 24 hours. Thereafter, the medium was replaced with osteogenic medium, obtained by adding to the control medium 10 nmol/L dexamethasone, 10 mmol/L of β-glycerophosphate 82 g/mL acorbato-2 phosphate, and cells were allowed for differentiation for 14 days. The osteogenic medium was replaced 2 times every 7 days for a total of four substitutions. Creating experimental bone defect and plant The second surgery mode of sedation and anesthesia were the same as the first. The flap was prepared for skeletonization of the parietal bones of the calvaria of rabbits. We proceeded to create five experimental bone defects (two on each parietal bone and one before that) of 6 mm in diameter and 2 mm in depth. Defects can be produced with standardized trephine burs, using continuous saline irrigation for cooling [ 15 ]. Using a special caliber created for the experiment, we checked the precise size of defects before insertion of the scaffolds alone and enriched with cells differentiated into the osteogenic line. Defects were grafted differently: on the right of the rabbit head, with Bio-Oss ® Block and mesenchymal stem cells in the caudal defect and with Bio-Oss Collagen ® and mesenchymal stem cells in the cranial one; on the left, Bio-Oss ® Block in the caudal defect and Bio-Oss Collagen ® in the cranial one, both without the addition of stem cells. The fifth defect, the frontal one, was left empty and used as internal control in each animal ( Supplementary Fig. 1, available online). The defects filled with different materials were coated with a collagen membrane (BioGide Geistlich ®) fixed with 5 specially modified titanium pins (Altapin ®). Finally, the flaps were sutured by wire Vicryl ® 3/0 and steel clips were applied outside on the skin, preventing the reopening of the flap. The skin was disinfected again with Betadine ® and animals were treated with antibiotic (enrofloxacin, Baytril ®), and anti-inflammatory/analgesic (car-profen, Rimadyl ®) therapy. No animal showed signs of suffering during the postoperative period, tightly controlled and recorded in audiovisual behavior. Thanks to the developed surgical technique, it was not necessary to complete any sacrifice before the scheduled date. Sacrifice and sampling Two animals were sacrificed by intravenous injection of Tanax ® after general anesthesia with Zoletil ® respectively at 2, 4, 6 and 10 weeks. The bone samples, taken immediately after sacrifice, were immersed in 4% buffered formaldehyde. Every sample was subjected to Micro-CT scan and then to histological analysis. Preparation of samples The marked samples were immersed in a solution of 40% formic acid and formate buffer in a 1:1 ratio for 5 days, long enough to ensure adequate decalcification. Samples were then post-fixed overnight in 10% buffered formalin, dehydrated through an ascending scale of alcohols (from 50% to 100%), clarified with xylene and fixed with permeating liquid paraffin at 60 °C. The material was then embedded in paraffin solidified at room temperature, so it was possible to obtain histological sections 5–10 µm thick, spread on glass slide. The sections were stained with hematoxylin to highlight nucleus and eosin for intra- and extra-cellular structures. Micro-CT data analysis X-ray microcomputed tomography (µ-CT) of samples was obtained by means of a cone-beam system called TOMOLAB (www. elettra. trieste. it/Labs/TOMOLAB). The device is equipped with a sealed microfocus X-ray tube, which guaranteed a focal spot size of 5 µm in an energy range from 40 up to 130 kV, and a maximum current of 300 µA. As a detector, a CCD digital camera was used with a 49. 9 mm×33. 2 mm field of view and a pixel size of 12. 5 µm×12. 5 µm. The samples were positioned onto the turn-table of the instrument and acquisitions were performed with the following parameters: distance source-sample (FOD) 100 mm; distance source-detector (FDD) 200 mm; magnification 2×; binning 2×2; resolution 12. 5 µm; tomographies dimensions (pixels) 1, 984×1, 024; slices dimensions (pixels) 1, 984×1, 984; number of tomographies 1, 440; number of slices 864; E=40 kV, I=200 µA; exposure time 2. 5 seconds. The slices reconstruction process achieved by means of commercial software (Cobra Exxim) started once the tomographic scan was completed and all the projections were transferred to the workstation. Input projections and output slices were represented by files (one file per projection and one file per slice) using arrays of 16-bit integers. Three-dimensional visualizations of the reconstructed slices were performed by means of OsiriX v. 3. 9. 4. 64bit Imaging Software. This software allowed identification of the correct angulation and segmentation of the samples from which the planar view was extracted. From the planar view, the percentage of newly formed bone was calculated as the rate between the volume of the newly formed bone on the volume of the original defect. Calculation was performed by means of Image ProPlus 6. 2 software. The image analysis procedure was assisted by a surgeon, expert of radiographies, to identify the orientation and the correct border between the newly formed bone and the native bone ( Supplementary Fig. 2, available online). Through the use of a code written in IDL, we cored three volumes of 128×128×36 voxels (the first on the original bone structure adjacent to the defect, the second on the scaffold and the third at the interface between the two), binarized by Otsu algorithm for evaluation of class separability threshold [ 16 ] implemented in software PORE3D. Later, through the program GEHC MicroView, selected volumes of interest were analyzed by the stereological parameters, using Euler number ( e ) as a selected parameter. e is a direct index of trabecular connectivity and can be defined as the maximum number of portions removable from the structure without losing its integrity. e was normalized for the parameter BVTV (bone volume/total volume) [ 17 – 19 ] Histological analysis Immediately after µ-CT scans, the samples were sent to the Department of Pathology of the Hospital of Monfalcone (Gorizia, Italy) to perform histological analysis. Samples were immersed in a solution of EDTA (ethylenediaminetetraacetic acid) disodium in acid buffer for five hours. Samples were then placed in histology cassettes properly oriented. The biocassettes were then included in the histoprocessor where, with a predetermined sequence and timing, the samples were post-fixed in 10% buffered neutral formalin, dehydrated through an ascending scale of ethanol (from 50% to 100%), clarified with xylene and permeated by liquid paraffin at 60 °C. The material was then embedded in paraffin and allowed to solidify on chilled plates. The obtained block was then sectioned by a microtome and 8 µm thick histological sections were spread on a glass slide and placed in an oven at 60 °C for 1 hour to ensure a good adhesion of the sections to the glass slides and at the same time to dissolve the paraffin excess. Subsequently, the sections were stained with hematoxylin and eosin. After the staining, the preparations were dehydrated and mounted with resin, which was placed on the coverslip. During histological analysis of the samples, we focused on the following fundamental aspects: - Comparison between the ossification with and without matching mesenchymal stem cells to two different scaffolds; - Comparison of the integration of the scaffolds in the context of the newly formed tissue, with and without using mesenchymal stem cells. Results Micro-CT data results Similar findings were registered in all the analyzed samples: the quantity of new bone (higher gray levels) showed a trend to increase with the animals’ age. When comparing the graphic curves obtained for both Bio-Oss ® and Bio-Oss ® with mesenchymal stem cells, the quantity of new bone for the Bio-Oss ® with mesenchymal stem cells samples was higher when age was considered. In particular, the 10-week samples with stem cells presented a more differentiated gray level than the 10-week samples without stem cells ( Fig. 1 ). Fig. 1 Comparison of pixel frequency of gray level in the samples treated with Bio-Oss without stem cells (A) or Bio-Oss and stem cells (B) at different weeks. Histological analysis results Specimens at 2 weeks: There were no inflammatory cells around the particles in animals treated with deproteinized bovine bone material associated with mesenchymal stem cells. Most particles were surrounded by newly formed connective tissue. In some areas, newly formed bone matrix was firmly adhering to the surface of the deproteinized bovine bone, with no spaces at the interface. In other areas, the islands of newly formed connective tissue were not closely related to the particles of deproteinized bovine bone material. In these cases, the surface was almost entirely covered by rounded cells that appeared to be active and wrinkled by a dense extracellular matrix. The findings of the left side defects were characterized by the presence of filler material with connective tissue around it. Deproteinized bovine bone units were easily distinguishable from newly formed connective tissue in the samples ( Fig. 2 ). Specimens at 4 weeks: The histological study of the side shows areas of connective tissue with clearly distinguishable areas of newly formed bone tissue. In particular, bone tissue presented a typical structure of bone tissue such as lamellar histological structure ( Fig. 3 ). Moreover, scaffold particles were still distinguishable from newly formed bone tissue and connective tissue. There was no inflammatory infiltrate at the interface or around the particles. In the control side, histology was consistent with that shown at 4 weeks ( Fig. 4 ). Specimens at 6 weeks: In the test side, particles of scaffold surrounded by maturing bone were easily visible ( Fig. 5 and 6 ). Fig. 2 Haematoxylin-eosin staining, at two-week sample. A: Bio-Oss® Block without stem cells (×100); B: Bio-Oss® Block with stem cells). Note the bone tissue (b) and connective tissue (c) (×200). Fig. 3 Newly bone tissue. The Haversian canal is located in the center of the image (×100). Fig. 4 At 4 weeks Bio-Oss Collagen ® with stem cells (×100). Fig. 5 Hematoxylin-eosin staining, showing the Bio-Oss® Block without stem cells at 6 weeks (×100). Fig. 6 At 6 weeks Bio-Oss® Block (a) with stem cells and bone tissue (b) (×100). Specimens at 10 weeks: In samples of last sacrificed animals complete ossification, demonstrated by trabecular bone structure still maturing, absence of connective tissue, completely replaced by bone in every regenerated defects. Only in the empty defect we can observe trabecular bone still surrounded by islands of connective tissue. As a very important result, histology confirmed the formation of the ossification front already deducible from micro-CT analysis. This ossification front was a zone of lower density of both healthy bone and scaffold, situated at the interface and growing with the temporal distance between the implant and animal sacrifice ( Fig. 6 and 7 ). Fig. 7 Micro-CT image of an intraosseus defect with Bio-Oss® Block and stem cells at 10 weeks. Note the interface (i) between bone (b) and biomaterial (B). There seems to be no significant differences in regeneration between Bio-Oss ® Block and Bio-Oss Collagen®. Both scaffolds were not easy to handle: Bio-Oss ® Block must be specially modified in order to adapt to the defect. Its solid structure was optimal to support the overlying tissues and to provide the needed support for the bone to regenerate. On the contrary, Bio-Oss Collagen® instead had a spongiform structure losing most of its mechanical ability in contact with the blood. It was also not indicated for supporting tissues. Nevertheless, it is extremely easy to handle by non expert operators and it is an excellent medium for growing the bone. Discussion Eight New Zealand rabbits were used for this study. After surgery and up to the date of sacrifice, the recovery of all the animals involved in this study was considered normal in terms of eating behavior, weight and quality of life. To the best of our knowledge, in the present literature, there is no evidence of a general agreement on the correct animal model to be used for scientific purposes similar to the ones involved in this study. Some authors claim that one should use a model in which bone biology and composition is very similar to that of humans ( i. e. dog, sheep, goat, pig or monkey) [ 20 ]. Indeed, they claim that the trabecular bone contained in rodent bones is poor also in the metaphysis of long bones and that re-modeling of Haversians channels by osteoclastic re-absorption does not happen in rodents [ 21 ]. On the contrary, other authors, without denying the previous observations, claim that the use of the above mentioned animal models present some drawbacks [ 2 ]. Rodents are the most common animals employed during in vivo studies reported in the literature; this is mainly related to practical reasons: large animals are more expensive, time consuming and more difficult to keep. There are multiple factors that affect osseous bone healing that are best evaluated in the in vivo environment, including biomechanical, cellular and vascular mechanics that comprise the healing process. The model chosen to assess a particular device should mimic the environment in which the device will be used therapeutically. One should choose a “critical-size defect” that will not spontaneously regenerate, to best characterize the contribution of the device to healing. Tissue engineering devices can be screened in various preclinical animal models to determine their potential. Depending on previous data on a product, one can choose the appropriate model to screen the potential of a new device. For new technology, it is advised to begin with small animal models, which can provide early data in a relatively fast and cost-effective way. This kind of research can progress with systems that simulate the human wound and therapeutic environment more closely in association with the planned clinical application of the device [ 22 ]. Both scaffolds exhibit open-pore structures allowing cell penetration and attachment [ 23 – 25 ]. Numerous studies have employed defects in the calvarium as the site in which to screen biomaterials for the bone response that they elicit, principally because the diameter of the critical-size defect is smaller than that in long bone sites. Another advantage is that multiple defects can be produced in the same surgical field. An incision is made through the scalp to expose the periosteum, which is then elevated and retracted to expose the bone. Defects can be produced with standardized trephine burs, using continuous saline irrigation for cooling. The biomaterial can then be placed into the defects [ 26 ]. Having shown no difference between two scaffold-groups regarding bone regeneration, micro-CT and histological examination revealed the presence of early bone changes. Modern techniques of bone regeneration are still very far from satisfactorily resolving all situations in which they are needed. Considerable help in the future could come from Tissue Engineering. This last can provide new and "targeted" tools enabling tissue-regenerating stimuli to arrive at the aimed cellular lines only. The new bone, formed with adipose derived stem cells, seems to be superposable with that obtained by traditional regenerative technique, both in the histological appearance and in the timing of neodeposition and calcification of the extracellular matrix. Further studies should be conducted in order to understand what the contribution is to regeneration of adult stem cells. The engineering of biomaterials plays a major role in this field. Indeed, the use of selected and differentiated cells should be combined with a specially designed scaffold, allowing amplifying the cells potential. Currently, there is no available scaffold meeting all the requirements of tissue engineering and research should be focused in this direction.

10. 7555/JBR. 34. 20200063

Journal of Biomedical Research

Mechanotransduction, nanotechnology, and nanomedicine

Mechanotransduction, a conversion of mechanical forces into biochemical signals, is essential for human development and physiology. It is observable at all levels ranging from the whole body, organs, tissues, organelles down to molecules. Dysregulation results in various diseases such as muscular dystrophies, hypertension-induced vascular and cardiac hypertrophy, altered bone repair and cell deaths. Since mechanotransduction occurs at nanoscale, nanosciences and applied nanotechnology are powerful for studying molecular mechanisms and pathways of mechanotransduction. Atomic force microscopy, magnetic and optical tweezers are commonly used for force measurement and manipulation at the single molecular level. Force is also used to control cells, topographically and mechanically by specific types of nano materials for tissue engineering. Mechanotransduction research will become increasingly important as a sub-discipline under nanomedicine. Here we review nanotechnology approaches using force measurements and manipulations at the molecular and cellular levels during mechanotransduction, which has been increasingly play important role in the advancement of nanomedicine.

Introduction Mechanotransduction is the process by which organisms perceive physical forces by producing biochemical signals in response. Such signals have been shown to control cell proliferation, migration, differentiation, and death [ 1 – 8 ]. The significant influence of mechanotransduction on cells suggests potential clinical importance in modulating, altering, and controlling the process. The recent discovery of the underlying molecular mechanisms of mechanotransduction advances the study closer to the point of clinical importance. Although mechanotransduction has been discovered for a long time and extensively studied at the cellular and tissue levels [ 9 ], it is just recently that its mechanism has been revealed at the single molecular level [ 4 ]. Although studies at cellular and tissue levels are still necessary to understand mechanotransduction, research on the process will increasingly apply nanotechnology. Clinical research into the applications of nanotechnology in mechanotransduction is becoming a sub-set of nanomedicine. One of nanomedicine uses nanoparticles (NPs) for site-specific delivery of drugs and diagnostic agents. Some NPs show promise in helping to reveal the workings of mechanotransduction for tissue engineering, drug delivery, diagnostics, and other medical advancements. Mechanotransduction Internal and external mechanical forces to the human body such as shear forces in blood vessels, stretching, actomyosin contraction, gravity, acoustic vibration, and pressure can regulate cellular development and various processes [ 4, 10 – 15 ] ( Fig. 1 ). For example, exposure to the microgravity environment of space travel reduces bone and muscle formation, and changes immune response and metabolism [ 25 ]. Mechanical niches of stem and cancer cells also regulate their fates such as differentiation and proliferation [ 26 – 27 ]. During aging, rigidity of human epithelial cells markedly increases, which also affects differentiation of their progenitor cells [ 28 – 29 ]. In these biological processes, DNA sequencing identified many expressed genes induced by the signals produced via mechanotransduction in response to mechanical inputs [ 30 – 31 ]. However, what is missing in mechanotransduction research is a complete understanding of how forces are sensed, transmitted, and transduced into gene expression. Fig. 2 illustrates possible mechanotransduction pathways that have been partially revealed as described below. The major obstacle to reveal the molecular mechanism of mechanotransduction and its translation into gene expression is the difficulty in reconstitution of the reaction in vitro, where mechanical force is a difficult parameter to reproduce and disappears once biological samples are lysed. Despite the challenges, methods have been developed to unravel the molecular mechanisms of mechanotransduction. Figure 1 Mechanical forces influence human physiology and pathophysiology. Shear flow in blood vessel influences physiology and pathophysiology of endothelial cells and blood cells (adapted from Nakamura et al [ 4 ] ). Stretching of lung tissue regulates the synthesis of extracellular matrix [ 16 – 17 ]. Exposure to microgravity is associated with atrophy in heart, muscle, and bone, which is also frequented in aging [ 18 – 19 ]. Exercise-induced mechanical stimulus regulates gene expression for muscle fiber hypertrophy [ 20 – 21 ]. Application of mechanical stress on periodontal ligament fibroblasts induces gene expression to regulate the development, differentiation, and maintenance of periodontal tissues [ 22 ]. Hydrostatic or osmotic pressure promotes chondrogenesis of mesenchymal stem cells and the transition and differentiation of notochordal cells into nucleus pulposus cells in the intervertebral disc [ 23 – 24 ]. Figure 2 A schematic overview of how mechanical forces are converted to gene expression in a cell. Some illustrated pathways were experimentally demonstrated. For example, external forces such as touching and stretching are sensed by mechanosensitive ion channels ( e. g. , Piezo channels) [ 32 ]. Internal forces such as actomyosin contraction trigger mechanotransduction as well. Actin cytoskeleton mediates sensing and transmission of forces to regulate nuclear pore size, which controls localization of a trans-acting factor (red) such as Yes-associated protein 1 and megakaryoblastic leukemia 1 [ 33 – 34 ]. Note that this diagram only depicts mechanotransduction pathways to gene expression. Mechanotransduction is also known to induce apoptosis, cell migration, and shape change [ 35 ]. Methods for studying mechanotransduction To understand a biological system, scientists apply "input" into the system and read "output" as a standard approach. For mechanotransduction research, input is the mechanical forces applied to a cell or tissue ( Fig. 1 ). Applying a controlled input of force can be achieved in several ways. For example, tissue culture cells can be repeatedly stretched on an elastic substrate in various directions and frequencies to mimic muscle, blood vessels, and lungs [ 36 – 38 ]. Hydrostatic pressure and membrane-stretching can activate channel proteins [ 39 – 40 ]. Cells can be compressed using a dynamic compressive bioreactor to engineer articular cartilage tissue and to mimic periodontal cells [ 41 – 43 ]. Shear stress can be generated by a pump to mimic blood flow [ 37, 44 ]. Even microgravity can be tested on satellites and space stations, or in a rotating wall vessel bioreactor on the Earth [ 12, 45 – 47 ]. Infrasound (0–20 Hz) and low-frequency noise (20–500 Hz) are also mechanical stimuli that influence physiology of cells and can be applied in research settings [ 48 ]. In addition to these external forces, internal forces can be controlled by inhibiting myosin. Stiffness of the substrate also affects the forces on adhesion molecules that link to the cytoskeleton and the nucleus [ 49 – 51 ]. "Output" in the experimental methods above includes gene expression, morphological changes, translocation of protein, and post-translational modification as these methods are well-established. However, as previously mentioned, these methods do not address how forces are sensed, transmitted, and transduced into gene expression, which have remained challenging questions in the field of mechanotransduction research. How mechanical forces are detected Accumulated evidence demonstrates that mechanical forces trigger conformational changes of proteins to activate them. For example, piezo cation channels use a lever-like mechanogating mechanism to function as a mechanotransduction channel [ 52 ]. Filamin A, an actin cross-linking protein, exposes a cryptic binding site for integrins and other binding partners when actomyosin contraction dissociates the domain-domain pair of filamin A [ 53 – 57 ]. Talin unfolding occurs in the R8 domain upon force application to activate downstream signaling [ 58 ]. Blood shear force can induce unfolding of the A2 domain of von Willebrand factor to expose the binding site for the glycoprotein Iα receptor in the A1 domain, cryptic A disintegrin and metalloproteinase with thrombospondin motifs 13 (ADAMTS13) binding sites, and the cleavage site in the A2 domain [ 59 – 60 ]. Deflection of stereocilia of hair cells by acoustic forces pulls open a calcium channel and activates the current through the channel [ 61 ]. All of these demonstrations rely on detailed structural information as no robust method is available to identify a mechanosensing molecule. A nanotechnology-based method that specifically recognizes mechanosensitive changes without recognizing non-mechanosensitive changes which could take place at the same time as mechanosensitive changes, is necessary to advance our knowledge of force sensing mechanisms. These changes could include not only conformational changes of protein but also other biological molecules such as membrane lipids and sugars. How mechanical forces are transmitted In theory, force transmission can be mediated by cell-matrix interaction, cell-to-cell interaction, cytoskeleton, pressure, and fluidic flow and vibration ( Fig. 2 ) [ 62 – 64 ]. The cytoskeleton-mediated transmission is fast even across long distances compared to molecular diffusion in cells, presumably due to the direct connection of the sensor to a target [ 65 ]. However, in theory, pressure change, stretching, and vibration can also transmit faster than molecular diffusion. The cytoskeleton is directly connected to the linker of nucleoskeleton and cytoskeleton (LINC) complex to regulate gene expression [ 51, 66 – 67 ]. In addition, osmotic shocks stretch the nucleus and nuclear pores to facilitate active transport of Yes-associated protein 1 (YAP), a transcriptional co-factor, into the nucleus [ 68 ]. The only means to identify a force transmitter is the perturbation of a candidate. For example, cytoskeleton can be perturbed by depolymerization agents such as latrunculin and nocodazole [ 69 – 70 ]. The LINC complex can be disrupted by targeting the component of the complex using siRNA and genome editing [ 68 ]. Another sophisticated method is necessary to identify and monitor force transmission at micro or nanoscales. For example, CRISPR screening may facilitate discovery of key mechanotransmitter and rationally designed fluorescent probe may monitor force transmission in real time [ 71 ]. How mechanical forces are converted to biochemical signals The mechanotransduction channels can directly convert mechanical forces into biochemical signals by incorporating ions into cytosol [ 52, 61 ] ( Fig. 2 ). Upon mechanical activation, filamin A can connect to integrin, smoothelin, and fimbacin, and dissociate FilGAP, a Rac-specific GTPase-activating protein [ 54 – 57 ]. The filamin A-integrin interaction regulates cell adhesion and migration, but functions of force-dependent interactions with smoothelin and fimbacin are not known [ 72 ]. Binding of FilGAP to filamin A targets FilGAP to sites of membrane protrusion, where it antagonizes Rac to control actin remodeling [ 73 ]. Mechanical forces also trigger proteolysis by exposing a cleavage site [ 59 – 60 ]. Enlargement of nuclear pore size regulated by mechanical forces such as substrate stiffness, indentation of plasma membrane by atomic force microscopy (AFM), and osmotic shocks enables the transport of YAP from cytosol to the nucleus to regulate gene expression [ 68, 74 ]. Other means to convert mechanical forces into biochemical signals use actin polymerization that is triggered by mechanical forces by unknown mechanisms [ 75 – 79 ]. Polymerization of actin dissociates myocardin-related transcription factor A from unpolymerized actin in cytosol facilitates nuclear localization of mitochondrial transcription factor A to associate with several serum response factor target promoters [ 80 – 82 ]. Although several other molecules such as cadherin, catenin, and merlin are shown to be involved in mechanotransduction pathways, how mechanical forces are exactly converted to biochemical signals through these molecules are not known [ 83 – 86 ]. The major obstacle is the lack of nanoscale structural information before and after activation by mechanical forces. Measurement of mechanical forces Measurement of forces applied to tissue, cells, and eventually a single molecule is the basis of mechanotransduction research and application ( Fig. 3A and B ). However, reported values of cell stiffness and viscosity vary substantially depending on methods even when different groups use the same instruments ( e. g. , elastic and viscous moduli of MCF-7 breast cancer cells can vary 1000-fold and 100-fold, respectively) [ 87 ]. Therefore, scientists need to be aware of the limitations of force measurement and confounding factors such as heat introduced by probes. Unfolding forces of protein domains such as fibronectin type Ⅲ and immunoglobulin domains also vary depending on methods due to different loading forces, loading rate, and other factors [ 88 – 91 ]. AFM measures the forces required to unfold individual domains of titin, ranging from 150 to 300 pN, whereas magnetic tweezers detected the critical force (~5. 4 pN) at which unfolding and folding have equal probability [ 90, 92 ]. This is the case for unfolding of filamin immunoglobulin domains too. The unfolding force ranged from 50 to 220 pN by AFM, whereas magnetic tweezers revealed two different modes of unfolding at <10 pN and >20 pN [ 88, 93 ]. The difference between AFM and the magnetic tweezers is that the loading rate of the magnetic tweezers is much lower than that used in the AFM experiments [ 93 – 94 ]. Although the magnetic tweezers is a choice to measure physical property of biological molecule in physiological condition, AFM can be used to characterize molecule under high-force pulling and to determine molecular-molecular interaction. Figure 3 Application of nanotechnology on mechanotransduction research. A: Measurement of mechanics and mechanical forces of aliving cell. Magnetic tweezers and optical tweezers measure mechanics of a cell using a magnetic bead and microscopic objects coated with specific ligand that attaches to cell surface receptor. Atomic forcemicroscopy (AFM) measures cell mechanics by directly touching a cell. Force sensor fused into adhesion molecule or attached on substrate measure traction force of a cell. Traction force microscopy measuresdisplacement of microbeads embedded in the substrate to determine traction force. Other methods, not shown in this figure, includemicropillar array to detect traction force and fluorescent resonance energy transfer (FRET)-based probe to map stress within a cell. B:Measurement of mechanical properties of a single molecule using magnetic and optical tweezers. AFM can be used for a single molecular analysis but loading rate is much higher than that used in the magnetic and optical tweezers. Fusing an internal control molecule whose mechanical property is known to a test molecule can be used to warrant single molecular analysis. C: Pattern of nanoscale scaffold regulates cell behaviors. D: Stiffness of nanoscale scaffold and mechanical stimulation regulate cell differentiation. For example, nanoparticles can be mechanically manipulated to activate a specific signaling pathway to induce differentiation, growth, and death. The optical tweezers demonstrated that application of 2–5 pN to filamin domain increases the affinity to its binding partners [ 95 ]. The computational calculation of the critical force required to denature an immunoglobulin domain is calculated to be 3. 5–5 pN using the energy difference between the folded and unraveled domain [ 96 ]. Since the forces generated by single myosin or kinesin molecules are 2- to 7-pN force, single and multiple motor molecules are capable of unfolding an immunoglobulin domain [ 97 – 98 ]. Nanoscale molecular sensors can also measure mechanical forces loaded on a single molecule in living cells. Fluorescent resonance energy transfer (FRET)-based probes can be genetically constructed and expressed in living cells. Since FRET changes correlate with forces, force can be measured in the cells. For example, a vinculin probe demonstrated that tension across vinculin in stable focal adhesion is about 2. 5 pN and changes as cells migrate [ 99 ]. In talin, the rod domain experiences a force gradient (in the single piconewton regime) upon integrin-mediated cell adhesion [ 100 ], which is consistent with single molecular analysis [ 101 ]. At cell-cell junctions, it was estimated that cadherin-catenin complex is subjected to a tension of ~5 pN under resting conditions rising to ~50 pN in stressed conditions consistent with experimental measurement [ 102 ]. FRET-based tension sensors can be used to determine traction forces at the cell surface that attaches to substrate as well [ 103 ]. Recently developed nanofiber optic force transducers have the potential to measure intracellular forces with sub-piconewton force sensitivity and a nanoscale footprint [ 104 ]. Manipulation of mechanical forces Nanotechnology is powerful to mechanically manipulate conformational changes of a molecule. For example, magnetic NP is a promising technique for activating a specific signaling pathway, controlling stem-cell differentiation, inducing cancer-cell death, and treating nervous system diseases [ 105 – 107 ]. The NPs can be made in customized sizes and surface characteristics with a high surface to volume ratio, and can be ingested by cells. Magnetic micropillar substrate can also be used for mechanotransduction research and application [ 108 ]. Optical and magnetic tweezers that control from nano- to 2–3 micron-particles can be attached to specific position of a molecule through linkers and internal control such as antibody and green fluorescent protein [ 109 – 111 ]. AFM allows a single molecule to be imaged but also to be manipulated using a cantilever tip [ 112 – 113 ]. Furthermore, high-speed AFM has recently been developed to record dynamic action of biological molecules (currently at 10–16 frames/s) [ 114 ]. These techniques can be applied to mechanically stimulate cells [ 68, 115 – 116 ]. Application of mechanotransduction research in nanomedicine Although mechanotransduction is essential during development and repair of tissues, the difficulty of mimicking the natural properties of tissues is one of the bottlenecks in applying mechanotransduction research to tissue engineering. Nanotechnology through customized nanomaterials has the potential to solve this problem. For example, stem cells can be differentiated into different types of cells by manipulating substrate rigidity and topography [ 117 – 119 ] ( Fig. 3C and D ). More specifically, a defined substrate topography induces chondrogenesis and osteogenesis from human mesenchymal stem cells [ 120 – 123 ]. Culturing myoblasts on aligned nanofibers engineer muscle (myotube) that can be used for skeletal muscle repair and generation [ 124 ]. Moreover, reduced graphene oxide, a low-weight 3D aerogel-like material with pore diameter in the range of approximately 5–10 μm, induces neuronal differentiation of human neuroblastoma cells by modulating RhoA/Hippo pathway [ 125 ]. Doping NPs into hydrogel improves the scaffold mechanics for tissue engineering such as treatment of myocardial infarction and skin scar, bone regeneration, proliferation and differentiation of bone marrow and periodontal stem cells [ 126 – 130 ]. These results suggest that nanotechnology can be used to manipulate matrix and tissue mechanics to control cell fate to repair and generate tissues. In this aspect, an organoid, a simplified and miniaturized version of an organ produced in vitro in 3D, can be a good model to study application of mechanotransduction regulated by nanotechnology [ 131 – 132 ]. Although NP-based diagnosis and therapy are the major topics of nanomedicine, it was recently demonstrated that cellular uptake of NPs are cell mechanics dependent [ 133 ]. Since mechanical properties of normal and diseased cells are significantly different [ 134 ], these results suggest the application of mechanotargeting of NPs in nanomedicine [ 135 ]. Moreover, the success of a shear-activated drug-delivery system inspired by platelet activation suggests that such a system can be applicable to perturb a specific mechanotransduction pathway [ 136 ]. Magnetic NP is a promising tool to remotely activate mechanotransduction. For example, mechanical activation of force-sensitive TWIK related potassium (TREK1 K + ) channel and integrin using magnetic NPs promotes mineralization of bones [ 137 ]. The magnetic NPs can also remotely control brain circuits [ 138 ]. Other nanomaterials possessing particular physical and chemical properties, such as carbon nanotubes and nanofibers can be used for tissue engineering. These materials are biocompatible, stable, easy to fabricate and functionalize, and have a potential effect on neurogenesis, osteogenesis, and stem cell differentiation due to their mechanical properties related to mechanotransduction [ 123 ]. Although application of nanotechnology in mechanotransduction research and tissue engineering is already beginning to happen and show promising results, safety of nanomaterial should be vigorously tested before medical application [ 139 – 140 ]. Conclusion and future perspectives Mechanical forces have a profound effect on human physiology and pathophysiology. Although research on tissue and cellular level is still necessary to understand mechanotransduction, its molecular mechanism at nano level should eventually be revealed for the potential of clinically significant findings. Nanotechnology provides a new set of tools for studying mechanotransduction and for its application in nanomedicine. Such nanomedicine includes force-induced therapeutics, diagnosis, and tissue engineering, which will offer unprecedented opportunities for innovative medicine. However, an understanding of mechanotransduction at the molecular level is still nascent, retaining its application in nanomedicine. Of critical importance is the need to identify a full-set of mechanosensing molecules and reveal how forces are sensed, transmitted, and converted to biochemical signals and gene expression. Such understanding should lead to the development of more selective drugs and treatment.

10. 7603/s40855-015-0001-2

Progress in Stem Cell

Dissecting asthma pathogenesis through study of patterns of cellular traffic indicative of molecular switches operative in inflammation

Background: Inflammation and degeneration are the two edged swords that impale a pulmonary system with the maladies like asthma and idiopathic pulmonary fibrosis. To explore critical role players that orchestrate the etiology and pathogenesis of these diseases, we used various lung disease models in mice in specific genetic knockout templates. Materials and methods: Acute and chronic allergic asthma and idiopathic pulmonary fibrosis model in mouse was developed in various genetic knockout templates namely α4 Δ/ Δ (α41-/-), β2-/-, and α4-/- β2 mice, and the following parameters were measured to assess development of composite asthma phenotype- (i) airway hyperresponsiveness to methacholine by measuring lung resistance and compliance by invasive and P enh by non-invasive plethysmography as well as lung resistance and compliance using invasive plethysmography, (ii) in situ inflammation status in lung parenchyma and lung interstitium and also resultant airway remodelling measured by histochemical staining namely Masson’s Trichrome staining and Hematoxylin&Eosin staining, (iii) formation of metaplastic goblet cells around lung airways by Alcian blue dye, (iv) measurement of Th1 and Th2 cytokines in serum and bronchoalveolar lavage fluid (BALf), (v) serum allergen-specific IgE. Specifically, ovalbumin-induced acute allergic asthma model in mice was generated in WT (wildtype) and KO (knockout) models and readouts of the composite asthma phenotype viz. airway hypersensitivity, serum OVA-specific IgE and IgG, Th2 cytokine in bronchoalveolar lavage fluid (BALf) and lymphocyte cell subsets viz. T, B cells, monocytes, macrophages, basophils, mast cells and eosinophils (by FACS and morphometry in H&E stained cell smears) were assessed in addition to lung and lymph node histology. Results: We noticed a pattern of cellular traffic between bone marrow (BM)→ peripheral blood (PB) → lung parenchyma (LP) → (BALf) in terms of cellular recruitment of key cell sub-types critical for onset and development of the diseases which is different for maintenance and exacerbations in chronic cyclically occurring asthma that leads to airway remodelling. While inflammation is the central theme of this particular disease, degeneration and shift in cellular profile, subtly modifying the clinical nature of the disease were also noted. In addition we recorded the pattern of cell movement between the secondary lymphoid organs namely, the cervical, axillary, ingunal, and mesenteric lymph nodes vis-à-vis spleen and their sites of poiesis BM, PB and lung tissue. While mechanistic role is the chief domain of the integrins (α4 i. e. VLA-4 or α4β1, VCAM-1; β2 i. e. CD18 or ICAM-1). Concluding remarks: The present paper thoroughly compares and formulates the pattern of cellular traffic among the three nodes of information throughput in allergic asthma immunobiology, namely, primary lymphoid organs (PLO), secondary lymphoid organs (SLO), and tissue spaces and cells where inflammation and degeneration is occurring within the purview of the disease pathophysiological onset and ancillary signals in the above models and reports some interesting findings with respect to adult lung stem cell niches and its resident progenitors and their role in pathogenesis and disease amelioration.

Introduction Inflammation is meant to re-establish a shift in the body’s homeostatic balance. Acute inflammation is the initial response to harmful stimuli and is achieved by the increased movement of plasma and leukocytes from blood into injured tissues concomitant with a cascade of biochemical events involving the systemic role of vascular and immune system and local role of other tissue specific cells within the injured tissue (ZF, 2007 ). Prolonged inflammation or chronic inflammation, leads to a progressive shift in the type of cells present at the site of inflammation and is characterized by simultaneous destruction and healing of the tissue from the inflammatory process. This is characterized by concurrent active inflammation, tissue destruction, and attempts at repair and may not be typically characterized by the aforementioned classic signs of acute inflammation. Instead, chronically inflamed tissue is characterized by the infiltration of mononuclear immune cells (monocytes, macrophages, lymphocytes, and plasma cells), tissue destruction, and attempts at healing, which include angiogenesis and fibrosis (Yang, 2006 ). Asthma and COPD although differ significantly in their underlying etiology, involve similar inflammatory changes in the respiratory tract. While the specific nature and the reversibility of these processes largely differ in each entity and disease stage, both are characterized by lung inflammation; though patients with asthma suffer largely from reversible airflow obstruction, whereas patients with COPD experience a continuous decline in lung function as disease progresses (Davidson et al. , 2010 ). By 2020 India alone will account for 18% of the 8. 4 million tobacco related deaths globally (Sharma P, 2009 ). In China, COPD is one of the high frequency causes of death followed closely by Ischemic heart disease and cardiovascular disease (Broide et al. , 1998 ). Inflammation is therefore key to etiology of most respiratory disorders and while it is critical for the body’s defence against infections and tissue damage, it has increasingly become clear that there is a fine balance between the beneficial effects of inflammation cascades and potential for tissue destruction in the long term. If they are not controlled or resolved, inflammation cascades lead to development of diseases such as chronic asthma, rheumatoid arthritis, psoriasis, multiple sclerosis and inflammatory bowel disease (Takizawa, 2007 ). The specific characteristics of inflammatory response in each disease and site of inflammation may differ but recruitment and activation of inflammatory cells and changes in structural cells remain a universal feature. This is associated with increase in the expression of components of inflammatory cascade viz. cytokines, chemokines, growth factors, enzymes, receptors, adhesion molecules and other biochemical mediators. The pathogenesis of allergic asthma involves the recruitment and activation of many inflammatory and structural cells, all of which release mediators that result in typical pathological changes of asthma. The chronic airway inflammation of asthma is unique in that the airway wall is infiltrated by T lymphocytes of the T-helper (Th) type 2 phenotype, eosinophils, macrophages/monocytes and mast cells. Accumulation of inflammatory cells in the lung and airways, epithelial desquamation, goblet cell hyperplasia, mucus hyper-secretion and thickening of submucosa resulting in bronchoconstriction and airway hyperresponsiveness are important features of asthma (Czarnobilska E, 2005 ; Murphy, 2010 ). Both cells from among the circulating leukocytes such as Th2 lymphocytes, mature plasma cells expressing IgE, eosinophils (Murphy, 2010 ) and neutrophils as well as local resident and structural cells constituting the ‘respiratory membrane’ (airway epithelial cells, fibroblasts, resident macrophages, bronchial smooth muscle cells, mast cells etc. ) contribute to the pathogenesis of asthma (Henderson et al. , 2002 ). Airway hyperresponsiveness of asthma is clinically associated with recurrent episodes of wheezing, breathlessness, chest tightness and coughing, particularly at night or in early morning. Furthermore, during exacerbations the features of “acute on chronic” inflammation have been observed. Chronic inflammation may also lead to the outlined structural changes often referred to as airway remodelling which often accounts for the irreversible component of airway obstruction observed in some patients with moderate to severe asthma and the declining lung function. Inflammation in COPD is associated with an inflammatory infiltrate composed of eosinophils, macrophages, neutrophils, and CD8 + T lymphocytes in all lung compartments (Henderson et al. , 2002 ) along with inflammatory mediators such as TNF-α, IL-8 (interleukin- 8), LTB4 (Leucotriene B4), ET-1 (Endothelin-1) and increased expression of several adhesion molecules such as ICAM-1 (Woodside and Vanderslice, 2008 ). The molecular mechanisms whereby inflammatory mediators are upregulated at exacerbation may be through activation of transcription factors such as nuclear factor (NF)- κB and activator protein-1 that increase transcription of proinflammatory genes (Erlandsen et al. , 1993 ). Acute exacerbations have a direct effect on disease progression by accelerating loss of lung functional though the inflammatory response at exacerbation is variable and may depend in part on the etiologic agent (Poole, 2001 ; Troosters et al. , 2010 ). Current therapies for COPD exacerbations are of limited effectiveness (Rennard et al. , 2007 ). Rational treatment depends on understanding the underlying disease process and there have been recent advances in understanding the cellular and molecular mechanisms that may be involved (Rennard et al. , 2007 ). Beyond the absence of curative therapy, current treatment options have inherent limitations, such as further complication by exacerbations, limitations of some orally available treatments and even refractoriness of the most effective treatment regimens such as inhaled corticosteroids (ICSs), long-acting beta2-agonists (LABAs), methylxanthines, leukotriene modifiers, chromones, and IgE blockers. Oropharyngeal adverse events and inadequate response to ICS in a lot of patients present a threat to continued therapy (Calverley et al. , 2007 ). Targeting oxidative damage using antioxidants such as N-acetylcysteine as shown efficacy in chronic bronchitis (Erlandsen et al. , 1993 ) but is relatively ineffective in established COPD (Holgate, 2007 ). Targeting TNF-α to ameliorate inflammation has also been disappointing (Cushley et al. , 1983 ; Nakajima, 1994 ). The use of inhaled steroids combined with longacting β2 agonists to reduce exacerbation rates in more severe disease is now widely accepted, but their effects on mortality are still in doubt (Laberge et al. , 1995 ) and presently there are no effective strategies beyond smoking cessation to slow disease progression in horizon (Schneider et al. , 1999 ). These data suggest that even relatively modest immunomodulators such as inhaled corticosteroids might further impact on local immunity already damaged by chronic inflammation and remodelling, rendering individuals to some degree more vulnerable to significant infections (Chin, 1997 ). Key to effective COPD therapy is prevention of loss of alveolar smooth muscle elasticity which is irreversible by early diagnosis and more effective intervention which is currently virtually non-existent. It is with an objective to identify targets that conventional therapy has obviously overlooked or underrated, that the pattern of cellular traffic is being studied under various patho-physiological situations. A number of genetic knockout models of mice were used and tissue specific (lung) inflammation under asthmatic (Th1-driven) condition was explored and immune cell traffic from their site of poiesis to their site of pathophysiological manifestation were studied (Henderson et al. , 1997 ). In our work with various genotype knockout models of mice, the data generated and interpreted from detailed analyses of cells traveling between bone marrow, peripheral blood, lung parenchyma, airways and the different secondary lymph organs or lymphoid tissues (cervical, axillary, inguinal, mesenteric lymph nodes and Peyer’s patch), we have detected some specific patterns. This is a report on the pattern of cellular traffic from which certain cell subsets were discernible to play rate limiting roles which have been presented along with comments on molecular implications of such directed movement of these key cell types. When specific molecules, critical for signaling the onset and/or development and maintenance of patho-physiology of acute allergic asthma and the associated inflammatory changes, are absent, as in the genotype knockout models, cell trafficking is drastically altered. As apparent from the data presented in this paper, the pattern of cell traffic can actually be molecular signatures for diagnosis of molecular causes of etiology in particular pathological manifestations of acute allergic asthma. The clear rationale for doing the work, that is meta-analysis of data and interpretation of cellular traffic are to understand- when and which cells mobilize from the site of poiesis (primary lymphoid organs); which cells are rate limiting for the sequence of steps required for onset, development, maintenance and exacerbation of acute asthma; which residual cells come back from the focal region of inflammation to secondary lymphoid organs/tissues which may be critical for generating “central memory” cell pool; key diagnostic as well as therapeutic differentiators in the etiology of several respiratory diseases with similar clinical symptoms. Information obtained shall be key to devising therapeutic/prophylactic strategies by targeting small molecules (pharmacological intervention), cells (cell based therapy through tissue engineering), antibody induced neutralization or arrest of cell activation of specific cells (cell targeting) etc. for personalized and translational medical treatment. Unless specific targets are identified in a strict spatio-temporal format, interfering of a target leads to undesirable side effects or even fatality. As outlined in the initial paragraphs, there are patient populations which are refractory to some drugs. So even for designing combination therapy, the timing and targeting is important. The same cell may behave quite differently at different times of the disease onset. The same cell may express different cellular proteins or secrete soluble proteins in its milieu and such changes, rapid and occurring in sequence, are extremely critical information to catch the correct target and at the correct time. Work embodied in this paper attempts to elucidate just these nodes of information throughout in a cellular factory in a specific disease template, acute asthma. Materials-Methods Animals C57BL6 mice were used as described previously (Banerjee, 2014 ; Banerjee and Henderson, 2012a ; Banerjee and Henderson, 2012b ; Banerjee and Henderson, 2013 ; Banerjee et al. , 2009 ; Banerjee et al. , 2012 ; Banerjee et al. , 2008 ; Ray Banerjee, 2011a ; Ray Banerjee, 2011b ; Ulyanova et al. , 2007 ). Mx. cre+α4 flox/flox mice were conditionally ablated by i. p. poly(I)poly(C) injection. cre- mice were used as WT (wild type) and α4 ablated mice were simply called α4-/-. CD18-/- mice on a C57BL6 background were called β2-/-. In total the following number of animals were used in each group: WT= 5 per experiment, +OVA= 5 per experiment, αa-/-= 5 per experiment, β2-/-= 5 per experiment, Rag2γC-/- (baseline)= 4 per experiment, Rag2γC-/- engrafted with WT BMC= 10 per experiment, Rag2γC-/- engrafted with α4-/- BMC= 10 per experiment. A total of three independent experiments for development and analyses of the OVA model and a total of four independent experiments for the engraftment and repopulation experiments in Rag2γC-/- mouse were performed. Data presented are mean ±SEM for all experiments and only p value less than 0. 01 have been considered. Experimental design for lymphopoiesis 5 million bone marrow cells in prewarmed HBSS were injected via tail vein in lethally irradiated (800 cGY) to 6-8 weeks old Rag2γC-/- recipients and reconstitution was followed at 5 weeks, 10 weeks and 6 months. Tissues were collected post sacrifice to assess the type of donorderived versus recipient’s own reconstituted cell types. In the repopulated animals, OVA-induced asthma was induced and composite asthma phenotype noted with detailed analysis of the cellular subtypes in the PLO, SLO and tissues- their structural identity and their functional propensity ( Fig. 1 ). Allergen sensitization and challenge Mice were sensitized and later challenged with OVA (Pierce, Rockford, IL) as described previously. Mice were immunized with OVA (100μg) complexed with aluminium sulfate in a 0. 2-ml volume, administered by i. p. injection on day 0. On days 8 (250 μg of OVA) and on days 15, 18, and 21 (125μg of OVA), mice were anesthetized briefly with inhalation of isoflurane in a standard anesthesia chamber and given OVA by intratracheal (i. t. ) administration. Intratracheal challenges were done as described previously. Mice were anesthetized and placed in a supine position on the board. The animal’s tongue was extended with lined forceps and 50 μl of OVA (in the required concentration) was placed at the back of its tongue. The control group received normal saline with aluminium sulfate by i. p. route on day 0 and 0. 05 ml of 0. 9% saline by i. t. route on days 8, 15, 18, and 21 ( Fig. 2 ). Figure 1. Study protocol for transplantation for hematopoietic reconstitution probing mobilization and homing. Figure 2. Study design to generate acute allergic asthma phenotype in mice. Pulmonary function testing In vivo airway hyperresponsiveness to methacholine was measured 24 hours after the last OVA challenge in conscious, free moving, spontaneously breathing mice using whole-body plethysmography (model PLY 3211; Buxco Electronics, Sharon, CT) as previously described. Mice were challenged with aerosolized saline or increasing doses of methacholine (5, 20, and 40 mg/ml) generated by an ultrasonic nebulizer (DeVilbiss Health Care, Somerset, PA) for 2 min. The degree of bronchoconstriction was expressed as enhanced pause (P enh ), a calculated dimensionless value, which correlates with the measurement of airway resistance, impedance, and intrapleural pressure in the same mouse. P enh readings were taken and averaged for 4 min after each nebulization challenge. Penh was calculated as follows: P enh = [(Te/Tr-1)X (PEF/PIF), where T e is expiration time, T r is relaxation time, PEF is peak expiratory flow, and PIF is peak inspiratory flow X 0. 67 coefficient. The time for the box pressure to change from a maximum to a user-defined percentage of the maximum represents the relaxation time. The T r measurement begins at the maximum box pressure and ends at 40%. BALf After pulmonary function testing, the mouse underwent exsanguination by intra-orbital arterial bleeding and then BAL (0. 4 ml three times) of both lungs. Total BAL fluid cells were counted from a 50μl aliquot and the remaining fluid was centrifuged at 200 g for 10 min at 4°C and the supernatants stored at –70°C for assay of BAL cytokines later. The cell pellets were re-suspended in FCS and smears were made on glass slides. The cells, after air drying, were stained with Wright-Giemsa (Biochemical Sciences Inc, Swedesboro, NJ) and their differential count was taken under a light microscope at 40X magnification. Cell number refers to that obtained from lavage of both lungs/mouse. Lung parenchyma Lung mincing and digestion was performed after lavage as described previously with 100μ/ml collagenase for 1 hr at 37°C, and filtered through a 60# sieve (Sigma). All numbers mentioned in this paper refer to cells obtained from one lung/mouse. Lung histology Lungs of other animals of same group were were fixed in 4% paraformaldehyde overnight at 4°C. The tissues were embedded in paraffin and cut into 5 μm sections. A minimum of 15 fields were examined by light microscopy. The intensity of cellular infiltration around pulmonary blood vessels was assessed by Hematoxylin and Eosin staining. Airway mucus was identified by staining with Alcian blue and Periodic Acid Schiff staining as described previously. Immunohistochemical staining of the lung Lungs of yet other animals of the same group were processed for immunohistochemical staining following standard procedures. They were stained with either anti-VCAM-1 (MK2) or anti β1 (9EG7) antibody and colour development done by HRP. Mouse tissues were prefixed in 4% paraformaldehyde in 100 mM PBS (pH 7. 4) for 6-12h at 4°C, washed with PBS 10 min 3 times and then the tissues were soaked in 10% sucrose in PBS for 2-3h, 15% sucrose in PBS for 2-3h, 20% for 3-12h at 4°C and then embedded in O. C. T. compound (Tissue-Tek 4583, Sakura Finetechnical CO. , Ltd, Tokyo, 103, Japan) and frozen in acetone cooled by dry ice. Frozen blocks were stored at – 70°C refrigerator until sectioned. Frozen blocks were cut on a freezing, sliding macrotome at 4 μm (LEICA CM1850 Cryostat) and air-dried for 30 min at RT. After washing in PBS 3 times for 10 min at RT, to block endogenous peroxidase activity, 0. 3% hydrogen peroxide was applied to each section for 30 min at RT. Each slide was incubated with blocking solution (normal serum from the specific secondary antibody was derived from) to block nonspecific reactions. Appropriately diluted primary antibody was applied to each slide and incubated for overnight at 4°C. After washing with PBS, slides were incubated with appropriately diluted specific biotin conjugated secondary antibody solution for 1h at RT. After washing with PBS, slides were incubated in AB reagent for 1h at RT (AB Complex/HRP, DAKO). After washing with PBS, slides were stained with 0. 05% DAB (3, 3’- diaminobenzidine tetrahydrochloride, Sigma) in 0. 05 M Tris buffer (pH 7. 6) containing 0. 01% H 2 O 2 for 5-40 min at RT. Slides were counterstained with Mayer’s hematoxylin and dehydrated in graded ethanol, xylene and mount with Mount-Quick (Daido Sangyo Co. Ltd. , Japan). CFU-c assay To quantitate committed progenitors, CFU-C assays were performed using methylcellulose semisolid media (Stemgenex, Amherst, NY) supplemented with an additional 50 ng of stem cell factor per ml (Peprotech, Rocky Hill, NJ) to promote growth of hematopoietic progenitors. Next, 0. 01 x 10 6 cells from lung were plated on duplicate 35-mm culture dishes and incubated at 37°C in a 5% CO 2 -95% air mixture in a humidified chamber for 7 days. Colonies generated by that time were counted using a dissecting microscope, and all colony types (i. e. , BFU-E, CFU-E, CFU-G, CFU-GEMM, CFU-GM, and CFU-M) were pooled and reported as total CFU-C. Aliquots of 1-10 x 10 4 cells were plated per 1 ml of semisolid methylcellulose (CFU-lite with Epo, Miltenyi Biotech, or complete human methylcellulose medium, Stem Cell Technologies, Vancouver, BC, Canada). CFU-C frequency was scored morphologically after 10 to 14 days in culture at 37°C, 5% CO 2, in a humidified incubator. Fluorescein-activated cell sorter (FACS) analysis Cells from hemolyzed peripheral blood (PB), bone marrow( BM), bronchoalveolar lavage (BAL), lung parenchyma (LP), spleen, mesenteric lymph nodes (MLN), cervical lymph nodes (CLN), axillary lymph nodes (LNX) and inguinal lymph nodes (LNI) were analyzed on a FACS Calibur (BD Immunocytometry Systems, San Jose, CA) by using the CELLQuest program. Staining was performed by using antibodies conjugated to fluorescein isothiocyanate (FITC), phycoerythrin (PE), allophycocyanin (APC), Peridinin Chlorophyll Protein (PerCP-Cy5. 5) and Cy-chrome (PE-Cy5 and PE-Cy7). The following BD pharmingen (San Diego, CA) antibodies were used for cell surface staining: APC-conjugated CD45 (30F-11), FITC-conjugated CD3(145-2C11), PE-Cy5 conjugated CD4 (RM4-5), PE-conjugated CD45RC (DNL- 1. 9), APC-conjugated CD8(53-6. 7), PE-Cy5 conjugated B220 (RA3-6B2), FITC-conjugated IgM, PE-conjugated CD19 (ID3), PE-conjugated CD21(7G6), FITC-conjugated CD23 (B3B4), APC-conjugated GR-1(RB6-8C5), and PE-conjugated Mac1(M1/70). PE-Cy5 conjugated F4/80 (Cl:A3-1(F4/80)) was obtained from Serotec Ltd. , Oxford, UK. PE-conjugated anti-α4 integrin (PS2) and anti-VCAM-1(M/K-2) was from Southern Biotechnology, Birmingham, Ala. Irrelevant isotype-matched antibodies were used as controls (Banerjee and Henderson, 2012b ). ELISA for cytokines Th2 cytokines (IL-4 and 5) and TNFα and IFNγ in BAL and serum (previously frozen at –70°C) were assayed with mouse Th1/Th2 cytokine CBA (BD Biosciences, San Diego, CA) following the manufacturer’s protocol. According to the manufacturer’s protocol, IL-13 and Eotaxin were measured by Quantikine M kits from R&D Systems, Minneapolis, MN (Banerjee, 2014 ). OVA specific IgE and IgG1 in ser Anti-mouse IgE (R35-72) and IgG1 (A85-1) from BD Biosciences, San Diego, CA were used for measuring OVA specific IgE and IgG1 (in serum previously frozen at –70°C) respectively by standard ELISA procedures as previously described. Results Rationale for the study The study was designed to develop a preclinical model of acute allergic asthma in C57Bl/6J mouse purchased from NIN under permission of the departmental animal ethics committee (approval dated 12-5-2010, renewed on 30-12-2013) and do a meta-analysis of unpublished earlier data generated in University of Washington, USA, (Banerjee, 2014 ; Banerjee and Henderson, 2012a ; Banerjee and Henderson, 2012b ; Banerjee and Henderson, 2013 ; Banerjee et al. , 2009 ; Banerjee et al. , 2012 ; Banerjee et al. , 2008 ; Ray Banerjee, 2011a ; Ray Banerjee, 2011b ; Ray Banerjee, 2007 ; Ulyanova et al. , 2007 ) together with new data generated in the University of Calcutta. Data from two main focused groups of experiments shall be shared and discussed in this work. Cell traffic from PLO to & from SLO to & from pulmonary tissue to & from PLO, post complete manifestation of composite asthma phenotype [some data published in which shall be further analyzed and discussed and new data presented here] and Lymphopoiesis, mobilization, homing and repopulation of PLO and SLO of lethally irradiated Rag2γC-/- recipients from WT (α4+/+) and αa ablated mouse bone marrow and then development (or not) of the composite asthma phenotype and inferences made thereof. [Here again some published data from shall be discussed in context with cellular traffic] α4 ablated and β2-/- mice were developed by other labs (Lee et al. , 2003 ; Scott et al. , 2003 ). α4+ cells in various tissues of α4 f/f and α4 Δ/Δ donors prior to transplantation Fig. 3 shows the distribution of alpha4+ cells in alpha+ versus alpha4 ablated mice. This is in donor mice itself (the mice from which BMC was prepared from femur for engraftment into bloodstream. Understandably PB has the lowest number of alpha4+ cells while surprisingly Peyer’s patch has the highest number of alpha4+ cells in alpha 4 ablated mice. In WT BMC showed a greater homing and engraftment property than the KO BMC transplanted into the recipients. Cells that showed greatest (in terms of shortest time of migration and in terms of the greatest number of viable cells) homing and engraftment were the ones in PP, LNI and MLN. Of note, these are total number of cells including progenitors and differentiated functionally mature cells. These may be therefore labelled as the “leaky” tissues were postablation, α4 expression still occurs, designating the transcriptome in these tissues as being non-permissive to ablation. Figure 3. Distribution of alpha4+ cells in primary and secondary lymphoid organs: donor neonatally ablated Mx. cre alpha4-/- vs. Clonogenic potential of bone marrow cells used in transplantation of Rag2γC-/- recipients Fig. 4 presents clonogenic potential of bone marrow of WT vs. α4-/- mice. α4-/- bone marrow cells show greater colony forming potential as found earlier (Ray Banerjee, 2007 ; Scott et al. , 2003 ; Ulyanova et al. , 2007 ). Figure 4. Donor BM CFU-C assessed before the transplant. Tissue distribution of hematopoietic progenitors posttransplantation and engraftment at 10 weeks As seen from data presented in Fig. 5 PLO show the maximum variation among the genotype groups with or without OVA treatment in the PLO. Compared to WT, in both KO groups α4-/- and β2-/-, there was 2-fold, 2. 3-fold and 2 fold increase in the number of progenitors in BM respectively post-OVA. In circulating blood, however, WT show 1. 7-fold increase in circulating progenitor number compared to 2. 3-fold in both KO groups. There was 3-fold increase in hematopoietic progenitors in spleen of WT compared to 9% and 13% respectively in α4-/- and β2-/- spleen. In the PP, WT and both KO groups show negligible change in progenitor number post-OVA. In tissue, a very curious thing is happening, compared to WT, where post-OVA increase in progenitor number in lung was 10-fold, in α4-/- it reduced half while in β2-/-, the number increased by 16. 5-fold. In the interstitial spaces of the lung, from the BALf, the hematopoietic progenitors that were detected were 4. 7-fold in WT compared to pre-OVA numbers being double in α4-/- but only a 1. 7-fold increase post-OVA. BALf in β2-/- however were a negligible number and obviously demonstrates a mechanistic inability to migrate across the interstitium. Tissue distribution of mature and differentiated hematopoietic cells Repopulation of thymus was significantly impaired in Rag 2–/– recipients of α4Δ/Δ cells, compared to those that received α4f/f donor cells ( Fig. 6A ). A decrease in total cellularity (by 43% at 6 months) was again demonstrable at 8 months post-transplantation, indicating no restorative evidence with time posttransplantation. Double-positive (DP, CD4+/CD8+) population was the predominant one in Rag2–/– recipients of α4Δ/Δ or α4f/f donor cells. The CD4:CD8 ratio greatly favored the CD4 population (w8:1). This suggests that the total repopulation of thymus was impaired, likely because of impaired migration of BM-derived progenitors to thymus, although their subsequent maturation (to DP) was not grossly impaired in the absence of α4 integrins. However, it is notable that CD8+cells were at very low levels in thymus and lower than controls, in contrast to levels in PB (w1. 9:1). Traffic between primary and secondary lymphoid tissue (differentiated immune cells only) (a) Proximal circulation Cellularity in cervical, axillary, and inguinal lymph nodes was similar to controls (i. e. , recipients of α4f/f donor cells). Detailed evaluation of subset distribution showed that there were modestly decreased proportions of mature B cells (B220+IgM+) in all LNs tested, or decreased proportions of activated T cells (CD3+/CD25+, CD3+/CD44+), but their absolute numbers were not significantly different from control groups ( Fig. 6 B-I ). There was a tendency for preferential migration of CD45RC–/CD4+ (memory) cells to lymph nodes, whereas CD45RC+/CD4+ (naive) cells instead preferentially migrated to spleen and thymus in α4 Δ/Δ recipients. In spleen, as noted above, the cellularity, especially of red pulp, was significantly increased and concerned all developmental stages of B cells and of total T cells. (b) Remote circulation In Rag–/– recipients of α4 Δ/Δ cells both at 6 and 8 months post-transplantation, there was a significant reduction in cell numbers recovered from these tissues compared to controls (about 17-fold in PPs, 67% less in MLNs) ( Tables 6B - I ). All subsets of B and T cells were severely reduced in PPs and MLNs repopulated by α4 Δ/Δ cells. The CD4:CD8 ratio in PPs favored a CD4+ profile, as seen in thymus. These data, like the ones in thymus, suggest significant homing impairment of all α4 Δ/Δ cells to these tissues. Rate limiting cells From the above results and detailed cellular subpopulation analyses in Tables 1 - 16, the following data stand out: In blood, WT lymphoid cells before OVA treatment increase to 1. 75 folds after OVA treatment compared to non-lymphoid cells which shows insignificant increase. Whereas in α4-/-, they increase by 2. 3-fold and in β2-/- increase is about 1. 6-folds. In both KO mice, however, there is significant increase in number of nonlymphoid cells post-OVA compared to WT 5-fold in α4-/- and 1. 6-fold in β2-/-. In lung, WT post-OVA lymphoid cells were a 106. 6 fold compared to control but in only 26. 57- fold more myeloid cells migrate to lung parenchyma due to OVA-induced inflammation. In both KO, non-lymphoid cells are 16-fold and 12-fold respectively compared to control. In BALf, there is 40-fold increase in number of lymphoid cells post-OVA in WT and 7. 6-fold increase in myeloid cells, but other than these same cells showing a slight increase (3. 5-fold) post-OVA in β2-/- cells, all KO cells basically failed to migrate across and show poor occurrence even compared to WT control ( Table 1 ). After the evaluation of the total and subpopulations of hematopoietic cells from PLO through blood to tissue site of inflammation (respiratory tissue), the obvious question was what fraction of the blood cells actually get recruited to SLO and thence to the inflamed tissue. Tables 3 onwards attempts to quantitate that in detail. Percent recruited from blood to tissue as shown in Table 2 show a post-OVA preferential recruitment of T cells in α4+ cells (60-fold) to α4- cells (28-fold) and none in β2- cells while B cells show a complete inability to migrate to tissues in both KO (decrease by 3-fold) compared to 2. 5-fold increase in WT cells. Macrophages (GR1-F4/80+) show no significant change in β2-/- compared to WT and α4-/- recruitment (20-fold and 15-fold respectively). As for a population of cells expressing both the myeloid markers, although in WT the number of cells detected were small, there was significant increase in their recruitment from blood (133- fold) but in α4-/- mice, there was no difference in their recruitment post-OVA as opposed to β2-/- that showed feeble recruitment to the tune of thrice the number of cells recruited before OVA treatment which however was 166- fold higher to begin with in both KO groups. As for the scenario in PLO (BM) other than T cells in whose number α4-/- cells show 10. 6-fold increase in synthesis of T cells as opposed to B cells, myeloid cells and indeed all other CD45+ cells compared to WT where an average of 2. 3±0. 36 fold increase in seen post-OVA ( Table 3 ). Tables 4 - 7 (A-D) detail the count of different cell populations and is much more elaborate readout of Table 1. The important highlight here is that lung parenchyma, BALf and trachea being the different structural components of the respiratory tissue, the last seems to be a key node of information throughput and compared to WT and β2-/- mice, α4-/- mice show a much exaggerated CD4:CD8 ratio 5:1 compared to 3:1 in the former. Additionally in all four tissues, Tables 5 - 7C, all three species of B cells, viz. B220+IgM+ mature plasma cells, B220+CD19+ memory cells, B220+CD23+ allergen specific plasma cells are severely decreased in number in β2-/-mice ab initio and although there is significant increase in these numbers after OVA treatment, the overall cell number in this KO mouse in respiratory tissue falls far short of even the numbers of these cells in the placebo treated WT. Table 8B shows that both KO mice show a Th2 skewed phenotype in untreated mice which respond to OVA in terms of ration but neither threshold number nor the composite phenotype is attained indicating that CD4:CD8 may not be absolutely essential for the etiopathophysiology at a critical level. The interesting finding in data presented in these levels however is probably the most important finding of the entire work because of the memory versus naïve or immature T and B cell numbers before and after OVA treatment ( Table 8C, D ). As regards the secondary lymphoid organs and tissues, the first to be considered was CLN as it is anatomically proximal to the lung. Table 9A-C displays an increase in total number of cells 26. 5% and 31. 3% were respectively contributed by T and B cells. Interestingly, post-OVA this increased by 5-folds in α4 Δ/Δ but remained same in β2-/-. Table 1. Values represent total number of different cells and their subsets migrated from bone marrow (BM) via circulation that is peripheral blood (PB) to lung parenchyma (LP) and interstitium (bronchoalveolar lavage fluid- BALf). Recruitment of all leukocytes and their subsets is less in α4-/- lung as well as BALf compared to control. In β2-/- lung, except T cells and eosinophils, all other cells were increased in number. (*P<0. 01 compared with post-OVA control) Recruited T cells post-OVA (both CD4+ and CD8+) were CD45RC negative (memory) in control lung and BAL but mostly CD45RC positive (na ïve) in α4- /- and β2-/- mice. Note the differences between α4 and β2 deficient mice in LP cell content. Mϕ denotes macrophage. (n=12/genotype group). (L=Lymphoid, NL+Non-lymphoid) (x10 6 ) Blood Lung BAL Trachea WT Lymphoid before 8. 01±1. 3 0. 015±0. 001 0. 06±0. 01 0. 004±0. 001 after 13. 9±2. 4 1. 6±0. 05 2. 4±0. 4 0. 25±0. 15 Non-Lymphoid before 3. 7±0. 7 0. 07±0. 001 0. 8±0. 003 0. 013±0. 01 after 4. 3±1. 2 1. 86±0. 06 6. 1±0. 098 0. 06±0. 03 α4-/- Lymphoid before 19±3. 2 0. 015±0. 001 0. 02±0. 009 0. 012±0. 005 after 44±1. 1 0. 3±0. 001 0. 05±0. 001 0. 086±0. 07 Non-Lymphoid before 6. 8±2. 1 0. 05±0. 001 0. 6+0. 016 0. 6+0. 016 after 34. 8±3. 3 1. 022±0. 05 0. 8±0. 001 0. 2±0. 1 β2-/- Lymphoid before 35. 9±5 0. 006±0. 0001 0. 021±0. 01 0. 014±0. 001 after 57. 9±8. 2 0. 104±0. 05 0. 08±0. 03 0. 08±0. 02 Non-Lymphoid before 50. 2±8. 3 0. 43±0. 07 0. 18±0. 001 0. 16±0. 09 after 79. 6±5. 1 5. 18±0. 03 0. 64±0. 03 0. 46±0. 2 Table 2A. Values represent percent cells of blood to be found in various tissues. The percentage blood cells that migrated to interstitial spaces of the lung (BALf). Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin (manufactured by Vision Scientific, South Korea, model Centurion Scientific C2 series) using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer (BD Accuri C6 cytometer) and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. WT α4-/- β2-/- Before After Before After Before After Total cells 7. 29 46. 6 2. 4 1. 03 0. 2 0. 5 T cells 0. 17 14 0. 014 0. 05 0. 02 0. 14 B cells 1. 68 22 0. 28 0. 15 0. 35 0. 136 GR1-F4/80+ 0. 7 27. 6 0. 07 0. 09 0. 4 2. 3 GR1+F4/80+ 13. 5 21. 5 0. 5 2. 3 0. 3 0. 7 GR1loF4/80hi 7. 6 16. 5 0. 5 2. 3 2. 6 7. 7 GR1loF4/80lo 16 13. 8 0. 5 1. 2 0. 5 0. 6 GR1hiF4/80lo 9. 6 73 0. 04 2. 16 0. 03 0. 35 GR1+F4/80- 15 46 0. 6 0. 3 0. 06 0. 045 Table 2B. Values represent percent cells of blood to be found in various tissues. The percentage blood cells that migrated to lung parenchyma. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin (manufactured by Vision Scientific, South Korea, model Centurion Scientific C2 series) using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer (BD Accuri C6 cytometer) and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. WT α4-/- β2-/- Before After Before After Before After Total cells 0. 75 19. 7 0. 27 1. 67 0. 55 3. 87 T cells 0. 26 18. 3 0. 09 1. 1 3 0. 009 B cells 0. 14 0. 54 0. 07 0. 05 0. 16 0. 7 GR1-F4/80+ 0. 05 1. 05 0. 017 0. 11 0. 42 1. 875 GR1+F4/80+ 0. 88 31 0. 48 1. 76 0. 8 5. 6 GR1loF4/80hi 0. 04 6. 15 0. 12 0. 2 5. 4 28 GR1loF4/80lo 1. 9 18. 5 1. 25 0. 89 1. 26 6. 4 GR1hiF4/80lo 7 1. 466 6 0. 9 0. 3 5. 7 GR1+F4/80- 0. 3 39. 5 0. 1 1. 1 0. 1 1. 67 Table 2C. Values represent percent cells of blood to be found in various tissues. The percentage of blood cells that migrated to trachea. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin (manufactured by Vision Scientific, South Korea, model Centurion Scientific C2 series) using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer (BD Accuri C6 cytometer) and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. Before After Before After Before After Total cells 1. 07 1. 35 0. 2 0. 2 0. 3 0. 3 T cells 0. 05 3 0. 013 0. 37 0. 02 0. 08 B cells 0. 07 0. 176 0. 18 0. 06 0. 15 0. 3 GR1-F4/80+ 0. 01 0. 2 0. 04 0. 6 0. 6 0. 5 GR1+F4/80+ 0. 0003 0. 04 0. 7 0. 3 0. 5 1. 5 GR1+F4/80- 0. 001 0. 17 0. 3 0. 4 0. 08 0. 2 Table 3. Lymphoid and myeloid cells in bone marrow. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer and analyzed by BD Accuri C6 software using monoclonal flurochrome tagged antibody as mentioned in Materials & Methods. Genotypes Total (x10 6 ) CD45+(x10 6 ) T cells(x10 6 ) B cells(x10 6 ) Myeloid cells(x10 6 ) WT+alum 28. 07±8. 3 25. 86±7. 6 0. 9±0. 05 2. 485±0. 1 22. 475±10. 3 WT+OVA 54. 13±11. 6 46. 5±2. 6 1. 895±0. 75 6. 49±1. 6 38. 085±9. 2 α4-/-+alum 34. 3±6. 5 28. 94±5. 7 0. 62±0. 2 3. 675±1. 4 24. 645±1. 3 α4-/-+OVA 104. 15±8. 8 85. 34±11. 9 6. 57±0. 5 7. 7±0. 7 71. 07±18. 9 β2-/-+alum 82. 05±11. 5 67. 7±1. 05 2. 7±0. 9 9. 7±0. 54 55. 3±1. 15 β2-/-+OVA 108. 17±13. 2 96. 7±13. 3 4. 15±1. 3 10. 4±1. 6 82. 15±4. 6 Table 4A. Total count (Lymphoid and myeloid cells) in Blood. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer and analyzed by BD Accuri C6 software using monoclonal flurochrome tagged antibody as mentioned in Materials & Methods. (x10 6 ) Total T cells B cells WT Control 11. 92±2. 4 5. 21±1. 3 2. 8±0. 6 OVA treated 18. 22±6. 4 7. 69±2. 4 6. 25±1 α4-/- Control 26. 2±8. 8 13. 8±3. 2 5. 32±1. 7 OVA treated 79. 04±22. 2 18. 5±1. 1 25. 6±7. 5 β2-/- Control 87. 5±19. 8 32. 2±5 3. 7±2. 3 OVA treated 138±60. 1 43. 3±8. 2 14. 6±1. 5 Table 4B. Blood T cell subset (x10 6 ) CD4+ CD8+ CD4+ CD8+ CD4:CD8 memory naive memory naive WT Control 4. 34±1. 7 0. 87±0. 02 0. 06±0. 01 4. 28±2. 3 0. 15±0. 03 0. 72±0. 1 5:1 OVA treated 9. 6±1. 2 0. 67±0. 5 8. 4±2. 4 1. 2±0. 3 0. 5±0. 1 0. 2±0. 01 10:1 α4-/- Control 9. 92±2. 3 0. 8±0. 1 0. 07±0. 001 9. 85±3. 4 0. 03±0. 001 0. 8±0. 1 12:1 OVA treated 26. 3±4. 3 2. 92±0. 6 21. 6±5. 3 4. 7±1. 4 2. 45±0. 8 0. 47±0. 1 9:1 β2-/- Control 21±2. 5 11. 2±1. 5 0. 53±0. 1 20. 5±2. 4 0. 45±0. 02 10. 75±4 1. 8:1 OVA treated 36±3. 7 7. 2±2. 2 29. 7±7 6. 3±0. 9 6. 7±1. 3 0. 5±0. 01 5:1 Table 4C. Blood (x10 6 ) B220+ B220+IgM+ Mature plasma cells B220+CD19+ Memory cells B220+CD23+ Allergen specific plasma cells WT Control 2. 8±0. 6 1. 84±0. 5 0. 27±0. 12 0. 79±0. 2 OVA treated 6. 25±1 3. 6±1. 3 8. 81±2. 6 3. 36±1. 08 α4-/- Control 5. 32±1. 7 5. 6±1. 9 1. 23±0. 7 2. 13±0. 75 OVA treated 25. 6±7. 5 17. 4±3. 5 16. 83±3. 6 17. 8±5. 7 β2-/- Control 3. 7±2. 3 15. 25±6. 7 0. 56±0. 01 0. 2±0. 1 OVA treated 14. 6±1. 5 27. 5±11. 9 8. 71±3. 9 14. 08±5. 6 Table 4D. Blood (x10 6 ) GR1 - F4/80 + GR1 + F4/80 + GR1 + F4/80 + GR1 + F4/80 - Gr1 lo F4/80 hi Gr1 lo F4/80 lo Gr1 hi F4/80 lo WT Control 0. 6±0. 5 2. 6±0. 9 0. 66±0. 01 0. 31±0. 01 0. 014±0. 001 0. 72±0. 13 OVA treated 0. 83±0. 2 3. 2±0. 6 1. 3±0. 1 1. 66±1. 5 0. 26±0. 01 0. 99±0. 02 α4-/- Control 0. 5±2. 1 6. 23±1. 4 3. 3±0. 2 1. 36±0. 25 0. 024±0. 001 0. 78±0. 03 OVA treated 9. 7±3. 3 18. 6±6. 6 11. 08+0. 5 6. 7±12. 9 0. 314±0. 01 10. 7±1. 97 β2-/- Control 0. 74±0. 2 47. 66±8. 3 3. 05±1. 8 7. 1±4. 5 14. 9±5 3. 4±11 OVA treated 5. 2±0. 8 54. 1±3. 04 8. 23±0. 9 19. 9±3. 74 25. 86±2. 6 19. 81±3. 97 Table 5 Total count (Lymphoid and myeloid cells) in BALf. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin (manufactured by Vision Scientific, South Korea, model Centurion Scientific C2 series) using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer (BD Accuri C6 cytometer) and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. Table 5A. BALf (x10 6 ) Total T cells B cells WT Control 0. 87±0. 01 0. 009±0. 0001 0. 047±0. 001 OVA treated 8. 5±0. 4 1. 05±0. 001 1. 398±0. 15 α4-/- Control 0. 62±0. 09 0. 002±0. 001 0. 015±0. 001 OVA treated 0. 82±0. 01 0. 009±0. 001 0. 04±0. 001 β2-/- Control 0. 2±0. 01 0. 008±0. 0001 0. 013±0. 0001 OVA treated 0. 72±0. 03 0. 06±0. 0002 0. 02±0. 001 Table 5B. BALf T cell subsets (x10 3 ) CD4+ CD8+ CD4+ CD8+ CD4:CD8 memory naive memory naive WT Control 7. 8±2. 5 0. 17±0. 01 0 7. 8±2. 2 0 0. 17±0. 01 46:1 OVA treated 1485±89 15±2. 6 1448±54 36. 5±7. 1 14. 5±2. 7 0. 46±0. 01 98. 6:1 α4-/- Control 2±0. 3 1±0. 7 0. 03±0. 001 1. 96±0. 1 0. 05±0. 01 0. 9±0. 01 2:1 OVA treated 7. 7±1. 2 0. 26±0. 1 2. 09±0. 1 5. 6±0. 1 0. 006±0. 001 0. 25±0. 05 29. 6:1 β2-/- Control 5±0. 3 1±0. 4 0. 125±0. 35 4. 875±1. 5 0. 03±0. 001 0. 96±0. 15 5:1 OVA treated 6±0. 2 2±0. 8 1. 69±0. 8 4. 3±1. 7 0. 8±0. 002 1. 16±0. 5 3:1 Table 5C. BALf (x10 3 ) B220+ B220+IgM+ Mature plasma cells B220+CD19+ Memory cells B220+CD23+ Allergen specific plasma cells WT Control 47±1 48. 5±16 5. 94±2. 1 0. 003±0. 001 OVA treated 1398±150 144. 5±67 98. 3±11. 5 102. 7±86 α4-/- Control 15+ 1 16. 2±6. 3 13. 6±6. 4 0. 01±0. 001 OVA treated 40±1 53. 3±12 52±11 42±7 β2-/- Control 13±1 8. 62±1. 5 4. 6±1. 13 0. 002±0. 001 OVA treated 20±1 48. 65±8. 4 31. 5±5. 9 5. 89±1. 8 Table 5D. BALf (x10 3 ) GR1 - F4/80 + GR1 + F4/80 + GR1 + F4/80 + GR1 + F4/80 - Gr1 lo F4/80 hi Gr1 lo F4/80 lo Gr1 hi F4/80 lo WT Control 13±3 350±120 55. 7±11. 6 52. 2±31 33. 05±11 305. 7±15 OVA treated 2100±98 977±98 215±98 302±156 760±2. 5 796±210 α4-/- Control 9. 5±1. 6 29. 8±9. 6 17. 8±1. 7 7. 4±2. 5 0. 1±0. 01 17. 2±2. 6 OVA treated 17. 8±1. 08 426. 5±120 260. 26±34 81±13. 6 6. 8±1. 2 34. 2±11. 9 β2-/- Control 15±1 160±28 80±28 37±1. 5 5±1. 4 20±1. 6 OVA treated 120±31 560±23 250±7 70±13 90±3. 7 25±1. 8 Table 5F. BALf (x10 3 ) Total Lym Mono Mac Mast PMN Eos WT Control 874±15 0 0 874±15 0 0 0 OVA treated 8500±418 2817±145 1141±153 501±21 40±8 784±19 3880±45 α4-/- Control 623±93 0 0 623±93 0 0 0 OVA treated 827±17 0. 6±0. 01 26±0. 4 2. 6±1. 1 0 36. 3±9 0. 9±0. 01 β2-/- Control 204±36 0 0 204±36 0 0 0 OVA treated 718±15 75. 9±12 141. 8±0. 5 64. 9±0. 6 0 66. 06±14 0 Table 6 Total count (Lymphoid and myeloid cells) in LP. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin (manufactured by Vision Scientific, South Korea, model Centurion Scientific C2 series) using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer (BD Accuri C6 cytometer) and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. Table 6A. LP (x10 6 ) Total T cells(CD3+) B cells(B220+) WT Control 0. 09±0. 014 0. 0135±0. 001 0. 004±0. 001 OVA treated 3. 59±0. 1 1. 41±0. 05 0. 034±0. 001 α4-/- Control 0. 07±0. 03 0. 0128±0. 001 0. 0038±0. 0001 OVA treated 1. 322±0. 04 0. 209±0. 01 0. 014±0. 00001 β2-/- Control 0. 44±0. 12 0. 0001±0. 00001 0. 006±0. 00001 OVA treated 5. 35±0. 97 0. 004±0. 0001 0. 1±0. 001 Table 6B. LP T cell subsets (x10 3 ) CD4+ CD8+ CD4+ CD8+ CD4:CD8 memory naive memory naive WT Control 13. 3±0. 85 0. 2±0. 01 0. 95±0. 02 12. 4±0. 07 0. 01±0. 01 0. 2±0. 01 66. 5:1 OVA treated 139±12. 5 10±0. 01 139±0. 7 0. 001 10±0. 1 0. 0001 14:1 α4-/- Control 11. 8±0. 07 0. 5±0. 01 0. 3±0. 01 11. 4±0. 03 0. 064±0. 01 0. 4±0. 01 24:1 OVA treated 202. 8±15 6. 2±4. 8 6. 7±3. 5 196±7. 1 0. 08±0. 03 6. 2±1. 5 33:1 β2-/- Control 0. 16±0. 1 0. 016±0. 001 0. 045±0. 001 0. 115±0. 01 0. 02±0. 01 0. 016±0. 001 10:1 OVA treated 2. 9±3. 3 0. 9±0. 01 1. 8±0. 02 1. 1±0. 005 0. 8±0. 01 0. 1±0. 004 3. 2:1 Table 6C. LP (x10 3 ) B220+ B220+IgM+ Mature plasma cells B220+CD19+ Memory cells B220+CD23+ Allergen specific plasma cells WT Control 4±1 8±5. 5 6. 7±2. 2 0. 002±0. 0001 OVA treated 34±1 28±1. 78 29. 5±1. 8 31. 2±15. 5 α4-/- Control 4±0. 1 6±3. 5 5. 5±1. 7 0. 0001±0. 0001 OVA treated 14±0. 001 18±8. 5 15. 7±3. 6 6. 3±3. 45 β2-/- Control 6±0. 001 6±3. 96 3. 2±1. 25 0. 0001±0. 0001 OVA treated 100±1 94±12. 4 76±13 2. 5±1. 2 Table 6D. LP (x10 3 ) GR1 - F4/80 + GR1 + F4/80 + GR1 + F4/80 + GR1+F4/80 - Gr1 lo F4/80 hi Gr1 lo F4/80 lo Gr1 hi F4/80 lo WT Control 1. 67±0. 5 23±12 0. 28±0. 1 6. 9±5. 6 1. 07±0. 08 6. 45±0. 35 OVA treated 80±13. 5 1430±5. 7 83±0. 4 430±12 50±24 680±78 α4-/- Control 2. 3±0. 85 33. 7±8. 5 4. 2±1. 07 17±0. 8 1. 466±0. 7 3. 5±0. 25 OVA treated 20±0. 2 328±15 24±0. 05 60±0. 45 3±0. 35 118±0. 5 β2-/- Control 20±5. 8 380±13. 5 165±34 90±27 50±19 37±8 OVA treated 97. 5±32. 5 4310±40 930±30 765±65 1480±190 920±80 Table 7 Total count (Lymphoid and myeloid cells) in trachea. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin (manufactured by Vision Scientific, South Korea, model Centurion Scientific C2 series) using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer (BD Accuri C6 cytometer) and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. Table 7A. Trachea (x10 6 ) Total T cells(CD3+) B cells(B220+) WT Control 18. 1±9. 8 2. 8±0. 3 2±0. 7 OVA treated 247. 5±59. 7 236. 8±82. 4 11. 3±0. 1 α4-/- Control 54±2. 37 1. 9±0. 2 10. 6±0. 8 OVA treated 189. 8±5. 8 69. 2±1 16. 9±0. 3 β2-/- Control 283±18. 9 8. 65±0. 4 5. 7±0. 7 OVA treated 474±3. 5 34. 5±2. 35 47. 4±3. 4 Table 7B. Trachea T cell subsets (x10 3 ) CD4+ CD8+ CD4+ CD8+ CD4:CD8 memory naive memory naive WT Control 2. 1±0. 3 0. 7±0. 05 0. 03±0. 0001 2. 07±0. 1 0. 002±0. 001 0. 5±0. 01 3:1 OVA treated 177. 5±0. 8 60±0. 1 169. 8±1. 14 0. 7±0. 03 57. 36±1. 1 2. 5±0. 8 3:1 α4-/- Control 1. 5±0. 03 0. 4±0. 01 0. 001±0. 0001 1. 46±0. 1 0. 001±0. 001 0. 4±0. 01 3. 75:1 OVA treated 57. 7±0. 16 11. 5±0. 02 47. 3±1. 6 10. 4±0. 5 5. 1±1. 6 4. 8±0. 4 5:1 β2-/- Control 6. 2±0. 01 0. 2±0. 04 0. 003±0. 0001 6±0. 05 0. 001±0. 001 0. 2±0. 01 31:1 OVA treated 27±0. 4 7. 5±1. 3 21. 3±7. 5 5. 7±0. 12 3. 6±2. 6 3. 7±0. 5 3. 6:1 Table 7C. Trachea (x10 3 ) B220+ B220+IgM+ Mature plasma cells B220+CD19+ Memory cells B220+CD23+ Allergen specific plasma cells WT Control 2±0. 7 0. 04±0. 001 0. 03±0. 001 0. 01±0. 001 OVA treated 11. 3±0. 1 0. 8±0. 01 1. 5±0. 6 3. 4±0. 5 α4-/- Control 10. 6±0. 8 0. 01±0. 001 0. 06±0. 001 0. 02±0. 07 OVA treated 16. 9±0. 3 0. 6±0. 5 3. 5±1. 5 4. 7±0. 8 β2-/- Control 5. 7±0. 7 0. 1±0. 04 0. 04±0. 001 0. 08±0. 001 OVA treated 47. 4±3. 4 0. 75±0. 01 4. 6±1. 7 5. 2±0. 65 Table 7D. Trachea (x10 3 ) GR1-F4/80+ GR1+F4/80+ GR1+F4/80- WT Control 0. 2±0. 02 1. 6±0. 5 12±2. 5 OVA treated 18±1. 7 34. 5±0. 5 9. 8±2. 4 α4-/- Control 0. 03±0. 001 41. 5±13 15±0. 75 OVA treated 7±0. 5 56±12. 5 162. 6±8. 9 β2-/- Control 0. 05±0. 001 135. 5±20 27±0. 5 OVA treated 9±2. 8 335±18. 5 123±7. 5 Table 8 Total count (Lymphoid and myeloid cells) in spleen. Number of cells in PLO and SLO and tissue were calculated as tC and DC by hemocytometric analysis in a standard Neubauer’s Hemocytometer, DC was analyzed by double blind counting from H&E stained smears prepared in a cytospin using Zeiss photo allotment and Axiostar plus software and by flow cytometry using BD flow cytometer and analyzed by BD Accuri C6 software using monoclonal fluorochrome tagged antibody as mentioned in Materials & Methods. Table 8A (x10 6 ) Total T cells(CD3+) B cells(B220+) WT Control 83. 95±25 38. 87±21 43. 08±2 OVA treated 113±21. 83 58. 56±13. 2 67. 02±6. 03 α4-/- Control 123. 81±18. 97 47. 73±7. 6 63. 86±12. 6 OVA treated 358. 8±8. 5 59. 31±15 76. 96±12 β2-/- Control 142. 5±42. 98 52. 94±5. 87 71. 03±3 OVA treated 290. 84±54 69. 22±15. 2 87. 54±19. 24 Table 8B. Spleen T cell subsets (x10 6 ) CD4+ CD8+ CD4+ CD8+ CD4:CD8 memory naive memory naive WT Control 25. 33±0. 323 10. 53±2. 01 1. 05±0. 37 20. 13±1. 19 0. 2±0. 07 10. 33±4. 5 2:1 OVA treated 52. 82±9. 05 8. 79±2. 2 39. 63±7. 79 12. 19±2. 53 6. 56±0. 78 2. 23±0. 76 3. 7:1 α4-/- Control 34. 25±6. 56 5. 18±1. 02 3. 76±5. 04 30. 49±5. 89 0. 41±0. 01 4. 77±2. 1 6. 6:1 OVA treated 30. 59±10. 09 14. 46±4. 3 26. 3±8. 43 4. 23±1. 19 12. 31±9. 5 2. 15±1. 08 2:1 β2-/- Control 34. 8±0. 06 6. 93±0. 04 0. 696±0. 43 34±7. 8 0. 22±0. 17 6. 71±3. 9 5:1 OVA treated 56. 39±0. 67 11. 9±2. 67 28. 27±2. 13 25. 12±4. 8 8. 05±2. 2 5. 22±1. 3 4. 7:1 Table 8C. Spleen (x10 6 ) B220+ B220+IgM+ Mature plasma cells B220+CD19+ Memory cells B220+CD23+ Allergen specific plasma cells WT Control 43. 08±2 25. 3±1. 8 1. 018±0. 56 0. 56±0. 12 OVA treated 67. 02±6. 03 39. 4±4. 8 54. 69±32. 16 21. 25±8. 1 α4-/- Control 63. 86±12. 6 42. 8±15. 3 15. 4±3. 6 1. 71±0. 85 OVA treated 76. 96±12 161. 3±67. 8 51. 87±6. 95 28. 76±4. 6 β2-/- Control 71. 03±3 65. 5±36. 7 2. 2±1. 5 1. 65±0. 53 OVA treated 87. 54±19. 24 104±65. 3 70. 7±11. 7 57. 69±31. 27 Discussion Figures 1 & 2 show the diagrammatic representation of the basic study protocols of transplantation for hematopoietic reconstitution and development of preclinical asthma respectively. C57Bl/6J mouse from NIN under permission from IAEC (University of Washington, Seattle, USA) and Departmental Animals Ethics Committee (Dept. of Zoology, University of Calcutta, India). Using this basic template of the study protocol of acute allergic asthma, migration of cells under no-disease condition (“clean” Rag2γc-/- where progenitors were killed off by lethal γ-irradiation and bone marrow cells from donors were allowed to repopulate all PLO and SLO) and under diseased condition where two genotype knockout mice were used and acute allergic asthma induced to track and study the various nuances of cell migration in the different PLO and SLO and their subsequent recruitment to the pulmonary tissue for orchestrating inflammation. At the outset, we must consider the significance of the donors that were chosen and the recipients’ pretreatment before transplantation. It may be conclusively deduced from Fig. 3 that some SLO are resistant to ablation of this integrin. The reason may be important metabolomic and transcriptomic pathways that disallows this resulting in the immunosecretome being readjusted and repositioned such that ablation of α4 is compensated. Fig. 4 however, unambiguously excludes BMC from such non-permissiveness to ablation ensuring a 100% αa free reconstitution. As seen in previous publications, (Banerjee et al. , 2009 ; Banerjee et al. , 2008 ; Ray Banerjee, 2007 ; Scott et al. , 2003 ), α4 has regulatory roles in mobilization, homing, and engraftment which explains their higher progenitor number in PLO (BM), circulation (PB), and spleen (SLO) in the KO groups but not appreciable numbers post OVA as was found in the WT ( Fig. 5 ). This is key. In all other secondary lymphoid organs assessed, namely, MLN, LNI and LNX, except CLN, pattern of cellular traffic in WT vs. KO groups and control vs. OVA treated groups do not follow a uniform pattern or any “trend”. The complete oppositeness of progenitor number modulation post-OVA in the two genotype groups indicate their diverse pathways action as enumerated in (Banerjee et al. , 2009 ; Ray Banerjee, 2007 ; Ulyanova et al. , 2007 ). But what is interesting here is the complete separation of the cell traffic (direction and number) as evident from Tables 1 - 8 and Fig. 6A - H. The picture of Tand B-cell distribution in spleen is more in line with what is present in PB, and contrasts that of BM described above most likely because of longer retention and maturation of these cells in the splenic environment compared to BM. CD40+ (dendritic cells) were lower in all organs except the spleen, where the proportion, but not the total number, was low. Fig. 7 shows the pathways of induction of differentiation of various mature functionally active immune cells that are key to inflammation. Additionally, the figure also outlines the switch from progenitor (pluripotent cells) towards lineage commitment and final differentiation into competent immune cells responsible for the immune activities within tissues. Important to note are the cytokines and other factors that not only induce differentiation but also attracts the cells from their sites of synthesis (PLO) to their sites of maturation (SLO) down to the tissues where they either perish in the onslaught or return to PLO or SLO (homing) with valuable information to be encoded as “central memory” cells that either become the effectors as and when they are recalled (mobilization) in the future during another exigency. Fig. 8 represents ramifications of the network of growth factors and their receptors, adhesion and signalling molecules and transcription factors, maturation, morphogenic and their guidance molecules, ECM proteins and proteinases. Fig. 9 shows schematically, the various functional cells populating the PLO, SLO and migrating to tissues, the known immune system lymphoid and myeloid subpopulations and their inducing cytokines and growth factors. Fig. 10 depicts the major PLO, SLO in humans and that denotes their exact counterparts in mouse as they will be extrapolated to represent. Figure 5. Progenitors in different tissues of the three genotype groups. Mortality being low and repopulation of transplanted cells being 100% by 48 weeks, we can say safely that the transplanted BMC did their job well, that is they found the regions to home to and then they settled down there and all progenitors generated new clones and differentiated into the type of cells that the tissues needed. In other words there was induction of differentiation in situ. As for the rate limiting steps which are key in cellular traffic resulting in onset and development of the actual disease, α4 was found to be a critical regulatory factor for myeloid cellular traffic compared to lymphoid. ( Table 1 ) Overall, both integrins seem to have regulatory roles in lymphoid and nonlymphoid cell migration to blood. The number of cells in BALf is the most striking and as described in (Banerjee et al. , 2009 ; Banerjee et al. , 2008 ; Ray Banerjee, 2011b ; Ulyanova et al. , 2007 ) that while α4 principally controls sensitization and signaling, β2 mainly control the migration from lung parenchyma to interstitium without which onset of the asthma immunopathology cannot occur. Table 2 reveals a curious thing. A subpopulation of cells GR1+F4/80+ (GR1 hi F4/80 hi ), probably newly formed macrophages migrated to pulmonary tissue from BM, which is a minus percent recruited from blood in normal untreated mice, shoot up in both KO groups, notwithstanding their lack of recruitment post-OVA. Fig 6A. Cellularity in thymus Figure 6B. Distribution of different development stages of B and T cells in BM, PB and Spleen. Graph of cell markers in BM, PB and Spleen, CD3+ Figure 6C. Distribution of different development stages of T cell subsets in BM, PB and Spleen; Graph of cell markers in BM, PB and Spleen, CD4+, CD8+ Figure 6D. Distribution of different developmental stages of T cells in BM, PB and spleen. Graph of cell markers in BM, PB and Spleen, CD4+CD25+ Treg. Figure 6E. Distribution of different developmental stages of T cells in BM, PB and spleen. Graph of cell markers in BM, PB and Spleen, Progenitor B cells Figure 6F. Distribution of different developmental stages of T cells in BM, PB and spleen. Graph of cell markers in BM, PB and Spleen, Pre B cells B220+CD34- Figure 6G. Distribution of different developmental stages of T cells in BM, PB and spleen. Graph of cell markers in BM, PB and Spleen, Pro B cells – Early B cells Figure 6H. Distribution of different developmental stages of T cells in BM, PB and spleen. Graph of cell markers in BM, PB and Spleen, Mature B cells (B220+IgM+) Figure 7. Mediators of differentiation pathways from stem cells. Figure 8. Ramifications of myriad factors operative in inflammation in complex network. Figure 9. Schematic representation of the key inflammasomes and their pathways. Figure 10. Summary findings and conclusion - Tentative scheme of cellular traffic. This may indicate either a sub-clinical inflammation in the KO mice which is spontaneous in nature or it may be cells newly recruited from bone marrow with other more extraneous and somewhat fortuitous functions. This group need better characterization. The GR1+F4/80- myeloid population which is definitely the new recruits from BM which have still not lost their GR1 expression may be short-lived granulocytes which show similar trend as this preceding group. The significance may be better grasped in the following tables where trends in PLO top SLO migration become clearer. The rate limiting cells are therefore lymphoid cells as they migrate from blood to lung tissue and thence to the interstitial spaces where the actual gas exchange occurs and this step is almost exclusively controlled by the β2 integrin. α4 controls the lymphoid signaling while β2 controls myeloid migration. So these are the rate limiting cells from PLO to tissue. Table 3 reveals conclusively that although in α4-/- BM, synthesis and sequestration of T cells fall quite tremendously and post-OVA their number is amplified manifolds more than in WT. As mentioned in (Scott et al. , 2003 ), this may either be due to a lack of “settling down” on the stroma of the PLO or a genuine increase in synthesis of these cells but it certainly means that there is possibly no impairment in the upstream IL-4-IL-4R and IL-5-induced recruitment from bone marrow in its absence. Data shown in Table 7D coupled with data presented in Fig. 6A highlight that α4 is probably key to the Th1-Th2 balance. In Table 5 - 7C, data shows that although IL-4/5 induced cell recruitment is probably still operative despite absence of β2- cells from these B cells, they are unable to cross the number threshold and therefore insufficient to mount a viable pathophysiological circuit. (Banerjee et al. , 2008 ; Farrell, 2003 ; Lee et al. , 2003 ; Ray Banerjee, 2007 ; Scott et al. , 2003 ; Ulyanova et al. , 2007 ; Weber et al. , 1992 ). There may be a role for epithelial progenitors, circulating and niche-dwelling (Gomperts et al. , 2006 ) which is beyond the scope of this paper and is being currently investigated. Summary findings & conclusion- Highlights Observations α-/- cells are slower to reconstitute Increased hematopoiesis in all primary & secondary lymphoid organs in KO gr Myeloid repopulation was similar in both groups B cells, pro B cells, pre B cells, mature B cells significantly decreased in KO gr Total T cells increased but activated T cells decreased in KO gr Repopulation of thymus was significantly impaired in KO gr CD+CD8+ was more in thymus and CD+ differentiation was skewed in KO gr OVA-induced acute allergic asthma phenotype was impaired in all aspects All PLNs show decreased mature B cells Enhanced tendency of memory T cells (CD45RC-CD4+) to migrate to LNs in KO gr Enhanced tendency of memory T cells (CD45RC+CD4+) to migrate to spleen & thymus Functionality & homing impairment in KO gr α-/- probably do not have a role to play during response to OVA per se, and neither to specifically T cell proliferation in response to the upstream cytokine signaling in response to OVA. Salient findings and clinical implications Various small molecules inhibitors for α4 integrin such as Tysabri (Roche) and integrin Beta2/CD18 antibodies products manufactured by Biolegend and Novus Biologicals and others which are specifically designed for modulation of the target molecule must take into account targeting strategies along with the timing of drug activation in a particular tissue (in pro drug formulation, through nano-vehicles for timed and tissue specific delivery. Positioning of PLO, SLO and inflamed tissue as nodes of cellular and information traffic and the network operative in specific temporal templates is crucial information for the success of such drugs. The role of α4 and β2 are very much cell specific and tissue specific, for example their role in bone marrow, blood, lung, interstitial tissue spaces, trachea, and lymph nodes are not uniform. The “kicking in” is initiated possibly by α4 during initiation and “priming” of the B cells and T cells. The “changing gear” is possibly orchestrated by β2 during the effector phase of the disease development. It is through the detailed mapping of such “switching on – switching off” information throughput of the cellular and non-cellular role players, translational and personalized medicine may devise effective strategies in disease amelioration. Conclusion α4 &β2 integrins control cellular migration of all lymphoid subpopulations from BM (site of poiesis) to PLO and of some sub-populations from PLO to SLO but has no effect on myelopoiesis or homing and functionality.

10. 7717/peerj. 10259

PeerJ

Polylactic acid as a suitable material for 3D printing of protective masks in times of COVID-19 pandemic

A critical lack of personal protective equipment has occurred during the COVID-19 pandemic. Polylactic acid (PLA), a polyester made from renewable natural resources, can be exploited for 3D printing of protective face masks using the Fused Deposition Modelling technique. Since the possible high porosity of this material raised questions regarding its suitability for protection against viruses, we have investigated its microstructure using scanning electron microscopy and aerosol generator and photometer certified as the test system according to the standards EN 143 and EN 149. Moreover, the efficiency of decontaminating PLA surfaces by conventional chemical disinfectants including 96% ethanol, 70% isopropanol, and a commercial disinfectant containing 0. 85% sodium hypochlorite has been determined. We confirmed that the structure of PLA protective masks is compact and can be considered a sufficient barrier protection against particles of a size corresponding to microorganisms including viruses. Complete decontamination of PLA surfaces from externally applied Staphylococcus epidermidis, Escherichia coli, Candida albicans and SARS-CoV-2 was achieved using all disinfectants tested, and human adenovirus was completely inactivated by sodium hypochlorite-containing disinfectant. Natural contamination of PLA masks worn by test persons was decontaminated easily and efficiently by ethanol. No disinfectant caused major changes to the PLA surface properties, and the pore size did not change despite severe mechanical damage of the surface. Therefore, PLA may be regarded as a suitable material for 3D printing of protective masks during the current or future pandemic crises.

Introduction COVID-19 (coronavirus disease 2019) is the designation of the disease caused by the SARS-CoV-2 infection. The World Health Organization (WHO) declared this epidemic a global pandemic affecting the whole world on 11 March 2020. The infection by SARS-CoV-2 was confirmed for the first time in Wuhan, China, but had a huge impact also in Europe and later in North and South America. Lombardy, Italy was the most severely affected region in Europe. Due to the risk of health care system collapse, the Italian government ordered a nationwide lockdown ( Spinelli & Pellino, 2020 ). Several studies showed that SARS-CoV-2, similarly to SARS-CoV-1, remains infectious for hours and days in aerosols and on surfaces, respectively ( Chin et al. , 2020 ; Kampf et al. , 2020 ; Van Doremalen et al. , 2020 ), emphasizing the need for efficient virucidal disinfection. The number of patients suffering from COVID-19 disease and the enormous rate of infection spread caused serious complications in many countries, including a desperate lack of protective equipment ( Swennen, Pottel & Haers, 2020 ). Sufficient production and distribution of protective equipment has been crucial for sustaining patient care during the pandemic. The current unsatisfactory situation regarding protective equipment in the USA has been described by Ranney, Griffeth & Jha (2020). Because of the lack of protective equipment including face masks, extended manufacturing facilities have become very important for supporting the health care system. In this regard, the production of protective masks using 3D printing has proven very promising. This technology, often based on Fused Deposition Modelling (FDM) due to its cost and technical benefits, has found various applications in the manufacturing of medical devices such as prosthetic and dental implants or scaffolds in tissue engineering ( Roopavath & Kalaskar, 2017 ; Tack et al. , 2016 ). The properties of 3D-printed objects render this technology attractive for manufacturing of protective masks. FDM provides adequate dimensional control, good surface finish and adaptability to use a variety of thermoplastic polymer filaments. The technology is based on high-temperature sintering of filaments and subsequent solidification of the printed product at room temperature. The polymers most commonly used for FDM are acrylonitrile-butadiene-styrene copolymers, polycarbonate, polyethylene terephthalate glycol (PETG) and polylactic acid (PLA) ( Chadha et al. , 2019 ; Ngo et al. , 2018 ). Due to its unique properties, PLA is one of the most attractive materials for 3D printing. Its main advantages include low printing temperatures of 200–210 °C, smooth appearance, low toxicity and favorable mechanical properties, especially a low warping effect and high geometric resolution ( Pajarito et al. , 2019 ; Vicente et al. , 2019 ). PLA is a biodegradable linear aliphatic polyester produced from renewable natural resources such as corn, wheat or sweet sorghum ( Nampoothiri, Nair & John, 2010 ). Nagarajan, Mohanty & Misra (2016) comprehensively reviewed its properties and applications. This polymer is produced by acid-catalyzed polycondensation of lactic acid monomers. Lactic acid of any chirality can be used, resulting in either poly-L-lactic acid, poly-D-lactic acid or poly-L, D-lactic acid (consisting of both isomers). Since L-lactic acid is the most common isomer in nature and is easily produced by lactic fermentation of various bio-wastes by bacteria (e. g. , Lactobacillus spp. ), it is also the most commonly used precursor for PLA manufacturing. The possibility of biotechnological production of the monomer significantly decreases its price, making the production of PLA very cheap. The glass transition temperature of PLA ranges between 50 and 80 °C, and the melting temperature reaches approximately 175 °C. Due to its natural precursor, PLA is easily biodegradable, for example, by thermal decomposition, enzymatic digestion, oxidation or photolysis. Ghorpade, Gennadios & Hanna (2001) studied the outcome of PLA-composting for 90 days and found that the compound was degraded by 70 %. The use of PLA is limited by its poor thermal stability and easy hydrolysis—it degrades more easily than other aliphatic polyesters. Nevertheless, PLA has found many applications in diverse areas including the packaging industry as a food packaging polymer for short shelf life products, the pharmaceutical industry for controlled drug delivery formulations and for tissue regeneration, and agriculture for better herbicide delivery management without negative effects on crop yield ( Auras, Harte & Selke, 2004 ; Aziz, Haq & Raina, 2020 ; Farto-Vaamonde et al. , 2019 ). Protective masks made by 3D printing from PLA are designed for repeated use, requiring frequent cleaning and disinfection. The low glass transition temperature and relatively low melting point of PLA makes heat sterilization in an autoclave at 121 °C impossible ( McKeen, 2014 ). The polymer can be sterilized using ethylene oxide, gamma radiation ( Fleischer et al. , 2020 ) or dry heat below 80 °C for no more than 20 min ( Zou et al. , 2011 ). Fleischer et al. (2020) examined the changes of PLA properties after cleaning with chemical disinfectants such as Cidex Opa (Johnson & Johnson) or chlorine solutions. Although these substances caused minor changes in stiffness and strength of 3D-printed PLA, 3D printing at appropriate conditions makes PLA objects mechanically amenable to cleaning and reuse. However, surface porosity of 3D-printed PLA medical tools should be minimized to prevent exposure of users to residual disinfectants by inhalation or skin contact. Oth et al. (2019) studied PLA object sterilization by low-temperature hydrogen peroxide gas plasma in the commercially available Sterrad ® apparatus (Johnson & Johnson). They observed only sub-millimeter deformations induced by this process, rendering it suitable for sterilization in different areas including surgical applications. In contrast to conventional steam autoclaving, sterilization by hydrogen peroxide prevents deformation of 3D-printed objects made from PLA or PETG. Swennen, Pottel & Haers (2020) presented a prototype of reusable custom-made 3D-printed face masks (produced by a selective laser sintering technique) from polyamide composite components. The authors proposed cleaning by 15 min exposure to a broad-spectrum antimicrobial solution, ANIOS CLEAN EXCEL, containing didecyldimethylammonium chloride and chlorhexidine digluconate. Nevertheless, material leakage and virus decontamination of the reusable face mask components have not been tested upon one or more disinfection cycles. In the present study, we have investigated FDM 3D-printed PLA structure and porosity after exposure to common chemical disinfectants including ethanol, isopropanol and a commercial disinfectant containing sodium hypochlorite, which are easily accessible. In addition, we examined the efficiency of PLA disinfection after artificial contamination with bacteria ( Staphylococcus epidermidis, Escherichia coli ), a yeast fungus ( Candida albicans ), viruses (SARS-CoV-2 and human adenovirus – HAdV) or natural contamination by wearing the masks. Materials and Methods PLA material and masks preparation by 3D printing Polylactic acid (PLA) was purchased in the form of filament for FDM 3D printing from Shenzen Creality 3D Technology Co. , LTD, China. Protective masks, circular plates (diameter of 10 cm and height of 0. 2 cm, printed vertically) and square carriers (1 × 1 cm, 0. 2 cm high) ( Fig. 1 ) were prepared using a 3D printer (Prusa i3 MK3, Czech Republic). The printing template was designed with Trimble Sketchup Pro, exported in a stereolithographic (stl) file (freely available at https://www. facebook. com/groups/1346383268879783/files/ ) and used to print the objects of investigation. The printing parameters were as follows: layer height = 0. 3 mm, shell thickness (perimeter) = 0. 4 mm, bottom/top thickness = 0. 2 mm, fill density = 10%, print speed = 90 mm/s, extrusion temperature = 215 °C, platform temperature = 60 °C, filament flow = 95%, machine nozzle size 0. 4 mm, the infill pattern was grid (i. e. , linear tilted 45°) and the total layers = 338. 10. 7717/peerj. 10259/fig-1 Figure 1 Objects made from PLA filaments using 3D printing by the FDM technology. (A) PLA carriers (1 × 1 cm). (B) Circular plate with a diameter of 10 cm (printed vertically). (C–E) Different types of PLA masks. Visualization of PLA mask structure using scanning electron microscopy The structure and porosity of PLA 3D-printed masks were examined using a scanning electron microscope (SEM) Nova NanoSEM 450 (Fei, USA). Approximately 1 × 1 cm pieces cut from printed masks were completely air-dried and visualized by SEM. Since the material is very sensitive to electron exposure, mild conditions had to be used, that is, voltage of 5 kV and low vacuum. Images of each visualized position were captured by LVD detector at gradual magnifications 2, 000×, 1, 000×, 500×, 100× (focusing on identical position), dwell time 5 µs and spot size 4. 5. The size of the pores between PLA filaments was measured and marked using a SEM operating software (xT microscope Control v6. 3. 4 build 3233), provided by the SEM manufacturer (Fei, USA). The SEM images shown in this study were selected as representative visualizations of the PLA microstructure. The SEM analysis merely illustrates the 3D-printed PLA surface morphology, and gap width measurements are not analyzed statistically. Visualization of PLA mask structure under stress conditions using scanning electron microscopy To investigate the impact of possible stress factors for PLA masks, cleaning with chemicals was performed and exposure to wearing-associated contamination was simulated, as outlined below. The effect of immersing in three chemical disinfectants (96% ethanol, 70% isopropanol and the commercial disinfectant and bleach SAVO Original, Unilever ČR s. r. o. , Czech Republic containing 0. 85% sodium hypochlorite diluted with water (2:9)) was tested by repeated (5 × 15 min) cycles and long-term (24 h) exposure. The simulation of human impact on the PLA structure was performed as follows: extensive exposure to fingers (to simulate incorrect application of the mask), abrasion with paper (minor mechanical stress) and dining fork (strong mechanical stress), immersion in 1. 9% sodium chloride solution for 4 h (to simulate perspiration). Short rinsing with 100% acetone was examined to investigate its effect on PLA surface properties. After each treatment, completely dried PLA carriers were examined using SEM, as described above. Aerosol particle passage through PLA material Surface contamination (and potential surface penetration) with infectious agents was simulated by an aerosol generator and a photometer (Lorenz Meβgerätebau FMP 03) with a differential pressure sensor ( Fig. S1 ), providing a suitable test system with stand for facial masks and flat filter materials. The device was certified as a test system according to the standards EN 143 (Respiratory protective devices—Particle filters—Requirements, testing, marking), and EN 149 (Respiratory protective devices—Filtering half masks to protect against particles—Requirements, testing, marking). A sample of the PLA material was attached in the standardized testing cartridge and was sealed with silicone to prevent false positive detection of penetrating particles passing along the edge of the PLA panel ( Fig. S2 ). The cartridge was mounted into the Lorenz Meβgerätebau FMP 03 device, between the aerosol generator and photometer. The aerosol generator produced a defined amount of aerosolized paraffin oil, the test system passed it through the material, and the photometer situated on the other side of the PLA sample measured the aerosol concentration, thereby indicating the retention efficiency. An integrated differential pressure sensor was used to determine the pressure loss during passage through the sample. The particle size distribution was approximately 0. 1–2 µm (geometric mean 0. 44 µm), which is close to the most frequently observed penetrating particle size ( Fig. S3 ). The output of the aerosol generator was set to 150% with flow 95 L/min, atomizer pressure 5 bar and oil temperature 60 °C. The test was performed for 270 s. Disinfection of PLA material artificially contaminated with bacteria and yeast fungus Wild strains of S. epidermidis, E. coli and C. albicans were used as representatives of gram-positive and gram-negative bacteria or yeast fungus, respectively. The concentration of bacteria was adjusted to approximately 1 × 10 7 colony forming units (CFU) per mL, the fungus concentration was 1 × 10 6 CFU/mL. Each PLA carrier with a size of 1 × 1 cm was contaminated with 10 µL of microbial suspension applied to the surface of carriers in 1 µL droplets for 1 h. The disinfection of contaminated carriers was carried out by immersing in three mL of 96% ethanol, 70% isopropanol, or 0. 85% sodium hypochlorite (SAVO Original, Unilever ČR s. r. o. , Czech Republic) for 15 min. After evaporation of disinfectant solutions, the carriers were immersed in one mL of sterile 0. 9% saline, vortexed, and the obtained suspensions were inoculated onto appropriate agar plates. Blood agar was used for S. epidermidis, Müller-Hinton (Oxoid, Czech Republic) agar for E. coli and Sabouraud agar (Oxoid, Czech Republic) for C. albicans. Samples not exposed to treatment by disinfectants were used as controls. The inoculated plates were incubated at 37 °C for 48 h. Each experiment was done in triplicate, and results were obtained by counting the average CFU/mL. Disinfection of PLA material artificially contaminated with viruses SARS-CoV-2, the causative agent of the COVID-19 pandemic, was isolated in a biosafety level 3 laboratory from a nasopharyngeal swab by inoculating Vero CCL81 cells (ECACC 84113001) and subsequent expansion by two additional passages in Vero CCL81 cells. Passage 3 was cleared by centrifugation at 1000 g for 5 min, passed through a 0. 45 µm filter, and stored at −80 °C until use. In addition to SARS-CoV-2, inactivation of a stable DNA virus, the Human Adenovirus 2 ATCC VR-846 (HAdV) obtained from the American Type Culture Collection (ATCC) was assessed. Similar to the previous set of experiments, PLA carriers of 1 × 1 cm size were contaminated with 20 µL of a SARS-CoV-2 suspension displaying a median tissue culture infectious dose (TCID50) of 10 6 IU/mL, which was applied to the surface of carriers in 1 µL droplets. An additional set of carriers was covered with 50 µL of HAdV suspension (10 6 virus copies) spread evenly over the entire surface. The contaminated carriers were then immersed in 96% ethanol, 70% isopropanol or 0. 85% sodium hypochlorite for 15 min. Subsequently, residual viruses—if present—were washed from the dried surface using 180–200 µL PBS. The solution was used directly for infection of Vero-E6 cells (ATCC CRL-1586), in case of SARS-CoV-2, or A-549 human lung carcinoma cells (DSMZ ACC107 from German Collection of Microorganisms and Cell Cultures), in case of HAdV, respectively. Recovered SARS-CoV-2 was titrated by an immunofluorescence (IF) assay using a 1:2. 5 serial dilution of Vero-E6 cells starting from 10 µL. Vero-E6 cells were incubated for 72 h at 37 °C in a CO 2 incubator prior to the IF assay. Briefly, medium was washed out, cells were fixed using 4% paraformaldehyde (PFA), cell membranes were perforated with 0. 2% Triton-X100, and SARS-CoV-2 was labeled with primary mouse anti-SARS-CoV-2 antibody. Secondary anti-mouse antibody was conjugated with a Cy3 fluorophore and a fluorescent microscope (Olympus IX 81, Germany) was used for signal detection. In the case of HAdV, serial dilutions of virus inoculum were used to infect A-549 cells and the cytopathic effect (CPE) was determined using Motic AE21 Inverted Phase Contrast Microscope (Zeiss, Germany). The titers of both recovered viruses infection particles were determined as TCID50 and calculated using the Spearman-Kärber method ( Kärber, 1931 ; Spearman, 1908 ). In addition, recovered HAdV genome copies were determined by real-time quantitative PCR (RQ PCR) as described previously ( Lion et al. , 2003 ) using the ABI Prism Fast 7500 Instrument (Thermo Fisher Scientific, MA, USA). Disinfection of PLA masks worn by test persons To investigate the feasibility of disinfecting PLA protective masks in practical use, three volunteers wore the protective masks of the same type for 4 h. Thereafter, smears from one half of the inner (approximately 80 cm 2 ) or outer surface (approximately 83 cm 2 ) of each mask were performed using sterile cotton swabs. These samples served as a control for natural mask contamination by manual handling, direct skin contact and exhalation. Each cotton swab was transferred into one mL of 0. 9% saline in a microtube, vortexed and inoculated onto a blood agar plate. Thereafter, the filters were removed from masks and the PLA skeletons of the masks were immersed in 96% ethanol for 15 min. After ethanol evaporation, cotton swab smears were taken from the second halves of the inner and outer mask surfaces, inoculated onto agar plates, and incubated at 37 °C for 48 h. The results were averaged and expressed as CFU/mL. Results Investigation of structure and porosity of 3D-printed PLA material The structure and porosity of PLA masks produced by 3D printing were investigated by SEM. Scanning electron micrographs of gaps between the PLA filaments were captured at four different magnifications ( Fig. 2 ). The PLA filament size determined was 312. 8 µm ( Fig. S4 ) and its surface appeared macroscopically very smooth ( Fig. 2A ). Further magnification showed only slight roughness of the surface and very small gaps between filaments ( Figs. 2B and 2C ). Additional increase of magnification revealed connecting filaments of PLA, resulting from the high temperature during 3D printing, with only very small pores (6. 049 µm in size) in between. The pores appeared to be completely closed deeper in the carrier, as observed at the highest magnification used (2, 000×) ( Fig. 2D ). To further test whether the pores were indeed closed and prevented particles from passing through the printed mask, we determined the number of paraffin oil aerosol particles displaying a size of 0. 1–2 µm using the aerosol generator and photometer, certified as a test system according to the common standards. Maximum pressure loss of the generated aerosol was detected, and absolutely no penetration occurred even though the PLA sample was printed with a diameter of 10 cm (corresponds approximately to the printed height of the masks) in the vertical position, simulated printing at a lower temperature in the upper layers (on the z -axis). 10. 7717/peerj. 10259/fig-2 Figure 2 Scanning electron micrographs depicting the structure and porosity of PLA material used for 3D printing of protective masks. (A) Magnification 100×, scale bar 500 µm. (B) Magnification 500×, scale bar 100 µm. (C) Magnification 1000×, scale bar 50 µm. (D) Magnification 2000×, scale bar 30 µm. The observed gaps were measured and marked by black lines. SEM parameters: low vacuum, 5 kV, LVD detector, dwell time 5 µs, spot size 4. 5. Images were taken at various magnifications at the same position. Effect of ethanol, isopropanol and sodium hypochlorite on disinfection of PLA material contaminated with bacteria, yeast fungus or viruses The results of disinfection of artificially contaminated PLA are summarized in Tables 1 and 2. Although the untreated PLA carriers were contaminated by highly concentrated bacterial suspensions of 1 × 10 5 CFU/mL, complete decontamination by all disinfectants used was achieved. Single colonies were observed in the samples of S. epidermidis and E. coli disinfected by isopropanol, but these isolated findings can reasonably be considered a contamination that occurred after treatment of the samples. The disinfection of PLA carriers contaminated with C. albicans (4 × 10 4 CFU/mL) was complete in all cases. 10. 7717/peerj. 10259/table-1 Table 1 PLA material contaminated by Staphylococcus epidermidis, Escherichia coli and Candida albicans, untreated or treated with ethanol, isopropanol or sodium hypochlorite. Untreated Ethanol Isopropanol Sodium hypochlorite Colony forming units on PLA material after contamination by bacteria and yeast (CFU/mL) S. epidermidis 1 × 10 5 0 1 0 E. coli 1 × 10 5 0 1 0 C. albicans 4 × 10 4 0 0 0 Note: Results are expressed in CFU/mL, as the mean of triplicate tests. Untreated samples indicate the CFU/mL count present on contaminated carriers. 10. 7717/peerj. 10259/table-2 Table 2 PLA material contaminated by SARS-CoV-2 and HAdV, untreated or treated with ethanol, isopropanol or sodium hypochlorite. Untreated Ethanol Isopropanol Sodium hypochlorite Virus titers recovered from PLA material after contamination by virus (10 3 IU/mL) SARS-CoV-2 114 0 0 0 HAdV 338 0. 8 4. 7 0 Note: Results are expressed in 10 3 IU/mL, as the mean of triplicate tests. Untreated samples indicate the virus count on contaminated carriers in IU/mL. Titers of SARS-CoV-2 and HAdV recovered from disinfected or untreated carriers were determined by IF- and CPE-based assays, respectively. All disinfection agents tested showed complete virucidal effects against SARS-CoV-2. Disinfectants per se exhibited a cytotoxic effects on Vero-E6 cells ( Table S1 ), but this effect was eliminated by serial dilutions during virus titer determination. HAdV infectivity was reduced by ethanol and isopropanol, and completely abolished by sodium hypochlorite. Similar trends were observed by RQ-PCR performed for detecting the HAdV genome copy numbers ( Table S2 ). Investigation of PLA structure after exposure to ethanol, isopropanol and sodium hypochlorite The effect of disinfectants on the PLA structure was investigated using SEM. PLA structure, gaps between filaments, and the structure of pores after five 15 min cycles of immersing the carrier in different disinfectants are shown in Fig. 3. Treatment with ethanol ( Fig. 3B ) resulted in slight melting of the PLA filaments, as compared with untreated PLA ( Fig. 3A ). The overall PLA structure and surface did not change, but, interestingly, the gap size between the filaments was reduced from the original 6 µm to approximately 850 nm ( Fig. 3B ). This indicates that ethanol treatment may improve the PLA mask properties with regard to structure density. Similarly, isopropanol treatment did not significantly affect the PLA structure ( Fig. 3C ). Only slight melting was detectable, resulting in decreased gap size to 3. 3–4 µm, in comparison to 6 µm in control samples. Moreover, the surface of filaments remained undamaged. Figure 3D depicts the effect of sodium hypochlorite, which did not alter the surface of filaments, but precipitated disinfectant filled the gaps between them, while the gap size remained almost the same as in the control sample (5–7 µm). 10. 7717/peerj. 10259/fig-3 Figure 3 Scanning electron micrographs depicting the structure and porosity of PLA material after short repeated treatment (5 times 15 min) with disinfectants. (A) Untreated sample. (B) Sample treated with 96% ethanol. (C) Sample treated with 70% isopropanol. (D) Sample treated with 0. 85% sodium hypochlorite. SEM parameters: low vacuum, 5 kV, LVD detector, dwell time 5 µs, spot size 4. 5. Images were taken at 2, 000× magnification (scale bar = 30 µm) at the same position. The observed gaps were measured and marked by black lines. Long-term treatment of PLA by immersion in disinfectants for 24 hours was also investigated using SEM ( Fig. 4 ). The effect of long-term treatment with ethanol ( Fig. 4B ) was similar to repeated exposure to sodium hypochlorite ( Fig. 3D ), that is, the gaps between filaments were significantly enlarged to 23. 84 µm ( Fig. 4B ), possibly filled with etched polymer. Investigation of aerosol particle passage through the PLA material after 24 hours in ethanol confirmed that the enlarged gaps were sealed, as no penetration was detected. PLA melting was also observed after prolonged isopropanol treatment ( Fig. 4C ). The gaps between filaments were sealed with the polymer in an irregular manner, resulting in variable gap sizes ranging from 1. 3 to 4. 1 µm. As in all previous tests with ethanol, the surface of PLA filaments remained unaffected. In contrast, long-term treatment with sodium hypochlorite damaged the surface of PLA filaments and revealed precipitation of the disinfectant on the surface ( Fig. 4D ). Similarly to short treatment with sodium hypochlorite, the gaps between filaments, ranging from 2 to 3. 5 µm, were completely filled with precipitated sodium hypochlorite ( Fig. 3D ). 10. 7717/peerj. 10259/fig-4 Figure 4 Scanning electron micrographs depicting the structure and porosity of PLA material after long-time treatment (24 hours) with disinfectants. (A) Untreated sample. (B) Sample treated with 96% ethanol. (C) Sample treated with 70% isopropanol. (D) Sample treated with 0. 85% sodium hypochlorite. SEM parameters: low vacuum, 5 kV, LVD detector, dwell time 5 µs, spot size 4. 5. Images were taken at 2, 000× magnification (scale bar = 30 µm) at the same position. The observed gaps were measured and marked by black lines. Disinfection of PLA masks by ethanol upon wearing by test persons To complement the results of disinfection upon artificial contamination ( Tables 1 and 2 ), disinfection of PLA masks after natural use was investigated. The disinfection efficiency with ethanol (96%) is summarized in Table 3. The microbial load detected on the inner surface of untreated masks varied significantly between different users, ranging from hundreds to thousands CFU/mL. Despite this variation, an average of 7 CFU/mL remained detectable after immersing the masks in ethanol for 15 minutes (short rinsing with ethanol was not sufficiently effective; Fig. S5 ). On the outer surface of untreated masks, 50–150 CFU/mL were detected, and an average of 2 CFU/mL remained detectable after disinfection ( Fig. 5 ). 10. 7717/peerj. 10259/table-3 Table 3 Natural contamination of PLA masks worn by test persons, and subsequently treated with ethanol (bacterial count expressed as CFU/mL). Inner surface (untreated) Inner surface (after disinfection) Outer surface (untreated) Outer surface (after disinfection) Effect of ethanol on disinfection of PLA masks from natural microbial contamination (CFU/mL) PLA mask 1 7, 000 19 85 0 PLA mask 2 257 0 153 0 PLA mask 3 108 2 59 2 average 2, 455 7 99 1 10. 7717/peerj. 10259/fig-5 Figure 5 Effect of ethanol on disinfection of PLA masks contaminated by wearing for 4 h. A representative set of blood agar plates is displayed. The plates were inoculated with material collected from (A) inner surface of the mask before treatment; (B) inner surface of the mask disinfected by immersion in ethanol for 15 min; (C) outer surface of the mask before treatment; (D) outer surface of the mask disinfected by immersion in ethanol for 15 min. Visualization of PLA structure upon mechanical and chemical challenge The impact on the PLA material by finger contact, abrasion by paper or metal and treatment by sodium chloride solution (mimicking perspiration) was analyzed using SEM ( Fig. 6 ). Although fingers may be greasy or sweaty, the contact did not cause any marks or alterations on the PLA surface ( Fig. 6A ). Similarly, gentle mechanical abrasion with paper did not affect the material ( Fig. 6B ). By contrast, intensive mechanical scraping with a dining fork significantly damaged the PLA structure ( Fig. 6C ), leading to compression of PLA filaments, reduction of inter-filament gaps, and shedding of PLA pieces ( Fig. 6D ). However, neither loosening of filaments, nor increase in gap size or other deformations were observed. Soaking in sodium chloride solution did not affect the structure, but salt crystals were present in the gaps between filaments ( Fig. 6E ). In addition, the effect of acetone, which is known to damage PLA, was evaluated. Virtually no gap was visible between filaments upon treatment, indicating that even short exposure to acetone smoothens the structure and seals the pores ( Fig. 6F ). 10. 7717/peerj. 10259/fig-6 Figure 6 Scanning electron micrographs depicting the simulation of human impact on structure and porosity of PLA protective masks. (A) Sample touched by finger. (B) Slightly mechanically stressed sample (paper abrasion). (C) Extremely mechanically stressed sample (scratching by dining fork). (D) Detail of a pore in extremely mechanically stressed sample (dining fork). (E) Sample after immersion in saline solution (perspiration and sweat simulation). (F) Sample after short rinsing with acetone. SEM parameters: low vacuum, 5 kV, LVD detector, magnification 100× or 500×, dwell time 5 µs, spot size 4. 5, scale bar 500 or 100 µm. Discussion The unexpected and sudden spread of SARS-CoV-2 infection, which resulted in the COVID-19 pandemic, has led to a desperate shortage of personal protective equipment, especially among the frontline workers. Because of this problem, many people started helping each other by manufacturing facial protection equipment from commonly available resources. An intriguing possibility is the production of protective face masks using FDM, the most widespread technique of 3D printing. A variety of polymers are suitable for FDM, including biodegradable PLA as the most affordable and environmentally friendly material because of its natural origin ( Ngo et al. , 2018 ). Despite the potential benefits, the suitability of PLA-based materials for protection against viruses was questioned due to their possible high porosity. To the best of our knowledge, this report provides the first data addressing this issue by testing 3D-printed PLA masks ( Fig. 1 ). The surface and other mechanical properties of products made from PLA or composite filaments were investigated previously ( Graupner, Herrmann & Müssig, 2009 ; Chi et al. , 2018 ; Ivanov et al. , 2019 ; Wang et al. , 2016 ). However, the microstructure of 3D-printed PLA objects is highly dependent on the printing parameters, and it is not possible to predict the structure and porosity of a particular object based on published data. To investigate the surface properties of protective face masks made from PLA, examination of structure and porosity is required. We showed by SEM that 3D-printed PLA masks have a compact structure, with small gaps between filaments. The gaps between individual filaments were 6 µm wide, but higher magnification showed that the pores were not continuous within the PLA carrier ( Fig. 2D ) and were actually completely closed. This finding was supported by measurements of the filtering efficiency of PLA, which revealed completely blocked passage of nanometer-sized paraffin aerosol particles. The mask material can therefore be considered impermeable for particles displaying the size range tested, including the fungus, bacteria, and viruses investigated. In combination with the obligatory single-use filters complying with FFP2/3 standards, which are inserted into the mask, spreading of the smallest viruses can also be prevented. Moreover, short exposure to acetone resulted in smoothening of the PLA surface ( Fig. 6F ). A similar 3D-printed reusable face mask prototype was reported by Swennen, Pottel & Haers (2020). The material (polyamide composite) and the printing method used (selective laser sintering technique) differ from the approach presented, but it provided a proof of principle for 3D printing of individualized 3D face masks with FFP2/3 filter membranes as a feasible and valuable alternative source for protective equipment. However, the authors of the cited study did not perform any virus decontamination testing of the reusable components of the face masks and were hence unable to assess the impact of repeated cycles of disinfection on the properties of the material. It was important therefore to determine the possibility of disinfecting the reusable face mask matrix. While SARS-CoV-2, being an enveloped RNA virus, belongs to the less challenging pathogens in terms of disinfection, HAdV (non-enveloped DNA virus) is highly resistant to commonly used disinfectants ( Gordon et al. , 1993 ; Lion & Wold, 2020 ). Adenoviruses mostly cause infections with only mild symptoms in immunocompetent hosts ( Lion, 2019 ), but due to their exceptional stability provide a perfect model for testing the inactivation efficiency. In addition, we examined the disinfection of PLA material from contamination with bacteria ( S. epidermidis and E. coli ) and yeast fungus ( C. albicans ). These microorganisms are part of the human microbiome and their persistence on the protective mask surface poses a risk for infection and a health threat to mask users ( Fisher & Shaffer, 2014 ). All bacterial and fungal microorganisms studied were successfully disinfected using either 96% ethanol, 70% isopropanol or 0. 85% sodium hypochlorite, after immersing contaminated PLA carriers in the respective disinfectant for 15 min ( Table 1 ). Ethanol disinfected the PLA masks contaminated from using by humans ( Fig. 5 ). In comparison to bacteria or fungi, viruses tend to be 1–2 orders of magnitude smaller, making them prone to enter deep into pores of the PLA material. Nevertheless, our data show that efficient disinfection of the PLA carriers from virus contamination is possible, as all tested disinfectants completely inactivated SARS-CoV-2 ( Table 2 ). Treatment with sodium hypochlorite for 15 minutes also completely inactivated the highly resistant HAdV, while ethanol and propanol only led to reduced loads of infectious virus ( Table 2 ). These data are in agreement with the reported sensitivity of both SARS-CoV-2 ( Chin et al. , 2020 ; Kampf et al. , 2020 ) and HAdV to specific disinfectants ( Gordon et al. , 1993 ; Lion & Wold, 2020 ). The present findings therefore provide evidence that PLA material disinfection can be performed with comparable efficiency to other surfaces by appropriate exposure to individual disinfectants. The results obtained can conceivably also help design efficient disinfection protocols for protective face masks made from different materials. Fleischer et al. (2020) examined the changes of PLA material after cleaning with chemical disinfectants (Cidex Opa, Johnson & Johnson and chlorine solutions), revealing mild alterations in the stiffness and strength of 3D-printed PLA samples. However, the authors concluded that high-quality 3D-printed surfaces generated with appropriate printer settings permit cleaning and reuse of 3D-printed medical tools, without compromising their mechanical properties. The authors also stated that immersion in cleaning agents can lead to their absorption into the PLA structure. Thus, additional research is needed to establish efficient and safe chemical cleaning of various 3D-printed surfaces, to prevent health risks associated with tactile and inhalation exposure to chemically cleaned materials. In general, we observed that five cycles of PLA treatment for 15 min with alcohol-based disinfectants resulted in decreased gap size between PLA filaments, without any remnants of disinfectant visible by SEM. By contrast, sodium hypochlorite precipitate was retained in the PLA structure, filling the gaps between PLA filaments. Disinfection of PLA masks with 0. 85% sodium hypochlorite therefore requires further medical investigation to determine whether exposure to the precipitate might be associated with any health risks. Long-term (24-h) treatment of PLA material with disinfectants resulted in partial melting of the filaments, but no erosions of the material were observed ( Fig. 4 ). Ethanol seems to be best suited for the disinfection of PLA masks because it evaporates and does not require removal by rinsing. Moreover, the barrier properties of the mask were not compromised even after long-term exposure, as determined by aerosol challenge. Although the surface of protective equipment should remain intact, inadvertent contacts with the hands and fingers often occur, and the possibility of inappropriate handling has to be considered. The pandemic setting requires medical staff to wear extensive protective equipment (e. g. , overalls, gloves, protective shields and face masks). Such equipment, together with high workload and stress, increases the body temperature and leads to excessive sweating. We mimicked such conditions by mechanical and chemical treatment in order to evaluate alterations of the protective masks. Touching the surface of the PLA material with fingers had no impact, but intensive mechanical stress caused alteration of the PLA filament surface, without affecting the inter-filament gap area. Treatment with sodium chloride (imitating perspiration and sweat) showed salt crystallization in the gaps between filaments ( Fig. 6E ). Crystallized salt compounds, such as sodium chloride or sodium hypochlorite ( Figs. 3D and 4D ), can cause discomfort by skin irritation and itching. This issue was described in detail by Payne (2020) and Wollina (2020) who stated that especially front-line workers obliged to wear a single face mask all day suffer from these problems. The exploitation of PLA may solve this issue, because the fast and cheap manufacturing of protective masks made from this material permits production on a large scale, thereby facilitating more frequent mask changes. Additionally, 3D-printed protective PLA masks are biodegradable, with relatively short decomposition time, thereby providing an environmentally friendly solution. Conclusions This study shows that PLA material is suitable for protection against various microorganisms as it is not permeable for submicroscopic particles. PLA can be efficiently disinfected from bacteria, yeast fungus, and SARS-CoV-2 by commonly available chemical disinfectants such as ethanol, isopropanol or sodium hypochlorite. However, contamination with HAdV, a highly resistant representative of non-enveloped viruses, could only be completely removed with sodium hypochlorite. PLA material is not altered by the immersion in disinfectant or by manual handling. Possible skin irritation after the use of certain disinfectants needs to be carefully evaluated. Single-use filters meeting the FFP2/3 standards are inserted into the mask structure and will be subject of further research and optimization. Overall, PLA can be recommended as suitable material for the manufacturing of protective face masks at times of epidemic spread of infections, such as the ongoing COVID-19 pandemic. Supplemental Information 10. 7717/peerj. 10259/supp-1 Supplemental Information 1 Testing setup of aerosol generator and photometer Lorenz Meβgerätebau FMP 03 certified as a test system according to the standards EN 143 and EN 149. The device consists of two parts, main control unit with aerosol generator and laser photometer (on the right of the image) and pneumatically operated filter-holder (on the left of the image). Click here for additional data file. 10. 7717/peerj. 10259/supp-2 Supplemental Information 2 PLA sample with diameter 10 cm printed vertically and placed in standardized cartridge for the aerosol generator and photometer Lorenz Meβgerätebau FMP 03. On the left of the image is a printed PLA sample. The middle part presents PLA sample attached into standardized Px filter cartridge with a sealer. On the right of the image is presented the whole setup with standardized Px filter cartridge, ready for testing. Click here for additional data file. 10. 7717/peerj. 10259/supp-3 Supplemental Information 3 Aerodynamic particle size distribution in the aerosol generator. On the x-axis is presented size of generated droplets (μm), on the y -axis is presented a percentage of the total dose. Red line represents geomean of particles, blue line represents a distribution of particles, defined by size and percentage. Click here for additional data file. 10. 7717/peerj. 10259/supp-4 Supplemental Information 4 Untreated PLA filament size. SEM parameters: low vacuum, 5 kV, LVD detector, dwell time 5 µs, spot size 4. 5. Image was taken at 100× magnification (scale bar = 500 µm). The observed gap was measured and marked by yellow line. Click here for additional data file. 10. 7717/peerj. 10259/supp-5 Supplemental Information 5 Effect of ethanol (short rinsing) on disinfection of PLA masks contaminated by wearing for 4 h. A representative set of blood agar plates is displayed. The plates were inoculated with material collected from (A) inner surface of the mask before treatment; (B) inner surface of the mask disinfected by short rinsing by ethanol; (C) outer surface of the mask before treatment; (D) outer surface of the mask disinfected by short rinsing by ethanol. Click here for additional data file. 10. 7717/peerj. 10259/supp-6 Supplemental Information 6 Cytotoxic effect (CTE) of used disinfectants on VERO E6 cells. Infectivity of control SARS-CoV-2 virus. Click here for additional data file. 10. 7717/peerj. 10259/supp-7 Supplemental Information 7 PLA material contaminated by HAdV, untreated or treated with ethanol, isopropanol or sodium hypochlorite. Results are expressed in 10 3 genome copies/mL, representing the mean of triplicate tests. Untreated samples represent genome copies/mL of carrier without treatment by disinfectant. Click here for additional data file. 10. 7717/peerj. 10259/supp-8 Supplemental Information 8 Raw data for bacteria and yeast fungus contamination. PLA material contaminated by Staphylococcus epidermidis, Escherichia coli and Candida albicans, untreated or treated with ethanol, isopropanol or sodium hypochlorite. Results are expressed in CFU/mL, as the individual values of triplicate tests. Untreated samples indicate the CFU/mL count present on contaminated carriers. Click here for additional data file. 10. 7717/peerj. 10259/supp-9 Supplemental Information 9 Raw data for SARS CoV-2 contamination. PLA material contaminated by SARS-CoV-2, untreated or treated with ethanol, isopropanol or sodium hypochlorite. Results are depicted as microimages from brightfield microscopy and immunofluorescence assay. Click here for additional data file. 10. 7717/peerj. 10259/supp-10 Supplemental Information 10 Raw data for adenovirus contamination. PLA material contaminated by HAdV, untreated or treated with ethanol, isopropanol or sodium hypochlorite. Results are expressed in 10 3 IU/mL, as the individual values of triplicate tests. Untreated samples indicate the virus count on contaminated carriers in IU/mL. Click here for additional data file. 10. 7717/peerj. 10259/supp-11 Supplemental Information 11 Raw data for SEM images. Click here for additional data file.

10. 7717/peerj. 10374

PeerJ

Wnt16 signaling promotes osteoblast differentiation of periosteal derived cells in vitro and in vivo

Background Periosteum plays critical roles in de novo bone formation and fracture repair. Wnt16 has been regarded as a key regulator in periosteum bone formation. However, the role of Wnt16 in periosteum derived cells (PDCs) osteogenic differentiation remains unclear. The study goal is to uncover whether and how Wnt16 acts on the osteogenesis of PDCs. Methods We detected the variation of Wnt16 mRNA expression in PDCs, which were isolated from mouse femur and identified by flow cytometry, cultured in osteogenic medium for 14 days, then knocked down and over-expressed Wnt16 in PDCs to analysis its effects in osteogenesis. Further, we seeded PDCs (Wnt16 over-expressed/vector) in β -tricalcium phosphate cubes, and transplanted this complex into a critical size calvarial defect. Lastly, we used immunofluorescence, Topflash and NFAT luciferase reporter assay to study the possible downstream signaling pathway of Wnt16. Results Wnt16 mRNA expression showed an increasing trend in PDCs under osteogenic induction for 14 days. Wnt16 shRNA reduced mRNA expression of Runx2, collage type I (Col-1) and osteocalcin (OCN) after 7 days of osteogenic induction, as well as alizarin red staining intensity after 21days. Wnt16 also increased the mRNA expression of Runx2 and OCN and the protein production of Runx2 and Col-1 after 2 days of osteogenic stimulation. In the orthotopic transplantation assay, more bone volume, trabecula number and less trabecula space were found in Wnt16 over-expressed group. Besides, in the newly formed tissue Brdu positive area was smaller and Col-1 was larger in Wnt16 over-expressed group compared to the control group. Finally, Wnt16 upregulated CTNNB1/ β -catenin expression and its nuclear translocation in PDCs, also increased Topflash reporter luciferase activity. By contrast, Wnt16 failed to increase NFAT reporter luciferase activity. Conclusion Together, Wnt16 plays a positive role in regulating PDCs osteogenesis, and Wnt16 may have a potential use in improving bone regeneration.

Introduction Periosteum, a bilayer fibrous soft tissue, is fundamental to both intramembranous bone formation and endochondral ossification during skeletal development and natural bone fracture healing ( Colnot, Zhang & Tate, 2012 ; Debnath et al. , 2018 ). Outer layer of periosteum comprises of fibroblasts and extracellular matrix, while inner layer acts as the reservoir of progenitor cells and osteoblasts ( Colnot, Zhang & Tate, 2012 ). Recent decades, periosteum as a central mediator of bone healing has become an optimal cell source for tissue engineering. Periosteum derived cells (PDCs), isolated from periosteum with an elevator and digested with enzyme, have been proved to possess highly proliferation ability and mesenchymal multipotency(osteogenic, chondrogenic, adipogenic and myogenic) ( Roberts et al. , 2015 ). PDCs contain plenty of stem cells that continuously give rise to osteoblasts, even in elderly patients ( Evans et al. , 2013 ; Ferretti et al. , 2015 ; Ferretti & Mattioli-Belmonte, 2014 ; Uematsu et al. , 2013 ). Those findings generate immense enthusiasm in its utilization in bone tissue regeneration. But its osteogenic and bone regenerative abilities are not well studied. Wnt16 has been regarded as a key regulator in bone metabolism. Osteoblast-specific overexpression of Wnt16 increases both cortical and trabecular bone mass and structure in mice ( Alam et al. , 2016 ). In addition, Wnt16 conditionally ablated mice showed reduced cortical bone thickness, and it was attributed to increased bone resorption and reduced periosteal bone formation ( Ohlsson et al. , 2018 ). Another study revealed that the periosteal bone formation rate and mineral apposition rate were reduced in Wnt16 knockout vs wild-type mice ( Wergedal et al. , 2015 ). Thus, Wnt16 may be a key player in femoral periosteum bone formation. However, the exact role of Wnt16 in PDCs osteogenic differentiation and its utilization in bone regenerative applications remain unclear. Together, our goal was to discover the function of Wnt16 in PDCs, as well as the capability of Wnt16 overexpressed-PDCs in bone regeneration in a critical size bone defect model. Materials & Methods Isolation and culture of PDCs This study was approved by the Research Ethics Committee of West China Hosipital of Stomatology Sichuan University, PR China. (No. WCHSIRB-D2018-093). The CD-1 mice were purchased from Animal experiment center of Sichuan University. Animals acclimated for at least a week from the day of arrival to day of surgical procedure. Animals were euthanized via CO2 inhalation followed by cervical dislocation. Adherent cells were isolated from femoral periosteum of postnatal 4 (P4) weeks CD-1 mice. PDCs were prepared by three sequential enzymatic digestions of femur. Femur were digested with 3 mg/mL of collagenase type I and 4 mg/mL of dispase (Invitrogen, Carlsbad, CA, USA), for 75 min at 37 °C, followed by passage through a 70-µm strainer (BD, Franklin Lakes, NJ, USA). The first digestion (15 min) was discarded as to prevent interference with the connective tissue. Cells from digestions II to III (30min each) were plated in T-75 flasks. PDCs were isolated and cultured in low-glucose Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) containing 20% fetal bovine serum (FBS; Gibco BRL). Passage 3 were used for all in vitro and in vivo experiments in consideration of a balance to obtain sufficient cells for study but also without extended culture that is associated with the loss of cellular phenotypes. Flow cytometry Flow cytometric analysis was used to evaluate the expression of CD105, stem cell antigen-1 (Sca-1), CD90 and CD45. Passage 3 PDCs were incubated individually and simultaneously with monoclonal mouse anti-mouse fluorochrome-conjugated antibodies for 30 min at room temperature. The antibodies were APC-CD105-647 (Cat. No. 120413, Rat anti-Mouse, Biolegend), PE-Cy7-Sca-1 (Cat. No. 108113, Rat anti-Mouse, Biolegend), FITC-CD90. 2 (Cat. No. 140303, Rat anti-Mouse, Biolegend) and PE-CD45 (Cat. No. 400607, Rat anti-Mouse, Biolgegend). After washing with 3% PBSF (3%FBS in phosphate buffer saline) twice, the samples were evaluated with 4-color Beckman Coulter FC 500. Real-time quantitative PCR Total mRNA of PDCs was extracted using TRIzol (Invitrogen). cDNA synthesis was done by reverse transcription of total RNA with TaqMan reverse transcription reagents (Applied Biosystems, CA, USA). Target gene expression was quantified using TaqMan Gene specific primers and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) by using ViiA 7 Real-Time PCR System (ThermoFisher, USA). The following primer/probe sets were used: mouse Wnt16, Mm00437347_m1; mouse Runx2, Mm00501584_m1; mouse osteocalcin/BGLAP (Ocn), Mm03413826_mH; Col-1/Co1a1, Mm00801666_g1 (Applied Biosystems, CA, USA). RNA interference and overexpression We used shRNA to knock down the expression of Wnt16 in PDCs. Liposome encapsulated PLKO. 1 RiWnt16/PLKO. 1 Scramble, PVSVG, Pdeta8. 9 was transfected into 80% confluent 239T cells, Virus supernatant was collected 2 days after transfection. PDCs were infected with the virus supernatant, and the stable positive cell lines were filtrated by using puromycin (1µl, Gibco, USA). Control pHBLV-CMV-EF1-RFP or Wnt16 overexpressed (OE) pHBLV-CMV-EF1-RFP-Wnt16 (6. 25 mg), pMD2. G (Plasmid #12259, Addgene) (0. 625 mg) and psPAX2 (Plasmid #12260, Addgene) (3. 125 mg) vectors were co-transfected into 293T cells using Calcium Phosphate Transfection Kit per manufacturer’s protocol (Invitrogen). Virus supernatant was filtered with 0. 45 mm membrane and purified with ViraBind™ Lentivirus Purification Kit (Cell Biolabs). PDCs were infected with lentivirus in 8 mg/mL polybrene (Santa Cruz Biotechnology). Infected cells and non-infected cells (blank group) were sorted by FACS based on RFP expression ( Figs. S1A – S1E ) and were further passaged 3–5 times. Cell proliferation and differentiation Control and Wnt16 knocking down PDCs were seeded in 96-well plates at a density of 3 × 103 cells per well. Cell numbers were analyzed on day 1, 2 and 3 by Cell Counting Kit-8 (CCK-8) (HY-K0301, MCE, USA) at 450-nm absorbance. For osteogenic differentiation, control and Wnt16 knocking down or overexpressed PDCs were seeded at 70% confluence and cultured with DMEM Medium for 24 h before switched the osteogenic differentiation medium consisting of 100 mM ascorbic acid, 2 mM b-glycerophosphate, 10 nM dexamethasone, with medium change every 2 days. RNA samples were collected at day 0, 2, 3, 7 and 14 for gene expression analysis and protein samples were collected at day 4 for western blotting. Alkaline phosphatase (ALP) staining (Stemgent) was performed at day 7, and Alizarin red staining at day 21. In vivo cells transplantation We used porous β -tricalcium phosphate ( β -TCP) scaffold (5 × 2 mm 3 ; pore size: 200–500 µm; porosity: 50–70%), which was produced by the National Engineering Research Center for Biomaterials, Sichuan University. As we reported previously, 1 × 10 6 control or Wnt16 over-expressed PDCs were planted on surfaces of each scaffold and were primed for osteogenic differentiation at 37 °C and 5% CO2 for 2 days ( Zhou et al. , 2015 ). 27 CD-1 mice (male, 4–5 wks, 6 mice per group) were used in the in-vivo experiments. These mice were randomly distributed into three groups: blank, control and Wnt16 OE. The mice were first anesthetized with isoflurane. Calvarial defect (diameter: 5 mm) were created on the calvarium of mouse by using a circular hollow bur, and the disconnected cranium was carefully lifted without destroying the underneath blood vessels. Both control and Wnt16 over-expressing PDCs were seeded on β -TCP scaffolds. Scaffolds without cells (blank group) and with control or Wnt16 OE cells were implanted into a calvarial defect. Upon recovery, animals were closely monitored every 2–3 mins until they maintained a sternal position. Animals were feed on soft food for next one week post-operatively. Animals continued to be observed twice daily for next 3 days post-operatively and every other day thereafter. If signs such as hunched posture, lethargy, lack of food intake, skin irritation or infection of the surgery site, dehiscence of the incision site are observed, a veterinarian would be contacted or the mouse would be euthanized by CO2 followed by cervical dislocation. Micro-CT imaging analyses Un-decalcified calvaria samples were harvested and fixed with 4% paraformaldehyde, then scanned using a micro-computed tomography (micro-CT) system (Scanco Medical, Bassersdorf, Germany) at a voxel size of 10. 5 µm3 to image bone. The cranium image was reconstructed via MicroView ABA 2. 2 software. The regions of interest (ROIs) were bone like structure in the defect. New bone formation rate was identified as the trabecular bone volume (BV)/scaffold tissue volume (TV). The quantitative structural parameters were trabecular number (TbN, 1/mm) and trabecular space (TbSp, µm). Histology Eight weeks later, animals were euthanatized by CO2 followed by cervical dislocation, the transplants were collected and fixed with 4% paraformaldehyde. After fixation, samples were demineralized in 0. 5 M EDTA (pH 7. 4) for 1 week and then dehydrated. Paraffin sections (5 mm) were used for hematoxylin-eosin (HE), Masson’s trichrome and immunohistochemical staining. We used Anti-Mouse HRP-DAB Cell & Tissue Staining Kit (R & D Systems) in the immunohistochemical staining. After deparaffinizing, heat-retrieving and blocking, sections were incubated with the primary antibodies (Brdu, ab270260, Abcam; Col11A1, ab64883, Abcam) overnight at 4 °C, then incubated with HRP conjugated secondary antibodies. Finally, images were developed with DAB. We stained three tissue sections per sample. The ROIs were calcified region, and we randomly selected three fields (400X). The BrdU and Col1 were analyzed by their integrated optical density (IOD) value via image pro plus 6. 0 software. When we calculated the IOD, we eliminated the color “blue” interference which stands for the nucleus. Western blot After washing with ice-cold PBS, protein was extracted by using RIPA Lysis Buffer (Thermo Scientific) with Protease/Phosphatase Inhibitor Cocktail (ab201119, Abcam USA). Samples were added on a NuPAGE® Novex® 4e12% BiseTris Protein Gel (1. 0 mm), transferred to nitrocellulose membrane, and incubated with anti-Pro-Col-1 (sc-8782, 1:200, Santa Cruz Biotechnology), and anti-GAPDH (sc-25778, 1:200, Santa Cruz Biotechnology) antibodies. Following using corresponding secondary antibodies, image was developed with IR fluorescence & Odyssey. Immunohistochemistry For immunocytochemical staining, control and Wnt16 overexpressed PDCs were cultured in the 6 well cell culture cluster until getting 70–80% confluence. After blocking, cells were incubated with anti-CTNNB1 antibody (ab32572, 1:200, Abcam) overnight at 4 °C. Transfections and luciferase reporter activity We seeded PDCs into 24-well plates and transiently transfected them with Topflash (Plasmid, 12457, Addgene) + Wnt16 (Plasmid, 42291, Addgene)/NFAT (Plasmid 11792, Addgene) + PGL4 [hRluc/sv40] (Plasmid, 64034, Addgene) and Topflash+ EGFP (Plasmid, 54762, Addgene) + PGL4 as control using Neon® Transfection System for Electroporation (Thermo Fisher Scientific) according to the manufacturer’s protocol. Post-transfection, cells were starved for 8 h, and then 50 ng/ml rhDKK1 (R&D Systems) or 50 ng/ml rhWnt3a (Abcam) or 50 ng/ml rhWnt5a (Abcam) or 5, 000 ng/ml INCA-6 (R&D Systems) was used to assay canonical Wnt and non-canonical Wnt pathways. After 16 hours‘ incubation, Cells were lysed and luciferase activities were measured according to the manufacturer’s instructions (Luciferase Reporter Assay System; Promega, CA, USA). Statistical analysis Experiments were performed in triplicate. We used SPSS software for our analyses. Quantitative data were analyzed for statistical difference using Student’s t -test or one-way ANOVA, with significance of p < 0. 05. The values are expressed as mean ± SD. Results PDCs isolation and morphological characterization PDCs are typically identified as mononucleated adherent cells isolated from femoral periosteum, which are thought to contain plenty of MSCs ( Debnath et al. , 2018 ; Ferretti & Mattioli-Belmonte, 2014 ). The cell surface antigens were characterized by flow cytometry analyses. Flow cytometry revealed that 89. 1% cells were CD90+CD105-, and 66. 6% cells were Sca-1+CD45-. In addition, within Sca-1+CD90+, there were 87. 2% cells were CD105-CD45-( Ferroni et al. , 2019 ; Ng et al. , 2020 ) ( Fig. 1A ). These results revealed that these isolated cells from femoral periosteum contain a proportion of stem cells. 10. 7717/peerj. 10374/fig-1 Figure 1 Wnt16 expression and function in osteogneisis of periosteum derived cells (PDCs). Wnt16 expression and function in osteogneisis of periosteum derived cells (PDCs). (A) Flow cytometry showed that approximately 89. 1% in the third passage of the isolated PDCs were CD90 + /CD105 -, and approximately 66. 6% of them were Sca-1 + /CD45 -. Within CD90 + /Sca-1 +, 87. 2% cells wereCD90 - /CD45 -. (B) Change of Wnt16 mRNA expression in PDCs under osteogenesis induction medium for 14 days ( n = 3). (C) mRNA of Wnt16, Runx2, OCN and collagen types I (Col-1) after Wnt16 knocking down (kd) by using wnt16 shRNA in PDCs after 7 days of osteogenic indction ( n = 3). (D) Whole plate mount of alkaline phosphataseactivity in PDCs with or without Wnt16 knocking down after 7 days of osteogenic indction. (E) Whole plate mount of Alizarin Red staining image in PDCs after 21 days of induction. mRNA of Wnt16 (F), Runx2 and OCN (H) upon Wnt16 overexpression (OE) by using lentivirus vector after 2 days of induction. ( n = 3) (G) Runx2 and Col-1 protein upon Wnt16 overexpression. * p < 0. 05, ** p < 0. 001, *** p < 0. 0001. Wnt16 involves in PDCs osteogenesis We analyzed the mRNA of Wnt16 at day 0, 2, 3, 7 and 14 days to uncover the expression pattern of Wnt16 during the osteogenesis of PDCs. We cultured the PDCs in osteogenic medium for 14 days, and detected Wnt16 mRNA expression via qPCR. Wnt16 mRNA expression up-regulated significantly in PDCs at day 7 and 14 ( p < 0. 0001) ( Fig. 1B ). In order to analyze the role of Wnt16 in the ostegenesis of PDCs, Wnt16 shRNA or Vector was transfected into PDCs. Endogenous Wnt16 was reduced to ∼1/8 of the original level in PDCs ( p < 0. 0001) ( Fig. 1C ). Our CCK-8 analysis resulted Wnt16 shRNA transfection did not affect the proliferation rate of those cells ( Fig. S1 ), suggesting perhaps its restricted roles in regulating mesenchyme differentiation and maturation without necessarily affecting cell growth. Because the expression of Wnt16 increased dramatically from day 7, so we checked the change of the mRNA expression of osteogenic transcriptional factors/genes, such as Runx2, collage type I (Col-1), and osteocalcin (Ocn) after osteogenic induction for 7 days. Interestingly, Wnt16 shRNA reduced significantly the mRNA expression of Runx2, Col-1 and Ocn ( p < 0. 05). ( Fig. 1C ) In addition, ALP staining intensity was also diminished a little in Wnt16 knocked down PDCs. ( Fig. 1D ) Surprisingly, after 21 days‘ostegenic induction Wnt16 shRNA dramatically weakened the intensity of alizarin red staining, stands for the mineralization of the extracellular matrix ( Aguila & Rowe, 2005 ). ( Fig. 1E ) Besides using shRNA to knock down Wnt16 expression, we used lentivirus infection to overexpress Wnt16 in PDCs to uncover whether up-regulated Wnt16 would promote the early process of osteogenesis. Wnt16 was enhanced to 8 times of the original level ( p < 0. 0001) ( Fig. 1F ), which significantly promoted not only mRNA expression ( p < 0. 05) ( Fig. 1H ) but also protein production ( Fig. 1G ) of Runx2 after two days of osteogenic stimulation. Meanwhile, OCN mRNA expression ( p < 0. 05) and Col-1 protein production were also enhanced by Wnt16 overexpression. Taken together, those data indicate that Wnt16 may up-regulate osteoblast differentiation and mineralization during PDCs osteogenesis. Wnt16 enhances collagenous fibers formation in critical size calvarial bone defect To further investigate the possibility of using Wnt16 altered-PDCs in bone regeneration, we established a mouse critical size calvarial defect. The genetically engineered PDCs was embedded in β -TCP scaffold. Then, we transplanted those graft into a calvarial defect (diameter: 5 mm). After 8 weeks of healing, we studied the capacity of calvarial defect restored and the quantity of newly formed bone in the defect sites by micro-CT. New bone formation was observed not only in the marginal but also in central regions of the defects in Wnt16 overexpression (OE) groups. ( Figs. 2A – 2F ) Morphometric analysis of the quantity of newly formed bone indicated that more BV/TV, TbN and less TbSp were found in Wnt16 OE group ( p < 0. 05). ( Figs. 2G – 2I ) In addition, the histological characteristics were observed in all groups. Blood vessel was observed in all groups, without any significant difference among groups. In HE and Masson staining images, abundant extracellular matrix and collagen was observed in Wnt16 OE groups. ( Figs. 2J – 2O & 2P – 2U ). In addition, Brdu positive area was attenuated by Wnt16 OE ( Figs. 3A – 3C, 3G ), but Col-1 was more intensive in Wnt16 OE group than in control and blank groups. ( Figs. 3D – 3F, 3H ). Taken together, our data suggest that Wnt16 promoted the transplanted cells undergoing osteogenic differentiation, leading to more collagenous fibers formation in the calvarial bone defect. 10. 7717/peerj. 10374/fig-2 Figure 2 Wnt16 enhanced mineralized tissue formation in vivo by PDCs/ β -TCP implantation. (A–F) Representative micro-CT images of bone formation. Yellow stands for newly formed bone like tissue. (G–I) Quantitative analysis of trabecular bone volume (BV)/scaffold tissue volume (TV) ratio, trabecular space (Tb. Sp), trabecular number (Tb. N). (J–Q, P–U) Histological analysis of bone formation at 8 weeks after implantation: H & E staining (J–Q); Masson’s trichrome stainings (P–U). (J, L, N, P, R, T) Magnification × 100; (K, M, O, Q, S, U) Magnification × 400; Red arrow: blood vessel; Blue arrow: mineralized matrix. Scale bar: 200 µm. * p < 0. 05, ** p < 0. 001, *** p < 0. 0001. 10. 7717/peerj. 10374/fig-3 Figure 3 Immunohistological staining of MSCs and osteogenesis marker proteins in calvarial defects at 8 weeks after surgery. (A–C) Brdu production in mouse calvarial defects. (G) Brdu expression was significantly lower than the control group ( n = 3). (D–F) Col-1production in mouse calvarial defects. (H) Col-1 expression was significantly higher than the control group ( n = 3). OE, over-expressed; m, matrix, Scale bar: 200 µm. * p < 0. 05. Wnt16 activating canonical β -catenin pathway To further understand the molecular mechanisms of Wnt16 regulation of osteogenic differentiation, we used Wnt16 OE PDCs to test the potential downstream signaling factors, which are β -catenin and WNT/Ca2+. Strikingly, immunofluorescence staining revealed that Wnt16 promoted CTNNB1 expression and nuclear translocation ( Figs. 4A – 4F ). Consistent with the increased nuclear translocation of CTNNB1, Wnt16 OE further increased the TCF/LEF reporter activity in a Topflash assay ( p < 0. 05), which indicates Wnt/ β -catenin signaling pathway activation ( Ye et al. , 2019 ). But Wnt16 OE along with Wnt3a, a canonical Wnt that is thought to activate Wnt/ β -catenin signaling pathway, showed almost similar effect when Wnt16 OE or Wnt3a existed alone; meanwhile DKK-1, a specific inhibitor of the target of the Wnt/ β -catenin pathway, restrained Wnt16‘s effect ( Fig. 4G ) ( Baron & Kneissel, 2013 ). On the other hand, Wnt16 OE failed to increase NFAT reporter activity, which indicates the activation of calcineurin/Ca2+ pathway ( Li et al. , 2016 ). However, when pretreatment with Wnt5a, a noncanonical wnt that has been proved to activate NFAT transcription factors, Wnt16 reduced the relative luciferase activity compared to the control group ( p < 0. 05) ( Fig. 4H ). In addition, no significantly effect was found when adding a calcineurin-NFAT interaction inhibitor INCA-6. Taken together, our data suggests that Wnt16 activates β -catenin signaling, but does not trigger off Wnt/Ca2+ pathway, and even may be a competitive inhibitor to Wnt5a regarding to the activation of Wnt/Ca2+ pathway. 10. 7717/peerj. 10374/fig-4 Figure 4 Wnt16 activate canonical wnt signaling pathway. (A–F) immunofluorescence staining of β -catenin in Wnt16 oeverxpressed (OE) PDCs or in control groups. Overexpression of Wnt16 in PDCs promote β -catenin translocating into nucles. (G) TCF/LEF promoter relative activity was quantified by transfecting the TCF/LEF reporter and using a luciferase assay in Wnt16 OE cells added with PBS, Wnt3a or DKK-1 ( n = 3). (H) NFAT-dependent transcriptional activity was quantified by transfecting the NFAT reporter and using a luciferase assay in Wnt16 OE cells added with PBS, Wnt5a or INCA-6 ( n = 3). Scale bar: 100 µm, * p < 0. 05. Discussion Wnt16 has been recognized as a positive regulator of cortical bone thickness and periosteal bone formation ( Alam et al. , 2016 ; Ohlsson et al. , 2018 ; Wergedal et al. , 2015 ). Our findings elaborate that Wnt16 appears to effect the whole stages of osteogenesis in periosteal derived cells (PDCs). Firstly, we identified the existence of stem cells in the PDCs. Flow cytometry showed 50. 6% cells in the PDCs were Sca-1+/CD45-, and 51. 9% cells were CD105-/34-. Secondly, under the induction of osteogenic medium, Wnt16 increased dramatically in PDCs from day 7 to day 14. Knocking down of Wnt16 in PDCs significantly reduced Runx2, Ocn and Col-1 expressions, while Wnt16 overexpression promoted both Runx2 mRNA expression and protein production. In addition, knocking down of Wnt16 has a tendency to reduce ALP activity, an earlier osteoblast differentiation marker, and dramatically reduced the alizarin red staining intensity, which refers to the later mineralization stage of osteogenesis ( Mortada & Mortada, 2018 ). Thirdly, using a model of bone injury, PDCs genetically engineered to express Wnt16 significantly regenerated more bone-like tissues in the critical size bone defects in vivo. Finally, we confirm that Wnt16 activates canonical Wnt β -catenin signaling, which is consistent with the previous studies ( Hendrickx et al. , 2020 ; Wergedal et al. , 2015 ). However, Wnt16 did not trigger off Wnt/Ca2+ pathway, but may indirectly inhibit Wnt/Ca2+ pathway by antagonizing Wnt5a. The present findings suggest that Wnt16 act positively in PDCs osteogenesis process, and Wnt16 modified PDCs may be promising for bone regeneration strategies. Recently, periosteal cells have generated immense enthusiasm regarding its utilization in tissue engineering and regeneration due to its essential roles in periosteum mediated de novo bone formation and fracture repair ( Bragdon & Bahney, 2018 ; Debnath et al. , 2018 ; Yang et al. , 2020 ). However, periosteal cell cultures have indicated the heterogeneous nature of PDCs with MSCs, fibroblasts and osteogenic cells ( Gao et al. , 2019 ). Previous study proved that bone-derived Sca-1+/CD45-/CD31- cells exhibiting trilineage osteoblastic, adipocytic, and chondrocytic differentiation ability and higher clonogenic efficiencies ( Holmes & Stanford, 2007 ). Interestingly, in our isolated PDCs, 66. 6% cells were Sca-1+/CD45- which indicates that half of the PDCs probably possess pluripotency capability. Furthermore, Debnath et al. identified three populations among periosteum derived cells, which were CD49 flow /CD51 low : CD200+/CD105- periosteal mesenchymal stem cells (PSCs), CD200-/CD105- periosteal progenitor 1 (PP1) cells, and CD105+/CD200 variable periosteal progenitor 2 (PP2) cells. They claimed PSCs are broadly different from both their PP1 and PP2 derivatives, and PSCs displayed increased per-cell bone formation capacity compared to PP1/PP2 cells ( Debnath et al. , 2018 ). Our results showed that 89. 1% cells were CD90+/CD105-, and within CD90+/Sca-1+ cells, 87. 2% cells were CD45-/CD105-, which means the majority of our PDCs were CD90+/Sca-1+/CD105-/CD45-. Recent studies reported that DDR2- PDCs, Nestin+ PDCs and LepR+ PDCs possess multipotent and self-renewal abilities ( Gao et al. , 2019 ; Yang et al. , 2020 ). Together, sorting a specific fraction of PDCs with highly pluripotency capability by using cell markers is promising and meaningful in bone tissue engineering and regeneration. Osteoblast differentiation involves several stages including commitment of mesenchymal progenitors to the osteoblast lineage, osteoprogenitor proliferation, matrix maturation and mineralization. Wnt signaling was thought to stimulate early osteoblasts in their capacity to differentiate, whereas mature osteoblasts were strongly inhibited in their capacity to induce mineralization ( Eijken et al. , 2008 ). Some inhibitors of WNT signaling even have a positive role in osteoblast maturation, such as Dkk2 ( Li et al. , 2005 ). Previous study demonstrated that Wnt16 may act in a stage-dependent manner to regulate osteoblast differentiation. Wnt16 up-regulated Bmpr1b, Bmp7 and Enpp1, which mainly function in osteoblast differentiation stage, but rWnt16 down-regulated Alpl, Rspo2 and Ibsp, genes that are mainly involved in osteoblast maturation and mineralization stage ( Sebastian et al. , 2018 ). However, our result showed that Wnt16 accelerate both osteoblast differentiation and matrix mineralization of PDCs osteogenesis process. This controversy clearly needs further studies to settle down regarding to whether Wnt16 inhibit or promote mature osteoblast mineralization. In the past years, researchers recognize that the effect of Wnt16-promoted osteogenesis in MSCs are coming from the activation of both canonical and non-canonical WNT signaling. Wnt16 activates Wnt/ β -catenin, Wnt/Ca2+ and Wnt/PCP signaling via a specific G α subunits, which plays an orchestrating role in the downstream activity in osteoblasts ( Gao et al. , 2019 ; Hendrickx et al. , 2020 ). Another paper suggested that non-canonical JNK pathway plays a key role in transcriptional activation of Wnt16 targets in osteoblasts ( Sebastian et al. , 2018 ). Our finding confirms that Wnt16 activates β -catenin signaling, yet suggests that Wnt16 may not trigger off Wnt/Ca2+ pathway but may indirectly inhibit Wnt/Ca2+ pathway by antagonizing Wnt5a. Reversely, Wnt5a is able to antagonize Wnt3a-induced β -catenin/TCF activity, reduce the stemness of hepatic progenitor cells, and promote hepatic differentiation of liver progenitors ( Fan et al. , 2017 ). Probably, Wnt16/ β -catenin signaling and Wnt5a/Ca2+ pathway may exist a regulatory feedback loop. Obviously, additional studies are needed to demonstrate the intricate regulatory network among Wnt16, Wnt5a and their intracellular downstream pathways. Our study has one limitation. The regenerated de novo bone formation in the present study is still unsatisfied. One reason may be the consequence of mixed cell population. Since our cells were not purified by cell markers such as CD90, CD105 and CD200, the un-purified cell population might contain fibroblasts, differentiated pre-osteoblasts/osteoblasts and were lack of progenitor cells, so that the transplant was not fully filled with new bone ( Debnath et al. , 2018 ; Sakai et al. , 2011 ). On the other hand, albeit effective gene expression is desirable, yet to date the exogenous gene is often driven by only one constitutive promoter, whose effect is restricted to only one facet, such as osteogenesis, osteoclastogenesis, or angiogenesis. In contrast to the conventional gene therapy approach, the Bac-CRISPRa system, which is programmable and capable of activating multiple target genes, may be a promising technology in bone regeneration ( Larouche & Aguilar, 2019 ). A recent study used CRISPRa-engineered BMSCs to activate Wnt10b and Foxc2, resulting in a remarkably improved calvarial bone healing ( Hsu et al. , 2020 ). Therefore, using cell markers to purified the primary PDCs and CRISPRs system therapy may further optimize bone healing. Conclusions In summary, our data uncover Wnt16 act positively in PDCs osteogenesis. In vivo study proves overexpression of Wnt16 in PDCs could promote orthotopic de novo bone-like tissue formation in bone defect. Supplemental Information 10. 7717/peerj. 10374/supp-1 Figure S1 Cell growth after knock down Wnt16 in PDCs Cell proliferation rate was determined by CCK-8 assay, no difference was found between control and Wnt16 knock down (kd) group. Click here for additional data file. 10. 7717/peerj. 10374/supp-2 Figure S2 Wnt16 overexpressed in PDCs tagged with RFP (A–B) immunofluorescence images showed that the transfected PDCs were RFP positive in Wnt16 OE and control group. (C–D) FACS images showed 4. 06% cells were RFP postive in blank group, while 26. 5% and 22. 1% cells were RFP positive in control and Wnt16 OE group. Click here for additional data file. 10. 7717/peerj. 10374/supp-3 Supplemental Information 3 Figure 1 raw data Click here for additional data file. 10. 7717/peerj. 10374/supp-4 Supplemental Information 4 Figure 2 raw data Click here for additional data file. 10. 7717/peerj. 10374/supp-5 Supplemental Information 5 Figure 3 raw data Click here for additional data file. 10. 7717/peerj. 10374/supp-6 Supplemental Information 6 Figure 4 raw data Click here for additional data file. 10. 7717/peerj. 10374/supp-7 Supplemental Information 7 Raw data exported from the PCR ViiA 7 Real-Time PCR software applied for data analysis and preparation for the detailed investigation shown in Fig. 1B Click here for additional data file. 10. 7717/peerj. 10374/supp-8 Supplemental Information 8 Raw data exported from the PCR ViiA 7 Real-Time PCR software applied for data analysis and preparation for the detailed investigation shown in Figs. 1F and 1H Click here for additional data file. 10. 7717/peerj. 10374/supp-9 Supplemental Information 9 Raw data exported from the PCR ViiA 7 Real-Time PCR software applied for data analysis and preparation for the detailed investigation shown in Fig. 1C Click here for additional data file. 10. 7717/peerj. 10374/supp-10 Supplemental Information 10 Raw data exported from the PCR ViiA 7 Real-Time PCR software applied for data analysis and preparation for the detailed investigation shown in Figs. 1F and 1H Click here for additional data file. 10. 7717/peerj. 10374/supp-11 Supplemental Information 11 Raw data exported from the PCR ViiA 7 Real-Time PCR software applied for data analysis and preparation for the detailed investigation shown in Fig. 1C Click here for additional data file. 10. 7717/peerj. 10374/supp-12 Supplemental Information 12 Raw data exported from the PCR ViiA 7 Real-Time PCR software applied for data analysis and preparation for the detailed investigation shown in Figs. 1F and 1H Click here for additional data file.

10. 7717/peerj. 10890

PeerJ

Use of conditioned media (CM) and xeno-free serum substitute on human adipose-derived stem cells (ADSCs) differentiation into urothelial-like cells

Background Congenital abnormalities, cancers as well as injuries can cause irreversible damage to the urinary tract, which eventually requires tissue reconstruction. Smooth muscle cells, endothelial cells, and urothelial cells are the major cell types required for the reconstruction of lower urinary tract. Adult stem cells represent an accessible source of unlimited repertoire of untransformed cells. Aim Fetal bovine serum (FBS) is the most vital supplement in the culture media used for cellular proliferation and differentiation. However, due to the increasing interest in manufacturing xeno-free stem cell-based cellular products, optimizing the composition of the culture media and the serum-type used is of paramount importance. In this study, the effects of FBS and pooled human platelet (pHPL) lysate were assessed on the capacity of human adipose-derived stem cells (ADSCs) to differentiate into urothelial-like cells. Also, we aimed to compare the ability of both conditioned media (CM) and unconditioned urothelial cell media (UCM) to induce urothelial differentiation of ADCS in vitro. Methods ADSCs were isolated from human lipoaspirates and characterized by flow cytometry for their ability to express the most common mesenchymal stem cell (MSCs) markers. The differentiation potential was also assessed by differentiating them into osteogenic and adipogenic cell lineages. To evaluate the capacity of ADSCs to differentiate towards the urothelial-like lineage, cells were cultured with either CM or UCM, supplemented with either 5% pHPL, 2. 5% pHPL or 10% FBS. After 14 days of induction, cells were utilized for gene expression and immunofluorescence analysis. Results ADSCs cultured in CM and supplemented with FBS exhibited the highest upregulation levels of the urothelial cell markers; cytokeratin-18 (CK-18), cytokeratin-19 (CK-19), and Uroplakin-2 (UPK-2), with a 6. 7, 4. 2- and a 2-folds increase in gene expression, respectively. Meanwhile, the use of CM supplemented with either 5% pHPL or 2. 5% pHPL, and UCM supplemented with either 5% pHPL or 2. 5% pHPL showed low expression levels of CK-18 and CK-19 and no upregulation of UPK-2 level was observed. In contrast, the use of UCM with FBS has increased the levels of CK-18 and CK-19, however to a lesser extent compared to CM. At the cellular level, CK-18 and UPK-2 were only detected in CM/FBS supplemented group. Growth factor analysis revealed an increase in the expression levels of EGF, VEGF and PDGF in all of the differentiated groups. Conclusion Efficient ADSCs urothelial differentiation is dependent on the use of conditioned media. The presence of high concentrations of proliferation-inducing growth factors present in the pHPL reduces the efficiency of ADSCs differentiation towards the urothelial lineage. Additionally, the increase in EGF, VEGF and PDGF during the differentiation implicates them in the mechanism of urothelial cell differentiation.

Introduction Congenital abnormalities, cancers as well as injuries may lead to irreversible damage to the urinary tract, which eventually requires reconstruction. Early attempts on the reconstruction of urinary tract have focused on the use of acellular matrix for bladder reconstruction. Acellular matrices are immunologically inert and act in vivo as scaffolds, to recruit the progenitor cells and infiltrate the matrix to produce bladder tissue ( Howard et al. , 2008 ). The two most commonly used acellular matrices for bladder and urethral reconstruction are the small intestinal submucosa (SIS) and the bladder acellular matrix (BAM) ( Staack et al. , 2005 ). In the year 2000, the FDA approved the use of porcine collagen matrix, derived from the small intestinal submucosa, in the reconstruction-based surgical procedures ( Hodde & Johnson, 2007 ). However, fibrosis, matrix shrinkage and the lack of the ability to perform urothelial anastomosis are the major hurdles that must be overcome before acellular matrices can be used in the bladder and urethral reconstruction ( Portis et al. , 2000 ; Campodonico et al. , 2004 ; Azadzoi et al. , 1999 ; Horst et al. , 2019 ). Adult stem cells represent an accessible source of unlimited repertoire of untransformed cells. Early attempts to incorporate stem cells in urinary tract tissue engineering culminated in using stem cells without transdifferentaition and directly implanting them in vivo ( Liao et al. , 2013 ; Chung et al. , 2005 ). Mesenchymal stem cells (MSCs) have been used in organ reconstruction by expanding them in vitro and then implanting them to induce their differentiation potential ( Sharma et al. , 2010 ; Sharma et al. , 2009 ). On the other hand, stem cells can be expanded and differentiated in vitro and then transplanted directly to the affected subject. Thus, direct differentiation of stem cells can reduce the time required for patient’s recovery. Smooth muscle cells, endothelial cells, and urothelial cells are the major cell types required for the reconstruction of lower urinary tract ( Qin et al. , 2014 ). Three induction protocols have been reported to induce the differentiation of stem cells towards the urothelial lineage: direct co-culture with urothelial cells, indirect co-culture system with urothelial cells, and culturing the stem cells in conditioned media (CM) derived from urothelial cell culture ( Becker & Jakse, 2007 ). The first protocol of direct co-culture system is inapplicable in cases of malignancies, infections, and inflammatory diseases ( Liu et al. , 2009 ; Shi et al. , 2012 ). The indirect co-culturing is applicable in small systems such as the filter well insert ( Liu et al. , 2009 ). To the current time, the use of conditioned media represents the most favorable and easy-to-use method for induction of MSCs to urothelial cells. However, several problems still need to be addressed before using these cells in clinical therapy, including the limited differentiation capability (transdifferentaition efficiency ranges from 40–70%) and the presence of xenogeneic substances such as fetal bovine serum (FBS) in cell culture media ( Zhang et al. , 2014 ; Shi et al. , 2012 ). Fetal bovine serum is the most vital supplement used in cell culture media for cell proliferation and differentiation. However, due to its limited supply and the increased demand on manufacturing xeno-free stem cell-based cellular products, optimizing the composition of culture media and the type of serum used are of critical importance. Human platelet lysate (HPL) is defined by Marx as “the volume of plasma fraction of autologous blood having a platelet concentration above baseline”. It has been reported that HPL enhances the proliferation and differentiation of MSCs compared to xenogenic FBS ( Kakudo et al. , 2008 ; Lucarelli et al. , 2003 ; Li et al. , 2013 ; Mishra et al. , 2009 ; Cervellia et al. , 2012 ). These effects make the HPL as an attractive alternative that can be used in MSCs culture with minimal adverse effects in clinical settings. In this study, the effects of FBS and pooled human platelet (pHPL) lysate were assessed on the capacity of huADSCs to differentiate into urothelial-like cells. Also, we aimed to compare the ability of both conditioned media (CM) and unconditioned urothelial cell media (UCM) to induce urothelial differentiation of ADCS. Methods Material and reagents Dulbecco’s Modified Eagle’s Medium DMEM (GIBCO, Waltham, MA, USA), collagenase type l (Worthington, Lakewood, NJ, USA), FBS (GIBCO, Waltham, MA, USA), streptomycin and penicillin and 20 mM L-glutamine (Euroclone, Italy), 0. 25% trypsin–0. 04% EDTA (GIBCO, Waltham, MA, USA), (SV-HUC-1) ATCC (CRL-9520), StemPro Adipogenesis differentiation media (Invitrogen, Hercules, CA, USA), Trizol reagent (Invitrogen, Hercules, CA, USA), Human MSC Analysis Kit (BD, Franklin Lakes, NJ, USA), iScript reverse transcription supermix (BioRad, Hercules, CA, USA), iQTM SYBR mix (BioRad, Hercules, CA, USA), cytokeratin-18 (Abcam, ab668, 1:200), uroplakin-2 (Santa Cruz, sc-15178, 1:50), ELISA kits (Abcam, UK). Cell culture All experimental protocols involving human tissues were approved by the Ethics Committee at the King Abdullah Hospital, School of Medicine, Jordan University of Science and Technology (IRB No: IRB/7/2019). After obtaining signed informed consents, human adipose tissue aspirates were obtained from three individuals, aged 30, 31 and 35, who underwent liposuction procedures. ADSCs were isolated as previously described ( Francis et al. , 2010 ). Briefly, adipose tissue aspirates were digested with 0. 1% collagenase type l (Worthington, Lakewood, NJ, USA) in PBS, for 45 min at 37 °C, with gentle shaking every five minutes. The enzyme was then diluted with an equal amount of complete cell culture medium consisting of DMEM (GIBCO, Waltham, MA, USA) supplemented with 10% FBS (GIBCO, Waltham, MA, USA), 1% streptomycin and penicillin and 20 mM L-glutamine (Euroclone, Italy). The suspension was centrifuged at 1200× g for 10 min and then the pellet was resuspended in 5 ml complete cell culture medium. After that, the cell suspension was passed through a 70 µm cell strainer and centrifuged at 500× g for 10 min. The obtained cells were counted and seeded at a density of 2 ×10 5 cells/cm 2 in a tissue culture flask and incubated at 37 °C and 5% CO 2. The medium was changed every two to three days until the adherent ADSCs became 70–80% confluent. Cells were detached with 0. 25% trypsin–0. 04% EDTA (GIBCO) solution, and the resulting ADSCs at passage 3–5 were used for further experiments. SV40 immortalized human ureter urothelium (SV-HUC-1) cell line was obtained from ATCC (CRL-9520). SV-HUC-1 cells were cultured in F-12K medium supplemented with 10% FBS and 1% streptomycin and penicillin. Conditioned medium derived from SV-HUC1 was collected and used to induce the ADSCs differentiation towards the urothelial-like cell lineage. Flow cytometry Cultured ADSCs at passage 3 and 70% confluency were utilized for cell surface marker assessment. Cells were detached using TrypLE (GIBCO) and washed twice with FACS buffer (PBS, 1% FBS). Then, cells were counted and adjusted to 10 6 cells/ml. Aliquots of 100µl were placed in test tubes and incubated with fluorochrome-conjugated antibodies using Human MSC Analysis Kit (BD, USA), which includes: CD-44, CD-105, CD-73, CD-90 and a negative cocktail includes: CD-34, CD-11b, CD-19, CD-45 and HLA-DR mix for 30 min in the dark, according to the manufacturer’s instructions. Cells were then centrifuged at 300× g for 5 min and resuspended in 500 µl FACS buffer. The analysis was performed using BD FACSCanto™ and the data were analyzed using Diva software. Multilineage differentiation Adipogenic differentiation was performed using StemPro Adipogenesis differentiation media (Invitrogen, Carlsbad, CA, USA) for 14 days. After that, cells were washed twice with PBS, fixed in 4% formaldehyde for 15 min and stained with oil red O stain, to confirm the presence of adipocytes. StemPro Osteogenic differentiation kit (Invitrogen) was used to induce ADSCs differentiation towards the osteogenic lineage. After 21 days in culture, differentiated cells were washed, fixed in 4% formaldehyde for 15 min, and stained with Alizarin Red S (ARS) stain, to verify the osteogenic differentiation. Cells under normal culture conditions were used as negative controls. Preparation of pooled human platelet lysate (pHPL) Platelet bags designated as platelet-rich plasma 1 (PRP1) were obtained from Jordan University Hospital/blood bank unit. Briefly, platelet bags from 17 donors were pooled in one container and centrifuged at 700× g for 17 min at 18 °C. After centrifugation, the platelets pellet was formed and the supernatant was designated as platelet-poor plasma (PPP). The latter was transferred into new sterile tubes, and platelets obtained from 1ml PRP1 were resuspended in 300µl of PPP, this was designated as PRP2. Following that, platelets concentration was adjusted to 2 ×10 6 platelets/µl with PPP, and lysed through two freeze/thaw cycles at −20 °C and then at 37 °C. Platelet fragments were removed by centrifugation at 3000× g for 20 min at 18 °C and filtrated through a 0. 2 µm filter. The obtained supernatant is now called pHPL. The pHPL was aliquoted and stored at −20 ° C. Upon supplementing the media with 10% pHPL, 2 IU/mL of heparin was added to prevent coagulation. Differentiation of ADSCs into urothelial-like cells When SV-HUC1 cells cultured in F-12K medium reached 80–90% confluency, medium was collected every 24 h and changed into fresh medium for two sequential days. Next, collected medium was centrifuged at 3000 rpm for 5 min, pooled and filtered through a 0. 22 µm filter and stored at −20 °C. For urothelial induction, ADSCs were seeded at a density of 500 cells/cm 2 andcultured with either conditioned or unconditioned medium (UCM) at a ratio of 1:4 F-12K: DMEM, supplemented with either 5% pHPL, 2. 5% pHPL or 10% FBS. Uninduced controls were also included. On day 14 post-induction, cells were utilized for immunofluorescence staining and RNA extraction for relative gene expression analysis. qRT-PCR gene expression analysis of urothelial markers Total RNA was isolated at day 14 of urothelial differentiation using Trizol reagent (Invitrogen, USA) and subsequently purified with RNeasy mini kit (Qiagen, Germantown, MD, USA). cDNA was synthesized using iScript reverse transcription supermix (BioRad, Hercules, CA, USA). Quantitative RT-PCR (qPCR) reactions were performed using iQ™ SYBR mix (BioRad, Hercules, CA, USA) and 300nM of each forward and reverse primers. Primer sequence and product size are provided in Table 1. 10. 7717/peerj. 10890/table-1 Table 1 qPCR primer sequences. Gene Forward Primer Reverse Primer Uroplakin-2 CGGAGAGCCGACAGCAAAC ACTGAGCCGAGTGACTGTGAAG Cytokeratin-18 GGTCAGAGACTGGAGCCATTA GGCATTGTCCACAGTATTTGC Cytokeratin-19 CGGGACAAGATTCTTGGT CCTTGATGTCGGCCTCCA GAPDH CAAGGTTGACAACTTTG GGGCCATCCACAGTCTTCTG Immunofluorescence staining of Urothelial Markers After 14 days of induction, cells on coverslips were fixed in 4% formaldehyde for 15 min and permeabilized with PBS/0. 1% Triton X-100 for 5 min. To prevent nonspecific binding, cells were incubated with blocking solution (3% BSA (wt/v) and 0. 3% Triton X-100 (v/v) in PBS) for 60 min. Cells were then incubated with the primary antibodies against cytokeratin-18 (Abcam, ab668, 1:200) and uroplakin-2 (Santa Cruz, sc-15178, 1:50), diluted in blocking buffer overnight at 4 °C in a humid chamber. Subsequently, cells were incubated with the appropriate secondary antibodies, either chicken anti-mouse IgG –FITC or donkey anti-Goat-IgG Cy3, for 1 h at room temperature followed by counterstaining with DAPI (4 ′, 6-diamidino-2-phenylindole) and mounting with mounting media (Invitrogen). Enzyme linked immunosorbent assay (ELISA) The secretion of epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), and platelet-derived growth factor (PDGF-BB) in conditioned and unconditioned culture medium was measured using ELISA kits (Abcam, Cambridge, UK), according to the manufacturer’s instructions. For growth factors measurement, fresh serum-free media was added on day 14 of urothelial induction and collected after 24 h. Triplicate samples were run in 96-well plates coated with an antibody specific to a particular growth factor mentioned above. The absorbance was measured at 450 nm and within 30 min of completing the assay. Statistical analysis All the experiments were performed at least three times, and statistical analysis was performed using SPSS 20. 0. The data were represented as the mean ± standard error of the mean (SEM) and tested for normality and equal variance before analysis using the Shapiro–Wilk test. Statistical differences were calculated using analysis of variance (ANOVA) and Post-hoc test for comparison between groups. The analysis of ELISA data was performed using Graphpad Prism for curve fitting and independent t -test for significance calculation. Differences were considered significant at P < 0. 05. Results Isolation and Characterization of ADSCs Cells with fibroblastic morphology were adhered to the tissue culture plate and reached the confluency within two weeks of the initial plating ( Fig. 1A ). Flow cytometry staining of the most common MSC markers showed positive expression of the following markers: CD-44 (100%), CD-105 (89. 8%), CD-73 (99. 9%), and CD-90 (100%) ( Figs. 1B – 1E ). To confirm the purity of the isolated ADSCs from the hematopoietic stem cell contamination, flow cytometry staining was performed and demonstrated minimal expression levels of the negative cocktail markers ( Fig. 1F ). To further validate the stemness of the isolated cells, ADSCs were transdifferentiated into the osteogenic and adipogenic cell lineages. Following 21 days of the osteogenic induction, cells exhibited flattened and more elongated morphology with extracellular calcium phosphate deposits as confirmed by Alizarin red S (ARS) staining. These deposits were absent in the uninduced ADSCs cultured in cell culture media ( Figs. 2A – 2D ). Additionally, adipogenic differentiation showed intracellular localization of lipid droplets. These droplets were positively stained with oil red O, and were absent in the uninduced negative control. Thus, the successful differentiation of ADSCs into osteoblasts and pre-adipocytes confirmed the multipotency of these cells ( Figs. 2E – 2H ). 10. 7717/peerj. 10890/fig-1 Figure 1 Characterization and differentiation potential of ADSCs. Characterization and differentiation potential of ADSCs. (A) Primary ADSCs morphology after 14 days in culture under the inverted phase contrast microscope. Scale bar = 100 µm. (B–F) Flow cytometry staining of ADSC markers. Cells showed positive staining for mesenchymal stem cells markers CD-44, CD-105, CD-73, CD-90 and negative for CD-34, CD-11b, CD-19, CD-45 and HLA-DR in the negative cocktail. 10. 7717/peerj. 10890/fig-2 Figure 2 Multilineage differentiation potential of ADSCs. (A–C) Osteogenic mineral deposition was observed after 21 days of osteogenic induction and positively stained with Alizarin Red S stain (ARS). (B–D) Uninduced ADSCs were used as a negative control and stained negatively with ARZ. (E–G) Lipid droplets were observed after 14 days of adipogenic differentiation of ADSCs and positively stained with Oil Red O. (F–H) ADSCs with normal culture media stained negatively for oil red O. Scale bar = 100 µm. All differentiation experiments were repeated at least three times. Urothelial cell markers detection by qRT-PCR To evaluate the capacity of ADSCs to differentiate towards the urothelial-like lineage, cells were cultured with either conditioned (CM) or unconditioned urothelial cell media (UCM), supplemented with either 5% pooled human platelet lysate (pHPL), 2. 5% pHPL or 10% FBS. After 14 days of induction, cells were utilized for further experiments. To analyze the effect of induction CM and UCM media on the differentiation of ADSCs towards the urothelial lineage, we measured the gene expression levels of two cytokeratin proteins; CK-18 and CK-19 and one uroplakin protein (UPK-2), expressed by the urothelial cells ( Fig. 3 ). CK-18 and CK-19 are considered as early markers of urothelial cell specification, and uroplakin proteins representing the terminal maturation stage. The levels of CK-18 expression were not significantly altered in groups treated with CM supplemented with either 5% or 2. 5% pHPL ( P > 0. 05). Meanwhile, a 6. 7-fold upregulation in CK-18 expression compared to the uninduced control was observed in CM FBS treated group ( P = 0. 006) ( Fig. 3A ). Additionally, UCM increased CK-18 expression of 3-fold in the presence of either FBS or pHPL relative to the uninduced cells (P-Value). On the other hand, culturing ADSCs in CM supplemented with FBS exhibited the highest upregulation level of CK-19 with a 4. 2-fold increase compared to the uninduced control ( P = 0. 01). In contrast, CM 2. 5% pHPL and CM 5% pHPL failed to upregulate the expression of CK-19. Whereas, UCM-FBS induced CK-19 expression, but to a lesser extent compared to CM-FBS. On the contrary, UCM supplemented with 5% and 2. 5% pHPL failed to upregulate the expression of CK-19 ( Fig. 3B ). Regarding the UPK-2 terminal differentiation marker, only CM-FBS and UCM-FBS cultures showed an increased level of expression of approximately 2-fold compared to control cells ( Fig. 3C ). 10. 7717/peerj. 10890/fig-3 Figure 3 Relative gene expression of urothelial markers. Real time-PCR performed on urothelial-like cells differentiated from ADSCs, cultivated in either conditioned medium (CM) or unconditioned urothelial cell media (UCM) in the presence of either 5% pHPL, 2. 5% pHPL or 10% FBS. Uninduced ADSCs were used as the calibrator sample. (A) Relative expression of cytokeratin-18, (B) cytokeratin-19 and (C) uroplakin-2. * P < 0. 05, ** P < 0. 01. All experiments were repeated at least three times. Detection of urothelial cell markers by immunofluorescence Since gene expression results suggest an enhanced cellular differentiation using CM culture conditions, we compared the expression of CK-18 (early differentiation marker) and UPK-2 (late differentiation marker) between pHPL and FBS supplemented cultures by immunofluorescence staining ( Fig. 4 ). Staining revealed that CM supplemented with FBS resulted in a 2. 5-fold increase in the expression of CK-18 early differentiation marker and a 2-fold increase in UPK-2 expression compared to non-induced ADSCs control ( Figs. 4A & 4B ). Whereas groups treated with CM & 5% pHPL or 2. 5% pHPL failed to elicit the same response ( Figs. 4A & 4C ). 10. 7717/peerj. 10890/fig-4 Figure 4 Immunofluorescence staining of urothelial markers at the cellular level. (A) ADSCs cultured in CM and different concentrations of pHPL (5% & 2. 5%) or with FBS, were assessed for the expression of cytokeratin-18 (FITC, green) and uroplakin- 2 (Cy3, red). SV-HUC cells were used as a positive control, meanwhile ADSCs were utilized as a negative control. Scale bar = 50 µm. (B & C) Semi-quantitative analysis of immunoflouresence representing percentage of positive cells relative to negative control. All experiments were repeated at least three times. Assessment of growth factor levels of induced ADSCs by ELISA To assess changes in the growth factor levels produced by the induced cells, we measured the levels of EGF, PDGF-BB, and VEGF, the main growth factor proteins secreted by cells into the culture medium. Following 14 days of urothelial induction, measurement of the growth factors in serum-free media collected after 24 h resulted in the detection of higher levels of these factors in cells treated with CM compared to their counterparts cultured in UCM ( Fig. 5 ). Additionally, cells cultured in 5% pHPL produced higher levels of EGF, PDGF-BB, and VEGF. Levels of VEGF were significantly higher in CM-5% pHPL, CM-2. 5% pHPL and UCM-pHPL ( P = 0. 009, 0. 023, 0. 004, respectively). Cells in FBS containing media showed the least amount of secretion of all three growth factors. Whereas, CM-FBS elicited higher levels of growth factors compared to UCM-FBS with 1. 4-fold, 2. 5-fold, and 2. 6-fold difference for EGF, PDGF-BB, and VEGF, respectively ( Fig. 5 ). 10. 7717/peerj. 10890/fig-5 Figure 5 Growth factor levels assessment in induced ADSCs. Enzyme linked immunosorbent assay (ELISA) was preformed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5%pHPL or FBS. (A) Measurement of epidermal growth factor levels (EGF). (B) Platelet derived growth factor-BB (PDGF-BB). (C) Vascular endothelial growth factor (VEGF). * P < 0. 05, ** P < 0. 01. All experiments were repeated at least three times. Discussion Currently, cell-based therapy and tissue engineering studies mostly rely on the ex-vivo expansion and differentiation of many cell types especially stem cells. The preferential use of MSCs over other types of stem cells is related to their ability to cross lineage commitment, they differentiate efficiently into many cell types, and their inability to form teratomas or tumors in vivo ( Qin et al. , 2014 ). Furthermore, MSCs exhibit low immunogenicity and possess immunosuppressive capabilities, facilitating their use in allogenic stem cell transplantation studies ( Klyushnenkova et al. , 2005 ; Le Blanc et al. , 2003 ). However, the clinical use of MSCs is hampered by their limited availability, growth variability, invasive collection procedures, and the use of xenogeneic sources of serum such as FBS during the expansion and differentiation procedures. Conditioned medium is widely used to induce MSC differentiation into chondrocytes, osteocytes, dopaminergic neurons, cholinergic neurons as well as urothelial cells ( Alves da Silva et al. , 2015 ; Heino, Hentunen & Väänänen, 2004 ; Aliaghaei et al. , 2016 ; Borkowska et al. , 2015 ; Zhang et al. , 2014 ). Different types of MSCs have been differentiated efficiently towards urothelial cells using urothelial cell-derived CM or FBS, making them an attractive source for urinary tract tissue regeneration ( Shi et al. , 2012 ; Zhang et al. , 2013 ). However, a cell line-derived CM might contain many undefined factors, which could influence the cells in a myriad of ways. In addition, the use of FBS carries the risk of infectious disease transmission and immunization, which results in restricting the use of such system in a clinical setting ( Heiskanen et al. , 2007 ; Sundin et al. , 2007 ). Thus, for the future translational purposes, we aimed to assess the ability of unconditioned urothelial cell media to substitute cell line-derived CM and pHPL as alternative to FBS, to induce ADSCs differentiation towards urothelial cells. Our results indicate that although ADSCs are capable of expressing early differentiation markers of urothelial like cells such as CK-18 and CK-19 when cultured in CM, this expression is minimal and could not be detected at the cellular level as observed by immunostaining. On the other hand, terminal differentiation phenotype associated with the expression of UPK-2 is only achievable in the presence of CM. This confirms the presence of certain signaling factors in the CM that act in a paracrine manner and are essential for the induction process. Zhang et al. (2013) reported an increased level of expression of a panel of cytokines and growth factors following 12 h of induction of ADSC into urothelial cells. However, after 21 days of induction, the levels normalized slowly until they reached a level similar to the baseline level seen before the induction process ( Zhang et al. , 2013 ). Thus, a critical timing window for the differentiation process is crucial during the early and intermediate stages of differentiation. Since UCM by itself lead to a minimal increase in the urothelial markers, an addition of growth factors and cytokines during this critical timing window might increase the differentiation potential. In a previous study, the addition of FGF-10 to Warton jelly-derived MSC induced their differentiation towards urothelial cells, as evident by the co-expression of CK-8 and UPK-III ( Chung & Koh, 2013 ). In a similar manner, discovering more of such inductive factors paves the way for the generation of a defined differentiation system. Here we report increased levels of EGF, VEGF, and PDGF-BB growth factors when cells were cultured in urothelial cell-derived CM. Several other studies also showed elevated levels of secretion when either ADSCs or BMSCs are cultured under the same conditions ( Tian et al. , 2010 ; Shi et al. , 2012 ; Zhang et al. , 2014 ). EGF plays a crucial role in inducing early ESC differentiation towards the endodermal lineage ( Cras-Méneur et al. , 2001 ; Kumar et al. , 2014 ). However, induction of terminal urothelial differentiation in vitro with PPAR γ activators requires the inhibition of EGF pathway ( Varley et al. , 2004a ; Varley et al. , 2004b ; Varley et al. , 2006 ). Thus, the requirement of EGF might be essential during the early stages of induction, to induce a basal urothelial phenotype and it should be followed by inhibition stage in order to produce the terminal superficial urothelial phenotype. It has been indicated that PDGF can induce proliferation and differentiation of cells originating from all three germ layers including lung, microvilli, gastrointestinal and endothelial cell development based on the normal PDGF signaling ( Boström, Gritli-Linde & Betsholtz, 2002 ; Karlsson et al. , 2000 ; Utoh et al. , 2003 ; Ding et al. , 2004 ; Calver et al. , 1998 ). Additionally, PDGF is known to be secreted by endothelial and epithelial cells and it regulates the proliferation and differentiation of neighboring smooth muscle cells ( Boström, Gritli-Linde & Betsholtz, 2002 ; Barkauskas et al. , 2013 ). On the other hand, VEGF mediates MSC differentiation towards the endothelial cell lineage via the Rho/MRTF-A pathway ( Wang et al. , 2013 ). Although the major role of VEGF is played during vasculoneogensis and cardiac development, it also contributes significantly to the development of organs of endodermal origin including liver, lungs and pancreas ( Carmeliet et al. , 1996 ; Giordano et al. , 2001 ; Lammert, Cleaver & Melton, 2001 ; Matsumoto et al. , 2001 ; Compernolle et al. , 2002 ). Even though the role of these growth factors during the differentiation, development and organogenesis is well-established, a need arises to characterize the extent of involvement of each of these factors in the induction of urothelial cell differentiation. The general consensus on the mechanism of action of conditioned medium is the presence of undefined soluble factors released by the differentiated cells as a way of intercellular communications. These biologically active factors act in a paracrine manner and trigger an internal signaling pathways in MSCs, directing them to differentiate towards a certain lineage ( Alves da Silva et al. , 2015 ; Zhang et al. , 2014 ). These differentiated cells in turn secrete bioactive factors, which act in an autocrine and paracrine manner to maintain the differentiated state of the cells, as well as inducing the differentiation of neighboring cells ( Alves da Silva et al. , 2015 ; Zhang et al. , 2014 ; Shi et al. , 2012 ). In terms of serum choice, we found that pHPL is not an efficient alternative for urothelial differentiation at a concentration of 5% and 2. 5%. Substituting FBS with pHPL during the differentiation reduces the expression of urothelial related markers. Several studies reported an increase in the differentiation potential of several types of MSCs towards the osteogenic, myofibroblastic, cardiomyogenic and adipogenic cell lineages in the presence of pHPL as a serum substitute ( Karadjian et al. , 2020 ; Samuel et al. , 2016 ; Chignon-Sicard et al. , 2017 ; Homayouni Moghadam, Tayebi & Barzegar, 2016 ). On the other hand, pHPL decreased the ability of certain types of stem cells to differentiate into osteoblasts, chondrocytes and adipocytes ( Gruber et al. , 2004 ; Lee et al. , 2014 ; Chignon-Sicard et al. , 2017 ). Thus, the response elicited towards the presence of pHPL in the culture media is affected by the stem cell type, the differentiation lineage, and the percentage of pHPL used in the culture media. MSC including ADSCs have an increased proliferative capacity in the presence of pHPL ( Karadjian et al. , 2020 ; Naaijkens et al. , 2012 ), ( Trojahn Kølle et al. , 2013 ). As urothelial cells differentiate from the basal layer towards the superficial layer, they lose their ability to regenerate. Thus, the molecular circuitry governing differentiation depends on a skew towards differentiation by reducing proliferation, which probably could not be achieved in the presence of pHPL. In conclusion, the induction of ADSCs towards the urothelial phenotype requires the presence of both CM and FBS. Substituting CM with UCM and FBS with pHPL significantly impacts the differentiation process. The levels of EGF, VEGF, and PDGF have increased during the differentiation process and thus might play an essential role in defining the mechanism of action of CM directed differentiation. Additionally, we found that pHPL at 5% and 2. 5% concentrations negatively influenced the differentiation process. This might be explained by a skew in cellular circuitry towards the proliferation rather than the differentiation process. Changing MSCs microenvironment to accommodate the urothelial cells microenvironment is a basic requirement for any successful differentiation protocol. This can be achieved by the addition of defined soluble factors, direct or indirect co-culture with differentiated cells, the use of conditioned media or direct implantation in vivo. Urothelial cell-derived CM represents a practical and efficient way for ADSCs differentiation into urothelial lineage. However, a demand arises to formulate a defined media that efficiently induces MSC differentiation towards the urothelial lineage for clinical purposes. Supplemental Information 10. 7717/peerj. 10890/supp-1 Supplemental Information 1 EGF ASC Raw data of enzyme linked immunosorbent assay (ELISA) measurements for epidermal growth factor levels (EGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5% pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-2 Supplemental Information 2 EGF Figure Figure generation setup for enzyme linked immunosorbent assay (ELISA) measurements of epidermal growth factor levels (EGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5%pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-3 Supplemental Information 3 PDGF Figure Figure generation setup for enzyme linked immunosorbent assay (ELISA) measurements for platelet derived growth factor levels (PDGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5%pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-4 Supplemental Information 4 PDGF ASC Raw data of enzyme linked immunosorbent assay (ELISA) measurements for platelet derived growth factor levels (PDGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5%pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-5 Supplemental Information 5 VEGF ANOVA One way ANOVA analysis file for enzyme linked immunosorbent assay (ELISA) measurements for vascular endothelial growth factor (VEGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5% pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-6 Supplemental Information 6 VEGF ASCs Raw data of enzyme linked immunosorbent assay (ELISA) measurements for vascular endothelial growth factor (VEGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5%pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-7 Supplemental Information 7 Final data analysis and graphing Click here for additional data file. 10. 7717/peerj. 10890/supp-8 Supplemental Information 8 VEGF Figure Figure generation setup for raw data of enzyme linked immunosorbent assay (ELISA) measurements for vascular endothelial growth factor (VEGF) performed on conditioned media collected after 24 h from cells induced for 14 days with urothelial cell derived conditioned medium (CM) or unconditioned urothelial cell media (UCM) with 5% pHPL, 2. 5%pHPL or FBS. Click here for additional data file. 10. 7717/peerj. 10890/supp-9 Supplemental Information 9 Output K18 IF Click here for additional data file. 10. 7717/peerj. 10890/supp-10 Supplemental Information 10 IF Data analysis Click here for additional data file. 10. 7717/peerj. 10890/supp-11 Supplemental Information 11 Output UP-2 Click here for additional data file.

10. 7717/peerj. 10929

PeerJ

Genes for degradation and utilization of uronic acid-containing polysaccharides of a marine bacterium

Background Oligosaccharides from polysaccharides containing uronic acids are known to have many useful bioactivities. Thus, polysaccharide lyases (PLs) and glycoside hydrolases (GHs) involved in producing the oligosaccharides have attracted interest in both medical and industrial settings. The numerous polysaccharide lyases and glycoside hydrolases involved in producing the oligosaccharides were isolated from soil and marine microorganisms. Our previous report demonstrated that an agar-degrading bacterium, Catenovulum sp. CCB-QB4, isolated from a coastal area of Penang, Malaysia, possessed 183 glycoside hydrolases and 43 polysaccharide lyases in the genome. We expected that the strain might degrade and use uronic acid-containing polysaccharides as a carbon source, indicating that the strain has a potential for a source of novel genes for degrading the polysaccharides. Methods To confirm the expectation, the QB4 cells were cultured in artificial seawater media with uronic acid-containing polysaccharides, namely alginate, pectin (and saturated galacturonate), ulvan, and gellan gum, and the growth was observed. The genes involved in degradation and utilization of uronic acid-containing polysaccharides were explored in the QB4 genome using CAZy analysis and BlastP analysis. Results The QB4 cells were capable of using these polysaccharides as a carbon source, and especially, the cells exhibited a robust growth in the presence of alginate. 28 PLs and 22 GHs related to the degradation of these polysaccharides were found in the QB4 genome based on the CAZy database. Eleven polysaccharide lyases and 16 glycoside hydrolases contained lipobox motif, indicating that these enzymes play an important role in degrading the polysaccharides. Fourteen of 28 polysaccharide lyases were classified into ulvan lyase, and the QB4 genome possessed the most abundant ulvan lyase genes in the CAZy database. Besides, genes involved in uronic acid metabolisms were also present in the genome. These results were consistent with the cell growth. In the pectin metabolic pathway, the strain had genes for three different pathways. However, the growth experiment using saturated galacturonate exhibited that the strain can only use the pathway related to unsaturated galacturonate.

Introduction Uronic acids are a class of sugar acids oxidized the hydroxyl group on C6 of aldoses. Uronic acids including D-glucuronic acid, D-galacturonic acid, D-mannuronic acid, L-guluronic acid, and L-iduronic acid are components of polysaccharides produced by animals (heparin), terrestrial plants (pectin), seaweed (alginate and ulvan), and bacteria (gellan gum) ( De Lederkremer & Marino, 2003 ). In general, these polysaccharides such as pectin, alginate, ulvan, and gellan gum, have broad potential in many applications due to their excellent properties of biocompatibility, non-toxic, immunogenicity, availability, and relatively low cost ( Morelli & Chiellini, 2010 ; Venkatesan et al. , 2015 ; Rahman, Dafader & Banu, 2017 ). Pectin is a polymer with a linear structure characterized by a backbone consisting of a few hundred to thousand D-galacturonic acid units linked together by α-(1 →4)-glycosidic linkages. It is found in the cell walls of the plant and intracellular layer of plant cells, mainly fruits, such as apples, oranges, and lemons ( Mudgil, 2017 ). Pectin contains a significant amount of neutral sugar, typically L-rhamnose, L-arabinose, D-galactose, D-xylose, and D-glucose, which are linked to the hydroxyl groups on the number 2 and 3 carbons of the main chain. Pectin attracted attention due to its gelling capabilities ( Penhasi & Meidan, 2015 ). Pectin is widely used as a thickener and stabilizing agent in the food industry ( Munarin, Tanzi & Petrini, 2012 ). Alginate, also known as alginic acid is an unbranched polymer composed of β-D-mannuronic acid (M) and α-L-guluronic acid (G), which are covalently (1-4)-liked. The residues are randomly arranged into MM-, GG- and MG-blocks ( Meng & Liu, 2013 ). Alginate is distributed widely in the cell wall of marine brown agar and has long been used in the industry such as the medical field, fabric, food, and beverage industries as thickening, gel-forming, and colloidal stabilizing agents ( Liakos et al. , 2013 ; Martău, Mihai & Vodnar, 2019 ). Ulvan is water-soluble polysaccharides found in the cell wall of green algae ( Ulva and Enteromorpha ) composed mainly of 3-sulfated rhamnose (Rha3S), glucuronic acid (GlcA), iduronic acid (IdoA), and xylose (Xyl) ( Kim, Thomas & Li, 2011 ). Ulvan has attracted pharmaceutical and medical applications for its anti-viral, anti-coagulant, and anti-proliferative activities towards cancer cells and its immune-stimulating properties ( Alves, Sousa & Reis, 2013 ). In addition, ulvan also showed to be an activator of plant defense and an inducer of plant resistance to fungal diseases ( Alves, Sousa & Reis, 2013 ). Therefore, it is a good potential for agricultural applications. Gellan gum is an exopolysaccharide produced from non-pathogenic, Gram-negative bacterium, Sphingomonas elodea (earlier Pseudomonas elodea ) using aerobic fermentation ( Vendrusculo, Pereira & Scamparini, 1994 ). Gellan gum consisting of repeating tetrasaccharide units of glucose, glucuronic acid, and rhamnose residues in a 2:1:1 ratio: [→3)- β-D-glucose-(1 →4)- β-D-glucuronic acid-(1 →4)- β-D-glucose-(1 →4)- α-L-rhamnose-(1 →] ( Jansson, Lindberg & Sandford, 1983 ). Gellan gum has been used in medicine, pharmaceutical formulations, cosmetics, or tissue engineering. As a biocompatible polysaccharide, gellan gum is used in contact with or inside the body. Besides, gellan gum is also useful in the food and biotechnology industry as immobilization of enzymes and yeast cells ( Iurciuc, Tincu)et al. (2015 ). These polysaccharides could also be alternative sustainable sources for fermentative biofuel production ( John et al. , 2011 ). The PLs and GHs play an important role in the saccharification of the polysaccharides in the process of biofuel production ( Edwards et al. , 2011 ; Takeda et al. , 2011 ; Li et al. , 2015 ). For example, endo- and exolytic alginate lyases from Saccharophagus degradans A1 were co-displayed on the yeast cell surface, and the co-displaying yeasts were able to effectively produce monosaccharides ( Takagi et al. , 2016b ). On the other hand, oligosaccharides from uronic acids generated by PLs are known to have multiple biological activities. For instance, alginate oligosaccharides stimulate the growth of human endothelial ( Kawada et al. , 1997 ) and keratinocytes cells ( Kawada et al. , 1999 ). In addition, the oligosaccharides also promote the growth and root elongation of rice and barley ( Tomoda, Umemura & Adachi, 1994 ; Hien et al. , 2000 ). Mandalari and colleagues reported that pectin oligosaccharides (POS) have probiotic effects through the improvement of bifidobacteria and lactobacillus ( Mandalari et al. , 2007 ). Besides, POS inhibited inflammation, fibrosis formation, as well as cancer progression, transformation, and metastasis ( Bonnin, Garnier & Ralet, 2014 ). Thus, PLs and GHs have attracted considerable interest in both academic and commercial spheres. Many bacterial species, such as the genera Agrobacterium, Bacillus, Cellulophaga, Clostridium, Erwinia, Flammeovirga, Flavobacterium, Microbulbifer, Pseudoalteromonas, Pseudomonas, Saccharophagus, Sphingomonas, Vibrio, Xanthomonas, and Zobellia, are known as bacteria that are capable of degrading uronic acid-containing polysaccharides, mainly alginate and pectin ( Liu et al. , 2019b ; Dubey et al. , 2016 ; Takagi et al. , 2016a ; Takagi et al. , 2016b ). The genus Catenovulum consisted of three species, C. agarivorans, C. maritimum, and C. sediminis, was also known to degrade agar (all three strains) and alginate ( C. maritimum, and C. sediminis ). Besides that, Catenovulum sp. LP was able to produce an ulvan lyase ( Li et al. , 2015 ; Qiao et al. , 2020 ). However, the degradation and utilization pathway of uronic acid-containing polysaccharides containing alginate, gellan gum, pectin, and ulvan by the genus Catenovulum are poorly understood. Our group isolated Catenovulum sp. CCB-QB4 (referred to hereafter as QB4) from Queens Bay of Penang, Malaysia, and the complete genome sequence was reported ( Lau et al. , 2019 ). From the study, it was reported that the QB4 genome contained 183 GHs and 43 PLs. Based on the information, this bacterium is predicted to have the ability to utilize many polysaccharides. To confirm the expectation, in this study, the cell growth of QB4 in the presence of polysaccharides containing uronic acids, namely, alginate, pectin, ulvan, and gellan gum, was confirmed using a shake flask fermentation method. In addition, genes involved in degrading and utilizing these polysaccharides were explored in the QB4 genome. This study is the first report to describe the degradation and potential metabolic pathways for utilization of four different uronic acid-containing polysaccharides in the genus Catenovulum. Materials and Methods Strain and chemicals The QB4 cells ( Lau et al. , 2019 ) was cultured using high nutrient artificial seawater medium (H-ASWM) [0. 5% tryptone, 2. 4% (w/v) artificial sea salt mix (Marine Enterprises International), 10 mM HEPES, pH 7. 6], as reported by Furusawa and co-workers ( Furusawa et al. , 2015 ). A total of 0. 2% of polygalacturonic acid (pectic acid) (Nacalai Tesque), sodium alginate (Sigma-Aldrich), Gelzan™ CM (gellan gum) (Sigma-Aldrich), and ulvan were used as carbon source throughout the research. Purification of ulvan The ulvan was purified according to the method described by Tabarsa et al. (2012) with slight modification. 5 g of the dry powder of Ulva pertusa was dissolved in 100 mL of water and stirred at 65 °C for 3 h. The mixture was cooled and centrifuged at 10, 000 g for 20 min at 15 °C using Sorvall™ RC 6 Plus Centrifuge (ThermoFisher Scientific). 4 volume of cold isopropanol was added into the supernatant to precipitate crude polysaccharide. The solution was left overnight at 4 °C. The crude ulvan was harvested by filtration, washed with 70% isopropanol several times, and dried overnight at 60 °C. Determination of bacterial growth The QB4 cells were inoculated in 10 mL of H-ASWM broth and cultured overnight at 30 °C. 0. 1 mL of the cell suspension was inoculated into 100 mL of H-ASWM medium with 0. 2% of each uronic acid as the carbon sources. Media with and without 0. 2% glucose (Fisher Scientific) were used as positive and negative controls, respectively. Each flask was inoculated with 0. 1 mL of pre-cultured bacterial cell suspension and incubated at 30 °C with an agitation speed of 200 rpm on the orbital shaker. To monitor the bacterial growth, the colony-forming unit (CFU) counts were conducted by culturing the cells on H-ASWM agar plates because the optical density measurement was not suitable for the sample with gellan gum in which the broth was solidified by calcium ion contained in H-ASWM medium. 100 l of the cell suspension from each sample was collected every 3 h. The suspension was diluted into 900 l of H-ASWM medium. After that, sequential 10-fold serial dilutions were made, and 100l of aliquots of each dilution were plated on H-ASWM agar plates. Colonies were scored after incubation for 48 h at 30 °C, and the growth curve was constructed. The experiment was performed in triplicates. The generation time (G) was estimated by the method described by Aparna et al. based on the growth curve ( Aparna, Parvathi & Kaniyassery, 2020 ). The number of generation (n) was determined by the method of n = (logb –logB)/log2, where B and b are the numbers of bacteria at the beginning of a time interval and at the end of the time interval, respectively. The generation time (G) was determined by the method of G = t/n, where t is the time interval (min). To confirm the utilization of saturated galacturonate, the QB4 cells were culture in an H-ASWM medium with 0. 2% saturated galacturonate. Media with and without 0. 2% glucose was used as positive and negative controls, respectively. To monitor the bacterial growth, the optical density (OD 600nm ) was measured at 3 h intervals for 30 h using UV spectrophotometer UV-1800 (Shimadzu). The experiment was performed in triplicates. Genomic analysis of Catenovulum sp. CCB-QB4 The complete genome of QB4 deposited at GenBank under the accession number CP026604 – CP026605 was determined by Lau et al. (2019). Genes involved in the carbohydrate-active enzyme (CAZymes) in QB4 were predicted using dbCAN pipelines ( Yin et al. , 2012 ). Several enzymes related to uronic acid metabolism in QB4 were predicted by Blastp at the National Center for Biotechnology Information (NCBI) server (Bethesda, MD, U. S. A. ) and were found using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database ( Kanehisa & Goto, 2000 ). Finally, the amino acid sequence similarity of all enzymes was confirmed using Blastp analysis with Protein Data Base (PDB). Signal peptide prediction was conducted by LipoP 1. 0 ( Juncker et al. , 2003 ) and SignalP 5. 0 ( Armenteros et al. , 2019 ). For sequence alignment of DEH reductase, the amino acid sequence of A1-R and A1-R’ from Sphingomonas sp. A1 and DEH reductase from Saccharophagus degradans 2-40 were obtained from GenBank ( https://www. ncbi. nlm. nih. gov/genbank/ ). The sequence alignment with 4 sequences was conducted by ClustalW ( Thompson, Higgins & Gibson, 1994 ) at PRABI Lyon-Gerland ( https://npsa-prabi. ibcp. fr/cgi-bin/npsa_automat. pl?page=/NPSA/npsa_clustalw. html ). Results Growth confirmation of QB4 using uronic acids To confirm the growth of QB4 in the presence of uronic acids, the cells of QB4 were cultured in H-ASWM with alginate, pectin, ulvan, and gellan gum. As shown in Fig. 1, the cells of QB4 showed good growth in the presence of the uronic acid-containing polysaccharides and glucose except for negative control. The cells of QB4 reached the early stationary phase at 9 h in H-ASWM with alginate and gellan gum. The generation time of alginate and gellan gum was 76. 76 and 74. 07 min. However, the cell number of the sample with gellan gum was 5. 7 times lower than that of the sample with alginate. On the other hand, the cells cultured by pectin and glucose reached the early stationary phase at 18 h. In the case of using ulvan, the cells reached stationary phase at 24 h, and the sample exhibited the highest cell number at 30 h in the experiment. The generation time of the samples with glucose, pectin, and ulvan was 91. 37, 95. 74, and 85. 10 min, which were longer than that of the samples with alginate and gellan gum. Thus, QB4 cells exhibited a robust growth in the presence of uronic acids. Subsequently, we focused on genes for degradation and utilization of uronic acids in the genome of QB4. 10. 7717/peerj. 10929/fig-1 Figure 1 Growth confirmation of QB4 cultured with four different uronic acid-containing polysaccharides, alginate (Alg), pectin (Pct), Ulvan (Ulv), and gellan gum (Gel). The growth was measured by cfu/mL. Media with and without 0. 2% glucose was used as positive (Glu) and negative controls (NC), respectively. All data shown are mean values from three replicate experiments. Error bars denote the standard deviation of triplicate samples. Genes for alginate degradation and utilization Alginate is a major polysaccharide found in the cell wall of brown algae and consisted of guluronate and mannuronate arranged as 1, 4-linked polysaccharides. First, alginate is degraded into oligomeric or monomeric units by alginate lyases. Eight alginate lyase genes were found in the QB4 genome ( Table 1 ). Ad1-PL6, Ad2-PL6 and Ad3-PL6, Ad4-PL7, Ad5-PL7, Ad6-PL7, Ad7-PL7, and Ad8-PL17 were classified into family PL6, PL7, and PL17 based on the CAZy database, respectively. As a result of Blastp search with Protein Data Bank (PDB), Ad proteins were similar to Alygc from Paraglaciecola chathamensis (Ad1-PL6 and Ad3-PL6), AlyQ from Persicobacter sp. CCB-QB2 (Ad2-PL6, Ad4-PL7, and Ad5-PL7), alginate lyase of Klebsiella pneumoniae (Ad6-PL7 and Ad7-PL7), and Saccharophagus degradans 2-40 (Ad8-PL17) ( Table S1 ). LipoP 1. 0 ( Juncker et al. , 2003 ) was used to predict the signal peptides and their type of each enzyme. Table 1 demonstrated that Ad1-PL6, Ad2-PL6, Ad4-PL7and Ad5-PL7 possessed type I signal peptide, which was cleaved by signal peptidase I, and Ad6-PL7, Ad7-PL7, and Ad8-PL17 have type II lipoprotein signal peptide, which was cleaved by signal peptidase II. The lipoprotein signal peptide referred to as “lipobox” plays an important role in anchoring the protein on the outer surface of the cell membrane after secretion and modification of N-terminal cysteine residue ( Pugsley, Chapon & Schwartz, 1986 ; Hutcheson, Zhang & Suvorov, 2011 ). This suggested that Ad6-PL7, Ad7-PL7, and Ad8-PL17 may localize on the cell surface. In contrast, Ad1-PL6, Ad2-PL6, Ad4-PL7, and Ad5-PL7 may be released into the culture medium. On the other hand, Ad3-PL6, which does not possess any signal peptides, may locate at the cytoplasm. 10. 7717/peerj. 10929/table-1 Table 1 Genes involving in alginate metabolism. Alginate-degrading enzymes Abbreviation Function CAZy Sp. GenBank Ad1_PL6 Alginate lyase PL6 Type I WP_108604939. 1 Ad2_PL6 Alginate lyase PL6, CBM32, CBM32 Type I WP_108601791. 1 Ad3_PL6 Alginate lyase PL6, CBM16 - WP_108605000. 1 Ad4_PL7 Alginate lyase PL7, CBM32 Type I WP_159084278. 1 Ad5_PL7 Alginate lyase PL7, CBM32, CBM32 Type I WP_108602212. 1 Ad6_PL7 Alginate lyase PL7 Lipobox WP_108602720. 1 Ad7_PL7 Alginate lyase PL7 Lipobox WP_108603319. 1 Ad8_PL17 Alginate lyase PL17 Lipobox WP_108601502. 1 Alginate-utilization proteins Abbreviation Function GenBank Au1 DEH reductase WP_108601028. 1 Au2 KDG kinase WP_108601027. 1 Au3 KDG kinase WP_108603513. 1 Au4 2-dehydro-3-deoxyphosphogluconate aldolase WP_108602314. 1 Au5 2-dehydro-3-deoxyphosphogluconate aldolase WP_108603883. 1 Notes. Sp. indicates signal peptide. Takase et al. (2010) described that an exotype oligoalginate lyase, A1-IV, from Sphingomonas sp. A1 degraded oligoalginates into monosaccharides, which are then nonenzymatically converted to 4-deoxy-L-erythro-5-hexoseulose uronic acid (DEH). Takagi and colleagues also reported that an alginate lyase, Alg7K, from Saccharophagus degradans showed exolytic activity and produced monosaccharides from oligoalginates ( Takagi et al. , 2016a ). The amino acid sequence of the alginate lyase, Ad6-PL7, showed high similarity (77. 3%) to Alg7K, suggesting that Ad6-PL7 might mediate hydrolysis of oligoalginate to produce monomers. The DEH is converted to 2-keto-3-deoxy-D-gluconate (KDG) by NADH or NADPH-dependent DEH reductase ( Takase et al. , 2010 ; Takase et al. , 2014 ; Kim et al. , 2016 ). A result of the BLASTp search demonstrated that the SDR family oxidoreductase, Au1, found in the QB4 genome ( Table 1 ) showed high similarity to NADH (A1-R’, 60. 9%) and NADPH (A1-R, 50. 8%)-dependent DEH reductase. In addition, the gene aliments of DEH reductases showed that the TGXXXGX motif and catalytic triad (Ser, Tyr, and Lys), which are highly conserved in SDR family enzymes ( Takase et al. , 2014 ), were conserved in the SDR family oxidoreductase ( Fig. S1 ). These results indicated that Au1 might function as DEH reductase. KDG kinase catalyzes the conversion of KDG to 2-keto-3-deoxy-phosphogluconate (KDPG), and then KDGP is converted into D-glyceraldehyde-3-phosphate and pyruvate by 2-dehydro-3-deoxy-phosphogluconate aldolase via the Entner -Doudoroff pathway. Two of the genes encoding the KDG kinase, Au2 and Au3, and the aldolase, Au4, and Au5, were found in the genome of QB4 ( Table 1 ). Finally, pyruvate produced by the pathway goes into further metabolic pathways for generating energy. The five Au proteins also demonstrated a high degree of amino acid sequence similarity (>50%) to DEH reductase A1-R’ from Sphingomonas sp. A1 (Au1), KDG kinase Shigella flexneri (Au2 and 3), 2-keto-3-deoxy-6-phosphogluconate aldolase from Thermotoga maritima (Au4), and KDGP aldolase from Escherichia coli (Au5) ( Table S1 ). As mentioned above, QB4 possessed many alginate lyases and enzymes responsible for alginate utilization. This is consistent with the robust growth of QB4 in the presence of alginate, as shown in Fig. 1. Genes for pectin degradation and utilization As a first step, pectin is depolymerized by pectin lyases or polygalacturonases. Pectin and pectate lyases are classified into five families of PLs. Although pectin lyases attack highly methyl-esterified pectin, pectate lyases specifically attack non-methylated polygalacturonate or methylated pectin with a very low degree ( Hugouvieux-Cotte-Pattat, Condemine & Shevchik, 2014 ). These enzymes degrade pectin to unsaturated pectic-oligosaccharides and disaccharides. In the CAZy database, seven pectate lyases belonging to family PL1 (Pd1_PL1, Pd2_PL1, Pd3_PL1, Pd4_PL1, and Pd5_PL1), PL3 (Pd6_PL3), and PL10 (Pd7_PL10) family pectate lyases were found in the genome of QB4. The result of Blastp search using the PDB database on Pd proteins also demonstrated that these Pd proteins were similar (>45%) to pectate lyase (Pd1_PL1, Pd2_PL1, Pd3_PL1, Pd5_PL1, Pd6_PL3, and Pd7_PL10) and pectinesterase (Pd4_PL1) from other bacterial species ( Table S2 ). CBM13 found in Pd1_PL1 and Pd6_PL3 and CBM35 found in Pd2_PL1 and Pd3_PL1 were capable of binding to multi-ligands, such as arabinan, arabinoxylan, and pectin ( Fujimoto, 2013 ; Dhillon et al. , 2018 ). However, CBM77 contained in Pd2_PL1 recognized homogalacturonan ( Fujimoto, 2013 ). Pectin methylesterase (PME) domain from family 8 Carbohydrate Esterase (CE8) was found in Pd4_PL1 and Pd5_PL1. PMEs produce de-esterified homogalacturonan by catalyzing the de-esterification of the methoxyl group of the pectin; as a result, the products are effectively degraded by pectate lyases ( Kashyap et al. , 2001 ). Besides, although Pd1_PL1, Pd3_PL1, Pd4_PL1, and Pd7_PL10 possessed type I signal peptide, Pd5_PL1 and Pd6_PL3 had lipobox, indicating that both two lyases localize on the cell surface. Hence, these results suggested that Pd5_PL1 plays an important role in the pectin degradation of QB4 due to its de-esterification and depolymerization activities and its localization. The unsaturated disaccharides generated by the pectate lyases are converted to 5-keto-4-deoxyuronate (DKI) by two distinct processes. One is that the disaccharides are degraded by oligogalacturonate lyase belonging to family PL22, and then the product, Δ-4, 5-unsaturated galacturonate, is linearized into DKI by KdgF ( Hobbs et al. , 2016 ). The other is that unsaturated galacturonyl hydrolases belonging to GH105 degrades the disaccharides and directly releases DKI ( Hobbs et al. , 2019 ). In the QB4 genome, no PL22 enzymes were found in the genome. However, three GH105 proteins, Pd8_GH105, Pd9_GH105, and Pd10_GH105, were present in the genome ( Table 2 ). These enzymes showed a high degree of similarity (49∼67%) on unsaturated rhamnogalacturonyl hydrolases (YteR) based on Blastp search with PDB database. These results indicated that oligogalacturonate might directly convert to DKI by GH105 proteins in QB4. DKI is converted to 2-keto-3-deoxygluconate (KDG) by two enzymes, a DKI isomerase (KduI) and a 2-dehydro-3-deoxy-D-gluconate 5-dehydrogenase (KduD) ( Hobbs et al. , 2019 ). The pathway analysis based on KEGG showed that four of KduI (Pu1, Pu2, Pu3, and Pu4) and KduD (Pu5, Pu6, Pu7, and Pu8) were present in the genome of QB4. Blastp search with the PDB database also exhibited that these enzymes showed a high degree of similarity (63∼71%) on KduD and KduI from Enterococcus faecalis, Escherichia coli, and a pectolytic bacterium, Pectobacterium carotovorum ( Table S2 ). These results indicated that QB4 uses pectin as a carbon source for cell growth and is consistent with the result shown in Fig. 1. 10. 7717/peerj. 10929/table-2 Table 2 Genes involving in pectin metabolism. Pectin-degrading enzymes Abbreviation Function CAZy Sp. GenBank Pd1_PL1 Pectate Lyase (Plasmid) PL1, CBM13 Type I WP_108605188. 1 Pd2_PL1 Pectate Lyase (Plasmid) PL1, CBM35, CBM77 - WP_108605291. 1 Pd3_PL1 Pectate Lyase (Plasmid) PL1, CBM35 Type I AWB69198 Pd4_PL1 Pectate Lyase (Plasmid) PL1, CE8 Type I WP_159084287. 1 Pd5_PL1 Pectate Lyase (Plasmid) PL1, CE8 Lipobox WP_108605272. 1 Pd6_PL3 Pectate Lyase (Plasmid) PL3, CBM13 Lipobox WP_108605187. 1 Pd7_PL10 Pectate Lyase (Plasmid) PL10 Type I WP_108605295. 1 Pd8_GH105 Unsaturated galacturonyl hydrolases (Plasmid) GH105 Lipobox WP_108605228. 1 Pd9_GH105 Unsaturated galacturonyl hydrolases (Plasmid) GH105 Type I WP_108605230. 1 Pd10_GH105 Unsaturated galacturonyl hydrolases (Plasmid) GH105 Lipobox WP_108605292. 1 Pd11_GH28 Polygalacturonases (Plasmid) GH28 Tat WP_108605238. 1 ] Pd12_GH28 Polygalacturonases (Plasmid) GH28 Type I WP_108605277. 1 Pectin-utilization proteins Abbreviation Function GenBank Pu1 5-dehydro-4-deoxy-D-glucuronate isomerase (KduI) WP_108601704. 1 Pu2 5-dehydro-4-deoxy-D-glucuronate isomerase (KduI) WP_108601929. 1 Pu3 5-dehydro-4-deoxy-D-glucuronate isomerase (KduI) WP_108602166. 1 Pu4 5-dehydro-4-deoxy-D-glucuronate isomerase (KduI) (Plasmid) WP_108605242. 1 Pu5 2-dehydro-3-deoxy-D-gluconate 5-dehydrogenase (KduD) WP_108601158. 1 Pu6 2-dehydro-3-deoxy-D-gluconate 5-dehydrogenase (KduD) WP_108601705. 1 Pu7 2-dehydro-3-deoxy-D-gluconate 5-dehydrogenase (KduD) WP_108602165. 1 Pu8 2-dehydro-3-deoxy-D-gluconate 5-dehydrogenase (KduD) (Plasmid) WP_108605243. 1 Pu9 Glucuronate isomerase WP_108601474. 1 Pu10 Glucuronate isomerase WP_108604964. 1 Pu11 Glucuronate isomerase (Plasmid) WP_108605248. 1 Pu12 Tagaturonate reductase WP_108601943. 1 Pu13 Altronate hydrolase WP_108601944. 1 Pu14 Tagaturonate epimerase (Plasmid) WP_108605252. 1 Pu15 Fructuronate reductase WP_108602164. 1 Pu16 Mannoate dehydratese WP_108601469. 1 Notes. Plasmid indicated that the gene is present in plasmid, not the genome. The pathway analysis also showed the other pathway using saturated galacturonate (monosaccharide) produced by polygalacturonases. As mentioned above, although QB4 possesses seven pectate lyases, only two polygalacturonases (GH28), Pd11_GH28 and Pd12_GH28, which were highly similar (>66%) to polygalacturonase from pectolytic bacteria, such as Erwinia carotovora and Thermotoga marítima ( Table S2 ), were found in the genome. Based on SignalP 5. 0 analysis, it was found that Pd11_GH28 and Pd12_GH28 contained Tat and type I signal peptide, respectively. Saturated galacturonate is converted to D-tagaturonate by glucuronate isomerase. Three glucuronate isomerases, Pu9, Pu10, and Pu11, which were highly similar (> %) to glucuronate isomerase of Salmonella enterica subsp. enterica serovar Typhimurium and uronate isomerase of Caulobacter vibrioides CB15 ( Table S2 ), were found in the QB4 genome. The conversion of D-tagaturonate to KDG occurs through two distinctly different pathways. One is that the process consists of two steps involving tagaturonate reductase and altronate hydrolase and one intermediate, D-altronate ( Richard & Hilditch, 2009 ). The other is that tagaturonate is converted via three steps involving tagaturonate epimerase, fructuronate reductase, and mannonate dehydratase and two intermediates, D-fructuronate and D-mannonate ( Valk et al. , 2020 ). As shown in Table 2, QB4 possessed all genes involved in the two different pathways, namely tagaturonate reductase (Pu12), altronate hydrolase (Pu13), tagaturonate epimerase (Pu14), fructuronate reductase (Pu15), and mannoate dehydratase (Pu16). Homologous enzymes of these proteins were also found in the PDB database ( Table S2 ). This result suggested that QB4 is also capable of using saturated galacturonate as a carbon source. In order to confirm this suggestion, the QB4 cells were cultured in H-ASWM broth with saturated galacturonate. As shown in Fig. 2, although the cells with glucose demonstrated robust growth, the cells were unable to grow in the broth with saturated galacturonate as well as the negative control in the incubation period. This result indicated that the unsaturated galacturonate utilization pathway is the main pathway to utilize polygalacturonic acid of QB4. Interestingly, genes for all pectate lyases, polygalacturonases, and one set of KduI (Pu4), KduD (Pu8), glucuronate isomerase (Pu11), and tagaturonate reductase were located in a plasmid ( https://www. ncbi. nlm. nih. gov/nuccore/NZ_CP026605. 1 ) found in QB4 ( Table 2 ). Genes for ulvan degradation and utilization Ulvan lyases classified into 5 families, PL24, PL25, PL28, PL 37, and PL40, in the CAZy database ( Li et al. , 2020 ) were isolated from several marine bacteria, such as genera, Alteromonas, Pseudoalteromonas, Formosa, and Nonlabens ( Kopel et al. , 2016 ; Qin et al. , 2018 ; Ulaganathan et al. , 2018a ; Ulaganathan et al. , 2018b ; Reisky et al. , 2019 ). Ulvan lyases cleave between L-rhamnose 3-sulfate (Rha3S) and D-glucuronic acid (GlcA) or L-iduronic acid (IdoA). Fourteen ulvan lyases in QB4 were classified into three PLs families, PL24 (Ud1_PL24, Ud2_PL24, Ud3_PL24, Ud4_PL24, Ud5_PL24, Ud6_PL24, Ud7_PL24, Ud8_PL24, and Ud9_PL24), PL25 (Ud10_PL25, Ud11_PL25, Ud12_PL25, and Ud13_PL25) and PL40 (Ud14_PL40). Ud proteins belonging to the PL24 family were highly similar (>62%) to short ulvan lyase of Alteromonas sp. LOR, and Ud proteins belonging to the PL25 family were highly similar (>57%) to ulvan lyase-PL25 of Pseudoalteromonas sp. PLSV ( Table S3 ). The number of the ulvan lyases was higher than that of Formosa agariphila (three PL40 proteins) ( Reisky et al. , 2019 ) and Alteromonas sp. LOR (one PL24 and one nonclassified protein) ( Foran et al. , 2017 ). Ud6_PL24 and Ud9_PL24 and the enzymes belonging to PL25 possessed lipobox. Although Ud2_PL24, Ud3_PL24, Ud4_PL24, Ud5_PL24, and Ud8_PL24 contained type I signal peptide, the remaining enzymes did not have any signal peptides. Although only one pectin lyase had a lipobox, many ulvan lyases possessed lipobox as well as alginate lyases, indicating that alginate and ulvan degradation activity of QB4 may be more effective than pectin degradation of the strain. 10. 7717/peerj. 10929/fig-2 Figure 2 Growth confirmation of QB4 cultured with saturated galacturonate (Sat. Gal). The growth was measured as optical density (OD600 nm). Media with and without 0. 2% glucose was used as positive (Glu) and negative controls (NC), respectively. All data shown are mean values from three replicate experiments. Error bars denote the standard deviation of triplicate samples. The unsaturated uronyl residue at the non-reducing end of the oligosaccharides produced by the ulvan lyases may be released by unsaturated glucuronyl hydrolases (GH105). Five putative unsaturated glucuronyl hydrolases, Ud15_GH105, Ud16_GH105, Ud17_GH105, Ud18_GH105, and Ud19_GH105 were found in the QB4 genome and showed similarity on unsaturated beta-glucuronyl hydrolase of Nonlabens ulvanivorans, which is known as a ulvan-degrading bacterium, based on BLASTp search ( Table 3 ; Table S3 ) ( Collén et al. , 2014 ). After forming DKI by the enzymes, the following process may be the same as pectin utilization. In the ulvan-degrading process of F. agariphila, Rha3S-Xyl-Rha3S was also the main product by ulvan lyases ( Reisky et al. , 2019 ). First, Rha3S at the non-reducing end is desulfated by sulfatases. The amino acid sequences of two sulfatases, Ulu1 and Ulu2, in the QB4 genome showed high similarity (77. 3 and 86. 2%) to the sulfatases, WP. 032096151. 1 and WP. 632096147. 1, located into ulvan utilization loci of Alteromonas sp. LOR ( Foran et al. , 2017 ), indicating that the two sulfatases may involve desulfation of the Rha3S. The Rha3 will remove the Rha3-Xyl-Rha3S by α-L-rhamnosidase. Five α-L-rhamnosidases classified into GH78 family, Ud20_GH78, Ud21_GH78, Ud22_GH78, Ud23_GH78, and Ud24_GH78, were found in the QB4 genome and showed similarity on rhamnosidase from Bacillus sp. GL1, Dictyoglomus thermophilum, Streptomyces avermitilis, and Bacteroides thetaiotaomicron ( Table S3 ). Ud21_GH78, Ud22_GH78, Ud23_GH78, and Ud24_GH78 contained lipobox, suggesting that these enzymes are located on the cell surface. In addition, Ud20_GH78 and Ud22_GH78 possessed the CBM67 domain, which binds L-rhamnose in a calcium-dependent manner (Fujimoto et al. , 2013), suggesting that the two α-L-rhamnosidases are the main components to generate Rha3. Rha3 metabolic pathway was described by Reisky et al. (2019). α-L-rhamnose is converted to β-L-rhamnose by L-rhamnose mutarotase. Next, isomerization of the β-L-rhamnose to L-rhamnulose is catalyzed by rhamnose isomerase. The product is converted to L-rhamnulose-1-phosphate by pentulose/hexulose kinase (rhamnulokinase), and subsequently, the L-rhamnulose-1-phosphate is cleaved into L-lactaldehyde and dihydroxyacetone phosphate by rhamnulose aldolase. Finally, the L-lactaldehyde is converted to pyruvate by aldehyde dehydrogenase and lactate dehydrogenase. Table 3 displayed that the corresponding genes (Ulu3, Ulu4, Ulu5, Ulu6, Ulu7, and Ulu8) were found in the QB4 genome, and these proteins showed similarity on corresponding proteins in the PDB database ( Table S3 ). This result indicated that Rha3 is metabolized by QB4 cells. The genes involving the metabolic pathway downstream of the mutarotation in F. agariphila form a gene cluster ( Reisky et al. , 2019 ). However, the genes in QB4 were randomly distributed in the genome. Xyl will be released by β-xylosidase classified into GH3 and GH43 families. Based on the CAZy database, seven GH43 (Ud25_GH43, Ud26_GH43, Ud27_GH43, Ud28_GH43, Ud29_GH43, Ud30_GH43, and Ud31_GH43) proteins were found in the QB1 genome while F. agariphila possessed two each of GH3 and GH43 proteins ( Reisky et al. , 2019 ). Ud25_GH43, Ud29_GH43, and Ud31_GH43 were similar to glycoside hydrolases from Zobellia galactanivorans (50. 00%) and Halothermothrix orenii H 168 (45. 75 and 39. 10%) ( Table S3 ). The D-xylose is converted to D-xylulose-5-phosphate via xylose isomerase and xylulose kinase, and then the product is passed to the pentose phosphate pathway. Two genes encoding xylose isomerase (Ulu9) and xylulose kinase (Ulu10) were found in the QB4 genome based on RAST server annotation and formed a gene cluster with transcription repressor, xylR (Ulu11) ( Table 3 ). These results suggested that QB4 cells may use not only glucuronic acid but also L-rhamnose and D-xylose as carbon sources. In addition, the presence of numerous genes encoding ulvan lyase may promote strong ulvan degradation and its utilization. The QB4 cells indeed exhibited a robust cell growth in the presence of ulvan as well as that in the presence of alginate, as shown in Fig. 1. Genes for gellan gum degradation and utilization Gellan gum is depolymerized gellan lyases classified into PL33 family proteins. Gellan lyases are found in several bacterial species, such as Bacillus sp. GL1, Geobacillus stearothermophilus 98, and Opitutaceae bacterium TAV5 ( Hashimoto et al. , 1997 ; Derekova et al. , 2006 ; Helbert et al. , 2019 ). Gellan gum is degraded to tetrasaccharides composed of [→3)- β-D-glucose-(1 →4)- β-D-glucuronic acid-(1 →4)- β-D-glucose-(1 →4)- α-L-rhamnose-(1 →] by gellan lyases. The tetrasaccharides are completely degraded to monosaccharides by β-D-glucosidase (GH1, GH2, GH3, GH5, GH9, GH16, GH30, GH39, and GH116), unsaturated glucuronyl hydrolase (GH88 and GH105), and α-L-rhamnosidase (GH78) ( Hashimoto et al. , 2003 ). 10. 7717/peerj. 10929/table-3 Table 3 Genes involving in ulvan metabolism. Ulvan-degrading enzymes Abbreviation Function CAZy Sp. GenBank Ud1_PL24 Ulvan lyase PL24 WP_108601530. 1 Ud2_PL24 Ulvan lyase PL24 Type I WP_108601531. 1 Ud3_PL24 Ulvan lyase PL24 Type I WP_108604943. 1 Ud4_PL24 Ulvan lyase PL24 Type I WP_108601549. 1 Ud5_PL24 Ulvan lyase PL24 Type I WP_108601554. 1 Ud6_PL24 Ulvan lyase PL24 Lipobox WP_108602230. 1 Ud7_PL24 Ulvan lyase PL24, CBM32 WP_108604992. 1 Ud8_PL24 Ulvan lyase PL24 Type I WP_108602276. 1 Ud9_PL24 Ulvan lyase PL24 Lipobox WP_108602290. 1 Ud10_PL25 Ulvan lyase PL25 Lipobox WP_108601547. 1 Ud11_PL25 Ulvan lyase PL25 Lipobox WP_108601636. 1 Ud12_PL25 Ulvan lyase PL25 Lipobox WP_108601685. 1 Ud13_PL25 Ulvan lyase PL25 Lipobox WP_108602265. 1 Ud14_PL40 Ulvan lyase PL40 WP_108602237. 1 Ud15_GH105 Unsaturated glucuronyl hydrolase GH105 Lipobox WP_108601620. 1 Ud16_GH105 Unsaturated glucuronyl hydrolase GH105 Lipobox WP_108601696. 1 Ud17_GH105 Unsaturated glucuronyl hydrolase GH105 Lipobox WP_108601928. 1 Ud18_GH105 Unsaturated glucuronyl hydrolase GH105 Lipobox WP_108602225. 1 Ud19_GH105 Unsaturated glucuronyl hydrolase GH105 Lipobox WP_108602277. 1 Ud20_GH78 α-L-rhamnosidase GH78, CBM67 Type I WP_108601540. 1 Ud21_GH78 α-L-rhamnosidase GH78 Lipobox WP_108604942. 1 Ud22_GH78 α-L-rhamnosidase GH78, CBM67 Lipobox WP_108601635. 1 Ud23_GH78 α-L-rhamnosidase GH78 Lipobox WP_108602268. 1 Ud24_GH78 α-L-rhamnosidase GH78 Lipobox WP_108602275. 1 Ud25_GH43 Putative β-xylosidase GH43 Type I WP_108601365. 1 Ud26_GH43 Putative β-xylosidase GH43 Lipobox WP_108601561. 1 Ud27_GH43 β-xylosidase GH43 Type I WP_108601631. 1 Ud28_GH43 β-xylosidase GH43 Lipobox WP_108602133. 1 Ud29_GH43 β-xylosidase GH43 Lipobox WP_159084088. 1 Ud30_GH43 β-xylosidase GH43 Lipobox WP_108602377. 1 Ud31_GH43 β-xylosidase GH43 Lipobox WP_108602397. 1 Ulvan-utilization proteins Abbreviation Function GenBank Ulu1 Sulfatase WP_108601682. 1 Ulu2 Sulfatase WP_108601697. 1 Ulu3 L-rhamnose mutarotase WP_108601622. 1 Ulu4 Rhamnose isomerase WP_108601699. 1 Ulu5 Pentulose/hexulose kinase (rhamnulokinase) WP_108602148. 1 Ulu6 Rhamnulose aldolase WP_108604932. 1 Ulu7 Aldehyde dehydrogenase WP_108601159. 1 Ulu8 Lactate dehydrogenase WP_108602249. 1 Ulu9 Xylose isomerase WP_108604489. 1 Ulu10 Xylulose kinase WP_108604490. 1 Ulu11 Transcriptional regulatory protein XylR WP_10860449. 1 Figure 1 demonstrated that QB4 might also use gellan gum as a carbon source for its growth. However, gellan lyases were not found in the genome. As shown in Table 3, QB4 possessed five unsaturated glucuronyl hydrolases and five α-L-rhamnosidases ( Table 3 ). The hydrolysates, DEH, and α-L-rhamnose, may be metabolized by the pathway described above. In addition, one β-D-glucosidase classified into the GH1 family (Gd1-GH1) was also found in the genome, and Gd1-GH1 showed high similarity (63. 23%) on β-glucosidase A of Hungateiclostridium thermocellum in the PDB database ( Table 4 ; Table S4 ). β-D-glucose produced by the β-D-glucosidase is converted to D-glucose-6-phosphate by glucokinase, and then the product is converted to β-D-fructose-6-phosphate, which is an intermediate of glycolysis, by glucose-6-phosphate isomerase. One of these genes, Gu1 and Gu2, were found in the QB4 genome ( Table 4 ) and were similar on glucokinase and glucose-6-phosphate Isomerase of E. coli in PDB database, respectively ( Table S4 ), indicating that β-D-glucose is converted to β-D-fructose-6-phosphate and is metabolized by glycolysis. Discussion Polysaccharides from plants, seaweed, and bacteria, especially their oligosaccharides, have attracted considerable attention due to various biological activities. Thus, polysaccharide lyases and glycosyl hydrolases were isolated from microorganisms, including fungi and bacteria. The genus Catenovulum have been known as bacteria that can degrade agar and polysaccharide containing uronic acids, such as alginate, pectin, and ulvan ( Cui et al. , 2014 ; Li et al. , 2015 ; Lee, Lee & Hong, 2019 ; Liu et al. , 2019a ; Xie et al. , 2013 ). Our previous report demonstrated that many genes encoding polysaccharide-degrading enzymes were present in the QB4 genome ( Lau et al. , 2019 ). In this study, expectedly, QB4 cells were able to degrade and used the polysaccharides as a carbon source for their growth ( Fig. 1 ). Especially, the QB4 cells exhibited a robust growth in the presence of alginate and ulvan, suggesting that polysaccharides from marine algae are effectively used by QB4 cells rather than that of plants and bacteria. The growth profile of QB4 was different from S. degradans of which the cells reached the early stationary phase at 9 h with alginate, glucose, and pectin ( Takagi et al. , 2016a ). 10. 7717/peerj. 10929/table-4 Table 4 Genes involving in ulvan metabolism. Gellan-degrading enzymes Abbreviation Function CAZy Sp. GenBank Gd1_GH1 β-D-glucosidase GH1 WP_159084202. 1 Gellan-utilization proteins Abbreviation Function GenBank Gu1 Glucokinase WP_108603884. 1 Gu2 Glucose-6-phosphate isomerase WP_108604264. 1 Three pathways of pectin metabolisms were known in bacteria. One is the polygalacturonate pathway metabolized unsaturated disaccharides. The pathway was found in bacteria, such as Escherichia coli and phytopathogenic enterobacterium, Erwinia chrysanthemi ( Chatterjee, Thurn & Tyrell, 1985 ; Richard & Hilditch, 2009 ). The other two pathways metabolize saturated galacturonate through uronate isomerase (isomerase pathway) ( Richard & Hilditch, 2009 ) or tagaturonate epimerase (epimerase pathway) ( Rodionova et al. , 2012 ). Table 2 displayed that the QB4 possessed genes involved in the three pathways. However, QB4 cells did not grow in the presence of saturated galacturonate ( Fig. 2 ). Although a pectin-degrading marine bacterium, Pseudoalteromonas sp. PS47, also has genes for the epimerase pathway, the strain was unable to grow on saturated galacturonate ( Hobbs et al. , 2019 ). Hobbs and colleagues proposed that GH28s of the strain are periplasmic proteins and, thus, that saturated galacturonate would be produced in the periplasm ( Hobbs et al. , 2019 ). It was suggested that the strain might be unable to transport extracellular saturated galacturonate. On the other hand, although Pd11_GH28 containing Tat signal peptide would be exported into periplasm through Tat (for twin-arginine translocation) system ( Stanley, Palmer & Berks, 2000 ), Pd12_GH28 would be secreted to the outside of the cells via Sec system, suggesting that saturated galacturonate might be generated in the culture broth in contrast to the strain PS47. However, the QB4 did not grow using saturated galacturonate. Thus, it seems like the BQ4 cells are also unable to transport extracellular saturated galacturonate as well as the strain PS47. In other cases, some investigators reported that the isomerase pathway is not crucial for utilizing pectin in Dickeya dadantii (formerly E. chrysanthemi ) ( Hugouvieux-Cotte-Pattat et al. , 1996 ; Pédron et al. , 2018 ). It was known that D. dadantii belonging to the class Gammaproteobacteria, which is a plant pathogen and pectinolytic bacterium, also possessed the polygalacturonate pathway and the isomerase pathway ( Hugouvieux-Cotte-Pattat et al. , 1996 ). Even though the enzyme production of the isomerase pathway was not impaired by a mutation in genes of the polygalacturonate pathway ( kduD and kduI ), these mutants did not exhibit their growth on polygalacturonate ( Hugouvieux-Cotte-Pattat et al. , 1996 ). In addition, the transcriptomic analysis of D. dadantii during the early colonization on the plant leaf demonstrated that genes involved in the polygalacturonate pathway were upregulated in the condition ( Pédron et al. , 2018 ). These results suggested that the isomerase pathway was not the main pathway for utilizing pectin in the bacterium. The third pathway, epimerase pathway, involving the conversion of D-tagatose to D-fructuronate by tagatose the epimerase (UxaE), was found in the hyperthermophilic bacterium, Thermotoga maritima, belonging to the phylum Thermotoga. However, the pathway was not found in other bacteria, such as Escherichia and Bacillus (Kuivanen et al. , 2019). uxaE and other epimerase pathway genes, such as hexuronate catabolism regulator ( uxaR ), fructuronate reductase ( uxaD ), predicted D-mannonate utilization enzyme ( gntE) constituted a regulon regulated by the GntR-like transcription factor UxaR in the genus Thermotoga ( Rodionova et al. , 2012 ). A regulon is a gene cluster or operon that is regulated by the same regulatory protein. However, Table 2 demonstrated that genes encoding Pu14, Pu15, and Pu16 were scattered in the chromosome and plasmid, suggesting that Pu14, Pu15, and Pu16 might not constitute a regulon. In addition, Fig. 2 demonstrated that the saturated galacturonate, which is a substrate of the epimerase pathway, was not used by QB4 cells. These results suggested that the epimerase pathway might not function in QB4. Figure 1 demonstrated that the QB4 cells exhibited a robust growth using ulvan as a carbon source. Qiao et al. (2020) fermented Catenovulum sp. LP using the shaking-flask method containing 1. 2% purified ulvan, of which the concentration was 6 times-higher than that of the QB4. The bacterial culture reached the stationary phase at 36 h incubation period ( Qiao et al. , 2020 ) that was relatively slower than that of the QB4. Table S5 showed that the QB4 genome contained 14 ulvan lyase genes, which is the largest number compared to Siansivirga zeaxanthinifaciens CC-SAMT-1 (8 genes) Tamlana sp. UJ94 (7 genes), Polaribacter sp. BM10 (7 genes), and Wenyingzhuangia fucanilytica CZ1127 (7 genes) in the CAZy database ( Table S5 ). The numerous genes in the QB4 genome may be crucial to efficiently degrade ulvan and to stimulate robust growth. Figure 1 demonstrated that the QB4 cells were capable of using gellan gum as a carbon source, however, no gallan lyases were found in the QB4 genome. It was known that Bacillus sp. GL1 and G. stearothermophilus 98, which possessed all genes for gellan gum degradation and utilization, were able to use gellan gum as a sole carbon source ( Hashimoto et al. , 1998 ; Derekova et al. , 2006 ). On the other hand, although Paludisphaera borealis PX4 belonging to the order Planctomycetales does not have any gellan lyases, the bacterium was capable of degrading gellan gum ( Ivanova et al. , 2017 ). The authors suggested that an unsaturated glucuronyl hydrolase and two α-L-rhamnosidases of the bacterium involve degrading gellan gum. Tables 3 and 4 showed that QB4 cells possess five unsaturated glucuronyl hydrolases, five α-L-rhamnosidases, and one β-D-glucosidase. We expected that these hydrolases play a critical role in degrading gellan gum in QB4 cells. The lack of gellan lyases may cause lower degradation efficiency. Thus, the cell growth with gellan gum was not robust compared to those with other uronic acids. Conclusion The agar-degrading bacterium, Catenovulum sp. CCB-QB4, used uronic acids, including alginate, pectin ulvan, and gellan gum, as carbon sources for its growth. Especially, QB4 cells exhibited robust cell growth in the presence of alginate and ulvan from seaweed. In gene analysis based on the CAZy database, a large number of polysaccharide lyases and hydrolases involved in degrading these uronic acids were found in the QB4 genome. Many alginate lyases, ulvan lyases, unsaturated glucuronyl hydrolases, α-L-rhamnosidase, and β-xylosidase contained lipobox, indicating that QB4 cells can effectively degrade alginate and ulvan and uptake the oligosaccharides. Of course, genes for metabolizing the uronic acids were also present in the QB4. These results were suggested that QB4 will become a source of novel uronic acid degradation enzymes. Supplemental Information 10. 7717/peerj. 10929/supp-1 Supplemental Information 1 Sequence alignment of homology regions of DFH reductase Red boxes indicate a TGXXXGX motif. Asterisks indicate catalytic triads of the enzyme. Sa. d. , Saccharophagus degradans ; Sp. Sphingomonas sp. The figure was drawn using the program ESPript (Robert and Gouet, 2014). Reference: Robert, X. and Gouet, P. (2014) ”Deciphering key features in protein structures with the new ENDscript server”. Nucleic. Acids Research 42 (W1), W320-W324 - doi: 10. 1093/nar/gku316. Click here for additional data file. 10. 7717/peerj. 10929/supp-2 Supplemental Information 2 The similarities of amino acid (aa) sequence of alginate metabolic enzymes of QB4 in PDB database Click here for additional data file. 10. 7717/peerj. 10929/supp-3 Supplemental Information 3 The similarities of amino acid (aa) sequence of pectin metabolic enzymes of QB4 in PDB database Click here for additional data file. 10. 7717/peerj. 10929/supp-4 Supplemental Information 4 The similarities of amino acid (aa) sequence of ulvan metabolic enzymes of QB4 in PDB database Click here for additional data file. 10. 7717/peerj. 10929/supp-5 Supplemental Information 5 The similarities of amino acid (aa) sequence of ulvan metabolic enzymes of QB4 in PDB database Click here for additional data file. 10. 7717/peerj. 10929/supp-6 Supplemental Information 6 List of bacterial species possessing 5 or more ulvan lyases in CAZy database Click here for additional data file. 10. 7717/peerj. 10929/supp-7 Supplemental Information 7 Fig. 1 raw data Click here for additional data file. 10. 7717/peerj. 10929/supp-8 Supplemental Information 8 Fig. 2 raw data Click here for additional data file.

10. 7717/peerj. 11022

PeerJ

Preparation and characterization of gelatin-polysaccharide composite hydrogels for tissue engineering

Background Tissue engineering, which involves the selection of scaffold materials, presents a new therapeutic strategy for damaged tissues or organs. Scaffold design based on blends of proteins and polysaccharides, as mimicry of the native extracellular matrix, has recently become a valuable strategy for tissue engineering. Objective This study aimed to construct composite hydrogels based on natural polymers for tissue engineering. Methods Composite hydrogels based on blends of gelatin with a polysaccharide component (chitosan or alginate) were produced and subsequently enzyme crosslinked. The other three hydrogels, chitosan hydrogel, sodium alginate hydrogel, and microbial transglutaminase-crosslinked gelatin (mTG/GA) hydrogel were also prepared. All hydrogels were evaluated for in vitro degradation property, swelling capacity, and mechanical property. Rat adipose-derived stromal stem cells (ADSCs) were isolated and seeded on (or embedded into) the above-mentioned hydrogels. The morphological features of ADSCs were observed and recorded. The effects of the hydrogels on ADSC survival and adhesion were investigated by immunofluorescence staining. Cell proliferation was tested by thiazolyl blue tetrazolium bromide (MTT) assay. Results Cell viability assay results showed that the five hydrogels are not cytotoxic. The mTG/GA and its composite hydrogels showed higher compressive moduli than the single-component chitosan and alginate hydrogels. MTT assay results showed that ADSCs proliferated better on the composite hydrogels than on the chitosan and alginate hydrogels. Light microscope observation and cell cytoskeleton staining showed that hydrogel strength had obvious effects on cell growth and adhesion. The ADSCs seeded on chitosan and alginate hydrogels plunged into the hydrogels and could not stretch out due to the low strength of the hydrogel, whereas cells seeded on composite hydrogels with higher elastic modulus, could spread out, and grew in size. Conclusion The gelatin-polysaccharide composite hydrogels could serve as attractive biomaterials for tissue engineering due to their easy preparation and favorable biophysical properties.

Introduction Biomaterials for tissue engineering have been widely studied, and play a pivotal role in providing platforms that facilitate cell adhesion, growth, and proliferation. However, one of the major challenges in designing functional scaffolds is to modify their properties to mimic the extracellular matrix (ECM) in the native tissues. The composition of ECM includes structural proteins, adhesion proteins, anti-adhesion proteins, and proteoglycans. The natural biomaterials used for the engineering of tissue constructs show many obvious benefits for mimicking ECM, including collagen, gelatin (GA), hyaluronic acid, laminin, chitosan, and alginate ( Schwach & Passier, 2019 ). Polysaccharides can increase the stability of scaffolds, whereas proteins can enhance the biological properties. Therefore, mixing protein components (such as GA) with polysaccharide components (such as chitosan and alginate) to form composite materials mimicking natural ECM has become an important strategy for tissue engineering applications ( Afewerki et al. , 2019 ). GA as a degradation product of collagen has good biocompatibility, high hydration degree, and low market cost. Therefore, GA has become a well-known biological material. However, natural GA hydrogel has low mechanical stability, thereby severely limiting its application in tissue engineering. Several methods have been used to overcome its defects, including physical mixing, chemical crosslinking, and enzymatic crosslinking. Chemical crosslinking agents, such as glutaraldehyde and methacrylic anhydride, were used to crosslink the composite materials of GA and polysaccharides to obtain better mechanical properties ( Majidi et al. , 2018 ; Miranda et al. , 2011 ; Rosellini et al. , 2009 ). However, incomplete removal of chemical crosslinking agents may lead to cytotoxicity ( Li, Liu & Liu, 2009 ). Therefore, in our previous study, microbial transglutaminase (mTG) was used in place of chemical crosslinking agents to crosslink the composite materials, thereby resolving the problem of cytotoxic side effects ( Yang et al. , 2016 ). Chitosan is a polysaccharide composed of glucosamine and N-acetyl-glucosamine. It is obtained by removing some acetyl groups from chitin, and it also an analog of glycosaminoglycan, which is a component of ECM. Therefore, chitosan exhibits many interesting biological properties, including biocompatibility, hydrophilicity, antibacterial, and antithrombotic effects. Chenite et al. (2000) first reported that nearly neutral chitosan/ β -GP aqueous solutions could gel quickly when heated. Thermosensitive chitosan hydrogel shows great potential in tissue engineering ( Zhou et al. , 2015 ). However, some authors reported that the thermosensitive chitosan hydrogel has insufficient mechanical strength when used as a tissue engineering material ( Sacco et al. , 2018 ; Supper et al. , 2014 ; Zhang et al. , 2019 ). Huang et al. (2014) constructed composite scaffolds comprising a chitosan hydrogel system and demineralized bone matrix, which exhibited an increased mechanical strength; the bone marrow stem cell (BMSC) retention of the hybrid scaffolds was more efficient and uniform than that of the other materials. Song et al. (2010) reported that ADSCs within the chitosan/ β -GP/collagen hydrogels displayed a typical adherent cell morphology and good proliferation with very high cellular viability after 7 days of culture. These experimental results indicated the possibility of overcoming the defects of chitosan gel by mixing with other materials, and composite scaffolds based on chitosan may be promising candidates for tissue engineering. Alginate is a natural polysaccharide isolated from brown algae and bacteria. It has good biocompatibility and non-antigenicity; it is also antithrombotic and biodegradable. Besides, alginate has the advantages of having abundant sources and being low cost. Among the strategies used for the obtention of alginate hydrogel, the most widespread is ionic crosslinking ( Sun & Tan, 2013 ). In the presence of multivalent cations, crosslinking is instantaneous and almost temperature-independent and allows solution/gel transformation under relatively mild conditions ( Cattelan et al. , 2020 ). Alginate hydrogel has been widely used in tissue engineering and has been approved for phase II clinical trials in the treatment of MI ( Lee & Mooney, 2012 ). However, alginate hydrogel has natural poor cell adhesion and poor in vivo degradation performance ( Bedian et al. , 2017 ; Tønnesen & Karlsen, 2002 ). To solve these problems, mixing alginate with natural proteins, such as collagen, fibronectin, and GA, has been proposed ( Hernández-González, Téllez-Jurado & Rodríguez-Lorenzo, 2020 ). Complexes of alginate and GA are interesting to use in tissue engineering, because they can provide suitable biological cues for hosting a variety of cells, including C2C12 myoblasts ( Rosellini et al. , 2018 ), HL1 cardiac muscle cell ( Xu et al. , 2009 ), neonatal rat ventricular myocytes (NRVMs) ( Möller et al. , 2011 ), and fetal rat myoblast H9C2 ( Saberianpour et al. , 2019 ). Composite methods of GA and polysaccharides have been proposed, and the materials obtained through such methods support cell adhesion, proliferation, and differentiation ( Miranda et al. , 2011 ; Rosellini et al. , 2009 ; Rosellini et al. , 2018 ). Generally, strategies for preparing composite hydrogels include physical mixing ( Liu et al. , 2013 ), crosslinking ( Rosellini et al. , 2019 ), in situ synthesis ( Wang et al. , 2009 ), bio-conjugation ( Ahadian et al. , 2015 ), and others. In our composite hydrogels, enzymatic modification of proteins mediated by mTG is applied and aims to improve the properties of target products. These enzymatic reactions occur under mild reaction conditions and do not produce toxic products ( Fatima & Khare, 2018 ). Da Silva et al. (2014) proved the feasibility of using mTG as a crosslinking agent for chitosan and GA hydrogel. Moreover, the microstructure of alginate GA-hydrogel microspheres was confirmed to be affected by GA content and mTG concentration ( Pilipenko et al. , 2019 ). Adipose-derived stromal stem cells (ADSCs) are adult stem cells with abundant cell sources and provide a potential source of stem cells for tissue engineering research and clinical application ( Suzuki et al. , 2015 ). In this study, the preparation of chitosan/mTG-crosslinked GA (C-mTG/GA) and alginate/mTG-crosslinked GA (A-mTG/GA) was proposed, which are two kinds of composite hydrogels. Swelling, enzymatic degradation, and mechanical tests were performed. Viability and proliferation tests of ADSCs seeded on the composite hydrogels were conducted to determine the applicability of the hydrogels in tissue engineering. Finally, the cytoskeleton features of the ADSCs’ distribution on hydrogels were evaluated ( Fig. 1B ). 10. 7717/peerj. 11022/fig-1 Figure 1 Schematic elucidating the preparation and characterization of the five hydrogels. (A) Schematic showing the five hydrogels fabrication process. (1) Preparation of chitosan hydrogel. (2) Preparation of alginate hydrogel. (3) Preparation of mTG/GA hydrogel. (4) Preparation of two composite hydrogels. (B) Characterization and biological assessments of the five hydrogels. (C–G) The appearance of the five tested hydrogels. Scale bar = 1 cm. Materials & Methods Materials Microbe transglutaminase (mTG, Bomei, China; enzyme activity, >100 U per gram), high-glucose Dulbecco’s modified Eagle’s medium (DMEM, Hyclone, UT, USA), fetal bovine serum (FBS, Gibco, NY, USA), and penicillin/streptomycin (P/S; Hyclone, UT, USA) were used. Trypsin (250 NFU/mg), ethylene diamine tetraacetic acid (EDTA), and collagenase type I (>125 NFU/mg) were purchased from Invitrogen (CA, USA). GA (type A, 300 Bloom), beta-sodium glycerophosphate ( β -GP), calcein-AM, propidium iodide (PI), thiazolyl blue tetrazolium bromide (MTT), dimethyl sulfoxide (DMSO), aqueous formaldehyde, Triton X-100, 4, 6-diamidino-2-phenylindole (DAPI), and fluorescein isothiocyanate (FITC)-phalloidin were purchased from Sigma (MO, USA). Chitosan (molecular weight: 100–300 kDa, degree of deacetylation: ≥85%), sodium alginate (Mw = 220, 000 and M/G ratio = 0. 38), calcium chloride (CaCl 2 ), sodium chloride (NaCl), and acetic acid were purchased from Kelong (Chengdu, China). Analytical- or chemical-grade reagents were used. Preparation of hydrogels Preparation of chitosan hydrogel Preparation of chitosan hydrogel was performed according to the method described elsewhere ( Chenite et al. , 2000 ) with minor modifications. Chitosan was placed on a piece of weighing paper, spread out, and exposed to ultraviolet (UV) light for 45 min, during which it was slightly turned over 3 times. The UV-disinfected chitosan particles were transferred to a sterile clean bench and dissolved in 0. 1 M acetic acid. After stirring with a glass rod, it was placed at 4 °C for 24 h and centrifuged at 8, 000 rpm for 10 min. The supernatant was collected to obtain the final chitosan solution concentration of 2% (w/v, weight volume ratio). This solution was stored at 4 °C. β -GP was dissolved in deionized water to obtain 50% (wt, weight ratio) solution, sterilized and passed through a 0. 22 µm filter, and stored at 4 °C. During the preparation of the chitosan hydrogel, the β -GP solution was slowly dripped into the chitosan solution on the ice bag. The ratio of chitosan solution to β -GP solution was 5:1 ( Xia et al. , 2010 ), and the pH value was adjusted to 7. 4. The mixed solution at 200 µL was added into a 24-well tissue culture plate (TCP) and placed into each well. The solution was incubated at 37 °C for 10 min to obtain chitosan hydrogel ( Fig. 1A ). Preparation of alginate hydrogel After disinfection by UV (as mentioned above), sodium alginate was dissolved in sterilized phosphate buffered saline (PBS) solution to form 1% (w/v) solution. CaCl 2 and NaCl were mixed with deionized water to form solutions at final concentrations of 100 and 150 mM, respectively. After high-pressure steam sterilization, CaCl 2 /NaCl solution was stored at 4 °C. To prepare the alginate hydrogel, the sodium alginate solution was placed into a 24-well TCP (200 µL placed in each well). Then, 1 mL of CaCl 2 /NaCl solution was slowly dropped into each well. After soaking for 10 min at room temperature, the excess liquid was removed, and sodium alginate hydrogel was obtained ( Fig. 1A ). Preparation of mTG/GA hydrogel mTG/GA hydrogel was prepared using a protocol described in our previous publication ( Long et al. , 2017 ; Yang et al. , 2016 ). GA was dissolved in PBS at 50 °C to obtain a solution, which was sterilized and passed through a 0. 22 µm filter. mTG was dissolved in PBS to prepare 10% (wt) solution, which was sterilized and passed through a 0. 22 µm filter. mTG/GA solution was prepared by adding mTG into the GA solution at 10 U/g⋅pro (enzymatic activity unit per gram of protein). Then, the mixed solution with a final GA concentration of 6% (w/v) was added into a 24-well TCP; 200 µL of the solution was placed into each well. The sample was incubated at 37 °C for 30 min for gelling ( Fig. 1A ). Preparation of C-mTG/GA and A-mTG/GA hydrogels mTG/GA, chitosan, and sodium alginate solutions were prepared as mentioned above. The 7. 5% (w/v) GA with 10 U/g⋅pro mTG and 2% (w/v) chitosan (or 1% (w/v) sodium alginate) solutions were mixed in a volume ratio of 4:1. The composite process was conducted on a hot platform (DB-H, Xinbao, China) at 37 °C. The mixtures with a final GA concentration of 6% (w/v) were added into 24-well TCPs; 200 µL of the mixtures was placed into each well. The solutions were incubated at 37 °C for 20min to obtain C-mTG/GA and A-mTG/GA hydrogels ( Fig. 1A ). Characterization of hydrogels Gelation time Ungelled mTG/GA mixture, β -GP/chitosan, C-mTG/GA, and A-mTG/GA solutions (1 mL) were added into 2 mL microtubes and incubated at 37 °C for gel. Sodium alginate solution at 1 mL was added into a 2 mL microtube, and CaCl 2 /NaCl solution was dropped for gelling. The onset of gelling was recorded as the gelation time, which was detected through the vial inverting method. Hydrogel degradation test In vitro enzymatic degradation property. The in vitro enzymatic degradation property of hydrogels was evaluated by exposing them to enzymes to assess degradation rate. Collagenase has been used previously as a mimic for some of the protease secreted by cells ( Mazzeo et al. , 2019 ), and trypsin is often used in cell isolation and culture. The material degradation process of these proteases must be evaluated to provide a basis for cellular inoculation and digestion on hydrogels. The pre-weighed hydrogels ( w 0 ) were then immersed in 0. 1% collagenase type I and 0. 25% trypsin/0. 01% EDTA for 12 h. At each time point (0. 25, 0. 5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 h), the liquid was removed completely. The hydrogels were weighed again ( w t ). The degree of degradation (D) was calculated as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\mathrm{D}(\text{%})= \left( {w}_{0}-{w}_{t} \right) /{w}_{0}\times 100. \end{eqnarray*}\end{document} D % = w 0 − w t ∕ w 0 × 100. Three repeated measurements were performed for each type of hydrogel. Hydrolytic and cellular degradation. The hydrolytic and cellular degradation of hydrogels were performed as described by our previous study ( Yang et al. , 2016 ). In brief, hydrogels were prepared into 35 mm culture dishes with ∼1 mL for each dish. To exclude the influence of swelling behavior of hydrogels, 2 mL of PBS was added into each dish and PBS was removed completely after 12 h; the hydrogels were weighed ( w 0 ). The hydrogels were incubated in PBS at 37 °C with 5% CO 2 for two weeks. PBS was changed every 2 days. At different time points (0, 2, 4, 6, 8, 10, 12, and 14 d), three samples of each kind of hydrogel were weighed ( w t ). The degree of degradation was calculated as above. Cell-mediated degradation was measured by seeding 1. 0 × 10 6 ADSCs (see ‘Primary culture of ADSCs’) on each hydrogel scaffold. The rest of the steps were the same as above, except that the PBS was changed to cell culture medium (high-glucose DMEM, 15% FBS, and 1% P/S). Swelling capacity Each type of hydrogel was prepared and weighed ( w 1 ). The hydrogels were immersed into PBS for 12 h at 37 °C, and excess PBS was blotted out with filter paper. The swollen hydrogels were obtained and re-weighed ( w 2 ). Swelling ratio (S) was calculated as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\mathrm{S} \left( \text{%} \right) = \left( {w}_{2}-{w}_{1} \right) /{w}_{1}\times 100. \end{eqnarray*}\end{document} S % = w 2 − w 1 ∕ w 1 × 100. Three repeated measurements were performed for each type of hydrogel. Mechanical property The mechanical property of the hydrogels was evaluated using a mechanical testing apparatus (HPB, Handpi, China). The effects of sample cutting were considered. Only the samples with the same dimensions were selected as test specimens. For this purpose, cylinder-shaped samples were cut to achieve a diameter of 15 mm and a thickness of 6 mm. The sample was fixed on a hot platform and examined at 37 °C. The detecting probe was a stainless steel cylinder (12. 5 mm in diameter) with a flat front attached to the mechanical testing machine. The hydrogels were compressed at a constant deformation rate of 1. 0 mm/s. Meanwhile, the value of loading force was recorded automatically by using a mechanical testing software (Handpi, China). The slopes of compressive stress–strain curves at 0% to 50% deformation were used to calculate the compressive modulus. The reported values are the mean of six specimens. Primary culture of ADSCs The study protocol was approved by the Institutional Animal Care and Use Committee of Sichuan University (approval number: KS2019006). ADSCs were isolated as described previously ( Yang et al. , 2016 ). Subcutaneous adipose tissue was obtained from a Sprague–Dawley rat (weighed 100 g, either male or female) and finely minced. The minced tissue was placed in a digestion solution containing 0. 1% collagenase type I and subjected to continuous agitation at 37 °C for 45 min. The cell suspension was filtered and centrifuged at 2000 rpm for 5 min. Cellular precipitation was resuspended with the cell culture medium and cultured in 25 mm 2 cell flasks. The cells were cultured at 37 °C in a 5% CO 2 incubator, and the medium was changed twice a week. Cultures were passaged every 5 days. The cells were observed daily under an inverted phase-contrast microscope (CKX41, Olympus, JAPAN). The cells were detached with 0. 25% trypsin/0. 01% EDTA and re-plated for cell passage. The third-passage ADSCs were used for the subsequent experiments. Cell viability 2D culture. Different types of hydrogels were prepared as described above. After washing with PBS, the hydrogels were ready for cell culture. ADSCs were seeded on the surface of the hydrogels at 1. 0 × 10 4 cells per well. ADSCs at same number were seeded on non-hydrogel TCPs as controls. 3D culture. ADSCs were prepared as described above and cell density was adjusted to 5. 0 × 10 6 cells/ml. The cell/hydrogel mixtures were obtained by mixing cell suspensions and different hydrogel solutions at 1:9 volume ratio and the mixtures were pipetted into the wells of a 24-well culture plate at 100 µL per well (∼5. 0 × 10 4 cells per well). The cell/hydrogel constructs were incubated at 37 °C for 2 h and then supplemented with cell culture medium. On day 5 of ADSC culture, the cell-cultured samples were washed thrice with PBS and incubated in 250 µL PBS containing 2 µM calcein-AM and 2 µM PI at 37 °C for 30 min. After re-washing with PBS, the cells were observed by using an inverted fluorescent microscope (XDS30; Sunny Instruments, China) equipped with a color digital camera (MD50; Mingmei, China). The viability of ADSCs was determined by staining with calcein-AM and with PI to label the live and dead cells, respectively. Cell proliferation studies ADSCs were seeded on the hydrogels and TCPs as described above at 1 × 10 4 cells per well. Cell proliferation was determined by MTT assay on days 0, 2, 4, and 6 of cell culture. At each time interval, the cell-cultured samples (three replicates) were rinsed thrice with PBS and treated with 800 µL of high-glucose DMEM containing 80 µL of MTT solution (5 mg/mL in PBS) at 37 °C for 4 h. The supernatant was removed after incubation, and the formazan crystals in the cells were dissolved in 400 µL of DMSO. Then, the absorbance of 100 µL of supernatant transferred to a new 96-well TCP was measured at 490 nm with a reference wavelength of 630 nm by using a microplate reader (Biotek ELx800, USA). Background absorbance from the control wells, which contained the culture medium but without cells, was subtracted. Cytoskeleton staining Cells were seeded on the hydrogels and TCPs at 1 × 10 5 cells per well (6-well TCPs) and cultured for 5 days. Cells were fixed with 4% (w/v) aqueous formaldehyde solution for 15 min at room temperature, and permeabilized with 0. 1% (v/v) Triton X-100 solution in PBS for 10 min. Afterward, they were stained with 5 µg/mL FITC- for 30 min followed by 1 µg/mL DAPI for 10 min. After incubation, fluorescent images were acquired using the inverted fluorescent microscope equipped with a digital camera. Statistical analysis Statistical analyses were performed using SPSS software (SPSS Inc. ). Each experiment was repeated ≥3 times. Data are presented as mean with standard deviation. Statistical significance between two groups was determined by an independent sample Student’s t -test. The level of statistical significance was P < 0. 05. Results Gelation time and appearance of hydrogels The chitosan samples at 1 mL of 2% (w/v) gelatinized after the addition of β -GP solution within 10 min at 37 °C, and the obtained hydrogel was opaque and light yellow. The GA samples at 1 mL of 6% (w/v) needed approximately 30 min to achieve gelling at 37 °C. The mTG/GA hydrogel was colorless and transparent. This result was consistent with that obtained in our previous study ( Yang et al. , 2016 ), in which we reported that the 6% solution of GA mixed with mTG took 25 ± 0. 55 min to gel. Alginate hydrogel samples at 1 mL of 1% (w/v) were obtained by Ca 2+ -crosslinking at room temperature within 10 min. When the calcium solution was dropped into the alginate solution, the edge started to shrink immediately and form wrinkles. After removing the additional liquid, we obtained a translucent white alginate hydrogel with uneven thickness. Samples of the two composite solutions (1 mL) prepared as mentioned in ‘Preparation of C-mTG/GA and A-mTG/GA hydrogels’ gelatinized within 20 min at 37 °C. The C-mTG/GA hydrogel was translucent yellow. The A-mTG/GA hydrogel was colorless and transparent ( Figs. 1C – 1G ). Swelling capacity Swelling capacity reflects the hydrophilic character and water retention capacity of hydrogels, and is also an important factor for predicting nutrient transfer within hydrogels ( Gu et al. , 2020 ). After being immersed in PBS for 24 h, mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels were slightly expanded, whereas the volumes of chitosan and alginates hydrogels almost remained unchanged. Swelling ability is an important index for hydrogels that are used in tissue engineering. We evaluated the swelling rate of the five hydrogels ( Table 1 ). Chitosan and alginate hydrogels showed maximum swelling rates of 1. 55 ± 0. 89% and 4. 68 ± 0. 87%, respectively, in PBS. Both swelling rates were significantly lower than that of the other three hydrogels. The swelling rate of mTG/GA hydrogel was 21. 33 ± 3. 56%. Although with the same concentration of GA, the swelling rate of C-mTG/GA hydrogel decreased to 14. 29 ± 2. 28% ( P < 0. 05). The swelling rate of A-mTG/GA hydrogel (20. 72 ± 0. 84%) did not differ significantly from that of mTG/GA hydrogel (21. 33 ± 1. 78%, P = 0. 62 >0. 05). 10. 7717/peerj. 11022/table-1 Table 1 Gelation time, swelling rate and compression modulus of the five tested hydrogels. Samples Chitosan hydrogels Alginate hydrogels mTG/GA hydrogels C-mTG/GA hydrogels A-mTG/GA hydrogels Gelation time (min) <10 <10 ∼30 <20 <20 Swelling rate (%) 1. 55 ± 0. 89 *, ** 4. 68 ± 0. 87 *, ** 21. 33 ± 1. 78 ** 14. 29 ± 2. 28 *, ** 20. 72 ± 0. 84 * Compression modulus (kPa) 3. 48 ± 0. 45 *** 7. 06 ± 1. 22 *** 17. 42 ± 1. 34 *** 14. 29 ± 2. 64 *** 19. 79 ± 1. 22 *** Notes. *, **, *** P < 0. 05 Mechanical property As cardiac tissue engineering substrates, the hydrogels need to have proper mechanical properties. We used a mechanical testing machine to evaluate the compression modulus ( Table 1 ). Chitosan and alginate hydrogels exhibited low compression moduli (3. 48 ± 0. 9 and 7. 06 ± 2. 44 kPa). Compared with mTG/GA hydrogel (17. 42 ± 2. 68 kPa), the strength of the composite hydrogel showed a decreasing trend after mixing with chitosan (14. 29 ± 5. 28 kPa, P < 0. 05), whereas the strength of the composite hydrogel showed an increased trend after mixing with sodium alginate (19. 79 ± 2. 44 kPa, P < 0. 05). These results may be related to the formation of intermolecular interactions (electrostatic interaction and hydrogen bonding) between the polypeptide chains of GA and the macromolecules of polysaccharides. The additional junctions in the complex gel network result in changes in elasticity compared with those of native GA ( Derkach et al. , 2020 ). Hydrogel degradation test Cells can produce different proteolytic enzymes via autocrine and/or paracrine, which may lead to the degradation of hydrogel. Therefore, the in vitro enzymatic degradation of hydrogels should be evaluated. Degradation tests in two parallel groups were conducted, namely, collagenase and trypsin degradation groups. The curves of enzymatic degradation are shown in Fig. 2. In the collagenase degradation tests, the C-mTG/GA hydrogel showed the fastest degradation rate, in which more than 30% of the original weight was lost in 15 min, and only 0. 20 ± 0. 36% remained after 5 h. During degradation by collagenases, mTG/GA and A-mTG/GA hydrogels also showed obvious degradation. Complete degradation of mTG/GA hydrogels occurred within 6 h, and complete degradation of A-mTG/GA hydrogels required 7 h. However, the enzymolysis rates of chitosan and alginate hydrogels were markedly slower than those of the other three hydrogels. For alginate hydrogel, only 13. 76 ± 3. 57% of the hydrogel was lost after 12 h. For chitosan hydrogel, approximately 96% remaining after digestion for 12 h. 10. 7717/peerj. 11022/fig-2 Figure 2 In vitro enzymatic degradation property of the five hydrogels. (A) The curves of 0. 1% collagenase degradation. (B) The curves of 0. 25% trypsin/0. 01% EDTA degradation. (C) The curves of hydrolysis degradation. (D) The curves of cellular degradation. Based on the trypsin degradation tests, mTG/GA and C-mTG/GA hydrogels almost completely dissolved after 2 h of enzymatic degradation. A-mTG/GA hydrogels showed slower degradation rate and were degraded completely by trypsin within 8 h. Alginate hydrogels showed nearly half of the mass loss after 12 h. However, chitosan hydrogel did not show degradation after 1 h, and its final mass after 12 h remained at approximately 90%. All hydrogels exhibited high capacity for hydrolytic resistance. After two weeks of immersion, all of the five kinds of hydrogels retained more than 95% of their original mass. In the test of cell-containing hydrogels, we found that all hydrogels degraded more rapidly than cell-free hydrogels. For two weeks, 25. 1 ± 3. 48% of mTG/GA hydrogel mass was lost. Moreover, the alginate hydrogel also showed significant degradation. At the second week, 12. 2 ± 2. 99% of the hydrogel mass was lost. However, the mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels did not show severe degradation. For C-mTG/GA, and A-mTG/GA hydrogels, approximately 5% of gel mass was lost after two weeks of incubation; meanwhile for chitosan hydrogel, mass loss was 1. 2 ± 0. 58%. These results may suggest that the incorporation of polysaccharides can help strengthen cellular degradation resistant capacity of GA-containing hydrogels. Cell morphological observation To observe the growth and adhesion of ADSCs on hydrogel surface and verify the biocompatibility of the hydrogels, cell images were recorded by using the inverted phase contrast microscope. Figure 3 shows the cell morphology of 2D cultures on the five hydrogels for 2 h, 1 day, and 3 days. After inoculation for 2 h, most of the ADSCs on the TCP were already attached and extended to the plate. Cells on the mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels showed some pseudopods and were stellate or irregular in shape. Some rounded cells still not attached to the surface of chitosan and alginate hydrogels were present. After inoculation for 1 day, most cells showed spreading activity, and cell proliferation was observed on the surface of C-mTG/GA and A-mTG/GA hydrogels. The elongation of ADSCs was also observed on the surface of mTG/GA hydrogel. Nevertheless, the cells on the surface of the chitosan and alginate hydrogels remained round in shape. Some of the cells on the surface of these hydrogels were surrounded by a ring shadow, indicating that cells were moving in and out within a relatively small space. These cells were trying to stretch out their pseudo feet. However, because no cell adhesion sites were present, these cells were unable to adhere to the surrounding hydrogel tightly, and hardly any stretching was maintained. Therefore, when these shadows appear, the range of cell activity can be inferred indirectly. The images at day 3 clearly showed the cells reached confluence and covered the surface of C-mTG/GA and A-mTG/GA hydrogels. The cells on the surface of chitosan hydrogel grew inward because of the inadequate strength for cell spreading. A few of the polygonal ADSCs were recorded on the surface of the alginate hydrogel. Cell viability The viability of ADSCs on the hydrogels was determined using live/dead staining assay by imaging live and dead cells under a fluorescent microscope. Cells that lost membrane integrity and were no longer viable were stained red (dead cells), whereas the viable cells were stained green (live cells) ( Fig. 4 ). The five hydrogels show good biocompatibility and were suitable for cell 2D culture. Although the proportion of dead cells was small in 2D cultures, cells seeded on chitosan and alginate hydrogel displayed negative growth (10–20 live cells per visual field at 10 × magnification) over the 5 days and remained circular in shape. The cells on mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels grew in size and clustered together. The numbers of live ADSCs cultured on mTG/GA and A-mTG/GA hydrogels in each visual field were more than 100 cells, and the numbers of dead cells in same visual fields were less than 10 cells. The average quantity of live cells on C-mTG/GA hydrogel was slightly less (∼75 live cells per visual field) than on the other hydrogels. Cell shape progressively assumed prickly or rhombohedral patterns. The cell morphology indicated that the cell development may progress to achieve a 3D shape instead of a flat 2D shape. The ADSCs on the C-mTG/GA hydrogel appeared to stretch out to different directions in a 3D hydrogel space (the circle of Fig. 4D ). The ADSCs developed on TCPs showed higher quantity and a spreading shape. 10. 7717/peerj. 11022/fig-3 Figure 3 Cell growth at 2 h, 1 day, and 3 days were observed by using an inverted phase contrast microscope. (A–C) Chitosan hydrogels, (D–F) alginate hydrogels, (G–I) mTG/GA hydrogels, (J–L) C-mTG/GA hydrogels, (M–O) A-mTG/GA hydrogels, and (P–R) TCPs. Scale bar = 200 µm. 10. 7717/peerj. 11022/fig-4 Figure 4 Live/dead cell staining results after 5 days of culture. The cytoplast of live cells emitted green fluorescence when stained with calcein-AM. The nuclei of dead cells emitted red fluorescence when stained with PI. (A) chitosan hydrogels, (B) alginate hydrogels, (C) mTG/GA hydrogels, (D) C-mTG/GA hydrogels, (E) A-mTG/GA hydrogels, and (F) TCPs. Scale bar = 200 µm. Live/dead staining of ADSCs embedded in the five hydrogels, shown in Fig. 5, revealed that >90% of cells survived the 3D culture. At day 5 in culture, it was evident that GA-containing hydrogels promoted spreading of embedded cells and resulted in more surviving cells compared to chitosan and alginate hydrogels. According to the images recorded by the inverted fluorescent microscope, ADSCs remained a rounded morphology in chitosan and alginate hydrogels after 5 days of culture. While in mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels, the cells assumed barbed-like morphology with numerous cell protrusions stretching out in different directions, which represented the three hydrogels supporting the cell survival and adhesion. 10. 7717/peerj. 11022/fig-5 Figure 5 Observing cell viability in 3D cultures at 5 days. (A–C) Chitosan hydrogels, (D–F) alginate hydrogels, (G–I) mTG/GA hydrogels, (J–L) C-mTG/GA hydrogels, (M–O) A-mTG/GA hydrogels. Scale bar = 200 µm. MTT assay results Significant difference in cell growth behavior was observed from day 0 to day 6 ( Fig. 6 ). Although the same number of cells were seeded on day 0, the number of cells cultured on each hydrogel was less than that of the TCP control group at different time intervals, and the difference was statistically significant ( P < 0. 05). There was an adaptation period, during which the cells grew on the materials. Some cells were detached from hydrogels. As a result, the number of cells on the hydrogels was certainly less than that of the control group. However, if the duration of measurement was longer, then the number of cells on the materials may increase. In our previous experiments ( Yang et al. , 2016 ), the number of cells growing on the materials could exceed the number of cells on TCPs with increasing measurement time. The purpose of this study was to determine the effects of different materials on cell compatibility and proliferation, but not to compare the difference of cell growth behavior between hydrogels and the control group. Therefore, long-term MTT assay (>2 weeks) was not performed. Compared with ADSCs on alginate and chitosan hydrogels, those on mTG/GA and the two composite hydrogels showed more cells and a faster proliferation rate. A decrease in cell proliferation was observed from samples on the chitosan and alginate hydrogels on day 4, and such a decrease may have been due to the poor adhesion of the cells on these two materials. Some cells were lost when the culture medium was changed routinely on day 3. 10. 7717/peerj. 11022/fig-6 Figure 6 The proliferation of the ADSC cultured on the hydrogels and TCPs after 0, 2, 4, and 6 days of culture, as determined by MTT assay. Cytoskeleton staining Cytoskeleton is a network system of protein fibers in eukaryotic cells. In a narrow sense, the cytoskeleton is composed of microtubules, microfilaments, and intermediate fibers. Microfilaments are spiral fibers composed of filamentous actin (F-actin). When the adherent cells spread out and became larger on the culture surface, the expression of actin increased, which led to the formation of the actin network. FITC-phalloidin is a kind of microfilament depolymerization inhibitor that has a strong affinity with actin filaments and only binds to F-actin. Therefore, only the F-actin of the ADSCs seeded on the hydrogels with excellent properties in cell adhesion and growth can be examined with FITC-phalloidin. Figure 7 shows the spreading and morphology of ADSCs grown on hydrogels and TCPs. ADSCs cultured on chitosan hydrogel for 5 days were still round and had no obvious cytoskeletal structure. The chitosan hydrogel showed poor properties for cell adhesion and growth, and the myofilament structure did not form in the cells. Few ADSCs on alginate hydrogel were polygonal, and few actin fibers were observed on the edge of the cells. ADSCs on C-mTG/GA hydrogel showed a radial arrangement of microfilament cytoskeleton. The staining images of ADSCs on A-mTG/GA and mTG/GA hydrogels were similar to those of the cells on TCPs. Most cells spread on the surface of the material, and the cytoskeleton network was orderly arranged. 10. 7717/peerj. 11022/fig-7 Figure 7 Cytoskeleton staining. The F-actin emitted green fluorescence when stained with FITC-phalloidin. The nuclei emitted blue fluorescence when stained with DAPI. (A) chitosan hydrogels, (B) alginate hydrogels, (C) mTG/GA hydrogels, (D) C-mTG/GA hydrogels, (E) A-mTG/GA hydrogels, and (F) TCP. Scale bar = 100 µm. Discussion To better mimic the physiological, biochemical, and physical cues of native tissues, the hybrid materials have extensively been explored. At a minimum, the preferred biomaterial for tissue engineering needs to meet the following essential criteria ( Yue et al. , 2020 ; Zhang et al. , 2019 ): (1) biodegradability, (2) proper elastic modulus, and (3) good biocompatibility. In the field of regenerative medicine, the scaffold material usually needs to be biodegradable ( Wang et al. , 2019 ). However, some exceptions exist, such as bone, articular cartilage, or cornea tissue engineering, which require stability of the implanted material ( Rastogi & Kandasubramanian, 2019 ). The blending of polymers may affect the degradation behavior. Therefore, degradation test of the scaffolds was conducted in vitro. Under the action of collagenase, the three hydrogels of mTG/GA, C-mTG/GA, and A-mTG/GA showed similar enzymatic degradation rates. Upon addition of trypsin, both mTG/GA and C-mTG/GA completely degraded in a short time. These results indicated that enzymatic cross-linking can provide stability of the scaffold but does not hinder the degradation of protein components. However, for A-mTG/GA hydrogels, the speed of enzymatic degradation by trypsin is slower than that of mTG/GA and C-mTG/GA hydrogels, which may be due to the fact that sodium alginate, as a natural anionic polymer, prevents the trypsin activity from entering GA through electrostatic action ( Lv et al. , 2014 ). Ruvinov et al. also observed that alginate-sulfate hydrogel protected the protein from the hydrolysis of trypsin ( Ruvinov, Leor & Cohen, 2010 ). GA and chitosan are biodegradable, whereas alginate shows high stability in vivo ( Bedian et al. , 2017 ). In our work, chitosan and alginate hydrogels were more stable to degradation of the two proteases, because the main components of chitosan and alginate gels are polysaccharides. However, alginate hydrogel lost almost half of its mass under the action of trypsin after 12 h. This phenomenon might be related to the addition of 0. 01% EDTA to trypsin. Dimerization of alginate chains is induced by calcium, and as a result, gel networks are formed. Depending on the amount of calcium present in the system, these inter-chain associations can be either temporary or permanent ( George & Abraham, 2006 ). EDTA can chelate Ca 2+ ( Hafer et al. , 2020 ); the content of calcium in the A-mTG/GA system is reduced, resulting in a thixotropic solution with high viscosity ( George & Abraham, 2006 ), thereby finally showing the decline of solid mass. The mechanical property of hydrogel governs final tissue engineering usage, e. g. , less stiffness and softer hydrogel can be used to soft tissues of the brain, and high stiffness and harder hydrogel may be effective for hard tissues of bones, thereby prompting us to characterize the mechanical properties of the five types of hydrogels in this study. Our results showed that the mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels had higher elastic moduli than that of chitosan and sodium alginate hydrogel. The addition of polysaccharides particles influenced the elastic modulus, while the effect differs depending on the type of polysaccharides used. After adding chitosan to GA, the structural heterogeneities in the composite’s network cause descending of the hydrogel’s mechanical property, but after adding sodium alginate to GA, the raising of the elastic modulus indicate the increase of physical crosslinking density ( Lewandowska-Łańcucka et al. , 2017 ; Li et al. , 2018 ). The stiffness of substrate also aids in cell functioning, i. e. , in growing and proliferating ( Caliari & Burdick, 2016 ). ADSCs grown on GA and on the two kinds of mixed hydrogels showed better adhesion and elongation, whereas ADSCs grown on alginate hydrogel showed relatively later adhesion. ADSCs cultured on chitosan hydrogel cannot adhere and grow due to low elastic modulus. Cytoskeleton staining also confirmed this conclusion. The cell morphology of ADSCs on alginate hydrogel was irregular without an obvious cluster-like branch. ADSCs cultured on chitosan hydrogel almost did not form filamentous actin, and the cell shape remained round. Hydrogels are water-insoluble networks of crosslinked hydrophilic polymers that exhibit swelling capacity in aqueous environments. The ability of water retention by the material strongly depends on the microstructure ( Yan et al. , 2005 ). After mixing chitosan into GA, the mixed hydrogel showed a decrease in swelling rate, which may be attributed to the formation of the tight microstructure of chitosan and GA. The swelling rate of the mixed hydrogels of sodium alginate and GA did not change significantly compared with the mTG/GA hydrogel, indicating that alginate molecules had little effect on the microstructure. Nadezhda et al. also considered that the swelling capacity of the microspheres prepared by alginate and GA was mainly regulated by the content of GA ( Lewandowska-Łańcucka et al. , 2017 ). The higher swelling capacity of the hydrogels enhances cell proliferation and cell viability by facilitating transport of nutrients into the hydrogels ( Annabi et al. , 2011 ; Gu et al. , 2020 ). That corresponds well with our findings. When the scaffold is implanted into the body, material biocompatibility becomes a key issue. The scaffold material must not induce adverse reactions, sensitization, carcinogenicity, and irritation in cells, tissues, and systems of humans. The material is supposed to degrade itself. The degradation products need to be non-toxic and can be absorbed or metabolized by the human body. Therefore, whether biomaterials can be successfully applied in tissue engineering depends on the biocompatibility of materials and the toxicity of degradation products ( Catalano et al. , 2013 ). The biocompatibilities of chitosan, sodium alginate, and GA have been confirmed ( Chatelet, Damour & Domard, 2001 ; Li et al. , 1999 ; Sosnik, 2014 ); however, whether the blending of these three abovementioned polymers affect the biocompatibility of materials needs to be further studied. We found that the composite hydrogel had no cytotoxic effect on ADSCS according to staining results on live and dead cells. In addition to supporting the survival of ADSCs, growth curves based on MTT assay suggested that ADSCs can proliferate in mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels. All these results indicated that the two composite hydrogels have good biocompatibility. The composite gels are obtained by physical mixing of GA and polysaccharides, thus they retain various cell-friendly active sites of GA, such as arginine-glycine-aspartic acid ( Lee et al. , 2003 ), which then provides cultured cells with a friendly environment for growth and proliferation. Conclusions To mimic the chemical composition of natural tissue, the combination of GA and chitosan or alginate was used to successfully fabricate composite hydrogels in this paper. Hydrogels used for cell culture must exhibit desirable characteristics, such as good swelling capacity, proper mechanical property, and biocompatibility. In this study, the physical properties of five hydrogels were assessed. The composite hydrogels showed different mechanical properties and swelling capacities that depended on different polysaccharide added. Most importantly, this study showed that mTG/GA, C-mTG/GA, and A-mTG/GA hydrogels have excellent biocompatibility and can support ADSC survival, adhesion, and proliferation. Therefore, we believe the biomimetic composite hydrogels of GA and polysaccharides could be suggested as promising materials to cell carriers in tissue engineering. Supplemental Information 10. 7717/peerj. 11022/supp-1 Supplemental Information 1 Compression modulus data Click here for additional data file. 10. 7717/peerj. 11022/supp-2 Supplemental Information 2 Enzymatic property data Click here for additional data file. 10. 7717/peerj. 11022/supp-3 Supplemental Information 3 Swelling rate data Click here for additional data file. 10. 7717/peerj. 11022/supp-4 Supplemental Information 4 Results of MTT assay Click here for additional data file. 10. 7717/peerj. 11022/supp-5 Supplemental Information 5 Supplementary experiment data Raw data of supplementary experiment (Live/dead staining images of 3D cultures and results of degradation test). Click here for additional data file.

10. 7717/peerj. 12091

PeerJ

Transcriptome analysis reveals the mechanism of stromal cell-derived factor-1 and exendin-4 synergistically promoted periodontal ligament stem cells osteogenic differentiation

Stromal cell-derived factor-1 (SDF-1) and Exendin-4 (EX-4) play beneficial roles in promoting periodontal ligament stem cells (PDLSCs) osteogenic differentiation, while the detailed mechanism has not been clarified. In this study, we aimed to evaluate the biological mechanism of SDF-1 and EX-4 alone or synergistic application in regulating PDLSCs differentiation by RNA-sequencing (RNA-seq). A total of 110, 116 and 109 differentially expressed genes (DEGs) were generated in osteogenic medium induced PDLSCs treated by SDF-1, EX-4, and SDF-1+EX-4, respectively. The DEGs in SDF-1 group were enriched in signal transduction related signaling pathways; the DEGs in EX-4 group were enriched in metabolism and biosynthesis-related pathways; and the DEGs generated in SDF-1+EX-4 group were mainly enriched in RNA polymerase II transcription, cell differentiation, chromatin organization, protein phosphorylation pathways. Based on Venn analysis, a total of 37 specific DEGs were identified in SDF-1+EX-4 group, which were mainly enriched in negative regulation of autophagy and cellular component disassembly signaling pathways. Short time-series expression miner (STEM) analysis grouped all expressed genes of PDLSCs into 49 clusters according to the dynamic expression patterns and 25 genes, including NRSN2, CHD9, TUBA1A, distributed in 10 gene clusters in SDF-1+EX-4 treated PDLSCs were significantly up-regulated compared with the SDF-1 and EX-4 alone groups. The gene set enrichment analysis indicated that SDF-1 could amplify the role of EX-4 in regulating varied signaling pathways, such as type II diabetes mellitus and insulin signaling pathways; while EX-4 could aggravate the effect of SDF-1 on PDLSCs biological roles via regulating primary immunodeficiency, tight junction signaling pathways. In summary, our study confirmed that SDF-1 and EX-4 combined application could enhance PDLSCs biological activity and promote PDLSCs osteogenic differentiation by regulating the metabolism, biosynthesis and immune-related signaling pathways.

Introduction Periodontal disease-induced tooth loss has become a global public health challenge that greatly affects people’s quality of life ( Peres et al. , 2019 ). Mesenchymal stem cell (MSC) based periodontal tissue regeneration has aroused great attention in the field of regenerative medicine ( Hu, Liu & Wang, 2018 ). Among all MSCs, periodontal ligament stem cells (PDLSCs) are the main candidate cells for periodontal regeneration. Several studies have demonstrated that transplanting autologous and allogeneic PDLSCs directly into periodontal defect areas or surgically created periodontal defect areas could result in periodontal tissue regeneration, which highlights that PDLSC-mediated tissue engineering is a promising therapy for periodontitis ( Bartold, Shi & Gronthos, 2006 ; Ding et al. , 2010 ; Liu et al. , 2008 ; Liu et al. , 2019a ; Liu et al. , 2019b ). An increasing number of researchers have focused on the recruitment of endogenous PDLSCs to the injury site to enhance healing by harnessing the innate regenerative potential of the body ( Lee et al. , 2017a ). Cytokines, chemokines, and adhesion molecules have been used to enhance cell migration, maintain tissue homeostasis, regulate immune responses, promote wound healing and facilitate periodontal tissue regeneration ( Lee et al. , 2017b ; Onizuka & Iwata, 2019 ; Wang et al. , 2013 ). Stromal cell-derived factor-1 (SDF-1), a member of the chemokine family, can promote the proliferation and migration of various MSCs by activating the G protein-coupled receptor C-X-C chemokine receptor type 4 (CXCR4) ( Du, Feng & Ge, 2016 ; Kimura et al. , 2014 ; Zhu, Dissanayaka & Zhang, 2019 ). Our previous study also demonstrated that topical application of SDF-1 could significantly recruit MSCs to the wound area and promote local vascular regeneration in a rat model ( Wang, Du & Ge, 2016 ). SDF-1 possesses great potential in promoting MSC migration and growth; however, the compromised osteogenic differentiation of these cells could not be induced by SDF-1. Therefore, the application of SDF-1 alone is insufficient for favorable bone regeneration, and the optimal method for potentiating periodontal bone regeneration is to combine SDF-1 with other osteogenic factors. Exendin-4 (EX-4), a full agonist of glucagon-like peptide-1 receptor (GLP-1R), has been widely used in the clinical treatment of type 2 diabetes mellitus (T2DM) ( Yap & Misuan, 2019 ). In addition, EX-4 plays key roles in promoting MSC proliferation and migration ( Zhang et al. , 2016 ; Zhou et al. , 2015a ). Recently, EX-4 has been confirmed to present the potential to promote osteogenic differentiation and bone formation in a variety of stem/precursor cells ( Feng et al. , 2016 ; Luciani et al. , 2018 ; Meng et al. , 2016 ). Moreover, in addition to enhancing the MSC osteogenic differentiation capability, EX-4 could promote the recruitment effect of SDF-1 ( Zhou et al. , 2015b ). Our previous study also confirmed that SDF-1 and EX-4 cotherapy synergistically promoted PDLSCs proliferation, migration and osteogenic differentiation ( Liang et al. , 2021 ). However, the mechanism of SDF-1 and EX-4 alone or synergetic application for PDLSCs osteogenic differentiation is not fully understood. High-throughput RNA sequencing (RNA-Seq) has been widely applied to analyze the whole transcriptomic changes of eukaryotes, which can provide progressively greater knowledge of both the quantitative and qualitative aspects of transcript biology ( Ozsolak & Milos, 2011 ). RNA-seq has been successfully applied to identify the potential transcriptional mechanisms of various diseases, such as cancers, metabolic diseases and retinal diseases ( Bakhtiarizadeh et al. , 2019 ; Demircioğlu et al. , 2019 ; Farkas et al. , 2015 ). In the present study, RNA-seq transcriptomic analysis was applied to identify the core dynamic differentially expressed genes (DEGs) signature affected by EX-4 and SDF-1 alone or synergistically application in osteogenic medium-induced PDLSCs. Additionally, an integrated network containing specific DEGs generated in EX-4+SDF-1-treated PDLSCs was constructed. The results revealed the whole alteration of gene expression in PDLSCs undergoing EX-4 and SDF-1 application during the osteogenic differentiation process, which establishes a foundation for further research investigating the synergistic application of SDF-1 and EX-4 to promote PDLSCs osteogenic differentiation. Materials and Methods Human subjects and ethics statements This study was approved by the Medical Ethical Committee of the Stomatology School, Shandong University (NO. 20170801). Five healthy individuals without systemic diseases aged from 18–30 who underwent premolar extraction at the Department of Oral and Maxillofacial Surgery were recruited for this project. All individuals agreed to participate in the research project and signed the informed consent forms according to the Helsinki Declaration of 1975. Cell isolation and culture The extracted teeth were stored in Dulbecco’s modified Eagle’s medium (DMEM, HyClone, Logan, UT, USA) with 5% antibiotics (100 U/mL penicillin, 100 mg/mL streptomycin, Sigma Aldrich, St Louis, MO, USA) and quickly transported from the clinic to the laboratory. Then, single PDLSCs were acquired as previously described in our previous study ( Du, Feng & Ge, 2016 ). Specifically, Primary PDLSCs were cultured with DMEM containing 20% fetal bovine serum (FBS, BioInd, Kibbutz, Israel) at 37 °C in a humidified atmosphere of 5% CO 2, and cells were trypsinized and passaged at a dilution ratio of 1:3 to expand the culture in 10% FBS medium upon the cell monolayer reached 80–90% confluence. Fourth passage cells were used in all experiments. RNA-seq analysis PDLSCs from 5 different individuals were cultured in osteogenic medium (OM) and treated with SDF-1, EX-4 or SDF-1+EX-4 at 21 d, and normal PDLSCs treated with OM served as a negative control (NC). Totally, we collected 20 samples and RNA-seq was used to analyze the whole genome expression at LC Sciences through the Illumina X10 platform (Hangzhou, Zhejiang, China). Firstly, the total RNA were performed quality control based on previous study ( Wang, Wang & Li, 2012 ), and then the clean reads were mapped to the reference genome (GRCh38) via hierarchical indexing for spliced alignment of transcripts (HISAT) (v2. 0. 4) ( Kim, Langmead & Salzberg, 2015 ). The mapped reads of each sample were assembled using StringTie (v1. 3. 0, Pertea et al. , 2015 ). Furthermore, all transcriptomes samples were merged to reconstruct a comprehensive transcriptome using Perl scripts. After the final transcriptome was generated, StringTie and edgeR were used to estimate the expression levels of all transcripts. StringTie was used to determine the expression level of mRNAs by calculating fragments per kilobase of exon model per million mapped fragments (FPKM). The DEGs were selected with statistical significance ( p value <0. 05) by R package (v 3. 2. 5). Volcano analysis of DEGs in PDLSCs Volcano analysis was used to identify the DEGs between each pair of groups ( Li, 2012 ). The up- and down-regulated genes were identified, and the total number of each pair of groups was visualized by the histogram. Gene ontology (GO) and kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis Based on the DEGs generated by SDF-1, EX-4 and SDF-1+EX-4 compared with NC, the overrepresented GO categories and the significant KEGG pathways were identified ( anonymous, 2019 ; Kanehisa & Goto, 2000 ). A q value <0. 05 was used as the cut-off criterion for the selection of significant GO terms and KEGG pathways. Venn and Upset analysis To identify the specific DEGs generated by every two different compared groups, overlapping analysis was performed according to the jvenn website ( http://jvenn. toulouse. inra. fr/app/example. html ), and an intersection UpSet diagram based on the UpSet R package was drawn ( Conway, Lex & Gehlenborg, 2017 ). The specific genes generated by each group were identified and the gene functions were analyzed according to Metascape website ( http://metascape. org/gp/index. html#/main/step1 ) ( Zhou et al. , 2019 ). Short time-series expression miner (STEM) analysis STEM software (version 1. 3. 8) ( Douglas et al. , 2019 ) was applied to identify the specific gene expression clusters in PDLSCs treated with SDF-1, EX-4 and SDF-1+EX-4. The genes in the upregulated clusters in SDF-1+EX4-treated PDLSCs were selected, and the expression of all genes in these clusters was shown by a heatmap. The interaction relationship of these genes was analyzed according to the GeneMANIA database ( http://genemania. org/search ) ( Franz et al. , 2018 ). In-depth mechanism analysis and functional network construction To identify the function of specific DEGs generated by SDF-1, EX-4 and SDF-1+EX-4-treated PDLSCs, a functional network was constructed according to STRING database ( https://string-db. org/ ) and GeneMANIA database ( http://genemania. org/search ) ( Franz et al. , 2018 ). Functional enrichment analysis of genes in the functional network was further performed by the Metascape database ( Zhou et al. , 2019 ). Gene set enrichment analysis (GSEA) GSEA is one of the functional class scoring analysis methods ( Subramanian et al. , 2005 ). To select the genes that were not significantly differentially expressed but were important for the function of biological pathways, GSEA was performed method according to all genes in SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1 groups according to the clusterProfiler and enrichplot R package. Results DEG analysis To investigate changes in gene expression profiles in SDF-1, EX-4 or SDF-1+EX-4- treated PDLSCs, FPKM expression values of the genes were calculated based on the read counts using featureCounts software. The fold change (FC) values of each gene at different time points post stimulation compared with NC were also calculated using DESeq2 R package. The thresh old value of —FC—>1 and FDR ≤ 0. 05 was used to identify DEGs between two different groups, and the results indicated that 110, 116, 109, 125, 103 and 100 DEGs were generated in different comparison groups SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, EX-4 vs SDF-1 ( Fig. 1 ). Among all compared DEGs, 56, 61, 54, 58, 63 and 54 upregulated genes and 54, 55, 55, 67, 40 and 46 downregulated genes were identified ( Fig. S1 ). 10. 7717/peerj. 12091/fig-1 Figure 1 Volcano diagram of differentially expressed genes (DEGs) in PDLSCs. The DEGs in SDF-1 vs NC (A), EX-4 vs NC (B), SDF-1+EX-4 vs NC(C), SDF-1+EX-4 vs EX-4 (D), SDF-1+EX-4 vs SDF-1 (E), and EX-4 vs SDF-1 (F). Each point represents individual genes. Black dots represent the genes that were not significantly differentially expressed, red dots indicate the genes that were significantly upregulated and blue dots indicate the genes that were downregulated (—FC— > 1 and p -adjusted value < 0. 05). The volcano plots were analyzed by DESeq2. GO and KEGG enrichment analysis The GO (biological process) enrichment results showed that the DEGs generated by SDF-1 stimulation were mainly enriched in signal transduction regulation of transcription, cell differentiation, RNA polymerase II transcription, and multicellular organism development signaling pathways ( Fig. 2A and Table S1 ). The KEGG analysis results also indicated that SDF-1-induced DEGs were mainly enriched in the glucagon signaling pathway, NF-kappa B signaling pathway, and TNF signaling pathway ( Fig. 2B and Table S1 ), and mostly involved in human diseases and organismal systems, such as viral infectious diseases, the immune system, and the endocrine system ( Fig. S2B and Table S1 ). In addition, the DEGs generated by the EX-4 stimulation were mainly enriched in the oxidation–reduction process, metabolic process, regulation of transcription, DNA-templated, protein phosphorylation, and lipid metabolic process pathways ( Fig. 2C and Table S1 ). KEGG analysis also indicated that EX-4-induced DEGs were mainly enriched in metabolism and biosynthesis-related pathways, such as nicotinate and nicotinamide metabolism, phenylpropanoid biosynthesis, and phenylpropanoid biosynthesis ( Fig. 2D and Table S1 ), and all these DEGs were significantly involved in energy metabolism, lipid metabolism and biosynthesis of other secondary metabolites ( Fig. S2D ). Moreover, we identified that the DEGs generated by SDF-1+EX-4 was also mainly enriched in RNA polymerase II transcription, cell differentiation, chromatin organization, and protein phosphorylation pathways according to the GO (biological process) analysis results ( Fig. 2E and Table S1 ). The KEGG enrichment analysis results indicated that these costimulated DEGs were mainly enriched in pathogenic Escherichia coli infection, mitophagy in animals, bacterial invasion of epithelial cells and cysteine and methionine metabolism signaling pathways ( Fig. 2F and Table S1 ). Additionally, we found that all the DEGs generated by the SDF-1, EX-4, SDF-1+EX-4 groups were significantly enriched in the nucleus, membrane, nucleosome and cytoplasm pathways according to the GO cellular component analysis, and based on the GO molecular function analysis, all these DEGs were involved in the DNA binding, protein binding, ATP binding and protein heterodimerization activity pathways ( Figs. 2A, 2C, 2E and Figs. S2A, S2C, S2E ). 10. 7717/peerj. 12091/fig-2 Figure 2 Functional enrichment analysis of DEGs in compared groups. (A) GO enrichment analysis of the DEGs in SDF-1 vs NC. (B) KEGG enrichment analysis of the DEGs in SDF-1 vs NC. (C) GO enrichment analysis of the DEGs in EX-4 vs NC. (D) KEGG enrichment analysis of the DEGs in EX-4 vs NC. (E) GO enrichment analysis of the DEGs in SDF-1+EX-4 vs NC. (F) KEGG enrichment analysis of the DEGs in SDF-1+EX-4 vs NC. Red bar represents the enriched signaling pathways according to the biological process. Orange bar represents the enriched signaling pathways according to the cellular component. Gray bar represents the enriched signaling pathways according to the molecular function (A, C, E). The size of the symbol represents the gene counts enriched in the signaling pathways. The color indicates the degree of statistical significance (B, D, F). Screening of specific DEGs and functional analysis To characterize the DEGs specifically generated by SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1, we performed an overlapped analysis of the six compared DEG groups, and the results indicated that 34, 30, 37, 39, 35 and 28 DEGs were specially generated ( Fig. 3A, Table S2, Fig. S3 and Table S3 ). In addition, the Metascape website was referenced to analyze the specific gene functions in different groups and the results indicated that the 34 specific DEGs generated in SDF-1 vs NC was mainly enriched in response to the wounding, and macromolecule methylation signaling pathways; the 30 specific DEGs generated in EX-4 vs NC were mainly enriched in the cellular response to external stimulus, regulation of protein complex assembly, and regulation of chromosome organization signaling pathways; the 37 specific DEGs generated in SDF-1+EX-4 vs NC were mainly enriched in the negative regulation of autophagy and cellular component disassembly signaling pathways; the 39 specific DEGs generated in SDF-1+EX-4 vs SDF-1 were mainly enriched in the Notch signaling pathway and Asparagine N-linked glycosylation signaling pathways; and the 28 specific DEGs generated by EX-4 vs SDF-1 group were mainly enriched in the response to inorganic substance, embryonic organ development and A positive regulation of cell migration signaling pathways ( Fig. 3B and Table S4 ). 10. 7717/peerj. 12091/fig-3 Figure 3 Screening of specific DEGs and function analysis of each compared group. (A) Venn analysis of the DEGs among SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1. (B) Functional enrichment analysis of the specific DEGs in each two compared groups (SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs SDF-1, EX-4 vs SDF-1), which were exhibited in the outermost layer in the Venn analysis. Screening of core DEGs generated by the SDF-1 and EX-4 combined application To identify the core DEGs generated by the SDF-1 and EX-4 combined application, STEM software was applied to perform a pattern analysis (mock infection was designated NC), and the results revealed 49 gene clusters among all DEGs generated by SDF-1, EX-4 and SDF-1+EX-4-treated PDLSCs ( Fig. S4 and Table S5 ). In all gene clusters, we focused on genes in clusters 14, 39, 47, 36, 7, 28, 45, 25, 18, and 41, which were significantly upregulated by the SDF-1 and EX-4 combined application compared with SDF-1 or EX-4 alone ( Fig. 4A ). Among the 10 clusters, a total of 25 genes, including NRSN2, CHD9, TUBA1A, AKAP13, VAMP7, NPIPA2 etc. , were identified, and all DEG expression in different groups is shown by a Heatmap ( Fig. 4B and Table S6 ). The GeneMANIA analysis network showed that these genes possessed an intricate tangle of connections through genetic interactions and co-expression pattern ( Fig. 4C and Table S7 ). Functional enrichment analysis indicated that our core DEGs and their related genes constructed in the network were mainly enriched in the metallothioneins binding metals, asparagine N-liked glycosylation, response to stimulus, detoxification, biological regulation and growth signaling pathways ( Fig. 4D and Table S8 ). 10. 7717/peerj. 12091/fig-4 Figure 4 Screening of core DEGs generated by the combined SDF-1 and EX-4 application. (A) Short Time-series Expression Miner (STEM) analysis in PDLSCs cocultured with SDF-1, EX-4 and SDF-1+EX-4. Gene clusters (including 14, 39, 47, 36, 7, 28, 45, 25, 18, 41) that were upregulated by the SDF-1and EX-4 combined application were selected out (mock infection was designated as NC, the first node was SDF-1 stimulated group, the second node was EX-4 stimulated group, and the final node was SDF-1+EX-4 combined stimulated group). (B) Heat map of the 25 upregulated DEGs generated by the combined SDF-1 and EX-4 application. (C) GeneMANIA analysis network based on these 25 DEGs; the genes in circles with a white slash are the actual DEGs, and the genes in circles without a slash are genes predicated based on physical interactions, coexpression, predications, colocalization, pathways, genetic interactions and shared protein domains. (D) Functional enrichment analysis of genes presented in the GeneMANIA network of C. Network construction and the mechanism analysis of the SDF-1 and EX-4 synergistic effect in PDLSCs To further clarify the mechanism of the SDF-1 and EX-4 combined application to promote PDLSC osteogenic differentiation, a detailed complex network analysis was performed based on the DEGs generated by the different groups through the STRING and GeneMANIA. A network analysis based on the STRING database showed that the DEGs generated by EX-4 and SDF-1 alone were centralized and converged into a network, while the DEGs in the SDF-1+EX-4 group were scattered in the network. The DEGs generated in SDF-1+EX-4 vs SDF-1, and SDF-1+EX-4 vs EX-4 were also centralized ( Fig. S5 ). In addition, compared with the NC group, EX-4 significantly elevated the expression of the core gene MAPK27, which plays central roles by activating MAPK-related signaling pathways; and SDF-1 mainly changed the expression of the key genes CREB1, MMP13, RHOQ and BIRC2, which exert their roles by activating ATP signaling pathways ( Figs. 5A and 5B ). More importantly, we identified that the gene expression levels of NEDD8, CHCHD1, LMO7, and ATP5L were significantly varied the SDF-1+EX-4 vs EX-4 groups, which played crucial roles in activating ATP-related signaling pathways and indicated that SDF-1 could amplify the EX-4 effect in elevating ATPase activity and promoting PDLSC osteogenic differentiation. Additionally, we found that the DEGs of SOX15, UBC, VAMP7 and ARPC5 were significantly changed in SDF-1+EX-4 vs SDF-1 groups, which played vital roles in promoting cytoskeletal protein formation and degradation ( Figs. 5C and 5D ). Although the DEGs generated in the SDF-1+EX-4 vs NC group were dispersed, most of these genes could be centralized by the genes HAP1, TP53, TAL1, RRPF40A and TALDO1 ( Fig. S6 ). 10. 7717/peerj. 12091/fig-5 Figure 5 Network construction and the mechanism analysis of SDF-1 and EX-4 synergistic effects in PDLSCs. The network constructed through the GeneMANIA database based on the core DEGs in SDF-1 vs NC (A), EX-4 vs NC (B), SDF-1+EX-4 vs EX-4 (C), and SDF-1+EX-4 vs SDF-1 (D). The genes in circles with a white slash are the actual dynamic DEGs. The dynamic DEGs and predicted genes interact based on physical interactions, coexpression, predictions, colocalization, pathways, genetic interactions and shared protein domains. Gene set enrichment analysis (GSEA) based on all genes All of the above analyses were based on the DEGs selected according to our threshold value and statistical analysis technique; however, genes that are not significantly differentially expressed but are important for the function of the SDF-1 and EX-4 biological pathways may be omitted, and thus, we performed a GSEA according to all genes in SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1. The results indicated that SDF-1 played its roles in regulating the metabolism of xenobiotics by cytochrome P450, regulation of autophagy, neuroactive ligand receptor interaction signaling pathways ( Fig. 6A and Table S9 ); EX-4 played roles mainly through affecting the metabolism related signaling pathways, such as, starch and sucrose metabolism, arginine and proline metabolism, and type II diabetes mellitus signaling pathways ( Fig. 6B and Table S9 ); and the combination of SDF-1 and EX-4 mainly regulating PLDSCs biological process through activating metabolism and immunity related signaling pathways, such as pantothenate and COA biosynthesis, vascular smooth muscle contraction, and maturity onset diabetes of the young ( Fig. 6C and Table S9 ). In addition, we confirmed that SDF-1 could amplify the role of EX-4 in regulating various signaling pathways, such as type II diabetes mellitus, the insulin signaling pathway, and the allograft rejection pathway ( Fig. 6D and Table S9 ), while EX-4 could aggravate the effect of SDF-1 on PDLSC biological roles by regulating signaling pathways, including primary immunodeficiency, tight junction, and basal transcription factors ( Fig. 6E and Table S9 ). Interestingly, we found that the signaling enrichment analysis based on the DEGs and GSEA shared similarity, which indicated that SDF-1 and EX-4 may regulate PDLSC biological activity by enhancing each other’s biological functions. 10. 7717/peerj. 12091/fig-6 Figure 6 GSEA enrichment based on all genes. GSEA enrichment plots of gene expression signatures of SDF-1 vs NC (A), EX-4 vs NC (B), SDF-1+EX-4 vs NC (C), SDF-1+EX-4 vs EX-4 (D), and SDF-1+EX-4 vs SDF-1(E), which are sorted according to the differences between the means of samples with high and low HOTAIR expression. The barcode plot indicates the position of the genes in each gene set; red and blue colors represent positive and negative Pearson correlations with HOTAIR expression, respectively. Discussion The effects of SDF-1 and EX-4 alone on bone regeneration have been widely reported; however, the synergetic effects of SDF-1 and EX-4 on PDLSC osteogenic differentiation and the potential mechanism have not been reported. In our previous study, we confirmed that the combined application of SDF-1 and EX-4 could significantly promote osteogenic differentiation of PDLSCs ( Liang et al. , 2021 ). In this study, we confirmed that SDF-1 could amplify the role of EX-4 by significantly activating metabolism-related signaling pathways, such as type II diabetes mellitus and insulin signaling pathways; and EX-4 could aggravate the effect of SDF-1 on PDLSCs biological roles by regulating primary immunodeficiency and tight junction signaling pathways. Briefly, our study confirmed that the SDF-1 and EX-4 combined application could enhance PDLSCs biological activity and promote PDLSCs osteogenic differentiation by regulating the metabolism, biosynthesis and immune-related signaling pathways. EX-4, a common glucagon-like peptide-1 receptor agonist, has been confirmed to possess excellent effects on treating patients with T2DM by significantly reducing HbA1c content compared to basal insulins ( Singh et al. , 2017 ). Our current study confirmed that EX-4-generated DEGs were enriched in the type II diabetes mellitus signaling, starch and sucrose metabolism, arginine and proline metabolism, and alanine aspartate and glutamate metabolism signaling pathways, while the combination of SDF-1 and EX-4 could significantly activate more metabolism-related signaling pathways, such as the valine leucine and isoleucine degradation, insulin signaling, phenylalanine metabolism, and pyruvate metabolism signaling pathways. In the plasma of type II diabetes, a pronounced postprandial rise in amino acids (such as leucine, isoleucine, valine, lysine, and threonine) and glucose-dependent insulinotrophic polypeptide was observed, which finally resulted in glycemic and insulinemic responses ( Nilsson, Holst & Björck, 2007 ). In our study, we confirmed that SDF-1 and EX-4 combination therapy could significantly increase the gene expression of EHHADH, HMGCL, IL4I1, PKLR, PIK3CG, PYGM, SLC2A4, RHOQ, PRKCZ, FBP1, and SH2B2, which play key roles in promoting amino acid degradation and insulin secretion; thus, SDF-1 assists EX-4 in controlling the blood glucose level of diabetes patients. EX-4 could promote the osteogenic differentiation of osteoblasts, adipose-derived stem cells, and PDLSCs by activating the Hedgehog/Gli1, Wnt and NF- κ B signaling pathways; however, the interactive relationship of EX-4 regulating PDLSCs metabolism-related pathways and osteogenic differentiation pathways and the mechanism by which SDF-1 amplifies the effect of EX-4 on PDLSCs osteogenic differentiation need to be further clarified ( Deng et al. , 2019 ; Gao, Li & Li, 2018 ; Liu et al. , 2019a ; Liu et al. , 2019b ). SDF-1 could promote PDLSCs proliferation, migration and osteogenic differentiation in vitro, play key roles in recruiting endogenic PDLSCs into the periodontal defect area and then contribute to angiogenesis in vivo ( Bae et al. , 2017 ; Du, Feng & Ge, 2016 ). In the current study, we confirmed that SDF-1 significantly activated metabolism-related signaling pathway of PDLSCs cultured in OM, such as the metabolism of xenobiotics by cytochrome P450, alanine aspartate and glutamate metabolism; however, the effects of metabolism on PDLSCs osteogenic differentiation have not been reported. In addition, SDF-1 activated the renin angiotensin system, basal transcription factor, and neuroactive ligand receptor interaction signaling pathways, which is consistent with previous studies ( Chu et al. , 2009 ; Wang et al. , 2015 ). Moreover, we found that the EX-4 and SDF-1 combined stimulation significantly activated the immunodeficiency, tight junction, complement and coagulation cascade signaling pathways compared to SDF-1 stimulation alone in osteogenic medium cultured PDLSCs. The adaptive immune system plays a prominent role in the development of heterotopic ossification ( Ranganathan et al. , 2016 ), and tight junctions (TJs) play a pivotal role in the modulation of paracellular permeability. For example, Cldn11 recombinant protein, a well-established component of TJs, could significantly decrease the resorption of lipopolysaccharide-induced calvarial bone and increase the osteogenic activity of calvarial bone formation ( Baek et al. , 2018 ). Moreover, stress-induced hematopoietic stem cell mobilization is enhanced by the fibrinolytic and complement cascades ( Nguyen, Lapidot & Ruf, 2018 ). In the current study, we confirmed that EX-4 could enhance the role of SDF-1 in PDLSCs osteogenic differentiation through RNA-seq analysis, although this process was limited due to the lack of biological experimental validation. In the future, we will design related biological experiments to validate the mechanism of the SDF-1 and EX-4 combined application in promoting PDLSCs osteogenic differentiation. In adipose-derived stem cells, EX-4 could activate the PI3K/AKT pathways and then augment the SDF-1α/CXCR4 cascade, which finally promotes cell migration ( Zhou et al. , 2015b ). In endothelial progenitor cells, EX-4 could activate the SDF-1β/CXCR7-AMPK/p38-MAPK axis and then ameliorate high glucose-induced EPC dysfunction ( Yang et al. , 2020 ). In the current study, we confirmed that EX-4 and SDF-1 coordinate to change PDLSCs gene expression of CHD9, MT1A, RNF145, ASPN, SIX1, TUBA1A, PRR5L, UBFD1, SEC14L1, ANP32E, AKAP13, and VAMP7, which could combine with TMEM159, CEACAM7, CEBPB, ABCG4, NPAT, UNG, TMEM256, MT1F, CCDC28A, DUSP19, ANGPTL2, PRPSAP1, C9orf116, MT1H, TIPRL, MAN1B1, FPGT, AHDC1, TMEM5, and C5orf15, and then activate various signaling pathways including response to stimulus, detoxification, biological regulation and growth pathways. However, the role of these DEGs and signaling pathways identified by RNA-seq in the current study on PDLSCs osteogenic differentiation needs to be further validated. Conclusion Our study confirmed that SDF-1 could amplify the role of EX-4 by activating more metabolism-related signaling pathways, such as type II diabetes mellitus and the insulin signaling pathways, and EX-4 could aggravate the effect of SDF-1 on PDLSCs biological roles by regulating primary immunodeficiency and tight junction signaling pathways. In addition, we confirmed that the SDF-1 and EX-4 combined application could enhance PDLSCs biological activity and promote PDLSCs osteogenic differentiation by regulating metabolism, biosynthesis and immune-related signaling pathways. Our current study lays a solid foundation for exploring the effects of SDF-1 and EX-4 synergistic application on periodontal tissue regeneration. Supplemental Information 10. 7717/peerj. 12091/supp-1 Supplemental Information 1 Bar plot of differentially expressed genes (DEGs) of PDLSCs Bar plot of differentially expressed genes (DEGs) of PDLSCs in SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1. The red histogram indicates the number of genes that were significantly upregulated and the blue histogram indicates the number of genes that were downregulated ( p -adjusted value <0. 05; FDR <0. 05). The numbers of the DEGs are indicated at the top of every bar. Click here for additional data file. 10. 7717/peerj. 12091/supp-2 Supplemental Information 2 Functional enrichment analysis of DEGs in different groups GO enrichment analysis of the DEGs in SDF-1 vs NC (A), EX-4 vs NC (C), and SDF-1+EX-4 vs NC groups (E). KEGG enrichment analysis of the DEGs in the SDF-1 vs NC (B), EX-4 vs NC (D), and SDF-1+EX-4 vs NC groups (F). The size of the symbol represents the gene counts enriched in the signaling pathways. The color indicates the degree of statistical significance (A, C, E). The different color bars represent the different enrichment methods of the KEGG catalog (B, D, F). Click here for additional data file. 10. 7717/peerj. 12091/supp-3 Supplemental Information 3 UpSet plot analysis of the DEGs UpSet plot analysis of the DEGs among the SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1 groups. Click here for additional data file. 10. 7717/peerj. 12091/supp-4 Supplemental Information 4 Short Time-series Expression Miner (STEM) analysis in PDLSCs Short Time-series Expression Miner (STEM) analysis in PDLSCs cocultured with SDF-1, EX-4 and SDF-1+EX-4 (mock infection was designated NC; the first node was the SDF-1 stimulated group, the second node was the EX-4 stimulated group, and the final node was the SDF-1+EX-4 combined stimulated group). Click here for additional data file. 10. 7717/peerj. 12091/supp-5 Supplemental Information 5 Network constructed through the STRING database Network constructed through the STRING database based on the DEGs in SDF-1 vs NC (A), EX-4 vs NC (B), SDF-1+EX-4 vs NC (C), SDF-1+EX-4 vs Ex-4 (D), SDF-1+EX-4 vs SDF-1 (E), and EX-4 vs SDF-1 (F). Click here for additional data file. 10. 7717/peerj. 12091/supp-6 Supplemental Information 6 Network constructed through the GeneMANIA database Network constructed through the GeneMANIA database based on the core DEGs in SDF-1+EX-4 vs NC (left panel) and the network based on the core genes generated by SDF-1+EX-4 (right). The genes in circles with white slash are the actual dynamic DEGs. The dynamic DEGs and predicted genes are interact based on physical interactions, coexpression, predications, colocalization, pathways, genetic interactions and shared protein domains. Click here for additional data file. 10. 7717/peerj. 12091/supp-7 Supplemental Information 7 GO and KEGG functional enrichment analysis results based on DEGs GO and KEGG functional enrichment analysis results based on DEGs in SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, SDF-1+EX-4 vs SDF-1, and EX-4 vs SDF-1. Related to Fig. 2 and Fig. S2. Click here for additional data file. 10. 7717/peerj. 12091/supp-8 Supplemental Information 8 Gene symbols of all DEGs generated in various compared groups Related to Fig. 3A. Click here for additional data file. 10. 7717/peerj. 12091/supp-9 Supplemental Information 9 Gene symbols of all DEGs generated in various compared groups Related to Fig. S3. Click here for additional data file. 10. 7717/peerj. 12091/supp-10 Supplemental Information 10 Functional enrichment analysis of the specific DEGs in each two compared groups Functional enrichment analysis of the specific DEGs in each two compared groups (SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs SDF-1, EX-4 vs SDF-1). Related to Fig. 3B. Click here for additional data file. 10. 7717/peerj. 12091/supp-11 Supplemental Information 11 Short Time-series Expression Miner (STEM) analysis in PDLSCs Short Time-series Expression Miner (STEM) analysis in PDLSCs cocultured with SDF-1, EX-4 and SDF-1+EX-4. Related to Fig. 4A. Click here for additional data file. 10. 7717/peerj. 12091/supp-12 Supplemental Information 12 Heat map of the 25 upregulated DEGs generated by the SDF-1 and EX-4 combined application Related to Fig. 4B. Click here for additional data file. 10. 7717/peerj. 12091/supp-13 Supplemental Information 13 GeneMANIA analysis network analysis results based on the 25 DEGs Related to Fig. 4C. Click here for additional data file. 10. 7717/peerj. 12091/supp-14 Supplemental Information 14 Functional enrichment analysis of genes presented in the GeneMANIA network of Fig. 4C Related to Fig. 4D. Click here for additional data file. 10. 7717/peerj. 12091/supp-15 Supplemental Information 15 GSEA enrichment based on all genes GSEA enrichment based on all genes in SDF-1 vs NC, EX-4 vs NC, SDF-1+EX-4 vs NC, SDF-1+EX-4 vs EX-4, and SDF-1+EX-4 vs SDF-1. Related to Fig. 6. Click here for additional data file.

10. 7717/peerj. 12188

PeerJ

The effect of cartilage decellularized extracellular matrix-chitosan compound on treating knee osteoarthritis in rats

Knee osteoarthritis (KOA) refers to a common disease in orthopaedics, whereas effective treatments have been rarely developed. As indicated from existing studies, chondrocyte death, extracellular matrix degradation and subchondral bone injury are recognized as the pathological basis of KOA. The present study aimed to determine the therapeutic effect of decellularized extracellular matrix-chitosan (dECM-CS) compound on KOA. In this study, rat knee cartilage was decellularized, and a satisfactory decellularized extracellular matrix was developed. As suggested from the in vitro experiments, the rat chondrocytes co-cultured with allogeneic dECM grew effectively. According to the results of the alamar blue detection, dECM did not adversely affect the viability of rat chondrocytes, and dECM could up-regulate the genes related to the cartilage synthesis and metabolism. As reported from the animal experiments, dECM-CS compound could protect cartilage, alleviate knee joint pain in rats, significantly delay the progress of KOA in rats, and achieve high drug safety. In brief, dECM-CS compound shows a good therapeutic effect on KOA.

Introduction KOA refers to a degenerative joint disease in orthopaedics. It is primarily characterized by synovitis, degeneration of cartilage, formation of osteophyte, as well as sclerosis of subchondral bone ( Mandl, 2019 ). The main clinical treatment complies with the mitigation of pain symptoms at the early stage of this disease. When osteoarthritis of the knee develops to advanced stage, knee replacement is recognized as the main existing treatment ( Bannuru et al. , 2019 ; Meiyappan et al. , 2020 ; Bijlsma, Berenbaum & Lafeber, 2011 ). Thus far, no effective drug is capable of preventing the progression of KOA. The incidence and prevalence of KOA will further rise over the coming decades as impacted by the aging population, the rising obesity rates and the high rates of traumatic knee injuries ( Martel-Pelletier et al. , 2016 ). This is recognized as a public health crisis, and rigorous high-quality KOA clinical research is urgently required to ensure patients receive safe and effective treatments. One of the OA progression mechanism is a reduction of chondrocyte numbers and ECM degradation in cartilage ( Carballo et al. , 2017 ). Chondrocytes are unique cells in articular cartilage, which are capable of synthesizing ECM components. ECM largely comprises COL2 and GAGs, which are of a critical significance for the biomechanical properties of cartilage. It protects the articular surface of the bone from abrasion, distributes the applied loads over a larger joint area, and creates a smooth, lubricated surface for the joint movement to reduce friction ( Ravindran et al. , 2015 ). dECM materials have been extensively employed in a wide range of biological and medical fields. The ECM obtained by decellularization exhibits low immunogenicity and retains the characteristics of ECM ( Lee et al. , 2020 ; Yao et al. , 2019 ). Chitosan (CS) is a non-toxic natural polymer, which has exhibits biodegradability and biocompatibility; it acts as a prominent bio-material ( Comblain et al. , 2017 ). By mixing chitosan with other natural or synthetic polymers, the multifaceted performance of the mentioned biological scaffolds can be effectively promoted ( e. g. , controlling the porosity and the water retention, reducing their biodegradation, enhancing their bioactivity and biocompatibility, and improving their mechanical properties) ( Wang et al. , 2020 ). The CHI/SF/ESM hydrogels with ECM-mimicking interconnected structures were successfully synthesized by Terin Adal etc. These hydrogels show superior mechanical properties for desired application and support growth, adhesion and differentiation of Human chondrocyte cells where high viability was observed under normal in vitro cell culturing conditions ( Adali, Kalkan & Karimizarandi, 2019 ). Most of the mentioned biological composite scaffolds made by CS exhibit no cytotoxicity and can promote the cell attachment and proliferation for the cartilage repair ( Kean & Thanou, 2010 ). The author has previously shown that the concentration of 40% (w/v) dECM suspension is the optimal in the rabbit KOA therapy ( Zhang et al. , 2020 ). In the present study, dECM of rat cartilage and chitosan was used to create dECM-CS compound to determine the therapeutic effect of dECM-CS compound to treat KOA. Furthermore, the experiments in vitro were continuously performed to explore the intervention effect of dECM on chondrocytes ( Fig. 1 ). 10. 7717/peerj. 12188/fig-1 Figure 1 Flow chart. Flow chart of dECM-CS compound preparation (picture is not copyrighted). Images adapted from BioRender Basic plan ( https://app. biorender. com/ ). Materials and Methods Decellularization In this study, the cartilage pieces were obtained from the femoral side articular surface of the fresh rat knee joints. Subsequently, the specimens were washed completely with phosphate buffered saline (PBS) to ensure that the specimen was free of impurities. The method adopted complied with the description reported by Kheir et al. (2011). Such a method was adopted in this experiment based on some improvements. In brief, the fresh rat knee cartilage underwent several freeze-drying cycles. Subsequently, the cartilage was treated with aprotinin, SDS, EDTA and tris-Hcl, then with the DNase and RNase reaction, and lastly washed with TritonX-100. To achieve the disinfection, Penicillin (100 U/ml), streptomycin (100 μg/ml), and fluconazole (2. 5 μg/ml) were added to all the solutions used. Decellularization evaluation Histology First, some fresh undealted rat cartilage pieces were obtained, and the cartilage pieces were decellularized. The specimen was fixed with paraformaldehyde (4% [w/v]) at normal temperatures for 24 h. Next, those tissues were decalcified with EDTA solution (25% [w/v], pH 7. 0) for 4–5 weeks till the hardness of cartilage tissue was exactly right to be sectioned. In this period, the EDTA solution was changed once per three days. Lastly, the hematoxylin-eosin (HE) staining, the alcian blue staining and the scanning electronic microscope (SEM) were used to identify whether the decellularization was successful. Biochemical analysis of GAG and DNA content The fresh or dealted cartilage pieces were placed in a papain solution to be digested at 60 °C for 12 h. Next, this reaction solution was centrifuged at 10, 000 g for 30 min. The GAG content was evaluated by employing the dimethylmethylene blue colorimetric quantitative detection kit. By complying with the instructions, a GAG standard curve was plotted by using the standard samples. Subsequently, the exact GAG content was obtained by using such a standard curve. The DNA content was determined by employing the dsDNA HS Assay Kit for Qubit. Likewise, a DNA standard curve was plotted by using the standard samples, Afterwards, the exact DNA content was obtained. The content of GAG or DNA was the wet weight per milligram of the sample, as expressed in micrograms. Scanning electron microscopy (SEM) The fresh and decellularized rat cartilage pieces were immersed in a glutaraldehyde solution (2. 5% [w/v]) for 2 h. Next, the tissue was flushed with a sodium dimethylarsenate buffer (pH 7. 4). Afterwards, the pieces were fixed with osmium tetroxide (1% [w/v]) and sprayed gold on the tissue for 30 s by the ion sputtering apparatusthe. Lastly, the samples were scanned with the SEM. Manufacturing of dECMs-CS compound, dECMs suspension After the decellularization, a tissue lapping apparatus (Servicebio, Wuhan, China) was adopted to grind the decellularized cartilage piece particle. Subsequently, the dECM particle was washed intensively with normal saline. So dECM is a non-immunogenic biological tissue material obtained from rat knee cartilage tissue through decellularized process, and its main component is GAGs. Finally, the dECMs were mixed with chitosan liquid or normal saline, respectively, at a concentration of 40% [w/v]. Measuring the dECM particle The dECM particles were stained with alcian blue, and then the particles were measured with imageJ. Identification of chondrocytes The collagen II immunofluorescence staining and the alcian blue staining were used to identify the cytology. Rat chondrocytes cultured with dECM The rat chondrocytes were cultured in a medium supplemented by 40% dECM suspension for 3–5 days to observe their growth status. Alamar blue detection The dECM was used to interfere with rat chondrocytes, and alarmarBlue™Cell Viability Assay Reagent (Solarbio, Beijing, China) was introduced to determine the chondrocytes viability. In brief, chondrocytes were seeded in 96-well plates at a concentration of 40, 000 cells/ml, and 40% dECM was adopted to interfere with the chondrocytes for 1 d, 3 d, 5 d and 7 d. Subsequently, 10 μl alamar blue reagent was added into the respective well, and the mentioned well was placed in the cell incubator for 4 h (37 °C, 5% CO 2 ). The reduction rate of alamar blue reagent was calculated by complying with the manual. Reverse transcription polymerase chain reaction (RT-PCR) The primer sequences are listed in Table 1. Based on the instructions, the total RNAs were extract by treating the chondrocytes of rats with the TRIzol reagent (Sigma, Ronkonkoma, NY, USA). Next, the cDNA was synthesized by using the Prime Script™ RT Master Mix Kit (Takara, Tokyo, Japan). With the GAPDH as the internal reference control, RT-PCR was performed with the TB Green™ Premix Ex TaqTM II Kit (Takara, Tokyo, Japan). The result was calculated by the instructions. Information regarding the primer sequences is given in Table 1. 10. 7717/peerj. 12188/table-1 Table 1 Primers used for RT-qPCR. Genes Primer sequence (5′–3′) Col2A1 S: GCCTCGCGGTGAGCCATGATC AS: CTCCATCTCTGCCACGGGGT ACAN S: TAGAGGATGTGAGTGGTCTT AS: TCCACTAAGGTACTGTCCAC SOX-9 S: GAGCTGAGCAGCGACGTCATCT AS: GGCGGCGCCTGCTGCTTGGACA PRG4 S: AACAGGGAAGATAGTGGC AS: CGTAGTAATCATAGCCGTCA GAPDH S: CACTGTGCCCATCTACGA AS: TGATGTCACGCACGATTT Note: S, sense; AS, antisence. Establishment of OA model and intra-articular injection treatment Specific to the in vivo experiment, sixty adults, Sprague–Dawley, male, rats weighing 260 ± 10 g were employed for this experiment. We obtained experimental rats from Yangzhou University, and all the rats were raised in the Comparative Medical Center of Yangzhou University. The feeding conditions were quiet, ventilated, clean and dry, the temperature was suitable, feed and drinking water were added regularly. Yangzhou University provided full approval for this research (YZUNSFC2020-LCYXY-24), and the experiment complied with the Guide for the Care and Use of Laboratory Animals (National Academies, 2011). The rats were anesthetized with 7% [w/v] chloral hydrate (0. 5 ml/100 g) based on the intraperitoneal injection. A post-traumatic OA model was built through the anterior cruciate ligament transection (ACLT) ( Aizah, Chong & Kamarul, 2019 ; Cohen-Solal, Funck-Brentano & Hay, 2013 ; Florea et al. , 2015 ). In brief, under general anesthesia, the anterior cruciate ligament of the right knee was transected. In addition, a sham operation was performed on the contralateral knee, while no ligament transection was performed. The rats randomly fell to six groups, i. e. , normal, sham-operated, ACLT-operated treated with physiological saline, and treated with 100 μl of chitosan, 100 μl of 40% [w/v] dECM-suspension, 100 μl of 40% [w/v] dECM-CS compound, respectively. To prevent the infection, cefazolin was injected intramuscularly 30 min before and three days after the surgery, and the celecoxib was given 1 week after surgery to relieve pain. The rats underwent an intra-articular injection therapy twice a week for 4 or 8 w. At the end of the experiment, all rats were killed with excessive anesthetic drugs to obtain knee joint specimens, and the animal carcasses were uniformly disposed of by the Animal specimen processing Center of Yangzhou University. Hot plate test Pain refers to a prominent clinical manifestation of KOA, and the hot plate test has been extensively described as a good tool to assess the pain of animal models ( Barrot, 2012 ). In this study, the hot plate test was performed to record the response time of rats to pain, and the therapeutic effect of KOA in the rats here was indirectly assessed. A hot plate (Softmaze, Shanghai, China) was employed at a pre-set plate temperature of 52. 5 °C as recommended for rats. Moreover, a cut-off time was set to 60 s. The time from when the rat was placed on the hot plate till the first sign of discomfort from the thermal stimulus was recorded immediately. To be specific, the licking, shaking, or stepping of the paws was observed. None showed the signs of thermally-induced damage to the paws once either study was conducted. The observers were blinded to the experiment. Weight-bearing asymmetry test To assess the pain and inflammation of the hind limbs of rats, an incapacitance tester (RWD Biotechnology, Shenzhen, China) was employed to measure the weight distribution of hind limbs ( Hamilton et al. , 2015 ). In the static load test, the rats were placed in a retainer, in which the animal could be comfortably maintained, while its rear paws were shelved on two separated sensor plates. By regulating the weight distribution on the rear paw, the animal stood, and the pain was alleviated naturally. The duration was set to 9 s, and the instrument readings indicated the different weight value of the sham (left) hind limb and the ACLT (right) hind limb. The observers were blinded to the experiment. Histological analysis The rat knee joints were obtained, and the specimens were placed in 4% [w/v] paraformaldehyde for 24 h. Subsequently, the specimens were incubated in the EDTA solution (decalcifying solution consisting of 25% [v/w] EDTA, pH 7. 0) at ambient temperatures. The EDTA solution was changed per three days. When the specimens were sufficiently soft to be sectioned, they were embedded in paraffin and then sliced. Next, the mentioned histological sections were stained with HE, the safranin O-fast green staining and alcian blue. Furthermore, the Osteoarthritis Research Society International (OARSI) grading system was adopted to score histopathologic variations in osteoarthritic cartilage ( Pritzker et al. , 2006 ). Identification of hepatorenal toxicity At 4 weeks and 8 weeks after the knee joint injection, the rats were randomly selected in the respective group. Their liver and kidney tissues were taken, and HE staining sections were performed to observe whether their liver and kidney cell structures were damaged, as well as determine the hepatorenal toxicity of dECM on the rats. Results Decellularization evaluation Histological The following figures present two groups of the cartilage tissue by using different histological staining processes. In the fresh undealt cartilage tissue, the HE staining showed the chondrocytes to be embedded in lacunae in the matrix. In the dealt pieces after the decellularization, however, almost all of the chondrocytes were removed away from the tissue. As demonstrated from the figure of the alcian blue staining, the staining intensity between the two group of cartilage pieces was consistent, indicating that GAG was not much to lose ( Fig. 2A ). 10. 7717/peerj. 12188/fig-2 Figure 2 Comparison of rat cartilage tissue before and after the decellularization. (A) HE staining, alcian blue staining and SEM scanning. (B) DNA content. *** P < 0. 001. (C) GAG content. (D) Cartilage tissue of rat knee joint, dECM particle characterization alcian blue. Biochemical analysis The results revealed that before and after the decellularization, the content of GAG per miligram of wet cartilage tissue was 5. 83 ± 0. 03 μg/mg and 5. 48 ± 0. 23 μg/mg, and the content of DNA per miligram of wet cartilage tissue was 0. 87 ± 0. 40 μg/mg and 0. 14 ± 0. 02 μg/mg. Moreover, no statistically significant difference was reported in GAG content ( t test, n = 3, P > 0. 05), ( Fig. 2C ) and a highly significant reduction was identified in the DNA content ( t test, n = 3, P < 0. 05) ( Fig. 2B ). Microstructural As demonstrated from the SEM scanning graph, the dealt cartilage lost the most of cells, the surface structure became loose, the chondrocytes were removed after the decellularization, and almost no cells in the pore after the decellularization ( Fig. 2A ). Staining and measurement of dECM particle The dECM particle was stained with alcian blue, and the its diameter was measured with Image J. It was indicated that the particle size was 49. 38 ± 9. 92 μm ( Fig. 2D ). Chondrocytes identification The immunofluorescence staining and the alcian blue staining were used to identify the chondrocytes harvest from the rat knee joint. Based on the alcian blue staining, the proteoglycan in the cytoplasm of chondrocytes was stained blue, which revealed that the primary rat chondrocytes were fusiform ( Fig. 3B ). In addition, type II collagen in the cytoplasm of rat chondrocytes could be stained red through the immunofluorescence ( Fig. 3A ). As suggested from the results of the two staining methods, the cells extracted from the articular cartilage were chondrocytes. 10. 7717/peerj. 12188/fig-3 Figure 3 dECM enhance the anabolism of chondrocytes. (A) The rat chondrocytes of knee joint were stained with type II collagen immunofluorescence staining. (B) The rat cartilage of knee joint was stained with alcian blue. (C) The growth state of rat chondrocytes on dECM (original magnification ⊆ 500). (D) The growth state of rat chondrocytes on dECM (original magnification ⊆ 1, 000). The effect of dECM on rat chondrocytes viability, toxicity As indicated by the different magnification electron microscopy, the chondrocytes in the dECM of the knee joint showed the significant growth ( Figs. 3C and 3D ). According to Alamar Blue staining, the number of chondrocytes in the knee joint of rats in both groups was elevated logarithmically after 3–5 days, and the dECM group achieved more active chondrocyte growth after 7 days ( Fig. 4A ). 10. 7717/peerj. 12188/fig-4 Figure 4 dECM enhance the anabolism of chondrocyteser. (A) Alamar blue assay was performed to verify the viability of rat chondrocytes at 1 d, 3 d, 5 d, and 7 d after the dECM intervention ( n = 3 for each group). (B) The mRNA levels anabolic genes in rat chondrocytes analyzed by RT-PCR after dECM treatment for 24 h. ( n = 3 for each group, ** P < 0. 01, *** P < 0. 001). The expression levels cartilage related genes Whether dECM up-regulated the expressions of extracellular matrix related genes in rat chondrocytes was investigated. The RT-PCR technology was adopted to indicate that the expressions of cartilage related genes were up-regulated after the dECM intervention. This study observed that the levels of Col2A1, ACAN, PRG4 and Sox9 mRNA increased through the dECM treatment ( Fig. 4B ). Therapeutic effect of dECM-CS compound on KOA in rats All experimental rats survived after operation and there were no complications such as wound infection. As indicated from the result, the dECM-CS compound treatment significantly delayed the progression of KOA in rats. As suggested from the figure, the articular cartilage erosion and the reduced proteoglycan loss were notably inhibited, which was assessed by using HE, the safranin O-fast green staining green and the alcian blue staining at 4 w or 8 w ( Figs. 5A and 5B ). This result was further objectively confirmed based on the results of OARSI scores ( Figs. 6A and 6B ). In comparison with the ACLT control group, the dECM-CS compound treatment group had the significantly lower scores of ACLT rats. By performing the hot-plate test ( Figs. 6C and 6D ) and the weight-bearing test ( Figs. 6E and 6F ), it was obviously found that the local injection treatment of dECM-CS compound in ACLT rats could significantly alleviate OA-induced pain after 4 w or 8 w of the treatment. 10. 7717/peerj. 12188/fig-5 Figure 5 Histological evaluation after intra-articular injection treatment. (A and B) Representative pictures of HE, safranin O-fast green and alcian blue from rats in normal, sham, vehicle, dECM suspension, CS, and dECM-CS compound groups at 4 w and 8 w ( n = 6 for each group). 10. 7717/peerj. 12188/fig-6 Figure 6 Evaluation after intra-articular injection treatment. (A and B) OARSI scores from rats in the normal, the sham, the vehicle, and the dECM suspension, the CS, and the dECM-CS compound groups at 4 w and 8 w ( n = 6 for each group). (C and D) Pain response times of rats in the respective group exposed to thermal stimulus at 4 w and 8 w after the treatment ( n = 6 for each group). (E and F) Weight difference of bilateral lower limbs of rats in the respective group after 4 w and 8 w of treatment ( n = 6 for each group). All data were expressed as mean ± standard deviation. * P < 0. 05, ** P < 0. 01, *** P < 0. 001. dECM hepatorenal toxicity in rats The HE staining of rat liver and kidney cells at 4 w and 8 w after the drug intervention indicated that the glomerular structure of hepatic lobules was intact, which demonstrated that no hepatorenal toxicity was produced ( Fig. 7 ). 10. 7717/peerj. 12188/fig-7 Figure 7 Evaluation of hepatorenal toxicity of drugs for intra-articular injection. (A and B) Liver and kidney tissue from rats in normal, dECM suspension, CS, and dECM-CS compound groups at 4 w and 8 w. Discussion KOA refers to a common disease in orthopedics, which mostly occurs in the elderly and results in a heavy medical burden ( Hunter, Schofield & Callander, 2014 ). At present, no drug is capable of significantly delaying or even reversing the progression of KOA. In this study, the decellularized technique was successfully applied to the rat knee cartilage, and the dECM was suggested to exert a positive effect on the growth of rat chondrocytes in vitro. The dECM-CS compound significantly helped relieve the pain of KOA in rats, and delayded the progression of KOA in rats. In addition, the local application of dECM-CS compound was also confirmed to have no obvious hepatorenal toxicity in rats. The results here confirmed the good effect of dECM-CS compound on the treatment of rat knee OA. At present, the step-by-step treatment has been adopted for KOA ( Abramoff & Caldera, 2020 ). Under more insights into the pathogenesis of KOA ( Hunter & Bierma-Zeinstra, 2019 ) and the emergence of novel technologies ( e. g. , regenerative medicine and tissue engineering), various studies aiming at delaying the progression of KOA have emerged ( Wang et al. , 2017 ; Zhang et al. , 2017 ). The metabolic imbalance of cartilage tissue is a vital link in the pathogenesis of OA ( Rahmati et al. , 2017 ). When chondrocytes are injured by trauma, inflammation or other factors, the synthesis and metabolism of chondrocytes are destroyed, the ECM components of cartilage are reduced, and the homeostasis of cartilage tissue is unbalanced, which causes the cartilage destruction to be gradually aggravated. Existing studies confirmed that the interaction between ECM and chondrocytes affects the differentiation and growth of chondrocytes ( Gentili & Cancedda, 2009 ). The decrease of transcriptional level of Col2A1, ACAN, PRG4 and Sox9 are closely related to OA ( Nham et al. , 2019 ; Ikegawa et al. , 2000 ; Haag, Gebhard & Aigner, 2008 ). The dECM was used in this study to interfere with rat chondrocytes, and the expressions of the mentioned genes were suggested to be up-regulated. This result indirectly indicated the protective effect of dECM on chondrocytes. Decellularized technology acts as the premise of the application of ECM. It is capable of reducing the immunogenicity of the tissue, while retaining the composition, structure and function of the ECM ( Cramer & Badylak, 2020 ). At present, the bone and cartilage decellularized technology has been applied in the tissue engineering technology, and it has achieved good research results ( Kheir et al. , 2011 ). In this study, the improved method was adopted to remove cells from cartilage tissue to obtain dECM ( Zhang et al. , 2020 ). This method was continuously used to treat rat knee cartilage. The rat knee cartilage was evaluated by the HE staining, the alcian blue staining and the scanning electron microscope, and the content of GAG and DNA in rat knee cartilage tissue were quantitatively detected before and after the acellular by using the biochemical methods. Next, a satisfactory dECM of rat knee cartilage was successfully obtained. To determine the effect of allogeneic cartilage dECM on the viability of chondrocytes, the in vitro experiments were performed, in which rat cartilage dECM was co-cultured with rat chondrocytes to observe the viability of chondrocytes. It was reported that chondrocytes grew well on cartilage dECM. Moreover, alamar blue was used to detect the toxicity of dECM to chondrocytes. As reported from the results, dECM exhibited no cytotoxicity to chondrocytes, indicating that allogeneic cartilage dECM did not adversely affect chondrocyte viability, and it was feasible to be used as a tissue engineering material for treating rat KOA. In our previous study, we found that rabbit knee cartilage dECM can significantly delay the progression of KOA ( Zhang et al. , 2020 ). This study hopes to combine cartilage dECM with other tissue engineering materials in the study of KOA. CS is characterized by non-toxic, anti-inflammatory, good biocompatibility, slow degradation rate, etc. ( Comblain et al. , 2017 ). As a tissue engineering material, it has been maturely applied in the basic research of orthopaedics. An experimental study termed as Chitosan-cartilage extracellular matrix hybrid scaffold induces chondrogenic differentiation to adipose-derived stem cells showed that adipose stem cells differentiated into chondrocytes in CS-dECM composite three-dimensional scaffolds and built considerable cartilage extracellular matrixes ( Lin et al. , 2020 ). Thus far, there has been no report on the efficacy of CS and dECM for treating KOA. As reported from the HE staining, the alcian blue staining and the saffron fast green staining, dECM-CS compound could reduce the degree of cartilage surface fibrosis, protect chondrocytes, avoid the degradation of the extracellular matrix in the cartilage tissue, and significantly delay the progression of KOA in the rats. The OARSI pathological grade score of dECM-CS group also objectively reflected its protective effect on cartilage. Some of the most important growth factors like TGFb, FGF, and IGF is retained in dECM. And the cartilage tissue is naturally inclined to respond to the growth factors ( Vinatier et al. , 2009 ). Cartilage tissue naturally lacks a supply of appropriate growth factors and nutrients owing to its avascular nature, so the retention of bioactive molecules will be especially beneficial in regenerating cartilage ( Benders et al. , 2013 ). These bioactive molecules may be involved in the up-regulation of genes related to cartilage synthesis and metabolism. Type II collagen and GAGs play an important role in cartilage formation in vivo and in vitro. Biomaterials containing chitosan can provide a suitable biochemical and biomechanical environment for chondrocytes to produce type II collagen and GAGs ( Muzzarelli et al. , 2012 ). From the above, it can be inferred that dECM-CS complex plays a synergistic role in the repair of cartilage tissue. Increasing pain in knee joint can seriously limit the normal function of the knee joint, the clinical use of non-steroidal anti-inflammatory drugs still have adverse consequences on the body ( Pelletier et al. , 2016 ). The causes of knee joint pain include knee arthritis, destruction of cartilage, etc. The hot plate test and the weight-bearing asymmetry test were performed to objectively assess the knee joint pain in rat KOA model. According to this experiment, the pain in dECM-CS group in ACLT was significantly less than that in Vehicle group, which demonstrated that dECM-CS compound could be effective in relieving pain in rat KOA model. The drug safety has always been an important aspect of concern ( Cao et al. , 2020 ). The HE staining was performed on tissue sections of liver and kidney of rats treated for 4 w and 8 w. As revealed from the results, the drugs prepared and applied here were relatively safe. The limitation is that the details of the interaction between dECM-CS compound and KOA require further research, and we will continue to explore the biological properties of dECM-CS compound in depth. Conclusion As a novel type of composite material, dECM-CS compound has achieved good results for treating KOA in rats. it is a potential treatment of KOA and provides a new idea to treat KOA. Statistical methods All the data in this experiment were processed and analyzed by SPSS19. 0. Pairwise comparisons were performed by t -test, and comparisons between multiple groups were performed by one-way ANOVA. The data were expressed as mean ± standard deviation, P < 0. 05 shows that the results are statistically different. Supplemental Information 10. 7717/peerj. 12188/supp-1 Supplemental Information 1 Author checklist. Click here for additional data file. 10. 7717/peerj. 12188/supp-2 Supplemental Information 2 Raw data. Click here for additional data file.

10. 7717/peerj. 1277

PeerJ

Pore size is a critical parameter for obtaining sustained protein release from electrochemically synthesized mesoporous silicon microparticles

Mesoporous silicon has become a material of high interest for drug delivery due to its outstanding internal surface area and inherent biodegradability. We have previously reported the preparation of mesoporous silicon microparticles (MS-MPs) synthesized by an advantageous electrochemical method, and showed that due to their inner structure they can adsorb proteins in amounts exceeding the mass of the carrier itself. Protein release from these MS-MPs showed low burst effect and fast delivery kinetics with complete release in a few hours. In this work, we explored if tailoring the size of the inner pores of the particles would retard the protein release process. To address this hypothesis, three new MS-MPs prototypes were prepared by electrochemical synthesis, and the resulting carriers were characterized for morphology, particle size, and pore structure. All MS-MP prototypes had 90 µm mean particle size, but depending on the current density applied for synthesis, pore size changed between 5 and 13 nm. The model protein α -chymotrypsinogen was loaded into MS-MPs by adsorption and solvent evaporation. In the subsequent release experiments, no burst release of the protein was detected for any prototype. However, prototypes with larger pores (>10 nm) reached 100% release in 24–48 h, whereas prototypes with small mesopores (<6 nm) still retained most of their cargo after 96 h. MS-MPs with ∼6 nm pores were loaded with the osteogenic factor BMP7, and sustained release of this protein for up to two weeks was achieved. In conclusion, our results confirm that tailoring pore size can modify protein release from MS-MPs, and that prototypes with potential therapeutic utility for regional delivery of osteogenic factors can be prepared by convenient techniques.

Introduction Mesoporous silicon (MS)-based materials are currently investigated in a variety of systems for drug delivery and tissue engineering applications ( Anglin et al. , 2008 ; Santos, 2014 ). Their main advantage lies on their outstanding surface area arising from the fine mesoporous structure that allows remarkable drug loadings to be achieved just by plain adsorption ( Prestidge et al. , 2008 ). MS is also biocompatible ( Canham, 1995 ; Godin et al. , 2008 ; Salonen et al. , 2008 ), and degrades in the body to silicates (SiO 2 ) ( Canham, 1995 ; Salonen et al. , 2008 ; Pastor et al. , 2009 ) that are eliminated by renal excretion ( Popplewell et al. , 1998 ). Silicates have FDA GRAS status, and even safety margins for silica nanoparticles administered intravenously start to be established ( Yu et al. , 2013 ). Inspired by these properties, researchers have investigated silicon-based carriers in a variety of formats (i. e. , scaffolds, microparticles, nanoparticles, etc. ) for delivering hydrophobic and hydrophilic drugs ( Anglin et al. , 2008 ; Prestidge et al. , 2008 ; Salonen et al. , 2008 ). MS-based materials have also been proposed for delivering drug-loaded nanoparticles within the concept of multistage delivery vehicles ( Tasciotti et al. , 2008 ). Devices composed of a crystalline mesoporous silicon matrix are alternatives to silica mesoporous structures ( Kresge et al. , 1992 ), but unlike those, they do not require mesophase template removal for their preparation. Mesoporous silicon can be prepared by stain-etching or electrochemical anodizing of silicon. Both methods result in suitable mesoporous (nanostructured) materials, but the stain-etching method is less controlled with respect to pore homogeneity, and often leaves an untreated crystalline silicon core inside the particles. Medical materials prepared from stain-etched mesoporous silicon should be additionally checked for complete removal of toxic nitric oxide residues. The electrochemical method for MS production is therefore more medical-friendly, and recently its scalability has been considerably improved ( Makushok, Matveyeva & Pastor, 2012 ). The desired nanostructure of MS fabricated by electrochemical methods can be easily achieved by a simple tuning of the preparation conditions; first of all, the applied current density. Even though these inner nanostructure parameters (pore size, overall porosity, particle size, etc. ) are important for MS silicon drug carriers, they cannot assure by themselves optimal drug payloads. The interaction between the drug and the carrier surface needs also to be engineered, and thus the surface modification and functionalization of MS nanostructures has been extensively studied in recent years ( Jarvis, Barnes & Prestidge, 2011 ; Jarvis, Barnes & Prestidge, 2012 ; Barnes, Jarvis & Prestidge, 2013 ). Among different techniques, a simple oxidation is frequently performed that converts the outer surfaces of crystalline mesoporous silicon to a mesoporous silica replica ( Kresge et al. , 1992 ). In a previous publication from our group, MS microparticles (MS-MPs) with an average pore size of 35 nm were prepared by an electrochemical method and stabilized by thermal oxidation. These MS-MPs were successfully loaded by absorption equilibrium with two model proteins, insulin and bovine serum albumin BSA ( Pastor et al. , 2011 ). Although these proteins were released from a vehicle in a controlled manner, the process was fast (∼80–100% release in less than 2 h), and consequently only suitable for some applications such as mucosal drug delivery. Previous studies with hydrogels ( Peppas et al. , 2000 ), solid polymers ( Sandor et al. , 2001 ), and other mesoporous materials ( Santos, Radin & Ducheyne, 1999 ) have shown that modulation of the inner nanostructure of the carrier can change the kinetics of drug release. We proposed that similar principles should apply for controlling the release of proteins from electrochemically synthesized MS-MPs. To address this hypothesis, we prepared MS-MPs with different pore sizes and explored how changes in inner nanostructure can influence the release of loaded proteins. This study was performed initially with the model protein α -chymotrypsinogen (aCT); then, considering the bioactivity of MS materials for orthopedic regeneration ( Canham, Reeves & Newey, 1999 ; Pastor et al. , 2007 ; Sun et al. , 2007 ), we loaded a protein of therapeutic interest for this application, bone morphogenetic protein-7 (BMP7). Materials and Methods Materials Boron doped silicon with different resistivity, 0. 01–0. 02 and 10–20 Ω cm, was purchased from Si Materials (Germany); wafer diameter was 100. 0 ± 0. 5 mm and thickness of 525 ± 25 µm ( pI = 2–3. 5). Fluoric acid (HF) (48%) was purchased from Riedel de Haën (Germany) and ethanol (96%) from Panreac (Barcelona, Spain). Synthetic air ( N 2 with 21% of O 2 ) was provided from AbelloLinde S. A. (Barcelona, Spain). Avidin-peroxidase conjugate, α -chymotrypsinogen A (aCT) from bovine pancreas ( pI = 9. 5; Mw = 25. 7 kDa), and 2, 2′-azino-bis (3-ethylbenzthiazoline-6-sulfonic acid) were obtained from Sigma Aldrich (Madrid, Spain). Recombinant human Bone Morphogenetic Protein-7 (BMP7) ( pI = 8. 1; Mw = 28. 8 kDa), polyclonal antibody rabbit anti-human BMP7, and biotinylated polyclonal antibody rabbit anti-human BMP7 were purchased from PeproTech (London, UK). All other solvents and chemicals used were high-grade purity. Preparation of mesoporous silicon microparticles (MS-MPs) MS-MPs were obtained by an electrochemical method similar to that previously described by us ( Pastor et al. , 2009 ). The main difference was the use of a 1:1 HF:Ethanol electrolyte, and special cyclic regimes with etch-stops in order to improve the homogeneity of pore sizes distribution along with the in-depth etching ( Bychto et al. , 2008 ). A constant current step (40 or 60 mA/cm 2 for 5–10 s) was followed by an etch-stop step (no current applied for 2–5 s) in cyclic periods. After obtaining a MS layer of ∼150 µm thickness, the electrochemical process was stopped, and the Si wafer was washed thoroughly with distilled water, dried, and the porous material was scratched from the remaining Si substrate. The obtained MS was subjected to a thermal oxidation under a flow of synthetic air for 1 h at 500 or 650 °C (Programat P200 equipped with a vacuum pump VP3 and gas inlet; Ivoclar-Vivadent, Inc. , Amherst, New York, US). To reduce the particle size to the micrometer scale, the MS material was milled and sieved in cascade. The fraction between 75 and 100 µm was selected for further studies. Henceforth, this fraction is referred to as MS-MPs. The preparation conditions for the three different MS-MP prototypes studied in this work are summarized in Table 1. For example, prototype B was prepared from Si wafer of 0. 01–0. 02 Ω cm resistivity, under a current density of 40 mA/cm 2 applied for 10 s, and then interrupted by a 2 s interval of zero current (etch-stop). This regime was cyclically repeated for a few hours until the 150 µm porous layer was grown. After recollecting the porous material, the material was thermally oxidized at 650 for one hour. 10. 7717/peerj. 1277/table-1 Table 1 Preparation conditions for different mesoporous silicon prototypes synthesized by the electrochemical method under special cyclic regimes with etch-stop (zero current) applied after each anodizing interval. Three different prototypes (A–C) were prepared and tested in this study, differing in silicon waver resistivity, current densities, etch-stop times, and thermal oxidation temperatures. Prototype Si wafer resistivity (Ω cm) Current density (mA/cm 2 )/ anodizing time (s) Etch stop time (s) Oxidation temperature (°C) A 0. 01–0. 02 40/5 5 500 B 0. 01–0. 02 40/10 2 650 C 10–20 60/5 2 550 Characterization of MS-MPs The porosity of the porous silicon materials was determined gravimetrically by comparing the mass of the silicon wafer before and after anodizing as previously described ( Pastor et al. , 2011 ). Particle sizes were analyzed with a Mastersizer 2000 (Malvern Instruments, Malvern, Worcestershire, UK). MS-MPs morphology was visualized by high resolution Scanning Electron Microscopy (SEM; Hitachi S4500; Hitachi, Tokyo, Japan). Additionally, the Brunauer- Emmett-Teller (BET) surface area of the MS-MPs was determined by N 2 adsorption–desorption isotherms (Micrometrics ASAP 2020 V3. 04H; Micromeritics France S. A. , Verneuil-en-Halatte, France). Pore size was calculated from the same N 2 adsorption data, by the Barroett-Joyner-Halenda (BJH) method. Protein loading Protein loading was carried out by solvent evaporation ( Prestidge et al. , 2008 ). Briefly, 20 µL of the model protein aCT (3 mg/mL) or BMP7 (5 µg/mL) in aqueous solutions were added to a fixed amount of MS-MPs (1 mg). The samples were gently vortexed for 10 s, and then incubated under mild agitation at 37 °C until total evaporation of solvent was reached and all amounts of proteins incorporated into the MS-MPs (about 7 h). The theoretical protein loadings were: 60 µg/mg of MS-MPs for aCT, and 0. 1 µg/mg of MS-MPs for BMP7. Loaded MS-MPs were freeze-dried and stored at −20 °C until use. In vitro release studies Samples comprising 1 mg of MS-MPs loaded with aCT or BMP7 were incubated with 500 µL of PBS (USP 38-NF 33, pH 7. 4) under agitation (100 rpm, Titramax 1000; Heindolf, Schwabach, Germany) at 37 °C (Inkubator 1000, Heindolf, Schwabach, Germany). At scheduled time points, release samples were collected, and centrifuged at 7000 RCF for 10 min at 4 °C (Microfuge 22R; Beckman Coulter, Brea, California, USA). The amounts of aCT in supernatants were determined by the bicinchoninic acid method (Micro BCA protein Assay Kit; Pierce Biotechnology Inc. , Rockford, Illinois, USA), and those of BMP7 by ELISA, as previously reported by us ( Reguera-Nuñez et al. , 2014 ). Amounts of released protein are expressed as percentage of a total protein mass added at the loading stage since the whole mass was considered as absorbed upon solvent evaporation. Results and Discussion Characterization of different MS-MPs carriers Mesoporous silicon microparticles (MS-MPs) were prepared by electrochemical etching, thermal stabilization, and milling to reduce the particle sizes. The resulting powder was sorted by sieving. The particles of the selected fraction (i. e. , the MS-MPs) were irregular in shape, but homogeneous in size ( Fig. 1A ). All the MS-MPs prototypes generated showed a normal distribution of sizes with a mean value around 90 µm ( Fig. 1B ). This normal particle distribution contrasted with our previous data where the particle distribution was log-normal ( Pastor et al. , 2011 ); this might be related to the different particle fractions selected on each work (90 µm vs. 33 µm mean size, respectively). The mesoporous structure of MS-MPs observed by high resolution SEM ( Fig. 1C ) revealed the regular and homogeneous pores propagated along a single direction, as it is common for electrochemically prepared MS. The SEM analysis, however, might not reveal the smallest pores of the materials due their well-known resolution limits. 10. 7717/peerj. 1277/fig-1 Figure 1 Morphological and physicochemical properties of mesoporous silicon microparticles (MS-MPs). (A) SEM image of MS-MPs (bar is 200 µm); (B) particle size distribution of the different MS-MP prototypes measured with a particle size analyzer; (C) example of a SEM image of the surface of MS-MPs (corresponding to prototype A, bar is 800 nm); (D) N 2 adsorption isotherms, volume adsorbed vs. relative pressure ( P / P 0), for the different MS-MP prototypes. The inner structure for three different MS-MP prototypes (A–C) prepared under the conditions summarized in Table 1 was characterized by N 2 adsorption–desorption experiments ( Fig. 1D ). The data revealed very high specific surface areas for prototypes A and B (>200 m 2 /g), but even more for prototype C (350 m 2 /g) ( Table 2 ). The porosity of all samples was high (>50%), and the mean pore diameter was ∼12 nm for prototypes A and B, and ∼6 nm for prototype C. These pore sizes were significantly smaller than MS-MPs prepared in our previous work ( Pastor et al. , 2011 ), a result of the different preparation conditions. Due to their tighter internal structure, we expected that the MS-MPs obtained in this work would be more suitable for the sustained release of proteins. 10. 7717/peerj. 1277/table-2 Table 2 Characteristics of the different mesoporous silicon microparticle prototypes. Data represent means ± S. D. , n = 3. Prototype Specific surface (m 2 /g) Porosity (%) Pore diameter (nm) A 210. 2 ± 13 72 ± 6 11. 4 ± 0. 7 B 224. 9 ± 16 53 ± 8 12. 4 ± 3 C 350. 8 ± 21 60 ± 5 5. 8 ± 0. 4 Due to the limited number of prototypes studied and the important difference in parameters observed, it is difficult to draw unequivocal conclusions on the relationships between the MS-MPs preparation parameters ( Table 1 ) and the resulting carrier properties ( Table 2 ). Still, under the tested preparation conditions, there is a positive correlation between the current density and the specific surface area. Also, an inverse correlation between the applied current density and the mean pore diameter can be noted, although the doping level of Si wafer might play a dominant role in this correlation. Globally, the study confirms the possibility to prepare MS-MPs with controllable mesoporous inner structures by the electrochemical method. Protein loading in MS-MPs After characterization of the different MS-MP prototypes, we studied how these systems are capable of loading and releasing two proteins, aCT and BMP7. The zymogen aCT was selected as a model protein for screening studies since it has very similar physicochemical properties (pI and Mw) to BMP7 (see data on ‘Materials’), and we have previously observed good correlation between encapsulation of both proteins ( Reguera-Nuñez et al. , 2014 ). ACT is a zymogen physiologically activated by the gut’s endopeptidases, and does not activate under the conditions of the loading procedures and release tests applied in this work. For protein loading in this work we decided to work under forcing conditions, and we evaporated a protein solution in the presence of the MS-MPs at 37 °C. This method has the main advantage of forcing protein encapsulation, which can be assumed to be close to 100%. Because MS-MPs cannot be degraded without harming the loaded protein, we were unable to quantify the loaded proteins. However, from the final release point of our release studies (see ‘Pore size can control the release of a model protein (aCT) from MS-MPs’ and ‘MS-MPs can achieve a 2-week sustained release of antigenically active BMP7’), we can conclude that >75% of aCT was loaded in all preparations, and >60% of BMP7. When using this loading method, the mechanisms that drive protein loading would be capillary forces and adsorption from a continually concentrating solution ( Karlsson et al. , 2003 ). Other possible mechanisms would be electrostatic interactions; after thermal oxidation the MS-MPs surface bears a negative charge as the silicon oxides cover the entire porous network ( Zangooie, Bjorklund & Arwin, 1998 ). This might affect the loading and release of cationic proteins such as aCT and BMP7. Under the tested conditions, the final protein payloads per mg of the carrier were 60 µg for aCT and 0. 1 µg for BMP7. Pore size can control the release of a model protein (aCT) from MS-MPs The release of loaded aCT from the three MS-MPs prototypes was analyzed in vitro (PBS, 37 °C). No burst release was observed for any of the tested prototypes, suggesting that most protein is inside the pores and not adsorbed on the outer MS-MP surface ( Fig. 2A ). This behavior is in agreement with our previous study on insulin and BSA, where despite of a faster release (<2 h), only a moderate burst effect was observed (∼30%) ( Pastor et al. , 2011 ). In the present work, the burst effect was drastically reduced, presumably because of the lower pore size of the carriers, and because of the different procedures for protein loading (solvent evaporation vs. adsorption equilibrium). 10. 7717/peerj. 1277/fig-2 Figure 2 In vitro release profile of (A) α -chymotrypsinogen and (B) BMP7 from MS-MPs prepared by the electrochemical method. Data represent means ± S. D. , n = 3. The MS-MPs investigated in this work were able to control protein release for longer periods of time than the carriers previously reported by us ( Pastor et al. , 2011 ): for prototypes A and B a ∼100% release was achieved in 30–40 h after incubation at 37 °C in PBS. Prototype C showed even more sustained kinetics, with high retention of aCT after 96 h ( Fig. 2A ). However, after 2-weeks, sample C had released 77. 2% ± 4. 2 ( n = 3) of the loaded aCT. The slower release should be associated with the nanostructure of the carriers, mainly to their pore size. Mean pore size was <15 nm for all prototypes studied here, and 33 nm in our previous work. Prototype C possesses pores with a mean size of ∼6 nm, half of those of prototypes A and B, and similar to the radius of gyration of aCT, 1. 76 nm ( Perkins et al. , 1993 ). As observed in other systems ( Santos, Radin & Ducheyne, 1999 ; Peppas et al. , 2000 ; Sandor et al. , 2001 ), when the drug’s radius of gyration is about the size of pores in the matrix, diffusion might be hindered, and more sustained release kinetics achieved. When comparing the different prototypes studied in this work, particle inner structure seems to be the critical factor modulating different release kinetics. When comparing the performance of the MS-MP prototypes from this work with those of our previous work ( Pastor et al. , 2011 ), two additional factors need to be considered. First, the effect of the chemical differences of the proteins tested. Insulin and BSA, used before, both bear negative charges in PBS, and therefore, their attachment to MS-MPs surfaced by adsoption shoud be driven mostly by hydrophobic interactions. On the other hand, aCT is positive in PBS, and therefore, ionic interactions with the silicon oxide on the surface of MS-MPs can be important to explain protein adsorption/desorption. Another parameter that could have some limited influence on protein release is the average particle dimensions, which was 33 µm in our previous work, and is 90 µm here ( Pastor et al. , 2011 ). Particle dimension will influence the diffusion length within the carrier for the protein. Recently, a new production method yielding planar mesoporous silicon microparticles with a controlled thicknesses, porosity and pore sizes has been reported ( Makushok, Matveyeva & Pastor, 2012 ). These new kind of materials might be interesting for release mechanism studies, since their lateral dimensions, perpendicular to the pore axis, will play no important role in the release process. MS-MPs can achieve a 2-week sustained release of antigenically active BMP7 Based on promising data obtained with aCT protein, we tested MS-MP prototype C for the controlled release of a therapeutic protein: BMP7. This protein is approved by FDA and other regulatory agencies for orthopedic applications (OP-1 Putty and OP-1 Implant; Stryker, Kalamazoo, Michigan, US), and it is delivered through a collagen sponge with limited controlled release properties. This limited controlled release has been linked to most of the treatment undesirable effects ( Lane, 2001 ). MS-MPs were loaded with BMP7 as described in ‘Protein loading in MS-MPs’, and the release kinetics of the protein was analyzed. Consistent with the data obtained with aCT, a 2-week sustained release was achieved ( Fig. 2B ). Once again, the release kinetics was characterized by low burst (<10%), and by a sustained release profile for at least 14 days. Maximum release observed over the experiment (28 days) was ∼70%. Significantly, the quantification of BMP7 in the supernatant was performed by ELISA, and thus it guarantees the presence of the protein in its antigenically-active form upon release. While antigenic activity is not a final proof of biological activity, previous studies from our group using the same ELISA kit have found a relation between antigenic BMP7 and bioactive protein in a glioblastoma cancer stem cell model ( Reguera-Nuñez et al. , 2014 ). The release profile was fitted to zero-order, first-order, Higuchi and to the Kosmeyer–Peppas models (Wizard—Statistics, Visualization, Data Analysis, Predictive Modeling, version 1. 4; Evan Miller©, US). Fitting to the first-order and Higuchi models was adequate ( p < 0. 008 and p < 0. 002, respectively), but the best fit was achieved with the Kosmeyer–Peppas model (BMP released% =10. 4 ⋅ t (days) 0. 64, p < 0. 001). The Kosmeyer–Peppas model is effective to describe release systems where release kinetics might depend on several factors. The diffusional exponent ( n = 0. 64) indicates a process of anomalous diffusion ( Korsmeyer et al. , 1983 ; Peppas, 1985 ). The similarities between aCT and BMP7 release kinetics reflect their similar physicochemical properties. Indeed, BMP7 has a radius of gyration ∼3. 5 nm (by analogy with other BMPs, ( Berry et al. , 2006 )) just slightly larger than aCT. It has also a basic isoelectric point (8. 1) close to that of aCT (9. 5). These similarities result in consistent profiles for both proteins, and suggest the robustness of the delivery technology. In summary, we have achieved sustained release of BMP7 for at least two weeks by using electrochemically synthesized MS-MPs. A preparation technology for the whole therapeutic system is convenient, since both components, protein solution and pre-formed empty MS-MPs, can be integrated together in an extemporaneous process. Due to the recently reported osteointegration properties of the MS-MP carrier itself ( Sun et al. , 2007 ), one of the immediate promising applications of this system would be in the bone regeneration area. Conclusions Mesoporous silicon microparticles with controlled inner structure (pore size) can be prepared by an electrochemical method, and loaded with proteins by simple adsorption and solvent evaporation. Under optimized electrochemical conditions, these microparticles present a nanostructure with pore sizes below 10 nm, and this small pore size is critical to provide sustained protein release over several days. The medical potential of the electrochemically synthesized mesoporous silicon microparticles is suggested by the two weeks sustained release profile of the osteogenic factor BMP7. Supplemental Information 10. 7717/peerj. 1277/supp-1 Figure S1 Mesoporous silicon Microparticles micrograph ( Fig. 1A ) Original scanning electron microscopy image of the mesoporous silicon microparticles ( Fig. 1A in the publication). Click here for additional data file. 10. 7717/peerj. 1277/supp-2 Figure S2 Surface of mesopous silicon microparticles ( Fig. 1C ) Original scanning electron microscopy image of the surface of mesoporous silicon microparticles ( Fig. 1C in the publication). Click here for additional data file. 10. 7717/peerj. 1277/supp-3 Supplemental Information 1 Distributive data for particle size analysis Excel data sheet showing the particle class interval (in µm) and the total volume of all the particles as calculated after analysis with ImageJ. The data was used to generate Fig. 1B. The distributions are calculated for Samples A, B and C. Click here for additional data file. 10. 7717/peerj. 1277/supp-4 Supplemental Information 2 Nitrogen adsorption isotherms Excel data sheet. N2 adsorbed volume (cm 3 /g) vs. the pressure/saturation pressure ratio for Samples A, B, C. This data was used to prepare Fig. 1D and the calculations for Table 1. Click here for additional data file. 10. 7717/peerj. 1277/supp-5 Supplemental Information 3 chymotripsinogen release data Excel data sheet. Percentage of chymotripsinogen released over time at 37 °C in PBS buffer for Samples A, B and C (60 µg per mg of microparticle was considered 100%). Data points for each replicate. Time expressed in days. Click here for additional data file. 10. 7717/peerj. 1277/supp-6 Supplemental Information 4 BMP7 release data Excel data sheet. Percentage of BMP7 released over time at 37 °C in PBS buffer for Sample C (0. 1 µg per mg of microparticle was considered 100%). Data points for each replicate. Time was expressed in days. Click here for additional data file.

10. 7717/peerj. 12819

PeerJ

Characterization of transcriptional landscape in bone marrow-derived mesenchymal stromal cells treated with aspirin by RNA-seq

Introduction Aspirin is a common antipyretic, analgesic, and anti-inflammatory drug, which has been reported to extend life in animal models and application in the treatment of aging-related diseases. However, it remains unclear about the effects of aspirin on bone marrow-derived mesenchymal stromal cells (BM-MSCs). Here, we aimed to analyze the influence of aspirin on senescence and young BM-MSCs. Methods BM-MSCs were serially passaged to construct a replicative senescence model. SA-β-gal staining, PCR, western blot, and RNA-sequencing were performed on BM-MSCs with or without aspirin treatment, to examine aspirin’s impact on bone marrow-derived mesenchymal stem cells. Results SA-β-gal staining, PCR, and western blot revealed that aspirin could alleviate the cellular expression of senescence-related indicators of BM-MSCs, including a decrease of SA-β-gal-positive cells and staining intensity, and downregulation of p16, p21, and p53 expression after aspirin treatment. RNA-sequencing results shown in the biological processes related to aging, aspirin could influence cellular immune response and lipid metabolism. Conclusion The efficacy of aspirin for retarding senescence of BM-MSCs was demonstrated. Our study indicated that the mechanisms of this delay might involve influencing immune response and lipid metabolism.

Introduction Tissue defects and functional impairment can be compensated by the coordination between differentiation and proliferation of specific stem cells or progenitors. Bone marrow-derived mesenchymal stromal cells (BM-MSCs) generated from mesoderm, one of the most well-characterized stem cells, possess self-renewal ability and multi-directional differentiation potential. They can differentiate into multiple lineages such as osteocytes, adipocytes, chondrocytes, neural cells, myocytes, and stromal cells supporting hematopoiesis ( Pittenger et al. , 1999 ). Moreover, BM-MSCs possess immunosuppressive activity ( Petrie Aronin & Tuan, 2010 ), which may benefit immunomodulation and tissue repair. Pioneering studies have shown that BM-MSCs can generate therapeutic results in bone defect ( Warnke et al. , 2004 ), skin regeneration ( Yang et al. , 2011 ), autoimmune diseases ( Parekkadan, Tilles & Yarmush, 2008 ), and cardiovascular disease ( Forrester, Price & Makkar, 2003 ). However, what is limiting is infrequent BM-MSCs obtained from the recipient’s bone marrow and restricted expansion capability accompanied by replicative senescence ( Ren et al. , 2013 ; Wagner, Ho & Zenke, 2010 ). This replicative senescence of BM-MSCs not only causes the cells to lose normal morphology but the functional decline also contributes to impairments in proliferation, differentiation, and stemness levels ( Galderisi et al. , 2009 ; Sethe, Scutt & Stolzing, 2006 ; Zheng et al. , 2016 ). For clinical therapeutic applications, overcoming cellular replicative senescence of BM-MSCs remain to be delineated. Aspirin is one of the world’s most widely used drugs. It is an oral antithrombotic agent for cardiovascular disease and in the treatment of chronic rheumatic diseases based on its anti-inflammatory, and anti-rheumatic activities. Nevertheless, the potential and real benefits of aspirin therapy go beyond its regular usage, including prevention and treatment of cancer ( Patrignani & Patrono, 2018 ), reduced incidence of postmenopausal osteoporosis ( Ma, Liu & Sun, 2019 ), and, in animal models, extended lifespan, an effect seen in Caenorhabditis elegans ( Wan et al. , 2013 ), Drosophila melanogaster ( Danilov et al. , 2015 ), and Mus musculus ( Strong et al. , 2008 ). This could happen through several mechanisms; for example, aspirin interferes with cytokine response processes, oxidant production, and blocking glycoxidation reactions that can improve lifespan in long-lived species by antioxidant therapies ( Phillips & Leeuwenburgh, 2004 ), and inhibition of cyclooxygenase I (COX I) and cyclooxygenase II (COX II), which results in downregulation of prostaglandins to reduce the inflammatory response ( Weissmann, 1991 ). To date, no research has concentrated on whether aspirin can delay the senescence of BM-MSCs by counteracting cellular inflammatory responses. In this present study, we aimed to analyze aspirin’s effects on early passages (EP) BM-MSCs and late passages (LP) BM-MSCs. Materials and Methods Cell culture and treatment of bone marrow-derived mesenchymal stromal cells Commercial Sprague Dawley rat BM-MSCs (RASMX-01001; Cyagen Biosciences, Guangzhou, China) were used for all the experiments. BM-MSCs were cultured at 37 °C and 5% CO 2 in SD Rat Mesenchymal Stem Cell Growth Medium (RASMX-90011; Cyagen Biosciences, Guangzhou, China), which contained basal medium, 10% fetal bovine serum, 1% penicillin-streptomycin, and 1% glutamine. Passage 3~5 (young) (EP) and passage 15~20 (aging) (LP) BM-MSCs were cultured until confluence reached 70~90%. EP and LP cells were treated with 400 μmol/L aspirin (A2093; Sigma-Aldrich, St. Louis, MO, USA) for 48 h. This dosage application used here was selected from our previous studies showing that it did not affect the apoptosis and proliferation of BM-MSCs ( Zhan et al. , 2018 ). Untreated cells were used as controls. Senescence-associated β-galactosidase (SA-β-gal) staining Cellular senescence was determined by SA-β-gal staining (C0602; Beyotime Biotech. , Jiangsu, China). Briefly, cells were seeded in a 6-well plate (Thermo, Waltham, MA, USA) at a density of 2 × 10 4 /well. After 24 h, EP and LP BM-MSCs were treated with 400 μmol/L aspirin for 48 h, respectively. Subsequently, cells were fixed with 4% paraformaldehyde for 15 min at room temperature and washed three times with PBS. PBS was replaced with SA-β-gal staining solution, the plate sealed with plastic wrap at 37 °C, and incubated overnight. The following day, cells were washed with PBS three times and then observed by light microscope in five random fields of view. The experiment was repeated three times. The percentage of SA-β-gal positive cells was calculated by counting the number of blue cells at least 100 cells per sample. Quantitative real-time PCR verification of senescence-associated gene With the intent to explore aging-related gene expression, real-time PCR was performed. Total RNA was extracted from cells using TRIzol (TaKaRa, Kusatsu, Japan) in terms of the manufacturer’s protocol. The extracted RNA was reverse transcripted into cDNA with the PrimeScript RT reagent kit (TaKaRa, Kusatsu, Japan). Subsequently, quantitative real-time PCR experiments were performed using the MxPro-Mx3000P real-time PCR System (Stratagene, La Jolla, CA, USA) based on SYBR Green I fluorescence. The reaction was performed in 20 μl using SYBR Premix Ex TaqTM kit (TaKaRa, Kusatsu, Japan) using the manufacturer’s protocol. The condition was set for 2 min at 95 °C, followed by 40 cycles of 95 °C for 15 s, 60 °C for 30 s, 72 °C for 30 s, accompanied with 95 °C for 10 s, 60 °C for 5 s and 95 °C for 5 s. For real-time PCR the sequences of the following primers were used: p53: 5′-TGA CTT TAG GGC TTG TTA TGA GAG–3′ (Forward), 5′-CAG CAG AGA CCC AGC AAC TAC–3′ (Reverse); p16: 5′-GAT GGG CAA CGT CAA AGT GG–3′ (Forward), 5′-TAC CGC AAA TAC CGC ACG AC–3′ (Reverse); p21: 5′-GGG AGG GCT TTC TTT GTG TA–3′ (Forward), 5′-GCA TCG TCA ACA CCC TGT CT–3′ (Reverse); GAPDH: 5′-GACAACTCCCTCAAGATTGTCAG–3′ (Forward), 5′-ATG GCA TGG ACT GTG GTC ATG AG–3′ (Reverse). Relative gene expression was calculated by the 2 −ΔΔ Ct method, and the level relative to GAPDH was normalized. Western blot Cells were collected and protein lysates were prepared with RIPA lysis buffer (Beyotime, Shanghai, China). Protein concentration was quantified with BCA Protein Assay Kit (Beyotime, Shanghai, China). Forty micrograms (μg) of protein lysates were split into 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to polyvinylidene difluoride (PVDF) membranes (Beyotime, Shanghai, China). Followed by blocking in 5% non-fat milk for 1 h at room temperature, the immunoblots were examined overnight 4 °C with the under antibodies: anti-p21 (Cat. No. ab109199, 1:1, 000; Abcam, Cambridge, UK), anti-p53 (Cat No. 10442-1-AP, 1:1, 000; Proteintech, Rosemont, IL, USA) and anti-β-actin (Cat. No. 8457S, 1:1, 000; Cell Signaling Technology, Inc. , Danvers, MA, USA). For secondary antibodies, peroxidase-conjugated IgG (#BL003A, 1:5, 000, Biosharp, Hefei, China) was used, followed by chemiluminescence detection with an enhanced chemiluminescence kit (Biosharp, Hefei, China). RNA sequencing Total RNA extraction was done from BM-MSCs after different treatments with TRIzol reagent. Subsequently, the RNA samples were sent to BGI Co. , LTD (Shenzhen, China), and BGISEQ-500 platform was applied to perform RNA-sequencing. Differentially expressed genes were determined based on Q value (Adjusted P -value) <= 0. 05 ( Love, Huber & Anders, 2014 ). Analysis of functional and pathway enrichment Biological function Gene Ontology (GO), Kyoto Gene and Genome Encyclopedia (KEGG) signal pathway, and gene set enrichment analysis (GSEA) enrichment analysis were performed in the Dr. Tom online software. In the GSEA analysis, normalized enrichment score (NES) ≥ 1 or ≤−1, nominal P -value (NOM P -val) <0. 05, and false discovery rate (FDR q-val) <0. 25 were considered as the cutoff values. Gene set variation analysis With the intent to explore pathway activity in the 50 hallmark pathways, we utilized gene set variation analysis (GSVA) analysis. GSVA analysis was capable of identifying pathways with significant differences in diverse treatment groups, which were more biologically far-reaching than genetic analysis ( Hänzelmann, Castelo & Guinney, 2013 ). Statistical analysis Throughout the text, we performed an unpaired Student’s t-test, and a P value < 0. 05 was examined statistically significant. Each experiment was repeated at least three times. GraphPad Prism 7. 0 was utilized for statistical analysis. Results Aspirin treatment reduces cellular senescence in bone marrow-derived mesenchymal stromal cells We began by performing SA-β-gal staining in BM-MSCs with or without aspirin treatment at early and late passages. To explore the potential relationship between aspirin treatment and senescence, we examined the SA-β-gal staining, which was optimally active at a pH of 6. 0 ( Figs. 1A, 1B ). The percentage of SA-β-gal-positive cells in EP BM-MSCs populations was initially low. In contrast, almost all LP BM-MSCs were SA-β-gal positive ( P < 0. 001), and the perinucleus was darkly blue stained. Cell replicative senescence produced by consecutive passages treated with aspirin produced a decrease in the ratio of the percentage of SA-β-gal-positive cells both in EP BM-MSCs and LP BM-MSCs ( P < 0. 05), and the perinucleus was more lightly blue-stained than without aspirin treatment. Further, to verify aspirin itself does not affect the SA-β-gal stain, we performed negative control experiments applied passages 1 BM-MSCs ( Fig. S1 ). 10. 7717/peerj. 12819/fig-1 Figure 1 Aspirin treatment reduces replicative senescence of BM-MSCs. (A) Senescence-associated β-galactosidase (SA-β-gal) staining of EP and LP BM-MSCs with or without aspirin treatment (400 μmol/L). Scale bar = 100 μm. (B) Percentages of cells that showed positive activity of SA-β-gal after each treatment scheme. At least 100 cells were examined in each of the three replicate experiments for different groups. * P < 0. 05, ** P < 0. 01, *** P < 0. 001. (C) Gene expression of p16, p21, and p53 in different groups. * P < 0. 05, ** P < 0. 01, **** P < 0. 0001, NS: no significance, with comparisons indicated by lines. (D) Western blot of p21 and p53 in EP and LP BM-MSCs with or without aspirin exposure, GAPDH served as an internal control. PCR analysis revealed that gene expression of senescence-associated molecules in LP BM-MSCs was down-regulated when cells were treated with aspirin ( Fig. 1C ). Consistent with the PCR results, western blot analysis of p21 and p53 expression levels were also reduced in cells treated with aspirin ( Fig. 1D ). Thus, these results indicated that aspirin can reduce cell senescence in LP BM-MSCs. Sequencing data of bone marrow-derived mesenchymal stromal cells We constructed 12 cDNA libraries from four groups (samples: p5_1, p5_2, p5_3 represented for EP BM-MSCs; samples: p5_ASA_1, p5_ASA_2, p5_ASA_3 represented for EP BM-MSCs treated with aspirin; samples: p20_1, p20_2, p20_3 represented for LP BM-MSCs; samples: p20_ASA_1, p20_ASA_2, p20_ASA_3 represented for LP BM-MSCs treated with aspirin) for RNA sequencing. The main sequencing characteristics of cell specimens collected here are annotated in Table 1. 10. 7717/peerj. 12819/table-1 Table 1 RNA sequencing reads filtering and reference genome alignment. Sample p5_1 p5_2 p5_3 p5_ASA_1 p5_ASA_2 p5_ASA_3 p20_1 p20_2 p20_3 p20_ASA_1 p20_ASA_2 p20_ASA_3 Total Raw Reads (M) 75. 15 74. 86 74. 87 75. 15 74. 86 74. 87 74. 53 74. 9 74. 74 71. 13 74. 65 74. 53 Total Clean Reads (M) 69. 1 69. 1 69. 16 69. 1 69. 1 69. 16 68. 38 69. 13 68. 84 65. 02 68. 58 68. 38 Total Clean Bases (Gb) 6. 91 6. 91 6. 92 6. 91 6. 91 6. 92 6. 84 6. 91 6. 88 6. 5 6. 86 6. 84 Clean Reads Q20 (%) 97. 88 97. 8 97. 83 97. 88 97. 8 97. 83 97. 89 97. 84 97. 88 97. 84 97. 89 97. 89 Clean Reads Q30 (%) 92. 28 91. 89 92. 12 92. 28 91. 89 92. 12 92. 49 92. 15 92. 47 92. 16 92. 56 92. 49 Clean Reads Ratio (%) 91. 95 92. 3 92. 38 91. 95 92. 3 92. 38 91. 75 92. 3 92. 11 91. 4 91. 86 91. 75 Total Mapping (%) 95. 39 95. 56 95. 46 95. 19 95. 28 95. 27 95. 04 95. 21 95. 25 95. 32 95. 3 94. 98 Uniquely Mapping (%) 88. 2 88. 39 88. 22 88. 16 88. 14 88. 15 87. 56 87. 79 87. 77 87. 92 87. 73 87. 63 Filtering out low-quality reads, linker contamination, and reads with too high N content in unknown bases, more than 65 million clean reads were obtained in each library. Subsequently, Q30 ratio (a base quality >30 and error rate <0. 001) among these clean reads was more than 91. 89%. Ultimately, 94. 98–95. 56% clean reads were mapped to Rattus_norvegicus reference genome (GCF_000001895. 5_Rnor_6. 0). Identification of DEGs After applying Bowtie2 to align clean reads to the reference gene sequence, and then using RSEM to calculate the gene expression level of each sample, we obtained differentially expressed genes (DEGs) between diverse groups via DESeq2 analysis (Q value <= 0. 05) ( Tables S1 – S3 ). Based on this screening criterion, LP BM-MSCs compared with EP, 9, 557 genes were found to be differentially expressed between these two groups ( Fig. 2A ). Among them, 4, 632 DEGs were up-regulated in LP BM-MSCs and 4, 925 DEGs were down-regulated. Further, we compared EP BM-MSCs before and after aspirin exposure, including 1, 245 DEGs, of which 569 genes were up-regulated and 676 DEGs were down-regulated after aspirin exposure ( Fig. 2C ). Compared with LP BM-MSCs with or without aspirin treatment, including 458 DEGs, of which 228 DEGs were up-regulated and 230 DEGs were down-regulated ( Fig. 2E ). Additionally, clustered heatmaps of the top 20 up-regulated and down-regulated DEGs expression were shown in Figs. 2B, 2D, 2F based on z-scores normalization, respectively. 10. 7717/peerj. 12819/fig-2 Figure 2 Identification of differentially expressed genes. (A) Volcano plot of EP and LP BM-MSCs. (B) Clustered heatmap of top 40 (20 up-regulated and 20 down-regulated) DEGs between the EP and LP BM-MSCs with gene annotations. (C) Volcano plot of EP BM-MSCs before and after aspirin exposure. (D) Clustered heatmap of top 40 (20 up-regulated and 20 down-regulated) DEGs between EP BM-MSCs before and after aspirin exposure with gene annotations. (E) Volcano plot of LP BM-MSCs before and after aspirin exposure. (F) Clustered heatmap of top 40 (20 up-regulated and 20 down-regulated) DEGs between LP BM-MSCs before and after aspirin exposure with gene annotations. The gradual color change from red to blue implies gene expression value from high to low; DEGs, differentially expressed genes. Functional annotation and pathway enrichment analysis of DEGs To get a sense of biological processes and pathways enriched in BM-MSCs treated with aspirin, we performed GO and KEGG pathway analysis. GO enrichment analysis results included three parts, biological processes (BP), cellular component (CC), and molecular function (MF). Detailed results of GO and KEGG enrichment analysis were summarized in Tables S4 – S9. For GO analysis, the rich ratio (Rich Ratio = Term Candidate Gene Num/Term Gene Num) was plotted on the X-axis, while the GO term was plotted on the Y-axis. According to the Q value, we demonstrated the top 20 terms from small to large. Biological processes analysis of LP BM-MSCs compared with EP BM-MSCs revealed DEGs in phosphorylation, biological process, protein phosphorylation, cellular response to DNA damage stimulus, cell migration, regulation of transcription, cell cycle, apoptosis, and others ( Fig. 3A ). Consistently, KEGG enrichment analysis showed similar changes in cell cycle and DNA replication after BM-MSCs reached replicative senescence ( Fig. 3B ). Also, we observed enrichment in focal adhesion, protein processing, cellular senescence, PI3K-Akt signaling pathway, MAPK signaling pathway, p53 signaling pathway, insulin signaling pathway, and others involved in cell senescence. We next explored the biological processes that were strongly associated with EP BM-MSCs when treated with aspirin. We found cell adhesion, cell proliferation, cell migration, wound healing, response to drug, regulation of apoptotic process, protein transport and others displayed significant changes ( Fig. 3C ). For KEGG enrichment included focal adhesion, ECM-receptor interaction, PI3K-Akt signaling pathway, lysosome, cell cycle, and others ( Fig. 3D ). Inspecting LP BM-MSCs treated with aspirin through GO analysis revealed response to calcium ion, extracellular matrix organization, response to estradiol, cell adhesion, regulation of cell proliferation, and others changes enriched in this group ( Fig. 3E ). Focal adhesion, regulation of actin cytoskeleton, PI3K-Akt signaling pathway, adherens junction, and other KEGG pathways were enriched in LP BM-MSCs treated with aspirin ( Fig. 3F ). Together, these analyses indicated after serially passaged replicative senescence, DEGs between the EP BM-MSCs and LP BM-MSCs were highly enriched in the cell cycle, DNA damage stimulus, and cellular senescence ( Figs. 3A – 3B ). After aspirin treatment, DEGs within the EP BM-MSCs and LP BM-MSCs were enriched in the focal adhesion, cell adhesion, and adherens junction pathways ( Figs. 3C – 3F ). 10. 7717/peerj. 12819/fig-3 Figure 3 Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. (A, C, E) GO enrichment analysis in different groups. The colors of circle dots illustrate the Q-values identified for each GO term (low: red, high: blue), with lower values for more significant enrichment. The size of dots indicates the number of the differentially expressed genes and the larger dots represent a larger gene number. ‘Rich Ratio’ represents the ratio of Term Candidate Gene Number to Term Gene Number. The greater the value of ‘Rich Ratio’, the more significant the enrichment. (A) GO biological process enrichment analysis of DEGs between EP and LP BM-MSCs. (C) GO biological process enrichment analysis of DEGs in EP BM-MSCs treated with aspirin. (E) GO biological process enrichment analysis of DEGs in LP BM-MSCs treated with aspirin. (B, D, F) KEGG enrichment analysis in different groups. The blue bar plots indicate −log10 Q value). The greater the value of −log10 Q value), the more significant the enrichment. The orange dots indicate the number of genes enriched to the term. (B) KEGG pathway enrichment analysis of DEGs between EP and LP BM-MSCs. (D) KEGG pathway enrichment analysis of DEGs in EP BM-MSCs treated with aspirin. (F) KEGG pathway enrichment analysis of DEGs in LP BM-MSCs treated with aspirin. GSEA and GSVA revealed biological functions of DEGs in bone marrow-derived mesenchymal stromal cells treated with aspirin Since GO and KEGG enrichment analysis often focuses on comparing DEGs between different treatment groups, it is easy to miss some not significant yet biologically meaningful DEGs. Here, we utilized gene set enrichment analysis (GSEA), which concentrates on interpreting gene expression data by analyzing shared common biological function gene sets ( Subramanian et al. , 2005 ). GSEA of biological process and KEGG pathways detected up-regulated cholesterol biosynthetic process, sterol biosynthetic process, and steroid biosynthesis pathways in LP BM-MSCs ( Figs. 4A – 4C ). GSEA of EP BM-MSCs treated with aspirin revealed that the addition of aspirin could down-regulated aging-related pathways such as longevity regulating pathway. After exposure to aspirin, we documented negative regulation of interleukin-6 production while regulation of lipolysis in adipocytes pathway was up-regulated ( Figs. 5A – 5C ). Besides, GSEA implicated enrichment of several inflammation pathways and lipid-related processes in LP BM-MSCs treated with aspirin including regulation of lipid metabolic process, linoleic acid metabolism, and alpha-Linolenic acid metabolism ( Figs. 6A – 6C ). To further verify the enrichment biological process, we profiled gene set variation analysis (GSVA) based on 50 hallmark gene sets pathways in different treatment groups. GSVA further implicated fatty acid metabolism, reactive oxygen species pathway, IL2 STAT5 signaling, and cholesterol homeostasis showed high score in LP BM-MSCs ( Fig. 7A ). After exposure to aspirin, the score of lipid metabolism synthesis and inflammation also decreased to varying degrees in EP and LP bone marrow-derived mesenchymal stem cells ( Figs. 7B – 7C ). 10. 7717/peerj. 12819/fig-4 Figure 4 The enrichment of DEGs between EP BM-MSCs and LP BM-MSCs were analyzed via gene set enrichment analysis (GSEA). Mainly including cholesterol ‘biosynthetic process’ term (A), ‘sterol biosynthetic process’ term (B), and ‘steroid biosynthesis’ pathway (C). The green broken line represents the change cure of the enrichment score (ES) of the group of genes, and the Y axis is the ES value. Each black vertical line represents one gene, and this part shows all genes under this pathway or GO term. The heat map shows the signal2 noise/log2 ratio value of all expressed genes. Red means value > 0, blue means value < 0, the darker the color, the greater the absolute value. 10. 7717/peerj. 12819/fig-5 Figure 5 The enrichment of DEGs in EP bone marrow-derived mesenchymal stem cells treated with aspirin was analyzed via gene set enrichment analysis (GSEA). Mainly including ‘negative regulation of interleukin-6 production’ term (A), ‘Longevity regulating pathway’ (B), and ‘regulation of lipolysis in adipocytes’ pathway (C). The green broken line represents the change cure of the enrichment score (ES) of the group of genes, and the Y axis is the ES value. Each black vertical line represents one gene, and this part shows all genes under this pathway or GO term. The heat map shows the signal2 noise/log2 ratio value of all expressed genes. Red means value > 0, blue means value < 0, the darker the color, the greater the absolute value. 10. 7717/peerj. 12819/fig-6 Figure 6 The enrichment of DEGs in LP BM-MSCs treated with aspirin was analyzed via gene set enrichment analysis (GSEA). Mainly including the ‘regulation of lipid metabolic process’ term (A), ‘alpha-Linolenic acid metabolism’ pathway (B), and ‘Linoleic acid metabolism’ pathway (C). The green broken line represents the change cure of the enrichment score (ES) of the group of genes, and the Y axis is the ES value. Each black vertical line represents one gene, and this part shows all genes under this pathway or GO term. The heat map shows the signal2 noise/log2 ratio value of all expressed genes. Red means value > 0, blue means value < 0, the darker the color, the greater the absolute value. 10. 7717/peerj. 12819/fig-7 Figure 7 Gene set variation analysis (GSVA). (A) GSVA of differentially expressed pathways between EP BM-MSCs and LP BM-MSCs. The dark blue column shows activated pathways in the LP BM-MSCs, and the green column indicates activated pathways in the EP BM-MSCs. (B) GSVA of differentially expressed pathways in EP BM-MSCs treated with aspirin. The dark blue column shows activated pathways in the EP BM-MSCs treated with aspirin, and the green column indicates activated pathways in the EP BM-MSCs. (C) GSVA of differentially expressed pathways in LP BM-MSCs treated with aspirin. The dark blue column shows activated pathways in the LP BM-MSCs treated with aspirin, and the green column indicates activated pathways in the LP BM-MSCs. The vertical axis represents each gene set, and the horizontal axis indicates the expression difference of each gene set in different groups. To further validate the results of RNA-sequencing, we performed RT-PCR analysis of the DEGs and the key genes found by GSEA analysis. We provided a conclusive table of these DEGs and the key genes in GSEA analysis in Table S10 and Fig. S2. Discussion Stem cells can repair damaged tissues and organs and reshape the biological functions of the body, so they are ideal seed cells for tissue engineering. Due to the limited amount of BM-MSCs obtained from the primary tissue, a substantice in vitro expansion is required to obtain a sufficient number of cells for clinical use. However, this process will cause the cells to undergo replicative senescence, which is also a major challenge in the current BM-MSCs research field. Previous research suggests that aspirin can impact senescence on BM-MSCs, but the mechanism has not been investigated yet. Therefore, we studied the gene expression profile changes in BM-MSCs treated with aspirin. To date, there are few studies that examine effects of aspirin’s impact on aging BM-MSCs. Aspirin, a non-steroidal anti-inflammatory drug (NSAID), is widely used to treat inflammation, cardiovascular diseases, and senile diseases. As noted in previous studies, aspirin can extend the lifespan of Caenorhabditis elegans ( Huang et al. , 2017 ), Drosophila melanogaster ( Danilov et al. , 2015 ), Mus musculus ( Strong et al. , 2008 ), as well as Homo sapiens ( Group, 2013 ). Moreover, emerging evidence indicates that long-term use of aspirin can improve health. In many cancer contexts, aspirin can significantly reduce the risk of many cancers, including enhancing the sensitivity to cisplatin in colorectal cancer ( Jiang et al. , 2020 ), down-regulating COX2 expression of lung cancer ( Chen et al. , 2019 ), activating AMP-activated protein kinase (AMPK) while inhibiting mTORC1 signaling in breast cancer ( Henry et al. , 2017 ), and other different mechanisms. In type 2 diabetes, high-dose aspirin promotes glucose metabolism and reduces fatty acid levels ( Hundal et al. , 2002 ). Besides, aspirin has considerable curative effects on neurodegenerative diseases, such as Parkinson’s disease ( Ren et al. , 2018 ), and Alzheimer’s disease ( Group et al. , 2008 ). Therefore, it is desirable to speculate that aspirin has anti-aging properties. A previous study explained that aspirin retards senescence of endothelial cells by reducing reactive oxygen species (ROS), increasing nitric oxide (NO) and cGMP levels ( Bode-Boger et al. , 2005 ). Cellular senescence in human and mouse fibroblasts was suppressed by aspirin via inhibiting COX2 expression ( Feng et al. , 2019 ). Together with our finding that aspirin could retard senescence in different species, including Homo sapiens, Mus musculus, and Rattus norvegicus. The difference between previous studies and our study was not only the species of experiment animal cell but also the concentration of aspirin. We applied 400 μmol/L aspirin, which was higher than the above-mentioned studies. In rat BM-MSCs, this dosage didn’t affect apoptosis and proliferation ( Zhan et al. , 2018 ). Hence, within different types of cells, aspirin could retard senescence to a certain extent with diverse concentrations. Besides, our results also showed aspirin could influence immune response and lipid metabolism in cells. In this study, we present a systematic comparative analysis of aspirin’s impact on EP and LP BM-MSCs. By using an original experiment and RNA sequencing approach, a comprehensive aspirin-BM-MSCs network for all differentially expressed genes was generated. Further, we performed GO functional enrichment and KEGG pathway analysis, GSEA, and GSVA to facilitate comparative studies in EP and LP BM-MSCs. Herein, this study offers new insights for alleviating the effects of aspirin on the senescence of BM-MSCs, and may open up the possibility of the application of aspirin in treating aging-related diseases. To analyze whether aspirin could slow cellular senescence, we conducted conventional measurements of aging, including gene and protein expression levels of p16, p21, p53, and SA β-gal staining. Cyclin-dependent kinase inhibitor 2a (CDKN2A/p16Ink4a) expression has been identified as a superior biomarker of aging ( Hudgins et al. , 2018 ; Krishnamurthy et al. , 2004 ). Direct evidence supporting a role for p16Ink4a in rat BM-MSCs has been widely applied, especially the increased expression in senescent BM-MSCs ( Yusop et al. , 2018 ; Zhang et al. , 2015 ; Zheng et al. , 2016 ). Further, cyclin-dependent kinase inhibitor p21 (CDKN1A) can act as a key molecular mediator of therapy-induced senescence ( Abbas & Dutta, 2009 ; Cazzalini et al. , 2010 ). P21 halts the cell-cycle progression after transcriptional activation by p53, which is a DNA damage response triggered by many senescence-inducing agents. The role of p21 in senescent cells has also been confirmed in rat BM-MSCs ( Wang et al. , 2020 ; Yang et al. , 2020 ; Zhang et al. , 2016 ). Typically, p53 is listed as an indicator of aging along with p16 and p21. Activation of p53 is involved in various processes including cell cycle arrest, apoptotic cell death, and cellular senescence ( Li et al. , 2012 ; Schmitt et al. , 2002 ; Vousden & Prives, 2009 ). Similarly, up-regulation of p53 is a consistent feature of cellular senescence in rat BM-MSCs ( Yang et al. , 2020 ; Zhang et al. , 2018 ; Zheng et al. , 2016 ). Our results suggested that adding aspirin to EP BM-MSCs and LP BM-MSCs could decrease aging indicators ( Fig. 1 ). Subsequently, we performed RNA-seq to explore the potential mechanism of aspirin’s impact on BM-MSCs. Biological process and KEGG enrichment analysis indicated that several pathways related to cellular processes were significantly enriched in LP BM-MSCs compared with EP BM-MSCs, including cellular senescence and p53 signaling pathways ( Figs. 3A – 3B ). This suggested that LP BM-MSCs were senescent. Further, GSEA and GSVA analysis of biological process and KEGG pathways identified cholesterol, sterol, steroid, and other lipids biosynthetic as significantly increased in LP BM-MSCs compared with EP BM-MSCs ( Figs. 4A – 4C, 7A ). Consistent with our findings, emerging evidence indicates the accumulation of lipids in aging cells including BM-MSCs ( Lu et al. , 2019 ; Seo et al. , 2019 ; Stolzing & Scutt, 2006 ). BM-MSCs underwent excessive adipogenic differentiation rather than osteogenic differentiation, which resulted in bone metabolism imbalanced and bone mass losing, during the abuse of hormones, menopause, and aging ( Liu, Xia & Li, 2015 ). Consistently, a trial conducted on bone marrow-derived mesenchymal stem cells, indicated that global lipid distribution and the relevant metabolic flows were disturbed when bone marrow-derived mesenchymal stem cells after serially passaging ( Lu et al. , 2019 ). For EP BM-MSCs treated with aspirin, analysis of biological process and KEGG enrichment, excluding broad pathways, found several pathways involved senescence, immune response, and lipid metabolism. For example, they included focal adhesion, ECM-receptor interaction, PI3K-Akt signaling pathway, cell cycle, protein digestion and absorption, MAPK signaling pathway, FoxO signaling pathway, and others ( Figs. 3C – 3D ). Among these, GSEA analysis showed the DEGs were significant with negative regulation of interleukin-6 (IL-6) production, longevity regulating pathway, and regulation of lipolysis in adipocytes ( Figs. 5A – 5C ). On the other hand, we found reactive oxygen species (ROS) pathway, inflammatory response retained the lowest t value of GSVA score in GSVA analysis ( Fig. 7B ). These findings suggested that aspirin might influence the immune response and lipid metabolism on EP BM-MSCs to slow cellular senescence. As noted in previous studies, the secretion of IL-6 is a characteristic of the secretion of senescence-associated inflammatory cytokines (SASP), which is promoted by persistent DNA damage ( Rodier et al. , 2009 ). Also, one common denominator of aging is increasing ROS production resulted from progressive mitochondrial dysfunction ( Harman, 1965 ). Aspirin has previously been shown to diminish ROS and inflammatory cytokines, including IL-1β, IL-6, and tumor necrosis factor-alpha (TNF-α) ( Liu et al. , 2019 ; Wang, Huang & Zuo, 2014 ). Regarding aspirin’s influence on lipid metabolism, we took note of some previous reports showing aspirin could inhibit the adipogenic differentiation of BM-MSCs and preadipocytes ( Su et al. , 2014 ; Zhan et al. , 2018 ). Although this observation was not verified in cells not induced to adipogenesis. Combined with our findings, aspirin might alleviate the accumulation of lipids in cells, reduce the accumulation of intracellular ROS, and eliminate inflammatory cytokines to retard aging in EP BM-MSCs. For LP BM-MSCs treated with aspirin, the analysis of biological process and KEGG enrichment were consistent with EP BM-MSCs treated with aspirin, and included several pathways involved in senescence, including focal adhesion, PI3K-Akt signaling pathway, ECM-receptor interaction, apoptosis, MAPK signaling pathway, cytokine-cytokine receptor interaction, protein digestion and absorption, and other pathways ( Figs. 3E – 3F ). GSEA analysis showed the DEGs were significant concerning linoleic acid metabolism, alpha-linolenic acid metabolism, and regulation of lipid metabolic process ( Figs. 6A – 6C ). On the other hand, GSVA analysis uncovered interferon α response retained the lowest t value of GSVA ( Fig. 7C ). These observations, consistent with EP BM-MSCs treated with aspirin, illustrated that aspirin might influence lipid metabolism and impact immune response. It was also possible that linoleic acid caused senescence through multiple mechanisms, as previous studies demonstrate increasing IKKβ activity and decreasing autophagic flux ( Kim et al. , 2005 ; Lee et al. , 2015 ; Raederstorff, Loechleiter & Moser, 1995 ). In our study, we found linolenic acid metabolism was downregulated by the addition of aspirin in LP BM-MSCs, in agreement with previous observations in human serum ( Ellero-Simatos et al. , 2015 ). In addition, interferon response can contribute to aging, as exemplified by studies with senescent rats in multiple tissues ( O’Brown et al. , 2015 ; Shavlakadze et al. , 2019 ). The administration of aspirin was capable of inhibiting IFN-γ production ( Cao et al. , 2015 ; Liu et al. , 2011 ), which indicated aspirin might be responsible for delaying senescence by inhibiting interferon response. Our study showed that aspirin could delay BM-MSCs senescence to a certain extent. The mechanisms of this delay might involve influencing immune response and lipid metabolism. These findings may allow better understanding and clinical application of aspirin-treated BM-MSCs. In this study, we performed an original experiment and RNA-seq to detect aspirin’s influence on bone marrow-derived mesenchymal stem cells, this research still existed restrictions. Initially, we chose a concentration of 400 μmol/L aspirin because it did not affect cell proliferation and apoptosis. However, this dosage is still greater than the blood concentration in the human body after oral administration of aspirin ( Hobl et al. , 2015 ; Levy, 1978 ; Lucker et al. , 1992 ). Therefore, if aspirin is applied to the clinic to delay aging, it is also necessary to increase the long-term use of low-concentration aspirin in vivo / in vitro experimental research. Second, due to the differential biological process and KEGG pathways both being involved in promoting senescence and delaying senescence, combined with our experiment finding, we only focused on inhibiting senescence-related pathways. This led to the analysis of the effect of aspirin on BM-MSCs that didn’t reach sufficient depth. Third, our previous research showed aspirin could interfere with adipogenic differentiation of BM-MSCs by inhibiting HDAC9 expression. In the present study, we found the potential mechanisms of aspirin delayed cell senescence included inhibiting lipid accumulation. As predicted, HDAC9 expression was elevated in aging BM-MSCs ( Fig. S3A ). After treatment with aspirin, HDAC9 expression was down-regulated ( Fig. S3B ). This likely indicates HDAC9 participates in the senescent process and is regulated aspirin. However, the exact role of HDAC9 in the aging process has not been verified. Further studies are needed to define these intricate mechanisms. Fourth, although aspirin was found to retard BM-MSCs senescence, it has not been verified in other cell types. Fifth, in our study, we have verified the meaningful genes by performing PCR experiment. Silencing and overexpression experiments of these meaningful genes will be necessary to verify the function in the future. Whether aspirin and its products could retard cell senescence and be applied locally in the clinical setting is an important topic for future studies. Conclusion In summary, aspirin could delay BM-MSCs senescence. The potential mechanisms include influencing immune response and lipid metabolism. Supplemental Information 10. 7717/peerj. 12819/supp-1 Supplemental Information 1 Senescence-associated β-galactosidase (SA-β-gal) staining of bone marrow-derived mesenchymal stem cells at passage 1. Senescence-associated β-galactosidase (SA-β-gal) staining of bone marrow-derived mesenchymal stem cells at passage 1 with or without aspirin treatment (400 μmol/L). Scale bar = 100 μm. Click here for additional data file. 10. 7717/peerj. 12819/supp-2 Supplemental Information 2 Validation of key genes after aspirin treatment based on GSEA analysis. Gene expression of Scd, Sirt1, and Pla2g2a in different groups. * P < 0. 05, * P < 0. 01, **** P < 0. 0001, NS: no significance, with comparisons indicated by lines. Click here for additional data file. 10. 7717/peerj. 12819/supp-3 Supplemental Information 3 Validation of Hdac9 gene expression after aspirin treatment. Gene expression of Hdac9 in different groups. * P < 0. 05, * P < 0. 01, with comparisons indicated by lines. Click here for additional data file. 10. 7717/peerj. 12819/supp-4 Supplemental Information 4 Different expression genes (DEGs) between p5 and p20. Click here for additional data file. 10. 7717/peerj. 12819/supp-5 Supplemental Information 5 Different expression genes (DEGs) between p5 and p5 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-6 Supplemental Information 6 Different expression genes (DEGs) between p20 and p20 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-7 Supplemental Information 7 GO analysis of different expression genes (DEGs) between p5 and p5 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-8 Supplemental Information 8 GO analysis of different expression genes (DEGs) between p5 and p20 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-9 Supplemental Information 9 GO analysis of different expression genes (DEGs) between p20 and p20 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-10 Supplemental Information 10 KEGG analysis of different expression genes (DEGs) between p5 and p20. Click here for additional data file. 10. 7717/peerj. 12819/supp-11 Supplemental Information 11 KEGG analysis of different expression genes (DEGs) between p5 and p5 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-12 Supplemental Information 12 KEGG analysis of different expression genes (DEGs) between p20 and p20 treated with aspirin. Click here for additional data file. 10. 7717/peerj. 12819/supp-13 Supplemental Information 13 Genes of interest and associated functions of interest arising after aspirin treatment in bone marrow-derived mesenchymal stem cells in vitro culture. Log2FC represents log2 fold change. Positive number of log2FC(p20/p5) is upregulated and negative is downregulated versus P5. Positive number of log2FC(p5+A/p5) is upregulated and negative is downregulated versus P5. Positive number of log2FC(p20+A/p20) is upregulated and negative is downregulated versus P20. Click here for additional data file. 10. 7717/peerj. 12819/supp-14 Supplemental Information 14 Raw data for Fig. 1B. Click here for additional data file. 10. 7717/peerj. 12819/supp-15 Supplemental Information 15 Raw data for Fig. 1C. Click here for additional data file. 10. 7717/peerj. 12819/supp-16 Supplemental Information 16 Uncropped blots of p21 in Fig. 1C. Click here for additional data file. 10. 7717/peerj. 12819/supp-17 Supplemental Information 17 Uncropped blots of p53 in Fig. 1C. Click here for additional data file. 10. 7717/peerj. 12819/supp-18 Supplemental Information 18 Uncropped blots of GAPDH in Fig. 1C. Click here for additional data file. 10. 7717/peerj. 12819/supp-19 Supplemental Information 19 DEGs between p5+ASA and p20+ASA. Click here for additional data file. 10. 7717/peerj. 12819/supp-20 Supplemental Information 20 DEGs between p5 and p5+ASA. Click here for additional data file. 10. 7717/peerj. 12819/supp-21 Supplemental Information 21 DEGs between p5 and p20. Click here for additional data file. 10. 7717/peerj. 12819/supp-22 Supplemental Information 22 DEGs between p20 and p20+ASA. Click here for additional data file. 10. 7717/peerj. 12819/supp-23 Supplemental Information 23 Raw data for Figures S2 and S3. Click here for additional data file. 10. 7717/peerj. 12819/supp-24 Supplemental Information 24 Primer sequences of validated genes in Figures S2 and S3. Click here for additional data file.

10. 7717/peerj. 12824

PeerJ

Effectiveness of interactive teaching intervention on medical students’ knowledge and attitudes toward stem cells, their therapeutic uses, and potential research applications

Background Stem cell science is rapidly developing with the potential to alleviate many non-treatable diseases. Medical students, as future physicians, should be equipped with the proper knowledge and attitude regarding this hopeful field. Interactive teaching, whereby the teachers actively involve the students in the learning process, is a promising approach to improve their interest, knowledge, and team spirit. This study aims to evaluate the effectiveness of an interactive teaching intervention on medical students’ knowledge and attitudes about stem cell research and therapy. Methods A pre-post test study design was employed. A six-session interactive teaching course was conducted for a duration of six weeks as an intervention. Pre- and post-intervention surveys were used. The differences in the mean scores of students’ knowledge and attitudes were examined using paired t-test, while gender differences were examined using an independent t-test. Results Out of 71 sixth-year medical students from different nationalities invited to participate in this study, the interactive teaching course was initiated by 58 students resulting in a participation rate of 81. 7%. Out of 58 students, 48 (82. 8%) completed the entire course. The mean age (standard deviation) of students was 24 (1. 2) years, and 32 (66. 7%) were males. The results showed poor knowledge about stem cells among the medical students in the pre-intervention phase. Total scores of stem cell-related knowledge and attitudes significantly improved post-intervention. Gender differences in knowledge and attitudes scores were not statistically significant post-intervention. Conclusions Integrating stem cell science into medical curricula coupled with interactive learning approaches effectively increased students’ knowledge about recent advances in stem cell research and therapy and improved attitudes toward stem cell research and applications.

Introduction The emerging stem cell (SC) biology discipline and the rapid revolution in SC research have radically transformed our thinking of cells, evolution, and disease. Using SCs for clinical applications represents the future of translational medicine since SCs can potentially be used to treat many kinds of difficult diseases that cannot currently be treated ( Chang et al. , 2018 ; Protze, Lee & Keller, 2019 ). Advances in SC research combined with tissue engineering techniques promise therapies to restore or replace damaged tissues ( Kwon et al. , 2018 ). This raises the need for medical education to introduce basic SC knowledge and the concept of translational medicine into the life sciences field. At the same time, SC research and applications still raise complex social, legal, ethical, and religious issues ( Al-Aqeel, 2005 ; Curley & Sharples, 2006 ; Pourebrahim, Goldouzian & Ramezani, 2020 ), especially in conservative societies ( Bouzenita, 2017 ). The emerging developments in SC applications are transforming the priorities of undergraduate and postgraduate medical educational programs ( Scott, 2015 ). Today, the traditional academic model for medical education is challenged by an evident gap between the rapidly changing disciplines in basic biomedical sciences and clinical practice. Although medical students have access to SC research theoretical advancements, traditional teaching approaches still fail to bridge this practice gap ( Brass, 2009 ). Thus, updated teaching techniques that facilitate the integration of SC research advancements with clinical practice are critical for medical students to achieve optimum patient care ( Knoepfler, 2013 ). Restructuring medical education to meet the current and future health care needs of SC-based interventions, including new curricula featuring the ethical, legal, and social implications of SC research, are thus a priority ( Pershing & Fuchs, 2013 ; Pierret & Friedrichsen, 2009 ). Since the early 1990s, many medical curricula have transitioned from traditional subject-based teaching toward integrated system-based teaching ( Ling et al. , 2008 ). Traditional didactic lectures for one hour become monotonous after 15–20 min as students’ participation in the learning process is minimal ( Gupta, BhattiK & AgnihotriP, 2015 ). On the other hand, the interactive teaching approach actively engages learners and interchanges ideas between learners and facilitators ( Kaur et al. , 2011 ). The effectiveness of educational interventions in increasing knowledge and attitudes towards SC applications was reported previously by a few studies ( Azzazy & Mohamed, 2016 ; Jin et al. , 2018 ; Kaya et al. , 2015 ). Although it is currently a hot research topic, SC education for students is uncommon ( Pierret & Friedrichsen, 2009 ). The interactive teaching modality was designed to introduce medical students to the pioneering area of SC biology and shed light on current advances in SC research. Medical students, as future physicians, are expected to answer patients’ questions regarding SCs and help them differentiate between what is realistic and unrealistic regarding SC-based therapies. Also, they should be able to use evolving discoveries in SC research and apply them in the care of patients. Thus, we aimed in this study to gauge the medical students’ knowledge and attitudes toward SCs, their therapeutic uses, and potential research applications and then evaluate the effectiveness of a six-session interactive teaching intervention on their knowledge and attitudes. Material and Methods Study design, participants, and setting A pre-post test design was employed for a sample of 71 sixth-year medical students, at the University of Science and Technology Yemen-Jordan branch (USTY-Jo), during the first semester of the academic year 2018-2019. An orientation lecture was held before the initiation of the study to explain the study aims, design, and details for the students and invite the students to participate. Study participation was voluntary, and the pre-intervention survey was distributed to all medical students who agreed to participate. After that, the participants were then invited to attend a six-session interactive teaching course, the intervention, for a duration of six weeks. This intervention was a part of phase I of the “Stem Cells: Hope or Hype?” project. Each interactive session lasted two to three hours and included brainstorming, learning by teaching, role-playing, class debate, panel discussions, reflections on stories, real-life situations, case-based scenarios, or videos. Details about the intervention are summarized in Table 1. After finishing the intervention, the same survey was distributed among the participants. 10. 7717/peerj. 12824/table-1 Table 1 Detailed study intervention. Number and title of intervention week Objectives of the intervention Interactive teaching methods Week one: Stem cell basic biology ✓ Reviewing the history of stem cell (SC) research. ✓ Understanding the basic biology of SCs and identifying characteristics that distinguish SCs from other types of cells. ✓ Classifying SCs according to source and potency. Brainstorming: The lecturer asked students an opening question: what do you know about SCs?, then he used the whiteboard to list all the ideas generated by the students and grouped them into few headlines. Visual aids: The lecturer presented a short video about the discovery of the microscope by Robert Hooke, and then he presented a diagram illustrating major historical events in SC research. Week two: Stem cell potential applications Recognizing potential applications of SCs in studying early human development, modeling diseases in a culture dish, testing new drugs, and restoring lost tissues. Group activity and learning by teaching: Students were divided into eight groups and were given one of four topics that cover potential applications of SCs. Each group had to read five articles about the topic and do a seminar for other students. Week three: Unproven stem cell therapies and stem cell tourism ✓ Listing current therapeutic uses of SCs such as bone marrow transplantation for leukemia. ✓ Shedding light on potential therapeutic uses of SCs such as limbal SCs for degenerative eye diseases. ✓ Increasing awareness about SC tourism and severe risks due to trying unproven SC therapies. Case-based scenarios: For patients who tried unproven SC therapies. Group activity: Students were divided into eight groups assigned to search for websites that promote unproven SC therapies. Week four: Stem cell research ✓ Understanding the induced pluripotent stem cells (iPSCs) and the role of transcription factors. ✓ Explaining SC-assisted technologies such as MRT, SCNT, and human/animal chimeras. Story: A reflection on Shinya Yamanaka’s story, who won the Nobel Prize for discovering induced pluripotent SCs. Week five: Cord blood banking and donation ✓Explaining techniques and procedures of cord blood collection, banking, and donation. ✓ Summarizing advantages and disadvantages of cord blood transplantation in comparison with bone marrow transplantation. ✓ Comparing between different types of cord blood banks. Role-playing: Students played different roles assigned to them: parents who are interested in cord blood banking and healthcare providers who should answer parents’ questions. Guest lecturer: To take about cord blood banking. Real-life situations: Students provided health education for pregnant women about cord blood banking. Week six: Bioethics of stem cell research Discussing ethical controversies surrounding SC research and their-assisted technologies. Panel discussion: With bioethics expert. Class debate: Class was divided into eight groups; four groups argued for another four groups against research involving embryonic SCs. Ethical considerations The study protocol was reviewed and ethically approved by the Institutional Review Board (IRB) of the research and ethics committee at USTY-Jo (IRB number, 9/120/2019). This study was conducted following the 1975 Helsinki declaration, as revised in 2008 and later amendments or comparable ethical standards. The study objectives and design were duly explained to the study participants during the orientation lecture and with each intervention session. They were informed about the study objectives, design, duration, interactive teaching methods, and the guest lecturers were invited to participate in the interactive teaching sessions. A written, signed, informed consent was obtained from each participant. Participants were informed that they could terminate the survey and interventions at any time desired. Participants did not receive any compensation or rewards for their participation in the study. To conduct the study with keeping the participant anonymity and survey confidentiality in light of its pre-post test design, we need a method of identifying the participant so that we can measure the change from the first survey to the second for the same participant without breaching the anonymity and confidentiality. One possible way is to anonymously generate a unique ID code for each participant. The codes should be easily recovered if needed, unlikely duplicated across multiple respondents, and unique for each participant. Thus, a coding system was created based on the participants’ names and birth dates. We asked each participant to generate their own ID code by providing the first letters of their first and family names (A–Z) plus a four-digit code that represents birthday (01-31) and month of birth (01-12). The participants’ ID codes were essential for data analysis to compare pre-and-post intervention scores and avoid duplicated data with preserving anonymity. Thus, the study was undertaken with complete confidentiality, and information provided by study participants was not disclosed to others. Study tools After detailed reviewing the literature regarding SC knowledge and attitudes, the researchers developed a structured, self-administered questionnaire. The questionnaire was not based on a particular study but preferably on information from various studies and recent guidelines from international organizations such as the International Society for Stem Cell Research (ISSCR) and the New York Stem Cell Foundation (NYSCF) ( Azzazy & Mohamed, 2016 ; Lovell-Badge et al. , 2021 ; NYSCF, 2017 ). The questionnaire was reviewed by a panel of experts in SC clinical practice and teaching, pilot-tested on 20 participants, and the necessary modifications were done. The questionnaire was designed and distributed in the English language as it is the official teaching language of the Jordanian medical schools. A soft copy of the distributed questionnaire is provided in File S1. The questionnaire was started by asking the participant to provide their ID code following specific instructions, as mentioned earlier. Then, the questionnaire included three major sections: demographic characteristics, SC knowledge, and SC attitudes. The demographics section included questions about age, gender, nationality of participants, name of the registered medical school, and student year level. The SC knowledge section began with a rating question about the participant perception of knowledge regarding SCs in general with a ten-point Likert scale, ranging from “zero = low knowledge” to “10 = high knowledge”. Then, a question about the participants’ preferred sources of knowledge about SCs with multiple choices included books, medical journals, workshops, social media, lectures, medical conferences, panel discussions, and other sources. After that, the SC knowledge section included 27 statements to measure the participants’ knowledge regarding SCs. These 27 statements consist of eleven correctly stated statements, seven false or misleading statements, and neither true nor false statements, with a total of nine statements. The SC knowledge section statements have an acceptable to excellent internal consistency and reliability with a Cronbach’s Alpha of 0. 61 and 0. 78 in pre-and post-intervention, respectively. The SC knowledge section was divided into four domains, including basic knowledge about SCs with a total of 13 statements (Cronbach’s Alpha =0. 42 and 0. 61 in pre-and post-intervention, respectively), potential applications of SCs with a total of four statements (Cronbach’s Alpha =0. 69 and 0. 66), therapeutic uses of SCs with a total of four statements (Cronbach’s Alpha =0. 44 and 0. 32), and lastly the participant knowledge about SC research with a total of six statements (Cronbach’s Alpha =0. 86 and 0. 75). The third section was designed to assess the medical students’ attitudes toward SCs via a total of ten statements with Cronbach’s Alpha of 0. 76 and 0. 68 in pre-and post-intervention, respectively. Participants responded to each statement of the SC knowledge and attitude scales described above, using a 5-point Likert scale ranging between “Strongly Disagree” and “Strongly Agree” for each statement to provide high-resolution data and detailed information. After that, each response was scored from “Zero = Strongly Disagree” to “Four = Strongly Agree” except for the seven false statements, where the code was reversed to be “Four = Strongly Disagree” and “Zero = Strongly Agree”. Responses to statements were summed to create scores for the total knowledge, each of the four knowledge domains, and total attitude. Thus, higher scores indicated good knowledge and positive attitude, while lower scores indicated poor knowledge and negative attitude. Knowledge scores ranged from zero to 108 for “total SC knowledge”, zero to 52 for “SC basic knowledge”, zero to 16 for “SC potential applications”, zero to 16 for “SC therapeutic uses”, and zero to 24 for “SC research”. The total attitude score ranged from zero to 40. After that, the scores of scales were converted into mean scores ranging from zero to four by dividing the scale score on the number of scale statements. Statistical analysis Data were analyzed using IBM Statistical Package for Social Sciences (SPSS), Windows Version 25. 0 (IBM Corp. , Armonk, NY, USA). Internal consistency for scales and subscales were tested using Cronbach’s alpha. Descriptive statistics were presented as means and standard deviations (SD) for continuous variables after verifying the normality of the dataset. Categorical variables were presented as proportions and frequencies. A Paired-samples t -test was used to examine the mean differences (MD) in students’ knowledge and attitude scores pre- and post-educational intervention, and the statistical significance and 95% confidence intervals of the difference in means were reported. Independent-samples t -test was used to examine mean gender differences in students’ knowledge and attitude scores. A p-value was set at or less than 0. 05 to be significant. Results Out of 71 sixth-year medical students invited to participate in this study, 58 initiated the interactive teaching course resulting in a participation rate of 81. 7%. The final sample consisted of 48 medical students who initially enrolled and completed the entire six-week course sessions with a completion rate of 82. 8%. Out of 48 medical students, 32 (66. 7%) were males, and more than half (56. 3%) were of Jordanian or Yemeni nationalities. The enrolled Students’ mean age (SD) was 24. 0 (1. 2) years. Demographic characteristics of study participants are summarized in Table 2. 10. 7717/peerj. 12824/table-2 Table 2 Respondents’ Characteristics ( n = 48). Characteristics Value Gender, N (%) Male 32 (66. 7%) Female 16 (33. 3%) Age, M (SD) 24. 0 (1. 2) Nationality, N (%) Jordanian 14 (29. 2%) Palestinian 3 (6. 3%) Syrian 8 (16. 7%) Iraqi 6 (12. 5%) Yemeni 13 (27. 1%) Others 4 (8. 3%) Notes. Abbreviations: M, Mean; SD, Standard Deviation. Knowledge regarding stem cells The three most common sources of knowledge regarding SCs before the intervention course were lectures (56. 3%), social media (45. 8%), and books (41. 7%), while any participant did not report panel discussions as a source of knowledge ( Fig. 1 ). Detailed information about pre-and post-educational intervention knowledge scores is summarized in Table 3. Pre-intervention, the lowest mean score among knowledge domains was observed with SC research section (1. 76 (0. 89)), and therapeutic uses (1. 84 (0. 63)), followed by SC basic knowledge (2. 14 (0. 30)) and their potential applications (2. 66 (0. 77)). The mean scores of all knowledge domains were lower than three in the pre-intervention phase of the study. The mean (SD) total knowledge score was 2. 09 (0. 30) pre-intervention, which is significantly improved to 3. 09 (0. 41) post-intervention ( p < 0. 001). Similarly, all knowledge domains’ scores significantly increased following the intervention. 10. 7717/peerj. 12824/fig-1 Figure 1 The reported sources of knowledge about stem cells before the intervention course. 10. 7717/peerj. 12824/table-3 Table 3 Pre- and post-educational intervention mean knowledge scores and differences ( n = 48). Pre- and post-educational intervention mean scores of students’ knowledge regarding stem cells, their potential applications, therapeutic uses, and research involving them Differences between pre-and post-educational interventions Score M (SD) Min–Max MD 95% CI p-value Lower Upper Stem cells: basic knowledge 1- I have sufficient knowledge of different types of stem cells, such as adult and embryonic stem cells. Pre 1. 88 (0. 94) 0–4 1. 89 1. 57 2. 23 <0. 001 * Post 3. 77 (0. 43) 3–4 2- I have sufficient knowledge of the sources of stem cells. Pre 2. 17 (0. 78) 0–4 1. 50 1. 22 1. 78 <0. 001 * Post 3. 67 (0. 52) 2–4 3- I have sufficient knowledge of the therapeutic uses of stem cells. Pre 2. 15 (1. 05) 0–4 1. 54 1. 22 1. 86 <0. 001 * Post 3. 69 (0. 51) 2–4 4- I have sufficient knowledge of the three germ layers (endoderm, mesoderm and ectoderm), and organs and tissues generated from each layer. Pre 2. 67 (1. 02) 0–4 1. 02 0. 70 1. 34 <0. 001 * Post 3. 69 (0. 59) 2–4 5- Cell differentiation is the process by which stem cells become more specialized cell types (true). Pre 2. 85 (0. 97) 1–4 0. 65 0. 38 0. 92 <0. 001 * Post 3. 50 (0. 72) 1–4 6- As a stem cell differentiates, it gradually loses potency and becomes unipotent (true). Pre 2. 23 (1. 06) 0–4 0. 48 0. 13 0. 83 0. 009 * Post 2. 71 (1. 32) 0–4 7- Self-renewing is the ability of a stem cell to produce more stem cells with identical characteristics as the “parent” cell (true). Pre 2. 46 (0. 82) 0–4 0. 89 0. 60 1. 20 <0. 001 * Post 3. 35 (0. 84) 1–4 8- Adult stem cells are pluripotent cells that have the potential to make all cell types of the body (false). † Pre 1. 58 (1. 16) 0–4 1. 34 0. 75 1. 92 <0. 001 * Post 2. 92 (1. 49) 0–4 9- Bone marrow is the only source for adult stem cells (false). † Pre 2. 17 (1. 24) 0–4 1. 10 0. 67 1. 54 <0. 001 * Post 3. 27 (1. 25) 0–4 10- Stem cells can differentiate into many cell types within a germ layer (true). Pre 2. 73 (0. 94) 0–4 0. 52 0. 12 0. 93 0. 013 * Post 3. 25 (1. 02) 0–4 11- Embryonic stem cells are derived from leftover blastocysts after in vitro fertilization (true). Pre 2. 06 (0. 76) 0–4 0. 90 0. 48 1. 31 <0. 001 * Post 2. 96 (1. 34) 0–4 12- Embryonic stem cells are derived from the umbilical cord after childbirth (false). † Pre 1. 35 (0. 91) 0–3 0. 38 −0. 16 0. 91 0. 165 Post 1. 73 (1. 65) 0–4 13- Embryonic stem cells are derived from the trophoblast of blastocysts (false). † Pre 1. 54 (0. 65) 0–4 0. 25 −0. 25 0. 75 0. 316 Post 1. 79 (1. 64) 0–4 The total score of stem cell basic knowledge Pre 2. 14 (0. 30) 0. 95 0. 83 1. 09 <0. 001 * Post 3. 09 (0. 47) Stem cells: potential applications 14- Stem cells can be used to study early human development (true). Pre 3. 02 (0. 84) 1–4 0. 40 0. 13 0. 66 0. 004 * Post 3. 42 (0. 71) 1–4 15- Stem cells can be used to understand the pathophysiology and analyze disease mechanisms by modeling disease in a culture dish outside the human body (true). Pre 2. 56 (1. 09) 0–4 0. 92 0. 48 1. 35 <0. 001 * Post 3. 48 (0. 83) 0–4 16- Stem cells can be used to test and screen new drug candidates and toxins to figure out their potential side effects (true). Pre 2. 21 (1. 09) 0–4 1. 27 0. 84 1. 70 <0. 001 * Post 3. 48 (0. 83) 0–4 17- Stem cells can be used to replace or restore tissues that have been damaged by disease or injury, such as diabetes, heart attacks, Parkinson’s disease, skin burns, or spinal cord injury (true). Pre 2. 85 (1. 09) 0–4 0. 73 0. 38 1. 08 <0. 001 * Post 3. 58 (0. 85) 0–4 The total score of stem cell potential applications Pre 2. 66 (0. 77) 0. 80 0. 53 1. 09 <0. 001 * Post 3. 46 (0. 59) Stem cells: therapeutic uses 18- There is a wide range of conditions or diseases for which stem cell therapies have been proven to be safe and effective such as osteoarthritis and multiple sclerosis (false). † Pre 1. 54 (0. 97) 0–4 0. 69 0. 23 1. 15 0. 004 * Post 2. 23 (1. 53) 0–4 19- There is nothing to lose from trying unproven stem cell therapies since they can provide hope for hopeful patients (false). † Pre 1. 71 (1. 07) 0–4 0. 67 0. 24 1. 10 0. 003 * Post 2. 38 (1. 39) 0–4 20- Bone marrow-derived stem cells will spontaneously regenerate into different cell types such as hepatocytes and neural cells without manipulation in the lab (false). † Pre 1. 88 (1. 00) 0–4 0. 45 −0. 01 0. 93 0. 055 Post 2. 33 (1. 53) 0–4 21- If the balance is skewed between differentiation and self-renewing properties of stem cells, it may result in tumor formation (true). Pre 2. 25 (0. 79) 0–4 0. 63 0. 26 0. 99 <0. 001 * Post 2. 88 (1. 04) 1–4 The total score of stem cell therapeutic uses Pre 1. 84 (0. 63) 0. 61 0. 36 0. 85 <0. 001 * Post 2. 45 (0. 80) 22- I would be confident to explain the induced-Pluripotent Stem Cells (iPSCs). Pre 1. 65 (1. 16) 0–4 1. 75 1. 38 2. 12 <0. 001 * Post 3. 40 (0. 71) 2–4 23- I would be confident to explain the transcription factors. Pre 1. 85 (1. 19) 0–4 1. 28 0. 87 1. 67 <0. 001 * Post 3. 13 (0. 89) 0–4 24- Adult cells can be “reprogrammed” genetically to assume stem cell-like state (true). Pre 1. 85 (1. 05) 0–4 1. 46 1. 04 1. 88 <0. 001 * Post 3. 31 (0. 83) 1–4 25- I would be confident to discuss the Somatic Cell Nuclear Transfer (SCNT). Pre 1. 58 (1. 07) 0–4 1. 48 1. 07 1. 89 <0. 001 * Post 3. 06 (0. 10) 0–4 26- I would be confident to explain the differences between therapeutic cloning and reproductive cloning. Pre 1. 81 (1. 20) 0–4 1. 40 0. 96 1. 83 <0. 001 * Post 3. 21 (0. 82) 0–4 27- I would be confident to discuss the mitochondrial replacement therapy. Pre 1. 83 (1. 19) 0–4 1. 69 1. 28 2. 10 <0. 001 * Post 3. 52 (0. 74) 1–4 The total score of stem cell research Pre 1. 76 (0. 89) 1. 51 1. 20 1. 82 <0. 001 * Post 3. 27 (0. 56) The total knowledge score Pre 2. 09 (0. 30) 1. 00 0. 86 1. 15 <0. 001 * Post 3. 09 (0. 41) Notes. M, Mean; SD, Standard Deviation; Min, Minimum score; Max, Maximum score; CI, Confidence Interval; MD, Mean Difference; Pre, Pre-educational intervention; Post, Post-educational intervention * Significant at p < 0. 05 based on paired-samples t -test. † The code was reversed for the false or misleading statements to be “Four = Strongly Disagree” and “Zero = Strongly Agree”. The total score of knowledge is the sum of the statements’ scores for the four major domains (stem cell basic knowledge, potential applications, therapeutic uses, and research) for each participant divided by 27. The mean SC basic knowledge domain score significantly increased from 2. 14 (0. 30) pre-intervention to 3. 09 (0. 47) post-intervention ( p < 0. 001). Post-intervention vs. pre-intervention, participants reported improved the knowledge with different types of SCs (3. 77 (0. 43) vs. 1. 88 (0. 94), p < 0. 001), sources of SCs (3. 67 (0. 52) vs. 2. 17 (0. 78), p < 0. 001), therapeutic uses of SCs (3. 69 (0. 51) vs. 2. 15 (1. 05), p < 0. 001) and three germ layers from which tissues and organs are generated (3. 69 (0. 59) vs. 2. 67 (1. 02), p < 0. 001). Students’ knowledge of sources of embryonic SCs significantly improved for statements related to leftover blastocysts after in vitro fertilization (2. 96 (1. 34) post-intervention vs. 2. 06 (0. 76) pre-intervention, p < 0. 001), but not for statements related to umbilical cord (1. 73 (1. 65) vs. 1. 35 (0. 91), p =0. 165) or trophoblast of blastocyst (1. 79 (1. 64) vs. 1. 54 (0. 65), p =0. 316). For potential applications of SC domain, the mean score significantly increased from 2. 66 (0. 77) pre-intervention to 3. 46 (0. 59) post-intervention ( p < 0. 001). Post-intervention vs. pre-intervention, students reported significantly higher knowledge scores regarding potential applications of SCs such as replacing or restoring damaged tissues (3. 58 (0. 85) vs. 2. 85 (1. 09), p < 0. 001), screening new drugs and toxins (3. 48 (0. 83) vs. 2. 21 (1. 09), p < 0. 001), modeling disease in a culture dish (3. 48 (0. 83) vs. 2. 56 (1. 09), p < 0. 001) and studying early human development (3. 42 (0. 71) vs. 3. 02 (0. 84), p =0. 004). For SC therapeutic uses domain, the mean total score significantly increased from 1. 84 (0. 63) pre-intervention to 2. 45 (0. 80) after the intervention course ( p < 0. 001). Compared to pre-intervention results, the students became significantly more aware about the side effects of trying unproven SC therapies after the intervention course (2. 38 (1. 39) vs. 1. 71 (1. 07), p = 0. 003 ) and tumor formation potential if the balance is skewed between cell differentiation and self-renewing properties of SCs (2. 88 (1. 04) vs. 2. 25 (0. 79), p =0. 001). In the SC research domain, the mean total score significantly increased from 1. 76 (0. 89) to 3. 27 (0. 56) ( p < 0. 001). Post-intervention vs. pre-intervention, students became more comfortable in giving an explanation of induced pluripotent SCs (3. 40 (0. 71) vs. 1. 65 (1. 16), p < 0. 001), transcription factors (3. 13 (0. 89) vs. 1. 85 (1. 19), p < 0. 001), and differences between therapeutic cloning and reproductive cloning (3. 21 (0. 82) vs. 1. 81 (1. 20), p < 0. 001). Moreover, participants became more knowledgeable that adult cells can be “reprogrammed” genetically to assume an SC-like state (3. 31 (0. 83) vs. 1. 85 (1. 05), p < 0. 001). Students were also more comfortable discussing mitochondrial replacement therapy (3. 52 (0. 74) vs. 1. 83 (1. 19), p < 0. 001) and somatic cell nuclear transfer (3. 06 (0. 10) vs. 1. 58 (1. 07), p < 0. 001). Attitudes toward stem cells As described in Table 4, the total attitude score significantly increased from 2. 66 (0. 56) to 2. 85 (0. 53) ( p =0. 048). Post-intervention vs. pre-intervention, students became more interested in expanding their knowledge regarding SCs (3. 77 (0. 43) vs. 3. 29 (0. 92), p =0. 001), and considered a well-structured program or training focusing on SC science (3. 48 (0. 68) vs. 2. 83 (0. 91), p < 0. 001). Students reported improved positive attitudes regarding integration of SC education in medical college curricula (3. 35 (0. 93) vs. 2. 83 (0. 10), p =0. 010), translational research (3. 27 (0. 84) vs. 2. 83 (0. 93), p =0. 009), and spending more money by government to support SC research (3. 69 (0. 72) vs. 3. 38 (0. 82), p =0. 046). In addition, participants’ improvements in attitude were statistically significant towards umbilical cord blood donation (3. 27 (1. 13) vs. 2. 85 (0. 10), p =0. 049), but not for bone marrow donation (3. 10 (1. 23) vs. 2. 81 (0. 94), p =0. 212). Participants’ negative attitudes regarding religious controversies surrounding SCs did not improve as the pre-intervention mean significantly decreased from 1. 88 (1. 10) to 1. 13 (1. 30), ( p =0. 003). However, similar reductions reported in attitude mean scores related to ethical controversies surrounding SCs (1. 13 (1. 20) post-intervention vs. 1. 29 (1. 09) pre-intervention, p =0. 420) and preserving umbilical cord blood in a private bank (2. 35 (1. 52) post-intervention vs. 2. 63 (1. 20) pre-intervention, p =0. 322) but they were not statistically significant. 10. 7717/peerj. 12824/table-4 Table 4 Pre- and post-educational intervention mean attitude scores and differences ( n = 48). Pre- and post-educational intervention mean scores of students’ attitudes regarding stem cells Differences between pre-and post-educational interventions Statements Score M (SD) Min–Max MD 95% CI p-value Lower Upper 1- I am interested in expanding my knowledge about stem cells (positive). Pre 3. 29 (0. 92) 0–4 0. 48 0. 22 0. 74 0. 001 * Post 3. 77 (0. 43) 3–4 2- Stem cell education should be integrated into medical college curricula (positive). Pre 2. 83 (0. 10) 0–4 0. 52 0. 13 0. 91 0. 010 * Post 3. 35 (0. 93) 0–4 3- I would consider a well-structured program or training focusing on stem cell science (positive). Pre 2. 83 (0. 91) 0–4 0. 65 0. 38 0. 92 <0. 001 * Post 3. 48 (0. 68) 2–4 4- I think stem cell therapies give rise to ethical controversies (negative). Pre 1. 29 (1. 09) 0–4 −0. 16 −0. 58 0. 25 0. 420 Post 1. 13 (1. 20) 0–4 5- I think stem cell therapies give rise to religious controversies (negative). Pre 1. 88 (1. 10) 0–4 −0. 75 −1. 23 −0. 27 0. 003 Post 1. 13 (1. 30) 0–4 6- Government should spend money to support stem cell research (positive). Pre 3. 38 (0. 82) 1–4 0. 31 0. 01 0. 62 0. 046 * Post 3. 69 (0. 72) 0–4 7- Transitional process of taking stem cell therapy from the laboratory through clinical trials should be encouraged (positive). Pre 2. 83 (0. 93) 1–4 0. 44 0. 12 0. 76 0. 009 * Post 3. 27 (0. 84) 1–4 8- People should consider the donation of bone marrow for a public bank (positive). Pre 2. 81 (0. 94) 1–4 0. 29 −0. 17 0. 76 0. 212 Post 3. 10 (1. 23) 0–4 9- People should consider the donation of their babies’ umbilical cord blood for a public bank (positive). Pre 2. 85 (0. 10) 0–4 0. 42 0. 00 0. 83 0. 049 * Post 3. 27 (1. 13) 0–4 10- I am willing to pay money for preserving the umbilical cord blood of my baby in a private bank for later use if a therapeutic need arises (positive). Pre 2. 63 (1. 20) 0–4 −0. 28 −0. 82 0. 27 0. 322 Post 2. 35 (1. 52) 0–4 The total attitude score Pre 2. 66 (0. 56) 0. 19 0. 02 0. 38 0. 048 * Post 2. 85 (0. 53) Notes. Abbreviations: M, Mean; SD, Standard Deviation; Min, Minimum score; Max, Maximum score; CI, Confidence Interval; MD, Mean Difference; Pre, Pre-educational intervention; Post, Post-educational intervention * Significant at p < 0. 05 based on paired-samples t -test. The total attitude score is the sum of the scores of the ten statements for each participant divided by ten. Gender differences As shown in Table 5, male students at baseline scored higher knowledge levels in comparison with female students with regard to SC potential applications (2. 85 (0. 66) vs. 2. 28 (0. 85) respectively, p =0. 014) and SC research (1. 95 (0. 81) vs. 1. 39 (0. 95) respectively, p =0. 036). Accordingly, the total knowledge score of males was higher than females (2. 16 (0. 27) vs. 1. 95 (0. 30) respectively, p =0. 017). However, after the intervention, gender differences were not statistically significant. 10. 7717/peerj. 12824/table-5 Table 5 Gender differences in mean knowledge and attitude scores pre-and post- educational intervention. Score Pre-intervention differences between males and females ( n = 48) Post-intervention differences between males and females ( n = 48) Males (n = 32 ) M (SD) Females (n = 16 ) M (SD) p-value Males (n = 32 ) M (SD) Females (n = 16 ) M (SD) p-value Total score of stem cell basic knowledge 2. 15 (0. 32) 2. 12 (0. 24) 0. 734 3. 15 (0. 45) 2. 99 (0. 49) 0. 279 Total score of stem cell potential applications 2. 85 (0. 66) 2. 28 (0. 85) 0. 014 * 3. 52 (0. 58) 3. 35 (0. 61) 0. 369 Total score of stem cell therapeutic uses 1. 82 (0. 65) 1. 89 (0. 59) 0. 719 2. 50 (0. 86) 2. 35 (0. 68) 0. 571 Total score of stem cell research 1. 95 (0. 81) 1. 39 (0. 95) 0. 036 * 3. 30 (0. 62) 3. 20 (0. 41) 0. 588 Total knowledge score 2. 16 (0. 27) 1. 95 (0. 30) 0. 017 * 3. 14 (0. 42) 3. 00 (0. 39) 0. 267 Total attitude score 2. 66 (0. 60) 2. 66 (0. 50) 1. 000 2. 81 (0. 52) 2. 92 (0. 55) 0. 517 Notes. Abbreviations: M, Mean; SD, Standard Deviation. * Significant at p < 0. 05 based on independent-samples t -test. Discussion The current study sheds light on the effectiveness of an interactive educational intervention in improving the knowledge and attitudes of medical students toward SCs, their therapeutic uses, and their potential research applications. The intervention course was conducted for six weeks, and different interactive teaching methods were used. The study results indicated that participants’ knowledge about SCs was insufficient in the pre-intervention phase as the mean scores for most knowledge domains and total knowledge were ≤ 2. These findings are concordant with previous studies that revealed poor knowledge regarding various aspects of stem cells banking, donation, and transplantation among the public, university students, and healthcare providers ( Azzazy & Mohamed, 2016 ; Kaya et al. , 2015 ; Lye et al. , 2015 ; Perlow, 2006 ; Suen et al. , 2011 ; Tuteja, Agarwal & Phadke, 2016 ). SC knowledge and attitude scores significantly improved following the intervention course. Post-intervention, participants were more interested in expanding their knowledge about SCs and considered well-structured programs or training courses as a successful approach to improve their understanding of SCs. The participants reported positive attitudes regarding the integration of SC education in medical college curricula after the intervention. This study provides a shred of landmark evidence from the Middle East and North Africa to implement the interactive learning approach in the SC teaching field. The excellent knowledge and attitude regarding SCs could be of great benefits not only to medical students but also to the overall health system as it will reflect on future healthcare providers being more informed and better guided to serve their patients with up-to-date information and improve the decision making power regarding SCs as an innovative method of therapy ( Perlow, 2006 ; Tork et al. , 2017 ; Tuteja, Agarwal & Phadke, 2016 ). This study could enhance medical curriculum development and teaching approaches and bridge the gap between basic sciences and clinical practice. As future health care leaders, medical students represent a source of information, or misinformation, which may influence patients’ behaviors and serve as a valuable source of information ( Davies et al. , 2002 ). This makes the medical school an ideal place to address information misconceptions and emphasize positive attitudes toward SC applications. Therefore, improvements in the region’s medical curricula should seriously consider interactive session models and introduce broader and more scientific resources for students in the healthcare field. This is especially true to follow-up on rapidly advancing scientific topics in the medical fields, where relying merely on available evidence from textbooks may introduce delays in transferring knowledge to medical students. Before conducting the intervention course, the enrolled students reported poor knowledge about SCs, with the lowest scores observed for SC research knowledge, therapeutic uses, and basic knowledge. However, their knowledge regarding the SC potential applications was relatively good. Following the intervention course, the students’ knowledge was significantly enhanced with better-reported scores, such as SC research knowledge. Significant improvements also spread to the other addressed knowledge domains regarding SCs, including SC basic knowledge, and that knowledge related to SC potential applications, and to a lesser extent, SC therapeutic uses’ knowledge. Previous educational interventions successfully increased the knowledge about SC transplantation and banking among medical, nursing, and law students and showed more positive attitudes toward SC donation following a particular intervention ( Azzazy & Mohamed, 2016 ; Kaya et al. , 2015 ). Innovative SC education using practical experiments to master SC culture and differentiation techniques were also reported to deepen medical students’ understanding of regenerative and translational medicine ( Jin et al. , 2018 ). After reviewing the educational interventions that were used in other studies to enhance the students’ knowledge regarding SCs, we found that their educational interventions were for a shorter duration than ours, and they did not use such interactive teaching methods ( Azzazy & Mohamed, 2016 ; Jin et al. , 2018 ; Kaya et al. , 2015 ). Thus, in our study, we were keen to provide a more comprehensive and detailed interactive teaching course for a longer duration that will cover more topics related to SC education, research, and potential applications, unproven SC therapies and tourism, and bioethics in order to design the interactive teaching course in an innovative way that will be more engaging to the medical students ( Azzazy & Mohamed, 2016 ; Jin et al. , 2018 ; Kaya et al. , 2015 ). Besides, study material developed by our research team could be adopted by other schools interested in establishing similar courses, and our interactive teaching courses could be integrated within curricula. In the current study, the most common sources of knowledge regarding SCs were lectures followed by social media. Social media has created an opportunity to disseminate information regarding unproven SC-based therapies directly to consumers to legitimize providers and their products by using solid emotional appeals such as patient testimonials ( Lyons, Salgaonkar & Flaherty, 2021 ). Other than social media, mass media, including newspapers, television, and radio, are considered a primary source of scientific communication to the public as it can significantly influence public attitudes toward controversial emerging technologies in regenerative medicine, such as the use of leftover blastocysts as a source for embryonic SCs and genome editing ( Sharpe, Di Pietro & Illes, 2016 ). Also, the portrayal of translational SC research in newspapers is highly optimistic and may foster unrealistic expectations regarding clinical translation speed ( Kamenova & Caulfield, 2015 ). Medical students should consider other sources for knowledge based on scientific evidence, such as medical journals and conferences. Unfortunately, none of the medical students in our study chose panel discussions as a source for SC knowledge, despite being considered a valuable way to trigger an exchange of viewpoints regarding ethical controversies surrounding SCs ( Arráez-Aybara et al. , 2018 ). Therefore, medical schools are invited to further invest in students’ knowledge about SCs by enhancing exposure to updated medical literature and medical conferences. This study indicates that the interactive teaching approach effectively improved the levels of knowledge and positive attitudes toward SCs. Furthermore, our interactive learning approach was sufficient to reduce gender gaps in SC knowledge scores, especially the scores related to SC potential applications and SC research since females were significantly less knowledgable than males in the pre-intervention course; however, the gender differences were reduced after the intervention course and became statistically insignificant. Despite the evolving amount of literature indicating the merits of the interactive learning approach, there is still a large gap between educational research and what happens in practice. The traditional didactic lectures still predominate in university classrooms ( Liebert et al. , 2016 ; Merideno, Antón & Prada, 2015 ; Saroyan & Snell, 1997 ). Previous studies that compared active learning to the traditional approach using passive learning indicated that interactive teaching methods generally result in longer retention of material, superior problem-solving and higher-thinking skills, more positive attitudes, and higher motivation the students to learn ( DIJK & JOCHEMS, 2002 ; Dodiya, Vadasmiya & Diwan, 2019 ; McKeachie, 1990 ; Wheijen, Jones & Rainer, 2002 ). Also, the students found such classes more fun and less tedious, and they were more satisfied with this teaching approach ( DIJK & JOCHEMS, 2002 ). Miller et al. reported a statistically significant higher average of students’ performance on exams using engaging lectures compared with traditional didactic lectures ( Miller, McNear & Metz, 2013 ). Also, the authors observed increased effectiveness of lectures, decreased distractions, and increased students’ confidence with the material using interactive teaching methods ( Miller, McNear & Metz, 2013 ). Based on our observations while conducting this study and the enrolled students’ feedbacks, the experience of interactive teaching technique was interesting for both students and researchers, and many of the students were enthusiastic about more courses designed with this approach. However, many challenges facing the incorporation of interactive teaching methods, including the limited amount of scientific content that could be covered within the class time, the significant efforts of preparation required by the instructor to create the interactive activities, and the effects of this approach on time available for traditional lectures. Given the rapidly growing amount of knowledge and the unlimited access of students to the information through the internet and other technologies, it may be more important to teach students how to use information rather than learning specific facts. This objective could be achieved by implementation of interactive teaching methods. Thus, medical educators may need to shift the importance of concept over content. In our study, the participants showed relatively positive attitudes toward SCs before the intervention, and furthermore, their attitudes improved following the intervention even to a lower extent than that observed in knowledge domains. However, negative attitudes related to SC religious controversies actually worsened the post-intervention course. Complex social, legal, ethical, and religious issues arise when emerging biotechnology involves human subjects ( Al-Aqeel, 2005 ), especially in conservative societies. However, Islamic teachings have paid attention to disease prevention and health promotion, and it is crucial to focus more on increasing our understanding of how SC applications could advance the health of human beings to facilitate the adoption of these technologies ( Bouzenita, 2017 ). Within this context, future SC-related interventions should focus on incorporating religious leaders from the medical community to present their points of view related to scientific facts from ethical, moral, and religious perspectives ( Aksoy, 2005 ; Al-Tabba, Dajani & Al-Hussaini, 2020 ; Fadel, 2012 ). The current Statute regarding SC use in developing countries is still unclear and not updated; thus, experts of different political, religious, scientific, and medical aspects who are familiar with laws are invited to develop a more comprehensive juridical system ( Al-Tabba, Dajani & Al-Hussaini, 2020 ; Pourebrahim, Goldouzian & Ramezani, 2020 ). However, although the negative attitudes toward ethical controversies surrounding SC therapies worsened following the intervention; the change in mean attitude scores was not statistically significant. Ethical concerns may be tightly connected to religious concerns and can only be mitigated by openly discussing the lack of religious restrictions related to medical improvements. Notably, our findings regarding religious and ethical controversies call for incorporating bioethics into the medical curriculum when addressing SC-related topics as ethical concerns were reported to be the obstacle that have obscured the proper potential use of SCs for revolutionizing medicine and treatment options in the future ( Hug, 2005 ). Medical curricula need to be restructured to include SCs or other emerging technologies in biomedicine and include research and healthcare ethics ( Abdulrazeq et al. , 2019 ; Brass, 2009 ; Sarkadi & Schatten, 2012 ). Adopting new technologies for patient care is challenging since many ethical dilemmas surround it, and future physicians should be prepared to deal with such dilemmas when they arise ( Curley & Sharples, 2006 ). A few limitations should be mentioned. The sample size was relatively small, and the participants were selected from a single medical school, limiting our results’ generalizability. While response and enrolment rates were not optimal, they are considered sufficient among medical students. Data collection in this study was also limited to quantitative methods; utilizing a qualitative approach supported by quantitative methods would be recommended to provide a richer analysis of the phenomena. A parallel-group with no intervention was not utilized which may introduce testing effects and exacerbate the results. The same survey was utilized pre-and post-intervention that was self-reported by the participants, which might explain that the improvements in knowledge and attitudes scores could be by chance and not due to intervention effects. Finally, this study used an interactive educational approach over a six-week course to improve the students’ knowledge and attitudes toward the SC field without comparison with other learning approaches. It would be interesting for future studies to compare the effectiveness of this interactive educational approach with other types of learning, such as flipped classrooms and traditional dedicated lectures. Conclusions This study demonstrates the crucial role of the interactive teaching approach on medical students’ knowledge and attitudes toward SCs, therapeutic uses, and potential research applications. The study results have proven poor knowledge about SCs in the pre-intervention phase, with a significant improvement after the interactive intervention course. As SC utilization becomes more common across specialties, having pre-clinical and clinical curricula educate future physicians on SCs is necessary. Thus, the SC concept such as research skills and SC therapeutic uses should be incorporated into the medical school curriculum to overcome the current shortcomings in SC knowledge and update future physicians with evidence-based medical practice. The study findings also indicate the effectiveness of the intervention course in achieving more positive attitudes toward SCs by the medical students. In addition, the differences in knowledge regarding SCs between males and females have reduced after the intervention course. Thus, the study concluded that an interactive teaching approach might be feasible under the same teaching resources and students’ situations, and thus, it could be considered beneficial to the medical students regardless of gender. Educators can capitalize on the available opportunities to improve the areas of the current SC curricula. Although the interactive intervention course was as short as six weeks, the study outcomes were promising. The appropriate intervention methods should be further tested by their implementation on a large scale in most medical subjects. Also, medical educators and schools are calling for integrating new interactive teaching approaches to address the life sciences instead of traditional teaching methods. Further studies with a larger sample are recommended to evaluate the needed curriculum content development, practical teaching approaches, and the most effective practice matters. Moreover, developing educational programs considering social, ethical, legal, religious, and cultural issues are recommended. Supplemental Information 10. 7717/peerj. 12824/supp-1 Supplemental Information 1 The distributed questionnaire Click here for additional data file. 10. 7717/peerj. 12824/supp-2 Supplemental Information 2 The participants’ responses data pre- and post-intervention Click here for additional data file.

10. 7717/peerj. 13355

PeerJ

lncRNA MALAT1 mediates osteogenic differentiation of bone mesenchymal stem cells by sponging miR-129-5p

Background Bone mesenchymal stem cells (BMSCs) have good osteogenic differentiation potential and have become ideal seed cells in bone tissue engineering. However, the osteogenic differentiation ability of BMSCs gradually weakens with age, and the regulatory mechanism is unclear. Method We conducted a bioinformatics analysis, dual-luciferase reporter (DLR) experiment, and RNA binding protein immunoprecipitation (RIP) to explore the hub genes that may affect BMSC functions. Results The expression level of long non-coding RNA (lncRNA) metastasis-associated lung adenocarcinoma transcript 1 ( Malat1 ) was significantly higher in the BMSCs from elderly than younger mice, while miR-129-5p showed the opposite trend. The results of alkaline phosphatase staining, quantitative reverse transcription PCR and western blot experiments indicated that inhibiting the expression of Malat1 inhibits the osteogenic differentiation of BMSCs. This effect can be reversed by reducing the expression of miR-129-5p. Additionally, DLR and RIP experiments confirmed that Malat1 acts as a sponge for miR-129-5p. Conclusion Overall, our study findings indicated that lncRNA Malat1 may play a critical role in maintaining the osteoblast differentiation potential of BMSCs by sponging miR-129-5p.

Introduction Bone mesenchymal stem cells (BMSCs) are widely studied multipotent stem cells found in bone marrow. Their strong plasticity and differentiation potential make them useful for tissue repair ( Chu et al. , 2020 ; Fu et al. , 2019 ). During the aging process, the osteogenic differentiation ability of BMSCs gradually weakens, resulting in a decrease in osteoblasts and the accumulation of adipocytes, eventually leading to diseases such as senile osteoporosis. Several key factors have been identified in the osteogenesis and adipogenesis of BMSCs. Peroxisome proliferator-activated receptor gama ( PPARG ) and core-binding factor alpha1 ( CEBPA ) are the most important regulators of BMSC adipogenic differentiation. Runt-related transcription factor ( RUNX2 ) and Osterix play essential roles during osteogenic differentiation of BMSCs ( Qadir et al. , 2020 ). Although many studies have investigated the differentiation fate of BMSCs, the exact mechanism is still unknown. Long non-coding RNA (lncRNA) is a type of non-coding RNA longer than 200 nucleotides. With the continuous development of high-throughput sequencing technology, a growing number of studies have shown that lncRNA participates in gene expression regulation and cell differentiation ( Huynh et al. , 2017 ). For example, Zhuang et al. (2015) found that lncRNA MEG3 promotes osteogenic differentiation of BMSCs in patients with multiple myeloma by targeting the transcription of bone morphogenetic protein. LncRNAs are widely involved in chromosome silencing, chromatin modification, transcriptional activation, and transcriptional interference. Among these, competitive endogenous RNA (ceRNA) is one of the main mechanisms of lncRNA regulation. CeRNAs have miRNA response elements that enable them to bind target miRNAs, thereby regulating their quantity and function. For instance, according to the findings of Yang et al. (2019a), lncRNA ORLNC1 functions as a ceRNA for miR-296 and regulates the balance between osteogenesis and adipogenesis of BMSCs. The study also found that overexpression of ORLNC1 led to decreases in the amount and density of trabecular bone in mouse femurs ( Yang et al. , 2019a ). MicroRNA, another representative non-coding RNA of 18–25 nucleotides in length ( Bartel, 2009 ), is also involved in Wnt/β-catenin and TGF-β/BMP/Smad signaling pathways to control osteogenesis ( Zhang et al. , 2021 ). This current study used bioinformatics analysis to investigate the differential expressed non-coding RNAs in BMSCs from middle-aged and elderly adults, finding lncRNA Malat1 highly expressed in the elderly group. Previous research found that Malat1 functioned as a miRNA decoy for miR-30 and thus influenced the cell fate determination of adipose-derived mesenchymal stem cells ( Yi, Liu & Xiao, 2019 ). Furthermore, Malat1 was also determined to promote osteogenesis differentiation and possibly act as an inhibitory regulator in steroid-induced avascular necrosis of the femoral head by sponging miR-214 (10). However, the role of Malat1 in regulating the osteogenic differentiation of BMSCs is largely unknown. In this research, we found that Malat1 promoted the osteogenic differentiation of BMSCs, and its function was related to miR-129-5p. These results contribute to a better understanding of the possible mechanism of stem cell differentiation and advance the use of stem cells in tissue engineering. Materials and Methods GEO data collection Human expression profiles (accession number: GSE35955 ) were downloaded from the NCBI GEO ( Edgar, 2002 ; Oghumu et al. , 2016 ), including the lncRNA and mRNA expression data of BMSCs derived from five middle-aged (42–67 year old) and four elderly (78–89 year old) donors. The data were then analyzed using the GPL570 Affymetrix Human Genome U133 Plus 2. 0 Array. The associated web link is displayed in File S1. Data quality control The quality of the nine samples was investigated using the R package AffyPLM ( Heber & Sick, 2006 ). The boxplot representations of the relative log expression and normalized unscaled standard errors are two commonly used chip quality control methods. Data pre-processing The transcripts were divided into two categories (1, 205 lncRNAs and 16, 418 mRNAs), utilizing the comprehensive gene annotation files downloaded from GENCODE ( Yin et al. , 2020 ) ( File S1 ). After the raw expression data were normalized, the limma package ( Ritchie et al. , 2015 ) was used to screen differentially expressed genes (DEGs, i. e. , differentially expressed lncRNAs (DELs) and differentially expressed mRNAs (DEMs)) between samples isolated from elderly and middle-aged groups. The screening criteria were |log2 (fold change)| >1 and adjusted P < 0. 05. lncRNA-miRNA-mRNA triples prediction and functional enrichment analysis The target miRNAs of DELs were collected using a putative microRNA target database miRcode ( Jeggari, Marks & Larsson, 2012 ) and the results were used to predict downstream mRNAs. The interactions between miRNA and mRNA were supported by at least two of the following databases–miRDB ( Wong & Wang, 2015 ), TargetScan 7. 0 ( Agarwal et al. , 2015 ) and miRTarBase 6. 0 ( Hsu et al. , 2011 ). The intersection of mRNA targets and DEMs were used for further analysis. A DEL-miRNA-DEM network was built with Cytoscape 3. 7. 1 ( Shannon et al. , 2003 ) software. Nodes with a high node degree (>5 connections), which was one of the network topological features, were considered to be the significant nodes in the network ( Han et al. , 2004 ). Using the clusterProfiler package ( Yu et al. , 2012 ), Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of the involved DEMs were performed to further explore the potential biological functions of lncRNAs. The thresholds of the functional categories were P < 0. 05 and a Benjamini–Hochberg corrected p -value < 0. 05. Animals and cell cultivation This study was approved by the Ethics Committee of Tongji University affiliated Stomatological Hospital (Approval sequence: -DW-015). Female C57BL/6 mice were purchased from the Model Animal Research Center of Nanjing University (China). On arrival, all mice were housed in stainless plastic cages and were maintained under controlled environmental conditions throughout the study: 21–23 °C, with a relative humidity of 60 ± 10%, 12:12 h light–dark cycles (lights on from 0700 to 1900 h), and with commercial diet and water available. Cages were changed twice per week. When the mice reached a set age ( i. e. , 6 weeks or 12 months), they received a single intraperitoneal injection of Zoletil in distilled water at a dose of 50 mg/kg and were sacrificed by cervical dislocation. We used six mice for each group as biological replicates. Primary BMSCs derived from mouse femurs were cultured in the above-described medium at 37 °C in a 5% CO 2 incubator. When the cell density reached 70–80%, osteoblast differentiation was induced in osteogenic medium (OM) which contained 100 nM dexamethasone, 100 nM ascorbic acid and 10 mM β-glycerophosphate (Sigma-Aldrich, USA). The medium was refreshed every three days and cells were harvested at the indicated times following treatments. During the experiment and routine feeding, animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health ( Anonymous, 2011 ). This study was approved by the Ethics Committee of Tongji University Affiliated Stomatological Hospital. The 293T cells were purchased from the Cell Bank of China and cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% (v/v) fetal bovine serum (Gibco Life Technologies, USA) and, 1% (v/v) penicillin/streptomycin (HyClone, USA). Alizarin Red staining and Alkaline phosphatase staining BMSCs maintained in osteogenic induction medium for seven and twenty-one days were stained with Alizarin Red (Nanjing Jiancheng Bioengineering Institute, China) and BCIP/NBT alkaline phosphatase staining kit (Beyotime, China), respectively, to detect the osteogenic capacity of mouse BMSCs. Quantitative analysis of ALP activity was determined by an Enhanced BCA Protein Assay Kit (Beyotime, China) combined with an Alkaline phosphatase assay kit (Jiancheng, Nanjing, China). Oil Red O staining BMSCs were cultured in adipogenic induction medium (Cyagen, USA) for twenty-eight days and stained with Oil Red O (Sigma-Aldrich, USA) according to the provided instructions to detect the adipogenic capacity of mouse BMSCs. Vectors and transfection The miR-129-5p inhibitor/mimic vector and Antisense oligonucleotide (ASO) -MALAT1 were designed and synthesized by Ribobio Co. , Ltd. Cell transfection was conducted with the help of a transfection Kit (Ribobio, China) and in accordance with the manufacturer’s instructions. Twenty-four hours post-transfection, quantitative reverse transcription PCR (qRT-PCR) was performed to confirm transfection efficiency, and samples were collected for further investigation. qRT-PCR Total RNA was extracted from BMSCs with RNAiso Plus (Takara, Japan) as per the manufacturer’s instructions. 1, 000 ng of total RNA was reverse-transcribed to complementary DNA using Takara PrimeScript RT reagent kit (Takara, Japan) and following the manufacturer’s protocol. MiRNA were reverse-transcribed using the transcriptor first strand cDNA synthesis kit (Roche, German). Equal amounts of complementary DNA were added to a 10 µL reaction mixture using Universal SYBR Green Master (Roche, German) and were subsequently used for real-time PCR reactions on a LightCycler96 Instrument (Roche, German). The primer sequences of alkaline phosphatase ( Alp ), runt-related transcription factor 2( Runx2 ), osteocalcin ( Ocn ), U6, osteopontin ( Opn ) and D-glyceraldehde-3-phosphate dehydrogenase ( Gapdh ) are listed in Table 1. The sequences of miR-129-5p and lncRNA MALAT1 were synthesized by Ribobio (Guangdong, China). U6 was used as an internal control for miRNA, while Gapdh was used as an internal control of mRNA and lncRNA. All experiments were repeated in triplicate, and the relative RNA expression rates were calculated using the 2 −ΔΔCt method. 10. 7717/peerj. 13355/table-1 Table 1 A list of primers. Gene and primer type Primer sequences STAT1 Forward primer 5′-TCACAGTGGTTCGAGCTTCAG-3′ Reverse primer 5′-GCAAACGAGACATCATAGGCA-3′ Alp Forward primer 5′-GCCCTCTCCAAGACATATA-3′ Reverse primer 5′-CCATGATCACGTCGATATCC-3′ Runx2 Forward primer 5′-CAAAGCCAGAGTGGACCCTT-3′ Reverse primer 5′-AGACTCATCCATTCTGCCGC-3′ Ocn Forward primer 5′-CTGACCTCACAGATCCCAAGC-3′ Reverse primer 5′-TGGTCTGATAGCTCGTCACAAG-3′ Opn Forward primer 5′-GATCAGGACAACAACGGAAAGG-3′ Reverse primer 5′-GCTGGCTTTGGAACTTGCTT-3′ Gapdh Forward primer 5′-AGGTCGGTGTGAACGGATTTG-3′ Reverse primer 5′-TGTAGACCATGTAGTTGAGGTCA-3′ U6 Forward primer 5′-GCTTCGGCAGCACATATACTAAAAT-3′ Reverse primer 5′-CGCTTCACGAATTTGCGTGTCAT-3′ Western blotting BMSCs were lysed in RIPA lysis buffer containing protease/phosphatase inhibiting cocktails (Beyotime, China). Total protein concentration was detected by the Enhanced BCA Protein Assay Kit (Beyotime, China). Equivalent amounts of protein lysates were loaded onto 12. 5% SDS-PAGE gels (Beyotime, China) and then transferred to polyvinylidene fluoride membranes (Sangon Biotech, China). Membranes were blocked with 5% non-fat milk and hybridized with primary antibodies against signal transducer and activator of transcription 1 (STAT1) (1:1000; Abcam, UK), glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (1:2000; BIOSS, China), lamin b1 (LAMIN B1) (1:10000; Abcam, UK) and RUNX2 (1:2000; Abcam, UK). GAPDH and LAMIN B1 were used as internal controls. The membrane was then hybridized with secondary antibodies for two hours. The blots were visualized and quantified using Odyssey CLx (LICOR, USA). Dual-luciferase report assay The lncRNA Malat1 and Stat1 sequences containing the predicted binding sites of miR-129-5p were amplified by PCR and inserted into the pmirGLO vector (Promega, USA). The resultant constructs were tagged as lncRNA Malat1 -wild-type (lncRNA Malat1 -wt)/ lncRNA Malat1 -mutated-type (lncRNA Malat1 -mut) and Stat1 -wild-type ( Stat1 -wt)/ Stat1 -mutated-type ( Stat1 -mut). The Malat1 -mut and Stat1 -mut were generated by site-directed mutagenesis, replacing the first six ribonucleotides of the miR-129-5p complementary sequence. The 293T cells were seeded into 6-well plates (1. 0 × 10 5 ) and co-transfected with miR-129-5p negative control (NC)/mimic and Malat1 -wt/mut or Stat1 -wt/mut using Lipofectamine 3000 (Invitrogen, USA) following the manufacturer’s instructions. Twenty-four hours after co-transfection at 37 °C, cells were harvested for luciferase assay using the dual-luciferase reporter (DLR) Assay System (Promega, USA). Transfection was performed three times. Nuclear/cytoplasmic RNA and Nuclear/cytoplasmic protein extraction Nuclear, cytoplasmic and total RNA were extracted by the PARIS™ Kit (Life technologies, USA). Simply put, cells were harvested and rinsed with phosphate buffered saline, and then centrifuged at 1, 2000 rpm for five minutes. The cell sediments were mixed with 300 ul of cell fractionation buffer. Following centrifugation, cells were placed on ice for ten minutes. The solutions were then centrifuged at 500 g for five minutes (at 4 °C) and the supernatants containing cytoplasmic lysate were collected. The remaining sediments which contained nuclear protein/RNA were collected and mixed with 300 ul of cell disruption buffer and placed on ice. The cytoplasmic lysate and nuclear lysate fractions were used for RNA extraction, with the remainder being used for protein extraction. The lysis buffer used for RNA extraction was mixed with 300 µl of 2X lysis/binding solution and the same volume of ethanol. After being centrifugated, the mixture was transferred into a filter cartridge/collection tube assembly. Following three washing steps, the RNA was finally eluted with Elution Solution which was preheated to 95–100 °C. The ratio of nuclear gene expression to cytoplasmic expression of each gene is used for comparison (2 −ΔCt method) ( Khudayberdiev et al. , 2013 ). Micro-computed tomography(micro-CT) Bone mass at the distal femur was assessed using micro-computed tomography (micro-CT) imaging (µCT50, Scanco Medical). After reconstructing image slices, the region of interest was manually selected in the marrow cavity. Ultra-high-resolution images (18 µm) of the specimens were obtained and relevant bone morphometric parameters, including BMD (mg/cc) and bone volume relative to total volume (BV/TV, %), were assessed. Immunofluorescence staining The cells were fixed with 4% Paraformaldehyde solution. Rabbit anti-mouse RUNX2 (Affinity Biosciences, OH, USA) was diluted 1:100 in PBS, and hybridized overnight at 4 °C. The samples were subsequently loaded with donkey anti-rabbit IgG immunofluorescence secondary antibodies (Beyotime, Shanghai, China) at 1:1000 dilution in PBS. The samples were also mounted with DAPI (1:10000, Sigma, US). Images were obtained under a confocal microscope (Leica Microsystems, Geremany). Fluorescence in situ hybridization (FISH) assay Fluorescence-labeled probes for lncRNA Malat1, 18S rRNA, and U6 RNA were designed and synthesized by Ribobio Co. , Ltd. Fluorescence in situ hybridization staining was performed using a Ribo™ Fluorescent in situ hybridization Kit (RiboBio, Guangzhou, China). Images were acquired on an ECLIPSE Ts2R laser-scanning confocal microscope (Nikon, Japan). RNA immunoprecipitation (RIP) RIP was carried out by an EZ-Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore, USA) in accordance with manufacturer guidelines. BMSCs were lysed using RIP lysis buffer and aliquoted for analysis. 100 µL of the lysate was mixed with RIP buffer containing beads conjugated with Argonaute-2 (Ago2) (Abcam, UK) antibody and IgG (Abcam, UK) antibody, followed by the addition of proteinase K buffer. The ensuing protein samples and purified target RNA were used for WB and qRT-PCR assays, respectively. Statistical analysis Statistical analyses were performed using GraphPad Prism 8. 0 software. Unless indicated otherwise, the data are presented as means ± standard deviations (SDs). Student’s t tests were used to compare results, with P < 0. 05 indicating a statistically significant difference. One-way analysis of variance was used for multiple comparisons. Results Data quality control We first conducted a data quality assessment. Regression analysis of raw data was performed using the affyPLM package, which provided us with relative log expression (RLE) boxplots, normalized unscaled standard errors (NUSE) boxplots, and RNA degradation plots ( Heber & Sick, 2006 ). The RLE boxplot reflects the consistent trend of gene expression, which is defined as the logarithm of the expression value of a probe set in a sample divided by the median expression value of the probe set in all samples. The RLE plot revealed that the levels of gene expression in GSE35955 were relatively consistent, and the median value was close to 0 ( Fig. 1A ). This indicates that the sample quality of the chip data was reliable. NUSE is a more sensitive quality control method than RLE. NUSE is defined as the standard deviation of the perfect-match value of a probe set in a sample divided by the median standard deviation of the perfect-match value of the probe set in each sample. Consistent with the results of RLE, the nine samples’ NUSE values were approximately one, indicating their standard deviations were very close ( Fig. 1B ). Next, we normalized the data and annotated the probe names. A total of 16, 418 mRNAs and 1, 205 lncRNAs were identified in the microarray data through the human comprehensive gene annotation file ( File S1 ). 10. 7717/peerj. 13355/fig-1 Figure 1 Data quality assessment of the raw data sets. (A) Boxplot representation of the relative log expression (RLE). (B) Boxplot representation of the normalized unscaled standard errors (NUSE). The screening results of DELs and DEMs To screen for genes that might function in BMSCs, we performed a differentially expressed gene analysis using the limma package ( Ritchie et al. , 2015 ). The screening criteria were |log2 (fold change)| >1 and adjusted P < 0. 05. As a result, 64 differentially expressed lncRNAs (DELs) and 651 differentially expressed mRNAs (DEMs) were identified between the samples isolated from middle-aged and elderly groups ( File S2 ). The expression of these DELs and all DEGs are visualized in a heatmap ( Fig. 2 ) and a volcano plot ( Fig. 3 ), respectively. MiRNAs associated with DELs were predicted through the miRcode database. At the same time, the mRNA targets of these miRNAs were predicted through three databases ( i. e. , miRTarBase ( Hsu et al. , 2011 ), TargetScan ( Agarwal et al. , 2015 ) and miRDB ( Wong & Wang, 2015 )) and were intersected with the previously obtained DEMs. Consequently, a total of 706 reliable miRNA-mRNA pairs and 347 predicted lncRNA-miRNA pairs (including 28 lncRNAs, 50 miRNAs and 281 mRNAs) were obtained for further analysis ( File S3 ). 10. 7717/peerj. 13355/fig-2 Figure 2 Heatmap of differentially expressed lncRNAs between BMSCs from middle-aged and old donors. The horizontal axis shows the names of nine samples. The vertical axis presents the gene names. 10. 7717/peerj. 13355/fig-3 Figure 3 Volcano plot of all differentially expressed genes between BMSCs from middle-aged and old donors. FC are fold-change. Downregulated genes are green and upregulated genes are red. Construction of the ceRNA network and Functional enrichment analysis Using Cytoscape software ( Shannon et al. , 2003 ), a DEL-miRNA-DEM network was built based on the miRNA-DEM and DEL-miRNA pairs ( Fig. 4A ). The network consisted of 28 DEL nodes, 50 miRNA nodes and 281 DEM nodes. 10. 7717/peerj. 13355/fig-4 Figure 4 The lncRNA associated ceRNA network and barplots of function enrichment analyses. (A) The lncRNA-miRNA-mRNA ceRNA network. The parallelograms represent lncRNAs, the ellipses represent mRNAs, and the triangles represent miRNAs. (B) The top 10 most significant Gene ontology terms. (C) The top 10 most significant pathway terms. Functional enrichment analyses based on DEMs, including GO analysis and KEGG analysis, may provide an approach to understanding the potential functions of DELs. We constructed a GO interaction network using the BiNGO tool ( File S3 ). It uncovered ten significantly enriched terms ( Fig. 4B and File S4 ), the top three of which were protein binding, focal adhesion and positive regulation of transcription from RNA polymerase II promoter. Twelve pathways were obtained from KEGG, as partially shown in Fig. 4C, including the terms of the AMPK, FoxO, and TGF-β signaling pathways ( File S5 ). Interestingly, the AMPK ( Wang et al. , 2013 ; Zhou et al. , 2019 ) and TGF-β signaling pathways ( Zhang et al. , 2017 ) were closely related to osteogenic differentiation of BMSCs. Topological analysis of the ceRNA network We computed the node degrees to identify the hub genes in the ceRNA network. In accordance with previous research ( Han et al. , 2004 ), we set the screening criterion as node degree ≥5. A total of 113 nodes were screened as hub genes, including 24 DELs, 49 miRNAs and 40 DEMs ( Table 2 ). We found that lncRNA MALAT1 had higher node degrees, suggesting a potential role for BMSCs among elder donors. 10. 7717/peerj. 13355/table-2 Table 2 The list of differentially expressed genes(node degree > 5). Number Gene type Gene name Node degree 1 miRNA hsa-miR-17-5p 50 2 miRNA hsa-miR-129-5p 48 3 miRNA hsa-miR-24-3p 47 4 miRNA hsa-miR-3619-5p 40 5 miRNA hsa-miR-761 39 6 miRNA hsa-miR-429 39 7 miRNA hsa-miR-1297 38 8 lncRNA MALAT1 35 9 miRNA hsa-miR-20b-5p 33 10 miRNA hsa-miR-137 32 11 miRNA hsa-miR-206 31 12 miRNA hsa-miR-23b-3p 31 13 miRNA hsa-miR-613 30 14 miRNA hsa-miR-363-3p 27 15 miRNA hsa-miR-125b-5p 27 16 miRNA hsa-miR-125a-5p 26 17 miRNA hsa-miR-107 26 18 miRNA hsa-miR-27a-3p 26 19 lncRNA DIRC3 24 20 miRNA hsa-miR-217 24 21 miRNA hsa-miR-142-3p 23 22 miRNA hsa-miR-33a-3p 22 23 miRNA hsa-miR-507 22 24 miRNA hsa-miR-338-3p 22 25 miRNA hsa-miR-216b-5p 21 26 miRNA hsa-miR-135a-5p 20 27 miRNA hsa-miR-375 20 28 lncRNA AC008088 19 29 lncRNA COX10-AS1 18 30 miRNA hsa-miR-876-3p 18 31 miRNA hsa-miR-139-5p 17 32 miRNA hsa-miR-301b-3p 17 33 lncRNA AC010145 17 34 lncRNA RRN3P2 17 35 lncRNA LINC00520 16 36 lncRNA JPX 16 37 lncRNA AC017002 16 38 lncRNA C20orf197 16 39 miRNA hsa-miR-22-3p 16 40 miRNA hsa-miR-1244 16 41 lncRNA AP000462 15 42 lncRNA RUSC1-AS1 15 43 miRNA hsa-miR-449c-5p 15 44 miRNA hsa-miR-490-3p 15 45 miRNA hsa-miR-508-3p 15 46 lncRNA AC106801 14 47 lncRNA CECR3 14 48 miRNA hsa-miR-193a-3p 14 49 miRNA hsa-miR-140-5p 14 50 miRNA hsa-miR-212-3p 14 51 lncRNA LINC00269 13 52 lncRNA C22orf24 13 53 miRNA hsa-miR-590-5p 12 54 miRNA hsa-miR-10a-5p 12 55 lncRNA TDRG1 11 56 miRNA hsa-miR-455-5p 11 57 miRNA hsa-miR-4262 11 58 lncRNA ITPKB-IT1 10 59 miRNA hsa-miR-146b-5p 10 60 mRNA BACH2 9 61 miRNA hsa-miR-425-5p 9 62 miRNA hsa-miR-4465 9 63 mRNA SOWAHC 8 64 mRNA KCTD15 8 65 mRNA CNN3 8 66 mRNA PRDM1 8 67 lncRNA AP000347 8 68 miRNA hsa-miR-4500 8 69 miRNA hsa-miR-4458 8 70 miRNA hsa-miR-518a-3p 8 71 mRNA KIAA0232 7 72 mRNA NXT2 7 73 mRNA ENPP4 7 74 mRNA FLOT2 7 75 mRNA CDC42SE1 7 76 mRNA MKNK2 7 77 mRNA KLF13 7 78 mRNA TMEM170B 7 79 lncRNA AC073321 7 80 lncRNA LINC00208 7 81 mRNA HIPK1 6 82 mRNA E2F5 6 83 mRNA PDPK1 6 84 mRNA DDIT4 6 85 mRNA ETV1 6 86 mRNA ANKRD12 6 87 mRNA CAPRIN2 6 88 mRNA UHMK1 6 89 mRNA S1PR1 6 90 mRNA BNIP3L 6 91 mRNA JMY 6 92 mRNA NBEA 6 93 mRNA STX16 6 94 mRNA STRN3 6 95 mRNA VLDLR 6 96 lncRNA KCNQ1-AS1 6 97 miRNA hsa-miR-4735-3p 6 98 mRNA GPC4 5 99 mRNA FZD3 5 100 mRNA ZNF516 5 101 mRNA PKIA 5 102 mRNA TULP4 5 103 mRNA TAF1D 5 104 mRNA PPP3R1 5 105 mRNA ZNF281 5 106 mRNA EP300 5 107 mRNA SKI 5 108 mRNA MET 5 109 mRNA ATP6V1A 5 110 lncRNA DSCR10 5 111 miRNA hsa-miR-3666 5 112 miRNA hsa-miR-4295 5 113 lncRNA AL022341 5 lncRNA Malat1 was upregulated but its target miR-129-5p was downregulated in BMSCs from old mice compared to young ones To verify the predicted results of the bioinformatics analysis, we isolated the femurs of young (6 weeks) and old mice (12 weeks). The micro-CT images of the femurs indicated that bone mass was lower in old than young mice ( Fig. 5A ). The old mice consistently displayed a lower bone volume per tissue volume (BV/TV) ( Fig. 5B ). We further extracted BMSCs from mice femurs, which have a long filamentous structure ( Fig. 5C ). Alizarin Red and Oil Red O staining indicated that the BMSCs were able to differentiate into osteoblasts and adipocytes under corresponding induction conditions ( Fig. 5C ). Moreover, we found that the expression of Malat1 was significantly up-regulated in the BMSCs derived from femurs of aged mice ( Fig. 5D ), suggesting that Malat1 may play an important role. Based on previously obtained lncRNA-miRNA pairs, we also examined the expression of miRNAs with potential binding sequences to Malat1. As a result, miR-124a-5p, miR-125b-5p, miR-20b-5p, and miR-24-3p were actually upregulated in aging mice-derived BMSCs, with only miR-17-5p and miR-129-5p showing opposite trends ( Fig. 5D ). As mentioned above, ceRNAs are capable of reducing the abundance of miRNAs, so these two down-regulated genes were the potential targets of lncRNA Malat1. However, we found no apparent binding between Malat1 and miR-17-5p ( File S6 ). Thus, we decided to explore the potential relationship between Malat1 and miR-129-5p. According to the obtained lncRNA-miRNA-mRNA triples ( File S3 ), Malat1 / miR-129-5p / Stat1 caught our attention. Stat1 is regulated by miR-129-5p and involved in BMSCs osteoblast differentiation ( Xiao et al. , 2016 ). Moreover, our previous study also indicated that Stat1 is an important transcription factor, which might contribute to the aging-related changes in BMSCs ( Wu et al. , 2019 ). 10. 7717/peerj. 13355/fig-5 Figure 5 The expression of LncRNA Malat1 in BMSCs from young and aged mice. (A)Representative micro-CT images showing the midshaft architectures of femurs from young and aged mice. N = 3. Scale bar, 1 mm. (B) Bone histomorphometric analysis of BMD and BV/TV in femurs from young and aged mice. N = 3. The mean ± s. e. m. is shown. (C) BMSCs showed a typical cobblestone-like morphology (a). Differentiation potential of BMSCs assessed by Alizarin Red staining (b) and Oil Red O (c). (D) The expression blot of lncRNA Malat1 and targets miRNA. At least three animals in each group were analysed, each experiment was repeated twice, and representative images are shown. To determine the subcellular localization of Malat1, we performed a FISH assay and cytoplasmic/nuclear fractionation with BMSCs. Surprisingly, the amount of Malat1 observed in the nucleus was higher than in the cytoplasm ( Figs. 6A and 6B ). However, several studies ( Huang et al. , 2020 ; Yang et al. , 2019b ; Li, 2022 ) have proved that Malat1 had a potential ceRNA regulatory mechanism in BMSCs, so Malat1 expressed in the nucleus might regulate the biological function of BMSCs through other mechanisms. 10. 7717/peerj. 13355/fig-6 Figure 6 LncRNA Malat1 was primarily localized in the nucleus. (A) FISH analysis of Malat1 in BMSCs (The nuclei were stained with DAPI, and 18S rRNA was used as a cytoplasmic marker). (B) Cell nuclear/cytoplasmic fractionation and qRT-PCR showed the cellular districution of lncRNA Malat1 in BMSCs ( U6 and Gapdh were used as separation quality standards and endogenous controls). Three independent experiments were performed for qRT-PCR assays. lncRNA Malat1 was upregulated during osteoblast differentiation of BMSCs Many important regulatory factors are involved in the process of osteoblast differentiation of BMSCs. Runx2 is essential for osteoprogenitor cell development and is an early transcription factor that determines the osteoblast differentiation of BMSCs ( Kawane et al. , 2018 ; Komor, 2010 ). When osteoprogenitor cells differentiate into pre-osteoblasts, the intracellular alkaline phosphatase (ALP) activity increases significantly, which is another marker for the early stage of osteoblast differentiation ( Aubin, 2001 ). ALP can promote bone mineralization, and reflect the degree of differentiation and functional status of osteoblasts ( Capulli, Paone & Rucci, 2014 ). With the further maturation of pre-osteoblasts, cells begin to produce osteopontin (OPN), bone sialoprotein (BSP), osteocalcin (OCN) and type 1 collagen, which are essential for bone formation and mineralization ( Zoch, Clemens & Riddle, 2016 ). Among them, OCN is a hormone-like polypeptide produced and secreted by osteoblasts, which is considered as a marker for late-stage osteoblastogenesis ( Nakashima & De Crombrugghe, 2003 ). To further understand the role of lncRNA Malat1 in osteoblast differentiation of BMSCs, we examined the dynamic changes of gene expression during the osteogenic differentiation of BMSCs. We found that Malat1 expression gradually increased over time ( Fig. 7A ). The dynamic expression of osteoblastic markers Runx2, Alp and Ocn during osteogenesis are shown in Figs. 7B – 7D. The expression of Alp, an early osteogenic indicator, reached the highest peak after seven days of osteogenic induction, and then gradually decreased. However, another early osteogenic marker, Runx2, continued to show an upregulated trend after seven days. At the same time, the expression level of Ocn, an indicator of late-stage osteogenesis, increased with induction time. In view of the positive correlation between Malat1 expression and the expression of osteoblastic markers of BMSCs, we speculated that Malat1 may promote the osteogenic differentiation of BMSCs. 10. 7717/peerj. 13355/fig-7 Figure 7 Expression patterns of lncRNA Malat1 during osteoblast differentiation of BMSCs. (A) qRT-PCR analysis was used to detect the expression of lncRNA Malat1 during osteoblast differentiation of BMSCs at days 0, 3, 7, 14 and 21. RNA expression at the indicated time points was normalized to day 0. Gapdh was used as an internal control. (B) qRT-PCR detection shows the expression of osteoblastic marker Runx2 on selected days. Gapdh was used as internal control. (C) qRT-PCR detection shows the expression of osteoblastic marker Ocn on selected days. Gapdh was used as internal control. (D) qRT-PCR detection shows the expression of osteoblastic marker Alp on selected days. Gapdh was used as internal control. Data are presented as means ± SD. Asterisk indicates P < 0. 05, double asterisks indicate P < 0. 01, triple asterisks indicate P < 0. 001, quadruple asterisks indicate P < 0. 0001 vs. control. Runx2, runt-related transcription factor 2; Ocn, osteocalcin; Alp, alkaline phosphatase. Three independent experiments were performed for qRT-PCR assays. Inhibition of Malat1 suppressed osteogenic differentiation of BMSCs To determine whether Malat1 expression played a role in BMSCs osteogenic differentiation, ASO against Malat1 was constructed and transfected into BMSCs. Compared with non-transfected control cells and the NC vector group, Malat1 expression was significantly downregulated in the ASO group ( Fig. 8A ), while the expression of miR-129-5p was upregulated ( Fig. 8B ). As a target gene of miR-129-5p, the expression level of Stat1 decreased after the disturbance of Malat1 expression ( Fig. 8C ). The results of qRT-PCR also showed that ASO- Malat1 also inhibited the expression of osteoblastic markers Alp, Runx2 and Ocn ( Figs. 8D – 8G ). Consistent with these findings, the WB assay results indicated that the expression of RUNX2 and STAT1 in BMSCs decreased after treatment with ASO- Malat1 ( Figs. 8H – 8I ). These results preliminarily demonstrated that Malat1 is indispensable for the osteogenic differentiation function of BMSCs. Each group of cells was then cultured in osteogenic medium (OM) or general medium (GM) for seven days. Osteogenic induction significantly enhanced the activity of the ALP enzyme, while inhibition of lncRNA Malat1 suppressed this effect ( Figs. 8J – 8K ), suggesting that Malat1 played an important role in the early stage of osteogenic differentiation of BMSCs. Immunofluorescence staining showed that ASO- Malat1 reduced the expression of RUNX2 protein in BMSCs ( Fig. 9A ). As a transcription factor, Runx2 regulates the transcriptional activity by binding to the promoters or enhancers of the target genes, so the process of Runx2 translocation into the nucleus is crucial for osteoblast differentiation ( Nakashima et al. , 2002 ). In that case, we also extracted the nuclear/cytoplasmic protein to detect the expression level of RUNX2 in different regions. The content of RUNX2 protein in the nucleus as well as the cytoplasm was reduced, which further supports the effect of Malat1 on osteogenic differentiation ( Figs. 9B – 9C ). Similar to RUNX2, STAT1 translocates into the nucleus after being activated in the cytoplasm, and regulates the expression of target genes, including nuclear factor-kappa B, myelocytomatosis oncogene, and matrix metallopeptidase 9 ( Görlich et al. , 1995 ; ten Hoeve et al. , 2002 ). As shown in Fig. 9D, the total content of STAT1 in BMSCs was downregulated in BMSCs after treated with ASO. The protein expression of intranuclear STAT1 decreased in the ASO group, while the expression of intracytoplasmic STAT1 increased. This may result from reduced activated STAT1 or blocked STAT1 nuclear translocation. These results suggest that Malat1 supports osteoblast differentiation of BMSCs, but whether this effect is related to miR-129-5p and Stat1 remains unclear. 10. 7717/peerj. 13355/fig-8 Figure 8 Inhibition of lncRNA Malat1 suppressed osteogenic differentiation of BMSCs. (A) qRT-PCR detection shows expression of lncRNA Malat1 in BMSCs after transfection with siRNA against lncRNA Malat1 (ASO) or the negative control (NC) vector for 24 h. Gapdh was used as an internal control. Data are presented as means ± SD. (B) qRT-PCR detection shows expression of miR-129-5p in BMSCs after transfection with siRNA against lncRNA Malat1 (ASO) or the negative control (NC) vector for 24 h. U6 was used as an internal control. Data are presented as means ± SD. Relative mRNA expression of Stat1 (C), Alp (D), Runx2 (E), Ocn (F) and Opn (G) was measured by qRT-PCR at day 7 of OM induction. Gapdh was used for normalization. Data are presented as means ± SD. (H) The expression of STAT1 and RUNX2 was detected by Western blot analysis when lncRNA Malat1 was knocked down in BMSCs. (I) Protein expression of STAT1 and RUNX2 in BMSCs transfected with ASO or NC vector. Asterisk indicates P < 0. 05, double asterisks indicate P < 0. 01 vs. Blank. (J) Images of ALP staining in BMSCs after culture in GM or OM for 7 days. Three independent experiments were performed for all of the blots and qPCR assays, and representative images are shown. (K) Histograms show ALP activity by spectrophotometry. Data are presented as means ± SD. 10. 7717/peerj. 13355/fig-9 Figure 9 Inhibition of lncRNA Malat1 impeded the nuclear translocation in Runx2. (A) Immunofluorescence staining was applied to test the effects of siRNA against lncRNA Malat1 (ASO) on the protein levels of RUNX2. (B) Western blot assay was conducted to test protein levels of total, nuclear, cytoplasmic RUNX2 in BMSCs transfected with ASO or NC vector. Three independent experiments were performed, and representative images are shown. (C) Quantitative analysis of protein expression of RUNX2 in BMSCs transfected with ASO or NC vector. Double asterisks indicate P < 0. 01, triple asterisks indicate P < 0. 001, quadruple asterisks indicate P < 0. 0001 vs. blank group. (D) Quantitative analysis of protein expression of RUNX2 in BMSCs transfected with ASO or NC vector. An asterisk (*) indicates P < 0. 05 vs. blank group. 10. 7717/peerj. 13355/fig-10 Figure 10 Inhibition of miR-129-5p promoted differentiation of BMSCs. (A) qRT-PCR detection shows expression of miR-129-5p in BMSCs after transfection with miR-129-5p inhibitors or NC for 24 h. Gapdh was used as an internal control. Data are presented as means ± SD. (B) qRT-PCR detection shows expression of Malat1 in BMSCs after transfection with miR-129-5p inhibitors or NC for 24 h. U6 was used as an internal control. Data are presented as means ± SD. Relative mRNA expression of Stat1 (C), Alp (D), Runx2 (E), Ocn (F) and Opn (G) was measured by qRT-PCR at day 7 of OM induction. Gapdh was used for normalization. Data are presented as means ± SD. (H) The expression of STAT1 and RUNX2 was detected by Western blot analysis when miR-129-5p was knocked down in BMSCs. (I) Protein expression of STAT1 and RUNX2 in BMSCs transfected with miR-129-5p inhibitor or NC vector. An asterisk (*) indicates P < 0. 05, double asterisks indicate P < 0. 01 vs. control. (J) Images of ALP staining in BMSCs after culture in GM or OM for 7 days. Three independent experiments were performed for all of the blots and qPCR assays, and representative images are shown. (K) Histograms show ALP activity by spectrophotometry. Data are presented as means ± SD. 10. 7717/peerj. 13355/fig-11 Figure 11 Inhibition of miR-129-5p did not directly affect the nuclear translocation in Runx2. (A) Immunofluorescence staining was applied to test the effects of miR-129-5p inhibitor on the protein levels of Runx2. (B) Western blot assay was conducted to test protein levels of total, nuclear, cytoplasmic RUNX2 in BMSCs transfected with miR-129-5p inhibitor or NC vector. Three independent experiments were performed, and representative images are shown. Inhibition of miR-129-5p promoted osteogenic differentiation of BMSCs MiRNAs play an important role in mesenchymal stem cells. For example, overexpression of miR-485-5p can induce a senescence-like phenotype and proliferation inhibition of adipose tissue-derived mesenchymal stem cells ( Kim et al. , 2012 ). Another research indicated that age-related bone loss can be rescued by regulating miR-142-3p / Bmal1 / YAP signaling axis in BMSCs ( Cha et al. , 2022 ). To investigate the role of miR-129-5p in the osteogenic differentiation of BMSCs and its underlying mechanism, a miR-129-5p inhibitor was constructed and transfected into BMSCs to determine the effect of miR-129-5p on the osteogenic differentiation of BMSCs. The results showed an acceptable inhibition effect ( Fig. 10A ). In addition, with the inhibition of miR-129-5p, the relative mRNA level of Malat1 and Stat1 were significantly upregulated ( Figs. 10B – 10C ). Similarly, Runx2, Opn and Ocn expression were also upregulated by miR-129-5p inhibition ( Figs. 10D – 10G ), suggesting that the effect of miR-129-5p on the BMSC osteoblast differentiation is opposite to that of Malat1. The western blot results indicated that the protein amount of both STAT1 and RUNX2 increased in the experimental group ( Figs. 10H – 10I ). The ALP enzyme activity of BMSCs increased after osteogenic induction, and inhibition of miR-129-5p further enhanced this effect ( Figs. 10J – 10K ). This suggests that miR-129-5p has the function of promoting the osteogenic differentiation of BMSCs. The results of immunofluorescence staining were similar to those of WB, and miR-129-5p inhibitor significantly up-regulated the expression of RUNX2 ( Fig. 11A ). To clarify whether miR-129-5p regulates the osteogenic differentiation of BMSCs by affecting the nuclear translocation of RUNX2, we also performed nucleocytoplasmic separation experiments. The results showed that the content of RUNX2 in the nucleus increased significantly, while the cytoplasmic content decreased ( Figs. 11B – 11C ), suggesting that miR-129-5p may affect the osteogenic differentiation of BMSCs through Runx2 signaling pathways. The result of nuclear/cytoplasmic protein extraction assay also showed that the protein expression of intracytoplasmic STAT1 increased in the miR-129-5p inhibitor treated group ( Fig. 11D ). However, the expression level of nuclear STAT1 in the miR-129-5p inhibitor group showed no significant difference. These results suggest that miR-129-5p inhibit osteoblast differentiation of BMSCs, but whether this effect is related to Stat1 still needs further research. miR-129-5p is a potential target of lncRNA Malat1 Most miRNAs directly bind to the target gene sequence through the RNA-induced silencing complex. Therefore, we used an online database ( http://starbase. sysu. edu. cn/ ) to predict the potential binding site between miR-129-5p and Malat1, finding that Malat1 has a broadly conserved binding site with miR-129-5p. Next, we constructed an original version of lncRNA Malat1 and a mutated one with seven altered complementary nucleotides ( Fig. 12A ). The luciferase reporter gene plasmids containing the two sequences were co-transfected into 293T cells with miR-129-5p mimics or scramble mimics. The results showed that miR-129-5p mimics could significantly downregulate the luciferase activity of the wild-type Malat1 groups ( Fig. 12B ). However, this phenomenon was not observed in the two Malat1 -mutation groups, indicating that miR-129-5p specifically targets lncRNA Malat1. The result of the qRT-PCR assay further confirmed that Malat1 inhibition significantly enhanced miR-129-5p expression in BMSCs ( Fig. 12C ). 10. 7717/peerj. 13355/fig-12 Figure 12 miR-129-5p is a potential target of lncRNA Malat1. (A) Complementary bases between the sequences are labeled with red font. The sequence of the mutant lncRNA Malat1 construct is also shown as underlined. (B) Dual-luciferase reporter assay of 293T cells co-transfected with lncRNA Malat1 -wt, or lncRNA Malat1-mut and with miR-129-5p mimic or NC. (C) MiR-129-5p expression in BMSCs was measured by qRT-PCR following transfection with ASO for 24 h. Data are presented as means ± SD. Two independent experiments were performed. Potential association of Stat1 with miR-129-5p Although we found that interfering with miR-129-5p did not affect the expression level of Stat1 in BMSCs, we still examined the binding relationship between the two genes, to potentially verify the results of a previous study ( Xiao et al. , 2016 ). Following the previous study’s methods ( Xiao et al. , 2016 ), we assessed the possible interaction between miR-129-5p and Stat1, by first predicting the potential binding site of Stat1 using the Starbase database. However, the binding site was in the coding sequence (CDS) instead of the 3′UTR region. As described above, luciferase reporter gene plasmids containing two versions of Stat1 were constructed and co-transfected into 293T cells along with miR-129-5p mimics or scramble mimics ( Fig. 13A ). It appeared that miR-129-5p mimics also downregulated the luciferase activity of the wild-type Stat1 groups, while this phenomenon was not observed in both Stat1 -mutation groups ( Fig. 13B ). However, the degree of downregulation of luciferase activity in 293T cells in the wild-type Stat1 group after transfected with miR-129-5p was different from the previous study ( Xiao et al. , 2016 ). Conversely, qRT-PCR analyses indicated that overexpression of miR-129-5p suppressed Stat1 expression in BMSCs ( Fig. 13C ), while the protein level of STAT1 showed no significant difference ( Fig. 13D ). Overall, it is still uncertain whether Stat1 is a direct target of miR-129-5p. 10. 7717/peerj. 13355/fig-13 Figure 13 Potential association of Stat1 with miR-129-5p. . (A) Sequence alignment between miR-129-5p and Stat1. (B) Dual-luciferase reporter assay of 293T cells co-transfected with Stat1-wt or Stat1-mut and miR-129-5p mimic or NC. QRT-PCR (C) and western blot (D) detection show the expression of Stat1 in BMSCs following transfection with miR-129-5p mimic or NC for 24 h. Data are presented as means ± SD. Two independent experiments were performed. MiR-129-5p inhibitor reversed the inhibitory effect of ASO- Malat1 on osteogenic differentiation of BMSCs To prove that the Malat1 / miR-129-5p signal axis regulated the osteogenic differentiation ability of BMSCs through miR-129-5p, we co-transfected miR-129-5p mimics and ASO- Malat1 into BMSCs using osteogenic induction. The findings of the qRT-PCR assay indicated that inhibition of miR-129-5p reversed the inhibitory effect of ASO- Malat1 on osteogenic differentiation by promoting Alp, Opn, Ocn, and Runx2 expression ( Figs. 14A – 14D ). Consistently, BMSCs transfected with only ASO- Malat1 showed a decreased level of ALP activity, while the addition of miR-129-5p reversed the inhibitory effect ( Fig. 14E and 14F ). MiRNA degrades mRNA or hinders its translation by forming a so-called RNA-induced silencing complex (RISC) ( Klein et al. , 2017 ). Ago2 is the core component of RISC, linking miRNAs and their mRNA target sites ( Klein et al. , 2017 ). Therefore, immunopurification of Ago2 under suitable conditions can obtain mutually bound miRNA and mRNA ( Yang et al. , 2019b ). In the RIP quality control experiment, the anti-Ago2 group successfully pulled down some Ago2 protein ( Fig. 15A ). In contrast, the anti-IgG group had no obvious bands ( Fig. 15A ), suggesting that the quality of the RIP assay was reliable. The RIP assay showed that anti-Ago2, an important part of the RNA silencing inducing complex, significantly enriched Malat1 and miR-129-5p in contrast with anti-IgG, indicating an endogenous combination between miR-129-5p and lncRNA Malat1 ( Fig. 15B ). 10. 7717/peerj. 13355/fig-14 Figure 14 Inhibition of miR-129-5p reversed the inhibitory effect of ASO on osteogenic differentiation of BMSCs. (A–D) Relative mRNA expression of Alp (A), Runx2 (B), Ocn (C) and Opn (D) were measured by qRT-PCR at day 7 of OM induction. Gapdh was used for normalization. Data are presented as means ± SD. (E) Images of ALP staining in BMSCs transfected with ASO or miR-129-5p inhibitor after culture in OM for 7 days. (F) Histograms show ALP activity by spectrophotometry. Data are presented as means ± SD. Three independent experiments were performed. 10. 7717/peerj. 13355/fig-15 Figure 15 lncRNA- Malat1 sponges with miR-129-5p. (A) RIP assay shows that the miRNAs were successfully pulled-down using an Ago2 antibody. IgG was used as the negative control. (B) RNA levels of lncRNA Malat1 and miR-129-5p after immunoprecipitation were determined by qRT-PCR. Gapdh was the reference gene of lncRNA Malat1. U6 was used as the reference gene of miR-129-5p. Numbers are mean ± SD. Three independent experiments were performed. Discussion The age-related degradation of bone microstructure can lead to fragility fractures, and BMSCs play a critical role in bone mineralization changes ( Ensrud & Crandall, 2017 ). The proliferation and effector functions of BMSCs are regulated by intracellular and extracellular stimuli, including exosomes, metabolites, and lncRNAs, some of which are related to aging ( Concha et al. , 2019 ; Jiang et al. , 2021 ). Therefore, we screened differentially expressed lncRNAs based on the expression profile of BMSCs from elderly and young donors, finding that lncRNA Malat1 expression was significantly increased in the elderly group. Several studies have investigated how lncRNA Malat1 regulates the osteogenic differentiation of BMSCs through the ceRNA mechanism. Huang et al. (2020) found that Malat1 could promote the osteogenic differentiation of BMSCs by sponging miR-214. Yang et al. (2019b) determined that Malat1 functioned as a sponge for miR-34c, and thus enhanced the osteoblast activity of BMSCs. Moreover, Malat1 was also capable of sponging miR-124-3p to exert the same effect in BMSCs ( Li, 2022 ). Interestingly, another study found that Malat1 can inhibit the alkaline phosphatase activity of BMSCs by activating the MAPK signaling pathway ( Zheng et al. , 2019 ). These results suggest that Malat1 has a dual-effect on the osteogenic differentiation of BMSCs, and its mechanism may be far more complicated. In this current study, the osteogenic differentiation ability of BMSCs was suppressed by ASO- Malat1, which could also inhibit the expression of Malat1. Given that the ceRNA mechanism has received increasing attention in recent years, we also sought to explain this inhibitory effect by searching for another lncRNA/miRNA signal axis. At first, we chose miR-17-5p and miR-129-5p as hub genes, both of which had binding sites for Malat1 and their expression was negatively correlated with Malat1 expression. However, DLR experiments indicated that only miR-129-5p interacts with Malat1. A recent study also found that MALAT1 can bind to miR-129-5p and downregulate its expression in pancreatic cancer cells ( Xu et al. , 2021 ). Additionally, the inhibition of BMSC osteogenic ability caused by a lack of Malat1 can be reversed by miR-129-5p inhibitors, suggesting that Malat1 plays a regulatory role through sponging miR-129-5p. It is worth noting that through the FISH experiment and the nuclear/cytosol fractionation experiment, we found that Malat1 expression was higher in the nucleus than cytoplasm, which contradicts the findings of a recent study ( Li, 2022 ). Since the ceRNA regulation pattern of Malat1 has been confirmed by several studies, it is possible that these nuclear-expressed Malat1 may also regulate the osteogenic differentiation of BMSCs by affecting important transcription factors. Research has suggested that miR-129-5p can negatively regulate Runt-related transcription factor 1 and subsequently inhibit the cartilage differentiation ability of BMSCs ( Zhu et al. , 2021 ). An earlier study found that miR-129-5p reduced the expression level of Stat1, resulting in a dramatic increase in RUNX2 protein levels in murine BMSCs ( Xiao et al. , 2016 ). Our study determined that overexpression of miR-129-5p inhibited the expression of Stat1 to a certain extent. However, the dual-luciferase reporter gene experiment and the RIP assay showed that the interaction between miR-129-5p and Stat1 is not obvious. Interestingly, the predicted bind site of Stat1 (located in the CDS of Stat1 ) was consistent with a previous study ( Xiao et al. , 2016 ). While several studies have pointed out that a large number of miRNA binding sites exist in the CDS region ( Helwak et al. , 2013 ; Brümmer & Hausser, 2014 ), few results were related to these binding sites. Zhang et al. (2018) discovered and confirmed a novel class of miRNA response elements that function in the CDS region. MiRNAs induce transient ribosome stalling to repress translation without affecting mRNA levels ( Zhang et al. , 2018 ). This discovery provides a new idea for improving the regulatory mechanism of miRNA. In summary, we conducted in vitro experiments and found that lncRNA Malat1 could affect the osteogenic differentiation process of BMSCs by regulating miR-129-5p. However, further research is needed to determine the specific mechanism through which miR-129-5p affects osteogenic differentiation and applications for the use of lncRNA Malat1 in bone tissue engineering. Supplemental Information 10. 7717/peerj. 13355/supp-1 Supplemental Information 1 Data banks/repositories corresponding to all datasets analyzed in this study Click here for additional data file. 10. 7717/peerj. 13355/supp-2 Supplemental Information 2 All differentially expressed genes identified between young and old donors Click here for additional data file. 10. 7717/peerj. 13355/supp-3 Supplemental Information 3 Potential co-expression competing triples (lncRNA-miRNA-mRNA triples) Click here for additional data file. 10. 7717/peerj. 13355/supp-4 Supplemental Information 4 The enriched Gene ontology(GO) terms of linked mRNAs in the ceRNA network Click here for additional data file. 10. 7717/peerj. 13355/supp-5 Supplemental Information 5 The enriched pathway terms of linked mRNA in the ceRNA network Click here for additional data file. 10. 7717/peerj. 13355/supp-6 Supplemental Information 6 miR-17-5p is not a potential target of lncRNA Malat1 (A) Complementary bases between the sequences are labeled with red font. The sequence of the mutant lncRNA Malat1 construct is also shown as underlined. (B) Dual-luciferase reporter assay of 293T cells co-transfected with lncRNA Malat1, or lncRNA Malat1-Mut and with miR-17-5p mimic or miR-NC. Click here for additional data file. 10. 7717/peerj. 13355/supp-7 Supplemental Information 7 Full-length uncropped blots Click here for additional data file. 10. 7717/peerj. 13355/supp-8 Supplemental Information 8 Raw data for Fig. 4 (Go Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-9 Supplemental Information 9 Raw data for Fig. 4 (KEGG Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-10 Supplemental Information 10 Raw data for Fig. 5 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-11 Supplemental Information 11 Raw data for Fig. 5 (microCT Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-12 Supplemental Information 12 Raw data for Fig. 6 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-13 Supplemental Information 13 Raw data for Fig. 7 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-14 Supplemental Information 14 Raw data for Fig. 8 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-15 Supplemental Information 15 Raw data for Fig. 8 (AKP Activity) Click here for additional data file. 10. 7717/peerj. 13355/supp-16 Supplemental Information 16 Raw data for Fig. 10 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-17 Supplemental Information 17 Raw data for Fig. 12 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-18 Supplemental Information 18 Raw data for Fig. 12 (Dual luciferase) Click here for additional data file. 10. 7717/peerj. 13355/supp-19 Supplemental Information 19 Raw data for Fig. 13 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-20 Supplemental Information 20 Raw data for Fig. 13 (Dual luciferase) Click here for additional data file. 10. 7717/peerj. 13355/supp-21 Supplemental Information 21 Raw data for Fig. 14 (PCR Analysis) Click here for additional data file. 10. 7717/peerj. 13355/supp-22 Supplemental Information 22 Raw data for Fig. 14 (AKP Activity) Click here for additional data file. 10. 7717/peerj. 13355/supp-23 Supplemental Information 23 Raw data for Fig. 15 (PCR Analysis) Click here for additional data file.

10. 7717/peerj. 13356

PeerJ

Comparison between hydroxyapatite and polycaprolactone in inducing osteogenic differentiation and augmenting maxillary bone regeneration in rats

Background The selection of appropriate scaffold plays an important role in ensuring the success of bone regeneration. The use of scaffolds with different materials and their effect on the osteogenic performance of cells is not well studied and this can affect the selection of suitable scaffolds for transplantation. Hence, this study aimed to investigate the comparative ability of two different synthetic scaffolds, mainly hydroxyapatite (HA) and polycaprolactone (PCL) scaffolds in promoting in vitro and in vivo bone regeneration. Method In vitro cell viability, morphology, and alkaline phosphatase (ALP) activity of MC3T3-E1 cells on HA and PCL scaffolds were determined in comparison to the accepted model outlined for two-dimensional systems. An in vivo study involving the transplantation of MC3T3-E1 cells with scaffolds into an artificial bone defect of 4 mm length and 1. 5 mm depth in the rat’s left maxilla was conducted. Three-dimensional analysis using micro-computed tomography (micro-CT), hematoxylin and eosin (H&E), and immunohistochemistry analyses evaluation were performed after six weeks of transplantation. Results MC3T3-E1 cells on the HA scaffold showed the highest cell viability. The cell viability on both scaffolds decreased after 14 days of culture, which reflects the dominant occurrence of osteoblast differentiation. An early sign of osteoblast differentiation can be detected on the PCL scaffold. However, cells on the HA scaffold showed more prominent results with intense mineralized nodules and significantly ( p < 0. 05) high levels of ALP activity with prolonged osteoblast induction. Micro-CT and H&E analyses confirmed the in vitro results with bone formation were significantly ( p < 0. 05) greater in HA scaffold and was supported by IHC analysis which confirmed stronger expression of osteogenic markers ALP and osteocalcin. Conclusion Different scaffold materials of HA and PCL might have influenced the bone regeneration ability of MC3T3-E1. Regardless, in vitro and in vivo bone regeneration was better in the HA scaffold which indicates its great potential for application in bone regeneration.

Introduction Bone tissue has the ability to spontaneously heal through bone deposition and remodeling ( Fernandez-Yague et al. , 2015 ). However, in a larger bone defect due to trauma, surgical treatment of tumor and craniofacial defect such as cleft palate, a bone repair can only be done by bone graft ( Fishero et al. , 2015 ; Robey et al. , 2015 ). Bone graft in cleft palate repair is important for tooth eruption and orthodontic tooth movement ( Wahab et al. , 2020 ). Craniofacial defect repair by surgeons often requires sophisticated treatment strategies and multidisciplinary input with ideal situations using autologous bone. However, this option is limited by a finite supply of available bone, potential donor site morbidity, particular attention to growing patients, prolonged surgeries in ‘hostile defect‘ that may be associate with free flap loss, anesthetics/patient-related risks, and contour deformities ( Lee et al. , 2013 ). In the event of autologous bone being impractical or not feasible, the application of tissue engineering can be a promising concept within the craniofacial surgery field utilizing the engineered materials with a combination of cells to improve or replace biological functions. Tissue regeneration aims to help the body heal naturally by implanting a scaffold to serve as a temporary matrix that would degrade over time while allowing the regeneration of the host tissue at the implant site. Cellular response and osteoblast differentiation can be affected by morphology, size, surface topography, surface chemistry, porosity, interconnected structure, and fibrous pore wall of the scaffold ( Tavakol et al. , 2012 ). Therefore, the selection of scaffold plays a crucial role in ensuring the success of bone regeneration. The chosen scaffold must allow the cells to migrate, proliferate, and differentiate into osteoblasts for the correct development of the bone tissue ( Bose, Roy & Bandyopadhyay, 2012 ). Synthetic scaffolds have advantages over natural scaffolds as they can be manufactured under controlled conditions that allow large scale production with uniform size and design as well as exhibiting reproducible physical and chemical properties ( Farinawati et al. , 2020 ). Hydroxyapatite (HA) is known for its excellent biocompatibility due to its similarity in composition to the apatite found in natural bone. In biological systems, HA occurs as the inorganic constituent for normal calcification such as on bone, teeth, fish enameloid, some species of shell, and in pathological calcification such as on dental and urinary calculus or stone ( Hench & Thompson, 2010 ; Kattimani, Kondaka & Lingamaneni, 2016 ). Natural occurring HA appears to be brown, yellow, or green in coloration while pure or synthetic HA appears in white coloration. HA contains only calcium and phosphate ions, therefore, no adverse local or systemic toxicity has been reported in any study ( Kattimani, Kondaka & Lingamaneni, 2016 ). Biocompatibility, bioactivity, osteoinductivity, and osteoconductivity are good properties of HA that make them extensively being used as a scaffold for bone regeneration. Moreover, the different forms of HA scaffold that are actively being used include granules ( Dorozhkin, 2015 ), paste and cement ( Ben-Nissan, 2014 ), coatings ( Eliaz & Metoki, 2017 ), porous ( Al-Naib, 2018 ), and dense blocks ( Megat Wahab et al. , 2020 ). Nonetheless, concerns have been raised regarding the brittleness and limited degradation properties of HA, including the slow degradation rate ( Fiume et al. , 2021 ). Polycaprolactone (PCL) is much preferred in terms of degradation. PCL is a synthetic polymer that can undergo degradation by hydrolysis of ester bonds in physiological conditions. PCL is an aliphatic semi-crystalline polymer with a melting temperature above body temperature. Hence, at physiological temperature, PCL attains a rubbery state resulting in its high toughness and superior mechanical properties ( Dwivedi et al. , 2020 ). PCL appears to be non-toxic and tissue compatible which makes it suitable as scaffolds for bone regeneration. Dwivedi et al. (2020) believed that PCL has easy availability, is relatively inexpensive, and can be modified to adjust its chemical and biological properties, physiochemical state, degradability, as well as mechanical strength. PCL exhibits a degradation time of approximately two to three years and it can be degraded by microorganisms or under physiological conditions ( Anderson & Shive, 1997 ). Its degradation time makes it appropriate for the replacement of hard and load-bearing tissues by enhancing stiffness while decreasing the molecular weight and degradation time for soft tissues. However, several reports have shown that PCL is lack of osteoconductive property due to its poor hydrophilic nature ( Hajiali, Tajbakhsh & Shojaei, 2018 ; Torres et al. , 2017 ; Zhao et al. , 2015 ). There is still a need to investigate the biological performance of HA and PCL scaffolds in terms of bone integration between the implanted scaffold and surrounding host tissues as well as the difference of bone retention in the defect between these scaffolds. In addition, the question arises of whether different types of scaffold material will affect the cellular and osteogenic potential of the transplanted cells. The effects after transplantation such as tissue rejection and bone viability also need to be considered in determining the success of bone regeneration. Therefore, this study assessed the potential of HA and PCL scaffolds in supporting in vitro cell viability, attachment, morphology, and osteoblast differentiation compared to the accepted model outlined for two-dimensional (2D) systems. Preosteoblast MC3T3-E1 cells were used as cell sources and were cultured on scaffolds and 2D culture plates. The comparison during in vivo bone regeneration of HA and PCL scaffolds was also assessed in this study using a rat model with a maxillary bone defect. Materials & Methods Scaffolds preparation HA scaffold was obtained from GranuMaS ® (Granulab, Selangor, Malaysia) with a granule size range from 0. 2–1. 0 mm. Meanwhile, PCL scaffold was obtained from Osteopore™ (Singapore) with a size range from 3 mm × 1. 5 mm (diameter by height). Both HA and PCL scaffolds were sterilized using 75% (v/v) ethanol for 30 min, washed three times in sterile phosphate buffer saline (PBS) (Gibco, Thermo Fisher Scientific, USA), and exposed to 15 min of ultraviolet radiation for each side of the scaffolds. Sterilized scaffolds were immersed in α-Minimum Essential Medium ( α -MEM) (Gibco, Thermo Fisher Scientific, USA) overnight prior to the cell culture. Characterization of scaffolds HA and PCL scaffolds were sputter-coated with gold in order to obtain sufficient conductivity on the surface and to avoid surface charging during the viewing process. The morphology of the coated scaffolds was viewed under a field emission scanning electron microscope (FESEM, Supra 55VP, Zeiss) and elemental analysis was conducted using energy dispersive X-ray (EDX). Cell culture Mouse MC3T3-E1 subclone 14 preosteoblast cells (ATCC No: CRL-2594™) were cultured in a complete medium consisting of α -MEM supplemented with 10% (v/v) fetal bovine serum (Gibco, USA), 1 mM sodium pyruvate (Sigma, USA) and 1% (v/v) penicillin-streptomycin (Gibco, USA). The 2D culture of MC3T3-E1 was conducted as previously done by Yazid et al. (2019). The scaffolds and 96-well plates were seeded with 50, 000 MC3T3-E1 cells suspended in a complete medium and incubated for overnight to permit cell attachment. MC3T3-E1 cells-seeded scaffolds were then transferred to new 96-well plates to prevent a false positive result from cells attached at the bottom of wells. All cultured cells were maintained in a humidified atmosphere of 5% (v/v) CO 2 at 37 °C and the medium was changed every three days. For osteoblast differentiation, MC3T3-E1 cells on HA scaffolds, PCL scaffolds, and 2D culture plates were cultured in osteogenic media ( α -MEM supplemented with 10% (v/v) fetal bovine serum, 1% (v/v) penicillin-streptomycin, 50 µg/mL ascorbic acid (Sigma, USA) and 10 mM β -glycerophosphate (Sigma, USA)). Meanwhile, MC3T3-E1 cells on scaffolds and 2D culture plate in a complete medium without differentiation factors were used as a negative control for osteoblast differentiation. The differentiation and complete medium were changed every three days. In vitro analysis MTT assay for cell viability The viability of MC3T3-E1 cells on scaffolds and 2D culture plates at 80–90% confluency was evaluated on days 0, 7, 14, and 21 of culture by using the 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) substrate (Sigma, USA) which was reduced to formazan that accumulated in the cytoplasm of viable cells. Briefly, MTT solution (5 mg/mL) and complete medium at a ratio of 1:9 were added to each well containing MC3T3-E1 cells. Cells with MTT solution were incubated at 37 °C in a humidified atmosphere for 4 h. After incubation, the MTT solution was removed and a glycine buffer solution containing dimethyl sulfoxide (Sigma, USA) was added to dissolve formazan salts produced by the enzymatic reaction. After 10 min of agitation, the supernatants were collected and transferred into a new 96-well plate. Then, the absorbance at 570 nm was measured with an ELISA microplate reader (Varioskan Flash Model 680, Thermo Fisher, USA). A viable cell number was obtained through a standard calibration curve determined by correlating a known cell number with the optical density of the solution. For the standard calibration curve, an increasing number of cells from 100, 500, 1, 000, 5, 000, 10, 000. 50, 000, 100, 000, 500, 000 and 1, 000, 000 of MC3T3-E1 cells were directly cultured and assessed. MTT optical density was normalized to the number of cells on scaffolds and 2D culture plates. MTT assay was performed for five biological replicates and the technical tests were run in triplicates. Cell attachment and morphology MC3T3-E1 cell attachment and morphology on HA and PCL scaffolds were examined using FESEM. EDX spectroscopy was conducted to analyze the elemental composition on HA and PCL scaffolds after 21 days of osteoblast differentiation. At osteoblast differentiation culture intervals of 0, 7, 14, and 21 days, cell-seeded scaffolds were fixed overnight in 2. 5% (v/v) glutaraldehyde (Polysciences, Inc. , Warrington, PA, USA) with PBS and stored at 4 °C. The fixed samples were then washed with PBS three times and subjected to sequential dehydration for 10 min in a graded ethanol series (30% (v/v), 50% (v/v), 70% (v/v), 80% (v/v), 90% (v/v), and 100% (v/v)). Samples of HA scaffold were dried using a critical point drying while samples of PCL scaffold were allowed to dry in air for 24 h at room temperature. Both samples were sputter coated (Quorum, Q150RS) for 30 s with gold and observed under FESEM, at 3–15 kV accelerating voltage. Alkaline phosphatase specific activity and total protein content The ALP specific activity was measured by Sensolyte ® pNPP alkaline phosphatase assay kit (AnaSpec, USA) according to the manufacturer’s protocol. Briefly, the cells were homogenized in the lysis buffer provided in the kit. Lysate produced was centrifuged for 10 min at 2, 500 g at 4 °C. The supernatant was collected and incubated with p-nitrophenyl phosphate (pNPP) at 37 °C for 30 min. Stop solution was added after 30 min of incubation and absorbance measurement was taken at a wavelength of 405 nm using an ELISA microplate reader. Total protein content was measured by Bradford assay with bovine serum albumin used as a standard ( Kruger, 2009 ). ALP activity results were normalized to the total protein content and were represented as U/mg. All samples were run for five independent experiments and repeated three times. In vivo analysis Cells and scaffolds preparation MC3T3-E1 cells were cultured in vitro on HA and PCL scaffolds for 14 days before transplantation. Cells on scaffolds were cultured in a complete media supplemented with osteoblast differentiation factors. The medium was changed every three days. After 14 days of in vitro culture, cells-seeded scaffolds were transplanted into rats with a surgically made maxillary bone defect. Animals A total of 24 mature female Sprague Dawley rats (age: 6–8 weeks, body weight: 200–300 g) were used in this study. The housing, care, and experimental protocol were approved by the Universiti Kebangsaan Malaysia Animal Ethical Committee with the approval number FD/2018/ROHAYA/26-SEPT. -2018-JUNE-2019. The animal study was reported according to the ARRIVE guidelines concerning the relevant items. The rats were obtained from the Laboratory Animal Resource Unit, Faculty of Medicine, Universiti Kebangsaan Malaysia (UKM). Prior to the transplantation’s surgery, the rats’ health was monitored for a week. All of the rats were kept in pairs per cage at the animal house of the Faculty of Health Science, UKM, with 12 h light-dark cycle at 21–25 °C and were fed with food pellets. Water was supplied on an ad libitum basis. The activity of rats was observed once daily throughout the study. Surgery and transplantation To create the bone defect on the rat’s left maxilla, each rat was first anesthetized with an intravenous injection of 80 mg/kg ketamine (TROY Laboratories PTY Limited, Glendenning, Australia) combined with 7. 5 mg/kg xylazine (Indian Immunological Limited, Telangana, India) and 12 mg/kg tramadol (Y. S. P, Kuala Lumpur, Malaysia). A buccal sulcular incision was made to expose the maxillary bone. A bone defect with four mm length and 1. 5 mm depth was created in the anterior part of the left maxilla using a trephine bur under constant irrigation. Constant irrigation with cooled sterile PBS was performed to prevent overheating of the bone. The rats were randomly divided into four groups each containing six animals ( n = 6). The four groups were as follows: implantation of HA scaffolds (group 1); implantation of PCL scaffolds (group 2); transplantation of MC3T3-E1-HA (group 3); and transplantation of MC3T3-E1- PCL (group 4). After the transplantation, the mucosal flaps were closed with a simple interrupted suture pattern using 4-0 non-absorbable black silk suture. Postoperatively, each rat received a subcutaneous injection of 2 mg/kg dexamethasone (Duopharma, Malaysia) and 20 mg/kg amoxicillin (Duopharma, Malaysia) for a week to prevent tissue rejection and perioperative infection. The rats were maintained on a soft high-glucose diet for a week. A regular diet was resumed one week postoperatively. Rats were monitored daily by visual observation for signs of infection, inflammation, lack of food and water intake, as well as lethargy. Euthanasia and maxilla sample collection All the surviving animals ( n = 24) were sacrificed six weeks after the transplantation surgery. Prior to euthanasia, the rats received an intravenous injection of anesthesia (80 mg/kg ketamine, 7. 5 mg/kg xylazine, and 12 mg/kg tramadol). Approximately 10 min after the anesthesia induction, rats were then sacrificed via drug overdosing using a commercial euthanasia solution of 390 mg sodium pentobarbital (Vetoquinol SA, France) and 50 mg/ml sodium phenytoin (Duopharma, Malaysia) that were administered intravenously at the lateral vein. Maxilla with intact surrounding tissues from all the rats was dissected and immediately placed in 10% (v/v) neutral buffered formalin (R&M Chemical, UK) for 24 h before being rinsed with PBS. The samples were maintained in a buffer solution consisting of PBS with penicillin and streptomycin at 4 °C until further use. Micro-computed tomographic analysis The maxillary section of the rats was scanned using in vivo high-resolution micro-computed tomography (Skyscan 1176; Skyscan, Belgium). The micro-computed tomography (micro-CT) analysis was performed as previously described by Kim et al. (2013). Specifically, the micro-CT projection images were acquired at a source voltage of 70 kV, and a current of 142 µA using a 1 mm aluminium filter with a resolution of 12. 32 µm pixels. Scanning was performed by a rotation angle of 360° around the vertical axis, camera exposure time of 580 ms, and a rotation step of 0. 7°. Each specimen was scanned for a total of 40 min. The micro-CT images were then reconstructed in the CTAnalyser Skyscan software using approximately 400 scan slices per sample with an image pixel size of 17. 56 m. The 2-dimensional projections acquired were elaborated to generate the region of interest (ROI). A 3-dimensional ROI of the bone defect was obtained by manually tracing the margin of each bone defect through the software. The ROI of each treatment group ( n = 6) was analyzed for percentages of new bone volume (%), bone surface (mm 2 ), and bone surface density (mm −1 ). The results were assessed by a trained examiner, who was blinded to the experimental groups (S. F. L. ). Hematoxylin and eosin analysis Tissue samples were decalcified in 10% (v/w) buffered ethylene diamine tetraacetic acid (Sigma, USA) (pH 7. 4) for 5–6 weeks then dehydrated with a graded sequence of increasing ethanol concentrations (from 70% to 100%) and embedded in paraffin. Serial 5 µm-thick sections were generated using a microtome and stained with hematoxylin and eosin (H&E) following the standard protocols. The stained section was observed under a light microscope (BX51; Olympus, Tokyo, Japan) and a digital image was obtained using CellB software. The results of new bone formation within the defect area were taken under 40x and 100x magnifications. Histological level of new bone formation was assessed using a seven-point scale outline by Salkeld et al. (2001) : 0 = no incorporation and no new bone formation, 1 = some incorporation and a small amount of new bone, 2 = some incorporation and a moderate amount of new bone formation, 3 = some corporation with new bone formation continuous with host bone and early remodeling changes in new bone, 4 = good graft incorporation and ample new bone, 5 = good graft incorporation of graft and new bone with host and ample new bone, 6 = excellent incorporation and advanced remodeling of new bone with graft and host. The analysis was validated by an experienced pathologist blinded to the study groups (N. S. N. ). Immunohistochemistry After deparaffinization, 5 µm-thick tissue sections were subjected to immunohistochemistry (IHC) staining for osteogenesis makers of ALP, and osteocalcin (OCN). Rabbit polyclonal antibody against ALP (1:100, ab65834, Abcam, Cambridge, UK), mouse monoclonal antibody against OCN (1:200, MAB1419, R&D Systems, USA), and Mouse and Rabbit Specific HRP/DAB IHC Detection Kit-Micro-Polymer (ab236466, Abcam, Cambridge, UK) were employed for the study. These primary antibodies were diluted in a solution containing 3% (w/v) BSA in PBS. The specimens were incubated with sodium citrate buffer (Sigma, USA) (pH 6. 0) and heated in a microwave oven at the lowest temperature for 10 min. Following a 10-minute incubation with 3% hydrogen peroxide solution, the slides were blocked with Protein Block for10 min to block nonspecific antigens. The specimens were then incubated overnight with a primary antibody at 4 °C. After washing the specimens with PBS buffer, the secondary antibody (goat anti-rabbit HRP-conjugate) incubation steps were conducted following the kit instructions. For visualization, the antigens were detected with DAB chromogen supplied by the kit. All sections were counterstained with Mayer’s hematoxylin, and the slides were dried, then, cover-slipped with DPX mounting medium (VWR Chemicals, France). The results were observed and documented using an Olympus BX51 microscope at 100 × magnification. The immunohistochemical evaluation of the ALP and OCN expression was performed semi-quantitatively using an immunoreactive score (IRS). The IRS was assessed according to a range of 0–12 as a product of multiplication between the percentage of positive cells score (0–4) and staining intensity score (0–3) as done previously by Koerdt et al. (2013) and Ulu et al. (2018) ( Table S1 ). Statistical analysis Multiple comparisons for in vitro cell viability and osteoblast differentiation potential of scaffolds as well as in vivo bone morphometric, histology grading, and IRS analyses were evaluated using Bonferroni-corrected by one-way analysis of variance (ANOVA). Values of p less than 0. 05 were considered to be statistically significant. The data were expressed as the mean ± standard deviation of five independent experiments for the in vitro study while a repeated experiment using six rats were used as in vivo biological replicates. Statistical analyses were performed using the SPSS 21. 0 software (SPSS Inc. , Chicago, IL, USA). Ethical approval This research was carried out according to the ethical and legal requirements of the Universiti Kebangsaan Malaysia Animal Ethical Committee (UKMAEC). This permission allowed us to use rats as experimental animals while abiding by the legal and ethical guidelines. Experiments utilized rats were performed humanely throughout this research. The euthanasia method was performed following the guideline from the American Veterinary Medical Association ( Underwood & Anthony, 2020 ). All described experimental protocols involving rats were designed and performed according to the animal ethical guidelines approved by the UKMAEC with approval reference number FD/2018/ROHAYA/26-SEPT. /944-SEPT. -2018-JUNE-2019. Results The observations and characterizations of HA and PCL scaffolds FESEM micrographs of HA and PCL scaffolds are presented in Fig. 1A. The HA scaffolds are composed of irregular granules with diameters between 0. 2 to 1. 0 mm. HA granules showed the morphology of spherical agglomerates with very limited contact areas among granules. The presence of micro-sized grains with a fully interconnected macroporosity could be observed at a higher magnification. Meanwhile, the PCL scaffolds had a typical honeycomb structure and a relatively smooth surface with interconnected equilateral triangles of regular porous morphology. The pore size of the PCL scaffold was measured, and the average pore size was 500 µm. At a higher magnification, the surface of PCL showed the presence of a small groove that may increase scaffold surface area. 10. 7717/peerj. 13356/fig-1 Figure 1 Characterization of HA and PCL scaffolds. (A) FESEM morphology of HA and PCL scaffolds at 23x, 100x, and 1000x magnifications. (B) EDX analysis of HA and PCL scaffolds. Figure 1B shows the EDX analysis for the HA and PCL scaffolds. The constituents are identified by the peaks concerning their energy levels. HA scaffolds have higher peaks of calcium (Ca) and phosphate (P) than PCL scaffolds. In addition, other essential elements such as carbon (C) and oxygen (O) can also be observed on both scaffolds. In vitro analysis Cell viability MC3T3-E1 cells viability on HA and PCL scaffolds as well as 2D culture plates were measured by the increased number of viable cells throughout the culture period of 21 days. Cells grown in all culture conditions showed a continuous increase with culture time, reflecting good cell viability ( Fig. 2 ). The number of viable MC3T3-E1 cells on scaffolds was markedly higher than the control 2D culture plate, significantly during the initial day of culture (days 0; after a 24-hour attachment period) between MC3T3-E1-HA and MC3T3-E1-2D ( p = 0. 034). Meanwhile, MC3T3-E1-HA ( p = 0. 000) and MC3T3-E1-PCL ( p = 0. 0004) showed a significantly higher number of cell viable on day 7 compared to MC3T3-E1-2D ( Table S2 ). Interestingly, MC3T3-E1 cells grown in HA and PCL scaffolds continued to grow for up to 14 days but this growth decreased at 21 days. MC3T3-E1 cells showed a significantly increased number of viable cells on HA scaffold compared to PCL scaffold on days 7 ( p = 0. 000) and 14 ( p = 0. 002) ( Fig. 2 and Table S2 ). 10. 7717/peerj. 13356/fig-2 Figure 2 Viability of MC3T3-E1 cells on HA scaffolds, PCL scaffolds, and control 2D culture plates. The viability of MC3T3-E1 cells on 2D culture plates showed an ongoing increase with culture time, whereas the viability of MC3T3-E1 cells cultured on HA and PCL scaffolds was increased up to 14 days and reduced at 21 days of culture. Values were plotted as a mean number of viable cells ± standard deviation ( n = 5). 10. 7717/peerj. 13356/fig-3 Figure 3 Osteoblast differentiation potential of MC3T3-E1 cells on HA and PCL scaffolds for 21 days culture period. (A) Field emission scanning electron microscope image. Arrows indicate mineralized nodules following osteoblast differentiation. Scale bar: 1 µm and 2 µm for HA scaffold, 2 µm for PCL scaffold. (B) EDX analysis for the detection of mineralization. Comparison of EDX elemental analysis on both scaffolds showed a higher level of calcium and phosphorus after 21 days of osteoblast induction which indicates a higher quantity of minerals. (C) ALP specific activity of MC3T3-E1 cells culture on both scaffolds and control 2D culture plates. Asterisks (*) indicate significant differences after a Bonferroni correction for n = 5, at p < 0. 05 between different culture conditions. Data represent mean ± standard deviation shown in units per microgram of total cellular protein. Cell attachment and morphology FESEM results indicated that both types of scaffolds allowed the attachment and spreading of the cells while maintaining a normal cellular morphology ( Fig. 3A ). As it can be seen in Fig. 3A, MC3T3-E1 cells on day 0 (a day after osteoblast induction) were already well attached and spread to the surface of both scaffolds, presenting a round shape configuration in HA scaffolds while a cluster of cells with extended cytoplasm in all directions was observed in the PCL scaffolds. After 7 days of culture, cells on both scaffolds showed a typical morphology presenting a flat configuration with more lamellipodia connecting to the neighboring cells and starting to form a continuous cell layer. Mineralized nodules can be observed as early as day 7 in PCL scaffolds. On day 14 of culture, MC3T3-E1 cells on HA scaffolds start to aggregate and form a monolayer of connected cells while a dense cell layer can be seen covering the surface of PCL scaffolds. At 21 days of culture, a dense cell layer could be observed on both scaffolds with some mineralized nodules appearing in between the cell layer especially on HA scaffolds (yellow arrows). During an early culture day, the density of the attached cells is higher in PCL scaffolds compared to HA scaffolds. However, as days of culture increased to day 21, the FESEM image showed that more cells were attached, and an intense appearance of mineralization nodule was observed over HA scaffolds compared to PCL scaffolds. EDX elemental analysis EDX spectroscopy results showed the presence of a higher level of calcium and phosphorus after 21 days of MC3T3-E1 cultured on HA and PCL scaffolds ( Fig. 3B ). HA scaffolds had a higher ratio of calcium/phosphorus (Ca/P) level from 2. 29 to 2. 4. Meanwhile, PCL scaffolds showed an increase in the Ca/P ratio from 2. 22 to 2. 36. Both scaffolds showed a slightly higher Ca/P ratio than the theoretical pure hydroxyapatite which is 2. 15 ( Venkatasubbu et al. , 2011 ). Oxygen and carbon peaks present on the EDX indicate by-products excreted during the extracellular matrix production of MC3T3-E1 cells on the scaffold. The concentration of oxygen and carbon on both scaffolds were observed to be lower than calcium and phosphorus throughout 21 days of osteoblast differentiation. ALP specific activity The ALP specific activity was evaluated at days 0, 7, 14, and 21 after MC3T3-E1was cultured in an osteogenic medium ( Fig. 3C ). A trend of ALP specific activity was increased up to 21 days of osteoblast differentiation in all culture conditions. There were no statistically significant differences ( p > 0. 05) between MC3T3-E1 cells on scaffolds and 2D culture plates at day 0. The result also showed that the ALP specific activity of MC3T3-E1-PCL (0. 19 ± 0. 03 U/mg) on day 7 was approximately 1. 24 and 3. 91 times higher than MC3T3-E1-HA scaffolds (0. 15 ± 0. 02 U/mg; p = 0. 001) and 2D culture plates (0. 05 ± 0. 03 U/mg; p = 0. 000). Interestingly, prolonged osteoblast induction of MC3T3-E1 cells on HA scaffolds resulted in a significantly higher ALP specific activity compared to 2D culture plates especially on days 14 ( p = 0. 000) and day 21 ( p = 0. 019) compared to PCL scaffolds. In vivo analysis Macroscopic analysis All animals ( n = 24) survived and remained healthy for the entire six weeks of the transplantation period, showing no noticeable sign of toxicity or other adverse effects. Moreover, examined samples present no significant complications such as dehiscence or fistula in the area of the maxilla defect. The maxillary bone defects were healing well without the presence of necrosis or obvious inflammation detected in any fresh maxilla specimen. A stable scaffold fixation with no migration was observed in all rat samples. HA and PCL scaffolds remained after six weeks of transplantation. Some parts of the HA and PCL scaffolds surface were covered by callus ( Fig. 4 ). 10. 7717/peerj. 13356/fig-4 Figure 4 Photographs illustrating the surgical transplantation of MC3T3-E1-HA and MC3T3-E1-PCL. The surgically-made bone defect was created in the anterior part of the rat’s left maxilla. (B) A maxillary bone defect measuring four mm length and 1. 5 mm depth was created. (C) Defect received transplantation of MC3T3-E1-HA. (D) Defect received transplantation of MC3T3-E1-PCL. (E) Harvested tissue samples of HA scaffold. (F) Harvested tissue samples of PCL scaffold. Throughout the study period, animals showed a good healing response without adverse tissue reactions. Red circles indicate the presence of scaffolds after six weeks of transplantation. Micro-computed tomographic analysis Bone regeneration potential of HA and PCL scaffolds were also investigated using rat’s maxillary bone defect. Surrounding tissues with scaffolds were imaged and analyzed using high resolution micro-computed tomographic (micro-CT), and 2D images were reconstructed ( Fig. 5A ). Limited to no bony bridge could be observed on empty scaffold groups. As shown in the 2D reconstructed image, treatment of rat maxillary bone defect with MC3T3-E1-HA (group 3) and MC3T3-E1-PCL (group 4) demonstrated bony bridges with an increased amount of filled new bone compared to the control groups. Bridging of the defects with new bone occurred extensively in group 3. Some of the HA granules have consolidated and its radio density increased. Group 4 showed minimal new bone formation and lower radio-density. Defects in group 4 showed a translucent bony bridge at six weeks postoperatively. 10. 7717/peerj. 13356/fig-5 Figure 5 Micro-CT analysis showed maxillary bone defect area with the transplantation of cells-seeded scaffolds displaying new bone and defect area after six weeks. (A) 2D reconstructed image of rat’s maxillary bone treated with scaffolds. No bony bridge formation could be observed in the empty HA and PCL control groups. Bridging of the defects with new bone occurred extensively with increase radio-density on the MC3T3-E1-HA group. No residual of the PCL scaffold can be observed during scanning. Moderate new bone formation and lower radio density with translucent bony bridge could be observed on the MC3T3-E1-PCL group. (B) Graphs display bone morphometric analysis in the form of a percentage of new bone volume, bone surface area, and bone surface density. Data were expressed as mean ± standard deviation. Asterisks (*) indicate significant differences after a Bonferroni correction for n = 6, at p < 0. 05. As shown in Fig. 5B, bone morphometric analysis of the micro-CT images was used to quantify the percentage of the total new bone volume, new bone surface area, and the surface density of newly formed bone. Rats on treatment group 3 (42. 74% ± 9. 45%) showed a significantly increased new bone volume compared to group 4 (5. 43% ± 1. 82%; p = 0. 002), HA scaffolds control group (12. 8% ± 7. 08%; p = 0. 012) and PCL scaffolds control group (2. 22% ± 0. 36%; p = 0. 001). There is no significant difference ( p > 0. 05) observed in the increment of bone surface area between treatment groups. Meanwhile, the new bone surface density level of group 3 (7. 91 ± 1. 07 mm −1 ) was increased significantly compared to group 4 (1. 69 ± 0. 29 mm −1 ; p = 0. 001), HA scaffolds control group (3. 45 ± 1. 43 mm −1 ; p = 0. 014) and PCL scaffolds control group (0. 9 ± 0. 1 mm −1 ; p = 0. 000). Microscopic analysis New bone formation and biocompatibility of the chosen scaffolds in rat’s maxillary bone defect model were evaluated histologically using H&E staining. Defects treated with HA and PCL scaffolds showed cells and tissues infiltration with new bone formation through six weeks of transplantation period ( Fig. 6 ). The newly formed bone in the defect area of MC3T3-E1-HA (groups 3) and MC3T3-E1-PCL (group 4) was higher compared to the empty scaffold control groups. The connective tissue within the bone bridge in groups 3 and 4 was less prominent than in empty scaffolds control groups. 10. 7717/peerj. 13356/fig-6 Figure 6 Histological analysis on rat’s maxillary bone defect after six weeks of transplantation. Histological images from each treatment group at low magnification and high magnification. The arrow symbol indicates newly formed bone. The new blood vessel is marked by an arrowhead. Scale bar 200 µm and 100 µm. HA, hydroxyapatite; PCL, polycaprolactone. Empty HA scaffold groups showed the defect was surrounded by fibrous connective tissue and the new bone formation was growing from the edge of the cavity towards the center. A moderate amount of bone islands with numerous osteoblast cells as well as a small number of osteocytes in the irregular lacuna could be observed at the edge of the defect. Group 3 showed considerably enhanced new bone formation compared to the empty HA control group. Effective scaffolding property of HA in osteoconduction could be seen with new bone ingrowth that was well developed throughout the pore channels of the HA scaffolds. This can be observed with connective tissues migrated within HA scaffolds. Significant deposition of osteoblast at the marginalized parts of the new bone surrounding the periphery of HA scaffolds could also be observed extensively in group 3 than the empty HA control group. The empty PCL group showed small new bone formation in the middle of the defect areas with the appearance of extensive fibrous connective tissue ingrowth. Group 4 showed a presence of new bone that was extended into the scaffolds from the edge of the defect towards the center. A moderate amount of bone islands could be observed peripherally. This confirmed that bone formation started at the periphery of the PCL scaffolds. However, intense fibrous and connective tissues were still present in the defect region treated with PCL scaffolds compared to HA scaffolds. Granulation tissue and neovascularization were visible at the defect areas treated with both scaffolds. Both scaffolds demonstrated good biocompatibility with no significant inflammatory reaction. Histological grading for new bone formation in the defect area treated with MC3T3T-E1-HA and MC3T3T-E1-PCL were higher compared to the empty scaffold control groups. In comparison to their respective controls, the MC3T3-E1-PCL scaffold showed the highest new bone formation with a 1. 64-fold increase while only a 1. 25-fold increase was observed on MC3T3-E1-HA scaffold. Defects treated with HA scaffolds showed the highest histological grading of new bone formation compared to PCL scaffolds. However, the mean difference between scaffold treatment groups was not statistically significant ( p > 0. 05) ( Table 1 ). Immunohistochemistry evaluation The ability of HA and PCL scaffolds to enhance in vivo bone regeneration was further evaluated by immunohistochemistry analysis. IHC staining revealed that all treatments group treated with HA and PCL scaffolds showed positive expression for ALP and OCN ( Fig. 7A ). For both scaffold materials, ALP immunoreactive areas appeared mainly in osteoblast at the interface between newly formed bone and scaffolds. Fibroblasts of the connective tissue were mainly unstained or weakly stained. However, fibroblasts close to the scaffolds also showed weak to moderate immunoreactivity. The expression of ALP was more intensive in treatment groups with HA scaffolds compared to PCL scaffolds prominently at the edge of a new bone, scaffolds interface, and in the extracellular matrix. OCN immunoreactive area was localized in the osteoblasts, osteocyte, and unmineralized fibrous matrix. Fibrous connective tissue was strong to moderately immunostained. Stronger expression of OCN was observed primarily in the peripheral parts of scaffolds and interfaces between the scaffold and newly formed bone. The immunoreactivity pattern of OCN was similar to ALP, except for intense staining at the interfaces between scaffolds. A newly formed bone and within the connective tissue could be noticed significantly in treatment groups with HA scaffolds rather than PCL scaffolds. Connective tissues inside the HA scaffolds also showed positive expression of OCN. 10. 7717/peerj. 13356/table-1 Table 1 Histological grade of bone formation at maxillary bone defect after six weeks of transplantation. Data expressed as mean difference ± standard deviation ( n = 6). No significant ( p > 0. 05) difference for MC3T3-E1-HA and MC3T3-E1-PCL compared to empty HA and PCL control groups. Treatment groups New bone formation (0 to 6) Empty HA scaffold 2. 00 ± 0. 68 Empty PCL scaffold 1. 83 ± 0. 60 MC3T3-E1 - HA scaffold 2. 50 ± 0. 43 MC3T3-E1-PCL scaffold 3. 00 ± 0. 52 10. 7717/peerj. 13356/fig-7 Figure 7 Immunohistochemistry analysis on rat’s maxillary bone defect after six weeks of transplantation. (A) Immunohistochemical staining for alkaline phosphatase (ALP) and osteocalcin (OCN) in each treatment group. The arrow symbol shows strong positive staining. Scale bar 100 µm. (B) Immunoreactive score (IRS) for ALP and OCN. Asterisks (*) indicate significant differences between different treatment groups after a Bonferroni correction at p < 0. 05. Data represent the mean IRS ± standard deviation of the six rats observed in at least two different visual fields. Mean IRS for ALP increased significantly in the MC3T3-E1-HA and MC3T3-E1-PCL compared to the empty HA scaffold ( p = 0. 001 and p = 0. 002, respectively) ( Fig. 7B ). Moderate IRS grade for ALP was demonstrated prominently on MC3T3-E1-HA followed by MC3T3-E1-PCL, PCL, and HA. Meanwhile, MC3T3-E1-HA and MC3T3-E1-PCL showed a significantly higher mean IRS for OCN compared to empty PCL scaffolds groups ( p = 0. 010 and p = 0. 026, respectively) ( Fig. 7B ). Moderate IRS grade for OCN was demonstrated prominently on MC3T3-E1-HA followed by MC3T3-E1-PCL, HA, and PCL. Discussion In this study, MC3T3-E1 cell viability on HA scaffold, PCL scaffold, and 2D culture plate was measured and compared to confirm the ability of chosen scaffolds to support the growth of cells during 3D in vitro culture. Through MTT assay, we identified that the number of viable MC3T3-E1 cells on scaffolds was significantly higher than the control 2D culture plate even during the initial day of culture, indicating that the 3D structure of scaffolds may provide an optimum growth environment for cells by facilitating more space for nutrient and metabolic waste exchange ( Hoveizi et al. , 2014 ; Lim, Sun & Sultana, 2015 ; Mehendale et al. , 2017 ). Moreover, the viability of MC3T3-E1 cells on the HA scaffold was significantly increased with a peak number of viable cells observed on day 14. The cellular response and cell’s behavior can be influenced by the characteristic of the material surface ( Seebach et al. , 2010 ; Shamsuddin et al. , 2017 ). The HA scaffold used in this study exhibits a rough surface that may enhance cell growth. A previous study by Ling et al. (2015) demonstrated that cells seeded on the HA-composite scaffold had higher cell proliferation compared to the β-tricalcium phosphate-composite scaffold due to the rough texture present on the surface of HA. This report is consistent with our finding that the HA scaffold promotes higher cell growth during the initial period. However, it was significantly decreased with a prolonged culture which reflects the dominant occurrence of osteogenic differentiation. This pattern has also been observed in cell viability on the PCL scaffold. The obvious reduction in cell viability on HA and PCL scaffolds compared to control 2D culture plate may be due to the transition of the cells from a proliferative phase to the differentiation phase induced by direct interactions with the scaffolds ( Weissenböck et al. , 2006 ). This finding suggests that HA and PCL scaffolds are cytocompatible as they support and enhance MC3T3-E1 cell viability. MC3T3-E1 cells differentiation toward osteoblast was observed through FESEM image, EDX analysis, and ALP specific activity. FESEM image showed intense attachment, well-spread morphology and extensive growth of osteoblast differentiated MC3T3-E1 cells on both scaffolds. These results coincide with studies by of Seebach et al. (2010) and Jamal et al. (2018) who also found strong attachment, growth, and proliferation of mesenchymal stem cells on HA and PCL scaffolds. Although the HA has the potential to naturally induce osteoblast differentiation, MC3T3-E1 cells on the PCL scaffold showed signs of osteoblast differentiation as early as day 7 of induction. On day 7 of osteoblast differentiation, mineralization nodules were detected on the PCL scaffold while it was not present on the HA scaffold. When cells come into proximity to one another, presumably, they are stimulated to differentiate and mineralization begins once multilayers of cells formed ( Bellows et al. , 1986 ; Jamal et al. , 2018 ). A study by Bellows et al. (1986) demonstrated that the ability of cells to form multilayers seems to be the basis of mineralization due to the incapability of single-cell layers to produce a mineralized matrix. In this study, MC3T3-E1 cells on PCL scaffolds at day 7 became close to each other as more cells can be observed attaching to the scaffold compared to HA scaffolds. These results are in agreement with Jamal et al. (2018) which demonstrated early osteoblast differentiation of dental pulp stem cells on PCL scaffolds was influenced by intense cell-to-cell contact. However, mineralized nodules can be observed more prominently on HA scaffolds with prolonged osteoblast differentiation. These results are likely to be related to the rough texture present on the surface of HA that coherently supports extensive MC3T3-E1 cell growth. An increased number of MC3T3-E1 cells may permit more cell-to-cell contact that directly induced osteoblast differentiation and mineralization. This finding raises the possibility that the surface characteristic of the scaffold could provide a suitable microenvironment for cell attachment, cellular interaction, and osteoblast differentiation. These results were further supported by EDX analysis which allowed assessment and quantification of the presence of different bone types based on elemental analysis of calcium, phosphorus, and nitrogen ( Prati et al. , 2020 ). Therefore, EDX spectroscopy was utilized to detect the amount of elemental calcium and phosphorus produced by cells grown on HA and PCL scaffolds. EDX results demonstrated that MC3T3-E1 cells were able to attach and grow on HA and PCL scaffolds and form mineralized nodules or tissues consisting of calcium and phosphorus deposits. HA scaffold showed the higher ratio of calcium to phosphorus after 21 days of osteoblast differentiation which indicates a higher quantity of minerals. In addition to the initial evaluation of MC3T3-E1 cell morphology using FESEM following osteoblast differentiation, an increase in ALP specific activity of these cells on scaffolds is evidence of successful induction of osteogenic differentiation. ALP is one of the generally recognized biochemical markers for bone cell activity and is considered to play a role in bone mineralization ( Megat Wahab et al. , 2020 ). Scaffolds for bone regeneration application must support the differentiation of cells to functional bone tissue. Our results showed an overall increase of ALP specific activity throughout 21 days of osteoblast differentiation for all culture conditions. Higher ALP specific activity of MC3T3-E1 cells has been observed on HA and PCL compared to the control 2D culture plate. Both scaffolds were able to support osteoblast differentiation of MC3T3-E1 cells. The results of this study also showed that MC3T3-E1 cells on PCL scaffolds required a shorter time to undergo osteoblast differentiation with a high ALP specific activity detected as early as day 7. However, as the day of osteoblast differentiation increased, ALP specific activity of MC3T3-E1 cells on HA scaffolds showed a more prominent result. A previous study by Deligianni et al. (2000) suggested that there was a delay in the expression of ALP activity on HA scaffolds with rougher surfaces. These results are consistent with the FESEM images that showed mineralized nodules of MC3T3-E1 cells on PCL can be observed as early as day 7 and an increase of mineralized nodules was observed with a prolonged culture on HA scaffolds. An increase in ALP specific activity of MC3T3-E1 cells on HA scaffolds at the end of the osteoblast differentiation period was further enhanced by osteoinductive property present on the scaffolds. Studies by Usui et al. (2010) and Wang et al. (2015) indicate the potential of calcium and phosphate ions released from HA could directly induce and up-regulates osteoblast differentiation which promotes bone formation through calcification. Taken together, these findings suggest the potential of HA and PCL scaffolds in promoting in vitro osteoblast differentiation which can be observed more prominently in HA scaffolds. Since this study also aims to test different scaffolds material on enhancing in vivo bone regeneration, it is important to consider that our animal model is ideal for scaffold implantation in the maxillary bone defect treatment. No synchronization for rat oestrus cycles was performed in this study, which may have a significant effect on osteogenesis. Even though additional studies in more diverse populations of animal gender should be conducted, recent analyses of variability in male and female’s animal models demonstrate that unstaged females are not more varied than males across diverse traits, from gene expression to hormone levels in multiple species ( Becker, Prendergast & Liang, 2016 ; Dayton et al. , 2016 ; Itoh & Arnold, 2015 ; Prendergast, Onishi & Zucker, 2014 ; Smarr et al. , 2017 ). Beery (2018) suggested that estrous variability in a female animal is no greater than intrinsic variability in males. The variability may arise from different sources such as consistent estrous-cycle dependent variability that may be exhibited by female animals, while male animals showed variability on different timescales, or more variability between individuals. No animal death was observed during the transplantation period. Postoperatively, all rats showed immediate recovery as daily activities resumed within 24 h. There were no complications such as dehiscence or fistula in the area of transplantation. A stable scaffold fixation with no migration was observed in all rat samples. HA and PCL scaffolds remained after six weeks of transplantation. This indicates that both scaffolds have slow biodegradation under physiologic conditions with high mechanical resistance. A similar finding was also reported by Wongsupa et al. (2017) which suggests that slow biodegradation of PCL-biphasic calcium phosphate scaffold could have advantages for highly loaded areas that required long-term critical functional support. In the presented study, in vivo bone regeneration of transplanted MC3T3-E1 cells on HA and PCL scaffolds were evaluated by micro-CT and histology analyses. The reconstructed micro-CT image revealed a radio-dense appearance, suggesting bone ingrowth within HA and PCL scaffolds that filled the defect area. These results corroborated the ideas of Crovace et al. (2020) who showed that new bone formation could be observed by radio-dense aspect from X-ray images. Meanwhile, the bone morphometric analysis revealed that HA scaffolds significantly enhanced new bone volume and surface density, although no significant differences was found on bone surface area compared to PCL scaffolds. Our results are similar to a previous study by Jang et al. (2017), which found bone formation was indicated by a higher bone volume present in granular and porous HA scaffolds. Low bone surface area occurs when trabecular thickness, bone surface density to tissue volume are at a high level ( Kim et al. , 2004 ). Micro-CT can be over-estimated depending on the characteristics of bone substitutes or scaffolds. Therefore, there are chances that an increase in bone formation is due to the inherent radio-opacity contained in HA scaffolds ( Jang et al. , 2017 ). Meanwhile, the low value of bone volume, surface area, and surface density on PCL scaffold is due to the translucent property of PCL. This led us to use a qualitative histological approach to further evaluate the new bone formation. The results from the histological observation showed that HA and PCL scaffolds are capable of enhancing new bone formation with intensive vascularity at the defect area, indicating the bone vitality and beginning of new bone formation. Both scaffolds revealed good biocompatibility with no adverse inflammatory side effects. Furthermore, the defect area with transplanted HA scaffold showed extended bone formation not only on its surface but also in the pores of the scaffold with the infiltration of connective tissues present within HA particles. This showed that the HA scaffolds used in this study is a porous scaffold that allows cell migration and proliferation. The presence of connective tissue inside the HA scaffold may also resulted from the interaction of MC3T3-E1 cells and host bone-forming cells with the osteoconductive property of HA scaffolds. Wypych (2018) referred to osteoconduction as the ability of bone-forming cells in the grafting area to move across a scaffold and slowly replace it with a new bone over time. These results are in agreement with Sulaiman et al. (2013) and Shao et al. (2018) that demonstrated bone formation for ceramic scaffold, started on the surface and proceeded to the center of the pores. Meanwhile, bone tissue formed on PCL scaffold as a result of the extensive proliferation of differentiation of cells from the surface of the PCL fibers that further grow in between several fibers and encapsulate them over time. There is no evidence to suggest that PCL has osteoinductive or osteoconductive properties. Therefore, factors such as enhanced cell-to-cell and cell-to-substrate interaction might contribute to the observed new bone formation on the PCL scaffold ( Rumiński et al. , 2018 ). Nonetheless, H&E staining is insufficient to conclude the types of tissue relevant to osteogenesis. Hence, the identification of new bone formation in scaffolds during in vivo study should be determined using more complex staining methods. However, in this study, new bone formation was not evaluated using Masson-Goldner trichrome, modified Masson-Goldner trichrome, or Movat’s pentachrome staining but by a stronger expression of ALP and OCN markers using immunohistochemistry (IHC) analysis. Other studies did not perform complex staining methods but used IHC staining to demonstrate the changes in osteoblast activity by the expression of ALP and OCN markers as an osteogenic maturation and occurrence of mineralization marker at the bone-connective tissue interface ( Abbasi et al. , 2020 ; Christenson, 1997 ; Jeon et al. , 2014 ; Yasui et al. , 2016 ). The IHC analysis supported micro-CT and histological observation which revealed the HA scaffold has a greater potential to augment maxillary bone regeneration in rats. The IHC staining showed that new bone tissues formed at the maxilla defect area were ALP and OCN positive. ALP is considered an intermediate marker and is secreted during osteoblast maturation and matrix mineralization ( Khanna-Jain et al. , 2012 ). Meanwhile, OCN expression is presented as a late marker of osteogenesis and is consistent with characteristics of mature bone ( Freire et al. , 2015 ). Stronger expressions of ALP and OCN were observed in all treatment groups that received HA scaffolds suggest that the osteogenic process was more advanced in the HA scaffold than the PCL scaffold. Nevertheless, OCN showed intense expression compared to ALP which prominently could be observed on the MC3T3-E1-HA group followed by MC3T3-E1-PCL, empty HA, and empty PCL groups. Significant differences were found between these markers, especially on the scaffold with cells groups compared to empty scaffold groups. Similar results have been observed in other studies evaluating the performance of HA and PCL-based scaffolds for in vivo bone models ( Abbasi et al. , 2020 ; Johari et al. , 2016 ). Thus, the results from these studies strongly support the stronger expression of OCN compared to ALP due to the enhanced osteogenic behavior of the scaffold. An increase in the cell populations residing in the scaffolds could lead to the intense production of osteogenic markers and subsequently accelerate the differentiation process ( Zhou et al. , 2020 ). The results from the present study indicate that HA and PCL scaffolds have the potential to repair rat’s maxillary defects with properties applicable for bone tissue engineering. Moreover, it was concluded that the transplantation of cells and HA scaffolds showed better in vivo bone regeneration potential with enhanced new bone volume, increment of bone surface density, and new bone formation as shown by micro-CT, histology, and IHC analyses. Conclusions In conclusion, HA and PCL scaffolds used in this study can support in vitro cell viability, attachment, morphology, and osteoblast differentiation following the accepted model of the two-dimensional system. Both scaffolds demonstrated good bone regeneration properties with no significant inflammatory reaction. However, HA scaffolds showed better new bone formation when transplanted on the maxillary bone defect of rats compared to PCL scaffolds. This confirmed in vitro growth and osteoblast potential of HA scaffolds. Based on the obtained result, it is suggested that HA could be considered as a potential scaffold for clinical use in maxillary bone regeneration. Supplemental Information 10. 7717/peerj. 13356/supp-1 Supplemental Information 1 Immunoreactive score (IRS) assigned for the semi-quantitative immunohistochemical evaluation of the ALP and OCN expression Click here for additional data file. 10. 7717/peerj. 13356/supp-2 Supplemental Information 2 Comparison of a number of viable MC3T3-E1 cells when cultured on HA scaffolds, PCL scaffolds, and 2D culture plates One-way ANOVA was conducted to compare MC3T3-E1 cell viability when cultured on scaffolds and 2D culture plates in terms of a number of viable cells (1 × 10 4 cells). Values were mean difference ± standard deviation. *Asterisks indicate significant differences after a Bonferroni correction for n = 5, at p < 0. 05. MTT assays for cell viability were carried out in triplicate. Click here for additional data file. 10. 7717/peerj. 13356/supp-3 Supplemental Information 3 Raw data In vitro cell viability analysis using MTT assay as well as ALP specific activity using ALP assay. In vivo analysis comprising micro-CT and hematoxylin, eosin, and immunoreactive score (IRS) analyses are also shown. Click here for additional data file. 10. 7717/peerj. 13356/supp-4 Supplemental Information 4 ARRIVE Checklist Click here for additional data file.

10. 7717/peerj. 13386

PeerJ

Discovery of protein-based natural hydrogel from the girdle of the ‘sea cockroach’

Hydrogels are widely used materials in biomedical, pharmaceutical, cosmetic, and agricultural fields. However, these hydrogels are usually formed synthetically via a long and complicated process involving crosslinking natural polymers. Herein, we describe a natural hydrogel isolated using a ‘gentle’ acid treatment from the girdle of a chiton species ( Chiton articulatus ). This novel hydrogel is shown to have a proliferative effect on mouse fibroblast cells (cell line, L929). The swelling capacity of this natural hydrogel was recorded as approximately 1, 200% in distilled water, which is within desired levels for hydrogels. Detailed characterizations reveal that the hydrogel consists predominantly (83. 93%) of protein. Considering its non-toxicity, proliferative effect and swelling properties, this natural hydrogel is an important discovery for material sciences, with potential for further applications in industry. Whether the girdle has some hydrogel activity in the living animal is unknown, but we speculate that it may enable the animal to better survive extreme environmental conditions by preventing desiccation.

Introduction Hydrogels are three-dimensional hydrophilic networks that can absorb large amounts of water without dissolving ( Liao & Huang, 2020 ; Schiller & Lai, 2020 ; Varaprasad et al. , 2017 ). This ability of hydrogels to take up large amounts of water by swelling makes them desirable materials for drug delivery ( Beninatto et al. , 2019 ; Hoare & Kohane, 2008 ), as a matrix for the growth of artificial organs, wound dressing, cancer treatments and tissue engineering ( Chao, Chen & Liu, 2020 ; Fan et al. , 2019 ; Griffin et al. , 2021 ; Ji et al. , 2021 ; Zhong et al. , 2020 ), food coating, agriculture ( Klein & Poverenov, 2020 ), antimicrobial delivery ( Jayaramudu et al. , 2019 ) and wastewater treatment ( Jing et al. , 2013 ). However, most commercially available hydrogels are formed synthetically from petroleum-based plastics or by combining natural biopolymers with crosslinking agents ( Fan et al. , 2019 ; Zhong et al. , 2020 ). These synthetic hydrogels are often preferred due to their high processability for customization, both in terms of their chemical composition and the ease with which mechanical properties can be adjusted, greater reproducibility, long term stability and the wide variety of easily accessible raw materials ( Samadian et al. , 2020 ; Voorhaar & Hoogenboom, 2016 ). However, excessive use of synthetic hydrogels can in some instances give rise to enormous health and environmental problems ( Liao & Huang, 2020 ). In addition, the biocompatibility and non-toxicity of synthetic hydrogels cannot be guaranteed ( Gajendiran, Rhee & Kim, 2018 ; Samadian et al. , 2020 ). Commercially available ‘natural’ hydrogels (such as cellulose, chitosan, starch, alginate and gelatin) have some superior physicochemical and biological properties when compared with synthetic hydrogels ( Elvira et al. , 2002 ; García-Astrain et al. , 2016 ; Jiang & Kobayashi, 2017 ; Paukkonen et al. , 2017 ; Zhang et al. , 2019 ). The most important of these are their sustainability, environmentally friendly composition, low immunogenicity, excellent biocompatibility and cytocompatibility, biodegradability, specific cellular responses, presence of antigens, cell proliferation controllability and 3D geometric structures ( Choi et al. , 2019 ; Liao & Huang, 2020 ; Samadian et al. , 2020 ; Shi et al. , 2016 ; Zhong et al. , 2020 ). Although these hydrogels are described as ‘natural’, they are not naturally sourced, but are formed in a laboratory by crosslinking of natural polymers ( e. g. cellulose and chitosan) after dissolution in a chemical solution. These processes can lead to poor mechanical properties and increased toxicity ( Jiang & Kobayashi, 2017 ; Paukkonen et al. , 2017 ). In addition, difficulties in obtaining sufficient quantities of natural starting polymers, the necessity for two or more steps in the production process and high production costs of the ‘natural’ crosslinked-polymer hydrogels further limit their utility ( Hoare & Kohane, 2008 ; Taylor & in het Panhuis, 2016 ). It is extremely desirable to identify a completely natural hydrogel that would eliminate the need for toxic synthetic polymers and chemicals in their production. For this reason, marine organisms that are abundant in nature may serve as an alternative source to produce hydrogels with desirable properties ( Varaprasad et al. , 2020 ). Herein, we report a completely natural hydrogel obtained directly from chiton girdle tissue with a ‘gentle’ acid treatment, and without the use of any cross-linkers or synthetic polymers. Chitons (Polyplacophora: Mollusca) are slow-moving, bilaterally symmetrical animals with eight calcareous plates overlying a dorsoventrally flattened body and a large muscular foot. Most species live on hard substrates and although they feed by grazing, seven different ecological feeding strategies have been identified ( Sigwart & Schwabe, 2017 ; Sirenko, 2000 ). Within the Polyplacophora, there are more than 922 recent and 368 fossils valid species worldwide ( Schwabe, 2005 ). The edible chiton, Chiton articulatus, known locally as the ‘sea cockroach’ or ‘dog’s tongue’, is a large species, occurring in high densities on the rocky intertidal splash zone along the Mexican Tropical Pacific ( Bullock, 1988 ; Reyes-Gómez, 2004 ). It occurs most commonly in sites where animals are exposed to strong wave action, although their distribution is likely affected by human fishing efforts, which may reduce abundances in more easily accessible areas ( García-Ibáñez et al. , 2013 ; Holguin-Quiñones & Michel-Morfín, 2002 ). The tissue used in this study is the integument tissue around the chiton body scleritome known as the girdle ( Schwabe, 2010 ). Girdle morphology is modified in some chiton species, reflecting adaptations to different lifestyles; in some carnivorous species it is expanded and acts as a net to trap small prey items, other species have a slit in the girdle to facilitate excretion of waste ( Schwabe, 2010 ). The chiton girdle is covered by a thick, organic, glycoproteinaceous cuticle ( Eernisse & Reynolds, 1994 ) composed of a very thin, electron-dense, outer layer and a thick inner electron translucent layer composed of lamellae ( Checa, Vendrasco & Salas, 2017 ). The girdle is often ornamented with calcareous and chitinous elements including scales, spines, spicules, needles, bristles and hairs that differ in appearance and occurrence among species ( Eernisse & Reynolds, 1994 ). The appearance of armature provided by these small, mineralised scales has inspired the design of flexible, man-made armour ( Connors et al. , 2019 ), although their biological function for the animal is uncertain ( Leise & Cloney, 1982 ). During an attempt to isolate chitin from the girdle of C. articulatus, a new natural hydrogel was discovered, which we discuss in this study; however, no hydrogel formation was observed in the girdle of two other chiton species: Plaxiphora aurata (Mopaliidae, Mopaliinae) or Rhyssoplax olivacea (Chitonidae, Chitoninae). The properties of the sea cockroach hydrogel were determined using Fourier-transform infrared spectroscopy (FTIR), thermogravimetric analysis (TGA), X-ray diffraction analysis (XRD), scanning electron microscope (SEM) and energy-dispersive X-ray spectroscopy (EDS). In addition, 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-Diphenyltetrazolium Bromide (MTT) assay and Real-Time, Quantitative Cell Analysis, xCELLigence system were used to test for cytotoxic effects. Materials and Methods Sample collection and materials Adult specimens of Chiton articulatus ( n = 80; 45 mm ≤ total length ≤ 80 mm) were collected from rocky intertidal shore at the northern limit of its geographical distribution, in Barras de Piaxtla, Sinaloa, Mexico (23°38′51. 0″N 106°48′16. 6″W), following collecting regulations established under Mexican law (NOM-126-SEMARNAT-2000). Individuals were relaxed by leaving them in a refrigerator at 7 °C for 30 min before dissection, following regulations for the humanitarian killing of animals as established under Mexican law (NOM-033-SAG/ZOO-2014). Subsequently, the soft tissues (foot, gills, gonads and viscera) were removed to allow the chiton scleritome (the eight interconnecting shelly plates) together with the girdle to be isolated ( Fig. S1 ). Each chiton scleritome and girdle were sun-dried for 2 days prior to further processing. Once dry, samples were stored away from further exposure to heat and light until the demineralization process. Protein-based hydrogel production The scleritome and girdle were washed several times with distilled water to remove any possible dirt and particles. Shell plates were dissected from the body structure and the softened girdle structure was separated intact. Then to remove the calcareous shell layer on the surface of the girdle, specimens were treated with 0. 5 M HCl (Sigma Aldrich, St. Louis, MO, USA) over 24 h. The integrity of the girdle structure was disrupted after the acid treatment, and pieces of girdle tissue were placed on dialysis membrane (Seamless Cellulose Tubing, size: 16/32, lot: 208001; Viskase Sales Corp. , Chicago, IL, USA) and kept in distilled water with frequent changes of water over 3 days to ensure complete removal of the acid. Afterwards, the hydrogel samples were dried in an oven at 60 °C for 1 day. Although the same procedures were attempted, hydrogel formation was not observed in two other chiton species Plaxiphora aurata and Rhyssoplax olivacea. Characterization Infrared spectra of hydrogel were obtained using a Perkin Elmer Spectrum FT-IR Spectrometer fitted with a Universal Attenuated Total Reflectance at 8 cm −1 resolutions in the wavelength range of 600–4, 000 cm −1. This analysis, based on the principle of refracting X-rays in a characteristic order depending on the atomic sequences of each crystal phase, was carried out with Bruker AXS D8 Advance in the range of 40 kV, 30 mA, 2θ at scanning range of 5–45°, with the final result based on the average of 64 scans to improve the signal-to-noise ratio. The percent crystallinity of the hydrogel sample was determined by using the intensity of the peaks obtained from XRD analysis. Crystallinity was calculated according to the following formula; (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$${\rm CrI}_{110} = [({\rm I}_{110}-{\rm I}_{\rm am}) / {\rm I}_{110}] \times 100$$\end{document} C r I 110 = [ ( I 110 − I a m ) / I 110 ] × 100 CrI = % crystallinity value, I 110 = maximum intensity value at 2θ = 20°, I am = maximum intensity value of the amorphous peak at 2θ = 13°. The thermal stability of the hydrogel was determined using a Thermogravimetry/Differential Thermal Analyzer (TGA Exstar-TG/DTA 7300 Instruments). The analysis was carried out under a nitrogen atmosphere at a heating rate of 10 °C/min in the range of 30–730 °C using a platinum crucible. The surface morphology of the hydrogel was revealed by scanning electron microscopy (SEM) at 5 kV over a range of different magnifications (500×–30, 000×). The swollen protein-based hydrogel sample sheets were freeze-dried for 24 h at −20 and −80 °C. To improve image quality, the material was gold-plated before scanning, using a Cressington sputter-coated 108 Auto. Energy dispersion spectrum (EDS) of protein-based hydrogel was measured at 20 kV and 5, 000× magnification (EDAX-Octane Pro). To determine the percentages of C, N, O and H elements in the structure of the protein-based hydrogel, high precision elemental analysis using SEM was conducted with a Thermo Flash 2000 microscope. The Kjeldahl method was used to measure the nitrogen content of the dried hydrogel (1 g) following official methods ( Bradstreet, 1954 ). After determining the total nitrogen content of the samples, crude protein content was calculated using a conversion factor of 6. 25 to convert % nitrogen to % crude protein since most meat proteins characteristically contain 16% nitrogen ( Salo-väänänen & Koivistoinen, 1996 ). Each sample was analysed in duplicate. Swelling study The swelling behaviour of hydrogel samples was gauged by placing samples in distilled water and aqueous solutions of 0. 9% NaCl, MgCl 2, CaCl 2 and FeCl 3 by weight using the gravimetric technique. Until swelling equilibrium was reached, dry and constant weight hydrogel (0. 05 g) samples were kept in distilled water and salt solutions at room temperature. The swollen hydrogel was then placed on blotter paper and weighed after removing excess water from the surface. The swelling ratio was calculated according to the following equation; (2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$${\rm Swelling\ ratio}\ (\%) = \displaystyle{{Ws - Wd} \over {Wd}} \times 100\%$$\end{document} S w e l l i n g r a t i o ( % ) = W s − W d W d × 100 % where wd is the dry sample weight; ws is the weight of the swollen sample. The water retention rate of the protein-based hydrogel was measured at different temperatures. Samples reached swelling equilibrium in distilled water were incubated in the oven at 30, 50 and 80 °C and weighed every hour for 12 h. Water retention rate (WR); (3) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$${\rm WR}\ (\%) = \displaystyle{{Wt - Wd} \over {Ws - Wd}}\times 100\%$$\end{document} W R ( % ) = W t − W d W s − W d × 100 % where wt is the weight of the sample at time t. Other arguments are the same as those defined earlier. To evaluate the reusability of the protein-based hydrogel, the dried products (0. 05 g) were suspended in distilled water (100 mL) at room temperature until reaching swelling equilibrium. The swollen samples were then weighed and the water holding capacity calculated according to the equation ( Eq. (2) ). Afterwards, the swollen sample was dehydrated in the oven at 60 °C until it reached a constant weight. For the next swelling experiment, an equal volume of distilled water was added to the recovered hydrogel sample. To determine the re-swelling capacity of the protein-based hydrogel, the same procedures were repeated five times and recorded. Determination of cytotoxic effects of the protein-based natural hydrogel obtained from the girdle of C. articulatus Cell culture Mouse fibroblast cells (cell line L929) obtained from Sap Institute, Ankara, Turkey were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Biological Industries ®, Cromwell, CT, USA) supplemented with 10% Fetal Bovine Serum (FBS; Biowest ®, Riverside, MO, USA), 1% Penicillin-Streptomycin solution (Biowest ®, Riverside, MO, USA) at 37 °C, 5% CO 2 humidified incubator and checked daily by using an inverted microscope (Leica DM IL LED; Leica Microsystems, Wetzlar, Germany). The cells were passaged at 80–85% confluence. Cytotoxicity assay The cytotoxic effect of hydrogel was evaluated using an MTT analysis, with L929 cells seeded in 96-well plates at 1 × 10 4 cells per well in 100 µl of complete DMEM. After 24 h of incubation, three different amounts of sterilized blown protein-based hydrogel (1, 2, and 4 mg) were placed in contact with the cells and then incubated from 24 h to 72 h at 37 °C in a humidified atmosphere of 5% CO 2. Following the incubation period, sterilized hydrogel pieces were removed and MTT (0. 5 mg/mL) solution was added to each well and further incubated for 4 h at 37 °C. Afterward, the reaction mixture was removed from each well, replaced by 100 µl of dimethyl sulfoxide (DMSO) solution, and the optical density (OD) was measured at 492 nm by ELISA reader (ChroMate ® ; Awareness Technology, Inc. , Palm City, FL, USA). The measured absorbances were directly proportional to the number of living cells and the viability in hydrogel-free control groups was calibrated to be 100%. Cell viability in the treatment groups was compared with the control group and calculated by the following equation; (4) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$$\% {\rm{Viable}}\;{\rm{cells}} = (({{\rm{A}}_1} - {{\rm{A}}_0})/({{\rm{A}}_2} - {{\rm{A}}_0})) \times 100\% $$\end{document} % Viable cells = ( ( A 1 − A 0 ) / ( A 2 − A 0 ) ) × 100 % where A 0 is the absorbance of the blank (medium without mouse cells), A 1 absorbance of 1, 2 and 4 mg hydrogels, A 2 is the absorbance of the control (cells in solution, grown without any hydrogel). Data analysis of MTT results was performed using GraphPad Prism software version 8 (GraphPad Software ®, San Diego, CA, USA). The data represent the mean ± SD (standard deviation). Statistical differences were evaluated by two-way ANOVA with a Bonferroni correction (95% confidence interval). Values with p < 0. 05 were regarded as statistically significant. An xCELLigence Real-Time Cell Analysis instruments (RTCA; ACEA Biosciences, Roche, Germany) system was also used for real-time and label-free online monitoring of cytotoxicity ( Stefanowicz-Hajduk & Ochocka, 2020 ). The 1 × 10 4 cells/well were seeded in a 16-well e-plate (RTCA; ACEA Biosciences, Roche, Germany) and then the cell index (CI) was measured every 15 min for 96 h. After the growth period of cells, the hydrogel pieces (1 and 2 mg) were added to 16-wells in the e-plate. The CI value of the control (cell with DMEM) and hydrogel treated samples were graphed by using the xCELLigence RTCA software version 2. 1. 0. Results A novel hydrogel As a serendipitous result of an attempt to characterize chitin in girdle tissue from the ‘sea cockroach’, Chiton articulatus, we discovered a new, natural hydrogel, which can be produced in just a few, simple steps and without the need for expensive equipment or toxic chemicals. Figure 1 describes the process of converting the chiton girdle to a protein-based hydrogel. 10. 7717/peerj. 13386/fig-1 Figure 1 Process for producing hydrogel from chiton girdle. First, the girdle structure is separated from the chiton body scleritome by dissection. The separated girdles are subjected to 0. 5 M HCl. Then the samples are filtered by rinsing with distilled water and the natural hydrogel is produced. Characterization The FTIR spectrum of C. articulatus hydrogel isolated from the girdle structure is shown in Fig. 2A. For the hydrogel isolated from C. articulatus, Amide I, Amide II and Amide III bands were recorded at 1, 625 cm −1, 1, 537 cm −1 and 1, 235 cm −1, respectively. These recorded peaks suggest that the hydrogel is protein-based. Other peaks recorded during FTIR analysis are given in Table S1. 10. 7717/peerj. 13386/fig-2 Figure 2 Characterization of the protein-based hydrogel by (A) FT-IR, (B) XRD, (C) TGA, (D) SEM, −20 °C freeze-dried, (E) SEM, −80 °C freeze-dried, (F) EDS and elemental analysis. The crystal structure of the protein-based hydrogel isolated from C. articulatus was revealed by XRD measurement ( Fig. 2B ). The XRD peaks of the hydrogel obtained from the girdle structure of C. articulatus were found to be compatible with previous protein studies and the crystallinity was calculated as 84. 9% ( Alashwal, Gupta & Husain, 2019 ; Ki et al. , 2007 ; Meng et al. , 2012 ). Results of Thermogravimetry (TGA) and Differential Thermal analyses (DTA) of the protein-based hydrogel isolated from C. articulatus are given in Fig. 2C. The peak clearly observed in the figure is due to degradation of the protein. The maximum decomposition temperature value was observed as 305. 9 °C and 77. 5% mass loss was recorded in this decomposition. Surface morphology and pore structure of hydrogels are the main characteristics used to determine their potential applications ( Bashir et al. , 2020 ). Therefore, to reveal the surface properties of the protein-based hydrogel, SEM analysis was conducted after drying samples using two different lyophilization techniques, and the surface morphologies are presented in Figs. 2D and 2E. The pore sizes range in size, with the largest reaching 100 µm. The elemental composition of the hydrogel was determined to be 11. 63% nitrogen, 44. 45% carbon, 6. 46% hydrogen and 0. 68% sulfur ( Fig. 2F ). In the EDS analysis, C, N, H and S as well as very small amounts of Na, Mg and Cl were detected in protein-based hydrogel. The result of Kjeldahl analysis revealed that the chiton hydrogel contains as 83. 93 g/100 g protein by dry weight. Swelling study The swelling rate (%) of the protein-based hydrogel over time in distilled water and saline solution types is given in Fig. 3A. It clearly shows a similar swelling behaviour when placed in distilled water, and 0. 9% by weight aqueous solutions of FeCl 3, MgCl 2, CaCl 2 and NaCl, to that exhibited by other hydrogel samples. In the saline solutions, it is evident that the swelling of the hydrogel samples is dramatically reduced compared to its uptake of distilled water. The maximum water absorption values of the samples in distilled water, FeCl 3, MgCl 2, CaCl 2 and NaCl were 1, 200%, 450%, 400%, 350%, 325%, respectively. Figures 3D – 3F shows the morphology after swelling. 10. 7717/peerj. 13386/fig-3 Figure 3 Swelling ratio. (A) Swelling ratio of hydrogels in distilled water and different salt solutions (0. 9% by weight): NaCl, CaCl 2, MgCl 2 and FeCl 3, (B) reswelling capability of protein-based hydrogel, (C) effect of temperature on water-retention (%) of protein-based hydrogel, (D) the natural hydrogel swells and keeps their structural integrity, (E) the image of swelled hydrogel taken by light microscopy, (F) an inverted test tube holds the protein-based hydrogel. The water absorbency of the protein-based hydrogel in distilled water and 0. 9 wt% NaCl solution as a function of the number of reswelling time was shown in Fig. 3B. Looking at the swelling capacity of five cycles in distilled water and 0. 9% NaCl solution, it was seen that the reswelling capacity of the hydrogel samples gradually decreases as the number of reswelling cycles increases ( Fig. 3B ). It can be seen that protein-based hydrogel samples were still able to maintain high water content even after five cycles of swelling and drying: approximately 546% in distilled water and 213% in 0. 9 wt% NaCl solution. Figure 3C showed the water-retention capacity of the protein-based hydrogel at three temperatures (30, 50 and 80 °C). The absorbency of the samples kept at 32 and 10 wt% after heating at 30 and 50 °C for 10 h, respectively, and 0. 8% at 80 °C for 6 h. In addition, it is understood from Fig. 3C that the water absorbed in the hydrogel can be released with increasing temperature. Cytotoxicity assay To assess the level of biocompatibility of the C. articulatus hydrogel for medical uses, different amounts of hydrogel were applied to the L929 mouse fibroblast cells ( Figs. 4A and 4B ). Cell viability according to MTT assay results for control (without any hydrogel), 1, 2 and 4 mg hydrogel-treated wells at 24 h were determined as 100%, 124. 2%, 98. 15%, and 83. 01% respectively. Viability changed to 100%, 132. 3%, 73. 46%, and 22. 15% at the end of 48 h, and 100%, 144. 9%, 62. 99%, and 20. 13% after 72 h. Accordingly, the viability of the cells treated with 1 mg hydrogel was significantly increased in all three incubation times (24, 48 and 72 h) compared to the control ( p < 0. 05). In addition, a significant decrease in cell viability was observed at 48 and 72 h in 2 and 4 mg hydrogel applications ( p < 0. 05). 10. 7717/peerj. 13386/fig-4 Figure 4 (A) Cell viability assays of mouse fibroblast L929 cells after being cultured with chiton hydrogel (1, 2, and 4 mg) for 24 h, 48 h and 72 h (bars represent mean cell viability ± SD; n = 3 statistical difference: * p < 0. 05, ** p < 0. 01, *** p < 0. 001), (B) real-time analysis. The MTT assay results were then validated with a real-time experiment examining cell viability in the xCELLigence system. The xCELLigence system measures cellular changes under real-time conditions as an electric impedance of the e-plate. We treated L929 non-cancerous mouse fibroblast cells with the 1 and 2 mg concentrations of hydrogel by xCELLigence system. The real-time analysis demonstrated a gradual increase of the cell index with 1 and 2 mg concentrations of hydrogel for 25 h on L929. Moreover, 1 mg hydrogel showed a proliferative effect on L929 cells until the 91st h (67 h after the addition of hydrogel) ( p < 0. 05). MTT and xCELLigence assay results are given in Fig. 4B. In the light of these important results, the absence of cytotoxic effects of hydrogel-based biological materials suggests C. articulatus hydrogel shows a significant potential for biotechnological and biomedical applications. Discussion Characterization Amide I (1, 600–1, 700 cm −1 ), Amide II (1, 504–1, 582 cm −1 ) and Amide III (1, 200–1, 300 cm −1 ) absorption bands are characteristic for protein-based materials ( Castrillón-Martínez et al. , 2017 ). The wide peak recorded at 3. 275 cm −1 in the spectrum is due to the vibration of the О–Н group connected by intermolecular hydrogen bonding of the absorbed water ( Pourjavadi et al. , 2006 ). These peaks are consistent with the results of the FTIR analysis of hydrogels prepared from collagen and silk fibroin proteins in the literature ( Montalbano et al. , 2018 ; Motta et al. , 2004 ). According to the XRD analysis, the peaks clearly observed for the hydrogel are 9. 7 and 19. 6°. It has been reported in the literature that the peaks recorded for other structural proteins such as collagen, keratin and silk are in the range of 7. 8–10. 7° and 19. 6–21. 8° ( Alashwal, Gupta & Husain, 2019 ; Ki et al. , 2007 ; Meng et al. , 2012 ). The maximum decomposition temperature recorded for protein-based hydrogel isolated from the girdle structure of C. articulatus was 305. 9 °C, which was slightly lower than decomposition temperatures reported for other structural proteins (collagen, keratin, fibroin, etc. ) in the literature ( Chomachayi et al. , 2020 ; Kakkar et al. , 2014 ; Mekonnen, Ragothaman & Palanisamy, 2017 ). The maximum decomposition temperatures recorded for hemicellulose-based hydrogels produced in a previous study were 300–375 °C ( Guan et al. , 2014 ). In another study, the degradation temperature for starch based hydrogel was recorded as 317–322 °C ( Vakili & Rahneshin, 2013 ). Synthetic hydrogels have higher thermal stability than biopolymer-based composite hydrogels. In an earlier study, decomposition of hydrogel backbone for surfmer-co-poly acrylates crosslinked hydrogels occurred 315–430 °C ( El-Hoshoudy et al. , 2019 ). In another study, in the thermogram recorded for hybrid polyacrylamide hydrogels, the degradation of cross-linked polymers occurred at 405–560 °C, while the crosslinker decomposed at 560–794 °C. In the same study, it was noted that the decomposition temperature increased as the crosslinker ratio increased ( Nadtoka et al. , 2018 ). More than one degradation step was recorded in all of the TGA analyzes performed on hydrogels in the literature ( El-Hoshoudy et al. , 2019 ; Guan et al. , 2014 ; Kumar, Kaith & Mittal, 2012 ; Lessa, Nunes & Fajardo, 2018 ; Nadtoka et al. , 2018 ; Peng et al. , 2015 ; Rusu et al. , 2015 ; Vakili & Rahneshin, 2013 ). These results reveal that synthetic hydrogels are composed of many different components, including crosslinkers. In the present study, the presence of a single degradation peak for the protein-based hydrogel isolated from C. articulatus, suggests that it consists of almost entirely of protein. This is confirmed by the result of the Kjeldahl analysis, which revealed that the chiton hydrogel contains 83. 93% protein by dry weight. Similarly high concentrations of crude proteins have been observed in fish protein-based hydrogel ( Hwang & Damodaran, 1997 ) and soy protein-based hydrogels, which are typically purified for industrial and research fields ( Abaee, Mohammadian & Jafari, 2017 ). As seen in previous studies of natural hydrogels ( Polat, Duman & Tunç, 2020 ; Qi et al. , 2020 ; Zhang et al. , 2020 ), a porous and rough morphology was observed in both protein-based hydrogel samples, however the first drying technique ( Fig. 2D ) makes the hydrogel more porous than the second drying technique ( Fig. 2E ). The pore sizes range in size, with the largest reaching 100 µm. Surface morphologies and pore structures of hydrogels vary considerably according to the drying technique and the solvents used ( Kumar & Han, 2017 ; Laftah, Hashim & Ibrahim, 2011 ; Mahinroosta et al. , 2018 ; Ullah et al. , 2015 ). Hence the effects of using other drying techniques and solvents on C. articulatus hydrogel need to be investigated in further studies. Our elemental analysis values are similar to those reported for some other structural proteins such as keratin, sericin and fibroin ( Jena et al. , 2018a ; Jena et al. , 2018b ; Shavandi et al. , 2017 ; Xia & Lu, 2008 ). It has been reported in the literature that Mg and Na elements are associated with calcium phosphates ( León-Mancilla et al. , 2016 ). These elements are thought to be present because the upper layer of the girdle structure, from which the hydrogel is obtained, is completely covered with shell plates ( Connors et al. , 2019 ). In a previous study, C, H, N, O and S were similarly detected in chitosan, silk fibroin and egg shell membrane hydrogels ( Adali, Kalkan & Karimizarandi, 2019 ). In addition to these, C, O, N and H were also detected in the synthetically produced bioconjugated graphene oxide hydrogel ( Soleimani, Tehrani & Adeli, 2018 ). In another study, high levels of P as well as C and H were recorded in the synthetically produced phosphate-containing hydrogel ( Wang et al. , 2003 ). This result shows that these detected elements in the synthetic hydrogels are related to the plastic polymeric materials added to the composition of the hydrogel. Swelling study The ability of hydrogels to take up large amounts of liquids means they have the potential for the sustainable release of absorbed molecules, and this capacity is of primary importance in many practical applications such as water release systems in agriculture ( Hosseinzadeh, 2013 ). Properties such as charge of cations and salt concentration greatly affect the swelling behaviour of hydrogels ( Tanan, Panichpakdee & Saengsuwan, 2019 ). Looking at the literature to compare the swelling capacities of synthetic and natural hydrogels, in the study of Tanan, Panichpakdee & Saengsuwan (2019), it was shown that the maximum swelling capacity in water was determined as 794% in biodegradable hydrogel based on ‘natural’ polymers. However, in synthetic hydrogels, the percentage of maximum swelling ratio of the hydrogel has been shown to be up to 2, 000% the dry mass ( Kiran, Krishnamoorthi & Kumar, 2019 ). The swelling rate of hydrogels in salt solutions is thought to depend not only on salt concentration but also on ionic charge. In the study of Tanan, Panichpakdee & Saengsuwan (2019), it was clearly shown that as the charge of cations increases, the swelling capacity decreases accordingly, but in our study, on the contrary, swelling after exposure to FeCl 3, MgCl 2 and CaCl 2 solutions was greater than that observed after exposure to NaCl. Similar results have been reported in the literature, with the swelling capacity of some hydrogels increasing with increasing charge density and decreasing salt concentration ( Chang et al. , 2011 ; Liu, Tong & Hu, 1995 ). However, it has been shown in many studies that the swelling ratio of cations decreases as the salt concentration and value increase ( Li et al. , 2017 ; Namazi, Hasani & Yadollahi, 2019 ; Wang et al. , 2018 ). In addition, it has been shown that the absorbance of nanocomposite hydrogels in salt solutions is relatively higher than that of pure hydrogel ( Namazi, Hasani & Yadollahi, 2019 ). Nevertheless, the swelling rate of hydrogel samples also differs in different salt solutions ( Chang et al. , 2011 ). Similar to this study, Chang et al. (2011) clearly showed that the swelling ratio was significantly decreased with increasing ionic charge in some hydrogels, whereas other hydrogels showed minimal swelling in CaCl 2 instead of FeCl 3. This proves that hydrogels have smart swelling behaviors in aqueous solutions such as NaCl, CaCl 2 and FeCl 3. The reswelling capacity is one of the most important characteristics of hydrogels for the application as a superabsorbent in practice, which shows the stable water absorption ability and water retention ( Li et al. , 2012 ; Tanan, Panichpakdee & Saengsuwan, 2019 ). The reason why the reswelling capacity of hydrogel samples gradually decreases as the number of re-swelling cycles increases is probably due to damage in polymeric network structures ( Tanan, Panichpakdee & Saengsuwan, 2019 ), which may affect the water holding capacity of the hydrogel. But at the same time, it can be seen that protein-based hydrogel samples are still able to retain a high-water content, even after five cycles of swelling and drying. Judging by our results, the C. articulatus hydrogel will prove to be useful as a recyclable and reusable superabsorbent material, as a certain degree of water absorption ability is conserved, even after repeated reswelling cycles. The water-retention capacity can be determined by the interaction of H bond and van der Waals forces between water molecules and hydrogels ( Vudjung et al. , 2014 ; Wen et al. , 2016 ). Our results demonstrated that the water-retention capacity of the chiton protein-based hydrogel decreases almost linearly with the increase in temperature and time. Lv, Wu & Shen (2019) proved that super-absorbent hydrogels (SAH) show poor water-retention capacity when temperature rises. Obviously, the product has extensive properties that can be useful in many areas, primarily in agricultural applications where it could be used to retain moisture and provide nutrients to plants. Cytotoxicity assay Cytotoxicity tests are important assays to determine the biocompatibility of materials to be used in medical applications ( Assad & Jackson, 2019 ) ( e. g. as scaffolds for drug delivery ( Motta et al. , 2004 )). In this study, cytotoxicity analyzes were conducted using different amounts of a novel chiton protein hydrogel, which was obtained without the use of toxic chemicals that could adversely affect natural molecules in the biological structure. Traditional MTT analysis and xCELLigence system showed that the chiton protein-based hydrogel, used at an appropriate level (1 mg), is not only non-toxic, but has a proliferative effect on cells. Additionally, we predict that the hydrogel in this study may show a further increase effect on cell proliferation with the protein-based bioactive components in its structure when used at an appropriate doses and times. The observed time-dependent increase in cytotoxicity in cells with high hydrogel application (for 2 and 4 mg) was associated with the increase in swelling capacity over time and, accordingly, with inadequate media and gas transportation to the cells. Biological function A natural hydrogel, that does not require laboratory polymerisation, was isolated in this study from the girdle of Chiton articulatus (Chitonidae, Chitoninae) ( Fig. 1 ). However, hydrogel was not found in two other chiton species examined: Plaxiphora aurata (Mopaliidae, Mopaliinae) and Rhyssoplax olivacea (Chitonidae, Chitoninae). The fact that hydrogel does not occur in all chiton species examined suggests that the tissue and structures responsible for hydrogel may be a species or clade-specific adaptation. Organisms, like C. articulatus, that live in the tropical, rocky intertidal splash zone must tolerate fluctuating and at times extremely high temperatures, levels of ultraviolet radiation and salinity. The ability to thrive in such extreme conditions relies on adaptations that provide protection against or a means to overcome these challenges. For instance, some studies have shown significant differences in levels of heat shock proteins in chiton species Katharina tunicata and Chaetopleura angulata sampled in different seasons ( Burnaford, 2004 ; Madeira et al. , 2017 ). The chiton Mopalia mucosa has been shown to undertake whole-animal volume regulation as evidenced by weight gain to control osmolarity ( Leise & Cloney, 1982 ; Moran & Tullis, 1980 ), and in some species, girdle elements, such as overlapping scales and hairs that entrap mud and detritus, may help prevent desiccation during low tides ( Leise & Cloney, 1982 ; Moran & Tullis, 1980 ). Whether the girdle in C. articulatus retains in the living animal some of the hydrogel characteristics observed in this study is unknown. We speculate that it may play a biological role, enabling the animal to better survive the extreme environmental conditions it experiences in the rocky intertidal ( Flores-Garza et al. , 2011 ) by preventing desiccation and helping to regulate osmolarity of cells. In fact, in the field, C. articulatus have been observed with the girdle subtly swollen and detached from the body scleritome, revealing a part of each plate (sclerite or valve); this in comparison between the groups of chitons observed (O. H. Avila-Poveda, 2016, personal observation). Furthermore, it has been suggested that changes to the tissue volume of the foot would affect an animal’s ability to attach firmly to its substrate ( McGill, 1975 ), and our finding that tissues associated with the girdle, but not the foot, produce hydrogel is consistent with this idea. Exactly how the hydrogel would function in the animal is unknown and is worthy of future study. Conclusion The biological properties of the novel C. articulatus hydrogel isolated in this study make it suitable for use as an industrial, three-dimensional scaffold material due to its behaviour and high biocompatibility with mouse fibroblast L929 cell growth. Further biological analyses, however, in particular more biocompatibility tests using different cell lines, are needed to determine its potential for use in biomedical applications such as biomedical engineering and regenerative medicine. Of relevance when considering the feasibility of using a chiton protein-based hydrogel for such applications, C. articulatus is currently being considered as a target for aquaculture ( Avila-Poveda, 2020 ). Although reasonably abundant in nature, farmed animals could provide a steady source of raw materials and offers the potential for selective breeding or genetic modification to change or improve hydrogel properties. Although we did not find hydrogels in two other species, surveys of other chiton species may also identify other sources for further novel hydrogels. Supplemental Information 10. 7717/peerj. 13386/supp-1 Supplemental Information 1 Raw data exported from the swelling study (swelling capacity/reswelling/water retention) applied for data analyses and preparation for Fig. 3. Click here for additional data file. 10. 7717/peerj. 13386/supp-2 Supplemental Information 2 Raw data exported from microplate reader ChroMate®ELISA (cytotoxicity assay) applied for data analyses and preparation for Fig. 4A. Click here for additional data file. 10. 7717/peerj. 13386/supp-3 Supplemental Information 3 Supplemental Figure and Table. Click here for additional data file.

10. 7717/peerj. 13442

PeerJ

Lipopolysaccharide-activated macrophages regulate the osteogenic differentiation of bone marrow mesenchymal stem cells through exosomes

Background Periodontal tissue regeneration is the ultimate goal of periodontitis treatment. Exosomes are nanoscale vesicles secreted by cells that participate in and regulate the physiological activities between cells. However, the relationship between inflammatory macrophage-derived exosomes and osteoblast differentiation in periodontitis has not been thoroughly reported. Here, we attempt to explore the role of inflammatory macrophage-derived exosomes in crosstalk with osteoblasts. Methods Porphyromonas gingivalis lipopolysaccharide was used to stimulate macrophages and inflate their inflammatory cellular state. Exosomes were extracted from inflammatory macrophages using supercentrifugation, and their characteristics were detected by transmission electron microscopy, particle size analysis, and Western blotting. Exosome uptake bybone marrow mesenchymal stem cells (BMSCs) was observed by fluorescence microscopy. The effects of exosomes on the BMSC inflammatory response and on osteogenic differentiation were detected by quantitative polymerase chain reaction and Western blot analysis. Alkaline phosphatase activity was tested for verification. Results We successfully extracted and identified inflammatory macrophage-derived exosomes and observed that BMSCs successfully took up exosomes. Inflammatory macrophage-derived exosomes upregulated the expression levels of the inflammatory factors interleukin-6 and tumour necrosis factor-alpha in BMSCs and mediated inflammatory stimulation. Additionally, they inhibited the transcription levels of the osteogenic genes alkaline phosphatase, type I collagen, and Runt-related transcription factor 2 as well as the alkaline phosphatase activity, while the use of the exosome inhibitor GW4869 attenuated this effect. Conclusion Our study shows that macrophages in periodontitis can mediate inflammatory stimulation and inhibit the osteogenic differentiation of bone marrow mesenchymal stem cells through the exosome pathway. Interference with exosome secretion is likely to be a promising method for bone tissue regeneration in inflammatory states.

Introduction Chronic periodontitis is one of the most common diseases in the oral and maxillofacial regions. Gram-negative anaerobic bacteria, especially Porphyromonas gingivalis, are the main pathogenic bacteria of periodontitis. Chronic periodontitis is characterized by the loss of periodontal supporting tissue caused by bacterial infection and the host immune-inflammatory response ( Cochran, 2008 ), leading to alveolar bone absorption and tooth loosening and shedding ( Chen et al. , 2020c ). Reasonable control of periodontitis can help patients regain chewing function, improve quality of life and promote mental health. The core goal of clinical periodontal therapy is still to remove calculus, plaque, and lesions from root surfaces, eliminate local inflammation of periodontal tissue and prevent disease progression. However, the ultimate goal of periodontal therapy, namely the reconstruction of supporting tissue structure, cannot yet be achieved ( Barnes et al. , 2013 ). Periodontal tissue engineering with stem cells as the core element is a promising future treatment of periodontitis. Alveolar bone regeneration is dependent on bone marrow mesenchymal stem cells (BMSCs), periodontal ligament stem cells and osteoblast differentiation. However, BMSCs are also affected by inflammatory processes. It has been reported that inflammatory or immune-related cytokines such as tumour necrosis factor-α (TNF-α), interferon-γ, and interleukin-17 (IL-17) affect BMSC differentiation and induce apoptosis ( Han et al. , 2017 ). Therefore, clarifying the effect of the inflammatory environment on BMSC osteogenic differentiation and exploring the specific mechanism behind it may help reduce the adverse impact of the inflammatory environment on BMSCs and promote their osteogenic differentiation, thus providing a potential treatment for periodontal tissue regeneration. When inflammation is confined to periodontal tissue, macrophages are the earliest immune cells recruited to the inflammatory site. A large number of macrophages can be seen in the periodontal tissue of patients with periodontitis, and a large number of proinflammatory factors and bone resorptive factors, such as interleukin 1β, interleukin 6 (IL-6), and TNF-α, can be detected in gingival crevicular fluid ( Lam et al. , 2014 ). Macrophages play an indispensable role in the occurrence and development of periodontitis. It has been shown that the coculture of macrophages and BMSCs can lead to the upregulation of the receptor activator of NF-κB ligand in an inflammatory environment, thereby affecting osteoclast formation and bone remodelling ( Xiao et al. , 2018 ). Macrophage paracrine signalling is an important pathway that regulates BMSC osteogenic differentiation. The uptake of paracrine vesicles, exosomes, and other cytokines by BMSCs can upregulate the transcription levels of alkaline phosphatase (ALP), Runt-related transcription factor 2 (RUNX2), osteocalcin, and osteopontin and the protein expression levels of RUNX2 and osteocalcin, enhancing the activity of ALP ( Chen et al. , 2021 ). Additionally, lipopolysaccharide-activated macrophages can promote the migration of BMSCs through paracrine IL-6 and inducible nitric oxide synthase ( Lei et al. , 2021 ). In periodontitis, infiltrating macrophages can regulate the function of BMSCs through paracrine cytokines. However, further studies are needed to determine whether exosomes mediate the regulation of BMSCs by macrophages. Exosomes are nanoscale vesicles with a diameter of 30–140 nm that are secreted by cells and contain complex RNA and proteins ( Yang et al. , 2017 ). They function mainly as carriers of information between donor and recipient cells ( Milane et al. , 2015 ). Under normal circumstances, most cells, including immune cells, can secrete exosomes, such as lymphocytes, macrophages, and dendritic cells. Additionally, the secretion amount and contents of exosomes may change under the stimulation of pathogens ( Chen et al. , 2020b ). It has been shown that in the microenvironment of hypoxia and serum deprivation, exosomes from M1-type macrophages can induce the apoptosis of BMSCs by delivering miRNA-222 to BMSCs ( Qi et al. , 2021 ). Additionally, M2-type macrophage-derived exosomal miRNA-5106 induces BMSCs to develop into osteoblasts by targeting salt-induced kinases 2 and 3 ( Xiong et al. , 2020 ). However, the effect of inflammatory macrophage-derived exosomes on BMSCs under the conditions of periodontitis has not been further reported. Therefore, exploring the role of macrophage-derived exosomes in regulating BMSC osteogenic differentiation in periodontitis is expected to provide a new solution for the realization of ideal periodontal tissue regeneration. In this study, we successfully extracted, identified, and characterized inflammatory macrophage-derived exosomes activated by Porphyromonas gingivalis lipopolysaccharide ( P. g -LPS). Additionally, we verified that macrophages could regulate BMSCs by secreting exosomes and inhibiting their osteogenic differentiation in the periodontitis environment, which may provide a potential treatment for periodontal tissue regeneration. Materials and Methods Cell culture and treatment Mouse mononuclear RAW264. 7 macrophages (purchased from Shanghai Cell Bank, Chinese Academy of Sciences) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Waltham, MA, USA) containing 10% foetal bovine serum (Gibco, Waltham, MA, USA) and 1% penicillin–streptomycin in a humid environment containing 5% CO 2 at 37 °C. The cells were inoculated into 6-well plates at a density of 2 × 10 5 cells per well for 12 h and were then pre-treated with 100 ng/ml P. g -LPS (Sigma Aldrich, USA) or phosphate-buffered saline (PBS) (Gibco, Waltham, MA, USA) for 0 h, 1 h, 3 h, 6 h, 12 h, 24 h. The pre-treatment and control group samples were collected at different time points, using a quantitative polymerase chain reaction to analyse the expression differences of pro-inflammatory cytokines IL-6 and TNF- α. The time point with the highest expression of inflammatory genes was selected as the P. g -LPS stimulation time for the follow-up experiment. After 6 h of P. g -LPS stimulation, the cells were washed three times with PBS, then the medium was repalced without P. g -LPS, and the cells continued to culture for 24 h or 48 h. The cell culture supernatants were collected, then the enzyme-linked immunosorbent assay (ELISA) was used to determine the IL-6 and TNF-α protein levels in the supernatants and to verify that the extracellular inflammatory environment in macrophages persisted for more than 48 h after P. g -LPS stimulation. Mouse bone marrow mesenchymal stem cells (purchased from Zhongqiaoxinzhou Biotech, Shanghai, China) were cultured in DMEM/F12 medium (Gibco, Waltham, MA, USA) containing 10% foetal bovine serum and 1% penicillin–streptomycin in a humid environment containing 5% CO 2 at 37 °C. To evaluate the osteogenic differentiation of BMSCs, cells were incubated with corresponding CM supplemented with the ingredients of the osteogenic (OS) induction medium containing 0. 1 µM dexamethasone, 10 mM β-glycerophosphate, and 50 µM ascorbic acid. BMSCs treated with the OS induction medium were then used for the Western blotting and Alkaline phosphatase (ALP) staining assays. Real-time quantitative PCR Total RNA was extracted from cells using the Cell Total RNA Extraction Kit (Novizan, Nanjing, China) and then further reverse transcribed to cDNA using the PrimeScript™ II 1st Strand cDNA Synthesis Kit for real-time polymerase chain reaction (Novizan, Nanjing, China). The qRT-PCR was carried out using ChamQ Universal SYBR qPCR Master Mix (Novizan, Nanjing, China). Every reaction was performed in a final volume of 10 µl containing 4 µl of cDNA, 0. 5 mM of each primer, and 5 µl of ChamQ Universal SYBR qPCR Master Mix. The amplification was carried out as outlined in the instructions. The relative expression levels of the target genes, including IL-6, TNF-α, ALP, COL-I, and RUNX2 were then calculated by the comparative 2-ΔΔCt method using StepOne Software version 2. 1. 22. All the samples were run in triplicate and normalized to GAPDH. Primer sequences are shown in Table 1. 10. 7717/peerj. 13442/table-1 Table 1 Sequences of primers used in quantitative real-time PCR (qRT-PCR) analysis. Gene Primer sequence Forward Reverse TNF-α GGAGGGGTCTTCCAGCTGGAGA CAATGATCCCAAAGTAGACCTGC IL-6 CTTGGGACTGATGCTGGTGACA GCCTCCGACTTGTGAAGTGGTA ALP ACGGCGTCCATGAGCAGAACTA CAGGCACAGTGGTCAAGGTTGG COL-1 GAGTCAGCAGATTGAGAACATCC AGTCAGAGTGGCACATCTTGAG RUNX2 CCCAGGCAGTTCCCAAGCATTT GGTAGTGAGTGGTGGCGGACAT GAPDH AGGTCGGTGTGAACGGATTTG GGGGTCGTTGATGGCAACA Enzyme-linked immunosorbent assay (ELISA) After 6 h of P. g -LPS stimulation, the cells were washed three times with PBS, then the medium was replaced without P. g -LPS, and the cells continued to culture for 24 h or 48 h. The cell supernatant was then collected, and soluble protein concentrations of IL-6 and TNF-α in the supernatant were detected using ELISA kits (Enzyme Immunoassay, Wuhan, China) according to the instructions. The determination ranges were as follows: TNF-α, 25–800 pg/mL; IL-6, 3–120 pg/ml. Samples were evaluated in triplicate. Conditioned culture medium collection The cells were inoculated at a density of 2 × 10 7 cells per well in a 150 mm dish overnight and then stimulated with 100 ng/mL P. g -LPS for 6 h. After three washes with PBS, the cells were placed in DMEM containing 10% foetal bovine serum which hads been centrifuged at 11, 000× g for 20 h to eliminate exosomes and 1% penicillin–streptomycin with or without 10 µM GW4869 (MCE, Princeton, New Jersey, USA) and further cultured for 24 h. The cell culture supernatants were collected. Cells and cell debris in the supernatant were removed by centrifugation at 4 °C at 1, 500× g for 20 min, and particles with a particle size greater than 200 nm were removed by filtration with a 0. 22-µm filter. Conditioned medium (CM) with or without GW4869 was labelled CM+GW4869 and CM, respectively. Extraction, quantification, and identification of exosomes The collected supernatant CM was centrifuged with a high-speed centrifuge (Thermo, Waltham, MA, USA) at 4 °C for 17, 000× g for 15 min, and then residual organelles were removed from the supernatant. A superspeed centrifuge (Thermo, Waltham, MA, USA) was then used at 4 °C and 110, 000× g for 80 min. After discarding the supernatant, the trace white precipitate at the bottom of the centrifuge tube was considered as exosomes, which were resuspended in PBS. Ten-microlitre exosome samples were subjected to lysis on ice for 10 min with the same amount of RIPA lysate, followed by the protein quantification of exosomes according to the instructions of the BCA kit (Beyotime, Shanghai, China). Then, 10 µl of exosomes was dropped onto the copper wire for precipitation for 1 min, and the float was absorbed by filter paper. Then, 10 µl of 2% uranium acetate was dropped onto the copper wire for staining for 1 min. After natural drying at room temperature, exosomes were characterized by a transmission electron microscope (Hitachi, Chiyoda, Japan) at 100 kV. Dynamic light scattering was performed using a particle size analyser (NanoFCM, Xiamen, China) to measure the size distribution of the exosomes. Western blotting, with 12 µl samples of protein per well, was used to detect the expression of the exosome surface-enriched proteins CD9, CD81, and TSG101. Calnexin was used as the internal reference. Labelling of exosomes and incubation of recipient BMSCs The quantitative exosome suspension was labelled with 6 µl of fluorescent labelling probe PKH67 (Merck, Darmstadt, Germany) and incubated at 37 °C for 5 min. Then, the mixture was centrifuged at 190, 000× g for 2 h at 4 °C. The supernatant was discarded and suspended again with PBS to remove the excess dye. BMSCs were inoculated at 5 × 10 5 cells per well in 6-well plates. The labelled exosomes with a final concentration of 50 µg/mL were incubated with BMSCs for 24 h. Fifty microlitres of Hoechst solution was added to each well, and the plates were placed at room temperature for 30 min under dark conditions. The uptake of exosomes was observed under a 600x fluorescence confocal microscope (FV3000; Olympus, Tokyo, Japan). Exosomes and conditioned medium transfection The experiment was divided into four groups: ordinary exosome group, inflammatory exosome group, CM group, and CM+GW4869 group. BMSCs were inoculated in 6-well plates at a density of 2 × 10 5 cells per well for 12 h. Then, for the inflammatory exosome group, cells were treated with 2. 5 ml exosome-free medium containing 50 µg/ml exosomes secreted by macrophages stimulated by P. g -LPS. The CM and CM+GW4869 groups were treated with 2. 5 mL of undiluted CM or CM+GW4869. The ordinary exosome group was treated with 2. 5 ml of exosome-free medium containing 50 µg/ml exosomes secreted by macrophages under normal conditions. Forty-eight hours after transfection, cells were collected for subsequent experiments. Western blotting Proteins in exosomes or cells were lysed and extracted using a whole-protein extraction kit (KeyGEN, Nanjing, China). Then, the proteins were separated by SDS–PAGE and transferred to NC membranes, blocked in 5% skim milk at room temperature for 1. 5 h, and incubated with the following primary antibodies: anti-CD9 (1:300; Abcam, Cambridge, MA, USA), anti-CD63 (1:300; Abcam, Cambridge, MA, USA), anti-TSG101 (1:300; Abcam, Cambridge, MA, USA), anti-calnexin (1:300, Abcam, Cambridge, MA, USA), anti-TNFα (1:1, 000; Abcam, Cambridge, MA, USA), anti-IL-6 (1:1000; Abcam, Cambridge, MA, USA), anti-ALP (1:1, 000, Boaosen, China), anti-RUNX2 (1:1, 000; Abcam, Cambridge, MA, USA), anti-COL-1 (1:1, 000; Abcam, Cambridge, MA, USA), and anti-GAPDH (1:5, 000; Abcam, Cambridge, MA, USA). The blot was then stained with horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody (KGAA35; KeyGEN, Nanjing, China). Finally, the protein expression level was detected using a chemiluminescence detection system. Each experiment was repeated 3 times. Determination of ALP and ALP staining BMSCs were plated in 6-well plates at a density of 2 × 10 5 cells per well for 12 h and incubated for 7 days and 14 days under the same treatment conditions as above, and the solution was changed every 3 days. After 7 days, the cells were collected and lysed, and the total protein amount of each group was determined using the BCA kit (Beyotime, Shanghai, China). After verifying that the total protein amount of the ordinary exosome group, the inflammatory exosome group, the CM group, and the CM+GW4869 group were all at the same level, the alkaline phosphatase activity of each group was detected using the ALP kit (Built, Nanjing, China). Each experiment was repeated 3 times. After 14 days, a BCIP/NBT alkaline phosphatase color development kit (Beyotime, Shanghai, China) was used according to the provided directions. The cells were washed three times with PBS and fixed with 4% paraformaldehyde for 30 min, then treated with BCIP/NBT substrate for 10 h, A microscope was used to analyse the colorimetric changes and a scanner was used to image the stained cells. Absorbance was then measured at 450 nm. Experiments were repeated in triplicate. Data analysis All numerical data are presented as the mean and standard deviation. Comparisons between two groups were conducted using independent samples t -tests. One-way ANOVA was used to analyse differences among more than two groups followed by Tukey’s post-hoc test when data were normally distributed and group variances were equal. All statistical analyses were performed using the GraphPad Prism 8 (GraphPad Software, Inc. , La Jolla, CA, USA). P < 0. 05 was considered significant. Results Release of exosomes from RAW264. 7 cells under P. g -LPS stimulation P. gingivalis LPS-stimulated (100 ng/mL) RAW264. 7 cells produced an inflammatory response. The qPCR results revealed that the gene expression levels of inflammatory factors TNF-α and IL-6 increased with time with the expression levels of these inflammatory factors reaching their highest level at about 6 h. ( Figs. 1A – 1B ). RAW264. 7 cells were pretreated for 6 h for subsequent experiments. ELISA results showed that compared with the control group, the soluble protein TNF-α and IL-6 contents in the 24 and 48 h groups were higher than those in the control group ( Figs. 1C – 1D ), suggesting that the inflammatory state of RAW264. 7 cells lasted for at least 48 h after 6 h of pretreatment with P. gingivalis LPS. 10. 7717/peerj. 13442/fig-1 Figure 1 Macrophage secreted exosomes after P. g -LPS stimulation. (A–B) The mRNA expression of proinflammatory mediators IL-6 and TNF-α of macrophages after stimulation with 100 ng/mL P. g -LPS for 1–24 h. (C–D) The expression of soluble proteins IL-6 and TNF-α in macrophage extracellular environment at 24 h and 48 h after stimulation by 100 ng/mL P. g -LPS for 6 h. (E) Microscopic structure of exosomes (scale bar: 1 µm, 500 nm, 200 nm, 100 nm) in the transmission electron microscope. (F) Exosome particle size distribution. (G) Exosome concentration determination diagram. (H) The expression of exosome surface enriched proteins CD81, CD9, TSG101 and negative marker protein Calnexin, the control group was the RAW264. 7 cell lysate. An asterisk (*) indicates P < 0. 05, two asterisks (**) indicate P < 0. 01, three asterisks (***) indicate P < 0. 001, four asterisks (****) indicate P < 0. 0001, ns indicates P > 0. 05, ns indicates no significant difference. All * are compared with the control group. After 325 ml samples of supernatant from RAW264. 7 exosomes were extracted using the ultracentrifugation ultrafiltration method, the samples were suspended in 200 µl PBS. The protein concentration was approximately 2. 23 mg/mL by BCA protein determination, and the collected RAW264. 7 exosomes were approximately 446 µg. Subsequently, the exosomes were characterized and identified by transmission electron microscopy, particle size analysis, and Western blotting. Under transmission electron microscopy, exosomes were observed as vesicles with a double-layer membrane structure outside and low electron density materials with uneven density inside ( Fig. 1E ). Most exosomes have particle sizes ranging from 70 to 100 nm, with an average particle size of 81. 20 nm ( Figs. 1F – 1G ). Compared with purified exosomes, the exosomal surface-enriched proteins CD9 and TSG101 were expressed in the extracted exosomes with high abundance, while the negative marker protein calnexin was not expressed ( Fig. 1H ). These results indicated that exosomes from RAW264. 7 cells were successfully extracted. Effect of inflammatory macrophage-derived exosomes on BMSC osteogenic differentiation To verify whether inflammatory macrophage-derived exosomes mediate the communication between BMSCs and the regulation of osteogenic differentiation in BMSCs, we incubated BMSCs with inflammatory macrophage-derived exosomes for 24 h, and confocal microscopy results showed that BMSCs successfully absorbed exosomes into the cytoplasm around their nuclei ( Fig. 2A ), which is a prerequisite for exosome-mediated communication. 10. 7717/peerj. 13442/fig-2 Figure 2 Exosome uptake by BMSCs and exosomes mediating the inflammatory response. (A) Images of exosome uptake by BMSCs in fluorescence confocal microscopy. BMSCs nuclei were stained with Hoechst (blue), BMSCs Cytoskeleton were stained with FITC-Phalloidin (red), macrophage-derived exosomes were labelled with PKH67 (green). (B–D) The protein expression of TNF-α and IL-6. (E–F) The mRNA expression levels of TNF-α and IL-6. An asterisk (*) indicates P < 0. 05, two asterisk (**) indicates P < 0. 01, three asterisks (***) indicate P < 0. 001, four asterisks (****) indicate P < 0. 0001, ns indicate P > 0. 05, ns indicates no significant difference. Subsequently, we examined the changes in the gene and protein expression levels ofinflammatory factors TNF-α and IL-6 to determine whether exosomes can mediate the inflammatory state in recipient cells. As shown in Figs. 2B – 2F, BMSCs were treated with ordinary exosomes, inflammatory exosomes, or CM. Compared with the ordinary exosome group, TNF-α, IL-6 gene, and protein expression levels were significantly upregulated in the inflammatory exosome group. When GW4869 inhibited exosome release, compared with the CM group, the expression levels of the TNF-α and IL-6 genes in the CM+GW4869 group decreased. However, there was no significant difference in TNF-α and IL-6 protein levels between the two groups. These results suggest that exosomes are involved in the mediating process of BMSC inflammatory stimulation. The effects of ordinary exosomes, inflammatory exosomes, and CM on BMSC osteogenic differentiation are summarized in Fig. 3. Compared with ordinary exosomes, inflammatory exosomes had a reduced expression of ALP, RUNX2, and collagen type I (COL-1) at both the gene and protein levels. Additionally, the expression levels of osteogenic genes ALP, RUNX2, and COL-1 were increased in BMSCs compared with CM after GW4869 treatment ( Figs. 3A – 3G ). The alkaline phosphatase activity test and staining results were consistent with BMSC gene expression ( Figs. 3H – 3K ). 10. 7717/peerj. 13442/fig-3 Figure 3 Effects of inflammatory macrophage-derived exosomes on osteogenic indices and ALP activity of BMSCs. (A) ALP mRNA expression, (B) COL-1 mRNA expression, (C) RUNX2 mRNA expression in BMSCs after transfection of ordinary exosomes, inflammatory exosomes, CM, and CM+GW4869 for 48 h. (D–G). The protein expression after BMSC transfection with ordinary exosomes, inflammatory exosomes, CM, and CM+GW4869 for 7 days. (H) BCA quantification of ALP activity. (I) ALP activity. (J) ALP staining images. (K) Quantitative mineralization assay of panel J. * indicates P < 0. 05, two asterisk (**) indicate P < 0. 01, three asterisks (***) indicate P < 0. 001, four asterisks (****) indicate P < 0. 0001, ns indicates no significant difference. Discussion Periodontitis disrupts the balance between bone formation and bone resorption in the alveolar bone, which increases bone loss in the alveolar bone ( Xu et al. , 2018 ). As an essential part of innate tissue immunity, the number of macrophages in the periodontal tissues of patients with chronic periodontitis is much higher than that in normal periodontal tissues. As the progenitors of osteoclasts, macrophages can induce bone resorption by secreting proinflammatory cytokines, and excessive cytokines can inhibit the differentiation, proliferation, and mineralization of osteoblasts. The degradation of the bone matrix provides attachment sites for osteoclasts, leading to further loss of alveolar bone tissue ( He et al. , 2018 ). Macrophages play a significant role in alveolar bone loss caused by chronic inflammation ( Lam et al. , 2016 ). Macrophages influence and regulate the reconstruction and loss of alveolar bone through paracrine signalling, but intercellular communication depends on not only cytokines and other extracellular vesicles but also exosomes. In this study, BMSC uptake of macrophage-derived exosomes was successfully observed through staining and confocal microscopy. As carriers of transmitted information, exosomes can effectively protect their contents from the influence of the extracellular environment in order to achieve directional delivery to recipient cells ( Fais et al. , 2013 ). Studies have shown that exosomes are involved in many inflammatory processes, such as acute lung injury, inflammatory bowel disease, and asthma inflammation. ( Canas et al. , 2021 ; Monsel et al. , 2016 ; Zhang et al. , 2019 ). We therefore hypothesized that exosomes play a role in the crosstalk between macrophages and BMSCs under inflammatory stimulation. The composition of macrophage exosome proteins is mainly divided into two categories. One category includes the standard proteins that are ubiquitous in the formation and secretion of vesicles. These proteins include the transmembrane transport proteins, fusion-related proteins (such as Rab and GTPases), heat shock proteins (such as HSP70 and HSP90), tetrapeptide transmembrane proteins (such as CD63, CD81, and CD9), and ESCRT complex-related proteins (such as Tsg101 and Alix). Among these proteins, CD63, CD81, CD9, and TSG101 are highly enriched in exosomes and have become commonly used marker proteins of exosomes ( Pegtel & Gould, 2019 ). The other category includes specific components closely related to macrophages. Compared with synthetic vectors such as liposomes and nanoparticles, exosomes have extensive and unique advantages in disease diagnosis and treatment due to their endogeneity and heterogeneity. However, the impurity and low yield of exosomes limit their clinical application as well as their application in scientific research. The commonly used techniques for exosome separation include ultracentrifugation, ultrafiltration, polymer precipitation, and immunoaffinity chromatography with different techniques being used for different purposes and applications. The most widely used separation technology is supercentrifugation, which is also currently considered the gold standard for exosome extraction and separation ( Zhang et al. , 2020 ). In this study, ultrafast centrifugation was used to extract macrophage-derived exosomes, and the marker proteins CD9, CD81, and TSG101 were identified. The International Society for Extracellular Vesicles (ISEV) pointed out that the identification of two characteristic proteins can prove that the extracted substances are exosomes ( Zhang et al. , 2020 ), and the reason why CD81 is not expressed in our extracted exosomes is that, although the positive indicator is the exosome marker protein, which is widely reported in many studies, some indicators may have no obvious bands due to low protein concentration or antibody immune typing. Therefore, our experimental results can also prove that we successfully extracted exosomes derived from macrophages, but our low extraction concentration and large sample size were still limitations of this study. The transmission electron microscopy results ( Figs. 1E – 1F ) showed that the average particle size of macrophage-derived exosomes extracted in this study was 81. 20 nm. Compared to nanomaterials of various sizes, exosomes of this size can effectively cross barriers (such as the plasma membrane and the blood/brain barrier), with lower metabolic efficiency and a longer-acting time. At the same time, exosomes have great potential to become drug delivery vectors suitable for delivering various small molecule drugs, proteins, nucleic acids, and gene therapy agents due to their natural material transport properties, inherent long-term recycling capacity, and excellent biocompatibility ( Cho et al. , 2018 ; Ohno et al. , 2013 ; Pascucci et al. , 2014 ). Based on confocal microscopy results, we observed an increased fluorescence intensity after adding macrophage-derived exosomes compared to the control group, indicating the successful internalization of exosomes and distribution of cells in the cytoplasm. The main internalization pathway of exosomes is through clathrin-independent endocytosis and pinocytosis ( Costa Verdera et al. , 2017 ), which involve endosomal acidification and membrane fusion between exosomes and target cells ( Bonsergent et al. , 2021 ). The efficiency of exosome uptake depends on cholesterol and tyrosine kinase activity as well as PH, temperature, and freezing and thawing times during the experiment ( Cheng et al. , 2019 ). The exosomes extracted in this study were stored at −80 °C, and follow-up experiments were conducted within 10 days to ensure the activity of exosomes and the uptake efficiency of target cells. Gram-negative anaerobic bacteria, especially Porphyromonas gingivalis, are the main pathogenic bacteria of periodontitis, and their main pathogenic component, lipopolysaccharide, can induce bone destruction by triggering the release of proinflammatory mediators such as IL-6 and TNF-α. Stimulated by LPS, macrophages contribute to the development of inflammation by secreting many biologically active molecules, including proinflammatory cytokines and proteolytic enzymes. Additionally, studies have pointed out that there are significant differences in the contents of exosomes released by LPS-activated macrophages, indicating that exosomes released under different conditions may be involved in various functional and pathological processes ( Raeven, Zipperle & Drechsler, 2018 ). LPS exosomes have been proven to contain a large number of proinflammatory factors. Among the protein components transmitted by exosomes, TNF, CCL3, CD40, and Serpine1 are closely related to inflammatory responses, and LPS exosomes are involved in the activation of various inflammatory signalling pathways. These pathways include the Nod-like receptor, IL-17, TNF, and Toll-like receptor signalling pathways ( Wang et al. , 2019 ). Previous studies have shown that 100 ng/mL LPS is sufficient to induce inflammatory responses in macrophages ( Bode, Ehlting & Haussinger, 2012 ), which was consistent with our experimental results. After LPS treatment, the gene and protein expression levels of IL-6 and TNF-α in macrophages were increased, and this inflammatory state persisted after the removal of the stimulus. Additionally, our study also confirmed that inflammatory exosomes are involved in the inflammatory mediations of macrophages to BMSCs, which is consistent with previous studies ( Wang et al. , 2017 ). Recent studies have found that bone formation, resorption, and remodelling are inseparable from the regulation of the immune system, and the communication between macrophages and other osteocytes plays an essential role in bone tissue homeostasis and new bone formation. Previous studies believed that macrophage products such as TNF-α, IL-6, IL-1β, etc. , were the main immune factors involved in bone destruction and that they reduced the osteogenic differentiation potential of stem cells. At the same time, the secretion and delivery of exosomes was thought to be a wasteful mechanism, but there is growing evidence that exosomes are ubiquitous mediators of cellular communication in all cell types. In this study, we proved that without the participation of the above-mentioned inflammatory factors, exosomes also greatly affected the osteogenic differentiation of BMSCs, which is consistent with the research reported by Zhu et al. (2019). Still, some studies have concluded that in the early stages of inflammation, M1-type macrophage exosomes were able to promote BMSC osteogenic differentiation, which may be due to the lower concentration of exosomes used in this study. Interestingly, we found that the expression levels of inflammatory markers IL-6 and TNF-α decreased in the CM group compared with the ordinary exosome group, while the expression levels of the osteogenic markers ALP, RUNX2, and COL-1 increased. Therefore, we speculated that macrophage-derived exosomes mainly mediate inflammation and inhibit osteogenic differentiation during the crosstalk between macrophages activated by lipopolysaccharide and BMSCs. However, other substances that exist in the macrophage extracellular environment that can inhibit inflammation and promote bone differentiation were not found in this study. Other studies have confirmed that all macrophage subtypes, including M0, M1, and M2 macrophages, can promote the osteogenic differentiation of BMSCs, and M1 macrophages have the most significant influence on bone formation ( Chen et al. , 2020a ), which is consistent with our observations in this study. Cells in the macrophage extracellular environment and their factor-specific roles require further research, which will be the goal of our future studies. In this study, our experimental results clarify that macrophage-derived exosomes are involved in inhibiting the osteogenic differentiation of stem cells in inflammatory states, and that the use of exosome inhibitors can attenuate this effect. But as we mentioned before, there are multiple ways of communication and crosstalk between cells, and whether blocking exosome secretion can be used to restore the osteogenic function of stem cells in the inflammatory state needs further verification. In addition, based on the similar size of exosomes and nanomaterials and the advantages exosomes have of lower immunogenicity, higher biocompatibility, natural targeting ability, and good biological barrier permeability, their carrier functions are used for more complex scenarios than nanomaterials. However, it is necessary to understand the specific mechanism of exosomes for intercellular communication under physiological and pathological conditions. In follow-up experiments, we will further improve the specific mechanism of the exosome pathway to inhibit the osteogenic differentiation of stem cells, such as the main components and mode of action of inflammatory exosomes to support this study. Conclusion In conclusion, our results suggest that P. g -LPS-activated macrophage-derived exosomes can inhibit BMSC osteogenic differentiation by mediating inflammatory stimulation. At present, our study is expected to provide a potential treatment plan to improve the treatment effect in the application scenario of in vitro stem cell therapy and to improve the reconstruction of alveolar bone in periodontitis by regulating exosomes in the periodontal microenvironment. Supplemental Information 10. 7717/peerj. 13442/supp-1 Supplemental Information 1 Raw data for Figures 1A, 1B Click here for additional data file. 10. 7717/peerj. 13442/supp-2 Supplemental Information 2 Raw data for Figures 1C, 1D Click here for additional data file. 10. 7717/peerj. 13442/supp-3 Supplemental Information 3 Raw data for Figure 1G Click here for additional data file. 10. 7717/peerj. 13442/supp-4 Supplemental Information 4 Raw data for the BCA quantification Click here for additional data file. 10. 7717/peerj. 13442/supp-5 Supplemental Information 5 Raw data for Figures 2E, 2F, 3A, 3B, 3C Click here for additional data file. 10. 7717/peerj. 13442/supp-6 Supplemental Information 6 Raw data for Figures 3H, 3I Click here for additional data file. 10. 7717/peerj. 13442/supp-7 Supplemental Information 7 Images of western blots Click here for additional data file.

10. 7717/peerj. 13475

PeerJ

Long noncoding RNA

Background Long noncoding RNA Gm31629 can regulate hypothalamic neural stem cells (htNSCs) senescence and the aging process. However, the effect of Gm31629 on the senescence of bone marrow mesenchymal stem cells (BMSCs) and bone regeneration is unclear. In the present study, we investigated the effects of Gm31629 on the senescence of BMSCs and bone regeneration. Methods Gm31629 knockout ( Gm31629 -KO) and wild-type (WT) mice were used to establish a bone regeneration model. The Brdu labelling, CCK8 assay, wound healing assay, β-gal staining and osteogenic differentiation assay were used to assess the effects of Gm31629 on the functions of BMSCs. Micro-computed tomography (CT), histochemical and immunohistochemical staining were used to evaluate the ability of bone regeneration. The mimic of Gm31629, theaflavin 3-gallate, was used to investigate its role on the senescence of BMSCs and bone regeneration. Results The expression of Gm31629 reduced in BMSCs of middle-aged mice was compared with that of young mice. The deletion of Gm31629 was sufficient to drive the senescence of BMSCs, resulting in impaired bone regeneration in mice. Mechanistically, Gm31629 could interact with Y-box protein 1(YB-1) and delay its degradation, decreasing the transcription of p16 INK4A of BMSCs. We also found that theaflavin 3-gallate could alleviate the senescence of BMSCs and promote bone regeneration in middle-aged mice. Conclusion These results indicated that Gm31629 played an important role on BMSCs senescence and bone regeneration and provided a therapeutic target to promote bone regeneration.

Introduction Bones, the primary structural material of mammals, are often damaged throughout life and undergo constant modeling, remodeling and repair ( Borrelli Jr et al. , 2012 ; Taylor, Hazenberg & Lee, 2007 ; Zheng et al. , 2019 ). Bone is a powerful self-healing tissue, but the ability to self-heal in the elderly can be reduced by complex changes at the molecular, cellular, and systemic levels ( Gruber et al. , 2006 ). Bone repair is a complex biological process involving the synergistic participation of vascular and skeletal precursor cells within the bone marrow ( Dimitriou, Tsiridis & Giannoudis, 2005 ). Bone marrow mesenchymal stem cells (BMSCs), which can self-renew and differentiate into multiple cell types, make great contributions to the regeneration of mesenchymal tissues such as cartilage, adipose and bone ( Lu, Li & Cheng, 2002 ; Pittenger et al. , 1999 ). Moreover, it has been suggested that BMSCs can act as potent microenvironmental regulators, which exert influence on various tissues, including bone ( Liu et al. , 2015 ; Su et al. , 2019 ; Sudres et al. , 2006 ; Xiao et al. , 2021 ; Xu et al. , 2018b ; Yu et al. , 2021 ). For example, Yu et al. (2021) reported that BMSCs-derived exosomal miR-136-5p promoted osteoblast proliferation, differentiation, thus facilitating fracture healing. Accordingly, BMSCs have been widely used in bone regeneration including bone tissue engineering for their close involvement in bone formation ( Fernandes & Yang, 2016 ; Lu, Li & Cheng, 2002 ; Zhang et al. , 2019 ). With age, various cell types, including BMSCs, undergo senescence ( Aguayo-Mazzucato et al. , 2019 ; Wang et al. , 2020 ; Wiley et al. , 2021 ). Senescent BMSCs not only showed decreased ability to differentiate to osteoblasts, but also showed a declining capacity for proliferation and migration ( Geissler et al. , 2012 ; Li et al. , 2015 ; Moerman et al. , 2004 ; Sethe, Scutt & Stolzing, 2006 ). Moreover, exosomal miR-31a-5p secreted by senescent BMSCs can not only inhibit osteogenic differentiation, but also promote osteoclast differentiation ( Xu et al. , 2018b ). All of these may lead to impaired therapeutic effects of senescent BMSCs in bone regeneration. However, the exact mechanism of BMSCs senescence remains unclear. Long non-coding RNAs (LncRNAs), which are characterized by transcripts more than 200 nucleotides in length, play a variety of regulatory roles through interactions with DNA, RNA and proteins ( Huo et al. , 2018 ). They have been observed to participate in the regulation of many biological processes and diseases, involving cell senescence, apoptosis, differentiation, proliferation, and tumorigenesis ( Guo et al. , 2020 ; Klattenhoff et al. , 2013 ; Ng, Johnson & Stanton, 2012 ; Xiao et al. , 2020 ; Yang et al. , 2011 ). Recently, several studies have revealed that lncRNAs are involved in regulating osteogenic differentiation of BMSCs and bone repair ( Liu et al. , 2022 ; Ouyang et al. , 2020 ). Our previous study showed that lncRNA, Gm31629 is down-regulated in the hypothalamic neural stem cell (htNSCs) of middle-aged mice compared with that of young mice ( Xiao et al. , 2020 ). Deletion of Gm31629 accelerated the senescence of htNSCs and leaded to aging-associated phenotype in mice ( Xiao et al. , 2020 ). Gm31629 could regulate the senescence of htNSCs by delaying the degradation of YB-1 ( Xiao et al. , 2020 ). YB-1 is a DNA/RNA-binding protein ( Lyabin, Eliseeva & Ovchinnikov, 2014 ) and has been reported to bind to the promoter region of p16 INK 4 A and inhibit its expression ( Kotake et al. , 2013 ; Xiao et al. , 2020 ), a maker of cellular senescence ( Ogrodnik et al. , 2019 ; Omori et al. , 2020 ). However, the role of Gm31629 in the senescence of BMSCs and bone regeneration has not been investigated. In the present study, we expanded our research and demonstrated that Gm31629 could also regulate the senescence of BMSCs and bone regeneration. Deletion of Gm31629 accelerated the degradation of YB-1, promoted the senescence of BMSCs, and impaired the ability of bone regeneration. We also found that the natural compound, theaflavin 3-gallate (TF2A), could mimic the activity of Gm31629 and alleviate the senescence of BMSCs. Treatment of TF2A could promote bone regeneration in middle-aged mice. Materials and Methods Animals and treaments Gm31629 knockout ( Gm31629-KO ) mice were obtained from Cyagen Biosciences as previously reported ( Xiao et al. , 2020 ). We purchased 3-month-old and 12-month-old C57BL/6J male mice from the Laboratory Animal Center of Central South University (Changsha, China). The model of bone regeneration was established as described before ( Chen et al. , 2019 ; Fukuda et al. , 2013 ; Yang et al. , 2020 ). Briefly, after anesthesia, the anterior medial approach was used to expose the distal femoral. Then, a 25-guage needle was used to drill a hole at the distal femur along the long axis of the femur and a 0. 6 mm diameter Kirschner wire was used to ablate trabecular bone of distal femur. This was minimally invasive injury and we made great effort to reduce the sufferings of the mice. One week after the injury, the mice were euthanized by cervical dislocation after anesthesia to collect the bone samples. For TF2A administration, mice were treated with TF2A or vehicle by gavage at a dosage of 8 mg/kg every day for three weeks before the establishment of bone regeneration model. After that, TF2A treatment continued for one week before the mice were euthanized. All the mice in this study were healthy and C57BL/6 background and kept in the Experimental Animal Research Center of Central South University with specific pathogen-free standard. The mice were housed in individual ventilated cage with six mice per cage. The mice were kept in room temperature with 12 h light-dark cycle and had free access to food and water. No animal was excluded from the experiments. Xiangya Hospital of Central South University Ethics Committee (Changsha, Hunan, China) approved this research (2019030350). All animal experiments conformed to all ethical requirements relating to animal research. BMSCs isolation, culture and senescence assays BMSCs were isolated as previously described ( Li et al. , 2015 ; Yu et al. , 2018 ). The isolated BMSCs were cultured with α-MEM supplemented with 15% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin in a humidified atmosphere of 5% CO2 at 37 °C to reach 80% confluence. Then the first-passage BMSCs were harvested and seeded in culture dishes for enrichment of cell populations. When the second-passage reach confluence after 1–2 week, they were subcultured. Only third-passage BMSCs were applied to perform further study unless specified otherwise. The senescent BMSCs were stained by a senescence β-galactosidase staining Kit (Solarbio Science & Technology) according to the manufacturer’s instructions. Briefly, after washing with PBS, the cells were fixed with 4% paraformaldehyde for 15 min at room temperature. Then the cells were stained with working solution overnight at 37 °C. Five different fields were randomly selected under a microscope to count the SA-βGal-positive (blue cells) and the percentage of SA-βGal-positive were calculated. Cell transfection For Gm31629 overexpression, the adenovirus particles expressing mouse Gm31629 were purchased from OBiO Technology Corporation (Shanghai, China). For Yb-1 overexpression, pcDNA3. 1-mYb-1 was purchased from Sino Biological Inc (Beijing, China). PcDNA3. 1-mYb-1 and negative control were transfected into BMCSs with lipofectamine 2000 (Invitrogen, Thermo Fisher Scientific, Waltham, MA) by a standard method. Wound healing assay BMSCs were seeded in 6-well plates at a density of 1 ×10 6 cells per well for each group. A linear wound was made using a sterile 200 µl pipette tip to scratch across the confluent cell layer. Images of wound healing were observed at 0 h and 24 h and the migration rate was calculated using ImageJ software (National Institutes of Health, Bethesda, MD, USA). CCK8 assay CCK8 assay was used to evaluate the growth of BMSCs according to the manufacture’s protocol (MedChemExpress, LLC). BMSCs were seeded in 96-well plates at a density of 5, 000 cells per well for each group. Then, we added 10 µl of CCK-8 solution into each well and incubated the plate at 37 °C for 2 h. At last, the OD value of each well was examined at 450 nm using a spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Brdu staining assay Brdu staining assay was conducted using standard methods. Briefly, BMSCs were seeded in 24-well plates at a density of 1 × 10 5 cells per well for each group and incubated with 10 µM Brdu labeling solution for 24 h in a cell incubator. After that, the cells were fixed with 4% paraformaldehyde for 20 min. After permeabilization using 0. 2% triton, the cells were incubated with 3% BSA for blocking. Then, the cells were incubated with the primary (Cell Signaling Technology, Danvers, MA; 5292, 1:400) and secondary antibody (Invitrogen, Thermo Fisher Scientific, Waltham, MA; A21202, 1:500), and counterstained DAPI. Osteogenic differentiation assay BMSCs were seeded in 6-well plates at a density of 5 × 10 5 cells per well for each group and cultured with osteogenic induction conditional medium (10 mM β-glycerol phosphate, 50 µM ascorbate-2-phosphate, and 0. 1 µM dexamethasone) for three weeks. We changed the osteogenic medium every other day. To assess the mineralization of cell matrix, 2% Alizarin Red S (Cyagen Biosciences, Santa Clara, CA) was used to stain the cell matrix. Alizarin Red S was destained with cetyl-pyridinium chloride solution and the OD value was quantified by spectrophotometry at 562 nm. Osteoclast differentiation assay We perform osteoclasts differentiation assay as described before ( Yang et al. , 2020 ). Briefly, bone marrow was flushed out of bone marrow cavity of mice. Isolated bone marrow cells were cultured with complete media for 14 h. Then the unattached cells were collected and treated with α-MEM containing 10% FBS, 30 ng/mL M-CSF (R&D Systems Inc. , Minneapolis, MN), 100 µg/mL streptomycin, and 100 U/mL penicillin for 72 h to gain pure monocytes and macrophages. After that, the monocytes and macrophages were cultured with osteoclastic induction medium (30 ng/mL M-CSF, 60 ng/mL RANKL) for 1 week. Osteoclasts were stained with TRAP staining kit (Sigma-Aldrich, St Louis, MO) according to manufacturer’s instructions. RT-qPCR analysis Extraction of total RNA was performed with Trizol (Invitrogen, Thermo Fisher Scientific, Waltham, MA) following standard methods and reverse transcription was conducted using 1 µg total RNA. RT-qPCR was conducted in duplicate using SYBR Premix Ex Taq II (Takara). We normalized the Ct value of Gm31629 to that of Gapdh and calculated ΔCt value (ΔCt=Ct ( Gm 31629) - Ct ( Gapdh ) ) for both 3-moth-old and 12-moth-old group. Then we normalized the ΔCt value of 12-moth-old group to the ΔCt value of 3-moth-old group and calculated ΔΔCt value (ΔΔCt = Δ Ct (12-moth-old) −Δ Ct (3-moth-old) ). The relative gene expression was calculated using the 2 −ΔΔCT method. All experiments were repeated three times. The primer sequences are listed in Table S1. Western blot Western blotting was performed as previously described ( Liu et al. , 2021b ; Peng et al. , 2019 ). Total cell proteins were separated by SDS-PAGE and blotted on PVDF membranes (Millipore, Sigma, Burlington, MA). After blocking with 5% milk, the membranes were incubated with specific antibodies to YB-1 (Cell Signaling Technology, Danvers, MA; 4202, 1:1000), p16 INK4A (Sigma-Aldrich, St Louis, MO, SAB4500072, 1:1000) and GAPDH (Proteintech, Rosemont, IL, USA; 10494-1-AP, 1:5000). Blots were visualized using an ECL Kit (Thermo Fisher Scientific, Waltham, MA; 32, 106). RNA pull-down assay and RNA immunoprecipitation We performed RNA pull down assay as previously reported ( Xiao et al. , 2020 ). Briefly, biotin-labeled full-length Gm31629 and antisense Gm31629 were incubated with nuclear lysate of BMSCs for 1 h. After that, the streptavidin agarose beads (Invitrogen, Thermo Fisher Scientific, Waltham, MA) were added and incubated at 25 °C for another 1 h. After washing with cold NT2 buffer, the pulled-down proteins were used for western blot analysis. RNA immunoprecipitation was performed using a Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, Sigma, Burlington, MA; 17-700) following the manufacturer’s instructions. The precipitated RNA was extracted, reversed transcribed and analyzed by RT-qPCR. All experiments were repeated three times. The primer sequences are listed in Table S1. Chromatin immunoprecipitation (ChIP) assay ChIP assay was performed with SimpleChip Kit (9003; Cell Signaling Technology, Danvers, MA) following the manufacturer’s instructions as previously described ( Yang et al. , 2019 ). Briefly, chromatin was crossed-linked (1 % formaldehyde, 10 min) and sheared to 100- to 500-bp fragments by sonication. The relevant protein-DNA complex was immunoprecipitated by YB-1 antibody (Santa Cruz Biotechnology, Dallas, TX; sc-398146) or IgG control. The ChIP DNA was used to perform standard PCR or RT-qPCR. All experiments were repeated three times. The primer sequences are listed in Table S1. µCT analysis µCT scanning was performed using a high-resolution micro-CT (SCANCO Medical AG, VIVACT 80; Wangen-Brüttisellen, Switzerland) with a resolution of 12 µm per pixel at 55 kV and 145 µA. We reconstructed a 3D model and analyzed the structure indices as previous reported ( Li et al. , 2021 ; Yang et al. , 2019 ; Yang et al. , 2020 ). Trabecular bone volume (Tb. BV/TV) in the regeneration region was calculated. Histochemistry and immunohistochemistry Histochemical and immunohistochemical staining were conducted as previously described ( Cai et al. , 2022 ; Yang et al. , 2020 ). Femora were collected and fixed with 4% paraformaldehyde for 1 day at 4 °C. Then we decalcified the bones with 10% EDTA and embedded them in paraffin. For histochemistry, 4 micrometer-thick slides were subjected to HE and TRAP staining according to a standard protocol. For immunocytochemistry, after antigen retrieval the samples were incubated with primary antibodies against osteocalcin (Takara M173) at 4 °C overnight and Horseradish peroxidase-streptavidin detection system (Dako Agilent, Santa Clara, CA) was used to detect immuno-activity. Statistics analysis Statistical analysis was performed using GraphPad Prism software 8. 0. Data are expressed as the mean ± standard deviation (sd). Unpaired Student’s t test was applied to compare two groups. One-way ANOVA was employed while comparing multiple groups. The difference was considered to be statistically significant at p < 0. 05. In order to avoid the type II error, we used G*Power 3. 1 to perform the statistical power analysis and the minimum power required in this study was set at 0. 8. All the samples were randomly assigned and no blinding was used. Results BMSCs undergo senescence during aging with reduced ability of bone regeneration To study the characteristics of senescent BMSCs, we compared the function and phenotype of BMSCs from middle-aged (12-month old) mice with that from young mice (3-month old) in vitro. The Brdu staining assay and CCK8 assay revealed that the proliferation ability of BMSCs from middle-aged mice was significantly reduced in comparison with BMSCs from young mice ( Figs. 1A – 1C ). The wound healing assay indicated that the migration ability of BMSCs from middle-aged mice was markedly declined in comparison with BMSCs from young mice ( Figs. 1D, 1E ). As expected, there were more SA-βGal-positive BMSCs in the middle-aged group than in the young groups ( Figs. 1F, 1G ). BMSCs from middle-aged mice also showed decreased osteogenic differentiation ability compared to that of young mice ( Figs. 1H, 1I ). These results indicated that an aging phenotype of BMSCs presented in the middle-aged mice, resulting in significant impaired function of BMSCs. 10. 7717/peerj. 13475/fig-1 Figure 1 The increased cell senescence in BMSCs of middle-aged mice. (A) Representative images of Brdu assay. Scale bar:100 µm. (B) Quantification of Brdu positive cells. ( n = 3). (C) CCK8 assay ( n = 3). (D) Representative images of BMSCs migration in wound healing test. Scale bar:200 µm. (E) Quantitative analysis of migration rate. ( n = 3). (F) SA-βGal staining of BMSCs. Scale bar: 50 µm. (G) The percentage of SA-βGal positive cells. ( n = 3). (H) ARS staining of BMSCs under osteogenic induction. Scale bar: 100 µm. (I) Quantification of calcium mineralization ( n = 3). Data are expressed as mean ± sd and statistical differences were analyzed by Student’s t test. ∗ p < 0. 05; ∗∗ p < 0. 01; ∗∗∗ p < 0. 001. To investigate the change of bone regeneration ability during aging, a bone regeneration model was established by surgical ablation of trabecular bone in distal femur. As expected, the bone volume in the regeneration area of middle-aged mice was lower than that of young mice at 7 days after ablation ( Figs. 2A – 2D ). The number of osteocalcin positive (ocn + ) osteoblasts in the bone regeneration area of middle-aged mice was less than that of young mice at 7 days after ablation ( Figs. 2E, 2F ). The number of tartrate-resistant acid phosphatase positive (TRAP + ) osteoclasts in the bone regeneration area of middle-aged mice was also less than that of young mice at 7 days after ablation ( Figs. 2G, 2H ). Altogether, these results indicated that middle-aged mice had reduced bone regeneration ability compared to young mice. 10. 7717/peerj. 13475/fig-2 Figure 2 The ability of bone regeneration decreases during aging. (A) Time plan for surgical ablation of trabecular bone in distal femoral of mice. (B) Representative micro-CT images. The white square was selected to measure trabecular bone volume in bone regeneration region. (C) Quantification of trabecular bone volume in bone regeneration region. ( n = 6). (D) HE staining of distal femora. Scale bar: 200 µm. (E) Immunohistochemical staining of osteocalcin positive cells. Black arrows represent osteocalcin positive cells. Scale bar: 50 µm. (F) Quantitative analysis of osteocalcin positive cells. ( n = 6). (G) TRAP staining images. Scale bar: 50 µm. (H) Quantitative analysis of TRAP positive cells. ( n = 6). Data are expressed as mean ± sd and statistical differences were analyzed by Student’s t test. ∗∗ p < 0. 01; ∗∗∗ p < 0. 001. Gm31629 regulates the senescence of BMSCs Gm31629 expression decreased significantly in BMSCs of middle-aged mice compared with that of young mice as analyzed by RT-qPCR ( Fig. 3A ). To study the function of Gm31629 in the regulation of BMSCs senescence, we isolated BMSCs from 3-month-old Gm31629 knockout ( Gm31629-KO ) mice and wild type (WT) mice. The proliferation and migration ability of BMSCs isolated from Gm31629-KO mice were markedly reduced in comparison with that of BMSCs isolated from WT mice ( Figs. 3B – 3F ). In addition, there were more SA-βGal-positive BMSCs in Gm31629-KO group than in WT group ( Figs. 3G, 3H ). BMSCs of Gm31629-KO mice also showed decreased osteogenic differentiation capacity compared to that of WT group ( Figs. 3I, 3J ). These data revealed that BMSCs from Gm31629-KO mice showed an aging phenotype with significant impaired function of BMSCs. 10. 7717/peerj. 13475/fig-3 Figure 3 Gm31629 regulates the senescence of BMSCs. (A) Relative Gm31629 expression in BMSCs of 3-month-old and 12-month-old mice as analyzed by RT-qPCR. The expression level of Gm32129 in BMSCs of 3-month-old mice was set at an arbitrary value = 1. ( n = 3). (B) Representative images of Brdu assay. Scale bar: 100 µm. (C) Quantification of Brdu positive cells. ( n = 3). (D) CCK8 assay. ( n = 3). (E) Representative images of BMSCs migration in wound healing test. Scale bar: 200µm. (F) Quantitative analysis of migration rate. ( n = 3). (G) SA-βGal staining of BMSCs. Scale bar: 50 µm. (H) The percentage of SA-βGal positive cells. ( n = 3). (I) ARS staining of BMSCs under osteogenic induction. Scale bar: 100 µm. (J) Quantification of calcium mineralization. ( n = 3). Data are expressed as mean ±sd and statistical differences were analyzed by Student’s t test. ∗ P < 0. 05; ∗∗ P < 0. 01; ∗∗∗ P < 0. 001; N. S, no significance. Gm31629 knockout mice show impaired bone regeneration ability To further study the function of Gm31629 in bone regeneration, we established the bone regeneration model in Gm31629-KO mice and WT mice at 3-month-old by surgical ablation of trabecular bone in distal femur ( Fig. 4A ). We found that the bone volume in bone regeneration region of Gm31629-KO mice was significantly lower than that of WT controls at 7 days after ablation ( Figs. 4B – 4D ). The number of ocn + osteoblasts was markedly reduced in bone regeneration region of Gm31629-KO mice in comparison with that of WT controls at 7 days after ablation ( Figs. 4E, 4F ). There is no significant difference of TRAP + osteoclasts in bone regeneration region of Gm31629-KO mice in comparison with that of WT controls 7 days after ablation ( Figs. 4G, 4H ). These data showed that the bone regeneration ability of Gm31629-KO mice was lower than that of WT mice. 10. 7717/peerj. 13475/fig-4 Figure 4 Gm31629-KO mice show impaired ability of bone regeneration. (A) Time plan for surgical ablation of trabecular bone in distal femoral of mice. (B) Representative micro-CT images. The white square was selected to measure trabecular bone volume in bone regeneration region. (C) Quantification of trabecular bone volume in bone regeneration region. ( n = 6). (D) HE staining of distal femora. Scale bar: 200 µm. (E) Immunohistochemical staining of osteocalcin positive cells. Black arrows represent osteocalcin positive cells. Scale bar: 50 µm. (F) Quantitative analysis of osteocalcin positive cells. ( n = 6). (G) TRAP staining images. Scale bar: 50 µm. (H) Quantitative analysis of TRAP positive cells. ( n = 6). Data are expressed as mean ± sd and statistical differences were analyzed by Student’s t test. ∗ P < 0. 05; ∗∗ P < 0. 01; N. S, no significance. Gm31629 regulates BMSCs senescence through YB-1/P16 INK 4 A pathway Our previous study demonstrated that Gm31629 directly interacted with YB-1 and increased the protein level of YB-1 by preventing its degradation, further reducing the expression of p16 INK 4 A and suppressing the senescence of htNSCs ( Xiao et al. , 2020 ). YB-1 is a DNA/RNA-binding protein ( Lyabin, Eliseeva & Ovchinnikov, 2014 ), and has been reported to bind to the promoter region of p16 INK 4 A and inhibit the expression of p16 INK 4 A ( Kotake et al. , 2013 ; Xiao et al. , 2020 ), a maker of cellular senescence ( Ogrodnik et al. , 2019 ; Omori et al. , 2020 ). Furthermore, several studies have demonstrated that the expression of p16 INK 4 A is much higher in BMCSs of older mice than in young controls ( Hu et al. , 2022 ; Li et al. , 2017 ). To verify the interaction between YB-1 and Gm31629 in BMSCs, the RNA pull-down assay was repeated and the binding of Gm31629 to YB-1 in BMSCs was confirmed ( Fig. 5A ). RNA immunoprecipitation assay further confirmed that Gm31629 could bind to YB-1 in BMSCs ( Fig. 5B ). In addition, Gm31629 knockout markedly reduced YB-1 protein level and Gm31629 overexpression significantly increased YB-1 protein level ( Figs. 5C, 5D ). To confirm that Gm31629 increases YB-1 protein level by preventing the degradation of YB-1 in BMSCs, we inhibited protein synthesis in BMSCs with cycloheximide (CHX) and found that Gm31629 knockout accelerated the degradation of YB-1 ( Fig. 5E ). These results indicated that Gm31629 could prevent the degradation of YB-1 in BMSCs. Moreover, ChIP-PCR assays showed that YB-1 could directly bind to the promoter of p16 INK 4 A in BMSCs ( Fig. 5F, 5G ). Then, we observed that Gm31629 knockout not only reduced YB-1 protein level, but also increased the expression of p16 INK 4 A ( Fig. 5H ). However, the overexpression of YB-1 in Gm31629 knockout BMSCs rescued the reduced YB1 protein level and reduced the expression of p16 INK 4 A ( Fig. 5H ). Moreover, the overexpression of YB-1 rescued the increased level of senescence in BMSCs derived from Gm31629- KO mice ( Figs. 5I, 5J ). In addition, osteogenic differentiation assay revealed that YB-1 overexpression rescued the reduced osteogenic differentiation of BMSCs derived from Gm31629- KO mice ( Figs. 5K, 5L ). Thus, these results indicated Gm31629 could stabilized YB1 protein and inhibited the expression of p16 INK 4 A, a possible mechanism for Gm31629 in regulating the senescence of BMSCs. 10. 7717/peerj. 13475/fig-5 Figure 5 Gm31629 regulates BMSCs senescence through YB-1/P16 INK4A pathway. (A) Western blot analysis of YB-1 pulled-down by Gm31629 (1-1301S) and antisense Gm31629 (1-1301AS) or other controls. (B) YB-1-retrieved Gm31629 RNA as determined by RT-qPCR analysis. The level of IgG-retrieved Gm31629 was set at an arbitrary value = 1. (C) Western blotting analysis of YB-1 protein in WT and Gm31629-KO BMSCs. (D) Western blotting analysis of YB-1 protein in adenovirus vector-driven Gm31629 overexpressed or control BMSCs. (E) Western blotting analysis of YB-1 protein in WT and Gm31629-KO BMSCs treated with CHX. (F) The binding of YB1 to the p16 INK4A promoter was detected by ChIP -PCR assay with an antibody against YB1 or IgG. (G) The abundance of YB-1 binding on the promoter of p16 INK4A was determined by ChIP assay followed by RT-qPCR analysis. (H) Western blotting result of YB-1 and P16 INK4A protein in WT and Gm31629-KO BMSCs with or without YB-1 overexpressed. (I) Representative images of SA-βGal staining of BMSCs. Scale bar:50 µm. (J) The percentage of SA-βGal positive cells. ( n = 3). (K) ARS staining of BMSCs under osteogenic induction. Scale bar:100 µm. (L) Quantification of calcium mineralization. ( n = 3). Data are expressed as mean ± sd and statistical differences were analyzed by Student’s t test or one-way ANOVA. ∗∗ P < 0. 01; ∗∗∗ P < 0. 001. TF2A treatment in vitro alleviates the senescence of BMSCs In the previous study, we also identified a natural compound, TF2A, which mimics the ability of Gm31629 to increased YB-1 protein level, reducing the senescence of htNSCs ( Xiao et al. , 2020 ). As expected, treatment of TF2A could mimic the ability of Gm31629 to increase YB-1 protein level and reduce the expression of p16 INK 4 A in BMSCs ( Fig. 6A ). We then found that TF2A attenuated the senescence of BMSCs and promoted osteogenesis of BMSCs ( Figs. 6B – 6E ). TF2A did not affect osteoclast differentiation, as evaluated by TRAP staining ( Figs. 6F, 6G ). These data demonstrated that TF2A could mimic the activity of Gm31629 to increase the protein level of YB-1, thus alleviating the senescence of BMSCs. 10. 7717/peerj. 13475/fig-6 Figure 6 TF2A treatment in vitro alleviates BMSCs senescence. (A) YB-1 and P16 INK4A protein levels in BMSCs treated with vehicle or TF2A. (B) Representative images of SA-βGal staining of BMSCs treated with TF2A or vehicle. Scale bar: 50 µm. (C) The percentage of SA-βGal positive cells. ( n = 3). (D) ARS staining of BMSCs under osteogenic induction. Scale bar: 100 µm. (E) Quantification of calcium mineralization. ( n = 3). (F) TRAP staining of bone marrow monocytes and macrophages under osteoclast differentiation. Scale bar: 200 µm. (G) Quantification of multinuclear TRAP positive cells per well. ( n = 3). Data are expressed as mean ± sd and statistical differences were analyzed by Student’s t test. ∗∗ P < 0. 01; N. S, no significance. TF2A treatment promotes bone regeneration in middle-aged mice To investigate whether treatment of TF2A could promote bone regeneration in middle-aged mice, 12-month-old C57BL/6J mice were orally treated with TF2A at a dosage of 8 mg/kg every day or with vehicle for three weeks. There weeks after TF2A or vehicle treatment, the mice were performed with surgical ablation of trabecular bone in the right femur, and continued the treatment with TF2A for one week ( Fig. 7A ). Administration of TF2A had no obvious influence on the weight of the mice ( Fig. 7B ). Treatment with TF2A promoted bone regeneration in 12-month-old mice in comparison with the control group ( Figs. 7C – 7E ). The number of ocn+ osteoblasts in bone regeneration region was also increased after TF2A administration, which indicated increased bone formation ( Figs. 7F, 7G ). There was no significant difference of TRAP+ osteoclasts in bone regeneration region of TF2A treated group compared with that of vehicle treated group ( Figs. 7H, 7I ). These data indicated that TF2A could promote bone regeneration in middle-aged mice. 10. 7717/peerj. 13475/fig-7 Figure 7 TF2A treatment promotes bone regeneration in middle-aged mice. (A) Time point at which mice were treated with TF2A or vehicle, and performed with surgical ablation of trabecular bone in distal femoral. (B) The body weight of 12-month-old mice before and after treated with TF2A or vehicle. ( n = 6). (C) Representative micro-CT images. The white square was selected to measure trabecular bone volume in bone regeneration region. (D) Quantification of trabecular bone volume in bone regeneration region. ( n = 6). (E) HE staining of distal femora of middle-aged mice. Scale bar: 200 µm. (F) Immunohistochemical staining of osteocalcin positive cells. Black arrows represent osteocalcin positive cells. Scale bar: 50 µm. (G) Quantitative analysis of osteocalcin positive cells. ( n = 6). (H) TRAP staining images. Scale bar: 50 µm. (I) Quantitative analysis of TRAP positive cells. ( n = 6). Data are expressed as mean ± sd and statistical differences were analyzed by Student’s t test. ∗ P < 0. 05; N. S, no significance. Discussion Bone has natural healing ability that is sufficient to repair bone injuries and the capacity of bone repair is compromised during aging ( Lin et al. , 2019 ). BMSCs are capable of self-renewal and can differentiate into various tissues, and the therapeutic potential of BMSCs for bone repair has been widely accepted ( Park et al. , 2012 ; Squillaro, Peluso & Galderisi, 2016 ). However, BMSCs undergo senescence during aging and show an obvious impairment in their proliferation, migration and differentiation ability ( Li et al. , 2017 ; Sepulveda et al. , 2014 ; Xu et al. , 2018b ). In addition, senescent cells can secret substantial chemokines, proinflammatory cytokines, proteases, and other factors ( Xu et al. , 2018a ; Xu et al. , 2015 ). These factors are termed the senescence associated secretory phenotype (SASP) ( Xu et al. , 2015 ), which may contribute to impaired therapeutic effects of senescent BMSCs ( Sepulveda et al. , 2014 ; Turinetto, Vitale & Giachino, 2016 ). In this study, we showed that there was increased senescent BMSCs from 12-month-old mice in comparison with that from 3-month-old mice. BMSCs from 12-month-old mice exhibited an obvious impairment in their proliferation, migration and osteoblastic differentiation ability. Compared with 3-month-old mice, 12-month-old mice had compromised bone regeneration ability, accompanied by reduced osteoblast in bone regeneration area. Accordingly, the prevention of BMSCs senescence or rejuvenation of aged BMSCs is a promising strategy to improve bone regeneration. Recently, multiple studies have focused on the mechanism of BMSCs senescence ( Guo et al. , 2021 ; Hu et al. , 2022 ; Liu et al. , 2021a ). The emerging roles of lncRNAs in regulating cellular senescence have also been documented in previous studies ( Lee et al. , 2020 ; Xia et al. , 2017 ). Our previous studies demonstrated that Gm31629 could regulate the senescence of htNSCs, and loss of Gm31629 accelerated aging-like phenotype ( Xiao et al. , 2020 ). Here, we extended our research and demonstrated that Gm31629 could also regulate the senescence of BMSCs and bone regeneration. BMSCs from Gm31629-KO mice showed a premature aging phenotype and their proliferation, migration, and osteogenic differentiation abilities were reduced. Gm31629-KO mice had compromised bone regeneration ability with reduced osteoblast in bone regeneration area. We did not observe significant changes in osteoclast between Gm31629-KO mice and WT mice, which indicated Gm31629 had no effect on osteoclasts. Previously, Sun et al. (2019) reported that lncRNA lnc-ob1 could regulate osteoblast activity and bone formation via upregulating the expression of Osterix in osteoblast. Since Gm31629 -KO mice are global Gm31629 knockout and the compromised bone regeneration ability of Gm31629 -KO mice may also result from loss of function of osteoblasts, osteocytes or other bone cells. A tissue-specific mouse model will be more convincing to elucidate the role of Gm31629 in regulating BMSCs senescence and bone regeneration. At the mechanistic level, we found that Gm31629 regulated the senescence of BMSCs and bone regeneration via interacting with YB-1 protein to delay its degradation. YB-1 is a multifunctional protein that can bind RNA and DNA ( Lyabin, Eliseeva & Ovchinnikov, 2014 ). By binding to nucleic acids, YB-1 participates in basic gene expression process, including transcription, mRNA stabilization and translation ( Lyabin, Eliseeva & Ovchinnikov, 2014 ). At the cellular level, YB-1 has been reported to regulate a variety of biological activities including cell proliferation, differentiation, senescence and apoptosis ( Kotake et al. , 2013 ; Lyabin, Eliseeva & Ovchinnikov, 2014 ). For example, Kotake et al. (2013) demonstrated that YB-1 could bind to the promoter region of p16 INK 4 A, inhibit its expression and prevent cellular senescence. In this study, our results confirmed that YB-1 could bind to p16 INK 4 A promoter, repress the expression of p16 INK 4 A and prevent BMSC senescence. These findings suggest that Gm31629 -YB-1 signaling axis plays a critical role in BMSC senescence and bone regeneration. Previously, Evans et al. (2020) reported that YB-1 could fine-tunes Polycomb repressive complex2 (PRC2) activities to control embryonic neural development. The findings of Schmid et al. (2013) suggested YB-1 could act as a mediator of Melanoma inhibitory activity (MIA)/cartilage-derived retinoic acid-sensitive protein (CD/RAP) dependent chondrogenesis. These studies suggest that Gm31629 -YB-1 signaling axis may also affect chondrogenesis and neurogenesis of BMSCs, which requires further study. TF2A is one of the isomeric monomers of black tea theaflavins and theaflavins have been reported to have many beneficial effects for the health ( Anandhan et al. , 2012 ; Lin, Huang & Lin, 2007 ; Tong et al. , 2018 ; Zhang et al. , 2016 ). Previously, we identified that TF2A could mimic the activity of Gm31629 and reduce the senescence of htNSCs, thus further alleviating age-related physiological decline ( Xiao et al. , 2020 ). In this study, we further demonstrated that TF2A also could alleviate the senescence of BMSCs and improve bone regeneration ability of middle-aged mice. Consistent with the function of Gm31629, TF2A had no obvious effects on osteoclasts. 10. 7717/peerj. 13475/fig-8 Figure 8 Schematic representation of Gm31629 regulating BMSCs senescence and bone regeneration. Gm31629 interacts with YB-1 and delays its degradation, thus decreasing the transcription of p16 INK4A and suppressing the senescence of BMSCs. In old subjects, the decreased expression of Gm31629 drives the senescence of BMSCs and leads to impaired bone regeneration. In summary, we showed the important role of Gm31629 in regulating BMSCs senescence and bone regeneration. Gm31629 could interact with YB-1 and delay its degradation, thus decreasing the transcription of p16 INK 4 A and suppressing the senescence of BMSCs ( Fig. 8 ). Hence, this study provides a potential new approach to attenuate BMSCs senescence and improve bone regeneration ability in aged subjects. Supplemental Information 10. 7717/peerj. 13475/supp-1 Supplemental Information 1 Primer sequence used for qRT-PCR and ChIP assays Click here for additional data file. 10. 7717/peerj. 13475/supp-2 Supplemental Information 2 Raw data Click here for additional data file. 10. 7717/peerj. 13475/supp-3 Supplemental Information 3 Uncropped WB image Click here for additional data file. 10. 7717/peerj. 13475/supp-4 Supplemental Information 4 Checklist Click here for additional data file.

10. 7717/peerj. 13513

PeerJ

Microfluidics-assisted electrospinning of aligned nanofibers for modeling intestine barriers

During electrospinning, the fibers deposited on the collector are usually randomly oriented in a disordered form. Researchers hope to generate periodic structures to expand the application of electrospinning, including improving the sensing properties of electronic and photonic devices, improving the mechanical properties of solid polymer composites and directional growth of human tissues. Here, we propose a technique to control the preparation of aligned foodborne nanofibers by placing dielectric polymers on microfluidic devices, which does not require the use of metal collectors. This study was conducted by introduced PEDOT:PSS polymer as a ground collector to prepare aligned foodborne nanofibers directly on the microfluidic platform. The fluidity of the electrolytic polymer collector makes it possible to shape the grounding collector according to the shape of the microcavity, thus forming a space adjustable nanofiber membrane with a controllable body. The simplicity of dismantling the collector also enables it extremely simple to obtain a complete electrospun fiber membrane without any additional steps. In addition, nanofibers can be easily stacked into a multi-layer structure with controllable hierarchical structures. The Caco-2 cells that grow on the device formed a compact intestinal epithelial layer that continuously expresses the tightly bound protein ZO-1. This intestinal barrier, which selectively filters small molecules, has a higher level of TEER, reproducing intestinal filtration functions similar to those of in vivo models. This method provides new opportunities for the design and manufacture of various tissue scaffolds, photonic and electronic sensors.

Introduction Nanofibers can mimic the structure of the extracellular matrix structure due to their high specific surface area, which is a research focus in the field of biomedicine and biomaterials. Organized various types of nanofiber structures controllable and orderly is one of the crucial challenges in this field. Electrospinning is a simple method that can control the layout of nanofibers. This technology makes it possible to prepare nanofibers with diameters from 10 nanometers to several microns simply and multifunction by selecting a variety of materials, including metals, ceramics, synthetic and natural polymers ( Naghdi et al. , 2020 ; Xue et al. , 2017 ; Shan, Li & Huang, 2020 ). For example, polyvinylpyrrolidone (PVP) is widely used in the preparation of materials containing functional nanofibers of pharmaceutical nanofibers, inorganic organic composites and liposomes due to its remarkable properties, such as high hydrophilicity, complexing capacity, biocompatibility ( Biswas et al. , 2006 ; Li et al. , 2022 ). Although functional fibers are widely used in materials science, such as many standard processes established in the food industry, they are still less used in the food production and engineering. Nanofibers can be used to regulate the sensory and physicochemical properties of delicious foods, such as texture, flavor, appearance and stability ( Hada & Goli, 2019 ; Aytac et al. , 2017 ; Alehosseini et al. , 2019 ; Shababdoust et al. , 2020 ). Electrospinning nanofibers can also be suitable packaging materials, used to package and protect bioactive ingredients from external environmental factors, and has been used as an ideal platform for health food packaging and spoiled food monitoring ( Aytac et al. , 2017 ; Alehosseini et al. , 2019 ; Choi et al. , 2020 ; Lin, Ni & Pang, 2019 ; Lin et al. , 2019 ). Microfluidic devices, also known as lab-on-chips, are characterized by the flexible combination and scale integration of various cell technologies on the micro controllable platform ( Whitesides, 2006 ; Su, Liang & Tan, 2021a ; Su, Liang & Tan, 2021b ). The platform can accurately control and manipulate the microscale fluid, especially suitable for preparing electrospun nanofibers with specific patterns and arrangements ( Hu et al. , 2017 ). In a typical process, electrospun nanofibers based on microfluidics are collected on the grounded metal electrode. Due to the bending instability of the high charge jet, electrospun fibers are usually deposited on the surface of the electrode plate randomly to form two-dimensional (2D) randomly oriented nanofiber mats ( Reneker et al. , 2000 ; Park et al. , 2018 ). Park & Kim (2015) prepared spatially patterned and disordered nanofiber mats on the surface of potassium chloride solution using microfluidics-assisted electrospinning technology. In addition to the above disordered nanofiber mats, electrospun nanofibers with specific patterns and arrangements have attracted increasing attention due to their outstanding potential in creating functional structures or devices. For example, the spatial arrangement of nanofibers can induce anisotropic optical polarization ( Yao et al. , 2007 ), and provide guidance clues for cell arrangement, differentiation and migration through external internal signals ( Hu et al. , 2017 ; Li et al. , 2019 ; Chen et al. , 2019 ). The arrangement and pattern making of electrospun nanofibers able to create advanced products, such as multifunctional sensors ( Guo & Ding, 2021 ) and cell culture media with micro/nano hybrid structure ( Park et al. , 2018 ; Jia et al. , 2014 ). However, the electrospun materials used in the above literature need to be dissolved with additional toxic chemicals, which limits their applications in bioengineering, health care and food industry. Furthermore, the adhesion between the electrospun nanofiber felt and the metal surface is very strong, which will lead to the morphological change of the nanofiber felt in the stripping process. Therefore, it is ideal for developing a new method to generate well-aligned food-grade nanofibers and to deposit or transfer them to the surface of the solid substrate for device manufacturing. In this article, we propose a novel method for the preparation of edible aligned fibers based on microfluidic devices simply, which needn’t to use metal collectors and toxic chemicals. This study introduced patterned PEDOT:PSS polymer as a ground collector to prepare ordered nanofibers directly on the microfluidic platform. The fluidity of the electrolytic polymer collector makes it possible to shape the grounding collector according to the shape of the microcavity, thus forming a space controllable nanofiber membrane with a complex structure. The simplicity of dismantling the collector also enables it extremely simple to obtain a complete electrospun fiber membrane independently without any additional steps. In addition, nanofibers can be easily stacked into a multilayer structure with controllable hierarchical structure for modeling intestine barriers This method provides further opportunities for the design and manufacture of new health food customization, tissue scaffolds, photon and electronic sensors. Materials and Methods Materials Poly(3, 4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS) was provided by aladdin as a surfactant-free aqueous dispersion with 1. 5 wt% in deionized water. Polyvinylpyrrolidone (PVP) (1, 300 kDa, Macklin), Zein (99. 5%, Shandong Jiuyu Biology Development Co. Ltd, Shandong, China), β-Cyclodextrin (98%, Macklin), Photoresist (SU-8 3035; MicroChem Co. Ltd, Euless, TX, USA), SYLGARD™ 184 silicone elastomer kit (Dow Corning Co. Ltd, Midland, MI, USA). Preparation of microfluidic devices via soft lithography Microfluidic device was fabricated by the standard soft lithography method we previously reported ( Jiang et al. , 2012 ). In short, a clean silicon wafer was first spin coated with a negative photoresist (Microchem Co. Ltd, Euless, TX, USA). After the photoresist is baked at 95 °C for 20 min, the resist is exposed to ultraviolet light through a photomask made by CAD software, and developed in the developer 2-Acetoxy-1-methoxypropane (Aladdin Co. Ltd, Shanghai, China) solution. Then, PDMS was poured into the silicon wafer and cured on a hot plate at 80 °C. Finally, the replica of the microfluidic device is stripped from the silicon wafer. Electrospinning of aligned foodborne nanofibers on microfluidic devices In this study, the electrospinning equipment used which included high pressure supply units, pinning nozzle, syringe pump and collection unit ( Jiang et al. , 2012 ). The polymer solution containing 5–20 wt% PVP or zein solution were prepared by using anhydrous ethanol as solvent in a certain volume ratio for electrospinning. β-cyclodextrin solution was prepared for electrospinning at the concentration of 180 wt% in deionized water as reported by Celebioglu & Uyar (2020). The distance between the syringe needle and the collector is about 12 cm, and the applied voltage is set at 13 kV. PVP, zein and hydroxypropyl-beta-cyclodextrin (HP-β-CyD) solutions were continuously supplied to the needle by a syringe pump at a certain flow rate, and the electrospinning time was usually 1 min. Humidity maintained at 75–85% in the whole experiments. The highly concentrated (180%, w/v) of HP-β-CyD were prepared in distilled water ( Celebioglu & Uyar, 2020 ), while zein (20%, w/v) were prepared in ethanol/water with the ratio of 70:30 (v/v) ( Yao, Li & Song, 2007 ). A microfluidic device with microchannel and microcavity was fabricated by using the soft lithography technology. The surface of microchannel was modified by plasma treatment (300 W, 20 s) to obtain hydrophilic surface. A conductive solution is added at inlet of the microchannel (PEDOT:PSS), and the electrolyte solution was evenly distributed on the hydrophilic pattern on count of the wetting characteristics of the microchannel. Then drying for 24 h, which enable to establish the position-selective patterning of the electrolyte solution in microfluidic device. Next, high voltage is applied between the metal needle and the patterned electrode on the microfluidic device for electrospinning, and the aligned nanofiber mats are generated on the surface of the dielectric patterned electrode. As the control experiment, copper strips was used to collect the nanofibers. Morphological characterization of intestinal barrier Human colorectal cells (Caco-2, Cell Bank of the Chinese Academy of Sciences, Shanghai, China) were cultured in high glucose DMEM medium containing 10% fetal bovine serum and 1% penicillin-streptomycin solution at 37 °C and 5% CO 2. After the preparation of the chip, ultraviolet disinfection was carried out for 2 h before the start of the experiment. The hose was connected to the chip and the DMEM medium was perfused overnight. Finally, the Caco-2 cells were digested with trypsin and made into a mixed suspension, and the cell suspension with a density of 1 × 10 6 cell/mL was inoculated into the channel. After the cells were settled for 2 h, the microchip was placed in a cell incubator for 24 h until the cells grew to form an intact intestinal barrier. The cell viability was detected by CCK-8 method, that is, 10 μL CCK-8 reagent and 90 μL cell culture medium were added into the chip channel, and the absorbance at 450 nm was measured by microplate reader after 2 h of incubation ( Li et al. , 2022 ). The resulting intestinal barrier was subsequently morphologically characterized by immunohistochemistry. To assess tight junctions between cells at the top of the intestinal barrier, confocal immunofluorescence microscopy was used to stain the tight junction protein ZO-1, After rapid cell rinsing using 1 × PBS, Caco-2 cells were fixed with 4% paraformaldehyde for 15 min and permeabilized with 0. 2% Triton-X 100 for 2 min before incubation with fluorescently labeled ZO-1 antibody at 4 °C for 12 h. Samples were further incubated with antibody for 1 h and counterstained with 4′, 6-diamidino-2-phenylindole (DAPI) for 15 min. After three final washes with PBS, microscopy was performed using a Leica laser confocal microscope. Functional characterization of intestinal barrier On the basis of establishing the morphological integrity of the intestinal microarray barrier, the functional integrity of the human intestinal barrier was assessed by quantifying the specific activity of apical brush border alkaline phosphatase (ALP) and transepithelial electrical resistance (TEER). The specific activity of ALP was determined by obtaining the culture medium from the upper and lower chambers after 5 days of cell culture, transferring the solution to a 96-well plate, and cleaving the product ( e. g. , 4-nitroaniline) was quantified at 405 nm in a microplate reader. Then, the TEER value was obtained by multiplying the specific resistance which subtracted baseline resistance value (without cells) by the total cell culture surface area on the membrane. DMEM medium served as blank control group. After establishing tight junction integrity, the apparent permeability coefficient (Papp, cm/s) of intestinal cell monolayers was determined. Papp measurement was achieved primarily by measuring the transit of isosulphate luciferin (FITC)-labeled glucose (FD20, 20 kDa) over time on the intestinal barrier chip. FD20 solution (1 mg/mL) was added to the intestinal barrier on-chip chamber and samples collected hourly from the lower chamber outlet were analyzed to quantify the amount of FD20 that crossed the intestinal barrier. The fluorescence intensity (490 nm excitation/520 nm emission) of the sample collected from the lower chamber was immediately measured to quantify the amount of FD20 transported from the cell apex to the basolateral surface. Papp was calculated according to Papp (cm/s) = (dQ/dt)(1/AC0) where A is the culture surface area (cm 2 ), dQ/dt is the steady state flux (g/s) and C0 is the initial concentration (mg/mL) of the FD20 solution applied on the apical cell surface. Morphology and distribution analysis The morphology of the nanofibers was studied by inverted microscope (Olympus IX-71; Olympus, Tokyo, Japan) and field emission scanning electron microscope (SEM, TM3000; Hitachi, Tokyo, Japan). The fibers used in SEM were collected on the conductive adhesive, dried in air, and plated with gold by sputtering gold plating machine (SBC-12; Kyky, Beijing, China) for 60 s (to obtain a gold plating layer of about 10 nm), so as to improve its conductivity. The scanning electron microscope was operated under 15 kV accelerating voltage. Image J software was used to measure the size, orientation angle and distribution of nanofibers ( http://rsb. info. nih. gov/ij/ ). Results and Discussion Microfluidic device fabrication for aligned nanofibers electrospinning Figure 1 shows the sequential steps of fabricating spatially patterned aligned nanofibers on the surface of the dielectric polymer by microfluidic devices. Firstly, the mask pattern with microchannel is generated by laser micromachining, and microfluidic devices are fabricated by mask exposure combined with soft lithography. Secondly, the surface of the microchannel was modified by plasma to obtain a hydrophilic surface. Plasma technology can be compatible with a variety of insulators without changing the overall characteristics of the material so that it can be used for material self-assembly and micro-contact printing. The surface area of the dielectric polymer treated with oxygen plasma becomes more hydrophilic, mainly by introducing oxygen functional groups on the surface of the polymer to increase the surface wettability ( Martins et al. , 2009 ). The conductive solution PETO:PSS is added at one end of the microchannel, and the electrolyte solution is selectively positioned on the hydrophilic pattern by virtue of the wetting characteristics of the microchannel ( Fig. 1 ). Then drying for 24 h, which enable to establish the position-selective patterning of the electrolyte solution in microfluidic device. 10. 7717/peerj. 13513/fig-1 Figure 1 Schematic diagram of the nanofiber scaffold array integrated with microfluidic channels. Next, a high-voltage patterning device is applied between the electrode surface and the metal substrate to control the fluidization. Because the aligned nanofibers are suspended above the microfluidic device, after the nanofibers are collected, they can be easily transferred to the surface of another substrate by moving the collecting electrode vertically for further device processing and manufacturing. An ungrounded carrier made of insulator or metal can also be placed in the gap to directly collect the aligned nanofibers and help prevent the suspended fibers from breaking. Since both ends of the fibers are physically fixed on the surface of the collector, their position and direction can be easily controlled by moving the collector. These fibers can be stacked layer by layer into a multi-layer structure, with the fibers of different layers facing different directions. In order to form the nanofibers regularly, it is needed to find the optimal concentration of the solution. Results as shown in Fig. 2A, PVP nanofibers were prepared under the condition of 11 kV voltage and 0. 5 mL/h injection rate. At very low concentration (<3%), the polymer falls onto the surface by electrospray in the form of droplets before being converted into fibers, and the spinning structure cannot be obtained by electron microscopy. The low concentration polymer (5%) solution has low viscoelastic force and the low macromolecular chain entanglement, so the electrostatic repulsion force and coulomb repulsion force needed for spinning cannot match. Although the low concentration polymer is helpful to stretch the electrospun jet, the jet is partially broken, forming irregular curls or spinning fracture. In the case of high tension, a low tension spherical solution is formed on the surface of the polymer. When the solution concentration increases from 5% to 15%, the viscosity increases, resulting in the increase of viscoelasticity and chain entanglement. The increase of solution concentration is related to the production of larger diameter fibers ( Doshi & Reneker, 1995 ). The increase of viscoelastic force prevents the jet from breaking partially, and the solvent molecules are distributed on the entangled polymer molecules, thus forming smooth fibers and improving the uniformity of fibers. When the concentration value is very high (>15%), the fiber becomes very thick, sometimes forms a foil structure, and the spinning diameter becomes more and more uneven, whose phenomenonis consistent with the previous result with high concentration at 21% and 24% ( Tiyek et al. , 2019 ). In addition, proper concentration is also a necessary condition to ensure the consistent orientation of nanofibers ( Park et al. , 2018 ; Chen et al. , 2019 ; Zhuang et al. , 2020 ). 10. 7717/peerj. 13513/fig-2 Figure 2 SEM images of nanofibers dissolved in ethanol at various (A) concentration; (B) applied voltages; (C) injection speed. To form nanofibers, the applied voltage should be sufficient to overcome the surface tension of the solution. Once the electrostatic force overcomes the surface tension, the spinning process begins. Results as shown in Fig. 2B, when the concentration exceeds 10 wt% and the injection speed is 0. 5 mL/h, the increase of applied voltage leads to the increase of fiber diameter. The insufficient electric field force in electrospinning process means that the stretching force is inadequate to elongate the solution. When the applied voltage is low ( e. g. , 9 kV), the Coulomb force is not enough to overcome the surface tension, resulting in the formation of spinning branches. When the applied voltage exceeds a limit voltage ( e. g. , 20 kV), the thickness of the fiber increases and the irregular structure is formed. The reason is that the Coulomb force is much greater than the viscoelastic force and the jet velocity is much faster. Therefore, there is not enough time for solvent evaporation to form coarse and irregular fibers, which is consistent with the previous results ( Baumgarten, 1971 ). In addition, when the PVP concentration exceeds 10 wt% and the applied voltage exceeds 11 kV, the number of fibers increases sharply with the increase of injection speed ( Fig. 2C ). It can be seen from Fig. 2C that different injection speeds have little effect on the fiber diameter, only the fiber shape and parallelism, which is consistent with previous reports ( Tiyek et al. , 2019 ). Microfluidic electrospinning for aligned nanofibers The electric field control near the integrated collector can attract highly positively charged electrospun nanofibers to form a patterned or aligned nanofiber mat structure ( Park et al. , 2018 ). We compared the effect of microfluidic device and common copper foil on the orientation of nanofibers. The SEM images of PVP electrospun nanofibers are fabricated in the aid of microfluidic patterned electrodes ( Fig. 3A ) and copper foil electrodes ( Fig. 3B ). The average diameter of PVP nanofibers produced by microfluidic patterned electrode and copper foil electrode is shown in Fig. 3. Both microfluidic patterned electrodes and copper foil electrodes collector can produce nanofibers, and the morphology of PVP nanofibers will not change with the collector type. The average orientation angle of PVP nanofibers produced by microfluidic patterned electrodes and copper foil electrodes is also shown. Compared with the copper foil electrodes, the rotation angle of the long axis of the nanofibers collected by the microfluidic patterned electrode is smaller, which means that the orientation of the nanofibers is better. 10. 7717/peerj. 13513/fig-3 Figure 3 The SEM image of the aligned nanofiber fabricated via a microfluidic device (A) and copper foil collectors (B). The SEM image of the aligned nanofiber fabricated via a microfluidic device (A) and copper foil collectors (B). The distribution of the normal angle between the long axis of the fiber and the edge of the parallel electrode was also statistically analyzed. The results displayed on each board came from measurements of more than 120 fibers. These results indicated that the electrolyte and the metal collector played similar role in the manufacturing process of electrospun nanofibers. In the traditional electrospinning process, the metal collector received the applied high voltage and generated a high-density electric field around the metal needle to facilitate the generation of a charged high-density polymer jet from the Taylor cone of the needle ( Li & Xia, 2004a ). The microfluidic patterned electrodes in opposite directions can make the positively charged polymer jets be distributed between the electrodes in parallel. During the jetting process of the polymer passing through the needle, especially the electric repulsion between the positive charges within the needle, the polymer jet is reduced to nanometer level ( Doshi & Reneker, 1995 ). The polymer jet undergoes whip like motion induced by the electric repulsion force between positive charges in the jet, which reduces the polymer jet to nanoscale ( Liu et al. , 2021 ). Therefore, the microfluidic patterned electrode also allows a continuous process of high charge polymer jet and whipping motion through the repulsive force between charges in the polymer jet through the concentrated electric field at the metal needle. Using the device shown in Fig. 1, we have also successfully prepared uniaxially oriented nanofibers from other food materials, such as cyclodextrin and zein. This method seems to be a general method by which many types of food materials can be electrospun into uniaxially aligned nanofibers without harmful organic solvents. Figure 4 shows an SEM image of the aligned nanofibers with different components and properties ( e. g. , β-cyclodextrin and zein) with ~500 nm and ~200 nm average fiber diameter, respectively. Edible nanofibers have good biocompatibility and exhibit good proliferative capacity of human hepatocytes and mouse fibroblasts, but are limited to very poor mechanical properties ( Dong, Sun & Wang, 2004 ). By introducing an insulator in the collector, the aligned composite nanofiber array can be easily collected. These fibers can also be directly deposited on glass and PDMS substrates, or transferred to glass and PDMS substrates for calcination or device manufacturing. The combination of nanospinning and template oriented technology can further expand the range of materials that can be made into uniaxially oriented nanofibers ( Yang et al. , 2007 ; Ye et al. , 2018 ). The Rayleigh scattering of these food derived nanofiber arrays is also anisotropic, which may be an effective optical polarizer ( Li & Xia, 2004b ). In addition to the production of nanofibers made from pure organic polymers, biomacromolecules ( e. g. proteins or enzymes), nanoparticles ( e. g. , superparamagnetic iron oxide), inorganic nanowires, etc ( Xue et al. , 2017 ; Alehosseini et al. , 2019 ; Wang, Su & Tan, 2020 ). These materials can be easily incorporated into electrospinning to obtain uniaxial oriented nanofibers with the required functional combination. 10. 7717/peerj. 13513/fig-4 Figure 4 SEM images of uniaxially aligned nanofibers made of different foodborne materials: (A) zein; (B) β–cyclodextrin. Multilayered structures of aligned nanofibers on microfluidic devices Layers of nanofibers can be stacked in different directions. As shown in Fig. 5, by transferring the uniaxially oriented nanofibers suspended in the gap to the same substrate (in a layer-by-layer manner), the array cross connection can be easily achieved. The aligned nanofibers can be easily stacked into a multilayer structure fiber membrane with a controllable hierarchical structure using a similar method. When multiple pairs of PEDOT:PSS electrodes are used as collectors, PVP nanofibers can also be directly deposited to form a multi-directional parallel fiber array. The spatial orientation and position of PVP nanofibers are determined by the position and structure of the electrodes. By controlling the position of the electrode and applying high voltage, the aligned nanofibers can be easily stacked into multilayer films. By changing the grounding mode of microfluidic patterned electrode pairs in situ, the double-layer network structure of nanofibers can be obtained ( Fig. 5A ). After collecting the first layer on the substrate, we collect the second layer in the electrode direction with 90° rotation angle. Figure 5A shows a SEM image of a rectangular pattern with a rotation angle of 90° formed by the method. The three-dimensional image shows the spatial outline of a grid composed of arranged fibers. We generated a two-layer grid height profile along a straight line through the nanofibers. 10. 7717/peerj. 13513/fig-5 Figure 5 Schematic illustrations of multilayered structures of aligned nanofibers that were composed of four (A) and six (B) electrodes deposited on microfluidic devices. During collection, the electrodes were alternately grounded for ~30 s. The three-dimensional image shows the spatial outline of a grid composed of arranged fibers. The three-layer grid was generated by alternating the grounding electrode pairs (1–4, 2–5 and 3–6) and placing three layers of uniaxially aligned nanofibers with their long axis rotated by 60 degrees. Figure 5B shows the patterning method of microfluidic electrode composed of six gold electrodes. The three-dimensional image shows the spatial outline of a grid consisting of arranged fibers. We generated a three-layer grid height profile through the nanofibers ( Fig. 5B ). In these cases, the measured angle between the fibers in the different layers is consistent with the corresponding rotation angle. By controlling the electrode structure and/or the sequence of applying high pressure, more composite structures composed of well-aligned nanofibers can be prepared. A remarkable feature of this technology is that in the production process of nanofibers, the arrangement and assembly of nanofibers can be completed simultaneously on the microfluidic device. By using the specially designed electrode as the collector, it is convenient to assemble multi-layer nanofibers with different components on the solid substrate, so as to form a highly ordered layered structure. This kind of structure has potential application value in the manufacture of food, electronic and photonic devices in the future. In addition, different nanofiber arrays can also be integrated into the same device, because a large number of electrode patterns can be read on a single substrate using traditional microfabrication technology. Reconstitution of intestinal barrier functions on microfluidic devices The chemical properties, surface morphology and structure of biomaterials can affect the function of cells ( Lu et al. , 2020 ). To further evaluate the barrier function of the intestinal chip, immunofluorescence microscopes were used to evenly stain the tightly connected protein ZO-1 and to assess the integrity of the tight connection at the top of the intestinal barrier. As shown in Fig. 6A, Caco-2 cells were found formed confluent polygonal epithelial monolayers on nanofibers of microfluidic devices. CCK-8 test also showed PVP non-oriented nanofibers and PVP oriented nanofibers have good biocompatibility and do not affect the viability of Caco-2 cells ( Fig. 6B ). The Caco-2 cells edges were well delimited by green junctions and well-organized in a characteristic geometry, indicating the integrated tight junction assembly of intestinal Barrier ( Liang, Su & Tan, 2021 ). 10. 7717/peerj. 13513/fig-6 Figure 6 Characterization of intestinal viability of the Caco-2 monolayer under different conditions. (A) Immunofluorescent images show the expression of tight junction protein, ZO-1 (green), in the Caco-2 monolayer. Bars: 50 μm. (B) Quantified analysis of the Caco-2 cell viability by cell counting kit-8 (CCK-8) kit. Data are shown as mean ± SD ( N = 3). In addition to cytomorphic characterization, we further characterize the intestinal barrier function of the intestinal chip. Alkaline phosphatase (ALP) is commonly used marker for intestinal epithelial cell differentiation as a brush border enzyme. The results showed that cell cultured in the microfluidic device with oriented nanofibers got 2. 6-fold higher ALP activity when compared to the cells cultured in control ( Figs. 7A, 7B ). The results indicated that the oriented nanofibers may promote the differentiation of intestinal cells due to the orderly arrangement structure. Fluorescent glucose was used to measure the Papp value of the intestinal epithelial, so as to characterize the effect of spinning polarity on the intestinal epithelial dysfunction. Compared to the control group, the proliferation of fluorescent dextran in the oriented nanofibers group was significantly reduced due to the presence of intestinal barrier on the chip ( Fig. 7C ). TEER is an important functional parameter for monitoring the quality of the in vitro intestinal barrier. As shown in Fig. 7D, we measured the TEER value of the intestinal barrier chip and subtracted the baseline resistance value of the control sample. The results showed that cells growing under all three cultures showed an increase in TEER values over time during the first 6 days after vaccination, and then maintained similar high levels during at least eight additional days of culture. This small molecule of intestinal filtering barrier selectively as well as the higher TEER levels reproduced the function of intestinal filtration similar to the models in vivo. 10. 7717/peerj. 13513/fig-7 Figure 7 Evaluation of intestinal differentiation and barrier functions under different conditions. (A) The ALP vitality of intestinal cells under distinct cultures were assessed on apical (AP) and basolateral (BL) sides, respectively. Results were obtained from triplicate independent experiments. (B) The apparent permeability coefficient (Papp) was measured by quantifying FD20 transport across the Caco-2 monolayer under different conditions. (C) The barrier integrity of the Caco-2 monolayer was quantified by TEER. Data are shown as mean ± SD ( N = 3). ** P < 0. 01, * P < 0. 05. Conclusion In conclusion, we show a simple and effective method to prepare edible micro nanofibers with a large area and suitable arrangement. This method does not require the use of metal collectors and toxic chemicals. Only a small amount of PEDOT:PSS aqueous solution is involved to perform in situ spinning within the microfluidic devices. The electrospun fibers are arranged in parallel along the direction of the microfluidic patterned electrode. This method also allows the construction of multidimensional structures based on other complex shapes of parallel arrays (such as meshes). The simplicity of dismantling the collector also enables the electrospun nanofiber membrane to exist independently without any additional steps. The ability of microfluidic devices to produce periodically ordered structures can broaden the application of nanofibers, such as in food manufacturing, nanoelectronic devices, polymer composites and tissue engineering. Supplemental Information 10. 7717/peerj. 13513/supp-1 Supplemental Information 1 Raw data of ALP ratio, CCK-8, Papp, TEER, etc. Click here for additional data file.

10. 7717/peerj. 13862

PeerJ

Histone H3K9 demethylase JMJD2B/KDM4B promotes osteogenic differentiation of bone marrow-derived mesenchymal stem cells by regulating H3K9me2 on RUNX2

Background A variety of proteins including epigenetic factors are involved in the differentiation of human bone marrow mesenchymal stem cells. These cells also exhibited an epigenetic plasticity that enabled them to trans-differentiate from adipocytes to osteoblasts (and vice versa) after commitment. Further in-depth study of their epigenetic alterations may make sense. Methods Chromatin Immunoprecipitation-PCR (ChIP-PCR) was used to detect the methylation enrichment status of H3K9me2 in the Runx2 promoter, alizarin red and alkaline phosphatase (ALP) staining were used to detect osteogenic differentiation and mineralization ability, western blot and quantitative RT-PCR were used to measure the differential expression of osteogenesis-related proteins and genes. Recombinant Lentivirus mediated gain-of-function and loss-of-function study. The scale of epigenetic modification was detected by laser confocal. Results Our results showed that compared with human bone marrow mesenchymal stem cells (hBMSCs) without osteogenic differentiation treatment, hBMSCs after osteogenic differentiation significantly promoted osteogenic differentiation and mRNA expression such as JMJD2B/KDM4B, osteogenesis-related genes like Runx2 and FAM210A in hBMSCs cells, suggesting that upregulation of JMJD2B/KDM4B is involved in the promoting effect of osteogenesis. After overexpression and silencing expression of JMJD2B, we found a completely opposite and significant difference in mRNA expression of osteogenesis-related genes and staining in hBMSCs. Overexpression of JMJD2B/KDM4B significantly promoted osteogenic differentiation, suggesting that JMJD2B/KDM4B could promote osteogenesis. In addition, ChIP-PCR showed that overexpression of JMJD2B/KDM4B significantly reversed the methylation enrichment status of H3K9me2 in Runx2 promoter. Furthermore, overexpression of JMJD2B/KDM4B significantly reverses the inhibitory effect of BIX01294 on H3K9me2, suggesting that JMJD2B/KDM4B regulates the osteogenic differentiation of hBMSCs by changing the methylation status of H3K9me2 at the Runx2 promoter. Conclusions Taken together, these results suggest that JMJD2B/ KDM4B may induce the osteogenic differentiation of hBMSCs by regulating the methylation level of H3K9me2 at the Runx2 promoter.

Introduction The research on the regulation of human bone marrow mesenchymal stem cells(hBMSCs) towards bone formation has been going on for decades. These cells exist in the bone marrow cavity and can differentiate into a variety of cell phenotypes, including osteoblasts, chondroblasts, stromal cells that support hematopoiesis, and adipocytes. The mainstream control methods found in current research can be roughly divided into the following categories: (1) Cell-to-cell regulation. For example, the co-culture of hAMSCs/hBMSCs mixed at the optimal ratio of 3/1 showed significantly higher cell proliferation, antioxidant properties, osteogenic and angiogenic differentiation than hBMSCs or HUVECs alone ( Zhang et al. , 2018 ). (2) Hormone. Like insulin ( Zhang et al. , 2020 ), in addition, there are some xenogeneic hormones that also play a role in inducing differentiation to a certain extent ( You & Xu, 2020 ). (3) Biomaterials. Through anodization and subsequent dip coating treatment, two highly ordered nano-pits with two different sizes were successfully constructed on the surface of 316LSS and then stimulated osteogenesis and angiogenesis ( Ni et al. , 2020 ). Shrestha et al. (2020) also found that both the composition and form of the hydrogel can determine the success of tissue formation, and these two factors have a complex interaction with cell behavior and matrix deposition. This has important implications for tissue engineering. In addition, a natural osteoinductive biomaterial nacre shell also has a similar function ( Green, Kwon & Jung, 2015 ). (4) Large amounts of Chinese medicine monomers. e. g. Polydatin Chen et al. , 2019a ; Chen et al. , 2019b ; Shen et al. , 2020 ), Chrysosplenetin ( Hong et al. , 2019 ), Hydroxysafflor yellow A ( Wang et al. , 2021 ) and Icariin ( Xu et al. , 2021 ). In addition, there are many kinds of exosome research. Such as, miR-375 ( Chen et al. , 2019a ; Chen et al. , 2019b ), There are also some regulatory factors that are inconvenient to be placed elsewhere. Classics such as Ocn, Opn, Osx, and some newly discovered ones such as TIMP-1 ( Liang et al. , 2019 ), brain-derived neurotrophic factor ( Liu et al. , 2018 ). Some researchers have also discovered the epigenetic mechanism of hBMSCs ( Cho et al. , 2005 ; El-Serafi, Richard Oreffo & Roach, 2011 ; Lee et al. , 2021 ). Some changes in the cellular environment can also cause hBMSCs to differentiate towards osteogenic direction. Pulsed electromagnetic field (PEMF) has also been successfully applied to accelerate fracture repair since 1979 ( Fu et al. , 2014 ). There are so many ways to regulate the directed differentiation of hBMSCs. Research on the effectiveness of promoting osteogenesis has become less attractive. It is also not easy to identify which type of regulatory means is most effective in promoting the osteogenic differentiation of hBMSCs. Therefore, we turned our attention to the changes in the cells themselves caused by various ways of promoting osteogenic differentiation. Most interestingly, we found that during the MSCs-transdifferentiation process were in relationship to the profiles of histone marks ( Meyer et al. , 2016 ). Changes of this type are called epigenetic modifications. Epigenetic refers to genetic changes in gene expression without altering the DNA sequence, including DNA methylation, histone modification and small non-coding RNA related regulation ( Wu & Sun, 2006 ), Lyu et al. (2019) found that large-grit, acid-etched (SLA) method-treated surfaces and mechanically processed surfaces have different effects on genome-wide DNA methylation and osteogenic pathway enrichment in hBMSCs. Among the existing epigenetic regulation of hBMSCs, MI192 and its regulated HDAC enzyme ( Man et al. , 2021b ), Setd7 ( Yin et al. , 2018 ), CUDC-907 ( Ali et al. , 2017 ), RG108 ( Assis et al. , 2018 ), have been discovered. Man et al. (2021a) significantly changed the epigenetic function of osteoblasts by using the HDAC inhibitor trichostatin A (TSA) to reduce HDAC activity and increase histone acetylation. Accelerate their mineralization and promote the osteoinductive ability of secretory extracellular vesicles (EVs). Epigenetics is reversible and susceptible to factors such as environment and drugs. There are many traditional Chinese medicines that can promote the differentiation of hBMSCs towards the osteogenic direction. Wang et al. (2021) discovered that one of the essential compounds of safflower: Hydroxysafflor yellow A, could promote osteogenesis and bone development via epigenetically regulating β -catenin and prevent ovariectomy-induced bone loss. Some scholars have also discovered that the epigenetic landscape of 3D cell models of human primary articular chondrocytes (hPACs) and human bone-marrow derived mesenchymal stem cells (hBMSCs) exhibits a huge difference in DNA methylation landscape ( Bomer et al. , 2016 ). This may indicate the future research direction of tissue engineering. In our opinion, epigenetics has been fairly close to the osteogenic differentiation nature of hBMSCs. In another recent epigenetic study we found that vascular smooth muscle cells (VSMCs) exhibit osteoblast-like characteristics in response to various stimuli such as oxidized cholesterol and inflammation. In the study of Kurozumi et al. (2019), it was found that IL-6 and sIL-6R induces STAT3-dependent differentiation of human VSMCs into osteoblast-like cells through JMJD2B/KDM4B -mediated histone demethylation of Runx2. This had caused us to think about the osteogenic epigenetic regulation mechanism of hBMSCs based on Runx2. To further investigate the role of JMJD2B/KDM4B in regulating the differentiation of hBMSCs into osteoblasts, we performed the following experiments. Materials & Methods Cell culture Human bone marrow-derived mesenchymal stem cell line was obtained from Saliai Biotechnology (G02007; Saliai Biotechnology, Guangzhou, China). Cells were maintained in α -MEM (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS), 50 µM ascorbate, 1% penicillin-streptomycin, and 10 mM β -glycerophosphate in a humidified atmosphere with 5% CO 2 at 37 °C. Cells at the third or fourth passage were used in the experiments. For osteogenesis treatment, cells were cultured in human bone marrow mesenchymal stem cells osteogenic medium (HUXMA-90021; Cyagen). Plasmid construction and recombinant lentivirus packaging Human JMJD2B was amplified using primers 5′- gcatgcATGGGGTCTGAGGACCACG -3′ (sense) and 5′- atgcatTCATCTGCAAGGGTCTTGAGTTG -3′ (antisense) and cloned into LV5-GFP/Puro vector (Clontech, Mountain View, CA, USA). The clones were confirmed by DNA sequencing. Add the shuttle plasmid and packaging plasmid (LV5-JMJD2B, pRev, pVSV-G) containing the target sequence to 293T cells in proportion. The hJMJD2B shRNA was designed according to the shRNA design rules and other literature ( Ye et al. , 2012 ). Bbs I; and BamH I; restriction sites were added to the shRNA end. The designed shRNA sequence was created by Jimma Gene(Shanghai, China). sh-hJMJD2B F: 5′-CACCGCCTGCCTCTAGGTTCATAATTCAAGAGATTATGAACCTAGAGGCAGG TTTTTTG-3′; Sh-hJMJD2B R: 5′-GATCCAAAAAACCTGCCTCTAGGTTCATAATCT CTTGAATTATGAACCTAGAGGCAGGC-3′. hJMJD2B shRNA was amplified and cloned into pGPU6/GFP/Neo vector (Clontech, Mountain View, CA, USA). The clones were confirmed by DNA sequencing. Add the lentiviral vector (Double-stranded GPR78 shRNA) containing the target sequence to 293T cells in proportion. Transfection Overexpression and silencing recombinant lentiviruses were purchased from Shanghai Hanheng Biotechnology. Collect hBMSCs cells in logarithmic growth phase, count, resuspend cells in complete medium, adjust the cell concentration to 1 ×10 5 cells/ml, inoculate into six-well plates, add 2ml of cell suspension to each well, and store at 37 °C, 5% CO 2 conditions Incubate overnight. Incubate overnight until the cell density is about 40% to 60%, and carry out virus infection by 1/2 small volume infection method. Aspirate the original medium in the culture well, add 1/2 volume (1 ml) of α -MEM low-glucose medium, and add polybrene to a final concentration of 5 µg/ml, add the corresponding volume of virus stock solution at MOI = 50, shake well, and put Incubate for 4 h at 37 °C, 5% CO 2. After 4 h of incubation, add 1/2 volume (1 ml) of α -MEM low-glucose medium to the culture wells, and incubate overnight at 37 °C under 5% CO 2 conditions. On the second day after infection, the virus-containing medium was aspirated, replaced with fresh α -MEM low-glucose medium (without polybrene), and the culture was continued at 37 °C, 5% CO 2. After hBMSCs were infected with virus for 72 h, the cells were digested and collected, and the cellular protein and RNA were extracted. The expression of JMJD2B in hBMSCs cells was detected by WB and qPCR. The method is the same as before. Alkaline phosphatase staining Data were collected as previously described in Wu et al. (2021). Specifically, Alkaline phosphatase staining was performed using a BCIP/NBT Alkaline Phosphatase Color Development Kit (C3206; Beyotime, China) following the manufacturer’s manual. hBMSCs cells were cultured in osteogenic medium for 1, 3, 7 or 14 days, washed with ice-cold PBS, and fixed with 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) for 15 min. Add BCIP solution (300X), NBT solution (150X), and BCIP/NBT dyeing working solution in sequence according to the following proportions. Add the BCIP/NBT dyeing working solution obtained after mixing to ensure that the sample can be fully covered. Incubate in the dark at room temperature for 5–30 min or longer (up to 24 h) until the color develops to the expected depth. The ALP positive cells stained blue/purple. For each experiment, a minimum of three washes was counted and the experiments were repeated three times. ALP staining was observed using a 450-fluorescent inverted phase-contrast microscope (Olympus, Tokyo, Japan) Alizarin red staining Data were collected as previously described in (Wu et al. , 2021). Specifically, cells were cultured in osteogenic medium for 1, 3, 7 or 14 days and then fixed in 95% ice-cold ethanol for 15 min. After three washes with ice-cold PBS, cells were stained with alizarin red (Sigma-Aldrich) for 5 min. Alizarin red staining was observed using an Olympus 450-fluorescent inverted phase-contrast microscope (Olympus, Tokyo, Japan). Quantitative real-time PCR (qRT-PCR) Total RNA was isolated from hBMSCs cells using Trizol reagent (DP424, Tiangen, Beijing, China). All quantitative real-time PCR assays were carried out using three technical replicates and three independent cDNA syntheses. Reverse transcription reaction using cDNA synthesis kit (KR118; Tiangen, Beijing, China), qRT-PCR was performed using a fluorescence quantitative PCR kit (QP002, Fulengen, Guangzhou, China) and gene-specific primers ( Table 1 ). GAPDH was used as an internal reference. The relative gene expression was determined using the 2 − ΔΔCt method. The PCR reactions were performed in triplicates. 10. 7717/peerj. 13862/table-1 Table 1 Primer sequence details. Gene name Gene ID Primer sequence (5′–3′) Amplification length (bp) h-JMJD2B NM_001370093. 1 Forward: GGACTAGAGGCCGTCTAAATTG 91 Reverse: ACTTCCTGCGTGCAAAGA h-Runx2 NM_001015051. 3 Forward: GCTTCATTCGCCTCACAAAC 112 Reverse: GTAGTGACCTGCGGAGATTAAC h-FAM210A NM_001098801. 2 Forward: CTGATGGGCGTAAGGAAGAAA 110 Reverse: TGGGTCTTTCCCAAGCATAC h-GAPDH NM_001289745. 3 Forward: CAAGAGCACAAGAGGAAGAGAG 102 Reverse: CTACATGGCAACTGTGAGGAG h-Osteocalcin NM_199173. 6 Forward: TCACACTCCTCGCCCTATT 114 Reverse: CCTCCTGCTTGGACACAAA h-Osteopontin NM_000582. 3 Forward: CATATGATGGCCGAGGTGATAG 108 Reverse: AGGTGATGTCCTCGTCTGTA h-Osterix NM_001173467. 3 Forward: CATTCTGGGCTTGGGTATCT 93 Reverse: GGCCTGAGATGAGAGTTTGT h-Runx2 (ChIP-PCR) NG_008020 Forward: TAATCTCCGCAGGTCACTAC 238 Reverse: ATACAAACCATACCCAAACC Western blot analysis Data were collected as previously described in (Wu et al. , 2021). Specifically, proteins were isolated from hBMSCs cells using RIPA buffer. Protein samples were separated using 5% SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane. The membrane was incubated with primary antibody against target protein ( Table 2 ), and β -actin (1:2000; B1033; Biodragon) Incubated overnight at 4 °C. The membrane was washed three times with TBST and incubated with HRP-linked secondary antibody (1:1, 000; 7074P2; Cell Signaling Technology) for 1 h at room temperature. 10. 7717/peerj. 13862/table-2 Table 2 Antibody details. Manu. Cat. Isotype Dilution Rabbit anti-KDM4B antiobdy Abcam Ab191434 Rabbit 1:1000 Hrp-Goat Anti-Rabbit IgG Jackson 111-035-003 Goat 1:10000 β -actin Polyclonal Antibody Biodragon B1033 Rabbit 1:2000 Rabbit anti-H3K9me2 antibody Biovision 6814-25 Rabbit 1:1000 Rabbit Anti-RUNX2 antibody Abcam ab76956 Rabbit 1:1000 Rabbit anti-FAM210A antibody novusbio NBP2-13992 Rabbit 1:1000 Chromatin immunoprecipitation (ChIP) Treatment of cells with ultrasound cross-linking. ChIP was performed using a ChIP Assay Kit (22202-20; Beyotime, China) following the manufacturer’s instruction. Briefly, cell lysates were centrifuged at 12, 000 rpm for 5 min at 4 °C. Take out 20 µl sample as Input for subsequent testing. Add 70 µl protein A+G Agarose/Salmon Sperm DNA to the remaining nearly 2 ml sample (about 35 µl is precipitation and 35 µl is liquid), and mix slowly or shake at 4 °C for 30 min. The supernatant was collected, and incubated with 0. 5–1 µg anti-H3K9me2 (6814-25; Biovision) and protein A+G Agarose/Salmon Sperm DNA (of which about 30 µl is precipitation and 30 µl is liquid) overnight at 4 °C. The immunoprecipitates were collected and washed three times with wash buffer. Next, the DNA fragments are purified for subsequent PCR experiments. Wash the sample three times with Elution buffer. Add 20 µl of 5M NaCl to 500 µl of supernatant, mix well, and heat at 65 °C for 4 h to remove crosslinks between protein and genomic DNA. For the 20 µl sample obtained as Input, add 1 µl 5M NaCl, mix well, and heat at 65 °C for 4 h. Then add an appropriate amount of EDTA, proteinase K, Tris, chloroform, glycogen, NaAc, and absolute ethanol for further purification, and the obtained DNA precipitate is used for PCR detection of the target gene. Immunofluorescence staining hBMSCS cells were seeded in a 6-well plate at a density of 1 ×10 5 cells/well. Cells were washed three times with PBS. After treatment, Cells were fixed in 4% paraformaldehyde solution for 15 min at room temperature, followed by fixation with pre-cooled acetone for 15 min. Later, they were washed twice in PBS followed by a 30-minute incubation in 2% bovine serum albuminsolution (Sigma-Aldrich, Cat# A7906-50G). Cells were then treated with 0. 5% PBS-Triton X-100 for 20 min and incubated with 3% H 2 O 2 for 15 min. The samples were then blocked with 5% serum for 20 min and incubated with JDJM2B (1:1000; Ab191434; Abcam) and FAM210A(1:200; NBP2-13992; Novusbio) overnight at 4 °C. Then add 200 µL of APC-labeled secondary antibody (1:500) to the sample, and incubate for 1 h at 37 °C, 5% CO 2, and saturated humidity. The nucleus was stained with DAPI (S2110; Solarbio) for 1 min. Finally, the samples were washed to remove excess solution and mounted on microscope slides for imaging Images were collected using VistarImage 3. 0 software (Nikon, Japan) software connected to a Olympus CKX53 inverted microscope. Images were acquired using an UPlanFLN objective at magnification 40 ×. Statistical analysis Data are expressed as the mean ±standard deviation. Statistical analysis was performed using SPSS 23. 0 software (SPSS Inc. , Chicago, IL, USA). Comparisons among different groups were performed using one-way analysis of variance followed by Tukey’s post hoc test. A P value less than 0. 05 was considered statistically significant. Results Osteogenesis promotes JMJD2B expression in hBMSCs cells To investigate the effect of osteogenic medium on hBMSCs proliferation, we treated hBMSCs cells at different time points. To further examine the effect of osteogenic medium treatment on osteogenesis and the involvement of histone modification, we determined the mRNA levels of JMJD2B as well as osteogenesis- and myogenesis- genes in hBMSCs cells at 0, 1, 3, 7, 14 days. The results showed that the osteogenic medium significantly upregulated the mRNA expression of Runx2 during the initial (day 3) osteogenic induction process, and furthermore dramatically upregulated the mRNA expression of FAM210A and JMJD2B at 7 and 14 days simultaneously ( Fig. 1A ). 10. 7717/peerj. 13862/fig-1 Figure 1 The effects of osteogenic medium on osteogenic response of hBMSCs cells. (A) hBMSCs cells were cultured in osteogenic medium as indicated. Quantitative real-time PCR (qRT-PCR) was performed at d0, d1, d3, d7, and d14 after treatment to measure mRNA expression of JMJD2B as well as osteogenesis- and myogenesis-related genes. GAPDH was used as an internal reference. Data are expressed as the mean ±standard deviation (SD). ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001 vs. D0. All experiments were independently repeated at least three times. (B) Western blot analysis was conducted to measure protein expression of Runx2, JMJD2B, FAM210A protein levels. β -actin was used as an internal reference. (C) Immunofluorescence staining was performed based on JMJD2B and FAM210A in hBMSCs. Scale bar = 20 µm. To further investigate whether osteogenesis affects related protein level. We determined the related protein level in hBMSCs cells treated with osteogenic medium. As shown in Fig. 1A, compared with cells at day 0, the proper osteoinduction time markedly elevated the protein levels of total Runx2, JMJD2B and FAM210A but slightly inhibited Runx2 at day 1, suggesting that osteogenesis may be induced by JMJD2B-mediated histone modification and the expression of FAM210A and Runx2. Next, we detected structure of histone demethylase regulating epigenetic modification in hBMSCs cells to investigated whether JMJD2B-mediated histone demethylation affects hBMSCs differentiation. JMJD2B immunofluorescent staining showed that osteogenesis substantially promoted JMJD2B level at 7, 14 days compared with the control (day 0), whereas no obvious changes in fluorescence signal were shown in 1 day and 3 days. Of note, osteogenesis also significantly promoted FAM210A level ( Fig. 1C ), which plays a vital role in regulating the structure and function of bones and muscles. These data suggest that osteogenesis may be induced by regulating JMJD2B-mediated histone modification and Runx2 and FAM210A-mediated differentiation of hBMSCs. Histone H3K9 Demethylase JMJD2B promotes osteogenic differentiation of hBMSCs To examine whether overexpression of JMJD2B could reverse the effect of silence JMJD2B lentivirus-transfected hBMSCs on osteogenesis, we overexpressed/silenced JMJD2B in hBMSCs cells and then treated the cells with vehicle or osteogenesis Induction Medium. As shown in Figs. 2A and 2B, qRT-PCR and western blot revealed that in general, overexpression of JMJD2B remarkably enhanced mRNA expression and protein level of JMJD2B, while silence of JMJD2B attenuating mRNA/protein expression of JMJD2B ( Figs. 2A – 2B ), regardless of the status of initial protein level. Compared with empty vector transfection (Scr and NC), overexpression of JMJD2B enhanced the intensities of ALP and alizarin red staining, in the presence of osteogenic medium, significantly in 7, 14 days. Silence of JMJD2B treatment resulted in reductions in ALP and alizarin red staining in empty vector-transfected cells. Importantly, slience of JMJD2B reduced ALP and alizarin red staining in the presence of osteogenic medium, suggesting that slience of JMJD2B abrogates the facilitate effects of osteogenic medium on osteogenic differentiation and mineralization of hBMSCs. According to the outcome of staining experiments based on the overexpression of JMJD2B in hBMSCs, that are contrary to silence effect further verify the key role of JMJD2B in the induction of osteogenic differentiation. 10. 7717/peerj. 13862/fig-2 Figure 2 JMJD2B regulates the methylation enrichment status of H3K9me2 in RUNX2 promoter to influence the osteogenic differentiation of hBMSCs. hBMSCs were transfected with empty or JMJD2B-overexpressing/silencing vectors and incubated in osteogenic medium or H3K9me2 inhibitor BIX01294/DMSO for 3, 7 or 14 days hBMSCs were transfected with empty or JMJD2B-overexpressing/silencing vectors and cultured in osteogenic medium (A) hBMSCs were treated for 14 days and the Runx2 promoter was ChIP-ed with anti-H3K9me2 or IgG control. Data are expressed as the mean ±SD. ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001, ∗∗∗∗ P < 0. 0001 vs. LV5-JMJD2B. (B) Alkaline phosphatase (ALP) staining and alizarin red staining were conducted to examine osteogenesis and mineralization, respectively. (C) Quantitative real-time PCR (qRT-PCR) was performed at 3 d after treatment to measure mRNA expression of Runx2 and JMJD2B genes. GAPDH was used as an internal reference. Data are expressed as the mean ±standard deviation (SD). ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001. All experiments were independently repeated at least three times. (D–E) Western blot analysis was conducted to measure protein expression of JMJD2B, Runx2, H3K9me2 protein levels. β -actin was used as an internal reference. (F) Quantitative real-time PCR (qRT-PCR) was performed at 3 d after treatment to measure mRNA expression of Runx2 and JMJD2B genes. GAPDH was used as an internal reference. Data are expressed as the mean ±standard deviation (SD). ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001. All experiments were independently repeated at least three times. Then, we measured the mRNA levels of osteogenic markers in hBMSCS cells in response to JMJD2B overexpression/slience. As shown in Fig. 2C, compared with Scr/NC, JMJD2B overexpression dramatically promoted mRNA expression of osteogenic markers, including Runx2, Ocn, Opn, and Osx, whereas JMJD2B slience generally showed counter effects on mRNA expression of these genes ( Fig. 1C, upper and middle panels), regardless of the presence of osteogenic medium. Collectively, the combined results of Figs. 1 and 2 suggest that histone H3K9 Demethylase JMJD2B might promotes osteogenic differentiation of hBMSCs by regulating the expression of Runx2, Ocn, Opn, Osx. JMJD2B regulates the methylation enrichment status of H3K9me2 in RUNX2 promoter to influence the osteogenic differentiation of hBMSCs Considering the critical role of JMJD2B in osteogenesis and the essential role of histone modification in hBMSCs differentiation regulation, we examined the effect of JMJD2B overexpression/slience on the methylation enrichment status of H3K9me2 in the Runx2 promoter in hBMSCs exposed to osteogenic medium, we wanted to determine whether the altered transcriptional activity of JMJD2B may be attributed to differences in DNA binding. To do this we compared the chromatin association of H3K9me2 with the Runx2 promoter using Chromatin Immunoprecipitation (ChIP). LV5-JMJD2B significantly increased H3K9me2 occupancy at the Runx2 promoter. However, Si-JMJD2B decreased H3K9me2 recruitment to Runx2 promoter ( Fig. 3A ). These results strongly suggest that LV5-JMJD2B has a different activity of co-regulator recruitment than Si-JMJD2B. The ChIP results revealed that binding levels of H3K9me2 at the promoter regions of Runx2 were significantly increased when JMJD2B overexpressed, whereas they were reduced at the promoter region while slienced ( Fig. 3A ), regardless of the presence of osteogenic medium. 10. 7717/peerj. 13862/fig-3 Figure 3 JMJD2B Is Required for Osteogenic Differentiation of hBMSCs. hBMSCs were transfected with empty or JMJD2B-overexpressing/silencing vectors and incubated in osteogenic medium for 0, 1, 3, 7 or 14 days (A) Quantitative real-time PCR (qRT-PCR) was performed at 72 h after treatment to measure mRNA expression of JMJD2B as well as osteogenesis-related genes. GAPDH was used as an internal reference. Data are expressed as the mean ±standard deviation (SD). ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001 vs. 0 M. All experiments were independently repeated at least three times. ALP staining and alizarin red staining were performed to examine osteogenesis and mineralization, respectively (B) Western blot analysis was conducted to measure protein expression of JMJD2B protein levels. β -actin was used as an internal reference. (C) Alkaline phosphatase (ALP) staining and alizarin red staining were conducted to examine osteogenesis and mineralization, respectively. (D) Quantitative real-time PCR (qRT-PCR) was performed at 3 d after treatment to measure mRNA expression of osteogenesis- related genes. GAPDH was used as an internal reference. Data are expressed as the mean ±standard deviation (SD). ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001. All experiments were independently repeated at least three times. As shown in Fig. 3B, compared with control, H3K9me2 inhibitor BIX01294 impaired the intensities of ALP and alizarin red staining, in the presence of osteogenic medium. Osteogenic medium treatment resulted in enhancement in ALP and alizarin red staining in control cells. Importantly, inhibition of H3K9me2 reduced ALP and alizarin red staining in the presence of osteogenic medium, suggesting that inhibition of H3K9me2 abrogates the promotion effects of osteogenic medium on osteogenic differentiation and mineralization of hBMSCs. To further investigate whether JMJD2B activates H3K9me2 binding levels, we measured the mRNA levels of Runx2 and JMJD2B in hBMSCs cells in response to H3K9me2 inhibition. As shown in Fig. 3C, compared with control, H3K9me2 inhibition dramatically reduced mRNA expression of Runx2 and JMJD2B. Next, we determined the protein level in hBMSCs cells inhibiting H3K9me2. As shown in Fig. 3D, compared with control, inhibition of H3K9me2 markedly cut down the protein levels of Runx2 and H3K9me2 but showed no significantly effects on JMJD2B in the presence of osteogenic medium. To examine whether compensatory expression of JMJD2B could reverse the effect of BIX01294 on these proteins, we overexpressed JMJD2B in hBMSCs and then treated the cells with DMSO medium dilution solution or BIX01294. Treatment with LV5-JMJD2B significantly overexpressed JMJD2B and elevated protein levels of Runx2 and JMJD2B in the absence or presence of BIX01294. Of note, inhibition of H3K9me2 significantly reduced Runx2 and H3K9me2 protein levels but does not affect the compensatory increase of JMJD2B overexpressing JMJD2B ( Fig. 3E ). qRT-PCR revealed that in general, overexpression of JMJD2B remarkably enhanced mRNA expression of Runx2 and JMJD2B. This enhancement slightly hindered by BIX01294 ( Fig. 3F ). These results suggest that by affecting the methylation enrichment status of H3K9me2 in the Runx2 promoter, JMJD2B may regulate the expression of Runx2 and mediate the differentiation of hBMSCs. Discussion DNA is the essence of heredity. This view has almost become a genetic truth. The rise of epigenetics has changed this phenomenon to a certain extent. People are surprised to find that the emergence of new traits is not necessarily the result of changes in DNA sequences. The environment may induce changes in biological traits, and this change can also Inherited to the next generation. Epigenetic refers to genetic changes in gene expression without altering the DNA sequence, including DNA methylation, histone modification and small non-coding RNA related regulation ( Wu & Sun, 2006 ). Lysine methylation is one of the most prominent histone post-translational modifications that regulate chromatin structure ( Berry & Janknecht, 2013 ). All four members of the JMJD2 family (also known as KDM4) have the ability to demethylate H3K9me2/3 and/or H3K36me2/3. The JMJD2 family belongs to a group of proteins containing the Jmj domain. JMJD2 family members are involved in the regulation of stem cell self-renewal and differentiation ( Mak et al. , 2021 ). Of note, JMJD2B also participates in the osteogenic differentiation of mesenchymal stem cells by mediating the demethylation of H3K9me3 on the DLX promoter ( Ye et al. , 2012 ). Chromatin immunoprecipitation (ChIP) coupled with PCR (ChIP-PCR) also identified a STAT-binding site in Runx2 promoter region containing a transcriptional repressor trimethylated histone 3 lysine 9 (H3K9me3), which can be demethylated by JMJD2B ( Kurozumi et al. , 2019 ). These findings suggest that histone H3K9 Demethylase JMJD2B might promotes osteogenic differentiation of bone marrow-derived mesenchymal stem cells by regulating H3K9me2 on Runx2, JMJD2B possibly playing an important role in bone formation and resorption. To verify the effect of JMJD2B on expression and osteogenic response, we treated hBMSCs with osteogenic medium. We found that mRNA/protein/fluorescence signal expression of JMJD2B was significantly upregulated in response to osteogenic medium stimulation, along with significantly facilitated/suppressed osteogenic differentiation in the following gain/loss of function study, suggesting that osteogenic medium promotes osteogenesis possibly by upregulating JMJD2B-mediated histone demethylation of osteogenesis-related chromatin. The canonical transcription factor Runx2 facilitates bone formation. Of note, JMJD2B-mediated histone demethylation is essential for Runx2 expression. Thus, we hypothesized that osteogenic medium promotes osteogenesis possibly through upregulating JMJD2B-mediated histone demethylation of H3K9me3/H3K9me2 in Runx2 promoter and that inhibitory expression of H3K9me2 might suppress the effects of osteogenic medium. Indeed, our results showed that overexpression/silence of JMJD2B increased/reduced the enrichment status of H3K9me2 in Runx2 promoter, in addition, inhibitory expression of H3K9me2 reversed the promoted effects of osteogenic medium on osteogenic differentiation as well as mRNA/protein expression of Runx2, which could be compensated by overexpression of JMJD2B. Together, these events facilitate osteogenic differentiation of hBMSCs, as evidenced by enhanced ALP and alizarin red staining and mRNA expression of osteogenic markers in hBMSCs. JMJD2B-mediated histone demethylation plays an essential role in the activation of Runx2 expression. Dysregulation of JMJD2B blocks the activation of Runx2 expression. Conclusions In conclusion, osteoinduction promotes JMJD2B expression in hBMSCs. Overexpression of JMJD2B reversed the inhibitory effects of negative inhibition of silent lentivirus on osteogenic differentiation as well as mRNA expression of osteogenic markers in hBMSCs. JMJD2B regulates the methylation enrichment status of H3K9me2 in the Runx2 promoter to influence the osteogenic differentiation and Runx2 expression in hBMSCs. These findings suggest that overexpression of JMJD2B is as a potential therapeutic approach to regulate the osteogenic differentiation of hBMSCs through activating histone demethylation of H3K9me3/H3K9me2 in the Runx2 promoter. Supplemental Information 10. 7717/peerj. 13862/supp-1 Supplemental Information 1 Raw data applied for data analyses and preparation for Fig. 1B for the time period of 0, 1, 3, 7, 14D Click here for additional data file. 10. 7717/peerj. 13862/supp-2 Supplemental Information 2 Raw data applied for data analyses and preparation for Fig. 2B Click here for additional data file. 10. 7717/peerj. 13862/supp-3 Supplemental Information 3 Raw data applied for data analyses and preparation for Fig. 3D Click here for additional data file. 10. 7717/peerj. 13862/supp-4 Supplemental Information 4 Raw data applied for data analyses and preparation for Fig. 3E Click here for additional data file. 10. 7717/peerj. 13862/supp-5 Supplemental Information 5 Raw data exported from the microscope CKX41 applied for data analyses and preparation for Fig. 1C FAM210a for the time period of 0, 1, 3, 7, 14D Click here for additional data file. 10. 7717/peerj. 13862/supp-6 Supplemental Information 6 Raw data exported from the microscope CKX41 applied for data analyses and preparation for Fig. 1C JMJD2B for the time period of 0, 1, 3, 7, 14D Click here for additional data file. 10. 7717/peerj. 13862/supp-7 Supplemental Information 7 Histograms and raw data showing mRNA fold expression and grayscale value of RUNX2, JMJD2B, FAM210A for Fig. 1 Click here for additional data file. 10. 7717/peerj. 13862/supp-8 Supplemental Information 8 Histograms and raw data showing mRNA fold expression and grayscale value of JMJD2B and other osteogenesis-related genes for the time period of 72 h for Fig. 2 Click here for additional data file. 10. 7717/peerj. 13862/supp-9 Supplemental Information 9 Raw data exported from the microscope applied for data analyses and preparation for Fig. 2C left panel Alkaline phosphatase (ALP) staining for the time period of 1, 3, 7, 14D Click here for additional data file. 10. 7717/peerj. 13862/supp-10 Supplemental Information 10 Raw data exported from the microscope applied for data analyses and preparation for Fig. 2C left panel Alizarin red staining for the time period of 1, 3, 7, 14D Click here for additional data file. 10. 7717/peerj. 13862/supp-11 Supplemental Information 11 Raw data exported from the microscope applied for data analyses and preparation for Fig. 3B left panel Alkaline phosphatase (ALP) staining for the time period of 7D Click here for additional data file. 10. 7717/peerj. 13862/supp-12 Supplemental Information 12 Raw data exported from the microscope applied for data analyses and preparation for Fig. 3B left panel Alizarin red staining for the time period of 14D Click here for additional data file. 10. 7717/peerj. 13862/supp-13 Supplemental Information 13 Histograms and raw data showing CHIP-qPCR, mRNA fold expression and grayscale value of Fig. 3 Click here for additional data file. 10. 7717/peerj. 13862/supp-14 Supplemental Information 14 Statistical Reporting Click here for additional data file.

10. 7717/peerj. 14550

PeerJ

Epigenetic regulation of dental-derived stem cells and their application in pulp and periodontal regeneration

Dental-derived stem cells have excellent proliferation ability and multi-directional differentiation potential, making them an important research target in tissue engineering. An increasing number of dental-derived stem cells have been discovered recently, including dental pulp stem cells (DPSCs), stem cells from exfoliated deciduous teeth (SHEDs), stem cells from apical papilla (SCAPs), dental follicle precursor cells (DFPCs), and periodontal ligament stem cells (PDLSCs). These stem cells have significant application prospects in tissue regeneration because they are found in an abundance of sources, and they have good biocompatibility and are highly effective. The biological functions of dental-derived stem cells are regulated in many ways. Epigenetic regulation means changing the expression level and function of a gene without changing its sequence. Epigenetic regulation is involved in many biological processes, such as embryonic development, bone homeostasis, and the fate of stem cells. Existing studies have shown that dental-derived stem cells are also regulated by epigenetic modifications. Pulp and periodontal regeneration refers to the practice of replacing damaged pulp and periodontal tissue and restoring the tissue structure and function under normal physiological conditions. This treatment has better therapeutic effects than traditional treatments. This article reviews the recent research on the mechanism of epigenetic regulation of dental-derived stem cells, and the core issues surrounding the practical application and future use of pulp and periodontal regeneration.

Introduction Dental-derived stem cells are obtained from dental papilla or dental follicles and can be isolated and cultured in teeth or periodontal soft tissue. Dental pulp stem cells (DPSCs), stem cells from exfoliated deciduous teeth (DFPCs), and stem cells from apical papilla (SCAPs) are all derived from dental papillae. Dental follicle precursor cells (DFPCs) and periodontal ligament stem cells (PDLSCs) are derived from dental follicles. In recent years, the discovery of dental-derived stem cells, their abundant sources, and their safety and effectiveness have won them increasing attention in the field of tissue regeneration. Epigenetic regulation refers to the regulation of gene expression without changing the DNA sequence. This plays an important role in the self-renewal and differentiation capacity of adult and embryonic stem cells ( Chen et al. , 2017 ). Epigenetic regulation exists widely in natural organisms and participates in many biological processes, such as embryogenesis, germ cell formation, hematopoietic stem cell differentiation, and tumor formation ( Canovas et al. , 2017 ; Chen, Yan & Duan, 2016 ; Mohammad, Barbash & Creasy, 2019 ). More importantly, epigenetic modifications can not only regulate tooth formation, development, and aging, but also affect the differentiation of dental-derived stem cells ( Hodjat, Khan & Saadat, 2020 ; Liu et al. , 2021 ; Townsend et al. , 2009 ). Epigenetic modifications are advantageous because they do not cause permanent DNA damage, off-target effects, or deleterious mutations. Therefore, an increasing number of studies have focused on the role of epigenetic modifications in regulating the proliferation and differentiation of dental-derived stem cells. This article focuses on the epigenetic regulation of dental-derived stem cells, describes the application of epigenetic regulation based on dental-derived stem cells in dental pulp and periodontal regeneration, summarizes the shortcomings of the existing research, and proposes possible future research directions. Why this review is needed and who it is intended for Endodontic and periodontal diseases are common and frequently-occurring diseases of the oral cavity. Traditional treatments for endodontics include root canal therapy and apical surgery. These treatments provide good relief of symptoms but cannot avoid tooth discoloration and pulp inactivation and necrosis. Periodontitis can cause destruction of the alveolar bone and also tooth loss. Traditional periodontal treatment may meet the needs of most patients with periodontitis, but fail to achieve periodontal tissue regeneration. In recent years, with the development of tissue engineering and the discovery of dental-derived stem cells, pulp and periodontal regeneration have become a potential treatment for these two diseases. The discovery of the mechanism of epigenetic modification of dental-derived stem cells has led to the potential use of epigenetic regulation in dental pulp and periodontal regeneration. However, no article targeting the same subject has been published to provide a clear guidance for clinic trials. This article aims to provide perspectives for dentists and researchers investigating dental pulp and periodontal regeneration using epigenetic regulated dental-derived stem cells, in order to reduce failures and improve the prognosis for patients. It is our hope that these techniques can develop more secure and effective treatment approaches. Survey Methodology To ensure an inclusive and unbiased analysis of the literature and to accomplish the review’s objectives, we searched the following literature databases: PubMed, Science Direct, Research Gate, and Google Scholar. The search terms included: dental-derived stem cells, epigenetic regulation, dental pulp regeneration, periodontal regeneration, DPSCs, SHEDs, SCAPs, DFPCs, PDLSCs, the targets were searched together with Boolean operators such as “AND” and “OR”. It is important to note that the keywords used and their variants and relevant words could be classified into categories and any combination of words from different categories was used for the search. The categories we used were as follows: 1. About dental-derived stem cells: dental-derived stem cells; DPSCs; SHEDs; SCAPs; DFPCs; PDLSCs; odontogenic stem cells. 2. About epigenetic regulation: epigenetic regulation; epigenetics; DNA methylation; histone modifications; noncoding RNA. 3. About pulp and periodontal regeneration: pulp regeneration; periodontal regeneration; odontogenic differentiation; nerve regeneration; angiogenesis; osteogenic differentiation. This article is based on published literature. The aspects of the inclusion criteria are the retrieval keywords, information not covered by previous literature, and most importantly the use of a clear and credible source. Properties of dental-derived stem cells Stem cells have significant proliferative capacity and multi-directional differentiation potential. The development of regenerative medicine has led to the use of stem cells in the repair of damaged cells, tissues, and organs, which have low self-healing abilities, with excellent safety and efficacy. Dental-derived stem cells mediate the process of tooth regeneration by upregulating odontogenic and angiogenic capacity in the form of secreted exosomes (Exo) ( Mai et al. , 2021 ). Dental-derived stem cells have been used in relevant clinical trials in the fields of periodontal tissue, maxillofacial bone tissue repair, and apical pulp disease treatment ( Feng et al. , 2010 ; Giuliani et al. , 2013 ; Nakashima & Iohara, 2014 ). However, the preclinical models for the use of stem cells in nerve regeneration, diabetes, and autoimmune diseases have only been preliminarily validated ( Kanafi et al. , 2013 ; Mead et al. , 2017 ; Shimojima et al. , 2016 ). Properties and potential of DPSCs Dental pulp stem cells have a relatively high proliferation rate, a low cellular senescence, multi-directional differentiation potential, and immunomodulatory properties ( Ma et al. , 2019 ; Mortada & Mortada, 2018 ). Damage to dental pulp causes dental pulp stem cells to induce the formation of various cellular components, including odontoblasts, to replenish damaged cells. In addition, in vitro studies have shown that DPSCs can differentiate into neural-like cells, osteoblasts, chondrocytes, adipocytes, muscle cells, endothelial cells, hepatocytes, and renal pericytes, etc. ( Barros et al. , 2015 ; Gandia et al. , 2008 ; Saito et al. , 2015 ). DPSCs can effectively promote pulp and periodontal regeneration. Guo et al. (2021) combined decellularized tooth matrix (DTM) with human dental pulp stem cells and successfully achieved the regeneration of dental pulp and periodontal tissue. Properties and potential of SHEDs SHEDs are isolated from the pulp tissue of exfoliated deciduous teeth, whose expression level of osteocalcin and alkaline phosphatase activity are higher compared to DPSCs ( Koyama et al. , 2009 ). SHEDs can express osteocalcin and RUNX-2 markers, resulting in the differentiation potential of osteoblasts and odontoblasts ( Miura et al. , 2003 ). SHEDs can induce the migration of naive bone marrow mesenchymal stromal cells (BMSCs) by secreting extracellular vesicles (EVs) containing various cytokines, thereby promoting the bone healing process ( Luo, Avery & Waddington, 2021 ). SHEDs can also induce pulp and periodontal regeneration. Yang et al. (2019) combined SHEDs cell sheets and DFSCs cell sheets with dentin matrix (TDM) and implanted them into the orthotopic jawbone of nude mice. The results indicated that SHED/TDM successfully achieved periodontal tissue regeneration with better migration ability and neurogenic differentiation potential ( Yang et al. , 2019 ). Properties and potential of SCAPs SCAPs exist in the apex of the developing tooth before tooth eruption and differentiate to odontoblasts, which mainly secrete apical dentin. SCAPs are less resistant to immune cell-mediated toxicity compared with other dental-derived stem cells, but can induce high levels of pro-inflammatory cytokine secretion ( Whiting et al. , 2018 ). Since SCAPs are derived from developing odontogenic tissues, they are more widely used in the field of tissue regenerative. For example, in dental pulp engineering, the regeneration process of dental pulp can be achieved by inducing endogenous stem cells to move to the regeneration site ( Rombouts et al. , 2017 ). Wei, Sun & Hou (2021) successfully used the silk fibroin-RGD-stem cell factor scaffold (the RGD peptide was arginine-glycine-aspartic acid polypeptide) to promote the migration and proliferation of SCAPs. This approach is promising for the further use of cell homing in dental pulp regeneration. Properties and potential of DFPCs The dental follicle contains a large number of undifferentiated precursor cells. In 2005, DFPCs were first isolated from the dental follicles of human third molars ( Morsczeck et al. , 2005 ). DFPCs are derived from neural crest and are direct precursor cells of periodontal tissue ( Zhang et al. , 2019a ). DFPCs can promote pulp regeneration through the paracrine pathway. Hong et al. (2020) found that DFPCs could effectively enhance the proliferation, migration, and odontogenic differentiation of inflammatory dental follicle cells (DPCs) in vitro and their ectopic dentinogenesis in vivo. Properties and potential of PDLSCs Periodontitis often leads to the destruction of periodontal tissue and may even lead to tooth loss. After root canal treatment, the periodontal tissue is the only source of nutrition for the root canals. Multiple preclinical studies have demonstrated the effectiveness of PDLSCs to restore damaged periodontal tissue in periodontal regenerative therapy ( Li et al. , 2020a ). PDLSCs are mainly isolated from the periodontal tissue of permanent teeth and can be further differentiated into osteoblasts, chondrocytes, and adipocytes under appropriate conditions ( Deng et al. , 2018a ). Studies have shown that PDLSCs have the strongest osteogenic ability, followed by DPSCs, and DFPCs ( Qu et al. , 2021 ). Epigenetic regulations of dental-derived stem cells Epigenetics refers to changing the expression level and function of a gene without changing its sequence. Its regulatory processes are regulated by signaling molecules whose interactions with neighboring cells induce appropriate transcriptional and epigenetic responses ( Surani, Hayashi & Hajkova, 2007 ). The way in which epigenetic mechanisms regulate gene expression related to environmental factors plays an important role in the development of various diseases, such as tumors and inflammation ( Yuan, Dong & Shen, 2022 ; Zarzour, Kim & Weintraub, 2019 ). In addition, epigenetic regulation may also affect the tooth number, size, and shape ( Fernández et al. , 2020 ). The common epigenetic regulations include DNA methylation, histone modification, and non-coding RNA regulations. Tables 1 – 3 summarize the regulatory mechanisms and potential applications of DNA methylation, histone modifications, and ncRNAs in dental-derived stem cells. 10. 7717/peerj. 14550/table-1 Table 1 Regulation of DNA methylation in dental-derived stem cells. Modification Stem cell Locus Pathway mechanism Target protein Potential applications Ref DNA methylation DPSCs p16 PSAT1 provides reduced SAM and decreased p16 methylation PSAT1, PHGDH Improve the pulp regeneration potential of aging DPSCs Yang et al. (2021) DNA demethylation DPSCs Wnt WNT-3A activates Wnt signaling by diminishing their 5mC content. NNMT Induce epigenetic remodeling and pulp regeneration potential of DPSCs Uribe-Etxebarria et al. (2020) DNA demethylation DPSCs OSX, DLX5, RUNX2 5-Aza-CdR induced the expression of OSX, DLX5 and RUNX2 by decreasing DNA methylation. DSPP, DMP1 Promote the odontogenic growth and differentiation of DPSCs Zhang et al. (2015) DNA methylation DPSCs KDM6B Alcohol suppressed KDM6B through dysregulating DNA methylation ALP, BMP2, BMP4, DLX2, OCN, OPN Promote osteogenic and odontogenic growth of dental mesenchymal stem cells Hoang et al. (2016) DNA methylation PDLSCs MIR31HG Mechanical force downregulates MIR31HG through DNA methylation. IL-6 Inhibite hPDLSCs proliferation Han et al. (2021) DNA demethylation PDLSCs CALAL POSTN attenuated the AGE-induced CALAL methylation RUNX2, OSX, OPEN, RANGE inhibit the osteogenic differentiation of PDLSCs Wang et al. (2022) DNA methylation PDLSCs DKK-1 Downregulation of Tet1 and Tet2 leads to hypermethylation of DKK-1 promoter, activating WNT pathway. FasL Promote immunomodulation of PDLSCs Yu et al. (2019) DNA methylation PDLSCs TNFR-1 High-glucose upregulates TNFR-1 via CpG island hypomethylation. TNFR-1 protein Aggravate viability reduction in hPDLSCs Luo et al. (2020) DNA methylation SHEDs IGF2 IGF2 was induced via DNA methylation and RXR/RAR pathways activation. RUNX2, ALP, BGLAP, DLX5 Promote osteogenic differentiation of SHED Fanganiello et al. (2015) DNA demethylation DFSCs HOXA2 HOTAIRM1 induced HOXA2 via DNA hypomethylation DSPP, DMP1 Induce osteogenic differentiation of DFSCs Chen et al. (2020b) Notes. DPSCs dental pulp stem cells PDLSCs periodontal ligament stem cells SHEDs stem cells from exfoliated deciduous teeth DFSCs dental follicle stem cells PSAT1 phosphoserine aminotransferase 1 PHGDH phosphoglycerate 5mC 5methyl-cytosine NNMT Nicotinamide-N-methyltransferase 5-Aza-CdR 5-Aza-20-de-oxycytidine kinase 1 TNFR1 tumor necrosis factor-alpha receptor-1 RXR/RAR Retinoid X Receptor/ Retinoic Acid Receptor DSPP dentin sialophosphoprotein DMP1 dentin matrix protein 1 KDM6B lysine (K)-specific demethylase 6B IL-6 interleukin- 6 ALP alkaline phosphatase 10. 7717/peerj. 14550/table-2 Table 2 Regulation of histone modifications in dental-derived stem cells. Modification Stem cell Locus Pathway mechanism Target protein Potential applications Ref Histone deacetylation DPSCs P21 IGFBP7 activated SIRT1, resulting in a deacetylation of H3K36ac and reduction of p21 transcription SIRT1 deacetylase Prevent DPSCs senescence and promote tissue regeneration Li et al. (2022) Histone acetylation DPSCs HAT/KAT8 WNT-3A activates Wnt by induces HAT expression and increased H3AC. ACLY Reduce the ability of DPSCs to differentiate into osteoblasts Uribe-Etxebarria et al. (2020) Histone acetylation and methylation DPSCs WNT3A, DVL3 Ferutinin regulates Wnt/ β -catenin pathway by H3K9 acetylation and H3K4 trimethylation. Osteocalcin, collagen 1A1 Direct DPSCs towards the osteogenic lineage Rolph et al. (2020) Histone demethylation DPSCs BMP2, HOX KDM6B catalyzes the demethylation of H3K27me3 and activates BMP2 and HOX ALP, BMP2, BMP4, DLX2, OCN, OPN Promote osteogenic and odontogenic growth of dental mesenchymal stem cells Hoang et al. (2016) Histone demethylation DFSCs Wnt Downregulated MEG3/EZH2 activated Wnt/ β -catenin signaling pathway via demethylation on H3K27 β -catenin and Wnt5a protein Promotes the osteogenesis of DPSCs and DFPCs Deng et al. (2018b) Histone methylation DFSCs PTH1R CHD7 activates PTH/PTH1R signaling pathway and interaction with H3K4me RUNX2, SP7, BGLAP, DLX5, BMP2, COL1A1 Promote osteogenic differentiation of DFSCs Liu et al. (2020a) Histone methylation DFSCs SFRP1 WAY-316606 inhibits SFRP1 via histone H3K4me3 and activates Wnt pathway β -catenin, RUNX2, ALP, osteocalcin, collagen Maintain the nonmineralized state of PDL Progenitors Gopinathan et al. (2019) Histone methylation PDLSCs COL1A1, RUNX2, IL-1 β, CCL5 LPS downregulated COL1A1, COL3A1, RUNX2 by H3K4me3 and upregulated CCL5, DEFA4, IL-1 β gene expression by H3K27me3 COL1A1, COL3A1, RUNX2, CCL5, DEFA4, IL-1 β Regulate periodontal lineage differentiation and the coordination of the periodontal inflammatory response Francis et al. (2019) Histone methylation PDLSCs RUNX2, MSX2, DLX5 The H3K4me3 active methyl mark globally switch to the H3K27me3 repressive mark under osteogenic induction conditions. DSPP, DMP1 Induce osteogenic differentiation of DFSCs Francis et al. (2020) Histone demethylation PDLSCs IGFBP5 BCOR inhibits IGFBP5 through histone K27 methylation. ALP Promote the odontoblast differentiation, proliferation, migration and mineralization of PDLSCs Han et al. (2017) Histone methylation SCAPs p15 INK4B. p27 Kip1 KDM2A increased H3K4 trimethylation at loci p15 and p27 cyclin-CDK Inhibite cell proliferation of SCAPs Gao et al. (2013) Notes. DPSCs dental pulp stem cells DFSCs dental follicle stem cells PDLSCs periodontal ligament stem cells SCAPs stem cells from apical papilla H3AC acetylated-Histone 3 H3K4me3 he histone H3 methylated at lysine 4 H3K27me3 the histone H3 methylated at lysine 27 H3K9me3 the histone H3 methylated at lysine 9 ACLY ATP-citrate lyase enzyme BMP2 bone morphogenic protein 2 HDAC3 histone deacetylase 3 MEG3 maternally expressed 3 EZH2 the enhancer of zeste homolog 2 PTH1R parathyroid hormone receptor-1 PCNA Proliferating cell nuclear antigen CHD7 Chromodomain helicase DNA-binding protein 7 α -SMA alpha-smooth muscle actin TNNT2 cardiac muscle troponin T ACTC1 cardiac muscle IGFBP5 insulin-like growth factor binding protein 5 BCOR BCL6 co-repressor p15INK4B cyclin-dependent kinase inhibitor 2B p27Kip1 cyclin-dependent kinase inhibitor 1B 10. 7717/peerj. 14550/table-3 Table 3 The regulatory role of ncRNAs in dental-derived stem cells. Modification Stem cell Locus Pathway mechanism Target protein Potential applications Ref miRNA DPSCs TLR-4 LPS activates lipopolysaccharide/TLR-4 signaling pathway by downregulating miR-140-5p. TLR-4 Enhance differentiation of DPSCs and inhibit proliferation Sun et al. (2017a) miRNA DPSCs Rac1 miR-224-5p targets the 3′-untranslated region of Rac1 gene and downregulates Rac1. MAPK8, caspase-3, caspase-9, Fas ligand Potect DPSCs from apoptosis Qiao et al. (2020) miRNA DPSCs TGFBR1 miR-24-3p and LEF1-AS1 sponged to regulate TGFBR1 expression. RUNX2, OSX, ALP Promote osteogenic differentiation of DPSCs Wu, Lian & Sun (2020) miRNA DPSCs CAB39 miR-34a-3p activates AMPK/mTOR signaling pathway by downregulating CAB39 AMPK, mTOR Downregulate alleviates senescence in DPSCs Zhang et al. (2021) miRNA DPSCs Foxq1 miR-320b mediated Foxq1 upregulation after calcium hydroxide stimulation. cyclin E1, cyclin D1 Pomote proliferation of DPSCs Tu et al. (2018) miRNA DPSCs TLR4 lncRNA-Ankrd26 promotes migration and osteogenesis via regulating miR-150-TLR4 signaling in MSCs OSX, ALP Promote dental pulp repair Li & Ge (2022) miRNA DFSCs Runx2, ALP and SPARC miR-204 negatively targets the gene of Runx2, ALP and SPARC. Runx2, ALP and SPARC Promote osteogenic induction in DFSCs Ito et al. (2020) miRNA PDLSCs IL-17, IL-35 Overexpression of miRNA-146a downregulates IL-17 and IL-35 expression under periodontitis IL-17, IL-35 Inhibit proliferation of hPDLSCs Zhao, Cheng & Kim (2019) miRNA PDLSCs PTEN miR-181b-5p regulates PTEN/AKT pathway and promotes BMP2/ Runx2 PKB, BMP2, Runx2 Promote hPDLSCs proliferation and osteogenic differentiation Lv et al. (2020) miRNA PDLSCs Satb2 miR-31 promotes Satb2 siRNA and inhibits osteogenic differentiation Runx2 Promote osteogenic differentiation of PDLSCs Zhen et al. (2017) miRNA PDLSCs Spry1 Upregulating miR-21 repressing Spry1 and inhibits TNF- α Spry1; TNF- α Suppress adipogenic and osteogenic differentiation of PDLSCs Yang et al. (2017) miRNA PDLSCs Notch2 miR-758 regulated Notch2 and interacts with lncRNA-ANCR Notch2 Regulate the osteogenic differentiation of PDLSCs Peng et al. (2018) circRNA DPSCs SATB2, RUNX2, OCN Exosome circLPAR1 induced osteogenic differentiation via downregulation of hsa-miR-31. SATB2, RUNX2, and OCN Promote osteogenic differentiation of DPSCs Xie et al. (2020) circRNA DPSCs RUNX1, Beclin1 hsa_circ_0026827 promotes osteoblast differentiation via Beclin1 and the RUNX1 signaling pathways by sponging miR-188-3p Beclin-1, RUNX1, ALP, OCN and OSX Promote osteogenic differentiation of DPSCs Ji et al. (2020) circRNA SCAPs ALPL CircSIPA1L1 is sponge for miR-204-5p, which upregulates ALPL. ALPL Promote the osteogenic differentiation of SCAPs Li et al. (2020b) circRNA PDLSCs SMAD5 circFAT1 inhibits miR-4781-3p targeting SMAD5. SMAD5 Mediate the periodontal bone regeneration of PDLSCs Ye et al. (2021) circRNA PDLSCs ERK CDR1as functioned as an miR-7 sponge to activate the ERK signal pathway. ERK Inhibites the proliferation of PDLSCs Wang et al. (2019b) lncRNA SHEDs BMP2 lncRNA C21orf121 competes with BMP2 binding to miR-140-5p, upregulates BMP2 expression. BMP2, Nestin, β III-tubulin, MAP2, NSE Promote neurogenic differentiation of SHEDs Liu et al. (2018a) lncRNA SCAPs ALP, RUNX2 LncRNA-H19 bound to miR-141, elevating phosphorylated levels of p38 and JNK. SPAG9 Promote the odontoblast differentiation of SCAPs Li et al. (2019b) Notes. DPSCs dental pulp stem cells PDLSCs periodontal ligament stem cells SHEDs stem cells from exfoliated deciduous teeth DFSCs dental follicle stem cells SCAPs stem cells from apical papilla TLR-4 toll-like receptor 4 Rac1 the Rac family small GTPase 1 RUNX1 runt-related transcription factor 1 CAB39 calcium-binding protein 39 AMPK AMP-activated protein kinase mTOR mammalian target of rapamycin BMP2 bone morphogenetic proteins 2 MAP2 microtubule-associated protein 2 MAPK8 mitogen-activated protein kinase 8 NSE neuron-specific enolase ALPL alkaline phosphatase SPARC secreted protein acidic and rich in cysteine ZEB2 zinc finger E-box binding homeobox 2 SMAD5 a receptor-regulated SMAD protein in SMAD family member CDR1as circRNA CDR1as TNF- α Tumor necrosis factor-alpha PHD2 prolyl hydroxylase domain–containing protein 2 LY294402 small interfering RNA for AKT AKT a phosphoinositide 3 kinase (PI3K)-dependent serine/threonine DNA methylation DNA methylation is one of the important epigenetic modifications and it is common in most eukaryotes ( Lin et al. , 2018 ). The formation of 5-methylcytosine (5mC) from cytosine-phosphorothioate-guanine (CpG) dinucleotides by DNA methyltransferase (DNMT) leads to the silencing of gene expression ( Radhakrishnan, Kabekkodu & Satyamoorthy, 2011 ). The level of methylated CpG is regulated by DNMT and DNA demethylase ten-eleven translocation (TET) ( Ren, Gao & Song, 2018 ; Li et al. , 2015 ). DNA methylation levels are associated with stemness and the differentiation potential of dental-derived stem cells. For instance, DPSCs exhibits low DNA methylation levels and repressive mark H3K9Me2 enrichment, which is mediated by increased DNMT3B and G9a expression, respectively. This leads to decreased AKT phosphorylation and promotes osteogenesis ( Shen et al. , 2019 ). The decreased expression of the serine metabolism-related enzyme phosphoserine aminotransferase 1 (PSAT1) provides less methyl donor S-adenosylmethionine (SAM) for the methylation of the aging marker p16 (CDNK2A), resulting in the reduced stemness and osteogenic differentiation capacity of DPSCs ( Yang et al. , 2021 ). Wnt can effectively regulate the epigenetic mechanism of DPSCs. The short-term activation of Wnt signaling by Wnt-3A can cause a decrease in the content of 5-methylcytosine (5-mC) in DPSCs, which reduces the ability of DPSCs to differentiate into osteoblasts ( Uribe-Etxebarria et al. , 2020 ). Different odontogenic stem cell genes have varied methylation levels and differentiation potentials. In DPSCs, PDLSCs, and DFPCs, the methylation of genes CD109 and SMAD3 are significantly different. At the transcriptional level, PDLSCs showed significantly higher expression levels of CD109, SMAD3, ALP, and RUNX2, which were identical to the differences in their DNA methylation profiles. The transcription levels of osteogenic differentiation-related factors and their osteogenic differentiation potential are higher in PDLSCs ( Ai et al. , 2018 ). The osteogenic differentiation process can be altered by modulating the methylation levels of specific genes in PDLSCs. Advanced glycation end-products (AGE) can increase the expression of DNMT1 and inhibit the methylation activation of calcitonin-related polypeptide α (CALCA) promoter, which inhibits the osteogenic differentiation of PDLSCs ( Wang et al. , 2022 ). Additionally, periostin (POSTN) can reduce the level of AGE receptors and DNA methylation of the CALCA promoter, thereby attenuating the inhibitory effect of AGE induction ( Wang et al. , 2022 ). Histone modifications Histones are located in the nucleus of eukaryotic cells and can form nucleosomes, the basic structure of chromatin, when bound to DNA. Modifications of amino acid residues in histone tails can cause structural changes in histones, which provide sites that can be recognized by specific proteins ( Strahl & Allis, 2000 ). The regulation of specific genes can also be achieved through the binding of specific proteins to sites. Many studies have been conducted on the modification of histones, specifically on methylation and acetylation. Histone methylation Histone methylation refers to the transfer of the methyl group of S-adenosylmethionine (SAM) to arginine or lysine site under the action of histone methyltransferases (HMTs) ( Wang & Jia, 2009 ). The expression or repression of genes is associated with specific residues catalyzed by HMTs. For example, histone H3-lysine 4 (H3K4) methylation promotes gene expression, while H3K9 and H3K27 methylation inhibit gene expression ( Blanc & Richard, 2017 ). However, histone demethylases can cause histone demethylation. For example, histone demethylase lysine (K)-specific demethylase 1A (KDM1A) targeting H3K4 and H3K9 can affect the differentiation of embryonic stem cells ( Pedersen & Helin, 2010 ). Histone modifications play key roles in the lineage commitment and differentiation of DFPCs and DPSCs. The H3K27me3 mark in DFSCs can strongly suppress the expression of two dentinogenic genes, dentin sialophosphoprotein (DSPP) and dentin matrix protein 1 (DMP1), whereas the H3K27me3 mark is almost absent in the promoters of the genes DSPP and DMP1 in DPSCs and the gene expression levels are significantly higher ( Francis et al. , 2020 ; Gopinathan et al. , 2013 ). The histone methylation-modifying enzyme enhancer of zeste homolog 2 (EZH2) mainly acts on H3K27 and regulates the osteogenic differentiation of DPSCs and DFPCs through the Wnt/ β -catenin signaling pathway. The reduction of EZH2 directly causes the downregulation of H3K27me3 and further leads to the accumulation of β -catenin, which activates the Wnt/ β -canonical signaling pathway and ultimately promotes the osteogenesis of DPSCs and DFPCs ( Deng et al. , 2018b ; Li et al. , 2018a ). Histone demethylases such as KDM6B, KDM1A, and KDM2A also play a regulatory role in the gene expression of dental-derived stem cells. KDM6B catalyzes the demethylation of histone H3K27me3 located near the promoter of bone morphogenetic protein-2 (BMP2). This activates BMP2 expression and promotes osteogenic and odontogenic growth of dental mesenchymal stem cells ( Hoang et al. , 2016 ; Liu et al. , 2022 ). In addition, KDM6B decreases the level of histone K27 methylation in the promoter of insulin-like growth factor binding protein 5 (IGFBP5), thereby promoting the odontoblast differentiation, proliferation, migration and mineralization of PDLSCs ( Han et al. , 2017 ). KDM1A can cooperate with 2-oxoglutarate 5-dioxygenase 2 (PLOD2) to regulate the differentiation process of SACPs ( Wang et al. , 2018a ). The knockdown of KDM1A or PLOD2 reduces ALP activity, promotes the expression of DSPP, DMP1 and RUNX2, and enhances bone/dentin production in SCAPs ( Wang et al. , 2018a ). Homeobox C8 (homeobox, HOXC8) significantly inhibits the osteogenic differentiation ability of SCAPs by directly binding to the KDM1A promoter and enhancing its transcription ( Yang et al. , 2020b ). KDM2A is able to increase histone H3 lysine 4 (H3K4) trimethylation at the p15 INK4B (cyclin-dependent kinase inhibitor 2B) and p27 Kip1 (cyclin-dependent kinase inhibitor 1B) loci ( Gao et al. , 2013 ). On the other hand, the attenuation of KDM2A prevents cell cycle progression in the G1/S phase of SCAPs ( Gao et al. , 2013 ). Inflammation and hypoxia can also cause the upregulation of KDM2A expression and repress the secreted frizzled-related protein 2 (SFRP2) transcription by reducing histone methylation in the SFRP2 promoter ( Yang et al. , 2020a ). SFRP2 can inhibit the Wnt/ β -catenin signaling pathway and further inhibit the target genes of the nuclear factor kappa B (NF-kB) signaling pathway. This enhances the bone/odontogenic differentiation capacity of SCAPs ( Yang et al. , 2020a ). Similarly, histone demethylase KDM3B is also capable of regulating the bone/dental differentiation, cell proliferation, and migratory potential of SCAPs ( Zhang et al. , 2020 ). Histone acetylation Histone acetylation is mainly related to histone acetyltransferases (HATs) and histone deacetylases (HDACs). Under the catalysis of HDACs, the acetyl group of acetyl-CoA is transferred to the amino acid residues of histone tails and promotes gene transcription ( Galvani & Thiriet, 2015 ). In addition, histone deacetylases cause chromatin condensation by deacetylating amino acids in histone tails, thereby repressing gene transcription ( Meier & Brehm, 2014 ). Histone acetylation regulates the stemness and differentiation process of dental-derived stem cells. For instance, histone acetyltransferases such as p300, general control non-arrestin 5 (GCN5), and lysine acetyltransferase 6B (KAT6B, also known as MORF) regulate the stemness or osteogenic differentiation of cells by modifying histones on target genes of DPSCs and PDLSCs. Among them, p300 can regulate the expression of genes DMP-1, DSPP, DSP, NANOG, and SOX2 in different ways. p300 promotes the odontogenic differentiation of DPSCs by catalyzing acetylation and promoting the expression of the histone, H3K9, within the promoter regions of DMP-1, DSPP, and DSP ( Liu et al. , 2015b ). Furthermore, MORF and GCN5 are mainly involved in the osteogenic differentiation process of PDLSCs under inflammatory conditions, among which GCN5 regulates DKK1 expression through the acetylation of the H3K9 and H3K14 promoter regions ( Li et al. , 2016 ). DKK1 can inhibit the Wnt/ β -catenin pathway and promote the osteogenic differentiation of PDLSCs. Chronic periodontal inflammation reduces the expression of MORF in PDLSCs ( Xue et al. , 2016 ). Methoxy-parvacrol (osthole) upregulates MORF in PDLSCs and catalyzes the acetylation of H3K9 and H3K14, which promotes the osteogenic differentiation of PDLSCs under inflammatory conditions ( Sun et al. , 2017b ). In addition, silencing HDAC expression or using histone deacetylase inhibitor (HDACi) can regulate gene expression by inhibiting HDAC activity. The inhibition of HDAC1, HDAC3, and HDAC6 expression can contribute to the odontogenic differentiation of DPSCs. The HDAC inhibitor MS-275 can act on HDAC1 and HDAC3, causing the up-regulation of the gene expression of odontogenic differentiation-related proteins in DPSC, including RUNX2, DMP1, ALP, and DSPP ( Lee et al. , 2020 ). Similarly, silencing HDAC6 can induce the expression of odontogenic marker genes such as OSX, OCN, and OPN in DPSCs, while inhibiting osteoclast differentiation ( Wang et al. , 2018c ). In addition, HDAC6 is also involved in the development and differentiation of PDLSCs. For instance, HDAC6 participates in the aging process of PDLSCs by regulating the acetylation of p27 Kip1. The inhibition of HDAC6 promotes senescence in PDLSCs and attenuates their osteogenic differentiation and migration abilities ( Li et al. , 2017 ). HDAC9, which is mainly involved in the osteogenic differentiation of PDLSCs under inflammatory conditions, impairs the osteogenic differentiation capacity of PDLSCs, whereas miR-17 induces osteogenic differentiation by inhibiting HDAC9 ( Li et al. , 2018b ). Finally, HDACi can regulate the differentiation process of dental-derived stem cells by inhibiting HDAC. Luo et al. (2018) found that HDACi, trichostatin A, and valproic acid could enhance the acetylation of histones H3 and H4, to promote the proliferation, migration, and adhesion of DPSCs. Noncoding RNA Non-coding RNAs (ncRNA) are transcripts with no or low coding potential, including ribosomal RNA (rRNA), transfer RNA (tRNA), and microRNA (miRNA) ( Ren & Wang, 2021 ). miRNA is the most studied in the field of epigenetics ( Ren & Wang, 2021 ). miRNAs directly interact with partially complementary target sites located in the 3′ untranslated region of target mRNAs and repress their expression ( Hombach & Kretz, 2016 ). Endogenous competing RNAs (ceRNAs) mainly regulate gene expression by competitively binding to miRNA ( Qi et al. , 2015 ). ceRNAs typically include long noncoding RNAs (lncRNAs) and circular RNAs (circRNAs). miRNA miRNAs regulate dental-derived stem cells by affecting the expression of related genes in RUNX2, BMP, Wnt, MAPK, and Notch1 signaling pathways. miRNAs regulate the differentiation process of SCAPs, DPSCs, and PDLSCs by affecting the expression of the gene RUNX2, which mainly mediates osteogenic/odontogenic differentiation in stem cells ( Hussain, Tebyaniyan & Khayatan, 2022 ). miR-450a-5p and miR-28-5p can affect the expression of signal transducer and activator of transcription 1 (STAT1), which is mainly involved in the negative regulation of RUNX2 ( Dernowsek et al. , 2017 ). An in vitro model system study found that STAT1 mRNA was gradually down-regulated and RUNX2 mRNA was gradually up-regulated as SHEDs differentiated into osteoblasts ( Dernowsek et al. , 2017 ). Similarly, miR-218 regulated the mineralization and differentiation process of DPSCs through the ERK1/2 pathway. ERK1/2 signaling converge at Runx2 to control the differentiation of DPSCs ( Chang et al. , 2019 ). The transforming growth factor beta (TGF- β )/ BMP signaling pathway plays an important role in the odontogenic/osteogenic differentiation of dental-derived stem cells. miR-132 inhibits the growth differentiation factor 5 (GDF5) of the TGF- β family and activates the NF- κ B axis, which attenuates the osteogenic differentiation ability of PDLSCs ( Xu et al. , 2019 ). CD105 is a co-receptor for the type I transmembrane glycoprotein and TGF β -1, which is associated with the osteogenic differentiation of cells. Ishiy et al. (2018) compared the mineralization degree of the SHED matrix with the low/high expression of CD105 and found that the high expression of CD105 reduced osteogenic potential, while miR-1287 was negatively correlated with CD105. miRNA can affect cell differentiation by regulating the expression of the Smad gene, which is an essential transcription factor in the TGF- β /BMP signaling pathway. miR-135b can inhibit the expression of the Smad4 and Smad5 genes, which hinder the odontoblast-like differentiation of dental pulp cells ( Song et al. , 2017 ). In PDLSCs, miR-23a acts on the bone morphogenetic protein receptor type 1B (BMPR1B) gene and inhibits the phosphorylation of Smad1/5/9, which attenuates the osteogenic differentiation of PDLSCs ( Zhang et al. , 2019b ). The Smad ubiquitination regulator (Smurf) regulates TGF- β /BMP signaling through ubiquitination, causing the degradation of signaling molecules and preventing the overactivation of TGF- β /BMP signaling ( Kushioka et al. , 2020 ). In SCAPs, miR-497-5p promotes bone/odontogenic differentiation by targeting SMAD-specific E3 ubiquitin protein ligase 2 (Smurf2) and regulating the Smad signaling pathway ( Liu et al. , 2020b ). Furthermore, the expression of miR-26a can be upregulated in the exosomes secreted from SHED, and miR-26a can improve angiogenesis in SHED by regulating TGF- β /SMAD2/3 signaling ( Wu et al. , 2021 ). Wnt/ β -catenin signaling can regulate the proliferation, development, and cell fate aspects of dental-derived stem cells. The overexpression of miR-140-5p represses the Wnt1 gene, which affects Wnt/ β -catenin signaling and ultimately inhibits the odontoblast differentiation of DPSCs ( Lu et al. , 2019 ). Chromodomain helicase DNA-binding protein 8 (CHD8) plays an essential role in maintaining the active transcription of nerve-specific genes and can be targeted and regulated by miR-221 ( Wen et al. , 2020 ). For example, in SHED, upregulated miR-221 activates the Wnt/ β -catenin pathway by inhibiting CHD8, which promotes the neurogenic differentiation of cells ( Wen et al. , 2020 ). In addition, both p38-mitogen-activated protein kinase (MAPK) and neurogenic locus Notch homolog 1 (Notch1) signaling pathways are involved in the osteogenic/odontogenic differentiation process of dental-derived stem cells. miR-143-5p can regulate the expression of MAPK pathway-related genes in DPSCs. To be specific, the downregulation of miR-143-5p increased the expression of p38 MAPK signaling pathway-related genes such as MAPK14 and MKK3/6, and odontoblast differentiation markers such as ALP and OCN ( Wang et al. , 2019a ). IGF-I can enhance the odontogenic/osteogenic differentiation ability of mesenchymal stem cells (MSCs) by activating the MAPK pathway, while the IGFBPs/IGF-I complex is regulated by matrix metalloproteinase 1(MMP1) ( Wang et al. , 2018b ). In SCAPs, miRNA let-7b inhibits bone/odontogenic differentiation of SCAP by targeting MMP1 ( Wang et al. , 2018b ). Notch1 is a transmembrane receptor, and the downregulation of Notch signaling inhibits self-renewal of DPSCs and induces their differentiation ( Wang et al. , 2011 ). miR-146a-5p can inhibit the expression of Notch1 and regulate the osteogenic/odontogenic differentiation process of DPSCs ( Qiu et al. , 2019 ). ceRNA lncRNA. lncRNA can regulate the differentiation of dental-derived stem cells by directly acting on GDF5, distal-less homeobox 3 (DLX3), and Kruppel-like factor 2 (KLF2). lncRNA growth arrest specific transcript 5 (GAS5) can enhance the expression of GDF5 in cells and promote the phosphorylation of the p38 MAPK/JNK signaling pathway, which enhances the osteogenic differentiation of PDLSCs ( Yang et al. , 2020c ). lncRNA H19 inhibits DNMT3B-mediated methylation of the DLX3 gene through S-adenosyl-L-homocysteine hydrolase (SAHH), which regulates odontoblast differentiation of DPSCs ( Zeng et al. , 2018a ). The direct interaction of lncRNA SNHG1 with EZH2 regulates KLF2 promoter H3K27me3 methylation and inhibits the differentiation of PDLSCs to osteoblasts ( Li, Guo & Wu, 2020c ). By inhibiting the expression of miRNAs, lncRNAs can also play a regulatory role. During the osteogenic differentiation of PDLSCs, lncRNAs can act as ceRNAs and form networks to regulate the Wnt/ β -catenin signaling pathway ( Lai et al. , 2022 ). lncRNA-ANCR competitively binds miR-758 and inhibits the expression of Notch2, which further affects the Wnt/ β -catenin signaling pathway and inhibits the osteogenic differentiation of PDLSCs ( Peng et al. , 2018 ). FoxO1 promotes bone formation in PDLSCs by competing with TCF-4 for β -catenin and inhibiting the Wnt pathway ( Wang et al. , 2016 ). lncRNA-POIR can inhibit the expression of the miR-182 target gene FoxO1 and affect the osteogenic differentiation process of PDLSCs ( Wang et al. , 2016 ). As ceRNAs, lncRNAs can affect the expression of genes related to the MAPK and BMP signaling pathways. LncRNA-H19 can competitively bind to miR-141 and prevent the miRNA-mediated degradation of SPAG9, thereby increasing the phosphorylation levels of p38 and JNK, which promotes the bone/odontogenic differentiation of SCAPs ( Li et al. , 2019b ). In SHEDs, lncRNA C21 or f121 can compete with BMP2 to bind with miR-140-5p and promote the neurogenic differentiation of SHEDs by upregulating BMP2 expression ( Liu et al. , 2018a ). lncRNA-CCAT1 combined with miR-218, and lncRNA G043225 combined with miR-588 can promote the odontogenic differentiation of DPSCs ( Chen et al. , 2020a ; Zhong et al. , 2019 ). circRNA In PDLSCs, circRNAs can indirectly regulate osteogenic differentiation by binding to miRNAs ( Gu et al. , 2017 ). circRNA cerebellar degeneration-related protein 1 transcript (CDR1as) and miR-7 can regulate the osteogenic differentiation and stemness of PDLSCs. CDR1as may promote the upregulation of GDF5 and the phosphorylation of Smad1/5/8 and p38 MAPK by inhibiting the expression of miR-7, inducing the differentiation of PDLSCs to osteoblasts. In addition, the interaction of CDR1as with miR-7 can also upregulate the expression of KLF4 to maintain the stemness of PDLSCs, while RNA-binding protein hnRNPM regulates its expression in PDLSCs by interacting with CDR1as ( Gu et al. , 2021 ). During the osteogenic differentiation of SCAPs, the expression profiles of circRNAs are significantly altered, and circRNAs mainly function as ceRNAs ( Li et al. , 2019a ). circ SIPA1L1 can promote the expression of the gene ALPL (alkaline phosphatase alkaline phosphatase) by binding to miR-204-5p, which causes the osteogenic differentiation of SCAPs ( Li et al. , 2020b ). Epigenetic regulatory network In the epigenetic regulation of dental-derived stem cells, there are multiple links between histone modifications, DNA methylation, and ncRNA, which interact with each other and participate in genetic regulation together. ncRNAs participate in the regulation of gene expression in stem cells by regulating DNA methylation. For example, lncRNA H19 can inhibit the activity of DNMT3B, which reduces the methylation of the distal-less homeobox (DLX3) of the gene, thereby promoting the odontogenic differentiation of DPSCs ( Zeng et al. , 2018a ). Similarly, miR-675 can also promote the odontogenic differentiation of DPSCs by inhibiting DNMT3B ( Zeng et al. , 2018b ). The lncRNA HOTAIRM1 inhibits the expression and enrichment of DNMT1 on the HOXA2 promoter and mechanically binds to the CpG island in the HOXA2 promoter region, leading to hypomethylation and the induction of HOXA2 and DFSC differentiation into osteoblasts ( Chen et al. , 2020b ). ncRNAs can also play a role in histone modifications, including histone methylation, histone acetylation, and histone deacetylation. miR-153-3p inhibits the transcription of ALP, Runx2, and OPN by targeting KDM6A, which results in the attenuated osteogenic differentiation of PDLSCs ( Jiang & Jia, 2021 ). miRNAs are involved in the aging and differentiation process of dental-derived stem cells by regulating the expression of HAT or HDAC. The upregulation of miR-152 represses HAT sirtuinc 7 (SIRT7) expression and affects the degree of histone acetylation, which accelerates the aging process of DPSCs ( Gu et al. , 2016 ). The upregulation of miRNA-383-5p can promote the down-regulation of the HDAC9 mRNA level, which leads to increased alkaline phosphatase activity, mineral node formation, and the expressions of RUNX2, osteocalcin, and Smad4 in PDLSCs and other osteogenic markers ( Ma & Wu, 2021 ). Similarly, miR-22 can inhibit HDAC6 expression and promote the osteogenic differentiation of PDLSCs ( Yan et al. , 2017 ). Therapeutic application of dental-derived stem cells in dental pulp and periodontal regeneration In 1971, Nygaard-Ostby & Hjortdal (1971) proposed the concept of pulp tissue regeneration. Pulp regeneration refers to the formation of new pulp tissue through tissue engineering to replace the infected or necrotic pulp tissue, thereby restoring the structure and function of the pulp-dentin complex under physiological conditions. Conventional apexogenesis may result in the thinning of the dentin wall and underdevelopment of the root, which greatly increases the risk of long-term root fracture. However, pulp regeneration can effectively form healthy pulp tissue and promote the formation of dentin. Periodontitis, a chronic inflammation of the periodontal tissue caused by dental plaque, can cause the destruction and absorption of the alveolar bone and tooth loss. Traditional periodontal treatments, such as scaling, focuses on controlling the occurrence of inflammation, but fails to restore the structure and function of periodontal tissue entirely. Periodontal tissue regeneration reconstructs periodontal tissue damaged by periodontitis and restores its structure and function by means of tissue engineering ( Chen et al. , 2010 ). The key elements of tissue regeneration are stem cells, scaffolds, and signaling molecules. The biological behavior of stem cells is regulated by epigenetics. Further research will likely lead to the discovery of an increasing number of new factors. These factors may regulate the development of stem cells towards odontogenic differentiation, angiogenesis, neurogenesis, and osteogenic differentiation through epigenetic mechanisms, and may facilitate the application of pulp regeneration and periodontal regeneration. Figure 1 demonstrates the epigenetic regulation of dental-derived stem cell differentiation and its application in pulp regeneration and periodontal regeneration. 10. 7717/peerj. 14550/fig-1 Figure 1 Multilineage potential of human dental-derived stem cells. Four kinds of human dental-derived stem cells have the capacity to differentiate under epigenetic modification into different somatic cell and tissue types, and finally contribute to regeneration of pulp or periodontal tissue. DPSC, Dental pulp stem cell; SCAP, Stem cells from apical papilla; SHED, Stem cells from human exfoliated deciduous teeth; PDLSC, Periodontal ligament stem cell. Odontogenic differentiation Transcription factor RUNX2 mainly mediates osteogenic/odontogenic differentiation and may effectively promote the expression of dentin matrix proteins or induce the transdifferentiation of cells into osteoblasts ( Li et al. , 2011 ). HDACi can affect the expression of the RUNX2 gene in stem cells by acting on HDAC. MS-275 can inhibit the expression of HDAC1 and HDAC3 and may induce the up-regulation of odontogenic related proteins in DPSCs, including RUNX2, DMP1, ALP, and DSPP, which promotes the odontogenic differentiation of DPSCs ( Lee et al. , 2020 ). Sultana et al. (2021) showed that without the induction of mineralized medium, MS-275 alone could increase the expression levels of BMP2, DMP1, DSPP, and Runx2 mRNA of mouse odontoblast-like cell line MDPC-23, and improved ALP activity. Therefore, MS-275 can effectively promote the odontogenic differentiation of DPSCs. BMPs signal through canonical Smad and non-Smad signaling pathways, in which BMP-Smad signaling can be involved in the formation of coronal dentin ( Omi et al. , 2020 ). The HDACi inhibitor TSA can significantly upregulate the levels of Smad and NFI-C in DPSCs by inhibiting HDAC3. Jin et al. (2013) treated DPSCs with TSA and found that the expression of BSP, DMP1, and DSPP was significantly increased compared with the control group; the level of Smad2/3 was also significantly up-regulated 21 days after mineralization induction. In contrast, neonatal mice that were maternally exposed to TSA exhibited thicker dentin and more dentin cells in their postpartum molars, with a greater ability to secrete DSP ( Jin et al. , 2013 ). In addition to regulating gene expression in DPSCs, Duncan et al. (2017) demonstrated that TSA could also promote the release of dentin matrix components from dentin. These two studies show that HDACi can promote the differentiation of DPSCs into odontoblasts as well as the release of the dentin matrix, which is very beneficial to the repair of dental pulp-dentin complex. The Wnt/ β -catenin signaling pathway may regulate the process of dentin formation and tooth development and the amplification of Wnt signaling can significantly improve the survival rate of damaged dental pulp cells and promote tertiary dentin formation ( Hunter et al. , 2015 ). miR-140-5p can repress the Wnt1 gene and affect the Wnt/ β -catenin signaling process ( Lu et al. , 2019 ). Lu et al. (2019) collected impacted third molars from patients aged 14–22 years and divided the extracted DPSCs into an miR-140-5p inhibition group, a negative control group (NC), and a blank control group. After 14 days of inducing cells to differentiate into odontoblasts, Alizarin Red S staining showed that the mineralized matrix deposition was greatest in the inhibitor group and least in the mock group. Western blotting showed that the inhibitor group had the highest expressions of DSPP and DMP-1 proteins while the mock group had the lowest ( Lu et al. , 2019 ). These results indicate that miR-140-5p can affect the odontoblast differentiation process of DPSCs. The p38 MAPK pathway is central to the transcriptional control of odontoblasts and its activation is critical for apical morphogenesis and enamel secretion ( Greenblatt et al. , 2015 ). The activation of the MAPK signaling pathway is also associated with the osteogenic/odontogenic differentiation of DPSCs ( Wu et al. , 2019 ). lncRNA-H19 can competitively bind to miR-141 and upregulate the phosphorylation levels of p38 and JNK. Li et al. (2019b) induced transfected SCAPs in an osteoblast differentiation medium, and the Western blot results showed that the protein expressions of OCN, RUNX2, ALP, and DSP in the H19-infected SCAP group were significantly higher than those in the control group. The SCAP that had a stable expression of H19, and the control group, were further loaded on Bio-Oss collagen scaffolds and implanted in the subcutaneous tissue of nude mice. H&E and Masson staining showed that the abundance of bone-like structures, collagen deposition, and dentin-like structures in the H19-infected SCAP group was higher than that in the control group ( Li et al. , 2019b ). This indicates that lncRNA-H19 can promote the odontoblast differentiation process of SCAPs by activating the p38 and JNK signaling pathways. Nerve regeneration CHD8 can affect neural progenitor cells and neurons, and also plays a role in maintaining the active transcription of neural-specific genes ( Wilkinson et al. , 2015 ). Meanwhile, CHD8 can also alter neurogenesis and cortical development by regulating the Wnt/ β -catenin signaling pathway ( Platt et al. , 2017 ). In SHEDs, upregulated miR-221 can bind to CHD8 and activate the Wnt/ β -catenin pathway ( Wen et al. , 2020 ). Wen et al. (2020) divided the SHEDs in the third-generation logarithmic growth phase into six groups: blank group, NC group (transfected with miR-221 negative sequence), miR-221 mimic group (transfected with miR-221 mimic), miR-221 inhibitor group (transfected with miR-221 inhibitor), siRNA-CHD8 group (transfected with siRNA into CHD8 vector), and miR-221 inhibitor + siRNA-CHD8 group (co-transfected with miR-221 inhibitor and siRNA-CHD8). The results of Western blot analysis revealed that the expressions of NSE, NESTIN, MAP-2, NF-M, and TH in the miR-221 inhibitor group were significantly lower than those in the NC group, while the miR-221 mimic group and siRNA-CHD8 group were both lower than those in the NC group. Immunofluorescence examination showed that the expressions of NSE and MAP-2 in the miR-221 inhibitor + siRNA-CHD8 group were higher than that in the miR-221 inhibitor group ( Wen et al. , 2020 ). Among them, neuron-specific enolase (NSE) was a highly specific marker of neurons and peripheral neuroendocrine cells, NESTIN was a key early neural progenitor cell marker, NF-M and microtubule-associated protein 2 (MAP2) was a neuron-associated marker, and TH was the rate-limiting enzyme in dopamine neurotransmitter biosynthesis. These results suggest that miR-221 can promote SHED differentiation into neurons by inhibiting CHD8. BMP2 is a neurotrophic factor that induces the growth of brain dopaminergic (DA) neurons in vitro and in vivo, whose induction depends on the Smad signaling pathway ( Hegarty, Sullivan & O’Keeffe, 2013 ). In SHEDs, lncRNA C21 or f121 competitively binds to miR-140-5p and upregulates BMP2 expression ( Liu et al. , 2018a ). The results of bioinformatics analysis conducted by Liu et al. (2018a) showed that there was a targeting relationship between the second spliceosome of lncRNA C21 or f121 and miR-140-5p, the same as miR-140-5p and BMP2. This suggested that lncRNA C21 or f121 competed with BMP2 to bind to miR-140-5p. Liu et al. (2018a) grouped and experimented with SHEDs in the third-generation logarithmic growth phase. The results showed that the protein expressions of both Nestin and β III-tubulin decreased, but increased in the transfected miR-140-5p inhibitor group compared with the NC group (transfected with lncRNA C21 or f121 negative sequence), the BMP2 and MAP2 in the si-C21 or f121 group, the miR-140-5p group, and the si-C21 or f121+miR-140-5p group. Further experiments showed that the up-regulation of lncRNA C21 or f121 or down-regulation of miR-140-5P increased the frequency of social behavior in rats and decreased the cumulative time of repetitive stereotyped movements in young rats ( Liu et al. , 2018a ). All of the abovementioned studies show that lncRNA C21 or f121 can effectively promote the neurogenic differentiation of SHEDs. Finally, some HDACi have the effect of inducing neurogenic differentiation of cells. For instance, Okubo et al. (2016) found that the total number of mRNAs of mature neuronal markers, neurofilament medium polypeptide (NeFM), and microtubule-associated protein 2 (MAP2) significantly increased to approximately 80% in VPA-treated rats compared with untreated rats ( Okubo et al. , 2016 ). Other studies have shown that the neurite number on the cells increased and branched processes were elongated after treating MSCs with combinations of MS-275 or NaB (a kind of HDACi). The cells were visualized by immunofluorescence staining of the neuronal markers ( Jang et al. , 2019 ). The above studies demonstrate the role of HDACi in inducing neurogenic differentiation. Further research may reveal whether dental-derived stem cells can be inducted to differentiate into neural cells. Angiogenesis TGF- β /SMAD2 signaling can promote angiogenesis and the secretion of vascular endothelial growth factor ( Ji et al. , 2014 ). Wu et al. (2021) discovered that the expression of miR-26a was up-regulated in SHED-secreted exosomes (SA-Exo), and miR-26a could promote the expression of TGF- β /SMAD2/3 signaling. The expression of angiogenesis-related proteins (VEGF, angiopoietin 2 and PDGF) of SHEDs was up-regulated, and the endothelial differentiation potential was increased after being treated with SA-Exo. SA-Exo treatment also increased the expression levels of angiogenesis-related proteins in HUVECs. Wu et al. (2021) implanted SHED aggregates into immunodeficient mice and performed histological analysis, which revealed the formation of a new, continuous dentin layer and blood vessels, and the regeneration of the dentin-pulp complex. In addition, dentin and blood vessel formation were enhanced by the combined implantation of SHED aggregates and SA-Exo. The expression level of the angiogenic marker CD31 was also higher. The inhibition of SA-Exo repressed dentin-pulp complex regeneration; however, supplementation with exogenous SA-Exo could restores this process. The results of qRT-PCR confirmed that the expression of miR-26a was significantly increased in SA-Exo, and the inhibition of miR-26a in SA-Exo could not cause the endothelial differentiation of SHED and HUVECs. Western blot analysis revealed that the overexpression of miR-26a upregulates TGF- β /SMAD2/3 signaling, and the inhibition of these two pathways led to reduced endothelial differentiation in SHEDs and HUVECs ( Wu et al. , 2021 ). These studies confirm that miR-26a in SA-Exo promote angiogenesis in SHEDs through the TGF- β /SMAD2/3 signaling pathway. ncRNAs have a strong ability to regulate endothelial cell migration, proliferation, and differentiation. miR-30a-3p targets the epigenetic factor methyl-CpG-binding protein 2 (MeCP2), and the overexpression of MeCP2 damages important genes involved in the regulation of endothelial function such as sirtuin1 ( Volkmann et al. , 2013 ). Volkmann et al. (2013) transfected endothelial cells with miR-30a-3p precursors, which significantly reduced MeCP2 protein levels and increased the migratory ability of endothelial cells. This suggests that miR-30a-3p has the ability to regulate endothelial cells. In addition, lncRNAs also have a role in regulating endothelial cells. Neumann et al. (2018) discovered that lncRNA GATA6 could inhibit the action of the epigenetic regulator, LOXL2, reduce the endothelial-mesenchymal transition in vitro, and promote the formation of blood vessels in mice. These two studies demonstrated the potential of ncRNAs in promoting angiogenesis. Future studies may reveal whether they can regulate dental-derived stem cells for angiogenesis. Osteogenic differentiation Different dental-derived stem cells have a diverse range of DNA methylation levels and unique osteogenic differentiation potentials. Compared with DPSCs and DFPCs, PDLSCs have lower methylation levels of genes related to osteogenesis, higher expression levels of factors such as SMAD3, ALP, OCN, and RUNX2, and a higher osteogenic differentiation potential ( Ai et al. , 2018 ). After culturing PDLSCs, DFPCs, and DPSCs14 in osteoinductive medium, Ai et al. (2018) used Alizarin Red S positive staining and found that the relative intensity of staining in PDLSCs was significantly higher than that in DPFCs and DPSCs. The simultaneous subcutaneous transplantation of cell deposits mixed with hydroxyapatite onto the dorsal surface of immunocompromised male mice found that PDLSCs formed more osteoid. This study demonstrated that DNA methylation can regulate the osteogenic differentiation potential of dental-derived stem cells by affecting the expression of related genes. Insulin-like growth factor (IGF) and its binding proteins play an important role in promoting bone formation ( Nguyen et al. , 2013 ). KDM6B can promote IGFBP5 transcription by reducing histone K27 methylation ( Han et al. , 2017 ). Han et al. (2017) found that by administering a local injection of rhIGFBP5 into the periodontitis area of a piglet model, they could significantly promote the regeneration of periodontal tissues such as alveolar bone and gingiva after 12 weeks. The Wnt/ β -catenin pathway can promote/inhibit the osteogenic differentiation of cells under various conditions ( Wagner et al. , 2011 ). HAT GCN5 inhibits the Wnt/ β -catenin signaling pathway by increasing the levels of H3K9ac and H3K14ac in the DKK1 promoter region ( Li et al. , 2016 ). Li et al. (2016) found that more active osteogenic differentiation was presented in cell populations with higher GCN5 expression, and GCN5 downregulation may lead to defective osteogenic differentiation of PDLSCs. GCN5 knockdown resulted in increased expression of β -catenin and decreased expression of genes and proteins related to osteogenic differentiation, such as RUNX2 and ALP. ChIP assays indicated that GCN5 binds to the promoter region of DKK1. Alveolar bone loss in the first and second maxillary molars was significantly reduced with increased GCN5 expression in periodontital rats ( Li et al. , 2016 ). Therefore, HAT GCN5 can promote the osteogenic differentiation of PDLSCs to regenerate alveolar bone by inhibiting the Wnt/ β -catenin signaling pathway. Both the MAPK and TGF- β /Smad signaling pathways may be involved in BMP-mediated osteogenesis ( Kim, Park & Choung, 2018 ; Zhu et al. , 2018 ). CDR1as is an inhibitor of miR-7 that can cause the upregulation of TGF- β family member GDF5, and the phosphorylation of p38 MAPK ( Li et al. , 2018c ). In PDLSCs, the knockdown of CDR1as or the overexpression of miR-7 significantly suppressed the mRNA and protein levels of GDF5. In contrast, a lower expression of GDF5 resulted in a decrease in the osteogenic markers ALP and RUNX2, as well as phosphorylated p38 MAPK. Li et al. (2018c) loaded CDR1as siRNA and negative control siRNA-treated PDLSCs onto scaffold material and implanted this into the calvarial defect area of nude mice. The results showed that the CDR1as knockdown group had less bone formation and significantly lower new bone formation than the control group. In the control group, bone tissue was generated at the edges of the bone defects, but in the CDR1as knockout group, little new bone was observed ( Li et al. , 2018c ). This study strongly demonstrates that CDR1as can promote the osteogenic differentiation of PDLSCs. Conclusions Dental-derived stem cells, as mesenchymal stem cells, play an essential role in pulp and periodontal regeneration. Epigenetic regulation can adjust gene expression independent of DNA sequence changes, which can affect the proliferation, differentiation, and function of dental-derived stem cells. DNA methylation, histone modifications, and ncRNAs constitute a grand epigenetic regulatory network that can function independently or coherently. Pulp regeneration and periodontal regeneration can be well achieved through epigenetic regulation. Considering the various types of epigenetic modifications and different mechanisms, the epigenetic research on dental-derived stem cells is still lacking at this stage. Current research has focused on classical epigenetic modifications and modification sites, while some potential modifications such as DNA 6mA modification, mRNA m6A modification, and modification on tRNA still need more experimental verifications. Epigenetic modifications are widespread in eukaryotes; some modification mechanisms in mesenchymal stem cells can be further studied in dental-derived stem cells. As new functions of epigenetic modification are revealed, we can also focus on their regulatory roles in dental-derived stem cells. Current research has focused on the regulatory role of specific epigenetic modification mechanisms in dental-derived stem cells. However, less attention has been paid as to whether there are interactions between epigenetic modifications, which limits our further exploration of epigenetic regulatory networks. Studies on the epigenetic regulation of dental-derived stem cells have also been influenced by the cells themselves. The special odontogenic potential of DPSCs and the excellent osteogenic potential of PDLSCs make them the best choice for pulp regeneration and periodontal regeneration. Therefore, the epigenetic mechanisms of DPSCs and PDLSCs are currently the most studied. SCAPs and SHEDs, which are obtained from tooth roots and exfoliated deciduous teeth respectively, have also been widely studied due to their abundant sources and low immunogenicity. In contrast, the sources of DFPCs are more limited, leading to fewer studies. Finally, the application of the epigenetic regulation of dental-derived stem cells is influenced by its mechanism of action. Tissue engineering can be accomplished through stem cells, scaffolds and signaling molecules, all of which are applicable. However, epigenetic regulation needs to act on specific modification sites, and most of the current experiments are done in the form of virus transfection, greatly limiting the application of epigenetic regulation in tissue regeneration. In addition, dental-derived stem cells interfere with the proliferation and differentiation of surrounding cells by secreting exosomes. Therefore, further research on exosomes will be beneficial to the application of epigenetic regulation in dental-derived stem cells.

10. 7717/peerj. 15164

PeerJ

Metabolic shift and the effect of mitochondrial respiration on the osteogenic differentiation of dental pulp stem cells

Background Metabolism shifts from glycolysis to mitochondrial oxidative phosphorylation are vital during the differentiation of stem cells. Mitochondria have a direct function in differentiation. However, the metabolic shift and the effect of mitochondria in regulating the osteogenic differentiation of human dental pulp stem cells (hDPSCs) remain unclear. Methods Human dental pulp stem cells were collected from five healthy donors. Osteogenic differentiation was induced by osteogenic induction medium. The activities of alkaline phosphatase, hexokinase, pyruvate kinase, and lactate dehydrogenase were analyzed by enzymatic activity kits. The extracellular acidification rate and the mitochondrial oxygen consumption rate were measured. The mRNA levels of COL-1, ALP, TFAM, and NRF1 were analyzed. The protein levels of p-AMPK and AMPK were detected by western blotting. Results Glycolysis decreased after a slight increase, while mitochondrial oxidative phosphorylation continued to increase when cells grew in osteogenic induction medium. Therefore, the metabolism of differentiating cells switched to mitochondrial respiration. Next, inhibiting mitochondrial respiration with carbonyl cyanide-chlorophenylhydrazone, a mitochondrial uncoupler inhibited hDPSCs differentiation with less ALP activity and decreased ALP and COL-1 mRNA expression. Furthermore, mitochondrial uncoupling led to AMPK activation. 5-Aminoimidazole-4-carboxamide ribonucleotide, an AMPK activator, simulated the effect of mitochondrial uncoupling by inhibiting osteogenic differentiation, mitochondrial biogenesis, and mitochondrial morphology. Mitochondrial uncoupling and activation of AMPK depressed mitochondrial oxidative phosphorylation and inhibited differentiation, suggesting that they may serve as regulators to halt osteogenic differentiation from impaired mitochondrial oxidative phosphorylation.

Introduction Human dental pulp stem cells (hDPSCs), a source of adult multipotent stem cells, can differentiate into multiple cell types, including odontoblasts/osteoblasts, chondrocytes, adipocytes, myogenic and neural cells ( Sui et al. , 2020 ; Tsutsui, 2020 ). Due to their easy isolation and great potential in tissue engineering and regenerative medicine, DPSCs are widely used in various fields ( Fernandes et al. , 2020 ). Numerous studies have been dedicated to uncovering the detailed mechanisms involved in their self-renewal ability and multilineage differentiation potential. The capacity of dental pulp stem cells to differentiate is essential for dental pulp repair and dentin rebuilding ( Sui et al. , 2020 ). Energy metabolism plays a critical role in regulating the proliferation, differentiation, and many biological processes of stem cells. The relationship between biological activities and mitochondrial aerobic oxidative phosphorylation has been studied since the first study dated 1957 ( Green, Lester & Ziegler, 1957 ). In stem cells, mitochondria produce energy not only to maintain homeostasis but also for differentiation ( Tsutsui, 2020 ). The metabolism shifts between mitochondrial oxidative phosphorylation (OXPHOS) and glycolysis change along with the cell status and levels of mitochondrial maturation ( Khacho & Slack, 2018 ; Wanet et al. , 2015 ). For instance, glycolysis is predominant in undifferentiated stem cells, while differentiated cells rely more on OXPHOS ( Ly, Lynch & Ryall, 2020 ). In mesenchymal stem cells, the mitochondrial process of OXPHOS is activated during osteogenic differentiation. However, the levels of glycolysis are maintained similarly to those in undifferentiated cells ( Shum et al. , 2016 ). Our previous study demonstrated that at the initial stage of hDPSCs differentiation, OXPHOS and glycolysis were upregulated ( Wang et al. , 2016 ). It should be determined whether mitochondrial and glycolytic changes exist during hDPSCs’ continuous differentiation. Moreover, the dynamic distribution, subcellular content, and structure of mitochondria have been shown to exhibit peculiar functions in the processes of cellular differentiation and reprogramming ( Khacho & Slack, 2017 ). Mitochondrial-mediated shifts accompanied by mitochondrial remodeling and dynamics are vital to neural stem cell differentiation and fate ( Coelho et al. , 2022 ). Other questions that need to be addressed are the role of mitochondria in the differentiation of hDPSCs. Some studies are underway to unveil the bidirectional crosstalk between mitochondria and cell differentiation, such as reactive oxygen species production, energy-sensing pathways, and the hypoxia-inducible factor pathway. Adenosine monophosphate (AMP)–activated protein kinase (AMPK), as a central metabolic sensor sensitive to the AMP/ATP ratio, can be activated by various types of cellular stress ( Aslam & Ladilov, 2022 ; Li et al. , 2013 ). AMPK regulates the fate of stem cells and serves as a critical regulator of differentiation ( Liu et al. , 2021 ; Sun et al. , 2017 ; Yang et al. , 2016 ). Previous reports have shown that AMPK is required for immune cell differentiation and adipogenic differentiation ( Rambold & Pearce, 2018 ; Son et al. , 2019 ; Yang et al. , 2016 ). AMPK influenced the osteogenic differentiation of human dental pulp mesenchymal stem cells ( Pantovic et al. , 2013 ). The study also detected the change in AMPK when OXPHOS was intervened. Metabolism is a vital regulator of stem cells and their fate. Anaerobic glycolysis may be an adaptation to the low oxygen niche for stem cells and maintain their stemness. The shift from glycolysis to OXPHOS is required for stem cells to differentiate. The reverse transition from OXPHOS to glycolysis is required to induce pluripotency from somatic cells ( Zhang, Menzies & Auwerx, 2018 ). Regulating the metabolism of stem cells could be a potential target for tissue engineering or regeneration medicine. This study aimed to investigate the changes in mitochondrial OXPHOS and glycolysis during hDPSCs osteogenic differentiation and their possible regulatory factors. Materials & Methods Reagents and antibodies Hexokinase (HK), pyruvate kinase (PK), lactate dehydrogenase (LDH), and alkaline phosphatase (ALP) enzyme activity kits were obtained from Jiancheng (Nanjing, China). MitoTracker Red CMXRos (M7512) and Alexa Fluor 488 Phalloidin (A12379) were obtained from Invitrogen (Carlsbad, CA, USA). The XF Cell Energy Phenotype Test Kit, XFp extracellular flux analyzer, XFp culture microplates, and bicarbonate-free DMEM were obtained from Seahorse Bioscience Agilent Technologies (North Billerica, MA, USA). Hoechst 33342 (B2261) and carbonyl cyanide 3-chlorophenylhydrazone (CCCP, C2759) were purchased from Sigma–Aldrich (St Louis, MO, USA). The following antibodies were used: CD90 (FITC, mouse anti-human, BioLegend, 328107), CD105 (PerCP/Cy5. 5, mouse anti-human, BioLegend, 323215), CD19 (PE, mouse anti-human, BioLegend, 982402), CD34 (PE, mouse anti-human, BioLegend, 343505), AMPK (Abcam, ab80039), p-AMPK (Abcam, ab133448), anti-GAPDH (Abcam, ab9485, ab8245), goat anti-rabbit IgG (H + L)-HRP (Bio-Rad Laboratories, 1706515), goat anti-mouse IgG (H + L)-HRP (Bio-Rad Laboratories, 1706516) and goat anti-rabbit IgG (H+L) Alexa Fluor 594-conjugated secondary antibody (Invitrogen, R37117). Cell culture and treatment After receiving written informed consent from patients and their parents, human dental pulp tissues were collected from caries- and periodontitis-free third molars extracted for orthodontic purposes from healthy donors ( n = 5). The experimental procedures were conducted according to the Declaration of Helsinki. The protocol was approved by the Institutional Ethics Committee of West China Hospital of Stomatology (WCHSIRB-D-2017-052; WCHSIRB-D-2020-075-R1). After washing with sterile PBS, dental pulp tissues were cut into fragments, digested with 3 mg/mL type I collagenase at 37 °C for 30 min, and placed on 25 cm 2 culture dishes. Later, they were maintained in standard medium, which was low glucose Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) (Gibco, Carlsbad, CA, USA), penicillin (100 units/mL), and streptomycin (100 mg/mL) at 37 °C in 5% CO 2. Cells were sub-cultured when they reached 80% confluence and used at passages 3∼6 ( Hu et al. , 2022 ). Osteogenic differentiation was induced by osteogenic induction medium (OIM), the standard medium supplemented with 10 mM b-glycerophosphate, 50 mg/mL ascorbic acid, and 100 nM dexamethasone (Sigma–Aldrich, St. Louis, MO, USA) for 1, 3, 5, and 7 days ( Chen et al. , 2013 ). All media were replaced every two days. To inhibit OXPHOS activity, cells were stimulated with a mitochondrial uncoupler, carbonyl cyanide 3-chlorophenylhydrazone (CCCP, 2 µM). A final concentration of 2 µM was used for this study according to a previous lethality analysis (LD50) ( Mandal et al. , 2011 ). To activate AMPK, cells were treated with a purine nucleoside analog, 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR, 500 µM). When cells achieved 50%∼60% confluency, AICAR was added to the cell medium at a concentration of 500 µM for 1, 3, 5, and 7 days ( Lee et al. , 2010 ). Flow cytometry analysis Flow cytometry analysis was used to identify stem cell surface markers ( Aydin & Şahin, 2019 ; Ponnaiyan, 2014 ). The primary cells at the third passage were detached from flasks using trypsin and washed twice with PBS. A total of 1 ×10 6 cells in 100 µL of PBS were labeled with 5 µL of antibodies against the surface markers CD90 (1:20), CD105 (1:20), CD19 (1:20), and CD34 (1:20) for 1 h at 4 °C protected from light. One sample without additional antibodies was used as a negative control. After the cells were washed twice and resuspended in 500 µL PBS, the labeled cells were analyzed using a Cytomics FC 500 flow cytometer (Beckman Coulter, Brea, CA, USA) and FlowJo software (Tree Star, Ashland, OR, USA). Alizarin red staining For osteogenic differentiation, 1 ×10 5 hDPSCs per well were seeded in 6-well plates. At 60%∼70% confluence, cells were cultured with OIM for 21 days and then washed with PBS. The cells were then stained with 40 mmol/L alizarin red S staining solution (pH 4. 2, Sigma-Aldrich, St Louis, MO, A5533) for 10 min after being fixed with 4% paraformaldehyde for 20 min at room temperature ( Yin et al. , 2021 ). Enzymatic activity assay For the ALP, PK, and LDH activity assays, 1 ×10 6 cells were resuspended in 0. 5 mL of ice-cold PBS and ultrasonically decomposed 5 times at 300 W for 5 s/time with 30s intervals on ice. For the HK activity assay, 5 ×10 6 cells were resuspended in one mL of extracting solution and ultrasonically decomposed. Then, the supernatant was collected at 8, 000 g/min for 10 min at 4 °C. The supernatant was used for protein quantification by the BCA protein assay (KGP902; keyGEN, Nanjing, China) and subsequent activity assays according to the manufacturer’s instructions. The optical density values were determined at 450 wavelengths by a Varioskan Flash multi-plate reader (Thermo Scientific, Waltham, MA, USA). The mean values of the mean absorbance rates from three wells were calculated. Enzyme activities were calculated from optical density values and normalized by the total protein of each sample. Cell energy phenotype analysis Our previous study showed that hDPSCs initiated differentiation on Day 3 ( Wang et al. , 2016 ). With further differentiation, extracellular flux rates were measured in this study. Cells (4 ×10 3 cells/well) were plated on XFp cell culture microplates and divided into experimental and control groups. The experimental group was cultured in OIM, while the control group was grown in standard medium at 37 °C and 5% CO 2. On Days 1, 3, 5, and 7, the extracellular acidification rate (ECAR, as an indicator of the glycolysis potential) and the mitochondrial oxygen consumption rate (OCR, as a parameter of mitochondrial respiration) were analyzed using the Seahorse XF Cell Energy Phenotype Test Kit on the Seahorse XFp Bioanalyzer. One hour before beginning the measurement, cells were switched to Seahorse bicarbonate-free DMEM, adjusted to pH 7. 4 with NaOH, and subsequently incubated in a non-CO 2 incubator at 37 °C for 1 h. Then, OCR and ECAR tests were performed under baseline and stressed conditions after oligomycin (1 µM, an ATP synthase inhibitor) and carbonylcyanide-p-trifluoro-methoxyphenyl hydrazone (FCCP, 2 µM, a mitochondrial uncoupling agent) were added to the microplates. Extracellular flux rates were analyzed by using Seahorse XF software after normalization to the total protein units, and the ratio of OCR/ECAR was calculated. Real-time quantitative PCR (qPCR) Total RNA was extracted from hDPSCs (4∼5 passages) using TRIzol (Invitrogen). RNA quantification and quality control were done using a 2000c Nanodrop spectrophotometer (Thermo Scientific, Waltham, MA, USA). cDNA was reverse transcribed with the PrimeScript RT Reagent Kit (Takara, Osaka, Japan). A negative control without reverse transcriptase was also implemented. Real-time PCR was performed using a SYBR Green PCR kit (Takara) and a Roche LightCycler480 Real-Time PCR System (Roche, Basel, Switzerland), as described previously ( Takeda et al. , 2008 ). Average crossing threshold (CT) values were calculated from the triplicate cDNA samples. Relative gene expression was calculated as 2 −△△CT ( Livak & Schmittgen, 2001 ). GAPDH was used as the internal reference gene, and levels were relativized to the control group. The sequences of specific primers are shown in Table S1. Fluorescence microscopy Cells were seeded on coverslips at 1 ×10 4 cells/well in 6-well plates and incubated with MitoTracker Red CMXRos (MRC) in FBS-free DMEM at a concentration of 100 nM at 37 °C for 30 min. After washing, cells were fixed in 4% paraformaldehyde for 10 min and then permeabilized with 0. 05% Triton X-100 in PBS for 10 min at room temperature. Then, the cells were incubated with Alexa Fluor 488 phalloidin (1:50) to visualize the cytoskeleton for 1 h at room temperature. Nuclei were stained with Hoechst 33342 (200 µg/mL) for 5 min. For the immunofluorescence experiment, cells were first fixed in 4% paraformaldehyde for 10 min and were permeabilized with 0. 05% Triton X-100 in PBS for 10 min. After washing, cells were blocked with 5% BSA in PBS for 30 min and incubated with primary antibody against p-AMPK ( Khorraminejad-Shirazi et al. , 2020 ) (1:200) overnight at 4 °C. Following further washing, the cells were incubated with Alexa Fluor 594-conjugated secondary antibody (1:1200). Nuclei were also visualized by Hoechst 33342 (200 µg/mL) for 5 min. Fluorescence signals were obtained using an Olympus microscope (Olympus IX73; Olympus, Munster, Germany). Immunoblotting According to the manufacturer’s recommendation, total proteins from hDPSCs (5∼6 passages) were extracted using a total protein extraction kit (KeyGEN BioTECh, Nanjing, China). SDS/PAGE was performed on Bio-Tris 5%–10% gradient polyacrylamide gels. Proteins were transferred to PVDF membranes (Bio-Rad Laboratories, Hercules, CA, USA), and membranes were blocked with 5% BSA or 3% nonfat milk in Tris-buffered saline containing 0. 1% Tween (TBS-T) for 60 min at room temperature, according to the manufacturer’s instructions. The membranes were incubated overnight with the appropriate primary antibodies, AMPK (1:1000), p-AMPK (1:1000; Abcam, Cambridge, UK), and anti-GAPDH (1:1000). Detection was subsequently performed with HRP-conjugated secondary antibodies for 1 h at room temperature, goat anti-rabbit IgG-HRP (1:10, 000) and goat anti-mouse IgG-HRP (1:10, 000). The membranes were scanned using the GelDoc XR+ System (Bio-Rad Laboratories, Hercules, CA, USA). The density of Western blot band signals was monitored using Quantity One software. Statistical analysis Every experiment was independently replicated a minimum of three times. ANOVA and Student’s t test were performed to determine statistical significance between the control and experimental groups using SPSS version 17. 0 software. P value s were considered to be significant when p < 0. 05 (* p < 0. 05, ** p < 0. 01, *** p < 0. 001). Results During osteogenic differentiation, glycolysis function declined after an initial slight increase For the identification of hDPSCs, flow cytometry was used to analyze cell surface marker expression, and the cells were also incubated in OIM for alizarin red staining. The results showed that the mesenchymal stem cell markers CD90 and CD105 were 99. 1% positive and 78. 9% positive in hDPSCs, respectively. The hematopoietic stem cell markers CD19 and CD34 were merely 2. 1% positive and 4. 6% positive in hDPSCs. Alizarin red staining of the cells incubated in OIM was positive ( Fig. S1 ). To assess osteogenic differentiation, the mRNA expression of osteogenic differentiation markers ( ALP and COL-1 ) and ALP activity were detected on Days 0, 1, 3, 5, and 7. By q-PCR, the expression of ALP mRNA increased at 3 days after differentiation induction ( p < 0. 01), and COL-1 started to be upregulated 5 days after treatment ( p < 0. 05) compared with the control ( Figs. 1A and 1B ). ALP activity, a marker of osteogenic differentiation, was also monitored ( Qian et al. , 2015 ). The increase in ALP activity occurred from Day 3 and continued to increase during the process of differentiation (1. 98 ± 0. 1 vs. 5. 63 ± 0. 07; 2. 86 ± 0. 08 vs. 7. 74 ± 0. 07; 2. 29 ± 0. 11 vs. 11. 25 ± 0. 18, p < 0. 001, Fig. 1C ). 10. 7717/peerj. 15164/fig-1 Figure 1 Glycolysis increased first and then decreased during hDPSCs differentiation. Cells were cultured in control and OIM for 0, 1, 3, 5, and 7 days, respectively. (A, C) mRNA expression and activity of ALP were upregulated from day 3. (B) COL-1 was upregulated from day 5. (D) HK, a key glycolytic enzyme, increased at day 3 when cells were initiated to differentiate and decreased from day 5. (E, F) The same trend was detected in PK and LDH. Data were mean ± SD, * p < 0. 05, ** p < 0. 01, *** p < 0. 001. The activity of several enzymes in glycolysis was measured. The activities of HK ( p < 0. 05), PK ( p < 0. 05), and LDH ( p < 0. 01) slightly increased at Day 3 and then decreased to a significantly lower level than that of the control at Day 7 during the differentiation of hDPSCs ( Figs. 1D, 1E, and 1F ). Mitochondrial OXPHOS predominated in the energy production of the differentiating hDPSCs The data showed that mitochondrial transcription Factor A ( TFAM, controlling mtDNA expression and mitochondrial biogenesis) expression was elevated at Days 3, 5, and 7 compared to the control ( Fig. 2A, p < 0. 05, p < 0. 01, p < 0. 01). The expression of TFAM in the OIM was also gradually upregulated beginning on Day 3 ( Fig. 2A, p < 0. 001). However, the expression level of nuclear respiratory Factor 1 ( NRF1, controlling nuclear-encoded respiratory chain component expression) was unchanged compared to that in control cells ( Fig. 2B ). 10. 7717/peerj. 15164/fig-2 Figure 2 Cell energy phenotype shifted to mitochondrial OXPHOS as hDPSCs differentiated. (A) TFAM was upregulated when hDPSCs differentiated. (B) NRF1 was not changed upon differentiation. (C) On days 1, 5, and 7, the OCR and ECAR of cells growing in basal condition were first measured. Then those of stressed condition without Oligomycin and FCCP were assayed to show the max metabolic potential. The ECAR was on the horizontal axis, and the OCR was on the vertical axis. The hollow rhombus was for basal condition, while the solid was for stressed condition. (D) For the differentiating cells, the ratio of basal OCR/ECAR, which was calculated to show the major energy phenotype, increased from day 5. (E) The stressed OCR/ECAR ratio had a similar tendency. Data were mean ± SD, * p < 0. 05, ** p < 0. 01, *** p < 0. 001. Later, the mitochondrial OCR (mitochondrial respiration indicator) and ECAR (glycolysis indicator) under baseline and stress conditions were analyzed by a Seahorse XF Cell Energy Phenotype Test Kit. Two stressor compounds, oligomycin, which causes an increase in glycolysis, and FCCP, which drives OCR rates higher, were used to induce stress conditions. Under basal conditions, consistent with upregulated TFAM mRNA expression, the OCR of differentiating cells was greater than that of the control on Day 3 ( Fig. 2C ). Nevertheless, the ratio of OCR/ECAR was not different ( Fig. 2D ). At Days 5 and 7, ECAR was significantly downregulated, and OCR was significantly increased. Moreover, the cells in OIM exhibited higher OCR/ECAR compared to the control at Days 5 and 7 ( Fig. 2D, Day 5, 1. 03 ± 0. 08 vs. 2. 71 ± 0. 39, p < 0. 01; Day 7, 1. 06 ± 0. 17 vs. 2. 29 ± 0. 09, p < 0. 01). Under stressed metabolic conditions, a similar pattern of specific dynamic changes in mitochondrial OXPHOS and glycolysis was observed ( Fig. 2E ). Mitochondrial uncoupling interfered with hDPSCs’ osteogenic differentiation A mitochondrial uncoupler, CCCP, was used to attenuate mitochondrial respiration ( Kane et al. , 2018 ). The cells were cultured in OIM with or without CCCP. Upon differentiation, both the basal and stressed mitochondrial OCRs were depressed after CCCP treatment compared to cells cultured in OIM ( Fig. 3A ). The ECAR of cells growing in the presence of CCCP showed a decrease compared to the control at Day 3 in the baseline condition (21. 21 ± 0. 9 vs. 13. 71 ± 1. 25, p < 0. 01). At the same time, there was no difference on other days ( Fig. 3B ). With MitoTracker Red CMXRos staining, the mitochondrial morphology was punctate. It failed to form a functional network (yellow arrow, Fig. 3C ). Importantly, this suppression of mitochondrial respiration attenuated the differentiation of hDPSCs, manifested as inhibited ALP activity and restrained ALP and COL-1 mRNA expression, compared to cells growing in the absence of CCCP ( Fig. 3D ). 10. 7717/peerj. 15164/fig-3 Figure 3 As hDPSC differentiation, cell energy phenotype shifted to mitochondrial OXPHOS. Cells were cultured in the OIM with or without CCCP (2 µM) for 1-7 days. (A) The change in OCR was monitored over time. With the CCCP treatment, the cells’ OCR was depressed compared to the control. (B) Except for day 3, the ECAR of the CCCP group had no significant change. (C) On day 3, fluorescence was used to detect the morphology of mitochondria: mitochondria - MitoTracker Red CMXRos (MRC), red; F-actin-Phalloidine, green; nucleus-Hoechst 33342, blue. The local magnification showed CCCP inhibited the form of mitochondria functional network. (D) The osteogenic differential markers, the mRNA expression of ALP and COL-1, and the activity of ALP were suppressed by CCCP. Scale bars:20 µm. Data were mean ± SD, * p < 0. 05, ** p < 0. 01, *** p < 0. 001. The activation of AMPK impaired OXPHOS-driven differentiation The activation of AMPK was detected in cells cultured in OIM with or without CCCP (2 µM) for 1-7 days. The ratio of AMPK phosphorylation/AMPK protein significantly increased at Days 3, 5, and 7 in the CCCP+OIM group compared to the OIM group ( Fig. 4A, Day 3, 1. 7 ± 0. 08; Day 5, 1. 55 ± 0. 18; Day 7, 1. 79 ± 0. 32). Later, whether AMPK influences the osteogenic differentiation of hDPSCs was examined. As shown in Fig. 4B, the activation of AMPK was elevated by AICAR treatment. Crucially, consistent with CCCP treatment, AICAR suppressed hDPSCs differentiation ( Fig. 4C ). Furthermore, the mRNA expression of TFAM was significantly decreased compared to the control on Days 5 and 7 after treatment with AICAR ( Fig. 4D, Day 5, 3. 94 ± 0. 29 vs. 1. 32 ± 0. 07; Day 7, 6. 08 ± 0. 19 vs. 1. 69 ± 0. 4). Mitochondrial fragmentation was also observed ( Fig. 4E ). These results indicated that the activation of AMPK contributed to hDPSCs differential inhibition and impaired OXPHOS. 10. 7717/peerj. 15164/fig-4 Figure 4 The activation of AMPK contributed to differential inhibition. (A) Immunoblot analysis for AMPK and phosphorylated AMPK in cell lysates from hDPSCs treated with CCCP (2 µM) or OIM only for 1, 3, 5, and 7 days. GAPDH served as the standard. (B) Cells were treated with AMPK agonist AICAR (500 µM) for 3 days. The p-AMPK was observed using immunofluorescence: p-AMPK, red; nucleus, blue. Scale bar: 50 µm. (C) The ALP, COL-1 expression and ALP activity were detected to unveil the effect of AICAR on cell differentiation. (D) qPCR analysis of mitochondrial factor TFAM in AICAR-treated (500 µM) cells. (E) Mitochondrial morphology in AICAR-treated and control cells were evaluated by fluorescence staining. Mitochondria-MitoTracker Red CMXRos (MRC), red; F-actin -Phalloidine, green; nucleus - Hoechst 33342, blue. Scale bar: 20 µm. Data were mean ± SD, * p < 0. 05, ** p < 0. 01, *** p < 0. 001. Discussion DPSCs have been used in regenerative medicine due to their multipotent capacity to differentiate into osteoblasts, adipocytes, chondrocytes, neural cells, and even endothelial cells ( Kawashima, 2012 ; Luzuriaga et al. , 2020 ). Metabolism is closely related to the physiological process of DPSCs, for example, aging and differentiation ( Macrin et al. , 2019 ; Wang et al. , 2016 ). This study demonstrated that a dynamic shift in glycolysis and mitochondrial OXPHOS occurred during the osteogenic differentiation of hDPSCs. Inhibiting mitochondrial respiration with the uncoupler CCCP suppressed hDPSCs differentiation and activated the central metabolic sensor AMPK. Metabolic shift accompanied by hDPSCs’ osteogenic differentiation Recent studies have shown that stem cells depend primarily on anaerobic glycolysis for ATP supply while differentiating cells depend on mitochondrial aerobic respiration. The metabolic shift and mitochondrial resetting into mature bioenergetic states are considered hallmarks of stem cell differentiation ( Wanet et al. , 2015 ). ALP activity and gene expression of Col-1 and ALP were used to evaluate the osteogenic differentiation potential of MSCs. In particular, ALP activity usually evaluates the early differentiation osteogenic potential. ALP activity and mRNA expression were upregulated during osteogenic differentiation of bone marrow mesenchymal stem cells ( Jiang et al. , 2022 ) and hDPSCs ( Qian et al. , 2015 ). In this study, the enhanced ALP activity and the increased mRNA levels of Col-1 and ALP also verified the osteogenic differentiation of hDPSCs. It also verified that early osteogenic differentiation occured when the osteogenic induction lasted for 7 days. When hDPSCs differentiated, cells presented distinct energy phenotypes. On Day 3, the level of ECAR and activities of glycolysis enzymes (HK, PK, and LDH) were elevated, showing that glycolysis increased. The results were consistent with our previous study ( Wang et al. , 2016 ). At this point, cells had improved glycolytic and mitochondrial function. Studies have shown that glycolysis supplies rapid energy generation and substrates for biosynthesis. Increased glycolysis can provide quick production of ATP and suitable substrates for biosynthesis, which meet the anabolic demands of initial differentiation ( Palmer et al. , 2015 ). This implies that differentiation requires the supply of sufficient energy and substrates. The combined effects of glycolysis and mitochondrial OXPHOS may support the tremendous energy demand on Day 3. Simultaneously, a higher level of OCR and an increased ratio of OCR/ECAR indicate a shift toward mitochondrial OXPHOS. The results showed that the enzyme activities of HK, PK, and LDH were decreased. Mitochondrial OXPHOS was the predominant source of energy production of the differentiating hDPSCs on Day 5 and Day 7 when a metabolic shift of hDPSCs was observed. The trend was similar to neural stem cells, mouse-induced pluripotent stem cells(iPSCs), and human iPSCs ( Coelho et al. , 2022 ; Folmes et al. , 2011 ; Prigione et al. , 2010 ). Some studies have suggested that mitochondrial content, such as mitochondrial DNA (mtDNA) copy number, mitochondrial mass, and biogenesis, is closely associated with changes in cell metabolism during pathological or physical activity ( Clark & Parikh, 2020 ; Procaccio et al. , 2014 ). To examine the mitochondrial content-related modulating factors during hDPSCs differentiation, the mRNA expression levels of TFAM and NRF1 were detected. As hDPSCs differentiated on Days 5 and 7, upregulated TFAM expression was observed, but NRF1 had no significant changes. NRF-1 and TFAM are involved in proliferator-activated receptor-γ coactivator-1α (PGC-1α)/NRF-1/TFAM signaling, the main pathway controlling OXPHEN and mitochondrial biogenesis. NRF1 is a transcriptional regulator of nuclear genes that contributes to oxidative phosphorylation. TFAM prompts mtDNA transcription, replication, and maintenance ( Kang, Chu & Kaufman, 2018 ; Scarpulla, 2008 ). NRF-1 and TFAM are also related to the differentiation of stem cells. In a previous study, NRF-1 and TFAM were increased during the osteogenic differentiation of rat dental papilla cells ( Jiang et al. , 2019 ). Highly expressed TFAM in DPSCs enhanced glutamate metabolism and OXPHOS activity. Bone regeneration of DPSCs was enhanced through the activation of mitochondrial aerobic metabolism when TFAM was highly expressed ( Guo et al. , 2022 ). It seemed that NRF1 was not changed during hDPSC differentiation. AMPK participated in differential inhibition when mitochondrial activity was attenuated The metabolic shift is essential in regulating cellular function, particularly stem cell self-renewal, pluripotency, and plasticity ( Andre et al. , 2019 ). Mitochondrial biogenesis and metabolic shifts toward OXPHOS are deemed early events in multiple stem cell differentiation processes ( Khacho & Slack, 2018 ; Wanet et al. , 2015 ). Consistent with previous reports, our data showed that inhibiting mitochondrial function by CCCP from the beginning prevented osteogenic differentiation of hDPSCs. Mitochondrial OXPHOS was required for differentiation. CCCP-treated hDPSCs exhibited an apparent fragmentation of mitochondrial tubules, likely due to a block in mitochondrial fusion. It was reported that CCCP treatment could promote the fission of the mitochondrial network in skeletal muscle cells and a human neuroblastoma cell line ( Park, Choi & Koh, 2018 ; Seabright et al. , 2020 ). The decreased intracellular ATP activated the energy sensor AMPK due to impaired OXPHOS. It was speculated that AMPK might be the direct regulator of the mitochondrial fission and fusion machinery to mediate subsequent events. As an intracellular energy sensor, the vital role of AMPK is to restore the energy balance. Once activated, AMPK induces catabolic pathways to produce energy and prevents anabolic pathways, such as lipid and protein synthesis, to save energy ( Wanet et al. , 2015 ). On the other hand, considerable evidence shows that AMPK regulates cell fate, including controlling stem cell pluripotency and differentiation ( Afinanisa, Cho & Seong, 2021 ; Fernandez-Veledo et al. , 2013 ; Liu et al. , 2021 ). In this study, CCCP-treated hDPSCs exhibited an elevated level of p-AMPK, showing that the energy sensor AMPK was involved in the process of metabolism that influenced differentiation. AMPK activation was a potential modulator when the osteogenic differentiation of hDPSCs initiated. To further verify the effect of AMPK activation, AICAR, a direct AMPK activator, was applied. AICAR resulted in mitochondrial fragmentation and differential inhibition, similar to CCCP-induced conditions. The results were consistent with previous studies in which AMPK activation has been reported to inhibit the differentiation of osteoblasts, chondrocytes, adipocytes, and myoblasts ( Bandow et al. , 2015 ; Kasai et al. , 2009 ). For cell differentiation, numerous anabolic pathways, including the biosynthesis of proteins and lipids, are required to support the specific function of differentiated cells. Mitochondrial biogenesis and stimulation of metabolism are also needed to produce adequate energy for the processes. CCCP-induced or AICAR-induced AMPK activation may be a regulator for the cells to stop differentiation during impaired OXPHOS, which might help maintain stemness or homeostasis. Mitochondrial dysfunction as a regulator may be a viable therapeutic target for stem cell-based therapies and interventions for cognitive defects ( Khacho, Harris & Slack, 2019 ). Limitations The detailed mechanism between mitochondria and osteogenic differentiation requires further investigation. Whether the metabolic shift and OPHOXS can regulate osteogenic differentiation in vivo needs to be further studied. Conclusion It was demonstrated that increased mitochondrial function was indispensable for the osteogenic differentiation of hDPSCs. It was also revealed that glycolysis gradually decreased in the stage of energy supply. The mitochondrial uncoupler CCCP depressed mitochondrial OXPHOS and inhibited hDPSCs differentiation. Activation of AMPK also interferes with mitochondrial morphology, mitochondrial OXPHOS, and osteogenic differentiation. The findings helped to reveal the relationship among glycolysis, mitochondrial OXPHOS, and osteogenic differentiation. CCCP or AMPK activation may be a potential regulator to quit differentiation from impaired OXPHOS or to maintain homeostasis. Supplemental Information 10. 7717/peerj. 15164/supp-1 Supplemental Information 1 The sequences of specific primers and Tm value Click here for additional data file. 10. 7717/peerj. 15164/supp-2 Supplemental Information 2 Culture, osteogenic capacity and cell surface markers of hDPSCs (A) phase-contrast microscopy images of hDPSCs at primary passage. (scale bar, 200 µm) (B) Alizarin red staining of hDPSCs cultured in osteogenic mineralized medium. Accumulation of mineralized nodules was observed. (Yellow arrow, scale bar, 100 µm) (C) The expressions of cell surface markers were investigated by using flow cytometry. It indicated that cells were mostly positive for CD90 and CD105 (MSCs surface markers), while they were negative for CD19 and CD34 (HSCs surface markers). Click here for additional data file. 10. 7717/peerj. 15164/supp-3 Supplemental Information 3 Figure 1 data Click here for additional data file. 10. 7717/peerj. 15164/supp-4 Supplemental Information 4 Figure 2 data Click here for additional data file. 10. 7717/peerj. 15164/supp-5 Supplemental Information 5 Figure 3 data Click here for additional data file. 10. 7717/peerj. 15164/supp-6 Supplemental Information 6 Figure 4 data Click here for additional data file. 10. 7717/peerj. 15164/supp-7 Supplemental Information 7 The raw data of WB in Figure 4 Click here for additional data file.

10. 7717/peerj. 15711

PeerJ

Development, physicochemical characterization and

This study aimed to produce hydroxyapatite from the dentine portion of camel teeth using a defatting and deproteinizing procedure and characterize its physicochemical and biocompatibility properties. Biowaste such as waste camel teeth is a valuable source of hydroxyapatite, the main inorganic constituent of human bone and teeth which is frequently used as bone grafts in the biomedical field. Fourier Transform infrared (FTIR), and micro-Raman spectroscopy confirmed the functional groups as-sociated with hydroxyapatite. X-ray diffraction (XRD) studies showed camel dentine-derived hydroxyapatite (CDHA) corresponded with hydroxyapatite spectra. Scanning electron micros-copy (SEM) demonstrated the presence of dentinal tubules measuring from 1. 69–2. 91 µm. The inorganic phases of CDHA were primarily constituted of calcium and phosphorus, with trace levels of sodium, magnesium, potassium, and strontium, according to energy dispersive X-ray analysis (EDX) and inductively coupled plasma mass spectrometry (ICP-MS). After 28 days of incubation in simulated body fluid (SBF), the pH of the CDHA scaffold elevated to 9. 2. in-vitro biocompatibility studies showed that the CDHA enabled Saos-2 cells to proliferate and express the bone marker osteonectin after 14 days of culture. For applications such as bone augmentation and filling bone gaps, CDHA offers a promising material. However, to evaluate the clinical feasibility of the CDHA, further in-vivo studies are required.

Introduction Bone is a connective tissue composed of an extracellular matrix. The organic part, which makes up 35% of the matrix, consists of collagen fibres and non-collagen fibres such as “osteocalcin, osteonectin and osteopontin, glycosaminoglycans (GAGs), lipids and plasma proteins” ( Marie, 1992 ). Bone contains four different cell types: osteoblasts, bone lining cells, osteocytes, and osteoclasts ( Marie, 1992 ). These cells are responsible for bone growth, function and maintenance ( Yu & Wei, 2021 ). The inorganic part, which makes up 65% of the matrix, consists of hydroxyapatite ( Fiume et al. , 2021 ; Chen et al. , 2021 ). Hydroxyapatite (HA), has a Ca/P ratio of 1. 667 and it has compositional similarity to human bone and teeth and demonstrates exceptional biocompatibility ( Chen et al. , 2021 ). Due to properties such as osteoconductivity, bioactivity and non-toxic nature, HA is commonly used as a bone grafting material ( Öksüz, 2018 ; Zhao et al. , 2021 ). Hydroxyapatite (HA) has been used clinically in various forms, such as powder, blocks, granules, used as coatings for biomedical implants, and in hybrid formulations with other biomaterials for increasing their osteoconductivity, mechanical properties and handling characteristics ( Ghiasi et al. , 2020 ; Khurshid et al. , 2022 ). However, there are some limitations of HA such as a slow rate of resorption and ability to undergo remodelling. Nevertheless, additional modifications can overcome these issues, such as incorporating ions into the HA lattice and increasing the porosity for specific requirements ( Mohd Pu’ad et al. , 2019 ; Ghiasi et al. , 2020 ). Bone and teeth are known to be the toughest and fully calcified tissue in the human body. HA produced from teeth has received attention recently as an emerging biomaterial ( Seo & Lee, 2008 ; Öksüz, 2018 ; Ratnayake et al. , 2020 ; Doğdu et al. , 2021 ). The calcium phosphates found in human and animal teeth have attracted significant attention as they are relatively stable when placed clinically. Hydroxyapatite (HA) can be extracted from many biological sources (such as animal bones and teeth), synthetically generated using precipitation methods using calcium and phosphate ions, and isolated utilising bioinspired approaches ( Mohd Pu’ad et al. , 2019 ). Numerous studies have shown that biologically generated hydroxyapatite is a strong possibility for creating efficient and affordable xenografts for bone repair ( Gao et al. , 2006 ; Londoño Restrepo et al. , 2016 ; Rahavi et al. , 2017 ; Ratnayake et al. , 2017 ; Öksüz, 2018 ; Luthfiyah et al. , 2022 ). In order to produce synthetic hydroxyapatite, the composition and purity of the starting material, the Ca/P mole ratio, the pH level, and the temperature of the solution should be strictly controlled. Consequently, extraction from biowaste such as animal hard tissues has been shown to provide an economical and sustainable hydroxyapatite which can be used for biomedical applications ( Mohd Pu’ad et al. , 2019 ). Based on our previous study, CBHA scaffolds were shown to be structurally stable, they retained their original morphology and provided a promising bone tissue engineering material for tissue repair ( Khurshid et al. , 2022 ). Therefore, we conducted an in-vitro study on extracted hydroxyapatite from camel dentine and camel cancellous bone HA as a control group. According to our knowledge there is no study reported which has investigated the chemical and biological properties of extracted hydroxyapatite from camel dentine. Therefore, the main aim of this study was to develop a xenograft material, namely hydroxyapatite, from the dentine portion of camel (Camelus Dromedarius) teeth using a defatting and deproteinization method ( Ratnayake et al. , 2020 ) and characterize the physiochemical and biological properties of the camel dentine derived hydroxyapatite. Materials & Methods Camel tooth dentine extraction and processing The study methodology was approved by the ethical committee of the Deanship of Scientific Research (DSR), King Faisal University, Saudi Arabia (KFU-REC-2022-JAN-ETHICS451). The camel ( Camelus dromedarius ) teeth were collected from a local slaughterhouse in Al-Ahsa, Saudi Arabia. Teeth were washed using distilled water, disinfected with 5. 25% sodium hypochlorite and all enamel and cementum was carefully removed using a MT plus Dental Plaster (Wet & Dry) trimmer (Renfert, Hilzingen, Germany). Using an ultrasonic cleaning device at a frequency of 40 to 60 kHz for 30 min, dental pulp tissues, lipids, and blood debris were eliminated ( Ratnayake et al. , 2020 ). After cleaning, the dentine was dissected into tube structures using a dental high-speed handpiece. Defatting and deproteinization of camel dentine Dentine samples were then subjected to defatting and deproteinisation procedures to remove all organic matter ( Ratnayake et al. , 2017 ; Ratnayake et al. , 2020 ). Camel bone derived hydroxyapatite scaffolds (CBHA) were used as a control in this study which was prepared according to Khurshid et al. (2022). Initially, the prepared dentine was pressure cooked for two hours at 15 psi (103. 4 kPa) in a conventional stainless steel pressure cooker and the water ratio was set at 50 mL of water per dentine sample, as previously described in Ratnayake et al. (2020). Following that, the dentine samples were dried for two hours in a desiccator. All dentine samples were soaked in a 0. 1 M sodium hydroxide (NaOH) (Sigma Aldrich, St. Louis, MO, USA) solution for 8 h (NaOH solution was changed after 4 h) at 70 °C ( Uklejewski et al. , 2015 ; Tanwatana, Kiewjurat & Suttapreyasri, 2019 ) to remove majority of the fat. Prior to calcination, the dentine blocks were dried for 30 min at 50 °C in the oven. Extraction of hydroxyapatite using low heat treatment Thermolyne, Burlington, USA, employed low heat sintering in a muffle furnace for 8 h at 750 °C to deproteinize camel dentine. Subsequently, samples were manually grounded using a laboratory mortar and pestle at room temperature under sterile conditions before being subjected to chemical analyses. Representative samples were preserved for morphological and structural analysis. Physiochemical characterization of the CDHA scaffold Fourier transform infra-red spectroscopy (FTIR) A Fourier transform infrared spectrophotometer (Perkin Elmer spectrum 100, UK) was used to identify the functional groups contained in the camel dentine derived hydroxyapatite (CDHA) samples. The FTIR spectra for CDHA sample was conducted in the mid-infra-red region in the frequency range of 400–4, 000 cm −1. Micro-Raman analysis Raman spectroscopy was used to extract data on the effect of alkaline and thermal treatment on structural and compositional changes occurring in the hydroxyapatite phase of the CDHA samples. A uniform confocal geometry of the Raman spectrometer (Horiba labRAM Evolution-2; Horiba France, Palaiseau, France) was used during the acquisition of Raman spectra deploying a fixed laser power using the microscope objective at 500 mW. The laser beam was directed from the blue emission laser (He-Cd gas laser) at a fixed excitation wavelength ≈ 448 nm. All the Using a 100x microscopic objective lens, Raman signals were captured through a 100 µm diameter confocal hole. The Raman data collecting and spectrum charting were both done using the Labspec-6 software (Horiba, Palaiseau, France), which was also utilized for post processing the spectrum. X-ray diffraction (XRD) analysis Malvern Panalytical X1 Pert MRD system (PW3040/60) was used to analyze the CDHA’s crystal structure and phase composition. Analysis was conducted in the region 10° < 2 θ < 70° using Cu-K α radiation. The generator was set to 40 kV and 40 mA. Scanning electron microscopy (SEM) analysis SEM analysis was used to assess the morphology of the CDHA scaffolds. To reveal the dentinal tubules, representative CDHA samples were cut with a scalpel blade. For the SEM and EDX analyses, the samples were subsequently coated with gold and palladium. The scanning electron microscope (Oxford JED-2300; Jeol, Tokyo, Japan) equipped with an energy dispersive X-ray (EDX, Cambridge, UK) and set at 15 kV and 15 mA was used to investigate the morphological structure and elemental composition of the CDHA. Inductively coupled plasma mass spectrometry (ICPMS) An Agilent 7500 ce quadrupole ICP-MS was used to determine the elemental composition and the calcium phosphorous molar ratio (Ca:P). The results were compared with the EDX analysis. Thermogravimetric analysis (TGA) Thermogravimetric analysis was conducted to determine the CDHA samples’ thermal degradation. TGA was conducted using a TGA analyser (Q50, TA instruments) within a N 2 atmosphere. The heating rate was 10 °C per minute, from 20 °C to 1, 000 °C. Chemical stability and biodegradation in simulated body fluid (SBF) Firstly, a stock solution of simulated body fluid (SBF) was made following the previously reported Kokubo et al. protocol ( Kokubo & Takadama, 2008 ). A scalpel blade was used to trim the samples such that their volume and mass were comparable. Each sample’s mass was noted ( n = 3). The falcon tubes were filled with the prepared CDHA sample and 10 ml of the SBF solution. The tubes were placed in an incubator at 37 ± 1 °C for 1, 5, 7, 14 and 21 days. At the conclusion of each time period, the pH of the solution was measured using a pH conductometer (Ionode, Acorn Scientific Ltd. , Auckland, New Zealand). Subsequently, the samples were removed from the pH solution and thoroughly dried in an incubator at 37 ± 1 °C for 24 h before measuring the dry weight. Weight change was calculated using the following formula and reported as a percentage WL = (W0 − W1)/W0 × 100%. W0 and W1 denotes the weights of sample before and after immersion, respectively. in-vitro biocompatibility testing of the CDHA scaffold Cell viability/cytotoxicity testing as well as immunocytochemical analysis was conducted on both the CDHA and CBHA samples using human osteoblast-like cells (Saos-2) ( Ratnayake et al. , 2023 ). The cells were purchased from the American Type Culture Collections (ATCC, Manassas, VA, USA). Cell culture Cell culture medium was prepared with minimum essential medium alpha (MEM-alpha; Invitrogen, Auckland, New Zealand) supplemented with 10% foetal bovine serum (FBS; Thermo Fisher Scientific, Auckland, New Zealand) and 5% antibiotic-antimycotic (Life Technologies, Auckland, New Zealand). The Saos-2 cells (ATCC ® HTB-85™; ATCC, Manassas, VA, USA) were cultured in 25 cm 2 cell culture flasks within a humidified incubator at 37 °C and 5% CO 2. For cell seeding onto scaffolds, once cultures in the flasks reached 70–80% confluence, all the media was removed from the flasks. The cells were then trypsinised and directly seeded onto the scaffolds. Finally, the CBHA samples were prepared using an eight mm biopsy bunch. The CDHA scaffolds were prepared to a similar size to that of CBHA cutting using a scalpel blade along the axis of the dentinal tubules (∼8 mm diameter). All dissected samples were sterilised by immersion in 70% ethanol for 30 min. Samples were then placed under UV light for another 30-min cycle and rinsed in Dulbecco’s Phosphate-Buffered Saline (DPBS). Finally, the sterilised and rinsed scaffolds were placed in one mL of the prepared culture medium and equilibrated over the next 24 h in a 37 °C humidified atmosphere with 5% CO 2. Cells were seeded at 6 × 10 3 per scaffold. The seeded scaffolds were incubated for 30 min to allow for attachment to the scaffold before adding one mL of culture media to each well and returning the seeded scaffolds with media into the cell culture incubator. The media was replaced daily. In addition, due to autofluorescence associated with the CDHA scaffolds, an indirect method using the elution media of the CDHA scaffold was used for the LIVE/DEAD and immunohistochemical analysis. In brief, 15 mg of the CDHA scaffold was placed in 10 mL of culture media and incubated in a humidified incubator at 37 °C and 5% CO 2 for 24 h. After that, the media was sterile-filtered, and the resulting elution media was used for future analysis. In addition, 6 × 10 3 cells were seeded on glass coverslips and incubated for 30 min to allow the cells to attach to the glass coverslip before adding one mL of the prepared elution media. The elution media was replaced daily. Cell fixing and SEM analysis 6 × 10 3 cells/scaffold was seeded and incubated for 48 h on the CDHA scaffold for SEM analysis. The cell-scaffold construct was fixed with 2. 5% glutaraldehyde in 0. 1 M cacodylate buffer (CB), pH 7. 4, for one hour with subsequent PBS and water rinse. The scaffolds were post-fixed with 1% Osmium tetroxide to improve the conductivity. Following that, the CDHA samples were washed with de-ionised water and dehydrated for 10 min in a serially diluted ethanol solution and critical point dried in a Bal-Tec CPD030 critical point dryer (Bal-tec, Balzers, Leichtenstein). The samples were mounted on aluminum stubs, sputter coated in an Emitech K575X sputter coater with 10 nm of gold palladium. Finally, The SEM images were captured at 15 kV accelerating voltage using a SEM (Oxford JED-2300, Jeol, Tokyo, Japan). Cell viability/cytotoxicity After each time period of 24, 48 and 72 h the cell-seeded scaffolds were stained using the LIVE/DEAD ® cytotoxicity kit according to the manufacturer’s instructions ( Ratnayake et al. , 2023 ). The experiments were observed using a confocal laser scanning microscope (Carl Zeiss Micro Imaging GmbH; Carl Zeiss, Jena, Germany) in order to determine the cell viability. Using the Zen 2009 software (Carl Zeiss), images were captured during visualisation. MTS cell proliferation assay To analyse the proliferation of the Saos-2 cells on both the CBHA and CDHA scaffolds ( n = 3), MTS [3-(4, 5-dimethylthiazol-2-yl)—5-(3-caebozymethoxyphenyl)—2-(4-sulphonyl)—2H-tetrazolium] assay was performed. Calorimetric measurement was conducted using a spectrophotometer (Synergy 2 Multi-Mode Microplate Reader; Biotek, Berlin, Germany) at a wavelength of 490 nm. Analyses were performed after 24, 48 and 72 h. The experiment was repeated in triplicate. Immunohistochemical analysis To investigate the expression of the bone-associated protein, osteonectin, using immunofluorescence, 20 × 10 3 cells were seeded on to glass coverslips ( n = 3). Once the cells were attached to the coverslip, the cells were fed with elution media supplemented with 10 nM dexamethasone, 5 mM β -glycerophosphate ( β -GP) and 100 µM L-ascorbic acid 2 –phosphate supplements for up to 14 days to induce cell differentiation. After the culture period (14 days), the coverslips were fixed with 10% neutral buffered formalin for four hours and then rinsed twice with PBS. Subsequently, the coverslips were blocked with donkey serum (Sigma Aldrich, Auckland, New Zealand) and then incubated with mouse anti-osteonectin (1:500 dilution; BIoGenex, USA) overnight in a refrigerator at 4 °C. The following day, the coverslips were incubated with AlexaFluor donkey anti-mouse 488 secondary antibody (Life Technologies, NZ) with excitation/emission of 495/519 nm. 4′, 6-diamidino-2-phenylindole (DAPI) (Duolink; Sigma Aldrich) was used to counterstain the coverslip and visualize the cell nucleus. Finally, coverslips were observed using an Olympus fluorescence microscope (Bahns Enterprise Microscopes, Bern, Germany). Images were captured using Q-capture 3. 1. 2 software. Results Processing of camel dentine and weight change during dentine processing Freshly prepared dentine disc from camel tooth requires a series of treatments for removal of the organic fats and proteins. This was achieved by pressure cooking the camel dentine samples for 2 h which removed majority of the fat and collagen. After that, the dentine samples were soaked in a 0. 1 M Sodium hydroxide (NaOH) (Sigma Aldrich, St. Louis, MO, USA) solution for 8 h (NaOH solution was changed after 2 h) at 70 °C which resulted in liquefied fat pouring from the dentine matrix. Weight changes indicted the removal of impurities and fats. Figure 1, shows the weight changes from (a–c) 111 g to 106 g following treatment with 0. 1 M sodium hydroxide (NaOH) (Sigma Aldrich, St. Louis, MO, USA) solution for 8 h. Furthermore, a weight loss of 44% was observed after heat treatment at 750 °C. 10. 7717/peerj. 15711/fig-1 Figure 1 Weight changes of camel dentine after 0. 1 M NaOH treatment (A, B) and 750 °C heat treatment (C). Physicochemical characterization of the CDHA samples Fourier transform infra-red spectroscopy (FTIR) Figure 2 illustrates the FTIR spectra of CDHA obtained after sintering at 750 °C. Hydroxyl (3, 571 cm −1 ), phosphate (1, 092 to 1, 040 cm −1, 962 cm −1, 633 to 566 cm −1 and 473 cm −1 ) and carbonate peaks (1, 455 to 1, 418 cm cm −1 ) associated with hydroxyapatite were observed for the CDHA sample, and the peaks associated with fat and collagen were not observed ( Ratnayake et al. , 2017 ; Ratnayake et al. , 2020 ). Micro-Raman spectroscopy Micro-Raman results for the CDHA samples indicated that CDHA is primarily comprised of HA after the defatting and deproteinising procedures. The observed HA phase was confirmed by the presence of intense Raman signature of the phosphate group (PO 4 ) arising from the stretching vibrational mode and observed in the Raman spectrum, as also shown in the Fig. 3, at 962 cm −1. The observed formation of the hydroxyapatite compositional phase has also been previously reported in the literature at similar spectral position of ≈ 960 cm −1 ( Timlin et al. , 2000 ; Sofronia et al. , 2014 ; Jaber, Hammood & Parvin, 2018 ). The small change in the vibrational energy in our samples can be related to the higher heat treatment temperature used, which has also been previously reported and is due to the bond softening among phosphate bands at higher annealing temperatures ( Campillo et al. , 2010 ; Marques et al. , 2018 ). The Raman peaks observed at 433 cm −1, 594 cm −1 and 1, 032 cm −1 are also found to be directly related to the lower vibrational modes of the hydroxyapatite phase of the composite ( Timlin et al. , 2000 ; Jaber, Hammood & Parvin, 2018 ). The C-H and C-O vibrational modes have also been found in the samples in the form of peaks positioned at 1, 046 cm −1 and 1, 079 cm −1, respectively, and that has further validated the presence of HA phase in the CDHA sample ( Timlin et al. , 2000 ; Londoño Restrepo et al. , 2016 ). 10. 7717/peerj. 15711/fig-2 Figure 2 FTIR spectra for the CHDA sintered at 750 °C. 10. 7717/peerj. 15711/fig-3 Figure 3 Micro-Raman spectra shows the presence of peak at 962 cm −1 for the CDHA sample. X-ray diffraction (XRD) analysis The results of the XRD analysis of the sintered CDHA powder and phase pure HA pattern (JCPDS 01-080-7085) is presented in Fig. 4. The XRD diffract grams obtained for the CDHA powder after sintering at 750 °C are very similar to the pattern for phase pure hydroxyapatite pattern (JCPDS 01-080-7085). the XRD patterns were sharper and narrower, demonstrating a high crystalline structure. 10. 7717/peerj. 15711/fig-4 Figure 4 X-ray diffraction patterns of CDHA (red) after sintering at 750 °C and phase pure HA (blue) (JCPDS 01-080-7085). Scanning electron microscopy (SEM) analysis Figure 5 illustrates the microscopic porous structure of the CDHA scaffold. The presence of the dentinal tubules indicated that the fat and collagen were successfully removed from the matrix of the CDHA scaffold. The dentinal tubules (Blue arrows), measured between 1. 69–2. 91 µm in diameter ( Fig. 5 ). However, some structural deterioration and debris were observed in this particular specimen (yellow arrows). 10. 7717/peerj. 15711/fig-5 Figure 5 Representative scanning electron micrograph (SEM) of the CDHA scaffold. Scale Bar = 10 µm. Blue arrows indicate dentinal tubules. Yellow arrows indicate the debris. Energy dispersive x-ray (EDX) and inductively coupled plasma mass spectrometry (ICPMS) analysis Figure 6 shows the elemental composition of the CDHA sample obtained by EDX analysis. The inorganic component of the CDHA scaffold mainly consists of calcium and phosphate with trace amounts of sodium, magnesium and strontium. 10. 7717/peerj. 15711/fig-6 Figure 6 EDX analysis of the CDHA sample. The chemical composition of the CDHA sample analysed using ICPMS and the Ca/P mole ratio is summarized in Table 1. All values are reported in mg/kg. The ICPMS analysis further showed that toxic elements such as cadmium, arsenic and lead are lower than the concentration limits suggested by American Society for Testing and Materials (ASTM) standards (F1185-03). Thermogravimetric analysis (TGA) Figure 7 shows the TGA for the CDHA when heated from 0 °C to 1, 000 °C. Approximately 2% of the weight was lost from the CDHA sample when the temperature had reached 1, 000 °C indicating excellent thermal stability upon sintering. Chemical stability and biodegradation in simulated body fluid Chemical stability The initial pH of the SBF solution was 7. 42. The pH of the buffer medium without the sample, which acts as a control, ranged between 7. 4–7. 5 throughout the experimental period. However, after day 1, the pH of the SBF solution containing the CDHA scaffold reached 9. 6 by day 5 (* p < 0. 05) and decreased to 9. 2 after 21 days (** p < 0. 01) ( Fig. 8 ). 10. 7717/peerj. 15711/table-1 Table 1 Chemical composition of CDHA as determined by ICP-MS analysis and EDX analysis. ICP-MS analysis of the CDHA sample. Sample (mg/Kg) Na Mg Ca P Zn K Sr Cd As Pb CDHA ASTM maximum limit (F1185-03) 1. 15 × 10 4 2. 01 × 10 4 3. 94 × 10 5 1. 8 × 10 5 207 325 894 <0. 25 5 <0. 25 211 0. 35 5 Ca/P ratio (SEM-EDX) = 1. 74 Ca/P ratio (ICP-MS) = 1. 69 10. 7717/peerj. 15711/fig-7 Figure 7 TGA of the CDHA scaffold to 1, 000 °C. 10. 7717/peerj. 15711/fig-8 Figure 8 Graphical representation of pH change over time following incubation in SBF ( n = 3). Error bars represent ± SE of the mean after 1-way ANOVA with Tukey’s multiple comparison (* p < 0. 05, ** p < 0. 001). Degradation The CDHA scaffold maintained its original morphology over 21 days when immersed in SBF. Although the scaffold exhibited increased weight loss during the time period investigated, only a minimal weight loss was observed from day 1 to day 21 (∼1. 4%) and this result was statistically significant (**** p < 0. 0001) Fig. 9. 10. 7717/peerj. 15711/fig-9 Figure 9 The percentage weight loss of the CDHA scaffolds vs. the investigated time period. ( n = 3. * p < 0. 05, **** p < 0. 0001, error bars represent +SE of the mean after one-way ANOVA with Tukey’s multiple comparison). in-vitro biocompatibility testing Cell fixing and SEM analysis Figure 10 shows an SEM image of the cellular attachment of the Saos-2 cells on the CDHA and CBHA scaffolds after 72 h culture. The cells seeded on both scaffolds exhibited slender cytoplasmic extensions (red arrows). 10. 7717/peerj. 15711/fig-10 Figure 10 SEM image of the CDHA (A) and CBHA (B) scaffold with attached Saos-2 cells after seeding for 72 h. Yellow arrows indicate Saos-2 cells. Red arrows indicate cytoplasmic projections. Bar = 100 µm. LIVE/DEAD assay Figure 11 shows the CDHA and CBHA scaffolds seeded with Saos-2 cells after performing the LIVE/DEAD Assay at 24, 48 and 72 h. Results indicated excellent cell viability for both types of scaffolds for all time periods investigated. The cells seeded on glass coverslips using the elution media of the CDHA scaffold exhibited a spindle-like morphology. 10. 7717/peerj. 15711/fig-11 Figure 11 Fluorescent images from the LIVE/DEAD ® assay of Saos-2 cells growing on CDHA scaffolds. Image shows cell viability at 24 h, 48 h and 72 h culture. (A) Saos-2 cells seeded directly on the CDHA scaffold. (B) Saos-2 cells seeded directly on CBHA scaffold. (C) Saos-2 cells seeded with the elution media of the CDHA scaffolds. Green, live cells (calcein), Red, dead cells (ethidium homodimer-1). MTS cell proliferation assay The CDHA and CBHA scaffolds cell proliferation was investigated using the MTS assay ( Fig. 12 ). Both scaffold types showed an increase in cell density over time. Despite the fact that the CDHA scaffold had numerically more cells than the CBHA scaffold, there was no statistically significant difference at any of the time points examined. 10. 7717/peerj. 15711/fig-12 Figure 12 MTS cell proliferation assay on CDHA and CBHA scaffolds after 24, 48 and 72-h culture. Error bars represent ± SE of the mean after 1-way ANOVA with Tukey’s multiple comparison test. ∗ p < 0. 05, ** p < 0. 01, *** p < 0. 001, **** p < 0. 0001. n = 3. Immunohistochemical analysis Immunohistochemical analysis showed that the Saos-2 cells expressed the bone marker osteonectin on the surface of the CDHA scaffold after 14 days culture ( Fig. 13 ). 10. 7717/peerj. 15711/fig-13 Figure 13 Representative images for the immunohistochemical analysis for osteonectin after 14 days on the CDHA scaffold after seeding of 20 × 103 Saos-2 cells. (A) Positive immunofluorescence staining (green) for osteonectin. (B) Nuclear staining (DAPI staining, blue) of the Saos-2 cells. Discussion A xenograft material, hydroxyapatite, was successfully produced from the dentine portion of camel teeth using a simple, cost-effective and reproducible defatting and deproteinisation method ( Ratnayake et al. , 2017 ; Ratnayake et al. , 2020 ; Khurshid et al. , 2022 ). Pressure cooking is an efficient method of removing fat and collagen from the dentine matrix as it denatures collagen. NaOH acts as a detergent removing fats and proteins. Fat absorbs microwave energy faster and is a cost-effective method of removing fat. It has previously been reported on the effectiveness of heat-treatment for deproteinization of bone cubes or other animal biowaste for use in osseous healing, such as bone graft and voids filling ( Khurshid et al. , 2022 ). Consequently, the samples were treated at 750 °C for 8 h in a dry heat furnace. It was evident that mass and colour changes (from yellow into chalky white) of the camel dentine samples subsequently occurred ( Fig. 1 ). The FTIR spectra of the CDHA showed the characteristic peaks associated with hydroxyapatite ( Fig. 2 ) ( Joschek et al. , 2000 ). In this study, peaks associated with collagen and fat were not observed, suggesting that all the organic matter was removed ( Barakat et al. , 2009 ). The sharp peak observed at 3, 571 cm −1 is due to the presence of the hydroxyl group. The peaks observed at 1, 092 to 1, 040 cm −1, 962 cm −1, 633 to 566 cm −1 and 473 cm −1 represent the phosphate bands ( Elkayar, Elshazly & Assaad, 2009 ; Ramesh, Ratnayake & Dias, 2021 ). In addition, carbonate was retained within the CDHA structure, which is confirmed by the carbonate bands observed at 1, 455 to 1, 418 cm cm −1 and 873 cm −1, respectively. Similar studies which used mammalian sources to produce hydroxyapatite, bone matrix-associated phosphate, carbonate and hydroxyl peaks were also observed in their respective FTIR spectra ( Ratnayake et al. , 2017 ; Ratnayake et al. , 2020 ; Khurshid et al. , 2022 ). Raman analysis was performed on the CDHA sample and the spectral positions of the Raman peaks suggest a strong presence of hydroxyapatite phase owing to the observation of all the four characteristic vibrational modes of phosphate bands in the CDHA sample ( Timlin et al. , 2000 ; Sofronia et al. , 2014 ; Jaber, Hammood & Parvin, 2018 ). Furthermore, all the four major Raman vibrational modes of the sample were also compared with the high quality spectral data obtained from the RRUFF hydroxyapatite database (ID number R100225) ( Lafuente et al. , 2016 ). A perfect Raman peak matching was observed for all the four peaks with that of the Ca 5 (PO 4 ) 3 OH’s reference Raman database shown in Fig. 3. Furthermore, the presence of 1, 032 cm −1 Raman mode in our sample indicates a significant degree of stoichiometric retention in the hydroxyapatite phase during the chemical and thermal treatments of the CDHA samples ( Timlin et al. , 2000 ). The retention mode ≈ 1, 032 cm −1 has been reported to be Raman inactive in the non-stoichiometric HA phase ( Timlin et al. , 2000 ). The XRD diffractograms showed sharp and narrow peaks indicating increase crystallinity and corresponded to the peaks associated with synthetic hydroxyapatite ( Fig. 4 ). In addition, secondary phases such as calcium oxide (2 θ = 37°) and tricalcium phosphate ( β -TCP) (2 θ ≈ 31. 13°) were not observed suggesting that the CDHA has undergone minimal decomposition after sintering at 750 °C leaving the carbonate in the lattice which was further confirmed by the FTIR spectra ( Fig. 2 ). Several other studies found similar findings when hydroxyapatite was synthesized from biowaste ( Barakat et al. , 2009 ). SEM analysis of the CDHA scaffold showed an interconnected micro-porous structure consisting of dentinal tubules. These dentinal tubules ranged from 1. 69–2. 91 µm, and these micropores could assist in enabling angiogenesis and cellular attachment ( Timlin et al. , 2000 ). The debris observed ( Fig. 5, yellow arrows) in the CDHA scaffold is likely due to the defatting and deproteination processes. The trace elements found in the CDHA scaffold could play a vital role in bone metabolism ( Landi et al. , 2008 ; Krishnamurithy et al. , 2014 ). Sodium plays a crucial role in cell adhesion and the bone mineralisation process. Magnesium stimulates the proliferation of osteoblasts, and strontium enhances osteoblast activity whilst inhibiting osteoclast activity ( Bigi et al. , 2007 ; Landi et al. , 2008 ). A slightly higher Ca:P mole ratio was calculated for CDHA compared with stoichiometric HA (1. 67) through EDX and ICP-MS analysis ( Table 1 ). The slightly higher value was due to the carbonate groups and trace elements in the CDHA sample ( Bahrololoom et al. , 2009 ; Ratnayake et al. , 2017 ). The result observed in this study is similar to our previous studies which used bovine teeth and bone as a. According to the TGA analysis, the CDHA scaffold only decreased by ∼2% of the weight suggesting an excellent thermal stability for the CDHA scaffold. The initial weight loss (endothermic loss) from 50 −200 °C was attributed to the evaporation of absorbed water. The minimal weight loss (∼1%) between 250–600 °C (exothermic loss) was due to the dissociation of the carbonate groups in the CDHA lattice ( Saalfeld et al. , 1994 ). A similar finding was observed in our previous study which utilized the dentine portion of bovine teeth ( Ratnayake et al. , 2020 ). Simulated body fluid (SBF) mimics the intracellular compartment of human body fluids and tissues ( Kokubo & Takadama, 2006 ). The pH increased significantly to 9. 6 by day five and dropped to 9. 2 by day 21 following sample treatment. This significant increase in pH could be due to the decomposition of the carbonate groups in the CDHA into CaO which reacts with the water to form Ca(OH) 2. Another factor could be the leaching of Ca 2+ ions and trace amounts such as Na, Mg, K and Sr from the CDHA lattice into the SBF solution to form an alkaline solution (NaOH, Mg(OH) 2, KOH, Sr(OH) 2 ) ( Ratnayake et al. , 2017 ; Ratnayake et al. , 2020 ). The degradation study showed that the CDHA scaffold decreased by ∼1. 5% of its original weight after 21 days treatment. Several studies have shown minimal HA degradation ( Gomi et al. , 1993 ; Monchau et al. , 2002 ). For example, Legeros et al. showed that physical properties such as a higher porosity and interconnected porous network increase the biodegradation of a scaffold ( LeGeros et al. , 1988 ). However, the CDHA scaffold exhibited a dense architecture consisting of micropores ranging from 1. 69–2. 91 µm, which would have minimised the biodegradation. In a previous study, a HA scaffold produced from bovine dentine showed a similar result, losing only ∼2. 5% of its original weight. However, the bovine dentine-derived HA (BDHA) pore sizes were larger than in the CDHA studied here ( Ratnayake et al. , 2020 ). Limitations of the biodegradation experiment included the absence of primary enzymes, such as lysozymes, responsible for the scaffold’s degradation in physiological environments. In addition, the scaffold was not placed in a true physiological environment where mechanical forces of perfusion and cellular waste products would degrade the scaffold. Human osteosarcoma (Saos-2) cells are commonly used to assess the biocompatibility of orthopedic materials due to its osteoblast like properties. Cells were seeded at 6 × 10 3 directly on to the surface of the scaffolds to evaluate their adhesion and proliferation. This specific cell density has been identified as reaching an appropriate confluency over the investigated time period (22 h) ( Ratnayake et al. , 2020 ). However, due to the autofluorescence and cells being adhered in different planes on the CDHA scaffolds, the elution media of the CDHA scaffold was used for the LIVE/DEAD assay. The SEM image ( Fig. 10 ) showed that the cells adhered to the CDHA and CBHA scaffolds and exhibited cellular projections which indicates intracellular communication and proliferating. The LIVE/DEAD assay showed the CDHA scaffold is non-toxic, allowing cells to adhere and proliferate ( Figs. 11A, 11C ). In addition, no dead cells were present indicating increased growth and viability over the cultured period. Indeed, the cells penetrated deeper into the pores of the CBHA scaffold, and the cells were viable ( Fig. 11B ). The MTS assay showed that the cells proliferated during the investigated time period for both the CDHA and CBHA scaffolds ( Fig. 12 ). Although the cell numbers were marginally numerically higher for the CDHA scaffold compared with the CBHA scaffold, this result was not statistically significant. The increased cell proliferation for the type of scaffolds indicated that the cells adhered to the scaffolds and proliferated. One of the reasons for this may be due to the trace elements (Carbonate, Mg, Na, K and Sr) present in the scaffolds as well as the topography ( Parsons et al. , 1988 ; Nair & Laurencin, 2007 ; Ratnayake et al. , 2020 ). Due to the dentinal tubules, the CDHA scaffold exhibited an uneven topography. Deligianni et al. (2000) and Golub & Boesze-Battaglia (2007) reported that a porous architecture and a rough surface promote a significantly increase in cell attachment. In addition, the LIVE/DEAD assay and SEM analysis ( Fig. 10 ) showed that the cells exhibited cytoplasmic projections known as filopodia, which act as anchorage points for cells to facilitate cell adhesion and migration. For immunohistochemical analysis, 20 × 10 3 cells were cultured on the well plate using the elution of the CDHA scaffold. An indirect method was used to eliminate autofluorescence associated with the CDHA scaffolds. A 14-day culture period was chosen according to a study Gao et al. who found that a cultivation period of 14 days was necessary to express osteonectin on a nano-porous silicon substrate ( Gao et al. , 2006 ). Osteonectin is a bone matrix protein associated with bone mineralization and is used as a bone marker. Immunohistochemical/immunofluorescence analysis demonstrated expression of the bone marker osteonectin after 14 days indicating in-vitro osteogenic differentiation leading to extracellular bone matrix formation. Conclusions This study showed that waste camel teeth are a potential source to extract hydroxyapatite using a simple, cost-effective, reproducible method. Although the CDHA material showed excellent biocompatibility properties in in vitro, further in vivo research is warranted to evaluate its feasibility as a bone substitute for clinical applications such as bioactive coatings for orthopaedic implants, bone void fillers and dental bone grafting. Supplemental Information 10. 7717/peerj. 15711/supp-1 Supplemental Information 1 FTIR data Click here for additional data file. 10. 7717/peerj. 15711/supp-2 Supplemental Information 2 XRD raw data Click here for additional data file. 10. 7717/peerj. 15711/supp-3 Supplemental Information 3 EDX raw data Click here for additional data file. 10. 7717/peerj. 15711/supp-4 Supplemental Information 4 Raw data exported from the SEM image of the Camel Dentine Hydroxyapatite (CDHA) scaffold with attached Saos-2 cells after seeding for 72 h Click here for additional data file. 10. 7717/peerj. 15711/supp-5 Supplemental Information 5 Raw data exported from the SEM image of the Camel Bone Hydroxyapatite (CBHA) scaffold with attached Saos-2 cells after seeding for 72 h Click here for additional data file. 10. 7717/peerj. 15711/supp-6 Supplemental Information 6 24 h cell proliferation Click here for additional data file. 10. 7717/peerj. 15711/supp-7 Supplemental Information 7 48 h cell proliferation Click here for additional data file. 10. 7717/peerj. 15711/supp-8 Supplemental Information 8 72 h cell proliferation Click here for additional data file.

10. 7717/peerj. 1650

PeerJ

Ultra-structural changes and expression of chondrogenic and hypertrophic genes during chondrogenic differentiation of mesenchymal stromal cells in alginate beads

Chondrogenic differentiation of mesenchymal stromal cells (MSCs) in the form of pellet culture and encapsulation in alginate beads has been widely used as conventional model for in vitro chondrogenesis. However, comparative characterization between differentiation, hypertrophic markers, cell adhesion molecule and ultrastructural changes during alginate and pellet culture has not been described. Hence, the present study was conducted comparing MSCs cultured in pellet and alginate beads with monolayer culture. qPCR was performed to assess the expression of chondrogenic, hypertrophic, and cell adhesion molecule genes, whereas transmission electron microscopy (TEM) was used to assess the ultrastructural changes. In addition, immunocytochemistry for Collagen type II and aggrecan and glycosaminoglycan (GAG) analysis were performed. Our results indicate that pellet and alginate bead cultures were necessary for chondrogenic differentiation of MSC. It also indicates that cultures using alginate bead demonstrated significantly higher (p < 0. 05) chondrogenic but lower hypertrophic (p < 0. 05) gene expressions as compared with pellet cultures. N-cadherin and N-CAM1 expression were up-regulated in second and third weeks of culture and were comparable between the alginate bead and pellet culture groups, respectively. TEM images demonstrated ultrastructural changes resembling cell death in pellet cultures. Our results indicate that using alginate beads, MSCs express higher chondrogenic but lower hypertrophic gene expression. Enhanced production of extracellular matrix and cell adhesion molecules was also observed in this group. These findings suggest that alginate bead culture may serve as a superior chondrogenic model, whereas pellet culture is more appropriate as a hypertrophic model of chondrogenesis.

Introduction Chondrogenic differentiation is a unique process that starts with cell–cell interactions and is influenced by the presence of specific growth and differentiation factors. During embryonic limb formation, this process is described as condensation. Based on this phenomenon, different chondrogenic induction techniques have been developed to recapitulate the in vivo chondrogenesis with the hopes that tissue repair can be achieved in clinical applications. It has been shown that the use of monolayer cultures is inadequate to reproduce chondrogenesis and there is a need for three-dimensional (3D) culture systems in order for this to occur. As such, cultures using high density of cells in the form of pellet or aggregates or a combination of cells with biomaterials are used to provide a cell-embedded 3D structure. It is said that this produce highest cell–cell and cell–matrix interactions and responsible for cell differentiation process ( Penick, Solchaga & Welter, 2005 ). However, in addition to inducing chondrogenesis, most of these culture conditions also lead towards a hypertrophic differentiation after the initial chondrogenic induction, similar to that observed in the terminal differentiation of hypertrophic chondrocytes during endochondral ossification ( Ma et al. , 2003 ; Steinert et al. , 2003 ; Ichinose et al. , 2005 ; Xu et al. , 2008 ; Bian et al. , 2011 ). This unstable chondrogenic phenotypic expression after the initial induction was considered to be the major hurdle to in vitro chondrogenesis for application in cartilage tissue engineering. During the initial stages of chondrogenesis and condensation of mesenchymal stromal cells (MSCs) in limb bud formation, higher expressions of two major cell adhesion molecules N-CAM and N-cadherin have been reported ( Monroy & De Leon, 1999 ; Woodward & Tuan, 1999 ; Hall & Miyake, 2000 ). These molecules, however, appear to be downregulated after chondrogenic differentiation and only expressed in the periphery of the limb anlagen in vivo or chondrogenic aggregate in vitro ( Widelitz et al. , 1993 ; Tavella et al. , 1994 ). In articular cartilage, each chondrocyte has been shown to be responsible for the production of extracellular matrix (ECM), which, when fully formed, creates a functional unit of cartilage or chondron that ultimately leads to the formation of cartilage matrix ( Poole, 1997 ) and there appears no direct cell–cell contact. Chondrocyte serves as the only responsible cell type for tissue homeostasis or synthesis and degradation of ECM ( Pearle, Warren & Rodeo, 2005 ); therefore, a profound understanding of the morphology and physiology of the engineered chondrocyte-like cell is required to determine the likelihood of cartilage regeneration outcomes. Although alginate bead culture system has been shown to provide beneficial microenvironment for evaluating chondrogenesis of MSCs in vitro ( Yang et al. , 2004 ; Ichinose et al. , 2005 ; Xu et al. , 2008 ; Duggal et al. , 2009 ; Diekman et al. , 2010 ), it has not been studied in greater detail. This is especially true when determining the cell adhesion molecules, hypertrophic genes, and detailed ultrastructural studies relating to cell–matrix interactions during the chondrogenic differentiation of MSCs in alginate beads. In addition, the advantage of one group over (pellet culture vs alginate culture) the other has not been previously demonstrated. Therefore, the present study was conducted to examine chondrogenic, hypertrophic, and cell adhesion molecule gene expressions and ultrastructural changes of chondrogenic differentiated MSCs in alginate beads and compare them with pellet and monolayer cultures. Materials and Methods Isolation and characterization of human bone marrow stromal cells Human bone marrow samples were obtained from healthy adults (male, age =21 + 2. 6 years) who were undergoing fracture fixation involving the long bones. After providing the patient information sheet and explaining the patients on bone marrow collection, written informed consent was obtained and bone marrow was collected. This study was approved by the University of Malaya Medical Center Ethics Committee (reference no. 602. 22). Human bone marrow was collected ( N = 6) in sterile 3-ml BD Vacutainer blood tubes (K2 EDTA, BD, Franklin Lakes, NJ, USA) by orthopedic surgeons and was kept at 4 °C until isolation. The mononuclear cells were isolated using Ficoll density gradient method as described earlier ( Krishnamurithy et al. , 2014 ). Isolated cells were cultured and expanded until P3 in T-75 flasks. The isolated cells were characterized as MSCs through flow cytometry using cell surface markers (positive and negative markers) and its ability to undergo trilineage differentiation (adipogenic, chondrogenic, and osteogenic differentiation) was described earlier ( Tan et al. , 2013 ). MSCs isolated form three biological samples were used for gene expression analysis and three biological samples were used for imaging analysis. Chondrogenic differentiation and experimental groups Pellet culture: MSCs were harvested at P3, and 2. 5 × 10 5 cells were pelleted in a 15-ml propylene centrifuge tube at relative centrifugal force (RCF) of 260 × g 5 min. After removal of the supernatant, the pellet cultured with 2 ml of chondrogenic medium contains the following: DMEM high glucose (4. 5 mg/ml d -glucose) with sodium pyruvate (110 µg/ml) (Invitrogen, Carlsbad, CA, USA), 50 mg/ml ITS (Sigma) (1×) (Invitrogen), l -ascorbate 2 phosphate (50 µg/ml) (Sigma, St. Louis, MO, USA), 10 ng/ml TGF β 3 (Invitrogen), 100 nM dexamethasone (1 × 10 −7 M) (Sigma), 100 µg/ml penicillin/streptomycin (Invitrogen), and 40 µg/ml l -proline (Sigma). The media were changed every 3 days. Alginate cell constructs: 1. 2% alginate prepared from alginic acid powder, low viscosity (Sigma-Aldrich) in 0. 9% sodium chloride (NaCl) and filtered sterile by a 0. 2-µm filter. MSCs at P3 were harvested, and a concentration of 4 × 10 6 cells per milliliter of alginate was obtained before dropping into sterile calcium chloride solution (CaCl 2 ) using a pipette. Alginate bead constructs were cross-linked in this solution for 10 min in an incubator at 37 °C, were rinsed in 0. 9% normal saline 2–3 times, and then transferred to the culture dishes (ultra-low attachment 6-well plates; Corning, New York, NY, USA). Three beads were placed per well (about 80, 000 cells per bead) and supplemented with 2 ml of chondrogenic medium. The medium was changed every 3 days. Alginate cell constructs were dissociated in a buffer solution containing 0. 015 M sodium citrate and 0. 15 M sodium chloride (Na 3 C 6 H 5 O 7 dehydrate 2H 2 O, MW = 294. 10), pH 7. 2, and centrifuged at 1, 100 rpm for 5 min before gene expression studies were conducted. Monolayer culture: MSCs at P3 were cultured in chondrogenic medium at a density of 4, 000 cells per centimeter in 6-well plates. Glycosaminoglycan (GAG) analysis and DNA quantification Alginate beads and pellets from days 3, 12, and 21 were washed with phosphate-buffered saline. Alginate bead was dissolved in sodium citrate and the pellets were crushed and the samples were digested in papain buffer overnight at 60 °C. The digested samples were subjected to biochemical analyses to determine the glycosaminoglycan and DNA content. GAG production was determined by using Blyscan assay kit (Biocolor, arrickfergus, Northern Ireland), according the manufacturer’s protocol. GAG content was determined using a standard curve drawn using standard solutions containing chondroitin 4-sulfate. Alginate beads without cells at the respective time point was analyzed in the same manner and the values are used as blank. DNA content was determined by using the fluorescent picoGreen dsDNA quantification assay (Invitrogen). GAG in alginate bead and pellet culture system was expressed as µg GAG per µg of DNA. Morphological studies On day 21, samples were fixed in 10% formalin and processed for histological studies stained with Safranin O Fast green or immunohistochemistry for Collagen type II (Mouse monoclonal anti-Human; Merk, Darmstad, Germany) in 1/100 dilution and aggrecan (Mouse monoclonal anti-Human; Abcam, Cambridge, UK) in 1/50 dilution using EnVision+ System-HRP (DAB) Dako kit (Dakocytomation, Glostrup, Denmark) according to the manufacturer’s protocol. RNA isolation and cDNA synthesis Total RNA was isolated from cultures of pellet, alginate, and monolayer on days 3, 12, and 21. RNA was isolated using SV total RNA isolation system (Promega, Madison, WI, USA) according to the manufacturer’s protocol. One hundred nanograms of RNA was used to generate cDNA with iScriptTM Reversed Transcription Supermix for RT-qPCR (BioRad) according to the manufacturer’s protocol. Quantitative real-time polymerase chain reaction PCRs were carried out in duplicate for each three biological sample in a final volume of 20 µl containing SYBR Green mastermix (BioRad, Hercules, CA, USA) with 160 nM concentration of each primer and 100 ng cDNA in 0. 2-ml PCR tubes using CSFX96TM Real Time System, Bio-Rad, under the following conditions: 3 min at 95 °C followed by 40 cycles at 58. 6 °C for 0. 20 s as annealing temperature and 72 °C for 0. 30 s as extension. The reactions were ended by 0. 1 min at 95 °C and a melt curve by increasing temperature from 65 °C, 0. 05 min, to 95 °C, 0. 5 min, stepwise. No template controls were used in each reaction as negative control. The data were presented as a time fold change relative to the internal control gene expression. The data were then normalized to transcription levels of day 0 culture using ΔCT and ΔΔCT methods. Values below 1 were considered downregulated. The following primer sets were applied in this experiment ( Table 1 ). 10. 7717/peerj. 1650/table-1 Table 1 Primer sequence. Gene Access no. Primer pairs 5′-3′ Amplicon size/reference Aggrecan NM_001135. 3 F CTACGACGCCATCTGCTACA 141 R TCAGTGATGTTTCGAGGCAG Beta-actin NM_001101. 3 F CTCTTCCAGCCTTCCTTCCT 116 R AGCACTGTGTTGGCGTACAG Collagen II NM_033150. 2 F GAAAGCCTGGTGATGATGGT 138 R GGCCTGGATAACCTCTGTGA Collagen X NG_008032. 1 F CACCTGTGGTCCTGAATGTG 163 R TCTGAGTGCCTGGATGTCTG N-Cadherin NM_001792. 3 F GGAAAAGTGGCAAGTGGCAG 159 R GGAGGGATGACCCAGTCTCT NCAM1 NM_000615. 6 F AGGAGACAGAAACGAAGCCA 161 R GGTGTTGGAAATGCTCTGGT Run X2 NM_001015051. 3 F TTACTTACACCCCGCCAGTC 139 R CACTCTGGCTTTGGGAAGAG SOX9 NM_000346. 3 F AGACAGCCCCCTATCGACTT 108 R CGGCAGGTACTGGTCAAACT GAPDH Bian et al. (2011) F AGGGCTGCTTTTAACTCTGGTAAA 111 R GAATTTGCCATGGGTGGAAT Transmission electron microscopy (TEM) Samples for TEM study were fixed with 4% glutaraldehyde in cacodylate buffer for 24 h at 4 °C and postfixed in buffered 1% osmium tetra oxide for 2 h at 4 °C, washed with cacodylate buffer 2–3 times, and kept overnight at 4 °C in cacodylate buffer before en bloc staining with 4% uranyl acetate (Agar Scientific Ltd R1043, Essex, England) in double distilled water for 10 min. Samples were then dehydrated in ascending series of ethanol and embedded in Epon (Agar 100 resin kit; Agar Scientific Ltd R1043). Semi-thin 1 µm sections were obtained using a Leica ultramicrotome (Reichert Ultracuts; Leica Microscystem, Vienna, Austria) and stained with toluidine blue for 1–2 min before viewing with light microscopy. Ultrathin sections were cut at 70–80 nm, collected on copper grids (300 meshes), and stained with uranyl acetate and lead citrate (Agar Scientific). Images were viewed using TEM (Leo Libra 120; Carl Zeiss SMT AG, Oberkochen, Germany). Statistical analysis The difference between experimental groups was calculated using nonparametric Kruskal-Wallis H test and the difference between two independent experimental groups using Mann–Whitney U test, available on the statistical software package SPSS (version 18. 0), with p ≤ 0. 05 being considered significant. Results Chondrogenic differentiation Chondrogenic differentiation of MSCs in different experimental groups was shown in Fig. 1. Safranin O Fast green staining of pellet culture ( Fig. 1A ) on day 21 showed red staining positive for GAG and green area indicating non-GAG depositions. In alginate culture on day 21, Safranin O Fast green showed intense staining of red color ( Fig. 1B ), indicating high GAG production in this group compared with control group ( Fig. 1C ). 10. 7717/peerj. 1650/fig-1 Figure 1 Safranin O Fast green staining of MSCs. Images showing GAG production in Pellet culture (A), Alginate beads (B) and Monolayer (C) for 21 days (10 X). Immunohistochemistry for Collagen type II and aggrecan Monolayer, pellet, and alginate bead cultures showed positive for Collagen type II staining on 21 days of culture ( Figs. 2A – 2C ), compared with the negative controls ( Figs. 2G – 2I ). Immunochemistry of aggrecan showed positive for pellet and alginate bead cultures ( Figs. 2D – 2E ); however, no positive staining was observed in monolayer ( Fig. 2F ) compared with the negative control ( Figs. 2J – 2L ). 10. 7717/peerj. 1650/fig-2 Figure 2 Immunohistochemistry of Collagen II and Aggrecan. Immunohistochemistry of Collagen II (A)–(C) and Aggrecan (D)–(F) in Alginate bead, pellet culture and monolayer after 21 days of culture in chondrogenic medium. (G)–(I) and (J)–(L) represents negative control for Collagen II and Aggrecan. Glycosaminoglycan/DNA quantification Figure 3 represents the GAG/DNA content of alginate bead and pellet cultures. The results indicate that hMSCs differentiated in alginate beads has produced significantly higher ( p < 0. 05) glycosaminoglycan than cells in the pellet culture system. The increase in GAG content was also significant over time, day 3 to day 12 and day 12 to day 21, in both alginate bead and pellet culture. 10. 7717/peerj. 1650/fig-3 Figure 3 Comparison of glycosaminoglycan content from alginate bead and pellet culture system. GAG content in the alginate bead and pellet culture system was normalized to respective DNA content. Data shown are mean ± SD. ∗ − p < 0. 05. Chondrogenic gene expression Expressions of chondrogenic genes, Collagen II, Sox 9, and aggrecan, were shown in Figs. 4A – 4C, respectively. In all time points (days 3, 12, and 21), alginate culture represents significantly higher ( p < 0. 05) chondrogenic gene expression compared with monolayer and pellet cultures. In alginate group, there was a significant increase between days 3 and 12 and no significant increase was observed between day 12 and 21 for Collagen II and Sox 9, whereas no significant increase was observed between days 12 and 21 for aggrecan. For pellet culture, there was a significant increase from days 3 to 21 for Collagen II and Sox 9 and expression of Sox 9 and aggrecan remains non-significant from day 12 to 21. 10. 7717/peerj. 1650/fig-4 Figure 4 Expression of Chondrogenic genes. Gene expression of Collagen II (A), Sox 9 (B) and Aggrecan (C) in alginate, pellet and monolayer cultures at day 3, day 12 and day 21. Hypertrophic gene expression The expressions of hypertrophic genes Collagen X and Runx2 are shown in Figs. 5A and 5B, respectively. The expressions of these genes were significantly higher ( p < 0. 05) in alginate beads at the earlier time point (day 3) compared with pellet culture; on the other hand, these genes were found to be significantly down regulated in the later time points of days 12 and 21. At day 3 no significant difference was observed between the expression of these genes between monolayer and alginate bead culture. In pellet culture, although the expressions of hypertrophic genes were significantly lower on day 3, the expression level was significantly increased over time. 10. 7717/peerj. 1650/fig-5 Figure 5 Expression of Hypertrophic genes. Gene expression of Collagen X (A) and Run X2 (B) in alginate, pellet and monolayer cultures at day 3, day 12 and day 21. Cell adhesion molecule expression N-CAM1 and N-cadherin ( Figs. 6A and 6B ) remained downregulated on days 3 and 12 in pellet culture and upregulated on day 21, but in alginate group, N-CAM1 was upregulated over time from days 3 to 21, and for alginate and pellet cultures, N-cadherin remained downregulated on days 3 and 12 but is upregulated on day 21. In monolayer, the expression of N-CAM1 was not significant on days 3 and 12, whereas it was seen upregulated on day 21 and N-cadherin was increased significantly over time. 10. 7717/peerj. 1650/fig-6 Figure 6 Expression of cell adhesion molecules. Gene expression of N-CAM1 (A) and N-Cadherin (B) in alginate, pellet and monolayer cultures at day 3, day 12 and day 21. TEM of MSCs in pellet and alginate cultures Resin-embedded samples of pellet and alginate cultures on day 21 and MSCs on day 0 were cut into semithin 1-µm sections, stained with toluidine blue, and studied under light microscopy ( Fig. 7 ). Ultrathin sections of 70 nm thickness on copper grids were cut, stained with uranyl acetate and lead citrate, and studied with electron microscopy ( Figs. 8 and 15 ). 10. 7717/peerj. 1650/fig-7 Figure 7 Toluidine blue staining of semi-thin sections (1 µm) of Alginate bead (B) and pellet culture (C) after 21 days of culture in chondrogenic medium and MSCs in day 0 (A). 20X. 10. 7717/peerj. 1650/fig-8 Figure 8 Transmission electron microscope images of mesenchymal stromal cells at day 0. TEM images of MSCs day 0. (A) MSC with a convoluted nucleus and prominent nucleoli (arrow) 1151X. (B) Blebs in the cell surface (arrow) 1600X N, Nucleus. (C) MSC with a round nucleus and uneven cell surface 1233X. (D) Higher magnification of MSCs shows mitochondria (M) and rough endoplasmic reticulum (rER) 4000X. The prominent features of semithin sections of alginate beads compared with pellet culture were lower cellularity and higher intercellular spaces ( Fig. 7B ). In alginate culture, cells were arranged in small groups of two or three cells per group and the matrix between them was stained purple using toluidine blue ( Fig. 7B ), whereas in pellet culture, cells were in closer contact with each other and less purple color indicated lower extracellular deposition ( Fig. 7C ). The presence of abundance of polyanions in the ECM of cartilage gives a purple color to the metachromatic dye such as toluidine blue. Samples of MSCs on day 0 as a control group stained blue, whereas samples of chondrogenic differnetiated mesenchymal stromal cells (CMSCs) in alginate and pellet cultures showed positive metachromatic areas (purple color). However, no difference was observed between alginate and pellet culture and this might be due to the limitation of ultra-thin sections used. Ultrathin sections of pellet and alginate beads containing chondrogenic MSCs In monolayer, MSC surface formed filopodia or blebs associated with dense bodies ( Figs. 8B and 8C ) and cytoplasm occupied with rER, free ribosomes, mitochondria (M), and vacuoles ( Fig. 8 ). Nuclei contains euchromatin, usually with a distinct nucleolus, demonstrating an active protein-synthesizing cell; in some cells, the nucleus was convoluted ( Fricker et al. , 1997 ) ( Fig. 8A ). In alginate bead culture, ultrastructure of nucleus in CMSC showed euchromatin with a variety of phenotypes, round, oval, or slightly indented ( Figs. 9 and 12 ). Chondrogenic alginate shows active protein synthesis, with euchromatic nuclei, prominent nucleoli, and abundant rER filled with electron lucent materials producing abundant ECM, containing collagen fibers ( Figs. 9, 10 and 12 ) and molecules with side branches similar to proteoglycan aggregates among fibers ( Fig. 10B ). Chondrogenic differentiated cells are arranged in small groups as they were seen in lower magnification in semithin ( Fig. 7B ) and ultrathin sections ( Figs. 9B and 12A ). Higher magnification of adjacent cells did not show any junctional complexes between the cells. Smaller vesicles contained electron lucent material near cell membrane ( Fig. 12A ) that might originate from rER releasing their content to ECM, causing the cells grow apart from each other, similar to that seen during the interstitial growth in cartilage. The ECM was rich in collagen fibers with banding patterns ( Fig. 10 ). In active protein synthesis, chondrogenic cells in alginate were difficult to distinguish between distended rER and Golgi apparatus ( Fig. 9A ); however, in higher magnification, the rER can be easily recognized with studded ribosomes ( Fig. 11B, black arrow). 10. 7717/peerj. 1650/fig-9 Figure 9 TEM images of MSCs differentiated in alginate bead after 21 days in chondrogenic medium. (A) The cytoplasm is filled with distended RER, Golgi apparatus in the supra-nuclear region (G) 2000X. (B) Three cells similar to isogenic groups in cartilage deposit their products in the 2520X. (C) A cell undergoing cell death with no clear nucleus and abundant cytoplasmic vesicles (V) and cytoplasmic inclusions or multilayer whorled membrane (MW) 1575X. (D) A cell with a euchormatic nucleus, rER and mitochondria (M) 1575X. N, Nucleus; RER, Rough endoplasmic reticulum; ECM, Extracellular matrix. 10. 7717/peerj. 1650/fig-10 Figure 10 TEM images of extra cellular matrix of MSCs differentiated in alginate bead after 21 days in chondrogenic medium. (A) Extracellular matrix in CMSC. ECM, Extracellular matrix; C, Cell 6300X. (B) Higher magnification shows striation of collagen fibres (arrow) and branched molecules probably Aggrecan (asterisk) 10000X. 10. 7717/peerj. 1650/fig-11 Figure 11 TEM image of perinuclear cytosol of MSCs differentiated in alginate bead after 21 days in chondrogenic medium. (A) A part of nucleus (N) and cytoplasm containing rough endoplasmic reticulum (rER) and Mitochondria (M) 6300X. (B) High magnification of cytoplasm. G, Golgi, black arrow indicates ribosomes, Red arrow shows microtubules, rER, Rough endoplasmic cytoplasm 10000X. 10. 7717/peerj. 1650/fig-12 Figure 12 TEM image of divided cells in alginate bead after 21 days in chondrogenic medium. (A) Two daughter cells resulted of cell division 1260X. (B) Higher magnification of the inset shows ECM (asterisk) in the inter-cellular space. Note small vesicles near cytoplasmic membrane (arrows), rER, rough endoplasmic reticulum 6300X. Two subpopulations of cells can be distinguished in ultrathin sections of pellet culture; In one group, signs of cell death appeared with abundant cytoplasmic vesicles, lipid droplets, free ribosomes, swelled or fused mitochondria ( Fig. 13A ), vacuoles, expelling of cytoplasmic organelles to ECM ( Fig. 13B ), indented nucleus ( Fig. 13D ), or cells without a prominent nucleus ( Fig. 13C ). The second population consisted of active protein-synthesizing cells with a euchromatin and round nucleus and abundance of collagen fibers secreted in the ECM ( Figs. 14 and 15 ). Golgi apparatus can be observed in both groups ( Figs. 13D and 14A ) as machinery for synthesis of carbohydrates ( Alberts et al. , 2002 ) or GAGs in chondrogenic-induced MSCs. 10. 7717/peerj. 1650/fig-13 Figure 13 TEM images representing cell death in pellet culture after 21 days in chondrogenic medium. (A) Increase of lipid droplet, N, nucleus; MW, Multivesicular membrane; L, lipid droplet; M, mitochondria; F, cytoplasmic fibrils, 1600X. (B) Fused mitochondria (M), expelled cell organelles including mitochondria (arrow) 2520X. (C) No clear nucleus, 1000X. (D) Cell with U-shape nucleus, free ribosomes (FR) in cytoplasm, Golgi apparatus (G), and lipid droplets (L) 1984X. 10. 7717/peerj. 1650/fig-14 Figure 14 TEM image of MSCs in pellet culture after 21 days in chondrogenic medium. (A) Well-defined Golgi apparatus and Centriole (arrow) in the perinuclear cytosol 4000X. (B) Collagen fibers (F) secreted to the ECM directly from the cytoplasm (white arrow) 3969X. 10. 7717/peerj. 1650/fig-15 Figure 15 TEM images of extracellular matrix of MSCs in pellet culture after 21 days in chondrogenic medium. (A) Collagen fibres (F) in the ECM 2500X. (B) Higher magnification 10000X shows striation (arrow) on fibres. Discussion Gene expression relating to chondrogenic markers, which include Collagen type II and Sox 9, and aggrecan in alginate beads in this study was found to be higher than that of pellet or monolayer culture. This is due to the fact that newly formed chondrogenic construct in alginate culture resembles an immature fetal cartilage in which anabolic activity surpluses the catabolic activity in chondrocytes and results in a higher production of ECM, even when compared with an adult cartilage tissue ( Aigner, Soeder & Haag, 2006 ). These results suggest that chondrogenic differentiation of MSCs in 3D culture, i. e. , alginate, is superior as compared with pellet and 2D monolayer cultures. Our results are consistent with those of previous researches ( Yang et al. , 2004 ) but contradict others ( Ichinose et al. , 2005 ). In several studies, chondrogenic differentiation of MSCs in alginate beads induced with TGF- β 3 calcification was reported. In our study, hypertrophic genes Runx2 and Col X were downregulated in alginate group but they were upregulated in pellet and monolayer cultures using similar culture medium. Expression of transcription factor Sox 9 is accompanied by the expression of chondrogenic gene, such as Col2a1, and aggrecan ( De Crombrugghe et al. , 2000 ). In this study, increase of expression of Sox 9 overlaps with the expression of Col 2 and aggrecan genes in 3D chondrogenic systems. However, in monolayer cultures, it was not accompanied by an increase in aggrecan. In this experiment, expression of Sox 9 remained low in monolayer and chondrogenic pellet culture models, whereas the expressions of Col X and Runx2 were obvious in these groups. Considering the fact that Sox 9 inhibits hypertrophic changes in chondrocytes ( Murakami et al. , 2004 ; Ikegami et al. , 2011 ; Dy et al. , 2012 ) in this study, higher expression of Sox 9 in alginate culture may have protected the cells from being differentiated to a hypertrophic state. Cell adhesion molecules N-CAM1 and N-cadherin were found to be downregulated in 3D cultures of alginate beads and pellet cultures on day 3. These results are consistent with downregulation of these genes during embryonic chondrogenesis ( Widelitz et al. , 1993 ; Tavella et al. , 1994 ), in which N-CAM1 and N-cadherin downregulated after the expression of chondrocyte-specific genes. On the other hand, during chondrogenic differentiation of MSCs in periosteal membranous bone in craniofacial skeleton such as avian quadratojugal joint (an equivalent to mammalian mandibular condylar cartilage), it was shown that N-CAM was not necessary before chondrogenesis ( Fang & Hall, 1999 ). MSCs in gel-like biomaterials, such as Col I, fibrin glue, Matrigel, and PuraMatrix peptide hydrogel, underwent proper chondrogenesis without a direct cell–cell communication ( Dickhut et al. , 2008 ). Therefore, it seems that direct cell–cell interactions both in vitro and in vivo are not always necessary for chondrogenic differentiation ( Boeuf & Richter, 2008 ). However, indirect paracrine communication between the cells in the gel-like material might play a role as it has been shown that MSCs express cytokines and growth factors ( Kim et al. , 2005 ). N-cadherin remained downregulated in both groups of pellet and alginate cultures until third week of differentiation and was later expressed in both groups on day 21. In a limb bud experiment, it was shown that N-cadherin was expressed before chondrogenesis in condensed MSCs and then disappeared from the center of condensation while the cells continued their differentiation and then reexpressed in perichondrium, an indication of appositional growth ( Oberlender & Tuan, 1994 ). As a comparison, the expression of N-cadherin in our experiment on day 21 can be justified with appositional growth of chondrogenic model of pellet in which the peripheral undifferentiated layer of cells behave as perichondrium ( Hillel et al. , 2010 ) or fibroblast ( Alberts et al. , 2002 ). TEM analysis revealed prevalence of cell death in pellet culture, as compared with alginate culture. It can be due to an immature hypertrophy and cell death, as it may happen during endochondral ossification ( Adams & Shapiro, 2002 ; Mackie et al. , 2008 ). The type of cell death can be nonapoptotic or physiologic cell death because it lacked the whole characterization of apoptotic cells, such as crescent heterochromatin nucleus ( Zamli & Sharif, 2011 ). However, there is a possibility that distinctive morphology of apoptotic nucleus is missing in our experiment due to its occurrence in an earlier time point. The morphology and abundance of mitochondria in MSCs and alginate culture were found to be similar ( Figs. 8 and 11 ). Hypoxic microenvironment in cartilage may have caused lower number of mitochondria in articular cartilage as compared with metabolically active cells because chondrocytes rely on glycolytic metabolism rather than oxidative phosphorylation ( Milner, Wilkins & Gibson, 2012 ). In this study, we did not perform any quantitative method for comparing mitochondria in MSC and CMSC; however, the qualitative assessment using pictures did not show any difference in mitochondrial density between undifferentiated MSCs and chondrogenic MSCs. This may be because culture conditions at ambient oxygen of 20% do not provide similar conditions to normal habitat of chondrocyte in cartilage, which is usually low at an oxygen concentration of 2–10% ( Zhou, Cui & Urban, 2004 ). It has been shown that chondrocytes were isolated from joint and cultured in vitro expressed mitochondrial biosynthesis ( Milner, Wilkins & Gibson, 2012 ). Further studies using immunofluorescence for detecting mitochondria can verify mitochondrial quantities during in vitro chondrogenic differentiation of MSCs. The swollen and fused mitochondria from the pellet culture might be due to hypertrophy and cell death. Mitochondrial fusion was shown to be accompanied with MSCs undergoing cell death. Fusion of mitochondria is described as a reaction of the cells to damaged mitochondria to repair by intermixing DNA and protein between mitochondria during damage or senescence ( Chan, 2006 ). Swollen mitochondria were also reported in fibrillated cartilage in OA patients ( Roy & Meachim, 1968 ). The morphology of nucleus in alginate and pellet cultures varied. It appears that cells are more elongated and indented in pellet, whereas in alginate, cells were mostly oval and spherical. Although a typical chondrocyte may have a spherical nucleus, previous ultrastructural studies demonstrated that variation in shapes including oval, spherical, elongated, or indented in human articular cartilage is common, which appears to be acceptable for chondrocyte morphology ( Roy & Meachim, 1968 ). Although the study conducted was well designed, several limitations need to be mentioned to ensure that the findings of the study are not overstated. One of the limitations is that of the time points used for the investigation. The study is only limited to three weeks, which may not be sufficient to demonstrate hypertrophy in our alginate cultures because there has been at least one study that demonstrated that hypertrophy may need up to 45 days of culture ( Dickhut et al. , 2008 ). The second limitation of the study is the fact that the study was limited to gene expression and selected protein investigations. The use of other investigative tools, which includes Western blotting, or protein profile of the constructs would have strengthened the paper further. This is especially true for hypertrophic markers, such as Collagen X and Runx2 and adhesion molecules, such as N-CAM1 and N-cadherin. Conclusion The present study suggests that alginate bead culture provides superior chondrogenic differentiation while decreasing the hypertrophic markers as compared with pellet and monolayer cultures. Alginate bead cultures resemble the articulate cartilage model, whereas pellet cultures resemble the endochondral ossification; therefore, alginate culture would be more suited as an articular cartilage model, whereas the pellet culture would be more appropriate for hypertrophic model of chondrogenesis. Supplemental Information 10. 7717/peerj. 1650/supp-1 Data S1 Raw data Click here for additional data file.

10. 7717/peerj. 16856

PeerJ

Electrical stimulation promoting the angiogenesis in diabetic rat perforator flap through attenuating oxidative stress-mediated inflammation and apoptosis

Background Skin flap transplantation is one of the effective methods to treat the diabetes-related foot ulceration, but the intrinsic damage to vessels in diabetes mellitus (DM) leads to the necrosis of skin flaps. Therefore, the discovery of a non-invasive and effective approach for promoting the survival of flaps is of the utmost importance. Electrical stimulation (ES) promotes angiogenesis and increases the proliferation, migration, and elongation of endothelial cells, thus being a potential effective method to improve flap survival. Objective The purpose of this study was to elucidate the mechanism used by ES to effectively restore the impaired function of endothelial cells caused by diabetes. Methods A total of 79 adult male Sprague-Dawley rats were used in this study. Gene and protein expression was assessed by PCR and western blotting, respectively. Immunohistochemistry and hematoxylin-eosin staining were performed to evaluate the morphology and density of the microvessels in the flap. Results The optimal duration for preconditioning the flap with ES was 7 days. The flap survival area percentage and microvessels density in the DMES group were markedly increased compared to the DM group. VEGF, MMP2, and MMP9 protein expression was significantly upregulated. ROS intensity was significantly decreased and GSH concentration was increased. The expression of IL-1β, MCP‑1, cleaved caspase-3, and Bax were downregulated in the DMES group, while TGF-β expression was upregulated. Conclusions ES improves the angiogenesis in diabetic ischemic skin flaps by attenuating oxidative stress–mediated inflammation and apoptosis, eventually increasing their viability.

Introduction Diabetes mellitus (DM) is a common chronic metabolic disease characterized by hyperglycemia ( Tomic, Shaw & Magliano, 2022 ). DM is linked to a substantial risk of foot complications, such as ulceration, gangrene, and consequent amputation. Amputation due to diabetes-related foot ulceration represents a significant majority (up to 85%) of non-traumatic amputations, with an annual incidence of ulceration of approximately 2% and a lifetime incidence of 34% ( Armstrong, Boulton & Bus, 2017 ; Tomic, Shaw & Magliano, 2022 ). Currently, skin flap transplantation is considered one of the effective methods to treat ulcers ( Lee et al. , 2014 ). Kim et al. (2021) reported that the incidence of flap necrosis used to repair wounds in diabetic patients (50%) is significantly greater than in non-diabetic patients (13. 3%). Additionally, Demiri et al. (2020) demonstrated that the necrotic rate of the ischemic skin flaps in diabetic patients reaches 35. 2%. Despite preoperative measures such as vessel skeletonization and flap viability assessment, flap necrosis is still the prevailing and consequent complication of ischemic skin flaps under diabetes ( Demiri et al. , 2020 ; Kim et al. , 2021 ). Diabetes-related metabolic dysfunction leads to the overproduction of mitochondrial superoxide and other reactive oxygen species (ROS) in the endothelial cells (ECs) of both macro- and microvasculature ( Giacco & Brownlee, 2010 ; Kohnert, Freyse & Salzsieder, 2012 ). The hyperglycemic state associated with diabetes leads to an increase in the levels of markers indicating DNA damage induced by oxidative stress, such as 8-hydroxy-2’-deoxyguanosine (8-OHdG) ( Oguntibeju, 2019 ). Hyperglycemia-induced oxidative stress increases the expression of pro-inflammatory cytokines, with infiltrating macrophages releasing inflammatory cytokines that subsequently trigger local and systemic inflammation ( Wellen & Hotamisligil, 2005 ). Increased intracellular ROS results in decreased angiogenesis under ischemia, various proinflammatory pathways are activated, and cell apoptosis increases, ultimately resulting in vascular damage ( Giacco & Brownlee, 2010 ; Vecchié et al. , 2019 ; Lee, Yun & Ko, 2022 ). Furthermore, increased oxidative stress and enhanced inflammatory responses are contributing factors of the development of diabetic vascular dysfunction, particularly in relation to microvascular complications ( Shi & Vanhoutte, 2017 ). Additionally, DM negatively impacts endothelial cell function, resulting in a decrease in the expression of vascular endothelial growth factor (VEGF) in the skin ( Costa et al. , 2013 ; Shi & Vanhoutte, 2017 ). Diabetic individuals suffer from vasculopathy, characterized by both macrovascular and microvascular injuries, leading to reduced blood perfusion in the flaps and an increased necrotic area ( Kim et al. , 2021 ). Thus, the primary cause for the high rate of necrosis in ischemic flaps among diabetic patients may be attributed to the intrinsic damage to vessels in DM ( Kim et al. , 2021 ). To date, surgical flap delay remains the most effective means of enhancing skin flap survival area, but it is an invasive two-step procedure associated with several complications ( Gersch et al. , 2017 ; Li et al. , 2019 ). Other techniques are available to enhance the viability of the skin flap, such as venous super-drainage ( Wang et al. , 2020a ; Zhu et al. , 2023 ), and arterial supercharging ( Fang et al. , 2020 ; Wang et al. , 2020a ), but they come with disadvantages such as significant tissue trauma, prolonged surgical duration, and the requirement for advanced microvascular anastomosis skills. Pharmacologic preconditioning, also known as chemical delay, increases the survival area of the skin flap. However, it is important to note that chemical drugs are frequently toxic and may induce other undesirable reactions ( Temiz et al. , 2016 ; Chen et al. , 2019b ). Despite attempts to enhance angiogenesis through pharmacological means to improve flap survival, no effective drug for surgical flap delay has yet been discovered to regulate this complex process ( Doğan & Özyazgan, 2015 ). Thus, the search of a non-invasive and effective method for managing ischemic necrosis in diabetic ischemic flap transplantation has emerged as a pressing concern in the field of flap surgery. Electrical stimulation (ES) is considered a promising approach for promoting tissue regeneration ( da Silva et al. , 2020 ), such as bone tissue engineering treatments ( Mobini et al. , 2017 ; Leppik et al. , 2020 ), ES-induced vascularization in tissue ( Wang & Meng, 2023 ), and wound healing ( Rabbani et al. , 2023 ). In addition, ES is extensively used in clinical settings as a rehabilitation method, effectively mitigating muscle atrophy and alleviating pain ( Allen et al. , 2023 ). Previous study demonstrated that ES in rat ischemic limb contributes to a significant increase in the expression of vascular endothelial growth factor (VEGF) and an evident enhancement in capillary density (angiogenesis) in the stimulated muscle ( Kanno et al. , 1999 ). Jeong et al. (2017) observed that ES accelerates the formation of capillaries and arterioles in the ischemic area of athymic mice hindlimb, leading to a reduced muscle necrosis and fibrosis, ultimately inhibiting the necrosis of the ischemic hindlimb. It is worth mentioning that vascular ECs play a crucial role in the process of angiogenesis. Previous in vitro studies showed that ES (electric field: 150–400 mV/mm) enhances the production of VEGF and membrane metalloproteinases ( Tzoneva et al. , 2016 ). Additionally, ES induces favorable physical changes in ECs, including increased cell proliferation, elongation, altered cell shape, reorientation of the long axis of the cell, alignment, and directional cell migration ( Chen et al. , 2019a ; Cunha, Rajnicek & McCaig, 2019 ; Geng et al. , 2019 ; Luo et al. , 2021 ). These alterations are essential prerequisites for the occurrence of angiogenesis. Furthermore, Long et al. (2019) performed a study using a rat model of middle cerebral artery occlusion ischemia-reperfusion injury, and they observed an evident decrease in the level of malondialdehyde (MDA) in the ES group. Conversely, the activity of glutathione (GSH) and superoxide dismutase (SOD) was significantly increased in comparison to that in the middle cerebral artery occlusion group ( Long et al. , 2019 ). Although Doğan & Özyazgan (2015) demonstrated that the use of ES in the normal area prior to skin flap elevation increases the blood flow following the skin flap elevation and increase the skin flap viability compared to that in the control group, but their findings did not provide sufficient evidence to elucidate the underlying mechanism of increased vascularity. On this basis, our research focused on ischemic skin flaps of diabetes, and aimed to elucidate the mechanism used by ES to effectively restore the impaired function of endothelial cells caused by diabetes, as well as to explore strategies for enhancing angiogenesis in the perforator flap. We found that ES promoted angiogenesis in diabetic rat perforator flaps by attenuating oxidative stress-mediated inflammation and apoptosis, increasing the viability of the multi-territory perforator flap. Materials and Methods Animal and study protocols This study was approved by the research ethics committee of the Second Hospital of Shandong University (KYLL-2019(KJ)A-0203). The experiments were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. In addition, this research also followed the “Animal Research: Reporting In Vivo Experiments” (ARRIVE) Guidelines. A total of 79 healthy adult male Sprague–Dawley rats (weight: 300–350 g) were purchased from Jinan Pengyue Experimental Animal Co. , Ltd (Jinan, China). Each cage contained one rat, rats were housed under standard conditions (humidity: 45–55%, temperature: 23–25 °C, and 12 h light/dark cycle) and fed with food and tap water at libitum. The feces were removed from the cages every 3 days. All rats were anesthetized before the experiment using 2% pentobarbital sodium (40 mg/kg, intraperitoneal injection) after isoflurane induction. All rats were euthanized by an intraperitoneal injection of pentobarbital sodium overdose (150 mg/kg) at the end of the experiment. The rats with the direct anastomotic vessels in the flap donor site were excluded prior to the experiment. This study was performed in two parts. Part one consisted of the evaluation of the optimal duration of the ES preconditioning flap, performed using 28 normal rats randomly divided into the four groups: control group (no ES), ES 3-day group, ES 5-day group, and ES 7-day group. Part two consisted of the use of ES to improve the multi-territory flap viability by the induction of angiogenesis in diabetic rats, performed using 51 rats randomly divided into three groups: 17 normal rats used as control group (control: no ES), 17 diabetic rats as the DM group (DM, no ES), and 17 diabetic rats pretreated with ES used as the DM ES group (DMES). Flap viability, histological analysis and molecular biology analysis were performed in each group. Animal model Vascular anatomy : according to a previous report ( Li et al. , 2019 ), the flap can be classified into five zones: deep circumflex iliac (DCI) artery angiosome, proximal choke zone (PCZ), posterior intercostal (PIC) artery angiosome, distal choke zone (DCZ), and thoracodorsal (TD) artery angiosome from the caudal to the cranial part ( Fig. 1A ). Based on the anatomical theory between adjacent angiosomes ( Cormack & Lamberty, 1984 ), the DCI is an anatomical territory, the PIC is a dynamic territory, and the TD is a potential territory in this flap. Thus, DCZ was selected for histological and molecular biology examination. 10. 7717/peerj. 16856/fig-1 Figure 1 Flap model, and scheme of the electrostimulation equipment. (A) Vascular anatomy: the perforated vessels of the three angiosomes were observed following the flap elevation, and three angiosomes and two choke zones were present in the dorsal flap, including the deep circumflex iliac (DCI), proximal choke zone (PCZ), posterior intercostal (PIC), distal choke zone (DCZ), and thoracodorsal (TD) vessels. (B) Flap design. (C) The flaps were raised based on the DCI artery perforator. (D) Scheme of the electrostimulation equipment: the equipment comprises two electrode stickers, each measuring 4. 0 × 4. 0 cm, strategically affixed to the PCZ and DCZ of the flap, respectively. Flap design : after shaving, a rectangular flap of approximately 11 cm × 3 cm in dimension over the unilateral dorsum of the rat was designed based on the DCI artery perforator vessels ( Fig. 1B ). The caudal demarcation was situated at the superior edge of the musculus gluteus maximus and its cranial demarcation at the 7 th cervical spinous process; the medial demarcation of the flap was located at the midline of the spine. Flap harvest : all surgical procedures were performed under standard sterile conditions. The skin flap was undermined above the panniculus carnosus, and the perforators of the TD and PIC arteries were ligated and cut off. The flaps were raised based on the DCI artery perforator vessels ( Fig. 1C ) and then were sutured in situ using 4–0 nonabsorbable suture. Generation of the DM rats Sprague–Dawley rats were fed with tailor-made high-sugar and high-fat fodder for 4 weeks. Afterwards, streptozotocin (50 mg/kg) dissolved in citrate buffer (pH = 4. 5) was intraperitoneally injected for 2 days. Blood samples were collected form the tail vein to measure the random plasma glucose. The random plasma glucose value was ≥16. 7 and ≤33. 3 mmol/L, suggesting that the diabetic rat model was successfully constructed ( Tam et al. , 2014 ). A total of 34 diabetes rats were generated. ES After completing the flap design, both electrodes stickers of the electrostimulation equipment (Xiangyu Medical Equipment Co. Ltd. , Anyang, China) placed the PCZ and DCZ of the flap respectively ( Fig. 1D ). ES was performed under isoflurane anesthesia for 40 min daily ( Fig. 1D ; Frequency: 1, 000 Hz, current intensity: 10 mA). Flap viability The assessment of flap survival area was performed on day 7 after the surgery. Images of the skin flaps were captured using a digital camera (Canon EOS800D; Canon, Tokyo, Japan), and Adobe Photoshop CS6 imaging analysis was used to determine the survival area. This measurement was expressed as a percentage by dividing the survival area by the total area of the flap and multiplying by 100%. Histological analysis The skin flap was raised at the scheduled time and flattened on the wooden board under a standard condition to maintain its accurate size. The flap was photographed using a digital camera (Canon EOS800D; Canon, Tokyo, Japan) to allow the vascular morphological analysis, using a plastic rule as a reference. A tissue specimen measuring 3 × 0. 5 cm was obtained from the DCZ for the purpose of histological analysis. The specimen was fixed in a solution of neutral buffered formalin for a duration of 24 h, after which it underwent dehydration. Subsequently, the tissue was embedded in paraffin and sectioned into slices measuring 3 μm in thickness. Hematoxylin and eosin (H&E) staining, as well as immunohistochemical (IHC) staining, were conducted. The stained sections were then captured using the Nanozoomer Digital Pathology (NDP) scanner S60 (NDP Scan C13210-01; Hamamatsu Photonics K. K. , Hamamatsu, Japan) in order to convert the images into the NDP format. The quantification of vessels with a diameter (D) greater than 0. 1 mm (circumference [C] exceeding 0. 314 mm; C = πD) in the subdermis and muscle layers was conducted through the utilization of H&E staining. The measurement of vascular circumference was accomplished using the NDP view software on the NDP image. Microvessels density (N/mm2) in the subdermis was assessed through IHC staining of CD31 (rabbit anti-rat CD31, #: ab281583, 1:100 dilution; Abcam). Any aggregation of brown endothelial cells was regarded as a single quantifiable microvessel. Five randomly chosen high power fields (20× magnification, 0. 85 × 0. 48 mm/field) were examined, and the microvessel count was determined for each NDP image. DNA damage by oxidative stress was assessed using the IHC staining of 8-OHdG antibody (#ab48508; Abcam, Cambridge, MA, USA; 1:200 dilution). Apoptosis was measured by the IHC staining of cleaved caspase-3 (CC3; #9664; CST, Danvers, MA, USA; 1:1, 000 dilution). Any brown endothelial cell cluster was considered as a single countable positive cell. Five high power fields (20× magnification, 0. 85 × 0. 48 mm/field) including the vessels were randomly selected in the subdermis in each slice, and the number of positive cells and total cells were counted in each field. Western blotting A tissue sample of 3 × 1 cm in size was collected from the DCZ. The sample was lysed using radioimmunoprecipitation assay lysis buffer to extract the total proteins, which were subsequently separated by gel electrophoresis using gels at different concentrations (range, 8 to 15%) and transferred onto a polyvinylidene fluoride membrane. The membrane was incubated with the following primary antibodies at 4 °C overnight: VEGF (#sc-7269, 1:500 dilution; Santa Cruz Biotechnology), matrix metalloproteinase-2 (MMP2: ab92536, 1:1, 000 dilution; Abcam), matrix metalloproteinase-9 (MMP9: ab76003, 1:1, 000 dilution; Abcam), monocyte chemoattractant protein-1 (MCP-1: #AF7437, 1:500 dilution, Beyotime), transforming growth factor β (TGFβ: #AF0297; 1:1, 000 dilution; Beyotime), Bax (#14796, 1:500 dilution; Cell Signaling Technology, Danvers, MA, USA), β-actin (#AC026; 1:5, 000 dilution; ABclonal, Wuhan, China). The membrane was treated with the corresponding secondary antibody and the bands were visualized using the automatic chemiluminescence/fluorescence image analysis system (Tanon 4800). The intensity of the bands was assessed by ImageJ2 software (NIH, Bethesda, MD, USA) and represented as fold change relative to that of β-actin. Reverse transcription-polymerase chain reaction (RT-PCR) Total RNA was extracted from the DCZ tissue. Reverse transcription was performed to obtain cDNA. qPCR was carried out using SYBR® Premix Ex Taq (TaKaRa) and the fluorescence qPCR instrument (ABI QuantStudio 5). Each amplification was performed in triplicate. Relative gene expression was calculated by the 2 –ΔΔCT method, using β-actin as the normalization control. The primer sequences used in this study were the following: interleukin-1β (IL-1β): CCAGGATGAGGACCCAAGCA (forward), TCCCGACCATTGCTGTTTCC (reverse);β-actin:CACCATGTACCCAGGCATTG (forward), TCGTACTCCTGCTTGCTGAT (reverse). Determination of ROS and GSH A tissue sample of 3 × 0. 5 cm in size was isolated from the DCZ and mechanically homogenized. The supernatant was collected to measure the amount of ROS and glutathione peroxidase (GSH) according to the manufacturer’s instructions. Tissue ROS intensity was evaluated using the tissue ROS assay kit (#BB-460512, Bestbio, Shanghai), and expressed as florescence intensity (RFU)/protein concentration (μg/L). Tissue GSH concentration (μmol/g protein) was measured using the reduced GSH assay kit (#A006-2-1; Nanjing Jiancheng Bioengineering). Tissue GSH concentration (μmol/g protein) was calculated as follows = [(Measured OD value- Blank OD value)/(Standard OD value-Blank OD value)] × standard sample concentration (20 μmol/L) × dilution fold (two folds) ÷ protein concentration (g prot/L). Statistical analysis Statistical analysis was performed using GraphPad Prism 9. All measurements were performed by two researchers in a double-blind manner. Results were expressed as mean ± standard deviation. The gaussian distribution was evaluated using the Shapiro-Wilk test. Groups were compared using the t- test or Mann-Whitney U test. A value of P < 0. 05 was considered statistically significant. Results Optimal duration of the ES preconditioning flap The flap survival area percentage was 80. 3% in the control group, 87. 3% in the 3-day group, 88. 2% in the 5-day group, 94. 0% in the 7-day group ( Fig. 2 ; original data: Table S1 ). The flap survival area percentage in the 7-day group was significantly increased in contrast to that in the control group ( Fig. 2, P = 0. 0026). Therefore, the optimal duration of ES preconditioning flap was 7 days, since it corresponded to the maximum survival area ( Fig. 2 ). 10. 7717/peerj. 16856/fig-2 Figure 2 Optimal duration of electrical stimulation (ES) preconditioning flap. Control group: no ES; 3-day group: ES duration for 3 days; 5-day group: ES duration for 5 days; 7-day group: ES duration for 7 days. NS: not significant. ( N = 7 per group). ES increased the diabetes flap viability The flap survival area percentage was 80% in the control group, 72% in the DM group, and 92% in the DMES group ( Fig. 3A ). The statistical analysis revealed that the flap survival area percentage in the DM group was significantly less than that in the control group ( Fig. 3B, P = 0. 0043; original data: Table S2 ), while the flap survival area percentage in the DMES group was remarkably more than that in the control group ( Fig. 3B, P = 0. 0022; original data: Table S2 ) and DM group ( Fig. 3B, P = 0. 0022; original data: Table S2 ). 10. 7717/peerj. 16856/fig-3 Figure 3 Electrical stimulation (ES) significantly improves diabetes ischemic skin flap viability. (A) Flap survival area; Black circle: flap necrotic area. (B) Flap survival area percentage. Control: no ES, DM: diabetes mellitus group (no ES), DMES: diabetes mellitus ES group. The vascular morphological changes showed that the vascular network of the DM group was much sparser than that of the control group, while the vascular network of the DMES group was much denser than that of the control and DM group ( Fig. 4A ). The results of the H&E staining revealed that the number of vessels greater than 0. 1 mm in diameter in the DM group was significantly lower than that in the control group ( Fig. 4B, P = 0. 024; original data: Table S3 ), while the number of vessels greater than 0. 1 mm in diameter in the DMES group was markedly greater than that in the control group ( Fig. 4B, P = 0. 023; original data: Table S3 ) and DM group ( Fig. 4B, P = 0. 003; original data: Table S3 ). IHC staining of CD31 showed that CD31-positive microvessels density in the DM group was less than that in the control group, while CD31-positive microvessels density in the DMES group was more than that in the control and DM group ( Fig. 5 ; original data: Table S4 ). 10. 7717/peerj. 16856/fig-4 Figure 4 Electrical stimulation (ES) significantly increases the number of vessels with a vascular diameter > 0. 1 mm. (A) Changes in vascular network. (B) H&E staining results: number of vessels with a vascular diameter > 0. 1 mm. Control: no ES; DM: diabetes mellitus group (no ES); DMES: diabetes mellitus ES group ( N = 4 rats per group). 10. 7717/peerj. 16856/fig-5 Figure 5 Electrical stimulation (ES) significantly increases the microvascular density. (A–C) Immunohistochemical staining of CD31 in the control (A), diabetes mellitus (B) and diabetes ES (C) group. (D) CD31-positive microvessels density. Control: no ES; DM: diabetes mellitus group (no ES); DMES: diabetes mellitus ES group ( N = 4 rats per group). MMP-2 and MMP-9 are widely considered as the main proteolytic enzymes involved in the degradation of the extracellular matrix in the vascular basement membrane ( Zhang et al. , 2020 ). The proangiogenic factor VEGF activates MMP2 and MMP9, regulating the remodeling of the endothelial extracellular matrix, promoting the degradation of the EC basement membrane, favoring the migration of ECs and creating an advantageous environment to tubule formation ( Zhang et al. , 2020 ). It is worth noting that the expression of MMP-2 and MMP-9 is consistent with the change of neovascularization ( Zhang et al. , 2020 ). Thus, MMP2 and MMP9 were considered as angiogenic markers and measured. Western blotting showed that VEGF protein expression in the DM group was significantly reduced compared to the control group (1. 0 ± 0. 14 vs 0. 42 ± 0. 16, P = 0. 0306; Fig. 6A ), while VEGF protein expression in the DMES group was significantly increased compared to the DM group (1. 3 ± 0. 43 vs 0. 42 ± 0. 16, P = 0. 0029; Fig. 6A ; original band: Fig. S1 ). MMP2 protein expression in the DMES group was significantly upregulated compared to the control group (1. 4 ± 0. 3 vs 1. 0 ± 0. 08, P = 0. 0240; Fig. 6B ; original band: Fig. S1 ) and DM group (1. 4 ± 0. 3 vs 0. 76 ± 0. 1, P = 0. 0018; Fig. 6B ; original band: Fig. S1 ). MMP9 protein expression in the DMES group was much higher than in the control group (2. 13 ± 0. 21 vs 1. 0 ± 0. 45, P = 0. 0014; Fig. 6C ; original band: Fig. S1 ) and DM group (2. 13 ± 0. 21 vs 0. 71 ± 0. 17, P = 0. 0003; Fig. 6C ; original band: Fig. S1 ). These results indicated that DM significantly reduced vascular density and ES increased vascular density by the improvement of angiogenesis. 10. 7717/peerj. 16856/fig-6 Figure 6 Western blotting showing the expression and quantification of the optical density of VEGF (A), MMP2 (B), and MMP9 (C) in the control, DM and DMES group. Gel electrophoresis was performed under the same experimental conditions, and tailored blots are shown. VEGF: vascular endothelial growth factor; MMP2: matrix metalloproteinase-2; MMP9: matrix metalloproteinase-9; Control: no ES; DM: diabetes mellitus group (no ES); DMES: diabetes mellitus ES group ( N = 4 rats per group). * P < 0. 05, ** P < 0. 01, *** P < 0. 001, NS: No signficance. The IHC staining of 8-OHdG was performed to measure the oxidative stress damage in the flap tissue among the control, DM and DMES group since it is a marker for DNA oxidation during oxidative stress ( Wang et al. , 2020b ). The results showed that the percentage of positive 8-OHdG in vascular endothelial cells of the flap tissue in the DM group was markedly increased compared to that in the control and DMES group, while no significant difference was found between the control and DMES group ( Fig. 7A ). In addition, the overproduction of ROS is the main cause of oxidative stress; thus tissue ROS intensity was also evaluated. ROS intensity in the DM group was much higher than that in the control group (1. 8 ± 0. 25 vs 0. 38 ± 0. 01, P < 0. 0001; Fig. 7B ; original data: Table S5 ) and DMES group (1. 8 ± 0. 25 vs 0. 35 ± 0. 04, P < 0. 0001; Fig. 7B ; original data: Table S5 ), while no significant difference was observed between the control and DMES group (0. 38 ± 0. 01 vs 0. 35 ± 0. 04, P = 0. 9733; Fig. 7B ; original data: Table S5 ). GSH is an endogenous antioxidant enzyme that protects the tissue from oxidative damage. GSH concentration in the DMES group was significant higher to that in the control group (3. 5 ± 0. 48 vs 2. 7 ± 0. 12, P = 0. 0480; Fig. 7C ; original data: Table S6 ) and DM group (3. 5 ± 0. 48 vs 1. 5 ± 0. 28, P = 0. 0007; Fig. 7C ; original data: Table S6 ), and GSH concentration in the DM group was much less than that in the control group (1. 5 ± 0. 28 vs 2. 7 ± 0. 12, P = 0. 0114; Fig. 7C ; original data: Table S6 ). The pro-inflammatory cytokines IL-1β and MCP‑1, and the anti-inflammatory cytokine TGFβ were measured. RT-PCR results showed that IL-1β gene expression in the DM group was much higher than that in the control group (2. 0 ± 0. 24 vs 1. 0 ± 0. 37, P = 0. 0427; Fig. 7D ; original data: Table S7 ) and DMES group (2. 0 ± 0. 24 vs 0. 90 ± 0. 51, P = 0. 0341; Fig. 7D ; original data: Table S7 ), while no significant difference was found between the control and DMES group (1. 0 ± 0. 37 vs 0. 90 ± 0. 51, P = 0. 8871; Fig. 7D ; original data: Table S7 ). Western blotting showed that MCP-1 protein expression in the DM group was much higher than that in the control group (2. 38 ± 0. 53 vs 1. 0 ± 0. 31, P = 0. 0021) and DMES group (2. 38 ± 0. 53 vs 1. 40±0. 31, P = 0. 0163), while no significant difference in MCP-1 protein expression was observed between the control and DMES group ( Figs. 7E and 7G ; original band: Fig. S2 ). TGFβ protein expression in the DM group was significantly downregulated compared to the control group (0. 43 ± 0. 16 vs 1. 0 ± 0. 36, P = 0. 0282) and DMES group (0. 43 ± 0. 16 vs 1. 0 ± 0. 19, P = 0. 0276), while no significant difference in TGFβ protein expression was found between the control and DMES group ( Figs. 7F and 7H ; original band: Fig. S2 ). These results indicated that ES attenuated oxidative stress-mediated inflammation. 10. 7717/peerj. 16856/fig-7 Figure 7 Electrical stimulation (ES) attenuates oxidative stress-mediated inflammation. (A) Immunohistochemical staining of 8-hydroxy-2’-deoxyguanosine (8-OHdG); black arrow: positive cells. (B) Reactive oxygen species (ROS) intensity in the tissue ( N = 3 per group). (C) Glutathione (GSH) concentration in the tissue ( N = 3 per group). (D) Interleukin-1β (IL-1β) gene expression ( N = 3 per group). (E and H) Expression and quantification of the optical density of MCP-1 (E and G), and TGFβ (F and H) in the control, DM and DMES group ( N = 4 per group) by western blotting. Gel electrophoresis was performed under the same experimental conditions, and tailored blots were shown. * P < 0. 05, ** P < 0. 01, *** P < 0. 001, **** P < 0. 0001. NS: No signficance. The expression of the pro-apoptotic proteins cleaved caspase-3 and Bax were measured. Caspase-3 is a pivotal regulatory protein of apoptosis and nuclear changes in apoptosis. It is cleaved to form cleaved caspase-3, which regulates the morphological and biochemical alterations in apoptosis ( Silva et al. , 2022 ). The immunohistochemical staining of cleaved caspase-3 showed that the percentage of positive cleaved caspase-3 in vascular endothelial cells of the flap tissue in the DM group was significantly increased compared to that in the control and DMES group, while no significant difference was observed between the control and DMES group ( Fig. 8A ). Western blotting showed that Bax protein expression in the DM group was significantly upregulated compared to the control group (1. 6 ± 0. 32 vs 1. 0 ± 0. 19, P = 0. 0181) and DMES group (1. 6 ± 0. 32 vs 0. 64 ± 0. 42, P = 0. 0142), while no significant difference in Bax protein expression was found between the control and DMES group ( Figs. 8B and 8C ; original band: Fig. S3 ). 10. 7717/peerj. 16856/fig-8 Figure 8 Electrical stimulation (ES) attenuates apoptosis. (A) Immunohistochemical staining of cleaved caspase-3. (B and C) Expression and quantification of the optical density of Bax in the control, DM and DMES group ( N = 4 per group) by western blotting. Gel electrophoresis was performed under the same experimental conditions, and tailored blots were shown. * P < 0. 05, ** P < 0. 01, NS: No signficance. Discussion Our findings demonstrated that ES significantly upregulated the expression of VEGF, MMP2, and MMP9 in the ischemic skin flaps of diabetic patients, increased the microvessels density and the number of vessels with a diameter longer than 0. 1 mm, eventually increasing the survival area percentage of the ischemic skin flaps. Further investigation revealed that ES significantly upregulated the expression of the antioxidant GSH and the anti-inflammatory cytokine TGFβ in the ischemic skin flaps of diabetic patients. Moreover, ES markedly downregulated the expression of the oxidative injury markers ROS, pro-inflammatory cytokines IL-1β and MCP‑1, as well as pro-apoptotic proteins cleaved caspase-3 and Bax. Although Doğan & Özyazgan (2015) demonstrated that the use of ES in the normal area prior to skin flap elevation increases the blood flow following the skin flap elevation and increase the skin flap viability compared to that in the control group, our research focused on ischemic skin flaps of diabetes and reveal the mechanism of ES promoting angiogenesis in this specific pathological microenvironment. ES is a non-invasive and non-pharmacological physical stimulus. At the molecular level, it is able to enhance the transportation of biomolecules, whether charged or uncharged, across biological membranes through the electrophoresis and electroosmosis ( Gratieri, Santer & Kalia, 2017 ). At the cellular level, ES has an effect on a diverse range of cellular components, including ion channels, membrane-bound proteins, cytoskeleton, and intracellular organelles ( Zhao, Mehta & Zhao, 2020 ). The underlying mechanisms responsible for the reaction of the cell to ES are currently being extensively investigated. Several hypotheses have been proposed, including the disruption of structural water, electroosmotic fluid flow, asymmetric ion flow and the activation of voltage-gated channels, mechanosensation, as well as redistribution of membrane components and lipid rafts ( Zhao, Mehta & Zhao, 2020 ). These interactions ultimately lead to alterations in cellular behavior and functions, such as migration, contraction, orientation, and proliferation ( Zhao, Mehta & Zhao, 2020 ). The most remarkable morphological changes in the microvasculature of diabetic individuals are a thickened capillary basement membrane, reduced luminal dimension and number of capillaries, as well as the degeneration of pericytes ( Sharma, Schaper & Rayman, 2020 ). The loss of autoregulation in microvascular blood flow leads to an increase in capillary pressure, which may trigger inflammatory responses within the microvascular endothelium. This results in endothelial injury and subsequent thickening of the capillary basement membrane, arteriolar hyalinosis, and reduced vasodilatory capacity ( Sharma, Schaper & Rayman, 2020 ). In our experiments, the vascular density of the flap of the DM group was significantly reduced compared to that of the control group, and the expression of the angiogenetic protein VEGF was also remarkably downregulated. The positive 8-OHdG rate in vascular endothelial cells and ROS intensity in the DM group was increased. The concentration of GSH, which works as a protective agent against oxidative damage to the tissue, was decreased. The expression of the pro-inflammatory cytokines IL-1β and MCP‑1 was upregulated and the expression of the anti-inflammatory cytokine TGFβ was downregulated. The expression of the pro-apoptotic proteins cleaved caspase-3 and Bax was upregulated. The electron transport chain in the mitochondria is impaired in a chronic high glucose environment, resulting in the generation of ROS, stimulating the increase of proton leakage and altering the mitochondrial membrane potential, subsequently causing the release of cytochrome c, ultimately leading to apoptosis ( Zhang et al. , 2023 ). In short, oxidative stress-mediated inflammation induced by hyperglycemia was increased, thereby exacerbating cellular apoptosis, and reducing angiogenesis, ultimately leading to a decrease in the viability of the multi-territory perforator flap. However, our results showed that the optimal duration for preconditioning the flap with ES was 7 days. ES significantly enhanced the survival area percentage of the diabetic ischemic skin flaps. Vascular density (microvessels density and number of vessels with a diameter longer than 0. 1 mm) significantly increased in the DMES group compared to that in the DM group, and the angiogenesis-related protein expression (VEGF, MMP2, MMP9) was significantly upregulated. Li et al. (2023) performed a study in which they created a set of cell-scale to perform experiments to examine the electro-mechanical coupling phenomenon in human umbilical vein endothelial cells (HUVECs) during angiogenesis. They found that the stimulation by an external electrical field (E ex ) polarizes intracellular calcium ions, generating a rear-to-front concentration gradient in HUVECs, and establishing an internal electric field (E in ) opposed to the E ex. Cells affected by changes in local calcium ion actively contract the cytoskeleton to activate Piezo1 channels, leading to the influx of extracellular calcium ions and gradually establishing a balance between E in and E ex. In addition, they found that the electro-mechanical coupling feedback loop guides pre-angiogenic activities, such as elongation and migration of HUVECs. The external electrical field also promotes the expression of the angiogenesis-related genes VEGF and eNOS in HUVECs. Our experiments showed that the 8-OHdG positive cell rate and the ROS intensity in the DMES group were significantly decreased compared to those in the DM group, and GSH concentration was increased. Previous studies demonstrated that external suitable ES leads to the change in cell membrane potential ( Zhao et al. , 2004 ), resulting in the polarization of the cell, inducing influx of extracellular Ca 2+ ( Emerson & Segal, 2001 ; Li et al. , 2023 ), triggering calcium signal cascade reaction, and reducing the level of the oxidative stress. ES also promotes the entrance of the NFE2-related factor 2 (Nrf2) into the nucleus, thus upregulating the production of the protective factors heme oxygenase 1 (HO-1) and NADPH quinone oxidoreductase 1 (NQO1) through the anti-oxidative stress signaling pathway Kelch-like ECH-associated protein 1 (Keap1)/Nrf2 ( Weng et al. , 2023 ). Sha et al. (2015) concluded that ES significantly enhances the enzyme activity of superoxide dismutase, GSH, and other endogenous oxidation free radical scavenging systems, inhibits lipid peroxidation in tissue cell membranes, reduces the malondialdehyde content, and plays a protective role in tissues. In addition, our results showed that IL-1β and MCP‑1 expression was downregulated, while TGF-β expression was upregulated. ES changes the function of inflammatory cells, leading to a reduction in the activity of extracellular matrix modifier enzyme and matrix metalloproteinase-1, inhibiting the secretion of inflammatory factors, while maintaining unchanged the immune cell count ( Kim et al. , 2019 ; Shin et al. , 2019 ). Thus, ES attenuated oxidative stress-mediated inflammation in ischemic skin flaps of diabetic rats. Interestingly, our results also revealed that the expression of the pro-apoptotic proteins cleaved caspase-3 and Bax in the DMES group was downregulated compared to the DM group, but the exact mechanism is unclear. ES mitigates apoptosis by regulating intracellular protein expression and activating several signaling pathways, such as the MAP kinase pathway, and the ERK signaling pathway ( Zhao et al. , 2021 ). Yang et al. (2014) demonstrated that ES mitigates the apoptosis of ischemic cardiomyocytes in rats by upregulating Bcl-2 gene expression and downregulating Bax gene expression, although the precise mechanism remains unclear. ES also promotes cell survival by inducing the opening of voltage gated calcium channels, which causes Ca 2+ influx into the cell, resulting in the activation of AKT by Ca 2+. Akt induces cell survival by activating the transcription factor NF-KB, which translocates to the nucleus to induce the transcription of pro-survival genes. NF-KB also activates the anti-apoptotic protein Bcl-2 and inhibits the tumor suppressor p53 ( Love et al. , 2018 ). In addition, ES is nothing more than a flow of electrons. This flow of electrons mimics the flow of ions in the biological system, mainly altering the membrane potential (V mem ) of the cells. This alteration in V mem is the root cause of the inhibition of apoptosis. This is the key point explaining why the survival area was increased after application of ES. In addition, the use of medium frequency current offers the advantage of diminishing skin impedance, thereby minimizing the peripheral dissipation of electrical energy and facilitating a deeper penetration into the muscle ( De Oliveira et al. , 2018 ). Also, medium frequency current elicits multiple nerve fiber action potentials per burst, resulting in firing rates that are multiples of the burst frequency ( De Oliveira et al. , 2018 ). Conversely, the application of high frequency current may induce muscle fatigue ( Szecsi & Fornusek, 2014 ). Conclusions ES improves the angiogenesis in ischemic skin flaps of diabetic rats through the reduction of oxidative stress–mediated inflammation and apoptosis, eventually increasing the ischemic skin flap viability. Thus, ES is a non-invasive, inexpensive, and effective method to increase the survival area of ischemic skin flaps in diabetic individuals. Supplemental Information 10. 7717/peerj. 16856/supp-1 Supplemental Information 1 The original Data. Click here for additional data file. 10. 7717/peerj. 16856/supp-2 Supplemental Information 2 Original western blotting band figures. Click here for additional data file. 10. 7717/peerj. 16856/supp-3 Supplemental Information 3 Author Checklist - Full. Click here for additional data file.

10. 7717/peerj. 16897

PeerJ

3D bioprinting in bioremediation: a comprehensive review of principles, applications, and future directions

Bioremediation is experiencing a paradigm shift by integrating three-dimensional (3D) bioprinting. This transformative approach augments the precision and versatility of engineering with the functional capabilities of material science to create environmental restoration strategies. This comprehensive review elucidates the foundational principles of 3D bioprinting technology for bioremediation, its current applications in bioremediation, and the prospective avenues for future research and technological evolution, emphasizing the intersection of additive manufacturing, functionalized biosystems, and environmental remediation; this review delineates how 3D bioprinting can tailor bioremediation apparatus to maximize pollutant degradation and removal. Innovations in biofabrication have yielded bio-based and biodegradable materials conducive to microbial proliferation and pollutant sequestration, thereby addressing contamination and adhering to sustainability precepts. The review presents an in-depth analysis of the application of 3D bioprinted constructs in enhancing bioremediation efforts, exemplifying the synergy between biological systems and engineered solutions. Concurrently, the review critically addresses the inherent challenges of incorporating 3D bioprinted materials into diverse ecological settings, including assessing their environmental impact, durability, and integration into large-scale bioremediation projects. Future perspectives discussed encompass the exploration of novel biocompatible materials, the automation of bioremediation, and the convergence of 3D bioprinting with cutting-edge fields such as nanotechnology and other emerging fields. This article posits 3D bioprinting as a cornerstone of next-generation bioremediation practices, offering scalable, customizable, and potentially greener solutions for reclaiming contaminated environments. Through this review, stakeholders in environmental science, engineering, and technology are provided with a critical appraisal of the current state of 3D bioprinting in bioremediation and its potential to drive forward the efficacy of environmental management practices.

Introduction The United States Environmental Protection Agency (USEPA) reports that as of September 2023, 1, 336 uncontrolled hazardous waste sites are registered on the National Priorities List (NPL) alone. The NPL is a crucial tool for identifying the contaminated sites requiring long-term remedial action through the Superfund program, and inclusion in the NPL is a critical step towards securing federal funding for the extensive cleanup operations required to remediate these hazardous waste sites ( US Environmental Protection Agency (US EPA), 2023 ). Similarly, the Canadian government’s Federal Contaminated Sites Inventory has listed 4, 503 active contaminated sites in Canada as of November 2023 ( Environment and Climate Change Canada (ECCC), 2016 ; Treasury Board of Canada Secretariat, 2023 ). These numbers underscore the ongoing challenge of managing hazardous waste in North America. The European Union, for instance, is grappling with the daunting task of addressing pollution in approximately 2. 8 million potentially affected land sites, as stated by the World Health Organization in July 2023 ( World Health Organization (WHO), 2023 ). Meanwhile, the Global Alliance on Health and Pollution has identified over 5, 000 toxic hotspots worldwide in low- and middle-income countries that require immediate remediation efforts ( Global Alliance on Health and Pollution (GAHP), 2023 ). Therefore, addressing the significant environmental and public health risks posed by hazardous waste and contaminated sites remains an urgent and complex global issue that demands sustained commitment and resources. Traditional remediation methods, such as excavation and incineration, can be expensive, generate hazardous waste, and have limited effectiveness. In contrast, bioremediation utilizes microorganisms or materials of biological origin, such as enzymes, biocomposites, biopolymers, or nanoparticles, to biochemically degrade contaminants into harmless substances, making it an environmentally friendly and cost-effective alternative. Bioremediation is a beacon of environmental sustainability, harnessing the power of biological processes and biomaterials to confront the escalating challenge of anthropogenic pollution. In the age where technological innovation is rapidly reshaping various industries, the field of environmental engineering is experiencing a renaissance with the advent of 3D printing technology, also known as additive manufacturing ( Amorim et al. , 2021 ; Gkantzou, Weinhart & Kara, 2023 ). This convergence can revolutionize bioremediation by offering novel solutions to complex environmental problems. 3D printing technology introduces unparalleled precision and customization to the fabrication of objects, operating under the principle of layer-by-layer construction from digital models. This technology is particularly promising for bioremediation, as it allows for the design and creation of intricate structures tailored to support microbial life or hold materials that are conducive to the removal of pollutants, facilitating the degradation of contaminants in diverse environmental matrices ( Schubert, Van Langeveld & Donoso, 2014 ; Schaffner et al. , 2017 ). The adaptability of 3D printing can be leveraged to enhance the efficiency of bioremediation strategies through the optimization of habitat architecture for microbial communities, thereby accelerating the biodegradation process. This confluence of biotechnology and additive manufacturing holds significant promise for developing innovative bioremediation strategies ( Gross et al. , 2014 ). In recent years, there has been a significant surge of interest in using 3D printing and 3D bioprinting for bioremediation research. This is evident from the exponential increase in publications, with countries like China, USA, India, United Kingdom, Germany, and Spain leading the way ( Fig. 1 ; Elsevier, 2023 ). These nations have invested heavily in advancing additive manufacturing technologies to support the development of cutting-edge bioremediation processes. Recent advancements in 3D printing have introduced materials and techniques specifically tailored for environmental applications. For instance, developing 3D-printed bioreactor media that can be customized to site-specific conditions, thereby maximizing microbial degradation activities, is a poignant illustration of the synergies between these technologies ( Elliott et al. , 2017 ). The high degree of customization enables the fabrication of structures with increased surface areas for microbial growth, optimizing the exposure of pollutants to degradative biofilms. By precisely controlling the spatial arrangement of cells and biomaterials, 3D bioprinting can create bioremediation devices with enhanced cell-cell interactions, improved nutrient and oxygen transport, and a more accurate representation of the physiological microenvironment, significantly enhancing their bioremediation performance ( Chimene et al. , 2016 ). 10. 7717/peerj. 16897/fig-1 Figure 1 Number of publications between 2014–2023 using the combination of keywords “3D Printing” and “Bioremediation, ” “3D Bioprinting” and “Bioremediation” found in Scopus. The number of publications between 2014–2023 using the combination of keywords “3D Printing” and “Bioremediation, ” “3D Bioprinting” and “Bioremediation” (A) and publications per country using the keywords keywords “3D Printing” and “Bioremediation, ” (B) and keywords “Bioprinting” and “Bioremediation, ” (C) found in Scopus (accessed on 10 Nov 2023; Elsevier, 2023 ). This review critically examines the principles, applications, and future directions of 3D printing in bioremediation. By evaluating the current state of research, this article aims to provide insights into the potential environmental benefits and challenges associated with implementing 3D printing technologies in bioremediation. Survey/search methodology To ensure the inclusion of the most relevant and recent advancements, our search methodology encompassed a thorough literature review spanning the last two decades, focusing on publications from the last 5 years. Utilizing databases such as Scopus, Web of Science, PubMed, and Google Scholar, we employed keywords such as “3D Bioprinting, ” “Bioremediation, ” “3D Printing, ” “Environmental Remediation, ” and others as explained below to narrow down bioremediation research that utilized additive manufacturing processes. Priority was given to recent experimental and review articles that directly contribute to the understanding of 3D bioprinting applications in bioremediation, ensuring our review reflects the latest trends and technological developments in this rapidly evolving field. A strategic combination of keywords and Boolean operators was employed to provide a thorough and precise retrieval of pertinent literature. Primary keywords were initially used to identify relevant works and secondary key terms were used to ensure that potentially relevant results were not missed. Conventional search engines such as Google, Bing, and DuckDuckGo were also utilized to ensure recent non-indexed works were also captured. Primary key terms included: (“3D Printing” OR “Additive Manufacturing”) AND “Bioremediation” “3D Bioprinting” AND “Bioremediation” “Environmental Remediation” AND (“3D Printing” OR “Bioprinting”) Examples of secondary key terms included but are not limited to: “Biocarriers” AND “3D Printing” “Enzyme Immobilization” AND (“3D Printing” OR “Bioprinting”) (“3D Printing” OR “Bioprinting”) AND “Heavy Metal Remediation” Inclusion and exclusion criteria In order to maintain scholarly rigor, the following criteria were established: Inclusion Criteria: Peer-reviewed articles published within the last two decades emphasizing the most recent 5 years to capture cutting-edge developments were included. Studies that explicitly discuss the utilization of 3D bioprinting within the scope of environmental bioremediation were prioritized. Additionally, we included papers contributing to the understanding of principles, applications, and prospective trajectories of 3D bioprinting in bioremediation. Exclusion Criteria: Articles outside the realm of peer-reviewed literature were generally excluded, except where they provided unique and critical insights not available in peer-reviewed sources. Studies predating the 20-year window or those diverging from the core focus on bioremediation and 3D bioprinting were omitted. Systematic selection process Our literature search was executed in multiple phases to ensure depth and breadth. Initial searches using broad keywords yielded a diverse collection of articles, which were then scrutinized based on titles and abstracts for relevance. Subsequently, full-text assessments were conducted to ascertain the suitability of these studies against the defined inclusion criteria. To safeguard against selection bias and ensure a holistic perspective, the selection of articles was grounded in their scientific robustness and relevance to the subject matter, irrespective of their specific outcomes or the nature of their findings. Cross-referencing citations within these articles further augmented the breadth of our literature review. This exhaustive and methodically structured approach assured a nuanced and comprehensive review of the existing research landscape of 3D bioprinting in the context of bioremediation. What is bioremediation? Traditionally, bioremediation has encompassed using natural microorganisms or other life forms to accumulate and break down environmental pollutants to clean up contaminated areas. This includes methods such as natural attenuation, bioaugmentation, phytoremediation, and landfarming, among others. More recently, the term has been expanded to include techniques incorporating genetically modified organisms, biomaterials that mimic biological processes, and other customized approaches for environmental remediation. Examples of these techniques include pollutant degradation using genetically engineered organisms, bioventing, in situ bioreactors, and nanobioremediation. Throughout this article, we will use the term bioremediation to refer to any of the above-mentioned methods for environmental remediation. What is 3D bioprinting? 3D bioprinting is an additive manufacturing technology that involves the precise layer-by-layer positioning of biological materials, biochemicals, and living cells to fabricate three-dimensional structures ( Murphy & Atala, 2014 ). This approach provides spatial control of the placement of functional components, enabling the creation of complex and functional constructs. This field of bioprinting is a rapidly developing area that focuses on printing materials of biological origin, commonly referred to as bioinks ( Fu et al. , 2022 ), and while it has traditionally been applied in tissue engineering, the evolution from traditional to modern bioprinting techniques underscores significant technological advancements and a more comprehensive range of potential applications, including bioremediation, as explored in this article. This article will also show how biomaterials have been successfully incorporated into various conventional additive manufacturing technologies such as material extrusion, vat polymerization, powder bed fusion, material jetting, and binder jetting to create unique bioprinting tactics that could be used for bioremediation. Considerations of 3D printing in bioremediation Incorporating 3D printing technology into bioremediation practices offers several advantages that improve the efficacy and efficiency of conventional remediation methods. In this section, we will delve into the fundamental principles that govern the application of 3D printing technology in bioremediation efforts. Design flexibility and customization One of the critical advantages of utilizing 3D printing for bioremediation lies in the ability to create customized structures tailored to suit unique environmental conditions and contaminant profiles. This cutting-edge technology offers unparalleled control over the shape, size, and surface area of the printed objects, thereby enabling the design of intricate features that can enhance microbial colonization and the efficient degradation of pollutants ( Duty et al. , 2017 ). Additive manufacturing is revolutionizing the manufacturing industry by allowing the creation of highly customized designs and enabling rapid prototyping ( Strack, 2019 ). In bioremediation, this technology could facilitate scale-up and mass production while simultaneously reducing costs, lead times, and waste. Additionally, additive manufacturing uses less energy than traditional manufacturing processes, making it a more sustainable solution for various environmental remediation techniques. Material selection and sustainability The use of 3D printing technologies has made it possible to incorporate a wide range of materials, such as biopolymers and recycled plastics, into the printing process. These materials are carefully chosen based on their ability to be printed, compatibility with biological systems, and ability to decompose naturally. By selecting sustainable materials, the 3D-printed structures themselves do not contribute to pollution but rather contribute to the overall remediation process ( Zhang et al. , 2023 ). An example is the inherent challenge in integrating functional bacteria with 3D bioprinting, which lies in achieving a delicate equilibrium between the manufacturability of the material, minimizing damage during the bioprinting procedure, and preserving bacterial activity and function ( Zhao et al. , 2023 ). Optimizing bioink selection, considering 3D printability, microbial and chemical compatibility, and contaminant degradation, is crucial for bioremediation. Compatibility of the materials with microbial systems is, therefore, an essential criterion for material selection, which facilitates successful integration with microbial systems. The design of 3D-printed structures for microbial-based bioremediation is closely aligned with the biological requirements of the microorganisms involved. This involves considering factors such as nutrient flow, aeration, and maintaining optimal environmental conditions for microbial growth and enzymatic activity ( Cao et al. , 2022 ). Using 3D-printed biocarriers has been shown to improve the nitrification efficiency of designed systems by facilitating the growth of sluggish bacterial species, highlighting the importance of ensuring microbial compatibility when considering their effectiveness for remediation purposes ( Noor et al. , 2023 ). It has been established that successful 3D printing of living materials with high performance relies on the development of new ink materials and 3D geometries that promote long-term cell functionality ( Qian et al. , 2019 ). Increased surface area The effectiveness of bioremediation processes is often influenced by the surface area available for microbial activity. 3D printing makes it possible to create porous structures with a favorable surface-to-volume ratio, allowing for significant space for chemical or microbial reactions. This innovative technology also boosts the accessibility of pollutants for degradation on functional surfaces ( Aguirre-Cortés et al. , 2023 ). Bioprinting can provide high surface area per unit volume, lightweight structure, high porosity, and roughness that are essential for the growth of biofilms, a widely used biocarrier for bioremediation purposes ( Sfetsas, Patsatzis & Chioti, 2021 ). Scalability and reproducibility 3D printing technology enables the production of bioremediation tools that can be scaled from small laboratory prototypes to larger structures suitable for field applications. Moreover, the digital nature of 3D printing ensures reproducibility, allowing for the consistent manufacture of bioremediation devices across different locations ( Thompson et al. , 2016 ). Recent research has identified key factors that can be improved to enhance the reproducibility and reliability of bioprinting, which holds great potential for future applications ( Grijalva Garces et al. , 2024 ). Life-cycle and ecological impact assessment A principle that is becoming increasingly significant in the application of 3D printing to bioremediation is the assessment of the life-cycle impact of the printed structures. Evaluating the ecological footprint of these materials from production to degradation is crucial for ensuring that the bioremediation strategy is truly sustainable ( Roy et al. , 2009 ). Life cycle assessment is a comprehensive approach that considers various factors, such as the type and amount of raw materials used, energy consumption throughout the technology/activity’s life cycle, and the amount of waste released to the environment. It aims to evaluate and quantify the environmental impact associated with a particular technology/activity in a detailed and rigorous manner. Additive manufacturing is a cost-effective solution for producing intricate and lightweight geometries, particularly in small batch quantities and situations where it can reduce lead times, which is highly relevant for bioremediation, but its overall economic potential could still be limited by factors such as expensive printers, lower production capacity, and slower build rates, and the societal impact of additive manufacturing on various stakeholders, including workers, local communities, society, consumers, and value chain actors, has yet to be fully assessed as this research is still in its early stages ( Kokare, Oliveira & Godina, 2023 ). Multi-material and function integration Modern 3D printing technologies can incorporate multiple materials into a single print, creating complex devices with integrated functions such as embedded sensors for monitoring remediation progress or channels for optimized distribution of nutrients and microorganisms ( Nazir et al. , 2023 ). A system for modeling microbes in 3D geometries using projection SLA to bioprint microbes within hydrogel matrices was able to show promise for engineering biofilms with dual functionality: metal sequestration and the uranium sensing capability using Caulobacter crescentus strains ( Dubbin et al. , 2021 ). A previous study used a gelatin/alginate (5% and 2% w/v, respectively) biomaterial ink containing B. subtilis 2569 that was genetically tailored to fabricate engineered multifunctional biofilms for fluorescence detection, conjugation chemistry, single-substrate bioremediation, and multi reaction bioremediation cascades incorporating nanoparticles ( Huang et al. , 2019 ). Therefore, utilizing the capabilities of 3D printing technology has the potential to transform the field of bioremediation by offering improved, innovative, and practical solutions for addressing environmental contamination. By adopting combinative bioremediation approaches such as this, we can fully leverage the ability of various bioremediation strategies and effectively combat environmental challenges with greater efficiency. Additive manufacturing techniques in bioremediation 3D printing encompasses a range of technologies that create objects by adding material layer-by-layer based on digital models. Each of these technologies offers distinctive benefits that can be leveraged in bioremediation to develop structures with characteristics tailored to environmental cleanup needs. Even though the adoption of this technology in bioremediation is in its early stages, the area is seeing massive growth. Below is a description of several key 3D printing methods, how they can be relevant to bioremediation, and examples where they are used for remediation purposes. Figure 2 depicts the specific 3D bioprinting methods utilized in bioremediation, highlighted in green. 10. 7717/peerj. 16897/fig-2 Figure 2 3D bioprinting methods utilized in bioremediation. Stereolithography (SLA) SLA is one of the oldest and most precise 3D printing techniques, which uses an ultraviolet (UV) laser to cure and solidify photopolymer resin layer by layer. SLA leverages photopolymerization, wherein a vat of liquid photopolymer resin cures upon exposure to a targeted UV light source. The resins typically blend reactive monomers and oligomers, photoinitiators, and various functional additives. During printing, photoinitiators absorb UV light to generate reactive species (free radicals or cations) that propagate a chain reaction, leading to crosslinking and forming a solid polymer matrix. The high resolution of SLA can produce parts with smooth surface finishes and intricate details. In bioremediation, the precision of SLA is particularly beneficial for creating microfluidic devices used in lab-on-a-chip applications that simulate environmental conditions for research and development of remediation strategies ( Huang et al. , 2015 ). In the context of bioremediation, SLA can be used as a tool for creating the necessary structures and devices to support the growth and activity of biological agents that break down pollutants ( Kadiak & Kadilak, 2017 ; Liu et al. , 2023b ). A recent study reported on the development of SiO 2 /TiO 2 /polymer scaffolds using SLA technology. These scaffolds incorporated sugarcane leaf-derived SiO 2 as the adsorbent, multi-phase TiO 2 synthesized through a solution combustion technique as the photocatalyst, and a photocurable resin as the structural material. The scaffolds demonstrated an average total removal efficiency of 81. 9% for methylene blue and 60% for rhodamine B dyes, which shows potential for use in wastewater treatment applications ( Bansiddhi et al. , 2023 ). However, the practicality of using SLA-printed items in bioremediation depends on developing and using appropriate materials that align with environmental safety and sustainability goals. Selective laser sintering (SLS) SLS uses a laser to sinter powdered material, typically nylon or polyamide, to form solid structures. This method can produce durable and complex geometries without supporting structures. SLS has also been used to manufacture structurally complex miniaturized photobioreactor parts using polyamide ( Krujatz et al. , 2016 ). SLS-printed parts can fabricate sturdy components for bioremediation processes that may require good mechanical properties and chemical resistance from harsh environmental conditions ( Hopkinson, Hague & Dickens, 2006 ). In recent times, there has been an increased interest in the development of polymeric nanocomposites for water treatment applications using SLS to create durable, efficient, and cost-effective polymer nanocomposites that are monodisperse, highly reactive and have minimal surface or structural defects ( Adeola & Nomngongo, 2022 ). Fused deposition modeling (FDM) FDM, also known as fused filament fabrication (FFF), is a widely used 3D printing method that extrudes thermoplastic polymers through a heated nozzle to form layers. FDM is highly versatile and allows for the printing of large parts at a lower cost. For bioremediation, FDM can be utilized to create custom housings for biofilters or frameworks for biofilm reactors that are scalable and cost-effective ( Rocha et al. , 2017 ). Studies using natural biopolymers and biopolymer-based materials, including chitosan, polylactic acid (PLA), alginate, and cellulose acetate (CA), for potential applications within the water treatment industry with emphasis on oil separation and metal removal, are being done using FDM ( Fijoł, Aguilar-Sánchez & Mathew, 2022 ). Digital light processing (DLP) Similar to SLA, DLP 3D printing also uses a light source to cure photopolymers, but it does so by projecting an entire layer’s image at once, which can result in faster print times. DLP is particularly suited for manufacturing small to medium-sized intricate structures that require high precision, such as scaffolds for microbial attachment in bioremediation systems ( Melchels, Feijen & Grijpma, 2010 ). Recently, researchers have been working on developing a platform for extrusion 3D bioprinting of hydrogel-based bio-inks loaded with diatoms, where a digital light processing (DLP) bioprinting platform was used to shape photolabile polymers containing dinoflagellates or diatoms that were responsive to contaminants (salt, antimicrobial agents and herbicide), even though this was developed for biosensing, platforms such as this could be easily adapted to created biohybrid materials that could be used for bioremediation ( Boons et al. , 2023 ). Material jetting and binder jetting Material Jetting involves jetting droplets of photopolymer, which are then cured by UV light. It is known for its ability to produce parts with high accuracy and smooth surfaces and its capacity to print with multiple materials simultaneously. This could be advantageous for creating multi-material bioremediation devices with structural and functional elements integrated into a single print ( Derby, 2010 ). Binder jetting involves selectively depositing a liquid binding agent onto a powder bed, bonding these areas together to form a part. Since it can use various materials, including metals, sands, and ceramics, this method can produce components for bioremediation that require specific material properties, such as catalyst supports for the chemical degradation of pollutants ( Gibson, Rosen & Stucker, 2015 ). These techniques are used in tandem with other additive manufacturing technologies for bioremediation applications. Multi-material 3D bioprinting Advanced 3D bioprinting technologies can handle multiple materials within a single printable bioink formulation. This allows for the fabrication of complex devices with varying material properties, including combining biodegradable materials with functional additives that enhance microbial growth or pollutant adsorption in bioremediation processes ( Sun et al. , 2013 ). Bioprinting, a specialized form of multi-material 3D printing that is extrusion-based, involves the precise layering of bioinks composed of cells, growth factors, and functional biomaterials to construct biofunctional structures. For experts in bioremediation, bioprinting opens a frontier for fabricating bio constructs tailored to degrade environmental pollutants. These living or biochemically functional architectures can be engineered to optimize the viability, functionality, and performance of the encapsulated particles, which may be cells, nanoparticles, enzymes, or other functional materials, thereby enhancing the efficiency and specificity of biodegradation pathways. By manipulating the composition and spatial distribution of different cell types within a bioink, bioprinted constructs can be customized to target specific contaminants. Additionally, integrating sensing components within bioprinted matrices can lead to the development of intelligent bioremediation systems capable of real-time monitoring and response. Advancements in bioink development, focusing on immobilizing microbes, enzymes, nanoparticles, metal-organic frameworks, or particles with catabolic prowess, are pivotal for extending bioprinting applications towards eco-restoration and pollution abatement. For bioremediation, these additive manufacturing techniques can be strategically selected based on the requirements of the remediation task, such as biodegradability, biocompatibility, chemical resistance, mechanical properties, and the complexity of the structures required for optimizing the degradation of contaminants. Table 1 provides a quick snapshot of how each of the above techniques could potentially play a role in bioremediation and their relative advantages and disadvantages. As the demand for innovative and sustainable bioremediation solutions continues to grow, the potential of various additive manufacturing technologies to revolutionize the way we tackle bioremediation is becoming increasingly evident and poised to become a key player in developing next-generation bioremediation techniques. 10. 7717/peerj. 16897/table-1 Table 1 Comparative view of different 3D printing techniques in the context of bioremediation. 3D printing technique Material compatibility Structural complexity Durability in tough environmental conditions Economic feasibility Suitability for bioremediation References Stereolithography (SLA) Photopolymers High Moderate Moderate-high Suitable for intricate bio-scaffolds and microfluidic devices for controlled bioremediation environments Huang et al. (2015), Kadiak & Kadilak (2017), Liu et al. (2023b), Bansiddhi et al. (2023) Selective laser sintering (SLS) Nylon, polyamide Moderate-high High Moderate-high Ideal for producing robust components for harsh environmental conditions Krujatz et al. (2016), Hopkinson, Hague & Dickens (2006), Adeola & Nomngongo (2022) Fused deposition modeling (FDM) Thermoplastics ( e. g. , PLA, ABS) Moderate Moderate Low-moderate Cost-effective for large-scale bioreactors and support structures Rocha et al. (2017), Fijoł, Aguilar-Sánchez & Mathew (2022) Digital light processing (DLP) Photopolymers High Moderate Moderate-high Suitable for precision structures like biofilm scaffolds and micro-bioreactors Melchels, Feijen & Grijpma (2010), Boons et al. (2023) Material jetting Photopolymers, waxes High Moderate High Applicable for multi-material structures with gradients for selective bioremediation Derby (2010) Binder jetting Metals, sands, ceramics Moderate High Moderate Useful for creating catalyst supports and filtration systems in water and soil remediation Gibson, Rosen & Stucker (2015) Materials for 3D printing in bioremediation The selection of materials in 3D printing for bioremediation is critical, as the materials must not only be suitable for the printing process but also conducive to bioremediation activities. For example, when used in bioremediation, certain materials must support microbial life for the degradation of pollutants and be mechanically stable while also being environmentally sustainable. Exploring new materials and techniques to achieve efficient and cost-effective bioremediation processes is also crucial. Some common materials for 3D bioprinting that could play a role in bioremediation are discussed below with examples. Biodegradable polymers Biodegradable polymers are favored in bioremediation applications for their ability to break down naturally over time, minimizing environmental impact. Polylactic acid (PLA) is one such polymer, popular in 3D printing for its ease of use and compostable properties. PLA can be used to create frameworks for microbial films in water treatment or soil remediation, gradually degrading into harmless lactic acid ( Farah, Anderson & Langer, 2016 ). Another research work using PLA has led to the creation of a bioremediation system based on using a native isolate of Chlorella vulgaris immobilized onto an alginate matrix inside a PLA device, where the researchers were able to successfully demonstrate the reduction of all inorganic nitrogen forms and total phosphorus by 90% after 5 days, and a 85% decrease in aerobic mesophilic bacteria ( Marconi et al. , 2020 ). Polyhydroxyalkanoates (PHAs) are another class of biopolymers produced by bacterial fermentation of sugars or lipids and are completely biodegradable, making them ideal for temporary structures in ecosystem restoration projects ( Kourmentza et al. , 2017 ). Ongoing research in this field will likely lead to even more innovative uses for these materials in the future. Composites Composites that blend biodegradable polymers with natural fibers or fillers can enhance the mechanical properties and biodegradability of printed objects. For instance, a composite of PLA and natural fibers like cellulose can be designed to provide structural support in bioremediation systems while maintaining biodegradability ( Benini, de Bomfim & Voorwald, 2023 ). In another recent work, researchers combined microgel-based granular inks that were 3D printable to fabricate bacteria-induced biomineral composites that were biomimetic comprising 93wt% calcium carbonate and the ability to withstand pressures up to 3. 5 Mpa ( Fig. 3, reused with permission from Hirsch et al. , 2023 ) for potential use as artificial corals to help in the regeneration of marine reefs and ocean remediation applications ( Hirsch et al. , 2023 ). Nanocellulose is another popular choice for creating 3D printable functional composites ( Finny, Popoola & Andreescu, 2021 ). 3D printable oil/water separators that could act as sponges to remove oil and other microorganisms from polluted sites have been developed using nanocellulose composites ( Firmanda et al. , 2023 ). 3D printable composites using polycaprolactone (PCL) and sodium alginate were found to have heavy metal adsorption properties, and the authors were able to demonstrate that sodium alginate retained its heavy metal adsorption properties within the PCL filament and was able to remove 91. 5% of copper ions from a 0. 17% w/w copper sulfate solution in 30 days thus making thermoplastic composite filaments such as these an exciting option for complex contaminated sites needing tailored solutions ( Liakos et al. , 2020 ). From the examples above, it is clear that the development of such functional composites opens up exciting bioremediation possibilities for tailored solutions in complex contaminated sites. 10. 7717/peerj. 16897/fig-3 Figure 3 3D printed biomineral composites. (A) Formulation of the bioink. (B) 3D printing. (C) Mineralization. (D) Schematic representation of the microbially-induced calcium carbonate precipitation process mediated by S. pasteurii. (E) 3D printed biomineral composite after four days of microbially-induced calcium carbonate precipitation process; the displayed scale bar is 10 mm. Image source credit: Hirsch et al. (2023), CC BY 4. 0 DEED, https://creativecommons. org/licenses/by/4. 0/. Functionalized materials Functionalized materials that contain adsorbents like activated carbon can be used to fabricate filters and membranes. These specialized materials are engineered to capture specific pollutants while allowing the proliferation of microorganisms that can degrade these pollutants ( Fan et al. , 2022 ). Researchers have constructed a 3D printing platform that uses rudimentary alginate chemistry for printing a bacteria-alginate bioink mixture onto calcium-containing agar surfaces, which resulted in the formation of bacteria-encapsulating hydrogels with varying geometries with the potential to be used as biofilms for environmental detoxification purposes such as bioremediation, heavy metal removal, removal of assimilable organic carbon, and wastewater treatment ( Balasubramanian, Aubin-Tam & Meyer, 2019 ). A functional material encapsulating Pseudomonas putida, a bacteria, in a biocompatible and functionalized 3D printable ink consisting of sodium hyaluronate and glycidyl methacrylate to print a “living material” capable of degrading phenol, a common pollutant, was demonstrated to show total phenol degradation after 40 h into harmless biomass ( Schaffner et al. , 2017 ). The ability of hydrogels to form hydrophilic aqueous microenvironments maintaining the reactivity of various catalysts along with their advantageous properties such as biocompatibility, swelling ability, and resistance to dissolution, make hydrogels ideal candidates for bioimmobilization and functionalization, as they provide improved stability of the immobilized components preventing leakages and the diffusion of substrate molecules and their reaction products. Sodium alginate and bentonite clay were used to create 3D printable nanocomposite hydrogels for the adsorption of the pesticide paraquat, and the removal tests indicated that the adsorption process was due to spontaneous adsorption mechanisms involving physisorption, showing a maximum adsorption capacity at equilibrium of 2. 29 mg/g with an ability to be reused for at least six cycles ( Baigorria et al. , 2023 ). Overall, these examples demonstrate the promising potential of functionalized materials and hydrogels in environmental remediation. By using specialized engineered materials and 3D printing technology, researchers can create innovative solutions for pollutant removal and wastewater treatment. These advancements in materials science and biotechnology offer hope for a cleaner and more sustainable future. Even though various materials could be used for environmental remediation, materials for 3D printing in bioremediation must be carefully chosen to ensure they do not introduce new contaminants, support the life cycle of the encapsulated or immobilized biological/chemical component, and have a negligible environmental footprint after their useful life. Design and modeling The conceptualization and execution of 3D structures tailored for bioremediation necessitate an interdisciplinary collaboration of environmental engineering, chemistry, biology, and materials science expertise. This intricate design and modeling process must encapsulate the multifaceted interactions between biological consortia and their physicochemical surroundings, ensuring that the created habitats not only foster microbial growth but also provide an active environment for the encapsulated active degradants to biodegrade the pollutants optimally. In a study aimed at removing drugs from water, researchers fabricated a device using SLA where they immobilized laccase sourced from Trametes Versicolor within a poly(ethylene glycol) diacrylate hydrogel and found that when the device was configured in the shape of a torus, it removed 95% of diclofenac and ethinylestradiol from aqueous solution within 24 and 2 h, respectively, and was much more efficient than free enzyme ( Xu et al. , 2022a ). This highlights the significance of creating and tuning optimal material geometries that favor pollutant removal when fabricating adsorbents. Computational design Computational tools are essential in the design process, enabling the simulation of different scenarios and the optimization of structures for maximum efficiency. Software such as computer-aided design (CAD) programs allows the creation of detailed 3D models that can be tested virtually under different conditions. Computational fluid dynamics (CFD) can simulate the flow of water or air through the structures, helping optimize nutrient distribution and waste removal, essential factors for microbial growth ( Versteeg & Malalasekera, 2007 ). Bioprinting is also fundamentally interdisciplinary, and therefore, it provides an opportunity for scientists and engineers to collaborate to apply engineering design and standardization parameters to the printing and analysis processes ( Correia Carreira, Begum & Perriman, 2020 ). Therefore, the incorporation of this technique into bioremediation methodology provides an avenue for pioneering interdisciplinary investigations. Design considerations When designing 3D structures for bioremediation, considerations include maximizing surface area for microbial colonization, enzyme/particle reactivity, and creating pore sizes that allow optimal flow rates and structural integrity to withstand environmental stresses. The design must also account for the ease of scaling up from laboratory to field sizes and the adaptability to different pollutants and ecological conditions ( Pant et al. , 2010 ). Integrating advanced design and modeling techniques ensures that 3D printed structures for bioremediation are optimized for environmental applications, promoting effective pollutant degradation and efficiency. Bio-inspired design Bio-inspired design, which emulates natural structures such as honeycombs or plant roots, can be particularly effective in bioremediation. These structures can be modeled to create complex geometries that mimic biological systems, offering high surface areas and efficient nutrient distribution pathways for microorganisms ( Wang, Chen & Chen, 2020 ). Research on bioinspired nanosurfaces with tailored multifunctionality, such as hydrophobicity, has attracted significant attention for scientific exploration and practical applications inspired by natural phenomena. As a result, 3D printing has emerged as an up-and-coming method for producing biomimetic materials with diverse applications due to its numerous advantages, including customizability, affordability, and accessibility ( Wang et al. , 2023a ). Applications of 3D printing in bioremediation This section delves into the ways in which 3D printing can be used in bioremediation, exploring topics such as microbial support structures, enzyme immobilization, heavy metal adsorption and filtration, case studies, and potential obstacles. Microbial support structures The success of bioremediation often hinges on the health and stability of microbial colonies. 3D printing has revolutionized the development of microbial support structures by enabling the creation of complex geometries tailored for microbial growth. These structures are designed to provide a high surface area-to-volume ratio, which is crucial for the colonization and bioactivity of microbes. Research has demonstrated that the porosity and the interconnectivity of the pores can be finely tuned to control the distribution of nutrients and the removal of metabolites, thus optimizing the bioremediation process ( Bhattacharjee et al. , 2016 ). Studies focusing on water treatment have utilized 3D-printed lattice structures that facilitate the growth of biofilms, which are integral in the degradation of organic pollutants ( Dzionek, Wojcieszyńska & Guzik, 2016 ). Researchers have formulated a dual-network bioink for 3D printing of “living materials” with enhanced biocatalysis properties, where the printable bioinks provide a biocompatible environment along with desirable mechanical performance; integrating microbes into these bioinks enabled the direct printing of catalytically living materials with high cell viability and optimal metabolic activity, for potential use in the bioremediation of chemicals; this study showed more than 90% degradation of methyl orange and acrylamide in 48 h using a bacteria-microalgae within the bioink matrix ( He et al. , 2022 ). A novel dual-crosslinking poly(ethylene glycol) diacrylate-alginate-poly (vinyl alcohol)-nanoclay (PAPN) bio-ink containing one heterotrophic bacterium ( Oceanimonas sp. XH2) was reported, where the authors used extrusion-based 3D printing to create a functional biomaterial with the capabilities of ammonia removal; the authors showed that the 3D printed PAPN functional material could remove 96. 2 ± 1. 3% ammonia within 12 h, and they also observed that the removal rate of ammonia increase with repeated use due to the rise in bacteria within the bio-scaffolds over time ( Li et al. , 2022 ). Similarly, various bacterial and microbial species are now being mixed with polymers creating functional complex bioinks, and these systems show enormous potential in applications such as bioremediation, and sometimes they can even respond to pollutants serving as sensors that can detect toxic chemicals and also potentially as oil spill filters as discussed earlier, making these 3D printed “minibiofactories” outstanding candidates for biotechnology-based bioremediation ( Kyle, 2018 ). Researchers have recently introduced a new micromodel technology that has been designed to investigate bacterial biofilm formation in porous media. This technology is particularly useful for understanding biofilm dynamics in various applications, including wastewater treatment and soil bioremediation. The heart of this technology is a 3D-printed micromodel that enables the growth of biofilm within a perfusable porous structure. By utilizing high-precision additive manufacturing techniques, particularly stereolithography, the authors have developed a system that allows for precise control over the microenvironment, including flow channels and substrate architecture. One of the key advantages of this technology is the ability to monitor crucial parameters such as oxygen consumption, pressure changes, and biofilm detachment, which are essential for comprehending and optimizing biofilm behavior. The authors have demonstrated how this technology can be used to study Pseudomonas aeruginosa biofilm development for several days within a network of flow channels ( Papadopoulos et al. , 2023 ). Studies like these demonstrate the benefits of using additive manufacturing techniques to create consistent 3D porous microarchitectures, and these approaches act as ideal platforms for examining the dynamics of biofilm development in 3D porous media and quickly refining processes that promote bioremediation. Enzyme immobilization The field of enzyme immobilization has greatly benefited from the advent of 3D printing. The technique allows for the precise placement of enzymes on various substrates, which can be used to catalyze the breakdown of pollutants. This spatial control not only improves the stability and reusability of enzymes but also enhances the efficiency of the bioremediation process. Researchers have leveraged 3D printing to develop bioreactors where enzymes are immobilized on printed scaffolds, resulting in increased degradation rates of pollutants like phenol and other aromatic compounds ( Shao et al. , 2022 ; Bellou et al. , 2022 ). 3D printing has also been used to create an enzyme-immobilized platform for biocatalysis by formulating a printable hydrogel ink comprising of dimethacrylate-functionalized Pluronic F127 (a non-ionic copolymer surfactant) and sodium alginate with the enzyme laccase for possible uses in environmental remediation. A piece of work using 3D bioprinting utilized a bioink made of sodium alginate, acrylamide, and hydroxyapatite with immobilized laccase for biodegradation of p-chlorophenol where the immobilized laccase exhibited excellent storage stability and reusability and retained over 80% of its initial enzyme activity after three days of storage, and was able to be reused for treating seven batches of phenolic compounds ( Liu et al. , 2020 ). Another recent work using laccase reported a biocatalytic system using immobilized laccase to 3D printed open-structure biopolymer scaffolds that were shown to remove 35-40% of estrogen group hormones such as 17β-estradiol and 17α-ethynylestradiol from municipal wastewater containing 56 ng/L of 17α-ethynylestradiol and 187 ng/L of 17β-estradiol ( Rybarczyk et al. , 2023 ). Research has also demonstrated that these estrogen group hormones could bind onto 3D-printed (SLS) filters made from commonly used polymers, such as polyamide-12 (PA), thermoplastic polyurethane (TPU), polypropylene (PP), and polystyrene (PS), and these filters showed enhanced surface morphology ( Fig. 4, reused with permission from ( Frimodig & Haukka, 2023 )) and removal capacities of 35, 32 and 37 μg g−1 for estrone, 17β-estradiol, and 17α-ethinylestradiol, respectively ( Frimodig & Haukka, 2023 ). These developments underscore the potential of 3D printing in creating more effective bioremediation tools and systems. 10. 7717/peerj. 16897/fig-4 Figure 4 Surface of each 3D printed filter imaged using SEM. Scanning electron microscope images of surfaces of each 3D printed filter using (A) polyamide-12, (B) polystyrene, (C) thermoplastic polyurethane, (D) polypropylene, (E) A simplified cross-section of the 3D-printed material illustrating solvent flow-through; notice the abundance of available surface area. Image source credit: Frimodig & Haukka (2023), CC BY 3. 0 DEED, https://creativecommons. org/licenses/by/3. 0/. Heavy metal adsorption and filtration Heavy metal contamination is a critical environmental issue, and 3D printing has emerged as a promising approach to developing novel adsorption and filtration systems. For example, a study that looked at sediment samples collected from three locations in Port Everglades, Florida, USA, indicated elevated ecological risk because of moderate-to-significantly high heavy metal contamination [As (0. 607–223 ppm), Cd (n/d–0. 916 ppm), Cr (0. 155–56. 8 ppm), Co (0. 0238–7. 40 ppm), Cu (0. 004–215 ppm), Pb (0. 0169–73. 8 ppm), Mn (1. 61–204 ppm), Hg (n/d–0. 736 ppm), Mn (1. 61–204 ppm), Ni (0. 232–29. 3 ppm), Se (n/d–4. 79 ppm), Sn (n/d–140 ppm), V (0. 160–176 ppm), and Zn (0. 112–603 ppm); n/d = not-detected] ( Giarikos et al. , 2023 ). 3D-printed structures can be embedded with materials like biochar, activated carbon, or metal-organic frameworks, which have a high affinity for heavy metals and can serve as potential remediation solutions for such issues. The design flexibility of 3D printing allows for the optimization of these structures, maximizing contact time and enhancing the removal efficiency of heavy metals such as lead, cadmium, and arsenic from contaminated water and soil ( Ignatyev, Thielemans & Vander Beke, 2014 ; Fee, Nawada & Dimartino, 2014 ). Researchers have also reported a polylactic acid-hydroxyapatite biocomposite prepared through a solvent-assisted blending and thermally induced phase separation technique, which was processed into highly permeable 3D biofilters using FDM showing maximum adsorption capacities of 112. 1 and 360. 5 mg/g for the metal salts of lead and cadmium respectively ( Fijoł et al. , 2021 ). A new work using a chitosan-hydroxyapatite coupled with PLA to create monolithic filters utilizing 3D printing demonstrated robust Cu 2+ removal performance with a maximum adsorption capacity of 119 mg/g, exhibiting the ability to remove more than 80% Cu 2+ from their sample in less than 35 min ( Wang et al. , 2023b ). The graphical representation of how the authors fabricate the filter can be seen in Fig. 5 (Reused with permission from ( Wang et al. , 2023b )). A previous work where a reusable monolithic 3D porous adsorbing filter was 3D printed using chitosan for heavy metal removal showed an adsorption capacity of 13. 7 mg/g with adsorption kinetics of 2. 2 mg/m per minute for Cu 2+ removal and this is further proof for the role of 3D bioprinting in the field of bioremediation ( Zhang et al. , 2019 ). 10. 7717/peerj. 16897/fig-5 Figure 5 Graphical representation of the PLA-Chitosan(CS)-Hydroxyapatite(HAP) filter fabrication. Image source credit: Wang et al. (2023b), CC BY 4. 0, https://creativecommons. org/licenses/by/4. 0. Multifunctional, robust, reusable, and high-flux filters are needed for sustainable water treatment and bioremediation, and to accomplish this, biobased and biodegradable water purification filters were developed and processed through 3D printing, more specifically using FDM; here, the authors used polylactic acid (PLA) based composites reinforced with homogenously dispersed (2, 2, 6, 6-Tetramethylpiperidin-1-yl)oxyl -oxidized cellulose nanofibers (TCNF) and chitin nanofibers (ChNF), and they have an adsorption capacity towards copper ions as high as 234 (TCNF) and 208 mg/g (ChNF) and maximum separation efficiency of 54% (TCNF) and 35% (ChNF) towards microplastics in laundry effluent water ( Fijoł et al. , 2023a ). 3D printing has also been combined with surface segregation and vapor-induced phase separation process to create structured adsorbents using composite inks consisting of polysulfone, polystyrene-block-poly(acrylic acid) and carbon nanotubes coupled with poly(ethyleneimine) (PEI) and terpyridine-COOH to get sorbents with copper ion removal capabilities of up to 31. 3 mmol/m 2 ; however, they observed degradation in copper removal in the presence of other ions ( Xu et al. , 2022b ). Biopolymer-based 3D printable hydrogels have also been explored for heavy metal removal from water, where a bioink consisting of shear-thinning hydrogels was fabricated by mixing chitosan with diacrylated Pluronic F-127, which showed 95% metal removal within 30 min in some cases ( Appuhamillage et al. , 2019 ). A one-step 3D printing method ( Fig. 6, reused with permission from ( Finny et al. , 2022 )) using 3D printable hydrogel-based adsorbents using alginate, gelatin, and polyethyleneimine-based bioink has also been reported to show excellent heavy metal ion removal adsorption capacities of 90. 38%, 59. 87%, 46. 27%, 38. 66%, and 6. 45% for Cu 2+, Ni 2+, Cd 2+, Co 2+, and Pb 2+ ions respectively from the tested samples ( Finny et al. , 2022 ). A 3D printable nanocomposite hydrogel was fabricated through electron beam crosslinking of alginate/nanoclay to remove inorganic micropollutants from wastewater for heavy metal removal applications where the authors note a maximum removal capacity of 532 mg/g for Pb(II) ions ( Shahbazi et al. , 2020 ). 10. 7717/peerj. 16897/fig-6 Figure 6 One-step 3D-printing of heavy metal removal hydrogel tablets. Illustration of the one-step 3D-printing fabrication (A) and removal (B) process of the hydrogel tablets, showing the interaction between PEI and Cu 2+ ions, as an example. The hydrogel turns blue in the presence of Cu 2+ due to the chelation process leading to the formation of cuprammonium complexes within the printed hydrogel. (Reused with permission from Finny et al. , 2022 ). Hydrogel filters containing algae cells have been 3D printed and experimentally shown to remove copper from test solutions by about 83% in 1 h ( Thakare et al. , 2021 ). A 3D-printed monolith fabricated using DLP using polyethylene glycol diacrylate, a plant-based resin, and chitosan exhibited removal efficiencies of 20. 8 % to 90. 4 % for methyl orange dye with an equilibrium uptake capacity ranging from 1 to 12. 7 (mg/g) after 2 h ( Husna et al. , 2022 ). In a recent work, cellulose and metal-organic frameworks were combined to create a 3D printed composite material that exhibited CO 2 and heavy metal ions adsorption capacities of 0. 63 mmol/g (27. 7 mg/g) and 8 to 328 mg/g, respectively while also displaying complete (>99%) removal of organic dyes in 10 min with high selectivity toward anionic dyes like methylene blue ( Nasser Abdelhamid, Sultan & Mathew, 2023 ). 3D printed biobased filters anchored with a green metal-organic framework have shown to have maximum adsorption efficiencies of 42. 3% for Pb (II), 72. 8% for Mn (II), 21. 1% for As (III), 47. 1% for Cd (II) and 41% for Zn (II) after 24 h, making them potential candidates for effluent treatment ( Fijoł et al. , 2023b ). Polydopamine (PDA) and bovine serum albumin (BSA) were added to a graphene-based ink to 3D print graphene-biopolymer aerogels for water contaminant removal as a proof of concept and preliminary results showed that the aerogel removed 100% organic solvents over 10 cycles of regeneration and reuse ( Masud, Zhou & Aich, 2021 ). Researchers have recently created recyclable 3D printed hydrogel composites that incorporate biochar sourced from rice husk for removing organic contaminants from tap water and have experimentally demonstrated that the hydrogel containing 10% w/w biochar (Alginate/Biochar) demonstrated significant adsorption capacities of 111. 4 mg/g for ibuprofen (IBU) and 214. 6 mg/g for methylene blue (MB) which represents an increase in adsorption capacities of 48% (IBU) and 58% (MB) compared to conventional hydrogels without biochar. This innovative development highlights the potential of novel composites and underscores the importance of continuing to explore new avenues for improving water quality ( Silva et al. , 2023 ). Case studies Real-world case studies illustrate the practical applications of 3D printing in bioremediation. One such example is the deployment of 3D-printed biofilters for the treatment of industrial wastewater, where the specificity of the printed matrix improved the reduction of nitrogen and phosphorus levels ( Mohd Yusoff et al. , 2023 ). Another case involved using 3D-printed sponges for oil spill management, where the porous structures enhanced the absorption of hydrocarbons, facilitating the subsequent biodegradation by marine microbes ( Walker & Humphries, 2019 ). These case studies, paired with the multiple works discussed previously, showcase the potential of 3D printing technology in environmental remediation. The ability to customize the matrix of the printed materials offers a high degree of control in designing effective and efficient bioremediation systems; the success of these case studies provides a promising outlook for the future of 3D printing in bioremediation and highlights the importance of interdisciplinary collaborations between engineering and environmental science. Analyzing patents could also be indicative of the commercialization potential for these technologies and could provide valuable insights. The University of Rochester has patented a low-cost and efficient 3D printing method for creating genetically modified Escherichia coli biofilms, which can be used for environmental detoxification and bioremediation ( Meyer, 2020 ). Princeton University has filed a patent for a method of manufacturing a 3D porous medium that has the potential to utilize motile bacteria to move toward and break down contaminants that are trapped in soils, sediments, and subsurface formations ( Datta & Bhattacharjee, 2020 ). Tianjin University has filed a patent for a high-affinity and high-mechanical double-network printing ink that enables the creation of a high-functional 3D microbial material with improved bioremediation efficiency and resistance to complex environmental impacts ( Zhao et al. , 2022 ). These case studies, complemented by patents from leading universities, illustrate the burgeoning role of 3D printing in bioremediation. They not only validate the efficacy and commercial potential of these technologies but also highlight the synergy between engineering and microbial ecology. This conjunction is paving the way for innovative, effective, and adaptable environmental remediation strategies, marking a significant advancement in the application of 3D bioprinting technologies for ecological restoration. Challenges and limitations Despite promising advancements, integrating 3D printing in bioremediation faces several challenges. One of the main concerns is the economic feasibility, particularly the high costs associated with certain 3D printing technologies and the research-intensive material development phase, which might be time and cost-prohibitive and may not be justified by the scale of many bioremediation projects ( Ngo et al. , 2018 ). Scalability remains a hurdle, as translating laboratory-scale successes to field applications is often challenging due to the complexities of real-world environmental conditions ( Park et al. , 2022 ). Sustainability issues also arise, especially in the life cycle assessment of the materials used for printing, focusing on the energy consumption and potential waste generated by the printing process ( Nadagouda, Ginn & Rastogi, 2020 ). Despite the challenges, researchers continue to work towards overcoming these obstacles and advancing the use of 3D printing in bioremediation, and the future looks promising. Future directions In this section, we explore the possibilities of using new materials and methods that incorporate advanced technologies to tackle evolving challenges and opportunities. We investigate the potential synergy between 3D bioprinting and emerging fields and how they can possibly be leveraged to create innovative solutions. Advanced materials Exploring advanced materials in 3D printing holds significant promise for enhancing bioremediation strategies. Smart polymers that respond to environmental stimuli such as pH, temperature, or the presence of specific contaminants could revolutionize the way bioremediation is approached by enabling more dynamic and responsive cleanup processes. Nanomaterials, such as nanoparticles with catalytic properties, can also be integrated into 3D-printed structures to boost the efficiency of pollutant degradation; however, the impact of using such particles needs to be assessed from a sustainability perspective. Research into biodegradable and bio-based printing materials further aligns with the sustainability goals of bioremediation, minimizing the environmental footprint of the remediation tools themselves ( Wei et al. , 2017 ; Shafranek et al. , 2019 ). Researchers are also investigating the 3D printability of algae-based materials and have found that PHAs derived from algae could be a sustainable alternative while maintaining excellent mechanical properties and being environmentally friendly ( Grira et al. , 2023 ). Researchers are exploring the use of 3D printing to create eco-friendly geopolymer materials that can remove methylene blue from wastewater. These materials made using reduced graphene oxide (rGO) and zinc oxide (ZnO) had achieved an impressive 92. 56% removal efficiency of MB within just 30 min, and they used the same geopolymer ink, which contained 56% rGO@ZnO, to 3D print a scaffold using Direct Ink Writing technology ( Liu et al. , 2023a ). These newer materials can further help improve the field by improving the multifunctionality of the constructs while enhancing their robustness, durability, and environmental sustainability. Utilizing these materials can potentially increase the efficiency and effectiveness of bioremediation processes, ultimately leading to a cleaner and healthier environment. Integration with other technologies Integrating 3D printing with the Internet of Things (IoT) and artificial intelligence (AI) presents exciting opportunities to create more intelligent and autonomous bioremediation systems. IoT devices can provide real-time monitoring of environmental conditions and pollutant levels, feeding data into AI algorithms to predict and adjust the bioremediation process for improved results. The potential for self-regulating bioremediation systems, which adapt to changing conditions without human intervention, could be realized through the convergence of these technologies, significantly increasing the efficacy and reducing the cost of bioremediation operations ( Lawless et al. , 2019 ; Salam, 2020 ). Machine learning is also being explored to optimize processes, applied materials, and biomechanical performances to enhance bioprinting and bioprinted constructs ( Sun et al. , 2023 ) and someday could help tailor 3D printable materials specific to the contaminated sites. Bioprinting in space missions to produce engineering living materials capable of oxygen production and wastewater treatment could significantly impact the development of bioregenerative life support systems ( Krujatz et al. , 2022 ). Additive manufacturing has the potential to contribute significantly to the field of bioremediation for terraforming applications and one day might play a pivotal role in making the atmosphere, volatile components, temperature, surface topography, or ecology of astronomical bodies habitable for human settlement. Automation of bioprinting processes coupled with robotic platforms brings a new dimension of functionality to the field of bioremediation. As suggested by one study, incorporating an advanced artificial intelligence-based control system into in situ bioremediation of petroleum-contaminated groundwater systems significantly improved the efficiency and effectiveness of a process, leading to better remediation results ( Hu, Huang & Chan, 2003 ). As 3D printing technology continues to evolve, it could be used to produce customized bioremediation systems that incorporate advanced AI-based control mechanisms, leading to more effective and efficient remediation outcomes. Policy and regulation Supportive policy and regulatory frameworks are essential for 3D printing technologies to become a mainstay in environmental management. Policies encouraging research and development and incentives for adopting green technologies can accelerate the integration of 3D printing in bioremediation. Regulations will need to evolve to ensure the safe deployment of these technologies, especially considering the use of novel materials and the potential generation of byproducts from 3D printing processes ( Papaconstantinou & Polt, 1997 ; Baiano, 2022 ). The European Union, for example, funds multiple water security projects that lead to the widespread implementation of novel solutions and innovation ( Community Research and Development Information Service (CORDIS), 2020 ). Conclusion 3D printing technology stands on the brink of revolutionizing bioremediation, offering unparalleled precision in fabricating structures that support intricate microbial ecosystems, enhance enzyme stability, and facilitate heavy metal sequestration. Nevertheless, while its potential is profound, several critical challenges and questions remain unaddressed, casting a shadow on the path to its widespread implementation. One of the most pressing issues lies in the economic viability of upscaling 3D-printed bioremediation solutions. Currently, the costs associated with 3D printing advanced materials, particularly at the scale required for impactful environmental applications, are not insignificant. This economic barrier must be surmounted to enable broader adoption of these technologies. Moreover, scalability extends beyond cost to the technical challenges of producing and deploying large-scale bioremediation structures in diverse environmental contexts. Regulatory frameworks also lag behind technological advancements, with current policies often ill-equipped to manage the nuanced risks and benefits of deploying 3D-printed materials in ecological settings. The development of comprehensive regulations that both promote innovation and ensure environmental safety is a critical need that must be met to foster public trust and industry growth. Looking to the future, unanswered scientific questions beckon for research into the long-term stability and functionality of 3D-printed bioremediation systems. The environmental impact of these materials, the degradant byproducts that they might produce, and the potential for nanoparticle or chemical leaching present a significant gap in our current understanding. Furthermore, while integrating IoT and AI holds promise for real-time monitoring and responsive bioremediation strategies, the practicalities of such systems under variable environmental conditions are yet to be fully explored. In conclusion, the pathway for 3D printing in bioremediation has immense potential and adoptive challenges. As this technology advances, it is imperative that research continues to address these economic, scalability, and regulatory challenges, as well as the pressing environmental safety and technical questions. Only through a concerted effort to bridge these gaps can we harness the full potential of 3D printing, steering the future of bioremediation toward more intelligent, effective, and sustainable practices. As the field of bioremediation continues to evolve, it is becoming increasingly clear that 3D printing has a crucial role to play in the development of more sophisticated and effective remediation technologies. By taking advantage of the unique capabilities of 3D printing, researchers and engineers can create highly customized and precise structures that can optimize the delivery of remediation agents to contaminated sites. Furthermore, 3D printing can be used to create complex microenvironments that mimic the natural conditions of soil and groundwater, allowing for more accurate testing and validation of new remediation techniques. As a result, 3D printing is poised to revolutionize the way we approach bioremediation, unlocking new opportunities for sustainable environmental management and protection.

10. 7717/peerj. 2040

PeerJ

Development of an angiogenesis-promoting microvesicle-alginate-polycaprolactone composite graft for bone tissue engineering applications

One of the major challenges of bone tissue engineering applications is to construct a fully vascularized implant that can adapt to hypoxic environments in vivo. The incorporation of proangiogenic factors into scaffolds is a widely accepted method of achieving this goal. Recently, the proangiogenic potential of mesenchymal stem cell-derived microvesicles (MSC-MVs) has been confirmed in several studies. In the present study, we incorporated MSC-MVs into alginate-polycaprolactone (PCL) constructs that had previously been developed for bone tissue engineering applications, with the aim of promoting angiogenesis and bone regeneration. MSC-MVs were first isolated from the supernatant of rat bone marrow-derived MSCs and characterized by scanning electron microscopic, confocal microscopic, and flow cytometric analyses. The proangiogenic potential of MSC-MVs was demonstrated by the stimulation of tube formation of human umbilical vein endothelial cells in vitro. MSC-MVs and osteodifferentiated MSCs were then encapsulated with alginate and seeded onto porous three-dimensional printed PCL scaffolds. When combined with osteodifferentiated MSCs, the MV-alginate-PCL constructs enhanced vessel formation and tissue-engineered bone regeneration in a nude mouse subcutaneous bone formation model, as demonstrated by micro-computed tomographic, histological, and immunohistochemical analyses. This MV-alginate-PCL construct may offer a novel, proangiogenic, and cost-effective option for bone tissue engineering.

Introduction In bone tissue engineering, seed cells play crucial roles in secreting growth factors and directly differentiating into the target tissue ( Wang et al. , 2015 ). However, for large bone defects that typically lack initial vascularization, implanted seed cells are often located a few hundred microns away from the nearest capillary supply and thus suffer from hypoxia and undergo apoptosis, resulting in a necrotic core of the implant ( Moon & West, 2008 ). Therefore, it is necessary to improve the proangiogenic ability of bone tissue engineering scaffolds. Recently, three-dimensional (3D) printed scaffolds have been widely studied for tissue engineering applications owing to their precise shape design and abundant choice of components ( Oryan et al. , 2014 ). Polycaprolactone (PCL) is a biodegradable polymer which has high mechanical strength and a low rate of degradation ( Shor et al. , 2009 ). It is generally considered that the degradation, mechanical strength and biocompatibility characteristics of PCL are suitable for bone tissue engineering applications ( Cheung et al. , 2007 ; Petrie Aronin et al. , 2008 ). However, PCL has a low cellular activity because it does not possess any biological molecules ( Rath et al. , 2012 ). To address these drawbacks, alginate is often used in combination with PCL scaffolds because it is structurally similar to extracellular matrix and can encapsulate various bioactive molecules ( Rufaihah & Seliktar, 2015 ). Growth factors, such as vascular endothelial growth factor (VEGF) and bone morphogenetic proteins (BMPs), can easily be incorporated into hydrogels and then seeded onto porous 3D printed scaffolds, thereby promoting angiogenesis and cell differentiation. For example, Kim, Jung & Kim (2013) investigated the osteoinductive potential of PCL scaffolds, and showed that it can be enhanced by coating PCL with alginate and BMP-2. However, the potential of growth factors in bone tissue engineering applications has been limited by their short half-life, low protein stability, high cost of production and restricted spatialtemporal effects due to lack of appropriate delivery approaches ( Mitchell et al. , 2016 ). Microvesicles (MVs) are spheroidal particles enclosed by a phospholipid bilayer with a diameter typically ranging from 30 to 1, 000 nm ( Ratajczak et al. , 2006 ; Lee et al. , 2011 ). Although MVs comprise a heterogeneous group, there are two common types: exosomes and microparticles. Exosomes originate from the endosomal compartment by fusion of multivesicular bodies with the plasma membrane, while microparticles (also called ectosomes) form by direct budding from the plasma membrane ( Cocucci, Racchetti & Meldolesi, 2009 ; Gatti et al. , 2011 ). During their developmental process, MVs ‘hijack’ both the membrane components (including antigens, receptors, and lipid rafts) and the cytoplasmic contents (including proteins, lipids, and nucleic acids) of their parent cells ( Ratajczak et al. , 2006 ). Upon release from the parent cells, MVs may interact with or enter their target cells and deliver their bioactive cargoes to them ( Van der Pol et al. , 2012 ). Cells may be changed by direct interactions, transfer of cell surface receptors, or epigenetic reprogramming ( Thery, Ostrowski & Segura, 2009 ; Quesenberry et al. , 2014 ). Accumulating evidence indicates that mesenchymal stem cell-derived MVs (MSC-MVs) possess potent proangiogenic potential ( Zhang et al. , 2012 ; Bian et al. , 2014 ; Chen et al. , 2014 ; Lopatina et al. , 2014 ). It has been reported that MSC-MVs can promote the proliferation, migration, and tube formation ability of endothelial cells in vitro ( Zhang et al. , 2012 ; Bian et al. , 2014 ; Lopatina et al. , 2014 ). In addition, the proangiogenic potential of MSC-MVs was demonstrated in several ischemic models in vivo ( Zhang et al. , 2012 ; Bian et al. , 2014 ). Moreover, analyses of MSC-MVs revealed the enrichment of angiogenesis-promoting growth factors and angiogenesis-associated mRNAs and miRNAs ( Chen et al. , 2014 ; Eirin et al. , 2014 ). As MVs are generally considered to be miniature versions of their parent cells, and because MSCs incorporated into hydrogels have been successfully used in regenerative medicine ( Yao et al. , 2015 ), we hypothesized that MSC-MVs could also be incorporated into hydrogels to promote neovascularization. In the present study, we developed an MV-alginate-PCL construct with enhanced proangiogenic ability for bone tissue engineering applications. The characteristics and proangiogenic properties of MVs released from rat bone marrow-derived MSCs (BMSCs) were examined in vitro. The procedures for fabricating the MV-alginate-PCL construct are shown in Fig. 1A. Subsequently, the effects of the MV-alginate-PCL construct on vascularization and tissue-engineered bone regeneration in vivo were investigated in a subcutaneous bone formation model in nude mice. 10. 7717/peerj. 2040/fig-1 Figure 1 Schematic design of the fabrication of MV-alginate-PCL constructs. (A) The procedures for fabricating the MV-alginate-PCL construct. MSC-MVs were isolated and resuspended with sodium alginate solution. Sterilized PCL scaffolds were loaded with MV-alginate composite solution and cross-linked with CaCl 2 solution. The MV-alginate-PCL constructs were implanted subcutaneously into nude mice for micro-CT, histological and immunohistochemical analyses. (B) A sketch of the structure of the 3D printed PCL scaffold. Materials and Methods Generation of 3D printed porous PCL scaffolds A PCL scaffold with a honeycomb-like pattern was fabricated using a fused deposition modeling technique, leading to triangular pores with a porosity of 70% and an average pore size of 0. 523 mm, as previously described ( Zhang et al. , 2010 ). The PCL scaffold was cut into 3-mm cubes and immersed in 75% ethanol for 2 h. Then the scaffolds were washed three times with PBS and dried at room temperature. A sketch of the structure of PCL scaffold is shown in Fig. 1B. Cell culture Primary culture of BMSCs All animal experiments were approved by the Ethical Committee of Tongji Medical College, Huazhong University of Science and Technology. Bone marrow was harvested from male Sprague–Dawley rats aged 2–3 weeks. Bone marrow was flushed out from the femurs and tibias with Dulbecco’s modified Eagle’s medium (DMEM; Hyclone, Logan, UT, USA) containing 10% fetal bovine serum (FBS; Hyclone) using a 1-mL syringe. The cells were centrifuged at 500 × g for 5 min. The cell pellet was resuspended in 10 mL of DMEM supplemented with 10% FBS (Hyclone) and 1% penicillin-streptomycin antibiotic (Gibco, Carlsbad, CA, USA), and the cells were seeded in a culture dish. After 48 h, the medium was changed and nonadherent cells were discarded. Cell passaging was performed until the monolayer of adherent cells reached 70–80% confluence. All of the experiments described below were performed using BMSCs from the third to fourth passage. Culture of human umbilical vein endothelial cells Human umbilical vein endothelial cells (HUVECs) were purchased from the American Type Culture Collection (ATCC, Rockville, MD, USA). The cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin–streptomycin antibiotic in a humidified incubator under an atmosphere of 5% CO 2 /95% air at 37 °C. Cell passaging was performed when the monolayer of adherent cells reached 90% confluence. Characterization of BMSCs Trilineage differentiation of BMSCs The trilineage differentiation potentials of BMSCs were measured as previously described ( Zhang et al. , 2009 ). All chemicals were purchased from Sigma (St. Louis, MO, USA) unless otherwise stated. For osteogenic induction, BMSCs were cultured in osteogenic differentiation medium (DMEM supplemented with 10 mM β -glycerophosphate, 0. 1 µM dexamethasone, and 50 µM ascorbic acid) for up to two weeks, with the medium changed twice a week. The extracellular accumulation of calcium was assayed by alizarin red staining. For adipogenic induction, BMSCs were cultured in adipogenic differentiation medium (DMEM supplemented with 5 µg/mL insulin, 200 µM indomethacin, 1 µM dexamethasone, and 0. 5 mM 3-isobutyl-1-methylxanthine) for 3 weeks, with the medium changed twice a week. The presence of lipid vacuoles was confirmed by oil red O staining. For chondrogenic induction, 1 × 10 6 BMSCs were pelleted by centrifugation at 500 × g for 5 min in a 15 mL centrifuge tube and incubated overnight in a humidified incubator with 5% CO 2 at 37 °C. The pelleted BMSCs were then cultured in DMEM supplemented with 0. 1 µM dexamethasone, 0. 17 mM ascorbic acid, 1. 0 mM sodium pyruvate, 0. 35 mM L-proline, 1% insulin-transferrin sodium-selenite, 1. 25 mg/mL bovine serum albumin, 5. 33 µg/mL linoleic acid, and 0. 01 µg/mL transforming growth factor- β (Cell Science, Canton, MA, USA) for four weeks, with the medium changed twice a week. The micromass pellets were formalin-fixed, paraffin-embedded, and cut into 10-µm sections. The sections were dewaxed and rehydrated before safranin O staining. Immunophenotype BMSCs were fixed in 10% formalin for 15 min and washed with phosphate-buffered saline (PBS). The expression of CD73, CD105, CD29, CD44, CD34, and CD45 was detected using rabbit anti-rat CD73, CD105, CD29, CD44, CD34, and CD45 monoclonal antibodies (Abcam, Cambridge, UK), respectively, followed by goat anti-rabbit IgG conjugated with fluorescein isothiocyanate (FITC; Invitrogen, Carlsbad, CA, USA). 4′6-Diamidino-2-phenylindole (DAPI; Beyotime, Beijing, China) was used for staining nuclei. Isolation and characterization of MSC-MVs MSC-MVs were harvested from the supernatant of BMSCs after 24 h of culture in DMEM without FBS, as described previously ( Hergenreider et al. , 2012 ) with some modifications. After centrifugation at 2, 000 × g for 20 min to remove cellular debris, the supernatant was centrifuged at 20, 000 × g for 1 h at 4 °C. The supernatant was then discarded, and the pelleted MVs were washed with ice-cold PBS and pelleted again by centrifugation at 20, 000 × g for 1 h at 4 °C. Finally, the supernatant was discarded, and the pelleted MVs were resuspended with PBS and stored at −80 °C until further experiments. The morphology of MSC-MVs was visualized using a scanning electron microscope (SEM; Hitachi, Tokyo, Japan), as previously described ( Sokolova et al. , 2011 ) with some modifications. Briefly, MSC-MVs were fixed with 2. 5% glutaraldehyde in PBS. After 2 h of fixation, glutaraldehyde was discarded, and the fixed MVs were washed twice with PBS and pelleted by centrifugation at 20, 000 × g for 1 h at 4 °C. MVs were then dehydrated in a series of ethanol solutions with increasing concentrations. The samples were dried at room temperature and then subjected to gold-palladium sputtering, followed by SEM analysis. For confocal microscopic analysis, MVs were stained with carboxyfluorescein succinimidyl ester (CFSE; Beyotime) in accordance with the manufacturer’s instructions, and then observed with a confocal microscope (Leica, Wetzlar, Germany). The phenotypic profile of MSC-MVs was determined by flow cytometry with an array of antibodies commonly used for MSC identification, as previously described ( Sun et al. , 2014 ) with some modifications. Standard microbeads with a diameter of 1 µm (Sigma) were used to set the upper size limit for MVs. Calcein AM (Molecular Probes, Eugene, OR, USA) was used to avoid the staining of cell debris. MVs were co-stained with calcein AM and phycoerythrin (PE) or peridinin-chlorophyll-protein (PerCP)-conjugated anti-CD73, -CD105, -CD29, -CD44, -CD90, -CD34, and -CD45 antibodies (BD Biosciences, San Jose, CA, USA), and analyzed using a FACSAria II flow cytometer (BD Biosciences). MVs were defined by the presence of calcein AM positivity and forward scatter signals lower than those of the 1-µm standard microbeads. Flow cytometric data were analyzed using FLOW JO software version 7. 6 (Tree Star Inc. , Ashland, OR, USA). Tube formation assay HUVECs (5 × 10 4 cells/well) were seeded onto the Matrigel (BD Biosciences)-coated wells of a 24-well plate and cultured in serum-free DMEM in the presence of various concentrations of MVs (1, 20, and 50 µg/mL) or PBS (control). Three replicated wells were set up for each group. Tube formation was examined using a phase-contrast microscope (Olympus, Tokyo, Japan) and the total length of the network was evaluated in five randomly selected fields for each well. The total length of the network was measured using Image-Pro Plus 6. 0 software (Media Cybernetics, Silver Spring, MD, USA), and expressed as a ratio to that of the respective control. Bone regeneration in vivo To investigate the effects of the MV-alginate-PCL constructs on promoting vascularization and tissue-engineered bone regeneration, four groups were prepared and implantation was performed subcutaneously into 4-week-old male nude mice ( n = 10 per group). The four groups were as follows: BMSC-MV-alginate-PCL group; BMSC-alginate-PCL group; MV-alginate-PCL group; and Alginate-PCL group. For the BMSC-MV-alginate-PCL group, BMSCs were subjected to osteogenic induction (as described above) for 2 weeks in culture dishes. Thawing of MSC-MVs was carried out in a preheated water bath at 37 °C and the thawed MVs were pelleted by centrifugation at 20, 000 × g for 1 h at 4 °C. The pelleted MVs were resuspended with 1. 5% sodium alginate (Sigma, product number W201502) solution at a final MV density of 1 µg/µL. Osteodifferentiated BMSCs were then harvested and mixed with the MV-alginate composite solution at a final density of 2 × 10 7 cells/mL. Twenty microliters of the BMSC-MV-alginate composite solution was seeded onto each PCL scaffold, and the constructs were completely immersed in 100 mM CaCl 2 solution for about 2 min to allow cross-linking. To investigate the distribution of BMSCs in BMSC-MV-alginate-PCL constructs, BMSCs were labeled with 1, 1′-dioctadecyl-3, 3, 3′, 3′-tetramethylindocarbocyanine dye (CM-Dil; Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions, and then mixed with MV-alginate composite solution and seeded onto PCL scaffold. The constructs were immersed in 100 mM CaCl 2 solution for 2 min and then observed by phase-contrast microscopy and confocal microscopy to determine whether the seeded cells were homogeneously distributed throughout the scaffold. For the BMSC-alginate-PCL group, BMSCs were subjected to osteogenic induction for 2 weeks, harvested, and mixed with 1. 5% sodium alginate solution at a final density of 2 × 10 7 cells/mL. Twenty microliters of the cell-alginate composite solution was then seeded onto each PCL scaffold and the constructs were completely immersed in 100 mM CaCl 2 solution for about 2 min to allow cross-linking. For the MV-alginate-PCL group, pelleted MVs were resuspended with 1. 5% sodium alginate solution at a final MV density of 1 µg/µL. Each PCL scaffold was then loaded with 20 µL of the MV-alginate composite solution and cross-linked with CaCl 2 solution. For the Alginate-PCL group, each PCL scaffold was loaded with 20 µL of 1. 5% sodium alginate solution and cross-linked with CaCl 2 solution. After 1 and 2 months of implantation, the animals were euthanized by an overdose of anesthesia and specimens were harvested for micro-computed tomographic (micro-CT), histological, and immunohistochemical analyses. Micro-CT analysis Micro-CT analysis was performed with a µCT-80 machine (Scanco Medical, Bassersdorf, Switzerland). Samples were fixed in 4% formalin and placed in the sample holder. The region of interest was set as a cylinder (36 mm in diameter and 5 mm in height) including all of the samples from a single group. The samples were three-dimensionally reconstructed, and the parameters of bone volume (BV) and bone volume per tissue volume (BV/TV) were obtained with micro-CT auxiliary software (Volume Graphics GmbH, Heidelberg, Germany). Histological and immunohistochemical analyses After the micro-CT analysis, the specimens were decalcified in 10% ethylene diamine tetraacetic acid solution for one week, dehydrated through an ethanol series, and embedded in paraffin for sectioning. The specimens were cut into 10-µm sections, mounted on glass slides, and stained with hematoxylin and eosin (HE). Immunohistochemical staining was performed on 10-µm sections. Antigen retrieval was performed prior to incubation with a rabbit anti-mouse CD31 (commonly used endothelial marker) monoclonal antibody (Abcam). The sections were then incubated with a horseradish-peroxidase-conjugated goat anti-rabbit antibody (Invitrogen), followed by color development with diaminobenzidine tetrahydrochloride (Santa Cruz Biotechnology, Santa Cruz, CA, USA) as the substrate. Five randomly selected 200 × fields in each slice (n = 3/group) were captured using a light microscope (Olympus). The number of CD31-positive vessels was calculated using Image-Pro Plus 6. 0 software (Media Cybernetics). Statistical analysis All data are presented as means ± standard deviation and were analyzed by one-way analysis of variance. Values of p < 0. 05 were considered statistically significant. Results Characterization of BMSCs Culture-expanded rat BMSCs displayed a spindle-like morphology ( Fig. 2A ). Extracellular calcium deposition was confirmed by alizarin red staining after 2 weeks of osteogenic induction of BMSCs ( Fig. 2B ). Exposure to adipogenic induction medium resulted in the accumulation of lipid vacuoles in the cytoplasm of BMSCs, as identified by oil red O staining ( Fig. 2C ). After four weeks of cell pellet culture in chondrogenic induction medium, BMSCs underwent chondrogenic differentiation, as confirmed by safranin O staining ( Fig. 2D ). Immunofluorescence staining revealed an immunophenotype that was positive for mesenchymal markers (CD73 and CD105) and cell adhesion molecules (CD29 and CD44) and negative for hematopoietic markers (CD34 and CD45) ( Fig. 2E ). 10. 7717/peerj. 2040/fig-2 Figure 2 Characterization of rat BMSCs. (A) Basic morphology of rat BMSCs. Scale bar: 250 µm. (B) BMSCs underwent osteogenic differentiation (demonstrated by alizarin red staining). Scale bar: 50 µm. (C) BMSCs underwent adipogenic differentiation (demonstrated by oil red O staining). Scale bar: 50 µm. (D) BMSCs underwent chondrogenic differentiation (demonstrated by safranin O staining). Scale bar: 25 µm. (E) Immunofluorescent staining of rat BMSCs showed that they were positive for CD73, CD105, CD29 and CD44 and negative for CD34 and CD45. Scale bars: 100 µm. Characterization of MSC-MVs MSC-MVs showed a spheroidal shape with a diameter of 100–1, 000 nm when observed under an SEM ( Fig. 3A ). They could be observed by confocal microscopy after staining with the fluorescent dye CFSE ( Fig. 3B ). For flow cytometric analyses, 1-µm standard microbeads were used as an internal size standard and calcein AM was used to avoid concomitant staining of cellular debris. Only particles with forward scatter signals below the level of the 1-µm standard microbeads and positively stained for calcein AM were defined as intact MVs. Our data showed that MSC-MVs exhibited an immunophenotype that was positive for CD73, CD105, CD29, CD44, and CD90 and negative for CD34 and CD45 ( Fig. 3C ). 10. 7717/peerj. 2040/fig-3 Figure 3 Characterization of MSC-MVs. (A) An SEM image revealing MSC-MVs (arrows) as spheroidal vesicles 100–1, 000 nm in diameter. Scale bar: 500 nm. (B) Confocal microscopy image of CFSE-stained MSC-MVs (arrows) with green fluorescence. Scale bar: 7. 5 µm. (C) Flow cytometric analysis of MSC-MVs revealed that they were positive for CD73, CD105, CD29, CD44 and CD90 and negative for CD34 and CD45. MSC-MVs promote tube formation of HUVECs in vitro The effect of MSC-MVs on in vitro capillary network formation was determined by a tube formation assay in Matrigel. Microscopic observation revealed that the difference between HUVECs treated with or without MSC-MVs was evident after 12 h of incubation. Therefore, the total length of the network structure was analyzed at the time point of 12 h. The results showed that MSC-MVs stimulated tube formation of HUVECs in a dose-dependent manner ( Fig. 4 ), suggesting that MVs secreted by rat BMSCs could promote angiogenesis in vitro. 10. 7717/peerj. 2040/fig-4 Figure 4 MSC-MVs promoted tube formation of HUVECs. (A) Representative images of tube formation assay in Matrigel. Scale bars: 200 µm. (B) Quantitative analysis of total length of vessel-like structures. Three replicated wells were set up for each group and five randomly selected views from each well were analyzed. * p < 0. 05 vs. control. 10. 7717/peerj. 2040/fig-5 Figure 5 Micro-CT analysis of bone formation. (A) 3D reconstruction of micro-CT images of the specimens from all of the four groups. (B) Quantitative analysis of bone volume (BV) and bone volume/tissue volume (BV/TV) in each group at 1 and 2 months ( n = 5/time point). * p < 0. 05. MV-alginate-PCL constructs promote tissue-engineered bone regeneration in vivo To investigate the distribution of the seeded cells in BMSC-MV-alginate-PCL construct, BMSCs were labeled with CM-Dil and the whole construct was observed by phase-contrast microscopy and confocal microscopy. As shown in Figs. S1A and S1B, BMSCs were incorporated in alginate and the alginate solution filled up the triangular pores of PCL scaffold. Additionally, confocal microscopy showed that CM-Dil labeled BMSCs were homogenously distributed around the PCL struts ( Figs. S1C and S1D ). To determine whether MV-alginate-PCL constructs could promote vascularization and tissue-engineered bone regeneration, four groups were prepared and implantation was performed subcutaneously into nude mice. After one and two months of implantation, samples from all groups were harvested and scanned by micro-CT to evaluate bone formation. As shown in Fig. 5, after one month of implantation, the BV and BV/TV in the BMSC-MV-alginate-PCL group (2. 29 ± 0. 38 mm 3, 8. 18% ± 1. 42%) were significantly increased compared with those in the BMSC-alginate-PCL (1. 37 ± 0. 36 mm 3, 4. 95% ± 1. 12%), MV-alginate-PCL (0. 09 ± 0. 03 mm 3, 0. 34% ± 0. 12%), and Alginate-PCL (0. 07 ± 0. 02 mm 3, 0. 27% ± 0. 08%) groups. The differences between the BMSC-MV-alginate-PCL group and the other three groups were even greater after 2 months of implantation, at which time the BV and BV/TV of the BMSC-MV-alginate-PCL group had increased to 6. 82 ± 0. 91 mm 3 and 24. 62% ± 3. 55%. In contrast, the BV and BV/TV of the other three groups after 2 months were lower BMSC-alginate-PCL: 3. 85 ± 0. 60 mm 3 and 13. 97% ± 2. 15%, MV-alginate-PCL: 0. 24 ± 0. 05 mm 3 and 0. 90% ± 0. 21%, and Alginate-PCL: 0. 18 ± 0. 04 mm 3 and 0. 66% ± 0. 13% ( Fig. 5 ). For further evaluation of the bone formation, HE staining was carried out on samples from each group after 2 months of implantation. As shown in Fig. 6, in the MV-alginate-PCL and Alginate-PCL groups, the constructs were primarily occupied by cord-like fibrotic tissue (high-magnification images in Fig. 6 ), and very little new bone formation was observed. In the BMSC-MV-alginate-PCL and BMSC-alginate-PCL groups, bone formation with mineralized tissue (high-magnification images in Fig. 6 ) was observed. Overall, more bone formation was observed in the BMSC-MV-alginate-PCL group than in the other groups. 10. 7717/peerj. 2040/fig-6 Figure 6 Histological analysis. Representative images of HE staining of the specimens from all of the four groups at low magnification (A) and high magnification (B). Scale bars: 100 µm. NB: new bone; FT: fibrotic tissue. MSC-MVs promote vascularization in vivo The ability of MSC-MVs to promote in vivo vascularization was evaluated by quantification of immunohistochemically-stained vessels. Specimens from all four groups showed positive staining for CD31. The mean number of CD31-positive vessels was significantly higher in the BMSC-MV-alginate-PCL group than in the BMSC-alginate-PCL, MV-alginate-PCL, and Alginate-PCL groups ( Figs. 7A and 7B ). Importantly, in the MV-alginate-PCL group, there was a three-fold increase in the number of CD31-positive vessels compared with that in the Alginate-PCL group ( Fig. 7B ). These results demonstrated the ability of MSC-MVs to promote vascularization in vivo. 10. 7717/peerj. 2040/fig-7 Figure 7 Immunohistochemical analysis. (A) Representative images of CD31 expression in the specimens from each group. Scale bars: 100 µm. (B) Quantitative analysis of CD31-positive vessels in each group after 2 months of implantation. Three samples in each group and five randomly selected views from each sample were analyzed. * p < 0. 05. Discussion The survival rate and repair efficacy of tissue-engineered grafts after implantation in vivo have been shown to be strongly associated with the extent of neovascularization ( Saran, Gemini Piperni & Chatterjee, 2014 ). The aim of this study was to investigate the possibility of incorporating MSC-MVs into an alginate-PCL construct previously developed for bone regeneration, to enhance its therapeutic potential by promoting angiogenesis. Notably, the MV-alginate-PCL construct mixed with osteodifferentiated BMSCs facilitated the most bone formation in the subcutaneous bone formation model in nude mice ( Fig. 5 ). Immunohistochemical staining confirmed the presence of enhanced formation of blood vessels in the MV-alginate-PCL and BMSC-MV-alginate-PCL groups ( Fig. 7 ). These results indicated that it is feasible to incorporate MSC-MVs into alginate-PCL constructs to enhance tissue-engineered bone regeneration by promoting vascularization. It is widely accepted that synthetic polymer scaffolds are poor at mimicking the natural microenvironment of autologous tissues, especially bone. Appropriate mechanical properties and porosity for the in-growth of vessels, as well as biocompatibility, are required for bone tissue engineering scaffolds ( Rath et al. , 2012 ). Alginate-PCL constructs can partly fulfill these requirements. PCL is an FDA-approved bioresorbable polymer for implantation. Porous PCL scaffolds produced by 3D printing with a controlled diameter, range of shape design options, and high mechanical strength have been investigated for their potential in the repair of bone defects ( Bao et al. , 2015 ). In addition, alginate, as a natural hydrogel with good biocompatibility and biodegradability, is structurally similar to extracellular matrix and can encapsulate various growth factors ( Lee et al. , 2013 ; Rufaihah & Seliktar, 2015 ). Alginate-PCL constructs have been used as dental and orthopedic implants in several studies ( Jang et al. , 2013 ; Kim, Jung & Kim, 2013 ; Kim & Kim, 2014 ). However, both PCL and alginate are poor at promoting angiogenesis, which is essential for bone repair. To the best of our knowledge, the present work describes for the first time the encapsulation of MSC-MVs in an alginate-PCL construct, thus enhancing tissue-engineered bone formation by improving angiogenesis. Several studies have shown that MSC-MVs can mimic the beneficial effects of MSCs, such as the ability to promote angiogenesis ( Zhang et al. , 2012 ; Bian et al. , 2014 ; Lopatina et al. , 2014 ). Owing to the fact that MSCs can be incorporated directly into a hydrogel, we hypothesized that MSC-MVs with a phospholipid bilayer may also work in a hydrogel system. Immunohistochemical analyses showed that MSC-MVs improved vascularization in vivo and thereby confirmed our hypothesis. We couldn’t rule out the possibility that MSC-MVs might also promote osteogenic differentiation of the seed cells, so we cultured MSCs in osteogenic differentiation medium in the presence of MSC-MVs or PBS (control) and replaced the medium and MSC-MVs every 3 days for 2 weeks. Then quantitative real-time polymerase chain reaction (qRT-PCR) was performed to detect mRNA expression of osteogenesis-associated genes like Runt-related transcription factor 2 (RUNX2), osteocalcin (OCN), and osteopontin (OPN). The results showed no significant difference in the expression of the osteogenesis-associated genes between the two groups (data not shown), suggesting that the significantly increased newly regenerated bone in the BMSC-MV-alginate-PCL group might be mainly due to the proangiogenic ability of MSC-MVs. To promote vascularization, various approaches were developed in previous studies ( Moon & West, 2008 ). Delivery of proangiogenic growth factors like VEGF through their encapsulation or incorporation into scaffolds for controlled release was one of the most commonly used approaches. For example, alginate microparticles loaded with VEGF were incorporated into freeze-dried, collagen-based scaffolds to ensure sustained release of bioactive VEGF ( Quinlan et al. , 2015 ). However, current approaches for delivering growth factors are often associated with limited success, on account of their uncontrolled release of proteins, short half-life, high cost, and potential safety risks for clinical application ( Hettiaratchi et al. , 2014 ). The arteriovenous (A-V) shunt loop strategy, involving a microsurgical approach to achieve an anastomosis between arteries and veins, has been proven to be advantageous for axial vascularization of a scaffold ( Rath et al. , 2012 ). Nevertheless, this microsurgical method is also limited by the diameter and distribution of vessels. Compared with the above-mentioned approaches, MSC-MVs have several inherent advantages. First, they are enriched in angiogenesis-promoting biomolecules and angiogenesis-related mRNAs and miRNAs ( Chen et al. , 2014 ; Eirin et al. , 2014 ). Specifically, Chen et al. (2014) reported that MVs released by human umbilical cord-derived MSCs contained a variety of angiogenesis-promoting factors, including VEGF, interleukin-6, basic fibroblast growth factor, angiogenin, and monocyte chemotactic protein-1. In addition, Eirin et al. (2014) discovered that MVs released by porcine adipose tissue-derived MSCs preferentially expressed mRNAs and miRNAs involved in angiogenesis, which might induce genetic alteration of the recipient cells. Second, the isolation of MSC-MVs is more economical than the use of expensive growth factors like VEGF. MSCs are usually greatly expanded because large numbers of seed cells are essential to construct a tissue-engineered graft. Since MSCs secrete large numbers of MVs during culture, recycling of their culture supernatant, which is usually discarded during cell passaging, for the isolation of MSC-MVs would avoid unnecessary waste and reduce the cost. Third, accumulating evidence proves that allogeneic and even xenogeneic MVs have little or no toxicity and immunogenicity in immune-competent animals ( Gyorgy et al. , 2015 ). Furthermore, it has been suggested that MV-induced cell-to-cell communication can occur across species ( Gatti et al. , 2011 ; Aliotta et al. , 2012 ), which might solve the problem of the severe shortage of appropriate donors. Although MSC-MVs hold great potential in bone tissue engineering applications, there are still many issues needed to be addressed. Firstly, the exact mechanism underlying the proangiogenic effect of MSC-MVs remains unclear. Future studies could investigate the genetic and epigenetic changes in target cells induced by MSC-MVs and the bioactive cargos enclosed in MSC-MVs. Secondly, a number of studies have demonstrated that MSC-MVs could exert anti-apoptotic effects on injured cells ( Bruno et al. , 2012 ; Lin et al. , 2014 ). Since the seed cells in the core area of the scaffold often suffer from hypoxia and poor nutrient supply due to the lack of vascularization ( Moon & West, 2008 ), it is worthwhile to investigate whether MSC-MVs may exert other beneficial effects on the seed cells in future. Thirdly, we did not compare the effects of MSC-MVs with proangiogenic growth factors in the present study, future study could investigate whether MVs could replace growth factors to achieve better outcomes. Although many issues remain to be addressed, the MV-alginate-PCL construct described herein represents a promising approach for promoting vascularization in tissue-engineered grafts. This composite graft has enormous potential for tissue engineering and regenerative medicine, and may be applied to the regeneration of other tissues and organs. Conclusions The present study developed a novel MV-alginate-PCL construct that takes advantage of the proangiogenic properties of MSC-MVs. MVs could be harvested from the culture medium of MSCs during their expansion and incorporated into alginate-PCL constructs. When combined with osteodifferentiated MSCs, theses constructs led to notable increases in vessel formation and tissue-engineered bone regeneration in a subcutaneous bone formation model in nude mice. Taking all of the findings into consideration, these constructs may offer a novel proangiogenic strategy for bone tissue engineering applications. Supplemental Information 10. 7717/peerj. 2040/supp-1 Figure S1 Characterization of BMSC-MV-alginate-PCL constructs. Representative images of the core area (A) and margin area (B) of the BMSC-MV-alginate-PCL constructs observed by phase-contrast microscopy, scale bars: 200 µm. Representative images of the BMSC-MV-alginate-PCL constructs observed by confocal microscopy (C and D), scale bars: 100 µm. Click here for additional data file.

10. 7717/peerj. 2229

PeerJ

Background The receptor activator of nuclear factor kappa-B (RANK)/RANK ligand/osteoprotegerin (OPG) system plays a critical role in bone remodelling by regulating osteoclast formation and activity. OPG has been used systemically in the treatment of bone diseases. In searching for more effective and safer treatment for bone diseases, we investigated newly formulated OPG-chitosan complexes, which is prepared as a local application for its osteogenic potential to remediate bone defects. Methods We examined high, medium and low molecular weights of chitosan combined with OPG. The cytotoxicity of OPG in chitosan and its proliferation in vitro was evaluated using normal, human periodontal ligament (NHPL) fibroblasts in 2D and 3D cell culture. The cytotoxicity of these combinations was compared by measuring cell survival with a tetrazolium salt reduction (MTT) assay and AlamarBlue assay. The cellular morphological changes were observed under an inverted microscope. A propidium iodide and acridine orange double-staining assay was used to evaluate the morphology and quantify the viable and nonviable cells. The expression level of osteopontin and osteocalcin protein in treated normal human osteoblast cells was evaluated by using Western blot. Results The results demonstrated that OPG in combination with chitosan was non-toxic, and OPG combined with low molecular weight chitosan has the most significant effect on NHPL fibroblasts and stimulates proliferation of cells over the period of treatment.

Introduction Osteoprotegerin (OPG) is a secretory glycoprotein of the tumour necrosis factor (TNF) receptor, which is highly expressed in adult bone, lung, heart, kidney and placenta ( Nagasawa et al. , 2002 ; Baharuddin et al. , 2015 ). The role of OPG in the pathological aspects of bone diseases, such as osteoporosis associated with estrogen deficiency and periodontal disease, has been well established ( Bostanci et al. , 2007 ; Koide et al. , 2013 ; Hofbauer et al. , 1999 ). In the medical field, OPG therapy has been used to reduce bone resorption and to enhance osseous healing ( Kostenuik et al. , 2001 ; Bekker et al. , 2001 ); the therapeutic strategies are based on OPG’s potent inhibitory action on osteoclast differentiation and function. In tumour-bearing mice, OPG treatment reduced osteoclast activity. Bekker et al. (2001) extended gene therapy investigations into human clinical trials by investigating the safety and tolerability of OPG as well as the bone anti-resorptive effects. This study showed that a single dose of OPG rapidly decreased bone resorption in post-menopausal women. Thus, blocking RANKL using OPG may be effective in the treatment of bone diseases characterized by increased bone resorption, such as osteoporosis. Chitosan is a natural, cationic carbohydrate polymer derived from chitin by partial deacetylation ( Ilium, 1998 ). It has been shown to be biocompatible and biodegradable both for in vitro and in vivo conditions ( Joshi et al. , 2012 ; Coimbra et al. , 2011 ). Chitosan also has high affinity to proteins, adheres well to mucosa and demonstrates antifungal effects; thus, it makes an ideal material for biomedical applications. To our knowledge, there is no study investigating the use of a drug delivery system with a polymer/polysaccharide matrix, such as chitosan, to deliver OPG locally. Therefore, in this study, we attempt to evaluate the cytotoxicity of low, medium and high molecular weights of chitosan (LMW, MMW and HMW respectively) and new combinations of OPG and chitosan (OPG-chitosan complexes) and their in vitro effect on normal, human periodontal ligament (NHPL) fibroblast cells and study the effect on bone marker production from normal human osteoblast cells. Materials and Methods Materials Low molecular weight (LMW), medium molecular weight (MMW) and high molecular weight (HMW) chitosan and human OPG protein (Recombinant Human OPG; PeproTech, Rocky Hill, New Jersey, USA) were used in this study. Tris buffer (5 mmol L −1, pH 7. 5) and dimethyl sulfoxide (DMSO) (Fisher Scientific, Leics, UK) were used throughout the experiment. Normal, human periodontal ligament (NHPL) fibroblasts and normal, human osteoblasts were obtained from Lonza (Lonza Inc. , Walkersville, MD, USA). Penicillin-streptomycin (Bioscience Ltd, Buckingham, UK), 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) propidium iodide and acridin orange (Sigma-Aldrich, St. Louis, MO, USA) were also used. Cell culture NHPL fibroblast cells were cultured and maintained in Dulbecco’s Modified Eagle’s Medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA). The medium was supplemented with 10% foetal bovine serum (FBS) with 1% antibiotics (penicillin-streptomycin) and incubated in 5% CO 2 at 37 °C. The medium was changed twice a week until a confluent cell monolayer was formed and observed under an inverted microscope. Cell viability assay We evaluated the effect of different molecular weights of chitosan (LMW, MMW and HMW) and OPG on cell viability after treatment. The cultured cells were trypsinized, seeded in 96-well micro plate (8 × 10 3 cells/well) and incubated at 37 °C in 5% CO 2 for 24, 48 and 72 h to allow cell attachment as described earlier ( Souza et al. , 2010 ). The medium was freshened and treated with serial dilutions of chitosan (100, 50, 25, 12. 5, 6, 3, 1. 5, μg mL −1 ) and OPG (30, 15, 7. 5, 3, 1. 5, 0. 75, 0. 35, 0. 19, 0. 09, 0. 045, 0. 024 μg mL −1 ) then incubated for 24, 48 and 72 h. Following incubation, 20 μL of tetrazolium bromide (MTT) (5 mg mL −1 ) solution was added to each well followed by incubation for 4 h. All remaining supernatant was removed and 100 μL of DMSO was added to dissolve the crystal formation. The optical density was measured at a wavelength of 570 nm using a microplate reader (Tecan Infinite M 200 PRO; Tecan, Männedorf, Switzerland). Cell proliferation assay A cell proliferation assay was carried out by using MTT on various concentrations of OPG-chitosan combinations. The different MWs of chitosan (at fixed concentration) combined with high, moderate and low concentrations of OPG were selected from the results of the cell viability assay. NHPL cells (8 × 10 3 /well) were treated with the OPG-chitosan combinations together with control samples (cells without treatment) and incubated for 24, 48 and 72 h. Each sample was assayed in triplicate. At the end of each incubation period, the previously explained procedure was used in another viability assay to measure the optical density by MTT assay. Morphological observation NHPL cells were seeded into 24-well microtiter plate (with a density 3 × 10 4 cells/well) and were incubated for 24 h at 37 °C and 5% CO 2. Then the cells were treated with different MWs of chitosan, combined with 0. 024 μg mL −1 OPG and incubated for 24, 48 and 72 h. The cellular morphological changes of the treated and untreated cells were observed under a Leica DM IRB (Leica Microsystems, Wetzlar, Germany) inverted microscope and compared with untreated, viable cells. Acridine orange and propidium iodide (AOPI) double-staining assay Acridine orange (AO) and propidium iodide (PI) are nuclear DNA staining (nucleic acid binding) dyes. AO is permeable to both live and dead cells and stains all nucleated cells, generating green fluorescence. This assay was conducted to assess the morphology and quantify the viable and nonviable (apoptotic) cells. Viable cells are indicated by green nuclei with round intact structure and nonviable cells will display orange-to-red areas. NHPL fibroblast cells were quantified using AOPI staining, according to standard procedures and examined under a fluorescence microscope (Lieca attached with Q-Floro Software, Solms, Germany). The treatment was carried out in a 25 mL culture flask (Nunc, Roskilde, Denmark). NHPL fibroblast cells were cultured at a concentration of 2 × 10 5 cell mL −1 and treated with different MWs of chitosan (LMW, MMW and HMW) combined with a 0. 024 μg mL −1 concentration of OPG. Flasks were incubated in an atmosphere of 5% CO 2 at 37 °C for 24, 48 and 72 h. The cells were then spun down at 220 g for 5 min. The supernatant was discarded and the cells were washed twice, using cold PBS after being centrifuged at 220 g for 5 min to remove the remaining media. Five microliters of fluorescent dye containing AO (10 μg mL −1 ) and PI (10 μg mL −1 ) were added into the cellular pellet at equal volumes. The freshly stained cell suspension was dropped onto a glass slide and covered with a cover slip. The slides were then observed under the fluorescence microscope within 30 min before the fluorescent colour begin to fade. The percentages of viable and nonviable cells were determined based on the morphological criteria assessed under the UV-fluorescence microscope. 3D cell encapsulation in cell culture plates In 96 well plates, BD PuraMatrix/cell/sucrose mixture was prepared according the protocol to the center of the well carefully, without introducing bubbles ( Abu-Yousif et al. , 2009 ). The cells were seeded into 96-well plate at a density of 10 × 10 3 per well. After the cells have been plated in all wells, gelation of the PuraMatrix was initiated by gently running culture media down the side of the well on top of the hydrogel. The media was changed gently two times over the next one hour to further equilibrate the pH of the hydrogel. The treatment was started at 7days post cell seeding and exposed for 24, 48 and 72 h. At the end of the treatment, cell viability was estimated by AlamarBlue assay (ThermoFisher Scientific, Hempstead, UK). Western blot Normal human osteoblasts were incubated in osteogenic medium at controlled conditions (5% CO2, 95% air and 37 °C). Osteoblast of second passage was used in this experiment. 75 mL cell culture flasks were used to seed the control and the cells were treated with OPG (0. 024 μg mL −1 and LM chitosan (100 μg mL −1 ). Subsequently, these cells were incubated for 24, 48 and 72 h. PRO-PREPTM (iNtRON, Biotechnology, Korea) was used to extract the whole proteins, and NanoOrange protein quantitation kit (Invitrogen) was employed for protein quantification. The proteins were transformed with nitrocellulose paper. The nitrocellulose blot was blocked with a solution of 4% (W/V) dry milk for 3 h. The blot was incubated with monoclonal osteopontin (clone 1B20; Novus Biologicals, Cambridge, UK) and osteocalcin (OCG3; Abcam, Cambridge, UK) at a 1:1000 dilution for overnight. The goat anti-mouse secondary antibody was added at dilution of 1:5000 for 3 h. After the final wash, the proteins bands were detected on the membrane with calorimetric horseradish peroxidase (HRP) substrate, 4-chloro-1-naphthol (4CN) (Bio-Rad) kit. To capture images of the membrane, a UV gel documentation system (Biospectrum 410; UVP) was used. Real-time PCR analysis The NHPL fibroblast cells were seeded and cultured as described earlier, then the cells were treated with OPG (0. 024 μg mL −1 and LM chitosan (100 μg mL −1 ). Subsequently, these cells were incubated for 24, 48 and 72 h. RNA was extracted by RNeasy Mini Kit (Qiagen, Valencia, CA, USA). The concentration and purity of RNA were measured by using the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). cDNA was synthesized from the RNA following the manufacturer’s instructions using High Capacity RNA-to-cDNA Kit for RT-PCR (Applied Biosystems, Foster City, CA, USA). For the PCR reaction, TaqMan ® Fast Advanced Master Mix was used (ThermoFisher Scientific, Hempstead, UK). The primers (for caspase 8, BCL2) that were used in this experiment. Glyceraldehyde phosphate dehydrogenase (GAPDH), a housekeeping gene, was used as an internal control. All of the assays were conducted in 96-well PCR plates using the StepOne Plus Real-Time PCR system (Applied Biosystems Inc, Waltham, MA, USA). The reactions were performed in triplicate. The mRNA level of each gene relative to that of GAPDH was calculated using the comparative quantification method Statistical analysis For each microplate, reading values calculated from the exposed cells were converted into percentages with the negative control values considered to be 100%. The data was reported as the mean ± standard deviation (SD). Result The cell viability after treated with different MW chitosan Table 1 summarises and compares the percentage of viable NHPL fibroblasts following treatment with different MWs of chitosan (LMW, MMW and HMW) over different time exposures and under control conditions using the MTT assay. Regardless of chitosan MW, the viability of the NHPL cells was ≥90%, following 24, 48 and 72 h of exposure as compared to the untreated cells. 10. 7717/peerj. 2229/table-1 Table 1 The percentage of NHPL fibroblast cells viability treated with different MWs of chitosan. Time exposure (h) Percentage of cell viability LMW of chitosan 0. 1 mg/mL MMW of chitosan 0. 1 mg/mL HMW of chitosan 0. 1 mg/mL 24 ≥100% ≥92% ≥90% 48 ≥100% ≥100% ≥96% 72 ≥100% ≥100% ≥100% Notes. The viability percentage values were obtained from the MTT assay. The viability assay of OPG Figure 1 summarises the percentages of viability according to different OPG concentrations at different exposure times. In general, the viability of cells decreased gradually as the OPG concentration and exposure time increased. At 30 μg mL −1, the viability was reduced to less than 80% after 48 h and to less than 60% after 72 h. The viability of cells treated with 0. 024–1. 5 μg mL −1 OPG was ≥90% after 72 h of exposure. 10. 7717/peerj. 2229/fig-1 Figure 1 Percentage of NHPL fibroblast cells viability after treatment with different OPG doses (0–30 µg/mL) after 24, 48 and 72 h. Proliferation assay of OPG The effect of OPG on cell proliferation was studied in vitro. The cells treated with OPG showed optical densities for two concentration levels (0. 024 μg mL −1 and 0. 18 μg mL −1 ) that were higher than cells treated with 1. 5 μg mL −1 OPG and the controls (zero concentration OPG). Figure 2 shows the growth rates of the treated cells compared to the untreated cells. The highest cell proliferation rate was displayed at the concentration of 0. 024 μg mL −1 after 72 h of incubation. 10. 7717/peerj. 2229/fig-2 Figure 2 Comparison of cell proliferation rates at control, low (0. 024 µg mL −1 ), moderate (0. 15 µg mL −1 ) and high (1. 5 µg mL −1 ) doses of OPG compared to control. Proliferation assay of LMW chitosan combined with different concentrations of OPG When combined with 0. 024 μg mL −1 OPG the LMW chitosan showed a higher cell proliferation rate than LMW chitosan combined with 1. 5 and 0. 18 μg mL −1 concentrations of OPG ( Fig. 3 ). 10. 7717/peerj. 2229/fig-3 Figure 3 The effect of LMW chitosan combined with low (0. 024 µg mL −1 ), moderate (0. 15 µg mL −1 ) or high (1. 5 µg mL −1 ) doses of OPG on NHPL fibroblasts proliferation in-vitro. NHPL Fibroblasts were treated for 24, 48 and 72 h. Proliferation assay of MMW chitosan combined with different concentrations of OPG The cells were treated with MMW chitosan (100 μg mL −1 combined with different concentration of OPG (1. 5, 0. 18, 0. 024 μg mL −1 ), the proliferation of cells was evaluated at 24, 48, 72 h. Chitosan combined with 0. 024 μg mL −1 OPG showed a higher cell proliferation rate than chitosan combined with 1. 5, 0. 18 μg mL −1 OPG over three different exposure time points ( Fig. 4 ). Proliferation assay of HMW chitosan combined with different concentrations of OPG The HMW chitosan combined with the same concentrations of OPG (0. 024, 0. 18 and 1. 5 μg mL −1 ) were used to evaluate the growth of the cells. The 1. 5 μg mL −1 OPG-chitosan combination induced a greater proliferation of cells than the other two combinations after 24 h. However, there was no marked difference in the proliferation of cells at 0. 024 μg mL −1 and 1. 5 μg mL −1 concentration levels after 72 h of incubation, and they showed an increase of cell proliferation, when compared to the control ( Fig. 5 ). 10. 7717/peerj. 2229/fig-4 Figure 4 The effect of MMW chitosan combined with OPG (low (0. 024 µg mL −1 ), moderate (0. 15 µg mL −1 ) or high (1. 5 µg mL −1 ) dose) on NHPL fibroblasts proliferation in-vitro. NHPL fibroblast cells were treated for 24, 48 and 72 h. 10. 7717/peerj. 2229/fig-5 Figure 5 The effect of HMW chitosan samples in different concentrations of OPG (low (0. 024 µg mL −1 ), moderate (0. 15 µg mL −1 ) or high (1. 5 µg mL −1 ) dose) on NHPL fibroblasts proliferation in-vitro. NHPL fibroblast cells were treated for 24, 48 and 72 h. Proliferation assay of three different MWs of chitosan combined with 0. 024 µg/mL OPG concentration using 3D culture system Previous results of the viability and proliferation assays ( Figs. 3 – 5 ) showed 0. 024 μg mL −1 concentration of OPG is the optimum concentration to use. It has been shown to be nontoxic and enhance the proliferation of cells. Another proliferation assay was carried out to determine the appropriate MW of chitosan (100 μg mL −1 ) to be used in combination with the 0. 024 μg/mL OPG. The result revealed that LMW chitosan in combination with the 0. 024 μg/mL OPG demonstrated higher cell proliferation, compared to MMW and HMW chitosan ( Fig. 6 ). 10. 7717/peerj. 2229/fig-6 Figure 6 The effect of HMW, MMW and LMW chitosan combined with 0. 024 µg/mL of OPG on NHPL fibroblasts proliferation in-vitro. NHPL fibroblast cells were treated for 24, 48 and 72 h. Morphological changes of cells after treatment with OPG-chitosan combinations In general, regardless of the MW of chitosan, the treated cells appeared to have normal morphology, such as flattened surfaces, and they adhered to the surface of well with very minimal rounded cells (not attached cells) at the various exposure times. Figure 7 is a phase contrast image showing morphological changes of NHPL fibroblast cells, following treatment with LMW, MMW and HMW chitosan in combination with 0. 024 μg/mL concentration of OPG. We noticed that there was no difference between the morphology of cells treated with gels compared to the control group of untreated cells. We suggest that OPG had no effect on the morphology of the cells. 10. 7717/peerj. 2229/fig-7 Figure 7 The morphological changes of NHPL fibroblast cells observed under an inverted microscope after 24, 48 and 72 h. (A) Control (untreated cells), (B) OPG-LMW chitosan combination, (C) OPG-MMW chitosan combination and (D) OPG-HMW chitosan combination. Quantification of the cell viability using AOPI double-staining In general, following exposure to different OPG-chitosan combinations at different time exposures, most of the cells were viable and showed fluorescent green with intact cell walls and minimal amounts of dead cells. Figure 8 shows the fluorescent images of the NHPL fibroblast cells, following treatment with LMW, MMW and HMW chitosan in combination with 0. 024 μg/mL OPG after 24, 48 and 72 h. Table 2 shows cell viability percentages after they were treated with OPG in combination with different MWs of chitosan. The cell viability percentages ranged from 87% to 92%, 89% to 92% and 89% to 92% after 24, 48 and 72 h respectively 10. 7717/peerj. 2229/fig-8 Figure 8 AOPI Viability: dual fluorescence for viable and nonviable cells treated with OPG-chitosan combinations and untreated cells. Images were observed at 50× magnification. (A) Control (untreated cells), (B) OPG-LMW chitosan combination, (C) OPG-MMW chitosan combination and (D) OPG-HMW chitosan combination. Osteopontin and osteocalcin protein levels Western blot analysis was used to examine the expression level of osteopontin and osteocalcin protein in treated normal human osteoblast cells with LMW chitosan combined with 0. 024 μg mL −1 of OPG. Results in Fig. 9 revealed that LMW chitosan combined with 0. 024 μg mL −1 of OPG induce the expression levels of bone formation marker proteins osteopontin and osteocalcin in time-dependent manner. At 24 h post treatment, proteins level were lower and increased after 48 and 72 h treatment. mRNA expression of caspase 8 and BCL2 A real-time PCR analysis was performed to investigate the differences in the mRNA expression levels of caspase 8 and Bcl-2 which are apoptosis-related genes. The results showed the downregulation of caspase 8 and upregulation of Bcl-2 expression after treatment with LMW chitosan combined with 0. 024 μg mL −1 of OPG ( Fig. 10 ). 10. 7717/peerj. 2229/table-2 Table 2 The cell viability percentages treated with OPG combined with different MW chitosan. Combinations 24 (h) 48 (h) 72 (h) Control (untreated cells) 87 ± 10% 89 ± 13% 92 ± 14% OPG in LMW chitosan 91 ± 16% 85 ± 15% 89 ± 20% OPG in MMW chitosan 92 ± 19% 92 ± 18% 90 ± 11% OPG in HMW chitosan 88 ± 12% 90 ± 9% 91 ± 7% Notes. The viability percentages of different MW chitosan were combined with OPG on NHPL cell lines in vitro at 24, 48 and 72-hour treatments. The viability percentage values were obtained from the AOPI double-staining assay. Data are reported as means ± SD for measurements in triplicate. 10. 7717/peerj. 2229/fig-9 Figure 9 Effect of LMW chitosan combined with 0. 024 μg/mL of OPG on osteopontin (A) and osteocalcin (C) proteins expression at 24, 48 and 72 hours. GAPDH (B) was used as a loading control. (D) Quantitative analysis of treatment. All data were expressed as means ± standard deviation (SD). 10. 7717/peerj. 2229/fig-10 Figure 10 Downregulation of caspase 8 and upregulation of Bcl-2 expression in NHPL fibroblast cells. The expression of caspase 8 and Bcl-2 after 24, 48 and 72 h of treatment was studied by RT-PCR. GAPDH was used as internal (positive). Discussion Different MWs of chitosan showed no marked inhibition on the viability of the NHPL fibroblast cells. This study also confirmed that chitosan has the ability to induce proliferation of NHPL fibroblast cells, which serve a role in bone healing. We used NHPL fibroblasts in this study as Fibroblasts are one of the important cells involved in the healing process. The NHPL fibroblast cell is one the main cells that contribute in the periodontal bone regeneration as it has osteoblast-like properties such as alkaline phosphatase activity, vitamin D-dependent production of osteocalcin and initiation of mineral-like nodules in the presence of a supportive medium ( Jönsson et al. , 2011 ; Scanlon et al. , 2011 ; Basdra & Komposch, 1997 ). The extent of chitosan’s ability to regenerate bone is still debated. Spin-Neto and co-workers ( Spin-Neto et al. , 2012 ; Spin-Neto et al. , 2010 ) revealed that there was no significant bone formation following chitosan and chitosan hydrochloride gel application in critical sizes of bone defects; the defects were repaired by connective tissue with variable degrees of inflammation. On the other hand, Muzzarelli et al. (1994) reported that chitosan enhanced osseous healing of defects created in sheep. Jung et al. (2000) also reported that chitosan has significant effects on the regeneration of bone tissue in calvarial defects in rats. Other studies have reported that chitosan and chitosan-based biomaterials, tested in the treatment of bone defects, have a high degree of biocompatibility, osteoconductivity and increased the density of newly formed bone ( Bojar et al. , 2014 ; Florczyk et al. , 2013 ; Jung et al. , 2013 ; Lee et al. , 2000 ), and chitosan derivatives have the ideal properties of biocompatible materials tested on normal human fibroblasts ( Spin-Neto et al. , 2012 ; Spin-Neto et al. , 2010 ; Muzzarelli et al. , 1994 ). Biocompatibility is one of the most important criteria in selecting biomaterials. Clinically, chitosan also has high potential in dental applications, and it has used to repair socket after dental extraction ( Ezoddini-Ardakani, 2011 ). Other research also reported that biodegradable dental chip containing chlorhexidine or thymoquinone was applied for management of chronic periodontitis in patients ( Al-Bayaty et al. , 2013 ; Jothi et al. , 2009 ). With regards to OPG, the viability of NHPL fibroblast cells was significantly high at concentrations of 0. 024 and up to 3 μg mL −1 of OPG. Previous studies have reported that OPG acts as a survival factor, at least in vitro, by blocking TRAIL-induced apoptosis ( Lane et al. , 2012 ; Holen et al. , 2002 ). Several preclinical studies that used OPG systemically for treatment of bone disorders have revealed that OPG inhibits bone resorption and improves osteoblastogenesis and new bone formation ( Yao et al. , 2011 ; Jin et al. , 2007 ; Lamoureux et al. , 2007 ). Other studies involving clinical use of OPG to treat bone loss in post-menopausal women revealed that biochemical markers of bone resorption were reduced and OPG was able to lessen the amount of bone turnover ( Bekker et al. , 1999 ). Body and co-workers ( Body et al. , 2003 ) stated that OPG was accepted as a treatment for patients with bone disease related to breast carcinoma or multiple myeloma and is thus effective in reducing levels of bone resorption markers. This is the first study to report on the toxicity evaluation of OPG in cells. The results demonstrated that a low concentration of OPG (0. 024 mg mL −1 ) has the greatest ability to induce proliferation of NHPL fibroblast cells compared to other concentrations. On the other hand, the proliferation of the cells was greater when OPG (0. 024 mg mL −1 ) was combined with chitosan (low and moderate molecular weights) compared to separate treatments with chitosan and OPG; but the combination with high molecular weight chitosan exhibited no difference among three doses of OPG. 3D cell culture models produce a pragmatic microenvironment and simulate an in vivo system, which aids to understand cell–cell interactions ( Yamada & Cukierman, 2007 ). Cells cultured in a 3D environment have the ability to acquire phenotypes and respond to stimuli similar to in vivo biological systems ( Prestwich, 2007 ; Godugu et al. , 2013 ). This study showed that the LMW chitosan combined with OPG had the greatest ability to induce cells proliferation compared to the moderate and high molecular weights of chitosan combined with OPG. This is in accordance with Chen et al. (2002) and Nor Asiah et al. (2013) who have reported that low molecular weight chitosan significantly promoted growth of normal fibroblasts. Other studies also reported that fibroblasts treated with low molecular weight chitosan stimulated fibroblasts proliferation compared to chitosan at higher molecular weights ( Tangsadthakun et al. , 2007 ; Wang et al. , 2007 ). The results also verified that the rate of NHPL fibroblast cells proliferation increased with time exposure to the OPG-chitosan matrixes and did not cause toxicity effects on fibroblast cell growth. This is also in agreement with a study reported by Nor Asiah et al. (2013). Regarding the effect of on the activity of NH osteoblast cells, human osteoblasts were treated with LMW chitosan combined with OPG with optimal concentration confirmed by the proliferation assays. The results showed that the osteocalcin and osteopontin levels were increased as time of exposure increased. Based on these findings, we may postulate this treatment has ability to enhance the differentiation of cells as Celic and coworkers have demonstrated that the production of some bone protein such as osteocalcin was increased in differentiated cells as it is late marker of bone formation ( Celic et al. , 1998 ). Osteocalcin is a mineralization-specific marker because its expression increases as the mineralization increase ( Yamada et al. , 2013 ). Xiao et al. (2004) reported that there is a correlation between calcification and the distribution of bone sialoprotein and osteopontin. OPG seems also play a key role in cell survival, via its interaction with TNF-related apoptosis-inducing ligand (TRAIL). OPG can act as a decoy receptor for TNF-related apoptosis inducing ligand, because it is efficiently binds with TRIAL. TRAIL signaling leads to cell death by activation of caspase-8 leads to caspase cascade that culminate in cell death ( Baud’huin et al. , 2013 ; Crowder & El-Deiry, 2012 ). Also, overexpression of Bcl-2 was found to inhibit the TRIAL-induced caspase 8, thus inhibiting the TRAIL-induced apoptosis in many cells. The results of RT-PCR revealed that the downregulation of caspase 8 and upregulation of Bcl-2 that may promote cell proliferation. These are in agreement with previous studies that reported the exogenous application of recombinant OPG has indeed been shown to be capable of inhibiting TRAIL-induced apoptosis and subsequently downregulate the caspase 8 expression ( Lemke et al. , 2014 ; Miyashita et al. , 2004 ; Shipman & Croucher, 2003 ). The LMW chitosan has structural characteristics similar to those of the glycosaminoglycans that facilitate the migration and proliferation of cells and the also the OPG has effects on the survival of cells, thereby the OPG-chitosan combination facilitate the tissue regeneration. This is the first study that evaluated the combination effect of OPG with different molecular weights of chitosan on osteoblast and NHPL fibroblast. The results of this study indicate the LMW chitosan combined with OPG has potential to be used as a biomaterial for bone tissue engineering. Conclusion Our results have suggested OPG in low molecular weight chitosan matrixes enhances cell growth and proliferation, and induce the production of osteopontin and osteocalcin protein levels. It can be used in different local preparations for potential bone defect application. Further study is necessary to clarify the effect of combining OPG and chitosan for bone management applications. Supplemental Information 10. 7717/peerj. 2229/supp-1 Table S1 The viability assay of OPG Click here for additional data file. 10. 7717/peerj. 2229/supp-2 Table S2 Proliferation assay of OPG Click here for additional data file. 10. 7717/peerj. 2229/supp-3 Table S3 Proliferation assay of LMW chitosan combined with different concentrations of OPG Click here for additional data file. 10. 7717/peerj. 2229/supp-4 Table S4 Proliferation assay of LMW chitosan combined with different concentrations of OPG Click here for additional data file. 10. 7717/peerj. 2229/supp-5 Table S5 Proliferation assay of HMW chitosan combined with different concentrations of OPG Click here for additional data file. 10. 7717/peerj. 2229/supp-6 Table S6 Proliferation assay of three different MWs of chitosan combined with 0. 024 µg/mL OPG concentration using 3D culture system Click here for additional data file. 10. 7717/peerj. 2229/supp-7 Table S7 Osteopontin and osteocalcin protein levels Click here for additional data file.

10. 7717/peerj. 2243

PeerJ

Improved cartilage regeneration by implantation of acellular biomaterials after bone marrow stimulation: a systematic review and meta-analysis of animal studies

Microfracture surgery may be applied to treat cartilage defects. During the procedure the subchondral bone is penetrated, allowing bone marrow-derived mesenchymal stem cells to migrate towards the defect site and form new cartilage tissue. Microfracture surgery generally results in the formation of mechanically inferior fibrocartilage. As a result, this technique offers only temporary clinical improvement. Tissue engineering and regenerative medicine may improve the outcome of microfracture surgery. Filling the subchondral defect with a biomaterial may provide a template for the formation of new hyaline cartilage tissue. In this study, a systematic review and meta-analysis were performed to assess the current evidence for the efficacy of cartilage regeneration in preclinical models using acellular biomaterials implanted after marrow stimulating techniques (microfracturing and subchondral drilling) compared to the natural healing response of defects. The review aims to provide new insights into the most effective biomaterials, to provide an overview of currently existing knowledge, and to identify potential lacunae in current studies to direct future research. A comprehensive search was systematically performed in PubMed and EMBASE (via OvidSP) using search terms related to tissue engineering, cartilage and animals. Primary studies in which acellular biomaterials were implanted in osteochondral defects in the knee or ankle joint in healthy animals were included and study characteristics tabulated (283 studies out of 6, 688 studies found). For studies comparing non-treated empty defects to defects containing implanted biomaterials and using semi-quantitative histology as outcome measure, the risk of bias (135 studies) was assessed and outcome data were collected for meta-analysis (151 studies). Random-effects meta-analyses were performed, using cartilage regeneration as outcome measure on an absolute 0–100% scale. Implantation of acellular biomaterials significantly improved cartilage regeneration by 15. 6% compared to non-treated empty defect controls. The addition of biologics to biomaterials significantly improved cartilage regeneration by 7. 6% compared to control biomaterials. No significant differences were found between biomaterials from natural or synthetic origin or between scaffolds, hydrogels and blends. No noticeable differences were found in outcome between animal models. The risk of bias assessment indicated poor reporting for the majority of studies, impeding an assessment of the actual risk of bias. In conclusion, implantation of biomaterials in osteochondral defects improves cartilage regeneration compared to natural healing, which is further improved by the incorporation of biologics.

Introduction Articular cartilage is a specialized tissue that covers joint surfaces and provides a low-friction and load-bearing surface for a smooth motion of joints. The structure and function of the tissue can be compromised by traumatic injuries and degenerative joint diseases. Due to its avascular nature, damaged cartilage tissue does not heal spontaneously and it remains a challenge to fully restore tissue function ( Ahn et al. , 2009 ; Cao et al. , 2012 ). The surgical options to treat patients with a localized cartilage defect are limited to cartilage regeneration approaches such as autologous chondrocyte implantation and microfracture surgery ( Aulin et al. , 2013 ; Bal et al. , 2010 ). The latter strategy, also known as bone marrow stimulation, is relatively simple, minimally invasive and inexpensive. During this procedure the subchondral bone plate below the cartilage lesion is perforated to initiate bleeding and induce a reparative response. The principle behind this regenerative resurfacing strategy is the migration of non-differentiated bone marrow-derived multipotent stem cells from the subchondral bone into the defect site leading to the formation of new cartilage tissue ( Buma et al. , 2003 ; De Mulder et al. , 2014 ; Erggelet et al. , 2009 ). Patients treated with bone marrow stimulation generally show clinical improvements up to 1. 5–3 years after surgery. However, five years after surgery higher incidences of clinical failures are observed ( Hoemann et al. , 2010 ; Van der Linden et al. , 2013 ). The newly formed tissue generally consists of fibrocartilage repair tissue rather than hyaline cartilage, has limited filling of the defect, integrates poorly with the surrounding tissue and has inferior mechanical properties compared to hyaline cartilage ( Dai et al. , 2014 ). Therefore, the need for regeneration of more durable cartilage tissue persists. Regenerative medicine and tissue engineering may offer promising alternatives and/or additions to clinical strategies that aim to restore damaged cartilage tissue. The construction of biomaterials and the incorporation of cells and biologics in these implants have been widely investigated for this purpose. Biomaterials can be implanted in osteochondral defects created by applying marrow stimulating techniques (microfracture and subchondral drilling ( Falah et al. , 2010 )) to guide and stimulate the formation of cartilage tissue ( Seo et al. , 2014 ). During microfracture surgery, an arthroscopic awl is used to penetrate the subchondral bone, while with subchondral drilling a high speed drill is applied to penetrate the trabecular bone. Different strategies have been applied including the implantation of biomaterials with and without cells. Acellular biomaterials offer various advantageous properties such as lack of donor-site morbidity, absence of cell culture costs, off the shelf availability, fewer regulatory issues, and application of one-stage surgical procedures ( Brouwer et al. , 2011 ; Efe et al. , 2012 ). Many researchers have explored the approach of implanting acellular biomaterials and investigated the use of various biomaterials in vivo, such as natural (e. g. , collagen ( Breinan et al. , 2000 ; Buma et al. , 2003 ; Enea et al. , 2013 ; Wakitani et al. , 1994 ), chitosan ( Abarrategi et al. , 2010 ; Bell et al. , 2013 ; Guzman-Morales et al. , 2014 ; Hoemann et al. , 2007 ), alginate ( Igarashi et al. , 2012 ; Mierisch et al. , 2002 ; Sukegawa et al. , 2012 ) and hyaluronic acid ( Aulin et al. , 2013 ; Kayakabe et al. , 2006 ; Marmotti et al. , 2012 ; Solchaga et al. , 2000 )) and synthetic polymers (e. g. , polycaprolactone ( Christensen et al. , 2012 ; Martinez-Diaz et al. , 2010 ; Mrosek et al. , 2010 ), polyvinyl alcohol ( Coburn et al. , 2012 ; Holmes, Volz & Chvapil, 1975 ; Krych et al. , 2013 ) and poly(lactic-co-glycolic acid) ( Athanasiou, Korvick & Schenck Jr, 1997 ; Chang et al. , 2012 ; Cui, Wu & Hu, 2009 ; Fonseca et al. , 2014 )). To combine the advantageous properties of these materials, multilayered biomaterials (e. g. , β -tricalcium phosphate-hydroxyapatite/hyaluronate-atelocollagen ( Ahn et al. , 2009 ), ceramic bovine bone-gelatin/gelatin-chondroitin sulfate-sodium hyaluronate ( Deng et al. , 2012 )), blends (e. g. , poly(glycolic acid)-hyaluronic acid ( Erggelet et al. , 2009 ) and type I collagen-hyaluronic acid-fibrinogen hydrogel ( Lee et al. , 2012 )) have been constructed. Biologics are natural factors that can be used to stimulate tissue regeneration, e. g. , by inducing proliferation and differentiation of cells. Biologics such as growth factors of the transforming growth factor β (TGF- β ) superfamily and others have been incorporated in biomaterials to guide and stimulate the formation of hyaline cartilage tissue ( Richter, 2009 ). Moreover, it has been reported that the animal model of choice may have a significant impact on study outcome of articular cartilage regeneration ( Reinholz et al. , 2004 ). Currently, there is no systematic overview of the current literature assessing the effect of various parameters (e. g. , applied biomaterials, incorporated biologics and animal models) on cartilage regeneration. The aim of this systematic review and meta-analysis is to assess all current evidence for the efficacy of articular cartilage regeneration using acellular biomaterials implanted in the knee and ankle joint after microfracture and subchondral drilling in animal models. Additionally, we strive to provide transparency on the quality of performed in vivo studies, in order to aid the design of future animal experiments and clinical trials. We provide a systematic and unbiased overview of the current literature addressing regeneration of articular cartilage using a wide range of acellular biomaterials containing various biological cues (as illustrated in Fig. 1 ). Results of semi-quantitative histological scoring systems are used as a quantitative outcome parameter for outcome assessment of cartilage regeneration. Although microfracture surgery and subchondral drilling strive to stimulate cartilage and osteochondral regeneration, respectively, both are generalized in this study as cartilage regeneration. Moreover, the evaluation of different subgroups (natural and synthetic origin of the biomaterials, structure of the materials (scaffolds vs. hydrogels), incorporated biological cues, and animal models) was included to gain insights in which parameters affect cartilage regeneration and to what extent. 10. 7717/peerj. 2243/fig-1 Figure 1 Illustration of cartilage regeneration by implantation of biomaterials after bone marrow stimulation. The implanted biomaterials provide a template to guide cartilage regeneration by bone marrow derived mesenchymal stem cells. Materials and Methods Search strategy To identify relevant peer-reviewed articles, a comprehensive search of the literature using PubMed and EMBASE (via OvidSP) was conducted, using the methods defined by De Vries et al. (2012) and Leenaars et al. (2012). The last search date was April 3rd 2015. In both databases, a tissue engineering search component developed by Sloff et al. (2014), consisting of equivalents for tissue engineering (e. g. , tissue regeneration, regenerative medicine, bio-engineering or biomatrices), was combined with a cartilage search component, consisting of equivalents for cartilage and cartilage-related surgeries (e. g. , chondral, chondrogenic, surgery, microfracturing or implants). The search components were constructed using MeSH terms (PubMed) and EMTREE terms (EMBASE) and additional free-text words from titles or abstracts ([tiab] or ti, ab). The obtained tissue engineering-related and cartilage-related results were filtered for animal studies using previously described animal search filters ( De Vries et al. , 2011 ; Hooijmans et al. , 2010 ). The complete search strategy is attached in Supplemental Information 1. No language restrictions were used. Study selection References from the PubMed and EMBASE search strategies were combined and duplicates were manually removed from EndNote, with the preference of PubMed over EMBASE. All screening phases were performed by two independent reviewers (MP and VG) and reported according to the “Preferred Reporting Items for Systematic Reviews and Meta-Analysis” (PRISMA) guidelines ( Higgins & Green, 2011 ). References were first screened based on title and were excluded based on the following criteria: (1) titles showed no relevance to regeneration of articular (hyaline) cartilage, (2) it was specifically stated in the title that the conducted experiment was an in vitro study only, (3) osteoarthritis animal models were used, (4) only ex vivo studies were performed, and (5) deceased animals were used. In case of doubt or disagreement, references were included for further screening. The second screening phase consisted of a title/abstract screening in Early Review Organizing Software (EROS, Institute of Clinical Effectiveness and Health Policy, Buenos Aires, Argentina; www. eros-systematic-review. org ). References were included based on the following inclusion criteria: (1) primary study, (2) animal model, (3) bone marrow stimulation by microfracturing or creation of an osteochondral defect, and (4) biomaterial implantation. Articles were only excluded when it was specifically stated in the abstract that the study was performed without healthy animals or acellular biomaterials, or if biomaterials were not implanted in the knee or ankle joint. Articles were not excluded in case important information in the abstract was missing. These articles were assessed in the full-text screening phase. For the full-text screening, articles were included if they met all of the following inclusion criteria: (1) primary study, (2) animal model, (3) healthy animals, (4) articular cartilage regeneration, (5) knee or ankle joint, (6) bone marrow stimulation by microfracturing or creation of an osteochondral defect, and (7) implantation of an acellular biomaterial. In general, if results of the two reviewers were different, articles were discussed until consensus was reached. In case of double publication, one of the studies was removed. During the screening phase, no selection was made based on publication language. The risk of bias assessment and meta-analysis was applied to studies with a comparison between a non-treated empty defect control and biomaterial implantation, and with semi-quantitative histological scoring system results as outcome data. Study characteristics From the studies included after the full-text screening, the following details were obtained: general information (author and year of publication), animal characteristics (species, strain, sex, age, weight and the number of animals), information related to the surgical defect (size, depth and location), experimental conditions, biomaterial, biologics, evaluation time points and all outcome measures used, i. e. , macroscopic evaluation, semi-quantitative macroscopic evaluation, histology, immunohistochemistry, semi-quantitative histological scoring, and biomechanical tests. Data from semi-quantitative histological scorings were used in the meta-analysis (described in ‘Analysis preparations and meta-analysis’). Histological scoring systems applied in different studies consisted of scoring parameters like cell morphology, Safranin-O staining, integrity of surface, thickness, surface of area filled with cells, chondrocyte clustering, degenerative changes, restoration of the subchondral bone and integrity. Risk of bias assessment A risk of bias analysis was performed to assess the methodological quality of the studies included in the meta-analysis, using an adapted version of the risk of bias tool described by Hooijmans et al. (2014) (for all included studies containing a ‘non-treated empty defect’ as control group and studies using semi-quantitative histological scoring systems as outcome measure). A flowchart was constructed ( Supplemental Information 2 ) to score for selection, performance, detection and attrition bias, where the scores ‘−’, ‘?’ and ‘+’ indicate a low, unknown and high risk of bias, respectively. The questions addressed are specified in the Supplemental Information 2. Articles were scored independently by MP and VG, and if the results of the two reviewers were different, results were discussed until consensus was reached. All articles written in Chinese (16 studies) were excluded from the risk of bias assessment only, due to limited resources to independently translate these articles by two native Chinese speakers. However, the data of these studies were extracted and used in the meta-analysis. Analysis preparations and meta-analysis Analysis preparations The statistical analyses were restricted to those studies containing the outcome measure semi-quantitative histology, making a comparison between a ‘non-treated empty defect’ as control group and implanted biomaterials as experimental group. Data (mean, standard deviation (SD) and number of animals) of the control and experimental group were extracted from the studies, for all available time points. When results were not given numerically, but depicted graphically, the mean and SD were measured using ImageJ (1. 46r, National Institutes of Health USA). For studies presenting results in boxplots, the mean and standard deviation were recalculated from the median, range and the sample size according the method described by Hozo, Djulbegovic & Hozo (2005). When data were described by a mean and confidence interval (CI), the CI was recalculated to a standard deviation by the following equation: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\text{standard deviation}=\sqrt{N}\times \frac{\text{upper limit}-\text{lower limit}}{3. 92} $\end{document} standard deviation = N × upper limit − lower limit 3. 92 for a 95% CI ( Higgins & Green, 2011 ). For some studies, data were unclear and assumptions were made, which are listed in Supplemental Information 3. To compare studies with different histological score system scales, means and standard deviations were converted to a 100% scale by dividing the result by the maximum achievable histological score and multiplying by 100%. In case of missing or unclear data, authors were e-mailed to retrieve the data. When data could not be obtained, these studies were excluded from the meta-analysis (reasons for exclusion are also given in Supplemental Information 3 ). Results of studies with several experimental groups were combined, following the approach described in the Cochrane Handbook, table 7. 7 ( Higgins & Green, 2011 ). The same approach was followed to combine results of different animals on several time points in the same group in the same study. One study ( Hamanishi et al. , 2013 ) had an SD of zero, which caused problems in the analyses. Therefore, the SD was changed to 4. 29, equal to the SD of the experimental group of the same study at the same time point. The resulting data were used to calculate the treatment effect and corresponding standard error (SE) per study. Meta-analysis The following main research question was assessed: Does an overall beneficial effect exist of implanting acellular biomaterials in osteochondral defects compared to non-treated empty defects? First, in order to select the appropriate statistical random-effects meta-analysis model, we compared a univariate approach to the bivariate approach. In the bivariate approach, separate outcomes for control and experimental group were used with their respective SEs. The correlation between these two outcomes was modeled with a compound symmetry covariance matrix, as this resulted in a much lower Akaike Information Criterion value than the use of an unstructured covariance matrix. Results were compared with those of the univariate approach, based on the treatment effect and SE per study. Results of the univariate and bivariate approaches were very similar and we therefore proceeded with the univariate approach, when applicable in combination with likelihood ratio tests. Restricted to the experimental groups, the following sub-questions were addressed to evaluate whether the treatment effect depended on specific variables: (1) Is there a difference between the use of natural and synthetic biomaterials?; (2) Does the structure of the biomaterials affect cartilage regeneration?; (3) Do differences among various material subgroups exist?; (4) Does incorporation of biologics have a beneficial effect on cartilage regeneration compared to control biomaterials?; (5) Do differences among subgroups of biologics exist?; (6) Do different animal models result in variations in cartilage regeneration? Results are shown as % cartilage regeneration (95% CI: [lower CI, upper CI]. Some studies have more than one experimental group. Therefore, the total number of studies and number of experimental groups (no. of studies/groups) are provided. Sensitivity analyses were performed to evaluate the effect of time (e. g. , all time points, short (≤8 weeks), long time points (>8 weeks), or the maximum time point), outliers (excluding consecutively the studies with the 10% highest/lowest pooled SD, and studies with the 10% highest/lowest SE), implant location, bone marrow stimulating technique applied (microfracturing vs. subchondral drilling), language (excluding studies reported in Chinese as the risk of bias of these studies was not assessed), and excluding studies where assumptions had to be made. Based on a pilot analysis, data of all time points were used for subgroup analyses. Subgroup analyses were only performed for subgroups consisting of more than two groups. The statistical analyses were performed with SAS/STAT ® software version 9. 2 for Windows (SAS Institute Inc. , Cary, NC, USA). The funnel plot shows the overall outcome of the pooled effect size of each study. I 2 was used as a measure of heterogeneity. The forest plot was created with ReviewManager (RevMan, Version 5. 3, 2014; The Cochrane Collaboration, The Nordic Cochrane Centre, Copenhagen, Denmark). Results Search and study inclusion The searches conducted in PubMed and EMBASE ( Supplemental Information 1 ) resulted in 4, 401 and 5, 986 studies, respectively, leaving 6, 688 studies after removal of duplicates. These studies were screened by title and title/abstract, which resulted in 1, 088 included studies after the title screening and 517 included studies after the title/abstract screening. Screening articles by full-text and subsequently selection for studies with empty defect controls as well as semi-quantitative histology as outcome measure resulted in 283 included studies after full-text assessment, of which 151 and 135 articles could be used for the meta-analysis and risk of bias assessment, respectively ( Fig. 2 ). The studies from Xie et al. (2014), Yao, Ma & Zhang (2000) and Zhou & Yu (2014) could not be retrieved as a full text and these studies were therefore excluded. An overview of all included studies after full-text assessment as well as studies included for the risk of bias assessment and meta-analysis is provided in Supplemental Information 3. All references and abbreviations can be found in Supplemental Information 4. In this table, remarks are provided related to exclusion reasons for risk of bias assessment and meta-analysis (e. g. , duplicate publication and incomplete data). Assumptions made for certain studies are also stated in this table. 10. 7717/peerj. 2243/fig-2 Figure 2 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) flowchart of the systematic search of literature. Study characteristics The study characteristics ( Supplemental Information 3 ) clearly show substantial variation among studies. A wide range of animal species was used, from small (rat and rabbit) to larger animal models (dog, minipig, goat, pig, sheep and horse). A large variation was observed between the ages of animals (e. g. , the age of rabbits ranged from 6 weeks to >2 years). Often ages were not described or specified specifically (e. g. , as adult or mature). Generally, the animals were older (range of years) in large animal models compared to animals used in small animal models (range of months). The defects were created at different locations in the knee joint, such as the trochlea, condyle (medial and lateral), femur and intercondylar fossa. In addition, a large variation was found in the dimensions of the prepared defects, e. g. , the dimensions of the defects created in rabbits ranged from 4–7 mm in diameter and 0. 8–9 mm in depth. Microfracture surgery and subchondral drilling was performed in 25 and 258 studies, respectively. The implanted biomaterials were of natural or synthetic origin or combinations thereof, and consisted of single-layered or multilayered implants or blends thereof. Implants were constructed from a wide range of materials or combinations thereof, such as collagen, chitosan, hyaluronic acid, alginate, fibrin, hydroxyapatite, poly(lactic-co-glycolic acid), polycaprolactone, poly(glycolic acid) and poly(ethylene glycol), and used in different states: scaffolds, hydrogels, or hybrid mixtures of both. Various biological cues were incorporated in the biomaterials prior to implantation or administered afterwards by injection into the knee joint, mostly growth factors of the TGF- β superfamily such as bone morphogenetic protein 2 (BMP-2) and TGF- β 1, but also fibroblast growth factor (FGF) and platelet-rich plasma (PRP). The maximum follow-up time was 1 year, but studies mainly investigated relatively short-term effects of implanted biomaterials on cartilage regeneration (up to 6 months). Risk of bias assessment A risk of bias assessment was performed to assess risks of bias (selection, performance bias, detection and attrition bias) in studies included for the meta-analysis ( Fig. 3 ). An overview of all scores per individual study is provided in Supplemental Information 6. 10. 7717/peerj. 2243/fig-3 Figure 3 Risk of bias of all included studies in the meta-analysis. The green, orange and red colors depict the percentages of studies with low, unknown or high risk of bias of the total number of assessed studies. The risk of bias assessment indicated a general lack of details regarding the experimental setup, as indicated by the orange bars. The green bars represent a low risk of bias, mainly for the difference between groups at the moment of surgical intervention and addressing incomplete outcome data. High risk of bias was infrequently scored, as indicated by the red bars. Q4–Q6 are not depicted in the graph, but are described in Supplemental Information 6. The risk of bias assessment showed that details with respect to the randomization method were not provided (Q1). It was often described that animals were randomized across different groups without describing the method of randomization, thereby limiting assessment of the adequacy of randomization and therefore the actual risk of selection bias. Another notable observation from the experimental designs studied was that only in a limited number of studies it was described that power calculations were performed, whereas sufficient power in animal experiments is a requirement for performing adequate studies. The actual power analyses were never provided in the studies. Due to a lack of information, it was also difficult to assess possible bias by differences in implantation sites (with differences in load-bearing conditions, Q2. 1) and differences between groups related to the age, sex and weight of the animals at the start of the experiment (Q2. 2). Generally, baseline characteristics of animals prior to implantation of biomaterials (e. g. , some animals received additional surgery related to harvesting of cells for biomaterials combined with cells, Q2. 3) were similar. When implanting biomaterials, no details were described on blinding different biomaterials (Q3). Blinding of the empty defect and biomaterial conditions should be performed to limit bias. However, blinding between the empty defect and biomaterial group is impossible in case only one biomaterial is implanted. More than half of the studies conducted blinded outcome assessment while performing the histological scoring, resulting in low risk of detection bias, whereas the other studies had an unknown risk (Q7). For most studies, no incomplete outcome data were described/found, resulting in low risk of attrition bias. For some studies, dropouts were described/found, resulting in differences between groups and high risk of bias (Q8). Overall, the risk of bias analysis generally revealed poor reporting of the experimental design for the majority of the studies, impeding an assessment of the actual risk of bias. Data synthesis For an overview of the meta-analysis and results obtained, see Table 1. The histological scores of defects implanted with biomaterials and non-treated empty defects are presented as a percentage on a 100% scale, where 0% and 100% indicate poor and perfect cartilage regeneration, respectively. Data are presented as the effect (%) with 95% CI. 10. 7717/peerj. 2243/table-1 Table 1 Overview of the meta-analysis results for the main research question assessing the overall beneficial effect of implanting acellular biomaterials in osteochondral defects compared to non-treated empty defects and sub-questions evaluating the effect of specific variables on the treatment effect. The total number of studies and number of experimental groups included in the meta-analysis are shown (some studies have >1 experimental group, no. of studies/groups). The quality of cartilage regeneration is presented on a 100% scale, where 100% represents the maximum achievable histological score and thus the best cartilage regeneration. Implantation of biomaterials significantly improved cartilage regeneration compared to non-treated empty defects, which was further improved by the incorporation of biologics. No significant differences were found between natural and synthetic materials, between the various material subgroups, and between the biomaterial structures (hydrogels versus scaffolds versus blends), and between animal species. Meta-analysis No. of studies/groups Subgroups Cartilage regeneration (% (95% CI)) Mean difference (% (95% CI)) p -value 1. Overall effect 127/400 Biomaterial 53. 6 [50. 7, 56. 6] 15. 6 [12. 6, 18. 6] 127/247 Empty defect 38. 1 [35. 1, 41. 0] p < 0. 0001 2. Origin materials 76/222 Natural 53. 0 [49. 3, 56. 6] −0. 73 [−6. 5, 5. 0] 39/137 Synthetic 53. 7 [48. 8, 58. 7] p = 0. 887 3. Material subgroups 20/68 Collagen 49. 5 [41. 1, 57. 8] p = 0. 804 6/17 Chitosan 57. 5 [40. 8, 74. 2] 5/11 Hyaluronic acid 47. 9 [31. 7, 64. 1] 5/16 Alginate 63. 0 [46. 9, 79. 00] 3/10 Fibrin 55. 3 [34. 4, 76. 3] 5/11 Bone 51. 2 [35. 2, 67. 2] 15/52 PLGA 58. 5 [49. 0, 68. 0] 6/21 PAMPS-PDMAAm DN 47. 9 [31. 7, 64. 1] 4. Scaffold structure 78/258 Scaffolds 53. 1 [49. 5, 56. 7] p = 0. 973 41/127 Hydrogels 54. 2 [49. 4, 59. 1] 7/17 Blends 55. 7 [42. 0, 69. 3] 5. Biologicals 113/291 No biologicals 51. 7 [48. 6, 54. 9] 7. 56 [2. 1, 13. 0] 35/109 Biologicals 59. 3 [54. 0, 64. 6] p = 0. 007 6. Biological cues 9/35 BMP 56. 6 [−6. 3, 119. 6] p = 0. 780 5/20 FGF 51. 8 [−43. 9, 147. 4] 8/14 PRP 55. 9 [−20. 9, 132. 8] 6/16 TGF 60. 2 [−7. 5, 128. 0] 7. Animal models 3/5 Dogs ED: 31. 9 [14. 5, 49. 4]; B: 50. 6 [33. 0, 68. 2] 18. 7 [−0. 0, 37. 3] 5/13 Goats ED: 58. 5 [43. 4, 73. 7]; B: 61. 6 [47. 6, 75. 6] 3. 1 [−13. 2, 19. 4] 1/3 Macaques ED: 12. 2 [−18. 2, 42. 6]; B: 6. 8 [−23. 3, 37. 0] −5. 4 [−37. 6, 26. 8] 10/20 Minipigs ED: 42. 4 [32. 4, 52. 4]; B: 56. 1 [46. 3, 66. 0] 13. 6 [3. 1, 24. 1] 94/333 Rabbits ED: 37. 7 [34. 2, 41. 1]; B: 52. 5 [49. 0, 55. 9] 14. 8 [11. 1, 18. 5] 13/23 Sheep ED: 35. 3 [26. 4, 44. 3]; B: 61. 3 [52. 7, 70. 0] 26. 0 [16. 3, 35. 7] p = 0. 348 Notes. ED Empty defect B Biomaterials Overall effect biomaterial implantation The meta-analysis indicates a significant improvement of cartilage regeneration using acellular biomaterials implanted after applying marrow stimulating techniques compared to non-treated empty defects (15. 6% (95% CI [12. 6, 18. 6], p < 0. 0001). The forest plot ( Supplemental Information 7 ) depicts the outcome effect of each individual study. In 73 studies cartilage regeneration significantly improved by the incorporation of biomaterials. In 48 studies no effect was found, whereas in only six studies a negative effect on cartilage regeneration was observed. A similar significant effect was observed taking into account the maximum follow-up only (16. 3% [13. 1, 19. 6], p < 0. 0001). Also for short and long term follow-up cartilage regeneration was significantly improved (≤8 weeks: 12. 5% [9. 3, 15. 7], >8 weeks: 17. 1% [13. 9, 20. 2]). No notable differences in cartilage regeneration were found between the results based on the maximum follow-up time per study versus those based on all time points per study. Therefore, further subgroup analyses were made using results from all time points together. Natural and synthetic materials The subgroup analysis assessing cartilage regeneration using materials of different origin, natural and synthetic, indicated no significant differences ( p = 0. 887) between natural (53. 0% [49. 31, 56. 63]) and synthetic materials (53. 7% [48. 75, 58. 65]). Dividing the group of materials into subgroups allows comparison of cartilage regeneration using different biomaterials. The following subgroups were studied: (1) collagen, (2) chitosan, (3) hyaluronic acid–based biomaterials), (4) alginate, (5) fibrin), (6) bone material-based, (7) PLGA, and (8) PAMPS-PDMAAm DN hydrogel. No significant differences between the biomaterial subgroups were found ( Table 1 ). Material structure Materials were divided in three groups based on their structure: (1) scaffolds, (2) hydrogels, and (3) blends. Cartilage regeneration was similar after use of scaffolds (53. 1% [49. 53, 56. 74]), hydrogels (54. 2% [49. 39, 59. 07]) and blends (55. 7% [42. 0, 69. 3), p = 0. 973. Biologics Incorporation of biologics in the biomaterials resulted in a statistically significant improvement in cartilage regeneration of 7. 6% [2. 1, 13. 0], p = 0. 007, compared to the implantation of control biomaterials. Including only those studies with a direct comparison between control biomaterials and biomaterials loaded with biologics resulted in an improved cartilage regeneration of 14. 6% [5. 9, 23. 4], p = 0. 003. Comparing various biological cues including BMP, FGF, PRP and TGF indicated no significant differences in improvement of cartilage regeneration between these biologics. Animal models Evaluation of the animal models used showed no significant differences ( p = 0. 348) between the effects of biomaterials implanted in dogs, goats, macaques, minipigs, pigs, rabbits, rats or sheep ( Table 1 ). Sensitivity analyses Sensitivity analyses were performed to assess the robustness of the meta-analysis with respect to the overall effect. The sensitivity analyses indicated that exclusion of studies with assumptions and studies written in Chinese (no risk of bias assessment analyzed) had no effect on the estimated difference in biomaterial regeneration. Moreover, including only studies with SDs or SEs in the 10–90% range did not notably change of the overall outcome effect. In a post-hoc analysis, we investigated cartilage regeneration using biomaterials implanted at different locations including condyles, femur, intercondylar fossa and the trochlea. No differences were found comparing these implant sites ( p = 0. 143). In another post-hoc analysis, we compared cartilage regeneration of empty defects or defects filled with biomaterials after applying microfracturing or subchondral drilling. For empty defects ( p = 0. 152) and biomaterial implants ( p = 0. 063) no significant differences between the two bone marrow stimulating techniques were found. Publication bias A funnel plot ( Fig. 4 ) was prepared for all included studies to analyze the overall comparison between acellular biomaterials and non-treated empty defect controls. No extensive asymmetry was observed, indicating an absence of considerable publication bias. 10. 7717/peerj. 2243/fig-4 Figure 4 Funnel plot of included studies to assess the overall effect of the implantation of acellular biomaterials compared to non-treated empty defect controls. The figure indicates no substantial asymmetry. Discussion The regeneration of damaged cartilage has been widely investigated using preclinical models. However, the efficacy of cartilage regeneration using implantation of acellular biomaterials has never been assessed using a systematic review and meta-analysis. This systematic review aimed (a) to provide an overview of currently existing knowledge and identify knowledge gaps, (b) to provide transparency on the quality of performed in vivo studies, and (c) to aid the design of future animal studies and clinical trials. The results could provide insight in strategies for future (pre) clinical research related to biomaterial properties, incorporation of biologics, choice of a suitable animal model, and their effects on cartilage regeneration. The general findings of this systematic review and meta-analysis are that the implantation of biomaterials improves cartilage regeneration compared to non-treated osteochondral defects by 16% (95% CI). There were only six out of 151 studies that showed a negative effect of biomaterial implantation on cartilage regeneration. In 48 studies no significant effect on cartilage regeneration was found. For those studies with improved cartilage regeneration (73 studies), clinical studies will have to confirm the beneficial effect of implantation of biomaterials on cartilage regeneration in human patients. Filardo et al. described the implantation of an osteochondral biomimetic scaffold consisting of a type I collagen cartilage-like layer, a type I collagen/hydroxyapatite intermediate layer, and a mineralized blend of type I collagen and hydroxyapatite as a subchondral bone compartment, to treat patients with osteochondritis dissecans. For these patients, clinical scores improved significantly after the first two years and evaluation by MRI indicated good defect filling and implant integration, but also heterogeneous tissue regeneration and changes of the subchondral bone ( Filardo et al. , 2013 ). In two studies included in this systematic review and meta-analysis, this osteochondral biomimetic scaffold was also implanted in sheep. Cartilage regeneration after six months was 81. 8% ± 8. 9% (empty defect: 23. 2% ± 20. 7%) and 81. 2% ± 5. 1% (empty defect: 23. 4% ± 6. 7%). A direct comparison between the degree of cartilage regeneration described in the preclinical studies and clinical study is not possible since no histological results were described in the clinical study. In addition, outcome measures used in preclinical studies may not predict the clinical outcome. For example, a randomized controlled clinical trial with BST-CarGel, a chitosan-based medical device, showed greater lesion filling and superior repair tissue quality compared to bone marrow stimulation after twelve months implantation, but without notable clinical differences related to pain, stiffness and physical function between both groups ( Stanish et al. , 2013 ). A remarkable observation is the difference in follow-up between the studies, which may explain the good histological scores in the preclinical studies after six months and heterogeneous tissue regeneration and changes of the subchondral bone after two years in human patients. In general, clinical studies demonstrated improved cartilage regeneration by the implantation of biomaterials after bone marrow stimulation, but there is still room for improvement regarding clinical outcome and tissue quality. The only subgroup analysis that showed a statistically significant result between the groups was between control biomaterials and biomaterials loaded with biologics. In future clinical studies assessment of the beneficial properties of implanting biomaterials loaded with biologics is of interest, since a significant improvement of 8% (95% CI) compared to control biomaterials was found and even 14. 6% when using studies that directly compared biomaterials with and without biologics. We were not able to perform analyses for the effect of the concentration or subtype of the growth factors due to the small size of these subgroups, although these factors may have a large effect on the outcome. In the study by Ishii et al. (2007) a positive effect of FGF-2 was observed by the addition of at least 183 ng to the biomaterials, while Maehara et al. (2010) showed significant improvements of impregnating biomaterials in 10 µg/ml and not for 100 µg/ml FGF-2. Loading biomaterials with different BMPs including BMP-2 ( Aulin et al. , 2013 ; Reyes et al. , 2012 ; Reyes et al. , 2014 ; Reyes et al. , 2013 ; Tamai et al. , 2005 ) and BMP-7 ( Mori et al. , 2013 ), or TGF subtypes including TGF- β ( Mierisch et al. , 2002 ) and TGF- β 1 ( Reyes et al. , 2012 ; Reyes et al. , 2014 ), resulted in significantly improved cartilage regeneration. However, for clinical application of these medical devices, one should take safety of the products into account as side effects of TGF- β in a joint environment, including fibrosis and osteophyte formation, have been described ( Blaney Davidson, Van der Kraan & Van den Berg, 2007 ) and patients suffered from major complications after spinal surgery and implantation of high concentrations of BMP/INFUSE ( Epstein, 2013 ). The study characteristics of all included studies were tabulated to provide an extensive overview of the available literature. Besides the internal validity of the studies, the generalizability (external validity) of the study results is of great importance. The latter is affected by factors related to the animal model (species, strain, weight, age, and sex), surgery (location and size of the defect) and follow-up, resulting in heterogeneity between studies. This was also indicated by the relatively high level of heterogeneity ( I 2 ) for the main meta-analysis (99. 4% [99. 4, 99. 4]), and the heterogeneity was almost similar for subgroup analyses. We chose to include only healthy animals receiving biomaterials. The screened studies also contained osteoarthritis models that were not included, which may be relevant for future applications to treat patients with osteoarthritis. Therefore, results from this systematic review and meta-analysis may be different compared to results found for osteoarthritis models and future clinical studies with osteoarthritis patients. We assumed that in order to assess the effect of implanted biomaterials on cartilage regeneration, reduction of the influence of confounding parameters would aid the validity of the results and conclusions. In this study, the meta-analysis included all available data of the effect of implanting biomaterials after applying bone marrow stimulating techniques (microfracture and subchondral drilling) compared to empty defects on cartilage regeneration. During microfracture surgery the subchondral bone is penetrated using an arthroscopic awl, whereas during subchondral drilling the trabecular bone is penetrated using a high speed drill, which may result in thermal necrosis ( Falah et al. , 2010 ). Remarkably, more studies applied subchondral drilling (258 studies) compared to microfracture surgery (25 studies), while microfracture surgery was developed to overcome problems associated with thermal necrosis from subchondral drilling in the treatment of human patients ( Kane et al. , 2013 ). We did perform a post-hoc meta-analysis to investigate differences in cartilage regeneration after applying both marrow stimulating techniques and subsequent implantation of biomaterials, which resulted in no significant differences between microfracturing and subchondral drilling. A reason for the larger number of animal studies performing subchondral drilling compared to microfracture surgery may be the ease to perform subchondral drilling over microfracture surgery in animals. Although in the included studies various implant locations (i. e. , trochlea and condyles) were used, we grouped the results in the meta-analysis. A post-hoc subgroup analysis was performed to compare defect locations, but no overall significant differences were found for biomaterials implanted at different implant locations. Our analysis did not confirm a finding of Chen et al. (2013) showing improved chondrogenesis in trochlear versus condylar cartilage defects after bone marrow stimulation in rabbits. This may be explained by various parameters affecting the degree of cartilage regeneration at different implant locations, such as the animal model, follow-up period and rehabilitation protocol. Different outcome measures such as macroscopic and histological evaluation, semi-quantitative macroscopical and histological evaluation using scoring systems, histomorphometry, PCR and biochemical assays were used to assess the regenerative potential of implanting biomaterials. In this systematic review and meta-analysis, only data from semi-quantitative histological scoring systems were used as outcome measure. We chose to use these data as most authors presented their results by this method and it allows quantitative comparison of different studies in a meta-analysis. Various histological scoring systems have been used by the authors of included studies, such as the O’Driscoll, Pineda, Wakitani and ICRS scoring system, which were also reviewed by Rutgers et al. (2010). Depending on the histological scoring system, parameters such as cell morphology, matrix staining, surface regularity, structural integrity, defect filling and the restoration of the subchondral bone were evaluated. A limitation of this outcome measure is that the specific topics addressed in the scoring systems greatly differ, i. e. , some studies focus on the regeneration of cartilage only, cartilage as well as subchondral bone, or include a biomaterial component (e. g. , scoring degradation of the implant). Other outcome measures including macroscopic evaluation, biochemical analysis and biomechanical aspects of the tissue may complete the overview of the tissue quality and provide valuable insights in articular cartilage regeneration, but these outcome measures were only used in a limited number of studies, and therefore not assessed in this analysis. The risk of bias assessment provided insights in the quality of the experimental design of the studies. Most studies scored a low or unknown risk of bias, however, also little high risk of bias was scored. Low methodological quality (internal validity) may result in an overestimation or underestimation of the intervention effect ( Higgins et al. , 2011 ). In general, details regarding the randomization procedure were not described. Moreover, an observation during the risk of bias assessment was that only few studies included in the systematic review described that power calculations were performed, which is a crucial aspect in conducting experimental studies to ensure sufficient power of experimental designs. As a consequence, studies may lack sufficient power and thereby run the risk of false negative results. Due to the poor reporting of the experimental design for the majority of the studies the assessment of the adequacy of randomization and power calculations, and thus the assessment of the actual risk of selection bias, was inadequate. However, it may also hold true that studies were well designed but there was only poor reporting of the experimental designs ( Hooijmans et al. , 2012 ). Most researchers scoring the histology sections were blinded and sections were randomized. However, when biomaterials are not (completely) degraded, blinding between biomaterials and empty defects is practically impossible. A lack of blinding of outcome assessors implies the risk of detection/observer bias ( Bello et al. , 2014 ). Bias may have been introduced by the lack of blinding and randomization and detracts from the overall validity of the results ( Bebarta, Luyten & Heard, 2003 ; Hirst et al. , 2014 ). There is a risk that the positive results found are an overestimation of the true effect of using biomaterials. Introducing standardized protocols such as the golden standard publication checklist ( Hooijmans et al. , 2011 ) or the ARRIVE guidelines ( Kilkenny et al. , 2012 ) may improve reporting of animal studies. Funnel plots represent the precision of the measured effects, which increases by an increase in study size. Therefore, for small and large studies scatter will be relatively large and little, respectively. As a consequence, generally, in the absence of bias the plot resembles a symmetrical pyramid (a funnel) ( Higgins & Green, 2011 ). An important limitation may be publication bias, since multiple studies were included from the same author and negative results may not be published. It was described in a study by ter Riet et al. that researchers themselves estimate that only 50% of the conducted animal experiments are published. This problem may be solved by statistical corrections for publication bias ( ter Riet et al. , 2012 ). In our study, the funnel plot did not show asymmetry and therefore did not indicate the presence of publication bias. The translational value of animal studies depends on the comparability to the clinical situation. One of the limitations of the performed animal experiments is the short follow-up times. The maximum follow-up time was one year, but most studies investigated cartilage regeneration up to six months. This limits the translational value since clinical improvements in humans are generally observed up to 1. 5–3 years after microfracture surgery ( Hoemann et al. , 2010 ; Van der Linden et al. , 2013 ). Moreover, many variations were present in the applied animal models, i. e. , animal characteristics (species, strain, sex, age, weight), surgical defects (size, depth and location), applied biomaterials, and incorporated biologics. A review by Chu, Szczodry & Bruno (2010) extensively reflects on benefits and limitations of different animal models used in cartilage repair studies. They state that for humans the volume of a cartilage defect is approximately 550 mm 3 and treatment is required for defects with a surface larger than 10 mm 2. Due to the limited joint size of many animals, larger animal models such as minipig, goat and horse therefore offer superior translational value than smaller animals such as rats, rabbits and dogs. However, all studies contained defect volumes smaller than 550 mm 3 and only few studies had defects surfaces larger than 10 mm 2. Additionally, cartilage thickness differs among various species, with goat, rabbit, minipig and dogs having thinner cartilage than humans. Another drawback for some animal models is the large endogenous repair potential. In humans, untreated defects show little to no regeneration while rabbits display a large regenerative potential, limiting clinical translation. Dog, goat, minipig and horse do not have this large endogenous repair and the use of these animals may therefore be favorable. The maturity of the animals is of great importance when designing animal experiments since open growth plates can impede with the applied treatment. Animal species are skeletally mature at different ages; i. e. , rabbits at the age of 16–39 weeks, pigs at 42–52 weeks, dogs at 12–24 months, sheep and goat at 24–36 months and horses at 60–72 months ( Ahern et al. , 2009 ; Chu, Szczodry & Bruno, 2010 ). In this study we did not group studies based on animal maturity. In addition to clinical relevance, other reasons to select an animal model are related to logistical, financial, and ethical considerations. A systematic review conducted by Ahern et al. (2009) investigated the strengths and shortcomings of different animal models and compared these with common clinical lesions in clinical studies. They remarked that smaller animal models are often used due to feasibility, while large animal models may more closely resemble humans. However, no differences were found between animal models in this systematic review and meta-analysis, which may be explained by various parameters affecting the degree of cartilage regeneration such as implant location, defect size, follow-up period and rehabilitation protocol. In this systematic review and meta-analysis the efficacy of cartilage regeneration using acellular biomaterials was compared to the natural healing response of defects treated with microfracture surgery and subchondral drilling. The risk of bias assessment indicated poor reporting in animal studies, which may be improved in future animal studies. Moreover, to improve the translation towards clinical trials animal experiments should be comparable to the clinical situation. As described in this systematic review a relatively high level of heterogeneity exists between studies related to the animal model, surgery and follow-up, with a need to resemble current clinical settings more closely. In this study we only addressed bone marrow stimulating techniques (microfracture and subchondral drilling) and subsequently the incorporation of biomaterials, but also the regeneration of partial thickness cartilage defects may be beneficial to prevent progression to full-thickness cartilage defects, limit the progression towards osteoarthritis and improve quality of life in patients. In many studies also cell-laden biomaterials have been implanted and the beneficial effect of cellular biomaterials versus acellular biomaterials and the natural healing response has been studied. Although acellular biomaterials offer various advantageous properties over cellular biomaterials such as no donor-site, no cell culture, off the shelf availability, less regulatory issues, and application of one-stage surgical procedures ( Brouwer et al. , 2011 ; Efe et al. , 2012 ), studying the additive value of cellular biomaterials may aid further improvement of marrow stimulating techniques. Conclusion The systematic review and meta-analysis resulted in a structured, thorough and transparent overview of literature related to the current evidence for the efficacy of cartilage regeneration using acellular biomaterials implanted after microfracturing in animal models. Cartilage regeneration is more effective by implantation of acellular biomaterials in microfracture defects compared to microfracturing alone. The efficacy is further improved by the incorporation of biologics. Supplemental Information 10. 7717/peerj. 2243/supp-1 Supplemental Information 1 Search strategies Click here for additional data file. 10. 7717/peerj. 2243/supp-2 Supplemental Information 2 Methodological quality assessment tool Click here for additional data file. 10. 7717/peerj. 2243/supp-3 Supplemental Information 3 Study characteristics Click here for additional data file. 10. 7717/peerj. 2243/supp-4 Supplemental Information 4 Study characteristics; abbreviations and references Click here for additional data file. 10. 7717/peerj. 2243/supp-5 Supplemental Information 5 Raw data Click here for additional data file. 10. 7717/peerj. 2243/supp-6 Supplemental Information 6 Risk of bias results per individual study Risk of bias results per individual study Click here for additional data file. 10. 7717/peerj. 2243/supp-7 Supplemental Information 7 Forest plot Click here for additional data file. 10. 7717/peerj. 2243/supp-8 Supplemental Information 8 Prisma checklist Click here for additional data file.

10. 7717/peerj. 2497

PeerJ

Enzymatically crosslinked gelatin hydrogel promotes the proliferation of adipose tissue-derived stromal cells

Gelatin hydrogel crosslinked by microbial transglutaminase (mTG) exhibits excellent performance in cell adhesion, proliferation, and differentiation. We examined the gelation time and gel strength of gelatin/mTG hydrogels in various proportions to investigate their physical properties and tested their degradation performances in vitro. Cell morphology and viability of adipose tissue-derived stromal cells (ADSCs) cultured on the 2D gel surface or in 3D hydrogel encapsulation were evaluated by immunofluorescence staining. Cell proliferation was tested via Alamar Blue assay. To investigate the hydrogel effect on cell differentiation, the cardiac-specific gene expression levelsof Nkx2. 5, Myh6, Gja1, and Mef2c in encapsulated ADSCs with or without cardiac induction medium were detected by real-time RT-PCR. Cell release from the encapsulated status and cell migration in a 3D hydrogel model were assessed in vitro. Results show that the gelatin/mTG hydrogels are not cytotoxic and that their mechanical properties are adjustable. Hydrogel degradation is related to gel concentration and the resident cells. Cell growth morphology and proliferative capability in both 2D and 3D cultures were mainly affected by gel concentration. PCR result shows that hydrogel modulus together with induction medium affects the cardiac differentiation of ADSCs. The cell migration experiment and subcutaneous implantation show that the hydrogels are suitable for cell delivery.

Introduction In the field of tissue engineering and regenerative medicine, researchers have been searching for ideal biomaterials that mimic the structure and composition of the extracellular matrix (ECM) and can be used for cell 3D culture and cell transplantation in vivo. Hydrogel materials are one of the most important areas of research in biological materials because of their high moisture content and high plasticity ( Johnson & Christman, 2013 ; Radhakrishnan, Krishnan & Sethuraman, 2014 ; Toh & Loh, 2014 ). Numerous kinds of hydrogel materials have emerged, including natural materials, such as collagen, gelatin, hyaluronic acid, laminin, chitosan, and sodium alginate, and synthetic materials, such as polylactide, polylactide-co-glycolic acid copolymer, polyethylene glycol, polycaprolactone, and polyacrylamide ( El-Sherbiny & Yacoub, 2013 ; Gibbs et al. , 2016 ). Gelatin, a well-known biological material because of its good biocompatibility, is the degraded product of collagen, which is the main component of native ECM. Compared with that of collagen, gelatin manufacturing process is relatively simple, and its market price is cheaper than the former. However, the natural gelatin hydrogels are highly hydrated and possess poor mechanical stability and durability, which critically limit their wide-spread application; thus, the manufacture of gelatin-based biomaterials frequently requires a certain degree of crosslinking for stabilizing gelatin macromolecules. Several approaches for crosslinking gelatin, such as physical crosslinking, chemical crosslinking, and enzymatic crosslinking, were used. In conventional physical crosslinking of gelatin, an aqueous solution of several percent gelatin turns to a transparent elastic hydrogel upon cooling below ∼35 °C, and crosslinking occurs via random coil gelatin molecules turning to the ordered triple helix conformation of collagen. However, the thermoreversibility of the hydrogel makes it melt at physiological temperature. Other physical crosslinking methods, such as plasma treatment, often result in low crosslinking extent of gelatin macromolecules because crosslinking occurs only at the surface of the material ( Ratanavaraporn et al. , 2010 ). Numerous researchers have aimed to build more stable hydrogels by using UV light or chemical crosslinkers. Photocrosslinked hydrogels usually present short gelation time and are chemically stable and mechanically strong, but both photo-initiators and UV light required for the photopolymerization reaction may lead to cell death ( Jin et al. , 2010 ). Chemical crosslinking of gelatin refers to the use of chemical reagents, such as glutaraldehyde, formaldehyde, 1-(3-dimethylaminopropyl)-3-ethyl-carbodimide hydrochloride, and genipin, as a chemical crosslinker ( Tseng et al. , 2013 ). Despite the improved mechanical strength and proteolytic stability of crosslinked gelatin hydrogels, chemical crosslinkers often elicit either cytotoxic side-effects or immunological responses from the host ( Yung et al. , 2007 ; Li & Liu, 2009 ). Enzymatic crosslinking is a new approach to biomaterial crosslinking. The enzymes, including tyrosinase and transglutaminase, are currently known as crosslinkers for numerous kinds of proteins, including gelatin and collagen ( Yung et al. , 2007 ; Spurlin et al. , 2009 ; Kuwahara et al. , 2010 ; Taddei et al. , 2013 ). Transglutaminase is highlighted as the best studied enzyme system involved in protein-based hydrogel crosslinking for tissue engineering approaches, because it can offer intimate integration between the in situ formed hydrogel and the native host tissue ( Teixeira et al. , 2012 ). Moreover, the hydrogel catalyzed by transglutaminase is mechanically stronger and more stable than that catalyzed by tyrosinase ( Chen et al. , 2003 ). However, owing to the relatively high price of these enzymes compared with chemical crosslinkers, the enzymatic crosslinking method for gelatin had been rarely used until microbial transglutaminase (mTG) was discovered. mTG, which is derived from streptomycetes, exhibits high specific activity over a wide range of temperature and pH and is Ca 2+ independent. mTG has been extensively utilized in the food industry, enhancing the functional properties of protein-rich food through covalent crosslinking ( Halloran et al. , 2008 ; Wangtueai, Noomhorm & Regenstein, 2010 ). At present, few studies have reported on gelatin hydrogel crosslinked by mTG as a cell scaffold material ( Paguirigan & Beebe, 2007 ; Yung et al. , 2007 ; Kuwahara et al. , 2010 ; Bode et al. , 2011 ; De Colli et al. , 2012 ; Bode et al. , 2013 ; Da Silva et al. , 2014 ). Numerous issues remain worth studying. For example, we know that transglutaminase is non-toxic and exerts no side-effects on several cell types, but we do not know its effects on other cell types, such as adipose tissue-derived stromal cells (ADSCs). ADSCs are a kind of adult stem cells with rich cell sources and can be obtained by minimally invasive surgery, such as subcutaneous liposuction. ADSCs present multiple differentiation potential and can differentiate into osteoblasts, chondrocytes, adipocytes, and cardiomyocytes ( Wilson, Butler & Seifalian, 2011 ; Wankhade et al. , 2016 ). Therefore, ADSCs present a considerable potential source of stem cells for tissue engineering research and clinical applications ( Pikula et al. , 2013 ; Suzuki et al. , 2015 ; Naderi et al. , 2016 ; Pak et al. , 2016 ). How will the degradation of gelatin/mTG hydrogels be affected by cell secretion after ADSCs are inoculated on the hydrogels? On the other hand, how does the degradation of materials affect cell growth? Extremely little knowledge on these topics is available. In this study, we will analyze the degradation of gelatin/mTG hydrogels in vitro and in vivo with or without cell inoculation and evaluate cell growth in 2D or 3D culture to determine whether the material is suitable as a cell scaffold. At present, whether gelatin/mTG hydrogel can be used as a cell carrier for in vivo transplantation after inoculated with ADSCs and whether the release of cells from the hydrogel is controllable remain unclear. If the cells are released too quickly, rapid cell loss from the implantation site will occur, thereby undermining the purpose of tissue repair and regeneration. In addition, whether the material is conducive to cell migration is unclear. Cell migration often facilitates the organization of the capillary network surrounding the implanted hydrogel to establish blood supply. In this study, we will design an in vitro 3D model to simulate cell migration inside the hydrogel with the aim of providing evidence for in vivo animal experiments in the future. Materials and Methods Preparation of gelatin hydrogels Gelatin gel formation was initiated by mTG addition. For hydrogel preparation, gelatin powder (type A, 300 Bloom; Sigma–Aldrich, MO, USA) was weighed and dissolved in phosphate-buffered saline (PBS) at 50 °C and then sterilized as rapidly as possible through 0. 22 µm filters to prevent filter blockage by the cooling gel. The mTG (Bomei, China, enzyme activity units > 100 U per gram) solution was prepared by dissolving mTG in PBS to obtain 10% (wt, weight ratio) solution and then sterilizing through 0. 22 µm filters. Gelatin/mTG hydrogels were prepared by mixing a certain amount of 10% mTG solution with different concentration of gelatin solutions according to the experimental need. To determine the effects of different gelatin concentration (1%, 2%, 4%, 6%, 8%, and 10% (w/v, weight/ volume)) on the gelation time and gel strength of resultant hydrogels, and the mTG dosage was retained at 10 U/g pro (enzyme activity units per gram of protein). Here, the protein is gelatin. To determine the effects of mTG dosage (2, 5, 10, 20, and 40 U/g pro) on gelation time and gel strength of the resultant hydrogels, and the concentration of gelatin solution was maintained at 4% (w/v). Gelation time and gel strength test For gelation time test, 2 ml of gelatin/mTG solution was added in a transparent glass vial and incubated at 37 °C, and the onset of gelling detected through the vial inverting method was recorded as gelation time. Six repeated measurements for each type of gelatin/mTG hydrogels were performed. For the gel strength test, 6 ml of gelatin/mTG mixing solution were added into a 35 mm culture dish and incubated at 37 °C for 2 h; then, the dish was fixed on the platform of a mechanical testing apparatus (HPB, Handpi, China). The detecting probe was a flat-head stainless-steel cylinder (12. 5 mm in diameter, 10 mm in height) attached to a digital pull-and-push dynamometer (HP-20 Handpi, China). The probe was loaded down onto the hydrogel at a constant rate of 1. 0 mm/s at room temperature (∼24 °C). The advance was stopped when the probe front reached to a 5 mm depth from the gel surface. Meanwhile, the value of loading force was recorded automatically by a mechanical measurement software (Yueqing Handpi Instruments Co. , Ltd, Zhejiang, China). The peak value of the recorded curve was acquired and converted to gel strength. Six repeated measurements for each type of gelatin/mTG hydrogels were performed. Hydrogel degradation test The degradation of hydrogel materials in two groups was tested. The first group consisted of cell-free hydrogels, and the second group consisted of cell-containing hydrogels. To prepare cell-free hydrogels, twelve 35 mm culture dishes were pre-weighed with an electronic balance and respectively marked. Subsequently, 2%–8% concentrations of gelatin/mTG solution were loaded into the dishes with 2 ml for each dish, and the dishes were incubated at 37 °C for 2 h. After gelling, the hydrogel-containing dishes were weighed again; 2 ml of PBS was added into each dish; and the dishes were cultured at 37 °C with 5% CO 2 for eight weeks. PBS was changed every 3 days. At different time points, three dishes were removed for testing; PBS was removed completely; and the hydrogel-containing dishes were weighed. The remaining hydrogel mass was obtained by calculating the weight variation of each dish. Finally, the degree of degradation was expressed as a percentage of the remaining hydrogel mass versus the initial hydrogel mass. To prepare cell-containing hydrogels, 2 ml of gelatin/mTG solution mixed with 5. 0 × 10 6 cells was added into a 35 mm culture dish. The rest of the steps were the same as the previous methodology, except that the culture medium was changed to cell expansion medium (see ‘Primary culture of ADSCs’). Cell culture medium was replaced every 3–4 days. Primary culture of ADSCs Animal study was approved by the Institutional Animal Care and Use Committee (IACUC) of Sichuan University, all experiments were performed in accordance with the guidelines of IACUC of Sichuan University. Subcutaneous adipose tissue was obtained from a 150 g Sprague–Dawley rat. Harvested tissue was enzymatically dissociated using 1 mg/ml collagenase type I (Sigma, MO, USA) in high-glucose Dulbecco’s modified Eagle’s medium (DMEM; Hyclone, UT, USA). Digestion was carried out under continuous agitation for 45 min at 37 °C and followed by centrifugation at 283 g for 7 min. The pelleted cells were then harvested and plated on 25 mm 2 cell flasks. The initial plates were denoted by Passage 0 (P0). At 24 h intervals, cultures were washed with PBS to remove contaminating erythrocytes and other non-attached cells. This procedure was repeated every 24 h for three days. The plating medium consisted of high-glucose DMEM, 15% fetal bovine serum (FBS; Invitrogen, CA, USA), and 1% penicillin/streptomycin (P/S; Hyclone, UT, USA). Cells were maintained at 37 °C with 5% CO 2 and fed two times per week. When cultures reached confluence within 5–7 days after the initial plating, the adherent cells were detached with 0. 25% trypsin/EDTA and then either replated at 5. 0 × 10 4 cells/cm 2 or immediately used in experiments. Cultures were passaged every 3–5 days, and the culture medium was replaced by expansion medium, which consists of high-glucose DMEM, 10% FBS, and 1% P/S. Cell viability in 2D culture Gelatin/mTG solutions with various gelatin concentrations (2%, 4%, 6%, 8% and 10%) were prepared as described above: the solutions were added into six-well tissue culture plates (TCPs), with 2 ml placed in each well. The solutions were incubated at 37 °C for gelling for 2 h. After washing thrice with sterile PBS, the hydrogels were ready for cell 2D seeding. ADSCs were digested and centrifuged at 283 g, and the supernatant was discarded. ADSCs were then seeded on the hydrogel surface at 1. 0 × 10 5 cells per well. Cells were cultured in the expansion medium and observed daily under an inverted optical microscope (CKX41, Olympus, Tokyo, Japan). After 14 days of cultivation, live/dead staining assay was employed to assess the viability of cell populations on 2D hydrogel substrates. Samples were washed thrice in sterile PBS and incubated at 37 °C for 30 min in a solution containing 2 µM calcein—AM (Sigma, St. Louis, MO, USA) and 2 µM propidium iodide (PI; Sigma, St. Louis, MO, USA) in PBS. After incubation, samples were washed again and pictures were obtained by using an inverted fluorescent microscope (XDS30; Ningbo Sunny Instruments Co. , Ltd. , Zhejiang, China) equipped with a video camera (MD50; Mingmei, China). Cell viability in 3D culture ADSCs were prepared as described above and cell density was adjusted to 5. 0 × 10 6 cells/ml. Cell suspensions were mixed with different concentrations (4%, 6%, and 8%) of gelatin/mTG solution at 1:9 volume ratio. Aliquots of 100 µl were loaded into six-well TCPs and incubated at 37 °C for 2 h. After gelling, 2 ml of culture medium was added in each well and the TCPs were incubated at 37 °C in 5% CO 2. The medium was changed every three days. Cell morphology was monitored daily under an inverted optical microscope. After two weeks of cultivation, cell viability was tested via the calcein—AM/PI assay as mentioned before, and images were captured using a confocal laser scanning microscope (A1si, Nikon, Japan). Cell proliferation assay ADSCs were seeded on the surface of 2–8% hydrogels (2D culture) or embedded into hydrogels (3D culture) as described above at 5. 0 × 10 4 cells per well of 24-well TCPs. Cells were cultured in expansion medium for four weeks, and medium was changed every 3 days. At each time point, cells/hydrogel constructs from three parallel samples were rinsed thrice with PBS and incubated in 100 µl of 10% Alamar Blue solution (Yeasen, China) for 3 h. Subsequently, the incubation solutions were transferred to a 96-well TCP, and fluorescence was measured with a plate reader using excitation/emission wavelengths of 530/590 nm. ADSCs seeded on TCPs served as a negative control. Cardiac differentiation of ADSCs To investigate the differentiation status of encapsulated stem cells, ADSCs from the third passage were encapsulated in 6% gelatin/mTG hydrogels with the method described above (‘Cell viability in 3D culture’). After forming hydrogels, encapsulated cells were divided into two groups: one group was treated with normal cell expansion medium (see ‘Primary culture of ADSCs’), and the other was treated with differentiation medium. The differentiation medium consisted of DMEM with 1% P/S supplemented with 2% horse serum (HS; Hyclone, UT, USA) and 50 µg/ml L -ascorbic acid. The medium was changed twice a week, and a 14-day culture was performed. Control groups were seeded on TCPs supplemented with cell expansion medium or with differentiation medium for the same period. The expression of cardiac-specific genes NK2 homeobox 5 (Nkx2. 5), alpha cardiac myosin heavy chain 6 (Myh6), gap junction protein alpha 1 (Gja1), and myocyte enhancer factor 2C (Mef2c) in hydrogel-encapsulated cells and in control groups were measured by real-time reverse transcription-polymerase chain reaction (RT-PCR). Total RNA was extracted using Trizol reagent (Life Technologies). Complementary DNA was synthesized from 1 mg total RNA by employing RevertAid First Strand cDNA Synthesis Kit (K1622; Thermo Scientific, Vilnius, Lithuania). Complementary DNA samples were subjected to PCR amplification using Luminaris Color HiGreen qPCR Master Mix (K0391; Thermo Scientific, Vilnius, Lithuania). The DNA sequences of primers are listed in Table 1. PCR was performed with a real-time PCR Detection System (iCycler IQ5; Bio-Rad, Hercules, CA, USA). Cycles were programmed as follows: 95 °C for 10 min, 40 cycles of 15 s denaturation at 95 °C, 30 s at an annealing temperature of 57 °C, 30 s extension at 72 °C, with a final extension at 72 °C for 10 min. The 2 −ΔΔ Ct method was used to evaluate relative mRNA expression levels for each target gene. The expression of the housekeeping gene β -actin was employed for internal normalization. The product size was confirmed by running 10 µl of samples on 2% agarose gel electrophoresis. 10. 7717/peerj. 2497/table-1 Table 1 Sequences of PCR primers. Gene Primer sequence (5′-3′) Accession no. Product (bp) Nkx2. 5 F: CCTCGGGCGGATAAGAAAG R: ACTTGTAGCGGCGGTTCT NM_053651. 1 262 Myh6 F: ACACCAGCCTCATCAACCA R: CCTTCTCCTCTGCGTTCCT NM_017239. 2 105 Gja1 F: TGTGATGAGGAAGGAAGAGAAG R: TTGAAGAGGATGCTGATGATGT NM_012567. 2 192 Mef2c F: CGGACTGATGAAGAAGGCTTAT R: GGCTGTGACCTACTGAATCG XM_003749164. 1 254 β -actin F: GGACCTGACAGACTACCTCAT R: GAACCGCTCATTGCCGATA NM_031144. 3 217 Cell migration study in a 3D culture model The migration ability of ADSCs in 3D hydrogels was evaluated by using a specially designed model in which cells were seeded. In brief, 2–8% gelatin/mTG solutions were loaded in six-well TCPs at 2 ml volume in each well and incubated at 37 °C for 3 h. Subsequently, a 2 mm radius hole was made at the center of each hydrogel by using a punch, and 25 µl of 8% gelatin/mTG solution mixed with 1. 0 × 10 7 cells was filled into the hole. The added gelatin/mTG/cells constructs were set to solidify for 2 h and then each well was supplemented with 2 ml expansion medium. The cells were cultured at 37 °C with 5% CO 2 for 14 days, and the culture medium was changed every 3 days. Meanwhile, the gelatin/mTG/cells constructs were monitored daily to observe the cell migration from the transplanted site into the surrounding matrix. Cell migration images were captured, and cell migration distances from the hole edge to the cell outgoing front in all directions were measured via image analysis software (Image Pro Plus 6. 0; Media Cybernetics, Rockville, MD, USA). Afterward, the average migration distance was calculated for statistical analysis. In vivo implantation of hydrogel The in vivo biological response to gelatin/mTG hydrogels was assessed at four weeks in subcutaneous SD rat models (age 6–7 weeks) with a total of four rats included in this study. Hydrogels with 4% and 8% gelatin were prepared as described above, and the hydrogels were cut into sheets of 10 mm × 10 mm × 1. 5 mm size. Prior to surgery, the rats were anesthetized with 10% chloral hydrate and the areas of surgery were shaved and disinfected with iodophors. A pair of hydrogel sheets with 4% and 8% gel concentrations was surgically placed within subcutaneous pockets located on both sides of an adult rat dorsum. Three rats were implanted with hydrogels, and the last rat was subjected to surgery, but no hydrogel was implanted. The incision was covered in Betadine ointment after surgical suture and assessed for signs of infection for three days after surgery. All rats survived the surgical procedure without surgical complications. After four weeks the rats were sacrificed and the implants were harvested and sectioned at 10 µm thick with a cryosection microtome (Leica CM1950; Leica, Wetzler, Germany). The sections were fixed in 4% paraformaldehyde and processed for histological analysis using hematoxylin/eosin staining. Statistical analysis Data are presented as mean ± SD. Statistical analyses were performed using SPSS software (version 14. 0). Statistical significance between two groups was determined by Student’s t -test. Results for more than two groups were evaluated by one-way ANOVA with least-significant difference (LSD) test. P < 0. 05 was considered statistically significant. Results Evaluation of gelation time and gel strength Gelation time of the gelatin/mTG solution was evaluated at 37 °C via a bottle-invert method ( Fig. 1 ). Depending on mTG dosage or gelatin percentage in gelatin/mTG solution, the gelation time was verified to be controllable. Decrease in either mTG dosage or gelatin percentage extended gelation time. When mTG dosage was kept constant at 10 U/g pro, the gelation time was inversely proportional to gelatin concentration (shown as Table 2 ). The 10% gelatin/mTG solution only took 15 s before arriving gel state. By contrast, the 1% gelatin/mTG solution retained a flow state and did not become a hydrogel. 10. 7717/peerj. 2497/fig-1 Figure 1 Morphology of 4% gelatin/mTG hydrogel. (A) Gelatin/mTG mixing solution. (B) The gelatin/mTG solution turned to a hydrogel at 37°C and the gel state was evaluated by bottle-invert method. (C) With sufficient gel strength, a piece of 4% gelatin/mTG hydrogel can be held by a pair of forceps. 10. 7717/peerj. 2497/table-2 Table 2 Gelation time and gel strength determined by gelatin percentage (mTG dosage = 10 U/g pro). Gelatin percentage 1% 2% 3% 4% 5% 6% 8% 10% Gelation time (min) – 124. 33 ± 6. 28 ∧ 89. 5 ± 2. 59 ∧ 50. 33 ± 4. 27 ∧ 38. 83 ± 0. 98 ∧ 25. 5 + 0. 55 ∧ 20. 5 ± 0. 55 ∧ 15. 67 ± 0. 52 ∧ Gel strength (kPa) – – 2. 80 ± 0. 28 # 5. 49 ± 0. 31 ∗ 12. 42 ± 0. 83 #∗ 20. 94 ± 1. 48 #∗ 34. 99 ± 1. 49 #∗ 54. 59 ± 5. 12 #∗ Notes. ∧, #, ∗ P < 0. 01, when compared to each other. To investigate the mechanical properties of gelatin/mTG hydrogels, we employed a digital pull and push dynamometer to evaluate gel strength. We found that gel strength is proportional to the concentration of gelatin solution. For example, gel strength increased rapidly as gelatin concentration increased from 2% to 6% and retained a relative constant strength thereafter even if increasing gelatin concentration to 10%. On the other hand, considering that gel strength may also be correlated with mTG dosage we applied a variety of mTG dosages with 4% gelatin solution to test their gelation time and gel strength (shown as Table 3 ). The results showed 10 U/g pro could be a compromised dosage that could bring about adequate gel strength and minimize the side effects possibly caused by excessive mTG. 10. 7717/peerj. 2497/table-3 Table 3 Gelation time and gel strength determined by mTG dosage (Gelatin percentage = 4%). mTG dosage (U/g pro) 2 5 10 20 40 Gelation time (min) 180. 33 ± 8. 02 $ 97. 33 ± 4. 27 $ 50. 33 ± 4. 27 $ 34 ± 1. 41 $ 20. 66 ± 0. 82 $ Gel strength (kPa) 1. 65 ± 0. 10 #∗∧ 3. 55 ± 0. 56 #∗∧ 5. 49 ± 0. 31 # 5. 50 ± 0. 29 ∗ 5. 58 ± 0. 34 ∧ Notes. $, #, ∗, ∧ P < 0. 01, when compared to each other. Hydrogel degradation performance Given that cells seeded in a hydrogel may secrete several proteases, such as collagenase, which could lead to hydrogel degradation, the degradation performance of gelatin/mTG hydrogel should be evaluated. The degradation test was carried out in two parallel groups: cell-free hydrogels and cell-containing hydrogels. In the test of cell-free hydrogels ( Fig. 2A ), the degradation rate of 2% concentration of hydrogels was inversely proportional to culture time. Within three weeks, the hydrogels lost more than half of their original mass. At the eighth week, only approximately 5. 6% of original mass remained. For the other three kinds of hydrogels (4%, 6% and 8%), their degradation curves were different. Apparent degradation in these hydrogels was not observed until the eighth week. For the hydrogels with 4% concentration gelatin, 3. 6% mass loss was observed 10. 7717/peerj. 2497/fig-2 Figure 2 Degradation properties of gelatin/mTG hydrogels containing with or without cells. (A) Degradation curves of cell-free hydrogels. (B) Degradation curves of cell-containing hydrogels. In the test of cell-containing hydrogels ( Fig. 2B ), we found that 2% hydrogels degraded more rapidly than cell-free hydrogels. For three weeks, nearly 80% of hydrogel mass was lost. Moreover, the 4% hydrogel also showed significant degradation. At the eighth week, nearly half of the hydrogel mass was lost. However, the 6% and 8% hydrogels did not show severe degradation. For 6% hydrogel, approximately 12% of gel mass was lost after eight weeks of incubation; meanwhile for 8% hydrogels, mass loss was less than 4%. Cell morphological observation from 2D culture We assessed the biocompatibility of gelatin/mTG hydrogels and its effect on cell adhesion. ADSCs were seeded on the surfaces of various hydrogels for two weeks. Cell viability was evaluated via live/dead staining assay. Cell staining showed that the vast majority of ADSCs cultured on 2%–10% gels were stained green (on behalf of living cells), and only few cells were stained red (on behalf of dead cells). This finding means that these hydrogels present good biocompatibility and are suitable for cell 2D culture. In addition, regarding cell morphology, the effects of different concentrations of hydrogel materials on cell growth and adhesion behavior were different. The cells seeded on 2% hydrogels plunged into hydrogels and grew inside owing to weaker gel strength, which led to more rounded or stick-shape cells appearing than other hydrogels ( Fig. 3A ). As the hydrogel concentration increased to 4%, cell shape progressively assumed more of barbed-like or caltrop-like patterns, which represented the cell pseudopodia stretching out in different directions in a 3D gel space ( Fig. 3B ). With further increasing hydrogel concentration ( Figs. 3C and 3D ), cells gradually spread out and grew in size. When the hydrogel concentration increased to 10% ( Fig. 3E ), the cell growth pattern on the gel was similar to that on TCP ( Fig. 3F ). 10. 7717/peerj. 2497/fig-3 Figure 3 Cell live/dead fluorescent staining of ADSCs seeded on the surface of gelatin/mTG hydrogels. ADSCs cultured on (A) 2%, (B) 4%, (C) 6%, (D) 8%, and (E) 10% gelatin hydrogel and on (F) TCP at day 14 after seeding. Living cells were stained by calcium—AM (Green), and dead cells were stained by PI (Red). Scale bars = 100 µm. Cell morphological observation from 3D culture Compared with hydrogel-surface culture, the 3D culture of embedding ADSCs into a hydrogel better mimicked in vivo cell growth. Three concentrations (4%, 6% and 8%) of hydrogels were included in 3D culture experiments. As for 2% hydrogels, owing to their relatively low gel strength and susceptibility to degradation, we did not consider the hydrogels as experimental candidates. As for 10% hydrogels, we also excluded them from the experiment because in both the hydrogel degradation test and the 2D cell culture experiment, their performances did not show apparent difference with that of 8% hydrogels. In 3D culture, we found that the encapsulated ADSCs from all tested hydrogels present good viability ( Fig. 4 ). After two weeks of culture, the slice views photographed by laser scanning confocal microscopy showed that the cells distributed evenly on the hydrogels with numerous cell protrusions. In the synthesized 3D images, we clearly saw that the ADSCs almost filled all areas inside the hydrogels. However, with the increase in hydrogel concentrations, some of the cells residing deep in these hydrogels died because of insufficient nutrient metabolism caused by the compact hydrogel. 10. 7717/peerj. 2497/fig-4 Figure 4 Observing cell viability in 3D cultures by a confocal laser scanning microscope. (A, C, E) In slice views, ADSCs embedded into the 4–8% concentrations of gelatin/mTG hydrogel exhibited good viability with abundant protrusions. (B) In 3D views, ADSCs are distributed in hydrogels extensively and exhibit vigorous proliferation in 3D culture. Scale bars = 100 µm. Evaluation of cell proliferation in 2D and 3D culture ADSC proliferation capabilities were assessed by Alamar Blue assay. During four weeks of 2D culture, significant difference in cell growth behavior was observed ( Fig. 5A ). On day 2, evident differences were not observed among the cells cultured on different substrates. However, from day 4 to day 12, the cells on TCP apparently showed a higher proliferation rate compared with cells cultured on the gel surface. However, after 12 days of culture, the cell number on TCP almost linearly decreased as the cells stopped growth caused by cell contact inhibition, and some cells were detached from the TCP surface. From the 2D culture experiments we found cell proliferation ability was almost proportional to the concentration of hydrogel. We also found that cell growth on lower hydrogel concentration (2% and 4%) decreases on day 18 possibly because of degradation of these materials, which led to the loss of the corresponding mechanical support. For 8% or 10% hydrogels, the cell number remained increasing despite the cell growth rate started to slow down after three weeks of culture. 10. 7717/peerj. 2497/fig-5 Figure 5 Cell proliferation capabilities assessed by Alamar Blue assay after four weeks of 2D or 3D culture. (A) Growth curves of ADSCs cultured on the surface of different hydrogels; (B) growth curves of ADSCs embedded into different hydrogels. ADSC proliferation in 3D culture was carried out using Alamar Blue assay ( Fig. 5B ). Cellular proliferation behavior in 3D culture presented a similar situation as that in 2D culture. The hydrogels at 8% or 10% concentration shows good performance in supporting cell growth because the cell number in these hydrogels continued to increase even after four weeks of culture. Cells on hydrogels with 4% gelatin concentration exhibit fast growth rate in the first two weeks; however, cells begin to fall down from day 18 after cell seeding. The 2% hydrogels exhibit poor performance and few cells survived in hydrogels after two weeks of culture. We speculate that the 2% hydrogels present weaker mechanical properties, together with the degradation of collagenase secreted by the residing cells, leading to poor performance in supporting cell 3D growth. Relative mRNA expression of encapsulated ADSCs To measure the effects of gelatin/mTG hydrogel on cardiogenic differentiation of ADSCs, we evaluated cardiac markers Nkx2. 5, Myh6, Gja1, and Mef2c were evaluated by real-time quantitative PCR. All values are normalized to the expression level of β -actin. The mRNA expression level of ADSCs cultured on TCPs in cell expansion medium (negative control) is indicated as “1. ” Each measure of gene expression in the encapsulated ADSCs is presented as a fold change over that of negative control. The results are shown in Fig. 6. The encapsulated ADSCs in differentiation medium present the highest mRNA expression level among all detected samples; their expression levels of Nkx2. 5, Myh6, Gja1, and Mef2c are 15. 26 ± 2. 06-, 20. 80 ± 2. 79-, 31. 01 ± 1. 62-, and 14. 64 ± 0. 82-fold, respectively. For the encapsulated ADSCs in normal cell expansion medium, their mRNA expression levels of Nkx2. 5, Myh6, Gja1, and Mef2c are 5. 56 ± 0. 67-, 5. 97 ± 1. 22-, 15. 69 ± 2. 05-, and 4. 60 ± 0. 91-fold higher compared with negative controls, respectively. No statistical difference was found in cells cultured on TCPs in the differentiation medium compared with negative controls. 10. 7717/peerj. 2497/fig-6 Figure 6 Cardiogenic gene expression of encapsulated ADSCs in cell expansion medium (3D+EM) or differentiation medium (3D+DM) for two weeks; ADSCs cultured on TCPs in expansion medium (TCP+EM) or differentiation medium (TCP+DM) were also evaluated. The expression levels of cardiac markers (A) Nkx2. 5, (B) Myh6, (C) Gja1, and (D) Mef2c were detected by qPCR. All values are normalized to the expression level of β -actin. (**, p < 0. 01). Evaluation of cell migration in 3D hydrogels For repairing the damaged tissue and organ, cells/scaffold construct implanted into the body should be able to release cells from the scaffold into the surrounding tissue. In this study, we designed a 3D cell migration model to simulate the in vivo transplant environment. The experiments show that cell migration activities started as early as the 2nd day after in vitro hydrogel implantation and continued to the end of the experiments. Low concentration of hydrogels favors cell migration. With the increase in gel concentration, the hydrogels became dense, and the average migration distances of ADSCs encapsulated in the hydrogels were decreased. After two weeks cells in the 2% hydrogels migrated by 9. 36 ± 0. 58 mm with an average migration speed of about 0. 67 mm/day. By contrast, in the other three hydrogels cells migrated by about 3–5 mm in the same period ( Fig. 7A ). The experiments also showed that the released cells from the cells/hydrogel construct maintained an extremely strong vitality and proliferative capacity during two weeks of experiments ( Figs. 7B – 7E ). 10. 7717/peerj. 2497/fig-7 Figure 7 Evaluation of cell migration in 3D hydrogels. (A) Migration distance of the released ADSCs crawled outward from the edge of the in vitro implantation site into the surrounding hydrogels of various hydrogel concentrations. (B–E) At day 10 after implantation, the cells were released from the injected gelatin/mTG hydrogel and crawled into the surrounding hydrogels at concentrations of (B) 2%, (C) 4%, (D) 6%, and (E) 8%. “Inject” denotes the injected gel; the green dash line denotes the edge of the injected gel. Scale bars = 200 µm. In vivo response to gelatin/mTG hydrogels To characterize the local tissue response to gelatin/mTG materials, we implanted 4% and 8% concentration hydrogels subcutaneously into an adult SD rat model and explanted after four weeks for analysis. Approximately 6–8 h after surgery, the rats could eat and drink, and no breathing difficulties and unusual activities were observed. Their wounds were healed without the formation of scar tissue within a week ( Fig. 8A ). The 8% hydrogels could be clearly observed under the skin throughout the experiment period, indicating that the hydrogels are not degraded significantly. However, the 4% hydrogels could only be observed in the first two weeks and are difficult to identify after another two weeks. We speculated that material degradation had occurred. After four weeks, the rats were sacrificed, and the implants were harvested, thus confirming our speculation. All 4% concentration samples were severely eroded, but no macroscopic signs of inflammation or toxicity were evident in the tissue surrounding the implants. As for 8% hydrogel samples, apparent evidence of capsule formation was found around the implants, and only slight degradation occurred. Hematoxylin and eosin staining of sections throughout the explanted hydrogels showed no apparent infiltration of cells into the 8% hydrogels, but a small number of cells had invaded into the 4% hydrogel samples ( Fig. 8B ). We can see from morphological observation, that the cells were not macrophages or inflammatory cells but may be endothelial cells instead, migrating from surrounding tissues, to form capillary networks inside hydrogels. 10. 7717/peerj. 2497/fig-8 Figure 8 In vivo biological response to the gelatin/mTG hydrogels was assessed in subcutaneous SD rat models. (A) Wounds of the rats were healed without the formation of scar tissue in a week. (B) A small number of cells had invaded into the samples with 4% hydrogel, as detected by hematoxylin and eosin staining when implants were harvested after four weeks of experiments. Scale bars = 100 µm. Discussion Evaluation of the performance of a biological scaffold usually involves two important aspects: whether the hydrogel can support cell adhesion and cell proliferation and determining whether the mechanical strength of the hydrogel material can meet the requirements of practical applications. For example, collagen-based hydrogel exhibits excellent cell adhesion property but poor mechanical performance. Synthetic hydrogel materials, such as polyethylene glycol and poly lactic acid, exhibit excellent mechanical properties, but the lack of cell binding domains leads to weak cell adhesive ability. Moreover, the degradation rate of most synthetic materials in vivo is slow, thereby affecting tissue regeneration. One of the advantages of choosing gelatin as a biological scaffold is its abundance of cell binding domains. Gelatin is a common structural protein and is obtained from the degraded product of collagen protein, but the majority of the amino acid sequences of collagen is retained. In particular, the molecular weight of gelatin is smaller than that of collagen, and the space conformation of the gelatin protein is simpler, thus exposing more cell binding sequences of Arg–Gly–Asp (RGD) that can be identified by cellular integrin proteins. Therefore, gelatin hydrogel possesses excellent cell-response features, which can support cell proliferation, migration, differentiation, and development into a complex organizational structure. Given that the mechanical properties of the non-crosslinked gelatin solution are extremely weak, gelatin cannot be directly used as a biological scaffold material. Gelatin is often used as substrate material in cell culture to promote cell adhesion on TCP. Gelatin can also be used as a part of composite material to increase cell biocompatibility ( Han et al. , 2015 ; Nieto-Suarez, Lopez-Quintela & Lazzari, 2016 ). Enzymatic crosslinking is a new approach in preparing gelatin hydrogel. Transglutaminases are a class of natural enzymes; there are at least eight genetically distinct transglutaminases in mammalian tissue, blood, ECM and on cell surfaces. All of these are Ca 2+ dependent and are related to essential homeostatic processes such as blood coagulation (Factor XIIIa); cell differentiation, death and ECM stabilization (tissue transglutaminase); and maintenance of epidermal integrity (keratinocyte transglutaminase). Transglutaminases have a transamidation activity and catalyzes the acyl-transfer reaction between the ε-amino group of lysine and the γ -carboxyamide group of glutamine in proteins. mTG is Ca 2+ independent and has transamidation ability similar to that of the mammalian versions ( DM et al. , 2006 ; Paguirigan & Beebe, 2007 ). The process of crosslinking gelatin/mTG is irreversible in physiological temperature range. The approach overcomes the shortage of reversibility in the obtained hydrogels via conventional physical crosslinking method. In addition, our experiments showed that the gelatin/mTG hydrogels possess high plasticity, and their mechanical properties are adjustable. These hydrogels can be used as preshaping hydrogels for cell inoculation or mixed with cells in solution and implanted in vivo by injection; subsequently, hydrogels are formed according to the spatial shape of the implantation site. In our experiments, we found that the mTG is not cytotoxic, given that enzymatic crossslinking prevents potential toxicity caused by the residue of chemical crosslinkers. In the study of mTG dosage screening, we found that 2 U/g pro could crosslink 4% gelatin solution into a hydrogel; however, the process of crosslinking was time-consuming. Increasing mTG dosage could shorten the gelation time and could promote gel strength. A 10 U/g pro could be a proper dosage; a dosage higher than this does not significantly increase the gel strength. We also tried to use the dosage of 40 U/g pro to crosslink the gelatin hydrogel for cell culture; however, we did not found the significant differences in cell morphology and proliferation between the experiment group with 40 U/g pro dose and the control group with 10 U/g pro dose. Therefore, we concluded that a suitable mTG dosage is 10 U/g pro. In evaluating the intrinsic properties of hydrogels, gelation time is a notable factor. For example, when mixing gelatin hydrogel with cells in 3D culture, we expect cells to be evenly suspended in the hydrogel. However, in the actual operation, if the gelation time is poorly controlled, a uniform cell distribution is difficult to achieve. Given that gelation time is related to the concentration of gelatin solution, the low-concentration gelatin solution exhibits long gelation time. When the gelatin/mTG solution was mixed with cells in the state prior to forming a gel, the cells will be precipitated by gravity, and the aim of uniform cell distribution cannot be achieved. High-concentration gelatin solution exhibits short gelation time. Therefore, the cells must be quickly mixed with the gelatin solution prior to gelling; otherwise, abundant bubbles will appear in the resultant hydrogel. With further stirring of the gel, the mechanical structure of the hydrogel will be undermined. In this study, we found that hydrogel concentration affects cell growth morphology and cell proliferation. In 2% gelatin hydrogels, the cell growth morphology showed a round or stick-like shape because of the relatively low gel strength. Coupled with faster hydrogel degradation rate, low-concentration hydrogels cause significant cell loss in the cell proliferation experiments. Therefore, 2% gel was not used for cell migration and subcutaneous implantation. In 2D cell culture experiment, the high-concentration hydrogel exhibited a similar cell growth pattern to that on TCP. However, we found certain growth behavior differences between the cell cultured on hydrogel and on TCP: upon reaching confluence, the cells stopped growth and began detaching from the TCP culture surface. By contrast, cells on hydrogel maintained excellent adhesion owing to the abundant integrin binding domains provided by the gelatin hydrogel. In addition, the cells gradually grew into the hydrogel inside and spread out all over the gel space. This growth mode enhanced cell adhesion on the hydrogel surface, expanded cell growth space, and promoted cell proliferation. As seen from the 3D culture experiment results, the gelatin/mTG hydrogels exhibit excellent biocompatibility with extremely high cell survival rate. Moreover, the shape of encapsulated cells changed from a round shape to a barb-like shape, indicating that cell pseudopodia began stretching out and that cells were in a good growth status. Studies indicated that anchorage-dependent cells that remain round or in a non-adhesive state will eventually undergo cell apoptosis ( Chen et al. , 1997 ; Kuwahara et al. , 2010 ). ADSCs have attracted considerable attention because of their abundant stem cell sources. ADSCs compose a plastic-adherent cell population that includes vascular and adipocyte progenitor cells and adult multipotent mesenchymal stem cells (MSCs) ( Madonna & De Caterina, 2010 ; Nguyen et al. , 2016 ). In our previous study ( Li et al. , 2015 ), we examined the surface protein expression of rat ADSCs at Passage 3 via flow cytometry analysis. The results showed that CD29, CD44h, CD49d, and CD90 were positive in expression, whereas CD45 and CD106 were negative. Our results were in agreement with the reports in the literature ( Bourin et al. , 2013 ; Guo et al. , 2016 ). In a non-differentiating medium, ADSCs can retain strong proliferation capability, maintain their phenotypes, and exhibit stronger multidirectional differentiation potential even after being passaged for 25 times ( Zhu et al. , 2008 ). ADSCs can differentiate into osteoblasts, chondrocytes, adipocytes, and cardiomyocytes via appropriate induction condition. Previous studies have indicated that the altering modulus of a hydrogel exerts a profound effect on the lineage commitment of stem cell ( Chatterjee et al. , 2010 ; Li et al. , 2012 ). Li et al. (2012) encapsulated human MSCs in composited acrylamide hydrogels with three different moduli: 16, 45 and 65 kPa, after 14 days of in vitro culture, they found more than 76% of MSCs expressed cardiac specific markers in the 45 and 65 kPa hydrogels, and MSCs in the 65 kPa hydrogel had the highest differentiation efficiency. In our study, we evaluated the in vitro cardiac differentiation capacity of encapsulated ADSCs in 6% gelatin/mTG hydrogel (∼20. 94 kPa) with or without cardiac induction medium. For the encapsulated ADSCs in normal cell expansion medium, the mRNA expression levels of Nkx2. 5, Myh6, Gja1, and Mef2c were higher than those of cells cultured on TCPs. This result shows that the modulus of gelatin/mTG hydrogel can affect cell differentiation. Furthermore, the encapsulated ADSCs in differentiation medium presented higher levels of cardiogenic gene expression compared with those in the expansion medium, further confirming that the synergistic effect of the hydrogel modulus and differentiation medium can contribute to the cardiac differentiation of encapsulated ADSCs. By contrast, the differentiation medium alone could not induce ADSCs on TCPs toward cardiomyocyte fate. In the cell migration experiments, we found that the cells were not only released from hydrogel encapsulation but also invaded the surrounding hydrogels. This finding shows that hydrogels are able to release entrapped cells and simultaneously accept the migrated cells from the surroundings. In short, cell release from the hydrogel encapsulation is controllable, and the cell migration rates can be regulated by changing the hydrogel concentration. However, increasing the hydrogel concentration will engender a certain degree of cell apoptosis in the hydrogel inside (under the hydrogel surface ∼200–300 µm). Therefore, if the thickness of cells/hydrogel construct exceeds 1 mm, we recommend that a multi-layered overlay should be considered for hydrogel design to facilitate the exchange of nutrients. When implanted into the body, the hydrogel can be invaded by the surrounding capillary network, thus solving the problem of blood supply to a certain extent. As we have seen in subcutaneous implantation experiments, some of endothelial-like cells had invaded the 4% hydrogel. If the experiment time is prolonged, capillary network may appear within the hydrogel. Moreover, the gelatin/mTG hydrogel can not only be used as a cell scaffold material, but also has a protective effect on the encapsulated cells against an inflammation response. Acute inflammatory response can produce unfavorable microenvironment for the injected stem cells. Inflammatory responses are time dependent, in the case of neural tissue, the most intense inflammatory response is within three days since stem cell transplantation ( Okano, 2002 ; Kuwahara et al. , 2010 ). From the experiment results of in vivo hydrogel implantation in our study, even the 4% hydrogel is able to maintain the gel state in the body for more than two weeks. This shows that gelatin/mTG hydrogel may serve as a cell barrier against acute inflammation response. Our next experiment will observe whether gelatin/mTG hydrogel can block inflammatory effects on the encapsulated cells. In conclusion, the gelatin/mTG hydrogel is a potential scaffold material, either as a cell vehicle for in vivo implantation for wound healing, and soft and hard tissue repair, or as a drug delivery carrier for drug screening. This hydrogel can also be combined with other biological materials to form a composite material with more functions. We believe that this hydrogel material will be widely used in the field of tissue engineering and regenerative medicine as well as the field of drug release study. In our future study, the effect of different hydrogel moduli on ADSCs lineage commitment will be investigated, and immunofluorescence assay and flow cytometry analysis should be performed to further evaluate the cardiac differentiation status of encapsulated ADSCs. In addition, chromosomal karyotyping to ADSCs should also be examined for detecting chromosomal abnormality. Moreover, in vivo hydrogel/cells transplantation experiment will help us to better understand the function of gelatin/mTG hydrogel in tissue repair and organ reconstruction. Supplemental Information 10. 7717/peerj. 2497/supp-1 Supplemental Information 1 Raw data for Table 2 Gelation time and gel strength determined by gelatin percentage (mTG dosage = 10 U/g pro) Click here for additional data file. 10. 7717/peerj. 2497/supp-2 Supplemental Information 2 Raw data for Table 3 Click here for additional data file. 10. 7717/peerj. 2497/supp-3 Supplemental Information 3 Raw data for Figure 2 Click here for additional data file. 10. 7717/peerj. 2497/supp-4 Supplemental Information 4 Raw data for Figure 5 Click here for additional data file. 10. 7717/peerj. 2497/supp-5 Supplemental Information 5 Raw data for real time RT-PCR in Figure 6 Click here for additional data file. 10. 7717/peerj. 2497/supp-6 Supplemental Information 6 Raw data for Figure 7 Click here for additional data file.

10. 7717/peerj. 2661

PeerJ

On the intrinsic sterility of 3D printing

3D printers that build objects using extruded thermoplastic are quickly becoming commonplace tools in laboratories. We demonstrate that with appropriate handling, these devices are capable of producing sterile components from a non-sterile feedstock of thermoplastic without any treatment after fabrication. The fabrication process itself results in sterilization of the material. The resulting 3D printed components are suitable for a wide variety of applications, including experiments with bacteria and cell culture.

Introduction Mass-produced, disposable products are ubiquitous in research laboratories. Roughly three billion microcentrifuge tubes are manufactured each year ( Hashemi, 2006 ). The ubiquity of these products has helped to standardize molecular methods by reducing variability from experiment to experiment and from laboratory to laboratory. However, the proliferation of these products has come at the cost of in-house expertise in fabrication. Without these skills, researchers are increasingly dependent on vendors to anticipate and provide for their needs. If an experiment calls for a component that is unusual or unique, researchers are forced to improvise or to redesign the experiment using more readily available components. These restrictions are not necessarily detrimental; standardized materials are crucial for reproducibility. Nevertheless, there are experiments in which the need for a custom component cannot be avoided. Many researchers have turned to 3D printing, a process by which three- dimensional objects are built up additively, to fill these needs. In some respects, the technology is more limited than traditional fabrication techniques used for laboratory equipment, such as metalworking or glassblowing; it is mostly limited to materials that can be melted and extruded at relatively low temperatures (150 °C–300 °C), such as thermoplastics. At the time of this writing, there are few inexpensive machines capable of combining more than one material. In other respects, 3D printing is more powerful than traditional fabrication techniques; additive manufacturing permits the creation of geometries that are impossible by other means, such as captured free moving parts. However, the principal advantage of additive manufacturing is the ability to move directly from a digital design to a finished part. It is not necessary to have a wide variety of specialized shop tools or the personnel and skills needed to operate and maintain them. One of the most important properties of basic labware in the biological sciences is sterility, and one of the most frequent questions laboratory biologists ask when they first learn of 3D printing is, “Can I autoclave these things?” Unfortunately, most thermoplastics that are widely used in biomedical applications, particularly polylactic acid (PLA) and polyglycolic acid (PGA), will not survive a standard autoclave cycle ( Rozema et al. , 1991 ). Sterilization with γ -radiation is effective, but causes drastic changes to the biochemical properties of the material ( Gilding & Reed, 1979 ). 1 1 For a detailed review of these studies, we recommend the review by Athanasiou & Niederauer (1996). Here we detail our work demonstrating that the 3D printing process itself appears to be sufficient for ensuring sterility. We note that the fused deposition modeling (FDM) 3D printing process, in which a thermoplastic filament is heated to melting and forced through a narrow tube under high pressure, resembles a sort of extreme pasteurization. Figure 1 compares the FDM 3D printing to several sterilization processes (note that thermal contact time is in log scale). The 3D printing process holds the material at a higher temperature for longer duration than both Ultra-High Temperature (UTH) pasteurization, which is used to produce shelf-stable milk (138 °C for two seconds) and high-temperature, short-time (HTST) pasteurization used for dairy, juice and other beverages and liquid ingredients (71. 5 °C–74 °C for 15–30 s). The only legal pasteurization method that exceeds the thermal contact time typical of FDM 3D printing is mentioned in Title 21, Sec. 1240. 61 of the Code of Federal Regulations, which permits milk to be treated at 63 °C for 30 min. This is a convenient sanitation regime for milk in non-commercial settings (indicated in Fig. 1 as “stovetop” pasteurization). 3D printing is also both hotter and longer duration than thermization, a process used to extend the shelf life of raw milk that cannot be immediately used, such as at cheese making facilities. 10. 7717/peerj. 2661/fig-1 Figure 1 Temperature and duration of various sterilization processes. Temperatures and durations for various methods of sterilization compared to fused deposition modeling (FDM) 3D printing. The extrusion process most closely resembles pasteurization, in which non-sterile liquid is forced through a narrow, heated tube. High-temperature, short-time (HTST) pasteurization is used for milk, fruit juices and other beverages and ingredients. Ultra-high temperature (UTH) processing is used to produce products such as shelf-stable milk that do not require refrigeration. Stove- top pasteurization (30 min at 63 °C) is indicated as “stovetop” pasteurization. Thermization, a process used to extend the shelf life of raw milk that cannot be immediately used, such as at cheese making facilities. Typical autoclave cycles using prevacuum, and gravity displacement are indicated as “prevacuum” and “gravity, ” respectively. A typical “flash” sterilization cycle for a gravity displacement sterilizer is also indicated. Pasteurization processes are indicated in black, autoclave processes in red, and thermization in orange. For most materials and toolpaths, FDM 3D printing is also hotter than typical autoclave cycles for both gravity displacement steam sterilization and prevacuum steam sterilization. “Flash” steam sterilization using a gravity displacement sterilizer must reach 132 °C for 3 min. The Centers for Disease Control guidelines for gravity displacement steam sterilization require that the cycle reaches 121 °C for 30 min, or 4 min at 132 °C using prevacuum steam sterilization. 3D printing thermoplastics using FDM typically requires temperatures between 190 °C and 240 °C, depending on the material and the print parameters. Because the fabrication process calls for different extrusion rates over the course of a print, the thermoplastic generally dwells in the melt region of the nozzle for between ten seconds and several minutes. To calculate the thermal contact time for an FDM 3D printer, we use the formula (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}T(f)= \frac{m\pi { \left( \frac{{d}_{f}}{2} \right) }^{2}}{{d}_{n}hf} \end{eqnarray*}\end{document} T f = m π d f 2 2 d n h f where f is the feed rate in millimeters per second, h is the layer height, d f is the filament diameter, d n is the nozzle diameter, m is the length of the melt zone. Because the length of the melt zone can be difficult to measure directly, but it may be inferred by using the area within the nozzle that has to be cleaned of melted plastic after a jam. For a feed rate of 50 mm/s, the thermal contact time in our 3D printer is about 16 s at 220 °C, although the print plan for a given part usually involves non-printing travel commands and regions where printing is carried out at a slower feed rate, resulting in longer thermal contact times. Besides contact time and temperature, many sterilization protocols stipulate that high pressure must also be achieved. Depending on the protocol, pressures may range from about 40 to 220 kPa (6–31 PSI). The pressure inside the melt zone of a 3D printer nozzle is more difficult to calculate, as it depends on the fluid dynamics within the nozzle. Many common thermoplastics, such as PLA, are non-Newtonian fluids when melted, which further complicates the question. With those caveats in mind, we offer some rough estimates of the pressure within the nozzle. At one extreme, the maximum possible pressure would occur when the force from the viscous fluid exiting the nozzle equals the maximum holding force of the stepper motor driving the extruder. For our printer, this is about 50–60 Newtons distributed over the area of the nozzle, which has a diameter of 0. 4 mm. In principle, this would translate to a pressure of about 400, 000 kPa (57, 000 PSI) at the aperture, about two thousand times the pressure of an autoclave cycle. In practice, the holding force of the motor is distributed over a larger area by the hydrodynamics of the melted plastic. If the force were distributed over the whole inner surface of the nozzle (about one square centimeter), that would result in a pressure of about 600 kPa (87 PSI), or about triple the highest autocalve pressure. Normally, printers operate at some fraction of maximum flow rate, and of course melting thermoplastic is not a simple fluid, and so the pressure is not distributed evenly. In our experiments, it is likely that the pressure was often below autoclave pressures. Nevertheless, the glass transition for materials like PLA occurs very abruptly, with only a few degrees separating the solid and liquid phase. Lowering the print temperature by a small amount can lead massive increases in pressure. With some experimentation, it should be relatively easy to operate a 3D printer with nozzle pressures well in excess of autoclave pressures. For example, the control firmware could be modified to make small adjustments to the temperature to match the flow rate, or the user could specify the temperature and flow rates in the print planning software to maintain a minimum pressure in the nozzle. Here we report our findings for a battery of culturing experiments conducted with 3D printed parts manufactured with consumer 3D printers. Several variations of sterile technique were tested; we printed parts onto surfaces treated with ethanol, onto flame-treated aluminum foil, and under UV light. Finally, we printed onto non-sterile carpenter’s tape, and then handled the parts with flamed forceps. To our surprise, all of these methods seem to be at least somewhat effective at producing sterile parts. We found that the resulting parts appear to be sterile under a wide variety of culture conditions known to enrich for a broad spectrum of microorganisms. This work was carried out in three laboratories across the United States, with experiments coordinated and results shared openly using Twitter. Much of this correspondence is directly referenced by this manuscript so that readers may follow how the research actually unfolded. The two 3D printers used are installed at the UC Davis Genome Center and the BEACON Center at Michigan State University, and most of the culturing work was done at the University of Michigan Medical School. After the initial experiments at UC Davis ( Fig. 2 ), researchers at Michigan State University independently developed variations on those techniques. When the initial results were reproduced, a battery of test parts were prepared using several variations on the technique and mailed to the University of Michigan for culturing. The work was conducted in this way in order to reduce the “in our hands” effect, so that we could be reasonably confident that others could successfully achieve the same results. 10. 7717/peerj. 2661/fig-2 Figure 2 Preliminary experimental results. Growth after 96 h at 37 °C in a shaking incubator. The beaker labeled (A) contains LB media inoculated with PLA plastic extruded from the printer nozzle at 220 °C. The beaker labeled (B) contains LB media inoculated with a segment of unextruded PLA plastic filament from the same spool. The beaker labeled (C) contains uninoculated LB. Results In all of experiments described, the material used for 3D printing was non-sterile polylactide, or poly-lactic-acid (PLA) filament sourced from suppliers that primarily serve the hobbyist market. This material was selected for a number of reasons. PLA is very easy to work with in 3D printing, with good layer-to-layer adhesion and very little shrinking or warping. It is also biodegradable, which is attractive for environmental reasons. PLA and related polymers are also known to be non-toxic, bio-compatible, and are widely used in medical applications, notably soluble medical sutures. When noted, UV treatment was carried out by placing the 3D printer inside a laminar flow hood equipped with a 15 watt germicidal florescent bulb (Philips, model G15T8). The bulb remained activated during the printing process, and completed prints were collected in a sterile dish exposed to the UV while other test parts were printed. As a result, the UV doses were variable but substantial. Enrichment experiments To assess the potential for contamination after printing, 10 mm diameter hollow cylinders were printed under a variety of conditions and incubated in several types of liquid media at different temperatures and under aerobic, microaerophilic and anaerobic conditions. Initially, cylinders from UC Davis and MSU were grown in lysogeny broth (LB) for 96 h. No growth was observed in the experimental tubes or in the negative control, but high turbidity was observed in the positive control. Throughout this study, positive controls were prepared by dropping cylinders onto the laboratory floor followed by retrieval using ungloved hands from underneath the refrigerator or a similarly inconvenient location. After these initial experiments indicated no growth on the 3D printed parts, several more cylinders were printed. At UC Davis, test parts were printed onto flame-treated aluminum foil and transferred into conical tubes using flamed forceps. One group of cylinders was printed while the printer was situated on an open lab bench, a second group was printed in a laminar flow hood, and a third group was printed in a laminar flow hood under a UV lamp. In the process, an ample supply of positive controls were created inadvertently. At Michigan State, test parts were printed onto an ethanol-treated build platform. Further growth assays were conducted with each group of cylinders in LB, nutrient-rich ACES-buffered yeast extract (AYE) ( Feeley et al. , 1979 ) and Terrific Broth in aerobic conditions at 37 °C and 30 °C, revealing no growth from UV treated parts up to seven days post-inoculation ( Table 1 ). Growth was observed with one non-UV treated part at 96 h, which was determined to be contaminated with flora typical of human skin via selective plating and light microscopy. This was likely due to a handling mistake after printing. See ‘Identification of contaminating organisms’ for methods of identification. 10. 7717/peerj. 2661/table-1 Table 1 Summary of experiments conducted. Experiment Material Part Media Δt °C Oxy. Fab. Cult. Repl. Result Preliminary ( Neches, 2014a ; Neches, 2014d ; Neches, 2014e ) Orange PLA blob LB 96 h 37 + UCD UCD 1 – First trial ( Neches, 2014b ; Neches, 2014c ) Orange PLA tube LB 96 h 37 + UCD UCD 6 – Small vessel ( Neches, 2014f ) Orange PLA vessel LB 96 h 37 + UCD UCD 1 – Terrific Broth ( Flynn, 2014a ; Flynn, 2014b ) Orange PLA tube TB 96 h 37 + UCD UM 2 – AYE Broth ( Flynn, 2014b ) Blue PLA tube AYE 96 h 37 + MSU UM 1 + (error?) First MSU trial ( Zaman, 2014c ; Zaman, 2014b ) Blue PLA tube LB 96 h 30 + MSU MSU 3 – Filament trial ( Zaman, 2014e ) Blue PLA filament LB 96 h 30 + MSU MSU 4 – AYE Broth 2 ( Flynn, 2014c ) Orange PLA tube AYE 48 h 37 + UCD UM 2 – Cell culture Orange PLA tube RPMI-1640 6 d 37 + UCD MSU 1 – Swimmer plate ( Zaman, 2014a ) Blue PLA track plate Soft LB agar n/a 37 + MSU MSU 1 n/a Meat Broth, Anaerobic ( Flynn, 2014d ) Orange & Blue PLA tube Meat broth 2w 37 – UCD & MSU UM 2 + growth in non-UV at 2w Swimmer plate, redesign ( Zaman, 2014d ) Blue PLA 3-track round plate soft LB agar 48 h 37 + MSU MSU 1 + (handling error) Printed on blue tape cleaned with etoh ( Lewis, 2014b ) Orange PLA tube RCM 7 d 37 – UCD UCD, Mills Lab 2 – To test for the presence of anaerobic organisms, parts printed with and without UV were incubated in anaerobic conditions at 37 °C using two growth media, AYE and a custom chopped meat broth (CM Broth) ( Hehemann et al. , 2012 ). After seven days, no growth was observed in any tube except the positive control. After 14 days, a tube containing a sample that had been printed without UV became turbid. The positive control and cells incubated from the non-UV treated part were analyzed via 16S rRNA sequencing and found to contain bacteria associated with human skin ( Supplemental Information 1 ). The germinants sodium taurocholate and glycine, known to germinate Clostridium difficile and some Bacillus spores, respectively, ( Wilson, Kennedy & Fekety, 1982 ; Sorg & Sonenshein, 2008 ) were added to Brain Heart Infusion (BHI) medium and incubated anaerobically with 3D printed parts for 28 days at 37 °C. Microscopy and plating revealed no germination of these types of spores at weekly examinations. Cell culture experiments Sterile cell culture is a requirement for a variety of biological research applications. The biocompatibility of PLA and PLA-copolymers has been studied in vitro since at least 1975 ( Schwope, Wise & Howes, 1975 ) and in vivo since at least 1966 ( Kulkarni et al. , 1966 ). These materials have been used for sutures and surgical implants in humans since at least 1974 ( Horton et al. , 1974 ). More recently, there has been a shift towards using 3D printed scaffolds in combination with cell culture for tissue engineering ( Bose, Vahabzadeh & Bandyopadhyay, 2013 ). If the 3D printing process is sufficient to create sterile scaffolds, researchers could create useful scaffolds without damaging them with heat, steam, radiation or chemical sterilization programs. We performed a simple assay to ascertain if 3D printed parts are sterile under cell culture conditions. 3D printed parts that had been printed either with or without UV treatment were cultured with bone marrow-derived mouse macrophages for six days. Contamination was assessed by plating on LB and charcoal yeast extract thymidine (CYET) agar plates and examination under light microscopy. No evidence of contamination was found either in cells alone ( Fig. 3A ) or cells cultured with parts either printed with ( Fig. 3A ) or without UV ( Fig. 3A ) as judged by growth on agar plates and microscopy. Cell morphology and growth rate appeared to resemble the control cells grown in the absence of a 3D printed part and no visible contaminants were observed. Additionally, cells grown in the presence of 3D printed parts were competent for infection by Legionella pneumophila (data not shown). Cells appeared to grow normally immediately adjacent to the part, though the opacity of the 3D printed part prevented inspection for growth directly on the printed surface. Thus, 3D printed parts do not appear to contaminate or affect the growth of bone marrow-derived macrophages under these conditions. 10. 7717/peerj. 2661/fig-3 Figure 3 Mouse macrophage growth in the presence of 3D printed parts. Macrophages derived from mouse bone-marrow after incubation with 3D printed parts that had been treated with UV (A), without UV treatment (B), and treated with UV after handling and before incubation (C) and a control set of cells grown without 3D parts (D). Photos representative of three replicates in two independent experiments. Cell size, morphology and confluency were determined to be consistent across all experimental groups. Motility assay To demonstrate the utility of directly 3D printing sterile labware, we designed a simple four-well plate that could be used to assay bacterial motility ( Lewis, 2014a ). Each well is 70 mm long and 10. 3 mm wide, and holds approximately 2. 5 mL of liquid media. The four-well plates were removed from the build platform by gloved hand, and were kept sterile in an empty 100 mm Petri dish. We filled each well with 2 mL of 0. 2% w/v LB agar and allowed them to solidify for approximately 20 min. Then, we spotted 2 µl of a bacterial culture that had incubated for twelve hours into three lanes, and left the fourth as a control for contamination. In this experiment, we used three bacterial strains, JW1183, BW25113, and REL606. The first two are from the Keio Collection of single-gene knockouts, and REL606 is an E. coli B strain that was used to initiate the E. coli Long-term Experimental Evolution Project ( Lenski et al. , 1991 ); JW1183 is a ycgR deletion, and BW25113 is the ancestral strain of the Keio Collection ( Baba et al. , 2006 ). The choice of the ycgR knockout was suggested by Chris Watters as a potential bacterial “superswimmer” ( Waters, 2014 ). Indeed, using this 3D printed plate, we were able to identify a strong swimming phenotype of the ycgR mutant ( Fig. 4 ). Contamination was not observed in the control wells from several plates printed on painter’s tape, abraded foil, or abraded and flamed foil that was used to wrap the part after printing and stored overnight before use. These results demonstrate that direct 3D printing of sterile parts is a viable and useful approach for applications that may require a non-standard part. 10. 7717/peerj. 2661/fig-4 Figure 4 A 3D printed motility assay device. A custom device for a motility assay fabricated using 3D printing. The device was found to be sterile without autoclaveing if contamination during post-fabrication handling is avoided. Materials and Methods Preliminary experiment A sterile glass beaker containing roughly 20mL of LB media was placed under the nozzle of a fused deposition modeling (FDM) 3D printer. The nozzle was heated to 220 °C, and the extruder drive motor was driven forward until about 20 mm of polylactide (PLA) filament had been melted and expelled through the nozzle and into the beaker. A tangle of molten and cooled PLA detached from the nozzle and fell into the beaker. The mouth of the beaker was then covered with sterile aluminum foil. An unopened sterile beaker of LB was prepared as the negative control. A positive control was prepared with a length of un-melted PLA filament from the spool. The three beakers were placed into a shaking incubator at 37 °C for 96 h. The experiment, the progress and the reuslt were announced in real-time on Twitter to generate feedback and suggestions from the community, which sparked collaboration described in this paper ( Neches, 2014a ; Neches, 2014d ; Neches, 2014e ; Neches, 2014b ; Neches, 2014c ). No growth was observed in the negative control or the beaker inoculated with extruded material, and robust growth was observed in the positive control. Experimental setup and results were posted on Twitter as they occurred. 3D printing The preliminary experiment seemed to indicate a potentially useful killing effect from the nozzle’s heat and pressure, and so a slightly more realistic assay was conducted. A simple model was created using the OpenSCAD ( Kintel & Wolf, 2000–2004 ) modeling language consisting of a cylinder of radius 4 mm and height 10 mm ( Fig. 5 ). 10. 7717/peerj. 2661/fig-5 Figure 5 Design and fabrication of test parts. A very simple model of a cylinder was created in OpenSCAD and exported in STL format. (A) The G-code toolpath visualization of test part in Cura. The slicing engine was set to a 0. 4 mm wall width (equal to the diameter of the nozzle), cooling fans inactive, no infill, a top and bottom layer height of zero, and a spiralized “Joris Mode” outer wall. (B) Test parts were then 3D printed on abraided and flamed aluminum foil at 220 °C with a feed rate of 50 mm/s. cylinder( r=4, h=10 ); The model was exported in Standard Tessellation Language (STL) format ( Burns, 1993 ). The manifold was then converted into G-code commands ( Thomas, Frederick & Elena, 2000 ) using Cura (version 13. 12-test on Linux), using a wall width of 0. 4 mm (equal to the nozzle diameter), cooling fans inactive, no infill, a top and bottom layer height of zero, and a spiralized outer wall (“Joris mode, ” after Joris van Tubergen) to produce a small, open tube. The G-code was stored on a SD card and printed on an Ultimaker kit-based FDM 3D printer (standard, current firmware builds distributed by Ultimaker were used). A small patch of aluminum foil was lightly abraded with fine-grit sandpaper to improve surface adhesion properties, and flamed over a Bunsen burner until signs of melting appeared. The foil patch was then affixed to the build platform, so that the build area indicated in the G-code toolpath would be entirely within the untouched center of the patch. The G-code toolpath was also examined to insure that the nozzle would contact no surface except the build area on the foil. Printing was then initiated with a feed rate of 50 mm/s at 220 °C. Once printing was complete, finished parts were immediately removed from the build area using flamed forceps and transferred to culture tubes or conical tubes for storage and shipping. Independent reproduction of growth experiment on printed component The experiment described in ‘3D printing’ was replicated at Michigan State University on a kit-built Ultimaker 3D printer modified with an E3D all-metal hot-end with a 0. 4 mm nozzle. A cylinder was designed using OpenSCAD with a radius of 4 mm and a height of 12 mm. The model was exported in STL format and sliced with Cura SteamEngine 13. 12. The cylinder was printed with a wall thickness of 0. 4 mm, a feed-rate of 10 mm/second (the effective speed with the minumum layer cooling time set to 5 s), and a nozzle temperature of 225 °C. The print surface was prepared with 3M Scotch Blue painters tape, and was lightly wiped with ethanol before printing began. Two printed cylinders were transfered to sterile glass tubes filled with 4 mL of LB media with flamed tweezers. A fragment of unused filament was used as a positive control, and an uninoculated tube was used as a negative control. Tubes were transfered to a shaking incubator set at 30 °C. No growth was observed after 24 h in any of the tubes with printed parts, while the unused filament contaminated the media. After two days, another cylinder was printed and incubated in LB broth. Again, after 24 h no growth was observed. None of the tubes with printed parts showed signs of growth after 96 h ( Fig. 6 ). 10. 7717/peerj. 2661/fig-6 Figure 6 Independent replication of results. After 48 h, only the positive control (A) was contaminated. Printed cylinders in LB did not appear to contaminate the media (C and D). Tube (B) contains uninoculated media. Terrific broth experiments Printed cylinders from UCD and MSU were dropped into glass culture tubes with 3 mL of AYE or TB broth in independent experiments and transferred to a roller in a 37 °C warm room. After 96 h, one of the “no UV” tubes in AYE broth became turbid with a mixed population of bacterial growth as examined by microscopy and plating on CYET agar ( Fig. 7 ). Repeated experiments did not yield growth for these parts, and so the contaminaiton was likely due to a handling error. No growth was observed for any parts grown in TB ( Fig. 8 ). 10. 7717/peerj. 2661/fig-7 Figure 7 Test on solid media. A total of 10 µl of each AYE tube (positive control, PLA plastic and negative control) was struck out on Charcoal Yeast Extract solid media and incubated at 37 °C to grow for 24 h. Growth revealed that the PLA test part (A) appeared to contain a different bacterial species than the positive control tube (B). Media from the negative control part was also plated (C). Using a light microscope, both bacterial growths appear coccoid, with the yellow colonies forming clumps more often. Experiment was repeated with parts from UC Davis and Michigan State, plus controls. Contamination was not observed. Meat broth experiments Test parts were incubated for two weeks under anaerobic conditions at 37 °C in chopped meat broth (CM Broth) ( Hehemann et al. , 2012 ). A non-UV treated test part fabricated at UC Davis exhibited evidence of growth ( Fig. 9 ). The contaminated media was plated on BHI+blood media and allowed to grow overnight (see Fig. 7 ), and 16S rRNA sequencing was performed on resulting colonies (see ‘Identification of contaminating organisms’). 10. 7717/peerj. 2661/fig-8 Figure 8 Test in Terrific Broth. 3D printed parts from UC Davis with and without UV treatment (B and C) were suspended in sterile Terrific Broth supplemented with potassium salts, along with a negative control (A) and a positive control (D). After 24 h at 37 °C, no growth was observed for parts treated with UV. Tube (F) contains a part from Michigan State after 48 h of incubation in Terrific Broth, along with negative and positive controls (E and G). No growth was observed 96 hours after inoculation. 10. 7717/peerj. 2661/fig-9 Figure 9 Test in anerobic conditions. After two weeks in anaerobic chamber at 37 °C in “meat broth, ” a non-UV treated part from UC Davis exhibited evidence of growth. All other parts were limpid, aside from the positive control. Contaminated media was plated on BHI+blood agar overnight (see Fig. 7 ), and 16S rRNA sequencing was performed on resulting colonies. Cell culture Sterility of 3D printed parts was assessed by incubating each part with bone marrow-derived macrophages from femurs of C57BL/6 mice (Jackson Laboratories) cultured in RPMI-1640 containing 10% heat-inactivated fetal bovine serum (FBS) (Gibco) ( Swanson & Isberg, 1996 ). Microscopy was performed by culturing macrophages in plastic dishes with 3D parts for 6 days after the initial isolation from bone marrow in L-cell conditioned media and examining under light microscope. The University Committee on Use and Care of Animals approved all experiments conducted in this study (principal investigator Michele Swanson; protocol reference number PRO00005100). Identification of contaminating organisms Contaminated media (see ‘Meat broth experiments’) was streaked onto BHI agar plates ( Fig. 7 ) supplemented with 10% defibrinated horse blood (Quad Five, Catalog No. 210; Ryegate, MT, USA) for colony isolation. Bacterial colonies with unique morphologies were picked into chopped meat broth and genomic DNA was extracted from a 1 mL cell pellet using Phenol:Chloroform and ethanol precipitation after bead beating. Nearly full length 16S rDNA was amplified using primers 8F and 1492R ( Eden, Schmidt & Blakemore, 1991 ) and run on a 1% agarose gel to confirm amplification and size. PCR products were purified using the Qiagen MinElute PCR purification kit (Catalog No. 28006), quantified and bidirectional sequenced at the University of Michigan DNA Sequencing Core. Sequencing reads were analyzed using the DNASTAR Lasergene software suite (DNASTAR, Inc. , Madison, WI, USA). Results were used to search the nr database ( Pruitt, Tatusova & Maglott, 2005 ) to using NCBI’s BLAST online search tool determine the closest relatives. A total of three unique bacterial colonies were analyzed; two from a positive control and one from a non-UV treated 3D printed part. All three were 99% similar to their closest database hit, and found to be common skin associated microflora. The positive control yielded sequences related to Staphylococcus epidermidis and Propionibacterium acnes. Similarly, the non-UV treated 3D printed part was also a Propionibacterium acnes indicating that the bacteria present were likely introduced to the 3D parts post printing. Bacterial strains, culture conditions and reagents For AYE growth experiments, 3D parts were cultured on a rolling spinner at 37 °C in N-(2-acetamido)-2- aminoethanesulfonic acid (ACES; Sigma)-buffered yeast extract (AYE) broth supplemented with 100 µg/mL thymidine (Sigma) ( Feeley et al. , 1979 ). Terrific Broth (TB) experiments were conducted on a rolling spinner at 37 °C in media containing yeast extract, tryptone and glycerol supplemented with 0. 17 M KH2PO4, 0. 72 M K2HPO4. Chopped Meat Broth and BHI-Blood agar experiments were performed in a Coy anaerobic chamber (Grass Lake, MI) at 37 °C ( Hehemann et al. , 2012 ). Anaerobic experiments were performed in anaerobic chambers from Coy Laboratories (Grass Lake, MI) in Brain-Heart Infusion broth supplemented with yeast extract (5 g/L). 0. 1% cysteine and 0. 1% taurocholate were added as germinants. 3D printed parts from UC Davis The following materials were prepared in Jonathan Eisen’s laboratory at the UC Davis Genome Center and shipped to Michele Swanson’s laboratory at the University of Michigan. All printed parts were printed using Printbl Orange 3 mm PLA filament at 220 °C with a feed rate of 50 mm/s, using the same G-code files described in ‘3D printing’, and placed into sixteen 50 mL conical tubes using flamed forceps. The contents of the conical tubes was as follows : • Test objects, printed under biosafety hood (10×) • Test objects, printed under biosafety hood with UV (10×) • Test objects, printed under biosafety hood with UV, then dropped onto no-sterile surface during handling (2×) • Test object, printed under biosafety hood with UV (1×) • Test object, printed under biosafety hood without UV, dropped during handling (1×) • Empty, unopened conical tube • Test object, printed under biosafety hood without UV (1×) • Test object, printed under biosafety hood without UV (1×) • Test object, printed under biosafety hood with UV (1×) • Test object, printed under biosafety hood with UV (1×) • Test object, printed under biosafety hood with UV, handled with ungloved hands (1×) • Test objects, printed on open bench and left on lab bench overnight (2×) • Unused Printbl Orange 3 mm PLA filament (3×) • Unused Laywoo-D3 cherrywood 3 mm printable wood filament (3×) • Unused Protoparadigm White 3 mm PLA filament (3×) • Unused Printbl Crystal Blue 3 mm PLA filament (3×). 3D printed parts from Michigan State University Several printed parts were prepared at Michigan State University and sent to the Michele Swanson lab at the University of Michigan. Cylinders were printed using the same G-code and parameters described in ‘3D printing. ’ All printed parts from Michigan State University were printed using Ultimaker translucent blue PLA. Each part was removed from the printbed using flamed forceps and transferred to a sterile 15 mL plastic tube. The contents of the tubes was as follows: • Test objects, printed on blue painters tape wiped down with ethanol (3×) • Test objects, printed on abraded foil wiped with ethanol and flamed (3×) • Unused Ultimaker translucent blue PLA filament (3×). 3D printing systems and materials The 3D printing systems and materials used in this study are relatively inexpensive and available to the public. While it is likely that nearly any 3D printer that uses thermoplastic extrusion will perform similarly for these purposes, the exact devices and materials used in this study are available from the following suppliers: • Ultimaker Original with v3 hot-end (UC Davis). https://www. ultimaker. com/pages/our-printers/ultimaker-original • Ultimaker Original modified with a E3D hot-end (Michigan State University). http://e3d-online. com/ • PLA (Poly-Lactic-Acid) filament, Blue-Translucent, 0. 75 kg. 2. 85 mm diameter. https://www. ultimaker. com/products/pla-blue-translucent • PLA filament, Orange, 1. 0 kg 3 mm diameter (2. 85 actual). http://diamondage. co. nz/product/pla-standard-colours. Discussion This work was inspired by the observation that, while most 3D printed products cannot be autoclaved, the extrusion temperatures typically used in 3D printing are significantly higher than temperatures used in most autoclave cycles. This led us to wonder if 3D printing is an intrinsically sterile process. Sterility is a difficult property to judge due to the impossibility of proving a negative. In the experiments we have presented here, we endeavored to create advantageous conditions for growth for a reasonably wide range of organisms, and particularly organisms likely to be problematic for experiments in clinical microbiology, cell culture and molecular biology. We used the “richest” rich media available to us, and attempted to induce germination of spores under aerobic and anaerobic conditions. Of course, this is not exhaustive, and the culturing conditions used would not detect the presence of (for example) Sulfolobus or Methanococcus maripaludis. We did not perform culture-independent sampling, which would be of obvious interest. However, as a practical matter, we find that the printing process does indeed produce functionally sterile parts which should be suitable a wide variety of experiments. While 3D printing is likely not the ideal method for producing all labware under all circumstances, there are nevertheless a wide variety of applications and settings in which the ability to produce small batches of sterile parts would be extremely useful. The ability to manufacture sterile parts on premises during extended fieldwork in remote locations can reduce logistical risks. Schools can print materials for student laboratory projects. Researchers in developing countries can reduce their reliance on costly imported disposable labware. Otherwise well-equipped laboratories can more cheaply obtain fully custom sterile components. Our experiments indicated that there are several reasonable approaches to sterile technique, though we did not attempt to establish which among them is optimal. We anticipated much higher rates of contamination than were actually observed. In more than twenty incubations, we found only two contaminated parts. Based on plating, light microscopy and 16S rRNA sequence obtained from the culture and on the fact that other parts prepared in the same way failed to produce growth, it is likely that the part was contaminated after printing. These experiments are not intended to establish a quantitative measure of the rate of contamination characteristic of the process, but rather to demonstrate that sterile parts can be produced by direct 3D printing of non-sterile thermoplastic feedstock. Future work While fused deposition modeling printers are by far the most common, widely available and inexpensive printers at present, there are several other 3D printing technologies. For example, there are a number of technologies based on materials that undergo photopolymerization. We happened to have two machines available to us that use photopolymerization, an Objet Eden 260, which uses an inkjet-like print head and a UV lamp, and a Formlabs Form 1, which uses stereolithography. We performed a variation of our preliminary experiment using cylinders printed using these machines, and found they were also able produce sterile parts ( Fig. 10 ). 10. 7717/peerj. 2661/fig-10 Figure 10 Test with alternative 3D printing technology (Objet Eden 260). A group of cylinders were printed on an Objet Eden 260. 24 cylinders were transfered directly from the printing plate to culture tubes by scraping them from the build plate with the open tube. Two cylinders were removed with an ungloved hand to act as positive controls. Tubes were incubated for 96 h at 37 °C in LB media, revealing one contaminated tube. The mechanism of sterilization in for these technologies is likely to be very different from FDM devices. It is possible that cells are destroyed by radiation; the Objet machine repeatedly exposes the build surface to intense UV radiation, and the Form 1 uses a 120 mW, 405 nm (violet) laser. However, the more likely killing mechanism is chemical, as the cross-linking chemistry of many photopolymerization systems is driven by high concentrations of free radicals. Unfortunately, the chemical composition of the input material and the precise nature of the reactions is proprietary. Formlabs was kind enough to point us to the catalog of their supplier of raw materials, but we were not able to deduce the chemistry of their system from this information alone. It is our hope that researchers more familiar with these polymer systems will take up this question, and perhaps design materials for these printers that can be certified for manufacturing sterile parts. Supplemental Information 10. 7717/peerj. 2661/supp-1 Supplemental Information 1 16S Sanger sequence data for contaminating organisms Click here for additional data file. 10. 7717/peerj. 2661/supp-2 Table S1 Temperature and duration of various sterilization processes The table from Fig. 1 is provided here to make it easier to access the data within the table. Click here for additional data file.

10. 7717/peerj. 2821

PeerJ

Background Electrical stimulation (ES) has been successfully used to treat bone defects clinically. Recently, both cellular and molecular approaches have demonstrated that ES can change cell behavior such as migration, proliferation and differentiation. Methods In the present study we exposed rat bone marrow- (BM-) and adipose tissue- (AT-) derived mesenchymal stem cells (MSCs) to direct current electrical stimulation (DC ES) and assessed temporal changes in osteogenic differentiation. We applied 100 mV/mm of DC ES for 1 h per day for three, seven and 14 days to cells cultivated in osteogenic differentiation medium and assessed viability and calcium deposition at the different time points. In addition, expression of osteogenic genes, Runx2, Osteopontin, and Col1A2 was assessed in BM- and AT-derived MSCs at the different time points. Results Results showed that ES changed osteogenic gene expression patterns in both BM- and AT-MSCs, and these changes differed between the two groups. In BM-MSCs, ES caused a significant increase in mRNA levels of Runx2, Osteopontin and Col1A2 at day 7, while in AT-MSCs, the increase in Runx2 and Osteopontin expression were observed after 14 days of ES. Discussion This study shows that rat bone marrow- and adipose tissue-derived stem cells react differently to electrical stimuli, an observation that could be important for application of electrical stimulation in tissue engineering.

Introduction Large segment bone defects, caused by open fractures, non-unions, infections and tumor resection are a major challenge in trauma and orthopedic surgery. Complications associated with current treatments and the projected increase in the number of cases due to ageing populations in developed countries give urgency to the search for alternative treatments ( Amini, Laurencin & Nukavarapu, 2012 ). Tissue engineering (TE) approaches that deliver osteoprogenitor cells with osteoconductive scaffolds directly into these large defects hold great potential for achieving optimal bone healing while eliminating the associated drawbacks of conventional treatments ( Petite et al. , 2000 ). Mesenchymal stem cells (MSCs) have been shown to be an attractive cell source for clinical bone tissue engineering applications. MSCs possess a great capacity for self-renewal and multi-lineage differentiation, including osteogenic differentiation. Several in vitro, in vivo preclinical and clinical studies have shown that MSCs are able to facilitate bone mineralization ( Colnot, 2011 ; Gamie et al. , 2014 ). While the main source for harvesting MSCs has been primarily bone marrow (BM), they have also been isolated from other tissues such as adipose tissue (AT) ( Grottkau & Lin, 2013 ). MSCs harvested from different sources have been reported to exhibit different proliferation and differentiation characteristics ( Musina et al. , 2006 ). Adipose tissue derived MSCs (AT-MSCs) are able to proliferate rapidly in culture, and facilitate expansion into large numbers of cells required for clinical effectiveness ( Cowan et al. , 2004 ). Recent studies confirm that AT-MSCs do differentiate into osteogenic lineage in vitro ( Bunnell et al. , 2008 ; De Girolamo et al. , 2007 ; Pelto et al. , 2013 ). However, a comparative study between AT-MSCs and BM-MSCs revealed that AT-MSCs have more capacity to proliferate and slightly less capacity for osteogenic differentiation. In addition, AT-MSCs exhibit high tolerance to serum deprivation-induced apoptosis in comparison with BM-MSCs ( Peng et al. , 2008 ). Although AT-MSCs are able to form osteoid matrix and regenerate bone in vivo, their capacity to differentiate into hematopoietic marrow is still in question ( Robey, 2011 ). Electrical stimulation (ES) has been shown to be effective in nerve and cardiac tissue engineering applications, primarily due to the electric nature of these tissues ( Ghasemi-Mobarakeh et al. , 2011 ; Tandon et al. , 2009a ). The long recognized piezoelectric characteristics of bone (electricity resulting from mechanical pressure), together with the known links between ES and bone growth, make bone an attractive target tissue for investigating the role of ES in bone regeneration and healing ( Tofail, Zhang & Gandhi, 2011 ). Electrical stimulation has been successfully used to treat bone defects clinically for more than 40 years ( Ryaby, 1998 ). Several common modes of electrical stimulation, such as pulsed electromagnetic fields, capacitive coupling and direct current (DC), have been used both experimentally and clinically to promote bone healing in different orthopedic applications ( Griffin & Bayat, 2011 ). While the mechanism by which ES promotes bone healing is still poorly understood, recent in vitro studies show that bioelectrical signals play a key role in cellular pathways involved in healing ( Levin, 2009 ). Recently, in vitro studies showed that DC ES, acting partially via electrochemical reaction at the cathode, alters several MSC behaviors, such as migration, proliferation, and differentiation ( Hardy et al. , 2015 ; Tandon et al. , 2009b ; Tsai et al. , 2009 ; Zhao et al. , 2011 ). Specifically, correlations between ES and the rate of osteogenic differentiation have been reported ( Fukada & Yasuda, 1957 ; Shamos, Lavine & Shamos, 1963 ). Balint et al. (2013), exposed BM-MSCs to pulsed DC ES and reported significant alterations in the expression of osteogenic marker genes. Hammerick et al. (2010), exposed AT-MSCs to pulsed DC ES and observed an increase of osteoblast-specific markers, including Runx2, Osteopontin and Type I Collagen. Recently, various different types of physical stimuli such as static magnetic fields, cyclic strain, low frequency vibration, and electric signals have been used to improve both the proliferative and the differentiation potential of stem cells. Specifically, ES has been shown to influence cell proliferation and differentiation in tissue engineering applications ( Balint, Cassidy & Cartmell, 2013 ; Marycz et al. , 2016 ). In our laboratory we study the possibility of combining ES with tissue-engineering methods to see if the combination can provide benefits not previously seen in either approach individually. We investigated the changes in expression of a few osteogenic differentiation key markers; Runx2, Osteopontin and Col1A2. Runt-related transcription factor-2, which is known as Runx2 (Cbfa1/PEBP2αA/AML-3/Osf2), is often referred to as the master switch of osteogenic differentiation. It interacts with the promoter regions of the key osteoblast specific genes such as osteocalcin, osteopontin, collagen I, bone sialoprotein, alkaline phosphatase and TGFβ receptor 1 ( Nakamura et al. , 2009 ). Secreted phosphoprotein 1 which is known as Osteopontin (OPN) has important role as a regulator of cytoskeleton dynamics and gene expression. In bone remodeling, up-regulation of expression of OPN has been observed. OPN is also reported to up-regulate during the osteogenic differentiation of MSCs. Collagen I is an important component of bone extra-cellular matrix. Col1A2 up-regulation has been observed as an early response to a number of different methods of inducing in vitro osteogenic differentiation ( Florencio-Silva et al. , 2015 ). In the present study we exposed rat BM-MSCs and AT-MSCs to DC ES and compared osteogenic differentiation behavior in both cell types. Materials and Methods Groups We designed our experiments to compare electrically stimulated (ES) versus not stimulated (Control) cell groups. Each group included both BM- and AT-derived rat MSCs, cultivated in osteogenic differentiation supplemented medium, harvested at 3 different time points (three, seven and 14 days) for three different analyses (Calcium deposition staining, viability test and osteogenic gene expression). All experiments were run in triplicate. Cell preparation and culture Sprague-Dawley (SD) rat MSC from bone marrow (RASMX-01001) and adipose tissue (RASMD-01001) were both obtained from Cyagen (CA, USA). Frozen vials of cells were thawed, cultured, and expanded to reach the desired number, based on the cell provider’s instructions. To achieve the appropriate number, cells were cultured at a density of 2. 5 × 10 4 cell/cm 2 until 80% confluency and then expanded over five passages. Cells from passage 6 were seeded in 6-well cell culture plates (TPP, Trasadingen, Switzerland) at a density of 10 4 cell/cm 2 in cell growth medium consisting of Dulbecco’s Modified Eagle Medium (DMEM) + GlutaMAX + 1 g/L D-Glucose + 10% Fetal Calf Serum (FCS) and 1% Penicillin/Streptomycin (10. 000 U/ml) all obtained from (Gibco ®, Gaithersburg, MD, USA), placed in a humidified incubator at 37 °C, 5% CO 2 and 3% Oxygen ( Haque et al. , 2013 ). The culture medium was changed initially, one day after seeding and then twice weekly. From the second day on, the cell growth medium was supplemented with 10 −7 M dexamethasone, 10 mM β-glycerophosphate, and 0. 05 mM ascorbic acid-2-phosphate, all obtained from Sigma-Aldrich (Heidelberg, Germany) ( Jaiswal et al. , 1997 ). Electrical stimulation of cells Electrical stimulation was applied by means of a purpose built DC ES cell culture chamber ( Mobini, Leppik & Barker, 2016 ). Briefly, the chamber consists of L-shaped platinum electrodes, separated by a distance of 22 mm, and secured to the lid of 6-well cell culture plates, (TPP, Trasadingen, Switzerland), and connected to a standard electrical power supply (e. g. , Triple Output Programmable DC Power Supplier (Supply-Model 9130; B&K Precision, Yorba Linda, CA, USA)), Fig. 1. For sterilization, electrodes were submerged in 70% ethanol for 10 min and washed by sterile calcium-magnesium free phosphate buffer saline (PBS) (Gibco ®, Gaithersburg, MD, USA) and finally exposed to UV light overnight. The cells that received electrical stimulation were exposed to 100 mV/mm of DC ES for 1 h per day. All evaluations and assays were performed 24 h after the last exposure. 10. 7717/peerj. 2821/fig-1 Figure 1 Setup for delivering direct current electrical stimulation to the cells. L-shaped platinum electrodes, 22 mm apart, secured to the lid of a 6-well cell culture plate and connected to a standard DC power supply. The electrodes are in contact with the bottom of the cell culture plate and are fully covered by culture medium. Cell viability and activity To confirm that the oxidation–reduction and electrochemical reactions of the metallic electrodes in the DC ES chamber were not cytotoxic, cell viability and metabolic activity were assessed by 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay (Kit I MTT; Roche Diagnostics, Manheim, Germany) after being exposed to ES for 7 and 14 days. Cell culture medium was completely removed and cells were washed twice with 1 mL PBS. One ml of fresh culture medium was added to the wells immediately prior to running the test. Absorbance was read at 550 nm with the reference wavelength of 650 nm, using an Infinite 200PRO NanoQuant device, with TECAN i-control™ software (Tecan, Crailsheim, Germany). We calculated fold change in optical density of electrically stimulated groups, relative to control. Osteogenic differentiation Alizarin Red stains calcium deposits in the cells, indicating the presence of functional osteocytes. Cultured cells were washed twice with PBS and fixed with 4% paraformaldehyde (Sigma Aldrich, München, Germany) solution in PBS for 30 min. Alizarin Red S (Sigma-Aldrich, München, Germany) solution (2% in PBS) was added to the fixed cells, incubated at room temperature for 30 min, and rinsed with deionized water repeatedly. Images were captured with a light microscope (CKX53, cellSens Entry 1. 9 Software; Olympus, Tokyo, Japan) at a magnification of 10X. Temporal osteogenic marker expression was evaluated by two-step Reverse-Transcription quantitative Polymerase Chain Reaction (RT-qPCR) technique. In brief, total RNA was isolated using an Aurum RNA isolation kit (BioRad, München, Germany) according to the manufacturer’s instructions. The quality and quantity of RNA were measured using gel electrophoresis and an Infinite 200PRO NanoQuant device (Tecan, München, Germany). Genomic DNA contamination was removed through digestion using RNase-free DNase-I following the manufacture’s protocol (New England BioLabs GmbH, Germany). DNase-treated RNA samples were reverse transcribed using iScript Select cDNA Synthesis Kit (Bio-Rad, München, Germany) according to the manufacturer’s instructions. The RT-PCR reaction was performed using cDNA equivalent to 1. 25 ng RNA and the SsoAdvanced Universal SYBR Green Supermix (BioRad, München, Germany). All samples were amplified in triplicate and PCR was performed using a CFX96 Touch Real Time PCR Detection System (BioRad, München, Germany) under the following conditions: 95 °C for 3 min followed by 40 cycles of 10 s of denaturation at 95 °C and 30 s of annealing and elongation at primer corresponding temperature. The rat gene specific primers used in this study were: Runt-related transcription factor 2 (RunX2, 60 °C, forward CTACTCTGCCGAGCTACGAAAT; reverse TCTGTCTGTGCCTTCTTGGTTC), Osteopontin (SPP1, 62 °C forward GATGAACAGTATCCCGATGCC; reverse TCCAGCTGACTTGACTCATGG) and Collagen type I, alpha 2 protein (Col1A2, 62 °C, forward TTCCCGGTGAATTCGGTCT; reverse ACCTCGGATTCCAA TAGGACCAG ), all purchased from Sigma Aldrich (Sigma Aldrich, München, Germany). Ribosomal protein P1 (RPLP1, 64 °C forward GCATCTACTCCGCCCTCATC, reverse AAGCCAGGCCAGAAAGGTTC) was used as a reference gene ( Curtis et al. , 2010 ). A melting curve analysis was applied to ensure the specificity of the PCR; amplification products were also analyzed by gel-electrophoresis. Data analysis All experiments were performed in triplicate and statistical significance of differences between groups was analyzed by one-way ANOVA and student t -test using GraphPad Prism (GraphPad Software Inc, La Jolla, CA, USA). Significance level was set at p < 0. 05. Relative quantification of messenger RNA (mRNA) levels of the target genes was analyzed using the comparative C T (threshold cycle values) method (2 −ΔΔCq ) ( Livak & Schmittgen, 2001 ). The results are presented as relative quantification (RQ), which is expression fold change compared to the calibrator (Our calibrator was the non-stimulated cells, cultivated in growth medium at day 0). Standard deviation (SD) was calculated with the ΔC q value of technical triplicates. Results Electrical stimulation optimization In order to optimize the electrical stimulation regime, first we exposed cells in culture to 10, 50, 100, and 200 mV/mm of DC ES for 1 h and found that 200 mV/mm caused cell lysis due to electro-chemical reactions in the vicinity of the electrodes. We found that 10 and 50 mV/mm for 1 h caused no significant changes in proliferation and differentiation (see Figs. 2C and 2D ). Based on these preliminary studied we chose 100 mV/mm of DC ES for 1 h per day for the present experiment. 10. 7717/peerj. 2821/fig-2 Figure 2 Calcium deposition. Calcium deposition stained using Alizarin Red S for; (A) BM-MSCs and AT-MSCs exposed to no electrical stimulation (controls) at days 7 and 14; (B) BM-MSCs and AT-MSCs exposed to 100 mV/mm of electrical stimulation, at days 7 and 14; Different degrees of staining are visible in electrically stimulated vs. non-stimulated controls; (C) BM-MSCs exposed to 10 and 50 mV/mm of electrical stimulation at day 7; (D) AT-MSCs exposed to 10 and 50 mV/mm of electrical stimulation at day 7 (Magnification = 10×). Cell viability and activity MTT assay was performed to compare viability and activity of electrically stimulated cells vs. non-stimulated controls. None of the cells exposed to 10, 50 and 100 mV/mm showed signs of toxicity. Figure 3 shows changes in optical density of groups exposed to 100 mV/mm DC ES compared to controls, indicating differences in cell viability between the different sources in response to electrical stimulation. At day 7, electrically stimulated AT-MSCs had approximately 30 percent increase in cell activity, which can be correlated to higher cell numbers in ES vs. control groups. At this time point BM-MSCs had less cell activity/number in ES vs. control groups. At day 14, the cell number/activity of ES AT-MSCs was slightly higher than controls, while BM-MSCs maintained the same trend. 10. 7717/peerj. 2821/fig-3 Figure 3 Cell viability. Measured by MTT assay, compared between electrically stimulated and non-stimulated controls. No significant difference in cell viability was detected between ES and non-stimulated control AT- and BM-derived MSC at 7 and 14 days (Values are shown as means ± standard deviations ( n = 3) ** p < 0. 01). Osteogenic differentiation The influence of electrical stimulation on osteogenic differentiation of rat AT- and BM- MSCs in culture were investigated after seven and 14 days and compared to controls. Figures 2A and 2B shows Alizarin Red S stained calcium deposits in the electrically stimulated (100 mV/mm) cells, initiating at day 7 and becoming significant at day 14, versus controls. Exposure to ES caused distinct changes in morphology and calcium deposition in BM-MCSs at day 14 ( Fig. 2B ) compared to control cells ( Fig. 2A ). Calcium deposition in AT-MSCs and BM-MSCs after seven days of exposure to lower (10 and 50 mV/mm) electrical fields, was not significant. However, initial morphological changes in both cell types were present at day 7, in cells exposed to electrical field of 50 mV/mm ( Figs. 2C and 2D ). Osteogenic phenotype gene expression was investigated by means of RT-qPCR analysis, in both ES and control groups of AT-MSCs and BM-MSCs at three, seven, and 14 days, ( Fig. 4 ). In both BM- and AT- MSCs, time-dependent gene expression patterns significantly differed in ES vs. control cells. Moreover, there was a difference in expression of osteogenic markers (Runx2, Osteopontin and Col1A2) between the ES BM- and AT-MSCs groups. However, temporal mRNA expressions of the same markers were similar in both AT- and BM-MSC, non-stimulated controls. In BM-MSCs, ES resulted in an increase in the expression of Runx2, Osteopontin and Col1A2 mRNA, three, 1, 500 and 14 times, respectively, in comparison to the controls. In the ES AT-MSC groups, the intensification of mRNA levels, Runx2 (six times) and Ostepontin (25 times) was observed on day 14. However there was no difference in expression of collagen type I mRNA (Col1A2) in AT-MSCs in ES and non-stimulated controls. 10. 7717/peerj. 2821/fig-4 Figure 4 RT-qPCR results. Temporal changes in messenger RNA (mRNA) of (A) Runx2, (B) Osteopontin, (C) Collagen Type1 (Col1A2) in BM-MSCs; (D) Runx2, (E) Osteopontin and (F) Col1A2 in AT-MSCs, in both ES and non-stimulated control cells. Total RNA extracted from cultured cells at days 3, 7 and 14 were transcribed into complementary DNA and subjected to real-time quantitative polymerase chain reaction analysis. The relative mRNA levels are expressed as arbitrary units normalized according to the corresponding levels of Ribosomal Protein P1 mRNA (Values are shown as means ± standard deviations ( n = 3) * p < 0. 05; *** p < 0. 001). Discussion Early studies exposing bone cells to DC electrical fields were in 1980s when Ferrier et al. (1986) exposed osteoblast-like cells to a 100 mV/mm electrical stimulation and observed cell migration toward the cathode. In 1997 and 1998, changes in TGF-β, BMP-2, and 4 mRNA levels were demonstrated in osteoblasts using capacitive coupling and pulsed electromagnetic field techniques ( Bodamyali et al. , 1998 ; Zhuang et al. , 1997 ). Later, several groups described using both indirect (electromagnetic fields or capacitive coupling) and direct (DC) electrical stimulation to promote MSC osteogenic differentiation ( Fu et al. , 2014 ; Ross et al. , 2015 ; Sun et al. , 2009 ). For example, in 2010 McCullen et al. reported experiments in which they exposed human AT-MSCs, in osteogenic supplemented medium, to DC ES (100 mV/mm at 1 Hz for 4 h/day) and after 14 days observed a significant increase in mineralization. In another study, rat BM-MSCs were seeded onto conductive polypyrrole films and exposed to 35 mV/mm of DC ES for 4 h and after 14 days significant osteogenic differentiation was observed ( Hu et al. , 2014 ). With the aim of better understanding these DC ES induced changes in MSCs osteogenic differentiation behavior, we exposed rat BM-MSCs and AT-MSCs to 100 mV/mm 1 h/day DC ES for three, seven and 14 days in the presence of osteogenic differentiation supplements in culture medium. The rationale for choosing this stimulation regime was based on reports in the literature ( Zhao et al. , 2011 ) and our own previous experiments. In the present study we observed that in most cases ES caused significant changes in osteogenic marker expression patterns both in BM- and AT-derived MSCs when compared to non-stimulated controls. We also found that when BM- and AT-derived MSCs are exposed to DC ES they express osteogenic markers differently. Finally we observed that ES induced elevation of osteogenic markers, Runx2, Osteopontin and Col1A2 at day 7, only in BM-MSC. We showed that in the presence of DC ES, BM-MSCs and AT-MSCs behave differently as it relates to osteogenic marker expression. These observations could be related to the known differences in osteogenic differentiation capacity between BM- and AT-derived MSC. Namely, studies that suggest AT-MSCs have less osteogenic potential than BM-MSCs ( Ratanavaraporn et al. , 2009 ). Moreover, reports indicate that, while a given set of common genes may be involved in early differentiation of MSC from both sources, a different set of genes could be involved in maturation into fully differentiated cells ( Liu et al. , 2007 ). Our results indicate that in control groups (osteogenic supplemented medium, without electrical stimulation), both in BM- and AT-MSCs, only a slight difference exists in temporal expression patterns of Runx2, Osteopontin and Col1A2. However, in the presence of ES, no common expression pattern was detectable. This suggests that ES might activate different cellular mechanisms in BM- vs. AT- MSCs. Runx2 is known as a master osteogenic transcription factor. Runx2 activates and regulates osteogenesis as the targeted gene of many signaling pathways, including transforming growth factor-beta 1 (TGF-β1), BMP, Wingless type (Wnt), etc. ( James, 2013 ). Osteopontin is an extracellular matrix protein that is highly negatively charged; however regulation of the osteopontin gene is poorly understood. Our findings showed that in the first week ES caused a significant up-regulation of Runx2, Osteopontin and Col1A2 mRNA expression in BM-MSC, while in AT-MSC, the same was observed only in the expression of Runx2 and Osteopontin, after 14 days. We believe that the future of clinical tissue engineering trusts on the techniques and treatments which are able to save time and deliver economic benefits. Electrical stimulation as an old concept with modern applications in medicine appears to be able to shorten the time frame of cell proliferation and differentiation. This means saving time, while reducing costs of cell-based therapies. Moreover, electrical stimulation has the potential to trigger natural procedure of healing and regeneration ( Leppik et al. , 2015 ) which could revolutionize the reconstructive and regenerative medicine by implanting electrical stimulation devices which can deliver controlled electrical stimulation doses. Conclusions We have demonstrated that DC ES promotes Runx2, Osteopontin and Col1A2 expression in BM-MSCs already at 7 days. Our results indicate that DC ES effects osteogenic gene expression, in both BM- and AT-MSCs at specific time points. Moreover, these gene expression patterns differ between BM- and AT-derived MSC. These effects and differences should be taken into consideration when applying these two cell sources in tissue engineering applications. Studying ES-induced changes in cellular behavior could lead to methods to control and optimize cell behavior in tissue engineering treatments. Supplemental Information 10. 7717/peerj. 2821/supp-1 Data S1 Raw Data of MTT Click here for additional data file. 10. 7717/peerj. 2821/supp-2 Data S2 Raw Data of PCR Click here for additional data file.

10. 7717/peerj. 3079

PeerJ

Isolation and characterization of human articular chondrocytes from surgical waste after total knee arthroplasty (TKA)

Background Cartilage tissue engineering is a fast-evolving field of biomedical engineering, in which the chondrocytes represent the most commonly used cell type. Since research in tissue engineering always consumes a lot of cells, simple and cheap isolation methods could form a powerful basis to boost such studies and enable their faster progress to the clinics. Isolated chondrocytes can be used for autologous chondrocyte implantation in cartilage repair, and are the base for valuable models to investigate cartilage phenotype preservation, as well as enable studies of molecular features, nature and scales of cellular responses to alterations in the cartilage tissue. Methods Isolation and consequent cultivation of primary human adult articular chondrocytes from the surgical waste obtained during total knee arthroplasty (TKA) was performed. To evaluate the chondrogenic potential of the isolated cells, gene expression of collagen type 2 (COL2), collagen 1 (COL1) and aggrecan (ACAN) was evaluated. Immunocytochemical staining of all mentioned proteins was performed to evaluate chondrocyte specific production. Results Cartilage specific gene expression of COL2 and ACAN has been shown that the proposed protocol leads to isolation of cells with a high chondrogenic potential, possibly even specific phenotype preservation up to the second passage. COL1 expression has confirmed the tendency of the isolated cells dedifferentiation into a fibroblast-like phenotype already in the second passage, which confirms previous findings that higher passages should be used with care in cartilage tissue engineering. To evaluate the effectiveness of our approach, immunocytochemical staining of the evaluated chondrocyte specific products was performed as well. Discussion In this study, we developed a protocol for isolation and consequent cultivation of primary human adult articular chondrocytes with the desired phenotype from the surgical waste obtained during TKA. TKA is a common and very frequently performed orthopaedic surgery during which both femoral condyles are removed. The latter present the ideal source for a simple and relatively cheap isolation of chondrocytes as was confirmed in our study.

Introduction Damage to articular cartilage has important clinical implications since the cartilage tissue possesses a limited intrinsic healing potential and tends to an incomplete regeneration by local chondrocytes, accompanied with an inferior fibrocartilage formation ( Camp, Stuart & Krych, 2014 ; McNickle, Provencher & Cole, 2008 ; Richter et al. , 2016 ). Surgical intervention is often the only option, although the repair of damaged cartilage is often less than satisfactory, and rarely restores full function or returns the tissue to its native state ( Kerker, Leo & Sgaglione, 2008 ; Kock, Van Donkelaar & Ito, 2012 ; Tuli, Li & Tuan, 2003 ). Over the past decade a number of viable options of cartilage regeneration have been introduced into clinical practice ( Camarero-Espinosa et al. , 2016 ; Hettrich, Crawford & Rodeo, 2008 ; Schrobback et al. , 2011 ). Among these, autologous chondrocyte implantation (ACI) seems the most promising since it relies on the use of biodegradable materials that serve as temporary cell-carriers, enabling in vitro cell growth and subsequent implantation into the defective cartilage ( Bomer et al. , 2016 ; Niemeyer et al. , 2016 ; Robb et al. , 2012 ). Tissue engineering of articular cartilage remains challenging due to the specific structure of cartilage tissue, i. e. , its multiphasic cellular architecture together with remarkable weight-bearing characteristics (e. g. , resistance to mechanical stress and wear) ( Kim, Shin & Lim, 2012 ; Su et al. , 2012 ). Good understanding of the cartilage structure, physiology, and the molecular basis of chondrogenesis is key to in vitro cartilage production, either for use in tissue engineering or clinics ( Bhat, Tripathi & Kumar, 2011 ; Lee et al. , 2013 ; Li et al. , 2012 ). The state-of-the-art concept of in vitro cartilage tissue development combines the use of biocompatible and biodegradable carrier materials, the application of growth factors, the use of different cell types (stem or already differentiated) and different approaches to simulate the native mechanical stimulation ( Gardner et al. , 2013 ; Hildner et al. , 2011 ; Khan et al. , 2013 ; Naranda et al. , 2016 ). More specific challenges of articular cartilage tissue engineering remain the high consumption of cells and related costs, as well as the preparation of an ideal host scaffold. Although solutions to both mentioned challenges have been introduced in recent years ( Bassleer, Rovati & Franchimont, 1998 ; Stellavato et al. , 2016 ), the cell part is gaining far less research momentum. Therefore, it comes to no surprise that novel approaches for chondrocyte isolation are highly desired, especially considering the high prices of ordered cells. Optimisation of isolation yields, abundant cell sources and efficient culturing procedures that lead to preparation of desired, reproducible and relatively affordable cell cultures or/and material-cell constructs with good durability are therefore highly rated novelties in recent research ( Dehne et al. , 2009 ; Naranda et al. , 2016 ; Otero et al. , 2012 ). Several methods for chondrocyte isolation from various tissue parts and organisms were introduced over the last decades ( Hu et al. , 2002 ; Li et al. , 2015 ; Mirando et al. , 2014 ; Shortkroff & Spector, 1999 ; Strzelczyk, Benke & Gorecki, 2001 ; Xu & Zhang, 2014 ). Although their cell source varies, the crucial steps of these reported isolation protocols have a lot of common ground. One of the main similarities to digest the harvested tissue during the preparation of the primary culture is the use the enzyme type 2 collagenase ( Hayman et al. , 2006 ; Lagana et al. , 2014 ). Variations in the time of the tissue exposure to the enzyme ( Hayman et al. , 2006 ), as well as combining it with other enzymes (trypsin, pronase, hyaluronidase etc. ) is not unusual ( Jakob et al. , 2001 ). Several examples of effective chondrocyte isolation procedures including the source tissue and organism, the digestion enzyme, time of tissue exposure and the cell yield, were summarized by Oseni, Butler & Seifalian (2013). In their study, Oseni, Butler & Seifalian (2013) evaluated the necessary isolation and characterization procedures that would give a maximum yield with optimal cell viability for the engineering of large cartilaginous constructs such as the human nose and ear. At this point it is important to mention that the state of the source tissue has also to be accounted for Lagana et al. (2014). In this context, Lagana et al. performed characterization of basic parameters of articular chondrocytes isolated from 211 osteoarthritic patients. They concluded that a systematic characterization of the cellular yield and chondrocyte proliferation rates is very useful in view of a possible autologous cell therapy ( Lagana et al. , 2014 ). Therefore, it is very important to determine the quality of the cell source, which is known to greatly influence the outcome of engineered tissue ( Lagana et al. , 2014 ). The most demanding part in the process of in vitro culturing still presents the preservation of the desired phenotype to a high enough passage to yield sufficient cells to perform planned experiments ( Pei & He, 2012 ; Rosenzweig, Solar-Cafaggi & Quinn, 2012 ; Schnabel et al. , 2002 ; Wu et al. , 2014 ). Since the latter depends on numerous factors and can therefore be confirmed only by a combination of (often) expensive techniques (different microscopies, molecular analysis, immunocytochemistry etc. ), it is important to prepare protocols for an easier and cheaper preliminary phenotype confirmation by means of methods, available in most cell laboratories around the world. Since the desired phenotype can be identified by chondrocyte specific production ( Chen et al. , 2014 ; Han et al. , 2010 ), we believe that the easiest and safest preliminary method to prove phenotype preservation could be the analysis of gene expression. More specifically, this analysis should include the evaluation expression of genes related to cartilage specific markers (e. g. , collagen type 2 and aggrecan). To follow-up possible dedifferentiation towards the fibroblastic phenotype ( Duan et al. , 2015 ; Goldring, Tsuchimochi & Ijiri, 2006 ; Haudenschild et al. , 2001 ; Makris et al. , 2015 ; Otero et al. , 2012 ), we propose simultaneous measurement of up-regulation of collagen type 1. Based on all mentioned it is clear that chondrocyte isolation from an abundant source with a high yield, together with an effective and cheap preliminary phenotype confirmation method, would be greatly beneficial to boost the development of cartilage tissue engineering ( Cetinkaya et al. , 2011 ; Goepfert et al. , 2010 ; Schrobback et al. , 2011 ). This study was therefore designed to provide a relatively simple, yet effective procedure for isolation and culturing of human tissue derived primary chondrocytes up to the second passage. As the preliminary method of phenotype confirmation, we chose the evaluation of chondrocyte specific gene expression, together with morphological evaluation of cells. Such an approach provides a cheap and effective protocol to be considered an alternative to other available methods ( Hu et al. , 2002 ; Li et al. , 2015 ; Strzelczyk, Benke & Gorecki, 2001 ; Xu & Zhang, 2014 ), especially suitable for other laboratories to boost their respective entry level cartilage tissue engineering studies. To confirm our claims and the overall effectiveness of the used approach, immunocytochemical analysis of the most important chondrocyte specific extracellular matrix products (aggrecan and collagen type 2) were evaluated after one week of cell growth (for the second passage). To observe the tendency of the chondrocyte cells towards differentiation into fibroblast like cells, collagen type 1 was also evaluated using the same approach. Materials and Methods Materials All used materials and chemicals were of laboratory grade and purchased from Sigma-Aldrich, Germany, if not stated otherwise. For specific parts of the isolation process and cultivation, all used labware and chemicals were additionally sterilized using the standard autoclavation procedure (Avtoklav A-21, Kambič, Slovenia). Isolation of primary chondrocytes Full-thickness cartilage was surgically removed from the femoral condyle of arthritic knee of a 50 years old patient who underwent total knee arthroplasty (TKA) performed at the University Medical Centre Maribor, Slovenia (application reference: 123/05/14). Prior to surgery, no systemic disease or any treatment was reported for the donor patient. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments and was approved by the Republic of Slovenia National Medical Ethics Committee (Ljubljana, Slovenia). The patients’ written consent was obtained. The cartilage tissue was surgically removed under sterile conditions during TKA procedure. The standard cutting blocs for femoral resection were used and resection was performed in the usual manner. Distal and/or posterior femoral condyles were used for chondrocyte isolation depending on the macroscopic condition of the cartilage tissue (due to e. g. , osteoarthritis). Immediately after the removal, the bone cuts were transferred into a previously sterilized 250 ml glass bottle filled with phosphate buffered saline (PBS; Sigma-Aldrich, Munich, Germany) and immediately brought to the cell isolation laboratory. The cartilage-bone tissue was transferred to a petri dish filled with PBS to prevent drying of the tissue. In the cell isolation laboratory, the cartilage tissue was carefully removed from the bone cuts surface using a No 11 blade to obtain approximately 2 x 2 mm pieces of cartilage tissue. PBS was carefully removed by a pipette and the petri dish was immediately filled with 10 mL solution of 0. 25 wt. % Trypsin/EDTA (Sigma, France). The as-prepared cartilage pieces were incubated for 3 h at 37°C and 5 wt. % CO 2 (CO 2 Incubator MCO-19AICUVH-PE; Panasonic, Tokyo, Japan), followed by addition of 20 mL of Advanced Dulbecco’s modified Eagle’s medium (Advanced DMEM; Gibco, Grand Island, NY, USA) to the cell suspension. The suspension was transferred to a 50 mL falcon tube and centrifuged at 300 x g for 10 min (Centrifuge 5804 R; Eppendorf, Hamburg, Germany). The supernatant was carefully discarded and the cell pellet was re-suspended in 20 mL of Advanced DMEM and centrifuged at 200× g for 5 min (Centrifuge 5804 R; Eppendorf, Hamburg, Germany). The supernatant was again carefully discarded and the cell pellet re-suspended in 10 mL Advanced DMEM supplemented with 100 IU/ml Penicillin, 1 mg/ml Streptomycin, 2mM L-glutamine and 5 wt. % foetal bovine serum (FBS; Gibco, Grand Island, NY, USA) and plated on 25 cm 2 flasks (in triplicates). In the cell pellet, very small fragments of cartilage were also present. Besides primary chondrocytes, these fragments were also seeded and after a week of incubation, the cells were observed crawling from the tissue fragments. Together with the primary chondrocytes these were then left until confluence was reached. Growing cells were regularly observed with an Axiovert 40 inverted optical microscope (Zeiss, Oberkochen, Germany) at several magnifications. The culturing medium was changed every three days. The general steps of the procedure are schematically depicted in Fig. 1. 10. 7717/peerj. 3079/fig-1 Figure 1 Chondrocyte isolation from cartilage in a short overview of the most important preparation steps. Gene expression analysis Gene expression analysis of cartilage specific markers collagen type 2 (COL2) and aggrecan (ACAN) was performed in order to determine the primary chondrogenic phenotype. Possible dedifferentiation to a more fibroblast like cell type was evaluated by monitoring the expression of collagen type 1 (COL1). After confluence was reached in all respective samples (triplicates) (see above ‘Isolation of primary chondrocytes’. for details), the cell suspension was transferred to micro-centrifuge tubes, and 1. 4 mL of TRI reagent (Sigma-Aldrich, Munich, Germany) was added. The tubes were vortexed for 30 min at room temperature. Afterwards, 280 µL of chloroform (Sigma-Aldrich, Munich, Germany) was added and the tubes were further vortexed for 15 min and centrifuged at 12. 000 rpm and 4°C. RNA extraction was carried out according to the manufacturer’s instructions ( Chomczynski, 1993 ; Louveau, Chaudhuri & Etherton, 1991 ). Concentration and purity of the extracted cellular RNA was determined using NanoDrop 2000c (Thermo Scientific, Waltham, MA, USA) through optical density readings at 260 nm and a 260/280 nm ratio. cDNA was obtained by using a cDNA reverse transcription kit (Applied Biosystems, California, USA). Primer sequences for cartilage target genes ACAN and COL2 were obtained from Caterson et al. (2001), while the corresponding mRNA sequences were retrieved from PubMed Nucleotide database ( http://www. ncbi. nlm. nih. gov/nuccore/ ) and the AceView database ( Thierry-Mieg & Thierry-Mieg, 2006 ). Primers for the target gene COL1 were designed using IDT oligo analyser ( http://eu. idtdna. com/calc/analyzer ). The primer sequences with the corresponding mRNA sequences and the corresponding NCBI accession numbers are given in Table 1. 2 µL of each cDNA sample with concentration of 15 ng/µL was used for quantitative real time PCR (qPCR) analysis performed using LightCycler 480 thermocycler (Roche, Switzerland) and with 2 × Maxima SYBR Green qPCR master mix (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. The quality and specificity of PCR amplicons were checked using melting curve analyses and agarose gel electrophoresis. All shown results are presented as average values with the standard errors. 10. 7717/peerj. 3079/table-1 Table 1 Primer sequences with the corresponding mRNA sequence and the corresponding NCBI accession numbers. Gene Gene name Accession number Primer sequence 5′ → 3′ ACAN Aggrecan NM_013227. 3 TGAGGAGGGCTGGAACAAGTACC NM_001135. 3 GGAGGTGGTAATTGCAGGGAACA COL1 Collagen type 1, alpha 1 NM_000088. 3 CGGCTCCTGCTCCTCTTAG CACACGTCTCGGTCATGGTA COL2 Collagen type 2, alpha 1 NM_001844. 4 TTTCCCAGGTCAAGATGGTC NM_033150. 2 CTGCAGCACCTGTCTCACCA GAPDH Glyceraldehyde-3-phosphate dehydrogenase NM_001289745. 1 GGGCTGCTTTTAACTCTGGT NM_002046. 5 TGGCAGGTTTTTCTAGACGG NM_001289746. 1 NM_001256799. 2 Immunocytochemistry We characterized cells according to the expression of specific surface proteins (COL1, COL2, ACAN). Additional staining was performed in order to analyse the cells’ general morphology (cytoskeleton (actin)—using Phalloidin—iFluor 555 Reagent (Abcam, Cambridge, UK); nucleus—using mounting medium with 4′, 6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, Munich, Germany)). Some more details about respective methods are described below. All micrographs were taken using either Floid Cell Imaging Station (Thermo Fisher Scientific, Waltham, MA, USA) or EVOS FL Cell Imaging System (Thermo Fisher Scientific, Waltham, MA, USA). General protocol for immunocytochemistry Round glass slides (2 r = 12 mm) were placed on the bottom of wells in a P24 plate (in triplicate for each used dye) similar to the procedure used by Oseni, Butler & Seifalian (2013). Isolated cells (from the second passage) at a density of 50, 000 cells / well were placed on each of the glass slides and incubated at 37°C, 5 wt. % CO 2 for seven days. The medium (Advanced DMEM, supplemented with foetal bovine serum (FBS, Gibco, Grand Island, NY, USA)) was removed and the cells were washed with phosphate buffered saline (PBS; Sigma-Aldrich, Munich, Germany) once. Fixation of cells was performed using the Fixation Solution (Millipore, Billerica, MA, USA) for 10 min at room temperature, followed by washing of the cells three times with cold PBS (∼4°C). Further sample handling differed for respective staining procedures. Namely, ACAN, COL1 and COL2 were stained using primary and secondary antibodies (the manufacturers protocols were followed for this purpose), whereas actin was stained in a single step (again, according to the manufacturers protocol). Actin staining Following the general protocol for immunocytochemistry, the working solution of the conjugated Phalloidin (1, 000× Phalloidin stock solution in dimethyl sulfoxide DMSO (Abcam, UK), 1/1, 000 dilution in PBS with 1 wt. % bovine serum albumin (BSA; Sigma-Aldrich, Munich, Germany) and 0. 1 wt. % Tween 20 (Sigma-Aldrich, Munich, Germany)) was added. Incubation was performed for 90 min at room temperature and in a dark room. Rinsing was performed with PBS and was repeated three times. The final step was the addition of the Fluoroshield Mounting Medium with DAPI. Micrographs were taken at the suitable wavelengths for respective used dyes (excitation/emission: DAPI = 306/460 nm and Phalloidin = 556/574 nm). Staining of COL1, COL2 and ACAN Following the general protocol for immunocytochemistry described above, the cells were incubated for 30 min with PBS, supplemented with 1 wt. % BSA and 0. 1 wt. % solution of Tween 20 to block nonspecific binding of antibodies. All incubations with the primary antibodies was performed overnight at 4 ∘ C. Respective dilutions (in PBS with 1 wt. % BSA and 0. 1 wt. % solution of Tween 20) of the primary antibodies were as follows: 1. ACAN: Anti-Aggrecan antibody [6-B-4] (Abcam, UK), 1:50, 2. COL2: Anti-Collagen 2 antibody (Abcam, UK), 1: 200, 3. COL1: Anti-Collagen 1 antibody (Abcam, UK), 1: 500. After incubation, the cells were washed three times with PBS for 5 min. Incubation of cells with the secondary antibodies was performed in a dark at room temperature for 1 h (the same procedure was used also as the control for the attachment of respective secondary antibodies). The dilutions of the secondary antibodies (in PBS with 1 wt. % BSA and 0. 1 wt. % solution of Tween 20) were as follows: 1. ACAN: Rabbit Anti-Mouse IgG H & L (Alexa Fluor 488) preabsorbed (Abcam, UK), 1: 1, 000, 2. COL2 and COL1: Goat anti-rabbit IgG H & L (Alexa Fluor 594) (Abcam, UK), 1: 1, 000. After incubation, the cells were washed three times with PBS for 5 min. Finally, three drops of the Mounting Medium Fluoroshield with DAPI were added and the solution was left on the cells for 5 min. Micrographs were taken at the suitable wavelengths for respective used dyes (excitation/emission: ACAN = 495/519 nm and COL2/COL1 = 590/617 nm). Results Isolation of primary chondrocytes As mentioned in the Materials and methods section, the full-thickness cartilage was obtained from the femoral condyle of an arthritic knee during knee arthroplasty (TKA) performed at the University Medical Centre Maribor, Slovenia. TKA is a common procedure at the mentioned hospital, considering that approximately 700 such surgeries are performed each year ( Univerzitetni klinicni Center, 2014 ). Since the removed cartilage tissue is considered surgical waste, this presents a reliable and continuous source for isolation of primary chondrocytes. The primary chondrocytes were isolated as described in the Materials and methods section. During their cultivation, their morphology and proliferation were regularly observed using inverted optical microscopy ( Fig. 2 ). Figure 2A shows the thin slice of cartilage that was used for their cultivation, while Figs. 2B – 2D present the primary human chondrocytes in a monolayer culture at different cultivation times. This initial examination was performed to follow possible morphological changes in the cell shapes, which would indicate possible dedifferentiation. 10. 7717/peerj. 3079/fig-2 Figure 2 (A) Thin slice of cartilage for primary chondrocyte isolation; (B–D) the primary human chondrocyte culture in a monolayer after 3, 6 and 9 days, respectively. The magnification of all shown images is 50×. The full confluence of the isolated cells for the first and second passage was reached after two (14 days) and after one (seven days) week, respectively. Cell growth stopped presumably due to contact inhibition ( Lackie, 2013 ). A comparison between the primary chondrocyte culture and the obtained chondrocyte cultures after the first and second passages is shown in Fig. 3. The cells formed confluent monolayers (after the above mentioned cultivation times) and appeared polygonal in shape ( Figs. 3A – 3C ). It can be observed that the chondrocyte morphology became more spindle-like in the second passage ( Fig. 3C ), showing their tendency for dedifferentiation, most likely towards fibroblast-like cells ( Hong & Reddi, 2013 ). Observing the mentioned changes was an indication that the third passage will not yield a high percentage of chondrocytes only using the proposed cultivation conditions. 10. 7717/peerj. 3079/fig-3 Figure 3 Human chondrocyte culture: (A) the explant culture of chondrocytes (“primary culture”), (B) monolayer of chondrocytes after first passage, and (C) monolayer of chondrocytes after the second passage. The magnification of all shown images is 50× (the inlay images were taken with a magnification of 100×). Gene expression analysis of the isolated chondrocytes Now that we determined the suitable number of passages presumably yielding a high percentage of chondrocyte cells, we performed additional characterization to confirm the chondrocytes’ desired phenotype. Analysis of gene expression was chosen due to its affordability and availability. The isolated cells from the human articular cartilage were characterized in regard of the genes related to specific chondrogenic production, namely collagen type 2 (COL2) and aggrecan (ACAN). To detect possible dedifferentiation towards fibroblast like cells, expression of collagen type 1 (COL1) was also determined. Expression of all three mentioned genes was performed after the confluence was reached for the second passage (after seven days). As shown in Figs. 4 and 5, both cartilage specific genes (COL2 and ACAN) and also the marker of fibrocartilage (COL1) were expressed in the isolated chondrocytes in both passages. qPCR results are presented as absolute Ct values. Reference gene GAPDH was used as an internal control ( Chen et al. , 2016 ). 10. 7717/peerj. 3079/fig-4 Figure 4 cDNA products of analysed genes (GAPDH, collagen type 1, collagen type 2 and aggrecan) at the end-point of qPCR on agarose gel electrophoresis. Analyzed genes: GAPDH (702 bp), COL2 (377 bp), COL1 (137 bp), ACAN (350 bp) and DNA markers (433 bp, 245 bp, 203 bp, 114 bp). 10. 7717/peerj. 3079/fig-5 Figure 5 Results of qPCR analysis presented as absolute Ct values of target genes expression (ACAN, COL1, COL2 and GADPH). The results are presented as average values with the standard errors of a triplicate. Immunocytochemistry We performed immunocytochemistry on the isolated cells to investigate chondrocyte phenotype alterations (ACAN, COL1 and COL2). Additionally, the cytoskeleton (actin) and cell nucleus were stained to show the overall healthy morphology of the cells. All staining was performed in three repetitions. As the negative control, staining only with the respective secondary antibodies, as well as with the Mounting medium with DAPI (after one day and after two days), were used. Production of all three proteins was confirmed ( Figs. 6A – 6C ), which is in agreement with the results from the molecular analysis. All negative controls have shown no fluorescence, confirming the effectiveness and specificity of the used protocols. Staining of actin ( Fig. 6D ) confirmed the expected morphology of healthy cells, which is in agreement with the micrographs using optical microscopy. 10. 7717/peerj. 3079/fig-6 Figure 6 Micrographs of the stained samples: (A) for ACAN, (B) for COL2 and (C) COL1. Additionally, (D) shows the cells with a stained cytoskeleton (actin). For all samples a mounting medium with DAPI was used to stain the nuclei. The magnification of all shown images is 460× (according to the manufacturers microscope specifications). Discussion The development of novel solutions related to any tissue engineering application consumes a huge number of cells to prove safety and efficiency ( Groeber et al. , 2012 ; Maver et al. , 2015 ; Mohd Hilmi & Halim, 2015 ; Rodriguez-Vazquez et al. , 2015 ). Cartilage tissue engineering is no exception, and hence large scale expansion of chondrocytes is required either for novel scaffold testing, determination of potential cytotoxic effects of medical devices and implants for orthopaedic use ( Bomer et al. , 2016 ; Camarero-Espinosa et al. , 2016 ; Makris et al. , 2015 ). Cultivation of such high cell counts is a demanding task, especially considering the low number of obtained cells in the primary culture, and an often limited amount of available tissue. Consequently, further expansion and consecutive passages are needed, which on the other hand can lead to dedifferentiation ( Mirando et al. , 2014 ; Shortkroff & Spector, 1999 ; Thirion & Berenbaum, 2004 ). The latter is evident by morphological changes of the cells from polygonal to more elongated, as well as through a reduction in the growth rate ( Cetinkaya et al. , 2011 ; Haudenschild et al. , 2001 ; Otero et al. , 2012 ). For example, development of novel scaffolds for cartilage tissue engineering often requires a million cells per sample scaffold (the number depends on the size of the scaffold to be tested), exposing the high demand for cells and at the same time one of the major bottlenecks in development of novel tissue engineering solutions. At later passages, the quality of chondrocytes gradually decreases and is characterized with many of the phenotypic traits of fibroblast like cells and an increased synthesis of collagen type 1, rather than type 2 ( Bonaventure et al. , 1994 ; Diekman et al. , 2010 ; Schnabel et al. , 2002 ). A sufficient number of cells can be ensured either through significant expenses (purchase of cells from different cell banks) or isolation of desired cells from tissues. While the first scenario requires sufficient funds, the latter requires appropriate tissue sources, an approval of respective Committees of Medical Ethics, and a rigorous final analysis to confirm the isolation of the desired cell type only. Since we work in the close proximity and in tight collaboration with the local University Medical Centre, the second scenario was more convenient. Our goal was to prepare a simpler and generally available protocol, which would include the isolation of primary chondrocytes from full-thickness cartilage that is surgically removed from the femoral condyle of an arthritic knee during total knee arthroplasty (TKA). As a preliminary prove of the protocols’ efficiency, we considered gene expression analysis as the best option, since it is affordable and the required instrumentation (PCR, inverted optical microscope) is most likely available in most cell biology laboratories. The set of analysed genes was carefully chosen considering the available literature to monitor cartilage phenotype alterations ( Caterson et al. , 2001 ; Diekman et al. , 2010 ; Grogan et al. , 2014 ; Jonitz et al. , 2012 ; Seda Tigli et al. , 2009 ; Shi et al. , 2014 ). Based on the mentioned, the correlation between COL2 and COL1 in addition to ACAN, seemed to be the most suitable. For confirmation of the effectiveness of the proposed approach in terms of chondrocyte specific production besides the gene expression, immunocytochemical staining of COL2, COL1 and ACAN was used as well. The latter confirmed the chondrocyte specific productions (ACAN and COL2), as well as the presence of COL1, which could be an indication of ongoing dedifferentiation to more fibroblast like cells. In general, the chondrocyte isolation protocol can be divided into different stages: isolation, seeding and chondrocytes grow in culture, although description and number of steps can vary ( Gosset et al. , 2008 ; Thirion & Berenbaum, 2004 ). After initial plating of the primary cultures, the chondrocytes spread out after 2–3 days and after 4–7 days the sufficient amounts of total RNA may be extracted. Primary cartilage phenotype (often confirmed by evaluating the presence of COL2 and ACAN mRNAs) may be initially preserved, but the expression of nonspecific collagens (e. g. , COL1) begins to appear already 7 days after isolation ( Otero et al. , 2005 ). Moreover, adult articular chondrocytes are strongly contact-inhibited and undergo a rapid change in phenotype and gene expression, termed “dedifferentiation”, when isolated from cartilage tissue and cultured on culturing plastics ( Haudenschild et al. , 2001 ). Therefore, primary chondrocyte cultures should be used for experimental analyses immediately before or just after confluence is reached to assure optimal matrix synthesis and cellular responsiveness ( Schneevoigt et al. , 2016 ). In the last two decades, several chondrocyte isolation protocols were developed and reported on ( Hayman et al. , 2006 ; Hu et al. , 2002 ; Jakob et al. , 2001 ; Lagana et al. , 2014 ; Oseni, Butler & Seifalian, 2013 ; Strzelczyk, Benke & Gorecki, 2001 ). For example, an important recent study was conducted by Lagana et al. (2014), who isolated chondrocytes from 211 osteoarthritic (OA) patients undergoing total joint replacement. The authors of this study analysed specific features of chondrocytes such as cellular yield, cell doubling rate and the dependence between these parameters and patient-related data (e. g. , joint type, age and gender). They concluded that such a systematic characterization of important cell source parameters could be useful in view of a possible autologous cell therapy for osteoarthritis, since the cell source quality is known to greatly influence the outcome of engineered tissue ( Lagana et al. , 2014 ). Another crucial study that we studied in details prior to our experimental design, was performed by Oseni, Butler & Seifalian (2013). In this study, the authors focused on a very important factor related to possible clinical use of cartilage tissue engineered products, namely the optimization of the isolation protocol to allow for a large-scale production. The result of their study was an optimized protocol with exactly defined isolation parameters (e. g. , enzyme and concentration to be used, time of digestion and the seeding density for tissue culturing). Two other studies have to be mentioned in this context as well. Namely, the studies from Jakob et al. (2001) and Hayman et al. (2006), respectively. Jakob et al. focused on the research of possible chondrocyte isolation yield improvement by using various combinations of enzymes and reagents. Their results indicated that chondrocyte yields and capacity to attach and proliferate are not highly sensitive to the specific isolation protocol used ( Jakob et al. , 2001 ). Finally, Hayman et al. (2006) conducted a study, in which they tested combinations of three different enzymes and variable incubation/digestion times. A very important discussion point raised by the authors of this study was that different isolation protocols are to be used, if the focus is only on the yield or the goal is to produce preferentially “native” chondrocytes ( Hayman et al. , 2006 ). The protocol of chondrocyte isolation described in this article led to successful growth and proliferation of cells with a proven chondrogenic potential up to the second passage as shown using molecular and immunocytochemical analysis. The characterization of primary human chondrocytes by molecular analysis showed the expression of cartilage specific genes (COL2 and ACAN), as well as a sign of dedifferentiation towards fibrocartilage for the second passage (indicated by the expression of COL1). In comparison with other available chondrocyte isolation protocols, we introduced some changes to the general protocol. As mentioned before, our target was a simple protocol with a high enough yield to conduct preliminary cartilage tissue engineering experiment, like testing of suitability of novel materials ( Naranda et al. , 2016 ). According to previous studies, the most commonly used enzyme in chondrocyte isolation, is type 2 collagenase ( Hayman et al. , 2006 ; Lagana et al. , 2014 ; Oseni, Butler & Seifalian, 2013 ). Various incubation times are used to allow for tissue digestion, but in our experience, longer enzyme exposure times of tissues (and with longer exposures, an increasing number of cells as well) often lead to an increased number of dead cells and/or a lower yield of the cells with a desired phenotype. Considering all mentioned, we used a Trypsin/EDTA combination and an incubation time of 3 h. Although this is not the first research study reporting the use of trypsin for chondrocyte isolation ( Hidvegi et al. , 2006 ; Jakob et al. , 2001 ), to the best of our knowledge it is the only one to use only this enzyme during the isolation protocol. Also, the reported incubation time is different to the mentioned studies. In addition, our protocol does not include the use of any growth factors like reported in some studies ( Lagana et al. , 2014 ), again with the focus to simplify the overall protocol. Moreover, no enzyme predigestion step was introduced in our protocol, like in some studies ( Oseni, Butler & Seifalian, 2013 ). The purpose of our study was not to revolutionize the chondrocyte isolation procedures, but rather to push the evolution of cartilage tissue engineering. As such, our desire was to present an alternative, affordable and relatively simple approach of chondrocyte isolation, especially suitable for laboratories working closely together with orthopaedic clinics. Such laboratories have the unique opportunity to use surgical waste materials, occurring during TKA. Since TKA is a very common surgery (considering the present demographics, the incidence will only increase ( Peterson et al. , 2015 )), this approach could make cartilage-related studies far more available also for laboratories with limited resources, and hence push the overall development of this field towards novel and cheaper therapeutic solutions. Based on our results, we can claim that the combination of the use of surgical waste tissue occurring during TKA, and analysis by inverted optical microscopy and chondrocyte specific gene expression, as well confirmation of chondrocyte specific production, indeed results in an alternative and affordable means to boost cartilage-related research in the future. Conclusion In this study, we describe a simple and affordable procedure of isolation and cultivation of human articular chondrocytes demonstrated a high chondrogenic potential to the second passage. As the source material, we propose the surgical waste tissue occurring during total knee arthroplasty (TKA). Chondrocyte cells are crucial not only for development of therapeutic approaches in cartilage repair (e. g. , autologous chondrocyte implantation—ACI), but are necessary in cartilage tissue engineering to allow the development of functional cell models and novel scaffolds. For this purpose, chondrocytes have to be isolated in sufficient quantities and their phenotype should be preserved. Since all mentioned challenges are related to very high costs, we propose alternative isolation and testing protocols that are cheaper and could especially boost the preliminary studies related to cartilage research.

10. 7717/peerj. 3301

PeerJ

Preservation media, durations and cell concentrations of short-term storage affect key features of human adipose-derived mesenchymal stem cells for therapeutic application

Background Adipose-derived mesenchymal stem cells (ADSCs) have shown great potential in the treatment of various diseases. However, the optimum short-term storage condition of ADSCs in 2∼8 °C is rarely reported. This study aimed at optimizing a short-term storage condition to ensure the viability and function of ADSCs before transplantation. Methods Preservation media and durations of storage were evaluated by cell viability, apoptosis, adhesion ability and colony-forming unit (CFU) capacity of ADSCs. The abilities of cell proliferation and differentiation were used to optimize cell concentrations. Optimized preservation condition was evaluated by cell surface markers, cell cycle and immunosuppressive capacity. Results A total of 5% human serum albumin in multiple electrolytes (ME + HSA) was the optimized medium with high cell viability, low cluster rate, good adhesion ability and high CFU capacity of ADSCs. Duration of storage should be limited to 24 h to ensure the quality of ADSCs before transplantation. A concentration of 5 × 10 6 cells/ml was the most suitable cell concentration with low late stage apoptosis, rapid proliferation and good osteogenic and adipogenic differentiation ability. This selected condition did not change surface markers, cell cycle, indoleamine 2, 3-dioxygenase 1 (IDO1) gene expression and kynurenine (Kyn) concentration significantly. Discussion In this study, ME + HSA was found to be the best medium, most likely due to the supplement of HSA which could protect cells, the physiological pH (7. 4) of ME and sodium gluconate ingredient in ME which could provide energy for cells. Duration should be limited to 24 h because of reduced nutrient supply and increased waste and lactic acid accumulation during prolonged storage. To keep cell proliferation and limit lactic acid accumulation, the proper cell concentration is 5× 10 6 cells/ml. Surface markers, cell cycle and immunosuppressive capacity did not change significantly after storage using the optimized condition, which confirmed our results that this optimized short-term storage condition of MSCs has a great potential for the application of cell therapy.

Introduction The use of mesenchymal stem cells (MSCs) is a potential regenerative therapeutic strategy because of their regenerative and immune-regulatory properties ( Kaplan, Youd & Lodie, 2011 ). Currently MSCs are widely used in treating various diseases, including immune disorders, degenerative diseases, and tissue injuries ( Venkataramana et al. , 2010 ; Wei et al. , 2013 ). Although MSCs can be derived from almost every tissue of the body ( da Silva Meirelles, Chagastelles & Nardi, 2006 ; Kern et al. , 2006 ; Mosna, Sensebé & Krampera, 2010 ), adipose-derived MSCs (ADSCs) are ideal cells for future use in regenerative medicine due to the high abundance of ADSCs in adipose tissue and the minimal morbidity associated with harvesting MSCs from adipose tissue ( Bajek et al. , 2016 ; Gomez-Mauricio et al. , 2013 ). Large-scale application of MSCs in regenerative medicine demands clinically acceptable “off-the-shelf” cell therapy products. Stem cells cryopreserved using dimethyl sulfoxide (DMSO) are commonly used in regenerative medicine; however, a great number of observed adverse reactions were tenuously or convincingly associated with the cryoprotectant DMSO. Cerebral infarction and myocardial injury occurred in two patients after intravenous injection of autologous stem cells with DMSO ( Chen-Plotkin et al. , 2007 ). Neurotoxicity was observed in a patient who suffered from a generalized tonic seizure upon infusion of DMSO-cryopreserved peripheral blood stem cells ( Mueller et al. , 2007 ). During the infusion of hematopoietic stem cells without washing the DMSO a patient developed bradycardia, abdominal pain and nausea, and 24 h later he developed anasarca and hypertension ( Ruiz-Delgado et al. , 2009 ). Other side effects caused by DMSO include cardiac arrest ( Rapoport et al. , 1991 ), severe respiratory arrest ( Benekli et al. , 2000 ), paradoxical embolism ( Darabi, Brown & Kao, 2005 ), transient consciousness loss ( Schlegel et al. , 2009 ) and so on. There are several alternative cryoprotectants, such as ethylene glycol, methanol and polymer hydroxyethyl starch, but these would cause cell injury and researchers have to focus on how to minimize or eliminate their toxicity ( Marquez-Curtis et al. , 2015 ). In addition, sometimes brief (i. e. , 24–48 h) storage of MSCs is needed, but cryopreservation of MSCs is not a practical way for brief storage of MSCs ( Haack-Sorensen et al. , 2007 ; Lazarus et al. , 2005 ; Kim et al. , 2004 ; Lane et al. , 2009 ). The short-term storage of fresh MSCs in 2∼8 °C does not require cryoprotectants which have underlying safety issues. Also it does not require complicated liquid nitrogen device, which means it can be used to improve the transportability of MSCs products. In order to maintain high quality of MSCs during the time between harvesting and administration, the surrounding environment needs to be strictly controlled and some key factors must be taken into account ( Gálvez-Martín et al. , 2014 ). Several factors including preservation media, durations of storage and cell concentrations may affect the viability and function of MSCs when suspended in liquid storage medium ( Lane et al. , 2009 ; Kao, Kim & Daley, 2011 ; Chen et al. , 2013 ). Different kinds of preservation media including M199 ( Mohamadnejad et al. , 2007 ), PBS ( Wang et al. , 2011 ), NS ( Venkataramana et al. , 2010 ), PlasmalyteA ( Chen et al. , 2013 ), 1% HSA in DMEM ( Lane et al. , 2009 ), 20% HSA and 5% glucose in Ringer’s lactate ( Gálvez-Martín et al. , 2014 ) have been used in previous studies. However, M199 and PBS are not approved vehicles for safe injections thus they could not be used clinically. The viability of cells stored in 1% HSA in DMEM decreased rapidly ( Lane et al. , 2009 ). There was no quality evaluation of the cells suspended in NS ( Venkataramana et al. , 2010 ). Transplantation of cells immediately after the harvest could receive best clinical outcomes because the quality of cells before administration affects therapeutic efficacy greatly. However, it takes hours or days to progress from harvest to transplantation inevitably ( Sohn et al. , 2013 ). Thus, it is of great importance to optimize an appropriate duration of storage with clinically acceptable cell viability and function. Although cell viability in 20% HSA and 5% glucose in Ringer’s lactate was high (>80%) till 48 h, there was no research on the proliferation and immunosuppressive capacity of MSCs ( Gálvez-Martín et al. , 2014 ). It has been reported that cell concentrations may affect biological properties of hematopoietic stem cells and cell viability of non-MSC cell lines ( De Loecker et al. , 1998 ; Espina et al. , 2016 ). Thus, we also evaluated the effects of cell concentrations during short-term storage on the characteristics of MSCs. Short-term storage condition with high viability and function of ADSCs has not been studied systematically so far. We aimed to optimize a short-term storage condition to ensure the viability and function of ADSCs for therapeutic application. Materials & Methods Study design This study consisted of four consecutive parts in which preceding results were applied in the subsequent steps. In part I, the impact of different media was measured and the most suitable medium was subsequently used throughout the study. NS and PlasmalyteA are commonly used vehicles, and the supplement of HSA could protect cells from environmental stress and prevent adherence to the tubes or vials ( Ikebe & Suzuki, 2014 ). Dextrose provides a source of energy for cell metabolism ( Anderson et al. , 1992 ). Previous study reported that the best preservation medium for short-term storage was 5% dextrose ( Pal, Hanwate & Totey, 2008 ). Thus, we decided to study 5% human serum albumin in 0. 9% normal saline (NS + HSA), 5% human serum albumin in multiple electrolytes (ME + HSA, as Baxter Healthcare Co. , Ltd. stopped production of PlasmalyteA here in China, ME with totally the same formula as PlasmalyteA was chosen to substitute PlasmalyteA. ), dextrose and growth medium (GM) by measuring cell viability and cluster rates, adhesion ability, apoptosis and CFU capacity. In part II, two durations (24 h and 48 h) of storage were evaluated by parameters described above and optimized duration was applied in the following study. In part III, cell concentrations were investigated by adopting measurement of proliferation and differentiation. In part IV, quantification of surface markers, cell cycle, IDO1 gene expression and Kyn concentration of ADSCs suspended in optimized concentration were studied. Cells were stored in 2ml cryogenic vials (Corning Incorporated, Corning, NY, USA) and then placed in a cold chain shipping container designed to ensure stable cooled products transport (2∼8 °C; more than 50 h). A continuous temperature monitoring device was embedded in the cold chain shipping container. ADSCs were suspended after storage in 2∼8 °C for the following research. Unstored cells were fresh cells that did not undergo storage. Preparation of storage media 5% (500 ml: 25 g) dextrose injection was purchased from Baxter Healthcare (Shanghai) Co. , Ltd. , China. NS + HSA was prepared by adding 5 ml 20% HSA (Shanghai Institute of Biological Products Co. , Ltd. , China) into 15 ml NS (Hunan Kelun Pharmaceutical Co. , Ltd. , China). ME + HSA was prepared by adding 5 ml 20% HSA into 15 ml ME (Sichuan Kelun Pharmaceutical Co. , Ltd. , China). GM was α -modified minimal essential medium ( α -MEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gemini, Australia). Isolation and culture of ADSCs This study was approved by the Ethics Committee of the Shanghai First People’s Hospital at Shanghai Jiao Tong University (number 2013KY080). The 33-year-old woman volunteer signed the contract and permitted adipose tissue to be used for storage and scientific research. The first passage of ADSCs was separated and cultured in GMP condition in Shanghai Kun’ai Biological Technology Co. , LTD. Cells were seeded in 6-well plates, cultured with α -MEM supplemented with 10% FBS and 1% penicillin/streptomycin and maintained at 37 °C in a humidified atmosphere at 5% CO 2. Medium was changed every three days. Cells were passaged at approximately 80% confluence and passages 3–5 were used for the experiments. Cell viability and apoptosis Cell viability and cluster rates were determined on an automatic cell counter (Countstar, Shanghai Ruiyu Biotech Co. , Ltd. , China) using trypan blue (Gibco, Gaithersburg, MD, USA) staining method. Unstained cells were counted as live cells. Cell apoptosis was analyzed using a FITC-conjugated Annexin V/PI assay kit (SAB, USA) as our previous study has reported ( Ren et al. , 2016 ). 0. 5 million cells were rinsed twice with PBS. After centrifugation, 500 µl buffer was added to suspend cells. A total of 5 µl Annexin V-FITC was added to the cell suspension and cells were incubated in the dark for 20 min at 4 °C, followed by addition of 10 µl PI and incubation in the dark for 5 min at 4 °C. Cell apoptosis was determined by flow cytometry (BD Bioscience, San Jose, CA, USA) and analyzed using FlowJo software (TreeStar, Ashland, OR, USA). Adhesion ability A total of 1 million cells were seeded in a 100-mm cell culture dish (Corning Incorporated, Corning, NY, USA) and allowed to attach for 24 h at 37 °C in a humidified atmosphere at 5% CO 2. Cells were then observed and evaluated under an inverted microscope IX51 (Olympus, Tokyo, Japan). Colony-forming unit (CFU) capacity A total of 250 cells suspended in 2 ml α -MEM supplemented with 10% FBS were seeded in a 60-mm dish (Corning Incorporated, Corning, NY, USA) or per well of a six-well plate (Corning Incorporated, USA). After culture for 10 days at 37 °C in a humidified atmosphere at 5% CO 2, cells were rinsed twice with PBS and fixed with methanol for 20 min at −20 °C. Then methanol was removed and cells were rinsed twice with PBS. Cells were stained with 1 ml 0. 2% crystal violet (Sinopharm Chemical Reagent Ltd. , Shanghai, China) for 1 h at room temperature. The plates were rinsed twice with PBS. Stained colonies with >50 cells were scored as CFU and counted under an inverted microscope. Cell proliferation Proliferation of ADSCs was assessed by a nontoxic metabolic indicator Alamar Blue (Life Technologies, Carlsbad, CA, USA) as our previous study has reported ( Chen et al. , 2016 ). In brief, cells were seeded in a 24-well plate (Corning Incorporated, Corning, NY, USA) at a concentration of 2 × 10 4 cells/well. After culture for 24 h, culture medium was changed into fresh medium containing 10% (v/v) Alamar Blue indicator and then cells were incubated in the dark for 3 h at 37 °C. Absorbance of the extracted dye was measured by an enzyme immunoassay analyzer (Thermo, USA) at wavelengths of 570 and 590 nm. Population-doubling time (PDT) was calculated according to the following equation: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}PDT=t\times lg2/(lg{N}_{t}-lg{N}_{0}). \end{eqnarray*}\end{document} P D T = t × l g 2 ∕ l g N t − l g N 0. In the equation, t indicated duration of proliferation, N t and N 0 represented harvesting cell number and initial seeding cell number, respectively. Differentiation assay Adipogenic differentiation Cells were seeded in a 12-well plate (Corning Incorporated, Corning, NY, USA) at a concentration of 1 × 10 4 cells/ well and cultured with α -MEM supplemented with 10% FBS until 80% confluency. Cells in differentiation group were incubated with adipogenic induction medium consisted of α -MEM supplemented with 10% FBS, 1 µM dexamethasone (Sigma, St. Louis, MO, USA), 0. 5 mM isobutylmethylxanthine (Sigma, St. Louis, MO, USA), 10 µM insulin (Sigma, St. Louis, MO, USA) and 200 µM indomethacin (Sigma, St. Louis, MO, USA). Cells in control group were cultured with α -MEM supplemented with 10% FBS. Medium was changed every three days. After incubation for three weeks, cells were stained by Oil Red O (Sigma, St. Louis, MO, USA) and observed under an inverted microscope. Then, adipogenic differentiation was quantified by an enzyme immunoassay analyzer at 510 nm after the elution with isopropyl alcohol for 10 min at room temperature. Osteogenic differentiation Cells were seeded in a 12-well plate at a concentration of 5 × 10 3 cells /well and cultured with α -MEM supplemented with 10% FBS until 80% confluency. Cells in differentiation group were incubated with osteogenic induction medium consisted of α -MEM supplemented with 10% FBS, 0. 1 µM dexamethasone, 10 mM β -glycerophosphate (Sinopharm Chemical Reagent Ltd. , Shanghai, China) and 200 µM ascorbic acid (Sigma, St. Louis, MO, USA). Cells in control group were cultured with α -MEM supplemented with 10% FBS. Medium was changed every three days. After incubation for three weeks, cells were stained by Alizarin red (Chroma-Schmidt GmbH, Köngen, Germany) and observed under an inverted microscope. Then osteogenic differentiation was quantified by an enzyme immunoassay analyzer at 570 nm after the solubilization with 10% cetylpyridinium chloride (Sigma, St. Louis, MO, USA) for 10 min at room temperature. Quantification of surface markers The surface markers of the cells were examined by flow cytometry. Briefly, 2. 5 × 10 5 cells were rinsed twice with PBS, and then suspended in PBS supplemented with 2% FBS. Cells were stained in the dark for 30 min at 4 °C with the antibodies (BD Biosciences, San Jose, CA, USA): R-phycoerythrin-(PE)-labeled HLA-DR, CD 34, CD45, CD73, CD90 and CD105. The control for PE-coupled antibodies was isotypic mouse IgG1. The data were evaluated using CellQuest software (BD Biosciences, San Jose, CA, USA) and analyzed by FlowJo software. Cell cycle The proportion of cells in different phases of cell cycle was analyzed by cell cycle staining buffer (Multisciences Biotech, Hangzhou City, China). Cells were rinsed twice with PBS and then incubated with buffer at a concentration of 1 × 10 6 cells/ml for 30min in the dark at room temperature. Cell cycle was determined by flow cytometry and analyzed using FlowJo software. Immunosuppressive capacity IDO1 is a rate-limiting enzyme in the kynurenine pathway which plays an important role in the induction of immune tolerance. To assess the immunosuppressive capacity of ADSCs, IDO1 gene expression and IDO1 activity assay were performed. Kyn concentration was evaluated as an index which could reflect IDO1 activity. After storage, cells were seeded in a six-well plate at a concentration of 5 × 10 5 cells/well. After culture for 24 h at 37 °C in a humidified atmosphere at 5% CO 2, culture medium was changed into fresh medium supplemented with 500 U/ml IFN- γ (PeproTech, Rocky Hill, NJ, USA) and then induced for 24 h. In control group, unstored cells were seeded in a six-well plate, and then treated using the same method as stored cells. Cells were used for IDO1 gene expression assay and supernatant was collected for Kyn concentration assay. IDO1 gene expression After incubation for 24 h, total RNA was extracted from ADSCs using RNAiso Plus Kit (Takara, Kusatsu, Shiga, Japan) following the manufacturer’s protocol. Total RNA concentrations were quantified using NanoDrop1000 (Thermo, Waltham, MA, USA). 1 µg total RNA was reserved. Real-time polymerase chain reaction (PCR) was achieved by SYBR green system (Takara, Japan). Amplifications for cDNA samples were carried out at 95 °C for 30 s, followed by 40 cycles at 95 °C for 5 s and at 60 °C for 30 s. Primer sequences were listed in Table 1. The relative quantification of target gene was calculated using the 2 −▵▵ t method and normalized to the transcript levels of glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Melting curve profiles were produced at the end of each PCR to confirm the specific transcriptions of amplification ( Chen et al. , 2016 ). 10. 7717/peerj. 3301/table-1 Table 1 Real-time PCR primer sequences. Gene Forward primer Reverse primer IDO1 CTGGGCATCCAGCAGACT TGAGCTGGTGGCATATATCTTCT GAPDH AACAGCGACACCCACTCCTC CATACCAGGAAATGAGCTTGACAA Kyn concentration A total of 1 ml cell supernatant was mixed with 250 µl 30% trichloroacetic acid (Sinopharm Chemical Reagent Ltd. , Shanghai, China) and the mixture was vortexed and centrifuged at 12, 000 rpm for 20 min at 4 °C. After filtration with 0. 22 µM membrane (Life Sciences, St. Louis, MO, USA), supernatant was analyzed by high performance liquid chromatography LC-20AT (HPLC, Shimadzu, Japan) equipped with a Phenomenex Gemini C18 (250 × 4. 6 mm, 5 µm) column. Mobile phase A was 1 µmol/l potassium dihydrogen phosphate (pH = 4; Sinopharm Chemical Reagent Ltd. , Shanghai, China) and mobile phase B was methanol (Merck, Kenilworth, NJ, USA). The rate of mobile phase A and B was 3:1 and the flow rate was 1 ml/min. Kyn concentration was detected by the UV-detector at a wavelength of 360 nm at room temperature. Statistical analysis All data were shown as mean ± standard deviation, and difference and significance were verified by an one-way analysis of variance (ANOVA), followed by the least-square difference (LSD) for multiple comparisons test. A level of significance of P < 0. 05 was used to indicate statistical differences. The statistical analysis was performed using SPSS 19 (SPSS Inc. , Chicago, IL, USA). Results Part I: evaluation of preservation media Harvested ADSCs were suspended in dextrose, ME + HSA, NS + HSA and GM at a concentration of 1 × 10 6 cells/ml. After storage for 24 h at 2∼8 °C, cells were used for the following research. Unstored cells were used as control. The viability of cells in ME + HSA (95. 88 ± 0. 69%) and NS + HSA (91. 96 ± 1. 53%) were significantly higher than in dextrose (67. 81 ± 6. 37%) ( Fig. 1A, P < 0. 05). Cell viability in dextrose dropped off dramatically after storage for 24 h. The cluster rate of cells in GM (26. 99 ± 1. 84%) was significantly higher than that of cells in ME + HSA (3. 44 ± 0. 81%), NS + HSA (4. 23 ± 2. 46%) and dextrose (3. 82 ± 0. 04%) ( Fig. 1A, P < 0. 05). As shown in Fig. 1A, cells in GM clustered obviously. These results indicated that dextrose, with low cell viability, and GM, with extremely high cluster rate, were not suitable media for the storage of ADSCs. 10. 7717/peerj. 3301/fig-1 Figure 1 Optimization of preservation media. (A) Photos of cells on counting chamber and analysis of viability and cluster rate. White arrow indicated cell cluster. (B) Apoptosis analysis by flow cytometry. (C) Morphology of cells re-plated on 100-mm dish after storage. (D) Photos and analysis of CFU. Results were presented as the means ± standard deviation for n = 3, ∗ P < 0. 05. In our subsequent observation, we found the viability of cells after storage measured by trypan blue staining was not precise and sensitive enough, so we performed Annexin V/PI binding assay ( Fig. 1B ). The proportions of normal cells in dextrose, ME + HSA, NS + HSA and GM all decreased but there were no significant differences among them. The proportions of early stage apoptotic cells and late stage apoptotic cells both increased in dextrose, ME + HSA, NS + HSA and GM, but there were no significant differences among them, respectively. The adhesion ability of ADSCs after storage was observed under an inverted microscope ( Fig. 1C ). Attached cells in the four groups after storage all showed similar spindle-shaped morphologies to cells in the unstored group. However, a mass of detached cells were obviously observed in NS + HSA, which indicated that a lot of cells lost their adhesion ability after storage in NS + HSA. An evaluation of CFU capacity was performed on ADSCs ( Fig. 1D ). All groups could form colonies with >50 cells after culture for 10 days. However, the CFU of cells in ME + HSA (13. 33 ± 2. 05) was significantly higher than that of cells in NS + HSA (2. 40 ± 1. 06). These results indicated that ME + HSA was a better preservation medium than NS + HSA. Based on the study of different preservation media, ME + HSA was selected as a proper preservation medium for high cell viability, low cell cluster rate, good adhesion ability and high CFU capacity. Part II: evaluation of durations of storage ME + HSA was selected as the storage medium for further study. The storage of ADSCs in ME + HSA for durations of 24 h and 48 h at 4 °C at a concentration of 1 × 10 6 cells/ml were studied. Unstored cells were used as control. The cell viability after storage for 48 h (95. 34 ± 4. 72%) was very high and there was no significant difference compared to cells stored for 24 h (98. 11 ± 1. 33%), as data were shown in Fig. S1. The cluster rate was lower after storage for 48 h (7. 98 ± 1. 20%) than after storage for 24 h (15. 06 ± 1. 34%). It seemed that cells could be stored in ME + HSA with high viability. Apoptosis was evaluated at 24 h and 48 h after storage ( Fig. 2A ). The proportion of late stage apoptotic cells increased notably over storage time from 24 h (29. 13 ± 3. 22%) to 48 h (41. 53 ± 1. 15%). However, no significant difference in early stage apoptotic cells was shown over storage time from 24 h (19. 8 ± 4. 16%) to 48 h (21. 3 ± 0. 36%). These results indicated that extending the duration of storage from 24 h to 48 h would accelerate the apoptosis especially from early to late stage apoptosis. 10. 7717/peerj. 3301/fig-2 Figure 2 Optimization of durations. (A) Apoptosis analysis of ADSCs in different durations by flow cytometry. (B) Morphology of cells re-plated on 100-mm dish. (C) CFU of cells in different durations. Results were presented as the means ± standard deviation for n = 3, ∗ P < 0. 05. Although the spindle-shaped morphology of attached cells did not change over the storage time ( Fig. 2B ), there were significantly fewer attached cells following storage for 48 h than for 24 h. The number of cells lost their adhesion ability increased obviously from 24 h to 48 h. After storage for 48 h, cells could form colonies with > 50 cells ( Fig. 2C ); however, the number of these colonies formed after storage for 48 h (8. 67 ± 1. 67) was obviously lower than that for 24 h (17. 07 ± 4. 01). In conclusion, cells could not be stored in ME + HSA for 48 h due to high level of apoptosis, poor adhesion ability and low CFU capacity although viability of cells suspended in ME + HSA for 48 h was very high. 24 h was shown to be an appropriate duration of storage, with relatively low proportion of late stage apoptosis, high adhesion ability and CFU capacity. Part III: evaluation of cell concentrations ADSCs suspended in ME + HSA were stored for 24 h at 2∼8 °C at various concentrations: 1 × 10 6 cells/ml, 5 × 10 6 cells/ml and 10 × 10 6 cells/ml. Unstored cells were used as control. Apoptosis of cells in different cell concentrations was shown in Fig. 3A. The proportion of normal cells decreased obviously as the cell concentration increased from 1 × 10 6 cells/ml (50. 6 ± 3. 66%) to 10 × 10 6 cells/ml (37 ± 0. 75%). Cells at a concentration of 5 × 10 6 cells/ml (30. 40 ± 2. 87%) showed obviously higher level of early stage apoptosis than cells at 1 × 10 6 cells/ml (19. 80 ± 4. 16%) and 10 × 10 6 cells/ml (23. 33 ± 3. 66%) ( P < 0. 05). Cells at concentrations of 1 × 10 6 cells/ml (29. 13 ± 3. 22%) and 5 × 10 6 cells/ml (26. 37 ± 7. 43%) showed lower level of late stage apoptosis than cells at 10 × 10 6 cells/ml (40. 60 ± 3. 78%). These results indicated that cells stored at the concentration of 10 × 10 6 cells/ml would cause the highest level of late stage apoptosis. 10. 7717/peerj. 3301/fig-3 Figure 3 Optimization of cell concentrations. (A) Apoptosis analysis of ADSCs by flow cytometry. (B) Morphology of cells re-plated on 100-mm dish. (C) CFU of cells at different concentrations. Results were presented as the means ± standard deviation for n = 3, ∗ P < 0. 05. Attached cells of all three groups could form spindle-shaped morphology ( Fig. 3B ), and it seemed that there were no significant differences in adhesion ability among cells at three cell concentrations. Cells could form colonies with >50 cells at all three concentrations ( Fig. 3C ). There were no obvious differences in CFU numbers among cells suspended at 1 × 10 6 cells/ml (17. 07 ± 4. 01), 5 × 10 6 cells/ml (13. 00 ± 1. 40) and 10 × 10 6 cells/ml (15. 47 ± 1. 29). These results indicated that cell concentrations during the storage period did not impact on the CFU capacity. Proliferation ability was shown in Fig. 4. The fluorescence value reached the peak at the 7th day in unstored cells and the PDT was 57. 57 ± 4. 77 h. The fluorescence values increased slowly until the 4th day in group 5 × 10 6 cells/ml and group 10 × 10 6 cells/ml, and then increased rapidly until reaching the peak on the 9th day with the PDT of 79. 05 ± 6. 74 h and 81. 07 ± 7. 84 h, respectively. There was no significant difference between group 5 × 10 6 cells/ml and group 10 × 10 6 cells/ml, in terms of either fluorescence values or PDT. As the fluorescence value in group 1 × 10 6 cells/ml increased slowly with the time, it did not reach its peak before we stopped our measurement on the 10th day. These results indicated that cells in group 1 × 10 6 cells/ml had low proliferation potential. All of these three groups at different concentrations showed osteogenic differentiation ( Fig. 5A ) and adipogenic differentiation abilities ( Fig. 5B ). Osteogenic differentiation in group 5 × 10 6 cells/ml (0. 09 ± 0. 01) was slightly higher than in group 1 × 10 6 cells/ml (0. 07 ± 0. 01) or group 10 × 10 6 cells/ml (0. 07 ± 0. 01). Adipogenic differentiation ability in group 5 × 10 6 cells/ml (0. 34 ± 0. 03) was obviously higher than in group 1 × 10 6 cells/ml (0. 18 ± 0. 02) and group 10 × 10 6 cells/ml (0. 16 ± 0. 01) ( P < 0. 05). These results suggested that group 5 × 10 6 cells/ml had the best differentiation potential. 10. 7717/peerj. 3301/fig-4 Figure 4 Proliferation of ADSCs at different cell concentrations. Results were presented as the means ± standard deviation for n = 3, ∗ P < 0. 05. 10. 7717/peerj. 3301/fig-5 Figure 5 Multidifferentiation of ADSCs at different cell concentrations (A) Osteogenic differentiation. (B) Adipogenic differentiation. Results were presented as the means ± standard deviation for n = 3, ∗ P < 0. 05. Level of apoptosis increased when the cell concentration increased from 1 × 10 6 cells/ml to 10 × 10 6 cells/ml. It seemed there were no obvious differences among the three groups in terms of adhesion ability or CFU capacity. Thus, we adopted the assays of proliferation and differentiation as two further evaluation parameters. Proliferation assay of group 1 × 10 6 cells/ml showed that low concentration would slower the growth of ADSCs and cells in group 5 × 10 6 cells/ml showed best osteogenic and adipogenic differentiation potential. These results suggested that 5 × 10 6 cells/ml was a suitable cell concentration for short-term storage. Part IV: evaluation of optimized condition After the selection of preservative media, durations of storage and cell concentrations, we decided to store ADSCs in ME + HSA for 24 h at a concentration of 5 × 10 6 cells/ml in 2∼8 °C. In order to give a comprehensive evaluation of our optimized condition, we studied the surface markers, cell cycle and immunosuppressive capacity of ADSCs after storage. Unstored cells were used as control. After storage, HLA-DR, CD34 and CD45 were all negatively expressed (<2%), while CD73, CD90 and CD105 were all positively expressed (>95%, Fig. 6A ). These results indicated that the storage of this optimized condition did not affect the expression of surface markers. 10. 7717/peerj. 3301/fig-6 Figure 6 Evaluation of optimized solution. (A) HLA-DR, CD34, CD45, CD73, CD90, CD105 expression. (B) Cell cycle distribution. (C) IDO1 gene expression by RT-PCR. (D) Kyn concentration by HPLC. Results were presented as the means ± standard deviation for n = 3. Flow cytometric analysis of cell cycle distribution was shown in Fig. 6B. After the storage, there were no significant differences in G0/G1, S and G2/M compared with unstored cells, respectively. These results indicated that this optimized condition did not change the cell cycle distribution. There was no significant fold difference in IDO1 gene expression between stored cells (3393. 54 ± 653. 65) and unstored cells (3654. 41 ± 136. 30, Fig. 6C ). Also, there were no obvious differences in Kyn concentration between unstored cells (25. 24 ± 1. 75) and stored cells (27. 45 ± 2. 31, Fig. 6D ). These results indicated that this optimized condition did not change gene expression and activity of IDO1. Discussion Adipose tissue was considered to be merely a passive energy store in previous years, before ADSCs could be isolated from adipose tissue as a new source of stem cells in 2001 ( Zuk et al. , 2001 ). ADSCs are multipotent cells with the ability to differentiate into both mesodermal and non-mesodermal lineages, similar to bone marrow-derived stem cells (BM-MSCs). In addition, ADSCs have a great number of advantages over BM-MSCs. ADSCs could be collected in large quantity with minimal morbidity ( Uzbas et al. , 2015 ), and derivation of ADSCs is easier (less invasive) and much more efficient than that of BM-MSCs. Thus ADSCs are attractive stem cells for regenerative application. In order to manufacture a clinical-scale large number of ADSCs for cell therapy in regenerative medicine and tissue engineering, strictly quality control is required which means that a cGMP-compliant clean room is needed. It seems impossible to build an expensive cGMP-compliant clean room in the hospital set-up, thus cell products should be produced in a central laboratory for up-scaling cells and then transported to the bedside of the patient ( Pal, Hanwate & Totey, 2008 ). Although cryopreservation is an alternative for long-term storage of ADSCs, its requirement for toxic cryoprotectants (i. e. , DMSO) and low recovery rate of cells demonstrated that it is not the best or the safest condition for the short-term storage of cell products ( Chen et al. , 2013 ; Grein et al. , 2010 ). There is no standard protocol for short-term storage of fresh cells before transplantation as well as no relatively comprehensive evaluation of cell quality after storage. Previous study reported that MSCs stored in saline or dextrose for more than 2 h lost cell viability significantly. MSCs lost CFU capacity and differentiation ability rapidly as storage time increased. Thus duration of storage was limited to 2 h to ensure the quality of MSCs ( Sohn et al. , 2013 ). Although Plasmalyte A, 1% HSA and 5% HSA were FDA approved injections and are typically used as preservation media prior to MSCs transplantation, none of these single components supported the survival of MSCs ( Chen et al. , 2013 ). No duration information was given in a clinical trial of MSCs stored in saline for the treatment of ischemic stroke ( Bang et al. , 2005 ). Cell concentration was considered to have an impact on cell quality, but researchers only measured MSCs counts among cells at the concentrations over a range of 0. 5 × 10 6 –20 × 10 6 cells/ml as the evaluation of cell quality ( Lane et al. , 2009 ). Other researchers reported few or no details about preservation conditions or the effect of short-term storage on MSCs ( Li et al. , 2002 ; Li et al. , 2005 ; Kim et al. , 2006 ; Shen et al. , 2006 ). The primary objective of this study was to optimize preservation media, durations of storage and cell concentrations of ADSCs to provide a feasible short-term storage condition for cell therapy. Results showed that cells formed clotted cell pellet after storage in GM for 24 h, which confirmed the natural preference of MSCs to form aggregates ( Potapova et al. , 2008 ). Thus it would be unsafe clinically to inject cells suspended in GM. In addition, FBS in GM remains associated with safety issues including transmission of viral disease, anaphylactic reactions and production of anti-FBS antibodies ( Ikebe & Suzuki, 2014 ; Mackensen et al. , 2000 ; Sundin et al. , 2007 ). Therefore, GM is not a suitable preservation medium for the short-term storage of ADSCs. When stored in dextrose, dark blue stained cells could be seen obviously. The viability of ADSCs suspended in dextrose decreased to 67. 81 ± 6. 37%, which was lower than the minimum viability (70%) acceptable by FDA for cell therapy. The deterioration of ADSCs survival in dextrose may be caused by the low-pH level (3. 2–5. 5). Additionally, high concentration of 5% dextrose (50 g/L) affects regenerative potential of MSCs and induces replicative senescence ( Chen et al. , 2013 ). The adhesion ability and CFU capacity of ADSCs in NS + HSA were obviously lower than in ME + HSA respectively, despite the fact that the viability and apoptosis of cells in NS + HSA had no significant differences compared to cells in ME + HSA respectively. Thus, a comprehensive evaluation system was needed to test the quality of cells before administration as rapid detection of viability and apoptosis would not always be reliable parameters. The mechanism of different effects on cell quality between ME + HSA and NS + HSA was not clear. We supposed that pH of two media may cause the difference as ME has physiological pH (7. 4) while NS has lower pH (4. 5–7. 0). Also the sodium gluconate ingredient in ME could provide energy for cell survival. Another point to discuss is the proper duration of short-term storage. We thought 24 h was enough for the transport of cell products to another city thousands of kilometers away and for the preparation of both patients and doctors. We also wanted to investigate the longest duration with >70% cell viability. Results showed that the viability of cells in 48 h was high (>90%), however, the late stage apoptosis rate in 48 h was also high (41. 53 ± 1. 15%). The adhesion ability was poor and CFU capacity was low (8. 67 ± 1. 67). High level of late stage apoptosis, poor adhesion ability and low CFU capacity showed that cells could not adequately be stored for 48 h. These results may be a consequence of reduced nutrient supply and increase in waste and lactic acid accumulation during prolonged storage ( Robinson, Picken & Coopman, 2014 ). The late stage apoptosis of cells (29. 13 ± 3. 22%) stored for 24 h was acceptable, and the CFU number (17. 07 ± 4. 01) of cells stored for 24 h was relatively high. All these results suggested that cells suspended in ME + HSA could be stored for 24 h before administration. Cell concentration during the storage is also an important factor which may affect cell quality. Compared to intravenous injection, subcutaneous injection and intramuscular injection require higher concentrated cell products with smaller volumes. As previous study has reported ( Espina et al. , 2016 ), MSCs suspension injected into small tendinous lesions would leak outside the defect and into the peritendinous tissue even the volume of cell suspension was only 1 ml. Thus cell products of high cell concentration are needed. We compared three cell concentrations 1 × 10 6 cells/ml, 5 × 10 6 cells/ml and 10 × 10 6 cells/ml. Late stage apoptotic rate increased as the cell concentration increased. We didn’t observed significantly differences among these three groups in terms of adhesion ability and CFU capacity. Thus, we adopted the evaluation of proliferation and differentiation. The proliferation of cells stored at a concentration of 5 × 10 6 cells/ml was very close to that of cells stored at 10 × 10 6 cells/ml and significantly faster than that of cells stored at 1 × 10 6 cells/ml. Cells stored at 5 × 10 6 cells/ml showed the best osteogenic and adipogenic differentiation potential. The high level of late stage apoptosis of cells stored at 10 × 10 6 cells/ml may be explained that the highest cell concentration may be associated with the fastest lactic acid accumulation ( Kilkson, Holme & Murphy, 1984 ). To some extent, decreasing cell concentration to limit lactic acid accumulation could enhance cell viability and improve cell function ( Kao, Kim & Daley, 2011 ). Long-term proliferation kinetics results indicated that the proliferation potential of cells stored at 1 × 10 6 cells/ml was impaired. Although increasing cell concentration could improve the proliferation potential, it seemed 5 × 10 6 cells/ml was high enough as the curve of it was very close to that of 10 × 10 6 cells/ml. Osteogenic and adipogenic differentiation results also suggested that 5 × 10 6 cells/ml was a suitable cell concentration. As previous study have reported the same cell concentration of 5 × 10 6 cells/ml as our result for cell therapy ( Garvican et al. , 2014 ; Godwin et al. , 2012 ), we thought cells suspended at 5 × 10 6 cells/ml could have the highest quality among these three concentrations. After the evaluation of preservation media, durations of storage and cell concentrations, we thought cells suspended in ME + HSA at a concentration of 5 × 10 6 cells/ml could be stored for 24 h at 2∼8 °C before administration. We then evaluated this condition by studying surface markers, cell cycle and immunosuppressive capacity of ADSCs after storage. Results showed that surface markers, cell cycle and immunosuppressive capacity did not change after the storage, which confirmed our result that this optimized condition has a great potential for the short-term storage of MSCs for cell therapy. Conclusions This is the first study to comprehensively optimize a short-term storage condition of ADSCs. Key factors during short-term storage including preservation media, durations and cell concentrations are studied. Comprehensive evaluation is needed to reflect real cell status before transplantation. Our results show that ADSCs suspended in ME + HSA, at a concentration of 5 × 10 6 cells/ml could be stored for 24 h at 2∼8 °C, which provides a reliable short-term storage condition for cell therapy. Future studies are still needed to improve cell viability, extend duration of storage, and verify the therapeutic effect of ADSCs after short-term storage in vivo. Supplemental Information 10. 7717/peerj. 3301/supp-1 Figure S1 Photos of ADSCs on counting chamber and analysisof viability and cluster rate in different durations Results were presented as the means ± standard deviation for n = 3, ∗ P < 0. 05. Click here for additional data file. 10. 7717/peerj. 3301/supp-2 Data S1 Raw data Click here for additional data file. 10. 7717/peerj. 3301/supp-3 Supplemental Information 1 IDO1 gene Click here for additional data file. 10. 7717/peerj. 3301/supp-4 Supplemental Information 2 GAPDH gene Click here for additional data file.

10. 7717/peerj. 3498

PeerJ

Microplasma-assisted hydrogel fabrication: A novel method for gelatin-graphene oxide nano composite hydrogel synthesis for biomedical application

Toxicity issues and biocompatibility concerns with traditional classical chemical cross-linking processes prevent them from being universal approaches for hydrogel fabrication for tissue engineering. Physical cross-linking methods are non-toxic and widely used to obtain cross-linked polymers in a tunable manner. Therefore, in the current study, argon micro-plasma was introduced as a neutral energy source for cross-linking in fabrication of the desired gelatin-graphene oxide (gel-GO) nanocomposite hydrogel scaffolds. Argon microplasma was used to treat purified gelatin (8% w/v) containing 0. 1∼1 wt% of high-functionality nano-graphene oxide (GO). Optimized plasma conditions (2, 500 V and 8. 7 mA) for 15 min with a gas flow rate of 100 standard cm 3 /min was found to be most suitable for producing the gel-GO nanocomposite hydrogels. The developed hydrogel was characterized by the degree of cross-linking, FTIR spectroscopy, SEM, confocal microscopy, swelling behavior, contact angle measurement, and rheology. The cell viability was examined by an MTT assay and a live/dead assay. The pore size of the hydrogel was found to be 287 ± 27 µm with a contact angle of 78° ± 3. 7°. Rheological data revealed improved storage as well as a loss modulus of up to 50% with tunable viscoelasticity, gel strength, and mechanical properties at 37 °C temperature in the microplasma-treated groups. The swelling behavior demonstrated a better water-holding capacity of the gel-GO hydrogels for cell growth and proliferation. Results of the MTT assay, microscopy, and live/dead assay exhibited better cell viability at 1% (w/w) of high-functionality GO in gelatin. The highlight of the present study is the first successful attempt of microplasma-assisted gelatin-GO nano composite hydrogel fabrication that offers great promise and optimism for further biomedical tissue engineering applications.

Introduction Tissue engineering is an emerging field that exists at the interface of material science, chemical engineering, and life science to develop alternatives to restore, improve and maintain diseased or damaged tissues ( Lanza, Langer & Vacanti, 2011 ); thus, it represents a fascinating trend in regenerative medicine. Broadly, the therapeutic approach in orthopedic tissue engineering focuses on the regeneration of a variety of connective tissues such as bone, cartilage, ligament, tendons, and muscle tissues ( Lu & Thomopoulos, 2013 ). Amongst the main challenge in orthopedic tissue engineering are the selection of appropriate cells (differentiated or progenitor cells) followed by fabrication and utilization of biocompatible and mechanically suitable scaffolds with enhanced potential to target major unresolved issues from the past ( Kuo et al. , 2006 ). Despite the intrinsic capability of connective tissues in the body to regenerate, they fail to regenerate themselves during injury or some diseases that ultimately lead to the loss of, or damage to, connective tissues. Connective tissue degeneration is one of the most common causes of pain, limited movement, deformity, and eventually progressive disability if not treated in time. Traditional surgical reconstruction fails to fully repair lost connective tissues and often causes donor site morbidity ( Cezar & Mooney, 2015 ). Recently, polymer-based scaffold fabrication gained popularity in tissue engineering during scaffold designing ( Piskin, 1995 ; Ji et al. , 2006 ) for repairing and regenerating desired tissues. Polymeric hydrogels, due to their unique biocompatibility and desirable physical characteristics, have a long history of use as scaffold material of choice for tissue engineering. Besides serving as matrices for tissue engineering and regenerative medicine, these polymer based hydrogels are capable of mimicking the extracellular matrix topography and can thus facilitate the delivery of required bioactive agents that promote tissue regeneration ( Amini, Laurencin & Nukavarapu, 2012 ; Park, 2011 ). Gelatin is a suitable polymer that has been extensively used in tissue engineering hydrogel scaffolds fabrication due to its high viscosity, density, excellent biocompatibility and tunable properties ( Golden & Tien, 2007 ). More interestingly, gelatin-based materials due to the presence of arginine-glycine-aspartic acid (RGD) adhesion peptide sequences are promising scaffolds for cell-based repair and facilitate cellular attachment, proliferation, and growth ( Shin, Jo & Mikos, 2003 ) with better biocompatibility ( Elzoghby, 2013 ; Hersel, Dahmen & Kessler, 2003 ; Ladage et al. , 2011 ); and have already been approved by the US Food and Drug Administration (US-FDA) for clinical use. The polymerization of gelatin occurs at mild conditions (room temperature, neutral pH, in aqueous environments) which facilitate cross-linking and hydrogel formation ( Aubin et al. , 2010 ; Nichol et al. , 2010 ). However, limitations of gelatin, such as poor mechanical strength and easy to get contaminate, are needed to be addressed to make it an ideal scaffold material for tissue engineering. Recently, graphene oxide (GO) has been gaining popularity as additive material along with biopolymers to improve their biocompatibility, mechanical strength, cell adhesion, and proliferation properties, specifically for various tissue engineering applications ( Chang et al. , 2011 ). When mixed with gelatin, due to presence of oxygen-containing hydrophilic groups in GO aromatic chains reduces the irreversible agglomeration of graphite sheets through π–π stacking and Van der Waals interactions for ease of making homogenous dispersions with gelatin solutions ( Ge et al. , 2012 ; Liu et al. , 2011 ). The effective fabrication of gelatin-GO based hydrogels is still lacking due to less well known mechanisms of cross-linking behind it. To date, chemical cross-linking by methylacrylate, glutaraldehyde, carbodiimide etc. are established chemical cross-linking techniques ( Barbetta et al. , 2006 ; Farris, Song & Huang, 2009 ; Park et al. , 2002 ). However, physical cross-linking methods may provide easier and non-toxic way over chemical methods to obtain tunable cross-linked polymers ( Rowan et al. , 2002 ). Plasma-induced ( Gomathi, Sureshkumar & Neogi, 2008 ) (highly energetic fourth state of matter) cross-linking could be a better choice than chemical cross-linking methods due to its nontoxic chemical free nature. Also, the shorter duration of cross-linking process by plasma technique ( Kitching, Pan & Ratner, 2004 ) could make it novel and unique process for scale-up during commercial industrial applications and could also provide sterile products with better flexibility ( Ohl & Schröder, 1999 ) for clinical applications than those of non-plasma techniques. Furthermore, no study has been reported using plasma for the cross linking and fabrication of Gelatin-GO based nano composite hydrogel systems. Lesser dimensions microplasmas have been reported as useful tools for materials synthesis and processing previously. Confining the plasma to a micronscale leads to its stability at atmospheric pressure ( Mariotti & Sankaran, 2010 ), making microplasma easy to implement and highly desirable tool for industrial applications. The generated gas discharges contain a high density of energetic electrons (>10 eV) that allows efficient material synthesis and processing. The previous reports also have demonstrated the feasibility of the microplasma-based process to produce metal ( Chiang & Sankaran, 2009 ; Chiang & Sankaran, 2008 ) and semiconductor nanoparticles ( Sankaran et al. , 2005 ), oxides ( Mariotti, Bose & Ostrikov, 2009 ), and carbon nanostructures ( Ghezzi et al. , 2014 ; Luo et al. , 2016 ). To the present, in the biomedical field, plasma research has been used for various applications such as surface sterilization ( Kvam et al. , 2012 ), promotion of hemostasis ( Schmidt et al. , 2015 ), enhancement of tissue regeneration ( Lee et al. , 2015 ), acceleration of wound healing ( Arjunan & Clyne, 2011 ), and anticancer therapy ( Fridman et al. , 2008 ; Schlegel, Köritzer & Boxhammer, 2013 ). However, more research is needed to establish plasma induced hydrogels synthesis and its successful implementation in biomedical tissue engineering (with or without cell based therapy) scaffold fabrication for tissue regeneration and repair. In the present study, we report that inert argon (Ar) microplasma treatment could be beneficial for modifying and reorganizing chemical groups in gelatin polymers for cross-linking and production of biomimetic nanocomposite gel-GO hydrogel system, which in turn would reduce the adversity of traditional chemical and other methods of cross-linking and polymerization (schematically illustrated as per Fig. 1 ). 10. 7717/peerj. 3498/fig-1 Figure 1 Schematic presentation of Ar- microplasma mediated gel-GO nano composite hydrogel synthesis and its’ biomedical applications (PRISMA flow diagram). Ar- microplasma helps in formation of gel-GO hydrogel system by free radical initiated molecular interaction between the polymer gelatin and graphene oxide resulting cross-linking and polymerization in a safe and tunable way intended for biomedical applications. The objectives of this work are to optimize, formulate, characterize, and evaluate gel-GO nano composite hydrogel system for its biocompatibility by systematic material characterization methods such as cross-linking index measurement, scanning electron microscopy (SEM), rheology, swelling behavior, Fourier transformation infrared (FTIR) spectroscopy, contact angle measurement, 3-[4, 5-dimethylthiazol-2-yl]-2, 5-diphenyl tetrazolium bromide (MTT) assay, microscopy, and a live/dead assay. Our study in resonance with Ar microplasma could be a useful tool for gel-GO nanocomposite hydrogel scaffold synthesis for tissue repair and regeneration. Further, it may prove to be a better approach for rendering an appropriate biomimetic scaffold designing platform for tissue engineering preventing chemical toxicity and related adverse effects. Materials and Methods Gelatin purification Gelatin type B, isolated from bovine skin, was purchased from Sigma-Aldrich (St. Louis, MO, USA). Gelatin samples with an approximate isoelectric point of 5 and Bloom strength of 225 were used. Gelatin type-B powder was dissolved in distilled water for cross-linking the material to preserve the hydrogel structure. Various concentrations of gelatin solutions were prepared (7%, 8%, 9%, and 10%) in double-distilled water at 50 °C with continuous stirring for 30 min. Gelatin was purified to exploit the large number of functional side groups. We used acetone (Sigma-Aldrich, St. Louis, MO, USA) as a desolvating agent by combining it with the dissolved gelatin solution in this study in a ratio 1:1. The supernatant was discarded and the high-molecular-weight gelatin was re-dissolved by adding an equal volume of distilled water and stirred at 400 rpm at 40∼50 °C. The pH of the gelatin solution (5. 7) was adjusted to near-neutral values of 7. 4 in consideration of the biomimetic property of scaffold for body adaptability. Preparation of GO (high functionality) Graphite (−325 mesh, 99. 995% pure) microcrystalline powder was purchased from Alfa Aesar, USA. Potassium permanganate (KMnO 4, 98%), Hydrogen peroxide (H 2 O 2, 35%), and Ether [(C 2 H 5 )2O, 99 + %] were obtained from ACROS, Belgium. Potassium nitrate (KNO 3, 95%) was purchased from JT-Baker (Center Valley, PA, USA). Hydrochloric acid (HCl, 37%) and Sulfuric acid (H 2 SO 4 > 95%) were purchased from Scharlau, Spain. The graphene oxide used in this study was synthesized by a modified Hummer’s method. Similar details of the preparation can be found elsewhere ( Wang et al. , 2015 ). Briefly, 0. 1 g of graphite was added in 10 mL of H 2 SO 4 containing KNO 3 (1 g) and magnetically stirred (300 rpm, 2 hrs) until a visually homogeneous dark gray solution formed. Then, KMnO 4 (0. 5 g) was slowly added to the previously formed solution and further stirred for 2 hrs at room temperature. After that, the solution mixture was put in water bath (IKA-HS7 digital; IKA Works, Staufen, Germany) at temperature 70 °C for 2 hrs. 200 g of composite precipitate mixture was removed from water bath, allowed to cool to room temperature by 350 g ice containing 5 ml of 35% H 2 O 2 (to prevent precipitation of insoluble MnO 2 ). Then, the mixture was centrifuged at 24, 500 rpm for 30 min to get crude GO. The crude GO was then bath sonicated in 60 mL deionised water for 30 min followed by bath-sonication in 30 mL HCl and in 60 mL ether for 30 min. Finally, purified GO solid mixture was obtained by centrifugation of the dispersion mixture (24, 500 rpm, 30 min) and was dried ( Li et al. , 2016 ; Kosynkin et al. , 2009 ). GO characterizations Ex situ characterization methods for natural graphite and as-produced GO were including X-ray diffraction (XRD), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), and micro Raman spectroscopy. For XRD, the dried GO powder was used as sample. The XRD was performed by BRUKER D2 PHASER- X-ray Powder Diffractometer (Bremen, Germany) (Cu Kα, λ = 1. 54Å). For TEM, samples were prepared by dispersing GO in ethanol and then dropped onto 300 mesh holey lacy carbon grids on cupper support (Ted Pella, Inc. , Redding, CA, USA) at ambient condition. The TEM images were observed by Hitachi H-9500 system, Japan. For XPS and Raman, the samples were prepared by dispersing GO in ethanol. Then thin sample films were prepared on Si wafer and dried in a hot air oven at 60 °C. XPS (VG ESCALAB 250; Thermo Fisher Scientific, Waltham, MA, USA) was performed using a monochromatic Al Kα X-ray radiation (10 kV and 10 mA). The source power was set to 72 W, and pass energies of 200 eV for survey scans and 50 eV for high-resolution scans. Raman scattering studies were performed at room temperature with a JASCO 5100 spectrometer (533 nm; JASCO, Tokyo, Japan). GO encapsulation into a gelatin matrix High-functionality GO was weighed and grinded into a fine powder with a mortar and pestle. The finely powdered GO was added to distilled water and sonicated in ultrasonic water bath (Elmasonic P, 110 V, 720 W, 7 A; GmbH & Co, Weißenburg, Germany) for 30 min at 50 °C to obtain a uniformly dispersed solution. Dispersion of GO in biological media often requires surfactant stabilization or sonication to prevent aggregation ( Ge et al. , 2012 ). Various concentrations of gelatin-GO solutions were prepared by adding the GO solution drop-wise into different concentrations of previously melted purified gelatin solutions. For proper mixing and composite solution formation by covalent bonding between the GO and gelatin, the composite solutions were again sonicated for 1 h under above-described conditions. After sonication, we obtained uniformly dispersed GO in gelatin solution. The pH value was again measured and adjusted to a near-neutral value (7. 4). Synthesis of gel-GO hydrogels by Ar microplasma Argon microplasma was used here as a physical cross-linking tool for the gel-GO composite modification to form hydrogels. The anode is a platinum (Pt) foil (99. 95%; Alfa Aesar, Ward Hill, MA, USA), which was immersed into the gel-GO solutions. The cathode is a stainless-steel capillary tube (with an inner diameter of 178 µm), which is also the gas inlet to create direct-current microplasma. Optimization of the argon microplasma process and gelatin and GO concentrations were the key factors in the entire process of gel-GO hydrogel fabrication along with various parameters including current, voltage, gas flow rate, conductivity, time of treatment, etc. , were thoroughly investigated periodically to obtain the desired scaffold. The current and voltage are interdependent and affect the ionization of the plasma gas as well as free radical production ( Bunshah & Deshpandey, 1985 ). It is important to maintain a steady current and voltage for uniform glow plasma discharge. Further, due to the conductive nature of gel-GO solutions, we used a resistor (300 W, 150 KΩ) and copper mesh as a barrier (to avoid surface burning) to stabilize the treatment process for the optimized microplasma conditions (2, 500 V, 8. 7 mA, 15 min, and a gas flow rate of 100 standard cm 3 /min). After plasma treatment, the resulting material gel-GO nanocomposite hydrogel was formed. Characterization of the gel-GO hydrogel Degree of cross-linking The degree of cross-linking was determined by a Ninhydrin assay, which is a direct way to determine the amount of free amino groups of untreated and plasma-treated gel-GO samples. The test sample was weighed and boiled with a Ninhydrin solution for 20 min. After that, the solution was cooled to room temperature, 95% ethanol was added, and the optical absorbance of the solution was recorded with a UV-visible spectrophotometer (Thermo Fisher Scientific, Rockford, IL, USA). At 570 nm using glycine at various concentrations as the standard. The amount of free amino groups in the gelatin before plasma treatment (Ci) and after (Cf) cross-linking is proportional to the optical absorbance of the solution. The degree of cross-linking of the various concentrations of gelatin were calculated as per Eq. (1). Results were the average of five independent measurements. (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\text{Cross-linking index}(\text{%})=(\mathrm{Ci}-\mathrm{Cf})/\mathrm{Ci}\times 100\text{%}. \end{eqnarray*}\end{document} Cross-linking index % = Ci − Cf ∕ Ci × 100 %. Surface morphology by SEM Gel-GO hydrogels were prepared for SEM after lyophilization for 72 h. Small pieces of the hydrogel discs were cut off and mounted onto stubs using double-sided adhesive tape, and then gold-coated in a sputter coater (Hitachi E-1010; Tokyo, Japan) at 20 mA, 9 Å for 90 s. The cross-section morphologies of the gelatin discs were examined using a Hitachi S-3500 SEM with an accelerating voltage of 15 kV. Fifteen different pores were randomly selected, and the average pore diameters were calculated. Results of five independent runs were averaged. Spectral change observation by FTIR spectroscopy The FTIR spectra for all samples were obtained in KBr pellets using a Perkin-Elmer Precisely-FTIR spectrophotometer (Melville, NY, USA) in transmission mode at a wavelength range of 400∼4, 000 cm-1 Gel-GO nanocomposite visualization by confocal microscopy For three-dimensional, high-resolution, non-destructive imaging of the gel after treatment, confocal microscopy was used to visualize the colloidal structure of the gel-GO composite. Confocal microscopy offers several advantages over conventional wide-field optical microscopy, including the ability to control the depth of field, reduction or elimination of background information away from the focal plane (that leads to image degradation), and the capability to collect serial optical sections from thick specimens. Furthermore, confocal microscopy X–Z plane imaging was used to analyze morphological changes of the gelatin matrix with or without GO. Test samples were evaluated on a confocal dish using confocal laser scanning microscopy {MitoTracker Red 580 (Invitrogen [Molecular Probes], Gibco, Carlsbad, CA, USA} to check the structural changes. Swelling behavior To study the swelling behavior, untreated gel-GO nanocomposites and microplasma-treated hydrogels were immersed in deionized water at 37 °C. Samples were taken out from deionized water at selected time intervals, wiped with tissue paper to remove surface water droplets, and weighed further. Wet and dry weights were measured and considered to evaluate the swelling ratio. The swelling ratio ( S ) was calculated using the following equation: (2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}S=({W}_{t}/{W}_{0});\end{eqnarray*}\end{document} S = W t ∕ W 0 ; where, W t is the weight of the swollen sample at a certain time point, and W 0 is the initial weight of the sample. Rheological analyses All rheological experiments were conducted using a Thermo Scientific™ HAAKE RheoStress 1 rotational cone plate rheometer with an angular resolution of 300 nanorad and a low-inertia drag cup motor. The plate diameter used was 15 mm. The sample was placed on horizontal plate and a shallow cone placed into it. The angle between the surface of the cone and the plate was around 1 degree. The storage modulus (G′) and loss modulus (G″) were evaluated for gelatin w/o GO, untreated and plasma-treated gel-GO hydrogel samples in a temperature range of 7∼60 °C to check the gel strength at general physiological conditions specifically at 37 °C. The viscoelasticity was observed at a frequency 1 Hz for up to 3, 000 s. The sol-to-gel and gel-to-sol state changes were correlated with an ideal biomimetic scaffold for orthopedics and specifically for soft tissue engineering fields with tunable property. Water contact angle for hydrogel hydrophilicity Water contact angles of untreated and microplasma-treated hydrogels were measured using the DIGIDROP-GBX contact angle measurement system (Bourg-de-Peage, France). The sample materials were cut into pieces ( n = 3), then deionized water droplet was gently deposited on each sample through a micro-syringe, images were captured for up to 30 s after the water droplet was dropped on the surface of the material, and the average contact angle was measured. 3-[4, 5-Dimethylthiazol-2-yl]-2, 5-diphenyl tetrazolium bromide (MTT) assay of cell proliferation The MTT cytotoxicity assay is a standard method prior to detailed in vitro study. Equal sizes, weights, and volumes (5 mm × 5 mm × 5 mm), (0. 3 g), (300 µL)) of untreated and treated hydrogel scaffold materials in triplicate were incubated in 2 mL Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Sigma-Aldrich) with 10% v/v of fetal bovine serum in 24 well cell culture plates at 37 °C, in 5% CO 2 and fully humidified air for 1, 3, 5, 7, 10, and 14 days. The liquid extraction medium (10 µL each) were collected for MTT assay. Here, we used an indirect method of measurement to assess cell viability. The L929 mouse fibroblast cell line (ATCC (NCTC clone 929; L cell, L-929 derivative of Strain L; ATCC CCL-1)) with an initial density of 5, 000 cells/cm 2 per well at a volume 100 µL were seeded in 96-well plates and were cultured for 24–48 h for proper attachment and growth. Then, periodically collected liquid extracts (10 µL each) were added to each well and incubated for 24 h. Finally, the cell metabolic activity was determined using a MTT cell proliferation kit (Sigma-Aldrich, Gibco, St. Louis, MO, USA). Further, the microplsma treated hydrogel system was compared with traditional genipin crosslinked gelatin hydrogel as control group for cytotocity assay by MTT assay with same experimental conditions as mentioned previously for untreated and microplasma treated gelatin hydrogel. Microscopy to observe cellular and hydrogel interactions Microscopic surface images of MG-63 cells seeded on various samples of gels and hydrogels for 24 h were observed under an inverted microscope (Nikon Eclipse TE2000-U Inverted Microscope) to clearly visualize cellular growth and proliferation as direct method for primary biocompatibility study. Live/dead assay The viability of cells on the surface of the gel-GO was assessed by live/dead staining. In brief, MG-63 cells were seeded directly onto the surface of gelatin, microplasma-treated gelatin, untreated gel-GO composite gels, and microplasma-treated gel-GO hydrogels in 6-well plates. After 24 h of incubation, cells/specimens were rinsed three times with phosphate-buffered saline (PBS). The cell laden hydrogels were cryosectioned. The slides were then incubated with live/dead stain (2 mM calcein AM and 4 mM ethidium homodimer-1) (Thermo Fisher Scientific, Dreieich, Germany) for 30 min at room temperature (RT). Viable cells (green) and dead cells (red) were counted under a fluorescence microscope (Olympus IX70, Tokyo, Japan). 10. 7717/peerj. 3498/fig-2 Figure 2 XRD, Raman, XPS and TEM for GO characterization. (A) XRD patterns of natural graphite and as-produced GO. (B) Micro Raman spectra of natural graphite and as-produced GO. (C) C1s peak of GO in the HRXPS spectrum. C1s peak was deconvoluted to C–C, C–O, C=O, and COOH surface functionalities at 284. 4, 286, 287, and 289 eV, respectively. (D) TEM image of GO. Statistical analysis Statistical analyses of all correlated data along with the MTT cytotoxic assay of the untreated gel-GO and microplasma-treated gel-GO hydrogel samples were analysed with the help of Graph pad prism software preferably by a two-way analysis of variance (ANOVA). Each experiment was independently performed and duplicated. Differences were considered significant at the p < 0. 05 level. Results GO characterization After GO synthesis, it was charactrerized by XRD, Raman, XPS and TEM as shown in Fig. 2. The microstructure of natural graphite and the as-produced GO was obtained from the XRD characterization. In the XRD patterns, the natural graphite showed a peak at ∼26. 3°( Fig. 2A ), attributed to the (002) plane of the interplanar graphite with a d spacing of 0. 34 nm according to the Bragg’s Law ( Futaba et al. , 2011 ). For the GO, a characteristic peak of oxidized GO structure at ∼9. 4°was observed ( Kosynkin et al. , 2009 ), suggesting the existence of a mixture of graphite and oxidized GO. The calculated d spacing was increased, which was due to the functional groups generated between the adjacent layers of as-produced GO during the process of oxidation. Further structural information of the as-produced GO was obtained from micro Raman spectroscopic characterization. As shown in Fig. 2B, the representative Raman spectra of natural graphite and as-produced GO. Three bands were observed around 1, 349 cm −1, 1, 596 cm −1, and 2, 679 cm −1, which were respectively assigned to the D-band, G-band, and 2D-band of carbon. The surface functionalities of the as-produced GO were further studied by XPS. Figure 2C and the C1s peaks of HRXPS spectra of natural graphite and as-produced GO were found to be deconvoluted to several peaks at 284. 4, 286, 287, and 289 eV, corresponding to sp 2 C–C, C–O, C=O, and COOH surface functionalities respectively. The results suggested that the as-produced GO possessed oxygen-containing functional groups in agreement with previous XRD and Raman results ( Cataldo et al. , 2010 ; Higginbotham et al. , 2010 ). Further, the topological TEM image of high functionality graphene oxide was shown in Fig. 2D. 10. 7717/peerj. 3498/fig-3 Figure 3 Cross-linking index. Cross-linking index of various concentrations of purified gelatin. An asterisk indicates statistically significant difference of 8% purified gelatin group cross-linking ( ∗ P > 0. 05 and n = 5) as compared with other groups (7%, 9%, 10%). Gel-GO hydrogel characterization Cross-linking index/degree of cross-linking A cross-linking index measurement here was used as an analytical tool to survey additional modifications to the initial polymer gelatin. Here, 8% purified gelatin during microplasma treatment achieved the highest degree of cross-linking of 61 ± 2% ( Fig. 3 ). Thus, in view of the uniform distribution of plasma energy into the composite solution and to avoid surface blocking and burning of a dense solution, 8% purified gelatin was selected with the optimal viscosity and density suitable for the desired gel-GO nanocomposite tissue engineering scaffold fabrication. Morphology and pore size analysis Detailed images of the morphology of the samples were observed by SEM. As per SEM images, the pore size of the 8% purified gel-GO hydrogel formed after plasma treatment was found to be better 287 ± 27 µm ( Figs. 4 and 5 ) among all the groups, adequate for cellular penetration and proliferation. 10. 7717/peerj. 3498/fig-4 Figure 4 Scanning electron microscopy. Morphological observations and representative scanning electron microscopic images of various concentrations of gelatin—graphene oxide nanocomposite material before (A) 7%, (B) 8%, (C) 9%, (D) 10% and after cross-linking (E) 7%, (F) 8%, (G) 9%, (H) 10% showing 8% purified gelatin achieving highest pore size 287 ± 27 µm after microplasma treatment. Scale bars: 100 µm. 10. 7717/peerj. 3498/fig-5 Figure 5 Pore size. SEM characterization of porous structure of Gelatin. Pore size of various concentrations of gelatin- graphene oxide nano composite before and after Ar-microplasma treatment. An asterisk indicates statistically significant differences ( ∗ P < 0. 05; n = 5) between the non-cross-linked and cross-linked groups for each concentration of gelatin groups. Further, in Fig. 6, GO encapsulation was clearly visualized in the gel-GO nano composite hydrogel matrix system. The SEM images showed that the graphene oxide nano particles were embedded in the polymeric gelatin matrix system. 10. 7717/peerj. 3498/fig-6 Figure 6 SEM of gel-GO nano composite. Scanning electron micrographic images of plasma treated nano graphene oxide encapsulated in gelatin matrix observed at various magnifications (lower to higher magnification from A to D). FTIR Analysis FTIR spectroscopy offers a vast assay of analytical tools. Different spectral changes during cross-linking and modification were observed by FTIR spectroscopy in our study. The FTIR spectrum of GO is shown in Fig. 7, and the appearance of characteristic stretching and bending vibrations confirms the presence of various functional groups in the structure of GO. Bands at 1, 053 and 1, 366 cm −1 correspond to stretching vibrations of C–O bonds. The intense band at 1, 227 cm −1 is due to stretching vibration of epoxy C–O bonds. The C=C (aromatic) and carbonyl stretching bonds were observed at 1, 627 and 1, 707 cm −1, respectively. A broad band at 3, 415 cm −1 can be attributed to O–H stretching vibrations. The FTIR spectrum of gelatin shows the presence of C=O stretching vibrations of amide I at 1, 639 cm −1, N–H bending of amide II at 1, 514. 2 cm −1 and amide III at 1, 211. 8 cm −1, respectively, along with an N–H stretching vibrational band due to a primary amine at 3, 442. 4 cm −1. After incorporation of gelatin into GO, there was a decrease in the intensity and shift in the vibrational frequency of C=O stretching of GO which could be attributed to the possible formation of an ammonium carboxylate complex through protonated amino groups of gelatin and carboxyl groups of GO. The merging of the vibrational frequency of amide bonds of gelatin and aromatic C=C vibrational stretching of GO led to a highly intense band at 1, 632 cm −1. A shift in the stretching vibration of amide III of gelation at 1, 211. 8–1, 241. 5 cm −1 was also observed. The FTIR of microplasma treated gelatin-graphene oxide showed that, the chemical architecture of gelatin and graphene oxide assumed to be linked via electrostatic interactions and some amide bonds was intact as no drastic changes in the vibrational frequencies of the key functional groups were observed. The intact functional groups are highly desired so as to prevent the loss of biomedical applicability of the fabricated hydrogel in case of significant changes in the structure. 10. 7717/peerj. 3498/fig-7 Figure 7 FTIR spectra analysis. FTIR spectra of: original high functionality graphene oxide (red line), gelatin (blue line) untreated gelatin-graphene oxide gel sample (gray line) and Ar- microplasma treated gelatin-graphene oxide cross-linked sample (hydrogel) (yellow line). The FTIR spectrum of the un-cross-linked gelatin and graphene oxide revealed a structure that was quite similar to raw gelatin and graphene oxide after gelatine grapheme oxide nano-composite hydrogel preparation. Confocal microscopy for gel-GO composite matrix Confocal microscopy revealed the true 3D resolution X–Z plane images of samples. It can be observed from Fig. 8 that the gelatin matrix without GO appeared flat and transparent in distinct layers. It was further elucidated using the technology of differential interference contrast from the confocal microscope. In contrast, the gel-GO group showed GO embedded as a crystal-dot-like structure within the gelatin matrix. 10. 7717/peerj. 3498/fig-8 Figure 8 Confocal microscopy. (A) gelatin w/o GO and (B) gelatin w/ GO nano composite hydrogel system (X–Z plane imaging) indicating graphene oxide (indicated by red arrows) encapsulated in gelatin matrix system (scale bar 20 µm). Swelling properties for degradation The degree of cross-linking, solvation, and degradation were primarily measured through the swelling behavior. The water-absorption capability was evaluated by monitoring the swelling ratio of untreated and treated gel-GO composite samples as a function of time. As shown in Fig. 9, the two-way ANOVA revealed a significant effect of microplasma treatment and its interactive effect on the swelling properties of the gel-GO nanocomposite hydrogels. Untreated samples exhibited swelling ratios of 5. 6∼6. 4 within 15 min, and showing disintegration into fragments by water uptake with a swelling ratio of 1. 4 within 120 min. In plasma-treated hydrogel groups, the swelling ratio reached a level 8. 3 from 6. 5 within 1 h and then equaled the initial value of 6. 5 in 2 h of incubation in deionized water at 37 °C ( p < 0. 05). 10. 7717/peerj. 3498/fig-9 Figure 9 Swelling property. The swelling property of untreated Gel-GO matrix and microplasma treated Gel-GO nanocomposite hydrogels in deionized water (pH 7. 4 and temp. 37 °C) showing significantly betterswelling property, stability and visco-elasticity of microplasma treated Gel-GO hydrogel in comparison to untreated Gel-Go material ( n = 3). Rheology for gel property analyses at body temperature Modifications of in situ gelling were measured by rheometry. The rheological study mainly focused on the storage modulus (G′) and loss modulus (G″) at a specific frequency (1 Hz) and temperature range of (7∼60 °C) with particular focus at 37 °C which is the same as the body’s physiological temperature. For analytical purposes, if G ′ > G″, then the material is more solid than liquid, and it maintains a gel state. The resulting gel-GO nano composite hydrogel was found to be with better and adequate gel strength (as shown in Fig. 10 ) in comparison to only gelatin w/o graphene oxide and untreated gel-GO nano composite. Interestingly in the case of microplasma-treated gel-GO samples, the storage modulus increased by 50% more than that of the untreated group, which was found to be even more than that of gelatin w/o GO at 40 °C temperature resulting in increased viscosity and gel strength properties ( Table 1 ). 10. 7717/peerj. 3498/fig-10 Figure 10 Rheology. Rheological tests for Storage modulus: G′ (G prime), Loss modulus, G ″ (G double prime) of gelatin without GO (red curve), gelatin-graphene oxide nano composite samples before plasma treatment (green curve) and after microplasma treatment (blue curve) at various temperature conditions indicating the differences between gelatine w/o GO, non-cross-linked and cross-linked groups. 10. 7717/peerj. 3498/table-1 Table 1 Rheology. Showing storage modulus: G′ (G prime) > Loss modulus, G″ (G double prime) up to temperature 40 °C in case of microplasma treated Gel-GO nanocomposite hydrogel with better visco-elasticity in comparison to untreated Gel-GO and gelatine w/o GO samples ( n = 5). Temp. (° C) Gel w/o GO Untreated Treated G′ G″ G′ G″ G′ G″ 5 32 ± 2 2 ± 3 42 ± 3 4 ± 0. 5 299 ± 6 16 ± 3 10 76 ± 3 3 ± 3 78 ± 4 6 ± 0. 3 329 ± 3 17 ± 4 20 198 ± 3 6 ± 1 205 ± 4 9 ± 0. 5 386 ± 7 18 ± 2 30 177 ± 1 6 ± 3 189 ± 5 9 ± 0. 5 335 ± 5 16 ± 3 37 104 ± 3 7 ± 3 108 ± 4 7 ± 0. 3 236 ± 7 12 ± 3 40 79 ± 3 5 ± 2 83 ± 3 10 ± 0. 7 182 ± 5 10 ± 1 Water contact angle for wettability and hydrophilicity The water contact angle measurement is the inverse measurement procedure to know the wettability, hydrophilicity, and hydrophobicity. The water contact angles for both untreated and plasma-treated samples revealed that both were in the hydrophilic range of <90 ° ( Figs. 11A and 11B ). However, after plasma treatment, the water contact angle increased from 49°(untreated) to 78°(treated) as observed in Fig. 12. 10. 7717/peerj. 3498/fig-11 Figure 11 Water contact angle. Water contact angle measurements of untreated Gel-GO nano composite material (A) and microplasma treated Gel-GO hydrogel (B) ( n = 5) showing decreased in hydrophilicity due to increase in water contact angle from 49° ± 7. 8°to 78° ± 3. 7°. 10. 7717/peerj. 3498/fig-12 Figure 12 Water contact angle (graphical). Water contact angle measurements of untreated Gel-GO nano composite material and microplasma treated Gel-GO hydrogel ( n = 5). MTT assay Metabolic activities of L929 cells were assessed in an indirect cytotoxicity test according to ISO 10993-5 guidelines before going to detailed in vitro cell specific studies. Cytotoxicity results were analyzed periodically from day 1 to maximum 14 days as shown graphically in Fig. 13 for GO and its effect on cell viability. At the intermediate 0. 5 wt% GO concentration, there was less-profound toxicity toward cells, while increases in cell viability and proliferation were observed at a concentration of 1 wt% GO along with gelatin. 10. 7717/peerj. 3498/fig-13 Figure 13 MTT assay (gel-GO untreated and gel-GO treated). Indirect MTT analysis of (L929 fibroblast) untreated and microplasma treated Gel-GO hydrogel (various weight percentage of graphene oxide in gelatine) after 1, 3 and 5, 7, 10 and 14 days of incubation with DMEM. MTT assay revealed the highest cell viability for microplasma treated Gel-GO hydrogels with 1 wt% of graphene oxide ( n = 6). 10. 7717/peerj. 3498/fig-14 Figure 14 MTT assay (plasma treated hydrogel and genipin treated hydrogel). Comparative MTT analysis of microplasma treated gelatin hydrogel and genipin cross-linked gelatin hydrogel by collection of liquid extraction medium at 1, 3 and 5, 7, 10 and 14 days of incubation with DMEM. MTT assay (L929 cell line) revealed the highest cell viability for microplasma treated gelatin hydrogel ( ∗ P < 0. 05; n = 6). 10. 7717/peerj. 3498/fig-15 Figure 15 Inverted microscopy. Inverted microscopy of untreated and microplasma treated gelatin–graphene oxide hydrogel seeded by osteosarcoma cell lines MG63 at 24 h. Effect of gelatin untreated (A) and gelatin microplasma treated (B) and gelatin-graphene oxide untreated (C) and microplasma treated (D) on MG63 cell lines and its proliferation was found comparatively with better result in microplasma treated gelatin—graphene hydrogel group (scale bar 100 µm). 10. 7717/peerj. 3498/fig-16 Figure 16 Live/ Dead assay (fluorescence microscopy). Live/Dead assay to check cytotoxicity of gelatin untreated (A), gelatin treated (B), gelatin-graphene oxide untreated (C), and microplasma treated gelatin-graphene oxide (D) using fluorescence staining methods (live/dead assay). Live cells and dead cells were fluorescently labelled green and red, respectively as visualized in the figure (scale bar 100 µm). On comparing microplsma assisted hydrogel with traditional genipin crosslinked hydrogel for the cytotoxicity by MTT assay, microplasma treated gelatin hydrogel system showed significantly better cell viability even on 14th day of treatment as shown in Fig. 14. Microscopy for cell-hydrogel interactions and cellular proliferation visualization Untreated and microplasma-treated gel-GO hydrogels seeded with osteosarcoma cells of the MG63 osteosarcoma cell line were microscopically observed for up to 24 h. The effects of gelatin and GO on the MG63 cell line revealed important prospects by microscopy for this study. Microscopy showed ( Fig. 15 ) better proliferation of the MG63 human osteosarcoma cell line in both gelatin- and gel-GO-treated hydrogel groups than those of the untreated gel groups. Live/dead assay We performed live/dead tests against MG-63 cells for all sample groups. We assessed the cytotoxicity of gelatin and gel-GO groups using a fluorescence staining method (live/dead assay). Live cells and dead cells were fluorescently labeled green and red, respectively. As shown in Fig. 16, almost all cells were found to be alive after a 24-h exposure to microplasma-treated gel-GO nanocomposite hydrogel. From quantitative live cell analysis under the fluorescence microscope at different fields, maximum cells (89%) were found to be alive with plasma treated gel-GO nanocomposite hydrogel as shown in Fig. 17. 10. 7717/peerj. 3498/fig-17 Figure 17 Live/Dead assay (graphical presentation). Quantitative Live/Dead assay to check primary biocompatibility and cell proliferation ability in presence of gelatin untreated G (u), gelatin treated G (t), gelatin-graphene oxide untreated G-GO (u), and microplasma treated gelatin-graphene oxide G-GO (t). Discussion Recently many polymeric hydrogel systems have been encouraged for biomedical applications and specifically in clinics ( Censi et al. , 2012 ; Shinde, Yeon & Jeong, 2013 ). The important characteristic properties to determine the quality of hydrogel systems such as mechanical strength, porosity, degradation kinetic, and bioactivity can be well tailored and controlled through chemical or physical methods ( Duan et al. , 2016 ; Kharkar, Kiick & Kloxin, 2013 ; Yesilyurt et al. , 2016 ; Lim et al. , 2014 ; Tibbitt et al. , 2015 ). Apart from this, the components of the hydrogel as well as the environmental condition are prime important things interdependent on each other for successful fabrication of desired hydrogel material. Gelatin is the processed form of collagen to be used as suitable polymer due its peculiar characteristics such as it is a high molecular weight polypeptide and the primary protein component of animal connective tissues, such as bone, hide, skin and tendon ( Nam & Park, 1999a ; Nam & Park, 1999b ). During hydrogel syntheses by chemical or physical processes, chemically crosslinked networks may result permanent junctions in irreversible manner, while physical networks have self modified controlled junctions that arise from either polymer chain entanglements or physical interactions such as ionic interactions, hydrogen bonds, or hydrophobic interactions ( Jen, Wake & Mikos, 1996 ). To be more specific during physical methods of cross-linking, high energetic ionizing radiation such as gamma rays ( Karadag et al. , 2001 ) and electron beams ( Ajji et al. , 2008 ), has been profoundly used as hydrogels initiator providing threshold energy. The study was undertaken to assess the feasibility of microplasma as an emerging tool to fabricate crosslinked gel-GO nanocomposite hydrogels for biomedical applications. As illustrated in the Fig. 18 hydrogen bonding, interaction between gelatin and graphene oxide resulted the formation of gelatin-graphene oxide nano composite gel matrix. To make it stronger, elastic and viscous, further argon microplasma was used here in our study for free radical production and cross-linked elastic network formation by molecular entanglements and ionic hydrogen bonding or covalent interactions between gelatin polymer chains and graphene oxide molecules in gel-GO nano composite hydrogel synthesis. 10. 7717/peerj. 3498/fig-18 Figure 18 Molecular mechanism behind Ar-microplasma mediated gel-GO nanocomposite hydrogel synthesis. Schematic illustration of mechanism behind Ar-microplasma assisted gelatin-graphene oxide nano composite hydrogel synthesis. The hydrogel system was characterized by cross-linking index measurements showing better cross-linking with increased porosity in the case of 8% purified gelatin. The large pore size and high porosity of the hydrogel materials enhanced the collision frequency of bio-macromolecules with free radicals produced from the plasma, which possibly helped promote the formation of cross-links between lysine and glutamic acid residues on gelatin chains along with the free and ionized radicals ( Lai et al. , 2013 ) produced by the highly energetic plasma. In the case of 7% purified gelatin, pore size was smaller compared to that of 8% plasma-treated purified gelatin due to interactions of more water molecules present within it, with the excess free radicals produced during plasma treatment resulting in neutralization and a smaller extent of cross-linking. However, the overall mechanism and scientific artifacts are not yet clear. Also, compared to the dense structure of 10% purified gelatin, the 8% purified gelatin possessed increased contact areas between gelatin molecules leading to a higher cross-linking degree and larger pore size due to the ease of interactions. It was demonstrated that the scaffold pore size governs many things in orthopedic tissue engineering such as cellular encapsulation, attachment, organization, and delivery to maintain the cellular integrity to encourage natural reparative and regenerative processes of tissues. If the pore size is too small, pore blocking may occur by cells, and no further cellular penetration happens, thus inhibiting tissue growth and proliferation. In recent studies, the effective pore size for orthopedic tissue engineering purposes was reported to be in the range of 200∼350 µm for regeneration ( Whang et al. , 1999 ). SEM also confirmed better porosity to meet the orthopedic tissue engineering scaffold criteria to be extended for other tissue engineering applications. Confocal microscopy X–Z plane imaging confirmed GO encapsulation and gel-GO composite formation due to covalent cross-linking between amino groups of gelatin and carboxyl groups of GO at an optimized time of treatment, stirring speed, and temperature. Studies have shown that graphene and its chemical derivatives have the ability to support cellular proliferation, adhesion, and differentiation with little or no cytotoxic effects ( Bai, Li & Shi, 2011 ; Shi et al. , 2012 ). Hydrogel swelling is an important parameter to determine the crosslinking density of hydrogels, and it affects cell adhesion and proliferation. Gelatin hydrogels should swell rapidly to a size sufficient to facilitate the attachment of cell grafts at implant sites. It was noted that the untreated gel-GO sample began to dissolve at 1 min due to weak viscoelastic and binding forces. Hydrogels used as tissue engineering scaffolds should not be quite dry, because the total water in the gel is comprised of both “bound” and “free” water ( Hoffman, 2012 ). Our findings suggest that the cross-linked porous gel-GO prepared by plasma cross-linking seemed suitable for use as orthopedic tissue engineering scaffolds due to their stability and swelling capability in aqueous environments without structural disintegration for 1 h. Successful cross-linking of the gelatin and GO was confirmed by FTIR spectroscopy. FTIR measurements were conducted on un-cross-linked gelatin and GO and cross-linked gel-GO to determine whether cross-linking of the gelatin affected the primary gelatin structure. The FTIR spectrum of the un-cross-linked gelatin and GO revealed a structure that was quite similar to raw gelatin and GO after composite preparation. This demonstrated that Ar-microplasma did not greatly alter the structure of raw gelatin. This controlled microplasma method was found to provide good functionally additive energy during the cross-linking process, which is an essential condition to limit insolubility and increase the biocompatibility of gelatin for applications in the biomedical field over toxic chemical processes. Rheology confirmed that better viscoelasticity and stability were well maintained in the gel state at bodily physiological pH and temperature conditions. In both the untreated and plasma-treated groups, the good viscoelasticity and gel strength were possibly due to the addition of GO. The water contact angle measurements confirmed the decreased surface hydrophilicity of the plasma-treated gel-GO hydrogel films to make them more as a viscoelastic seal to attach to tissue surfaces for cellular adhesion and proliferation. This is attributed to Ar plasma treatment that induced modifications of the polymer to make it adequately hydrophilic. Water contact angle measurements proved it to be a suitable carrier material that provides a platform for cellular attachment and proliferation. The MTT assay revealed better proliferation with 1% GO in the composite hydrogel system even at up to 14 days compared to the untreated hydrogel groups. A higher number of covalent cross-links between adjacent polymer chains along with GO caused formation of a potential elastic network which is useful for cell penetration and growth. The additive binding forces for many-fold enhancement of the mechanical gel strength in treated samples were due to microplasma-induced cross-linking. Cell proliferation positively indicated an initial adaptability of fibroblast cells, as well as cell attachment and adhesion. With an increase in the time period, cells began penetrating and proliferating due to the porous structure of the GO content in the composite hydrogel scaffolds. Furthermore, GO has several physiochemical properties such as an ultra-large surface area and possesses many functional groups including hydroxyl (OH), epoxy (C–O–C), and carboxyl (COOH) groups on its surface ( Nezakati, Cousins & Seifalian, 2014 ). Therefore, if added to biomaterial tissue engineering scaffolds along with plasma as ionization energy source, GO can adsorb some biomolecules to improve their chemical and biological properties allowing bio-functionalization, and outstanding water solubility that makes the scaffold a promising material ( Nguyen & Nguyen, 2016 ) for cellular proliferation. In our study we also compared our plasma treated gelatin scaffold system with traditionally synthesized genipin mediated crosslinked gelatin hydrogel system for cytotoxicity assay. Plasma mediated gelatin hydrogel was found to be with better cell viability and cellular proliferation for fibroblast cell line L929 and significantly with less cytotoxicity. The use of plasma process for tissue engineering polymer scaffold fabrication may be ideal for future biomedical application with less toxicity. Microscopy and the live/dead assay also observed with better results of cell proliferation and survival prospects with the hydrogel system. It is noteworthy to mention for the first time that Ar-microplasma-induced gel-GO composite hydrogels supported cellular spreading and alignment with improved viability and proliferation. Suitable mechanical strength and enhanced tunable properties also represent desirable attributes of this composite hydrogel system, especially as a scaffold material in orthopedic tissue engineering. The resulting biodegradable, soft, elastic gel-GO nanocomposite hydrogel material was shown to cover a wide range of suitable properties for tissue engineering such as the cross-linking degree, pore size, hydrophilicity, viscoelasticity, and tunable mechanical properties; these are all imperative in controlling biological responses to implanted materials along with cells and growth factors at the defect site during healing and regeneration. After characterizing the hydrogels, we obtained future insights for the application of our hydrogel system as a simple, novel, thin sealing plug material scaffold with the anticipation of better reliability during surgical interventions in orthopedic clinics for cartilage or bone regeneration as well as for biomedical tissue engineering without toxic chemical additives. The gel-GO hydrogel fabrication system with mechanical durability and improved cellular performance may provide an effective tool for the healing of complex defective and degenerative tissues. Conclusions The present study optimizes the microplasma-mediated cross-linking process to overcome toxicity issues associated with fabrication of hydrogels in tissue engineering by chemical cross-linking. Further, this study explores the effect of Ar-microplasma in gelatin hydrogel formation containing GO. The gel-GO nanocomposite hydrogel was characterized by various methods such as the degree of cross-linking, FTIR spectroscopy, SEM, confocal microscopy, swelling behavior, contact angle measurement, and rheology, and the cell viability was also examined by an MTT assay, live/dead assay, and microscopy. The pore size of the hydrogel was found to be 287 ± 27 µm which is optimum for orthopedic tissue engineering purposes with future direction to be used in other tissue engineering fields. The contact angle of 78° ± 3. 7°indicated the controlled hydrophilic nature of the hydrogel. Rheological data revealed improved storage as well as loss modulus of up to 50% with tunable viscoelasticity, gel strength, and mechanical properties at 37 °C body temperature conditions in the microplasma-treated groups. Better cell viability at 1% (w/w) of high functionality GO in gelatin was demonstrated by the MTT assay, microscopy, and live/dead assay as well as directly by inverted microscopy. As observed, the aforementioned plasma strategy is suitable to enhance the soft tissue engineering scaffold fabrication and tissue regeneration for promoting the clinical and biomedical applications in relevant fields. These encouraging results highlight the uniqueness of the Ar-microplasma process for gel-GO nanocomposite hydrogel scaffold fabrication and its promising attributes. This study moves forward the novel use of an electrically neutral beam of pure argon plasma from the bench top to the clinic for biomedical material fabrication with a tissue engineering approach to assist and accelerate the regeneration and repair of defective and damaged tissues. Keeping these exciting findings of the biomedical applicability endowed by Ar-microplasma in view, in vitro and in vivo experimental studies of hydrogels in various fields of basic and applied biomedical tissue engineering will be carried out in the near future. However, specific biomedical applications of plasma still require detailed investigations. Supplemental Information 10. 7717/peerj. 3498/supp-1 Supplemental Information 1 Cross-linking Index Cross-linking index of various concentrations of purified gelatin. An asterisk indicates statistically significant difference of 8% purified gelatin group cross-linking ( ∗ P > 0. 05 and n = 5) as compared with other groups (7%, 9%, 10%). Click here for additional data file. 10. 7717/peerj. 3498/supp-2 Supplemental Information 2 FTIR spectra analysis FTIR spectra of: original high functionality graphene oxide (red line), gelatin (blue line) untreated gelatin-graphene oxide gel sample (gray line) and Ar- microplasma treated gelatin-graphene oxide cross-linked sample (hydrogel) (yellow line). The FTIR spectrum of the un-cross-linked gelatin and graphene oxide revealed a structure that was quite similar to raw gelatin and graphene oxide after gelatine grapheme oxide nano-composite hydrogel preparation. Click here for additional data file. 10. 7717/peerj. 3498/supp-3 Supplemental Information 3 Swelling property The swelling property of untreated Gel-GO matrix and microplasma treated Gel-GO nanocomposite hydrogels in deionized water (pH 7. 4 and temp. 37 °C) showing significantly betterswelling property, stability and visco-elasticity of microplasma treated Gel-GO hydrogel in comparison to untreated Gel-Go material ( n = 3). Click here for additional data file. 10. 7717/peerj. 3498/supp-4 Supplemental Information 4 Water contact angle (graphical) Water contact angle measurements of untreated Gel-GO nano composite material and microplasma treated Gel-GO hydrogel. ( n = 5). Click here for additional data file. 10. 7717/peerj. 3498/supp-5 Supplemental Information 5 MTT assay (gel-GO untreated and gel-GO treated) Comparative MTT analysis of microplasma treated gelatin hydrogel and genipin cross-linked gelatin hydrogel by collection of liquid extraction medium at 1, 3 and 5, 7, 10 and 14 days of incubation with DMEM. MTT assay (L929 cell line) revealed the highest cell viability for microplasma treated gelatin hydrogel ( ∗ P < 0. 05; n = 6). Click here for additional data file. 10. 7717/peerj. 3498/supp-6 Supplemental Information 6 MTT assay (plasma treated hydrogel and genipin treated hydrogel) Comparative MTT analysis of microplasma treated gelatin hydrogel and genipin cross-linked gelatin hydrogel by collection of liquid extraction medium at 1, 3 and 5, 7, 10 and 14 days of incubation with DMEM. MTT assay (L929 cell line) revealed the highest cell viability for microplasma treated gelatin hydrogel ( ∗ P < 0. 05; n = 6). Click here for additional data file. 10. 7717/peerj. 3498/supp-7 Supplemental Information 7 Prisma checklist Summarised manuscript Click here for additional data file. 10. 7717/peerj. 3498/supp-8 Supplemental Information 8 PRISMA flow diagram Schematic presentation of the gel-GO nano-composite hydrogel synthesis by Ar-microplsma and its biomedical application. Click here for additional data file.

10. 7717/peerj. 3513

PeerJ

Local application of osteoprotegerin-chitosan gel in critical-sized defects in a rabbit model

Background Osteoprotegerin (OPG) is used for the systemic treatment of bone diseases, although it has many side effects. The aim of this study was to investigate a newly formulated OPG-chitosan gel for local application to repair bone defects. Recent studies have reported that immunodetection of osteopontin (OPN) and osteocalcin (OC) can be used to characterise osteogenesis and new bone formation. Methods The osteogenic potential of the OPG-chitosan gel was evaluated in rabbits. Critical-sized defects were created in the calvarial bone, which were either left unfilled (control; group I), or filled with chitosan gel (group II) or OPG-chitosan gel (group III), with rabbits sacrificed at 6 and 12 weeks. Bone samples from the surgical area were decalcified and treated with routine histological and immunohistochemical protocols using OC, OPN, and cathepsin K (osteoclast marker) antibodies. The toxicity of the OPG-chitosan gel was evaluated by biochemical assays (liver and kidney function tests). Results The mean bone growth in defects filled with the OPG-chitosan gel was significantly higher than those filled with the chitosan gel or the unfilled group ( p < 0. 05). At 6 and 12 weeks, the highest levels of OC and OPN markers were found in the OPG-chitosan gel group, followed by the chitosan gel group. The number of osteoclasts in the OPG-chitosan gel group was lower than the other groups. The results of the liver and kidney functional tests indicated no signs of harmful systemic effects of treatment. In conclusion, the OPG-chitosan gel has many characteristics that make it suitable for bone repair and regeneration, highlighting its potential benefits for tissue engineering applications.

Introduction The gold standard for bone regeneration is an autologous bone graft. However, the procurement of autogenous bone comes with some disadvantages, such as creating an additional surgical area, significant morbidity, and limited source material. Thus, bone tissue engineering and regenerative medicine can be employed to improve bone regeneration ( Arrigoni et al. , 2013 ; Muschler et al. , 2010 ; Sasso et al. , 1998 ). The discovery of osteoprotegerin (OPG) as an inhibitor of osteoclast activity and maturation has led to new research exploring the applicability of OPG as a potential therapeutic agent for the treatment of bone diseases and to induce bone formation ( Bekker et al. , 2001 ; Fili, Karalaki & Schaller, 2009 ; Hofbauer et al. , 2001 ; Kostenuik et al. , 2001 ). The first clinical trial evaluated the efficacy of recombinant Fc-OPG, used systemically as a drug for the treatment of osteoporosis in postmenopausal women ( Bekker et al. , 2001 ) and another study evaluated a different formulation of OPG, known as AMGN-0007, in patients with lytic bone lesions associated with multiple myeloma or breast carcinoma ( Body et al. , 2003 ). Both of these studies reported that Fc-OPG treatment resulted in reduced bone turnover markers when administered at a low dose, and had a longer antiresorptive effect when administered at an equivalent dose. The authors of these studies cited two potential concerns with Fc-OPG therapy. The first is the generation of anti-Fc-OPG antibodies, which might cross-react with endogenous Fc-OPG, neutralising its activity. The second potential concern is the binding of Fc-OPG to TNF-related apoptosis-inducing ligands, which could inhibit their role in tumour surveillance ( Schwarz & Ritchlin, 2007 ). These side effects are, in part, due to systemic administration of the treatment. Recently, it was reported that twice-weekly injections with a high dose of OPG-Fc (5. 0 mg/kg) into the mesial and distal mucosa of the first molars during orthodontic movement improved bone quantity and orthodontic anchorage in a rat model ( Fernández-González et al. , 2016 ). In addition, OPG-chitosan matrices have been shown to enhance cell growth and proliferation, as well as increasing the production of osteopontin (OPN) and osteocalcin (OC) protein levels ( Jayash et al. , 2016 ). In order to optimise the OPG concentration and prolong the duration of protein release, aimed at avoiding the side effects associated with systemic application, a controlled drug delivery system for the local application of OPG is required. Chitosan was selected as the matrix material for this method of drug delivery. Bone matrix proteins such as OPN function to induce osteoclast migration and adhesion, while OC functions to regulate mineralisation. These functions highlight the importance of measuring bone matrix proteins when characterising osteogenesis processes, as well as the influence of the drug on their expression ( Bondarenko et al. , 2014 ). Cathepsin K is one of the biomarkers expressed by osteoclasts during active bone resorption, making it a useful and specific biomarker of osteoclastic activity. Cathepsin K is expressed by osteoclasts and a small number of osteoclast precursors, but it is not expressed by osteoblasts or osteocytes. Therefore, its expression is specific to the resorption phase of bone metabolism ( Drake et al. , 1996 ). We previously formulated and investigated an OPG-chitosan gel for its cytocompatibility ( Jayash et al. , 2017a ); however, the in vivo effects on bone regeneration were not extensively investigated and discussed in our previous work. In the present paper, we investigated the efficacy of an OPG-chitosan gel in bone regeneration in terms of the expression of osteoblast- and osteoclast-specific proteins, and tested the toxicity of the OPG-chitosan gel by biochemical assays (liver and kidney function tests). Methods Formulation of the osteoprotegerin-chitosan gel The OPG-chitosan and chitosan gels were prepared using water-soluble chitosan (25 kDa). Recombinant human OPG protein (1 mg/mL) was used to prepare the OPG-chitosan gel (PeproTech, Rocky Hill, NJ, USA) with a chitosan binder (85 kDa). The methodology for preparing the OPG-chitosan gel has been described previously ( Jayash et al. , 2017b ) and patented under the title, “An osteoprotegerin-chitosan gel for bone tissue regeneration” (PI 2016701598, UM). Animal model The animal experiment was authorised by the Institutional Animal Care and Use Committee at the University of Malaya (FOM IACUC), and registered under number 20150115/DENT/R/NAB. Critical-sized defects were created in the calvarial bone of 18 New Zealand white female rabbits (6 months old; 3. 5–4 kg). The rabbits were divided into three groups: (i) untreated control group (group I; n = 6), (ii) chitosan-only gel (group II; n = 6), and (iii) OPG–chitosan gel (group III; n = 6). The anaesthesia, surgery, medical treatment, and euthanasia procedures have been described previously ( Jayash et al. , 2017a ). Three rabbits from each group were examined at 6 and 12 weeks. The bone samples were fixed in 10% neutral buffered formalin for 24 h, decalcified in 10% EDTA for 3 weeks, then embedded in paraffin according to the established technique. Three 4-µm thick sections from each paraffin block were cut using a microtome (Leica, Wetzlar, Germany). For histological studies, the sections were stained with haematoxylin and eosin (H&E). Immunohistochemistry was performed according to established methods ( Gruber & Ingram, 2003 ; Bondarenko et al. , 2014 ). The anti-OPN (clone 1B20; Novus Biologicals, Cambridge, UK), anti-OC (OCG3; Abcam, Cambridge, UK), and cathepsin K (Biovision, Milpitas, CA, USA) polyclonal antibodies have previously been verified in rabbit bone tissue ( Arrigoni et al. , 2013 ; Bondarenko et al. , 2014 ). Quality control A negative control reagent was used with each specimen to identify any non-specific staining. If non-specific staining could not be clearly differentiated from specific staining, the labelling of the test specimen was considered invalid. In this experiment, rabbit immunoglobulin fraction (normal; Dako, Industrial RowTroy, Michigan, USA) was used as the negative control. Histological evaluation Verification of the immunohistochemical reaction was performed using a light microscope and scanned using a digital slide scanner (3DHISTECH Ltd. , Budapest, Hungary). The results were assessed using computer-assisted image analysis (ImageJ; National Institutes of Health, Bethesda, MD, USA). The image was opened and the image threshold was adjusted until all stained areas were selected. A histogram was displayed to provide assistance. Staining was assessed by setting a threshold using the threshold tool. The threshold tool settings that successfully quantified the staining in a positive-stained specimen were repeated in every image for comparison. The analysed–set measurement was selected, and the parameters to be measured were chosen ( Jensen, 2013 ). The mean and standard deviation were calculated for each sample. Serum biochemical parameters Blood samples were collected from all rabbits before surgery and before sacrifice. The blood samples were allowed to clot at room temperature before centrifuging at 1, 000 g for 10 min. The serum was separated and analysed for markers of kidney function, including creatinine and urea nitrogen, and markers of liver function, including alkaline phosphatase (ALP) and alanine aminotransferase (ALT). Analyses were conducted using a clinical chemistry analyser (902; Hitachi, Tokyo, Japan) with standard diagnostic kits (Hitachi 902, Roche). Statistical analysis Statistical analyses were performed using the parametric one-way ANOVA test and non-parametric Mann–Whitney U and Kruskal–Wallis tests for comparison between the mean values of different groups. The significance value was set at p < 0. 05. Results are presented as the mean (arithmetic mean) and standard deviation. Results Clinical findings At 6 weeks, rabbits in group III (OPG-chitosan gel) showed partially healed surgical defects, which were filled with a dense, opaque structure (bone). On the other hand, the surgical defects of groups I (surgery only) and II (chitosan gel) were completely filled with thin, transparent soft tissue. At 12 weeks, the surgical defects of groups II and III were completely filled with hard tissue, whereas the defects of group I were still only partially healed, with regions of soft tissue ( Fig. 1 ). 10. 7717/peerj. 3513/fig-1 Figure 1 Gross appearance of surgical defects at 6 and 12 weeks. (A) Group I (control), (B) group II (chitosan gel), and (C) group III (OPG-chitosan gel). Histological results (haematoxylin and eosin staining) At 6 weeks, the defects of rabbits in group III were filled with new bone and osteoid tissue, while the group II defects were filled with new bone and fatty marrow. However, the group I defect was only partially filled, with the least amount of new bone compared to groups II and III. The connective tissue of the bone bridge in groups II and III was less prominent than that of group I. No graft particles were seen at 6 weeks, which suggests that the particles may have been completely resorbed. The new osteoid bone filling each of the defects was formed within the region of interest. The trabecular bone in group II appeared thick and dense, while the trabecular bone in group III had become lamellar bone. At 12 weeks, newly formed bone had completely filled the defect of rabbits in group III. The newly formed bone in these defects resembled a bridge, and was arranged as lamellae in some areas. These areas contained large osteons and Haversian canals, as well as highly cellularised connective tissue, especially within the central region. Moreover, cortical organisation was evidenced by the presence of trabecular projections, in addition to the maturation of lamellar bone indicated by thickening at the margin of the defect compatible with the original bone structure of the region ( Fig. 2 ). 10. 7717/peerj. 3513/fig-2 Figure 2 Photomicrographs of defect sites of (A) Group I, (B) Group II (C) Group III at 6 and 12 weeks. Sections were stained with haematoxylin and eosin. NB, new bone; OS, osteoid; FM, fatty marrow; HC, Haversian canal; TB, trabecular bone. (Scale bar represents 500 µm. ) Immunohistochemistry results Osteoblast markers At 6 weeks, OPN immunolabelling was observed in the osteoclasts, osteoblasts, and fibroblasts. In almost all specimens, OPN immunoreactivity was located within the matrix of compact bone, cancellous bone, and osteoid. Osteocalcin immunoreactivity appeared in the matrix of compact bone, but was weaker in cancellous bone and osteoid ( Fig. 3 ). 10. 7717/peerj. 3513/fig-3 Figure 3 Photomicrographs of immunostaining for osteocalcin and osteopontin in (A) Group I, (B) Group II (C) Group III at 6 weeks. The pictures are arranged by staining technique (columns) and the investigated treatment (rows). Areas that stained positive for osteocalcin and osteopontin are indicated by red arrowheads. NB, new bone; OS, osteoid; FM, fatty marrow. At 12 weeks, OPN expression was characterised by osteoid staining in all investigated groups. Immunolabelling of OPN was observed in osteoblasts. The bone matrix also showed stained areas. Osteocalcin was detected in the matrix of compact bone, but showed reduced expression in osteoid and bone cells ( Fig. 4 ). 10. 7717/peerj. 3513/fig-4 Figure 4 Immunohistological results at 12 weeks for (A) Group I, (B) Group II (C) Group III. The pictures are arranged by staining technique (columns) and by the investigated treatment (rows). Areas that stained positive for osteocalcin and osteopontin are indicated by red arrowheads. NB, new bone; OS, osteoid; FM, fatty marrow; CB, compact bone; CAB, cancellous bone. Based on the parametric one-way ANOVA, there was a significant difference in the mean percentage of OPN expression between groups I and III ( p < 0. 05). The highest expression of OPN was observed in group III, followed by group II and group I, which showed the lowest OPN expression. Based on the parametric one-way ANOVA, there was no significant difference in the mean percentage of OPN expression between groups II and I ( p > 0. 05). However, there was a significant difference in the mean percentage of OPN expression between groups III and II ( p < 0. 05; Fig. 5 ). 10. 7717/peerj. 3513/fig-5 Figure 5 Statistical analysis of the percentage expression of osteopontin as a bone-formation marker between groups I, II and III at 6 and 12 weeks. Data are presented as the average of three independent experiments ( n = 3). ∗∗ Significant difference between the control (group I) and experimental groups (groups II and III). ∗ Significant difference between groups III and II. A p -value of <0. 05 was considered significant in all analyses. Statistical analysis of the percentage of OC expression at 6 weeks showed a significant difference between groups I, II and III ( p < 0. 05). The highest OC expression was found in group III, followed by group II ( Fig. 5 ). Regarding OC expression at 12 weeks, there was a significant difference in the mean percentage of OC expression in group III when compared to groups II and I ( p < 0. 05). There was also a significant difference in the mean percentage of OC expression between groups III and I ( p < 0. 05). The highest OC expression at 12 weeks was found in group III, followed by group II ( Fig. 6 ). 10. 7717/peerj. 3513/fig-6 Figure 6 Statistical analysis of the percentage expression of osteocalcin as a bone-formation marker between groups I, II and III at 6 and 12 weeks. Data are presented as the average of three independent experiments ( n = 3). ∗∗ Significant differences ( p < 0. 05) between control (group I) and experimental groups (groups II and III). In the intragroup comparison, there was a significant difference in OPN expression between 6 and 12 weeks in all groups ( p < 0. 05), with the highest expression of OPN observed at 6 weeks. Similarly, there was a significant difference in OC expression between 6 and 12 weeks in all groups, with the highest expression observed at 6 weeks. The expression of OPN and OC between 6 and 12 weeks was compared for all groups. We found no significant difference in the expression of OPN and OC in group III ( p > 0. 05). In group II, however, there was a significant difference in the expression of OPG and OC at 6 weeks ( p < 0. 05), when OC was more highly expressed. In group I, there was a significant difference between the expression of OPN and OC at 12 weeks, with higher expression observed for OPN. As the sample size was small, non-parametric tests were also carried out, and similar results were obtained. Osteoclast marker More cathepsin K-positive areas were observed in the medullary region of the defect site than the cortical bone region at both 6 and 12 weeks. Cathepsin K-positive multinuclear cells were also detected within the newly-formed bone. In addition, fewer osteoclasts were detected at 12 weeks than at 6 weeks for all groups ( Fig. 7 ). 10. 7717/peerj. 3513/fig-7 Figure 7 Cathepsin K immunostaining of osteoclasts in groups (A) Group I, (B) Group II (C) Group III at 6 and 12 weeks after surgery. Cathepsin K-positive multinuclear cells are indicated by red arrowheads. MR, medullary region. Serum biochemical parameters The serum biochemistry results of rabbits in groups I, II and III at baseline and at 6 and 12 weeks are summarised in Tables 1 and 2. There were no significant changes in the levels of creatinine and urea nitrogen. In addition, no significant differences in serum electrolytes including calcium, potassium, and chloride were noted. The effects of OPG and/or chitosan on liver function parameters such as albumin, ALT, G-glutamyl transferase, ALP and total bilirubin were also examined. Animals in groups I, II and III showed no difference in hepatic markers. The effects of treatment on triglyceride, high-density lipoprotein (HDL) cholesterol, and total cholesterol levels are shown in Table 2. Rabbits in all groups showed no significant changes in triglyceride, HDL cholesterol, or total cholesterol levels after treatment. Discussion Allografts and autografts present several clinical problems that need to be addressed, such as transmission of infection, unpredictable resorption, pain at the donor site, and limited donor source. To eliminate these problems, it would be advantageous to develop a material that can promote bone regeneration at the defect site. The intention of this study was to investigate the local of application of OPG-chitosan gel in a critical-sized defect in the parietal bone in a rabbit models. To investigate this, three treatments (OPG-chitosan gel, chitosan gel, and untreated control groups) were tested at 6 and 12 weeks. The clinical results of the current study showed that treatment of surgical defects with the OPG-chitosan gel was not associated with any signs of inflammation or infection. This indicates biocompatibility of the newly formulated OPG-chitosan gel. At 12 weeks, the surgical defects of rabbits treated with the chitosan gel were filled with new bone, as observed from the histological sections. However, the defects in this group were only partially filled, indicating that chitosan enhances bone formation. This observation is in agreement with previous studies that reported a higher amount of bone formed in defects treated with chitosan compared to untreated controls ( Ezoddini-Ardakani et al. , 2012 ). In addition, the amount of bone formed was reduced when compared to those defects treated with chitosan combined with other active components ( Bush et al. , 2016 ; Oktay et al. , 2010 ). 10. 7717/peerj. 3513/table-1 Table 1 Serum biochemical data for rabbits treated with OPG-chitosan (group III) or chitosan (group II) gels and untreated control rabbits (group I) at baseline and 6 weeks after treatment. Parameter Baseline 6 weeks Normal range Group III Group II Group I Group III Group II Group I Sodium mmol/L 142 ± 0 142 ± 1 141 ± 1 143 ± 0 141 ± 1 141 ± 0. 7 139–146 Potassium mmol/L 4 ± 0. 1 5 ± 0. 4 5 ± 0. 7 3. 8 ± 0. 1 3. 95 ± 0. 1 5 ± 0. 6 3–5 Chloride mmol/L 103 ± 0. 7 102 ± 0 101. 5 ± 0. 2 100 ± 4. 2 101 ± 1 99. 5 ± 4 104–116 Urea nitrogen mmol/L 8 ± 0. 42 7 ± 1. 5 8 ± 0. 2 6 ± 1. 5 7 ± 0. 07 9 ± 0. 07 6. 35–16 Creatinine µmol/L 77 ± 6 78 ± 2 77 ± 5 102 ± 15 77 ± 7 85 ± 4 60–140 Albumin (g/L) 41 ± 0. 0 38 ± 1 41 ± 1 39 ± 4 40 ± 4 40 ± 1 20–40 Total Bilirubin (µmol/L) <2 <2 <2 <2 <2 <2 2–5 Alkaline Phosphate (U/L) 102 ± 6 111 ± 8 110 ± 4. 9 50 ± 10 81 ± 3 69 ± 3 17–192 Alanine Aminotransferase (U/L) 30. 5 ± 3 94. 5 ± 2 36 ± 15 35 ± 0 53 ± 12. 7 52. 5 ± 3. 1 38–86 G-Glutamyl Transferase (U/L) 5. 5 ± 0. 7 5. 5 ± 0. 9 2. 5 ± 0. 7 6 ± 1. 5 6. 5 ± 1. 4 5 ± 1. 4 6–22 Triglyceride (mmol/L) 0. 5 ± 0. 1 0. 7 ± 0. 1 0. 65 ± 0. 2 0. 5 ± 0. 1 0. 7 ± 0. 15 0. 75 ± 0 – Total Cholesterol (mmol/L) 1 ± 0. 2 0. 9 ± 0. 1 1. 1 ± 0 1. 3 ± 0. 3 0. 85 ± 0. 2 1. 2 ± 0. 1 – HDL Cholesterol (mmol/L) 0. 8 ± 0. 2 0. 61 ± 0. 1 0. 8 ± 0. 2 0. 2 ± 0. 1 0. 62 ± 0. 1 0. 8 ± 0. 1 – Notes. Data are expressed as the mean ± SD of three rabbits per group. 10. 7717/peerj. 3513/table-2 Table 2 Serum biochemical data for rabbits treated with OPG-chitosan (group III) or chitosan (group II) gels and untreated control rabbits (group I) at baseline and 12 weeks after treatment. Parameter Baseline 12 weeks Normal range Group III Group II Group I Group III Group II Group I Sodium mmol/L 142 ± 2 142 ± 1 143 ±2 142 ± 2. 6 143 ± 1 143 ± 2 139. 3–145. 7 Potassium mmol/L 4 ± 1 4 ± 0. 3 4 ± 0. 4 4 ± 0. 6 4 ± 0. 4 4 ± 0. 5 3–5 Chloride mmol/L 100 ± 0. 5 101 ± 0. 6 98 ± 3. 5 102 ± 0. 2 101 ± 2 104 ± 4 105–116 Urea nitrogen mmol/L 8 ± 1 8 ± 1 9 ± 0. 4 8 ± 1 7 ± 1 8 ± 1 6–16 Creatinine µmol/L 85. 7 ± 10 87 ± 1 89 ± 2 89 ± 10 83 ± 2 87 ± 1 60–140 Albumin (g/L) 43. 3 ± 3 40. 7 ± 1 42 ± 1 41 ± 2. 9 40 ± 3 39 ± 3 20–41 Total Bilirubin (µmol/L) <2 <2 <2 <2 <2 <2 2–5 Alkaline Phosphate (U/L) 68. 5 ± 12 44 ± 3 69. 5 ± 8 58 ± 1 44 ± 9 48 ± 8 17–192 Alanine Aminotransferase (U/L) 58 ± 6. 9 60 ± 9. 7 48. 7 ± 10 53. 7 ± 15. 8 67 ± 12 54 ± 7 38–86 G-Glutamyl Transferase (U/L) 6 ± 1. 2 7 ± 1. 5 3 ± 1. 4 4 ± 0. 6 7 ± 1. 7 5 ± 1. 2 6. 00–22. 00 Triglyceride (mmol/L) 0. 7 ± 0. 5 0. 6 ± 0. 2 0. 4 ± 0. 01 1. 2 ± 0. 2 0. 6 ± 0. 2 0. 5 ± 0. 2 – Total Cholesterol (mmol/L) 1 ± 0. 2 1 ± 0. 2 1 ± 0. 2 1 ± 0. 2 1 ± 0. 2 1 ± 0. 2 – HDL Cholesterol (mmol/L) 0. 7 ± 0. 1 0. 5 ± 0. 03 0. 7 ± 0. 1 0. 9 ± 0. 04 0. 6 ± 0. 1 0. 6 ± 0. 1 – LDL Cholesterol (mmol/L) 0. 2 ± 0. 1 0. 2 ± 0. 08 0. 1 ± 0. 1 0. 1 ± 0. 1 0. 2 ± 0. 1 0. 26 ± 0. 1 – Notes. Data are expressed as the mean ± SD of three rabbits per group. Our observation that the surgical defects treated with OPG-chitosan gel (group III) were completely filled with new bone suggests that the OPG protein represents an active component that could improve bone regeneration. We also observed a large number of osteons in the bone tissue of the OPG-chitosan group at 6 and 12 weeks. This observation is in agreement with an earlier study ( Allegrini et al. , 2006 ), and supports the notion that active bone formation is taking place during the healing process. In any bony defect, the most intensive cellular reactions occur during the first 6 weeks, when the bone defect is initially healed with trabecular bone consisting of primitive woven bone. After 6 weeks, there is a reduction in the number of cells in the area of the defect, as well as increased calcium deposition, which is in agreement with a study by Gehrke (2013). Bone remodelling was also active after 8 weeks of healing in the rabbit, with various degrees of bone maturation in addition to uneven osseous formation ( Jansen et al. , 1991 ). The compatibility and osteoconductivity potential of the OPG-chitosan gel was evident from the presence of progressive new bone formation and remodelling in the surgical defects at 6 and 12 weeks. A previous study reported an osteoinductive effect of treatments used to enhance bone formation in rabbits through the increased expression of bone-formation proteins (e. g. , osteopontin and collagen type I) in bone defects treated with the bioconstruct ( Arrigoni et al. , 2013 ). The current study showed significantly stronger expression of OPN in the OPG-chitosan gel group compared to the control group, which supports the osteogenic potential of the OPG-chitosan gel. All groups showed higher expression of OPN at 6 weeks compared to 12 weeks, but no significant difference was observed for OC expression in the OPG-chitosan gel group. This is in agreement with other studies that found that OPN is expressed throughout matrix maturation, and OC seems to play a role in the early stages of bone formation, showing positive expression at the area of bone formation at 6 weeks ( Al-Ghani, Al-Hijazi & AL-Zubaydi, 2011 ; Kim et al. , 2015 ). In this study, OC expression was associated with mature bone, confirming its expression as a late marker of osteogenesis. Osteocalcin is commonly localised adjacent to osteoblasts, osteoid, and within osteocyte lacunae. Tera et al. (2014) observed OC staining in bone defects in a rat model, and concluded that the staining was relatively weaker in the regenerating bone compared to mature bone. The expression of OC in OPG-chitosan, chitosan, and control groups was higher at 6 weeks compared to 12 weeks, except for OC expression in the OPG-chitosan gel group, for which there was no significant difference. Osteocalcin appears to play a role in the early stages of bone formation, and showed positive expression at the area of bone formation at 6 weeks. The OC expression level was directly related to the bone formation activity. The results of this study confirm the important roles of OC and OPN in bone healing, consistent with a study by Bondarenko et al. (2014). Evaluation of the biocompatibility and osteoconductivity of the OPG-chitosan gel demonstrated the superiority of this gel in terms of biocompatibility (no adverse reactions were observed), osteoconductivity, and progressive osteogenesis during the entire bone-healing period. Complete wound closure of the critical-sized bone defect filled with OPG-chitosan gel compared to chitosan gel and the unfilled control further supports the favourable biological properties of this biomaterial in a clinical situation. The immunohistological investigations presented here are consistent with enhanced new bone formation in the OPG-chitosan group, as we observed increased expression of OPN and OC when compared to the chitosan gel group and the untreated control group. This can be explained by the characteristics of the OPG-chitosan gel, which allows prolonged retention of OPG at the defect site. We demonstrated that OPG is released from the OPG-chitosan gel in vivo, and it is possible that the controlled release of OPG at the defect site enhances the recruitment of active osteoblasts and suppresses the recruitment of osteoclasts, thereby improving bone regeneration. We previously reported that the OPG-chitosan gel increases the in vitro proliferation of normal human osteoblasts ( Jayash et al. , 2017a ). Cathepsin K is expressed by osteoclasts and osteoclast precursors, but is absent in osteoblasts and osteocytes. Therefore, its expression is specific to the resorption phase of bone metabolism ( Drake et al. , 1996 ). Osteoblasts control the formation and activity of osteoclasts, as well as the resorption of bone through coupling mechanisms such as RANK, RANKL, and OPG ( Jeon et al. , 2016 ; Pederson et al. , 2008 ; Teti, 2013 ). The cathepsin K immunohistochemical staining showed fewer osteoclasts in the defect treated with OPG-chitosan gel than in the chitosan gel and control groups. This may be due to the action of OPG, which acts as a decoy receptor for RANKL, blocking the activation and maturation of osteoclasts. This observation is in agreement with previous studies in which chitosan has been demonstrated to enhance bone growth in critical-size defects in rat, rabbit, and sheep models ( Muzzarelli et al. , 1994 ; Pang et al. , 2005 ; Wang et al. , 2002 ). Previous studies have also shown a significant increase in bone formation in the osseous healing area of rats, rabbits, dogs, and humans where the OPG or chitosan-based biomaterial was applied ( Ezoddini-Ardakani, 2011 ; Florczyk et al. , 2013 ; Jung et al. , 2014 ; Lamoureux et al. , 2007 ; Yao et al. , 2011 ). Furthermore, the serum levels of markers of liver and kidney function indicated no signs of any harmful systemic effects of treatment. This is in agreement with a previous study that reported no adverse effects in the liver and kidney resulting from administration of a chitosan drug for treating bone fractures ( Ho et al. , 2015 ). Conclusion We conclude that the bioresorbable OPG-chitosan material induced the formation of a significant quantity of bone in a critical-sized parietal bone defect in a rabbit model. Although the chitosan gel group also yielded a greater amount of bone compared to the control group at all time points, the amount of bone generated was significantly lower than the OPG-chitosan gel at both 6 and 12 weeks. This study also reports that the OPG-chitosan gel showed the best behaviour, both clinically and histologically. It is biocompatible when used in vivo, and showed no adverse reactions. The combination of OPG and chitosan in a gel resulted in significantly enhanced new bone formation compared to chitosan alone or the untreated control in a rabbit model. This suggests that the enhanced new bone formation can be attributed to local release of OPG. The OPG-chitosan gel has many characteristics that make it suitable for bone repair and regeneration, highlighting its potential benefits for tissue engineering applications. Supplemental Information 10. 7717/peerj. 3513/supp-1 Table S1 Figure 4 raw data Summary of the means of OPN expressions in percentages for Groups I, II and III at 6 weeks. Click here for additional data file. 10. 7717/peerj. 3513/supp-2 Table S2 Figure 5 raw data Summary of the means of OC expressions in percentages for Groups I, II and III at 6 weeks. Click here for additional data file. 10. 7717/peerj. 3513/supp-3 Table S3 Figure 7 raw data The percentage of OPN expression percentages in groups I, II and III at 12 weeks. Click here for additional data file. 10. 7717/peerj. 3513/supp-4 Table S4 Figure 8 raw data The percentage of OC expressions percentages in groups I, II and III at 12 weeks. Click here for additional data file.

10. 7717/peerj. 353

PeerJ

Culture of equine fibroblast-like synoviocytes on synthetic tissue scaffolds towards meniscal tissue engineering: a preliminary cell-seeding study

Introduction. Tissue engineering is a new methodology for addressing meniscal injury or loss. Synovium may be an ideal source of cells for in vitro meniscal fibrocartilage formation, however, favorable in vitro culture conditions for synovium must be established in order to achieve this goal. The objective of this study was to determine cellularity, cell distribution, and extracellular matrix (ECM) formation of equine fibroblast-like synoviocytes (FLS) cultured on synthetic scaffolds, for potential application in synovium-based meniscal tissue engineering. Scaffolds included open-cell poly-L-lactic acid (OPLA) sponges and polyglycolic acid (PGA) scaffolds cultured in static and dynamic culture conditions, and PGA scaffolds coated in poly-L-lactic (PLLA) in dynamic culture conditions. Materials and Methods. Equine FLS were seeded on OPLA and PGA scaffolds, and cultured in a static environment or in a rotating bioreactor for 12 days. Equine FLS were also seeded on PGA scaffolds coated in 2% or 4% PLLA and cultured in a rotating bioreactor for 14 and 21 days. Three scaffolds from each group were fixed, sectioned and stained with Masson’s Trichrome, Safranin-O, and Hematoxylin and Eosin, and cell numbers and distribution were analyzed using computer image analysis. Three PGA and OPLA scaffolds from each culture condition were also analyzed for extracellular matrix (ECM) production via dimethylmethylene blue (sulfated glycosaminoglycan) assay and hydroxyproline (collagen) assay. PLLA coated PGA scaffolds were analyzed using double stranded DNA quantification as areflection of cellularity and confocal laser microscopy in a fluorescent cell viability assay. Results. The highest cellularity occurred in PGA constructs cultured in a rotating bioreactor, which also had a mean sulfated glycosaminoglycan content of 22. 3 µg per scaffold. PGA constructs cultured in static conditions had the lowest cellularity. Cells had difficulty adhering to OPLA and the PLLA coating of PGA scaffolds; cellularity was inversely proportional to the concentration of PLLA used. PLLA coating did not prevent dissolution of the PGA scaffolds. All cell scaffold types and culture conditions produced non-uniform cellular distribution. Discussion/Conclusion. FLS-seeding of PGA scaffolds cultured in a rotating bioreactor resulted in the most optimal cell and matrix characteristics seen in this study. Cells grew only in the pores of the OPLA sponge, and could not adhere to the PLLA coating of PGA scaffold, due to the hydrophobic property of PLA. While PGA culture in a bioreactor produced measureable GAG, no culture technique produced visible collagen. For this reason, and due to the dissolution of PGA scaffolds, the culture conditions and scaffolds described here are not recommended for inducing fibrochondrogenesis in equine FLS for meniscal tissue engineering.

Introduction The knee menisci are semilunar-shaped fibrocartilages with extracellular matrix (ECM) composed primarily of types I and II collagen, glycosaminoglycans (GAGs), and water ( Fithian, Kelly & Mow, 1990 ). It is now well established that intact menisci are crucial for the maintenance of normal joint function, however these critical structures are frequently injured in humans and animals. Meniscal tears are the most common knee injury in people, and arthroscopic meniscectomy represents the most common human orthopedic surgery performed annually ( Burks, Metcalf & Metcalf, 1997 ). Meniscal injuries are also a significant cause of lameness and decreased performance in horses ( Peroni & Stick, 2002 ; Walmsley, 1995 ; Walmsley, Phillips & Townsend, 2003 ); equines affected by naturally occurring meniscal tears may also be a viable model for the study of human meniscal injury. As the axial, avascular portion of the meniscus has a limited ability to heal spontaneously, ( Arnoczky & Warren, 1983 ; Kobayashi et al. , 2004 ), the majority of meniscal injuries are treated with partial menisectomy. However, this also results in eventual articular cartilage damage of the tibia and femoral condyles, and progression of debilitating osteoarthritis ( Arnoczky & Warren, 1983 ; Cox et al. , 1975 ). Thus tissue engineering new meniscal fibrocartilage is being investigated as a treatment for avascular meniscal injuries. Synovium may be an ideal cell source for meniscal tissue engineering. Synovium plays an important role in attempted vascular zone healing and regeneration ( Cisa et al. , 1995 ; Kobuna, Shirakura & Niijima, 1995 ; Ochi et al. , 1996 ; Shirakura et al. , 1997 ). Synovium has the ability to form fibrocartilaginous-like tissue in vivo in response to meniscectomy ( Cox et al. , 1975 ). In addition, synoviocytes have been reported to be an important element in cellular repopulation of meniscal allografts ( Arnoczky & Warren, 1983 ; Rodeo et al. , 2000 ). Synovial tissue progenitor cells, grossly indistinguishable in culture from type B or fibroblast-like synoviocytes (FLS), can undergo chondrogenesis in vitro ( De Bari et al. , 2001 ; Nishimura et al. , 1999 ). Taken together, these data indicate that synovium may be able to serve as a source for functional fibrocartilage in engineering meniscal tissue, provided the chondrogenic potential of synoviocytes can be optimized. Tissue engineering scaffolds must provide substrate and stability for cellular retention, intercellular communication, and cellular growth to allow seeded cells to proliferate extracellular matrix (ECM). As the scaffolds naturally degrade, the cellular ECM must be able to take on the biomechanical function and form previously designated by the scaffolds to maintain construct integrity. Thus a scaffold must be hydrophilic enough to allow cell adhesion but have a long enough half-life to not prematurely dissolve, which would prevent ECM proliferation and cell death. PGA (polyglycolic acid) and PLLA (poly-L-lactic acid) are biodegradable, biocompatible, polyesters, that are attractive for tissue engineering because they are readily available, can be easily processed into a variety of structures, and are approved by the Food and Drug Administration for a number of biomedical applications ( Lavik et al. , 2002 ). PGA has been successfully used as a scaffold for meniscal fibrochondrocytes in vivo ( Kang et al. , 2006 ) and cultured in vitro ( Aufderheide & Athanasiou, 2005 ) to form meniscal-like tissue. PLLA has been successfully used for in vitro tissue engineering of leporine meniscal fibrochondrocytes ( Esposito et al. , 2013 ; Gunja & Athanasiou, 2010 ), chondrocytes ( Sherwood et al. , 2002 ), and human fibroblasts ( Hee, Jonikas & Nicoll, 2006 ). PGA–PLLA combinations have also been successfully used for in vitro meniscal culture ( Ionescu & Mauck, 2013 ). In addition, chondrocytes cultured on PGA-PLLA mixtures versus collagen sheets contain more collagen type II and have stronger mechanical properties ( Beatty et al. , 2002 ) than single polymer scaffolds. Further investigation of combination use of PLLA combined with PGA for in vitro synoviocyte culture is warranted. Cartilage and fibrocartilage engineering with biodegradable scaffolds is most successful if uniform cell distribution is achieved ( Davisson, Sah & Ratcliffe, 1999 ; Pazzano, Mercier & Moran, 2000 ; Smith, Dunlon & Gupta, 1995 ), which is optimized through the use of rotating bioreactors ( Aufderheide & Athanasiou, 2005 ; Kim, Putnam & Kulik, 1998 ; Pazzano, Mercier & Moran, 2000 ). In addition, rotating bioreactors provide mechanical stimulation of cultured cells. This has a positive effect on cell differentiation, cell viability, extracellular matrix production, and compressive biomechanical properties, through mechanotransductive effects ( Davisson, Sah & Ratcliffe, 1999 ; Imler, Doshi & Levenston, 2004 ; Pazzano, Mercier & Moran, 2000 ; Smith, Dunlon & Gupta, 1995 ). Thus scaffold culture in a rotating bioreactor may represent a useful technique for synoviocyte-based engineering of functional meniscal tissue. Based on this prior research, we believe that both PGA and PLLA would be viable synthetic scaffolds for the in vitro culture of FLS for application in meniscal fibrocartilage tissue engineering. Thus, the first objective of this study was to (1) determine cell distribution and ECM formation of equine FLS seeded and cultured dynamically in a rotating bioreactor versus static seeding and culture, on two synthetic scaffold types, PGA and open-cell PLLA (OPLA). The second objective was to compare cell viability, distribution, and ECM formation of FLS cultured on 2% vs 4% PLLA coated PGA scaffolds, cultured for 14 or 21 days. Our hypothesis was that we would see no difference in equine FLS content, FLS distribution, and ECM formation between scaffold type, biomechanical culture environment, and culture duration. Materials and Methods Experiment 1 Tissue collection and monolayer cell culture Six 8. 0 mm × 8. 0 mm biopsies of synovial intima and subintima were obtained from both stifles of an adult American Quarter Horse, euthanatized according the American Veterinary Medical Association’s guidelines for humane euthanasia, for reasons unrelated to the study. The horse was determined to be free of orthopedic disease based on pre-mortem physical examination and post mortem gross examination of the joint. Tissue was placed in Dulbeccos’ Modified Eagle’s Media (DMEM) with 10% fetal bovine serum, 0. 008% Hepe’s buffer, 0. 008% non-essential amino acids, 0. 002% penicillin 100 I. U. /mL streptomycin 100 ug/mL, amphoterocin B 25 ug/mL, 0. 002% L-ascorbate, and 0. 01% L-glutamine in preparation for monolayer culture. Synovium was sectioned into 2. 0 mm × 2. 0 mm pieces using a #10 Bard Parker blade under sterile conditions. The tissue fragments were combined with sterile Type 1A clostridial collagenase solution (Type 1A Clostridial Collagenase; Sigma, St. Louis, MO) at a concentration of 7. 5 mg/mL of RPMI 1640 solution. The mixture was agitated at 37 °C, 5% CO2, 95% humidity for six hours. Cells were recovered through centrifugation, the supernatant decanted and the cellular pellet re-suspended in 5 mL of supplemented DMEM. The cell solution was transferred to a 25 cm 2 tissue culture flask containing 5 mL of supplemented DMEM. The flasks were incubated at 37 °C, 5% CO2, 95% humidity, with sterile medium change performed every 3 days. Synovial cells were monitored for growth using an inverted microscope until observance of 95% cellular confluence per tissue culture flask. At second passage cells were transferred to 75 cm 2 tissue culture flasks containing 11 mL of media. At 95% confluence the cells were subcultured until the 4th cell passage had been reached. At 4th passage cells were removed from flasks, counted using the Trypan Blue exclusion assay ( Strober, 2001 ), and transferred to scaffold culture as described below. Scaffolds A non-woven polyglycolic acid (PGA; Tissue Scaffold, Synthecon, Houston, TX) felt, 3 mm thick, with 10 µm diameter fibers was utilized for this study. The open-cell poly-lactic acid (OPLA sponge, BD Biosciences, Bedford, MA) utilized were 5. 0 mm × 3. 0 mm, non compressible, cylindrical sponges. The average OPLA sponge pore size was 100–200 µm with a hydration capacity of 30 µl/scaffold. PGA and OPLA scaffolds were sterilized in ethylene oxide. Following sterilization, the PGA felt was cut using a sterile Baker’s biopsy punch to create 5. 0 mm diameter discs prior to cell culture. Dynamic culture Twelve PGA scaffolds (PGA-D group) and 12 OPLA sponges (OPLA-D group) were placed in separate 110 mL vessel flasks of a rotating bioreactor system (Rotating Bioreactor System, Synthecon, Houston, TX ( Fig. 1 ) containing 110 mL of supplemented DMEM. The scaffolds were presoaked for 24 h in the bioreactor at 37 °C, 5% CO2, 95% humidity, prior to cell introduction. Fourth passage FLS were removed from the tissue culture flasks enzymatically (Accutase Innovative Cell Technologies, San Diego, CA) and counted. Cells were added to the 110 mL bioreactor flasks at a concentration of 1 million cells/scaffold via a 60 cc syringe, slowly injected over several minutes. For the duration of the study the bioreactor vessels were rotated at 51. 1 rpm to allow the scaffolds to free-float and rotate within the culture medium, without contacting the inner bioreactor surfaces. Cultures were maintained at 37 °C, 5% CO2, 95% humidity. Fifty percent of the cell culture medium volume was changed using sterile technique every 3 days. Cell counts were performed on discarded media for the first two media changes. 10. 7717/peerj. 353/fig-1 Figure 1 Dynamic culture: rotating bioreactor apparatus. Rotating wall bioreactor flask (110 mL) containing media and PGA scaffolds seeded with equine fibroblast-like synoviocytes (A). Flasks loaded on the rotating base apparatus; flasks rotate around their longitudinal axis (B). Static culture Twelve PGA scaffolds (PGA-S group) and 12 OPLA sponges (OPLA-S group) were placed individually in non-treated 24 well tissue culture plates, each well containing 2 mL of supplemented DMEM ( Fig. 2 ). The scaffolds were presoaked for 24 h at 37 °C, 5% CO2, 95% humidity, prior to cell introduction. Then FLS were transferred from monolayer culture as described above, and slowly over 3 min, pipetted on top of the scaffolds in solution, at 1 million cells per scaffold in each well. The plates were maintained at 37 °C, 5% CO2, 95% humidity, with 50% cell culture medium changed every 3 days. Cell counts were performed on discarded media for the first 2 media changes. 10. 7717/peerj. 353/fig-2 Figure 2 Static culture of equine fibroblast-like synoviocytes on PGA scaffolds in a 24 well tissue culture plate, with each well containing 2 mL of supplemented DMEM. Histologic analysis All scaffolds were harvested on the 12th day of culture. Six scaffolds from each group (PGA-S, PGA-D, OPLA-S, OPLA-D) were fixed in 10% buffered formalin, embedded in paraffin, sectioned, and stained with Masson’s Trichrome, Safranin–O, and Hematoxylin and Eosin. Histologic specimens were examined at 10× magnification (Zeiss Microscope; Carl Zeiss, Thornwood, NY). Images of each section, (three from the scaffold periphery and three from the scaffold center) at 2 o’clock, 6 o’clock and 10 o’clock positions ( Fig. 3 ) were digitally captured by a digital camera (Olympus DP-70 Olympus, Melville, NY) and saved as tagged-image file format images. Digital image analysis was performed as previously validated ( Amin et al. , 2000 ; Benzinou, Hojeij & Roudot, 2005 ; Girman et al. , 2003 ; Goedkoop, Rie & Teunissen, 2005 ) whereby cellular density was assessed using a thresholding algorithm ( Loukas et al. , 2003 ) using computer image analysis (Fovea 3. 0, Reindeer Graphics, Asheville, NC). This algorithm allows quantification of cellular nuclei based on their histogram values. All cell counts were additionally validated by hand counts. Safranin-O staining, indicating presence of GAG, and Masson’s Trichrome staining, indicating presence of collagen, were subjectively evaluated and recorded. 10. 7717/peerj. 353/fig-3 Figure 3 Histologic cell counting method. Method for viewing all scaffolds to standardize cell counts and determine regional cell count differences between the scaffold center and periphery. Cells were counted at the periphery and central regions (dark dotted circles) of each scaffold (cross- hatched circle) using digital image analysis; peripheral cell counts (light dotted circles) were obtained at the 2 o’clock, 6 o’clock and 19 o’clock positions. Circles represent a low power (10× objective) field of view. Biochemical ECM analysis Three cultured scaffolds from each group were analyzed for glycosaminoglycan (GAG) and collagen production. Wet weight of each scaffold was obtained. GAG content of the scaffold was performed using the Dimethyl-methylene Blue Sulfated Glycosaminoglycan assay ( Farndale, Buttle & Barrett, 1986 ). Collagen content of the cultured scaffolds was assessed using the hydroxyproline assay, as described by Reddy et al. ( Reddy & Enwemeka, 1996 ). Statistical methods Data were tested for normality using a Shapiro–Wilk test. Data were then analyzed using a one way analysis of variance followed by a Tukey’s test, to compare the effect of scaffold type and seeding technique on cell counts and ECM quantity. To determine significance between periphery and central cell counts within each scaffold, a paired, 2-tailed student’s t-test was performed. For all tests significance was set at P < 0. 05. All statistical analyses were performed using a statistical software program (GraphPad Prism Version 6, San Diego, CA). Experiment 2 Scaffolds PLLA was dissolved in methylene chloride as a 2% or 4% solution. The 2% and 4% PLLA solution each was applied to a 3. 0 mm thick sheet of the same, above- described, non-woven PGA felt, using an eye-dropper. Following PLLA treatment, the treated felt was placed in a vacuum dessicator overnight and then sterilized in ethylene oxide. Following sterilization, the 2% and 4% PLLA modified PGA felts were cut into fourteen 5 mm × 7 mm × 3 mm square scaffolds using sterile scissors and a #10 bard parker blade ( Fig. 4 ). 10. 7717/peerj. 353/fig-4 Figure 4 PLLA coated scaffolds. Scanning electron microscopy of a 2% PLLA coated PGA scaffold (A) and a 4% PLLA coated scaffold (B) prior to cell seeding; bar = 100 μm. Tissue collection and monolayer cell culture Synovial intima/subintima was harvested from the stifles of two mixed breed, adult horses euthanatized according the American Veterinary Medical Association’s guidelines for humane euthanasia, for reasons unrelated to the study. These horses were also determined to be free of orthopedic disease based on pre-mortem physical examination and post mortem gross examination of the joint. The tissue was transported, minced and digested as described above. Cells were recovered through centrifugation, the supernatant decanted and the cellular pellet re-suspended in 5 mL of supplemented DMEM. The cell solution was transferred to a 25 mL tissue culture flask containing 5 mL of supplemented DMEM. Cells were grown in monolayer culture, under the conditions described above, until the 4th cell passage had been reached. Dynamic culture Fourteen 2% PLLA coated PGA scaffolds and fourteen 4% PLLA coated PGA scaffolds were placed in separate 110 mL vessel flasks of the rotating bioreactor system containing 110 mL of supplemented DMEM. The scaffolds were presoaked for 24 h in the bioreactor at 37 °C, 5% CO2, 95% humidity, prior to cell introduction. After this time it was noted that the scaffolds were floating at the apex of the flasks. Using sterile surgical technique, scaffolds were sterily removed from the flasks, pierced centrally, and strung on loops of 3-0 nylon surgical suture with knots placed adjacent to the scaffolds to prevent bunching on the line. Seven scaffolds were placed per suture. The strings of scaffolds were then placed back in to the bioreactors and presoaked for another 12 h, at which time complete hydration and submersion were achieved ( Fig. 5 ). 10. 7717/peerj. 353/fig-5 Figure 5 Positioning of PLLA coated scaffolds in the rotating bioreactor. Rotating wall bioreactor flask containing 2% PLLA coated PGA scaffolds, strung on suture to ensure equal submersion and positioning in the rotating flask. Scaffolds were then dynamically seeded. Synovial membrane cells were removed from the tissue culture flasks using as described above and counted using the Trypan Blue exclusion assay ( Strober, 2001 ). Cells were added to the bioreactor flasks at a concentration of 1 million cells/scaffold. For the duration of culture, the bioreactor was maintained at 37 °C, 5% CO2, 95% humidity at 51. 1 rpm. Fifty percent of the cell culture medium volume was changed using sterile technique every 3 days. Seven scaffolds were harvested on day 10 of culture, and 7 scaffolds were harvested on day 21 of culture. Determination of cell viability Cell viability was determined with the use of ethidium homodimer-1 (4 ul/ml PBS) and Calcein AM (Acetoxymethylester) (0. 4 ul/ml PBS) fluorescent stains (Invitrogen, Carlsbad, CA) and the use of Confocal Laser Microscopy. The Confocal Laser Microscope consists of the BioRad Radiance 2000 confocal system coupled to an inverted microscope (Olympus IX70 Olympus, Melville, NY) equipped with Krypton–Argon and red diode laser. Approximately 1. 0 mm sections were made from the halved scaffold using a rotary paper cutter. A section from each scaffold’s cut center and a section from each scaffold’s periphery was examined. Sections were incubated with the staining agents for 30 min at room temperature, placed on a glass microscope slide, moistened with several drops of PBS, and stained using the fluorescent double labeling technique. The sections were examined under 10 × magnification. Images were taken of each specimen as described above, (three from the section periphery and three from the section center) at the 2 o’clock, 6 o’clock and 10 o’clock positions. Images were digitally captured as described above. Live and dead cell counts were determined by hand counts. DNA quantification One half of each construct was lyophilized and a dry weight obtained. Samples were incubated in 1. 0 ml Papain Solution (2 mM Dithiothreitol and 300 ug/ml Papain) at 60 °C in a water bath for 12 h. A double stranded DNA quantification assay (Quant-iT PicoGreen™ Invitrogen, Carlsbad, CA) was performed. Double stranded DNA extracted from bovine thymus was mixed with TE buffer (Invitrogen, Carlsbad, CA) to create standard DNA concentrations of 1, 000, 100, 10, and 1 ng/mL. The standards and 100 uL of each papain digested sample (used in the above GAG and hydroxyproline assays) were added to a black 96 well plate. 100 uL of 2 ug/mL of Pico Green reagent was added to each well and the plate was incubated for 5 min. Sample fluorescence was read at 485 nm excitation/528 nm emission by a spectrophotometric plate reader (Synergy HT–KC-4; BioTek, Winooski, VT). Absorbances were converted to ng/mL concentrations and total double stranded DNA yield in ng using FT4 software (BioTek, Winooski, VT). Statistical methods Data were tested for normality using a Shapiro–Wilk test. Scaffold weights were compared using a 2-tailed paired t-test. Scaffold dsDNA content was analyzed using a repeated- measures analysis of variance with a Geisser-Greenhouse correction. Significance was set at p < 0. 05. All statistical analyses were performed using a statistical software program, (GraphPad Prism Version 6, San Diego, CA). Results Experiment 1 As determined by the Trypan Blue exclusion assay, viability of cells at the time of transfer from monolayer culture to static or dynamic seeding was 98. 6%. No live cells were detected in any of the media changes for either static or dynamically cultured scaffolds, indicating that viable cells rapidly adhered to the scaffolds. At the time of harvest upon gross examination, the fibers of the PGA scaffolds and the sponge surface of the OPLA scaffolds were still visible. PGA scaffolds subjectively appeared more translucent. Despite equal cell seeding concentrations, the effect of dynamic bioreactor culture on cell content of PGA scaffolds (PGA-D versus PGA-S) was to increase scaffold cellularity ( P < 0. 001). This was also found in OPLA-D versus OPLA-S scaffolds ( P = 0. 028). The effect of scaffold type also significantly increased scaffold cellularity of PGA-D versus OPLA-D ( P = 0. 017), while OPLA-S had great cellularity than PGA-S ( P = 0. 0217; Table 1 ). 10. 7717/peerj. 353/table-1 Table 1 The effect of seeding and cell culture biomechanical environment and the effect of scaffold type on scaffold cellularity. Cell count (Mean number of cells per 10× objective field ±SD) Effect of biomechanical environment (dynamic vs static culture) Scaffold type Biomechanical environment Dynamic culture Static culture PGA 1128 ± 575 cells 54 ± 34 cells P < 0. 001 OPLA 375 ± 118 cells 301 ± 65 cells P = 0. 028 Effect of scaffold type (PGA vs OPLA) P = 0. 017 P = 0. 0217 All groups, with the exception of OPLA-S, showed increased cellular distribution to the periphery of the scaffolds ( Table 2 ). Due to the shape of the OPLA-S on histological sectioning, there was overlap of central and peripheral fields of view, precluding accurately localized cell counts; peripheral cell count was 307 ± 52 and central cell count was 287 ± 80 ( P < 0. 464). Cells grew in whorls, strands, and sheets on the PGA scaffolds, while cells grew in clumps on the surface pores of the OPLA sponges ( Fig. 6 ). 10. 7717/peerj. 353/fig-6 Figure 6 Histologic cell distribution on PGA and OPLA scaffolds. Micrographs of scaffolds seeded with equine fibroblast-like synoviocytes; Hematoxylin and Eosin staining, 10× objective magnification; bar = 100 μm. (A) PGA scaffold cultured in a static environment; (B) PGA scaffold cultured in a dynamic environment (rotating bioreactor); (C) OPLA scaffold cultured in a dynamic environment (rotating bioreactor); (D) OPLA scaffold cultured in a static environment. Note the intact PGA fibers (open arrow) and the cells located in clumps in the pores of the OPLA scaffold (closed arrows). 10. 7717/peerj. 353/table-2 Table 2 Peripheral and central cell count (Mean number of cells per 10× objective field ±SD). Scaffold Peripheral cell count Central cell count P -value PGA-D 1433 ± 487 724 ± 314 P < 0. 001 PGA-S 80 ± 28 28 ± 11 P < 0. 001 OPLA-D 476 ± 90 295 ± 55 P < 0. 001 Staining for collagen and glycosaminoglycan using Masson’s Trichrome and Safranin-O, respectively, was negative for extracellular matrix production in all sections of all scaffold types and culture conditions evaluated. In the PGA-D group, the dimethylmethylene blue assay detected a mean of 22. 29 µg of GAG per scaffold, (range 19. 34–28. 13 µg), with a mean % GAG scaffold content of 0. 0345% (µg GAG per µg scaffold wet weight). No GAG was detected in OPLA constructs or PGA-S constructs. The hydroxyproline assay did not detect collagen production in any group. Experiment 2 Post PLLA modification, mean scaffold dry weights before soaking and seeding were 1. 01 mg for 2% PLLA coating and 1. 52 mg for 4% PLLA coating ( P < 0. 001). Scaffold dry weights decreased over time. Mean lyophilized weight on day 10 for 2% PLLA coating was 0. 533 mg, which decreased to 0. 257 mg on day 21 ( P = 0. 02). Mean lyophilized weight on day 10 for 4% PLLA coating was 0. 481 mg, which decreased to 0. 381 mg on day 21 ( P = 0. 043). Scaffold cellularity as measured by dsDNA content increased over time: for the 2% group, day 10 cellularity was 102. 6 ng dsDNA/mg dry weight, and on day 21 it was 281. 79 ( P = 0. 021). On day 10 for the 4% group, dsDNA content was 111. 01 ng dsDNA/mg dry weight and on day 21 it was 140. 2 ng dsDNA/mg dry weight ( P = 0. 032; Fig. 7 ). 10. 7717/peerj. 353/fig-7 Figure 7 Double stranded DNA content of PLLA coated scaffolds. Mean ± Standard Error of the Mean (SEM) of dsDNA content of PGA scaffolds coated in 2% PLLA and 4% PLLA, seeded dynamically and cultured in a rotating bioreactor for 14 days and 21 days. A bar and (*) indicates a significant difference between two treatment groups ( P < 0. 05). PLLA coating also affected scaffold dsDNA content. Scaffolds with the 2% PLLA coating had greater dsDNA content than the 4% PLLA coating on day 21 ( P = 0. 003), but not on day 10 ( P = 0. 602; Fig. 7 ). As visible under confocal microscopy, cells only adhered to the surface of exposed PGA fibers and had poor to no penetration to the scaffold centers in all PLLA coated scaffolds. Viable cell numbers were estimated only because of the marked cellular clumping; all scaffolds showed mixtures of viable and non-viable cells localized in clumps on the scaffold outer margins ( Figs. 8 – 11 ). Histologic examination of H + E stained constructs revealed minimal cellular adhesion to the PLLA, in all groups at all times, with cells growing primarily on the exposed PGA scaffold, in tightly packed clumps, or adhering to exposed fibers of PGA. No extracellular matrix was observed in any scaffolds on histologic analysis, which also reflected the uneven cellularity ( Fig. 12 ). 10. 7717/peerj. 353/fig-8 Figure 8 Cell viability: 2% PLLA scaffolds. Photomicrographs of 2% PLLA coated PGA constructs harvested on day 10, under standard light (column A ) and under laser confocal microscopy (column B ), using the calcein AM-ethidium homodimer live-dead assay. Images represent scaffold transverse cross sections (row T ) and scaffold surface coronal sections (row C ). Green stained cells are alive, red stained cells are dead. 10× objective magnification; bar = 100 μm. 10. 7717/peerj. 353/fig-9 Figure 9 Cell viability: 2% PLLA scaffolds. Photomicrographs of 2% PLLA coated PGA constructs harvested on day 21, under standard light (column A ) and under laser confocal microscopy (column B ), using the calcein AM-ethidium homodimer live-dead assay. Images represent scaffold transverse cross sections (row T ) and scaffold surface coronal sections (row C ). Green stained cells are alive, red stained cells are dead. Note the spurious red staining of scaffold PGA fibers. 10× objective magnification; bar = 100 μm. 10. 7717/peerj. 353/fig-10 Figure 10 Cell viability assay: 4% PLLA scaffolds. Photomicrographs of 4% PLLA coated PGA constructs harvested on day 10, under standard light (column A ) and under laser confocal microscopy (column B ), using the calcein AM-ethidium homodimer live-dead assay. Images represent scaffold transverse cross sections (row T ) and scaffold surface coronal sections (row C ). Green stained cells are alive, red stained cells are dead. Note the spurious red staining of PGA fibers. 10× objective magnification; bar = 100 μm. 10. 7717/peerj. 353/fig-11 Figure 11 Cell viability: 4% PLLA scaffolds. Photomicrographs of 4% PLLA coated PGA constructs harvested on day 21, under standard light (column A ) and under laser confocal microscopy (column B ), using the calcein AM-ethidium homodimer live-dead assay. Images represent scaffold transverse cross sections (row T) and scaffold surface coronal sections (row C ). Green stained cells are alive, red stained cells are dead. Note the spurious red staining of PGA fibers. 10× objective magnification; bar = 100 μm. 10. 7717/peerj. 353/fig-12 Figure 12 Distribution of cells on 2% and 4% PLLA coated PGA scaffolds. Photomicrographs of 2% PLLA coated PGA scaffolds harvested on day 10 (row 1 ) and day 21 (row 2 ), and 4% PLLA coated PGA scaffolds harvested on day 10 (row 3 ) and day 21 (row 4 ), H+E staining. Column A represents images of the center of the construct and column B represents images taken of the scaffold periphery. Note that the cells have grown in dense clusters; 10× objective magnification; bar = 100 μm. Discussion The current study analyzed the effect of scaffold type, biomechanical stimuli, and culture duration on FLS seeding and production of specific meniscal ECM constituents. We found that FLS-seeded PGA constructs cultured in a rotating bioreactor had the highest cellularity, with a mean sulfated glycosaminoglycan content of 22. 3 µg per scaffold. PGA constructs cultured in static conditions had the lowest cellularity. For PLLA coated PGA, increasing concentration of PLLA decreased scaffold cellularity, while increased culture time increased scaffold cellularity, as determined by the dsDNA assay. A non-uniform cellular distribution was observed for all scaffold types and culture conditions. Bioreactor culture provides a number of benefits over static culture which would account for the higher cellularity of PGA-D and OPLA-D versus PGA-S and OPLA-S scaffolds. The rotating wall bioreactor used in this study provided a dynamic, laminar fluid shear, which perfuses scaffold cultured cells ( Bilodeau & Mantovani, 2006 ), and thereby encourages cell survival and proliferation by providing efficient transport of nutrients, gases, catabolites, and metabolites and maintaining physiologic media pH ( Gooch et al. , 2001 ; Vunjak-Novakovic et al. , 1998 ). Mixing of culture media also promotes cell seeding by creating matched relative velocities of cells and scaffolds, particularly on non-woven PGA scaffolds ( Vunjak-Novakovic et al. , 1998 ). In addition, the rotating wall bioreactor limits cellular stress by reducing strong shear forces and cellular impact on the walls of the bioreactor ( Bilodeau & Mantovani, 2006 ). However, in our study, scaffold characteristics such as scaffold density and hydrophilicity may have negated the advantages of bioreactor culture, as seen with OPLA or PLLA coated scaffolds, which had fewer cells and markedly uneven cell distribution, respectively. A higher cell count was found on PGA-D versus OPLA-D, indicating either better adherence or cell proliferation on PGA. Non-woven PGA scaffolds favor cellular capture and retention because of their polar surface properties and high surface area for cellular adhesion ( Day, Boccaccini & Shurey, 2004 ; Moran, Pazzano & Bonassar, 2003 ). Cellularity of PGA-D was further increased by the open weave and low density (45–77 mg/cc) of PGA scaffolds supports cellular proliferation through superior flow-through of culture media and nutrient delivery ( Vunjak-Novakovic et al. , 1998 ). This is in contrast to the highly dense (871 mg/cc) OPLA sponges with non-communicating pores, which could inhibit nutrient and gas transfer to seeded cells ( Pazzano et al. , 2004 ; Pazzano, Mercier & Moran, 2000 ; Wu, Dunkelman & Peterson, 1999 ). For PLLA covered PGA scaffolds, cells were located primarily on exposed PGA fibers, and scaffold cellularity was inversely proportional to the concentration of PLLA. Although PLLA is widely used in tissue-engineering applications because of its slower degradation characteristics, strength, and mechanical properties, its hydrophobic, inert nature can affect cell–matrix interactions and decrease cellular adhesion ( Moran, Pazzano & Bonassar, 2003 ). While the PLLA coating of PGA scaffolds was intended to protect from premature scaffold dissolution, we observed that with longer duration of culture, scaffolds appeared to be more fragile to disruption with forceps manipulation, particularly on the outer edges as well as around the centrally placed suture. In agreement with this observation, all scaffold dry weights dropped over time, indicating scaffold dissolution. Thus PLLA did not prevent PGA hydrolysis and decreased scaffold integrity. PLLA coating also provided a hydrophobic barrier to centralized cell seeding and ingrowth. Thus, for the future study of scaffold seeded equine FLS, use of PLLA type scaffolds is not recommended. Cell distribution across all scaffolds was uneven, in contrast to previous reports on bioreactor chondrocyte culture ( Mahmoudifar & Doran, 2005 ; Pazzano et al. , 2004 ). Lower central cell density in our scaffolds may have indicated poor axial cell penetration and in-growth. Alternatively, higher peripheral cellularity could reflect increased peripheral cell division caused by increased exposure to media nutrients, gas exchange, and mechanotransductive effects ( Mahmoudifar et al. , 2002 ). Additionally the OPLA scaffolds had clumped cell distribution in the outermost pores. The OPLA sponge porosity may not allow uniform cell distribution; the 100–200 µm pores do not consistently communicate with each other. While OPLA-S did not have different peripheral and central cell counts, this was due to an artifact of the sponge shape and precluded distinction of peripheral cells from central cells. To increase central scaffold cell content, flow-through bioreactors ( Bilodeau & Mantovani, 2006 ) may have greater cell seeding efficiencies than rotary bioreactors. Alternatively, cells may be seeded at the time of scaffold formation, such as during hydrogel synthesis, to insure central scaffold cellularity ( Narita et al. , 2009 ). The culture conditions utilized in the present study resulted in minimal to no ECM formation, in contrast to other studies. The mean GAG content of the PGA-D scaffold of 0. 0345% (wet weight basis) was lower than the 0. 6–0. 8% wet weight in the normal meniscus, and thus represents a sub-optimal response for engineering purposes ( AufderHeide & Athanasiou, 2004 ). Synoviocytes typically produce collagen type I constitutively, ( Garner, 2000 ; Levick, Price & Mason, 1996 ), however production and deposition of hydroxyproline was not detected in this study. The most likely reason for this failure of ECM formation was lack of culture with a specific fibrochondrogenic media. For example, culture with recombinant transforming growth factor-beta, insulin-like growth factor-1 and basic fibroblast growth factor have been shown to induce in vitro collagen formation in human synoviocytes ( Pei, He & Vunjak-Novakovic, 2008 ; Pei et al. , 2008 ). Treatment of equine FLS with recombinant chondrogenic growth factors, in addition to the scaffold and bioreactor culture conditions used in the present study, resulted in greater type II collagen and aggrecan gene expression ( Fox et al. , 2010 ). Reported scaffold seeding concentrations for cartilage tissue engineering include 30, 000 fibroblasts/mL ( Day, Boccaccini & Shurey, 2004 ); 600, 000 chondrocytes/mL ( Stading & Langer, 1999 ); 5 million chondrocytes/mL ( Griffon, Sedighi & Sendemir-Urkmez, 2005 ); and 10 million chondrocytes/mL ( Hu & Athanasiou, 2005 ). Our seeding density of 1 million equine FLS per scaffold may have been too low, as dense cell aggregates are required for meniscal developmental fibrochondrogenesis ( Clark & Ogden, 1983 ) due to the embryonic community effect ( Gurdon, Lemaire & Kato, 1993 ). In the present study the FLS were exposed to the mild shear forces and hydrostatic pressurization in a rotating bioreactor ( Mauck et al. , 2002 ) which may not have been the optimal type of forces required for synovial collagen I formation. A combination of in vitro tensile and compressive forces ( AufderHeide & Athanasiou, 2004 ; Benjamin & Ralphs, 1998 ) may be required to support formation GAG ( Valiyaveettil, Mort & McDevitt, 2005 ) and types I and II collagen ( Kambic & McDevitt, 2005 ), the major ECM components of fibrocartilage. Cell culture on scaffolds may also result in cellular stress shielding, thereby resulting in suboptimal matrix formation ( Huey & Athanasiou, 2011 ). Synovial macrophages may have contaminated our FLS cultures, thereby also decreasing ECM formation ( Pei, He & Vunjak-Novakovic, 2008 ; Bilgen et al. , 2009 ) and future studies should include negative isolation of macrophages. Additionally, co-culture with meniscal fibrochondrocytes as decribed by Tan and workers ( Tan, Zhang & Pei, 2010 ) may have also helped fibrochondrogenic differentiation of equine FLS and will be the focus of future studies. Increased culture time may also be beneficial to ECM formation; other studies show time dependent ECM expression ( Griffon, Sedighi & Sendemir-Urkmez, 2005 ; Mueller et al. , 1999 ; Sha’ban et al. , 2008 ). One study of synovial chondrogenesis on PGA scaffolds utilized a longer culture duration of 60 days, with successful ECM formation ( Sakimura et al. , 2006 ). Despite better cellularity, PGA scaffolds began losing integrity over the culture period, even when coated with PLLA. Unless rapid ECM formation can be achieved before dissolution occurs, PGA hydrolyzes too quickly (t1/2 = 16 days) for the purpose of long term meniscal fibrocartilage synthesis. Treatment with chondrogenic or fibrochondrogenic media may induce production of ECM, thus making the culture systems described here more feasible for meniscal tissue engineering. Conclusion In conclusion, we reject the null hypothesis; dynamic cell seeding and culture, as well as increased culture duration, increased scaffold cellularity. Scaffold type also affected cellularity; for bioreactor culture, PGA had higher cell counts versus OPLA, while OPLA had higher cell counts versus PGA in static culture. Cells could only grow unevenly in the pores of the OPLA sponge, and cells could not adhere to the PLLA coating of PGA scaffolds. Increasing the concentration of PLLA coating on a PGA scaffold decreased the cellularity of the scaffold, and did not prevent scaffold dissolution. While PGA culture in a bioreactor produced measureable GAG, no culture technique produced visible collagen. For this reason, and due to the dissolution of PGA scaffolds, the exact culture of conditions described here are not recommended for inducing equine fibrochondrogenesis towards meniscal tissue engineering. Further research is recommended to enhance extracellular matrix production through additional biomechanical and biological stimulation, including treatment with chondrogenic media, increased culture duration, and increased cell seeding concentrations.

10. 7717/peerj. 3665

PeerJ

Preparation and characteristics of gelatin sponges crosslinked by microbial transglutaminase

Microbial transglutaminase (mTG) was used as a crosslinking agent in the preparation of gelatin sponges. The physical properties of the materials were evaluated by measuring their material porosity, water absorption, and elastic modulus. The stability of the sponges were assessed via hydrolysis and enzymolysis. To study the material degradation in vivo, subcutaneous implantations of sponges were performed on rats for 1–3 months, and the implanted sponges were analyzed. To evaluate the cell compatibility of the mTG crosslinked gelatin sponges (mTG sponges), adipose-derived stromal stem cells were cultured and inoculated into the scaffold. Cell proliferation and viability were measured using alamarBlue assay and LIVE/DEAD fluorescence staining, respectively. Cell adhesion on the sponges was observed by scanning electron microscopy (SEM). Results show that mTG sponges have uniform pore size, high porosity and water absorption, and good mechanical properties. In subcutaneous implantation, the material was partially degraded in the first month and completely absorbed in the third month. Cell experiments showed evident cell proliferation and high viability. Results also showed that the cells grew vigorously and adhered tightly to the sponge. In conclusion, mTG sponge has good biocompatibility and can be used in tissue engineering and regenerative medicine.

Introduction Tissue engineering aims to repair and reconstruct damaged tissues and organs. Biological scaffolds designed for tissue engineering not only provide mechanical support for cell growth but also provide a micro environment that can regulate cell behavior and tissue regeneration. One of the challenges in scaffold design is in ensuring good mechanical structure and properties, rich active groups, as well as excellent biocompatibility and biodegradability. In addition, the design of biomaterials should mimic the physical characteristics and biological attributes of natural extracellular matrix (ECM) as much as possible ( Bencherif et al. , 2012 ; Bencherif, Braschler & Renaud, 2013 ; Kennedy et al. , 2014 ; Bencherif et al. , 2015 ; Xiao et al. , 2015 ). ECM is a secretory product of cells, which can form highly ordered insoluble aggregates that combine with cells to form various tissues and organs. ECM provides a 3D space for cell growth, affects cell adhesion, migration, proliferation, apoptosis, and signal transduction, and is involved in the development of organisms and the process of inflammatory response. Collagen is the major structural protein of ECM. In vivo, they can self-assemble to form collagen fibers of highly ordered three-helix structure with high mechanical strength. Gelatin is a hydrolysis product of collagen, which retains many functional groups that can be identified by cells. Therefore, gelatin is good for cell adhesion. Gelatin also has good hydrophilicity, biodegradability, and low antigenicity. Thus, it is a popular source of many composite biomaterials. However, pure gelatin material is brittle and soluble in water, which hinders its application in tissue engineering. To date, findings of previous work show that the mechanical and thermal properties of gelatin material can be improved by crosslinking gelatin molecules. The crosslinking methods include physical crosslink (dehydrothermal treatment ( Nadeem et al. , 2013 ; Siimon et al. , 2014 ); UV irradiation ( Bhat & Karim, 2014 ); chemical crosslink, such as glutaraldehyde ( Dainiak et al. , 2010 ; Ito et al. , 2014 ; Poursamar et al. , 2016 ), carbodiimide ( Pieper et al. , 1999 ; Chimenti et al. , 2011 ; Yeh et al. , 2012 ), and genipin ( Nadeem et al. , 2013 ; Poursamar et al. , 2016 ); enzymatical crosslink, such as transglutaminase ( Yung et al. , 2007 ; Kuwahara et al. , 2010 ; Yamamoto et al. , 2013 ) and horseradish peroxidase ( Sakai et al. , 2009 ; Chuang et al. , 2015 )). Physical crosslinking methods, such as dehydrothermal treatment, may cause degeneration of gelatin protein. Furthermore, UV irradiation does not easily penetrate thick material, leading to incomplete crosslinking. In the use of chemical crosslinking agents, the side effects of residual crosslinkers in scaffolds need to be considered. Chemical crosslinking agents, such as glutaraldehyde, glycerin aldehyde, formaldehyde, epoxy resin, and carbodiimide, have cytotoxicity, which affects cell growth and even lead to cell apoptosis. Transglutaminase has a remarkable ability of protein crosslinking. This enzyme catalyzes acyl-transfer reactions between λ-carboxyamide groups of glutamine residues and ε-amino groups of lysine residues, resulting in the formation of ε-(λ-glutaminyl) lysine intra and intermolecular crosslinked proteins ( Halloran et al. , 2008 ). In the past, mammalian-derived transglutaminase was limited by its high production cost. At present, the transglutaminase extracted from microorganisms have high yield and low price, which make it suitable as a crosslinking agent ( Collighan & Griffin, 2009 ). Our previous study ( Yang et al. , 2016 ) has shown that mTG crosslinked gelatin hydrogel has good biocompatibility and can support the growth and proliferation of adipose-derived stromal stem cells (ADSCs). We hypothesized that the preparation of mTG crosslinked gelatin sponges (mTG sponges) by freeze drying may also be suitable for tissue engineering applications. To the best of our knowledge, no study has been conducted from this perspective. In the current study, mTG was used as a crosslinking agent in the preparation of porous gelatin sponges. The physical properties of the sponges were evaluated by measuring their material porosity, water absorption, and elastic modulus. The stability of the sponges were assessed via hydrolysis and enzymolysis. To study the material degradation in vivo, the morphology and histochemical staining of the subcutaneously implanted sponges were analyzed in the animal experiments of rats for 1–3 months. To further evaluate the cell compatibility of mTG sponges, ADSCs were cultured and inoculated into the scaffold. Cell proliferation and viability were measured via alamarBlue assay and LIVE/DEAD fluorescence staining, respectively. Cell adhesion on the sponges was observed by scanning electron microscopy (SEM). Materials and Methods Preparation of gelatin sponges Gelatin sponge is a freeze-dried product of gelatin hydrogel. Thus, the gel needs to be prepared first. The mTG crosslinked gelatin hydrogel was prepared as described in our previous publication ( Yang et al. , 2016 ). Gelatin powder (type A, 300 Bloom; Sigma, St. Louis, MO, USA) was dissolved in deionized water at 50 °C to make 4% solution and sterilized through a 0. 22 μm syringe filter. mTG powder (enzymatic activity >100 U per gram; Bomei, China) was dissolved in phosphate-buffered saline (PBS) to make 10% solution and also sterilized through a syringe filter. Gelatin hydrogel was obtained by mixing gelatin and mTG solution with a ratio of 40 μL of mTG per milliliter gelatin solution. The mixture was incubated at 37 °C for gelatinization. The resultant hydrogel was frozen at −20 °C for 8 h and freeze dried for 48 h to produce the mTG sponges. Un-crosslinked gelatin sponge was made by cooling the 4% gelatin solution at 4 °C for 2 h, freezing at −20 °C for 8 h, and freeze drying for 48 h. Samples of the aforementioned sponges in dry and wet states (after immerging in PBS for 2 h) were observed under a stereoscopic microscope (Jinteng, Tianjin, China) attached to a CMOS camera (MD50; Guangzhou Ming-Mei Technology Co. , Ltd. , Guangzhou, China). Porosity of mTG sponges The porosity of gelatin sponges was measured by a liquid-displacement method described in the literature ( Kanokpanont et al. , 2012 ). A pre-weighed mTG sponge was placed into a known volume ( V 1 ) of absolute ethanol and degassed for 5 min by a vacuum pump. The full volume of ethanol and the soaking sponge was recorded as V 2. The ethanol-soaking sponge was then abandoned, and the remaining volume of ethanol was recorded as V 3. The porosity (ε 1 ) of the sponge was calculated as follows: (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}{\varepsilon }_{1}=({V}_{1}-{V}_{3})/({V}_{2}-{V}_{3})\times 100\text{%}\end{eqnarray*}\end{document} ε 1 = V 1 − V 3 ∕ V 2 − V 3 × 100 % Uniaxial tensile test The mechanical property of dry and wet mTG sponges was measured on a uniaxial mechanical testing apparatus (HPE; Yueqing Handpi Instruments Co. , Ltd, Wenzhou, Zhejiang, China) equipped with 20 N capacity. Wet sponges were prepared by dipping the dry sponges in PBS for 1 h. Rectangular samples (20 mm × 5 mm) were prepared from the dry and wet mTG sponges prior to testing, and the sample thickness was measured by a vernier caliper. A small tare load of 0. 01 N was applied to ensure that each sample received the same initial tension. Samples were tested up to failure at a crosshead speed of 10 mm/min. Data on stress versus strain were calculated from the measurements. The linear slope of the curve was calculated to obtain the elastic modulus. The ultimate tensile stress and elongation at break were also determined. Water absorption To measure the water absorption capacity of mTG sponges, a pre-weighed freeze-dried sponge ( W 0 ) was immersed into deionized water at room temperature for 1 h. The water-absorbing sponge was taken out. Excess water on its surface was carefully removed using filter paper, then weighed again. The weight ( W 1 ) was then noted. The water absorption ratio (ε 2 ) was calculated as follows: (2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}{\varepsilon }_{2}=({W}_{1}-{W}_{0})/({W}_{0})\times 100\text{%}\end{eqnarray*}\end{document} ε 2 = W 1 − W 0 ∕ W 0 × 100 % Material degradation Hydrolysis test The hydrolysis performance of the mTG sponges was assessed by immersing the materials in deionized water for a certain period. Their degradation rates were also analyzed. Pre-weighed freeze-dried mTG sponges were sterilized in 75% ethanol for 20 min, followed by 30 min UV irradiation and several washings with deionized water. The sterilized sponges were kept in water and placed in a cell incubator at 37 °C. At the specified time points (i. e. , one, two, and three months), the remaining sponges were obtained, freeze dried, and weighed again. The percentage of remaining weight divided by original weight is the material hydrolysis rate. Enzymolysis test: Pre-weighed wet sponges were exposed to 0. 1% collagenase type I (>125 CDU/mg; Invitrogen, Carlsbad, CA, USA) at 37 °C for 6 h in a horizontal shaker. The remaining sponges were obtained and weighed at the specified time points. The extent of enzymatic degradation was determined by calculating the percentage of the remaining weight versus the original weight. Sponges immersed in PBS served as negative control. Subcutaneous implantation Animal study was approved by the Institutional Animal Care and Use Committee (IACUC) of Sichuan University, all experiments were performed in accordance with the guidelines of IACUC of Sichuan University. In vivo biocompatibility of the mTG sponges was assessed in subcutaneous implantation of Sprague–Dawley (SD) rats (age 7 −8 weeks) as described in our previous publication ( Yang et al. , 2016 ). The mTG sponges were sterilized in 75% ethanol, followed by UV irradiation and several washings with PBS, and equilibrated in DMEM for 12 h. The sponges were then cut into a size of 7 mm × 7 mm × 4 mm using a scalpel in a sterile bio-safety cabinet and surgically placed within subcutaneous pockets located on the dorsum of SD rats. Each of the four rats in the experimental group received two dorsal subcutaneous implants. Meanwhile, two rats in the control group were subjected to surgery, but no sponge was implanted. The rats were sacrificed at the designated time points (one and three months). The implanted sponges were resected from the underlying muscle, and their sizes were measured using a Vernier caliper. For material degradation analysis, the harvested implant was cut to expose its cross section for imaging. Images were analyzed using Image Pro Plus software (Media Cybernetics, Rockville, MD, USA). The average thickness of the coating tissue on the sponges was measured, and the volume of the coating tissue was estimated. The volume of the remaining sponges was calculated by subtracting the volume of the coating tissue from the full volume of the implants. The extent of in vivo biomaterial degradation was determined by calculating the percentage of the remaining volume versus the original volume of the sponges. Cell proliferation assay The isolation method of primary ADSCs was described in our previous publications ( Li et al. , 2015 ; Yang et al. , 2016 ). mTG sponges were prepared as mentioned in the previous section. All of the sponges (7 mm × 7 mm × 4 mm) were aspirated dry before cell seeding. The sponges were then placed in each well of a 24-well plate. Each gelatin sponge was seeded with 100 μL of ADSC suspension solution (∼5. 0 ×10 4 cells). After 1 h of incubation at 37 °C with 5% CO 2, the culture medium was supplemented to each well. The culture was conducted for a month and the medium was replaced twice a week. Cell growth was observed daily using an inverted microscope (CKX41; Olympus, Tokyo, Japan). At each time point, the detected samples were refreshed with culture medium supplemented with 10% alamarBlue solution (Yeasen, Shangha, China). After 3 h of incubation, the incubation solutions were transferred to a 96-well plate, and fluorescence was measured with a plate reader using excitation/emission wavelengths of 530/590 nm. ADSCs seeded on culture flask (25 cm 2 ) served as negative control. LIVE/DEAD cell assay The survival status of ADSCs in mTG sponges was detected by LIVE/DEAD staining. Cell/sponge constructs were prepared (see section ‘Cell Proliferation Assay’) and cultured for a month. The constructs were washed in PBS and cut at 1-mm intervals along the transverse planes of the sponge. Representative slices from the middle of the sponge were incubated at 37 °C for 30 min in a solution containing 4 μM calcein-AM (Sigma, St. Louis, MO, USA) and 4 μM propidium iodide (Sigma, St. Louis, MO, USA). After incubation, samples were washed and observed at an inverted fluorescent microscope (XDS30; Sunny Optical Technology, Yuyao, Zhejiang, China) equipped with a digital camera. SEM observation To study cell adhesion on scaffold, SEM images were taken from the slices of cell/sponge constructs (see previous section). The samples were washed with PBS and fixed in 2. 5% glutaraldehyde for 2 h at room temperature. Scaffolds were then dehydrated with gradient ethanol aqueous solutions (10%–100%) and dried overnight under vacuum. Before observation, all of the specimens were coated with gold and observed under a scanning electron microscope (Hitachi S3400+EDX; Hitachi, Tokyo, Japan) at an accelerating voltage of 20 kV. Statistical analysis Data are presented as mean ± SD. Statistical analyses were performed using SPSS software (version 14. 0). Statistical significance was evaluated using one-way ANOVA with least significant difference test. The level of statistical significance was set at P < 0. 05. Results Material appearance and physical characteristics The un-crosslinked gelatin hydrogel is colorless. It can maintain a gel state at 25 °C but melt at 37 °C. Figures 1A and 1B show the morphological change of the hydrogel. After crosslinking by mTG, the gelatin hydrogel showed a milky white color. It can maintain a gel state either at 25 °C or at 37 °C ( Figs. 1E and 1F, respectively). Both the crosslinked and un-crosslinked hydrogels can be freeze dried into white porous sponges ( Figs. 1C and 1G, respectively). After freeze drying, the stiffness of the gelatin material increased significantly. However, the un-crosslinked gelatin sponge is considerably brittle. Once the sponge is exposed to water, its material structure begins to collapse ( Fig. 1D ). Therefore, it cannot be used as a scaffold material. The mTG sponge can maintain its porous scaffold structure even under 37 °C warm water. However, it is gradually restored to the porous hydrogel after water absorption ( Fig. 1H ). 10. 7717/peerj. 3665/fig-1 Figure 1 Appearance of gelatin gel and sponge. The morphology of un-crosslinked gelatin gel on a thermostat aluminum stage at 25 °C (A) or at 37 °C (B); the un-crosslinked gelatin sponge at 25 °C (C) or at 37 °C (D); the mTG crosslinked gelatin gel at 25 °C (E) or at 37 °C (F); and the mTG crosslinked gelatin sponge at 25 °C (G) or at 37 °C (H). One of the major advantages of sponge scaffold is its high porosity. The high porosity contributes to cell growth and promotes the exchange of nutrients and metabolites. The porosity of mTG sponge material was 53. 51 ± 3. 45%, and the pore size of the sponge is small and even. Moreover, as gelatin sponge has high hydrophilicity, it can absorb water more than its own weight. Thus, it is often used as a hemostatic sponge. The water absorption rate of the mTG sponge is 1, 209. 3 ± 57. 8%. Notably, although the weight of the sponge increased by more than 10 times after water absorption, the sponge decreased in volume. On the one hand, the surface tension of water molecules in the sponge is strong, leading to a certain degree of shrinkage. On the other hand, the material has good hydrophilicity. The elastic modulus is an important index for measuring the strength of a material. In the uniaxial tensile test, the gelatin sponge scaffold has a relatively high elastic modulus in dry state, which is 133. 4 ± 7. 9 kPa. It can bear a maximum tensile force of 67. 9 ± 8. 1 kPa, and its elongation at break is 57. 1 ± 10. 0%. When the material absorbs water, its elastic modulus decreases to 59. 9 ± 3. 0 kPa ( P < 0. 01, when compared to that of the dry sponges). The maximum compressive capacity was 110. 8 ± 5. 4 kPa ( P < 0. 01), and the elongation at break was 177. 2 ± 6. 2% ( P < 0. 01). Representative tensile curves of dry and wet gelatin sponges were shown in Fig. 2. 10. 7717/peerj. 3665/fig-2 Figure 2 Uniaxial mechanical testing of mTG sponge. (A) Tensile curve of dry gelatin sponge and (B) tensile curve of wet gelatin sponge. Hydrolysis and enzymolysis The mTG sponge is highly resistant to hydrolysis. After three months of soaking in water at 37 °C, its mass loss became less than 5% ( Fig. 3A ). However, the sponge is sensitive to collagenase. Under the catalysis of 0. 1% collagenase type I, the quality of the mTG sponge is decreased exponentially, and it can be completely dissolved after only about 6 h, whereas no obvious mass loss could be seen in control group ( Fig. 3B ). Gelatin is a product of collagen hydrolysis, and collagenase is a digestive enzyme that is specific for collagen and collagen derivatives. Therefore, gelatin sponges can be degraded by collagenase solution. 10. 7717/peerj. 3665/fig-3 Figure 3 In vitro degradation performance of mTG sponges. (A) Hydrolysis curve and (B) enzymolysis curve. Analysis of subcutaneous implantation The mTG sponges were implanted subcutaneously in rats to evaluate the degradation of the material in vivo and observe whether they would cause inflammatory responses. The results showed that all animals in the experimental and control groups had no inflammatory reaction after operation. Figure 4A shows the morphology of the material in the subcutaneous tissue after one month of implantation. The implant is located between the skin and the muscle layer, the surface of which is wrapped by a connective tissue. No inflammatory reaction nor pus was observed. After three months of implantation, the mTG sponges were completely degraded and absorbed without causing an inflammatory response, as shown in Fig. 4B. 10. 7717/peerj. 3665/fig-4 Figure 4 Photographs of gelatin sponges implanted subcutaneously in rats. The implanted gelatin sponges with attached skin were removed from the experimental rats after one month of subcutaneous implantation (A) or after three months of subcutaneous implantation (B); the one-month-old implant without attached skin (C); and cross section of the one-month-old implant was observed under a stereomicroscope (D). To assess the tissue rejection in the subcutaneous implantation, the one-month-old implant was removed and analyzed, as shown in Fig. 4C. The implant is like a flesh-color ellipsoid. The implant material is cuboid, and then becomes ellipsoid, indicating that after implantation, the material has shown a certain degree of degradation at least on its edge. Subsequently, the implant was centrally cut to observe and measure the thickness of the coating layer and calculate the remaining volume of the sponge. After one month of implantation, the average residual volume of mTG sponges was 27. 73 ± 2. 68%, and the average thickness of their coating tissue was 0. 19 ± 0. 16 mm. In addition, tiny capillaries could be seen on the surface of some implants under observation of a stereomicroscope ( Fig. 4D ). Accordingly, the material can promote angiogenesis. In the images of H&E staining, we observed an abundance of cells and fibrous tissue in the coating tissue of the implant, including a few of the megakaryocytes, as shown in Fig. 5A. In the central slices of the implant, only the gelatin sponge material can be seen, whereas no cells were observed ( Fig. 5B ). This condition can be explained by the fact that the implantation time is short; the cells have not yet penetrated into the interior of the material. 10. 7717/peerj. 3665/fig-5 Figure 5 Histological observations of the gelatin sponges after one month of subcutaneous implantation in rats. (A) H&E staining of marginal slice of the implant and (B) H&E staining of the central slice of the implant. Observation cell growth via inverted optical microscope The cell compatibility of gelatin sponge was detected by cell/scaffold co-culture. The ADSCs were seeded into the scaffold. After one month of culture, the cell growth and morphology were observed using an inverted optical microscope. Observation via microscope clearly shows that the gelatin sponge contains numerous fiber connections, which form the interconnected holes and mesh structures. The interconnected meshes facilitate cell growth and the delivery of nutrients, and many cells are evenly distributed at all levels of the material, as shown in Fig. 6. The cells are tightly attached to the gelatin fibers and grow up into many cell aggregates, which vary in size and shape. In addition, some of the cells grow laterally, covering many meshes and forming thin cell sheets. 10. 7717/peerj. 3665/fig-6 Figure 6 Cell growth in mTG sponges observed under optical microscope. (A) The cell morphology and distribution were observed under a microscope field of 4×, and (B) the adhesion of the cells on the scaffold was observed under a microscope field of 10×. Evaluation of cell proliferation AlamarBlue assay was used to investigate the effect of gelatin sponge on cell proliferation. From 2D to 3D culture, a period for cell adaptation was allowed to cope with the environmental changes. During this period, the cellular behaviors were mainly adhesion and migration, whereas the cell proliferation was not evident. The number of cells in mTG sponges began to increase on the third day after inoculation. On Day 30, the number of cells was approximately 50 times that of the original number. Therefore, the material can support cell adhesion and proliferation. The cell growth curve is shown in Fig. 7. The mTG sponge has many spaces for cell accommodation. Thus, we speculate that with the extension of culture time, cell growth will continue. In control groups, the cells apparently showed a higher proliferation rate compared with cells cultured in sponges from day 0 to day 12. However, the cell number on culture flasks almost linearly decreased after 12 days of culture, as the cells stopped growth caused by cell contact inhibition, and some cells were detached from the flask surface. Analysis of LIVE/DEAD cell fluorescence staining The survival and distribution of the cells on the mTG sponge can be observed directly by LIVE/DEAD cell fluorescence staining. Figures 8A – 8C show that the cells are evenly distributed in the sponge scaffold. After one month of culture, the cells formed numerous cell aggregates in the material, which is consistent with the observation under the light microscope. Figure 8A shows that living cells are numerous and distributed throughout the material. The number of apoptotic cells is small. Only a few scattered red dots can be seen in the picture ( Fig. 8B ). Next, a high-powered microscope was used to observe the viability of the cell aggregates. We found that some of the cell aggregates showed strong green fluorescence, indicating an accumulation of many living cells ( Fig. 8D ). Within the aggregate, a small number of cells were dead ( Fig. 8E ). This condition may be due to the high cellular density in the center of the aggregate, which hinders the metabolism and leads to cell apoptosis. The results of LIVE/DEAD staining further confirmed that the mTG sponge had good cell compatibility. 10. 7717/peerj. 3665/fig-7 Figure 7 Evaluation of cell proliferation in mTG sponges and on culture flasks. 10. 7717/peerj. 3665/fig-8 Figure 8 LIVE/DEAD cell fluorescence staining on mTG sponges. (A and D) Live cells are stained with calcein-AM (green), (B and E) dead cells are stained with propidium iodide (red), and (C and F) merged fluorescence images of cellular LIVE/DEAD staining. Magnification of the microscope is (A–C) 4× and (D–F) 20×. Observation cell adhesion by SEM The cell adhesion and morphology on the mTG sponge was observed by SEM. Figure 9 shows that the scaffold material presents a lamellar structure that forms large pores. The approximative pore size based on the SEM images of mTG sponge is 100 μm. We estimate that each large pore can accommodate dozens of cells. Some of the cells on the material surface grow together to form a cell aggregate. Other cells are tightly attached to the surface of the material. In addition, several cells show a small ball shape. Thus, we speculate that these are the cells undergoing cell division. 10. 7717/peerj. 3665/fig-9 Figure 9 SEM observation of cell adhesion on mTG sponge. Discussion mTG is a promising material crosslinking agent. First, the enzyme-mediated crosslinking reaction can be achieved under physiological conditions, which is particularly suitable for the formation of in situ injectable hydrogels. Second, the residual enzyme in the crosslinking reaction will be degraded in vivo, and the harm to the cell is substantially less than that of the chemical crosslinking agent. Actually, mammalian transglutaminases (TG2) are found extensively in the extracellular matrix of animal tissues; they act as a matrix stabiliser through protein crosslinking activity and as an important cell adhesion protein involved in cell survival ( Collighan & Griffin, 2009 ). As a homologous enzyme of TG2, mTG has good biocompatibility. Small amounts of residual mTG in hydrogels exhibit no cytotoxicity ( Schloegl et al. , 2012 ). Third, mTG can selectively recognize specific amino acid sequences of gelatin materials and form a stable interconnection between protein molecules. Consequently, the physical and chemical properties of the crosslinked materials were improved. In the preparation and application of gelatin sponge scaffolds, the stiffness and strength of materials, porosity and permeability, hydrophilicity, biocompatibility, as well as biodegradability and vascularization should be considered. Stiffness and strength are important parameters of biomaterials. The stiffness of the material significantly influences cell growth and differentiation ( Xia, Liu & Yang, 2017 ). The appropriate material stiffness can provide good mechanical support for cells and tissues. However, the high stiffness will make the material brittle and easily breakable. The ideal biomaterial is designed according to the stiffness of the replaced tissue. For example, scaffold materials for bone tissue engineering should have a high stiffness, and the stiffness of the materials used for skin repair should be relatively small. In the dry state, the gelatin sponge scaffold has high stiffness. However, when the material is immersed in aqueous solution, its stiffness decreases significantly. Therefore, the pure gelatin sponge scaffold is more suitable for soft tissue repair. In this case, the main concern for the material is its strength. The strength of sponge scaffolds can be characterized by measuring the elastic modulus and yield stress. Moreover, the strength of the material can be adjusted by changing the manufacturing process, crosslinking method, and polymer concentration. Hence, the strength of gelatin sponge can be adjusted by changing the concentration of gelatin solution or the amount of crosslinking agent. The pore size and porosity of materials affect the cellular migration, growth, and proliferation ( Wang et al. , 2016 ). In sponge scaffolds, the interconnected porous structure contributes to the transport of nutrients and metabolites, leading to faster cell growth and proliferation. In particular, the large pore size is beneficial to the ingrowth of cells on the material. The pore size and porosity of sponge are related to the physicochemical properties of material composition and the manufacturing process. Freeze drying is the most common method for preparing gelatin sponge. The pore size and porosity of gelatin sponge can be adjusted by changing the gel concentration, freezing temperature, and freezing time. The gelatin sponge usually has high porosity, which makes it suitable for use as a tissue engineering scaffold. However, the pore size of the sponge should be determined according to the actual requirement. At present, the optimal pore size is yet to be determined. Some studies have suggested that scaffold material with a small pore size is optimal because it is closer to the nanostructure of natural tissue. Other studies suggested that the large pore structure may be more beneficial for cell proliferation and ECM secretion ( Griffon et al. , 2006 ; Nava et al. , 2016 ). In addition, studies have shown that cells in material with small pore size will undergo dedifferentiation, whereas those in the large pore size stop proliferating and secrete more ECM instead ( Lien, Ko & Huang, 2009 ). The gelatin sponge has high hydrophilicity. It can absorb water weighing more than 10 times of its own. Thus, the gelatin sponge is often used as hemostatic material in clinics. Although the highly hydrophilic scaffold material is beneficial for cell growth and proliferation, the high water absorption causes the scaffold to expand excessively and accelerates its degradation. As the high concentration of gelatin can form more compact hydrogel, changing the gelatin concentration can adjust the water absorption rate of the mTG sponge to stabilize its mechanical structure. Biocompatibility refers to the ability of scaffolds to support cell growth and proliferation without causing any toxic or immune response. It is also defined as the interaction between biomaterials and adjacent natural tissues. The ideal scaffold material does not cause rejection after implantation in vivo. In the current study, we implanted mTG sponges into the subcutaneous layer of rats. One month after implantation, no inflammation was observed at the implant site. However, the mTG sponge was wrapped by a thin layer of connective tissue, indicating that the material has some minor tissue rejection in the early stage of implantation. Along with the prolongation of implantation time, the connective tissues and the implanted materials will be absorbed by the host. Consequently, the coating layer was not observed, and the materials were absorbed after three months of implantation. In the field of regenerative medicine, the scaffold material usually needs to be biodegradable. Material degradation occurs in some biological processes, such as enzymolysis or hydrolysis, resulting in the destruction of the material structure or the loss of material functional groups. If the material degradation rate is properly designed, the newborn tissue will fill into the room left by the degraded material. This condition will not cause the collapse of the material structure. Therefore, the degradation rate of an ideal material needs to be equal to or comparable to the regeneration rate of the newborn tissue. However, some exceptions exist, such as regenerative bone tissue, articular cartilage, or cornea, which do not require complete degradation of the implanted material. For these applications, the most important requirement is that the implant material should be firmly combined with the surrounding tissue ( Slaughter et al. , 2009 ). For biodegradable materials, another point must be considered, that is, the final condition of the material degradation products. Ideally, the degradation products need to be absorbed by the host or excreted by metabolism. If this case is not possible, the degradation products should at least be non-cytotoxic or not trigger the immune response. Angiogenesis is an important criterion for evaluating the performance of scaffold materials. Almost all of the tissues in the body, except the cartilage and cornea, require blood supply for growth and development. Therefore, biomaterials designed by mimicking the natural tissue should have the ability to promote angiogenesis. For the porous scaffolds implanted in the body, the degree of vascularization is related to the pore size, porosity, degradation rate of the material, as well as the secretion factors of the host ( Walthers et al. , 2014 ). To improve the degree of vascularization of scaffold materials, growth factors, such as vascular endothelial growth factor, fibroblast growth factor, or drugs, should be added to promote angiogenesis ( Martino et al. , 2015 ). When the material is degraded, the factors or drugs embedded in the material will be released. In the mTG sponges of subcutaneous implantation, some capillaries form on the coating tissue. Results show that these sponges are beneficial to angiogenesis. In addition, ADSCs have the potential to differentiate into adipocytes, osteoblasts, chondrocytes, cardiomyocytes, vascular cells, and endothelial cells. The combination of gelatin sponge scaffolds with ADSCs may further promote the vascularization of sponge scaffolds by cell autocrine or paracrine effects. Moreover, we speculate that inflammatory response following the biomaterial implantation may also contribute to the angiogenesis. Certainly, more studies are needed to uncover the underlying mechanism of angiogenesis. Taken together, we believe that gelatin sponges encapsulated with cells or drugs will have great potential for tissue regeneration. Supplemental Information 10. 7717/peerj. 3665/supp-1 Data S1 Raw data Experimental data Click here for additional data file.

10. 7717/peerj. 3927

PeerJ

Augmented cartilage regeneration by implantation of cellular versus acellular implants after bone marrow stimulation: a systematic review and meta-analysis of animal studies

Bone marrow stimulation may be applied to regenerate focal cartilage defects, but generally results in transient clinical improvement and formation of fibrocartilage rather than hyaline cartilage. Tissue engineering and regenerative medicine strive to develop new solutions to regenerate hyaline cartilage tissue. This systematic review and meta-analysis provides a comprehensive overview of current literature and assesses the efficacy of articular cartilage regeneration by implantation of cell-laden versus cell-free biomaterials in the knee and ankle joint in animals after bone marrow stimulation. PubMed and EMBASE (via OvidSP) were systematically searched using tissue engineering, cartilage and animals search strategies. Included were primary studies in which cellular and acellular biomaterials were implanted after applying bone marrow stimulation in the knee or ankle joint in healthy animals. Study characteristics were tabulated and outcome data were collected for meta-analysis for studies applying semi-quantitative histology as outcome measure (117 studies). Cartilage regeneration was expressed on an absolute 0–100% scale and random effects meta-analyses were performed. Implantation of cellular biomaterials significantly improved cartilage regeneration by 18. 6% compared to acellular biomaterials. No significant differences were found between biomaterials loaded with stem cells and those loaded with somatic cells. Culture conditions of cells did not affect cartilage regeneration. Cartilage formation was reduced with adipose-derived stem cells compared to other cell types, but still improved compared to acellular scaffolds. Assessment of the risk of bias was impaired due to incomplete reporting for most studies. Implantation of cellular biomaterials improves cartilage regeneration compared to acellular biomaterials.

Introduction Articular cartilage facilitates joint loading and movement by resisting compressive and shear forces ( Swieszkowski et al. , 2007 ). For patients, localized cartilage defects can have detrimental long term effects such as joint dysfunction, pain, and degenerative osteoarthritis. Upon cartilage damage, its avascular nature prevents spontaneous healing ( Buckwalter, Saltzman & Brown, 2004 ). Clinical treatments for full-thickness cartilage defects and osteochondral lesions include bone marrow stimulation techniques, e. g. , microfracture and subchondral drilling, and autologous chondrocyte implantation. Defect size generally determines treatment, where microfracture and autologous chondrocyte implantation are used to treat small (<2. 5 cm 2 ) and large lesions (>2. 5 cm 2 ), respectively ( Cucchiarini et al. , 2014 ). Microfracture surgery is a minimally invasive and inexpensive one-step approach, where multiple perforations, microfractures, are made in the subchondral bone plate to induce bleeding and provoke a reparative response. The formed blood clot consists of bone marrow-derived mesenchymal stem cells (BM-MSCs), growth factors and other proteins, supporting cartilage formation ( Steadman, Rodkey & Rodrigo, 2001 ). The repaired tissue, however, generally consists of fibrous cartilage, which lacks the mechanical properties of native hyaline cartilage ( Dai et al. , 2014 ). Microfracture results in temporary clinical improvement only ( Saris et al. , 2014 ), and the demand for improved cartilage regeneration persists. Cartilage regeneration may be improved by tissue engineering and regenerative medicine (TERM) in addition to bone marrow stimulating techniques. TERM encompasses the development of biomaterials, which can be loaded with cells and biologics ( Seo et al. , 2014 ). Upon implantation and infiltration of BM-MSCs, the biomaterial may act as a template to guide/stimulate cartilage regeneration ( Cucchiarini et al. , 2014 ). In a previous systematic review and meta-analysis on animal models, we showed that acellular biomaterials in addition to bone marrow stimulation was more effective in regenerating cartilage in vivo than bone marrow stimulation alone, which was further improved by use of biologics ( Pot et al. , 2016 ). When biomaterials are loaded with cells, bone marrow stimulation may be even more effective. Biomaterials loaded with cells after bone marrow stimulation has been widely investigated in vivo, and included loading of chondrocytes ( Ahn et al. , 2009 ; Caminal et al. , 2016 ; Christensen et al. , 2012 ), BM-MSCs ( Araki et al. , 2015 ; Igarashi et al. , 2012 ; Wakitani et al. , 1994 ), synovium-derived mesenchymal stem cells (SD-MSCs) ( Pei et al. , 2009 ; Lee et al. , 2013 ; Shimomura et al. , 2014 ), adipose-derived stem cells (ADSCs) ( Xie et al. , 2012 ; Masuoka et al. , 2006 ; Kang et al. , 2014 ), periosteal cells ( Perka et al. , 2000 ; Schagemann et al. , 2009 ), fibroblasts ( Yan & Yu, 2007 ), umbilical cord stem cells (UCSC) ( Yan & Yu, 2007 ; Chung et al. , 2014 ) and embryonic stem cells (ESC) ( Cheng et al. , 2014 ). Cells are either used directly after harvesting ( Betsch et al. , 2014 ; Getgood et al. , 2012 ) or after an additional in vitro step of cell expansion ( Guo et al. , 2010 ; Dorotka et al. , 2005 ) and/or differentiation ( Sosio et al. , 2015 ; Necas et al. , 2010 ). In this systematic review and meta-analysis, we present a comprehensive overview of all current literature regarding regeneration of articular cartilage by implantation of cell-laden versus cell-free biomaterials in the knee and ankle joint after bone marrow stimulation in animal models ( Fig. 1 ). We further investigated the effect of loading biomaterials with (1) stem cells versus somatic (differentiated) cells, (2) different cell types (e. g. , chondrocytes, MSCs, ADSCs), and (3) culture conditions of cells (e. g. , use after harvesting, in vitro expansion and/or differentiation). In the meta-analysis, histological scores from semi-quantitative histological scoring systems were used to assess the effect on cartilage regeneration. 10. 7717/peerj. 3927/fig-1 Figure 1 Illustration of articular cartilage regeneration by implantation of cellular and acellular biomaterials after applying bone marrow stimulation. The figure was adapted from Pot et al. (2016). Materials and Methods Search strategy An extensive literature search was performed in PubMed and EMBASE (via OvidSP) to identify relevant peer-reviewed articles until June 29, 2016, using methods defined by De Vries et al. (2012) and Leenaars et al. (2012). The search strategy ( Supplemental Information 1 ) consisted of search components for tissue engineering ( Sloff et al. , 2014 ) and cartilage ( Pot et al. , 2016 ). Results were refined for animal studies by applying animal search filters ( Hooijmans et al. , 2010 ; De Vries et al. , 2011 ). No language restrictions were applied. Study selection After obtaining all references, duplicates were manually removed in EndNote X7 (Thomson Reuters, Philadelphia, PA, USA) by one author (MP). Resulting references were screened for relevance by two independent authors (MP and VG/WD) based on title, title/abstract and full-text using Early Review Organizing Software (EROS, Institute of Clinical Effectiveness and Health Policy, Buenos Aires, Argentina, http://www. eros-systematic-review. org ). In case of disagreement between authors or any doubt, references were included for further screening. An overview of all exclusion criteria per screening phase is provided in Supplemental Information 2. Studies were included for risk of bias assessment and meta-analysis when semi-quantitative histological scoring was used as outcome measure. Study characteristics Study characteristics were extracted from the studies by MP. Basic information (author, year of publication), animal model characteristics (species, strain, sex, etc. ), experimental characteristics (surgery, biomaterial, follow-up, etc. ), cell characteristics (cell type, culture conditions, etc. ) and outcome characteristics (macroscopic evaluation, histology and semi-quantitative histological scoring, etc. ) were obtained. Risk of bias assessment The methodological quality was assessed for studies included in the meta-analysis. A risk of bias analysis was performed according to an adapted version ( Pot et al. , 2016 ) of the tool described by Hooijmans et al. (2014). Selection, performance, detection and attrition bias were scored independently by MP and VG/WD using questions and a flowchart ( Pot et al. , 2016 ), where ‘-’, ‘?’ and ‘+’, indicating low, unknown and high risk of bias. In case of differences between authors, results were discussed until consensus was reached. Unfortunately, 16 articles were published in Chinese and we did not have the resources to obtain certified translations of these articles. We were, however, able to successfully extract the data of these studies using Google Translate ( https://translate. google. com/ ) and used the data in the meta-analysis. A sensitivity analysis was performed to evaluate the effect of language (exclusion of Chinese articles, see ‘Meta-analysis’). Analysis preparations and meta-analysis Analysis preparations Meta-analyses were performed for outcome measure semi-quantitative histology; data were used from studies that compared biomaterials with (experimental group) and without cells (control group). In general, these histological scoring systems and their components, extensively reviewed by Rutgers et al. (2010), evaluate the degree of cartilage regeneration by scoring parameters like Safranin-O staining (which stains negatively charged glycosaminoglycans, an important component of cartilage tissue), surface integrity and cartilage thickness. Outcome data (mean, standard deviation (SD) and number of animals) were extracted from the studies for all time points as follows: (1) numerical data from the text/tables, (2) graphical results by measuring the mean and SD using ImageJ (1. 46r, National Institutes of Health USA), (3), boxplot results by recalculating from median, range and sample size to mean and SD ( Hozo, Djulbegovic & Hozo, 2005 ), and (4) for results presented as mean and confidence interval (CI) per group, the following equation was used to recalculate CI to a standard deviation: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$SD=\sqrt{N}\times \frac{upper~limit-lower~limit}{3. 92} $\end{document} S D = N × u p p e r l i m i t − l o w e r l i m i t 3. 92 for a 95% CI ( Higgins & Green, 2011 ). When data were missing or unclear, authors were contacted to provide data. Studies were excluded from meta-analysis in case data could not be retrieved or remained unclear (e. g. , missing SD, all SD’s similar to corresponding mean, and histological scores exceeding maximum), unless data were sufficiently clear to make assumptions (i. e. , group size and number of animals per time point and analyses, see Supplemental Information 3 ). A sensitivity analysis was performed to evaluate the effect of assumptions (exclusion of articles with assumptions, see ‘Meta-analysis’). Histological scoring systems describe the degree of cartilage regeneration with different scoring scales. To compare data from different studies, all data were converted to a 100% cartilage regeneration scale by dividing both the mean and SD by the maximum score of the scoring system and multiplying the outcome by 100%. In this systematic review, healthy tissue is represented as 100% cartilage regeneration. Lower percentages indicate less regenerated cartilage. When results of experimental groups could be combined per study (i. e. , outcome of various biomaterials seeded with one cell type), we did so, followed the approach described in the Cochrane Handbook, table 7. 7 ( Higgins & Green, 2011 ), which means that we calculated a weighted average of the results with an appropriate standard deviation. Time points of treatment groups were combined using the same approach. The mean and corresponding standard error (SE) per treatment group were subsequently calculated per study. Meta-analysis The main research question was: Is there an overall beneficial effect on cartilage regeneration of implanting biomaterials loaded with cells compared to acellular biomaterials? We used a bivariate approach to model a random effects meta-analysis, i. e. , separate outcomes for the control and experimental group were used with their respective SEs. The correlation between these two outcomes was modeled with a compound symmetry covariance matrix, as this resulted in a lower Akaike Information Criterion value than an unstructured covariance matrix. To evaluate the effect of specific variables on treatment outcome for the experimental group (biomaterials loaded with cells), the following sub-questions were addressed: (1) Is there a difference between the use of stem cells and somatic (differentiated) cells (stem cells vs. somatic cells); (2) Do differences among various cell subgroups exist (e. g. , chondrocytes vs. other cells); (3) Is there a difference between biomaterials loaded with cells which were not cultured in vitro, were expanded in vitro or were differentiated in vitro (during surgery vs. expansion, surgery vs. differentiation, and expansion vs. differentiation)? Results are depicted as % cartilage regeneration (95% CI: [lower CI, upper CI]. The mean difference (% [95% CI]) is presented as condition A–condition B. Based on a previous study, data of all time points were used ( Pot et al. , 2016 ). Subgroup analyses were performed in case subgroups consisted of more than five experimental groups in at least three studies. Most studies contained more than one experimental group, therefore the total number of studies and number of experimental groups (no. of studies/groups) is provided in the analysis. No adjustment for multiple testing was applied in analyses of sub-questions. Sensitivity analyses were performed on the main research question to evaluate the effect of language (excluding Chinese articles, as the risk of bias for these articles was not investigated), and the effect of assumptions (excluding articles for which assumptions were made) in the meta-analysis. SAS/STAT ® software version 9. 2 for Windows, copyright© 2002–2008 by SAS Institute Inc. , Cary, NC, USA, was used to perform statistical analyses. R software version 3. 0. 1 ( R Core Team, 2011 ) with package meta ( Schwarzer, 2015 ) was used to create the funnel plot, which illustrates effect sizes of all studies versus their precision, and test for the asymmetry, using the method of moments estimator for the between study variation ( Thompson & Sharp, 1999 ). I 2 was used as a measure of heterogeneity. I 2 measures the percentage of variability in treatment effect estimates that is due to between study heterogeneity rather than chance ( Higgins et al. , 2003 ). If I 2 is 0%, this suggests that the variability in the study estimates is entirely due to chance. If I 2 is >0% there might be other reasons for variability. ReviewManager ( The Cochrane Collaboration, 2014 ) was used to create the forest plot. Results Search and study inclusion Searching PubMed and EMBASE databases for references regarding cartilage regeneration by implantation of cellular and acellular biomaterials in the knee and ankle joint in combination with bone marrow stimulation resulted in a total of 11, 248 references (Pubmed 4, 743, Embase 6, 505). Removal of duplicates left 7, 354 references. Screening by title and title/abstract resulted in exclusion of 6, 744 references. Full-text of 610 studies resulted in 146 included studies. The full-text of some studies ( Xie et al. , 2014 ; Yao, Ma & Zhang, 2000 ; Zhou & Yu, 2014 ) could not be retrieved and these were excluded. In the meta-analysis, studies were used which applied semi-quantitative histology as outcome measure, resulting in 117 included studies. A risk of bias assessment ( Fig. 2 ) was performed for 101 of 117 studies (excluding Chinese studies). Supplemental Information 3 provides an overview of all included studies after full-text screening, risk of bias assessment and meta-analysis, as well as detailed information regarding reasons for exclusion and assumptions made for certain studies. Supplemental Information 4 contains the reference list and abbreviations of Supplemental Information 3 studies. 10. 7717/peerj. 3927/fig-2 Figure 2 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) flowchart of the systematic search of literature. Of the 117 studies included for the meta-analysis, a risk of bias assessment was performed for 101 studies, excluding Chinese articles. Study characteristics A large variation between studies was observed regarding animal model characteristics (species, strain, sex, etc. ), experimental characteristics (surgery, biomaterial, follow-up, etc. ), cell characteristics (cell type, culture conditions, etc. ) and outcome characteristics (macroscopic evaluation, histology and semi-quantitative histological scoring, etc. ), as can be appreciated from Supplemental Information 3. Various animal species were used including rabbit, dog, sheep, pig, rat, horse, minipig, goat and macaques. A large range was found in animal age, e. g. , the age of rabbits ranged from six weeks to >2 years. Small animals were generally younger (in the range of months) compared to larger animals (in the range of years). In many studies, no detailed information was provided regarding the animal’s absolute age, but merely e. g. , adult or mature. The method for bone marrow stimulation was mostly subchondral drilling (142 studies), where only four studies used microfracture. Defects were created at various locations (trochlea, condyles, femur and intercondylar fossa) and with diverse dimensions (e. g. , for rabbits: diameter 4–7 mm and depth 0. 8–9 mm). Implanted biomaterials were prepared from natural (e. g. , alginate and collagen), synthetic (e. g. , poly(lactic-coglycolic acid) and polycaprolactone) or mixtures of natural and synthetic materials. In 27 studies biologics, such as bone morphogenetic protein 2 and transforming growth factor beta, were loaded in the biomaterials. Different cell types were applied, including chondrocytes, bone marrow-derived mesenchymal stem cells (BM-MSCs), bone marrow-derived progenitor cells, synovium-derived stem cells (SD-MSCs), bone marrow-derived mononuclear cells, adipose-derived stem cells, adipose-derived stromal vascular fraction cells, endothelial progenitor cells, embryonic stem cells, umbilical cord blood stem cells, fibroblasts, and periosteal cells, while in some studies undefined cell populations like bone marrow aspirate concentrate were used. Cells were either seeded directly after harvesting on biomaterials and implanted in the created defect or cultured in vitro to expand and/or differentiate the cells, followed by seeding on biomaterials and implantation. In vitro differentiation was performed with cells cultured in monolayer (without biomaterials), followed by seeding of the cells onto the biomaterials and implantations, or by directly culturing the cells on biomaterials prior to implantation. In most studies, short-term cartilage regeneration was investigated: the follow-up time was generally less than 6 months with a maximum follow-up of 12 months. Risk of bias assessment The methodological quality was assessed for all studies included in the meta-analysis except Chinese articles. The overview of the results in Fig. 3 indicates a general lack of information regarding the experimental setup of the studies, limiting the assessment of the actual risk of bias. Please see Supplemental Information 5 for all scores per individual study. 10. 7717/peerj. 3927/fig-3 Figure 3 Results of the risk of bias analysis. Low, unknown or high risk of bias are presented in green, orange and red, respectively, where the percentages indicate the percentage of studies scoring low, unknown or high risk of bias of the total number of investigated studies per question. Low risk of bias was mainly found for addressing incomplete outcome data and baseline characteristics at the moment of surgical intervention. Unknown risk of bias was generally the result of limited details described in the studies regarding the experimental set-up. High risk of bias was only occasionally scored. Questions 4–6 are not depicted graphically, but are described and explained in Supplemental Information 4. In the assessed studies, details regarding the application and method of randomization (Q1) were generally lacking. As a result, assessment of the actual risk of selection bias was practically impossible. Assessment of the actual risk of bias due to differences in baseline characteristics was difficult since no details regarding randomization were described. Differences may have been present in load-bearing between implantation sites (Q2. 1) and age, sex and weight of animals (Q2. 2). In most studies, few differences were found between animals at the moment of surgical intervention since animals were treated similarly (Q2. 3). Details regarding blinding of experimental conditions at the moment of implantation were generally not provided, which may have resulted in bias (Q3). Random housing of animals was generally not (well) described (Q4). Caregivers and/or investigators did not know which intervention each animal received during the experiment (Q5). No details were presented regarding the random selection of animals for outcome assessment (Q6). The method of blinding during analysis, however, was well described in most studies (Q7). Incomplete outcome data were identified or described in a few studies only, which resulted in studies with high risk of bias (Q8). Generally, most studies lacked reporting of important details and therefore adequate assessment of the actual risk of bias was difficult. Data synthesis Semi-quantitative histological scores were used as outcome data to compare biomaterials with cells (experimental group) and without cells (control group) and to address sub-questions related to the use of type of cells and culture conditions. An overview of all meta-analysis results is provided in Table 1 ; an overview of all raw data is given in Supplemental Information 6. 10. 7717/peerj. 3927/table-1 Table 1 Overview meta-analysis results; the effect on cartilage regeneration of (1) the addition of cells to biomaterials, (2) loading of stem cells vs. somatic cells, (3) loading of specific cell types, e. g. , chondrocytes vs. all cells except chondrocytes, and (4) culture conditions. The total number of studies and number of groups included in the meta-analysis are depicted (studies may have >1 experimental group, no. of studies/groups). Results are presented on a 100% cartilage regeneration scale, where 100% indicates ‘maximum’ cartilage regeneration. The addition of cells to biomaterials significantly improved cartilage regeneration compared to acellular biomaterials. The use of stem cells or somatic cells resulted in comparable cartilage regeneration. Cartilage regeneration was significantly lower for biomaterials seeded with adipose-derived stem cells compared to other cell types. Cartilage regeneration was not affected by the method of cell manipulation. Meta-analysis No. of studies/groups Subgroups Cartilage regeneration (% [95% CI)] Mean difference (% [95% CI]) p -value 1. Overall effect 98/265 Cellular scaffolds 61. 5 [58. 5–64. 5] 18. 6% [15. 2–22. 0] 98/208 Acellular scaffolds 43. 0 [40. 0–46. 0] p < 0. 0001 2. Stem cells or somatic cells 57/148 Stem cells 61. 5 [58. 1–65. 0] −1. 28 [−6. 5–4. 0] 36/101 Somatic cells 62. 8 [58. 5–67. 1] p = 0. 622 3. Type of cells 30/81 Chondrocytes 63. 6 [58. 1–69. 0] 2. 7 [−3. 4–8. 9] p = 0. 373 44/117 Bone marrow-derived MSCs 61. 5 [57. 1–65. 9] −0. 3 [−6. 0–5. 4] p = 0. 919 3/6 Synovium-derived MSCs 7. 4 [36. 7–98. 2] −6. 0 [−8. 5–20. 5] p = 0. 412 11/19 Adipose-derived stem cells 56. 3 [49. 9–62. 6] −5. 9 [−11. 3–−0. 4] p = 0. 036 8/14 Bone marrow aspirate 54. 7 [39. 8–69. 6] −7. 6 [−20. 5–5. 2] p = 0. 239 3/7 Bone marrow-derived mononuclear cells 74. 1 [27. 9–100. 0] 12. 9 [−8. 6–34. 3] p = 0. 238 4. Cell manipulation 14/27 During surgery: harvesting, implantation 58. 9 [51. 3–66. 5] Surgery vs. Expansion −2. 4 [−10. 8–5. 9] p = 0. 564 59/180 Expansion: harvesting, expansion in vitro, implantation 61. 4 [57. 6–65. 1] Surgery vs. Differentiation −4. 2 [−13. 5–5. 1] p = 0. 374 27/58 Differentiation: harvesting, differentiation in vitro, implantation 63. 1 [57. 6–68. 6] Expansion vs. Differentiation −1. 7 [−8. 2–4. 7] p = 0. 594 Data are presented as the effect (%) with 95% CI, where 100% cartilage regeneration represents healthy tissue and lower percentages indicate less regenerated cartilage tissue. Overall effect implantation of cellular and acellular biomaterials The meta-analysis indicates that implantation of cellular and acellular biomaterials resulted in 61. 5% (95% CI [58. 5–64. 5]) and 43. 0% (95% CI [40. 0–46. 0]) cartilage regeneration, respectively. The addition of cells to biomaterials significantly improved cartilage regeneration by 18. 6% (95% CI [15. 2–22. 0], p < 0. 0001). An overview of results for each individual study is displayed in the forest plot ( Supplemental Information 7 ), presenting improved cartilage regeneration by loading biomaterials with cells in 66 studies, similar cartilage regeneration in 30 studies, and a negative effect on cartilage regeneration in two studies. The heterogeneity ( I 2 ) for the comparison between cellular and acellular biomaterials was very high (99. 4% (95% CI [99. 3%–99. 4%])). Stem cells and somatic cells No significant differences ( p = 0. 622) were found between biomaterials loaded with stem cells (61. 5% (95% CI [58. 1–65. 0])) and somatic cells (62. 8% (95% CI [58. 5–67. 1])). Cell type Biomaterials were loaded with various cell types. Subgroup analyses were only performed when subgroups consisted of more than five experimental groups in at least 3 studies. Seeding biomaterials with adipose-derived stem cells significantly decreased cartilage regeneration, while no other significant differences were observed ( Table 1 ). Only for scaffolds seeded with adipose-derived stem cells (ADSCs), reduced cartilage regeneration was found (56. 3% (95% CI [49. 9–62. 6])) compared to cellular scaffolds. However, cartilage regeneration using ADSCs-seeded scaffolds still improved regeneration compared to acellular scaffolds. Cell manipulation Comparing differences in cartilage regeneration between biomaterials loaded with cells which were not cultured in vitro (implanted immediately after harvesting of cells) or were expanded and/or differentiated in vitro indicated that cell manipulation did not affect cartilage regeneration ( Table 1 ). Sensitivity analyses To investigate the robustness of the meta-analysis, sensitivity analyses were performed regarding the overall effect of the addition of cells to biomaterials. The overall outcome effect for cellular scaffolds was not notably affected by the exclusion of studies (1) with assumptions (2) or written in Chinese (no risk of bias assessment performed). Also for acellular biomaterials, the exclusion of these studies had no effect on cartilage regeneration. Publication bias Publication bias was assessed for all studies included in the meta-analysis comparing cartilage regeneration using acellular versus cellular biomaterials. Although the funnel plot ( Fig. 4 ) is rectangular in shape, no major asymmetry was observed, giving no indication for publication bias ( p -value 0. 866). 10. 7717/peerj. 3927/fig-4 Figure 4 Funnel plot of the studies included in the meta-analysis comparing cartilage regeneration using cell-laden and acellular biomaterials. No substantial asymmetry was found. Discussion Bone marrow stimulation can be applied to induce cartilage regeneration. Despite therapy, the formed neotissue generally consists of fibrous cartilage, which lacks mechanical and biological properties of native tissue ( Dai et al. , 2014 ). Therefore, microfracture results in temporary clinical improvement only ( Saris et al. , 2014 ). To regenerate more durable cartilage tissue, regenerative medicine and tissue engineering may offer a promising addition to bone marrow stimulation by the implantation of scaffolds, which can act as a template to guide and stimulate cartilage regeneration ( Cucchiarini et al. , 2014 ). In a previous systematic review, the quality of newly formed cartilage in animals was improved by the implantation of biomaterials after bone marrow stimulation, which was further enhanced by loading biomaterials with biologics ( Pot et al. , 2016 ). The aim of this systematic review was (a) to provide a comprehensive and systematic overview of all current literature regarding animal studies on cartilage regeneration using cellular versus acellular biomaterials and to identify knowledge gaps, (b) to assess the efficacy of cartilage regeneration using cellular versus acellular biomaterials and to investigate the effect of various parameters (i. e. , stem/somatic cells, cell source, cell culture conditions), (c) to gain insight in the methodological quality of animal studies, and (d) to improve the design of future animal models and eventually clinical trials. In animal studies, the implantation of cellular biomaterials in animal models significantly improved cartilage regeneration by 18. 6% compared to acellular biomaterials. Seeding of cells is a major component of the tissue engineering paradigm, which may stimulate healing by the production of many bioactive components. Therefore, the addition of cells to biomaterials enhanced the regenerative process ( Wang et al. , 2017 ). The heterogeneity ( I 2 ) for the main research question and subgroup analyses was very high. Results should therefore be interpreted with caution, especially for subgroup analyses with a limited number of studies. Further clinical studies are required to assess the potential beneficial effect of cellular biomaterials versus acellular biomaterials in patients. Marcacci et al. (2005) published promising results of a multicenter clinical phase III retrospective cohort study in which patients were treated with an implant consisting of autologous chondrocytes grown on Hyalograft C, a hyaluronic acid derivative, with a 3-year follow-up. Assessment indicated major clinical improvements and hyaline-like cartilage for the majority of biopsies. In a subgroup analysis, no significant differences were found between somatic cells and stem cells. Differences were found between various cell types. Adipose-derived stem cells (ADSCs) reduced cartilage regeneration in the subgroup analysis. However, cartilage regeneration using biomaterials seeded with ADSCs was still superior to biomaterials without cells. As compared to other cell types, the origin of ADSCs from fatty tissue may have resulted in significantly reduced cartilage regeneration compared to cells derived from cartilage or subchondral bone. MSCs and chondrocytes have distinct advantages. MSCs are not limited by donor-site morbidity and matrix production after expansion in vitro ( Bernhard & Vunjak-Novakovic, 2016 ), can be harvested from numerous sources, maintain their multipotency after expansion in vitro, can differentiate into chondrocytes that produce cartilage matrix and may suppress proinflammatory cytokines by their immunoregulatory properties. Chondrocytes on the other hand do not terminally differentiate after chondrogenic differentiation, which results in bone formation ( Bernhard & Vunjak-Novakovic, 2016 ), and are more easy to manipulate ( Deng et al. , 2016 ). In clinical trials, the addition of MSCs or chondrocytes to biomaterials resulted in comparable cartilage regeneration ( Nejadnik et al. , 2010 ; Lee et al. , 2012 ). In this study no subgroup analysis was performed to investigate the culture of cell-loaded scaffolds in bioreactors. Bernhard & Vunjak-Novakovic (2016) described the beneficial effects of culturing cell-loaded scaffolds in bioreactors with mechanical loading protocols, as these scaffolds more closely resembled the native compressive properties of cartilage tissue and as the applied force steered the location and alignment of cartilage matrix deposition by chondrocytes ( Bernhard & Vunjak-Novakovic, 2016 ). Study characteristics showed a large heterogeneity between studies due to differences in animal model, performed surgery, implanted biomaterial and follow-up period. To reduce the influence of possible confounding parameters, we excluded studies using healthy animals in which created defects were not filled during the first surgery and osteoarthritis animal models, despite their greater relevance for future applications to treat patients with osteoarthritis. Various outcome measures were used to investigate cartilage regeneration, including MRI, macroscopic and histological evaluation (more extensively discussed in Pot et al. , 2016 ). We selected data from semi-quantitative histological scoring systems as outcome measure, because histological scores are frequently used and allow for quantitative comparisons between studies. However, different scoring systems are available (extensively reviewed by Rutgers et al. ( Hooijmans et al. , 2014 )) that assess different processes, e. g. , cartilage regeneration only, cartilage and subchondral bone regeneration, and additional biomaterial degradation. Not discriminating between these parameters may be considered as a limitation, but usage of all scoring systems may provide an extensive and complete overview of all aspects affecting the regenerative process. Additionally, evaluation of cartilage regeneration using semi-quantitative histological scoring may still be observer-dependent and subjective, possibly inducing observer (detection) bias. Therefore, it may be better to combine histological scores with biochemical parameters and biomechanical properties, but the ideal combination of outcome parameters remains unknown ( Hooijmans et al. , 2014 ). The methodological quality assessment was performed to evaluate the experimental designs and reliability of the results of included studies. The methodological quality (internal validity) is of great importance since a low methodological quality may result in an overestimation or underestimation of the effect size ( Higgins et al. , 2011 ). No studies were included in or excluded from the meta-analysis based on methodological quality assessment results. Generally, the possibility of assessing the actual risk of bias was limited due to the absence of important details regarding the experimental set-up in most studies and method of randomization. It may be that the animal studies were performed well, but that experimental designs were only reported poorly ( Hooijmans et al. , 2012 ). For the analysis of the histological sections, however, most studies described that sections were randomized and that outcome assessors were blinded. Detection/observer bias may be introduced in case blinding was not performed and can result in an overestimation of the actual effect of the therapy ( Bello et al. , 2014 ). The overall validity of the study results may be impaired by bias due to the lack of blinding and randomization ( Bebarta, Luyten & Heard, 2003 ; Hirst et al. , 2014 ). Reporting of animal studies may be improved by using standardized protocols, including the ARRIVE guidelines ( Kilkenny et al. , 2012 ) or golden standard publication checklist ( Hooijmans et al. , 2011 ). A high translational value of animal studies is crucial to take treatments forward to clinical practice. Therefore, validated and predictive animal models are required. Many challenges and limitations are associated with the use of animal models for cartilage defects. Chu, Szczodry & Bruno (2010) and Ahern et al. (2009) extensively described strengths and shortcomings of different animal models related to e. g. , joint size, cartilage thickness, defect size, intrinsic healing potential and animal maturity, in comparison to lesions in clinical studies. In most animal experiments, the follow-up period was maximally six months, while in patients clinical improvements are generally observed up to 1. 5–3 years after microfracture surgery ( Hoemann et al. , 2010 ; Van der Linden et al. , 2013 ). The translational value and considerations to select animal models were extensively discussed before ( Pot et al. , 2016 ). Improved reporting of animal studies is required in future studies and studies should strive to resemble the clinical situation to facilitate translation. For clinical application of new regenerative medicine and tissue engineering strategies, including the use of biomaterials, biologics and cells, the effectiveness needs to be proven both in animal models and clinical studies ( Cousin et al. , 2016 ). Moreover, the cost-effectiveness of new interventions in clinical practice may be assessed using early health economic models ( De Windt et al. , in press ). Considerations for the addition of cells to biomaterials are of great importance and limitations (including donor-site morbidity, cell culture costs, regulatory issues, limited off the shelf availability, and potential multiple-stage surgical procedures ( Pot et al. , 2016 ; Efe et al. , 2012 )) should be weighed against potentially superior cartilage regeneration by applying cellular biomaterials. Difficulties in controlling cell culture and the development of novel materials stimulating tissue regeneration may justify the use of acellular biomaterials. Future research focusing on biomaterials properties, source and manipulation of cells, and possibly patient profiling, may allow selection of the best treatment for each individual patient ( Kon et al. , 2015 ). Conclusion This systematic review and meta-analysis provides an extensive overview of all animal studies applying regenerative medicine and tissue engineering approaches to regenerate articular cartilage by implantation of cellular versus acellular biomaterials after applying bone barrow stimulation. Cartilage regeneration was more effective by implantation of cellular biomaterials compared to acellular biomaterials. This study together with a previous study on the beneficial properties of scaffolds and growth factors implies that all components of the tissue engineering paradigm can be valuable for improved regeneration of articular cartilage. Supplemental Information 10. 7717/peerj. 3927/supp-1 Supplemental Information 1 Rationale and contribution systematic review and meta-analysis Click here for additional data file. 10. 7717/peerj. 3927/supp-2 Supplemental Information 2 PRISMA flow diagram Click here for additional data file. 10. 7717/peerj. 3927/supp-3 Supplemental Information 3 Supplemental Information 1 Click here for additional data file. 10. 7717/peerj. 3927/supp-4 Supplemental Information 4 Supplemental Information 2 Click here for additional data file. 10. 7717/peerj. 3927/supp-5 Supplemental Information 5 Supplemental Information 3 Click here for additional data file. 10. 7717/peerj. 3927/supp-6 Supplemental Information 6 Supplemental Information 4 Click here for additional data file. 10. 7717/peerj. 3927/supp-7 Supplemental Information 7 Supplemental Information 5 Click here for additional data file. 10. 7717/peerj. 3927/supp-8 Supplemental Information 8 Supplemental Information 6 Click here for additional data file. 10. 7717/peerj. 3927/supp-9 Supplemental Information 9 Supplemental Information 7 Click here for additional data file. 10. 7717/peerj. 3927/supp-10 Supplemental Information 10 PRISMA checklist Click here for additional data file.

10. 7717/peerj. 4076

PeerJ

The role of biomaterials in the treatment of meniscal tears

Extensive investigations over the recent decades have established the anatomical, biomechanical and functional importance of the meniscus in the knee joint. As a functioning part of the joint, it serves to prevent the deterioration of articular cartilage and subsequent osteoarthritis. To this end, meniscus repair and regeneration is of particular interest from the biomaterial, bioengineering and orthopaedic research community. Even though meniscal research is previously of a considerable volume, the research community with evolving material science, biology and medical advances are all pushing toward emerging novel solutions and approaches to the successful treatment of meniscal difficulties. This review presents a tactical evaluation of the latest biomaterials, experiments to simulate meniscal tears and the state-of-the-art materials and strategies currently used to treat tears.

Introduction The knee is considered a hinge joint; however, because it also features characteristics of an arthrodial joint, it is a more complex joint than other hinge joints such as the elbow and ankle. The knee consists of two articulations which form the tibiofemoral joint (further separated into the medial and lateral tibiofemoral joints) and the patellofemoral joint. The articulations are not entirely congruent and this arrangement allows for the combination of gliding and rolling motions which is constrained mainly by the ligaments of the knee. The menisci are fibrocartilagenous structures that sit on top of tibia to deepen the plateaus with the primary functions transmitting load through the joint and also serve to increase joint stability and lubrication of the articular cartilage ( Seedhom, Dowson & Wright, 1974 ; Walker & Erkman, 1975 ; McDermott, Masouros & Amis, 2008 ). The menisci are commonly injured due to traumatic events and/or degenerative stresses. In the United States alone, it was estimated that approximately 6. 6 million patient visits to the emergency department between 1999 and 2008 were due to knee injuries equating to 2. 29 knee injuries per 1, 000 people ( Gage et al. , 2012 ; Reid et al. , 2017 ). Furthermore, by 2060 the percentage of people reaching an age of 50 will reach 50% representing a change in population demographic and likelihood for pressures in the knee. A large proportion of knee injuries in the general population are meniscal related and meniscal injuries are even more common in a physically active population ( Baker et al. , 1985 ; Nielsen & Yde, 1991 ). Given the role meniscal tears, and subsequent partial or full removal of the meniscus, play in development of osteoarthritis ( Englund, Roos & Lohmander, 2003 ; Roos et al. , 1998 ) there is an increased interest in preservation of these structures following injury. For this reason, there is also an increased interest the role biomaterials play in meniscal repair, regeneration and replacement options. Advances in materials technology have brought about an increased usage of biomaterials and medical devices in the body ( Hallab, Link & McAfee, 2003 ; Chevalier, 2006 ; Yamamoto, Takagi & Ito, 2016 ). A biomaterial is a material or substance or combination of substances, other than drugs, synthetic or natural in origin, which can be used for any period of time ( Bochyńska et al. , 2016c ; Brannigan & Dove, 2017 ; Kaur, 2017 ), which augments or replaces partially or totally any tissue, organ or function of the body in order to improve the quality of life of an individual ( Bergmann & Stumpf, 2013 ). The biomaterial must be able to interact with the surrounding human tissue and body fluids to improve or replace the anatomical defect. Some examples of the recent advances for biomaterial use in medicine include knee and hip replacement ( Walczak, Shahgaldi & Heatley, 1998 ; Bahraminasab & Farahmand, 2017 ), ocular implants ( Lloyd, Faragher & Denyer, 2001 ; Baino et al. , 2017 ; Mota et al. , 2017 ), heart valves ( Vongpatanasin, Hillis & Lange, 1996 ; Vander Roest & Merryman, 2016 ; Emmert & Hoerstrup, 2017 ), bone implants ( Bròdano et al. , 2014 ; Apicella et al. , 2017 ), dental implants ( Tamimi et al. , 2014 ), biosensors ( Sun et al. , 2014 ; Calvo et al. , 2017 ), orthopaedic screws and sutures ( Waizy et al. , 2014 ; Zhao et al. , 2017 ) and tissue allografts ( Cameron & Saha, 1997 ; Sanen et al. , 2017 ). The achievements, in terms of biocompatibility, to lower risk of failure and improved surgical outcomes have contributed to the expanding use of biomaterials. For these reasons advancements in biomaterial development is and has been a significantly fast-growing area of research. This review article will focus on providing a general review of the menisci and meniscal injuries. We also discuss biomaterials and the subsequent role biomaterials play in the surgical treatment options for meniscal repair, regeneration and replacement as well as future directions. While other reviews have been developed, their focus has been to provide an overview of materials only, this review provides significant detail on cell lines used, models and materials to support research momentum for future developmental medical breakthroughs. Meniscus Biomechanics and function The menisci are fibrocartilage structures, composed mainly of type 1 collagen, that sit on top of tibia, Fig. 1. 10. 7717/peerj. 4076/fig-1 Figure 1 Superior view of the right tibia in the knee joint illustrating the menisci and cruciate ligaments. (A) anterior cruciate ligament, (B) articular cartilage on medial tibial condyle, (C) medial meniscus, (D) posterior cruciate ligament, (E) lateral meniscus, (F) articular cartilage on lateral tibial condyle. The lateral meniscus (e) is a C-shaped structure that covers approximately 80% of the lateral tibial plateau whereas the medial meniscus (c) is a U-shaped structure and covers only 60% of the medial tibial plateau. The menisci are relatively avascular with only 10–30% of the peripheral region of the medial meniscus and 10–25% of the lateral meniscus being vascular ( Arnoczky & Warren, 1982 ). Based on its vascularisation, the menisci can be divided into three zones: the red-red vascular zone (outer peripheral region), the white-white avascular zone (inner region) and the red-white zone which lies between of the two other zones and has characteristics of other two zones. The red vascular region is thick and convex and attaches to the capsule of the joint whereas the white-white inner region is thin, concave and is a free edge unattached to the joint. The menisci effectively deepen the tibial plateau and allow smooth articulation between the tibial and femoral condyles and the transmission of loads across the tibiofemoral joint. In full knee extension, the medial meniscus transmits approximately 50% of the load on the medial compartment, while lateral meniscus transmits approximately 70% of the load in the lateral compartment ( Walker & Erkman, 1975 ). As knee flexion increases the amount of load transmitted to the lateral meniscus increases such that when the knee is flexed beyond 75°the entire load that passes through the lateral compartment, is transmitted by the lateral meniscus ( Walker & Erkman, 1975 ). For the medial meniscus the increase in load transmission as the knee flexes is less apparent ( Walker & Erkman, 1975 ). When the meniscus is intact, the load is well distributed across the tibiofemoral compartment; however when part or the entire meniscus is removed there is considerable alterations to load distribution such that there is a decrease in the contact area and increases in peak contact forces ( Bedi et al. , 2012 ; Lee et al. , 2006 ; Ihn, Kim & Park, 1993 ). Meniscal tears Meniscal tears are one of the most common intra-articular knee injuries ( Clayton & Court-Brown, 2008 ; Majewski, Susanne & Klaus, 2006 ) and is typically the result of an axial loading and rotational forces which result in a shear load on the meniscus ( Browner, 2009 ). This may be a result of a traumatic event or cumulative stress leading to degenerative tears. The medial meniscus is more often injured than the lateral ( Majewski, Susanne & Klaus, 2006 ); however, lateral meniscal tears are more often associated with acute ACL tear ( Bellabarba, Bush-Joseph & Bach Jr, 1997 ). Although there is no uniformly accepted classification of meniscal tears, the classifications typically involve a description of the tear pattern and location. Common tear patterns that typically originate from traumatic events include longitudinal, bucket-handle, and radial tears ( Greis et al. , 2002 ). Whereas horizontal, flap and complex tears are typically seen in older adults and due to cumulative stress resulting in degeneration ( Greis et al. , 2002 ). The location of the tears may be classified based on the zone classification system purposed by Cooper, Arnoczky & Warren (1990) in which the menisci are divided into three radial zones (anterior, medial and posterior) and four circumferential zones (meniscosynovial junction or periphery, outer third, middle third and inner third of the menisci) ( Fig. 2 ). 10. 7717/peerj. 4076/fig-2 Figure 2 Schematic diagram highlighting the various types of meniscal tears, Bucket Handle MRI image taken from Han et al. (2015) (CC BY NC 3. 0), Radial Tear, MRI image taken from Jung et al. (2012), and longitudinal (photograph taken from Feucht et al. (2015) (CC BY 4. 0)) and horizontal tears (MRI taken from Ohishi et al. (2010) (CC BY 2. 0)) all with permission. In a similar fashion to the zone classifications, tears may be graded as partial or full-thickness tears or using a grading scheme 0–III in which 0 indicates a normal intact menisci and III a full-thickness tear ( Cooper, Arnoczky & Warren, 1990 ; Pihl et al. , 2017 ). Available treatment options Meniscal tears account for a significant portion of surgical procedures performed by orthopaedic surgeons, the patient experiences significant pain and sometimes complete disability from these tears and resulting procedures ( Berthiaume et al. , 2005 ). The surgical procedures involved in the treatment of a meniscal tear may include a partial or full meniscectomy or a meniscal repair. The meniscectomy procedures involve either part or all of the damaged meniscus being removed which in turn leads to higher rates of osteoarthritis in subsequent years. Surgical treatment of meniscal injuries has undergone a number of developments over the past two decades, moving from open arthroscopic surgery; from total to partial meniscectomy and adding novel treatments; such as repair using a variety of devices, materials, transplants, collagen implants or xenografts ( Kim et al. , 2013 ; Jiang et al. , 2012 ; Grogan et al. , 2017 ; Baek et al. , 2017 ). If meniscectomy takes place or insisted upon, this procedure is mainly due to changes in load distribution across the articular cartilage as studies have shown that following total meniscectomy peak contact pressures increase by 253% and 165% following partial meniscectomy ( Lee et al. , 2006 ; Baratz, Fu & Mengato, 1986 ; Beamer et al. , 2017 ; Van Egmond et al. , 2017 ). Following meniscectomy, there is also evidence of reduced muscle strength, altered gait patterns and clinical outcomes ( Hall et al. , 2013 ; Hall et al. , 2014 ; McLeod et al. , 2012 ; Sturnieks et al. , 2008a ; Sturnieks, Besier & Lloyd, 2011 ; Sturnieks et al. , 2008b ; Salata, Gibbs & Sekiya, 2010 ; Scholes et al. , 2017 ). For these reasons, there are an increasing number of interests in performing meniscal repair. What needs to be remembered is that not all meniscal tears are suitable for repair, and thus other treatment options such as meniscal replacement and regeneration are of considerable interest when a surgical intervention is necessary to improve any pain and symptoms. Biomaterials Current treatment modalities for meniscal repair tears still carry their drawbacks and novel, robust and effective solutions are required. Some recent advances in meniscus research suggest that low cellularity, ( King et al. , 2017 ) dense ECM and poor vascularisation coupled with the inflammatory responses ( King et al. , 2017 ) in the knee joint are responsible for a lack of healing. Recently, biomaterials in the form of tissue adhesives have become available for clinical use: fibrin glue, ( Bochyńska et al. , 2016a ) TissuGlu ®, Dermabond ®, ( Balakrishnan et al. , 2017 ) where the development of these new adhesive biomaterials has improved the properties of existing biomaterials alone (TissuGlu ®, Raleigh, NC, USA; Ethicon Inc. , Somerville, NJ, USA). Furthermore, these materials and strategies are not always a given success, presenting limitations to the accomplishment of the meniscal reparation. Tissue engineering using biomaterials Of late, tissue-engineering and cellular biomaterial interactive concepts have been introduced to develop cellular-based reparation for cartilage regeneration ( Temenoff & Mikos, 2000 ). The type of cell used to engineer cartilage is critical as a future goal of biomaterial development. Various cell populations that have been investigated for these roles include: chondrocytes ( King et al. , 2017 ; Chen & Cheng, 2006 ), mesenchymal stem cells, bone marrow stromal cells and perichondrocytes ( Bruns et al. , 1998 ). The choice of biomaterial is critical to the success of tissue engineering approaches for cartilage repair. The concept of ‘tissue engineering’ was first introduced and postulated by Green Jr (1977) where chondrocytes grown ex vivo could be transplanted into a region of tissue defect. Recently, tissue and biomaterial engineering concepts have been initiated to develop cellular based approaches for tissue repair ( Freed et al. , 1993 ). Typically, the process for engineering tissue involves the isolation of chondrocytes which are then seeded into a biocompatible matrix or scaffold and finally cultivated for implantation into the defected region. A large variety of biomaterials, natural and synthetic, have been employed as potential cell-carriers for tissue regeneration. The most common naturally occurring materials include type I and type II collagen-based biomaterials. Furthermore, some of the contrasting synthetic approaches include: polyglycolic acid or poly-L-lactic acid or other various composite mixtures ( Chen & Cheng, 2006 ). In essence, an ideal candidate biomaterial would be a cell-carrier substance which closely mimics the natural environment in the surrounding matrix—as given by the definition of a biomaterial. Regenerative approaches to meniscus repair occurs in a series of precise stages. It is typically understood that the low cellularity (endogenous meniscus cells and meniscus progenitors) ( Mauck & Burdick, 2015 ), the dense ECM, poor vascularisation potential and the inflammatory responses typically linked to meniscus wounds all contribute to the success or failure of the meniscus healing and regeneration alone. This success of healing process is without a biomaterial introduced into the site. Based on these principles, the potential use of a biomaterial to develop and deliver a viable solution requires thought around this repair process. Biomaterials are typically promoters of tissue repair through provision of scaffold layers for cellular attachment and growth and differentiation further acting as a vehicle for protein and gene transfer to regenerate functional tissue approaches ( Chen, Zhang & Wu, 2010 ). Biomaterials in this area should have several properties to support viable repair. Typically, this is achieved through: (1) The material must act as a support structure for cell lines (i. e. , cells that are seeded in vitro are compatible, adhere to the material if required or certain cell lines are not required; filtered out). For meniscal repair the biomaterial must provide appropriate biomechanical functions after implantation to shield cells from damaging or compressive forces; (2) Possess sufficient mechanical strength to protect the surrounding cells (cells should be mechanically stable i. e. , cell attachment is maintained). For meniscal repair the biomaterial must maintain their shape and integrity, mechanical stability and strength for the defect area in question until new host tissue has been regenerated. Furthermore, it may be important to provide biological and mechanical context for cell differentiation, proliferation and attachment when a biomaterial is introduced into the knee. For example, it is now very well understood that cells are influenced by the local external environment including the adhesive and biophysical properties ( Engler et al. , 2006 ); (3) Withstand in vivo forces during the joint movement operation (mechanical and structural stability of the biomaterial in the meniscus area needs to be able to withstand compressive and tensile forces (these forces have been aptly described in Paschos et al. (2017) ); (4) Bioactivity should be provided to accommodate cellular attachment and cellular migration (the biomaterial in the meniscus will therefore be able to promote tissue regeneration). Furthermore, providing directional cues, such as chemotactic gradients to guide cells like endogenous cells to the injury site. Recently, some studies have shown that allowing migration of cells provides a motivation for the cells to attach and drives the cellular colonisation process ( Mauck & Burdick, 2015 ; Greiner et al. , 2014 ); (5) The biomaterial should have biodegradable properties and be able to remodel as the novel cartilage grows, embeds and replaces the original construct; therefore, the matrix must be non-toxic, non-adhering and non-stimulating for inflammatory cells. The biomaterial for a meniscus should therefore facilitate host tissue integration and provide the appropriate biomechanical function in the knee. (6) Furthermore, they should be non-immunogenic as this is catastrophic for the biomaterial insertion. For any biomaterial, this is important, to prevent rejection the appropriate level of biocompatibility and non-toxic ability needs to be considered. Biocompatibility One of the most important non-mechanical requirements of orthopaedic biomaterials is biocompatibility. Biocompatibility is the ability of a substrate to exist in contact with tissues of the human body without causing an unacceptable degree of harm in the body. The biomaterial domain has been aptly described by Mardis and Kroeger, “the utopian state where a biomaterial presents an interface with physiologic environment without the material adversely affecting the environment or environment adversely affecting the biomaterial” ( Mardis & Kroeger, 1988 ). An understanding of biocompatibility requires an appreciation of tissue cell, bacterial cell and host defence response to the insertion of a biomaterial in particular for this review—for meniscal interventions. Once the biomaterial has been placed into the body, a conditioning film containing biomolecules such as; water, electrolytes, cholesterol, vitamins, lipids and proteins ( Chapman et al. , 2013 ) (albumin, igG, fibronectin, fibrinogen, laminin, collagen and osteopontin) form on the surface long before cells are present and reach the state of equilibrium ( Thevenot, Hu & Tang, 2008 ). In the very early implantation period or injury for this matter, inflammatory cells begin to proliferate, this is an immediate response ( Anderson, Rodriguez & Chang, 2008 ). The first contact with tissue, proteins in blood and the interstitial fluids adsorb on to the biomaterial surface. An injury to vascularised connective tissue initiates the inflammatory response but also leads to the process of thrombus formation involving the activation of the extrinsic and also intrinsic coagulation, complement, fibrinolytic, kinin-generating systems and platelets ( Anderson, Rodriguez & Chang, 2008 ). The conditioning layer represents a dynamic, ever-changing layer due to differential diffusion and mass transport of molecules in and out of the implant surface. Later stages of competitive binding then occur on the surface of the material owing to functional groups within the molecules. Cells therefore never see the ‘true’ surface of the biomaterial, but more correctly, respond and interact to a conditioned film that has consequently developed in-situ. Following the conditioning sequence of the biomaterial, attachment cells secure themselves to the protein and protein matrices using integrin receptors. Thus, this conditioning layer is vital to the reaction of cells to the surface of the implanted biomaterial. The introduction of the biomaterial, the conditioning and immune response sequence is not always obvious as proteins have the ability to conform and expose epitopes that are not always identified as self-produced by the body’s immune system. Immune cells react as they detect what were once normal proteins and recognise them as foreign bodies. This process can result in a cascade of blood coagulation and chronic inflammation that can lead to occlusion of nutrients, changes in oxygen and fibrous capsule formation—operating toward total rejection by the body of the implanted biomaterial ( Nasab & Hassan, 2010 ). The extent of the deformation process for proteins has been remedied based on the selection of material type. Surfaces are made more “passive” where chemical treatments are added to the manufacturing process. Passivation with acids such as nitric acid of stainless steel creates a less reactive oxide layer; this has been shown to improve the biocompatibility process. One added benefit to passivation is it serves as a means for removing foreign material from the surface, such as bacteria or biofilms ( Blumenfeld & Bargar, 2006 ). Passivation can also be used to surface-modify natural or synthetic polymer biomaterial substrates for meniscal tear applications. For example, albumin, where the resulting surface passivation has been shown to reduce and prevent clotting ( Kaur, 2017 ; Hanker & Giammara, 1988 ). Role of biomaterials in meniscal repair An article by Abrams et al. (2013) has shown that while there was no increase in the overall number of meniscal procedures, over a seven-year period there has been an 11. 4% increase in isolated meniscal repairs and a 48. 3% increase in meniscal repairs in combination with ACL reconstruction. This sharp increase in meniscal repair treatment is mainly due to the increased knowledge in the importance of the preservation of the meniscus to maintain normal knee function and prevent osteoarthritis. It has been shown that following meniscal repair, peak contact pressures are similar to that experienced with an intact meniscus ( Bedi et al. , 2010 ). Unfortunately, it is estimated that currently only 20% of all meniscal tears are repairable. Tears in the meniscal periphery (i. e. , the red-red vascular zone) are most likely to heal whereas those in the meniscal avascular zone (i. e. , the white-white zone) are unlikely to heal and those in the red-white zone have the potential to heal ( Belzer & Cannon, 1993 ; Noyes & Barber-Westin, 2012 ). Besides vascularisation, tear type and various patient characteristics can influence decision making on treatment options and success of a meniscal repair. Typically, tears that are less than 2 cm in length, longitudinal and acute are more amendable to repair than larger tears ( Taylor & Rodeo, 2013 ; Laible, Stein & Kiridly, 2013 ). Meniscal repairs are also not typically recommended for degenerative tears and thus repair success is typically superior in young patients (less than 50 years of age) ( Laible, Stein & Kiridly, 2013 ). When appropriately performed, meniscal repairs provide considerable improvements in terms of clinical outcome and osteoarthritis prevention compared to a partial meniscectomy ( Stein et al. , 2010 ). Thus, finding ways to increase the number of meniscal tears that can be treated by meniscal reparation is of great importance. Vascularisation in the meniscus tissue is of high relevance to biomaterial design. From prenatal development up until after birth, the meniscus is fully vascularised. Following this, from the age of ten, vascularisation reduces to 30% of the meniscus and at maturity the meniscus only in the peripheral region of approximately 10% of the tissue. Vascularisation represents another challenge in meeting the requirements of success for biomaterial implantation as a meniscus operation. Vascular endothelial growth factor enhances the blood vessel density in peri-implant spaces. Biomaterial scaffolds of knee menisci exist in a highly challenging environment as little vascular support is provided in this region of the body. Electrospinning of polymeric fibres can be produced to support other engineering applications such as blood vessel, tendons, meniscus and cartilage ( Xu et al. , 2004 ). Some authors have used unique biodegradable nanofibers as a scaffold to support blood vessel engineering. They have demonstrated that fibres of 500 nm with an aligned topography is able to mimic the circumferential orientation of cells and fibrils ( Leong et al. , 2009 ). The authors have postulated that macrophage within the CES produce angiogeneic growth factors that potentially stimulates vascularisation. Role of biomaterials in meniscal replacement/regeneration Owing to the limited percentage of meniscal tears that can be repaired and the poor clinical results with untreated symptomatic meniscal injuries and partial meniscectomy, biomaterial synthetic and allogeneic (genetically dissimilar) interacting biomaterials have been investigated to serve as a matrix to lead meniscal regeneration medicine, particularly as a cellular support. Hydrogels Using a biomaterial that has the ability to seamlessly integrate in to water matrices is another attractive property in regenerative medicine applications for meniscal repair. Hydrogels are one such material with a considerable water based content; using hydrated polymer networks capable of absorbing and retaining fluids. Hydrogels are determined by their monomeric composition, crosslinking density and polymerisation ability. Due to the crosslinking chemistry the polymer remains insoluble in solution. The insolubility, along with the high hydration threshold, make hydrogels appealing to use for human tissue mimetics ( Kobayashi, Chang & Oka, 2005 ). As an example, some authors have used a poly (vinyl alcohol) hydrogel with a water content of approximately 90% to produce knee implants using a rabbit model ( Makris, Hadidi & Athanasiou, 2011 ). The implant replaced the whole lateral meniscus over two years. In a subsequent study, the hydrogel implant was not able to prevent damage to articular cartilage but was able to reduce progression of meniscal decay. Some of the new and emerging biomaterial types have been shown in Table 1. 10. 7717/peerj. 4076/table-1 Table 1 Summary of biomaterial studies used in meniscus research. Biomaterial used Author Engineering region Success(es) Species model Ramifications Synthetic polymers Hydrogels Kim & Healy (2003) Meniscus tissue engineering Maintained 90% water content that are not degraded by proteases. The hydrogels used in this study were incorporated with non-reducible collagen crosslink, pyridinoline. Mammalian Peptide linked hydrogels have the ability to be tailored to create environment responsive artificial extracellular matrices that are degraded by proteases. Rey-Rico, Cucchiarini & Madry (2017) Meniscus tissue engineering Review article providing results on specific 3D microenvironments using hydrogels. Many hydrogel polymers were used in this paper. Human and animal models Hydrogels can be used as a platform for precision and targeted meniscus tissue engineering Polygolic acid Buma et al. (2004) Meniscus tissue repair Optimal pore geometry realised (15–25 µm) Canine Autologous meniscus cells seem to be the optimal cell source for tissue engineering. Research should be stimulated to demonstrate suitability of other cell lines for meniscal repair. Ibarra et al. (1997) Tissue engineered meniscal tissue repair Used to replace massive tears or completely resected menisci Bovine A pivotal paper to show that autologous cells could eventually be used to replace allografts for meniscus transplantation. Natural Polymers Collagen-glycosaminoglycan (GAG) Mueller et al. (1999) Regeneration approaches to knee meniscus Type II GAG matrix increased DNA content and cellular response to the matrix over 3 weeks Canine Type II matrix for the number of cells and the higher GAG synthesis of type II matrices commend further investigation and regeneration of meniscus in vivo. McCorry & Bonassar (2017) Tissue engineering for meniscal repair Mesenchymel stem cells increased the GAG and collagen production in both co-culture and monoculture groups in a 4 week study. Bovine MSC lacks fibre organisation capability. The study suggests that GAG production and fibre formation are largely linked, therefore co-culture techniques can be used to balance synthetic properties and matrix modelling capability. Collagen sponge Walsh et al. (1999) Medial meniscus repair Collagen sponge acted as a scaffold producing abundant tissue repair. Canine Degenerative changes were present in both groups indicating biomechanical function was compromised. Murakami et al. (2017) Meniscal scaffold structure and repair Collagen sponges demonstrated greater strength. At 12 weeks stress and compression testing was performed, lower inflammation was noted in all samples coated with PGA. Foreign body multinucleated giant cells in implanted groups appeared in weeks 8. Lapine Meniscal scaffolds using PGA should possess biological and biomechanical functions. The PGA coating was a beneficial property of the scaffold and offers excellent biomechanical function, regeneration and ultimately less inflammation in this material type. Copolymeric (L-lactide/epsilon-caprolactone) De Groot et al. (1997) Meniscal repair Copolymer implants demonstrated improved adhesion; fibrocartilage was affected of the compression modulus. The copolymer was degradable. Canine Tearing problems usually associated with sutures were partly circumnavigated, this paper paves the way for more work in meniscal prostheses including transplantation. Chitin Chitin sutures are an emerging material of choice for improvement of the mechanical properties of a knee healing process ( Brittberg et al. , 1994 ). Owing to its favourable mechanical properties, chitin has been used for applications that require exceptional integrity and physical strength in surgical sutures, some new medical textiles and even as bone substitute materials. Nanofibres Electrospun scaffolds are also another emerging biomaterial that has begun to be used for cellular adhesion applications in regenerative medicine. The fibres have the ability to mimic both anisotropy of fibrous tissues and withstand high load forces that are imposed on the tissue during physiological motions ( Ionescu & Mauck, 2012 ). The electrospun biomaterials can also be tailored to produce various size, shapes and makeup (for example coaxial materials) will influence cell interactions and the cells will begin to proliferate and adhere and finally deposit matrix on to the fibre network. These interactions provide improved mechanical properties for the biomaterial scaffold over time. Fibres can be collected on to rotating drums or flat collection plates, depending on the order, orientation and architectures that they are required. Cells typically are seeded on to these scaffolds and cultured over time in vitro. In a study by Passaretti et al. (2001) tensile modulus was seen to improve on fibre aligned scaffolds some 7-fold higher than disorganised fibres approaching the value of a normal meniscus. Essentially, the authors determined that cells prefer to align on ordered scaffold fibres rather than disorganised arrangements. Further to these findings, internal organisation in the form of sheet fibres can also be arranged for tissue-mimicking structures. Specifically, for meniscal tissue engineering, cells can be isolated, expanded and manually seeded on to the surfaces of electrospun scaffolds prior to an implantation operation, expediting the regenerative process. Cells along with host cells will migrate on to the newly implanted scaffold and deposit proteoglycan and collagen. Some implantation methods require surgery prior to this implant step to isolate the cells prior to seeding, maturation and implantation. Biodegradable polymers Some of the more current treatment methods for repair of meniscal tears are somewhat indifferent for positive results and outcomes. Tissue adhesives are a promising alternative, owing to their ease of application and minimal tissue trauma. Co-polymeric tissue adhesives have been shown to have adhesive strengths of 40–50 kPa and hold edges of meniscal tears together during healing periods. These results indicate that copolymers are able to improve tissue capacity for self-repair specific for meniscal applications. Other authors have used amphiphilic copolymers based on polyethylene glycol, trimethylene carbonate and citric acid to synthesise end-functionalised hexamethylene diisocyanate to form biodegradable hyper-branched tissue adhesives. The work showcases resorbable tissue materials for meniscus repair. The materials have excellent mechanical and adhesive properties that could be adjusted through variation of the composition of the copolymers ( Bochyńska et al. , 2016c ). Regenerative engineering converges a number of research areas and is truly multidisciplinary inclusive of tissue engineering, advanced materials, stem cell science and developmental biology to regenerate complex tissues from menisci to whole limbs ( Narayanan et al. , 2016 ). Clinical applications of tissue engineering technologies are still relatively restricted owing in part to the limited number of biomaterials that are approved for human use. While many biomaterials have been developed, their translation into practice has been extremely slow. Consequently, many researchers are still using biodegradable choices that were approved some 30 years ago. Most degradable biomaterials used to date comprise of synthetic polyesters: • Poly(L-lactic acid) PLLA; • Poly(L–glyolic acid) PLGA; and • Biological polymers such as: alginate or chitosan, collagen or fibrin ( Middleton & Tipton, 2000 ). Polyester-based polymers are clearly an excellent candidate as a synthetic biodegradable and bio-absorbable material for medical applications. The use of synthetic polyesters as biomaterials allow the unique control of the morphology, mechanical properties and degradation profiles measured through the monomer selection, polymer composition informed through the copolymer and homopolymer, stereo-complexation and also the molecular weight. In an excellent review by Brannigan and Dove, degradation mechanisms has been discussed in detail, in a clinical research capacity—these parameters are of paramount importance to understand the behaviour of the material in vitro or in vivo. The authors discuss enzymatic, oxidative, and physical degradation. Brannigan & Dove (2017) discuss the use and importance of polyester type Poly-HDPE scaffolds with an interconnected porous structure for cartilage regeneration. In their work, neocartilage formation within a synthetic polyester scaffold based on polymerisation of high internal phase emulsions were used. The fabrication of polyHIPE polymers (PHP) was ordered to have highly porous giving structure to the cartilage with a higher potential in force wear. Another example of the use of biodegradable polymers in meniscal repair research includes, poly lactic acid or L-PLA is used in menisca reconstruction in a study using canines, the presence of macrophages, fibroblasts, giant cells and lymphocytes were observed to be attaching to the material. From this study it seems that biocompatibility reduces when the degradation process ensues. This degradation property therefore promoted inflammatory responses and thus rejection ( Jones et al. , 2002 ). Discovery of new biomaterials—beyond state of the art The next phase in developing knee meniscal biomaterials for replacement and or regeneration applications extends to the design, discovery and evaluation of bioactive materials. Bioactive meniscal materials have been used with some significantly exciting and promising results. For example, bioactive scaffolds have been shown to modulate local ECM density to improve repair ( Shin, Jo & Mikos, 2003 ). A novel biphasic collagen scaffold and shown to support meniscal repair in vivo to support meniscal cell ingrowth but also producing ECM in vitro by Howard et al. (2016). The authors have shown that the addition of PRP enhanced scaffold enhanced healing ( Howard et al. , 2016 ). Other emerging materials which could show potential in meniscal repair include: cartilage matrix is also a promising material for cartilage regeneration given the emerging evidence supporting its chondroinductive character. The cartilage matrix is a promising material for hyaline cartilage tissue engineering applications and has been shown that cell derived matrix and ECM materials and have been demonstrated to show established decellularisation, representing an excellent and promising choice of new material for future direction. ( Redman, Oldfield & Archer, 2005 ). A drawback so far is that the FDA regulatory approval may affect the decision to use a native or cell-derived matrix. To expedite FDA approval, a full chemical decellularisation of allogeneic matrix may be used—this way, removal of cells ensures no cross-species interactions ( Sutherland et al. , 2015 ). For example, allogeneic cells from bone marrow can be used in cardiac repair ( Lemcke et al. , 2017 ). Initially, this is a relatively straightforward process whereby advanced synthesis of new materials can be performed. The difficulty lies with producing the novel activity and evaluation of the behaviour of the material in the biological system. Adapting the surface properties through the addition of synthetic peptides and or molecular drugs can yield thousands of candidate materials for testing. This approach has already been realised in the form of library derived screening techniques using commercially available methacrylate monomers—influencing attachment, growth, proliferation and differentiation of human embryonic stem cells ( Anderson, Levenberg & Langer, 2004 ). Further developments in biomaterials will continue to expand at the interface of nanotechnology. Understanding the tribological interaction with the surrounding interface of the human body is an approach that is being realised using the “bottom-up” approach ( Zur et al. , 2011 ). The bottom up approach will develop novel, self-assembling and environment reactive biomaterials. In particular, self-assembling peptides offer a new approach owing to the large variety of sequences that can be produced by chemical synthesis. These advances include the design of short peptides that have the ability to resemble nanofilaments which are compatible in vitro, without rejection. The use of peptides in polymeric materials allows for resistance in concentration, pH or level of divalent cation variability ( Hartgerink, Beniash & Stupp, 2002 ). The use of combinational gene therapy and biomaterial approaches is a recent technique to remedy meniscal lesions formed when orthopaedic surgery and loss of the meniscus has accelerated in the patient. The lack of therapeutic options suggests there is a need for improved treatments to enhance meniscal tear repair treatments/operations. Combinational approaches may also provide strategies to support this remedy ( Cucchiarini et al. , 2016 ). Gene therapy, can be directly applied as a combination or direct approach to meniscal repair strategies. A recent evaluation on gene therapy with cell and tissue engineering-based approaches demonstrates a six strategy approach: (a) directly using gene transfer vectors ( Elsler et al. , 2012 ), (b) administering genetically modified cells ( Nakagawa et al. , 2015 ), which could be fraught upon in some researching countries, continents, (c) implanting the biocompatible material that can deliver the recombinant factor, as we have seen rejection may be a potential problem with this result, (d) applying autologous platelet-rich plasma or fibrin clotting factors, (e) providing a biomaterial that delivers a gene transfer vector, (f) transplanting a material seeded with cells, again, we envisage a potential rejection with this treatment. Stem cell approaches Exciting new techniques are emerging as non-invasive approaches to meniscal tear correction using stem cells. The promising use of new tissue engineering approaches have incorporated natural biomaterials made from extracellular matrices of decellularised tissues from the heart, lung and bone for example ( Yuan et al. , 2017 ). The use of a scaffold or ‘shell’ to align stem cells upon in a given feature is fast becoming attractive. Decullularisation preserves the molecular composition with tissue specific molecules including structural and mechanical features present in the original tissue. The preservation step will aid in guiding the behaviour of the therapeutic cells and facilitate tissue development when implanted, non-invasively to the meniscal tear region. In vitro studies have also been used to investigate tissue surface modification with collagenase to prime the surface where the addition of the TGF-beta3 cells has been proven to increase the number of cells present in meniscal tears repaired with newly developed tissue adhesives such as isocyanate-terminated block polymers. For example, Bochynske et al. have used cylindrical explants harvested from bovine menisci, the explants were simulated to possesses a full thickness-tear where the explants were then removed and glued back to the defect. In addition, the repair constructs were then culture with and without the addition of TGF-beta3 and assessed for their histological appearances. The histological staining of the constructs confirmed that cytotoxicity was not an issue and after 28 days, meniscal cells were present in the contact glues ( Bochyńska et al. , 2016b ). The results demonstrate that the use of TGF-beta 3 induces thicker cell numbers round the edges of the annulus of the explants and also appears to be a promising treatment for tears using these glue types. Biomimetics One final, prominent field emerging in material science lies with biomimetic biomaterial approaches ( Chapman et al. , 2014 ). Biomimetic materials are materials that have been directly replicated from nature to produce a solution to a specific problem. Some synthetic polymers may be able to provide a more biomimetic environment than the previously discussed hydrogel approach. Functionalising hydrogels using chemistry is one strategy that requires future investigation. Hydrogels have the ability to create a more ‘native’ microenvironment for cells in a particular area of the body—i. e. , the knee. For example, scaffolds with biomimetics have been developed for tissue engineering based on a multidisciplinary approach using engineering of biomaterials and nano/micro structuring of the defect tissues. The use of 3D bioprinting is considered to be conventional however, the technique has allowed for traditional fabrication methods for porous bone and cartilage regeneration to be taken in new directions using gas forming, soluble particle leaching or freeze drying. Newer methods to generate porous scaffolds using biodegradable polymers include using gas forming of porogens (ammonium bicarbonate particles). Injectable hydrogels using click chemistry (high yielding, wide in scope molecules) have also shown to be highly advantageous for local delivery of bioactive molecules, ease of handing and reduced invasiveness, these techniques have been demonstrated to be potentially used in 3D bioprinting ( Jo, Kim & Noh, 2012 ). The use of hand held 3D matrix printing using a bio-pen has allowed for in-situ printing and repair to take place. This will be a major development in regenerative medicine ( Di Bella et al. , 2017 ). The most recent and emerging areas for biomimetic medical materials are ( Chen et al. , 2016 ): (1) 3D bioprinting (focussed on medical materials); (2) designing nano/micro technologies; (3) surface modification of biomaterials for their cellular interaction ability; (4) clinical aspects of biomaterials for cartilage focussing on cells, scaffolds and cytokines ( Fratzl & Weinkamer, 2007 ). The traditional methods still have many advantages ( Chen et al. , 2016 ), but as 3D printing techniques develop coupled with new developments in chemistry of the biomaterial, the use of biomimetic design and the inherent properties linked to biocompatibility will enable more advanced developments in the future of meniscal repair. Conclusions Evidently, the diversity of biomaterials for meniscal applications is immense. Many approaches to mimicking the structure and function of the ECM have been conceived. It is crucial that these advances continue to be investigated for their ability to interact within a biological system. As biomaterials advance and new methods of delivery develop, inclusive of minimal invasive surgery move forward—the field of meniscal tears and treatment will be greatly advanced and if not greatly reduced in the coming decade.

10. 7717/peerj. 4177

PeerJ

Murine pluripotent stem cells that escape differentiation inside teratomas maintain pluripotency

Background Pluripotent stem cells (PSCs) offer immense potential as a source for regenerative therapies. The teratoma assay is widely used in the field of stem cells and regenerative medicine, but the cell composition of teratoma is still elusive. Methods We utilized PSCs expressing enhanced green fluorescent protein (EGFP) under the control of the Pou5f1 promoter to study the persistence of potential pluripotent cells during teratoma formation in vivo. OCT4-MES (mouse embryonic stem cells) were isolated from the blastocysts of 3. 5-day OCT4-EGFP mice (transgenic mice express EGFP cDNA under the control of the Pou5f1 promoter) embryos, and TG iPS 1-7 (induced pluripotent stem cells) were generated from mouse embryonic fibroblasts (MEFs) from 13. 5-day OCT4-EGFP mice embryos by infecting them with a virus carrying OCT4, SOX2, KLF4 and c-MYC. These pluripotent cells were characterized according to their morphology and expression of pluripotency markers. Their differentiation ability was studied with in vivo teratoma formation assays. Further differences between pluripotent cells were examined by real-time quantitative PCR (qPCR). Results The results showed that several OCT4-expressing PSCs escaped differentiation inside of teratomas, and these escaped cells (MES-FT, GFP-positive cells separated from OCT4-MES-derived teratomas; and iPS-FT, GFP-positive cells obtained from teratomas formed by TG iPS 1-7) retained their pluripotency. Interestingly, a small number of GFP-positive cells in teratomas formed by MES-FT and iPS-FT (MES-ST, GFP-positive cells isolated from MES-FT-derived teratomas; iPS-ST, GFP-positive cells obtained from teratomas formed by iPS-FT) were still pluripotent, as shown by alkaline phosphatase (AP) staining, immunofluorescent staining and PCR. MES-FT, iPS-FT, MES-ST and iPS-ST cells also expressed several markers associated with germ cell formation, such as Dazl, Stella and S tra8. Conclusions In summary, a small number of PSCs escaped differentiation inside of teratomas, and these cells maintained pluripotency and partially developed towards germ cells. Both escaped PSCs and germ cells present a risk of tumor formation. Therefore, medical workers must be careful in preventing tumor formation when stem cells are used to treat specific diseases.

Introduction Pluripotent stem cells (PSCs), including embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs), have the potential to differentiate into all cell types of the body in vitro through embryoid body formation or in vivo through teratoma formation. Due to these characteristics, stem cells provide an option for treating a multitude of clinical problems, such as myocardium damage after heart infarction, spinal cord damage after mechanical injury, brain damage after stroke, age-related macular degeneration of the retina, liver damage, extensive skin burns, Parkinson’s disease, and diabetes ( Abdelalim et al. , 2014 ; Lodi, Iannitti & Palmieri, 2011 ; Orlic et al. , 2001 ; Ratajczak, Bujko & Wojakowski, 2016 ). When transplanted into immune-compromised mice, undifferentiated PSCs can form teratomas, consisting of multiple tissue types derived from all three germ layers ( Przyborski, 2005 ; Takahashi & Yamanaka, 2006 ). As such, there have been many efforts to differentiate pluripotent cells to cells with medical applications in an in vivo developmental environment. For example, neural stem cells (NSCs) have been differentiated in vivo through teratoma formation, and pure NSC populations exhibit properties similar to those of brain-derived NSCs ( Hong et al. , 2016 ). Similarly, fully functional and engraftable hematopoietic stem/progenitor cells (HSPCs), along with functional myeloid and lymphoid cells, have been isolated from teratomas when human iPSCs were transplanted into immunodeficient mice ( Amabile et al. , 2013 ; Suzuki et al. , 2013 ). In addition, the teratoma assay can be applied to assess the safety of human PSC-derived cell populations that are used for therapeutic application since a small number of undifferentiated cells contaminating a given transplant material can be efficiently detected by their multi-lineage differentiation ability ( Stachelscheid et al. , 2013 ). However, the intrinsic self-renewal and pluripotency qualities of PSCs that make them therapeutically promising are responsible for an equally fundamental tumorigenic risk ( Lee et al. , 2013 ). Studies on teratomas will contribute to a better understanding of their stepwise development processes and underlying molecular mechanisms and may provide helpful information for the development of tissue engineering technologies ( Aleckovic & Simon, 2008 ). These facts prompted us to address the additional characteristics of teratoma growth and differentiation after PSCs injection. In the present study, we aimed to isolate OCT4-expressing cells that escaped differentiation inside of growing teratomas and to determine whether OCT4-expressing cells still possess self-renewal and pluripotency abilities. Materials & Methods All animal experiments were approved by the Animal Care and Use Committees of the State Key Laboratories for Agrobiotechnology, College of Biological Sciences, China Agricultural University (Approval number: SKLAB-2016-05-01). Briefly, mice were bred in a 12/12 h light/dark period and sacrificed by cervical vertebra dislocation. Mouse strains OCT4-GFP transgenic mice (Model Animal Research Center of Nanjing University) express EGFP (enhanced green fluorescence protein) cDNA under the control of the Pou5f1 promoter, which is active in pluripotent stem cells. This strain is useful for isolating pluripotent stem cells, as they specifically express green fluorescent protein. These OCT4-GFP transgenic mice were the source of the OCT4-MES and OG2 MEFs (mouse embryonic fibroblasts of 13. 5-day OCT4-EGFP mice embryos) used in this study. Derivation of MES and generation of iPSCs To obtain OCT4-MES, uteri containing E3. 5 embryos were isolated from timed pregnancies and transferred individually to the wells of a 24-well plate with irradiated mouse embryonic fibroblast (MEF) feeders. After five days of incubation, embryo outgrowths were separated from trophectoderm, individually picked, and expanded in MES medium (Dulbecco’s modified eagle medium (DMEM) supplemented with 15% fetal bovine serum (FBS), L-glutamine, nonessential amino acids, β-mercaptoethanol, and 1, 000 U/ml leukemia inhibitory factor). OG2 MEFs were cultured in MEF medium (DMEM supplemented with 10% FBS, L-glutamine and nonessential amino acids); infected with retroviruses generated from pMX retroviral vectors encoding mouse Pou5f1, Sox2, Klf4 and c-Myc ; and cultured on irradiated MEF feeder cells in MES medium. Subsequently, a single ESC-like colony was individually picked and expanded on feeders to establish stable lines. Both OCT4-MES and iPSCs originated from male embryos. Additional details can be found in our previous study ( Pei et al. , 2015 ). Immunofluorescence Cells were fixed with 4% paraformaldehyde, permeabilized with 0. 1% Triton X-100, and blocked with 2% BSA. The cells were then stained with primary antibodies against OCT4 (ab19857, 1:500; Abcam, Cambridge, UK), SOX2 (ab97959, 1:1, 000; Abcam, Cambridge, UK), NANOG (ab80892, 1:500; Abcam, Cambridge, UK) and SSEA1 (ab16285, 1:200; Abcam, Cambridge, UK), followed by staining with the respective secondary antibodies conjugated to Alexa Fluor (A-11008, A-11037, A-21044, 1:1, 000; Invitrogen, Carlsbad, CA, USA). Finally, cells were counterstained with DAPI (D9542, Sigma, St. Louis, MO, USA). RNA purification and cDNA preparation Feeders were removed by plating ESCs on a gelatin-coated dish for 30 min, and unattached cells were collected by centrifugation. Total RNA was extracted from pure PSCs using Trizol reagent according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). RNA was reverse-transcribed using oligo-dT and M-MLV Reverse Transcriptase (Promega, Madison, WI, USA). Real-time quantitative PCR qPCR was performed on a LightCycler 480 II Real-Time PCR System (Roche, Basel, Switzerland) using the LightCycler 480 SYBR Green I Master Mix (Roche, Basel, Switzerland 4887352001). The qPCR data was analyzed using the comparative CT (2 −ΔΔ CT ) method as the description by Livak & Schmittgen (2001). The ΔCT was calculated using Gapdh, EF1α and β-tubulin as internal control. The primers used for qPCR and PCR are listed in Table 1. 10. 7717/peerj. 4177/table-1 Table 1 Sequence of primers used in this study. Gene Sequence (5′ − 3′) Gapdh Forward AGGTCGGTGTGAACGGATTTG Reverse TGTAGACCATGTAGTTGAGGTCA β-tubulin Forward TGAGGCCTCCTCTCACAAGTA Reverse CCGCACGACATCTAGGACTG EF1α Forward GTGTTGTGAAAACCACCGCT Reverse AGGAGCCCTTTCCCATCTCA Pou5f1 Forward GTTGGAGAAGGTGGAACCAA Reverse CTCCTTCTGCAGGGCTTTC Sox2 Forward AAGGGTTCTTGCTGGGTTTT Reverse AGACCACGAAAACGGTCTTG Utf1 Forward GTCCGGACCCTTCGATAACC Reverse CTCGGCCTCTTGCTCCAC Nanog Forward TTCTTGCTTACAAGGGTCTGC Reverse AGAGGAAGGGCGAGGAGA Rex1 Forward CAGTTCGTCCATCTAAAAAGGGAGG Reverse TCTTAGCTGCTTCCTTGAACAATGCC Tbx3 Forward ATCGCCGTTACTGCCTATCA Reverse TGCAGTGTGAGCTGCTTTCT Lin28a Forward GTCTTTGTGCACCAGAGCAAG Reverse ATGGATTCCAGACCCTTGGC Nr5a2 Forward TAGGACCGGAAAGCGTCTGC Reverse GCTTCCGTCTCCACTTTGGG Dazl Forward GCCCGCAAAAGAAGTCTGTG Reverse ACCAACAACCCCCTGAGATG Stella Forward GAGAAGACTTGTTCGGATTGAGC Reverse CATCGTCGACAGCCAGGG Stra8 Forward CTCCTCCTCCACTCTGTTGC Reverse GCGGCAGAGACAATAGGAAG Vasa Forward ACCAAGATCAGGGGACACAG Reverse TAACCACCTCGACCACTTCC Teratoma production and analysis Approximately 1×10 6 PSCs were suspended in 150 µl of PBS (phosphate buffered solution) and subcutaneously injected into the hind limb of NOD/SCID mice to form teratomas. Three weeks after injection, the teratomas were harvested, fixed overnight with 4% paraformaldehyde, embedded in paraffin, sectioned, HE stained or immunostained (primary antibodies against GFP, Cat. 2956, 1: 200 (Cell Signaling Technology, Danvers, MA, USA); Biotin-Streptavidin horseradish peroxidase detection kit, Cat. SP-9001 (Beijing Zhongshan Golden Bridge Biotechnology Company, Beijing, China)), and analyzed. Statistical analysis All results are presented as the mean ± standard deviation. Results were statistically analyzed using SAS (Statistics Analysis System) program. Significance of differences between samples was determined (at the significance level p < 0. 05) using Kruskal–Wallis test. Results Both OCT4-MES and TG iPS 1-7 are pluripotent OCT4-EGFP mice express green fluorescent protein under the control of the pluripotency-associated Pou5f1 promoter and are widely used to study the function of PSCs ( Pei et al. , 2015 ). These mice were used to generate mouse embryonic stem cells (MES) and iPSCs. OCT4-MES were isolated from the blastocysts of 3. 5-day OCT4-EGFP mice embryos, while other mice were selected to prepare MEFs after day 13. 5. The isolated MEFs were used to generate iPSCs by infecting them with a virus carrying OCT4, SOX2, KLF4 and c-MYC. Then, TG iPS 1-7 was selected from the isolated iPSC clones. Both OCT4-MES and TG iPS 1-7 were maintained on feeder cells in the presence of leukemia inhibitory factor. They both exhibited typical MES-like morphologies ( Figs. 1A and 1B ). Immunofluorescent staining confirmed the expression of the three master transcription factors (OCT4, NANOG and SOX2) as well as ESC-specific surface marker SSEA-1 in OCT4-MES and TG iPS 1-7 ( Figs. 1C – 1J ). The PCR results further demonstrated that these cells expressed pluripotency marker genes, including Pou5f1, Sox2, Nanog, Rex1, Tbx3, Nr5a2, Utf1 and Lin28a ( Fig. 2 ). Next, in vivo teratoma formation assays were performed to further validate the pluripotency of OCT4-MES and TG iPS 1-7. Approximately 1×10 6 PSCs were suspended in 150 µl of PBS and injected into non-obese diabetic/severe combined immunodeficient (NOD/SCID) mice to form teratomas. Three weeks after injection, OCT4-MES and TG iPS 1-7 formed teratomas in vivo, and hematoxylin and eosin (H&E) staining confirmed the formation of all three germ layers in each teratoma ( Figs. 3A – 3F ). These results revealed that OCT4-MES and TG iPS 1-7 were pluripotent. Interestingly, we observed OCT4-positive cells growing in clusters in the teratoma masses formed by OCT4-MES and TG iPS 1-7 ( Figs. 3G, 3H ). 10. 7717/peerj. 4177/fig-1 Figure 1 OCT4-MES and TG iPS 1-7 are pluripotent. (A and B) Phase contrast images of OCT4-MES (A) and TG iPS 1-7 (B). Both types of cells exhibit typical MES-like morphologies. (C–J) Immunofluorescent staining showing that both OCT4-MES (C–F) and TG iPS 1-7 (G–J) express the pluripotency markers OCT4 (C and G), NANOG (D and H), SOX2(E and I) and SSEA1 (F and J). 10. 7717/peerj. 4177/fig-2 Figure 2 OCT4-MES, MES-FT, MES-ST, TG iPS 1-7, iPS-FT and iPS-ST express pluripotency genes. Expression of pluripotency marker genes was evaluated by PCR, showing that OCT4-MES, MES-FT, MES-ST, TG iPS 1-7, iPS-FT and iPS-ST all express pluripotency marker genes, including Pou5f1, Sox2, Nanog, Rex1, Tbx3, Nr5a2, Utf1 and Lin28a. 10. 7717/peerj. 4177/fig-3 Figure 3 GFP-positive pluripotent cells are present in teratomas generated by OCT4-MES and TG iPS 1-7. (A–F) Hematoxylin and eosin staining of teratomas derived from OCT4-MES (A–C) and TG iPS 1-7 (D–F). Products of all three germ layers are seen in the image: Ectoderm: epidermis with keratin (A and D). Mesoderm: smooth muscle (A and E). Endoderm: gastrointestinal lining cells/glands (C and F). Specified cells are indicated by arrows. (G and H) Immunohistochemistry to detect the presence of GFP-positive pluripotent cells in teratomas generated by OCT4-MES (G) and TG iPS 1-7 (H). GFP-positive pluripotent cells were stained in brown with anti-GFP primary antibodies. OCT4-positive cells from OCT4-MES and TG iPS 1-7 teratomas have self-renewal and pluripotency qualities To quantify the fraction of OCT4-positive pluripotent cells in teratomas generated by OCT4-MES and TG iPS 1-7, we cut the teratomas into pieces and digested them with trypsin and then cultured the cells in MEF medium. Three days later, we found that most of these cells separated from OCT4-MES and that TG iPS 1-7-derived teratomas had the morphology of mouse embryonic cells, but a small number of cells were round and expressed GFP ( Figs. 4A, 4B, 4D and 4E ). After picking these cells and culturing them in MES medium, we found that they had typical MES-like morphologies, and they were AP-positive ( Figs. 4C, 4F ). We named these cells MES-FT and iPS-FT, which were derived from OCT4-MES and TG iPS 1-7, respectively. OCT4-expressing MES-FT and iPS-FT cells were grown in the presence of leukemia inhibitory factor, and they expressed pluripotency marker genes, including Pou5f1, Sox2, Nanog, Rex1, Tbx3, Nr5a2, Utf1 and Lin28a ( Fig. 2 ). The immunostaining results showed that these colonies were positive for OCT4, NANOG, SOX2 and SSEA-1 ( Figs. 5A – 5H ). We performed in vivo teratoma formation assays to further validate the pluripotency of MES-FT and iPS-FT. MES-FT and iPS-FT formed teratomas in vivo, and the hematoxylin and eosin (H&E) staining results confirmed the formation of all three germ layers in each teratoma ( Figs. 5I – 5N ). As in the results described above, there were also OCT4-positive pluripotent cells in the teratomas formed by MES-FT and iPS-FT ( Figs. 5O, 5P ). 10. 7717/peerj. 4177/fig-4 Figure 4 OCT4-positive pluripotent cells isolated from teratomas have typical mouse embryonic cell morphology. (A, B, D and E) A small number of OCT4-GFP positive cells were found among teratoma cells generated by OCT4-MES (A and B) and TG iPS 1-7 (D and F) cultured in MEF medium. (C and F) MES-FT (C) and iPS-FT (F) have typical mouse embryonic cell morphology and are AP-positive when cultured in MES medium. 10. 7717/peerj. 4177/fig-5 Figure 5 MES-FT and iPS-FT own self-renewal and pluripotency ability. (A–H) Immunofluorescent staining of pluripotency markers OCT4, NANOG, SSEA1 and SOX2 in MES-FT (A–D) and iPS-FT (E–H). Both types of cells expressed all four markers. (I–N) Hematoxylin and eosin staining of teratomas derived from MES-FT (I–K) and iPS-FT (L–N). Products of all three germ layers are seen in the image: Ectoderm: epidermis with keratin (I and L). Mesoderm: smooth muscle (J and M). Endoderm: gastrointestinal lining cells/glands (K and N). Specified cells are indicated by arrows. (O and P) Immunohistochemistry to detect the presence of GFP-positive pluripotent cells in teratomas generated by MES-FT (O) and iPS-FT (P). GFP-positive pluripotent cells were stained in brown with anti-GFP primary antibodies. OCT4-positive cells from MES-FT and iPS-FT teratomas are still pluripotent We discovered several round and bright cells expressing OCT4-GFP under a microscope in cells separated from teratomas formed by MES-FT and iPS-FT cells ( Figs. 6A, 6B, 6D and 6E ). These round and bright cells formed AP positive clones ( Figs. 6C, 6F ). We named these cells MES-ST and iPS-ST. OCT4-expressing MES-ST and iPS-ST cells also expressed pluripotency marker genes, including Pou5f1, Sox2, Nanog, Rex1, Tbx3, Nr5a2, Utf1 and Lin28a ( Fig. 2 ), and the immunostaining results demonstrated that they expressed the stemness regulators OCT4, NANOG, SOX2 and SSEA-1 ( Figs. 6G – 6N ). 10. 7717/peerj. 4177/fig-6 Figure 6 MES-ST and iPS-ST owned PSCs characteristic. (A, B, D and E) A small number GFP-positive cells were found among teratoma cells generated by MES-FT (A and B) and iPS-FT (D and E) cultured in MEF medium. (C and F) MES-ST (C) and iPS-ST (F) have typical mouse embryonic cell morphology and are AP-positive when cultured in MES medium. (G–K) Immunofluorescent staining of pluripotency markers OCT4, NANOG, SSEA1 and SOX2 in MES-ST (G–J) and iPS-ST (K–N). Both MES-ST (G–J) and iPS-ST (K–N) expressed all four markers. The above results showed that OCT4-MES, TG iPS 1-7, MES-ST, iPS-ST, MES-ST and iPS-ST had pluripotency characteristics. However, MES-FT, iPS-FT, MES-ST and iPS-ST were survivors of the differentiation environment, so we wanted to know whether there were differences among these cells. Thus, we next investigated their differences. OCT4-positive cells separated from teratomas express germ cell marker genes To explore the gene expression patterns of OCT4-MES, TG iPS 1-7, MES-FT, iPS-FT, MES-ST and iPS-ST, cDNA was prepared from these cells without feeders for gene expression analysis. First, we detected the expression of pluripotency genes. When normalized to the values for OCT4-MES cells, the expression level of Pou5f1 was higher in MES-FT, and that of Lin28a was higher in both MES-FT and MES-ST cells, but there were no differences in the Nanog expression levels between these three cell lines ( Fig. 7A ). When normalized to the values for TG iPS 1-7 cells, iPS-FT and iPS-ST both highly expressed Pou5f1 and Nanog ( Fig. 7B ). However, there were no differences in the expression level of Lin28a ( Fig. 7B ). The expression of pluripotency marker genes in these cells varied slightly, but they were all within reasonable levels. Thus, these cell types were all pluripotent. 10. 7717/peerj. 4177/fig-7 Figure 7 OCT4-MES, TG iPS 1-7, MES-FT, iPS-FT, MES-ST and iPS-ST express pluripotency genes; MES-FT, iPS-FT, MES-ST and iPS-ST more highly express several markers associated with germ cell formation. (A) The expression levels of Pou5f1, Nanog and Lin28a in OCT4-MES, MES-FT and MES-ST were determined by qPCR. Both MES-FT and MES-ST highly expressed Lin28a, and MES-FT also highly expressed Pou5f1. (B) The expression levels of pou5f1, Nanog and Lin28a in TG iPS 1-7, iPS-FT and iPS-ST were determined by qPCR. iPS-FT and iPS-ST highly expressed pou5f1 and Lin28a. (C) The expression levels of Dazl, Stella, Stra8 and Vasa in Oct4-MES, MES-FT and MES-ST were determined by qPCR. Both MES-FT and MES-ST highly expressed Stra8. MES-ST also highly expressed Dazl and Stella. (D) The expression levels of Dazl, Stella, Stra8 and Vasa in TG iPS 1-7, iPS-FT and iPS-ST were determined by qPCR. iPS-FT and iPS-ST highly expressed Dazl, Stra8 and Vasa. iPS-FT also highly expressed Stella. Relative expression was quantified using the comparative threshold cycle (Ct) method (2 −ΔΔ Ct ). n = 3, Gapdh, EF1α and β-tubulin were used as references. *, p < 0. 05. Previous results have shown that PSCs that escape from differentiation inside of embryonic bodies express several markers associated with germ cell formation ( Attia et al. , 2014 ). As such, we further assayed the differences between the expression levels of important germ cell-specific genes ( Dazl, Stella, Stra8, Vasa ) in MES-FT, iPS-FT, MES-ST and iPS-ST ( Figs. 7C, 7D ). When normalized to the values for OCT4-MES, Dazl and Stella were more highly expressed in MES-FT cells, and the expression level of S tra8 was elevated nine-fold and ten-fold in MES-FT and MES-ST, respectively. Similarly, iPS-FT and iPS-ST highly expressed Dazl, Stra8 and Vasa than TG iPS 1-7. iPS-FT also highly expressed Stella. The above results show that OCT4-positive cells separated from teratomas have elevated expression of several markers associated with germ cell formation, such as Dazl, Stella and S tra8. Discussion ESCs and iPSCs are characterized by their ability to develop into any cell type of the adult organism. As such, they can be widely applied to the treatment of many diseases. This is especially true for iPSCs, as they do not present ethical issues. A previous report demonstrated the presence of undifferentiated human ESCs expressing the surface marker CD133 ( Ritner & Bernstein, 2010 ). However, no additional research has been performed to investigate the characteristics of those undifferentiated cells in teratomas. Therefore, in this study, we isolated OCT4-GFP positive cells, MES-FT and iPS-FT, from teratomas generated by OCT4-MES and TG iPS 1-7, respectively. MES-FT and iPS-FT exhibit MES-like morphologies, express pluripotency marker genes and proteins, and can generate all three germ layers in an in vivo differentiation model. We discovered that there were still pluripotent cells in the teratomas formed by MES-FT and iPS-FT, so we separated them from the teratoma mass and named them MES-ST and iPS-ST. Further study confirmed that these isolated cells (MES-ST and iPS-ST) retained pluripotency and were capable of differentiation. From these results, it can be inferred that a subset of PSCs escape differentiation during in vivo differentiation, and the escaped cells retain their PSC characteristics in the appropriate environment. Since the escaped PSCs (MES-FT, iPS-FT, MES-ST and iPS-ST) still possessed PSC-like characteristics, these cells may progress to tumor formation at an undefined later time point. Bottai et al. (2010) reported that they used 5 × 10 5 undifferentiated murine ESCs to cure spinal cord injury. However, some of the transplanted ESCs were found as dense aggregates in the tissue. This result supports our view that ESCs can be maintained in vivo. Another study showed that transplantation of 1 − 2 × 10 6 MES cells into SV129 mice led to tumor formation in 100% of cases, whereas transplantation of 5 × 10 5 cells produced tumors in two of six mice and transplantation of 1 × 10 5 ESCs gave rise to tumor formation in one of six transplanted mice within 100 days ( Dressel et al. , 2008 ). It can be deduced that there is likely a niche within teratomas that nurse PSCs, and the number of cells determines the niche environment. The more PSCs used for transplantation, the higher probability of tumor formation. The escaped PSCs (MES-FT, iPS-FT, MES-ST and iPS-ST) showed slight similarities to primordial germ cells (PGCs), as shown by the high expression of Pou5f1, Dazl, Stella and Stra8 in MES-FT, MES-ST, iPS-FT, and iPS-ST. Pou5f1, Dazl, Stella, Stra8 and Vasa are well-known germ cell markers, and they are also commonly expressed in ESCs ( Cauffman et al. , 2005 ; Kehler et al. , 2004 ; Tedesco et al. , 2009 ; Toyooka et al. , 2000 ; Wongtrakoongate et al. , 2013 ). Stra8 is required for the chromosomal program of meiotic prophase ( Soh et al. , 2015 ). Dazl, an intrinsic meiotic competence factor, is required for Stra8 -mediated initiation of meiosis in germ cells ( Lin et al. , 2008 ). Overexpression of Stra8 and Dazl genes promotes the transdifferentiation of mesenchymal stem cells and ESCs in vitro toward PGCs ( Li et al. , 2017 ; Shi et al. , 2014 ). The elevated expression of Pou5f1, Dazl, Stella and Stra8 might indicate that the GFP positive cells separated from teratomas partially develop towards germ cells. This suggests that it is possible to isolate PGCs from teratoma differentiation models. Conclusions In summary, we found a small number of OCT4-expressing PSCs that escaped differentiation inside teratomas. The escaped cells kept their unique properties of self-renewal and pluripotency and were able to form teratomas in vivo. They also expressed several markers associated with germ cell formation, such as Pou5f1, Dazl, Stella and Stra8, suggesting that these cells may partially differentiate into germ cells. Therefore, this study serves as a warning that medical workers using stem cells to treat specific diseases must pay careful attention to prevent tumor formation because OCT4-expressing cells retain pluripotency, and it is feasible to isolate germ cells from teratomas. This study of PSCs that remain undifferentiated within teratomas has provided critical information for further investigation of the applications of stem cell therapy and for obtaining germ cells from in vivo differentiation models. Supplemental Information 10. 7717/peerj. 4177/supp-1 Data S1 Realtime PCR raw data Click here for additional data file.

10. 7717/peerj. 4306

PeerJ

Establishment of an immortalized mouse dermal papilla cell strain with optimized culture strategy

Dermal papilla (DP) plays important roles in hair follicle regeneration. Long-term culture of mouse DP cells can provide enough cells for research and application of DP cells. We optimized the culture strategy for DP cells from three dimensions: stepwise dissection, collagen I coating, and optimized culture medium. Based on the optimized culture strategy, we immortalized primary DP cells with SV40 large T antigen, and established several immortalized DP cell strains. By comparing molecular expression and morphologic characteristics with primary DP cells, we found one cell strain named iDP6 was similar with primary DP cells. Further identifications illustrate that iDP6 expresses FGF7 and α-SMA, and has activity of alkaline phosphatase. During the process of characterization of immortalized DP cell strains, we also found that cells in DP were heterogeneous. We successfully optimized culture strategy for DP cells, and established an immortalized DP cell strain suitable for research and application of DP cells.

Introduction Hair follicles have the characteristic of periodical growth, which provides a nice model for the research of tissue regeneration. Dermal papilla (DP) cells have contact with hair follicle stem cells regularly and may play important roles in the regeneration of hair follicle ( Su et al. , 2017 ; Woo et al. , 2017 ). The signals from DP may regulate the regeneration of hair follicles and melanocyte ( Guo et al. , 2016 ; Li et al. , 2013 ). Dissociated human DP cells induce hair follicle neogenesis in grafted dermal-epidermal composites ( Thangapazham et al. , 2014 ). The limitation for DP research lies in the difficulty for culture of DP cells ( Morgan, 2014 ). As so far, the human intact dermal papilla transcriptional signature can be partially restored by growth of papilla cells in 3D spheroid cultures ( Topouzi et al. , 2017 ). When the culture environment was changed into 2D environment, very rapid and profound molecular signature changes were discovered ( Higgins et al. , 2013 ; Lin et al. , 2016 ). The isolation method of DP by surgical microdissection has been established in mouse vibrissae follicles and in human hair follicles ( Gledhill, Gardner & Jahoda, 2013 ), but the isolated DP cells can not be long-term cultured. Since the isolation of primary DP cells is time-consuming and has limited population doubling. There are also several inter-individual and intra-individual variations. It is necessary to establish stable DP cell lines to investigate hair biology. Immortalized DP cell lines of human have been established, and had hair growth promoting effects ( Shin et al. , 2011 ; Won et al. , 2010 ). In rodent animal models, immortalized rat DP cells have already been obtained ( Kang et al. , 2015 ). However, an effective immortalized mouse DP cell line is to be constructed. The goals of this project are optimize the isolation and culture condition of DP from mouse skin and establish an immortalized DP cell line for future research. Materials and Methods Isolation and culture of DP cells C57BL/6 mice were obtained from and housed in the laboratory animal center of the Army Medical University, Chongqing, China. All the animal-related procedures were conducted in strict accordance with the approved institutional animal care and maintenance protocols. All experimental protocols were approved by the Laboratory Animal Welfare and Ethics Committee of the Army Medical University. Permission number for producing animals: SCXK-PLA-20120011. Permission number for using animals: SYXK-PLA-20120031. A 9-day old C57BL/6 mouse was sacrificed according to standard protocol. The vibrissa pads were cut off bilaterally with an iris scissor in a 100-mm plate. Vibrissa pads were rinsed with PBS, and then hair follicles were dissected together with their connective tissue sheath using 27G syringe needles under dissecting microscope. The dissected hair follicles were rinsed with PBS and incubated with 0. 25% dispase for 20 min at room temperature. Dissected hair follicles were transferred into a new 100-mm plate and thoroughly washed with PBS. A horizontal cut directly above dermal papilla was made. After that, dermal papilla was dissected out of dermal sheath using 27G syringe needles under dissecting microscope. Then the dissected DP tissues were transferred into a 10 μg/cm 2 collagen I coated 24-well plate. DP media were added after 30 min incubation in 37 °C. DP cells presented at about 3 days later. Cells reach confluence after 2 weeks and were passaged onto collagen I coated plates. DP medium should include α-MEM (Gibco, Waltham, MA, USA), 10% FBS (Gibco, USA), 1 × sodium pyruvate (Gibco, USA), 1 × non-essential amino acid (Gibco, USA), 1 × penicillin-streptomycin, 10 ng/ml bFGF (PeproTech, Rocky Hill, NJ, USA). During the optimization process, the classical DP medium was used as control. The control medium is consisted of DMEM (Gibco, USA), 10% FBS (Gibco, USA), 1 × penicillin-streptomycin. Another control medium was the classical DP medium with the addition of 10 ng/ml bFGF (PeproTech, USA). Establishment of immortalized DP cell line Retrovirus with SV40 large T antigen which was flanked with FRT sites was prepared as formerly reported ( Yang et al. , 2012 ). Primary DP cells were plated in a 60 mm-dish at 50% confluency in the morning. After attachment, polybrene (final concentration 10 μg/mL) and retrovirus (3. 0 × 10 7 TU) were added together. The second day, the supernatant was aspirated out of dish, and new DP medium was refilled. At the same time, hygromycin was added at a final concentration of 200 μg/mL. The culture medium was changed every two days until all the cells in control group died. Antibiotic-selected DP cells were diluted with DP medium to 1–2 cells per 100 μL, and 100 μL diluted cells were added into every well of 96-well plates. Wells with only one cell were labeled and were monitored every 2 days. Cells in the labeled wells were passaged when the cell number of clones reached 20 or more. RT-PCR Total RNA of immortalized DP cells were extracted with Eastep™ super total RNA extraction kit (Promega, Beijing, China) according to manufacturer’s protocol. Complementary DNA was synthesized from RNA using Rever Tre Ace cDNA synthesis Kit (Toyobo, Osaka, Japan) according to the manufacturer’s protocol. Several gene expressions were determined by PCR machine (Bio-Rad, Hercules, CA, USA) with the synthesized cDNA as template. The primers used were shown in Table 1. PCR mastermix (Novoprotein, Shanghai, China) were used when amplifying. The reannealing temperatures (Tm) and product size for the primers were also shown in Table 1. 10. 7717/peerj. 4306/table-1 Table 1 The information of primers in RT-PCR experiment. Primers Sequence Tm (°C) Product size Noggin-f 5′ AGCACCCAGCGACAACCT 3′ 61 343 bp Noggin-r 5′ CAGCCACATCTGTAACTTCCTC 3′ Tbx18-f 5′ GCTGCTAACCAGACCCAC 3′ 58 537 bp Tbx18-r 5′ GTCCATGTCGCCAATACTC 3′ Bmp6-f 5′ TGCCTTAAACCACGAACAA 3′ 58 345 bp Bmp6-r 5′ GCTGGGAATGGAACCTGAA 3′ Fgf7-f 5′ AGCGGAGGGGAAATGTTCG 3′ 61 238 bp Fgf7-r 5′ TCCAGCCTTTCTTGGTTACTGAGA 3′ Sostdc1-f 5′ CCCCCATCCCAGTCATTTCTT 3′ 58 308 bp Sostdc1-r 5′ CAGGGGGATAATTTCACACTGAGA 3′ Sox2-f 5′ AAAACCGTGATGCCGACTA 3′ 58 431 bp Sox2-r 5′ ATCCGAATAAACTCCTTCCTTG 3′ SMA-f 5′ AGGGAGTAATGGTTGGAATGG 3′ 58 351 bp SMA-r 5′ CATCTCCAGAGTCCAGCACAA 3′ Gapdh-f 5′ ACCACAGTCCATGCCATCAC 3′ 52 450 bp Gapdh-r 5′ TCCACCACCCTGTTGCTGTA 3′ Immunocytochemistry staining Cover slides were placed on a 24-well plate, and cells were plated on cover slides. Twenty-four hours later, cover slides were rinsed with PBS and fixed with acetone. Then, the cover slides were rinsed with PBS and incubated with 5% goat serum in PBS at room temperature for 1 h. After that, slides were incubated with a rabbit anti-FGF7 antibody (1:100; Boster, Wuhan, China) or a rabbit anti-α-SMA antibody (1:200; Bioss, Beijing, China) at 4 °C overnight and subsequently with appropriate secondary antibodies (1:500; ZSGB-bio, Beijing, China). The slides were counterstained with DAPI (1:10, 000; Beyotime, Shanghai, China) for 10 min. At last, the cover slides were moved to microscope slides, mounted with antifade mounting medium (Beyotime, Shanghai, China), and observed under fluorescent microscope. The immunostaining experiments were repeated three times. Alkaline phosphatase staining Cover slides were placed on a 24-well plate, and cells were plated on cover slides. Twenty-four hours later, cover slides were rinsed with PBS and fixed with in situ fixation solution (Beyotime, Shanghai, China) for 10 min. Then the cover slides were rinsed with PBS five times. Fresh made NBT/BCIP staining buffer (Beyotime, Shanghai, China) or BM purple (Roche, Indianapolis, IN, USA) were added into the wells. The plate was covered with aluminium foil in the dark. Color change was monitored every 15 min to avoid non-specific staining. After the colour change appeared, the staining solution was aspirated out and the cells were washed twice with 1 × PBS. At last, the cover slides were dehydrated, cleared, moved to microscope slides, mounted with permount (ZSGB-bio, Beijing, China), and observed under microscope. The AP staining experiments were performed twice. Detection of immortalization Primary DP cells and iDP6 cells were cultured. The iDP6 cells were treated with AdGFP (adenovirus with the ability to express GFP protein), AdFlip (adenovirus with the ability to express flip recombinase, which can interact with FRT thus remove the expression of SV40) or PBS. Forty-eight hours later, cells were collected and total proteins were extracted with RIPA lysis buffer (Beyotime, Shanghai, China). Then, total proteins were loaded to 1% SDS-PAGE gel (Beyotime, China) and transmitted to PVDF membrane (Bio-Rad, Hercules, CA, USA). The PVDF membrane were incubated with anti-SV40 (1:1, 000; Santa Cruz Biotechnology, Dallas, TX, USA) and anti-GAPDH (1:500; ZSGB-bio, Beijing, China) antibodies. HRP labelled secondary antibodies were used, and the results were observed under ChemiDoc™ Touch Imaging System (Bio-Rad, Hercules, CA, USA). The experiment on reversing immortalization was performed twice. Results DP cells can be long-term cultured with the optimized strategy We optimized the culture strategy for DP cells from three dimensions, plate coating, dissecting method, and culture media ( Fig. 1 ). The optimized dissecting method worked well in obtaining primary DP cells. DP cells grew better on plate coated with collagen I than on uncoated plate. The morphology of DP cells did not have any significant difference between classical DP culture medium (DMEM with 10% FBS) and classical DP culture medium with the addition of bFGF (data not shown). Compared with classical DP culture medium, primary DP cells grew better in the optimized culture medium ( Figs. 2A – 2D ). The morphology of passaged DP cells was much more resemble in primary DP cells in the optimized culture medium. The cultured DP cells still had the characteristics of agglutinative growth in the optimized culture medium, but not in the control medium ( Figs. 2E – 2H ). 10. 7717/peerj. 4306/fig-1 Figure 1 Optimized strategy for the isolation and culture of DP cells. At first, the whole skin of vibrissa area was cut, then the DP tissue was separated from the skin together with vibrissa pad, and then the DP tissue was collected after dispase digestion. After that, the collected DP tissue was cultured with our optimized culture medium in collagen I-coated plate. 10. 7717/peerj. 4306/fig-2 Figure 2 Optimization of culture media for DP cells. Cells in (A, C, E, G) are cultured in DMEM culture medium with 10% FBS, cells in (B, D, F, H) are cultured in optimized culture medium. (A)–(D) are primary DP cells. (A) and (B) are 2 days after culture; (C) and (D) are 4 days after culture. (E)–(H) are DP cells after one generation of passage. (E) and (F) are 2 days after passage; (G) and (H) are 4 days after passage (100×). Scale bar = 100 μm. DP cells are heterogeneous Primary DP cells were immortalized by SV40 system. DP cells before antibiotic-selection were named with 0#. After antibiotic-selection, DP cell strains were selected by infinite dilution method. Not every single cell grew to clone at last. Cell strains were named with the time sequence when they grew to clone beginning with just one single cell. Totally 19 cell strains survived at last, named with iDP1 to iDP19 (1#–19#). The morphologic characteristics of the selected cell strains were different from each other ( Fig. 3 ). Some cells still look like fibroblast, whereas some cells changed into epithelial-like cells ( Fig. 3G ). iDP6 still had the characteristic of agglutinative growth, while others lost this characteristic. Specially, iDP10 grew clonally, which implied that the cell line was more primitive. For these cell strains, the expression patterns of the markers for DP cells were also determined by RT-PCR, including FGF7, BMP6, Sox2, Tbx18, Sostdc, α-SMA and noggin ( Fig. 4 ). All these data indicate that the cell strains were totally different from each other. Since each cell line grew from one single DP cell, the DP cells from one DP tissue were heterogeneous. 10. 7717/peerj. 4306/fig-3 Figure 3 Morphology of immortalized cell strains. (A)–(L) represent cell strains named with 0#, 3#, 4#, 5#, 6#, 9#, 10#, 11#, 12#, 14#, 17#, 19#. Scale bar = 100 μm. 10. 7717/peerj. 4306/fig-4 Figure 4 Expression pattern of the iDP6. The expression of several known DP markers were determined by RT-PCR. GAPDH was used as internal control. Each lane represents a DP cell strain. pDP, primary DP cells. iDP6 keeps the molecular characteristics of primary DP cells Taken morphology and mRNA expression characteristics together, iDP6 is the one that most similar to the primary DP cells. To determine whether iDP6 can be used in DP research, the activity of alkaline phosphatase was determined by AP staining, and the expression of FGF7 and α-SMA were determined by immunocytochemistry as well. At protein level, just like in situ, some iDP6 cells still had high AP activity ( Figs. 5A – 5F ). FGF7 was expressed in the cytoplasm of all iDP6 cells ( Figs. 5G – 5L ). α-SMA was expressed in both the cytoplasm and the nucleus of all iDP6 cells ( Figs. 5M – 5R ). Although the AP activity in iDP6 was lower than primary DP cells, the expression patterns of FGF7 and α-SMA was similar to primary DP cells. 10. 7717/peerj. 4306/fig-5 Figure 5 Characterization of the iDP6. (A)–(C), (G)–(I), (M)–(N) Characterization of iDP6. (D)–(F), (J)–(L), (P)–(R) Characterization of primary DP cells. (A)–(F) Alkaline phosphatase staining. (G)–(R) Immunocytochemistry, red color denotes positive expression, blue color denotes the counterstaining of DAPI. (G)–(R) The right panels are the merge of the left two panels. Arrowheads denote the positive expression. Scales bars for (A) and (D) are 100 μm. Scale bars for (B), (E), and (M)–(R) are 50 μm. Scale bars for (C), (F) and (G)–(L) are 25 μm. The immortal process of DP cells is reversible At first, the expression of SV40 were determined to make sure that iDP6 were immortalized. Western blot showed that iDP6 cells expressed SV40, while primary DP cells did not express SV40 ( Fig. 6A ). Then, AdFlip was used to remove the expression of SV40 in iDP6 cells. AdGFP and PBS were used as control. Results showed that compared with control groups, the expression of SV40 decreased at 48 h after being treated with AdFlip ( Fig. 6B ). These results demonstrated that iDP6 was successfully immortalized and the immortal process was reversible. 10. 7717/peerj. 4306/fig-6 Figure 6 Reversible immortalization of DP cells. The expressions of SV40 large T antigen were determined by western blot. (A) The expression of SV40 in iDP6 and primary DP. (B) The expression of SV40 in adenovirus treated iDP6. DP, primary dermal papilla cells; iDP6, immortalized DP cells #6; AdGFP, AdFlip; PBS, iDP6 cells treated with AdGFP, AdFlip or PBS. Proteins were collected at 48 h after treatment. Discussion Primary cell culture needs to simulate the in vivo environment of the cells. In anagen, DP cells reside in the center of hair bulb. They are circumstanced with a single layer of dermal cells. Usually, DP cells periodically interact with epithelial cells outside of the single layer of dermal cells. In telogen, DP tissue is a little far from hair follicle stem cells (HFSCs). In anagen onset, it begins to move close to HFSCs, and keeps interaction with HFSCs during early anagen. In the late anagen, it begins to move away from HFSCs. In catagen, it keeps away from HFSCs. Thus, the environment for DP cells in vivo varies with hair cycle ( Bassino et al. , 2015 ). In addition, exogenous connective tissue may also impact the function of DP cells ( Zhang et al. , 2014b ). To exclude contamination from other mesenchymal cells, epithelial cells, and adipose tissue, the use of microdissection techniques is preferred ( Zhang et al. , 2014a ). To culture DP cells in vitro, all the conditions should be taken into consideration. As the main component of connective tissue, collagen is mostly secreted by fibroblast. Collagen is widely used in tissue engineering and cell culture. We coated the plate with collagen, and found it was good for the growth of DP cells. DP is relatively independent in the anagen hair follicle. However, it is too small to be isolated quickly. So we used a stepwise method to isolate DP. DP cells grew from the isolated DP. The most important condition for cell culture is the culture medium. Since the classical DMEM with 10% FBS can not long-term culture mouse DP cells, we seek to find culture medium for special fibroblast. We found that a mesenchymal stem cell culture medium α-MEM worked well. Additionally, bFGF is a critical component of human embryonic stem cell culture medium. In conjunction with BMP4, bFGF promotes differentiation of stem cells to mesodermal lineages ( Yuan et al. , 2013 ). DP originates from mesodermal, so we added bFGF to the medium. However, since the classical medium and the classical medium with the addition of bFGF did not have significant difference on the culture of primary DP, the main effective additions in the optimized medium maybe sodium pyruvate and non-essential amino acids. Based on these data, the optimized strategy works well in isolating and long-term culture of DP cells. We are the first to use this strategy to culture DP cells. There are several molecules reported to be expressed in DP cells, including Sox2, Tbx18, Sostdc, α-SMA and noggin ( Weber & Chuong, 2013 ). To characterize immortalized DP cells, all the markers were tested. Recently, we found that two secretive proteins, FGF7 and BMP6 were also expressed in DP cells in vivo. FGF signaling was reported to regulate the size of dermal papilla ( Yue et al. , 2012 ), and BMP7 was reported to attenuate fibroblast-like differentiation of DP cells ( Bin et al. , 2013 ). Thus the expressions of FGF7 and BMP6 are also tested. Both the expressions of markers and morphology indicate that the immortalized DP cell strains are heterogeneous and iDP6 is a good cell strain to represent primary DP cells. It is reported that human DP cells have stem cell-like phenotypes ( Kiratipaiboon, Tengamnuay & Chanvorachote, 2016 ), neural crest stem cell-like cells were also isolated from rat vibrissa DP ( Li et al. , 2014 ), dermal stem cells also lies in mouse dermal sheath ( Rahmani et al. , 2014 ). So it is reasonable that DP cells are heterogeneous. But exactly how many kinds of cells are in DP remain to be discovered ( Yang et al. , 2017 ). Single cell assay technologies may help. Conclusions From the results of present study, it can be concluded that we optimized the dissection and culture of mouse DP cells from three dimensions: stepwised dissection, collagen I coated plate and α-MEM based culture medium. Based on the optimized strategies, we successfully immortalized the cultured primary DP cells with addition of SV40 large T antigen. We successfully selected several cell strains, characterized them, and found iDP6 cell strain similar to primary DP cells. In addition, the SV40 large T antigen in iDP6 can be removed by the addition of AdFlip. In a word, we establised an immortalized DP cell strain that can be used in future research.

10. 7717/peerj. 4656

PeerJ

Behavior and biocompatibility of rabbit bone marrow mesenchymal stem cells with bacterial cellulose membrane

Background Tissue engineering has been shown to exhibit great potential for the creation of biomaterials capable of developing into functional tissues. Cellular expansion and integration depends on the quality and surface-determinant factors of the scaffold, which are required for successful biological implants. The objective of this research was to characterize and evaluate the in vitro characteristics of rabbit bone marrow mesenchymal stem cells (BM-MSCs) associated with a bacterial cellulose membrane (BCM). We assessed the adhesion, expansion, and integration of the biomaterial as well as its ability to induce macrophage activation. Finally, we evaluated the cytotoxicity and toxicity of the BCM. Methods Samples of rabbit bone marrow were collected. Mesenchymal stem cells were isolated from medullary aspirates to establish fibroblast colony-forming unit assay. Osteogenic, chondrogenic, and adipogenic differentiation was performed. Integration with the BCM was assessed by scanning electron microscopy at 1, 7, and 14 days. Cytotoxicity was assessed via the production of nitric oxide, and BCM toxicity was assessed with the MTT assay; phagocytic activity was also determined. Results The fibroblastoid colony-forming unit (CFU-F) assay showed cells with a fibroblastoid morphology organized into colonies, and distributed across the culture area surface. In the growth curve, two distinct phases, lag and log phase, were observed at 15 days. Multipotentiality of the cells was evident after induction of osteogenic, chondrogenic, and adipogenic lineages. Regarding the BM-MSCs’ bioelectrical integration with the BCM, BM-MSCs were anchored in the BCM in the first 24 h. On day 7 of culture, the cytoplasm was scattered, and on day 14, the cells were fully integrated with the biomaterial. We also observed significant macrophage activation; analysis of the MTT assay and the concentration of nitric oxide revealed no cytotoxicity of the biomaterial. Conclusion The BCM allowed the expansion and biointegration of bone marrow progenitor cells with a stable cytotoxic profile, thus presenting itself as a biomaterial with potential for tissue engineering.

Introduction Researchers have been studying bone marrow mesenchymal stem cells (BM-MSCs) for their applicability in regenerative medicine, and for improving current methodologies ( DiMarino, Caplan & Bonfield, 2013 ; Wei et al. , 2013 ; Kobolak et al. , 2016 ; Li et al. , 2016 ). BM-MSCs are widely used in clinical and therapeutic use due to several factors: they are easily accessible; it is possible to achieve the necessary volume of cells in a short time, through culture replication; they allow autologous use or the treatment of several patients with a single sample, since the expression of HLA antigens is poor; they can be used without the need for HLA typing, making them ready for use in any patient. Even after being frozen, they preserve their characteristics, which allows the creation of bio-banks ( Wabik & Jones, 2015 ). The use of mesenchymal stem cells (MSCs) has shown promise in the field of regenerative medicine. Studies have investigated the use of MSCs in cardiovascular events ( Castellanos et al. , 2016 ), immunological dysfunctions ( Kaplan, Youd & Lodie, 2011 ; Zhao, Ren & Han, 2016 ), bone repair ( Emmet et al. , 2016 ), cartilaginous and intervertebral discs ( Blanquer, Grijpma & Poot, 2015 ), tendinosis ( Peach et al. , 2017 ), and hematological malignancies ( Wang, Qu & Zhao, 2012 ), among others ( Schnitzler et al. , 2016 ; Squillaro, Peluso & Galderisi, 2016 ). Tissue engineering is a promising multidisciplinary field that involves the development of materials or devices capable of specific interactions within biological tissues ( Langer & Vacanti, 2016 ). Advances in research have demonstrated biocompatibility between stem cells and biopolymers in the development of in vitro tissues capable of repairing injured areas ( Lima et al. , 2017 ; Park et al. , 2017 ; Weinstein-Oppenheimer et al. , 2017 ). Several biomaterials with different physicochemical and mechanical properties have been developed, with biomedical purposes including tissue regeneration, drug delivery systems, new vascular grafts, or in vitro and in vivo tissue engineering supports ( Lin et al. , 2013 ; Xi et al. , 2013 ; Soheilmoghaddam et al. , 2014 ; Zulkifli et al. , 2014 ; Kim & Kim, 2015 ; Pires, Bierhalz & Moraes, 2015 ; Urbina et al. , 2016 ). The scaffold surface can generate cellular responses which can affect adhesion, proliferation, migration, biointegration, and cellular function ( Abbott & Kaplan, 2016 ). This interaction is especially important to define the degree of rejection of medical implants ( Achatz et al. , 2016 ). Bacterial cellulose is an extracellular polysaccharide secreted primarily by Gluconacetobacter xylinus, an aerobic, Gram-negative, and chemoheterotrophic bacterium that can be grown in liquid medium from various sources of carbon and nitrogen, and basically uses glucose as the substrate. In culture medium, this microorganism produces very fine fibers that intertwine, forming a film with a nanofibrillar structure ( Moosavi-Nasab & Yoursefi, 2011 ; Li et al. , 2012 ; Panesar et al. , 2012 ). Nanofibrils of length from 20 to 100 nm intertwine, forming a three-dimensional network, resulting in a high degree of hydrophilicity ( Jozala et al. , 2015 ; Rajwade, Paknikar & Kumbhar, 2015 ), water retention capacity, and porosity, which allows selective permeability, adhesion of cell culture, and diffusion of the culture medium ( Cavka et al. , 2013 ; Zepon et al. , 2013 ; Ashok et al. , 2015 ; Kirdponpattara et al. , 2015 ). Many studies have used the bacterial cell membrane in vitro, in preclinical studies investigating drug, hormone, and protein release systems, artificial skin ( Fu, Zhang & Yang, 2013 ), cartilage ( Cruz et al. , 2016 ), menisci ( Achatz et al. , 2016 ), intervertebral discs ( Fávaro et al. , 2016 ), valvular prostheses, artificial corneas, and the urethra ( Rajwade, Paknikar & Kumbhar, 2015 ). However, it will be necessary to improve our knowledge of bacterial cellulose membrane (BCM) biointegration and biodegradation, especially with respect to BM-MSCs. This purpose of this study was to characterize and evaluate rabbit BM-MSC behavior in vitro when associated with a BCM, by analyzing adhesion, expansion, and cellular integration with the biomaterial, as well as the ability to induce macrophage activation. BCM cytotoxicity and toxicity were also evaluated. Material and Methods Study design Bone marrow samples were collected from three adult rabbits and used for isolation and cryopreservation of MSC. A Mus musculus mouse was used as a source of peritoneal macrophages. To determine cellular viability, Trypan Blue staining and growth curve analysis were performed. For the fibroblastoid colony-forming unit assay, cells collected from the bone marrow (BM) cultured in 24-well plates at passage 6 were used. Chondrogenic, osteogenic, and adipogenic induction were used to assess the potential for differentiation into mesenchymal lineages. To verify BM-MSC biointegration with the BCM, inverted light microscopy and scanning electron microscopy (SEM) were used to analyze the phagocytic capacity, toxicity, and cytotoxicity of the BCM. This study was performed in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Ethics Committee on the Use of Animals of the Federal University of Piauí (permit number: 268/16). Anesthetic protocol for bone marrow collection After solid anesthetic fasting of 4 h, and 2 h of liquids, the rabbit was chemically restrained with a combination of 35 mg/kg of ketamine hydrochloride and 3 mg/kg of midazolam maleate. Trichotomy of the major trochanter region was performed, followed by antisepsis by femoral puncture with a 5 mL syringe; a heparinized 40 × 12 mm needle was used to obtain a BM sample. For antibacterial prophylaxis, 10 mg/kg of enrofloxacin was given twice daily for 7 days, and 25 mg/kg of sodium dipyrone plus 3 mg/kg of tramadol was administered twice daily for 3 days for pain control ( Ninu et al. , 2017 ). BM-MSC isolation, cultivation, and expansion The methodology presented was adapted from Argôlo Neto et al. (2016). Medullary aspirate (1. 5 mL) was diluted in phosphate-buffered saline (PBS) at a ratio of 1:1 in 15 mL conical tubes. The resulting contents were filtered through 100 µm mesh, deposited in a 15 mL conical tube containing Ficoll Histopaque at a ratio of 1:1 (Ficoll:BM), and centrifuged at 2, 000 rpm for 30 min at 20 °C to separate the cellular constituents by density gradient. The whitish halo, rich in mononuclear cells, was aspirated with an automatic pipettor (Houston; Swiftpro, HTL Lab solution, Warsaw, Poland), immediately diluted in sterile PBS with 1% antibiotic (100 U/mL penicillin and 100 µg/mL streptomycin) for cell lavage, and re-centrifuged at 1, 500 rpm for 10 min at 20 °C. BM samples were resuspended in complete Dulbecco’s modified Eagle’s medium (DMEM) containing 3. 7 g/L sodium bicarbonate and 10–15 mM HEPES (Invitrogen, no. 15630080; Invitrogen, Carlsbad, CA, USA), pH 7. 5, 15% fetal bovine serum (Invitrogen), 1% penicillin–streptomycin, 1% L-glutamine (Invitrogen), and 1% non-essential amino acids (Sigma, St. Louis, MO, USA), and cell viability was assessed. For this purpose, a 50 µL aliquot of each sample was diluted in 50 µL 0. 2% Trypan Blue dye, and mixed in a sterilized glass vial for cell counting in a Neubauer chamber. Cells were seeded in a six-well cell culture plate (TPP) at a density of 10 6 cells/well in 2. 0 mL of low-glucose DMEM, and kept in an incubator (Thermo Scientific Series II Water Jacket; Thermo Scientific, Waltham, MA, USA) at 37 °C in 5% CO 2 and 95% humidity. The wells were washed twice every 3 days with PBS solution containing 1% antibiotic (100 U/mL penicillin and 100 µg/mL streptomycin), followed by exchange of the culture medium until the cultures reached 80% confluency. Subsequently, the wells were subjected to trypsinization with 2. 0 mL 1× trypsin (Invitrogen, no. 25200-114, 10× Trypsin–EDTA solution), and incubated at 37 °C for 5 min. Following this, trypsin was inactivated with the addition of 4. 0 mL low-glucose DMEM. The solution was transferred to a 15 mL conical bottom tube, and centrifuged (FANEM refrigerated Cytocentrifuge MOD. 280R Excelsa 4) at 20 °C and 1, 500 rpm for 10 min. The supernatant was discarded, the pellet was resuspended in 1. 0 mL of DMEM, and a new cell count was performed. The cells in suspension were used for expansion. To do this, 10 6 cells/mL in 25 cm 2 tissue culture bottles with 3. 0 mL of supplemented DMEM were incubated at 37 °C in 5% CO 2 and 95% humidity. The cultures were expanded and photographed with an inverted phase-contrast microscope (COLEMAN NIB-100), and peaked with twice the original area; cell concentration was verified at each passage. Fibroblastoid colony-forming unit assay After plating 1 ×10 4 cells/mL of the BM-MSC rabbit fraction in 24-well plates, plates were observed daily to monitor the establishment of colonies with more than 30 cells. Cells were then fixed with 4% paraformaldehyde for 30 min, and stained with Giemsa for 10 min at room temperature. Any excess stain was washed away with distilled water. The colonies were observed and macroscopically counted on the 24-well plates ( Paramasivam et al. , 2016 ). Cell viability Cell count, which determines concentration and viability, was performed using the Trypan Blue exclusion method. After mixing 30 µL of the cell suspension with 30 µL Trypan Blue solution (50 µL of 4. 25% sodium chloride in 200 µL of Trypan Blue), a 10 µL aliquot was observed in a Neubauer chamber under an optical microscope (10× objective). The BM-MSC growth curve was performed in duplicate by plating 1 ×10 4 cells/mL in five six-well plates, and counting two wells every 24 h over the course of 15 days. The culture medium of the plates was changed every 3 days to maintain nutrient availability ( Sangeetha et al. , 2017 ). Cell differentiation For cell differentiation assays, we use the protocols provided by Stem Pro ®. Analysis of cell differentiation potential was performed with sixth-passage BM-MSCs cryopreserved in liquid nitrogen for 12 months. They were thawed and grown in 25 cm 2 bottles for cell expansion until 80% confluency was reached. Cultures were then trypsinized and seeded at the concentration according to the manufacturer’s instructions, for chondrogenic, osteogenic, and adipogenic differentiation. For chondrogenic differentiation, 3 ×10 5 cells per well were seeded in a 96-well plate. After 48 h, formation of spheroid bodies was observed, and the culture medium was replaced with that from a Stem Pro ® Chondrogenesis Differentiation Kit. Exchange of the medium was performed every 3 days during a 21-day period. Analysis was performed with histological sections stained with Alcian Blue. For osteogenic differentiation, 6 ×10 4 cells were seeded in a 24-well plate. Initially, the supplemented culture medium was removed and replaced with the osteogenic induction medium, and changed every 3 days during a 21-day period. During this period, morphological characteristics of the cells were evaluated. After osteogenic differentiation, cells were stained with Alizarin Red, which identifies the calcium-rich extracellular matrix, and is characteristic of the presence of osteoblasts. To do this, the cell monolayer was washed with PBS, and fixed with 10% alkaline phosphatase (AP) for 30 min at room temperature. The AP was then removed, the cell monolayer was washed with distilled water, and Alizarin Red was added for 5 min. Subsequently, the dye was removed, and five washes were performed with distilled water; the calcium-rich extracellular matrix and the amount of calcium deposits were recorded with an inverted light microscope. For adipogenic differentiation, 2 × 10 4 cells per well were seeded in a 24-well plate, and Stem Pro Adipogenesis Differentiation Kit induction medium was added once the cells reached 80% confluency. The culture medium was exchanged every 3 days over a period of 10 days. Once differentiation occurred, the culture was stained with Oil Red to visualize lipid vacuoles. BM-MSC biointegration with the BCM The bacterial cellulose membrane used in this study was developed in partnership with the pharmacy department of Sorocaba University - UNISO (Sorocaba, São Paulo-Brazil). MCB was obtained from the culture of G. xylinus ATCC 53582 prepared using 100 ml of the Hestrin and Schramm medium at 30 °C for 48 h under agitation of 150 rpm. After that, 10 6 cells/mL-1 were withdrawn from the culture medium. For the production of MCB, 24 well plates were used. Each well was filled with 1mL of inoculated culture medium. Plates were maintained at 30 °C in a static culture for 0, 24, 48, 72 and 96 h. MSCs were seeded onto the membrane ( Jozala et al. , 2015 ). To study BM-MSC expansion and biointegration with the BCM, 2 ×10 4 cells were cultured in 12-well plates on BCM for three distinct periods (1, 7, and 14 days). The BM-MSCs were fixed to the BCM using 3% glutaraldehyde, washed once with PBS, and dehydrated by slow water exchange using a series of ethanol dilutions (30%, 55%, 70%, 88%, 96%, and 100%) for 20 min at each concentration. For analysis by SEM (FEI Quanta FEG 250), samples were fixed to the stub with double-sided carbon tape, placed in a dehumidifier for 2 h, and metalized with gold. Phagocytic activation Phagocytic activity was assessed by collecting resident macrophages from the mouse peritoneum. The animal was euthanized by cervical dislocation after being reassured and sedated by intraperitoneal injection of a combination of xylazine hydrochloride and ketamine hydrochloride (10 and 80 mg/kg body weight, respectively). Macrophage removal was performed in a laminar flow hood with the animal affixed in the dorsal decubitus position by administering 8 mL of sterile PBS at 4 °C into the abdominal cavity. The abdominal region was softly massaged, and aspiration was performed using a needle coupled to a sterile syringe. The cells were counted in a Neubauer chamber by the Trypan Blue exclusion colorimetric method, and a minimum of 95% of living cells was obtained. The cells were counted using Neutral Red to obtain the desired concentration of macrophages (2 ×10 5 cells/mL). Peritoneal macrophages were plated in each well, and incubated on the BCM. After 48 h of incubation at 37 °C and 5% CO 2, 10 µL of stained zymosan solution was added, and incubation continued for 30 min at 37 °C. Following this, 100 µL of Baker’s fixative was added to paralyze the phagocytic process, and after 30 min the plate was washed with 0. 9% saline solution to remove the zymosan and Neutral Red that were not phagocytized by macrophages. The supernatant was removed, 100 µL of extraction solution was added, and after solubilization on a Kline shaker, absorbance was measured at 550 nm in a BioTek plate reader (model ELx800) ( Souza et al. , 2017 ). Toxicity To assess toxicity, the nitric oxide (NO) induction test was performed. Peritoneal macrophages (2 × 10 5 per well) were plated and incubated with the BCM after 24 h of incubation at 37 °C and 5% CO 2. Cell supernatants were transferred to another 96-well plate for nitrite dosing. The standard curve was prepared with sodium nitrite diluted in Milli-Q water at 1, 5, 10, 25, 50, 75, 100, and 150 µM in the appropriate culture medium. At the different timepoints, the standard curve was determined with the same volume of Griess reagent (1% sulfanilamide in 10% H 3 PO 4 [v:v] in Milli-Q water, added in equal parts to 0. 1% naphthylenediamine in Milli-Q water), and the absorbance was read on a BioTek plate reader (model ELx800) at 550 nm. Lipopolysaccharide (LPS) was used as a positive control ( Sundaram et al. , 2016 ). BCM cytotoxicity The basis of cytotoxicity assays is the evaluation of biomaterial-induced interference in cellular metabolic processes, and the investigation of processes that may intervene in cell growth/multiplication, or even culminate in cell death ( Ávila et al. , 2014 ). According to Boersema et al. (2016), cytotoxicity can be evaluated by different methods according to the type of cell damage: alterations in plasma membranes can be evaluated by means of dyes such as Trypan Blue and alamarBlue; alterations in the metabolic functions of mitochondria can be measured by the MTT (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium bromide) colorimetric method. The experiments were performed separately in 24-well plates. In the first plate, 2 × 10 5 macrophages per well were plated, and 500 µL of supplemented RPMI 1640 medium was added. In the second plate, 1 × 10 5 BM-MSCs in low-glucose DMEM were added. The plates were incubated at 37 °C and 5% CO 2 for 4 h to allow for cell adhesion. Two washes were performed with their respective media for removal of nonadherent cells. Subsequently, 500 µL of each medium was added, and the BCM (diameter 15. 4 mm) was added. Macrophages were incubated for 48 h, and BM-MSCs for 7 days, followed by the addition of 10% 5 mg/mL MTT (diluted in medium). The macrophages and BM-MSCs were incubated for another 4 h in an incubator at 37 °C with 5% CO 2. The supernatant was discarded, and 100 µL of dimethyl sulfoxide (DMSO) was added to all wells. The BCM was removed, and the plate was shaken for 30 min on a Kline shaker (model AK 0506) at room temperature for complete dissolution of the formazan. The colorimetric reading was performed in a spectrophotometer at 550 nm in a BioTek plate reader (model ELx800). In the control group, the same conditions were applied to the culture media and the respective cultured cells ( Barud et al. , 2015 ). Statistical analysis For analysis of phagocytic capacity, Student’s t -test was used for independent samples of the cytotoxicity (MTT) and NO induction assays. GraphPad Prism version 5. 0 was used to generate the graphs. These tests were performed in triplicate. Results Immediately after isolation, cells from the BM appeared rounded and dispersed, and floated in the culture medium. From the first day of culture, it was possible to identify undifferentiated cells with a fibroblastoid morphology that had adhered to the plastic. On day 2, cells appeared to still be in the adhesion process ( Fig. 1A ). The formation and proliferation of fibroblast colonies were evident on day 5 of culture. Colonies were of varying sizes, surrounded by empty spaces, and distributed throughout the culture plate. The cells showed well-defined cytoplasmic boundaries, and nuclei with regions of condensed chromatin; the closer they were to one another, the more elongated the cells were, and they were arranged parallel to one another ( Fig. 1B ). 10. 7717/peerj. 4656/fig-1 Figure 1 CFU-F assay, bone marrow mesenchymal stem cell (BM-MSC) in culture and growth curve of stem cells. (A) Cells in the adhesion process on day 2 of cell culture performed in 12-well plates (objective 20×, bar: 25 µm). (B) CFU-F assay in a 24-well plate: photomicrography of Giemsa-stained BM-MSC colonies after 5 days of cell culture at 80% confluency, and colonies with more than 30 cells per field (objective 20×, bar: 25 µm), (C) cells arranged in parallel with fibroblastoid morphology at 80% confluency on day 10 of cell culture in 12-well plates (objective 10×, bar: 50 µm), (D) and (E) cytoplasmic adhesion and expansion with 80% confluency in 25 cm 2 bottles after trypsinization on day 15 of culture (objective 10×, bar: 50 µm), (F) cells with fibroblastoid morphology arranged in parallel and in colonies at 80% confluency in 25 cm 2 bottles after trypsinization on day 20 of culture (10×objective, bar: 50 µm) and (G) growth curve of stem cells derived from rabbit bone marrow during 15 days of culture after thawing, at a concentration of 1 ×10 4 cells/mL. Phases identified: lag (days 1–4), log (days 5–11), and culture decline (days 12–15). In the observation on day 10, the cells adhered and arranged in colonies with 80% confluency in a 12-well plate ( Fig. 1C ). After the first passage, cells reached confluency more rapidly, with only a 5-day interval until 80% confluency was reached in 25 cm 2 bottles ( Figs. 1D – 1F ). After thawing, cell cultures exhibited viability of 96%, with similar morphological characteristics and maintenance of differentiation as the primary culture. The observed time to confluency was superior to that of the first passage of the primary culture. At day 3, the culture showed 80% confluency. In the growth curve, we identified two phases (lag and log) which corresponded to the adaptation period of the cells to the culture conditions, the exponential growth period, and the stability period with a reduction in cell growth. Data regarding cell concentration were used to evaluate cell kinetics, and are presented in Fig. 1G. Differentiation into BM-MSC mesodermal lineages The cell differentiation assay showed the potential of BM-MSCs to differentiate into chondrogenic, osteogenic and adipogenic lineages. Following chondrogenic differentiation, cells were stained vibrant blue by Alcian Blue, and control cells presented some spontaneous differentiation. During osteogenic induction, the culture demonstrated increased deposition of calcium in the extracellular matrix from day 13 of culture. On day 21 of induction, the culture exhibited osteogenic characteristics, which were confirmed with Alizarin Red staining. The negative control showed adhered cells with morphology indicative of spontaneous differentiation foci. During adipogenic differentiation, the cells gradually changed to a fibroblastoid morphology, and the cytoplasmic lipid vacuoles became bulky ( Fig. 2 ). 10. 7717/peerj. 4656/fig-2 Figure 2 Photomicrographs showing BM-MSC differentiation. (A) Negative control for 14 days of chondrogenic differentiation (objective 10×, bar: 25 µm), (B) negative control for osteogenic differentiation for 21 days (objective 10×, bar: 25 µm), (C) negative control for adipogênica differentiation for 10 days (objective 10×, bar: 25 µm), (D) BM-MSC chondrogenic differentiation (objective 20×, bar: 25 µm), (E) BM-MSC osteogenic differentiation showing calcium deposits in the extracellular matrix (objective 10×, bar: 25 µm) and (F) photomicrograph showing the adipogenic differentiation of BM-MSCs, with lipid vacuoles present in the cytoplasm stained red with Oil Red (objective 40×, bar: 25 µm). BM-MSC biointegration with the BCM In the BCM-associated cell culture, BM-MSCs with a fibroblastoid shape integrated with the biomaterial, and proliferation of the colonies was evident at 14 days of culture ( Figs. 3A, 3B ). 10. 7717/peerj. 4656/fig-3 Figure 3 Photomicrographs of BM-MSCs adhered to the bacterial cellulose membrane (BCM) and scanning electron microscopy showing BM-MSC anchorage and biointegration with the BCM. (A) BM-MSC adhesion after 7 days of cell culture, highlighting the formation of CFU-F on the BCM (objective 20×, bar: 25 µm), (B) BM-MSC colonies after 14 days of culture (objective 10×, bar: 50 µm), (C) analysis after 24 h of cell culture (40, 000×), (D) and (E) with after 7 (10, 000× and 15, 000× respectively) and (F) 14 days of culture (40, 000×). Using SEM, it was possible to observe that the rounded shape of the cells after 24 h of culture was maintained after being subtly anchored to the randomly arranged fibers of the BCM. After 7 days of culture, the cells presented themselves in groups, forming colonies with several fixation points, generating greater adhesion to the biomaterial. Micrographs recorded after 14 days of cell culture show BM-MSCs with their cytoplasm fully adhered to the BCM ( Fig. 3 ). Macrophage activation and BCM cytotoxicity In the phagocytic activity assay, Student’s t -test was performed to determine the difference between the absorbance resulting from the association of macrophages with cellulose, and the control group (macrophages in the presence of 0. 2% DMSO in RPMI 1640 medium). In the presence of the BCM, macrophage activity was significantly increased ( Fig. 4A ). 10. 7717/peerj. 4656/fig-4 Figure 4 Macrophage activation and BCM cytotoxicity. (A) Zymosan particle phagocytosis by macrophages in the presence of the BCM. The graph represents the mean ± standard error of the mean of three independent experiments performed in triplicate (control: mean 0. 28567, standard deviation 0. 03161; BCM, mean 0. 36100, standard deviation 0. 03474). ABS, absorbance; C, control; BCM, bacterial cellulose membrane; * p < 0. 05. (B) Colorimetric nitrite dosage produced by macrophages treated with lipopolysaccharide (LPS) in the presence of the BCM. The plot represents the mean ± standard error of the average of three independent experiments performed in triplicate (control, mean 100. 0000, standard deviation 0. 0000; LPS, mean 150. 8889, standard deviation 1. 0541; BCM, mean 109. 6300, standard deviation 11. 0047). Student’s t -test was performed for comparison between groups and the control (0. 2% dimethyl sulfoxide [DMSO] in RPMI 1640 medium). C, control; LPS, lipopolysaccharide; BCM, bacterial cellulose membrane; * p < 0. 05. (C) Formazan crystals in BCM cultured with peritoneal macrophages and (D) BM-MSCs. Increasing view 40×. (E) BM-MSC viability in the BCM (control, mean 100. 0000, standard deviation 0. 0000; BCM, mean 94. 4533, standard deviation 1. 1926), and (F) viability of murine macrophages in the BCM (control, mean 100. 0000, standard deviation 0. 0000; BCM, mean 97. 7867, standard deviation 3. 3200). The plot represents the mean ± standard error of the mean of three independent experiments performed in triplicate. Student’s t -test was performed to compare the groups with the control (0. 2% DMSO in DMEM/RPMI medium). C, control; BCM, bacterial cellulose membrane; * p < 0. 05. The colorimetric reading of NO release showed that the levels remained at a non-cytotoxic concentration for the cells in the presence of the BCM ( Fig. 4B ). The difference in NO release between the control and BCM was statistically significant at p < 0. 05 ( p -value 0. 0184, t 0. 05 -critical: 2. 6252), as was that between LPS and BCM ( p -value: 0. 0001; t 0. 05 -critical: 11. 1963). The tetrazole salt (MTT), incubated with cells with full metabolic activity, showed intense mitochondrial activity ( Figs. 4C, 4D ). In this trial, the metabolism of MTT by BM-MSCs showed a statistically significant difference ( p -value: 0. 0001; t 0. 05 -critical: 2. 6252) but there was no statistically significant difference ( p -value: 0. 0628; t 0. 05 -critical: 2, 000) between the BCM associated with murine macrophages and with the control. In both conditions, cell viability was greater than 94% ( Figs. 4E, 4F ). Discussion After isolation, BM-MSCs exhibited a rounded shape in culture. During the adhesion and expansion process, their morphology modified, becoming gradually fusiform, and proliferating in parallel in colonies; the exclusion of hematopoietic cells in the medium exchanges was perceptible. Similarly, Zhang et al. (2014) stated that MSCs adhere to favorable surfaces with rapid morphological changes, ranging from rounded to elongated shapes. According to Ikebe & Suzuki (2014), adhesion to plastic is the first criterion for the characterization of MSCs. In the cellular adhesion phase, physicochemical connections occur between the cells and the contact surface, including ionic forces that rapidly alter cell morphology, and which are evident after 1 h of culture ( Bakhtina et al. , 2014 ; PU & Komvopoulos, 2014 ; Wang et al. , 2016 ). The organization of cells in fibroblastoid colonies has been considered by Kisiel et al. (2012) as the second major characteristic of MSCs. In this experiment, colony formation was evident after 5 days of primary culture, suggesting that these interactions can occur without cellular differentiation, and therefore allow fibroblastoid morphology to be maintained. Regarding cell viability after thawing, the lag phase was evident from day 1 to day 4 of the growth curve, and the log phase occurred between days 5 and 11, with exponential mitotic divisions evident mainly between days 9 and 11; a decline in the number of cell divisions occurred between days 12 and 15. Secunda et al. (2015) defined the lag phase as a relatively short stage characterized by onset of the release of cell proliferation factors. The exponential cell growth (log) phase is the second phase, in which the growth rate and duration depend on the medium used. When cellular metabolism can no longer be maintained, cells undergo apoptosis. Levels of confluence above 90% induce cell death through a mechanism of inhibition by contact, triggering apoptosis, in addition to the reduction of substrate levels of the culture medium, due to the high cellular concentration, as described by Meirelles & Nardi (2003). The ability to differentiate into more than one mesenchymal lineage (chondrogenic, osteogenic, or adipogenic) is an important multipotentiality feature of MSCs, and is a fundamental requirement for their characterization ( Wuchter, wagner & Ho, 2016 ). According to Kolf et al. (2015), the tissue formed by chondrogenic cell differentiation acquires a vibrant blue color when stained with Alcian Blue; during osteogenic differentiation, it is possible to observe the gradual deposition of calcium in the extracellular matrix, which is attributable to the presence of osteoblasts. Alizarin Red staining showed a fairly characteristic reddish coloration, providing evidence of this potential. According to Munir et al. (2017), formation of lipid vacuoles in the cell cytoplasm and staining by Oil Red characterize the formation of adipocytes; during adipogenic differentiation, several independent vacuoles can be found, and fuse as they expand inside the cell. In this study, cell culture using specific media for differentiation into mesodermal (chondrogenic, osteogenic, or adipogenic) lineages demonstrated the multipotentiality of rabbit BM-MSCs. Cell adhesion and proliferation largely depend on the characteristics of the biomaterial surface, since interactions that occur on the surface will drive the biological responses ( Chahal et al. , 2016 ; Khayyeri et al. , 2016 ). After 7 days of culture, cells showed organization in a fibroblastoid format with a tendency for cell grouping. In the analysis performed at 14 days, the BM-MSCs were present in colonies, and covered the BCM surface. Using SEM, we verified that BM cells maintain their rounded shape on the BCM surface in the first 24 h, with few biomaterial fixation bridges. A delay in BM-MSC anchoring to the BCM was observed when compared to adhesion in culture plates, and this anchoring onset was evident in a few hours. According to Silveira et al. (2016), the three-dimensional structure of BCM nanofibers exhibits an arrangement similar to that of the collagen fibers of the extracellular matrix, and a surface with different pores can provide variable times for cell adhesion to the biomaterial. BM-MSC anchoring and proliferation on the BCM were evident on day 7 of culture with grouped cells, and several cytoplasmic projections were evident in the BCM. On the 14th day of culture, fixation of the BM-MSCs occurred by interaction with the biomaterial. Consistent with the studies of Alberti & Xu (2016) and Santana, Neto & Sá (2014), the presence of cytoplasmic projections and normal cell morphology are factors that confirm cytocompatibility between the BCM scaffold and cells. Equilibrium in immune system cell activation also reflects a tissue’s regenerative quality. In the presence of the BCM, the macrophages presented a statistically significant increase ( p -value 0. 0002; t 0. 05 -critical: 4. 8118) in their activity compared with the control group. Qiu et al. (2016) clarified that maintaining the scaffold intact during the period of adhesion and cell proliferation is important for the regenerative process and the architecture of the tissue to be repaired. The implanted biomaterial should gradually biodegrade to give rise to newly formed tissue without exacerbating an inflammatory response that compromises the repair quality. Thus, the adequate inflammatory response of the host in specific situations makes the biomaterial compatible with its use. The ability of bacterial cellulose to be degraded has not yet been fully elucidated. In animal and human tissues, it is considered limited due to the absence of hydrolases that rupture the ß (1, 4) binding of the cellulose chain, which is responsible for the solubility of the biomaterial ( Oliveira, Rambo & Porto, 2013 ). Although the idea of a completely degradable scaffold is interesting from the point of view of tissue engineering, there remain difficulties with materials that exhibit this property, since the timing of degradation and tissue repair combined with the mechanical properties acquired by the newly formed tissue have led researchers to believe that a material with a low rate of degradation may respond better when the scarring process requires more time-consuming conditions ( Bhattacharjee et al. , 2015 ). After inflammation, macrophages release NO as a way to eliminate pathogens. In addition, NO is known as an inflammatory response mediator, inhibiting or inducing inflammation according to the concentration of NO released ( Taraballi et al. , 2016 ). The colorimetric nitrite dosage produced by macrophages in the presence of the BCM showed a non-cytotoxic concentration, approaching the value obtained in the control group. The MTT assay is a method to assess cell viability widely used to evaluate the metabolism of MTT in the mitochondria of viable cells when incubated with cells with full metabolic activity crossing the plasma membrane, and which, when coming in contact with the superoxide produced by mitochondrial activity, is reduced by succinate dehydrogenase present in MTT-formazan-containing mitochondria. The crystals formed are insoluble in water; however, they are solubilized in DMSO medium, and show violet coloration. Thus, cell viability is directly proportional to the intensity of staining ( Toh, Yap & Lim, 2015 ). According to Li, Zhou & Xu (2015), a material is considered non-cytotoxic and biocompatible when cell viability is greater than 70%. In this study, the MTT assay presented intense violet staining, showing that the BCM does not produce a toxic effect on the cells; 94% cellular viability is considerably favorable for non-interference of cellular activity. Conclusion The expansion and cellular integration of biomaterials depends greatly on the quality and suitability of the biomaterial surface. The BCM allowed the adhesion, expansion, and biointegration of BM-MSCs, and the cytotoxicity and toxicity of the BCM were low enough to maintain considerable viability in cell culture. Macrophage activation and the rate of BCM degradation make the BCM an ideal biomaterial for slow healing processes in which reconstructed tissues require a scaffold with longer durability. Considering the interaction demonstrated between BM-MSCs and the BCM, it can be stated that the BCM is a promising biomaterial in tissue engineering and regenerative medicine. However, it will be necessary to test the behavior of BCM implants in vivo. Supplemental Information 10. 7717/peerj. 4656/supp-1 Data S1 Results obtained for the macrophage activation and bacterial cellulose membrane cytotoxicity This file contains Unpaired t test results between the variables of Phagocytic capacity; Tetrazole salt (MTT) assay (Formazan crystals in bacterial cellulose membrane cultured with BM-MSCs and peritoneal macrophages); Nitric oxide and their p value; estatistical significance; confidence interval, intermediate values used in calculations and the descriptive statistic. Click here for additional data file.

10. 7717/peerj. 469

PeerJ

Augmented reality in healthcare education: an integrative review

Background. The effective development of healthcare competencies poses great educational challenges. A possible approach to provide learning opportunities is the use of augmented reality (AR) where virtual learning experiences can be embedded in a real physical context. The aim of this study was to provide a comprehensive overview of the current state of the art in terms of user acceptance, the AR applications developed and the effect of AR on the development of competencies in healthcare. Methods. We conducted an integrative review. Integrative reviews are the broadest type of research review methods allowing for the inclusion of various research designs to more fully understand a phenomenon of concern. Our review included multi-disciplinary research publications in English reported until 2012. Results. 2529 research papers were found from ERIC, CINAHL, Medline, PubMed, Web of Science and Springer-link. Three qualitative, 20 quantitative and 2 mixed studies were included. Using a thematic analysis, we’ve described three aspects related to the research, technology and education. This study showed that AR was applied in a wide range of topics in healthcare education. Furthermore acceptance for AR as a learning technology was reported among the learners and its potential for improving different types of competencies. Discussion. AR is still considered as a novelty in the literature. Most of the studies reported early prototypes. Also the designed AR applications lacked an explicit pedagogical theoretical framework. Finally the learning strategies adopted were of the traditional style ‘see one, do one and teach one’ and do not integrate clinical competencies to ensure patients’ safety.

Introduction Augmented reality (AR) supplements the real world with virtual objects, such that virtual objects appear to coexist in the same space as the real world ( Zhou, Duh & Billinghurst, 2008 ). It has the potential to provide powerful, contextual, and situated learning experiences, as well as to aid exploration of the complex interconnections seen in information in the real world. Students can use AR to construct new understanding based upon their interactions with virtual objects, which bring underlying data to life. AR is being applied across disciplines in higher education, including; environmental sciences, ecosystems, language, chemistry, geography and history ( Johnson et al. , 2011 ; Klopfer & Squire, 2007 ). Clinical care is also interested in AR because it provides doctors with an internal view of the patient, without the need for invasive procedures ( Bajura, Fuchs & Ohbuchi, 1992 ; Chris, 2010 ; De Paolis et al. , 2011 ; De Paolis et al. , 2008 ; Pandya, Siadat & Auner, 2005 ). Since students and medical professionals need more situational experiences in clinical care, especially for the sake of patient safety, there is a clear need to further study the use of AR in healthcare education. The wide interest in studying AR over recent years ( Rolland et al. , 2003 ; Sielhorst et al. , 2004 ; Thomas, John & Delieu, 2010 ) has highlighted the following beliefs: • AR provides rich contextual learning for medical students to aid in achieving core competencies, such as decision making, effective teamwork and creative adaptation of global resources towards addressing local priorities ( Frenk, Chen & Bhutta, 2010 ). • AR provides opportunities for more authentic learning and appeals to multiple learning styles, providing students a more personalized and explorative learning experience. • The patients’ safety is safeguarded if mistakes are made during skills training with AR. While information technology has been presented as a driver for educational reforms, technology, by itself, is not a vehicle for learning ( Merrill, 2002 ; Salomon, 2002 ). To prevent AR from being a gimmick with tremendous potential, it is important to understand the new capabilities that technology offers, including how those capabilities may impact learning ( Garrison & Zehra, 2009 ; Salinas, 2008 ). Therefore, a necessary first step is to analyze the current research on AR in healthcare education to determine its’ strengths and weaknesses. There are two systematic reviews about AR; one is on AR in rehabilitation to improve physical outcomes ( Al-Issa, Regenbrecht & Hale, 2012 ), and the other is focused on AR tracking techniques ( Rabbi, Ullah & Khan, 2012 ). In addition to these, Lee (2012) published a literature review to describe AR applied in training and education, and discussed its potential impact on the future of education. Carmigniani & Furht (2011) developed an overview of AR technologies and their applications to different areas. Shuhaiber (2004) discussed augmented reality in the field of surgery, including its potential in education, surgeon training and patient treatment. Thomas, John & Delieu (2010), provided a brief overview of AR for use in e-health within medicine, and specifically highlighted issues of user-centered development. Ong et al. (2011) presented the use of AR in assistive technology and rehabilitation engineering, focusing on the methods and application aspects. Of the studies include in these reviews, only two focused on medical or healthcare education. The first reviewed the current state of mixed reality manikins for medical education ( Sherstyuk et al. , 2011 ). The second analyzed applying AR in laparoscopic surgery with a focus on training ( Botden & Jakimowicz, 2009 ). Both were helpful in understanding AR from different perspectives, but lacked a broader view of AR in healthcare education. Furthermore, few reviews focused on analyzing AR in relation to learning and teaching, which is important for ensuring that AR has an appropriate instructional design adapted to medical education. The aim of this study was therefore to investigate of the current state of AR in healthcare education and its reported strengths and weaknesses of the reported AR applications for education in healthcare. Material and Methods We conducted an integrative review, as described by Whittemore (2005). Integrative reviews are the broadest type of research review method and allow for the inclusion of various research designs to more fully understand a phenomenon of interest. In contrast, systematic reviews combine the evidence of primary studies, with similar research designs, to study related or identical hypotheses ( Whittemore, 2005 ). They are more useful at providing insights about effectiveness, rather than seeking answers to more complex search questions ( Grant & Booth, 2009 ). Scoping reviews identify the nature and extent of research evidence to provide a preliminary assessment of the potential size and scope of available research ( Grant & Booth, 2009 ). However, due to lack of a process for quality assessment, the findings from scoping reviews are weak resources for recommending policy/practice. Performing an integrative review helped us to understand how AR has been applied in healthcare education, and our findings will be used to help guide better practice. Our review included multi-disciplinary research publications, reported until 2012, which were related to the construct of AR in healthcare education. Inclusion and exclusion criteria In order to present a comprehensive overview of AR relative to healthcare education, we used broad inclusion criteria. An article was included in the review if it was a research paper (*), on AR (**), or on AR in healthcare education (***), and was written in English. The criteria of inclusion and exclusion were further defined as follows in Table 1 : 10. 7717/peerj. 469/table-1 Table 1 The inclusion and exclusion criteria. Criterion Inclusion criteria Exclusion criteria Research • Clearly described the goal or research question • Neither goal nor research question described • A scientific study design • Review papers were put in introduction • The data collection and analysis methods were clearly described • The results were clearly described Focus of the technology • Combination of real and virtual environments • Used mixed reality in name, but was only virtual reality. • Interactive in real-time • Real or perceived registration in 3D Content • Healthcare education • Education without medicine • Health science education • Medicine without education • Medical education • Veterinary medicine education Clarification of criteria terms: • Research paper. There is no widespread accepted set of criteria with which to assess the quality of studies. Further, research paradigm is different across the various members of the academic community, such as developer, educator and doctor. We have not restricted the methodology and the writing style of the research papers but they should contain the following core information; clear description of the context, study aims, research question, study design, sampling, data collection and analysis, and findings. Papers were excluded if they did not describe the core information mentioned above. • AR. Augmented reality, which sometimes is referred to as ‘mixed reality‘, or ‘blended reality, ’ is a technology that allows a live real-time direct or indirect real-world environment to be augmented/enhanced by computer-generated virtual imagery information ( Carmigniani & Furht, 2011 ; Lee, 2012 ). It is different from virtual reality that completely immerses the user in a computer-generated virtual environment. We did not make a clear distinction between augmented reality and augmented virtuality (AV) where AR is closer to the real world and AV is closer to a pure virtual environment ( Milgram & Colquhoun, 1999 ). Studies focusing on enhancing the user’s perception of and interaction with the real world through virtual information were included. It would be excluded if it only discussed the virtual environment. • Healthcare education. According to the glossary of medical education terms from AMEE, medical education is “the process of teaching, learning and training of students with an ongoing integration of knowledge, experience, skills, qualities, responsibility and values which qualify an individual to practice medicine” ( Wojtczak, 2002, p 36). “With the growing understanding of the conditions for learning within medical care and health care, and the increasing focus on the ‘lifelong’ nature of medical education, medical education now, more so than in the past, needs to span three sectors: undergraduate, postgraduate and the continuing professional development of established clinicians” ( Swanwick & Buckley, 2010, p 123). The two definitions represents to current established perspectives on medical education, the first with a process and outcome focus, while the second is acknowledging education as a lifelong continuum. Search strategy and inclusion procedure Agreement about the review protocol, and inclusion and exclusion criteria was reached through discussion between EZ, NZ and IM. Relevant computerized databases were searched for eligible studies, including: ERIC, CINAHL, Medline, Web of Science, PubMed and Springer-link. Separate searches were completed for each database with no date restrictions, no methodological filter, and the language limited to English. The searches were updated until November 2012. Word groups representing the key characteristics of our study were created and combined in several ways. The first group was ‘augmented reality’ and included terminology with similar meaning such as ‘mixed reality, ’ or ‘blended reality. ’ The second group was ‘medical education’ and included terms like ‘healthcare education, ’ ‘health science education’ and so on. The two key groups of terms used the Boolean operator ‘(and)’ to combine with the terms one another when searching for papers to include. Also, we used symbols like ‘medic * education’ to include more related articles with potentially different endings. EZ independently searched for eligible studies in the six databases using the methods above and identified each article meeting the inclusion criteria. ‘Medical education’ and synonyms were searched in ‘all areas’ in the six databases throughout the search procedure. We began by searching for ‘augmented reality, ’ or synonyms, plus ‘medical education’ in all areas to get the overall data. Next, ‘augmented reality, ’ or its’ synonyms, were searched within the title or abstract field, but with ‘medical education’ in ‘all areas. ’ One reason for this is that we felt the focus terms should be placed in the title or abstract. Another reason is that the papers in which augmented reality was neither in the title nor abstract, were not studying augmented reality when we reviewed them. When the abstract contained insufficient information we sometimes referred to the full text to assess eligibility. This was then discussed with NZ and IM. After confirming that the paper’s title and abstract discussed augmented reality on medical education, the full text was downloaded and printed to re-read and analyze, if it met the review criteria. EZ examined and marked the full texts to select the articles that met the inclusion criteria. AH checked the excluded papers by EZ to ensure we did not leave out any papers that should include. NZ checked the full text and discussed with EZ. IM was involved in the discussions and selection process when necessary. The quality of the studies was then reviewed by all the co-authors for final inclusion. Data extraction and analysis We extracted information specifically on research, technology and learning from the included studies. The characteristics and the results of the included studies were recorded with a standardized data-extraction form. Data were extracted independently and in tripartite for all characteristics. Three main characteristics, including research, technology and learning, and eleven sub-characteristics were described through qualitative content analysis for each of the included studies ( Appendix SI ). Also, we used content analysis to describe the study design for each study ( Appendix SI ). Thematic analysis was used to identify the prominent themes that describe current use of AR in healthcare education. The themes are then presented in the result section in terms of strength and weakness of AR. Results Identification of relevant studies We found 2, 529 papers on AR in medical education in the above-mentioned six databases. After screening the titles and abstracts, we found 270 citations in the titles and 179 in the abstracts that included ‘augmented reality’, ‘mixed reality’ or ‘blended reality. ’ These terms were selected to keep focus on the key characteristics that we wanted to scrutinize and identify. After further reading of the title and abstracts, and removal of any duplicate papers, 77 full-text papers were retrieved and reviewed in more detail. Twenty-five articles met our inclusion criteria for data extraction and were analyzed. Figure 1 shows the selection process. Papers were mainly excluded if their research aim and context were not clearly described. Some articles which seemed to discuss medical education were later excluded because they only focused on medicine or treatment, and not on healthcare education ( Bruellmann et al. , 2012 ; Di Loreto et al. , 2011 ; Pagador et al. , 2011 ), and vice versa, one was excluded because it discussed education of another discipline that could contribute to the health of students ( Hsiao, 2012 ). 10. 7717/peerj. 469/fig-1 Figure 1 Literature search and selection flow. From the included 25 research papers focusing on AR in healthcare education, 20 were based on quantitative research methods, 3 on qualitative research methods and 2 on mixed research methods. In these studies, AR was applied on 15 healthcare related subjects. Most of studies used their own AR system and 5 groups used the same system. Methodological quality of the identified studies We chose to apply a broad inclusion criteria and no restriction with regard to the papers’ methodology since research on AR is still in an early innovative phase. Methodological quality was presented adapting the Medical Education Research Study Quality Instrument (MERSQI) ( Reed et al. , 2007 ). Quality ( Table 2 ) was assessed purely for descriptive purposes, not as grounds to exclude. 10. 7717/peerj. 469/table-2 Table 2 Characteristics of the included studies. Characteristics Types No of studies Study design Experiment 1 group post-test only 2 1 group pre-test and post-test 1 2 groups randomized 6 2 groups non-random 5 3 groups non-random 4 Descriptive Interviews 3 Questionnaire 10 Case 2 Type of data Self-reported (participants) 10 Measured 18 Data analysis Descriptive analysis 19 Other types of analysis 3 Outcomes Satisfaction, attitudes, perceptions, opinions 10 Knowledge, skills 16 Experiences 2 Healthcare outcome 0 Not reported 3 Use of augmented reality in healthcare education The earliest study on AR in healthcare education was published in 2002 but publications in the field take off starting in 2008 (see Appendix SI ). Fig. 2 was developed to map our results on AR in healthcare education, and to give us a clearer understanding of learning paradigms and the capabilities of AR offered in current research. Across the studies we saw high variability in the research aims and also the role of AR in healthcare education ( Fig. 2 ). Twelve studies focused on evidence that AR can improve learning. Seven studies were aimed at developing AR systems for healthcare education. Two studies investigated the user’s acceptance of AR as a learning technology. Six studies tested AR applications. The main use of AR for learning has been to provide feedback, and eight studies used AR as a means to provide feedback to students. Two studies used AR as an innovative interface and two studies used it for simulator practice. The other studies tried AR as navigation, regenerative concept, remote assessment and training, and as a meaningful information tool. One used it to reduce resources, while another group used it to offer immersion in a scenario, and one tried to give participatory reality. The research results showed that learners can accept AR as a learning technology, and that AR can improve the learning effect by acquisition of skills and knowledge, understanding of spatial relationships and medical concepts, enhancing learning retention and performance on cognitive-psychomotor tasks, providing material in a convenient and timely manner that shortens the learning curve, giving subjective attractiveness, and simulating authentic experiences (see Appendix SI and Fig. 2 ). 10. 7717/peerj. 469/fig-2 Figure 2 Characteristics of AR in medical education. * This number is the total of unique participants for all the included papers. We used the largest number given for two groups ( Botden et al. , 2007 ; Botden et al. , 2008 ; Botden, Hingh & Jakimowicz, 2009a ; Botden, Hingh & Jakimowicz, 2009b ; Leblanc et al. , 2010a ; Leblanc et al. , 2010b ; Leblanc et al. , 2010c ; Leblanc et al. , 2010d ), who published 4 papers; ** This number shows the type of computer system that was used in the included papers. Three papers did not describe a computer system ( Karthikeyan et al. , 2012 ; Sakellariou et al. , 2009 ; Yudkowsky et al. , 2012 ). Technical specifications Most of the included papers (50%) employed mobile laptops. Four studies used smaller mobile devices such as smart phone, tablet, PDA and e-book readers. Seven papers used stationary desktop computers. Three papers did not mention which computing system they used in their studies. Of the included papers, 68% used a camera and marker as a tracking device. Two papers used an electromagnetic tracker but different markers; one a radiographic marker and one used anatomical landmarks. Two papers used sensors. Other tracking systems, such as hybrid optical tracker and Wi-Fi signal, were found in at least one of the included papers. One paper described using a head-and-hand tracking system, but did not provide details on the technology ( Yudkowsky et al. , 2012 ). One did not use a tracking device because they projected the virtual picture on a manikin ( Pretto et al. , 2009 ). Strengths of AR in healthcare education We identified three themes that related to the strengths of AR in healthcare education. AR implemented in several healthcare areas and aimed at all level of learners AR was applied in various subjects, such as: joint injection, thoracic pedicle screw placement, laparoscopic surgery, administering local anesthesia, endotracheal intubation, ventriculostomy, forensic medicine, inguinal canal anatomy, diathermy, tissue engineering, alimentary canal physiology and anatomy, disease outbreak, clinical breast examination, cardiologic data, and life support training, all of which are applicable to healthcare education (see Fig. 2 ). We found that 64% of the included papers were within surgery, primarily laparoscopic surgery, which represented 44% (11/25). Two groups provided the majority of publications of laparoscopic surgery ( Botden et al. , 2007 ; Botden et al. , 2008 ; Botden, Hingh & Jakimowicz, 2009a ; Botden, Hingh & Jakimowicz, 2009b ; Leblanc et al. , 2010a ; Leblanc et al. , 2010b ; Leblanc et al. , 2010c ; Leblanc et al. , 2010d. Other healthcare subject areas had only one paper included in this research. While two studies did not mention participants, the remaining 23 studies included 713 participants representing medical staff, medical students, high school students and children, (see Table 2 and Fig. 2 ). Participants used AR to learn healthcare skills and aquire knowledge. Most of the participants were, or will be, healthcare staff, however the children and high school student participants may not pursue an education or career in healthcare in their future. AR seems useful for improving healthcare education Ninety-six percent of the papers claimed that AR is useful for improving healthcare education. Several aspects were elicited in the different studies such as decreased amount of practice needed, reduced failure rate, improved performance accuracy, accelerated learning, shortened learning curve, easier to capture learner’s attention, better understanding of spatial relationships, provided experiences with new kinds of authentic science inquiry and improved assessment of trainees. Broad focus of research—from user acceptance, system development and testing, to the study of learning effects Even though every paper in this study had its own research aim and focus, together they gave us a more complete perspective of how AR is being used in healthcare education ( Fig. 2 ). Two papers investigated user acceptance of AR and they claimed that participants would like to use AR instructions in their future professional life, primarily due to the perceived usefulness of AR. Six papers focused on developing AR systems and two of them tested the usefulness of the systems. One of the six studies, in addition to two other studies focused on evaluating the validity of AR systems. One paper described the usefulness, reliability and applicability of the AR system, and one tested the system value. Fourteen out of the twenty-five papers presented AR for various learning aims. Weaknesses of AR in healthcare education We also identified three themes around the weaknesses of AR in healthcare education. Lack of learning theories to guide the design of AR Of the included papers, 80% did not clearly describe which kind of learning theory was used to guide design or application of AR in healthcare education. One claimed that they used activity-based learning but did not tell us how they used it; moreover, the learning strategies are not clearly described in the paper ( Sakellariou et al. , 2009 ). Two groups used standard skills, such as the manual skills of fundamentals of laparoscopic surgery or expert illustration of what is done in practice, to guide design of AR systems. The participants in these groups used the standard skills identified to perform a task. One group, which used situated learning, allowed the participants to explore and navigate with AR environments, but did not show any learning effect ( Rasimah, 2011 ). Only one group used on location learning theory and the learning strategy of collaborative inquiry and role play ( Rosenbaum, Klopfer & Perry, 2007 ). The results indicated that incorporating the affordances of AR games and the dynamic models of participatory simulations make possible new kinds of authentic science inquiry experiences. Traditional learning strategies applied In 64% of the included papers it was shown that they are still using traditional methods of teaching practical skills in medical education, whether or not AR was used as a guidance system or as feedback tool. Three included papers (12%) did not describe how the participants used AR to learn. One wrote that students can explore and navigate with AR environments, but that the time allotted was only half an hour and no learning effect was shown ( Rasimah, 2011 ). However, a few studies explored other methods. One study investigated AR in teaching using different forms such as; group setting, self-learning or revision of cases ( Jan, Noll & Albrecht, 2012 ). One research group used interactive story and another group used game play to attract students ( Karthikeyan et al. , 2012 ; Nischelwitzer et al. , 2007 ). One group used collaborative inquiry and role-play strategies ( Rosenbaum, Klopfer & Perry, 2007 ). Mostly AR applications prototypes reported Fifty-six percent of the papers presented an AR prototype without studying its impact. Five groups studied the ProMIS AR simulator, which was used by colorectal surgeons in their training to improve laparoscopic colorectal skills. The 5 groups contributed with 11 papers. The usefulness, reliability and applicability of the ProMIS AR simulator system were examined, and the systems’ value and validity were also evaluated. ProMIS AR was additionally compared to other systems. Discussion In this paper, we have shown an overview of the use of AR in healthcare education, additionally, we have identified the currently reported strength and weakness. The findings suggests a potential role in healthcare education even if most of the AR applications were still in a prototype stage. Most studies said AR is useful for healthcare education, with one exception that did not mention the learning effect of AR. AR is useful because it helps the healthcare learner to understand spatial relationships and concepts, to acquire skills and knowledge, to strengthen cognitive-psychomotor abilities, and to shorten their learning curve and prolong learning retention. Further, it increases subjective attractiveness by providing students with authentic simulated experiences. Moreover, AR offers more conveniences, such as with time. Most of the studies used AR for learning through feedback or as a navigation system. However, a few used AR to offer immersion into a scenario, a participatory reality or a regenerative concept. Some used AR as an innovative interface or meaningful information tool. The others tried AR for remote assessment and training, or simulator practice. One used it to reduce resources. Comparison with existing literature Two pieces of literature relevant to healthcare education, focused on introducing several examples of using AR systems. Sherstyuk et al. (2011) introduced human manikins with augmented sensory input for medical education, while Botden & Jakimowicz (2009) compared three AR systems that allow the trainee to use the same instruments currently being used in the operating room for laparoscopic surgery. Al-Issa, Regenbrecht & Hale (2012) used systematic review to investigate the effectiveness of physical outcomes through use of AR in rehabilitation. AR is not currently included in rehabilitation training and the study also showed that research on AR in rehabilitation is still in its infancy. Rabbi, Ullah & Khan (2012) attempted a systematic review of AR tracking techniques but did not show a result. Carmigniani & Furht (2011) focused on analysis of the technical specifications of different types of AR and pointed out the advantages and disadvantages. They also discussed AR for use in medicine and education. This review searched six different databases to determine the characteristics of AR in healthcare education, and to distinguish the strength and weakness found in current research. It particularly focused on including studies related to healthcare education. Most of the AR applications found in this review are based on mobile computing systems, especially on laptops. It is different with Carmigniani and Furht’s study where the medical AR application systems are fixed in-doors. While light mobile AR has been predicted to be feasible to develop as real-time AR applications that are locally processed, our findings show that there are still very few examples of light mobile AR ( Carmigniani & Furht, 2011 ). In our review, we aimed to not only describe the research outcome and learning effect of included papers, but also to check which kind of learning theory was used and how they used it. AR and educational theory Although each study presented a clear research aim, few suggestions were given for choosing an AR model that is better for healthcare education. Moreover, there is not enough evidence to inform the design of suitable learning activities with AR system, where knowledge and skill development could be integrated into the learner’s world. Thus, further research in this area should be taken to clarify the appropriate AR model, instructional designs and how to effectively use AR for healthcare education. Study strengths and limitations To our knowledge, this is the first integrative review that specifically addresses AR in healthcare relative to education. We explored how AR was applied in healthcare education encompassing a broad range of learners, learning strategy, outcomes and study designs. Content analysis and thematic analysis were useful to provide a comprehensive understanding on AR in healthcare education. This review tries to provide a comprehensive description of AR in healthcare education with no research methodology filter. However, it is possible that some studies were missed if the key words did not appear on the title or abstract. The studies were also limited by excluding any non-English studies. This was not only because most of papers were published in English, but also because the authors come from different countries, and English allowed them to reach a consensus on the articles to include in the analysis. It is useful to minimize bias, but we possibly excluded some important papers. Further, an interesting AR application could have been missed because it was not published. Conclusions AR is in the early stages of application within healthcare education but it has enormous potential for promoting learning in healthcare based on this review of preliminary AR studies. The infancy of AR in healthcare education requires more than the testing and improvement of prototype products, but also needs to identify appropriate learning theories to better guide application of AR in healthcare education. Supplemental Information 10. 7717/peerj. 469/supp-1 Appendix SI Appendix I: Description of 25 comparative studies included in the integrative review of AR in medical education Click here for additional data file.

10. 7717/peerj. 4815

PeerJ

High correlation between skin color based on CIELAB color space, epidermal melanocyte ratio, and melanocyte melanin content

Background To treat skin color disorders, such as vitiligo or burns, melanocytes are transplanted for tissue regeneration. However, melanocyte distribution in the human body varies with age and location, making it difficult to select the optimal donor skin to achieve a desired color match. Determining the correlations with the desired skin color measurement based on CIELAB color, epidermal melanocyte numbers, and melanin content of individual melanocytes is critical for clinical application. Method Fifteen foreskin samples from Asian young adults were analyzed for skin color, melanocyte ratio (melanocyte proportion in the epidermis), and melanin concentration. Furthermore, an equation was developed based on CIELAB color with melanocyte ratio, melanin concentration, and the product of melanocyte ratio and melanin concentration. The equation was validated by seeding different ratios of keratinocytes and melanocytes in tissue-engineered skin substitutes, and the degree of fitness in expected skin color was confirmed. Results Linear regression analysis revealed a significant strong negative correlation ( r = − 0. 847, R 2 = 0. 717) between CIELAB L * value and the product of the epidermal melanocyte ratio and cell-based melanin concentration. Furthermore, the results showed that an optimal skin color match was achieved by the formula. Discussion We found that L * value was correlated with the value obtained from multiplying the epidermal melanocyte ratio (R) and melanin content (M) and that this correlation was more significant than either L * vs M or L * vs R. This suggests that more accurate prediction of skin color can be achieved by considering both R and M. Therefore, precise skin color match in treating vitiligo or burn patients would be potentially achievable based on extensive collection of skin data from people of Asian descent.

Introduction Evaluating skin color match for treating skin color disorder requires an objective measurement to quantify visual skin color into numerous levels. The principles of color measurement established by the Commission International d’Eclairage (CIE) have been widely applied to skin. Color values have been obtained by using reflectance spectroscopy, expressed in terms of color space L ∗ value, hue angle, and chroma values ( Weatherall & Coombs, 1992 ). Differences of trichromatic color vision between individuals are identified in terms of one, two, or all three CIE color-space parameters: L ∗ value, a ∗ value, and b ∗ value (CIELAB). The L ∗ value, which correlates to perceived lightness and ranges from absolute black (0) to absolute white (+100), is the most sensitive of the trichromatic values to skin color change. Because the method is quantitative and the principles are internationally recognized, these color-space parameters are proposed for the unambiguous communication of skin color information that relates directly to visual observations of clinical importance or scientific interest ( Stevenson et al. , 2012 ). Medical therapy for hypopigmentation disorders has improved in recent years; however, complete repigmentation and perfect skin color match seem to be unsatisfactory in most patients treated ( Iannella et al. , 2016 ). Therefore, a variety of surgical grafting techniques have been performed to treat skin color disorders that do not respond to medical treatment, such as split thickness grafts, cultured autologous melanocytes, minigrafts, suction blister grafts, and non-cultured epidermal suspension, among others ( Gauthier & Benzekri, 2012 ). These techniques contain different advantages and disadvantages with respect to cost, time consumption, treatment area and location, need for equipment, and possible outcome of abnormal appearance ( Gauthier & Benzekri, 2012 ). A previous study reported that cultured skin substitutes (CSS) fabricated from autologous keratinocytes and fibroblasts seeded onto collagen-glycosaminoglycan substrates could be applied to excised, full-thickness burns on five patients. Spontaneous repigmentation of CSS treatment from passenger melanocytes in keratinocyte culture was found within 2 months after grafting ( Harriger et al. , 1995 ). Kahn & Cohen (1995) used very thin epithelial sheet grafts harvested by dermatome from pigmented donor areas and then covered it with petrolatum gauze on five patients with stable vitiligo, which all resulted in excellent repigmentation and no scarring developed. Gupta, Shroff & Gupta (1999) collected the donor epidermal sheets from the blisters developed through the cutaneous suction apparatus, and then grafted onto the denuded skin lesion site. Alternatively, Falabella (2001) implanted very small dermo-epidermal grafts on recipient sites prepared with minipunches of similar size. Mulekar (2003) and Van Geel et al. (2001) both grafted non-cultured melanocyte-keratinocyte suspensions onto previously dermabraded vitiligo lesions and achieved a high repigmentation. Transplantation of pure primary melanocyte cultures has also been proposed with a better response, relatively homogeneous skin color, and capability for use on larger lesion areas ( Chen et al. , 2004 ; Kaufmann et al. , 1998 ; Lontz et al. , 1994 ; Olsson & Juhlin, 2002 ). Currently, the outcomes of repigmentation through pigment skin tissue engineering or pigment cell transfer are still unpredictable. Previous studies indicated that regulating cutaneous pigmentation in cultured skin substitutes was feasible by titration of human melanocytes and keratinocytes ( Swope, Supp & Boyce, 2002 ; Swope et al. , 1997 ). However, epithelial melanocytes used in Swope’s experiments came from a single donation; thus, the influence of different sources on skin pigmentation could not be determined. The aim of this work was to identify how melanocyte numbers and activity from different Asian people modulate skin color via a well-defined correlation between skin color and melanocytes. A total of 15 human adult foreskin samples donated from Asian individuals were analyzed and the correlations among skin color determined by CIELAB color-space parameter: L ∗ value, epidermal melanocyte ratio, and melanin content per 10 6 melanocytes was investigated. Finally, we evaluated the feasibility of applying this skin color relationship in skin tissue engineering. Materials and Methods Materials Gibco Company (New York, NY, USA) supplied Trypsin-EDTA solution (10 ×) and Penicillin-Streptomycin (10, 000 units/ml penicillin G sodium, 10, 000 µg/ml streptomycin sulfate in 0. 85% saline). Melanocyte culture medium contains Medium 254 (Cascade Biologics Inc. , Portland, OR, Portland), 1% Human Melanocyte Growth Supplement (HMGS, Cascade Biologics Inc. , Portland, OR, Portland), and 1% Penicillin-Streptomycin. The EpiLife ® keratinocyte medium (contain 0. 06 mM calcium chloride) was obtained from Cascade Biologics Inc. (Portland, OR, Portland). Poly (ε-caprolactone) (PCL, CAPA 6500, MW 50, 000 g/mol) was purchased from Solvay (Warrington, UK). Type I collagen from calfskin was purchased from Sigma (St. Louis, MO, USA). Dichloromethane (DCM) was purchased from J. T. Baker (Phillipsburg, NJ, USA). Inclusion criteria for skin samples collection A total of 15 Asian young adult foreskin samples were collected during circumcision surgery in Tri-Service General Hospital, R. O. C. Mean age of patients was 24. 47 ± 1. 03 years old, ranging from 21 to 38 years. The study protocol was reviewed and approved by the Institutional Review Board (IRB) in the Tri-Service General Hospital, R. O. C. (TSGHIRB No. : 095-05-0068). Written informed consent was obtained from each donor. Primary culture of human epidermal keratinocytes and melanocytes Primary human epidermal keratinocytes (PHEKs) and melanocytes (PHEMs) were isolated from equal size (1 cm × 1 cm in square) of human young adult foreskin samples obtained in the surgery of circumcision. For culture of PHEKs, the foreskin sample was initially immersed in 10 ml of 0. 2% Dispase II solution (Sigma, St. Louis, MO, USA) at 4 °C for 48 h and diced in pieces, followed by incubation in 0. 05% Trypsin-EDTA solution for 15 min. The pelleted cells were obtained by a centrifugation at 1, 300 rpm for 5 min, seeded in a fibronectin/collagen (AthenaES, Baltimore, MD, USA) coated flask and cultured in EpiLife ® keratinocyte medium at 37 °C in 5% CO 2. For primary culture of PHEMs, the epidermal cell cultured in the EpiLife ® medium were transferred to melanocyte culture medium after the epidermal primary culture and incubated at 37 °C in 5% CO 2. Highly selected culture of PHEMs was then obtained after passage 2. Skin color measurement Foreskin samples were obtained immediately after surgery, blood and adipose tissues were removed, and color was measured in triplicate for each sample ex vivo using a color reader CR-10 (Konica Minolta, Osaka, Japan). The color differences were displayed in terms of trichromatic L ∗, a ∗, and b ∗ values as determined by the CIE. Since the L ∗ value correlates to perceived color brightness (black vs. white) and is the most sensitive trichromatic value for measuring skin pigmentation in the pilot study, it was chosen rather than a ∗ and b ∗ values to represent measured skin color in this study. The L ∗ value was measured in triplicate by detecting the top surface of the collected foreskin sample and recorded as mean ± SE (standard error). Analysis of melanocyte ratio in epidermal cells Fluorescent-activated cell sorting (FACS) was used to distinguish melanocytes from epidermal cells. Epidermal cell suspensions of foreskin were immediately prepared after surgery, which were then centrifuged at 300 × g for 5 min and the cell pellets were treated with a Cytofix/CytoPerm Plus kit (BD, Franklin Lakes, NJ, USA) for subsequent flow cytometry. Briefly, melanocytes were isolated from epidermal cells by adherence. Next, the cells were permeabilized with 200 µl Cytofix/CytoPerm solution for 20 min at 4 °C and washed with 1 ml Perm/Wash Buffer (BD, Franklin Lakes, NJ, USA) twice. For labeling melanocytes, permeabilized cells were stained with mouse anti-human melan-A IgG (Santa Cruz Biotechnology, CA, USA) and incubated with fluorescein isothiocyanate-conjugated (FITC) goat anti-mouse IgG (Jackson, PA, USA). Negative controls for melan-A staining consisted of cells stained with FITC goat anti-mouse IgG only. The samples were then analyzed with five replicates being used for each sample, the samples were analyzed on a FACS Calibur flow cytometer using CellQuest software (BD, Franklin Lakes, NJ, USA); the mean value was then obtained. Determination of melanin concentration The melanin production of melanocytes from distinct foreskin samples in term of total melanin content per 10 6 melanocytes was measured by a spectrophotometric assay. Briefly, Purified primary human epidermal melanocytes (PHEMs) (10 6 cells per pellet) were lysed with 1 M NaOH at 80 °C for 2 h. After centrifugation at 12, 000 × g for 10 min at room temperature, the supernatants were transferred to fresh tubes and melanin content was determined in triplicate for each sample by measuring the absorbance at 490 nm in a spectrophotometer and expressed as microgram of melanin per 10 6 cells. Synthetic melanin (Sigma, St. Louis, MO, USA) was used to plot a standard curve. Preparation of collagen/PCL membranous scaffolds Collagen/polycaprolactone (PCL) scaffolds were prepared as described previously ( Dai et al. , 2004 ). Briefly, type I collagen was dissolved in 1% acetic acid, generating a 0. 25% w/v collagen solution. The collagen solution was poured into a round glass vial (diameter 2. 5 cm and height 4. 5 cm) and frozen at −20 °C for 50 min, followed by lyophilizing in a freeze-dryer (DRC-1100, Eyela, Japan) for 24 h. The PCL/dichloromethane (DCM) solution (2. 5% w/v) was then added to the freeze-dried collagen matrix to prepare a 1:20 w/w collagen/PCL scaffold. The glass vial was kept open overnight to allow DCM evaporation. Color measurement of pigmented tissue-engineered skin substitutes based on collagen/PCL scaffold Pigmented tissue engineered skin substitutes (diameter 2. 5 cm) with primary human epidermal keratinocytes (PHEKs) and/or PHEMs (total cell density of 4. 72 ×10 4 cells/cm 2 ) in various ratios were prepared based on the L ∗ value of calculated skin color relationship including L40 (melanocyte ratio was 1. 79), and L50 (melanocyte ratio was 4. 83) ( L ∗ values as 40 and 50 respectively) groups. Epilife serum-free medium was used to incubate the skin substitute at 37 °C in 5% CO 2. The 6-well tissue culture plastics without cells served as the blank group. The color of pigmented tissue-engineered skin substitutes was measured with a color reader at various time points (10, 20, 30, 40, 50, 60, and 70 days). The analysis of correlation for skin color and melanocytes To estimate the relationships of epidermal melanocyte ratio or melanin concentration vs skin lightness based on foreskin CIELAB L ∗ value, linear regression analysis was performed. The coefficient of determination ( R 2 ) and Pearson’s correlation coefficient ( r ) were calculated to measure goodness of fit of a statistical model and the strength and direction of the linear relationship. Statistical analysis The continuous data such as L ∗ value of foreskin samples and tissue-engineered skin substitutes, melanin amount, and the melanocyte ratio in epidermal cells were shown as “mean ± SD (SD: standard deviation)”. The variables were grouped first, comparing mean values in categories, Pearson’s correlation coefficient (r) and coefficient of determination consecutively. Simple linear regression analysis was used afterwards to evaluate the combined effect of several variables and to estimate the coefficient of determination. Statistical Product and Service Solutions (SPSS) was used to perform a stepwise forward selection procedure. Thus, for each iterative loop of this procedure one more variable was integrated into the new formula. All statistical result is statistically significant when the P value is less than 0. 05 ( P < 0. 05). Statistical analysis was performed using Statistical Package for the Social Sciences, Version 12. 0 (SPSS Inc. , Chicago, IL, USA). Results Measurement of skin color ( L ∗ value) The study design of measuring skin color of foreskin samples, melanin concentration, and epidermal melanocyte ratios for establishing a relationship with skin color is shown in Fig. 1. The skin color of 15 human young adult foreskin samples was measured immediately after surgery by a color reader ex vivo yielding values from 39. 43 ± 0. 21 to 52. 37 ± 1. 90 as determined by CIELAB color-space parameter: L ∗ value ( Table 1 ). 10. 7717/peerj. 4815/fig-1 Figure 1 The illustration of the process for measurement of skin color and formulation of the relation among L ∗ value, epidermal melanocyte ratio and melanin concentration. 10. 7717/peerj. 4815/table-1 Table 1 Patient profile of foreskin and demographic data ( n = 3 for L ∗ value and melanin concentration; n = 5 for epidermal melanocyte ratio; values are mean ± SD) a. Samples (No. ) Age (years) L ∗ value Epidermal melanocyte ratio (%) Melanin concentration (µg/10 6 melanocytes) 1 23 39. 43 ± 0. 21 1. 96 ± 0. 09 68. 96 ± 0. 20 2 25 47. 27 ± 1. 90 3. 18 ± 0. 24 34. 61 ± 0. 39 3 24 41. 20 ± 2. 17 3. 13 ± 0. 10 41. 50 ± 0. 34 4 24 46. 97 ± 1. 52 1. 40 ± 0. 28 81. 91 ± 1. 92 5 24 52. 13 ± 1. 62 1. 97 ± 0. 31 25. 88 ± 0. 11 6 22 52. 37 ± 1. 90 1. 44 ± 0. 09 21. 97 ± 0. 62 7 22 45. 93 ± 0. 59 2. 70 ± 0. 10 47. 43 ± 0. 72 8 26 49. 30 ± 0. 60 3. 49 ± 0. 14 24. 31 ± 0. 58 9 24 44. 07 ± 0. 15 2. 44 ± 0. 10 43. 32 ± 2. 66 10 22 47. 53 ± 2. 37 2. 71 ± 0. 52 26. 39 ± 0. 68 11 21 48. 70 ± 2. 81 3. 40 ± 0. 22 26. 97 ± 0. 27 12 38 49. 73 ± 1. 15 3. 58 ± 0. 07 23. 39 ± 0. 32 13 23 47. 87 ± 3. 26 3. 23 ± 0. 22 17. 99 ± 0. 50 14 25 45. 87 ± 2. 18 1. 68 ± 0. 12 67. 66 ± 0. 46 15 24 50. 60 ± 1. 40 1. 41 ± 0. 14 36. 07 ± 0. 20 Notes. a n, the number of tests performed on each individual sample. Melanocyte ratio in epidermal cells (R) The concentration of melanocytes in foreskin epidermal cell suspensions was measured using FACS. The results showed ratios of melanocytes to epidermal cells ranging from 1. 40 ± 0. 28 to 3. 58 ± 0. 07% ( Table 1 ). Epidermal melanin production (M) To determine the melanin productivity of melanocyte isolated from each foreskin, PHEMs were cultured for two passages reaching a purity of 99. 4% and analyzed using immunohistochemistry assays and FACS ( Fig. 2 ). The results showed that the melanin production of 15 PHEMs ranged from 17. 99 ± 0. 50 to 81. 91 ± 1. 92μg/10 6 cells ( Table 1 ). 10. 7717/peerj. 4815/fig-2 Figure 2 Determination of epidermal melanocyte ratio (100×; scale bar: 100 µm). (A) The foreskin samples were treated immediately with the FACS method. The results showed up to 2. 44% of melanocytes (M1) in a human adult foreskin sample. (B) The results of fluorescent-activated cell sorting (FACS) showed up to 99. 4% purity of melanocytes (M2) in a selected PHEMs culture. (C) Selected primary culture of human epidermal melanocytes (PHEMs) was shown in bright field. (D) The PHEMs in selected primary culture were labeled by an anti-melan antibody incorporated with fluorescein isothiocyanate (FITC). The analysis of correlation for skin color We next examined the correlation between skin color and melanin concentration. Plots of L ∗ value against epidermal melanocyte ratio and L ∗ value against melanin concentration per 10 6 melanocytes were generated based on the data shown in Table 1 ( Fig. 3 ). The results revealed no correlation between L ∗ value and epidermal melanocyte ratio ( r = − 0. 081, R 2 = 0. 0066). However, L ∗ value was negatively correlated with melanin concentration ( r = − 0. 592, R 2 = 0. 3503) ( P = 0. 02) ( Fig. 3 ). However, there were significant associations between the L ∗ value with the production of epidermal melanocyte ratio and melanin content. A further linear regression analysis showed a significantly strong negative correlation ( r = − 0. 847, R 2 = 0. 717) between L ∗ value and the epidermal melanocyte ratio multiplied by cell-based melanin concentration ( Fig. 4 ). Given these results, an equation describing skin color relationship was generated: as L ∗ = a × ( M × R ) + b ( M : melanin concentration per 10 6 melanocytes; R: epidermal melanocyte ratio; a = − 0. 095; and b = 55. 872). 10. 7717/peerj. 4815/fig-3 Figure 3 Statistical correlations between skin color-space parameter: L ∗ value and epidermal melanocyte ratio or melanin concentration. (A) L ∗ value corresponding to epidermal melanocyte ratio expresses a random distribution where the relation between them seem to be fairly weak. Correlation coefficient, r = − 0. 081 ( P = 0. 774) and coefficient of determination, R 2 = 0. 0066. (B) A slight trend of a higher L ∗ value corresponding to a higher melanin concentration per 10 6 melanocytes was observed. Correlation coefficient, r = − 0. 592 ( P = 0. 02) and coefficient of determination, R 2 = 0. 3503. 10. 7717/peerj. 4815/fig-4 Figure 4 Statistical correlations between skin color-space parameter: L ∗ value and the product of epidermal melanocyte ratio and melanin concentration. A strong correlation was shown between L ∗ value and the value of epidermal melanocyte ratio (R) (%) multiplied by melanin concentration (M) (µg/10 6 melanocytes). Correlation coefficient, r = − 0. 847 ( P < 0. 001) and coefficient of determination, R 2 = 0. 717. Application of the skin color relationship in skin tissue engineering As shown in Fig. 5A, collagen/PCL constructs at 70 days were ranked in order of different shades of black as PHEMs > L40 > L50 > PHEKs > Blank. The result of the L ∗ value profile of the collagen/PCL constructs is shown in Fig. 5B. Obvious changes of the L ∗ values were noted in 20 days post-culture for all groups except the blank control group. After 70 days, the L ∗ value of L40, and L50 groups were 43. 50 ± 1. 87, and 50. 67 ± 0. 55, respectively, whereas PHEMs and PHEKs exhibited values of 30. 30 ± 0. 56 and 79. 73 ± 0. 60, respectively. 10. 7717/peerj. 4815/fig-5 Figure 5 (A) The gross appearance of the collagen/PCL constructs at 70 days; (B) the L ∗ value profiles of the collagen/PCL constructs during a period of 70 days. The 6-well tissue culture plastics without cells served as the blank group. Discussion To date, there is no method to accurately estimate transplanted skin pigments via a skin color formula used to fabricate tissue-engineered skin. Numerous previous studies have shown that skin constitutive pigmentation is determined by melanin production levels. For example, Alaluf et al. (2002b) performed correlation analysis and found the best correlation between the L ∗ value and total melanin content in the epidermis. Del Bino et al. (2015) reported that total melanin content, including eumelanin and pheomelanin content, determines the constitutive skin pigmentation. Wakamatsu et al. (2006) indicated that cell-based melanin production and the predominant biological forms of melanin produced by melanocytes affect skin pigmentation. Additionally, implanting different numbers of melanocytes influences skin color. Swope, Supp & Boyce (2002) implanted 1. 1 × 10 2, 1. 1 × 10 3, and 1. 1 × 10 4 human melanocytes/cm 2 into athymic mice and found that mice with the highest density of melanocytes were significantly darker than mice in the other groups. Duval et al. (2014) confirmed that dermal fibroblasts influence the degree of skin pigmentation by measuring quantitative parameters related to skin color, melanin content, and melanocyte numbers in an in vitro skin system. Brankov, Prodanovic & Hurley (2016) found that pigmented basal cell carcinomas (BCCs) have a higher mean melanocyte count compared to non-pigmented BCCs, indicating that the pigment is increased not only because of increased melanin, but also because of increased melanocyte counts. The density of melanocytes varies with the body site, with approximately 900 melanocytes per square mm on the back and approximately 1, 500 melanocytes per square mm in the genital region ( Thingnes et al. , 2012 ). Therefore, for the practical treatment of patients with skin color disorder, the implanted skin should be evaluated for both melanin production and melanocyte count from the donor site for precise color matching. We predict that regulating skin color can be determined by native cell-based melanin production ( M ) and melanocyte numbers in the epidermis ( R ). We found that the L ∗ value was correlated with the value of the multiplicative product of the epidermal melanocyte ratio ( R ) and melanin content ( M ), and these values were much higher than those of the two aforementioned parameters individually, suggesting that considering the epidermal melanocyte ratio and melanin content strengthens the prediction of skin color. Moreover, we performed statistical analysis and found a negative correlation between M and R ( r = − 0. 5359, P = 0. 04). On normalizing M to R ( M ∕ R ) vs L ∗, a low correlation ( r = 0. 3340, P = 0. 22) was observed. This result also agreed with the result that the product of M and R is close to a fixed value. Although skin pigmentation is known to be regulated by melanocytes, the factors affecting and regulating ethnic skin color are further to be determined ( Hoath & Leahy, 2003 ; Snell & Bischitz, 1963 ). Based on our previous research, skin color is essentially affected by the ethnicity of the individual, while melanocyte density and differentiation are also influenced by the environmental factors, such as ultraviolet radiation (UVR) and factors secreted by neighboring keratinocytes and fibroblasts ( Dai et al. , 2018 ). The melanocortin 1 receptor (MC1R), a G protein-coupled receptor that regulates the quantity and quality of melanin production, is the major determinant of the pigment phenotype of the skin. Three agonists, such as alpha-melanocyte stimulating hormone, adrenocorticotrophic hormone, and proopiomelanocortin, can activate MC1R via the cyclase/cAMP/protein kinase A signaling pathway. Next, the cAMP response element binding protein is phosphorylated, resulting in the transcriptional induction of microphthalmia-associated transcription factor (MITF). MITF is involved in regulating the expression of melanogenic proteins, such as tyrosinase (TYR), tyrosinase-related protein 1, and tyrosinase-related protein 2 to regulate skin color ( Gillbro & Olsson, 2011 ). MITF activity can also be regulated by different transcription factors or mediators secreted by keratinocytes and fibroblasts, such as basic fibroblast growth factor, stem cell factor, endothelin-1, prostaglandins, and leukotrienes. Based on a genome-wide association study, polymorphisms in three genes, including SLC24A5, TYR, and SLC45A2, showed highly significant associations with melanin content in the skin ( Stokowski et al. , 2007 ; Wilson et al. , 2013 ). The SLC24A5 gene encoding the NCKX5 protein, as a potassium-dependent sodium-calcium exchanger, exhibits lower exchange activity, resulting in reduced melanogenesis and lighter skin in individuals ( Wilson et al. , 2013 ). These regulatory genes in melanocytes are suggested to differ from race to race or even among individuals and lead to different melanin production capabilities ( Han, Choi & Son, 2006 ). Alaluf et al. (2002a) studied the ethnic variation of melanin content and composition in the human skin from photoprotected and photoexposed areas on human bodies. They found that melanosome size plays a significant role in the variation of different ethnic skin types; African skin had the largest melanosomes followed in turn by Indian, Mexican, Chinese, and European. In addition, the levels of light-colored, alkali-soluble melanin including pheomelanin and DHICA-enriched eumelanin in photoprotected skin areas for European, Chinese, and African skin are estimated as 43. 2, 34. 4, and 15. 1%, respectively ( Alaluf et al. , 2002a ). In our study, L ∗ value was correlated with melanin content but not epidermal melanocyte ratio. A similar density of pigment-producing melanocytes in the skin (∼1, 000/mm 2 ) has been found in skin types from different races ( Staricco & Pinkus, 1957 ). However, the skin color of people from the same group may vary because of different living habits, such as the time spent outdoors, diet, or the use of sunscreen, among other factors. Taiwanese people comprise a multi-racial population, including aborigines and immigrants from the mainland. This area was colonized by the Dutch and Japanese in the past. Therefore, the skin color may reflect different racial characteristics showing high variation in skin samples in this study. In fact, there are few more crucial factors affecting final skin color, including the amount of melanin, melanin composition, and melanosome size across skin from a range of ethnicities. In addition, the higher level of melanin production in darker skin was due to the continuously higher level of tyrosinase activity in melanocytes. Taken together, the amount of melanin in the epidermis plays an important role in skin color of individuals from different races, which agrees with our result that melanin content was significantly correlated with L ∗ value of CIELAB color space. Given that the native melanin-producing capability of a melanocyte depends on genetic factors ( Fitzpatrick, Miyamoto & Ishikawa, 1967 ; Szabo et al. , 1969 ; Wakamatsu et al. , 2006 ), in clinical applications for autogenous pigment cell therapy, the cell-based melanin content is constant in the same individual, and the predicted skin color matches represented as L ∗ value could be detected from the normal skin area adjacent to the hypopigmented lesion site. Thus, a meticulously estimated epidermal melanocyte ratio for the purpose of autogenous pigment cell therapy applications would be attained based on the color formula, which presents as L ∗ = 0. 095 × ( M × R ) + 55. 872 ( M : melanin concentration per 10 6 melanocytes; R : epidermal melanocyte ratio). Moreover, when culturing various melanocyte ratios in the collagen/PCL scaffolds, the results from in vitro experiments confirmed that the use of this skin color formula is feasible. The L ∗ value of L40 and L50 groups was 43. 50 ± 1. 87 and 50. 67 ± 0. 55 respectively, very near the predicted value. Therefore, predicting skin color is possible via collecting skin data for establishing the skin color formula. In the clinic, the product of melanin production and melanocyte distribution (M × R) is indicative of the darkness of skin color. The density of melanocytes differs in different parts of the body. For skin autograft, a plastic surgeon often obtains skin grafts from different body sites for implantation at the recipient site. Based on the skin color formula derived in our study, we can more precisely predict skin color by measuring melanin production by the donor skin and adjusting the ratio of melanocytes and keratinocytes to produce a specific color in transplanted skin to treat patients with skin color disorders, such as vitiligo or burn. The cell sheet technique or plasma gel can be used, adjusting the ratio of melanocytes and keratinocytes to prepare transplanted skins suitable for different skin color lesions. The skin color formula still has some limitations that should be addressed before its clinical application. Swope et al. found that cutaneous pigments would be present as a function of melanocyte density and time after grafting due to the depletion of human melanocytes in CSS in an animal model ( Swope et al. , 1997 ; Swope, Supp & Boyce, 2002 ). Furthermore, skin color is suggested to be determined by several other factors including the total quantity of melanin, the proportion between the brown–black eumelanin and the yellow–red pheomelanin, and its distribution involved in the epidermis ( Naysmith et al. , 2004 ). The log e values of eumelanin/pheomelanin ratio were inversely related to the color variables b ∗ (yellow–blue) ( R 2 = 0. 51, P < 0. 001), a ∗ (red–green)( R 2 = 0. 47, P < 0. 001), and to a lesser extent L ∗ values ( R 2 = 0. 22, P < 0. 001) ( Naysmith et al. , 2004 ). On the other hand, three of cell types including melanocytes, keratinocytes, and fibroblasts actively participate in regulating skin pigmentation through secreted factors and their receptors, and their interactions may determine skin pigmentation ( Yin et al. , 2014 ). The other chromophores present in the skin include oxyhemoglobin, reduced hemoglobin, and carotene, which may influence the true skin color. Skin blood flow increases with increases in hemoglobin ( Tsuchida, Fukuda & Kamata, 1991 ), and a ∗ values correlate linearly well with hemoglobin levels. The a ∗ value indicates the redness of the skin color and is mainly influenced by the degree of vascularization and the stretching of the skin over surrounding tissues. Therefore, all three CIE color-space parameters: L ∗ value, a ∗ value, and b ∗ value should be discussed in the future for perfect match of the skin color used in clinical application. Conclusions A formula for estimating skin color based on CIELAB L ∗ value, epidermal melanocyte ratio, and melanin concentration was generated from people of Asian descent. This skin color formula may serve as a useful methodology for determining distinct pigment cell concentrations for cell therapy or for developing a pigmented skin tissue engineering model for transplantation and pharmaceutical screening applications in vitro. Supplemental Information 10. 7717/peerj. 4815/supp-1 Supplemental Information 1 Analysis of skin pigmentation Click here for additional data file. 10. 7717/peerj. 4815/supp-2 Supplemental Information 2 Measurement of L value in tissue engineering skin Click here for additional data file.

10. 7717/peerj. 4939

PeerJ

3D skeletal muscle fascicle engineering is improved with TGF-β1 treatment of myogenic cells and their co-culture with myofibroblasts

Background Skeletal muscle wound healing is dependent on complex interactions between fibroblasts, myofibroblasts, myogenic cells, and cytokines, such as TGF-β1. This study sought to clarify the impact of TGF-β1 signaling on skeletal muscle cells and discern between the individual contributions of fibroblasts and myofibroblasts to myogenesis when in co-culture with myogenic cells. 3D tissue-engineered models were compared to equivalent 2D culture conditions to assess the efficacy of each culture model to predictively recapitulate the in vivo muscle environment. Methods TGF-β1 treatment and mono-/co-cultures containing human dermal fibroblasts or myofibroblasts and C2C12 mouse myoblasts were assessed in 2D and 3D environments. Three culture systems were compared: cell monolayers grown on 2D dishes and 3D tissues prepared via a self-assembly method or collagen 1-based hydrogel biofabrication. qPCR identified gene expression changes during fibroblast to myofibroblast and myoblast differentiation between culture conditions. Changes to cell phenotype and tissue morphology were characterized via immunostaining for myosin heavy chain, procollagen, and α-smooth muscle actin. Tissue elastic moduli were measured with parallel plate compression and atomic force microscopy systems, and a slack test was employed to quantify differences in tissue architecture and integrity. Results TGF-β1 treatment improved myogenesis in 3D mono- and co-cultures containing muscle cells, but not in 2D. The 3D TGF-β1-treated co-culture containing myoblasts and myofibroblasts expressed the highest levels of myogenin and collagen 1, demonstrating a greater capacity to drive myogenesis than fibroblasts or TGF-β1-treatment in monocultures containing only myoblasts. These constructs possessed the greatest tissue stability, integrity, and muscle fiber organization, as demonstrated by their rapid and sustained shortening velocity during slack tests, and the highest Young’s modulus of 6. 55 kPA, approximate half the stiffness of in situ muscle. Both self-assembled and hydrogel-based tissues yielded the most multinucleated, elongated, and aligned muscle fiber histology. In contrast, the equivalent 2D co-culture model treated with TGF-β1 completely lacked myotube formation through suppression of myogenin gene expression. Discussion These results show skeletal muscle regeneration can be promoted by treating myogenic cells with TGF-β1, and myofibroblasts are superior enhancers of myogenesis than fibroblasts. Critically, both TGF-β1 treatment and co-culturing skeletal muscle cells with myofibroblasts can serve as myogenesis accelerators across multiple tissue engineering platforms. Equivalent 2D culture systems cannot replicate these affects, however, highlighting a need to continually improve in vitro models for skeletal muscle development, discovery of therapeutics for muscle regeneration, and research and development of in vitro meat products.

Introduction Muscle regeneration occurs when muscle fibers are damaged during exercise or injury and quiescent satellite cells are activated to a proliferative myoblast phenotype ( Hill, Wernig & Goldspink, 2003 ). Terminal differentiation is promoted through expression of the myogenic transcription factor myogenin (MYOG) and exit from the cell cycle. Cell fusion with an injured muscle fiber or other myoblasts forms a nascent syncytial myofiber ( Charge & Rudnicki, 2004 ; Le Grand & Rudnicki, 2007 ). Fibroblasts support and stabilize muscle fiber architecture and biomechanics through basement membrane synthesis and facilitate muscle regeneration with extracellular matrix (ECM) deposition and remodeling ( Murphy et al. , 2011 ; Sandbo & Dulin, 2011 ; Sanes, 2003 ). The interaction between myoblasts and fibroblasts, two predominant cell types involved in skeletal muscle regeneration, with surrounding ECM and trophic factors determine healing outcomes. Transforming growth factor beta 1 (TGF-β1), produced by multiple cell types during wound healing, is a primary mediator of the mechanical, biochemical, and cellular behaviors observed in response of muscle to injury ( Karalaki et al. , 2009 ). TGF-β1 signaling differentiates fibroblasts into myofibroblasts, which are significant producers of collagen I (COL I), the main protein component of scar tissue, and α-smooth muscle actin (α-SMA), a highly contractile form of actin ( Mendias et al. , 2012 ). Normally these proteins are transiently expressed during muscle regeneration and contribute to tissue remodeling by temporarily providing physical substrates and biochemical cues for muscle fiber regeneration. However, excessive TGF-β1-mediated myofibroblast activation leads to ECM accumulation and tissue stiffening that decreases the ability of myoblasts to regenerate muscle ( Gilbert et al. , 2010 ; Smith et al. , 2011 ). The use of culture models consisting solely of myoblasts to investigate skeletal muscle regeneration is narrow in scope and validity and prevents effective screening of therapeutics. Using co-culture systems allows the bi-directional signaling between fibroblasts and myoblasts that attenuates and stabilizes myogenesis. Additionally, while 2D in vitro cell culture systems have demonstrable value, they are limited by a number of factors that have unwanted influence on cell behavior. After isolation, primary cells quickly lose their in situ characteristics in response to a mechanically and biochemically alien environment ( Janson et al. , 2013 ). Fibroblasts and muscle stems cells are particularly sensitive to mechanical stimulation, and the rigidity of substrata can mask the cellular responses under investigation ( Engler et al. , 2006 ; Godbout et al. , 2013 ). It is therefore highly valuable to further optimize the in vitro recapitulation of the in vivo environment. In this regard, myogenesis should be studied within the context of tissues rather than cell culture plates. Engineered tissue that contains aligned muscle fibers embedded within connective tissue is biomimetic to muscle fascicles observed in vivo ( Turrina, Martinez-Gonzalez & Stecco, 2013 ; Dennis et al. , 2001 ). The inclusion of fibroblasts and myofibroblasts is therefore our design target for modeling skeletal muscle. Although myofibroblasts are primarily known for their role in tissue regeneration, their inclusion in engineered skeletal muscle tissues has not been investigated, and the effectiveness of using myofibroblasts has not been compared to the known performance of fibroblasts. Signaling of profibrotic factors such as TGF-β1 can deregulate the regenerative capacity of muscle and drive fibrosis when in excess ( Lieber & Ward, 2013 ; Mann et al. , 2011 ) but regularly contribute to muscle regeneration at moderate levels. The use of TGF-β1 may therefore be a means to organize tissue-engineered skeletal muscle development in vitro. A comprehensive and unifying characterization of the impact of TGF-β1 on myogenesis in different culture models has not been established, however. Consequently, this study sought to investigate the influence of TGF-β1 treatment on myogenesis in 2D and 3D culture models, and whether myofibroblasts outperform fibroblasts in accelerating myogenesis when co-cultured with myoblasts. 3D conditions included self-assembled, scaffoldless tissue constructs first developed by Gwyther ( Gwyther et al. , 2011 ; Strobel et al. , 2018 ) and collagen 1-based hydrogels, while 2D models included cells grown on plastic culture plates. Monocultures consisted of murine C2C12 myoblasts, human dermal fibroblasts, or TGF-β1-differentiated myofibroblasts, and co-cultures were composed of myoblasts with either fibroblasts or myofibroblasts. Our results show TGF-β1 signaling in 3D improves myogenesis in myoblast monocultures and co-cultures with myofibroblasts. In co-cultures, myofibroblasts enhanced muscle differentiation to a greater extent than fibroblasts. Our scaffoldless self-assembled tissues and hydrogels containing C2C12s both similarly displayed enhanced myogenesis with inclusion of myofibroblasts and TGF-β1 treatment, indicating this technique has wide applicability across a variety of 3D platforms. However, 2D experiments demonstrated TGF-β1-mediated inhibition of myogenesis in both mono- and co-culture conditions. Our data highlights that 2D in vitro models of skeletal muscle obscure a complete understanding of mechanisms of muscle regeneration, in addition to the usefulness of TGF-β1 treatment and inclusion of myofibroblasts in improving tissue engineered models. Materials and Methods Cell and tissue culture As shown in the experimental design listed in Fig. 1A, cultures consisting solely of human neo-natal dermal fibroblasts (passage 4 hDFs, PCS-201-010; ATCC, Manassas, VA, USA) or mouse myoblasts (pre-passage 4 C2C12s, CRL-1772, ATCC) were plated on standard cell culture plates at 1 k/cm 2 in growth media (GM) consisting of DMEM (Gibco, Waltham, MA, USA), 10% fetal bovine serum (FBS; Hyclone, Pittsburgh, PA, USA), 1% penicillin/streptomycin (P/S, Gibco). To promote differentiation of fibroblasts to myofibroblasts, some fibroblasts were incubated with 1 ng/mL TGF-β1 (Peprotech, Rocky Hill, NJ, USA) for 6 days between P4 and P5. C2C12s and P4 fibroblasts and myofibroblasts were trypsinized, resuspended in GM, and seeded at a 1:1 ratio on standard cell culture plates at 20 k/cm 2 or self-assembled into 3D tissue constructs. To develop 3D self-assembled tissues, cell suspensions were added into custom 2% agarose molds that were prepared by casting molten agarose onto a patterned PDMS mask. After solidification of the gel, the molds were sectioned and added to cell culture plates. Each tissue mold is composed of a hollow well with a 2 mm diameter central agarose post ( Fig. 1A ). When a cell suspension is added to the well, the cells will prefer to adhere to each other instead of the agarose, and self-assemble into a ring-shaped tissue structure that contracts around the central post. A cell seeding density of 350, 000 cells/millimeter of post diameter is used per tissue sample. This density was selected through a pilot experiment determining the concentration of cells required to self-assemble tissue around the circumference of the agarose post, which is a function of the post’s diameter (C Malcuit, pers. comm. , 2013). Hydrogels were produced from a protocol adapted from Langelaan et al. (2011) where 1. 5 × 10 6 C2C12s alone or a 1:3 ratio of fibroblasts/myofibroblasts and C2C12s were resuspended in 42% GM, 54% rat collagen type 1 (3 mg/mL; Corning Life Sciences, Corning, NY, USA), and 2. 7% NaOH (0. 1 M). The hydrogel mixture was poured between house-shaped Velcro anchors super glued to the bottom of a six well plate and gelled for 45 min before adding GM. After 24 h, the media were switched in both 2D and 3D conditions to differentiation medium (DM) composed of DMEM, 2% horse serum (Sigma-Aldrich, St. Louis, MO, USA), and 1% P/S, with or without 1 ng/mL TGF-β1. Media was changed every 2–3 days for 7 days, after which cells and tissues underwent qPCR, ICC or IHC, or mechanical characterization. See Fig. 1B for listing of experimental groups. 10. 7717/peerj. 4939/fig-1 Figure 1 Technical protocol for 2D/3D culture systems and experimental conditions design. (A) Technique : 2D cell culture and 3D tissue engineering protocols involve an initial high serum culture period on standard cell culture plates to separately amplify myoblast and fibroblast populations, with a subset of ‘preconditioned’ fibroblasts being treated with TGF-β1 to differentiate them into myofibroblasts. After 6 days the cells are passaged and either replated on cell culture plates for 2D studies or used for 3D systems. 3D tissue-engineered models were produced either via scaffoldless self-assembly or collagen 1 hydrogel biofabrication. A low serum culture period subsequently followed, with some groups treated with TGF-β1 over 7 days. The black arrow denotes the location of a tissue construct around the annulus inside an agarose mold; scale bar = 2 mm. (B) Conditions : experimental design is defined by comparisons between culture systems (2D culture plates vs 3D agarose gels/collagen 1-based hydrogels), cellular content (co-culture or monoculture of myoblasts, fibroblasts, or myofibroblasts), and their biochemical treatment in culture (±TGF-β1). The resulting group identification terms code for the conditions investigated in these experiments. Gene expression analysis Total RNA was isolated from cultures using TriZol reagent (Life Technologies, Carlsbad, CA, USA) according to the manufacturers recommended instructions. cDNA was reverse transcribed from 1 µg total RNA using qScript SuperMix (Quanta Biosciences, Gaithersburg, MD, USA) according to manufacturers instructions. qPCR was carried out using a 5 ng equivalent of cDNA in a 1X reaction of PerfeCTa SYBR Green SuperMix (Quanta Biosciences, Beverly, MA, USA) and 250 nM each (forward and reverse) custom oligonucleotide primers (Integrated DNA Technologies, Coralville, IA, USA) generated by using PrimerBlast (National Center for Biotechnology Information, Bethesda, MD, USA). Reactions were carried out on an Eppendorf RealPlex2 qPCR system, and fold changes in gene expression were calculated using the ΔΔCT method using species-specific GAPDH primers. Primer sequences are listed in Table S1. Human collagen 1 and α-SMA expression levels in the 2D FibCon condition were used as the reference group for fibroblast gene expression, and mouse myogenin gene expression levels of 2D MyoCon were used as a reference for myoblasts. N = 5 for all groups. Immunostaining 2D and 3D cultures were assessed qualitatively with immunostaining. 2D cultures were fixed with 4% formaldehyde for 20 min, and 3D cultures were fixed for 2 h. 3D self-assembled tissues were then submerged in OCT media, frozen for cryosectioning, sectioned at 30 μm thick, and applied to glass slides. Hydrogels remained unsectioned for whole-mount imaging. After fixing, 2D and 3D samples were permeabilized with 0. 1% Triton X-100 in PBS for 20 min, and then incubated with a blocking solution consisting of 5% FBS in PBS for 20 min. Following blocking, samples containing myoblasts were stained with rhodamine phalloidin (R415, 1:40 dilution; Life Technologies, Carlsbad, CA, USA) and/or myosin heavy chain (MF-20, 1:50; Developmental Studies Hybridoma Bank, Iowa City, IA, USA), procollagen (M-38, 1:50; Developmental Studies Hybridoma Bank) or α-SMA (A 2547, 1:1, 000; Sigma-Aldrich, St. Louis, MO, USA) primary antibodies. Primary antibody staining was followed by incubation with Alexa-Fluor 488 or 564 secondaries (1:250; Life Technologies, Carlsbad, CA, USA) and DAPI (1:50, 000; Invitrogen, Carlsbad, CA, USA). Imaging of 2D samples was completed with an Olympus IX81 microscope (Olympus corporation of the Americas, Center Valley, PA, USA), 3D self-assembled tissues were imaged with an Olympus Fluoview FV1000 Confocal Microscope, and hydrogels were imaged with an Olympus Fluoview FV1000MPE multiphoton microscope. All images were processed with ImageJ (Version 1. 48; National Institute of Health, Bethesda, MD, USA). The total area of procollagen staining was measured in 2D FibCon and FibTGFβ images and divided by the total number of nuclei to find the average area of procollagen staining per cell. Increased area procollagen signal was extrapolated as a change of in cell shape and increased cell spreading. Circularity of DAPI stained fibroblast nuclei was measured to compare changes in cyto- and nucleoskeletal shape. N = 13 images for each group. 2D MyoCon and CoCon images were also processed to measure changes in myotube alignment. In each image, the myotubes formed angles with a common axes ( x or y ). The standard deviations of the angles from each image were measured to compare the variability in each group, with smaller averaged standard deviations for one sample image indicating higher myotube alignment. N = 8–9 images for each sample. Slack tests The slack test method has been adapted by our lab to indirectly measure and compare tissue integrity, ECM maturity, and muscle fiber development between engineered samples. Two slack test methods were employed. In the whole tissue slack test, self-assembled tissue constructs were released from tension in culture and their shortening lengths were recorded over time. The initial lengths (μm, L i ) of rings were recorded across their longest axis while they remained on their posts. Rings were then removed from the agarose posts, suspended in PBS (37 °C) on a submerged horizontal microbeam from the Microsquisher system (CellScale, Ontario, Canada), and points across their longest axis for 27 min. The final lengths ( L f ) were recorded and percent shortening (%) was calculated with the formula % = 100*( L f / L i ). Tissues were then stored in their respective medias for 24 h at 37 °C and imaged with a stereoscope. N = 3–6 samples per group. The cleaved tissue slack test was performed to release the tissues from internal tension, which may arise from tissue structuring patterns building passive forces. 3D MyoCon, MyoTGFB, CoCon, and CoTGFB tissue rings were cut with a scalpel to generate linear muscle constructs freed from their agarose posts and placed in a 37 °C bath of cell culture media. Tissue lengths were measured every 3 min for 15 min, then samples were allowed to slack over 24 h at 37 °C to give the final length ( L f ). The initial length (μm, L i ) for each construct, the length of the uncut tissue sample still mounted on its post, was calculated as the circumference of the 2 mm agarose posts using the formula: L i = 2 πr. The percent of original size (%) was determined with the formula: % = 100*( L f / L i ). N = 4–5 samples per group. Tissue elastic moduli assessments For calculations of the elastic stiffness of self-assembled engineered tissues, two methods of mechanical characterization were used: a compression-based system to generate slopes from stress–strain curves and atomic force microscopy (AFM). Tissue constructs were sectioned into 4–5 pieces with diameters between 250 and 500 μm and were compressed in a PBS fluid bath (pH 7. 4, 37 °C) using a Microsquisher (Cellscale, Waterloo, Ontario, Canada). Tissues were compressed to 40% of their original diameter at a rate of 1% per second using microbeams with diameters between 0. 2 and 0. 3048 mm. Force–displacement curves were generated and the slope of the linear portion of the curve was extrapolated as the elastic modulus of the tissue. Slopes were averaged for each 3D condition with N = 8–12 section samples per group. Tissue elastic moduli were measured by an Asylum MFP3D-Bio AFM (Asylum Research, Santa Barbara, CA, USA) through a nano-indention method using MFP-3D software (Version 13. 04. 77). Force-distance curves were determined using Eqs. (1) and (2) : (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}F \left( \delta \right) = \frac{4\sqrt{R}}{3} \frac{E}{1-{\nu }^{2}} {\delta }^{3/2}\end{eqnarray*}\end{document} F δ = 4 R 3 E 1 − ν 2 δ 3 ∕ 2 (2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}F \left( \delta \right) = \frac{E}{1-{\nu }^{2}} \frac{2\tan \nolimits \alpha }{\pi } {\delta }^{2}. \end{eqnarray*}\end{document} F δ = E 1 − ν 2 2 tan α π δ 2. Where F is the measured force, E is the local Young’s modulus, R is the cantilever’s tip radius (for a spherical tip), α is the cantilever’s tip angle (for a cone tip), ν is the Poisson’s ratio of the sample (assumed as 0. 5 for an incompressible material), and δ is the sample indentation. Pyramidal tips with a nominal tip radius 20 nm, 200 µm in length, 20 µm in width, and a tip semi-angle of 15°on silicon nitride triangular V-shaped cantilevers with a nominal spring constant of 0. 06 N/m (DNP-10; Bruker Inc. , Camarillo, CA, USA) were employed. The recorded force-distance curves were analyzed in MATLAB and statistical analysis was done using SPSS (Ver. 17. 0; IBM, Somers, NY, USA). Statistics Statistical analyses were performed with two-tailed unpaired t -tests and one-way ANOVAs using GraphPad Prism (Version 4; GraphPad Software) and statistical significance was defined as p < 0. 05. Means are reported with standard error bars in bar graphs. Skew coefficients were calculated in Excel (Version 14. 6. 4; Microsoft). Results Fibroblast to myofibroblast differentiation is observed in 2D cultures treated with TGF-β1, regardless of myoblast presence 2D fibroblast-only cultures displayed increased α-SMA staining in TGF-β1-treated samples ( Figs. 2A and 2B ) and increased α-SMA gene expression ( p < 0. 01, Fig. 3A ), indicating their differentiation to myofibroblasts. There was no significant difference in collagen 1 transcription between any 2D conditions containing fibroblasts or myofibroblasts ( Fig. 3B ), but post-translational control of the protein differed with TGF-β1 treatment. Fibroblasts in the control group were spindle-shaped and displayed smaller areas of procollagen signal per cell nuclei ( Figs. 2C and 2I ), while TGF-β1 treatment significantly increased the area of procollagen staining per nuclei by 32. 2 % ( p < 0. 0001, Figs. 2D and 2I ) and enhanced the secretion of procollagen out of the cell (red arrow in 2D). Nuclei circularity was also increased in 2D FibTGFβ samples ( p < 0. 0001, Fig. 2J ), indicating a widening of the nucleoskeleton accompanying increased cell spreading, compared to the more elongated nuclear shape in control conditions. Enhanced cell proliferation was also observed with TGF-β1 treatment ( p < 0. 0001, Fig. 2K ), as recorded by average number of nuclei per field of view. α-SMA gene expression in 2D CoCon samples was similar to transcript levels of the fibroblast-only control group, but was significantly upregulated in 2D CoTGFβ conditions ( p < 0. 01, Fig. 3A ). Myogenesis is suppressed in 2D cultures with TGF-β1 supplementation and in the presence of fibroblasts and myofibroblasts, yet fibroblasts increase myotube alignment Robust myotube formation was present in 2D MyoCon samples ( Fig. 2E ), decreased with exposure to TGF-β1 ( Fig. 2F ) or in co-culture with fibroblasts ( Fig. 2G ), and completely inhibited in myofibroblast co-cultures supplemented with TGF-β1 ( Fig. 2H ). MYOG gene expression in the 2D MyoTGFβ and 2D CoCon conditions were similar to each other and both displayed significant MYOG downregulation in comparison to 2D MyoCon. 2D co-cultures exposed to TGF-β1 showed the lowest MYOG expression of any group ( p < 0. 01, Fig. 3C ). Despite diminishing myogenesis, fibroblasts organized myotube formation by increasing their alignment in 2D CoCon samples ( p < 0. 001, Fig. 2L ). 10. 7717/peerj. 4939/fig-2 Figure 2 Immunostaining and morphological characterization of 2D cultures. 2D FibCon (A) and FibTGFβ (B) stained for α-SMA and DAPI, and FibCon (C) and FibTGFβ (D) stained for procollagen. 2D MyoCon (E), MyoTGFβ (F), CoCon (G), and CoTGFβ (H) conditions stained for myosin heavy chain (MyHC) and DAPI. Scale bars = 100 µm for all images. Quantitative analysis of average area of procollagen signal/average number of nuclei per field of view (I), nuclei circularity (J), and average number of total nuclei per field of view (K) in FibCon and FibTGFβ samples. Quantitative analysis of myotube alignment (L) in MyoCon and CoCon cultures, reported as the standard deviation of myotube alignment angles. Graphs display group averages with standard error bars. 10. 7717/peerj. 4939/fig-3 Figure 3 Gene expression of fibroblast collagen 1 and α-SMA and myoblast myogenin in 2D cell cultures and 3D self-assembled tissue conditions. Species-specific gene expression for all groups is normalized to gene expression for group (a), 2D FibCon or MyoCon samples. Average fold change values are reported with standard error bars. Significance values were defined at p < 0. 01. (A) α-SMA: (a) is significantly different from (b), (d), (e), (f), (g), and (h); (b) is significantly different from (c), (e), (f), (g), and (h); (c) is significantly different from (e), (f), (g), and (h); (d) is significantly different from (e), (f), (g), and (h); (e) is significantly different from (g) and (h); and (f) is significantly different from (g) and (h). (B) COL1: (a) is significantly different from (e), (f), and (h); (b) is significantly different from (c), (e), (f), and (h); (c) is significantly different from (e), (f), and (h); (d) is significantly different from (e), (f), and (h); (e) is significantly different from (g) and (h); (f) is significantly different from (g) and (h); and (g) is significantly different from (h). (C) MYOG: (a) is significantly different from (b), (c), (d), (e), (g), and (h); (b) is significantly different from (d), (f), (g), and (h); (c) is significantly different from (d), (f), and (h); (d) is significantly different from (e), (f), (g), and (h); (e) is significantly different from (f) and (h); (f) is significantly different from (g) and (h); and (g) is significantly different from (h). In comparison to fibroblasts grown on plastic, self-assembled tissues containing only fibroblasts have a suppressed ability to assume a myofibroblast phenotype In comparison to 2D fibroblast-only controls, expression of procollagen was dramatically downregulated in fibroblast-only tissues, and this was not improved by TGF-β1 exposure ( p < 0. 01, Fig. 3B ). α-SMA staining was absent in 3D FibCon samples ( Fig. S1A ), but was somewhat increased with TGF-β1 treatment ( Fig. S1C ). 3D FibCon and FibTGFβ α-SMA gene expression was transcriptionally downregulated to an even greater degree than procollagen, yet these values were not significantly different from each other ( p < 0. 01, Fig. 3A ). Fibroblast/myofibroblast-only tissues were quite fragile; they would easily rupture during handling (arrow, Fig. 4D ) and would sometimes collapse, losing their annular shape ( Figs. 4C and 4F ). Despite their fragility, their surfaces were smooth and homogenous ( Fig. 4S ). However, the average thickness of constructs containing myofibroblasts treated with TGF-β1 was 33. 2% larger than respective controls ( p < 0. 01, Fig. 5F ), similar to the myofibroblast hypertrophy seen in 2D FibTGFβ. 10. 7717/peerj. 4939/fig-4 Figure 4 Zero force shortening of self-assembled tissues and their surface characteristics. Tissue shortening was assessed when removed from agarose molds after 3 min, 27 min, and 24 h. 3D FibCon (A–C), FibTGFβ (D–F), MyoCon (G–I), MyoTGFβ (J–L), CoCon (M–O), and CoTGFβ tissue rings (P–R), with scale bars = 1, 000 µm. The filled arrow in (D) indicates a structural rupture in the tissue, and the black arrow in (M) indicates thinned myoblast-free region of sample. Higher magnification images of FibCon (S), MyoTGFβ (T), and CoTGFβ tissues (U) show surface texture. In (T), the filled arrow is a large nodule on a MyoTGFβ sample, and unfilled arrows indicate smaller syncytium; scale bars = 300 µm. 10. 7717/peerj. 4939/fig-5 Figure 5 Biomechanical and elastic properties of 3D tissue constructs. (A) An agarose gel mold with central post and tissue of tissue skewered onto a microbeam from the Microsquisher system. The white arrow denotes the longest axis of the tissue constructs used to track slacking from initial length in the whole tissue slack test, scale bar = 1, 000 μm. (B) Percent (%) of initial length of the longest axis of tissue samples in the whole tissue slack test 27 min after their removal from agarose gels. All groups shortened from their original size significantly ( p < 0. 001), with significant differences between each group ( p < 0. 05), except for (a) and (b). (C) The shortening length of tissue samples over 3 min intervals in the cleaved tissue slack test. Note the final data point is at 24 hours, and the Y -axis origin is 2 mm. (D) Image of the Microsquisher system used to obtain elastic’s modulus of tissue sections (white asterisk); scale bar = 300 μm. (E) Percent (%) of initial length of tissue samples in the cleaved tissue slack test after 24 h. All groups shortened from their original size significantly ( p < 0. 001), with significant differences between each group ( p < 0. 05), except for (b) and (c). (F) Initial thickness of tissue samples, asterisks indicate p < 0. 001. (G) Screenshot of AFM cantilever and sample (white asterisk), scale bar = 100 μm. (H) Young’s modulus of tissues generated from Microsquisher force-displacement data. (a) is significantly softer than (e) and (f); (b) is significantly softer than (c), (d), (e), and (f); and (c) and (d) are significantly softer than (f), with p < 0. 001. (I) Young’s modulus of MyoCon, CoCon, and CoTGFβ tissues calculated by AFM, asterisks indicate p < 0. 01. Graphs display group averages with standard error bars. TGF-β1 supplementation improves myoblast differentiation and alignment in myoblast-only self-assembled tissues 3D myoblast-only groups differed in their degree of myogenesis and their myotube formation patterns. 3D MyoCon contained myotubes that had some degree of organization but generally weren’t aligned ( Figs. 6A and 6B ), whereas myoblast samples treated with TGF-β1 had myotubes that were aligned along their circumferential axis (white filled arrow, Figs. 6E and 6F ). Gene expression profiles indicated that 3D MyoCon samples had significantly lower myogenin expression than both the 2D MyoCon control and 3D MyoTGFβ condition ( p < 0. 01, Fig. 3C ), but the 2D control and 3D TGFβ treated condition were not significantly different from each other. α-SMA was present in both untreated ( Figs. 6C and 6D ) and TGF-β1 treated ( Figs. 6G and 6H ) myoblast constructs. Observable in the α-SMA stained sections are fissures and breaks in the tissue (white asterisks, Figs. 6C and 6G ). Additionally, the surfaces of myoblast-only tissues presented nodular syncytium of fused myoblasts lacking the anchorage necessary to elongate and form myotubes (arrow, Figs. 6I – 6K ), and this was macroscopically observable (unfilled arrows, Fig. 4T ). Heterogeneous tissue patterning is also seen here, with many ripples along the tissue surface, large budges (filled arrow, Fig. 4T ), and regions of varying density. 10. 7717/peerj. 4939/fig-6 Figure 6 Histology of self-assembled myoblast monocultures. 3D MyoCon (A–D, I–K) and MyoTGFβ constructs (E-H) are displayed with myosin heavy chain (A, E, I), MyHC merged with DAPI (B, F, K), and DAPI alone (J). α-SMA (C, G) and α-SMA merged with DAPI (D, H). White filled arrows indicate direction of alignment and anisotropy of myotubes (E), while unfilled arrows indicate unanchored syncytium (A, I), and white asterisks identify regions of tissue breakage (C, G), scale bars = 100 µm. In self-assembled co-cultures, fibroblasts and myofibroblasts homogenize tissue surfaces, myofibroblasts improve myotube formation, and TGF-β1 enhances myogenesis Addition of fibroblasts or myofibroblasts smoothened the surface of co-culture tissues ( Fig. 4U ), similarly to fibroblast- or myofibroblast-only samples, and these constructs lacked the studding of nodular syncytium seen in myoblast-only tissues. Co-culture conditions ( Figs. 4M and 4P ) appeared visually denser than their fibroblast- ( Figs. 4A and 4D ) and myoblast-only ( Figs. 4G and 4J ) counterparts. Similar to the hypertrophy observed in 2D and 3D myofibroblast conditions treated with TGF-β1, CoTGFβ tissues were 27. 9% thicker than CoCon ( p < 0. 01, Fig. 5F ). While human fibroblast/myofibroblast α-SMA gene expression was decreased in all 3D conditions compared to the 2D FibCon group, it was significantly higher in 3D CoCon and CoTGFβ compared to 3D fibroblast-only tissues ( p < 0. 01, Fig. 3A ), and co-culture transcription levels were not influenced by TGF-β1 supplementation. A non-species specific α-SMA antibody was visible within myotubes, where CoCon samples showed α-SMA within punctate and short myotubes ( Fig. 7A ), while those in CoTGFβ tissues were elongated and more mature ( Fig. 7C ), indicating that TGF-β1 facilitated the development of bundles of myotubes aligned in parallel. Additionally, collagen 1 expression was increased in 3D co-cultures compared to 3D fibroblast/myofibroblast-only cultures ( p < 0. 01, Fig. 3B ), but CoCon constructs were not significantly different from the 2D FibCon condition. Interestingly, collagen 1 gene expression was highest in 3D CoTGFβ samples, which was also the condition of highest MYOG expression ( p < 0. 01). 10. 7717/peerj. 4939/fig-7 Figure 7 Histology of self-assembled co-cultures. 3D CoCon (A, B, E, F, H, I) and CoTGFβ tissues (C, D, G, J, K). α-SMA (A, C) and α-SMA merged with DAPI (B, D). Myosin heavy chain staining of CoCon constructs (E, H), f-actin staining of E (F), and DAPI staining of H (I). MyHC staining of CoTGFβ tissues (G, J, K) with DAPI and f-actin staining (J, K). (J) is an inset of G. White arrows indicate cross sections of myotubes observed with MyHC and actin staining, while unfilled arrows show elongated myotubes fused along circumferential axis of tissue rings, scale bar = 100 µm. 3D CoCon constructs containing fibroblasts frequently presented dense bulges with thickened regions of cells composed of differentiating myoblasts ( Figs. 7E and 7F ) between thinner regions absent of MyHC staining ( Figs. 7H and 7I ); this heterogeneous tissue architecture can also be seen macroscopically (arrow, Fig. 4M ). CoCon constructs contained myotubes that were short but numerous. Some myotubes fused from one wall of the tissue to another, so that during sectioning, cross sections of myotubes are apparent (filled arrows, Figs. 7E and 7F ). In contrast, CoTGFβ samples containing myofibroblasts had myotubes that were thicker and more multinucleated ( Figs. 7G, 7J and 7K ). These tissues also had many cross-sections of myotubes (filled arrow, Fig. 7J ), but additionally contained much longer circumferentially aligned myotubes than in control co-culture samples (unfilled arrow, Fig. 7G ). MYOG expression in 3D CoTGFβ constructs was significantly higher than all other groups ( p < 0. 01, Fig. 3C ), while MYOG expression in 3D CoCon constructs was suppressed compared to 2D MyoCon ( p < 0. 01), and was not significantly different from 3D MyoCon. This data demonstrates that addition of fibroblasts with myoblasts in 3D co-culture does not improve myogenesis beyond what is observed in cultures consisting solely of myoblasts, and TGF-β1 supplementation is a myogenesis-promoting factor. Zero-force velocity is greatest in co-culture with TGF-β1 supplementation in slack tests Slack tests are a measurement of zero force velocity tissue shortening resulting from the recoil of stretched elastic elements and myofibrillar filament sliding in sarcomeres. Slack tests assess acto-myosin kinetics that are independent of Ca 2+ activation, which provides biomechanical information about extracellular matrix networking, MHC isoform and muscle fiber type, and muscle organ characteristics ( Claflin & Faulkner, 1985 ; Josephson & Edman, 1998 ; Reggiani, 2007 ). In whole tissue slack tests, the initial maximum lengths of self-assembled tissue constructs were recorded prior to removal from their posts (white arrow, Fig. 5A ). Constructs were then released from tension ( Figs. 4A, 4D, 4G, 4J, 4M and 4P ), their lengths measured after 27 min ( Figs. 4B, 4E, 4H, 4K, 4N and 4Q ), and allowed to slack over 24 h ( Figs. 4C, 4F, 4I, 4L, 4O and 4R ). While all groups shortened significantly from their original size after 27 min ( p < 0. 001, Fig. 5B ), tissues containing only fibroblasts or myofibroblasts shortened the least, with 3D FibCon remaining 87. 3% and FibTGFβ 85. 4% their original size. Fibroblast/myofibroblast tissues were the only cell-content matched tissue group that did not display significantly different final length values due to TGF-β 1 treatment. Myoblast tissues underwent significantly greater shortening than fibroblast/myofibroblast samples: MyoCon rings shrank to 69. 4% and MyoTGFβ to 62. 2% their initial diameter. Co-culture tissues underwent the greatest alteration, however. CoCon samples had the most significant shortening compared to all other groups (48. 7%), and CoTGFβ was the second most shortened group (55. 1%). Interestingly, after 24 h, CoTGFβ tissues ( Fig. 4R ) showed a greater degree of sample shortening than CoCon ( Fig. 4O ). Since reference points for tracking changes in length were not conserved overnight in ring-shaped tissues, a cleaved tissue slack test was performed by cutting tissues containing C2C12s into linear constructs and plotting slacking lengths over time ( Fig. 5C ). Most of the length shortening occurred within the first 3 min and slowed considerably by 9 min, with co-culture samples displaying the highest velocities. After 24 h, 3D MyoCon samples shortened to 53. 2% of their initial length, MyoTGFB to 49. 8%, CoCon to 47. 2%, and CoTGFB to 42. 7% ( Fig. 5E ). Each experimental group was significantly different from the others ( p < 0. 01), except for MyoTGFB and CoCon. The slack tests demonstrated that while TGF-β1 exposure for fibroblast-only samples did not lead to significantly more zero-force shortening in comparison to untreated controls, TGF-β1 positively impacted shortening of myotubes within myoblast-containing tissues, and the interactions between TGF-β1-treated myofibroblasts and myoblasts promoted the greatest degree of shortening over time. Tissue elastic moduli increase in co-cultures and with TGF-β1 exposure Using the parallel plate compression system, the Young’s modulus for sectioned self-assembled tissues was determined to be 2. 04 kPa for FibCon, 0. 81 kPa for FibTGFβ, 3. 32 kPa for MyoCon, 3. 86 kPa for MyoTGFβ, 4. 47 kPa for CoCon, and 6. 55 kPa for CoTGFβ ( Fig. 5H ). FibCon and FibTGFβ tissues were significantly softer than CoCon and CoTGFβ ( p < 0. 001). While MyoCon and MyoTGFβ were also significantly softer than CoTGFβ ( p < 0. 001), TGF-β1 did not significantly increase stiffness of samples containing only myoblasts. Similarly, while TGF-β1 treatment increased the stiffness of CoTGFβ, it was not a significant increase with respect to CoCon. Consequently, AFM was used to increase the sensitivity of measurements of elasticity at the tissue surface. AFM generated Young’s moduli of 2. 66 kPa for MyoCon, 3. 64 kPa for CoCon, and 6. 02 kPa for CoTGFβ, with all groups being significantly different from one another ( p < 0. 01, Fig. 5I ). Supplementation with TGF-β1 and addition of myofibroblasts enhances myogenesis in collagen 1-based hydrogels MyoCon, MyoTGFβ, CoCon, and CoTGFβ hydrogels were imaged for myosin heavy chain. In the myoblast control condition lacking TGF-β1 supplementation, some MyHC was present in unfused mononuclear cells, but observable myotube formation was absent ( Fig. 8A ). In contrast, TGF-β1-treated myoblast hydrogels showed robust myotube formation, with somewhat inconsistent alignment orientations ( Fig. 8B ). Without supplementation with TGF-β1, co-cultures with fibroblasts and myoblasts showed small and immature myotube-shaped cells ( Fig. 8C ), in contrast to its TGF-β1-treated counterpart containing myofibroblasts ( Fig. 8D ). CoTGFβ hydrogels yielded the thickest and longest muscle fibers of any condition and possessed the highest degree of multinucleation. These muscle fibers also demonstrated the best alignment. CoTGFβ hydrogels were also more visibly contracted in their culture wells than other conditions ( Fig. 8E ), although MyoTGFβ hydrogels displayed a lesser degree of contraction. Contraction of MyoCon and CoCon hydrogels was not observed. 10. 7717/peerj. 4939/fig-8 Figure 8 Histology of mono- and co-cultures in collagen 1-based hydrogels. Myosin heavy chain staining of MyoCon (A), MyoTGFβ (B), CoCon (C), and CoTGFβ (D) hydrogels. Scale bar = 100 µm. (E) are images of the hydrogels from culture wells of a 6-well plate, in descending order: MyoCon, MyoTGFβ, CoCon, and CoTGFβ hydrogels. Discussion These experiments investigated the impact of tissue remodeling and ECM deposition on the biomechanical tissue niche; TGF-β1 signaling; and cell–cell signaling between fibroblasts, myofibroblasts, and myoblasts on in vitro simulations of myogenesis. Our data demonstrates that transitioning from 2D to 3D culture systems has a profound impact on gene and protein expression involved in mechanotransduction. Consistent with other literature, TGF-β1 differentiated fibroblasts to myofibroblasts in 2D cultures containing only fibroblasts ( Dahl, Ribeiro & Lammerding, 2008 ; Dong et al. , 2013 ; Sandbo & Dulin, 2011 ). This was accompanied by an upregulation of α-SMA gene and protein expression. While myofibroblasts in 2D did not show upregulated transcription levels of collagen 1, post-transcriptional control mechanisms increased collagen translation and secretion into the extracellular space. Steady collagen gene expression levels in 2D may be maintained by mechanotransduction signaling, which was altered in the equivalent 3D self-assembled tissue model, and resulted in radical suppression of collagen 1 and α-SMA transcription. Additionally, myoblasts have been reported to assume a myofibroblast-like phenotype with TGF-β1 exposure in 2D ( Cencetti et al. , 2010 ; Charge & Rudnicki, 2004 ; Filvaroff, Ebner & Derynck, 1994 ; Li et al. , 2004 ), and our TGF-β1 treated 2D cultures showed similar decreases in myogenecity in monocultures, with the myofibroblast co-culture displaying the most downregulated myogenin expression of any group. In vivo, however, while muscle fiber regeneration can be inhibited by exogenous TGF-β1 ( Filvaroff, Ebner & Derynck, 1994 ; Mendias et al. , 2012 ), endogenous TGF-β1 signaling plays an important role in myogenesis. Genetically truncating the type II TGF-β receptor in myoblasts arrests their ability to differentiate in vivo, indicating that TGF-β1 signaling contributes to muscle development ( Karalaki et al. , 2009 ; Myhre & Pilgrim, 2012 ). In agreement with these studies, we found that TGF-β1 supplementation improved myogenesis in our 3D mono- and co-culture muscle models, a reversal of the observations in 2D cultures. The mechanism for TGF-β1-mediated increases in myogenic differentiation and contractility highlights a differential effect of TGF-β1 on myofibrillogenesis in 2D and 3D systems. Myofibrillogenesis initiates sarcomere development by increasing the dense packing of contractile and cytoskeletal proteins in muscle cells. During myofibrillogenesis, myofibrils are assembled from the framework of the actin cytoskeleton and undergo maturation by successive incorporation of proteins with increasing contractility. This process is driven by transmitting mechanical tension in the actin cytoskeleton to surrounding ECM through costameres ( Sanger et al. , 2010 ) and results in the development and alignment of sarcomeres ( Weist et al. , 2013 ). TGF-β receptor activation increases cytoskeletal polymerization of α-SMA stress fibers in 2D culture platforms ( Filvaroff, Ebner & Derynck, 1994 ), but also simultaneously has an inhibiting effect on myogenesis. However, since α-SMA is a precursor protein to sarcomeric actin in myofibrillogenesis, moderate TGF-β1 levels in 3D may accelerate myofibrillogenesis through an α-SMA-mediated upregulation mechanism that could increase the tensional forces cells exert on their environment. In 3D, fibroblasts and myofibroblasts benefit muscle regeneration through depositing ECM, organizing myoblast differentiation and myotube fusion, and stimulating myofibrillogenesis ( Sanger et al. , 2010 ; Turrina, Martinez-Gonzalez & Stecco, 2013 ). Myotube formation dynamically remodels the biomechanical environment within muscle to further attenuate fibroblast and myofibroblast activity. In vitro cell traction studies have shown that myotubes exert five to eight times greater traction forces on their substrates than fibroblasts, and these forces increase as myotubes mature ( Dahl, Ribeiro & Lammerding, 2008 ). Contracting myotubes transmit their forces to surrounding ECM through costameres and exert tensional forces throughout the muscle organ to facilitate contraction ( Costa, 2014 ). This mechanism of force transmission could explain the greater recoil and shortening capacity co-culture tissues in contrast to myoblast-only samples with a suppressed ability to disperse forces to the surrounding ECM. Lower levels of supportive ECM for myotubes in self-assembled monocultures decreases the availability for integrin connections to the tissue microenvironment and reduces tensional forces applied to surrounding structures, resulting in less elastic recoil after being released from tension. Additionally, our data showing lower elastic moduli in 3D myoblast monocultures than co-cultures is congruent with the findings of Meyer and Lieber, who found the elastic stiffness of skeletal muscle fibers to be significantly dependent on the presence of ECM. In their study, dissected muscle fibers stripped of their extracellular matrix sheaths had four-fold lower elastic moduli than fascicles that retained their ECM ( Meyer & Lieber, 2011 ). 3D fibroblast and myofibroblast co-cultures were distinct in their morphology, force transmission capacities, and biomechanical properties. The bundling of highly multinucleated myotubes in parallel within CoTGFβ constructs was more similar to in vivo muscle fascicle morphology than fibroblast co-cultures, and was accompanied by the greatest upregulation in myogenin expression of all groups. Enhanced muscle fiber maturation and patterning likely increased the magnitude of zero-force shortening in CoTGFβ samples due to acto-myosin kinetics. This compliments a study by Larkin et al. , who found that TGF-β1 treatment of heterogeneous fibroblast and satellite cell populations in engineered muscle constructs increased their force generation capacities during electrical stimulation ( Weist et al. , 2013 ). However, this study did not investigate the differentiation of fibroblasts to myofibroblasts within a co-culture system ( Weist et al. , 2013 ). Our experiments identify that TGF-β1 has both individual and synergistic benefits to skeletal muscle engineering. These data show that TGF-β1 treatment in 3D monocultures of myoblasts boosts myoblast fusion and differentiation. In co-cultures with myoblasts, TGF-β1 activates fibroblasts to differentiate into myofibroblasts, which are better accelerators of myogenesis than fibroblasts, due in part to their enhanced ECM synthesis. The tenacity of myofibroblasts in co-culture is demonstrated through the highest collagen 1 gene expression of all conditions, where collagen 1 reinforcement likely improved tissue integrity and stiffness. Here we have demonstrated a robust synergistic effect between the incorporation of myofibroblasts, not fibroblasts, and TGF-β1 supplementation in tissue-engineered models of skeletal muscle myogenesis. Conclusions These experiments have demonstrated that fibroblasts and myofibroblasts are not interchangeable cell types in in vitro models. To the author’s knowledge, we are the first to report that myofibroblasts enhance the outcome of myogenesis to a greater degree than fibroblasts when in co-culture with myogenic cells in a tissue engineered model of skeletal muscle. Additionally, our comparison between 2D and 3D studies highlights the importance of using a culture model that appropriately recapitulates the in vivo environment. In literature, many mechanistic, functional, and differentiation studies of myogenic cells treated with TGF-β1 performed in 2D lead to erroneous conclusions about the impact of TGF-β1 on myogenesis. We are the first to methodologically describe this distinction in literature, as well as show that TGF-β1 signaling in 3D plays an innate role in promoting skeletal muscle differentiation in monocultures of myogenic cells. The tissue-engineered TGF-β1-treated myoblast and myofibroblast co-culture model is a promising candidate for therapeutic treatment of volumetric muscle loss and can be employed as a screening tool for pharmaceutical treatments that support muscle regeneration. Additionally, this model can be used to improve product quality of in vitro meat, which has experienced accelerated interest in recent years ( Langelaana et al. , 2010 ; Post, 2012 ). Because of growing awareness of the livestock sector’s contribution to worsening climate change, deforestation of the rain forests, reductions in biodiversity, environmental degradation ( Steinfeld et al. , 2006 ), and increased bacterial resistance to antibiotics ( Frieden, 2013 ), devising a sustainable solution for biomanufactured meat production is imperative. In vitro meat is being recognized as a sustainable alternative to traditional factory farming as it uses less food, water, and land resources; produces less greenhouse gases; requires less energy for production ( Tuomisto & de Mattos, 2011 ); and does not require systemic overuse of antibiotics. Since meat quality is dependent, in part, on sarcomere development, muscle fiber type, and connective tissue content ( Joo et al. , 2013 ), including TGF-β1 as a cell culture media supplement and co-culturing muscle cells with myofibroblasts can produce meat with flavor and texture increasingly similar to that derived from livestock. Although our self-assembled CoTGFβ model of skeletal muscle achieved structural similarity to muscle fascicles, utilizing this self-assembly strategy for scale up of tissue engineering is not practical. Here we have shown that the CoTGFβ technique can be used in hydrogel-based systems, which possess flexibility in design parameters to incorporate vasculature ( Koffler et al. , 2011 ; Levenberg et al. , 2005 ). Additionally, electrical stimulation of muscle fibers further matures in vitro skeletal muscle models by accelerating sarcomere development ( Langelaan et al. , 2011 ) and increasing contraction force ( Fuoco et al. , 2015 ; Ito et al. , 2014 ). Accordingly, applying vascularization and electrical stimulation regimens to TGF-β1-treated myofibroblast and myoblast co-cultures shows promise to improve engineered skeletal muscle’s recapitulation of native muscle. Supplemental Information 10. 7717/peerj. 4939/supp-1 Supplemental Information 1 Fig 2I raw data—procollagen area/nuclei This spreadsheet contains the total area (um 2 ) of procollagen staining and total number of nuclei in 13 images of 2D FibCon and 2D FibTGFB each, as measured in ImageJ. The total area was divided by the number of nuclei for each image to give an average value of procollagen staining area per nuclei. Click here for additional data file. 10. 7717/peerj. 4939/supp-2 Supplemental Information 2 Fig 2J raw data—nuclei circularity The spreadsheet contains the average circularity values of the nuclei for 2D FibCon and 2D FibTGFB conditions in 13 images, respectively. A value of 1 denotes a perfect circle, and value of 0 denotes a line, or completely noncircular object. Click here for additional data file. 10. 7717/peerj. 4939/supp-3 Supplemental Information 3 Fig 2K raw data—total nuclei This dataset includes the total nuclei count for 13 images each of 2D FibCon and 2D FibTGFB samples and the calculated average number of nuclei per image. Click here for additional data file. 10. 7717/peerj. 4939/supp-4 Supplemental Information 4 Fig 2L raw data—myotube angle This spreadsheet contained the angle ( θ ) formed between the x axis of an image and each myotube within that image. The SD of all θ measurements was calculated and averaged for 8 images each of 2D MyoCon and 2D CoCon conditions to assess the magnitude of alignment of myotubes. Click here for additional data file. 10. 7717/peerj. 4939/supp-5 Supplemental Information 5 Fig 3A-C raw data—qPCR This dataset contains calculated fold changes in gene expression for three genes: human collagen 1, human alpha-smooth muscle actin, and mouse myogenin. The ΔΔCT method was employed to calculate fold changes from critical threshold (CT) values, and normalized to GAPDH housekeeping genes in each species. Click here for additional data file. 10. 7717/peerj. 4939/supp-6 Supplemental Information 6 Fig 5B raw data—% shortening in whole tissue slack tests This spreadsheet notes the shortening length (mm) of 3D samples during whole tissue slack tests after 27 minutes to calculate the total % of tissue shrinkage from their original size. Click here for additional data file. 10. 7717/peerj. 4939/supp-7 Supplemental Information 7 Fig 5C and 5E raw data—rate of shortening over 24 hours and % shortening in cleaved tissue slack tests This spreadsheet notes the shortening length (mm) of 3D samples during cleaved tissue slack tests over time (min) and the calculation of total % of tissue shrinkage from their original size after 24 h. Click here for additional data file. 10. 7717/peerj. 4939/supp-8 Supplemental Information 8 Fig 5F raw data—tissue thickness This spreadsheet contains the recorded thickness (um) of tissue samples. Click here for additional data file. 10. 7717/peerj. 4939/supp-9 Supplemental Information 9 Fig 5H raw data—Young’s modulus from compression This spreadsheet contains force and displacement data from compression tests, and the extrapolated Young’s modulus (Pa) values. Click here for additional data file. 10. 7717/peerj. 4939/supp-10 Supplemental Information 10 Fig 5I raw data—Young’s modulus from AFM This spreadsheet contains Young’s modulus values for engineered tissues via AFM testing. Click here for additional data file. 10. 7717/peerj. 4939/supp-11 Figure S1 Histology for self-assembled fibroblast monocultures 3D FibCon (A, B) and FibTGFβ (C, D) tissues with α-SMA staining (A, C) and α-SMA merged with DAPI (B, D). Scale bars = 100 μ m, 30 μ m thick sections. Click here for additional data file. 10. 7717/peerj. 4939/supp-12 Table S1 Primer sequences for quantitative PCR Human and mouse GAPDH were used as reference genes for species-specific gene expression. Click here for additional data file.

10. 7717/peerj. 5805

PeerJ

Recent advances and challenges on application of tissue engineering for treatment of congenital heart disease

Congenital heart disease (CHD) affects a considerable number of children and adults worldwide. This implicates not only developmental disorders, high mortality, and reduced quality of life but also, high costs for the healthcare systems. CHD refers to a variety of heart and vascular malformations which could be very challenging to reconstruct the malformed region surgically, especially when the patient is an infant or a child. Advanced technology and research have offered a better mechanistic insight on the impact of CHD in the heart and vascular system of infants, children, and adults and identified potential therapeutic solutions. Many artificial materials and devices have been used for cardiovascular surgery. Surgeons and the medical industry created and evolved the ball valves to the carbon-based leaflet valves and introduced bioprosthesis as an alternative. However, with research further progressing, contracting tissue has been developed in laboratories and tissue engineering (TE) could represent a revolutionary answer for CHD surgery. Development of engineered tissue for cardiac and aortic reconstruction for developing bodies of infants and children can be very challenging. Nevertheless, using acellular scaffolds, allograft, xenografts, and autografts is already very common. Seeding of cells on surface and within scaffold is a key challenging factor for use of the above. The use of different types of stem cells has been investigated and proven to be suitable for tissue engineering. They are the most promising source of cells for heart reconstruction in a developing body, even for adults. Some stem cell types are more effective than others, with some disadvantages which may be eliminated in the future.

Introduction Congenital heart diseases (CHD) refer to the abnormal formation of the infant’s heart, great thoracic vessels and heart valves during intra-uterine development ( National Congenital Heart Disease Audit Report 2012–2015, 2016 ). CHD is different from the acquired heart diseases which occur because of lifestyle or aging ( British Heart Foundation, 2016 ). The abnormalities are structural defects, such as valve defects, intravascular or intracardial stenosis, congenital arrhythmias or cardiomyopathies which greatly affect the early and future life of a CHD patient ( National Congenital Heart Disease Audit Report 2012–2015, 2016 ; Okudo & Benson, 2001 ). People born with CHD need immediate medical care after birth which further continues throughout their lives. In 2010, it was estimated that only in the USA approximately 2. 4 million people suffered from CHD and more than half of them were adults ( Gilboa et al. , 2016 ). In Europe, for instance, for the period of 2000 to 2005, about 36, 000 live births per year were diagnosed with CHD ( Dolk, Loane & Garne, 2011 ). The number grows bigger when including the unborn children who were diagnosed with CHD and died either by termination of pregnancy or by intrauterine death or even neonate death ( Dolk, Loane & Garne, 2011 ). In the UK, about 8 in every 1, 000 live babies born have a heart or circulatory condition ( National Congenital Heart Disease Audit Report 2012–2015, 2016 ). Some estimate those numbers to be higher and, commonly, percentages of each type of CHD change depend on the geographical area of investigation ( Hoffman & Kaplan, 2002 ; Van der Linde et al. , 2011 ), Table 1 further mentions other CHD frequencies for other countries. CHDs not only have an effect on the individual’s and their family’s lives but also have a huge financial impact on the healthcare system. According to NHS England for the financial year 2013/14 the total spent on CHD was £175 million ( NHS England, 2015 ) and in the US the total cost for CHD treatment in 2008 was approximately $298 billion ( Lundberg, 2013 ). In general, the number of children and adults being diagnosed with CHD increases due to the improved technology of diagnostic tools ( Hoffman & Kaplan, 2002 ). 10. 7717/peerj. 5805/table-1 Table 1 Frequencies of CHDs in some regions. United States Affects 1% of live births ( Krasuski & Bashore, 2016 ) South America Colombia: 1. 2 per 1, 000 live births Brazil (Minas Gerais): 9. 58 in 1, 000 live births Brazil (Londrina): 5. 49 in 1, 000 live births ( Pedra et al. , 2009 ) Mexico Affects 6–8 per 1, 000 newborns. Drawing to the conclusion that there about 12, 000 or 16, 000 babies living with CHD ( Calderón-Colmenero et al. , 2013 ) Asia Affects 9. 3 per 1, 000 live births ( Van der Linde et al. , 2011 ) Europe Affects 8. 2 per 100 live births ( Van der Linde et al. , 2011 ) United kingdom Affects about 9 in every 1, 000 babies ( NHS, 2018 ) Russia Affects 2. 7–3. 8 per 1, 000 newborns estimating as 86 newborns per year being affected with CHD ( Postoev, Talykova & Vaktskjold, 2014 ) Australia Affects 8–10 cases per 1, 000 live births. Resulting in 2, 400–3, 000 newborns with CHD each year. About 65, 000 adults are living with CHD ( HeartKids, 2018 ) Africa Mozambique: 2. 3 in 1, 000 live births Northern Nigeria: 9. 3% (122 of 1, 312 patients) ( Zühlke, Mirabel & Marijon, 2013 ) Canada Affects 1 in 80–100 live births ( Canadian Heart Alliance, 2018 ) CHD can be diagnosed using transabdominal fetal Doppler echocardiography. Such prognostic protocols are performed in high-risk groups of pregnant women, like those with a family history of CHD ( Nayak et al. , 2016 ). In adults with CHD, the most effective diagnostic practice is transesophageal echocardiography, electrocardiogram, pulse oximetry, X-rays, cardiac catheterization and MRI ( Sun et al. , 2015 ). The CHDs are managed by surgery, and the efficiency of this approach is largely dependent on the materials which are used during the surgery. These materials are expected to be close to the native cardiac tissue in both structure and function. In structures, CHD could present extremely complicated malformations which cannot be spontaneously or by singular surgical procedure reconstructed, hence the dire need for more research into biomaterials for Tissue Engineering (TE). The recent extensive research focuses on possible ways to fabricate a near ideal tissue. So far, TE appears to be the way forward in creating ideal tissue that can probably mimic the native heart tissue both in structure and function. TE refers to creation of functional three-dimensional tissue using biomaterials and cells for replacement or restoration of damaged organs and/or parts of them. TE is the most promising approach at the present for CHD, as treatment can be “patient-specific” and the engineered tissue could adjust to the developing body of the recipient. Many would think that TE is an idea which conceived and developed in a very recent past. However, it has been proven that tissue regeneration and TE is a concept which was born thousands of years ago, and it has inspired Greek mythology, history, arts, and religion. In art, religion inspired the well-known painting of “Healing of Justinian” based on the miracle of St. Cosmas and St. Damian, physicians and Christian martyrs who appear to have transplanted the leg of an Ethiopian to the body of a patient ( Durant, 2018 ). The closest to an artificial replacement of a body part was discovered in Egypt on a mummy which had a wooden replacement of the hallux ( Finch, 2011 ). However, today, TE involves a combination of creating scaffolds and cell seeding. With regards to the heart, the most commonly used and known artificial parts are the mechanical heart valves and conduits ( Zilla et al. , 2008 ; Gott, Alejo & Cameron, 2003 ). The first artificial heart valve was placed on live patients only in the last century ( Zilla et al. , 2008 ; Perry et al. , 2003 ). In contrast to the adult heart, infants’ and children’s hearts regenerate in a larger capacity because the regenerative ability is proportionally correlated to age ( Rupp & Schranz, 2015 ). Additionally, there is an insufficient number of heart donors which becomes more challenging because of the heterogeneous relation of recipient-donor and the diverse range of CHD. These points result in high mortality rates and further financial costs to the healthcare systems ( Saxena, 2010 ). The mortality rate among all patients who are waiting for any type of organ transplantation is highest in infants who wait for heart transplant ( Dodson, 2014 ; Homann et al. , 2000 ). In 2003, some evidence was presented to support the regenerative ability of the adult heart ( Beltrami et al. , 2003 ). This evidence shows the existence of cardiac stem/progenitor cells which can differentiate into new cardiomyocytes and participate in cardiac regeneration ( Beltrami et al. , 2003 ). However, only a small number of preclinical studies have focused on CHD treatments ( Ebert et al. , 2015 ; Tarui, Sano & Oh, 2014 ). Stem cells (SC) have been widely investigated mainly for myocardial infarction (MI), as it is currently the leading cause of morbidity and mortality worldwide ( Fakoya, 2017 ). Cell seeding is a fundamental component of TE. Several studies have examined the possibility of direct cell delivery in the damaged area, cardiac patch implantation and engineered heart tissue, with the former being the most popular ( Ebert et al. , 2015 ; Fakoya, 2017 ; Feric & Radisic, 2016 ). All possible types of stem cells are under investigation to identify the most appropriate cell types for tissue engineering using in corrective surgery of CHD. This review looks into the congenital heart diseases, biomaterials and scaffolds, and, types of stem cells used in TE. Method This paper was based on review articles and reports in reputable peer-reviewed journals and government websites. The research was conducted using Medline on OvidSP, PubMed, google scholar, website, books, e- books, and reports. The words “congenital heart disease”, “tissue engineering”, “surgical treatment”, “stem cells”, “scaffolds”, “biomaterials” and a combination of those were used to retrieve literature from the databases. Congenital heart disease: types, malformations, presentations and interventions CHD includes a diverse range of conditions which shows a variety of symptoms, indications, and malformations detected during pregnancy or after birth ( Sun et al. , 2015 ). However, these malformations are much influenced by the age of diagnosis ( Hoffman & Kaplan, 2002 ). The etiology of CHD is unknown, but it is generally accepted that many factors or a combination of them could contribute to CHD and considered to be caused by multifactorial inheritance. These factors could be genetic, epigenetic or environmental factors such as alcohol and drugs consumption, as well as viral infections like Rubella ( Sun et al. , 2015 ). The severity of the disease varies, and a number of malformations could be present in each case. Based on the severity of CHD, they are categorized to mild, moderate, and severe CHDs, which the latter is subcategorized to Cyanotic and Acyanotic lesions ( Saxena, 2010 ). The most frequent type of severe CHD is Ventricular Septal Defect (VSD) ( Penny, 2011 ). VSD could cause myocardial defects which disappear in the first year of the infant’s life ( Penny, 2011 ). Nevertheless, the VSD could also cause some malformations which can be managed only by surgical intervention, that is, infant pulmonary hypertension ( Hoffman & Kaplan, 2002 ; Penny, 2011 ). The other CHD type is Atrial Septal Defect (ASD) which is usually asymptomatic and in most of the case will only be diagnosed in adulthood ( Hoffman & Kaplan, 2002 ). Atrioventricular septal defects (AVSD) is mainly observed in trisomy 21 ( Hoffman & Kaplan, 2002 ). AVSD is usually characterized by “complete AV-canals with one common AV valve for both ventricles and an interatrial and intraventricular communication” and requires surgical correction. The results of long-term patient follow up after operation have shown very satisfactory survival rate ( Boening et al. , 2002 ). Another type of CHD, tetralogy of Fallot (ToF), is characterized by VSD, pulmonary stenosis, right ventricular hypertrophy and over-riding of the aorta ( Apitz, Webb & Redington, 2009 ). Infants who suffer from ToF will require immediate surgical intervention for better survival rates and avoid cyanosis, a result of inadequate pulmonary blood flow ( Apitz, Webb & Redington, 2009 ). Calcific Aortic Valve (CAV), another type of CHD, is a disease which progresses slowly and results to a mild valve thickening and obstructing blood flow, aortic sclerosis or severe calcification with impaired leaflet motion ( Freeman & Otto, 2005 ). CAV presents many similarities with arteriosclerosis in adults which is caused by lifestyle or aging ( Freeman & Otto, 2005 ). However, CAV is a congenital, progressive disease which could be diagnosed in patients less than one year of age, and those in childhood or even adulthood ( Smith et al. , 2015 ; Medscape, 2015 ). Table 1 summarizes the frequencies of congenital heart diseases, while Table 2 summarizes the types of CHDs, presentations and possible management. 10. 7717/peerj. 5805/table-2 Table 2 Types, presentations and management of CHDs. Types of CHD Presentations of CHD Medical & surgical therapeutic approach to CHD Atrial Septal Defects ⮚ This defect manifests as a hole in the wall (septum) that separates the chambers above (atria) from those below (ventricles) ⮚ The volume of blood that flows through the lungs is increased over time due to the hole caused by the atrial defect resulting in damage to the blood vessels in the lungs ⮚ Frequent respiratory or lung infections ⮚ Difficulty breathing ⮚ Tiring when feeding (infants) ⮚ Shortness of breath when being active or exercising ⮚ Skipped heartbeats or a sense of feeling the heartbeat ⮚ A heart murmur, or a whooshing sound that can be heard with a stethoscope ⮚ Stroke ⮚ Swelling of legs, feet, or stomach area ( CDC, 2018a ). ⮚ Medical monitoring: the patient is monitored to see if the atrial septal defect would close on its own ⮚ Medications: beta blockers (to maintain a regular heartbeat) or anticoagulants (to help reduce blood clots) ⮚ Surgery: can be done through Cardiac catheterization or open-heart surgery ⮚ Follow-up care ( CDC, 2018a ; Mayo Clinic, 2018a ) Hypoplastic Left Heart Syndrome ⮚ This defect affects the normal blood flow through the heart. The left side of the heart does not form correctly and as such it is considered a critical congenital heart defect ⮚ The following structural malformations are observed in the left side of heart a. The left ventricle is underdeveloped. b. The mitral valves are not formed. c. The aortic valve is not formed. d. The ascending aorta is underdeveloped. ⮚ The left side of the heart cannot pump oxygen-rich blood. ⮚ Rapid, difficult breathing ⮚ Pounding heart ⮚ Weak pulse ⮚ Poor feeding ⮚ Being unusually drowsy or inactive ⮚ Ashen or bluish color ⮚ Dilated pupils ⮚ Lackluster eyes that seem to stare ( CDC, 2018b ; AHA, 2018 ). ⮚ Medication: inpatient medications include prostaglandin E1, Dopamine and Potassium Chloride and outpatient medications are Furosemide, Digoxin and Captopril ( Patnana & Turner, 2018 ). ⮚ Nutrition: feeding tube or special high-calorie formula. ⮚ Surgery: a. Norwood Procedure: performed on the infant within 2 weeks of a baby’s life. b. Bi-directional Glenn Shunt Procedure: done on an infant around 4 to 6 months of age. c. Fontan Procedure: performed on an infant around 18 months to 3 years of age ( CDC, 2018b ) Tricuspid Atresia ⮚ Occurs in which the tricuspid valve is not formed leading to the underdevelopment of the right ventricle. ⮚ The right side of the heart can’t pump sufficient blood to the lungs. ⮚ Problems breathing ⮚ Ashen or bluish skin color (cyanosis) ⮚ Poor feeding ⮚ Extreme sleepiness ⮚ Slow growth and poor weight gain ⮚ Edema of the abdomen, legs, ankles and feet ( Mai et al. , 2012 ; Mayo Clinic, 2018f ). ⮚ Medications: prostaglandins like Alprostadil IV to keep open the ductus arteriosus. ⮚ Nutrition: feeding tube ⮚ Surgery: a. Atrial Septostomy: performed in the first few days or weeks of a baby’s life b. Banding c. Shunt Procedure: done within the first 2 weeks of a baby’s life. d. Bi-directional Glenn Procedure: performed around 4 to 6 months of the baby’s life. e. Fontan Procedure: done around 2 years of age ( Mai et al. , 2012 ). Tetraogy of Fallot ⮚ Has a combination of four heart defects. This defect is a combination of pulmonary stenosis, ventricular septal defect, overriding aorta and right ventricular hypertrophy. ⮚ Cyanosis ⮚ Shortness of breath ⮚ Rapid breathing especially during feeding or exercise ⮚ Fainting ⮚ Clubbing of fingers and toes ⮚ Poor weight gain ⮚ Fatigue during play or exercise ⮚ Prolonged crying ⮚ Irritability ⮚ Heart murmur due to pulmonary stenosis ( Mayo Clinic, 2018e ; Baffa, 2018 ). ⮚ Medication: Prostaglandin E 1 infusion. ⮚ Surgery: a. Temporary surgery (palliative surgery): improve blood flow to the lungs. b. Intra-cardiac repair: done during the first year after birth ( Mayo Clinic, 2018e ). Bicuspid Aortic Valve ⮚ Has only two (bicuspid) cusps instead of three. ⮚ A bicuspid aortic valve may result in the heart’s aortic valve to narrow (aortic valve stenosis) which prevents the valve from opening completely, which reduces or blocks blood flow from the heart to the body. ⮚ Trouble breathing ⮚ Chest pain or pressure ⮚ Fatigue ⮚ Heart racing ⮚ Light-headedness ⮚ Fainting ( Northwestern Medicine, 2018 ). ⮚ Surgery: a. Aortic valve replacement b. Balloon valvuloplasty c. Aortic valve repair d. Aortic root and ascending aorta surgery ( Mayo Clinic, 2018b ) Patent Ductus Arteriosus ⮚ A persistent opening between the two major blood vessels leading from the heart. ⮚ Large patent arteriosus can cause poorly oxygenated blood to flow in the wrong direction. ⮚ Poor eating leads to poor growth. ⮚ Sweating with crying or eating ⮚ Persistent fast breathing or breathlessness ⮚ Easy tiring ⮚ Rapid heart taste ( Mayo Clinic, 2018c ) ⮚ Medications: NSAIDS (Advil, Infant’s Motrin), or indomethacin (Indocin) ( Mayo Clinic, 2018c ) ⮚ Surgery: Video-assisted thoracic surgical (VATS) repair ⮚ Catheter procedure: Trans-catheter occlusion ( Cleveland Clinic, 2018 ) ⮚ Watchful waiting Pulmonic Valve Stenosis ⮚ This defect affects the pulmonic valve in which a deformity on or near the valve causes it to be smaller and as such slows the blood flow. ⮚ The narrowing is due to the underdevelopment of the valve during fetal growth. The cusps maybe defective or too thick or may not separate from each other well. ⮚ Heart murmur ⮚ Fatigue ⮚ Shortness of breath, especially during exertion ⮚ Chest pain ⮚ Fainting ( Mayo Clinic, 2018d ) ⮚ Surgery: a. Balloon valvulplasty b. Open-heart surgery ( Mayo Clinic, 2018d ) Ventricular Septal Defect ⮚ A fissure connecting the two ventricles of the heart. Size varies with each patient ( Spicer et al. , 2014 ) ⮚ It can occur isolated or in association with other CHDs. ⮚ There are three kinds: muscular, periventricular and supra-crystal. These are based on location within the septum ( Carminati et al. , 2007 ) ⮚ Blood shunting: Depending on the size of the hole, blood flows from the left ventricle to the right. ⮚ Pulmonary hypertension : The shunting of blood flow leads to increased ventricular output to the pulmonary artery. With time, this can lead to pulmonary hypertension. ⮚ Eisenmenger’s Syndrome: Rise in pulmonary vascular resistance leads to increase in right ventricular pressure. This can lead to reverse shunting of blood from the right to left. This leads to cyanosis ( Spicer et al. , 2014 ) ⮚ Patients could also display clubbing ( Carminati et al. , 2007 ) ⮚ Growth retardation: Increased blood flow to the lungs results in an increase in lung compliance. This increases the energy demand for respiration. Thus, an energy deficit is created where the infant does not consume as much calories as is burned. This impedes growth. ⮚ Airway Obstruction: a. Increased pulmonary blood flow increases the size of the pulmonary arteries. This can cause the physical obstruction of large and small airways. b. There is also the possibility of the incidence of pulmonary edema due to increased blood flow. The combination of these events can lead to respiratory distress. Thus, symptoms such as wheezing, and tachypnea can be observed ( Spicer et al. , 2014 ). c. Holosystolic/pansystolic murmur on auscultation ( Minette & Sahn, 2006 ) ⮚ Smaller holes resolve themselves with time. ⮚ Surgery: Usually done on larger fissures. It is indicated where patients express symptoms of heart failure, left heart overload and history of endocarditis. ⮚ Percutaneous techniques: a. These do not require opening the patient up. b. Trans catheter approach: A catheter is threaded from an artery in the legs, or groin into the heart. A device is then placed to obstruct the hole in the ventricle ( Spicer et al. , 2014 ; Carminati et al. , 2007 ). Total Anomalous Pulmonary Venous Condition ⮚ The pulmonary veins are attached to the right atrium instead of the left. ⮚ Usually associated with atrial septal defect. ⮚ Cyanosis: Oxygenated blood from the lungs is pumped into the right atrium. It mixes with deoxygenated blood and passes through the atrial septal defect into the left atrium. This decreases oxygen supply to the body leading to cyanosis. ⮚ Pulmonary hypertension: Some patients have constricted pulmonary veins that lead to pulmonary hypertension. This leads to pulmonary effusion. ⮚ Hypovolemia : Some patients manifest with a narrow or restrictive atrial septal defect. This significantly reduces the blood flow to the body leading to hypovolemia ( NIH, 2018 ). ⮚ Surgery: The pulmonary veins are surgically reattached to the left atrium. ⮚ Cardiac catheterization: For the patients with a restricted atrial septal defect, a balloon pump is used to widen the fissure until corrective surgery can be carried out ( NIH, 2018 ) Trans-Position of the Great Arteries ⮚ A condition whereby the aorta and the pulmonary arteries are transposed. The aorta arises from the right ventricle and leads to the lungs. The pulmonary artery arises from the left ventricle and leads to the body. ⮚ It is comorbid with ventricular septal defect and patent ductus arteriosus. ⮚ Cyanosis: Mixing of blood leads to supply of poorly oxygenated blood to the body ( Martins & Castela, 2018 ) ⮚ Medications: Prostaglandin E1 is administered to keep the ductus arteriosus open. ⮚ Surgery: a. Balloon atrial septostomy: A catheter is threaded through the foramen ovale. A balloon is inflated to rip a fissure in the atrium ( Martins & Castela, 2018 ) Truncus Arterious ⮚ A condition where the truncus arteriosus of a fetus does not differentiate into an aorta and pulmonary vein. Thus, the patient only has one vessel exiting the heart ⮚ Cyanosis: This leads to mixing of oxygenated and deoxygenated blood. Thus, the oxygen supply to the body is decreased ⮚ Congestive heart failure: the excess of volume of blood flow to the heart increases pressure in the lungs. This would eventually lead to cardiac failure. ⮚ Usually comorbid with ventricular septal defect ( Cincinnati Children Health, 2018 ) ⮚ Surgery: The truncus arteriosus must be separated into two vessels. This would allow separate blood flow channels to the heart and body ( Cincinnati Children Health, 2018 ) Ebstein’s Anomaly ⮚ A congenital malformation of the tricuspid valve ⮚ The posterior and septal leaflets are displaced downwards. This leads to a downward enlargement of the right atrium. ⮚ The walls of the right atrium become thin. ⮚ It can be comorbid with patent foramen ovale or atrial septal defect. ⮚ Patients can be asymptomatic but could also present with symptoms. ⮚ Cyanosis: Shunting of the blood between patent foramen ovale and atrial septal defect leads to blood mixing between the left and right sides of the heart. This can lead to cyanosis. ⮚ Conduction irregularities: Some patients present with arrhythmias ( Khan et al. , 2018 ) ⮚ Surgery: a. Only required where patient manifests severe symptoms. b. Cone procedure: Where the anterior septal leaflet is maneuvered and sewn to the true annulus. This attachment causes it to be conical in shape. c. Valve replacement d. Where the defective valves can be surgically replaced ( Khan et al. , 2018 ) Pulmonary Atresia ⮚ Characterized by a restriction to blood flow from the right ventricle to the pulmonary artery. It could be due to malformation of the pulmonary valve or of the pulmonary artery itself. ⮚ It can manifest with a ventricular septal defect where there are collateral arteries supplying the lungs. ⮚ It could also manifest without a ventricular septal defect. Here the right ventricle is usually hypoplastic. It would usually be comorbid with a patent ductus arteriosus ( Safi, Liberthson & Bhatt, 2016 ). ⮚ Surgery: a. A shunt must be created between the pulmonary artery and the aorta. This can be done by administration of prostaglandin E to keep the ductus arteriosus open. It could also be doe surgically. b. Fontan’s procedure: Done for patients with a hypoplastic right ventricle. The right atrium is surgically connected to the pulmonary artery ( Safi, Liberthson & Bhatt, 2016 ). Aortic Stenosis ⮚ Defect of the aortic valve that restricts its opening. ⮚ It can lead to ventricular hypertrophy which can eventually lead to heart failure. ⮚ There is also a possibility of development of atrial fibrillation. ⮚ Presence of systolic murmur ( Skybchyk & Melen, 2017 ) ⮚ Valve replacement: There are two approaches: a. Surgically b. Trans-catheter approach ( Skybchyk & Melen, 2017 ). Coarctation of the Aorta ⮚ It is the constriction of the proximal end of the aorta leading to restriction to blood flow. ⮚ Patients can present with acidosis, cardiac failure, as well as shock after ductus arteriosus closes ( Johnson et al. , 1998 ) ⮚ Surgery: a. Coarcted potion can be resected. And the ends of the artery can be re-anastomosed. b. A patch can also be used to surgically dilate the artery. ⮚ Balloon Angioplasty: A catheter is threaded in to the aorta. A balloon is inflated to enlarge the aorta. Some types of the CHDs are misdiagnosed, undiagnosed, or diagnosed very late in life which could make a successful treatment challenging ( Torok, 2015 ). CHD patients usually require medication, cardiac catheterization or a series of surgical interventions throughout their lives ( Sun et al. , 2015 ; Homann et al. , 2000 ). This, among other risks, increases the chances of HLA-sensitization which eventually makes it more difficult to find a cross-match ( Homann et al. , 2000 ). The main challenge with artificial materials is the alterations in size and function of the heart from the neonatal period, infanthood, and to adulthood, to which current artificial materials cannot adjust to ( Sun et al. , 2015 ). The need for early intervention is essential for normal physical and cognitive development ( Torok, 2015 ). The use of bioprosthetics including allografts and xenografts is widely used for treatment of CHD as they present good survival rates in the patients and fewer interventions throughout their lifetimes ( Dodson, 2014 ). The observed symptoms, defects, and complications upon bioprosthetic treatment of CHD patients can differ from case to case ( Penny, 2011 ; Torok, 2015 ). Heart, a complex organ for tissue engineering Unlike other human tissue, cardiac tissue is a more complex tissue considering not only because of its mechanical and structural function but also due to its electrical properties as well. The human heart mainly consists of cardiomyocytes, functions as a blood pump which is regulated by the electrical signal generated by the pacemaker cells in the sinoatrial nodes. This signal is directed and spread through the atrioventricular node to Purkinje fibers, and this is highly important for the direction of blood flow ( Files & Boucek, 2012 ). The diastolic and systolic function of the heart is necessary to be synchronized and adjusted according to the body’s needs. The isolation of this signal from the rest of cardiac tissue is as important as this electrical signal by itself. This importance is achieved by the extracellular matrix (ECM) of heart which is also responsible for mechanical support and endurance ( Files & Boucek, 2012 ). Aside from the functional aspects briefly mentioned above, the structural aspects (the cardiomyocyte and its ECM) are other factors for strong consideration in the success of TE. To yield an engineered tissue with the best functionality, this engineered tissue must be similar in every sense to the native tissue. The ECM of the heart is mainly a complex mesh of structural elements such as cardio fibroblasts and collagen fibrils, and non-structural elements such as proteoglycans, glycosaminoglycans, and glycoproteins ( Kaiser & Coulombe, 2016 ; Rienks et al. , 2014 ) among other components. Repairing the heart using materials which do not have or do not comply with the above characteristics and cannot work in harmony with the host-heart, will result in a non-efficient functioning heart accompanied by a series of complications. The concept and act of repairing the heart using various engineered techniques has evolved over the years from the use of artificial heart valves and grafts to bioprosthesis, and currently forward to the use of biomaterials and scaffolds cells. Artificial heart valve and grafts Charles Hufnagel was the first to experiment on animals with an artificial valve which he designed in the 1940’s ( Gott, Alejo & Cameron, 2003 ). A few years later, the same type of valve was transplanted into humans ( Gott, Alejo & Cameron, 2003 ). Nonetheless, Hufnagel’s valve required changes which were succeeded by Harken-Soroff and later by Starr-Edwards ball valve ( Gott, Alejo & Cameron, 2003 ). Following several improvements, the latest version is made of pure carbon as a lighter, smoother material for blood flow and more durable in comparison to other materials ( Zilla et al. , 2008 ). Due to the position of the valve, there is a high transvascular pressure which leads to ‘impact wear’ and ‘friction wear’ ( Zilla et al. , 2008 ). Further complications with the use of valve replacements include inflammation around the prosthesis and calcification of the valve itself ( Brown, 2005 ). The main disadvantage of artificial valves is the thromboembolic risk which leads to lifetime treatment with anticoagulants. This type of treatment involves various complications and brings different risks to patient ( Zilla et al. , 2008 ). Similarly, to valves, artificial vascular grafts have been used for many years as a surgical treatment for CHD ( Chaikof, 2007 ). Nevertheless, they provoke an inflammatory response, and they are much less flexible than the body’s natural tissue ( Torok, 2015 ). Originally porous fabric knitted of Dacron and polytetrafluoroethylene (PTFE) was used for stenosis treatment. Later alterations in the porosity of fabric were introduced to prevent the material’s corruption ( Torok, 2015 ). The first attempt to combine biomaterials with patient’s cells was made by Wesolowki and colleagues by “preclotting the graft with the patient’s blood” ( Torok, 2015 ). However, future attempts to produce successful results created many doubts about the actual function of cells following citation on clot ( Torok, 2015 ). Nonetheless, both artificial heart valves and conduits require continuous use of anticoagulants. Remarkably, they do not grow with the patients’ heart as patients with CHD are highly likely to require surgical intervention when they are an infant or a child ( Torok, 2015 ; Ratner et al. , 2004 ). Bioprosthetics: allograft, xenografts, and autografts Allograft heart valves and arterial grafts are collected from deceased humans. In contrast, xenografts are harvested from porcine and bovine animals including heart valves and carotid arteries ( Zilla et al. , 2008 ; Perry et al. , 2003 ; Brown, 2005 ). Allografts and xenografts came into the picture as an alternative to artificial valves and conduits. Their main advantage is that they do not require lifetime treatment with anticoagulants ( Zilla et al. , 2008 ; Hibino et al. , 2010 ; Schmidt & Baier, 2010 ). Animal tissue is treated with glutaraldehyde ( Hibino et al. , 2010 ). Glutaraldehyde is a five-carbon bifunctional aldehyde used to stabilize tissue to protect from chemical and enzymatic degradation and maintain “its mechanical integrity and natural compliance” ( Hibino et al. , 2010 ). Also, treatment is necessary to reduce immunogenicity of the xenograft by decellularization and sterilization of the tissue ( Hibino et al. , 2010 ). Like xenograft, allografts also need treatment before transplantation and can even be cryopreserved ( Dodson, 2014 ). Despite the great advantage, a number of complications are related to bioprosthetic grafts related to their preparation ( Hibino et al. , 2010 ). The risk of cytotoxicity leading to inflammation, as well as the partial loss of mechanical properties of tissue, have been reported ( Hibino et al. , 2010 ). Moreover, calcification is often observed in infants and children with bioprosthetics. Many efforts are being made to find out an alternative treatment for heart bioprosthetics therapy. However, there is a controversy regarding their efficiency ( Hibino et al. , 2010 ). In a study carried out by Homann and colleagues, the outcomes of 25 years using allografts and xenografts for reconstruction of the right ventricular outflow tract showed 66% survival at 10- years’ follow up. Furthermore, patients with allografts had a mean reoperation-free interval time of 16 years in contrast to the xenograft recipients which this interval time is 10. 3 years ( Dodson, 2014 ). Allografts may present better outcomes, but they are not in abundance like xenografts. Therefore, many studies are concentrating on the development of tissue valves and vascular grafts created by stem cell-seeded on artificial or natural scaffold ( Perry et al. , 2003 ). The relatively recent “CorMatrix” patch fabricated from the decellularized porcine small intestinal submucosa extracellular matrix (SIS-ECM) mainly composed of collagen, elastin, glycan, and glycoproteins have been introduced into cardiac surgery. SIS-ECM has not only been used in animal models for cardiac surgery ( Dodson, 2014 ; Homann et al. , 2000 ; Files & Boucek, 2012 ; Kaiser & Coulombe, 2016 ; Rienks et al. , 2014 ; Brown, 2005 ; Chaikof, 2007 ; Ratner et al. , 2004 ; Hibino et al. , 2010 ; Schmidt & Baier, 2010 ; Kurobe et al. , 2012 ; Johnson et al. , 1998 ; Ruiz et al. , 2005 ), but also in human studies, for cardiac and vascular reconstructions such as augmentation of the tricuspid valve ( Wainwright et al. , 2012 ), vascular repair of ascending aorta, aortic arch, right ventricular outflow tract, pulmonary artery, valvular reconstruction ( Scholl et al. , 2010 ), and closure of septal defects ( Quarti et al. , 2011 ). The study by Witt et al. (2013) reported a small risk of stenosis when SIS-ECM is used in the reconstruction of the outflow tracts and great vessels. Interestingly, the SIS-ECM has effectively proved the function in the high-pressure vessels ( Quarti et al. , 2011 ). The pitfall of this study, however, was the short follow up period. A very common surgical practice for CHD is the use of pericardial patches for repairing the septal defect ( Torok, 2015 ). The autologous pericardium is the best choice for infants as it is free, it does not provoke any immune-response, and it is sterile. Even though it requires some preparation before application, autologous pericardium creates less fibrotic tissue in comparison to Dacron ( Torok, 2015 ). Allograft pericardium is available, but quite a few risk factors are associated with its use ( Torok, 2015 ). Biomaterials and scaffolds for tissue engineering Generally, scaffolds work as a primary base for cells to enhance and produce relevant tissue. The scaffolds should have specific morphological, functional, and mechanical properties to support cells survival and differentiation ( Feric & Radisic, 2016 ). Biomaterials used to produce scaffolds should be made of components which will accommodate the above characteristics and create a friendly cell microenvironment ( Feric & Radisic, 2016 ). The previously mentioned characteristics apply for all different types of engineered tissue, and the goal is to mimic host tissue in the best possible way. With regards to artificial and bioprosthetic cardiac correction choices, scaffolds and biomaterials should contain various properties such as being biodegradable, biocompatible, flexible and durable and absence of immunogenicity and calcification. Due to the variety of sizes of patients’ hearts and defects, designed scaffolds should have various sizes and be able to grow and adapt to the heart ( Schmidt & Baier, 2010 ). The biomaterial should allow neo-vascularization for adequate oxygenation of the tissue, create minimal scarring tissue and thrombotic risk, as the latter could lead to life treatment with anticoagulants ( Witt et al. , 2013 ; Miyagawa et al. , 2011 ). Furthermore, these biomaterials scaffolds should be bioactive, meaning they should enhance cellularization in vitro and in vivo, and optimize cell efficiency and degrade at a desirable rate ( Miyagawa et al. , 2011 ). What’s more, biomaterials and scaffolds should be in abundance and cost-effective, as high cost could restrict development and use of it in TE as a routine therapeutic choice ( Avolio, Caputo & Madeddu, 2015 ). The most common biomaterials for cardiac and vascular TE used today are synthetic, and natural polymers ( Files & Boucek, 2012 ), and the electrospinning technique has been proven to be the most efficient way to produce scaffolds with these biomaterials ( Williams, 2004 ). Synthetic polymers for cardiac scaffolds The easiest way to have materials in abundance is to manufacture them. The need for suitable biomaterials for cardiac tissue repair has triggered development of synthetic polymers which are easy to fabricate and manipulate. These polymers can be manipulated with respect to their physical properties, molecular weight, heterogeneity index, and degradation speed ( Zhao et al. , 2015 ). Many synthetic polymers are biocompatible and have excellent mechanical properties which make them a popular choice for sutures and mesh production ( Files & Boucek, 2012 ). Frequently used polymers in cardiac surgery are polyglycolic acid (PGA) and polylactic acid (PLA). These two polymers have been used as a single biomaterial or as 50:50 composite to reconstruct tissue-engineered vascular grafts for treating children with congenital heart disease ( Maeda et al. , 2015 ). Carrier et al. (1999) presented acceptable ultrastructural features and metabolic cell ability when cells were cultured on PGA scaffold in a rotating bioreactor ( Sugiura et al. , 2018 ). The rotating bioreactor increases cell adherence and decreases cell damage ( Sugiura et al. , 2018 ). The highest concern with synthetic polymers is their toxicity. Therefore, the use of poly-L-lactic acid (PLLA) has increasingly become more of interest. PLLA has demonstrated very good results when combined with bone marrow mesenchymal SC (BM-MSC) for vascular tissue engineering ( Masuda et al. , 2008 ). In vivo studies by Hashi et al. showed that nanofibrous scaffolds created with PLLA can remodel in both cellular and ECM content, similar to that of the native artery ( Masuda et al. , 2008 ). Both acellular and cellular scaffolds were implanted into the common carotid artery of live animal models (rats). The cellular scaffold was seeded with MSC and “exhibited very little platelet aggregation on the luminal surfaces” compared to the acellular grafts. This is due to the antithrombogenic property of MSC ( Masuda et al. , 2008 ). PLLA, when degraded in the body, can “be excreted in carbon dioxide and water” ( Files & Boucek, 2012 ). Polyurethane, unless copolymerized, is biocompatible but not biodegradable. Polyurethane has been successfully experimented in combination with other materials for cardiac tissue repair, such as siloxane films ( Hashi et al. , 2007 ; Baheiraei et al. , 2015 ), cellulose ( Baheiraei et al. , 2016 ), urea ( Su et al. , 2016 ), PLLA ( Hernández-Córdova et al. , 2016 ). Poly (ε- caprolactone) in combination with other biomaterials have also been proven to be efficient composite for cardiac tissue repair. They have been used in combination with PLLA alone ( Tomecka et al. , 2017 ; Sharifpanah et al. , 2017 ), PLLA and collagen ( Centola et al. , 2010 ), polypyrrole and gelatin ( Mukherjee et al. , 2011 ), polyglycolic acid ( Kai et al. , 2011 ), poly (hydroxymethyl glycolide) ( Sugiura et al. , 2016 ), chitosan and gelatin ( Castilho et al. , 2017 ). Based on our understanding of the heart as an electroactive tissue, Hitscherich and colleagues have created a piezoelectric scaffold fabricated by electrospinning Polyvinylidene Fluoride-Tetrafluoroethylene (PVDF-TrFE) for cardiac tissue engineering ( Pok et al. , 2013 ). The combination of synthetic with natural polymers has been suggested to increase cell adherence. However, pure natural polymers have also been examined as an option for polymers ( Files & Boucek, 2012 ). Natural polymers for cardiac scaffolds Natural polymers are biodegradable, biocompatible and easily manipulated matrices composed of complex elements which make up the native tissue ( Hitscherich et al. , 2016 ). The natural polymers used so far for cardiac repair include collagen, gelatin, alginate, silk, fibrin, chitosan and hyaluronic acid ( Files & Boucek, 2012 ). Despite their poor mechanical properties, they are good biomaterial for heart TE, as they have high biocompatibility, promote cell-binding and could biodegrade with no “additional treatment or modifications” ( Files & Boucek, 2012 ). Collagen is the most widely utilized natural polymer which is the most abundant ECM protein. It functions to guide biological processes, provide structural scaffolding, and tensile integrity ( Castilho et al. , 2017 ). Several kinds of literature have reported the use of various collagen types and their modifications in cardiac tissue repair ( Nelson et al. , 2011 ; Herpel et al. , 2006 ; Serpooshan et al. , 2016 ; Dawson et al. , 2011 ; Yu et al. , 2017 ; Sun et al. , 2017 ; Frederick et al. , 2010 ; Hsieh et al. , 2016 ). Fibrin can be manipulated to create gels, microbeads, and hydrogels ( Zhao et al. , 2015 ). Likewise, biological molecules like the growth factors can be incorporated ( Nie et al. , 2010 ). Fibrin glue can be used as a stand- alone therapy in cardiac tissue repair as it possesses intrinsic regenerative properties ( Menasché et al. , 2014 ). The success of fibrin patch seeded with human embryonic stem cell-derived cardiac progenitor cells (hESC-CPC) in non-human primate model ( Gil-Cayuela et al. , 2016 ) has resulted in its translation to the first case report of using hESC-CPC in severe heart failure with an encouraging patient functional outcome ( Menasché et al. , 2015 ). Other studies on fibrin have demonstrated its efficacy as a sealant after intramyocardial injection ( Terrovitis et al. , 2009 ); for myocardial tissue repair when seeded with adult stem cells, neonatal cardiac cell, and mesenchymal stem cells ( Llucià-Valldeperas et al. , 2016 ; Tao et al. , 2014 ; Ichihara et al. , 2017 ); to form aortic valves in tissue engineering ( Moreira et al. , 2016 ). Chitosan has been experimented with and appears in the literature as a biomaterial for cardiac regeneration ( Camci-Unal et al. , 2014 ; Wang et al. , 2014 ). Overall, in cardiomyogenesis, many researchers have agreed on the fact that chitosan seems to be more effective when combined with other factors enhancing integration of stem cells into cardiac tissues ( Chopra et al. , 2006 ). Alginate, when used alone, has proven to have a remarkable effect on the function of heart models with myocardial infarction. Furthermore, seeding alginate with stem cells has proven to be more efficient in repair of the cardiac tissue ( Wang et al. , 2012 ; Landa et al. , 2008 ; Leor et al. , 2009 ; Sabbah et al. , 2013 ). The use of hyaluronic acid has been shown to be largely dependent on its molecular weight, and several kinds of literature have reported its successful use in cardiac tissue repair ( Bonafè et al. , 2014 ; Yoon et al. , 2009 ; Yang et al. , 2010 ; Göv et al. , 2016 ). Evidence of gelatin scaffold placed subcutaneously, and/or on infarcted myocardium in adult rat hearts have shown a good survival of the graft, vessel formation and junctions with recipient rat heart cells ( Li et al. , 1999 ). Gelatin was reported to sustain neonatal rat cardiomyocyte tissue in vitro for three weeks ( McCain et al. , 2014 ); supported the growth of human induced pluripotent stem cell (iPSC)-derived cardiomyocytes ( McCain et al. , 2014 ); its hydrogel seeded with autologous human cardiac-derived stem cell and basic fibroblast growth factor (bFGF) effectively released bFGF for repair of ischemic cardiomyopathy ( Yacoub & Terrovitis, 2013 ); and several other studies have shown the efficacy of gelatin as a scaffold for cardiomyogenesis when seeded with cells ( Takehara et al. , 2008 ; Navaei et al. , 2016 ; Kudová et al. , 2016 ; Cristallini et al. , 2016 ). Fibrinogen/Thrombin-based Collagen Fleece (TachoCombo) have been successfully used to secure hemostasis and enhance complete reconstruction of a large pulmonary artery defect in a canine model ( Okada et al. , 1995 ). Hence, this biomaterial may be used in reconstruction of the low-pressure pulmonary vessels during a cardiac surgery for a total anomalous pulmonary venous return or transposition of the great vessels. Nevertheless, not all of these polymers can tune well for cardiac TE, and risk of inflammation still exists ( Sakai et al. , 2001 ). Native extracellular Matrix as scaffold Native-specific ECM could be a category itself or part of natural polymers. ECM is collected from animal or donor tissue and processed for culturing cells ( Zimmermann, Melnychenko & Eschenhagen, 2004 ). Studies have shown that the ratio of native ECM in culture could play a key role in stem cells (SCs) enhancement, differentiation, survival, and phenotype ( Duan et al. , 2011 ). Other studies have shown that contractible engineered heart patches cultured in ECM mixture and implanted in syngeneic Fischer 344 rats can vascularize, become innervated and survive up to 8 weeks in vivo ( Zimmermann, Melnychenko & Eschenhagen, 2004 ). Furthermore, ECM of decellularized and repopulated hearts and other organs are being used for drug testing ( Lewis, 2016 ). An experiment on decellularized mouse hearts which were repopulated with human cells through coronary vessels exhibited myocardium, vessel-like structures and intracellular Ca2+ transients contracted spontaneously and responded as expected to various drug interventions ( Lu et al. , 2013 ). It was concluded that heart “ECM could promote proliferation, specific cell differentiation and myofilament formation” ( Lu et al. , 2013 ). The option of ECM for TE could help overcome the challenges faced using synthetic and other natural biomaterials to replace tissue, valves or organs ( Iop et al. , 2009 ). Scaffoldless cell sheet Another technique, which is independent of scaffolds, has been developed based on the cells’ ability to connect via cell-to-cell junction proteins and create ECM ( Carrier et al. , 1999 ). The cells are cultured in normal conditions at 37 °C in a temperature-responsive polymer cultures dish. When the culture temperature conditions change, the cells detach from the polymer culture dish as one cell sheet ( Carrier et al. , 1999 ). This technique was developed to avoid inflammatory reactions and fibrotic deposits in the area of graft where scaffold was placed following degradation ( Carrier et al. , 1999 ). A study has shown that contractile chick cardiomyocyte sheets could function effectively around rat thoracic aorta when applied on host myocardium ( Witt et al. , 2013 ). This cell sheet could synchronize within 1 h of implantation with the host tissue ( Witt et al. , 2013 ). Similar results have been shown in 3D structures using several cell sheets aiming to create a thick cardiac patch ( Yamato & Okano, 2004 ). The number of sheets is limited, as more than three exhibits poor vascularization. However, the combination of endothelial cells and cardiomyocytes is being examined to promote vascularization before implantation ( Carrier et al. , 1999 ). Table 3 summarizes the pros and cons of the various materials and biomaterials used in tissue engineering. 10. 7717/peerj. 5805/table-3 Table 3 Advantages and disadvantages of materials and biomaterials used in TE. Artificial prosthesis Biological prosthesis Biomaterial scaffolds Scaffoldless tissue Advantages • Available in abundance • Many different sizes • Long term results available • Available in abundance • No requirement for life-long treatment with anticoagulants • Good mechanical properties • Ultrastructural features • Cell adherence • Biocompatibility • Biodegradable • 3D-priting allows any shape and size • No need for scaffold • Spontaneous and synchronous pulsation • Could create tubular construct • Can grow with host Disadvantages • Impact & friction wear • Inflammation • Calcification of valve • Less flexible than natural tissue • Life-long treatment with anticoagulants • Do not grow with the patients’ heart • Risk of cytotoxicity • Inflammation • Loss of mechanical properties • Calcification in infants and children • Immunological reactions • Some present toxicity • Risk of inflammation • Not all tune well with the heart • Limited number of cell sheets (max 3) • Poor vascularization in more than 3 cell-sheets Stem cells for tissue engineering An equally important point in choosing a suitable biomaterial, scaffold or scaffoldless cell sheet, is the choice of the most appropriate cell types suitable for the TE. Stem cells (SCs) as a known cell source possesses the ability to differentiate toward Cardiac Muscle Cells/cardiomyocytes (CMCs), smooth muscle cells (SMCs) and endothelial cells (ECs), can regenerate cardiac tissue. Based on these properties, they play a key role in TE field. The currently used SCs in TE are summarized in Table 4. 10. 7717/peerj. 5805/table-4 Table 4 Scaffolds and SCs used for TE in some study models. Engineered tissue Scaffold Type of SCs Study models Reference Heart valve Synthetic biodegradable non-woven PGA mesh Human Chorionic villi-derived cells & hCB- EPCs Culture in bioreactor 164 Synthetic biodegradable hAFSCs Culture in bioreactor 166 Porcine decellularized scaffold BM-MSCs & BM-MSCs Lambs 176 Vascular graft Various synthetic biodegradable Human Umbilical CB-EPCs Static conditions & biomimetic flow system 165 Biodegradable non-woven PGA BM-MNCs Mice 174 Biodegradable PLA & PGA BM-MNCs Human 79 Embryonic stem cells Embryonic stem cells (ESCs) are one of the cell sources which are used in TE approaches. ESCs are derived from the inner cell mass of preimplantation blastocyst ( Boroviak et al. , 2014 ). They can differentiate into all different cell types of three germ layers. Human ESC (hESC) could be a good candidate for cardiac tissue engineering. In a study conducted by Landry et al. , hESC-derived cardiomyocytes (hESC-CMC) showed very good phenotype including myofibril alignment, density, morphology, contractile performance and gene expression profile which, however, was only confirmed after 80–120 days in vitro culture ( Lundy et al. , 2013 ). Various groups apart from Landry and colleagues conducted studies to confirm the successful differentiation of ESC to cardiomyocytes as presented in a review by Boheler and colleagues in 2002 ( Boheler et al. , 2002 ). Duan et al. (2011) investigated how native cardiac ECM could affect hESC differentiation. This group processed porcine hearts to collect digested cardiac ECM which then mixed with collagen to create a hydrogel for cell cultivation purposes. The cultured hESCs on biomaterials comprised of 75% native cardiac porcine ECM and 25% hydrogel with no additional growth factors have shown a great differentiation with cardiac troponin T expression and contractile behavior, compared to the hydrogel with a smaller ratio of native ECM. Based on various studies, ESCs could be a good option for cardiac TE ( Zimmermann, Melnychenko & Eschenhagen, 2004 ). Additionally, some factors such as ethical concerns, provoked immunogenicity, and risk of tumorigenesis make the ESCs a very controversial choice of cell source for TE ( Schmidt & Baier, 2010 ). Induced pluripotent stem cells Another type of SCs which are used in TE is induced pluripotent stem cells (iPSCs). The iPSCs are somatic cells which are reprogrammed to behave like ESC and show the same properties. The iPSCs can differentiate into all three germ layers ( Takahashi et al. , 2007 ; Yu et al. , 2007 ). Takahashi and his group was the first group who was able to reprogram the somatic adult cells like fibroblasts to iPSCs using viral vectors to introduce four key factors OCT4, SOX2, c-Myc, and KFL-4 to fibroblasts ( Takahashi et al. , 2007 ). This method was used to reprogram fibroblasts to embryonic-like cells and from this state differentiate them into a relevant type of cells ( Takahashi et al. , 2007 ). Ludry et al. have also shown that human iPSC-derived cardiomyocytes (hiPSC-CMCs) present the same characteristics as hESC-derived cardiomyocytes in long-term in vitro culture ( Lundy et al. , 2013 ). Lu and colleagues presented the successful repopulation of decellularized cadaveric mouse heart with hiPSC-derived multipotential cardiovascular progenitors ( Lundy et al. , 2013 ). They also demonstrate that the heart ECM promotes proliferation, differentiation and myofilament formation of CMs from the repopulated hiPSC-derived cells. Furthermore, they have checked the electrical coupling of these cells and also examined the constructive ability of repopulated heart using electrocardiogram which presented arrhythmia. Lu et al. have further examined the effects of pharmacological agents on repopulated heart and observed remarkable responses ( Zimmermann, Melnychenko & Eschenhagen, 2004 ). This model is explored as an option to personalized medicine concerning drug testing/discovery ( Zimmermann, Melnychenko & Eschenhagen, 2004 ; Bosman et al. , 2015 ). Specifically, for CHD which presents such a variety of profiles, individual patient-specific human models’ development could help to understand how each patient would respond to existing pharmaceutical treatments. Even though it may not be possible to create actual organ heart models with individual clinical features of the disease ( Caputo et al. , 2015 ). Nonetheless, like ESC, iPSC has demonstrated tumorigenesis ( Schmidt & Baier, 2010 ). Ieda and colleagues showed a direct transdifferentiation of fibroblasts to functional cardiomyocytes using three key factors, Gata4, Mef2c, and Tbx5, within a very short time and suggested that direct reprogramming could reduce the risk of tumorigenesis ( Ieda et al. , 2010 ). Still, using viral vectors for reprogramming procedure is problematic and involves various risks ( Schmidt & Baier, 2010 ). Therefore, today more different ways for iPSC production are being used and investigated to find out safer and more effective alternatives for this reprogramming procedure ( Tarui, Sano & Oh, 2014 ; Lüningschrör et al. , 2013 ). In the event this problem is solved, iPSC could be the safest type of cell sources for TE as they will not provoke any immune-response and cell harvesting procedure to produce iPSCs is not life-threatening for the patients. Moreover, iPSC raises less ethical concerns in comparison to ESC or fetal SC. Prenatal, perinatal, and postnatal stem cells Prenatal, perinatal and postnatal SCs are other cell sources used in TE, which include chorionic villi derived multipotent SCs, amniotic fluid-derived SCs (AFSCs), umbilical cord blood derived-endothelial progenitor cells (UCB-EPCs) and umbilical cord- or cord blood-derived- multipotent SCs ( Webera, Zeisbergera & Hoerstrup, 2011 ). UCB progenitors, like endothelial progenitor cells (EPCs), have distinctive proliferative properties in comparison to other cells sources ( Murohara et al. , 2000 ). This category of SCs is exceptionally important as the child’s own SCs could be used for heart TE, for CHD patients who are diagnosed before birth. Immunogenicity or an additional procedure to harvest autologous SCs from the infant or child could be avoided in this procedure. Furthermore, it has been proven that AFSCs do not form teratomas in contrast to ESCs and iPSCs ( De Coppi et al. , 2007 ). All categories of these cells have been investigated with remarkable results on engineered valves and vascular grafts ( Webera, Zeisbergera & Hoerstrup, 2011 ; Schmidt et al. , 2006 ; Schmidt et al. , 2004 ; Schmidt et al. , 2007 ; Yao, 2016 ; Petsche Connell et al. , 2013 ). This type of SCs is not applicable to adults who have been diagnosed later in life with CHD. Adult stem cells Bone marrow-derived stem cells Apart from UCB, EPCs can be found in the peripheral blood (PB-EPCs) and bone marrow (BM- EPCs) of adults ( Asahara et al. , 1997 ; Zisch, 2004 ). However, bone marrow is a richer source of EPCs in comparison to peripheral blood. Asahara et al. (1997) identified the CD34+ mononuclear hematopoietic progenitor cells in the peripheral blood which in vitro presented endothelial-like characteristics. The EPCs which are originated from BM considered to play a crucial role in endothelial repair, and they have been suggested for treatment of ischemia patients and vascular TE with very encouraging results ( De Coppi et al. , 2007 ; Schmidt et al. , 2004 ; Zisch, 2004 ; Olausson et al. , 2014 ). A successful complete endothelium regeneration of decellularized canine carotid arteries has been reported in animal studies using PB-EPCs ( Zhou et al. , 2012 ). In addition to vascular TE, the EPCs have been assessed for tissue-engineered heart valves ( Sales et al. , 2011 ). Bone marrow derived EPCs are greatly involved in de novo vessel formation and neovascularization in pathological conditions like ischemia and cancer ( Asahara et al. , 1997 ). Similar applications to PB-EPCs and prenatal EPCs have been recorded for vascular graft in the congenital heart surgery using bone marrow-derived stem cells ( Mirensky et al. , 2010 ). Mirensky and colleagues used sheets of non-woven PGA mesh as a scaffold to create vascular graft in combination with human bone marrow mononuclear cells (BM-MNCs). The results were very encouraging as no aneurysm or thrombotic incidence were reported, despite the absence of anticoagulants. This group suggests this method as a suitable vascular treatment for CHD based on their results from 6 weeks follow up after graft implantation, which has shown signs of degradation, and it was fully accommodated by the host’s cells ( Mirensky et al. , 2010 ). Nonetheless, the host mice were at full-growth which makes it questionable how successful this application could be in a developing animal model. Interestingly only one week after implantation no human BM-MNCs were detected, which suggests that BM-MNCs play a paracrine role rather than cell replacement ( Zisch, 2004 ). More studies have reported similar results with the same conclusion ( Schmidt & Baier, 2010 ; Fernandes et al. , 2015 ). Furthermore, an investigation on 25 young patients under 30 years old who underwent extracardiac total cardiopulmonary connection with BM-MNCs engineered vascular graft has also presented convincing results. A long-term patient follow-up has shown zero deaths in relation to the implanted grafts, no thromboembolic, hemorrhagic, or infectious complications, however, 6 of them developed grafts stenosis which was treated successfully ( Schmidt & Baier, 2010 ). Despite the encouraging results with BM-MNCs, BM-MSCs still present more advantages. These advantages are, for example, their ability to differentiate into a variety of cell types even progenitor cells; relatively easy procedure for their collection, isolation, storage, and proliferation; presenting a similar phenotype to the valve cells; present anti-thrombogenic properties; and their immunogenicity is manageable ( Masuda et al. , 2008 ; Iop et al. , 2009 ). In a comparative study, Vincentelli et al. (2007) examined short- and long-term characteristics of the porcine decellularized scaffold which were processed with in-situ injections of BM-MNCs and BM-MSCs, before transplanting in a lamb of animal models. Short-term results did not show any significant differences. However, the 4 months (long-term) follow up has shown a significant decrease of transvalvular and distal gradients, more inflammatory reaction, more structural deterioration as well as calcification, and a thick fibrous pannus around the suture line in the BM- MNCs group. These observations in the BM-MNCs group were significantly different from the BM-MSCs group ( Vincentelli et al. , 2007 ). Cardiac progenitor cells Cardiac progenitor cells (CPCs) or also known as Cardiac Stem Cells (CSCs) are a type of cells which are found in the adult heart and express stem cell factor receptor kinase c-kit+, also shown to be negative for the markers of blood/endothelial cell lineage ( Beltrami et al. , 2003 ). CPCs have only been in the spotlight for about 15 years now with the credit given to Beltrami and colleagues as they demonstrated the self-renewing and multipotent characteristics of these cells; differentiating into all three different cardiogenic cell types which are cardiomyocytes, smooth muscle cells and endothelial cells ( Beltrami et al. , 2003 ). More recently, a study carried out by Vicinanza and colleagues helped to lay to rest the controversies about c-kit+ cardiac cells by demonstrating that only a small fraction of the c-kit+ adult cardiac stem cells possess the tissue-specific progenitor properties ( Vicinanza et al. , 2017 ). The study was carried out in adult Wistar rats in which an experimental acute myocardial infarction (MI) was induced and the border of the infarcts were injected with GFP+c-kit+ for the first group, “GFP-expressing CD45-c-kit+ CSCs (CSC GFP )” for the second group, and the placebo group were injected with PBS. The study showed that CSC GFP only significantly reduced apoptosis and hypertrophy of the cardiomyocytes, it also significantly reduced scarring and improved ventricular functions ( Vicinanza et al. , 2017 ). The study further demonstrated that about 10% of the overall c-kit+ cardiac cells are CD45-c-kit+, and only about 10% of these are clonogenic and multipotent. Therefore, it was inferred that only about 1–2% of the total c-kit+ myocardial cells have a demonstrable multipotent CSC phenotype ( Vicinanza et al. , 2017 ). These c-kit+ CD45-(deletion of which also renders the cells CD31-, CD34-) CPCs were also shown to express to some degree Sca-1, Abcg2, CD105, CD166, PDGFR- α, Flk-1, ROR2, CD13, and CD90 ( Vicinanza et al. , 2017 ). The number of CPCs have however been shown to be significantly higher in neonates but dramatically decreases after the age of 2 ( Chaikof, 2007 ; Bosman et al. , 2011 ). Thus, these CPCs in combination with a suitable scaffold, could be the answer to treat a number of CHD, as CPCs could be collected during palliative surgery or via endomyocardial biopsy ( Saxena, 2010 ; Torok, 2015 ). 10. 7717/peerj. 5805/fig-1 Figure 1 Schematic of the different types of stems that can be used on the biomaterial backbone for cardiovascular tissue engineering (TE). This schema represents the different types of stem cells that can be used on the biomaterial backbone (depicted as the background characters) for cardiovascular Tissue Engineering (TE). (A) Induced pluripotent stem cells (iPSCs) derived from fibroblast. (B) Prenatal, Perinatal, and Postnatal Stem cells (PPPSCs) are derived from amniotic fluid, umbilical cord, and chorionic villi. (C) Bone Marrow Stem Cells (BMSCs) such endothelial progenitor cells (EPCs) and mesenchymal stem cells (MSCs) can easily be isolated from the bone marrow. (D) Cardiac progenitor cells (CPCs) can be harvested during palliative surgery or endomyocardial biopsy. (E) Embryonic stem cells (ESCs) derived from the inner cell mass of the blastocyst. 10. 7717/peerj. 5805/fig-2 Figure 2 Promising strategies for CHDs treatment. The schematic diagram represents the potential of stem cells (SCs) and tissue engineering (TE) for corrective surgical treatment of infants as well as adolescent patients with Congenital Heart Disease (CHD). Various sources for stem cells (SCs) are presented here as alternatives to harvesting the appropriate stem cells (SCs) which can be used to seed on clinically certified biomaterial scaffolds for reconstructing functional cardiac tissue-engineered grafts. These grafts could be implanted via the corrective surgery into the heart of infants and adolescent patients with congenital heart disease (CHD) for definitive correction of cardiac defects. These optimized cardiac-tissue engineered grafts should have the potential to grow in parallel with the child, while lacking any tumorigenicity, immunogenicity, thrombogenicity, calcification, or other risk factors. The potential SCs and biomaterials for TE in CHDs are represented in the Fig. 1 below. Also, Fig. 2 below shows the promising strategies for the treatment of CHDs. Standpoint Due to the high number of patients as well as newborns who are suffering from CHD and also high costs of their implications, it is necessary and vital to making intensive research for finding an effective treatment for CHD patients. Stem cell research has shown remarkable results in all kinds of tissue engineering including skin ( Alessandri, Emanueli & Madedu, 2004 ), cartilage ( Metcalfe & Ferguson, 2007 ), vascular ( Makris et al. , 2015 ), ocular ( Chlupác, Filová & Bacáková, 2009 ) and cardiac tissues ( Miyagawa et al. , 2011 ; Sugiura et al. , 2018 ; Karamichos, 2015 ; Kane et al. , 2014 ). The importance of engineered cardiac tissue lies in the fact that synthetic non-degradable materials cannot adjust to the patient’s developing body. A number of patients who suffer from CHD are adults, and they are more suitable for this type of therapy. However, many patients with severe CHD are infants and children whose bodies’s are constantly developing. Although various synthetic and natural biodegradable biomaterials have been used so far which have shown good results, the one with the best degradation rate is yet to be found. There are various complications related to existing surgical treatments and scaffolds which cannot be ignored. Calcification, inflammatory reaction and life-long anticoagulants treatment are the most important known complications for the conventional methods of CHD treatment ( Zilla et al. , 2008 ; Brown, 2005 ). The complexity of CHD makes TE possibly the most suitable solution for treatment of patients with CHD. Conclusion The replacement or correction of a malformation in a complex system like the cardiovascular system could only be successful with tissues which can mimic the native heart and vascular tissues. SCs have opened the door to such treatments. The best SC candidates and biomaterials are yet to be identified, despite the encouraging results. All different types of SCs which have been investigated so far still present some disadvantages. Extensive research would be required to enable deeper understanding, solve drawbacks, and promote SCs use for tissue engineering in the future. All the efforts channeled at obtaining proper legal regulation for using SCs, developing new technologies for scaffold production as well as scaffoldless techniques, developing faster and safer methods for producing patient-specific iPSCs, and research into the effectiveness of SCs in TE for treatment of CHD, predicts a very positive future for patients, researchers and surgeons.

10. 7717/peerj. 5809

PeerJ

Chondrogenic differentiation of adipose-derived mesenchymal stem cells induced by L-ascorbic acid and platelet rich plasma on silk fibroin scaffold

Articular cartilage is an avascular tissue with limited regenerative property. Therefore, a defect or trauma in articular cartilage due to disease or accident can lead to progressive tissue deterioration. Cartilage tissue engineering, by replacing defective cartilage tissue, is a method for repairing such a problem. In this research, three main aspects—cell, biomaterial scaffold, and bioactive factors—that support tissue engineering study were optimized. Adipose-derived mesenchymal stem cells (ADSC) that become cartilage were grown in an optimized growth medium supplemented with either platelet rich plasma (PRP) or L-ascorbic acid (LAA). As the characterization result, the ADSC used in this experiment could be classified as Mesenchymal Stem Cell (MSC) based on multipotency analysis and cell surface marker analysis. The biomaterial scaffold was fabricated from the Bombyx morii cocoon using silk fibroin by salt leaching method and was engineered to form different sizes of pores to provide optimized support for cell adhesion and growth. Biocompatibility and cytotoxicity evaluation was done using MTT assay to optimize silk fibroin concentration and pore size. Characterized ADSC were grown on the optimized scaffold. LAA and PRP were chosen as bioactive factors to induce ADSC differentiation to become chondrocytes. The concentration optimization of LAA and PRP was analyzed by cell proliferation using MTT assay and chondrogenic differentiation by measuring glycosaminoglycan (GAG) using Alcian Blue at 605 nm wavelength. The optimum silk fibroin concentration, pore size, LAA concentration, and PRP concentration were used to grow and differentiate characterized ADSC for 7, 14, and 21 days. The cell morphology on the scaffold was analyzed using a scanning electron microscope (SEM). The result showed that the ADSC could adhere on plastic, express specific cell surface markers (CD73, CD90, and CD105), and could be differentiated into three types of mature cells. The silk fibroin scaffold made from 12% w/v concentration formed a 500 µm pore diameter (SEM analysis), and was shown by MTT assay to be biocompatible and to facilitate cell growth. The optimum concentrations of the bioactive factors LAA and PRP were 50 µg/mL and 10%, respectively. GAG analysis with Alcian Blue staining suggested that PRP induction medium and LAA induction medium on 12% w/v scaffold could effectively promote not only cell adhesion and cell proliferation but also chondrogenic differentiation of ADSC within 21 days of culture. Therefore, this study provides a new approach to articular tissue engineering with a combination of ADSC as cell source, LAA and PRP as bioactive factors, and silk fibroin as a biocompatible and biodegradable scaffold.

Introduction Unlike other connective tissues, articular cartilage is an avascular tissue which lacks a nervous and lymphatic system ( Zhang, Hu & Athanasiou, 2009 ). This unique property make it difficult for the cartilage to regenerate tissue after being damaged. Following disease, its structure tends to degrade progressively ( Hunziker, 2002 ). The absence of blood supply may also inhibit the wound healing process ( Baugé & Boumédiene, 2015 ), leading to necrosis ( Mason et al. , 2000 ), consequently causing a permanent defect in the area. Some methods have been established to treat patients with cartilage tissue defects, such as microfracture, autologus implantation, autograft, allograft, and joint replacement ( Zhang, Hu & Athanasiou, 2009 ); however, these methods have some disadvantages and cause side effects. Tissue engineering is promising for creating new cartilage tissue that is expected to become functional and ready to be implanted to replace damaged cartilage tissue. Four main parameters support the success of articular cartilage tissue engineering: cell type, growth factors or bioactive factors, mechanical stimuli, and scaffold material ( Kock, Donkelaar & Ito, 2012 ). The cell source for tissue engineering can be obtained from several tissues or organs, one of which is adipose tissue. Adipose-derived mesenchymal stem cells (ADSC) are mesenchymal stem cells (MSC) obtained from perivascular white adipose tissue, including subcutaneous adipose tissue ( Kishi, Imanishi & Ohara, 2010 ). The isolation of ADSC is relatively easy and produces a higher yield of cells compared to other adult stem cell source tissues ( Johnstone et al. , 2013 ). Bioactive factors used in tissue engineering control cell proliferation and differentiation ( Brochhausen et al. , 2009 ); In their study, L-ascorbic acid (LAA) and platelet rich plasma (PRP) were used as bioactive factors for chondrogenic differentiation of ADSC. Temu et al. (2010) showed that LAA could induce the differentiation of a ATDC5 cell line to chondrocytes. Moreover, LAA can regulate adult stem cell differentiation to some mesenchymal tissues derivatives, such as adipocytes, osteocytes, myocytes, and chondrocytes ( Choi et al. , 2008 ). LAA has a role as co-factor in post-translational modification of collagen molecules ( Gessin et al. , 1988 ), which is a component of the extracellular matrix (ECM) of chondrocytes. Previous study had shown that PRP could be a good candidate for a bioactive factor due to its composition, such as transforming growth factor-β1 (TGF-β1), platelet-derived growth factor (PDGF), epidermal growth factor (EGF), insulin-like growth factor-1 (IGF-1), and vascular endhotelial growth factor (VEGF) that are important for cell differentiation and proliferation ( Pawitan et al. , 2014 ). Moreover, PRP had been shown to stimulate cell proliferation and matrix biosynthesis of mammalian chondrocytes as well as the expression of a chondrogenic marker in a MSC 3D culture ( Akeda et al. , 2006 ; Drengk et al. , 2009 ). The objective of this research was to induce ADSC differentiation to become chondrocytes which were seeded on silk fibroin scaffolds in a PRP induction medium and a LAA induction medium. Silk fibroin was chosen as the biomaterial of the scaffold because it is biodegradable and biocompatible ( Wang et al. , 2006 ). Materials and Methods Silk fibroin scaffold fabrication We used silk fibroin fabricated using a salt leaching method as the scaffolds, as previously characterized by Judawisastra & Wibowo (2017). The silk fibroin was obtained from a Bombyx mori cocoon that was degummed to remove sericin protein which can cause biocompatibility and hypersensitivity problems in vivo ( Altman et al. , 2003 ). The cocoon was cut and immersed in 0. 05% Na 2 CO 3 solution for 1 h. The cocoon was washed in deionized water to remove residual Na 2 CO 3 solution and then dried in a fume hood overnight. Dried silk fibroin was diluted in 8 wt% CaCl 2 -Formic acid solution at room temperature with constant stirring for 15–30 min. The silk fibroin concentration was optimized by the addition of 6 gr, 8 gr, 10 gr, and 12 gr silk fibroin into 8 wt% CaCl 2 -Formic acid solution. NaCl with a specific particle size was added into the fibroin solution and homogenized. The NaCl particle size was important because the scaffold pore size was determined by it. The optimization of pore size to support ADSC proliferation was performed using MTT assay for 100, 300, and 500 µm pore size on days 1, 3, 5, 7, and 14. Data were taken in triplicate for each scaffold on each observational day. The ratio of NaCl and fibroin solution was 5:1, and then the mixture was dried in the fume hood overnight. The mixture was immersed in a 70% alcohol solution for ±30 min to induce β-sheet formation ( Terada et al. , 2016 ), and then the fibroin was immersed in distilled water for 3 days to remove salt residues. Successfully obtained silk fibroin was stored at −80 °C for 30 min for easier cutting. Before further analysis, th silk fibroin was sterilized using an autoclave for 15–20 min at 121 °C ( Sommer et al. , 2016 ). Adipose-Derived Stem Cells (ADSC) isolation and culture The ADSC was obtained from human adipose tissue with ethical approval (No. Reg. 0417060790) from The Health Research Ethics Committee from the Faculty of Medicine, Padjajaran University. The ADSC isolation method followed that of Remelia et al. (2016). Adipose tissue was processed enzymatically using H-Remedy recombinant enzyme which was added to adipose tissue in 10% v/v concentration, and then incubated at 37 °C for 1 h. After 1 h, the growth medium Dulbecco’s Modified Eagle’s Medium (DMEM) low glucose (1 g/L) (Sigma)and L-glutamin (4 mM) were added to the sample to inactivate enzymes. Then, the sample was centrifuged for 5 min with 600 g. The supernatant was discarded, and a 10 mL red blood cell lysis buffer was added into the pellet and incubated for 5 min at room temperature. The sample was centrifuged for 10 min at 600 g and the supernatant was then discarded. The pellet, which is called stromal vascular fraction (SVF), was cultured at 37 °C, 5% CO 2 to increase the number of cells or to facilitate cell proliferation. After the cells’ culture reached 80–90% confluency, they were harvested enzymatically using Trypsin-EDTA (0. 25%) and cryopreserved in liquid nitrogen. ADSC characterization ADSC characterization included specific cell surface marker analysis and differentiation potency evaluation. The specific cell surface marker refers to the protocol in the Human MSC Analysis Kit (BD Stemflow™). ADSC passage 1 that reached >80% confluency were harvested enzymatically using Trypsin-EDTA (0. 25%) (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). Cell concentration at 5 × 10 6 –10 7 cells/mL was resuspended in 1 ml staining buffer. After resuspension, a 100 µL cells solution in the staining buffer was taken and added into the 5 µL hMSC positive cocktail (CD90FITC, CD105PerCP-Cy5. 5, CD73 APC), and hMSC negative cocktail (CD34 PE, CD11b PE, CD19 PE, CD45 PE, HLA-DR PE) antibody or isotype and incubated in the dark for 30 min. Then, the sample was washed twice in the staining buffer and resuspended in 300–500 µL staining buffer. Finally, the sample was analyzed using a flowcytometer. For the differentiation analysis, ADSC passage 2 that reached >80% confluency was harvested enzymatically in Trypsin-EDTA (0. 25%) (Gibco). ADSC were cultured in 24 well-plate (1 × 10 4 sel/ well ) in the growth medium DMEM Low Glucose (Gibco) supplemented with 10% FBS (Gibco). After reaching 80% confluency, the growth medium was replaced with a chondrogenic induction medium (MesenCult™ Chondrogenesis Differentiation Kit, Vancouver, Canada), osteogenic induction medium (MesenCult™ Osteogenesis Differentiation Kit), and adipogenic induction medium (MesenCult™ Adipogenic Differentiation Medium (Human)). After 7–14 days incubation (for adipogenesis) and/or longer than 14 days (for chondrogenesis and osteogenesis), the cells were observed using an inverted microscope. The cells were fixed in 4% formaldehyde in saline and stained, using Alcian Blue staining which is specific for glycosaminoglycan, one of the components in chondrocytes extracellular matrix; Alizarin Red staining which is specific for mineralized matrix expression as osteoblast marker; and Oil Red O staining for lipid vacuoles in adipocyte marker. The excess dye stain was washed in PBS. The cell observation was performed by inverted microscope. Biocompatibility analysis of silk fibroin scaffold Biocompatibility of the silk fibroin scaffold was analyzed using 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromidefor (MTT Assay). ADSC (1 × 10 5 cells/mL) were grown on a sterile scaffold (0. 5 cm × 0. 5 cm × 1 mm) in 96 well-plate. The cells were maintained in growth medium (supplemented with DMEM) and incubated at 37 °C; 5% CO 2. The effect on ADSC at days 1, 3, 5, 7, and 14 after being cultured was evaluated using MTT assay. For the cytotoxicity assay, the growth medium was discarded, and then 10 µL of MTT reagent (5 mg/mL) was added into 100 µL growth medium. The cells were incubated in MTT solution for 4 h at 37 °C in the dark. After that, the MTT solution was removed, and 100 µL dimethyl sulfoxide (DMSO) was added to dilute the formazan crystal that had formed. The absorbance of the solution was read at 570 nm wavelength using a microplate reader (Bio-Rad). The observation was repeated three times. Scanning Electron Microscope (SEM) analysis ADSC morphology was analyzed using a scanning electron microscope (SEM) (SU 3500; Hitachi, Krefeld, Germany; Center of Advanced Science ITB) after being seeded on the scaffold. ADSC (10 6 cells/mL) were grown on the sterile scaffold (0. 5 cm × 0. 5 cm × 1 mm) in 96 well-plate. The cells were maintained in the growth medium and incubated at 37 °C; 5% CO 2. SEM analysis was done on cells cultured for 1 and 21 days. Cells were fixed in 100 µL of 2, 5% (v/v) glutaraldehyde in 0. 1 M cacodylate buffer (Electron Microscopy Science; Hatfield, PA, USA), and incubated for 24 h at 4 °C. The sample was dehydrated using an ascending alcohol series and dried by the freeze drying method for 3 h. The dried sample was sputtered with gold coating and observed under SEM. Optimization of LAA and PRP concentration Cell proliferation in optimization LAA and PRP concentration was analyzed using MTT assay, and the cell differentiation was analyzed using Alcian Blue staining. ADSC (1 × 10 5 cells/mL) were seeded on 96 well-plate. The concentrations of LAA solution in the induction medium were 25 µg/mL, 50 µg/mL, 100 µg/mL, and 200 µg/L. The concentrations of PRP solution (platelet content approximately 1. 086 × 10 6 /µL) in the induction medium with 1% heparin added were 5%, 10%, and 20%. The cell cultures were placed in an incubator at 37 °C; 5% CO 2 in the induction medium, and replaced every 2 days. Proliferation analysis was observed on days 1, 3, 5, 7, and 14. In addition, the differentiation potency was analyzed using extracellular matrix staining for sulfated-GAG as a chondrocytes marker. ADSC (10 4 cells/mL) were grown on 24 well-plate, and the cells were incubated at 37 °C; 5% CO 2, and the medium was replaced every 2 days. On days 14 and 21, the cells were prepared for Alcian Blue staining and observed using an inverted microscope to evaluate chondrogenic differentiation. Data were taken in triplicate for each group on every observational day. The intensity of blue colour as the staining result indicated that glycosaminoglycan extracellular matrix had formed. Blue colour intensity was quantified using the software Digimizer (MedCalc, Ostend, Belgium). Analysis of Glycosaminoglycan (GAG) content ADSC (10 6 cells/mL) were grown on a scaffold in LAA induction medium and PRP induction medium. On days 7, 14, and 21, the cell culture was washed in PBS before being fixed using an acetone:methanol (1:1) solution at 4 °C for 3 min. One percent Alcian Blue in 3% acetic acid was added into the cell culture. The cells were incubated for 30 min and the overstaining dye was washed in 3% acetic acid and deionized water. One percent of Sodium dodecyl sulfate (SDS) was added to the cell culture and homogenized using a shaker at 200 rpm for 30 min. The absorbance was read using a microplate reader at 605 nm wavelength. The observation was repeated three times. Result ADSC characteristics Observation using a light microscope showed that cells had a spindle-like morphology, which is flattened. The cells had formed lamellipodium, indicating their complete adherence on the polystyrene flask in the standard culture condition. According to Lotfy et al. (2014), ADSC morphology in their standard culture condition was fibroblast-like and fusiform (spindle-like shape). The population doubling time (PDT) of ADSC was 42. 9 h, while that of Griffin et al. (2017) was 39, 88 ± 4. 4 h. This difference in PDT could be due to the cell passage and genetic variation of donors determined by PDT of ADSC. The cell surface markers that were used to confirm MSC were cluster of differentiation 90 (CD90), CD73, and CD105 as positive markers, and CD45, CD34, CD11b, CD19, HLA-DR, as negative markers. The singlet and sharp peaks in the graphs show that there was a single population of the cells. The positive and negative markers had the singlet and sharp peak. The ADSC expressed the positive markers CD90 (91. 38%), CD73 (96. 25%), and CD105 (61. 33%). However, the sample and isotype peak of the negative markers overlapped, decreasing the percentage (0. 38%) ( Fig. 1 ). The MSC must express those three positive markers exceeding 90% of cell population, while the MSC must express negative markers until below 2% ( Dominici et al. , 2006 ). Our result showed that ADSC expressed CD90 (91. 38%), CD73 (96. 25%), and CD105 (61. 33%), and the negative markers were expressed only in a small percentage (0. 38%) ( Fig. 1 ). 10. 7717/peerj. 5809/fig-1 Figure 1 Flow cytometry result of specific ADSC cell surface markers. Cell expressed positive markers (A) CD90, (B) CD73, (C) CD 105, and (D) negative marker CD45, CD34, CD11b, CD19, HLA-DR. The multipotency analysis showed that ADSC complied with one of the MSC criteria, which could be differentiated into three mesenchymal stem cell derivative cells: chondrocyte, osteocyte, and adipocyte ( Fig. 2 ). 10. 7717/peerj. 5809/fig-2 Figure 2 Multipotency evaluation of ADSC. (A) Alcian Blue staining, (B) Alizarin Red and (C) Oil Red O in ADSC culture which was induced with differentiation induction medium, and (D) non-staining group. Black arrow showed positive result from each staining group. Optimization of silk fibroin scaffold To promote stronger cell adhesion and faster cell growth, often some modifications are made on scaffolds. To achieve the best support of ADSC growth, we optimized silk fibroin concentration and pore size. Figure 3 depicts that the percentage of cell viability increased gradually within 14 days in different silk fibroin concentrations, indicating that the cells seeded on all scaffolds proliferated actively. On day 14, the scaffold made from 8% w/v, 10% w/v, and 12% w/v had higher percentages of cell viability than the control group, and the 12% w/v silk fibroin scaffold had the highest score. 10. 7717/peerj. 5809/fig-3 Figure 3 Growth curve of ADSC on scaffold in different silk fibroin concentration. Figure 4 shows that all scaffolds supported cell proliferation and maintained cell viability. Apparently, pore size could affect ADSC proliferation. The percentage of cell viability in scaffold with 500 µm pore size was the highest; thus, this scaffold appears to have been the optimal one for supporting ADSC proliferation. 10. 7717/peerj. 5809/fig-4 Figure 4 Growth curve of ADSC on scaffolds that have different pore size. Biocompatibility analysis of silk fibroin scaffold SEM images were taken to characterize the structure of the silk fibroin scaffold. Based on the SEM analysis presented in Fig. 5, the scaffold made from 12%w/v silk fibroin formed 500 µm pores. The scaffold had an average pore size 536 ± 97 µm and an interconnected porous structure. 10. 7717/peerj. 5809/fig-5 Figure 5 SEM image represented morphology of scaffold. Scaffold was made from 12% w/v silk fibroin and had 500 µm pore size ( Judawisastra & Wibowo, 2017 ). (l—l) represents pore size. It is known that the surface roughness of a scaffold is an important factor in promoting cell attachment ( Acharya, Ghosh & Kundu, 2008 ). Figures 6A – 6B shows that ADSC cells formed a fillipodia structure which might be one of the factors that promoted fast attachment of ADSC cells in this study. The SEM image also shows the cell structure that attached to the scaffold on days 1 and 21 ( Figs. 6C – 6D ). The images indicate that the scaffold could support ADSC proliferation, which is in line with the growth curve ( Fig. 7 ). 10. 7717/peerj. 5809/fig-6 Figure 6 Morphology of ADSC on scaffold made from 12% w/v silk fibroin and had 500 µm pore size. Single cell was on (A) day 1 and (B) day 21. Cell population was on (C) day 1 and (D) day 21. Red arrow shows cytoplasm extension, called filopodia. White arrow and the area marked with yellow stripe line show the area covered by cells. The growth curve of ADSC determined by the MTT assay on the optimized scaffold showed the desired increase of viable cells between days 1–21 ( Fig. 7 ). The curve shows a lag phase of cell growth from days 1 to 7 and the log phase after day 7. 10. 7717/peerj. 5809/fig-7 Figure 7 Growth curve of ADSC on scaffold made from 12% w/v Silk Fibroin and had 500 µm pore size. SEM images of tissue formation on the scaffold were obtained on day 21 to examine the tissue structure as the cell had reached full density on the scaffolds surface. The SEM result showed the presence of two cell populations on the scaffolds. One formed a monolayer and the other showed aggregates. Based on the scaffolds optimization experiments above related to ADSC cell viability, growth, and proliferation, we found that 12% w/v silk fibroin with 500 µm pore size was the best candidate to promote ADSC cell adhesion and tissue formation ( Fig. 8 ). 10. 7717/peerj. 5809/fig-8 Figure 8 ADSC population observed on day 21. There are two types of cell population in (A) monolayer formation and (B) aggregation. Optimization of LAA and PRP concentration To assess the best concentration of bioactive factors to induce cell proliferation in our experiment, we added LAA or PRP at various concentrations to growth medium and examined the cell proliferation by MTT assay independently. Figure 9 shows the growth curve of ADSC in the presence of LAA at various concentrations. The graphs depict that 25 µg/mL, 50 µg/mL, 100 µg/mL, and 200 µg/mL LAA increased the percentage of cell viability from day 1 to 21 of observation. Our data showed that the addition of LAA promoted faster cell proliferation than in cells not treated with LAA, and the best LAA concentration was 50 µg/mL. However, there were no significant differences among various concentrations of LAA. This result indicates that LAA supported ADSC cell proliferation. 10. 7717/peerj. 5809/fig-9 Figure 9 Growth curve of ADSC in L-Ascorbic Acid (LAA) supplemented medium in various concentrations. The quantification using Digimizer software showed the blue color intensity in various LAA concentrations ( Fig. 10 ). The result shows that ADSC in LAA supplemented medium secretes more GAG than the cells grown in standard medium. However, the amount of GAG was not significantly different among groups ( p > 0. 05), and blue color intensity in the 50 µg/mL treatment group tended to be higher compared to the other groups. 10. 7717/peerj. 5809/fig-10 Figure 10 Graph of blue colour intensity xomparison from Alcian Blue staining in various LAA concentrations. Figure 11 showed ADSC proliferation in various PRP concentrations in the induction medium. The cells grown in the induction medium supplemented with 5%, 10%, and 20% PRP showed an increasing percentage of cell viability from days 1 to 21. The percentage of cell viability from various PRP concentrations were higher than that without PRP (supplemented with FBS). This result indicated that PRP supported ADSC cell proliferation. Among various PRP concentrations, cell viability in 10% PRP was higher than the other concentration on every observational day. From this result, we concluded that 10% PRP in the induction medium was the optimum concentration to promote cell proliferation. 10. 7717/peerj. 5809/fig-11 Figure 11 Growth curve of ADSC in various Platelet Rich Plasma (PRP) concentration of medium. Similar to the LAA experiment, GAG production in cells treated with PRP was assessed with Alcian Blue staining. The Alcian Blue intensity of cells treated with various PRP concentrations was quantified ( Fig. 12 ). As shown in Fig. 12, 10% and 20% PRP concentration caused higher production of GAG than the control group (supplemented with FBS); the blue color intensity in the 5% PRP concentration and the FBS group did not show any siginificant difference ( p > 0. 05). Furthermore, the blue color intensity in 20% PRP concentration was lower than the blue color intensity in 10% PRP concentration. Based on these results, we conclude that 10% PRP concentration had the best potency in chondrogenic differentiation. 10. 7717/peerj. 5809/fig-12 Figure 12 Graph of blue colour intensity comparison from Alcian Blue staining in various PRP concentrations. Evaluation of ADSC differentiation on scaffold in LAA induction medium and PRP induction medium Further, we examined the best concentration of LAA and the PRP effect on ADSC differentiation grown on optimized scaffold. To assess cell differentiation, production of GAG on the grown culture was analyzed based on Alcian Blue absorbance at 650 nm ( Fig. 13 ). Both the absorbance value of Alcian Blue stained cells in 50 µg/mL LAA and 10% PRP increased gradually from day 7 to 21. The absorbance value in LAA induction medium or PRP induction medium were higher than that of the control group (supplemented with FBS). Based on the graph, 50 µg/mL of LAA or 10% PRP induction medium could induce chondrogenic GAG production of ADSC grown on scaffold. 10. 7717/peerj. 5809/fig-13 Figure 13 Graph of glycosaminoglycan (GAG) content in ADSC cultured on scaffold in 50 µg/mL LAA and 10% PRP supplemented medium. Discussion Cartilage tissue damage can be overcome by implanting cartilage tissue in the damaged area, but sources of cartilage tissue are limited. In this research, we utilized MSC from adipose tissue (ADSC) to be differentiated to chondrocytes. Based on multipotency assessment and morphology observation, we confirmed that the ADSC used in this study complied with criteria described for healthy and potent ADSC. Our ADSC showed an ability to differentiate into chondrocytes, adipocytes, and osteocytes. Furthermore, it also complied with the standards of the International Society for Cellular Therapy(ISCT) ( Dominici et al. , 2006 ). Following the international standard for MSC surface markers declared by ISCT, further assessment of ADSC used in this study showed it complied with those criteria, except lower expression of endoglin marker (CD105) than the standard. However, CD105 expression of ADSC has been described as inconsistent and depends on many factors such as cell source, cell passage, isolation method, incubation duration, and growth phase of cells ( Baer & Geiger, 2012 ). Early passage cells used in our experiment might be the reason for lower expression of CD105 ( Crisan et al. , 2008 ). However, based on other markers and morphology analysis, we concluded our grown culture matches ADSC. CD105 is not the main marker to characterize MSC from adipose tissue, and it is recommended only as an alternative or additional marker ( Bourin et al. , 2013 ). The multipotency evaluation of ADSC showed that the cells could be differentiated into chondrocytes, osteocytes, and adipocytes. The same result was obtained by Hamid et al. (2012) showing that ADSC has the differentiation capacity to become chondrocyte, osteocyte, and adipocyte when cultured in a standard induction medium. Therefore, ADSC used in this research complied with MSC criteria. Based on the optimization of the scaffold, the optimum cell proliferation was obtained on the 12% silk fibroin scaffold. Based on previous study, the mechanical strength of a scaffold is an important factor in articular cartilage tissue engineering because cartilage has a role as facilitator in body mass transmission in movement ( Fox, Bedi & Rodeo, 2009 ). We speculatethat higher silk fibroin concentration increased the mechanical strength of scaffold. The optimum cell proliferation was also facilitated by the scaffold pore size 500 µm, which is the best pore size obtained in a previous experiment ( Judawisastra & Wibowo, 2017 ). The result of our research was in line with Murphy, Haugh & O’Brien (2010) concluding that a scaffold with a bigger pore size facilitates the cells to proliferate faster than the cells seeded on a scaffold with small pore size. Bigger pore size may also facilitate faster cell migration. The pore structure also affects the roughness of the scaffold’s surface, which is one of the important factors for cell attachment ( Acharya, Ghosh & Kundu, 2008 ). This research showed that within 21 days there were only two phases of growth curve pattern, compared to in the standard culture condition which has three phases in shorter observational days. This pattern was different from the ADSC growth pattern in the standard culture condition. According to ( Christodoulou et al. , 2013 ), ADSC was in lag phase on days 1–3 and entered log phase in days 3–8; and after day 8, ADSC started entering the stationary phase. The delayed growth of ADSC on the scaffold could be due to the influence of the scaffold used in this research, including the pore size. Therefore, the role of a scaffold in facilitating ADSC growth needs further investigation. Moreover, on day 21, two types of cell population were found on the scaffold: monolayer and aggregates; however, the existence of those cellpopulation types needs further investigation. According to Zhang et al. (2014), aggregation formation is very important in the initial stage of chondrogenic differentiation because it can facilitate extracellular matrix condensation. Here, we confirmed that chondrogenesis can be induced by LAA in standard growth medium. Potdar & D’Souza (2010) also showed that an LAA addition in 250 µM concentration into subcutaneous adipose tissue (SCAT) hMSC culture can increase the proliferation rate. The difference in LAA concentration that can increase cell proliferation rate could be due to the difference of cell source donor. Choi et al. (2008) also showed that the role of LAA in MSC proliferation was dose-dependent. Moreover, chondrogenesis evaluation of cells treated with LAA has shown that 50 µg/mL (170 µM) of LAA could induce optimum chondrogenic differentiation. L-ascorbic acid was well known as inducer in chondrogenesis ( Choi et al. , 2008 ; Kao et al. , 1990 ), however, the mechanism has not been studied. Based on analysis of proliferation and chondrogenic differentiation potency, the cells in 50 µg/mL LAA showed the best proliferation rate and differentiation potency. Therefore, 50 µg/mL LAA concentration was chosen as the optimum concentration and used for the next investigation in this research. The LAA concentration in 50 µg/mL (approximately 170 µM) is also a general concentration of LAA used in standard chondrogenic induction medium. In this research, PRP could also induce chondrogenesis because it contains many growth factors that can induce chondrogenesis in ADSC. In agreement with previous study, our study showed that PRP increases the proliferation rate of ADSC; specifically 10% PRP induces the accumulation of chondrogenic GAG ( Liao et al. , 2015 ). As well as LAA, ADSC response to PRP was dose dependent with an optimum concentration of 10% to stimulate ADSC proliferation and differentiation Mardani et al. (2013). Spreafico et al. (2009) showed that PRP addition in culture medium was proved to increase the proteoglycan production, although the mechanism is still unknown. The result of our research was supported by Kim et al. (2012) who observed ADSC differentiation on silk fibroin porous scaffold and showed that ADSC could grow and produce chondrogenic GAG. In comparison to Kim et al. (2012), we also evaluated the PRP addition and found that chondrogenic differentiation showed the best when the cells were induced with PRP. Here, we have shown that the engineered silk fibroin scaffold with certain concentrations of LAA and PRP can be applied to maintain the survival of ADSC and stimulate chondrogenesis, thereby providing a new and optimized method for cartilage tissue engineering. Conclusion In conclusion, 50 µg/mL of LAA or 10% PRP induction medium can induce optimum chondrogenic differentiation and growth of ADSC on 12% w/v silk fibroin scaffold which had 500 µm pore size. The combination of scaffold and induction medium could faciltate the proliferation and differentiation of ADSC, leading to chondrocytes. In the future, the result of this research could be developed for cartilage tissue engineering. Supplemental Information 10. 7717/peerj. 5809/supp-1 Supplemental Information 1 Raw data exported from the absorbance score resulted from microplate reader Biorad iMark applied for growth curve analyses for Fig. 7 The cells grown on optimized scaffold, which made from 12% w/v Silk Fibroin and had 500 µm pore size on day 1, 3, 5, 7, 14, and 21. Each observational group was quantified triplicate. The growth curves were plotted from the cell viability ( y -axis) and observational time ( x -axis). The cell viability was obtained from the absorbance value of MTT assay. Click here for additional data file. 10. 7717/peerj. 5809/supp-2 Supplemental Information 2 Raw data exported from the absorbance score resulted from microplate reader Biorad iMark applied for growth curve analyses for Figs. 3 and 4 to determine the optimum scaffold There are two parameters that observed in determining the optimum scaffold, which are pore size and natural polymers concentration in developing scaffold. The growth curves were plotted from the cell viability ( y -axis) and observational time ( x -axis). The cell viability was obtained from the absorbance value of MTT assay. Each observational group was quantified triplicate. Click here for additional data file. 10. 7717/peerj. 5809/supp-3 Supplemental Information 3 Raw data represent graph of blue color intensity comparison from alcian blue staining in various ( Fig. 10 ) LAA Concentration and ( Fig. 12 ) PRP Concentration The blue color intensity was obtained from the absorbance value ( y -axis) exported from Biorad iMark. There are 5 groups in LAA optimization process and 4 groups in PRP optimization process. Each group was repeated triplicate. Click here for additional data file. 10. 7717/peerj. 5809/supp-4 Supplemental Information 4 Raw data represent graph of glycosaminoglycan (GAG) content in ADSC cultured on scaffold in 50 µg/mL LAA and 10% PRP supplemented medium ( Fig. 13 ) The blue color intensity was obtained from the absorbance value ( y -axis) exported from Biorad iMark. There are 3 observational groups, which are ADSC on optimum scaffold with optimum LAA supplemented medium, ADSC on optimum scaffold with optimum PRP supplemented medium and ADSC on polystyrene plate. Each group was repeated triplicate. Click here for additional data file. 10. 7717/peerj. 5809/supp-5 Supplemental Information 5 Raw data represents growth curve of ADSC in various L-Ascorbic Acid (LAA) concentration of medium ( Fig. 9 ) and Platelet Rich Plasma (PRP) concentration of medium ( Fig. 11 ) To optimize the optimum LAA and PRP concentration, the ADSC were grown on LAA supplemented medium and PRP supplemented medium. The growth curves were plotted from the cell viability ( y -axis) and observational time ( x -axis). The cell viability was obtained from the absorbance value of MTT assay. Each observational group was quantified triplicate. Click here for additional data file.

10. 7717/peerj. 581

PeerJ

Fibrochondrogenic potential of synoviocytes from osteoarthritic and normal joints cultured as tensioned bioscaffolds for meniscal tissue engineering in dogs

Meniscal tears are a common cause of stifle lameness in dogs. Use of autologous synoviocytes from the affected stifle is an attractive cell source for tissue engineering replacement fibrocartilage. However, the diseased state of these cells may impede in vitro fibrocartilage formation. Synoviocytes from 12 osteoarthritic (“oaTSB”) and 6 normal joints (“nTSB”) were cultured as tensioned bioscaffolds and compared for their ability to synthesize fibrocartilage sheets. Gene expression of collagens type I and II were higher and expression of interleukin-6 was lower in oaTSB versus nTSB. Compared with nTSB, oaTSB had more glycosaminoglycan and alpha smooth muscle staining and less collagen I and II staining on histologic analysis, whereas collagen and glycosaminoglycan quantities were similar. In conclusion, osteoarthritic joint—origin synoviocytes can produce extracellular matrix components of meniscal fibrocartilage at similar levels to normal joint—origin synoviocytes, which makes them a potential cell source for canine meniscal tissue engineering.

Introduction Meniscal injury is a common cause of lameness and pain of the dog. Due to the virtually absent healing response in the majority of the meniscus, injured meniscal tissue is commonly removed to relieve the clinical signs of lameness, joint locking, and painful popping. Unfortunately, partial meniscectomy hastens the development of secondary arthritis ( Berjon, Munuera & Calvo, 1991 ; Connor et al. , 2009 ; Cox et al. , 1975 ) and thus patient lameness. Tissue engineering methods are being investigated to address this challenge of meniscal injury and loss. One of the great obstacles to achieving the reality of tissue engineered menisci is determination of an ideal cell source for in vitro culture and extracellular matrix (ECM) formation. Because cells cannot be synthesized de novo, they must be harvested autologously, or obtained from living or deceased tissue donors. When determining ideal cell sources for tissue engineering, location of the source tissue, quantity of donor tissue available for harvest, and ability to harvest the cells in a minimally invasive fashion must be considered. Autologous cells are particularly attractive because they have a low potential for infectious disease transmission ( Pessina et al. , 2008 ) and immunogenic tissue rejection ( Hamlet, Liu & Yang, 1997 ; Ochi et al. , 1995 ; Rodeo et al. , 2000 ). While producing normal, healthy menisci in vitro is the ultimate goal of tissue engineering, use of normal meniscal cells from a healthy donor site would cause irreversible patient harm and thus is a poor choice for meniscal tissue engineering. Mesenchymal stem cells were recently identified in normal joint—origin canine synovium ( Zhang, Dietrich & Lopez, 2013 ), which could be used towards in vitro meniscal fibrochondrogenesis. When cultured in monolayer, cells obtained from the synovial membrane of normal joints are primarily positive for CD90 (marker for stemness), CD29 ( β - integrin), CD44 (hyaluronic acid receptor) and negative for markers of hematopoetic progenitors (CD34) and leukocyte antigens (CD45; Zhang, Dietrich & Lopez, 2013 ). These cells are also able to undergo chondrogenesis when cultured in pellet form ( Zhang, Dietrich & Lopez, 2013 ). However, clinical use of normal autologous synoviocytes as a tissue engineering cell source would require surgery on another unaltered joint within the patient’s body. Autologous, osteoarthritic joint—origin synovium has been investigated as a cell source for fibrocartilage tissue engineering in dogs, because of its abundance and ease of harvest during clinically required surgical procedures ( Warnock et al. , 2012 ). In vitro, cultured osteoarthritic joint—origin canine synovial membrane cells are plastic adherant and fibroblast—like, and contain populations of cells that can undergo chondrogenesis ( Warnock et al. , 2011 ; Warnock et al. , 2013 ), suggesting the presence of mesenchymal stem cells. In vivo synovium also has the ability to form fibrocartilage ECM ( Smith et al. , 2012 ; Tienen et al. , 2006 ). Conversely, synoviocytes in osteoarthritic joints secrete a number of inflammatory mediators and destructive matrix metalloproteinases ( Benito et al. , 2005 ; Fiorito et al. , 2005 ; Sutton et al. , 2007 ), which could inhibit in vitro fibrochondrogenic potential. For example, canine osteoarthritic joint—origin synoviocytes produce less total collagen than normal joint—origin synoviocytes in monolayer culture ( Warnock et al. , 2011 ). This limitation may not be present with improved culture conditions ( Warnock et al. , 2013 ). For instance, osteoarthritic joint—origin synovial fluid stem cells require culture as a micro-mass to undergo efficient in vitro chondrogenesis, compared to cells derived from healthy joint fluid ( Krawetz et al. , 2012 ). Thus, the purpose of this study was to evaluate and compare the fibrochondrogenic potential of synoviocytes from osteoarthritic and normal canine joints that were cultured as tensioned bioscaffolds under conditions previously shown to increase meniscal-like ECM content in canine osteoarthritic joint—origin synoviocytes ( Warnock et al. , 2013 ). We hypothesized that with the use of this tensioned culture system, there would be no difference in cell viability and fibrocartilage-like ECM formation between tensioned synoviocyte bioscaffolds from normal joints (normal joint—origin tensioned synoviocyte bioscaffolds, “nTSB”) and osteoarthritic joints (osteoarthritic joint—origin tensioned synoviocyte bioscaffolds, “oaTSB”). Materials and Methods Tissue harvest With informed owner consent, synovium was obtained from 12 dogs with naturally occurring (stifle) osteoarthritis as per Institutional Animal Care and Use Committee approval. Dogs were treated for degeneration of the cranial cruciate ligament and medial meniscal injury via exploratory arthroscopy, partial meniscectomy if indicated, and tibial plateau leveling osteotomy. Synovial villi were arthroscopically harvested during routine partial synovectomy using a tissue shaver (Stryker, San Jose, CA) as previously described ( Warnock et al. , 2012 ). Synovial villi from the osteoarthritic joints were immediately placed in a 50 ml polypropylene tube containing 40 mL of Dulbeccos’ Modified Eagle’s Media (DMEM, Invitrogen) with 10% fetal bovine serum (FBS, Invitrogen), warmed to 37 °C. The tube was transported immediately to the laboratory and centrifuged at 313 xg, media was decanted, and tissue fragments transferred by pipette and sterile forceps into a digestion solution as described below. Normal synovium was also harvested from six dogs which were euthanatized via sodium pentobarbital overdose for reasons unrelated to the study, as per the Institutional Animal Care and Use Committee Protocol and in accordance with the American Veterinary Medical Association Humane Euthanasia Guidelines. Dogs were assessed by a Diplomate of the American College of Veterinary Surgeons—Small Animal to not have any orthopedic disease based on medical history, pre-mortem physical examination, and post-mortem gross joint evaluation. Post-mortem, a lateral arthrotomy and patellar tendon transection was performed on each stifle joint. The parapatellar, suprapatellar, lateral, and medial wall synovium were dissected off the joint capsule using a #15 bard parker blade. Synovium was transported as described above. In the laboratory, synovium harvested from normal joints was additionally was minced into 2 × 3 mm pieces using sterile technique. Cell culture Osteoarthritic joint—origin synovial villi and normal joint—origin synovium tissue fragments were completely digested with sterile Type 1A clostridial collagenase 10 mg/mL in RPMI 1640 solution (Invitrogen) over 2–6 h at 37 °C. Tissue was deemed to be completely digested when no ECM could be visualized microscopically at 20 × objective magnification. Cells were cultured in monolayer for four passages to isolate Type B fibroblast-like synoviocytes ( Vasanjee et al. , 2008 ) and Type C intermediate synoviocytes ( Vasanjee et al. , 2008 ) as described previously ( Warnock et al. , 2012 ). The following media formulation was used for the duration of culture: high glucose DMEM, supplemented with 17. 7% FBS, 0. 021 mg/mL glycine, 0. 025 mg/mL L-alanine, 0. 037 mg/mL L- asparagine, 0. 038 mg/mL L-aspartic acid, 0. 042 mg/mL L-glutamic acid, 0. 033 mg/mL L-proline, 0. 030 mg/mL L-serine, 0. 23 mg/mL pyruvate, 0. 52 mg/mL L-glutamine, 6. 75 mg/mL HEPES buffer, 0. 15 mg/mL L-ascorbic acid -2 phosphate, 177. 0 units/mL penicillin, 177. 0 µg/mL streptomycin, and 0. 44 µg/mL amphoterocin. The flasks were incubated at 37. 6 °C, 5% CO 2, 95% humidity, with sterile media change performed every 24 h. Cell flasks were observed under 10x objective magnification every 24 h to assess confluency. Cells were passaged upon reaching 95% confluence, which was defined as monolayer cell culture with no visible exposed flask surface in between cells, and no overlap of the cells on each other. At harvest and at each passage cell viability counts were performed using the trypan blue exclusion assay ( Strober, 2001 ). At the 4th passage, cells from each joint were transferred into eight 150 cm 2 flasks and allowed to become hyperconfluent cell sheets, defined as cells overlapping each other in greater than 100% confluency ( Fig. 1 ). TSB were then made as previously described ( Warnock et al. , 2013 ). Briefly, hyperconfluent cell sheets were dislodged off the flask floors ( Ando et al. , 2008 ), and each sheet was wrapped over 2. 0 cm diameter, 22 ga cerclage wire hoops in three layers, with approximately 0. 5 N of tension to avoid tearing, to synthesize TSB. The TSB were placed in 6-well plates in 9. 0 mL of the above described culture media, with the free end of the cell sheet facing down to prevent loosening. Bioscaffolds were harvested for analysis after a total of 30 days in culture ( Ando et al. , 2008 ; Tan, Zhang & Pei, 2010 ). Bioscaffold analyses Bioscaffold analyses examined presence of ECM components responsible for meniscal form and function. These include type I and type II collagen, ( Kambic & McDevitt, 2005 ; Eyre & Wu, 1983 ); α - smooth muscle actin (ASM), ( Ahluwalia et al. , 2001 ; Kambic, Futani & McDevitt, 2000 ; Spector, 2001 ); and glycosaminoglycans (GAG), ( Adams & Ho, 1987 ; Nakano, Dodd & Scott, 1997 ; Stephan, McLaughlin & Griffith, 1998 ), including aggrecan ( Valiyaveettil, Mort & McDevitt, 2005 ). Differences in expression of inflammatory mediators or presence of macrophages were investigated as these factors may be associated with decreased in vitro ECM synthesis in osteoarthritic joint—origin synoviocytes ( Fiorito et al. , 2005 ; Pei et al. , 2008 ). Cell viability One TSB per dog was washed three times in sterile phosphate buffered saline and immersed in 4 µM ethidium homodimer and 6 µM acetomethoxy calcein (calcein –AM) solution (Ethidium homodimer and Calcein AM Live/Dead Viability Assay; Invitrogen, Carlsbad, CA) for 20 min at 37. 6 °C, 5% CO 2, 95% humidity. Cells were then visualized in at least five regions of the bioscaffolds, (and two in the center and three on the periphery, at approximately the 2, 6, and 10 o’clock positions) using a laser microscope (Eclipse Ti-u Laser Microscope; Nikkon, Japan). The number of viable (green) and non-viable (red) cells per each field counted by hand. Due to the complex three-dimensional nature of the bioscaffolds, these cell counts provided an estimate of cell viability. Immunohistologic analysis Two TSB per dog were fixed in 10% buffered formalin, paraffin embedded, and tissue blocks cut in 4 µm sections for histologic and immunohistologic analysis. All slides were labelled with randomly generated acquisition numbers and analyzed in a blinded fashion. Sections were stained with Hematoxylin and Eosin (“H&E”), Masson’s Trichrome, and Toluidine Blue. Cell morphology and general ECM architecture was assessed using H&E; organization and intensity of collagen staining was described using Masson’s Trichrome, and intensity of GAG staining was assessed using Toluidine Blue. Immunohistochemistry Immunohistochemistry was performed as previously described ( Warnock et al. , 2012 ) for type I collagen (AB749P; 1:100 dilution; Millipore), type II collagen (AB746P; 1:100: Millipore), macrophage MAC387 receptor to determine type A synoviocyte content, (CBL260; 1:200 dilution; Millipore); and alpha smooth muscle actin (M0851; 1:30; Dako). Extracellular and intracellular immunoreactivity intensity and prevalence was scored as previously described ( Wakshlag et al. , 2011 ) with some modifications: immunoreactivity was localized to intracellular or extracellular staining, and ECM immunoreactivity intensity was described and scored, as negative (0), mild (1), moderate (2), or strong (3) staining. As determined by hand count, intracellular immunoreactivity and extracellular immunoreactivity was categorized as positive in <10%, 10–50%, or >50% of cells and sample area, respectively. Each of these histologic observations was assigned a score ( Table 2 ). Then a histologic intensity coefficient was calculated for each ECM component, as follows: [[(Extracellular matrix staining intensity score) × (percentage area coverage of positive staining score)] + [(Intracellular staining intensity score) × (percentage positive staining cells score)]]/2 ( Table 1 ). 10. 7717/peerj. 581/table-1 Table 1 Histologic scoring system. Extracellular and intracellular immunoreactivity intensity was localized to intracellular or extracellular staining, and ECM immunoreactivity intensity was described and scored as negative (0), mild (1), moderate (2), or strong (3) staining. As determined by hand count, intracellular immunoreactivity and extracellular immunoreactivity was categorized as positive in <10%, 10–50%, or >50% of cells and sample area, respectively. Each of these histologic observations was assigned a score ( Table 2 ). Then a histologic intensity coefficient was calculated for each ECM component, as follows: [[(Extracellular matrix staining intensity score) × (percentage area coverage of positive staining score)] + [(Intracellular staining intensity score) × (percentage positive staining cells score)]]/2 ( Table 1 ). None <10% 10–50% >50% None Mild Moderate Strong % Positive staining cells Intracellular staining intensity Intracellular score 0 1 2 3 0 1 2 3 % Positive stained extracellular area Extracellular staining intensity Extracellular score 0 1 2 3 0 1 2 3 10. 7717/peerj. 581/table-2 Table 2 Gene expression in tensioned synoviocyte bioscaffolds. The effect of osteroarthritis on fibrochondrogenic gene expression of tensioned synoviocyte bioscaffolds (fold-changes ± SEM). Fold changes were calculated using the following formula: fold change = 2 −ΔΔ C T = [( C T gene of interest − C T housekeeping gene GAPDH) oaTSB − ( C T gene of interest − C T housekeeping gene GAPDH) nTSB ]. Tensioned synoviocyte bioscaffolds (TSB) Dog Normal Osteoarthritis SEM P -value Gene: N = 4 N = 7 SOX-9 0 Reference + 1. 17 1. 54 0. 72 Collagen type I α 1 0 + 6. 88 2. 62 0. 04 Collagen type II α 1 0 + 71. 1 4. 48 0. 02 Aggrecan 0 −1. 15 1. 77 0. 84 Interleukin-6 0 −19. 0 2. 01 0. 001 Tumor Necrosis Factor α 0 + 1. 49 2. 55 0. 77 Tissue weight One TSB per dog was lyophilized and a dry weight obtained. Samples were digested in 1. 0 ml Papain Solution (2 mM Dithiothreitol and 300 µg/ml Papain) at 60 °C in a water bath for 24 h. This papain digest solution was used to obtain double stranded DNA (dsDNA), GAG and collagen content of the bioscaffolds. DNA quantification Double stranded DNA quantification assay (The Quant-iT PicoGreenTM Assay; Invitrogen) was performed per manufacturer’s instructions; double stranded DNA extracted from bovine thymus was used to create standards of 1, 000, 100, 10, and 1 ng/mL. Standard and sample fluorescence was read by a fluoromoter (Qubit; Invitrogen) at 485 nm excitation/ 528 nm emission, and dsDNA concentration was determined based on the standard curve. Biochemical ECM analysis Glycosaminoglycan content was determined by the di-methyl-methylene blue sulfated glycosaminoglycan assay ( Farndale, Buttle & Barrett, 1986 ) using a spectrophotometer (Synergy HT– KC4 Spectrophotometric Plate Reader and FT4software, BioTec, Winooski, VT). Collagen content was determined by Erlich’s hydroxyproline assay, as described by Reddy & Enwemeka (1996). Hydroxyproline content was converted to collagen content using the equation: µg hydroxyproline x dilution factor/0. 13 = µg collagen ( Ignat’eva et al. , 2007 ), because hydroxyproline consists of approximately 13% of the amino acids in human meniscal collagen ( Fithian, Kelly & Mow, 1990 ). Collagen and GAG content were standardized to tissue dry weight as percentage of dry weight, to compare the experimental neotissues to previously reported normal meniscal ECM content ( Eyre & Wu, 1983 ). Total GAG and collagen content were also reported in µg/neotissue to measure total synthetic activity over the course of 30 days in each TSB. GAG and collagen content were additionally standardized to dsDNA content using the following equations: [µg GAG/µg dsDNA] ( Li & Pei, 2011 ) and [µg collagen/ µg dsDNA] to identify chondrogenic cellular activity of each tested cell origin. Real-time RT-PCR One TSB per dog was snap frozen in liquid nitrogen and stored at −80 °C. Total RNA was isolated using the phenol-chloroform extraction ( Chomczynski & Sacchi, 1986 ) with slight modifications. Samples were pulverized using a liquid nitrogen-cooled custom-made stainless steel pulverizer and homogenized in trizol (Trizol; Qiagen Sciences, 0. 025 mL/mg of tissue) and mixed with chloroform. The aqueous phase was then treated with isopropanol to precipitate nucleic acids. RNA of samples was purified using on-column DNAse digestion (RNeasy; Qiagen Sciences). The RNA quality and quantity was determined using capillary electrophoresis (RNA 6000 Nano LabChip Kit, Agilent 2100 Bioanalyzer; Agilent Technologies), and RNA integrity numbers (Imbeaud S, 2005) were determined (2100 Expert software, Agilent Technologies). First-strand cDNA synthesis was performed from 400 ng total RNA (SuperScript III First-Strand Synthesis System; Invitrogen Life Technologies, Carlsbad, CA) and Oligo-(dT) 20 primers. To control for possible genetic DNA contamination, non reverse-transcribed samples were also processed. Pre-designed primers and probes (Taq-Man ® Primers and Probes; Applied Biosystems Inc. , Foster City, CA) were obtained for each of the genes of interest: IL-1 β, IL-6, TNF- α, SOX-9 (an embryonic chondrogenic transcription factor), collagen type I α 1, collagen type II α 1, aggrecan, and the reference gene GAPDH (see Appendix S1 ). All assays were confirmed to amplify their targets at 95% or greater efficiency using RNA from tissues of interest. Quantitative real-time PCR was performed (StepOnePlus RT-PCR System; Applied Biosystems Inc. ) using a proprietary reagent system (TaqMan Gene Expression Master Mix; Applied Biosystems Inc. ) Controls included template-free negative controls and non reverse-transcribed negative controls. All samples were run in triplicates and all negative controls were run in duplicates for 40 cycles (15 s at 95 °C, 1 min at 60 °C) after 2 min of incubation with Uracil-DNA Glycosylase at 50 °C, and 10 min at 95 °C of enzyme activation. Quantitative gene expression was determined in triplicates using the comparative CT method ( Schmittgen & Livak, 2008 ). The gene GAPDH was used as internal control (housekeeping gene). Threshold cycles (CT) for each gene were defined by recording the cycle number at which fluorescence reached a gene-specific threshold. Fold changes for gene expression data were calculated using the following formula: fold change = 2 −ΔΔ CT = [( C T gene of interest − C T housekeeping gene GAPDH) n TSB − ( C T gene of interest − C T housekeeping gene GAPDH) oa TSB ]. Statistical methods A D’Agostino & Pearson omnibus normality test was performed on all data to test for normality. Cell harvest data was non-parametric data and was analyzed with a Wilcoxon matched-pairs signed rank test, and data reported as median and interquartile range. Significance was declared at P < 0. 05. Data were analyzed with a statistical software program Graph Pad Prism, San Diego, CA. The effect of osteoarthritis (osteoarthritic versus normal joint status) on gene expression and ECM composition was analyzed using a 2-tailed Student’s t -test, assuming unequal variances. The effect of osteoarthritis on the histologic scoring of TSB extracellular matrix formation was analyzed using a non-parametric Mann–Whitney U-test; ranking of the histologic scores was performed using a Kruskal-wallis analysis on ranks followed by a Fisher’s exact test. Significance was declared at P < 0. 05. Data were analyzed using Statistical Analysis System, version 9. 3 (SAS Institute Inc. , Cary, NC). Results Cell harvest The mean age of dogs with stifle osteoarthritis was 4. 7 years (range: 2–8 years). Breeds represented included: Golden Retriever (1), American Staffordshire Terrier (2), Labrador Retriever (3), Australian Shepherd (1), Rottweiler (2), Boston Bull Terrier (1), Goldendoodle (1), and mixed breed (1), with 7 neutered males, 4 spayed females, and one intact female dog. As observed by a Diplomate of the American College of Veterinary Surgeons–Small Animal, all dogs had marked villous synovial hyperplasia and osteophytosis, and grade 1–2 Outerbridge cartilage lesions of the medial femoral condyle and tibial plateau ( Outerbridge, 1961 ). Cell yield from arthroscopic synovial debris was 1. 9 × 10 6 ± 3. 7 × 10 5 cells per joint, and cells were 99. 5% ± 0. 002 viable at harvest. Mean age of dogs with normal stifles was 4. 3 years (range: 3–6 years); breeds represented included: Red Tick Hounds (4), Labrador Retriever (1), and American Staffordshire cross (1), with 3 female intact dogs, 2 male intact dogs, and one neutered male. Cell yield per joint was 1. 4 × 10 7 ± 2. 6 × 10 6 per joint and cells were 99. 5% ± 0. 01 viable. As the entire stifle joint synovial membrane could be harvested post mortem, a greater volume of tissue and thus greater cell numbers were obtained from the normal joints versus arthroscopic harvest of the osteoarthritic joints ( P = 0. 01). Cell culture and cell characterization At 4th passage, cells were transferred into eight 150 cm 2 flasks in order to have enough TSB for tissue analyses. This, however, resulted in greater cell seeding numbers for nTSB versus oa TSB. Thus, normal joint—origin synoviocytes were seeded at 1. 49 × 10 7 cells per flask, wherease 6. 52 × 10 6 osteoarthritic joint—origin cells were seeded per flask. At 4th passage, normal joint—origin cells were 99. 0 ± 0. 4% viable compared with 98. 8 ± 0. 4% viability of osteoarthritic joint—origin cells ( P = 0. 85). Culture duration from tissue harvest to hyperconfluent cell membrane formation and synthesis of TSB was 37. 6 days and similar for both cell origins (range 20–49 days). During the first week of tensioned bioscaffold culture, the culture media phenol red pH indicator changed to yellow by the time the 24 h media change was required, indicating marked increase in media acidity. In addition, during the first 7–10 days of culture, approximately 2–3 bioscaffolds per normal and osteoarthritic joint unraveled or slipped off their wire hoops (no group differences observed), and were not analyzed in this study. The typical appearance of intact nTSB and oaTSB is pictured in Fig. 2 ; thickness of TSB was 2–3 mm. 10. 7717/peerj. 581/fig-1 Figure 1 Hyperconfluent cell sheets. (A) Representative example (A) of a hyperconfluent cell sheet in a cell culture flask, in monolayer culture, just prior to harvest for formation of a tensioned synoviocyte bioscaffold; (B) phase contrast photomicrograph of the hyperconfluent cell sheet, 10 × objective magnification, bar = 100 µm. 10. 7717/peerj. 581/fig-2 Figure 2 Tensioned synoviocyte bioscaffolds. Representative samples of a tensioned synoviocyte bioscaffold synthesized from normal joint origin synoviocytes, or nTSB (“N”), and a tensioned synoviocyte bioscaffold from osteoarthritic joint origin synoviocytes, or “oaTSB, ” (“OA”). At harvest, tensioned bioscaffolds from normal dogs had a dry weight of 39. 3 mg (range 27. 5–50. 4 mg), which was more than for oaTSB (23. 6 mg, range 10. 2–50. 1 mg; P = 0. 008). Mean estimated cell viability of nTSB and oaTSB was similar, with 78% of cells viable (range: 72%–86%). Cell viability was not associated with peripheral versus central location on the TSB. Laser microscopy revealed cells with fusiform, fibroblastic morphology, oriented parallel with the vector of tension, as well as the presence of acellular, circular regions in the bioscaffolds. Hematoxylin and eosin staining revealed highly cellular bisocaffolds, with layers of fibroblastic cells organized in parallel, as sheets or bands, or variably arranged in whorls, with eosinophilic ECM ( Fig. 3 ). Subjectively, both nTSB and oaTSB had heterogeneous extracellular matrix architecture and cell distribution, with regions of dense cellularity and regions of dense extracellular matrix ( Fig. 3 ). 10. 7717/peerj. 581/fig-3 Figure 3 Bioscaffold cellularity and alpha-smooth muscle actin expression. Hematoxylin and Eosin stain of normal joint—origin synoviocyte bioscaffolds (“N H&E”) and osteoarthritic joint—origin synoviocyte bioscaffolds (“OA H&E”). Immunohistochemisty for alpha smooth muscle actin (ASM) of normal joint—origin synoviocyte bioscaffolds (“N ASM”) and osteoarthritic joint—origin synoviocyte bioscaffolds (“OA ASM”). Note the more extensive expression of ASM in the osteoarthritic joint—origin synoviocyte bioscaffold cells, and regional strong ASM expression along the periphery of spontaneously forming defects ( * ) in the normal and osteoarthritic joint—origin bioscaffolds. Immunohistochemistry negative controls of normal joint—origin synoviocyte bioscaffolds (“N NC”) and osteoarthritic joint—origin synoviocyte bioscaffolds (“OA NC”). 10 × objective magnification, bar = 100 µ m. Percent dsDNA content was used to quantify tissue cellularity. Despite an initial higher seeding cell count at 4th passage, dsDNA accounted for 0. 11 ±. 02% of nTSB dry weight, versus 0. 21 ± 0. 03% of oaTSB dry weight ( P = 0. 01). Immunohistologically, oaTSB had more ASM positive cells than nTSB; the median histologic score for nTSB was 6 versus 9 for oaTSB ( p = 0. 0102, Fig. 4 ). Nine of 12 oaTSB had the highest possible ASM histologic scores of 9, whereas none of the nTSB achieved a perfect score of 9 ( P = 0. 009, Fig. 4 ). In 50% of all bioscaffolds, ASM positive cells were concentrated around the bisocaffold periphery and around the margins of what appeared to be spontaneously forming bioscaffold defects ranging from 70–600 µm ( Fig. 3 ). These defects corresponded with the circular acellular regions viewed on laser microscopy. The other 50% of bioscaffolds did not contain circular defects, nor did ASM expression seem to be geographically localizable. 10. 7717/peerj. 581/fig-4 Figure 4 Histology scores for tensioned synoviocyte bioscaffolds. Scatter plots for histology scores for type I collagen, type II collagen, glycosaminoglycan, and alpha-smooth muscle actin in normal joint—origin tensioned synoviocyte bioscaffolds versus osteoarthritic joint—origin tensioned synoviocyte bioscaffolds, showing the median (thin line) and interquartile range (bolded bars). Histologic scores for collagens type 1 and 2 were calculated as follows: Histologic score= [[(% positive staining cells × intracellular staining intensity) + (% positive stained extracellular area × extracellular staining intensity)]/2]. The histologic score for alpha smooth muscle actin (ASMA) was calculated by (% positive staining cells × intracellular staining intensity). The histologic score for glycosaminoglycan (GAG) was calculated by (% positive stained extracellular area × extracellular staining intensity). An (∗) denotes statistical significance. Gene expression The oaTSB had a greater gene expression of type I collagen (7-fold increase; P = 0. 04) and type II collagen (71-fold increase; P = 0. 02) and a lower gene expression of interleukin-6 (19-fold decrease; P = 0. 001) versus nTSB. No significant changes were observed for relative expression of SOX-9 ( P = 0. 72), aggrecan ( P = 0. 84), and tumor necrosis factor- α ( P = 0. 77; Table 2 ). Interleukin-1 β was not expressed at detectable levels in any bisocaffolds. Glycosaminoglycan content The total GAG content of oaTSB was lower than the GAG content of nTSB ( P = 0. 02; Table 3 ). After adjustment for dry weight or DNA content, no significant group differences were observed. 10. 7717/peerj. 581/table-3 Table 3 Extracellular matrix and dsDNA content of tensioned synoviocyte bioscaffolds. The effect of osteoarthritis on extracellular matrix and double stranded- DNA composition of tensioned synoviocyte bioscaffolds. Data is reported as mean ± SEM. Tensioned synoviocyte bioscaffolds (TSB) Dog Normal Osteoarthritis P -value N = 6 N = 12 Concentrations (µ g/neotissue): Glycosaminoglycan 684 ± 74 434 ± 44 0. 02 Collagen 4855 ± 1270 3302 ± 392 0. 29 DNA 47. 4 ± 11. 9 42. 3 ± 5. 2 0. 71 Proportion (% dry weight): Glycosaminoglycan 1. 73 ± 0. 11 2. 05 ± 0. 16 0. 11 Collagen 12. 1 ± 2. 4 16. 6 ± 2. 4 0. 20 DNA 0. 111 ± 0. 020 0. 214 ± 0. 030 0. 01 Index (µ g/µ g dsDNA): GAG 18. 7 ± 3. 9 11. 4 ± 1. 6 0. 13 Collagen 132 ± 36 92. 0 ± 17. 5 0. 35 Glycosaminoglycan was deposited regionally in all bioscaffolds but more GAG staining was observed in oaTSB than in nTSB. Median GAG histologic score was 1. 0 for nTSB and 3. 0 for oaTSB ( P = 0. 0007, Figs. 4 and 5 ). Only 1 of 6 nTSB had a GAG histologic score above 1, whereas 11 of 12 oaTSB had GAG histologic score above 1 ( P = 0. 004, Fig. 4 ). 10. 7717/peerj. 581/fig-5 Figure 5 Histologic analysis of glycosaminoglycan content of tensioned synoviocyte bioscaffolds. Toluidine Blue staining for glycosaminoglycan of normal joint—origin tensioned synoviocyte bioscaffolds (“NTB”) and osteoarthritic joint—origin tensioned synoviocyte bioscaffolds (“OATB”). 10 × objective magnification, bar = 100 µm. Collagen content There was no difference in quantified total collagen content of oaTSB and nTSB ( Table 3 ). Similar results were observed after adjustment for dry weight or DNA content. Masson’s Trichrome staining revealed collagen deposited in bands, sheets, and whorls, containing and surrounded by numerous fibroblastic cells lined in parallel with the orientation of the collagen ( Fig. 6 ). A significant difference in the median type I collagen histologic scores of nTSB and oaTSB could not be detected, which were 7. 5 and 6. 0, respectively ( P = 0. 11, Fig. 4 ). However, 4 of 6 nTSB had a type I collagen histologic score greater than 7. 5, vs. only 1 of 12 oaTSB had a score of 7. 5 ( P = 0. 02, Fig. 4 ). Histologically, nTSB had more type II collagen than oaTSB ( Fig. 6 ); median type II collagen histologic scores were 4. 0 in nTSB and 2. 5 in oaTSB, ( P = 0. 03, Fig. 4 ). None of the oaTSB had a score greater than 2. 5 whereas 5 of 6 nTSB had a collagen type II histology score of 2. 75 ( P = 0. 0007, Fig. 4 ). 10. 7717/peerj. 581/fig-6 Figure 6 Histologic analysis of collagen content of tensioned synoviocyte bioscaffolds. Masson’s Trichrome staining for collagen of normal joint—origin synoviocyte bioscaffolds (“NMT”) and osteoarthritic joint—origin synoviocyte bioscaffolds (“OAMT”). Immunohistochemistry for type I collagen and type II collagen of normal joint—origin synoviocyte bioscaffolds (“NCOL1” and “NCOL2”) and osteoarthritic joint—origin synoviocyte bioscaffolds (“OACOL1” and “OACOL2”). In this example the type I collagen ECM of both bisocaffolds is moderately positive. For type II collagen, the cells are moderately immunoreactive and the ECM is mildly immunoreactive in the normal joint—origin synoviocyte bioscaffold, while the cells and ECM of the osteoarthritic joint—origin bioscaffold are mildly immunoreactive. 10 × objective magnification, bar = 100 µm. Synovial macrophage content Based on immunohistochemistry, no macrophages (Type A synoviocytes) were found in any bisocaffolds ( Fig. 7 ). 10. 7717/peerj. 581/fig-7 Figure 7 Immunohistochemical analysis for macrophages. Immunohistochemistry for the macrophage MAC387 receptor, in a lymph node (positive control, “PC MAC ”), a normal joint—origin bioscaffold (“ N MAC ”) and an osteoarthritic joint—origin synoviocyte bioscaffold (“OA MAC ”). Negative controls are as pictured in Fig. 3. Note the lack of immunoreactivity in the bioscaffold samples. The positive control, a canine lymph node, is shown at 20 × objective magnification, bar = 100 µm, to allow clearer viewing of the macrophages. To show adequate representation of the bioscaffold neotissues, N MAC and OA MAC are pictured at 10 × objective magnification, bar = 100 µ m. Discussion Previous studies comparing in vitro canine synoviocyte fibrochondrogenesis in monolayer culture ( Warnock et al. , 2011 ), and canine synoviocyte chondrogenesis in micromass culture ( Krawetz et al. , 2012 ) concluded that osteoarthritic synoviocytes had inferior in vitro fibrochondrogenic potential, compared with normal synoviocytes. Fiorito and colleagues ( 2005 ) came to a similar conclusion in a study comparing in vitro chondrogenesis of human synoviocytes grown in pellet culture, as determined by histologic analysis. In contrast, with the culture conditions in the present study, especially providing conditions for self-tensioning, cells originating from osteoarthritic joints increased type I and II collagen gene expression, and oaTSB contained similar total collagen content, as compared to nTSB. While tissue dry weight and thus total GAG content of oaTSB was lower than nTSB, a significant difference in GAG content standardized to dry weight and cellularity could not be detected between oaTSB and nTSB. Histologic analysis using toluidine blue, a semi-quantitative measure of GAG, revealed more GAG deposition in oaTSB than nTSB. Thus, the greater dry weight of nTSB versus oaTSB was likely due to unmeasured ECM components such as fibronectin, type III and VI collagen, and vitronectin ( Ando et al. , 2007 ; Ando et al. , 2008 ), which are found in native synovium ( Okada et al. , 1990 ; Price, Levick & Mason, 1996 ). These findings also indicate that given the chance to self-tension, autologous, diseased synoviocytes can produce the ECM components of fibrocartilage in vitro at a comparable level of normal joint—origin synoviocytes. The unstable mechanical environment and inflammatory environment of the cranial cruciate ligament deficient joint favors synovial intimal hyperplasia and synovial membrane and joint capsule fibrosis ( Bleedorn et al. , 2011 ; Buckwalter, 2000 ; Oehler et al. , 2002 ; Smith et al. , 1997 ), all of which were encountered in the osteoarthritic joints in the present study. The in vivo pathogenic synovial hyperplasia may have accounted for the collagen gene upregulation seen in oaTSB. Rat and human osteoarthritic synoviocytes spontaneously express TGF β -1 and its receptor ( Fiorito et al. , 2005 ; Mussener et al. , 1997 ), which is a pro-collagen and chondrogenic growth factor ( Daireaux et al. , 1990 ; Leask & Abraham, 2004 ; Miyamoto et al. , 2007 ; Pangborn & Athanasiou, 2005a ; Pangborn & Athanasiou, 2005b ; Pei, He & Vunjak-Novakovic, 2008 ). Upregulation of TGF β -1 and its receptor may also be a plausible mechanism for oaTSB collagen gene upregulation. Collagen II upregulation seemed to occur independently of SOX-9 expression, a finding duplicated in cultured human osteoarthritic chondrocytes ( Aigner et al. , 2003 ). Additionally, decreased expression of IL-6 gene may be a mechanism for the observed upregulation of type II collagen genes in oaTSB; IL-6 has been found to inhibit chondrogenic differentiation of murine marrow mesenchymal cells ( Wei et al. , 2013 ). Further research is required to confirm the mechanism of hyaline chondrogenic ECM formation in canine TSB, through immunohistochemistry of TGFbeta-receptor and SMAD-family protein expression ( Xu et al. , 2012 ). Despite equal quantities of non-specific collagen in nTSB and oaTSB, immunohistologic analysis revealed less type I and type II collagen in oaTSB, particularly in the ECM. Post translation regulation by prolyl-4-hydroxylases ( Grimmer et al. , 2006 ) or ECM degradation by synovial matrix metalloproteinases ( Fiorito et al. , 2005 ) may have decreased oaTSB accumulation of type I and II collagen accumulation, despite increased collagen gene expression. IL-6 has also been found to increase gingival fibroblast synthesis of type I collagen in vitro ( Martelli-Junior et al. , 2003 ), and increase type I collagen synthesis by tenocytes in vivo ( Andersen et al. , 2011 ). It is possible that the decreased IL-6 gene expression in oaTSB synoviocytes also decreased type I collagen formation as seen on histologic analysis. One weakness of our study was that expression of type I and II collagen was not corroborated with a Western blot, nor quantified via ELISA, to further our understanding of this discrepancy between histologic collagen expression and collagen gene expression. Additionally we did not characterize the percentage and type of mesenchymal progenitor cells present in normal versus osteoarthritic synovium; difference in number and chondrogenic potential of these cells may have also accounted for a difference in collagen ECM formation. Other osteoarthritic cell types, such as chondrocytes, have reduced cell proliferation compared to normal cells in monolayer culture ( Acosta et al. , 2006 ). In contrast, oaTSB contained more dsDNA per dry weight than nTSB, despite the lower harvest cell yield and lower cell seeding density at 4th passage of osteoarthritic joint—origin synoviocytes. There was an intrinsic weakness of our study; by clinical necessity, synovium from osteoarthritic joints was harvested using a different technique (arthroscopy) than the normal joints (arthrotomy), and more synoviocytes can be obtained via arthrotomy. Although cell growth kinetics was not the focus of this study, cell culture media containing 17. 7% FBS likely provided mitotic stimuli to support and increase oaTSB cellular proliferation. The markedly hyperplastic state of the synovium in vivo may also have primed the osteoarthritic cells to continue to proliferate in vitro. Cell viability was high at harvest and at the start of 4th passage, but declined in all TSB possibly due to the long culture period. Additionally, as evidenced by media color changes, inadequate nutrient delivery to TSB in the culture wells and daily shifts in pH may have also led to nTSB and oaTSB cell mortality. This cell mortality may have affected ECM formation in both groups: the collagen content of nTSB (12%) and oaTSB (16%) did not reach that of the healthy meniscus, at 60–70% of dry weight ( McDevitt & Webber, 1990 ), although the GAG content of nTSB (1. 7%) and oaTSB (2%) did approximate the 2–3% GAG per dry weight of the whole meniscus ( McDevitt & Webber, 1990 ; Stephan, McLaughlin & Griffith, 1998 ). Consistent with prior studies ( Warnock et al. , 2013 ), all oaTSB and nTSB in the present study were negative for any macrophages, which have been reported to contaminate human osteoarthritic synoviocyte monolayer cultures and reduce in vitro chondrogenic activity ( Pei et al. , 2008 ) by contributing to the inflammatory milieu. In the present study, 4 passages and long term culture as TSB likely eliminated any non-adherent cells, including synovial macrophages ( Krey, Scheinberg & Cohen, 1976 ). Synovium from osteoarthritic joints has also been found to express inflammatory cytokines ( Fiorito et al. , 2005 ). Both nTSB and oaTSB expressed similar RNA quantity of the TNF α gene, indicating an inflammatory response in in vitro culture ( Lindroos et al. , 2010 ), independent of the diseased status of the cell origin. Paradoxically, IL-6 expression was decreased in oaTSB. Although the exact reason for this is unclear, decreased IL-6 gene expression may represent the response of synoviocytes from osteoarthritic joints to the change in environment; from the high motion, inflamed stifle containing multiple injured cell types (ligament, cartilage, meniscus, synovium) to the static tension of TSB culture and high FBS concentration cell culture media. Decreased IL-6 in oaTSB may have reflected better mechanical homeostasis ( Asparuhova, Gelman & Chiquet, 2009 ; Chan et al. , 2011 ; Gardner et al. , 2012 ) in the cells in oaTSB: the majority of cells in oaTSB were uniformly positive for ASM, while 10–50% of nTSB cells were ASM positive. Synoviocytes increase expression of intracellular ASM in response to TGF β -1 ( Xu et al. , 2012 ). Endogenous receptivity in osteoarthritic origin-joint synoviocytes to TGF β -1 present in FBS ( Goddard, Grossman & Moore, 1990 ; Mussener et al. , 1997 ) may explain increased ASM in the cells of oaTSB. In the present study, staining for ASM was positively associated with the formation of circular defects, indicating that the ECM was not strong enough to prevent tears from forming during ASM-mediated self-tensioning ( Kambic, Futani & McDevitt, 2000 ; Vickers et al. , 2004 ; Warnock et al. , 2013 ). Given the higher dsDNA content of oaTSB and the high cellularity of the TSB, these defects may have also been caused by increased cell turnover. Conclusion When cultured as TSB in high concentrations of FBS, osteoarthritic joint—origin synoviocytes can produce ECM components of meniscal fibrocartilage at similar levels to normal joint—origin synoviocytes. Potential reasons for this include increased collagen and decreased IL-6 gene expression and the greater GAG and ASM staining in oaTSB compared with nTSB Osteoarthritic joint—origin synoviocytes are a viable cell source toward meniscal tissue engineering. Further investigation of culture environments to optimize cell viability and ECM formation and strength are justified due to the promising data reported here. Supplemental Information 10. 7717/peerj. 581/supp-1 Supplemental Information 1 Supporting data file for open data policy Data sheet showing cell viability, ECM content, gene expression, and histologic scores for tensioned bioscaffolds originating from cell harvested from normal and osteoarthritic joints. Click here for additional data file. 10. 7717/peerj. 581/supp-2 Supplemental Information 2 Informed consent form for animal owners Click here for additional data file. 10. 7717/peerj. 581/supp-3 Supplemental Information 3 Animal care and use exemption form Click here for additional data file. 10. 7717/peerj. 581/supp-4 Appendix S1 Appendix 1 Assays used for RT-PCR Click here for additional data file.

10. 7717/peerj. 6178

PeerJ

Artificial liver research output and citations from 2004 to 2017: a bibliometric analysis

Background Researches on artificial livers greatly contribute to the clinical treatments for liver failure. This study aimed to evaluate the research output of artificial livers and citations from 2004 to 2017 through a bibliometric analysis. Methods A list of included articles on artificial livers were generated after a comprehensive search of the Web of Science Core Collection (from 2004 to 2017) with the following basic information: number of publications, citations, publication year, country of origin, authors and authorship, funding source, journals, institutions, keywords, and research area. Results A total of 968 included articles ranged from 47 citations to 394 citations with a fluctuation. The publications were distributed in 12 countries, led by China ( n = 212) and the US ( n = 207). There were strong correlations of the number of citations with authors ( r 2 = 0. 133, p < 0. 001), and countries ( r 2 = 0. 275, p < 0. 001), while no correlations of the number of citations with the years since publication ( r 2 = 0. 016, p = 0. 216), and funding ( r 2 < 0. 001, p = 0. 770) were identified. Keyword analysis demonstrated that with the specific change of “acute liver failure, ” decrease in “bioartificial livers” and “hepatocyte, ” and increase in “tissue engineering” were identified. The top 53 cited keyword and keyword plus (including some duplicates counts) were identified, led by bioartificial liver (405 citations) and hepatocyte (248 citations). The top 50 cited keywords bursts were mainly “Blood” (2004–2008), “hepatocyte like cell” (2008–2015), and “tissue engineering” (2014–2017). All keywords could be classified into four categories: bioartificial livers (57. 40%), blood purification (25. 00%), clinical (14. 81%), and other artificial organs (2. 78%). Discussion This study shows the process and tendency of artificial liver research with a comprehensive analysis on artificial livers. However, although it seems that the future of artificial livers seems brighter for hepatocyte transplantation, the systems of artificial livers now are inclined on focusing on blood purification, plasma exchange, etc.

Introduction Owing to the high mortality of liver failure, numerous studies had investigated its treatment. Artificial livers were a promising kind of treatment for liver failure. In clinical practice, different kinds of bioartificial systems, blood purification and hepatocyte transplantation had already been applied in liver failure ( Naruse, Tang & Makuuch, 2007 ; Wallin, 2005 ). However, a great number of obstacles and setbacks were experienced in investigating artificial livers ( Carpentier, Gautier & Legallais, 2009 ). In 1992 and 1994, Ash et al. (1992) and Ash (1994) set a precedent in performing the first liver dialysis which was previously named as Biologic-DT. Subsequently, Stange and Mitzner ( Stange et al. , 2000 ) initially developed the Molecular Adsorbent Recirculating System (MARS). Furthermore, various bioartificial liver systems had been reported in nine clinical studies since 1990 ( Park & Lee, 2005 ). Although the artificial liver was considered as one of the most effective treatment for liver failure, the definition of an artificial liver still remains obscure. Additionally, in 2006, Onodera et al. (2006) classified artificial livers into three categories: bioartificial liver, blood purification, and hepatocyte transplantation. In the early works, liver support devices were considered as the treatments of liver failure, and these devices have developed into two different strategies: blood purification and bioartificial liver. The purpose of the former was the removal of toxins in the blood related to coma and cerebral edema, while the latter’s intention was to provide metabolic, detoxic, and synthetic function of hepatocytes. However, both of these devices were restricted to the lack of both suitable animal models of liver failure and complete understanding of the pathophysiology of liver failure. Thus, hepatocyte transplantation had developed. Bibliometric sciences provide a statistical and quantitative analysis of publications and offer a convenient way to visibly measure researchers’ efforts in the investigation of a specific field ( Ashok et al. , 2016 ; Yao et al. , 2018 ). Though bibliometric methods aim to make comments on qualitative features, the major purpose of their analysis is to transform something intangible (scientific quality) into a manageable entity ( Wallin, 2005 ). Citations analysis is a bibliometric process that determine the influence of an article and make further explanations according to the original information about authors and journals ( Schmidt et al. , 2014 ). The more cited an article is, the greater influence it makes in a specific area ( Eyre-walker & Stoletzki, 2013 ). Despite the inevitable limitations ( Wallin, 2005 ) of bibliometric analysis in assessing research quality, bibliometric analysis is widely known as one of the best measurements of research trend ( Zhou et al. , 2018 ). This study is the first multicenter retrospective study on artificial livers with bibliometric analysis, based on Onodera et al. ’s (2006) definition of artificial livers. It aimed at evaluating artificial livers output and citations from 2004 to 2017 with a bibliometric analysis, which helps researchers understand the development process of artificial livers and provide guidance for the future direction of further researches. Materials and Methods Search strategy With strict inclusion criteria, Web of Science (WoS) provides adequate researches for literature analysis, including keywords, authors, institutions, countries, and publication years, which were vital for bibliometric analysis. Therefore, we conducted a citation search of the WoS Core Collection from 2004 to 2017. The keywords and free words were “Bioartificial Liver” or “Bioartificial Livers” or “Liver, Bioartificial” or “Livers, Bioartificial” or “Artificial Liver” or “Artificial Livers” or “Livers, Artificial” on topic field. All retrieval was conducted on 25th May 2018 to avoid possible changes in citation rate. After all data were extracted, articles were ranked by citation number. Inclusion and exclusion Two independent investigators (M. Z. He and X. H. Bian), respectively, screened all the titles and abstracts to select eligible articles according to the inclusion criteria ( Fig. 1 ). Any questions were resolved by discussion and the help of the third independent investigator (Y. Li). Data were retrieved from literature if meeting the following criteria: (1) the main content should be related to the study; (2) literature could be any type of researches (such as article, review, editorial material, and meeting abstract); (3) literature involving other topics related to the study. The exclusion criteria were: (1) duplicates; (2) the titles and abstract of papers were not related to the study; (3) the abstract of paper was not accessible. 10. 7717/peerj. 6178/fig-1 Figure 1 Overview of article selection process. Evaluation of included articles The following data were extracted from the 968 included articles by one investigator (M. Z. He) as followed: (1) the number of publications; (2) citations; (3) publication year; (4) country of origin; (5) authors and authorship; (6) funding source, (7) journals; (8) institutions; (9) keywords and research fields. Articles that were collaborative work of authors from multiple countries were also identified. In addition, the top 18 cited (T18) journals of the top 100 cited (T100) articles were evaluated with impact factors (Ifs) which were in accordance with the 2017 edition of journal citation reports: science edition (2017–2018). Statistical analysis With GraphPad Prism version 6. 0, the Spearman test was used to evaluate the strength and direction of the linear relationship between the number of T100 citations and the number of authors, year since publication, funding, and countries in each publication. Furthermore, the Spearman test with GraphPad Prism was also applied to evaluate the correlation of citation index between different databases (WoS Core Collection and Scopus) and the correlations of the number of impact factor with the number of T100 articles per journals and the number of average citations. The citation index was measured as the true impact of an article independent of short-lived trends ( Liu et al. , 2016 ) and the impact factor usually serves as indicator for reflecting the average number of yearly citations for recent papers published in the journal ( Garfield, 2006 ). Both of them were used to assess the equality of research output in most bibliometric analyses. All probability values were two-tailed, meanwhile, the threshold of the number of T100 citations, countries, and authors for significance was set at p < 0. 001, while others were not. Bibliometric analysis Using the online analysis platform of literature metrology ( http://bibliometric. com/ ), the trend of the number of publications, the trend of country of origin, the cooperation between countries, and the trend of keywords were shown. All the data were standardized by the frequency of occurrence. With CiteSpace, the network map of authors, co-authors and institutions were shown by publication year, and the number of citations via citation-tree rings, while the network map of keywords and keyword plus were shown only according to number of citations ( Chen, 2004 ). Results Total number of published items The number of published items on artificial livers was considered as an index of research productivity. A total of 2, 954 papers were identified after the initial record in the WoS from 1986 to 2017. From 2, 954 records, 1, 622 titles were excluded, because 1, 532 titles were unrelated to artificial livers, 90 titles were not mainly discussing artificial livers, and four titles were duplicates. In addition, 364 articles were excluded, because 231 abstracts of which were not accessible, 89 abstracts were not mainly discussing artificial livers, and 44 abstracts were unrelated to artificial livers. Thus, 968 articles were extracted in WoS during the period from 2004 to 2017. The number of articles reached a peak of 95 in 2005 and rapidly decreased to a nadir of 63 in 2007, then fluctuated up and down during 2008–2015 ( Fig. 2A ). However, it reached the lowest point in 2016 and grew again in 2017. 10. 7717/peerj. 6178/fig-2 Figure 2 Numbers of included articles (A) and the growth trends of countries (B) from 2004 to 2017. Country of origin A total of 12 different countries contributed to the literature on artificial livers ( Fig. 2B ). China and the US produced similar number of articles during 2004–2017, while a decrease in the number of articles was identified for Japan. Among the international cooperation ( Fig. 3 ), and the collaboration between the US and China was the most frequent, followed by US-Germany. 10. 7717/peerj. 6178/fig-3 Figure 3 Interactions between countries of the included articles. Citation count and possible factors influencing citations Among the 968 included articles, T100 publications on artificial livers were identified by WoS and ranked by the number of citations ( Table 1 ; Table S1 ). The median number of citations was 198 (ranged from 47 to 394), with 20 papers cited over 100 times. The median number of citation index was 36. 31 (ranged from 3. 61 to 69), and there was strong correlation between citation index and the number of the citations ( r 2 = 0. 343, p < 0. 001; Fig. S1A ) per article in WoS. In addition, the number of citations and citation index of every article in Scopus database were strongly correlated, and the number of citation index between WoS and Scopus was also strongly correlated ( r 2 = 0. 350, p < 0. 001; r 2 = 0. 990, p < 0. 001, respectively; Figs. S1B and S1C ). Both the mean values and the standard deviations (SDs) of the number of citations and citation index for WoS and Scopus Database were evaluated. The mean values of the number of citations for these two databases were 79. 270 (SD = 45. 425, ) and 83. 000 (SD = 49. 560), respectively, while the mean values of citation index were independently 0. 040 (SD = 0. 023) and 0. 041 (SD = 0. 025). The oldest cited paper was written by Chan et al. (2004) and was published in 2004. The latest paper written by Larsen et al. (2016) was published in 2016. 10. 7717/peerj. 6178/table-1 Table 1 T100 most-cited articles ranked by the number of Times cited. Rank Articles Publication year Cited times (Web) Citation index (Web) Cited times (Scopus) Citaitons index (Scopus) 1 Demetriou AA, Brown RS, Busuttil RW, et al. Prospective, randomized, multicenter, controlled trial of a bioartificial liver in treating acute liver failure. [J]. Annals of Surgery, 2016, 239(5):667–670. 2004 394 0. 197 427 0. 213 2 Lee PJ, Hung PJ, Lee LP. An artificial liver sinusoid with a microfluidic endothelial-like barrier for primary hepatocyte culture [J]. Biotechnology & Bioengineering, 2007, 97(5):1340–1346. 2007 220 0. 110 236 0. 118 3 Kelm JM, Fussenegger M. Microscale tissue engineering using gravity-enforced cell assembly [J]. Trends in Biotechnology, 2004, 22(4):195–202. 2004 191 0. 095 200 0. 100 4 Medical applications of membranes: Drug delivery, artificial organs and tissue engineering 2008 177 0. 088 197 0. 098 5 Stamatialis DF, Papenburg BJ, Gironés M, et al. Medical applications of membranes: Drug delivery, artificial organs and tissue engineering [J]. Journal of Membrane Science, 2008, 308(1–2):1–34. 2006 161 0. 080 162 0. 081 Note: See Table S1 for a complete list of T100. To identify the factors that determined the number of citations of T100 articles, we studied possible correlations between the number of citations and years since publication, authors, funding, and countries ( Fig. 4 ). There were strong correlations of the number of citations with authors ( r 2 = 0. 133, p < 0. 001) and countries ( r 2 = 0. 275, p < 0. 001). While no correlations of the number of the citations with years since publication ( r 2 = 0. 016, p = 0. 216), and funding ( r 2 < 0. 001, p = 0. 770) were identified. 10. 7717/peerj. 6178/fig-4 Figure 4 Correlations between the number of citations and the number of funding (A), the number of years since publication (B), the number of countries (C), and the number of authors (D). Journals More than 60 journals contributed to T100 publications, and T18 journals of T100 publications are listed in Table 2. Articles were most frequently published in Biomaterials ( n = 9), followed by Tissue Engineering ( n = 7), Liver Transplantation ( n = 6), and Biotechnology and Bioengineering ( n = 6). T18 journal IFs of T100 articles ranged from 1. 59 to 17. 016 with the median number of IFs (9. 303). Many of the T100 articles were published in high-IF journals, however, these IFs were poorly correlated with the number of T100 articles ( r 2 = 0. 023, p = 0. 556; Fig. S2A ), and the number of average citations ( r 2 = 0, 060, p = 0. 345; Fig. S2B ). 10. 7717/peerj. 6178/table-2 Table 2 T18 Journals of the T100 publications were included. Journal No. of articles (citations) Impact factors 2017–2018) Biomaterials 9(686) 8. 806 Tissue engineering 7(445) 3. 508 Biotechnology and bioengineering 6(536) 3. 952 Liver transplantation 6(502) 3. 752 Journal of hepatology 3(209) 14. 911 Biomedical microdevices 3(159) 2. 077 Biotechnology letters 3(138) 1. 846 Annals of surgery 2(504) 9. 203 Journal of membrane science 2(228) 6. 578 Hepatology 2(198) 14. 079 Trends in biotechnology 2(184) 13. 578 Transplant immunology 2(179) 1. 655 Gut 2(160) 17. 016 Journal of cellular and molecular medicine 2(157) 4. 302 Journal of bioactive and compatible polymers 2(154) 1. 598 Tissue engineering part B-reviews 2(121) N/A Stem cells 2(115) 5. 587 World journal of gastroenterology 2(109) 3. 300 Note: N/A, not available. Authorship and institutions More than 3, 500 authors contributed to the included articles and the majority (89%) of T100 articles were produced by cooperation work involving ≥3 authors. The top five cited authors ranked by the number of articles were listed in Table 3. The most frequently appearing authors was Chamuleau, RAFM, who authored 32 included articles (two as first and five as corresponding author) with total 387 citations, followed by Li, LJ, who authored 31 included articles (three as first and 28 as corresponding author) with a total of 127 citations. 10. 7717/peerj. 6178/table-3 Table 3 The top five cited authors ranked by the number of articles. Rank Authors No. of articles First Citations of first Correspond Citations of correspond Total citations 1 Chamuleau, RAFM 32 2 60 5 88 387 2 Li, LJ 31 3 26 28 122 127 3 Hoekstra, R 29 3 17 16 121 346 4 Ding, YT 24 0 0 19 97 113 5 Ijima, H 24 7 25 13 43 76 CiteSpace detected the information on author and co-cited authors and presented them through a network map ( Figs. 5A and 5B ). Among the cooperation between authors, Li, LJ ranked the first (cooperated with 30 authors), followed by Chamuleau, RAFM (cooperated with 28 authors). According to the network map of co-cited authors, Demetriou AA (210 citations) ranked first, followed by Nyberg SL (154 citations), and Van de K (153 citations). 10. 7717/peerj. 6178/fig-5 Figure 5 Network map of authors (A), co-cited authors (B), institutions (C), and keywords (D). Among the top five institutions of included articles ( Table 4 ), the leading institutions with the most productive articles was University Amsterdam ( n = 85, with 1, 066 citations), followed by Zhejiang University ( n = 53, with 160 citations). In addition, University Pittsburgh, National University Singapore, and Harvard University possessed 49, 44, and 40 included articles, respectively (with 211, 211, and 225 citations, respectively). 10. 7717/peerj. 6178/table-4 Table 4 The top five cited institutions of included articles on artificial livers. Rank Institution No. of articles No. of citations 1 University of Amsterdam, AZ Amsterdam, Netherlands 85 1, 066 2 Zhejiang University, Hangzhou City, Zhejiang Province, Peoples of Republic China 53 160 3 University of Pittsburgh, Pennsylvania, US 49 211 4 National University of Singapore, Singapore 44 211 5 Harvard University, Cambridge, Massachusetts, US 40 225 CiteSpace identified the information on institutions and showed them through a network map ( Fig. 5C ). Among the network map, Zhejiang University ranked the first with the cited numbers of 45, followed by Harvard University (with 33 citations), Kyushu University (with 33 citations), and University of Amsterdam (with 28 citations). Keywords and research fields Keywords with bibliometric analysis provided information about directions and trend of research. A rough estimate of research changes could be found with Fig. 6. The keywords of “bioartificial liver” and “hepatocyte” appeared less and less during 2004–2017, however, “tissue engineering” was appeared more frequently after 2011, and almost became the most frequently used keywords during the period from 2015 to 2017. In addition, it was not difficult to find that each time the keyword of “acute liver failure” increase, either “bioartificial livers” or “tissue engineering” follows the rise. 10. 7717/peerj. 6178/fig-6 Figure 6 Growth trends of keywords on artificial livers from 2004 to 2017. Top 53 cited keyword and keyword plus (including some duplicates counts) ranked by the count (the total citations of the publications in which the keyword and keyword plus appeared) were identified and analyzed with CiteSpace ( Fig. 5D ; Fig. S1 ), among which bioartificial liver ranked the first ( n = 405), followed by hepatocyte ( n = 248), in vitro ( n = 172) and transplantation ( n = 151). Using CiteSpace software, we extracted top 50 cited (T50) keywords with the strongest citation bursts ( Fig. S3 ), clearly indicating the research frontiers over time. The time interval was a blue line and the time period that represents a burst keyword category was a red line, suggesting the beginning and the end of the time interval of each burst ( Bornmann & Mutz, 2015 ). The keywords with the strongest citation bursts were firstly ignited by “blood” from 2004 to 2007, followed by “blood purification” (2006–2010) and “hepatocyte like cell” (2008–2015). Keywords with the strongest citation bursts after 2010 were listed as follows: “prometheus” (2012–2013), “plasma exchange” (2014–2017), and “tissue engineering” (2014–2017). More than 3, 900 keywords were extracted from the 968 publications. After the data were combined with consent word, 1, 732 keywords were ranked by the number of occurrences. Among them, 108 keywords appearing at least seven times, were identified and analyzed with Table 5. These keywords showed the main content of studies on artificial livers and research trend of this field. These keywords were classified into four domains based on two independent investigators (Meizhi He, Yan Li): bioartificial livers, blood purification, clinical, and other artificial organs. Among the 108 keywords, the bioartificial livers domain ranked the first with the highest percentage (57. 40%). The blood purification domain was followed with 25. 00%, the clinical was 14. 81%, and other artificial organs was 2. 78%. 10. 7717/peerj. 6178/table-5 Table 5 Keyword on artificial livers from 2004 to 2017. Domain Topic Percentage with keywords, % Frequency of keyword occurrence ( n ) Bioartificial livers domain (total, 57. 40%) Human liver technology 15. 74 Academic medical center bioartificial liver (83), artificial liver (58), 3D cell culture (55), artificial liver support system (27), 3D models (26), Artificial (17), 3D co-culture (14), Hepatocyte cultivation (14), Liposome (12), engineered liver (11), artificial implantable devises (10), 3D (9), Genomics (8), Histology (8), 3D visualization (7), Cell culture (7), Human artificial mini chromosome (7), Human liver tissues and organs 14. 81 Hepatocyte (39), adipose-derived stromal cell (18), Anatomy (14), Induced pluripotent stem (iPS) cells (12), Embryonic stem (ES) cell (11), Endothelial cell (10), Liver (9), Stem cell (9), adipose-derived stem cells (8), HepG2 (8), Liver progenitor cell (8), Mesenchymal stem cell (8), Bone marrow marrow mesenchymal stem cells (7), Feeder cells (7), Human hepatoblastoma cell line (7), High cell-density structures (7) Bioartificial livers device 10. 19 Flat-bed configuration (28), Fluidized bed (24), Scaffold (10), Biomaterials (11), Bioartificial liver assist device (11), Hybrid artificial liver (10), Bioartificial organs (9), Bioartificial (8), Bioartificial liver (8), Xenotransplantation (8), PERV (8) Liver function index 6. 48 A hepatic time (36), Autoimmune (25), Albumin synthesis (31), Growth factor (14), Glycosylation (10), Hepatic function (9), ALT (7) Hepatocyte related substance 5. 56 Amanitaphalloides (95), Bilirubin (68), Alginate beads (62), Antioxidant (59), Endotoxin (29), Apheresis (20) Bioreactor 4. 63 Airlift Reactor (78), Galactose (37), Flat membrane bioreactor (29), Bioreactor (12), Galactosylated membrane (9), Blood purification Domain (total, 25. 00%) In vitro technique 11. 11 Extracorporeal (30), Bioaffinity separation (18), Animal cell culture engineering (17), Hemodiafiltration (15), Fluid shear stress (14), Fractionated plasma separation and absorption (13), A combined rotational mold system (12), Molecular adsorb entre circulating system (12), Mass transfer (9), Extra corporeal liver perfusion (9), Hypothermia (8), Fluidization (7) Molecular substance 7. 41 Ammonia (21), EROD (12), Medium (10), hemoglobin-based oxygen albumin (8), carrier (8), Acetaminophen (7), Midazolam (7), Glutathione S-transferase expression (7) Extracellular substance 3. 70 Extracellular matrix (24), EGF (20), E-Cadherin (18), Hepatic growth factors (13) External device 2. 77 Adsorption columns (17), Extracorporeal Liver Assist Device (14), Energy systems (9) Clinical domain (total, 14. 81%) Clinical treatment 8. 33 Major hepatectomy (16), intra operative shunt (14), Extended liver resection (13), Liver cell therapy (12), Liver support treatment (12), Gene therapy (10), Hepatocyte transplantation (8), Multi objective optimization (8), FDA guidelines (7) Disease 6. 48 Acute liver failure (15), Alcoholic hepatitis (13), end-stage liver disease (11), acute-on-chronic liver failure (10), Hepatic failure (9), Hepatocellular carcinoma (8), Acute poisoning (7), Other artificial organs domain (total, 2. 78%) Other artificial organs 2. 78 Artificial heart blood pump (20), Artificial heart (clinical) (13), Artificial bone (12) The human liver technology topic of the human liver domain showed the highest percentage (15. 74%), followed by human liver tissues and organs (14. 81%). Among them, the most frequent keyword was “Amanitaphalloides” ( n = 95) in hepatocyte related substance topic, followed by “academic medical center bioartificial liver” ( n = 83) in human liver technology topic, “airlift reactor” ( n = 78) in Bioreactor topic, “bilirubin” ( n = 68), “alginate beads” ( n = 62), and “antioxidant” ( n = 59) in Hepatocyte related substance topic, “artificial liver” ( n = 58) in human liver technology topic. Additionally, “extracorporeal” ( n = 30), “hepatocyte” ( n = 39), “a hepatic time” ( n = 36), “albumin synthesis” ( n = 31), “galactoses” ( n = 37), and “3D cell culture” ( n = 55) appeared 30 or more than 30 times. While in an analysis of research area of T100 publications ( Table 6 ), Biochemistry ranked first with the number of 30 publications, followed by Applied Microbiology ( n = 24), Cell Biology ( n = 21), and Gastroenterology & Hepatology ( n = 21), respectively. 10. 7717/peerj. 6178/table-6 Table 6 The top five cited research area. Rank Research area No. of articles 1 Biochemistry 30 2 Applied microbiology 24 3 Cell biology 21 4 Gastroenterology & hepatology 21 5 Engineering 16 Discussion Total number of published items This study is the first bibliometric analysis on artificial livers. The results of the number of articles each year had demonstrated that the research on artificial livers fluctuated from 2005 to 2014. There were twice descents in 2005–2007 and 2014–2016, probably for the following reasons: (1) the average annual incidence rate of liver failure decreased ( Ichai & Samuel, 2011 ); (2) some investigators had finished their projects on artificial livers and changed their research orientation ( Sugawara, Nakayama & Mochida, 2012 ; Todo & Furukawa, 2004 ); (3) WoS only covered a small part of publications ( Tanaka et al. , 2011 ). In fact, comparing with the growth trend of countries, the number of publications of each country remained stable, except for Japan which demonstrated a sharp decline since 2005. A possible explanation was that the incidence of diseases that can further develop into liver failure, such as acute liver injury, decrease by years in Japan and most researchers in Japan would like to investigate liver transplantation rather than artificial livers ( Ogura et al. , 2010 ). All in all, we concluded that the research on artificial livers was still progressing in a relatively slow pace but was extremely promising which requires researchers to solve the temporary problems in a better way. Country of origin and institutions In growth trends of countries, the result showed that China played an important part in the research on artificial livers. This was probably due to the largest population of domestic hepatitis B patients ( Kjaergard et al. , 2003 ) and an undesirable liver transplantation situation in China ( Nielsen, 1998 ). The high incidence rate of hepatitis B in China largely led to this increasing number of investigations of interest ( Chen, Shen & Xiang, 2011 ). US, Japan, and Germany collectively contributed to a predominant number of publications, verifying that countries with higher economic ranking were associated with larger quantity and better quality of biomedical publications ( Qiu et al. , 2010 ). In addition, USA, Japan, and Germany were developed countries and had a huge disease burden of liver failure, promoting more research interest and more research funding as well in artificial liver researches ( Gillum et al. , 2011 ). Figure 2B indicated that artificial live researches in the US were influential worldwide. The results of country of origin also lead to the results of institution. This finding confirmed that the great demand in effective treatment strategies could encourage the development of scientific researches in this field. Citation count and possible factors influencing citations Among the 968 included publications, we extracted T100 publications and made a list in a table for advanced analysis. T100 publications on artificial livers were cited 47–394 times, with only 20 publications cited over 100 times. This number was quite small, for the reason that citations differed between different professional domains, mainly contributing to the number of researchers in specific medical fields ( Tang et al. , 2016 ). In McDowell et al. ’s (2017) research on pediatric liver transplantation, the citations of T10 publications ranged from 175 to 635, significantly fewer than those of hypertension (2, 242–7, 248) ( Oh & Galis, 2014 ) and diabetes (3, 420–10, 292) ( Zhao et al. , 2016 ). In order to minimize the effect of publication time on citations, the citation index was applied to estimate publications impact on their field in a short period. The results showed that there is high correlation between citations and citation index, indicating that citations were hardly affected by citation period. Additionally, high correlations between WoS and Scopus on citations were found with the premise that the citation index was strongly related to citations in Scopus. Besides, through investigating four possible factors that influence citations, the result demonstrated that the more authors and more countries work together, the higher quality of an articles could be. However, this phenomenon contradicted Ahmed et al. (2016) research, which concluded that two-authored article was the most suitable. It seemed that the effect of the number of authors differs between specific research fields. No correlation between the number of citations and the number of years since publication was identified, which might be related to the tendency of citing particular papers in researchers ( Azer & Azer, 2016 ). There was also no correlation found between the number of citations and the number of funding, indicating that the scientific impact of researchers was only weakly limited by the number of funding ( Fortin & Currie, 2013 ). Journals Numerous studies had shown that the impact factor of journal was the best indicator for citations ( Saha, Saint & Christakis, 2003 ). However, citations were not clearly affected by IFs of the journals in this analysis ( Hecht, Hecht & Sandberg, 1998 ). They found that the majority of the considered as “highest-impact journals” did not report fresh research results. The issue that caused the most trouble was that IF was a quantitative measure of a quality that cannot be quantified. Authorship Analysis of authorship describes the cooperation between authors and the high productive authors, of which the top five cited authors listed, produced more than 400 publications. Thus, they were called “productive authors” ( Zongyi, Dongying & Baifeng, 2016 ). Unfortunately, a small number of prolific authors appeared in the network map of co-cited authors, suggesting that not only the number of publications but also the quality of publications should be assessed in the evaluation of these authors. Demetriou AA, Nyberg SL, and Van de K were cited more than 100 times. Although none of them belong to the prolific authors, they played an important part in artificial liver research. Keywords and research fields Keyword analysis consists of three parts: (1) a rough estimation of the trend of research directions (2) analysis of the research hotspots and frontiers (3) detailed classification and analysis of research area. The increase of “bioartificial livers” and “tissue engineering” usually followed by the increase of “acute liver failure, ” because acute liver failure usually developed into serious liver failure, and artificial livers were considered to be the most effective treatment. The keyword of “bioartificial livers” with more than 400 citations decreased during the period from 2004 to 2017. It reached a peak (54 articles) at 2005, and most of them focused on the topic of hollow fiber ( Abu-Absi et al. , 2005 ; Lu et al. , 2005 ; Nguyen, Brotherton & Chau, 2005 ), bioartificial cell ( Aoki et al. , 2005 ; Cheng et al. , 2005 ; Gerlach, 2005 ; Monga et al. , 2005 ), and transplantation ( Garkavenko et al. , 2005 ; Liu & Chang, 2005 ). This finding suggested that many researchers showed great interest in bioartificial livers in 2005. They made every effort to culture human hepatocytes, rat hepatocytes, and stems cells to support liver functions and improve the device of bioreactors ( Park et al. , 2005 ). The keyword “bioartificial livers” covered the meaning of “hepatocyte, ” and thus the trend of “hepatocyte” was similar with “bioartificial livers. ” Tissue engineering studies on 3D technology ( Arai et al. , 2017 ), mathematical model ( Chapman et al. , 2017 ), liver microencapsulation technique ( Chapman et al. , 2017 ) and stems cells ( Kadota et al. , 2014 ) were mainly carried out in 2016 and 2017. According to T50 keywords of artificial livers, we extracted the top three research hotspots and listed them as follows: Bioartificial liver: A bioartificial liver system incorporated hepatocytes into a mechanical, albumin dialysis (AD)-based artificial liver support device to replace liver function ( Nicolas et al. , 2017 ). The two main elements of bioartificial livers were bioreactors and cell material. There was no definition of which kind of cell is the most ideal, and many researchers were still in search of the most ideal cells. Hepatocyte: Hepatocyte was known as the function unit of liver. It was ideal to investigate a kind of hepatocyte with regeneration ability and necessary functions for substitution of necrotic or dysfunctional hepatocytes. Transplantation: With the development of liver regenerative medicine, cell transplantation was becoming an increasingly popular topic, and many researchers showed great interest in it. The keywords bursts were considered as research frontiers over time. We extracted some keyword burst and divided them into three parts as follows: Blood and blood purification: Acute liver failure would be found variety of toxic substances in blood. Therefore, researches on artificial livers mainly focus on extracorporeal blood purification (hemodialysis, hemofiltration, hemodiafiltration, plasmapheresis, hemodsorption, cell-based therapy, etc. ) ( Thongboonkerd, 2010 ; Nie et al. , 2015 ). Plasma exchange and Prometheus: Since relatively simple detoxification devices use in 1999 (AD, molecular adsorbent recirculating system (MARS)) and 2003 (fractionated plasma separation, Prometheus), the treatment of liver failure had improved a lot ( Davenport et al. , 2015 ). However, prospective trials of extracorporeal support with AD, superflux dialyzers in series with absorption columns, and bioartificial devices containing hepatocytes have not demonstrated a significant survival advantage. Therefore, researchers found that before liver failure, plasma exchange improved cardiovascular stability, extending survival, and increase overall survival. Tissue engineering: “Tissue engineering” appeared in 2014 for the first time and continued until 2017. It is the best prediction that liver function could be restored with the technology of tissue engineering, and even now, many researcher are working hard in perfecting the technology of tissue engineering. The future is bright for patients with serious liver failure in receiving a “new” liver through tissue engineering. Among the four domains mentioned, the domain of “bioartificial livers” occupied the largest portion and the other artificial organ domains occupied the smallest portion. This suggested that bioartificial liver would still be a treatment strategy of great potential in liver failure and needed more researchers to promote its development. At the same time, the occupation of clinical domain suggested that clinical requirement was one of the strongest motivations for stimulating the development of artificial livers. The result of the research fields was similar to the four domains of keywords. There were some limitations in this study. Firstly, the data were only extracted from the WoS database. Secondly, it was difficult to make definite predictions on the development trend as well as the most popular topic of interest since researches on artificial liver was still developing slowly though with great potential. Despite its limitations, this study has provided the tendency and main context of researches in artificial livers, indicating that hepatocyte transplantation might be the frontier in artificial livers but now the liver support system could gain better clinical effect. Conclusions The number of publications on artificial livers fluctuated from 2004 to 2017, most of which are Chinese publications. After excluding several confounding factors (database, number of authors, countries, years since publication, and funding), the citations roughly estimated the quality of the included articles. The analysis of authorship and institutions also contributed to evaluating the quality of the included articles. In addition, through a detailed analysis which roughly assessed the research tendency, hotspots and frontiers of keywords, this study showed the progress and research trend of artificial livers. However, although it seems that the future of artificial livers seems brighter for hepatocyte transplantation, the systems of artificial livers now are more focusing on blood purification, plasma exchange, etc. Supplemental Information 10. 7717/peerj. 6178/supp-1 Supplemental Information 1 The relation between citation index and the number of citations respectively in WoS(A) and Scopus(B), and the citation index between WoS and Scopus(C). The relation between citation index and the number of citations respectively in WoS(A) and Scopus(B), and the citation index between WoS and Scopus(C). Click here for additional data file. 10. 7717/peerj. 6178/supp-2 Supplemental Information 2 The relation between the number of articles and the impact of factors in T18 journals (A), and between the number of citations and the impact of factors in T18 journals (B). The relation between the number of articles and the impact of factors in T18 journals (A), and between the number of citations and the impact of factors in T18 journals (B). Click here for additional data file. 10. 7717/peerj. 6178/supp-3 Supplemental Information 3 Top 50 keywords with the strongest citation bursts. Click here for additional data file. 10. 7717/peerj. 6178/supp-4 Supplemental Information 4 T100 most-cited articles ranked by the number of times cited. Click here for additional data file. 10. 7717/peerj. 6178/supp-5 Supplemental Information 5 Keywords ranked by the number of citations on artificial livers from 2004 to 2017. Click here for additional data file.

10. 7717/peerj. 6358

PeerJ

A gelatin/collagen/polycaprolactone scaffold for skin regeneration

Background A tissue-engineered skin substitute, based on gelatin (“G”), collagen (“C”), and poly(ε-caprolactone) (PCL; “P”), was developed. Method G/C/P biocomposites were fabricated by impregnation of lyophilized gelatin/collagen (GC) mats with PCL solutions, followed by solvent evaporation. Two different GC:PCL ratios (1:8 and 1:20) were used. Results Differential scanning calorimetry revealed that all G/C/P biocomposites had characteristic melting point of PCL at around 60 °C. Scanning electron microscopy showed that all biocomposites had similar fibrous structures. Good cytocompatibility was present in all G/C/P biocomposites when incubated with primary human epidermal keratinocytes (PHEK), human dermal fibroblasts (PHDF) and human adipose-derived stem cells (ASCs) in vitro. All G/C/P biocomposites exhibited similar cell growth and mechanical characteristics in comparison with C/P biocomposites. G/C/P biocomposites with a lower collagen content showed better cell proliferation than those with a higher collagen content in vitro. Due to reasonable mechanical strength and biocompatibility in vitro, G/C/P with a lower content of collagen and a higher content of PCL (GC L P H ) was selected for animal wound healing studies. According to our data, a significant promotion in wound healing and skin regeneration could be observed in GC L P H seeded with adipose-derived stem cells by Gomori’s trichrome staining. Conclusion This study may provide an effective and low-cost wound dressings to assist skin regeneration for clinical use.

Introduction Biocomposites, biocompatible and/or eco-friendly composites, can be formed by different varieties of natural and synthetic polymers including polysaccharides, proteins and biodegradable synthetic polymers. Generally, biocomposite materials have better structural properties than either constituent material alone. Therefore, biocomposite materials are often used in contact with living tissues such as scaffolds for cell-based therapy, biomedical implants and controlled drug delivery devices. Collagen based materials are the most well-known biocomposites in clinical use. Collagen-based wound dressings have been applied in the treatment of burn and ulcer patients over the last 30 years ( Doillon & Silver, 1986 ; Peters, 1980 ). Highly innovative tissue-engineered skin substitutes have been developed to mimic normal skin recently with melanocytes, a capillary-like network, sensory innervation and adipose tissue ( Bechetoille et al. , 2007 ; Regnier et al. , 1997 ; Tremblay et al. , 2005 ; Trottier et al. , 2008 ). By designing and incorporating specific therapeutic factors in skin substitutes, the promotion of wound healing as well as reduction of morbidity and mortality for large wounds may be achieved. The concept of design for the tissue engineered skin equivalent is based on existing models comprising a stratified epithelium grown on a matrix populated with dermal fibroblasts ( El-Ghalbzouri et al. , 2002 ). The biomaterials used for supporting skin cell growth include natural biodegradable polymers such as collagen and gelatin, as well as synthetic biodegradable polymers such as α-polyester and poly(ε-caprolactone) (PCL) ( Hajiali et al. , 2011 ; Ng, Khor & Hutmacher, 2004 ). For the demand of the durability, elasticity, cosmetic appearance of normal skin and wound repair, several functional dermal layers were developed. Integra ® artificial skin is a bi-layered structure consisting a dermal replacement layer covered with a silicone sheet as an outer layer. The dermal layer includes a porous fiber matrix with cross-linked bovine tendon collagen and glycosaminoglycan (chondroitin-6-sulfate) ( Burke, 1984 ; Burke, 1987 ; Burke et al. , 1981 ; Yannas & Burke, 1980 ). Comp Cult Skin ® is a type of composite cultured skin consisting of neonatal keratinocytes and fibroblasts cultured in bovine type-I collagen scaffold. Apligraf ® is a bi-layered living skin consisting an epidermal layer of neonatal keratinocytes and a dermal layer composed by neonatal fibroblasts in a type I bovine collagen gel ( Bell et al. , 1981 ; Trent & Kirsner, 1998 ; Wilkins et al. , 1994 ). The dermal components for most of these modified co-cultured skin constructs are mainly composed of collagen. However, pure collagen products are fragile and have difficulty in handling during clinical applications. To solve this problem, biodegradable polymers such as PCL were incorporated with collagen to increase the mechanical strength, which shows better resistance to external force for wound treatment. On the other hand, the well designed collagen scaffolds have been proven to cancel the skin wound contraction and prevent scar formation formed by contractile cells ( Dai et al. , 2004 ; Dai et al. , 2009 ; Powell, Supp & Boyce, 2008 ; Soller et al. , 2012 ; Yannas et al. , 1989 ). Meanwhile, animal studies also revealed that the wound size was effectively reduced by the use of collagen in vivo ( Dai et al. , 2004 ; Dai et al. , 2009 ; Powell, Supp & Boyce, 2008 ; Soller et al. , 2012 ; Yannas et al. , 1989 ). It may be associated with the fast growing epidermal layer of mice. In our previous study ( Dai et al. , 2004 ; Dai et al. , 2009 ), we developed collagen:PCL (C/P) biocomposites in two ratios (1:8 and 1:20). They both exhibited a similar porous structure that facilitated cell proliferation. Meanwhile, the smaller pore size may prevent the direct contact between keratinocytes and fibroblasts, and allow for cell interaction via signaling through existing pores. It is worth noting that collagen-based wound dressings were expensive. To solve this problem, we chose gelatin, the degradation of collagen, for the preparation of skin biocomposites. Therefore, the purpose of this study was to prepare a cost-effective, mechanically strong and biodegradable biocomposites based on gelatin, collagen, and PCL (G∕C∕P). Moreover, G∕C∕P biocomposites with a lower collagen content were designed and tested in this study because of good cell attachment and proliferation in our preliminary testing. Therefore, high and low GC mats with two different ratios of PCL were fabricated. The structure, thermal characteristics and biocompatibility were evaluated by in vitro testing. The potential in promoting large-sized wound healing was confirmed in the full-thickness skin defect model of nude mice. Materials and Method Preparation of G∕C∕P and C/P biocomposites The preparation of the G∕C∕P and C/P biocomposites was listed in Table 1. Aliquots of type B gelatin (from bovine skin, Sigma-Aldrich, St Louis, MO, USA) with specific concentration of collagen (Sigma-Aldrich, St Louis, MO, USA) solution was prepared by dissolving gelatin in double distilled water followed by mixing collagen (dissolved in acetic acid) to produce two ratios of collagen in gelatin/collagen biocomposites respectively. The dissolution of gelatin/collagen was facilitated by stirring using a heating magnetic stirrer in a 25 ml glass shell vial at the temperature of 40 °C. After complete dissolution, aliquots (0. 25 ml) of the gelatin/collagen solution were added to 7 ml glass vials followed by elimination of air bubbles and frozen at −20 °C for approximately 45–50 min. In the second stage, samples were transferred to a freezer at −72 °C for 35 min. The frozen samples were placed in a freeze dryer (Edwards Modulyo ® ) at −44 °C under 42 mbar vacuum for 24 h to prepare for the G/C mats. Aliquots (0. 5 ml) of PCL (Mw=115000; Solvay Interox, Warrington, UK) /dichloromethane (DCM with HPLC grade; Fisher Scientific, Loughborough, UK) solution were added carefully to the freeze dried G/C mats with low collagen content to produce GC L P L (G/C:PCL is 1:8) and GC L P H (G/C:PCL is 1:20) biocomposites respectively, whereas aliquots (0. 5ml) of PCL/dichloromethane (DCM) solution were added to the freeze dried gelatin/ collagen mats with high collagen content to produce GC H P L (G/C:PCL is 1:8) and GC H P H (G/C:PCL is 1:20) biocomposites respectively. The vials were kept for 30 min and then removed the lids to allow solvent evaporation overnight. 10. 7717/peerj. 6358/table-1 Table 1 The composition and preparation of four types of GCP biocomposites. Gelatin/ddH 2 O (200 µl) Collagen/1% HAc (50 µl) PCL/DCM (500 µl) C/G + C (%, w/w) G + C:P (w/w) GC L P L 2 mg/ml 1. 25 mg/ml 7. 4 mg/ml 14 1:8 GC L P H 2 mg/ml 1. 25 mg/ml 18. 5 mg/ml 14 1:20 GC H P L 0. 8 mg/ml 6. 05 mg/ml 7. 4 mg/ml 35 1:8 GC H P H 0. 8 mg/ml 6. 05 mg/ml 18. 5 mg/ml 35 1:20 C/P(1:8) 1. 85 mg/ml (250 µl) 7. 4 mg/ml 1:8 C/P(1:20) 1. 85 mg/ml (250 µl) 18. 5 mg/ml 1:20 Notes. G gelatin C L collagen with lower proportion C H collagen with higher proportion P L PCL with lower proportion P H PCL with higher proportion On the other hand, C/P biocomposite was fabricated described by previous study ( Hajiali et al. , 2011 ). Briefly, collagen solution was prepared in 1% acetic acid by stirring with a magnetic stirrer overnight at room temperature. Aliquots (0. 25 ml) of the collagen solution were added to 7 ml glass shell vials (Fisher Scientific) and frozen at –20 °C for approximately 45–50 min. Samples were transferred to a freezer at –70 °C for 35 min before freeze drying. Aliquots (0. 5 ml) of a solution of PCL in DCM were added carefully to the freeze dried collagen mats to prepare 1:8, 1:20 collagen:PCL materials, respectively. The vials were kept stopped for 30 min before removing the lids to allow solvent evaporation overnight. The component ratios of the four G∕C∕P and C/P biocomposites are shown in Table 1. Protein release from G∕C∕P biocomposites To determine the release of gelatin and collagen from biocomposites ( n = 3), the bicinchoninic acid (BCA) assay was used to estimate the amount of protein release from G∕C∕P biocomposites after incubation in phosphate-buffered saline (PBS) at 37 °C for 12 days. Individual samples were added to 7 ml glass shell vials containing 1 ml PBS and then incubated at 37 °C in a water bath. The release media was replaced completely by fresh PBS periodically and analysed for total protein content using the BCA assay. The absorbance of the calibration samples measured at 562 nm was used to produce a calibration curve used to calculate the protein (gelatin/collagen) concentration of the test samples. Thermal analysis by differential scanning calorimetry (DSC) The thermal characteristics of GC L P L, GC L P H, GC H P L, and GC H P H, biocomposites with weight between 2–10 mg were recorded using a Perkin-Elmer Pyris Diamond differential scanning calorimeter. All the samples were lightly pressed into the bottom of the pan to ensure good thermal contact. Sealed DSC pans were used in the study. Triplicate samples were heated at a rate of 10 °C/min from 10 °C to 100 °C. Peak melting temperature (Tm) and heat of fusion data for the PCL component of the materials were determined using the built-in software of the DSC. The latter measurement was subsequently used to estimate the percentage crystallinity of PCL in the composites from the reported heat of fusion of 139. 5 J/g for fully crystalline PCL ( Burke, 1987 ). Indium was used as a standard. The tensile strength analysis by a universal testing machine The mechanical analysis was conducted on GCP and C/P biocomposites using a universal testing machine (Instron) under axial loading. The biocomposite specimens were cut as the suitable size of rectangular block and were carefully clamped at the center of the cross-head with its end faces exactly perpendicular to the longitudinal axis. The crosshead speed of 50 mm/min was applied for this test. The tensile strength (MPa) was calculated as the force at failure divided by the cross-sectional area of the biocomposite. Results were the average from 3–6 measurements and analyzed by one-way analysis of variance and the t -test ( P < 0. 05). Observation of porous structure by scanning electron microscopy (SEM) The scanning electron microscope was used to observe the surface morphology of G∕C∕P biocomposites. Samples were attached to aluminum SEM stubs using carbon tabs (Agar Scientific). Specimens were sputter coated with gold prior to examination using a HITACHI ® S-3000N scanning electron microscope. Cell culture on G∕C∕P Primary human epidermal keratinocytes (PHEK) and primary human dermal fibroblasts (PHDF) were isolated and primarily cultured from the donated human foreskin samples after the surgery of circumcision in the study, which were approved by the institutional review board (IRB). The study protocol was reviewed and approved by the Institutional Review Board (IRB) in the Tri-Service General Hospital, R. O. C. (TSGHIRB No. : 100-05-251). Then the written informed consent was obtained from each donor. EpiLife ® HKM (Cat No. M-EPIcf-500 with addition of 0. 06 mM calcium chloride and HKGS kit: Cat No. S-001-5) was used for keratinocyte culture. The cell isolation from skin samples, cell expansion, and cell counting followed those described previously ( Burke, 1984 ). PHEK (child foreskin; P4) were seeded on the top surface of G∕C∕P (1:8 and 1:20) and C/P biocomposites and 24-well tissue culture plastics (TCP) at a cell density of 1. 7 × 10 5 or cells per cm 2 for up to 9 days, PHDF (adult foreskin; P3-4) were seeded on the same materials at a cell density of 2. 0 × 10 4 cells per cm 2 for up to 10 days. Human adipose-derived stem cells (ASCs; passages 3–4) were seeded on the biocomposites and 24-well TCP at a cell density of 2. 0 × 10 4 cells per cm 2 for up to 12 days. Trypsin-EDTA solution was used for cell detachment followed by cell counting at 1, 3, 6, and 9 days for PHEK and cell at 1, 4, 7, and 10 days for PHDF using a Weber’s haemocytometer. The experiments were repeated at least twice with similar results. The adhesion, growth, and distribution of cells seeded on G∕C∕P biocomposites were investigated using the immunohistochemical assay. PHEK (child foreskin; P4) were seeded on the top surface of G∕C∕P biocomposites (1:8 and 1:20), C/P biocomposites and 24-well tissue culture plastics (TCP) at a cell density of 1. 7 × 10 5 cells per cm 2 for time intervals up to 9 days. PHEK on G∕C∕P biocomposites (1:8 and 1:20) at 1 and 3 days were labeled with the (primary) monoclonal rabbit anti-human cytokeratin (CK)-14 1:100 (Bioworld Technology), and the (secondary) rhodamine-conjugated polyclonal goat anti-rabbit IgG 1:200 (Chemicon, Billerica, Mass) examined under a fluorescent microscope. PHDF (adult foreskin; P3-4) were seeded on the top surface of 1:8 and 1:20 GC10, GC25 and collagen:PCL biocomposites and TCP (24-well tissue culture plastics) at a cell density of 2. 0 × 10 4 cells per cm 2 for time intervals up to 10 days. PHDF on G∕C∕P biocomposites (1:8 and 1:20) at 1 and 3 days were labeled with the (primary) monoclonal mouse anti-human α-tubulin antibody 1:200 (Cat No. sc-5286; Santa Cruz Biotechnology, Inc. ), and the (secondary) fluorescein (FITC)-conjugated polyclonal goat anti-mouse IgG 1:100 (Lot No. 62686; Jackson ImmunoResearch Laboratories, Inc. , Baltimore, MD, USA) and examined under the fluorescent microscope. Preparation of human adipose-derived stem cells (ASCs)-seeded G∕C∕P biocomposites for animal experiments Human adipose tissues were obtained using lipoaspirate after syringe-assisted liposuction from abdomens of adult Taiwanese female patients who had no systemic metabolic diseases or lipid disorders. The procedures and protocols were conducted at the Tri-Service General Hospital, Taipei, Taiwan, and were approved by IRB. The isolated fat tissues were washed using PBS. After washing, the tissues were digested with an equal volume of PBS including 0. 075% type I collagenase (Sigma-Aldrich Company Ltd, Poole, UK) at 37 °C for 1 h. After centrifugation at 2, 500 rpm for 10 min, the cell pellet was resuspended in Dulbecco modified Eagle medium-low glucose (DMEM-LG; Invitrogen) containing 10% fetal bovine serum (FBS; Invitrogen) and 1% penicillin-streptomycin (Invitrogen), and placed in the incubator. After 1 day of culture, the dishes were washed with Hanks buffered salt solution (HBSS; Thermo Scientific, Agawam, MA, USA) to discard blood cells and nonadhesive cells and fresh culture medium was then added. The medium was refreshed every 3 days. When the cells reached 70% to 90% confluence, the cells were trypsinized (0. 25% trypsin; Sigma), neutralized with cultured medium and then passaged at a ratio of 1:3. Cells of passages 3 to 4 were used in the experiments. G∕C∕P biocomposites were cut to fit 12-well tissue culture plastics and then placed into the wells of TCP. For each well, human adipose-derived stem cells (ASCs) at a cell density of 10 6 cells per cm 2 were seeded onto the biocomposite with the addition of cultured medium. After incubation for 24 h, the ASC-seeded biocomposite was washed with HBSS to remove the culture medium and nonattached cells. Animal model and wound healing experiments A skin wound healing model of nude mice was employed ( Huang et al. , 2012 ). All mice were acquired, housed, and studied under a protocol approved by the Institutional Animal Care and Use Committee of National Defense Medical Center, Taiwan, R. O. C. The three eight-week-old nude mice (BALB/c-nu; BioLASCO) were chosen for each group including GC L P H scaffold only and GC L P H scaffold seeded with ASCs in this study. All of the surgical instruments were sterilized and the surgical procedures were performed under laminar flow. The surgical sites were sterilized with Easy Antiseptic Liquid 2% (Panion & BF, Taipei, Taiwan) before surgery. After anesthesia, a square of full-thickness cutaneous wound (12 mm ×12 mm) was made by surgery using scalpel on each dorsum of the hind thighs, followed by grabbing, pulling the region with a forceps, and excising of the full-thickness tissue with scissors. The wounds were divided into the following two groups: blank GC L P H biocomposite and ASCs-seeded GC L P H biocomposites. The GC L P H biocomposites (0. 1–0. 2 mm thick) for ASCs seeding were used. Both grafts were placed on each wound, sewn with 6 to 8 stitches using NC125L Nylon 5-0 surgical sutures (UNIK, Taipei, Taiwan), and covered with Tegaderm films (3M Health Care, MN) to prevent catching, biting, or wound infections. The wounds were continuously observed for a period of 3 weeks. The area of the wounds was blindly measured twice using ImageJ v1. 44p software ( http://imagej. nih. gov/ij ; NIH, Bethesda, MD, USA). After 2 weeks, the tissues of healed wounds were excised, fixed in 10% formalin for at least 24 h at room temperature, subsequently embedded in paraffin, and sectioned in 5-µm increments. The sections were stained with Gomori’s trichrome staining and examined by an optical microscope (Olympus BX50, Hamburg, Germany) ( Luna, 1992 ). The collagen component of the extracellular matrix deposited in skin substitutes was stained green. Statistical analysis Results were presented as the mean ± standard deviation of three replicates for each experiment. Statistical analysis was performed using Statistical Package for the Social Sciences, Version 12. 0 (SPSS Inc. , Chicago, Illinois). The statistically significant differences between groups were assessed by one-way ANOVA analysis of variance, followed by Tukey post-hoc test. P < 0. 05 was considered statistically significant. Results Protein release from the G∕C∕P The compositions and fabrication of G∕C∕P biocomposites are present in Table 1 and the cumulative release of protein from the biocomposites is shown in Fig. 1 and Table 2. For G∕C∕P with a lower collagen content, a higher cumulative protein release was present for GC L P L biocomposite in comparison with GC L P H group. At 6 h, the former showed about 78% average release, whereas the latter showed 52% average release. The protein was released quickly in 12 h for both biocomposites but slow down afterwards. The cumulative release (w/w) at 24 h was 100% and 76% for two biocomposites respectively. For G∕C∕P with a higher collagen content, a higher cumulative release of protein was found for GC H P H biocomposite relative to GC H P L biocomposite. The former exhibited a higher average release rate of about 62% average release at 6 h, compared to 52% average release of the latter. The average cumulative release (w/w) after 24 h was 81% and 62% respectively for GC H P H and GC H P L biocomposites. The cumulative release reached a plateau for both biocomposites after 24 h. PCL is the polymer with slow degradation rate and low water uptake and hence GC L P H showed better collagen and gelatin encapsulation due to high PCL content. Therefore, unencapsulated gelatin or collagen were released within 3 days and the remained gelatin or collagen would be released slowly depending on the PCL degradation rate. In addition, the amount of protein release significantly increased by gelatin amount in G∕C∕P biocomposites with low PCL content of polymer but no effect in the those with high PCL content of polymer. 10. 7717/peerj. 6358/fig-1 Figure 1 Cumulative protein release from G/C/P biocomposites in PBS at 37 °C. 10. 7717/peerj. 6358/table-2 Table 2 Cumulative protein release from G/C/P biocomposites in PBS at 37 °C. Gelatin (%) Collagen (%) PCL (%) 6 h (%) 12 h (%) 24 h (%) 72 h (%) 144 h (%) 216 h (%) 288 h (%) GC L P L 9. 61 1. 50 88. 89 78 ± 2 96 ± 3 100 ± 4 107 ± 4 107 ± 4 107 ± 4 108 ± 4 GC L P H 4. 12 0. 64 95. 24 52 ± 2 73 ± 1 76 ± 1 82 ± 3 82 ± 3 82 ± 3 83 ± 3 GC H P L 7. 99 3. 12 88. 89 52 ± 1 59 ± 2 62 ± 2 64 ± 2 66 ± 3 67 ± 4 71 ± 5 GC H P H 3. 42 1. 34 95. 24 62 ± 15 78 ± 16 81 ± 15 85 ± 15 86 ± 15 92 ± 18 94 ± 17 Notes. G gelatin C L collagen with lower proportion C H collagen with higher proportion P L PCL with lower proportion P H PCL with higher proportion h hours (time) Thermal and tensile analyses for G∕C∕P biocomposites The thermal properties of G∕C∕P biocomposites ( n = 3) are listed in Table 3. For G∕C∕P made with GC L mats, the melting point of the PCL component was close to the value of 60 °C normally found for pure PCL. The crystallinity of the PCL in these biocomposites (C/P and G∕C∕P) was estimated by comparing the fusion heat with that of 100% crystalline PCL as—139 J/g. From our previous DSC study ( Dai et al. , 2004 ), the 1:20 collagen:PCL biocomposite displayed the highest crystallinity of 63. 5 ± 3. 8%, followed by 48. 5 ± 4. 4% for 1:8 collagen:PCL biocomposites. However, the crystallinity of the PCL films prepared by solvent casting was only around 54. 2 ± 5. 4% in comparison with 1:20 w/w collagen:PCL biocomposites at 63. 5 ± 3. 8% ( Table 3 ). For G∕C∕P biocomposites, the average crystallinity of GC H P H was found to increase by 2. 9% relative to the solvent cast PCL film. Furthermore, the mean crystallinity of the PCL phase in the higher gelatin/collagen content (1:8 w/w) tended to be lower than that of 1:20 films. However, there was no statistically significance between these groups (Paired t -test: P = 0. 173). 10. 7717/peerj. 6358/table-3 Table 3 The analysis of G/C/P biocomposites for thermal characteristics ( n = 3, values are expressed as mean ± SD). Tm (°C) Crystallinity (%) Reference GC L P L 59. 2 ± 1. 5 53. 9 ± 1. 8 GC L P H 61. 5 ± 2. 0 54. 8 ± 2. 4 GC H P L 58. 3 ± 0. 2 55. 1 ± 1. 2 GC H P H 59. 8 ± 0. 3 57. 1 ± 1. 4 PCL (100%) film 62. 8 ± 0. 2 54. 2 ± 5. 4 Dai et al. (2004) C/P (1:8) 60. 7 ± 0. 2 48. 5 ± 4. 4 Dai et al. (2004) C/P (1:20) 63. 7 ± 0. 4 63. 5 ± 3. 8 Dai et al. (2004) The mechanical properties of G∕C∕P vs. C/P are listed in Table 4. The enhanced tensile strength and elongation and lower Young’s modulus were detected in GC L P H when compared with C/P biocomposites (1:20 mixture). 10. 7717/peerj. 6358/table-4 Table 4 The tensile properties of four GCP and C/P biocomposites ( n = 3; values are expressed as mean ± SD). Young’s modulus (MPa) Elongation (%) Tensile strength (MPa) GC L P H 35. 5 ± 8. 3 10. 7 + 1. 8 2. 5 ± 1 GC H P H 46. 6 ± 6. 2 6. 7 ± 1. 7 2. 2 ± 0. 4 GC L P L 30. 8 ± 9. 1 5. 1 ± 1. 5 0. 8 ± 0. 1 GC H P L 59. 5 ± 8. 3 7. 4 ± 1. 8 2. 2 ± 0. 5 C/P (1:20) 75. 8 ± 4. 2 9. 4 + 1. 3 3. 7 + 0. 4 C/P (1:8) 5. 7 + 3. 8 18. 1 + 6. 1 0. 5 ± 0. 1 Structure of G∕C∕P biocomposites The SEM images of G∕C∕P biocomposites are shown in Fig. 2. From these image, an irregular pore structure (20–100 µm) was apparent on the surface of G∕C∕P biocomposite. The biocomposites with a lower PCL content (GC L P L ) exhibited more porous structure. On the other hand, biocomposites with a higher collagen content (GC H P L and GC H P H ) showed less pore structure when compared to those with a lower collagen content (GC L P L and GC L P H ). The GC H P L biocomposite with highest content of collagen and lowest content of PCL showed a rough, fibrous surface due to the underlying collagen mat structure. The higher ratio of PCL was applied, the smoother surface was present because smooth overlying areas could be formed by the PCL phase. In addition, the surfaces of biocomposites with higher ratio of collagen seemed to have gel-like material filling the pores. All biocomposites exhibited the rough, fibrous structure of G/C mat overlaid by the smooth PCL phase. Little difference in morphology was observed. 10. 7717/peerj. 6358/fig-2 Figure 2 SEM images of GC L P L, GC L P H, GC H P L, and GC H P H biocomposites. Attachment and proliferation of PHEK, PHDF and ASC on G∕C∕P biocomposites The attachment and proliferation of PHEK, PHDF and ASCs on G∕C∕P are shown in Fig. 3. The number of PHEK on GC L P H and GC H P L was greater than that on the other biocomposites at 1 day. The cell density of PHEK was greater on C/P biocomposites (1:8) than on G∕C∕P biocomposites at 3 and 6 days ( P < 0. 05). PHEK almost had equal cell density on C/P and G∕C∕P biocomposites at 9 days ( P > 0. 05). Moreover, the GC L P L biocomposite appeared to have the greatest number of cells at 9 days. However, no significant differences were present among these groups. The number of PHDF on GC H P L was greater than on the other biocomposites after 1 day ( P < 0. 05). There was no significant difference in the cell number among all groups at 4, 7, and 10 days. Results of cell proliferation on different biocomposites showed that PHEK exhibited faster cell growth than PHDF at 1 day. Meanwhile, the numbers of ASCs on different biomaterials were also investigated. Cell proliferation on GC L P H was similar to the other groups in 10 days. Based on these results, PHEK, PHDF and ASCs well attached and proliferated on all G∕C∕P biocomposites. 10. 7717/peerj. 6358/fig-3 Figure 3 Attachment and proliferation of (A) Primary human epidermal keratinocyte (PHEK) and (B) primary human dermal fibroblasts (PHDF) (C) human adipose-derived stem cells (ASCs) grown on G/C/P and C/P biocomposites. The blank wells (tissue culture polystyrene, TCP) served as the control. ∗: indicates P < 0. 05 between TCP and other groups. The morphology of PHEK and PHDF is shown in Fig. 4. Keratinocytes with similar granular shape were observed on the biocomposites. A higher expression of cytokeratin-14, the structural protein of mature keratinocyte, was induced as the incubation time increased. On the other hand, α tubulin, a major component of fibroblast cytoskeleton, was stained and appeared as the spindle-like shape. As the time increased, more α tubulin was expressed in PHDFs. The fluorescent expressions of cytokeratin-14 and α tubulin were parallel to the numbers of attached and proliferated cells shown in Fig. 3. In addition, the α tubulin expression of fibroblasts and cytokeratin-14 expression of keratinocytes seeded on GC L P H and GC H P L were higher than those of other groups at three days. Therefore, the GC L P H biocomposite with the lowest collagen content in the 4 GCP scaffolds was selected as the skin substitute for the pilot animal study because gelatin was a cheaper material than collagen. 10. 7717/peerj. 6358/fig-4 Figure 4 Immunohistochemical assay for (A) PHEK grown on four G/C/P biocomposites at 1, 3 days and for (B) PHDF grown on four biocomposites at 1, 3 days, as the staining controls lacking primary antibody were present in the second set of columns. CK-14 was a specific protein for PHEK, whereas alpha tubulin was a specific protein for PHDF. w/o, without. (scale bar: 250 µm). In vivo animal study Rapid wound closure was observed in the both groups covered with GC L P H biocomposites in comparison with the control group, as shown in Fig. 5. In both groups covered with GC L P H biocomposites, the wound size dramatically decreased apparently was observed from 8 to 12 days and then gradually decreased until 21 days. The rate of wound healing was almost the same for both groups within 16 days. However, the wound healing based on wound area exhibited statistical difference at day 21 ( P < 0. 05). The incomplete wound closure also observed even up to 21 days for the group covered with GC L P H only but not ASC-seeded GC L P H. Furthermore, Gomori’s trichrome staining was performed in control, groups covered with GC L P H only and ASC-seeded GC L P H, as presented in Fig. 6. The open wound covered with GC L P H only exhibited loose collagen deposition even after 14 days ( Figs. 6B & 6E ). In the ASC-seeded GC L P H group, complete wound closure with differentiated epidermis and abundant dermal parallel-arranged fibrous collagen deposition was observed at 14 days ( Figs. 6C & 6F ). At a larger magnification, the ASC-seeded GC L P H group showed the largest thickness of epidermis, followed by the GC L P H only group, and then the open wound. The thick epidermis layer was obviously observed in the outside portion of the wound for the ASC-seeded GC L P H. 10. 7717/peerj. 6358/fig-5 Figure 5 Comparison of wound closure: (A) the appearance and (B) the area of wound in the full-thickness skin defect of nude mice covered with GC L P H scaffold and those seeded with ASCs until 21 days. The untreated wounds were served as the control. ∗: indicates P < 0. 05 between control and other groups. 10. 7717/peerj. 6358/fig-6 Figure 6 Collagen deposition in the wound bed by Gomori’s trichrome staining for the histology of cultured skin model. Collagen deposition in the wound bed by Gomori’s trichrome staining for the histology of cultured skin model in the groups of (A) open wound (B) GC L P H only (C) ASCs-seeded GC L P H (40×; scale bar: 30 µm) on the full-thickness skin defect of nude mice for 14 days. (D) to (F) were the magnified images of the box area (red rectangle) in (A) to (C) respectively (100×; scale bar: 5 µm). Discussion Based on the lower collagen content, we developed a cheaper skin substitute integrating the advantages of natural and synthetic biopolymers to promote the growth of skin cells in this study. Berillis has mentioned that collagen-based biopolymers are critical for tissue engineering and regenerative medicine due to its superior biocompatibility and low immunogenicity which is depended on the source of collagen ( Berillis, 2015 ). Collagen not only provides the building block for elastin and collagen fiber formation, but also acts as ligands for dermal fibroblasts to stimulate the production of new collagen, elastin and hyaluronic acid ( Sibilla et al. , 2015 ). In addition, collagen hydrolysate exhibits bioactivities including antihypertensive activity, lipid-lowering activity, as well as reparative properties in injured skin ( Fan, Zhuang & Li, 2013 ). Gelatin, which is derived from the partial hydrolysis of collagen, could be effective in promotion of granulation and epithelialization in the skin wound and also modulate drug delivery in targeted tissues ( Albuquerque-Jr et al. , 2009 ; Dantas et al. , 2011 ; Ikada & Tabata, 1998 ). Gelatin contains integrin binding sites for cell adhesion with the characteristics of low cost, natural abundance, biodegradability and biocompatibility ( Lakshminarayanan et al. , 2014 ; Ma et al. , 2005 ). Poly- ε-caprolactone (PCL), approved by the Food and Drug Administration (FDA), is a hydrophobic, biodegradable and biocompatible synthetic polyester and which are widely applied in drug delivery and tissue repair ( Spearman, Rivero & Abidi, 2014 ). Therefore, our approach could combine the characteristics including cell binding properties of natural polymers such as gelatin and collagen and the structural stability of PCL. These composite films are highly flexible which may allow effective draping over the wound site to improve the graft application. Tissue engineering scaffolds should have enough residence time for tissue development. PCL was chosen in this study because of its long resorption time, over 1 year ( Pitt et al. , 1981 ). Additionally, the stability of PCL reduces the possible toxicological effect associated with chemical crosslinking of natural polymers. Tear resistance is also conferred by PCL which facilitates the manipulation during surgery. We also expect the stability of PCL may inhibit wound contraction by fibroblasts for optimal dermal regeneration of full thickness wounds ( Werner et al. , 1994 ). However, PCL is known to be susceptible to acid and enzymatic degradation ( Leffler & Muller, 2000 ). Mochizuki et al. (1995) reported that certain lipases enhanced the degradation of polycaprolactone (PCL) when compared with incubation in buffer only. Based on DSC data, the degree of PCL crystallinity was greater for the biocomposites with a lower G/C content in which implicated that the gelatin/collagen phase could impede PCL crystallization. The notable reduction of PCL crystallinity in the biocomposites with GC H mat may be associated with the effect of GC H mat to interfere with the nucleation of PCL crystals. This phenomenon may also facilitate hydrolytic chain scission in this semi-crystalline polymer. In addition, the solvent evaporation may be conducive to develop high crystallinity due to the differences of crystal nucleation process between the pellet and the thin membranous form of PCL. Gelatin may not be able to develop a very strong architecture during the process of lyophilisation thereby exerts a minor effect on crystal nucleation of PCL phase. From DSC results, the similar GC content had almost the same percentage of crystallinity in PCL ( Table 1 ). Therefore, collagen content may play a critical role in the degree of the PCL crystal nucleation pattern. As the PCL content decreased, C/P biocomposites showed a more open, porous structure in a previous study ( Coombes et al. , 2002 ; Dai et al. , 2004 ). In previous study, the cell attachment and spreading on collagen:PCL composites found that 1:8 composites exhibited a greater masking or coating effect of the fibrous structure, whereas 1:20 composites showed a roughened, dimpled surface texture and separated pores. A similar trend in G∕C∕P was observed based on SEM images of this study. The pore number and size were slightly greater for the lower collagen content. Our previous study reported that the pore number and size seemed to increase after incubation in PBS, suggesting that protein may be released from the sub-superficial layers of the biocomposites ( Dai et al. , 2004 ). As shown in Fig. 1, GC H P L and GC L P H remained higher gelatin/collagen amount than other two G∕C∕P biocomposites after 24 h incubation and Fig. 3 also indicated that GC H P L and GC L P H had slightly better cell attachment than others. Therefore, cell attachment and proliferation may be more related the amount of remained gelatin or collagen rather surface morphology. The cost effect, quality of materials, and simple fabrication procedures are critical for the development of tissue-engineered skin model. Although collagen is more bioactive than gelatin, the price for collagen is much higher than that of gelatin. Therefore, the replacement of collagen by gelatin or other biomaterials may be cost-effective. However, pure gelatin was relatively fragile and weak during the lyophilization procedure, which made the fabrication difficult to control and caused one-third failure preparation rate in the early stage of this work. From Table 4, the C/P biocomposite also showed lower tensile strength, lower elongation, and higher Young’s modulus when compared with GC L P H biocomposites. Therefore, the mixture of gelatin and collagen as G/C mat was employed. The G∕C∕P biocomposites were found to possess proper biocompatibility. ASCs are widely used in tissue engineering and stem cell therapy because of their advantages in accessibility and immunosuppressive characteristics ( Gonzalez-Rey et al. , 2009 ). Effective would healing was improved by ASCs in literature ( Lam et al. , 2013 ; Lin et al. , 2013 ; Nauta et al. , 2013 ; Shokrgozar et al. , 2012 ). The adipose lineage cells could serve as a new cell source that promoted reepithelialization and angiogenesis in dermal wound healing from previous literature ( Ebrahimian et al. , 2009 ). However, stem cells directly applied on the wound site are not well survived ( Lam et al. , 2013 ), which reduces the efficacy of wound closure. An abundant amount of dermal collagen deposition was shown by histological staining. These findings suggested that G∕C∕P biocomposites are good carrier for ASCs. To verify the fate of adipose stem cells seeded, a tracing dye has to label the cells, which will be a future subject of study. In addition, G∕C∕P biocomposites may further be incorporated with growth factors to accelerate the wound repair ( Shin & Peterson, 2013 ). Finally, we hypothesize that the stronger mechanical strength of G∕C∕P may help resisting the contraction generated by fibroblasts and prevent possible scar formation. Therefore, the G∕C∕P biocomposites (in particular GC L P H ) with advantages in biocompatibility as well as mechanical properties are potential scaffolds for clinical wound treatment. Conclusions G∕C∕P biocomposites were fabricated and characterized as tissue engineered skin substitutes in this study. A similar fibrous/porous structure and thermal properties were present in G∕C∕P biocomposites. A reasonable mechanical strength and biocompatibility in vitro were achieved for the G∕C∕P biocomposites containing the lower collagen content (GC L P). Furthermore, the rapid closure of the skin wounds in nude mice were observed for those treated with GC L P H only or ASC-seeded GC L P H. Therefore, GC L P H biocomposites could be applied as low cost wound dressings based on high content of inexpensive gelatin and reasonable healing capacity. Supplemental Information 10. 7717/peerj. 6358/supp-1 Supplemental Information 1 Fig 1 Protein release. Click here for additional data file. 10. 7717/peerj. 6358/supp-2 Supplemental Information 2 Fig 3A PHEK cell growth. Click here for additional data file. 10. 7717/peerj. 6358/supp-3 Supplemental Information 3 Fig 3B PHDF cell growth. Click here for additional data file. 10. 7717/peerj. 6358/supp-4 Supplemental Information 4 Fig 3C ASCs cell growth. Click here for additional data file. 10. 7717/peerj. 6358/supp-5 Supplemental Information 5 Fig 4C Quantitative analysis of area percentage of fluorescent-positive cells. Click here for additional data file. 10. 7717/peerj. 6358/supp-6 Supplemental Information 6 Fig 5B Wound closure. Click here for additional data file. 10. 7717/peerj. 6358/supp-7 Table S1 The analysis of G/C/P biocomposites for thermal characteristics Click here for additional data file. 10. 7717/peerj. 6358/supp-8 Table S2 The tensile properties of four GCP and C/P biocomposites Click here for additional data file. 10. 7717/peerj. 6358/supp-9 Supplemental Information 7 Differential scanning calorimetry scan of four G/C/P biocomposites (Tm, melting temperature; Endo, endothermic; data H, enthalpy). Click here for additional data file.

10. 7717/peerj. 6589

PeerJ

Osteogenic capacity and cytotherapeutic potential of periodontal ligament cells for periodontal regeneration in vitro and in vivo

Background The periodontal ligament cells (PDLCs) contain heterogeneous cell populations and possess stem-cell-like properties. PDLCs have attracted considerable attention as an option for periodontal regeneration. However, the osteogenic differentiation of PDLCs remains obscure owing to variable osteo-inductive methods and whether PDLCs could be directly used for periodontal regeneration without stem cell enrichment is uncertain. The aim of the present study was to clarify the osteogenic differentiation capacity of PDLCs and test PDLCs as an alternative to stem cells for periodontal regeneration. Methods We tested the performance of human PDLCs in osteo-inductive culture and transplantation in vivo while taking human bone marrow derived mesenchymal stem cells (hMSCs) as positive control. Proliferation of PDLCs and hMSCs in osteo-inductive condition were examined by MTT assay and colony formation assay. The osteogenic differentiations of PDLCs and hMSCs were assessed by Alkaline phosphatase (ALP) activity measurement, von Kossa staining, Alizarin red S staining and quantitative RT-PCR of osteogenic marker gene including RUNX2, ALP, OCN, Col I, BSP, OPN. We transplanted osteo-inductive PDLCs and hMSCs with hydroxyapatite/tricalcium phosphate (HA/TCP) scaffolds to immunodeficient mice to explore their biological behaviors in vivo by histological staining and immunohistochemical evaluation. Results After 14 days of osteo-induction, PDLCs exhibited significantly higher proliferation rate but lower colony-forming ability comparing with hMSCs. PDLCs demonstrated lower ALP activity and generated fewer mineralized nodules than hMSCs. PDLCs showed overall up-regulated expression of RUNX2, ALP, OCN, Col I, BSP, OPN after osteo-induction. Col I level of PDLCs in osteo-inductive group was significantly higher while RUNX2, ALP, OCN were lower than that of hMSCs. Massive fiber bundles were produced linking or circling the scaffold while the bone-like structures were limited in the PDLCs-loaded HA/TCP samples. The fiber bundles displayed strong positive Col I, but weak OCN and OPN staining. The in vivo results were consistent with the in vitro data, which confirmed strong collagen forming ability and considerable osteogenic potential of PDLCs. Conclusion It is encouraging to find that PDLCs exhibit higher proliferation, stronger collagen fiber formation capacity, but lower osteogenic differentiation ability in comparison with hMSCs. This characteristic is essential for the successful periodontal reconstruction which is based on the synchronization of fiber formation and bone deposition. Moreover, PDLCs have advantages such as good accessibility, abundant source, vigorous proliferation and evident osteogenic differentiation capacity when triggered properly. They can independently form PDL-like structure in vivo without specific stem cell enrichment procedure. The application of PDLCs may offer a novel cytotherapeutic option for future clinical periodontal reconstruction.

Introduction Periodontal diseases are common and are the main reasons for teeth loss worldwide. The progression of periodontal disease results in destruction of tooth supporting tissues involving alveolar bone, periodontal ligament (PDL) and cementum. Retaining or improving PDL function is crucial for restoring periodontal defects. Many periodontal tissue engineering approaches have been proposed such as guided tissue regeneration (GTR), growth factor utilization like EMD, BMPs, FGF and cell transplantation, etc. ( Andrei et al. , 2018 ; Siaili, Chatzopoulou & Gillam, 2018 ). GTR and growth factor utilization are currently applied clinically but only result in partial regeneration which may be due to the unique periodontal anatomical structure and different alveolar bone defect morphology ( Ripamonti & Petit, 2009 ; Bottino et al. , 2012 ; Aimetti et al. , 2016 ; Soltani Dehnavi et al. , 2018 ). Therefore, suitable seed cells and efficient cytotherapeutic approaches are imperative for the regeneration of periodontium. Several mesenchymal-stem-like cell populations have been successfully derived from different dental tissues such as dental pulp stem cells, stem cells from exfoliated deciduous teeth, PDL stem cells, stem cells from apical papilla and dental follicle progenitor cells ( Rodríguez-Lozano et al. , 2011 ). These stem cells, as well as those from bone marrow ( Hasegawa et al. , 2006 ; Yu et al. , 2013 ) and adipose tissue ( Sugawara & Sato, 2014 ; Suzuki et al. , 2017 ) are good therapeutic candidates for periodontal regeneration. But their application is greatly limited by the source shortage, low harvesting efficiency and difficulty in isolation, culture and proliferation. The PDLCs consisted of heterogeneous cell populations including fibroblasts, endothelial cells, epithelial cell rests of Malassez, sensory cells (such as Rufini-type end organ receptors), osteoblasts and cementoblasts ( D’Errico et al. , 1999 ; Lekic et al. , 2001 ; Xiong, Gronthos & Bartold, 2013 ). PDLCs possess distinct functional characteristics and differ from cells in other connective tissues in a number of respects. Many studies showed that PDLCs had stem cell properties and performed osteogenic differentiation when triggered appropriately ( Seo et al. , 2004 ; Nagatomo et al. , 2006 ; Silvério et al. , 2010 ; Moshaverinia et al. , 2014a ). PDLCs have attracted considerable attention as an alternative to mesenchymal stem cells owing to their promising osteogenic differentiation ability, high proliferation rate, easy accessibility and abundant quantity ( Akizuki et al. , 2005 ; Iwata et al. , 2010 ; Zhou et al. , 2012 ). However, the osteogenic differentiation of PDLCs remains obscure owing to variable osteo-inductive methods and whether PDLCs could be directly used for periodontal regeneration without the enrichment of mesenchymal-stem-like cells is uncertain ( Yang et al. , 2013 ; Seubbuk et al. , 2017 ; Wei et al. , 2017 ; Proksch et al. , 2018 ). The aim of the present study was to clarify the osteogenic differentiation capacity of PDLCs comparing with human bone marrow derived mesenchymal stem cells (hMSCs) under easy-handling, practical and standardized osteo-inductive procedures in vitro and in vivo. We tested PDLCs as an alternative source for periodontal tissue regeneration by transplanting PDLCs-loaded hydroxyapatite/tricalcium phosphate (HA/TCP) scaffold into immunodeficient mice. We anticipated that this study may provide us with a deeper understanding of the biological properties and application potential of PDLCs for future periodontal tissue engineering and regeneration. Materials and Methods Cell culture and osteogenic differentiation Human PDLCs were isolated from healthy permanent premolars extracted for orthodontic reasons (10–14-year-old donors) in Department of stomatology, Beijing Friendship Hospital, Capital Medical University. Following surgical extraction, soft tissues close to root neck or apical zone were removed. Extracted tooth roots with the middle part of PDL were digested with an enzyme mixture (two mg/mL dispase, one mg/mL collagenase type I) at 37 °C for 1. 5 h. PDLCs pellets were then collected by centrifugation at 1, 000 rpm for 10 min. The pellets were re-suspended with dulbecco’s modified eagle media (DMEM) containing 10% fetal bovine serum and 100 U/mL penicillin and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA, USA), and cultured in a humidified atmosphere of 5% CO 2 at 37 °C. The study was approved by the Ethics Review Board of Beijing Friendship Hospital, Capital Medical University, Beijing, China (2017-P2-109-01). Human bone marrow derived mesenchymal stem cells were obtained from inpatients of Beijing Friendship Hospital, Capital Medical University who suffered alveolar cleft (11–20-year-old donors). Approximately 10 mL bone marrow of iliac crest was obtained and kept in 10 mL Hanks solution with 100 U/mL sodium heparin. The hMSCs-enriched low-density fraction was collected by centrifuging in lymphocyte separation medium at 1, 000 rpm for 10 min. The hMSCs were re-suspended with DMEM containing 10% fetal calf serum, 100 U/mL penicillin and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA, USA). The four to six passages of PDLCs and hMSCs were used for the following experiments. Osteogenic differentiation was performed by changing the culture medium to differentiation medium, comprising of DMEM supplemented with 100 nM Dexamethasone, 10 mM sodium β-glycerophosphate and 50 nM ascorbic acid-2-phosphate (Thermo Fisher, Waltham, MA, USA). Differentiation medium was replenished every 3 days during a differentiation period of 14 days. MTT assay Periodontal ligament cells and hMSCs were seeded in 96-well plates at a density of 1 × 10 3 cells/well and induced to osteogenic differentiation. Growth profiles of the cells were determined using MTT method, during a course of 8 days. At each time point, 20 μL of MTT solution (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide) was added to the cells in each well, incubated at 37 °C for 4 h, replaced the medium with 150 μL dimethyl sulfoxide (DMSO) (Sigma, St. Louis, MO, USA) at the end of incubation. Analysis were performed with a microplate reader (ELX800; BioTek Instruments Inc, Winooski, VT, USA) at 570 nm, to calculate the profiles of cell proliferation. Colony formation assay To assess the colony-forming capability of PDLCs and hMSCs, single cell suspensions were seeded into Φ6 cm dishes (Costar, Washington, D. C. , USA) at a density of 500 cells/dish. After induced osteogenic differentiation of 14-day, the dishes were washed with PBS for two times, fixed in ice-cold methanol for 10 min, and stained with crystal violet (Sigma, St. Louis, MO, USA). Photographs were taken with a phase-contrast microscope (Nikon, Tokyo, Japan) and quantitative analysis of colony formation was carried out using ImageJ ( Rasband, 2016 ). Alkaline phosphatase staining Periodontal ligament cells and hMSCs were seeded in 12-well plates (Costar, Washington, D. C. , USA) at a density of 5 × 10 4 cells/well. After osteogenic differentiation performed as described, Alkaline phosphatase (ALP) staining was carried out. In brief, culture wells were rinsed twice with PBS, fixed in 95% alcohol for 1 min, washed twice with distilled water, incubated with substrates mixture (2% sodium β-glycerophosphate 25 mL, 2% sodium barbiturate 25 mL, distilled water, 50 mL, 2% calcium chloride five mL, 2% magnesium sulfate two mL, several drops of chlorform) at 37 °C for 3 h, washed twice with distilled water, incubated in 2% cobalt nitrate for 2 min, and finally incubated with 2% amine sulfide for 1 min. Staining pictures were taken using a phase-contrast microscope (Nikon, Tokyo, Japan). Measurement of ALP activity Periodontal ligament cells and hMSCs were seeded in 24-well plates (Costar, Washington, D. C. , USA) at 1 × 10 4 cells/well. Osteogenic differentiation was performed as described. After a culture of 14 days, cells were collected and lysed in 0. 2% Triton X-100 lysis buffer. With cell lysates, ALP activity was measured using a StemTAG™ Alkaline Phosphatase Activity Assay Kit (CBA-301; Cell biolabs, San Diego, CA, USA). Protein concentration was determined using a Pierce BCA Protein Assay Kit (Thermo Fisher, Waltham, MA, USA), and was used to normalize ALP activity for each sample. Mineralization assay—Von kossa staining and Alizarin red S staining The PDLCs and hMSCs were seeded at 5 × 10 4 cells/well in 12-well plates (Costar, Washington, D. C. , USA) respectively. After 14 days of osteo-incubation, Von kossa staining was performed by means of a modified method ( Cheng et al. , 1994 ). In brief, culture wells were rinsed with PBS, fixed with 95% alcohol for 1 min, rinsed three times with deionized water, incubated in flesh 2% silver nitrate at 37 °C for 30 min, washed three times with deionized water, and exposed to UV-light for 30 min to present the results. For Alizarin Red S Staining, cells were fixed with cold methanol at room temperature for 10 min, rinsed twice with distilled deionized water, stained with 1% Alizarin Red S (pH4, Sigma, St. Louis, MO, USA) at room temperature for 30 min. Then, the excess dye was removed and the cells were rinsed with distilled deionized water until the background became clear. Osteogenic marker genes expression—Quantitative RT-PCR assay To analyze the osteogenic marker genes expression in PDLCs and hMSCs following the induction of osteogenic medium for 14 days, total RNAs were isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). cDNA was synthesized using Script M-MLV reagent (Invitrogen, Carlsbad, CA, USA). Briefly, 15 μL reverse transcription reactions containing one μg total RNA, two μL oligo-dT primer, two μL deoxynucleotide (dNTP) (10 mM each), four μL 5 × first-strand buffer, 0. 5 μL RNase inhibitor (40 U/μL), and one μL TIANScript M-MLV (200 U/μL). For reverse transcription, the mixture was incubated at 70 °C for 5 min and moved onto ice for 2 min and then incubated at 25 °C for 10 min, 42 °C for 50 min, and finally 95 °C for 5 min. Osteogenic markers of RUNX2 (runt-related transcription factor 2), ALP, OCN (osteocalcin), COL1 (collagen type I), BSP (bone sialoprotein), OPN (osteopontin) were examined by quantitative real-time polymerase chain (PCR) reaction. β-actin was used as an internal reference in all applications. Specific primers were synthesized by Sangon Biotech, China ( Table 1 ). The real-time PCR (TIANGEN, Beijing, China) was performed using a 25 μL PCR reaction with two μL RT product, 2. 5 μL 10 × HotMaster Taq buffer, one μL dNTP (2. 5 mM each), and 0. 2 μL HotMaster Taq polymerase (2. 5 U/μL). The reactions were carried out at 95 °C for 2 min, followed by 40 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s. Relative gene expression was calculated by the 2 −ΔΔCT method. 10. 7717/peerj. 6589/table-1 Table 1 Specific primers used for real-time polymerase chain reaction of osteogenic markers. Primer Abbreviation GenBank accession Amplicon Size (bp) Sequence (5′ ≥ 3′) Runt-related transcription factor 2 RUNX2 NM_001015051 101 Forward: TGGTTACTGTCATGGCGGGTA Reverse: TCTCAGATCGTTGAACCTTGCTA Alkaline phosphatase ALP NM_001632 81 Forward: GTGAACCGCAACTGGTACTC Reverse: GAGCTGCGTAGCGATGTCC Osteocalcin OCN NM_199173 112 Forward: CACTCCTCGCCCTATTGGC Reverse: CCCTCCTGCTTGGACACAAAG Collagen type I Col I NM_000088 140 Forward: GAGGGCCAAGACGAAGACATC Reverse: CAGATCACGTCATCGCACAAC Bone sialoprotein BSP NM_004967 106 Forward: CACTGGAGCCAATGCAGAAGA Reverse: TGGTGGGGTTGTAGGTTCAAA Osteopontin OPN NM_001251830 230 Forward: CTCCATTGACTCGAACGACTC Reverse: CAGGTCTGCGAAACTTCTTAGAT Beta actin β-actin NM_001101 250 Forward: CATGTACGTTGCTATCCAGGC Reverse: CTCCTTAATGTCACGCACGAT Transplantation Preparation for transplantation vehicles Hydroxyapatite/tricalcium phosphate ceramics scaffolds were incubated in serum-free DMEM medium for 2 days and then sterilized under 121 °C for 15 min. Osteogenic differentiation was performed with PDLCs and hMSCs for a period of 14 days, 2 × 10 5 cells were then loaded to a piece of HA/TCP of 30 mm 3 volume and transplanted to immunodeficient mice respectively. Surgical implantation procedures Immunodeficient male severe combined immunodeficient mice (SCID) mice of 5 week-old ( n = 6) were used for transplant recipients of PDLCs or hMSCs. Operation was performed under local anesthesia made by lidocaine hydrochloride injection. Skin incision was made on the dorsal surface of each mouse, subcutaneous pockets were made in which cell-loaded HA/TCP scaffolds were placed. Up to four transplants were inoculated per animal. The experimental protocol was approved by the Institutional Animal Care and Use Committee of Beijing Friendship Hospital, Capital Medical University, Beijing, China (18-2004). Histological and immunohistochemical assay Mice were sacrificed 12 weeks after transplantation. Tissue samples were surgical isolated, fixed in 4% formalin for 24 h, decalcified in 14% EDTA solution (pH 7. 0) for 7–10 days, and embedded in paraffin. After tissue sectioning (four μm in thickness), slides were deparaffinized, hydrated and stained with hematoxylin and eosin (HE) using standard techniques. For immunohistochemical staining, slides were heated in a 60 °C oven for 30 min, and subsequently hydrated to water through a series of decreasing concentrations of ethanol. Then immunohistochemical staining was performed using a Immunohistochemical kit SP-9001 (Zhongshan Biotech Co. , Zhongshan, China). The used anti-OCN, anti-COL1, anti-OPN antibodies were purchased from Cell Signaling Technology (CST, Danvers, MA, USA). Results were observed and documented using an Olympus BX60 microscope. Statistical analysis of data Statistic analyses were performed with Student’s t -test, n ≥ 3. Data are presented as mean ± standard deviation. Differences between groups are considered statistically significant if P < 0. 05. Results Morphology and proliferation of PDLCs and hMSCs in osteo-inductive condition Periodontal ligament cells presented fibroblast-like morphology while after 14 days of osteo-inductive culture ( Fig. 1A ) while about half percentage of hMSCs exhibited relatively flattened broad shape ( Fig. 1B ). PDLCs demonstrated significantly higher proliferation rate but lower colony-forming ability comparing with hMSCs in osteo-inductive condition ( Figs. 1C – 1E ; P < 0. 01). The higher self-renewal efficiency of PDLCs provided a potential superiority in periodontal tissue regeneration. 10. 7717/peerj. 6589/fig-1 Figure 1 Morphology and proliferation of PDLCs and hMSCs in osteo-inductive condition. (A) PDLCs’ morphology after 14 days of osteo-inductive culture. (B) hMSCs’ morphology after 14 days of osteo-inductive culture. (C) Proliferation curve of PDLCs and hMSCs under osteo-inductive culture. (D) Colony formation assay of PDLCs and hMSCs after 14 days of osteo-inductive culture. (E) Quantitative analysis of colony formation. Scale bars, 100 μm. (Student’s t -test, n ≥ 3; * P < 0. 05, ** P < 0. 01). Osteogenic differentiation in vitro Periodontal ligament cells and hMSCs were exposed to osteogenic medium for 14 days and then osteogenic differentiations were assessed by measuring ALP activity, calcium deposition and osteogenic marker gene expression. Alkaline phosphatase activity Under normal culture condition, both PDLCs and hMSCs showed very weak ALP staining and low ALP activity. While in osteo-inductive medium, PDLCs and hMSCs both demonstrated strong ALP-positive staining and revealed higher ALP activity compared to their control, respectively ( P < 0. 05). No significant difference was found between PDLCs and hMSCs control group ( P = 0. 45), but the ALP activity of hMSCs was higher than that of PDLCs in osteo-inductive groups ( Figs. 2A and 2B ; P < 0. 05). 10. 7717/peerj. 6589/fig-2 Figure 2 Osteogenic differentiation of PDLCs and hMSCs after 14 days of osteo-inductive incubation in vitro. (A) ALP staining. The PDLCs and hMSCs kept in normal growth medium were used as control, while the PDLCs and hMSCs kept in osteo-inductive medium were assessed as test groups. (B) ALP activity assay. (C) von Kossa staining of PDLCs and hMSCs after 14 days of osteo-inductive incubation. (D) Quantitative analysis of von Kossa staining. (E) Alizarin red S staining of PDLCs and hMSCs after 14 days of osteo-inductive incubation. (F) Quantitative analysis of Alizarin red S staining. Scale bars, 100 μm. (Student’s t -test, n ≥ 3; * P < 0. 05, ** P < 0. 01). Calcium deposition Calcium deposition were further assessed by von Kossa staining and Alizarin red S staining after 14 days of osteogenic culture. Osteo-induction produced mineralized nodules in both PDLCs and hMSCs ( Figs. 2C – 2F ). PDLCs yielded smaller and fewer mineralized nodules comparing with hMSCs under osteo-inductive condition. Quantitative analysis also convinced the relatively inferior calcified capacity of PDLCs compared to hMSCs ( P < 0. 01). Osteogenic marker genes expression Quantitative RT-PCR showed that the expression levels of osteoblast-specific genes, such as RUNX2, ALP, OCN, Col I, BSP, OPN, were generally higher in both PDLCs and hMSCs after 14 days of osteo-inductive medium incubation (Osteo group) than those in normal medium (Control group) ( Fig. 3 ; P < 0. 05). No significant differences were found in control groups of PDLCs and hMSCs when they were cultured in normal medium ( P > 0. 05) except hMSCs exhibited a slightly increased expression of RUNX2 ( Fig. 3A ; P < 0. 05). When cultured in Osteo-inductive medium, PDLCs showed a promising osteogenic ability by the overall up-regulated expression of RUNX2, ALP, OCN, Col I, BSP, OPN ( P < 0. 05). Comparing with hMSCs counterparts, PDLCs (Osteo group) demonstrated a little bit lower expression of RUNX2, ALP, OCN ( Fig. 3A – 3C ; P < 0. 05) but almost the same level of BSP, OPN ( Figs. 3E and 3F ; P > 0. 05). This confirmed that PDLCs may differentiate toward osteoblastic direction under proper inductive condition. Furthermore, Col I level of PDLCs (Osteo group) was significantly higher than that of hMSCs which indicated PDLCs possessed better collagen forming capacity ( Fig. 3D ; P < 0. 01). 10. 7717/peerj. 6589/fig-3 Figure 3 Osteogenic marker genes expression in PDLCs and hMSCs after 14 days of osteo-inductive incubation in vitro. The PDLCs and hMSCs kept in normal growth medium were used as control, while the PDLCs and hMSCs kept in osteo-inductive medium were assessed as test groups. The expression of RUNX2(A), ALP(B), OCN(C), Col I(D), BSP(E), OPN(F) were analyzed by quantitative real-time PCR. (Student’s t -test, n ≥ 3; * P < 0. 05, ** P < 0. 01). Osteogenic differentiation in vivo To explore the application possibility of PDLCs and hMSCs in periodontal regeneration, we triggered the osteogenic differentiation of PDLCs and hMSCs by culturing them in osteo-inductive medium for 14 days in vitro. And then we transplanted the osteo-inductive cells with HA/TCP scaffolds to SCID mice is to examine their biological behaviors in vivo. A total of 12 weeks later, the transplants were taken out and put through histological evaluation by HE staining ( Fig. 4 ) and Col I, OCN and OPN immunohistochemical staining ( Fig. 5 ). There was no evident bone formed or fiber cluster appeared in cell-free HA/TCP control ( Fig. 4A ). Ectopic bone formation was found in the 12-week hMSCs-loaded HA/TCP, while no fiber bundles were formed ( Fig. 4B ). On the contrary, massive fiber bundles were produced linking or circling the scaffold while the bone-like structures were limited in the PDLCs-loaded HA/TCP samples ( Figs. 4C and 4D ). The fiber bundles displayed strong positive Col I, but weak OCN and OPN staining ( Figs. 5B – 5D ). The ectopic bone structure in hMSCs-loaded HA/TCP samples exhibited positive staining of Col I, OCN and OPN ( Figs. 5F – 5H ). The in vivo results were consistent with the in vitro data which confirmed strong collagen forming ability and considerable osteogenic potential of PDLCs. 10. 7717/peerj. 6589/fig-4 Figure 4 Transplantation of osteo-inductive PDLCs and hMSCs with HA/TCP scaffolds to SCID mice. After 12 weeks, the transplants were taken out and performed HE staining. (A) Cell-free HA/TCP control. No evident bone formed or fiber cluster appeared. (B) In hMSCs-loaded HA/TCP samples, ectopic bone formation was found, but no obvious fiber bundles appeared. (C–D) In the PDLCs-loaded HA/TCP samples, massive fiber bundles linked or circled the scaffold (C, arrow) were found while the bone-like structure was limited (D, arrows). Scale bars, 100 μm. 10. 7717/peerj. 6589/fig-5 Figure 5 Osteogenic differentiation of PDLCs and hMSCs in vivo. The expression of Col I, OCN and OPN in the transplants were evaluated by immunohistochemical staining. (A–D) In the PDLCs-loaded HA/TCP samples, the fiber bundles displayed strong positive Col I(B), but weak OCN(C) and OPN(D) staining. (E–H) In the hMSCs-loaded HA/TCP samples, the ectopic bone structure exhibited positive staining of Col I(F), OCN(G) and OPN(H). This result confirmed strong collagen forming ability and considerable osteogenic potential of PDLCs in vivo. Scale bars, 100 μm. Discussion Periodontal diseases are mainly inflammatory diseases and eventually lead to the destruction of periodontal tissues and teeth loss. Aside from the initial periodontal therapy such as scaling, root planning and periodontal surgical treatment to remove the infectious tissue, the ultimate therapeutic goal is to restore the lost fibrous PDL attachment and form new bone and cementum ( Estrela et al. , 2011 ; Han et al. , 2014a ). For this purpose, several periodontal tissue engineering approaches have been proposed, including genetic engineering, growth factor utilization, and cell transplantation. An ideal cell transplantation approach requires adequate source cells which are accepted by the recipient immune system ( Maeda et al. , 2011 ; Fu et al. , 2014 ; Liu et al. , 2015 ). Considering that periodontal tissue consists of gingiva, alveolar bone, cementum and PDL, it is better to choose periodontal-derived cells and the heterogeneous resource cell population may be favorable for periodontal regeneration ( Akizuki et al. , 2005 ; Iwata et al. , 2010 ). To repair periodontal defects, the regeneration of the PDL is as important as reparation of the bone. Focus should be put on both bone regeneration and functionally oriented PDL. To this aspect, PDLCs may provide an ideal resource cells due to their abundance and heterogeneous components ( Huang, Gronthos & Shi, 2009 ; Hynes et al. , 2012 ; Iwasaki et al. , 2014 ). But the effectiveness is unclear and the successful rate is unpredictable. Our objective was trying to illuminate the regenerating capacity of PDLCs and optimize the procedures. The PDLCs are heterogeneous mixture in which the predominant cell type is fibroblasts ( McCulloch & Bordin, 1991 ). The PDL fibroblasts originate from the ectomesenchyme of the investing layer of the dental papilla and dental follicle. Whether these fibroblast subsets are derived from one single type of progenitor cell is unknown but they exhibit different functional characteristics ( McCulloch, 1995 ; Lekic & McCulloch, 1996 ). We found that PDLCs exhibited considerably higher proliferation rate but lower colony-forming ability comparing with hMSCs in osteo-inductive condition. This may be partially due to the major percentage of fibroblasts in PDLCs mixture. The higher self-renewal efficiency of PDLCs provided a potential superiority in periodontal tissue regeneration. Moreover, PDLCs are relatively abundant in extracted teeth and can be non-invasively obtained comparing with hMSCs. However, the regenerative properties of PDLCs are affected by donor age, disease condition and tissue quality. PDLCs derived from young donors revealed greater cementum- and PDL-like tissue formation than those from aged donors. The self-renewal capacity of PDLCs decreased while donor aging ( Gao et al. , 2013 ; Tomokiyo et al. , 2018 ). Inflammation is another factor to affect the regenerative potential of PDLCs. PDLCs obtained from the patients with periodontitis showed significantly lower bone formation than the cells acquired from the patients with healthy PDL tissues ( Tang et al. , 2016 ). Therefore, how to obtain large numbers of PDLCs and maintain good self-renewal capacity is significantly important to achieve PDL regenerative therapy especially in elderly patients since prevalence and severity of periodontal disease are increasing with age. More efforts should be devoted to this issue for future clinical applications. In addition, different cell harvesting methods lead to different cell components which eventually result in different tissue engineering outcomes ( Seo et al. , 2004 ; Proksch et al. , 2018 ). We obtained the PDLCs from the extracted roots digested for 1. 5 h at 37 °C with enzyme mixture. Under this condition, we might harvest more fibroblasts from the PDL than cementoblasts which attach to the root surface or osteoblasts which anchor to the alveolar bone. After transplantation, these isolated PDLCs got a typical PDL-like structure in vivo. Because of the heterogeneity of PDLCs, there is contradictory evidence about whether PDLCs have the capacity to differentiate into osteoblast or form hard tissue following osteogenic induction. Many researchers reported that PDLCs had the ability to differentiate into osteoblast-like cells when triggered appropriately ( Fuchigami et al. , 2016 ; Seubbuk et al. , 2017 ; Furue et al. , 2017 ). However, the osteo-inductive methods were highly variable and certain experimental perturbations existed which might impair the validity of the results ( Hu et al. , 2014 ; An et al. , 2015 ; Seubbuk et al. , 2017 ; Wei et al. , 2017 ). Therefore, in this study, we uniformed the osteo-inductive condition for both PDLCs and hMSCs in order to make the results more clear and definite. We found that PDLCs expressed significantly elevated ALP activity, calcium deposition and osteogenic marker gene expression (RUNX2, ALP, OCN, Col I, BSP, OPN) after 14 days of osteo-induction. This confirmed that PDLCs actually had the osteogenic differentiation ability under proper inductive condition, although which was a little bit inferior to hMSCs since the mineralized nodules in vitro and bone-like structure in vivo of PDLCs samples were fewer than that of hMSCs. A wide variety of animal models (dog, rat, pig and sheep) and defect types were created to explore the successful periodontal regeneration with use of PDLCs periodontal ligament stem cells and scaffolds ( Han et al. , 2014b ; Bright et al. , 2015 ; Tassi et al. , 2017 ). Owing to the wide variability in defect type, cell source and cell scaffold, no meta-analysis or evident conclusion was possible ( Bright et al. , 2015 ). Due to the different surgical procedures to establish animal models for periodontal defects, it was also hard to tell whether the bone/cementum formation owes to the transplanted PDLCs differentiation or just repairation from the residue PDL structure. We performed xenotransplantation of human PDLCs-loaded scaffold to immunodeficient mice subcutaneously to clarify the pure self-rebuild function of PDLCs and excluded the interference from adjacent tissue cells. Evident bone formation was found in the 12-week hMSCs-loaded HA/TCP transplants, while no fiber bundles were formed. In contrast, massive fiber bundles were produced linking or circling the scaffold while the bone-like structure was limited in the PDLCs-loaded HA/TCP samples. The fiber bundles displayed strong positive Col I and weak OCN, OPN staining. Our in vitro and in vivo data showed that PDLCs demonstrated superior collagen forming ability comparing to hMSCs. The abundant collagen type I-positive oriented fiber bundles formed by PDLCs were preferable for the future reconstruction of PDL structure. This was consistent with the studies conducted by other scholars in which the favorable fiber-forming ability of PDLCs has already been successfully applied in the repair of tendon injuries ( Moshaverinia et al. , 2014b ; Hsieh et al. , 2016 ). Besides the inherent variant properties of PDLCs and hMSCs, another reason accounting for the different differentiation capacity may be the local regeneration niche ( Dangaria et al. , 2009 ). We even transplanted PDLCs into the periodontal defect animal models, we could not simulate the micro-environment exactly the same as that of the human periodontium. The scaffold architecture should be optimized to fulfill proper spacious distribution and physiological force loading function. This requires further and intensive research. In summary, the successful reconstruction of periodontium is based on the synchronization of fiber forming and bone depositing which enables the anchorage of PDL into cementum and alveolar bone. Abundant source, good proliferation, combination of powerful fiber forming capacity and considerable osteogenic differentiation ability provide PDLCs as a novel candidate for periodontal regeneration. Conclusions Compared with hMSCs, PDLCs are easier to acquire and provide an abundant source for cell-based periodontal regeneration. PDLCs exhibit vigorous proliferation, evident osteogenic differentiation capacity and strong fiber-forming ability when triggered properly. They can independently form PDL-like structures in vivo without specific stem cell enrichment procedure. The application of PDLCs may offer a novel cytotherapeutic option for future clinical periodontal reconstruction. Supplemental Information 10. 7717/peerj. 6589/supp-1 Supplemental Information 1 Raw data. Click here for additional data file.

10. 7717/peerj. 6986

PeerJ

Enhanced mechanical, thermal and biocompatible nature of dual component electrospun nanocomposite for bone tissue engineering

Traditionally, in the Asian continent, oils are a widely accepted choice for alleviating bone-related disorders. The design of scaffolds resembling the extracellular matrix (ECM) is of great significance in bone tissue engineering. In this study, a multicomponent polyurethane (PU), canola oil (CO) and neem oil (NO) scaffold was developed using the electrospinning technique. The fabricated nanofibers were subjected to various physicochemical and biological testing to validate its suitability for bone tissue engineering. Morphological analysis of the multicomponent scaffold showed a reduction in fiber diameter (PU/CO—853 ± 141. 27 nm and PU/CO/NO—633 ± 137. 54 nm) compared to PU (890 ± 116. 911 nm). The existence of CO and NO in PU matrix was confirmed by an infrared spectrum (IR) with the formation of hydrogen bond. PU/CO displayed a mean contact angle of 108. 7° ± 0. 58 while the PU/CO/NO exhibited hydrophilic nature with an angle of 62. 33° ± 2. 52. The developed multicomponent also exhibited higher thermal stability and increased mechanical strength compared to the pristine PU. Atomic force microscopy (AFM) analysis depicted lower surface roughness for the nanocomposites (PU/CO—389 nm and PU/CO/NO—323 nm) than the pristine PU (576 nm). Blood compatibility investigation displayed the anticoagulant nature of the composites. Cytocompatibility studies revealed the non-toxic nature of the developed composites with human fibroblast cells (HDF) cells. The newly developed porous PU nanocomposite scaffold comprising CO and NO may serve as a potential candidate for bone tissue engineering.

Introduction The commonly occurring bone trauma diseases were tissue destruction or bone fracture ( Qi et al. , 2018 ). The autograft or allografts are the traditional methods commonly used to treat bone defects. However, they possess certain disadvantages such as donor scarcity, supply limitation, immunogenicity risk and infections which makes their usage limited in bone tissue repairing ( Wang & Yeung, 2017 ). Recently, the development of tissue engineered bone substitutes was widely used in biomedical applications because it overcomes the drawbacks associated with the above traditional methods. Tissue engineering (TE) comprises of three components, namely cells, growth factors and scaffolds ( Pereira et al. , 2015 ). Among these, the scaffolds play a vital role in supporting cellular responses and tissue regeneration ( Bružauskaite et al. , 2016 ). The scaffold used in the bone tissue engineering should resemble the native extracellular matrix (ECM) structure, and it is still one of the major challenges in the tissue engineering applications ( Venugopal et al. , 2007 ). The 3D structure of the natural ECM mainly contains fibrous proteins having a diameter in nano-metric range ( Heydarkhan-Hagvall et al. , 2008 ). Recently, the use of electrospinning is widely reported in the tissue engineering applications which produces the nano-scale matrices. The scaffolds produced using electrospinning technique contains nanofibers which can mimic the native function of the ECM ( Wang, Ding & Li, 2013 ). Electrospinning is a cost-effective and versatile technique which uses high voltage to fabricate the nanofibers from polymeric solution. In electrospinning, there are many parameters are involved like flow rate, the viscosity of the polymer, high voltage and collector distance which may affect the quality of fibers produced ( Subbiah et al. , 2005 ). The nanofibers fabricated using electrospinning technique possess high surface area with interconnected pores makes them an attractive choice for biomedical applications ( Huang et al. , 2003 ). In this research, the nanofibers were fabricated using Tecoflex EG-80A (medical grade polyurethane). Properties like biodegradability, good barrier properties and better oxidation stability of polyurethane make them as an undisputed choice in tissue engineering applications ( Lamba, Woodhouse & Cooper, 1998 ; Ma et al. , 2011 ). In this work, a polyurethane nanocomposite comprising canola oil (CO) and neem oil (NO) was fabricated to enhance its biological properties. Traditionally, the use of oils to alleviate the bone-related pain is widely practiced. Generally, the essential oils derived from plants possess anti-inflammatory, antiseptic and antispasmodic properties ( Ali et al. , 2015 ). These properties would play a vital role in reducing muscle and joint pain. CO is a vegetable oil derived from the seed of Brassica napus which is used as the cooking oil for making food varieties ( Dupont et al. , 1989 ). It is considered good for people’s health owing to low saturated and plenty of polyunsaturated fats ( Lin et al. , 2013 ). The chemical constituents present in the CO were tocopherols, phytosterols, polyphenols and carotenoids. The active constituents present in the CO were reported to improve oxidative stability and also possess antimicrobial and antioxidant property ( Ghazani & Marangoni, 2013 ). The second candidate employed in this work is NO. Neem or the Margosa trees is generally called as Azadirachta indica and the active constituents present in the NO were known for its antimicrobial and numerous medicinal properties. NO contains various active constituents such as nimbin, nimbolide, nimbidin and limonoids ( Alzohairy, 2016 ; Asif, 2012 ). Further, every part of the neem trees was reported to very useful in the medical field against various human diseases owing to the wide range of its pharmacological attributes ( Kumar & Navaratnam, 2013 ). The blood compatibility evaluation of the fabricated materials is one of the main factors which decides their role in clinical applications. The developed materials should reduce thrombus formation and less toxic to red blood cells (RBC’s) ( Jaganathan & Prasath, 2018 ). In this study, a novel electrospun bone scaffold based on polyurethane (PU) mixed with CO and NO was fabricated. For the first time, we have investigated the combined effect of these oils in bone tissue engineering. This works aims on the determination of physico-chemical properties, blood compatibility and cyto-compatibility of the newly developed scaffolds to analyze their potential for bone tissue engineering. Materials and Methodology Materials Tecoflex EG-80A (medical grade polyurethane) was supplied from Lubrizol and dissolved in dimethylformamide (DMF) solvent (Sigma Aldrich, Dorset, UK). NO (Thurgas Trading, Pulau Pinang, Malaysia) and CO (Naturel, Johor, Malaysia) were procured from the local market. The chemicals such as sodium chloride physiological saline (0. 9% w/v) and phosphate buffered saline (PBS) were obtained from Sigma-Aldrich, Kuala Lumpur, Malaysia. Blood compatibility reagents as calcium chloride (0. 025 M), rabbit brain activated cephaloplastin and thromboplastin (Factor III) were procured from Diagnostic Enterprises, Solan, India. Preparation of solution PU pellets in DMF was prepared at a concentration of 9 wt% and stirred for 12 h to obtain a homogeneous solution. Similarly, the NO and CO in DMF was prepared at a concentration of 9 wt% and stirred for 1 h maximum. The PU solution with 9 wt% was mixed with 9 wt% CO at a ratio of 7:2 (v/v%) and the PU (9 wt%) with CO and NO (9 wt%) solution was mixed at a ratio of 7:1:1 (v/v%) respectively. The mixture is stirred for 2 h maximum to attain even dissolution. Electrospinning process The electrospinning process was utilized to fabricate the nanofibers. To fabricate the nanofibers, the obtained homogeneous solution is electrospun at conditions of applied voltage 10. 5 kV, flow rate of 0. 5 ml/h and collector distance 20 cm. The nanofibers collected on the aluminum foil were removed and vacuum dried to eliminate the remaining residual content present in the electrospun membranes. Figure 1 represents the schematic image involving the materials, preparation of solution and electrospinning process. 10. 7717/peerj. 6986/fig-1 Figure 1 Schematic image involving the materials, preparation of solution and electrospinning process. Images generated by Saravana Kumar Jaganathan. Physio-chemical characterizations Scanning electron microscopy The SEM apparatus (Hitachi Tabletop scanning electron microscopy) was used to investigate the morphologies of the electrospun nanofibrous membrane. The samples were gold plated at 20 mA for 2 min before performing SEM analysis. The analysis was performed at a voltage of 10 kV and the microphotographs were captured at different magnifications. Finally, fiber diameter and its distributions were calculated using Image J by choosing 30 locations randomly from the captured image. IR analysis The chemical groups in the electrospun membranes were investigated using Fourier-transform infrared (FTIR) analysis through Thermo Nicolet, Waltham, MA. To begin the experiment, a small piece of sample was inspected and the spectra were measured between a wavelength of 600 and 4, 000 cm −1 with an average of 32 scans per minute at 4 cm −1 resolution. Contact angle measurements The water contact angle measurement determines the surface wettability which was tested using VCA (video contact angle) unit. Initially, the water droplet was placed on the electrospun membrane and the static image of the droplet was captured through a high-resolution video camera. The computer integrated software was used to measure the angle between the water droplet and the surface. Mechanical properties test The uniaxial load test machine was used to determine the mechanical properties of the electrospun nanofibrous membrane according to ASTM D882-10. All the samples were prepared with a size of 40 mm * 15 mm and gripped at the clamp ends of the tensile machine. The testing was performed with a strain rate of 5 mm/min and load cell of 500 N. The machine record data predict the tensile stress–strain curves and the tensile strength and elongation was calculated from the curve. The gauge length used was 20 mm. The thickness of the electrospun PU, PU/CO and PU/CO/NO were 0. 193 mm, 0. 086 mm and 0. 043 mm while their cross section areas were 2. 895 mm 2, 1. 29 mm 2 and 0. 645 mm 2 respectively. TGA analysis The thermal properties of the electrospun nanofibers were investigated by the PerkinElmer TGA 4000 unit. A sample weight of about 3 mg was placed on the aluminum and the testing was performed under dry nitrogen atmosphere. A heating rate of 10 °C/min with a temperature range of 30–1, 000 °C was applied to the specimen and the data obtained were explored into the excel sheet to draw the TGA curves and derivative weight loss curves. Surface measurement analysis Quantitative surface roughness analysis of the electrospun nanofibrous membrane was investigated by atomic force microscopy (AFM) unit (Nanowizard; JPK Instruments). To begin, the samples are fixed on a specimen holder and the scanning is performed at room temperature in a normal atmosphere. The scanning was performed in a 20 µm × 20 µm area by tapping mode and the 3D image was obtained with 256 * 256 pixels using JPKSPM data processing software. Coagulation assay Activated partial thromboplastin time (APTT) and Prothrombin time (PT) assay To evaluate the antithrombogenicity of the developed nanofibrous mats, the coagulation assays APTT and PT were utilized. Activated partial thromboplastin time is an intrinsic pathway to measure the time taken for the activation of the blood clot. To start the assay, the fabricated electrospun membranes were incubated with PBS at 37 °C for 30 min. After incubation, the samples were added with 50 µl of platelet-poor plasma (PPP), followed by adding 50 µL of the APTT agent (rabbit brain activated cephaloplastin) and incubated at 37 °C for 3 min respectively. Further, 50 µL of calcium chloride (CaCl 2 ) (0. 025 M) solution was added to the above mixture which initiates the blood and APTT were measured. Similarly, the prothrombin assay was also done to measure the time taken for blood clot through the extrinsic pathway. For the PT assay, the steps involved were similar to the APTT assay. For PT assay, the samples were added with 50 µl of PPP, followed by the adding 50 µL of the PT agent (thromboplastin (Factor III)) respectively. To initiate the blood clot, the mixture was stirred using the needle and the PT was measured ( Manikandan et al. , 2017 ). Hemolysis assay Hemolysis assay was carried out to analyze the toxicity of fabricated membranes against red blood cells (RBCs). Initially, the mixture of citrated blood and diluted saline was done at a ratio of 4, 5 (v/v%). Next, the sample with a size of (1 × 1 cm 2 ) was cut and was soaked in physiological saline at 37 °C for 30 min. After 30 min, the samples added to the prepared mixture of citrated blood and diluted saline for 1 h at 37 °C. Then, the samples were detached and centrifuged at 3, 000 rpm for 15 min. Finally, the optical density (OD) was measured from the aspirated supernatant at 542 nm to evaluate the hemoglobin release. The percentage of hemolysis or hemolytic index was calculated as described earlier ( Manikandan et al. , 2017 ). Cytocompatibility studies Human Dermal Fibroblast (HDF) cells were cultured in the medium of Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and incubated with 5% carbon dioxide (CO 2 ) at 37 °C. For every 3 days, the medium was changed. Prior to the cell seeding, the prepared electrospun scaffolds were cut into the small size and placed in 96-well plates. The electrospun membrane placed in the culture plates were sterilized with 75% alcohol solution for 3 h and washed with PBS. Then, HDF cells with a density of 10 × 10 3 cells/cm 2 were placed on each electrospun membrane and fixed in CO 2 incubator. After 3 days culturing, the cellular electrospun mats were added with 20% of 3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H tetrazolium, inner salt (MTS) and incubated for 4 h. After 4 h, the culture plates were retrieved and OD was measured at 490 nm using spectrophotometric plate reader to determine the viability rates. Statistical analysis One way ANOVA followed by Dunnett post hoc test was performed for experiments results with 3 trails to determine statistical significance. The analyzed data were expressed as mean ± SD. A representative of three images was presented for qualitative experiments. Results SEM investigation Figures 2A – 2C indicate SEM images of PU and their nanocomposites fabricated with CO and NO. SEM images depicted that the fibers of electrospun PU, PU/CO and PU/CO/NO membranes were randomly oriented and rendered beadles morphology. The fiber diameters of PU, PU/CO scaffold, PU/CO/NO scaffold were estimated as 890 ± 116. 911 nm, 853 ± 141. 27 nm and 633 ± 137. 54 nm and their fiber distribution curve as shown in Figs. 3A – 3C. 10. 7717/peerj. 6986/fig-2 Figure 2 SEM photographs of (A) polyurethane, (B) polyurethane/CO composites, and (C) polyurethane/CO/NO composites. Sample with size of 1 cm * 1 cm was cut and imaged in Hitachi Tabletop TM3000. 10. 7717/peerj. 6986/fig-3 Figure 3 Fiber diameter distribution of (A) polyurethane, (B) polyurethane/CO composites, and (C) polyurethane/CO/NO composites. Measured using Image J software by choosing 50 locations randomly. IR analysis Attenuated total reflectance-Fourier-transform infrared (ATR-FTIR) analysis was carried out to identify the characteristic peaks present in the fabricated PU, PU/CO and PU/CO/NO membranes as presented in Fig. 4. The spectra of PU membrane include 3, 323 cm −1 (NH stretching), 2, 939 cm −1 and 2, 854 cm −1 (CH stretching), 1, 730 cm −1 and 1, 703 cm −1 (carbonyl stretching), 1, 597 cm −1 and 1, 531 cm −1 (vibrations of NH stretching), 1, 413 cm −1 (vibrations of CH 2 stretching) and 1, 221 cm −1 and 1, 104 cm −1 (CO stretch corresponding to alcohol group) ( Manikandan et al. , 2017 ). In the spectra of PU/CO and PU/CO/NO scaffolds, the peaks were similar; however, decrease in peak intensity was found with the addition of canola and neem oil. Further, the electrospun nanocomposites exhibited a peak shift of CH stretch in pure PU from 2, 939 cm −1 to 2, 930 cm −1 in PU/CO and 2, 929 cm −1 in PU/CO/NO scaffold. 10. 7717/peerj. 6986/fig-4 Figure 4 IR spectrum of polyurethane, polyurethane/CO composites, and polyurethane/CO/NO composites. Sample with size of 1 cm * 1 cm was cut and measured in wavelength range between 600 and 4, 000 cm −1 in Nicolet iS5. Wettability measurements The results of water wettability for the developed PU, PU/CO and PU/CO/NO scaffolds were listed in Table 1. The results of water wettability test indicated that PU scaffold exhibited the contact angle of 100° ± 0. 58 and for electrospun PU/CO scaffold, it was increased to 108. 7° ± 0. 58 and reduced to 62. 33° ± 2. 52 in PU/ CO/NO scaffold. Thermal analysis The results of the thermal analysis for developed PU, PU/CO and PU/CO/NO scaffolds were presented in Figs. 5A – 5C. TGA demonstrated the initial temperature decomposition of PU membrane starts at 276 °C, while for the developed PU/CO and PU/ CO/NO scaffold, the initial decomposition temperature was 296 °C and 306 °C respectively. At 1, 000 °C, the residue weight percentage of PU was found to be only 0. 47%, whereas for the developed PU/CO and PU/ CO/NO scaffold were observed to be 2. 23% and 1. 22% respectively indicating the interaction of PU with CO and NO. Further, the DTG curve of PU/CO and PU/CO/NO scaffold was shown in Figs. 6A – 6C. In PU, there was three weight loss namely two major loss and one minor loss. The first and second major loss occurs at 223 °C to 348 °C and 348 °C to 446 °C, while the third minor loss was seen at 557 °C to 684 °C respectively. In the case of PU/CO scaffold, it shows two losses only, the first weight loss at 234 °C to 377 °C and final loss at 377 °C to 545 °C. While in the PU/CO/NO scaffold, it showed three weight loss in which starts from 223 °C to 378 °C (first loss), 378 °C to 529 °C (second loss) and 529 °C to 697 °C (third loss) respectively. 10. 7717/peerj. 6986/table-1 Table 1 Contact angle Measurement of PU, PU/CO and PU/CO/NO composites. Sample with size of 1 cm * 1 cm was cut and a water droplet of size 0. 5 µl was placed on it. The static image of the droplet was captured through a high-resolution video camera in VCA equipment and the manual contact angle was determined. S. No Sample Average contact angle in degrees 1 Pure Polyurethane 100° ± 0. 5774 2 Polyurethane/CO composites 108. 7° ± 0. 5774 * 3 Polyurethane/CO/NO composites 62. 33° ± 2. 517 * Notes. * mean differences were significant compared with pure PU ( p < 0. 05). 10. 7717/peerj. 6986/fig-5 Figure 5 Thermal behavior of polyurethane, polyurethane/CO composites and polyurethane/CO/NO composites. Sample weighing 3 mg was heated between temperature range of 30–1, 000 °C under nitrogen atmosphere in Perkin-Elmer TGA unit. 10. 7717/peerj. 6986/fig-6 Figure 6 DTG curve of polyurethane, polyurethane/CO composites and polyurethane/CO/NO composites. Sample weighing 3 mg was heated between temperature range of 30–1, 000 °C under nitrogen atmosphere in Perkin-Elmer TGA unit. Surface measurements The measurement of surface roughness for developed PU, PU/CO and PU/CO/NO scaffold were shown in Figs. 7A – 7C. The average roughness of PU membrane was found to be 576 nm, while the electrospun PU/CO/NO and PU/CO/NO scaffold exhibited roughness of 389 nm and 323 nm respectively. 10. 7717/peerj. 6986/fig-7 Figure 7 Surface measurements of polyurethane, polyurethane/CO composites, polyurethane/CO/NO composites. Sample with size of 1 cm * 1 cm was cut and scanned in 20 * 20 µm size with 256 * 256 pixels under normal atmosphere in Nanowizard, JPK instruments. Tensile analysis The stress strain curves of the developed PU, PU/CO and PU/CO/NO scaffold were presented in Figs. 8A – 8C. The PU nanofibers exhibited the tensile strength of 7. 373 ± 0. 4917 MPa with elastic modulus of 7. 033 ± 2. 129 MPa. The tensile strength of electrospun PU/CO and PU/CO/NO scaffolds were found to be 8. 370 ± 0. 2433 MPa and 9. 747 ± 0. 4061 MPa with elastic modulus of 6. 267 ± 0. 1155 MPa and 8. 497 ± 1. 479 MPa respectively. 10. 7717/peerj. 6986/fig-8 Figure 8 Tensile strength of polyurethane, polyurethane/CO composites and polyurethane/CO/NO composites. Sample with size of 4 cm * 1. 5 cm was stretched at a cross head speed of 5 mm/min with a 500 N load cell in Gotech Testing Machines, AI-3000. Coagulation results The APTT and PT of the developed PU, PU/CO and PU/CO/NO scaffold were measured to analyze their clot formation through intrinsic and extrinsic pathway as indicated in Figs. 9 and 10. In APTT assay, the blood clotting of PU membrane was found to 152. 7 ± 3. 055 s and for electrospun PU/CO and PU/CO/NO scaffold, the blood clotting was observed to be 173. 3 ± 3. 215 s and 154. 7 ± 4. 163 s respectively. Similarly, in PT assay, the blood clotting time of PU membrane was found to 88. 67 ± 2. 517 s and for electrospun PU/CO and PU/CO/NO scaffold, it was observed to 96 ± 3 s and 94 ± 3. 606 s respectively. Further, hemolysis assay was done to evaluate the osmotic stress of the developed membranes on the RBCs. For the PU membrane, the hemolytic percentage was found as 2. 48% and for the PU/CO and PU/CO/NO scaffold, it was 1. 33% and 1. 24%, respectively, as indicated in Fig. 11. 10. 7717/peerj. 6986/fig-9 Figure 9 APTT assay of polyurethane, polyurethane/CO composites, and polyurethane/CO/NO composites (mean differences were significant compared with pure PU ( p < 0. 05)). Sample with size of 1 cm * 1 cm was added with 50 µL of platelet-poor plasma (PPP) followed by incubating with 50 µL of reagent (rabbit brain cephaloplastin) and 50 µL CaCl 2 (0. 025 M) to calculate the blood clotting time. 10. 7717/peerj. 6986/fig-10 Figure 10 PT assay of polyurethane, polyurethane/CO composites, and polyurethane/CO/NO composites (mean differences were significant compared with pure PU ( p < 0. 05)). Sample with size of 1 cm * 1 cm was added with 50 µL of platelet-poor plasma (PPP) followed by incubating with 50 µL of thromboplastin (Factor III) to calculate the blood clotting time. 10. 7717/peerj. 6986/fig-11 Figure 11 Hemolytic assay of polyurethane, polyurethane/CO composites, and polyurethane/CO/NO composites (mean differences were significant compared with pure PU ( p < 0. 05)). Samples with size of 1 cm * 1 cm added to the mixture of citrated blood and diluted saline (4:5 v/v%) for 1 h at 37 °C. After this, the samples were centrifuged and optical density (OD) was measured at 542 nm. Cell viability analysis The nanofibers toxicity was assessed through HDF cells and evaluated after 3 days of cell culturing. Figure 12 indicates the MTS assay results of electrospun PU, PU/canola oil and PU/canola/neem oil scaffold. After 3 days of culture, the cell viability of PU/canola oil and PU/canola/neem oil scaffold were reported to 226. 3 ± 29. 94% and 248. 7 ± 23. 97% respectively while the pristine PU membrane exhibited viability rate of 179. 7 ± 15. 04%. 10. 7717/peerj. 6986/fig-12 Figure 12 MTS assay of polyurethane, polyurethane/CO composites, and polyurethane/CO/NO composites (mean differences were significant compared with pure PU ( p < 0. 05)). Samples with size of 0. 5 cm * 0. 5 cm was cut and placed in the 96 well plates. The scaffold were seeded with fibroblast cells with 10 × 10 3 cells/cm 2 density and cultured for 3 days. After 3 days, the medium was added with 20% of MTS reagent for 4 h and optical density (OD) was measured at 490 nm. Discussion The performance of the scaffolds used in tissue regeneration mainly depends on certain features like microstructure and host cell responses. Many materials used in tissue engineering possess desirable characteristics but are lacking in blood compatibility or bioactivity. Electrospun PU scaffolds were reported to mimic the ECM matrix but their utilization in tissue regeneration was limited due to poor bioactivity. In this work, CO and NO were added to the electrospun PU fibers to enhance its physico-chemical and cell response. The electrospun PU/CO and PU/CO/NO nanocomposite displayed a reduction in fiber diameter which might be due to the bioactive constituents present in the CO and NO oil. The bioactive constituents may have a putative role in altering the solution parameters which would have resulted in the reduced fiber diameter. A similar effect was observed in a recent study ( Chao, Mani & Jaganathan, 2018 ). In their work, electrospun PU scaffold added with grape seed oil, honey and propolis for bone tissue engineering. It was observed that the fabricated composites showed a reduction in the fiber diameter than the pristine PU due to the bioactive constituents from the additives. Our developed nanocomposites showed smaller fiber diameter than the pristine PU suggesting its suitability for bone tissue growth. In IR spectrum, the intensity of PU was decreased with the incorporation of canola and neem oil. This was owing to the hydrogen bond formation ( Unnithan et al. , 2012 ). Moreover, the prepared nanocomposites showed CH peak shift which concludes the presence of CO and NO content in the PU matrix ( Tijing et al. , 2012 ). The wettability analysis revealed that the addition of CO renders the scaffold to be hydrophobic whereas the addition of NO improved the wettability nature. The contact angle which is optimum for adhesion of osteoblast cell is reported to be 0–106° ( Wei et al. , 2009 ). Hence, the reported wettability of PU/CO scaffold might reduce the osteoblast cell adhesion. However, the addition of NO into PU/CO scaffold renders hydrophilic nature which lies in the reported wettability which might facilitate the adhesion of osteoblast cells. A scaffold with hydrophilic nature may be suitable for bone tissue engineering applications as depicted in some recent researches. Hassan, Sultana & Hamdan (2014) fabricated poly (ε-caprolactone) scaffold fabricated with hydroxyapatite nanofibers for bone tissue engineering. It was observed that the addition of hydroxyapatite into PCL matrix improve the wettability and suggested hydrophilic surface might enhance the osteoblast growth. To conclude, the ambivalent nature of the developed scaffolds may find potential application in tailoring its wettability to the need. In TGA analysis, the thermal behavior of the PU membrane was enhanced by the addition of CO and NO. Manikandan et al. (2017) developed nanocomposite utilizing polyurethane incorporated with murivennai oil and was reported that the incorporated murivennai oil enhanced the thermal behavior of the pure polyurethane and our findings resembled their observations. The results of DTG depicted addition of CO and NO reduced the weight loss of the PU. The reduced weight loss was observed the decrease in intensity of the weight loss peak of the prepared nanocomposites compared to the pristine PU. The surface measurements suggested the developed nanocomposites exhibited smoother surfaces than the pristine PU. Kim et al. (2016) electrospun poly (ε-caprolactone) scaffolds and investigated the effect of fiber diameter on surface roughness. It was reported that the scaffolds with smaller fiber diameter show the smoother surfaces compared to the larger fiber diameter. Hence, the decrease in the surface roughness of the fabricated composite was might be due to their smaller fiber diameter. Ribeiro et al. (2015) studied the effect of surface roughness on osteoblast cell response in poly (L-lactide) electrospun membranes. It was observed that the electrospun membranes with lower surface roughness showed enhanced osteoblast cell proliferation compared to the higher roughness surfaces. Hence, the reduced surface roughness values of the electrospun nanocomposites might favor the enhanced osteoblast cell adhesion and proliferation. Mechanical results showed that the encapsulation of CO and NO improved the tensile strength of the pristine PU. Few works of literature have been reported that the smaller fiber diameter would result in the improvement of the mechanical strength ( Mani et al. , 2019 ; Sheikh et al. , 2010 ). The addition of CO and NO oil resulted in the reduction of the fiber diameter which might have favored the enhancement of the tensile strength. Shanmugavel et al. (2014) fabricated a bone scaffold based on polycaprolactone incorporated with aloe vera and silk fibroin. They reported the tensile strength was enhanced by the adding of aloe vera and silk fibroin into the PCL matrix. Further, the observed tensile strength of the prepared nanocomposite was found to be 4 MPa and concluded a suitable candidate for bone tissue engineering. Our fabricated electrospun membranes exhibited better tensile strength than those reported values indicating the superiority of the fabricated scaffolds for bone tissue engineering. The blood compatibility assessments revealed the prolonged blood clotting time of the developed nanocomposites. The prolonged blood clotting time was because of the addition of CO and NO into the PU matrix. An important requirement for a scaffold in tissue engineering applications is their ability to retard the thrombus formation. Initially, when CO is introduced in the pristine PU, the surface turns to hydrophobic which facilitate the adhesion of plasma protein irreversibly resulting in prolonged blood clotting times. However, when there is an addition of NO, it introduces surface hydrophilicity in the nanocomposite resulting in blood clotting times reduction than the developed PU/CO scaffolds. This may be due to the trade-off between the polar and apolar regions of the PU/CO/NO scaffold ( Szycher, 1991 ). Strikingly, still the blood compatibility of PU/CO/NO scaffold is still better than the control indicating its suitability in bone tissue engineering application. Recently, Jaganathan et al. (2017) fabricated polyurethane scaffold mixed with castor oil nanofibers and investigated the blood compatibility behavior of the fabricated nanocomposites. They reported that the blood compatibility of the pristine PU was increased with the incorporation of castor oil which resembles our findings. Polyphenols compounds are reported to be found in the castor oil ( Chakravartula & Guttarla, 2007 ). These constituents appear to be one of the components present in CO ( Ghazani & Marangoni, 2013 ) and NO ( Nahak & Sahu, 2011 ). Moreover, the fabricated PU/CO and PU/CO/NO scaffold showed less toxic to RBC. The developed nanocomposites were reported to be non-hemolytic materials because their calculated hemolytic index was below 2% ( Manikandan et al. , 2017 ). In the MTS assay, it was observed that the cell viability of the electrospun membranes was enhanced compared to the control plates. The cellular viability of the fabricated composites was better than pure PU. The cell adhesion and proliferation is a multifactorial process and various properties influence it. It have been reported that the cell adhesion is influenced by properties such as fiber diameter ( Unnithan et al. , 2012 ), wettability ( Miguel et al. , 2017 ), surface roughness ( Chou et al. , 2009 ), surface chemistry ( Anselme, Ploux & Ponche, 2010 ) and surface energy ( Hallab et al. , 2001 ). Hence, any of these factors would have positively influenced the cell adhesion of the fabricated composites. The addition of NO to the PU/CO scaffold slightly increased the cell adhesion and proliferation. This is because NO rendered the surface to be hydrophilic (62. 33°). This value was found to be in the optimal wettability range (40–70°) during which the adhesion and proliferation of fibroblast cells reported to be maximum ( Miguel et al. , 2017 ). Conclusion In this work, PU nanocomposites based on CO and NO were electrospun successfully. The fabricated PU nanocomposites showed randomly oriented nanofibers with reduced fiber diameter compared to control. The existence of CO and NO in PU matrix was confirmed by the formation of hydrogen bond. The developed PU/CO scaffold rendered hydrophobic while the PU/CO/NO scaffold exhibited hydrophilic behavior and both scaffolds exhibited enhanced thermal stability spotted than the pristine PU. The surface roughness of PU/CO and PU/CO/NO scaffold were found to reduce compared to pure PU. Further, the developed nanocomposites showed enhanced blood compatibility behavior than the pristine PU. Moreover, the newly electrospun scaffolds were observed to non-toxic against RBC and HDF cells than the pristine PU as noted in the hemolytic assay and MTS assay. Finally, this research suggests that the novel developed PU nanocomposite comprising CO and NO with better physio-chemical characteristics, improves blood compatibility behavior and non-toxic nature may be beneficial for repairing the bone defects caused by trauma. Supplemental Information 10. 7717/peerj. 6986/supp-1 Dataset S1 Raw data for AFM analysis Click here for additional data file. 10. 7717/peerj. 6986/supp-2 Dataset S2 Raw data for FTIR analysis Click here for additional data file. 10. 7717/peerj. 6986/supp-3 Dataset S3 Raw data for tensile analysis Click here for additional data file. 10. 7717/peerj. 6986/supp-4 Dataset S4 Raw data for TGA analysis Click here for additional data file. 10. 7717/peerj. 6986/supp-5 Dataset S5 Raw data for morphology analysis Click here for additional data file.

10. 7717/peerj. 6993

PeerJ

Effect of GARP on osteogenic differentiation of bone marrow mesenchymal stem cells via the regulation of TGFβ1

Mesenchymal stem cells (MSCs), which have multipotential differentiation and self-renewal potential, are possible cells for tissue engineering. Transforming growth factor β1 (TGFβ1) can be produced by MSCs in an inactive form, and the activation of TGFβ1 functions as an important regulator of osteogenic differentiation in MSCs. Recently, studies showed that Glycoprotein A repetitions predominant (GARP) participated in the activation of latent TGFβ1, but the interaction between GARP and TGFβ1 is still undefined. In our study, we successfully isolated the MSCs from bone marrow of rats, and showed that GARP was detected in bone mesenchymal stem cells (BMSCs). During the osteogenic differentiation of BMSCs, GARP expression was increased over time. To elucidate the interaction between GARP and TGFβ1, we downregulated GARP expression in BMSCs to examine the level of active TGFβ1. We then verified that the downregulation of GARP decreased the secretion of active TGFβ1. Furthermore, osteogenic differentiation experiments, alkaline phosphatase (ALP) activity analyses and Alizarin Red S staining experiments were performed to evaluate the osteogenic capacity. After the downregulation of GARP, ALP activity and Alizarin Red S staining significantly declined and the osteogenic indicators, ALP, Runx2, and OPN, also decreased, both at the mRNA and protein levels. These results demonstrated that downregulated GARP expression resulted in the reduction of TGFβ1 and the attenuation of osteoblast differentiation of BMSCs in vitro.

Introduction Implant supported dentures have become a mainstream treatment in repairing dentition defects and deletions ( Alkan et al. , 2018 ; Greenberg, 2017 ). However, anodontia caused by serious decay, periodontal disease, congenital anodontia, and trauma may result in severe alveolar bone resorption, which can be a disadvantage of implantation. Although autogenous bone grafts are generally thought to be the gold standard for bone regeneration ( Moses et al. , 2007 ; Sbordone et al. , 2014 ; Zhang et al. , 2014 ), this procedure is limited due to local complications and an insufficiency of bone volume in donor sites ( Jensen, Jensen & Worsaae, 2016 ; Nkenke & Neukam, 2014 ). A novel and easily available treatment is therefore necessary for bone regeneration. Mesenchymal stem cells (MSCs), which can be isolated from tissues such as blood, adipose tissue, bones, and teeth, have the multipotential to differentiate into osteoblasts, chondrocytes, and adipocytes, and have become a promising vector for bone tissue engineering ( Ma et al. , 2017 ). Because they are easy to cultivate and expand, and maintain their pluripotency after serial subcultivation in vitro, bone marrow MSCs (BMSCs) are an ideal vehicle for tissue regeneration in severe bone defects and bone remodeling ( Yang, FMV & Putnins, 2010 ; Han et al. , 2013 ). However, bone formation is a complex process controlled by many factors ( Loeffler et al. , 2017 ; Schroeder & Mosheiff, 2011 ), especially transforming growth factor β1 (TGFβ1). TGFβ1, which is secreted by bone marrow stromal cells and hematopoietic progenitors, is abundant in bone matrix, and functions as a modulator of cell growth, inflammation, and matrix synthesis ( Kirk & Kahn, 1995 ; Kim & Niyibizi, 2001 ; Rahman et al. , 2015 ; Majidinia, Sadeghpour & Yousefi, 2018 ), so we assumed that the use of TGFβ1 may be an alternative approach in stem cell-based tissue engineering. However, TGFβ1 is secreted as an inactive complex, which includes the mature TGFβ1 dimer and the latency-associated proteins (LAPs). Furthermore, LAPs remain inactive. It is the removal of TGFβ1 from LAPs that activates the TGFβ1 function. It is generally accepted that the latent TGFβ1-binding proteins (LTBPs) participate in the transportation and activation process. LTBPs, associated with latent TGFβ1, assist the complex transfer and then are anchored to the extracellular matrix (ECM). The involvement of protease, thrombospondin-1, and integrin, and changes in the condition of reactive oxygen species and pH then trigger the activation process of latent TGFβ1. Another hypothesis is that membrane Glycoprotein A repetitions predominant (GARP) also participates in the activation of TGFβ1. GARP, which encodes a leucine-rich repeat containing 32 (LRRC32) protein, was first isolated from a breast carcinoma ( Ollendorff et al. , 1992 ). Most proteins of GARP with the LRR motif are membrane bound, and the remaining proteins can be secreted to extracellular sites or are localized to the cytoplasm or nucleus ( Rothberg et al. , 1990 ; Ohkura & Yanagida, 1991 ). Many relevant studies have focused on the biological function of GARP, and have proposed that GARP is involved in the activation of latent TGFβ1. It was reported that GARP and latent TGFβ1 are co-localized on the membrane of BMSCs ( Carrillo-Galvez et al. , 2015 ; Niu et al. , 2017 ), but the interaction between them and whether GARP regulates the bioactivities of TGFβ1 in BMSCs still remain to be elucidated. Furthermore, whether GARP can affect the osteogenic differentiation ability of BMSCs is unknown. We therefore downregulated the expression of GARP in BMSCs, and found that the secretion of TGFβ1 was decreased, and the osteogenic results showed attenuated osteogenesis while the expression of GARP was knocked-down. Materials and Methods Animals All the BMSCs were isolated from 4-week-old male Sprague-Dawley (SD) rats under SPF raising conditions. The experiments were approved by the Animal Research Committee of Zhongshan Hospital, Fudan University, Shanghai, China (2016-128). Isolation and culture of BMSCs Rats were sacrificed by cervical dislocation. The femur and tibia were isolated, and both ends of the bones were removed to expose the marrow, which was flushed with a 10 mL syringe filled with alpha-Minimum Essential Medium (α-MEM; Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific), and 1% penicillin/streptomycin. The fresh marrow suspension was filtered through a 70 µm cell strainer and centrifuged at 500 × g for 5 min. The supernatant was discarded, and the sediment was resuspended and seeded in 25 cm 2 culture flasks. The culture medium was replaced after 48 h. BMSCs were trypsinized using 0. 25% trypsin/1mM EDTA (Gibco; Thermo Fisher Scientific) to subculture the cells when they grew to 80% confluency. Cells at passages 3–6 were used for all experiments. Identification and GARP expression of rat BMSCs by using flow cytometry To identify the target cell, we used fluorescein isothiocyanate (FITC)-conjugated CD90 (561973; 1:100; BD biosciences; Franklin, Lakes, NJ, USA), CD45(561867; 1:100; BD biosciences), CD105(NB500-453; 1:100; novus biologicals; USA) and phycoerythrin (PE)-conjugated CD 44 (MA5-16908; 1:100; Thermo Fisher Scientific Inc. ) to label the BMSC membranes. For analysis of GARP expression, we used a LRRC32/GARP antibody (NBP2-24664; 1:1, 000; Novus Biologicals) followed by Alexa Fluor 647 AffiniPure secondary antibody (111-605-003; 1:100; Jackson Immuno Research Laboratories, West Grove, PA, USA). When cells grew to 80% confluency, they were harvested by gently washing once with phosphate-buffered saline (PBS), trypsinizing for 2 min at 37 °C, and centrifuging for 5 min at 500 × g. The cells were resuspended to a concentration of 1 ×10 6 cells/mL in PBS, and 100 µL suspensions were used for each sample. One µg of antibody was added to each sample, then incubated on ice in the dark for 1 h. One mL of PBS was added to stop the reaction, then the suspension was centrifuged for 5 min at 500 × g, and the supernatant was discarded. The pellet was washed twice by resuspending the cells in PBS and centrifuged for 5 min at 500 × g, followed by discarding the supernatant. The pellet was resuspended in 500 µL of PBS per sample, and used for the flow cytometry (BD Biosciences) analysis. Immunofluorescent staining The third passage BMSCs were seeded in a 6-well culture plate with prepared cell sheets at a density of 1. 5 ×10 4 cells/mL, and the cells were incubated overnight. When the cells were 50% confluent, the cell sheets were washed three times by PBS. The cells were then fixed with 4% formalin for 30 min, followed by washing three times with PBS, and blocked with 5% bovine serum albumin for 30 min. The experimental and control group cells were then incubated with anti-GARP antibody (NBP2-68740; 1:100; Novus Biologicals) and homologous anti-IgG antibody (A7016; 1:100; Beyotime Institute of Biotechnology, Haimen, China) at 4 °C in the dark overnight respectively, and the cell sheets washed three times with PBS. The cells were then incubated with FITC secondary antibody (A0562; 1:100; Beyotime Institute of Biotechnology) at room temperature for an hour. Cell sheets were washed three times with PBS. Then cells were incubated with 4′, 6-diamidine-2′-pheynylindole dihydrochloride (DAPI; Invitrogen, Carlsbad, CA, USA) for 10 min in the dark, and then washed as previously described. The cell sheets were imaged using an immunofluorescence microscope (Olympus, Tokyo, Japan). Lentivirus gene vector production and transfection of BMSCs The GARP short hairpin RNA (shRNA)-encoded lentivirus vector was packaged by Shanghai Hanyin (Shanghai, China). Lentiviral vectors for rat GARP-shRNA carried a green fluorescent protein (GFP) sequence. The target sequences for rat GARP knockdown are listed in Table 1. The recombinant GARP-shRNA lentivirus and the negative control lentivirus were prepared and titered to 10 9 TU/mL (transfection unit). Before transfection, BMSCs were seeded at a density of 5 × 10 4 cells per well in a 6-well culture plate and incubated overnight. The cells were approximately 50% confluent on the day of infection. A mixture of medium with polybrene (Shanghai Hanyin) was prepared at a final concentration of 5 µg/mL. The medium was removed from the plate wells and replaced with 1 mL of this polybrene/media mixture. The cells were infected by adding 10 µL shRNA lentivirus to the culture. The plate was gently swirled to mix and incubated in 5% CO 2 at 37 °C. The medium was changed after 8 h. Stable clones expressing a GARP shRNA (GARP-sh) were selected using 3 µg/mL puromycin dihydrochloride (Shanghai Hanyi). At the same time, negative control cells (NC) were infected by the same dose of empty vector virus. 10. 7717/peerj. 6993/table-1 Table 1 Primers for RT-qPCR and GARP shRNA sequences. This table showed the primers for RT-qPCR in the present study and GARP shRNA sequences. Gene Forward (3′–5′) Reverse (5′–3′) GARP GGCAGAGAACAGCCTCACTC AAGGCACCATCCTCAATGTC ALP GATGGACAAGTTCCCCTTTG CCTTCACGCCACACAAGTAG Runx2 CCTCTGACTTCTGCCTCTGG CCTCTGACTTCTGCCTCTGG OPN CCAAGCGTGGAAACACACAGCC GGCTTTGGAACTCGCCTGACTG β-actin GCAGGAGTACGATGAGTCCG ACGCAGCTCAGTAACAGTCC GARP-sh1 GGCTCAACCTACAGGGAAA GARP-sh2 GGTTAAAGGCTCAGAGAAC GARP-sh3 GGCTGTACTTGCAGGGAAA GARP-sh4 GCACTTCGCCACCTGGATTTA GARP-sh5 GCAACAGCATTGAGACCTTCC Notes. GARP Glycoprotein A repetitions predominant ALP Alkaline phosphatase OPN Osteopontin Runx2 Runt-related transcription factor 2 Cell Counting Kit-8 (CCK-8) analysis of transfected BMSCs To further investigate the differences of proliferation and viability between transfected and controlled BMSCs, CCK-8 experiments were performed. The transfected and controlled BMSCs were seeded in a 96-well culture plate at a density of 1 × 10 3 cells/mL. At the same time, there were six wells with only culture medium, which were used as the blank controls. The plates were then incubated at 37 °C. According to the manufacturer’s instructions, 10 µL CCK-8 solution (Dojindo Molecular Technologies, Kumamoto, Japan) was added to each well. The optical density (OD) value was measured at a wavelength of 450 nm, with the results presented as the mean. Quantitative real time-polymerase chain reaction (RT-qPCR) Extractions of total RNA were obtained from each group using TRIzol (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) according to manufacturer’s instructions. Reverse transcription of RNA was conducted using the PrimeScript RT Master Mix (Takara Bio, Ostu, Japan). RT-qPCR analyses were performed by using the SYBR-Green Real-Time PCR Master Mix (Takara Bio). Rat β-actin, a single housekeeping gene, was selected as an internal control for normalization. All primers used in the RT- qPCR are listed in Table 1. The RT-qPCR procedure was conducted using the following program: one cycle at 95 °C for 30 s; 40 cycles at 95 °C for 5 s; and one cycle at 60 °C for 30 s. The total volume of the reaction was 20 µL, and the relative expression of the target gene was analyzed by the comparative cycle threshold (2 − ΔΔ Ct ) method. Western blot analysis For western blot analysis, cells from each group were washed with PBS on ice. Cells were lysed using RIPA lysis buffer (Beyotime Institute of Biotechnology). The lysate was collected and then centrifuged at 12, 000× g for 5 min. The supernatant was collected and the protein concentration was determined using a BCA protein assay kit (Beyotime Institute of Biotechnology). Twenty µg of protein lysate from each sample was resolved using a 10% SDS-PAGE gel and then transferred to a polyvinylidene fluoride membrane. The membranes were blocked with 5% nonfat milk for 2 h at room temperature (RT), and the membrane was incubated with primary antibodies including anti-GARP (NBP2-24664; 1:1, 000; Novus Biologicals), anti-ALP (08337; 1:10, 000; Abcam, Cambridge, UK), anti-OPN (91655; 1:1, 000; Abcam), anti-Runx2 (23981; 1:1, 000; Abcam), and anti- β-actin (8227; 1:1, 000; Abcam) at 4 °C overnight. The next day, the membranes were washed in TBST (20 mM Tris-HCL, 137 mM NaCl and 0. 1% Tween-20; pH 7. 6) three times, for 10 min each time. The membrane was then blocked with goat-anti-rabbit IgG-HRP antibody (SC-2370; 1:5, 000; Santa Cruz Biotechnology, Dallas, TX, USA) for 1 h, followed by washing in TBST three times, with 10 min for each wash. Proteins were detected by Chemiscope5600 (Shanghai Clinx Science Instruments, Shanghai, China) using BeyoECL Plus Reagent (Beyotime Institute of Biotechnology). TGFβ1 analysis by ELISA of transfected BMSCs When cells grew to 80% confluency, the medium was changed to SD rat mesenchymal stem cell osteogenic differentiation medium (Cyagen Biosciences, Santa Clara, CA, USA). Supernatants were collected at 6 h, 12 h, 18 h, 24 h, and 48 h after the change of medium. All samples were centrifuged at 2, 500 rpm/min for 20 min, and cell debris was discarded to obtain the supernatant. TGFβ1 levels were analyzed using a rat TGFβ1 ELISA kit (Dakewe Bioengineering, Shenzhen, China), according to the manufacturer’s instructions. Osteogenic differentiation Cells were seeded at a density of 5 × 10 4 cells per well in a 6-well culture plate and incubated. When cells grew to 80% confluency, the medium was replaced with SD rat mesenchymal stem osteogenic differentiation medium (Cyagen Biosciences). The differentiation medium was changed every 3 days. Alkaline phosphatase (ALP) activity analysis After osteogenic differentiation, cells were washed gently with PBS, fixed with 4% formalin for 30 min, and stained with a BCIP/NBT alkaline phosphatase color development kit (Beyotime Institute of Biotechnology) for 30 min at RT. Finally, they were washed with PBS and then visualized with a microscope (Olympus, Tokyo, Japan). Alizarin Red S staining experiment On the days 14 after the osteogenic differentiation, cells were gently washed with PBS, fixed with 4% formalin for 30 min and then washed with ddH 2 O. Each well was added about 1mL Alizarin Red S staining reagent (Cyagen Biosciences) for 3–5 min at RT. Then cells were gently washed with ddH 2 O and visualized with a microscope (Olympus, Tokyo, Japan). Statistical analysis All data are expressed as the mean ± standard deviation. P values < 0. 05 were considered as statistically significant differences. The comparisons of two groups were performed using the two-tailed t -test, and the results of more than two groups were analyzed by one-way analysis of variance. Statistical analysis was performed using Prism 6. 0 software (GraphPad, La Jolla, CA, USA). Results Identification of rat BMSCs and GARP expression of BMSCs The mature BMSCs adhered to the bottom of the culture flask and showed a spindle shape and typical fibroblast-like appearance ( Fig. 1 ). To identify the purity of BMSCs, we assessed the expression of bone marrow mesenchymal and hematopoietic stem cell membrane epitopes, CD44, CD90, CD105 and CD45. Based on the cytometry results ( Fig. 2 ), positive percentage of BMSCs for the CD44 was 95. 82%, the CD90 99. 85%, the CD105 99. 09%, while the CD45 positive ratio was 5. 53%. This indicated that we obtained a high purity of BMSCs. For the detection of GARP expression of BMSCs, we used anti-GARP antibody to react with GARP protein, homologous anti-IgG antibody as isotype control, and FITC-conjugated secondary antibody for immunofluorescence staining, followed by imaging with a fluorescence microscope, which showed that GARP was localized to both the membrane and cytoplasm of BMSCs ( Fig. 3 ). 10. 7717/peerj. 6993/fig-1 Figure 1 BMSCs morphological characters observed under light microscope. The images revealed that BMSCs showed a spindal shape and typical fibrablast appearance. BMSCs, bone marrow mesenchymal stem cells. (A) 40X. Bar = 500 µm (B) 100X. Bar = 100 µm (C) 200X. Bar = 50 µm (D) 400X. Bar = 20 µm. 10. 7717/peerj. 6993/fig-2 Figure 2 Cytometry results of BMSCs. Cytometry results demonstrated that BMSCs were positive for CD 44 (95. 82%), CD 90 (99. 85%), CD 105 (99. 09%) and negative for CD 45 (5. 53%). BMSCs, bone marrow mesenchymal stem cells; FITC, fluorescein isothiocyanate; PE, phycoerythrin. (A) count; (B) Blank Control; (C) CD 44; (D) CD 90; (E) CD 105; (F) CD 45; ( t test). 10. 7717/peerj. 6993/fig-3 Figure 3 Immunofluorescent staining experiment. For immunofluorescent staining experiment, we used GARP antibody to show the expression of GARP protein, homologous anti-IgG antibody as isotype control and DAPI to show the location of nucleus. GARP, Glycoprotein A repetitions predominant; DAPI, diamidino-phenyl-indole. (A) and (D) showed the merge image of two groups; (B) and (E) showed the FITC immunofluorescent staining; (C) and (F) showed the DAPI staining; 200X; Bar = 50 µm. To verify the expression level of GARP during osteogenic differentiation, RT-qPCR and western blot experiments were performed at different stages of osteogenic differentiation on days 0, 7 and 14. GARP expression levels increased significantly with the increasing osteogenic differentiation time, indicating that GARP may play an important role in osteogenic differentiation ( Fig. 4 ). 10. 7717/peerj. 6993/fig-4 Figure 4 GARP expression during osteogenic differentiation. According to the qRT-PCR and western blot experiments results, on days 7 and 14 after osteogenic differentiation, GARP mRNA and protein levels increased significantly, compared to the days 0. (A) qRT-PCR experiment of GARP mRNA expression; (B) western blot experiment of GARP protein expression; ( t test, * P < 0. 05). GARP expression was knocked-down by lentivirus transfection To further investigate the bioactivities of GARP, we constructed the GARP-shRNA encoded lentivirus vector to silence the expression of GARP. After selection of successfully transfected cells by puromycin dihydrochloride, samples were extracted to analyze the GARP mRNA and protein levels via RT-qPCR and western blotting, respectively. We tested five targets for RNA interference to choose the most the effective one. Based on the RT-qPCR results, mRNA levels of GARP were significantly decreased with transfection of GARP-sh4 and GARP-sh5 ( Fig. 5 ). The subsequent experiments were therefore conducted using GARP-sh5. GARP total protein expression levels were examined by western blotting. The results showed a dramatic decline of total protein expression ( Fig. 6 ). We hypothesized that the membrane localized GARP, which bound latent TGFβ1 to the cell surface, was a key component in regulating the bioactivities of BMSCs, so we performed flow cytometry to determine GARP membrane protein levels of the treatment and control groups, which showed that the GARP membrane protein expression level in the NC group was 26. 1%, and 10. 7% in the GARP-sh group. There was a 15. 4% expression difference between the transfected and NC groups ( Fig. 7 ). 10. 7717/peerj. 6993/fig-5 Figure 5 qRT-PCR results of GARP expression. qRT-PCR results showed that GARP-sh4 and GARP-sh5 shRNA could significantly down-regulate the mRNA expression of GARP in BMSCs. ( t test, * P < 0. 05). 10. 7717/peerj. 6993/fig-6 Figure 6 Western blot experiments of GARP expression. Western blot experiments showed that GARP total protein expression in GARP-sh group was decreased, compared to control group. 10. 7717/peerj. 6993/fig-7 Figure 7 Cytometry results of GARP expression. Cytometry results demonstrated that the GARP expression was 26. 1% in NC group and 10. 7% in GARP-sh group. ( t test); (A) the count number of tested cells; (B) blank group; (C) NC group; (D) GARP-sh group. Based on the CCK-8 analysis result, cells of GARP-sh group and NC group were in the slow growth period on the first two days after seeding in the plate. On the days 3, the growth pattern of cells began to display an exponential growth phase. On the days 6, the cells reached to the plateau phase ( Fig. 8 ). Compared to the NC group, No significant difference was showed and the silence of GARP didn’t affect the viability and proliferation of BMSCs. 10. 7717/peerj. 6993/fig-8 Figure 8 Cell Counting Kit-8 (CCK-8) analysis of transfected BMSCs. CCK-8 analysis result demonstrated that cells of GARP-sh group and NC group were in the slow growth period on the first two days after seeding in the plate. On the days 3, the growth pattern of cells began to display an exponential growth phase. On the days 6, the cells reached to the plateau phase. There was no significant differences between two groups ( t test. ). Decreased levels of active TGF β 1 were detected after transfection. It was reported that GARP bound latent TGFβ1 on the cell surface ( Tran et al. , 2009 ), so we proposed that a decreased GARP protein level may affect TGF β1 secretion. To confirm this possibility, we developed an ELISA to determine the total TGFβ1 expression of supernatants from the control and GARP-sh groups. We collected supernatants at 6 h, 12 h, 24 h, and 48 h after the change of fresh medium. The results showed that the TGFβ1 secretion level was gradually increased before 24 h and then reached a plateau ( Fig. 9 ). Based on comparison of the two groups, silencing of GARP expression decreased the extracellular secretion of active TGFβ1. 10. 7717/peerj. 6993/fig-9 Figure 9 ELISA experiments of mature TGF β 1 level in GARP-sh and NC groups. ELISA experiments showed that mature TGFβ1 level was down-regulated in GARP-sh group. ( t test, * P < 0. 05). Inhibition of osteogenic differentiation of BMSCs after transfection with the GARP-sh lentivirus Numerous factors were involved during bone ossification, and it was confirmed that ALP, Runx2, and OPN were essential to this process. A high concentration of ALP denoted the start of ossification. Runx2 is the main osteogenic differentiation component of BMSCs. OPN participates in the proliferation and calcification of BMSCs. Thus, to evaluate the osteogenic differentiation ability of BMSCs, we performed ALP activity staining, RT-qPCR, and western blotting to analyze ALP, Runx2, and the OPN protein expression levels, respectively, at different stages of osteogenic differentiation on days 0, 7, and 14. RT-qPCR results showed that ALP, Runx2, and OPN expressions showed no difference on days 0 of differentiation. But on days 7, the expressions of these factors increased to varying degrees, when compared with those on days 0. In addition, the expressions of these factors in the GARP-sh group were lower than the NC group. On days 14, ALP, Runx2, and OPN expressions were higher than on days 0 and days 7, and there was a significant decline of these factors in the GARP-sh group ( Fig. 10 ). 10. 7717/peerj. 6993/fig-10 Figure 10 Osteogenic differentiation related factors’ mRNA expression results. Osteogenic differentiation results demonstrated ALP, Runx 2 and OPN mRNA expression was decreased in GARP-sh group at osteogenic differentiation 7 days and 14 days. ( t test, * P < 0. 05); (A) mRNA expression of ALP; (B) mRNA expression of Runx2; (C) mRNA expression of OPN. Western blot analysis indicated that ALP, Runx2, and OPN protein levels showed no significant difference between the two groups during the pre-induction of osteogenesis. On days 7 and 14, the results displayed remarkable differences in the ALP, Runx2, and OPN expression levels between the GARP-sh and NC groups ( Fig. 11 ). 10. 7717/peerj. 6993/fig-11 Figure 11 Western blot analysis of osteogenic related factors’ protein expression. Western blot analysis showed that ALP, Runx2 and OPN protein expressions were decreased in GARP-sh group. On the days 7, the ALP activity staining assay showed that the color of ALP staining was much more obvious than that of the GARP-sh group. On days 14, the Alizarin Red S staining level of NC group was much darker than GARP-sh group. And there are more mineralized nodule formation in the NC group, compared to the GARP-sh group ( Fig. 12 ). Together, the results showed that ALP levels and mineralizing ability were decreased in the GARP-sh group during osteogenic differentiation. 10. 7717/peerj. 6993/fig-12 Figure 12 ALP and Alizarin Red S staining analysis of osteogenic differentiation. After 7 days osteogenic differentiation, ALP staining experiments showed that the staining of NC was much higher than the GARP-sh group. on 14 days after osteogenic differentiation, Alizarin Red S staining results showed that there was more mineralized nodule formation in the NC group, compared to the GARP-sh group. (A) ALP and Alizarin Red S staining with the gross appearance. (B) ALP and Alizarin Red S staining under the microscopic view (40X). Bar = 500 µm. Discussion BMSCs show a multilineage potential, allowing differentiation into osteoblasts, chondrocytes, adipocytes, and other tissue cells, which are easily isolated and cultured. Thus, BMSCs could be suitable cells for bone tissue engineering as osteogenic stem cells ( Cancedda et al. , 2000 ). The osteogenic differentiation process is activated by a series of signal transduction processes, of which the most important is the TGFβ pathway ( Jian et al. , 2006 ). Regulation of TGFβ may therefore provide a novel approach for the differentiation of BMSCs. GARP, which has been mostly reported in the study of activated Treg cells, is expressed on the cell surface of activated functional FOXP3 + Tregs, and functions as a receptor for latent TGFβ1. Binding of TGFβ1 to the membrane is essential for the functioning of Tregs ( Tran et al. , 2009 ). With the assistance of GARP, latent TGFβ1 is localized on the cell membrane, and is then ready to release its active form ( Stockis et al. , 2009 ). However, the exact mechanism of activation of membrane latent TGFβ1 in Tregs is not known. GARP mRNA expression has been recently detected in MSCs ( Barbet et al. , 2011 ). The interaction between GARP and latent TGFβ1 has also been described in a follow-up report ( Carrillo-Galvez et al. , 2015 ). In the present study, we conducted further studies to determine the effect of GARP on the osteogenic development of BMSCs via the regulation of latent TGFβ1. Our study determined the existence of GARP in BMSCs. To further investigate the bioactivity of GARP, we performed transfection experiments to silence the expression of GARP. The results showed that both total and membrane GARP protein expressioxns were knocked down. Furthermore, we showed that the secretion of TGFβ1 was decreased with the reduction of GARP. Previous studies have reported that TGFβ1 is secreted as an inactive form (latent TGFβ1) ( Annes, Munger & Rifkin, 2003 ), and both LTBPs and GARP bind latent TGFβ1. LTBPs bind latent TGFβ1 by disulfide bonds. In this manner, LTBPs assist the transmembrane transportation and anchorage in the ECM, of latent TGFβ1. Furthermore, the storage of latent TGFβ1 in the ECM is functionally activated under specific conditions ( Taipale et al. , 1994 ). GARP can anchor latent TGFβ1 noncovalently to the cell membrane. The membrane GARP/latent TGFβ1 complex serves as a source of active TGFβ1. The membrane latent TGFβ1 can be activated by mediation of integrins α v β6 ( Wang et al. , 2012 ). In our study, the decrease of membrane GARP provided an insufficient amount of latent TGFβ1, so eventually the secretion of TGFβ1 was lower than untreated BMSCs. In contrast to the results of Carrillo-Galvez et al. , our data indicated that silencing of GARP resulted in the reduction of TGFβ1. It was reported that GARP competed better with LTBPs when binding to latent TGFβ1 ( Wang et al. , 2012 ). We speculated that membrane-associated GARP plays a more important role than the LTBPs in the process of TGFβ1 activation. TGFβ1 can be produced by BMSCs and participates in osteoblast differentiation of stem cells as a pleiotropic molecule. The activation of the SMAD signaling pathway by TGFβ1 can increase the osteoblast differentiation ( Jian et al. , 2006 ; Kulterer et al. , 2007 ; Ng et al. , 2008 ; Miron et al. , 2013 ). A previous study reported that TGFβ1 actions can be attributed to the foundation of bone mass and quality through the regulation of perilacunar/canalicular remodeling ( Dole et al. , 2017 ). In the present study, we showed that osteogenic-related genes of ALP, Runx2, and OPN were significantly downregulated with decreased secretion of TGFβ1, and that the suppression of osteogenic differentiation was also confirmed at the translational level. According to our western blot results, all ALP, Runx2, and OPN protein expression levels were significantly attenuated compared to the untreated group. Moreover, the ALP activity of GARP-sh cells was lower than that of untreated BMSCs. These results indicated that the decreased amount of TGFβ1 inhibited the osteogenesis of BMSCs. However, the biological mechanism remains to be definitively elucidated. Further animal experiments are necessary to demonstrate that the osteogenic differentiation is affected by the GARP/latent TGFβ1 complex in vivo. It has been shown that MSCs reside in the periodontal ligament, which may be a promising reservoir for regeneration ( Beertsen, McCulloch & Sodek, 1997 ). In addition, TGFβ1 can cooperate with other growth factors, such as bone morphogenetic proteins (BMPs), to induce osteoblast differentiation ( De Gorter et al. , 2011 ; Hyun et al. , 2017 ). In the present study, GARP was responsible for the regulation of the activation and secretion processes of latent TGFβ1. Furthermore, GARP indirectly regulated osteogenic differentiation of BMSCs. These results may provide promising applications in the field of tissue engineering. Conclusions We verified that the downregulation of GARP decreased the level of mature TGFβ1 and the osteogenic ability in rat BMSCs. However, bone regeneration is a rather complicated process that is regulated by various factors. It still remains to be elucidated whether a decrease in the levels of mature TGFβ1 reduces the osteogenic ability in vivo. Based on the results above, we hypothesize that the osteogenic ability can be improved by increased TGFβ1 levels via the upregulation of GARP. The regulation of GARP on levels of TGFβ1 may therefore provide a novel solution to challenges in tissue engineering and bone regeneration. Supplemental Information 10. 7717/peerj. 6993/supp-1 Dataset S1 Raw data of BMSCs images under light microscope Click here for additional data file. 10. 7717/peerj. 6993/supp-2 Dataset S2 Raw data of GARP expression Click here for additional data file. 10. 7717/peerj. 6993/supp-3 Dataset S3 Raw data of osteogenic differentiation Click here for additional data file.

10. 7717/peerj. 7160

PeerJ

Main histological parameters to be evaluated in an experimental model of myocardial infarct treated by stem cells on pigs

Myocardial infarction has been carefully studied in numerous experimental models. Most of these models are based on electrophysiological and functional data, and pay less attention to histological discoveries. During the last decade, treatment using advanced therapies, mainly cell therapy, has prevailed from among all the options to be studied for treating myocardial infarction. In our study we wanted to show the fundamental histological parameters to be evaluated during the development of an infarction on an experimental model as well as treatment with mesenchymal stem cells derived from adipose tissue applied intra-lesionally. The fundamental parameters to study in infarcted tissue at the histological level are the cells involved in the inflammatory process (lymphocytes, macrophages and M2, neutrophils, mast cells and plasma cells), neovascularization processes (capillaries and arterioles) and cardiac cells (cardiomyocytes and Purkinje fibers). In our study, we used intramyocardial injection of mesenchymal stem cells into the myocardial infarction area 1 hour after arterial occlusion and allowed 1 month of evolution before analyzing the modifications on the normal tissue inflammatory infiltrate. Acute inflammation was shortened, leading to chronic inflammation with abundant plasma cells and mast cells and complete disappearance of neutrophils. Another benefit was an increase in the number of vessels formed. Cardiomyocytes and Purkinje fibers were better conserved, both from a structural and metabolic point of view, possibly leading to reduced morbidity in the long term. With this study we present the main histological aspects to be evaluated in future assays, complementing or explaining the electrophysiological and functional findings.

Introduction During the last decade, many attempts have been made to develop a treatment for heart failure using experimental models of myocardial infarction (MI). Most of the attempts have been based on advanced therapies. Although promising advances have been achieved, trial results have limitations in the clinical phase ( Chen et al. , 2014 ; Kim et al. , 2017 ; Liu et al. , 2016 ). The attempts that have reached clinical trial phases have used mesenchymal stem cells (MSCs). From diverse origins, MSCs are immunoprivileged cells that are suitable for allogeneic and xenogeneic transplantation ( Aggarwal & Pittenger, 2005 ). As immunomodulatory cells, they act on the injured or inflamed area by secreting various growth factors. To do so, they stimulate the proliferation of local cells which also helps with remodeling the local matrix. In summary, they improve both healing and repairing of post-MI tissue ( Wu et al. , 2007 ). Currently, MSC-based therapies present a promising treatment plan for decreasing morbidity and mortality from chronic diseases with poor wound healing ( García-Gómez et al. , 2010 ). To design new therapeutic approaches based on experience to date, we should consider two important factors: selecting the correct experimental model and choosing the most appropriate parameters to analyze. Experimental models must meet minimum requirements, including being feasible and reproducing human disease as closely as possible. A porcine model has proven to be the most effective model for studying ischemic heart disease, due to the anatomical, physiological and pathological similarities between human and pig hearts ( Dixon & Spinale, 2009 ). Most of these models are based on surgical techniques for coronary artery occlusion after thoracotomy ( Hoffmann et al. , 2004 ; Klocke et al. , 2007 ; Litvak, Siderides & Vineberg, 1957 ) or are based on less invasive techniques ( Krombach et al. , 2005 ; Yoshimizu et al. , 1986 ) that reproduce the ischemic lesion as accurately as possible. The most frequently used model involves anterior descending coronary artery occlusion (due to its simple and reproducible approach) with occlusions at various levels and a selective venous return. In all cases, an arterial occlusion lasting more than 1 h causes an irreversible effect similar to a permanent occlusion ( Verdouw et al. , 1998 ). Considering the parameters to be analyzed, most studies have focused mainly on the acute phase of the infarction, and results have been based on clinical and electrophysiological parameters ( Chen et al. , 2014 ; Kim et al. , 2017 ; Liu et al. , 2016 ). In order to prevent and treat pathologies derived from the affectation of the coronary arteries and consequent cardiac failure, it is necessary to understand not only the clinical-electrophysiological parameters but also the pathophysiology of the disease and the progress of the tissues involved in MI. Different types of cells have been used in experimental assays on pigs ( Castro et al. , 2019 ; Johnson & Singla, 2017 ; Mohsin & Houser, 2019 ; Shah et al. , 2018 ; Van der Spoel et al. , 2011 ; Wang et al. , 2017 ). The types of cells that have been tested in various clinical trials ( http://www. clinicaltrials. gov ) with disappointing results have included bone marrow mononuclear cells (BMNCs), bone marrow mesenchymal stem cells(BMSCs), mesenchymal cells derived from adipose tissue (ASCs) and, more recently, pre-derived embryonic cells (ESCs), induced pluripotent stem cells (iPSCs) and cardiac progenitor cells (CSCs) ( Higuchi et al. , 2017 ). We used a model of descending coronary artery occlusion to describe the pathophysiology of the process and the intracardiac injection of adipose-derived mesenchymal stem cells (ASCs), a type of MSCs that are abundant and easy to obtain, to analyze histological changes in both the infarcted tissue and the adjacent area. In this way, we were able to observe the effects of the intralesional application of ASCs. Our research involved analyzing the changes that occurred during the inflammatory process, cicatrization and neovascularization, all of which are characteristic events in post-MI tissue regeneration. In addition, we studied the state of conservation of cardiac cells, cardiomyocytes and Purkinje fibers close to the infarcted region. To that end, we created an MI occluding the left anterior descending coronary artery and we compared the experimental treatment (intralesional injection of 5 × 10 6 cells, MI + ASC group) with the natural evolution of infarcted tissue (saline injection, Myocardial Infarction Control (MIC) group). All the animals were euthanized after 4 weeks. Our objective was to determine, from a histopathological point of view, the effect of ASCs on the previously discussed patterns that are fundamental to understanding the functional consequences of long-term post-MI. With this study we explored the main histological aspects to be evaluated in assays in MI animal models treated by stem cells. These studies can add rationality to understanding the electrophysiological and functional findings. Materials and Methods Animals Nine Landrace-Large White pigs 23–28 kg were used: one was male -used to obtain ASCs from subcutaneous adipose tissue- and the rest were females destined for the intervention. All procedures were performed in the Experimental Surgery Department of La Paz University Hospital in Madrid, Spain. We followed the protocol approved by the Animal Welfare Ethics Committee (CEBA 14-2011) and complied with the EU Directive on experimental animals (63/2010 EU) and related Spanish legislation (RD 53/2013). Surgery Twenty-four hours before surgery, the animals were pre-medicated with a fentanyl patch, and were administered 12 mg/kg ketamine (im), midazolam 0. 5 mg/kg (im), and tramadol 5 mg/kg (im) 15 min before the intervention. The animals were anesthetized using isoflurane 2% and a continuous infusion of morphine (12 mg), ketamine (30 mg) and lidocaine (15 mg) in 500 ml saline solution at a rate of 10 ml/kg/h throughout the intervention. A lateral thoracotomy was performed, and the anterior descending artery was ligated by 6/0 silk suture (Ethicon). One hour after ligation, 5 × 10 6 stem cells or saline solution (1 ml) were delivered by syringe into the myocardial ischemic tissue around the infarct area (at 3 points). After surgery, the animals were kept individually isolated in a 2 m 2 space with controlled food and water ad libitum. Analgesia was provided with a fentanyl patch every 2 days for a week, and the animals tolerated food the day after surgery. Ceftriaxone 40 mg/kg (im) was used as antibiotic prophylaxis from 3 days before until 72 h after the intervention. Isolation of Adipose-Derived Mesenchymal Stem Cells Adipose tissue-derived stem cells (ASCs) were obtained from an animal of the same breed according to a previously described protocol with minor modifications ( Zuk et al. , 2001 ). After that, the cells were expanded in culture with Dulbecco’s Modified Eagle’s Medium, supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C and 5% CO 2. The cells were characterized by flowing cytometry and differentiated as adipocytes, osteocytes and chondrocytes, confirming that we were working with mesenchymal stem cells according to International Federation for Adipose Therapeutics criteria ( Bourin et al. , 2013 ). Finally, the cells were expanded in vitro and aliquots of 5 million were frozen with 10% dimethyl sulfoxide and stored in liquid nitrogen after use. One week before the intervention, the required aliquots were defrosted and cultured until a sufficient number was obtained (5 × 10 6 cells/animal). Before use, the cells were detached from the culture with trypsin-EDTA and were washed three times with phosphate buffered saline (PBS, Gibco). Prior to injection, they were marked with CelltrackerDil (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions in order to identify ASCs in animal tissue samples. The cells were located in red spectra (553/570 nm) by fluorescence microscopy. Obtaining Samples In all cases, a 2 mL sample of blood was drawn prior to surgery and 24, 48 and 72 h post-surgery, for the purpose of analyzing troponin and biochemical parameters. All animals were euthanized 1 month after surgery by intravenous injection of 5M potassium chloride, having previously been anesthetized using isoflurane 2%. The hearts were extracted and washed with 10% formaldehyde through the coronary artery and mitral valve. Once the tissues were fixed, samples were obtained from the infarct area and periphery. Histological analysis Five mm 3 samples were fixed in 10% formaldehyde at room temperature, embedded in paraffin and cut into 5-micron-thick slices in a Micron HM360 microtome. Sections were stained with hematoxylin-eosin to identify plasma cells, lymphocytes, neutrophils and capillaries, using toluidine blue for the identification of mast cells. For the immunohistochemical studies, histology sections were deparaffinized and rehydrated before endogenous peroxidase activity was blocked with H 2 O 2 (0. 3%) in methanol. The slides were rinsed with PBS and incubated with primary antibodies in a moist chamber at room temperature. The sections were subsequently incubated with biotinylated anti-rabbit IgG and LBA (DAKO) for 25 min at room temperature, rinsed with PBS and immersed for 25 min in avidin peroxidase. The immunostaining reaction product was developed using diaminobenzidine. Counter staining was performed with hematoxylin. The specificity of the immunohistochemical procedure was confirmed by incubation of sections with non-immune serum instead of a primary antibody. The primary antibodies used were anti-CD14 antibody (MyBioSource, MBS-2027456, 1/500), anti-CD163 antibody (Serotec, MCA242GA, 1/200), muscle specific actin monoclonal antibody (Novocastra, A7811, 1/100), Desmin Monoclonal Antibody (Novocastra, DE-R11, 1/50), Connexin 43 mouse monoclonal antibody (Cell Signaling Technology, cst-3512, 1/50), and HIF1-α antibody (Gene Tex, GTX 30105, 1/1000). The Masson trichrome technique was used to evaluate the degree of fibrosis (percentage of collagen area [blue] versus tissue area [red]). All histological slides were studied under a Zeiss Axiophot 2 microscope and photographed with an AxiocamHRc camera. Twenty contiguous non-overlapping fields (200 × or 400 ×) per slide from each group were counted according to the Novotny et al. (2018) and the Yuetal2018 protocols. All the cells were quantified by the same researcher without knowledge of the groupings. Statistical analysis Heart rate differences between groups were assessed using the corrected chi-squared test. Results with a p value of less than 0. 05 were considered significant. Results Animal model The experimental animal model we created showed high infarction reproducibility due to occlusion of the left anterior descending coronary artery in its upper third, as well as a high animal survival rate (90%). In all cases, the necrotic tissue was in the same anatomical region, was approximately 1. 4 cm (±0. 2) in diameter, and was associated with ST elevation as shown in the electrocardiogram. Subsequently, MI was confirmed by a >20% increase in basal troponin levels at 24 h post occlusion and a decrease of those levels by about 50% at 48 h. ASC location after injection Immunohistological detection of the DiL signal in the cicatricial region was observed in the IM + ASC group after one month. As expected, no positive DiL signal was observed in the control hearts ( Fig. 1 ). 10. 7717/peerj. 7160/fig-1 Figure 1 Adipose Stem Cells in the myocardium of MI+ASC heart group. (A, C and E) Fluorescence Microscopy Celltracker Dil 400 ×; (B, D, and F) Hematoxylin-eosin, 400 ×. Histological results ( Tables 1 and 2 ) Inflammatory reaction After the quantitative analysis of the infarcted area, we observed a greater infiltration of plasma cells and mast cells (2:1) in the animals treated with ASCs (MI + ASCs). In addition, the quantity of the neutrophils was less (1:3) and the number of lymphocytes was similar in both groups. 10. 7717/peerj. 7160/table-1 Table 1 Histological results in the scar area. SCAR AREA MI+ASCs MIC PLASMA CELLS 20X * ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 8. 73 ± 0. 50 3. 7 ± 0. 41 NEUTROPHILS 40X * ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 9. 05 ± 0. 68 12. 86 ± 0. 74 LYMPHOCYTES 40X ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 8. 46 ± 0. 75 7. 06 ± 0. 55 MAST CELLS 20X * ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 1. 34 ± 0. 17 0. 51 ± 0. 11 MACROPHAGES 20x ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 13. 97 ± 0. 94 15. 12 ± 0. 81 MACROPHAGES M2 20X * ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 2. 65 ± 0. 38 0. 75 ± 0. 13 CAPILLARIES ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 9. 02 ± 0. 57 9. 38 ± 0. 64 ARTERIOLES ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x} \pm $\end{document} x ¯ ± se) 3. 13 ± 0. 23 2. 31 ± 0. 19 SCAR AREA (µm 2 ) 26. 15 ± 1. 96 23. 18 ± 2. 02 SCAR FIBROSIS DENSER LAX Notes. * P < 0. 05. se standard error \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\bar {x}$\end{document} x ¯ average 10. 7717/peerj. 7160/table-2 Table 2 Histological condiction of cardiac cells in the pericicatricial zone. PERICICATRICIAL ZONE CARDIOMYOCYTES PURKINJE FIBERS MI+ASCs MIC MI+ASCs MIC LOSS OF THE ACTINE PATTERN + ++ ++ +++ LOSS OF THE DESMIN PATTERN + ++ + ++ CHANGES IN GAP JUNCTION + ++ + ++ NUCLEOUS POSITIVEFOR ANTI- HIF-1α ANTIBODY * (% positive nuclei) 34. 75(45. 006%) 13. 75(18. 1%) 5. 63 (56. 44%) 4. 32(23. 64%) Notes. - NONE + MILD ++ MODERATE +++ SEVERE * P < 0. 05. 10. 7717/peerj. 7160/fig-2 Figure 2 Macrophages in the peri-cicatricial area. The number of macrophages (arrow) in the peri-cicatricial area of MI+ASC heart (A) is similar to MIC group (B). Immunohistochemistry anti-CD14, PAP. 400 ×. M2 macrophages (arrows) are in greater proportion in MI+ASC group (C) than in MIC (D). Immunohistochemistry anti-CD163 PAP. 400 ×. The amount of CD14+ macrophages quantified in the infarcted area was greater in the MIC group. CD163+ macrophages were present in greater amounts in the MI + ASC group (15. 17% of CD14+ macrophages were CD163+). In the MIC group, 4. 95% of the CD14+ macrophages were CD163+ ( Fig. 2 ). We did not find inflammatory infiltrate outside the infarct area in any of the animal groups studied. Vascular density The number of capillaries was similar in both groups. We found more arterioles in the scarring region of the hearts belonging to the MI + ASC group than in the MIC group (at a ratio of 3:2) ( Figs. 3A – 3B ). 10. 7717/peerj. 7160/fig-3 Figure 3 Scarring region. Capillaries (asterisk) and arterioles (arrows) in the scarring region of MI+ ASC hearts (A) and MIC hearts (B). Hematoxylin-eosin, 200 ×. Denser scar in ASC group (C) than in Control group (D). Masson, 200 ×. Collagen deposition A denser and more organized scar was found in the MI + ASC group than in the MIC group. In both groups, fibrosis extension was similar ( Figs. 3C – 3D ). Cardiomyocytes in the peri-cicatricial region The MI + ASC group had better conservation of the cytoskeleton than the MIC group based on actin/desmin staining ( Figs. 4A – 4D ). 10. 7717/peerj. 7160/fig-4 Figure 4 Cardiomyocytes in the peri-cicatricial zone. Labeling of actin demostrate little areas with loss of actine expression in MI+ASC hearts (A), in MIC group (B) actin pattern is lost in bigger zones (asterisk). We can see cells with contraction-bands (yellow arrows). Immunohistochemistry anti-actin, PAP. 200 × Desmin expression is better conserved (white arrows) in MI+ASC group (C) than MIC group (D, yellow arrows). Immunohistochemistry anti-desmin, PAP. 200 ×. Greater number of HIF-1α-positive nuclei (white arrows) in MI+ASC (E) than in MIC group (F, yellow arrows). Immunohistochemistry anti- HIF1-α, PAP. 400 ×. Connexin 43 is better conserved in MI+ASC hearts (G, white arrows) than in MIC hearts (H) where it is diminished and granular (yellow arrows). Immunohistochemistry anti connexin 43, PAP. 200 ×. Studying the distribution of Connexin 43 cells, we observed that, in the MI + ASC group, there was protein expression, and its distribution was conserved. In the MIC group, the expression of Connexin 43 was reduced and its distribution was altered ( Figs. 4E – 4F ). Immunohistochemistry revealed HIF1-α protein throughout areas of perinfarcted myocardium. In the MI + ASC group, 45% of the cardiomyocytes had positive nuclei for HIF-1α, whereas in the MIC group, only 18. 1% of them were positive ( p < 0. 05) ( Figs. 4G – 4H ). Purkinje fibers in the peri-cicatricial region The structure of the cytoskeleton of the MI + ASC group was well preserved, whereas in the MIC group we observed a decrease of filaments ( Figs. 5A – 5D ). 10. 7717/peerj. 7160/fig-5 Figure 5 Purkinje fibres in the peri-cicatricial zone. Actine filaments in the Purkinje fibers are better conserved in MI+ASC (A, white arrows) than in MIC ones (B, yellow arrows). Immunohistochemistry anti-actin, PAP. 200 ×. Most of the desmine filaments are better conserved in MI+ASC group (C, white arrows) than in the MIC group hearts (D, yellow arrows). Immunohistochemistry anti-desmin, PAP. 200 ×. More positive nuclei (white arrows) in the cells of pericicatricial zone in MI+ASC group (E) than in control group (F, yellow arrows). Immunohistochemistry anti- HIF1-α, PAP. 400 ×. Better conserved pattern of connexin 43 between Purkinje fibers in MI+ASC group (G, white arrows) than in control group where, also, it is diminished (H, yellow arrow). Immunohistochemistry anti-connexin 43, PAP. 400 ×. The immunohistochemical technique reflected a decrease in the expression of Connexin 43 cells in the gap junctions of the Purkinje fibers in the MI + ASC group, a decrease that was even more marked in the MIC group, in which its distribution was irregular ( Figs. 5E – 5F ). Based on the expression of HIF-1α, immunoreactivity was observed in the nuclei of the Purkinje fibers throughout the areas of perinfarcted myocardium and was not present in non-infarcted myocardium. In the MI + ASC group, 56. 44% of the Purkinje fibers had positive nuclei, whereas in the MIC group, only 23. 64% had positive HIF-1α nuclei ( p < 0. 05) ( Figs. 5G – 5H ). Discussion Various studies have been published on myocardial infarction models after a permanent ligation, especially in the acute phase. Most of these have focused on evaluating the functionality and electrophysiology of the proposed treatments ( De Siena et al. , 2010 ; Kim et al. , 2017 ). In our study, we emphasized the pathological aspects more than the electrophysiological ones. At the electrophysiological level, we observed a decrease in systolic and diastolic pressure after the infarction which, although it did not normalize completely, improved hours after the infarction. Although this normalization was not significant, it was observed earlier in the animals treated with cells. Also, we observed a decrease in heart rate in animals treated with cells less than one hour post-cell injection, which implies a lower cardiac output and an improvement in the state of the animals. At the blood protein level, the most significant parameter was troponin, and we observed post-infarct that troponin levels decreased at 48 h more than 50% in animals treated with cells, while the control group did not reach a 50% decrease in that period of time. These better data in the group treated with cells coincides with the pathophysiological data described: a lower inflammatory infiltrate, less fibrosis and a better conservation of the parenchyma. For our study, we analyzed the three main factors that could lead to an increase in post-infarction morbidity: a chronic inflammatory process, scarring characteristics and the status of myocardial cells remaining around the infarcted area, at a time specific to its evolution and comparing a normal evolution with experimental cell therapy treatment. One of the main goals in regenerative medicine and tissue engineering in the MI field is initial inflammatory response modulation. As Kocher et al. (2001) had already reported on the anti-inflammatory effect of Mesenchymal Stem Cells (MSCs) in myocardial infarct. Since then, numerous research groups have published articles referring to multiple reasons for the infusion of MSC to shorten and regulate the inflammatory process, for example the regulating T or T-native cells, the M1/M2 macrophage transition, the secretion of interleukins like IL10 or IL-4. ( Najar et al. , 2016 ; Krampera et al. , 2003 ; Yañez et al. , 2010 ; Hirose et al. , 2018 ). Actually, many studies have been carried out and they all have found different reasons to explain this shortening of the acute inflammatory process, but there may be many other aspects that generate this effect as suggested in 2012 by Georgiev-Hristov et al. (2012). The principal problem is a prolongation of the inflammatory phase during the wound healing process leading to adverse scarring and causing medium/long term cardiac failure ( Chen & Frangogiannis, 2016 ; Leor et al. , 2016 ). Along these lines, it is necessary to analyze the qualitative histological characteristics of that reaction: it should attract reparative cells, such as M2 macrophages, which favor the formation of well vascularized tissue, and with the right proportion of collagen for proper ventricular function. Our results show that plasma cells decrease in the infarcted tissue in an untreated heart. However, they were attracted to the inflammatory-reparative area by the ASCs and they infiltrated the infarcted tissue of the hearts treated with ASCs. Similar to other authors ( Mazo et al. , 2012 ; Rasmusson et al. , 2005 ; Rasmusson et al. , 2007 ), we consider that this increase in plasma cells in the treated animals might be due to the paracrine activity of mesenchymal cells. When we analyzed the various cells involved in the inflammatory process, we found that the control group maintained a high number of neutrophils and an increase in macrophages, which reveals an acute inflammation or an initial phase of a chronic inflammatory stage. On the other hand, ASC treatment caused late chronic phase development of a decrease in neutrophil number and an increase in type 2 macrophages. This outcome might be due to the anti-inflammatory effect of ASCs ( García-Gómez et al. , 2010 ; Kuo et al. , 2012 ; Van den Akker, De Jager & Sluijter, 2013 ) and the shortening of the inflammatory-reparative process generated by the ASCs. Another interesting result that we observed was that mast cell infiltration in the infarcted area was three times higher in the ASC-treated animals. We believe the lack of statistical significance between both groups was due to the fact that this infiltration by mast cells in an MI persists during the inflammatory process from the chronic stage ( Levick et al. , 2011 ; Reid et al. , 2011 ). The mast cells seem to be involved in the paracrine regulation of growth factors (GF) in the infarcted area, although their exact functions must be further studied ( Gentek & Hoeffel, 2017 ). Macrophages change their phenotype and function in response to signals from the microenvironment. The M1/M2 balance may influence cardiac repair improvement and post-MI function ( Leor et al. , 2016 ; Gombozhapova et al. , 2017 ). Therefore, this approach could be used as a therapeutic tool. Interactions between ASCs and macrophages are known: ASCs increase the expression of the M2 phenotype ( Ryabov et al. , 2018 ). In post-MI healing, a prolonged presence of M1 macrophages can lead to an increase in the size of the MI area and prevent correct resolution. We observed this increase of the MI area in our control animals. On the other hand, in treated animals we observed an increase in M2 macrophages to implicate a diminution in MI area. Our results, in accordance with previous data, showed a better condition of the infarcted tissue in the IM-ASC group animals. Also, ASCs present angiogenic effects when implanted in infarcted tissue. Numerous published studies have demonstrated this approach ( Chou et al. , 2014 ; Citro et al. , 2014 ; Kim et al. , 2014 ). In our study, after 1 month of evolution of the ASC group, we did not observe an increase in capillaries. Although we detected a greater number of arterioles, the difference was not statistically significant. We believe that 1 month of evolution of the infarction led to stabilization of the angiogenesis of the infarcted region. A study with analysis of shorter periods could possibly clarify this difference. In our study we observed a dense scar associated with better organization of collagen type I in the ASC group, that possibly explained the better remodeling of the perinfarcted tissue and the generation of a more organized scar. Results are in accordance with those observed by other authors ( Nong et al. , 2011 ; Yu et al. , 2018a ; Yu et al. , 2018b ). Free radical cardiac myoglobin and other sources play an important role in myocardial infarction ( Zhu & Zuo, 2013 ). These oxygen radicals can potentially influence cardiac inflammation and the survival rates of injected stem cells. Thus, further studies into the correlation between reactive oxygen species and inflammation should be carried out in the future. Finally, we determined the survival conditions of cardiac cells, cardiomyocytes and Purkinje cells. Different studies had previously implicated both cell types in remodulation of infarct tissue and a positive response to MSC treatment ( Li et al. , 2010 ; Muguruma et al. , 2006 ; Shafei et al. , 2017 ; Yoon et al. , 2005 ). In general, we can conclude that in the MI + ASC group after 4 weeks of post infarction evolution, the cardiomyocytes and Purkinje cells were better conserved, and we can affirm this point by Connexin 43 and HIF1-α status. From both qualitative and quantitative points of view, the cytoskeleton filaments of these cells and the gap junctions (Connexin 43) in ASC group were similar to non-pathological conditions. After tissue hypoxia, affected cardiac cells respond to various mechanisms aimed at restoring cellular oxygen levels. One of the main response pathways is the inhibition of hydroxylation of HIF1-α by prolyl hydroxylases (PHDs), which translocate to the nucleus, initiating the transcription of factors that support normoxia: promoting angiogenesis, increasing cell proliferation and migration, stimulating glycolysis, etc. With all this activity, cell survival is more likely ( Szade et al. , 2015 ). Thus, stabilization and accumulation of HIF1-α is considered cardioprotective, helping preserve myocardial structure and function ( Cheng et al. , 2016 ; Lee et al. , 2000 ; Townley-Tilson, Pi & Xie, 2015 ; Wu et al. , 2015 ). After 4 weeks of postinfarction evolution, we observed how the presence of ASCs coincided with a greater proportion of cardiomyocytes and Purkinje fibers expressing nuclear HIF1-α with respect to the control group. Therefore, in these ASC-treated hearts, there was a more intense adaptive response to hypoxia conditions. Given that HIF1- α is a mechanism which starts immediately after the decrease in oxygen levels, it was maintained 4 weeks after MI, especially in the MI-ASC group. We have also demonstrated that the infarction model carried out by the occlusion of the anterior coronary artery was reproducible. One month of evolution represents a period that is similar to what we observe in the clinic. In conclusion, in this study we have shown the main histological parameters to be assessed after the generation of an infarction: the cells involved in the inflammatory process, cicatricial and neovascularization processes characteristic of post-MI tissue regeneration as well as in the conservation of cardiac cells, cardiomyocytes and Purkinje fibers adjacent to the infarcted area. Finally we used Stem Cell treatment to demonstrate the implications of these histological parameters in infarct tissue remodulation. Thus, in this study, we propose what we consider to be the main histological aspects to be evaluated in future assays, and provide complementary explanations for the electrophysiological and functional findings. Supplemental Information 10. 7717/peerj. 7160/supp-1 Dataset S1 Raw Data Arterioles (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-2 Dataset S2 Raw data Capillaries (200 ×) Click here for additional data file. 10. 7717/peerj. 7160/supp-3 Dataset S3 Raw data CD14 (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-4 Dataset S4 Raw Data CD163 (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-5 Dataset S5 Raw Data HIF Cardiomyocytes (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-6 Dataset S6 Raw Data HIF Purkinje (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-7 Dataset S7 Raw Data Lymphocytes (400 ×) Click here for additional data file. 10. 7717/peerj. 7160/supp-8 Dataset S8 Raw Data Mast Cells (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-9 Dataset S9 Raw Data Neutrophils (400 ×) Click here for additional data file. 10. 7717/peerj. 7160/supp-10 Dataset S10 Raw Data Scar área (200 × field) Click here for additional data file. 10. 7717/peerj. 7160/supp-11 Dataset S11 Raw Data plasma cells (200 × field) Click here for additional data file.

10. 7717/peerj. 7233

PeerJ

An artificial-vision- and statistical-learning-based method for studying the biodegradation of type I collagen scaffolds in bone regeneration systems

This work proposes a method based on image analysis and machine and statistical learning to model and estimate osteocyte growth (in type I collagen scaffolds for bone regeneration systems) and the collagen degradation degree due to cellular growth. To achieve these aims, the mass of collagen -subjected to the action of osteocyte growth and differentiation from stem cells- was measured on 3 days during each of 2 months, under conditions simulating a tissue in the human body. In addition, optical microscopy was applied to obtain information about cellular growth, cellular differentiation, and collagen degradation. Our first contribution consists of the application of a supervised classification random forest algorithm to image texture features (the structure tensor and entropy) for estimating the different regions of interest in an image obtained by optical microscopy: the extracellular matrix, collagen, and image background, and nuclei. Then, extracellular-matrix and collagen regions of interest were determined by the extraction of features related to the progression of the cellular growth and collagen degradation (e. g. , mean area of objects and the mode of an intensity histogram). Finally, these critical features were statistically modeled depending on time via nonparametric and parametric linear and nonlinear models such as those based on logistic functions. Namely, the parametric logistic mixture models provided a way to identify and model the degradation due to biological activity by estimating the corresponding proportion of mass loss. The relation between osteocyte growth and differentiation from stem cells, on the one hand, and collagen degradation, on the other hand, was determined too and modeled through analysis of image objects’ circularity and area, in addition to collagen mass loss. This set of imaging techniques, machine learning procedures, and statistical tools allowed us to characterize and parameterize type I collagen biodegradation when collagen acts as a scaffold in bone regeneration tasks. Namely, the parametric logistic mixture models provided a way to identify and model the degradation due to biological activity and thus to estimate the corresponding proportion of mass loss. Moreover, the proposed methodology can help to estimate the degradation degree of scaffolds from the information obtained by optical microscopy.

Introduction Currently, the analysis of images of cell growth and differentiation from one type of lineage to another is to a great extent qualitative ( Basiji et al. , 2007 ). This method of analysis, which is based on observation of images, does not yield robust quantification of the changes produced during the cell growth and differentiation. The qualitative analysis is currently performed on images obtained by biomedical imaging techniques such as transmission electron microscopy and scanning electron microscopy. This analysis is often based on examining (in a subjective way) the electrodensity of related objects that appear in the obtained images (cells, cell nuclei, cytoplasm, biomaterials such as type I collagen, and extracellular material, among others) ( Sanjurjo-Rodríguez et al. , 2017 ; Sanjurjo-Rodríguez et al. , 2014 ; Sanjurjo-Rodríguez et al. , 2016 ; Sanjurjo-Rodríguez et al. , 2013 ; Martínez-Sánchez et al. , 2013 ; Gashti et al. , 2012 ). Therefore, the information obtained with this type of microanalytical techniques is largely of a qualitative nature (elements observed in a certain area of the image are analyzed by specialized personnel) ( Gashti et al. , 2012 ). Specifically, such studies determine how electrodense the particles and/or areas contained in the image are ( Sanjurjo-Rodríguez et al. , 2014 ). Implementation of a quantitative analysis is also necessary, to evaluate the observable changes in a more general, reproducible, and reliable way ( Wootton et al. , 1995 ; Appel et al. , 2013 ; Ong, Jin & Sinniah, 1996 ; Tayebi et al. , 2012 ; Teverovskiy et al. , 2004 ; Han et al. , 2012 ; Appel et al. , 2013 ; Tabesh et al. , 2007 ), independently of the imaging technique that is used (e. g. , ultrasonography, photoacoustic microscopy, magnetic resonance imaging, optical imaging, X-ray imaging, and nuclear magnetic resonance imaging) ( Appel et al. , 2013 ; Nam et al. , 2014 ; Tayebi et al. , 2012 ). A method based solely on the criterion of the observer, even an expert observer, may be not able to quantitatively verify the level of reliability of the information present in the image. This situation poses a risk of wrong and skewed decisions and conclusions based on the analyzed results. In addition, these tasks are often difficult (requiring skilled personnel) and highly time consuming. Thus, the implementation of image segmentation techniques and statistical analysis of the image information that automatize this process are necessary ( Han et al. , 2012 ; Appel et al. , 2013 ; Tarrío-Saavedra et al. , 2017 ). Nowadays, there are techniques related to image analysis that allow researchers to overcome the limitations of qualitative analysis. In this field, segmentation of images is a fundamental technique for identification of objects, such as cells in tissues. Among the methodologies of image segmentation, there are the thresholding method, clustering approach, region-based approaches, edge detection approaches, and others ( Patil & Deore, 2013 ). The techniques based on artificial intelligence also allow for automation of processes and for improving decision making based on the components and objects of each image. Thus, this approach constitutes a new image-diagnostic technology capable of quantifying the changes related to the biological activity shown in images. That is, the techniques of image analysis become scientific support during analysis of images of cellular tissues and their components. The imaging techniques facilitate the task of technical specialists when they deal with the analysis of observable biological changes in the images of cells in tissues. Image segmentation Within this framework, image segmentation techniques provide information about the number of cells and size, shape, and extension of the tissue in each micrograph. The study of the number, shape, and extension of the cells as a function of time provides valuable information about the degree and rate of degradation of scaffolds such as collagen for the culture of bone cells. Application of statistical regression models to these variables can extract information about the mechanisms of biological degradation of the material, thus predicting the level of degradation due to biological activity and, as a result, making decisions and creating selection criteria ( Tarrío-Saavedra et al. , 2017 ; Janeiro-Arocas et al. , 2016 ). In the research on tissues, the image analysis is a complex task that requires—in addition to broad knowledge of analytical techniques—deep knowledge about the problem at hand ( Teverovskiy et al. , 2004 ; Suvarna, Layton & Bancroft, 2018 ; Bozzola & Russell, 1999 ). Namely, tasks such as differentiation between an extracellular matrix and scaffold (biomaterial) require well -trained and experienced personnel ( El-Jawhari et al. , 2016 ; Sanjurjo-Rodríguez et al. , 2016 ). Modern imaging techniques such as those based on optical and digital microscopy, scanning electron microscopy, and flow cytometry, among others, involve a long process for which the knowledgeability of technicians and researchers is crucial ( Appel et al. , 2013 ; Nam et al. , 2014 ; Gashti et al. , 2012 ; Tayebi et al. , 2012 ). In addition, the analysis performed to identify the processes in which certain types of cells grow and differentiate into other lineages is generally based on the subjective opinion of the analyst. Therefore, as indicated by Basiji et al. (2007), there is a lack of morphological studies that support these analyses. In fact, as Meijering (2012) suggests, the main problem of image analysis is image segmentation, one reason being that the type and quality of the acquired images have a strong influence on the success of cell segmentation (correct identification of objects) ( Kasprowicz, Suman & OToole, 2017 ). On the other hand, the choice of an appropriate segmentation procedure and its parameters also depends on the work of personnel specialized in the concrete problem that needs to be solved. Among all the techniques of image segmentation, there are several approaches such as thresholding and its adaptive variants ( Zhao et al. , 2014 ; Cheng et al. , 2013 ), approximations based on clusters ( Wang & Pan, 2014 ; Gong et al. , 2013 ), and certain mixed models ( Cheng et al. , 2013 ), and other alternatives. Statistical modeling The statistical techniques applied to the extracted vector of features after segmentation are another key topic in this domain. Now, the statistical analysis performed on two-dimensional (2D) and 3D images is often based on methods mainly focused on a descriptive analysis of the data. This exploratory analysis often includes calculation based on measurements of a position (mean) and dispersion (standard deviation) in addition to unsupervised classification methodologies for identifying groups of objects (e. g. , cells, membranes, extracellular material, and scaffold biomaterial) segmented in the images. At a population level, the exploratory data analysis does not provide information about the relation between cell growth and differentiation from stem cells or about degradation of the bioscaffold. Therefore, there is a lack of application of inferential statistical models. These could identify and explain possible correlations between, on the one hand, predictors that account for the cell growth and differentiation and, on the other hand, the degradation level of biomaterials. The statistical study of scaffold degradation under the action of cell growth is necessary to choose a more adequate biomaterial for each application. In fact, there are works involving applied mathematics and statistical techniques that estimate the degradation level of biomaterials depending on critical-for-degradation variables such as the type of biomaterial, cell group, culture medium, cell growth duration, and time of degradation of the biomaterial ( Chen, Zhou & Li, 2011 ; Hoque, Yong & Ian, 2012 ; Pitt & Zhong-wei, 1987 ; Sandino, Planell & Lacroix, 2008 ). Chen, Zhou & Li (2011) developed a numerical model taking into account stochastic hydrolysis and mass transport to simulate a process of degradation of biomaterials and their erosion. Hoque, Yong & Ian (2012) modeled the mass loss using an exponential expression, assuming that the diffusion of water and hydrolysis are the main causes of the biomaterial degradation processes. For our purposes, tools of statistical learning (the field dealing with the interrelation of statistics and computing) for complex data have been applied to model the trends of degradation corresponding to the materials being analyzed ( Friedman, Hastie & Tibshirani, 2008 ). The aforementioned studies show the need for statistical modeling of the mechanical, physical, and rheological phenomena associated with biomaterials such as type I collagen as well as the modeling of degradation to which these biomaterials are subject in cell culture (in addition to the growth and cellular differentiation modeling). In the present work, the data obtained from the segmentation process were analyzed by applying nonparametric and parametric regression models to determine the degree of degradation of type I collagen. The adjusted models were evaluated and compared using criteria of goodness of fit such as the coefficient of determination, R 2, according to the definition proposed by Hayfield & Racine (2008). The computational tool that was used for the model estimation and evaluation is the R statistical software ( R Core Team, 2017 ). Specifically, packages nls2, grofit, and minpack were employed to fit nonlinear regression models, whereas the investr package ( Greenwell & Kabban, 2014 ) were used to obtain an estimate of prediction intervals for the fitted model. In addition, an evolutionary algorithm for global optimization was applied via the DEoptim package to obtain an initial solution for parameters of the adjusted nonlinear models ( Ríos-Fachal et al. , 2014 ). Alternatively, the grofit library could be utilized. It allowed us to fit parametric and nonparametric regression models from an initial solution for the model parameters obtained by means of locally weighted polynomial regression models (LOWESS). Furthermore, the mgcv library was used to evaluate generalized additive models (GAMs) based on the fitting of a basis of penalized regression splines ( Wood, 2006 ), whereas packages ggplot2, ( Wickham, 2009 ), and RColorBrewer were utilized to implement graphics. The scheme of the proposed methodology The main contributions of this work to bioscaffold degradation analysis are summarized below. All of them will be conveniently described throughout this document: • The proposed automatic procedure, based on the random forest supervised classification, to identify different regions of interest (ROIs) in the images obtained by optical microscopy. It is important to emphasize that the random forest classification model is trained taking into account the information provided by qualified laboratory personnel in a qualitative study of this type of images. Therefore, we present a model that is capable of automating a complex and time-consuming task in the context of this type of analysis. • Once the ROIs were identified, i. e. , the extracellular matrix, collagen, cellular nuclei, and background, the present methodology allowed us to separately study their evolution. • The proposed methodology performs extraction of features related to cell growth in the ROIs of each image. • Statistical learning methods such as regression model fitting were employed to explain the evolution of relevant characteristics (of cell growth) and of biodegradation of collagen. • The growth of cells can be characterized by statistical modeling of relevant features such as the total area of cells. These models can help to determine the rate of cell growth, predict cell growth evolution, and to evaluate the use of support materials (in this case collagen) as scaffolds. • Modeling and estimating the relation between the features of cell growth and of differentiation (from stem cells) and the collagen mass loss (the index of material biodegradation level) were performed too. • A method for estimating the degree of degradation (collagen mass loss) from the information in micrographs is proposed as well. Materials and Methods In this work, we studied the degradation of a biomaterial, namely type I collagen, as a result of cellular activity. Biomaterials are considered mechanically functional and physiologically acceptable products used to safely replicate the function of living tissues in biological systems and are implanted temporarily or permanently into the body. The goal is to restore an existing function and, in some cases, to regenerate tissues ( Alfaro & Fernández, 2011 ). Within this framework, because of high biocompatibility, type I collagen is regarded as the gold standard in tissue engineering ( Silvipriya et al. , 2015 ) owing to its biocompatibility. Collagen is a biomaterial (the most common protein in the body) that can act as a scaffold for regeneration of bone and cartilage tissues, among other applications in tissue engineering, e. g. , injectable matrices and scaffolds intended for bone regeneration ( Sanjurjo-Rodríguez et al. , 2017 ; Geiger, Li & Friess, 2003 ; Behring et al. , 2008 ; Inzana et al. , 2014 ; O’brien, 2011 ; Silvipriya et al. , 2015 ). Indeed, it has a high potential for cultivation of cells for producing bone because, among other reasons, it is one of the two main components of bone (along with hydroxyapatite), accounting for 89% of its organic matrix and 32% of its volumetric composition ( O’brien, 2011 ). To improve its performance as a scaffold, collagen can be employed in combination with other types of substrates such as synthetic materials or bovine bone ( El-Jawhari et al. , 2016 ). Human mesenchymal stem cells 3A6 (CMMh-3A6), a line of immortalized mesenchymal stem cells, were provided by Stem Cell, Department of Medical Research & Education and Orthopedics & Traumatology, Veterans General Hospital, Taipei, Taiwan. A total of 4, 200, 000 CMMh-3A6 cells of passage × + were cultured. Two changes of the culture medium and one subculturing (passage 2 or reseeding) were carried out per week because high confluence (90%) can be reached within a relatively short period (approximately between 3 and 4 days) in the plates of an inverted microscope. The biomaterial (type I collagen, commercial product) was prepared to obtain disc-shaped cuts (with a diameter of 8 mm) performed using a biopsy punch. Type I collagen sponges were arranged in three groups, each consisting of 21 samples. For technical reasons, we call groups with the labels CCO, CCT, and CO groups #1, #2, and #3, respectively. The 21 samples in group 1 (CCO) were made up of type I collagen, CMMh-3A6 cells, and a commercial osteogenic culture medium, namely the hMSC Osteogenic Differentiation BulletKit™ Medium (Lonza, Spain). Group #2 (CCT) was composed of 21 samples consisting of Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 1 g/L D-glucose and pyruvate (Gibco, USA), 5% of GlutaMax (Gibco), and 10% of fetal bovine serum (Gibco) as well as CMMh-3A6 cells seeded on the biomaterial (type I collagen). Finally, group #3 (CO) was composed of 21 samples consisting of type I collagen and the commercial osteogenic culture medium (hMSC Osteogenic Differentiation BulletKit™ Medium; Lonza), but without cells. That is, the third group was the control group. The experiment was conducted for 44 days, under conditions that mimic human organism conditions (in a culture oven): pH = 7. 4, temperature = 37 °C, and 5% CO 2. Then, histomorphological analysis was performed on the samples, and they were embedded in paraffin, deparaffinized, and stained with hematoxylin and eosin. After that, 2D images were captured by optical microscopy at 4 × magnification. Overall, 60 micrographs were obtained, which were segmented by machine learning algorithms for each analyzed group, thus resulting in binary images. The analyzed parameters that were taken into account for the classification were type I collagen, the extracellular matrix, nuclei, and image background. It is necessary to extract relevant features from images to train the computational system. These attributes constitute the basis for the learning process. Then, a set of binarized images is extracted (one for each segmentable attribute), from which relevant information for the calculation of these features is also extracted. Proposed methodology This work focuses on the modeling of the degree of degradation of type I collagen and cell growth by segmentation analysis (identification and separation of objects) of images. These images correspond to cultures of bone cells and are obtained by optical microscopy, after staining of the cells with hematoxylin and eosin. The process map in Fig. 1 shows the proposed methodology that was developed here to characterize and model collagen degradation and cell growth. It includes several stages that are roughly outlined below. 10. 7717/peerj. 7233/fig-1 Figure 1 A general outline of the proposed modeling of cellular growth and the degree of degradation of type I collagen. To study the biodegradation of collagen as a scaffold for regeneration of bone tissue, CMMh-3A6 stem cells were seeded to differentiate them into osteocytes. Then, images were captured by optical microscopy after staining with hematoxylin and eosin. The images were obtained as a time series within short periods, 3 to 4 days. Based on the expertise of a team of researchers in the biomedical area, an ROI was defined on which it was desirable to perform the analysis via the extraction of various characteristics (e. g. , the area, number of particles, and ratios). These are the extracellular matrix, background, nuclei, and type I collagen. Once the ROIs were defined, the feature extraction from the images was carried out. These features contain information about the shape, size, and extension of the cells, as related to the progression of collagen degradation. At this stage, the tensor structure and entropy ( Sahoo, Soltani & Wong, 1988 ) corresponding to the pixels related to different ROIs of each image were obtained. With these attributes, a training sample was built, when we knew the actual type of ROI corresponding to each pixel. This sample was used later to perform the segmentation of the images. After obtaining the training sample, we proceed to apply supervised classification algorithms based on machine learning such as the random forest classifiers ( Breiman, 2001 ). To achieve this goal, first, the classes of ROIs were determined. Via the identification of these classes, we estimated a model on the basis of the training sample by associating the ROI of each pixel with its corresponding feature vector (composed of entropy and energy values). Next, the models (based on random forest) assigned the ROI class (resulting from generation of probability maps ( Arganda-Carreras et al. , 2017 )) to each pixel of the test sample (the sample in which we did not know the actual ROI). Once the different classes were identified within each image, and the probability maps were built, we proceed to extract the representative characteristics related to the cellular activity. These features are the area, major and minor axes of a fitted ellipse, circularity, and histogram intensity mode, among other features corresponding to the objects of each image. The extracted characteristics were selected taking into consideration that they provide relevant information about the cellular activity and, by extension, biodegradation of the support material or scaffolding, in this case, type I collagen. Software tools ImageJ ( Schneider, Rasband & Eliceiri, 2012 ; Schindelin et al. , 2015 ) and Waikato Environment for Knowledge Analysis ( Weka ) ( Hall et al. , 2009 ) were used to perform the segmentation, classification, and feature extraction tasks. On the one hand, ImageJ was employed to extract the image attributes in order to apply thresholding processes and to generate the corpus of data related to each ROI. On the other hand, Weka was applied to training of the random forest classifier that provided information about the identification and categorization of different ROIs. Finally, a statistical learning-model approach based on parametric (linear and nonlinear) and nonparametric regression models was applied to estimate the value of relevant characteristics of the biological activity extracted as a function of time. These estimates are an indicator of the degree of collagen degradation. These estimated models provide information about the type and degree of dependence between variables. They also offer prediction of the degradation degree from image features. More information about texture analysis of original micrographs, the segmentation process via the random forest classifier, and application of statistical model ing of the scaffold degradation degree is provided in the text below. Texture analysis The approach proposed in this study is to quantitatively determine cell growth and differentiation and the degree of degradation of type I collagen due to a biological activity. The first step for achieving these objectives is to associate a numerical vector of features with each pixel. In fact, once the different ROIs were identified by expert personnel, readers can see that the different regions can be differentiated according to the texture of the image. For this purpose, scale-invariant texture filters were applied to 2D images by means of the ImageJ software. Image texture can be defined as a spatial arrangement of the color or intensities of an image in a given region, whereas texture analysis is often performed to separate different regions of an image. Therefore, in this paper, we propose to apply an image texture analysis consisting of extraction of representative features, such as a structure tensor and entropy, to identify different ROIs. Once the characteristic vectors are extracted, they will serve as a training sample for the classifier, in this case, the random forest algorithm. 1. Structure tensor: This is a matrix representation of the image’s partial derivatives defined as second-order symmetric positive matrix J: (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}J= \left[ \begin{array}{@{}cc@{}} \displaystyle \left\langle {f}_{x}, {f}_{x} \right\rangle w&\displaystyle \left\langle {f}_{x}, {f}_{y} \right\rangle w\\ \displaystyle \left\langle {f}_{x}, {f}_{y} \right\rangle w&\displaystyle \left\langle {f}_{y}, {f}_{y} \right\rangle w \end{array} \right] \end{eqnarray*}\end{document} J = f x, f x w f x, f y w f x, f y w f y, f y w where f x and f y are images of the partial spatial derivatives: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$ \frac{\partial f}{\partial x} $\end{document} ∂ f ∂ x and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$ \frac{\partial f}{\partial y} $\end{document} ∂ f ∂ y, respectively ( Budde & Frank, 2012 ). From this matrix, all major and minor eigenvalues are separated for each pixel and channel in the image: (2) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}T(v)=\lambda v\end{eqnarray*}\end{document} T v = λ v where λ is the eigen values, and v denotes the eigen vectors. 2. Entropy: For its calculation, a circle of radius r is drawn around each pixel. The image intensity histogram corresponding to the separated circle is obtained as binarized image fragments. Finally, entropy is calculated for each particle, where p is the probability of each chunk in the histogram corresponding to each channel of the image, both in RGB and in Hue, Saturation Brightness formats. The random forest classifier for identifying different ROIs It is possible to apply a wide variety of supervised classification models based on the descriptors that were extracted from the images in order to identify different ROIs. For the specific case of this research, a random-forest–type classifier ( Criminisi, Shotton & Konukoglu, 2012 ) was chosen. It has the advantages of operating on attributes (or modular characteristics), of preventing overfitting of certain classes, and an optimized computational cost. In recent years, decision trees proved to be some of the most promising techniques in machine learning, computer vision, and medical analysis ( Criminisi, Shotton & Konukoglu, 2012 ). Random forest classifiers operate by constructing several decision trees (predictive processes that map observations of an item to conclusions about the objective value of the item) in the training phase, to then produce a class fashion (by its nature as a classifier) for each tree. Image segmentation Now, the benefits of the technology based on artificial vision allow us to determine the biological behavior of cells such as cell growth and differentiation as well as degradation of biomaterials such as type I collagen. Image segmentation is a reference within the support techniques for achieving this goal. In general, image segmentation is defined as the process of dividing an image into different segments or groups of pixels that share certain common characteristics ( Mallik et al. , 2011 ; Janeiro-Arocas et al. , 2016 ). This is a fundamental task in image analysis for the detection of objects ( Mallik et al. , 2011 ; Janeiro-Arocas et al. , 2016 ). In fact, it is a central problem in many studies dealing with image analysis ( Meijering, 2012 ). It is important to note that the type and quality of the acquired images influence the success of cell segmentation (identification and separation of objects) ( Kasprowicz, Suman & OToole, 2017 ). Therefore, adequate quality of images is one of the existing requirements of this set of techniques. In fact, an important feature of image segmentation is that although it is conceptually simple, it lacks generality; consequently, it cannot be implemented reliably and easily for all cell lines, image modalities, and cell densities without preprocessing of images ( Alanazi et al. , 2017 ). Limiting factors for the image segmentation process are, e. g. , the type of objects to be analyzed in the image, the aims pursued by the research team, and limitations of the knowledge of the technician in charge of the segmentation process. All these factors necessitate devising specific strategies for the processing and analysis of images for each particular problem. Despite the absence of a universal segmentation procedure ( Alanazi et al. , 2017 ), today it is possible to analyze 2D images of the behavior of stem cells in vivo, with information on sequential growth and differentiation as a function of time, via a set of different segmentation techniques. Among the most popular segmentation and image-processing techniques are thresholding, region-growing methods, edge detection, and Markov random field color algorithms ( Grys et al. , 2017 ). In this work, we propose a segmentation method based on a random forest classifier trained on vectors of attributes and on properties obtained from filters that are applied to labeled ROIs in the images. In fact, different studies have revealed the great potential of the random forest method as a support tool in the segmentation process. Thus, Schroff, Criminisi & Zisserman (2008) demonstrated the potential of random forest as a technique for segmentation processes of objects present in images. They worked with the MSRC image dataset, incorporating both features that describe objects locally and those that characterize the environment of those objects. In the medical domain, Dhungel, Carneiro & Bradley (2015) identified formation of unusual masses in mammograms. These masses varied in size, texture, and the area that delimits them, merging with the rest of the tissues of the breast. The classifiers based on deep belief networks (first filtering in which suspicious areas are located) and random forest (last filtering) were applied in a cascade. This approach significantly reduced the number of false positives. Furthermore, Khan, Hanbury & Stoettinger (2010) proved that the random forest technique is highly effective at segmenting human skin images. Those authors used random forest in combination with Improved Hue, Luminance and Saturation color space and compared this classifier with others based on Bayesian networks, multilayer perceptron, support vector machine, AdaBoost, naïve Bayes, and Radial Basis Function neural networks. The classifier surpassed all other techniques, yielding 0. 877, 0. 738, and 0. 740 in terms of accuracy, precision, and recall, respectively. Finally, Ghose et al. (2012) demonstrated the effectiveness of segmenting magnetic resonance images of human prostates. To this end, those authors propose to use a framework of decision trees in order to obtain a probabilistic representation of those voxels that define the prostate. The supervised classification process was applied by means of the Weka software ( Hall et al. , 2009 ). This computational tool allows for the application of a wide range of techniques from the field of data mining, e. g. , data visualization techniques, feature selection, unsupervised classification or clustering algorithms, supervised classification such as random forest, and regression models ( Frank et al. , 2004 ). Regression model fitting Once the vectors of features related to the growth and cellular differentiation of osteocytes were obtained through the process of image segmentation (mentioned above), the degree of degradation of type I collagen was estimated via application of statistical regression models. In addition, it is important to note that the extracellular matrix ROI is extracted from the image and its features. They are the area, objects’ circularity (4 πarea ∕ perimeter 2 ), and the intensity histogram mode, among others that are modeled as a function of time to estimate cellular growth. The relation between collagen mass loss and the area or circularity was studied here to model the degree of degradation of collagen as well as the biological activity. Two main regression approaches were employed: nonparametric and parametric nonlinear regression. In the text below, the two aforementioned approaches are briefly introduced. Nonparametric methods have attracted the attention of academia and industry ( Hayfield & Racine, 2008 ). They have relevant advantages, e. g. , they do not assume any parametric function that constrains the relation between variables or even a specific distribution of those variables of interest ( Tarrío-Saavedra et al. , 2011 ). Indeed, nonparametric methods provide estimates of flexible models that do not impose any prespecified function ( Maity, 2017 ), thereby allowing investigators to model a wide range of possible complex nonlinear functions ( Shokrzadeh, Jozani & Bibeau, 2014 ). There are many types of nonparametric models dedicated to regression tasks, for example, kernel smoothing ( Wand & Jones, 1994 ), local polynomial regression ( Breidt & Opsomer, 2000 ), regression splines ( Wood, 2006 ; Mammen & Van de Geer, 1997 ), and LOWESS ( Cleveland & Devlin, 1988 ). Specifically, penalized regression spline modeling within the framework of nonparametric GAMs ( Wood, 2006 ) is proposed here to obtain information about the type of dependence between collagen mass loss and time. For the sake of simplicity, a penalized B-spline basis was fitted. A generic GAM as a function of two linear predictor variables ( X 1 and X 2 ) and two smooth predictors ( T 1 and T 2 ) can be defined as follows: (3) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}Y={\beta }_{0}+{\beta }_{1}{X}_{1}+{\beta }_{2}{X}_{2}+{s}_{1}({T}_{1})+{s}_{2}({T}_{2})+{s}_{12}({T}_{1}, {T}_{2})+\epsilon \end{eqnarray*}\end{document} Y = β 0 + β 1 X 1 + β 2 X 2 + s 1 T 1 + s 2 T 2 + s 12 T 1, T 2 + ϵ where the effects on the response Y of X 1 and X 2 are linear, whereas the effects of T 1 and T 2 are only assumed to be smooth. This model additionally allowed us to include linear or smooth effects for the interaction between variables, as is the case for S ( T 1, T 2 ). The discrepancy between Y and Y estimates, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\hat {Y}$\end{document} Y ˆ, was measured by the model residuals, ϵ. When the Y variable is discrete, e. g. , binomially or Poisson distributed, Y is replaced by g ( η ). In the present case, Y = MassLoss and X = Time ; thus, (4) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\hat {Y}=E(Mass loss{|}Time)={\beta }_{0}+{s}_{Time}(Time)\end{eqnarray*}\end{document} Y ˆ = E M a s s l o s s | T i m e = β 0 + s T i m e T i m e As mentioned above, package mgcv provides computational tools to fit GAMs. This library uses penalized regression splines for the estimation of smooth effect s of each predictor variable x. To define the concept of splines, consider the set of nodes a < t 1 < t 2 < ⋯ < t L < b in interval [ a, b ]. A spline function is a polynomial function in each subinterval of a certain degree (e. g. , three in the case of a cubic spline), which is continuous in the knots up to order degree L − 1. With the aim of determining a spline function, we must know L + degree + 1 values, known as the degrees of freedom of the spline family defined for the set of knots. The degree + 1 number is known as spline order. Thus, we can write the smooth effect as (5) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}{s}_{j}(x)=\sum _{k=1}^{L+degree}{\beta }_{jk}{\phi }_{jk}(x)={\beta }_{j}^{{^{\prime}}}{\phi }_{j}\end{eqnarray*}\end{document} s j x = ∑ k = 1 L + d e g r e e β j k ϕ j k x = β j ′ ϕ j for the basis composed of L + degree functions ϕ jk that define the B-splines, where L is the number of knots (10 by default) and degree = 3 (cubic regression splines). Given the s j definition, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\hat {Y}=E(Mass loss{|}Time)$\end{document} Y ˆ = E M a s s l o s s | T i m e is now a parametric mean function with parameters \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\beta =({\beta }_{0}, {\beta }_{Time}^{{^{\prime}}})^{{^{\prime}}}$\end{document} β = β 0, β T i m e ′ ′ and predictor Time that define the intercept and the splines that define s j. The penalized least squares objective function for estimating β is (6) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\parallel Y-X\beta {\parallel }^{2}+\sum _{j}{\lambda }_{j}{\beta }_{j}^{{^{\prime}}}{B}_{j}{\beta }_{j}\end{eqnarray*}\end{document} ∥ Y − X β ∥ 2 + ∑ j λ j β j ′ B j β j The values of λ j are selected by an iterative algorithm to minimize a generalized cross-validation criterion. Parametric models provide the transfer function that determines the dependence relation between critical variables of a process. In addition, the constitutive parameters usually have a physical meaning. Thus, the chemical, physical, and biological processes of materials and substances can be characterized and compared by studying the parameter values. To illustrate a parametric-model expression in a simple way, we can define (7) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}Y=m \left( {t}_{i} \right) +{\varepsilon }_{i}, \mathrm{with} i=1, 2, \ldots, n, \mathrm{and} \mathrm{E} \left( {\varepsilon }_{i} \right) =0\end{eqnarray*}\end{document} Y = m t i + ε i, with i = 1, 2, …, n, and E ε i = 0 where Y is the response variable to be modeled, and m is the regression function fitted to Y. The latter can be the mass loss, area, and circularity, among other features critical for collagen degradation and cell growth. Finally, ε i are the independent, identically distributed model residuals. In a fixed design, 0 ≤t 1 ≤t 2 ≤… ≤t n, the parametric functions can be defined as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\mathcal{M}= \left\{ {\mathrm{m}}_{\mathrm{\theta }} / \mathrm{\theta }\in \circleddash \right\} $\end{document} M = m θ ∕ θ ∈ ⊝, where θ is the parameter vector that defines the regression model, and ⊝ is a subset of ℝ k. The linear models are the simplest and most typical parametric models, but in chemistry and biology domains, nonlinear models are also important and useful. In this regard, there are many processes characterized by exponential and sigmoid types of relations between critical variables such as material degradation ( Robles-Bykbaev et al. , 2018 ; Ríos-Fachal et al. , 2014 ), crystallization ( López-Beceiro, Gracia-Fernández & Artiaga, 2013 ), oxidation ( Tarrío-Saavedra et al. , 2013 ), and material dimensional variation due to magnetic changes ( Tarrío-Saavedra et al. , 2017 ). In fact, the requirements of biological growth have popularized sigmoid monotonic nonlinear functions, characterized by exponential growth and a point of inflection from which the speed of growth gradually decreases until reaching a saturation zone ( Kahm et al. , 2010 ; Román-Román & Torres-Ruiz, 2012 ; Kingsland, 1985 ; Tarrío-Saavedra et al. , 2014 ). Some of the most popular functions are those based on logistic and Gompertz functions: 1. Gompertz function: (8) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}y \left( t \right) =A\exp \nolimits \left[ -\exp \nolimits \left( \frac{\mu e}{A} \left( \mathrm{\lambda }-t \right) +1 \right) \right] \end{eqnarray*}\end{document} y t = A exp − exp μ e A λ − t + 1 2. Logistic function: (9) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}y \left( t \right) = \frac{A}{1+\exp \nolimits \left( \frac{\mu -t}{scal} \right) } \end{eqnarray*}\end{document} y t = A 1 + exp μ − t s c a l where A is the asymptote, µ represents the inflection point or time corresponding to the maximum rate of change, λ is the delay until the beginning of exponential growth, and scal or the scale parameter accounts for the growth (or decay) rate of the function. The logistic function is used to explain a great number of real processes related to biology, chemistry, physics, informetrics, and economy, among other fields, taking into account that it is their underlying model ( Limoges, 1987 ; Román-Román & Torres-Ruiz, 2012 ). Furthermore, nonlinear mixture models are created by the addition of two or more nonlinear functions. They serve for identifying, separating, and estimating overlapping processes often accompanying further kinetic analysis ( Francisco-Fernández et al. , 2012 ; Sánchez-Jiménez et al. , 2013 ; López-Beceiro et al. , 2010 ; López-Beceiro et al. , 2012 ). This is the case for the mixture logistic model: (10) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}m \left( \theta ;t \right) =\hat {Y}(t)= \sum _{i=1}^{k} \frac{A}{1+\exp \nolimits \left( \frac{{\mathrm{\mu }}_{i}-t}{{scal}_{i}} \right) }, \end{eqnarray*}\end{document} m θ ; t = Y ˆ t = ∑ i = 1 k A 1 + exp μ i − t s c a l i, where k denotes the number of logistic components, t is usually time or even temperature, and θ is the vector of parameters θ = ( A i, μ i, scal i ). Logistic mixture models have been successfully applied in many studies on overlapping processes (e. g. , degradation and crystallization processes) of different materials such as wood, polymers, composites, ceramics, and metal-organic frameworks ( Rios-Fachal et al. , 2013 ; López-Beceiro et al. , 2011 ; López-Beceiro et al. , 2010 ; Pato-Doldán et al. , 2012 ; Francisco-Fernández et al. , 2012 ). Results Once the images were obtained, four ROI classes were identified based on the information that the photographs provided about type I collagen and the extracellular matrix: type I collagen, extracellular matrix, background, and nuclei. The next step was the automatic classification of each pixel into one of the four groups or ROIs identified in the image. This approach allowed us to study the evolution of the collagen mass or extracellular matrix independently, preventing the analysis errors related to the inclusion of information corresponding to the background or artifacts (representation errors). As indicated in Section 3, the four ROIs present can be distinguished from one another according to the texture of the image. Therefore, a texture analysis was carried out in which each pixel was defined by a feature vector formed by the parameters of the structure tensor and entropy. Once each area of the images of the training sample was defined by a feature vector, a random forest supervised classification model was developed (trained) to identify the ROI corresponding to each pixel of all the images being studied. Once the results of the classifier were validated, probability maps were generated for each of the 60 images and for each above-mentioned class. Each map, estimated for each ROI, contained information on the probability that each pixel belongs to the layer that represents the corresponding map. This section shows the main results of applying the proposed methodology to model collagen biodegradation. In accordance with the different steps of the proposed procedure, it was divided in two subsections: image analysis with feature extraction and statistical modeling. Image analysis and feature extraction The probability maps corresponding to each ROI were applied as masks to the original images in order to quantify the type I collagen data extracted from each image and cellular growth and differentiation. New datasets composed of features extracted from the image (mean area in μm 2, mean circularity, the major axis of ellipses, and the mode of the greyscale histogram) were extracted from the ROIs of images corresponding to each period (measured in days). In the case of the type I collagen group, it was observed that the probability matrices obtained by applying the random forest classifier decrease as a function of time when the ROI corresponding to type I collagen was studied, whereas those corresponding to the ROI of the extracellular matrix increase as a function of time, as expected due to osteocyte growth. These findings support the use of ROI features such as mean area for characterizing the collagen degradation due to cellular activity. Figures 2 – 3 show the original micrographs and those obtained by the application of segmentation techniques. Segmentation algorithms based on detection of a threshold value and random forest machine learning algorithms were used. The segmentation of images by threshold ing, based on image conversion to the grayscale, is aimed at finding the ideal values that separate different ROIs by filtering regions of the histogram of a black-and-white image. The values of the filters can be found using local minima of the histogram ( Bulgarevich et al. , 2018 ). Regarding application of the random forest classifier, changes in the area and shape of objects could be observed, based on the analysis of extracellular-matrix ROIs (see Figs. 2 – 3 ). The area of the extracellular matrix seems to increase with time (violet or white area). Moreover, the region where the extracellular matrix is placed (collagen scaffold) seems to decrease because the mass of the collagen scaffold decreases with time. The extracellular matrix becomes more concentrated with time. Consequently, the degradation of type I collagen due to cellular activity and cellular growth can be studied through representative features of these images. In addition, cellular growth and the degradation degree of collagen under the influence of cellular activity are estimated by nonparametric and parametric statistical models. 10. 7717/peerj. 7233/fig-2 Figure 2 Images corresponding to cell culture samples of the CCO group: collagen, stem cells (CMMh-3A6), and the osteogenic culture medium. The original (A, D, G), segmented images by thresholding (B, E, H), and segmented images by the random forest classifier (C, F, I) are presented. Each row refers to the time points at which the images were captured: T0 (the start of the experiment), T21 (the middle point of the experiment, 21 days), and T44 (the end of the experiment, 44 days). To justify the use of random forest instead of other techniques such as thresholding, examples of image segmentation with two images are shown below. In Figs. 4A and 4D, two images obtained by optical microscopy are depicted, corresponding to the 18th and 38th days of observation from the beginning of the experiment, respectively. Figs. 4B and 4E present application of the threshold segmentation technique to the images corresponding to the 18th and 38th days, respectively. In addition, Figs. 4C and 4F correspond to segmentation of the images in Figs. 4A and 4D, respectively, by the random forest method. All the images are from the CCO group. 10. 7717/peerj. 7233/fig-3 Figure 3 Images corresponding to cell culture samples of the CCT group: collagen, stem cells (CMMh-3A6), and the nonosteogenic culture medium. The original (A, D, G), segmented images by thresholding (B, E, H), and segmented images by the random forest classifier (C, F, I) are presented. Each row refers to the time points at which the images were captured: T0 (the start of the experiment), T21 (the middle point of the experiment, 21 days), and T44 (the end of the experiment, 44 days). 10. 7717/peerj. 7233/fig-4 Figure 4 (A) An image obtained by optical microscopy and belonging to the CCO group (the osteogenic medium with cells). It was obtained at time = 18 days, that is, it corresponds to an early stage, where collagen is mostly found. (B) An image of collagen resulting from the application of segmentation based on the classical threshold (Threshold) to the image of (A). The area of collagen obtained with this technique is 178965 (189 pixels represent the 300 micrometers seen in the image). (C) An image of collagen resulting from the application of segmentation based on random forest to the image of (A). The area of collagen obtained with this technique is 175261 (189 pixels represent the 300 micrometers seen in the image). (D) An image obtained by optical microscopy and belonging to the CCO group (the osteogenic medium with cells). It was obtained at time = 38 days, that is, it corresponds to a stage where a large quantity of extracellular mass and nuclei is found, in addition to a lesser quantity of collagen. (E) A collagen image resulting from the application of segmentation based on the classical threshold (Threshold) to the image of (D). The area of collagen obtained with this technique is 384170 (189 pixels represent the 300 micrometers seen in the image). (F) A collagen image resulting from the application of segmentation based on random forest to the image of (A). The area of collagen obtained with this technique is 193125 (189 pixels represent the 300 micrometers seen in the image). When the random forest and thresholding techniques were compared, we noted that there was a significant loss of data ( Fig. 4E ) with respect to the actual collagen shown in the image obtained from the laboratory ( Fig. 4D ). In fact, the image of Fig. 4D is much more difficult to segment because it contains subtle variations of intensity (color) in collagen; these variations can be detected only by the random forest technique as shown in Fig. 4F. In addition, the segmentation technique based on random forest can be more precise, given that it allows us to obtain a collagen area that is much closer to reality. In fact, the thresholding technique also regards the nuclei and some of the extracellular matrix as a part of collagen. Therefore, the collagen area estimates by the Threshold technique are higher than what they should be. Tenfold cross-validation techniques were applied too to measure the performance of the random forest classifier on the segmentation task in the present case study. Table 1 shows the measurements of goodness of classification of each ROI from the images in Fig. 4D. Considering only the overall precision (proportion of pixels correctly classified), 97. 3% of pixels were correctly classified. This good performance is supported by Table 2, which presents the confusion matrix corresponding to application of the random forest classifier to the image in Fig. 4D. The number of pixels correctly classified was much higher than that of incorrectly identified pixels for each ROI. Figure 5 compares the threshold and random forest segmentation methods for COO and CCT groups in the range “days 0–44”. The misclassification error (square root of the number of misclassified squared pixels of each image) of the threshold method tend to be greater than the corresponding to the random forest. This finding supports the use of random forest in this application. Data, random forest parameters, segmented images and validation measures of these examples are shown in supplementary material (see Dataset S1 ). Two tutorial videos are supplied in order to show the advantages of using the random forest classifier for image segmentation with respect to the threshold segmentation method. They show how to reproduce the segmentation tasks from an optical micrograph (e. g. , the Fig. 4F ) by random forest classifier or threshold method, using ImageJ and Weka software (see the Video S1 and Video S2 ). 10. 7717/peerj. 7233/table-1 Table 1 Measurements of goodness of the ROI classification when the random forest method was applied to Fig. 4D. Detailed accuracy by class TP rate Fp rate Precision Recall F-Measure MCC ROC area PCR area ROI 0. 957 0. 008 0. 983 0. 957 0. 970 0. 957 0. 998 0. 997 Collagen 0. 975 0. 023 0. 947 0. 975 0. 961 0. 944 0. 997 0. 993 Extracell. matrix 0. 914 0. 004 0. 939 0. 914 0. 926 0. 922 0. 998 0. 979 Nuclei 0. 997 0. 002 0. 995 0. 997 0. 996 0. 995 1. 000 1. 000 Background Weighted average 0. 973 0. 010 0. 973 0. 973 0. 973 0. 963 0. 999 0. 996 10. 7717/peerj. 7233/table-2 Table 2 The confusion matrix corresponding to application of the random forest classifier to the image of Fig. 4D. The number of pixels is indicated. Confusion matrix Predicted Collagen Extracell. matrix Nuclei Background Real Collagen 5727 220 25 10 Extracell. matrix 92 5630 36 16 Nuclei 7 90 1073 4 Background 1 7 9 6361 10. 7717/peerj. 7233/fig-5 Figure 5 (A) Square root of the number of misclassified squared pixels for each image (when collagen ROI is analyzed) obtained at each time point and corresponding to the CCO group; (B) square root of the number of misclassified squared pixels for each image (when collagen ROI is analyzed) obtained at each time point and corresponding to the CCT group. As shown in the tutorials, the random forest classifier provides a more accurate identification of the different ROIs in an automatic way. The random forest segmentation method is more useful to identify the different regions of interest, even when the differences between regions are very slight and can be overlapped with the noise of the image. Conversely, the threshold method has some drawbacks: 1. It is necessary to manually adjust the thresholds corresponding to the different ROIs, and 2. this can lead in the losing of important information related to the area and edges of the different ROIs; 3. in addition, threshold method is highly depended on the operator skills (not as automatic as the random forest method). Summarizing, out of two techniques, the random forest classifier is the more automatic and accurate alternative for segmentation task, preventing in a greater extent undesirable effects such as the image noise. Statistical regression modeling The first step is to apply descriptive analysis techniques in order to evaluate the dependence structure between the variables critical for collagen degradation and osteocyte growth. Then, the collagen degradation modeling task was done. Further regression analysis of image feature variables related to cell growth and differentiation is necessary to determine the influence of cell growth on collagen degradation. Finally, the cell growth influence on collagen degradation was modeled via the degree of collagen degradation (measured by the mass loss percentage) and the mean area, mean circularity, and grayscale histogram mode corresponding to each image captured at each time point. Exploratory correlation analysis Identifying and characterizing dependence relations between features extracted from an image, time, and the degree of degradation (collagen mass loss) are the goals here. Scatterplots are the most intuitive way to check the dependence structure of a dataset. Figure 6 illustrates the scatterplot matrix for mean area, mean circularity, collagen mass loss, and time for groups CCO and CCT. Other interesting variables (such as the major axis of the fitted ellipse or even mean roundness of the objects in each image) were not included due to their strong dependence on the area or circularity. The area and circularity of the extracellular matrix provide information about cellular growth and shape. 10. 7717/peerj. 7233/fig-6 Figure 6 The scatterplot matrix corresponding to the mean area, mean circularity, collagen mass loss, and time within the CCO and CCT groups. The strongest dependence was found between the mass loss and time variables. In fact, the collagen mass loss strongly depends on time in a relatively complex nonlinear way. It seems that collagen degradation is composed of two main steps that could be related to two main degradation processes. This trend was observed in both groups CCO and CCT, with only slight differences between them. In addition, an asymptotic type of dependence between circularity and time was observed, whereas mean area seemed to increase with time. This increase seemed to be of the sigmoid type when the CCO group was studied, but in the case of the CCT group, the type of trend was not clear due to high dispersion. If the dependence structure of the features of the CO group is studied and compared with that of groups CCO and CCT (see Fig. 7A ), we note that there are no trends between variables (either linear or nonlinear), and scattering is substantial. This finding is in agreement with the fact that CO is the control group without stem cells. 10. 7717/peerj. 7233/fig-7 Figure 7 (A) The scatterplot matrix corresponding to the mean area, mean circularity, collagen mass loss, and time within groups CO, CCO, and CCT; (B) the scatterplot matrix corresponding to the neperian logarithm of the mean area, neperian logarithm of mean circularity, collagen mass loss, and time within groups CO, CCO, and CCT. Furthermore, Fig. 7B depicts the relations between the variables when a logarithmic transform is applied to the mean area and mean circularity. We see that there could be linear relations between the logarithmic transform of the area and circularity as a function of time. In addition, there could be slightly linear relations of the collagen mass loss with the logarithmic transform of the area and circularity. This information will help to estimate the proper statistical models for the degree of collagen degradation. Collagen degradation modeling Estimating the underlying model that explains the degree of collagen degradation depending on time is the goal here. This model can provide information about the different steps of degradation and can serve for forecasting tasks. The mass loss feature is the critical variable chosen for characterizing the degree of type I collagen degradation. According to the results of the exploratory analysis, the use of a sigmoid-type nonlinear function is needed to describe the relation between mass loss and time. Moreover, it seems that the main trend can be the result of the sum of two sigmoid trends. A nonparametric GAM with penalized regression splines was applied to confirm the results of descriptive analysis. In fact, in Figs. 8A and 8B, the GAM estimates for groups CCO and CCT are shown. Taking into account the fit and bootstrap confidence intervals, we infer that the main trend could be composed of two sigmoid curves, each one corresponding to different degradation processes. 10. 7717/peerj. 7233/fig-8 Figure 8 (A and B) depict the fittings (via a nonparametric GAM) of mass loss as a function of time in groups CCO and CCT, respectively. (C and D) show the fittings of the parametric logistic mixture model to estimate the mass loss with respect to time in groups CCO and CCT, respectively. (E and F) illustrate the two logistic components of the logistic mixture model applied to groups CCO and CCT, respectively. A more accurate fit was obtained for the CCT group and explained 99% of the collagen mass loss information. In any case, it seems that we can obtain accurate estimates of mass loss as a function of time taking into account the high values of the determination coefficient and the relatively narrow confidence intervals. The next step was to fit a parametric model based on the adequate nonlinear function. The aim was to obtain a transfer function that, in addition to interpretable parameters, could provide predictions and allow us to compare different groups. The sigmoid trend and the fact that the mass loss increases monotonically support the use of logistic functions. In accordance with the GAM results, a nonlinear regression model based on a mixture of two logistic functions (defined by three parameters) is proposed. This type of nonlinear regression model has been successfully applied to separate overlapping degradation processes within the framework of thermal analysis in degradation studies ( Tarrío-Saavedra et al. , 2014 ). Figures 8C and 8D shows the logistic-mixture-model fittings that estimate the mass loss depending on time for groups CCO and CCT, respectively. Estimation and prediction asymptotic confidence intervals are provided too. Table 3 shows signification analysis and the model parameters. Signification analysis, fitted trends, confidence intervals, and the achieved goodness of fit ( R 2 = 0. 94 and R 2 = 0. 99 for groups CCO and CCT, respectively) justify the use of the logistic mixed model to explain the mass loss of collagen. The first component is characterized by an inflection point at ∼2 days and a saturation asymptote at 50% of collagen mass loss. On the contrary, the second logistic component is defined at an inflection point of ∼16–17 days and an overall mass loss of between the 21–13 of weight percent loss. The meaning of the two logistic components can be guessed taking into consideration control sample CO. In fact, the samples of the CO group lost mass almost immediately, and the mass remained constant at ∼30% mass loss. Thus, this mass loss is related to another factor apart from cellular activity. Deeper research is necessary to identify the nature of the degradation process (either hydrolytic or otherwise). In addition, we infer that the first logistic component depends, to a large extent, on another degradation factor aside from cellular differentiation and growth. In fact, considering the porosity of collagen, the degradation due to cell growth and differentiation should take place mainly after longer periods. Regarding the second logistic component, it may be more related to cell activity (given that the inflection point is ∼16–17 days) that corresponds to ∼20% collagen mass loss after 28 days. 10. 7717/peerj. 7233/table-3 Table 3 The signification analysis of fitted parameters of the mixed logistic model and bootstrap intervals corresponding to groups CCO and CCT at the 95% confidence level. Parameters Estimates Bootstrap 95% interval lower limit Bootstrap 95% interval upper limit Standard error t value p value CCO A 1 50. 223 42. 841 54. 690 3. 900 12. 88 4e−06 μ 1 1. 7340 1. 0042 2. 4513 0. 481 3. 604 0. 0087 scal 1 1. 0000 0. 4259 1. 6945 0. 446 2. 240 0. 0600 A 2 23. 463 17. 667 34. 728 5. 463 4. 295 0. 0036 μ 2 16. 261 14. 273 18. 803 1. 329 12. 24 5. 6e−6 scal 2 1. 3233 0. 3400 4. 1071 1. 245 1. 063 0. 3230 CCT A 1 52. 502 36. 131 58. 400 7. 699 6. 819 0. 0003 μ 1 2. 4768 2. 0578 3. 0790 0. 370 6. 694 0. 0003 scal 1 1. 1546 0. 5631 1. 6311 0. 337 3. 430 0. 0110 A 2 21. 653 13. 539 55. 246 12. 52 1. 729 0. 1274 μ 2 16. 881 10. 753 26. 856 3. 040 5. 553 0. 0009 scal 2 3. 6554 0. 6866 10. 410 3. 777 0. 968 0. 3654 The progression of degradation of type I collagen can be compared between groups CCO and CCT through examination of the logistic components. Figure 8E shows the fitted first logistic component for groups CCO and CCT. There are only slight differences, but it seems that the degradation process tends to begin even before the inflection point μ CCO (=1. 73 days ) < μ CCT (=2. 47 days ). The results on bootstrap confidence intervals for parameter µ(implemented using the nlstools package ( Baty et al. , 2015 ) with 1, 000 resamplings) suggest that CCO samples tend to begin to degrade before the CCT samples do. There are, however, no differences between the groups in terms of the mass loss percentage and the rate of mass loss for this first degradation process. Furthermore, in Fig. 8F, the second logistic components of groups CCO and CCT are compared. The rate of degradation corresponding to the CCO group is significantly greater than that corresponding to the CCT group. This finding can be interpreted via analysis of the scale parameter. In fact, the bootstrap confidence interval for the scale parameter is placed at lower values than those corresponding to group CCT (at lower values of scale, the logistic curve has a higher rate of change). Therefore, the osteogenic medium slightly promotes the degradation of collagen, advancing the degradation process and increasing the rate of degradation. Cell growth modeling The mean area and mean circularity of the identified objects of the extracellular matrix obtained from each image at each time point provide important information about cell growth and differentiation. Accordingly, the dependence relation of the area and circularity with time was characterized and statistically modeled. Figure 9 shows the best models fitting the area and circularity variation depending on time, which were chosen according to goodness of fit and parameter signification criteria. 10. 7717/peerj. 7233/fig-9 Figure 9 Regression models fitted to the mean area or mean circularity of extracellular-matrix images as a function of time. (A) Mean area as a function of time corresponding to the CCO group (logistic model). (B) Mean area as a function of time in the CCT group (logistic model). (C) The natural logarithm of mean circularity of the extracellular matrix as a function of time in the CCO group (linear model). (D) The natural logarithm of mean circularity of the extracellular matrix as a function of time in the CCT group (linear model). Collagen degradation depending on cell growth modeling The next step is to determine the dependence relation between collagen degradation (through its mass loss percentage) and cell growth and cell differentiation indicators such as mean area and circularity of the extracellular matrix. On the basis of the results in the previous sections, logarithmic transformation of the area and circularity was proposed above to obtain more informative, simpler, and more accurate models. Indeed, linear regression models are proposed to explain the collagen mass loss with respect to circularity and the area (after natural logarithms are applied). Figure 10 presents different linear models fitted to the real data obtained from groups CCO ( Figs. 10A, 10C ) and CCT ( Figs. 10B, 10D ). Very informative linear models were fitted to estimate the mass loss versus area ( massloss = − 5. 217 + 10. 578ln( area ) with R 2 = 0. 70) and depending on circularity ( massloss = − 19. 4 − 14. 9ln( circularity ) with R 2 = 0. 88) for the CCO group. It is important to stress the strong relation between the mass loss and circularity. The obtained model could be useful even for prediction tasks. The same trends were found in the CCT group, but high dispersion prevented obtaining more informative models. In conclusion, the degree of degradation, measured as the mass loss percent, is proportional to the logarithm of extracellular-matrix area and inversely proportional to the logarithm of cell circularity. 10. 7717/peerj. 7233/fig-10 Figure 10 Linear models fitted to the collagen mass loss obtained in groups CCO and CCT. (A) Collagen mass loss as a function of the natural logarithm of mean area of the extracellular matrix in the CCO group. (B) Collagen mass loss as a function of the natural logarithm of mean area of the extracellular matrix in the CCT group. (C) Collagen mass loss as a function of the natural logarithm of mean circularity of the extracellular matrix in the CCO group. (D) Collagen mass loss as a function of the natural logarithm of mean circularity of the extracellular matrix in the CCT group. A more complete multivariate linear model that accounts for all the collagen degradation sources of variation was also estimated and validated by a threefold cross-validation procedure. The estimated model expression (see Table 4 ) is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\widehat{Mass loss}=-17. 0396-0. 215CCT-2. 1751CO-0. 0599\ln \nolimits (circularity)\nonumber\\\displaystyle +3. 7804\ln \nolimits (mode)+0. 1760\ln \nolimits (area) \end{eqnarray*}\end{document} M a s s l o s s ^ = − 17. 0396 − 0. 215 C C T − 2. 1751 C O − 0. 0599 ln c i r c u l a r i t y + 3. 7804 ln m o d e + 0. 1760 ln a r e a where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\widehat{Mass loss}$\end{document} M a s s l o s s ^ is the mass loss estimates; CCT and CO are dichotomous variables equal to 1 if the corresponding group is CCT or CO, respectively; and mode is the grayscale histogram mode corresponding to each image. It is characterized by R 2 = 0. 89. The latter variable is included to take into account information about the image color (possibly related to cellular growth) in the model. Increasing the extracellular-matrix area and histogram mode increases the collagen mass loss, whereas decreasing circularity increases the mass loss. The effects in CO and CCT groups are a decrease in the collagen mass loss due to cell activity, when compared with the CCO reference group. 10. 7717/peerj. 7233/table-4 Table 4 Signification analysis of multivariate-linear-model parameters and bootstrap intervals at the 95% confidence level. Parameters Estimates Bootstrap 95% interval lower limit Bootstrap 95% interval upper limit Standard error t value p value Intercept −17. 0396 −25. 88 2. 79 4. 5 −6. 1 2. 3e−6 CCT −0. 2150 −0. 346 −0. 078 0. 07 −3. 2 0. 0033 CO −2. 1751 −2. 587 −1. 641 0. 23 −9. 6 7e−10 ln(circularity) −0. 0599 −0. 173 0. 008 0. 0514 −1. 7 0. 1080 ln(mode) 3. 7804 2. 509 5. 387 0. 52 7. 2 1. 5e−7 ln(area) 0. 1760 0. 065 0. 249 0. 04 4. 1 0. 0004 Figure 11A illustrates relative importance of the group, extracellular-matrix histogram mode, circularity, and area for the type I collagen mass loss, in % of R 2. For this purpose, the R 2 contribution averaged over orderings of regressors and lmg metrics ( Grömping, 2006 ) were used. Almost all the variability in the collagen mass loss could be explained by group (∼56% of R 2 ) and mean area variation (∼23% of R 2 ) (with preapplied logarithm transformation), but the contributions of circularity (∼8% of R 2 ) and mode (∼14% of R 2 ) are significant too. The overall determination coefficient of the multivariate linear model is 0. 89. 10. 7717/peerj. 7233/fig-11 Figure 11 (A) Relative importance of predictors (in terms of R 2 ; in the multivariate linear model) that explain the collagen mass loss; (B) observed values vs. model predictions when 2/3 of data was utilized as a training sample via a cross-validation procedure to evaluate predictive accuracy of the model. Figure 11B shows the results of the cross-validation procedure for measuring predictive accuracy of the multiple linear regression model. The data are randomly assigned to three different folds. During three iterations, each fold is removed (test sample), while the remaining data (training sample) are employed to evaluate the model and then to predict the mass loss corresponding to the removed observations (corresponding to the test group). In fact, observed versus predicted values are depicted in a scatterplot. The cross-validation predictions for each of the three groups in which the data are divided are indicated with different colors. The data points revealed linear trends almost coincident with each other and very close to the bisector. These findings support high predictive accuracy of the newly developed multivariate model for analysis of collagen mass loss. In conclusion, prediction of the degradation degree of collagen (measured as the percentage of mass loss) could be done via the predictors based on image analysis of the extracellular matrix. Data, workplace and scripts corresponding to the statistical analysis can be found in Dataset S2, Dataset S3 and Code S1. Discussion The combination of microscopy, image analysis, classification machine learning methods, and statistical learning tools of regression modeling allowed us to model the degradation level of collagen as a function of time and depending on cell growth and differentiation features. Consequently, the main goal of this work is to provide a systematic and quantitative procedure for estimating the degree of degradation of collagen scaffolds. This work includes both (1) a proposed strategy for obtaining relevant data by imaging techniques, and (2) a comprehensive statistical approach to model the collagen degradation and cell growth. This approach provides information complementary to the results of relevant works where the collagen degree of degradation has been studied but not modeled ( Alberti & Xu, 2015 ; Ma et al. , 2003 ). The statistical modeling also provides tools for characterizing degradation of a material and for separating the effects of different sources of degradation (hydrolysis and biological degradation). In addition, the relation between mass loss and features extracted from micrographs (using optical microscopy) was studied here, opening up a way to characterize collagen degradation in a more automatic manner by image analysis. Besides, this methodology could be useful for research into the degradation of scaffolds other than type I collagen ( Alizadeh et al. , 2013 ; Han et al. , 2014 ). First of all, the degree of degradation of type I collagen was studied, characterized, and modeled via experimental data (see Fig. 8 ). The mass loss percentage was found to be a critical variable for collagen degradation, and its dependence on time was studied. The statistical exploratory analysis and application of a GAM revealed that there are at least two sigmoid processes of degradation. Thus, a logistic mixture regression model was proposed and successfully applied to determine this relation, which represents a mode of degradation of collagen. The first logistic component (it accounts for 50% of collagen mass) depends on another degradation factor apart from cellular differentiation and growth. Further research is needed to identify the type of the degradation processes involved, namely a hydrolysis process ( Schliecker et al. , 2003 ). The second logistic component may be more clearly related to the cell activity and corresponds to 20% of collagen mass. Differences between fittings corresponding to groups CCO and CCT were uncovered by studying the first and second logistic functions. Thus, the osteogenic medium plays a significant role in collagen degradation. Namely, the first logistic degradation process tends to take place earlier in the CCO group than in group CCT. In addition, the rate of degradation in the second logistic degradation process is higher in the CCO group. This comparison was made via examination of model parameter estimates such as logistic inflection point µand scale (see Table 3 ). The logistic mixture models could help to separate the degradation modes due to different causes (cell growth and hydrolysis among others) and to study the collagen biodegradation more properly. Once the degree of degradation of collagen was modeled as a function of time, this study also addressed the problem of estimating the progression of collagen degradation with respect to the growth and differentiation of osteocytes from stem cells. This relation modeling is necessary to develop tissue engineering applications. Therefore, methodologies that can extract relevant quantitative information about cell growth are necessary. The application of microscopy, combined with image analysis and supervised classification, is proposed in this work to perform this task. In fact, the random forest supervised classification algorithm permits the identification of different regions of interest in each micrograph automatically: extracellular matrix, collagen, nuclei, and background. Thus, the extracellular matrix areas of each image and their evolution depending on time can be studied separately. The random forest methodology can be applied if previous image texture analysis and feature extraction (pixel energy and tensor matrix) can be implemented. From segmented images of the extracellular matrix, another feature extraction process is implemented to obtain characteristics such as the mean area and mean circularity of the identified object in each image. These are relevant features of osteocyte cell growth. Images show how the extracelular matrix ROI increases in terms of area, whereas the collagen region (that supports the cells) decreases. This can be observed in Figs. 2 – 4 – 9, in which the region containing the extracellular matrix decreases, although its concentration (and its area) increase significantly. This relevant information about cell growth is summarized in the extracted features shown in Figs. 6 and 7. The statistical modeling of representative features of cell growth and differentiation with respect to time is intended. Identifying the type of relationship allow us to obtain information about the degree of degradation due to cell activity. In the case of the CCO group, the area varies in a sigmoid way with respect to time, in fact the logistic model explain the 75% of the overall variability ( Fig. 9A ). The bootstrap 95% confidence interval for µparameter, 6. 756 < μ < 12. 177, shows that the extracellular mean area growth could be more related with the second logistic function of the fitting of mass loss versus time. Thus, this second logistic component is also related with the biological degradation. Moreover, the evolution of the image objects mean circularity is studied. Taking into account its asymptotic decay, a logarithmic transformation is proposed. A clear linear trend can be observed in the Fig. 9C, indeed the fitted linear model ln( circularity ) = − 3. 03 − 0. 136 Time explain the 80% of ln( circularity ). Therefore, the extracellular matrix circularity is time dependent. The same trends can be observed for area and circularity of extracellular matrix of CCT group but with higher dispersion. In fact, the measurements uncertainty prevents obtaining an informative model for the area depending on time. Informative linear regression models were fitted to define the relationship between collagen degradation and cellular growth ( Fig. 10 ). The collagen mass loss linearly varies with respect to area and circularity when logarithmic transformation of predictors is applied. In fact, mass loss increases significantly at a constant rate when ln(area) increases. Otherwise, mass loss decreases significantly at a constant rate when ln(circularity) increases. Thus, increasing osteocyte extracellular matrix area and decreasing its circularity are related with the increase in collagen mass loss, i. e. , the degradation degree. When the circularity of the extracellular matrix is known, the scaffold mass loss (and thus the degree of degradation) can be estimated through the application of linear models, mainly when the CCO group is studied ( R 2 = 0. 88). A relatively strong linear dependence was also identified between mass loss and extracellular matrix area ( R 2 = 0. 7), providing a tool to estimate the degree of collagen degradation in a reasonable way. In the CCT group, the linear relationships between mass loss with respect to extracellular matrix area and circularity are weaker than in the CCO group ( Fig. 10 ). Thus, the osteogenic medium is an influencing factor of the degree of degradation. Furthermore, a multivariate linear regression model was proposed to estimate the collagen mass loss for all the groups studied (including or not osteogenic medium and/or stem cells), including also the area, circularity and grayscale histogram mode predictors (see Table 4 ). This model allows us to estimate the degree of degradation taking into account a comprehensive set of sources of variation. The fitted model explains about the 89 % of collagen mass loss total variability. The more influencing predictors on collagen degradation are the group and area. This means that the collagen degradation mainly depends on the variation of osteogenic medium and the addition of stem cells. Moreover, the area of extracellular matrix accounts for more than the 20% of the explained model variance. Thus, the collagen degradation (mass loss) is highly dependent on cell growth. In a less extent, histogram mode and circularity also influence the level of degradation of collagen. Finally, the mass loss prediction accuracy of the model was successfully tested by a 3-fold cross validation process ( Fig. 10 ). Thus, this model could be used to estimate the collagen degree of degradation from the features obtained from image analysis. In conclusion, the prediction of the degree of degradation of collagen (measured as the percentage of mass loss) could be done from the predictors based on the image analysis of extracellular matrix. Conclusions A methodology based on image analysis, random forest classification, and statistical learning was proposed for estimating the degree of degradation of collagen scaffolds due to osteocyte cell growth and differentiation. The mass loss percentage is defined as the critical variable for collagen degradation and its dependence on time and features extracted from micrographs (extracellular matrix area, circularity, histogram mode) was studied. The random forest classifier has been proposed to perform the segmentation task, i. e. , the identification of the different ROIs. In fact, the random forest classifier is a more automatic and accurate alternative for identifying the different ROIs (collagen, extracellular matrix, nuclei) than the classical thresholding method, preventing in a greater extent undesirable effects such as the image noise. The statistical descriptive analysis and the nonparametric GAM regression provide information about two different degradation processes. They are identified as two overlapped sigmoid type steps in the type I collagen mass loss trend. Consequently, a logistic mixture regression model was proposed to define this relationship that represents the path of degradation of collagen. Using this type of statistical models allows the separation of the effects of different types of degradation. The first logistic component accounts for other degradation procedures apart from cellular differentiation and growth, whereas the second logistic component is more related to the cell activity, which corresponds with around the 20% collagen mass. Moreover, by studying the parameters of the first and second logistic components of the model, differences between CCT and CCT groups were found. Accordingly, the osteogenic medium was found to be an influencing factor in collagen degradation and its influence could be estimated quantitatively. The relationship between the degree of collagen degradation with respect to the growth and differentiation of osteocytes from stem cells was also modeled. The combined use of microscopy, image analysis, and random forest supervised classification was proposed to extract relevant quantitative information about cell growth. The different regions of interest of each micrograph were identified automatically: extracellular matrix, collagen, and background artifacts. The extracellular matrix areas (representative of cell growth) of each image and their evolution depending on time can be studied separately. As a result, the area, circularity and histogram mode, among other relevant features, were extracted to summarize the osteocyte growth information. The evolution of these features with respect to time was modeled using nonlinear (based on logistic function), and linear models (after logarithmic transformation of predictor). The relation between the advancement of collagen degradation and cellular growth was modeled. Increases in osteocyte extracellular matrix area and decreases in its circularity are related to the increase in the collagen mass loss and, thus, the degree of degradation. The mass loss changes linearly way with respect to the extracelular matrix area and circularity when logarithmic transformation is applied. These relationships were modeled through linear regression models. The mass loss versus circularity relationship was found to be stronger than those between mass loss and extracellular matrix area. This could be because the increase in the extracellular matrix area was related to collagen degradation only when collagen pores are filled. A multivariate linear regression model was used to estimate more accurately the degree of collagen degradation, including all the influencing variables, in addition to evaluating the importance of each predictors. The osteogenic medium type (and stem cell addition) and extracellular matrix area are predictors exerting the greatest influence on collagen degradation. Taking into account that the area accounts for more than the 20% of the R 2, collagen degradation is highly dependent of cell growth. After prediction accuracy testing, this model could be used to estimate the collagen degree of degradation from the features obtained from image analysis. Supplemental Information 10. 7717/peerj. 7233/supp-1 Dataset S1 Raw data, metadata, random forest parameters, comparison with thresholding segmentation, segmented images and validation measurements 1. Description of the folder that has been sent with all the details of the files. 2. Comparison between random forest and thresholding. 3. Hyperparameters of the experiment. 4. Random Forest validation measurements: Precision, Recall, Confusion Matrix. 5. Files to reproduce the experiment (classif and comp folders), including raw data. Click here for additional data file. 10. 7717/peerj. 7233/supp-2 Dataset S2 Image features (cell growth) and collagen mass loss (degree of degradation) depending on time The mass loss proportion of collagen in addition to different features extracted from images corresponding to the extracellular matrix (area, circularity of detected objects, mode, among others), and the time variable are included. Click here for additional data file. 10. 7717/peerj. 7233/supp-3 Dataset S3 The R workspace with all the datasets and objects used in the scripts Click here for additional data file. 10. 7717/peerj. 7233/supp-4 Supplemental Information 1 Code and outputs in R A html report is developed with R markdown and included with the corresponding files. Code and outputs are included. Click here for additional data file. 10. 7717/peerj. 7233/supp-5 Video S1 Video with a tutorial that explain how to identify the different regions of interest form an optical micrograph by using the random forest segmentation method with ImageJ software Click here for additional data file. 10. 7717/peerj. 7233/supp-6 Video S2 Video with a tutorial that explain how to identify the different regions of interest form an optical micrograph by using the threshold segmentation method with ImageJ software Click here for additional data file.

10. 7717/peerj. 7271

PeerJ

Three-dimensional printing with biomaterials in craniofacial and dental tissue engineering

With the development of technology, tissue engineering (TE) has been widely applied in the medical field. In recent years, due to its accuracy and the demands of solid freeform fabrication in TE, three-dimensional printing, also known as additive manufacturing (AM), has been applied for biological scaffold fabrication in craniofacial and dental regeneration. In this review, we have compared several types of AM techniques and summarized their advantages and limitations. The range of printable materials used in craniofacial and dental tissue includes all the biomaterials. Thus, basic and clinical studies were discussed in this review to present the application of AM techniques in craniofacial and dental tissue and their advances during these years, which might provide information for further AM studies in craniofacial and dental TE.

Introduction The development of tissue engineering (TE) and regeneration constitutes a new platform for translational medical research. It has already been an important kind of therapeutic method in craniofacial and dental field, such as trauma, skeletal disease, wound surgery and periodontal disease ( Rai et al. , 2017 ). There are several approaches to develop scaffolds, such as electrospinning, mold casting, salt leaching, sintering and freeze drying. Some of these methods are easy and inexpensive, such as mold casting and salt leaching. Some can fabricate three dimensional scaffolds with good structure with a comparatively high speed, such as electrospinning, however, none of them can solve the problem of solid freeform fabrication. Solid freeform fabrication of three-dimensional scaffolds with complex space structure, not only the irregularly curved external structure, but also the internal porous structure, is important in craniofacial and dental regeneration because of its anatomical limitations. Therefore, attempts to improve design and fabrication of bio-active scaffolds, especially on freeform fabrication comprise majority of studies in biomaterial researches. Recently, additive manufacturing (AM) has been applied for scaffold developing ( He et al. , 2015 ). This method was firstly introduced by Herver Voelcker in 1970 to describe the algorithms for the purposes of 3D solid modeling. AM has been widely used in industry because of its accuracy of shaping ( Torres et al. , 2011a ; Torres et al. , 2011b ). It helps researchers to meet the demands of solid freeform fabrication in TE, too ( Warren et al. , 2003 ; Obregon et al. , 2015 ). It also has unique advantages in fabrication of patient-specific scaffolds with multiple materials. In some recent advances, materials with live cells were used, making it possible to construct organ and tissue using AM ( Mannoor et al. , 2013 ). Another hots spot of study in the field of tissue engineering combined with material manufacturing methods is electrospinning. Electrospinning uses electrostatic principle to manufacture the nanofibers required for TE applications ( Zamani et al. , 2018 ). There are mainly three types of technique: blending electrospinning, coaxial electrospinning, and emulsion electrospinning; they share the same basis ( Lu et al. , 2016 ; Tong, Wang & Lu, 2012 ). There is a high electric field applied to draw a polymer solution between the injection needle and a collector. The polymer forms a suspended drip and is stretched into a conical shape called “Taylor Cone” by the high voltage power. Then, the charged droplet forms a charged jet by breaking free from the surface tension of the top droplet. Due to the evaporation of the solvent or the curing cooling of the solute and melt, the charged jet finally condenses into filaments and deposits on the collecting plate in the form of nonwovens ( Barnes et al. , 2007 ; Nair, Bhattacharyya & Laurencin, 2004 ; Chan et al. , 2009 ). The nanofibers prepared by electrospinning have large specific surface area and high porosity in three-dimensional structure, which makes electrospinning nanofiber membranes have a wide application value in many fields ( Qian et al. , 2011 ; Chung et al. , 2010 ). It is worth mentioning that bio-electrospraying and cell electrospinning, both based on this principle, were firstly used to deal with living cells and whole organisms in 2005/06 ( Jayasinghe, Qureshi & Eagles, 2006 ; Townsend-Nicholson & Jayasinghe, 2006 ). A series of studies have confirmed that this high-strength electric field drive technology, naming bio-electrospraying, showed no significant side effect on the bioactivity of living samples ( Jayasinghe, 2011 ). Cell electrospinning is a leading technology in the formation of cell fibers and stents that can be used to create a variety of biological structures, from simple cell stents and diaphragms to more complex structures ( Jayasinghe, 2013 ). In the recent years, bio-electrospraying and cell electrospinning have attracted significant increasing amount of interest. Here we review the application of AM techniques in craniofacial and dental TE. First, we will describe the types and strategies of four typical AM printers used by tissue engineering researchers most frequently, along with their advantages and limitations. Then, we will present recent advances of AM related with craniofacial bone, craniofacial cartilages and dental tissue. Finally, we will look ahead to recommend the future possible AM research field in craniofacial and dental TE. Survey Methodology PubMed and Web of Science databases were searched (until January 2018) using the following free-text terms: additive manufacturing, craniofacial/dental tissue engineering. AM Approaches in craniofacial and dental TE Selective Laser Sintering (SLS) SLS was developed by Carl Deckard at the University of Texas and described in his master’s thesis ( Deckard, 1991 ; Deckard, Beaman & Darrah, 1992 ; Beaman & Deckard, 1990 ). Its fundamental principle is to control the laser concentrated infrared heating beam to melt free powders together to generate a precise structure. In a SLS printer, a fabrication chamber is settled at the base, filling with tightly compacted plastic powder. The temperature of the chamber is kept just below the melting point of free powder. While the laser beam moves under the guidance of scanner system and computer code, precisely shaped monolayer is printed by causing the temperature to rise above the melting point of plastic powder ( Melchels, Feijen & Grijpma, 2010 ) ( Fig. 1A ). As a result, morphology and melting temperature of the powder are considered as the two crucial parameters in laser sintering ( Mazzoli, 2013 ). According to the mechanism of SLS, the heating temperature should be able to melt the surface layer. The molten materials on the surface then work as binder to connect neighboring non-molten particle cores ( Mazzoli, 2013 ). This so-called “partial melting” phenomenon was modeled first by Fischer et al. (2002). The laser sintering powder is commercially available. They are polymeric materials such as poly(L-lactide) (PLLA) /carbonated hydroxyapatite (CHA) ( Zhou et al. , 2008 ), polyvinyl alcohol (PVA) ( Chua et al. , 2004 ) and poly-e-/caprolactone (PCL) ( Williams et al. , 2005 ). In a SLS printer, polymeric powder have a 50 µm mean particle size diameter ( Mazzoli, 2013 ). 10. 7717/peerj. 7271/fig-1 Figure 1 Four kinds of typical AM printers. (A) Schematic of SLS. The fabrication chamber is settled at the base, filling with tightly compacted plastic powder. When the laser beam moves under the guidance of the scanner system and computer code, precisely shaped monolayer is printed by causing the temperature to rise above the melting point of plastic powder. (B) Schematic of SLA. A computer-controlled laser beam moves and cures the top liquid resin by photopolymerisation. The polymerized resin will adhere to a building platform for support. After finishing the first layer, the building platform drops a defined distance under the liquid surface and the laser repeats the above steps to cure a second layer. (C) Schematic of FFF. Thermoplastic polymeric filament is extruded as the “ink” from a high temperature nozzle (typically 95 °C–230 °C) because of a solid-semiliquid state transition. After printing the pattern of the first layer on a surface, either the nozzle rises, or the platform descends in the Z -axis direction at a thickness of a mono by the control of computer. The process is repeated until structure generation is complete. (D) Schematic of binder jetting: Liquid binder is printed as ink onto powder container. Then a new consecutive solid thin layer of free powder will be put on the binder. This printing process repeats until finishing the work. Many advantage of SLS method, such as accuracy, fast fabricating, low price, elective powder type, no need of supporting material, can be documented ( Mazzoli, Germani & Moriconi, 2007 ). The disadvantage of SLS is that with crucial laser power and scanning speed, there is limit in the size of object fabricated with the commercially obtained machines. What’s more, this method cannot fabricate scaffolds with hydrogel material ( Duan & Wang, 2011 ). Stereolithography (SLA) SLA printing was firstly published in 1986, in U. S. patent Apparatus for production of three-dimensional objects by stereolithography ( Hull, 1986 ). He first exploited the spatially controlled solid transition of liquid-based resins by photopolymerization to produce complex structures layer-by-layer in SLA approach ( Skoog, Goering & Narayan, 2014 ). In brief, a computer-controlled laser beam moves and cures the top liquid resin by photopolymerization. The polymerized resin adhere to a building platform for support. After finishing the first layer, the building platform drops a defined distance under the liquid surface and the above steps repeats to cure a second layer ( Fig. 1B ). This technique was later modified by application of digital light projector, known as digital light processing (DLP). It enables architectures built from the bottom of the building platform. After finished the first layer, the platform raises a short distance from the liquid surface and curing procedure repeats. It looks like the structure is lift by the platform, so that the resin required is significantly reduced. Since DLP derived initially from SLA and they share close concepts, in this review, we use SLA to refer to them both. Taking advantage of the extreme accuracy of laser light, SLA printer has been largely used to build complex and precise structures. Most commercial systems have the capacity to fabricate structures with a resolution of 50 µm. On the other hand, the major limitation of SLA also lies on stereolithography, which limits choices of resins. Most of SLA resins are based on low molecular weight, multi-functional monomers for they formed highly cross-linked networks. Poly (propylene fumarate) (PPF) is the most often used polymer in the fabrication of tissue scaffolds with SLA because of its favorable biocompatibility and photo-cross linking functionality. Although only a limited selection of photocurable resins have been used in SLA, such as PPF and polyurethane (PU) ( Hung, Tseng & Hsu, 2014a ), efforts have been made to improve the features of photocurable materials for TE usage, in order to create biodegradable materials ( Skoog, Goering & Narayan, 2014 ) and cell-compatible photocurable hydrogels, in the past decade. Fused deposition modeling (FDM) FDM is another common AM technique, which was first used in the 1990s ( Cai, Azangwe & Shepherd, 2005 ). The printing process of FDM is based on layer-by-layer deposition of thermoplastic polymers. Due to a solid-semiliquid state transition, thermoplastic polymeric filament is extruded as the “ink” from a high temperature nozzle (usually 95 °C–230 °C). After printing the pattern of the first layer on a surface, either the nozzle rises, or the platform descends in the Z -axis direction at a thickness of a monolayer under the control of computer. The process is repeated until structure generation is completed ( Korpela et al. , 2013 ). Depending upon the polymer material and the design, the FDM printer usually prints 3D structures with a typical thickness of 100–300 µm ( Cai, Azangwe & Shepherd, 2005 ) ( Fig. 1C ). This technique has unique advantages because of its wide-ranged operating temperature, user friendly control system, and large number of commercial platforms. Several kinds of biodegradable materials have been used in the process, including polylactic acid (PLA), PVA, PCL, poly (D, L-lactide-co-glycolide) (PDGA) and poly (D, L-lactide) (PDLLA). Several polymers, such as PLA, PCL and PVA, are extensively utilized for their considerable biocompatibility and biodegradation. With some modification of the printer, hydrogels such as alginate, collagen, decellularized ECM, and marine products such as biogenic polyphosphate (Bio-PolyP) and biogenic silica (Bio-Silica) ( Wang et al. , 2013 ; Wang et al. , 2014 ) can be used as well, providing possibility of loading live cells in printing progress. However, FDM has a significant drawback, which is the lowest precision among the four methods. The minimal scale of the printing bar is about 0. 1 mm ( Cai, Azangwe & Shepherd, 2005 ). It is also difficult to generate micro-porous structures for bone TE without further modifications. In addition, as it is printing in an open space, external supports is needed to get rid of the collapse of structures. After finishing the printing, those supports must be removed carefully. Binder Jetting Binder jetting is a technology developed at almost the same period with FDM. Its first development is in the early 1990s ( Sachs, Cima & Cornie, 1990 ). In 2010, the first binder jetting machine was commercially obtained. Its basic working process shares many similarities with inkjet printing ( Meteyer et al. , 2014 ). In a binder jetting printer, liquid binder is printed as “ink” onto powder container. Then a new consecutive solid thin layer of free powder will be put on the binder. This printing process repeats until work finishes. The structures printed by binder jetting printers have layer thickness among 76–254 µm ( Torres et al. , 2011a ; Torres et al. , 2011b ) ( Fig. 1D ). The advantage of this method is that binder jetting printer has various choices of printable materials: high-performance composites are used to produce tough, strong, colored, and best resolution models, elastomeric materials which give rubber-like properties or casting material which enables the creation of metal prototypes ( He et al. , 2015 ). Another advantage is parts can be produced with no need of supporting structure, so it is more applicable in complicated 3D structure establishment ( Gokuldoss, Kolla & Eckert, 2017 ). This method has a faster printing speed than other AM methods, which can be accelerated by using multiple print heads. On the other hand, the disadvantage of this method is also clear. A lot of post-printing treatment increased the time and financial cost. The control of pore existence, size and shape is difficult because material is stacked, not melted together. Current status and challenges of AM applications for craniofacial bone, cartilage and dental tissues AM application in craniofacial bone TE Polymer biomaterials for craniofacial bone TE. Fabricating maxillofacial bone scaffold is a major application of AM technology in craniofacial usage. The selection of an ideal bone graft material relies on multiple factors such as material viability, graft size, porosity, hydrophilic, biodegradability, osteoconductivity and osteoinductivity. It was first reported that synthetic polymeric materials could generate AM bone scaffolds. Many polymers are printable, for they often have proper melting ranges to fulfill the technique requirement of shaping with FDM or binder jetting. As far back as in 1996, PLA was used as AM material in computer aided design (CAD) bone generation ( Giordano et al. , 1996 ). After that, other polymeric scaffolds have been increasingly developed in AM techniques, such as PCL ( Williams et al. , 2005 ; Lohfeld et al. , 2012 ; Korpela et al. , 2013 ; Van Bael et al. , 2013 ; Temple et al. , 2014a ), poly(lactic-co-glycolic acid) (PLGA) ( Luangphakdy et al. , 2013 ), poly(trimethylene carbonate) (PTMC) ( Blanquer, Sharifi & Grijpma, 2012 ) and so on. As a widely used biomedical material, PLA has good biocompatibility as implants with FDA clearance. Printed PLA bars have physical properties of maximum measured tensile strength. The maximum measured tensile strength of low molecular weight PLLA (53 000) is 17. 40 ± 0. 71 MPa, while that of high molecular weight PLLA (312, 000) is 15. 94 ± 1. 50 MPa ( Giordano et al. , 1996 ). PCL is an alternative with PLA because it does not release acid in PLA remodeling. This means it is more resistant in vivo. PCL also has a lower glass transition temperature and melting temperature, making it superior to PLA in certain bone grafting applications. For instance, PCL can be easily blended with other materials, including tricalcium phosphate (TCP), hydroxyapatite (HA) and bioactive glass (BAG), due to its low melting temperature ( Korpela et al. , 2013 ). In addition, the compressive module of PCL can be increased up to 30–40% by adding 10 wt % of BAG. As modifications for the mechanical performances ( Duan & Wang, 2010 ), polymers are also blended in defined ratios to make printable composites, such as PCL/PLGA by FDM ( Shim et al. , 2014 ) and PLGA/PVA by binder jetting ( Ge et al. , 2009 ). PVA also serves as a porogen in the printed architectures by taking advantage of its water-soluble properties. PVA-blended HA was printed by SLS to study the feasibility of composite scaffold ( Simpson et al. , 2008 ). SEM observations showed significant improvements in the sintering effects and to be a suitable material when processed by SLS for TE scaffolds. Cells and animal models used in craniofacial bone TE. The selection of cell is important for bone TE. For orthopedic and maxillofacial researches, primary stem cells as bone marrow stromal cells (BMSC) ( Fedorovich et al. , 2009 ; Rath et al. , 2012 ) and adipose derived stem cells (ADSC) ( Temple et al. , 2014a ) are wildly applied to seed cell types. Fibroblasts are used for viability test and proliferation essay, as well as human multi-potent dental neural crest-derived progenitor cells (dNC-PCs) ( Fierz et al. , 2008 ). Multiple bone cell lines are applied in AM studies, including MC3T3-E1 ( Leukers et al. , 2005 ; Khalyfa et al. , 2007 ; Lan et al. , 2009 ; Melchels, Feijen & Grijpma, 2010 ; Blanquer, Sharifi & Grijpma, 2012 ), SaOS-2 ( Duan & Wang, 2010 ; Wang et al. , 2013 ), C3H/10T1/2 cells ( Inzana et al. , 2014 ) and MG-63 ( Feng et al. , 2014a ; Feng et al. , 2014b ). With osteogenic induction, the attached bone cells not only exhibited cell viability around 60%–90%, but also kept potential of osteogenic differentiation which is confirmed by observing bone metabolism related RNA and protein expression, such as runt-related transcription factor 2 (RUNX2), bone morphogenetic proteins (BMPs), alkaline phosphatase (ALP) and osteonectin (ON) activity. For cells used in craniofacial bone TE, there are different advantages for different cells. Bone cell lines as MC3T3-E1, SaOS-2, c3h/10T1/2, MG-63 were often used for initial screening of biological activity of materials ( Przekora, 2019 ). Since these cells are tumor-derived cell lines or immortalized osteoblast cell lines, their gene expressions are quite different from those of primary cells ( Pautke et al. , 2004 ). The best seed cells for craniofacial bone TE are still considered to be primary OBs because of their behavior in studying osteoconductive and osteopromotive properties ( Przekora, 2019 ). The advantage of using stem cells also include testing the osteoconductive ability of printing materials ( Temple et al. , 2014b ). What’s more, many kinds of tissue can be the source of autologous stem cells. Several animals had been taken in AM mandible scaffold research. Rabbits are most frequently used in the study of mandibular bone repair ( Alfotawei et al. , 2014 ). A protocol described the usage of three-dimensional printed scaffolds with multipotent mesenchymal stromal cell (MSCs) in mandibular reconstruction of rabbits. They used BMSC and ADSC from rabbits ( Fang et al. , 2017 ). One of the previous studies was performed on six mature minipigs ( Fig. 2 ). The researchers created four mandibular defects on each pig. After the defect sites were modelled by CAD/CAM techniques, scaffolds with complex geometries and very fine structures were produced by AM technology. Then the autologous porcine bone cells were seeded on these polylactic acid/polyglycolic acid (PLA/PGA) copolymer scaffolds. Implanting these tissue-constructs into the bone defects supported bone reconstruction ( Meyer, Neunzehn & Wiesmann, 2012 ). What’s more, in a recent study, researchers proved that the craniofacial reconstruction including mandible could be achieved through 3D bioprinting. They presented an integrated tissue-organ printer (ITOP) that can fabricate stable, human-scale tissue constructs of any shape. They also found vascularized bone growth in the central and peripheral portion in vivo trails of rats ( Kang et al. , 2016 ). For periodontal bone regeneration, at least 4 mm augmentations of craniofacial bone had already been achieved with synthetic monetite blocks. 3D printing TCP plates were used as onlay grafts in periodontal surgery. The 4. 0- and 3. 0-mm high blocks were filled with newly formed bone with 35% and 41% of respective volumes ( Torres et al. , 2011a ; Torres et al. , 2011b ). These 3D-printed customized synthetic onlay grafts were further used in dental implant surgery to achieve bone augments ( Tamimi et al. , 2014 ). Direct writing (DW) technology had been applied to produce a TCP scaffolds to repair the rabbit trephine defect. The scaffolds had micropores ranging from 250 × 250 µm up to 400 × 400 µm. After 16 weeks, 30% of the scaffold was remodeled by osteoclast activity with new bone filling in the scaffolds and across the defects ( Ricci et al. , 2012 ). These studies suggested that AM scaffold with tissue engineering could be used in human craniofacial defect repair in the future. 10. 7717/peerj. 7271/fig-2 Figure 2 Chart of the different working steps done in this investigation. Chart of the different working steps done in this investigation. (A–C) Fabrication of the scaffolds. (D–F) cell cultivation. (G–I) implantation of cell-loaded scaffolds and healing. Histology of bone regeneration 3 days after implantation (arrows mark regions of mineralized matrix; original magnification X10) (J). Defect site 30 days post implantation (arrows mark regions of mineralized matrix; original magnification X10) (K). © Springer ( Meyer, Neunzehn & Wiesmann, 2012 ). Technique challenges for craniofacial bone printing and current strategies. Although cell migration and proliferation inside the porous scaffold were observed in an AM HA scaffolds with inner-connective pores ( Fierz et al. , 2008 ), for all the porous scaffolds, it is still a big challenge to keep good cell viability in the central area. Insufficient nutrition and oxygen in static culture lead to cell necrosis and make low cell density area. The method of dynamic cultivation can partly solve this problem. A dynamic cultivation system by perfusion containers strongly increased the MC3T3-E1 population compared to the static cultivation method in a 7-day in vitro cultivation. Close contact between cells and HA granules were observed deeply in the printed structure ( Leukers et al. , 2005 ). In another study, application of perfusion bioreactor system to a BCP binder jetting fabricated scaffold not only successfully reversed the decreased OB and BMSC cell numbers but also increased their differentiation potential ( Rath et al. , 2012 ). Incomplete healing is another current limitation to AM bone grafts. Therefore, growth factors are applied in scaffolds. Bone morphology protein-2 (BMP2), a bone growth factor with strong bone induction property, is often used. The controlled release of BMP2 can be achieved by surface coating or nanoparticles embedding. More consideration is required according to the printing procedure for AM scaffolds. BMP2 loaded gelatin microparticles (GMPs) was used as a sustained release system and dispersed in hydrogel-based constructs, comparing with direct inclusion of BMP2 in alginate or control GMPs ( Poldervaart et al. , 2013 ). In another study with a multi-head deposition system (MHDS), rhBMP2 was loaded by either gelatin (for short-term delivery within a week) or collagen (for long-term delivery up to 28 days) and dispensed directly into the hollow microchannel structure of PCL/PLGA scaffold during the printing process ( Shim et al. , 2014 ). The in vivo micro-computed tomography (micro CT) and histological analyses indicated that CL/PLGA/collagen/rhBMP2 scaffolds lead to superior bone healing quality at both 4 and 8 weeks, without inflammatory response. Transforming growth factor-β (TGF-β) was another important growth factor widely used in osteoblast differentiation and animal models ( Nikolidakis et al. , 2009 ). Due to the hydrophobic feature of most printable materials, surface modification can be exploited to improve biocompatibility. Collagen is a widely used coating material for AM bone scaffold coating. The flexural strength and toughness of a calcium phosphate scaffold was significantly improved by coating a 0. 5 wt% collagen film ( Inzana et al. , 2014 ). Biomimetic and β-TCP ( Luangphakdy et al. , 2013 ) can enhance the surface roughness and increase bone differentiation, thus may minimizing the need for expensive bone growth factors ( Gibbs et al. , 2014 ) ( Table 1 ). 10. 7717/peerj. 7271/table-1 Table 1 Comparison of various printed bone scaffolds in several in vitro and in vivo studies. Authors Materials Strategies Evidence Model of study Periods Effects Leukers et al. (2005) HA DP+ Sintered In vitro MC3T3-E1 7 days The cells proliferated deep into the structure forming close contact HA granules. Williams et al. (2005) PCL SLS In vitro In vivo BMP7 transduced HGF, Mice 4 weeks SLS printed PCL scaffolds enhance bone tissue in-growth. Mapili et al. (2005) PEGDMA SLA In vitro Acryl-PEG-RGD 24 h Heparan sulfate allows efficient cell attachment and spatial localization of growth factors. Arcaute, Mann & Wicker (2006) PEGDMA SLA In vitro Human dermal fibroblasts 24 h Cell viability reaches at least 87% at 2 h and 24 h following fabrication. Li et al. (2007) epoxy resin (SL, 7560, Huntsman); CPC(scaffold) SLA In vitro OB 7 days Negative molds were generated by SLA. Cell density increased. Khalyfa et al. (2007) TCP/TTCP 3DP, Sintered, polymer infiltration In vitro MC3T3-E1 3 weeks Objects with high compression strengths are obtained without sintering. Cell proliferation and osteogenic differentiation are achieved. Goodridge et al. (2007) SLS In vivo Rabbit tibiae 4 weeks Bone was seen to have grown into the porous structure of the laser-sintered parts. Habibovic et al. (2008) Bioceramic 3DP In vivo 12 adult Dutch milk goats 12 weeks Bone formation within the channels of both monetite and brushite, indicate osteoinductivity of the materials. Lee et al. (2008) PPF/DEF SLA In vitro Fibroblasts 1 week Cells were adhering to and had proliferated at the top surface of the scaffold. Geffre et al. (2009) Polymer (NG) FDM In vivo Femoral condyles (animal NG) 5 months Biomimetic porous design largely enhances bone ingrowth. Lan et al. (2009) PPF/DEF SLA In vitro MC3T3-E1 2 weeks MC3T3 pre-osteoblast compatibility with PPF/DEF scaffolds is greatly enhanced with biomimetic apatite coating Fedorovich et al. (2009) photosensitive hydrogel (Lutrol) Hydrogel extrusion, UV In vitro MSCs 3 weeks MSCs embedded in photopolymerizable Lutrol-TP gels remain viable of 60% and keep potential of osteogenic differentiation. Zigang et al. (2009) PLGA/PVA 3DP In vitro Human Osteoblasts CRL-11372 3 weeks Expression of ALP and osteonectin remain stable whilst collagen type I and osteopontin decrease. Ge et al. (2009) PLGA/PVA 3DP In vivo Rabbit: 1 intra-periosteum model. 2 bone defect of Ilium. 24 weeks In both models, the implanted scaffolds facilitated new bone tissue formation and maturation. Duan & Wang (2010) Customized Ca–P/PHBV SLS In vitro SaOS-2, C3H10T1/2 cells 3 weeks Affinity of rhBMP2 on immobilized heparin facilitated the osteogenic differentiation of C3H10T1/2 cells during the whole period. Warnke et al. (2010) TCP, HAP 3DP+ Sintered In vitro Primary human osteoblasts. 1 week Superior biocompatibility of HAP scaffolds to BioOss@ is proved, while BioOss@ is more compatible than TCP. Melchels, Feijen & Grijpma (2010) poly(D, L-lactide) resin SLA In vitro MC3T3 11 days Pre-osteoblasts showed good adherence to these photo-crosslinked networks. Detsch et al. (2011) HA, TCP, HA/TCP 3DP In vitro RAW 264. 7 cell line 21 days The results show that osteoclast-like cells were able to resorb calcium phosphate surfaces consisting of granules. Torres et al. (2011) b-TCP powder 3DP In vivo Rabbit calvaria vertical bone augmentation 8 weeks Synthetic onlay blocks achieve vertical bone augmentations as as high as 4. 0 mm. Rath et al. (2012) biphasic calcium phosphate (BCP) 3DP + Sintered In vitro OB BMSC 3 weeks, 6 weeks Application of a bioreactor system increases the proliferation and differentiation potential Blanquer, Sharifi & Grijpma (2012) PDLLA 3-FAME/NVP SLA In vitro MC3T3 NG Mouse preosteoblasts readily attach and spread onto porous structures with the well-defined gyroid architectures by SLA. Korpela et al. (2013) PCL/bioactive glass(BAG), PLA FDM In vitro Fibroblasts 2 weeks FDM printed PLA has better cell friendly surface than PCL and PCL/BAG. Luangphakdy et al. (2013) PLGA TCP PPF HA TyrPC MCA 3DP VS SLA VS PL VS CM In vivo Canine Femoral Multi-Defect Model 4 weeks TyrPCPL/TCP and PPF4SLA/HAPLGA Dip are better in biocompatibility than PLGA and PLCL scaffolds. MCA remains the best. Wang et al. (2013) biogenic polyphosphate (bio-polyP) and biogenic silica (bio-silica) SFF/ indirect 3DP/ direct 3DP In vitro SaOS-2 cells, RAW 264. 7 cells 10 days Bio-silica ans bio-polyP increase release of BMP2 while bio-polyP inhibits osteoclasts activity. Van Bael et al. (2013) PCL SLS In vitro hPDCs 2 weeks The double protein coating increased cell metabolic activity and cell differentiation Feng et al. (2014a) Feng et al. (2014b) β -TCP SLS In vitro MG-63 5 days, 4 weeks The mechanical and biological properties of the scaffolds were improved by doping of zinc oxide (ZnO). Feng et al. (2014a) Feng et al. (2014b) nano-HAP SLS(NTSS) In vitro MG-63 5 days Cells adhered and spread well on the scaffolds. A bone-like apatite layer formed. Temple et al. (2014a) Temple et al. (2014b) PCL FDM In vitro hASCs 18 days ASCs seeded on the PCL scaffold are successfully induced in to both vascular and osteogenic differentiation. Shim et al. (2014) PCL/PLGA FDM In vitro in vivo hTMSCs Rabbit radius defect 4 weeks 8 weeks PCL/PLGA/collagen released rhBMP2 over one month in vitro, induced the osteogenic differentiation of hTMSCs in vitro and accelerated the new bone formation in the 20-mm rabbit radius defect. Inzana et al. (2014) Calcium phosphonate powder CPS 3DP In vitro In vivo C3H/10T1/2 cells, Murine critical size femoral defect. 9 weeks 3D printed CPS are enhanced through alternative binder solution formulations. Tween improve the flexural strength of CPS. Implants are osteoconductive. Pati et al. (2015) PCL/PLGA ECM FDM In vitro In vivo hTMSCs, Rat calvarial defect. 8 weeks The differentiation and mineralization may be augmented by combined effect of cell-laid extracellular matrix, exogenous osteogenic factors, and flow-induced shear stress AM application in craniofacial cartilage Polymer biomaterials for craniofacial cartilage TE. Cartilage is one of the few tissues that are not vascularized, which makes its regeneration unique. The most widely applied techniques in cartilage printing included FDM, SLA and SLS. For cartilage repair, polymeric materials like PLA, PCL as well as PLGA were most common cartilage scaffolds. Another kind of major material was the hydrogel. Hydrogel could mimic the elastic module of cartilage and have been applied for cartilage reparation for a long time. Recent study showed PEG hydrogel had promising potential for cartilage bioprinting ( Cui et al. , 2012 ). Cells for craniofacial cartilage TE in AM approaches. Chondrocytes were the standard seed cells in cartilages TE, but chondrocytes from different cartilage subtypes exhibited different differentiation. In AM cartilage regeneration, to generate different cartilage subtypes, chondrocytes were harvested from several kinds of cartilages. In one research, rib cartilage cells were co-cultured with adherent stromal cells in a porous PCL scaffolds fabricated by FDM, making a culture system which may have potential of clinical usage ( Cao, Ho & Teoh, 2003 ). In one research, porcine articular chondrocytes were seeded in PLGA scaffold fabricated with liquid-frozen deposition manufacturing, cultured for a total of 28 days. Final results showed that cells proliferated well and secreted abundant extracellular matrix ( Yen et al. , 2009 ). Not only chondrocytes, but also stem cells were also applied in cartilages TE, such as MSCs and so on ( Pati et al. , 2015 ). Interestingly, bone marrow clots (MC) as a promising resource proved to be a highly efficient, reliable, and simple cell resource that improved the biological performance of scaffolds as well. The FDM printed PCL-HA scaffold incubated with MC exhibited significant improvements in cell proliferation and chondrogenic differentiation. This study suggested that 3D printing scaffolds, MC could provide a promising candidate for cartilage regeneration ( Yao et al. , 2015 ). Stem cell-based approach and chondrocyte-based approach were common choices for cartilage regenerations. The major advantage of using stem cells is that autologous transplantation can be implemented ( Walter, Ossendorff & Schildberg, 2019 ). Unlike chondrocytes, autologous stem cells, such as BMSCs or ADSCs, are rich in source. Xenografts of chondrocytes is not a good choice for human cartilage repair for there are immunological reactions ( Stone et al. , 1997 ). It is also reported that chondrocytes lost the chondrogenic differentiation after several passages ( Von der Mark et al. , 1977 ; Frohlich et al. , 2007 ). On the other hand, the stem cells may form fibrocartilage-like tissue in defect without grows factors ( Yoshioka et al. , 2013 ). Differences in depth of the defect also affect the cartilage regeneration, which should be selected according to research purposes ( Nixon et al. , 2011 ). AM application for TMJ cartilage. The temporal mandibular joint (TMJ) disc is a heterogeneous fibrocartilaginous tissue which plays a vital role in its function. It was reported recently that researchers had developed TMJ disc scaffold with spatiotemporal delivery of connective tissue growth factor (CTGF) and transforming growth factor beta 3 (TGFβ3) which induced fibrochondrogenic differentiation of MSCs. They used layer-by-layer deposition printing technique with polycaprolactone (PCL) to fabricate the scaffold. CTGF and TGFβ3 were used as growth factors and human MSCs were used as seeding cells. After 6 weeks of cell culture, it resulted in a heterogeneous fibrocartilaginous matrix which was similar with the native TMJ disc in structure. Due to the possible effect of remaining PCL scaffold structure, the mechanical properties of the engineered TMJ discs by 6 weeks were approximated to the native properties ( Legemate et al. , 2016 ). Schek et al. (2005) used image-based design (IBD) and solid free-form (SFF) fabrication techniques to generate biphasic scaffolds. They found the growth of cartilaginous tissue and bone tissue after seeding different cells which demonstrated the possible therapy to regenerate TMJ joints ( Fig. 3 ). In another study, researchers found that poly (glycerol sebacate) (PGS) might be potential scaffold material for TMJ disc engineering ( Hagandora et al. , 2013 ). Considering the complex geometries of TMJ cartilage, AM techniques have great potential in its fabrication, and further exploration is needed in customized TMJ cartilage engineering. 10. 7717/peerj. 7271/fig-3 Figure 3 Image-based design allowing creation of defect site- specific scaffolds. The revised legend: Image-based design allowing creation of defect site-specific scaffolds. The patient image (A) is used in conjunction with appropriate microstructure architecture to create the design for the implant (B). This design can then be produced using solid free-form fabrication, as in this prototype constructed from a single polymeric material (C). Scaffolds were demineralized prior to sectioning, resulting in empty areas (marked with *) that were previously occupied by HA. Safranin O and fast green staining showed a large area of pink-stained cartilage (arrow) in the polymer sponge, in contact with the green–brown-stained bone that formed in the ceramic phase (E). Small pockets of cartilage were also observed within the pores of the ceramic phase of the scaffold (E, arrow). Hematoxylin and eosin staining of the ceramic phase showed the formation of bone (F, arrow) with marrow space within the pores of the HA. The assembled composite: the upper polymer phase (white) and the lower ceramic phase (blue) are transversed by the two PLA struts, one of which is visible on the front of the construct (G). © John Wiley & Sons ( Schek et al. , 2005 ). AM application for other craniofacial cartilages: ear, nose and throat. Other than TMJ, in craniofacial area, cartilage also forms ear, nose, and larynx. Anatomically shaped ear, nose and throat were already printed through PR approaches. PCL-based ear and nose scaffold were printed and perfused with type I collagen containing chondrocytes. The samples were implanted into adult Yorkshire pigs for 8 weeks and histologically analyzed. Histological evidences present that they resulted in the growth and maintenance of cartilage-like tissue ( Zopf et al. , 2015 ). A bionic ear was printed with precise anatomic geometry of a human ear by alginate as matrix with 60 million chondrocytes per milliliter. An electrically conductive silver nanoparticle (AgNP) was also printed and infused inductive coil antenna as the sensory part of the ear, connecting to cochlea-shaped electrodes supported on silicone. After in vitro culture, this printed bionic ear not only demonstrated good biocompatibility, but also exhibited enhanced auditory sensing for radio frequency reception, which mimicked the functional human ears ( Mannoor et al. , 2013 ). Functional tissue-engineering tracheal reconstruction has also been reported on rabbits by 3D printed PCL scaffolds. The shape and function of reconstructed trachea were restored successfully without any graft rejection. Histological results showed proper cartilage regeneration ( Chang et al. , 2014 ). Technique challenges for cartilage printing and current strategies. A highlight in cartilage printing is that cells can be printed together with gels as cell vectors. For printing of cell-laden material, the important criterions lay on the suitable shear force and temperature. Otherwise, damage may occur to cells and reduce the viability in the printed constructs ( Derby, 2012 ; Pati et al. , 2015 ). Some studies have been paying attention to modification of the printer nozzle and materials. In one study, an electrospun head was added on an inkjet printer and print electrospun PCL film with fibrin–collagen hydrogel-based cartilage layers inside. It was designed for printing a fibrin-collagen hydrogel of five layers in only 1 mm thickness. With this multi-layer scaffold, this research successfully enhanced the strength of printed materials and overcame the major limitation of inkjet printer in material’s loading ability. Therefore, it is possible to be used to print some load bearing tissue such as cartilage ( Xu et al. , 2013 ) ( Table 2 ). 10. 7717/peerj. 7271/table-2 Table 2 Comparison of various printed cartilage scaffolds in several in vitro and in vivo studies. Authors Materials Strategies Evidence Model of study Periods Effects Cao, Ho & Teoh (2003) PCL (NaOH treated) FDM In vitro hOB(iliac crest) hChondrocytes (rib cartilage) 50 days Osteogenic and chondrogenic cells can grow, proliferate, distribute, and produce extracellu-lar matrix in these PCL scaffolds. Smith et al. (2007) PCL SLS In vivo Yucatan minipig mandibles 3 months Cartilaginous tissue regeneration along the articulating surface with exuberant osseous tissue formation. Yen et al. (2009) PLGA (type II collagen) FDM In vitro Chondrocytes (condyles of Yorkshire pigs) 4 weeks Scaffolds swell slightly. The cartilaginous tissue formation was observed around but not yet in the interior of the constructs. Yen et al. (2009) PLGA (lyophilized for 48 h) LFDM In vitro Chondrocytes (condyles of Yorkshire pigs) 4 weeks Decrease swelling significantly. Mechanical strength is closer to native articular cartilage. Proliferate well and secret abundant ECM. Soman et al. (2012) ZPR PEG SLA In vitro hMSCs 1 week Zero Poisson‘s ratio (ZPR) material PEG has been printed to generate 3D printed scaffolds. The hMSCs adhere and proliferate well. Grogan et al. (2013) GelMA SLA In vitro Ex vivo human avascular zone meniscus cells; Human meniscus ex vivo repair model 6 weeks Micropatterned GelMA scaffolds are non-toxic, produce organized cellular alignment, and promote meniscus-like tissue formation. Mannoor et al. (2013) Alginate, silicon, (AgNP infused) syringe extrusion In vitro Chondrocytes (articular cartilage of calves) 10 weeks The ears are cultured in vitro for 10 weeks. Audio signals are received by the bionic ears. Lee et al. (2013) PCL, hyaluronic acid, gelatin SLS In vitro Chondrocytes (New Zealand white rabbit) 4 weeks This study successfully forms a soft/hard bi-phase scaffold, which offers a better environment for producing more proteins. Xu et al. (2013) PCL, FN, Collagen Inkjet, Electrospun In vitro In vivo Rabbit elastic chondrocytes; Immunodeficient mice subcutaneous model 8 weeks The hybrid electrospinning/inkjet printing technique simplifies production of complex tissues. Schuller-Ravoo et al. (2013) PTMC SLA In vitro Bovine chondrocytes 6 weeks The compression moduli of the constructed cartilage increases 50% to approximately 100 kPa. Gao et al. (2014) PEG Inkjet, UV In vitro human chondrocytes 4 weeks Printed neocartilage demonstrated excellent glycosaminoglycan (GAG) and collagen II production with consistented gene expression. Pati et al. (2015) dECM, PCL Extrusion, FDM In vitro hASCs hTMSCs 2 weeks Tissue-specific dECM bioinks achieve high cell viability and functionality. Chen et al. (2014) PCL (coating with collagne) SLS In vivo Subdermally dorsal model of female nude mice 8 week Collagen as a surface modification material is superior to gelatin in supporting cells growth and stimulating ECM protein secretion. Chang et al. (2014) PCL FDM In vivo Rabbit half-pipe-shaped tracheal defect. Rabbit MSCs 8 weeks The 3DP scaffold with fibrin/MSCs served as a resorbable, chondro-productive, and proper cartilage regeneration strategy. Zhang et al. (2014) PEG/ β -TCP SLA & hydrogel In vivo Rabbit trochlea critical size osteochondral defects. 52 weeks The repaired subchondral bone formed from 16 to 52 weeks in a “flow like” manner from surrounding bone to the defect center gradually. Yao et al. (2015) PCL/HA FDM in vitro in vivo Bone marrow clots and BMSC from 30 female New Zealand white rabbits (5-6 months old). 60 Female nude mice (6-7 weeks old). 4 weeks Combination with MC is a highly efficient, reliable, and simple method that improves the biological performance of 3D PCL/HA scaffold. Zopf et al. (2015) PCL SLA In vitro In vivo Yorkshire pigs Supraperichondrial soft tissue flaps 2 months The histological evidence present that anatomically PCL based ear and nose resulted in the growth and maintenance of cartilage-like tissue. AM applications in dental tissue TE strategies for tooth and periodontal tissue regeneration have been increasingly explored recently even though the implanting of titanium artificial tooth root is clinically more and more mature ( Ohazama et al. , 2004 ; Monteiro & Yelick, 2017 ). By now, two tissue regeneration surgical procedures, guided bone regeneration (GBR) and guided tissue regeneration (GTR), have already been applied in dental clinics and proved to have a reliable effect on bone and gingival regeneration ( Bottino et al. , 2012 ). Few clinical methods can be applied in dental tissue regeneration; however, a lot of AM researches were done in this field. Multiple kinds of cells involve in the progress of dental tissue formation, including ameloblasts for enamel, odontoblasts for dentin, cementoblasts for cementum, and cells of multiple lineages including mesenchymal, fibroblastic, vascular, and neural cells that form dental pulp ( Fisher, Dean & Mikos, 2002 ; Xue et al. , 2013 ; Park et al. , 2014a ; Jensen et al. , 2014 ). Dental tissue includes composites of enamel, dentin and pulp, periodontal ligament, cementum, and so on. Since the dental tissue are related with each other, some researches chose to establish combined dental tissue like scaffolds with AM technology, such as cementum/dentin interface ( Lee et al. , 2014 ) or cementum/PDL interface ( Cho et al. , 2016 ). Various materials can be used in AM technology for dental tissue ( Table 3 ). As a result, we divide the load of press into one (single) tissue regeneration and multi (combined) tissue regeneration and reviewed them one by one. 10. 7717/peerj. 7271/table-3 Table 3 Comparison of various printed dental scaffolds in several in vitro and in vivo studies. Authors Materials Strategies Evidence Model of study Periods Effects Kim et al. (2010) PCL/HA (Infused SDF1- and BMP7-loaded collagen) FDM In vivo 22 male (12-week-old) Sprague-Dawley rats: 1 Rat’s dorsum subcutaneous pouches for human mandibular molar scaffolds, 2 right mandibular central incisor for rat central incisor teeth 9 weeks A putative periodontal ligament and new bone regenerate at the interface of rat incisor scaffold with native alveolar bone by cell homing. Lee et al. (2014) PCL/HA 100 um, 300 um, 600 um. FDM In vitro In vivo 1 DPSCs, 2 PDLSCs, 3 ABSCs. The dorsum’s mid-sagittal plane for 10-week-old immunodeficient mice (Harlan) 4 weeks DPSC-seeded multiphase scaffolds yield aligned PDL-like collagen fibers. The fibers inserted into bone sialoprotein-positive bone-like tissue and putative cementum matrix protein 1-positive/dentin sialophosphoprotein-positive dentin/cementum tissues. Xue et al. (2013) Alginate/ gelatin Hydrogel extrusion In vitro hDPCs Self-defined shaped 3D constructs are printed and achieve the cell viability of 87%. Jensen et al. (2014) PCL FDM In vitro hDPCs S3 weeks The HT-PCL scaffold promotes cell migration and osteogenic differentiation. Rasperini et al. (2015) PCL SLS In vivo Clinical case on a periodontitis patient‘s canine. 13 months The case demonstrated a 3-mm gain of clinical attachment and partial root coverage. However, the scaffold became exposed at the 13th month. Cho et al. (2016) PCL, collagen I gel FDM Ex vivo PDLSCs seeded PCL was placed on tooth root surface defect. 6 weeks The new mineralized tissue layer seen in BMP-7 treated samples expressed cementum protein 1 (CEMP1) Jung, Lee & Cho (2016) PEG, PCL, cell-laden Alginate Hydrogel extrusion and FDM In vitro Multiple-layer bioprinting teeth was fabricated with a frame, two kinds of cell-laden hydrogel and a support. Single dental tissue regeneration. Lee et al. ’s ( 2014 ) group has done tooth and periodontal regeneration by cell homing. The research starts from bioprinting of PCL-HA material into two kinds of anatomically tooth shaped scaffold by SLA technology, one is human molar scaffold, and another is rat incisor scaffold. Growth factors of bone morphogenetic protein-7 (BMP7) and stromal cell-derived factor-1 (SDF1) were added into the scaffold to active cell homing in vivo. These two scaffolds were orthotopically and ectopically implanted into mandibular incisor extraction socket and dorsum subcutaneous pouches of rats. After 9 weeks, tooth-like structures and periodontal integration were successfully generated by their study with endogenous cell homing and angiogenesis ( Kim et al. , 2010 ). High survival rates were reported in a self-defined shape engineered pulp, which was as high as 87% ± 2%. This research was done to establish a dental pulp like tissue with human dental pulp cells (hDPCs) in sodium alginate/gelatin hydrosol (8:2), and an amount of 1 × 10 6 cells/ml were seeded ( Xue et al. , 2013 ). In a recent study to generate artificial periodontal ligament (PDL) tissue, human PDL cells were seeded on anatomically FDM printing PCL/HA scaffolds. In periodontal osseous fenestration defects on nude mice, guided fiber alignment was later observed oblique orientation to the root surface 6 weeks post implant, which mimics the mature PDL fiber aliment ( Park et al. , 2014b ). Another study invested the osteogenic potential of human dental pulp stem cells (hDPSCs) on different porous PCL printing scaffolds. This research used a specially designed double-layer scaffold system for better osteogenic differentiation. The first layer was nanostructured porous PCL (NSP-PCL) scaffold, and the second layer was PCL coating with a mixture of hyaluronic acid and beta-TCP (HT-PCL) scaffold. With 21 days of in vitro cultivation, the NSP-PCL and HT-PCL scaffolds promoted osteogenic differentiation and Ca 2+ deposition, showing promising application periodontal tissue regeneration ( Jensen et al. , 2014 ). A very recent clinic case first showed the SLS printed PCL scaffolds’ application on a periodontal tissue regeneration in a periodontitis patient. The case demonstrated a 3 mm gain of clinical attachment and partial root coverage. However, the scaffold became exposed at the 13th month and been removed. However, it showed huge potential of AM applications for dental tissues ( Rasperini et al. , 2015 ) ( Fig. 4 ). 10. 7717/peerj. 7271/fig-4 Figure 4 Design and fabrication of anatomically shaped human and rat tooth scaffolds by 3D bioprinting. Design and fabrication of anatomically shaped human and rat tooth scaffolds by 3D bioprinting. Anatomic shape of the rat mandibular central incisor (A) and human mandibular first molar (B) were used for 3D reconstruction and bioprinting of a hybrid scaffold of poly- ϵ -caprolactone and hydroxyapatite, with 200-µm microstrands and interconnecting microchannels (diam. , 200 µm), which serve as conduits for cell homing and angiogenesis (C, D). A blended cocktail of stromal-derived factor-1 (100 ng/mL) and bone morphogenetic protein-7 (100 ng/mL) was delivered in 2 mg/mL neutralized type I collagen solution and infused in scaffold microchannels for rat incisor scaffold (E) and human molar scaffold (F), followed by gelation. (G) In human mandibular molar scaffolds, cells populated scaffold microchannels without growthfactor delivery. (H) Combined SDF1 and BMP7 delivery induced substantial cell homing into microchannels. (I) Combined SDF1 and BMP7 delivery homed significantly more cells into the microchannels than without growth-factor delivery ( p < 0. 01; N = 11). (J) Combined SDF1 and BMP7 delivery elaborated significantly more blood vessels than without growth-factor delivery ( p < 0. 05; N = 11). (K, L) Mineral tissue in isolated areas in microchannels adjacent to blood vessels and abundant cells, and confirmed by von Kossa staining. (M) Tissue sections from coronal, middle, and two root portions of human molar scaffolds were quantified for cell density and angiogenesis. s, scaffold; GF, growth factor(s). Scale: 100 µm. © SAGE Publications Kim et al. (2010). Combined dental tissue regeneration. Lee et al. (2014) established a multiphase scaffold mimicking cementum/dentin interface, PDL and alveolar bone by 3D printing blended polycarprolactione/hydroxyapatite (90:10) materials. By adding adequate growth factor and culturing cells, they established PDL-like tissue, the fiber of which connects from one side dentin/cementum tissue to another side bone-like tissue, which is just similar to living PDL’s anatomical property ( Lee et al. , 2014 ). Another recent 3D bioprinting research showed BMP7 was benefitional for cementum formation. This research established an interface between cementum and human PDL like tissue, which is novel in combining natural tissue with artificial AM tissue in vitro. The AM scaffold was fabricated with PLGA, and then seeded human PDLSCs. After 6 weeks of culturing, they found that cementum-like layer can be successfully formed in this interface between cementum and human PDL like tissue. They also found that BMP7 helped in cementum matrix protein 1 secretion in vitro, which may be good for cementum tissue establishment ( Cho et al. , 2016 ). Conclusions The transition of new techniques from a novel experimental phase to be regularly available to any laboratory has frequently driven step-changes in the progress of science ( Hung, Tseng & Hsu, 2014b ). Considering the rapid development of commercial printers and open-resource software, the AM technique has great potential to facilitate the next generation TE. Despite some limitations on current AM scaffolds, there have been recent exciting advances in AM technique microstructure control, porosity, porous interconnectivity, and surface modification, bioactivity in vitro and in vivo. Its development may lead to a promising future to functional tissue and organ regeneration. The following fields are recommended for further AM studies in craniofacial and dental TE: The long-term healing effects on animal models. Pre-clinic studies and clinical application on patients, including the whole procedure from the collection of defect image data of patients to the long-term morphological and functional evaluation of the AM conducted patient-specific scaffolds. All-in-one manufacturer protocol for printing complex tissue structures with customized materials, porosity, surfaces and pattern designs. Tissue and (or) organ printing with live cells.

10. 7717/peerj. 7651

PeerJ

Magnetically directed antioxidant and antimicrobial agent: synthesis and surface functionalization of magnetite with quercetin

Oxidative stress can be reduced substantially using nanoantioxidant materials by tuning its surface morphological features up to a greater extent. The physiochemical, biological and optical properties of the nanoantioxidants can be altered by controlling their size and shape. In view of that, an appropriate synthesis technique should be adopted with optimization of the process variables. Properties of magnetite nanoparticles (IONP) can be tailored to upgrade the performance of biomedicine. Present research deals with the functionalization IONP using a hydrophobic agent of quercetin (Q). The application of quercetin will control its size using both the functionalization method including in-situ and post-synthesis technique. In in-situ techniques, the functionalized magnetite nanoparticles (IONP@Q) have average particles size 6 nm which are smaller than the magnetite (IONP) without functionalization. After post functionalization technique, the average particle size of magnetite IONP@Q2 determined was 11 nm. The nanoparticles also showed high saturation magnetization of about 51–59 emu/g. Before starting the experimental lab work, Prediction Activity Spectra of Substances (PASS) software was used to have a preliminary idea about the biological activities of Q. The antioxidant activity was carried out using 2, 2-diphenyl-1-picrylhydrazyl (DPPH) assay. The antibacterial studies were carried out using well diffusion method. The results obtained were well supported by the simulated results. Furthermore, the values of the half maximal inhibitory concentration (IC50) of the DPPH antioxidant assay were decreased using the functionalized one and it exhibited a 2–3 fold decreasing tendency than the unfunctionalized IONP. This exhibited that the functionalization process can easily enhance the free radical scavenging properties of IONPs up to three times. MIC values confirms that functionalized IONP have excellent antibacterial properties against the strains used ( Staphylococcus aureus, Bacillus subtilis and Escherichia coli ) and fungal strains ( Aspergillus niger, Candida albicans, Trichoderma sp. and Saccharomyces cerevisiae ). The findings of this research showed that the synthesized nanocomposite has combinatorial properties (magnetic, antioxidant and antimicrobial) which can be considered as a promising candidate for biomedical applications. It can be successfully used for the development of biomedicines which can be subsequently applied as antioxidant, anti-inflammatory, antimicrobial and anticancer agents.

Introduction The field of nanotechnology is rapidly emerging as it has elicited much interest among the research community due to a diverse range of applications in different fields such as industry, medicine, and cosmetics. Rather than the bulk materials their nanostructured counterparts display unique optical, physiochemical, electrical and magnetic characteristics. This is attributed to their higher surface to volume ratio, size and shape effects. This innovative feature of nanoparticles enables them to be used extensively for biomedical applications. Different types of nanoparticles including metallic ( Liao, Nehl & Hafner, 2006 ), fluorescent (quantum dot), magnetic ( Chatterjee, Gnanasammandhan & Zhang, 2010 ; Corr, Rakovich & Gun’ko, 2008 ), protein-based nanoparticles ( Hawkins, Soon-Shiong & Desai, 2008 ; Kogan et al. , 2007 ) and polymeric ( Kumari, Yadav & Yadav, 2010 ; Soppimath et al. , 2001 ) nanoparticles are utilized for biomedical applications. However, most of the research has been focused on developing the magnetic nanoparticles. The dimension of the magnetic nanoparticles can be few nanometers up to tens of nanometers. Usually, they have a similar or smaller size than the protein molecule; it is easier for the cells or viruses to interact or attach/infiltrate inside the biological matrix of concern ( Varadan, Chen & Xie, 2008 ). The functionalized IONP can interact and bind with different types of biological molecules including enzymes, proteins, nucleotides or antibodies. It can even interact with the drugs based on functionalization techniques. Thus, an external magnetic field can be used to release it inside the targeted tissues, organ or a tumor ( Chorny et al. , 2010 ; Gao, Gu & Xu, 2009 ; Gupta & Gupta, 2005 ; Lacramioara et al. , 2016 ; Laurent et al. , 2008 ). It can be used as a curing agent for cardiovascular diseases ( Chorny et al. , 2010 ). It has the potential for curing oxidative damage ( Erica, Daniel & Silvana, 2011 ). The presence of magnetic nanoparticles can ensure targeted drug delivery to specific organs ( Verma et al. , 2013 ). The presence of reactive oxygen species (ROS) inside the body can damage several biological activities including DNA-protein cross-links, protein fragmentation/oxidation and enzyme activation/deactivation ( Kareem et al. , 2015 ). The endogenous antioxidants are naturally produced inside the human organ. Some of the anti-oxidants can be provided externally through food and termed as exogenous antioxidants ( Yehye et al. , 2015 ). Recently the development of biocompatible nanoparticles having antioxidant properties has gained a great deal of attention. Among different types of IONP, Magnetite (Fe 3 O 4 ) nanoparticles are extensively used in magnetic separation, targeted drug delivery, magnetic resonance imaging, tissue engineering, bio-separation, magnetic hyperthermia and cell tracking ( Mirzajani & Ahmadi, 2015 ; Thomas et al. , 2015 ; Torres et al. , 2010 ; Tudisco et al. , 2015 ). Earlier research in vivo has demonstrated that Magnetite nanoparticles are comparatively benign due to their non-accumulating tendencies inside the vital organs. It can be promptly eliminated from the body ( Boyer et al. , 2010 ). Polymeric coating such as polyethylene glycol (PEG) over the IONP can reduce its’ toxicity level when used for human fibroblasts ( Wang et al. , 2008 ). Thus, numerous process optimization techniques have been undertaken to functionalize or coat IONPs. This has been done mainly by controlling the synthesis parameters or choosing suitable groups to incorporate with them ( Barreto et al. , 2011 ). Flavonoids are hydrophobic substances and used as natural antioxidants in several studies. This can be classified as flavones, flavonols, flavanones, flavan-3ols, anthocyanidins, and isoflavones ( Ross & Kasum, 2002 ). Quercetin is a kind of natural flavonol and can be extracted from berries, tea, red wine apples, citrus fruits, and red onions. It has exhibited antioxidant ( Casas-Grajales & Muriel, 2015 ; Gormaz, Quintremil & Rodrigo, 2015 ), anti-inflammatory, anti-obesity, ( Williams et al. , 2013 ) anticancer ( Khan et al. , 2016 ), anti-viral and antimicrobial properties ( Aziz et al. , 1998 ; Liu et al. , 2017 ). The coplanar structure coupled with their hydrophobicity enables them to interact with phospholipid bilayer of bio-membranes. The -OH and -C 6 H 5 groups of flavonol can be specific or non-specific in binding to the functional proteins (enzymes, hormone receptors, and transcription factors). However, quercetin is sparingly soluble in water and unstable in physiological systems ( Sun et al. , 2015 ). Thus, its direct applications are somewhat restricted. To resolve these limitations, quercetin can be used as a functionalizing agent for nanoparticles. For instance, magnetite-quercetin nanoparticles have been studied as a drug delivery system ( Barreto et al. , 2011 ). Quercetin functionalized rare earth oxides have been demonstrated to exhibit synergistic antibacterial and hydroxyl radicle scavenging properties ( Wang et al. , 2013 ). Quercetin and Gallic acid have been used for consecutive coating of the bimetallic nanoparticles. The coating enables it to be used successfully as antioxidant, antimicrobial and antitumor agents ( Mittal, Kumar & Banerjee, 2014 ). The coating provided by quercetin can give a protective layer over the nanoparticles to inhibit cellular damage, cytotoxicity and apoptotic death ( Sarkar & Sil, 2014 ). In this research, we have prepared quercetin functionalized IONP, using in-situ synthesis and post-synthesis method. Both the methods used here provided nano-particle samples with controlled particle sizes. The functionalization has been carried out successfully and the sample has shown great potential to be used as an antimicrobial and antioxidant agent. The antioxidant activity of the synthesized sample has been checked using 2, 2-diphenyl-1-picrylhydrazyl (DPPH) assay. Some commonly available pathogens which can easily resist different types of drugs have been chosen for antibacterial studies (e. g. , Gram-positive Staphylococcus aureus, Bacillus subtilis, and Gram-negative Escherichia coli ) in vitro. The antifungal activity of IONP@Q against Aspergillus niger, Candida albicans, Trichoderma sp. and Saccharomyces cerevisiae has been investigated. The biological activity of the synthesized sample has been analyzed using the PASS program. The values of the half maximal inhibitory concentration (IC50) of the DPPH antioxidant assay decreased using the functionalized one and it exhibited a 2–3 fold decreasing tendency than the unfunctionalized IONP. MIC values confirm that functionalized IONP@Q have excellent antibacterial against strains used and fungal strains. Our findings illustrated that the synthesized quercetin functionalized nanoparticles can be a promising candidate as a nano antioxidant and an antimicrobial therapeutic agent. Experimental Method Chemicals Ferrous chloride tetrahydrate (FeCl 2. 4H 2 O, Merck), ferric chloride hexahydrate (FeCl 3. 6H 2 O, Sigma ≥ 97%), Quercetin (R & M) and ammonia solution (R & M, 28%) were used to synthesize the samples. The chemicals thus obtained were of analytical grade. Thus, those were used without further purification. Deionized water was used throughout the entire study. The surface morphological features with a particle size of IONP@Q samples were studied using a JEOL JEM-2100F High-Resolution Transmission Electron Microscope with a field emission gun operating at 200 kV. Samples for TEM measurements were prepared by evaporating a drop of the colloid onto a carbon-coated copper grid. The mean particle size of the was calculated as from the HRTEM image from an observation of 100 particles using Gatan Digital MicroGraph software. The identity of the phase and the degree of crystallinity of the magnetite samples were investigated using a PANalytical X-ray diffractometer (model EMPYREAN) with a primary monochromatic high-intensity Cu-K α ( λ = 1. 54060 A°) radiation. A range of 2 θ (from 10. 00 to 90. 00) was scanned. FTIR of the samples was recorded using a Perkin Elmer FTIR-Spectrum 400. EDX was studied using EDX (INCA Energy 200 (Oxford Inst. )) under a vacuumed condition with a working distance of 6 mm. The surface area method was employed to calculate the percentage composition. Raman spectra of the synthesized samples were analyzed using 514 nm Argon gas laser. The magnetism hysteresis loop measurement was conducted using Lake Shore vibrational sample magnetometer (VSM) in the solid state. The measurement was carried out under room temperature where the magnetic field range was kept at −10 to +10 kOe. Preparation of IONP Both Ferrous and Ferric salts having a molar ratio of 1:1. 5 were first dissolved in 100 ml deionized water (DI). NH 4 OH (3. 0 M) was added dropwise (5 mL min −1 ) to the ferrous/ferric solution. The mixture was stirred at 600 rpm and the final pH of the solution was kept at 11 for 90 min, the solution thus obtained was stirred and heated to 80 °C under oxidizing environment. At the final stage of the reaction, a black precipitate was formed, and it was isolated by magnetic decantation. The precipitate was consecutively washed with deionized water (DI) and ethanol. The resultant sample thus obtained was freeze-dried. Functionalization using in-situ technique Preparation of organic IONP@Q1 Ferrous and ferric salts with a molar ratio of 1:1. 5 were dissolved in 100 ml DI. Quercetin dihydrate (1 g) was dissolved in 3 ml of acetone. Subsequently, the quercetin solution was added to the ferrous/ferric solution with continues stirring. NH 4 OH (3. 0 M) was added dropwise (5 mL min −1 ) to the ferrous/ferric solution at 600 rpm until the solution has a final pH value of 11. The reaction was carried out at 80 °C with continuous stirring for 90 min under the oxidizing environment. The resultant magnetite nanoparticles were washed thrice with DI water and acetone collected using an external magnet and dried using freeze drier. Post functionalization Preparation of IONP@Q2 Quercetin functionalized IONP were synthesized by the nanoprecipitation method ( Kumar et al. , 2014 ) with some modifications. Quercetin (0. 5 g) was dissolved in minimum amount of acetone. Then, 1 g of IONP were dissolved in 50 ml of DI and sonicated for 30 min. Quercetin solution was continuously added during sonication followed by stirring for 24 h. The resultant magnetite nanoparticles were washed thrice with DI water and acetone collected using an external magnet and dried using freeze drier. Antioxidant activity A standard DPPH method with some minor modification was carried out to observe the antioxidant activity of the synthesized sample ( Deligiannakis, Sotiriou & Pratsinis, 2012 ; Sotiriou, Blattmann & Deligiannakis, 2016 ). 300 µL of sample stock methanolic suspensions and 1 mL of a methanolic solution of DPPH (0. 2 mM) were mixed. The mixture was placed inside the 1 cm quartz cuvettes. After 30 min, absorbance was recorded. The absorbance was decreasing at 517 nm continuously. The experiments were performed in duplicate. The sample and the DPPH solution were allowed to mix exactly for 30 min. The absorbance measurements were then taken out precisely within 30 min after mixing. The radical scavenging activity is expressed in percentage by the given relation: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\mathrm{Percentage~ of~ Inhibition} (\text{%})= \frac{(Ac-As)}{Ac} \times 100. \end{eqnarray*}\end{document} Percentage of Inhibition % = A c − A s A c × 100. In which As = Absorbance of the compounds/ positive control and Ac = Absorbance of control (DPPH solution). To determine the concentration required to achieve 50% inhibition (IC50) of DPPH radical; the percentage of DPPH inhibition for each compound was plotted against different concentrations. Antimicrobial activity Antibacterial activity Agar well diffusion method was used to study the antibacterial properties of the IONP@Q. Precultures of all bacterial strains were spread on nutrient agar (NA) and wells (diameter = 6 mm) were filled with 100 µl of the test samples (100 mg/ml) and incubated at 37 °C for 24 h. Sterile distilled water was used as negative control. Positive controls used were streptomycin 100 mg/disc and ampicillin 100 mg/mL for Gram-positive and Gram-negative bacteria, respectively. Antibacterial activity i. e. , the formation of halo (inhibition) zone and the diameter of inhibition zones were measured. Antifungal properties For antifungal activities of IONPQ well diffusion method was used to test it against various fungal strains. Aspergillus niger & Trichoderma sp, a filamentous fungus (multicellular); Saccharomyces cerevisiae and Candida albican, a yeast (unicellular);, yeast were used to study antifungal properties. Potato dextrose agar (PDA) plates were inoculated with fungal strains under aseptic conditions and wells (diameter = 6 mm) were filled with 100 µL of the test samples (100 mg/mL) and incubated at 25 °C for 48 h. Sterile distilled water was used as negative control. The positive control used was nystatin at 100 mg/mL. The percentage of inhibition (POI) of mycelia growth was calculated using Eq. (2): \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*}\mathrm{POI}= \frac{R1-R2}{R1} \times 100 \end{eqnarray*}\end{document} POI = R 1 − R 2 R 1 × 100 where R 1 = radius of the pathogen away from the antagonist. R 2 = radius of the pathogen towards the antagonist. Results Surface functional groups analysis using FTIR techniques The surface functional groups over the synthesized samples were identified using the FTIR analysis and are illustrated in Fig. 1. Figures 1A – 1D illustrates the FTIR spectra of IONPs, Q, In-situ IONPs and Post situ IONPs. The sharp peak around 550, 554 and 549 cm −1 confirms the presence of magnetite ( Shah et al. , 2017 ). After in-situ and post situ functionalization, the peaks were shifted slightly due to IONPs-Q complex formation. The broad peak around 3, 000 cm −1 in all the samples showed the presence of -OH stretching vibration. The presences of the sharp peak for carbonyl groups were observed at 1, 623 and 1, 621 cm −1 IONP@Q1, and IONP@Q2, respectively ( Bukhari et al. , 2009 ). After DPPH assay was carried out, surface functional groups for IONP@Q samples were again identified and shown by Figs. 1E – 1H. Excess DPPH solution was added with IONP@Q samples. The resultant mixture was stored inside the dark for 30 min. After that, the samples were washed three times with ethanol. N-O stretching band was present, and it gave rise to peaks at 1, 530, 1, 310, 1, 513, 1, 335, 1, 530, 1, 329 cm −1 for IONP@Q1 and IONP@Q2 respectively. This reflected that the functionalized surface of IONPs was attached with DPPH radical. The representative peaks for DPPH radicals were absent on IONPs ( Kumar et al. , 2014 ; Shah et al. , 2017 ). 10. 7717/peerj. 7651/fig-1 Figure 1 FTIR spectra of (A–D) IONP@Q before DPPH assay (E–H) IONP@Q after DPPH assay. Raman spectra The structural phase of the functionalized nanoparticles has also been illustrated by Raman spectra ( Fig. 2 ). The Magnetite phase of IONPs is confirmed by the presence of the main vibrational mode around 671 cm −1 (A1g) ( De Faria, Venâncio Silva & De Oliveira, 1997 ). Nevertheless, all the samples have the main band around 671 cm −1, 466 cm −1 and 348 cm −1 that have been assigned to A1g, T2g and Eg vibrations of magnetite. The absence of maghemite was also confirmed by the Raman Spectra ( Francisco et al. , 2011 ; Shebanova & Lazor, 2003 ). The functionalized IONPs showed some additional bands around 1, 217 and 1, 337 cm −1 which are attributed to the aromatic ring of Q. The band around 1, 450 and 1, 519 cm −1 over IONP-Q are due to the OH bending band whereas the band at 1, 597 cm −1 indicates C=O stretching ( Numata & Tanaka, 2011 ). 10. 7717/peerj. 7651/fig-2 Figure 2 Raman spectra of IONP@Q. X-ray Crystallographic Data (XRD) analysis The XRD curves for all the samples showed diffraction peaks at the 2 θ value of 30, 35, 43, 57 and 63, which is around, and Brag reflections, respectively ( Fig. 3 ). The result confirmed the presence of pure magnetite nanoparticle having the cubic inverse spinal structure (JCPDS No. 82-1533) ( Dorniani et al. , 2012 ). The absence of superlattice diffractions around 210, 213, and 300 exhibits the absence of maghemite in all samples. Overall it reflected that the functionalization could not change the phase of iron oxide. 10. 7717/peerj. 7651/fig-3 Figure 3 XRD spectra of IONP@Q. Magnetic properties The saturated mass magnetization was determined by vibrating sample magnetometer. The magnetization values obtained were 64. 19, 51. 04 and 59. 07 emu g −1 for IONP, IONP@Q1, and IONP@Q2, respectively. Figure 4 illustrates the hysteresis loops of the synthesized samples under the magnetic field at room temperature. Table 1 summarizes the values for the saturation magnetizations for the synthesized samples. The results reveal that the synthesized nanoparticles exhibit super-paramagnetic behavior. The functionalized samples exhibited lower saturation magnetization values than the bulk magnetite (∼92 emu g −1 ) and unfunctionalized magnetite nanoparticles ( Cornell & Schwertmann, 2006 ). Lower magnetization values for IONP@Q is because of organic coating and impurities accumulated over the surface of the synthesized nano-composites ( Dorniani et al. , 2014 ; Dorniani et al. , 2012 ; Ma et al. , 2007 ). 10. 7717/peerj. 7651/fig-4 Figure 4 Magnetic hysteresis loops of IONP@Q. Morphology and structure High-Resolution Transmission Electron Microscopy (HRTEM) was carried out to observe the particle size and morphology of the synthesized IONP@Q samples and is illustrated in Fig. 5. The average particle size observed was 10, 6 and 11 nm, for IONP, IONP@Q1, and IONP@Q2 correspondingly ( Fig. 5 ). The particles were the spherical shape and they have uniform size distribution. The nanoparticles were aggregated to form a cluster. This occurred due to the magnetic characteristics of the synthesized particles. The crystal lattice fringe spacing 0. 26 nm confirmed cubic magnetite nanoparticles ( Iyengar et al. , 2014 ). HRTEM images reflected that in-situ functionalized IONP@Q1 has a smaller particle size than the unfunctionalized IONP as well as post functionalized IONP@Q2. This reflects that the in-situ functionalization was efficient enough to control the size of the nanoparticles compared to other synthesis routes. This might be attributed to the formation of the metal-Q complex ( Barreto et al. , 2011 ) as shown in Fig. 6. The presence of hydroxyl groups, over the surface of IONPs allows the attachment of different functional groups. In aqueous medium, IONPs are hydroxyl functionalized which are amphoteric and may react with acids or bases ( Laurent et al. , 2008 ). EDX analysis The existence of magnetite was confirmed by the presence of Fe and O content inside the sample. The organic phase is confirmed by the presence of C contents as shown in Table 2. The uniform distribution of the atoms was identified using the mapping images ( Figs. 7, 8 and 9A ). After DPPH assay, N signals were also observed due to the attachment of DPPH which agrees with the FTIR. Figures 7B, 8B and 9B show the EDX of IONP@Q1, IONP@Q2, and IONP respectively. Percentage of Nitrogen increased from 0 to 0. 3 and 0 to 0. 5 for IONP@Q1 and IONP@Q2 respectively, while carbon contents increased from 4. 6 to 7. 6, and 6. 1 to 8. 3 f for IONP@Q1 and IONP@Q2 respectively. However, nitrogen content was not varied for unfunctionalized IONP. This reflected that the free radicals failed to attach with the IONP surface. 10. 7717/peerj. 7651/table-1 Table 1 Magnetic properties. Sample Ms IONP 64. 19 IONP@Q1 51. 04 IONP@Q2 59. 07 10. 7717/peerj. 7651/fig-5 Figure 5 HRTEM images (A, B, D, E, G and H) and particle size distribution (C, F and I) of IONP@Q. 10. 7717/peerj. 7651/fig-6 Figure 6 Proposed structured of the Magnetite–quercetin complex. PASS-predication For designing novel free radical inhibitors for antioxidant drugs PASS prediction is useful ( Kareem et al. , 2015 ). PASS analysis can predict more than 4, 000 kinds of biological activity having a mean accuracy of about 95% ( Anzali et al. , 2001 ; Kadir et al. , 2014 ; Kadir et al. , 2013 ; Stepanchikova et al. , 2003 ). The results are depicted as the values obtained for appropriate Pa (probability “to be active”) and Pi (probability “to be inactive”) ratio. It is rational that the compounds having values Pa > Pi can exhibit those types of activities ( Way2drug, 2015 ). Table 3 summarizes some portion of the predicted biological activity spectra (Lipid peroxidase inhibitor, antioxidant, a free radical scavenger, and anti-inflammatory) for the Q functionalized samples. Probable activities by PASS to be validated by experimental bioassay. Only predicted antioxidant and free radical scavenging activities of the PASS program were experimentally verified by DPPH assay. Pa > 0. 7 indicated that the corresponding compound was very likely to reveal activity in experiments, 0. 5 < Pa < 0. 7 suggested that the compound was likely to reveal activity in experiments, while Pa < 0. 5 implied that the compound was unlikely to reveal activity in experiments. However, predictive antioxidant values and other biological activities for the Q with Pa > 0. 7, indicated that Q functionalized IONP surface could exhibit improved activity than the unfunctionalized iron oxide. This is due to its’ biocompatibility which can aid in drug transportation system as well as bioimaging. Activities predicted by PASS were validated by experimental bioassay. 10. 7717/peerj. 7651/table-2 Table 2 EDX elemental composition (A) Before DPPH assay (B) After DPPH assay. Sample A B Fe O C N Fe O C N IONP 69. 6 30. 4 – – 77. 5 21. 2 1. 3 IONP@Q1 69. 7 25. 7 4. 6 – 63. 2 28. 9 7. 6 0. 3 IONP@Q2 69. 4 24. 5 6. 1 – 65. 2 26. 1 8. 3 0. 5 10. 7717/peerj. 7651/fig-7 Figure 7 FESEM image (inset: EDX elemental map of Fe, O, C and N) of IONP@Q1 for the following elements: Fe, O, C and N (A) before DPPH assay and (B) after DPPH assay. 10. 7717/peerj. 7651/fig-8 Figure 8 FESEM image (inset: EDX elemental map of Fe, O, C and N) of IONP@Q2 for the Fe, O, C and N (A) before DPPH assay and (B) after DPPH assay. 10. 7717/peerj. 7651/fig-9 Figure 9 FESEM image (inset: EDX elemental map of Fe, O, C and N) of IONP for the Fe, O, C and N (A) before DPPH assay and (B) after DPPH assay. Antioxidant activity The UV–Vis absorption curve obtained for the synthesized samples are illustrated in Fig. 10. From the curve, it can be seen that the peak intensity of DPPH is continuously dropping. The decrease in absorbance around 517 nm was used to calculate the free radical scavenging percentage. Table 4 refers to the IC50 values for DPPH scavenging assay. The inhibition of stable DPPH free radicals of the in-situ synthesized organic nano-compounds were found to be IONP@Q1 (IC50 3 ± 0. 002 mg/ml; 58%), IONP@Q2 (0. 7 ± 0. 002 mg/ml; 75%) and at a 10 −4 M. The values obtained are 1–3 times more than the unfunctionalized magnetite (IC50 4. 7 ± 0. 002 mg/ml; 50%) ( Fig. 11 ). 10. 7717/peerj. 7651/table-3 Table 3 Part of the Predicted Biological Activity Spectra of the Q on the Basis of PASS Prediction Software. Biological Activity a Pa b Pi Antioxidant 0. 878 0. 003 Free Radical Scavenger 0. 816 0. 002 Lipid Peroxidase Inhibitor 0. 813 0. 003 Anti-inflammatory 0. 704 0. 015 Notes. a Probability “to be active”. b Probability “to be inactive”. 10. 7717/peerj. 7651/fig-10 Figure 10 DPPH Scavenging percentage by nanomagnetite at different concentrations. 10. 7717/peerj. 7651/table-4 Table 4 IC50 of IONP@Q. IC50 a Values (mg) ± S. E. M b and Max. inhibition % Sample IC50 mg/ml % Inhibition IONP 5 mg 4. 7 ± 0. 002 50 IONP@Q1 5 mg 3. 0 ± 0. 002 58 IONP@Q2 5 mg 0. 7 ± 0. 002 75 Notes. a IC50, 50% effective concentration. b S. E. M, standard error of the mean. 10. 7717/peerj. 7651/fig-11 Figure 11 DPPH Scavenging percentage by nanomagnetite at different concentrations. The sequence followed for DPPH scavenging properties can be shown as Q > IONP@Q2 > IONP@Q1 > IONP. The free radical scavenging phenomenon can be ascribed for the transfer of electrons from IONP@Q towards the free radicals located at the central nitrogen atom of the DPPH molecule. The scavenging properties of functionalized magnetite were increased as compared with that in the case of naked IONP. Quercetin itself was more potent scavenger than both, functionalized magnetite and naked IONP. The observed moderate AOX enhancement of functionalized magnetite could be attributed to the Q on nanoparticle surface. Antimicrobial activity Antibacterial activity Agar well diffusion test was carried out and the results obtained are illustrated by Fig. 12A. The results are displayed in terms of percentage inhibition of diameter growth (PIDG) of bacterial strain using the concentration of IONP around 100 mg/ml. Both the synthesized sample functionalized using different routes exhibited antibacterial activity for both Gram-positive and negative strains. Nevertheless, the highest PIDG values were obtained after using IONP@Q2. Minimum Inhibition Concentration (MIC) is given in Table 5. This showed its’ efficient and prevailing bactericidal activity. Gram-positive bacteria have a thick peptidoglycan layer (10–30 nm), whereas Gram-negative bacteria have an additional outer layer with a thin peptidoglycan layer (10 nm). IONP@Q has shown different antibacterial activity for different bacterial strains. This was expected and can be explained depending on their composition of the cell wall for each type of strain. After adding the IONP@Q nanoparticles, bacterial growth inhibition will take place. This phenomenon takes place due to the internalization of functionalized IONPs inside the cell. This would destroy the cell wall by damaging the 1, 4 glycosidic bonds. 10. 7717/peerj. 7651/fig-12 Figure 12 Percentage of inhibition (POI) of (A) bacterial growth and (B) fungil growth, after treatment with IONP@Q. Antifungal activity Agar well diffusion test was carried out and the results obtained are illustrated by Fig. 12B. The antifungal activity of IONP@Q against Aspergillus niger, Candida albicans, Trichoderma sp. and Saccharomyces cerevisiae was analyzed. The potential antifungal activity of the synthesized compounds among all fungi strains used. IONP@Q2 compound had exhibited the highest percentage of inhibition (POI) for Aspergillus niger, Candida albicans, and Saccharomyces cerevisiae, while IONP@Q1 showed highest fungal activity against Trichoderma sp. MIC values reveal that antifungal activity of activity of IONP@Q2 was higher for some fungal strains compared to IONP@Q1 as shown in Table 6. The nanoparticles were attached with the respiratory system which leads to the antifungal activity of the synthesized sample. The attachment caused cell death ( Rudramurthy et al. , 2016 ). In general, smaller nanoparticles exhibit higher antimicrobial activity ( Padmavathy & Vijayaraghavan, 2008 ). However, activities also depend on the formulation process and physical characteristics of nanoparticles ( Slavin et al. , 2017 ). 10. 7717/peerj. 7651/table-5 Table 5 MIC value for antibacterial strains. Both ampicillin and streptomycine were used as standard for Gram positive and Gram negative bacteria respectively. Bacterial species Mean value true MIC (mg/mL) Q IONP@Q1 IONP@Q2 Staphylococcus aureus 25 10 5 Bacillus subtilis 25 10 5 Escherichia coli 10 10 5 10. 7717/peerj. 7651/table-6 Table 6 MIC value for antifungal strains. Nystatin was used as standard drug. Fungal species Mean value true MIC (mg/mL) Q IONP@Q1 IONP@Q2 Aspergillus niger 25 25 5 Candida albicans 25 10 5 Trichoderma sp. 25 25 5 Saccharomyces cerevisiae 25 10 5 Conclusions In this research, IONP nanoparticles were successfully functionalized. Both in-situ and post-situ technique was used for functionalization. HRTEM analysis was carried out to determine the average particle size for both the sample functionalized in different techniques. For in-situ functionalized IONP, the particle size obtained was 6 nm whereas, after post-functionalization, the synthesized IONP@Q2 had the particle size 11 nm-slightly bigger than the in-situ one. The particle size obtained here followed the sequence of IONP@Q2 > IONP > IONP@Q1. Due to the use of the in-situ process, the particle size of IONP was controlled by Q. This is owing to the formation of a kind of IONP-Q complex. Predictive values for antioxidant activities and other biological activities for the Q with Pa > 0. 7 were presented. Owing to its biocompatibility, it can be used as a promising candidate for drug delivery and bio-imaging agent. In this research, probable activities by PASS were validated by experimental bioassay. The scavenging activity of the sample using the DPPH was found to be in the order of IONP@Q2 > IONP@Q1 > IONP. Both the methods used here for functionalization of IONP has increased its’ free radical scavenging capacity more than one- to three- fold compared to the unfunctionalized one. The synergistic effect of the magnetite coated by Q had not only increased the free radical scavenging capacity but also controlled the particle size of IONP. The formation of IONP@Q-DPPH was confirmed by FTIR and EDX. Finally, IONP@Q showed potential antifungal and antibacterial effects on the strains under observation. The antimicrobial of IONP@Q2 was higher than IONP@Q1. This might be taken place through the destruction of the membrane. The findings here clearly reveal that the synthesized IONP@Q has combinatorial properties (magnetic, antioxidant and antimicrobial). Thus, the application of the synthesized sample can be promising for further clinical applications. Supplemental Information 10. 7717/peerj. 7651/supp-1 Data S1 EDX Spectra and Pass Studies of IONP@Q Click here for additional data file.

10. 7717/peerj. 8212

PeerJ

In vitro histomorphometric comparison of dental pulp tissue in different teeth

Background Dental pulp (DP) represents an accessible and valuable source promising of stem cells for clinical application. However, there are some disadvantages associated with the isolation of dental pulp stem cells (DPSCs), which include the size and weight of the pulp tissue needed to yield sufficient cells for culturing in vitro. Therefore, the objective of this study was to compare in vitro histomorphometry of DP from permanent (premolars, third molar), supernumerary and deciduous teeth of patients between 5 and 25 years old with regards to weight, length, width and the cell density in the four regions of the DP in order to obtain quantitative parameters in a tissue that represents a valuable source of stem cells. Methods DPs were obtained from 10 central incisors deciduous, 20 permanent teeth (10 premolars, 10 third molars) and 10 supernumeraries (six mesiodents and four inferior premolar shapes). The pulps were carefully removed, and the entire tissue was weighed. The pulp length and the width were measured with a digital Vernier caliper. The cellular density analysis was performed according to the four regions of the DP (coronal, cervical, medial and apical) in histological slides using photography and the ImageJ ® program for quantification. Results The Pearson correlation test revealed that DP weight among different types of teeth is correlated with age in male patients. A significant positive correlation was noted between length and width of the DP with age in both genders. The mean DP weight for supernumerary and third molar teeth was greater than deciduous and premolar teeth. Finally, the histological analysis showed that the coronal and apical portions of DP in supernumerary and premolar teeth have the highest cell density. Conclusions The DP of supernumerary teeth has quantitatively the best morphometric parameters and cell density comparable with the quality of DP obtained from deciduous teeth.

Introduction Dental pulp (DP) is an innervated, highly vascularized soft tissue that provides vitality to the tooth ( Rodas-Junco et al. , 2017 ). DP is located inside each primary or permanent tooth, and its main functions include the generation of dentin and maintenance of its biological and physiological vitality in response to traumatic injuries, physical stimulus or bacterial infections ( Marrelli et al. , 2018 ; Ravindra et al. , 2015 ; Tatullo et al. , 2015 ). The regenerative function of DP suggests that it contains odontogenic progenitor cells or stem cells that are involved in the regeneration process. In this context, DP has drawn attention in dental research as an accessible and valuable source of stem cells known as dental pulp stem cells (DPSCs). Moreover, DPSCs are going to be ideal for tissue engineering and regenerative medicine ( Chalisserry et al. , 2017 ; Honda, Sato & Toriumi, 2017 ; Ledesma-Martínez, Mendoza-Núñez & Santiago-Osorio, 2016 ). However, there are some disadvantages associated with the isolation of DPSCs, which might be directly related to the size and weight of DP tissue by limiting the number of stem cells isolated from it ( Raoof et al. , 2014 ). Besides, some studies have also focused on methods to correlate some morphometric parameter of the pulp tissue with the patient’s age. For example, Ravindra et al. (2015) observed through intra-oral radiography that the total area of the pulp decreased with age. Other authors focus on establishing regression equations using the number of cells in DP to also predict age ( Hossain et al. , 2017 ; Von Böhl et al. , 2016 ). Conversely, until now there are no reports that relate morphometric parameters such as weight, length, width and cellular density of DP among different types of teeth and its influence on tissue quality for the isolation of DP cells. Therefore, the objective of this study was to compare the histomorphometry of DP in temporal and permanent teeth and evaluate the cell density in four regions in this tissue with to purpose of generating quantitative parameters that would have important applications in the DPSCs isolation. Materials and Methods Patient recruitment and tooth storage The DP tissue was obtained from 10 deciduous central incisors, 20 permanent teeth (10 premolars, 10 third molars) and 10 supernumeraries (six mesiodents and four inferior premolar shape). An informed patient consent was obtained from patients or parents of minors. The collection of the material was performed at the Clinics of the Master in Pediatric Dentistry and Oral Surgery, Faculty of Dentistry, Autonomous University of Yucatan. The age of patients ranged from 5 to 25 years, and a slight prevalence of females (21/40, 55. 5%) was noted. The protocol was approved by the Ethics Research Committee of Hideyo Noguchi Regional Research Center, Autonomous University of Yucatán (Approval Number: CIE-06-2017). The extracted deciduous teeth exhibited one-third to two-thirds root resorption with well-defined roots. After extraction, all the teeth were rinsed for 5 min in a conical tube containing phosphate-buffered saline (PBS 1X: 138 mM NaCl, 3 mM KCl, 8. 1 mM Na 2 HPO 4 and 1. 5 mM KH 2 PO 4, pH 7. 4) and labeled with the donor’s age and tooth type. Tissue removal and processing The teeth were immersed in sterile phosphate buffer saline (PBS 1X pH 7. 4), stored on ice pack and transported to the cell culture lab for sample processing. After cleaning the surface of the tooth, a vertical cut of the dental organ was performed using a rotary electric micro motor (String ® ) with a diamond disc (diameter: 22 mm and thickness: 0. 4 mm; ATK ® ). During this process, constant irrigation was maintained with cold PBS 1X pH 7. 4 to reduce overheating of dental tissue ( Figs. 1A – 1C ). Thereafter, the entire pulp tissue was carefully extracted from the cavities of the tooth using a metal clamp and it was weighed using an analytical balance (Citizen CX 200) ( Fig. 1D ). The pulp length and width were measured with a digital Vernier caliper with a 0. 01 mm calibration. 10. 7717/peerj. 8212/fig-1 Figure 1 Cutting technique of vertical dental organ to obtain pulp. (A) Cutting vertical of the dental organ using electric rotatory micro motor with diamond disc and irrigation with cold PBS 1X pH 7. 4. (B and C) Breaking-up of the tooth along the vertical axis using a metal spatula and divided into two equal halves, one mesial and the distal. (D) Isolation of the complete DP tissue from the cavities of the tooth using a metal clamp. Histomorphometry The pulp tissues were fixed in 10% formaldehyde solution. Subsequently, the DP was dehydrated in increasing concentration of alcohol: 60%, 70%, 80% and 100%. After that, the DPs were embedded in paraffin and dissected in sections of 5 µm of thickness with a sliding microtome (Leica LM2500). The slices were then dewaxed and stained with hematoxylin and eosin. Four histology slides from each tooth were selected for analysis. Images of the pulp were captured through the digital microscope at a resolution of 1, 280 × 720 (Leica DM750 camera MC170H) and 4–100× magnifications connected to a computer. A cell counting was performed manually under high-power (100× magnification) at four regions of each DP tissue: coronal, cervical, middle and apical. For each cell population, the number of cells was normalized to the total area of the pulp sample (325 µm × 402 µm). Afterwards, each image obtained from the histology slides were analyzed by ImageJ v1. 49 Software. Statistical analysis A one-factor analysis of variance with Tukey’s post hoc test was used. The Pearson correlation coefficient was calculated to determine the correlation between morphometric measurements of the DP and the relationship with the patient’s age. Statistical significance was defined as p ≤ 0. 05. Results Correlation between morphometric measurements of the dental pulp and patients’ age A correlation study was undertaken to examine the weight, length and width of DP with patients’ age ( Figs. 2A – 2F ). The DP weight was correlated with age in male patients, whereas significant changes ( p ≤ 0. 05) were not noted in females ( Figs. 2A and 2B ). The Pearson correlation test revealed a significant positive ( p ≤ 0. 05) correlation between length and width of the DP with age in both genders ( Figs. 2C – 2F ). In general, the data showed that the DP obtained from males had a greater weight, length and width compared with that from females. The results indicate an optimal age interval in males (15–20 years) and females (20–25 years) to obtain 10 mg of DP. 10. 7717/peerj. 8212/fig-2 Figure 2 Correlation of weight, length and width measurements of the dental pulp and its relationship with the patient’s age. DP tissue from males and females were isolated and evaluated by (A and B) Weight (C and D) Length and (E and F) Width to determinate its correlation with the patient’s age by using the Pearson correlation coefficient. Data from all of the investigated third molar, premolars, deciduous and supernumerary are shown. The value r was calculated for all data at a significance level of p ≤ 0. 05. Comparison of morphometric measurements of dental pulp among different types of teeth Linear data correlation and one-factor variance analyses were applied to the variables of weight, length and width ( Table 1 ). The weight of DP was significantly increased ( p ≤ 0. 05) in supernumerary teeth (20. 5 ± 3. 56 mg) and third molars (18. 7 ± 11. 5 mg) from male patients compared with those from females. In contrast, the mean differences in length and width measurements of DP in the third molars were highly significant ( p ≤ 0. 05) compared with the other types of teeth in both genders ( Table 1 ). 10. 7717/peerj. 8212/table-1 Table 1 Comparison of weight, length and width of dental pulp from different teeth between males and females. The values of weight, width and length of the pulp of each 10 samples per tooth type were calculated. Different letters indicate significant differences between each measurement and type of teeth. p -value <0. 05. Parameters Male Female Total Mean SD Mean SD Mean SD Deciduos Weigth (mg) 2. 40 0. 011 1. 7 0. 0013 2. 05 0. 006 Length (mm) 3. 75 4. 500 4. 33 1. 03 4. 04 a, b, c 2. 765 Width (mm) 1. 00 0. 557 1. 0 0 1. 0 c, d 0. 278 Third molar Weigth (mg) 18. 7 0. 011 14. 6 0. 007 16. 65 0. 009 Length (mm) 9. 66 4. 500 9. 14 2. 600 9. 40 b 3. 550 Width (mm) 3. 66 0. 557 2. 57 0. 786 3. 115 a, b, c 0. 672 Premolar Weigth (mg) 9. 10 0. 003 7. 1 0. 003 8. 10 0. 003 Length (mm) 9. 75 1. 250 7. 83 2. 850 8. 79 a 2. 050 Width (mm) 1. 75 0. 500 1. 66 0. 515 1. 705 a 0. 507 Supernumerary Weigth (mg) 20. 5 0. 035 7. 2 0. 005 13. 85 0. 020 Length (mm) 8. 75 3. 320 9. 0 5. 650 8. 875 c 4. 485 Width (mm) 1. 87 0. 640 1. 5 0. 707 1. 685 b, d 0. 674 Comparison of dental pulp histology and cellular density of dental pulp tissue in different types of teeth Due the DP of the different teeth analyzed showed different morphometric measurements, we hypothesized that the density in the cell-rich zone may also be different. Thus, a histological analysis was performed in coronal, cervical, middle and apical regions of the DP to evaluate the cell density ( Figs. 3A – 3N ). The histological evaluation revealed a three-layer structure consisting of a layer of odontoblasts (OB; Figs. 3A – 3N, yellow arrowhead) with regularly arranged columnar cells in the contour of the pulp and a dispersed layer of cells in all DP samples. Conversely, in the subodontoblastic zone (SOB), a thin and cell-free layer zone (CFZ; Figs. 3A – 3N, green arrowhead) was observed. The cell-rich zone (CRZ; Figs. 3A – 3N, red arrowhead) showed a dense layer of cells per unit, especially in fibroblasts and undifferentiated mesenchymal cells that continue in the central zone of the pulp, in which the presence of blood vessels (BV), nerve fibers (NV) and connective tissue (CT) is highlighted ( Figs. 3A – 3N ). The DP of the supernumeraries presented a limited presence of NV ( Figs. 3I – 3L ) compared with the pulp obtained from the premolars, third molars and deciduous. The histology of the pulp in the coronal and cervical regions of deciduous teeth showed an irregular shape, making layers difficult to identify during staining with hematoxylin and eosin ( Figs. 3M and 3N ). 10. 7717/peerj. 8212/fig-3 Figure 3 Photomicrographs of histologic sections of different regions in dental pulp from premolars, third molar, supernumerary and deciduous teeth. Microscopic image demonstrating a typical tissue from (A–M) coronal, (B–N) cervical, (C–K) middle and (D–L) apical regions of DP in different teeth. All sections were stained with hematoxylin and eosin. OB, odontoblast layer (yellow arrowhead); D, dentin; CFZ, cell-free zone (green arrowhead); CRZ, cell-rich zone (red arrowhead); BV, blood vessels; NF, nerve fibers; D, dentin; P, pulp and CT, connective tissue. UD: denotes Not determined, this regions are not found in deciduous teeth (O and P). A greater cellular density was observed in the coronal region of deciduous and premolar teeth compared with supernumerary and third molars ( Table 2 ). In contrast, a high cell density was observed in the apical region of supernumeraries and premolars compared with third molar teeth ( Table 2 ). Together, these results indicate that the apical region of the DP in supernumerary and premolar teeth potentially represents the ideal location to obtain cells. 10. 7717/peerj. 8212/table-2 Table 2 Cell density values in specific regions of dental pulp tissue from different teeth. All values were calculated from three histological sites for each one of the regions of DP among the different teeth expressed as the mean ± standard deviation. (a, b and c) indicate significant differences in each region of DP and the type of tooth. (A, B, C, D and E) indicate significance between different tooth and each region of DP, p ≤ 0. 05. UD: denotes Not determined, this regions are not found in deciduous teeth. Teeth Cell density (cel/mm 2 ) Coronal Cervical Middle Apical Mean total Premolar 40. 90 ± 7. 46 36. 80 ± 9. 09 25. 80 ± 8. 07 51. 60 ± 10. 12 38. 78 ± 12. 57 a Third molar 19. 20 ± 4. 42 19. 40 ± 4. 43 12. 20 ± 5. 22 30. 30 ± 2. 00 20. 28 ± 7. 70 a, b, c Supernumerary 39. 70 ± 12. 63 38. 50 ± 10. 16 24. 90 ± 5. 71 61. 00 ± 10. 08 41. 03 ± 16. 21 b Deciduous 47. 30 ± 14. 71 46. 80 ± 14. 74 UD UD 47. 05 ± 14. 34 c Mean total 36. 78 ± 14. 77 A, B 35. 38 ± 14. 12 C, D 20. 97 ± 8. 87 A, C, E 47. 63 ± 15. 34 B, D, E Discussion Knowledge of DP histomorphometry in teeth is important for identifying the tooth type that could provide the best source of cells. Several researchers have focused on obtaining DPSCs from exfoliated deciduous or permanent third molars because stem cells from these teeth exhibit a high proliferation capacity ( Daud et al. , 2016 ; Shekar & Ranganathan, 2012 ). It is therefore necessary to identify other source of DPs with quality cellular characteristics for the isolation of cells. In the literature, reports on morphometric and histological measurements of the DP regions among temporary and permanent teeth are quite limited. Besides, pulp weight is a parameter that few authors have considered to obtain DPSCs ( Alsulaimani et al. , 2016 ; Kellner et al. , 2014 ; Singh et al. , 2016 ). Our research group considers that DP weight could be important as a starting point for the isolation of DPSCs. The results of the present study show that the DP weight was greater in the supernumerary and third molars teeth in male patients compared with female patients ( Table 1 ). The weight observed in supernumerary ( mesiodents ) teeth was may be due to the dense fibrous DP, which could indicate a greater amount of organic substance in this tooth type. The third molar DP of females exhibited an increased weight compared with males likely because female DP was obtained from upper teeth with fused roots. This characteristic facilitated the procurement of a larger homogenous pulp tissue compared with males whose third molars were primarily from the mandible and with separated roots. In Méndez & Zarzoza (1999) reported a mean weight pulp of 13. 10 ± 4. 33 mg in eight premolars fragmented with a hammer. In our study, we used a vertical cutting method to obtain DP from all types of teeth. This variation in weight could be explained by the fact that the fractionation by impact could lead to the loss of pulp tissue. In addition, some premolars had separated roots, so the size of the pulp tissue may reduce. On the other hand, it is also important to note that the DP also undergoes age-related changes, and several studies have focused on aging. Recently, Kellner et al. (2014) determined that the relation between pulp vs. hard tooth tissue in third molar decreases with aging. This finding was also observed in our study ( Fig. 2 ). Unfortunately, no reports on deciduous pulp weight were found to compare our observations. The length of DP was three-fold higher in permanent compared with deciduous teeth ( Table 1 ) because deciduous teeth showed a physiological reabsorption that does not occur in permanent teeth. Regarding DP width, measurements of deciduous and permanent teeth are obtained using radiological techniques, such as pericapical X-rays or orthopantomography. For example, Kazmi, Anderson & Liversidge (2017) used conventional radiology to measure the mesio-distal crown width of deciduous teeth in males and females, revealing no significant differences, which was consistent with our findings ( Table 1 ). However, there are no reports about the in vitro length of the pulp for comparison with our results. On the other hand, the cell-rich zone contains progenitor cells that exhibit plasticity and pluripotency. For instance, Lizier et al. (2012) indicated that DPSCs are located in multiple niches, which are associated with capillaries and the nerve network of the central region in the CRZ and in the outer layer of pulp tissue ( Graziano et al. , 2008 ; Pagella et al. , 2015 ). We observed that the apical region of the DP of the supernumerary teeth, showed a higher cell density compared to the other permanent teeth. Interestingly, in the total analysis of cell density, the supernumerary teeth have similar cellular density compared to deciduous teeth ( Table 2 ). Although the pulp tissue of the deciduous teeth was smaller, this pulp exhibited the highest cell density compared with the other dental organs ( Table 2 ). This finding could be explained because the coronal and apical regions are held together after a physiological resorption in deciduous teeth. Gronthos et al. (2002) showed that the DP derived from lower deciduous central incisors contains a large number of cells able to form adherent colonies similar to mesenchymal stem cells in in vitro culture. In contrast, the small number of cells available for isolation due to the size of the pulp, especially in exfoliated deciduous teeth represents a potential problem with obtaining DPSCs ( Marrelli et al. , 2018 ). Thus, one of the advantages of supernumerary teeth for the isolation of cells is that these teeth are extracted at an early age, which apparently retain embryogenic characteristics as demonstrated for another source of stem cells from the oral cavity ( Dunaway et al. , 2017 ). These results indicate that supernumerary teeth in patients between 5–20 years of age have the best morphometric parameters. However, the determination of the biological properties such as proliferation and differentiation potential of the isolated cells in the different regions of this tissue requires further studies. Conclusions In this study, in vitro histomorphometric comparison and cellular density of the DP from temporary and permanent teeth of patients from 5 to 25 years of age were addressed. It was shown that supernumerary DP has the best morphometric parameters and its cell density is comparable to that of deciduous tooth pulp. This phenomenon has not been described before and could have important applications in the isolation of stem cells in this tissue. Supplemental Information 10. 7717/peerj. 8212/supp-1 Supplemental Information 1 Raw data parameter analysis. Click here for additional data file. 10. 7717/peerj. 8212/supp-2 Supplemental Information 2 Raw Data of pulp parameters. Click here for additional data file. 10. 7717/peerj. 8212/supp-3 Supplemental Information 3 Raw data cell density in different teeth. Click here for additional data file.

10. 7717/peerj. 8663

PeerJ

Suppression of osteogenic differentiation and mitochondrial function change in human periodontal ligament stem cells by melatonin at physiological levels

N-Acetyl-5-methoxytryptamine (melatonin, MT) at pharmacological concentrations promotes the osteogenic differentiation of human bone marrow-derived mesenchymal stem cells; however, its role at physiological concentrations (1 pM–10 nM) remains unclear. We explored the effects of 1 pM–1 µM MT on the osteogenic differentiation of human periodontal ligament stem cells (hPDLSCs) and its underlying mitochondrial dynamics-mediated mechanisms. T he PDLSC phenotype was detected by flow cytometry and evaluated for three-line differentiation. Alkaline phosphatase activity assay and Alizarin red staining were used to evaluate osteogenic differentiation. Osteogenesis-related gene and protein expression levels were measured by quantitative reverse transcription -polymerase chain reaction and western blotting. Mitochondrial function assays were performed using reactive oxygen species, ATP and NAD + /NADH kits and molecular mechanisms of mitochondrial dynamics-related proteins were assessed by western blotting. Our results have shown that physiological MT concentrations induced differentiation of hPDLSCs and down-regulated osteopontin (OPN) and osteocalcin (OCN) expression levels, which were restored or even up-regulated by 1 µM MT (lowest pharmacological concentration). Compared to the osteogenic induction alone, this treatment decreased the intracellular ATP content, whereas the intracellular reactive oxygen species level and NAD+/NADH ratio were increased. Mitochondrial function- and dynamics-related protein expression levels were consistent with those of osteogenic genes following osteogenic induction and MT treatment of hPDLSCs at various physiological concentrations. Physiological MT concentrations inhibited the osteogenic differentiation of hPDLSCs and simultaneously altered mitochondrial function. These findings provide insights into the stem cell tissue engineering and functions of MT.

Introduction Human periodontal ligament stem cells (hPDLSCs) can be isolated from the human periodontal ligament (hPDL). Stem cell therapy and tissue engineering are promising treatments for bone defects, with hPDLSCs as possible ideal seed cells ( Xu et al. , 2019 ). hPDLSCs have good osteogenic, neural chondrogenic, and adipogenic differentiation abilities ( Zhu & Liang, 2015 ). Cell differentiation is associated with the mitochondria, particularly its dynamics and biogenesis ( Seo, Yoon & Do, 2018 ). Forni et al. (2016) suggested that mitochondrial fusion and fission play a vital role in mitochondrial network remodeling during differentiation. Mitochondria, as subcellular structures, have various functions including ATP production, cellular Ca 2+ buffering and apoptosis ( Arruda & Hotamisligil, 2015 ; Kato et al. , 2017 ). Cell differentiation requires a considerable amount of energy, in which the mitochondrion plays a vital role ( Shen et al. , 2018 ). Furthermore, the loss of stem cell pluripotency and differentiation is associated with an increased mitochondrial oxidative capacity ( Mitra, 2013 ). Mitochondrial biogenesis and dynamics are involved in the osteogenic differentiation of mesenchymal stem cells ( Chen et al. , 2008 ). Additionally, mitofusin-2 (MFN2), dynamin-related protein1, mitochondrial fission factor (MFF), and translocase of outer membrane 20 (TOM20) may regulate mitochondrial dynamics ( Grey et al. , 2000 ). N-Acetyl-5-methoxytryptamine (melatonin, MT) is mainly synthesized by the pineal gland and affects sleep induction, antitumor, anti-inflammatory and free radical scavenging activities ( Mauriz et al. , 2013 ). MT plays an important role in bone metabolism by promoting osteoblast osteogenesis ( Zhou et al. , 2019 ). The osteoblast-inducing, bone-enhancing effects of MT involved osteoblasts and osteoclasts through the regulation of melatonin type 2 (MT 2 ) receptor, MEK1/2 and MEK5 ( Maria et al. , 2018 ). MT was also found to suppress osteoclastic and osteoblastic activities in goldfish scales ( Suzuki & Hattori, 2002 ) and protect against the oxidative stress-induced hPDL osteoblast senescence phenotype and osteoclast differentiation ( Bae et al. , 2018 ). Two ranges of MT concentrations, 0. 01–10 and 1–100 nM, are considered as physiological and pharmacological, respectively ( Moriya et al. , 2007 ). MT dosage of 5 and 100 nM directly stimulates hBMSCs and osteoblast osteogenic differentiation via the MT 2 receptor pathway ( Radio, Doctor & Witt-Enderby, 2006 ; Sharan et al. , 2017 ). It has been shown that a continuous exposure for 21 days to 50 nM MT is required to induce osteoblast differentiation from human mesenchymal stem cells (hAMSCs) through the formation of MT 2 R/Gi/β-arrestin/MEK/ERK1/2 complexes and subsequently induce osteogenesis ( Sethi et al. , 2011 ). Studies of MT have typically used pharmacological concentrations ( Lee, Le & Kang, 2018 ; Li et al. , 2019 ); however, few studies have used physiological concentrations in normal cells such as hPDLSCs. The development of endosymbiotic evolution theory supports that MT is synthesized and metabolized by the mitochondria and chloroplasts ( Zhao et al. , 2019 ). A previous study showed that MT inhibits proliferation and promotes differentiation via mitochondrial complex I and IV activity ( Liu et al. , 2013 ). Additionally, MT treatment enhanced cisplatin and radiation cytotoxicity in head and neck squamous cell carcinoma by stimulating mitochondrial reactive oxygen species (ROS) generation, apoptosis and autophagy ( Fernandez-Gil et al. , 2019 ). Furthermore, MT may restore normal mitochondrial function ( Reiter, Tan & Galano, 2014 ). Thus, strategies for achieving MT-induced changes in mitochondrial function are needed. We investigated the effects of physiological MT concentrations on osteogenic differentiation of hPDLSCs and preliminarily explored the associated mechanisms of mitochondrial dynamics. Materials & Methods Isolation and culture of hPDLSCs The study protocol was approved by the Medical Ethics Committee of the Hospital of Stomatology Sun Yat-sen University, China (KQEC-2019-10). We extracted and collected 15 healthy human premolars from 15 healthy adult patients aged 18–25 years undergoing orthodontic therapy at the Hospital of Stomatology, Sun Yat-sen University. All participating patients signed a written informed consent form. hPDLSCs were isolated and cultured as previously described ( Seo, Miura & Gronthos, 2004 ). In our study, cells grown in alpha modified Eagle’s medium (α-MEM) containing 10% fetal bovine serum (FBS) were used at passage 3–5 for subsequent experiments. hPDLSCs were cultured in osteogenic medium (OS, 10% FBS, 0. 1 mM dexamethasone, 0. 2 mM ascorbic acid, and 10 mM β-glycerophosphate; Sigma-Aldrich, St. Louis, MO, USA). The hPDLSCs were cultured in basal growth medium or OS with MT at various concentrations (0, 1 pM, 0. 1 nM, 10 nM and 1 µM, Sigma-Aldrich) for 3, 7 or 21 days. All above media and reagents were purchased from Gibco (Grand Island, NY, USA) unless otherwise noted. Multipotential differentiation hPDLSCs were induced in OS, adipogenic medium (Cyagen Biosciences, Santa Clara, CA, USA), and chondrogenic medium (Cyagen Biosciences) for 21 days. Next, the cells were fixed in 4% paraformaldehyde and stained with 1% Alizarin red staining (ARS) solution (Cyagen Biosciences) for the osteogenesis assay. Additionally, 3 mg/mL oil red O (Sigma-Aldrich) was used for adipogenesis analysis and Alcian blue staining for the chondrogenic assay. The images were viewed using an inverted microscope (Zeiss AG, Oberkochen, Germany). Cell surface marker expression of hPDLSCs Passage 3 hPDLSCs were prepared for flow cytometric analysis to detect stem cell surface makers. The cells were incubated in TrypLE buffer (Gibco) to obtain single-cell suspensions and then resuspended in phosphate-buffered saline (PBS) containing 2% FBS. The cells were incubated with the mesenchymal markers CD34, CD45, CD73, CD90, CD164 and CD166 (BD Biosciences, Franklin Lakes, NJ, USA) for 30 min at 4 °C. Cell samples were analyzed by flow cytometry (Beckman Coulter, Brea, CA, USA). MT affects hPDLSCs viability The hPDLSCs were treated with different concentrations of MT (0, 1 pM, 0. 1 nM, 10 nM, and 1 µM). Cell viability was tested with the cell counting kit-8 (CCK8, Dojindo, Kumamoto, Japan) after 0, 24, 48, 72 and 96 h according to the manufacturer’s instructions. The absorbance was determined at a wavelength of 450 nm. Alkaline phosphatase (ALP) activity assay hPDLSCs were cultured in OS supplemented with different concentrations of MT for 3 days. Next, a commercial ALP kit (Jiancheng, Nanjing, China) was used to measure ALP activity following the manufacturer’s instructions. The generation of p -nitrophenol in the presence of ALP was determined by measuring the OD at a wavelength of 520 nm. The OD was normalized to the protein concentration. Quantitative assay of ARS hPDLSCs were exposed to MT at various concentrations in OS for 21 days. The cells were stained with ARS and images were acquired with a camera (Nikon, Tokyo, Japan). Next, 500 µL of 0. 1 M hexadecylpyridinium chloride monohydrate (Sigma) was added to each well to dissolve the contents at 37 °C for 30 min. Subsequently, 100 µL of samples was transferred to the wells of a 96-well plate, and the absorbance of the supernatant was measured at 562 nm. Quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) analysis The hPDLSCs were cultured for 7 days, and total mRNA was extracted using a RNA-Quick purification kit (YISHAN Biotechnology, Shanghai, China) according to the manufacturer’s instructions. Next, first-strand cDNA was synthesized using PrimeScript™ reverse transcription (RT) Master Mix (TaKaRa, Shiga, Japan). PCR was performed with SYBR Green I Master Mix (Roche Applied Science, Basel, Switzerland) following the manufacturer’s protocol. Gene-specific primers used in this study were commercially synthesized, and their sequences are listed in Table 1. 10. 7717/peerj. 8663/table-1 Table 1 Primer sequences. Gene abbr. Species Forward primers, 5′–3′ Reverse primers, 5′–3′ OPN Human GTGATTTGCTTTTGCCTCCT GCCACAGCATCTGGGTATTT OCN Human AGCAAAGGTGCAGCCTTTGT GCGCCTGGGTCTCTTCACT GAPDH Human TCTCCTCTGACTTCAACAGCGACA CCCTGTTGCTGTAGCCAAATTCGT Western blotting The hPDLSCs were lysed for 30 min at 7 days post osteogenic induction. Cell extracts containing 40 µg total protein were loaded and separated by 4–20% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SurePAGE, Ubiotechnology, Guangdong, China) and transferred onto a polyvinylidene fluoride membrane (Millipore, Billerica, MA, USA). The membranes were blocked in Tris-buffered saline +0. 1% Tween-20 (TBST) containing 5% bovine serum albumin (BioFroxx, Einhausen, Germany) for 2 h at 20 °C–25 °C after transfer. The membranes were immunoblotted with the following primary antibodies at a 1:1, 000 dilution: polyclonal rabbit anti-OPN (Novus Biologicals, Littleton, CO, USA); polyclonal rabbit anti-OCN (Novus Biologicals), mitochondrial dynamics antibody sampler kit (Cell Signalling Technology, Danvers, MA, USA); and monoclonal mouse anti-β-actin (Beyotime, Shanghai, China) overnight at 4 ° C. This was followed by incubation with secondary antibodies at 20 °C–25 °C for 1 h. Immunoreactive bands were visualized by chemiluminescence detection reagents (Millipore, Temecula, CA) according to the manufacturer’s instructions. Band intensities were quantified using ImageJ 1. 36b (NIH, Bethesda, MD, USA). Measurement of intracellular ATP content The hPDLSCs were cultured for 7 days, after which the ATP level was measured with an enhanced ATP assay kit (Beyotime) according to the manufacturer’s instructions. Absorbance was measured with a luminometer (GloMax Multiplus plate reader, Promega, Madison, WI, USA). Luminescence intensity was divided by the total number of cells. Measurement of intracellular ROS hPDLSCs were cultured for 7 days in TrypLE buffer to obtain single-cell suspensions, resuspended in α -MEM containing 2% FBS, incubated with 5 µM CellROX ® green reagent (Invitrogen, Carlsbad, CA, USA) at 37 °C for 30 min, and then washed with PBS three times. The fluorescence intensity of 1 × 10 4 cells was recorded by flow cytometry. NAD + /NADH assay hPDLSCs were cultured for 7 days. Intracellular NAD + and NADH levels were measured using an NAD + and NADH assay kit (Beyotime) according to the manufacturer’s instructions. The absorbance of the supernatant was measured at 450 nm, and the NAD + /NADH ratio was calculated as follows: [NAD + − NADH]/NADH. Statistical analysis All data are expressed as the means ± S. D. One-way analysis of variance was used to compare groups, and Fisher’s least significant difference (LSD) was used for post-hoc analysis. P < 0. 05 was considered to indicate statistical significance. All experiments were performed in triplicate and SPSS 22. 0 software (SPSS, Inc. , Chicago, IL, USA) was used for statistical analysis. Results Isolation and characterization of PDLSCs hPDLSCs were successfully obtained and exhibited a homogeneous, large, fibroblast-like morphology ( Fig. 1A ). hPDLSCs showed morphological changes and osteogenic/adipogenic/chondrogenic differentiation potential after 3 weeks of induction ( Figs. 1B – 1D ). The mesenchymal stem cell (MSC) properties of the hPDLSCs were characterized by identifying cell surface markers. hPDLSCs were positive for the mesenchymal markers cluster of differentiation 73 (CD73, 96. 54%), CD90 (99. 78%), CD146 (71. 30%), and CD166 (93. 74%) but negative for the hematopoietic markers CD34 and CD45 in flow cytometric analysis ( Figs. 1E – 1J ). These results suggest that hPDLSCs were successfully isolated and had good osteogenic differentiation ability. 10. 7717/peerj. 8663/fig-1 Figure 1 Characterisation of hPDLSCs and effect of MT on cell viability in hPDLSCs. (A) Primary hPDLSCs were obtained by tissue block culture (scale bar = 200 µm). (B) Alizarin red staining of hPDLSCs after 3 weeks of osteogenic induction (scale bar = 100 µm) (C) Oil red o staining of hPDLSCs after 3 weeks of adipogenic differentiation (scale bar = 50 µm). (D) Alcian blue staining of hPDLSCs after 3 weeks of chondrogenic differentiation (scale bar = 50 µm). (E–J) Flow cytometry showed hPDLSCs were positive for the MSC markers CD73, CD90, CD146 and CD166 but negative for the haematopoietic markers CD34 and CD45. (K) Cell viability was measured by CCK8 assay and showed no significant difference in cell viability between the MT treatment groups and basal growth medium treatment group (NS: no significant difference). MT does NOT affect the viability of hPDLSCs There was no significant difference between the MT groups and basal growth medium group ( P > 0. 05) ( Fig. 1K ). The cell growth rate was relatively stable, and the growth of each group relative to the previous time point was significantly different (data shows in Dataset S3 ). These results indicate that MT has no apparent cytotoxicity. Physiological concentrations of MT affect hPDLSCs osteogenic differentiation The ARS experiments indicated that 1 pM–10 nM MT reduced the effect of the OS ( P < 0. 05, Figs. 2A, 2B ). We also found that treatment with MT at 1 pM, 0. 1 nM, and 10 nM caused a greater decrease in ALP activity than OS alone. Particularly, the highest decrease in ALP activity was observed in the 0. 1 nM MT-treated group ( P < 0. 05, Fig. 2C ). In contrast, treatment with the pharmacological MT concentration of 1 µM increased ALP activity, but there was no significant difference compared to OS alone. 10. 7717/peerj. 8663/fig-2 Figure 2 Effect of MT on osteogenic differentiation of hPDLSCs. (A) Gross appearance was observed with a camera; (B) Quantitative evaluation of Alizarin red staining in hPDLSCs. (C) ALP activity was measured. (D–E) The mRNA levels of OPN and CON were determined by qRT-PCR. (F–I) Effect of MT on osteogenic differentiation in protein levels of OCN and OPN in hPDLSCs. Representative immunoblots are shown (F) β -actin was used as the internal control. (G–H) Relative quantitative analysis of grey values of different proteins; the optical density was examined with ImageJ software. Each value is presented as the mean ± SD; n = 3; # P < 0. 05, ## P < 0. 01, ### P < 0. 001, respectively, vs. basal growth medium groups or between the two groups indicated by a solid black line. Next, we detected the expression levels of OPN and OCN, which are considered as closely associated with osteogenic differentiation, by qRT-PCR and western blotting after day 7 of osteogenic induction. The analysis showed that physiological MT treatment downregulated the osteogenic markers OCN and OPN ( Figs. 2D, 2E – 2H ). In the differentiated groups, 0. 1 nM MT showed the highest reduction compared to OS alone ( P < 0. 05), and the suppressive effect was abrogated at the lowest pharmacological concentration (1 µM). These results indicate that physiological concentrations of MT but not 1 µM reduced the differentiation-promoting effect of OS on hPDLSCs. Changes in mitochondrial respiratory function of hPDLSCs following osteogenic induction and MT treatment We detected the intracellular ROS activity of hPDLSCs by flow cytometry. We found that ROS generation in the osteogenic induction group was lower than that in basal growth medium group, whereas the 0. 1 nM MT-treated group showed higher levels than any other MT-treated group with OS ( P < 0. 05, Figs. 3A, 3B ). Furthermore, measurement of the intracellular ATP and NAD + /NADH levels showed that osteogenic differentiation increased the production of intracellular ATP, whereas 1 pM, 0. 1 nM and 10 nM MT reduced this effect ( P > 0. 05, Fig. 3C ). We also found that the NAD + /NADH ratio was reduced in the OS group ( P <0. 001, Fig. 3D ). Notably, the increase in the ratio was not concentration-dependent at physiological concentrations of MT. Furthermore, the 1 pM and 0. 1 nM MT-treated groups showed significant differences from the OS alone group ( P <0. 05). 10. 7717/peerj. 8663/fig-3 Figure 3 Effect of MT on mitochondrial respiratory function in hPDLSCs during osteogenic induction. (A) ROS production of hPDLSCs was measured by flow cytometry, and the mean fluorescence values for the vehicle control cells are indicated by vertical lines in each graph; (B) Quantitative evaluation of ROS in hPDLSCs; (C) Intracellular ATP content of hPDLSCs in each group was detected by chemiluminescence microplate reader; (D) NAD + /NADH ratio in each group was detected. Each value is presented as the mean ± SD; n = 3; # P < 0. 05, ## P < 0. 01, ### P < 0. 001, respectively, vs. basal growth medium or between the two groups indicated by the solid black line. Effects of MT on mitochondrial dynamics during osteogenesis The protein levels of mitochondrial dynamics were assessed by western blotting. The expression of MFF and P-MFF was initially increased and then decreased with increasing MT concentrations ( Figs. 4A, 4C, 4D ). In addition, the expression levels of MFN-2 and TOM20 first decreased and then increased with increasing MT concentrations ( Figs. 4A, 4B, 4E ). The expression of MFN-2 and TOM20 was significantly decreased at 0. 1 nM MT compared to in the OS alone group ( P <0. 01, P <0. 001, respectively, Figs. 4B, 4E ). We also found that at MT concentrations >0. 1 µM, the expression of MFN-2 and TOM20 was gradually increased, which continued up until 1 µM. Furthermore, the expression levels of MFN-2 and TOM20 at 1 µM MT were comparable to or even higher than those in the OS alone group ( Figs. 4B, 4E ). 10. 7717/peerj. 8663/fig-4 Figure 4 Effect of MT on mitochondrial dynamics within hPDLSCs osteogenic induction. (A–E) Effect of osteogenic medium on protein levels of MFN-2, P-MFF, MFF, and TOM20 in hPDLSCs. Representative immunoblots are shown (A), β-actin was used as the internal control; (B–E) Relative quantitative analysis of grey values of different proteins; the optical density was examined with ImageJ software. Each value is presented as the mean ± SD; n = 3; # P < 0. 05, ## P < 0. 01, ### P < 0. 001, respectively, vs. basal growth medium or between the two groups indicated by the solid black line. Discussion Studies have indicated that MT promotes the osteogenic differentiation of cells in vitro and in vivo at pharmacological concentrations ( Lee, Le & Kang, 2018 ; Li et al. , 2019 ). Our results showed that at physiological concentrations (1 pM–10 nM), MT decreased the expression of osteogenesis-related genes and mineralized nodules in hPDLSCs. In addition, we verified that at pharmacological concentrations, MT increased the mineralized nodules in hPDLCs ( Figs. S1A, S1B ), which is consistent with reported results ( Lee, Le & Kang, 2018 ). These findings indicate that MT exhibits dual effects: MT not only promoted the osteogenic differentiation of BMSCs, but also suppressed osteogenic differentiation. The normal PDL must to maintain certain periodontal ligament spaces and elasticity to prevent excessive hyperplasia of the alveolar bone and cementum. Failure to achieve this can lead to reduced periodontal ligament spaces and disappearance of the PDL ( Menicanin et al. , 2015 ). We speculated that the low concentrations of MT from blood circulation can reach the PDL, where it plays a role in balancing bone formation and absorption. One study showed that MT at pharmacological concentrations inhibits osteoclast differentiation ( Zhou et al. , 2017 ). A possible explanation for these results is that the action of MT is concentration- and cell type-dependent. The energy supply of MSCs shifted from glycolysis to aerobic metabolism during differentiation into osteoblasts ( Chen et al. , 2008 ). Hypoxia inhibited the osteogenesis of MSCs ( Pattappa et al. , 2011 ). Mitochondria play a key role in the osteogenesis of MSCs ( Zhang et al. , 2017 ). The mitochondria are best known for their ability to produce ATP in eukaryotic cells. NADH, which has an oxidation state of NAD +, is one of the major electron donors in the mitochondrial respiratory chain. NAD + /NADH represents the reductant-oxidant state in mitochondria, which is one of the best parameters for characterizing mitochondrial function ( Ni et al. , 2014 ). Our results revealed an increase in ATP levels and decrease in the NAD + /NADH ratio during hPDLSC differentiation, which was inhibited by physiological MT concentrations. However, at <10 nM, the weakening of cellular aerobic respiration was obvious compared to in the osteogenic induction group. Therefore, MT may affect aerobic respiratory function in the mitochondria. Mitochondria may be involved in both producing MT and blocking cytochrome c release, which drives the selective G protein- coupled receptor (GPCR) in the outer membrane ( Suofu et al. , 2017 ). Cellular oxidative phosphorylation is reflected by not only cellular ATP production, but also by ROS ( Zhang et al. , 2017 ). We observed a decrease in ROS levels in hPDLSCs following osteogenic induction, which is consistent with some previous studies ( Chen et al. , 2008 ; Zhang et al. , 2017 ). During mitochondrial function enhancement, ROS accumulation can lead to cell damage, which is not conducive to osteogenesis ( Tahara, Navarete & Kowaltowski, 2009 ). MT has been suggested to scavenge ROS ( Lee, Le & Kang, 2018 ); however, in this study, ROS levels were increased after MT treatment, indicating that MT does not play a scavenging role at physiological concentrations. Thus, the role of MT in ROS clearance remains to be confirmed. MFN2 is a mitochondrial fusion protein belonging to the transmembrane GTPases family and is conserved from yeast to humans. Mitochondrial fission accessory proteins such as MFF and P-MFF both play pivotal roles in mitochondrial fission ( Seo, Yoon & Do, 2018 ). The expression of MFN2 in this study was down-regulated, whereas that of P-MFF and MFF was increased by physiological concentrations of MT. This phenomenon is consistent with the down-regulation of osteogenic genes and changes in the mitochondrial function of hPDLSCs. Furthermore, our results are consistent with those of previous studies showing elevation of MFN2 expression with decreased P-MFF expression following enhanced cell differentiation ( McBride, Neuspiel & Wasiak, 2006 ). TOM20 is a protein encoded by TOMM20 in humans. This protein is anchored to the outer membrane by an N-terminal hydrophobic segment and exposes the receptor domain to the cytosol. Its main function is to induce cytosolic synthesis of mitochondrial proprotein recognition and transportation, and then promote the movement of proprotein to the TOM40 translocation pore with TOM22 ( Yamamoto et al. , 2011 ). We found that as the osteogenic capacity declined, the expression of TOM20 decreased. TOM20 in the outer mitochondrial membrane (OMM) influences the kinetics of material import into the mitochondria and thus contributes to importing malate dehydrogenase (MDH) ( Grey et al. , 2000 ). Inhibition of MDH is shown to inhibit NADH expression ( Yang et al. , 2018 ). It has been reported that the melatonin type 1 (MT 1 ) receptors are expressed in the OMM and mitochondria release MT, which is bound to high-affinity MT 1 located in OMM with its ligand-binding domain ( Suofu et al. , 2017 ). Therefore, we hypothesized that at physiological concentrations, MT may affect TOM20 expression and MT 1 receptor. TOM20 protein further affects MDH in the mitochondria, thereby decreasing NADH levels. Decreased expression of reducible NADH levels may increase ROS levels and thus decrease osteogenic differentiation in hPDLSCs. The potential regulation of TOM20 expression by physiological concentrations of MT and its effects on the mitochondrial dynamics of osteogenic differentiation in hPDLSCs require further evaluation. Conclusions Our study showed that administration of MT at physiological concentrations inhibited osteogenic differentiation and associated mitochondrial functions in hPDLSCs. Furthermore, increasing concentrations of MT weakened this inhibition and showed a tendency to promote osteogenic differentiation. At physiological concentrations, MT may also affecte the mitochondrial dynamics-related proteins TOM20, MFN2, MFF and P-MFF during cell differentiation. Our study indicates that MT exhibited dual effects that were concentration- and cell type-dependent. These findings provide insight into stem cell tissue engineering strategies and MT function. Supplemental Information 10. 7717/peerj. 8663/supp-1 Supplemental Information 1 The original figure of Click here for additional data file. 10. 7717/peerj. 8663/supp-2 Supplemental Information 2 The original figure of Click here for additional data file. 10. 7717/peerj. 8663/supp-3 Supplemental Information 3 The original figure of OPN protein band detected using an Image Quant Las 4, 000 mini imaging system shown in Fig. 2F Click here for additional data file. 10. 7717/peerj. 8663/supp-4 Supplemental Information 4 The original figure of MFF protein band detected using an Image Quant Las 4, 000 mini imaging system shown in Fig. 4A Click here for additional data file. 10. 7717/peerj. 8663/supp-5 Supplemental Information 5 The original figure of OCN protein band detected using an Image Quant Las 4, 000 mini imaging system shown in Fig. 2F Click here for additional data file. 10. 7717/peerj. 8663/supp-6 Supplemental Information 6 The original figure of P-MFF protein band detected using an Image Quant Las 4, 000 mini imaging system shown in Fig. 4A Click here for additional data file. 10. 7717/peerj. 8663/supp-7 Supplemental Information 7 The original figure of TOM 20 protein band detected using an Image Quant Las 4, 000 mini imaging system shown in Fig. 4A Click here for additional data file. 10. 7717/peerj. 8663/supp-8 Figure S1 The effect of of MT at pharmacological concentration on the osteogenic differentiation of hPDLSCs (A–B) Cells were treated with MT at pharmacological concentrations with osteogenic medium for 21 days; (A) Gross appearance is displayed by camera; (B). Quantitative evaluation of Alizarin red staining for hPDLSCs. (each value is presented as the mean ±SD; n = 3; # P < 0. 05, ## P < 0. 01, ### P < 0. 001, respectively, vs. basal growth medium group; ∗ P < 0. 05, ∗∗ P < 0. 01, ∗∗∗ P < 0. 001, respectively, vs. OS alone group). Click here for additional data file. 10. 7717/peerj. 8663/supp-9 Figure S2 TEM images of mitochondrial dynamics in hPDLSCs (A) bar: 0. 5 Click here for additional data file. 10. 7717/peerj. 8663/supp-10 Datasets S1 Band intensities of OPN These data were quantified by using ImageJ 1. 36b (NIH Freeware, USA). Click here for additional data file. 10. 7717/peerj. 8663/supp-11 Datasets S2 Band intensities of OCN These data were quantified by using ImageJ 1. 36b (NIH Freeware, USA) Click here for additional data file. 10. 7717/peerj. 8663/supp-12 Datasets S3 Band intensities of MFN2 These data were quantified by using ImageJ 1. 36b (NIH Freeware, USA). Click here for additional data file. 10. 7717/peerj. 8663/supp-13 Datasets S4 Band intensities of p-MFF These data were quantified by using ImageJ 1. 36b (NIH Freeware, USA). Click here for additional data file. 10. 7717/peerj. 8663/supp-14 Datasets S5 Band intensities of MFF These data were quantified by using ImageJ 1. 36b (NIH Freeware, USA). Click here for additional data file. 10. 7717/peerj. 8663/supp-15 Datasets S6 Band intensities of Tom20 These data were quantified by using ImageJ 1. 36b (NIH Freeware, USA). Click here for additional data file. 10. 7717/peerj. 8663/supp-16 Supplemental Information 8 Multiple Comparisons between each time point in different processing groups Click here for additional data file.

10. 7717/peerj. 8970

PeerJ

Exosomes derived from M0, M1 and M2 macrophages exert distinct influences on the proliferation and differentiation of mesenchymal stem cells

Background Different phenotypes of macrophages (M0, M1 and M2 Mφs) have been demonstrated to play distinct roles in regulating mesenchymal stem cells in various in vitro and in vivo systems. Our previous study also found that cell-conditioned medium (CM) derived from M1 Mφs supported the proliferation and adipogenic differentiation of bone marrow mesenchymal stem cells (BMMSCs), whereas CM derived from either M0 or M2 Mφs showed an enhanced effect on cell osteogenic differentiation. However, the underlying mechanism remains incompletely elucidated. Exosomes, as key components of Mφ-derived CM, have received increasing attention. Therefore, it is possible that exosomes may modulate the effect of Mφ-derived CM on the property of BMMSCs. This hypothesis was tested in the present study. Methods In this study, RAW264. 7 cells were induced toward M1 or M2 polarization with different cytokines, and exosomes were isolated from the unpolarized (M0) and polarized (M1 and M2) Mφs. Mouse BMMSCs were then cultured with normal complete medium or inductive medium supplemented with M0-Exos, M1-Exos or M2-Exos. Finally, the proliferation ability and the osteogenic, adipogenic and chondrogenic differentiation capacity of the BMMSCs were measured and analyzed. Results We found that only the medium containing M1-Exos, rather than M0-Exos or M2-Exos, supported cell proliferation and osteogenic and adipogenic differentiation. This was inconsistent with CM-based incubation. In addition, all three types of exosomes had a suppressive effect on chondrogenic differentiation. Conclusion Although our data demonstrated that exosomes and CM derived from the same phenotype of Mφs didn’t exert exactly the same cellular influences on the cocultured stem cells, it still confirmed the hypothesis that exosomes are key regulators during the modulation effect of Mφ-derived CM on BMMSC property.

Introduction During the past several decades, stem cell therapy has been the focus of tissue engineering and regenerative medicine ( Wei et al. , 2013 ; Thurairajah, Broadhead & Balogh, 2017 ). Attributed to their advantages, mesenchymal stem cells (MSCs) stand out from multiple stem cells and become the most promising choice for both autologous and allogeneic transplantation ( Maxson et al. , 2012 ; Wei et al. , 2013 ; Poltavtseva et al. , 2018 ). However, the application of MSCs from bench to bedside encounters many challenges, such as low cell dose, low survival rate and poor potency ( Silva et al. , 2018 ; Regmi et al. , 2019 ). Considerable efforts have been made to improve the regenerative efficiency of MSCs in vivo. Currently, the significance of macrophages (Mφs) in the recruitment and modulation of MSCs is well recognized ( Zhang et al. , 2017a ; Ma et al. , 2018 ; Cai et al. , 2018 ). Macrophages, essential components of innate immunity, play important roles in tissue regeneration ( Mosser & Edwards, 2008 ; Krzyszczyk et al. , 2018 ). In response to various stimuli, Mφs can switch phenotype from an unpolarized (M0) to a polarized (M1 and M2) state ( Mantovani et al. , 2013 ) and play unique roles in different stages of tissue healing ( Oishi & Manabe, 2018 ; Pajarinen et al. , 2019 ). Generally, M1 Mφs contribute to the debridement of wounds and exert pro-inflammatory functions. In contrast, M2 Mφs exert anti-inflammatory functions and facilitate tissue repair ( Murray et al. , 2014 ; Shapouri-Moghaddam et al. , 2018 ). During the past few years, accumulative studies have confirmed the modulating effect of Mφs on MSCs ( Yu et al. , 2016 ; Grotenhuis et al. , 2016 ). Our previous study also revealed that cell-conditioned medium (CM) generated by differently polarized Mφs exerted different influences on BMMSC cellular behaviors in vitro ( He et al. , 2018 ). However, the exact mechanism remains unclear. The dominant opinion is that cytokines are the main contributors to Mφ function ( Champagne et al. , 2002 ; Zhang et al. , 2017b ); however, Ekström et al. (2013) found a modulating effect of LPS-stimulated monocyte-derived exosomes on MSCs. Therefore, to better illustrate the mechanism of Mφ-MSC cross-talk, Mφ-derived exosomes should be considered. Exosomes are special endosomal-derived membranous microvesicles with a diameter of 50–150 nm. They are crucial mediators in intercellular communication and participate in many biological activities ( Zhang et al. , 2019 ). Exosomes, which carry proteins, lipids, nucleic acids and other cargos, can be released into the extracellular milieu and internalized by target cells, in which they modulate cellular behaviors ( Jan et al. , 2019 ). The biocompatibility, stability and capacity to transport bioactive components and overcome biological barriers indicate the great potential of exosomes as suitable therapeutic agents ( Liu & Su, 2019 ). For example, it was found that maturing osteoclast-derived exosomes transport RANK which can bind osteoblastic RANKL and promote bone formation ( Ikebuchi et al. , 2018 ). Actually, increasing evidence has suggested that Mφ-derived exosomes are important regulators in many physiological processes ( McDonald et al. , 2014 ; Saha et al. , 2016 ; Wei et al. , 2019 ). However, the role of different phenotypes of Mφ-derived exosomes in the regulation of MSC functions remains ambiguous. Based on these previous studies, we hypothesized that exosomes are key regulators during the modulation effect of Mφ-derived CM on bone marrow mesenchymal stem cells (BMMSCs). Our study also aims to further clarify the function of exosomes derived from different phenotypes of Mφs (M0-Exos, M1-Exos and M2-Exos) on the proliferation and differentiation of BMMSCs. The outcomes are expected to improve our understanding of Mφ-MSC cross-talk and contribute to better modulation of MSC potency in tissue regeneration. Materials and Methods Isolation and culture of mouse BMMSCs Male C57BL/6 mice (6–8 weeks) were purchased from the Laboratory Animal Research Centre of the Fourth Military Medical University. Animals used in this study were approved by the Animal Use and Care Committee of the Fourth Military Medical University (IACUC-20180804). According to previously reported methods ( Huang et al. , 2015 ), the mice were sacrificed by cervical dislocation and the femurs and tibias dissected. After two washes with phosphate buffer solution (PBS), bone marrow cells were flushed from the bones into 10-cm culture dishes using alpha-minimal essential medium (α-MEM; Invitrogen, Carlsbad, CA, USA) supplemented with 20% fetal bovine serum (FBS, Hangzhou Sijiqing Biological Engineering Materials, Zhejiang, China) and 1% penicillin and streptomycin (Sigma–Aldrich, St. Louis, MO, USA). Then, the dishes were incubated at 37 °C in a 5% CO 2 incubator. The medium was refreshed every 3 days to remove nonadherent cells. When the primary cells reached 70–90% confluence, they were digested with 0. 25% trypsin (Invitrogen, Carlsbad, CA, USA) and passaged. Cells from the 2nd or 3rd passage were used in the following experiments. Identification of mouse BMMSCs The isolated primary BMMSCs were subjected to flow cytometry analysis, colony-forming assay, EdU (5-ethynyl-2′-deoxyuridine) incorporation assay, cell counting kit-8 (CCK-8) assay, Alizarin red S staining, Oil red O staining and Alcian blue staining to identify the phenotype. The results were showed in the Supplemental File. The flow cytometry analysis revealed that these cells were strongly positive for MSC markers, such as CD105, Sca-1, CD73 and CD90, but were negative for hematopoiesis-related markers, CD34 and CD45 ( Fig. S1A ). Toluidine blue staining revealed that these cells possessed the ability to form new colony units ( Fig. S1B ). Furthermore, these cells exhibited decent proliferative potential, as evidenced by the results of the EdU and CCK-8 assays ( Figs. S1C and S1D ). Alizarin red S staining showed that mineralized nodules formed after osteogenic induction ( Fig. S1E ). Oil red O staining showed the formation of lipid-rich vacuoles after adipogenic induction ( Fig. S1F ), and Alcian blue staining showed acidic proteoglycan formation after chondrogenic induction ( Fig. S1G ). All of these observations confirmed the multipotent differentiation ability of the BMMSCs. Culture of the Mφs and Mφ polarization The mouse Mφ cell line RAW264. 7 (ATCC Cat# TIB-71, RRID: CVCL_0493) was used in the present study and cultured in α-MEM supplemented with 10% FBS. For each sample, 2 × 10 6 cells were seeded into a 10-cm culture dish. As previously reported ( He et al. , 2018 ), lipopolysaccharide (LPS) at a concentration of 200 ng/ml plus interferon-gamma (IFN-γ) at a concentration of 10 ng/ml was used to induce Mφ polarization into the M1 phenotype, while IL-4 at a concentration of 20 ng/ml was used to induce Mφs toward the M2 polarization. All cytokines were purchased from PeproTech, Princeton, NJ, USA. As a control, RAW264. 7 cells incubated in medium supplemented with PBS were considered M0 Mφs. Following a 24-h induction, the phenotypes of polarized cells (as stimulated by LPS plus IFN-γ or IL-4) and the unpolarized cells (PBS) were identified by flow cytometry analysis, quantitative real-time polymerase chain reaction (qRT-PCR) and enzyme-linked immunosorbent assay (ELISA). Isolation and characterization of M0, M1 and M2 Mφ-derived exosomes To isolate exosomes from the M0, M1 and M2 Mφs, exosome-depleted FBS was generated by ultracentrifugation of FBS at 100, 000× g for 70 min with an L-80 ultracentrifuge (45 Ti rotor) from Beckman Coulter (Brea, CA, USA) ( Hoshino et al. , 2015 ), which removed the bovine exosomes from the FBS. Although we do not have nanoparticle tracking data of FBS before and after ultracentrifugation, we trust that FBS was effectively depleted of exosomes since other studies used similar methods ( Kobayashi et al. , 2014 ; Lee et al. , 2019 ). Following a 24-h incubation with or without M1/M2 induction, the culture medium was discarded, and the cells were washed twice with PBS to remove remaining cytokines. Then, α-MEM supplemented with 10% exosome-depleted FBS was used to further culture the cells. After 24 h, the CM generated by the M0, M1 or M2 Mφs (termed CM0, CM1 and CM2) was collected separately. Each CM sample was first centrifuged at 2, 000× g (30 min at 4 °C) to remove cells and debris. After the CM was transferred into a new tube, 0. 5 volume of the total exosome isolation (from cell culture medium) reagent (Invitrogen, Carlsbad, CA, USA) was added to each CM supernatant. Then, the CM/reagent mixtures (CM0, CM1 and CM2) were vortexed and incubated at 4 °C overnight as described in the instructions. Finally, each mixture was centrifuged at 10, 000× g for 60 min at 4 °C. Following the removal of the supernatant, the exosomes at the bottom of each tube (M0-Exos, M1-Exos or M2-Exos) were resuspended with PBS (exosomes isolated from 1 ml of CM of each sample were suspended in 100 μl of PBS). Transmission electron microscopy Freshly isolated exosomes were dropped on special copper grids, where they dried at room temperature. Then, the samples were subjected to negative staining with 1% aqueous uranyl acetate for 5 min and washed twice with deionized water. The grids were dried at room temperature before TEM analysis. The samples were visualized with a JEM-1400Plus transmission electron microscope from JEOL (Tokyo, Japan). Nanoparticle tracking analysis The M0-Exos, M1-Exos and M2-Exos were sent to a company (Wuhan GeneCreate Biological Engineering Co. , Ltd. , Wuhan, China) for nanoparticle tracking analysis. In brief, the exosome samples were diluted to an optimal concentration, and the size and number were determined with a NanoSight NS 300 system (NanoSight Technology, Malvern, UK). Western blotting analysis Western blot assays were performed to measure the exosome surface markers CD9, CD63, CD81 and Alix on M0-Exos, M1-Exos and M2-Exos. The exact procedures of the Western blot assay were previously described ( Xu et al. , 2019 ). The primary antibodies used in this study were anti-mouse CD63 antibody (Abcam, Cambridge, UK), anti-mouse CD81 antibody (Cell Signaling Technology, Danvers, MA, USA), anti-mouse CD9 antibody (Abcam, Cambridge, UK) and anti-mouse Alix antibody (Cell Signaling Technology, Danvers, MA, USA). Horseradish peroxidase (HRP)-conjugated goat anti-rabbit and goat anti-mouse (Cell CWBIO) secondary antibodies were used. Internalization of the exosomes by BMMSCs PKH67 fluorescent cell linker kits (Sigma–Aldrich, St. Louis, MI, USA) were used to label the exosomes according to the manufacturer’s instructions. Then, the mixture of the exosomes and PKH67 dye was subjected to exosome spin columns (MW3000) (Invitrogen, Carlsbad, CA, USA) to remove excess dye. Finally, the PKH67-labeled exosomes were incubated with BMMSCs at 37 °C for 4 h. A confocal laser microscope (FV1000; Olympus, Tokyo, Japan) was used to observe the uptake of the exosomes by the BMMSCs. Cell treatment and grouping To investigate the effects of M0, M1 and M2 Mφ-derived exosomes on the proliferation and differentiation of BMMSCs, the M0-Exos, M1-Exos or M2-Exos at a concentration of 100 μl/ml were supplemented into the complete medium or inductive medium used to culture BMMSCs, respectively. For each group, the culture medium was refreshed every other day and exosomes were added at the same time. The culture medium supplemented with the same volume of PBS were used as blank control. Effects of Mφ-derived exosomes on BMMSC proliferation The proliferation ability of the BMMSCs cultured in different complete medium (Supplemented with M0-Exos, M1-Exos, M2-Exos or PBS) was determined on the basis of cell colony-forming, CCK-8 and EdU incorporation assays. Effects of Mφ-derived exosomes on BMMSC osteogenic differentiation Osteogenic medium was generated by complete medium supplemented with 50 μg/ml vitamin C, 10 nM dexamethasone and 10 mM β-glycerophosphate (all purchased from Sigma–Aldrich, St. Louis, MO, USA). To assess the osteogenic potency of the BMMSCs, the cells were seeded in 12-well culture plates at a density of 2 × 10 5 cells/well. When the cells reached 80–90% confluence, the culture medium was replaced with different osteogenic medium (Supplemented with M0-Exos, M1-Exos, M2-Exos or PBS). After 7 days of osteogenic induction, the CM of each group was collected to measure the ALP activity with an alkaline phosphatase assay kit (Jiancheng Bioengineering, Nanjing, China). The cells were fixed with 4% paraformaldehyde for 30 min and stained with a BCIP/NBT alkaline phosphatase color development kit (Beyotime Institute of Biotechnology, Haimen, China) as described in the instructions. The expression level of osteogenesis-related genes ( ALP, BMP-2, COL-1, OCN and Runx2 ) were assessed by qRT-PCR analysis. In addition, the Alizarin red S staining were also conducted after induction for 14 days. Effects of Mφ-derived exosomes on BMMSC adipogenic differentiation The adipogenic induction medium was generated by complete medium supplemented with 0. 5 mM 3-isobutyl-1-methylxanthine, 1 μM dexamethasone, 0. 1 mM indomethacin and 10 μg/ml insulin. To detect the adipogenesis ability of the BMMSCs, the cells were seeded in 12-well culture plates at a density of 2 × 10 5 cells/well. When the cells reached 80–90% confluence, the culture medium was replaced with different adipogenic induction medium. After induction for 7 days, the adipogenic differentiation of the BMMSCs were determined with Oil red O staining and qRT-PCR assay. Effects of Mφ-derived exosomes on BMMSC chondrogenic differentiation To assess the chondrogenic differentiation ability of the BMMSCs in response to various exosomes, the cells were seeded in 12-well culture plates at a density of 2 × 10 5 cells/well. When the cells reached 80–90% confluence, the complete medium was replaced with chondrogenic differentiation medium (Cyagen Biosciences, Inc. , Guangzhou, China) supplement with M0-Exos, M1-Exos, M2-Exos or PBS. After chondrogenic induction for 7 days, the chondrogenic differentiation ability were analyzed by Alcian blue staining and qRT-PCR assay. Flow cytometry analysis The cell surface markers of the BMMSCs and Mφs were analyzed by flow cytometry. Briefly, the cells were trypsinized and washed with PBS. To block Fc receptors, the cells were incubated with 2% anti-mouse CD16/32 (BioLegend, San Diego, CA, USA) on ice for 10 min. Then, the cells were washed twice with PBS and incubated with specific antibodies for 30 min at 4 °C in the dark. Excess antibody was removed by washing the cells with PBS. Untreated cells were used as blank controls. The samples were then analyzed with a Beckman Coulter Epics XL cytometer (Beckman Coulter, Fullerton, CA, USA). For the characterization of BMMSCs, the following antibodies were used: PE anti-mouse Ly-6A/E (Sca-1), PE anti-mouse CD90. 2, PE-anti-mouse CD73, PE anti-mouse CD105, PE anti-mouse CD34 and FITC anti-mouse CD45 (all from BioLegend, San Diego, CA, USA). FITC anti-mouse CD86 and PE anti-mouse CD206 (both from BioLegend, San Diego, CA, USA) were used for the identification of Mφs. In this experiment, pulse width measurements were used to eliminate the possibility that the detected cells were doublets of RAW 267 cells. 7-aminoactinomycin D (7-AAD), as well as isotype controls were used to exclude dead cells and the nonspecific binding of the monoclonal antibodies. Each group with no less than 10 6 cells was gated for flow cytometric analysis. Quantitative real-time polymerase chain reaction To measure the mRNA expression levels, qRT-PCR was conducted. Total RNA from the cultured cells was extracted with TRIzol reagent (Invitrogen, Carlsbad, CA, USA). According to the manufacturer’s instructions, cDNA was synthesized using PrimeScript™ RT Master Mix (Perfect Real-Time; TaKaRa, Tokyo, Japan). Then, qRT-PCR was performed using SYBR ® Premix Ex Taq™ II (Tli RNaseH Plus; TaKaRa, Tokyo, Japan). The primers used are listed in Table 1. The β -actin housekeeping gene was used to normalize the expression level of the related genes. 10. 7717/peerj. 8970/table-1 Table 1 Primer sequences for quantitative real-time polymerase chain reaction (qRT-PCR) analysis. Primer Full name Sequence (5′–3′) IL-1 β Interleukin 1- β Forward AAGGAGAACCAAGCAACGACAAAA Reverse TGGGGAACTCTGCAGACTCAAACT iNOS Inducible nitric oxide synthase Forward CAAGCTGAACTTGAGCGAGGA Reverse TTTACTCAGTGCCAGAAGCTGGA TNF- α Tumor necrosis factor- α Forward TATGGCCCAGACCCTCACA Reverse GGAGTAGACAAGGTACAACCCATC Arg-1 Arginine-1 Forward AGCTCTGGGAATCTGCATGG Reverse ATGTACACGATGTCTTTGGCAGATA CD206 CD206 Forward AGCTTCATCTTCGGGCCTTTG Reverse GGTGACCACTCCTGCTGCTTTAG IL-10 Interleukin 10 Forward GCCAGAGCCACATGCTCCTA Reverse GATAAGGCTTGGCAACCCAAGTAA ALP Alkaline phosphatase Forward CTTCTTGCTGGTGGAAGGA Reverse AAAACGTGGGAATGATCAGC BMP-2 Bone morphogenetic protein 2 Forward TGACTGGATCGTGGCACCTC Reverse CAGAGTCTGCACTATGGCATGGTTA COL-1 Collagen-1 Forward GCTGGAGTTTCCGTGCCT Reverse GACCTCGGGGACCCATTG Runx2 Runt-related transcription factor-2 Forward AGGGAATAGAGGGGATGCATTAG Reverse AAGGGAGGACAGAGGGAAACA OCN Osteocalcin Forward CTGACAAAGCCTTCATGTCCAA Reverse GCGCCGGAGTCTGTTCACTA Adiponectin Adiponectin Forward TTCTGTCTGTACGATTGTCAGTGGA Reverse GGCATGACTGGGCAGGATTA PPAR- γ Peroxisome proliferator activated receptor- γ2 Forward TCAGGTTTGGGCGGATG Reverse CAGCGGGAAGGACTTTATGTATG Col-2a1 Collagen type II α1 Forward CTGACCTGACCTGATGATACC Reverse CACCAGATAGTTCCTGTCTCC Cdh2 Cadherin 2 Forward CCGTGAATGGGCAGATCACT Reverse TAGGCGGGATTCCATTGTCA Sox9 SRY (sex determining region Y)-box 9 Forward TACGACTGGACGCTGGTGCC Reverse CCGTTCTTCACCGACTTCCTCC β -actin β -actin Forward CTCTTTTCCAGCCTTCCTTCTT Reverse GAGGTCTTTACGGATGTCAACG Enzyme-linked immunosorbent assay After induction for 24 h, the polarized cells (M1 and M2 Mφs) were washed with PBS to remove remaining cytokines and then were cultured with fresh medium for another 24 h. Then, the culture media of M0, M1 and M2 Mφs were collected and centrifuged to remove the cells and debris. The concentrations of two different cytokines (TNFα and IL-10) secreted into the CM were then detected with ELISA kits (Neobioscience, Guangzhou, China) according to the manufacturer’s instructions. Colony-forming assay Briefly, 800 cells were seeded in a 60-mM culture dish. After 14 days, the cells were washed with PBS and fixed for 30 min in 4% paraformaldehyde. Then, the cells were stained with 1% toluidine blue for 20 min at room temperature. After the cells were washed three times with PBS, photos were taken of these dishes, and the colony-forming units (CFUs) were counted under a microscope. Each CFU with ≥50 cells was quantified for statistical analysis. EdU (5-ethynyl-2′-deoxyuridine) incorporation assay Bone marrow mesenchymal stem cells were seeded in a 12-well culture plate at a density of 2 × 10 5 cells per well. When the cells reached 70–80% confluence, the EdU incorporation assay was performed according to the manufacturer’s protocol with a BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 594 (Beyotime Institute of Biotechnology, Haimen, China). The cells were visualized with a LeicaTCS SP5 X confocal microscope (Leica, Germany) and photographed. Cell counting kit-8 assay Bone marrow mesenchymal stem cells were seeded in a 96-well culture plate at a density of 3, 000 cells per well, and the culture medium was refreshed every other day. Every day at the same time-point for 7 days, a CCK-8 assay was performed with a Cell Counting Kit-8 (Beyotime Institute of Biotechnology, Haimen, China). Briefly, the medium in each test well was replaced with 180 μl of fresh medium supplemented with 20 μl of CCK-8 reagent and incubated at 37 °C for 3 h in the dark. Then, the medium was transferred to a new 96-well plate, and the absorbance was measured at 450 nm with a microplate reader (EL×800; BioTek Instruments Inc. , Highland Park, FL, USA). Alizarin red S staining After induction for 14 days, the BMMSCs were washed with PBS and fixed in 4% paraformaldehyde for 30 min. Then, the cells were stained with Alizarin red S for 30 min. To remove excess staining solution, the cells were washed with PBS three times, and the stained samples were observed and photographed with an inverted microscope (Olympus, Shinjuku City, Tokyo, Japan). Quantitative analysis was conducted by dissolving the stained samples into 2% cetylpyridinium chloride and measuring the OD values of the solutions at 560 nm. Oil red O staining After 7 days of adipogenic induction, the cells were fixed with 4% paraformaldehyde and stained with Oil red O for 30 min. The stained cells were observed and photographed under a microscope. For quantification, the stained samples were dissolved with isopropanol and the OD values of the solutions were measured at 560 nm. Alcian blue staining After chondrogenic induction for 7 days, the cells were fixed with 4% paraformaldehyde and stained with Alcian blue for 20 min. Then, the stained cells were observed and photographed. Statistical analysis All data were collected from at least three independent experiments (biological replicates) with three repeats (technique repeats). We combined the quantitative data from three biological replicates and analyzed the results with GraphPad Prism 5 software (San Diego, CA, USA). Statistical significance between groups was determined by one-way analysis of variance (ANOVA) and Tukey’s post-hoc test. Data are presented as the mean ± standard deviation (SD). Values of p < 0. 05 were considered statistically significant. Results Identification of the Mφ phenotypes After stimulation with different cytokines, the expression of the cell surface markers in Mφs with different phenotypes was analyzed by flow cytometry. Compared to PBS- and IL-4-treated Mφs, the cells treated with LPS plus IFN-γ showed significant upregulation of CD86 (specific surface marker of M1 Mφs, Figs. 1A – 1C ). Cells treated with IL-4 displayed higher expression of CD206 (specific surface marker of M2 Mφs) than did cells treated with PBS or LPS plus IFN-γ ( Figs. 1D – 1F ). The mRNA expression levels of specific genes known as M1 and M2 markers ( Murray & Wynn, 2011 ) were quantified by qRT-PCR. Compared with cells treated with PBS or IL-4, the cells treated with LPS plus IFN-γ had significantly upregulated expression levels of IL-1 β, iNOS and TNF- α (M1-specific markers) ( Figs. 1G – 1I ; p < 0. 01 or 0. 001). The levels of Arg-1, CD206 and IL-10 (M2-specific markers) in the cells treated with IL-4 were obviously higher than they were in the cells stimulated with PBS or LPS plus IFN-γ ( Figs. 1J – 1L ; p < 0. 01 or 0. 001). In addition, the expression levels of IL-1 β, iNOS and TNF- α in PBS- and IL-4-treated cells showed no significant difference, and there was no significant difference in the expression levels of Arg-1, CD206 or IL-10 between the cells treated with PBS and those treated with LPS plus IFN-γ ( Figs. 1G – 1L ). The ELISA data revealed that the concentration of TNF-α in the CM generated by the LPS plus IFN-γ treated cells was significantly higher than it was in the CM treated by PBS and IL-4 ( Fig. 1M ; p < 0. 001), a finding that was consistent with the results of PCR. Similarly, the concentration of IL-10 in the CM generated by IL-4-treated cells was significantly higher than it was in the CM for each of the other two groups ( Fig. 1N ; p < 0. 001). All of these results suggested that the RAW264. 7 cells were successfully induced to M1 polarization by LPS plus IFN-γ and to M2 polarization by IL-4. 10. 7717/peerj. 8970/fig-1 Figure 1 Identification of the macrophage phenotypes following stimulation with LPS plus IFN-γ (M1 induction) or IL-4 (M2 induction); unpolarized cells (incubated with medium supplemented with PBS) are considered M0 cells. (A–F) Results from the flow cytometry analysis of CD86 (M1 marker) and CD206 (M2 marker) in LPS plus IFN-γ-stimulated, IL-4-stimulated or unstimulated (PBS) cells. (G–L) Gene expression levels in LPS plus IFN-γ-stimulated, IL-4-stimulated or unstimulated (PBS) cells detected by qRT-PCR assay. IL- 1β, iNOS and TNF- α were used as M1-related markers, while Arg-1, CD206 and IL-10 were applied as M2-related markers (values were normalized to β -actin and relative to PBS group (unstimulated cells)). (M and N) ELISA results of cytokine levels in the supernatants generated by LPS plus IFN-γ-stimulated, IL-4-stimulated or unstimulated (PBS) cells. TNF-α was used as an M1-polarization marker, and IL-10 was used as an M2-polarization marker. Data are presented as the mean ± SD for n = 3; ** p < 0. 01 and *** p < 0. 001 indicate significant differences between the indicated columns. Characterization of different Mφ-derived exosomes The exosomes were isolated from the CM generated by M0, M1 or M2 Mφs. The images viewed by transmission electron microscopy showed that exosomes released from Mφs were small round nanometer-sized particles with bilayer membranes ( Fig. 2A ). Nanoparticle tracking analysis revealed that the diameters of these exosomes ranged from 30 to 150 nm ( Fig. 2B ). Moreover, the M0-Exos, M1-Exos and M2-Exos all expressed the exosomal markers CD9, CD63, CD81 and Alix ( Fig. 2C ). These all confirmed the successful extraction of exosomes. 10. 7717/peerj. 8970/fig-2 Figure 2 Identification of the exosomes derived from the polarized macrophages (M1-Exos and M2-Exos) or the unpolarized macrophages (M0-Exos). (A) Representative TEM images of the M0-Exos, M1-Exos and M2-Exos (scale bar: 100 nm). (B) Size distribution profiles of the M0-Exos, M1-Exos and M2-Exos, as determined by nanoparticle tracking analysis. (C) The presence of exosome marker proteins (CD81, CD63, CD9 and Alix) in the M0-Exos, M1-Exos and M2-Exos (Western blot assay results). (D) Representative confocal images showing exosomes endocytosed by the BMMSCs (the BMMSCs were incubated with PKH67-labeled M0-Exos, M1-Exos or M2-Exos for 4 h; the cells cultured without exosomes served as the blank control; scale bar: 10 μm): green, the PKH67-labeled exosomes; red (phalloidin), the framework of the BMMSCs. TEM, transmission electron microscopy. Internalization of the exosomes by BMMSCs To visualize whether M0-Exos, M1-Exos and M2-Exos can be internalized by BMMSCs, PKH67-labeled exosomes were incubated with BMMSCs. The endocytosis of the exosomes was observed under a confocal microscope. The green fluorescence of the PKH67-labeled exosomes could be observed in the cells cocultured with exosomes, while the cells not cocultured with exosomes presented only the blue DAPI fluorescence and the red phalloidin fluorescence ( Fig. 2D ). These demonstrated that M0-Exos, M1-Exos and M2-Exos all could be untaken and internalized by BMMSCs. Effects of Mφ-derived exosomes on the proliferation of the BMMSCs To detect the proliferation ability of BMMSCs in different culture conditions, colony-forming assay was conducted firstly. Toluidine blue staining revealed that all of the groups had the ability to form new colony units ( Figs. 3A and 3B ). The data from the quantified analysis of CFU numbers showed that the supplement of M1-Exos promoted the BMMSCs to form the most CFUs ( p < 0. 01 or 0. 001). However, M2-Exos obviously decreased the number of CFUs formed by the BMMSCs as compared with that of control ( p < 0. 01). The supplementation of M0-Exos did not have a significant influence on CFU formation of the BMMSCs ( Fig. 3C ). The data from CCK-8 assays conducted during the 7-day time course indicated no significant difference between different groups. However, the trend of the cell growth histograms was consistent with that of the colony-forming assays ( Fig. 3D ). The effects of exosomes on BMMSC proliferation were further confirmed by EdU incorporation assays. It is obvious that cells cultured with M1-Exos have the most EdU-positive cells, suggesting that cells in this group have the greatest proliferative potential. While cells cultured with M0-Exos and M2-Exos showed fewer EdU-positive cells than did the control group, revealing low proliferative potential ( Fig. 3E ). 10. 7717/peerj. 8970/fig-3 Figure 3 Cell proliferation in response to exosome-based incubation (control: cells in normal culture); BMMSCs were incubated in normal medium supplemented with exosomes (M0-Exos, M1-Exos or M2-Exos). (A) Representative general images of the colony formed by the BMMSCs after 14 days in various cultures. (B) A single colony formed by the BMMSCs (scale bar: 250 μm). (C) Results from the quantitative analysis of CFUs. (D) Proliferative potential of the BMMSCs in different cultures (during a 7-day incubation period) in terms of the CCK-8 assay results. (E) Proliferative potential of the BMMSCs in different cultures in terms of the EdU incorporation assay results; representative images showing EdU-positive cells (labeled with red fluorescence; scale bar: 100 μm). Data are presented as the mean ± SD for n = 3; ** p < 0. 01 and *** p < 0. 001 indicate significant differences between the indicated columns. Effects of Mφ-derived exosomes on the osteogenic differentiation of the BMMSCs To analyze the osteogenic differentiation ability of the BMMSCs under different culture conditions, Alizarin red S staining was conducted on day 14. It can be found that BMMSCs cultured with supplement of M1-Exos appeared to form the most calcium deposits. However, the cells cultured with supplement of M0-Exos and M2-Exos formed fewer calcium deposits than that of the control ( Figs. 4A and 4B ). The results of quantitative analysis also indicated that M1-Exos significantly promoted the mineralized nodule formation of BMMSCs ( p < 0. 001), while M0-Exos and M2-Exos showed no significant influence on the mineralized nodule formation of BMMSCs as compared with the control group ( Fig. 4C ). These were further demonstrated by ALP staining ( Figs. 4D and 4E ) and the detection of ALP activity on day 7. The quantitative analysis of ALP activity revealed that M0-Exos, M1-Exos and M2-Exos all increased the ALP activity of BMMSCs ( Fig. 4F ; p < 0. 01 or 0. 001), and it can be found that cells cultured with supplement of M1-Exos exhibited the highest ALP activity ( p < 0. 001). The mRNA expression levels of osteogenesis-related genes were analyzed by qRT-PCR after cells were subjected to osteogenic induction for 7 days. The data showed that neither M0-Exos or M2-Exos significantly influenced the expression levels of ALP, COL-1, OCN and Runx2 but reduced the expression level of BMP-2 ( p < 0. 05 or 0. 01). While the expression levels of both ALP and Runx2 in the cells cultured with M1-Exos were increased ( p < 0. 05 or 0. 01). In addition, the expression levels of ALP, BMP-2, OCN and Runx2 in the cells cultured with M1-Exos were higher than those of cells cultured in M0-Exos or M2-Exos ( Figs. 4G – 4K ; p < 0. 05 or 0. 01). 10. 7717/peerj. 8970/fig-4 Figure 4 Osteogenic potential of BMMSCs in response to exosome-based incubation (control: cells in normal osteogenic cultures); BMMSCs were incubated in normal osteogenic medium supplemented with exosomes (M0-Exos, M1-Exos or M2-Exos). (A) General view of Alizarin red S stained cells after 14 days of osteogenic induction. (B) Representative images of Alizarin red S staining captured under the microscope (scale bar: 250 μm). (C) Quantitative analysis results of mineralized nodules. (D) General view of ALP staining after 7 days of osteogenic induction. (E) Representative images of ALP staining captured under the microscope (scale bar: 250 μm). (F) Statistical analysis results of the ALP activity. (G–K) Expression levels of osteogenesis-related genes ( ALP, BMP-2, COL-1, OCN and Runx2 ) in the BMMSCs (qRT-PCR assay). Values were normalized to the level of β -Actin. Data are presented as the mean ± SD; n = 3; * p < 0. 05, ** p < 0. 01 and *** p < 0. 001 indicate significant differences between the indicated columns. Effects of Mφ-derived exosomes on the adipogenic differentiation of the BMMSCs To investigate the adipogenic differentiation ability of BMMSCs under different culture conditions, after induction for 7 days, the cells were stained with Oil red O. The results revealed that the cells treated with M1-Exos formed the most lipid droplets, while both the M0-Exo- and M2-Exo-incubated cells formed fewer lipid droplets than were found in the control group ( Figs. 5A and 5B ). The quantitative analysis data were consistent with the staining results. Both M0-Exos and M2-Exos had negative effects on the lipid droplet formation of the BMMSCs ( p < 0. 001 or 0. 05). While M1-Exos significantly promoted more lipid droplets to form in the BMMSCs ( Fig. 5C ; p < 0. 001). The effects of exosomes on the adipogenic differentiation of BMMSCs were further confirmed by qRT-PCR analysis. It can be found that M1-Exos dramatically upregulated the adiponectin and PPAR- γ gene (adipogenesis-related markers) expression of BMMSCs ( p < 0. 01), while cells cultured with M0-Exos and M2-Exos showed no difference in adiponectin and PPAR- γ gene expression compared with that of the control cells. Moreover, the PPAR- γ expression level in the cells cultured with M1-Exos was also higher than that of the cells cultured with M0-Exos or M2-Exos ( Figs. 5D and 5E ; p < 0. 05 or 0. 01). 10. 7717/peerj. 8970/fig-5 Figure 5 Adipogenic potential of the BMMSCs in response to exosome-based incubation (control: cells in normal adipogenic cultures); BMMSCs were incubated in normal adipogenic medium supplemented with exosomes (M0-Exos, M1-Exos or M2-Exos). (A) General view of the Oil red O-stained lipid droplets after 7 days of adipogenic induction. (B) Representative images of Oil red O staining captured under the microscope (scale bar: 100 μm). (C) Quantitative analysis results of the lipid droplets. (D and E) Relative mRNA expression levels of adipogenesis-related genes ( adiponectin and PPAR- γ) in the BMMSCs (qRT-PCR assay results). Values were normalized to β- Actin and relative to the level of the control. Data are presented as the mean ± SD; n = 3; * p < 0. 05, ** p < 0. 01 and *** p < 0. 001 indicate significant differences between the indicated columns. Effects of the Mφ-derived exosomes on the chondrogenic differentiation of the BMMSCs The chondrogenesis ability of the BMMSCs in different cultures was detected by Alcian blue staining. It is obvious that cells cultured in normal culture medium formed more significant and larger multilayered aggregates than did the other groups of cells. However, the cells cultured in M1-Exos only formed small monolayered aggregates ( Figs. 6A and 6B ). The expression level of chondrogenesis-related genes Cdh2 (cadherin 2), Col2-a1 (collagen II-encoding gene) and Sox9 (SRY (sex-determining region Y)-box 9) were also analyzed by qRT-PCR. It was discovered that M0-Exos, M1-Exos and M2-Exos downregulated the expression of chondrogenesis-related genes. Concretely, both M0-Exos and M2-Exos had a negative effect on the expression levels of Cdh2 and Col2-a1 ( p < 0. 001, 0. 01 or 0. 05) but showed no significant influence on Sox9 expression. M1-Exos significantly downregulated the expression level of all the detected chondrogenesis-related genes compared with the level expressed in the control group cells ( Figs. 6C – 6E ; p < 0. 001 or 0. 05). These results indicate that Mφ-derived exosomes have a negative effect on BMMSC chondrogenic differentiation. 10. 7717/peerj. 8970/fig-6 Figure 6 Chondrogenic potential of the BMMSCs in response to exosome-based incubation (control: cells in normal chondrogenic cultures); BMMSCs were incubated in normal chondrogenic medium supplemented with exosomes (M0-Exos, M1-Exos or M2-Exos). (A) General view of Alcian blue staining after 7 days of chondrogenic induction. (B) Representative images of Alcian blue staining captured under the microscope (scale bar: 250 μm). (C–E) Expression levels of chondrogenesis-related genes ( Cdh2, Col-2a1 and Sox9 ) in the BMMSCs after 7 days of chondrogenic induction (qRT-PCR assay). Data are presented as the mean ± SD; n = 3; * p < 0. 05, ** p < 0. 01 and *** p < 0. 001 indicate significant differences between the indicated columns. Discussion Mφs are key mediators of host defense and participate in a range of physiological process ( Oishi & Manabe, 2018 ). They are heterogenous cells which can influence the microenvironment through their polarization into different phenotypes ( Brown, Sicari & Badylak, 2014 ). Over the last few years, the importance of Mφs in stem cell survival and tissue repair has been recognized ( Cai et al. , 2018 ). Since Mφs can switch their phenotypes during the process of tissue regeneration ( Novak & Koh, 2013 ), it’s essential to explore the regulating effect of different phenotypes of Mφs on MSCs. In this study, RAW264. 7 cells were stimulated with different cytokines and different methods were used to identified their phenotypes. The results of flow cytometry analysis, ELISA and qRT-PCR all revealed that RAW264. 7 cells were successfully polarized to M1phenotype with the stimulation of LPS plus IFN-γ or polarized to M2 phenotype with the stimulation of IL-4. This was consistent with the results of our previous study ( He et al. , 2018 ). Recently, several studies have confirmed the regulation effect of macrophage-derived exosomes. For example, it was reported that Mφ-derived extracellular vesicles (EVs) are essential for intestinal stem cell self-renewal, proliferation and intestinal homoeostasis ( Saha et al. , 2016 ). Another study established that Mφ-derived exosomes can accelerate wound repair by inducing endothelial cell proliferation and migration ( Li et al. , 2019 ). However, it still remains unclear whether Mφ-derived exosomes also play a role in the regulation of BMMSC property. To test this assumption, exosomes were isolated from the CM of M0, M1 or M2 Mφs, separately. The results of TEM and NTA showed that the size and morphology of these isolated exosomes meet the standards mentioned in the literature ( Kalluri & LeBleu, 2020 ) and the exosmal markers are also positive in these exosomes. This demonstrated that exosomes were successfully isolated from M0, M1 and M2 Mφs. In addition, it was found that the exosomes secreted by different phenotypes of Mφs all could be internalized by the BMMSCs. Based on these results, the effects of exosomes derived from M0, M1 and M2 Mφs on BMMSC proliferation and osteogenic, adipogenic and chondrogenic differentiation were investigated. Our data demonstrated that the exosomes secreted by M1 Mφs promoted the proliferation of BMMSCs, however, the exosomes secreted by M2 Mφs impaired the proliferation of BMMSCs. M0-Exos didn’t exhibit significant influence on the proliferation of BMMSCs ( Figs. 3A – 3C ). This observation is consistent with our previous findings ( He et al. , 2018 ) and indicates that exosomes are key mediators during the regulating of Mφ-derived CM on BMMSC proliferation. However, it’s in contrast to previously reported data that IFN-γ-activated Mφs negatively regulated the proliferation and activation of hematopoietic stem cells ( McCabe et al. , 2015 ), while M2 Mφs positively regulated the proliferation of BMMSCs ( Yu et al. , 2016 ). This discrepancy can be attributed to the differences in cell lineage and culture conditions. Thus far, the exact role of differently polarized Mφs in osteogenesis has not reached a consensus ( Pajarinen et al. , 2019 ). It has been demonstrated that CM of classically activated monocytes can increase the expression of osteogenic genes in human MSCs ( Omar et al. , 2011 ). Zhang et al. (2017b) also found that M1 Mφs can promote the osteogenic of the MSCs during the early and middle stages. In this study, we confirmed that Mφ-derived exosomes also play key roles in the osteogenic differentiation of BMMSCs. The data of Alizarin red S staining, ALP activity assays and gene expression measures all demonstrated that M1-Exos can significantly promote the osteogenic differentiation of BMMSCs ( Fig. 4 ). This supported the results of the aforementioned studies. However, we found that neither M0-Exos nor M2-Exos exhibited an obvious effect on the osteogenesis differentiation of the BMMSCs according to the Alizarin red S staining or gene expression evidence. It was inconsistent with our previous study which proved that CM generated by M0 or M2 Mφs can positively regulate the osteogenesis of BMMSCs ( He et al. , 2018 ). In fact, several papers have reported that M2 Mφs can promote osteogenic differentiation of MSCs ( Schlundt et al. , 2015 ; Jin et al. , 2019 ; Zhu et al. , 2019 ). In this study, only the results of the ALP activity assay showed a slight promotion effect by M0-Exos and M2-Exos on BMMSC osteogenic differentiation. This revealed that the exosomes and CM derived from the same phenotype of Mφs didn’t exert the same influence on the osteogenic differentiation of BMMSCs. It’s a reminder that other mediators in the CM of Mφs may also affect the property of BMMSCs and Mφs with different phenotypes may modulate the osteogenesis of BMMSCs via different paracrine components. M1 Mφs may promote the early and middle stages of osteogenesis mainly through exosomes. M0 and M2 Mφs may regulate the osteogenic differentiation of BMMSCs by secreting cytokines during the late stage. Mφs also play crucial roles in adipose tissue and influence the adipogenesis. It was reported that M1 Mφs suppressed the adipogenesis of PDGFRα+ preadipocytes, but M2 Mφs showed no influence ( Cheng et al. , 2019 ). In addition, researchers discovered that Mφs in adipose tissue can transport miRNAs through exosomes which finally influence the insulin sensitivity and glucose homeostasis ( Ying et al. , 2017 ). Evidence also confirmed that Mφ-derived microRNA can influence adipocyte metabolism ( Tryggestad et al. , 2019 ). These all indicate that Mφ-derived exosomes may also participate in the adipogenic differentiation of MSCs. This hypothesis was confirmed in the present study. We discovered that M1-Exos promoted the adipogenic differentiation of the BMMSCs, according to Oil red O staining and gene expression levels. However, M0-Exos and M2-Exos did not have a significant influence on adipogenesis-related gene expression and even reduced the lipid droplet formation in the BMMSCs. The result of M1-Exos was in contradiction with that of Cheng et al. (2019), this could be due to the difference of cocultured cells and the involvement of other mediators in the co-culture condition of Cheng et al. (2019). Nevertheless, it supports the data of our previous study which showed that the CM of M1 Mφs promoted the adipogenic differentiation of BMMSCs ( He et al. , 2018 ). This demonstrated that exosomes also modulate the effect of Mφ-derived CM on the adipogenic differentiation of BMMSCs. Apart from proliferation and osteogenic and adipogenic differentiation, Mφs are also involved in the chondrogenic differentiation of stem cells. Several studies have confirmed the negative effect of M1 Mφs on chondrogenesis and the chondrogenesis-inductive effect of M2 Mφs ( Han et al. , 2014 ; Dai et al. , 2018 ; Hu et al. , 2018 ). However, the exact mechanism for these effects remains unclear. Sesia et al. (2015) found that Mφs only promoted the chondrogenic differentiation of BMMSCs in the condition of direct coculture and the CM of Mφs did not modulate chondrogenesis. However, others reported that the CM generated by M1 Mφs inhibited the chondrogenesis of MSCs and that the CM of M2 Mφs did not influence the expression of the COL2 gene but reduced the expression of the ACAN gene ( Fahy et al. , 2014 ). In this study, the effects of exosomes derived from different phenotypes of Mφs on the chondrogenic differentiation of BMMSCs were also investigated. The results of cell staining and the detect of chondrogenesis-related genes both demonstrated that exosomes derived from M0, M1 and M2 Mφs all had a negative effect on BMMSC chondrogenesis. The effect of M1-Exos was consistent with the results of Fahy et al. (2014), but the beneficial effect of M2-Exos was not found in our study. On one hand, it may be attributed to the different origin of Mφs, and on the other hand, it’s probably because of the involvement of other cytokines in the CM. However, our study detected chondrogenic differentiation of BMMSCs only at the early stage (on the 7th day). The long-term effect of Mφ-derived exosomes on chondrogenesis remains to be explored. Taken together, the aforementioned data demonstrated that exosomes derived from different phenotypes of Mφs could exert various influences on the proliferation and osteogenic, adipogenic and chondrogenic differentiation of BMMSCs. It was found that M1-Exos exhibited more robust effects on the proliferation and osteogenic and adipogenic differentiation of BMMSCs than does M0-Exos or M2-Exos. In addition, all three types of exosomes had a suppressive effect on the chondrogenic differentiation of BMMSCs. This indicates that Mφ-derived exosomes may be explored as active reagents to improve the property of MSCs in the regenerative microenvironment. Conclusions This study confirmed our hypothesis that exosomes modulate the effect of Mφ-derived CM on the proliferation and differentiation of BMMSCs. It was also found that even when derived from the same Mφ phenotype, the CM and exosomes do not necessarily exert similar cellular influences on the cocultured stem cells. This provides new insight into the interaction between Mφs and MSCs and indicates that Mφ-derived exosomes may be used in an efficient therapeutic strategy for tissue regeneration. However, the process of tissue repair and regeneration is successive, and it is difficult to determine the exact dose and timing for exosome application. In addition, the exact mechanisms of the exosomes still need to be further investigated. Supplemental Information 10. 7717/peerj. 8970/supp-1 Supplemental Information 1 Characterization of the BMMSCs isolated from the bone marrow of C57BL/6 mice. (A) Results from the flow cytometry analysis of cell surface markers of the BMMSCs. (B) Colony formation ability of the BMMSCs: a general view of colonies (top) and a single colony observed by microscopy (bottom; scale bar: 250 μm). (C) Representative images of EdU-positive cells (cell viability in terms of the EdU assay results; scale bar: 100 μm). (D) Growth curve of the BMMSCs during 7-days in culture in terms of the CCK-8 assay results. (E) A representative image showing the potential of the BMMSCs toward osteogenic differentiation (Alizarin red staining; scale bar: 250 μm). (F) A representative image showing the potential of the BMMSCs toward adipogenic differentiation (Oil red O staining; scale bar: 100 μm). (G) A representative image showing the potential of the BMMSCs toward chondrogenic differentiation (Alcian blue staining; scale bar: 100 μm). Click here for additional data file. 10. 7717/peerj. 8970/supp-2 Supplemental Information 2 Raw data for Figures 1B, 1C, 3B, 3C, 4B 4D-I, 5B-D, 6B-D. Click here for additional data file. 10. 7717/peerj. 8970/supp-3 Supplemental Information 3 Images of western blots. Click here for additional data file.

10. 7717/peerj. 9799

PeerJ

Characterization of primary cilia features reveal cell-type specific variability in in vitro models of osteogenic and chondrogenic differentiation

Primary cilia are non-motile sensory antennae present on most vertebrate cell surfaces. They serve to transduce and integrate diverse external stimuli into functional cellular responses vital for development, differentiation and homeostasis. Ciliary characteristics, such as length, structure and frequency are often tailored to distinct differentiated cell states. Primary cilia are present on a variety of skeletal cell-types and facilitate the assimilation of sensory cues to direct skeletal development and repair. However, there is limited knowledge of ciliary variation in response to the activation of distinct differentiation cascades in different skeletal cell-types. C3H10T1/2, MC3T3-E1 and ATDC5 cells are mesenchymal stem cells, preosteoblast and prechondrocyte cell-lines, respectively. They are commonly employed in numerous in vitro studies, investigating the molecular mechanisms underlying osteoblast and chondrocyte differentiation, skeletal disease and repair. Here we sought to evaluate the primary cilia length and frequencies during osteogenic differentiation in C3H10T1/2 and MC3T3-E1 and chondrogenic differentiation in ATDC5 cells, over a period of 21 days. Our data inform on the presence of stable cilia to orchestrate signaling and dynamic alterations in their features during extended periods of differentiation. Taken together with existing literature these findings reflect the occurrence of not only lineage but cell-type specific variation in ciliary attributes during differentiation. These results extend our current knowledge, shining light on the variabilities in primary cilia features correlated with distinct differentiated cell phenotypes. It may have broader implications in studies using these cell-lines to explore cilia dependent cellular processes and treatment modalities for skeletal disorders centered on cilia modulation.

Introduction Primary cilia are non-motile sensory organelles that protrude from most mammalian cell surfaces ( Singla & Reiter, 2006 ). They transduce various extracellular cues essential for proliferation, differentiation and homeostasis via Hedgehog (Hh), Wnt, Notch, Receptor Tyrosine Kinase, Transforming growth factor β (TGFβ), G protein-coupled receptor and calcium signaling pathways ( Christensen et al. , 2012 ; Clement et al. , 2013 ; Ezratty et al. , 2011 ; Huangfu et al. , 2003 ; Lancaster, Schroth & Gleeson, 2011 ; Nauli et al. , 2003 ; Wheway et al. , 2015 ). The importance of cilia in regulating embryonic and postnatal development is underscored by a wide spectrum of human disorders termed as ciliopathies caused by aberrations in cilia formation and function. The ciliary backbone or axoneme is a microtubular structure that emanates from a modified centriole, the basal body within the ciliary pocket and is encased in a ciliary membrane contiguous with the plasma membrane of the cell. The transport of proteins into the cilia and the bidirectional intra-ciliary movement of cargo is driven by the intraflagellar transport (IFT) machinery composed of multimeric protein complexes, including but not limited to IFT-A/B complexes and associated motor-proteins, kinesin-2 and dynein 2 ( Berbari et al. , 2008 ; Cole et al. , 1998 ; Follit et al. , 2006 ; Nonaka et al. , 1998 ; Omori et al. , 2008 ; Pazour, Wilkerson & Witman, 1998 ; Piperno & Mead, 1997 ; Porter et al. , 1999 ; Signor et al. , 1999 ; Snow et al. , 2004 ; Yoshimura et al. , 2007 ). This not only facilitates the receipt and transduction of signals at the primary cilium but is required for its assembly and maintenance ( Bhogaraju et al. , 2013 ; Cole et al. , 1998 ; Wren et al. , 2013 ). A growing body of evidence has revealed that bone and cartilage are dynamic entities that assimilate a broad range of sensory cues to facilitate skeletal formation and repair. This is augmented by primary cilia extending from the cell surface of osteoblasts, osteocytes, chondrocytes and their precursor preosteoblast and mesenchymal stem cells (MSCs) ( Federman & Nichols, 1974 ; Scherft & Daems, 1967 ; Tummala, Arnsdorf & Jacobs, 2010 ). Cilia serve as critical mechanosensors of fluid flow in osteoblasts and osteocytes ( Temiyasathit et al. , 2012 ), whereas in chondrocytes their main function is sensing peripheral tissue deformation ( McGlashan, Jensen & Poole, 2006 ). In this context mechanical stimulation serves as a crucial anabolic signal; it promotes osteogenic (OS) differentiation in MSCs and mineralization of the chondrocyte matrix ( Chen et al. , 2015 ; Hoey, Kelly & Jacobs, 2011 ; O’Conor et al. , 2014 ). Primary cilia are also required for osteoblast, osteocyte polarity, alignment in bone development and regulate chondrocyte cell polarity at the growth plate ( Lim et al. , 2020a ; Song et al. , 2007 ). They are chemosensitive and serve as a nexus of signaling activity important for OS and chondrogenic (CH) differentiation, such as Hh, TGFβ, Fibroblast growth factor (FGF), Wnt, platelet derived growth factor and parathyroid hormone related peptide ( Cai et al. , 2012 ; Caplan & Correa, 2011 ; Chen, Deng & Li, 2012 ; Jiang et al. , 2016 ; Kozhemyakina, Lassar & Zelzer, 2015 ). The essential role of ciliary function in osteoblast and chondrocyte differentiation is substantiated by impaired differentiation following cilia abrogation caused by knockdown of constituents of the IFT machinery, for example, Ift80, Ift88 and Kif3a ( Qiu et al. , 2012 ; Tummala, Arnsdorf & Jacobs, 2010 ; Wang, Yuan & Yang, 2013 ; Yang & Wang, 2012 ). Moreover, many ciliopathies manifest with altered cilia length and preponderance, underscoring the importance of cilia in cellular function ( Failler et al. , 2014 ; Zhang et al. , 2016 ). The basal levels of ciliation and cilia length have been described for several skeletal cell-lines, for example, MC3T3-E1, MLO-Y4, chondrocytes and mesenchymal stem cells ( Brown et al. , 2014 ; Malone et al. , 2007 ; Wann & Knight, 2012 ; Xiao et al. , 2006 ). Ciliary attributes may be tailored in accordance with distinct differentiated cell states. This is corroborated by a previous study that demonstrated lineage specific changes in primary cilia length and frequencies in response to chemically induced differentiation in human MSCs (hMSCs) over a 7 day period ( Dalbay et al. , 2015 ). Primary cilia length are malleable and modulated by several factors such as cytoskeletal actin organization and inflammatory cytokines ( Kim et al. , 2010 ; McMurray et al. , 2013 ; Pitaval et al. , 2010 ; Wann & Knight, 2012 ). Modulation of the anterograde IFT transport speed by intracellular cyclic AMP levels have also been shown to control ciliary length ( Besschetnova et al. , 2010 ). In chondrocytes compressive loading was reported to reversibly reduce cilia length and incidence ( McGlashan et al. , 2010 ). Moreover, variable stimulation with growth factors may produce disparate consequences on cilia length. While the constitutive activation of FGF signaling was shown to cause primary cilia shortening, its transient stimulation resulted in cilia elongation, in growth plate chondrocytes and limb bud derived mesenchymal cells ( Kunova Bosakova et al. , 2018 ; Martin et al. , 2018 ). C3H10T1/2, MC3T3-E1 and ATDC5 are murine mesenchymal stem cells, preosteoblast and prechondrocyte cell-lines, respectively, that are commonly employed in various in vitro studies, interrogating the mechanisms underlying osteoblast and chondrocyte differentiation, bone disorders and repair ( Hino et al. , 2018 ; Lee et al. , 2017 ; Wang et al. , 2019 ; Yao & Wang, 2013 ). Despite the wide use of these cell-lines, how ciliary features vary in them during native OS and CH differentiation has not been so far investigated. Here we evaluate the primary cilia length and frequency during OS differentiation in C3H10T1/2 and MC3T3-E1 and CH differentiation in ATDC5 cells, over a period of 21 days. Our study revealed cell-type and lineage specific modulation of ciliary characteristics during extended periods of differentiation. These findings expand our existing knowledge and shine light on the primary cilia features in these cell-lines correlated with distinct differentiated cell fates, and may be significant for clinically relevant and explorative studies evaluating cilia dependent molecular mechanisms in these cellular models. Materials and Methods Cell culture ATDC5 cell-line (gift from Dr. Uwe Kornak, Charité—Universitätsmedizin Berlin) were propagated in complete media consisting of DMEM/F-12 (1:1) with 1% L-Glutamine, 10% heat inactivated fetal bovine serum (FBS), one mM sodium pyruvate, 100 U/ml penicillin and 100 μg/ml streptomycin (HiMedia, Mumbai, India). Cells were induced with CH media comprising of complete media supplemented with 10 −7 M dexamethasone, 50 μg/ml ascorbic acid, 10 mM β glycerophosphate and 1X insulin-transferrin-sodium selenite supplement (all from Sigma–Aldrich, St. Louis, MO, USA). Current method of CH differentiation was adapted from previous studies ( Newton et al. , 2012 ; Weiss et al. , 2012 ). C3H10T1/2 was obtained from the cell repository at National Center for Cell Science, India and MC3T3-E1, sub-clone 4 was a gift from Dr. Uwe Kornak, Charité—Universitätsmedizin Berlin and were cultured in complete media comprising of DMEM and modified Eagle’s minimum essential medium, respectively, with 1% L-Glutamine, 10% heat inactivated FBS, one mM sodium pyruvate, 100 U/ml penicillin and 100 μg/ml streptomycin (HiMedia, Mumbai, India). OS media comprised of complete media supplemented with 10 −7 M dexamethasone, 50 μg/ml ascorbic acid and 10 mM β glycerophosphate. All cells were incubated in a humidified atmosphere (37 °C, 5% CO 2 ) and media was replaced every second or third day for 21–30 days, as indicated. Alizarin red staining Calcium deposition in the extracellular matrix (ECM) was estimated by Alizarin red dye that combines with calcium in the cellular matrix, as described previously ( Ovchinnikov, 2009 ; Yamakawa et al. , 2003 ). Briefly, cells were seeded in triplicates at the following densities: 18, 000/cm 2 (C3H10T1/2), 11, 000/cm 2 (MC3T3-E1) and 12, 000/cm 2 (ATDC5), respectively in multi-well plates, and were induced ~24 h later with OS or CH differentiation media. Uninduced cells were propagated in complete media for the same period of time as those induced. At 7, 14 and 21 days post differentiation, cells were fixed in 4% paraformaldehyde (PFA) and stained with 2% Alizarin red, pH 4. 2 (Sigma–Aldrich, St. Louis, MO, USA) for 40 min at room temperature (RT). Stained monolayers were washed with distilled water and images were captured with inverted phase contrast microscope (Olympus, Tokyo, Japan). Levels of mineralization were quantified by extraction of the stain using 10% (v/v) acetic acid and absorbance was measured at 405 nm ( Gregory et al. , 2004 ). Alkaline phosphatase staining C3H10T1/2 and MC3T3-E1 cells were seeded in triplicates at 18, 000/cm 2 and 11, 000/cm 2 per well, respectively in multi-well plates. They were induced ~24 h later with OS differentiation media or left uninduced for the same duration as those induced. At 7, 14 and 21 days post differentiation, cells were fixed in 4% PFA, and alkaline phosphatase ( Alp ) was detected by incubating with nitro blue tetrazolium (NBT)/5-bromo-4-chloro-3-indolyl phosphate substrate (Sigma–Aldrich, St. Louis, MO, USA) for up to 30 min at RT. After washing the stained cells with 1X PBS, images were captured using inverted phase contrast microscope (Olympus, Tokyo, Japan). Alcian blue staining Glycosaminoglycan (GAG) deposition in the ECM was ascertained following CH differentiation in ATDC5 cells plated at a density of 12, 000/cm 2. Subsequently, the cells were fixed in 95% methanol, followed by incubation in 0. 1 M HCl and stained with 1% Alcian blue 8GX solution in 3% acetic acid (Sigma–Aldrich, St. Louis, MO, USA). Stained cells were washed with distilled water and images were captured with inverted phase contrast microscope (Axiovert A1 FL; Zeiss, Oberkochen, Germany). Staining intensity was estimated using Fiji ( http://imagej. net/Fiji ) and represented in arbitrary units. Primary cilia detection during osteogenic and chondrogenic differentiation For detection of cilia by immunostaining, cells were seeded on glass coverslips in multi-well plates at the following densities: 18, 000/cm 2 (C3H10T1/2), 11, 000/cm 2 (MC3T3-E1) and 12, 000/cm 2 (ATDC5). They were induced ~24 h later with OS (C3H10T1/2 and MC3T3-E1) or CH (ATDC5) differentiation media and propagated for 7, 14 and 21 days. The uninduced cells were propagated in complete media for the same period of time. To enhance ciliogenesis, all cells were serum starved (ss) as described earlier ( Prosser & Morrison, 2015 ). To this end induced and uninduced cells were cultured in differentiation or complete media containing 0. 5% FBS for 48 h prior to staining. Two sets of cells were fixed prior to differentiation: one was starved (day 0 uninduced), while the other remained non-starved (day 0 uninduced no ss). Immunocytochemistry and image analyses The cells were fixed in 4% PFA, permeabilized in 0. 2% Triton X-100, blocked in 5% normal goat serum and finally incubated overnight at 4 °C with primary antibodies—anti-acetylated α tubulin (mouse monoclonal, 1:4, 000, cat# T7451; Sigma–Aldrich, St. Louis, MO, USA) and anti-Arl13B (rabbit polyclonal, 1:2, 000, cat# 17711-1-AP; ProteinTech, Rosemont, IL, USA) diluted in blocking solution. Cells were incubated in the following secondary antibodies: Alexa fluor 488 goat anti-rabbit and Alexa fluor 568 goat anti-mouse IgG (cat. # A11034 and cat. # A11031; Molecular Probes, Eugene, OR, USA/ThermoFisher Scientific, Waltham, MA, USA), diluted at 1:500 for 2 h at RT. Nuclei were stained using DAPI and mounted in Prolong Diamond Antifade mountant (both from Invitrogen, Carlsbad, CA, USA; ThermoFisher Scientific, Waltham, MA, USA). Images were acquired using an inverted fluorescence microscope equipped with LD Plan-Neofluar 63X/0. 75 Corr Ph2 oil immersion objective and Axiocam 503 CCD camera (Axiovert A1 FL; Zeiss, Oberkochen, Germany). Primary cilia were discerned by acetylated α tubulin or Arl13B staining and their length were determined manually by tracing along them using Fiji ( http://imagej. net/Fiji ). Cilia lengths are represented in micrometer (μm). Immunolabeled entities with a minimal length of 1. 5 μm were ascertained as primary cilia. Overall length was assessed for a total of 100–178 cilia in three independent replicates. Ciliary frequencies were evaluated in ~100–200 cells per condition in each of three independent replicates. Quantitative real time PCR analyses Cells were seeded at the following densities for RNA extraction: 21, 000/cm 2 (C3H10T1/2), 15, 000/cm 2 (MC3T3-E1) and 17, 000/cm 2 (ATDC5). They were induced ~24 h later with OS (C3H10T1/2 and MC3T3-E1) or CH (ATDC5) differentiation media and propagated for 7, 14 and 21 days. The uninduced cells they were propagated in complete media for the same period of time. All cells were starved as described above. Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA; ThermoFisher Scientific, Waltham, MA, USA). For each sample, total RNA content was assessed by absorbance at 260 nm and purity by A260/280 ratios, and then reverse transcribed using Superscript IV Vilo mastermix™ (Invitrogen, Carlsbad, CA, USA; ThermoFisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Real time quantitative PCR was carried out using StepOne (Applied Biosystems, Foster City, CA, USA; ThermoFisher Scientific, Waltham, MA, USA) with a final reaction volume of 10 μl. All reactions were prepared with five μl of 2x PowerUP™ SYBR™ Green Mastermix (Applied Biosystems, Foster City, CA, USA; ThermoFisher Scientific, Waltham, MA, USA), and run in duplicates for each of three independent replicates. The mRNA levels for target genes were normalized to GAPDH using primer sequences indicated in Table 1. Quantification was carried out using the ΔΔCt method. 10. 7717/peerj. 9799/table-1 Table 1 Primers used in qRT-PCR (F: forward; R: reverse). Gene Primer sequence (5′–3′) Tm (°C) Product size (bp) Alp F: CCAACTCTTTTGTGCCAGAGA 58 110 R: GGCTACATTGGTGTTGAGCTTTT 60 Runx2 F: CTTTACCTACACCCCGCCAG 60 116 R: GTCCACTCTGGCTTTGGGAA 60 Ptch1 F: CCAGCGGCTACCTACTGATG 60 150 R: TGCCAATCAAGGAGCAGAGG 60 Sox9 F: TGAAGAACGGACAAGCGGAG 60 198 R: CAGCTTGCACGTCGGTTTTG 60 Mmp13 F: GGAGCCCTGATGTTTCCCAT 60 165 R: ATCAAGGGATAGGGCTGGGT 60 Gapdh F: CATGGCCTTCCGTGTTCCTA 60 172 R: GTTGAAGTCGCAGGAGACAAC 60 SAG mediated modulation of Hedgehog signaling For stimulation with Smoothened (Smo) Agonist (SAG), cells were seeded as described above and propagated in either complete or OS media supplemented with SAG at a final concentration of one μm. Culture media was replaced every second or third day during the period of induction. Cells were ss for 48 h as mentioned above prior to harvesting. Statistical analyses Data analyses was performed using GraphPad Prism (v8. 4) ( www. graphpad. com/scientific-software/prism/ ). One-way analysis of variance (ANOVA) was performed for identifying variation in cilia length among induced and uninduced groups. Two-way ANOVA was performed to assess effects of day to day variation on induction. Nested t -test was performed to evaluate whether induction has an overall effect on cilia length, considering day-wise variation is nested within the induction effect. Primary cilia length, frequencies and gene expression studies were performed in three independent replicates and p < 0. 05 was considered statistically significant. Cilia length and frequency values are shown as mean ± SEM. Results C3H10T1/2, MC3T3-E1 and ATDC5 undergo the anticipated osteogenic and chondrogenic differentiation in vitro The deposition of minerals in the form of hydroxyapatite in the ECM is a physiological characteristic of hard tissues such as bone and growth-plate cartilage. Alp expressed by osteoblasts and hypertrophic chondrocytes hydrolyzes pyrophosphate to generate inorganic phosphate to promote matrix mineralization ( Orimo, 2010 ). Consequently, high levels of Alp and matrix mineralization reflect osteogenic differentiation. C3H10T1/2 and MC3T3-E1 cells showed elevated OS differentiation following induction from day 14 onward, evidenced qualitatively by pronounced Alp staining in C3H10T1/2 ( Figs. 1A – 1D ) and MC3T3-E1 ( Figs. 1I – 1L ). Elevated calcium deposits in OS induced vs uninduced cells were revealed using Alizarin red staining (ARS) in C3H10T1/2 ( Figs. 1E – 1H ) and MC3T3-E1 ( Figs. 1M – 1P ). Quantitative estimation revealed significant increase in matrix mineralization by day 14 post OS induction that was further enhanced in 21 day induced monolayers ( Figs. 1Q and 1R ). We evaluated the transcript levels of osteoblast differentiation markers, Alp and Runx2 by qRT-PCR and found that both were significantly upregulated at 14 and 21 days post OS induction in C3H10T1/2 ( Fig. 1S ). In MC3T3-E1, significantly high Alp levels were detected at 14 and 21 days post induction, while elevated Runx2 was detected at day 21 after OS stimulation ( Fig. 1T ). 10. 7717/peerj. 9799/fig-1 Figure 1 Characterization of in vitro osteogenic differentiation in C3H10T1/2 and MC3T3-E1 cells. (A–P) Enhanced alkaline phosphatase ( Alp ) levels were qualitatively evident in (A–D) C3H10T1/2 and (I–L) MC3T3-E1 cells at 14 and 21 days post osteogenic (OS) induction compared to uninduced. Similarly robust calcium deposition in the ECM was revealed by Alizarin red staining (ARS) in (E–H) C3H10T1/2 and (M–P) MC3T3-E1 cells at 14 and 21 days following OS induction. (Q and R) Quantitative mineralization levels based on ARS confirmed significantly higher extracellular calcium deposition at 14 and 21 days following OS induction compared to day matched uninduced cells in (Q) C3H10T1/2 and (R) MC3T3-E1 cells (** p < 0. 01, *** p < 0. 001; Two-way ANOVA followed by Sidak’s post hoc test). (S and T) qRT-PCR analyses of the transcript levels of OS differentiation markers, Alp and Runx2. Levels were normalized to Gapdh (* p < 0. 05, ** p < 0. 01, **** p < 0. 0001; One-way ANOVA followed by Tukey’s post hoc analyses). (S) In C3H10T1/2, Alp and Runx2 are significantly upregulated at 14 and 21 day OS differentiated cells (T) In MC3T3-E1, Alp levels were elevated after 14 and 21 days of OS induction, however, significantly appreciable Runx2 was detected 21 days following induction. ATDC5 cells, when induced for CH differentiation, showed progressive increase in mineral deposition in the ECM from day 14 onward ( Figs. 2A – 2E ), evidenced by quantification of ARS in the induced cellular monolayers as compared to uninduced ( Fig. 2F ). Transcript levels of Sox9, an early marker of CH differentiation was significantly elevated at 7 and 14 days post CH induction and was subsequently diminished at day 21 ( Fig. 2G ). Mmp13, a marker of chondrocyte hypertrophy was upregulated after 7 days of CH media stimulation but was most strongly induced in 21 day CH induced monolayers ( Fig. 2G ). CH differentiated cells also showed significant GAG deposition in the ECM at 21 days after treatment, revealed by Alcian blue staining ( Figs. 2H – 2J ). Given that cells were seeded at a moderately higher density for gene expression analyses as compared to other assays, we compared cell proliferation using Trypan blue in C3H10T1/2. Cells were seeded at a density of 21, 000/cm 2 on plastic vs 18, 000/cm 2 on glass coverslips, followed by OS induction for 2, 4, 6, 8, 10, 12 days. No significant variability was observed in viable cell numbers between the two substrates ( Fig. S1 ). 10. 7717/peerj. 9799/fig-2 Figure 2 Characterization of chondrogenic differentiation in ATDC5 cells. (A–F) ARS mediated detection of calcium deposition in the ECM following chondrogenic (CH) induction of ATDC5 cells revealed (A–E) perceptible mineralization day 7 onward. (F) Quantitative assessment confirmed significantly high mineralization levels in CH induced monolayers at 14, 21 and 30 days as compared to uninduced cells (** p < 0. 01, **** p < 0. 0001; Two-way ANOVA followed by Sidak’s post hoc test). (G) qRT-PCR analyses revealed significant upregulation of the transcript levels of CH marker Sox9 at 7 and 14 days following differentiation (* p < 0. 05, ** p < 0. 01; One-way ANOVA followed Dunnett’s post hoc test), while CH hypertrophy marker Mmp13 was elevated at 7, 14 and 21 days following differentiation (** p < 0. 01, *** p < 0. 001, **** p < 0. 0001; One-way ANOVA followed by Tukey’s post hoc analyses). Transcript levels were normalized to Gapdh. (H–J) Significantly high extracellular glycosaminoglycan (GAG) deposition was detected at 21 days post CH induction (**** p < 0. 0001; Wilcoxon signed rank test). Cilia features demonstrate lineage and cell-type specific alteration in osteogenic and chondrogenic differentiation We investigated the changes in primary cilia characteristics, for example, length and frequency over 21 days following CH and OS differentiation in ATDC5, C3H10T1/2 and MC3T3-E1 cells. Primary cilia length was discerned following immunolabeling with acetylated α tubulin. To ascertain the reliability of measurement we co-labeled cilia with its markers Arl13b and acetylated α tubulin in a subset of conditions and compared ciliary length detected with each marker in uninduced and induced monoloayers. We found no significant differences in cilia length reported by either marker in these cell-lines ( Figs. S2 – S4 ). Chondrogenic induction results in primary cilia elongation in ATDC5 cells Representative images for cells at day 0 uninduced and 7, 14 and 21 days CH differentiated cells are shown ( Figs. 3A – 3D ; Fig. S5 ). Cilia length was increased at all time points following CH induction compared to day 0 and day matched uninduced controls but were longest after 14 days of differentiation ( n = 110–152; One-way ANOVA followed by Tukey’s post hoc test; Fig. 3E ). Assessing specifically the induced group, showed that cilia in 14 day CH stimulated cells were significantly longer than in 7 and 21 day CH induced monolayers (Two-way ANOVA followed by Tukey’s post hoc test; Fig. 3F ). Significant variability in cilia length was noted between day 0 uninduced no ss and day 14 uninduced monolayers (One-way ANOVA followed by Tukey’s post hoc test; Fig. S7A ). We also found the CH induction effect to be significantly greater despite day specific variabilities in cilia length ( p < 0. 0001, Nested t test). Finally primary cilia frequency did not vary significantly during CH differentiation in ATDC5 cells ( Fig. S8A ). 10. 7717/peerj. 9799/fig-3 Figure 3 Chondrogenic differentiation causes elongation of primary cilia in ATDC5 cells. Two sets of undifferentiated cells were considered at day 0, namely non-serum starved (day 0 uninduced no ss) and starved (day 0 uninduced). All other CH differentiated and day matched uninduced cells were starved. (A–D) Representative images of ATDC5 primary cilium at day 0 uninduced and days 7, 14 and 21 after CH induction. Primary cilia were labeled with acetylated α tubulin (green), while nuclei were stained with DAPI (blue). Scale bar: five μm. (E) At days 7, 14 and 21 after CH differentiation primary cilia were significantly longer than day 0 and their day matched uninduced control cells. n = 110–152 (** p < 0. 01, **** p < 0. 0001, One-way ANOVA followed by Tukey’s post hoc test). (F) Among the CH induced cells, cilia were longest at day 14 (**** p < 0. 0001, Two-way ANOVA followed by Tukey’s post hoc test). Osteogenic differentiation was associated with cell-type specific changes in primary cilia length and incidence C3H10T1/2 and MC3T3-E1 cells were treated to identical OS differentiation conditions. Cell-line specific changes in primary cilia length were noted and representative images are shown ( Figs. 4, 5A – 5D ; Fig. S6 ). Both cell-lines showed variability in cilia length even in the absence of induction factors ( Figs. S7B and S7C ). 10. 7717/peerj. 9799/fig-4 Figure 4 Osteogenic differentiation in C3H10T1/2 mesenchymal stem cells was associated with progressive cilia lengthening. (A–D) Representative images of primary cilium at day 0 uninduced and 7, 14 and 21 days OS induced cells. Primary cilia were labeled with acetylated α tubulin (green), while nuclei were stained with DAPI (blue). Scale bar: five μm. All differentiated and undifferentiated cells were serum starved (ss) except single set of uninduced cells at day 0 (day 0 uninduced no ss). (E) In day 7 OS differentiated cells primary cilia were significantly shorter than in day 0 uninduced; cilia were also shorter after 14 days of OS differentiation but were subsequently elongated in 21 day OS induced cells compared to day 0 and day matched uninduced cells. n = 111–177 (*** p < 0. 001, **** p < 0. 0001, One-way ANOVA followed by Tukey’s post hoc test). (F) Among the OS media treated cells, cilia were progressively elongated following differentiation that is, day 7 OS < day 14 OS < day 21 OS (**** p < 0. 0001, Two-way ANOVA followed by Tukey’s post hoc test). 10. 7717/peerj. 9799/fig-5 Figure 5 Osteogenic differentiation in MC3T3-E1 preosteoblast cells caused a distinct pattern of cilia elongation and reduction in frequencies. (A–D) Representative images of MC3T3-E1 primary cilium at day 0 uninduced and days 7, 14 and 21 after CH induction. Primary cilia were labeled with acetylated α tubulin (green), while nuclei were stained with DAPI (blue). Scale bar: five μm. All OS treated and untreated cells were serum starved (ss) except one set of uninduced cells at day 0 (day 0 uninduced no ss). (E) Serum starvation caused cilia length increase at day 0; at day 7 after osteogenesis cilia were longer than day 0 uninduced; at 14 and 21 days following OS differentiation the cilia are significantly longer than both day 0 and their corresponding day matched controls. n = 101–178 (** p < 0. 01, **** p < 0. 0001, One-way ANOVA followed by Tukey’s post hoc test). (F) Among OS media induced cells cilia were longest at 14 days of OS differentiation (** p < 0. 01, **** p < 0. 0001, Two-way ANOVA followed by Tukey’s post hoc test). (G) Starvation appeared to increase primary cilia prevalence in all induced and uninduced monolayers. However, cilia frequencies were significantly reduced only in 21 days OS differentiated cells, n = 326–710 (* p < 0. 05, **** p < 0. 0001; One way ANOVA followed by Tukey’s post hoc test). In C3H10T1/2, primary cilia in 14 day OS differentiated cells were significantly shorter compared to both day 0 and the corresponding day matched uninduced controls, but were longer than in 7 day OS induced cells ( n = 111–177; One-way ANOVA followed by Tukey’s post hoc test; Fig. 4E ). Cilia appeared longest at 21 days post OS differentiation. Overall, cilia seemed to be progressively lengthened with sustained OS differentiation in this cell-line (Two-way ANOVA followed by Tukey’s post hoc test; Fig. 4F ). Preponderance of ciliated cells did not vary significantly during OS differentiation in C3H10T1/2 ( Fig. S8B ). In MC3T3-E1, serum starvation caused significant increase in primary cilia length and incidence, including in uninduced cells at day 0 ( Figs. 5E – 5G ). Nevertheless, cilia in 14 and 21 days OS media treated cells were significantly elongated, compared to day 0 and day matched uninduced cells ( n = 101–178; One-way ANOVA followed by Tukey’s post hoc test; Fig. 5E ). Among OS media induced MC3T3-E1 cells, cilia were longer after 14 days of OS stimulation compared to 7 and 21 day differentiated cells (Two-way ANOVA followed by Tukey’s post hoc test; Fig. 5F ). At 21 days after OS induction, cilia prevalence in MC3T3-E1 was significantly reduced compared to day 0 and day matched uninduced cells, however no differences were observed at other time-points ( n = 326–710; One-way ANOVA followed by Tukey’s post hoc test; Fig. 5G ). For both C3H10T1/2 and MC3T3-E1 despite day-wise cilia length variabilities ( Figs. S7B and S7C ) OS induction had a significant effect ( p < 0. 001 and p < 0. 0001 respectively, Nested t test). SAG mediated Hedgehog activation is likely associated with increased cilia length and osteoblast differentiation in C3H10T1/2 cells We assessed the mRNA expression levels of Patched-1 ( Ptch1 ), a direct target of Hh signaling, a canonical ciliary pathway in C3H10T1/2 cells. It was found to be significantly upregulated at 14 days and subsequently diminished at 21 days following OS induction ( Fig. 6A ). We then tested whether Hh stimulation via treatment with Smo agonist/SAG influenced ciliary characteristics during OS differentiation. SAG treatment alone for 7 days significantly upregulated Alp mRNA levels in C3H10T1/2 cells and this effect was further enhanced when combined with OS induction for the same period of time ( Fig. 6B ). In addition, SAG treatment with or without OS induction produced significant elongation of cilia compared to 7 day OS induced cells ( n = 105–177; One-way ANOVA followed by Tukey’s post hoc test; Fig. 6C ). No change was observed in ciliary frequencies following SAG treatment ( Fig. S8C ). 10. 7717/peerj. 9799/fig-6 Figure 6 SAG treatment was associated with primary cilia elongation and increased osteoblast differentiation in C3H10T1/2 cells. (A and B) qRT-PCR analyses of the transcript levels of Ptch1 and Alp. Levels were normalized to Gapdh. (A) mRNA levels of Ptch1 was significantly elevated after 14 days of OS induction and then diminished at 21 days (* p < 0. 05, ** p < 0. 01; One-way ANOVA followed by Tukey’s post hoc analyses). (B) Both SAG treatment individually and combined with OS media for 7 days led to significant upregulation of Alp transcripts compared to day 7 OS differentiated cells; notably Alp levels were more strongly elevated with combinatorial SAG and OS induction as compared to SAG stimulation alone. (*** p < 0. 001, **** p < 0. 0001, One-way ANOVA followed by Tukey’s post hoc test). (C) SAG treatment also resulted in significant elongation of primary cilia compared to day 0 uninduced and 7 day OS differentiated cells. n = 105–177 (* p < 0. 05, **** p < 0. 0001, One-way ANOVA followed by Tukey’s post hoc test). Discussion Primary cilia are dynamically regulated sensory organelles that integrate diverse stimuli for mediating skeletal development and homeostasis. Ciliary properties, such as length and frequencies are often tightly correlated to cellular function. However, how cilia vary in native OS and CH differentiated monolayers in vitro is less understood. Here we characterized primary cilia features during 21 days of OS and CH differentiation in C3H10T1/2, MC3T3-E1 and ATDC5 cell-lines. In contrast to CH differentiation in hMSCs, which did not alter primary cilia length or shortened them in the presence of TGFβ3 ( Dalbay et al. , 2015 ), CH differentiation in ATDC5 cells without TGFβ3 caused cilia elongation ( Fig. 3 ). The mean cilia length observed in non-starved day 0 uninduced ATDC5 cells (2. 7 ± 0. 03 µm) was similar to that in primary cultures of mouse fetal proliferative chondrocytes (2. 82 ± 0. 05 µm) but was longer than cilia in proliferative zone of mouse growth-plate cartilage (1. 2 ± 0. 01 µm) ( Martin et al. , 2018 ). CH differentiation was associated with lengthening of primary cilia at all time-points evaluated. Cilia were longest in 14 day CH differentiated cells (4. 1 ± 0. 11 µm) correlating with intermediate levels of differentiation, compared to those at day 7 (3 ± 0. 04 µm) and day 21 (3. 2 ± 0. 04 µm) post CH induction ( Figs. 2 and 3 ). Highest levels of matrix mineralization, GAG deposition and expression of Mmp13, a chondrocyte hypertrophy marker were noted at day 21 post CH induction. The proportion of ciliated non-starved day 0 control ATDC5 cells (83. 9 ± 3%) were less compared to mouse fetal chondrocytes (92. 6 ± 3. 7%) ( Martin et al. , 2018 ). During CH differentiation in ATDC5 cells ciliary frequencies were not appreciably different and ranged from ~80% to 88% ( Fig. S8A ). Dalbay et al. (2015) further reported that OS differentiation in hMSCs for seven days caused cilia length increase and reduction in frequencies. Primary cilia elongation has also been associated with elevated OS differentiation in other contexts ( Zhang et al. , 2017 ). Congruently, we observed cilia lengthening, although in a cell-type specific manner, during OS differentiation. The mean cilia length in non-starved day 0 uninduced MC3T3-E1 preosteoblasts (2. 2 ± 0. 03 µm) and C3H10T1/2 mesenchymal stem cells (2. 3 ± 0. 03 µm) were shorter compared to mouse primary osteoblasts (~2. 9 µm) ( Lim et al. , 2020b ). Cilia length in MLO-Y4 osteocytes have been reported to range from 2 µm to 4 µm ( Xiao et al. , 2006 ). In C3H10T1/2 cilia were longest in 21 day OS induced cells (2. 9 ± 0. 04 µm); at 14 day OS differentiation cilia (2. 3 ± 0. 03 µm) remained shorter than in controls but were elongated compared to in 7 days of osteogenesis (2. 1 ± 0. 02 µm) ( Fig. 4 ). In MC3T3-E1, cilia lengths were significantly high in day 14 (3. 9 ± 0. 06 µm) and 21 (3. 1 ± 0. 05 µm) OS differentiated cells vs controls but were longest in the former ( Fig. 5 ). Frequencies of ciliated non-starved day 0 uninduced C3H10T1/2 (85. 3 ± 2. 6%) and MC3T3-E1 (81. 4 ± 3. 9%) cells were greater than in mouse primary osteoblasts (~70%) ( Lim et al. , 2020b ). Notably the preponderance of ciliated cells significantly decreased in 21 day OS differentiated MC3T3-E1 cells compared to controls (72. 5 ± 0. 6%). In contrast, cilia incidence remained unchanged during osteogenesis in C3H10T1/2 and ranged from ~79% to 86% ( Fig. S8B ). It is noteworthy that significant OS differentiation was first evident after 14 days of OS induction and was highest after 21 days in both cell-lines ( Fig. 1 ). These data reflect the occurrence of distinct cell-type specific molecular machineries that likely function by cilia modulation in disparate ways to elicit similar overall functional responses in each cell-line. Our current approach of primary cilia length evaluation could be influenced by cells being seeded at a moderately lower density and on glass substrate for cilia immunodetection as compared to plastic for gene expression analyses, as well as cell morphology alterations coincident with differentiation. Furthermore cilia length in cultured cells could be influenced by several factors including differentiation media constituents. For example, dexamethasone that is widely utilized for promoting osteogenesis ( Derfoul et al. , 2006 ; Langenbach & Handschel, 2013 ; Weiss et al. , 2012 ) is a potential enhancer of ciliogenesis and cilia length ( Forcioli-Conti et al. , 2015 ; Khan et al. , 2016 ). However, its effect may be dosage dependent. In previous studies, cilia elongation has been observed with one μm and 10 μm dexamethasone with some effect being noted at concentrations as low as 10 nM ( Forcioli-Conti et al. , 2015 ; Khan et al. , 2016 ). Despite this, using dexamethasone at identical concentrations to those employed here (100 nM), CH induction produced unaltered or shorter cilia, in hMSCs ( Dalbay et al. , 2015 ). This suggests that the inclusion of dexamethasone alone may not account for increased cilia length in all contexts. Testing each induction media component individually will be necessary to discern their role in cilia modulation, if any. While specific mechanism(s) driving cilia length modulation in response to OS and CH differentiation here are unclear, length increase may be essential to enhance ciliary mechanosensitivity, signal transduction and promoting differentiated cell fates. Cilia elongation involves coordinated ciliary membrane extension and modulation of its composition. Longer cilia could experience greater membrane strain, stimulating opening of stretch-activated ion channels in the ciliary membrane or lead to increased bending energy at their base, triggering mechanotransduction signaling ( Resnick, 2015 ; Resnick & Hopfer, 2007 ; Schwartz et al. , 1997 ; Spasic & Jacobs, 2017 ). Increase in cilia length also enhances its surface area leading to accelerated synthesis and trafficking of cilia-specific proteins and signaling molecules to concentrate them in the ciliary microdomain for greater signaling activity ( Breslow et al. , 2013 ; Kee et al. , 2012 ). The Hh pathway performs essential roles in early limb bud patterning, stimulation of osteoblast differentiation in endochondral and intramembranous ossification and postnatal bone homeostasis ( Abzhanov et al. , 2007 ; Horikiri et al. , 2013 ; Kronenberg, 2003 ; Zhu et al. , 2008 ). The expression of Sonic hedgehog (Shh) was shown to be upregulated during OS differentiation in rat MSCs ( Ma et al. , 2013 ). In mammals, the cilium forms an indispensable scaffold to concentrate Hh pathway components, modulating responsiveness to its ligands and regulating activator and repressor forms of the Glioma (Gli) family of transcription factors that control the expression of Hh-target genes ( Huangfu et al. , 2003 ; Liu, Wang & Niswander, 2005 ; May et al. , 2005 ). Accordingly, Gli2 and Gli3 that are essential for mouse skeletal development ( Hui & Joyner, 1993 ; Mo et al. , 1997 ) and their proteolytic processing machinery localize at the cilium ( Haycraft et al. , 2005 ; Mick et al. , 2015 ). Disruption of ciliary components, Ift80, Ift88 and Kif3a caused Gli2 depletion, in addition to defective OS differentiation, but osteogenesis was rescued by Gli2 overexpression ( Haycraft et al. , 2007 ; Koyama et al. , 2007 ; Yang & Wang, 2012 ). In the present study we determined that the expression of Ptch1, a direct target and negative regulator of Hh signaling ( Carballo et al. , 2018 ), were increased in 14 day OS differentiated C3H10T1/2 cells but subsequently diminished after 21 days of OS stimulation ( Fig. 6A ). How increased Ptch1 produces enhanced OS differentiation at 14 days is unclear but its subsequent decline in 21 day OS induced cells was indicative of potentially high Hh activity and was congruent with the highest levels of osteogenesis observed at this time-point. We treated C3H10T1/2 cells with SAG that can activate Smo, the essential transducer of Shh signaling by promoting its ciliary enrichment ( Chen et al. , 2002 ; Frank-Kamenetsky et al. , 2002 ; Rohatgi et al. , 2009 ). Alternatively, SAG may mediate Smo activation in a cilia independent manner ( Fan et al. , 2014 ). SAG stimulation individually and in combination with OS media treatment was associated with elevated Alp expression and significantly elongated cilia in 7 day OS differentiated cells ( Figs. 6B and 6C ). This is contrary to studies where SAG treatment in neuronal cell-types did not alter cilia length or neuronal activity ( Bansal et al. , 2019 ). Modifying cilia length in neural cell-types also did not alter Hh-dependent patterning ( Bangs et al. , 2015 ; Bonnafe et al. , 2004 ). Notably Shh treatment has also been shown to potentiate OS differentiation in MC3T3-E1, C3H10T1/2, and ST2 cells ( Spinella-Jaegle et al. , 2001 ; Tian et al. , 2012 ). These observations provide initial evidences likely suggesting that Hh activation could influence primary cilia properties and promote OS differentiation in C3H10T1/2, and mandate further dissection of its mechanistic basis. In the current study we provide a detailed characterization of cilia features in OS and CH differentiated C3H10T1/2, MC3T3-E1 and ATDC5 cells. All cell-lines evaluated here have been extensively utilized to study molecular processes underlying osteoblast, chondrocyte differentiation and bone disorders, for example, osteoporosis and ciliopathies. It is noteworthy that skeletal ciliopathies such as short-rib thoracic dysplasias and cranioectodermal dysplasias are characterized by changes in primary cilia length and frequencies ( Dupont et al. , 2019 ; Walczak-Sztulpa et al. , 2010 ). Moreover, degenerative conditions such as bone aging and osteoporosis are marked by the suppression of autophagy in osteocytes owing to their susceptibility to hypoxia and oxidative stress ( Onal et al. , 2013 ). Interestingly, primary cilia length modulation has been suggested to control autophagic activity in some contexts ( Pampliega et al. , 2013 ). Thus, modifying cilia length and thereby their sensory properties could serve as an attractive treatment modality in bone disorders. C3H10T1/2, MC3T3-E1 and ATDC5 have also been used in tissue engineering studies evaluating bone healing using various types of biomaterial scaffolds, ultrasound frequencies, modulation of three dimensional cellular environment, static magnetic fields and growth factors ( Arosarena et al. , 2011 ; Baudequin et al. , 2017 ; Cicuendez et al. , 2017 ; Martinez Sanchez et al. , 2017 ; Matsumoto et al. , 2018 ; Weiss et al. , 2012 ; Yang et al. , 2018 ). In this premise, our data inform on the presence of stable cilia to orchestrate signaling during extended periods of OS and CH differentiation, in these cells. We report unique cell-line specific ciliary characteristics that could be useful to define lineage specific in vitro differentiation phenotypes. Our findings illustrate the dynamic variability in ciliary attributes in native differentiated cell-states. This can serve as a useful reference for studies using these cell-lines to dissect cilia dependent cellular processes or therapies for skeletal disorders involving cilia regulation. Finally our results shed light on the extent of basal variation in ciliary features in the absence of differentiation that may inform on the utility of these cell-lines and appropriate controls depending upon experimental goals. Overall these findings warrant in-depth delineation of the underlying regulatory mechanisms and signaling events in each cell-type. Conclusion We characterized the variation in primary cilia features, length and frequencies in ATDC5, C3H10T1/2 and MC3T3-E1 cells following CH and OS differentiation over 21 days. Briefly both were associated with elongation of cilia but displayed distinct alterations correlating with unique in vitro differentiated cell states. Reduced cilia frequencies were noted in 21 day OS differentiated MC3T3-E1 cells. Further investigations are needed to uncover the specific molecular processes governing the observed cell-line specific variations in cilia features during differentiation. Supplemental Information 10. 7717/peerj. 9799/supp-1 Supplemental Information 1 Comparative evaluation of cell proliferation in C3H10T1/2. Cells were seeded at a density of 18000 cm 2 and 21000 cm 2 on glass and plastic, respectively followed by OS induction. Count of viable cells were estimated by Trypan blue at 2, 4, 6, 8, 10 and 12 days and no significant differences were observed (Two way ANOVA followed by Tukey’s post hoc analysis). Click here for additional data file. 10. 7717/peerj. 9799/supp-2 Supplemental Information 2 Primary cilia length evaluation in uninduced and CH differentiated ATDC5 cells by dual markers. (A) Representative images of primary cilium in ATDC5 at day 0 uninduced without serum starvation (day 0 uninduced no ss), day 0 and 7 uninduced and in 7 day CH differentiated cells. Cilia were co-immunolabeled with dual markers, acetylated α tubulin (red) and Arl13b (green); nuclei were labeled by DAPI (blue). (B) Ciliary length was measured for each marker and condition and no significant differences were observed, n=40 (Welch’s t Test). Click here for additional data file. 10. 7717/peerj. 9799/supp-3 Supplemental Information 3 Primary cilia length estimation by dual markers in uninduced and OS induced C3H10T1/2 cells. (A) Representative images of primary cilium in C3H10T1/2 at day 0 uninduced without serum starvation (day 0 uninduced no ss), day 0 and 7 uninduced and in 7 day OS induced cells. Cilia were co-immunolabeled with markers, acetylated α tubulin (red) and Arl13b (green); nuclei were labeled by DAPI (blue). (B) Ciliary length was measured for each marker and condition and no significant differences were noted, n=40 (Welch’s t Test). Click here for additional data file. 10. 7717/peerj. 9799/supp-4 Supplemental Information 4 Cilia length determination in uninduced and OS differentiated in MC3T3-E1 preosteoblast cells by dual ciliary labeling. (A) Representative images of primary cilium in MC3T3-E1 at day 0 uninduced without serum starvation (day 0 uninduced no ss), day 0 and 7 uninduced and in 7 day OS media stimulated cells. Cilia were co-immunolabeled with markers, acetylated α tubulin (red) and Arl13b (green); nuclei were labeled by DAPI (blue). (B) Ciliary length was measured for each marker and condition and no significant differences were noted, n=40 (Welch’s t Test). Click here for additional data file. 10. 7717/peerj. 9799/supp-5 Supplemental Information 5 Lower magnification images of primary cilia during chondrogenic differentiation in ATDC5 cells. Primary cilia were labeled with acetylated α tubulin (green), while nuclei were stained with DAPI (blue). Scale bar: 5 μm. Images were obtained at 40X magnification. Click here for additional data file. 10. 7717/peerj. 9799/supp-6 Supplemental Information 6 Lower magnification views of primary cilia during osteogenic differentiation in C3H10T1/2 and MC3T3-E1 cells. Primary cilia were labeled with acetylated α tubulin (green), while nuclei were stained with DAPI (blue). Images were obtained at 40X magnification. Scale bar: 5 μm. Click here for additional data file. 10. 7717/peerj. 9799/supp-7 Supplemental Information 7 Cilia length variability in the absence of osteogenic and chondrogenic differentiation. Two sets of undifferentiated cells were considered at day 0, non-serum starved (day 0 uninduced no ss) and starved (day 0 uninduced). All other day matched uninduced cells were starved. (A) Cilia length in day 14 uninduced ATDC5 monolayers was significantly longer than at day 0 uninduced no ss, n=110-152 (**** p<0. 0001, One-way ANOVA followed by Tukey’s post hoc analysis). (B) In C3H10T1/2 cells, primary cilia were significantly shorter and longer at 7 and 14 days, respectively compared to day 0 uninduced no ss, n=111-155 (** p<0. 01, **** p<0. 0001, One-way ANOVA followed by Tukey’s post hoc analysis). (C) Primary cilia length was significantly increased with starvation in day 0 uninduced MC3T3-E1 cells; at days 7, 14 and 21 uninduced cells displayed significantly longer cilia compared to day 0 uninduced, n=101-155 (** p<0. 01, **** p<0. 0001, One-way ANOVA followed by Tukey’s post hoc analysis). Click here for additional data file. 10. 7717/peerj. 9799/supp-8 Supplemental Information 8 Primary cilia prevalence was not altered with chondrogenic and osteogenic differentiation in ATDC5, C3H10T1/2 and SAG treatment. All differentiated and undifferentiated cells were serum starved (ss) except single set of uninduced cells at day 0 (day 0 uninduced no ss). No significant variation in primary cilia frequencies were observed in (A) ATDC5 with CH differentiation, n=317-527, and OS induction in (B) C3H10T1/2, n=311-432 and (C) SAG treatment with or without OS differentiation over a 7 day period in C3H10T1/2 cells, n=311-432 (One-way ANOVA followed by Tukey’s post hoc analysis). Click here for additional data file. 10. 7717/peerj. 9799/supp-9 Supplemental Information 9 Raw data for mineralization levels by Alizarin red staining in osteogenic and chondrogenic differentiation of C3H10T1/2, MC3T3-E1 and ATDC5 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-10 Supplemental Information 10 Raw data for transcript levels of Alp, Runx2 and Ptch1 during osteogenesis in C3H10T1/2. Click here for additional data file. 10. 7717/peerj. 9799/supp-11 Supplemental Information 11 Raw data for Alp and Runx2 gene expression during osteogenic differentiation in MC3T3-E1. Click here for additional data file. 10. 7717/peerj. 9799/supp-12 Supplemental Information 12 Raw data for glycosaminoglycan (GAG) deposition in the ECM following 21 days of chondrogenic differentiation in ATDC5 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-13 Supplemental Information 13 Raw data for Sox9 and Mmp13 transcript levels following chondrogenic differentiation in ATDC5 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-14 Supplemental Information 14 Raw data for primary cilia length and frequency variation during chondrogenesis in ATDC5 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-15 Supplemental Information 15 Raw data for primary cilia length estimation by dual markers in uninduced and following chondrogenesis in ATDC5 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-16 Supplemental Information 16 Raw data for primary cilia length and incidence during osteogenic differentiation and SAG stimulation in C3H10T1/2 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-17 Supplemental Information 17 Raw data for primary cilia length evaluation in uninduced and osteogenic differentiation in C3H10T1/2 by dual marker labeling. Click here for additional data file. 10. 7717/peerj. 9799/supp-18 Supplemental Information 18 Raw data for primary cilia length and frequency variation during osteogenic differentiation in MC3T3-E1. Click here for additional data file. 10. 7717/peerj. 9799/supp-19 Supplemental Information 19 Raw data for cilia length determination in uninduced and osteogenic differentiation in MC3T3-E1 by dual marker labeling. Click here for additional data file. 10. 7717/peerj. 9799/supp-20 Supplemental Information 20 Raw data for Alp expression levels following SAG treatment in C3H10T1/2 cells. Click here for additional data file. 10. 7717/peerj. 9799/supp-21 Supplemental Information 21 Raw data for comparison of viable cell numbers during osteogenic differentiation in C3H10T1/2. Click here for additional data file.

10. 7759/cureus. 10085

Cureus

Acceleration of Bone Healing by In Situ-Forming Dextran-Tyramine Conjugates Containing Basic Fibroblast Growth Factor in Mice

An enzymatic crosslinking strategy using hydrogen peroxide (H 2 O 2 ) and horseradish peroxidase (HRP) has been receiving increasing attention for use with in situ-formed hydrogels (IFHs). Several studies have reported the application of IFHs in cell delivery and tissue engineering. IFHs may also be ideal carrier materials for bone repair, although their potential as a carrier for basic fibroblast growth factor (bFGF) has yet to be evaluated. Here, we examined the effect of an IFH made of dextran (Dex)-tyramine (TA) conjugates (IFH-Dex-TA) containing bFGF in promoting bone formation in a fracture model in mice. Immediately following a fracture procedure, animals either received no treatment (control) or an injection of IFH-Dex-TA/phosphate-buffered saline (IFH-Dex-TA/PBS) or IFH-Dex-TA containing 1 μg bFGF (IFH-Dex-TA/bFGF) into the fracture site (n=10, each treatment). Fracture sites injected with IFH-Dex-TA/bFGF showed significantly greater bone volume, mineral content, and bone union than sites receiving no treatment or treated with IFH-Dex-TA/PBS alone (each n=10). This Dex-TA gel may be an effective drug delivery system for optimizing bFGF therapy.

Introduction About 5-10% of fractures result in delayed or poor non-union healing at the fracture site. These cases may lead to functional disability due to deformed healing or pseudoarthrosis [ 1 ]. Therefore, the use of bioactive materials that encourage the bone formation and healing may improve fracture healing. One method that is used to increase the speed of fracture healing involves the local application of growth factors [ 2 ]. Methods that aim to promote bone formation via the sustained release of growth factors using various carriers have been reported. One growth factor known to be active at fracture-healing sites is basic fibroblast growth factor (bFGF). Fibroblast growth factors (FGFs) consist of a family of 23 structurally related polypeptides that play a critical role in angiogenesis and mesenchymal cell mitogenesis [ 3, 4 ]. bFGF is expressed in periosteum during mesenchymal cell proliferation and chondrogenesis and promotes the growth of many types of cells, such as osteoblasts and chondrocytes [ 2, 5 - 7 ]. Among FGF family members, the accumulation of bFGF is greatest in the bone matrix, and it is expressed in periosteum early in bone formation [ 5, 8, 9 ]. In several animal-model studies, locally applied recombinant human bFGF (rhbFGF) has shown osteogenic properties in the regeneration of bone fractures and defects, as well as osteoporotic bone [ 10 - 12 ]. Moreover, several clinical trials have recently reported that bFGF accelerates bone union following osteotomy and in tibial shaft fractures [ 2, 7 ]. These properties indicate that bFGF is effective in promoting bone formation and is a growth factor with therapeutic potential in clinical settings. However, despite this osteogenic potential of bFGF, its efficiency diminishes rapidly following the diffusion in body fluid from bone defect sites [ 13 ]. Moreover, bFGF at high doses can produce adverse side effects, including thrombocytopenia, renal toxicity, and malignant cell activation [ 14, 15 ]. Accordingly, the use of bFGF should ideally be restricted to a form where it is combined with a carrier to promote retention at wound sites. This in turn highlights the need for growth factor delivery carriers that provide the sustained release of bFGF at fracture sites [ 10 - 12, 16 ]. Implantable carriers such as absorbable collagen sponge or hydroxyapatite have been used to aid fracture healing in clinical settings. However, these biomaterials require surgical incision for implantation, and the method is accordingly invasive [ 17 ]. In contrast, injectable materials have the advantage of being less invasive than implantable materials but, compared to implantable materials, generally diffuse only from the injection site [ 18 ]. Therefore, a material that is injectable and has the advantages of an implantable material may be an ideal candidate for a bFGF carrier. In this regard, attention has been recently focused on an enzymatic crosslinking strategy using hydrogen peroxide (H 2 O 2 ) and horseradish peroxidase (HRP) for use with in situ-formed hydrogels (IFHs) made of natural polysaccharides, such as dextran (Dex), pullulan, and hyaluronic acid [ 19 ]. IFHs have suitable properties for biomedical applications, including good cytocompatibility, tunable reaction rate, and substrate specificity, and several studies have reported their use in cell delivery and tissue engineering for bone or cartilage repair [ 20 - 22 ]. IFHs may also be ideal carrier materials for bone repair, although their potential as a carrier for bFGF has yet to be examined. Here, we examined the effect of an IFH made of Dex (IFH-Dex) containing bFGF for promoting osteogenesis in a fracture model in mice. Materials and methods Synthesis of dextran-tyramine conjugates (Dex-TA) Dextran-tyramine conjugates (Dex-TA) were synthesized by referring to previous reports [ 23 ]. Dextran was combined with PNC to form derivatives of p-nitrophenyl carbonate, which were treated with tyramine (TA) by aminolysis. Dextran produced by Meito Sangyo Co. (40 g, 471 mmol OH) (Meito Sangyo Co. , Ltd. , Nagoya, Japan) was dissolved in DMF (1, 600 mL, containing LiCl 30. 9 g) under nitrogen at 90 ˚C. After the dextran was dissolved, the mixture was allowed to cool and at 0 ˚C. PNC (23. 8 g, 120 mmol) and pyridine (9. 2 ml) were combined with the solution under stirring. The feeding molar ratio of PNC to hydroxyl groups with dextran was about 0. 25. The reaction was allowed to continue overnight. Dextran activated with p-nitrophenyl carbonate groups (denoted as Dex-PNC) was then precipitated in cold ethanol (2, 000 ml), followed by filtering and careful washing with ethanol and diethyl ether, and drying in a vacuum oven. Subsequently, Dex-PNC was dissolved in 740 mL of DMF, and TA (9. 1 g, 65 mmol) was added under nitrogen. The reaction was continued for three hours at room temperature. The product was then precipitated in cooled ethanol (800 ml), filtered, and washed carefully with diethyl ether and ethanol. The Dex-TA conjugates were purified further using ultrafiltration against deionized water and isolation following lyophilization. 1H NMR was used to establish the composition of the Dex-TA conjugates. The degree of substitution (DS) (1H NMR) was 12. 1H NMR (D2O): d 2. 60 and 2. 88 (m, -CH2-CH2-), 3. 20-3. 84 (m, dextran glucosidic protons), 4. 84 (s, dextran anomeric proton), 6. 72 and 7. 01 (m, TA aromatic protons). DS, defined as the number of substituents/100 anhydroglucosidic rings (AHG rings) in dextran, was evaluated using 1H NMR by comparison of signal integrals at d 5. 0 and d 6. 5-7. 5 for Dex-TA, in reference to the previous method [ 23 ]. Preparation of IFH-Dex-TA IFH-Dex-TA was prepared by cross-linking Dex-TA polymer in the presence of HRP as the catalyzing enzyme, and H 2 O 2 in 10 mM phosphate-buffered saline (PBS; pH 7. 4). Briefly, Dex-TA polymer solution (final concentration: 2% w/v) was combined with 0. 8 units/mL HRP solution (final concentration: 0. 8 units/mL) containing 1 µg bFGF (IFH-Dex-TA/bFGF) or PBS (IFH-Dex-TA/PBS) and H 2 O 2 solution (final concentration: 4 mM). Mouse fracture model The femur fracture model was produced in C57BL/6J mice aged nine weeks [ 24 ]. The mice were maintained at Nippon Charles River Laboratories (Kanagawa, Japan) in a semi-barrier system with controlled temperature (23 ±2 °C), humidity (55 ±10%) and lighting (12-h light/dark cycle), and received standard rodent chow (CRF-1; Oriental Yeast, Tokyo, Japan). The fracture model was generated by producing a 10-mm incision on the lateral side of the left thigh under sterile conditions. The left patella was medially dislocated by producing a 4-mm lateral parapatellar incision. Following the drilling of a 0. 5-mm hole in the intercondylar notch, a stainless steel needle (0. 5-mm diameter) was retrogradely inserted into the intramedullary canal. The osteotomy was conducted using a wire saw of 0. 22-mm diameter via a small lateral approach, and insertion of a stainless steel needle into the intramedullary canal was used for stabilization. Immediately following the fracture, the animals either received no treatment (control) or received an injection of IFH-Dex-TA/PBS or IFH-Dex-TA/bFGF in the fracture site (n=8, each treatment). All animal experiments were conducted in accordance with the guidelines of the Animal Ethics Committee, Kitasato University (approval number: 2019-127). Determination of new bone volume and bone mineral content All mice were sacrificed four weeks after treatment. Femurs along with the surrounding muscle were removed and fixed in 4% paraformaldehyde for 48 hours at 4 °C. The femurs were moved into PBS and imaged on a micro-focus X-ray CT system (inspeXio SMX-90CT; Shimadzu, Tokyo, Japan) using a 90 kV acceleration voltage, 110 mA current, 20 lm/pixel voxel size, and 1, 024 × 1, 024 matrix size. Using the micro-CT images of the whole femur, newly developed bone volume and bone mineral content were quantified in a 10-mm region of interest centered on the fracture site (500 slices) chosen at the shaft of the femur for each animal using a 3-dimensional (3D) image analysis software application (Tri-3D-Bon; Ratoc System Engineering, Tokyo, Japan), as reported previously. Regions of new bone were determined with a threshold density of 300 mg/cm 3 [ 18, 24 ]. Histology The bone formation mechanism induced by IFH-Dex-TA/bFGF was assessed by excising femurs from the control and treated animals four weeks after the production of fractures. They were dematerialized in a solution of 20% ethylenediaminetetraacetic acid (EDTA) for four weeks. Residual tissue was embedded in paraffin, and 3-µm coronal sections were cut along the long axis of each femur. These sections were processed by hematoxylin and eosin (HE) staining for morphological evaluation. Sustained in vitro release of bFGF To assess the sustained release of bFGF from IFH-Dex-TA, H 2 O 2 solution containing Dex-TA and HRP solution containing 1 µg bFGF were added to a 0. 5-mL plastic microcentrifuge tube. After curing IFH-Dex, 200 μl of PBS was added to the tube. To determine the release of bFGF from IFH-Dex-TA, bFGF-loaded microtubes were incubated in 200 μl of PBS for one, four, eight, 24, 48, and 72 hours. The supernatant was collected and kept at -30 °C until assay. The concentration of bFGF was estimated using a commercial ELISA kit (R&D Systems, Minneapolis, MN). Results Dex gel containing bFGF induced callus formation in vivo We evaluated callus formation in the fractured femurs following treatment with IFH-Dex containing bFGF using micro-CT image analysis at four weeks post-treatment (Figure 1 ). Figure 1 Representative 3D micro-CT image of femurs following injection of in situ-formed hydrogel made of dextran (IFH-Dex-TA) loaded with bFGF 3D micro-CT images of fractured femurs from (A) control, (B) IFH-Dex-TA/PBS-, and (C) IFH-Dex-TA/bFGF-treated groups after four weeks of recovery. Red: new bone formation; gray: existing bone CT: computed tomography; IFH: in situ-formed hydrogels; Dex: dextran; TA: tyramine; PBS: phosphate-buffered saline; bFGF: basic fibroblast growth factor Compared to sites that received no treatment (control) or were treated with IFH-Dex alone, fracture sites injected with IFH-Dex-TA/bFGF showed significantly greater bone volume and bone mineral content (Figure 2 ) (p<0. 05). In contrast, these variables were comparable between the IFH-Dex-TA and control groups. Figure 2 Quantification of callus area and bone mineral content at the fracture site four weeks following the creation of the fracture Analysis of (A) bone volume (mm 3 ) and (B) bone mineral content (mg) in calluses from control (white bars), IFH-Dex-TA/PBS- (black bars), and IFH-Dex-TA/bFGF-treated (gray bars) groups. Data are shown as the mean ± standard error (SE) (n=8) *p: <0. 05 versus the control group IFH: in situ-formed hydrogels; Dex: dextran; TA: tyramine; PBS: phosphate-buffered saline; bFGF: basic fibroblast growth factor Histomorphometric findings To evaluate bone union, we conducted a histological examination of the fracture site four weeks post-fracture. The IFH-Dex-TA/bFGF-treated group exhibited large calluses at the fracture site, and the fracture site was bridged by newly formed bone (Figure 3 ). In contrast, in the IFH-Dex and control groups, small calluses were observed at the fracture site (Figure 3 ). Figure 3 Hematoxylin and eosin (HE) staining of the femur and surrounding muscle (A–B) control, (C–D) IFH-Dex-TA/PBS, and (E–F) IFH-Dex-TA/bFGF. Scale bars indicate 2 mm (A, C, E) or 0. 5 mm (B, D, F) IFH: in situ-formed hydrogels; Dex: dextran; TA: tyramine; PBS: phosphate-buffered saline; bFGF: basic fibroblast growth factor Sustained release of bFGF from IFH-Dex-TA in vitro The in vitro profile of bFGF release from IFH-Dex-TA is shown in Figure 4. bFGF release from Dex-TA gel occurred with an initial burst in the first four hours followed by a gentler release pattern after eight hours. Thereafter, the sustained release rate was moderate, with 37% of the administered dose of bFGF gradually released across 72 hours. Figure 4 Sustained release of bFGF from IFH-Dex-TA gel in vitro bFGF concentration in PBS at different time points. Results are presented as mean ± standard error (SE) (n=5) IFH: in situ-formed hydrogels; Dex: dextran; TA: tyramine; PBS: phosphate-buffered saline; bFGF: basic fibroblast growth factor Discussion Previous studies have reported that bFGF combined with carriers having various forms, including powders, sheets, sponges, gels, has an effect on the bone to promote bone formation [ 2, 10 - 12, 16, 18, 25 ]. Dextran protects bFGF from acid and heat inactivation and proteolysis, and its protective effect is stronger than that of heparin, a known bFGF stabilizer [ 26 ]. Dextran gel is gradually released from bFGF and promotes angiogenesis [ 27 ]. In our present study, 1 μg bFGF with in situ-formed hydrogels composed of Dex-TA induced accelerated bone formation at the fracture site in mice. We previously showed that 1 μg bFGF combined with artificial collagen gel failed to accelerate bone formation in a mice fracture model [ 18 ]. In addition, even when 10 μg bFGF combined with collagen powder was administered to the fracture site, bone formation was not accelerated [ 25 ]. Accordingly, this IFH-Dex-TA gel may be useful as a carrier for bFGF to accelerate bone formation. When using various substances as carriers for growth factors, it is important that the growth factors be released slowly. bFGF is reported to have a growth-promoting effect on undifferentiated mesenchymal cells at an early stage in the process of fracture healing [ 11 ]. When administered directly into the body, it diffuses rapidly. However, because it is considered to produce its activity by affecting the initial stage of the bone union process [ 11, 13 ], it is important that the release occurs locally in order to minimize or prevent diffusion. The IFH-Dex-TA gel containing bFGF provided a large and sustained release of bFGF in the first four hours after injection. The amount released thereafter decreased, but the bFGF concentration in the PBS solution after 72 hours was 38. 7 ng/mL. In this regard, bFGF was reported to show proliferative activity on periosteal cells at a concentration of 1 ng/mL in vitro [ 28 ]. The proliferation of periosteal cells occurred from day one to three after the creation of a fracture in a fracture model in mice [ 29, 30 ]. Accordingly, we speculate that bFGF-containing IFH-Dex-TA gel could release a sufficient amount of bFGF to exert a cell-growth-promoting effect during fracture healing. There were two main limitations to this study. The release kinetics of bFGF in vivo remain unclear. The usage of fluorescently-labeled bFGF was needed to reveal the release kinetics. Moreover. extrapolating the results obtained from small animal models directly to man may not be clinically relevant. We recommend further investigation using large animals. Conclusions We examined the osteogenesis-promoting ability of Dex gel containing bFGF in a fracture model in mice. Fracture sites injected with Dex/bFGF showed significantly greater bone volume and bone mineral content than sites receiving no treatment or treated with Dex gel alone. The use of this Dex gel as a drug delivery system may be effective for optimizing bFGF therapy.

10. 7759/cureus. 10558

Cureus

Successful Full-Thickness Skin Regeneration Using Epidermal Stem Cells in Traumatic and Complex Wounds: Initial Experience

Skin grafts generated from cultured autologous epidermal stem cells may have potential advantages when compared to traditional skin grafting. In this report, we will share our initial experience with a new technique for the treatment of difficult cutaneous wounds. Eight patients with traumatic or complex wounds underwent full-thickness skin harvesting and processing of epidermal stem cells, followed by the application of our novel management protocol. The patients were at high risk for non-healing and/or severe scar formation due to large traumatic de-gloving crush injuries, wounds from necrotizing fasciitis, or chronic wounds from osteomyelitis. We examined the percent graft success, recipient to donor size ratios, the median time to epithelialization, and two-point sensory discrimination. An international scale (The Patient and Observer Scar Assessment Scale - POSAS) was used to evaluate wound cosmesis and included parameters such as pain, pruritus, vascularity, pigmentation, and thickness of the healing wound. In total, 10 out of 11 wounds had 100% survival of the graft, and one patient had an 80% graft take. The largest wound was 1600 cm 2, and all wounds were harvested from small-donor sites, which were closed primarily. The mean wound to donor ratio was >25:1. Most wounds were fully epithelialized within 30 days. Neurologically, four out of six patients studied exhibited two-point discrimination similar to the adjacent native uninjured skin. The majority of patients reported their wounds to have limited pain or pruritus, and similar pigmentation to adjacent skin.

Introduction Major advancements in skin grafting did not occur until the 19th century with the advent of general anesthesia. The limitations of the commonly utilized split-thickness skin grafts are well described: wounds often develop severe scarring, hyperpigmentation, and/or hyperalgesia [ 1 - 3 ]. With the exception of autologous full-thickness skin grafts, no split-thickness or skin substitute has been able to fully replicate the architecture, function, or cosmetic appearance of native skin. Stem cells have been shown to proliferate, differentiate, and survive when cultured, and may revolutionize the treatment of disease [ 4, 5 ]. A novel technique utilizing autologous epidermal stem cells may confer benefit in the coverage of major skin loss. Extensive skin constructs have been created from the harvesting of tiny portions of healthy skin prior to auto-grafting [ 2 ]. This homologous, autologous stem cell-derived graft material appears to not only accelerate wound healing but also to reinstate the native wound’s sensation, hair follicles, pigment, and glandular morphology [ 6 ]. We describe the first experience with ­­eight patients undergoing epidermal stem cell auto-grafting in the setting of complex wounds. There are no other case series in the literature describing the use of epidermal stem cells in the setting of traumatic or previously infected wounds. Case presentation Eight patients presented to our institution with difficult to heal wounds. These included large traumatic wounds, debrided wounds from severe necrotizing fasciitis, and chronic wounds, which had failed prior skin grafting. Following our own novel protocol, a full-thickness segment of skin for harvest was removed from the thigh, groin, or abdomen. Care was taken to avoid cautery during graft harvest, so cellular morphology was not altered. The specimen was excised and placed in a sterile cup that contained a gentamicin antibiotic. The size of the graft was typically a small fraction of the size of the wound. The specimen donor site was then irrigated, and hemostasis was obtained prior to being closed primarily. The donor sites are all listed as one dimensional because they were closed primarily. After processing of the donor specimen at a bioengineering laboratory (Polarity TE, Salt Lake City, USA), the graft material was returned as a tissue paste in a sterile package. The recipient wound bed, at this point, had a clean bed of healthy granulation tissue. During the deployment, the graft material was spread on the wound bed as a paste similar in consistency to peanut butter. The wound was then covered by a thin sheet of silicone that was meshed in a 1. 5:1 fashion using a Skin Graft Mesher (Zimmer® Surgical, Dover, USA) and placed directly over the graft material. The silicone sheet was secured by staples to the wound edge and then covered by a fibrin glue (Evicel™, Somerville, USA). Lastly, a white Granulofoam™ dressing (KCI, San Antonio, USA) was applied with an Ioban™ drape (3M™, St. Paul, USA) to secure the dressing. Negative pressure therapy was initiated and set to -75 mmHg. This dressing was then removed and reapplied every 5-7 days for three weeks. The dressing was then changed to a regular non-adherent material in between the wound and the overlying gauze bandage. Successful graft maturation was noted by the presence of dermal islands (see Figure 1 ). These were portions of healthy granulation that subsequently epithelialized and formed functional and glandular tissue (see Figure 2 ). The final graft resembled healthy skin by general appearance, pigmentation, and sensation. Additional images are provided in appendices. Figure 1 Dermal islands of growing tissue at two weeks Figure 2 Fully functional, sensate, and glandular tissue at four weeks The tissue was evaluated postoperatively in several ways. The Patient and Observer Scar Assessment Scale (POSAS) was used to evaluate the cosmesis and functionality of the wound. This assessment scale has been shown to have inter-observer reliability and is widely used amongst dermatologists and plastic surgeons worldwide [ 7 ]. It measures scar quality metrics, including pigmentation, pliability, thickness, and similarity to adjacent skin architecture. It is also unique in that it combines the patient’s viewpoint of the wound with the observer’s assessment. In addition to the POSAS evaluation, we assessed the grafts for two-point discrimination and time to epithelialization. The average patient age was 46 years (see Tables 1 - 2 ). All patients were at high risk for non-healing and severe scar formation. Nearly all of the wounds were acute and treated during their initial hospitalization. In total, 10 out of 11 patients had 100% graft survival. The single patient who had only an 80% graft survival had been extensively and repeated debrided for Fournier’s gangrene of the perineum (Table 1, patient 7), and likely failed 100% grafting because of a flare-up of pyoderma gangrenosum. The largest wound was over 1600 cm 2, and all wounds were harvested from a donor site, which was closed primarily. The wound to donor ratio exceeded 25:1 on most patients. The average donor site scar was seven centimeters in length. All donor site scars were healed without pain or deformity. The average time to epithelialization of the graft recipient wound beds was 30 days. Neurologically, more than 50% of the patients exhibited two-point discrimination, which was similar to the adjacent native skin. Overall, patients reported the grafted wounds to be low in pain, pruritus and had similar pigmentation to adjacent skin. Table 1 Wound characteristics LUE - left upper extremity; RUE - right upper extremity; LLE - left lower extremity; B/L - bilateral Patient Type/location Wound/donor ratio Two-point discrimination Overall scar score Percent take 1 Fournier’s gangrene 28:1 Yes 100% 100% 2 LUE traumatic degloving 37:1 Yes 70% 100% 3 RUE traumatic degloving 13:1 Yes 80% 100% 4 LLE traumatic degloving 75:1 Yes 50% 100% 5 Chronic osteomyelitis 10:1 No 100% 100% 6 Sacral ulcer and B/L heel 30:1 No 100% 100% 7 Fournier’s gangrene 18:1 Pending Pending 80% 8 RUE traumatic degloving 22:1 Pending 80% 100% Table 2 Demographics DM - diabetes mellitus; PVD - peripheral vascular disease; RSD - reflex sympathetic dystrophy Patient Type/location Time to epithelization Size of the wound Harvest site Size of donor site Risk factors for poor wound healing 1: 48-year-old female Necrotizing fasciitis - bilateral groin and perineum 3 weeks Left groin: 20x8cm; right groin: 4x3cm Right lateral thigh 6cm DM 2: 56-year-old female Traumatic degloving - left upper extremity 4 weeks 20x15cm Left lateral thigh 8cm none 3: 37-year-old male Traumatic degloving - right upper extremity 8 weeks 8x5cm, 15x5cm Right lateral thigh 8cm none 4: 48-year-old female Traumatic degloving - left lower extremity x2 4 weeks 50x20cm (thigh); 30x20cm (calf) Right medial thigh, right groin 12cm (right thigh), 9cm (right groin) none 5: 37-year-old male Chronic osteomyelitis - right lower extremity 4 weeks 7x6cm Lower abdomen 4cm PVD, DM, RSD 6: 52-year-old male Sacral ulcer and bilateral heel 4 weeks Bilateral heel: 8x4cm; sacrum: 10x16cm Lower abdomen 7cm DM 7: 65-year-old male Fournier’s gangrene - perineum 8 weeks 6x12cm Lower abdomen 4cm PVD, DM, pyoderma 8: 27-year-old male Right upper extremity degloving 4 weeks 8x20cm Right groin 7cm none Discussion The main findings of our study were: 1. epidermal stem cell skin grafts had a nearly 100% survival despite their use in highly complicated wounds; 2. the cosmetic appearance and functionality (including sensation) of the graft site was similar to that of adjacent native skin. Split thickness skin grafts are commonly employed for coverage of various types of granulating wound beds, but have a number of limitations. These grafts are absent of normal sensation, and also lack the durability, appearance (i. e. , pigmentation), and functionality of native skin [ 1, 3 ]. This is an inherent limitation of wound biology. Re-epithelization is limited by competing and opposing forces of fibroblasts during the proliferative phase of wound healing and is most apparent in large or deep wounds. This process promotes scar formation and secondary wound contracture without the functionality of normal skin [ 8 ]. Published literature has shown that high-risk patients with compromised wound biology have a high rate of failure (nearly 30%) with even the most advanced skin substitutes [ 6, 9 ]. On a cellular level, both chronic and complex wounds have dysregulated growth factors, cellular activity, and the persistence of a poorly differentiated state of keratinocytes [ 8 ]. An increased understanding of the nature of stem cells, molecular signaling, and tissue engineering in the past two decades has made the clinical application possible (see Figure 3 ) [ 5 ]. Anatomically, the reservoir for epithelial stem cells lies in the outer root sheath in the proximal bulge of a hair follicle. Cell lineage studies performed by Ito and colleagues in 2004 showed that these pluripotent bulge cells and their progeny have the ability to proliferate and undergo terminal differentiation to form the native full-thickness architecture. This includes the hair shaft, melanin-producing cells, and glandular cells [ 10 ]. The same investigator was able to show, in animal studies, that these isolated bulge cells rapidly divide and migrate in the setting of the acute wound-healing response [ 11 ]. Radio-labeled leucine-rich repeat-containing G-protein coupled receptor, a transmembrane protein marker for stem cells, has been extensively studied and found throughout epithelial surfaces, specifically in the follicular bulge. In animal studies, these isolated cells were shown to have the ability to proliferate, migrate, and differentiate into native hair follicles [ 12 ]. Figure 3 Timeline of advances in skin grafting A recent pediatric case report using the same skin construct from our investigation supported this cellular phenomenon. In this case, a biopsy of a child who had received the stem cell-derived graft five months prior demonstrated a normal distribution of melanocytes and structurally normal hair follicles. On further cell staining, this biopsy demonstrated a fully developed stratified epidermis with organized dermal plexus vasculature [ 2 ]. The ability for these pluripotent stem cells to differentiate has been studied extensively and is the result of precise signals from their microenvironment. This gives these stem cells the capacity to recognize, regenerate, and replace damaged or dying cells in their vicinity [ 5 ]. Our case series included only very complex wounds in a complicated patient population (Table 2 ). We focused on high risk, sometimes previously grossly infected, large wounds that traditional split-thickness skin grafting would have likely failed, and in two of our patients had failed. We feel that our excellent graft survival, along with the cosmesis and functionality of the grafts, would not have been possible without this epidermal stem cell-derived product. With regards to safety, the graft is autologous, so there does not appear to be a risk of immunogenic reactions. There are a number of limitations to our investigation. The subjectivity of the scar assessment and time to epithelialization were apparent. The wounds were examined on follow up appointments that were spread apart by at least a week after discharge, limiting our analysis to periodic observations. The percent graft survival was somewhat subjective as well because it reflects a visual estimation of how well the graft took throughout the wound. This was not a major limitation because almost all the wounds had 100% graft survival. Our study is retrospective and limited by the size of the study population and, as a result, warrants further study. Additionally, there is no long-term data because our patients have only recently received grafting; however, this is the first report of its kind, evaluating a series of patients with complex wounds receiving successful grafting from epidermal stem cells. Future studies would include looking at tissue specimens and performing a skin biopsy of the recipient wounds beds and comparing them to adjacent native skin. Conclusions We describe the index case series utilizing a novel tissue based, autologous, homologous epidermal stem cell derived product in complex surgical wounds. These wounds on average were very large and likely not amenable to traditional skin grafting. Despite requiring only a small full thickness donor skin specimen, more than 90% of the patients had a 100% graft survival covering large and complicated wound surfaces, and many experienced normal sensation at the healed graft site. Furthermore, our clinicians felt that cosmesis was far superior to the results seen with split thickness skin grafts. This is the first case series in the literature describing the use of epidermal stem cells.

10. 7759/cureus. 12647

Cureus

Using Immersive Technologies to Develop Medical Education Materials

Principles of modern surgical education for clerkship and residency were established by the novel approaches of Sir William Osler, MD, Flexner report, and Halsted's principles. The evaluation of surgical education has continued to benefit from the wisdom of the past by harnessing technologies. Rapidly changing and improving the nature of the surgery fostered that evaluation and enforced the institutions to find new solutions for surgical education. In the present descriptive technical report, our aim was threefold: (1) to share acquired educational materials based on immersive technologies involving 3D-printing, Augmented Reality (AR), and 360-degree video recording to improve ongoing pediatric surgery student training at our faculty, (2) to describe workflow underlying the construction of the materials, and (3) to provide approaches that may help other students and lecturers to develop their educational materials. The educational materials, including 3D-printed models, AR hybrid student book, a hydrogel-based simulation model of the kidney, and Mirror World Simulation, were constructed. The authors, who are medical students, led the construction of the educational materials, so the educational materials were shaped by a collaboration between students and pediatric surgeons. The materials constructed enabled the students to practice surgical procedures and experience different surgical environments. We believe these educational materials can serve as a valuable resource for training in many medical specialties in the future. This work was presented at the American College of Surgeons (ACS) Quality and Safety Conference Virtual, August 21-24, 2020.

Introduction Medical education has evolved dramatically from ancient times to the present. One of the milestones in that evaluation is the detailed illustrations drawn during the Renaissance, which enabled physicians to explore the human anatomy, make discoveries and inventions in various medical fields, and translate their knowledge to trainees [ 1 - 2 ]. Today, many tools and technologies involving simulations, smartphones, tablets, telemedicine, Augmented Reality (AR), Virtual Reality, 360-degree video recording, wearable devices, digital games, e-learning environments, atlases accompanied by AR, virtual patients, and 3 Dimensional (3D)-printed models are in use to facilitate students' essential knowledge acquisition and help them to gain required skills [ 3 - 4 ]. A 2011 study in Transactions of the American Clinical and Climatological Association demonstrated the exponential growth of medical knowledge: in 1950, the doubling time was 50 years; in 1980, 7 years; in 2010, 3. 5 years, and is projected to be 73 days by 2020 [ 5 ]. In addition to expanding the volume of medical knowledge, the characteristics of trainers have changed over the years through the changes in habits and attitudes. According to Prensky's famous article published in 2001, today's medical students, as he designated as "digital natives, " do not just have different hobbies or music preferences; they learn differently as well. Consequently, digital natives have obliged digital immigrants to transform their educational methods and materials [ 6 ]. Much more educational materials are in progress to meet the requirements of medical education that have been changing under the influence of many factors involving the changes in pedagogical methods, health care environment, roles of the physician, students' profile, and rapidly increasing volume of medical knowledge [ 3 - 6 ]. In this descriptive technical report, we wanted to describe educational materials developed using immersive technologies to improve ongoing pediatric surgery student training in our faculty [ 7 - 9 ]. We harvested computer-aided design software (CAD), 3D printers, AR development tools, hydrogel-combined molding techniques, and 360-degree video recording to construct our educational materials. Students and surgeons worked collaboratively on the construction of educational material; consequently, they embedded it into the theoretical and practical curriculum of pediatric surgery clerkship. The purpose of the present paper is not limited to demonstrating educational materials based on immersive technologies. In this paper, we also tried to share the workflow underlying the construction of the materials in a detailed and step-by-step manner as we aim to enable readers worldwide to construct these educational materials at their institutions. Lastly, these instructions were enhanced with videos, and the recordings from clerkship were given to demonstrate the usage of educational materials in daily practice. Technical report 1. Construction of 3D-printed models of congenital anomalies (a) Three-Dimensional Modeling: Five different pathologies in pediatric surgery involving anorectal malformation, esophageal atresia, vesicoureteral reflux, choledochal cyst, and jejunoileal atresia were modeled three-dimensionally and appropriately to anatomical characteristics according to two-dimensional drawings, radiological views, and surgical experience using Autodesk 3ds Max (Autodesk, San Rafael, USA) (Figure 1 ). Then, 3D models were exported as stereolithography (STL) files. They were imported into Meshmixer (Autodesk, San Rafael, USA) and repaired using that software to prepare them for printing. Figure 1 Three-dimensional modeling Anorectal malformation, esophageal atresia, vesicoureteral reflux, choledochal cyst, and jejunoileal atresia were modeled three-dimensionally. The virtual model of the anorectal malformation was demonstrated in this figure. (b) 3D Printing: Ultimaker 2 Extended + (Ultimaker B. V. , Geldermalsen, The Netherlands) was preferred for printing 3D models. Following the repair process, 3D models were imported into the Ultimaker Cura (Ultimaker B. V. , Geldermalsen, The Netherlands), a specific software used to prepare models for printing. The sizes of the models were rearranged according to the characteristics of models to make them more feasible and effective for clinical and educational indications. Then, settings involving printing material (filament type), nozzle speed, and infill/support futures were modified specific to models and saved as G-code files. Subsequently, models were 3D printed on Ultimaker 2 Extended + (Figure 2 ). Figure 2 3D printing Models of anorectal malformation, esophageal atresia, vesicoureteral reflux, choledochal cyst, and jejunoileal atresia were 3D printed. (c) Fine-tuning: As mentioned above, the models were supported with materials added by Ultimaker Cura to make the printing easier. After the printing step had been completed, the models were cleared from support materials. Then they were colored to differentiate different parts. That enabled us to demonstrate and teach pathologies more effectively (Figure 3 ). Figure 3 Fine-tuning Models of anorectal malformation, esophageal atresia, vesicoureteral reflux, choledochal cyst, and jejunoileal atresia were colored manually. 2. Construction of the simulation model based on 3D printing and hydrogel Figure 4 shows the process of constructing the simulation model based on 3D printing and hydrogel. Figure 4 The construction of the simulation model based on 3D printing and hydrogel was summarized using a process flow diagram. CAD: computer-aided design; PVA: polyvinyl alcohol (a) Design and Fabrication of the External Mold: The virtual model of a kidney can be produced through different methods involving creating on CAD software and generating on 3D calculation software by using data acquired through digital imaging and communications in medicine (DICOM) from computed tomographic angiography (CTA) images. We received the virtual model of the kidney using these methods and imported it as an STL file into the CAD software (Fusion 360, Autodesk, San Rafael, USA) (Figure 5 ). The Create Base feature was selected to enter the direct editing mode. In this mode, due to the importing as an STL file into the software, the virtual model of the kidney was in a mesh form. However, the model allows users to create objects in another form named solid body. As a result, it is impossible to use the mesh form of the virtual model as reference geometry for the external mold. The mesh form of the virtual model was converted to the solid body using the Mesh to the B-Rep conversion tool. Then, the sketches, 2D geometries used as a base to construct 3D geometries, were created and used as paths to add three solid pipes (an inlet runner duct and two outlet vents) that facilitate the unidirectional filling of the kidney. A box was created according to the earlier designed parts of the mold to be the target body. The inverse geometry of the kidney and additional technical features were created by combining them with the target body using cut operation. In other words, firstly, the parts that would be hollow in the completed mold were created; then, they were excluded from the box that would be the body of the mold. Then the mold was split into two pieces using a horizontal plate to facilitate the removal of the kidney model. The completed external mold was exported and converted into an STL file (Figure 6 ). It was then 3D printed on the Ultimaker 2 Extended+ (Figure 7 ). Figure 5 Designing the kidney model for the external mold Virtual model of the kidney was prepared. Figure 6 Designing the external mold The mold of the kidney model was prepared using the inverse of its geometry. Figure 7 Fabrication of the external mold The mold of the kidney model was printed. (b) Preparation of the PVA Solution: To prepare 20% w/v aqueous solution of PVA, 13. 2 g of PVA powder was added to 47 mL of deionized water. The solution was heated and stirred at 100°C for approximately one hour until it became clear. The solution was left to cool to room temperature for about one hour. Finally, 19. 6 mL of dimethyl sulfoxide (DMSO) was added to the solution. Throughout the preparation process, the solution was covered with stretch film to prevent drying. (c) Injection of the PVA Solution: The constructed mold was sealed with silicone to prevent leaking during freeze/thaw cycles. The PVA solution was injected into the inlet runner duct of the mold using a 1 mL syringe until the overflow at the outlet vents was seen. Overall, 15 mL of PVA solution was used per model. d) The Freezing and Thawing Process: Multiple freezing and thawing cycles were performed to convert the PVA solution into a gel. This allowed the material to form a porous network of hydrogen bonds and simulate characteristics of the parenchyma of the kidney. The mechanical properties of the material were improved by altering the concentration of the solution and the number of freeze-thaw cycles based on feedback received from surgical experiences. Seven freeze-thaw cycles were subsequently administered to the entire external mold. Each cycle consists of two steps: (1) placing the mold in a −20°C freezer for 24 hours, and (2) thawing the mold at 4°C for 24 hours The completed model was then separated from the mold (Figure 8 ). Figure 8 Kidney model based on 3D printing and hydrogel The completed kidney model was separated from the mold. 3. Creation of the hybrid pediatric surgery book based on AR (a) Compilation of the Pediatric Surgery Book: A pediatric surgery book was created from lecture notes prepared using reference textbooks for subjects involving gastrointestinal atresias, ileus, malrotation, abdominal anomalies, and trauma (Figure 9 ). Then parts that had been found appropriate for implementing AR were enhanced with visual instructions to make the book easier to use. Figure 9 Compilation of the pediatric surgery book The pediatric surgery book was created for subjects including gastrointestinal atresias, ileus, malrotation, abdominal anomalies, and trauma. (b) Creation of 3D Models of Related Disorders Using Cad Programs: 3D models of diseases related to subjects in the pediatric surgery book created were acquired using Autodesk 3ds Max; then, they were revised in response to the feedback of pediatric surgeons (Figure 10 ). Figure 10 Creation of 3D models of related disorders using CAD programs Three-dimensional models of diseases regarding subjects constituted the pediatric surgery book were acquired. CAD: computer-aided design (c) Transformation of Radiologic Records Into Video Format: Selected findings that would hold importance in the prognosis of diseases mentioned in the pediatric surgery book were highlighted on radiological views and recorded in video format. (d) AR Application Development: Prepared 3D models and videos of radiological views were imported to a developer program named Unity (Unity Technologies, Copenhagen, Denmark). The platform of the AR application was acquired into Unity using the Vuforia tool. Required arrangements were made in the settings window of Unity. There are two fundamental elements of AR application development. The first of them is the specificity of the activation of application in response to the detection of the desired surface or pattern. The second is the content, such as animation, 3D model, or video, that will be seen on the surface or pattern. At the first step of the application development, selected pages of the pediatric surgery book were used as an activating surface. In the next step, prepared 3D models and videos of radiological views were matched with activating surfaces in Unity. C# codes, which enabled users to move, magnify, or turn 3D models by finger movements, were written, and these codes, called scripts, were added to 3D models and videos to make our mobile application interactive. Designs of the mobile application interface involving button designs and backgrounds were acquired using Adobe Photoshop CC (Adobe, Mountain View, California, USA) and implemented in the mobile app by importing into Unity. All of the prepared components were then combined (Figure 11 ). Figure 11 AR application development Accompanying Augmented Reality application was prepared. AR: Augmented Reality Before launching the product in the market, we employed multiple methods to validate the functionality and identify bugs in the app. Following this, the final version of the app was tested on several devices with different resolutions. Then, we applied to launch our app in the market. It was offered as an open access app to students for testing. 4. Construction of the Mirror World Simulation (a) 360-degree video recording: 360-degree videos of three selected places, including classroom, operation room, and meeting room, were recorded (Figure 12 A, B, C). Then, video file sizes and color/light settings were adjusted using the Gear 360 video (Samsung, Seoul, Korea). Figure 12 360-degree video recording 360-degree videos of three selected places involving classroom (A), operation room (B), and meeting room (C) were recorded. (b) Construction of Stages in the Digital World: Double layer globes used as stages for 360-degree videos were acquired using Autodesk 3DsMax and positioned in a virtual environment (Figure 13 ). Figure 13 Construction of stages in the digital world Double-layer globes were acquired to use as stages for 360-degree videos. (c) Combining 360-degree Videos With Globes: 360-degree videos were imported into Unity to combine with globes and compressed a second time. Then 360-degree videos were implemented into the inner layer of globes using the video tool of Unity, and virtual lights were added into each of the globes. Autoplay and Loop settings were selected (Figure 14 ). Figure 14 Construction of the Augmented Reality application (d) Developing the AR Mobile App: The Google ARCore platform was settled. AR Camera and surface detection scripts were added. 360-degree videos combined with globes were added as child files. Then, player settings that were required to develop the AR app in Unity were adjusted. The AR app was then finalized and built for Android devices and exported as. apk file. Discussion 3D-printed models of congenital anomalies 3D printing was the first immersive technology we used to construct educational materials. It has robust indications in health care involving dentistry, anatomical models, medical devices, tissue engineering scaffolds, tissue models, and drug formulation [ 10 ]. Particularly in surgical specialties, 3D printing has wide-range applications, including preoperative planning, patient education, patient-specific prostheses, orthoses and grafts, surgical devices, training, education, and case presentation [ 10 - 12 ]. In the present descriptive technical report, we utilized 3D printing to acquire models of five different pathologies in pediatric surgery, including anorectal malformation, esophageal atresia, vesicoureteral reflux, choledochal cyst, and jejunoileal atresia [ 7 ] (Figure 15 ). Figure 15 3D printed models of congenital anomalies 3D printing enabled us to demonstrate complex anatomical relationships, implement active learning strategies in lectures, and construct case-specific and curriculum-specific models. The constructed models were used in pediatric surgery clerkship (Video 1 ). Video 1 Surgical education materials were used in clerkships Simulation models based on 3D printing and hydrogel In addition to 3D printed models, we harnessed 3D printing to construct a flexible model of the kidney [ 9 ]. Contrary to the earlier models, for the kidney model, we used 3D printed molds acquired using inverse geometry and filled with PVA-based hydrogel as a tissue-mimicking material. Different materials are used in literature as tissue-mimicking material for training surgical procedures and preoperative planning [ 13 - 15 ]. Our preferred tissue-mimicking material and workflow led to flexible and low-cost models that enabled students to train in the fundamental surgical skills, such as suturing. However, further research is required to validate the characteristics of the tissue-mimicking material and the effectiveness of the kidney model. Hybrid pediatric surgery book based on AR Even 3D printed models and hydrogel-based models had a third dimension: they weren't appropriate for "flipped classroom" or "self-paced learning" concepts and were limited in number. We constructed the hybrid pediatric surgery book based on AR to overcome these limitations. AR would be defined as a system that blends real and virtual worlds and enables users to perceive digital virtual visual content through real-life and real-time interaction. Compared to Virtual Reality, AR does not cut the connections of users with reality and surrounds them with a virtual world; instead, AR adds a layer onto real life. Another difference between AR and VR is device requirement. AR can be used with highly accessible devices, even with personal smartphones, whereas, VR requires relatively higher cost special devices, which was why we preferred to use AR to enhance the content of the books. The hybrid pediatric surgery book based on AR enabled students to examine spatial relationships and complicated pathologies and interact with them by walking around, zooming in/out, moving, and turning them. In addition to models, anonymized radiological views were added to the related content, which allowed students to easily access the radiologic views, look at their highlighted or raw versions, and examine them in detail by zooming in (Figure 16 ). More importantly, all content was usable through personal devices such as smartphones or tablets. Students were able to access models and radiological views when they wanted and where they wanted (Video 2 ). Despite the potential of AR to accompany the course curriculum, according to our knowledge, there are still very few examples in literature [ 8, 16 - 19 ]. However, we believe that enhancing the curriculum with AR would facilitate the learning of students and would be applicable in different surgical and non-surgical specialties with minor changes. Figure 16 Hybrid pediatric surgery book based on AR Video 2 A hybrid pediatric surgery book based on Augmented Reality was used in lectures 4. Mirror World Simulation Different internal and external factors involving restricted duty hours and a decreased indication of some surgical procedures led to a decrease in students' exposure to different surgical environments. The COVID-19 pandemic exacerbated that problem by limiting the number of surgeons in-house, case volumes, and in-person learning opportunities [ 16 ]. Mirror World is defined as a constructed copy of the real world in digital form [ 20 ]. We embedded 360-degree video recordings in Mirror World Simulation to acquire portals to different surgical settings including an operation room (OR), classroom, and a meeting room (Figure 17, Video 3 ). Using the operation room portal, students could experience the surgical procedure as if they were present in OR and review the surgical procedure from different angles. The classroom portal would increase student engagement in lectures, and the meeting room would help incorporate crowded student groups in meetings. We believe that Mirror World Simulation may be an effective solution for the new normal and may help the trainees on different levels, from undergrads to residents, to experience different surgical settings from various angles without time or place restriction. Figure 17 Mirror World Simulation Mirror World Simulation enabled students to experience different surgical environments. Video 3 Mirror World Simulation enabled students to experience different surgical environments Conclusions In our university, we constructed pediatric surgical education materials based on immersive technologies by the collaborative efforts of pediatric surgeons, lecturers, and medical students. We believe using immersive technologies in medical education may help lecturers and students in the new era with transforming pedagogy. In this paper, educational materials based on immersive technologies were described, and a detailed workflow was shared. We hope our workflow may serve as a valuable resource for other students and lecturers to develop their educational materials.

10. 7759/cureus. 13869

Cureus

Stem Cell Therapy in the Management of Fracture Non-Union – Evaluating Cellular Mechanisms and Clinical Progress

Bone, as a physiological and anatomical construct, displays remarkable intrinsic healing capacity. The overwhelming majority of fractures will heal satisfactorily, if aligned anatomically, compressed and immobilised appropriately. Of the 10% of fractures that do not heal, even under ideal mechanical and biological conditions, further consideration must be given to augment bone healing. Management strategies for non-union pose a significant clinical challenge to the practicing orthopaedic surgeon. Stem cell therapy is beginning to demonstrate significant potential for augmented bone repair in the context of non-union. This review attempts to contextualise the function of stem cells within this clinical setting, reviewing the relevant cellular mechanisms and clinical applications. From evaluating the literature base, there is a lack of high-quality evidence examining the role of mesenchymal stem cells (MSCs) within this research focus. Appropriately designed randomised controlled trials are required to evaluate this research area further, with a view to guiding future treatment options for the practicing orthopaedic surgeon.

Introduction and background Bone, as a physiological and anatomical construct, displays remarkable intrinsic healing capacity [ 1 ]. The overwhelming majority of fractures will heal satisfactorily, if aligned anatomically, compressed and immobilised appropriately [ 2, 3 ]. Of the 10% of fractures that do not heal, even under ideal mechanical and biological conditions, further consideration must be given to augment bone healing. Management strategies for non-union pose a significant clinical challenge to the practicing orthopaedic surgeon. A widely-accepted concept for augmenting bone healing in the non-union clinical setting involves employing the ‘diamond concept’; attempting to ensure a combination of an optimal mechanical environment, growth factors, osteoconductive scaffolds and stem cells [ 3 ]. Stem cell therapy is beginning to demonstrate significant potential for translational clinical application, across a variety of surgical and medical specialties. Both in vitro and in vivo experimental models have demonstrated promising results; positive therapeutic effects relating to augmented bone repair in the context of disease have been produced [ 1, 3 ]. This review aims to explain what a stem cell is, relating this to their physiological function and highlighting the variety of stem cells present, whilst discussing methods of stem cell isolation for use in bone repair models. This review will also aim to evaluate the fundamentals of bone repair at the cellular level, focusing on the context of fracture non-union. An insight into future stem cell application in bone repair and disease will also be included, with limitations of the current evidence base also explored. We also hope this paper serves as a useful primer and aide-memoire for the practicing orthopaedic surgeon on the evolving concepts of bone healing and stem cell therapy in the context of non-union. Review Methodology A focused literature search was performed using the Pubmed/MEDLINE index. Inclusion criteria included the use of human studies, papers written in the English language and full papers. Published abstracts, wholly non-human studies and non-English language papers were excluded. All relevant literature was screened for inclusion, with relevant papers proceeding to further full screen evaluation. Stem cell biology To evaluate the advances in stem cell therapy relating to bone repair and disease, the fundamentals of stem cells must be explained to provide clinical context. 'Stemness' is a term used to describe a cell with stem-cell qualities compared to a non-stem cell [ 4 ]. A succinct definition of a stem cell is that of a self-renewing cell that can differentiate in a symmetric and asymmetric fashion. This is to say that a stem cell can divide by clonal expansion leading to the production of more of the same stem cells and it can also differentiate into different cell progeny [ 5 ]. Concerning stem cell potency, a 'true' stem cell is multipotent, differentiating into more than one cell type. In contrast, a progenitor cell derives from a multipotent stem cell but is unipotent, dividing into a single progeny [ 6 ]. Dependent upon the type of stem cell, it can demonstrate pluri-, toti-, multi- or unipotent developmental potential [ 7 ]. Table 1 demonstrates definitions of these specific terms in relation to cell potency [ 3, 8, 9 ]. Table 1 Definition of specific terms relating to cell potency Totipotent Pluripotent Multipotent Unipotent Relative Potency High Medium Low Low Examples - Fertilised ovum - Embryonic Stem Cells - Induced Pluripotent Stem Cells - Haematopoietic Stem cells - All cell types Developmental potential Any cell type Can differentiate into cell types from 3 primordial germ layers Differentiate into limited range of cell types Can only give rise to one single cell type Self-renewal describes the mode of proliferation specific to stem cells. For a cell to be self-renewing, it must produce a daughter cell with the same replicative potential of the mother cell. An example of this type of cell lineage are neural stem cells, where both daughter and mother cells are multipotent [ 10 ]. Types of stem cell Stem cells can be broadly divided into two categories: those of embryonic origin and those of non-embryonic origin. Human embryonic stem cells (ESCs) are considered the ‘gold standard’ for developing understanding of stem cell pluripotency [ 11 ]. ESCs are obtained from the blastocyst stage of the embryo; cells obtained are pluripotent. Undifferentiated stem cells have also been sourced [ 12 ]. ESCs can differentiate into any cell derived from the three primordial germ layers, thus can be considered pluripotent [ 2 ]. Human embryonic stem cell research remains an area of intense ethical and political scrutiny, given the method of harvest ultimately requires embryonic destruction [ 3 ]. A variety of other stem cells from a non-embryonic origin exist, such as amniotic epithelial cells and fetal stem cells, however thorough explanation of these is beyond the scope of this review [ 13 ]. The focus of this review concerns adult stem cells (ASCs). An example of these are haematopoietic stem cells (HSCs), which have the capacity to derive all types of mature blood cell [ 14 ]. There are a range of cell types in this category, but this review will focus on stem cells which can differentiate into tissues of a mesodermal origin and assist in bone repair and disease. Cells of a mesodermal origin include chondrocytes, muscle cells, adipocytes and skeletal tissue cells [ 15 ]. One stem cell population that derives the aforementioned cell types are mesenchymal stem cells (MSCs), cells found primarily in adult bone marrow [ 16 ]. MSCs are low in number compared to the rest of the cell population in the bone marrow and they originate primarily from the stromal cell system; a system of non-haematopoietic connective tissues [ 17 ]. The mesenchyme is mesodermal-derived tissue that can develop into a variety of connective tissue types. The isolation of MSCs requires a bone marrow aspiration from the superior iliac crest of the donor. This is then centrifuged to separate the mononuclear layer of cells and then this cell layer is cultured to allow the MSCs present to adhere to a plastic surface, enabling extraction [ 18 - 20 ]. The integral clinically relevant property of MSCs is their ability to differentiate into various types of connective tissue, and their ability to sustain self-renewal. MSC cultures must be exposed to a different combination of substances to produce different cell populations. Table 2 highlights these combinations [ 18 ]. Table 2 Conditions needed for MSC differentiation into different cell types TGF: Transforming growth factor; PPAR: Peroxisome proliferator-activated receptor; EGF: Epidermal growth factor; MSC: Mesenchymal stem cell. Cell type Conditions Chondrocyte Requires TGF-beta, a nutrient medium which does not contain serum and a three-dimensional culture zone Osteoblast Requires the presence of fetal bovine serum, ascorbic acid-2-phosphate, beta-glycerol-phosphate and dexamethasone Adipocyte Requires PPAR-gamma and fatty acid synthetase Neuron Has been shown to occur with the presence of isobutylmethylxanthine and dibutyrul cyclic AMP. Also shown to occur with the presence of brain-derived neurotrophic factor and EGF. Concerning clinical musculoskeletal application, the scope of use is far-reaching [ 3, 13, 21 ]. Pre-clinical animal models have demonstrated that an intramuscular injection of MSCs into murine muscle can help treat Duchenne's muscular dystrophy [ 22 ]. Translational clinical benefit has been demonstrated by autologous transplantation of MSCs of a specified myogenic lineage in patients with Duchenne’s muscular dystrophy, successfully increasing the amount of dystrophin-competent cells [ 23 ]. The therapeutic effects of MSCs in graft-versus-host disease (GVHD) have been extensively studied and demonstrate the role of MSCs in immunomodulation. Studies have illustrated that MSC administration diminishes the immune response associated with GVHD by modifying specific unwanted actions of immune cells, thus decreasing chances of organ rejection [ 24 ]. Multiple further clinical uses of MSCs have been demonstrated, however their application in the fracture non-union setting appears less fully understood [ 3 ]. Mesenchymal stem cells are introduced into the clinical site either by direct injection into the clinical site, or along bioengineered scaffolds which seek to promote proliferation, MSC survival and are considered more osteoconductive surfaces in comparison to the host surface [ 25 ]. Principles of normal bone repair and maintenance: fracture healing Consolidation of a fracture site occurs through a combination or primary (direct) bone healing, secondary (indirect) bone healing, or a combination of both mechanisms. This can be augmented by micro motion at the fracture site, however an overload of motion can lead to non-union of the fracture [ 26 ]. A widely-accepted biomechanical concept governing fracture stability pertains to the fracture site accepting 2-10% strain, with strain applied outside of these margins promoting non-union [ 27 ]. Direct bone healing (primary healing) occurs when fracture fragments are compressed and are directly opposed. Compression is produced by surgical implantation, for example dynamic compression plating, or by immobilisation (for example splinting or cast immobilisation). Implants directly compress the fracture fragment, which on mechanical loading stimulates bone formation allowing the joint to bear load as time progresses. Initially the implant bears mechanical load, but as healing progresses and more bone is laid down, load bearing is transferred gradually to newly formed bone. Osteoclasts and osteoblasts produce bone by forming constructs known as ‘cutting cones’, whereby osteoclasts resorb bone and osteoblasts deposit bone following this. Only bone present at the implant-bone interface is restored with this method. Repair needs to occur outside of the interface to ensure mechanical integrity of the construct, thus a process known as gap healing occurs, whereby pockets of fractured bone not at the interface are repaired by periosteal osteoblastic activity. Approximately 12 weeks after implantation the newly formed bone acquires the tensile strength of normal bone [ 28 ]. If the fracture ends are not in continuity or under sufficient compression, an external callus normally forms at the fracture site, in the presence of a fracture haematoma which acts as a nutrient medium, osteoconductive scaffold and fibrin mesh for subsequent bone deposition. This differentiates through various connective tissue forms until bone is formed by osteoblasts. Fragile woven bone created by this process is remodelled into stronger lamellar bone. This can be considered the classical process whereby secondary (indirect) bone healing occurs [ 2, 26, 29 ]. Within the quoted literature, fracture healing is dependent principally upon five conditions at the fracture site. These include there being a sufficient blood supply, a fracture-haematoma, a population of cells diverse enough to enhance healing and provide nutrition to the site, a constant mechanical environment, and an osteoinductive scaffold for repair to occur on [ 1 - 3, 26 ]. Fracture non-union: definition and aetiology For the purposes of the practicing orthopaedic surgeon, a non-union is a fracture that will not heal without further intervention [ 30 ]. The causes of non-union can be broadly categorised into three separate categories. As per the work of Stewart, these can be summarised as host factors, biological factors and mechanical factors [ 31 ]. Host factors well validated within the literature include smoking, diabetes, immunosuppression, compliance with post-operative advice and steroid use. Biological factors include presence of active infection, adequacy of soft tissue coverage, bone defect and vascular supply to the fracture site. Mechanical factors include fracture configuration and adequacy/suitability of surgical fixation [ 21, 31 ]. Clinically, patients with established non-union report persistent pain around the fracture site, with the possible awareness of fracture fragment motion, and reduced functional capacity of the affected limb [ 31, 32 ]. Distinguishing between different types of non-union is important for clinical practice and guides surgical management. The classification of morphological type of non-union allows the clinician to target treatment appropriately. Non-union can be classically divided into hypertrophic union, whereby the fracture ends are viable and capable of mounting biologic reaction, and atrophic, whereby fracture ends are inert and incapable of mounting a biologic reaction [ 33 ]. The accepted method of classification divides non-unions by cause and fracture configuration into further sub divisions [ 34 ]. See Table 3 for further sub divisions [ 29, 33, 35 ]. Table 3 Types of non-union based on morphological classification of fracture site Type of non-union Hypertrophic vs Atrophic Causes Characteristics Elephant foot Hypertrophic Joint mobilisation and weight bearing before the repair process has a chance to start Hypertrophic and lots of callus present Horse foot Hypertrophic After ineffective surgery involving poor fixation with implants Slightly hypertrophic and very little callus present Oligotrophic Some debate exists – cannot be definitively considered atrophic/hypertrophic Fracture is significantly displaced Some vascularity exists with some callus present but no evidence of sufficient bridging callus Torsion wedge Atrophic Occur after the treatment of tibial fractures involving implants Presence of a small fracture fragment which has a poor blood supply Comminuted Atrophic Occur due to failing and breaking of implants used for internal fixation Presence of one or more necrotic fracture fragments Defect Atrophic Occur due to fracture or through infection Loss of a small part of the diaphysis. As time progresses, fracture fragments become atrophic Atrophic Combination of torsion wedge, comminuted and defect – atrophic End-stage result of torsion wedge, comminuted and defect Fragments of fracture are missing, scar tissue presence, lack of osteogenesis, limb inactivity has led to atrophy. Management of non-union is dependent on complex decision making relating to surgical, implant and patient factors. Broadly speaking, hypertrophic non-union is treated by ensuring adequacy of mechanical stability at the fracture site. As the fracture site is considered biologically active, callus formation will be considered sufficient without external modification [ 31, 36 ]. Atrophic non-union management has been principally focused on restoring a normal biological environment for bone production to occur, with a focus on autologous bone grafts (ABG) to bridge segmental bone defects. The use of MSCs in the context of atrophic non-union presents an alternative viable therapeutic option for such defects [ 3, 31 ]. Preparation and conditioning of MSCs The successful differentiation of MSCs to osteoblasts is the basis of potential treatment options in non-union; autologous stem cells could serve to provide a native osteoblast population for the recipient, stimulating bone growth at the fracture site. The preparation and conditioning of MSCs is fundamental in facilitating differentiation into osteoblasts. Historically, the vital factors governing successful differentiation were considered scaffold composition, the surface scaffold topography and the interaction between the stem cell and the scaffold. Two approaches to conditioning stem cells specifically down an osteoblastic lineage have been demonstrated in the literature: the ‘top-down’ and ‘bottoms-up’ approaches [ 37 ]. The first approach is considered a ‘top-down’ approach, whereby bioengineering focus is placed on optimising the extracellular environment to augment MSC differentiation down. Success has been demonstrated by including the combination of MSCs with hydroxyapatite (HA) granules and bone morphogenetic protein-2 (BMP-2), to promote osseointegration by utilising the biocompatibility and porosity of the added materials [ 32, 38 ]. Dilogo et al. appear to have demonstrated effective clinical results in promoting fracture healing and scaffold integration by utilising MSC and a ‘top-down’ approach in treating segmental bone defects in the non-union setting, utilising MSCs, HA granules and BMP-2 to facilitate both an osteogenic and osteoinductive environment [ 38 ]. The ‘top-down’ approach can be considered the more well-validated tissue engineering technique in stem cell differentiation setting [ 37 ]. As discussed by Kim and Mikos, the use of the ‘top-down’ approach invariably incorporates the use of external factors such as BMP-2 at supra-physiological levels, leading to possible complications reported as over-activation of osteoclasts, heterotopic ossification and autoimmune dysregulation [ 37 ]. To bypass these effects, the ‘bottoms-up’ approach can be considered. The ‘bottoms-up’ approach focuses on genetic modification to MSCs, leading to regulation of gene expression and thus theoretically controlling differentiation at the intrinsic genetic level. Novel gene delivery vehicles are currently being evaluated for their proposed efficacy in providing stable genetic modification for MSCs. The complex bioengineering concepts underpinning this are beyond the scope of this review. A further growth-area focuses on the isolation of specific MSC sub-populations [ 39 ]. MSCs can be considered an inherently heterogeneous cell population. Isolating surface markers and morphological characteristics, for example cell size, of MSCs that have successfully undergone osteoblastic differentiation could allow for a more reliable and robust differentiation process [ 37, 39 ]. Clinical applications The first reported clinical use of stem cell engineering in fracture repair was carried out by Vacanti et al. in 2001 [ 40 ]. A patient with an avulsed digital phalanx of the hand received stem cell therapy in addition to traditional reconstructive surgical techniques. The traditional reconstructive techniques of splinting and skin grafting occurred. Additionally, a pocket of periosteal cells was introduced into the surgical site on a bioengineered scaffold. The patient reported a successful functional outcome at three months. The use of periosteal cells to augment bone healing at the surgical site has led to further focused clinical work in this field [ 29, 40 ]. The work of Quarto et al. can be considered the seminal work in stem cell engineering and non-union [ 41 ]. Bone marrow was extracted from patients suffering from significant segmental bone defects. This was then cultured ex vivo and the MSCs present in the bone marrow differentiated into osteoprogenitor cells, which were then placed on a bio-engineered scaffold and transferred back to the patients. Abundant radiological callus formation was demonstrated, alongside functional recovery. Quarto et al. did not explicitly state the context of non-union but given the post traumatic setting and subsequent bone loss, the non-union is likely to be considered atrophic in nature and thus would be more appropriate for biologic augmentation [ 41 ]. Clinical studies have highlighted that patients injected with bone marrow of which the MSC concentration was less than 1000 MSCs/ml did not experience augmented bone growth, whereas patients injected with bone marrow aspirate of which the MSC concentration was around 1500 MSCs/ml experienced rapid healing [ 3, 41 ]. There does not appear to be an agreed consensus on the optimal concentration for inducing bone growth, with further clinical work required to ascertain the appropriate concentration for inducing augmented bone healing. Reflecting upon the five conditions for fracture healing highlighted earlier, each one needs to be fulfilled for bone repair to occur. An issue of concern relates to the fact that MSC numbers obtained from a bone marrow aspiration are highly variable from patient to patient, thus providing an effective, reproducible treatment is proving difficult. This variability can be overcome by isolating a small number of MSCs and allowing them to proliferate ex vivo, but this method of treatment is logistically difficult, technically demanding, and costly. Bioengineering solutions have been theorised to negate a lack of MSCs, including genetically modifying MSCs to over express BMP, leading to an enhanced osteogenic effect of each MSC. Another possible future approach involves implanting a scaffold with MSCs into a highly vascularised tissue like skeletal muscle, allowing the scaffold to grow in vivo and obtain sufficient vascularisation and growth factors. This could then be transferred to the site of non-union [ 42 ]. These future challenges present the next exciting research chapter in the bone repair field [ 21 ]. Multiple innovative surgical strategies have been developed for bridging bone defects, such as vascularised bone grafting, employing the Ilizarov technique and the Masquelet technique [ 43 ]. One such technique that has had limited success using stem cell therapy is the Masquelet technique. In summary; the Masquelet technique involves the use of a temporary cement spacer to bridge the bone defect, followed by staged grafting with non-vascularised cancellous bone graft [ 44 ]. This technique takes advantage of the ‘induced membrane’, whereby a membrane forms around the temporary spacer, allowing theoretically for low bone graft resorption, anatomical bridging of defect and an absence of invasion by surrounding soft tissues [ 43 ]. At the time of bone grafting, native stem cell populations have been introduced into the fracture site, demonstrating successful bone healing in an isolated case by Dilogo et al. [ 45 ]. The role of stem cells in the context of the Masquelet technique appears to be an area that requires further research development, realising the potential of the induced membrane effect to enhance the efficacy of MSCs, by theoretically providing a suitably osteogenic and inductive environment for bone formation. To date, no robust clinical data evaluating the role of stem cells in comparison to control subjects is available, with research into the Masquelet technique still relatively underdeveloped, in comparison to the well-established techniques of vascularised bone grafting and the Ilizarov technique. Another potential therapeutic option involving a different population of human stem cell has also been explored clinically. The use of human multipotent adipose-derived stem cells (hMADSCs) could potentially counteract some of the problems associated with isolation of MSCs. Obtaining hMADSCs from adipose tissue is non-invasive, and adipose tissue is generally abundant in-patient populations. Adipose-derived stem cells are also considered to have a high degree of cellular plasticity; meaning they can differentiate into a variety of cell populations [ 46 ]. One study illustrated that Adipose derived stem cells (ASCs) injected into a murine model with non-union fractures provided a substantial bone growth. At the initial stages of fracture consolidation, the ASCs differentiated into cells of haematopoietic and osteoblastic lineage. A further novel effect was demonstrated, whereby ASCs induced a local paracrine effect on the fracture site that stimulated osteogenesis and angiogenesis [ 47 ]. Veriter et al. have demonstrated positive effects clinically using hMADSCs to augment bone repair in critical bone defects, either due to oncological resection or congenital defect [ 46 ]. Grafts were developed with incorporated ASCs, alongside demineralised bone matrices, which were then incorporated into the site of defect. Four-year follow-up demonstrated osseointegration and biological integration of the grafts with evidence of union radiologically. These results, albeit promising, are not reflected in the underdeveloped research base relating to ASCs in clinical use. Concerns remain relating to immunomodulation and possible immunosuppressive capacity of ASCs; mechanisms relating to possible tumorigenicity have been proposed [ 48 ]. Percutaneous autologous bone-marrow grafting has now been used in several studies with success, illustrating that MSCs present in extracted bone marrow are playing a useful therapeutic role in aiding the repair of a fracture. A correlation has also been uncovered relating to the efficacy of the treatment and the number of osteoprogenitor cells present in the graft, merely underlining the importance of the initial stem cells in the treatment process [ 35 ]. Further promising clinical trials have been conducting using percutaneous bone marrow for long bone reconstruction, with injection performed under fluoroscopic guidance [ 49 ]. Emadedin et al. demonstrated a safe and effective approach by utilising percutaneous autologous bone marrow-derived MSC implantation in long bone atrophic union, but further randomised double-blind trials are required to fully assess effect-size and safety of this orthobiologic approach [ 49 ]. Strengths and limitations Significant clinical limitations exist relating to the long-term efficacy of MSCs in the treatment of non-union. Whilst a short-term positive clinical efficacy has been demonstrated by recent work, long-term 5-10 year follow-up of patients has not been evaluated within the current literature base [ 3, 21, 31, 38 ]. Long-term follow-up analyses of current patient cohorts are needed to assess the anatomical and functional integrity of any bone repair, and whether any adverse effects of MSC use can be identified in the medium-to-long-term follow-up period. Further, undertaking MSC research in this context appears to require translational academic expertise in musculoskeletal tissue engineering and orthopaedic surgery. Barriers to access are reflected in the paucity of robust translational literature, with most research produced in tertiary academic centres currently. Further to this end, there appears to be distinct heterogeneity relating to method of obtaining stem cells, isolation, and subsequent introduction into the defect. Such heterogeneity makes comparison between methods challenging. Critically, there does not appear to be an internationally agreed definition for ‘non-union’, either defined through radiological or clinical parameters. Thus, production of robust meta-analyses and quantitative analysis of stem cell use within this setting would appear unfeasible, given seemingly equivalent studies would have inherent methodological heterogeneity. Heterogeneity in the assessment of fracture consolidation makes discerning whether MSCs are clinically beneficial difficult, with direct comparison between study outcomes poor due to this lack of uniformity in outcome measurement. Outcome measurement should focus on not just radiological outcome, but patient reported outcomes, combining holistic measurement of pain and functional status [ 3 ]. Further, prospective cohort analysis appears to be the prevailing study methodology, with no control subjects appearing within the scope of any such trial of MSC application in non-union. Without control subjects, examining the true effect of MSCs is challenging both from a clinical and methodological standpoint. Conclusions This review references and appraises the key literature relating to stem cell usage in the non-union setting, contextualising the key research advances and areas of research focus. We have also summarised the salient clinical and basic science literature relating to bone repair at the cellular level. In summary, the role of MSCs in the treatment of non-union does not appear to be fully understood. Understanding surrounding isolation, culture, and implantation of stem cells in the context of non-union is based on principally heterogenous case studies, lacking randomisation and homogenous research techniques. There is a lack of high-quality evidence examining the role of MSCs within this research focus. Appropriately designed randomised controlled trials are required to evaluate this research area further, with a view to guiding future treatment options for the practicing orthopaedic surgeon.

10. 7759/cureus. 13989

Cureus

Are 3D Printing Templates an Advantage in Upper Thoracic Pedicle Screw Fixation?

Background This study aims to compare the clinical results of patients with upper thoracic vertebral fractures treated with pedicle screw and posterior spinal fusion with preoperative surgical planning and 3-dimensional (3D) modeling and patients treated with freehand screws. Methods Fifty patients who underwent pedicle screw placement with a diagnosis of upper thoracic fracture between June 2018 and October 2020 were included in our study. Pedicle screws were used in 25 patients (group 1) after the planning was completed with the help of 3D preoperative printing and modeling. Pedicle screws were applied in 25 patients in the control group (group 2) using the freehand technique. Intraoperative bleeding amount, operation time, and correct screw placement data in both groups were recorded. Results The operation time was 134 ± 22 minutes for group 1 and 152 ± 38 minutes for group 2. The difference in operation times was found to be statistically significant (p < 0. 05). Based on axial and sagittal reconstruction images, the accuracy rate of pedicle screw placement (grades 0 and 1) in group I was 96. 6% compared to 83. 6% in group II. The minor perforation rate (grade 1, <2 mm) was 5. 8% in group I compared to 11. 8% in group II. The moderate perforation rate (grade 2, 2-4 mm) was 3. 4% in group I compared to 14% in group II. The severe perforation rate (grade 3, >4 mm) was 2. 3% in group II; however, misplaced screws were not associated with neurological deficits. The difference in overall accuracy rates between the two groups was significant (p < 0. 05). Conclusions For 3D models of upper thoracic pedicle screw insertion, guide plates can be produced inexpensively and individually. It provides a new method for the accurate placement of upper thoracic pedicle screws with high accuracy and secure use in screw insertion.

Introduction Spinal cord injuries are common in patients with upper thoracic vertebral fractures due to the smaller diameter of the spinal canal compared to the cervical and lumbar regions. Posterior screw fixation is applied in the upper thoracic region and is used in trauma, segmental instability, kyphosis, scoliosis, infection, and tumor treatments [ 1, 2 ]. Pedicle screw fixation provides rigid intervertebral fixation but associated with complications, such as artery injury, nerve root damage, and dural damage [ 3, 4 ]. Advantages of surgical stabilization in upper thoracic vertebral injuries are correction of sagittal and coronal balance and neurological decompression in kyphotic fractures. Additionally, fixation and fusion prevent hyperkyphosis [ 5 ]. Anatomical studies indicate that thoracic pedicle screw insertion in the upper segments is more difficult than that in the lower segment due to the narrower diameter of the pedicles [ 6 ]. Moreover, the visualization of the upper thoracic region by X-ray and C-arm fluoroscopy during surgery is limited. This limitation of imaging in the preoperative period increases the risk of screw malposition. A number of studies reveal that inaccuracies in placement using these traditional techniques range from 10% to 50% [ 7 ]. In the authors' experience with image guidance in over 1500 cases, several potential pitfalls have been identified that could lead to sub-optimal results when using intraoperative spinal navigation [ 7 ]. Currently, 3-dimensional printing technology is used in the preoperative evaluation of patients' anatomy, prosthesis, and implant applications. 3D printers help to increase surgical success by providing a preoperative simulation of surgical approaches [ 8, 9 ]. Models created with 3D compression are used in medical and surgical fields, such as cranial surgery, maxillofacial traumas, tissue engineering, chest deformities, and complex spine surgery [ 8, 10 ]. Creating patient-specific 3D printing models reduces complications during surgery by providing the surgeon with preoperative surgical planning and application. This study aims to compare the clinical results of patients with upper thoracic vertebral fractures treated with pedicle screw and posterior spinal fusion with preoperative surgical planning and 3D modeling and patients treated with freehand screws. Materials and methods This article was previously posted to the Research Square preprint server on January 07, 2021. Patients diagnosed with upper thoracic fractures between June 2018 and October 2020 were evaluated according to Thoraco‐Lumbar Injury Classification and Severity score (TLICS). Patients were deemed as appropriate candidates for spinal stabilization based on TLICS score of 5 or greater; patients suffering from upper thoracic trauma at T1-T6 segment accompanied with incomplete or complete spinal cord injury, and requiring surgery; patients with senile osteoporotic vertebral fracture and other upper thoracic trauma, not requiring surgery. The final diagnosis was based on the thoracic vertebral fracture and spinal cord injury in thoracic magnetic resonance imaging. Patients who had previously undergone spine surgery, preoperative radiation, chemotherapy, or had recurrent tumors and younger than 18 years were excluded from the study. Fifty patients with pedicle screw implantation were included in this study. Pedicle screws were applied in 25 patients (group I) after planning was completed with the help of 3D preoperative modeling. Pedicle screws were applied in 25 patients in the control group (group II) using the freehand technique. 3D printing was performed in the Kutahya Health Sciences University Research and Application Laboratory. Digital design and 3D printing Preoperative computed tomography (CT) images of 25 patients diagnosed with an upper thoracic fracture in the Neurosurgery Clinic were used for 3D models. Preoperative digital imaging and communications in medicine (DICOM) images of each patient were reconstructed using 32-channel computed tomography at a slice thickness of 0. 625 mm and a planar resolution of 0. 35 mm (Aquilion™ Large Bore CT, Canon Medical Systems, Tustin, USA) (Figure 1 ). Figure 1 Preoperative CT view of the upper thoracic fracture. CT images containing approximately 450 sections for each model were transferred to the 3D Slicer (version 4. 10. 1, Boston, MA, USA) program to create a 3D vertebral model. Using this software, the images were used to create 3D models of the vertebral region related to the complex surface treatment method. Whereas only the vertebral model was created from preoperative DICOM images (Figure 2 ). The templates to be used for pedicle screw placement were modeled in SolidWorks 2015 SP5 software (SolidWorks Corporation, Waltham, MA, USA) according to the measurements obtained from CT images and 3D vertebra models. Figure 2 Modeling of the vertebrae in the 3DSlicer program. 3D vertebrae model and template data were saved in stereolithography (STL) format and transferred to Ultimaker Cura (version 4. 7. 1) (Ultimaker B. V. , Utrecht, Netherlands) software. Printing parameters for preoperative vertebral models and templates were prepared in Ultimaker Cura software. The following printing parameters were used for Ultimaker 2 Extended 3D printer and polylactic acid (PLA) in Ultimaker Cura software for the printing of preoperative models: 0. 4 mm nozzle diameter, 200°C nozzle temperature, 70°C build plate temperature, and 70% filling rate. Preoperative planning studies were performed on vertebra models and templates by the relevant surgeon (Figure 3 ). Figure 3 T4-T6 vertebrae prepared for preoperative planning. Operational methods Patients’ operations were performed by the same surgeons. Preoperative modeling and planning of the patients to be operated on was done with the help of 3D printing and modeling methods. Patient-specific full-scale spine models were available for reference at the time of surgery. Pedicle screws were placed from anatomic regions previously determined by planning and checked with fluoroscopy. In the control group patients, pedicle screws were placed using the freehand technique and under fluoroscopic control. Evaluation of efficacy Intraoperative bleeding amount, operation time, and correct screw placement data in both groups were recorded. Intraoperative bleeding was calculated by subtracting the volume of fluid used for flushing from the total fluid volume in the suction bag. The time of insertion of each pedicle screw was recorded. A control CT scan was performed after the operation. Screw malpositions and violations of the medial and lateral walls of the pedicles were noted. The position of the screws was evaluated according to the Gertzbein classification [ 11 ]. In this classification, there are four categories for screw placement: grade 0, screws are completely within the pedicle; grade 1, perforation < 2 mm; grade 2, perforation between 2 and 4 mm; and grade 3, perforation > 4 mm. In the current study, grades 0 and 1 were considered satisfactory, whereas grades 2 and 3 were regarded as perforated. Statistical analysis The statistical analysis was performed using SPSS 24. 0 (IBM Corp. , Armonk, NY, USA) software. Data are presented as the mean ± SD (x±s), and intergroup comparisons were performed with independent-samples t-tests. The enumeration data are expressed as a ratio, and intergroup comparisons were performed with the chi-square test; p = 0. 05 was used as the statistical inspection standard. Results Of the 50 patients diagnosed with upper thoracic fractures, 25 were female, and 25 were male. The mean age of these 50 patients was 37. 3 ± 5. 9 years. No statistically significant difference was found between the groups in terms of age and gender (Table 1 ). Table 1 Comparative demographic data of both groups. Ϯ Compared with between group 1 and group 2 Group 1 Group 2 p-value Ϯ Number of Patients 25 25 - Sex 14 Male / 11 Female 13 Male / 12 Female 0. 184 Age 36. 4 ± 6. 2 38. 1 ± 5. 7 0. 260 Based on axial and sagittal reconstruction images, the accuracy rate of pedicle screw placement (grades 0 and 1) in group I was 96. 6% compared to 83. 6% in group II (Table 2 ). The minor perforation rate (grade 1, <2 mm) was 5. 8% in group I compared to 11. 8% in group II. The moderate perforation rate (grade 2, 2-4 mm) was 3. 4% in group I compared to 14% in group II. The severe perforation rate (grade 3, >4 mm) was 2. 3% in group II; however, misplaced screws were not associated with neurological deficits (Table 2 ). The difference in overall accuracy rates between the two groups was significant (p < 0. 05). Table 2 Classification of patients according to Gertzbein scoring. Ϯ Accuracy = (grade 0 + grade 1)/n x 100% Misplacement (according to Gertzbein’s classification) Group 1 (n = 174 screws) Group 2 (n = 171 screws) Grade 0 (screws are completely within the pedicle) 158 (90. 8%) 123 (71. 9%) Grade 1 (screw perforation < 2 mm) 10 (5. 8%) 20 (11. 8%) Grade 2 (screw perforation between 2–4 mm) 6 (3. 4%) 24 (14%) Grade 3 (screw perforation > 4 mm) - 4 (2. 3%) Accuracy Ϯ 96. 6% 83. 6% The operation time was 134 ± 22 minutes for group 1 and 152 ± 38 minutes for group 2. The difference in operation times was found to be statistically significant (p < 0. 05) (Table 3 ). The amount of blood loss for group 1 was 962 ± 108 mL. For group 2, it was 992 ± 114 mL. The difference in the amount of blood loss was not statistically significant (p > 0. 05) (Table 3 ). Table 3 Surgical data. Ϯ Compared with between group 1 and group 2 Group 1 (n = 25) Group 2 (n = 25) p-value Ϯ Operation Time (min) 134 ± 22 152 ± 38 p < 0. 05 Blood Loss (mL) 962 ± 108 992 ± 114 p > 0. 05 The mean TLICS scores for group 1 and group 2 were 6. 3 ± 4. 2 and 6. 7 ± 4. 1, respectively. Of the 25 upper thoracic fracture patients (group 1), five (20%) were T3, eight (32%) were T4, and 12 (48%) were T6 fractures, which were operated on by 3D modeling with preoperative planning (Figures 4 - 6 ). For these 25 patients, the concordance rate between pedicle positions studied on preoperative models and postoperative pedicle screw positions was 93. 8% for T3 fractures, 94. 7% for T4 fractures, and 98. 4% for T6 fractures (Table 4 ). Figure 4 Postoperative CT image of a male patient operated with free hand technique. Figure 5 Postoperative CT image of a female patient operated with 3D modeling technique. Figure 6 Postoperative CT image of a male patient operated with 3D modeling technique. Table 4 Compliance of preoperative planning (Group 1) with postoperative pedicle screw positions in the vertebral model. Vertebral Model with Fracture Diagnosis Pedicle Screw Size Number of Pedicle Screws Inserted Model Integrity Rate T3 4. 0 x 26 mm 24 93. 8% T4 4. 0 x 26 mm 57 94. 7% T6 6. 0 x 45 mm 93 98. 4% The 3D printing time of vertebral templates produced for each patient was 14 ± 3 minutes. The amount of PLA used for the production of each template was 2 ± 0. 42 grams. The total production cost of 25 templates produced for each patient was $5. 31. Discussion Upper and middle thoracic fractures are rare among spinal fractures. Approximately 10-20% of general spinal traumas are observed in this region [ 12 ]. Upper thoracic vertebral injuries often result in axial stress and bending with rotation and dislocation. This type of injury is generally seen at T4-T6 levels in motorcycle riders [ 13 ]. With the development of spine surgery, posterior thoracic interpedicular screwing has become more important. The key to this operation is the correct placement of the pedicle screw in one step, but it has been difficult due to safety concerns relating to the upper thoracic pedicle properties. The upper thoracic spine pedicles are smaller in diameter, and there are different angles for each spine. Thoracic pedicles are short and narrow, and their cortex is thin and fragile; therefore, thoracic pedicles are easily broken during screwing [ 14 ]. Additionally, the angles of the thoracic pedicles are different from each other, which has made the rate of error in placing thoracic pedicle screws at one time very high, causing serious consequences by damaging the surrounding tissues [ 15 ]. The anatomy of the thoracic pedicles is more complex, and screw insertion is more difficult in complex thoracic fractures and vertebral malformations. The penetration rate can reach 30-40% in the insertion of the thoracic pedicle screw with the freehand technique [ 16 ]. CT-based navigation systems are used to guide the placement of pedicle screws on the spine. Its use is not common due to its disadvantages, such as intraoperative position changes and spinal instability, lack of real-time navigation, and high cost [ 17 ]. With the use of 3D printing in spine surgeries, the production of guide plates, and provision of preoperative simulation, the accuracy of operations has increased. The fact that the accuracy is not affected by the intraoperative position and the higher reliability of guide plates provides superiority to navigation systems. Providing preoperative simulation and using the model as a guide during surgery reduce the surgeon's margin of error and operation time [ 18 ]. Controls performed with 2D fluoroscopy in upper thoracic interpedicular screwing show high error rates. Guzey et al. retrospectively examined 113 pedicle screws between T2-T8 in 24 patients without coronal deformity [ 19 ]. The control of the pedicle screws was checked during the operation by C arm fluoroscopy and postoperative CT. The faulty pedicle screw insertion rate was found to be 20. 3%, 27. 4% between T2-T5 and 14. 5% between T6-T8 [ 19 ]. Pedicular screws were applied to T4-T12 levels by five experienced surgeons on five fresh cadavers. In postoperative CTs, a faulty screw placement rate was found at a rate of approximately 41%. Of these, 21 screws were observed to be in the vertebral canal by preparing the medial wall of the pedicle [ 4 ]. In our study, in patients who underwent surgery using the freehand technique, a total of 28 (16. 3%) incorrect pedicular screw placements were observed, 24 screws grade 2 (14%) and four screws grade 3 (2. 3%). With the application of 3D printing in spine surgeries, personalized production of guide plates, and preoperative simulation of the operation on the model have increased the accuracy of operations. Lu et al. used 3D modeling as an aid to cervical pedicle or vertebral plate screw placement and proved that it can provide correct placement of the screws [ 20 ]. Customized 3D spine models and screw insertion guide plates can be used to aid screw insertion and ensure the correct insertion of screws. Mizutani et al. designed 3D models to apply cervical pedicle screws and achieved good results with guide plates in placing cervical pedicle screws [ 21 ]. Sugawara et al. created personal 3D navigation models for thoracic pedicle screws and applied pedicle screws under their guidance simply and safely. In 103 patients, 813 screws were placed with 3D guides. In postoperative CT scans, 801 screws (98. 5%) were placed without cortical violation, and no injury to the vessels and nerves was observed [ 9 ]. Xu et al. placed 56 pedicle screws in seven patients with upper and middle thoracic trauma using the 3D printing-supported preoperative plan method [ 10 ]. Regarding the placement of 56 screws according to postoperative CT images, 33 were grade 0, 18 were grade 1, four were grade 2 (perforated sidewall), and one was grade 3 (perforated sidewall, no vascular nerve injury). The accuracy rate was 91% [ 22 ]. In our study, screw placement was performed according to postoperative CT images of 174 pedicle screws placed in the upper thoracic spine in 25 patients with preoperative 3D printing support and guidance, and 158 (90. 8%) were grade 0, 10 (5. 8%) were grade 1, and six (3. 4%) were grade 2. Grade 3 positioning was not observed in any screw, and the pedicular screw placement accuracy rate was 96. 6%. Comparing the pedicle screw placement accuracy of the upper thoracic vertebrae (96. 6%) and the pedicle screw placement accuracy (83. 6%) of the freehand technique in the 3D printing-supported group, the difference was statistically significant (p < 0. 05). In the study by Pan et al. , 37 patients with spinal deformities were operated on, with group 1 (25 patients, 396 screws) supported by 3D printing and group 2 (25 patients, 312 screws) supported by the freehand method. The operation time in group 1 was 283 ± 22. 7 minutes. In group 2, it was 285 ± 25. 8 minutes. The operation time was found to be shorter in group 1, although the difference was not statistically significant (p = 0. 89) [ 23 ]. In our study, whereas the operation time was 134 ± 22 minutes for group 1, it was 152 ± 38 minutes for group 2. The difference in operation times was statistically significant (p < 0. 05). In the study by Clifton et al. , for 40 C7, 40 T6, and 40 L5 pedicle screws, the rate of agreement between the pedicle positions studied on preoperative models and the postoperative pedicle screw positions was found to be 100% for C7, 100% for T6, and 93% for L5 [ 24 ]. In our study, five (20%) of the 25 upper thoracic fracture patients (group 1) were T3, eight (32%) were T4, and 12 (48%) were T6 fractures, which were operated on by preoperative planning using 3D modeling. For these 25 patients, the concordance rate between pedicle positions studied on preoperative models and postoperative pedicle screw positions was 93. 8% for T3 fractures, 94. 7% for T4 fractures, and 98. 4% for T6 fractures. Vertebral screw misplacement and vascular injuries are common in the upper thoracic region [ 25 ]. The ability to perform preoperative surgical simulation of the 3D printing-supported model, the application of the upper thoracic pedicle screw will become more efficient and easier. In our study, the accuracy rate obtained in the 3D printing-supported group was 96. 5%, which was higher than that of the freehand technique group. We think that the 3D printing-supported method in upper thoracic pedicle screw application will shorten learning time, provide easier learning on the model, and increase pedicular screw placement accuracy. Conclusions For upper thoracic pedicle screw insertion 3D models, guide plates can be produced inexpensively and individually. It provides a new method for accurate placement of upper thoracic pedicle screws with high accuracy and comfortable use in screw insertion.

10. 7759/cureus. 16749

Cureus

Constructing an Individualized Middle Cerebral Artery Model Using 3D Printing and Hydrogel for Bypass Training

The importance and complexity of cerebral bypass surgery (CBS) highlight the necessity for intense and dedicated training. Several available training models are yet to satisfy this need. In this technical note, we share the steps to construct a digital imaging and communications in medicine (DICOM)-based middle cerebral artery (MCA) model that is anatomically accurate, resembles handling properties of living tissue, and enables trainers to observe the cerebrovascular anatomy, improve and maintain microsurgical dexterity, and train in the essential steps of CBS. The internal and external molds were created from the geometry of DICOM-based MCA using Fusion 360 software (Autodesk, San Rafael, USA). They were then three-dimension (3D) printed using a polylactic acid filament. The 15% w/v solution of polyvinyl alcohol (PVA) was prepared and injected between the molds. Using five freeze-thaw cycles the solution was converted to tissue-mimicking cryo-gel. The model was then placed in a chloroform bath until the internal mold dissolved. To evaluate the accuracy of the MCA model, selected characteristics were measured and compared with the MCA mesh. The DICOM-based MCA model was produced using 3D printing that was available in the lab and the overall cost was less than $5 per model. The external mold required six and a half hours to be 3D printed, while the internal mold only required 23 minutes. Overall, the time required to 3D print the DICOM-based MCA model was just short of seven hours. The greatest statistically significant difference between the virtual MCA model and the DICOM-based MCA model was found in the length of the pre-bifurcation part of the M1 segment and the total length of the superior bifurcation trunk of M1 and superior branch of M2. The smallest statistically significant difference was found at the diameter of the inferior post-bifurcation trunk of the M1 segment and the diameter at the origin of the artery. This technical report aims to show the construction of a CBS training system involving the DICOM-based MCA model that demonstrates the shape of the vascular tree, resembles the handling/suturing properties of living tissue, and helps set up a homemade training station. We believe that our DICOM-based MCA model can serve as a valuable resource for CBS training throughout the world due to its cost-effectiveness and straightforward construction steps. Moreover, once the DICOM-based MCA model is used with our training station, it may offer an option for trainers to gain and maintain CBS skills despite limitations on time, cost, and space. This work was presented in February 2019 at the American Association of Neurological Surgeons/Congress of Neurological Surgeons (AANS/CNS) Cerebrovascular Section Annual Meeting held in Honolulu, Hawaii.

Introduction In neurosurgery, there are many procedures that a skilled neurosurgeon must master. One such procedure is the cerebral bypass. Any brain surgeon dealing with cerebral vasculature requires precise actions while navigating the blood vessels on the exterior surface of the temporal lobe [ 1 ], microanastomosis skills, familiarity with the shape and volume of the vessels, and the ability to use instruments properly for basic vessel repairment or complex revascularization procedures [ 2 ]. A recent study has demonstrated that the number of cerebral bypasses performed for adult revascularization has increased in the United States between the years 2002 and 2014. However, its application for the treatment of occlusive vascular disease and cerebral aneurysms has decreased [ 3 ]. This procedure comes with its challenges and a neurosurgeon must be highly trained in augmentation and restoration to overcome them. Training will not only help in reducing complications but will also add confidence to the neurosurgeon as they perform a high-risk procedure [ 2 ]. The importance and complexity of the cerebral bypass highlight the necessity of training to improve and maintain microanastomosis skills. Efficient training must simulate the real operative conditions using accessible, inexpensive, reusable, and anatomically correct models that provide haptic feedback. Many resources for training outside the operating room exist, including animal models, such as turkey/chicken brachial arteries [ 4 ]. Each model has its own strengths and weakness. In the initial stages of practicing cerebrovascular bypass surgery (CBS) skills, latex gloves or silastic tubes can be useful models that demonstrate high construct validity. However, they do not provide realistic handling properties of cerebral tissue since multiple factors, such as localized vessels, will affect this procedure [ 5 ]. On the other hand, live animal models are close to real surgery conditions because of their natural blood flow, the potential for thrombosis, and similarity to human tissue. Even so, ethical regulations on living animal models render this option unfeasible. Additionally, live animal models require vivarium and animal care staff, several pharmaceuticals, and knowledge of anesthesia for animals. Turkey and chicken wing/leg could be better options since they still have tactile feedback and are not restricted by ethical regulations and institutional review board approval [ 2 ]. Human cadavers provide anatomical certainty but are susceptible to infections and require dedicated laboratories. Moreover, fixation solutions make the vessel wall tissue stiff and friable. New training models have emerged in several fields of medicine with recent advances in three-dimensional (3D) printing and virtual reality technologies. Especially in the field of neurosurgery, creating 3D models for surgical training and preoperative planning is of paramount importance due to the complexity of procedures and insufficiency of 2D images for demonstrating spatial relationships between nerves, arteries, veins, brain, and skull [ 6 ]. Even as 3D printing and virtual reality become more relevant, it is important to note that they are unable to provide physical hands-on experience that is crucial for all neurosurgeons-in-training. Recently, 3D printing evolved with the use of biological materials as the substrate. This concept, known as 3D bioprinting, has made its presence in regenerative medicine, tissue engineering, drug and cancer research, and organ transplantation. Various studies have revealed that bioprinting mimics native vascular networks for tissue engineering in the growth of thick tissue [ 7 ]. Although bioprinting has strong potential for constructing vessels, limitations associated with cost and infrequent applications do not make it a preferred method for constructing models for CBS training and preoperative planning [ 8 ]. Despite significant improvements in current technologies, constructing vessels with high accuracy using 3D printing remains difficult due to the hollow shape of vessels, their curvature, small size, and special characteristics of wall tissue. Furthermore, even if 3D printers can facilitate the construction of anatomically correct models, there are not many satisfactory printing materials that mimic the handling properties of vessel wall tissue. To make up for what 3D printing lacks, we aimed to integrate our model with polyvinyl alcohol cryogel (PVA-C). PVA-C was chosen because it has human tissue-mimicking properties and can be easily manipulated due to its high elasticity [ 9 ]. Depending on the required treatment, various types of cerebral bypass procedures are performed on different arteries. In this study, we chose to construct models of the middle cerebral artery (MCA) due to its frequent applications in CBS training [ 2 ]. We highlight the workflow for constructing MCA models through 3D printing and PVA-C to generate models that are anatomically accurate, cost-effective, easily constructible, and resemble handling properties of living tissue, and aids in the training of CBS skills. Our workflow includes generating molds from a virtual model of MCA using computer-aided design tools and 3D printing, preparing and injecting polyvinyl alcohol (PVA) solution as a tissue-mimicking material, applying freeze/thaw cycles to convert the PVA solution to PVA-C, discarding the external mold, and dissolving the internal mold using chloroform. We also developed a CBS training station using a tripod and smartphone camera to further demonstrate how our MCA models can be utilized without expensive surgical equipment and tools. Technical report Material and methods Fabrication of the Internal Mold The virtual model of MCA can be generated by extracting data from digital imaging and communications in medicine (DICOM) and images of computed tomographic angiography (CTA). We downloaded a DICOM-based virtual model of MCA from an online resource ( thingiverse. com ) and smoothed it using computer-aided design (CAD) software namely Meshmixer (Autodesk, San Rafael, USA), to decrease the volume of the model without changing its shape and angles. The resulting model, which represents the virtual model of the internal mold, was exported as a stereolithography (STL) file and uploaded to Ultimaker Cura (Ultimaker B. V, Geldermalsen, Netherlands), an open-access 3D printing software. The virtual model was then rotated until its branches were placed on top and the opposite part was placed below to print anatomically accurate branches. Brim, which adds a single layer flat area around the base of the model to discourage warping, was chosen from the build plate adhesion settings on the 3D printing software. Following that, low-density supports that prevent the overhanging parts of the model from falling were generated. These supports were removed post-printing. Polylactic acid (PLA) thermoplastic was selected as the 3D printing material due to its high solubility in chloroform that facilitates the removal of the internal mold. Finally, the virtual model of the internal mold was 3D-printed on Ultimaker 2+ (Ultimaker B. V, Geldermalsen, Netherlands) (Figure 1 ). Due to the fragile character of the internal mold, the support material that enables the printing of branches was removed gently. Figure 1 Fabricating the internal mold via 3D printing Design and Fabrication of the External Mold The DICOM-based virtual model of MCA was imported as an STL file into the CAD software Fusion 360 (Autodesk, San Rafael, USA). Create Base Feature was selected to enter direct editing mode. In this mode, the STL file appears in mesh form. This mode also allows users to create objects in another form named solid body. As the mesh form of the DICOM-based virtual model could not be used as reference geometry for the external mold, it was converted to a solid body by using the Mesh to B-Rep conversion tool and then connected with boxes by the loft tool to create a shape that transitions between two or more edges or faces. Then came the sketches. 2D geometries that use a base to construct 3D geometries were created and used as paths to add three solid pipes (an inlet runner duct and two outlet vents) that facilitate unidirectional filling of the vessel wall. A box was created according to the earlier designed parts of the external mold to be the target body. The parts that would be hollow in the completed external mold were then created. These parts were then excluded from the box that would be the body of the external mold. The workspace was switched to the patch environment to create and extrude a fit point spline along an axis of arteries. It was used as a tool plane to split the external mold for easy removal of the DICOM-based MCA model with the internal mold inside. Small boxes were added to two sides of the external mold to enable easy handling. The completed external mold was exported and converted to STL file format. It was then 3D printed via Ultimaker 2+ (Ultimaker B. V, Geldermalsen, The Netherlands) (Figure 2 ). Figure 2 Fabricating the external mold via 3D printing Preparation of PVA Solution To prepare a 15% w/v aqueous solution of PVA, 5g of PVA powder was added to 23. 5ml of deionized water. The solution was heated and stirred at 100°C for approximately one hour until it became clear. The solution was left to cool to room temperature for approximately one hour. Finally, 9. 8ml of dimethyl sulfoxide (DMSO) was added to the solution. Throughout the preparation process, the solution was covered with stretch film to prevent drying. Injection of PVA Solution The internal mold was mounted between two parts of the external mold and assembled. The constructed double-layered mold was sealed with silicone to prevent leaking during freeze/thaw cycles. The PVA solution was injected into the inlet runner duct of the mold using a 1ml syringe until the overflow at the outlet vents was seen. Overall, 3ml of PVA solution was used per model. The Freezing and Thawing Process Multiple freezing and thawing cycles were performed to convert the PVA solution into a gel. This allowed the material to form a porous network of hydrogen bonds and simulate characteristics of the artery wall. The mechanical properties of the material were improved by altering the concentration of the solution and the number of freeze-thaw cycles based on feedback received from surgical experiences. Five freeze-thaw cycles were subsequently administered to the entire external mold. Each cycle consists of two steps. Step one involves placing the mold in a −20 °C freezer for 24 hours. And step two is the thawing of the mold at 4 °C for 24 hours. Removal of the Internal Mold The two parts of the external mold were separated and the model with the internal mold was taken out. The internal mold-model complex was placed in a chloroform bath. After 24 hours, the internal mold dissolved completely without impacting the physical model (Figure 3 ). Figure 3 The model was placed in the chloroform bath until the internal mold dissolved Training Station A training station was constructed using a smartphone and a tripod. The camera application on the smartphone was used to replicate the magnified view provided by the surgical microscope, and the tripod was used to stabilize the smartphone to arrive at the desired position (Figure 4 ). Figure 4 The training station was constructed using a tripod and a smartphone Statistical Analysis We evaluated the accuracy of our protocol by comparing measurements between the DICOM-based virtual model of MCA numbering one (n=1) and the mean of the DICOM-based MCA models (n=8). The following anatomical characteristics of MCA were measured using MeshLab software (ISTI-CNR, Pisa, Italy) for the DICOM-based virtual model of MCA and a digital caliper for the DICOM-based MCA models: (1) diameter of the artery at its origin, (2) length of the pre-bifurcation part of the M1 segment, (3) diameter of the artery at the bifurcation point, (4) diameter of the inferior post-bifurcation trunk of the M1 segment, (5) total length of the inferior post-bifurcation trunk of M1 and inferior branch of M2, (6) diameter of the superior post-bifurcation trunk of the M1 segment, (7) total length of the superior post-bifurcation trunk of M1 and superior branch of M2. These measurements can be seen in Figure 5. Figure 5 Depiction of the anatomical characteristics of MCA that were measured via MeshLab software (ISTI-CNR, Pisa, Italy) for the mesh and a digital caliper for the model The anatomical characteristics were adopted from the literature [ 10 ]. The MCA consists of four segments, M1 through M4. The M1 segment begins at the origin of the MCA, where the internal carotid artery divides and gives rise to the MCA and anterior cerebral artery, and terminates at the limen of the insulae. The M2 segment begins at the limen of the insulae and terminates at the circular sulcus of the insulae. The M1 segment may bifurcate or trifurcate before forming the M2 segment. In our DICOM-based virtual model of MCA, a bifurcation was examined. We divided our DICOM-based MCA model into two parts according to the bifurcation of the M1 segment for the measurements. The mean and standard errors were reported for each measured parameter. One-sample t-tests were performed to assess any statistically significant differences between measurements of the DICOM-based virtual model of MCA and DICOM-based MCA models. All statistical analyses were performed using John's Macintosh Project (JMP), version 14 (SAS, Cary, NC, USA). Results Suturable, hollow, DICOM-based MCA model was acquired with the above-proposed method. The filament cost was less than $1 for two pieces of the external mold. Compared to the external mold that required 32g of filament and six and a half hours for 3D printing, the internal mold required only 1g of filament and 23 minutes for 3D printing. A PVA solution of 33. 3ml was prepared using 5g PVA powder, 23. 5ml deionized water, and 9. 8ml DMSO. However, only 3ml of the PVA solution was spent per DICOM-based MCA model. The average costs were $15 per 100g of PVA powder, $15 per liter of deionized water, and $20 per 236. 6ml (8oz) of DMSO. The total material cost per DICOM-based MCA model was less than $5. We relied on in-house 3D printing services provided by our institution. As for the training station, the average cost of the tripod was $2. The cost of the smartphone wasn’t included as it is a widely available personal device. With the addition of surgical instruments (needle holder, forceps, micro scissor, and 10-0 suture) the total cost of constructing our CBS training system is $129 (Table 1 ). Table 1 List of costs incurred in constructing the DICOM-based MCA model that includes the training station, materials used, and microsurgical instruments and equipment. Cost of smartphone and 3D printer were excluded as the former is a personal device and for the latter, we relied on the one provided by our institution. Software and Materials Where to Find Average Cost Amount per Model Total Cost per Training Module Training Station Smartphone Personal device 1 - Phone application DSLR Zoom Camera (Peakercorp Dev Team) Play Store or similar apps at App Store Free app $0 Tripod ebay. com $2 1 $2 The total cost of the training station $2 Surgical equipment and expendables Microneedle holder ebay. com, amazon. com $12 1 $12 Tying micro forceps ebay. com, amazon. com $15 1 $15 Microscissor (or a normal scissor) ebay. com, amazon. com $10 1 $10 10-0 suture ebay. com, amazon. com, alibaba. com $3 1 $3 Pen Stationery store The total cost of surgical equipment $40 Materials used in model preparation Polylactic acid (PLA) filament ebay. com, amazon. com, alibaba. com $25/kg 32+1g $0. 83 Polyvinyl alcohol (PVA) powder ebay. com, amazon. com, alibaba. com $15/100g <1g Deionized water (low grade) ebay. com, amazon. com, alibaba. com $15/L <3ml Dimethyl sulfoxide (DMSO) ebay. com, amazon. com, alibaba. com $20/ 8oz (236. 6ml) <1ml Silicone tube sealant ebay. com $5 little Syringe ebay. com, amazon. com, alibaba. com $1 1 Veterinary needles ebay. com, amazon. com, alibaba. com $6/pack (12/pack) 1 Others (oven, refrigerator, pat, glassware, bar, stretch film, etc. ) Found in every home - - - The total cost of the CBS training module (including training station, neurosurgical equipment, and materials required to produce DICOM-based MCA model) $129 Statistically significant differences were observed when comparing measurements of the DICOM-based virtual model of MCA to those of the mean of the DICOM-based MCA models (p<0. 002). The greatest mean differences between the DICOM-based virtual model of MCA and DICOM-based MCA models were observed in the length of the pre-bifurcation part of the M1 segment (-0. 300mm) and total length of the superior post-bifurcation trunk of M1 and superior branch of M2 (-0. 275mm). The smallest differences were observed in the diameter of the inferior post-bifurcation trunk of the M1 segment (-0. 075mm) and the diameter of the artery at the origin (-0. 100mm). Moreover, the greatest standard deviation (SD) of means was observed for the total length of the superior post-bifurcation trunk of M1 and a superior branch of M2 (0. 128). And the smallest SD of means was observed in the diameter of the inferior post-bifurcation trunk of the M1 segment (0. 071) (Table 2 ). Table 2 Anatomical characteristics of the MCA were measured on the DICOM-based virtual model and DICOM-based MCA models. Paired t-tests were performed to assess any statistically significant differences between these models. Landmark DICOM-based virtual model of MCA DICOM-based model of MCA (1) DICOM-based model of MCA (2) DICOM-based model of MCA (3) DICOM-based model of MCA (4) DICOM-based model of MCA (5) DICOM-based model of MCA (6) DICOM-based model of MCA (7) DICOM-based model of MCA (8) Mean Standard Deviation Mean Difference Diameter of the artery at the origin 3. 2mm 3. 2mm 3mm 3. 1mm 3. 2mm 3mm 3. 1mm 3mm 3. 2mm 3. 1mm 0. 092582 -0. 1 Length of the pre-bifurcation part of the M1 segment 21. 9mm 21. 7mm 21. 5mm 21. 5mm 21. 7mm 21. 5mm 21. 6mm 21. 6mm 21. 7mm 21. 6mm 0. 092582 -0. 3 Diameter of the artery at the bifurcation point 2. 5mm 2. 4mm 2. 4mm 2. 3mm 2. 4mm 2. 3mm 2. 2mm 2. 2mm 2. 4mm 2. 325mm 0. 0886405 -0. 175 Diameter of the inferior post-bifurcation trunk of the M1 segment 2. 1mm 2. 1mm 2mm 2mm 2. 1mm 2mm 1. 9mm 2mm 2. 1mm 2. 025mm 0. 0707107 -0. 075 Total length of the inferior post-bifurcation trunk of M1 and inferior branch of M2 12. 6mm 12. 5mm 12. 4mm 12. 5mm 12. 6mm 12. 3mm 12. 4mm 12. 4mm 12. 5mm 12. 45mm 0. 092582 -0. 15 Diameter of the superior post-bifurcation trunk of the M1 segment 1. 9mm 1. 8mm 1. 8mm 1. 7mm 1. 8mm 1. 7mm 1. 6mm 1. 7mm 1. 8mm 1. 7375mm 0. 0744024 -0. 163 Ttotal length of the superior post-bifurcation trunk of M1 and superior branch of M2 16. 8mm 16. 6mm 16. 4mm 16. 5mm 16. 6mm 16. 3mm 16. 5mm 16. 6mm 16. 7mm 16. 525mm 0. 128174 -0. 275 Once the DICOM-based MCA models were compared to the DICOM-based virtual model of MCA one-by-one, the greatest p-value was observed for the DICOM-based MCA model (4) (p=0. 0453), whereas the smallest p values were observed for DICOM-based MCA model (6) and (7) (p=0. 0002). Discussion Cerebrovascular bypass surgery remains a complex procedure in neurosurgery. With the help of cutting-edge technology, the necessary training to maintain quality CBS skills is now easily within our reach. In this study, we constructed a DICOM-based MCA model that precisely simulates the shape of the vascular tree, mimics handling properties of living tissue, enables trainers to observe the cerebrovascular anatomy, improves and maintains microsurgical dexterity, and helps train in the essential steps of CBS. Our protocol involved creating models based on DICOM data, 3D printing, and casting. 3D printing has a wide range of applications, including implantable device and surgical tool fabrication, pharmaceutical product design, organ printing, model construction for training, and preoperative planning among many others in medicine [ 11 ]. These broad-spectrum applications are capable of providing opportunities to acquire individually designed, or DICOM-based products with various options. These options are also available in both cost-effective and time-effective ways. We benefitted from 3D printing in acquiring our external and internal molds. Molding techniques have been preferred in various studies ranging from vascular surgery [ 12 ] to neurosurgery [ 13 ] due to the need to produce soft models that have living tissue characteristics. These models enable trainers to practice surgical procedures involving clipping, cutting, suturing, stapling, and energy-based device use. Furthermore, it leads to the development of new imaging techniques and surgical approaches. In neurosurgery especially, these techniques are crucial for molding a highly skilled neurosurgeon. Different techniques and materials used for casting are selected according to the tissue and procedure involved. The material characteristic and usefulness of PVA polymer as a tissue-mimicking material has been described previously [ 14, 15 ]. We also chose to use PVA as the tissue-mimicking material. Acquiring a hollow model is essential for vascular training or research. In the literature, many methods based on 3D printing described how to construct hollow models that replicate the complex shape of the vascular tree [ 12, 16 ]. Those methods differ from each other in various ways such as choice of casting material, the number of layers on the mold, and internal mold (template) removing approach to name a few. Chee et al. established the construction of walled carotid phantom using 3D printed molds and PVA cryo-gel for vascular imaging investigations. In their study, a two-step process consists of snapping the core and retrieving the fragments to obtain the vessel model with a hollow lumen. Even if this method succeeded for their model, we found it inappropriate for our study as we aimed to construct the DICOM-based MCA model, which has complex geometry and is small compared to the CAD-based carotid model. We also preferred to use a solvent for removing the internal mold. The chloroform enables us to obtain a hollow lumen while providing minimal harm to the DICOM-based MCA model. However, our results demonstrated that there were statistically significant differences when comparing measurements of the DICOM-based virtual model of MCA to those of the mean of the DICOM-based MCA models. According to our measurements and statistical analysis, we believe that the DICOM-based MCA model may be an efficient option for resident and medical student education, although it may be inappropriate for preoperative training. Even if models are of paramount importance in training, the magnified view provided by a neurosurgical microscope is required for CBS training. However, conventional neurosurgical microscopes are expensive, immobile, limited in number, and not available at all times. Several systems that replicate the magnified view provided by a neurosurgical microscope were previously described to enable trainers to practice practically anywhere without spending a lot of money [ 17 ]. Moreover, the cost of the models and required systems/equipment are also important in training [ 18 ]. Our objective in this technical report was two-fold. One, to construct a vessel model that demonstrates the shape of the vascular tree and resembles the handling/suturing properties of living tissue. And two, to construct that model affordably and suggest a cost-effective CBS training system to be used in conjunction with the training station. Taken together, the material cost for an MCA model is less than $5. This price is substantially cheaper than the price of other effective models, and our model is far easier to replicate when compared to the one made by the University of Washington [ 19 ]. While the internal mold is dissolved in chloroform to acquire a hollow model, the external mold, however, can be used repeatedly. Due to the cost-effective and straightforward construction steps, we believe our MCA model can be widely used by residents and neurosurgeons throughout the world. Moreover, other models could be produced following our construction step. In other words, with minor changes, any disease morphology can be added to the model, or the vessel wall tissue characteristic can be altered according to age [ 20 ], patient, or disease. This would allow the model to become a valuable resource for a wide range of case-specific training. Moreover, the model can be integrated with current simulation systems to produce more holistic training. Study limitations Limitations were present in assessing the material characteristic of our MCA model in an objective way. We didn’t perform any mechanical tests such as stiffness test, uniaxial tensile strength test, or needle insertion deformation test to demonstrate the accuracy of the vessel wall tissue characteristics compared to real MCA. Also, we didn't get feedback from practicing cerebrovascular surgeons or trainers about the model's ability to replicate human cerebral vasculature. Another limitation was the lack of hemodynamic flow and thrombosis. However, such flow can be provided via basic pump-based infusion. And to compare with other training models, further studies measuring improvement in CBS skills are required. Moreover, our training system using the smartphone camera was limited to two dimensions, which is contrary to neurosurgical microscopes that provide three-dimensional viewing angles. Furthermore, measurements of the DICOM-based virtual model of MCA (n=1) were compared with the means of measurements of DICOM-based MCA models (n=8). Conclusions In this technical note, we aimed to report how to construct a CBS training system involving the DICOM-based MCA model that demonstrates the shape of the vascular tree and resembles the handling/suturing properties of living tissue along with the training station. We believe that our DICOM-based MCA model may revolutionize current training methods by providing a realistic and widely available option for CBS training throughout the world with its desirable features of cost-effectiveness, and straightforward construction steps. Moreover, once the DICOM-based MCA model is used with our training station, it may offer an option for trainers to gain and maintain CBS skills without limitations of time, cost, and space.

10. 7759/cureus. 17022

Cureus

Stem Cell Therapy for the Treatment of Myocardial Infarction: How Far Are We Now?

Myocardial infarction is one of the leading causes of death worldwide. Poor functional recovery of the myocardium is noticed after an event of myocardial infarction. Researchers and clinicians around the world have been engaged to regenerate the damaged human heart for a long time. Stem cell therapy is an exciting newer therapy to treat cardiovascular diseases. Various types of stem cells have been used to revive the damaged myocardium after myocardial infarction, and they have overall demonstrated safety and moderate efficacy. The specific mechanisms by which these cells help in improving cardiac function are still not completely known. There is growing evidence that intracoronary bone marrow cell transplantation in patients with myocardial infarction beneficially affects the remodeling of the damaged myocardium. Our systematic review article aims to assess the effects and the future of stem cell therapy in patients with myocardial Infarction. We searched articles in PubMed, ScienceDirect, and Google Scholar. Thirty-one studies that included 2171 patients in total were analyzed. Most of these studies showed stem cell therapy is safe and well tolerated in patients, and modest improvements are seen in left ventricular functions with no major adverse effects. However, some studies showed no positive and clinically significant outcomes. So, more high-quality studies on a larger scale are required to support and confirm its efficacy in remodeling damaged myocardium after myocardial infarction. We should also perform studies to determine the timing of cell delivery that is best suited for stem cell therapy.

Introduction and background One of the most common causes of morbidity and mortality globally is cardiovascular disease [ 1 ]. The one-year mortality is approximately 13% and the five-year prognosis for patients with heart failure is 50%, even though there has been tremendous advancement in the treatment of acute myocardial infarction (MI) [ 1 ]. The presence of any obstruction in the coronary arteries gives rise to acute myocardial ischemia [ 1 ]. Rupture of plaques, fissuring, or formation of any superimposed thrombus may be responsible for this obstruction formation [ 1 ]. Although there have been major advancements in the management of acute Myocardial Infarction including fibrinolysis and rapid revascularization, the prognosis remains poor due to the lack of self-repairing of the already damaged myocardium, which may result in complications like heart failure [ 1 ]. There are multiple methods to repair the damaged heart that include cell transplantation, gene therapy, stimulating innate repair pathways, direct reprogramming of cells, cardiac tissue engineering, and biomaterial delivery [ 2 ]. Among these, the most accepted strategy for heart repair is the delivery of exogenous cells [ 2 ]. Almost every cell type we can think of, such as skeletal myoblasts to pluripotent stem cells and their derivatives has been transplanted into the injured myocardium [ 2 ]. Stem cells are unspecialized immature cells that can divide and replicate themselves throughout the entire life of an organism [ 3 ]. Skeletal myoblasts (satellite cells) are classically the stem cell population within the non-cardiac musculature [ 4 ]. There are 2% to 7% improvements in ejection fractions (EF) with the administration of adult bone marrow cells (BMC) [ 4 ]. The exact mechanisms of improvement of damaged heart function by cell therapy are unclear, but it is assumed that the paracrine effect plays a central role [ 5 ]. Transplanted mesenchymal stem cells (MSCs) can engraft and differentiate into cardiomyocyte-like and endothelial cells and recruit endogenous cardiac stem cells [ 6 ]. As the viability and function of autologous adult stem cells decline with age, especially in patients with MI, alternative sources of stem cells such as Wharton’s jelly-derived mesenchymal stem cells (WJ-MSCs), cardiac progenitor cells can also be used [ 6 ]. Isolating and expanding resident cardiac progenitor cells present in the adult myocardium cells is a tough task. However, these are more beneficial than the other stem cell types because they are likely predestined to cardiovascular fate [ 5 ]. The first-ever encouraging study showing positive outcomes in MI patients with stem cell therapy was published by Strauer et al. in 2002, many other trials have been conducted since then [ 7 ]. The main objective of our article is to evaluate the safety and effects of transplanting stem cells in patients with acute myocardial infarction. Figure 1 given below illustrates the pathophysiology of MI. Figure 1 Pathophysiology of myocardial infarction Review Methods Study Protocol We implemented Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) 2020 Guidelines in our study process. Sources of Data and Search Strategy Articles were searched from three databases PubMed, Science Direct, and Google Scholar using specific keywords related to the research topic. The keywords used are as follows: myocardial infarction, ST-elevation myocardial infarction, non-ST-elevation myocardial infarction, stem cell transplantation, mononuclear bone marrow cell transplantation, adult population. The Medical Subject Headings (MeSH) is the National Library of Medicine (NLM) controlled vocabulary thesaurus that we specifically used to search articles in the PubMed database. We performed a search using a combination of MeSH terms and text words given below. The final MeSH search strategy used was as follows: Myocardial Infarction OR ST-Elevation Myocardial Infarction OR Non-ST-Elevation Myocardial Infarction OR Acute Myocardial Infarction OR ( "Myocardial Infarction/drug therapy"[Mesh] OR "Myocardial Infarction/mortality"[Mesh] OR "Myocardial Infarction/prevention and control"[Mesh] OR "Myocardial Infarction/therapy"[Mesh] ) AND Stem Cell Transplantation OR Stem cell therapy OR mesenchymal stem cell OR Progenitor Cell OR mononuclear bone marrow cell transplantation OR ( "Stem Cell Transplantation/instrumentation"[Mesh] OR "Stem Cell Transplantation/methods"[Mesh] OR "Stem Cell Transplantation/mortality"[Mesh] OR "Stem Cell Transplantation/therapeutic use"[Mesh] OR "Stem Cell Transplantation/therapy"[Mesh] )AND Adult. Other databases and the keywords used for the search are mentioned in Table 1. Table 1 Databases and search results Databases Keywords used for the search Search results Initial results Timeframe 2011-2021 PubMed The final MeSH search strategy as mentioned above 107, 272 57, 904 ScienceDirect Myocardial Infarction and Stem cell therapy 22, 211 12, 640 Google Scholar "Myocardial infarction" and "stem cell therapy" and "mononuclear bone marrow cell transplantation" 1, 300 547 Inclusion and Exclusion Criteria We only included the articles published in the English language, which were human studies and clinical trials. We selected articles published from 2011-2021. The inclusion criteria were: (1) Patients diagnosed with myocardial infarction; (2) Patients who received stem cell therapy after myocardial infarction; (3) Age of the patients 19 and above; (4) Both male and female patients were selected. Exclusion criteria were studies on animals, reviews, or studies for which the full text was unavailable or only abstracts were available. We did not include gray literature. Risk and Quality Assessment Two reviewers independently (RAB and RB) extracted and evaluated the quality of the included 31 studies. Revised Cochrane’s risk of bias assessment tool was used for randomized controlled trials (RCTs) and clinical trials. Data Extraction Two reviewers (RAB and RB) separately extracted relevant data from included 31 studies using standard data extraction forms and data was extracted under the following headings: name of the author, country of the study, the name of the journal where it was published, year of publication, study design, the title of the study, sample size, patient characteristics, size of the treatment group and control group, follow-up period, and outcome of the study. Results A total of 2650 articles from PubMed, 221 articles from ScienceDirect, and 547 articles from Google Scholar were collected using the search strategy we have mentioned in the method section and were then screened based on the title and abstract related to our study. We also removed the duplicates. Then, we filtered out a few papers based on the eligibility criteria and availability of full text. In the end, only 31 items were included, and these articles were checked for quality based on their study characteristics. A complete Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) flow diagram is given below in Figure 2. Figure 2 PRISMA flow diagram PRISMA: Preferred Reporting Items for Systematic Review and Meta-Analyses Study Characteristics Our systematic review includes patients from 31 studies. Table 2 shows the characteristics of the included studies and the outcomes of the studies. Table 2 Characteristics and outcomes of the included studies STEMI = ST-Elevation Myocardial Infarction, CAG = Coronary Angiography, PCI = Percutaneous Coronary Intervention, PPCI = Primary Percutaneous Coronary Intervention, LVEF = Left Ventricular Ejection Fraction, CABG = Coronary artery bypass grafting, CK = Creatine Kinase, H/o = History of, MI = Myocardial Infarction, CCT = Controlled Clinical Trial, CT = Clinical Trial, MNC = Mononuclear stem cells, CCTA = Coronary Computed Tomography Angiography, RCT = Randomized Controlled Trial, BMMNC = Bone Marrow Mononuclear Cell, BNP = B-type natriuretic peptide, WMSI = Wall Motion Score Index, N-BMC = Normoxia Bone Marrow Mononuclear Cells, HP-BMC = Hypoxia Preconditioned Bone Marrow Mononuclear Cells, ECG = Electrocardiogram, SPECT = Single-Photon Emission Computed Tomography, LV = Left Ventricle, MRI = Magnetic Resonance Imaging, CT scan = Computed Tomography scan, F-18-FDG-PET = F-18-Fluorodeoxyglucose Positron Emission Computed Tomography, 99mTc-SPECT= 99mTc-sestamibi Single-Photon Emission Computed Tomography, G-CSF = Granulocyte Colony-Stimulating Factor, MSCs = Mesenchymal Stem Cells, WJ-MSC = Wharton’s jelly-derived Mesenchymal Stem Cells, CDC = Cardio sphere-derived autologous stem cell, LVEDD = End-Diastolic Dimension of the Left Ventricle, LVESV = End-Systolic Volume of the Left Ventricle, LVEDV = End-Diastolic Volume of the Left Ventricle, SPCs = Stem/ Progenitor Cells, LIN- = Lineage-negative, CADUCEUS = CArdiosphere-Derived aUtologous stem CElls to reverse ventricUlar dySfunction, TIME trial = Timing in Myocardial Infarction Evaluation trial Author Country/Year Study design Patient characteristics Sample size Treatment group Control group Follow-up period Outcome Benedek et al. [ 8 ] Romania / 2014 CCT Adult with h/o STEMI and PPCI and abnormalities in wall motion and less than 50% stenosis 18 Autologous MNC = 9 Placebo = 9 Four years with clinical examinations, ECG, Echocardiography, 64-slice CCTA A small improvement in EF and the plaque burden is lower in coronary segments treated with stem cells Alestalo et al. [ 9 ] Finland / 2015 Double-blinded RCT H/o STEMI, < 75 years, hemodynamically stable and no cardiogenic shock or rescue PCI/CABG 26 BMMNC = 14 Placebo = 12 Cytokines after four days and LV angiogram after six months A balancing effect between the anti-inflammatory and proinflammatory cytokine BMMNC group at day four Bozdag-Turan et al. [ 10 ] Germany / 2012 CCT 18–80 years with h/o STEMI, absence of co-morbidities, cancer, and active bleeding or trauma in the last two months 24 BMC = 12 Placebo = 12 Six months with left ventriculography Infarct size and BNP level decreased, and global EF and infarct wall movement velocity were increased in the stem cell group Choudry et al. [ 11 ] 5 centers in Europe [United Kingdom (3); Switzerland (1); Denmark (1)] / 2016 Double-blinded RCT Acute anterior MI with anterior wall motion abnormality and h/o PPCI 100 BMMNC = 55 Placebo = 45 One year with cardiac MRI and Echocardiography Small non-significant improvement in LVEF Duan et al. [ 12 ] China / 2015 RCT H/o MI, < 75 years with planned CABG for triple-vessel disease, LVEF < 30%, and no aneurysm or valvular diseases 42 CABG + BMMNC = 24 Only CABG = 18 One year with Echocardiography Improvement in left ventricular functions in the treatment group Gao et al. [ 6 ] China / 2015 Double-blinded RCT 18-80 years with a h/o STEMI and reperfusion with stent implantation and LV local wall-motion abnormality. CK > three-fold the upper limit of the normal 116 WJ-MSC = 58 Placebo = 58 18 months with F-18-FDG-PET and 99mTc-SPECT and two-dimensional Echocardiography LVEF significantly increased and LVESV and LVEDV greatly decreased in the treatment group Hu et al. [ 13 ] China / 2015 RCT 18-75 years old with acute STEMI and PPCI with stent implantation or thrombolysis and LV local wall motion abnormality 36 N-BMCs = 11, HP-BMCs = 11 Standard therapy = 14 Six and 12 months Improvement in changes of LVEDV, LVESV, and WMSI in the HP-BMC group. Myocardial perfusion defect ratio was reduced in HP-BMCs and N-BMC groups at six months Huan et al. [ 14 ] China / 2015 CT 18-75 years old with a h/o acute STEMI and treatment with PCI, LVEF < 50% 104 Group A = BMMNC within two hours after PCI = 27, Group B = three-seven days after PCI = 26, Group C = seven to 30 days after PCI = 26 Placebo = 25 patients Six months with angiography. SPECT and Echocardiography at six and 12 months Effects of cell therapy given within 24 hours are the same as to given three-seven days after PPCI Kim et al. [ 15 ] South Korea / 2018 RCT STEMI 26 BMMNC = 14 Placebo = 12 Four and 12-months with SPECT Increase in the LVEF from baseline to the fourth month and twelfth month in the bone marrow mesenchymal stem cells group. Lee et al. [ 16 ] Korea / 2014 CT 18-70 years old with STEMI 69 MSC = 33 Placebo = 36 Six months with SPECT Safe with modest improvement in LVEF Makkar et al. [ 17 ] USA / 2012 CT H/o two to four weeks of MI and LVEF = 25–45% 25 Cardio sphere-derived autologous stem cells = 17 Standard therapy = 8 Six months with MRI Scar mass was reduced, increase in viable heart mass and regional contractility in the CDC group. LVEDV, LVESV, and LVEF were the same in the two groups Malliaras et al. [ 18 ] USA / 2014 RCT Patients of CADUCEUS trial were followed up for a year 25 Cardio sphere-derived autologous stem cells = 17 Standard therapy = 8 One year with MRI Scar size reduced, increased viable myocardium, and improved regional function of infarcted myocardium Micheu et al. [ 19 ] Romania / 2015 CT 18-81 years old with STEMI & h/o angioplasty with stent implantation, LVEF < 40%. 18 Autologous BMCs = 7 Standard therapy = 11 Six months with clinical examination, Echocardiography, 24 hours ECG Safe and LVEF was increased Moccetti et al. [ 20 ] Switzerland / 2012 CT Acute anterior STEMI treated by PPCI and LVEF < 50%. 60 Autologous BMMNC = 23 Standard therapy = 37 Five years with Echocardiography Safe and LV function improved Moreira et al. [ 21 ] Brazil / 2011 RCT 18-80 years old with h/o MI and reperfusion and involving more than 10% of the LV 30 BMMNC via anterograde intra-arterial coronary (IAC) delivery = 14, BMMNC via retrograde intravenous coronary (IVC) delivery = 10 Placebo = 6 Cardiac MRI was performed before cell injection The retrograde approach to deliver stem cells was safe and cell retention by cardiac tissue is more in the anterograde approach Nair et al. [ 22 ] India / 2015 RCT Anterior MI and LVEF = 20-50%, 20-65 years with h/o CAG between one to three weeks 250 Stem cell therapy + standard care = 125 Standard care = 125 Six months with Echocardiography Safe, but not clinically significant Naseri et al. [ 23 ] Iran / 2018 RCT 18-75 years old with a h/o acute MI infarction, eligible for elective CABG 77 CD133 (+) = 21, MNC = 30 Placebo = 26 Six and 18 months after CABG with SPECT Significant differences were seen between the MNC and placebo groups in LVEF and a decrease in the LV thickening Nicolau et al. [ 24 ] Brazil / 2018 RCT 30-80 years, LVEF ≤ 50%, and regional dysfunction in the infarct-related area 120 BMMNC = 66 Placebo = 55 Six months with MRI No significant effects Peregud-Pogorzelska et al. [ 25 ] Poland / 2020 CCT not randomized <65 years old with first MI and EF ≤ 45% 34 Standard therapy + autologous BM-derived LIN- SPCs = 15 Only standard therapy = 19 One, three, six months, and one year with Echocardiography Safe and > 10% improvement in LVEF is noticed at 12 months Quyyumi et al. [ 26 ] USA / 2017 RCT STEMI with a stent and LVEF ≤ 48% and ≥ four days post stent 161 Intracoronary infusion of autologous CD34 (+) cell = 78 Placebo = 83 Six months with SPECT Safe Rodrigo et al. [ 27 ] Netherlands / 2013 CCT First acute STEMI treated with PPCI and maximum CK level was > 1, 600 U/L 54 MSC = 9 Standard therapy = 45 matched but nonrandomized patients Three, six months, one year, four-five years with Echocardiography, Holter, and clinical examination Improvements in LV function but not significantly different when compared to controls Roncalli et al. [ 28 ] France / 2011 RCT Acute MI and successful reperfusion with LVEF ≤ 45%, age 18–75 years 101 BMMNC = 52 Placebo = 49 Three months with MRI, Echocardiography, and SPECT Multivariate analysis shows improvement of myocardial viability than univariate analysis San Roman et al. [ 29 ] Spain / 2015 RCT Adult, acute MI with PPCI or post-fibrinolysis PCI and rapamycin drug-eluting stent implantation 120 BMMNC = 30, GCSF = 30, G-CSF + cells = 29 Standard therapy = 31 12 months with cardiac MRI Not many differences among the four groups Shah et al. [ 3 ] India / 2014 CT 30-70 years old, acute MI with PCI 19 Autologous BMCS = 12 Standard therapy = 7 24 months with Echocardiography, ECG, Holter monitoring Increase in LVEF with LV function improvements in stem cell group Srimahachota et al. [ 30 ] Thailand / 2011 RCT H/o STEMI with LVEF < 50% and PCI 23 Autologous BMCs = 11 Standard therapy = 12 Six months with cardiac MRI Symptoms improved than baseline, but not many significant changes were noticed in the two groups Sürder et al. [ 31 ] Switzerland / 2013 RCT Acute MI 200 BMMNC five-seven days after STEMI = 66, BMMNC three to four weeks after STEMI = 67 Standard therapy = 67 Four months with cardiac MRI No significant improvements Traverse et al. [ 32 ] USA / 2012 RCT MI and PCI, LVEF < 45% 120 BMMNC at day three or day seven randomly = 79 Placebo = 41 Six months with cardiac MRI No significant improvement Traverse et al. [ 33 ] USA / 2011 RCT Acute MI and PCI, LVEF < 45% 87 BMMNC after two to three weeks of MI = 58 Placebo = 29 Six months with cardiac MRI No significant improvement Traverse et al. [ 34 ] USA / 2018 RCT Patients of TIME trial, acute MI and PCI, LVEF < 45% 120 BMMNC at day three or day seven randomly = 79. 58 patients were followed up Placebo = 41, 27 patients were followed up Two years with cardiac MRI No significant improvement Turan et al. [ 35 ] Germany / 2011 RCT 18-80 years old with MI and LV dysfunction 56 BMMNC = 38 Placebo = 18 Three, six months, and one year with left ventriculography Decrease in infarct size but an increase of global EF and infarct wall movement velocity in stem cell group Yerebakan et al. [ 36 ] Germany / 2011 RCT MI at least 14 days before admission and LV akinesia with an indication for CABG 55 Intramyocardial CD133 (+) BMCs + CABG = 35 Only CABG = 20 18 months with 24-hour Holter monitoring, echocardiography, MRI, and CT scan Intramyocardial stem cell therapy was tolerable but did not have significant improvements Discussion After an acute myocardial infarction, patients usually suffer from left ventricular remodeling even after having successful revascularization [ 8 ]. Remodeling of the heart means changes in the size, shape, structure, and function of the cardiac muscles [ 37 ]. Our systematic review had 31 studies and 2171 patients. We observed the effectiveness of stem cells in an injured heart muscle after myocardial infarction. Effect of Stem Cell Therapy on Heart Function After Myocardial Infarction Benedek et al. conducted a controlled clinical trial (CCT). They included 18 patients in this trial, out of which nine patients received autologous bone marrow-derived mononuclear cells (BMMNC) [ 8 ]. On follow-up after four years, this study showed a slight improvement in ejection fraction (EF) in the stem cell group, the number of coronary plaques in segments infused with stem cell vs placebo group was ten vs twenty-one, calcium scoring in stem cell group vs placebo group was 295 vs 796, plaques creating > 50% stenosis in stem cell group vs placebo group were two vs eight and the plaque burden was much lower in coronary segments treated with stem cells [ 8 ]. Bozdag-Turan et al. conducted a prospective nonrandomized CCT in 24 patients in which he noticed a reduction in infarct size (p < 0. 001), an increase in global EF (p = 0. 003), and an increase in infarct wall movement velocity. Additionally, B-type natriuretic peptide (BNP) level also decreased in the stem cell group (p < 0. 001) [ 10 ]. The clinical trial (CT) performed in 19 patients by Shah et al. demonstrated 12 patients who had received stem cell therapy, their echocardiography showed an increase in left ventricular ejection fraction (LVEF) from baseline at six months (3. 8%) which was sustained at two years (1. 63% increase), whereas in the control group LVEF was initially increased by 1. 5% but at follow-up, in two years LVEF was decreased by 7. 3% compared to the baseline [ 3 ]. Kim et al. concluded in their randomized clinical trial (RCT) of 26 patients that there was some improvement in LVEF [ 15 ]. Turan et al. in their RCT of 56 patients described there was a decrease in infarct size but an increase of global EF and infarct wall movement velocity in the stem cell group compared to the control group [ 35 ]. Similarly, Naseri et al. in their RCT of 77 patients noticed significant differences between the stem cell groups and placebo groups in LVEF and a decrease in the left ventricular (LV) thickening [ 23 ]. Gao et al. had 58 patients in their treatment group who had received 6 × 10 6 Wharton's Jelly-derived mesenchymal stem cells (WJ-MSC) dispersed in 10 mL heparinized saline and 58 patients on the control arm who received placebo [ 6 ]. 18 months later, follow-up revealed LVEF in the WJ-MSC group significantly increased in comparison to the placebo group. Also, left ventricular end-systolic and end-diastolic volumes were greatly decreased in the WJ-MSC group [ 6 ]. In the non-randomized CCT of Peregud-Pogorzelska et al. in 34 patients, they found that stem cell therapy is safe and 60% of patients from the bone marrow-derived lineage negative (LIN-) stem/progenitor cell group showed about > 10% improvement in LVEF after a year with no signs of unfavorable remodeling of the left ventricle (LV) [ 25 ]. Similarly, one RCT conducted by Quyyumi et al. on 161 patients, among which 78 patients received an intracoronary infusion of autologous CD34 (+) cell (CLBS10) (cell therapy 10, Caladrius Biosciences Inc, Basking Ridge, NJ) revealed that stem cell therapy was safe and at one year, 3. 6% and 0% deaths were observed in the control and treatment group, respectively [ 26 ]. Additionally, the clinical trials conducted by Lee et al. and Micheu et al. found stem cell therapy is safe [ 16, 19 ]. Hu et al. in their RCT included 36 patients out of which 22 patients in the treatment arm either received normoxia-bone marrow cells (N-BMCs) or hypoxia-preconditioned bone marrow cells (HP-BMCs) and 14 patients received standard therapy [ 13 ]. There was an improvement in changes of left ventricular end-diastolic volume (LVEDV) and left ventricular end-systolic volume (LVESV) in HP-BMC group than N-BMC or control group (P < 0. 05), wall motion score index (WMSI) got better in HP-BMCs and N-BMC group (P<0. 050), but not in the control group [ 13 ]. Additionally, the myocardial perfusion defect ratio was reduced in HP-BMCs and N-BMC groups at six months compared with baseline [ 13 ]. Makkar et al. conducted a CT in 25 patients, among which 17 patients in the control group were given cardio sphere-derived autologous stem cells; they showed at follow-up after six months [ 17 ] and one year [ 18 ] that the scar size was reduced, myocardial viability was increased along with the improved regional function of the damaged myocardium. Roncalli et al. in their RCT included 52 patients receiving BMMNC and 49 patients receiving placebo [ 28 ]. Myocardial viability improved in 16/47 (34%) patients in the treatment arm compared to 7/43 (16%) in the control group (P = 0. 06) and the number of non-viable segments becoming viable was 1. 2 ± 1. 5 in the BMMNC group and 0. 8 ± 1. 1 in the control group (P = 0. 13) [ 28 ]. At three months follow-up, the multivariate analysis showed improvement of myocardial viability than the univariate analysis (P = 0. 03) [ 28 ]. It also revealed that active smoking has a significant adverse effect (P = 0. 04), and a positive trend for microvascular obstruction (P = 0. 07) was observed as well [ 28 ]. Meanwhile, the double-blinded RCT of Alestalo et al. with 26 patients (14 receiving BMMNC and 12 receiving placebo) observed a harmonizing effect between the anti-inflammatory and pro-inflammatory cytokines in BMMNC treated ST-elevation myocardial infarction (STEMI) patients on day four [ 9 ]. The inflammation process of myocardial infarction (MI) was affected by this balancing effect, and it helped in remodeling and repair of the damaged heart muscles after an episode of acute MI [ 9 ]. In contrast, the double-blinded RCT of Choudry et al. in 100 patients showed although LVEF was increased compared with the baseline in both treatment and control groups, there was not much difference between the two groups (2. 2%; 95% confidence interval, CI: −0. 5 to 5. 0; P = 0. 10) at one-year [ 11 ]. The RCT conducted by Nair et al. in 250 patients revealed, even though it is safe, stem cell therapy has no benefit in STEMI [ 22 ]. The number of patients in this study receiving the stem cell therapy deviated from 125 to 71 patients and the follow-up period was relatively short, which might have affected the outcome [ 22 ]. Similarly, Nicolau et al. in their RCT of 120 patients found intracoronary infusion of autologous bone marrow-derived mononuclear cells (BMMC) to STEMI patients did not improve LV function or decrease scar size [ 24 ]. This study did not have a core cell-processing laboratory, there was an unbalanced enrollment by the centers, and they used LVEF as the endpoint which may not be the most suitable endpoint to investigate the effect of cell infusion due to its constant changes in the acute phase [ 24 ]. These all factors had influenced the result. San Roman et al. in their RCT divided their study population into four groups which include one group of 30 patients receiving bone marrow mononuclear cells, 30 patients assigned to granulocyte colony-stimulating factor (G-CSF), 29 patients receiving G-CSF + cells, and a placebo group of 31 patients receiving standard therapy [ 29 ]. Patients treated with any of these stem cell approaches experienced similar changes in LVEF and LVESV when compared to the control group, with a small but significant reduction in infarct area (p = 0. 038) [ 29 ]. One year later, cardiac magnetic resonance imaging (MRI) did not show much difference in these four groups [ 29 ]. But it was an open-labeled study and the study population was small which may have had an impact on the result of the study [ 29 ]. Srimahachota et al. concluded in their RCT in 23 patients that, stem cell therapy is safe but no improvement in LVEF can be described from the study [ 30 ]. The authors described a few reasons for not having a positive outcome such as the BMMNC cannot maintain at the infarcted area and a very few BMMNC remained at heart [ 30 ]. Also, cytokines may be needed to integrate the stem cells in the affected heart area to initiate the cells to trans-differentiate to cardiac myocyte, and the best cell type and timing for stem cell infusion is not yet known [ 30 ]. Similarly, Sürder et al. in their RCT of 200 patients [ 31 ] and Traverse et al. in their RCT of 120 patients explained that they did not notice any improvement in LV function in the stem cell group [ 32, 34 ]. Traverse et al. demonstrated the use of cardiac MRI led to greater dropout of patients during the follow-up period because of device implantation in patients with more severe LV dysfunction which negatively impacted the outcome of their study [ 34 ]. Also, two CTs by Huang et al. and Rodrigo et al. showed there is no significant improvement in left ventricular function clinically [ 14, 27 ]. The study of Rodrigo et al. was underpowered as they had a small number of patients and they used echocardiography rather than magnetic resonance imaging (MRI) to observe the effects on LV [ 27 ]. Even though single-photon emission computed tomography (SPECT) imaging showed an improvement in myocardial perfusion after three months of stem cell treatment, since SPECT imaging was not repeated in the control group, the effect of bone marrow-derived mesenchymal stem cells (MSC) therapy on myocardial perfusion can not be evaluated [ 27 ]. Two other RCTs focused on patients who received stem cells with coronary artery bypass grafting (CABG) [ 12, 36 ]. Duan et al. had a treatment group of 24 patients having CABG + BMMNC and a control group of 18 patients who only underwent CABG [ 12 ]. One year post-surgery follow-up with echocardiography showed significant improvement in LV function including improvements in the end-diastolic dimension of the left ventricle (LVEDD), end-systolic dimension of the left ventricle (LVESD), LVEDV indexed to body surface area (LVEDVI), LVESV indexed to body surface area (LVESVI), the mass of left ventricle (LV-mass) and LV-mass indexed to body surface area (LV-mass I) compared to the data collected before the operation in CABG+BMMNC group [ 12 ]. Similarly, Yerebakan et al. had 35 patients who received intramyocardial CD133 (+) bone marrow stem cell transplant + CABG, and 20 patients who only had CABG [ 36 ]. Follow-up after 18 months post-surgery showed intramyocardial stem cell therapy was well tolerated but did not have many significant improvements [ 36 ]. The authors mentioned that no follow-up angiography was performed, there was an unplanned withdrawal of patients which resulted in incomplete follow-up testing, and a limited number of patients were available for the final analysis [ 36 ]. In addition, MRI was not available in the preoperative assessment, so the results of this study should be accepted with caution [ 36 ]. Most of the mentioned studies agreed that there was a significant improvement in myocardial function, mainly the left ventricular end-diastolic volume and left ventricular end-systolic volume as well as ejection fraction after treatment with stem cell therapy. In addition, they showed no evidence of adverse effects in patients after receiving stem cell therapy. However, some of the included studies reported there was no improvement or benefit from stem cell therapy in myocardial infarction treatment, even though stem cell therapy was safe and well-tolerable to those MI patients. The reason behind it is the studies were not done in a larger population and many patients were lost during follow-up. In addition to that, the optimum time for the administration of stem cells is not yet established. So, from our systematic review, we can conclude that we need to perform more trials in a larger population and follow up with them closely to find out the effectiveness of stem cell therapy in patients with myocardial infarction. Types of Stem Cells Used One randomized clinical trial (RCT) performed by Gao et al. used Wharton's Jelly-derived mesenchymal stem cells (WJ-MSC) [ 6 ]. WJ-MSCs display more cardiovascular differentiation potential and as they are immune privileged, they can be transplanted into unrelated recipients [ 6 ]. Another two studies used cardio sphere-derived autologous stem cells [ 17, 18 ], whereas Naseri et al. in their RCT used CD133 (+) and mononuclear cells [ 23 ]. In the study of Rodrigo et al. bone marrow-derived mesenchymal stem cells (MSC), a subpopulation of bone marrow cells was used which can differentiate into several cell types including vascular cells, functional cardiomyocytes, etc [ 27 ]. In some preclinical models of acute myocardial infarction, it is seen that MSC transplantation promotes neovascularization and myogenesis, which in turn results in improved myocardial function [ 27 ]. San Roman et al. studied four groups which include one group of 30 patients receiving bone marrow mononuclear cells (BMMNC), 30 patients assigned to granulocyte colony-stimulating factor (G-CSF), 29 patients receiving G-CSF + BMMNCs, and a placebo group of 31 patients receiving standard therapy [ 29 ]. Quyyumi et al. in their RCT in 161 patients used autologous CD34 (+) cell (CLBS10) in 78 patients [ 26 ], whereas Peregud-Pogorzelska et al. delivered autologous bone marrow-derived lineage negative stem/progenitor cells in 15 patients in the treatment arm [ 25 ]. The researchers of the rest of our 24 included studies used autologous bone marrow mononuclear stem cells to treat the patients. We observed that the type of stem cell used did not have any influence on the outcome of this therapy. Figure 3 given below illustrates the different types of stem cells used to regenerate the damaged heart muscle. Figure 3 Different types of stem cells used to regenerate the damaged heart muscle Route and Time of Administration of Stem Cell Regarding the route of administration, intracoronary administration of the stem cell was performed in all our included studies. In addition to that, Moreira et al. in their randomized control trial (RCT) showed retrograde approach (intravenous coronary approach) to deliver stem cells was safe and cell retention by cardiac tissue is more in the anterograde (intra-arterial coronary) approach [ 21 ]. Timing of administration did not have much effect on the outcome. Huang et al. in their clinical trial (CT) divided the control group into three subgroups, and they got bone marrow mononuclear stem cells (BMMNC) infusion within two hours, three to seven days after percutaneous coronary intervention (PCI), and seven to thirty days after PCI respectively [ 14 ]. Effects of cell therapy given within 24 hours were noticed the same as given three to seven days after the primary PCI [ 14 ]. Similarly, Sürder et al. in their RCT demonstrated there is not much difference in the outcomes in groups where BMMNCs are administered five to seven days after ST-elevation myocardial infarction vs in groups where it is administered three to four weeks later [ 31 ]. Limitations There are some limitations of our study, we had a small number of people used in these studies, and a few of them were lost during the follow-up. Additionally, the optimal time of stem cell delivery has not been determined. Moreover, we included only the studies conducted from 2011- 2021 to concentrate more on the updated information. Conclusions Our study focused on evaluating the safety and effectiveness of stem cell therapy in patients with acute myocardial infarction. We found that most of our included studies showed significant improvement in myocardial function after stem cell therapy, but some of the studies failed to show the same improvement. Also, we observed that the stem cell therapy was safe, well-tolerated and no major adverse effects were reported. Because the result was still inconsistent and contradictory, we need to perform high-quality, well-designed clinical trials with a large sample size and more comparable results to assess and establish the efficacy of stem cell therapy in patients with acute myocardial infarction.

10. 7759/cureus. 17258

Cureus

Stem Cell Therapy and Its Significance in Pain Management

Pain management has always been a challenging issue, which is why it has been a major focus of many rigorous studies. Chronic pain which typically lasts for more than three months is prevalent at an astounding rate of 11% to 19% of the adult population. Pain management techniques have gone through major advances in the last decade with no major improvement in the quality of life in affected populations. Recently there has been growing interest in the utilization of stem cells for pain management. Advancement of stem cell therapy has been noted for the past few years and is now being used in human clinical trials. Stem cell therapy has shown promising results in the management of neuropathic, discogenic back, osteoarthritis, and musculoskeletal pain. In this article, we will discuss the role of stem cells in the pain management of the aforementioned conditions, along with the mechanism, adverse effects, and risks of stem cell therapy.

Introduction and background The term “chronic pain” refers to pain quality that remains persistent after the healing process or is present without the existence of tissue damage, which typically lasts more than three months [ 1 ]. This type of pain is a consequence of psychologic, biologic, and social factors in which the evaluation and management usually require a multifactorial approach [ 2 ]. In developed countries, about 30% of patients complain of moderately severe to severe pain which has lasted longer than six months [ 3 ]. A noted increment in chronic pain prevalence is correlated to adult life, estimating it to be at a rate of 11% to 19% [ 1 ]. On the other hand, the term “stem cells” points to the cells that are undifferentiated and have the capability to differentiate and replicate into variable types of tissue. Back in 1998, the initial cultivation of embryonic stem cells derived from a human was successfully established. Thereafter, stem cell attraction and interest have been growing. The transplantation of stem cells was applied just recently in the therapeutic measures of pain, as it has shown promising outcomes for the management of multiple disease entities including neuropathic pain, degenerative diseases of the joints, musculoskeletal pain that is unresponsive to the standard medical treatment, and intervertebral disc (IVD) disease [ 4 ]. Review Stem cell therapy Stem cells are defined as undifferentiated cells capable of perpetual self-renewal and have the ability to differentiate into specialized cell types. During the embryonic stage of development, these cells play an important role in forming organs, and in adults, they help in organ repair as well as renewal of tissue functions. A stem cell is comprised of many distinct cell types, namely embryonic, adult, and induced pluripotent stem cells (iPSCs). Embryonic stem cells can be obtained from the inner cell mass of blastocysts and are capable of differentiating into all three germ layers. Adult stem cells can be categorized according to the tissue of origin, such as placenta and umbilical cord stem cells, hematopoietic stem cells, bone marrow-derived mesenchymal stem cells (MSCs), and adipose-derived MSCs. They are present in all tissues of the body and are important for repair and renewal when tissues or organs are damaged [ 5 ]. Adult stem cells were recently induced to develop into a pluripotent state in cells called iPSCs. These cells share some of the same characteristics as embryonic stem cells such as proliferation, morphology and gene expression. This was made possible by retroviruses, which were used to carry genes for transcription factors in adult cells [ 6 ]. Furthermore, stem cells can be classified based on their ability to differentiate. Totipotent cells are found when the fertilized ovum starts to divide and can differentiate into any form of cells to build organs. Pluripotent cells are generally in the form of embryonic stem cells or iPSCs and are capable of differentiating into any cell type but cannot form complete organs. However, differentiation into specific cell types remains very poor. It is still a subject of immense interest with regards to what constitutes the optimal substrate and environment for differentiation into specific cell types. Studies in animal models using PSCs or embryonic stem cells resulted in the formation of unusual solid tumors called teratomas resulting in early setbacks of human trials. However, over the years PSCs have been modified to limit this unusual proliferative capacity and have been successfully used to treat animals with conditions like diabetes, acute spinal injury, and visual impairment [ 5 ]. Multipotent stem cells are found in the form of adult cells and have the plasticity to become all the progenitor cells for a particular germ cell layer or they could be restricted to differentiate into one or two specialized cell types depending on the substrate provided. One such example is how retinoic acid has been used to induce MSCs into neuronal cells [ 7 ]. Additionally, the lack of immunogenicity and no ethical jurisdiction makes them best suited for use in humans [ 8 ]. These qualities make MSCs an ideal candidate for clinical application. Different methods depending on the site can be used to harvest stem cells. Bone marrow stem cells can be harvested directly from blood, bone marrow (iliac crests/sternum) or the umbilical cord. Adipose tissue stem cells can be harvested by liposuction or excision of adipose tissue. Adipose tissue samples are digested with the help of collagenase and then they are centrifuged. This separates the precipitate layer containing adipose stem cells at the bottom and is called the stromal vascular fraction [ 7 ]. How does it work? MSCs can rapidly divide and repair damaged cartilage making them attractive for repairing damaged tissues. MSCs have several advantages. They can be harvested from one’s own body and are easy to prepare. Further, they are immunoprivileged, meaning they do not elicit a significant immune response. First off, tissue engineering is used to prepare scaffolds to generate functional tissues for the replacement of damaged tissues. Harvested MSCs are then seeded into the scaffolds where they begin their function. Direct engraftment of the host is one of the main working mechanisms of MSCs. MSCs are initially guided by the signaling cascades initiated at the sites of injury. As the platelets aggregate, they release cytokines which activate an influx of macrophages and neutrophils to the site of injury. The permeability of blood vessels at the injury site increases, setting the stage for homing of MSCs, and therefore increasing the aggregation of MSCs [ 4 ]. Besides direct engraftment, MSCs release a number of trophic factors in a paracrine fashion including growth factors like TGF-beta, factors inducing angiogenesis, transcription factors, and cytokines like PGE2, IL-2 and IL-6 [ 4 ]. In fact, studies have shown that paracrine function may be even more significant than direct host engraftment as the effects of MSCs continue to appear in vivo, even after the cells are displaced or resorbed. Another mechanism by which they exert their effect is immunomodulation. MSCs inhibit the differentiation of immature monocytes into dendritic cells, which are responsible for antigen presentation to naïve T cells. T cells secrete interferon-α and IL-4, which are responsible for inflammation. MSCs halt the development of T cells in G0/G1 phase and consequently the secretion of interferon-α and IL-4 as well. Finally, they act on natural killer cells making the local environment less susceptible to autoimmune regulation [ 9 ]. Stem cell therapy in neuropathic pain The Implementation of Stem Cells in Neuropathic Pain Therapy Chronic neuropathic pain (NP) is considered to be on the rise, especially with the increase in diabetes prevalence among the United States population [ 10 ]. NP is pain that results subsequently from a lesion or disease affecting the somatosensory system. NP was described to have a prevalence of 3% to 17% of the adult life population. Overall, 20% to 25% of all chronic pain is associated with NP. Stem cell therapy is a novel method of treatment that is gaining attention [ 1 ]. The classic qualities of NP were noted to be burning, shooting, and sharp [ 11 ]. NP can be classified as peripheral or central origin. Trigeminal neuralgia, radiculopathy, polyneuropathy, peripheral nerve injury, and postherpetic neuralgia are all subtypes of peripheral origin NP. On the other hand, subtypes of central origin NP include brain injury, multiple sclerosis, spinal cord injury, and central poststroke pain. All of the previously mentioned subtypes accompany chronic NP that consequently follow the lesion or the disease [ 1 ]. The cell-based method of treatment was described by Chakravarthy et al. to be a novel therapeutic approach in the chronic pain associated with neuropathies and degenerative joint disease. Replacement of cells that were previously injured has an impactful role along with delivering trophic factors [ 12 ]. MSCs have been shown to be capable of self-renewing and have the potential to differentiate into variable cell types including neurons, adipocytes, osteoblasts, and more [ 13 ]. As of now, there is no definite treatment of NP syndrome with concurrent promotion of nervous system repair. Despite that, the novel treatment with the use of stem-cell transplant repairs the nervous system instead of isolated palliation. NP related to disorders including sciatic nerve injury, neuropathies associated with diabetes, and spinal cord injury was demonstrated to be successfully treated with stem cell therapy [ 10 ]. Stem Cell Therapy in Diabetic Neuropathy The term diabetic neuropathy refers to the occurrence of clinical signs and symptoms resulting from a dysfunction in the peripheral nerves in patients with diabetes mellitus. Diabetic neuropathy is considered to be the most common complication of diabetes mellitus, with a rate of 30% to 50% of affected individuals [ 14 ]. Investigations on the utilization of stem cells in diabetic neuropathy were done in three studies conducted on animal models described by Vadivelu et al. , where administration of stem cells was through the intramuscular route in the hind leg. MSCs were used due to their ability to differentiate into multiple cell types and their cytokine secretion ability. Moreover, a recent study suggests that MSCs can promote neurotrophic factors. The loss of these factors was reported to be partly attributable to diabetic neuropathy, making the use of these stem cells an advantageous point. One of these studies has conducted a trial with the use of marrow mononuclear cells, due to the easy accessibility to this type of cells. These studies have concluded that improvement was noted in two to 15 weeks duration following the injury. Furthermore, these studies have reported this novel approach as a successful treatment method for neuropathic associated pain [ 10 ]. The neurotrophic factors secreted by stem cells were reported to accommodate neuronal protection along with providing neuronal regeneration [ 15 ]. Although not the routine treatment for diabetic neuropathy, the use of MSC treatment is indicated in patients with diabetic foot ulcers, acute relapses, intractable symptoms, and critical limb ischemic disease. Thereby, this novel treatment targets both vascular and neural components. The benefit of using MSC therapy is known to be attributed to the modulation of the immune response by the short-term effects of the paracrine and juxtacrine roles rather than by lesion site engraftment of the MSCs as a long-term effect. Generating anti-inflammatory MSCs has been shown to alleviate diabetic neuropathy pain. In a mice treatment module, MSC therapy has been shown to reduce proinflammatory cytokine concentrations in the mice’s serum [ 13 ]. Prior to and after receiving MSC therapy, the mice were evaluated for painful diabetic neuropathy through two established behavioural assays [ 16 ]. Growth factor therapy has shown to be beneficial in diabetic neuropathy due to its ability to promote regeneration of neural tissue and angiogenesis, overall improving nerve function. Bone marrow-derived stem cells were found to produce both angiogenic factors along with neurotrophic growth factors, supplementing selective cells in order to sequent the neuronal regeneration process, and having a more beneficial effect than the growth factor treatment. [ 14 ]. Stem Cell Therapy in Trigeminal Neuralgia Trigeminal NP encompasses variable states of diagnosis, this includes trauma resulting in maxillofacial NP, odontalgia which is atypical, and burning mouth syndrome. Trigeminal NP is considered to be a localized pain. Thereby, its patient population forms an ideal group to investigate the innovative novel therapy. NP is associated with poor response to over-the-counter analgesia and opioids, along with moderate pain relief in 40% of patients in response to anticonvulsant medications and tricyclic antidepressants. MSCs were found to exert anti-inflammatory impacts via cytokine secretion, combatting the ongoing inflammation manifesting as the NP. The previously proposed hypothesis has been studied on animal models with trigeminal NP showing a notable reduction in the inflammatory symptoms and promising outcomes [ 11 ]. Sacerdote et al. conducted a study on an animal model with hind paw NP. He reported that utilizing MSCs derived from adipose tissue resulted in the reduction of interleukin-1b pro-inflammatory cytokine and increased levels of IL-10, which is an anti-inflammatory cytokine in injured nerve tissue. A notable decrease in mechanical allodynia and the accompanied thermal hyperalgesia was noticed [ 17 ]. MSC therapy administration was previously evaluated and well established to be safe in trials involving both humans and animal models. No abnormal transformations were reported on the cranial nerve-associated nerve physiology. Also, there were no changes in the distributed area of sensation for the trigeminal nerve branches (V2 and V3) and lastly, no notable changes in the injection site. In addition, there were no abnormal reports on motor nerve abnormalities of the face or the jaw involving the seventh cranial nerve or the trigeminal nerve’s motor branch respectively [ 11 ]. Stem Cell Therapy in Spinal Cord Injury Two experimental studies were conducted on mouse models to study spinal cord injury treatment with the use of stem cell therapy. Embryonic stem cell oligosphere culture derivatives such as oligodendrocyte progenitor cells were selected to be utilized by one study group to cause remyelination in the lesioned nerve leading to NP inhibition. Neuregulin was downregulated afterwards through an interfered RNA. Thereafter, the myelinating process was noted to be reduced along with an increment in the allodynia functional measures. The other study used nanoparticles in correlation with co-cultured human stem cells that are derived from adipose tissue. Both in vivo and in vitro studies resulted in an increase in the expansion and self-renewing process of the administered stem cells. This was especially noted in the GABAergic neurons with a reported significant decrease in the inflammatory cells and the inflammatory mediators along with allodynia improvement after an interval duration of four weeks [ 10 ]. Stem cell therapy in discogenic back pain Chronic low back pain affects 68% of adults older than the age of 60 worldwide. Stem cell therapy has shown beneficial results as an alternative to conventional regimens in the management of degenerative disc disease (DDD). The objective of stem cell therapy is to restore the disc’s cellularity and minimization of the inflammatory response [ 18 ]. The causes of disc degeneration are multifactorial; they involve aging, smoking, genetics, nutritional factors, mechanical injury, and comorbidities. Normally, the amount of water and proteoglycan content of the disc increases from the outer annulus fibrosus (AF) to the inner nucleus pulposus. Opposingly, the amount of collagen decreases in the disc from out to in [ 18, 19 ]. In DDD histology, there is a progressive loss of the transition zone between AF and nucleus pulposus over years. This is due to a change from collagen type II to collagen type I, which is synthesized by the nucleus pulposus. This eventually leads to dehydration and loss of proteoglycans [ 20 ]. Disc narrowing, which can be observed radiographically, can be caused by many mechanisms such as matrix metalloprotease-mediated disc degeneration, diminished disc nutrition, etc. [ 18 ]. The features of DDD are osteophytes, joint space narrowing, and end plate sclerosis, eventually leading to nerve root compression resulting in symptoms of pain and numbness. Angiogenesis advances from the periphery, eventually extending centrally into the nucleus pulposus, causing discogenic pain [ 18, 21 ]. Both non-surgical and surgical interventions have failed to manage this condition effectively. A controlled study which was done on 1, 450 patients targeting return to work (RTW) as an outcome measure revealed that 67% of the control group had RTW within two years but only 26% of patients could RTW after two years following intervention by fusion surgery [ 22 ]. Stem cell therapy restores the cellularity of the IVD and reduces inflammatory mediators. Patient selection plays a pivotal role in the success of stem cell therapy. This intervention can improve the overall outcome in patients who fail to respond to conservative treatment or are in the early stages of DDD [ 22 ]. Patients with a disability, as proven by functional scores such as a Pfirrmann grading of Grade III or IV on MRI, and moderate chronic back pain are considered as ideal candidates for the therapy in many studies [ 23 ]. Stem cell therapy is currently using stem cells from other sources, autogenic or allogenic in origin or primary cells harvested from the IVD. Nucleus pulposus progenitor cells (NPPC), AF specific progenitor cells, autologous IVD cells, iPSCs, autologous chondrocytes, MSCs (derived from adipose tissue, bone marrow or umbilical cord Wharton's jelly) and embryonic stem cells are some of the different cell lines that have been used as stem cells for discogenic back pain. The different cell lines differ in their characteristics depending on the origin tissue [ 18 ]. Harsh environments such as nutrient scarcity, acidic condition and low cellularity make utilization of undifferentiated stem cells a major challenge. The microenvironment of the cultivation culture determines their effectiveness and production of stem cell phenotypes. Stem cell priming with growth factors or different environmental factors such as glucose, oxygen, etc. , has been rigorously researched by in vitro and in vivo studies [ 18 ]. Synergistic effect in making a favorable environment has been seen in in vitro studies that have combined IVD and MSCs. Inflammatory cytokines and matrix degenerating enzyme-related genes were suppressed in an in vitro study that combined human MSCs with rat IVD-NP cells. Similar effects were seen in some in vivo studies, with animal models using MSCs. 81% to 91% improvement in MRI signal intensities were noted in post-nucleotomy rabbit models due to suppression of type 1 collagen formation with the use of MSCs injection compared to sham-treated discs which showed 67% to 60%. Reduced disc tissue degeneration, microenvironment catabolism, recovered disc height and decreased pain has been evident in the growing number of studies. Culturing of human MSCs in vitro has proven that many of these effects are reproducible. Improvement in pain and function has been noted in limited clinical studies that have been done. Both animal and human studies have successfully shown evidence supporting disc regeneration and at least partial recovery in addition to safety and feasibility. To collect more data on human benefits, further clinical studies are required [ 24 ]. Gene therapy has been advanced in order to overcome the shortcomings of conventional methods. Over direct delivery of proteins, gene therapy has advantages such as enhanced efficacy and sustained anti-inflammatory factors and growth factors synthesis endogenously [ 18 ]. Image-assisted percutaneous injection through the AF has been used conventionally, although it has raised some safety concerns. Alternative routes of administration are through the pedicles (transpedicular approach) and use of delivery vehicles [ 25 ]. Delivery vehicles, hydrogels have been utilized to overcome retention issues and provide additional support for cell survival and phenotype retention. Scaffolding materials, such as hyaluronan, fibrin, and atelocollagen have been developed to improve efficiency of stem cell delivery into degenerated IVDs [ 18 ]. Stem cell therapy in osteoarthritis Osteoarthritis (OA) remains to constitute a large burden to healthcare and negatively impact the quality of life; along with other conditions ranked as 11th highest contributors of global disability [ 26 ]. The estimated prevalence of problematic hip/knee OA is approximately 242 million globally; 3. 8% when accounting for decreased quality of life (QoL) and societal burden [ 27, 28 ]. Knee OA (KOA) carries a higher incidence when compared to other joints (i. e. , hip), increasing to 60% among the obese population [ 29 ]. Of note, a population-based cohort study published in 2008 estimated a 45% lifetime risk of symptomatic knee-osteoarthritis [ 30 ]. Specifically, KOA was shown to be affecting nine million adults (over the age of 45 years) in the United States with symptoms ranging from moderate to severe [ 31 ]. With the escalation of an aging population, sedentary lifestyle, and obesity, the number of patients affected with KOA is undoubtedly on the rise. The concept of age-related joint degeneration in the event of KOA has been substituted by other theories and factors. In fact, KOA has been shown to be the end result of a chronic interplay between heterogeneous systemic and local reactions. Systemic factors include age, race, ethnicity, and diet, while local ones include body-mass-index (BMI), trauma history, occupational, and mechanical factors [ 32 ]. The aforementioned factors, along with indigenous biologic factors (i. e. , cytokine homeostasis) result in trivial changes at the knee joint, particularly at the area interspacing between the subchondral bone and the articular cartilage, which becomes evident over time [ 33 ]. Owing to this effect, an increase in bone mass and trabecular thickness ensues with reduced ability to withstand stress and compression impact on cartilage; the stage at which chondrocyte cell senescence occurs [ 30 ]. Exhibited degeneration of the articular cartilage is a pathognomic feature for OA associated with bone remodeling, osteophyte formation, capsular distribution, and periarticular muscle atrophy [ 30 ]. The complexity of cartilage is derived from multiple building blocks and constituents such as chondrocytes type II, collagen, proteoglycan, and an abundantly hydrophilic extracellular matrix containing highly complexed meshwork of cytokines and growth factors secreted by surrounding synovial cells and chondrocytes [ 30 ]. Prolonged exposure to stressors including reactive oxygen species (ROS) and nitric oxide (NO) in turn trigger macrophages and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kβ). They play an important role in immune regulation and activation of cytokines which are associated with inflammation and ultimately cause disruption of homeostasis in the synovial fluid [ 30 ]. Other important associated inflammatory molecules include IL-1 β, TNF-α, INF- γ, TGFβ, MMP-9, and MMP-13, all of which contribute to pathologic hallmarks in the pathogenesis of OA [ 34 ]. Current conventional therapy of OA is mainly focused on providing symptomatic control over hindering disease course progression. Commonly used radiographic grading systems for KOA include Kellgren-Lawrence (K/L). Along with exercise therapy, physical therapy, and weight reduction to strengthen adjacent muscles, current pharmacologic treatment includes acetaminophen, non-steroidal anti-inflammatory drugs (NSAIDs), gabapentin, pregabalin, and opioids, which are used for K/L grade 0-1. Other means include valgus directing force bracing [ 32, 35 ]. Intra-articular (IA) corticosteroid injections are commonly prescribed as a symptomatic treatment for grade ≥ 2 upon which osteophyte formation, joint space narrowing, subchondral sclerosis, and deformity of the joint are evident [ 32, 35 ]. Subsequently, total knee arthroplasty (TKA) becomes an option for K/L grade 3-4. It is an invasive procedure associated with a significant number of complications. It was shown that 20% of patients who underwent TKA will have persistent knee pain with a possible need for TKA revision and is subject to further risk and morbidity including loss of function within a year after the procedure [ 34 ]. Moreover, with the uncertainty of progression and symptomatic treatment following TKA, it was estimated that 61% of medical expenses are spent on TKA procedures [ 32 ]. Compared to conventional therapy which provides symptomatic treatment, multiple randomized control trials (RCTs) have demonstrated promising potential for stem cell therapies tackling OA morbidity and healthcare burden. The concept of stem cells arises from multipotent cells which have the ability to differentiate into different cell types and possess auto-regeneration according to the body’s needs. These cells can be readily extracted from tissues in the body including the bone marrow, adipose tissue, and the synovium [ 36 ]. Of note, MSCs are adult stem cells (ASCs) derived from the mesodermal origin and have the ability to differentiate into different connective tissue cells including osteocytes, adipocytes, and chondrocytes [ 37 ]. In vitro MSCs were shown to have the ability to differentiate into chondrocytes and enhance the proliferation of resident progenitor cells in vivo. With the help of growth factors and extracellular matrix proteins, MSCs demonstrated the ability to create a repair microenvironment that could explain the postulated theory of associated pain reduction and the termination of the disease progression cycle [ 38 ]. Subsequently, this method of autologous repairing gives rise to the activation of senescent, metabolically active chondrocytes’ ability to repair damaged tissue [ 38 ]. Trials suggested that transplanted MSCs in the joint are activated and subsequently expressing anabolic genes such as the Indian hedgehog and other ‘hit and run’ genes that enhance collagen type synthesis, analgesic peptide transcription, and the production of various anti-inflammatory cytokines and analgesic proteins [ 39, 40 ]. The aforementioned disease-modifying properties build an evidence-based tool for MSC therapy in the context of OA disease progression and stabilization [ 36, 40 ]. Performed clinical trials have subjectively depended on different scoring systems in assessing the effect of engrafted MSCs including the Knee Injury and Osteoarthritis Outcome Score (KOOS), Western Ontario, and McMaster Universities Arthritis Index (WOMAC), and the Visual Analog Scale (VAS). According to Buzaboon et al. , MSCs treatments demonstrated pain reduction and improved knee joint function post-intervention from baseline according to different parameters such as KOOS, VAS, and WOMAC. Most trials have demonstrated a peak in positive effects occurring between the sixth and 12th months following treatment [ 33, 39, 41 ]. However, subjectively assessed results were not statistically significant in most trials selected in the study. This was assumed to be due to the late intervention with MSCs to treat OA, which was probably at an advanced stage. The theory is that MSC implantation is more effective in OA when provided at early stages of disease [ 33 ]. Nonetheless, MSCs implantation in OA has been shown to be a more superior approach to conventional therapy. This can be explained by the fact that MSCs treatment can hinder inflammation, restore vital tissue components and damaged cartilage, and the negligible number of serious local or generalized systemic adverse events and complications. However, with the early promising results of MSCs therapy in OA, many questions regarding precise mechanisms remain unexplored. There is still a need for further clinical trials and studies to be performed at the early stages of OA (K/L Grade 0-2), examining different cell culture preparation, and exploring different dosing intervals, frequency of therapy, and appropriate delivery method. Stem cell therapy in musculoskeletal disease The World Health Organization (WHO) has enlisted musculoskeletal disease as one of the most common causes of severe long-term pain and physical disability. In 2016, it comprised the second-highest global volume of years lived with disability [ 42 ]. Musculoskeletal conditions may involve bones, muscle, tendons, cartilage, ligaments, joints, with pain often being the first presenting complaint. With increasing life expectancy, the burden of musculoskeletal diseases is on the rise. However, the treatment modalities are limited to managing the symptoms rather than curing it. The rapid development of regenerative medicine and stem cell therapy offers hope. MSCs can rapidly divide and repair damaged cartilage, making them suitable to treat conditions such as tendon, ligament, and cartilage damage. Stem cells are easy to extract, prepare, and inject making them suited for outpatient settings. MSCs are prepared via centrifugation of harvested tissue, and then separated to isolate MSC dense fluid. This fluid is then injected into the autologous therapeutic tissue of interest such as joint, tendon, cartilage, or intervertebral disc. Bone marrow aspirate concentrate (BMAC) has been commonly used for the treatment of musculoskeletal conditions, but there are reports of using stem cells derived from peripheral blood stem cells (PBSCs) [ 43 ]. There are limited data evaluating the safety and efficacy of stem cells in humans for musculoskeletal disease. In a systematic review by Law et al. , bone marrow-derived MSCs were found to be safe and feasible for chronic patellar tendinopathy and femoral head necrosis, but three out of eight studies did not reach statistical significance [ 44 ]. In a five-year study conducted by Pascual-Garrido et al. , stem cell therapy showed improvement in clinical scores, pain, as well as improvement in ultrasound grade in seven out of eight patients with chronic patellar tendinopathy. Maximal improvement was seen at two years and was maintained at five years [ 45 ]. In patients with femoral head necrosis, core decompression plus BMAC resulted in better pain relief than decompression alone as assessed by Harris Hip Score, but it did not reach statistical significance. However, the injection of PBSCs improved pain function, imaging and other patient-reported outcomes over control in patients with osteonecrosis of femoral head [ 46 ]. Most of the studies conducted involved a small sample size. Despite this limited evidence supporting the use of stem cells in musculoskeletal conditions, their use continues to grow in the United States and other parts of the world. Furthermore, there has been a recent interest in anterior cruciate ligament regeneration using MSCs. Although initial success has been achieved in animal models, its potential is yet to be realized in humans. There is a need for long-term randomized controlled trials before stem cell therapy gains wide acceptance in treatment of musculoskeletal diseases. Side effects/risks of stem cell therapy Stem cell transplantation is considered a safe option for the above-mentioned medical conditions. Many researchers have reported no adverse outcomes among patients following stem cell therapy. Lalu et al. conducted a meta-analysis on 1012 patients who underwent stem cell transplantation for multiple medical conditions such as inflammatory bowel disease, stroke, cardiomyopathies, cardiac ischemia and reported no adverse events during and 90 months after the procedure [ 47 ]. Other studies have reported mild adverse effects after the procedure among which include nausea, vomiting, infections, and endocrine dysfunction [ 48 ]. Yobu et al. conducted a meta-analysis on the use of MSC transplantation for the management of knee osteoarthritis. This study comprised 582 patients and reported that the majority of adverse events were minor such as pain at the site of injection, swelling of knee joint, reduced range of motion, infections. Some patients manifested small or large bowel obstruction which was managed accordingly [ 49 ]. El-Badaway et al. conducted a meta-analysis on the efficacy of stem cell therapy used for the management of diabetic patients. Their study results reported minor adverse effects such as nausea and abdominal pain in 21. 7% of patients [ 50 ]. However, some clinicians have reported some serious adverse events associated with it. An important adverse effect of stem cell therapy is the risk of tumor formation secondary to the malignant transformation of MSCs. Certain factors which account for tumor formation secondary to stem cell therapy consist of expression of oncogenic tumor cell markers, chromosomal instability or increased telomerase activity. Other factors which have been observed to play a role in tumor progression include the inhibitory effect of MSC on immune regulatory cells of the body. Some researchers have proven that if there are any existing tumor cells in the body, they can escape immune surveillance after the commencement of stem cell therapy [ 51 ]. If the MSC exhibits disease memory or has been excised from diseased tissue it can exhibit the expression of inflammatory markers which can lead to disease expression in transplanted patients. There have been previous studies that have demonstrated the development of malignancy after stem cell transplantation [ 52 ]. Pan et al. conducted a study by transplanting transformed MSC derived from human bone marrow and liver in immunodeficient mice. They observed the development of sarcoma-like tumors and concluded that efficient screening of MSC cultures can prevent tumorigenesis as an adverse event [ 53 ]. Conclusions The potential of human-based PSC therapy in a wide array of applications is promising. Clinical application of stem cell-based therapy has been described in approximately 14 diseases and trauma-related implications and is currently investigated by multiple clinical trials. The utilization of stem cell therapy directed towards the regeneration of certain cell types including senescent cells and other components of organ tissues is challenging yet very promising. Ongoing clinical trials have shown positive outcomes of stem cell therapy in treating chronic diseases associated with significant morbidity and reduced QoL. In this review article, we shed light on the current advancement of stem cell therapy in pain management. Along with trophic factors and other anabolic effects of stem cell therapy, the pain of chronic diseases including DM-associated neuropathic pain, OA, back pain, ligamentous pain, and other injuries can be effectively managed with this modality reflecting into decreased healthcare burden and improvement of QoL. In addition, their effect can be reasonably fast, and the long duration of their effect adds to their advantages. The strengths of this review include the fact that there are many supporting studies and reviews which encourage the advancement of stem cell therapy, the only limitation is that majority of these studies are in their earlier stages. Despite the fact that utilization of stem cell therapy is still in an early stage and most major studies are still at their preclinical phase, the prospect of this approach is very assuring given the advancement in technology and the increasing number of clinical studies and evidence. Future studies should focus on human-based clinical trials since the results of animal trials may not entirely be applicable in humans.

10. 7759/cureus. 17705

Cureus

Bone Grafts in Trauma and Orthopaedics

Worldwide, there are millions of patients each year suffering from bone-related illness due to trauma, degenerative diseases, infections or oncology that require orthopaedic intervention involving bone grafts. This literature review aims to analyse the characteristics of the different bone grafts: autografts, allografts and synthetic bone substitutes. The review will assess their medical value based on their effectiveness as well as scrutinising any drawbacks. The goal is to identify which options can give the optimal result for a patient being treated for a bone defect. Bone autografts remain the gold standard since there are no issues with histocompatibility or disease transmission while possessing the ideal characteristics: osteogenicity, osteoconductivity and osteoinductivity. However, synthetic options such as calcium phosphate ceramics are becoming popular as a viable alternative for treatment since they can be produced in desired quantitates and yield excellent results while not having the problem of donor site morbidity as seen with autografts. Furthermore, advancements in fields such as bone tissue engineering and three-dimensional printing are generating promising results and could provide a path for excellent treatment in the future. The emergence of such innovations highlights the importance and the constant need for improvement in bone grafting.

Introduction and background Worldwide, an estimated 2. 2 million orthopaedic procedures involving bone grafting take place annually [ 1 ], with the incidence rate projected to increase by 13% each year [ 2 ]. Bone substitutes can be natural, synthetic or composite materials. The first recorded surgery involving bone grafts was in 1668 [ 3 ]. However, American Orthopaedic surgeon Fred H. Albee is credited for pioneering bone graft surgery in 1906. In 1965, Dr. Marshall Urist discovered demineralized bone matrix and bone morphogenic proteins with the first marketed demineralized bone matrix available in 1991. Bone histology Bone is a specialized connective tissue characterized by a mineralized extracellular matrix and cellular components, which include osteoprogenitor cells, osteoblasts, osteocytes and osteoclasts. The bone matrix consists of organic and inorganic components with the latter accounting for about 70%-75% of bone mass [ 4 ]. The osteoid is the unmineralised organic component consisting of type 1-collagen fibers and acidic ground substance made up of proteins, carbohydrates, proteoglycan aggregates and osteonectin [ 5 ]. The inorganic component is primarily composed of calcium phosphate and calcium carbonate. Crystallisation of these minerals results in the formation of hydroxyapatite. The tensile strength of bone is a result of the collagen fibers and the hydroxyapatite crystals attribute to its compressional strength [ 6 ]. Osteoprogenitor cells found in the periosteum and endosteum are essentially stem cells of mesenchymal origin that give rise to osteoblasts. Osteoblasts are metabolically active bone-forming cells. They are responsible for producing and secreting osteoid, the organic component of the extracellular matrix, which subsequently becomes calcified [ 7 ]. Once the osteoblasts become trapped in the organic matter they differentiate into osteocytes. Osteocytes are mature cells that are responsible for maintaining bone tissue by providing transport channels for nutrients and waste products via their cytoplasmic processes. Lastly, osteoclasts are multinucleated cells derived from monocytes. They are responsible for bone resorption, therefore, are essential for the repair, remodelling and growth of bone [ 7, 8 ]. Ideal bone graft properties Bone grafts properties can be divided into three groups based on their biological actions: osteogenicity, osteoinductivity and osteoconductivity. Osteogenic capabilities of the bone graft reflect on their ability to synthesize new bone. Osteoinductive grafts can stimulate mesenchymal cells and osteoprogenitor cells in the surrounding host tissue to differentiate into osteoblasts, thus, promoting the formation of new bone. Osteoconductive grafts provide passive microscopic and macroscopic scaffolding that supports bone formation and bone growth by providing a porous structure through which cells such as osteoblasts can migrate through and blood vessels can grow into, consequently incorporating the graft tissue into the host’s bone [ 9, 10 ]. An ideal bone graft has all three of the aforementioned properties as well as being of low cost, lack risks for infection and being readily available in any desired quantities. Incorporation of bone graft Bone graft incorporation can be generalized into two phases. The first phase begins with the formation of hematoma around the implanted bone, followed by the release of cytokines and growth factors. Subsequently, inflammatory processes take place resulting in the development of fibrovascular tissue. Although, it is worth noting that the incorporation of grafts differs between cortical and cancellous bone grafts due to their structural difference. The vascular response seen with cancellous bone is much greater than cortical bone since it is much more porous. Review Natural bone grafts Autograft Autograft bone is harvested from the patient’s own body from a different unaffected site. The graft is most often obtained from distant sites, the most popular being the posterior iliac crest since graft tissue obtained here is said to have the highest osteogenic potential and it provides both cancellous and cortical bone. Other sites include femoral greater trochanter, proximal-distal tibia, calcaneus and distal radius. Autogenous bone grafts remain the gold standard therapy because to some extent they have all three of the desired properties: osteogenic, osteoinductive and osteopromotive [ 9 ]. Since the donor and recipient are the same individuals there will be no issues with histocompatibility and there are no risks of transmitting disease. Although patients can potentially suffer from donor site morbidities such as chronic pain, hypersensitivity, paresthesia, pelvic instability and infections. The complications rate following iliac crest bone graft harvesting ranges between 2% and 36% [ 2 ]. Allograft Allografts bone graft material is harvested from living donors or cadavers. Harvested bone tissue is processed to decrease the risk of host-versus-graft immune response and facilitates the removal of harmful substances from the bone tissue that may transmit disease. Through processing methods, the shape and size of the graft material can be altered to fit for their intended use. However, the processing step does seem to alter the structure of the graft thus affecting its mechanical competency and its ability to stimulate bone healing [ 11 ]. The graft material is preserved by deep freezing or freeze-drying. Deep frozen allografts are advantageous since they retain their structural properties. On the other hand, the advantage of freeze-drying is that it allows storage of the grafts at room temperature but freeze-dried allografts can become fragile to torsion or bending since microfractures form along the collagen fibers of the allograft, hence, deteriorating its structural strength [ 12 ]. Allografts have many advantages, namely the abundant supply of graft material, which can be obtained in the desired configuration. There is no need to compromise host structures to obtain the graft tissue, and subsequently, there will be no donor-site morbidity as seen with autografts. On the other hand, regardless of the processing and sterilization of the grafts, there will always be a risk of host immune response and disease transmission. That being said, disease transmission is extremely rare with the risk of contracting HIV or hepatitis B/C estimated to be about 1/1. 6 million. Synthetic bone substitute There has been significant research and development to generate alternative options in the form of synthetic bone substitutes. The aim is to provide a cost-effective graft substitute that is similar in structure and strength of human bone, support new bone formation, bioactive, biocompatible, osteoconductive, osteoinductive, and osteointergrative. Metal Trabecular metal technology is an innovative 3D material made from the metal tantalum that shows excellent biocompatibility and resistance to corrosion. It is structurally similar to cancellous bone. Trabecular metal technology surfaces have nano-textured topography while exhibiting high porosity of up to 80% that are consistent in size and shape [ 13 ]. Ceramic Ceramic graft options include calcium phosphate, calcium sulphate, calcium phosphate cement and bioactive glass. Ceramic based bone substitutes account for 60% of the synthetic graft market and have been very popular because they are bioactive providing good osteoconductivity, become integrated into the host tissue very well, can be obtained in any desired amount while there are no risks of disease transmission or donor site morbidity as seen with the natural bone grafts [ 14 ]. Drawbacks of ceramics are the relatively high cost and the fact that they are brittle with very low tensile strength. Calcium phosphate cement, which is available in an injectable form, allows the possibility of intraoperative moulding of the graft. The cement consists of a mixture of dicalcium phosphate anhydrite and tetracalcium phosphate. Once injected into the graft site using a dual-chamber syringe, the cement begins to harden and is converted to porous hydroxyapatite with osteoconductive properties. Calcium phosphate cement has been successfully used to fill bone defects but while the recipe for the cement is FDA approved it is still under research and development to be perfected. Bioactive glass is also a ceramic-based bone substitute. It exhibits very good osteoconductive and osteointergrative properties. Once implanted, the ions in the bioactive glass such as Na+ and K+ react with the extracellular fluid, resulting in the production of a silica-rich gel layer forming over the implant which is highly porous [ 15 ]. Subsequently, Ca2+ and PO43+ from the extracellular fluid react and then precipitate onto the silicone-rich layer forming a coat of hydroxyapatite on which blood proteins, growth factors and collagen will be adsorbed. The newly formed hydroxyapatite layer is like the naturally found version in bone and attracts macrophages to initiate tissue healing as well as osteoblasts and osteoprogenitor cells to begin new bone formation. One difficulty faced with bioactive glass is that they may be difficult to fix to the bone since they can be very difficult to shape and when attempted may break in the process [ 16 ]. Polymer Based Polymer-based bone substitutes provide a scaffold structure promoting osteoconductivity. Synthetic polymers can be manufactured in any amount as well as having their structure and composition being designed as desired. There is also no concern of immunogenicity or the presence of pathogenic agents. Non-degradable polymers such as ultra-high molecular weight polyethylene have been used in the production of acetabular cups used in total hip arthroplasty [ 17 ]. Over the past decade, degradable polymer bone substitutes have been used more and more since it is ideal to have a synthetic graft that is completely resorbed leaving no foreign material behind. However, polymers do not provide much mechanical strength, so it is best when they are used together with another bone substitute. Coralline Hydroxyapatite Coralline hydroxyapatite is obtained from calcium carbonate extracted from sea coral. The material is exposed to heat and pressure while in an aqueous phosphate solution, resulting in the conversion of the calcium carbonate exoskeleton to calcium phosphate [ 16 ]. Coralline hydroxyapatite has high compressional strength but does not offer much tensile strength like synthetic hydroxyapatite. It is generally accepted that the minimum requirement for the pore size in biomaterials for promoting cell migration and vascular ingrowth is 45-100μm [ 16 ], The problem with synthetic hydroxyapatite is that the pore size is not consistent, which ultimately affects its effectiveness as a bone graft. In contrast to this, the structure of coralline hydroxyapatite is very similar to cancellous bone and the naturally formed pores tend to have a more constant size. The commercially available products can have a mean pore size of 200μm or 500μm [ 9 ]. Coralline hydroxyapatite has been successful in treating metaphyseal defects such as tibial plateau fracture. In recent times, there have been studies on animals where coralline hydroxyapatite was used as a carrier for bone morphogenic proteins with promising results being achieved [ 9 ]. Clinical use of bone grafts Trauma In Europe, 5. 7 million patients are admitted every year due to trauma, costing health care services €78 billion each year [ 18 ]. In the UK alone, there were 186, 000 bone and joint-related emergencies between 2012 and 2013 [ 18 ]. There can be cases when there is significant bone loss overwhelming the bone’s capabilities to heal, thus resulting in non-unions or mal-union. This warrants the use of bone grafts to bridge the gap between the fracture ends and promote bone healing. Oncology Primary malignant bone cancer is rare, with about only 559 cases of all sarcoma subtypes in 2011 in the UK, accounting for 0. 2% of all cancers [ 19 ]. Secondary bone cancers, however, are much more common and arise due to metastasis spread from other organs, most commonly from the thyroid, breast, lungs, kidney and prostate. When surgically resected, large defects are left in the treated bone, thereby, warranting the use of bone grafts to fill the voids. Future of bone grafts Bone Tissue Engineering Bone tissue engineering aims to essentially act as a source for unlimited autogenous bone tissue. Human mesenchymal stem cells have been the subject of several studies and experiments aimed at bone tissue engineering with promising results being yielded. The patient’s bone marrow is used as the source for the human mesenchymal stem cells, which are multipotent undifferentiated cells that are capable of undergoing chondrogenesis and osteogenesis [ 20 ]. The idea is to use a polymer scaffold onto which these cells grow onto while in the presence of bone stimulating factors. 3D Printing 3D printing allows the production of custom-made implants unique to a patient’s bone defect as well as ensuring that the printed product has the desired structure and composition to mimic bone architecture. The uniformity and size of pores can be carefully controlled in order to produce a complex implant that offers a scaffold with prime osteoconductivity. The implant can be printed layer-by-layer using modified hydroxyapatite powder as the feed material together with a polymer-based binder [ 21 ]. Once the printing is complete, the product is dried, cleaned and sintered at 1, 250°C for two hours [ 22 ]. Exposure to the high temperature decomposes the polymer-based binder, leaving behind only the ceramic body. The aim is to have 3D printed implants clinically available in the next 10 to 20 years [ 22 ]. The problem with the high temperature used in the manufacturing process is that it makes it impossible to infuse additional agents such as bone growth factors or antibiotics. Moreover, there is the possibility of coating the 3D printed scaffolds with human stems cells harvested from the patient. This would significantly enhance the integration of the graft into the host tissue. Gene Therapy Gene therapy is a promising method to promote new bone growth. The idea is to transfer genetic encoding information to the target site and induce bone healing by manipulating the endogenous host cells to produce specific proteins such as growth factors. The use of virus vectors for the expression of bone morphogenic proteins for bone formation has been successfully demonstrated in vitro and in vivo animal models [ 23 ]. Unfortunately, there are limitations to gene therapy that inhibits progress being made in this field. Firstly, there is always a risk of immune response when introducing a viral vector into the host tissue and such response can drastically deteriorate the effectiveness of the gene therapy. The use of viral vectors also raises the risk of infections. The possibility of tumour development is present if the transferred gene is encoded into a wrong position in the host’s DNA. Furthermore, for gene therapy to be successful the right gene should be targeted in the right cells but there is potential for the viral vector to target unwanted host cells. Conclusions The prevalence of medical conditions related to bone or joint impairment is on the rise primarily due to the ageing population and the major impact they have on the lives of the affected individual is apparent. Despite the presence of several choices, autografts remain the ‘gold standard’. However, synthetic bone substitutes are gradually gaining momentum in Orthopaedics, especially calcium phosphate ceramics. The future holds tremendous promise because of current researches, especially in 3D printing and bone tissue engineering, which aim to provide grafts that have all the ideal characteristics and being available in unlimited amounts. For now, these innovations may be some years away from human trials and may not even be cost-effective initially. However, with the continuous advancement in technology, there are plenty of reasons to be optimistic. The fact that several studies such as these are taking place further emphasizes the importance of bone tissue regeneration in Orthopaedics and the constant need to improve the bone grafting options being available for treating patients.

10. 7759/cureus. 18755

Cureus

Evaluation of In Vivo Adhesion Properties of New Generation Polyglactin, Oxidized Regenerated Cellulose and Chitosan-Based Meshes for Hernia Surgery

Introduction Composite meshes coated with anti-adhesive barriers have been developed by taking advantage of the robustness of polypropylene meshes for use in hernia repair. We aimed to evaluate the effects of composite meshes containing polyglactin, polycaprolactone, oxidized regenerated cellulose and chitosan on the adhesion formation. Methods Forty-two Sprague Dawley male rats were divided into six groups of seven rats according to the content of the meshes used. A defect was created on the right abdominal wall of the rats and an oval composite mesh of 2 cm in diameter was placed over the defect and fixed. The rats were sacrificed under anesthesia on the 7th postoperative day. Macroscopic and histopathological examination was performed and the incorporation of the mesh with the abdominal wall and the presence of intraabdominal adhesions were evaluated. Results When the macroscopic findings of the rats were evaluated, there was a statistically significant difference between the rat groups in terms of the distribution of peritoneal adhesion scores (p<0. 05). There was no statistically significant difference between the rat groups in terms of the distribution of inflammation, fibrosis and macrophage levels (p>0. 05). Conclusion It was evaluated that the development of intraabdominal adhesion and the strength of adhesion decreased when biocompatible adhesion barriers with anti-adhesive properties such as oxidized regenerated cellulose and chitosan were used in the structure of composite meshes used in hernia repair. Hemostatic and antibacterial properties of these substances are promising to create the ideal mesh.

Introduction Intra-abdominal adhesions are seen after abdominal operations and these adhesions cause complications such as chronic abdominal pain, small bowel obstruction, infertility in women, and cause iatrogenic bowel injury in secondary operations performed for adhesions [ 1, 2 ]. For these reasons, patients are frequently admitted to hospitals, repeated hospitalizations may be required and patients often have to be operated. This situation brings with it a great financial burden. A method that prevents the formation of intra-abdominal adhesions will eliminate the subsequent operations and the financial burden it will bring. Prosthetic meshes made up of different materials have been developed to prevent complications caused by intra-abdominal adhesions, but the ideal mesh has not been created yet [ 3 ]. Today, the most commonly used prosthetic mesh for hernia repair is polypropylene mesh. Because the polypropylene mesh causes an intense inflammatory response compared to other meshes and causes adhesion to the visceral organs, especially the small intestine, composite meshes containing anti-adhesive barrier, especially suitable for intraperitoneal use in laparoscopic hernia repair have been developed in order to prevent adhesion on the visceral side. In this study, composite meshes including polyglactin, polycaprolactone, oxidized regenerated cellulose and chitosan were used, which accelerated tissue growth and had anti-adhesive activity. We aimed to evaluate the effects of these meshes which would accelerate tissue regeneration by preventing complications that might occur due to mesh use on adhesion formation. Materials and methods This experimental study was initiated after it was approved by the Local Ethics Committee for Animal Experiments of Istanbul Medipol University (9/30/2015-E-2500). The animals were cared for according to the principles of the National Institutes of Health publication “Guide for Care and Use of Laboratory Animals”. Forty-two Sprague Dawley male rats were used in the study. The rats were given standard rat food and water during the experiment. All rats were kept separately in cages where 12 hours of light and dark environment were provided at room temperature. The rats were divided into six groups of seven rats each: Control group; polypropylene mesh, group A; polycaprolactone coated polypropylene mesh, group B; polycaprolactone coated polypropylene mesh containing 20% oxidized regenerated cellulose, group C; polycaprolactone coated polypropylene mesh containing 40% oxidized regenerated cellulose, group D; polyglactin coated polypropylene mesh containing 30% chitosan, group E; polyglactin coated polypropylene mesh containing 10% chitosan. Surgical procedure Surgeries were performed under sterile conditions. Rats were administered 100 mg/kg Ketamine (Ketalar®) and 5 mg/kg Xylazine hydrochloride (Rompun®) intramuscularly, to provide long-term anesthesia and analgesia. Tail pinch and extremity withdrawal responses were examined to understand the depth of anesthesia. The abdominal hair was shaved and the anterior abdominal wall was stained with povidone-iodine. The abdomen was entered with a midline incision. In the right abdominal wall, a 1x1 cm area of parietal peritoneum and partial muscle was excised at a distance of 2 cm from the medial wall to the midline incision. An oval 2 cm diameter mesh was placed with a 6-0 polypropylene continuous suture, leaving the medial wall 1 cm away from the midline incision. The abdominal wall was closed with 4-0 polydioxanone and the skin with 5-0 polydioxanone. Flunixin (Flumexin®) 0. 01 mg/kg was administered subcutaneously once for postoperative analgesia following wound closure. The rats were sacrificed under anesthesia with Ketamine and Xylazine hydrochloride on the 7th postoperative day. Macroscopic examination was performed and the union of the mesh with the abdominal wall (incorporation) and the presence of intra-abdominal adhesions were evaluated and noted. Then, for histopathological evaluation, the specimens were numbered according to the groups and placed in 10% formaldehyde and sent for pathological examination. Macroscopic evaluation of adhesion The presence and degree of intra-abdominal adhesions were evaluated according to the scoring modified from Greca et al. [ 4, 5 ] (Table 1 ). The type of adhesion was evaluated according to the classification recommended by Zühlke et al. [ 6 ] (Table 2 ). In addition, the union of the mesh with the parietal peritoneum (incorporation) was graded. While evaluating incorporation, the mesh surface was divided into four areas and each area was graded as 25%. Table 1 Criteria and scores for peritoneal adhesions modified from Greca et al. [ 4, 5 ]. Scores Classification 1 No adhesion 2 Omentum adhesion at suture zone 3 Omentum adhesion up to 50% of the mesh surface 4 Omentum adhesion more than 50% of the mesh surface 5 Visceral adhesion at suture zone 6 Visceral adhesion at mesh Table 2 Characteristics of adhesion types according to Zühlke et al. [ 6 ]. Type Characteristics 1 Filmy adhesion, easy to separate by blunt dissection 2 Stronger adhesion; blunt dissection possible, partly sharp dissection necessary; beginning of vascularization 3 Stronger adhesion; lysis possible by sharp dissection only; clear vascularization 4 Very strong adhesion; lysis possible by sharp dissection only; organs strongly attached with severe adhesions; damage of organs hardly preventable Histopathological examination The mesh was excised 2 cm from its upper border. The specimen was washed lightly with distilled water and then fixed with 10% buffered formalin. Samples were divided into segments, each containing the abdominal wall, mesh, and transition zone. Tissues were routinely followed up with an automatic tissue tracking device and placed in paraffin blocks. Samples were cut 4-5 micrometers, stained with hematoxylin and eosin and Picro Sirius red, and examined blindly by the pathologist under both light microscopy and polarized microscopy. Under the light microscope, the presence of inflammation, fibroblast proliferation, angiogenesis, granulation tissue, fibrosis, giant cells, macrophages and edema were examined at 200 magnification and were classified as grade 0 (none), grade 1 (mild), grade 2 (moderate) or grade 3 (diffuse) using a scale. Scoring was performed at the end of this grading. The presence of neutrophils and giant cells as a foreign body reaction was expected to show the biological incompatibility of the mesh. Statistical analysis of data NCSS (Number Cruncher Statistical System) 2007 (Kaysville, Utah, USA) program was used for statistical analysis. While evaluating the study data, in addition to descriptive statistical methods (minimum, maximum, median, frequency, ratio), Kruskal Wallis Test was used for the comparison of quantitative data in three or more groups, and Mann Whitney U test was used for pairwise comparisons. Fisher Freeman Halton Test was used to compare qualitative data. Significance was evaluated at p<0. 01 and p<0. 05 levels. Results When the macroscopic findings of the rats were evaluated, there was a statistically significant difference between the rat groups in terms of the distribution of peritoneal adhesion scores (p<0. 05). The rate of visceral adhesion at the mesh, which was considered to be the strongest adhesion, was detected as 85. 7% in the control group in which the polypropylene mesh was used, while this rate was 57. 1% in group A in which only polycaprolactone coated polypropylene mesh was used and it was 28. 6% when oxidized regenerated cellulose was added to this mesh. Visceral adhesion was detected to only one mesh in groups D and E in which polyglactin-coated polypropylene patches containing chitosan were used. The distribution of macroscopic findings according to the groups is shown in Table 3 and the peritoneal adhesion degrees that were evaluated macroscopically are shown in Figure 1. Table 3 The distribution of macroscopic findings according to the groups. a Fisher Freeman Halton test; b Kruskal Wallis test; c Min-Max(Median): minimum-maximum and median scores in each group according to scales. *p<0. 05; **p<0. 01. n(%): number and percentage of rats in each group. Control group Group A Group B Group C Group D Group E p Peritoneal adhesion criteria and scores; n(%) Omentum adhesion at suture zone 0 (0. 0) 0 (0. 0) 0 (0. 0) 1 (14. 3) 2 (28. 6) 0 (0. 0) Omentum adhesion up to 50% 1 (14. 3) 0 (0. 0) 3 (42. 9) 1 (14. 3) 3 (42. 9) 2 (28. 6) Omentum adhesion more than 50% 0 (0. 0) 3 (42. 9) 2 (28. 6) 3 (42. 9) 1 (14. 3) 4 (57. 1) Visceral adhesion at mesh 6 (85. 7) 4 (57. 1) 2 (28. 6) 2 (28. 6) 1 (14. 3) 1 (14. 3) c Min-Max (Median) 3-6 (6) 4-6 (6) 3-6 (4) 2-6 (4) 2-6 (3) 3-6 (4) b 0. 031* Adhesion type; n(%) Type 1 0 (0. 0) 0 (0. 0) 1 (14. 3) 2 (28. 6) 5 (71. 4) 1 (14. 3) Type 2 0 (0. 0) 0 (0. 0) 3 (42. 9) 1 (14. 3) 2 (28. 6) 3 (42. 9) Type 3 3 (42. 9) 3 (42. 9) 2 (28. 6) 3 (42. 9) 0 (0. 0) 3 (42. 9) Type 4 4 (57. 1) 4 (57. 1) 1 (14. 3) 1 (14. 3) 0 (0. 0) 0 (0. 0) c Min-Max (Median) 3-4 (4) 3-4 (4) 1-4 (2) 1-4 (3) 1-2 (1) 1-3 (2) b 0. 001** Incorporation; n(%) 51%-75% of mesh surface 2 (28. 6) 2 (28. 6) 1 (14. 3) 0 (0. 0) 0 (0. 0) 0 (0. 0) a 0. 375 76%-100% of mesh surface 5 (71. 4) 5 (71. 4) 6 (85. 7) 7 (100. 0) 7 (100. 0) 7 (100. 0) Figure 1 Macroscopic evaluation of peritoneal adhesion degrees after sacrification. (a) Omentum adhesion at suture line, (b, c) omentum adhesion up to 50% of the mesh surface, (d) omentum adhesion more than 50% of the mesh surface, (e, f) visceral adhesion at mesh. When the distribution of types of adhesions was examined, a statistically significant difference was found between the groups. While the rate of type 4 adhesions was detected as 57. 1% in the control group and group A rats, this rate was found to be significantly higher than the rats in the B, C, D and E groups (p<0. 05). There was no statistically significant difference between the groups in terms of the incorporation grades of rats (p>0. 05). Fibroblast proliferation was moderately observed in 85. 7% of the rats in the control group which was significantly higher than the A, B and C groups (p <0. 05). However, moderate fibroblast proliferation was detected in all rats in groups D and E. While moderate angiogenesis was observed in 6 of the rats (85. 7%) in the control group, similarly, moderate angiogenesis was detected in all rats in the group D and in 85. 7% of the group E rats. Mild angiogenesis was detected in 5 (71. 4%) of the rats in groups A and B and in all rats in group C. While 85. 7% of the rats in the control group had moderate granulation, mild granulation was observed in most of the rats in groups A, B and C, and moderate granulation was detected in all rats in groups D and E. The distribution of histopathological findings according to the groups is shown in Table 4. Table 4 The distribution of histopathological findings according to the groups. a Fisher Freeman Halton test; b Kruskal Wallis test; c Min-Max(Median): minimum-maximum and median scores in each group according to scales. *p<0. 05; **p<0. 01. n(%): number and percentage of rats in each group. Control group Group A Group B Group C Group D Group E p Inflammation; n(%) Mild 2 (28. 6) 4 (57. 1) 6 (85. 7) 6 (85. 7) 5 (71. 4) 6 (85. 7) a 0. 161 Moderate 5 (71. 4) 3 (42. 9) 1 (14. 3) 1 (14. 3) 2 (28. 6) 1 (14. 3) Fibroblast proliferation; n(%) Mild 1 (14. 3) 5 (71. 4) 6 (85. 7) 5 (71. 4) 0 (0. 0) 0 (0. 0) a 0. 001** Moderate 6 (85. 7) 2 (28. 6) 1 (14. 3) 2 (28. 6) 7 (100. 0) 7 (100. 0) Angiogenesis; n(%) Mild 1 (14. 3) 5 (71. 4) 5 (71. 4) 7 (100. 0) 0 (0. 0) 0 (0. 0) Moderate 6 (85. 7) 2 (28. 6) 2 (28. 6) 0 (0. 0) 7 (100. 0) 6 (85. 7) Diffuse 0 (0. 0) 0 (0. 0) 0 (0. 0) 0 (0. 0) 0 (0. 0) 1 (14. 3) c Min-Max (Median) 1-2 (2) 1-2 (1) 1-2 (1) 1-1 (1) 2-2 (2) 2-3 (2) b 0. 001** Granulation; n(%) Mild 1 (14. 3) 5 (71. 4) 6 (85. 7) 3 (42. 9) 0 (0. 0) 0 (0. 0) a 0. 001** Moderate 6 (85. 7) 2 (28. 6) 1 (14. 3) 4 (57. 1) 7 (100. 0) 7 (100. 0) Fibrosis; n(%) Mild 7 (100. 0) 7 (100. 0) 5 (71. 4) 6 (85. 7) 7 (100. 0) 7 (100. 0) a 0. 407 Moderate 0 (0. 0) 0 (0. 0) 2 (28. 6) 1 (14. 3) 0 (0. 0) 0 (0. 0) Giant Cells; n(%) Mild 5 (71. 4) 1 (14. 3) 7 (100. 0) 7 (100. 0) 0 (0. 0) 2 (28. 6) Moderate 1 (14. 3) 0 (0. 0) 0 (0. 0) 0 (0. 0) 1 (14. 3) 5 (71. 4) Diffuse 1 (14. 3) 6 (85. 7) 0 (0. 0) 0 (0. 0) 6 (85. 7) 0 (0. 0) c Min-Max (Median) 1-3 (1) 1-3 (3) 1-1 (1) 1-1 (1) 2-3 (3) 1-2 (2) b 0. 001** Macrophages; n(%) Mild 4 (57. 1) 7 (100. 0) 6 (85. 7) 7 (100. 0) 7 (100. 0) 7 (100. 0) a 0. 066 Moderate 3 (42. 9) 0 (0. 0) 1 (14. 3) 0 (0. 0) 0 (0. 0) 0 (0. 0) Edema; n(%) None 7 (100. 0) 4 (57. 1) 1 (14. 3) 4 (57. 1) 7 (100. 0) 7 (100. 0) Mild 0 (0. 0) 2 (28. 6) 2 (28. 6) 1 (14. 3) 0 (0. 0) 0 (0. 0) Moderate 0 (0. 0) 1 (14. 3) 4 (57. 1) 1 (14. 3) 0 (0. 0) 0 (0. 0) Diffuse 0 (0. 0) 0 (0. 0) 0 (0. 0) 1 (14. 3) 0 (0. 0) 0 (0. 0) c Min-Max (Median) 0-0 (0) 0-2 (0) 0-2 (2) 0-3 (0) 0-0 (0) 0-0 (0) a 0. 001** There was no statistically significant difference between the groups in terms of the distribution of inflammation, fibrosis and macrophage levels of rats (p>0. 05). Discussion Abdominal adhesions are abnormal fibrotic bands between organ surfaces in the abdomen or between the walls of the abdominal cavity. In patients who have undergone abdominal or gynecological operations, intraabdominal adhesion formation is seen in 95% [ 7 ]. The main causes of adhesion formation are peritoneal trauma, ischemia and foreign bodies. Laparoscopic and minimally invasive techniques have been adopted to reduce the trauma that may occur during surgical intervention. Laparoscopic surgery is associated with less intraabdominal adhesion than classical open surgery, but postoperative adhesion still occurs in 37. 7% of patients operated with laparoscopic methods [ 8 ]. Surgical technique is not sufficient to reduce postoperative adhesions and related complications. In order to prevent adhesions, besides choosing the surgical technique to minimize peritoneal damage, it is necessary to reduce the inflammatory response, provide inhibition of coagulation, stimulate fibrinolysis, and protect the surfaces that may cause adhesions [ 9 ]. Polypropylene mesh, which is the most widely used prosthetic material for hernia repair today, causes intense inflammatory response compared to other meshes and causes adhesion in visceral organs, especially small intestines. It is not suitable for intraperitoneal use due to complications such as fistula and small bowel obstruction. For this reason, composite meshes have been developed for use in intraperitoneal hernia repair. Bilayer composite meshes are preferred to prevent intraabdominal adhesion due to the strength and integration properties of permanent meshes such as polypropylene on the parietal peritoneum side and the intestinal protective property of the anti-adhesive barrier on the visceral side [ 10 ]. There are clinical studies showing that composite meshes are associated with shortened hospital stay, moderate complication rates, and low rates of infection and recurrence of hernia [ 11 ]. Today, the most widely used composite meshes around the world are polyglactin-based ones. Polyglactin is an absorbable and biocompatible polymer. It has proven its reliability and it is widely used in materials developed for tissue engineering and temporary implants such as surgical threads [ 12 ]. Materials such as polyurethane, polytetrafluoroethylene (PTFE), oxidized regenerated cellulose, polyethylene glycol, sodium hyalulose, carboxy methyl cellulose and collagen are used as anti-adhesives on the visceral surfaces of composite meshes [ 13 ]. Among these, oxidized regenerated cellulose used in Proceed® meshes has been reported to be superior to other materials in terms of anti-adhesive effects in experimental studies [ 13 ]. Aramayo et al. did not find a statistically significant difference between the groups in terms of the presence and degree of adhesion in an experimental study on 40 rabbits. A polypropylene mesh was used in one group, a mesh containing polypropylene and absorbable polygleocapron 25 (Ultrapro®) in other group, and a mesh containing polypropylene, polydiaxanone and oxidized regenerated cellulose (Proceed®) in another group [ 14 ]. In our study, no statistically significant difference was found between the control group in which the polypropylene mesh was used and the groups A, B, and C in which polypropylene and polycaprolactone meshes were used, in terms of peritoneal adhesion criteria and scoring. When oxidized regenerated cellulose was added as an adhesion barrier to the polycaprolactone coated polypropylene mesh (group B and C), the decrease in the degree of adhesion was not statistically significant compared to group A in which the mesh without adhesion barrier was used. When oxidized regenerated cellulose was added as an adhesion barrier to the polycaprolactone coated polypropylene mesh, it was evaluated that the adhesion rate of type 4 showing strong adhesion was low and that the oxidized regenerated cellulose caused a decrease in the adhesion strength. Chitin, which is the most common polysaccharide-based biopolymer in the world after cellulose, is the main component of shellfish such as crab and shrimp, and is found in the skeleton of insects and the cell walls of fungi. Different molecules are obtained with the changes made in the structure of the chitin. The most important of these molecules is chitosan. The fact that chitosan has both antibacterial and hemostatic effects has made it attractive for use in the biomedical field [ 15, 16 ]. Due to the fragile structure of chitosan and its limited use due to its inadequate mechanical properties, this problem can be overcome by forming composites together with other polymers. When visceral organs come into contact with the polypropylene mesh, dense adhesions occur. In our study, in the control group in which polypropylene mesh was used, the rate of visceral adhesion to the mesh, which was the highest level of peritoneal adhesion score, was found to be higher than the B and C groups without reaching a statistical significance, but it was found to be significantly higher than the D and E groups. These findings showed that using the adhesion barrier together with the polypropylene mesh instead of using only the polypropylene mesh could reduce the adhesion macroscopically, and that the use of chitosan as the adhesion barrier significantly reduced the adhesion formation. In the experimental study by Altınel et al. in which polypropylene mesh, mesh containing polypropylene and absorbable polygleocapron 25 (Ultrapro®) and mesh created by adding chitosan to polypropylene and absorbable polygleocapron 25 (Ultrapro®) were used on rats; the groups were compared in terms of adhesion score and strength, and no significant difference was found between the adhesion scores of mesh groups and their chitosan-coated forms [ 17 ]. In another study, Paulo et al. observed the effects of chitosan on the formation of adhesion when combined with polypropylene mesh on rats in an experimental study using a polypropylene mesh and chitosan-coated polypropylene mesh, and stated that the amount of adhesion decreased when chitosan was added as an adhesion barrier in the mesh structure compared to the groups in which only the polypropylene mesh was used [ 18 ]. Utiyama et al. performed incisional hernia repair on rats and compared the histopathological characteristics of the groups (the group in which a mesh was not used, the group in which a polypropylene mesh was used and the group in which a mesh containing polypropylene and polygleocapron 25 [Ultrapro®] was used) in terms of inflammatory response. They stated that there was no difference between the groups in terms of fibrosis, macrophage, lymphocyte, neutrophil and giant cell counts, and granuloma [ 19 ]. Similarly, in our study, no statistically significant difference was found between the groups in terms of inflammation levels, distributions of fibrosis levels and macrophage levels. There are some limitations of our study. At first, although the seven-day period was sufficient to evaluate the differences between the groups on adhesion formation, it was not possible to evaluate the long-term anti-adhesion efficacy of the meshes used. Also, considering the effectiveness of bleeding on adhesion formation, we think that when a material with hemostatic properties such as oxidized regenerated cellulose is added to the structure of the mesh, the hemostatic effect can be evaluated with different parameters. The model must be easily applicable and reproducible in surgical research and especially in vivo applications of innovations in biomedicine. Rats are one of the most suitable animals in terms of surface area in adhesion formation models. It is also known that the response to peritoneal trauma is similar to humans. In addition to all these, it is unclear how the degree of adhesions will make a difference in humans thus, the effects of these meshes on humans should be evaluated. Conclusions Composite meshes with anti-adhesive barriers were developed for use in hernia repair and when placed intraperitoneally visceral surface of the mesh acts as a barrier for adhesion formation. This study showed that when adhesion barriers with biocompatible, anti-adhesion properties such as oxidized regenerated cellulose and chitosan are used in the structure of the composite meshes in hernia repair, the development of intra-abdominal adhesion and the strength of adhesion may decrease. Considering that these materials also have hemostatic and antibacterial properties, they can be promising to create the ideal mesh.

10. 7759/cureus. 1952

Cureus

Ultrasound-based Techniques as Alternative Treatments for Chronic Wounds: A Comprehensive Review of Clinical Applications

Ultrasound (US) waves have been recently developed for the treatment of different chronic wounds with promising therapeutic outcomes. However, the clinical efficacy of these techniques is still not fully understood and standard guidelines on dose ranges and possible side effects should be determined. This paper aims to comprehensively review the recent advances in US techniques for chronic wound treatment, their therapeutic efficacies, and clinical considerations and challenges. The databases of PubMed (1985-2017), EMBASE (1985-2017), Web of Sciences (1985-2017), Cochrane central library (1990-2017), and Google Scholar (1980-2017) were searched using the set terms. The obtained results were screened for the title and abstract by two authors and the relevant papers were reviewed for further details. Preclinical and clinical studies have shown strong evidence on the therapeutic efficiency of US in chronic wounds. The main limitation on developing clinical standard protocols of US for treatment of wounds is the lack of definite dose-response for each wound. However, spatial average temporal average is the main parameter for defining US dosage in wound treatment. The range of 0. 5 to 3 W/cm 2 is a range of dose exerting significant therapeutic outcomes and minimum adverse effects. Low-frequency US waves can accelerate the healing speed of open wounds as well as deep-tissue injuries. In addition, US waves show promising therapeutic efficacy for chronic wounds. To develop clinical US protocol for each wound type, further in vitro and in vivo preclinical and clinical trials are needed to reach an exact dose-response for each wound type.

Introduction and background During the recent years, along with the advancements in new medications, different techniques have been developed for the treatment of different chronic wounds such as pressure relieving beds, and medicinal plants [ 1 ]. However, high worldwide prevalence of wounds, high costs and side effects of conventional medications have necessitated the development of alternative or adjunctive techniques for wound treatment. In this regard, several methods have been developed for the treatment of different acute and chronic wounds including laser [ 2 ], direct current, electric and magnetic fields [ 3, 4 ], light and electromagnetic fields [ 5 ]. Ultrasound (US) waves have been recently proposed for the treatment of different wounds and showed promising outcomes. These mechanical waves have several intrinsic advantages over other non-medication techniques that make them a good candidate for wound healing. The capability of deep penetration to reach deep-seated tissues, being highly orienting and focusing, and low scattering are some of these advantages [ 6 ]. Different US-based techniques have been developed enjoying these features for the treatment of different disorders including skin wounds, musculoskeletal disorders, malignant tumors, and bone fractures [ 7, 8 ]. US waves have reportedly shown promising outcomes for soft tissue injuries than other disorders [ 7 ]. Several preclinical and animal studies have shown different physiological effects of US on living tissues [ 9 ]. In this regards, high-frequency US waves were used in tendon injuries treatment and short-term pain relieving [ 10 ], fresh fracture healing [ 11 ], venous and pressure ulcers, and surgical incisions [ 12, 13 ]. However, some studies have reported possible side effects of US waves under inappropriate parameters that can cause burns or damage the endothelial tissues [ 1 ]. In line with the studies on these fields and promising outcomes, different commercial US-based modalities have been developed. Most of these devices work in low frequencies as the use of high-frequency US in clinical medicine is restricted due to the risk of tissue heating. Low-frequency US waves are actually a slow release technique associated with low tissue heating so that these techniques may become the standard technique for delayed healing wounds, skin ulcers, and nonunion fractures. Surface acoustic wave (SAW) patch therapy is another US technique developed for wound treatment. It employs a different acoustic wave than traditional ultrasound, utilizing a scattered beam with a maximum penetration of 4 cm, while traditional ultrasound can penetrate 10 cm. Some studies have reported that the application of SAW patch therapy increases tissue oxygenation and saturation, which consequently facilitates the wound healing [ 14, 15 ]. US waves have emerged as a promising alternative or adjunctive strategy for chronic wounds. However, the clinical efficacy of these techniques for different chronic wounds is still not fully understood. In addition, the clinical guidelines on the allowed doses and possible side effects of these techniques should be determined. To address this issue, the present study was aimed to review the therapeutic effectiveness of US-based techniques for treatment of chronic wounds and the clinical challenges for the development of these techniques as routine approaches for wound treatment are discussed. Review Method The databases of PubMed (1985-2017), EMBASE (1985-2017), Web of Sciences (1985-2017), and Google Scholar (1980-2017) were searched using the set terms. The title and abstract of the obtained records were reviewed by two authors and they came to the consensus whether the studies are related to the review. Animal and human studies in both in vivo and in vitro designs that evaluate the therapeutic effects of US waves in chronic wounds were included for further review. Because of the immense body of literature in this field, various protocols and devices used in different wound types, this study was aimed to provide a comprehensive and descriptive overview of the recent advances in applications of US waves for the treatment of chronic wounds, therapeutic efficacies, and clinical considerations of US-based techniques for chronic wounds. Search strategy Scientific records were retrieved using a systematic searching of multiple bibliographic databases. The last update of the search was performed on May 30th, 2017 including PubMed (1985-2017), EMBASE (1985-2017), Web of Sciences (1985-2017), and Google Scholar (1980-2017). The language of search was limited to English. The search key words based on the MeSH heading included "ultrasound wave" OR "ultrasound" AND "chronic wound" AND "treatment" OR "clinical considerations" OR "therapeutic efficacy" OR "dose response". The titles and abstracts of all the records retrieved by the search strategy were carefully reviewed by at least two authors and the relevant records with full texts available were used for further assessments. In addition to the records identified in the systematic search of databases, the reference lists of the relevant papers were assessed manually to identify studies appropriate for the full reviews and the eligible studies were also included in the full review stage. Inclusion and exclusion criteria The identifying, screening, and eligibility stages of studies for inclusion or exclusion were performed independently by the three authors and disagreements were resolved by discussion. Those studies that presented an original research and the main criterion for eligibility of a study were human or animal studies in vivo or in vitro assessing therapeutic outcomes or a biological effect of US waves in any types on any kind of chronic wounds. The exclusion criteria were: (1) abstract only, (2) books, (3) letters, (4) conference papers, (5) editorials or (6) guidelines assessments (Figure 1 ). Figure 1 Flowchart of the study procedure. Results A total of 121 studies were retrieved from the searching process. After screening the abstracts and titles, 65 records remained. In the screening stage, 15 records were removed as being abstract only (5), book (3), editorial (2), and conference records (5). In the eligibility stage, 12 records were removed and 38 papers were included for full-text review. Two more papers were also added from the reference lists screening, and totally, 40 papers were included in the final review. Physical characteristics of therapeutic US The US waves delivered to the body and soft tissues undergo diffusion and through vibrating, the molecules progressively lose their energy as the waves pass through the tissue. The main phenomena causing the US waves attenuation are absorption, scattering or dispersion, reflection, and rarefaction [ 16 ]. Power expressed in Watts is the main parameter to assess the therapeutic outcomes of US-based techniques. The amount of energy transferred to a target tissue is determined by two main groups of parameters: the US waves' characteristics (frequency, intensity, amplitude, focus, and beam uniformity) and the type and physical characteristics of the target tissue as well as tissues the US waves pass them. The therapeutic US waves have a frequency range of 0. 75–3 MHz and most of the US devices are set at two frequencies: 1 or 3 MHz. Increasing the frequency decreases the penetration depth; however, low-frequency US waves are less focused. One-MHz US waves are adsorbed mainly by tissues located in depth of 3–5 cm making them an appropriate option for deeper injuries and in patients with higher content of subcutaneous fat. In contrast, 3 MHz US waves are appropriate for more superficial lesions at depths of l–2 cm [ 17 ]. Regarding medical applications, acoustic impedance is the main parameter describing a tissue, which is defined as the product of the tissue's density and the US wave's speed within the tissue. Tissues that are rich in fat have low US absorption and thus high penetration of US waves, whereas tissues with high protein content such as skeletal muscle have higher US adsorption and low penetration. The acoustic impedance difference between two tissues determines the percent of reflection at the interface of the two tissues where higher differences mean lower percents of transmission through the interface [ 16 ]. When reflected US meets further transmitted waves, a standing wave is generated that may impose adverse effects on tissue. These adverse effects can be minimized through producing a uniform wave, using pulsed waves, and moving the transducer during the treatment intervention. Transducers with greater diameter produce more focused US beam. Energy distribution within the US beam is not uniform and the greatest non-uniformity is formed near the transducer surface. The variation of the beam intensity is determined by the beam non-uniformity ratio (BNR), defined as the ratio of the maximum intensity of the transducer to the averaged intensity across its surface. Mechanisms of action In vitro studies have shown the therapeutic outcomes of US waves on tunneling or debilitation wounds are mainly through killing multi-drug resistant bacteria such as vancomycin-resistant Enterococcus and resistant Pseudomonas aeruginosa [ 18 ]. Several in vitro studies have shown that US waves improve cell proliferation, collagen production, bone formation, and angiogenesis [ 17 ]. One of the proposed mechanisms of action of US waves in wound treatment is reducing the wound-related pains. Pain associated with chronic wounds is always a challenging clinical issue with no definitive solution. Different studies have demonstrated the therapeutic efficacy of low-frequency US waves in chronic wounds, not only for curing but also for pain relieving, decreasing pigmentation and odor [ 19 ]. Clinical evidence has shown that the US intervention reduces wound-associated pain in the patients with painful chronic lower-extremity wounds [ 20 ]. However, there are controversial findings on the clinical outcomes of US waves. For instance, a systematic review of the efficacy of different US modalities on chronic wound treatment concluded inadequate evidence for clinical efficacy of therapeutic US in chronic wounds [ 1 ]. Clinical considerations Pressure Ulcers Several studies have been conducted to investigate the efficacy of different US waves in the treatment of pressure ulcers. In general, the evidence for the effectiveness of US waves for pressure ulcers is limited. Some randomized controlled trials conducted on US treatment in pressure ulcers showed no significant differences between the treatment groups [ 21, 22 ]. Similarly, Reddy, et al. in a randomized controlled clinical trial found no significant therapeutic efficacy for the US waves in the treatment of pressure ulcers. Flemming, et al. in a systematic review found no vigorous evidence on the therapeutic efficacy of US in the pressure ulcers. However, the main limitations of their study were inconsistency in the research methodology and physical parameters of the reviewed studies and the small sample size of the reviewed studies [ 23 ]. Akbari Sari, et al. reviewed the effectiveness of US therapy on pressure ulcers. They showed no reliable proof of advantage of treatment by US in the healing of pressure ulcers. However, their review suffered the heterogeneities in the research methods and the small sample size of the reviewed studies [ 23 ]. Therefore, to reach a definitive conclusion on the therapeutic efficacy and clinical value of US waves in the treatment of pressure ulcers further studies should be conducted. Combined US - Traditional Techniques Low-frequency US techniques have been used in combination with standard wound care medications for treatment of purulent wounds. The findings of these studies showed the therapeutic effectiveness of US technique as an adjunctive or alternative treatment for purulent wound. A case series study (n = 17) showed the effectiveness of the combination of low-frequency US together with gentamicin solution so that the purulent septic complications were reduced from 35. 7% to 5. 9% [ 24 ]. Several studies have investigated the efficacy of low-frequency US in combination with antibiotic agents in different chronic wounds or bacterial cultures. Rediske, et al. demonstrated that continuous US waves and systemic gentamicin administration significantly decreased the viable bacteria concentration in the simulated implant putridity [ 25 ]. Other studies have reported that application of US in the bacterial cultures of E. coli and P. aeruginosa increased the efficacy of antibacterial action of gentamicin [ 26, 27 ]. Diabetic Wounds A cross-sectional experiment comparing the therapeutic outcomes of low-frequency US and laser irradiation in patients with diabetes mellitus and purulent surgical wound (n = 112) reported higher-effectiveness of the US treatment in the first and second phases of wound healing process [ 28 ]. Swist-Chmielewska, et al. compared the efficacy of US waves at two power densities of 0. 5W/cm 2 and 1 W/cm 2 for the treatment of Venous crural ulceration [ 29 ]. They reported US waves speed up the ulceration healing process and US waves at 0. 5 W/cm 2 showed greater outcomes. However, they reported no significant difference in terms of granulation development rate and debridement of the wound between the two densities. Gottrup and Apelqvist carried out a review of the available literature on new methods for the treatment of diabetic foot ulcers [ 30 ]. They evaluated the therapeutic efficacies of several wound healing techniques including antimicrobial agents, dressings, topical negative pressure, hyperbaric oxygen treatment, electrical, electromagnetic, laser, shockwave, and US techniques, growth, and cell biology modulating factors, tissue engineering, bioengineered skin and skin grafts, and adjuvant therapies. Their review demonstrated a restricted proof on the level I evidence to recommend these techniques as usual clinical methods. However, some of the US-based techniques can be used as alternative or adjunctive treatment for some types of chronic wounds. The main reasons for the lack of strong evidence are insufficient sample size, short follow-up period, non-random allocation to treatment arms, non-blinded outcomes evaluation, poor description of control, and concurrent interventions. Therefore, it is necessary to enhance the quality and methodology of clinical trials [ 30 ]. Extremity Lower Wounds Extremity lower wounds are the most prevalent wounds worldwide and most of the US-based techniques have been proposed for this type of wounds. Johannsen, et al. in a meta-analysis on the efficacy of US waves in the treatment of chronic leg ulcer concluded that low doses of US administered around the ulcer edge exert the greatest therapeutic effects [ 31 ]. Callam, et al. compared the therapeutic outcomes of standard wound care with a pulsed US intervention for a 12-week period for chronic leg wounds. They reported that the ratio of wound closure area was 20% greater in the US intervention group [ 32 ]. Lundeberg, et al. in a randomized controlled trial (n = 44) investigated the outcomes of a combined treatment of pulsed US and a standard wound healing technique for chronic leg ulcers [ 33 ]. They compared the outcomes of a standard treatment (paste impregnated bandage and a self-adhesive elastic bandage) with a placebo US and with real pulsed US intervention applied three times a week for four consecutive weeks, followed by two times a week for other four-week period and then once a week for next four weeks. The wound healing rates were assessed after four, eight and 12 weeks [ 33 ]. They observed no significant differences in the percentage of cured ulcers in the pulsed US treatment as compared with the placebo group [ 33 ]. Eriksson, et al. in a randomized controlled trial compared the therapeutic efficacy of a US treatment consisting of 1. 0 W/cm 2 at 1 MHz, for 10 min twice a week for eight weeks in a standard model of chronic leg ulcers. They observed no significant differences between the real and placebo treatments in the percentage wound closure area and the number of cured wounds examined at two, four, six, and eight weeks after the start of treatment [ 34 ]. Peschen, et al. investigated the outcomes of low-frequency (30 kHz) low-intensity US on the chronic venous leg ulcers combined with a conventional outpatients’ therapy. Patients were randomly divided into two groups: conventional treatment with topical application of hydrocolloid dressings and combined US-conventional treatment (compression therapy). The US therapy consisted of a 10-min of foot-bathing with continuous US wave at 100 mW/cm 2 density three times a week for three months. The ulcer area was measured before intervention and after intervention at two, four, six, eight, 10, and 12 weeks post-intervention. In addition, the radius of ulcer was measured daily. After each session, adverse effects were evaluated. The results showed the mean decrease of ulcer area in the US group was 55. 4% compared to 16. 5% for the control group. In addition, daily decrease of ulcer size in the US-treated group was 0. 08 mm compared to 0. 03 mm for the control. Both US and control groups indicated minor adverse effects. The findings of this study confirmed the therapeutic efficacy of the low-frequency low-dose US technique in chronic venous leg ulcers [ 35 ]. The American Society of Plastic Surgeons evaluated the efficacy and feasibility of US treatment for leg and foot ulceration [ 36 ]. Their assessments, which were based on the clinical experiment guidelines on chronic wounds of the leg and foot ulcer, did not note the application of US as a choice of treatment [ 36 ]. Kavros, et al. in a retrospective analysis evaluated the clinical efficacy of MIST-US technique for chronic leg and foot ulcer. They reported that the efficacy of wound healing of a standard wound care in combination with MIST-US technique was significantly higher than the standard wound care alone. In addition, application of the MIST-US therapy accelerated the wound healing rate compared with the standard wound care [ 14 ]. Cullum, et al. conducted a Cochrane review on the efficacy of US on the rate of venous leg ulcer healing. They concluded that the conducted studies on US treatment in venous leg ulcers suffer small clinical evidence with low sample size, poor-quality, and heterogeneous. Their review concluded no significant evidence of the efficacy of US for venous leg ulcers healing. There was a number of weak evidence which showed enhanced therapeutic efficacy of US; however, to reach a more conclusive answer, further high-quality large sample size randomized controlled trials are needed [ 20 ]. Low-Frequency Non-Contact US One of the new US-based techniques for the wound treatment with promising outcomes is non-contact low-frequency US (NLFU) techniques which is also called MIST-US treatment. This technique has been approved by the United States Food and Drug Administration for wound treatment [ 37 ]. Several case-series, preclinical, and randomized controlled trials have been conducted on the efficacy of NLFU in the treatment of different chronic wounds including burns, digital ulcers, infected surgical wounds, and sacral pressure ulcers [ 38 - 44 ]. In a randomized, double-blinded, sham-controlled, multi-center study, Ennis, et al. compared the therapeutic efficacy of active and sham NLFU (40 kHz) US technique using a MIST-US device in recalcitrant diabetic foot wounds therapy (n = 55). The active US treatment showed a significant increased wounds healing ratio compared to the placebo group. Furthermore, the frequency and type of the reported side effects did not differ between the two groups [ 45 ]. Kavros, et al. investigated the effects of NLFU therapy on ischemic wounds and reported significant improvements in wound healing period in patients with critical limb ischemia following the administration of combined standard wound care and NLFU [ 46 ]. The treatment protocol consisted of daily five-minute treatment for three times a week for consecutive three months or until wounds reached a full recovery. The main outcome of remission was defined a more than 50% wound area reduction after three-month treatment period. The percentage of cured patients in the combined standard wound care and MIST-US group (63%) was significantly higher than the standard wound care group (29%). This study also proposed that the baseline transcutaneous oxygen pressure is a parameter that can predict the outcome of US waves on wound healing [ 46 ]. Kavros and Schenck investigated the efficacy of NLFU treatment in chronic, intractable lower-leg and foot ulcer (n = 51) using a non-randomized, baseline-controlled clinical series [ 15 ]. They compared the efficacies of a standard wound treatment and low-frequency US therapy alone and in combination. The patients had leg and foot ulcer different etiologies including diabetes mellitus, neuropathy, limb ischemia, chronic renal insufficiency, venous illness, and inflammatory connective tissue disease. The average wound healing period for the baseline standard wound care, control group, was 9. 8 weeks, whereas NLFU group showed healing period of 5. 5 ± 2. 8 weeks. They concluded NLFU treatment can significantly improve the wound healing in recalcitrant leg and foot ulcer [ 15 ]. Ennis, et al. investigated the efficacy of MIST-US on the wound closure in chronic non-healing lower extremity wounds with different etiologies. They showed appropriate and optimal treatment duration can enhance the wound healing process and the improvements were clinically significant [ 47 ]. They also investigated the effect of MIST-US on the micro circulatory flow patterns within the wound bed. The standard treatment period was two weeks and 69% of the wound was healed by applying the desired therapeutic model. When MIST US was applied alone, the average wound healing period was reduced to seven weeks, compared with the 10-week healing period in the control group. They concluded that using MIST-US treatment alone or in combination with moist wound care could completely heal 69% of chronic wounds [ 47 ]. Conclusions The preclinical in vitro and in vivo studies along with clinical studies show that US waves in specific frequencies, mainly low-frequency range, can shorten the healing period of open wound. In addition, these waves can be clinically effective for early treatment of deep-tissue injuries. Although early studies have been relatively promising, the main challenge for developing US-based techniques as standard treatment options for different wound is defining an exact dose-response for each wound. One of the main steps to define the dose response for US applications in wound treatment is defining the exact mechanisms of action as a function of main physical parameters of US waves as well as biological parameters of the target wounds. In this regard, conducting further controlled trials with big sample size is necessary to reach this goal.

10. 7759/cureus. 2087

Cureus

Pathologic Remodeling of Endoneurial Tubules in Human Neuromas

Background: Laminins are extracellular matrix proteins that participate in endoneurial tubule formation and are important in the regeneration of nerves after injury. They act as scaffolds to guide nerves to distal targets and play a key role in neurite outgrowth. Because there is evidence that laminin architecture affects nerve regeneration, we evaluated endoneurial tubules by examining the laminin structure in clinical samples from patients with nerve injuries. Methods: In a retrospective review of eight nerve injury cases, we evaluated nerve histology in relation to clinical history and injury type. The immunohistochemical delineation of the laminin structure in relationship with the neuroma type was performed. Results: Five cases of upper-trunk stretch injuries—four from childbirth injury and one from a motorcycle accident—and three cases of nerve laceration leading to neuroma formation were examined. In the upper-trunk stretch injuries, avulsed nerves demonstrated no neuroma formation with a linear laminin architecture and a regular Schwann cell arrangement, but increased fibrous tissue deposition. For neuromas-in-continuity after a stretch injury, laminin immunohistochemistry demonstrated a double-lumen laminin tubule, with encapsulation of the Schwann cells and axonal processes. Nerve laceration leading to stump neuroma formation had a similar double-lumen laminin tubule, but less severe fibrosis. Conclusions: In nerve injuries with regenerative capacity, endoneurial tubules become pathologically disorganized. A double-lumen endoneurial tubule of unclear significance develops. The consistency of this pattern potentially suggests a reproducible pathophysiologic process. Further exploration of this pathophysiologic healing may provide insight into the failure of programmed peripheral nerve regeneration after injury.

Introduction Traumatic peripheral nerve injury (PNI) occurs in a wide array of situations, including laceration, concussion, stretch, rupture, or avulsion of peripheral nerves. PNI occurs in 2. 8% of all trauma patients, with an incidence of 13–23:100, 000 persons/year in developed countries, which represents a notably higher incidence than spinal cord injury [ 1 - 2 ]. Various mechanistic patterns of injury, levels of injury severity, and patient-specific factors (e. g. , age, associated secondary injuries) can make PNI a heterogeneous disease process. A hallmark of severe injury is the development of a neuroma, which occurs with ineffective nerve fiber regeneration to its target tissue. The term “neuroma” was first coined by Odier of Geneva in 1811 to describe deep lesions of nerves but did not distinguish nerve tumors from lesions with other mechanisms [ 3 ]. Current thought suggests that neuromas are the result of sprouting axons that exit from disrupted perineurium to form a fibrous, disorganized mass of fibroblasts and macrophages [ 4 ]. A related aspect, the neuroma-in-continuity, as described by Sunderland [ 5 ], describes a partial nerve injury, thickened tissue, Wallerian degeneration of fascicles, and the growth of nerve tissue out of endoneurial growth tubes but within the nerve epineurium. Extracellular matrix (ECM) scaffold proteins are postulated to play a role in nerve regeneration by organizing endoneurial tubules, macrophages, and Schwann cells [ 6 ]. We suspected that the pathophysiology of a neuroma-in-continuity is more than fibrous tissue that prevents neurite outgrowth because we recognize that there is fibrosis from surgical manipulation after every nerve grafting surgery. To evaluate our hypothesis that the endoneurial tubule may reveal aspects of the failure of regeneration, we assayed clinical samples from various neuroma-in-continuity and stump neuroma pathologic specimens. One of the important ECM proteins involved in the promotion of neurite outgrowth is laminin. Thus, we suspected that laminin may have the most to reveal about neuromas. Materials and methods After receiving institutional review board approval with a waiver of informed consent, we performed a retrospective chart and pathological review to identify patients of the senior author (MM) who underwent surgical neurolysis and resection of neuromas as part of surgical repair. The pathological review was performed to ensure tissue was available. Histology was performed at the discretion of the senior author and neuropathologist (CP); it included hematoxylin and eosin (H&E), trichrome, and laminin immunohistochemistry (IHC) stains. Five-micrometer, formalin-fixed, paraffin-embedded sections of neuroma resections from all patients were cut at regular intervals and mounted on glass sides. H&E and trichrome stains were prepared on all specimens except for Cases Two and Four. Immunohistochemical staining was performed using laminin antibodies (Leica Biosystems, Wetzlar‎, Germany) at a dilution of 1:100. Staining was performed using the avidin-biotinylated peroxidase complex (ABC) method on a Ventana Staining system (Ventana Medical Systems, Inc. , Arizona, United States) and counterstained with hematoxylin. The means (ranges) of patient ages and follow-up, along with the mechanism of injury and descriptive histopathological findings, were analyzed. Statistical analysis was not performed with this limited sample. Results The summary of all eight patients is shown in Table 1. Case descriptions are presented in the Supplemental Material. Five patients, four infants, and one young adult, with a mean age of 5. 8±12. 1 years (95% confidence interval (CI) 0. 3, 27. 4 years; median 0. 5 years) had stretch-related/avulsion injuries; and three patients with a mean age of 40. 1±14. 4 years (CI 26. 9, 55. 5 years) had transection injuries. Stretch-related injuries occurred in two males and three females, mostly infants. Transection injuries occurred in two males and one female, all adults. At surgery, there were five neuroma-in-continuity injuries, four from birth injuries and one stab wound; two stump neuromas, both iatrogenic; and one spinal nerve avulsion injury, with discontinuity of the peripheral nerve from the spinal cord. Table 1 Summary of peripheral nerve injury cases and staining patterns Case Age (yr) Sex Mechanism Injury pattern classification Neuroma pattern Staining H&E Trichrome Laminin 1 0. 5 M Birth brachial plexopathy Stretch-related Neuroma-in-continuity Whorl-like perineurium Fibrotic interneuron areas Aberrant double-lumen 2 0. 5 F Birth brachial plexopathy Stretch-related Neuroma-in-continuity Whorl-like perineurium Aberrant double-lumen 3 0. 3 F Birth brachial plexopathy Stretch-related Neuroma-in-continuity Whorl-like perineurium Fibrotic interneuron areas Aberrant double-lumen 4 0. 5 F Birth brachial plexopathy Stretch-related Neuroma-in-continuity Whorl-like perineurium Aberrant double-lumen 5 27. 4 M Trauma to brachial plexus from motorcycle Avulsion No neuroma Normal pattern, wavy Fibrotic perineurial scar Wavy linear fibers 6 26. 9 F Orthopedic resection of common peroneal nerve Transection Stump neuroma Whorl-like perineurium Fibrotic perineurial scar Aberrant double-lumen 7 37. 9 M Iatrogenic femoral nerve injury from inguinal herniorrhaphy Transection Stump neuroma Granular-like perineurium Fibrotic perineurial scar Aberrant double-lumen 8 55. 5 M Work-related median nerve laceration Transection Neuroma-in-continuity Granular-like perineurium Fibrotic perineurial scar Aberrant double-lumen The four infants with stretch-related injuries demonstrated classic neuroma formation on histology, as did the three patients with laceration/iatrogenic injury. The single patient with the nerve root avulsion injury did not demonstrate neuroma formation. Stretch neuroma-in-continuity On H&E stains, neuromas from infants with stretch-related injuries showed a monotonous cellularity and whorl-like perineurium along extensive interfascicular fibrotic scars (Figure 1, A, D, F, I). Trichrome stain showed perineurial fibrotic scar and disorganized nonlaminar architecture (Figure 1, B, G). Laminin IHC showed aberrant ECM formation, with uneven and abnormal double-lumen endoneurial tubules (Figure 1, C, E, H, J). Figure 1 Histology of pediatric brachial plexus stretch neuromas-in-continuity (A, D, F, I) H&E, (B, G) trichrome, and (C, E, H, J) laminin staining of resected neuromas-in-continuity from infants with birth stretch-related injury is shown. Monotonous cellularity and whorl-like endoneurium (asterisk) and extensive intraneural fibrotic scar are seen on H&E and trichrome stains. Laminin staining showing uneven, abnormal double-lumen endoneurial tubules (arrows) reflective of severely aberrant regeneration. Case One = A, B, C, ×100); Case Two = D, E, ×200; Case Three = F, G, H, ×200; Case Four = I, J, ×100. Stretch avulsion injury Nerve tissue from the trunks of a brachial plexus after avulsion from the spinal cord showed parallel, wavy fibers with minimal fibrotic scar on H&E (Figure 2, A-C). Nerve fibers showed evidence of stretch, with relative straightening of the nerve fibers and wide fiber spacing. Laminin stain reflected the H&E showing laminar, parallel extracellular protein deposition with minimal disruption (Figure 2A ). Figure 2 Histology of brachial plexus avulsion injury (A) H&E, (B) trichrome, and (C) laminin staining (all ×200) of the proximal end of the lower trunk from an adult with avulsion is shown (Case Five). The lower trunk was trimmed prior to transfer to the C7 spinal nerve. Parallel, wavy fibers (black arrow, A) with minimal fibrotic scar (white arrow, B) and laminar laminin (black arrow, C) with minimal disruption is seen. Minimal regeneration is seen along with limited fibrosis. Laceration neuromas Neuromas from the transection injuries were similar in microscopic appearance to the stretch-related injuries. H&E stains showed well-encapsulated perineurial tissue, heterogeneous nuclei, with reduced interneural fibrotic scars (Figure 3, A, D, G). Trichrome reflected the presence of fibrotic intraneural tissue (Figure 3, B, E, H). Laminin IHC showed separated groupings of double-lumen endoneurial tubules in all three cases (Figure 3, C, F, I). Figure 3 Histology of nerve laceration injuries (A, D, G) H&E, (B, E, H) trichrome, and (C, F, I) laminin staining of neuromas from adult patients with transections is shown. Well-encapsulated endoneurial tissue, heterogeneous nuclei, with reduced interneural fibrotic scars (black arrows, A, D, G) as compared with samples from neonatal stretch injuries. Double-lumen endoneurial tubules are seen (black arrows, C, F, I) are similar to those from birth stretch injuries. Case Six = A, B, C, ×200; Case Seven = D, E, F, ×400; Case Eight = G, H, I, ×200. Discussion The ECM, Schwann cells, and signal transduction from trophic factors play an important role in axon regeneration and functional recovery [ 7 ]. These three factors have been suggested to work in concert in promoting successful regeneration; however, the tissue architecture mediating this process is not clearly understood. Upon losing contact with an axon because of Wallerian degeneration, Schwann cells revert to an immature, proregenerative state [ 8 ]. When transformed to this state, they proliferate and signal macrophage infiltration to clear debris. Schwann cells reorganize into columns along the ECM, termed the Bands of Büngner, to guide neuronal axon growth [ 9 ]. In the case of a neuroma-in-continuity, a distinct pathophysiology likely occurs to prevent the choreography of neurite, Schwann cell, and ECM from successful regeneration. We hypothesized that the endoneurial tubule may play a role in the failure of regeneration, and we selected laminin as a potential marker of the particular failure within the endoneurial tubule. Laminins play an important role in axonal guidance. Laminin has been used as a substrate for nerve conduits to guide axonal growth in a variety of settings [ 10 - 11 ] and has been used during tissue engineering approaches [ 12 - 13 ]. Laminin specifically induces signaling pathways, including PI-3-kinase [ 14 ]. Within Schwann cells, an early pro-myelinating pathway is driven by neuregulin 1 while a later anti-myelinating pathway is driven by laminin [ 14 ]. Schwann cells lacking laminin induction were shown to express decreased schwannomin (ser518) phosphorylation, as well as CDC42 and Rac1 activation [ 15 ]. Furthermore, decreased levels of these proteins reduced Schwann cell-dependent myelination. Laminins were able to enhance the phosphorylation of IκB and p65 NF-κB signaling proteins in schwannoma cells [ 16 ]. A study by Chen and Strickland [ 17 ] demonstrated the importance of laminins using a Cre-loxP system to disrupt laminin γ1 in Schwann cells. This model showed the motor deficits of laminin resulted in hind leg paralysis, tremor, and Schwann cell inability to differentiate and synthesize myelin proteins as well as Schwann cell apoptosis. In addition, after a sciatic nerve crush, axons showed significantly impaired regeneration. These results suggest that laminin plays a key role in organizing neurite outgrowth as well as proper signaling to regulate regeneration. Our pathologic specimens demonstrated a consistent double lumen of a laminin-encasing axon and Schwann cells (Figure 4 ) in all forms of neuromas. The origins and consequences of this histopathologic observation are unknown. One possible cause may be the pathophysiologic remodeling of damaged ECM by regenerating axons and Schwann cells. Alternatively, the invasion of inflammatory cells after trauma may also participate in the pathology of extracellular proteins. Figure 4 Double-lumen laminin pattern compared with typical laminin pattern (anti-laminin IHC and hematoxylin) (A, B) neuroma-in-continuity specimen presented with cross-sectional (A) and longitudinal (B) slicing of the endoneurial tubule. The outer lumen (arrowhead) typically has thinner staining and an eccentric nucleus outside the laminin. The inner lumen (arrow) laminin is irregular, appears to have multiple channels, and may contain one or more nuclei. (C, D) normal nerve specimen removed for non-pathologic indications presented with cross-sectional (C) and longitudinal (D) slicing of the endoneurial tubule. Laminin staining is around a single channel, no more than one internal nucleus (C) and is thin around Schwann cells (D) (all ×400). IHC: immunohistochemistry Two interesting features are notable. First, the pattern of a double-lumen tubule was consistent, regardless of whether the lesion was a neuroma-in-continuity or a stump neuroma and of the age of the patient, suggesting a conserved or consistent response to severe injury. Second, injured or regenerating axons seem to be required. As shown in the histology of the avulsed spinal nerves, where motor neurons are not present and the distal sensory axons are presumably intact, the architecture of the laminin tubules was unremarkable. Because Schwann cells produce the ECM and regenerating axons appear to be essential for the formation of the double-lumen tubule, it would seem to be a shared process. It is possible that loss of endoneurial tubule integrity produces axon–Schwann cell-mediated remodeling, as has been shown when large somatic fibers remodel the smaller endoneurial tubules of autonomic nerves [ 18 - 19 ]. However, much further work is necessary to identify the mechanism of this laminin deposition. There are several limitations of this study. The sample is currently a small sample of overall nerve injury patterns. There was some heterogeneity of IHC staining among samples. In addition, we have only one sample from an avulsion injury, which served as a key comparison. Conclusions Our study used laminin antibodies to assess pathophysiologic regeneration in neuromas. Laminin IHC showed disorganized double-lumen endoneurial tubules in pathologic specimens of neuroma-in-continuity and stump neuromas, whereas avulsion injury maintained good nerve architecture and a relatively normal laminin pattern. This observation of the pathologic remodeling of endoneurial tubules during neuroma formation suggests pair interaction between regenerating axons and Schwann cells in response to injury to the ECM. Better understanding these patterns may help to generate better-directed treatment approaches.

10. 7759/cureus. 23111

Cureus

Clinical Application of Stem Cell Therapy in Reconstructing Maxillary Cleft Alveolar Bone Defects: A Systematic Review of Randomized Clinical Trials

An alveolar cleft is the most common congenital bone defect. This systematic review aimed to investigate the use of stem cells for alveolar cleft repair and summarize the outcomes of clinical research studies. The electronic databases PubMed, Scopus, Web of Sciences, and Google Scholar were utilized to search the literature for relevant studies after administering specific inclusion and exclusion criteria. The search included articles that were published from 2011 to 2021 and specific keywords were used in the databases. The search was completed by two independent reviewers following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Only four studies satisfied both the inclusion and exclusion criteria and were included in this systematic review. These studies investigated different aspects of bone reconstruction in the maxillary alveolar bone by stem cells, including cell types, clinical applications, biomaterial scaffolds, and follow-up period. The accumulated evidence in this systematic review is limited and insufficient to support the role of stem cell use in bone regeneration of maxillary alveolar bone defects. The outcome of using stem cells was studied only in 57 subjects from the four included studies. Although the noninvasive methods of isolating stem cells make them attractive resources for bone regeneration, more research is required in order to standardize and investigate stem cell therapy. This should be done beforehand in adults in less invasive procedures such as bone defect repair in dentistry prior to considering this type of therapy in this vulnerable patient population.

Introduction and background Alveolar cleft reconstruction was first reported in 1901 by Von Eiselberg followed by Lexer in 1908 and Dratcher in 1914 with successful bone grafting attempts in cleft patients [ 1 ]. Since then, surgeons have been trying to achieve the best reconstruction outcomes by harvesting and implanting autologous bone in the cleft site at different time points. Primary alveolar bone grafting is performed at an early stage following lip repair [ 2 ]. Secondary alveolar bone grafting is usually performed during the mixed dentition before lateral incisor eruption in order to provide bony support for its eruption and stabilization of the maxilla [ 3 ]. Cleft patients start their therapeutic journey early in life with multiple maxillofacial reconstruction procedures [ 4 ]. Repairing the cleft bone defect requires a bone graft to fill the defect and regenerate the missing bone [ 4 ]. Autogenous bone is still considered the preferred graft for alveolar bone reconstruction and the most commonly used one [ 4 - 5 ]. This is due to an abundance of autogenous cells and signaling molecules that encourage healing and induce regeneration in implanted defect sites. However, the limited amount of available bone in pediatric patients and the invasive harvesting procedure have an additional negative impact on cleft patients, with the increased morbidity of having infections, paraesthesia, and scarring [ 6 ]. As an alternative, tissue-engineering strategies provide options that can overcome the aforementioned drawbacks by using customized bio-artificial grafts to fill the defect site and regenerate the missing or damaged tissues. Tissue engineering materials that have been used to replace autogenous bone include demineralized bone matrix (DBM), deproteinized bovine bone (DBB), synthetic polymers, and recombinant human bone morphogenetic protein (rhBMP) [ 7 - 8 ]. Then, cells with great growth potential such as stem cells, bioactive molecules, or growth factors can be added to activate the implanted grafts [ 9 ]. In cleft alveolus defect studies, stem cells derived from bone marrow, umbilical cord, dental pulp, and human exfoliated deciduous teeth have been isolated and inspected in terms of tissue regeneration [ 10 ]. The latter is considered a promising source as cells were harmlessly isolated from naturally exfoliated deciduous teeth pulp [ 11 ]. Although numerous experimental studies investigated the use of stem cells in regenerating bone defects in animal models and human clinical trials, few available studies examined the role and potential application of different types of stem cells to repair maxillary alveolar bone defects [ 12 - 15 ]. Only one systematic review was conducted in 2018 [ 16 ], they discussed the use of stem cells in bone regeneration and concluded that stem cells were effective in bone tissue repair and regeneration for clinical application in different experimental studies. Thus, the current systematic review aimed to collect, compare, and analyze the outcome of using different types of stem cells in regenerating maxillary alveolar bone defects. Also, it may provide clinicians with various options when selecting stem cells as a bioactive factor loaded within the tissue-engineered scaffold. Review Materials and methods Two independent reviewers carried out this systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [ 17 ]. Focus Review Question The review question was framed as the following: “Can stem cell therapy be used as a promising future approach in the field of bone reconstruction to treat children and young patients with maxillary alveolar bone cleft defect?” Information Sources An electronic search for articles in the English language was performed using PubMed, Scopus, Web of Sciences, and Google Scholar from 2011 to 2021 due to the lack of updated reviews that were covered by this research area in the dental field. Literature Search Strategy The literature search strategy was carried out in December 2021 and then updated in February 2022. The search was done by following the PRISMA guidelines using subsequent electronic databases: Public Medline (PubMed), Scopus, Web of Sciences, and Google Scholar. The search was conducted using the following combination of keywords: “cleft alveolus”, “maxilla”, “alveolar bone”, “graft”, “repair”, “stem cells”, “dental pulp”, “dental stem cells”; “human DPSCs”, “SHED”, “MSCs”, “mesenchymal stromal cells”; “deciduous tooth”, “deciduous teeth”, “tooth exfoliation”, “regeneration”, “tissue engineering”, “tissue regeneration”, “bone tissue engineering”, “bone transplantation”, “bone reconstruction” “tissue-engineered bone”, “bone regeneration”, “osteogenesis”, “osteoblast”, “bone substitute”, “scaffold”, and “tissue scaffolds”. A detailed summary of the search strategy can be found in Appendix No. 1. Inclusion Criteria Studies were included if they followed the applied criteria: scientific articles published between 2011 and 2021; scientific articles that were published in the English language; and articles conducted on human subjects only. Exclusion Criteria Studies were excluded if they met any of the following applied criteria: review articles; case reports; in vitro/in vivo studies; editorial or personal opinion articles; papers published in a non-English language; papers that illustrated clinical relevance about the regeneration of tissues other than bone; articles that studied stem cells that were not used for cleft alveolus; and articles that discussed the role of stem cells used in bone grafting for the maxillary cleft alveolus by percentages and samples taken from non-human sources. Critical Appraisal The reviewers independently assessed the titles and abstracts of the retrieved publications based on the eligibility criteria and PRISMA standards. Disagreements or contradictions between the two reviewers were resolved through discussion and consensus. Data Extraction After thoroughly reading the articles and taking into consideration the variables "title, abstract, methods, unilateral/bilateral cleft defects, type of stem cell, type of scaffold material, follow-up period following the surgery, and main results, " data were extracted. Both reviewers independently validated the data for completeness and correctness and entered it into standardized Microsoft Office Excel worksheets (Microsoft Corporation, Redmond, WA). Data Items Data from the selected studies were gathered and sorted into columns containing the following information: author and year, study design, the number of subjects, age of the patient, unilateral/bilateral cleft defects, type of stem cell, type of scaffold material, follow-up period following the surgery, type of scoring systems/volumetric measurements used, quantity and quality of bone formation measurements, and main outcomes. Methodological Quality and Risk of Bias Assessment of Included Studies The methodological quality of each study was performed using the risk of bias assessment tool outlined in the Cochrane Risk of Bias tool - VISualization (robvis) [ 18 ]. The Cochrane Collaboration recommends a specific tool to assess the risk of bias in each selected study. The two authors judged the risk of bias of the selected studies based on the following domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias. Each domain was assessed as “low, ” “unclear, ” or “high”. These assessments were reported for each selected study in the “risk of bias” figures. The overall risk of bias associated with each study was evaluated as follows: Low risk of bias: all domains were assessed as “low risk”; Unclear risk of bias: at least one domain was assessed as “unclear risk”; and High risk of bias: at least one domain was assessed as “high risk”. The risk of bias was assessed during the process of data extraction, which could influence the outcome of each selected study. The Cochrane Risk of Bias tool - VISualization (robvis) was used to assess bias present in chosen studies and identify papers with intrinsic methodological and design flaws [ 18 ]. Types of Outcome Measurements Primary outcomes: The status of the bone defect at the end of the alveolar cleft bone repair would either be significant new bone formation by stem cell therapies or failure. Secondary outcomes: The quality and quantity of the newly formed bone in the maxillary alveolar cleft defect by using different scoring systems and volumetric measurements. Synthesis of Results Two tables were developed to describe a variety of relevant data. The first table was prepared to include the study characteristics of each included study and the second table included the outcomes of bone formation by using different scoring systems and volumetric measurements. Statistical Analysis Meta-analysis was not possible due to the heterogeneity of the included studies. As a result, only parametric data relating to the age of the patients in the included studies are presented as a mean and standard deviation (M ± SD), as well as a descriptive evaluation of the findings. Results Study Selection Initially, keywords were used to get a total of 16018 articles from databases. A total of 15094 articles were excluded due to title and abstract duplicity or irrelevance. Following assessment for eligibility, only four papers were involved in this review. Figure 1 depicts a summary of the search flow chart for this systematic review. Figure 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart for study selection Study Characteristics Four human studies that met the inclusion criteria and were conducted during the previous 10 years were included in the search. These studies evaluated the effectiveness of stem cell application in bone regeneration within the maxillary cleft alveolar bone defect. This systematic review included four studies with a total sample of 57 subjects [ 19 - 22 ]. The age of the patients ranged from five months to 10 years in these studies [ 19 - 22 ]. Two studies reported the age of patients with a mean age and standard deviation (mean ± SD) [ 19, 21 ] while the age of patients was not reported in two studies [ 20, 22 ] in which one represented patients receiving implants [ 22 ]. All types of studies included in this systematic review were randomized clinical trials [ 19 - 22 ]. The stem cell types used included: bone marrow mesenchymal stem cells (BMMSCs) [ 20, 22 ], umbilical cord stem cells (UCSCs) [ 21 ], and deciduous dental pulp stem cells (DDPSCs) [ 19 ]. Scaffolds were used to support and seed stem cells for bone tissue engineering in the bone defect sites. The different types of scaffolds used in the included studies were summarized in Table 1. Studies showed that extracting autologous stem cells from different tissue types is safe and results in favorable outcomes presented clinically by supporting alveolar bone cleft defects regeneration [ 19 - 20, 22 ]. In addition, studies emphasized the importance of using scaffolds and membranes that have osteoinductive and osteoconductive properties to enhance the regeneration capacity of stem cells in the bone defect sites. For example, the use of platelet-rich fibrin (PRF) with BMSCs showed superior outcomes compared to an autogenous iliac crest bone graft [ 20 ]. Regarding the cleft defects in this systematic review, three studies included unilateral cleft defects [ 19 - 20, 22 ] while one study involved both unilateral and bilateral cleft defects [ 21 ]. Regarding the follow-up period following the surgery; studies reported different follow-up periods, including five years in one study [ 19 ], 10 years in one study [ 21 ], four months in one study [ 22 ], and one study did not report the follow-up period [ 20 ]. In this section, an informative summary of all included studies and their features is provided in Table 1. A summary of the different types of stem cells used in the included studies for maxillary alveolar bone cleft reconstruction is illustrated in Figure 2. Table 1 Summary of all included studies in this systematic review Authors Year Study Design Number of Subjects / Age of Patients (Mean ± SD) Unilateral / Bilateral Cleft Defects Type of Stem Cells Used Type of Scaffold Material Used Follow-Up Period Following the Surgery Tanikawa DYS, et al. [ 19 ] 2020 (Randomized controlled clinical trial) (n=6) / 10 ± 1. 41 years old Unilateral cleft defects “Deciduous dental pulp stem cell” (DDPSC) hydroxyapatite-collagen sponge (250 mg, Geistlich Biomaterials AG, Wolhusen, Germany) (5 years follow-up) Mossaad A, et al. [ 20 ] 2019 (Randomized controlled clinical trial) (n=24) / Not Reported Unilateral cleft defects “Bone marrow mesenchymal stem cells” (BMMSCs) Group A: Autogenous iliac crest bone + Group B: Nano calcium hydroxyapatite with a collagen membrane + Group C: Bone marrow stem cells extract and Platelet-rich fibrin (PRF) membrane. Not Reported Mazzetti MPV, et al. [ 21 ] 2018 (Randomized clinical trial) (n=9) / 5. 11 ± 0. 60 months newborns Unilateral + Bilateral cleft defects “stem cells from umbilical cord blood and placenta blood” Autologous stem cells (10 years follow-up) Bajestan MN, et al. [ 22 ] 2017 (Randomized controlled clinical trial) (n=18) / Not Reported (Patients receiving implants) Unilateral cleft defects Autologous “Bone marrow mesenchymal Stem cells” (BMMSCs) 2 groups: 1. Control group (n=8) 2. Stem cell therapy (n=10): beta-tricalcium phosphate (β-TCP) (4 months follow-up) Figure 2 Sources of stem cells used in alveolar cleft defect reconstruction The numbers refer to the number of included studies in this systematic review and summarized in Table 1. BMSCs: Bone marrow stem cells, DDPSCs: Deciduous dental pulp stem cell of human healthy extracted deciduous teeth, UCSCs: Umbilical cord stem cells Source: Refs. [ 19 - 22 ]. Primary Outcomes The primary outcomes demonstrated regeneration of the alveolar cleft defect with bone formation following stem cell therapy. Three studies reported neo-bone formation with stem cell application in maxillary alveolar reconstructions [ 19 - 20, 22 ]. All these three studies were controlled clinical trials that showed significant bone formation compared to control groups [ 19 - 20, 22 ]. Only one study showed a non-significant outcome, in which stem cells were injected into the bone defect without using any scaffold or membrane [ 21 ]. An informative description of all included studies and their bone formation outcomes are summarized in Table 2. Table 2 Outcomes of quality and quantity of bone formation and their measurements in this systematic review CBCT: cone-beam computed tomography, CT: computed tomography, HU: Housefield unit, SD: standard deviation, rhBMP-2: recombinant human bone morphogenetic protein-2, BMSCs: bone marrow mesenchymal stem cells, DDPSCs: deciduous dental pulp stem cells Authors Year Type of scoring systems/volumetric measurements used Quality and quantity of bone formation measurements Main Outcomes of stem cell therapy Tanikawa DYS, et al. [ 19 ] 2020 - Volumetric analysis of CT images. - 6 and 12 months’ time points. - Superimposition of the images on anatomical landmarks included the pyriform aperture superiorly, and the cement-enamel junction inferiorly. -The defect at the 6-month follow-up was smaller in the stem cells group (253. 2 mm 3, SD 85. 8) and group two (iliac crest bone graft) (260. 4 mm 3, SD 98. 5) compared to group one (rhBMP) (393. 6 mm 3, SD 144. 7, P=0. 048) - At the 12-month follow-up examination, the mean postoperative defect became similar in all groups. - Bone filling percentage at 6-month follow-up was significantly higher with DDPSCs (75. 6%, SD 4. 8) but at the 12-month follow-up examination, this difference disappeared. Significant results of bone regeneration compared with traditional iliac crest bone grafting and rhBMP-2. Mossaad A, et al. [ 20 ] 2019 Bone density measurement at the graft site from CT compared to normal side in Housefield unit (HU). Bone density was higher in the BMSCs group (mean ± SD 618 ± 60. 2) compared to the normal side (mean ± SD 375. 6 ± 67. 9), followed by nano calcium hydroxyapatite with collagen membrane group (mean ± SD 539. 9 ± 84. 5) compared to normal side with (mean ± SD 395. 3 ± 65. 9) - The autogenous iliac crest group (mean ± SD 461. 0 ± 66. 3) compared to normal side (mean ± SD 368. 5 ± 68. 3) showed resorption in some cases and gave the least values. Superior bone regeneration with bone marrow stem cells followed by nano calcium hydroxyapatite, both groups showed significant differences compared to the autogenous iliac crest group. Mazzetti MPV, et al. [ 21 ] 2018 Facial tomography in one patient, 2 years postoperatively. Not reported There was no evidence of neo-bone formation in cases injected with stem cells. Bajestan MN, et al. [ 22 ] 2017 Ridge width at re-entry was assessed clinically with open bone measurements and radiographically with CBCT. Bone width was 1. 5 ± 1. 5 mm in the stem cell therapy group and 3. 3 ± 1. 4 mm in the control group. Significant bone formation but there is limited osseous regeneration in large defects. Secondary Outcomes The secondary outcomes reported different scoring systems and volumetric measurements to evaluate the quality and quantity of new bone formation postoperatively as shown in Table 2. Three studies evaluated bone formation postoperatively by using computed tomography (CT) scans [ 19 - 20 ]. While one study used a facial tomography scan [ 21 ], another one used cone-beam computed tomography (CBCT) [ 22 ]. Regarding the quality and quantity of bone formation outcomes, three studies reported positive outcomes with stem cell therapy compared to controls and other test groups [ 19 - 20, 22 ]. The measurements of the new bone formation in the defect sites were not reported in one study [ 21 ]. Quality and Risk Assessment of the Included Studies The quality and risk assessment of all included studies were completed by two authors. Included studies were determined by following the Cochrane Risk of Bias tool - VISualization (robvis) [ 18 ] to determine the risk of bias. The majority of the included articles had a low risk of bias in the following domains: blinding of outcomes assessment (50%), incomplete outcome data (50%), selective reporting (50%), other sources of bias (50%). All articles demonstrated a low risk of bias (100%) in random sequence generation; allocation concealment; and blinding of participants and personnel domains (Figure 3 ). Overall, among the four studies, one study (25%) was found to have a low risk of bias [ 19 ] and three studies (75%) had an unclear risk of bias [ 20 - 22 ] as shown in Figure 4. The scoring of unclear risk of bias was given to three studies due to lack of sufficient information to make a clear judgment in the following domains: blinding of outcomes assessment, incomplete outcome data, selective reporting, and other sources of bias (Figure 4 ). Figure 3 Overall risk of bias summary of all selected studies Figure 4 Risk of bias tool of the selected studies (VISualization - (robvis)) Source: Refs. [ 19 - 22 ] Discussion This systematic review was conducted to describe and evaluate all research findings in the previous 10 years that satisfied our research objective. It included all the latest clinical studies on the role and application of stem cells for bone reconstruction in maxillary alveolar bone cleft defects. Our review demonstrates a comprehensive set of evidence extracted from four articles that fulfilled our inclusion and exclusion criteria. Up to date, there is only one recent systematic review that investigated the use of stem cells in clinical application for bone regeneration in bone defects covering the period from 1984 to 2017 and is summarized in Table 3 [ 16 ]. Fifty-six studies supported the role of human exfoliated deciduous teeth (SHEDs) and human dental pulp stem cells (hDPSCs) in repairing bone defects, including cranial/calvarial, mandibular, tibial bone, and femoral bone defects [ 16 ]. The included studies involved animal experimental models and four human studies, in which three human studies investigated repairing post third molar extraction defects with collagen sponge scaffold and one human clinical trial studied periodontal bone defect regeneration with beta-tricalcium phosphate (β-TCP) scaffold [ 23 - 26 ]. It was concluded that the majority of retrieved studies suggested that stem cells isolated from SHEDs and hDPSCs were effective in bone tissue repair and regeneration for clinical application in animal models or humans [ 16 ]. However, alveolar cleft defects were not included in their studies, which warranted the necessity of conducting this review. On the other hand, the only systematic review that investigated tissue engineering strategies for alveolar cleft reconstruction in humans up to the year 2012 included only one stem cell study [ 27 ]. However, neither the bone quantity nor quality data were provided [ 28 ]. In agreement with previously published systematic reviews, few studies used stem cells in bone regeneration in humans. Most of the included studies in our systematic review favored the use of stem cells in bone regeneration within the alveolar bone defects for cleft patients (Table 1 ). They used various techniques in extracting, isolating, culturing, and characterizing stem cells. Interestingly, none of the studies reported neither adverse nor negative effects on the clinical application of selected stem cells (Table 1 ). Table 3 Summary of the recent systematic review included in this systematic review SHEDs: human exfoliated deciduous teeth, hDPSCs: human dental pulp stem cells Authors Year Number of studies using Method summary Main Conclusions Leyendecker Junior A, et al. [ 16 ] 2018 56 studies The systematic review summarises and presents in vivo studies performed from 1984 to November 2017. Using two different databases (PubMed/MEDLINE and Web of Science databases), an electronic search was done. The use of SHEDs and hDPSCs appears to be effective for bone repair/regeneration as clinical applications for the cleft alveolus. The use of MSCs with a PRF membrane showed favorable bone regeneration represented in increased bone width and density (Table 2 ) [ 20 ]. CT measurements showed that bone density in the MSCs group was higher (mean ± SD 618 ± 60. 2 compared to normal side mean ± SD 375. 6 ± 67. 9) followed by the nano calcium hydroxyapatite with collagen membrane group (mean ± SD 539. 9 ± 84. 5 compared to normal side with mean ± SD 395. 3 ± 65. 9) [ 20 ]. Whilst the autogenous iliac crest group resulted in resorption in some cases and gave the least values (mean ± SD 461. 0 ± 66. 3 compared to normal side mean ± SD 368. 5 ± 68. 3) [ 20 ]. However, the use of MSCs without a scaffold or membrane showed limited osseous regeneration [ 22 ]. The same outcome was observed in umbilical cord stem cells (UCSCs) studies; adequate alveolar height when cells were seeded on a gelfoam scaffold while no evidence of neo-bone formation was detected by injecting UCSCs without scaffold or membrane [ 21 ]. However, no measurements were reported in this study [ 21 ]. The use of DDPSCs in the Tanikawa et al. study not only showed that stem cells harvested from shedding teeth present a reliable source of stem cells for bone regeneration in cleft defects, but it also demonstrated that the numbers of cells harvested from each tooth are sufficient to seed regenerating scaffolds [ 19 ]. In addition, the DDPSCs group showed comparable defect size regeneration to the traditional iliac crest graft group and superior bone filling percentage at the six months follow-up (Table 2 ). However, at the 12-month follow-up, both groups and the rhBMP-2 group showed the same outcome. Although autologous bone graft is considered the gold standard in bone regeneration, it involves second surgical site morbidity with a limited amount of bone to be harvested. In addition to the lengthy operative time and stay in the hospital, there is a risk of intraoperative blood loss, postoperative pain, and high cost [ 19 ]. On the other hand, rhBMP-2 adverse effects involved severe swelling in maxillofacial surgery and postoperative nasal stenosis in cleft children [ 29 - 30 ]. As a result, the Food and Drug Administration (FDA) warning was issued against the utilization of it in the pediatric population due to a lack of evidence that confirms long-term effectiveness or safety in children [ 19 ]. The use of stem cells isolated from shedding teeth could transform bone regeneration procedures in cleft alveolus cases. This is due to the enrichment of deciduous teeth with stem cells and the ease of isolating them compared to other sources in the human body [ 19 ]. Although most of the included studies in this review favor bone marrow stem cells in cleft alveolar bone regeneration in terms of clinical, radiological, and histopathological outcomes, the level of evidence remains low since there are few human studies. This warrants the development of standardized protocols to retrieve and process the different types of stem cells, standardize and report the timing of intervention, the dimension of the maxillary alveolar bone cleft defect to be repaired, and the method of radiological assessment procedure used during follow up stages after surgery. The three-dimensional radiographic evaluation with CT is still considered the most reliable method for the analysis of height and volume of the alveolar bone [ 15 ]. In summary, the current review demonstrates an overview of different studies showing the outcomes after using BMMSCs, UCSCs, and DDPSCs for bone regeneration in maxillary alveolar cleft defects (Table 2 ). It can be concluded that there are insufficient data to support the use of stem cells in alveolar bone cleft repair for now. However, the favorable results from the included studies in this review encourage the need for more clinical trials to be able to consider stem cells as one of the treatment approaches in dentofacial bone regeneration defects. Moreover, standardization and thorough clinical evaluation in adult patients is a prerequisite before investigating this form of therapy in pediatric patients. Study Strengths and Limitations of This Systematic Review Our review compiled and evaluated all peer-reviewed publications published in the previous 10 years that met our inclusion criteria. This is the only systematic review that has thoroughly explored the subject of using stem cells in bone regeneration for an alveolar cleft in depth. The systematic review conducted by Leyendecker Junior A et al. (2018) covered the period between 1984 and 2017 [ 16 ]. In addition, we used Public Medline [PubMed], Scopus, Web of Sciences, and Google Scholar as search engines. One advantage of using Google Scholar is that it prevents researchers from missing any valuable research that has been published in journals but has not yet been cited in PubMed, Scopus, or Web of Science. On the other hand, we were limited to perform a systematic review without meta-analysis on the selected articles because of the heterogeneity of the confounding factors in the currently involved human studies. In addition, only a few studies were conducted in human subjects that were part of the inclusion criteria and used stem cells in bone reconstruction to treat children and young patients with alveolar clefts. Conclusions This systematic review revealed that there were limited studies on humans using stem cells for alveolar cleft defect repair. They reported encouraging results from stem cell therapy, including significant bone formation in defect sites. A needle aspirate was used to harvest stem cells from bone marrow, and stem cells from extracted or exfoliated teeth were isolated from the dental pulp. Compared to the gold standard bone grafting procedure, which involves donor site morbidity, these noninvasive methods of isolating stem cells make them attractive options for bone repair in the near future. However, more research is needed before the FDA can set suitable standards and restrictions for employing stem cells in alveolar cleft clinical trials. In future clinical studies, it will be critical to standardize a methodology for extracting and processing stem cells in sufficient numbers to enhance bone regeneration within the implanted scaffold.

10. 7759/cureus. 23318

Cureus

Efficacy of Chitosan and Chlorhexidine Mouthwash on Dental Plaque and Gingival Inflammation: A Systematic Review

Mouthwash is the effective chemical plaque control mechanism being practiced globally. Teeth and tongue discoloration, a temporary change in taste perception, an increase in calculus deposits, a burning sensation, and genotoxicity of buccal epithelial cells are all possible side effects. This review evaluates the efficacy of chitosan mouthwash in comparison to chlorhexidine mouthwash in combating plaque accumulation and gingival inflammation. Electronic databases such as Medline, Cochrane, LILACS, TRIP, Google scholar, and clinical trial registries (CTRI) for ongoing trials were searched with appropriate medical subheadings (MeSH) and search terms. Randomized clinical trials comparing the efficacy of chitosan mouthwash and chlorhexidine mouthwash on dental plaque accumulation and gingivitis were included. The outcome variables of interest were plaque index, gingival index, gingival bleeding index, and colony-forming unit (CFU/ml). All data from the included studies were extracted in a customized extraction sheet. The risk of bias across the studies was assessed using the Cochrane tool for intervention (ROB-2), which consisted of six domains. Of the included three studies, we found one study with an overall low risk of bias and two studies with an overall high risk of bias across the domains. Though there was a significant reduction in plaque accumulation, gingival inflammation, and colony-forming units on the use of chitosan mouthwash and chlorhexidine mouthwash separately, all three included studies reported that a combination of both be more effective.

Introduction and background Plaque‐induced gingivitis is a very common periodontal disease that is seen in everyday dental practice. It is caused by the build up of microbial biofilms on the surfaces of teeth, and poor or insufficient oral hygiene is the primary cause [ 1 ]. Plaque-induced gingivitis is characterized by redness, puffiness, and a proclivity for easy bleeding when brushing or flossing. If left untreated, gingivitis, the first stage of periodontal disease, would spread and infiltrate the soft and bony supporting components of the teeth, eventually leading to tooth loss. Plaque-induced gingivitis treatment to prevent and reduce plaque accumulation by a number of approaches that enhance oral hygiene [ 2 ]. These include tooth brushing, flossing, tooth cleaning sticks, oral irrigators, and/or professional scaling and polishing to mechanically remove dental plaque [ 3 ]. However, due to subjective variability, the efficacy of mechanical plaque management remains debatable [ 4, 5 ]. In such circumstances, antimicrobial mouthwashes should be used in conjunction with mechanical oral hygiene methods [ 6 ]. Antimicrobial mouthwashes prevent plaque development by decreasing oral bacteria's growth, metabolism, and colonization [ 7 ]. Mouthwashes containing chlorhexidine gluconate are the most commonly used supplement to mechanical intervention in the treatment of gingivitis. It has been shown to be extremely effective in reducing plaque accumulation and is considered the gold standard for plaque control [ 8 ]. Long-term usage of chlorhexidine, however, has been linked to a variety of side effects, including tooth and tongue discoloration, a temporary change in taste perception, an increase in calculus deposits, a burning sensation, and genotoxicity of buccal epithelial cells [ 6, 9 ]. A variety of natural products have been incorporated into dental for plaque control and caries prevention due to increased antibiotic resistance and adverse effects of some antimicrobials on the one hand, and the safety, availability, and relatively low costs of natural products on the other hand [ 10 ]. Chitosan is a natural polymer made from the alkaline hydrolysis of chitin, a natural chemical present in exoskeletons of arthropods, crab shells, and insect cuticles. Chitosan and related nanoparticles have gotten a lot of attention in the pharmaceutical, food, agriculture, textile, and tissue engineering industries because of their inherent biocompatibility, biodegradability, and lack of toxicity. Chitosan contains antibacterial, antioxidant, wound healing, and mucoadhesive properties [ 11, 12 ]. Chitosan has an anti-adherence activity which causes bacterial surface modifications, alterations in bacterial surface ligands expression levels, and gets adsorbed to the hydroxyapatite crystals in the tooth surface. These characteristics are responsible for the bactericidal and bacteriostatic properties of chitosan [ 13 ]. The interaction of cationic chitosan with the anionic cell surface, increasing membrane permeability and cellular material leakage from the cell may be the antibacterial mechanism of chitosan. Chitosan may also interfere with the production of mRNA and the embedding of proteins [ 14 ]. Due to its outstanding features such as absorbability, malleability, and cohesive threshold concentration to store and gradually release pharmaceuticals with optimal resorption, it has previously been used as a carrier system for the local administration of various drugs [ 15 ]. It also has anti-inflammatory properties because it affects prostaglandin E2 levels [ 16 ]. Since evidence suggest chitosan to be less cytotoxic and genotoxic this review has been directed to find its efficacy against chlorhexidine [ 17, 18 ]. The objective of the review is to evaluate the efficacy of chitosan vs chlorhexidine Mouthwash on dental plaque and gingival inflammation. Review All randomized controlled and clinical trials comparing the efficacy of chitosan and chlorhexidine mouthwash were included. Case reports, case series, in-vitro, and animal studies that measured the plaque accumulation and gingivitis were excluded. Periodontally healthy individuals aged 30 years and above of both genders were the populations in each of the included studies. The primary outcomes of the review are plaque index, gingival index. The secondary outcome is the total bacterial count as colony forming units (CFU/mL). A detailed search strategy for each database to find out studies for this review was developed. Both free-text terms, MeSH terms and a combination of both was used to search in each database. Electronic searches were conducted on Medline, Cochrane, LILACS, TRIP, and Google scholar. No language restrictions were placed. However, studies published from 2011 to 2021 were included. The search strategy of the Medline database is given in Appendix 1. World Health Organization International Clinical Trials Registry Platform was searched to identify for ongoing trials. Also, a hand search was made with the help of a librarian in the following journals: Journal of Periodontology, Journal of Clinical Periodontology, Journal of Periodontal Research. The authors assessed the obtained studies to find whether they met the selection criteria. All randomized controlled and clinical trials that compared the effectiveness of chitosan and chlorhexidine on plaque accumulation and gingival inflammation which met the selection criteria were included. Case reports, case series, in-vitro studies, and animal trials that measured plaque and gingival index were excluded. Any disagreement between the authors was resolved by discussion. Each author screened for the title and abstract of each article identified through the search strategy. If there were insufficient data in the title and abstract, the full text of the publications was collected in order to make a clear choice. Each of the review authors independently extracted data with the help of a specially designed data extraction sheet. The following data were recorded for each study: study ID, study design, sample size, participants and group, methodology, parameters, statistical analysis, and results. Any disagreements in the data extraction were sorted out by discussion. Assessment of methodological quality of included studies was performed. Each author independently assessed the risk of bias for each included study, any disagreement was resolved by discussion to arrive at a consensus. Revised Cochrane Risk of Bias Tool for Randomized Trials (RoB‑2) [ 19 ] to assess the risk of bias for included studies. Bias arising from the randomization process, bias due to deviations from the intended interventions, bias due to missing outcome data, bias in the measurement of the outcome, and bias in the selection of the reported results were all domains used to assess the methodological quality of the included studies. The overall score was given for each study based on the scores of each domain. Low RoB for studies for which we identified all domains as being “low risk. ” Studies for which we identified one domain with some concerns come under “some concerns. ” High RoB for studies for which we identified one or more domains as being “high risk” and more than two domains as “some concerns. ” Since, there was difference in the methodological assessment for plaque index, gingival index and also availability of insufficient data to pool the results, meta-analysis has not been performed. Thus, forest plot for pooled results and funnel plot for publication bias were not performed. A search strategy yielded 112 publications from various databases. On removing 26 duplicate records, we landed up with 86 records. These 86 records were screened for title and abstract, of which 69 records were not suitable for this review. About 17 remaining records, 14 records were excluded for not meeting the inclusion criteria. Thus, three records remained for qualitative analysis in the review as shown in Figure 1. Figure 1 Flowchart of search The authors individually extracted data with the help of a data extraction form and entered all the details into a spreadsheet. The following attributes were recorded for each of the included studies: study ID, study design, sample size, participants and group, methodology, parameters, statistical analysis and results of the plaque index and gingival inflammation. The detailed characteristics of the included studies and the method of assessment and follow‑up data were presented in Table 1. Among the three studies, two studies are Interventional clinical trials and one study was a randomized controlled trial. The parameters analyzed were gingival index, plaque index, bleeding index and colony forming unit. Table 1 Summation table of the included studies Abbreviation: CHX- Chlorhexidine; CHT- Chitosan; CFU- Colony Forming Unit Study ID Year Evaluation period Study groups Method of evaluation Outcome Limitations/future scope Vilasan et al. [ 20 ] 2020 3 months 1. Group 1, 20 patients rinsing with 20 ml of 0. 2% chlorhexidine twice daily for 3 months were allocated. 2. Group 2-10 patients rinsing with 20 ml of 0. 5% chitosan twice daily for 3 months were allocated. 3. Group 3- 10 patients rinsing with 10 ml of chlorhexidine chitosan combination twice daily for 3 months were allocated 1. Plaque index using Turesky Gilmore Glickman modification (1970) of the Quigley and Hein (1962) index. 2. Gingival index using Loe and Silness index (Loe and Schiott, 1970) 3. Bleeding on probing using Ainamo and Bay index. The combination of chitosan and chlorhexidine showed a statistically significant reduction (p<0. 05) in plaque indices from baseline at all-time intervals when compared to that of chlorhexidine or chitosan alone. The study was carried out on a small sample size with short evaluation time. It did not have any microbiological analysis and toxicity testing. No attempt was made to find out the exact mechanism of chlorhexidine and chitosan combination. Mhaske et al. [ 21 ] 2018 4 days 1. Group I included 15 subjects who used 0. 2% CHX 2. Group II included 15 subjects who used 2% CHT solution 3. Group III involves 15 subjects who used 0. 2% CHX/2% CHT combination. Plaque index, Gingival index and Streptococcus mutans count 1. Plaque index was lowest in group I at day 0, while it was highest in group III. 2. At day 4, PI was highest in group II, while lowest in group III. 3. Gingival index was lowest in group I and highest in group II at day 0, and lowest in group I and highest in group III at day 4. 4. Both chitosan and chlorhexidine were found to be effective in controlling plaque. However, a combination of both provides even better results. Less sample size and low follow up days Nair et al. [ 22 ] 2017 7 Days 1. CHX group A: 30 seconds mouth rinsing with 15 ml of CHX mouthwash for 20 patients. 2. CHT group B: 30 seconds mouth rinsing with 20 ml of CHT in 10 ml of water for 20 patients. Colony forming unit (CFU) The mean CFU count reduction after using 0. 12% CHX and 2% CHT for one week were 3. 563X102 and 3. 714X102 respectively. Both the mouthwashes were effective in reducing the total bacterial count after one week. Less follow up days. Specific microorganism name was not mentioned in the study. In Figures 2, 3, the review authors' judgments on each RoB-2 were shown as percentages across all included research, and the RoB-2 for each included study was presented as a summary. Among the three included studies, one study had low risk of bias [ 20 ] and two other studies has high over all risk of bias [ 21, 22 ]. Figure 2 Risk of bias - 2 (ROB-2) presented as percentages across all included studies Figure 3 Risk of bias - 2 (ROB-2) for each included study Summary of the main results To the best of our knowledge, this is the first systematic review to compare the efficacy of chitosan mouth rinse versus chlorhexidine in preventing plaque and gingivitis. Overall, chitosan is helpful in reducing plaque and gingival irritation, according to the findings of this systematic review of 145 samples. The quality of the included studies, as well as the high heterogeneity among them, must be considered when weighing the results of this review. The reduction of plaque scores and gingival inflammation were the major endpoints of this study. Despite the fact that most individuals practice brushing their teeth at least twice a day, the prevalence of gingivitis and chronic periodontitis remains high in most populations around the world [ 23 ]. Effective plaque control is well acknowledged as a critical aspect in the prevention and treatment of periodontal disorders [ 2, 7 ]. Despite the fact that mechanical oral hygiene is the simplest and most effective technique for plaque reduction, the majority of adults do not brush or floss their teeth effectively [ 4, 5 ]. Mouth rinses have been shown to be efficient in blocking and lowering gingival plaque formation when used in conjunction with mechanical oral hygiene [ 7 ]. In the current review, two studies [ 19, 20 ] found that a combination of chlorhexidine mouthwash and chitosan mouthwash was effective in reducing plaque and gingival score, whereas Nair et al. [ 22 ] evaluated the in vivo effect of CH and chitosan on plaque microbial and found a mean colony-forming unit count reduction after using 0. 125% CHX and 2% chitosan for one week and concluded that both are effective and Van Strydonck et al. [ 24 ] compared 0. 12% CHX to 0. 05% cetylpyridinium chloride and 0. 2% CHX after three days and found no significant difference in plaque accumulation in either group. Costa et al. [ 25 ] stated that chitosan is efficient against the majority of bacteria and recommended it as a replacement for standard mouthwashes. Decker et al. [ 26 ] investigated the effects of CHX on plaque combinations in order to develop antiplaque techniques. In that investigation, CHX (0. 1%) was utilized as a positive control, saline was used as a negative control, and two CHT derivatives were linked to Streptococci sanguis for two minutes with their CHX combination. According to their findings, the CHX & CHT combination was more effective than CHX alone because it combined the bioadhesive qualities of CHT with the antibacterial activity of CHX, resulting in a synergistic antiplaque effect that was superior to CHX alone. The antiplaque action of chitosan, according to Decker et al. [ 27 ] and Costa et al. [ 28 ], is due to its antiadhesive property toward microbes. Chitosan, according to some researchers, can be used efficiently in dentifrices to promote oral hygiene since it decreases plaque by 70% [ 29 ]. In a randomized clinical experiment, Uraz et al. [ 30 ] investigated the clinical and microbiological effects of chitosan on dental plaque and discovered a reduction in microbiological count ( S. mutans and C. albicans levels) in both the CH and chitosan groups. Chen and Chung [ 31 ] tested the bactericidal activity of chitosan in vivo and in vitro at various temperatures (25-37°C) and pH levels (pH 5-8). They discovered that chitosan has antibacterial properties equivalent to commercial mouthwashes. They concluded that in the future, water-soluble chitosan could be a viable alternative to commercial mouthwashes. Costa et al. [ 28 ] investigated the possible use of high- and low-molecular-weight chitosan as an oral antibacterial agent and found that efficiency decreased only little after a week. They also discovered that chitosan could block the formation of biofilms by two microorganisms and could act on mature biofilms, resulting in a 94% reduction in biofilm survival. Giunchedi et al. [ 32 ] looked examined CHX buccal tablets made from drug-loaded CH microspheres. Combining CHT microspheres with CHX as a controlled drug delivery system not only extended the drug's release in the oral cavity, but also increased CHX's antibacterial effectiveness. Strengths and weaknesses of the review This review included all randomized controlled and clinical trials and excluded case reports, case series, in-vitro, animal studies and ex-vivo studies. Because this evaluation looked at all human in-vivo trials, it has a lot of therapeutic application. Every precaution was taken to reduce bias at every stage of the evaluation. To discover all relevant studies, we searched electronic databases and trial registries with no language constraints. We used the Cochrane Risk of Bias - 2 tool to assess the methodological quality of the included studies, which has five categories and one overarching area that offers exploratory information on the risk of bias. One included study had low risk of bias and two included studies has high risk of bias. The primary outcomes of the review are plaque index and gingival index. Though each study evaluated plaque index and gingival index as scores, the criteria of the indices used are different. Therefore, no quantitative analysis and data synthesis, investigation of heterogeneity, sensitivity analyses were performed. Also, the present review failed to search for other databases such as Excerpta Medica database (EMBASE) and EBSCO. All included studies are from one country (India), the results of the review may or may not be generalizable to other countries. Thus, the applicability of the results of this review is possible to the Indian population. Conclusions Based on the present review, both chitosan and chlorhexidine are found to be effective in controlling plaque and gingival inflammation. However, a combination of both provides even better results. Chitosan can be used as an alternative mouthwash. Further, randomized controlled trials following the CONSORT guidelines from a different population with different cultural and racial variations are needed to validate the effectiveness of chitosan on plaque accumulation and gingival inflammation.

10. 7759/cureus. 23419

Cureus

Top 50 Cited Bone Graft Orthopedic Papers

The purpose of this research is to recognize the highest 50 most-mentioned articles in the literature concentrating on bone grafts. That has been accomplished with the use of the Scopus database and the search slogan "bone grafts, " and we inquired for the 50 most-cited articles on bone grafting. The study was completed in September 2020. We investigated the articles issued between 1970 and 2020. The articles were organized and classified based on the total number of citations. We appraised the following information relating to each article: first author, year of publication, journal, and title. A total of 1, 580 studies matched our search standards, of which the 50 most-cited extended between 1, 862 and 403 citations. Seven articles were cited more than 1, 000 times. The article by Marx et al. was the maximum-cited article, with 1, 862 citations, followed by Younger et al. 's with 1, 461 and Giannoudis et al. 's with 1, 245. The majority of the studies originated from the United States (n = 30) and were published in the 2000s. Biomaterials was the most regular destination journal (n = 8), followed by the Journal of Bone and Joint Surgery American series (n = 7). A maximum of the articles focused on the different types of bone grafts and their alternatives including bone tissue engineering (n=29). Our investigation of the highest 50 articles linking to bone grafting has emphasized the most significant papers in the field. These cover a wide-ranging variety of topics including types, management, and mechanism of action of bone grafts. To recognize the present treatment guidelines and how the use of bone grafting has grown, it is vital to know the most-cited articles relating to this grafting.

Introduction and background The natural science of fracture healing is better recognized than ever before, with developments in orthopedic implants such as locked plates and bioabsorbable screws, and the osseous healing has become more expectable and less eventful. Nevertheless, occasionally one’s intrinsic biological response, or simultaneous surgical stabilization, is insufficient. With the hope of facilitating bone union, bone grafts, bone substitutes, and orthobiologics are being depended on more than ever before. The osteogenic, osteoconductive, and osteoinductive properties of these substrates have been illuminated in the basic science literature and authorized in the clinical orthopedic practice. Furthermore, business constructed around these substances is more fruitful and desirable than ever before. This analysis provides a wide-ranging overview of the basic science, clinical value, and economics of bone grafts, orthobiologics, and bone substitutes [ 1 ]. Within the academic medical field, the number of times an article is quoted by other writers has been commonly considered to be a dependable pointer of its academic influence and effect within this field [ 2 ]. Since Lefaivre et al. determined the 100 utmost-cited articles in the orthopedic field [ 3 ], there have been abundant reports categorizing the most-referenced articles across a wide range of orthopedic surgery subspecialties and subject ranges, including shoulder, hand, foot and ankle, arthroscopic surgery, hip arthroplasty, and trauma surgery [ 4 - 9 ]. The design of this research was to scrutinize the 50 most-cited articles in bone grafting in orthopedics and the features that make them significant to physicians and researchers within the orthopedic field. To achieve this goal, data from the Scopus citation indexing service were used to achieve an inclusive, organized citation search of all orthopedic-specific publications journal by journal. Given the nature of the field, we theorized that a noteworthy share of the detected citations would be basic science studies. Review Method The 50 most-cited articles linked to bone grafting were examined in the Scopus engine by using defined search terms. All forms of scientific papers, reviews, and conference papers with reference to our subject were graded along with the absolute number of citations and scrutinized for the following features: journal title, year of publication, number of citations, citation density, geographic origin, and article type. Mean citation number was considered as the total number of citations the article has established divided by the number of years since publication (total citations/years since publication) [ 10 ]. Results The highest 50 articles concerning bone grafting have been cited a total of 33, 895 times. The average number of citations per year is 753. 22. The maximum 50 articles, numbers of citations, and mean citation number are listed in Table 1. Table 1 Top 50 cited research papers relating to bone grafting. First author Title Citations Citations /year 1 R. E. Marx [ 11 ] Platelet-rich plasma: growth factor enhancement for bone grafts 1862 83 2 E. M. Younger [ 12 ] Morbidity at bone graft donor sites 1461 47. 13 3 P. V. Giannoudis [ 13 ] Bone substitutes: an update 1245 83 4 A. J. Salgado [ 14 ] Bone tissue engineering: state of the art and future trends 1120 70 5 S. Bose [ 15 ] Recent advances in bone tissue engineering scaffolds 1115 139. 38 6 E. Arrington [ 16 ] Complications of iliac crest bone graft harvesting 1096 45. 67 7 G. Ian Taylor [ 17 ] The free vascularized bone graft: a clinical extension of microvascular techniques 1045 23. 22 8 A. R. Amini [ 18 ] Bone tissue engineering: recent advances and challenges 995 124. 38 9 J. C. Banwart [ 19 ] Iliac crest bone graft harvest donor site morbidity: a statistical evaluation 992 39. 68 10 E. Carragee [ 20 ] A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned 906 100. 67 11 T. W. Bauer [ 21 ] Bone graft materials: an overview of the basic science 833 41. 65 12 C. Damien [ 22 ] Bone graft and bone graft substitutes: a review of current technology and applications 812 28 13 H. Burchardt [ 23 ] The biology of bone graft repair 755 20. 41 14 M. Kikuchi [ 24 ] Self-organization mechanism in a bone-like hydroxyapatite/collagen nanocomposite synthesized in vitro and its biological reaction in vivo 736 38. 74 15 G. E. Friedlaender [ 25 ] Osteogenic protein-1 (bone morphogenetic protein-7) in the treatment of tibial nonunions 733 38. 58 16 R. Dimitriou [ 26 ] Bone regeneration: current concepts and future directions 689 76. 56 17 J. M. Kanczler [ 27 ] Osteogenesis and angiogenesis: the potential for engineering bone 671 55. 92 18 J. Goulet [ 28 ] Autogenous iliac crest bone graft: complications and functional assessment 666 28. 96 19 C. G. Finkemeier [ 29 ] Bone-grafting and bone-graft substitutes 654 36. 33 20 A. W. Yasko [ 30 ] The healing of segmental bone defects, induced by recombinant human bone morphogenetic protein (rhBMP-2). A radiographic, histological, and biomechanical study in rats 641 22. 89 21 J. Silber [ 31 ] Donor-site morbidity after anterior iliac crest bone harvest for single-level anterior cervical discectomy and fusion 638 37. 53 22 M. Yaszemski [ 32 ] Evolution of bone transplantation: molecular, cellular, and tissue strategies to engineer human bone 606 25. 25 23 H. Wang [ 33 ] Biocompatibility and osteogenesis of biomimetic nano-hydroxyapatite/polyamide composite scaffolds for bone tissue engineering 600 46. 15 24 P. Hernigou [ 34 ] Percutaneous autologous bone-marrow grafting for nonunions: influence of the number and concentration of progenitor cells 597 39. 8 25 T. J. Herbert [ 35 ] Management of the fractured scaphoid using a new bone screw 589 16. 36 26 L. T. Kurz [ 36 ] Harvesting autogenous iliac bone grafts: a review of complications and techniques 587 18. 94 27 R. Dimitriou [ 37 ] Current concepts of molecular aspects of bone healing 578 38. 53 28 S. Boden [ 38 ] Use of recombinant human bone morphogenetic protein-2 to achieve posterolateral lumbar spine fusion in humans: a prospective, randomized clinical pilot trial 2002 Volvo award in clinical studies 554 30. 78 29 R. Murugan [ 39 ] Biomimetic nanocomposites for bone graft applications 552 36. 8 30 J. Woodard [ 40 ] The mechanical properties and osteoconductivity of hydroxyapatite bone scaffolds with multi-scale porosity recombinant human bone morphogenetic protein-2 542 41. 69 31 P. Warnke [ 41 ] Growth and transplantation of a custom vascularized bone graft in a man 525 32. 81 32 M. Geiger [ 42 ] Collagen sponges for bone regeneration with rhBMP-2 522 30. 71 33 S. Laurie [ 43 ] Donor-site morbidity after harvesting rib and iliac bone 515 14. 31 34 H. Mankin [ 44 ] Long-term results of allograft replacement in the management of bone tumors 506 21. 08 35 W. R. Moore [ 45 ] Synthetic bone graft substitutes 492 25. 89 36 J. Zins [ 46 ] Membranous versus endochondral bone: implications for craniofacial reconstruction 489 13. 22 37 H. Frost [ 47 ] A 2003 update of bone physiology and Wolff s law for clinicians 453 28. 31 38 W. Bonfield [ 48 ] Hydroxyapatite reinforced polyethylene - a mechanically compatible implant material for bone replacement 448 11. 49 39 H. Yuan [ 49 ] Osteoinductive ceramics as a synthetic alternative to autologous bone grafting 443 44. 3 40 P. Francis [ 50 ] Bone morphogenetic proteins and a signaling pathway that controls patterning in the developing chick limb 438 16. 85 41 S. Khan [ 51 ] The biology of bone grafting 435 29 42 D. Tadic [ 52 ] A thorough physicochemical characterization of 14 calcium phosphate-based bone substitution materials in comparison to natural bone 435 27. 19 43 E. Ahlmann [ 53 ] Comparison of anterior and posterior iliac crest bone grafts in terms of harvest-site morbidity and functional outcomes 434 24. 11 44 J. Inzana [ 54 ] 3D printing of composite calcium phosphate and collagen scaffolds for bone regeneration 433 72. 17 45 W. De Long [ 55 ] Bone grafts and bone graft substitutes in orthopedic trauma surgery: a critical analysis 423 32. 54 46 O. Bergland [ 56 ] Elimination of the residual alveolar cleft by secondary bone grafting and subsequent orthodontic treatment 421 12. 38 47 P. Hernigou [ 57 ] Treatment of osteonecrosis with autologous bone marrow grafting 416 23. 11 48 G. Daculsi [ 58 ] Biphasic calcium phosphate concept applied to the artificial bone, implant coating and injectable bone substitute 413 18. 77 49 A. Oryan [ 59 ] Bone regenerative medicine: classic options, novel strategies, and future directions 412 68. 67 50 A. Greenwald [ 60 ] Bone-graft substitutes: facts, fictions, and applications 408 21. 4 The most commonly cited paper was by R. E. Marx et al. in 1998 representing a greater bone density in bone grafts with platelet-rich plasma with a total of 1, 862 citations (mean citations 83/year) [ 11 ]. The most primitive publication was in 1975 by G. Ian Taylor et al. indicating a novel technique of free vascularized bone graft technique used and combined with a suitable soft tissue flap repairing method, where this system was established to salvage two injured legs which would otherwise have been amputated [ 17 ]. The newest publications were in 2014 by J. Inzana about a new category of bone graft technique which used low-temperature 3D printing of calcium phosphate scaffolds with greater functioning over old-style methods [ 54 ], and by A. Oryan who studied the literature of bone grafting and presented bone tissue engineering as an approach in the orthopedic surgery [ 59 ]. The maximum frequent decade in this list was the 2000s with 24 papers (Table 2 ). Table 2 Top 50 papers published by decade. Decade Number 1970s 1 1980s 8 1990s 10 2000s 24 2010s 7 Twenty-six journals were included in publishing the maximum of 50 articles (Table 3 ). Impact factors of these journals fluctuated between 0. 372 and 59. 102. Journal of Biomaterials occupied the upper position of this list with eight publications (16%) and chased closely by the Journal of Bone and Joint Surgery - American Volume (n = 7) (14%) and Clinical Orthopaedics and Related Research (n = 6) (12%). The English language was the common language in all papers. Table 3 Top 50 papers published per medical journal. Medical journal Number Impact factor 2018 Biomaterials 8 10. 273 Journal of Bone and Joint Surgery - Series A 7 4. 716 Clinical Orthopaedics and Related Research 6 4. 154 Spine 4 3. 024 Plastic and Reconstructive Surgery 3 3. 682 Injury 2 1. 620 Angle Orthodontist 1 2. 028 ANZ Journal of Surgery 1 1. 071 Advanced Drug Delivery Reviews 1 16. 663 BMC Medicine 1 8. 639 Cleft Palate Journal 1 1. 395 Composites Science and Technology 1 6. 808 Critical Reviews in Biomedical Engineering 1 0. 660 Development 1 5. 763 European Cells and Materials 1 3. 682 Journal of Applied Biomaterials: An Official Journal 1 0. 372 Journal of Bone and Joint Surgery - Series B 1 4. 301 Journal of Orthopaedic Surgery and Research 1 1. 907 Journal of Orthopaedic Trauma 1 1. 758 Lancet 1 59. 102 Macromolecular Bioscience 1 2. 895 Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontics 1 1. 791 Proceedings of The National Academy of Sciences of The United States of America 1 9. 553 Spine Journal 1 2. 903 The Journal of the American Academy of Orthopaedic 1 2. 441 Trends in Biotechnology 1 12. 068 Total 50 The highest 50 articles were created from 12 diverse countries (Table 4 ), where the USA was in the topmost with 30 articles (60%), then the UK with five articles (10%). Twenty-nine research papers are available as articles, while 15 reviews are involved in the uppermost cited papers and the conference papers are demonstrated six times (Table 5 ). Table 4 Countries of top 50 research papers. Country Frequency Percent USA 30 60. 0 UK 5 10. 0 Australia 3 6. 0 France 3 6. 0 Germany 2 4. 0 Iran 1 2. 0 Japan 1 2. 0 Netherlands 1 2. 0 Norway 1 2. 0 Portugal 1 2. 0 Singapore 1 2. 0 China 1 2. 0 Total 50 100. 0 Table 5 The origin of top 50 papers. Origin Frequency Article 29 Conference paper 6 Review 15 Total 50 A number of significant subjects are demonstrated in this list of top 50 papers. Twenty-nine articles (58%) scrutinize several categories of bone grafting. Besides these, seven papers are focused on bone tissue engineering, which points to inducing a novel practical bone regeneration method through a synergetic combination of biomaterials, cells, and numerous growth factors. Eight papers (16%) observe bone grafting complications that are frequently connected to the iliac bone graft donor site. The mechanism of action of the bone graft method in the acceleration of bone healing is clearly demonstrated in seven papers (14%); additionally, the same numbers of papers (seven) are focused on proving various techniques in applying bone grafts (Figure 1 ). Figure 1 The contents of the top 50 papers. The contents are techniques, bone tissue engineering, mechanism of action (MOA), complications, and types of bone graft (BG). Discussion Our study recognizes the topmost 50 research papers published on bone graft based on the number of citations recognized in several scientific studies. This research validates a wide range of valued information regarding the authors, topics, and time periods that have had a deep impact on the orthopedic specialty. It records the changes in information over 45 years. In this paper, the citation number was nominated as the marker of effect. This has been carried out for several further surgical specialties including numerous orthopedic topics. Citation analysis, although controversial, allows for the measurement of peer recognition and suggests insights into the readership of the article [ 61 ]. Regrettably, the citation number does not directly associate with study quality. Nevertheless, a high citation number specifies that various researchers have found an article beneficial and its material worthy for inclusion and more discussion in their work. The 50 uppermost cited articles on bone graft were cited 33, 895 times. The highest seven papers, which were cited more than 1, 000 times, according to absolute numbers were cited at nearly 9, 000 times. These numbers are higher than the uppermost cited papers in the numerous orthopedic fields such as hip and knee arthroplasty and oncology [ 62, 63 ]. This is even more obvious, when compared to the uppermost cited papers in hand or shoulder surgery [ 4, 5 ]. The most-cited paper illustrated the mechanism of action of platelet-rich plasma in improving the usefulness of bone grafts by creating a higher concentration of human platelets and platelet-derived growth factors (1998) issued in the Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontics [ 11 ]. This study has been cited 1, 862 times with a mean citation number of 38. 00/year. In this paper, Marx reached an assumption that the addition of platelet-rich plasma to various bone grafts augmented the radiographic maturation rate 1. 62 to 2. 16 times when compared to bone grafting without platelet-rich plasma [ 11 ]. The second maximum-cited paper was by Younger Edward M (1989) about complications at bone graft donor sites published in the Journal of Orthopaedic Trauma. This research was cited 1, 461 times (47. 13 citations/year). Younger studied the medical records of 239 patients with 243 autogenous bone grafts taken on to document the morbidity at the donor sites. He stated that the general major complications were deep infection, prolonged wound drain, hematomas collection, reoperation, pain lasting for more than six months, severe sensory loss, and unsightly scars, while the minor complications comprised superficial infection, minor wound problems, temporary sensory loss, and mild or resolving pain. He observed that there was a much higher complication rate if the incision used for the surgery was also the same incision used to harvest the bone graft [ 12 ]. A whole of 12 countries contributed to the uppermost 50 articles with the majority derived from the USA. Forty-four papers were created from countries where English is the first language. All countries characterized on the list are first-world countries with a large health-care expenditure [ 64 ]. Parallel results have been realized in other fields where the USA led most positions [ 3, 65, 66 ]. Remarkably, the uppermost five articles were published in a 23-year gap from 1989 to 2012. Consequently, they have had significant time to merge these top citation numbers and this appears to be a crucial factor in their top positions. When we investigate the mean citation number of the topmost two articles, their citation densities are obviously high at 83 and 47. 13 correspondingly. Though, the uppermost citation density is noticed in the fifth paper (Recent advances in bone tissue engineering scaffolds) at 139. 38 citations/year [ 15 ]. This recommends that these papers are highly significant in the field. Nevertheless, a limitation in mean citation number does not signify the progression of a paper's influence over time. For example, a paper that was published three decades ago about the free vascularized bone graft: A clinical extension of microvascular techniques by Professor Geoffrey Ian Taylor who was particularly recognized for his pioneering research in microsurgery and bone grafting and received extensive acknowledgment and frequent citations at that time may still hold a high mean citation number despite not being referenced for many years [ 17 ]. O’Neill (2014) recommended that the mean citation number may in fact be effective in evaluating the proximity of impact a paper has, when comparing articles from diverse time periods [ 67 ]. There are an additional number of boundaries related to this type of research documented by various authors. The Scopus search engine used in this work extends from 1996 to the present day. Hence, any articles published before this date will not be involved in our study, which likely results in numerous classic research articles being excluded. Citation analysis also brings with it some intrinsic faults. It does not take account of biased citing, self-citation, formal or informal influences not cited, technical limitations of citation indices, and not being able to add publications if not indexed in Scopus [ 68 ]. Alternative metrics, or Altmetrics, assess the influence of scholarly materials via online metrics, with an emphasis on data arising from social media outlets, for instance: mentions, views, shares, download, saves, tweeting, tags, and comments. Altmetrics will certainly provide a complimentary measurement through the internet to traditional citation metrics, which will certainly become an alternative dimension whereby the reach of a journal article can be evaluated [ 69 ]. Conclusions The scrutiny of the uppermost 50 articles connecting to bone grafting has emphasized the most significant papers in the field. These cover a wide range of issues including categories, management, and mechanism of action of bone grafting. Citation number was used to detect the influence of these papers. Although this may not directly associate with study quality, it does provide an insight into the effect that a research paper has had on the scientific community. This list may prove priceless to surgeons involved in the treatment of patients who need to use bone grafting in orthopedic surgeries, especially in replacing bone defects and motivating fracture healing and those actively advancing the progress of the field.

10. 7759/cureus. 25121

Cureus

Meniscus Tear: Pathology, Incidence, and Management

Meniscus tears are a common orthopedic pathology and planning a single, effective treatment is challenging. The diagnosis of meniscal tears requires detailed history-taking, physical examinations, special diagnostic tests, and most likely magnetic resonance imaging (MRI) to confirm the lesion. A good understanding of the meniscal structure including vascularity, zones, function, and affected movements with associated symptoms plays a crucial role in establishing an optimal management plan. A careful assessment of the patient's characteristics, comorbidities, post-repair rehabilitation, and patient’s overall function and satisfaction are also important for ideal management. While conservative management is commonly implemented and the only option for certain patients, partial meniscectomy remains to be the most performed treatment procedure. However, partial meniscectomy is no longer the first-line therapy due to the limitation of certain patient characteristics and side effects in the long run. Instead, meniscal repair has been shown to have better long-term outcomes and is therefore recommended for all tears, especially for young patients with acute traumatic lesions. Tissue engineering has been of high interest in the current research with promising therapeutic results. This review critically evaluates and compares the management of meniscal tears with surgical versus comprehensive management using the current literature.

Introduction and background A meniscus tear is one of the most common sports-related injuries and often requires surgery due to pain and dysfunction of the knee [ 1 ]. Initially, menisci were described as functionless vestigial remnants and were commonly resected [ 1 ]. Increasing scientific research in recent decades discovered the important role of menisci in anatomical and biomechanical functions, including load sharing, stabilizing, shock-absorbing, and lubrication [ 1 ]. The incidence of meniscal tears is estimated to be 60 per 100, 000 population approximately and the incidence of meniscal-related injuries is rising significantly due to increased sports participation and advanced diagnostic tools [ 2 ]. This made meniscus surgery one of the most common orthopedic operations with an incidence of 17 procedures per 100, 000 in the United States [ 2 ]. Studies found that patients with meniscal injuries have hastened cartilage wear, which predisposes them to early degenerative changes and poor long-term function [ 2 ]. In fact, more than 75% of patients with symptomatic osteoarthritis have known meniscal injuries [ 2 ]. The treatment options for meniscal tear include nonoperative management, meniscal repair, or meniscectomy [ 2 ]. Surgical management is the mainstay treatment of most meniscal tears. The first open surgical repair of a meniscal tear was performed by Annandale in 1885 followed by the development of various arthroscopic techniques [ 2 ]. Total meniscectomy was the gold standard management of meniscal tears until the 1970s due to the lack of understanding of the vital role of menisci [ 2 ]. However, it became prominent that the patients who underwent meniscectomy developed femoral condylar flattening and joint space narrowing, leading to degeneration [ 3 ]. Since then, studies confirmed that the meniscus is an important weight-bearing structure and its absence leads to knee instability and osteoarthritis [ 1 - 3 ]. Therefore, meniscal injury repair and preservation have become a big part of orthopedic research and have advanced significantly only in the last few decades [ 1 - 3 ]. A study by Abrams GD et al. found that the number of meniscal preservation procedures has doubled in the last five years [ 4 ]. Changing meniscal treatment from resection to preservation, either surgical or conservative, has shown promising results with shorter recovery times and functional outcomes [ 2 ]. Arthroscopic partial meniscectomy (APM) has been the gold standard treatment for meniscal injuries in the last few years around the globe [ 2, 5 ]. More than 350, 000 APMs have been performed from 2005 to 2011 in the United States [ 5 ]. However, recent studies suggest that the outcomes of APM are not significantly different from the outcomes of placebo surgery, and the risk of undergoing total knee replacement is three times higher among patients with previous APM [ 2, 5 ]. A meniscal repair is an effective alternative option to heal meniscal lesions without the adverse effects of partial and total meniscectomy [ 2 ]. Although the short-term outcome of meniscus repair shows less than a 10% failure rate at a two-year follow-up, the long-term failure rates have stayed consistent between 23% and 30% despite using various techniques [ 2 ]. Nonetheless, meniscal repair still is the preferred method with lower rates of radiographic degenerative changes compared to meniscectomies [ 2 ]. Conservative management can be the preferred option for certain patients with smaller tears, advanced osteoarthritis, and those who are unable to undergo operations. A number of clinical trials have concluded that the surgery is not superior to conservative management in degenerative meniscal lesions [ 5 ]. Over the years, clinical studies demonstrated that patients with knee osteoarthritis who underwent arthroscopic treatment showed more symptomatic relapse and lower satisfaction rates compared to patients with traumatic meniscal lesions [ 5 ]. Choosing the optimal management for meniscal tear remains controversial due to a lack of evidence directly comparing these treatment options. However, it requires a thorough assessment with several factors, including patient age, comorbidities, characteristics of the tear, and symptoms, to establish the appropriate treatment. This paper aims to review the current literature about the meniscus anatomy, pathology, incidence, and management options of meniscal tears, specifically to compare surgery to conservative management. Review Anatomy of the meniscus In the knee, menisci are wedges of fibrocartilage that are located between the tibial plateau and femoral condyle [ 6 ]. The most plentiful part of the menisci is collagen (75%), predominantly type I collagen (>90%), despite the fact that it additionally contains types II, III, V, and VI [ 6 ]. Collagen strands are organized for the most part along a longitudinal or circumferential bearing. The microanatomy of the meniscus is composed of thick fibrocartilage, which is called fibrochondrocyte, as it is a combination of fibroblasts and chondrocytes [ 7 ]. These cells are liable for the blend and support of the extracellular fibrocartilaginous matrix. The extracellular network additionally incorporates proteoglycans, glycoproteins, and elastin [ 7 ]. The larger, semilunar medial meniscus is firmly attached in comparison to the loosely attached, rounder lateral meniscus [ 8 ]. Both the anterior and posterior horns of the menisci attach to the tibial plateau. Anteriorly, a transverse ligament connects the menisci; posteriorly, the meniscofemoral ligament holds the posterior horn of the lateral meniscus to the femoral condyle [ 8 ]. The peripheral meniscus is connected to the tibia by the coronary ligaments. Although the lateral collateral ligament passes closer to the lateral meniscus, it is not attached to it (Figure 1 ) [ 6 ]. Figure 1 Menisci anatomy viewed in situ on the tibia Note. Image from Bryceland JK et al. (2017) [ 6 ] The joint capsule attaches to the entire periphery of each meniscus but is more firmly attached to the medial meniscus [ 6 ]. Popliteal hiatus is the area between the joint capsule and lateral meniscus that allows the popliteal tendon to pass through. Contraction by the popliteus during knee flexion pulls the lateral meniscus posteriorly, staying away from entanglement inside the joint space [ 7 ]. The medial meniscus does not have a direct solid association. The medial meniscus might move a couple of millimeters while the less steady lateral meniscus might move somewhere around 1 cm [ 6 ]. The vascular supply of the knee joint is regulated by the parameniscal capillary plexus whereby the lateral and medial geniculate arteries anastomose [ 6 ]. There are three distinct zones of the meniscus, distinguished based on blood supply: the peripheral vascularized red-red zone (zone 1), the avascular white-white zone (zone 3), and the transition between two called the red-white zone (zone 2) [ 7 ]. Healing and repair of tissue are directly related to vascularization with blood supply to the tissue, hence making the white-white zone prone to degenerative lesions [ 7 ]. Biomechanics and pathology of the meniscus The menisci are responsible for the majority of load transmission across the knee. Some other crucial functions include increasing joint conformity for maintaining fluid lubrication with synovial fluids [ 2 ]. This further maintains congruity between femoral and tibial condyles to assimilate their proper usage. Moreover, energy dissipation, or shock absorption, by the menisci is primarily important when it comes to trauma and high impact load over knee joints [ 9 ]. The knee extension and flexion biomechanics are directly related to the motion of the femoral condyle. During extension, the menisci are displaced anteroposteriorly due to force exerted by the femoral condyle [ 9 ]. During flexion, the menisci deform mediolaterally, maintaining joint congruity and maximal contact area [ 9 ]. The femur externally rotates on the tibia, and the medial meniscus is pulled forward. Meniscus tears can be divided into vertical, horizontal, and complex. Vertical lesions are asymptomatic often and could lead to longitudinal lesions in the periphery of the meniscus [ 2 ]. Horizontal tears could be more fatal, leading to complete cleavage between the meniscal edge layers. Complex lesions are related to degenerative changes in the knee, consisting of both vertical and horizontal lesions (Figure 2 ) [ 1 ]. Figure 2 Different types of meniscus tears Note. Image from Karia M et al. (2019) [ 1 ] Epidemiology All populations are subjected to meniscal tears when their knee is subjected to an external force that causes the knee to twist. The prevalence of meniscal tears is approximately 12% to 14%, with an approximate incidence of 61 cases in every 100000 people [ 10 ]. Acute trauma-related tears are more prevalent in active young populations and those engaging in sports activities. On the other hand, degenerative meniscus tears affect the elderly population, with the peak onset age in men being 41 to 50 years, while in females, it is 61 to 70. The approximate number of cases per year is 850000. The associated orthopedic surgeries to correct the meniscus tears are between 10% and 20% [ 10 ]. Several factors are risk factors for meniscal tears, increasing the likelihood of developing meniscal tears. The non-modifiable risk factors for meniscal tears include sex, where the incidence in men is 2. 5 times more than in women [ 11 ]. Meniscal tears are more in individuals with a biconcave tibial plateau, a discoid meniscus, those with lower extremity alignment, and those with ligamentous laxity. The modifiable risk factors that increase the risks of developing meniscal tears are a high body mass index, certain occupations, such as squatting, lifting and carrying weights, stairs climbing, and athletes, and those engaging in sports-related activities, including footballers, and those playing rugby. Several conditions are associated with meniscus tears. Patients with anterior cruciate ligament (ACL) injuries have increased incidences of having meniscal tears with an approximation of 22% to 86% [ 10 ]. Acute ACL injury was mainly associated with lateral meniscal tears while chronic ACL injury was associated with medial meniscal tears. According to Valdez et al. , although posterior cruciate ligament (PCL) incidence is lower than ACL, approximately 8% of PCL patients develop meniscal tears. Grade III PCL increases the prevalence of meniscus tears [ 10 ]. Diagnostics A diagnosis of meniscus tears involves a detailed history, physical examination, imaging, and special tests. History-taking information about the cause and presentation of the meniscus tears can be identified. A physical examination can diagnose a torn meniscus. The cause of the signs and symptoms of meniscus tears can be identified when the physicians move the affected lower limbs in different positions, watching the patient walk and squat [ 2 ]. Radiological studies are used to generate images to confirm meniscus tears. As the meniscus consists of cartilage, X-rays cannot be used to show a meniscus tear. However, X-ray images rule out other causes associated with symptoms similar to those with meniscus tears. The most common radiographic test used to diagnose meniscus tears is magnetic resonance imaging (MRI). In detecting a torn meniscus, MRI has a specificity of 88% and a sensitivity of 93% [ 2 ]. On MRI, the abnormal shape of the meniscus and high signal intensity contacting the surface edge can be appreciated (Figure 3 ) [ 2 ]. Figure 3 Sagittal MRI of posterior horn medial meniscus horizontal tear Note. Image from Bhan K (2020) [ 2 ] MRI: magnetic resonance imaging McMurray’s test is one of the clinical tests used to diagnose meniscal tears. It is a procedure that involves systemic rotation of the knee by the physician. The test is positive when there is clicking/popping and pain in the knee when the knee is rotated [ 12 ]. Another clinical test is Apley’s grinding test where the patient lies in a prone position with flexion of the knees to 90 degrees. This is followed by the physician's medial and lateral rotation of the knee followed by distraction. Compression, instead of a distraction, follows a repetition of the process. A meniscus tear is diagnosed when a decreased rotation is associated with a more painful knee when the knee is rotated and compressed [ 13 ]. Treatment The fibrocartilaginous meniscus is essential for the musculoskeletal stability of the knee joint [ 14 ]. Damage or loss of this vital structure can lead to significant articulatory morbidity and an accelerated course of osteoarthritis. Therefore, attempts should be made to preserve the meniscus [ 14 ]. Treatment and management of meniscal tears are dictated by multiple factors and include age, the complexity of the tear, tissue quality, the severity of symptoms, etiology (traumatic versus atraumatic tear), and quantified surgical risk [ 14 ]. For acutely painful and swollen knees with a suspected meniscus tear, the initial strategy is to follow the R. I. C. E. (rest, ice, compression, elevation) principle [ 7 ]. Oral medication, such as acetaminophen and nonsteroidal anti-inflammatory drugs (NSAIDs), can also be prescribed to alleviate pain and swelling [ 7 ]. For degenerative tears and simple traumatic meniscal tears, additional conservative management involves the use of a knee brace, activity modification, physical therapy, and quadriceps strengthening exercises. Physical therapy should be initiated early and should begin with pain-free range of motion exercises with progression to weight-bearing exercises as tolerated [ 7 ]. Endurance activities like biking and swimming that decrease mechanical load across the knee joint should also be encouraged [ 7 ]. In patients who decline surgery or in whom chronic NSAID use/surgery is contraindicated, intra-articular cortisone or hyaluronic injections can also be provided intermittently, every two to three months, for short-term relief [ 15 ]. For simple traumatic/degenerative tears, it is reasonable to continue conservative management for about four to six weeks [ 7 ]. A study found that quadriceps-strengthening exercises three times a week for 10 weeks have improved knee function by 35% in patients with osteoarthritis [ 16 ]. Another randomized controlled trial compared APM followed by supervised exercise and exercise therapy alone in patients with degenerative meniscus tear [ 17 ]. They found significant improvement in both groups after eight weeks but no significant difference in outcomes between the groups [ 17 ]. Therefore, the author suggested that conservative treatment with supervised exercise should be the first-line management [ 17 ]. However, if mechanical symptoms persist, are disabling, and significantly affect the quality of life, surgical intervention should be considered [ 7 ]. Generally, surgery is favored if the case scenario involves any of the following: (1) red zone tear, (2) complex and extensive meniscal rips > 1 cm, (3) young healthy candidates with age<40 years old, (4) acute tears that occurred <6 weeks, and (5) the presence of a concurrent ACL injury [ 14 ]. The current surgical approach to managing meniscal tear encompasses meniscectomy, meniscal repair, and meniscal reconstruction [ 14 ]. Meniscectomy A meniscectomy (or meniscal resection) can be done completely or partially via an open or arthroscopic approach [ 14 ]. In the current era, total meniscectomy is almost never performed due to well-established side effects, most importantly early-onset osteoarthritis [ 14 ]. An APM is more commonly performed because it is minimally invasive, associated with shorter recovery time, and lower morbidity comparatively [ 14 ]. Indications for APM include radial white-white zone (non-perfused) meniscus tears and degenerative meniscus injuries, which are unresponsive to conservative management [ 15 ]. Nevertheless, osteoarthritis still occurs in the long run. Extensive clinical research shows no significant long-term benefits of APM over non-operative management of traumatic as well as atraumatic meniscus tears [ 15 ]. Factors associated with poor outcomes of APM included obesity, female gender, and advanced osteoarthritis [ 15 ]. Therefore, per current guidelines, APM is no longer the first-line therapy and should be undertaken only in selective patients with non-repairable meniscal tears and those with persistent mechanical symptoms beyond 3 months [ 14 ]. Meniscal Repair A meniscal repair like meniscectomy can also be performed via an open surgical or arthroscopic approach [ 15 ]. Arthroscopic repair predominates over open repair due to a lower risk of neural damage. Tear patterns and adequacy of vascularity should be accessed before proceeding with meniscal repair [ 15 ]. Repair is most advantageous in the setting of acute traumatic meniscal tears within the well-perfused, peripheral red-red zones of the meniscus. Furthermore, longitudinal/horizontal and vertical tears are more amenable to repair compared to radial tears [ 14 ]. However, a repair can still be attempted with radial tears in partially perfused red-white zones [ 14 ]. Arthroscopic meniscus repair can be achieved via inside-out, outside-in, and all-inside techniques [ 14 ]. The inside-out approach is associated with the greatest success rates and is the golden standard in meniscal repair [ 14 ]. In this approach, the sutures are passed from inside the knee to an extra-capsular area through the extra-articular incision and a knot is then secured over the joint capsule [ 14 ]. The inside-out technique is commonly used for posterior horn meniscal damage [ 14 ]. The outside-in technique is more commonly used for anterior horn tears. In this approach, the spinal needle is passed through the meniscal rip in an outside-in manner [ 15 ]. The suture is passed through the arthroscopic portal once the tip of the needle is visible. The suture is pulled back after an interference knot is tied at the end. The operation is continued until the tear is stabilized [ 15 ]. The all-inside technique is most beneficial in case of extreme posterior meniscal rips. Instruments used for repair (such as screws, staples, etc. ) are commonly made of bioabsorbable compounds like poly-L lactic acid [ 15 ]. These implants are deformable and hence lowering the potential of chondral erosion during weight-bearing [ 15 ]. Although arthroscopic techniques aim to lower the risk of neurovascular problems, inadvertent damage can still occur in all the above [ 15 ]. Meniscal Reconstruction The least commonly performed is the meniscus reconstruction surgery, in which attempts are made to replace missing/resected components of the native meniscus with functional ones [ 14 ]. The aim of this procedure is to re-establish the functionality of the knee joints and mitigate degenerative processes that would otherwise result from poor knee biomechanics [ 14 ]. Reconstruction can be performed with the use of either meniscal scaffolds or via meniscal allograft transplantation (MAT) [ 14 ]. MAT involves the transplantation of preserved meniscus allograft. Meniscal scaffold surgery, on the other hand, uses synthetic biodegradable porous structures to fill meniscal defects [ 14 ]. The high porosity of the scaffolds allows vascular tissue to grow within them which provides additional reinforcement [ 14 ]. Finally, postoperative rehabilitation and gradual return to normal activity are encouraged to optimize outcomes with all of the above surgeries [ 14 ]. Cell-Based Tissue Engineering Tissue engineering (TE) is an upcoming technology that has the potential for use in the treatment of meniscal tears [ 18 ]. Regeneration of the meniscus is the primary concept employed, which is met by stimulating the differentiation of cells into tissue that has phenotypical features identical to the native meniscus [ 18 ]. This technology can be used not only for the repair of meniscal tears but also for the regeneration of a partial or complete meniscus following a partial, subtotal, or total meniscectomy [ 18 ]. The most common cell types used in meniscal TE include meniscal cells, articular cells, and mesenchymal stem cells (MSC) specifically, embryonic, bone marrow, and synovium derived [ 18 ]. These cells can proliferate and differentiate into cartilaginous cells with the deposition of extracellular matrix (ECM) found in the native meniscus [ 18 ]. As per a study by Marsano et al. , MSC led to tissue samples with the greatest resemblance to the human meniscus particularly, with larger depositions of ECM including glycosaminoglycans and collagen types I, II, and IV [ 18 ]. Hence, making MSC an attractive cell type for meniscal TE [ 18 ]. Additionally, MSCs also exhibit multilineage differentiation and self-renewal capacity, which make them the perfect substrate for regeneration [ 18 ]. They can also re-establish joint homeostasis and enhance tissue repair via the secretion of paracrine and anti-inflammatory factors [ 18 ]. Furthermore, bone-marrow-derived MSC remains the main cell source of MSC-based TE. Primarily, because they have a higher osteogenic, chondrogenic, and adipogenic potential and because they can be harvested with ease and limited morbidity [ 18 ]. Importantly, the bone marrow also encompasses hematopoietic stem cells; which are distinguished from MSC due to the presence of cell surface antigens, including CD45 and CD34 [ 18 ]. Moreover, growth factors have also been used in TE to further promote cell proliferation/differentiation and deposition of ECM while inhibiting tissue metalloproteinase and improving vascularization of the engineered tissue [ 18 ]. The commonly used growth factors include transforming growth factor β (TGF-β), hepatocyte growth factor (HGF), insulin-like growth factor 1 (IGF-1), fibroblast growth factor 2 (FGF-2), and platelet-derived growth factor (PDGF) [ 7 ]. There are two ways of administering MSC and growth factors for the purpose of meniscal TE: (1) intra-articular injections and (2) seeding onto a biomaterial meniscal construct called a scaffold. The main advantage of IA injections is that they can be performed with limited morbidity, can be repeated, and can prevent the systemic diffusion of injected cells [ 18 ]. In terms of a scaffold, the most ideal characteristics would be: (i) cell-instructive, allowing proliferation and differentiation of the seeded cells, (ii) architectural mimic of the native meniscus, (iii) resilience to mechanical forces acting on the joint while allowing deposition of ECM, (iv) biocompatibility to prevent an immunogenic reaction, and (v) easily implantable [ 18 ]. As per Chiari et al. study, both natural allogeneic and synthetic scaffolds have been used but are singly not sustainable [ 18 ]. Calling for the need for a hybrid coupling of the biocompatibility of allogeneic scaffold with the mechanical strength of synthetic scaffold in order to tailor a more sustainable scaffold [ 18 ]. As per available literature, preclinical studies on cell sources, scaffolds, and growth factors have shown several limitations along with no clear benefits; thereby making the clinical practice of TE controversial [ 18 ]. Advancements in genetics and bioengineering are, therefore, required to solve the clinical challenges [ 18 ]. Conclusions The menisci were thought to be a functionless structure and were resected completely decades ago. Since the menisci were actually known to play an essential role in the biomechanics of the knee, the orthopedic research on menisci shifted from resection to preservation. Today, a meniscus tear is a very common injury and its incidence is increasing in all ages either due to trauma or osteoarthritis. The recent advances in ongoing research studies and clinical trials contribute to the diagnosis and management of the condition. A thorough investigation of the patient history, physical examination, as well as meniscal tear characteristics, will facilitate a better understanding of the pathogenesis and therapy. The vascular supply of the knee joint plays a major role in the healing and repair of the tissue properly. Therefore, identifying the correct location of the lesion will contribute to an optimal treatment plan and postoperative rehabilitation. Although conservative management is preferred for some patients and has its role in a patient's functional improvement, surgery remains the main treatment for meniscal tears. Despite having the advantage of being minimally invasive, fast recovery, and low complications, partial meniscectomy still leads to osteoarthritis in the long run. Arthroscopic meniscal repair has been popularizing in the last decade and became the procedure of choice for meniscal lesions. Although meniscal repair has shown promising results and a low short-term failure rate, the long-term failure rate is reported to reach up to 30% consistently. Cell-based tissue engineering is a newer therapy with a concept of regeneration of the menisci by stimulating other cells. Studies have limitations and the therapy is yet to be proven to be beneficial in both the short and long run, thus more evidence-based research studies are required. Greater efforts in developing modern imaging and technologies will continue to provide advanced tools to further develop diagnostic and treatment interventions.

10. 7759/cureus. 25329

Cureus

Is a Bioengineered Heart From Recipient Tissues the Answer to the Shortage of Donors in Heart Transplantation?

With the increase in life expectancy worldwide, end-organ failure is becoming more prevalent. In addition, improving post-transplant outcomes has contributed to soaring demand for organs. Unfortunately, thousands have died waiting on the transplant list due to the critical shortage of organs. The success of bioengineered hearts may eventually lead to the production of limitless organs using the patient’s own cells that can be transplanted into them without the need for immunosuppressive medications. Despite being in its infancy, scientists are making tremendous strides in “growing” an artificial heart in the lab. We discuss these processes involved in bioengineering a human-compatible heart in this review. The components of a functional heart must be replicated in a bioengineered heart to make it viable. This review aims to discuss the advances that have already been made and the future challenges of bioengineering a human heart suitable for transplantation.

Introduction and background According to reports, currently, 64. 34 million people suffer from heart failure worldwide [ 1 ]. Furthermore, the number of patients with end-organ heart failure is rising, leading to an all-time high in the number of people waiting for an organ transplant [ 2 ]. Several strategies have been devised to increase this strained supply of heart for transplantation, including expanding donor criteria [ 3 ], use of advanced perfusion machines such as organ care systems (OCS) to improve viability [ 4 ], use of normothermic regional perfusion (NRP) in donor from cardiac death (DCD) hearts, and xenotransplantation. Recently, the focus has shifted to new procedures using regenerative cells, angiogenesis factors, biological matrices, biocompatible synthetic polymers, and online registry systems that utilize bioimplants. These advanced technologies are collectively referred to as tissue engineering [ 5 - 8 ]. Ultimately, the goal is to grow a heart de novo. In addition to the unlimited organ supply, the new organ would be antigenically identical to the recipient as the recipient’s cells would be used, eliminating the need for immunosuppressive agents. Even though bioengineering a fully functioning heart is in its infancy, huge strides have been made in achieving this goal. Scientists have been able to bioengineer models of the heart, lungs, pancreas, liver, and kidney. An important strategy for supporting the recipient’s cells and creating an autologous tissue/organ is to create a mechanical, geometrical, and biological environment that closely mimics the native organ’s properties. The breakthrough in growing an artificial heart was the invention of the decellularization of extracellular matrix (ECM), which maintains the native vascular network [ 9 ]. Numerous tissues and organs have been engineered using decellularization, including livers [ 10 ], lungs [ 11 ], kidneys [ 12 ], corneas [ 13 ], bladders [ 14 ], vasculature [ 15 ], articular cartilage [ 16 ], intestines [ 17 ], and hearts [ 18 ]. There has been some success in engineering a heart in the lab. Although technological innovations and biological model systems have resulted in great progress, constructing such complicated tissue structures effortlessly remains a challenge. This review aims to outline the techniques involved in bioengineering a heart in the lab and the challenges involved in developing it into a viable organ for transplantation (Figure 1 ). Figure 1 Outline of the processes involved in bioengineering a heart. The figure outlines the process of bioengineering a heart in the lab. The process starts with the decellularization of a human or animal heart which creates a decellularized extracellular matrix (ECM) scaffold. This ECM scaffold is then reseeded with cells (recellularization process) and then cultured in a bioreactor for the growth and migration of cells throughout the ECM with the use of growth factors and various stimuli. This would, hypothetically, create a functioning “bioengineered heart” that can be transplanted into a recipient. Review Structure of the heart The human heart comprises various cells, each specialized to perform a specific task. A human heart contains roughly 2-3 billion cardiomyocytes, making up only about one-third of its total cells [ 19 ]. Additionally, other cells include endothelial cells, fibroblasts, and specialized conducting cells like Purkinje fibers. On top of that, structural scaffolds support the functions of cells arranged into structures, such as vessels, muscles, and nerves. These scaffolds mainly consist of polysaccharides and proteoglycans embedded in complex sugars and chemokines matrix, allowing the heart to coordinate its mechanical and electrical functions [ 20, 21 ]. Sprawled around this is a collection of protein fibers such as collagen and elastin, which confers mechanical strength to the heart and allow for the constant loading and unloading forces [ 22, 23 ]. Thus, it is necessary to construct a scaffold around which the specialized cells can grow and maintain vitality through blood perfusion to recreate a functioning heart in a laboratory [ 24 ] (Figure 2 ). Figure 2 Components of a functional heart. The figure depicts the components of a functional heart. These components can be stratified into three parts. The heart has a myriad of cells. The heart is composed of predominantly cardiomyocytes along with endothelial cells, smooth muscle cells, and cardiac fibroblasts among others. The cells populate a scaffold of extracellular matrix (ECM) which is composed of protein fibers such as collagen and elastin surrounded by proteoglycans, laminins, and fibronectins. This gives the heart its biophysical properties like mechanical strength to undergo rapid muscle movement during the cardiac cycle following an electrical activity. A bioengineered heart must have all three components of the heart to be deemed functional. Types of ECM scaffolds available Extracellular matrix (ECM) and cells in an organ display a “dynamic reciprocity, ” whereby the ECM constantly adapts to the demands of the cells [ 25 ], and selecting the appropriate scaffold is the key component for growing a viable organ in the lab. Researchers have also studied various synthetic scaffolds as potential surrogates for the ECM, but none can replicate its intricacy or structure compared to native ECM. It is possible to “vascularize” synthetic materials such as polylactic acid (PLLA) and polylactic glycolic acid (PLGA) and to produce them consistently [ 26, 27 ]. The significant advantage of synthetic ECM is its production scalability as it does not require to be harvested from living tissue, but these do not match the native myocardium’s tensile strength. Hydrogels have also been studied extensively and even accepted by the Food and Drug Administration for drug delivery and adjunct for cell therapy. Hydrogels consist of a cross-linked hydrophilic polymer matrix with over 30% water content [ 28 ]. However, they have poor cell retention [ 29 ] or poor tensile strength [ 30 ]; hence, they are not feasible as a primary scaffold for constructing an organ. Decellularizing the whole heart and leaving the ECM serves as a potential solution to this problem with the particular advantage of having a balanced composition of all the proteins present physiologically [ 31 ]. Creating the “ideal” scaffold: decellularization of the heart The Badylak laboratory developed the first technique for decellularizing tissue [ 32 ]. This process involved the removal of the cell, leaving only the ECM, which retained its composition, architecture, and mechanical properties. There are several methods for removing cells from the ECM. These methods include physical methods (e. g. , freeze/thaw cycles), enzymatic degradation (e. g. , trypsin), and removal by using chemicals (e. g. , sodium dodecyl sulfate) [ 33 ]. Ott et al. noted that decellularization could be achieved with different detergent solutions. Comparative studies on decellularization methods have mixed results regarding the superiority of different techniques [ 34 - 37 ]. Based on the results, the sodium dodecyl sulfate (SDS) solution was found to be the best [ 18 ]. However, a few studies have suggested that SDS treatment causes degradation of the ECM with a reduction in elastin, collagen, and glycosaminoglycans (GAG) content [ 34 ]. The decellularization process utilizes 1% SDS perfused through the coronary circulation, followed by washing it with de-ionized water and subsequently 1% Triton-X-100 (Sigma). Finally, the organ remnant is washed with phosphate-buffered saline (PBS) wash buffer, antibiotic, and protease, leaving a decellularized ECM [ 38, 39 ]. Using this technique, they decellularized the heart, reseeded it with neonatal cardiac cells, and grew the first beating rodent heart in the lab [ 18 ]. Decellularized tissue provides a dynamic environment for the orientation and coupling of cells and facilitates the exchange of nutrients and oxygen throughout the depth of the tissue. Moreover, this process efficiently removes both allogeneic and xenogeneic antigens, possibly preventing the need for immunosuppressants [ 33 ], which is especially important as one of the causes of heart failure in transplanted hearts is myocardial fibrosis from chronic rejection [ 40 ]. This process can be potentially avoided by using a decellularized heart to generate an ECM scaffold which can then be repopulated using the recipient’s cells. Sources for creating ECM scaffolds Researchers have used animal heart ECM and human heart ECM scaffolds to provide this decellularized ECM scaffold. The porcine heart has often been deemed suitable for its similarity with the human heart [ 41 ]. As decellularization removes most of the cells, much of the antigen load is removed. However, the porcine heart ECM contains α-1, 3-galactose epitope (α-gal), which can stimulate an immune response [ 42, 43 ]. One way to circumvent this is to use pigs lacking α-gal epitope, but this technique needs further research. Another possible problem with using a porcine heart is the possible risk of horizontal transmission of porcine viruses like the porcine endogenous retrovirus, cytomegalovirus, HSB, circovirus, etc. [ 44, 45 ]. Although a few tests can detect the presence of these viruses, they have poor sensitivity, and hence further work has to be done [ 46 ]. A cadaveric heart that is unfit for transplant can also be used to harvest an ECM scaffold [ 47 ]. The only drawback to this is that it may not always be possible to achieve the desired level of tissue engineering fidelity with these matrices because they may be damaged or diseased. Moreover, there is an assumption that they are superior for the growth and differentiation of human cells, but there is no robust evaluation to support this assumption. The method for decellularization of the cadaveric human heart is similar to that of other animals, utilizing 1% SDS and 1% Triton X-100, with the only difference being a longer perfusion time for these chemicals [ 48, 49 ]. Recellularization of scaffolds These cells are highly specialized and terminally differentiated, and hence, they do not proliferate normally. Therefore, to repopulate a human-sized scaffold, autologous human cardioblasts must be isolated or expanded in large quantities. Hence, for the recellularization of ECM, a method of inducing progenitor cells had to be devised. Thus, the discovery of methods to reprogram or induce adult cells into pluripotent stem cells was a significant milestone in stem cell biology and tissue bioengineering [ 50 - 52 ]. Once we have the cells for repopulation of ECM, recellularization is required to achieve a functional organ product for implantation. For recellularization to be achieved, choosing appropriate cell sources, seeding cells optimally, and cultivating them using organ-specific cultures are needed [ 24 ]. Cells from fetuses and adults, embryonic stem cells (ESCs), mesenchymal stem cells (MSCs), and induced pluripotent stem cells (iPSCs) have all been used [ 24 ]. Obtained with ease and ethically, stem cells from bone marrow stroma or adipose tissue (MSC) have shown promise as the ideal cells for recellularization [ 53 ]. In addition, human somatic cells can be reprogrammed to produce iPSCs, and they exhibit properties similar to ESCs [ 54 ]. A potential solution to the problem of getting a large number of human cells for tissue engineering or other regenerative medicine approaches is the ability to produce iPSCs from readily available autologous cells such as fibroblasts or blood cells [ 55, 56 ]. The only drawback to using iPSCs is the possibility of teratoma formation due to its pluripotent nature [ 48, 57 ]. However, the potential solution to this problem is to allow controlled differentiation toward a cardiac lineage before implantation into the ECM [ 58 ]. Although previously any attempts to produce iPSCs would result in karyotype instability [ 59 ], recent advances have been made with iPSCs maintaining chromosomal integrity [ 60 ]. These advances have ushered a step forward in the pursuit of creating viable organs in the lab. Cell seeding techniques depend on the type of organ being engineered, and, for the heart, it usually involves seeding by perfusion through the vascular tree [ 24 ]. This step is called re-endothelization and is usually the first step to recellularization. A dynamic communication between endothelial cells and cardiomyocyte populations occurs via direct cell interactions and the secretion of various factors [ 61, 62 ]. It is evident from multiple reports that seeding endothelial cell populations and cardiomyocyte populations simultaneously provides functional benefits that aid in maintaining the recellularization process [ 63 ]. Interestingly, endothelial cells have also demonstrated the ability to differentiate into cardiomyocytes in other cardiomyocyte cells [ 64 ], which may aid in more efficient recellularization. Moreover, besides the advantage, the recellularization of both the vascular tree and the heart parenchyma must be uniform to prevent two key issues in the heart, namely, thrombogenesis [ 65 ] and arrhythmogenesis [ 66 ]. Improved cell concentration and diffusion over the scaffold can be achieved by optimizing the mechanical environment, scaffold coating, and cell perfusion systems by using multiple perfusion routes simultaneously, which for the heart involves both direct intramyocardial injections and perfusion of the vascular tree [ 67 ]. However, the potential problem with intramyocardial injections is that even though the injection site shows dense cellularity, the cells are generally poorly distributed throughout the scaffold [ 58 ]. Moreover, sequential injections of cardiac cells will likely be required to rebuild the chamber parenchyma, which may compromise matrix integrity [ 48 ]. Nevertheless, given that cardiac cells include fibroblasts, in which ECM is produced and secreted, there is a possibility that endogenous matrix repair may occur after cell seeding to help resolve this issue [ 62 ]. While sourcing cells for recellularization using stem cells is a work in progress, multiple studies have explored ways to develop mature cardiomyocytes derived from iPSCs that are more physiologically similar to native cardiomyocytes [ 68, 69 ]. One of the most recent cardiac constructs was engineered using PSC-derived cardiac cells in a ratio of equal cardiomyocyte and noncardiomyocyte cells, cultured in serum-free media [ 70 ]. Cardiomyocytes cultivated in this method were elongated, had organized sarcomeres and distinguished bands, and exhibited increased contractility [ 70 ]. It is encouraging to see these results that stem cells can be used to produce cardiomyocytes similar to native mature cells, reinforcing the notion that stem cells can be a cardiac cell source. Growing the heart in a bioreactor After enough cells have been seeded onto an organ scaffold, cell culture is required. A bioreactor is required for perfusion and provides a nutrient-rich environment that encourages organ-specific cell growth [ 24 ]. Bioreactors should allow nutrient-rich oxygen to be pumped with adjustable rates of flow and pressure and monitor and control the pH and temperature of the media. Moreover, mechanical stimulation is also an essential component for engineering organs of the musculoskeletal and cardiovascular systems [ 71 ]. A wide range of mechanical properties is employed in the design of bioreactors, including substrate stiffness and dynamic changes in stiffness throughout culture, pulsatile flow, and providing stretch to enhance cell maturation, alignment, and generation of force in engineered constructs [ 72 ]. Presently, there are several types of bioreactors available, with Radnoti [ 73 ] and BIOSTAT B-DCU II [ 74 ], to name a few. In addition, there has been an increase in bioreactor designs incorporating real-time monitoring to assess the status of engineered tissues. These designs may incorporate biochemical probes to assess transmural pressure changes or sampling ports to test cells’ viability and biochemical composition after recellularization [ 75, 76 ]. The incorporation of sampling methods within bioreactor designs will keep constructs sterile, allowing for modifications in stimuli to be made while maintaining a closed system, and providing researchers with valuable feedback on cell responses throughout bioengineering. Further research is being conducted to make bioreactors that can be used to maintain the perfect milieu for growing these bioengineered tissues and organs. Evaluating the organ for functionality For an organ to be viable for transplant, three things must be ensured: sterility of the process, structural integrity, and, lastly, patency for surgical anastomosis. Biological tissues are sterilized by gamma radiations or peracetic acid at low concentrations before the ECM is repopulated with cells [ 77 ]. Once the cells are added, antibacterial, antifungals, and other antibiotic drugs can be utilized. It is re-evaluated for integrity before the ECM is recellularized and only gets the green light for cell seeding if structural integrity is maintained. Interestingly, with the aid of endoscopy, decellularized constructs can be easily manipulated and visualized for macro and microstructure defects at the level of chambers, papillary muscle, and valves [ 47 ]. One of the most important aspects of evaluating the integrity of ECM is to check for intact coronary vasculature, which can be done by micro-optical coherence tomography [ 48 ]. Heart constructs engineered in the lab have been demonstrated to undergo cyclical muscular contraction but also have been shown to respond to drugs and exhibit electrical activity. However, electrocardiography analysis of the bioengineered hearts has shown irregular wave morphology due to loss of coupling between cardiomyocytes [ 78 ]. Therefore, it will be crucial to develop continuous monitoring of cardiac electrophysiology, function, and even vascular patency if these artificial constructs can be transplanted into patients. Limitations and future prospects Over the past decade, research in regenerative medicine has enabled us to understand better the challenges associated with developing a bioartificial heart. The first challenge was creating a biocompatible scaffold which has already been resolved with the development of various decellularization techniques, making it possible to generate an anatomically accurate and vascularized heart scaffold. With the advent of newer techniques for iPSC generation of stable karyotype, cell generation is also potentially resolved. Presently, research has to be aimed to address the challenges in reseeding the ECM scaffold. A potential solution might be the advancement in 3D-printed matrixes with embedded cells. However, decellularized ECM remains the gold standard for now as 3D-printed matrixes cannot replicate the complexity and structural integrity of the natural component of ECM. Another potential problem is the creation of a bioreactor that can efficiently maintain the environment required for the growth of cardiac and other differentiated cells around the decellularized ECM scaffold. Constructing organs is no easy feat and involves much technical expertise. Hence, many resources are required in every step of artificially reproducing tissues and organs. Thus, even if bioengineering a heart is a possibility in the near future, it may not be financially feasible to use them for transplantation until the cost of making such constructs is lowered. Additionally, we do not know the long-term viability of such constructs. These constructs use chemicals to decellularize ECM as well as induce the conversion of adult cells into pluripotent cells. Some questions arise on how the complex network of cells and ECM would interact over the long run. The heart is a complex organ that requires a highly specialized conduction system to ensure efficient, coordinated, and purposeful contraction of the heart chambers. Any deviance may lead to fatal arrhythmia or thrombus formation. We are yet to reproduce a perfect conduction system in the lab, let alone test its long-term functionality. Furthermore, the use of induced pluripotent cells also raises the prospect of long-term tumorigenesis and malignancy. Despite rapid advances in bioengineering and artificial hearts, research and clinical trials must be conducted to determine the long-term feasibility of using these organs. Conclusions It is an exciting era for biomedical engineering that carries considerable potential to address damaged organs, either via repair or replacement. The advances in heart bioengineering have been astounding. However, further research must be conducted till a mechanically, electrically, and physiologically well-coordinated organ can be constructed and ultimately transplanted into patients needing it. To propel the field forward in the quest for creating unlimited immunotolerant grafts, a coordinated approach should be fostered among researchers, clinicians, regulatory bodies, and society.

10. 7759/cureus. 25997

Cureus

Application of Collagen-Based Scaffolds for the Treatment of Spinal Cord Injuries in Animal Models: A Literature Update

SCI is regarded as one of the most devastating central nervous system (CNS) injuries, exhibiting an alarmingly rising incidence rate, indirectly connected with the expansion of the global economy. The consequences of SCI are multidimensional: SCI injuries may result in permanent voluntary motor dysfunction and loss of sensation while incurring heavy economic and psychological burdens as part of the treatment. Thus, it is crucial to develop effective and suitable SCI treatment strategies. Collagen-based scaffold application is one of the most promising methods of SCI treatment. This review compiles newer bibliographical data regarding the application of collagen scaffolds for the treatment of Spinal cord injury (SCI) in animal models. Recently, several relevant studies have been carried out using carefully selected animals with similar pathophysiology to humans. In mouse, rat and canine models that have undergone transection or hemisection, the stump connection, the transplanted cell differentiation, and the elimination of glial scar are promising. Also, encouraging results have been found regarding the increased neuronal growth, the decreased collagen deposition, the behavioral recovery, the improved electrophysiology, and the enhanced axonal regeneration.

Introduction and background Spinal cord injury (SCI) is regarded as one of the most devastating central nervous system (CNS) injuries, exhibiting an alarmingly rising incidence rate, indirectly associated with the expansion of the global economy [ 1 - 5 ]. The consequences of SCI are multidimensional: SCI injuries may result in permanent voluntary motor dysfunction and loss of sensation while incurring heavy economic and psychological burdens. Thus, developing effective and suitable SCI treatment strategies is crucial [ 1 ]. SCI causes neurological disabilities as the CNS central axons or the nerve fibres often fail to regenerate [ 4, 6 ] due to chronic inflammatory response, demyelination, and increased proteoglycans [ 6 ]. In mammalian CNS, the leading cause behind the limitation of central axons to regenerate is glial scar formation, which inhibits axonal remodelling and regrowth [ 4, 7 ]. Healthy glial cells support neuronal function and signal transmission [ 8 ]. However, when SCI occurs, the borders of SCI lesion are separated from healthy tissue by mechanical damage and secondarily by a glial scar densely populated by newly hypertrophic and proliferating cells [ 4 ]. These cells are mainly astrocytes, pericytes, non-pericyte perivascular cells, and Schwann cells [ 4 ]. This process hinders the regeneration of neurons’ myelin sheath and function, with glial scar constituting a physical and molecular barrier to developing CNS axons [ 8, 9 ]. Most therapeutic strategies developed in recent years focus on eliminating post-SCI inhibitors of regeneration. Contemporary approaches provide support and guidance toward the regeneration of affected neurons by applying neural scaffolds to the SCI lesion site [ 1 ]. Advanced tissue engineering (TE) technology has paved the way for SCI treatment [ 3, 10, 11 ]. The extracellular matrices (ECMs) allow living cell inoculation, growth, and differentiation, thus promoting axon and fibre regeneration. Matrices are co-cultured with cells and transplanted in the SCI area [ 4 ]. This way, the ECM of the spinal cord can be successfully mimicked, as scaffolds are rich in glycosaminoglycan, a gap-filling polysaccharide of staunch structure [ 4 ]. However, excessive amounts thereof contribute to the uncontrollable development of extensive and grave glial scar. The most suitable solution to combat this issue shall be collagen [ 4 ]. There is a great deal of attention regarding the characteristics of scaffolds, especially the biomaterials used to construct them. Wang et al. stress the importance of biocompatibility for cells, apposite porosity topography and permeability of scaffold materials [ 12 ]. As aforementioned, collagen-based scaffolds are a popular choice for biomaterials used for SCI treatment purposes. Collagen is a protein found in abundance in the ECM, provoking a minimal autoimmune response and promoting cell adhesion, proliferation, migration, and differentiation [ 13 ]. Collagen scaffolds come in various forms, including hydrogel, sponge, or guidance conduit, which serves as an instrument to administer therapeutic drugs and proteins to the SCI site [ 14 ]. This review constitutes a compilation of newer bibliographical data on collagen scaffolds as applied for SCI treatment purposes in animal models to provide a fresh insight into the available bibliography. Review Pathophysiology of SCI The pathophysiology of SCI is often broadly categorized as either “acute impact” or “compression”. Injury resulting from acute impact is essentially a spinal cord concussion that triggers a series of reactions localized in the grey matter, concluding in haemorrhagic necrosis. Grey matter hypoperfusion usually triggers the sequence of events mentioned in this section. Reperfusion and increased intracellular calcium occur soon after the injury and are crucial for the injury outcome. Thus, mechanisms transpiring in the initial stages of injury should be targeted for improved prognosis. Injury resulting from spinal cord compression occurs upon impingement of the spinal cord by a mass, increasing the parenchymal pressure. As far as the white matter is involved, the tissue response is gliosis, demyelination and axonal loss. At the same time, grey matter structures are preserved. A rapid or critical degree of compression will collapse the venous side of the microvasculature, resulting in vasogenic oedema and exacerbated parenchymal pressure. Therefore, leading to swift progression of the SCI process. Collagen scaffolds for SCI treatment In line with the above, several regeneration medicines (RM) and TE studies prove the effectiveness of injected collagen hydrogels. Notably, Sugai et al. in 2015 performed a transection of the spinal cords in rats and compared the effect of collagen-based scaffolds with fibrin and Chitosan. The collagen-based scaffolds showed the most proliferation after the transplant, while the neural stem and progenitor cells survived up to 11 weeks after the transplant [ 15 ]. Breen et al. in 2016 examined the role of injectable collagen hydrogel in administering neurotrophin-3 into rat models undergoing a lateral hemisection of the spinal cord between T9-T10. According to Breen et al. , functional recovery was significantly increased at four weeks postoperatively. The NG2 positive cells expressed within the lesion area were meaningfully reduced compared to the hemisection-only group. Additionally, they reported lowered macrophage and microglial response and reduced glial scarring of the SCI area [ 14 ]. Han et al. in 2017 studied the effect of linear ordered collagen scaffolds loaded with human mesenchymal placenta stem cells in canines with a complete spinal cord transection. Han et al. reported many host cells in the collagen scaffold group, while the new tissue was in a structured form. Contrary, the regenerated tissue was in structural disorder in the ungrafted group. Additionally, they reported an increase in the number of neurons and regenerated axonal fibers penetrating the lesion site with linear order and distinct distribution [ 16 ]. Li et al. have also presented a series of outstanding pieces of research on regeneration and overcoming inhibitory factors following SCI [ 17 - 20 ]. In 2015, they analyzed the delivery of proteins and drugs through scaffolds to enhance post-SCI recovery in rats, proving that collagen scaffolds could support the regeneration of the axons and their remyelination. They also showed that the rats that received the collaged-based scaffolds modified by CBD-EphA4LBD and CBD-PlexinB1LBD promoted the development of more axon fibres through the lesion site. The rats in that group exhibited significant improvements in locomotion from the first week [ 17 ]. In 2016, the same group of scientists presented a porous collagen scaffold with neurotrophic factors CBD-BDNF and CBD-NT3. They reported that the scaffold promoted the outgrowth of the cerebellar granular neurons in vitro. They also found that the cavities caused by the SCI injury were significantly reduced and suggested that the functionalized collagen scaffold could enhance the anti-inflammatory function. Combined with Cyclic Adenosine Monophosphate (cAMP), the scaffold aided the repair of a completely transected spinal cord in a rat model. Still, the locomotion outcome was unsatisfactory, suggesting that rebuilding an injured neural connection is exceptionally complex [ 18 ]. In 2017, Li et al. studied the effect of functional collagen scaffold with Cetuximab in rats and dogs with T8 SCI. The Modified Linear Order Scaffolds with Cetuximab showed a much higher number of newborn and mature neurons in rodents and dogs. Additionally, they reported that the cetuximab group had the highest distribution and density of neuronal nuclei, meaning that the functional scaffold promotes migration and neural production [ 19 ]. Most recently, Li et al. demonstrated that paclitaxel (PTX) reduced glial scarring attributed to SCI by rescuing myelin-inhibited differentiation of NSCs. The cells were co-cultured with PTX and transplanted via a functional collagen scaffold into a complete T8 transection of the spinal cord in a rat model. Improvement of sensation and locomotion was confirmed by Western Blot (WB) and mRNA-Seq results that showcased the ability of PTX to trigger neuronal differentiation via the Wnt/β-catenin signalling pathway [ 20 ] The effect of collagen-based scaffolds to release therapeutic substances was also discussed earlier by Snider et al. , whereby the effectiveness was demonstrated using rat models to provide relevant evidence during the acute and chronic SCI phase. As mentioned by Snider et al. , the control group showed slight movement in one or two joints and extensive movement in one. Additionally, they reported reduced inflammatory cells and a higher organization in the new tissue of the spinal cord [ 21 ]. Some studies have been published concerning the treatment of SCI injuries in human patients [ 22 - 24 ]. In 2016, the team led by Dai reported promising results after the transplant of collagen-based scaffolds in human patients with complete SCI at the cervical or thoracic level. According to their study, the resection of the scar or the scaffold transplant did not have any easily adverse effects. They reported improved penis reflex two months after the surgery; while in two cases, there was recovery of somatosensory evoked potentials six months after the operation. Additionally, skin sweating was partially recovered below the level of the injury in three cases [ 23 ]. In 2018, the same team performed transplants of collagen-based scaffolds with human umbilical mesenchymal stem cells in two cases of SCI injuries in humans. The first case, where the injury was at the T11 level, showed recovery of sensory function at two months, which was further improved at six months. They also reported sense function of the bowel and the bladder at nine months. The muscle function was progressively regained below the T11 level after the injury, and at 12 months, the patient could walk with a brace. The Cervical SCI patient began to recover sensory function at two months and increased up to S5 at nine months. At 12 months post-surgery, the patient regained accurate bowel and bladder sensory function. The muscle control also improved, and at six months, the patient could even raise his lower legs against gravity. In both cases, the ASIA score improved from A to C 12 months post-surgery [ 22 ]. The importance of animal models in SCI studies Al-Hoseini et al. , noticing the importance of using animal models in SCI studies, conducted a systematic review on “Animal Models of Spinal Cord Injury”. The researchers categorized 2, 209 injuries according to level, outcome, animal species and purpose of the study. Most of the reviewed studies examined drug effectiveness, while others observed pathophysiologic changes. Eighty-one per cent of SCI sites involved the thoracic region, whilst contusion, transection and compression were the most common injury types induced. The majority (72. 4%) of SCI assessments were conducted on rats, as the rodents' biological and behavioural outcomes and biomechanics and neuropathology highly resemble humans. According to the study, rodents, such as mice or rats, are optimal for preliminary SCI studies because of the low reproductory cost and resemblance to human beings in terms of pathology and genomes [ 25 ]. Cats are another popular choice in SCI studies, mainly because they are larger than rodents, allowing more effortless surgical manoeuvres. Another important preclinical model is the pig, which combines an intermediate size and a more remarkable resemblance to human physiology. Fish, lamprey and other vertebrates have also been deployed in novel SCI studies, owing to their unique regeneration capability. The study points out that the optimal choice for SCI studies would be the non-human primates and larger animals that represent human SCIs a lot better than other organisms [ 25 ]. Notwithstanding the above, these primates are not ultimately deployed in such studies because of costly care and regulatory and ethical considerations. As an alternative, canines can be studied in laboratory conditions after naturally occurring SCI (e. g. , due to accidents), causing less moral concern. As shown in Table 1, four studies [ 14 - 16, 26 ] carried out between 2014 and 2017 have applied collagen-based scaffolds for SCI treatment in mouse, rat and canine models. In these cases, the stump connection, the transplanted cell differentiation, the elimination of glial scar and the increased neuronal growth were noted. Additionally, the decreased collagen deposition, the behavioural recovery, the improved electrophysiology and the enhanced axonal regeneration were evident. Table 1 Review of SCI studies deploying collagen-based scaffolds in animal models Material Animal model Spinal cord injury type Outcome Functional recovery Reference Motor Sensory Collagen Mouse Transection Connection of stumps in the transected spinal cord N NM Sugai et al. (2015) [ 15 ] Rat Hemisection Decrease of glial scaring and collagen deposition, increase of neurons Y NM Breen et al. (2016) [ 14 ] Canine Recovery of behavioral and electrophysiology: preventing formation of glial scars; enhancing axon regeneration Y Y Han et al. (2014, 2017) Conclusions This review has aimed to compile the latest bibliographical data available concerning the application of collagen scaffolds to treat SCI in animal models. SCI is one of the most critical cases a patient and a surgeon may encounter, bearing significant economic and psychological implications. A few relevant studies have recently been carried out using carefully selected animals that resemble human pathophysiology. Collagen-based scaffold application is one of the most promising methods of SCI treatment.

10. 7759/cureus. 26042

Cureus

Role of Placental Extracts in Periodontal Regeneration: A Literature Review

Periodontium is a specialized tissue surrounding the teeth. It is made up of the gingiva, periodontal ligament, cementum, and alveolar bone. The healing of periodontal tissues when infected occurs through repair and regeneration. The central dogma of regenerative periodontics is to stimulate a cascade of healing events that, if coordinated well, can lead to proper tissue synthesis which in turn would play a very important part in managing periodontitis and preventing tooth loss. Many regenerative procedures are being followed in periodontics using newer and modified barrier membranes. Placental membranes like amnion, chorion and amnion-chorion membranes are one among these that serve the purpose because of their active components and therapeutic effects. This literature review highlights the benefits of placental extracts in regenerative periodontal therapy.

Introduction and background The placenta is an organ developed during pregnancy which is enriched with mesenchymal stem cells (MSCs) and its presence is vital for fetal growth and maturation. These MSCs have the capacity to differentiate into different cell types and also have the added advantage of renewing themselves, which places them in high demand for regenerative procedures. The fresh membrane is extracted from the placenta during delivery via the vaginal or caesarean procedure. The cleaned membrane is placed in a 0. 025 per cent solution of sodium hypochlorite and kept at 4°C in a sterile solution containing penicillin. Placental membranes can remain sterile for up to six weeks [ 1 ]. Cultured whole human amniotic membrane is a source of pluripotent stem cells which help in the formation of the primitive liver, lung, neural, epithelial, haematopoietic cells and digestive tract. The human amnion‑derived cells are capable of forming cells of all three germ layers [ 2 ]. The advances in the cell therapy approach combined with the option of auto banking provide us with a scope of using placental extracts in regenerative therapies [ 3 ]. The use of foetal membranes for skin transplantation has been in vogue since 1910 [ 4 ]. These membranes have been utilised in ophthalmology as well [ 5 ]. In animals afflicted with diabetes, it has been noticed that when these placental extracts were administered it led to accelerated healing of wounds which prompted their use in diabetic neuropathy and angiopathy [ 6 ]. The treatment of nonhealing trophic ulcers, enterocutaneous fistula, orthopaedic pathology and the like has been performed using amnion membrane transplantation [ 7 ]. The placental extract also contains nucleotides like polydeoxyribonucleotides (PDRNs), that are known for their regeneration. Thus, amnion or amniotic membrane (AM), chorion membrane, amnion chorion membrane (ACM) and gel forms of placental extracts have multiple applications in medicine and dentistry. Cryopreserved, dehydrated amnion-chorion laminate is the available source of placental allograft for dental use [ 8 ]. The loss of soft and hard tissues is the ultimate outcome of periodontitis and thus restoring these tissues back to health is of prime importance. The concept of guided tissue regeneration (GTR) and guided bone regeneration (GBR) have thus been used as treatment approaches and placental extracts have also carved a niche for themselves in regenerative medicine owing to their biological properties. Review Properties of placental extract membranes Amnion and chorion membranes constitute the fetal component of extraembryonic tissue. The outer layer of the AM is the chorion membrane (CM). The AM consists of three layers: the epithelium, basement membrane, and the stroma which consist of an inner compact layer, middle fibroblast layer, and outermost spongy layer [ 9 ]. The amniotic epithelial cells are in contact with the basement membrane whereas the amniotic MSCs are located in the deeper spongy layer of the membrane [ 10 ]. Chorion is composed of a reticular layer, the basement membrane, and trophoblasts. The AM is 0. 02-0. 5mm thick and its self-adherent property permits it to intimately adapt over the root contours and defect areas [ 8 ]. The resorption rate of BioXclude ACM (Snoasis Medical, Golden, CO) was 8-12 weeks, whereas a dehydrated CM resorbed in two to four weeks [ 11 ]. As the reticular layer interacts with the spongy layer, the chorion is three to four times thicker than the amnion. Collagens I, III, IV, V, and VI are present in the reticular network [ 12 ]. The amniotic membrane has an inhibitory effect on macrophages and polymorphonuclear neutrophils [ 13 ], Interleukin 1𝛼and interleukin 1β [ 14 ], due to which its anti-inflammatory action is potentiated. The ligands for CD44 are the glycosaminoglycans and hyaluronic acid present in the amniotic membrane which facilitate the adhesion and entrapment of inflammatory cells including lymphocytes onto their surface [ 15 ]. The wounds treated with AM heal by regeneration as the membrane downregulates transforming growth factor 𝛽 and its receptor expression by fibroblasts. Human amniotic membrane (HAM) has been tried in the reconstruction of temporomandibular joint ankylosis due to its antifibrotic property [ 16, 17 ]. The AM facilitates the migration of epithelial cells, reinforces basal cell adhesion, promotes epithelial differentiation, prevents epithelial apoptosis, and promotes epithelialization in the healing of wounds. Their good permeability in contrast to other synthetic materials provides sufficient oxygenation for epithelial cells [ 18 - 22 ]. Transplantation of fresh AM resulted in the expression of HLA I antigens causing a mild inflammatory response [ 23 ] which was not observed with the cryopreserved amniotic membrane where the epithelial cells are lost during the process of cryopreservation [ 24 ]. Thus, cryopreserved tissue grafts of placental membrane materials have a low risk of immune rejection [ 25 ]. The AM produces 𝛽-defensins, especially 𝛽3-defensin, which inhibits the release of MMPs and suppresses proteinase action, thus leading to decreased inflammatory cell infiltration [ 26, 27 ]. The apoptosis of leukocytes is promoted by the amniotic membrane and due to the presence of interstitial collagens in it, the membrane is resistant to proteolytic factors [ 28, 29 ]. The low molecular weight elastase and proteinase inhibitors and elafins, which are part of the innate immune system are responsible for the antimicrobial property of AM [ 30 ]. The ACM provides a biomatrix with various proteins like lactoferrin, laminin-5, platelet-derived growth factor (PDGF) α, and β, fibroblast growth factor; transforming growth factor-β. collagen types I, III, IV, V, and VI aid in wound healing. Its antiviral property is due to the presence of cystatin E, an analogue of cysteine proteinase inhibitor [ 31 - 34 ]. The amniotic membrane works as an excellent scaffold, providing the ideal environment for the growth and differentiation of cells. Due to the rapid angiogenesis which occurs the grafted area heals rapidly and uneventfully. Peptides present in the placental extract including fibronectin III, regulate trypsin activity. It has been demonstrated that one or more peptides from human placental extract including fibronectin type III, help in the regulation of trypsin activity which aids in debridement and prevents keloid formation during wound healing [ 35 ]. Clinical applications BioXclude is a second-generation placental allograft composed of amnion and chorion tissue (300um in thickness) and has been used in GTR and GBR procedures. It has the property of self-adherence thus suturing is avoided, which in turn reduces the operating time. Hence this material is the preferred option in recession defects, particularly in the posterior region. Ambio5™ (Katena, Denville, NJ) is a third-generation amniotic membrane, which is thicker and more amenable for transplantation and has yielded good results. Another study was conducted on 15 patients with 30 mandibular degree II furcation defects, who were randomly allotted into Group I (PRF and AM) and Group II (PRF only). The clinical parameters like PI, GI, PPD, RAL, and furcation depth were assessed at baseline, three months, and six months. The PRF+ amnion group showed better improvement in the treatment of grade II furcation [ 36 ]. A study was conducted on 16 patients with class I gingival recession. Patients were divided into two groups with those allocated in group I undergoing CAF +AM and those allotted in group II undergoing CAF with PRF. The clinical parameters were evaluated at six and 18 months postoperatively. It was concluded that CAF with AM and CAF with PRF were equally effective in providing clinically significant outcomes, however, with respect to root coverage the AM showed a better percentage of root coverage as compared to PRF [ 37 ]. Fifty-one subjects with bilateral class I gingival recession defects were randomly divided into two groups, wherein the test group was treated with an amniotic membrane and coronally positioned flap, while the control group was treated with a coronally positioned flap alone. Clinical parameters such as RD, RW, PD, RAL, WKG, and TKG were recorded at baseline and after five years of follow-up. Intergroup comparison showed a non-significant difference in all variables except the TKG. This study concluded that AM helped improve the TKG, which is beneficial in the maintenance of the gingival margin [ 38 ]. In a patient aged 40 years, a combination of bilaminar and CAF techniques with HAM was performed to treat a class I gingival recession. The study results favoured the use of HAM as the recession was completely covered, moreover, there was an improvement in the gingival phenotype [ 39 ]. Fifty patients with PPD ≥ 6 mm and an intra-bony component of ≥ 3 mm were randomly allocated to collagen membrane and biphasic calcium phosphate group as well as amniotic membrane and biphasic calcium phosphate groups. It was inferred that both the groups performed equally well and that an amniotic membrane with biphasic calcium phosphate could be preferred as a treatment option for intra-bony defects [ 40 ]. The anti-inflammatory effect of chorion as a barrier membrane in periodontal pocket therapy was evaluated by assessing interleukin-11 (IL-11) level in gingival crevicular fluid (GCF). Two sites in two quadrants from each of the 10 patients were selected and randomly allocated in Group 1 (flap surgery) and Group 2 (flap surgery and chorion membrane placement). Intergroup comparison showed a statistically significant decrease in SBI, PPD, CAL, and IL-11 in Group 2 compared to Group 1 at four weeks. It was concluded that adjunctive use of chorion membrane in flap surgery was effective in treatment outcomes [ 41 ]. Eighteen intra-bony defects in 9 patients with chronic periodontitis were randomly assigned to group 1 (FDBA and chorion membrane) and group 2 (DFDBA and chorion membrane) for periodontal therapy. Clinical and radiographic (RVG) measurements were made at baseline and 12 months. The results were not statistically significant, however, group 1 (FDBA) showed an increase in bone density which was statistically significant. Within the limitations of the present study, both the groups showed similar results with a significant increase in bone density in the FDBA group [ 42 ]. In a study 30 sites with Miller's Class, I and class II recession were taken and randomly allocated to the chorion membrane (test) PRF membrane (control) group. The clinical parameters recorded were CAL, REC-HT, REC-WD, WKG, and GTH. It was observed that there was a significant improvement in all the parameters in both the groups, however, on intergroup comparison the test group (chorion membrane) showed better results related to CAL, REC-HT, and GTH when compared to the control group [ 43 ]. In a case report a patient had a faulty post and core with a crown with respect to the maxillary right central incisor with a PD of 8 mm and HGR 2. 5 mm. After the crown was replaced, a flap was reflected by a semilunar incision and after debridement, a chorion membrane was placed and sutured back. The patient was evaluated at different points of time postoperatively and it was observed that all the clinical parameters improved over a period of six months [ 44 ]. Around 30 patients with Miller's Class I and class II GR-type defects were divided into three groups randomly. In Group A 10 patients were treated with only CAF, group B 10 patients were treated with CAF, CM and DFDBA. Similarly, in Group C 10 patients were treated with CAF, AM and DFDBA. Clinical parameters were assessed at baseline and three months postoperatively. The percentage of root coverage obtained within the study groups was highest in group B [ 11 ]. A cohort of 20 patients with 25 Class I/II interdental papillary recession defects were treated with ACM and coronal advancement of the gingiva papillary unit via a semilunar incision on the labial aspect followed by a sulcular incision within the area of interest. The black triangle height (BTH) and also the black triangle width (BTW) were calculated by image analysis software, which showed a statistically and highly significant difference from the baseline until three and six months postoperatively. It was inferred that ACM allograft in conjunction with a coronally advanced flap can be an appropriate minimally invasive alternative for papillary regeneration [ 45 ]. A study was done to compare treatments of the deficient ridge with ACM/dPTFE membranes in 22 non-molar sites on the identical arch. Postoperative clinical and radiographic ridge dimensions weren't significantly different between the two treatments. ACM sites had significantly more osteoid and better bone volume density compared with dPTFE. Additionally, the researchers remarked that ACM use may improve both patient and clinician-concerned outcomes related to implant placement [ 46 ]. Dehydrated amnion/ chorion membrane allograft allows it to be administered as a topical powder or mixed with saline to create an injectable solution or a topical gel. Local injection of Placentrex is a very good therapeutic option when administered in the early stages of oral submucous fibrosis without any side effects and contraindications [ 47 ]. Compared to povidone-iodine and saline dressing Placentrex gel provided faster healing without causing interference to granulation tissue and was found to be effective even in presence of pus serum, blood, and slough [ 48 ]. Another study was done to evaluate the healing efficacy of topically applied placental extract gel both clinically and histologically. 10 healthy patients in the age group of 18-35 years. who were indicated for depigmentation procedure were selected for the study. Depigmentation was done with the scalpel technique on the maxillary and mandibular anterior region. In group, A human placental extract gel was applied to the wound and a non-eugenol pack was placed whereas group B was covered with a non-eugenol pack only. Wound Healing Index and Visual Analogue Score were assessed after seven and 15 days. The epithelisation of the wound was assessed by using toluidine blue after seven days of surgery. Application of human placental extract (HPE) gel showed a statistically significant improvement clinically and histologically in Group A with a distinct parakeratinized stratified squamous epithelium with fibrous connective tissue and nil inflammatory cell infiltrates. Whereas in Group B eight individuals showed moderate inflammatory cell infiltrate. It was concluded that local administration of HPE directly onto wound margins promotes wound healing due to an increase in the amount of transforming growth factor in the early phase of wound healing and vascular endothelial growth factor in the late phase [ 49 ]. After depigmentation, it has been observed that the application of Placentrex gel helped in healing the wound better with better patient comfort when compared to a periodontal dressing alone [ 50 ]. Amniotic membrane in tissue engineering The scaffold is one of the triads required for tissue engineering. It has to be biocompatible with the host tissue to help in regeneration. The amniotic membrane is enriched with cytokines and growth factors which augment its benefits when used as a three-dimensional scaffold for sustained drug release or to initiate attachment of cells required for regeneration. Conclusions In this review, the benefits of using placental extracts were highlighted. As periodontal regeneration revolves around the presence of cells, signaling molecules, and scaffolds, the advent of amnion and chorion membranes has largely benefitted the field of regenerative periodontics, as both these membranes have a vast array of growth factors to promote healing and they can also work as biocompatible scaffolds to orchester the ingrowth of cells and blood vessels to initiate regeneration and repair. Many more studies should be done using these membranes to validate their role in regenerative medicine.

10. 7759/cureus. 27832

Cureus

Amniotic Membrane: An Approach to Periodontal Regeneration

Over the years, various materials have been used for scaffold-based periodontal tissue engineering to regenerate lost periodontal tissues. The use of amniotic membrane (AM) as a scaffold for periodontal regeneration has gained great interest among researchers. This narrative review aims to appraise the properties of AM and its potential clinical applications in periodontal regeneration. PubMed, ScienceDirect, Scopus, and Wiley Online Library databases were searched for relevant articles that highlighted the properties and applications of AM in periodontal regeneration. AM has a unique structure and components contributing to its exceptional properties such as anti-inflammatory (presence of anti-inflammatory factors), low immunogenicity (presence of human leukocyte antigen-G), anti-scarring (downregulation of transforming growth factor-β), antimicrobial (expression of antimicrobial factors), promotion of epithelialization (production of growth factors), and reduction of pain (protection of exposed nerve endings). Its use in the treatment of periodontal tissue defect has shown to be effective. AM showed various beneficial properties as an ideal scaffold. Future studies and long-term clinical trials on the efficacy and survival rate of AM are required to completely understand the potential application of AM in periodontal regeneration.

Introduction and background Periodontal disease is a major public health issue that distributes globally and comprises a wide spectrum of conditions ranging from mild gingivitis to severe periodontitis [ 1, 2 ]. According to the Global Burden of Disease Study 2016, severe periodontal disease was the 11th most prevalent condition in the world, with its prevalence ranging from 20% to 50% [ 3, 4 ]. It is a chronic inflammatory condition initiated by bacteria in dental biofilm in the susceptible host, which can be modified by the presence of risk factors. In general, it can be classified as gingivitis and periodontitis. The ultimate goals of periodontal therapy include the arrest of periodontal disease progression and complete reconstitution of all periodontal attachment to their original architecture and function that replicates its pre-disease structure [ 5 ]. Periodontal regeneration is defined as the restoration and reconstruction of the lost periodontium or supporting structures including the alveolar bone, cementum, periodontal ligament, and gingiva [ 6 ]. However, current conventional periodontal therapies show a limited potential for complete periodontal regeneration. Over the years, various methods have been used in achieving periodontal regeneration. The most common is guided tissue regeneration (GTR), whereby the membrane or other biomaterials are used as a barrier or scaffold in order to allow the desired cell to repopulate the periodontal defect area. The membranes that have been used include natural and synthetic biomaterials [ 7 ]. The amniotic membrane (AM) is the innermost layer of the fetal membranes, which is avascular and forms an amniotic fluid-filled sac that surrounds and protects the embryo. AM is translucent and is one of the thinnest membranes (approximately 0. 02-0. 5 mm) in the human body. It is made up of three distinct layers: (1) epithelium, (2) basement membrane, and (3) stromal matrix. The stromal matrix can be further divided into three layers, which are the inner acellular compact layer, the middle loose fibroblast layer, and the outermost spongy layer [ 8, 9 ]. AM is routinely discarded post-partum. It is obtained after normal or cesarean deliveries under informed consent, which usually poses little to no ethical concerns. Consequently, it is a readily available and cost-effective biomaterial for scaffolds in tissue engineering [ 10 ]. Scanning electron microscopy analysis of AM revealed rough surface architecture with the presence of microporosity, which may provide a suitable platform for cell attachment [ 11 ]. AM was used for wound treatment more than a century ago as a skin graft substitute for open wound for treating burnt and ulcerated skin surfaces by which it can accelerate epithelialization and reduce pain [ 12 - 14 ]. In 1940, de Rötth [ 15 ] first reported the use of fresh amnion and chorion in ophthalmology to reconstruct the ocular surface in patients with symblepharon [ 16, 17 ]. Since it was discovered that AM could be separated, sterilized, and safely used, amnion-derived cells have attracted much attention in dentistry, particularly for the regeneration of periodontal tissues [ 18 ]. From an updated review of the top five clinical applications of AM in regenerative medicine from 2015 to 2020, it was revealed that dermatology (specifically wound healing), orthopedics, ophthalmology, dentistry, urology, oncology, and otolaryngology used AM more compared to other specialties. AM only accounted for 6% in dentistry as compared to 32% in dermatology and 26% in orthopedics [ 19 ]. However, AM is one of the biomaterials that became an area of interest in periodontal application. The reports on its use in the management of gingival recession, furcation, and intrabony defects have shown positive outcomes [ 20 ]. Hence, due to the increasing number of studies in the field of regenerative medicine, studies are still needed to clarify the future prospect of AM in dentistry particularly periodontology [ 8 ]. Therefore, the purpose of this review is to appraise the properties of AM and its potential clinical applications in the field of regenerative periodontology. Review Article search A web search of all relevant literature was performed on the databases such as PubMed, ScienceDirect, Scopus, and Wiley Online Library. The following keywords were searched alone or in different combinations in the titles and abstracts: “amniotic membrane”, “periodontal regeneration”, “periodontal surgery”, “tissue engineering”, “regenerative medicine”. Relevant articles were identified, and duplicates were removed. Full texts of the identified articles that met the inclusion criteria were acquired and assessed. Articles were searched and retrieved from the reference lists of the initially selected articles for additional relevant studies. Searches were limited to articles in the English language and published from January 2001 until December 2021. The inclusion criteria for articles include clinical trials, case reports, and case series. The findings from the search are presented as a narrative review. Results During the initial search process, overall, 2, 108 articles were found from the databases ScienceDirect, Scopus, PubMed, and Wiley Online Library. However, after further screening, which included removing duplicates and reviewing publications based on titles, abstracts, and articles, only 16 articles published in the year 2014 to 2021 were identified and included. Figure 1 depicts the process of conducting a literature search and the number of articles found. Figure 1 Flowchart of the literature search and selection process Of the 16 articles, nine (56. 25%) were studies that originated from India. Two (12. 5%) were studies that originated from Iran. Other study origins were from Japan, Taiwan, Malaysia, Italy, United States of America, each (6. 25%) respectively. Out of 16 articles, six (37. 5%) were case reports, five (31. 25%) were randomized controlled clinical trials, three (18. 75%) were animal studies using a rat model, and one (6. 25%) was an in vitro study and case series, respectively. It was found that AM possesses many beneficial properties, which include antifibrotic, anti-inflammatory, antimicrobial, anti-scarring, mechanical strength, and flexibility [ 21 ]. One of the most widely reported properties of AM is its ability to reduce inflammation [ 22 - 25 ]. AM is also proven to have other properties such as low immunogenicity, pain reduction and promotion of epithelialization, self-adhesive, and aesthetics [ 26 - 28 ]. In regard to AM as a potential scaffold for periodontal regeneration, nine studies reported promising outcomes in the root coverage procedures for the treatment of gingival recession defects. Seven studies on GTR demonstrated that AM was an effective barrier able to enhance bone fill and improve periodontal parameters (probing pocket depth [PPD] and clinical attachment loss [CAL]). One laboratory study reported AM as a suitable scaffold for periodontal fibroblast cell growth [ 11 ]. The findings from all selected studies are summarized in Table 1. Table 1 Detailed summary of the selected studies AM, amniotic membrane; PPD, probing pocket depth; CAL, clinical attachment loss; HPDLFs, human periodontal ligament fibroblasts References Study Origin Type of Study Properties Clinical Applications of AM in Periodontal Regeneration [ 26 ] India Case report Self-adhesive, promotion of epithelialization, low immunogenicity, easily available, cost-effective Stable and full root coverage of a Miller class I gingival recession defect seven months post-surgery [ 21 ] Japan In vivo (rat model) Antifibrotic, anti-inflammatory, antimicrobial, anti-scarring, mechanical strength, flexibility AM acts as a scaffold for periodontal ligament stem cells to enhance periodontal regeneration and showed a monolayer of the cells on the amnion surface [ 22 ] India Case report Non-immunogenic, anti-inflammatory, antibacterial, reduction of pain, aesthetics AM allograft in conjunction with gingival flap showed a complete root coverage of a Miller class II gingival recession with improved tissue architecture six months post-surgery [ 27 ] India Case report Lack of immunogenicity, antibacterial, reduction of pain, aesthetics AM can be used as an allograft material in the treatment of root coverage to gain attachment level and reduce the length of the recession [ 23 ] India Randomized controlled clinical trial Anti-inflammatory, anti-infective, antimicrobial AM functions as a barrier for guided tissue regeneration to increase bone fill and reduce PPD and CAL [ 18 ] India Case series Self-adhesive AM as an autograft tissue in the treatment of shallow-to-moderate Miller’s class I and II recession defects showed a significant improvement in the clinical attachment level and width of keratinized gingiva six months postoperatively [ 28 ] Iran Randomized controlled clinical trial Self-adhesive, aesthetics Coronally advanced flap with AM in the treatment of Miller’s class I and II gingival recessions decrease surgical operation time and patient discomfort [ 24 ] Taiwan In vivo (rat model) Anti-inflammatory, anti-angiogenesis, immunosuppression AM and adipose-derived stem cell co-culture system increases bone regeneration in a periodontal osseous defect rat model by forming more hard tissues and showing better defect recovery [ 29 ] India Case report Promotion of epithelialization, anti-scarring, lack of immunogenicity, antimicrobial, antibacterial AM can be used as an effective barrier in conjunction with bone grafts to treat an intrabony defect [ 30 ] India Randomized control clinical trial Promotion of epithelialization, anti-scarring, lack of immunogenicity, self-adhesive Coronally advanced flap using AM showed a favorable outcome of root coverage percentage in the treatment of localized gingival recession defects by maintaining the structural and anatomical configuration of the regenerated tissues [ 31 ] India Case report Excellent handling properties, self-adhesive, easily available, uniform thickness The combined approach of the coronally advanced flap and AM in the treatment of multiple adjacent gingival recessions showed significant root coverage and an increase in thickness of keratinized gingiva [ 11 ] Malaysia In vitro Biocompatible for cell growth, porous surface AM serves as a scaffold for the attachment and proliferation of HPDLFs in periodontal tissue engineering [ 32 ] Italy Case report Promotion of epithelialization, reduction of pain, anti-scarring AM acts as an allograft in the treatment of gingival recession in conjunction with coronally advanced flap and can promote palatal wound healing [ 25 ] Iran Randomized controlled clinical trial Anti-inflammatory, reduction of pain, anti-scarring, aesthetics AM as a biological dressing on wound healing after free gingival graft surgery can prevent postoperative complications and help to accelerate healing [ 33 ] United States of America In vivo (rat model) Neovascularization, promotion of osteoconduction Periodontal regeneration was enhanced in surgically created rat periodontal furcation defects by preserving its structure during cultivation and healing periods, supporting cell attachment and bone deposition [ 34 ] India Randomized controlled clinical trial Promotion of epithelialization, reduction of pain AM as a barrier with biphasic calcium phosphate provides a better outcome in the management of periodontal intrabony defects by reducing PPD and CAL in chronic periodontitis patients Properties of AM Anti-inflammatory Out of the 16 studies, five (31. 25%) showed that AM has anti-inflammatory properties. Kumar et al. conducted a randomized controlled clinical trial to investigate the anti-inflammatory, anti-infective, and therapeutic effects of AM when utilized for GTR in confined interdental lesions [ 23 ]. The interleukin (IL)-1β and human beta-defensins (hBD)-2 levels were measured in the gingival crevicular fluid (GCF) of the test site (AM with bone graft) and control site (bone graft only). GCF is an inflammatory exudate that can be used as a non-invasive method to evaluate periodontal inflammatory reactions in a variety of clinical settings [ 35 ]. AM demonstrated a significant reduction in IL-1β level and an insignificant increase in hBD-2 expression in GCF. Increased hBD-2 levels play an important role in defense from periodontopathogens in human gingival tissues. The significant reduction of GCF IL-1β levels in AM-treated sites indicated that AM has a significant anti-inflammatory effect on periodontal tissues [ 23 ]. This finding is consistent with that of Kadkhoda et al. [ 25 ], whereby the inflammation was used as an objective measure of clinical healing. At all follow-up visits, the inflammation on the palatal donor site was more prominent in the control group, although the difference was significant only after 14 days post-surgery. On day 21 post-surgery, the inflammation score in AM group was “0, ” which indicates no inflammation [ 25 ]. Several studies have shown that the incorporation of AM into collagen scaffolds enhanced its anti-inflammatory properties through chemical and mechanical effects. Chemically, there is a presence of various anti-inflammatory factors and substances such as activin A, IL-1 and IL-2 receptor antagonists, IL-10, endostatin, and tissue inhibitors of metalloproteinase (TIMP)-1, TIMP-2, TIMP-3, and TIMP-4, which inhibit endothelial cell proliferation, angiogenesis, and tumor growth [ 13, 20, 36, 37 ]. Secretory leukocyte proteinase inhibitor (SLPI) and elafin have both anti-inflammatory and anti-microbial effects [ 8 ]. Chemical-mediated anti-inflammatory effect is also driven by the suppression of pro-inflammatory cytokines IL-1α, IL-1β, IL-2, IL-8, interferon-γ, tumor necrosis factor (TNF)-β, basic fibroblast growth factor, and platelet-derived growth factor [ 16 ]. Other than that, there is a decreased recruitment of inflammatory cells such as polymorphonuclear cells, CD3 cells, CD4 T cells, and CD11b cells [ 38, 39 ]. In addition to the chemically mediated anti-inflammatory effect, the mechanical effect was demonstrated by AM, which serves as a physical barrier that confines inflammatory cells to the affected area and decreases inflammatory mediators. AM stromal matrix entraps T lymphocytes and results in apoptosis of the inflammatory cells [ 39 ]. Therefore, AM has considered being a suitable allotransplantation tissue due to its anti-inflammatory effect. Low Immunogenicity and Immunomodulatory Six of the (37. 55%) 16 studies reported that AM has low immunogenicity and immunomodulatory properties. Rehan et al. studied the effectiveness of coronally advanced flap (CAF) with AM in the treatment of localized gingival recession defects [ 30 ]. The results were reported to be stable even after 18 months postoperatively, suggesting that AM forms a physiologic seal with the host tissue hindering bacterial contamination while supporting AM’s ability to decrease host immunologic response through localized suppression of polymorphonuclear cell migration. This finding is in accordance with the case reports of Shah et al. [ 22 ] and Shetty et al. [ 26 ] who reported stable results in AM-treated sites for six and seven months, respectively, post-treatment without recurrence of recession. The results from these reports are encouraging and demonstrated that amnion allograft is well-tolerated by the gingival tissues without any sign of immununorejection. In fact, immunosuppression is mandatory in skin allografts. However, AM transplantation for skin or corneal defects performed an exceptional lack of immunogenicity property by showing no signs of rejection in the absence of immunosuppression. This phenomenon result was most likely from the combination of anti-inflammatory, low immunogenicity, and immunomodulatory properties [ 40 ]. Low immunogenicity is important to create a biocompatible scaffold for tissue engineering. The occurrence of acute rejection after transplantation of AM is very rare due to the fact that amniotic epithelial cells do not express human leukocyte antigen (HLA)-A, HLA-B, HLA-D, and HLA-DR antigens. Instead, amniotic epithelial cells express immunoregulatory factors HLA-G and Fas ligand on their surfaces. The expression of HLA-G is the main factor that prevents the rejection of the trophoblast because it is involved in the induction of immune tolerance by acting as a ligand for inhibitory receptors that present on the natural killer (NK) cells and macrophages [ 8 ]. The presence of interferon-𝛾 and other immunologic factors has been observed in the AM [ 9, 20, 41 ]. The immunologic factors secreted by the epithelial cells reduce the host immunologic response to prevent a maternal immune attack [ 18, 23, 42 ]. It was reported that there was no immunorejection observed from the transplantation of allogeneic periodontal ligament stem cell (PDLSC)-transferred amnion into swine periodontal defect models. No enhancement of T-cell and B-cell proliferation and immunoglobulin production was shown, thus suggesting the possibility of periodontal regeneration using allogeneic PDLSC-transferred amnion [ 21 ]. AM is also said to be immunomodulatory due to its unique molecular arrangement, which makes it invulnerable to maternal immune system responses. The cellular components of AM have active suppression activity on the immune cells’ activity through a strong paracrine secretion. This suggests that AM may have an immunomodulatory effect after transplantation, preventing the cellular cargo from being rejected [ 43 ]. Due to its success to prevent an allogenic or xenogenic immunologic reaction, AM has gained great interest in transplantation and tissue engineering. Despite these promising results, questions remain on the long-term efficacy and stability of AM as an immunomodulatory biological dressing. Therefore, to establish the efficacy and stability of AM, more randomized controlled clinical trials involving immunological investigations with longer follow-up visits are required. Antimicrobial The results of three studies (18. 75%) showed that AM has antimicrobial properties. The study by Kumar et al. [ 23 ] reported that there was a minimal insignificant increase in the hBD-2 levels in sites treated with AM. This relatively small rise in the hBD-2 levels was caused by a significant reduction in the IL-1β levels. AM demonstrates an antimicrobial effect due to hBD production and by forming a biological “seal” with the host tissues, thus acting as a physical barrier against the outer environment. Defensins help in tissue proliferation, and the production of antimicrobial peptides by AM may promote periodontal regeneration [ 23 ]. It was suggested that the mechanism of antimicrobial action of AM is due to its role as a biological barrier against bacterial infiltration by closely adhering to the wound surface and preventing dead space formation and serous charge accumulation [ 37 ]. Other literature further explained that AM forms a barrier with the wound surface via fibrin and elastin linkages. This firm adherence helps in restoring lymphatic integrity, protecting circulating phagocytes from exposure, and allowing faster removal of surface debris and bacteria from the wound surface. There are two mechanisms mediating the antimicrobial activity: (1) direct, via secretion of antimicrobial factors such as human cathelicidin (LL-37), and (2) indirect, via secretion of immunomodulatory factors, which upregulate bactericidal activity and phagocytosis by immune cells. AM is also found to contain many bactericidal products of purine metabolism and lysozyme. A major group of antimicrobial peptides found in the AM is formed by defensins, mostly β3-defensin, that helps the epithelial surfaces to resist microbial colonization [ 38 ]. Apart from that, antimicrobial compounds found in amniotic cells, such as SLPI and elafin, act as components of the innate immune system to guard against infection. Treatment of AM with IL-1 receptor antagonist or lactoferrin also showed an antimicrobial effect [ 8 ]. Therefore, the antimicrobial property of AM has made it a suitable option for post-surgery applications in wound healing, burns, dental injuries, and ophthalmology because bacterial infection and biofilm growth are common in these sites [ 44 ]. Promotion of Epithelialization and Reduction of Pain A case report described Miller’s class III gingival recession treated with a palatal epithelial-connective tissue autograft and AM. It was reported that surgical treatment with palatal epithelial-connective tissue graft and AM can help accelerate the epithelialization of the wound at the palatal donor site, reducing morbidity. A positive resolution of the treated recession, absence of infection, and complete reepithelialization of the palate treated with AM were observed 30 days post-surgery [ 32 ]. AM may act as a basement membrane that promotes epithelialization by aiding epithelial cell migration, basal cell adhesion, epithelial differentiation, and epithelial apoptosis prevention. AM also produces growth factors that stimulate epithelialization and have a pain-reducing effect [ 7, 20, 37 ]. It reduces inflammation and hydrates the wound bed, thus promoting faster healing. This membrane was also proven to promote rapid epithelialization of the palatal donor site wound with a reduction of post-operational pain, thus leading to less discomfort experienced by the patient [ 32 ]. AM promotes healing and wound epithelialization while reducing granulation tissue formation in large open wounds without any adverse reaction, as reflected by decreasing analgesics intake and pain scores as well as minimal discomfort postoperatively [ 25, 27 ]. These results may be explained by the fact that the stromal surface closely adheres to the wound surface, and therefore the mucoid lining can protect the exposed free nerve endings in the wound area from external irritants and reduce pain sensation by preventing trauma and nerve stimuli [ 38, 41 ]. Another mechanism proposed is AM causes downregulating of the pro-inflammatory cytokines, such as TNF-α and IL-6, and activation of neutrophils and M1 and M2 macrophages, which help to relieve pain [ 19 ]. Anti-Scarring Scar tissue formation is a common occurrence during wound healing. Scar development is a complex biological process involving cell-cell and cell-matrix interactions driven by cytokines [ 45 ]. Kumar et al. [ 29 ] demonstrated the anti-scarring property of AM when used in conjunction with bone grafts to treat an intrabony defect. AM improves the overall regeneration due to its rich source of pluripotent stem cells, specialized proteins, and cytokines, thus promoting wound healing and reducing postoperative scarring [ 29 ]. In an 18-month clinical study to compare the efficacy of CAF using AM and platelet-rich fibrin (PRF) membrane in gingival recession, it has been demonstrated that CAF with AM is effective and showed better results than PRF membrane in providing clinically significant outcomes of root coverage by maintaining the structural and anatomical configuration of the regenerated tissues and enhancing healing through reduction of postoperative scarring [ 30 ]. This could be through secretion of vascular endothelial growth factor and hepatocytes growth factor that establishes a balance between transforming growth factors (TGF)-1 and TGF-2. Furthermore, there is a downregulation of TGF-β signaling modulated by hyaluronic acid, which suppresses the expression of TGF-β receptors such as TGF-β1, -β2, -β3 isoforms, and TGF-β type II receptor, inhibiting fibroblasts proliferation [ 18 ]. Differentiation of fibroblasts into myofibroblasts is also inhibited, thus reducing scarring [ 16 ]. Other contributing factor includes the reduction of protease activity due to the secretion of TIMPs [ 44 ]. Self-Adhesive Five (31. 25%) of the 16 studies have demonstrated that AM has a self-adhesive property. AM is able to adhere to the recipient's exposed root and proximal site upon placement on gingival recessions, thus eliminating the need for suturing [ 28 ]. AM can self-adhere and intimately adapt to contour around roots, thus contributing to the ease of root coverage procedure by making it less technically demanding and significantly reducing the surgical time [ 18 ]. AM used with CAF demonstrated stable results at the 18-month follow-up [ 30 ]. Besides, AM can be used to provide a significant root coverage outcome, increase the thickness of keratinized gingiva, and improve gingival biotype [ 22, 31 ]. It closely mimics the human mucosa basement membrane and contains laminin-5, which plays a role in the cellular adhesion of gingival cells. Other than the laminins, the basement membrane of AM contains collagen types III, IV, and V, and cell-adhesion bioactive factors including glycoproteins and fibronectins. The self-adhesive property of AM helps reduce operatory time because it does not require a second surgical site in the root coverage procedure [ 31 ]. Aesthetics Gingival recession appears clinically as the display of the root surface of the tooth due to the displacement of the gingival margin apically from the cementoenamel junction and is thus associated with multiple aesthetics and functional problems such as exposed root, cervical/root caries, tooth hypersensitivity, and pulp hyperemia. Therefore, the ultimate goal of any root coverage procedure is complete and stable coverage of the recession defect. AM provides excellent aesthetic results in terms of texture and color match to the recipient site and results in a complete root coverage for gingival recession defects [ 22 ]. The subepithelial connective tissue grafts technique is considered the “gold standard” of root coverage procedures. Remarkably, Lafzi et al. [ 28 ] observed that AM with CAF is relatively comparable with the gold standard. In fact, satisfaction with aesthetic results of AM was higher. In a randomized clinical control study, AM was used as a biological dressing at the palatal donor site after harvesting the soft tissue graft [ 25 ]. The observation after 21 days showed excellent color match and tissue texture of the palatal donor site with the adjacent tissue [ 25 ]. Owing to its aesthetic properties, AM could be one of the considered options in oral cavity defect reconstruction procedures. Improvement in Gingival Biotype AM is used in conjunction with CAF in root coverage procedures in Miller’s class I and class II gingival recession defects to provide stable and significant root coverage and increase the thickness of keratinized gingiva [ 26, 28, 30 - 32 ]. It is not surprising to note that some results showed a complete (100%) if not near-complete root coverage since AM has many exceptional properties that make it a membrane of choice to be used with CAF as a combined approach in treating gingival recessions [ 28, 31 ]. The root coverage was stable even after 18 months postoperatively by maintaining the structural and anatomical configuration of the regenerated tissues without any adverse effect. In another study, it was demonstrated that CAF with AM and PRF both achieved 100% root coverage and enhanced the gingival biotype in bilateral multiple Miller’s class I recession. Furthermore, the AM-treated sites demonstrated more stable results than the PRF-treated sites at the end of the seventh month [ 26 ]. This finding was consistent with that of Rehan et al. [ 30 ] who also compared the effectiveness of CAF with AM and PRF in the treatment of Miller’s class I recession defects in an 18-month study. The authors concluded that both membranes are equally effective in providing clinically significant outcomes with respect to root coverage in which AM shows a better percentage of root coverage as compared to PRF [ 30 ]. Because of its better stability and ease of handling, the application of AM as a novel approach to root coverage could be more desirable than PRF. Potential Scaffold for Regeneration Other than being used in a combined approach with CAF clinically, AM acts as a scaffold for periodontal ligament cell growth from in vitro studies [ 11, 21 ]. The basement membrane of the AM contains extracellular matrix components that produce a nearly native scaffold for cell seeding, thus suitable to be applied in the periodontal regenerative procedure [ 8, 42 ]. In the study of Iwasaki et al. , PDLSC-transferred AM was found to have a therapeutic potential on periodontal tissue regeneration by significantly enhancing the formation of periodontal tissues in vivo [ 21 ]. Meanwhile, Elahi et al. observed that human periodontal ligament fibroblasts (HPDLFs) can attach, proliferate, and integrate with AM, which indicated that AM is biocompatible and can be a promising scaffold for periodontal regeneration [ 11 ]. AM and adipose-derived stem cell-co culture systems could increase bone regeneration in a periodontal osseous defect rat model by forming more hard tissues and showing better defect recovery [ 24 ]. Therefore, the combination of tissue engineering technology utilizing AM and stem cell therapy to regenerate periodontal bone is very encouraging in patients with periodontal disease who suffer from tooth loss. Nevertheless, as described earlier, AM is a very thin membrane and delicate, thus requiring proper handling during application [ 7, 42 ]. Hence, having a thorough understanding of its physical characteristics as well as expert operators manipulating AM during periodontal regenerative procedures may aid in achieving good AM adaptation to the defect site. Conclusions Based on this review, it is evidenced that AM has unique structure and components contributing to its exceptional properties such as anti-inflammatory, low immunogenicity, anti-scarring, antimicrobial, promoting epithelialization, reduction of pain, and improving gingival biotypes, as well as a suitable platform for periodontal cells growth. Owing to these various beneficial properties, AM may serve as a potential alternative natural biomaterial that can be used for regenerative periodontal therapy. However, more clinical trials are recommended to further elucidate its efficacy and sustainability to act as a scaffold in periodontal regeneration.

10. 7759/cureus. 27946

Cureus

Dentin Matrix Metalloproteinases: A Futuristic Approach Toward Dentin Repair and Regeneration

Matrix metalloproteinases (MMPs) have been linked to modulating healing during the production of tertiary dentin, as well as the liberation of physiologically active molecules and the control of developmental processes. Although efforts to protect dentin have mostly centered on preventing these proteases from doing their jobs, their role is actually much more intricate and crucial for dentin healing than anticipated. The role of MMPs as bioactive dentin matrix components involved in dentin production, repair, and regeneration is examined in the current review. The mechanical characteristics of dentin, especially those of reparative and reactionary dentin, and the established functions of MMPs in dentin production are given particular attention. Because they are essential parts of the dentin matrix, MMPs should be regarded as leading applicants for dentin regeneration.

Introduction and background Regeneration of the tissue must closely resemble parent dentin due to the firmly associated structural and functional relationship in the physiologic dentin. In other words, maintaining the mechanical characteristics of the tissue provided by its biological structure is necessary for dentin regeneration. Dentin is an essential mineralized tissue that contains odontoblasts' biological functions within dentinal tubules and is in charge of reducing mastication pressures [ 1 ]. These pressures need to be transmitted from a rigid (96% mineral by weight enamel) to a much more elastic (70% mineral by weight dentin) substance. Collagenous (86% type 1, together with types 3, 5, and 6) and non-collagenous proteins make up the dentin matrix. Following pulpal injury from cavity preparations, carious lesions, erosion, and restorative dental materials, dentin is capable of limited healing. The circumpulpal dentin layer grows inward as a result of dentin healing by tertiary dentin deposition, enlarging the pulp chamber and root canals. The exact tubular structure of physiological dentin is lost in tertiary dentin. Reactionary dentin forms in the shape of tubular odontodentin or atubular and bone alike osteodentin after minor trauma, which does not affect the underlying Hoehl’s cells or odontoblast layer [ 2 ]. Despite being substantially mineralized, reactionary dentin is less elastic and rigid [ 3 ]. Progenitor cells are needed to fill up the deficiency after more severe injuries that cause cellular death by laying reparative dentin in the shape of a dentinal bridge. The quantity of tertiary dentin in a tooth may or may not have an impact on how well it functions. Dentin hardness is linearly correlated with tissue's lack of elasticity and directly proportional to the density of minerals present in the tissue. The tissue has different mineral densities, with peritubular dentin being the dentin that is closest to the dentinal tubule border, hardest, and least elastic [ 4 ]. The much more elastic intertubular dentin is found between tubules. Hardness is highest in circumpulpal dentin, lowest in the DEJ, and again falls toward the pulp [ 5 ]. Dentinal tubules increase in number, density, and size as they grow nearer to the cell body of odontoblasts [ 6 ]. Variations in the ratio of intertubular to peritubular dentin, which affects the tissue's hardness and mechanical qualities, are correlated with changes in dentinal tubule density. The effects of matrix metalloproteinase (MMP) activity on tube density and the structural and mechanical characteristics of tertiary dentin have not been demonstrated in research to our knowledge. Future regenerative models ought to take into account using these elements as success indicators because of how crucial they are to dentin function. Review Dentin regeneration and its implications The vitality of the pulp is always necessary for dentin repair and regeneration. This idea has been applied in the management of deciduous teeth that have open apices using regenerative endodontics. The center of the tooth is made up of the dental pulp, which is a loose connective vascular tissue. In close proximity to dentin, it is made up of cellular elements such as immune cells, odontoblasts, neuro-vascular networks, fibroblasts, and extracellular elements such as glycoproteins and collagen [ 1 ]. Vital pulp therapies help to regenerate vascularized and innervated dental pulp that can facilitate odontoblast differentiation and dentin neogenesis for the full development of roots [ 7 ]. Such a sort of dentin regeneration falls short of addressing the clinical crown loss brought on by caries. The amount of dentin that might be replaced in studies is unclear, as is the question of whether this would balance the amount lost via disease. Enamel is an acellular tissue; hence regeneration of this tissue presents much greater difficulties than dentin regeneration [ 8 ]. At the very least, tissue regeneration requires the capacity for cellular and matrix replacement through proliferation [ 9 ]. Replacement of the damaged dentin matrix by freshly generated, proliferating odontoblasts and differentiated dental pulp stem cell (DPSC) is necessary for dentin repair. In this regard, scaffolds and biological cues hold the most potential, enabling cellular infiltration, matrix deposition, and mineralization later on. Collagen, silk, chitosan, alginates, hyaluronic acid, hydrogels, and fibrin have all proved successful in promoting dental pulp cell maturation [ 10 ]. Despite the fact that these scaffolds claim to encourage pulpal regeneration, no research has demonstrated meaningful amounts of mineralization to compensate for the dentin that has been lost to caries. Additionally, the length of time needed to repair the bulk of the tissue that caries destroyed would be excessive and clinically inappropriate. To avoid massive forces from interfering with the process of regeneration, the tooth might need to be removed from its occlusal position. Additionally, a full-coverage restoration would definitely be necessary for a better long-term outlook. Dentin MMPs MMPs are a group of 28 modular endopeptidases that play an important function in the remodeling of extracellular matrix and the control of extracellular signaling networks that, among other things, regulate inflammation, bone growth, and angiogenesis [ 11 ]. The cellular elements of both hard and soft tissues, such as epithelial cells, fibroblasts, osteoclasts, osteoblasts, as well as hypertrophic chondrocytes, chondroclasts, inflammatory cells, and odontoblasts, create these substances [ 12 ]. The majority of MMPs have a propeptide domain, which is in charge of maintaining the enzyme's latent conformation, a zinc-binding catalytic domain, which is in charge of their proteolytic function, and a hemopexin-like domain, which is in charge of protein-protein interactions as shown in Figure 1. Figure 1 Structure of MMPs MMPs contain a zinc-binding catalytic domain, an specificity determining site, a propeptide domain, and a hemopexin-like domain. Various proteolytic and also nonproteolytic methods trigger these zymogens, which enables them to perform the many tasks that they are intended for [ 13 ]. Their substrates can be used to categorize them, and these substrates are mostly defined by specificity-determining sites on their catalytic domain [ 14 ]. MMP-2, -3, -7, -8, -9, -13, -14, -20, -23, and -25 are MMPs that are present in dentin as shown in Figure 2 [ 12, 15 - 17 ]. Figure 2 MMPs present in dentin and their functions MMP activation and inhibition dynamics are still poorly understood, despite being one of the significant aspects of tissue remodeling in reaction to illness. MMP’s proteolytic activity has previously been linked to tissue deterioration. A significant rise in the activity of MMP-14 has also been linked to a carious response, though its precise function is still unknown. MMP-8 has been recognized among the principal collagenases in human dentin linked to carious lesions [ 3 ]. These endogenous MMPs perpetuate the illness via enzymatic activity they have and could be produced from the extracellular matrix or triggered by the process of caries. Additionally, like in the case of MMP-13, their diminished capacity has been linked to lowered decay risk [ 16 ]. Inhibiting MMP activity has been the main focus of efforts to delay or prevent illness. It has been determined that MMPs' enzymatic destruction of collagen fibrils is what causes resin-based restorations to fail to owe to hybrid layer degradation. Exogenous MMP inhibitors have enhanced clinical results of the restorations, which are resin-bonded so far by maintaining the bond strength and hybrid layer. Examples include tetracycline antibiotics and chlorhexidine [ 18 ]. In human dentin, endogenous Tissue Inhibitors of Metallo Proteinases I to IV (TIMP-I, TIMP-II, TIMP-III, and TIMP) have been found. Even though this increase occurs at the same time as raising MMP expression, TIMP-II expression rises during the caries process [ 3, 12 ]. Despite the fact that the MMP/TIMP ratio and the substrate-inhibitor specificity might help to describe and regulate the activity of MMPs in tissues, the importance of their co-expression has not yet been shown. It has also been suggested that TIMP signaling occurs independently of its MMP-inhibitory effect. Important elements of scaffolds utilized in regenerative endodontic operations include TIMP-I and TIMP-II expression [ 19 ]. Additionally, it is known that substances that promote pulp cell proliferation also increase the expression of TIMP [ 20 ]. In these circumstances, MMP counter-regulatory effect on TIMP signaling may explain MMP co-expression. In the end, matrix turnover must be balanced for regeneration. Knowledge of MMP-TIMP protein interactions and biological processes is necessary for maximizing MMP activity. MMPs are essential for tooth growth. The earliest MMPs to express are MMP-2 and MMP-9, which may help with basement membrane breakdown and signal the beginning of ameloblast and odontoblast terminal differentiation [ 21 ]. MMP-2 (and MMP-20) loss of function causes larger levels and a better reach of non-collagenous proteins known to stimulate dentin mineralization, which is significant at early stages [ 22 ]. Still, as development progresses, root dentin abnormalities are caused by the dysfunction of other MMPs, such as MMP-14 [ 17 ]. MMPs are not absurd proteins with similar roles, according to the available evidence. Their unique modes of action have the potential to influence tissue development, and their imbalance has important consequences for tissue integrity. The propensity for cell maturation, as well as the necessary remodeling of freshly deposited dentin matrix, are required for dentin regeneration and tertiary dentin creation. For these processes, MMPs are essential. In order to achieve regeneration, it may not be possible to totally stop their proteolytic activity. In fact, even though they are present throughout the disease, some MMPs may operate as a protective mechanism by aiding in the healing process. The optimum proteolytic activity and calcium affinities of MMP-3 are pH-dependent (range pH 5. 2-5. 6). Thus, at the essential pH for demineralization of enamel and dentin, as those observed in carious settings, its proteolytic activity is at its peak [ 23 ]. However, angiogenesis and reparative dentin deposition have both been linked to MMP-3 [ 24 ]. Utilizing MMPs to promote dentin regeneration MMPs have been linked to the control of developmental processes, the release of physiologically active molecules, and the modulation of repair during the production of tertiary dentin, among other things [ 25 ]. MMPs open pathways for progenitor cells to enter and also activate growth factors that control angiogenesis, the immune system, and cellular differentiation. When a disease is present, MMP activity may be produced by endogenous (immune cells), exogenous (bacterial products), and dentin matrix reservoirs. Bacterial byproducts from decayed lesions may also activate odontoblast MMP release by activating signaling cascades. High MMP-2 expression in odontoblasts next to reactive dentin has been linked to higher proteolytic activity in this region [ 26 ]. The maturation of collagen fibers and the beginning of mineral production in the freshly created dentin may be caused by this action [ 27 ]. MMPs aid in the removal of detritus and the development of tissue. The significance of these enzymes in the regenerative process is supported by animal models of tissue regeneration. One of the first steps in new limb regeneration following amputation is MMP overexpression. MMP activity is necessary for limb regeneration, and its suppression slows it down [ 28 ]. The most highly expressed MMPs in this particular model were MMP-9, MMP-3, and MMP-13 [ 29 ]. Similar to other injuries, pulp damage triggers an inflammatory response that includes the invasion of PMN’s cells and the production of proteases like MMP-9 [ 30 ]. These MMPs aid in the breakdown of exposed carious dentin, angiogenesis, and cell migration, which activates the processes that result in the deposition of tertiary dentin [ 26 ]. As a diagnostic and predictive indicator of pulpal inflammation, MMP-9 is being employed in endodontics to help direct clinical decisions [ 30 ]. Because MMP-3 has anti-inflammatory characteristics, it can reverse mild irreversible pulpitis in vivo. These characteristics include a reduction in the invasion of macrophages and antigen-presenting cells, as well as the suppression of Interleukin-6 (IL-6) production [ 31 ]. Additionally, independent of its proteolytic activity, MMP-3 can increase the synthesis of the connective-tissue growth factor (CTGF), which promotes the migration of dental pulp cells [ 32 ]. In vivo pulp damage models, MMP-3 is also localized to endothelial cells and promotes angiogenesis and reparative dentin deposition [ 24 ]. The lack of increased MMP-3 activity in pulps with irreversible damage supports the protein's potential for regeneration [ 33 ]. Future models of dentin regeneration should pay close attention to MMP-3 as a regeneration mediator. A palisade layer of odontoblasts that lines the pulp chamber's perimeter is in charge of producing dentin. These cells release bioactive chemicals during dentin deposition that direct the tissue's mineralization. Similar to this, these cells are prompted to secrete new dentin in response to dentin breakdown brought on by attrition, carious exposures, and chemical assaults. The chemical signals and cellular dynamics necessary for primary dentin production, however, are not present during regeneration and repair. As a result, the tissue is dependent on bioactive chemicals to promote the differentiation and proliferation of cells necessary for dentin regeneration and repair [ 34 ]. The bioactive chemicals that were once engaged in the natural deposition of mature dentin are stored in the tissue. As a result, the tissue has a defense system against environmental assaults. Dentin contains sequestered forms of Transforming Growth Factor-1 (TGF-1), Platelet-Derived Growth Factor- AB (PDGF-AB), Vascular Endothelial Growth Factor (VEGF), Placenta Growth Factor (PlGF), and Fibroblast Growth Factor-2 (FGF-2) [ 35 ]. These growth hormones are made soluble, which encourages angiogenesis, odontoblast differentiation, and tertiary dentin deposition. As a result, dentinal bridges that are denser, thicker, and more structurally similar to physiological dentin are produced [ 36 ]. Additionally, removing these components from plasma can maintain the viability of tooth-bud cells, which has led to the regeneration of teeth in porcine animal models [ 37 ]. A promising method for releasing growth factors and activating the genes that promote odontoblast differentiation is dentin conditioning with ethylenediaminetetraacetic acid (EDTA) [ 38 ]. Another way to release these elements from tooth tissues is through MMPs. These proteases may be exposed while still serving their purpose when phosphoric acid etch-and-rinse adhesive methods are used [ 39 ]. Numerous of these proteases have been linked to the promotion of pulp repair. Similar regeneration qualities have been shown in vivo using direct pulp capping agents made of dentin matrix components that have undergone MMP digestion [ 40 ]. These investigations recognized MMP-1, MMP-9, MMP-13, and MMP-20 as pulpal healing boosters. In tissue regeneration models, endogenous MMP activity has been used to transport growth factors from scaffolds. In recent work, Huang et al. [ 41 ] constructed a scaffold including growth factor-binding sites and an MMP-2 cleavage site. When activated by MMP-2 in vivo, such hydrogel scaffolds would be helpful for the release of growth factors. It has been demonstrated that concentrated venous blood growth factors, including PDGF-BB, TGF-1, VEGF, and others, can stop dental pulp cells from releasing proinflammatory cytokines and stimulate the regeneration of dentine-pulp complex in vivo [ 42 ]. Melatonin-induced dental pulp TGF - secretion has also been used to immunomodulate the pulpal inflammatory response to damage [ 43 ]. By attracting dental pulp stem cells, promoting their proliferation and odontogenic differentiation, as well as promoting pulp angiogenesis, overexpression of PDGF-BB encourages regeneration [ 44 ]. Similar to Bone Morphogenetic Protein (BMPs), FGF-2 and bioactive pulp-capping agents like BMP-2 and BMP-4 promote cell differentiation and tertiary dentin deposition [ 45 ]. The tissue engineering trio includes scaffolds, signaling molecules, and cells. Successful scaffolds must endure in tissues for a sufficient amount of time to permit cellular colonization before being degraded by enzymes [ 46 ]. In order to maintain cell colonization and long-term proliferation, two factors necessary for neovascularization and angiogenesis, MMPs have been used to remove hydrogel scaffold systems at the appropriate moment [ 47 ]. The scaffolds must be kept in place by the native cells that are present in the regenerating tissues. Anti-inflammatory cytokines like IL-10 that are released cause TIMP expression and stop MMPs from degrading the scaffold [ 48 ]. On the other hand, it is also known that inflammatory cells in the regenerating tissues secrete MMPs that modify these scaffolds and freshly deposited extracellular matrices [ 49 ]. As was already mentioned, MMPs play a crucial role in the development of tertiary dentin. These enzymes enhance the bioavailability of signaling molecules, the cellular processes that result in dentin healing, and the maturation of the dentin collagen matrix. In the mineralization process, activation of signaling molecules is crucial. MMPs play a role in the maturation and function of a class of non-collagenous proteins known as small integrin-binding ligand, N-linked glycoprotein (SIBLINGs), in addition to the collagenous components of the dentin matrix. This family of proteins contains the proteins Osteopontin (OPN), dentin-matrix protein-1(DMP-1), dentin sialo phosphoprotein (DSPP), matrix extracellular phosphoglycoprotein (MEPE), and bone sialoprotein (BSP-2). Both latent and TIMP-inhibited MMPs are activated by SIBLINGs, which bind exclusively to MMPs [ 50 ]. OPN/MMP-3, DMP-1/MMP-9, and BSP/MMP-2 are examples of the known partners. DSPP can be broken down by MMP-9 into dentin sialoprotein (DSP) and dentin phosphoprotein (DPP), and MMP-2 can also cleave DMP-1 to liberate physiologically active peptides [ 51 ]. There is currently no recognized MMP companion for DSPP or MEPE [ 52 ]. SIBLINGs make up the majority of the extracellular matrix proteins in dentin that have been phosphorylated, and they have been linked to mineralization and odontoblast development [ 53 ]. These two protein families' connections might give a chance for dentin healing. Conclusions MMPs are a family of proteinases that are in charge of dentin repair by modulating non-collagenous proteins and signaling molecules as well as matrix formation and remodeling. These proteinases have as many different activities as the protein family itself, and they play a developmental role in both illness and healing. The available literature demonstrates that not all MMPs have the same characteristics. Some may function more destructively than others, while vice versa. Dentin regeneration has so far been tackled in one of two ways: either by encouraging stem cells to deposit dentin or by creating scaffolds that will help mineralized tissue deposit where the dentin deficiency is. These multifarious proteins are excellent candidates for stimulating dentin regeneration because of the spatiotemporal modulation of MMP production, their multifunctionality, and their capacity to autoregulate. Future studies should concentrate on utilizing these enzymes' characteristics to encourage dentin regeneration.

10. 7759/cureus. 28463

Cureus

A Review on Techniques and Biomaterials Used in 3D Bioprinting

Three-dimensional (3D) bioprinting is a cutting-edge technology that has come to light recently and shows a promising potential whose progress will change the face of medicine. This article reviews the most commonly used techniques and biomaterials for 3D bioprinting. We will also look at the advantages and limitations of various techniques and biomaterials and get a comparative idea about them. In addition, we will also look at the recent applications of these techniques in different industries. This article aims to get a basic idea of the techniques and biomaterials used in 3D bioprinting, their advantages and limitations, and their recent applications in various fields.

Introduction and background Three-dimensional (3D) printing of biological material is a revolutionary technology through which we can print various materials ranging from simple muscle tissue, neural tissue, and cartilage, to an entire organ. In this process, we first construct a 3D model of the structure we want to print using patient's scans such as X-ray, CT, or MRI, which will then be printed in a layer-by-layer model taking care of every microscopic as well as macroscopic detail of the tissue. This model is then printed in a layer-by-layer fashion, which is then further processed to hold it together to function as a single unit [ 1 ]. While printing a particular structure, we need to keep in mind the properties of biomaterials used, such as biocompatibility, strength, stability, and immunogenicity, before selecting the correct biomaterial [ 2 ]. Bioprinting is not a single-step process; it involves various complex processes to print customized 3D structures for the patient, such as designing the structure with the help of computers using the patient's radiological imaging reports and then prototyping using a technique known as solid free form fabrication, which will take care of every microscopic as well as macroscopic detail of the tissue. With the progress in bioprinting technology and the qualities of biomaterials, 3D bioprinting can lead to various advantages in the short and long run. Although, at present, it seems scary to a normal person to think about having a printed organ in his own body, if this technology succeeds, it can save so many people waiting for years for organs [ 3, 4 ]. Other uses of 3D bioprinting are the treatment of burn wounds using artificial skin, bioprinting of bones and cartilage, drug testing, preparing diseased tissue models to check the treatment's efficacy before actually giving it to a patient, bladder implants, and heart valve implants. Besides so many advantages that we can have from 3D bioprinting, there are many challenges ahead of us, such as the technology being too expensive. This technology will only be advantageous for only a few people, leaving behind the poor who will have to wait for a donor. Also, as this technology is not yet so advanced, that makes it a very risky procedure as we still do not have all the information about the types of complications that can occur from this procedure. Also, there is still a long road ahead of us, which requires years of research to make this procedure successful [ 5 ]. The main goal of 3D bioprinting is to replace the non-functioning or defective tissue/organ with the new bioprinted one, which will function the same as the native organ structurally and functionally. This bioprinted tissue must know how to regenerate and differentiate on its own when implanted inside the patient's body. With the proper use of technology and the correct type of biomaterial, adequate tissue can be printed, which will perform all these required functions; therefore, adequate research in the field of biomaterials is needed to find the correct material that can work as native tissue. In this review article, commonly used bioprinting technology, their application, advantages, and limitations, along with types of biomaterials used in the field of 3D printing (both natural and synthetic) and their advantages and limitations have been discussed, as well as their application in the various industries [ 6 ]. Review Typically used techniques in bioprinting Among all the types of techniques used in bioprinting, the most commonly used methods are described in Figure 1 and the biomaterials used in them are described in Table 1. Figure 1 Different Types of Techniques Used in Bioprinting Table 1 Summary of Commonly Used Techniques in Bioprinting PEG, polyethylene glycol; PVA, polyvinyl alcohol; TCP, tricalcium phosphate; UV; ultraviolet; 3D, three-dimensional Techniques Procedures Biomaterials Applications Fixed deposition modelling Heat-sensitive plastic filaments are melted down and arranged in a layer-by-layer fashion to build a 3D object [ 2 ] Nylon, PVA, polycarbonate Regeneration of cartilage tissues, bone tissue; delivery of antibiotics; prosthetics [ 3 ] Extrusion-based printing Extrusion of the material using pressure through the nozzle of the printer is done to form the desired shape [ 2 ] Collagen, hyaluronic acid, alginate, PEG, gelatin, chitosan Aortic valve; neural tissue; muscle tissue; bones; implants [ 2 ] Selective laser sintering Solid 3D structures are formed using a powder arranged in a layer-by-layer fashion using a high-power laser [ 4 ] Ceramics, metals, polyamide Drug delivery; tissue engineering Stereolithography Photopolymers of high sensitivity are bound together using a beam of UV laser, heat, or electron beam Photopolymers Medical models and prototypes Inkjet Alternate powder and liquid binding material layers are added in a layer-by-layer fashion [ 5 ] Hydroxyapatite, Alpha- TCP, beta -TCP, PVA, PEG, PEG hydrogel Printing of biomolecules such as protein and nucleic acid Laminated object manufacturing Thin sheets are coated with adherent material, glued together in a layer-by-layer fashion, and then cut into the desired shape using a laser or metal cutter [ 2 ] Metals, Plastic, Paper Prototypes The advantages and limitations of the methods are illustrated in Table 2. Table 2 Summary of Advantages and Limitations of Different Bioprinting Techniques Techniques Advantages Limitations Fixed deposition modelling Low cost, quick processing, easy to operate, high porous materials can be made Less compatibility, high temperature destroys the material, lack of mechanical strength, only thermoplastics can be used Extrusion-based printing Long viability, can print highly dense material, low cost Pressure may affect cell viability, cannot print complex tissue Selective laser sintering Good support offered from a powder bed, many types of materials can be used Highly expensive, printers are large and complex to install, process is slow Stereolithography High resolution, high viscous material can be printed [ 3 ]. Ultraviolet rays used are toxic and make skin cancer-prone, slow process, cell viability is short Inkjet Quick processing, high resolution, long viability, more compatible, multicolor printing is possible [ 5 ] Low mechanical strength, nozzle gets blocked frequently because of the highly dense material used Laminated object manufacturing Low cost, quick processing, easy to operate Difficulty in manufacturing complex tissues Biomaterials Typically used biomaterials used in 3D printing have been illustrated in Figure 2. Figure 2 Classification of the Biomaterials Used in Bioprinting PCL, polycaprolactone; PLGA, polylactic-co-galactic acid Natural Polymers Naturally occurring polymers can be derived using various physical, biochemical, or chemical methods. Natural polymers are compatible, can hold the fluid, and can be easily dissolved in different solvents such as phosphate buffers and cell culture solutions, making them more tissue friendly. Due to these qualities, it is possible to print it in a layer-by-layer manner, producing a model that will mimic a natural organ if placed in a stable environment [ 6, 7 ]. One of the critical properties of these naturally occurring polymers is that when provided with a controlled environment such as normal temperature, adequate water, and proper medium to grow, they can mimic cells or tissue, undergo proliferation, maturation, and differentiation, and coordinate with surrounding structures [ 8 - 10 ]. One major drawback of natural polymer is that all these activities are majorly affected if the surrounding environment becomes unstable, such as an increase in temperature, dehydration, or the nature of the solvent in which it is dissolved. Some commonly used natural polymers are alginate, gelatin, collagen, chitosan, and hyaluronic acid, and are described below. Alginate Alginate is derived from the cell walls of Phaeophyceae (brown algae) and is used in the form of salts of alginic acid. Wang first used alginate in the form of sodium alginate, but the problem faced was its gelation point, which is 0°C, while 3D bioprinting was done at room temperature; therefore, it was crosslinked with other metals such as calcium and barium to increase its compatibility and mechanical strength [ 11 ]. An important thing to take care of while using alginate is to use it in adequate concentration because if used in less concentration, the model's strength is majorly affected. All the activities such as proliferation, growth, and maturation are affected if used in high concentration. Therefore, it is crucial to use the proper guidelines regarding the concentration of alginate to be used for 3D bioprinting. Also, as alginate shows the property of delayed degradation, it is recommended to use alginate in an oxidized form, which is expected to show increased degradability and will be more suited for 3D bioprinting [ 12 ]. The chemical structure of alginate is given in Figure 3. Figure 3 Chemical Structure of Alginate Hyaluronic Acid Hyaluronic acid is an integral part of the extracellular matrix, which plays a major role in the proliferation of cells and angiogenesis. Due to its high cell adhesive property and water-absorbing quality, it can be used to change the viscosity of other polymers such as gelatin. As with other natural polymers, hyaluronic acid is crosslinked with synthetic polymers to increase its compatibility. One example is the crosslinking of hyaluronic acid with methyl acrylate forming a rigid non-biodegradable polymer known as HAMA (hyaluronic acid methylacrylate) [ 13, 14 ]. Another polymer formed via crosslinking is GeIMA which, when used in combination with HAMA (HAMA-GeIMA), will increase its mechanical strength and compatibility. It has been proven that the 1:4 ratio of GeIMA:HAMA is an adequate ratio to improve the compatibility of the polymer formed (np 101). As this Combination shows superior qualities, it has been applied in the bioprinting of musculoskeletal, cardiac, and neural tissues [ 15 ]. The chemical structure of Hyaluronic acid is given in Figure 4. Figure 4 Chemical Structure of Hyaluronic Acid Collagen Collagen is widely known to support the skin, ligaments, bone, tendon, and cartilage due to its resistance and toughness. Type 1 and type 2 are most often used in musculoskeletal repair using 3D printed models. Collagen has been observed to promote proliferation, maturation, and differentiation of bone and cartilage cells [ 16 ]. As seen in other polymers, using collagen in bioprinting is best done when combined with other polymers to increase its viscosity and decrease its degradation compared to using collagen only. It is commonly crosslinked with alginate, agarose, hyaluronic acid, and fibrin [ 17 ]. To increase its compatibility, collagen has also been crosslinked with heparin sulfate and polyurethane for printing conduits, which can help nerve repair [ 18 - 19 ]. However, the drawback of using collagen is its easy solubility in acids, making it temperature- and pH-dependent. Also, the rapid degradation of collagen by collagenase and metalloproteinase in the body makes it difficult to use [ 20 ]. The chemical structure of collagen is given in Figure 5. Figure 5 Chemical Structure of Collagen Gelatin Gelatin is a linear molecule that is obtained by breaking collagen. Being a natural substance, it is not toxic, and is low in immunogenic properties, hydrophilic, and highly degradable, which makes it a special polymer. Before printing gelatin, it is combined with culture media to make it denser [ 21 - 26 ]. Many agents, such as hormone growth-promoting factors, can be crosslinked with gelatin molecules. Heparin, at the time of gelation, and other naturally occurring polymers such as hyaluronic acid, agarose, fibrin, collagen, and chitin will increase its mechanical strength and compatibility [ 27 ]. The combination of gelatin with synthetic polymers in the presence of UV light has led to the formation of GeIMA, which, when used in combination with HAMA (HAMA-GeIMA), will increase its mechanical strength and compatibility. It has been proven that the 1:4 ratio of GeIMA:HAMA is an adequate ratio to increase the polymer formed compatibility. Another combination of gelatin is crosslinking of gelatin with chemical agents such as calcium chloride to improve the stability and degradation properties of gelatin hydrogels such as gelatin-fibrin or gelatin-alginate combination [ 28, 29 ]. The chemical structure of gelatin is given in Figure 6. Figure 6 Chemical Structure of Gelatin Fibrin Fibrin is a natural polymer as it is formed in blood in the presence of thrombin due to the rapid polymerization of fibrinogen [ 30 ]. Although fibrin has been found superior in its properties such as compatibility compared to other natural polymers to increase its efficacy, it is combined with other natural polymers to overcome its low strength, less viscosity, high degradation, and gelation properties when used alone [ 31 ]. The recent trend is to combine natural polymers and crosslink them using chemical agents to form a hybrid type of polymers in various combinations such as gelatin-chitosan-alginate-fibrinogen and gelatin-hyaluronic acid-glycerol-fibrin. This combination helps create a more stable structure that can print quickly, and those models can survive longer in the body's environment. Fibrin and its combination with other polymers are being used in bioprinting of skin, which will be helpful in early wound closure in many cases and early regeneration of tissue and its vasculature [ 32 ]. The formation of fibrin is illustrated in Figure 7. Figure 7 Diagram Illustrating Formation of Fibrin Chitosan Chitosan is usually derived from shrimp shells and is formed from the hydrolysis of chitin. Like other natural polymers, it is low in strength and has degradable properties; therefore, a similar combination of crosslinking with chemical agents is done with collagen, alginate, and gelatin to increase its viscosity and biodegradability, and to make it more compatible, it is used to repair rigid structure such as skin, bone, and cartilage [ 33 - 38 ]. The chemical structure of chitosan is given in Figure 8. Figure 8 Chemical Structure of Chitosan Synthetic Polymers Synthetic polymers are made artificially by humans in a laboratory using chemicals in the appropriate environment required for their production; they are high in strength and resistance. The main advantage of synthetic polymers is that we can modify them easily as they can withstand changes in temperature and pH and can be processed according to our needs due to their increased resistance and mechanical strength. Since the gelation temperature of synthetic polymers is shallow compared to natural polymers with a very high melting temperature, they are very suitable for models for 3D bioprinting; therefore, formed polymers are inert, are difficult to degrade, and have a high tensile strength. Polyethylene Glycol Polyethylene glycol is a linear synthetic polymer that is compatible, is low in immunogenicity, and has a high affinity for water, making it well qualified for bioprinting. Another name for polyethylene glycol is polyethylene oxide [ 39, 40 ]. Polyethylene glycol cannot adhere appropriately to the cells; it is crosslinked with other molecules such as carboxyl group, acrylate, or thiol group to make it more suitable for use in the repair of soft tissues. Polyethylene glycol can also be polymerized in the presence of UV light to increase cell encapsulation rate and its mechanical strength. Using the Inkjet bioprinting technique, PEG has also been crosslinked with GeIMA to increase its strength for the bioprinting of rigid structures such as cartilage and bone [ 41, 42 ]. Since polyethylene glycol is not degraded on its own, hydrolytic blocks such as polycaprolactone and PGA is used to increase its degradation rate. The chemical structure of polyethylene glycol is given in Figure 9. Figure 9 Chemical Structure of Polyethylene Glycol Polycaprolactone Polycaprolactone is a partially crystalline polymer that can be easily degraded naturally in our body [ 43 ]. It is a thermoplastic polymer produced at the temperature of -60°C when combined with other agents to change its mechanical structure and degradation rate. It can be called an ideal material to be used in fused deposition modelling technology of 3D bioprinting [ 44, 45 ]. As it is done in all other synthetic polymers, polycaprolactone is crosslinked with other bioagents such as polycaprolactone-alginate to increase its cell adhesive property for regeneration of cartilage. Polycaprolactone has also been combined with GeIMA using UV light to increase the strength and stability of the scaffold. GeIMA concentration is proportional to the hardness of the scaffold and is widely used in cartilage and bone regeneration [ 46 ]. Other uses of polycaprolactone are to form sutures and in devices such as drug delivery system [ 47 ]. The chemical structure of polycaprolactone is given in Figure 10. Figure 10 Chemical Structure of Polycaprolactone Polyurethane Polyurethane is a linear biodegradable polymer that shows outstanding compatibility and mechanical properties [ 48 ]. Polyurethane, when used alone, is inert and cannot be degraded. Therefore, it is crosslinked with other materials to increase its compatibility and stability. Waterborne polyurethane is one such type that removes its problem of temperature and pH dependency, which is mainly dependent on its short segment (diol segment). Waterborne polyurethane is now used to repair chondrocytes and nerve cells [ 49 - 51 ]. Polyurethane has also been crosslinked with other bioagents such as adipose stem cell-fibrin-alginate-gelatin and cryoprotectant to protect against the damage from low temperatures to synthesize them. Another form of polyurethane is an elastic variety of polyurethane, which has been widely used for nerve repair and vascular repair conduits. Combining polyurethane with polycaprolactone and polyethylene glycol increases its mechanical strength, stability, compatibility, and biodegradability [ 52 ]. The chemical structure of polyurethane is given in Figure 11. Figure 11 Chemical Structure of Polyurethane Polylactic-Co-Galactic Acid Polylactic-co-glycolic acid is formed using two polymers, lactic acid and glycolic acid, by copolymerization. It is usually seen that the transition temperature of polylactic-co-glycolic acid is around 40-60°C, and glycolic acid and lactic acid are used in the ratio of 1:3 [ 53 ]. It has been observed that the degradation rate of polylactic-co-galactic acid depends on the concentration of glycolic acid used while synthesizing. Polylactic-co-galactic acid is mainly used where high mechanical support is required [ 54 ]. It can also be combined with other agents such as growth-promoting factors or adipose stem cells to make it more useful and compatible for making the complicated structure of 3D bioprinted organs. PLGA can also be synthesized at low temperatures to create a complex organ structure with fibrin hydrogel to act as a native organ when transplanted [ 55 ]. The chemical structure of poly-co-galactic acid is given in Figure 12. Figure 12 Chemical Structure of Polylactic-Co-Glycolic Acid Recent advances in bioprinting technology The application of the internet of things (IoT) with the technology of bioprinting has led to breakthroughs in surgical techniques [ 56 ]. The ultimate goal is to break the chain of years and years of waiting for a donor organ and to print an entire organ that will be structurally and functionally similar. The main organs on focus to print are our heart, bone, skin, cartilage, and tendon. Besides focusing on printing an entire organ, 3D bioprinting has been used in various other branches, which have been described in Table 3 [ 57 - 60 ]. Table 3 Summary of Application of Bioprinting in Different Industries Industries Uses Dental Crowns, filling, implants, fixtures Pharmacy Drug delivery Medicine Pharmacy, prosthetics, hearing aids, orthopedic screws/plates Food Cookie, candy, pizza Automobile industry Prototypes, spare parts Conclusions Even though there is still a long road ahead of us to print an organ, this cutting-edge technology has shown a promising potential that will change the lives of thousands of people dying every day because of the need for a donor organ. However, implanting a printed organ in a human body is still scary for many people. If successful, it will solve many problems, such as a long waiting list for a transplant and issues of organ rejection, and will completely change the face of medicine. Since, at present, there are not enough biomaterials that can be used in 3D bioprinting, there is a high need for research in this matter as this shows the potential of saving the lives of many patients who require a transplant. Still, in its early phases, bioprinted organs have already proved functional in labs, but there is a long road in front of us until they will be transplanted into an actual human body.

10. 7759/cureus. 28647

Cureus

Injectable Platelet-Rich Fibrin - A Revolution in Periodontal Regeneration

As of a few years ago, platelet concentrates have been applied in a variety of medical and dental procedures. A notable aspect is that platelet-rich fibrin (PRF) is the most commonly utilized platelet concentrate in the field of dentistry. The most significant modification that was used over the years but had the biggest impact was injectable platelet-rich fibrin (I-PRF), which has more special properties. Additionally, the results of this I-PRF have been useful. The solid platelet-rich fibrin (PRF), which is a noticeable feature and has a low speed and duration in centrifugation, is the main advantage of I-PRF. I-PRF is primarily found in liquid form as PRF. It facilitates the quickening of increased vascularization and aids in accelerating the healing of wounds. An autologous blood concentration known as I-PRF has been known for many years. The advantage of I-PRF is that it exhibits constant release of growth factors and promotes cell migration by announcing the expression of type I collagen and transforming growth factor mRNA. The majority of the time, plastic and orthopedic operations use injectable platelet aggregates. It also reduces adverse reactions to transplanted material as compared to other grafting techniques. Additionally, it makes numerous other operations, like regenerative ones, much better options. In circumstances where it has been noticed, I-PRF is helpful and crucial in periodontics for bone regeneration and wound healing. It is therefore not difficult to predict that this fully autologous blood concentrate, which is now being utilized in numerous applications and requires little invasiveness, will become even more frequently used in the future. This review paper contains the differences between platelet-rich plasma (PRP) and PRF, the development of diverse platelets, and the use of I-PRF in periodontal therapy.

Introduction and background Higher amounts of peptide growth factors can now be transported by platelet concentrates into periodontal lesions. Platelet-rich fibrin (PRF), a second-generation preparation included in the category of platelet-rich plasma (PRP) and first stated by Choukroun et al. in 2001, is the most effective platelet concentrate out of all those available [ 1 ]. First, in 1954, Kingsley introduced platelet concentrate, which later became known as platelet-rich plasma (PRP) [ 2 ]. It was created as a thrombocytopenia treatment. The initial attempt to use concentrated plate growth factors to aid in the healing of wounds during and after surgery involved several tries. Marx et al. first proposed the growth factor that is present in PRP and its concentration in their paper published in they discuss the results of platelet-rich preparation used in maxilla face reconstruction applications in articles from the year 1988 [ 3 ]. PRP preparation techniques vary according to protocols and take between 30 and 60 minutes to complete. In PRP, there are essentially two key centrifugation techniques used. Regarding the first centrifugation process, tubes coated with ethylene diamine tetra acetic acid (EDTA) and citric acid primarily serve to inhibit natural coagulation. In the second centrifugation step, bovine thrombin, calcium chloride, or any other artificial coagulant is introduced to the plasma to create artificial coagulation after the erythrocytes have been settled by the first centrifuge. It is renowned for its quick centrifugation technique as well [ 4 ]. First-generation platelet concentrate, or PRP as it is commonly known, can be in gel or liquid form. Following thrombin and calcium activation of centrifuged blood, it also manifests as a frail fibrin network. The issue is that we artificially included bovine thrombin and calcium chloride during the procedure; therefore, the product is not entirely autologous. Ninety-five percent of the platelets in PRP come from the blood. These are the cells that directly influence other cells, such as osteoblasts, connective tissue cells, epithelium, and periodontal ligament cells. Even though platelet-rich plasma is crucial for delivering growth factors at various stages of wound healing, its major goal was to purge leukocytes from blood concentrates [ 5 - 6 ]. The two main drawbacks of platelet-rich plasma are that it is expensive and takes a long time to produce. The preparation process also involves a lot of steps. The second significant drawback is that the fibrin matrix structure created by artificial coagulation is stiffer than the fibrin matrix structure created by spontaneous coagulation [ 7 ]. Due to the drawbacks of first-generation PRP, platelet-rich fibrin, a second-generation blood product, was found. Since the discovery of platelet-rich plasma, platelet concentrates have been produced entirely autologously without the use of anticoagulants such as calcium chloride or artificial bovine thrombin throughout the production process. Simply put, it is created primarily by utilizing the patient's blood without the addition of any supplements. By manipulating the patient's blood, platelet-rich fibrin is essentially a surgical biological preservative; it also has a distinctive morphology. Additionally, it is crucial in preparing cells for tissue regeneration and is helpful in plastic surgery, maxillofacial surgery, and implant procedures. In the branch of periodontal therapy, PRF has many benefits in treating different types of periodontal defects [ 8 - 9 ]. PRF has another advantage: it has a complex three-dimensional fibrin framework. And when it comes to platelet rich fibrin (PRF) classification, it is done on the basis of centrifugal speed, time required for centrifugation and type of test tube used. Details of various types of PRF and their respective mode of preparation are given in Table 1. Table 1 Types of PRFs PRF - platelet-rich fibrin Centrifugal speed Centrifugal time Tube type Nature of the obtained PRF Platelet rich fibrin (L-PRF), Choukroun 2000 [ 1 ] 2700 rpm 12 minutes Glass tube Solid Titanium platelet-rich fibrin (T-PRF), Tunali 2014 [ 10 ] 2700 rpm 12 minutes Titanium tube Solid Advanced platelet-rich fibrin (A-PRF), Choukroun 2014 [ 11 ] 1300 rpm 14 minutes Glass tube Solid Albumin platelet-rich fibrin (Alb-PRF), Fujioka 2020 [ 12 ] 1300 rpm 8 minutes Glass tube Solid Injectable platelet-rich fibrin (I-PRF), Mourao 2015 [ 13 ] 700 rpm 3 minutes Plastic tube liquid Injectable platelet-rich fibrin (I-PRF) is the most recent and successful advancement in PRF. In essence, it was created by slowing down the liquid-based centrifugation approach and omitting the formation of a PRF membrane. I-PRF is referred to be an advanced type of PRF since it is injected (autologous PRF) into afflicted soft tissues, mucous membranes, or skin. It also has special qualities in the regeneration of human tissues. Gene therapy, tissue engineering, and platelets have all been demonstrated to be effective sites for signaling pathways such as platelet-derived factor growth. Biological intervention in regenerative therapies primarily comes in three forms. Injectable platelet-rich fibrin was created in 2001 by Choukroun et al. [ 1 ]. Leukocyte platelet-rich fibrin, a complex three-dimensional fibrin structure made primarily of 97 percent platelets and 50 percent leukocytes, was first created. Three layers are created after centrifugation: the uppermost layer is made up of platelet-poor plasma, the middle layer is made up of fibrin clots with a high concentration of platelets, and the lower layer is made up of red blood cells. But because these three coagulation layers formed without separation and the resorption duration is sufficient in soft healing, the desired outcome could not be achieved [ 10 ]. As a result, researchers began to design several types of PRF. To prevent the potential of silica particles hanging in the fibrin structure in a glass tube traveling to the patient, titanium platelet-rich fibrin (T-PRF) has been created. T-PRF has been seen to stay in tissue for longer than 30 days without causing any problems, and because of its lengthy resorption time and abundance of growth factors, this particular PRF was helpful in the repair of soft tissues [ 11 ]. Advanced platelet-rich fibrin (A-PRF) was created using a low centrifugation approach, which has resulted in a significant rise in the number of inflammatory cells and growth-promoting substances. As a result, in this instance, regenerative potential has grown [ 12 - 14 ]. In PRF, coagulation starts when blood and silica in a glass tube come into contact. T-PRF has a tighter fibrin network structure and contacts the titanium surface rather than silica when blood comes into touch with it. When compared to PRP, PRF's major drawback was that it was available in solid form. As a result, an I-PRF was produced; after a short time and low centrifugation speed, it was acquired in liquid form without building a PRF membrane [ 15 - 16 ]. Since injecting solid platelet-rich fibrin was not achievable, Miron et al. conducted fundamental research and discovered that liquid platelet-rich fibrin could be created by reducing the centrifugal speed and time duration below the specified levels. He claimed that a centrifugal speed of 60 g for three minutes allows for separation before clots have a chance to develop and also prepares the residual liquid in the state. Additionally, it was noted that only 1 to 1. 5 ml of I-PRF were volumetrically present in a test tube containing 10 ml. The injectable platelet-rich fibrin can be injected into the skin or scalp of the face, and it remains a liquid for 10 to 15 minutes before it solidifies into a clot [ 17 ]. Additionally, PRF is crucial for the healing of wounds because it controls immunity, promotes angiogenesis, traps circulating stem cells, induces collagen synthesis, and stimulates the growth of fibroblasts and osteoblasts as well as the prolonged delivery of growth factors to the area where a wound is actually present. Review Preparation of I-PRF There are different types of preparation methods given by different researchers. Take a test tube, fill it with 9 to 10 ml of blood without adding any preservatives, and centrifuge it for two to three minutes at a speed of 3300 rpm to produce an orange-colored fluid that is thought to contain injectable platelet-rich fibrin, according to Mourao et al. [ 18 ]. Then, in 2009, AL-Malawi declared that, in accordance with the low-speed configuration approach, blood must first be collected in a test tube and immediately kept in a centrifuge at 600 rpm, 44g, for eight minutes. Following this procedure, a yellow i-PRF was created at an upper level, and other components were created or were already existent at a lower level [ 19 ]. Properties of I-PRF The following are some of the characteristics of PRFs: i-PRF has demonstrated that it experiences increased cellular migration. Additionally, it was shown that PRP and i-PRF demonstrated comparable tissue compatibility. Additionally, we can see that i-PRF has a bone transplant bonding mechanism that helps with the right adaption of the defect area. Fibronectin, an extracellular glycoprotein, is the main component of injectable platelet-rich fibrin. Fibronectin has a big molecular weight. Additionally, applying fibronectin to the surfaces of roots promotes cellular growth. From supra-crestal components to periodontal ligaments, cellular growth spreads. Last but not least, I-PRF offers higher biologic qualities than PRP. Merits and demerits of I-PRF Merits It is simple to prepare and use, and there is no biological modification. Additionally, it facilitates cellular motility and cytokine enmeshing. The majority of medication is in injectable form, which also lessens potential consequences. Additionally, because more growth factors are produced, it has a greater ability to activate regenerative cells. In addition, it creates a tiny fibrin clot that allows it to function as a dynamic gel. Last but not least, and perhaps most crucially, it is a straightforward and affordable procedure regardless of one's financial situation. Additionally, it is crucial for the release of growth factors for 10 to 12 days. Demerits Because I-PRF is made in small amounts from autologous blood, it only applies to a small portion of general surgery. The primary clinical benefit of I-PRF is based on the short handling time between blood collection and centrifugation since platelet-rich plasma is created without the use of additional anticoagulants. Another significant drawback is that the fibrin matrix is only usable for that individual donor since it contains circulating immune cells and highly antigenic plasmatic chemicals. Additionally, if stored I-PRF is not used right away, it may get contaminated with germs. Application in periodontal therapy Mourao et al. , in 2015, stated that platelet-rich plasma can be replaced by I-PRF when used with biomaterials in bone grafting as a platelet concentrate for bone regeneration [ 18 ]. Injectable platelet-rich fibrin has the ability and potential to release larger amounts of a variety of growth factors, according to research by Miron et al. [ 17 ]. Miron et al. from 2017. Additionally, it involves the expression of increased quantities of platelet-derived growth factor (PDGF), transforming growth factor (TGF), collagen 1, and fibroblast migration [ 17 ]. Chenchev et al. (2017) demonstrated through successful radiographic and clinical outcomes that combining advanced platelet-rich fibrin (A-PRF) with injectable platelet-rich fibrin (I-PRF) is beneficial for bone argumentation of the alveolar ridge prior to or during implant placement [ 20 ]. According to Wang et al. (2018), in control tissue culture, PRP promotes osteoblast migration by a factor of two, whereas i-PRF displays a factor of three, indicating that i-PRF exhibits stronger osteoblast differentiation and proliferation [ 21 ]. In accordance with Varela et al. (2018), I-PRF, which contains platelets, leukocytes, type 1 collagen, osteocalcin, and growth factors, is an excellent or extremely helpful option for the healing of soft and mineralized tissue [ 22 ]. According to Gode et al. 2019, I-PRF improved the postoperative survival rate of diced cartilage [ 23 ]. According to Izol et al. I-PRF has a favorable impact on root coverage in free gingival graft surgery [ 24 ]. Ozsagir et al. 2020 found that for people with thin phenotypic, combining injectable platelet-rich fibrin with micro-needling had the greatest potential to increase gingival thickness. The results also revealed that the first step in non-surgical approaches for enhancing and improving gingival thickness can be thought of as a combination of injectable platelet-rich fibrin and micro-needling [ 25 ]. Turer et al. , in 2020, stated that gingival recession decreases more in group applied I-PRF in operations with coronally advanced flap with a connective tissue graft [ 26 ]. Adding I-PRF to a coronally advanced flap and combining it with a connective tissue graft resulted in the development of increasing keratinized tissue height and decreasing recessive depth when compared to combining the coronally advanced flap (CAP) only with a connective tissue graft, according to research by Turer et al. in 2020 [ 26 ]. Combining advanced injectable platelet-rich fibrin and injectable platelet-rich fibrin appears to improve bone formation in alveolar clefts while reducing bone resorption and increasing bone volume, according to Dayashankara Rao et al. in 2021 [ 27 ]. He added that secondary alveolar grafting, if necessary, improves or boosts periodontal health. Uses in various fields I-PRF is used as an injection for a variety of conditions, including osteoarthritis, meniscus healing, alopecia, sports injuries, tendon/ligament injuries, musculoskeletal regenerative producers, and acne. It is also used in areas such as facelift surgery, knee arthroplasty, and heart surgery to reduce the incidence of infections. Conclusions Clinicians and researchers in the field of dentistry need to conduct much more research in tissue transmission engineering to fully understand platelet concentration's benefits and applications in various fields. It has beneficial effects to allay worries about the disease and immunogenic reactions because it is an entirely natural, physiological, and affordable source of an autologous product. There is more to learn about platelet concentration, and more work needs to be done on tissue transmission. And in this regard, injectable platelet-rich fibrin-which introduced the usefulness and functionality of the application of platelet concentrates-was the cleverest development in the field of platelet-rich fibrin. Additionally, it affects osteoblastic behavior, which aids in the significant release of growth factors when combined with a variety of biomaterials. Therefore, the presence of platelets and growth factors can convert an osteoconductive graft into an osteopromotive one.

10. 7759/cureus. 28969

Cureus

Rapid Prototyping in Maxillofacial Rehabilitation: A Review of Literature

This review focuses on fast prototyping advancements in the field of maxillofacial prosthodontics, as well as the various methods for fabricating maxillofacial prostheses. As of date, the interface and software used for processing and designing maxillofacial prostheses are costlier, atypical for the specific purpose, and only reachable to highly trained dental specialists or computer-aided design (CAD) engineers. This review is a summary of all rapid prototyping trials conducted in the mentioned context of three-dimensional (3D) printing of maxillofacial prostheses, treatment modalities, and future perspectives relating to rapid prototyping in dentistry. We performed a search of relevant articles on Google Scholar and PubMed, which yielded a total of 21 articles for full-text reviews. After excluding some articles based on the exclusion criteria, a review was conducted. This study gives a comprehensive discussion of current issues and future ideas for integrating digital technology with conventional techniques.

Introduction and background Rapid prototyping (RP) is an industrial revolution that has evolved hastily. People are interested in innovation in general, particularly when the eventual result can give tangible benefits. RP is a valuable tool for prosthodontic design and simulation, and it is the technology of the future. The transition from visual to the visual and tactile depiction of bodily objects ushers in a new type of collaboration known as "Touch to comprehend. " The birth of the newer technology capable of directly manufacturing bodily items from graphical data generated using computer software is discussed in this article [ 1 ]. From graphical computer data, mechanical models are created and this type of computer-aided prototyping is RP. It can be done in two ways: subtractive and additive, a term that refers to a substance that is commonly utilized. The additive manufacturing (AM) method varied from old-style subtractive manufacturing principles and is now used in a variety of fields, including personalized medicine, aerospace, and dental specialty. This method of manufacturing allows for the rapid manufacture of custom-based complicated parts, making it a viable option for self-growing robot development [ 2 ]. Inkjet printing, fused deposition modeling (FDM), stereolithography, selective laser sintering, photopolymer jetting, and printing with the precipitate binder are examples of additive manufacturing technologies. RP is broadly categorized into additive and subtractive technology. An overview of various types of RP used is mentioned in Figure 1 [ 3 ]. Figure 1 Overview of Different Types of Rapid Prototyping Used in Dentistry 3D: three dimensional Review Material and methods The protocol for this review was registered with the International Prospective Register of Systematic Reviews (PROSPERO) with the registration number CRD42021251023. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to conduct this review. The PRISMA flowchart for the selected studies is given in Figure 2. Figure 2 PRISMA Flowchart for the Studies Included in the Systematic Review *The articles excluded after reading the title PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses The assessment was based on the population, intervention, control, and outcomes (PICO) study criteria. The electronic search on the Google Scholar database provided a total of nine articles that were considered potentially relevant. The texts found using the “[AND] & [OR]” Boolean operators in between the search words "Rapid Prototyping", "Maxillofacial Prostheses", "3D Printing", "Stereolithography", "Dentistry", "Dentofacial Prostheses" were 88. In the second phase of article selection, all articles selected needed to be in the English language. A total of 65 articles were excluded after reading the title, and all duplicate articles were excluded. A total of 21 articles were selected for the systematic review. Of these 21 articles, after reading the complete text, the most relevant nine articles were selected for the systematic review. The search strategy showing article search through PubMed database search is summarized in Table 1 Table 1 Flowchart Showing Article Search Through PubMed Database The keywords used were "Rapid Prototyping", "Maxillofacial Prostheses", "3D Printing", "Stereolithography", "Dentistry", and "Dentofacial Prostheses" using the “[AND] & [OR]” Boolean operators in between the search words Search results combined after screening the PubMed database 48 Articles not in the English language excluded 1 Articles excluded after reading the title 24 Duplicate articles excluded 2 15 articles were searched for full texts 9 articles were excluded, and a total of 6 articles were selected for the review The search strategy showing article search through Google Scholar database search is summarized in Table 2. Table 2 Flowchart Showing Article Search through Google Scholar Database The keywords used were "Rapid Prototyping", "Maxillofacial Prostheses", "3D Printing", "Stereolithography", "Dentistry", and "Dentofacial Prostheses" using the “[AND] & [OR]” Boolean operators in between the search words Search results combined after screening the Google Scholar database 40 Articles not in the English language excluded 1 Articles excluded after reading the title 21 Duplicate articles excluded 2 16 articles were searched for full texts 13 articles were excluded, and a total of 3 articles were selected for the review Discussion The use of 3D printing technology in several aspects of modern dental medicine has permitted the fabrication of sophisticated prosthodontic, surgical, and orthodontic devices that require the molding materials to be flexible and abrasion-resistant (Table 3 ). Different materials, e. g. , composites, polymers, ceramics, and metallic blends, are employed for additive manufacturing [ 4 ]. Innovations in molding materials and forming procedures have improved RP techniques to the point where this technology is used for more than just prototyping; it is also used to reproduce real functional elements [ 5 ]. The feasibility of this technique is increasing in a variety of dental practice fields, including oro-maxillofacial surgery and prosthesis, the production of surgical guides or physical models in dental implant therapies, and prosthodontics [ 6 - 10 ]. Table 3 Applications of Rapid Prototyping in Dentistry Applications of rapid prototyping in dentistry Prosthodontics Wax pattern fabrication, direct prosthesis milling, 3D graphic data for complete denture fabrication, fabrication of maxillofacial prostheses and obturators, guided Implant surgeries, training and research Endodontics 3D visualization of complex canals, accurate diagnosis and treatment planning, training and research Orthodontics Diagnosis and treatment planning, fabrication of appliances, aligners, lingualized orthodontics, 3D models for orthognathic surgery Oral and Maxillofacial Surgery Fabrication of surgical guides, assessment of cases Dental prostheses such as crowns, removable and fixed partial dentures (RPDs and FPDs), and metal copings can also be planned, manufactured, and developed using RP techniques. This technique saves time and intervention in traditional prosthesis fabrication and also aids in the elimination of any flaws caused by human skills. To create frameworks for cast partial dentures, digital dental surveying and RP-produced patterns can be used [ 11 ]. Furthermore, RP also reduces the amount of extra-oral time required for autogenous tooth transplantation. The dental practice has profited from RP in the accurate reconstruction of maxillofacial defects as well as in osteogenic distraction with promising results [ 12 - 15 ]. Further discussion is mainly focused on the reconstruction of maxillofacial defects and prosthesis fabrication using rapid prototyping. 3D Printing Technique Data from cone beam computed tomography (CBCT) and optical scanners (IOS) pictures form the basis of 3D printing technology. This information is then transformed into a standard tessellation language file (STL), which can then be imported into 3D modeling software and altered to match the clinician's manufacturing requirements. Clinicians then upload the files to their preferred printer after making these changes. The prosthesis might be printed directly or a mold could be made for more traditional silicone manufacture [ 12 ]. Stereolithography (SLA), selective laser sintering (SLS), inkjet-based systems, and fused deposition modeling (FDM) are the most frequently used technologies in dental practice. Digital light processing (DLP), SLA, and material jetting (MJ) are the three most prevalent categories of 3D printing used in dentistry. On top of the building printer platform, the machine uses additive fabrication processes to create a prosthesis. Prostheses can be printed using a variety of materials, including ceramic, metal, and thermoplastic resin. Following manufacture, post-manufacturing operations are carried out to verify that the product is free of flaws and correctly processed; the scope of these steps varies according to the printer type and the material being used. It should be highlighted that the correctness and precision of each printer type are greatly indicative of the quality of the printer, the technology used, the materials, the settings in the software, and the post-manufacturing refining process. The interconnection of all of the features has a greater impact on overall quality than the differences in production processes like SLA, DLP, and MJ [ 15 - 17 ]. Use of RP Techniques in Dental and Facial Prosthetics RP techniques are now observed as a promising and satisfactory alternative for the fabrication and manufacture of dental prostheses. Molding of a dental (facial) prosthesis and metal casting mold (shell) is now performed in a short period of time. Using an incremental printing method, 3D printing creates ceramic casting molds for metal casting. Many time-consuming steps and labor-intensive work of the traditional investment casting technique are eliminated with RP techniques. The technique also eliminates the need for wax and core tooling design and manufacturing, wax and core molding, wax assembly, shell dipping and drying, and wax elimination [ 4 - 6 ]. Facial Prosthesis Mold Over the last decade, RP techniques have been used successfully to fabricate facial prostheses. Although pattern fabrication with the aid of RP was a feasible procedure, the traditional flasking and investing procedures were still required to make the actual prosthesis. Using a mold would eliminate the need for traditional flasking and investment procedures, as well as shorten the process of creating the prosthesis. Furthermore, the generated resin mold can be kept because it is long-lasting and allows for multiple pourings [ 7 ]. Combining the digital and conventional techniques, a hybrid protocol for the fabrication of maxillofacial prosthesis has been described in Figure 3. A comparative evaluation of different techniques used for maxillofacial prostheses fabrication is done in Table 4. Figure 3 Hybrid Protocol for the Workflow of Maxillofacial Prostheses Combining the conventional and digital techniques for the fabrication of maxillofacial prostheses, a hybrid protocol has been formulated Table 4 Comparative Evaluation of Different Techniques used for Maxillofacial Prostheses Fabrication. Workflow Clinical efficacy Time Cost-effectiveness Edge quality and marginal adaptation Aesthetic outcomes Material characteristics CONVENTIONAL FABRICATION Manual Impression making and multiple try-ins Several complex steps, labor-intensive Time-consuming Cheaper, compared to digital technique Good The Patient relies on the skills of the Prosthodontist Medical grade silicone HYBRID 3D capture of facial topography Excellent; contactless Semi-automated Less time-consuming compared to Conventional fabrication Cuts off the additional digital fabrication costs Acceptable Acceptable Medical grade silicone DIGITAL FABRICATION 3D capture of facial topography Excellent; sometimes challenging and prone to errors Minimal time required Expensive Comparatively low Acceptable No material is clinically approved for direct fabrication Conventional Workflow: Choosing an appropriate impression technique and material (irreversible hydrocolloids or elastic silicones are the most commonly utilized impression materials) based on the type of defect, the size, and the presence of undercuts in the affected part, and a custom tray is important. To retrieve the impression without causing any damage to tissue in the surrounding, some anatomic undercuts are blocked. The gypsum cast is obtained when the impression is poured, and a wax pattern of the anatomic portion to be replaced is made up. The wax is carved to reproduce the defect's natural morphological details, followed by a try-in step of the prosthesis wax-up with the equivalent maxillofacial prosthesis [ 16 ]. Digital Workflow: The final prostheses are created using rapid prototyping, specifically additive manufacturing. Maxillofacial prostheses are fabricated indirectly by procurement of a mold or model of the prosthesis, followed by the traditional workflow for part processing, or directly manufacturing with the help of 3D printing with adequate material, depending on the anticipated digital workflow and the material being used (e. g. , acrylic resins, silicone-based elastomers, and others) [ 17 ]. Recent Advances 3D Bioprinting, a combination of 3D printing and tissue engineering is a rapidly expanding technology in the field of regenerative medicine for autograft production. Biomaterials, bioactive substances, and even cells that are carefully positioned and with spatial control can be 3D printed to reconstruct human tissues and organs that can imitate their native counterparts in terms of both shape and function. This process is known as 3D bioprinting [ 5, 7 ]. It's the result of combining 3D printing with tissue engineering. Tissue engineering is a field of regenerative medicine that tries to construct an autologous graft using the patient's own cells. Additive manufacturing technology, such as 3D printing, is now frequently used to improve the aesthetics of maxillofacial prostheses with precise 3D fabrication. It uses CAD software to create complicated facial shapes, which is then followed by layer-by-layer material deposition to create 3D objects. It can make not just complex craniofacial analogs but also can manufacture prototypes for osteotomy guides, bone grafts, and occlusal splints to be used intraoperatively, which increases efficiency and makes surgery easier. However, creating indistinguishable maxillofacial prostheses continues to be a challenge [ 7, 12 ]. Conclusions It is clear that 3D RP is an important tool for creating maxillofacial prostheses and 3D bioprinting is a boon in creating complex tissues and organs, such as muscular tissue, and for using biomaterials to manufacture and develop the extra segments. RP techniques are currently playing a larger role and will play an important role in prosthodontics the dominant digital fabrication technologies. There are, however, important problems in the process of innovation that must be addressed. The majority of currently available tissue products closely resemble genuine tissues and have emerged with a focus on tissue removal and multicellular systems. The first aspects of making biopolymer manufacture and obtaining standards should be in sync with the creative cycle. Learning and utilizing the 3D printing technique and achieving the standards requires immediate assistance. In the forthcoming era, clinicians should assume that 3D printing technology will have applications in a wide array of dentistry fields, especially maxillofacial rehabilitation.

10. 7759/cureus. 29248

Cureus

Injectable and Self-Invigorating Hydrogel Applications in Dentistry and Periodontal Regeneration: A Literature Review

Hydrogels are thought of as unique polymers utilized to build new materials, and two key factors that impact their features are their hydrophilicity and the degree of cross-linking of the polymer chains. An injectable hydrogel is based on the hypothesis that certain biomaterials can be injected into the body as a liquid and progressively solidify there. The scientific research community was intrigued and interested by its discovery. The hydrophilic polymers that are used to make hydrogels can typically be split into two groups: natural polymers derived from tissues or other sources of natural materials, and synthetic polymers produced by combining principles from organic chemistry and molecular engineering. A variety of organic and synthetic biomaterials, such as chitosan, collagen or gelatin, alginate, hyaluronic acid, heparin, chondroitin sulfate, polyethylene glycol, and polyvinyl alcohol, are used to generate injectable hydrogels. A promising biomaterial for the therapeutic injection of cells and bioactive chemicals for tissue regeneration in both dentistry and medicine, injectable hydrogels have recently attracted attention. Since injectable scaffolds can be implanted with less invasive surgery, their application is seen as a viable strategy in the regeneration of craniofacial tissue. Treatment for periodontitis that effectively promotes periodontal regeneration involves injecting a hydrogel that contains medications with simultaneous anti-inflammatory and tissue-regenerating capabilities. The advantages of injectable hydrogel for tissue engineering are enhanced by the capability of three-dimensional encapsulation. A material's injectability can be attributed to a variety of mechanisms. The hydrogels work well to reduce inflammation and promote periodontal tissue regeneration.

Introduction and background Due to the chemical or physical cross-linking of individual polymer chains, a hydrogel is a three-dimensional (3D) network of hydrophilic polymers that can swell in water and store a lot of water while keeping the structure. Whichterler and Lim made the first discovery of hydrogels in 1960 [ 1 ]. The degree of polymer chain cross-linking and hydrophilicity are two key elements that affect the properties of hydrogels. The ability of hydrogels to retain water is due to the presence of functional groups such as hydroxylic (-OH), carboxylic (-COOH), amidic (-CONH-), primary amidic (-CONH 2 ), and sulphonic (-SO 3 H) groups within the polymer network [ 2 ]. The hydrogel systems have been proposed as potential carriers or scaffolding for pharmaceuticals. They have also been thoroughly investigated for a number of biomedical applications due to their benefits, which include biocompatibility, permeability to oxygen and nutrients, physical qualities similar to those of the original extracellular matrix (ECM), and programmable physical and mechanical properties [ 3 ]. An injectable hydrogel is based on the concept that some biomaterials can be injected into the body as a liquid and then solidify in place. Injectable hydrogels have generated a lot of attention in the areas of medication delivery, tissue engineering, and dermal fillers because they have the necessary physicochemical qualities to be injected in situ into the body [ 4 ]. One especially intriguing family of self-healing hydrogels is one that may be injected or printed (in the context of 3D printing). As schematically shown in Figure 1, these self-healing injectable hydrogels are capable of momentarily fluidizing under shear stress and then regaining their original structure and mechanical characteristics after releasing the applied tension [ 5 ]. Figure 1 Shows a self-healing injectable hydrogel's behavior schematically At the first place, at rest, the material exhibits gel-like properties; in the second place, fluidization under shear as a result; and finally, self-healing of the original structure and mechanical characteristics after flow. Source: Ref. [ 5 ]. Review History of hydrogels The first hydrogel poly-2-hydroxyethyl methacrylate (PHEMA) was created and described in 1960 by Whichterler and Lim. They used it to make moisture-absorbing contact lenses. Modern hydrogels are demonstrated and resembled by their 3D crosslinking structure. The scientific research community was intrigued and interested by its discovery. Then, in the 1980s, Lim and Sun created calcium-alginate gel composites for islet-droplet microcapsule cell embedding [ 6 ]. According to Buwalda et al. , there have been three distinct generations of hydrogels. The first generation of hydrogels mostly consisted of gels with diverse crosslinking techniques created by chemically altering a monomer or polymer using an initiator. After this time, in the 1970s, the significance of hydrogels increased to a new level as stimuli-responsive properties were incorporated into the hydrogels, allowing second-generation hydrogels to react to a variety of highly specific stimuli, including changes in pH, temperature, or the concentration of certain biomolecules in a solution. The focus switched to the creation of stereo-complexed biomaterials and hydrogels joined through physical interactions in the third-generation hydrogels. These changes prompted scientists to focus their efforts more intently on creating the current "smart hydrogels, " which can be tailored to acquire specific qualities like stimulus responsiveness and adjustable mechanical and other physicochemical properties [ 7 ]. Building blocks for the preparation of injectable hydrogels A variety of materials are used to create injectable hydrogels. Generally speaking, there are two types of hydrophilic polymers that are utilized to make hydrogels: natural polymers taken from tissues or other natural sources, and synthetic polymers created utilizing organic chemistry and molecular engineering concepts. Building blocks of synthetic and biocompatible natural polymers are used to create the injectable hydrogels shown in Table 1 [ 7 ]. Table 1 Shows biocompatible natural polymer and synthetic polymer building blocks for the preparation of injectable hydrogels Source: Ref. [ 7 ]. Natural sources Synthetic polymers By using covalent or physical crosslinking (such as ionic or hydrogen bonding) to create injectable hydrogels, these organic polymers have been used as building blocks (e. g. , reaction of functional groups on modified polymers). Synthetic polymers are employed in conjunction with natural polymers or biomimetic peptides to promote cell adhesion, migration, and protein release because they lack the intrinsic biochemical cues for contact with cells. Hyaluronic acid Polyethylene glycol (PEG) Chitosan Polyvinyl alcohol (PVA) Heparin Poly N-isopropylacrylamide (PNIPAAm) Alginate Polycaprolactone (PCL) Fibrin Collagen Chondroitin sulfate Silk Types of injectable hydrogels Natural and synthetic biomaterials, such as chitosan, collagen or gelatin, alginate, hyaluronic acid, heparin, chondroitin sulfate, polyethylene glycol (PEG), and polyvinyl alcohol (PVA), are used to form injectable hydrogels. Chemical techniques that use covalent crosslinking produced by enzymes, physical techniques that use weak secondary forces, chemical techniques that use photo-cross-linking, physical techniques that use Michael addition, and chemical techniques that use click chemistry can all be used to create injectable hydrogels that are ion-sensitive, pH-sensitive, or temperature-sensitive [ 8 ]. Chitosan-Based Hydrogels Chitosan is a cationic polymer consisting of glucosamine and N-acetylglucosamine that is found naturally. Because of its biocompatibility and biodegradability, chitosan is frequently used in the pharmaceutical and medical fields. Because of the many amino groups on its backbone, chitosan is a great choice for creating injectable self-healing hydrogels based on imine linkages [ 9 ]. Phosphate-Based Hydrogels A fresh, injectable oligo-polyethylene-glycol fumarate (OPF) gel for bone healing with continuous release of phosphate ions, improved electrical conductivity, and mechanical strength. The mechanical stability and electrical conductivity of gel formulations were shown to be improved by a carbon nanotube nanocomposite, and the inclusion of two-dimensional (2D) black phosphorus nanosheets allowed a continuous release of phosphate ions by environmental oxidation [ 10 ]. Alginate-Based Hydrogels Alginate hydrogels with biomedical uses fall into two categories: "Physical" or "reversible" gels are held together by molecular entanglements, hydrophobic forces, and ionic or hydrogen bonding, as opposed to "chemical" or "permanent" gels, which are produced when stable covalent connections crosslink networks. Alginate hydrogels are excellent alternatives for drug carriers because of their high water content, nontoxicity, soft consistency, biocompatibility, and biodegradability. They can transport low-molecular-weight drugs as well as macromolecules like proteins and DNA either sustainably or locally [ 11 ]. Hyaluronic Acid (HA)-Based Hydrogels Because it may imitate tissue ECM and has the ability to influence cell behavior during tissue regeneration, such as cartilage and pulp regeneration, HA has been widely used to make injectable hydrogels. Since HA's molecular structure contains a variety of primary and secondary hydroxyl and carboxyl groups, a bifunctional small-molecule crosslinking agent can also be utilized for this purpose. This results in a HA hydrogel with the optimum composition, shape, hardness, and biological activity [ 12 ]. Collagen (col)- or Gelatin (gel)-Based Hydrogels Due to their high levels of biocompatibility, biodegradability, bioactivity, and diversity, gelatin methacryloyl (GelMA) hydrogels are frequently employed for tissue healing. By adding double bonds to the gelatin polymer chains, which under photoinitiation quickly form hydrogels, GelMA hydrogel is created. Lithium acylphosphinate salt (LAP), a blue light initiator, facilitates preparation and speeds up the gelation process. It poses no threat. Injectable hydrogel made from GelMA hydrogel works well and may be molded using 3D printing [ 13 ]. Fibrin-Based Hydrogels Known as a helpful cell-transplantation matrix, they can improve cell adhesion, proliferation, differentiation, and migration in a 3D scaffold. Fibrin is a naturally occurring fibrous protein implicated in blood clotting. Scaffolds have been constructed using fibrin, either by itself or in conjunction with other substances, for applications involving cartilage tissue engineering. A unique injectable hydrogel system has been created using PEG, stem cells generated from human amniotic fluid, and fibrin-based hydrogels that can promote in situ neovascularization and cause a fibrin-driven angiogenic host response [ 14 ]. Elastin-Based Hydrogels For numerous biological applications, injectable hydrogels produced from ECM proteins like elastin hold considerable promise. The primary problems of these hydrogels are their lack of mechanical strength, use of cytotoxic chemicals, and fixed gelling behavior [ 15 ]. Chondroitin Sulfate (CS)-Based Hydrogels The safety and high biocompatibility of CS, a common material for cartilage tissue engineering scaffolds that exhibits quick gelation, outstanding mechanical capabilities, and delayed degradation qualities, have been thoroughly explored and documented [ 16 ]. PEG-Based Hydrogels PEG-based hydrogels and their derivatives have received a lot of attention recently due to their capacity to be well tolerated in vivo in the context of drug administration and tissue engineering applications. Because they may be administered easily and painlessly by injecting low-viscosity precursor polymer solutions, injectable, in situ gelling counterparts improve the potential applications of these hydrogels [ 17 ]. PVA-Based Hydrogels For the in situ production of hydrogels under physiological conditions, new PVA compounds with a variety of pendant chemoselective characteristics have been developed. For the first time, PVA was modified by adding thiol, cysteine 1, 2-aminothiol, and aminooxy side chains by direct carbamate connections from protected nucleophilic functionalities to the hydroxyl groups of PVA [ 18 ]. Shape Memory (SM) and Self-Healing (SH) Hydrogels Both SM and SH have their roots in reversible interactions. It is quite difficult to synthesize hydrogels using both SH and SM [ 19 ]. Interpenetrating Polymer Network (IPN) Hydrogels To address the conundrum of choosing between a complex structure and an easy biodegradability, a novel method for creating biodegradable hydrogels with an IPN structure that consists of peptide self-assembling networks and a covalently cross-linked network has been proposed [ 20 ]. Double Network (DN) Hydrogels Due to their outstanding mechanical and chemical adaptability, DN hydrogels have become leading contenders for tissue engineering. DN hydrogel formulations combined with processing advancements (such as additive manufacturing and injection) have produced a remarkable set of findings that significantly advance the development of systems that can address the complex environment around tissues and allow for individualized fabrication techniques [ 21 ]. Programmable Hydrogels Programmable hydrogels are hydrogels that have the ability to periodically, reversibly, and sequentially alter their properties and functionalities. Using the aforementioned concepts, programmable hydrogels that are induced to undergo functional changes could be produced for a range of intriguing applications [ 22 ]. 3D Printed Hydrogels In tissue engineering and regenerative medicine, 3D bioprinting, one of the most recent biotechnologies, is frequently used to produce complex artificial organ and tissue designs that closely resemble genuine organs and tissues. In order to restore functional and site-specific tissues or organs, bioprinting is the additive deposition of cell-loaded hydrogels in a specified structural framework [ 23 ]. Injectability mechanisms of injectable hydrogels A material's injectability can be attributed to a variety of mechanisms. These mechanisms are divided into three major categories: The first mechanism is in situ gelling liquids - solutions or liquids that normally flow but harden into gels when injected into the body. The second mechanism is injectable gels - even though certain gels are created ex vivo, their shear thinning qualities and capacity to restore their hydrogel shape following relaxing procedures make them suitable for injection. The third mechanism is injectable particles - in addition to the two types of injectable systems previously discussed, the third class of injectable particles also has the ability to be injected when immersed in a liquid phase. Depending on the desired outcomes, particle sizes might be nano-, micro-, or macroscale [ 24 ]. Applications of injectable hydrogels Therapeutic Applications in Dentistry A promising biomaterial for the therapeutic administration of cells and bioactive compounds for tissue regeneration in dentistry and medicine is injectable hydrogels. They are suitable for minimally invasive surgical operations in a clinical context because they offer adaptable tissue-like qualities, regulated degradation and release behavior, and the ability to conform to the 3D defect upon gelling. A recent development in tissue engineering used a biomaterial scaffold and bioactive substances to encourage endogenous cell migration and tissue repair, further demonstrating the potential of the "homing" strategy for dentin-pulp and craniofacial regeneration [ 25 ]. Injectable biomaterials are regarded as great candidates for pulp and dentin regeneration because of the tooth root canal's small size and uneven shape. The first organic biomaterial employed as an injectable gel for pulp regeneration was collagen. As a dental stem cell transporter, fibrin was treated with PEG to create a poly-ethylene-Gylated fibrin hydrogel with a slower rate of breakdown [ 26 ]. By modifying its physical and chemical characteristics, the hydrogel's shelf life may be increased to three years, and it can be successfully utilized as the only maxillofacial material. As a denture adhesive, hydrogels have a number of advantageous characteristics [ 27 ]. Since injectable scaffolds can be inserted with less invasive surgery, lowering the risk of surgical problems and enhancing postoperative recovery, their usage is seen as a potential strategy for the regeneration of craniofacial tissue [ 28 ]. Laden HA injectable hydrogels as a biomaterial for the encapsulation of human dental pulp cells demonstrate significant clinical potential for endodontic regeneration therapy. In order to regenerate bone tissue, CS-based hydrogels can greatly increase cell proliferation and cell adhesion [ 29 ]. Therapeutic Applications in Periodontal Regeneration Treatment for periodontitis that effectively promotes periodontal regeneration involves injecting a hydrogel that contains medications with simultaneous anti-inflammatory and tissue-regenerating capabilities. The hydrogels showed outstanding self-healing abilities, an expedient gelation process, and injectability [ 30 ]. To reduce inflammation and encourage tissue regeneration, local medication delivery has been used as a successful method. The CS-based delivery approach can be depended upon to deliver the encapsulated active drugs to the disease site within periodontal pockets by utilizing injectable chitosan hydrogels with modulable physico-chemical characteristics [ 29 ]. Emdogain is based on a derivative of porcine enamel matrix, a combination of proteins supplied in an aqueous gel solution made of propylene glycol alginate and including amelogenin (90%) along with a few other nanomelogenin such as ameloblastin, enamelin, and tuftelin. Emdogain has been shown to regenerate a variety of periodontal tissues, including connective tissues like the periodontal ligament as well as osseo-like tissues, acellular cementum, and alveolar bone [ 31 ]. Aspirin/erythropoietin was used to fill the successfully created and proven to be helpful in periodontium regeneration CS/gelatin hydrogel [ 32 ]. The dual drug-loaded oxidized dextran and phenylboronic acid-functionalized polyethylene imine hydrogel, a novel potential therapeutic agent, may be beneficial for the therapy of chronic periodontitis with diabetes mellitus [ 33 ]. Innovative hybrid hydrogel offers enormous potential as an injectable platform technology with a variety of applications in the eradication of mouth infections such as periodontal disease and pulpal pathology [ 34 ]. Injectable Hydrogel Delivery for Tissue Engineering The use of injectable hydrogels as a delivery system for topical and localized medication delivery appears promising. The advantages of injectable hydrogel for tissue engineering are enhanced by the capability of 3D encapsulation. An injectable hydrogel must strike a compromise between mechanical properties, carrier capacity, and processability [ 35 ]. It has been demonstrated that among the several scaffolds for bone tissue engineering applications, hydrogels are attractive templates for bone regeneration due to their similarity to the natural ECM. Chitosan, a natural biopolymer, has attracted a lot of attention since it can be utilized to produce thermo/pH-responsive injectable hydrogels [ 36 ]. Future directions and current status Due to their adaptability, injectable hydrogels, a subset of hydrogel, have drawn significant interest in biological applications. According to reports, the injectable hydrogel can be used in a variety of biomedical treatments, such as tissue engineering for cartilage and bone as well as periodontal implants and submucosal fluid cushions. In addition to being simple to implant, this kind of hydrogel can be customized to fit certain purposes [ 37 ]. Conclusions Due to its potential for less invasive local drug administration, more precise implantation, and site-specific drug delivery into difficult-to-reach tissue regions and into interface tissues, where wound healing takes time, injectable hydrogels have attracted a lot of interest in the biomedical industry. It is still extremely desirable to create an injectable hydrogel with self-healing capabilities for continuous, controlled drug delivery. For prolonged protein release, targeted drug delivery, and tissue engineering, hydrogels, microgels, and nanogels have become efficient and useful platforms because of their high biocompatibility and microporous structure with adjustable porosity in periodontal regeneration. Among other things, the hydrogels are successful in reducing inflammation and promoting periodontal regeneration.

10. 7759/cureus. 29253

Cureus

Comparison of the Efficacy of Platelet-Rich Plasma (PRP) and Local Corticosteroid Injection in Periarthritis Shoulder: A Prospective, Randomized, Open, Blinded End-Point (PROBE) Study

Background Periarthritis or frozen shoulder, also called adhesive capsulitis, is characterized by stiffness and pain along with gradual loss of active and passive movement in the glenohumeral joint. More than 2-5% of the population suffers from periarthritis with a higher incidence in the age group of 40-60 years. The various treatment modalities used for its management include simple physiotherapy, short-wave therapy, ultrasonic therapy, transcutaneous electrical nerve stimulation, hydrotherapy, analgesics, intra-articular injections, manipulation under general anesthesia (MUA), and surgical management. The application of intra-articular steroid injection has been a common and efficacious option in rapidly diminishing shoulder pain and disability. Some recent studies reported a better outcome using platelet-rich plasma (PRP) injections in frozen shoulder cases. Hence, this randomized controlled trial was conducted to compare the efficacy of intra-articular injections of PRP and triamcinolone in patients of shoulder periarthritis in a population from the eastern region of India Methodology A total of 60 patients with periarthritis shoulder were allocated into two groups after randomization. Group A received 2 mL autologous PRP, and Group B received 2 mL of triamcinolone (40 mg/mL) intra-articular injection. Patients were followed up on the 4th week, 12th week, and 24th week. The assessment of pain and function using the visual analog scale (VAS) score and the Disabilities of Arm, Shoulder, and Hand (DASH) score, respectively, was done at each follow-up. The primary analyses of both primary and secondary outcomes were conducted in the intention-to-treat (ITT) population. SPSS version 24 (IBM Corp. , Armonk, NY, USA) was used for data analysis. Results The mean VAS score in the PRP and triamcinolone groups was 14. 33 ± 3. 79 and 31. 63 ± 7. 62, respectively (p = 0. 0001) after 24 weeks. The mean DASH score in the PRP and triamcinolone groups was 18. 08 ± 8. 08 and 31. 76 ± 3. 63, respectively (p = 0. 0001), which shows significant improvement in both pain and disability scores in the PRP group after 24 weeks. Conclusions The triamcinolone group showed better short-term outcomes whereas PRP showed better long-term outcomes in reducing pain and disability scores in terms of VAS and DASH scores.

Introduction Periarthritis of the shoulder is characterized by functional loss of passive and active shoulder motion. This condition was termed by Duplay in 1896 and later substituted by the term frozen shoulder by Codman in 1932. Subsequently, Nevaiser introduced the term adhesive capsulitis [ 1 ]. This disorder is defined by the American Shoulder and Elbow Surgeons as a condition of significant restriction of both active and passive motion of the shoulder joints because of an unknown etiology that occurs without an intrinsic shoulder disorder [ 2 ]. The definite pathophysiology of periarthritis remains unclear. The progressive fibrosis causing the contracture of the glenohumeral joint capsule results in pain and stiffness [ 3 ]. Periarthritis can be primary or secondary. The primary (or idiopathic) type occurs without any known trauma or provoking event. The secondary type is often observed after periarticular trauma, fracture, or dislocation of the glenohumeral joint [ 4 ]. According to recent studies, the incidence of periarthritis is 2-5% in the general population [ 5, 6 ]. The affected population includes 70% females. The idiopathic type often involves the non-dominant extremity, while 40-50% of cases have been reported as bilateral involvement. Regardless of the etiology, the condition is more prevalent in the 40-60-year age group [ 4, 7 ]. The risk factors for developing periarthritis include diabetes. Patients with type I diabetes have a 40% chance of developing periarthritis. Up to 29% of individuals with type II diabetes may develop this condition. Thyroid disease, Parkinson’s disease, cardiac disease, autoimmune disease, chronic obstructive pulmonary disorder, and myocardial infarction are also linked with increased incidence of periarthritis or adhesive capsulitis [ 3, 8 ]. In most cases, periarthritis resolves spontaneously or it can last for up to three years [ 9 ]. Various treatment approaches have been used and explored to treat this disorder. Physical therapy individually or in combination with short-wave therapy, ultrasonic therapy, transcutaneous electrical nerve stimulation, and hydrotherapy is used [ 10 ]. Pharmacological treatment includes the use of analgesic or non-steroidal anti-inflammatory drugs, oral or intra-articular use of corticosteroids, and sodium hyaluronate injections. Other approaches to treat periarthritis include manipulation under anesthesia (MUA), dilation or distension of the capsule, and arthroscopic or open capsular release (arthroscopic capsulotomy) [ 3, 4, 11 ]. Platelet-rich plasma (PRP) is an emerging entity in the field of tissue engineering and regenerative medicine due to its availability, affordability, and minimally invasive procedure. Its autologous nature prevents an immunological reaction and offers good therapeutic safety. Recently, evidence in immune-mediated disorders and inflammatory processes has garnered attention due to their anti-inflammatory effects through the inhibition of nuclear factor kappa B signaling in target cells and by tissue inhibitor of matrix metalloproteinase. The creation and remodeling of the extracellular matrix also encompass a function of platelet growth factors which further supports this treatment modality [ 12 ]. The application of intra-articular steroid injection has been a common and efficacious option in rapidly diminishing shoulder pain and disability [ 5 ]. Some recent studies show a better outcome using PRP injections in frozen shoulder cases [ 13 ]. A systemic review and meta-analysis by Sun et al. described that patients taking a single steroid injection for a frozen shoulder is effective and safe and improves functional outcomes and pain scores [ 14 ]. Corticosteroid injections have been associated with prominent side effects, which have led to the conception of modalities such as PRP. This randomized trial aimed to evaluate and compare the efficacy of intra-articular injections of PRP and steroid (triamcinolone) in periarthritis. We hypothesized that PRP would prove more effective in relieving pain and improving function. Several studies have reported comparative analyses of steroids and PRP. Most of these were conducted outside India. Studies by Upadhyay et al. [ 15 ], Kothari et al. [ 16 ], and Kumar et al. [ 17 ] reported the effect of PRP versus steroids in periarthritis in the Indian population. One study from the eastern part of India with a similar intervention was conducted by Barman et al. [ 18 ], but the follow-up period was only 12 weeks. Hence, this study was conducted to analyze the comparative efficacy of PRP versus steroids in periarthritis with a follow-up duration of 24 weeks in a population from the eastern region of India. Materials and methods Trial design This study was a parallel-group, prospective, randomized, open, blinded end-point (PROBE), single-center clinical study. Randomization was done in permuted blocks of varying sizes (2, 4, 6) using a sealed envelope website (computer-generated) [ 19 ]. There was central randomization, and the person doing randomization was not part of the study. The investigator assigning intervention telephonically contacted the randomizer on the recruitment of every new patient regarding the group to which the patient was assigned. Another investigator (other than the one assigning intervention) assessed the outcomes of the patients without any knowledge of the study group to which the patient belonged to. Patients were recruited to different treatment regimens following proper randomization. Unlike double-blind studies, the treatment regimens were recognizable to both physicians and patients. The trial was conducted according to the principles of the Consolidated Standards of Reporting Trials (CONSORT). Site of the study The study was conducted from December 2020 to December 2021 at the Department of Orthopaedics, Rajendra Institute of Medical Sciences (RIMS), Ranchi Jharkhand, India. Ethical approval was obtained (vide reference number: 123, dated November 23, 2020) from the Institutional Ethical Committee of RIMS, Ranchi. Participants A total of 60 patients from the outpatient department (OPD), Department of Orthopedics, RIMS who were clinically diagnosed to have periarthritis shoulder and willing to participate were randomized into two groups. A written informed consent regarding participation was obtained before recruitment. The complete procedure of the study was explained to all participants in their language by the investigator before recruitment. The inclusion and exclusion criteria are presented in Table 1 and Table 2, respectively. Table 1 Inclusion criteria. Serial number Criteria 1 Patients aged between 30 and 75 years 2 Patients having shoulder pain for at least one month and associated with more than one-third of loss of active shoulder flexion, abduction, and external rotation 3 A normal anteroposterior radiograph of the glenohumeral joint in neutral rotation 4 Willingness to refrain from any other auxiliary treatment modality Table 2 Exclusion criteria. Serial number Criteria 1 Patients with any previous treatment in the form of local injections 2 Suffering from symptoms of shoulder pain due to other reasons 3 Unwillingness to participate in the study 4 Any intrinsic glenohumeral pathology 5 History of shoulder trauma/surgery, and clinical evidence of complex regional pain syndrome 6 History of injection in the involved shoulder joint during the preceding six months 7 Non-steroidal anti-inflammatory drugs intake in the last seven days 8 Patients with hematological disorders or on antiplatelet or anticoagulant therapy 9 Patients with thyroid disorders, pulmonary disorders particularly emphysema and chronic bronchitis, neoplastic disorders 10 Pregnant or breastfeeding females Sample size The sample size was calculated by OpenEpi, Version 3, an open-source calculator based on the findings of the study by Kothari et al. , in which the mean VAS score for PRP and steroid group were reported [ 16 ]. The calculated sample size was 29 for each group (Table 3 ). Rounding off to the nearest, the total sample size was finally set as 60 (30 per group). Table 3 Sample size calculation. *Difference between the means. Group 1 Group 2 Difference* Mean 1. 9 3. 4 -1. 5 Standard deviation 1. 8 2. 2 Variance 3. 24 4. 84 Sample size of Group 1 29 Sample size of Group 2 29 Total sample size 58 Confidence interval (two-sided) 95% Power 80 Ratio of sample size (Group 2/Group 1) 1 Procedure All information about the history, clinical features, examination findings, and treatment (if any were taken before) were recorded in a predesigned proforma. All patients were subjected to routine blood investigation and radiographic examinations of the cervical spine and ipsilateral shoulder under study. Before administrating the injection, povidone-iodine and ethyl alcohol were applied to the skin. One milliliter of 2% lignocaine with adrenaline was injected at the injection site after administering the test dose. After 10 minutes, the proposed injection was injected. If any resistance was felt during the injection, the needle was withdrawn slightly and again injected. The first group of patients was administered 2 mL of triamcinolone (40 mg/mL). The second group was given 2 mL autologous PRP. To prepare PRP, about 15 mL of the patient’s blood was drawn through a scalp vein catheter. The PRP was prepared using a differential centrifugation technique with two spins. The blood was collected in three citrate tubes having 0. 9% sodium citrate as an anticoagulant. The first spin was performed at 1, 500 rpm for 15 minutes using a laboratory centrifuge. This spin separated the red blood cells from the rest of the components. The upper half of the supernatant was discarded. The lower halves of the supernatant from all three tubes were transferred into another plain tube for the second spin. The second spin was performed at 2, 500 rpm for 10 minutes. The upper half of the supernatant was discarded. Three milliliters of the lower half was taken into a syringe having 0. 1 mL of calcium chloride. At the end of the preparation of PRP, 1 mL of obtained PRP (as a sample) was sent for platelet count, and the count was compared with the patient’s platelet count. Another 2 mL was used for intra-articular injection. The platelet count in the PRP preparation was 860, 000 ± 74, 500 platelets per mm 3 which were 4. 2 ± 1. 37 times higher than whole blood values. In our study, we injected freshly prepared PRP (within 30 minutes of preparation), as a study by Blajchman [ 20 ] reported that platelets may alter the shape and reduce the functional properties, including the degranulation of α-granules due to prolonged storage. All patients were advised regarding post-injection care. The possibility of pain increasing during the initial two weeks was explained to the patient. Post-injection, patients were prescribed paracetamol (650 mg BD orally for five days) for pain relief in both groups. Patients were advised to rest during the initial two weeks and avoid strenuous activities by the extremity under study after the injection. Physiotherapy was advised for both groups. Bilateral cases were injected simultaneously, and the post-injection protocol was the same. Assessment and follow-up After inclusion in the study, demographic data, baseline clinical findings, duration of pain, dominancy of the affected side, and associated comorbidities were recorded. Any relevant X-ray findings were noted. Special investigations were performed as per comorbidity present in a case. The follow-ups were done in the 4th week, 12th week, and 24th week for all patients of both groups. The assessment of pain and function through the VAS and the Disabilities of Arm, Shoulder, and Hand (DASH) score, respectively, was done at each follow-up. Any adverse effects were noted and reported. All data were documented in case report form (CRF) designed for the project and in Excel sheets for analysis. Outcome measures The primary outcome of the study was the pain reduction assessed using the VAS after the injections. The DASH scores were assessed as a secondary outcome. Statistical analysis The primary analyses of both primary and secondary outcomes were conducted in the intention-to-treat (ITT) population (i. e. , all randomized participants for whom consent was given to use data). SPSS version 24 (IBM Corp. , Armonk, NY, USA) was used for data analysis. The data with categorical variables were expressed as numbers and percentages, while the continuous variables were expressed as the mean ± standard deviation (SD). An unpaired t-test was used for analyzing continuous variables in the intergroup analysis. The Fisher’s exact test and Pearson’s chi-square test were used for analyzing categorical variables. P-values of <0. 05 were considered to be significant. Results A total of 60 patients were recruited for the study and randomized equally into two groups. One patient from the PRP group and two patients from the triamcinolone group did not come for the last follow-up (24 weeks). However, analyses were done for a total of 60 patients as per the ITT analysis protocol (See Figure 1 ). Figure 1 CONSORT diagram. n: number of patients; ITT: intention to treat; NSAIDs: non-steroidal anti-inflammatory drugs The demographic data presented in Table 4 reveals that both groups were similar in characteristics. There was no significant difference between both groups in the baseline characteristics, e. g. , age, gender, the dominance of the affected side, duration of symptoms, and presence of diabetes mellitus. This revealed that patient variability was not present between both groups. Moreover, the inclusion and exclusion criteria were followed strictly during patient recruitment and randomization. Therefore, the possibility of patient variability in the study groups was negligible. Table 4 Clinicodemographic characteristics. #: Unpaired t-test was used for intragroup analysis; a: Fisher’s exact test/Pearson’s chi-square were used. PRP: platelet-rich plasma; SD: standard deviation Variables Triamcinolone (n = 30) PRP (n = 30) P-value Age (mean ± SD) 46. 70 ± 7. 13 47. 8 ± 9. 56 0. 615 # Sex (n) Male 13 12 0. 793 a Female 17 18 Involved side (n) Dominant 12 10 0. 592 a Non-dominant 18 20 Duration of symptoms in months (mean ± SD) 3. 217 ± 0. 887 3. 567 ± 1. 015 0. 160 # History of diabetes mellitus (n) Present 14 13 0. 7952 a Absent 16 17 The patients with frozen shoulders were aged from 33 to 67 years. The incidence of the disease was higher in the fifth decade of life (46. 67%). The mean age of the patients was 47. 25 ± 8. 38 years (in triamcinolone and PRP treatment groups). The incidence of the disease was higher in females (58. 33%) compared to males (41. 67%). In the triamcinolone group, there were 56. 67% females, while in the PRP group, there were 60% females. Among 60 patients, 30 received prolotherapy (PRP injection) and 30 received triamcinolone injection for frozen shoulder. Table 5 represents the outcome analysis of both groups. In the first follow-up (four weeks), the mean VAS score in the triamcinolone group was 46. 27 ± 8. 17 while it was in 51. 70 ± 6. 02 in the PRP group. This significantly shows better improvement of pain with triamcinolone injection (p = 0. 0048). Table 5 Outcome Assessment. #: p-value derived from unpaired t-test for intragroup analysis; *: statistically significant. PRP: platelet-rich plasma; VAS: visual analog scale; DASH: Disabilities of Arm, Shoulder, and Hand; SD: standard deviation Triamcinolone (n = 30) PRP (n = 30) Mean difference (95% CI) P-value # VAS score (mean ± SD) Baseline 69. 63 ± 6. 46 67. 40 ± 4. 87 2. 23 (-0. 73, 5. 18) 0. 136 4 th week 46. 27 ± 8. 17 51. 70 ± 6. 02 -5. 43 (1. 72, 9. 14) 0. 0048* 12 th week 31. 83 ± 10. 31 43. 23 ± 4. 01 -11. 40 (7. 36, 15. 44) 0. 0001* 24 th week 31. 63 ± 7. 62 14. 33 ± 3. 79 17. 30 (-20. 41, -14. 19) 0. 0001* DASH score (mean ± SD) Baseline 75. 36 ± 6. 49 77. 63 ± 7. 18 -2. 27 (-1. 26, 5. 81) 0. 2040 4 th week 42. 40 ± 5. 58 45. 03 ± 5. 45 -2. 63 (-0. 22, 5. 48) 0. 0699 12 th week 36. 50 ± 4. 86 34. 36 ± 4. 27 2. 14 (-4. 504, 0. 224) 0. 0752 24 th week 31. 76 ± 3. 63 18. 08 ± 8. 08 13. 70 (-16. 93, 10. 46) 0. 0001* In the second follow-up (12 weeks), the mean VAS score in the PRP group was 43. 23 ± 4. 01 while it was 31. 83 ± 10. 31 in the triamcinolone group. This significantly showed better improvement of pain with triamcinolone injection (p = 0. 0001) after 12 weeks. However, in the third follow-up (24 weeks), the mean VAS score in the PRP and triamcinolone groups was 14. 33 ± 3. 79 and 31. 63 ± 7. 62, respectively, which showed a significantly better improvement in the VAS score in the PRP group (p = 0. 0001). For DASH scores (see Table 5 ), after four weeks of injection, the triamcinolone group shows somewhat better improvement, although there was no significant difference in both groups (p = 0. 069). After 12 weeks of injection, the PRP group showed somewhat better improvement, although no significant difference was found between the groups (p = 0. 075). At the third follow-up (24 weeks), the mean DASH score in the PRP and triamcinolone groups was 18. 08 ± 8. 08 and 31. 76 ± 3. 63, respectively, which showed significant improvement in the DASH score in the PRP group (p = 0. 0001). Discussion Frozen shoulder or shoulder periarthritis is the most common cause of the gradual onset of pain and stiffness with loss of active and passive movement of the glenohumeral joint [ 16 ]. Various treatment modalities are used for the management of periarthritis, e. g. , physiotherapy, intra-articular injections, oral and injectable corticosteroids, MUA, hydrodilation, and surgery [ 1, 21 ]. Triamcinolone is a long-acting steroid with anti-fibrotic and anti-inflammatory properties [ 17 ]. This study compares the effect of intra-articular injections of triamcinolone versus PRP. In this study, 60 patients with shoulder periarthritis with ages ranging from 33 to 67 years were included. The incidence of the disease was higher in the fifth decade of life (46. 67%). The result was similar to previous studies [ 16, 22 ]. The mean age of the patients included in the study was 47. 25 ± 8. 38 years. The prevalence rate of frozen shoulder is expected to be 2-5% of the population, with the peak occurrence in persons aged 40-60 years [ 11, 23 ]. Our study reported a higher incidence (46. 67%) of the disease in the fifth decade of life. Our study reported that periarthritis mostly occurred in female patients than males, which is similar to a previous study [ 7 ]. The side of the joint affected by periarthritis was higher on the non-dominant side. A total of 38 (63. 33%) patients had affected joints by periarthritis on the non-dominant side. Moreover, the majority of the studies showed a higher prevalence rate on the non-dominant side [ 24 ]. About 45% of patients with periarthritis had diabetes mellitus as comorbidity, while 8. 33% of patients had hypertension. In our study, we assessed the VAS and DASH scores at baseline, 4th, 12th, and 24th weeks. We found that the VAS score showed significant improvement in the triamcinolone group (p = 0. 0048 and p = 0. 0001, respectively) than in the PRP group at four and 12 weeks. The DASH score was reduced in both groups in the 4th week (p = 0. 0699) and 12th week (p = 0. 0752), but the improvement was statistically not significant. However, in a study by Barman et al. , there was no significant difference at the end of three weeks after a single dose of PRP injection or steroid injection. However, PRP was found to be more effective than corticosteroid injection at 12 weeks in pain and disability score improvement [ 18 ]. At 24 weeks, both the VAS and DASH scores showed significant improvement in the PRP group to the triamcinolone group (p = 0. 0001). Our result was similar to previous studies by Kothari et al. and Kumar et al. that assessed triamcinolone and PRP [ 16, 17 ]. A case study by Aslani et al. in 2016 also reported good results with PRP in the frozen shoulder [ 25 ]. Evidence of PRP administration in periarthritis is continuously emerging [ 26, 27 ]. In their systematic review, Griesser et al. reported that the use of steroids significantly improved the forward elevation and abduction temporarily, as well as short-term and long-term pain reduction assessed through the Shoulder Pain and Disability Index (SPADI) and VAS scores [ 23 ]. Our study has added support to this growing technique. The study showed that at the 12th week, both the steroid and PRP groups improved the VAS and DASH scores. However, the steroid group had a better outcome in the 12th week, while in the 24th week, the PRP group showed better outcomes. Various studies have reported that the effect of steroids gradually decreases over a long-term follow-up. Blanchard et al. [ 28 ] compared the steroid injections and physiotherapeutic interventions for adhesive capsulitis and reported a good efficacy of corticosteroid injections in the short term (six weeks) and, to a lower magnitude, in the longer term (one year). Another study by Shah and Lewis [ 6 ] found that corticosteroid injections in adhesive capsulitis improved pain and range of motion for 6-16 weeks after the first injection. A systematic review that included 12 randomized controlled trials on using corticosteroids in adhesive capsulitis reported that the intervention was beneficial, although its effect was small and not well maintained [ 29 ]. It has been suggested that the efficacy of corticosteroids on periarthritis is exerted through anti-inflammatory properties and suppressing the granulomatous response in affected tissue which leads to clinical improvement. In contrast, a study reported that PRP releases a pool of several growth factors (transforming growth factor-β, platelet-derived growth factor, vascular and epidermal endothelial growth factor) which helps in tissue repair [ 18 ]. PRP also releases hepatocyte growth factor and tumor necrosis factor-alpha, which possess potent anti-inflammatory effects [ 30 ] In this study, long-term improvements in the PRP group could be explained by the fact that PRP might have effects on improving all phases of tissue repair, e. g. , inflammatory, proliferative, and remodeling phases of capsular healing in periarthritis [ 18 ]. Based on the above discussion, it can be concluded that the effect of steroid injections lasts for a shorter period, while PRP injections might have a longer effect. Strength and limitations In this study, the standardized technique for PRP preparation was used and comparisons were done with the conventionally used treatment. The actual platelet count in obtained PRP was compared to the whole blood or baseline platelet count. All intra-articular injections were administered by a single experienced clinician. Evaluation of pain and disability outcomes was done at several time points over up to 24 weeks for high-quality evidence of the effect of PRP and corticosteroid injections. Despite the carefully designed protocol for the study, there are some limitations to this study. The study did not explore cost analysis. All stages of periarthritis were included in our study; therefore, further studies are needed to compare the effect of these interventions in different stages of periarthritis. This study was conducted on single injections of steroids and PRP as most of the studies on periarthritis were based on single intra-articular injections [ 29 ]. Moreover, this is a standard protocol followed in the institution and approved by the ethical committee. Studies exploring the effect of multiple injections need to be conducted in the future. Conclusions Intra-articular injections of PRP and triamcinolone for periarthritis are effective in reducing pain and disability scores in terms of VAS and DASH scores. The triamcinolone group showed a better effect in short-term outcomes (12th-week analysis) whereas PRP showed better results in long-term outcomes (24th-week analysis). A large sample size study to enhance the power of the study with robust design must be conducted in the future that compares single versus multiple injections as well as both steroid and PRP injections simultaneously.

10. 7759/cureus. 29366

Cureus

The Use of Nanorobotics in the Treatment Therapy of Cancer and Its Future Aspects: A Review

The late Nobel Physicist Richard P. Feynman, in a dinner talk in 1959, very rightly said that there is enough room for the betterment of technology beyond our scope of imagination, proposing utilizing mechanical tools to make those that are relatively smaller than the others, which further can be rendered fruitful in making even more compact mechanical devices, all the way down to the level of the smallest known atom, emphasizing that this is "a progress which I believe cannot be avoided". Feynman proposed that nanomachines, nanorobots, and nanodevices may eventually be utilized to construct a huge range of atomically accurate microscopic instruments and manufacturing equipment, as well as a large number of ultra-small devices and other nanoscale and microscale robotic structures. Biotechnology, molecular biology, and molecular medicine could be used to create totally self-sufficient nanorobots/nanobots. Nanorobotics includes sophisticated submicron devices constructed of nanocomponents that are viewed as a magnificent desired future of health care. It has a promising potential in medication delivery technology for cancer, the top cause of mortality among those under the age of 85 years. Nanorobots might transport and distribute vast volumes of anticancer medications into diseased cells without hurting normal cells, decreasing the adverse effects of existing therapies such as chemotherapy damage. The ultimate development of this innovation, which will be accomplished via a close partnership among specialists in robotics, medicine, and nanotechnology, will have a significant influence on illness detection, therapy, and prophylaxis. This report includes a study on several ways to cancer therapy utilizing nanorobots. Furthermore, it offers insight into the future breadth of this area of research.

Introduction and background Researchers have emphasised nanotechnology as an outstanding technological trend in the last few decades, and it is characterized by the fast proliferation of electronics for applications in communication, known as nanomedicine, and environmental monitoring. Studies are now being conducted on the scientific bottlenecks that affect the lifespan of the living, particularly humans. Among these constraints are illnesses with few or no alternatives for treatment and healing. A drug delivery system (DDS) refers to an alternative diagnosis and/or therapy that has been shown in the medical fraternity [ 1, 2 ]. Nanorobots are nanoelectromechanical systems (NEMS), a recently developed chapter in miniaturisation, similar to microelectromechanical systems (MEMS), which is already a multibillion-dollar business. Designing, architecting, producing, programming, and implementing such biomedical nanotechnology are all part of nanorobotics and NEMS research. Any scale of robotics includes calculations, commands, actuation and propulsion, power, data-sharing, interface, programming, and coordination. There is heavy stress on actuation, which is a key prerequisite for robotics [ 1 ]. The similarity in size of nanorobots to that of organic human cells and organelles brings up a huge variety of its possible uses in the field of health care and environmental monitoring of microorganisms. Other potential uses, such as cell healing, may be possible if nanorobots are tiny enough to reach the cells. Furthermore, it is still to be realised that the tiny sensors and actuators' square measures are necessary for the growing concept of a strongly connected ascending information technology infrastructure; the envision of artificial cells (nanorobots) that patrol the cardiovascular system, thus, detecting and destroying infections in minute quantities. This might be a programmable system with approachable ramifications in medicine, creating a revolutionary replacement from therapy to bar [ 1 ]. Chemotherapeutic substances employed in cancer treatment measure disseminates non-specifically throughout the body, where they exert an influence on both malignant and normal cells, restricting the drug quantity feasible within the growth and also resulting in unsatisfactory medication due to excessive toxic hazards of the chemotherapy drugs on normal cells of the body. It is safe to say that molecularly focused medical care has evolved as a collaborative method to overcome the lack of specificity of traditional cancer therapy drugs [ 3 ]. With the help of nanotechnology, intercellular aggregation of the drugs in cancer cells can be increased while minimising the risk of unwanted drug toxicity in normal cells by utilising various drug targeting mechanisms [ 4 ]. Review This review article focuses on the recent advancements, technological growth, and expansion in the field of nanorobotics and nanotechnology and its application in the discipline of bio-healthcare systems, principally for the DDS in the medication of cancer. Existing research literature and relevant studies regarding the topic of concern were read and a detailed analysis was undertaken in the indexes of PubMed, Science Direct, MEDLINE, Scopus, and Google Scholar. Hardly any language or time constraints were applied. To obtain a detailed search, more articles, synonyms, and derivatives of the phrases were employed; the following evaluation phrases were used: "drug delivery", "cancer", "neoplasms", and "cancer therapy". Nanorobots and their types Nanorobots are miniaturised machines that have the ability to perform work at par with that of current existing machines, having applications in the aspects of medicine, industry, and other areas like the development of nanomotors employed for the conservation of energy; nanorobots have also proved to be serviceable in reducing infertility problems by acting as an engine and giving a boost to the sperm motility when attached to them [ 2 ]. Organic and inorganic nanorobots are by far the most commonly studied. Organic nanorobots, also known as bio-nanorobots, are created by combining virus and bacterium DNA cells. This type of nanorobot is less harmful to the organism. Diamond structures, synthetic proteins, and other materials are used to make inorganic nanobots, which are more hazardous than organic nanobots. To overcome this hurdle of toxicity, researchers have devised a way involving encapsulating the robot, thus decreasing its chances of being destructed by the body's self-defence mechanism [ 5, 6 ]. Scientists can gain an understanding of how to energise micro and nano-sized devices using reactionary processes if they understand the biological motors of live cells [ 7 ]. The Chemistry Institute of the Federal Fluminense University created a nano valve, which is made up of a tank covered with a shutter in which dye molecules are housed and may leave in a uniform fashion whenever the cover is opened. This gadget is also natural, made of silica (SiO2), beta-cyclodextrins, and organo-metallic molecules, and shall be used in therapeutic applications [ 1 ]. Proteins are employed in certain studies to feed nanomotors that can move huge objects, as well as the use of DNA hybridisation and antibody protein in the development of nanorobots. DNA hybridisation is defined as a process by which two complementary single-stranded DNA and/or RNA molecules bond together to form a double-stranded molecule. A nanorobot can be functionalized using a variety of chemical compounds [ 8 ]. It has been investigated in nanomedicine in DDS, which operates directly on targeted cells of the human body. Researchers create devices that can administer medications to precise places while simultaneously adjusting the dose and amount of release. This DDS using nanorobots can be used to treat joint disorders, dental problems, diabetes, cancer, hepatitis and other conditions [ 2, 9 - 12 ]. One of the benefits of this technology is the potential to diagnose and treat illnesses with minimal impact on normal tissues, minimizing the likelihood of negative effects and guiding healing and remodelling therapy at the cellular and sub-cellular levels [ 13, 14 ]. Chemotherapy drug delivery using nanorobots in cancer treatment New advances in medication delivery have resulted in greater quality in targeted drug delivery that uses nanosensors to detect particular cells and regulate discharges through the use of smart medicines [ 1 ]. Traditional chemotherapeutic drugs act by eliminating swiftly replicating cells, which is a primary feature of malignant cells. Most anticancer medications have a limited therapeutic boundary, often resulting in cytotoxicity to normal stem cells that proliferate quickly, such as bone marrow, macrophages, gastrointestinal tract (GIT), and hair follicles, causing adverse effects like myelosuppression (lower synthesis of WBCs, producing immunosuppression), mucositis (inflammation of the GIT lining), alopecia (hair loss), organ malfunction, thrombocytopenia/anaemia, and haematological side effects, among other things. Doxorubicin is used to treat numerous forms of cancer, including Hodgkin's disease, when it is combined with other antineoplastic medicines to minimize its toxicity [ 15, 16 ]. Paclitaxel is a drug that is injected intravenously and is used to treat breast cancer. Some of the significant side effects include bone marrow suppression and progressive neurotoxicity. Cisplatin is an alkylating drug that results in the intra-DNA binding filament. Its negative effects include giddiness and severe vomiting, and it can be nephrotoxic [ 1 ]. Camptothecin is applied to treat neoplasia by inhibiting type 1 topoisomerases, an enzyme required for cellular duplication of genetic information. Numerous initiatives have been launched with the goal of employing nanotechnology to build DDS that can reduce the negative impacts of traditional therapy. On the surface of single-walled carbon nanotubes (SWNTs), doxorubicin was layered [ 17 ]. Doxorubicin was used in metastatic tumour cells as a polymer prodrug/collagen hybrid. The use of polymeric pro-drug nanotechnology in the therapy of rapidly dividing abnormal cells is a novel advance in the field [ 18 ]. Nanotechnology is continually looking for biocompatible materials that may be used as a DDS. The nanoparticle hydroxyapatite (HA), a significant component of bone and teeth, was employed to deliver paclitaxel, an anti-neoplastic medication, and the out-turn implies that therapy should begin with hydrophobic medicines [ 19 ]. Various initiatives have been launched with the goal of employing nanotechnology to build DDS, which can reduce the negative influence of traditional chemotherapy. The limitation of conservative chemotherapeutics is that it is unable to target malignant cells exclusively. These above-listed adverse effects often result in a delay in treatment, reduced drug dose or intermittent stopping of the therapy [ 20 ]. Given the ability of nanorobots to travel as blood-borne devices, they can aid in crucial therapy procedures such as early diagnostics and smart medication administration [ 21 ]. A nanorobot can aid with smart chemotherapy for medication administration and give an efficient early dissolution of cancer by targeting only the neoplastic-specific cells and tissues and preventing the surrounding healthy cells from the toxicity of the chemotherapy drugs so being used. Nanorobots as drug transporter for timely dose administration allow chemical compounds to be kept in the bloodstream for as long as essential, giving expected pharmacokinetic characteristics for chemotherapy in the therapies for anti-cancer as shown in Figure 1 [ 22 - 25 ]. The clinical use of nanorobots for diagnostic, therapy, and surgery can be accomplished by injecting them via an intravenous route. The nanorobots may be getting intravenously injected into the body of the recipient. The chemotherapy pharmacokinetics comprises uptake, metabolism, and excretion, as well as a rest period to allow the body to re-establish itself ahead of the succeeding chemotherapy session. For tiny tumours, patients are often treated in two-week cycles [ 26 ]. As a primary time threshold for medical purposes, nanorobots can be used to assess and diagnose the tumour within a short span of time using proteomic-based sensors. The magnetic resonance contrast-agent uptake kinetics of a very small molecular weight can forecast the transport of protein medicines to solid tumours [ 27 ]. Testing and diagnostics are critical components of nanorobotics study. It provides speedy testing diagnosis at the initial visit, eliminating the need for a follow-up appointment following the lab result, and illness identification at an earlier stage. The demand for energy for propulsion is a restriction in the usage of nanorobots in vivo. Because small inertia and strong viscous forces are associated with less productivity and less convective motion, higher quantities of energy are required [ 28 ]. Drug retention in the tumour will decide the medication's effectiveness after nanorobots pass cellular membranes for targeted administration. Depending on its structure, medication transport pathways from plasma to tissue impact chemotherapy to achieve more effective tumour chemotherapy [ 27 ]. According to the latest research, nanotechnology, DNA production of molecular-scale devices with superior control over shape, and site-specific functionalisation assures interesting benefits in the advancement of nanomedicine. However, biological milieu uncertainty and innate immune activation continue to be barriers to in vivo deployment. Thus, the primary benefit of nanorobots for cancer medicine administration is that they reduce chemotherapeutic side effects. The nanorobot design integrates carbon nanotubes and DNA, which are current contenders for the latest types of nanoelectronics, as the optimum method [ 29 ]. As a compound bio-sensor with sole-chain antigen-binding proteins, a complementary metal oxide semiconductor (CMOS) is used for building circuits with characteristic sizes in tens of nanometres [ 30 ]. For medicament release, this approach employs stimulation elicited upon proteomics and bioelectronics signals. As a result, nanoactuators are engaged to adjust medication delivery whenever the nanorobot detects predetermined modifications in protein gradients [ 1, 31 ]. Thermal and chemical signal changes are relevant circumstances directly connected to significant medical target identification. Nitric oxide synthase (NOS), E-cadherin, and B cell lymphoma-2 (Bcl-2) are some instances of fluctuating protein aggregation within the body near a medical target under diseased conditions. Furthermore, temperature changes are common in tissues with inflammation [ 32 ]. The framework integrates chemical and thermal characteristics as the most essential clinical and therapeutic recommendations for nanorobot template analysis. It also integrates chemical and thermal characteristics as the most essential diagnostic and therapeutic recommendations for nanorobot framework evaluation. The simulation in a three-dimensional real-time setting attempts to provide a viable model for nanorobot foraging within the body. One of the breakthroughs describes a hardware structure rooted in nano-bioelectronics for the use of nanorobots in neoplasia therapy [ 33, 34 ]. The continuous venture in building medical micro-robots has led to the initial conceptual framework research of a full medical nanorobot until now issued in a peer-reviewed publication, "Respirocytes", detailed a theoretical unnatural mechanical red blood cell, or "Respiro-cytes", consisting of 18 billion perfectly ordered architectural atoms proficient in delivering 236 times extra oxygen to the tissues and cells of the body per unit volume than normal red blood cells [ 35 ]. Microbivores, or unnatural phagocytes, might monitor the circulation, searching for and eliminating pathogens such as bacteria, viruses, or fungi. These nanobots may use up to 200 pW continuously. This capability is employed to break down germs that have been entrapped. Microbivores have biological phagocytic defences that are either organic or antibiotic-assisted, and they can operate up to 1, 000 times quicker. Even the most serious septicaemic diseases will be eliminated by microbivores within a short span of time. Because virulent microorganisms are entirely digested into harmless sugars and amino acids, which are the nanorobot’s sole discharge, the nanorobots reject the advanced possibility of sepsis or septic shock [ 36, 37 ]. Figure 1 Challenges of nanorobots in drug delivery. This image throws light on the various challenges that different shaped nanorobots face when employed for drug delivery [ 38 ]. Image reprinted from [ 38 ] under Creative Commons Attribution 4. 0 International License. Future of nanotechnology in the area of medicine To bring in combination the required collaborative skills to produce these unique technologies, numerous conventional streams of science, such as medicine, chemistry, physics, materials science, and biology, have come together to form the expanding field of nanotechnology. Nanotechnology has a vast span of possible applications (Figure 2 ) [ 39 ], from improvements to current practices to the creation of entirely new tools and skills. The last few years have observed an exponential increase of interest in the topic of nanotechnology and research, which has led to the identification of novel applications for nanotechnology in medicine and the emergence of an advanced branch called nanomedicine. It includes the science and technology of diagnosing, treating, and preventing illness, traumatic injury, and alleviating pain; conserving and enhancing human health using nanoscale architectured materials, biotechnology, and genetic engineering; eventually, complex machine systems and nanorobots, known as "nanomedicine" (Figure 3 ) [ 40, 41 ]. Figure 2 Illustration showing various other applications of nanotechnology in medicine. The discovery of nanoparticles with the help of nanotechnology has led to its various uses in the area of medicine. The nanoparticle so created can be employed for various uses like in the manufacturing of nano implants, tissue engineering for drug delivery systems, gene delivery systems, drug screening, theranostics, cancer therapy, biomarker mapping, disease detection, and bio-imaging [ 39 ]. Image reprinted from [ 39 ] under Creative Commons Attribution 4. 0 International License. Figure 3 Schematic diagram of the current trends of micro/nanorobotics in precision medicine. Nanorobots are being used in various domains of pre-clinical and clinical medicine. In pre-clinical medicine, nanorobots are being employed in bioimaging and various delivery systems of drugs, gene therapy, living cells, and inorganic therapeutics. Similarly, nanorobots in clinical medicine are being extensively used in disease diagnosis and surgeries for biopsy, biofilm degradation, tissue collection, and sampling [ 41 ]. Image reprinted from [ 41 ] under Creative Commons Attribution 4. 0 International License. In vivo diagnostics, nanomedicine might create technologies that can act within the human body to diagnose ailments earlier and identify and measure toxic chemicals and tumour cells. In the surgical aspect, when launched into the body through the intravenous route or cavities, a surgical nanorobot controlled or led by a human surgeon might work as a semi-autonomous on-site surgeon. An inbuilt computer might manage the device's operations, such as looking for disease and identifying and fixing injury by nanomanipulation while maintaining communication with the supervising surgeon via coded ultrasonic signals [ 37 ]. By transforming mechanical energy from bodily movement, muscle stretching, or water flow into electricity, scientists were able to design a new generation of self-sustained implanted medical devices, sensors, and portable gadgets [ 39 ]. Nanogenerators generate electricity by bending and then releasing piezoelectric and semiconducting zinc oxide nanowires. Nanowires may be produced on polymer-based films, and the utilization of flexible polymer substrates may one day allow portable gadgets to be powered by their users' movement [ 39 ]. Fluorescent biological labelling, medication and gene delivery, pathogen identification, protein sensing, DNA structure probing, tissue engineering, tumour identification, separation and purification of biological molecules and cells, MRI contrast enhancement, and phagokinetic research are among the uses. The extended duration effect of nanomedicine study is to describe quantitative molecular-scale components called nanomachinery. Accurate command and manipulation of nanomachinery in cells can lead to a more diverse and advanced gain in the interpretation of cellular processes in organic cells, as well as the creation of new technologies for disease detection and medication. The advantage of this research is the formation of a platform technology that will affect nanoscale imaging methodologies aimed to investigate molecular pathways in organic cells [ 40, 42 ]. Conclusions The main target of writing this review was to provide an outline of the technological development of nanotechnology in medicine by making a nanorobot and introducing it in the medication of cancer as a new mode of drug delivery. Cancer is described as a collection of diseases characterised by the unregulated development and spread of malignant cells in the body, and the number of people diagnosed every year keeps adding up. Cancer treatment is most likely the driving force behind the creation of nanorobotics; it can be auspiciously treated using existing medical technology and therapeutic instruments, with the major help of nanorobotics. To decide the prognosis and chances of survival in a cancer patient, consider the following factors: better prognosis can be achieved if the evolution of the disease is time-dependent and a timely diagnosis is made. Another important aspect is to reduce the side effects of chemotherapy on the patients by forming efficient targeted drug delivery systems. Programmable nanorobotic devices working at the cellular and molecular level would help doctors to carry out precise treatment. In addition to resolving gross cellular insults caused by non-reversible mechanisms or to the biological tissues stored cryogenically, mechanically reversing the process of atherosclerosis, enhancing the immune system, replacing or re-writing the DNA sequences in cells at will, improving total respiratory capacity, and achieving near-instant homeostasis, medically these nanorobots have been put forward for use in various branches of dentistry, research in pharmaceuticals, and aid and abet clinical diagnosis. When nanomechanics becomes obtainable, the ideal goal of physicians, medical personnel, and every healer throughout known records would be realized. Microscale robots with programmable and controllable nanoscale components produced with nanometre accuracy would enable medical physicians to perform at the cellular and molecular levels to heal and carry out rehabilitating surgeries. Nanomedical doctors of the 21st century will continue to make effective use of the body's inherent therapeutic capacities and homeostatic systems, since, all else being equal, treatments that intervene the least are the best.

10. 7759/cureus. 2953

Cureus

Post-football Gonathrosis: Injuries and Surgeries are A Risk

Football is one of the most popular sports in the world. Many studies have shown there is a high incidence of gonarthrosis in football players. The reason for this increase is said to be injuries to the meniscus, the anterior cruciate ligament (ACL) and the resulting surgeries. The incidence is significantly increased in players with knee injuries. The knee is also the most commonly injured site in football and the most common cause of surgery in football players. Together these injuries, particularly of the ACL or meniscus and the resulting surgeries, increase the risk of developing gonarthrosis in post-football years.

Introduction and background Football is one of the most popular sports in the world with over 300 million players involved in it. Many studies performed show that there is a high prevalence rate (60%-80%) of osteoarthritis (OA) in footballers post-retirement [ 1 ]. Football is a very dynamic game involving physical activities like running, jumping, dribbling and tackling; however, there is no evidence that suggests any of these activities are directly linked to the development of osteoarthritis in their post-football days [ 2 ]. However, researches have generally shown that there is an increased occurrence of osteoarthritis in former elite players as compared to age-matched controls. A study published in the United Kingdom reported that knee injuries, especially of the cruciate ligament and meniscal injuries, are the cause of nearly half of all the injuries that resulted in forceful early retirement in football players during the period of 1987-1988 to 1994-1995 [ 3 ]. According to an article published in the journal named Clinics in Sports Medicine, it was suggested that the primary cause of osteoarthritis in football players is injuries to the meniscus, the anterior cruciate ligament, and the surgeries as a result of these injuries [ 4 ]. The reason for this may be the limited healing capacity of the cartilage and other intra-articular soft tissues such as anterior cruciate ligament (ACL) and meniscus that joint injuries often lead to the development of disabling osteoarthritis. This is one of the reasons why osteoarthritis is five to 12 times more common in football players as compared to the general population and is diagnosed four to five years earlier as well [ 5 ]. Review Osteoarthritis Osteoarthritis is a degenerative joint disease which involves the joint cartilage and its surrounding structures. During the course of the disease, there is damage of articular cartilage, remodeling of sub-articular bone, osteophyte formation, ligament laxity, weakening of peri-articular muscles and in some cases synovial inflammation may also be seen [ 6 ]. Osteoarthritis results from a failure of chondrocytes to maintain a balance between synthesis and degradation of the extracellular matrix proteins which include two components, tissue fluid and a framework of structural molecules consisting of collagen type II, proteoglycans, and non-collagenous proteins and glycoproteins [ 7 ]. About 13% of women and 10% of men aged 60 years or older have symptomatic knee osteoarthritis. These percentages are likely to increase due to increasing age of the population and increased rate of obesity in the general population along with other risk factors which include sports participation, injury to the joint, genetic susceptibility, female gender, bone density, muscle weakness, joint laxity. Mechanical forces exerted on joints are one of the most modifiable risk factors and it can be determined using a person’s basal metabolic index. Female sex, lower education levels, obesity, and poor muscular strength are important risk factors for symptomatic disease and subsequent disability [ 8 ]. The result of these changes in the joint includes joint pain, dysfunction and ultimately in advanced stages joint contractures, muscle atrophy, and limb deformity [ 7 ]. In advanced stages the patient may also be in psychological distress and it is important to screen the patient for signs of anxiety like abnormal posture to avoid excessive pain, insomnia, and signs of depression like early morning wakening, irritability, weight loss, problems remembering things or concentrating [ 9 ]. The diagnosis of osteoarthritis is usually made on the basis of history and clinical examination. The role of radiology is usually to confirm the diagnosis and rule out other conditions [ 9 ]. Treatment options for osteoarthritis are divided into non-pharmacological and pharmacological modalities [ 10 ]. Weight reduction is one of the first steps in the management of osteoarthritis [ 11 ]. In a study published by Huang et al. , it was seen that pain reduction and improvement of movement was seen in the population having OA who were undergoing weight reduction therapy [ 11 ]. Other non-pharmacological approaches include the avoidance of activities that put excessive stress on the joints. However, the most commonly used modality is rehabilitation and physical therapy [ 12 ]. Pharmacological management includes the use of acetaminophen, aspirin, and other non-steroidal anti-inflammatory drugs (NSAIDs) for the relief of pain. Glucosamine sulfate and chondroitin sulfate are used as nutritional supplements as they are used in the manufacture and repair of cartilage [ 13 ]. The surgical procedures performed include arthroscopic surgery for OA with meniscal tear, osteotomy to correct misalignment and total joint replacements. A new treatment modality that uses tissue engineering and biological therapy is called autologous chondrocyte implantation in which chondrocytes form a normal cartilage are harvested and implanted into the defected cartilage affected by OA [ 14 ]. Despite all these treatment modalities OA causes major disability in the population with an estimated 22. 5 million adults from the United States (US) having difficulty walking three city blocks and an estimated 21. 7 million adults from the US having difficulty climbing stairs [ 15 ]. Osteoarthritis in knee joint Knee is the most commonly affected joint in osteoarthritis with an involvement of 41% as compared to 30% in hands and 19% in hips [ 16 ]. According to a study published by Zhang et al. Symptomatic gonarthrosis affects 13% of women and 10% of men aged 60 years and older. In a study conducted by Duncan et al. the most common pattern of radiological osteoarthritis involvement of these two joints is in combination (40%), with isolated patello-femoral joint (PFJ) involvement seen in 24% and isolated tibio-femoral joint (TFJ) involvement seen in 4% only. [ 17 ] Prolonged squatting is an important risk factor for tibio-femoral knee osteoarthritis in the elderly population [ 18 ]. Along with this obesity and meniscectomy are also stronger risk factors for TFJ disease while Heberden’s nodes and family history are more closely associated with PFJ involvement [ 8 ]. The most important exogenous risk factors for developing Secondary OA are injury related macro-trauma, repetitive micro-trauma, increased BMI and previous surgery [ 19 ]. The direction of mal-alignment of knee serves as an indication for the type of OA in the knee joint, frontal plane mal-alignment may indicate patella-femoral joint OA and tibio-femoral joint OA, valgus mal-alignment is associated mostly with lateral patella-femoral joint OA, varus mal-alignment with medial tibio-femoral joint OA with the former one i. e. lateral PFJ OA being more common of the two [ 20 ]. The sign and symptoms of moderate to severe isolated PFJ OA include dramatic swelling in the past, valgus mal-alignment, marked reduction in quadriceps strength, and pain on compression while those of TFJ joint involvement include history of previous trauma, varus mal-alignment, bony enlargement, reduced knee flexion range of motion, and fixed flexion deformity [ 21 ]. The treatment options for knee osteoarthritis are similar as discussed above, which include weight reduction, avoidance of excessive stress on knee joints, analgesics, use of supplements like proteoglycans and chondroitin sulfate, physical therapy, and rehabilitation. In severe cases, surgery is opted which includes total and partial knee replacements [ 10 ]. Osteoarthritis in football players: Cause Football, Injuries or surgery OA has been seen to be a common occurrence in football players after their retirement from the sport. In a study conducted on 117 former top-level athletes, it was found that the prevalence of tibio-femoral and patella-femoral OA was 31% in weightlifters, 29% in football players, 14% in runners and 3% in shooters. It was also found that football players had the highest prevalence of tibio-femoral OA and weight lifters had the highest prevalence of patella-femoral OA [ 22 ]. Another study conducted by Krajnc et al. (2010) measured the prevalence of gonarthrosis in 40 former Slovenian football players and reported a 60% rate of gonarthrosis [ 23 ]. In another study conducted by Drawer et al. on 500 former players registered by the English Professional Footballers’ Association (PFA) with a response of 37% (185), he showed that around 47% respondents retired early due to injury out of which 42% were due to acute injuries and 58% were due to chronic injuries. He further showed that both of these acute (46%) and chronic (37%) injuries were most common of the knee. Out of those 185 respondents, the prevalence of medically diagnosed OA was 32%, among whom the rate of developing OA was double (51%) in the players who retired through injury in contrast to the players who didn’t retire through injury (25%) [ 24 ]. Roos et al. , in their comparative study between elite football players, non-elite football players, and normal age-matched controls found that the prevalence of gonarthrosis in normal age-matched controls was 1. 6%, non-elite players was 4. 2% and among elite players was 15. 5%. If the number of players with injuries from the non-elite group was excluded, there was no difference in the prevalence of gonarthrosis between the two groups but the prevalence was still high in the elite group [ 25 ]. Table 1 shows the prevalence of Radiographic gonarthrosis in football players in comparison to age-matched normal people. Results show that there is an increased prevalence of gonarthrosis in football players. Table 1 Prevalence of Radiographic Gonarthrosis Author In footballers In Normal Petrillo S. et al. [ 26 ] 53. 7% 31. 9% Paxinos O. et al. [ 27 ] 52% 33% Chantraine A. et al. [ 28 ] 56% - So should football be included as a cause of OA as an occupational illness? Spahn in 2015 analyzed multiple articles to find if playing football without injuries is a cause of development of OA and found out that there is only a slight increased risk (relative risk=1. 3) of gonarthrosis in players without injury [ 29 ]. In contrast in studies without differentiation between injured and non-injured knee the risk was significantly increased. (Relative risk=2. 9). Gillquist et al. in their study also showed that radiographic OA of knee is significantly increased after all knee injuries [ 30 ]. So exogenous, contact related trauma and the resulting injuries are listed as the major predisposing factors for the occurrence of early OA in football players [ 29 ]. In a review article, Blagojevic et al. stated that previous knee trauma increases the risk of OA 3. 86 times [ 31 ]. The incidence of knee injuries of all the football injuries is 15-19%, out of which 35%-37% are strains, 20-21% are sprains and 16-24% contusions; they constitute 58% of all the major injuries [ 32 ]. Of the ligamentous injuries, medial collateral ligament (MCL) is the most commonly injured ligament. In their study, Price et al. found that MCL injuries accounted for 85% of all the knee injuries suffered by football players [ 33 ]. ACL injury carries the highest morbidity with a successful return to play to pre-injury levels after surgery ranging from 50%-90% [ 34 ]. Janine et al. in their research performed on 217 players from eight Dutch clubs found out that the overall incidence of injuries was 6. 2 per 1000 player hours with training incidence of 2. 8 and match incidence of 32. 8. A team sustained an average of 1. 1 injuries per match. The results also showed that the most common body part injured was the knee with an incidence of 21. 3% [ 35 ]. Table 2 shows the incidence of knee injuries in football players and the incidence of knee surgeries as a result of these injuries. Results show there is a high incidence of knee injuries in football players and a high number of surgeries are done as a result of these injuries. Table 2 Incidence of Knee Surgeries and Injuries in Football Players Authors Knee injuries in football players Knee surgeries in football players Krajnc Z. et al. [ 36 ] 73% 43% Musumeci G. et al. [ 18 ] 17% 52% Soccer is one of the sports with the highest number of ACL injuries with an incidence rate of 0. 15%-3. 67% per person per year [ 37 ]. A study performed on elite European football showed that a high-level men’s team can expect 0. 4 ACL injuries per season [ 38 ]. Tears of the lateral meniscus are more common in football than any other sports. Knee injuries are also the most common reason for surgery in football players [ 39 ]. These injuries and the resulting surgeries increase the risk of gonarthrosis in football players. Gillquist et al. showed that isolated meniscus rupture and subsequent repair, partial or total ruptures of the ACL without other major injuries increase the risk tenfold to around 15%-20% incidence of OA as compared to the age-matched, uninjured population in which the incidence is only 1%-2% [ 30 ]. Table 3 shows that there is a high prevalence of radiographic gonarthrosis in football players with ACL injuries. Table 3 Prevalence of Radiographic Osteoarthritis with Anterior Cruciate Ligament Rupture Author Prevalence Von Porat A. et al. [ 40 ] 78% Lohmander LS. et al. [ 41 ] 51% In a retrospective review, Neyret et al. analyzed 91 knees having the same operation, rim-preserving meniscectomy, with a follow up of average 27 years and found that in patients operated with intact ACL the prevalence of radiological diagnosed OA was 24% and with ruptured ACL was 77% [ 42 ]. Meniscectomy in a join with intact ligaments doubles the risk of development of gonarthrosis to 30%-40% while 50%-70% patients with complete ACL rupture and associated injuries have radiographic changes of OA after 15-20 years [ 30 ]. A recent study has showed that all footballers under-going revision ACL surgery had OA when they were examined 37 months after reoperation [ 43 ]. Smith et al. , in their research, showed that the incidence of gonarthrosis was only 4% in those American football players without any previous knee surgery, 11% in those with history of meniscus repair and 24% of those with a history of ACL reconstruction. It was also noticed that in knees with previous ACL reconstruction, the rate of OA in the tibio-femoral compartment doubled in those patients who have had previous meniscal surgery [ 44 ]. The cost of ACL injuries is estimated to be US $4 billion for surgical treatment alone and US$ 7. 6 billion annually when treated with ACL reconstruction and US $17. 7 billion when treated with rehabilitation. Even after so much effort, 59%-70% will develop radiographic OA, 16%-19% will have symptomatic gonarthrosis in the course of their lifetime, and 13-15% will need total knee replacements [ 45 ]. Conclusions From the above observations, it can be concluded that football is one of the most common sports being played throughout the world at the present time. Playing football in itself increases the risk of developing gonarthrosis only slightly but is still higher than the age-matched normal population. The reason for this is said to be the repetitive micro-trauma suffered by the joint due to the vigorous training and playing hours. The risk for developing gonarthrosis is significantly increased due to injuries particularly the ACL injury or injury to the meniscus. Football is the sport with the highest number of ACL injuries. Surgeries, particularly for ACL reconstruction or meniscectomy, performed further increases the risk of developing gonarthrosis in the post-football life.

10. 7759/cureus. 29850

Cureus

Evaluation of the Chemical, Morphological, Physical, Mechanical, and Biological Properties of Chitosan/Polyvinyl Alcohol Nanofibrous Scaffolds for Potential Use in Oral Tissue Engineering

Background Chitosan is a biocompatible, biodegradable, and non-toxic natural polymer that can be fabricated by different methods for use in dental and biomedical fields. Electrospinning can produce polymeric nanofibrous scaffolds and membranes with desirable properties for use in tissue engineering. The objectives of this study were to investigate several morphological, physical, and biological characteristics of these nanofibrous scaffolds and evaluate their potential use in tissue engineering. Methodology Chitosan/polyvinyl alcohol nanofibrous scaffolds (CS/PVA NFS) in a ratio of 70/30 were fabricated by conventional electrospinning. The scaffolds were evaluated chemically by Fourier transformed infrared spectroscopy (FTIR) and morphologically by the atomic force microscope (AFM) and the field emission-scanning electron microscope (FE-SEM). These scaffolds were also evaluated mechanically by a tensile strength test and several investigations, including water contact angle, swelling ratio, and degradation ratio. Biological evaluations included protein adsorption, cell culture, and cell viability assay. Results The morphological evaluation revealed a homogenous, bead-free mat with an average fiber diameter of 172. 7 ± 56. 8 nm, an average pore size of 0. 54 ± 0. 17 µm, and porosity of 74. 8% ± 3. 3%; the scaffolds showed a tensile strength of 6. 67 ± 0. 7 Mpa. Scaffolds showed a desired hydrophilic property, as shown by the water contact angle test with a mean angle of 29. 5°, while the swelling ratio was 229%, and degradability in phosphate buffer solution after 30 days was 26. 9 ± 2. 9%. In-vitro cell culture study with adipose tissue mesenchymal stem cells and cell viability and cytotoxicity tests by MTT assay demonstrated well-attached cells with increasing proliferation rate with no signs of cytotoxicity. Conclusions Assessment of the CS/PVA NFS revealed randomly oriented bead-free and porous mats. The scaffolds were stable at aqueous solutions following thermal treatment. They were hydrophilic, biodegradable, and biocompatible, as shown by the cell culture and MTT assay, which suggest that the fabricated scaffolds have the potential to be used in tissue engineering applications either as scaffolds, bio-grafts, or barrier membranes.

Introduction Polymeric nanomaterials are natural, synthetic, and composite polymers manufactured by one of the nano- approached techniques and have at least one of their three dimensions in the nanoscale [ 1 ]. They have attracted great attention lately as many of them are widely available, cheap, easy to process and synthesize, non-toxic, biocompatible, biodegradable, non-antigenic, eco-friendly, and some of them possess intrinsic antimicrobial properties [ 2, 3 ]. They have been incorporated into various technological applications, such as electrical and electronic industries, water treatment, heavy metal chelating, food industries, and agriculture [ 4, 5 ], as well as in biomedical fields such as dentistry, pharmacology, and medicine [ 6, 7 ]. They can be synthesized in various shapes and sizes, such as nanoparticles, nanorods, nanofibers, nanohydrogels, nanosponges, and nanotubes [ 8, 9 ]. Among these different nanoforms, nanofibers of polymeric materials have been intensively investigated in the biomedical field as they can form scaffolds for tissue engineering, drug carriers, wound dressings, and bio-grafts [ 10 ]. The fabrication of these nanofibers can be achieved by various techniques, such as electrospinning, self-assembly, phase separation method, and template synthesis [ 11 ]. Electrospinning has been a widely accepted method for fiber fabrication because of its efficiency, simplicity, applicability, and ease of use [ 12 ]. During electrospinning, fibers are formed when applying a high-voltage electrostatic field between the needle tip and the collector while the solvent evaporates. The produced nanofibers can form scaffolds with better physical, chemical, and biological properties, such as a high surface area to volume ratio, high tensile strength, high compressibility, good abrasion resistance, high porosity, good biodegradability, and similarity to the extracellular matrix [ 13 ]. To fabricate a suitable scaffold for tissue engineering purposes, one should consider using these polymeric nanomaterials for their outstanding properties. Natural polymers such as collagen, silk fibroin, alginate, cellulose, hyaluronic acid, gelatin, elastin, chitin, and chitosan have been studied extensively in this field [ 14, 15 ]. Chitosan is a unique positively charged biopolymer found in the cell wall of some fungi and yeasts and can be produced by partial or full deacetylation of its ancestor, chitin [ 16 ]. In addition to the unique properties of the natural polymeric materials, chitosan has native antibacterial properties [ 17 ]; can adhere to negatively charged surfaces; attract cells, proteins, hormones, and growth factors [ 18 ]; and has multifunctional active groups that enable it to be modified, interact with, and be cross-linked with a wide variety of materials yielding new composites or enhancing the chemical, physical, and biological properties of chitosan [ 19 ]. The excellent biological and chemical properties of chitosan have spotted a light on it as a candidate material for biomedical field applications. Pure chitosan nanofibers can be produced either by using a high concentration of acetic acid [ 20 ] or by using trifluoroacetic acid (TFA) and dichloro methanol (DMO) [ 21 ] as solvents, either of which can be environmentally harmful and can have toxicity issues; additionally, the mechanical properties of the produced nanofibers are limited [ 22 ]. For these reasons, chitosan nanofibers have been prepared with other natural polymers such as silk fibroin [ 23 ], collagen [ 24 ], and gelatin [ 25 ], or synthetic polymers such as polyvinyl alcohol (PVA) [ 26 ], polyethylene oxide (PEO) [ 27 ], and polylactic acid (PLA) [ 28 ]. These composite nanofibers have better physical, chemical, and biological properties. PVA is a synthetic polymer that is biocompatible, biodegradable, non-toxic, and electro-spinnable [ 29 ]. When blended with chitosan, it causes interference with the rigid connections between chitosan molecules and binds with chitosan by hydrogen bonds and facilitating electrospinning and enhancing the scaffold’s properties [ 30 ]. Chitosan has been used in the biomedical field as a hemostatic agent, a wound dressing material, a drug carrier; in the delivery of proteins, growth factors, and vaccines; in anti-tumor therapy; and in biological imaging [ 31 ]. It can also be used in soft and hard tissue engineering. It can be used as a bio-scaffold for cell attachment and proliferation, as a grafting material with other polymers and bio-ceramics, or as a barrier membrane in guided bone regeneration and soft-tissue regeneration [ 32 ]. In the current work, chitosan/polyvinyl alcohol nanofibrous scaffolds (CS/PVA NFS) were produced by the electrospinning method and then modified by a heat treatment process. There is no previous report that has assessed the chemical, morphological, physical, mechanical, and biological characteristics of these materials. The suitability of these scaffolds in tissue engineering has not yet been investigated. Therefore, the current investigation aimed to assess the chemical composition, morphological structure, tensile strength, wettability, swelling ratio, and degradation rate of these scaffolds. In addition, an assessment of the biological characteristics such as protein adsorption, cell viability, and cell culture was also performed to examine the possibility of implementing these scaffolds in oral tissue engineering. Materials and methods This research work was approved by the University of Damascus Local Research Ethics Committee (ID: DN-30082022-12). Preparation of solutions Chitosan powder (medium molecular weight = 190-310 kilodaltons, degree of deacetylation (DDA) = 85%, viscosity 200-800 centipoise), PVA (molecular weight = 78. 000 g/mol; hydrolysis 99. 8%), glacial acetic acid 99%, absolute ethanol, and phosphate buffer saline (PBS) with a pH of 7. 4 (Sigma Aldrich, St. Louis, MO, USA). All chemicals were used without any further purification. Our distillation unit made deionized (DI) water. All the tests were performed at the nanotechnology laboratory, Department of Physics, Faculty of Science, Damascus University. Chitosan solution was prepared by dissolving 3% weight/volume (W/V) in 2% aqueous acetic acid and stirred using a magnetic stirrer at a speed of 400 rounds per minute (rpm) for 24 hours to obtain a homogenous solution, 10% of PVA was added to 90 mL of DI water and stirred at a speed of 400 rpm for two hours at 60˚C, and then for four hours at room temperature. Both solutions were centrifuged to eliminate undissolved and unsuspended particles, and then both chitosan and PVA were mixed in a ratio of 70/30 under 200 stirring speed for 24 hours at room temperature. Electrospinning of CS/PVA The electrospinning apparatus consists of an electrostatic/high-voltage generator (ES813 D50. 1, EsdEmc Technology Rolla, MO, USA), a syringe pump (SN 50C6T, Sino device Technology Co, Ltd, China), and a sheet of aluminum as a collector. CS/PVA solution was loaded onto the syringe and mounted on the pump, the applied voltage was set at a value of 23 kV, the distance between the needle tip and collector was 14 cm, and the pump rate was adjusted at 0. 02 mL/hour was kept constant throughout the experiment. The scaffolds were thermally treated at 100˚C for six hours and sterilized by gamma radiation (25 kGy). Characterization of the produced scaffolds Chemical Analysis: Fourier Transformed Infrared spectroscopy Chemical analysis with the Fourier transformed infrared spectroscopy (FTIR) was conducted using a Bruker spectrometer (Bruker Tensor 27 IR, US) with a wavenumber range between 400 and 4, 000 and a resolution of 4 cm -1. It is used to identify chemical substances and functional groups in different matter forms [ 33 ]. Morphological Analysis Field emission-scanning electron microscope (FE-SEM): Evaluation of the morphological microstructure of the prepared scaffolds was conducted using FE-SEM (MERA3 TESCAN, Brno, Czech). The average diameter of the nanofibers was calculated according to the measurement of 100 randomly selected fibers from different parts of the scaffolds and pore size using ImageJ software (ImageJ, U. S. National Institute of Health; Bethesda, Maryland, USA) [ 34 ]. Atomic force microscopy (AFM): Topographical evaluation of the prepared scaffolds was performed using AFM (EasyScan2 FlexAFM, Nanosurf, Leistal, Switzerland) in the tapping mode. Similarly, the average diameter and pore size of the scaffolds were calculated using the device corresponding software program Nanosurf Report Expert v 5. 0 (Nanosurf, Leistal, Switzerland). Mechanical Analysis: Tensile Strength Tensile strength for heat-treated and untreated CS/PVA mats was performed using a tensiometer (Model M250-2. 5CT, Testometric Co Ltd. , Lancashire, UK) and using the ASTM D882-12 standard test method [ 35 ]. In brief, the nanofibrous mats were cut into small specimens of 10 × 50 mm strips, mounted into the grips with an initial grip distance of 20 mm, and stretched until breakage with a strain rate of 10 mm/minute. Ten samples of each group were measured, and the averages of Young’s modulus, elongation at break, and the ultimate tensile strength were calculated from the stress/strain curve. Physical Analysis Water contact angle (WCA): A droplet of deionized water was deposited on the surface of the scaffold, and images were taken at different time intervals (1, 30, and 60 seconds) using a digital camera (Canon PowerShot A520, Canon Inc. , NY, USA). The image processing and angle measurements were done to determine the contact angles. Porosity, pore distribution, and nanofiber density: The percentage of porosity of the prepared scaffolds and the density of nanofibers were determined by the method described by Lim et al. [ 36 ]. In brief, absolute ethanol was used as displacement liquid. In brief, the scaffolds were cut into small pieces, and the volume of the scaffolds was taken (Vd). Then the dry weight (W1) was noted and immersed in a known volume of absolute ethanol until saturation. Subsequently, the final weight after immersion was token (W2), and the porosity percentage and the fiber density were calculated using the following equations: W eth in pores = w2 - w1 Veth = W eth in pores /ρV Porosity %(Ɛ) = {V eth /V d } × 100. True volume of fibers V f = V d - V eth The density of the fibers (d) = W 1 /V f Where W eth in pores is the weight of ethanol in pores, V eth is the volume of ethanol in the pores, and ρ is a constant that represents the density of alcohol. The experiment was repeated thrice. The mean and standard deviation values were calculated. Swelling behavior and degradation rate of the CS/PVA NFS: The swelling ratio of the scaffold was evaluated using the method from Meng et al. [ 37 ]. First, the scaffold was cut into five square shapes (about 10 × 10 mm). Then the initial weight of scaffolds (Wi) was noted, and then scaffolds were immersed in distilled water for 24, 48, and 72 hours. In each period, the samples were rinsed, and excessive water was removed by gently dipping them in a filter paper and then weighed in wet condition (Ww). The swelling ratio was calculated according to the following equation: Swelling ratio % = [(Ww-W i )/W i ] × 100. The degradation test was conducted by the method described by Agrawal and Pramanik [ 38 ]. In brief, the evaluation was done by noting the initial weight of the samples (Wi); the samples were then placed in 5 mL of PBS and incubated at 37°C for 30 days, the samples were taken out from the buffer at specific time intervals and rinsed with deionized water and dried, and their final weight was noted (Wf). The degradation rate was then calculated according to the following formula: Weight loss % = [(W i - W f )/W i ] × 100. Biological Analysis Protein adsorption: To determine the amount of protein adsorbed onto the CS/PVA scaffold by ultraviolet-visible (UV-vis) spectroscopy, we used the method described by Liao et al. [ 39 ]. First, three samples were soaked in ethanol 70% for one hour and then rinsed with PBS. The samples were then incubated in 10 mL PBS containing 10% fetal bovine serum (FBS) at 37°C incubators for 24, 48, and 72 hours. The samples were then washed with PBS three times to eliminate excess proteins. The washing solution was returned to the FBS solution and quantified using a UV-vis spectrometer. The quantity of protein adsorbed onto the scaffold was calculated by finding the difference between the initial total amount of protein and the non-adsorbed protein. Cell culture and cell viability assay: Cell culture and cell viability assays were performed using the method described by Ghorbani et al. [ 40 ]. Human adipose tissue mesenchymal stem cells (h AD-MSC, Rockville, MD) were trypsinized and cultured in a minimum modified eagle medium (MEM) (Biochrom, Berlin, Germany) containing 10% FBS (Biochrom), 2 mM of L-glutamine (Biochrom), 100 U/mL penicillin (Merck), 100 mg/mL streptomycin, and 1 mM sodium pyruvate (Merck) in a 5% CO 2 atmosphere at 37˚C incubator. The culture medium was replaced every other day. The fourth passage was used in all cellular experiments. The scaffolds were cut into small rectangular (for cell adhesion and morphology tests) and circular shapes (for cell viability and cytotoxicity test), sterilized by gamma radiation, and put in a culture medium to enhance cell seeding later. For cell morphology and adhesion tests, a certain amount of cell concentration (1 × 10 5 cells/mL) was seeded on the Cs/PVA scaffolds and tissue culture plate (TCP) as a control sample. The seeded cells were incubated for 72 hours (5% CO 2 and 37°C), and the medium was changed daily. Scaffolds with cells were prepared for SEM investigation (washed by PBS, fixed, dehydrated, and dried). For the cell viability assay, the tetrazolium salt [3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide] (MTT) was used because the living cells can reduce the MTT tetrazolium compound into a colored formazan product that can be quantified by a spectrophotometer and is directly proportional to the number of living cells in that culture. In brief, the cells were seeded on the CS/PVA mats with culture medium as test groups and on TCPs as a control sample and put in an incubator (5% CO 2 and 37°C) for 24, 48, and 72 hours separately. The assay was done by aspiring the spent medium and adding 1 mL of MTT solution and 9 mL of fresh medium and incubating for four hours at 37°C and in darkness; absorbance measures were taken using a microplate reader at 570 nm. Results Chemical analysis: FTIR The FTIR results of pure chitosan showed a large peak at 3, 442 cm -1, which refers to O-H and N-H stretching, the 2, 919 cm -1 peak refers to C-H stretching, while peaks at 1, 645 cm -1, 1, 422 cm -1, and 1, 384 cm -1 represent C=O stretching, N-H bending, and C-O stretching, respectively. The last peak at 1, 067 cm -1 represents the -C-O-C- glycosidic linkage of chitosan polysaccharide monomers. On the other hand, pure PVA shows typical bands of hydroxyl groups (O-H) at 3, 313 cm -1 due to O-H stretching, and three peaks at 2, 950 cm -1, 1, 636 cm -1, and 1, 273 cm -1 refer to C-H stretching, C=O and C-O stretching of acetate group, and C-O-H bending, respectively (Figure 1 ). Figure 1 FTIR analysis of pure chitosan (in blue) and polyvinyl alcohol (in red). Cs: chitosan; PVA: polyvinyl alcohol; FTIR: Fourier transformed infrared spectroscopy All characteristic peaks of chitosan and PVA appeared in all spectra with the only difference in the intensity and shape of the peaks or little shifting, as shown in Figure 2. Figure 2 The FTIR analysis of the chitosan/polyvinyl alcohol blend. (A) Before electrospinning. (B) After electrospinning. (C) After heat treatment. Cs/PVA: chitosan/polyvinyl alcohol; ES: electrospinning; Fourier transformed infrared spectroscopy Morphological analysis Characterization of the Morphology and Topography of Nanofibers The FE-SEM images of the CS/PVA NFS showed continuous, bead-free, and randomly oriented nanofibers with a diameter ranging from 50 to 400 nanometer (nm), with an average diameter of 172. 7 ± 56. 8 nm (Figures 3, 4 ), whereas the surface roughness (arithmetical mean height (Sa)) ranged from 68. 3 to 101 nm, with a mean Sa of 83. 15 ± 10. 82 nm, and the root mean square height (Sq) ranged from 86. 1 nm to 131 nm, with a mean Sq of 105. 3 ± 12. 7 nm (Figure 5 ). Figure 3 Field emission-scanning electron microscope micrographs of electrospun Cs/PVA obtained in different magnifications: (A) 5, 000×. (B) 50, 000×. (C) 100, 000×. (D) 200, 000×. D: diameter of the fiber; Cs/PVA: chitosan/polyvinyl alcohol Figure 4 Fiber diameter and pore size of chitosan/polyvinyl alcohol nanofibrous scaffolds. (A) Average fiber diameter. (B) Average pore size. Figure 5 Atomic force microscopy surface micrographs of the chitosan/polyvinyl alcohol nanofibrous scaffolds. (A) General morphology. (B) Three-dimensional view of the scaffold. The pore size of the scanned scaffolds ranged from 0. 1 to 1 µm, with an average of 0. 54 ± 0. 17 µm, while the porosity percentage of the test samples was between 69. 9% and 79. 5%, with a mean value of 74. 8% ± 3. 3. Fiber density was 0. 676 ± 0. 058 g/cm 3, as shown in Table 1. Table 1 Porosity percentage and fiber density of chitosan/polyvinyl alcohol nanofibrous scaffolds. Porosity % Mean ± standard deviation Density Mean ± standard deviation First sample 69. 9 74. 8 ± 3. 3% 0. 638 g/cm 3 0. 67 ± 0. 058 g/cm 3 Second sample 75. 03 0. 63 g/cm 3 Third sample 79. 48 0. 763 g/cm 3 Mechanical analysis: tensile strength The ultimate tensile strength of the heat-treated scaffolds was 6. 67 ± 0. 7 MPa compared to 8. 54 ± 1. 2 MPa of the untreated group, while the elongation at break decreased from 5. 21 ± 1. 1 mm to 1. 97 ± 1. 3 mm. The Young modulus was 35. 8 ± 2. 7 compared to 44. 7 ± 3. 5 for the other group, as shown in Figure 6. Figure 6 Stress/strain curves of treated and untreated chitosan/polyvinyl alcohol nanofibrous scaffolds. Cs/PVA: chitosan/polyvinyl alcohol Physical analysis Contact Angle, Swelling Ratio, and Degradation Rate of CS/PVA NFS The mean contact angle of CS/PVA NFS immediately after drop stabilization was 72. 3° ± 1. 9°, and the angle decreased to 54. 3° ± 0. 7° and 29. 5° ± 0. 7° after 30 and 60 seconds, respectively, as shown in Figure 7 and Table 2. Figure 7 Photographic images of the water droplets on a surface and the calculation of contact angles of the chitosan/polyvinyl alcohol nanofibrous scaffolds. (A) The instrument used for water dropping. (B) At one second. (C) At 30 seconds. (D) At 60 seconds. (E) Contact angle measured at one second. (F) Contact angle measured at 30 seconds. (G) Contact angle measured at 60 seconds. Table 2 The raw data of the contact angles of the three samples of chitosan/polyvinyl alcohol nanofibrous scaffolds along with the mean values and standard deviations. Samples 1 second 30 seconds 60 seconds First sample 70. 3° 53. 3° 28. 7° Second sample 75. 1° 55. 4° 30. 6° Third sample 71. 5° 54. 3° 29. 2° Mean 72. 3° 54. 3° 29. 5° Standard deviation 1. 9° 0. 7° 0. 7° The mean swelling ratio of CS/PVA NFS was 89% ± 6. 8 on the first day, and the ratio increased on the second and third day to 159. 4% ± 12. 72 and 225. 4% ± 17. 44, respectively (Figure 8 ). No further swelling occurred in the following days. Figure 8 Swelling ratio of the chitosan/polyvinyl alcohol nanofibrous scaffolds within three days. (A) Assessment of swelling ratio in each of the five samples. (B) The average values of the five samples on each assessment day. Cs/PVA: chitosan/polyvinyl alcohol The mean degradation rate of the CS/PVA NFS in PBS was 0. 5% ± 0. 02 on the first day. The degradation rate increased gradually over time to 2. 04% ± 0. 41, 10. 24% ± 0. 85, and 26. 85% ± 2. 89 on the 7th, 16th, and 30th days, respectively, as shown in Figure 9 and Table 3. Figure 9 Degradation over time of the chitosan/polyvinyl alcohol nanofibrous scaffolds. Cs/PVA: chitosan/polyvinyl alcohol Table 3 The degradation values (mean percentage of weight loss) of the chitosan/polyvinyl alcohol nanofibrous scaffolds over 30 days of observation. Days Mean percentage of weight loss Standard deviation Day 1 0. 50% 0. 02 Day 3 0. 50% 0. 04 Day 7 2. 04% 0. 41 Day 12 5. 63% 0. 97 Day 16 10. 24% 0. 85 Day 21 15. 17% 1. 25 Day 25 20. 63% 2. 62 Day 30 26. 85% 2. 89 Biological analysis Protein Adsorption and Cell Viability The results of the in vitro protein adsorption test showed that 60% of the protein was adsorbed onto the CS/PVA NFS on the first and second days, and a 54. 7% adsorption ratio was observed on the third day, as shown in Table 4. Table 4 Protein adsorption profile of chitosan/polyvinyl alcohol nanofibrous scaffolds over three days of observation. Time Initial protein concentration Free protein (mg/mL) Adsorbed protein (mg/mL) Adsorption (%) Day 1 10 mg/mL 4 6 ± 2. 0 60% Day 2 4. 01 5. 99 ± 0. 14 60% Day 3 4. 53 5. 47 ± 2. 35 54. 7% The proportion of cell viability of the hAD-MSCs on the CS/PVA NFS through the MTT assay was 98. 89% ± 14. 67 on the first day, 118. 94% ± 1. 08 on the second day, and 105. 11% ± 3. 64 on the third day, as shown in Figure 10. Moreover, cells were successfully attached and proliferated on these scaffolds, as shown in Figure 11. Figure 10 Cell viability of human adipose tissue-derived mesenchymal stem cells over 72 hours of observation. MTT: 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide Figure 11 Field emission-scanning electron microscope images for the cultured human adipose tissue-derived mesenchymal stem cells on the chitosan/polyvinyl alcohol nanofibrous scaffolds. (A-C) At 5, 000× magnification. (D) At 20, 000× magnification. Discussion To confirm the existence of both chitosan and PVA, the FTIR was carried out. The FTIR analysis revealed that chitosan and PVA were homogenously blended and successfully electrospun together, as evident by the presence of characteristic peaks of both polymers, and the shift of the peaks and/or the change in the intensity of their absorption peaks after electrospinning and heat treatment were due to increase in hydrogen bonding between chitosan and PVA [ 41 ]. One of the advantages of using electrospinning to produce nanofibers is the close similarity of these fibers to the fibers of ECM concerning shape, diameter, orientation, and porosity. Moreover, the nano-sized diameter of the produced fibers led to an increased surface area-to-volume ratio. These factors would enhance the biological and cellular responses [ 13 ]. The randomly oriented smooth nanofibers observed by the FE-SEM micrographs with their average diameter of 172. 7 ± 56. 8 nm and pore size of 0. 54 ± 0. 17 µm are similar to those of ECM along with the increase in surface area, at least theoretically supporting protein adsorption and cell adhesion, proliferation, and differentiation. Porosity can be defined as the volume occupied by voids of the total volume of the nanofibrous scaffold and can be obtained using the fluid displacement method; ethanol was chosen as a displacing solution instead of water as the latter caused swelling of the fibers. The optimal porosity for cell penetration is 60-90% [ 42 ]. The porosity of the fabricated scaffolds was within this range (74. 8% ± 3. 3); the randomly oriented fibers, along with layered deposition of the fibers during electrospinning, created interconnected pores which would help cells to communicate with each other and create channels and passages for the new blood vessels and oxygen, nutrient, and exudate exchange [ 42 ]. In this study, the tensile property of CS/PVA NFS decreased with heat treatment. The decrease in mechanical properties in the group subjected to heat treatment was due to loss of water and increased crystallinity, thus decreasing the elasticity of the fibers [ 43 ]. In general, the overall decrease in the mechanical properties of CS/PVA NFS is due to the increased content of chitosan and randomly oriented fibers [ 38 ]. Nonetheless, these scaffolds can be used in non-load-bearing areas. Contact angle tests can be used to evaluate the surface wettability and hydrophilicity of nanofibrous mats, which are crucial for protein adsorption and enhanced cell attachment, proliferation, and differentiation within the matrices of synthesized biomaterials. Scaffolds that possess high surface energy have better wettability and hydrophilicity than those with lower surface energy. In general, a surface with a WCA of more than 90° is considered hydrophobic, while those with WCA lower than 90° are hydrophilic [ 27 ]. Moreover, cell attachment on the mats is highly consistent with higher surface energy and WCAs of less than 75° [ 44 ]. In our study, CS/PVA scaffolds show a great hydrophilic nature and good wettability over time. The initial contact angle was 72. 3° ± 1. 9° immediately after drop stabilization, which decreased to 54. 3° ± 0. 7° and 29. 5° ± 0. 7° after 30 and 60 seconds, respectively. This could be attributed to the presence of amino and hydroxyl groups on the CS/PVA scaffold’s surface. Studying the swelling behavior of biomedical scaffolds is important as it indicates their structural stability in an aqueous environment and their capability to allow cells, body fluids, and proteins to penetrate them [ 36 ]. The swelling ratio of CS/PVA nanofibrous mats increased over time from 89% ± 6. 8 on the first day to 229% ± 21. 68 on the third day and became stable after that. This could be attributed to the hydrophilic nature of the scaffold, as indicated by the contact angle, as well as to the higher porosity that led to an increased swelling ratio [ 39 ]. The in-vitro degradation rate of the heat-treated CS/PVA scaffolds was investigated in PBS for 30 days. Weight loss increased with time, and the scaffolds lost 26. 9% ± 2. 9 of their original weight by 30 days; the increased chitosan content could explain this in this scaffold and increased hydrogen bonding between chitosan and PVA during thermal treatment [ 41 ]. Protein adsorption onto fabricated nanofibrous scaffolds is key to improving cell adhesion and attachment and regulating cellular response [ 39 ]. The results of the in-vitro protein adsorption study on the fabricated CS/PVA scaffolds on days one, two, and three were discussed in the Results section. From this study, it can be observed that protein was adsorbed successfully onto the scaffolds and achieved equilibrium, and all attachment sites were filled quickly from the first day. This could be attributed to the higher positively charged chitosan content in our scaffold, which could bind more negatively charged proteins, the increased wettability, and the nanoporous nature of the CS/PVA scaffolds that led to the increased surface area [ 40 ]. A cell viability study is crucial for any scaffold to be applied in tissue engineering [ 45 ]. The MTT assay performed to determine the proliferation rate of AD-MSC on the CS/PVA NFS showed a steady increase in cell viability over time which means that CS/PVA scaffolds supported cell attachment and proliferation, and no signs of cytotoxicity were observed. The morphology of the AD-MSC on the CS/PVA scaffold was observed on day three by FE-SEM. Cells spread well on the scaffold’s surface with a homogenous, flat pattern and cytoplasmic extensions (filopodia) of cells interacting with the scaffold. It is worth noting that the fabricated scaffolds retained their architecture within the culture media during the experiment. Higher magnification of the AD-MSC morphology provides a clear view of the importance to fabricate the scaffolds in a manner that resembles the ECM nanostructure of the tissues as it provides micropores, interlacing fibers, and rough sites for the cells to implant and extends their extensions to attach and interact with the surroundings. This could be attributed to the nanofibrous architecture that mimics the ECM and provides rough sites for cell adhesion, fiber’s diameter, increased porosity, and appropriate pore size, which provide mechanical interlocking, cell-to-cell communication, and the high surface area along with the biocompatibility, hydrophilic properties, and non-toxicity of the chitosan and PVA, which all favor cell implantation, attachment, proliferation, and migration [ 12, 45 ]. Our findings were in line with many researchers who used at least chitosan in their research and different cells such as mouse osteoblast cells (MC3T3E1) [ 45 ] and L-929 fibroblast cells [ 12 ]. On the other hand, Agrawal and Pramanik reported in their in-vitro biocompatibility study that mesenchymal stem cells did not attach to the fabricated PVA/CS scaffolds and chitosan/silk fibroin/polyethylene oxide scaffolds and attributed that to inadequate growth factors within the culture media and/or to remaining acetic acid within the scaffolds which turned the formulation more acidic and unsupportive for cell growth and proliferation [ 38 ]. Limitations of the study PVA is readily soluble in water and needs to be stabilized. In this study, heat treatment was used as an eco-friendly method. The current analysis did not evaluate the effect of different temperatures and treatment times to achieve the optimum conditions for scaffold stabilization. In addition, this work did not include a thermogravimetric analysis which would have given us detailed information regarding the physical and chemical behavior of the scaffolds during the thermal change over time. The degradability of CS/PVA NFS was tested in PBS only (pH = 7. 4) and not in acidic and alkaline media. Conclusions This study showed the possibility of preparing CS/PVA NFS by electrospinning. Chitosan was successfully blended and incorporated into the nanofibers. The scaffolds proved stable in an aqueous medium by thermal treatment with good wettability. Reasonable swelling ratios, degradation rates, and protein adsorption were obtained. Cultured cells attached and proliferated well on the nanofibrous mat with no signs of cytotoxicity, along with a good capability to adsorb proteins. The potential of these scaffolds to be used in tissue engineering applications in the oral cavity is evident.

10. 7759/cureus. 30027

Cureus

Smart Pacemaker: A Review

Since the first pacemaker was implanted, nearly 60 years have passed. Since then, pacemaker technology has made major advancements that have increased both its safety and effectiveness in treating people with bradyarrhythmias. The repeated stimulation of cells in specialized "pacemaker" regions of the mammalian heart and the transmission of stimulus via the ventricles serve as evidence that the electrical function of the mammalian heart is necessary for a regular mechanical (pump) role. The development of action potentials in individual cardiac cells is linked to myocardial electrical activity and the heart's regular cooperative electrical functioning. A container or pulse initiator that houses the battery and electronics, as well as lines that connect to the myocardium to deliver a depolarizing pulse and detect intrinsic cardiac stimulation, are all parts of a pacemaker. Defibrillators could be used with artificial hearts that have electrical pacemakers integrated into them in order to treat arrhythmia, heart failure, and cardiac arrest. Modern pacemakers have units for supporting patients with other disorders like "heart failure, " which happens when the heart does not pump as forcefully as it should. While many pacemakers are effective in treating different types of arrhythmias (irregular heartbeats), they also have units for treating them.

Introduction and background Despite significant advancements and the development of novel therapies, cardiovascular illnesses have remained the world's leading cause of death and morbidity for a decade [ 1 ]. Several treatments, including cell-based therapies, have been shouldered in the last 20 years; however, penurious subsister and injection of relocated cells in the ischemic environment of cardiac tissue have restricted their clinical usefulness [ 2 ]. One of the most pressing issues in the industry is the capacity to track the fate of modified tissue and its impact on the nursed organ after transplantation. Furthermore, when it is feasible to observe the activity of the implanted tissue, a tool to intercede in the therapy's outcome without the obligation for additional surgical mediation or ongoing medical attentiveness would prove to be highly advantageous to the therapy's effectiveness [ 3, 4 ]. It has been almost six decades since the initial pacemaker was placed. Since then, pacemaker technology has advanced significantly, boosting its safety and potency in treating individuals with bradyarrhythmias. Despite advancements, pacemaker therapy continues to be analogous with considerable peri- and post-procedural problems [ 5 ]. Leadless pacemaker therapy is a revolutionary method that addresses lead and pocket-associated issues in standard transvenous pacemaker treatment and was recently launched in clinical trials. These leadless pacemakers are independent, single-chamber suitable ventricular pacemakers inserted by a femoral percutaneous route [ 6 ]. Review Physiology The heart is a metrical electromechanical pump whose operation relies on the origination and transmission of action potentials, tailed by relaxation and a phase of refractoriness until a succeeding stimulus is produced. Inward (Na+ and Ca2+) and outward (K+) impulse-transporting ion channels are consecutively switched on and off during myocardial action potentials. Action potential waveforms vary in separate parts of the heart due to variances in Na+, Ca2+, and K+ channel articulation. These variations add up to a typical, unidirectional impulse circulation and the creation of usual cardiac rhythms [ 7 ]. The electrical function of the mammalian heart is required for a regular mechanical (pump) role, as evidenced by the successive stimulation of cells in specialized "pacemaker" areas of the heart and the transmission of stimulus via the ventricles. The creation of action potentials in discrete cardiac cells is attributed to myocardial electrical venture, and normal cooperated electrical working of the whole heart. Changes in channel function caused by inherited or acquired illness impact the action potential repolarization and can result in life-threatening arrhythmias [ 7 ]. Tissue engineering When the heart cannot keep up, technology comes up with a solution: an artificial pacemaker, a medical apparatus that provides electrical activity to keep the heartbeat stable. Artificial pacemakers have been used to treat various heart diseases that cause them to beat abnormally since the late 1950s, and they come with varying degrees of programmability. Pacemakers comprise a pulse initiator or can accommodate the battery and electronics, as well as lines that pass from the can to the myocardium to impart a depolarizing pulse and ascertain intrinsic cardiac stimulus. Insulation materials separate the conductor cables and the lead tip electrodes. The leads might be concentric (a tube within a tube) or co-radial, depending on the relationship between the wires (side-by-side coils). Active (with an electrically active helix at its tip for mechanical strength) or passive lead attachment to the myocardium (electrically inert tines harbor the lead). Short-circuiting causes high impedance (fracture) or low impedance (insulation breach) depending on how conductor elements and insulation materials are disrupted. When a potential difference (voltage) is enforced between the two electrodes, the pacing starts to ensue [ 8, 9 ]. The devices are made entirely of FDA-approved biocompatible materials; nevertheless, they are non-biodegradable and will persist in the patient's body for the rest of their lives until they are removed. Although not always harmful, there are times when the existence of old-fashioned electrical implants within the tissue is dispensable. If these non-biodegradable devices become a regular place, they could represent a danger to the patient's life [ 3, 10 ]. Pacemakers today use activity sensors (also known as accelerometers) to measure the patient's movement. However, these sensors are frequently insufficient to deliver real-time automated heart rate variation to the wearer. It might fluctuate depending on the rate of respiration; for example, current pacemakers are not sensitive enough to monitor and process specific brain activities related to heart rate control. As a result, even if the patient is cycling, the pacemaker indicates that they are at rest since precise signals are not refined. When a neuronal circuit at the base of the brain depreciates due to age or disease, it fails to provide the correct signals for the heart to pump correctly [ 11 ]. Hybrid technique Substituted hearts with built-in electrical pacemakers could be employed as a defibrillator in the event of arrhythmia or failure, as well as in the event of cardiac arrest. Of course, they would be merged accompanied by a network of electrodes that can provide a complete account of the organ's condition [ 3, 12 ]. Tian and colleagues provided the first example of a hybrid technique. A planar, slim, penetrable, and electronic mesh with amalgamated silicon nanowire field-effect transistors was created via lithography (nanoFETs). By removing an underlying sacrificial layer, this apparatus was subsequently removed from the silicon wafer and transformed into a discrete device. The mesh was designed to be highly penetrable, and its thin thickness of 2mm made it incredibly pliable and convenient to handle in 3D as shown in Figure 1. Afterward, the nanoelectronic webbing was embedded in a leveled electrospun biomaterial fiber mat and employed as a hybrid setting for sowed neurons and cardiomyocytes because the hybrid material is so pliable that it might be rolled and folded into a broader 3D-designed tissue following cell seeding. The research proved the production of engineered cardiac tissue in which the electronics had a negligible impact on its organization, as well as the scaffold's ability to observe the tissue's function from within a spatiotemporal manner [ 13, 14 ]. Figure 1 Hybrid technique [ 3 ] Two present-day examples have revealed the use of electrospun fiber scaffolds as the membrane and dielectric of the electronic net in order to construct a hybrid tissue made of biodegradable electronics [ 7 ]. The synthetic polymer polyurethane was utilized in one example (Lee et al. 2019) [ 15 ], whereas the innate protein albumin was used in the other (Feiner et al. , 2018) [ 16 ]. Both studies used biodegradable substances that can be used as a podium for tissue engineering and can be electrospun into fiber configuration to create flexible and elastic scaffolds. It was viable to build a mesh that functions as an electronic unit and a scaffold for modified tissue by evaporating metal electrodes onto electrospun fiber scaffolds through a shadow mask. Cardiomyocytes planted on these pieces of equipment are organized into utilitarian heart tissue just like their perfect counterparts. It was feasible to document extracellular potentials from the confines of the created tissues, activate them and deliver anti-inflammatory medications in a regulated approach using evaporated gold electrodes [ 3, 17, 18 ]. The Micra Transcatheter Pacing System and the Nanostim Leadless Cardiac Pacemaker (L. C. P. ; St. Jude Medical) are now leadless pacing systems in clinical use (T. P. S. , Medtronic). Both the devices are self-sufficient and can single chamber pace sensing, and rate reaction supply The Nanostim L. C. P got the C. E. mark in October 2013 [ 6 ]. However, Micra Transcatheter Pacing System now comes in an AV version with dual chamber sensing and RV pacing. The Nanotism Leadless Cardiac Pacemaker has been replaced by the AVIRA Leadless system. Cardiac resynchronization therapy In patients with left ventricular (LV) dysfunction, choosing the suitable implanted cardioverter defibrillator (ICD) or cardiac resynchronization treatment (CRT) device might be difficult [ 19 ]. ACC/AHA Guidelines for conduction system disorders and another set of ACC/AHA Guidelines for the management of patients with ventricular arrhythmia can assist doctors in order to select the best gadget for the patient. Certain patients gain from cardiac resynchronization therapy (CRT) with advanced congestive heart failure. The Left Ventricular (LV) Pacing Lead Implantation Procedure Using the Overlay Ref Technique may be facilitated [ 20 ]. A depiction of CRT is shown in Figure 2. An innovative electromagnetic navigation system that shows the real-time 3D location of delivery instruments with implanted sensors circumfused on previously recorded X-ray cine-loops of coronary sinus venograms has been created to aid in the installation of left ventricular (LV) leads. When the new guiding system is used, it is possible to implant CRTs safely, successfully, and with substantially less radiation exposure [ 21 ]. Figure 2 Working of CRT [ 22 ] Despite the promising results of CRT in bradycardia patients, not all patients experiencing RV pacing develop LV dysfunction; some may be impervious to the pacing-induced systolic dyssynchrony. Because CRT is expensive and has a high rate of complications, not every patient should receive it. It is crucial to pick people susceptible to systolic dyssynchrony from repeated ventricular pacing in bradycardia. Even though it might be ideal to find baseline indicators of pacing-induced systolic dyssynchrony, there is not enough information to guide patient selection [ 23 ]. Another recent problem is the CRT-induced pro-arrhythmia that may be connected to the LV lead inside the epicardial scar. It is an uncommon but significant consequence that antiarrhythmic medications cannot treat. A clinical conundrum arises when LV pacing is turned off because HF may deteriorate. Catheter ablation can treat recurrent arrhythmias; however, patients still need additional treatment. According to clinical data, women benefit from CRT more than males, but fewer women compared to men participated in the CRT clinical trials. Increased hiring of female prospects might lower the rate of non-responders. Although CRT is not advised for patients with narrow QRS, one of the research of discrete patient data found that females showed favorable results with CRT-defibrillator at a lower QRS length than males, which emphasizes the need for gender-specific medicine [ 24 ]. CRT is a costly therapy, which makes reimbursement difficult [ 25 ]. The relatively high complication rate associated with CRT as a result of the intricate structure of the coronary vein is another problem that affects its utilization in clinical practice. Compared to ICD or RV pacing, implanting takes more knowledge, expertise, and training. Lead displacement and dislodgement, implantation dissection, as well as phrenic nerve activation are risks connected with the insertion of CRT It is not surprising that the benefits of CRT have been disputed during its 20-year history, and its long-term effectiveness has been questioned. Despite having clear indications for CRT, a significant portion of patients received ICDs. Additionally, almost one-fourth of HF patients had RV pacing with frequent ventricular pacing percentage installed. More instruction with medical care based on guidelines is necessary for both patient and physician groups to address the low participation rate. CRT is now used in patients with congenital heart disease with various HF phenotypes (systemic LV failure, RV failure, and single ventricle failure), hypertrophic cardiomyopathy, pulmonary hypertension causing RV failure, and HF with intact ejection fraction [ 23 ]. A multifaceted therapy proposition, such as AV junction ablation or pulmonary vein isolation in conjunction with CRT, might enhance the feedback to CRT with the aid of rate control because the efficacy of CRT is diminished in atrial fibrillation. Creating new technology is necessary to overcome barriers that prevent CRT from being widely used. The LV lead access route, such as the percutaneous subxiphoid approach or transventricular passage, must be improved in challenging patients. It has been shown safe and effective to implant CRT with the aid of an electromagnetic tracking system based on sensors [ 26 ]. Complications Since the transcatheter aortic valve replacement (TAVR) technology is applied to healthier and lower-risk groups, the implications and anticipations of procedure-related problems, together with the requirement of permanent pacemaker (PPM) installation, need to be determined [ 27 ]. Due to the modification of a conduction system that is already ill, researchers have revealed a greater prevalence [ 28, 29 ]. All patients had paravalvular aortic regurgitation after PHV implantation. Echocardiography suggested that the PHV stent frame may not be ideally positioned against the sick native valvular structures in the area of calcific nodules. Even though the initial improvement in left ventricular function and clinical status following consolation of the aortic valve blockage was unaffected by paravalvular aortic regurgitation, severe paravalvular aortic regurgitation may have a deteriorating effect on long-term clinical outcomes succeeding PHV implantation. Future developments in stent design, such as larger maximal stent diameters, may lessen the frequency and graveness of paravalvular aortic regurgitation [ 30, 31 ]. Other crucial factors to take into account with any valve-replacement procedure include issues such as postoperative stroke, valve degeneration, PVL, and conduction abnormalities. With combined rates of 2. 7 percent at 30 days and 4. 8 percent at prolonged follow-up, the occurrence of postoperative neurological events was less and appeared to correspond to the previously described prevalence of strokes for traditional AVR [ 32 ]. To more fully evaluate the protectiveness and effectiveness of SURD-AVR in comparison to the previously mentioned options, more trials comparing SURD-AVR to traditional AVR and TAVR with lasting results, a minimum deficit of follow-up, and randomization are required. Future prospects Although there have been considerable advancements in the domain, each variety of tissue under investigation has its own set of hurdles that must be addressed before it can be used in clinical settings. The idea of incorporating electronics into all types of synthetic tissues has a lot of potential for improving tissue development and function [ 3, 33 ]. While many modern pacemakers are effective in treating various types of arrhythmias (irregular heartbeats), they have units for supporting patients with other disorders such as "heart failure", which occurs when the heart does not pump as powerfully as it should [ 34, 35 ]. In pacemaker (PM) patients, high abidance to remote monitoring (RM) improves outcomes; yet adherence remains unsatisfactory without a bedside console newer-generation PMs using Bluetooth low-energy (BLE) technology can communicate straightaway with patient-controlled intelligent devices via an app [ 36, 37 ]. A smart pacemaker is a sophisticated device that can detect even the smallest irregularity in the functioning of the heart. These smart pacemakers then not only help in rectifying the abnormality in the heart and reestablishing normal body functions but also alert the patients and their physicians about the patient's cardiac health. Hence, allowing them to take appropriate measures for the same. This feature of smart pacemakers had helped bring down the mortality rate in today's world. Conclusions The sector of biomedical electronic installs has progressed from the times of rigid implantable pacemakers to micro- and nanoscale, delicate electronic webs having traits as small as single cells and mechanical characteristics compared to the softest tissues. We have been able to bypass the considerable organic-inorganic hurdle that customarily subsists in joining electronics, tissues, and organs thanks to rapid breakthroughs in the sphere of flexible and stretchable electronics. These advancements have made it possible to counteract the detrimental effects that a foreign implant may have on a tissue and its denial by the body. The electronics could be removed after the organ function has been reinstated, and there is no obligation for ongoing inspection or mediation. On the other hand, if the existence of the electronics does not impair tissue behavior, they may be left in place indefinitely to ameliorate the patient's standard of life. The evolution of these sorts of substitutions would not only assist minimized donor organ scarcity but would also lessen the necessity for persistent follow-ups and operations. For carefully chosen individuals with heart failure, cardiac resynchronization treatment (CRT) is a beneficial therapeutic. Recent research indicates that re-coordinating left ventricular dyssynchrony may not be how CRT delivers the majority, if not all, of its advantages. Other probable mechanisms of effect that may differ from patient to patient and over time include atrioventricular resynchronization, decreased mitral regurgitation, and avoidance of bradycardia. No one treatment target exists; hence it is unlikely that any one metric will be able to predict benefit with any degree of accuracy. The purpose of this article was to create awareness among the medical society about smart pacemakers.

10. 7759/cureus. 30097

Cureus

A Review of Corneal Blindness: Causes and Management

Corneal blindness refers to a group of eye disorders that change the corneal transparency, causing corneal scarring and blindness. The leading causes of corneal blindness include infectious causes, i. e. , due to bacteria, viruses, fungi, and protozoa. The most common predisposing factors are trauma, contact lens usage, or the use of steroid medications. The various other diseases included are trachoma, dry eye disease, keratoconus, ophthalmia neonatorum, and non-infectious uveitis. Various clinical modalities are used for treating corneal blindness, including organ transplantation. Organ donation is cumbersome as various ethical and other factors are involved. Hence the concept of eye banking was introduced to meet the increasing demand for donors of the cornea. The eye bank's role is harvesting, processing, and keeping a record of the cornea being transplanted and donated. Furthermore, various recent advancements have been made for lamellar keratoplasty surgeries, including bioengineered corneas to fulfil the need for the unavailability of donors for the cornea. Various specific health interventions have been implemented to reduce the prevalence of corneal blindness globally. For proper management of corneal blindness, we have three components that are needed to be taken care of: prevention of corneal blindness, appropriate treatment modalities, and providing adequate rehabilitation services to the patients. This review encompasses the main reasons for corneal blindness and the management and treatment modalities available for the patients. The terms cornea, corneal blindness, treatment, management, causes, and complications were used for the review article on PubMed.

Introduction and background Although most of the eye's focusing power comes from the cornea, the focus is fixed. By altering the lens' shape, accommodation, i. e. , the focussing of light to provide a clearer vision of close objects, can be achieved. The epidemiology of cornea is too extensive, including viral ocular disorders, which lead to corneal scarring and inflammatory conditions and finally lead to functional blindness [ 1 ]. Globally, the foremost reasons for blindness are uncorrected refractive error, glaucoma, and diabetic retinopathy. With increasing age, the number of people affected by vision loss also increases [ 2 ]. To cope with the rising cases, we should take the initiative to set up large-scale eye camps to address these patients. The government should develop various public health programmes and awareness programmes to deal with the current havoc created due to loss of vision targeting the senior population and neonates and children. Approximately 1. 4 million children have blindness globally, meaning they are more likely to be in a lower socio-economic class and suffer from socio-economic deprivation [ 3 ]. Review Anatomy of cornea When the cornea is touched, an involuntary reaction to close the eyelid occurs because the cornea possesses unmyelinated nerve endings that are subtle to feel, temperature, and chemicals. A healthy cornea has no need for or necessity for blood vessels since transparency is of utmost significance. The anterior-most part of the eye is the transparent structure forming the anterior one-third of the outer layer called the cornea [ 4 ]. The cornea comprises six layers: the corneal endothelium, Dua's layer, Bowman's membrane, corneal stroma, and corneal epithelium. The thickness of the cornea in an adult is 550 microns [ 4 ]. The main functions of the cornea are the following: to protect structures inside the eye, the structural barrier, and against environmental infections, and to contribute to the eye’s two-third refractive power [ 5 ]. It is constituted of two components: the cellular component and the acellular component. Collagen and glycosaminoglycans are included in acellular details, whereas endothelial cells, keratocytes and epithelial cells are included in the cellular part [ 4 ]. Causes of corneal blindness The various causes leading to corneal blindness are depicted in Figure 1. Figure 1 Causes of corneal blindness Image credit: Author Shivangi C. Tidke Bacterial Keratitis Bacterial keratitis is the most common type of infectious keratitis in most countries worldwide. The most common bacteria responsible for causing bacterial keratitis is coagulase-negative staphylococci followed by Staphylococcus aureus, Streptococci spp. , Pseudomonas aeruginosa and Enterobacteriaceae spp [ 6 ]. Fungal Keratitis There are various species of fungi that cause fungal keratitis eventually leading to corneal blindness. The most common fungi that are involved are Candida, Aspergillus, and Fusarium [ 7 ]. In India, the most commonly involved species is Aspergillus which is then followed by Fusarium [ 8, 9 ]. The prevalence of specific species of fungi depends upon various specific risk factors such as temperature, climatic conditions, and urbanization of that region. The patient-specific risk factors include trauma and contact lens use [ 10 ]. The pathophysiology behind fungal keratitis is a defect in the corneal epithelium which gives entry to the corneal stroma. This is the reason for the increased prevalence of this disease in patients with trauma [ 11 ]. Herpes Simplex Keratitis (HSK) The infection caused by the herpes simplex virus is another cause which progressively leads to blindness due to corneal involvement. Keratitis instigated by herpes simplex virus (HSV) type 1 is a primary reason for contagious blindness. Epithelial or dendritic keratitis is the most typical manifestation. With recurrent illness, herpes stromal keratitis can cause progressive corneal opacification and visual loss [ 12 ]. The mouth, genitalia, and eyes, however, are the most typical sites of infection in a person with a healthy immune system. Very young toddlers and, in rare cases, adults may also get brain infections. In affluent nations, HSV eye illnesses are the principal infectious reason for corneal blindness [ 13 ]. The three types of stromal keratitis caused by HSV are endothelial subtype, epithelial subtype, and stromal subtype. In disciform HSK, the DNA value of the HSV is reduced as compared to the dendritic HSK. The laboratory boosts confidence in identifying HSK subdivisions with the combination of the HSV Immunoglobulin A by the enzyme-linked immunosorbent assay (ELISA) by the use of tear samples along with the procedure of real-time polymerase chain reaction (PCR) [ 14 ]. Trachoma Currently, trachoma is the leading cause of preventable blindness globally. Chlamydia trachomatis bacterial infection is spread through sexual contact, causing chlamydia. It is the furthermost often stated microbial disease in the United States. It is the most prevalent sexually transmitted disease (STD) in the entire world. It produces trachoma, the most pervasive infectious factor behind blindness worldwide, an eye infection [ 15 ]. Dry Eye Disease It was previously known as Keratoconjunctivitis sicca, which was given by Henrik Sjogren. He also established the triad, which included joint pain, dryness of mouth and Keratoconjunctivitis sicca [ 16 ]. The patient presents with the symptoms of foreign body existence, itchy and gritty eyes accompanied by excess tears and blurred vision. Furthermore, when the disease is not treated, it leads to further worsening of conditions causing discomfort and eventually causing blindness. The most common affected population is the senior population; the disease prevalence is seen more in females as compared to males. The various causes of dry eye disease are ocular surface dysfunction, blink rate, autoimmune diseases and other disorders [ 17 ]. The risk factors that are mainly responsible for the condition are the female gender and the advancement of age. The hormonal imbalance in the females further aggravates the symptoms like the decrease of tear production significantly around 60 years, along with the effect on meibomian gland function and the density of goblet cells of the conjunctiva [ 16 ]. The assessment and diagnosis of the diseases are based upon the questionnaires that are specifically developed for dry eye diseases like symptom assessment in dry eye (SANDE) and ocular comfort index (OCI) [ 18 ]. Keratoconus Keratoconus is an illness that causes thinning of the cornea, eventually causing reduced visual acuity and irregular astigmatism. It is a bilaterally asymmetrical disease associated with ocular inflammation. The various risk factors associated with keratoconus are allergy, atopy, eye rubbing and various environmental and familial factors, which are mediated by immunoglobulin E (IgE) [ 19 ]. Ophthalmia Neonatorum This is the conjunctiva disease, a kind of conjunctivitis observed in neonates. This disease frequently spreads following a vaginal birth and is associated with severe side effects such as ocular perforation and ulceration, which may result in lifelong blindness. According to research, this eye condition is caused by a number of microbes, including Chlamydia trachomatis, N. gonorrhoea, infection of the virus, and bacteria from the skin and intestinal system [ 20 ]. Non-infectious Neonatorum It is an autoimmune disease, which is one of the leading reasons for blindness that can be prevented by various measures. Management and treatment of corneal blindness For evaluation of the infectious keratitis, the primary step is the sample collection for culture and direct microscopy using various stains. There might be a need for corneal biopsy for deeper infections [ 11 ]. The gold standard techniques for diagnosing the infectious cause leading to corneal blindness remain the same, that is, Gram staining and culture methods which give the results instantaneously [ 21 ]. For confirmation of the diagnosis, a PCR is used occasionally because of its high sensitivity. Fungal Keratitis The medical treatment currently available includes antifungal agents which are fungistatic that increase the duration of treatment for complete eradication of the causative agent. The drug of choice is topical natamycin 5% and the other topical antifungal agents that can be given are amphotericin B, which is used specifically for yeasts; voriconazole; and itraconazole. For the patients who do not respond to the given medical treatment, surgical interventions have been used for their treatment. The interventions include lamellar keratoplasty and therapeutic keratoplasty [ 11 ]. Bacterial Keratitis The treatment of choice for bacterial keratitis is the use of topical antibiotics. Fluoroquinolones are most commonly used. Even anti-collagenases and steroids are used for the treatment [ 22 ]. In the case of bacterial keratitis, topical antibiotics are the most preferred and primary stay treatment regimen. Even anti-collagenases and steroids can be used in bacterial keratitis [ 22 ]. Voriconazole is used for fungal keratitis, which is a newer generation triazole because of its tremendous ocular penetration [ 23 ]. The ultimate target for managing the above condition is to decrease inflammation and avoid further eye complications [ 24 ]. Dry Eye Disease Questionnaires like Fluorescin break up time or Ocular Surface Disease Index can be used for diagnosing dry eye disease along with Schirmer's test for detecting decreased tear production; the assessment can also be carried out by stains and by cytology to find out the ocular damage [ 25, 26 ]. Dry eye disease management can be done by keeping the ocular environment in control, such as avoiding prolonged exposure to digital devices, avoiding the dry atmosphere and using external protection like contact lenses such as silicon hydrogel and scleral lenses [ 26 ]. Even lipids can be used with velocity enhancers for the effective treatment of dry eye disease (DED) [ 27 ]. Keratoconus The early detection of keratoconus can be done by more frequent monitoring of the disease progression and performing the indicated interventions at the time, leading to improved patient outcomes so that the use of transplantation of cornea is reduced to a significant amount [ 19 ]. The advances for the diagnosis are the corneal biomechanics, various biomarkers like tear inflammatory cytokine or levels of matrix metalloproteinases in the tear immunoglobulin A or by using artificial intelligence [ 28 ]. Various treatment modalities are available for the management of keratoconus and for preventing corneal blindness. To prevent further disease progression, the mainstay clinical modality is corneal cross-linking. Various strategies and new molecules have been implicated in the scleral cross-linking. One of the more unknown molecules isolated from the Streptoverticillium sp. , named transglutaminase, does not need photoactivation, which makes the cornea stiffer without causing much damage to the underlying layers of the cornea. The others include numerous keratoplasty procedures, transplantation of the Bowman's layer, additive keratoplasty, and cellular therapies [ 28 ]. Keratoconus causes the fragmentation of the Bowman’s layer in the earliest phase of the disease [ 29 ]. The biomechanical support to the cornea is provided by the Bowman's layer, which is also responsible for the maintenance of its shape and keeping it sturdy. Hence, if we replace this tissue, we can stop the further progression of the keratoconus to its later stages and prevent blindness [ 30 ]. A femtosecond laser was employed to produce the stromal compartment, which decreased the risk of micro-perforation during manual dissection. Recently, intraoperative optical coherence tomography has allowed for better visibility of the dissection plane [ 31 ]. The graft is localised at the mid-stromal level in the traditional approach; however, a recent variation describes inserting the graft as an onlay in the subepithelial region [ 32 ]. The additive keratoplasty helps to increase the thickness of the cornea along with biomechanical stability. To avoid immune rejection, an approach towards tissue engineering has been preferred, which enabled better outcomes in the case of visual acuity and in terms of biomechanical effects [ 28 ]. Diagnosis of non-infectious uveitis is based on clinical symptoms and the association with systemic diseases [ 33 ]. Corneal transplantation One of the treatment modalities that is used for treating corneal blindness is organ donation. The concept of corneal transplantation for the treatment of blindness was first stated by Himly in the year 1813, but the first transplantation surgery was actually performed by Von Hippel in the year 1886 by replacing the cornea of a rabbit [ 34 ]. Anterior corneal opacities were initially treated using lamellar keratoplasty, including the selective removal of layers of the cornea. This treatment modality was actually used to treat the disease keratoconus and also the scarring of the cornea, but it was halted since it didn't provide the best visual gain. This might have been because of the imperfect interface or any remaining opacities [ 35 ]. The various types of lamellar keratoplasty surgeries that can be done are anterior lamellar keratoplasty (ALK), superficial anterior lamellar keratoplasty (SALK), automated lamellar therapeutic keratoplasty (ALTK) and others. Immune Privilege The three barriers that contribute to the ocular immune privilege of the cornea are anatomical, cellular and molecular. The mentioned mechanisms facilitate immune tolerance to donor antigen. The predisposing factors destroying and interrupting the immune privilege are the following: the previous rejection of the graft, vascularized corneal tissue, and ocular inflammation. When the immune privilege is interrupted, it leads to corneal graft rejection. It is predominantly a cell-mediated pathway [ 36 ]. Recent Advancements Intra-operative Optical Coherence Tomography (iOCT) Continuous feedback on intraoperative surgical manoeuvres is provided by the iOCT. In Lamellar keratoplasty programmes like superficial anterior lamellar keratoplasty, automated lamellar therapeutic keratoplasty, deep anterior lamellar keratoplasty, Descemet stripping endothelial keratoplasty, and Descemet membrane endothelial keratoplasty, it is beneficial. The centre corneal thickness (CCT) of the donor and host corneas, both of which are significant criteria for choosing the blade size to be utilised in the microkeratome for dissection, may be measured using the iOCT. Additionally, it serves as an intraoperative directing aid when donor tissue is manually prepared so that the issues related to this will be further reduced. With the use of iOCT, appropriate coherence may be carried out in situations of superficial anterior lamellar keratoplasty and automated lamellar therapeutic keratoplasty. The iOCT directs every surgical step in the deep anterior lamellar keratoplasty (DALK) process, from the depth of trephination through graft-host apposition [ 37 ]. Femtosecond Laser-Assisted Lamellar Keratoplasty (FALK) The femtosecond laser can be used for laser and total thickness penetrating keratoplasty (PKP). This keratoplasty has various improvements when compared with the manual one [ 38 ]. Bioengineered Cornea This technique was developed for visual rehabilitation and for managing the gap in the availability of donors. Bioengineered cornea includes replacing the part of the cornea or the whole of the diseased cornea [ 39 ]. These include various methods ranging from the use of keratoprosthesis that actually supersedes the function of the cornea to the most recent advancement of tissue-engineered hydrogels, which assist in regenerating the tissue of the host [ 40 ]. Prevention Various public health interventions are needed to decrease the prevalence of corneal blindness. Measures like vitamin A supplements and advice regarding the modification of nutrition, i. e. , nutritional assessment and the vaccination against measles can prevent xerophthalmia, which is caused due to Vitamin A deficiency. Implementing the SAFE strategy, which includes Surgery for trichiasis, Antibiotics for infection, Facial cleanliness and Environmental improvement to control transmission, can successfully prevent trachoma, which is caused by Chlamydia trachomatis, causing corneal opacification leading to corneal blindness and decreasing its prevalence. Onchocerciasis can cause blindness due to inflammation caused by the Onchocerca volvulus, which can be controlled by the ivermectin distribution in public along with the control of the Simulium fly. In less developed and developing countries, traumatic corneal abrasion is the most common precipitating factor for blindness due to corneal involvement. Hence, to prevent such blindness due to trauma, a prophylactic topical antibiotic should be taken for a few days, especially in high-risk occupation people as farmers, who have an increased risk for trauma through vegetable matter [ 41 ]. Conclusions According to WHO, 1. 9 million people have corneal blindness due to the opacification of the cornea, which accounts for about 5% of the total patients who have blindness. The various conditions which progressively lead to corneal blindness are infectious such as herpes simplex keratitis, bacterial keratitis, fungal keratitis; glaucoma; trachoma; non-infectious uveitis; keratoconus; and dry eye disease. These conditions cause disruption and damage to the structural integrity of the cornea, which eventually leads to blindness and the disturbance of visual acuity. The management of corneal blindness depends on the precipitating disease that is leading to the blindness. One of the significant clinical modalities that are preferred is corneal transplantation. Hence setting up eye banks all over the country is the need of the hour to decrease the prevalence of corneal blindness. Other than corneal transplantation, conservative therapy is given, like avoiding the risk factors that are leading to the specific disease or opting for modified lifestyle modalities. Topical antibiotics and steroids are also provided for treatment purposes in case of infectious disease. Some of the recent advancements in the transplantation of cornea are iOCT, FALK, and the bioengineered cornea. The development of the technique of bioengineered cornea is helpful in coping with the increasing demand for cornea. Therefore, to reduce the prevalence of preventable blindness of cornea, proper measures should be taken and health interventions should be implemented globally.

10. 7759/cureus. 30116

Cureus

Recent Advancements in In Vitro Fertilisation

The field of assisted reproductive technologies has witnessed many new developments over the past 10 years. This review examines new stimulation techniques that might increase the number of fully developed oocytes derived during the in vitro fertilisation (IVF) cycle in addition to strategies for enhancing oocyte quality in older women. Before moving on to several fresh methods for determining endometrial receptivity, we talk about how preimplantation genetic screening (PGS) is currently being utilised. The main goal of this review is to highlight technological fields that might be debatable or are still sufficiently novel to require rigorous controlled trials for recognition. The use of IVF has been on the rise recently, mostly as a result of deferred childbearing, and there is no reason to believe that this trend will alter. Infertility therapies have advanced significantly thanks to the methods and techniques that were established via studies on animals and, more recently, people. Some technical discoveries in reproductive medicine have had a significant impact on innovations and treatment choices in other fields of medicine as well. The objective of this succinct review article is to quickly summarise and explain the advancements made in this intriguing area of medicine over the past 40 years.

Introduction and background A higher yield of mature oocytes during an in vitro fertilisation (IVF) cycle may be obtained by using novel stimulation techniques. Additionally, we are interested in devising techniques to raise oocyte quality, particularly in older women. In fact, according to recent projections, assisted reproductive technologies (ART) may keep 400 million people (3% of the world's population) alive by the year 2100 [ 1 ]. Treatments must be both secure and efficient as a result. IVF research will face new challenges in the future. The main difficulties are as follows: how to deal with the inevitable issues of egg ageing and female infertility, how to understand implantation problems and subsequently create remedies, and how to advance therapies for male infertility. Given the compelling motivations for caretakers, researchers, and most critically, infertile couples, only time will tell what chances and avenues the ensuing 40 years will offer for assisted reproduction. Since the world's first IVF baby was born around 40 years ago, it is estimated that over eight million infants have been born as a result of IVF infertility therapy. Before delving into some problematic new techniques for determining endometrial receptivity, we first explore the current debate around preimplantation genetic screening (PGS) [ 1 ]. Due to sociodemographic shifts and advances in technology, the demand for IVF has increased, changing how a significant portion of the population reproduces. The goal of this analysis is to emphasise the social and demographic factors that are fueling an increase in the demand for IVF on a global scale, in addition to providing an overview of emerging technologies that have the potential to considerably enhance IVF usage and lower its cost. Review Improving oocyte quality: role of mitochondria A woman's procreative capacity dramatically declines in the fourth decade of her life, which is directly tied to an ageing-related decline in oocyte quality and quantity [ 2 ]. After the age of 32, fecundity gradually decreases, and after the age of 38, it decreases fast [ 3 ]. Since the frequency of live births following oocyte donation in older women is proportional to the donor's age, oocyte quality is most likely the key factor causing a decrease in fecundity with age. Although greater DNA damage brought on by a less active DNA repair mechanism is a potential contributor to oocyte loss, the pathways leading to an increase in ovarian follicle loss in "aged" ovaries are yet unknown [ 4 ]. Chromosome aneuploidy increases in frequency when oocyte quality declines, primarily as a result of meiotic mistakes made during oocyte maturation. The oocyte must undergo nuclear, cytoplasmic, and epigenetic alterations in order to develop. These modifications all need energy, which the mitochondria provide by oxidative phosphorylation (OXPHOS) [ 5 ]. Mitochondria play an integral role in maintaining the quality of oocytes, the dysfunction of which is depicted in the flow chart given below [ 6 ] (Figure 1 ). Figure 1 Virulent cycle between mitochondria dysfunction and oxidative stress damage Mitochondria play an integral role in maintaining oocyte quality [ 6 ] Coenzyme Q10 supplementation Adenosine triphosphate (ATP) is made through an approach called OXPHOS, which involves five complexes that are positioned on the inner mitochondrial membrane [ 2 ]. The antioxidant properties of ubiquinone, also known as coenzyme Q10 (CoQ10), along with its capacity to control cellular redox and having an impact on several signalling pathways make it crucial in this process [ 7, 8 ]. After the age of 30 in humans, the majority of their tissues have lower CoQ10 concentrations [ 9, 10 ]. Due to its association with a decline in fertility and an increase in aneuploidies, CoQ10 loss may hasten the ageing process. In an elderly animal model, Ben-Meir et al. (2015) showed that supplementing with CoQ10 prevented the loss of ovarian reserve, enhanced mitochondrial function, and markedly decreased oocyte aneuploidy. In comparison to older animals on a placebo, these elderly mice stimulated produced more offspring and had more oocytes [ 2 ]. Afterwards, it was found that isolated CoQ deficiency brought on by conditional deletion of the PDSS-2 gene in young animal oocytes resulted in phenotypic changes resembling oocyte mitochondrial dysfunction linked to ageing [ 2 ]. If the animals received CoQ10, these alterations might be undone. It is reasonable to assume that CoQ10 supplementation would be beneficial to older women in a manner similar to how CoQ10 administration has been shown to improve reproductive outcomes in elderly animals [ 2 ]. Currently, a great deal of research is being conducted in this field. The ageing process is very different between mice and women because of the enormous differences in lifespan, despite the animal model's apparent promise. Given that giving CoQ10 to mice for 12-16 weeks is likely equivalent to years of human use, more clinical research is necessary [ 2 ]. Ultimately, we can say that CoQ10 is advantageous in IVF as it improves the oocyte quality with regard to ageing. The only disadvantages of this method are the side effects of CoQ10, which are very rare and include heartburn, nausea, diarrhoea, abdominal pain, fatigue, and dizziness. This procedure of supplementing CoQ10 can be accomplished by two methods: in vivo (Figure 2 ) and in vitro (Figure 3 ). In in vivo method, oral treatment methods are employed to enhance the oocyte quality. Whereas, in in vitro method, the culture media is supplemented with the enzyme, which can further be either standard culture or in vitro maturation culture [ 6 ]. Figure 2 In vivo method of CoQ10 supplementation CoQ10 is directly administered to the patient in the form of oral tablets [ 6 ] CoQ10: coenzyme Q10 Figure 3 In vitro method of CoQ10 supplementation CoQ10 is supplied to the culture media containing oocytes [ 6 ] CoQ10: coenzyme Q10 Mitochondrial transfer To combat ooplasmic ageing, subcellular oocyte modification has been used in other programmes. Cohen et al. performed ooplasmic transfers into mature oocytes of patients whose numerous IVF cycles had failed due to insufficient embryo development using donor oocytes. The findings indicated that the infants were alive and in good health [ 11 ]. Given that mitochondria are found in the cytoplasm, donor ooplasm's mitochondria were also present in recipient eggs, which was thought to be the primary factor influencing better development. Ooplasm donation is no longer practised because heteroplasmy testing on several healthy children revealed that their mitochondrial DNA (mtDNA) was derived from both the mother and the cytoplasm donor, indicating that the heteroplasmy was present in the oocytes [ 12 ]. But later research using autologous mitochondrial transfer has enhanced the earlier work with ooplasm transfer. The mitochondria are taken out of the patient's ovary's oocyte precursor cells in the superficial epithelial layer and injected into their oocytes during fertilisation in this cutting-edge procedure. It has been demonstrated that mitochondrial injection enhances embryo growth and aids in live births in women who had previously experienced poor embryo development [ 13 ]. There are some unethical demerits regarding the above-mentioned technique, such as germlines being modified during mitochondrial donation which leads to the passing down of such modifications to upcoming generations. This method also has the potential of psychological and emotional impact on the offspring leading to an effect on an individual's sense of identity. The efficacy of this method, like the existence of oocyte precursor cells, needs to be confirmed through appropriate randomised controlled trials. Medical advancements To enhance the production of oocytes available for IVF, controlled ovarian hyperstimulation, or COH is used. Numerous gonadotropin injections, visits to the fertility clinic, and transvaginal ultrasound examinations are required for COH. As a result, it takes a lot of time and effort to perform COH. The utilisation of transportable facilities and perhaps self-contained endovaginal telemonitoring might further simplify follicular and endometrial monitoring in light of recent improvements in portable, less expensive ultrasound systems [ 14 ]. Combining such methods would hasten and reduce COH intrusion. This method of treatment is known to have several demerits such as emotional stress, high economical costs, and lastly ovarian hyperstimulation syndrome (OHSS), which manifests as abdominal pain, nausea, vomiting, bloating, tenderness over the area of ovaries, diarrhea, and shortness of breath. Evaluation of patients for psychological issues during IVF is one therapy that may help lessen the burden of treatment even further, in addition to counselling and coping mechanisms like e-therapy [ 15, 16 ]. As per a randomised controlled trial conducted by van Dongen et al. , internet-based interventions carry great potential in relieving psychological distress, particularly when care is personalised to patients' personal risk profiles [ 17 ]. Technological advancements One could argue that automation and the miniaturisation of IVF laboratories are the two most promising technological advancements that have the potential to democratise access to IVF in the near future. An IVF facility's exorbitant costs, unequal access, and inconsistent operations are primarily the result of its manual hiring, construction, and operation. The fundamental steps conducted in the IVF laboratory are as follows. Firstly, there is determination and segregation of sperm and oocytes, followed by fertilisation and embryo culture. Next, we select embryo for transfer, and lastly, cryopreservation of surplus embryos and gametes is done. Significant progress has already been made toward automating each of these various stages with the help of novel techniques. However, the majority of the IVF process is still carried out manually. The revolutionary new IVF lab-on-a-chip concept has the potential to transform in vitro fertilisation by automating nearly all necessary steps in a single system [ 15, 16, 17 ]. In the multidisciplinary field of microfluidics, fluid dynamics is precisely controlled and manipulated under the influence of minute geometrical constraints that favour surface forces over volumetric counterparts. Earlier IVF laboratory procedures used macroscale methodologies to microscale cellular biological activities, despite their historical success [ 18 ]. At least four benefits that could result from integrating microfluidics into the IVF laboratory have been predicted. Firstly, fluidic gamete/embryo manipulations that are precisely controlled. The second involves creating biomimetic culture environments, while the third entails making microscale genetic and molecular bioassays easier. The last benefit involves allowing for miniaturisation and automation. On the contrary, it is difficult to standardise and scale up, which require external pumps and tubing, as well as connectors and valves to operate. Automated sperm analysers and microfluidic sperm-sorting equipment are frequently used in IVF procedures [ 19, 20, 21 ]. Using microfluidics, sperm and sperm-bearing tissue have been removed from testicular biopsies [ 22, 23, 24, 25, 26, 27, 28 ]. Even though the vast majority of IVF patients are candidates for conventional fertilisation, microfluidic in vitro insemination has been shown to be successful [ 29 ]. Potentially, intracytoplasmic sperm injection (ICSI) will not be necessary in the future thanks to microfluidic devices. ICSI has established itself as the de facto technique of insemination in human clinical IVF, demonstrating the significance of precise microfluidic push/pull cumulus-oocyte-complex cumulus cell removal in creating good visibility of the oocyte cytoplasm/orientation [ 30 ]. Despite its technological difficulties, the ICSI phase of fertilisation may be carried out effectively on a commercial scale [ 30 ]. In the future, automated ICSI is likely to be integrated with microfluidics, robotics, and high-tech optics [ 31, 32 ]. Scientific advancements Our understanding of the mechanisms that control folliculogenesis has continually improved as a result of research on fertility preservation [ 33 ]. The interactions between the oocytes and the somatic cells surrounding them, as well as the vital hormones and growth factors, have been revealed by follicular in vitro culture techniques. Recent developments in multi-step culture techniques have made it possible to activate, develop, and in vitro mature (IVM) ovarian cortex tissue primordia to produce metaphase II oocytes [ 34 ]. The advantages of IVM include reduced risk of OHSS and polycystic ovaries, lower medication costs, reduced stress, and lower monitoring burden. In contrast, it has been observed that chances of live birth with IVM are slightly lower than with IVF. Ovarian tissue cryopreservation and IVF have improved the chances of preserving fertility in prepubescent girls and adolescent women who are more likely to experience primary ovarian insufficiency (POI) from gonadotoxic chemotherapy for cancer or other serious illnesses. As long as some dormant follicles are still present in the ovarian cortex, fascinating developments in this technology may make it possible to isolate oocytes from females who have undergone POI or who have gone through natural menopause. An artificial ovary could be developed in a mouse model using scaffolds made through 3D printing for tissue engineering [ 35, 36 ]. Microfluidic culture techniques can be used to mimic the menstrual cycle by promoting follicle development [ 37 ]. Co-treatment with gonadotropins and letrozole in IVF Gonadotropin stimulation and the oral medication letrozole used during IVF cycles may be beneficial, especially for breast cancer women receiving fertility preservation treatment, according to recent research [ 38 - 41 ]. Letrozole and ovarian stimulation are employed in the treatment of breast cancer patients to reduce blood oestrogen levels. According to these studies, breast cancer patients who received letrozole and gonadotropins for the duration of the stimulation had lower estradiol concentrations than they would have anticipated but also had more mature oocytes available for cryopreservation than breast cancer-free controls who received conventional COH [ 41 ]. It has been associated with positive effects, including decreased gonadotropin doses that minimise the cost of IVF therapy and enhanced oocyte and mature oocyte counts while retaining the same pregnancy rate as conventional stimulation. On the other hand, gonadotropin stimulation may lead to OHSS, profound hypoestrogenemia, as well as more time-consuming and complex stimulation protocols. In 2005, the impact of letrozole on intraovarian testosterone levels and the success of IVF cycles was investigated. According to Garcia-Velasco et al. , the use of letrozole 2. 5 mg during the initial five days of gonadotropin stimulation significantly raised the levels of androstenedione and testosterone in follicular fluid and improved the success of IVF cycles. Letrozole considerably outperformed the control group in terms of both the number of recovered oocytes and the implantation rate. Pre-treatment with dehydroepiandrosterone/testosterone Numerous strategies have been investigated in an effort to raise intrafollicular androgen concentrations in people who do not respond well to medication because intraovarian androgens may significantly impact early follicular development. To increase the ovarian sensitivity to FSH and follicular response to gonadotrophin therapy in low-responder IVF patients, transdermal testosterone was utilised as a pre-treatment [ 42 ]. As the quantity of cumulus oocyte complexes grew, so did the frequency of clinical pregnancies and live deliveries. According to Gleicher et al. , individuals with low ovarian reserve received dehydroepiandrosterone (DHEA) for 30 to 120 days as a supplement (25 mg three times per day). They found that individuals who had received treatment had higher anti-mullerian hormone (AMH) levels and greater conception rates when compared to patients who had not received treatment. DHEA may lessen aneuploidy and miscarriage, according to the same study team's hypothesis [ 43, 44 ]. Wiser et al. carried out a prospective randomised controlled experiment to ascertain the effect of DHEA supplementation on the effectiveness of IVF in patients who had poor responses. They discovered that the DHEA group had much higher rates of live birth and higher-quality embryos compared to the controls. In both groups, the number of zygotes and eggs was the same. It is unknown whether DHEA helps older women or persons who respond slowly because not many randomised controlled studies have been conducted on these populations. New approaches to assess endometrial receptivity The endometrial receptivity array (ERA), a microarray study of implantation-associated gene expression, and ultrasound assessment of sub-endometrial wave frequency are novel techniques for evaluating endometrial receptivity in IVF facilities. Reproductive genetics IVF and reproductive genetics are long considered the industry's frontiers. Preimplantation genetic testing (PGT) of embryos to find chromosomal abnormalities has become more popular as a result of the introduction of next-generation sequencing. The use of PGT for monogenic disorders has expanded along with the popularity of infertile couple carrier screening. It has many advantages such as improved embryo selection, preventing transfer of embryos that will not implant, less time-consuming procedures, reduced costs, and lastly, it has a positive impact on psychological well-being. However, being an invasive procedure is one of its major disadvantages. Other demerits include a cycle with no transfer and embryo mosaicism. Future treatments for severe monogenic disorders may employ germline genome modification (GGM), thanks to advancements in micromanipulation methods and CRISPR-Cas9 gene-editing tools [ 45 ]. The UK is currently conducting clinical trials for mitochondrial replacement therapy (MRT), a more advanced treatment than GGM for heritable mtDNA problems [ 46 ]. Conclusions In the future, more and more people will use IVF, thereby changing the way a large portion of the human population reproduces. In the near future, IVF will likely be used in several regions worldwide to conceive up to 10% of all children. Given the rapid advancement of reproductive genetics and IVG science and technology, it is essential that regulatory organisations and the general public work together to develop a framework for evaluating the moral implications of emerging technologies. Emerging technologies should be incorporated into clinical practice with the help of a carefully planned clinical trial. The IVM technique reduces the risk of OHSS while enhancing patient safety. IVM may also be advantageous for patients who need to preserve their fertility or who do not respond well to other mentioned treatment modalities. It is now evident that technological advancement, the evolution of necessary tools, as well as the accumulation of experience and training among those performing the procedure have all contributed to success rates rising to as much as 56%. For women over the age of 35 years, this technique is feasible.

10. 7759/cureus. 30702

Cureus

Fibronectin and Its Applications in Dentistry and Periodontics: A Cell Behaviour Conditioner

The formation of biomaterials is a physical phenomenon that is primarily influenced by the material's chemical and physical characteristics, as well as by the availability of proteins and their mutual interactions. A common extracellular matrix (ECM) glycoprotein called fibronectin (FN) is a biomaterial that is essential for tissue repair. Cellular FN (cFN), also known as the "large external transformation sensitive (LETS) protein" or "galactoprotein, " was found during the quest for tumour markers twenty-five years ago and was later identified as the surface fibroblast antigen. Twenty different isoforms of the FN protein can be created by alternative splicing of a single pre-messenger ribonucleic acid (pre-mRNA) molecule. FN is an outstanding illustration of an ECM protein that intricately influences cell activity. FN is necessary for cell behaviours like cell adhesion, cell migration, and differentiation of cells as well as highly coordinated tissue processes like morphogenesis and wound repair. Plasma FN is absorbed by tissues and deposited in extracellular matrix fibrils along with locally generated cellular FN. cFN is produced by a wide range of cell types, including fibroblasts, endothelial cells, chondrocytes, synovial cells, and myocytes. FN and other cell adhesion proteins can promote cell attachment to tooth surfaces. Periodontal ligament (PDL) cell-ECM interactions, and consequently the regeneration of periodontal tissues, depends on FN. Specific FN segments serve as indicators of periodontal disease status and provide evidence for their potential involvement in the pathophysiology of the condition. FN is an all-purpose biomaterial that may be utilised for clinical applications ranging from tissue engineering to disease biology. Therefore, it would be desirable to develop materials that specifically bind to FN.

Introduction and background Protein adsorption from biological fluids, such as blood plasma, occurs shortly after a biomaterial has been implanted, mediating the interface between surface cells and the biomaterial. The composition of the adsorbed protein layer at the interface has a substantial impact on the nature of the tissue and material's fate, and it also has an impact on important aspects of cell response such as adhesion, spreading, migration, proliferation, and differentiation [ 1 ]. Protein availability and interactions between proteins that could lead to the adsorption of proteins that don't productively induce cells because of a deficient structure are two factors that drive the disorganized nature of the biomaterials process [ 2 ]. The selection of an appropriate biomaterial to provide a scaffold that replicates the native extracellular matrix and directs resident stem cells to restore functional tissue presents a significant challenge in tissue engineering (TE) [ 3 ]. A common extracellular matrix (ECM) glycoprotein called fibronectin (FN) is a biomaterial that is essential for tissue repair. The plasma form of FN circulates in the blood and, in response to tissue damage, it is absorbed into fibrin clots to influence the platelet function and mediates hemostasis [ 4 ]. FN promotes cell-ECM interactions during vital processes like development, wound healing, fibrosis, and tumour progression [ 5 ]. Review History of FN In vivo, there are two distinct groups of FNs that have been identified as a result of quite diverse research projects. During post-World War II studies on the fractionation of human blood plasma, plasma FN (pFN), which is primarily found as a soluble dimer circulating in distinct body fluids, was identified as the first cold-insoluble globulin. Cellular FN (cFN), formerly known as the "large external transformation sensitive (LETS) protein" or "galactoprotein, " was found during the quest for tumour markers 25 years ago. Later, it was identified as the surface fibroblast antigen [ 6 ]. By the middle of the 1970s, Vaheri and his colleagues united and chose a common name for the protein: fibronectin, which combines the Latin words fibra, which means fibre, and nectere, which means to bind or link. Reconsideration of the structural glycoproteins and ECM was driven by the discovery of FN. Most biologists had dismissed the ECM up to that point as a dull collection of inert, meaningless molecules [ 7 ]. Structure and biological activity of FN In humans, FN has more than 20 distinct alternative splicing isoforms. Cold-insoluble globulin (CIG), or the physically complicated dimeric high-molecular-weight glycoprotein FN, is composed of two nearly identical subunits (250 kiloDalton) that are covalently connected by disulfide bonds adjacent to their C-terminal region [ 1 ]. The interactions between cells and the extracellular milieu that surrounds them determine how they behave. Cell receptors and integrins are primarily solicited because cells are highly sensitive to the mechanical and bio-physicochemical features of this ECM environment. Cells literally "feel" the distinctiveness of their surroundings through the integrins and react to important factors including mechanical characteristics. For instance, a lack of stiffness, such as that found in an extremely soft environment, impairs cell adhesion and, as a result, results in a lack of intracellular tension and an ineffective intrinsic signalling system [ 8 ]. By alternative splicing, 20 distinct isoforms of the FN protein can be produced from a single pre-mRNA molecule. FN domains contain binding sites for ECM substances like collagen, heparin, fibrin, and other FN molecules, as well as for cell binding via integrin receptors. FN is secreted in both its more soluble pFN and less soluble cFN forms [ 9 ]. Figure 1 shows how the structure of FN is bounded to integrin at the surface of the cell. FN fibrils exhibit significant elastic properties. Cells have the ability to stretch FN fibrils up to four times their relaxed length. The three repeats of FN-I, FN-II, and FN-III make up the majority of its multimodular structure, which gives the material its mechanical responses. Disulfide bonds are used to cross-link the beta-strands of type I and type II modules; type III modules do not include any disulfide bonds. The seven-beta strand sandwich motif, which is seen in many mammalian proteins, is present in the structural features of FN-III modules. It has been proposed that the unfolding of individual FN-III modules gives the FN fibrils their flexibility. Furthermore, by stretching FN-III modules, nucleation sites for the assembly of FN into fibrils that were previously concealed might be seen. The most plausible locations for these hidden binding sites, also known as cryptic sites, are the FN-III cores or the spaces between two close FN-III modules' hinge regions. FN-III modules contain extremely little sequence homology, with a typical sequence identity of under 20%, but having very comparable tertiary structures. The simulation results show that FN-III modules can be pre-stretched before encountering the main unfolding barrier by making only a few small adjustments to their tertiary structures [ 1, 8, 9 ]. Figure 1 The cells bind and exert forces on fibronectin through transmembrane receptor proteins of the integrin family, which mechanically couple the actin cytoskeleton to the ECM via an elaborate adhesion complex structure of fibronectin bounded to integrin at the cell surface ECM: extracellular matrix; FN: fibronectin FN as Cell Behaviour Conditioner FN is an outstanding illustration of an ECM protein that intricately influences cell activity. It is a mediator in numerous cellular interactions with the ECM, possessing binding sites for integrins, other ECM proteins, growth factors, and itself. As a result, FN is essential for both highly coordinated tissue behaviours like morphogenesis and wound healing as well as cell behaviours like cell adhesion, migration, and differentiation. Growing data indicates that the selectivity of integrin engagement to this can lead to varied cell responses since FN can bind a variety of different integrins [ 10 ]. The first assembly of the basement membrane does not involve FN matrix assembly. This is because FN fibrils are one of the first ECM proteins to assemble during tissue development and wound healing. Additionally, FN molecules contain multiple domains that bind a variety of ECM proteins, growth factors, and small molecules [ 11 ]. Growth factors like platelet-derived growth factor (PDGF), Insulin-like growth factor (IGF-1), and transforming growth factor (TGF) are thought to be involved in periodontal regeneration and repair causing these cells to significantly upregulate the embryonic isoforms of fibronectin. The antigens that correspond to the fibronectin alternatively spliced variants do not reveal their functional activities, but the selective expression of these antigens suggests that they play very specific functions in periodontal regeneration and repair [ 12 ]. Properties of FN The ECM contains proteins called FN that have a variety of cell activities, including structural support and signalling signals for gene expression, cell contractility, migration, and differentiation. Table 1 shows a list of various properties of fibronectin [ 1, 4, 5 ]. Table 1 Important characteristic properties of fibronectin Properties of Fibronectin Cell migration Cell contractility Differentiation of cells Helps in growth factor signaling system Signaling cue for cell growth Possesses tissue-specific mechanical properties serving as both a molecular reservoir and a structural scaffold Types of FN pFN Disulphide (S-S) bond at the near C-terminus of pFN causes it to form a heterodimer that is made up of 230 kiloDalton and 235 kiloDalton chains [ 13 ]. Along with locally produced cellular FN, plasma FN is absorbed by the tissues and deposited in extracellular matrix fibrils. Given the plasma's abundance of FN, it has been thought that keeping plasma FN in a tight compact shape is essential to prevent abnormal interactions between FN protomers as well as between FN and other macromolecules and cell surface receptors [ 14 ]. cFN cFN, a stretched and insoluble isoform that affects both ECM homeostasis and overall ECM-cell interactions, is secreted and organised into dense complex fibril networks. Increased cellular connections and the formation of fibril networks are made possible by the FN molecule's stretching [ 15 ]. Numerous cell types, including fibroblasts, endothelial cells, chondrocytes, synovial cells, and myocytes, develop cFN [ 16 ]. FN in wound healing conditions The formation of ECM and re-epithelialization during wound healing depends upon FN. FN performs several functions in the healing of wounds as a result of the existence of numerous function domains and binding sites in its structure. It interacts with cytokines, the ECM, and many cell types. ECM formation is FN's primary function [ 17 ]. The early stages of wound healing are when pFN is most crucial since this is when it binds to platelets and fibrin, giving the fibrin clot more strength. Additionally, the FN in this clot is essential for the cell adhesion, migration, and aggregation of platelets. cFN is essential for the growth of granulation tissue and is added to the wound bed by fibroblasts and endothelial cells in the healing process [ 4 ]. By promoting platelet aggregation and adhesion to the damaged endothelium surface, FN reinforces the clot by creating a fibrin-FN network, also known as the "fibrin-fibronectin provisional matrix" [ 18 ]. Applications of FN in clinical dentistry Because demineralization exposes collagen, to which FN binds, the role of FN in the attachment of cells to the tooth surface has received a lot of attention. FN and other cell adhesion proteins can make it easier for cells to attach to tooth surfaces [ 19 ]. FN can influence fibroblast attachment in the gingiva favourably in dental applications and implant operations, preventing inflammation-driven disintegration of the tissues around the implant [ 20 ]. It has been shown that covering titanium surfaces with FN and glow discharge plasma (GDP) increases the surface hydrophilicity, roughness, and ability to promote cell adhesion, migration, and proliferation [ 21 ]. To prevent dental epithelial cells formed from postnatal day-1 molars from adhering to an FN-coated dish, the cyclic arginine-glycine-aspartate (RGD) and a 1-integrin-neutralizing antibody were used for its prevention [ 22 ]. This suggests that FN-1 integrin interactions play a key role in dental epithelial-cell binding. FN production in dental pulp cells is stimulated by calcium ions produced from calcium hydroxide. By coming into contact with other cells, FN may stimulate the development of dental pulp cells into mineralized tissue-producing cells, which are the primary cells responsible for forming dentinal bridges [ 23 ]. With a less significant impact on epithelial cells, FN coating of implant surfaces increased gingival fibroblast binding by two to three times [ 24 ]. The FN-immobilized chitosan scaffolds may work well as three-dimensional substrates for the attachment and growth of dental pulp stem cells. To stimulate odontogenic differentiation in the case of dental regeneration, they need to be supported by the proper biosignals [ 25 ]. Titanium has a remarkable affinity for human-FN (HFN) adsorption due to its hydrophilicity. Early cell adhesion, spreading, and proliferation are impacted by selective HFN adsorption [ 26 ]. The FN-conjugated, micro-grooved titanium (Ti) substrate is an effective surface to promote osteoblast formation and osteoblast marker gene expression in mesenchymal stem cells (MSCs) [ 27 ]. Root canal sealants showed biocompatibility and stimulated the expression of FN and tenascin (TN) [ 28 ]. Using a confocal laser scanning microscope and indirect immunofluorescence, it is possible to locate where FN was located in the dental pulp of developing and fully developed human teeth. Intense fluorescence was visible along the basement membrane in growing teeth that were exposed to the mesenchyme of Hertwig's epithelial sheath and freshly formed (mantle) predentine. The corkscrew-shaped FN fibres that crossed from the pulp into the predentine in a direction parallel to the odontoblasts' long axis could be detected between the cells as the odontoblasts becomes longer. In the odontoblast layer, fibronectin is present at every stage of dentogenesis. Von Korff fibres are fibrous structures between odontoblasts that are FN-positive. These structures are directly linked to dentinogenesis and odontoblast differentiation [ 1, 12, 13, 18 ]. Applications of FN in periodontics Periodontal ligament (PDL) cell-ECM interactions, and consequently the regeneration of periodontal tissues, depend on fibronectin [ 29 ]. Specific FN segments serve as indicators of periodontal disease status and provide evidence for their probable involvement in the pathophysiology of the condition [ 30 ]. Low quantities of FN fragments are present in healthy PDL sites. This is probably because of the tissue's high metabolic rate, which may result in a steady flow of ECM breakdown products in the form of fragments. Additionally, because FN is localised in the periodontal ECM, it can take part in tissue regeneration and wound healing. Type I and type III collagen fibrils in the PDL are coated with FN, and it also occasionally covers the gaps between neighbouring collagen fibrils and the cell membrane [ 29 ]. The fibrin glue system, a commercial product, includes FN, which promotes early wound attachment and healing [ 31 ]. A new approach to researching the impact of diabetes on periodontal disease is available through the efficient glycation of collagen type-I (COLI) and FN by methylglyoxal (MG) therapy. Protein glycation may have a role in the development of diabetes-related periodontal wound healing and the significant impact of glycated COLI and FN on human gingival fibroblasts (hGF) and human-periodontal ligament fibroblasts (hPDL) behaviour [ 32 ]. FN appears to have the potential to slow epithelial down growth, which in turn encourages root biomodification and regeneration. It is implicated in the attachment of gingival fibroblasts to root surfaces. The function of FN as a root biomodification agent in regeneration is discovered [ 33 ]. Citric acid and FN usage have the potential to encourage reattachment following periodontal therapy [ 34 ]. Faster healing results from surgery when FN and citric acid are used together. FN and citric acid together promote cellular proliferation [ 35 ]. The improved connective tissue attachment seen with tetracycline treatment alone appeared to be largely negated by the addition of fibronectin to tetracycline-treated roots [ 36 ]. When utilised as a regenerative therapy for intraosseous lesions in humans, enamel matrix derivative (EMD) and autologous fibrinogen or FN system (AFFS) combined with bovine-porous bone mineral (BPBM) had similar effects on decreasing periodontal probing pocket depth, clinical attachment loss, and defect fill [ 37 ]. The unique method of surface functionalization by FN deposition onto hydrolyzed polylactic-glycolic acid (PLGA) fibres is proposed to increase the bioactivity of scaffolds. According to the biocompatibility data, the FN deposition significantly enhanced both scaffold colonisation and cell behaviour. the impact of an FN gradient on electrospun surfaces, including the spreading behaviour and interactions between cells and scaffolds [ 38 ]. Connective tissue and alveolar bone regeneration increases after periodontal repair using guided tissue regeneration (GTR) surgery and results are somewhat improved by adjunct citric acid combined with autologous FN [ 39 ]. Diabetes-related chronic leg and foot wounds, as well as periodontal disease, frequently exhibit FN fragmentation. The theory that exposure to certain FN segments dramatically changes cell behaviour was further supported by cell culture assays [ 40 ]. Regardless of patient type, minimal FN could be found in the gingival crevicular fluid (GCF) of healthy sites. Additionally, the PDL's middle region is where collagen and new fibroblast production are both most active, indicating a connection between the locations of FN and fibroblast turnover. As a result, FN and its fragments are probably engaged in the interactions between cells and the ECM that support the preservation, regeneration, and wound healing of periodontal tissues [ 29, 30 ]. An analysis of the sites in each group, however, showed that the concentration of FN in the GCF was highest in healthy sites and decreased when there was gingival inflammation. The concentration of GCF and FN was never to be physiologically or biologically active [ 41 ]. In the pathophysiology of Porphyromonas gingivalis fimbriae in adult periodontal disease, salivary FN is the key regulator [ 42 ]. Future directions ECMs, both natural and synthetic, have been widely used in biomedical fields as one of the most successful elements in tissue regeneration [ 43 ]. The primary obstacle preventing the enrichment of biomaterials with FN or its fragments is the compound's high molecular weight, which limits its stability and bioavailability. Therefore, the development of substances that particularly bind FN would be desirable. Placing smaller molecules on scaffold surfaces would be a novel idea in this context and might be used to attract and retain FN from the extracellular environment [ 1 ]. FN is a versatile material platform used for a variety of purposes, including disease biology and tissue engineering, to finetune how an essential ECM protein is given to cells [ 44 ]. Conclusions The use of FN due to its cell behaviour conditioning has many applications in dentistry and periodontal therapy. Due to the structural support and signalling cues offered by FN, the ECM plays several roles in cell survival, migration, differentiation, gene expression, growth factor signalling, and cell contractility, among others. FN is working wonders in the field of dentistry, diagnosing periodontal disease. Specific FN fragments impair the activities of PDL cells in vitro, and FN fragments are identified in vivo in conjunction with periodontal disease. Specific FN segments serve as indicators of periodontal disease status and provide evidence for their potential involvement in the pathophysiology of the condition. The discipline of dentistry needs to use FN more in clinical aspects and more research on this adaptable biomaterial needs to be done. It controls cell behaviour and plays a key role in communication between the intracellular and extracellular environments, aiding in early attachment and healing of surgical wounds, and other functions that make it useful as well as a versatile biomaterial in dentistry and periodontal applications.

10. 7759/cureus. 30982

Cureus

Classic and Current Opinions in Human Organ and Tissue Transplantation

Graft tolerance is a pathophysiological condition heavily reliant on the dynamic interaction of the innate and adaptive immune systems. Genetic polymorphism determines immune responses to tissue/organ transplantation, and intricate humoral and cell-mediated mechanisms control these responses. In transplantation, the clinician's goal is to achieve a delicate equilibrium between the allogeneic immune response, undesired effects of the immunosuppressive drugs, and the existing morbidities that are potentially life-threatening. Transplant immunopathology involves sensitization, effector, and apoptosis phases which recruit and engages immunological cells like natural killer cells, lymphocytes, neutrophils, and monocytes. Similarly, these cells are involved in the transfer of normal or genetically engineered T cells. Advances in tissue transplantation would involve a profound knowledge of the molecular mechanisms that underpin the respective immunopathology involved and the design of precision medicines that are safe and effective.

Introduction and background Transplantation (or grafting) is a surgical or medical procedure involving grafting cells, tissues, or organs from one body part to another, thereby substituting or repairing the damaged, missing, or diseased cells, tissues/or organs. Therefore, a transplant (or a graft) is a group of cells, tissue, or organ grafted into a recipient. Transplants can save lives or restore function to a better quality of life for sick people with vital organ failure if correctly done [ 1, 2 ], but they can also bring untold challenges [ 3, 4 ]. The demand for organ transplants increases steadily, with the kidney being one of the most transplanted solid organs. The kidney, liver, heart, lung, pancreas, and small bowel were the most transplanted solid organs in 2019 and accounted for the 153, 863 transplants recorded [ 5 ]. The COVID-19 pandemic caused a decline in transplantation rates in the early periods of the outbreak. Still, the demand for transplants by diseased patients has not waned, thereby pointing to the continued relevance of tissue transplantation in the medical sphere [ 6 ]. Religion, societal behavior and beliefs, and medical ethics are challenges to the general acceptance of tissue or organ transplantation [ 3 ]. Furthermore, successful transplantation usually depends on the occurrence or absence of rejection [ 7 ], while a shortage of appropriate donor organs is still a major limiting factor in transplantation [ 7, 8 ]. A good understanding of the immunology of transplantation rejection is vital if more advances are made in this field. The favorable manipulation of the immune cells to promote graft tolerance will be advantageous to solving the problem of tissue rejection [ 8 - 10 ]. This review discusses some historical and relevant opinions and the mechanisms and immunology involved in tissue transplantation and graft rejection. Materials and methods The relevant works of literature were obtained by screening online databases, namely: Medline/PubMed, Google Scholar, Scopus, Web of Science, ProQuest, and grey works of literature, using some keyword combinations such as: "Tissue Transplant", "Graft", "Tumor immune response", "Graft tolerance", History of Transplantation", "Types of Transplantation", "Immunology of Transplantation", "Transplant Rejection", "Preventing Rejection" and "Regenerative Medication". Only English publications were included. Also, both original and review articles were used in preparing the study. Review Brief History of Tissue Transplantation The transfer of tissues and organs, based on needs, among humans is a practice that has its roots in the early centuries. Hamilton [ 11 ] narrated extensively how ancient man showed belief in replacing lost organs through procedures of magic and miracles. Hamilton's account is corroborated by documentation on skin transplants done between 3000 and 2500 BC (Before Christ) in India [ 12 ]. Early research on tissue transplantation among different species, especially between animals and man and humans, was filled with many challenges despite a few recorded successes [ 12, 13 ]. Nevertheless, the evolvement of science and better documentation has led to significant progress in the art of transplantation. Alexis Carrel's exploits in vascular science, which involved the transplantation of blood vessels, won him a Nobel Prize and led to better transplantation of other body organs by connecting the arteries and veins of a donor to the corresponding arteries and veins of the recipient [ 14 ]. The progress of transplantation up to the current age is better explained by how kidney transplantation has evolved. As recorded by Hakim & Papalois [ 15 ], kidney transplants in the early 20 th century involved transplantation amongst animals and later from animals to humans. In addition to the first kidney transplant between humans, these transplantation procedures were largely unsuccessful. The failed attempt for the first transplant amongst humans was recorded in Russia in 1936, and a post-mortem donor was involved [ 16, 17 ]. More attempts at kidney transplantation were later adjudged successful between the 1950s and 1960s. The work progressed from transplants involving identical twins to non-identical twins before climaxing with transplants involving non-siblings. A chronological flow of the significant landmarks [ 18 - 20 ] in kidney transplantation is shown in Figure 1. Figure 1 Timeline of landmark achievements in kidney transplantation. This figure has been developed using Biorender [ https://biorender. com /] license number: YP24IH1241. Image Credit: Susmita Sinha. Tissue transplantation is now attempted in almost all human body organs; this scientific venture has explored the bones, the eyes, the skin, and solid organs [ 12, 21 ]. A most recent account of how transplantation has evolved, especially genetic engineering, has been reported [ 22 ]. In what was described as a ground-breaking heart transplant, a male patient received a pig heart that was previously modified genetically. The new heart was said to have performed well for several weeks without rejection before the man eventually died [ 23 ]. Although not free of ethical concerns, such attempts at xenotransplantation point to a bright future for the science and art of tissue transplantation. Types of Transplantation There are 4 kinds of grafts or transplants (xenograft, isograft, allograft, and autograft) based on the genetic variations between the recipient's and donor's tissues (Table 1 ). The immunology of grafting is a very complex specialty in medicine [ 24 ]. Grafted organs/tissues may either be rejected or destroyed by the recipient's immune system, or the recipients may accept the organ or tissue. If there is rejection, medication to suppress the immunologic response from the recipient is most likely needed. Table 1 Categories of Organ/Tissue Transplantations with their possible unfavorable results. S/N Transplant Type Donor and Recipient Potentially unfavorable consequences 1 Xenotransplant The donor is an animal, while the recipient is human Rejection is highly possible 2 Allotransplant The donor and recipient may or may not be relatives but must be same species Rejection is potentially likely 3 Isotransplant The donor and recipient are identical twins Rejection may not be likely 4 Autotransplant The donor is the self, and the recipient is also self No envisaged rejection Xenografting or Xenotransplantation The word "Xenos" is a Greek word meaning foreign or strange. Xenografting is heterologous transplantation involving the grafting of viable cells, tissues, or organs between two species (e. g. , a dog and a pig). It is a cross-species transplantation method. The continued demand for viable organs, tissues, and cells brought about by end-stage organ failure and chronic diseases has been the driving force in this medical/scientific research and practice [ 25 ]. However, it has been confronted with the significant challenges of immunological barriers and ethical issues. Organ rejection is widespread in xenotransplantation. In humans, for instance, natural antibodies circulate in the blood, and these cause instant transplant rejection when the organ-donating species is, for example, a pig. Again, the complement systems are often activated each time organs from pigs are grafted into humans or primates and are highly prone to profound system toxicity due to the central role played by the complement system in body homeostasis and metabolism [ 26 ]. The porcine complementary proteins are foreign to primate complement regulatory systems. Studies have shown that genetic engineering may be a way out of this complementary system challenge if pigs are genetically modified to contain some human complement regulatory proteins in their cells [ 27 ]. Another fundamental challenge facing xenograft is ethical issues. Three ethical issues quickly come to mind when we talk of xenotransplantation: animal rights (effects on the donor animal), human rights (the impact on the human population and the impact on the individual recipient), and interference with nature. An animal rights issue arises because animals, like humans, also have rights to existence and should not be sacrificed in favor of humans [ 28 ]. Human rights regarding the recipient can quickly be cleared by obtaining the necessary informed consent. Still, the populace also needs authorization because of the possibility of transferring new pathogens from animal to human populations - a public health risk [ 1 ]. The ethical issue of interference with nature may not be so applied. It may be understandable that by interfering with nature, man can free himself from the extinction effects of some natural phenomena [ 28 ]. A few examples of xenograft include grafting human keratinocytes onto non-human cells (e. g. , mice) and then using "ZenSkin" (Reconstructed Human Epidermis) construction as a model for human skin physiology. ZenSkin has applications in pre-clinical and R&D for evaluating how a topical product will affect the human skin [ 29, 30 ]. Other examples include transfusing non-human blood into human patients and skin grafts from non-humans. Voronoff, in the 1920s, suggested that transplanting slices of chimpanzee testis into geriatric male patients with low sexual vigor would give new energy to such patients [ 31 ]. A French Surgeon, Alexis Carrel, developed a method of suturing blood vessels, thereby facilitating organ grafting from non-human primates into human patients [ 32, 33 ]. Isograft or Isotransplantation This refers to the inter or intra-transfer of viable tissue(s) or organ(s) between organisms of the same species. Intra-transfer involves the grafting of tissues or organs from a part of the body of an organism to another part of the same organism, while inter-transfer is between separate organisms but of the same species [ 34 ]. Corneal transplantation (or keratoplasty), Dacron vascular grafts, and cartilage and bone grafting are all examples of isografts. Renal transplantations are very common and rated as the most successful, primarily because artificial kidney machines are available and the kidney is a paired organ. There is tissue-type compatibility and less risk of fatal organ rejection by the recipient because of donor-recipient matching [ 35 ]. A transplant between identical twins is another example of isograft. It is implausible that a recipient will reject an isograft, so an immunosuppressant is unnecessary. Isograft is an allograft of tissue transplanted between genetically identical individuals of the same species. It refers to tissue grafted from genetically similar twins to another within a species. Autograft transplantation (or autologous grafting) is the grafting of tissue/organs from one area to another position in the same individual patient. Autograft Autologous grafting is the transplantation of viable cells, tissue, or organs from one area to another of the same individual or patient. It is frequently referred to as the "gold standard" in bone grafting due to its dependability [ 36 ]. The high success rate is due to the fact that bone autograft is a living tissue that contains osteogenic cells and growth factors needed for healing and bone regeneration [ 37 ]. Autograft mostly involves tissue transplant where occasionally tissues more desperately elsewhere are required (examples include skin grafts where a skin tissue can be removed from a part of the body with surplus or less important area and transplanted to another area where the tissue is, vein extraction for CABG, etc. ) can be extracted and transferred to another part of the same individual. Sometimes an autograft is done to remove the tissue and then treat it in-vitro or treating the person before returning it to the site of action [ 38 ]. Other common types of autografting include the reconstruction of the damaged anterior cruciate ligament, skin grafting used to replace damaged or lost skin, and blood vessel grafting used in heart bypass surgery to create an alternative route for blood flow to bypass a blocked coronary artery [ 39 - 41 ]. Autografts pose no risk of disease transmission or immune rejection. However, they have several limitations, which include a limited supply, surgical complications, donor-site pain, and high donor-site morbidity at the procurement site [ 42 ]. Allograft Allografts are tissues such as bone, skin, tendon, ligaments, and heart valves recovered from a human donor who is not an identical twin for transplantation into another person [ 43 ]. The transplant is called an allogeneic transplant (allograft) or homograft. Most human organ transplants are under allografting, where an organ is extracted from an individual (donor) and transferred to another individual (recipient). Due to the difference in genetic constituents of donor and recipient, allograft may result in a significant immune response that may trigger graft rejection [ 44 ]. Allografts have been successfully used in various medical procedures, especially when an autograft cannot be used. Allograft skin is beneficial in patients with burns that cover a large area of the body. It can be used as a temporary dressing while awaiting the healing of autograft donor sites between harvesting sessions [ 45 ]. Also, allografts are used in corneal transplantation when a patient has damaged or failed corneas [ 46 ]. Pretransplantation screening of allografts is performed to confirm the donor's tissue viability and the donor's health status to eliminate transmissible diseases such as HIV, Syphilis, hepatitis B, and hepatitis C [ 47 ]. To ensure the recipient's safety, the allograft is cleaned and aseptically processed using alcohol, antibiotics, and detergents to rid the tissue of as many cellular elements as possible. Chemical sterilization and electromagnetic radiation are also used to destroy microbes [ 48 ]. Unlike the autograft, it takes longer to incorporate into the recipient's body. Chronic rejection and toxicity of immunosuppressive drugs used to improve successful allograft acceptance are some challenges facing the clinical execution of allograft transplants [ 49 ]. An example of allografting rejection includes transplanting an organ, such as skin, between two parties who are not identical twins. Skin allografts are used for patients with widespread burns or other conditions causing such huge skin loss that the patient does not have enough intact skin to provide the graft. Skin allografts are eventually rejected due to T cell allorecognition leading to an inflammatory immune response. Still, the resultant wounded areas that are evident by the loss of epidermis, caused by prolonged moisture and friction, develop into well-vascularized granules that autografts from the patient have healed sites take readily [ 50 ]. However, an example of allografting without organ rejection is a cornea transplant. Cornea transplants are often not rejected because the cornea has no blood vessels resulting in the inability of the host immune system to recognize and reject the graft [ 51 ]. Immunology of Transplantation Rejection Organ rejection is known to result from the interactions between the adaptive and innate immune systems with the implicated lymphocytes, macrophages, neutrophils, and natural killer cells [ 7 ]. The histocompatibility antigens (HCA), encoded by histocompatibility genes (HCG), are implicated in the rejection of grafted tissues and organs [ 52 ]. Over 40 loci on the HCG are known to encode HCA. However, the loci on the major histocompatibility complex (MHC) have been remarkable for the most dangerous allograft rejection reactions [ 53 ]. The human MHC is found on the short arm of chromosome number 6, very close to the complement genes [ 54 ]. However, other antigens causing weaker reactions may exhibit strong rejection reactions in combination. An individual can manifest the MHC genes from both allelic pairs on the body cell surface, with each team coming from each parent. Each child is half identical to the mother and the father regarding the MHC complex. Therefore, it follows that an individual has a 25% likelihood of having a sibling with a similar MHC. This forms the basis of allograft between relatives. The human MHC genes complex encode-3 prominent Class I alleles, namely human leukocyte antigens (HLA)-A, HLA-B, and HLA-C, and 3 top-class II alleles, HLA-DR, HLA-DQ, and HLA-DP. The occurrence of two or more distinct forms (alternative phenotypes) of HLA-A, HLA-B, or HLA-DR loci is a known cause of failed transplantation. Closely HLA-matched transplant will most unlikely be recognizable and rejectable, and HLA mismatching has grave effects on the recipient's transplant survival [ 55 ]. The MHC molecules are classified as either Class I or Class II molecules. While class I molecules reside in cells with a nucleus, class II molecules reside in professional antigen-presenting cells (APCs) [ 56 ]. Physiologically, MHC molecules display antigenic peptides on the T cells, and t lymphocytes can only respond to processed and presented antigens that have complexed with the MHC molecules. The class I molecules offer antigenic peptides from within the cell (endogenous- and auto-antigens) to the cluster of differentiation (CD) 8 T cells [critical subpopulation of major histocompatibility complex (MHC) class I-restricted T cell]. Such antigens include intracellular bacteria, viruses, parasites, cancer cells, and self-antigens. The class II molecules process and present exogenous (extracellular) antigens like extracellular bacteria to CD4 T cells [ 57, 58 ]. Clinical Stages of Graft Rejection The clinical stages involved in graft rejections are summarized in Figure 2. Figure 2 Clinical Stages of Graft Rejection. Notes: APC=Antigen Presenting Cell, CD=Clusters of Differentiation. This figure has been developed using Biorender [ https://biorender. com /] License Number: DA24ILUA8K. Image Credit: Susmita Sinha. Hyperacute Rejection Hyperacute rejection appears within 24 hours after grafting and only in grafts with profound blood vessels such as the kidney. It is characterized by blood clots inside the blood vessels and graft necrosis. This kind of immunological response is mediated by humoral immunity; the recipient has pre-formed antibodies against the transplant [ 59, 60 ]. The antigen-antibody complexes cause the stimulation of the complement system, leading to profound clot formation in the capillaries and consequent death of the graft. The liver is relatively more resistant to hyperactive rejection than the kidney, possibly due to dual blood supply to the hepatic system. Proper ABO cross-matching with the exclusion of anti-donor human leukocyte antigen (HLA) antibodies mitigates hyperacute rejection [ 53 ]. Acute Transplantation Rejection Occur any time from the first week to 6 months after the transplant as acute cellular rejection or as acute humoral rejection. Acute Cellular Rejection This is an immunological response in the host's/recipient's lymphoid tissues due to lymphocytes stimulated against donor antigens. The donor's dendritic cells enter the recipient's systemic circulation to function as antigen-presenting cells (APCs) [ 50, 61 ]. It is common in renal grafts. Acute cellular rejection detection involves biopsy, B-lymphocyte antigen CD20 staining in cases not responding to treatment, negative kidney C4d staining, positive activating lymphocyte markers test, and proteomic study [ 62 ]. The first rejection instance is treated with pulse intravenous steroids and may be repeated in cases of recurring or refractory rejections. The second line of treatment (Thymoglobulin and a murine monoclonal antibody, OKT3) may be used for deteriorating grafts. The prognosis depends on the number of rejection episodes, potent drugs, time of rejection from transplantation, and response to treatment [ 62 ]. Acute Humoral Rejection This is also called acute vascular rejection. It is a severe organ transplant injury mediated by antibodies and complement. The antibodies may be pre-existing or represent anti-donor antibodies developing shortly after grafting [ 63, 64 ]. Willicombe et al. [ 65 ] demonstrated that even low donor-specific antibodies titer not detectable with flow cytometry or complement-dependent cytotoxic cross-matches is linked to lower-ranking renal allograft outcomes. Such patients will likely need augmented immunosuppression. Loupy et al. [ 66 ] posited a significant swing in the first-year post-graft in the C4d Banff scores, thus proving the humoral process's changing and painless nature of C4d is not a sufficiently sensitive marker. Still, inflammations in the microvessels and spotting of donor-specific antibodies are better markers of humoral rejection. Chronic Graft Rejection (CGR) This is also called chronic transplant rejection (CTR). The allograft function is lost several months to years after grafting. Although the graft may still be in place, graft function loss is due to persisting immune system attacks on the allo-MHC. CGR is mediated by humoral as well as cellular immunity. Although immunosuppressants and tissue-typing methods are helpful in the first-year post-graft, CGR is almost always not preventable. It appears to be fibrotic scarring in the grafted organs, although the specific histopathology image depends on the grafted organ [ 67 ]. Mechanisms of Rejection in Tissue Transplantation The immunological reaction to the grafted organ is both lymphocyte and antibody-mediated. Nevertheless, the central player in transplant rejection is the T cell/lymphocyte [ 68 ]. There are 2 phases in transplant rejection (Figure 3 ): a sensitization phase and an effector phase [ 50 ]. Figure 3 Clinical Stages of Graft Rejection. Notes: APC=Antigen Presenting Cell, MHC=Major Histocompatibity Complex, CD=Clusters of Diffrentiation, T Cell= A Subclass of Lymphocytes. This figure has been developed using Biorender [ https://biorender. com /] License Number: DA24ILUA8K. Image Credit: Susmita Sinha. Sensitization Phase Here, through their receptors, the helper (CD4) and cytotoxic (CD8) T-cells can identify the alloantigen displayed on the donor/foreign transplant cells. Antigen recognition begins with the T-cell receptor cross-talk with the antigen expressed by MHC molecules, followed by the costimulatory receptor/ligand cross-talk with the T-cell/APC surface [ 69 ]. One of the several costimulatory pathways involved in the sensitization phase is the communication between the T-cell surface CD28 with its APC surface ligands, B7-1 or B7-2 (referred to as CD80 or CD86, respectively) [ 70 ]. Also, CD8-associated antigen-4 (CTLA4) binds to B7-1 or B7-2 ligands to provide signals that cancel effects. CD40 and its ligand CD40L (CD154) equally serve for co-stimulation in this phase. Typically, the two convolutions of the MHC molecules form a peptide-binding groove to take up the peptides of normal cellular proteins origin. Thymic or central and peripheral tolerance mechanisms swing into action to ensure that the formed self-peptide-MHC complexes are unrecognizable by the T-cells, suppressing any possible autoimmune responses [ 71 ]. The two distinct but interrelated pathways of allorecognition are the direct and indirect pathways, generating specific groups of allospecific T-cell clones. Direct Pathway/Mechanism The direct mechanism is the primary pathway seen in early immunological response. Here, the host/recipient T-cells identify whole allo-MHC molecules found superficially on the donor or stimulator cell. The recipient T-cells see allo-MHC molecule + allo-peptide as having the self-MHC + non-self-peptide shape and determine the donating tissue as non-self [ 50, 72 ]. The grafted organ has an undefined number of passenger APCs that appear as dendritic cells occupying the interstices with intensely populated allo-MHC molecules. These can activate the recipient's T cells directly. When the allogeneic or donor cells interact with the T-cells, the T-cells proliferate profusely in comparison with the clone populations that target antigens displayed by auto-APC. This mechanism is suggested in acute allorejection [ 73 ]. Indirect Pathway/Mechanism T-cells identify refined alloantigens displayed as peptides by auto-APCs. Then, epitope switching or spreading in which T cells proliferate to a more variable repository, such as initially immunologically dormant peptides [ 74 ]. Ali et al. [ 75 ] demonstrated that the connection of self-MHC + allopeptide-primed T cells with acute vascular type rejection is partially modulated via the production of augmented alloantibody. In contrast, chronic allograft vasculopathy is modulated by primed T cells. Molecular Interactions in T-lymphocyte Activation During T-lymphocyte (T-cell) stimulation, inositol phospholipid molecules in the cell membrane are added to water molecules to form diacylglycerol (DAG) and IP3 [ 75 ], resulting in the influx of Ca2+ into the cytoplasm [ 76 ]. This provokes a series of events that form calcium-calmodulin complexes, stimulation of several kinases, protein phosphatase IIB or calcineurin, and calcineurin dephosphorylates cytoplasmic, nuclear factor of stimulated T cells (NFAT) and thus causing NFAT to relocate from the cytoplasm into the nucleus. In the nucleus, NFAT combines with the Interleukin-2 promoter sequence to activate the synthesis of Interleukin (IL)-2 mRNA from DNA [ 77 ]. Several other events also take place within the T cell, such as protein kinase C (PKC) stimulation by diacylglycerol (DAG) and stimulation of nuclear factor kappa B (NFkB) [ 76, 78 ]. Effector Phase The effector phase is the second phase in organ transplant rejection that involves alloantigen-dependent and independent factors. Reduced blood flow initially induces a nonspecific inflammatory reaction, leading to increased antigen presentation to T cells due to the upregulated expression of adhesion molecules [ 79 ]. Also, intact soluble MHC molecules are liberated to stimulate the indirect allorecognition pathway [ 80 ]. Within the first few weeks after tissue transplant, several T lymphocytes and their derived cytokines like IL-2 and IFN-γ are generated. Later, RANTES (Regulated on Activation, Normal T Cells expressed and secreted), MCP-1, and IP-10 are produced, leading to the influx of many macrophages into the allograft. The effector phase is also marked by upregulation of Interleukin-6, Tumor Necrotic Factor-α, inducible nitric oxide synthase (iNOS), and growth factors leading to rapid multiplication of smooth muscles, thickening of the inner lining of lymph and blood vessels, interstitial fibrotic scarring and, in the case of the kidney, scarring or hardening of the glomeruli [ 57, 58 ]. MHC class II molecules, costimulatory molecules, and adhesion molecules are expressed following the stimulation of the endothelial cells by T lymphocytes-derived cytokines and macrophages [ 81, 82 ]. Apoptosis Apoptosis is the last stage involved in tissue rejection. It is the usual mechanism for the cell-killing processes leading to the programmed death of the target cell [ 82 ]. Post-stimulation of the cytotoxic T lymphocytes involves the generation of cytotoxic granules containing (a) serine proteases (called granzymes) that induce programmed cell death and (b) pore-forming cytolytic proteins (perforin) [ 82, 83 ]. The cytotoxic granules join the effector cell membrane during target cell recognition and arrangement and liberate its content into the immune synapse. The granzymes insert into the target cell cytoplasm to induce programmed cell death (apoptosis). This is the common cause of apoptosis in allograft rejection [ 83 ]. The fas-dependent pathway is another important pathway CD8+ can employ to achieve cytolysis and apoptosis and limit T-lymphocytes' rapid multiplication in response to stimulations to antigens. Cell-mediated cytotoxicity plays active functions in acute allograft rejection [ 84, 85 ]. Role of Natural Immunity in Graft Rejection The T-lymphocytes unarguably play an essential role during acute organ rejection (Figure 4 ). However, the increase in pro-inflammatory mediators in the allograft occurs before the T lymphocytes response is seen as an innate response to tissue injury and does not depend on the acquired immunity [ 86, 87 ]. Figure 4 Role of Natural Immunity in Graft Rejection. Notes: APC=Antigen Presenting Cell, DAG=Diacyl Glycerol, LAT=Linker for Activation of T-Cell, MAPK=Mitogen-Activated Protein Kinase, RAS=Rat Sarcoma, PKC=Protein Kinase C, SAPK=Stress Activated Protein Kinasse, JNK=c-Jun N-terminal Kinase, CD=Clusters of Differentiation, PIP=Phosphatidyl Inositol Phosphate, PKB=Protein Kinase B, NFκβ= Nuclear Factor kappa beta, Th=T hepler cells, IL=Interleukin. This figure has been developed using Biorender [ https://biorender. com /] license number: VY24J04THQ. Image Credit: Susmita Sinha. Even though natural mechanisms alone do not lead to transplant rejection, they are necessary for optimal acquired immunological reactions to the transplant. They are also vital in resistance to tolerance induction [ 88, 89 ]. Although essential in particular disease management, cutting off the natural immune responses most assuredly impacts tissue grafting [ 86 ]. Natural Killer (NK) Cells NK cells can discriminate between allogeneic cells and self and have robust cytolytic effector mechanisms to establish as much effector response as possible, even without previous immune sensitization [ 90 ]. Unlike lymphocytes, NK cells can be stimulated even without MHC molecules. This is possible due to the several NK inhibitory receptors produced by specialized alleles of MHC class I antigens on cell surfaces. NK cells are also equipped with stimulatory receptors activated by antigens on non-self-cells. NK cells also assist CD28+ host T lymphocytes and encourage allograft rejection [ 91 ]. NK cells have been identified to play an active role in chronic and acute rejection of solid organ grafts [ 92 ]. In addition, they also modulate allograft outcomes of the heart. Neutrophils Because of their number and high motility, neutrophils are the prime white blood cells to migrate to grafted organs and have been recognized as potent markers of transplant injury [ 93 ]. The release from dead cells upregulates the stimulation and subsequent neutrophil infiltration into grafted tissues, and the extracellular matrix is of damage-associated molecular patterns (DAMPs) [ 94 ]. DAMPs also trigger the generation of inflammatory cytokines by activating pattern recognition receptors (PRRs) on macrophages. These inflammatory cytokines include ELR+ CXC chemokines and IL-1β, which play some critical functions in neutrophil recruitment [ 95 ]. In addition, neutrophils also exhibit PRRs. When activated by DAMPs, they evoke a series of events, including; the production and release of reactive oxygen species (ROS) and hydrolyzing enzymes that aggravate damage to transplanted organs/tissues. Although not counted among the professional antigen-presenting cells (APCs), neutrophils can migrate from peripheral sites to transport their antigens to lymph nodes [ 96 ]. They can also trigger T-cells differentiation by an exhibition of MHC and costimulatory molecules [ 97 ]. Neutrophils are also known to contribute to clearing inflammation and start the production of anti-inflammatory substances among other myeloid cells [ 98, 99 ]. Macrophages These are highly motile, naturally trained immune cells capable of detecting, ingesting, and destroying disease-causing and other harmful particles. They constitute most parts of host defense and tissue homeostasis mechanisms and initiate the development of other immune cells [ 100 ]. Tissue macrophages are localized inside tissues, while blood macrophages originate from the monocytes that circulate in the blood and develop into macrophages in the bone marrow. They are pivotal in the mediation of transplant immunopathology. Apart from mobilizing first-line defense against pathogenic organisms and functioning as APCs, they equally censure allografts as non-self-entity and encourage transplant loss by a similar mechanism [ 101, 102 ]. Macrophages are implicated in ischemia/reperfusion injury (IRI), the alloimmune response, and acute graft rejection [ 103, 104 ]. Macrophage mobilization happens immediately after reperfusion during organ grafting, and copious amounts of pro-inflammatory cytokines are generated to destroy the tissue [ 105, 106 ]. Macrophages may also trigger graft rejection by activating acquired alloimmune reactions. They also furnish costimulatory signals that ease and augment the stimulation of T lymphocytes [ 101 ]. Transplant injury could be alleviated and graft survival prolonged if macrophages are deleted or inhibited [ 107 ]. Both clinical and animal studies demonstrated some positive correlation between allograft rejection and macrophage infiltration [ 108, 109 ]. Also, in B cell-mediated rejection, there is demonstrable infiltration of macrophages and monocytes [ 110, 111 ]. Graft Tolerance and Minimizing Rejection Tissue/organ graft is recommended for end-stage tissue/organ failure patients. The clinical practice's goal and challenge are striking a balance between the allogeneic immune response, the unwanted consequences of the immunosuppressants, potentially fatal infection, malignancies, organ toxic­ity, hypertension, and diabetes. Mitigating long-term immunosuppression through immunologic tolerance is highly recommended to ensure long-term patient and allotransplant survival. That graft recipients enjoy a better quality of life and improved life expectancy [ 8 ]. Transplant tolerance conserves stable allotransplant functions without immunosuppressive treatment [ 8 ]. Although rejection cannot be ruled out completely, some immunological tolerance to the grafted tissue does occur. Some hypotheses on the development of transplant tolerance include adverse selection in the form of clonal deletion, absence of the normal immunological reaction to a particular antigen or allergen in donor-specific T and B cells, and formation of immune cells that blocks the actions of some other types of lymphocytes, or circumstances that decrease the immunological response against the transplanted organ and lingering dendritic cells (in the organ recipient) that are from an organ donor and which ensure immune-mediated chimeric state between the grafted organ and its recipient. Regulatory T Lymphocytes in Graft Tolerance Ensuring allograft tolerance has become an ideal treatment goal in clinical transplant practice. Mitigating immunological reactions in allotransplantation and suppressing infections and tumor formation are significant hurdles in transplant practice. Although immunosuppressants effectively suppress acute rejection [ 112 ], currently utilized options cannot ensure that the recipient's immune system responds to antigens except those from donor alloantigens after transplantation [ 113, 114 ]. Regulatory T cells (Tregs) refer to the specialized subset of T lymphocytes processing immunological reactions and ensuring homeostasis and self-tolerance. They suppress T lymphocytes' rapid multiplication and stimulation by cell-to-cell contact [ 115 ], modulate hyper-immune responses to non-self-antigens, and uphold self-tolerance [Figure 5 ] [ 115, 116 ]. Figure 5 Illustrating the regulatory effect of T regulatory cells on the immune system. Treg cells release anti-inflammatory cytokines like IL10 and TGFβ and also convert ATP to AMP, which together inhibits the proliferation of effector T lymphocytes. Treg cells release perforin that attacks effector T cells and causes their apoptosis. CD25 expression from Treg cells causes sequestration of IL 2 and decreases the proliferation of Natural Killer cells (NK cells). Treg cells also directly inhibit the proliferation of B lymphocytes and reduce the expression of CD 80 and CD 86. Treg also promotes the differentiation of monocyte to M2 macrophages and suppresses the conversion of monocyte to M1 macrophages, which is pro-inflammatory. Treg also causes neutrophils to reduce the secretion of IL 6 and CXCL. Notes: Treg cell: T regulatory cell. NK cell: Natural killer cell. IL: Interleukin. TGF: Transforming Growth Factor. CXCL: CXC chemokine Ligand. , ATP: Adenosine Triphosphate, AMP: Adenosine Monophosphate, CD: Clusters of differentiation, T Cell: Subclass of Lymphocytes, IL: Interleukin. This figure has been developed using BioRender [ https://biorender. com /] License Number: PL24IU7VJY. Image Credit: Rahnuma Ahmad Pellerin et al. [ 117 ] suggest that Tregs are important in ensuring allograft tolerance. Treatments targeting Treg function and survival are novel options for ensuring immuno-tolerance in patients with organ transplants. CD25 and MHC class II expressions are the two important Tregs markers [ 118 ]. It has been demonstrated that successful allografting in humans is linked to a robust CD4+CD25+ Tregs population [ 119 ]. CD25+CD4+FOXP3+ regulatory T cells function to modulate immunological reactions to alloantigens and prevent rejection in-vivo [ 120 ]. Naturally occurring CD25+CD4+FOXP3+ regulatory T cells are produced as separate subsets during the differentiation of T lymphocytes in the thymus [ 121 ]. During organ grafting, CD25+CD4+FOXP3+ regulatory T cells (phenotypically and physiologically related to those derived from the thymus) may be triggered either in-vivo or ex-vivo alloantigen exposure [ 122 ]. The mouse model has also demonstrated similar regulatory T-cell functions [ 123 ]. Innate Immune Cells in Transplantation Tolerance Monocytes and Macrophages: Monocytes are blood phagocytes that form macrophages - the tissue-resident dendritic cells (DCs). Macrophages can modulate acquired immune responses and exhibit pro- or anti-inflammatory effects [ 124 ]. It has been previously stated that macrophages can contribute to allotransplant rejection via several mechanisms. However, evidence suggests they are also implicated in transplant tolerance in the adoptive transfer of regulatory macrophages (Mregs) [ 125, 126 ]. These Mregs can inhibit the alloactivation of T lymphocytes via iNOS generation and function as critical mediators of transplant tolerance [ 126 ]. They are crucial in the induction of immuno-tolerance and have associated therapeutic involvement in tissue grafting [ 127 ]. Neutrophils: Neutrophils involved in programmed cell death (apoptosis) are also able to modulate inflammation by releasing Arginase-1 (a metabolic suppressor of T lymphocyte stimulation) and shedding microvesicles that bear anti-inflammatory mediators [ 128, 129 ]. A unique neutrophil subset through matrix metallopeptidase-9 (MMP-9) expression is required for optimal reperfusion of grafted islets [ 129 ]. Natural Killer Cells: Administration of anti-CD28 monoclonal antibodies causes NK cells to enhance tolerance during kidney allotransplant by inhibiting pro-inflammatory immunity [ 130 - 133 ]. López-Botet et al. [ 134 ] posited that the pathway of tolerance induction by NK cells depends on the nature of the graft or the immunosuppressant therapy. Distinct subpopulations of NK cells can induce tolerance through specific pathways, such as toxicity of the white blood cell or/and cytokine release. This can be observed during chronic inflammation or infection. Here, NK cells are triggered, on exposure to IL-12, to secrete IL-10 [ 135 ]. IL-10 cytokine secretion by NK cells ensures that the fetus is not rejected by maternal allospecific T lymphocytes and inhibits inflammatory responses in the brain, spinal cord, and eye [ 136 ]. NK cells indirectly also trigger regulatory T lymphocytes in anterior chamber-acquired immune deviation (ACAID), leading to a generalized antigen-specific immune digression in the body [ 114 ]. The modulation of homeostatic CD8+ effector memory (TEM) enlargement by NK cells was perforin-independent, possibly moderated through competition for IL-15 cytokine [ 137 ]. NK cells can modulate the generation of tolerance by several pathways because of their cytolytic actions, cytokinogenesis, and capacity to compete for stimulation with cells aggressive toward "other" cells [ 138 ]. Depending on the nature of the graft and the recipient's alloimmune reactions, distinct NK cell subpopulations and pathways may be involved in tolerance initiation [ 139 ]. Cross-Matching and Use of Immunosuppressants to Mitigate Graft Rejection Cross-matching is vital in the workup towards tissue transplantation as a lack of data on compatibilities between donor and recipient will result in a futile outcome. When a positive cross-match is obtained on testing, it implies a hyperacute rejection is a potential outcome in any recipient of such graft. The rejection is usually due to the presence of donor-specific antibodies (DSAbs) in the recipient's serum performed against one or several human leukocyte antigens (HLA) [ 140 ]. Despite their roles in graft rejection, the HLA proteins are important because they can help the immune cells differentiate themselves from non-self-proteins, preventing bodily harm. In addition, the variations in the HLA genes are numerous, leading to complexities in the immunology of transplants [ 141 ]. Pregnancy, blood transfusion, and previous transplantation are significant ways DSAbs usually develop [ 142 ]. While there are a couple of cross-matching techniques available, the occurrence of high graft loss despite negative cross-matches in high-risk patients caused a need for the development of more sensitive cross-matching methods [ 141 ], such as the enzyme-linked immunosorbent assay (ELISA) and Bead-based fluorescent assays [ 142 ]. One of the most straightforward techniques for cross-matching, as seen in the Complement-dependent cytotoxicity cross-matching, involves preparing a mix of the recipient's serum with T or B cells (T and B lymphocytes) from the donor with the addition of a complement. The presence of lysis and its proportion indicates whether the cross-match is assigned a weakly, moderately, or strongly positive grade [ 140 ]. From the preceding, the role of immunosuppressants in helping to mitigate graft rejection becomes clear. Research on immunosuppressive agents has increased steadily over the decades. The corticosteroids were first employed as far back as 1950, before the advent of antiproliferative agents such as azathioprine [ 143 ]. Cyclosporine A and tacrolimus, both calcineurin inhibitors, are the primary agents used around the globe. Other approved agents are sirolimus, mycophenolate mofetil, and belatacept, which were approved in the last decade by the Food and Drug Administration (FDA) [ 144 ]. There are many ongoing clinical trials for novel immunosuppressive agents with intended clinical relevance in organ transplants. Tocilizumab, fingolimod, and sotrastaurin are some current agents being investigated [ 143 ]. The alleviation of graft rejection through immunosuppressants could be through induction or maintenance therapy. The final aim of all agents in use is to diminish immune response to promote graft tolerance and suppress the effects of any positive cross-match, especially for sensitized patients. Blockade of T-cell activation, induction of apoptosis, prevention of T-cell proliferation, and inhibition of B lymphocyte differentiation into antibody-producing cells are common mechanisms of action of immunosuppressive drugs [Figure 6 ] [ 144 ]. Figure 6 Showing the mechanism of different immunosuppressive drugs on T and B lymphocytes. Drugs inhibit specific pathways, cell cycle, and DNA synthesis by inhibiting mTOR, NFkB, NAFT, and JAK, which decreases lymphocyte activation and proliferation and promotes graft tolerance. mTOR: mammalian target of rapamycin. NFkB: Nuclear Factor kB. NAFT: Nuclear factor of activated T cells. JAK: Janus Kinase. This figure has been developed using Biorender [ https://biorender. com /] license number: RC24IZS47Z. Image Credit: Rahnuma Ahmad The invention of an individualized treatment plan for organ recipients and the discovery of those agents which would reduce toxicity and side effects and increase therapeutic efficacy in graft tolerance are the properties expected of future immunosuppressive agents [ 143 ]. Regenerative Medicine and Tissue Engineering Tissue engineering, as a field, seeks to understand and explore bio-substitutes for the restoration, maintenance, and improvement of the physiology of human tissues. In contrast, regenerative medicine as a field in health science seeks to understand and explore the processes involved in substituting, devising, or restoring mammalian cells, tissues, or organs to restore normal physiology. Tissue engineering and regenerative medicine (TERM) share many similar intended outcomes, leading to the coining of the acronym "TERM" to represent the two fields [ 145 ]. TERM is intended to help solve the significant problems with traditional transplantation: shortages in organ donors and immunologically engineered graft rejection [ 146 ]. Three key elements are necessary for the science and art of tissue engineering: scaffolds that serve as the extracellular matrix, cell sources, and a stimulus that could be in the form of growth factors [ 147 ]. While the scaffolds are mainly biodegradable materials, the cell samples could be obtained from tissues to be regenerated or, most recently, are usually stem cells (hematopoietic stem cells, embryonic stem cells, induced pluripotent stem cells, etc. ). Growth factors will help in vascularization and cell differentiation [ 145 ]. Furthermore, in TERM, cells could be obtained from the same individual (autologous) or a different person (allogeneic). Xenogenic cells have also been experimented with, which, alongside allogeneic cells, can elicit immune reactions, resulting in a need for immunosuppressants [ 146 ]. There are variations in the regenerative capacities of different human tissues and organs, with the cornea and cartilage showing very limited or no regenerative abilities and the lung and liver having more abilities [ 146 ]. This notwithstanding, a vast amount of research has been done in tissue engineering in recent decades. However, they have yet to yield the desired bench-to-bed outcomes, especially in bone tissue engineering. In bone tissue engineering, this is primarily due to unsuccessful clinical trials, which are attributed partly to the manufacturing and designing ideal scaffolds [ 148 ]. Spinal cord injury is another infirmity requiring the innovation provided by TERM. Salgado et al. [ 145 ] hydrogels have been adequately researched to employ tissue engineering techniques to deliver human neural stem cells. 3D bioprinting has been a way of making better scaffolds because it allows biomaterials to integrate well into a patient's tissue and promote vascularization [ 146 ]. Future Perspectives This review suggests the need for more advancement in research toward fighting tissue rejection and improving tolerance. It points to the multifaceted role of the immune cells in the concepts of graft rejection. Understanding the molecular biology of tissue transplantation facilitates the identification of the different proteins and pathways involved. This would enhance these proteins' genetic engineering and production in commercial quantities for prophylactic and therapeutic purposes. Also, the design of novel proteins through quantum computing can be possible at the proteomic dimension. Conclusions Tissue transplantation is still a relevant area in medicine with the potential for more breakthroughs if the hindering challenges are overcome. Even when improved with genetic manipulations, xenotransplantation faces ethical and rejection concerns. The T lymphocytes involved in the sensitization and effector phases of tissue rejection are central to the immunology of tissue graft rejection. However, the regulatory Tregs are necessary alongside the regulatory macrophages to fight rejection and promote tolerance.