Lipoteichoic Acid Accelerates Bone Healing by Enhancing Osteoblast Differentiation and Inhibiting Osteoclast Activation in a Mouse Model of Femoral Defects

Lipoteichoic acid (LTA) is a cell wall component of Gram-positive bacteria. Limited data suggest that LTA is beneficial for bone regeneration in vitro. Thus, we used a mouse model of femoral defects to explore the effects of LTA on bone healing in vivo. Micro-computed tomography analysis and double-fluorochrome labeling were utilized to examine whether LTA can accelerate dynamic bone formation in vivo. The effects of LTA on osteoblastogenesis and osteoclastogenesis were also studied in vitro. LTA treatment induced prompt bone bridge formation, rapid endochondral ossification, and accelerated healing of fractures in mice with femoral bone defects. In vitro, LTA directly enhanced indicators of osteogenic factor-induced MC3T3-E1 cell differentiation, including alkaline phosphatase activity, calcium deposition and osteopontin expression. LTA also inhibited osteoclast activation induced by receptor activator of nuclear factor-kappa B ligand. We identified six molecules that may be associated with LTA-accelerated bone healing: monocyte chemoattractant protein 1, chemokine (C-X-C motif) ligand 1, cystatin C, growth/differentiation factor 15, endostatin and neutrophil gelatinase-associated lipocalin. Finally, double-fluorochrome, dynamic-labeling data indicated that LTA significantly enhanced bone-formation rates in vivo. In conclusion, our findings suggest that LTA has promising bone-regeneration properties.


Introduction
Delayed bone healing leads to a diminished quality of life, a personal financial burden and increased medical care costs. Thus, novel therapies to enhance bone healing are sorely needed. The U.S. Food and Drug Administration defines a non-union as a fracture that is at least nine months old, which has not shown any sign of healing for three consecutive months [1]. Clinically, non-unions are classified according to their morphology and radiographic appearance in atrophic and hypertrophic non-unions [2,3].
Current state-of-the-art therapeutic methods for impaired bone healing can be divided into four categories: conservative therapy, surgical therapy, molecular targets for local or systemic Hematoxylin and eosin (H&E) staining revealed an early-onset dense bone bridge by Day 7 in LTA-treated mice, whereas mice in the sham group showed hematoma formation ( Figure 2a). Safranin-O and Masson's trichrome staining images were acquired. Many undifferentiated mesenchymal cells were present in the callus and in the areas of inflammation. In the sham control group, we observed the bone bridge, hypertrophic chondrocytes, proliferating chondrocytes and osteoblasts (Obs) on Day 14 after introducing the femoral bone defects. In the LTA group, the bone bridge appeared earlier (Day 7) after introducing the femoral defects; moreover, mature osteocytes (Ocys) and Ocls were present in trabecular bone on Day 14. These results demonstrate that endochondral ossification was nearly complete on Day 14 in the LTA group, but not in the sham group (Figure 2a). Immunofluorescence analysis showed intense osterix and cathepsin K staining surrounding the trabecular bones in LTA-treated group, whereas diffuse staining was observed in the sham control group (Figure 2b). We also quantified the intensities of osterix and cathepsin K signals relative to that of DAPI (bone tissue containing area). Bone remodeling occurred earlier in the LTA group (Supplementary Figure S1). The results of micro-computed tomography (micro-CT) revealed that treatment with LTA (but not the vehicle control) enhanced the bone healing. (b) Body weights were measured daily in mice treated with PBS or LTA. All the results were normalized to the initial weight of each mouse. (c) Quantitative results of micro-CT analysis in mice treated with PBS (n = 5) or LTA (n = 6). Data are presented as the mean ± standard error of the mean. Analyses were conducted with two-way analysis of variance followed by Bonferroni's post-hoc test. * p < 0.05, *** p < 0.001. Abbreviations: d, day; Tb.Th, trabecular thickness; Tb.N, trabecular number; Tb.Sp, trabecular spacing; BV, bone volume; TV, tissue volume; BS, bone surface.
Hematoxylin and eosin (H&E) staining revealed an early-onset dense bone bridge by Day 7 in LTA-treated mice, whereas mice in the sham group showed hematoma formation ( Figure 2a). Safranin-O and Masson's trichrome staining images were acquired. Many undifferentiated mesenchymal cells were present in the callus and in the areas of inflammation. In the sham control group, we observed the bone bridge, hypertrophic chondrocytes, proliferating chondrocytes and osteoblasts (Obs) on Day 14 after introducing the femoral bone defects. In the LTA group, the bone bridge appeared earlier (Day 7) after introducing the femoral defects; moreover, mature osteocytes (Ocys) and Ocls were present in trabecular bone on Day 14. These results demonstrate that endochondral ossification was nearly complete on Day 14 in the LTA group, but not in the sham group (Figure 2a). Immunofluorescence analysis showed intense osterix and cathepsin K staining surrounding the trabecular bones in LTA-treated group, whereas diffuse staining was observed in the sham control group (Figure 2b). We also quantified the intensities of osterix and cathepsin K signals relative to that of DAPI (bone tissue containing area). Bone remodeling occurred earlier in the LTA group (Supplementary Figure S1).  , and osteoblasts (Obs) were observed on Day 14 after femoral bone defects were introduced in the sham control group. The bone bridge, HCs and Obs appeared earlier in the LTA group, on Day 7 after femoral defects were introduced. An extensive network of primary bone (osteocytes, Ocys) and osteoclasts (Ocls) had formed and woven bone was observed on Day 14 following the introduction of femoral defect in the LTA group. (b) Immunofluorescence was used to detect osterix (an Ob marker) and cathepsin K (an Ocl marker). Intense osterix and cathepsin K signals were observed surrounding the trabecular bones in the LTA-treated group, whereas the signals for these two proteins were diffusely distributed in the PBS group. The white arrows indicate that fewer newly formed bones were present in the sham group, whereas vigorous bone-formation activity was observed in the LTA group.

LTA Had Beneficial Effects on Bone Formation In Vitro
LTA directly enhanced osteogenic factor-induced MC3T3-E1 cell differentiation, including the ALP levels (Figure 3a,b) and calcium deposition (Figure 3c,d). LTA also increased osteopontin secretion during MC3T3-E1 cell differentiation (Figure 3e). During RANKL-induced Ocl activation, LTA inhibited the activity of the Ocl-associated protein, tartrate-resistant acid phosphatase (TRAP; Figure 3f,g). However, LTA did not alter the RANKL-enhanced expression of cathepsin K at either the protein-or mRNA-expression levels (Figure 3h-j).
following the introduction of femoral defect in the LTA group. (b) Immunofluorescence was used to detect osterix (an Ob marker) and cathepsin K (an Ocl marker). Intense osterix and cathepsin K signals were observed surrounding the trabecular bones in the LTA-treated group, whereas the signals for these two proteins were diffusely distributed in the PBS group. The white arrows indicate that fewer newly formed bones were present in the sham group, whereas vigorous bone-formation activity was observed in the LTA group.

LTA Had Beneficial Effects on Bone Formation In Vitro
LTA directly enhanced osteogenic factor-induced MC3T3-E1 cell differentiation, including the ALP levels (Figure 3a,b) and calcium deposition (Figure 3c,d). LTA also increased osteopontin secretion during MC3T3-E1 cell differentiation (Figure 3e). During RANKL-induced Ocl activation, LTA inhibited the activity of the Ocl-associated protein, tartrate-resistant acid phosphatase (TRAP; Figure 3f,g). However, LTA did not alter the RANKL-enhanced expression of cathepsin K at either the protein-or mRNA-expression levels (Figure 3h-j). (h,i) Immunofluorescence staining confirmed that LTA did not affect RANKL-induced differentiation of cathepsin K-positive Ocl-like cells. (j) Quantitative polymerase chain reaction analysis was performed to measure the mRNA-expression levels of cathepsin K. The data are presented as the mean ± standard error of the mean. Analyses were conducted with two-way ANOVA, followed by Bonferroni's post-hoc test. * p < 0.05, ** p < 0.01, *** p < 0.001.

LTA Positively Regulated Bone Healing by Promoting Protein Secretion from Obs
Antibody arrays enable high-throughput analysis of protein-expression levels on a large scale. To examine the possible molecules involved in LTA-enhanced osteogenesis, an antibody array was used. Culture supernatants were collected (with or without LTA pretreatment) and analyzed by performing protein-array assays. Each spot was measured based on changes in the signal intensity to identify potential secretory factors induced by LTA, resulting in enhanced bone formation (Figure (h,i) Immunofluorescence staining confirmed that LTA did not affect RANKL-induced differentiation of cathepsin K-positive Ocl-like cells. (j) Quantitative polymerase chain reaction analysis was performed to measure the mRNA-expression levels of cathepsin K. The data are presented as the mean ± standard error of the mean. Analyses were conducted with two-way ANOVA, followed by Bonferroni's post-hoc test. * p < 0.05, ** p < 0.01, *** p < 0.001.

LTA Positively Regulated Bone Healing by Promoting Protein Secretion from Obs
Antibody arrays enable high-throughput analysis of protein-expression levels on a large scale. To examine the possible molecules involved in LTA-enhanced osteogenesis, an antibody array was used. Culture supernatants were collected (with or without LTA pretreatment) and analyzed by performing protein-array assays. Each spot was measured based on changes in the signal intensity to identify potential secretory factors induced by LTA, resulting in enhanced bone formation (Figure 4a). The intensity of all spots was evaluated with respect to those of the references and negative spots. We found that six of the 111 spots showed changes between different groups and these spots were selected for further intensity measurements (Figure 4b). The expression levels of all the six proteins were higher in the LTA group than in the control group. These six proteins were monocyte chemoattractant protein 1 (MCP-1), chemokine (C-X-C motif) ligand 1 (CXCL1), cystatin C, growth/differentiation factor 15 (GDF15), endostatin and neutrophil gelatinase-associated lipocalin (NGAL), which may be related to LTA-accelerated bone healing during bone remodeling. 4a). The intensity of all spots was evaluated with respect to those of the references and negative spots. We found that six of the 111 spots showed changes between different groups and these spots were selected for further intensity measurements (Figure 4b). The expression levels of all the six proteins were higher in the LTA group than in the control group. These six proteins were monocyte chemoattractant protein 1 (MCP-1), chemokine (C-X-C motif) ligand 1 (CXCL1), cystatin C, growth/differentiation factor 15 (GDF15), endostatin and neutrophil gelatinase-associated lipocalin (NGAL), which may be related to LTA-accelerated bone healing during bone remodeling.

LTA Enhanced Dynamic Bone-Formation Rates In Vivo
To examine the impact of LTA on the dynamic bone-formation rate in vivo, we performed a double-labeling study of the mouse femoral defects, using xylene orange (XO) and calcein green (CG). Newly formed CG-labeled bone cells surrounded the trabecular surfaces in the sham group, whereas these cells were present inside the trabecular bones in the LTA group. Histomorphometric analysis revealed that static bone-formation parameters were significantly different between the LTA and sham groups. A fully connected bone bridge formed in the LTA group, but not in the sham group ( Figure 5). These results demonstrate that the dynamic bone-formation rates in the LTA group were higher than those in the sham group. Furthermore, our results show a higher rate of trabecular remodeling in the LTA group, compared with sham group. In general, labeling with XO quickly faded in the LTA group, but clear labeling was observed in the trabecular bone areas in the sham group. Calcein green labeling was found on the trabecular surface in the sham group, whereas CG signaling was present inside the trabecular region in the LTA group. These results show the potential for LTA to enhance endochondral ossification in mice with femoral bone defects.

LTA Enhanced Dynamic Bone-Formation Rates In Vivo
To examine the impact of LTA on the dynamic bone-formation rate in vivo, we performed a double-labeling study of the mouse femoral defects, using xylene orange (XO) and calcein green (CG). Newly formed CG-labeled bone cells surrounded the trabecular surfaces in the sham group, whereas these cells were present inside the trabecular bones in the LTA group. Histomorphometric analysis revealed that static bone-formation parameters were significantly different between the LTA and sham groups. A fully connected bone bridge formed in the LTA group, but not in the sham group ( Figure 5). These results demonstrate that the dynamic bone-formation rates in the LTA group were higher than those in the sham group. Furthermore, our results show a higher rate of trabecular remodeling in the LTA group, compared with sham group. In general, labeling with XO quickly faded in the LTA group, but clear labeling was observed in the trabecular bone areas in the sham group. Calcein green labeling was found on the trabecular surface in the sham group, whereas CG signaling was present inside the trabecular region in the LTA group. These results show the potential for LTA to enhance endochondral ossification in mice with femoral bone defects.

Discussion
LTA is a cell wall component of GP bacteria that is recognized by toll-like receptor 2 on the cell surface, which then initiates signaling cascades including the NF-κB and MAPK pathways [16]. LTA has been isolated from different species of GP bacteria, and it has various chemical structures and functional activities [17]. Our recent data indicate that PJI caused by GN bacteria indicates a higher risk of aseptic loosening after reimplantation, mainly because of lipopolysaccharide (LPS)-mediated effects on Ocl differentiation [14]. Interestingly, LPS (but not LTA) reduced both the number of trabeculae and the bone-mineral density in mice. Because LTA does not cause aseptic loosening, we were curious about its potential effects on bone healing. In this study, we demonstrated that LTA isolated from Staphylococcus aureus accelerated bone healing in vivo. In our in vitro studies, LTA directly enhanced Ob differentiation and inhibited Ocl activation in MC3T3-E1 and RAW264.7 cell models, respectively. Moreover, LTA stimulated Obs to secrete several proteins, including MCP-1, CXCL1, cystatin C, GDF15, endostatin and NGAL, which may provide a regulatory function through an autocrine mechanism. It was previously reported that the MCP-1 production by Obs in bone specimens from patients with S. aureus-associated osteomyelitis can stimulate the proliferation of osteoblastic cells [18]. It was also shown that Obs could release CXCL1, which attracted Ocl precursors to the bone environment [19]. Cystatin C is synthesized by bone cells and affects bone morphogenetic protein-signaling cascades in osteoblastic cells and then promotes Ob differentiation, mineralization and bone formation [20][21][22]. GDF-15, endostatin and NGAL were previously found to inhibit RANKL-induced Ocl formation [23][24][25][26]. Endostatin attenuated vascular endothelial growth factor-A-induced osteoclastic bone resorption [27]. Based on these findings, we speculate that LTA inhibited Ocl activation and bone resorption through LTA-induced expression of MCP-1, CXCL1, cystatin C, GDF15, endostatin and NGAL. These molecules may be related to LTA-accelerated bone healing during bone remodeling.

Discussion
LTA is a cell wall component of GP bacteria that is recognized by toll-like receptor 2 on the cell surface, which then initiates signaling cascades including the NF-κB and MAPK pathways [16]. LTA has been isolated from different species of GP bacteria, and it has various chemical structures and functional activities [17]. Our recent data indicate that PJI caused by GN bacteria indicates a higher risk of aseptic loosening after reimplantation, mainly because of lipopolysaccharide (LPS)-mediated effects on Ocl differentiation [14]. Interestingly, LPS (but not LTA) reduced both the number of trabeculae and the bone-mineral density in mice. Because LTA does not cause aseptic loosening, we were curious about its potential effects on bone healing. In this study, we demonstrated that LTA isolated from Staphylococcus aureus accelerated bone healing in vivo. In our in vitro studies, LTA directly enhanced Ob differentiation and inhibited Ocl activation in MC3T3-E1 and RAW264.7 cell models, respectively. Moreover, LTA stimulated Obs to secrete several proteins, including MCP-1, CXCL1, cystatin C, GDF15, endostatin and NGAL, which may provide a regulatory function through an autocrine mechanism. It was previously reported that the MCP-1 production by Obs in bone specimens from patients with S. aureus-associated osteomyelitis can stimulate the proliferation of osteoblastic cells [18]. It was also shown that Obs could release CXCL1, which attracted Ocl precursors to the bone environment [19]. Cystatin C is synthesized by bone cells and affects bone morphogenetic protein-signaling cascades in osteoblastic cells and then promotes Ob differentiation, mineralization and bone formation [20][21][22]. GDF-15, endostatin and NGAL were previously found to inhibit RANKL-induced Ocl formation [23][24][25][26]. Endostatin attenuated vascular endothelial growth factor-A-induced osteoclastic bone resorption [27]. Based on these findings, we speculate that LTA inhibited Ocl activation and bone resorption through LTA-induced expression of MCP-1, CXCL1, cystatin C, GDF15, endostatin and NGAL. These molecules may be related to LTA-accelerated bone healing during bone remodeling.
Immune responses and circulating immunity-related factors regulate skeletal cells during the processes of normal and pathological bone formation. Although bacterial factors can trigger immunity that induces pro-osteogenic pathways, these usually pale in significance due to osteolysis and concerns of systemic inflammation. Limited data have been published demonstrating that LTA may affect Obs and Ocls in vitro. Staphylococcal LTA inhibited the phosphorylation of extracellular signal-regulated kinase and c-Jun N-terminal kinase in Ocl precursors, which were treated with macrophage colony-stimulating factor and RANKL, concomitantly with a decreased DNA-binding activity of activator protein 1 [28]. The LTA molecule of Enterococcus faecalis is an Ocl inhibitor that significantly inhibited osteoclastogenesis of BMMs in the presence of RANKL [29]. In contrast, the mRNA-and protein-expression levels of osteogenesis markers were significantly upregulated after treatment with the staphylococcal LTA; enhanced ALP positivity was found in the LTA groups; and calcium nodule formation increased simultaneously [13]. LTA may function as an osteo-stimulatory factor through a synergy with osteoinductive signals [17]. Although these studies mentioned above indicate the impacts of LTA on Obs and Ocls in vitro, no reports have described the effects of LTA on bone healing in vivo. In this study, we demonstrated that LTA not only accelerated bone healing, but also enhanced dynamic bone formation.
Bone lining tissues contain residual osteal-macrophages called osteomacs, which interact with OBs and are located immediately adjacent to OBs; they regulate bone formation and play an important role in regulating the bone healing process [30]. The F4/80(+) Mac-2(−/low) TRACP(-) osteomacs are present within the bone injury site and persisted throughout the healing duration [31]. Osteomacs are required for deposition of collagen type 1 matrix and bone mineralization in a mouse model of tibial injury [31]. Additionally, induction of M2 macrophages through interleukin 4 and 13 significantly enhanced bone formation during fracture healing [32]. Depletion of M1 macrophages was also found to reduce callus properties, alter the cytokine expression profiles during early bone repair and impair the bone healing process [33]. These studies strongly indicate that macrophages influence bone healing.
Osteointegration refers to a process whereby bone cells come into direct contact with an orthopedic implant. Osteointegration includes mesenchymal stem cell attachment, proliferation, and differentiation into Obs on the implant surface, resulting in the formation of mineralized bone around the implant [34,35]. Because of the demand for artificial joint-replacement surgery and the common use of metal implants in orthopedic surgery, osteointegration is becoming a crucial research topic. Approximately 30,000 patients require joint-replacement surgery in Taiwan per year. For successful total-joint arthroplasty, the long-term fixation of implants (due to osteointegration) is crucial. Additionally, to improve the efficacy of other orthopedic surgical procedures, the metal implants used in many orthopedic treatments (such as fracture surgery, rod implantation and chiropractic therapy) require osteointegration. For example, posttraumatic bone fractures are usually fixed with implanted devices to correct the position of bone fragments and to assist in the healing process. Therefore, applying LTA on the surface of an orthopedic implant to improve osteointegration may represent a novel and viable treatment strategy. Moreover, new strategies to enhance bone healing are still needed. We believe that LTA may contribute to the process of bone healing. Using LTA to promote bone healing has the following advantages. First, because LTA comes from a GP bacterial cell wall, it is easy to obtain. A supplier would only need to culture a large number of bacteria and collect LTA from the cell wall. Consequently, the manufacturing cost should be relatively low. Finally, if LTA is used as a clinical drug, it should be easy for suppliers to provide LTA at an appropriate dosage and purity [36]. In conclusion, our findings suggest that LTA shows promise as an immuno-modulatory biomaterial for bone regeneration.

Experimental Animal Study
All animal procedures complied with the National Institute of Health guidelines and were reviewed and approved by the local Hospital Animal Care and Use Committee. Initially, 10-week-old male C57BL/6 mice were anesthetized via intraperitoneal injection (0.01 mL/kg body weight) of a 1:1 (vol/vol) mixture of tiletamine-zolazepam (Zoletil; Virbac, Carros, France) and xylazine hydrochloride (Bayer HealthCare AG, Leverkusen, Germany), and the surgical site was shaved and disinfected with povidone-iodine. An incision was made in the skin overlying the right knee joint. A medial parapatellar arthrotomy (with lateral displacement of the quadriceps-patella complex) was performed to access the distal femur. After locating the femoral intercondylar notch, the femoral intramedullary canal was manually pierced with a 25-gauge needle and intrafemorally injected with 10 mg/kg LTA (from S. aureus; Sigma-Aldrich, St. Louis, MI, USA) in phosphate-buffered saline (PBS; 10 µL). A stainless-steel rod (with a length of 0.9 mm and diameter of 0.4 mm) was surgically placed in a retrograde manner. A 1-mm defect was formed at the midshaft of the right femur using drill bits of different sizes. The quadriceps-patellar complex was repositioned to the midline, and the surgical site was closed with subcutaneous 6-0 Dexon sutures. Buprenorphine (0.2 mg/kg) was administered subcutaneously every 24 h as an analgesic throughout the experimental duration. The mice were sacrificed on Day 7 or 14 post-surgery. The femur was immediately fixed in formaldehyde (10%) and subjected to micro-CT analysis.

Micro-CT Bone Imaging
Nondestructive ultrastructural analysis was performed using a SkyScan 1176 micro-CT scanner (Bruker Microct, Kontich, Belgium). The samples were wrapped in saline-soaked gauze and scanned at 50 kV with a 0.5-mm aluminum filter. Images with a resolution of 9 µm were reconstructed using GUP-NRecon software (version 1.7.4.2) and analyzed using CTAn software (version 1.15.4.0, SkyScan). The grayscale was based on Hounsfield units, and validated calcium standards were scanned as a density reference (0.25 and 0.75 g/cm 3 Hydroxyapatite Phantoms). The BV, TV and BV/TV ratio (expressed as a percentage) were calculated for the sections of interest (n = 117), which were located between 1 mm above and below the damaged site, comprising all the woven callus and remodeled bone. A 3D image was constructed using CTVox software (version 3.3.0, SkyScan) for illustration purposes.
MC3T3-E1 cells (CRL-2593, American Type Culture Collection, Manassas, VA, USA) were plated in a 12-well plate (density: 5 × 10 3 cells/well) and then cultured in 10% FBS-differentiation medium with or without osteogenic factors (OS). The OS was composed of 5 mM glycerol 2-phosphate, 0.1 µM dexamethasone, and 50 mM ascorbic acid. The cells were treated with or without 100 ng/mL LTA in the osteogenic differentiation medium. Calcification of MC3T3-E1 cells was assessed via Alizarin Red S staining (ScienCell, Carlsbad, CA, USA) to monitor matrix mineralization. Calcium assays were performed using a Calcium LiquiColor Assay (Stanbio laboratory, Boerne, TX, USA) in accordance with the manufacturer's instructions. A Fast Violet B Salt capsule (catalog number 851-10 CAP, Sigma-Aldrich) was dissolved in naphthol AS-MX phosphate alkaline solution (catalog number 855, Sigma-Aldrich) and then used for ALP staining. Sigma 104 ® phosphatase substrate (catalog number 104105, Sigma-Aldrich) was used for ALP assays [37].

Enzyme-Linked Immunosorbent Assay (ELISA)
Osteopontin levels were measured using a Mouse Osteopontin DuoSet ELISA Kit (DY441, R&D Systems, Minneapolis, MA, USA), according to the manufacturer's protocol. Thresholds for intraand inter-assay coefficients of variation were set at <15%.

Protein Array
We used MC3T3-E1 cell culture medium to survey factors with protein array panels containing 111 well-categorized monoclonal antibodies to compare LTA-specific protein-expression patterns in two groups (with and without LTA). Proteome Profiler Antibody arrays (catalog number ARY028; R&D Systems) were used to simultaneously evaluate the expression levels of multiple factors, including cytokines, chemokines and soluble receptors in the cell culture medium of MC3T3-E1 cells. The signal intensities of spots in the protein array were quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Fluorochrome Labeling to Measure Dynamic Bone-Formation Rates
Double-fluorochrome in vivo labeling was performed using XO and CG, as described [38]. After introducing femoral bone defects, the mice were subcutaneously injected with XO (80 mg/kg body weight, Day 7) and CG (10 mg/kg body weight, Day 12), after which they were sacrificed on Day 14. The incorporation of fluorochromes into the undecalcified bone sections on glass slides was examined by fluorescence microscopy on methyl methacrylate-embedded sections (40 µm thickness).

Statistical Analysis
All data were obtained from at least three independent experiments. Quantitative data were analyzed with two-way analysis of variance (ANOVA), followed by Bonferroni's post-hoc test, and are presented as the mean ± standard error of the mean. Body weights were analyzed using two-way repeated-measures ANOVA, followed by Tukey's post-hoc test. Categorical variables were examined with GraphPad Prism software, version 7.0 (GraphPad Inc., San Diego, CA, USA). Two-tailed p values <0.05 were considered to reflect statistically significant differences.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.