Mechanisms underlying heterologous skin scaffold-mediated tissue remodeling

Biocompatibility of two newly developed porcine skin scaffolds was assessed after 3, 14, 21 and 90 days of implantation in rats. Both scaffolds showed absence of cells, preservation of ECM and mechanical properties comparable to non-decellularised skin before implantation. Host cell infiltration was much prominent on both scaffolds when compared to Permacol (surgical control). At day 3, the grafts were surrounded by polymorphonuclear cells, which were replaced by a notable number of IL-6-positive cells at day 14. Simultaneously, the number of pro-inflammatory M1-macrophage was enhanced. Interestingly, a predominant pro-remodeling M2 response, with newly formed vessels, myofibroblasts activation and a shift on the type of collagen expression was sequentially delayed (around 21 days). The gene expression of some trophic factors involved in tissue remodeling was congruent with the cellular events. Our findings suggested that the responsiveness of macrophages after non-crosslinked skin scaffolds implantation seemed to intimately affect various cell responses and molecular events; and this range of mutually reinforcing actions was predictive of a positive tissue remodeling that was essential for the long-standing success of the implants. Furthermore, our study indicates that non-crosslinked biologic scaffold implantation is biocompatible to the host tissue and somehow underlying molecular events involved in tissue repair.

by an equally complex signaling network involving numerous growth factors, chemokines and cytokines 13,14 . Transforming growth factor (TGF)-β , platelet-derived growth factor (PDGF), basic fibroblast growth factor (bFGF), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF) and the interleukin (IL) family 12 are of particular interest amongst innumerous proteins present in the wound milieu.
The first phase of wound healing begins with capillary damage and clot formation. It is then followed by a number of events, including the production of pro-inflammatory cytokines and chemokines 15 , recruitment of inflammatory cells such as neutrophils and macrophages 16 , and also fibroblasts and endothelial cells 17 . In the second phase (proliferative phase), the process of neovascularization with delivery of nutrients and oxygen to the wound bed will contribute to fibroblast proliferation 14 . The fibroblasts become activated and, triggered by TGF-β 1, acquire a smooth muscle cell-like phenotype called myofibroblasts 17,18 . These cells exhibit contractile properties, due to the expression of alpha -smooth muscle actin (α -SMA) in microfilament bundles, playing a major role in the contraction and maturation of granulation tissue 19 . During the remodeling phase, the number of vascular cells and myofibroblasts are dramatically reduced by apoptosis 20 . In this scenario, the key orchestrators and effectors of the remodeling progression -macrophages -are known to alter their function and phenotype to meet the needs of the healing tissue 21 .
During the proliferative phase, macrophages assume a wide spectrum of different functional phenotypes that can influence repair 22 . The exact mechanisms are still poorly understood, but it has been thought that the macrophage phenotype switch during tissue remodeling may be driven, in part, by phagocytosis of tissue debris and apoptotic cells 21 . It has also been observed that transition between the pro-inflammatory (M1) to the anti-inflammatory (M2) phenotype is followed by a decrease in VEGF and pro-inflammatory cytokine production (IL-1, IL-6), with an increase in IL-10 secretion 23,24 and TGF-β 25 .
In the context of regenerative medicine, the effect of macrophage phenotypes upon biologic scaffold remodeling has already been shown 26 . Our study however discuss the pleiotropic effect of macrophages upon different cell types and molecular responses evaluated by the host tissue reaction to two non-crosslinked skin scaffolds tailored by different decellularisation processes and one commercially available crosslinked dermal matrix well established for soft-tissue repair.

Results
Evaluation of scaffolds. Both decellularisation methods (described in Table 1) used to produce the acellular matrices proved to be time (< 3 days) and cost-efficient, producing 'off-the-shelf ' skin scaffolds, which is very desirable in regenerative medicine. Following decellularisation, samples were assessed histologically using HE-stained sections. Figure 1A shows the histological structure of native porcine skin, whereas the scaffold revealed the complete removal of cells and nuclear material from the matrix (Fig. 1B,C). The collagen fibers of the scaffolds showed the classical collagen fiber architecture and distribution (Fig. 1E,F), when compared to the non-decellularised control skin (Fig. 1D). Furthermore, DNA quantification showed that approximately 90% of nuclear material was depleted after both decellularisation protocols, with no visible DNA on the electrophoresis gel following staining by ethidium bromide (Fig. 1G,H).
The mechanical properties were evaluated in all the groups by tensile strength test, measured as the maximum stress that each sample could be stretched to withstand before breaking. The results showed that both decellularisation processes did not change the mechanical properties of the scaffolds when compared to the control group (Fig. 1K,L).
Histological analysis of acellular matrix and host tissues: biocompatibility. The porcine skin scaffolds (1 and 2) and Permacol were implanted subcutaneously in rats and its biocompatibility was assessed on the 3°, 14° and 21° days. Histological analysis indicated that the implants were biocompatible, with no evidence of rejection or fibrosis. At day 3 of implantation, both scaffolds 1 and 2 were surrounded by an accumulation of polymorphonuclear (PMN) cells ( Fig. 2A,E -insets); and from day 14, cell population was replaced by mononuclear cells, most likely macrophages and fibroblasts/ myofibroblasts distributed throughout both scaffolds ( Fig. 2B-D,F-H -insets). Histologic appearance of the Permacol (used as surgical control) at day 3 after surgery showed a mild reaction of the host tissue and the cell infiltration was limited almost exclusively to the edge of the matrix, and slowly increased up to 90 days ( Fig. 2I-L).
The collagen fibers of the scaffolds and Permacol were observed by picrosirius-haematoxylin polarization technique. At day 3 post-implantation, all implants displayed a uniform distribution of collagen fibers, which showed to be strongly birefringent under polarized light (Fig. 3A,E,I). In contrast, there was a significant decrease in collagen at the site of scaffold implantation at day 14 ( Fig. 3B,F) when compared to day 3 (Fig. 3A,E), showing degradation and rearrangement of the matrices. Histologic appearance of a disorganized connective tissue from day 14 to 21 ( Fig. 3B,C,F,G) and a significantly increased collagen deposition at day 90 (Fig. 3D,H), suggested that remodeling of connective tissue was taking place. The implanted Permacol matrices, however, were still identifiable throughout the study (Fig. 3I-L). Cellular changes after scaffolds implantation: immunohistochemistry. The angiogenesis was analyzed by immune staining of the multimeric glycoprotein Von Willebrand factor (vWF) (Fig. 4A-C) in all scaffolds. The newly formed vessels were seen from day 14, with a significant increase at day 21 (Fig. 4D). While in Permacol there was no increase to over time (day 3 = 0; day 14 = 0; day 21 = 0.6 ± 0.2; day 90 = 2.5 ± 0.6).
The ratio between M1/M2 phenotypes was calculated and showed in the Fig. 5I. Although the initial response to the scaffolds implantation showed a predominant inflammatory reaction (with values consistently above 1.0, which indicates a M1 macrophage response), the host tissue reaction seemed to equalize and displayed a prominent M2-profile thereafter.
The marker genes for collagen remodeling COL1A1, COL1A2 and COL3A1 displayed an antagonist behavior. A marked increase on the collagen 3 ( Fig. 6D) was seen at day 14 with a substantial decrease on day 21, when collagen 1 (COL1A1 and COL1A2; Fig. 6E,F) displayed its maximum expression (~8-fold increase in COL1A1 and 7-fold increase in COL1A2).
Permacol did not stimulate significantly any of those classical markers of wound healing during the studied time points (Fig. 6A-F).

Discussion
Biologic scaffolds are commonly used to promote repair and restoration of functional tissues 27 . Although most, if not all of these scaffolds are composed of ECM or components of ECM, the clinical and tissue remodeling outcomes may differ greatly 26 . The bioactivity of scaffolds can be altered depending on the type of decellularisation reagents and processes used, eliciting distinct host-tissue morphologic and molecular responses. Models of wound healing in rodents can provide reliable and reproducible information on the response of wound contaction to the experimental therapy 28 . Given the remarkable rate of advancement in wound healing in animal species, research may be able to develop more incisive models of the healing process and its key steps 28,29 .
In this study, we used a rat model to evaluate the cellular and molecular response to implantation of two decellularised non-crosslinked porcine skin scaffolds in comparison to a commercially available crosslinked dermal scaffold (Permacol). Permacol has been used for years as a surgical implant, providing durability and strength for ventral hernia repair and abdominal wall reconstruction 30,31 .
The surgical sites at which the scaffolds were placed showed a dense infiltration of polymorphonuclear (PMN) and some mononuclear cells, which persisted up to day 21. A possible systemic reaction was discarded by measuring leukocytes and inflammatory mediators in the blood of the animals after implantation (Supplementary Figure 1).    Both scaffolds displayed an uniform collagen remodeling when observed under polarized light (picrosirius-haematoxylin staining), but not the HDMI-crosslinked porcine dermal scaffold Permacol. Cross-linking agents, such as 1,6-hexamethylene diisocyanate (HDMI), are included in the processing of biologic materials to impart strength and to slow the degradation rate once implanted. Permacol being a crosslinked matrix could be distinctly seen with little reorganization of host connective tissue, and with insignificant infiltration of cells (IL-6-positive, macrophages and myofibroblasts) and blood vessels up to day 90. Our results are in agreement with other studies demonstrating that treatment with any crosslinking reagents can change the collagen architecture at a molecular level 32 . These molecular alterations impact directly on cell surface interactions 33 , and can change the quantity and identity of secreted pro-inflammatory cytokines and chemokinesm 34 , the gene expression pattern 35 , and downstream remodeling events 36,37 . Kulig and collaborators (2013) showed that various non-crosslinked scaffolds commercially available, such as Alloderm, Gelfoam and Stattice, have lower cell migration and proliferation when compared to Permacol, despite of the latter being a crosslinked matrix. These results suggest that biologic and structural properties of both type of scaffolds need to be carefully studied to assess new products and its uses in this promising class of biomaterials 38 .
The non-crosslinked scaffolds 1 and 2 showed a similar degree of biocompatibility, with a pronounced cell infiltration from day 3 and much higher biodegradability when compared to Permacol. The angiogenic response was initiated on day 14, and at day 21 the number of newly formed vessels, visualized by immunostaining of von Willebrand factor (vWF), had significantly increased.
Although the non-crosslinked scaffolds 1 and 2 showed a similar degree of biocompatibility, a slightly delayed integration with host tissue, cellular infiltration and biodegradability was seen on scaffold 2 when compared to scaffold 1. Although these differences were not significant, they may be attributed to the loss of the glycosaminoglycans (GAGs) content evidenced in the scaffold 1, probably caused by the reagents used in the decellularisation process. It has been shown that Triton-X, especially when combined with other detergents or enzymatic digestion (included in the manufacturing of scaffold 1), can lead to a certain loss on ECM components, resulting in early tissue degradation 33 .
The long-term remodeling outcome differs greatly amongst scaffold materials 26 . The maintenance of the ECM structure and biomechanical properties of the non-crosslinked scaffolds supported a satisfactory tissue remodeling when compared to the cross-linked Permacol. Nevertheless, depending on the particular application, it may be advantageous to use crosslinked scaffolds to limit the degree of subsequent scaffold degradation, which is crucial when scaffolds are intended to be used as permanent and non-absorbable matrices.
The temporal sequence of remodeling events of extracellular matrix devices, including the speed of scaffold degradation and the extent of new-tissue deposition by the host, may be predictive of the course of wound healing and outcome of the procedure 27 . In order for wounds to heal timely and properly, there must be a fine balance of interaction between various cell types, cytokines, growth factors, proteases and ECM components. Coordinated efforts of cells including leukocytes 16 , platelets, fibroblasts, endothelial cells 14 and macrophages 39,40 occur in order to promote formation of new tissue and wound closure.
In commonly used experimental models, the inflammatory phase is generally limited to the first few days after injury. Within hours after acute injury, inflammatory cells begin to accumulate at the affected site 41 . In common experimental models using healthy young male mice, neutrophil accumulation peaks around day 1 and returns to uninjured levels by days 5 to 10 after skin wounding 42 . The proliferative phase begins within days after injury and peaks within 1 week after injury; and the remodeling phase begins as proliferation starts to subside, and can last for months or more 21 .
Our results however indicated that, after skin scaffolds implantation, the remodeling process was somehow delayed. This could be attributed to an increased initial inflammatory response, which is clearly seen with an extensive infiltration of PMN cells and a concurrent amplified IL-1β mRNA expression, which showed a return to baseline conditions only after 14 days. This could have been caused by residues from reagents or components of the implanted scaffolds that could have directly or indirectly affected the cellular events. Collagen-based matrices have been used in several models for different medical applications 43 , but despite the desired effect, these materials can cause a relatively severe local inflammatory reaction after implantation, depending on the structure and functional characteristics of the material 44 . On the other hand, at a later stage our results did not show any complications related to the delayed proliferative phase, such as persistent inflammation or low-grade chronic inflammation, poor tissue organization, presence of foreign-body giant cells or fibrous encapsulation as shown by other authors 45,46 .
The different phases of wound healing is strictly regulated by multiple growth factors and cytokines 14 that regulate migration, proliferation, and differentiation of cells 47 . Of particular interest, IL-6 has been shown to be important in the remodeling response 12 . Our results showed that IL-6 positive cells were increased at day 14 and successively decreased thereafter. Those positive cells have the morphology of large mononuclear cells and we suggest that they may be monocytes that migrated into the tissue and started the transition into macrophage phenotypes. Sequential activation of classic, pro-inflammatory, M1 macrophages and alternatively activated or anti-inflammatory M2a, M2b, and M2c macrophages occurs during normal healing and facilitates transitions through the inflammatory, proliferative, and remodeling phases of repair 22 . Although morphologically indistinguishable using routine methods of examination, macrophages phenotypes can be identified and distinguished according to their cell surface markers and their cytokine and gene expression profiles 26,48 .
M1 phenotype (CCR7+ cells) accumulation occurred in our study on day 14. This is divergent with other studies showing that the M1 peak is somewhat earlier, around days 3 to 7, and declines significantly by days 10 to 14, although low levels of macrophage accumulation may persist for weeks 41,49,50 .
Depletion of macrophages during the inflammatory phase reduces granulation tissue formation and cell proliferation in mouse skin wounds 39 and prevents formation of new myofibers in injured muscle 51  Although the exact mechanisms by which macrophages regulate these processes remain to be determined, cytokine production may be one mechanism by which macrophages of the inflammatory phase initiate transition into the proliferative phase of healing 26 .
Interestingly, in our study we showed a timely gene expression shift marking this phenotype transition, which is in agreement with other authors 24,52,53 . mRNA expression of the vascular growth factor (VEGF-A) which showed to be strongly up-regulated until day 14, started subsiding as the transforming growth factor beta 1 (TGF-β 1) gene expression began increasing. This was mirrored by an augmented number of CCR7+ (M1 macrophages) at day 14 and a low number of CD163+ cells (M2 macrophages) observed at the implantation site. After day 14, however, the M1/M2 ratio indicated a predominant M2 response in both tested scaffolds.
Many growth factors and cytokines secreted by macrophages have pleiotropic influence on cell proliferation, angiogenesis, and extracellular matrix synthesis. To mention few, VEGF-A was shown to be strongly induced after cutaneous injury, with keratinocytes and macrophages being the major producers 53 . Its major function in wound healing is to promote cell differentiation and angiogenesis 39,54 . IL-6, also secreted by macrophages, initiates the healing response 55 and induces regulation of proliferation of myofibroblasts 24,56 . And also, not less importantly, the TGF-β 1 that is known as the signature mediator of myofibroblasts differentiation 57 .
For successful wound healing, it is important that inflammation, angiogenesis and reepithelialization processes take place 14,58,59 . Non-crosslinked porcine skin scaffolds may allow the sequence of wound healing events to take place leading to an enhancing successful reepithelialization in skin wound models. According to Hoganson and collaborators (2010) the retention of a wide array of ECM proteins and cytokines makes those non-crosslinked matrices very promising assets showing potential advantages in some surgical applications. The retention of angiogenic growth factors and TGF-β suggests that those non-crosslinked matrices may support excellent fibroblast infiltration and collagen production when used for hernia repair and other surgical reconstruction applications 60 .
In our study, the mRNA VEGF-A/TGF-β 1 switch seemed not only to hallmark the M1/M2 transition but also to overlap actions on myofibroblast differentiation, neovascularization and synthesis of different types of collagen. It has been shown that collagen deposition in skin wounds appears to be controlled by monocytes/macrophages of the inflammatory phase, rather than macrophages of later phases 39,61 . Although pronounced transcription of COL1α 1 and COL1α 2 can be up-regulated as early as 2 to 3 days after injury 42,61 , our results showed that those genes were up-regulated at day 21. Nonetheless, studies show that both early mRNA and late protein levels of collagen, when reduced by depletion of monocytes/macrophages during the inflammatory phase, seem to result in delayed healing but reduced scar formation 52 .
One of the characteristics of wound remodeling is the change of extracellular matrix composition. We therefore assessed the expression of the collagen marker genes COL1A1, COL1A2 and COL3A1 in response to skin scaffolds implantation and the results showed that the transition from collagen type III to I occurred around day 21, when we also see an increased number of myofibroblasts in the site of implantation.
Myofibroblasts are the predominant mediator of the contractile process because of their ability to extend and retract 14 . During granulation tissue formation, fibroblasts are gradually modulated into myofibroblasts, which are characterized by increased expression of smooth muscle differentiation markers such as alpha-smooth muscle actin (α -SMA), smooth muscle myosin, and desmin starting after the first week, and reaching a peak after day 15 62 .
Although there are many interconnected cells and molecular events to be further explored in more detail, our results suggest that modulation of myofibroblasts and ECM remodeling in the implantation site seems to be timely associated with the M2 response.
In the context of biomaterials, the large spectrum of events that underlines tissue homeostasis showed to be regarded, at least in part, to the well-known heterogeneity of macrophage phenotypes 26,39 . The relatively late responsiveness of those cells after non-crosslinked biological scaffolds implantation, in our study, seems to intimately affect various other cell responses; and this range of mutually reinforcing actions was predictive of a positive tissue remodeling result.
The classical view of wound healing shows a timed sequence of cellular events directed largely by cytokines and growth factors that predicate the functional outcome of the newly formed tissue. Our study suggested that non-crosslinked biologic scaffold implantation somehow delayed the sequence of cell responsiveness and underlying molecular events involved in tissue repair. The late macrophage response seemed to modulate a network of cellular events and molecular feedback, promoting a favorable remodeling response that is essential for long-standing success of the implant. Our data showed optimum biological properties of the non-crosslinked scaffolds, which may be useful tools as skin substitutes in reconstructive surgeries, and can also reduce morbidity, improving functional outcomes in clinical settings 6 .

Materials and Methods
Dermis harvesting and scaffold production. Fresh porcine skin was obtained from Large-White/ Landrace crossbreed pigs (6 months old) after euthanasia. Skin was cleansed with soap, shaved and washed with water and iodine based solution (10% w/w Cutaneous Solution -Iodinated Povidone, Videne, Garforth, UK). The intact skin (epidermis and dermis) was dissected from the animal's flank, washed in sterile phosphate buffered saline (PBS, Sigma-Aldrich, Dorset, UK) with an antibiotic/ antimycotic solution (AA; Sigma-Aldrich, Dorset, UK) and stored in sterile plastic bags at − 20 °C for 24-48 h. The skin samples were defrosted, cut (2 × 2 cm), and allocated randomly into 2 groups for the production of decellularised scaffolds 1 and 2 (Table 1). This study was approved by Ethics Committee in Animal Experimentation of São Paulo State University (UNESP) (protocol no. 062/2012 -CEUA) and was performed according to the regulatory guidelines of the UK Home Office.
Histological analysis of scaffolds. Samples from scaffolds (1 and 2) and Permacol (TSL, Hampshire, UK), a commercially available cross-linked porcine acellular dermis used as surgical control, were fixed for 24 h in 10% NBF solution. They were washed, dehydration in graded alcohol, embedding in paraffin wax and sectioned at 5 μ m. Separate sections were stained with Haematoxylin and Eosin (HE) and Picrosirius-haematoxylin (picrosirius-haematoxylin stained sections were analyzed under polarized light to assess degradation and tissue remodeling at the implantation site).

DNA quantification and gel electrophoresis of scaffolds. Samples were cut into small pieces, weighed
and stored aseptically in DNA-free tubes. DNA was extracted following manufacturer's instructions of GenElute mammalian genomic DNA miniprep kit (Sigma-Aldrich, Dorset, UK). Total DNA was quantitated by measuring the absorbance in a nano-drop spectrophotometer (NanoDrop ND1000, Thermo Scientific, Wilmington, USA), and the absolute amount of DNA per milligram of tissue was calculated.
The size, quality and purity of the extracted DNA were determined by electrophoresis. A 1.2% agarose (Agarose Type I, low EEO, Sigma-Aldrich, Dorset, UK) gel with 1x Tris-borate-ethylenediaminetetraacetic acid (TBE -Bio Reagent, 10x, Sigma-Aldrich, Dorset, UK) running buffer was run at 4 to 5 V/cm between the electrodes. Equal volumes of DNA (5 μ L) and 1 μ L of loading buffer (5x DNA loading buffer, Yorkshire Bioscience Ltd., York, UK) were loaded into each well. Visualization was achieved by staining with 1% of ethidium bromide and DNA samples were measured via ultraviolet transllumination against a 1-kb DNA ladder (Q-Step 4 quantitative DNA ladder, Yorkshire Bioscience Ltd., York, UK).

Glycosaminoglycan (GAG) quantification of scaffolds.
To quantify GAG content on both scaffolds and control (non-decellularised skin), the Blyscan GAG assay kit (Biocolor, Carrickfergus, Northern Ireland) was used 33 . The absorbance was measured using a plate reader (Versamax, Molecular Devices LLC, USA) at 656 nm and the absolute GAG content was calculated per milligram of tissue.
Collagen quantification of scaffolds. Collagen was quantified using Sircol collagen assay kit (Biocolor, Carrickfergus, Northern Ireland) 33 . Extracts were placed in a 96-well plate in triplets and spectrophotometric readings were taken at 555 nm on a microplate reader (Versamax, Molecular Devices LLC, USA). Absolute values were attained with a standard curve composed of type I bovine skin collagen solution (0.5 mg/mL) in the range of 5-100 μ g per 0.1 mL. Total collagen was normalized per milligram of tissue.
Biomechanical test of scaffolds. Mechanical properties of non-decellularised skin control and scaffolds 1 and 2 (n = 5) were evaluated by a tensile strength test measured as the maximum stress that each sample could with stand whilst being stretched before reaching breaking point. For each test, specimens were cut into "dumbbell" shapes using a standard mold with 2.5 cm in length and 0.4 mm in width. The tests were performed with the application of uniaxial tension using a tensiometer (InstronInspec 220 Benchtop Portable tester, Instron, Buckinghamshire, UK). Samples were clamped into the holders and loaded at a constant tension rate. The tensile tester recorded the real-time load and elongation to which the tissue was subjected. Parameters such as maximum load (N), testing time (sec) and extension at maximum load (mm) were recorded. Strain was defined as rate of change of sample deformation and calculated as a ratio of the length at the maximum load to the original length; and stress was defined as how much load was applied at each meter squared sample. Young's modulus (YM; MPa) was calculated by the ratio of stress (N/m 2 ) over strain (%). Experimental Animals. Male Wistar rats (250-300 g body weight), aged from 6 to 8 weeks old, were randomly distributed into 9 groups (n = 5 per group) and kept in a 12 h light/dark cycle with food and water ad libitum. Study Design. Both scaffolds and the control Permacol were cut (1 × 0.3 cm) and kept in saline prior to implantation. The animals were anesthetized with intramuscular administration of ketamine (100 mg/kg) and xylazine (10 mg/kg). The skin was shaved and cleaned using iodine based solution (10% w/w Cutaneous Solution -Iodinated Povidone, Videne, Garforth, UK). Ventral incisions (0.5 cm) were performed just inferior to the costal margin bilaterally. The subcutaneous tissue was dissected with scissors to expose a 'pocket' of approximately 1 cm where the scaffolds were placed. The scaffolds were positioned and fixed subcutaneously.
After 3, 14, 21 and 90 days after implantation, the animals (n = 5 for both tested scaffold and Permacol) were weighed and euthanized with an intra-peritoneal injection of sodium pentobarbitone. The scaffolds (Sc) as well as adjacent areas were collected and stored either in -80 °C or neutral buffer formalin (NBF).
Processing of skin fragments for histological analysis. The samples were processed for histology as described in "Histological analysis of scaffolds" session. Picrosirius-haematoxylin was used to enhance the natural birefringence of collagen and allow quantification of collagen fibers in the implantation site. HE staining was used for general assessment of the host tissue, implant integration, as well as general tissue reaction. An average of 5 fields were randomly selected and analyzed by two independent individuals using the commercial software Image-Pro Plus 6.0.
Immunohistochemical analysis. Sequential fragments of each specimen were cut at 4 μ m intervals.
Representative sections of each sample were prepared for immunohistochemical staining by deparaffinization with xylene and rehydration through a graded ethanol series. Sections were submitted to an antigen retrieval step at 96 °C for 30 min, according to Table 2. The endogenous peroxide activity was blocked with 3% hydrogen peroxide for 30 min followed by a blocked with 10% BSA/PBS and an incubation with primary antibodies. Some sections were incubated with 1% BSA instead of the primary antibody to provide a negative control for the reaction. After washing, the sections were incubated with a secondary biotinylated Ab (Histostatin ® Bulk Kit, Invitrogen, Frederick, MD, USA). Positive staining was detected using a peroxidase-conjugated streptavidin Scientific RepoRts | 6:35074 | DOI: 10.1038/srep35074 complex and the color was developed using the DAB Substrate Kit (Invitrogen, Frederick, MD, USA). Finally, the sections were washed in distilled water, counterstained with haematoxylin. To assess neovascularization, the number of blood vessels stained for Von Willebrand factor (vWB) in the implanted scaffolds quantified and showed as mean ± SEM of the positive vessels per cm 2 (5 random fields using a high-power objective x40). For all other immune-positive cells, data were expressed as cells per cm 2 counted in 5 randomly selected fields. All sections were analyzed using an Axioskop 2-Mot Plus Microscope (Carl Zeiss, Jena, GR), and the AxioVision software for quantitative analysis.
The details of antigen retrieval, washing buffers and primary antibodies dilution are described in the Table 2.

Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR). Fluorogenic
qRT-PCR-based (TaqMan) assay was used to detect amplification of the target genes. Briefly, total RNA was extracted using a commercially available kit (Qiagen RNeasy Mini Kit; Qiagen, Hilden, Germany), according to the manufacturer's instructions, with the following modifications to minimize RNA degradation by abundant skin RNAses. Samples were homogenized using bead-beating technology (Precellys, Bertin Technologies, Montigny-le-Bretonneux, France). Proteins potentially interfering with RNA isolation were removed by incubating the homogenate in 590 μ L distilled water and 5 μ L Proteinase K solution 20 mg/mL (Life Technologies, Paisley, UK) at 55 °C for 10 min then centrifuged at room temperature for 3 min. Supernatants were combined with 0.5 volumes of ethanol (96-100%) into a Rnase-Dnase free tube and RNA was isolated through a RNeasy mini column. The concentration and purity of the RNA was analyzed using the Nandrop ND-1000 (NanoDrop Technologies, Wilmington, DE). All RNA samples were examined for their purity. The absorbance ratio at A260/A280 nm of all samples was ranged from 1.85 to 2.1, indicating all the samples were pure during the RNA extraction procedure.
Complementary DNA (cDNA) was obtained by reverse transcription (RT) of 1 μ g of total RNA, with the Superscript III reverse transcriptase system (Invitrogen, Carlsbad, CA, USA) following the manufacturer's protocol and using oligo(dT)15 as a primer. Real-time PCR was performed with the Eco Real-Time PCR System (Illumina, San Diego, CA, USA). The following amplification profile was used: UDG Incubation 50 °C for 2 min, AmpliTaq Gold 95 °C for 10 min, PCR 40 Cycles -95 °C for 15 sec and 60 °C for 1 min. For each reaction, a total volume of 20 μ L was used, which consisted of 9 μ L of diluted cDNA (10 ng/μ L of RNA), 10 μ L of TaqMan Gene Expression Master Mix (2X) and 1 μ L of TaqMan Gene Expression Assay (20X) (Applied Biosystems, CA, USA). Commercially available primers (IL-1ß, Rn00580432_m1; TGF-ß1, Rn01475963_m1; VEGF-A, Rn01511604_m1; COL1A1, Rn01463870_g1; COL1A2, Rn00584426_m1; COL3A1, Rn01437683_m1 -Applied Biosystems, CA, USA) were used to probe for target mRNA.
Previous results showed that Hprt1 (Hypoxanthine phosphoribosyltransferase 1), CycA (Cyclophilin A), ACTs (actins) and Gapdh (Glyceraldehydes-3-phosphate dehydrogenase) are suitable potential reference genes for rats 63 . We therefore tested 3 of those potential reference genes (CycA, Actb and Gapdh) for validation of their gene expression stability and suitability as internal controls according to the requirements of the MIQE Guidelines. Our preliminary tests for the reference genes showed that although some genes that are implicated in basal cell metabolism may not be stable upon different conditions 64 , Gapdh amongst the other reference genes displayed the narrowest expression range with cycle threshold values between 23 and 25 (Ct = 24 ± 1.5). This gene had a very stable expression in all the experimental conditions along with comparable expression characteristics to target genes. For this reason the use of Gapdh alone as a reference gene shown in our study was sufficient. mRNA was then normalized relative to Gapdh (GAPDH, Rn01462662_g1; Applied Biosystems, CA, USA), which was subsequently used to calculate expression levels of the target genes. The comparative Ct method was used to measure the gene transcription in samples 65 . Results were expressed as relative units based on calculation of 2-Δ Δ Ct, which gives the relative amount of target gene normalized to endogenous control (GAPDH) and to the control (sham-operated) samples with the expression set as 1. Negative controls were either RT without enzyme or PCR without cDNA template.

Statistical analysis.
Results were shown as mean ± SEM. For samples with normal distribution, analysis of variance (ANOVA) was used for repeated measures followed by the Bonferroni test. The Kruskal-Wallis test followed by Dunn's test was used for samples with non-normal distribution. Statistical significance was defined as p < 0.05.

Antigen retrieval
Washing buffer Primary antibody reference