Human lung fibroblast WI-38 (CCL-75, ATCC) were seeded at 2 × 10⁴ cells/cm² into a T-175 flask and cultured for five to seven days until confluence (approx. 9 × 10⁴ cells/cm²) was reached in Dulbecco's Modified Eagle medium (DMEM; #11995-065, Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco) and 1% antibiotics (#15140122, Gibco). The medium was changed every 3 days. After reaching 100% confluency, the cells were washed twice with Dulbecco’s phosphate-buffered saline (DPBS; Sigma-Aldrich) and lysed with a 0.25% Triton X-100 solution (#H5141, Promega) containing 10 mM NH₄OH (#221228, Sigma-Aldrich) for 90 s at 37°C. After the decellularization process, 50 U/mL Deoxyribonuclease I (#18047-019, Invitrogen) and RNase A (#12091-039, Invitrogen) was added to the lysate for 90 min, and the resulting solution was frozen overnight in 50 mL conical tubes at -80°C and lyophilized.

For western blot analysis, dECM was washed with PBS and then scraped to collect the sample. The sample was extracted from RIPA buffer containing a protein inhibitor cocktail (#89900, Thermo Fisher Scientific). The total protein quantity before and after decellularization was measured by bicinchoninic acid assay (#23225, Thermo Fisher Scientific). Samples (20 µg) were separated by electrophoresis of a 7.5% polyacrylamide gradient gel, and the separated proteins were transferred to a nitrocellulose membrane (HATF29325, Millipore). The membrane was blocked with 2.5% bovine serum albumin (BSA) in Tris-buffered saline (TBS), and primary and horseradish peroxidase-labeled secondary antibodies: anti-collagen I (1:1,000, ab138492, Abcam), -collagen III (1:1,000, ab184993, Abcam), -collagen IV (1:1,000, ab6586, Abcam), -fibronectin (1:1,000, ab6328, Abcam), -laminin (1:800, ab11575, Abcam), -elastin (1:800, ab213720, Abcam), and -β-actin (1:1,000, ab8226, Abcam) were added, incubated overnight at 4°C and proteins were detected. The membrane was washed, and the protein bands were detected using a chemiluminescence reagent (#32106, Thermo Fisher Scientific). Images were captured using the ImageQuant LAS 500 (GE Healthcare).

Relative growth factor levels were determined using a sandwich immunoassay array kit (ARY007, R&D systems). Briefly, the extracted ECM total protein concentration was assessed using a bicinchoninic acid assay, and 200 μg of total protein was used as input for the array. Chemiluminescence was detected using an ImageQuant LAS 500, and the signal intensities were measured using ImageQuant TL software (GE Healthcare).

The bioactive scaffolds were fabricated using a collagen solution and ECM powder for lyophilization within the mold. Specifically, Type I porcine collagen (MS Collagen; MSBIO) was reconstituted in 0.001N HCl at 5 mg/mL at 4°C under continuous gentle magnetic stirring (COL group). After the collagen was completely dissolved, the solution was mixed with fdECM (0.1%, w/v) (COL-fdECM group). Afterward, the mixture solution either for the COL or COL-fdECM group was poured into the 12 mm diameter mold to form scaffolds, collagen patch or fdECM-supplemented collagen patch, and they were frozen for 1 day at −78°C prior to lyophilization.

All experimental procedures for in vivo full-thickness skin wound study were approved by the Institutional Animal Care and Use Committee of Konkuk University (KU18128). Male 6-week-old C57BL/6 mice were purchased from Orient Bio. The mice were housed under controlled conditions at a temperature of 23 ± 2°C, humidity of 50 ± 5%, and a light-dark cycle of 12/12 h. To establish the full-thickness skin wound model, the surgical area was clipped and aseptically prepared using povidone iodine. The surgical excision was made on the dorsal skin of mice using a 10 mm biopsy punch under inhalation anesthesia with isoflurane (3.5% for induction and 1.5% for maintenance). The mice were randomly allocated into three groups (n = 5 per group), including a CON group with non-treatment, a COL group with collagen patch treatment, and a COL-fdECM group with fdECM-collagen patch treatment. After applying the allocated treatment, the skin wound was covered with Tegaderm and Coban to maintain humidity and protect from infection. Treatment was repeated for three consecutive days after the wound was induced, and the wound size was measured after 7 and 14 days. Prophylactic antibiotics (enrofloxacin, 5 mg/kg) and analgesics (meloxicam, 1 mg/kg) were administered subcutaneously after surgery and for 2 days afterward.

All experimental procedures for in vivo diabetic skin ulcer healing study were approved by the Institutional Animal Care and Use Committee of Konkuk University (KU18042). To establish the diabetic skin ulcer model, leptin receptor knockout mice (DB), C57BL/6-Lepr^em1hwl/Korl, were obtained from the National Institute of Food and Drug Safety Evaluation (NIFDS, Cheongju, Korea). Male 10-week-old DB and misty (wild-type) were used, and the same surgical procedure described in section In Vivo Full-Thickness Skin Wound Healing Study was applied. The DB mice were randomly allocated into two groups (n = 4 per group), the DB mice with injury-only group (DB/I) and DB mice with fdECM-collagen patch treatment group (DB/COL-fdECM). Misty mice with injury-only group (Wild/I) were used to compare the normal skin wound and diabetic ulcer.

After 7 and 21 days after inducing the skin wound and treatment, dorsal skin samples were obtained and fixed with 10% neutral-buffered formalin (BBC Biochemical). The tissue samples were embedded in paraffin through a routine process and were sectioned at 4 µm thickness. The tissue sections were hydrated by gradient alcohols for routine histological analysis and stained with hematoxylin and eosin. In addition, tissue sections were stained with Masson’s trichrome, and specific protein expressions were visualized by immunohistochemistry for further analysis. Briefly, the hydrated tissue sections were boiled using a TintoRetriever pressure cooker (Bio SB) in sodium citrate buffer (pH 6.0) to detect α-SMA, CD34, PGP9.5, loricrin, E-cadherin, and CD31 or in Tris-EDTA buffer (pH 9.0) to detect collagen I and III. For the primary antibodies, anti-collagen I (1:200, ab6308, Abcam), -collagen III (1:200, ab7778, Abcam), -α-SMA (1:1,000, ab7817, Abcam), -CD34 (1:200, MA5-27901, Invitrogen), -PGP9.5 (1:100, ab8189, Abcam), -loricrin (1:200, ab85679, Abcam), -E-cadherin (1:100, #14472, Cell Signaling Technology) and -CD31 (1:100, sc-376764, Santa Cruz Biotechnology) were used. The anti-CD31 antibody-labeled sections were incubated with a Vectastain Elite ABC-Peroxidase kit (Vector Laboratories), and the reactions were visualized by Vector SG (Vector Laboratories). Finally, the sections were counterstained with a nuclear fast red solution (Vector Laboratories) and analyzed under a microscope (Leica Microsystems). To detect proteins other than CD31, the antibody-labeled sections were incubated with secondary antibodies conjugated with Alexa fluor 488 and/or Alexa fluor 594 (Abcam). Then the sections were mounted in 4′-6-diamidino-2-phenylindole (DAPI) containing media (H-1800, Vector Laboratories) and analyzed under a fluorescence microscope (Leica Microsystems) or confocal fluorescence microscope (Zeiss).

Human dermal fibroblasts (PCS-201-010, ATCC) were cultured using DMEM, supplemented with 10% FBS, 1% penicillin/streptomycin (#15140122, Gibco), and 2% Mycoeraser (ME05-100, RD tech) at 37°C in a 5% CO₂ humidified incubator. The cells were seeded on a 6-well plate at a density of 7.5 × 10⁵ cells/well and incubated for 24 h to reach the confluency. The scratch wound was made using a commercial scratcher (SPL Life Sciences). The recombinant human epidermal growth factor (rhEGF) (E9644, Sigma-Aldrich) or fdECM powder was dissolved in DPBS. The concentrated solution of rhEGF or fdECM was added to each well to a final concentration of 10 ng/mL. For the control group, the same volume of DPBS was added. Photomicrographs were taken at each time point (0, 18, 24, and 48 h) after applying the treatment. The relative scratch gap area was measured by using ImageJ software (National Institutes of Health). After 48 h when the scratch gap was restored, the cells were either fixed with 4% paraformaldehyde or the supernatant media and cells were separately harvested and stored at −78°C, for further analysis.

The fixed cells were permeabilized using 0.1% Triton X-100 for immunocytochemistry analysis. The samples were blocked with 2% normal serum and incubated with anti-α-SMA (1:1000), -collagen I (1:200), and -collagen III (1:200). The signals were detected with secondary antibodies conjugated with Alexa fluor 488 or 594, and nuclei were counterstained with DAPI (D9542, Sigma-Aldrich).

Growth factors secreted by fibroblasts in the media were determined using a human growth factor antibody array kit (ab134002, Abcam), according to the manufacturer’s protocol. Two milliliters of supernatant media were used for each quantification. Chemiluminescence was detected using a charged-coupled device (CCD) camera (Fusion Solo S, Vilber), and the signal intensities were measured using ImageJ.

The cells were harvested for mRNA expression analysis at -78°C. Total RNA was isolated using the RNeasy Mini kit (#74106, Qiagen) according to the manufacturer’s instructions. One microgram of extracted RNA was transcribed into cDNA using the QuantiTect Reverse Transcription Kit (#205311, Qiagen). The primers for the target genes are listed in Table 1. The PowerUp SYBR Green Master Mix (A25742, Applied Biosystems) was used to perform real-time PCR (initial denaturation, 5 min at 95°C; recurring denaturation, 15 s at 95°C, and amplification, 30 s at 60°C, for 45 cycles). All mRNA levels were normalized to mean values of reference genes, β-actin, GAPDH, 18s rRNA, β2-microglobulin, and TBP, and the results were calculated as fold changes of the cycle threshold (Ct) value relative to the controls using the 2^−ΔΔCt method.

To analyze the MAPK, TGF-β/SMAD, JAK/STAT, and PI3K/AKT signaling pathways, total protein was extracted using M-PER (Thermo Fisher Scientific) extraction buffer with Halt phosphatase protease inhibitor (Thermo Fisher Scientific) according to the manufacturer’s instruction. The protein concentration was determined using a protein assay kit (Bio-Rad), and the total protein (20 µg) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a 4-20% precast gel (Bio-Rad). The separated proteins were transferred onto polyvinylidene fluoride membranes, followed by western blot analysis. Briefly, the background signals were blocked with EveryBlot Blocking buffer (#12010020, Bio-Rad) and incubated with each primary antibody, purchased from Cell Signaling Technology and diluted to 1:1,000 in 3% BSA/TBST (SMAD3, #9513; p-SMAD3, #9520; ERK1/2, #4695; p-ERK1/2, #4370; JNK, #9252; p-JNK, #4668; p38, #9212; p-p38, #4511; STAT3, #30835; p-STAT3, #9145; AKT, #4691; p-AKT, #4060; β-actin, #4970). After several washing steps, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (Cell Signaling Technology). A CCD camera was used to detect immunoreactive bands using a chemiluminescent substrate (Clarity Western ECL Substrate; Bio-Rad). Quantification was performed using ImageJ, and each expression was standardized by the β-actin level.

Statistical analysis was conducted using Prism V5 (Graphpad Software Inc). Significant differences between groups were determined using one-way analysis of variance (ANOVA) followed by Bonferroni multiple comparisons post hoc test for parametric measures, Kruskal–Wallis test followed by Dunn post hoc test for non-parametric measures, and linear regression for construction of standard curves. Statistical significance was accepted at p < 0.05.

A commercially obtained human cell line, WI-38, was used to prepare fdECM. After the cells reached full confluency, we applied an optimal decellularization protocol to remove DNA and nuclei from the cells while maximally preserving ECM content (Figure 1A and Supplementary Figure S1). The extracted ECM was approximately 100 mg per confluent T-175 flask (571 μg/cm²) and the ECM yield was 220 mg/g protein (p < 0.001, compared to extracted protein without decellularization). To verify the preservation of the active ingredients, the protein contents of cells before and after decellularization were compared by western blot. As a result, collagens I, IV, and elastin proteins showed little change, while collagen III, fibronectin, and laminin were reduced to 20, 50, and 20%, respectively, compared to the original cells (Figure 1B). In the growth factor array analysis, coagulation factor III, Serpin E1, and dipeptidyl peptidase-4 were the most abundant remaining factors even though the original contents were reduced (Figure 1C). However, several growth factors, including epidermal growth factor (EGF), hepatocyte growth factor (HGF), insulin-like growth factor-binding protein (IGFBP), platelet derived growth factor (PDGF), and vascular endothelial growth factor (VEGF) remained without being significantly lost during the decellularization process (Figure 1D).

Based on the fdECM remaining ingredients, we hypothesized that fdECM could accelerate wound healing via angiogenesis (Werner and Grose, 2003). First, we made an fdECM supplemented collagen patch to confirm its effect by adding the extracted fdECM to collagen I for the in vivo experiment (Figure 2A). The fdECM-collagen patch was observed to have a more homogeneous surface than the control collagen patch. However, the mechanical properties of the fdECM-collagen patch were similar to the control collagen patch, except the slightly increased compressive modulus (Supplementary Figure S2). We applied this patch to the full-thickness skin wound model to confirm its biological activities. After 14 days, in the control group (CON), where nothing was applied, the hemorrhagic crust was on the wounded skin (Figure 2B). In contrast, the defect area was almost completely epithelialized in the control collagen patch (COL) or fdECM-collagen patch (COL-fdECM) groups, and the wound area appeared smaller in the COL-fdECM group. To confirm the angiogenic efficacy of the prepared fdECM-collagen patch, we analyzed the blood vessels in the granulation tissue by CD31 immunostaining in the sampled tissue on day 7 (Figure 2C). As a result, the COL-fdECM group exhibited about twice as much blood vessel area as the other groups because the size and number of the blood vessels were significant (Figure 2D). Histological analysis was performed on the sample tissue on day 21 when the proliferation ended and remodeling was active, to confirm the long-term effect on wound healing. In the CON group, most of the defects consisted of immature skin tissue where only a few skin adnexa were observed (Figure 2E). However, in the COL-fdECM group, more hair follicles were observed in the center of the wound bed than the CON or COL group (Figure 2F). Moreover, the normal skin was closer to the center of the defect in the COL-fdECM group than the control (Figure 2G). In addition, when the collagen or fdECM-collagen patch was applied, the thickness of both the epidermis and the dermis was reduced compared to the CON group (Figure 2H). These results collectively confirm that the scaffold’s regeneration efficacy created with fdECM promoted angiogenesis and contributed to wound contraction. Furthermore, the reduced thickness of the skin and increased hair follicles in the wound bed suggest that the treatment may help wound remodeling. Therefore, we attempted additional analyzes to evaluate it.

First, we analyzed the epidermal barrier function by staining cell adhesion molecule E-cadherin and the major component of the cornified envelope, loricrin (Figure 3A) (Tunggal et al., 2005; Abhishek and Palamadai Krishnan, 2016). The E-cadherin positive area seemed larger in the control group because of epidermal thickness. However, the signal intensity was higher in both the COL and COL-fdECM groups, meaning keratinocytes adhered to each other tightly. Loricrin expression in the CON group appeared diffuse and less intense, similar to the E-cadherin expression pattern. However, in the COL-fdECM group, loricrin was expressed in a constant line. Then, we analyzed the content of major components in the dermis, collagen I and III (Figure 3B). The immunofluorescence staining confirmed that collagen I and III content in the regenerated dermis was higher in the COL-fdECM group than in the CON and COL groups. We also checked whether the cut nerve fiber re-innervated from the ganglia to the skin. The PGP9.5 positive nerve fiber area was also higher in the superficial dermis of the COL-fdECM group compared to the other groups (Figure 3C). These results indicate that the fdECM also helps remodel the skin wounds to the original tissue. To analyze the effect and mechanism of fdECM on wound remodeling, we analyzed the phenotype change of the dermal fibroblast, the cell constituting most of the skin (Figure 4). After 7 days, the α-SMA⁺ was highest in the COL-fdECM group, indicating that CD34⁻ dermal fibroblasts were mostly α-SMA⁺ myofibroblasts. Conversely, their activity was lowest at day 21 in the COL-fdECM group, suggesting dermal fibroblasts reversed their phenotype.

To validate fdECM’s efficacy on fibroblasts, we treated human dermal fibroblasts with unfabricated fdECM and analyzed the behavior. As a positive control, we used rhEGF, commercially available and known for various physiological activities, including angiogenic effect. The fdECM group showed increased cell migration to the scratch gap than the control group 18 h after the scratch wound was induced, although the number of fibroblasts migrating was highest in the rhEGF group (Figure 5A). After 48 h, when the scratch wound was completely restored, the fibroblasts of the CON and rhEGF groups remained in their active form, α-SMA positive myofibroblasts (Figure 5B) (Pilcher et al., 1999). However, most fibroblasts of the fdECM group showed α-SMA negative quiescent form, consistent with the in vivo results. Although myofibroblasts play a major role in contracting skin wounds, the remaining myofibroblasts, after wound healing, are suspected of causing hypertrophic scar formation (Chitturi et al., 2015). Therefore, the reduced number of myofibroblasts indicates a more favorable state for wound remodeling.

Moreover, it was confirmed that ECM components secreted by fibroblasts were different for each group. The fdECM or rhEGF treated fibroblasts were observed to secrete collagen I more than the control fibroblasts (Figure 5C). In contrast, increased collagen III-positive cells were observed in the CON and rhEGF groups and less in the fdECM group (Figure 5D). The major dermal collagen type changes from collagen III to collagen I as the skin wound is remodeled (Volk et al., 2011; Tracy et al., 2016). Therefore, these results indicate fibroblasts treated with fdECM reverse their differentiation to a stable form and secrete collagen I more than collagen III. We also evaluated fibroblasts’ growth factor secretion pattern (Figure 6). Among the 41 factors tested, fibroblast growth factor 7 (FGF7), insulin-like growth factor 1 (IGF1), insulin-like growth factor-binding protein 3 (IGFBP3), transforming growth factor beta-2 (TGF-β2), and heparin binding epidermal growth factor (HB-EGF) increased in the fdECM group compared to the CON group, but other factors did not differ. In particular, FGF7 and IGF1 were higher than in the rhEGF group, suggesting that these two factors may play a role in wound migration (Haase et al., 2003; Kirfel and Herzog, 2004).

To elucidate the underlying mechanism of fdECM on wound remodeling, mRNA expression level and cell signaling were inspected. As shown in a previous study describing that FGFs mRNA increased rapidly after wounding and then decreased as the wound healed [6], FGF-2, -7, and -10 were downregulated in the fdECM treated group after 48 h (Figure 7). Matrix metalloproteinases (MMPs) mRNA pattern after 48 h of fdECM treatment was similar to that of the late phase in skin wound healing. Proinflammatory MMPs, such as MMP1 and MMP2, were downregulated, while MMP3, remodeling MMP, was upregulated. Tissue inhibitor of metalloproteinase-2 (TIMP2) was significantly downregulated in fdECM treated dermal fibroblasts, while TIMP1 and TIMP3 were stable.

We also measured the signaling pathway activated by fdECM using immunoblot analysis. Similar to rhEGF, fdECM treated fibroblasts showed increased AKT and JNK phosphorylation and decreased SMAD3 phosphorylation (Figure 8). However, the phosphorylation of ERK was decreased compared to the rhEGF group, and the p38 was decreased to the control group level. Together with the other results, these results suggest that adequate remodeling is related to regulating the p-38 and ERK pathways.

Finally, we investigated whether the fdECM efficacy can also be applied to the diabetic skin wound model (Figure 9A). After 15 days of inducing the skin wound, the wound closure rate of COL-fdECM treated diabetic mice (DB/COL-fdECM) was higher than the injury-only group (DB/I) (Figure 9B). Histopathologically, the wound bed size of the diabetic ulcer model was larger compared to normal mice, and hyperkeratosis was observed (Figure 9C). However, the wound closure of diabetic ulcers was improved after applying the fdECM-collagen patch. The CD31 positive vessel area was greater in the DB/COL-fdECM group compared to the DB/I group (Figure 9D). However, the increased vessel area was due to the increased size of the blood vessels rather than the number of blood vessels. Because of the progressed wound healing stage compared to previous experiments, this result suggests that the vessels are more mature after COL-fdECM treatment. Moreover, collagen quantity in the regenerated dermis was higher after being treated with an fdECM-collagen patch, suggesting a better quality of dermal regeneration (Figure 9E). The thickness of the epidermis and dermis was also significantly reduced compared to the DB/I group (Figure 9F). These results support that fdECM and its therapeutic efficacy can be applied to chronic skin wounds such as diabetic ulcers.