MicroRNA-129 and -335 Promote Diabetic Wound Healing by Inhibiting Sp1-Mediated MMP-9 Expression

Delayed wound healing is a common and serious complication in patients with diabetes, affecting 15–20% of all persons with diabetes (1–3). The reasons diabetic wounds do not heal properly are unknown. However, an imbalance between the synthesis and degradation of the extracellular matrix (ECM) leads to poor wound healing in patients with diabetes (4).

Matrix metalloproteinases (MMPs) are a zinc-dependent endopeptidase family that degrades ECM components involved in tissue remodeling (5). The detrimental effects of MMPs in diseased tissue are attributed to the rapid turnover of potential growth factors, receptors, and the newly formed ECM, which are essential for wound healing (6). Elevated MMP-9 levels are present in various chronic nonhealing wounds, including diabetic foot ulcers (DFUs) (7–9). Our previous studies show that abnormally high MMP-9 expression contributes to delayed wound healing because of the imbalance between ECM synthesis and degradation, and reducing MMP-9 expression promotes diabetic wound healing (10,11).

MMP expression and activity are tightly regulated at multiple levels (12), including gene transcription, posttranscriptional processing, and proenzyme activation (13,14). We recently demonstrated that methylation levels at specific DNA loci in the MMP-9 promoter region are affected by conditions resembling diabetes in the human keratinocyte cell line HaCaT. Additionally, we investigated the MMP-9 promoter using bioinformatics analyses and identified a specificity protein-1 (Sp1) binding site (15). However, Sp1 involvement in the regulation of MMP-9 promoter activity is not well studied. Similarly, the regulatory mechanisms of Sp1-mediated MMP-9 expression are unknown.

MicroRNAs (miRNAs) are highly conserved endogenous small noncoding RNA molecules involved in numerous biological processes including diabetic wound healing. miRNAs regulate posttranscriptional gene expression by binding to their target mRNAs, leading to mRNA degradation or translation suppression (16–19). Although recent studies revealed the important role of miRNAs in skin biology and diseases, their role in diabetes wound healing is in its infancy.

We used serum miRNA expression profiling to identify novel miRNAs in diabetic wounds and found that microRNA-129 (miR-129) and miR-335 were significantly downregulated in patients with diabetes compared with healthy patients. miR-129 and -335 play key roles in the regulation of cancer progression, chemoresistance, proliferation, and cell cycle control (20–23). However, their function in diabetic wound healing is unclear. Through bioinformatics analysis, we predicted that the transcription factor Sp1 was a common target of miR-129 and -335. Therefore, we explored the mechanism through which miR-129 and -335 regulated Sp1-mediated MMP-9 expression and studied their expression and function in the wound healing of patients with diabetes.

The Institutional Review Board of the Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University (Guangdong, China) approved the study protocol, which was in accordance with the principles of the Helsinki Declaration II. Written informed consent was obtained from each participant before data collection. The Sun Yat-Sen University Institutional Animal Care and Use Committee approved all animal studies.

Demographic information including age, sex, blood pressure, BMI, blood pressure (systolic and diastolic blood pressure), fasting blood glucose, hemoglobin A_1c (HbA_1c), serum creatinine, triglycerides, total cholesterol, HDL cholesterol, LDL cholesterol, and blood cell count (red blood cell, white blood cell, platelets, and hemoglobin) were recorded for all patients with DFUs and control subjects (Supplementary Table 1). Wound fluids were collected from the ulcer site of 21 patients with DFUs as previously described (24). Samples were stored at −80°C for further quantification of MMP-9 expression. The MMP-9 concentration was measured using an MMP-9 activity assay kit (Abcam, Boston, MA). The ulcer grade was assessed using the Wagner classification (25).

Diabetic perilesional skin samples were obtained from patients undergoing amputation surgery. Control (healthy) skin tissue samples were collected from the lower leg of patients without diabetes undergoing reconstruction due to foot injury. The wound edge area was excised in the perilesional area of the skin ulcer or the edge of the surgical incision during routine surgery. Samples were rapidly trimmed to small strips (size 1 × 1 × 1 mm), fixed in 4% paraformaldehyde at 48°C overnight, and embedded in paraffin. Each paraffin block was serially sectioned into 20 sections 4 μm thick and used for immunohistochemistry (IHC).

Skin tissues were stained with hematoxylin-eosin (HE). Epidermal and dermal thickness was measured on stained slides using pictures taken at ×10 magnification. Masson trichrome staining was performed using a staining kit (Shanghai Bogoo Biotechnology, Shanghai, China) according to the manufacturer’s instructions. For collagen quantification, Masson trichrome–stained skin areas were determined using ImageJ software (National Institutes of Health).

Paraffin-embedded sections of skin tissues were stained using the avidin–biotin complex method as previously described (26). Sections were incubated with polyclonal rabbit anti-Sp1 or anti–MMP-9 antibody at 4°C overnight followed by treatment with horseradish peroxidase–labeled anti-rabbit IgG. The streptavidin–biotin complex was used to visualize the staining. For each section, five high-power fields were observed, and 100 cells from each field were counted. Scoring was conducted according to the immunoreactive score (IRS) standard as previously reported (26). The antibodies were purchased from Millipore (Temecula, CA).

Primary human keratinocytes were obtained from the hospital during routine infant circumcision (26) and identified as described previously (Supplementary Fig. 1). HaCaT cells were purchased from American Type Culture Collection (Manassas, VA) and cultured in keratinocyte serum-free medium (Gibco, Gaithersburg, MD). After reaching 80% confluence, cells were incubated overnight in keratinocyte serum-free medium containing 0.5 mg/mL BSA (Sigma-Aldrich, St. Louis, MO) or advanced glycation end product (AGE)–BSA (Millipore). The supernatants of HaCaT cells were collected to measure MMP-9 activity using an MMP-9 activity assay kit (Abcam). HaCaT cells were harvested for analysis of Sp1, MMP-9, and miRNA expression. Where indicated, Sp1 small interfering RNA (siRNA)1–3 (GenePharma, Shanghai, China) was transfected into the cells, or the Sp1-specific inhibitor mithramycin A (100 nmol/L; Abcam) was added.

RNA isolation and quantitative real-time PCR (qPCR) were performed as previously described (27). Briefly, total RNA was extracted with TRIzol (Gibco) according to the manufacturer’s instructions. qPCR was performed using a LightCycler 480 SYBR Green Master (Roche Diagnostics, Mannheim, Germany) in a LightCycler System. All data were analyzed using the expression of GAPDH as an internal RNA standard.

To identify Sp1 and MMP-9 expression, lysates from cultured keratinocytes, HaCaT cells, and skin tissues were collected and analyzed. Nuclear proteins were separated using commercial NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, Waltham, MA). Western blot analysis was performed as described previously (28). For normalization, the same membrane was immunoblotted with anti–β-actin or anti-H3 antibody (Cell Signaling Technology, Danvers, MA).

HaCaT cells were incubated with annexin V–FITC in loading buffer with 10 μL propidium iodide, resuspended in binding buffer, and detected using an FACS (BD Biosciences, Franklin Lakes, NJ) analyzer. Analysis was carried out in triplicate for three separate experiments.

Chromatin samples from HaCaT cells or primary keratinocytes were prepared and incubated with an anti-Sp1 antibody or normal IgG control (Millipore). Pulled-down DNA fragments and input DNA were purified and used as templates for qPCR analysis using primers designed to amplify the −668- to −538-bp region of the MMP-9 promoter, which contains the putative Sp1 binding site.

HaCaT cells were cotransfected with the MMP-9 promoter luciferase reporter and the pRL-TK reporter plasmid (Promega, Madison, WI) in the presence of an empty vector or a plasmid expressing Sp1 (Sangon Biotech Co., Shanghai, China). The luciferase activities were assessed using a Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s protocol. The relative luciferase activity was normalized to the Renilla luciferase activity.

An anti-AGO2 antibody (Ab32381; Abcam) and the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (17-700; Millipore) were used to perform RNA immunoprecipitation of AGO2-containing RNA-induced silencing complex according to the manufacturers’ instructions.

Microarray hybridization, data generation, and normalization were performed by KangChen Bio-tech (Shanghai, China) using standard protocols (28–30). The threshold value used to screen differentially expressed miRNAs was a fold change of ≥2.0 or ≤0.5, P < 0.01, and a normalized signal value indicating a relative abundance to the transcript of ≥2.0. The Sp1 3′ untranslated region (UTR) sequences containing putative wild-type (WT) and mutated (MUT) target sites for miR-129 or miR-335 were chemically synthesized and inserted into the pGL3-luciferase reporter vector (Promega). Human embryonic kidney 293T (HEK-293T) cells were cotransfected with the reporter vector containing either the WT (Sp1-WT) or MUT (Sp1-MUT) putative Sp1 binding site and miR-129 or miR-335 mimics using the Lipofectamine 3000 transfection reagent (Invitrogen, Carlsbad, CA). After 48 h, cell lysates were prepared and used for the detection of the luciferase activity measured with a Dual-Luciferase Reporter Assay Kit (Promega).

miR-129 or -335 mimics, inhibitors, and scrambled miRNA (negative control [NC]) were designed and purchased from RiboBio (Guangzhou, China). HaCaT cells were cultured until 50–60% confluent and transiently transfected with miR-129 or -335 mimics, inhibitors, or NC. (50 μmol/L) using Lipofectamine 3000 (Invitrogen), according to the manufacturer’s instructions. After 48 h, the cells were harvested and used for further assays.

The 5-ethynyl-2′-deoxyuridine DNA proliferation assay was performed according to the manufacturer’s instructions (RiboBio). Images were obtained using an inverted fluorescence microscope. The percentage of 5-ethynyl-2′-deoxyuridine–positive cells was calculated from five random fields. Cell proliferation was determined using a CCK-8 kit (CWBiotech, Beijing, China) according to the manufacturer’s instructions. The vertical migration abilities of HaCaT cells were assessed using a transwell migration assay, as described previously (31). The stained cells from five randomly chosen fields were counted under a light microscope.

The hsa-miR-129 and hsa-miR-335 probes were labeled with 5′ digoxigenin. An enhanced sensitive in situ hybridization detection kit I (POD; Boster Biological Technology, Wuhan, China) was used to perform in situ hybridization, according to the manufacturer’s protocol. All related reagents were purchased from CWBiotech. For each section, five high-power fields were observed, and 100 cells from each field were counted. The percentage of miRNA-positive cells was calculated.

Sprague-Dawley rats (200–250 g each) were obtained from the Laboratory Animal Center of Sun Yat-Sen University. A diabetic wound was made as previously described (11). miR-129 and/or -335 agomirs (500 μL of a 1-nmol/L agomir solution in PBS) were injected intradermally into the wound edges of the rats immediately and 4, 7, and 10 days after wounding. A negative agomir injection group was used as the control. Wound images were immediately acquired after wounding (day 0) and on days 7, 10, and 14. Wound closure was quantified as the percentage of the initial wound area size. Wound healing rate was calculated as the percentage of the original wound size using the following formula: (initial area − final area)/initial area × 100%. Wound samples were collected 14 days after injury for histology, gene and protein expression, and collagen deposition analyses.

Continuous variables are presented as the mean ± SD, and differences among groups were tested using the Student t test or one-way ANOVA. Statistical analysis was performed using SAS software, version 9.2 (SAS Institute Inc., Cary, NC). All SPSSs were two-sided, and P < 0.05 was considered statistically significant.

MMP-9 levels in wound fluid were significantly higher in patients with DFUs than in control subjects. Additionally, an increase in the Wagner grade of the lesion was associated with a dramatic increase in MMP-9 level (Fig. 1A). Similarly, improvement in DFUs was accompanied by a gradual decline of MMP-9 levels (Fig. 1B). MMP-9 expression in skin tissues of patients with DFUs was in the epidermis (Fig. 1C). The MMP-9 staining intensity was significantly higher in the skin from patients with DFUs compared with control subjects. A significant increase in MMP-9 expression was also observed in skin tissues of diabetic rats compared with controls (Fig. 1C). Histological results of perilesional skin biopsies from patients and animal models showed that the epidermis and dermis from diabetic wounds were significantly thinner than from control subjects (Fig. 1C and D). The collagen fibers of the diabetic skin were thin, degenerated, and fractured; additionally, dermal collagen was less dense and disordered (Fig. 1E).

Sp1 was expressed predominantly in the cytoplasm of keratinocytes from normal skin tissues, whereas in diabetic skin tissues, Sp1 was expressed in both the nucleus and cytoplasm of keratinocytes (Fig. 2A). We also observed significantly increased Sp1 immunostaining in the skin samples from patients with diabetes and diabetic animals compared with controls (Fig. 2B).

After 24-h exposure to 200 mg/L AGE-BSA, the Sp1 mRNA expression in HaCaT cells and primary keratinocytes was unchanged. However, MMP-9 mRNA expression levels increased significantly after AGE-BSA treatment (Fig. 2C). The Sp1 and MMP-9 protein expression increased significantly after 48 h of AGE-BSA treatment compared with the control group (Fig. 2D). Sp1 was expressed predominantly in the cytoplasm of HaCaT cells, whereas AGE-BSA increased Sp1 activity and promoted its nuclear translocation (Fig. 2E and F).

Bioinformatics analysis indicated a putative Sp1 binding site is located in a region between −571 and −540 bp on the MMP-9 promoter (Fig. 3A). Therefore, we hypothesized Sp1 could act on the MMP-9 promoter and enhance MMP-9 expression. Sp1 inhibition via RNA interference or treatment with the specific Sp1 inhibitor mithramycin A significantly decreased MMP-9 expression (Fig. 3B) without influencing cell proliferation (Supplementary Fig. 2) or apoptosis (Fig. 3C).

To test whether Sp1 modulates MMP-9 promoter activity, we performed luciferase assays. Sp1 overexpression significantly increased MMP-9 promoter activity (approximately eightfold) (Fig. 3D). The results from chromatin immunoprecipitation (ChIP) and qPCR assays confirmed Sp1 binds directly to the MMP-9 promoter and activates its transcription in both HaCaT cells and primary keratinocytes (Fig. 3E). Moreover, the treatment with AGE-BSA induced a stronger binding of Sp1 to the MMP-9 promoter than BSA treatment (Fig. 3F).

To identify changes in miRNA expression in diabetic skin lesions, we conducted a comprehensive microarray analysis comparing miRNA expression profiles of serum samples from five patients with diabetic wounds and five matched control subjects (Supplementary Table 2). Hierarchical clustering identified 159 miRNAs significantly altered in the serum of patients with diabetic wounds (Fig. 4A). Six of the 58 downregulated miRNAs (miR-32, -106a, -106b, -129, -155, and -335) were validated using qPCR in AGE-BSA–treated HaCaT cells (Fig. 4B). We focused on miR-129 and -335 because they were significantly downregulated in patients with diabetes, and their biological function in diabetic wound healing was unclear. A significant downregulation of miR-129 and -335 expression was confirmed in the skin of patients and rats with diabetic wounds through qPCR (Fig. 4C) and in situ hybridization assays (Fig. 4D and E).

We then searched for possible miR-129 and -335 targets using three online bioinformatics algorithms: miRanda, TargetScan, and PicTar (Fig. 5A). The analysis indicated miR-129 and -335 had highly conserved binding sites in the 3′-UTR of Sp1 in several species. Argonaute-RNA immunoprecipitation (AGO-RIP) results showed miR-129, miR-335, and Sp1 were enriched in the AGO-RIP fraction of HaCaT cells compared with the IgG control (Supplementary Fig. 3). To verify whether miR-129 and miR-335 directly target the Sp1 3′-UTR, we performed 3′-UTR luciferase reporter assays. The luciferase activity of the Sp1 WT 3′-UTR reporter decreased twofold after transfection with miR-129 or -335 mimics compared with control. The mutation of the predicted target sites completely abolished the repressive effect of miR-129 or -335 mimics on the reporter gene expression, demonstrating miR-129 and -335 directly target Sp1 (Fig. 5B).

Because Sp1 is an important transcription factor for MMP-9, the downregulation of Sp1 by miR-129 and -335 might lead to the downregulation of MMP-9 transcription. To investigate this hypothesis, we evaluated Sp1 and MMP-9 protein levels in HaCaT cells after treatment with miR-129 or -335 mimics/inhibitors. miR-129 or -335 overexpression in keratinocytes significantly decreased Sp1 expression and MMP-9 activity, whereas opposite results were obtained when inhibiting miR-129 or -335 (Fig. 5C, D, and F).

Interestingly, miR-129 and -335 mimics had a synergistic inhibitory effect on Sp1 and MMP-9 expression. Meanwhile, miR-129 and -335 inhibitors had synergistic effects promoting Sp1 and MMP-9 expression (Fig. 5C, D, and F). After treatment with AGE-BSA, miR-129 or -335 ectopic expression also contributed to the downregulation of Sp1 expression and MMP-9 activity (Fig. 5E and F). Moreover, transfection of miR-129 or -335 mimics moderately induced the migratory capability of HaCaT cells, and their cotransfection significantly induced cell migration without affecting cell viability, proliferation (Supplementary Figs. 4–6), or apoptosis (Fig. 5G).

ChIP experiments were performed to investigate Sp1 binding to the MMP-9 promoter in the presence of miRNA inhibitors. We found that miR-129 or -335 knockdown resulted in increased binding of Sp1 to the MMP-9 promoter, and miR-129 and -335 inhibitors had a synergetic effect on Sp1 binding activity (Fig. 5H).

We injected miR-129 or -335 agomir or control oligos intradermally into the wound edges of diabetic rats immediately and 4, 7, and 10 days after the skin was injured (Fig. 6A). At 14 days, the wound healing rate was significantly higher in the miR-129 (79.5 ± 2.1%) or -335 (82.4 ± 4.5%) agomir group than in the controls (69.1 ± 2.2%; P < 0.05). The combination miR-129 and -335 agomir group (mix agomirs group) showed an 84.4% wound contraction (Fig. 6B and C).

Consistently, we observed decreased expression of Sp1 and MMP-9 in the skin of the rats injected with the agomir in comparison with controls (Fig. 6D and E). The combined injection of miR-129 and -335 showed synergistic inhibitory effects on Sp1 and MMP-9 expression in vivo.

Histological results showed that the epidermis and dermis of the control rats were significantly thinner than those of the rats injected with the miRNA agomirs (Fig. 6F). A clear and distinct collagen layering was found in the miRNA agomir-treated group, whereas in the control group, collagen bundles were loosely packed (Fig. 6F).

These results indicate that miR-129 and -335 promote wound healing through the downregulation of Sp1-mediated MMP-9 expression in vivo.

We report the novel finding that, in diabetes, the expression of the transcription factor Sp1 is increased, and its nuclear translocation is activated; Sp1 binds directly to the MMP-9 promoter, increasing MMP-9 expression in keratinocytes, which is associated with poor wound healing. miR-129 and -335 were markedly downregulated in the skin of patients with DFUs, as well as in animal models of diabetic wounds. These two miRNAs act cooperatively to decrease MMP-9 levels by directly targeting the Sp1 3′-UTR. In vivo upregulation of miR-129 and -335 promoted wound closure through the suppression of Sp1-mediated MMP-9 expression in a diabetic wound model. Thus, targeting miR-129 and -335 may be a valid approach to promote wound healing in patients with diabetes.

In the pathogenesis of diabetic wounds, MMP-9 is activated in skin cells, which impairs the balance of ECM synthesis and degradation, leading to unhealed wounds (7,8,32,33). In this study, we explored the association between MMP-9 levels in wound fluid and the severity of DFUs. The gradual decline of MMP-9 levels during the dynamic observation of the wound healing process further verified the essential role of MMP-9 in diabetic wound healing. In support of this, we and others (10,11) have shown that MMP-9 inhibition promotes wound healing. We then provided mechanistic details showing that MMP-9 is activated in the skin of patients with diabetes. MMP gene expression is regulated at the transcriptional level, and the MMP-9 promoter contains cis-acting regulatory elements that bind several transcription factors (34–36). In a previous study analyzing the transcriptional activity of a truncated MMP-9 promoter, we demonstrated that Sp1 is the most important transcription factor regulating MMP-9 expression (15). Sp1 is associated with the development and progression of various chronic complications in diabetes (37,38). However, the clinical significance and biological role of Sp1 in diabetic skin tissues are unknown. The present results show Sp1 is upregulated in diabetic skin tissues and AGE-BSA increased Sp1 activity and promoted its nuclear translocation in keratinocytes. At the same time, ectopic Sp1 expression increased MMP-9 levels in HaCaT cells. Using both inhibitors and siRNAs, Sp1 was a potent mediator of MMP-9 upregulation in response to AGE-BSA. This finding is similar to the recently reported Sp1 role in MMP-9 expression induction by reactive oxygen species in macrophages of pulmonary fibrosis (39). Our results clearly demonstrate that diabetes-induced Sp1 expression is closely related to MMP-9 activation. Further, ChIP and dual-luciferase assays verified that the binding of Sp1 to the MMP-9 promoter was functional: Sp1 could activate MMP-9 transcription via directly binding to the MMP-9 promoter region. Although previous studies suggested Sp1 might be involved in the regulation of MMP transcription (40,41), this is the first study reporting the direct binding of Sp1 to the MMP-9 promoter in diabetic skin cells.

Next, we focused on Sp1 posttranslational regulation. miRNAs play pivotal roles in different phases of wound healing (42). We analyzed the miRNA signature in the serum of patients with DFUs and identified 159 differentially expressed miRNAs. Consistent with previous studies, the wound-healing miR-32, -106a, -106b, and -155 (5,43–45) were altered in our comparative miRNA array. Among all of the downregulated miRNAs, miR-129 and -335 showed the most significant changes. miR-129 and -335 play key roles in tumorigenesis (20–23), whereas their roles in wound healing remain unknown. In this study, in diabetic conditions induced by AGE-BSA treatment, miR-129 and -335 expression significantly decreased in HaCaT cells and primary cultured keratinocytes.

The main function of miRNAs is to regulate posttranscriptional gene expression by binding to target mRNAs leading to their degradation, causing translation suppression or gene activation. First, the inverse correlation between miR-129 and miR-335 with Sp1 in skin samples of patients with diabetic wounds was demonstrated, suggesting that miR-129 and -335 may regulate Sp1 and are therefore clinically relevant. Then, Sp1 was identified as a common target of miR-129 and -335 through AGO-RIP experiments and 3′-UTR luciferase reporter assays. Interestingly, miR-129 and -335 showed combinatorial effects on Sp1 repression in keratinocytes. The synergetic effect cannot be attributed to the dose of miRNAs used in the experiments because the same total amount of miRNAs was used each time. Recent studies focusing on the combinatorial effect of miRNAs establish that a single mRNA molecule can be targeted by multiple miRNAs (46,47). For example, miR-30d, miR-181a, and miR-199a-5p act cooperatively to decrease the levels of 78-kDa glucose-regulated protein in cancer (48). The results presented in this study suggest that the combined action of multiple miRNAs might be important to achieve efficient Sp1 downregulation.

The modulation of miRNA expression through the administration of specific miRNA mimics or inhibitors might have therapeutic potential for nonhealing wounds (49,50). In the current study, subcutaneous injection of miR-129 or -335 agomir accelerated diabetic wound healing, improved the skin thickness in a diabetic wound animal model through decreased MMP-9 expression, increased collagen content, and enhanced migration of keratinocytes. As mentioned above, in vitro experiments showed that the overexpression or inhibition of miR-129 or -335 resulted in the downregulation or upregulation of Sp1 expression, respectively, and the concomitant change of MMP-9 protein levels in keratinocytes. Therefore, we hypothesized that the mechanism through which miR-129 and -335 promote wound healing in diabetic skin tissues might involve the suppression of Sp1-regulated MMP-9 expression.

The combined local application of miR-129 and -335 in vivo synergistically downregulated MMP-9 expression through targeting Sp1 and induced the recovery of skin thickness and collagen content. However, miR-129 and -335 did not act synergistically in wound contraction compared with the single miRNA injection. The combination of the two miRNA mimics may induce the overinhibition of MMP-9 expression. Despite excessive MMP activity delaying wound healing, their overdownregulation might inhibit ECM degradation and subsequent keratinocyte migration. This indicates that MMP-9 expression must be tightly controlled during wound healing.

In summary, our findings demonstrate an important role for Sp1-mediated MMP-9 expression in diabetic wound healing. miR-129 and -335 inhibit MMP-9 expression by targeting Sp1. Upregulation of miR-129 or -335 favors wound healing through MMP-9 downregulation. Collectively, these results establish the molecular mechanism by which miR-129 and -335 modulate the matrix-remodeling enzyme MMP-9 via Sp1 and suggest that this mechanism could be exploited to accelerate wound healing in vivo. These findings provide insight into the regulation of MMP-9 expression in diabetic wound healing by miRNAs and provide new therapeutic targets and strategies to ameliorate complications from diabetic wounds.

Funding. This work was supported by grants from the National Natural Science Foundation of China (81770827, 81471034, 81370910, and 81670764), the 863 Program for Young Scientists (S2015AA020927), Science and Technology Planning Project of Guangdong Province (2016B020238001), and the Special Fund for Science and Technology Development of Guangdong Province (2016A01010301).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. W.W. and C.Y. researched the literature, designed the methods and experiments, analyzed the data, performed the statistical analysis, interpreted the results, and wrote and approved the manuscript. X.y.W. performed the collection of clinical data and the measurement of MMP-9 in wound fluid. L.y.Z. and G.j.L. worked together on the collection of associated data and their interpretation, performed the molecular biological experiments, and collaborated with all of the other authors. L.y.Z. and T.t.Z. performed the animal experiments. D.L., C.W., and M.d.H. worked together on the isolation, culture, and associated assessment of keratinocytes; approved the manuscript; and collaborated with all of the other authors. L.Y. and M.R. also designed the experiments, discussed analyses and their interpretation, and helped to edit, revise, present, and approve the manuscript. L.Y. and M.R. are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.