GADD45a Promotes Active DNA Demethylation of the MMP-9 Promoter via Base Excision Repair Pathway in AGEs-Treated Keratinocytes and in Diabetic Male Rat Skin

GADD45a Promotes Active DNA Demethylation of the MMP-9 Promoter via Base Excision Repair Pathway... Abstract Diabetes elevates matrix metalloproteinase (MMP)-9 levels in the skin and its keratinocytes, and activated MMP-9 impairs skin wound healing. Epigenetic regulation of the DNA methylation status within the MMP-9 promoter plays an important role in the alteration of MMP-9 expression. Our aim was to investigate the role and mechanism of growth arrest and DNA damage-inducible 45a (GADD45a), a well-known DNA demethylation regulatory protein that mediates DNA methylation, in the regulation of MMP-9 expression. In this study, we showed that GADD45a was markedly upregulated in skin tissues from patients with diabetic foot ulcers, in diabetic rats, and in human keratinocyte (HaCaT) cells exposed to advanced glycation end products. We observed a substantial positive correlation between the levels of GADD45a and MMP-9 expression. Knockdown of GADD45a ameliorated the increase in MMP-9 transcription induced by a diabetic condition by inhibiting demethylation in the MMP-9 promoter and promoted diabetic HaCaT cell migration, but GADD45a knockdown did not affect HaCaT cell proliferation or apoptosis. Additionally, we demonstrated that overexpression of GADD45a activated MMP-9 expression by inducing promoter demethylation. Moreover, we found that GADD45a binds to the promoter of MMP-9 and recruits thymine-DNA glycosylase for base excision repair-mediated demethylation in diabetic HaCaT cells and diabetic rat skin. Our results reveal a mechanism in which GADD45a is required for demethylation of the MMP-9 promoter and the induction of diabetic wound healing. The inhibition of GADD45a might be a therapeutic strategy for diabetic foot ulcers. Diabetes mellitus (DM) is an ever-growing problem, and the number of diabetic adults worldwide is projected to reach 439 million in 2030 (1). Diabetic foot ulcers remain one of the major causes of amputation in working patients. Despite intensive therapy, only 60% of all diabetic foot ulcers are healed within 1 year of onset, and >10% of individuals with diabetic ulcers will eventually require lower extremity amputation (2). Studies have documented that diabetes activates matrix metalloproteinases (MMPs), especially MMP-9, which disrupts the homeostasis between extracellular matrix (ECM) synthesis and degradation, a phenomenon that precedes diabetic wound healing (3–7). Our previous studies have demonstrated that MMP-9 plays a pivotal role in the process of wound healing. Increased MMP-9 expression inhibits wound healing in patients with diabetic foot ulcers and in diabetic rats, whereas decreased MMP-9 expression improves diabetic chronic wound healing (8–10). Recently, many studies have shown that epigenetic mechanisms play a central role in controlling MMP-9 gene expression, including histone modification, DNA methylation, and noncoding RNAs (11). Increasing studies have suggested that epigenetic factors contribute to the pathogenesis of diabetic complications (12, 13). Genomic DNA methylation profiling of type 2 DM islets revealed that CpG loci exhibited a substantial hypomethylation phenotype, which may provide insight on DNA methylation and diabetic pathogenesis (14). Recently, Kowluru et al. (15) demonstrated that the hyperglycemic milieu induces hypomethylation of the retinal MMP-9 promoter, thereby regulating its transcription. Our previous studies have shown that advanced glycation end products (AGEs), a complex associated with diabetic complications (16–18), and tumor necrosis factor α induce MMP-9 promotor demethylation (19, 20). However, the molecular mechanisms underlying active DNA demethylation of the MMP-9 promoter in diabetic skin are complex and have not yet been clarified. Growth arrest and DNA damage protein 45a (GADD45a) is a member of a small family of stress-response genes and serves as a nexus between DNA repair and epigenetic gene regulation by mediating demethylation (21–23). GADD45a has been identified as a diabetes-associated gene that may link diabetic cardiomyopathy and baroreflex dysfunction, and 11 direct interacting proteins have been found to be highly associated with DM (24). However, whether GADD45a participates in MMP-9 promoter demethylation remains an important question. Here, we used skin tissues from patients with diabetic foot ulcers; the streptozotocin (STZ)-induced diabetic rat model, which is one of the most extensively characterized models of diabetic skin pathology; and diabetic human keratinocyte (HaCaT) cells to explore the expression levels of GADD45a and MMP-9, thereby providing a therapeutic target for diabetic foot ulcers. We found that GADD45a was significantly upregulated under diabetic conditions in vivo and in vitro, and its expression level was positively correlated with MMP-9 expression. Next, we showed that GADD45a targeted the promoter of MMP-9 and recruited thymine-DNA glycosylase (TDG) to induce MMP-9 promotor demethylation. Thus, our data corroborate a close link between the GADD45a- and 10–11 translocation protein (TET)-TDG–mediated DNA demethylation pathways, and our results provide a potential preventive strategy for chronic skins ulcers in DM. Materials and Methods Participants and tissue specimens Patient skin-tissue specimens were obtained from the Sun Yat-sen Memorial Hospital of Sun Yat-sen University (DM foot skin specimens, n = 5; non-DM skin specimens, n = 5). Samples were collected after written consent was obtained from the patients; the study received ethics approval and was conducted in accordance with the Declaration of Helsinki II. Diabetic foot ulcers from individuals with an active foot ulcer and nondiabetic foot specimens from healthy, nondiabetic subjects were obtained. The two groups were matched for age and sex. In addition, the diabetic groups were matched for the duration of diabetes and glycemic control, as defined by the levels of hemoglobin A1c. All of the diabetic foot patients had type 2 diabetes with neuropathy and peripheral artery disease, and one patient had an infection. The clinical information of the participants can be found in Supplemental Table 1. Animals and the establishment of an STZ-induced diabetic rat model All protocols for animal use and euthanasia were approved by the Sun Yat-sen University Institutional Animal Care and Use Committee. Twelve male Sprague-Dawley rats, each weighing ∼220 to 260 g, were randomized into the control (n = 6) or experimental (n = 6) group after adaptive feeding for 1 week. Animals were maintained on a 12-hour dark-light artificial cycle (lights on at 7:00 am) with food and water available ad libitum. Diabetes was induced by an intraperitoneal injection of 35 mg/kg STZ (Sigma-Aldrich, St. Louis, MO) in citrate buffer (pH 4.2) for 3 consecutive days. Animals injected with buffer alone were used as normal controls (n = 6). After a final injection of STZ at 72 hours, rats with a fasting blood glucose level of 16.7 mmol/L or higher, as determined from the tail vein blood using a OneTouch II® glucometer (LifeScan, Rochester, NY), were used as established DM models. STZ-treated rats were given daily injections of a small dose of protamine zinc insulin (∼0.5 to 3.0 U/d), as required to reduce mortality by maintaining blood glucose levels and avoiding ketosis. Rats were housed for 8 additional weeks before they were euthanized, and then, tissue was harvested. Skin specimens were fixed in 4% paraformaldehyde overnight at 4°C and then embedded in paraffin for morphometric studies. Paraffin-embedded tissues (4-μm sections) were stained for subsequent immunohistochemical examination. Cell culture The immortalized HaCaT line was cultured in Dulbecco’s modified Eagle’s medium, supplemented with 10% fetal calf serum and 1% antibiotics (100 U/ml penicillin and 100 mg/ml streptomycin) at 37°C in 5% CO2. Cells were grown to ∼30% to 40% confluence and treated with 100 µg/ml glycolaldehyde-modified AGE-bovine serum albumin (BSA; AGEs; Calbiochem, La Jolla, CA) or BSA (Calbiochem) at the same concentrations for ∼24 to 72 hours. DNA and RNA were isolated from cells or tissues using TRIzol (Life Technologies, Carlsbad, CA). Immunohistochemistry and data analysis, immunofluorescence, plasmid construct, and transfection For further details, please see Supplemental Materials and Methods. Quantitative real-time polymerase chain reaction analysis For quantitative real-time polymerase chain reaction (qPCR), total RNA was extracted, and 1 µg total RNA was converted into cDNA using a PrimeScriptTM RT Kit (TaKaRa, Otsu, Japan). Amplification and analysis were performed with a Roche LightCycler96 Real-Time PCR System (Roche, Basel, Switzerland). Values were normalized to the housekeeping gene actin B (ACTB). Alternatively, gene-specific primers were used (Supplemental Table 2). Cell proliferation assays and cell cycle and apoptosis analyses Cell proliferation was measured using the Cell Counting Kit-8 Assay (Dojindo Laboratories, Kumamoto, Japan). For cell analyses, cells were fixed and digested with RNase A, and propidium iodide was added before flow cytometry analysis. To examine cell apoptosis, cells were resuspended in binding buffer and stained with Annexin V and propidium iodide for 15 minutes in the dark before flow cytometry analysis. Details of the procedure are provided in Supplemental Materials and Methods. Wound-healing assay HaCaT cells were treated with AGEs or BSA and transfected with the indicated small interfering RNAs (siRNAs). After 48 hours, the attached cells were scratched with a 10-µl pipette tip, and images were captured at 0 hour using a microscope. The plates were returned to the 37°C incubator and cultured with 5% CO2 for 24 hours, at which time another set of images of the same wounds was captured. The wound widths were measured with ImageJ software (National Institutes of Health, Bethesda, MD) and were normalized and presented as a percentage of the wound measured at time zero. Western blotting analysis Total proteins extracted from each group of cells were resolved by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA). After blocking, the polyvinylidene difluoride membranes were washed three times for 10 minutes with tris(hydroxymethyl)aminomethane-buffered saline and Tween 20 at room temperature and incubated with primary antibodies (Table 1). Following extensive washing, membranes were incubated with secondary peroxidase-linked goat anti-rabbit immunoglobulin G (IgG) for 1 hour at room temperature. After washing three times for 10 minutes with tris(hydroxymethyl)aminomethane-buffered saline and Tween 20 at room temperature once more, the immunoreactivity was visualized by enhanced chemiluminescence, and membranes were exposed to the Imaging System G (Syngene, Cambridge, UK). Table 1. Antibodies Used Peptide/Protein Target  Antigen Sequence (if Known)  Name of Antibody  Manufacturer, Catalog No.  Species Raised in; Monoclonal or Polyclonal  Dilution Used  RRID  MMP-9  –  MMP-9 antibody  Abcam, ab76003  Rabbit; monoclonal  WB (1:1000)  AB_1310463  IHC (1:1000)  RNAP II  –  RNA polymerase II antibody  Abcam, ab817  Mouse; monoclonal  ChIP (1:500)  AB_306327  p300  –  KAT3B/p300 antibody  Abcam, ab14984  Rabbit; monoclonal  ChIP (1:500)  AB_306327  Thymine DNA glycosylase  –  TDG antibody  Abcam, ab106301  Rabbit; polyclonal  WB (1:1000)  AB_10973374  IP (1:50)  GADD45a  –  GADD45a antibody  Abcam, ab180768  Rabbit; polyclonal  WB (1:1000)  AB_2687480  IHC (1:200)  GADD45a  –  Gadd45a (H-165)  Santa Cruz Biotechnology, sc-797  Rabbit; polyclonal  IF (1:50)  AB_2232121  IP (1:50)  GADD45b  –  GADD45b antibody  Abcam, ab105060  Rabbit; polyclonal  WB (1:1000)  AB_10714129  GADD45g  –  GADD45g antibody human  Abcam, ab196774  Rabbit; monoclonal  WB (1:1000)  AB_2687481  FLAG  –  DYKDDDDK tag antibody  Cell Signaling Technology, 14793  Rabbit; monoclonal  IP (1:50)  AB_2572291  Peptide/Protein Target  Antigen Sequence (if Known)  Name of Antibody  Manufacturer, Catalog No.  Species Raised in; Monoclonal or Polyclonal  Dilution Used  RRID  MMP-9  –  MMP-9 antibody  Abcam, ab76003  Rabbit; monoclonal  WB (1:1000)  AB_1310463  IHC (1:1000)  RNAP II  –  RNA polymerase II antibody  Abcam, ab817  Mouse; monoclonal  ChIP (1:500)  AB_306327  p300  –  KAT3B/p300 antibody  Abcam, ab14984  Rabbit; monoclonal  ChIP (1:500)  AB_306327  Thymine DNA glycosylase  –  TDG antibody  Abcam, ab106301  Rabbit; polyclonal  WB (1:1000)  AB_10973374  IP (1:50)  GADD45a  –  GADD45a antibody  Abcam, ab180768  Rabbit; polyclonal  WB (1:1000)  AB_2687480  IHC (1:200)  GADD45a  –  Gadd45a (H-165)  Santa Cruz Biotechnology, sc-797  Rabbit; polyclonal  IF (1:50)  AB_2232121  IP (1:50)  GADD45b  –  GADD45b antibody  Abcam, ab105060  Rabbit; polyclonal  WB (1:1000)  AB_10714129  GADD45g  –  GADD45g antibody human  Abcam, ab196774  Rabbit; monoclonal  WB (1:1000)  AB_2687481  FLAG  –  DYKDDDDK tag antibody  Cell Signaling Technology, 14793  Rabbit; monoclonal  IP (1:50)  AB_2572291  Abbreviations: ChIP, chromatin immunoprecipitation; IF, immunofluorescence; IHC, immunohistochemistry; IP, immunoprecipitation; RNAP II, RNA polymerase II; RRID, Research Resource Identifier; WB, Western blot. View Large siRNAs, PCR primers, drugs, and antibodies siRNAs against GADD45a, GADD45b, GADD45g, TDG, TET2, and control siRNAs were purchased from GenePharma (Shanghai, China) and transfected with Lipofectamine 3000 (Invitrogen, Carlsbad, CA). All knockdown experiments showed at least a 70% reduction in target gene expression with a single siRNA. Primers were synthesized by GENEray biotechnology (Shanghai, China). The sequences of siRNAs used in this study are listed in Supplemental Table 3; we tested two siRNAs for every gene and screened for the most effective siRNA to use in the subsequent experiments. CRT 0044876 and 5-aza-2′-deoxycytidine were purchased from Sigma-Aldrich, diluted in dimethyl sulfoxide, and added to the cell culture media for a final concentration of 0.5 and 5 μM; the cells were then incubated for 48 or 72 hours. The following antibodies were used and listed in Table 1: MMP-9 (catalog no. ab76003; Abcam Cambridge, UK), RNA polymerase II (RNAP II; catalog no. ab81; Abcam), KAT3B/p300 (catalog no. ab14984; Abcam), TDG (catalog no. ab106301; Abcam), GADD45a (catalog no. ab180768; Abcam; catalog no. sc-797; Santa Cruz Biotechnology, Dallas, TX), GADD45b (catalog no. ab105060; Abcam), GADD45g (catalog no. ab196774; Abcam), and Flag (catalog no. 14793; Cell Signaling Technology, Danvers, MA). Quantitative DNA methylation analysis DNA methylation was analyzed using bisulfite-converted genomic DNA (EZ DNA Methylation Kit; Zymo Research, Irvine, CA) and the MassARRAY system (Sequenom, San Diego, CA), as previously reported. Details of the procedure are provided in Supplemental Materials and Methods. DNA methylation levels were quantified from mass spectra using Epityper Software v.1.2 (Sequenom) and were measured by the Beijing Genomics Institute. The MassARRAY primers for the MMP-9 locus are listed in Supplemental Table 4. Chromatin immunoprecipitation Cells were cross-linked with 1% paraformaldehyde at room temperature for 10 minutes, quenched with 0.5 M glycine, lysed in nuclei lysis buffer for 10 minutes, placed on ice, and sonicated seven times (15 seconds on and 59 seconds off) at 4°C to yield ∼150 to 250 base pair (bp) DNA fragments. The sheared chromatin was incubated with the indicated antibodies. Protein-DNA complexes were captured on protein A/G agarose, followed by washes in low-salt buffer, high-salt buffer, LiCI buffer, and Tris/EDTA buffer, according to the manufacturer’s recommendations (Millipore). After elution and reversal of the cross-link, DNA was extracted and amplified by qPCR. The amount of precipitated DNA was normalized to the amount of both the input DNA and the IgG-bound DNA. The primers were designed to amplify six regions upstream of the transcriptional start site (TSS; Supplemental Table 5). Coimmunoprecipitation HaCaT cells, in which Flag-tagged GADD45a was overexpressed, were lysed. The skin tissues of rats were cut and lysed overnight on ice with immunoprecipitation (IP) lysis buffer. Following preclearing of the lysate by centrifugation and incubation with control agarose resin for 1 hour at 4°C, lysates were incubated on a rotating wheel overnight at 4°C with specific antibodies or IgG as a control. Assay was performed using a coimmunoprecipitation (Co-IP) kit (Pierce, Rockford, IL), according to the manufacturer’s instructions. Statistical analysis Each experiment was performed at least three times. Data were presented as the means ± standard deviation (SD) where applicable and analyzed using SPSS software (IBM, New York, NY). Comparison of two groups was performed using an unpaired t test. Statistical significance was assessed with ANOVA, followed by a least significant difference test for multiple comparisons. P < 0.05 was considered statistically significant. Results GADD45a and MMP-9 are overexpressed in diabetic skin and cells exposed to AGEs To investigate the role of the GADD45a protein in diabetic skin, GADD45a expression and distribution in human and rat skin tissues were analyzed. We used immunohistochemistry to examine GADD45a expression in diabetic skin biopsy samples and normal skin tissue samples. Abundant GADD45a expression was observed in the epidermal skin tissues of DM patients and was localized in the dermal and subcutaneous connective tissues [Fig. 1(a)]. In contrast, GADD45a expression levels in diabetic skin were significantly higher than those in non-DM skin [P = 0.001, insulin receptor substrate (IRS):DM group 10.0 ± 2.8 vs non-DM group 0.6 ± 0.54; Fig. 1(c)]. In addition, we examined six diabetic rat skin samples and six nondiabetic rat skin samples. During the establishment of the diabetic rat model, no experimental animals died of hyperglycemia or other complications. We observed a substantial increase in GADD45a expression in the skin tissue of DM rats [P < 0.001, IRS:DM group 9.6 ± 2.2 vs non-DM group 0.6 ± 0.5; Fig. 1(b) and 1(c)]. We then examined MMP-9 expression and found that MMP-9 expression levels were significantly higher in diabetic skin tissues than in non-DM skin tissues [Fig. 1(d)–1(f)]. MMP-9 expression was positively correlated with GADD45a expression in human skin tissue (Supplemental Fig. 1). Figure 1. View largeDownload slide GADD45a is expressed in diabetic skin and cells exposed to AGEs. (a–c) Immunohistochemical staining of GADD45a in diabetic and nondiabetic (a) human or (b) rat skin tissues. (c) Quantification analysis is presented. Original magnification, ×200 (×400 in the next row).Data are presented as the means ± SD (n ≥ 5 per group). **P < 0.01 vs the normal group. (d–f) Immunohistochemical staining of MMP-9 in diabetic and nondiabetic (d) human or (e) rat skin tissues. (f) Quantification analysis is presented. Original magnification, ×200 (×400 in next row). Data are presented as the means ± SD of three independent experiments (n ≥ 5). **P < 0.01 vs normal group. (g–i) Expression of GADD45a, GADD45b, GADD45g, and MMP-9 in HaCaT cells treated with AGEs for 0, 24, 48, and 72 hours, as assayed by (g and h) qRT-PCR and (i) Western blotting. (j) Microscopy images show the expression of GADD45a (red) with nuclear staining by DAPI (blue) after treatment with AGEs or BSA for 72 hours. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs control group. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole; DM, diabetic skin tissue; IRS, insulin receptor substrate; NDM, normal, nondiabetic skin tissue; qRT-PCR, quantitative reverse transcription PCR. Figure 1. View largeDownload slide GADD45a is expressed in diabetic skin and cells exposed to AGEs. (a–c) Immunohistochemical staining of GADD45a in diabetic and nondiabetic (a) human or (b) rat skin tissues. (c) Quantification analysis is presented. Original magnification, ×200 (×400 in the next row).Data are presented as the means ± SD (n ≥ 5 per group). **P < 0.01 vs the normal group. (d–f) Immunohistochemical staining of MMP-9 in diabetic and nondiabetic (d) human or (e) rat skin tissues. (f) Quantification analysis is presented. Original magnification, ×200 (×400 in next row). Data are presented as the means ± SD of three independent experiments (n ≥ 5). **P < 0.01 vs normal group. (g–i) Expression of GADD45a, GADD45b, GADD45g, and MMP-9 in HaCaT cells treated with AGEs for 0, 24, 48, and 72 hours, as assayed by (g and h) qRT-PCR and (i) Western blotting. (j) Microscopy images show the expression of GADD45a (red) with nuclear staining by DAPI (blue) after treatment with AGEs or BSA for 72 hours. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs control group. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole; DM, diabetic skin tissue; IRS, insulin receptor substrate; NDM, normal, nondiabetic skin tissue; qRT-PCR, quantitative reverse transcription PCR. The superficial layer of the epidermis (∼60 μm) interfaces more directly with the outside world than the deeper dermal layer, and the epidermis is composed of ∼95% keratinocytes. We first confirmed that GADD45a and MMP-9 mRNA and protein were expressed in HaCaT cells treated with 100 µg/ml AGEs for 24, 48, and 72 hours. The mRNA levels of GADD45a were increased after 48 and 72 hours, and MMP-9 mRNA levels were increased after 72 hours [Fig. 1(g) and 1(h)]. GADD45a and MMP-9 protein expression levels were both increased after 24 hours, but they were significantly increased after treatment with AGEs for 72 hours [Fig. 1(i)]. In contrast, the mRNA and protein expression levels of GADD45b/g were not significantly altered after treatment with AGEs [Fig. 1(g) and 1(i)]. Therefore, we selected the time point of 72 hours for subsequent experiments. Fluorescence microscopy also showed that AGEs treatment increased the expression level of GADD45a, which was predominantly located in the nucleus [Fig. 1(j)]. GADD45a affects the biological function of HaCaT cells Next, we examined the impact of GADD45a on the biological function of HaCaT cells. We observed that the invitro proliferation and apoptosis rates of HaCaT cells transfected with siRNA targeting GADD45a were similar to those of cells transfected with a scrambled siRNA sequence (siRNA control) after treatment with AGEs [Fig. 2(a)–2(c)]; the interference efficiency is shown in Fig. 3(a) and 3(b). Next, we analyzed the proportion of cells in various stages of the cell cycle using flow cytometry. Likewise, the results were not different among the groups [Fig. 2(d)–2(f)]. Compared with control and mock cells, the wound-healing capacity of HaCaT–small interfering GADD45a (siGADD45a) cells cultured in normal medium was not different. Treatment with AGEs reduced the wound-healing capacity of HaCaT cells; however, siGADD45a transfection improved wound healing in cells treated with AGEs, even compared with controls [Fig. 2(g) and 2(h)]. These experiments showed that the silencing of the expression of GADD45a can effectively promote the migration ability of diabetic HaCaT cells. Figure 2. View largeDownload slide GADD45a affects the biological function of HaCaT cells. (a) Proliferation of BSA- or AGEs-treated cells following the downregulation of GADD45a or transfection of scrambled control siRNA (Mock), examined with cell counting kit-8. (b–f) HaCaT cells in a logarithmic growth phase (∼30% to 40% confluence) were treated with mock or siGADD45a and then exposed to 100 µg/mL AGEs or BSA. After 72 hours, the cells were harvested and then subjected to apoptosis and cell-cycle analysis. (b) Representative images of the apoptosis rate of the cells. (c) Quantification of apoptotic cells. (d) Representative images of the cell-cycle distribution profiles of the cells. (e and f) Quantification of the cell-cycle distribution. (g) Representative images of wound-healing assays. HaCaT cells were transiently transfected with siGADD45a or mock and treated with BSA or AGEs. After 72 hours, wound-healing assays were performed, as described in Materials and Methods. (h) Statistical analyses of wound-healing assays. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs control group. Ctrl, control; FITC-A, fluorescein isothiocyanate-area; N.S., not significant. OD, optical density; PE-A, phycoerythrin-area. Figure 2. View largeDownload slide GADD45a affects the biological function of HaCaT cells. (a) Proliferation of BSA- or AGEs-treated cells following the downregulation of GADD45a or transfection of scrambled control siRNA (Mock), examined with cell counting kit-8. (b–f) HaCaT cells in a logarithmic growth phase (∼30% to 40% confluence) were treated with mock or siGADD45a and then exposed to 100 µg/mL AGEs or BSA. After 72 hours, the cells were harvested and then subjected to apoptosis and cell-cycle analysis. (b) Representative images of the apoptosis rate of the cells. (c) Quantification of apoptotic cells. (d) Representative images of the cell-cycle distribution profiles of the cells. (e and f) Quantification of the cell-cycle distribution. (g) Representative images of wound-healing assays. HaCaT cells were transiently transfected with siGADD45a or mock and treated with BSA or AGEs. After 72 hours, wound-healing assays were performed, as described in Materials and Methods. (h) Statistical analyses of wound-healing assays. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs control group. Ctrl, control; FITC-A, fluorescein isothiocyanate-area; N.S., not significant. OD, optical density; PE-A, phycoerythrin-area. Figure 3. View largeDownload slide GADD45a is required for MMP-9 expression. (a–c) Relative levels of the GADD45 family genes and MMP-9 mRNA in HaCaT cells following siRNA-mediated knockdown of GADD45a, GADD45b, or GADD45g. Scrambled siRNA (Mock) was used as a control. RNA levels were normalized to ACTB mRNA. (d–f) Relative protein expression was analyzed by Western blotting. (g) qRT-PCR was used to analyze GADD45a and MMP-9 levels in HaCaT cells after inhibition of the GADD45a gene in the process of AGEs. (h) Relative levels of GADD45a/MMP-9 protein in HaCaT cells following siRNA-mediated knockdown of GADD45a and treatment with 100 µg/ml AGEs (48 hours). (i and j) Relative levels of (i) GADD45a/MMP-9 mRNA and (j) protein in HaCaT cells, 48 hours after infection with GADD45a-adenovirus (Ad-GADD45a). Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the control group. Ctrl, control; qRT-PCR, quantitative reverse transcription PCR. Figure 3. View largeDownload slide GADD45a is required for MMP-9 expression. (a–c) Relative levels of the GADD45 family genes and MMP-9 mRNA in HaCaT cells following siRNA-mediated knockdown of GADD45a, GADD45b, or GADD45g. Scrambled siRNA (Mock) was used as a control. RNA levels were normalized to ACTB mRNA. (d–f) Relative protein expression was analyzed by Western blotting. (g) qRT-PCR was used to analyze GADD45a and MMP-9 levels in HaCaT cells after inhibition of the GADD45a gene in the process of AGEs. (h) Relative levels of GADD45a/MMP-9 protein in HaCaT cells following siRNA-mediated knockdown of GADD45a and treatment with 100 µg/ml AGEs (48 hours). (i and j) Relative levels of (i) GADD45a/MMP-9 mRNA and (j) protein in HaCaT cells, 48 hours after infection with GADD45a-adenovirus (Ad-GADD45a). Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the control group. Ctrl, control; qRT-PCR, quantitative reverse transcription PCR. Knockdown of GADD45a promotes diabetic HaCaT cell migration via the transcriptional downregulation of MMP-9 Our previous experiments have shown that increased MMP-9 expression contributes to delayed wound healing in late-stage diabetic foot ulcers (8). As GADD45a is a powerful nuclear protein and is overexpressed in diabetic skin tissues, we proposed that it might be involved in regulating MMP-9 expression. To investigate further the association between GADD45a levels and MMP-9 expression, we transiently knocked down GADD45a expression in HaCaT cells by RNA interference without AGE treatment. Real-time PCR and Western blot results showed that GADD45a expression was significantly downregulated, as shown in Fig. 3(a) and 3(b). Downregulation of GADD45a resulted in decreased MMP-9 mRNA and protein expression compared with controls; however, downregulation of GADD45b/g did not affect the transcription or post-transcriptional levels of MMP-9 [Fig. 3(c)–3(f)]. Following exposure to AGEs, we found that the mRNA and protein levels of MMP-9 were also inhibited by siRNA GADD45a [Fig. 3(g) and 3(h)]. Then, we generated the recombinant adenovirus (Ad)-GADD45a and transfected HaCaT cells for 48 hours. The mRNA and protein levels of MMP-9 following the upregulation of GADD45a were significantly higher than those in the control group [Fig. 3(i) and 3(j)]. Based on the previous results, we concluded that GADD45a regulates MMP-9 expression. GADD45a activates MMP-9 transcription by inducing the demethylation of the MMP-9 promoter As our previous study showed that wound healing in diabetic skin was related to MMP-9 DNA demethylation (19), we reasoned that GADD45a might be involved in the activation of the transcription of MMP-9 via DNA demethylation. We detected MMP-9 promoter methylation in diabetic rat skin tissues and AGE-treated HaCaT cells. There are 12 CpG sites in the MMP-9 promoter of the rat [Fig. 4(a)]. Because the method of MassARRAY could not detect the −691-bp site, we evaluated the methylation levels of 11 CpG sites. In diabetic rat skin tissues, MMP-9 promoter methylation was significantly decreased at three CpG sites (−329, −97, and −56 bp) compared with nondiabetic skin tissues [Fig. 4(b)]. After treatment of HaCaT cells with AGEs for different durations, we evaluated 10 CpG sites in the MMP-9 gene promoter and observed DNA demethylation in several CpG sites; in particular, after 72 hours of treatment, there was a substantial decrease of methylation (Supplemental Fig. 2). Therefore, the time point of 72 hours may be important for substantial demethylation of the MMP-9 promoter by the demethylation protein complex. Among the 10 CpG sites of MMP-9 in HaCaT cells treated with AGEs for 72 hours, DNA methylation was significantly decreased at three CpG sites (−233, −223, and −36 bp) compared with those in control cells [Fig. 4(c) and 4(d)]. Figure 4. View largeDownload slide GADD45a activates MMP-9 transcription by promoter demethylation. (a) The position of CpG sites in the promoter of rat MMP-9 is indicated by “lollipops.” Arrows refer to the MMP-9 TSS. The location of two pairs of BSP primers is represented. (b) DNA methylation levels around the TSS of MMP-9 in diabetic or nondiabetic rat skin tissues were measured by MassARRAY. (c) The position of CpG sites in the promoter of human MMP-9 is indicated by lollipops. (d) DNA methylation levels around the TSS of MMP-9 in HaCaT cells treated with AGEs or BSA were measured by MassARRAY. (e and f) BSA- or AGEs-exposed HaCaT cells were transfected with siRNA against GADD45a or a scrambled control RNA (Mock). DNA methylation was measured by MassARRAY. (g) Methylation analysis of the MMP-9 promoter of HaCaT cells after treatment with Ad-GADD45a or control adenoviruses (Vector). Methylation levels of 10 CpG sites were measured by MassARRAY. (h) Location of DNA fragments (C1–C6) are indicated in the scheme. (i and j) ChIP assay monitoring the association of RNAP II and p300 at different regions of the MMP-9 promoter in AGE-treated HaCaT cells transfected with siGADD45a or control RNA (Mock). Values represent the enrichment of the bound protein fraction relative to IgG. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and #P < 0.01 vs the corresponding control group. BSP, bisulfite sequencing PCR; ChIP, chromatin immunoprecipitation; N.S., not significant. Figure 4. View largeDownload slide GADD45a activates MMP-9 transcription by promoter demethylation. (a) The position of CpG sites in the promoter of rat MMP-9 is indicated by “lollipops.” Arrows refer to the MMP-9 TSS. The location of two pairs of BSP primers is represented. (b) DNA methylation levels around the TSS of MMP-9 in diabetic or nondiabetic rat skin tissues were measured by MassARRAY. (c) The position of CpG sites in the promoter of human MMP-9 is indicated by lollipops. (d) DNA methylation levels around the TSS of MMP-9 in HaCaT cells treated with AGEs or BSA were measured by MassARRAY. (e and f) BSA- or AGEs-exposed HaCaT cells were transfected with siRNA against GADD45a or a scrambled control RNA (Mock). DNA methylation was measured by MassARRAY. (g) Methylation analysis of the MMP-9 promoter of HaCaT cells after treatment with Ad-GADD45a or control adenoviruses (Vector). Methylation levels of 10 CpG sites were measured by MassARRAY. (h) Location of DNA fragments (C1–C6) are indicated in the scheme. (i and j) ChIP assay monitoring the association of RNAP II and p300 at different regions of the MMP-9 promoter in AGE-treated HaCaT cells transfected with siGADD45a or control RNA (Mock). Values represent the enrichment of the bound protein fraction relative to IgG. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and #P < 0.01 vs the corresponding control group. BSP, bisulfite sequencing PCR; ChIP, chromatin immunoprecipitation; N.S., not significant. Knockdown of GADD45a led to increase methylation at the −36-bp CpG site with BSA exposure, thus supporting that GADD45a is required for the maintenance of MMP-9 promoter hypomethylation [Fig. 4(e)]. However, with AGEs treatment, siGADD45a transfection led to a 39.3% increase in DNA methylation around the TSS of MMP-9 [−233, −223, and −36 bp; Fig. 4(f)]. Then, we transfected HaCaT cells with Ad-GADD45a-Flag and detected the DNA methylation level of the MMP-9 promoter after 48 hours. We found that methylation was significantly reduced at three CpG sites [−712, −223, and −36 bp; Fig. 4(g)]. Indeed, knockdown of GADD45a in AGEs-treated HaCaT cells (expressing both GADD45a and MMP-9) led to decreased occupancy of both RNAP II and histone acetyltransferase p300 at the MMP-9 promoter [Fig. 4(h)–4(j)]. Therefore, knockdown of GADD45a inhibited the AGEs-induced demethylation of the MMP-9 promoter and inhibited MMP-9 expression in HaCaT cells. GADD45a mediates MMP-9 promotor demethylation by TDG-dependent base excision repair GADD45a removes methylated cytosine residues through nucleotide excision repair (NER) or base excision repair (BER) machinery (25–27). To detect the mechanism of MMP-9 demethylation, we examined the mRNA expression levels of the NER pathway [xeroderma pigmentosum (XP)A, XPC, XPF, and XPG], as well as the BER pathway (TDG), after AGEs exposure in HaCaT cells at different time points. The results showed that TDG expression was significantly higher in AGEs-treated HaCaT cells than in control cells at 72 hours, whereas the NER pathway mRNA expression levels did not change [Fig. 5(a)]. To clarify the role of TDG, we monitored MMP-9 expression after knockdown of TDG and showed that it was markedly reduced [Fig. 5(b) and 5(c)]; however, the MMP-9 mRNA expression levels were not significantly different after knockdown of NER genes (Supplemental Fig. 3). To show that GADD45a recruits TDG, an essential protein of the BER pathway, to mediate active demethylation, we treated cells with CRT0044876 (a BER inhibitor) and monitored MMP-9 expression. The BER inhibitor prevented the GADD45a-dependent expression of MMP-9, but GADD45a expression was not affected [Fig. 5(d)]. In addition, the GADD45a-mediated expression and DNA demethylation of MMP-9 were inhibited upon knockdown of TDG [Fig. 5(e) and 5(f)], further indicating that TDG cooperates with GADD45a to activate MMP-9 expression. To demonstrate that GADD45a and TDG regulate MMP-9 expression by adjusting the demethylation level of the MMP-9 promoter, we treated cells with 5-aza-2′-deoxycytidine, which causes DNA demethylation, and detected the MMP-9 expression levels. 5-Aza-2′-deoxycytidine promoted MMP-9 expression and prevented the siGADD45a- or siTDG-induced decrease in MMP-9 expression [Fig. 5(g) and 5(h)]. Figure 5. View largeDownload slide GADD45a-mediated MMP-9 promotor demethylation by TDG-dependent BER. (a) Relative mRNA levels of NER genes (XPA, XPC, XPF, and XPG) and the BER gene TDG in HaCaT cells treated with AGEs or BSA. RNA levels were normalized to ACTB mRNA. (b and c) HaCaT cells were transfected with siRNA against TDG or a scrambled control siRNA (Mock), and (b) qRT-PCR and (c) Western blots were used to analyze the level of MMP-9 and TDG expression. (d) Relative levels of MMP-9 and GADD45a mRNA in HaCaT cells treated with AGEs for 24 hours, followed by treatment with the BER inhibitor CRT0044876 (CRT) for 48 hours. DMSO was used as a solvent control, and GADD45a expression was considered a DNA damage-response control. RNA levels were normalized to ACTB mRNAs. (e) Relative levels of MMP-9 mRNA in HaCaT cells infected with Ad-GADD45a or control adenoviruses (Vector). Cells were cotransfected with siRNA against TDG or a control RNA (Mock). MMP-9 mRNA was measured by qPCR. RNA levels were normalized to ACTB mRNAs. (f) Methylation analysis of the MMP-9 promoter in HaCaT cells after treatment with siRNA against TDG (siTDG) or a scrambled control (Mock). Methylation of CpGs 1–10 was measured by MassARRAY. (g and h) Expression levels of GADD45a, TDG, and MMP-9 in HaCaT cells treated with 5-aza-2′-deoxycytidine (5aza) and transfected with siGADD45a or siTDG for 72 hours, as assayed by (g) qRT-PCR and (h) Western blotting. Data are presented as the mean ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs the corresponding control group. Ctrl, control; DMSO, dimethyl sulfoxide; NC, negative control; qRT-PCR, quantitative reverse transcription PCR. Figure 5. View largeDownload slide GADD45a-mediated MMP-9 promotor demethylation by TDG-dependent BER. (a) Relative mRNA levels of NER genes (XPA, XPC, XPF, and XPG) and the BER gene TDG in HaCaT cells treated with AGEs or BSA. RNA levels were normalized to ACTB mRNA. (b and c) HaCaT cells were transfected with siRNA against TDG or a scrambled control siRNA (Mock), and (b) qRT-PCR and (c) Western blots were used to analyze the level of MMP-9 and TDG expression. (d) Relative levels of MMP-9 and GADD45a mRNA in HaCaT cells treated with AGEs for 24 hours, followed by treatment with the BER inhibitor CRT0044876 (CRT) for 48 hours. DMSO was used as a solvent control, and GADD45a expression was considered a DNA damage-response control. RNA levels were normalized to ACTB mRNAs. (e) Relative levels of MMP-9 mRNA in HaCaT cells infected with Ad-GADD45a or control adenoviruses (Vector). Cells were cotransfected with siRNA against TDG or a control RNA (Mock). MMP-9 mRNA was measured by qPCR. RNA levels were normalized to ACTB mRNAs. (f) Methylation analysis of the MMP-9 promoter in HaCaT cells after treatment with siRNA against TDG (siTDG) or a scrambled control (Mock). Methylation of CpGs 1–10 was measured by MassARRAY. (g and h) Expression levels of GADD45a, TDG, and MMP-9 in HaCaT cells treated with 5-aza-2′-deoxycytidine (5aza) and transfected with siGADD45a or siTDG for 72 hours, as assayed by (g) qRT-PCR and (h) Western blotting. Data are presented as the mean ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs the corresponding control group. Ctrl, control; DMSO, dimethyl sulfoxide; NC, negative control; qRT-PCR, quantitative reverse transcription PCR. GADD45a binds to TDG to promote MMP-9 promoter demethylation To monitor the interaction between GADD45a and TDG, we performed Co-IP experiments with HaCaT cells overexpressing Flag-GADD45a and diabetic rat skin tissues and confirmed the association of TDG with GADD45a [Fig. 6(a) and 6(b)]. TDG is involved in a major active DNA demethylation pathway by excising the oxidation products, participating in TET-mediated oxidation of methylcytosines (28, 29). Our previous studies have shown that TET2 binds to the promoter of MMP-9, which expands its demethylation under diabetic conditions (19). To examine further whether GADD45a cooperates with the TET2–TDG demethylation pathway to activate MMP-9 expression, we examined the changes in 5-methylcytosine (5mc) and 5-hydroxymethylcytosine (5hmc) and found no substantial difference between the siGADD45a and mock groups [Fig. 6(c) and 6(d)] or between the Ad-GADD45a and vector groups treated with BSA or AGEs [Fig. 6(e) and 6(f)]; however, the levels of 5mc in the AGEs treatment groups were lower than those in the BSA treatment groups [Fig. 6(c)], which is consistent with our previous report (19). We also determined that there was no association between TET2 and GADD45a (data not shown), confirming that GADD45a did not directly affect the role of TET2 or the overall methylation level in diabetic HaCaT cells. Moreover, the chromatin immunoprecipitation (ChIP) experiment revealed that GADD45a has a substantial enrichment at the MMP-9 promoter in AGEs-treated HaCaT cells but not IgG or TDG [Fig. 6(g) and 6(h)], thus supporting that GADD45a binds to the MMP-9 promoter and recruits TDG to mediate demethylation. Figure 6. View largeDownload slide GADD45a binds to TDG to promote MMP-9 promoter demethylation. (a) FLAG-tagged GADD45a was immunoprecipitated from lysates of transfected HaCaT cells, and coprecipitated TDG was evaluated on immunoblots with antibodies against TDG and FLAG. The input sample was 10% of the lysate that was used for IP. The positions of molecular weight marker proteins are indicated. (b) GADD45a was immunoprecipitated from lysates of diabetic rat skin tissues. (c and d) 5mc and 5hmc levels, which reflect global gene methylation, were measured by the quantification of methylcytosine and hydroxymethylcytosine, respectively. 5mc and 5hmc levels in HaCaT cells, 48 hours after transfection with siGADD45a or control RNA (Mock) and treatment with BSA or AGEs. (e and f) 5mc and 5hmc levels in the MMP-9 promoter were measured using HaCaT cells by quantifying methylcytosine and hydroxymethylcytosine levels after treatment with Ad-GADD45a or control adenoviruses (Vector). (g and h) ChIP assays showing the enrichment of GADD45a and TDG at the MMP-9 promoter in HaCaT cells treated with BSA or AGEs. Coprecipitated DNA was analyzed by qPCR. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the corresponding control group. Figure 6. View largeDownload slide GADD45a binds to TDG to promote MMP-9 promoter demethylation. (a) FLAG-tagged GADD45a was immunoprecipitated from lysates of transfected HaCaT cells, and coprecipitated TDG was evaluated on immunoblots with antibodies against TDG and FLAG. The input sample was 10% of the lysate that was used for IP. The positions of molecular weight marker proteins are indicated. (b) GADD45a was immunoprecipitated from lysates of diabetic rat skin tissues. (c and d) 5mc and 5hmc levels, which reflect global gene methylation, were measured by the quantification of methylcytosine and hydroxymethylcytosine, respectively. 5mc and 5hmc levels in HaCaT cells, 48 hours after transfection with siGADD45a or control RNA (Mock) and treatment with BSA or AGEs. (e and f) 5mc and 5hmc levels in the MMP-9 promoter were measured using HaCaT cells by quantifying methylcytosine and hydroxymethylcytosine levels after treatment with Ad-GADD45a or control adenoviruses (Vector). (g and h) ChIP assays showing the enrichment of GADD45a and TDG at the MMP-9 promoter in HaCaT cells treated with BSA or AGEs. Coprecipitated DNA was analyzed by qPCR. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the corresponding control group. Discussion Our findings identified GADD45a as an important signature protein in diabetic skin ulcers. GADD45a was shown to be markedly upregulated in the skin tissue of patients with diabetic foot ulcers and in the skin of diabetic Sprague- Dawley rats; moreover, the upregulation of GADD45a was associated with MMP-9 expression. More importantly, our study showed that inhibiting GADD45a increased the migration ability of diabetic skin cells by downregulating the expression of MMP-9 in vitro. The suppression of TDG levels to prevent GADD45a-mediated MMP-9 demethylation may be a mechanism whereby overexpressed MMP-9 degrades the ECM in diabetic skin ulcers. Our study also demonstrated that GADD45a protein interacts with TDG in vitro and in vivo. Based on these findings, we propose the model illustrated in Fig. 7. The targeting of GADD45a, a demethylation-promoting protein, may be a good approach to preventing the amputation of a diabetic foot. Figure 7. View largeDownload slide A model of the role of GADD45a in the regulation of MMP-9 demethylation. (Lower) In normal cells, the promoter of MMP-9 is methylated, and MMP-9 expression levels are low. (Upper) In diabetic cells, GADD45a is overexpressed and binds to the MMP-9 promoter. GADD45a recruits TDG to induce promoter demethylation via BER. Demethylated MMP-9 harbors RNAP II (Pol II) and active chromatin marker p300, thus promoting MMP-9 expression. Figure 7. View largeDownload slide A model of the role of GADD45a in the regulation of MMP-9 demethylation. (Lower) In normal cells, the promoter of MMP-9 is methylated, and MMP-9 expression levels are low. (Upper) In diabetic cells, GADD45a is overexpressed and binds to the MMP-9 promoter. GADD45a recruits TDG to induce promoter demethylation via BER. Demethylated MMP-9 harbors RNAP II (Pol II) and active chromatin marker p300, thus promoting MMP-9 expression. Evidence has revealed that hyperglycemia could induce genomic hypomethylation and aberrant gene expression within the liver of type 1 DM rats (30), the skin of type 1 DM zebrafish (31), and human primary aortic endothelial cells (32, 33). However, the expression level and effect of the GADD45a protein remain poorly understood in diabetic skin tissues. Moreover, it is unknown whether GADD45a participates in regulating MMP-9 promoter methylation in diabetic skin cells. Herein, our data indicate that GADD45a is capable of upregulating MMP-9 expression in diabetic skin cells, but its downregulation leads to aberrant MMP-9 activation without AGEs stimulation, suggesting that GADD45a is an essential factor for MMP-9 expression in HaCaT cells, which warrants investigation in a large number of clinical samples. In our study, GADD45a/b/g expression was different in diabetic skin tissues and cells. These differences revealed that GADD45 family proteins were tissue specific in human tissues and diseases. To explain the roles of GADD45a in MMP-9 expression and promoter demethylation under the diabetic conditions, GADD45a interference RNA or Ad-GADD45a was constructed and transfected into HaCaT cells. We demonstrated that downregulated GADD45a could reverse the pathological effects of MMP-9 on the degradation of ECM proteins by promoting the migration of diabetic cells. Moreover, GADD45a interference RNA did not affect cell proliferation or apoptosis in our study, indicating that the extent of inhibition of GADD45a is appropriate to study the change of the demethylation level of the MMP-9 promoter. The expression of GADD45a leads to an important change in the CpG sites of the MMP-9 promoter, which undergoes up- or downregulation of methylation after the corresponding changes of GADD45a expression. Therefore, GADD45a expression should be inhibited in an appropriate range, which is a more effective reduction of MMP-9 demethylation and more beneficial to the treatment of a diabetic foot ulcer. Regarding the demethylation mechanism of GADD45a, in our study, we indicated that the GADD45a protein bound and interacted with the −668- to −538-bp segments of the MMP-9 promoter. The segments are rich with transcription factor binding sites, such as nuclear factor κB, Sp1 transcription factor, and AP-1 transcription factor; therefore, they may involve the process of GADD45a-mediated demethylation, thereby regulating MMP-9 expression. Actually, the role of DNA demethylation is to affect transcription factor binding (34, 35). Moreover, our results showed that MMP-9 and GADD45a protein levels increased after 24 hours by treatment with AGEs. This phenomenon suggests that GADD45a may be involved in the demethylation of MMP-9 in the early stage, and there may exist another mechanism that adjusts the expression of MMP-9, such as regulation by noncoding RNA, histone modifications, and transcription factor activity. Therefore, the function of this segment in the demethylation of the MMP-9 promoter and regulation of MMP-9 gene expression should be investigated in further studies. Multiple mechanisms underlying active demethylation have been proposed, including C–C bond cleavage on the exocyclic methyl group and BER and NER pathways of replacement of the methylated cytosine (36). Our study showed that GADD45a interacts with TDG and helps mediate the demethylation of the MMP-9 promoter using BER and TDG. TDG can replace 5mc with unmethylated cytosine by excising the interactive oxidation products of DNA methylation, including 5-formylcytosine and 5-carboxylcytosine (28, 37, 38). We observed that TDG was not enriched in the MMP-9 promoter. Therefore, GADD45a cannot stimulate TDG to undergo specific substrate binding; rather, GADD45a may act as a scaffold protein that directs TDG to the promoter of target genes and excises the oxidation products (5-formylcytosine and 5-carboxylcytosine). The interaction of GADD45a and TDG induced the demethylation of gene promoters, which is consistent with a recently proposed model (39, 40). We analyzed the methylation levels in the MMP-9 promoter following AGEs treatment of different durations. The results showed that methylation was reduced, not only after 72 hours but also at 48 and 96 hours (Supplemental Fig. 3). We found that TDG levels significantly increased only after 72 hours and that its levels were decreased after 96 hours; therefore, the location of demethylation proteins seems to be more important than their expression levels in the process of demethylation. Further investigation will be required to reveal the molecular mechanisms by which GADD45 proteins regulate the role of TDG enzymes in specific loci in the chromatin contexts. In addition, GADD45a interacts with TET-mediated oxidative DNA demethylation by directly binding to each other (41); therefore, many auxiliary factors may be involved in modulating the activity of the TET-TDG system. Moreover, we found that knockdown of GADD45a had no effect on the levels of 5mc and 5hmc but did influence MMP-9 promoter hypermethylation and gene downregulation. This is consistent with GADD45-mediated demethylation being restricted to single-copy genes, and global methylation appears to be unaffected by GADD45a expression (42, 43); this phenomenon was also demonstrated in GADD45a knockout mice (44). Our previous studies demonstrated that TET2 contributes to the development of diabetic skin ulcers by promoting MMP-9 promoter demethylation (19, 20). Therefore, we favor a model in which GADD45a promotes active DNA demethylation by recruiting TDG to target loci and further promotes the TET-TDG demethylation pathway. In summary, we propose that GADD45a interacts with target loci in the MMP-9 promoter and promotes HaCaT cells migration by inhibiting MMP-9 gene expression. These findings reveal a role of GADD45a in diabetic skin ulcers and provide a therapeutic target for chronic refractory skin ulcers. The changes of DNA methylation occur frequently in diabetic tissues and cells so the GADD45a-mediated DNA demethylation may not be limited to the MMP-9 promoter but may also be relevant to the global epigenetic changes observed during the occurrence of diabetes and its complications. Abbreviations: 5hmc 5-hydroxymethylcytosine 5mc 5-methylcytosine ACTB actin B Ad adenovirus AGE advanced glycation end product BER base excision repair bp base pair BSA bovine serum albumin ChIP chromatin immunoprecipitation Co-IP coimmunoprecipitation DM diabetes mellitus ECM extracellular matrix GADD45a growth arrest and DNA damage-inducible 45a HaCaT human keratinocyte IgG immunoglobulin G IP immunoprecipitation IRS insulin receptor substrate MMP matrix metalloproteinase NER nucleotide excision repair, qPCR, quantitative polymerase chain reaction RNAP II RNA polymerase II SD standard deviation siGADD45a small interfering growth arrest and DNA damage-inducible 45a siRNA small interfering RNA STZ streptozotocin TDG thymine-DNA glycosylase TET 10–11 translocation protein TSS transcriptional start site XP xeroderma pigmentosum. Acknowledgments We are grateful to Kai Huang of the Department of Vascular Surgery and Gang Zeng of the Department of Orthopaedics at the Sun Yat-sen Memorial Hospital of Sun Yat-sen University for providing human skin tissues. Financial Support: This study was supported by National Natural Science Foundation of China Grants 81370910 (to M.R.) and 81471034 (to L.Y.); Natural Science Foundation of Guangdong Province Grant S2013010016443 (to M.R.); Science and Technology Planning Project of Guangdong Province, China Grant 2014A020212161 (to J.Z.); Grant [2013]163 from the Key Laboratory of Malignant Tumor Molecular Mechanism and Translational Medicine of Guangzhou Bureau of Science and Information Technology; Grant KLB09001 from the Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institutes; and a grant from the Guangdong Science and Technology Department (2015B050501004). Author Contributions: L.Z. and W.W. contributed to the conception and design of the study; acquisition, analysis, and interpretation of the data; and drafting and writing of the article. C.Y., T.Z., M.H., X.W., N.L., K.S., C.W., and J.Z. participated in the acquisition of data and critical revision of the manuscript, collected data, and revised the article. M.R. and L.Y. analyzed data, planned the experiments, performed the literature search, and revised the article. M.R. and L.Y. are the guarantors of this work and, as such, had full access to all of the data in the study; they take responsibility for the integrity of the data and the accuracy of the data analysis. All of the authors approved the final version to be published. Disclosure Summary: The authors have nothing to disclose. References 1. Shaw JE, Sicree RA, Zimmet PZ. Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract . 2010; 87( 1): 4– 14. Google Scholar CrossRef Search ADS PubMed  2. Hinchliffe RJ, Andros G, Apelqvist J, Bakker K, Friederichs S, Lammer J, Lepantalo M, Mills JL, Reekers J, Shearman CP, Valk G, Zierler RE, Schaper NC. A systematic review of the effectiveness of revascularization of the ulcerated foot in patients with diabetes and peripheral arterial disease [published correction appears in Diabetes Metab Res Rev. 28(4):376]. Diabetes Metab Res Rev . 2012; 28( Suppl 1): 179– 217. Google Scholar CrossRef Search ADS PubMed  3. Lobmann R, Schultz G, Lehnert H. Proteases and the diabetic foot syndrome: mechanisms and therapeutic implications. Diabetes Care . 2005; 28( 2): 461– 471. Google Scholar CrossRef Search ADS PubMed  4. Chao CY, Cheing GL. Microvascular dysfunction in diabetic foot disease and ulceration. Diabetes Metab Res Rev . 2009; 25( 7): 604– 614. Google Scholar CrossRef Search ADS PubMed  5. Kyriakides TR, Wulsin D, Skokos EA, Fleckman P, Pirrone A, Shipley JM, Senior RM, Bornstein P. Mice that lack matrix metalloproteinase-9 display delayed wound healing associated with delayed reepithelization and disordered collagen fibrillogenesis. Matrix Biol . 2009; 28( 2): 65– 73. Google Scholar CrossRef Search ADS PubMed  6. Singh K, Agrawal NK, Gupta SK, Mohan G, Chaturvedi S, Singh K. Differential expression of matrix metalloproteinase-9 gene in wounds of type 2 diabetes mellitus cases with susceptible -1562C>T genotypes and wound severity. Int J Low Extrem Wounds . 2014; 13( 2): 94– 102. Google Scholar CrossRef Search ADS PubMed  7. Zhai J, Wang Y. MDI 301, a synthetic retinoid, depressed levels of matrix metalloproteinases and oxidative stress in diabetic dermal fibroblasts. Oncotarget . 2017; 8( 27): 43889– 43896. Google Scholar PubMed  8. Yang C, Zhu P, Yan L, Chen L, Meng R, Lao G. Dynamic changes in matrix metalloproteinase 9 and tissue inhibitor of metalloproteinase 1 levels during wound healing in diabetic rats. J Am Podiatr Med Assoc . 2009; 99( 6): 489– 496. Google Scholar CrossRef Search ADS PubMed  9. Li N, Luo HC, Yang C, Deng JJ, Ren M, Xie XY, Lin DZ, Yan L, Zhang LM. Cationic star-shaped polymer as an siRNA carrier for reducing MMP-9 expression in skin fibroblast cells and promoting wound healing in diabetic rats. Int J Nanomedicine . 2014; 9: 3377– 3387. Google Scholar CrossRef Search ADS PubMed  10. Li N, Luo HC, Ren M, Zhang LM, Wang W, Pan CL, Yang LQ, Lao GJ, Deng JJ, Mai KJ, Sun K, Yang C, Yan L. Efficiency and safety of β-CD-(D3)7 as siRNA carrier for decreasing matrix metalloproteinase-9 expression and improving wound healing in diabetic rats. ACS Appl Mater Interfaces . 2017; 9( 20): 17417– 17426. Google Scholar CrossRef Search ADS PubMed  11. Labrie M, St-Pierre Y. Epigenetic regulation of mmp-9 gene expression. Cell Mol Life Sci . 2013; 70( 17): 3109– 3124. Google Scholar CrossRef Search ADS PubMed  12. Ling C, Groop L. Epigenetics: a molecular link between environmental factors and type 2 diabetes. Diabetes . 2009; 58( 12): 2718– 2725. Google Scholar CrossRef Search ADS PubMed  13. Villeneuve LM, Reddy MA, Natarajan R. Epigenetics: deciphering its role in diabetes and its chronic complications. Clin Exp Pharmacol Physiol . 2011; 38( 7): 451– 459. Google Scholar CrossRef Search ADS PubMed  14. Volkmar M, Dedeurwaerder S, Cunha DA, Ndlovu MN, Defrance M, Deplus R, Calonne E, Volkmar U, Igoillo-Esteve M, Naamane N, Del Guerra S, Masini M, Bugliani M, Marchetti P, Cnop M, Eizirik DL, Fuks F. DNA methylation profiling identifies epigenetic dysregulation in pancreatic islets from type 2 diabetic patients. EMBO J . 2012; 31( 6): 1405– 1426. Google Scholar CrossRef Search ADS PubMed  15. Kowluru RA, Shan Y, Mishra M. Dynamic DNA methylation of matrix metalloproteinase-9 in the development of diabetic retinopathy. Lab Invest . 2016; 96( 10): 1040– 1049. Google Scholar CrossRef Search ADS PubMed  16. Peppa M, Stavroulakis P, Raptis SA. Advanced glycoxidation products and impaired diabetic wound healing. Wound Repair Regen . 2009; 17( 4): 461– 472. Google Scholar CrossRef Search ADS PubMed  17. Zhu P, Yang C, Chen LH, Ren M, Lao GJ, Yan L. Impairment of human keratinocyte mobility and proliferation by advanced glycation end products-modified BSA. Arch Dermatol Res . 2011; 303( 5): 339– 350. Google Scholar CrossRef Search ADS PubMed  18. Tan Q, Wang W, Yang C, Zhang J, Sun K, Luo HC, Mai LF, Lao Y, Yan L, Ren M. α-ketoglutarate is associated with delayed wound healing in diabetes. Clin Endocrinol (Oxf) . 2016; 85( 1): 54– 61. Google Scholar CrossRef Search ADS PubMed  19. Zhang J, Yang C, Wang C, Liu D, Lao G, Liang Y, Sun K, Luo H, Tan Q, Ren M, Yan L. AGE-induced keratinocyte MMP-9 expression is linked to TET2-mediated CpG demethylation. Wound Repair Regen . 2016; 24( 3): 489– 500. Google Scholar CrossRef Search ADS PubMed  20. Ling L, Ren M, Yang C, Lao G, Chen L, Luo H, Feng Z, Yan L. Role of site-specific DNA demethylation in TNFα-induced MMP9 expression in keratinocytes. J Mol Endocrinol . 2013; 50( 3): 279– 290. Google Scholar CrossRef Search ADS PubMed  21. Niehrs C, Schäfer A. Active DNA demethylation by Gadd45 and DNA repair. Trends Cell Biol . 2012; 22( 4): 220– 227. Google Scholar CrossRef Search ADS PubMed  22. Salvador JM, Brown-Clay JD, Fornace AJ, Jr. Gadd45 in stress signaling, cell cycle control, and apoptosis. Adv Exp Med Biol . 2013; 793: 1– 19. Google Scholar CrossRef Search ADS PubMed  23. Schäfer A. Gadd45 proteins: key players of repair-mediated DNA demethylation. Adv Exp Med Biol . 2013; 793: 35– 50. Google Scholar CrossRef Search ADS PubMed  24. Wang N, Yang C, Xie F, Sun L, Su X, Wang Y, Wei R, Zhang R, Li X, Yang B, Ai J. Gadd45α: a novel diabetes-associated gene potentially linking diabetic cardiomyopathy and baroreflex dysfunction. PLoS One . 2012; 7( 12): e49077. Google Scholar CrossRef Search ADS PubMed  25. Barreto G, Schäfer A, Marhold J, Stach D, Swaminathan SK, Handa V, Döderlein G, Maltry N, Wu W, Lyko F, Niehrs C. Gadd45a promotes epigenetic gene activation by repair-mediated DNA demethylation. Nature . 2007; 445( 7128): 671– 675. Google Scholar CrossRef Search ADS PubMed  26. Cortellino S, Xu J, Sannai M, Moore R, Caretti E, Cigliano A, Le Coz M, Devarajan K, Wessels A, Soprano D, Abramowitz LK, Bartolomei MS, Rambow F, Bassi MR, Bruno T, Fanciulli M, Renner C, Klein-Szanto AJ, Matsumoto Y, Kobi D, Davidson I, Alberti C, Larue L, Bellacosa A. Thymine DNA glycosylase is essential for active DNA demethylation by linked deamination-base excision repair. Cell . 2011; 146( 1): 67– 79. Google Scholar CrossRef Search ADS PubMed  27. Schmitz KM, Schmitt N, Hoffmann-Rohrer U, Schäfer A, Grummt I, Mayer C. TAF12 recruits Gadd45a and the nucleotide excision repair complex to the promoter of rRNA genes leading to active DNA demethylation. Mol Cell . 2009; 33( 3): 344– 353. Google Scholar CrossRef Search ADS PubMed  28. He YF, Li BZ, Li Z, Liu P, Wang Y, Tang Q, Ding J, Jia Y, Chen Z, Li L, Sun Y, Li X, Dai Q, Song CX, Zhang K, He C, Xu GL. Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science . 2011; 333( 6047): 1303– 1307. Google Scholar CrossRef Search ADS PubMed  29. Pastor WA, Aravind L, Rao A. TETonic shift: biological roles of TET proteins in DNA demethylation and transcription. Nat Rev Mol Cell Biol . 2013; 14( 6): 341– 356. Google Scholar CrossRef Search ADS PubMed  30. Williams KT, Garrow TA, Schalinske KL. Type I diabetes leads to tissue-specific DNA hypomethylation in male rats. J Nutr . 2008; 138( 11): 2064– 2069. Google Scholar CrossRef Search ADS PubMed  31. Olsen AS, Sarras MP, Jr, Leontovich A, Intine RV. Heritable transmission of diabetic metabolic memory in zebrafish correlates with DNA hypomethylation and aberrant gene expression. Diabetes . 2012; 61( 2): 485– 491. Google Scholar CrossRef Search ADS PubMed  32. Pirola L, Balcerczyk A, Tothill RW, Haviv I, Kaspi A, Lunke S, Ziemann M, Karagiannis T, Tonna S, Kowalczyk A, Beresford-Smith B, Macintyre G, Kelong M, Hongyu Z, Zhu J, El-Osta A. Genome-wide analysis distinguishes hyperglycemia regulated epigenetic signatures of primary vascular cells. Genome Res . 2011; 21( 10): 1601– 1615. Google Scholar CrossRef Search ADS PubMed  33. Dhliwayo N, Sarras MP, Jr, Luczkowski E, Mason SM, Intine RV. Parp inhibition prevents ten-eleven translocase enzyme activation and hyperglycemia-induced DNA demethylation. Diabetes . 2014; 63( 9): 3069– 3076. Google Scholar CrossRef Search ADS PubMed  34. Al-Romaih K, Sadikovic B, Yoshimoto M, Wang Y, Zielenska M, Squire JA. Decitabine-induced demethylation of 5′ CpG island in GADD45A leads to apoptosis in osteosarcoma cells. Neoplasia . 2008; 10( 5): 471– 480. Google Scholar CrossRef Search ADS PubMed  35. Hashimoto K, Oreffo RO, Gibson MB, Goldring MB, Roach HI. DNA demethylation at specific CpG sites in the IL1B promoter in response to inflammatory cytokines in human articular chondrocytes. Arthritis Rheum . 2009; 60( 11): 3303– 3313. Google Scholar CrossRef Search ADS PubMed  36. Walsh CP, Xu GL. Cytosine methylation and DNA repair. Curr Top Microbiol Immunol . 2006; 301: 283– 315. Google Scholar PubMed  37. Ito S, D’Alessio AC, Taranova OV, Hong K, Sowers LC, Zhang Y. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature . 2010; 466( 7310): 1129– 1133. Google Scholar CrossRef Search ADS PubMed  38. Maiti A, Drohat AC. Thymine DNA glycosylase can rapidly excise 5-formylcytosine and 5-carboxylcytosine: potential implications for active demethylation of CpG sites. J Biol Chem . 2011; 286( 41): 35334– 35338. Google Scholar CrossRef Search ADS PubMed  39. Arab K, Park YJ, Lindroth AM, Schäfer A, Oakes C, Weichenhan D, Lukanova A, Lundin E, Risch A, Meister M, Dienemann H, Dyckhoff G, Herold-Mende C, Grummt I, Niehrs C, Plass C. Long noncoding RNA TARID directs demethylation and activation of the tumor suppressor TCF21 via GADD45A. Mol Cell . 2014; 55( 4): 604– 614. Google Scholar CrossRef Search ADS PubMed  40. Xu GL, Walsh CP. Enzymatic DNA oxidation: mechanisms and biological significance. BMB Rep . 2014; 47( 11): 609– 618. Google Scholar CrossRef Search ADS PubMed  41. Kienhöfer S, Musheev MU, Stapf U, Helm M, Schomacher L, Niehrs C, Schäfer A. GADD45a physically and functionally interacts with TET1. Differentiation . 2015; 90( 1–3): 59– 68. Google Scholar CrossRef Search ADS PubMed  42. Gavin DP, Kusumo H, Sharma RP, Guizzetti M, Guidotti A, Pandey SC. Gadd45b and N-methyl-D-aspartate induced DNA demethylation in postmitotic neurons. Epigenomics . 2015; 7( 4): 567– 579. Google Scholar CrossRef Search ADS PubMed  43. Schäfer A, Schomacher L, Barreto G, Döderlein G, Niehrs C. Gemcitabine functions epigenetically by inhibiting repair mediated DNA demethylation. PLoS One . 2010; 5( 11): e14060. Google Scholar CrossRef Search ADS PubMed  44. Engel N, Tront JS, Erinle T, Nguyen N, Latham KE, Sapienza C, Hoffman B, Liebermann DA. Conserved DNA methylation in Gadd45a(−/−) mice. Epigenetics . 2009; 4( 2): 98– 99. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Endocrinology Oxford University Press

GADD45a Promotes Active DNA Demethylation of the MMP-9 Promoter via Base Excision Repair Pathway in AGEs-Treated Keratinocytes and in Diabetic Male Rat Skin

Loading next page...
 
/lp/ou_press/gadd45a-promotes-active-dna-demethylation-of-the-mmp-9-promoter-via-DoScoU1szl
Publisher
Endocrine Society
Copyright
Copyright © 2018 Endocrine Society
ISSN
0013-7227
eISSN
1945-7170
D.O.I.
10.1210/en.2017-00686
Publisher site
See Article on Publisher Site

Abstract

Abstract Diabetes elevates matrix metalloproteinase (MMP)-9 levels in the skin and its keratinocytes, and activated MMP-9 impairs skin wound healing. Epigenetic regulation of the DNA methylation status within the MMP-9 promoter plays an important role in the alteration of MMP-9 expression. Our aim was to investigate the role and mechanism of growth arrest and DNA damage-inducible 45a (GADD45a), a well-known DNA demethylation regulatory protein that mediates DNA methylation, in the regulation of MMP-9 expression. In this study, we showed that GADD45a was markedly upregulated in skin tissues from patients with diabetic foot ulcers, in diabetic rats, and in human keratinocyte (HaCaT) cells exposed to advanced glycation end products. We observed a substantial positive correlation between the levels of GADD45a and MMP-9 expression. Knockdown of GADD45a ameliorated the increase in MMP-9 transcription induced by a diabetic condition by inhibiting demethylation in the MMP-9 promoter and promoted diabetic HaCaT cell migration, but GADD45a knockdown did not affect HaCaT cell proliferation or apoptosis. Additionally, we demonstrated that overexpression of GADD45a activated MMP-9 expression by inducing promoter demethylation. Moreover, we found that GADD45a binds to the promoter of MMP-9 and recruits thymine-DNA glycosylase for base excision repair-mediated demethylation in diabetic HaCaT cells and diabetic rat skin. Our results reveal a mechanism in which GADD45a is required for demethylation of the MMP-9 promoter and the induction of diabetic wound healing. The inhibition of GADD45a might be a therapeutic strategy for diabetic foot ulcers. Diabetes mellitus (DM) is an ever-growing problem, and the number of diabetic adults worldwide is projected to reach 439 million in 2030 (1). Diabetic foot ulcers remain one of the major causes of amputation in working patients. Despite intensive therapy, only 60% of all diabetic foot ulcers are healed within 1 year of onset, and >10% of individuals with diabetic ulcers will eventually require lower extremity amputation (2). Studies have documented that diabetes activates matrix metalloproteinases (MMPs), especially MMP-9, which disrupts the homeostasis between extracellular matrix (ECM) synthesis and degradation, a phenomenon that precedes diabetic wound healing (3–7). Our previous studies have demonstrated that MMP-9 plays a pivotal role in the process of wound healing. Increased MMP-9 expression inhibits wound healing in patients with diabetic foot ulcers and in diabetic rats, whereas decreased MMP-9 expression improves diabetic chronic wound healing (8–10). Recently, many studies have shown that epigenetic mechanisms play a central role in controlling MMP-9 gene expression, including histone modification, DNA methylation, and noncoding RNAs (11). Increasing studies have suggested that epigenetic factors contribute to the pathogenesis of diabetic complications (12, 13). Genomic DNA methylation profiling of type 2 DM islets revealed that CpG loci exhibited a substantial hypomethylation phenotype, which may provide insight on DNA methylation and diabetic pathogenesis (14). Recently, Kowluru et al. (15) demonstrated that the hyperglycemic milieu induces hypomethylation of the retinal MMP-9 promoter, thereby regulating its transcription. Our previous studies have shown that advanced glycation end products (AGEs), a complex associated with diabetic complications (16–18), and tumor necrosis factor α induce MMP-9 promotor demethylation (19, 20). However, the molecular mechanisms underlying active DNA demethylation of the MMP-9 promoter in diabetic skin are complex and have not yet been clarified. Growth arrest and DNA damage protein 45a (GADD45a) is a member of a small family of stress-response genes and serves as a nexus between DNA repair and epigenetic gene regulation by mediating demethylation (21–23). GADD45a has been identified as a diabetes-associated gene that may link diabetic cardiomyopathy and baroreflex dysfunction, and 11 direct interacting proteins have been found to be highly associated with DM (24). However, whether GADD45a participates in MMP-9 promoter demethylation remains an important question. Here, we used skin tissues from patients with diabetic foot ulcers; the streptozotocin (STZ)-induced diabetic rat model, which is one of the most extensively characterized models of diabetic skin pathology; and diabetic human keratinocyte (HaCaT) cells to explore the expression levels of GADD45a and MMP-9, thereby providing a therapeutic target for diabetic foot ulcers. We found that GADD45a was significantly upregulated under diabetic conditions in vivo and in vitro, and its expression level was positively correlated with MMP-9 expression. Next, we showed that GADD45a targeted the promoter of MMP-9 and recruited thymine-DNA glycosylase (TDG) to induce MMP-9 promotor demethylation. Thus, our data corroborate a close link between the GADD45a- and 10–11 translocation protein (TET)-TDG–mediated DNA demethylation pathways, and our results provide a potential preventive strategy for chronic skins ulcers in DM. Materials and Methods Participants and tissue specimens Patient skin-tissue specimens were obtained from the Sun Yat-sen Memorial Hospital of Sun Yat-sen University (DM foot skin specimens, n = 5; non-DM skin specimens, n = 5). Samples were collected after written consent was obtained from the patients; the study received ethics approval and was conducted in accordance with the Declaration of Helsinki II. Diabetic foot ulcers from individuals with an active foot ulcer and nondiabetic foot specimens from healthy, nondiabetic subjects were obtained. The two groups were matched for age and sex. In addition, the diabetic groups were matched for the duration of diabetes and glycemic control, as defined by the levels of hemoglobin A1c. All of the diabetic foot patients had type 2 diabetes with neuropathy and peripheral artery disease, and one patient had an infection. The clinical information of the participants can be found in Supplemental Table 1. Animals and the establishment of an STZ-induced diabetic rat model All protocols for animal use and euthanasia were approved by the Sun Yat-sen University Institutional Animal Care and Use Committee. Twelve male Sprague-Dawley rats, each weighing ∼220 to 260 g, were randomized into the control (n = 6) or experimental (n = 6) group after adaptive feeding for 1 week. Animals were maintained on a 12-hour dark-light artificial cycle (lights on at 7:00 am) with food and water available ad libitum. Diabetes was induced by an intraperitoneal injection of 35 mg/kg STZ (Sigma-Aldrich, St. Louis, MO) in citrate buffer (pH 4.2) for 3 consecutive days. Animals injected with buffer alone were used as normal controls (n = 6). After a final injection of STZ at 72 hours, rats with a fasting blood glucose level of 16.7 mmol/L or higher, as determined from the tail vein blood using a OneTouch II® glucometer (LifeScan, Rochester, NY), were used as established DM models. STZ-treated rats were given daily injections of a small dose of protamine zinc insulin (∼0.5 to 3.0 U/d), as required to reduce mortality by maintaining blood glucose levels and avoiding ketosis. Rats were housed for 8 additional weeks before they were euthanized, and then, tissue was harvested. Skin specimens were fixed in 4% paraformaldehyde overnight at 4°C and then embedded in paraffin for morphometric studies. Paraffin-embedded tissues (4-μm sections) were stained for subsequent immunohistochemical examination. Cell culture The immortalized HaCaT line was cultured in Dulbecco’s modified Eagle’s medium, supplemented with 10% fetal calf serum and 1% antibiotics (100 U/ml penicillin and 100 mg/ml streptomycin) at 37°C in 5% CO2. Cells were grown to ∼30% to 40% confluence and treated with 100 µg/ml glycolaldehyde-modified AGE-bovine serum albumin (BSA; AGEs; Calbiochem, La Jolla, CA) or BSA (Calbiochem) at the same concentrations for ∼24 to 72 hours. DNA and RNA were isolated from cells or tissues using TRIzol (Life Technologies, Carlsbad, CA). Immunohistochemistry and data analysis, immunofluorescence, plasmid construct, and transfection For further details, please see Supplemental Materials and Methods. Quantitative real-time polymerase chain reaction analysis For quantitative real-time polymerase chain reaction (qPCR), total RNA was extracted, and 1 µg total RNA was converted into cDNA using a PrimeScriptTM RT Kit (TaKaRa, Otsu, Japan). Amplification and analysis were performed with a Roche LightCycler96 Real-Time PCR System (Roche, Basel, Switzerland). Values were normalized to the housekeeping gene actin B (ACTB). Alternatively, gene-specific primers were used (Supplemental Table 2). Cell proliferation assays and cell cycle and apoptosis analyses Cell proliferation was measured using the Cell Counting Kit-8 Assay (Dojindo Laboratories, Kumamoto, Japan). For cell analyses, cells were fixed and digested with RNase A, and propidium iodide was added before flow cytometry analysis. To examine cell apoptosis, cells were resuspended in binding buffer and stained with Annexin V and propidium iodide for 15 minutes in the dark before flow cytometry analysis. Details of the procedure are provided in Supplemental Materials and Methods. Wound-healing assay HaCaT cells were treated with AGEs or BSA and transfected with the indicated small interfering RNAs (siRNAs). After 48 hours, the attached cells were scratched with a 10-µl pipette tip, and images were captured at 0 hour using a microscope. The plates were returned to the 37°C incubator and cultured with 5% CO2 for 24 hours, at which time another set of images of the same wounds was captured. The wound widths were measured with ImageJ software (National Institutes of Health, Bethesda, MD) and were normalized and presented as a percentage of the wound measured at time zero. Western blotting analysis Total proteins extracted from each group of cells were resolved by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA). After blocking, the polyvinylidene difluoride membranes were washed three times for 10 minutes with tris(hydroxymethyl)aminomethane-buffered saline and Tween 20 at room temperature and incubated with primary antibodies (Table 1). Following extensive washing, membranes were incubated with secondary peroxidase-linked goat anti-rabbit immunoglobulin G (IgG) for 1 hour at room temperature. After washing three times for 10 minutes with tris(hydroxymethyl)aminomethane-buffered saline and Tween 20 at room temperature once more, the immunoreactivity was visualized by enhanced chemiluminescence, and membranes were exposed to the Imaging System G (Syngene, Cambridge, UK). Table 1. Antibodies Used Peptide/Protein Target  Antigen Sequence (if Known)  Name of Antibody  Manufacturer, Catalog No.  Species Raised in; Monoclonal or Polyclonal  Dilution Used  RRID  MMP-9  –  MMP-9 antibody  Abcam, ab76003  Rabbit; monoclonal  WB (1:1000)  AB_1310463  IHC (1:1000)  RNAP II  –  RNA polymerase II antibody  Abcam, ab817  Mouse; monoclonal  ChIP (1:500)  AB_306327  p300  –  KAT3B/p300 antibody  Abcam, ab14984  Rabbit; monoclonal  ChIP (1:500)  AB_306327  Thymine DNA glycosylase  –  TDG antibody  Abcam, ab106301  Rabbit; polyclonal  WB (1:1000)  AB_10973374  IP (1:50)  GADD45a  –  GADD45a antibody  Abcam, ab180768  Rabbit; polyclonal  WB (1:1000)  AB_2687480  IHC (1:200)  GADD45a  –  Gadd45a (H-165)  Santa Cruz Biotechnology, sc-797  Rabbit; polyclonal  IF (1:50)  AB_2232121  IP (1:50)  GADD45b  –  GADD45b antibody  Abcam, ab105060  Rabbit; polyclonal  WB (1:1000)  AB_10714129  GADD45g  –  GADD45g antibody human  Abcam, ab196774  Rabbit; monoclonal  WB (1:1000)  AB_2687481  FLAG  –  DYKDDDDK tag antibody  Cell Signaling Technology, 14793  Rabbit; monoclonal  IP (1:50)  AB_2572291  Peptide/Protein Target  Antigen Sequence (if Known)  Name of Antibody  Manufacturer, Catalog No.  Species Raised in; Monoclonal or Polyclonal  Dilution Used  RRID  MMP-9  –  MMP-9 antibody  Abcam, ab76003  Rabbit; monoclonal  WB (1:1000)  AB_1310463  IHC (1:1000)  RNAP II  –  RNA polymerase II antibody  Abcam, ab817  Mouse; monoclonal  ChIP (1:500)  AB_306327  p300  –  KAT3B/p300 antibody  Abcam, ab14984  Rabbit; monoclonal  ChIP (1:500)  AB_306327  Thymine DNA glycosylase  –  TDG antibody  Abcam, ab106301  Rabbit; polyclonal  WB (1:1000)  AB_10973374  IP (1:50)  GADD45a  –  GADD45a antibody  Abcam, ab180768  Rabbit; polyclonal  WB (1:1000)  AB_2687480  IHC (1:200)  GADD45a  –  Gadd45a (H-165)  Santa Cruz Biotechnology, sc-797  Rabbit; polyclonal  IF (1:50)  AB_2232121  IP (1:50)  GADD45b  –  GADD45b antibody  Abcam, ab105060  Rabbit; polyclonal  WB (1:1000)  AB_10714129  GADD45g  –  GADD45g antibody human  Abcam, ab196774  Rabbit; monoclonal  WB (1:1000)  AB_2687481  FLAG  –  DYKDDDDK tag antibody  Cell Signaling Technology, 14793  Rabbit; monoclonal  IP (1:50)  AB_2572291  Abbreviations: ChIP, chromatin immunoprecipitation; IF, immunofluorescence; IHC, immunohistochemistry; IP, immunoprecipitation; RNAP II, RNA polymerase II; RRID, Research Resource Identifier; WB, Western blot. View Large siRNAs, PCR primers, drugs, and antibodies siRNAs against GADD45a, GADD45b, GADD45g, TDG, TET2, and control siRNAs were purchased from GenePharma (Shanghai, China) and transfected with Lipofectamine 3000 (Invitrogen, Carlsbad, CA). All knockdown experiments showed at least a 70% reduction in target gene expression with a single siRNA. Primers were synthesized by GENEray biotechnology (Shanghai, China). The sequences of siRNAs used in this study are listed in Supplemental Table 3; we tested two siRNAs for every gene and screened for the most effective siRNA to use in the subsequent experiments. CRT 0044876 and 5-aza-2′-deoxycytidine were purchased from Sigma-Aldrich, diluted in dimethyl sulfoxide, and added to the cell culture media for a final concentration of 0.5 and 5 μM; the cells were then incubated for 48 or 72 hours. The following antibodies were used and listed in Table 1: MMP-9 (catalog no. ab76003; Abcam Cambridge, UK), RNA polymerase II (RNAP II; catalog no. ab81; Abcam), KAT3B/p300 (catalog no. ab14984; Abcam), TDG (catalog no. ab106301; Abcam), GADD45a (catalog no. ab180768; Abcam; catalog no. sc-797; Santa Cruz Biotechnology, Dallas, TX), GADD45b (catalog no. ab105060; Abcam), GADD45g (catalog no. ab196774; Abcam), and Flag (catalog no. 14793; Cell Signaling Technology, Danvers, MA). Quantitative DNA methylation analysis DNA methylation was analyzed using bisulfite-converted genomic DNA (EZ DNA Methylation Kit; Zymo Research, Irvine, CA) and the MassARRAY system (Sequenom, San Diego, CA), as previously reported. Details of the procedure are provided in Supplemental Materials and Methods. DNA methylation levels were quantified from mass spectra using Epityper Software v.1.2 (Sequenom) and were measured by the Beijing Genomics Institute. The MassARRAY primers for the MMP-9 locus are listed in Supplemental Table 4. Chromatin immunoprecipitation Cells were cross-linked with 1% paraformaldehyde at room temperature for 10 minutes, quenched with 0.5 M glycine, lysed in nuclei lysis buffer for 10 minutes, placed on ice, and sonicated seven times (15 seconds on and 59 seconds off) at 4°C to yield ∼150 to 250 base pair (bp) DNA fragments. The sheared chromatin was incubated with the indicated antibodies. Protein-DNA complexes were captured on protein A/G agarose, followed by washes in low-salt buffer, high-salt buffer, LiCI buffer, and Tris/EDTA buffer, according to the manufacturer’s recommendations (Millipore). After elution and reversal of the cross-link, DNA was extracted and amplified by qPCR. The amount of precipitated DNA was normalized to the amount of both the input DNA and the IgG-bound DNA. The primers were designed to amplify six regions upstream of the transcriptional start site (TSS; Supplemental Table 5). Coimmunoprecipitation HaCaT cells, in which Flag-tagged GADD45a was overexpressed, were lysed. The skin tissues of rats were cut and lysed overnight on ice with immunoprecipitation (IP) lysis buffer. Following preclearing of the lysate by centrifugation and incubation with control agarose resin for 1 hour at 4°C, lysates were incubated on a rotating wheel overnight at 4°C with specific antibodies or IgG as a control. Assay was performed using a coimmunoprecipitation (Co-IP) kit (Pierce, Rockford, IL), according to the manufacturer’s instructions. Statistical analysis Each experiment was performed at least three times. Data were presented as the means ± standard deviation (SD) where applicable and analyzed using SPSS software (IBM, New York, NY). Comparison of two groups was performed using an unpaired t test. Statistical significance was assessed with ANOVA, followed by a least significant difference test for multiple comparisons. P < 0.05 was considered statistically significant. Results GADD45a and MMP-9 are overexpressed in diabetic skin and cells exposed to AGEs To investigate the role of the GADD45a protein in diabetic skin, GADD45a expression and distribution in human and rat skin tissues were analyzed. We used immunohistochemistry to examine GADD45a expression in diabetic skin biopsy samples and normal skin tissue samples. Abundant GADD45a expression was observed in the epidermal skin tissues of DM patients and was localized in the dermal and subcutaneous connective tissues [Fig. 1(a)]. In contrast, GADD45a expression levels in diabetic skin were significantly higher than those in non-DM skin [P = 0.001, insulin receptor substrate (IRS):DM group 10.0 ± 2.8 vs non-DM group 0.6 ± 0.54; Fig. 1(c)]. In addition, we examined six diabetic rat skin samples and six nondiabetic rat skin samples. During the establishment of the diabetic rat model, no experimental animals died of hyperglycemia or other complications. We observed a substantial increase in GADD45a expression in the skin tissue of DM rats [P < 0.001, IRS:DM group 9.6 ± 2.2 vs non-DM group 0.6 ± 0.5; Fig. 1(b) and 1(c)]. We then examined MMP-9 expression and found that MMP-9 expression levels were significantly higher in diabetic skin tissues than in non-DM skin tissues [Fig. 1(d)–1(f)]. MMP-9 expression was positively correlated with GADD45a expression in human skin tissue (Supplemental Fig. 1). Figure 1. View largeDownload slide GADD45a is expressed in diabetic skin and cells exposed to AGEs. (a–c) Immunohistochemical staining of GADD45a in diabetic and nondiabetic (a) human or (b) rat skin tissues. (c) Quantification analysis is presented. Original magnification, ×200 (×400 in the next row).Data are presented as the means ± SD (n ≥ 5 per group). **P < 0.01 vs the normal group. (d–f) Immunohistochemical staining of MMP-9 in diabetic and nondiabetic (d) human or (e) rat skin tissues. (f) Quantification analysis is presented. Original magnification, ×200 (×400 in next row). Data are presented as the means ± SD of three independent experiments (n ≥ 5). **P < 0.01 vs normal group. (g–i) Expression of GADD45a, GADD45b, GADD45g, and MMP-9 in HaCaT cells treated with AGEs for 0, 24, 48, and 72 hours, as assayed by (g and h) qRT-PCR and (i) Western blotting. (j) Microscopy images show the expression of GADD45a (red) with nuclear staining by DAPI (blue) after treatment with AGEs or BSA for 72 hours. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs control group. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole; DM, diabetic skin tissue; IRS, insulin receptor substrate; NDM, normal, nondiabetic skin tissue; qRT-PCR, quantitative reverse transcription PCR. Figure 1. View largeDownload slide GADD45a is expressed in diabetic skin and cells exposed to AGEs. (a–c) Immunohistochemical staining of GADD45a in diabetic and nondiabetic (a) human or (b) rat skin tissues. (c) Quantification analysis is presented. Original magnification, ×200 (×400 in the next row).Data are presented as the means ± SD (n ≥ 5 per group). **P < 0.01 vs the normal group. (d–f) Immunohistochemical staining of MMP-9 in diabetic and nondiabetic (d) human or (e) rat skin tissues. (f) Quantification analysis is presented. Original magnification, ×200 (×400 in next row). Data are presented as the means ± SD of three independent experiments (n ≥ 5). **P < 0.01 vs normal group. (g–i) Expression of GADD45a, GADD45b, GADD45g, and MMP-9 in HaCaT cells treated with AGEs for 0, 24, 48, and 72 hours, as assayed by (g and h) qRT-PCR and (i) Western blotting. (j) Microscopy images show the expression of GADD45a (red) with nuclear staining by DAPI (blue) after treatment with AGEs or BSA for 72 hours. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs control group. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole; DM, diabetic skin tissue; IRS, insulin receptor substrate; NDM, normal, nondiabetic skin tissue; qRT-PCR, quantitative reverse transcription PCR. The superficial layer of the epidermis (∼60 μm) interfaces more directly with the outside world than the deeper dermal layer, and the epidermis is composed of ∼95% keratinocytes. We first confirmed that GADD45a and MMP-9 mRNA and protein were expressed in HaCaT cells treated with 100 µg/ml AGEs for 24, 48, and 72 hours. The mRNA levels of GADD45a were increased after 48 and 72 hours, and MMP-9 mRNA levels were increased after 72 hours [Fig. 1(g) and 1(h)]. GADD45a and MMP-9 protein expression levels were both increased after 24 hours, but they were significantly increased after treatment with AGEs for 72 hours [Fig. 1(i)]. In contrast, the mRNA and protein expression levels of GADD45b/g were not significantly altered after treatment with AGEs [Fig. 1(g) and 1(i)]. Therefore, we selected the time point of 72 hours for subsequent experiments. Fluorescence microscopy also showed that AGEs treatment increased the expression level of GADD45a, which was predominantly located in the nucleus [Fig. 1(j)]. GADD45a affects the biological function of HaCaT cells Next, we examined the impact of GADD45a on the biological function of HaCaT cells. We observed that the invitro proliferation and apoptosis rates of HaCaT cells transfected with siRNA targeting GADD45a were similar to those of cells transfected with a scrambled siRNA sequence (siRNA control) after treatment with AGEs [Fig. 2(a)–2(c)]; the interference efficiency is shown in Fig. 3(a) and 3(b). Next, we analyzed the proportion of cells in various stages of the cell cycle using flow cytometry. Likewise, the results were not different among the groups [Fig. 2(d)–2(f)]. Compared with control and mock cells, the wound-healing capacity of HaCaT–small interfering GADD45a (siGADD45a) cells cultured in normal medium was not different. Treatment with AGEs reduced the wound-healing capacity of HaCaT cells; however, siGADD45a transfection improved wound healing in cells treated with AGEs, even compared with controls [Fig. 2(g) and 2(h)]. These experiments showed that the silencing of the expression of GADD45a can effectively promote the migration ability of diabetic HaCaT cells. Figure 2. View largeDownload slide GADD45a affects the biological function of HaCaT cells. (a) Proliferation of BSA- or AGEs-treated cells following the downregulation of GADD45a or transfection of scrambled control siRNA (Mock), examined with cell counting kit-8. (b–f) HaCaT cells in a logarithmic growth phase (∼30% to 40% confluence) were treated with mock or siGADD45a and then exposed to 100 µg/mL AGEs or BSA. After 72 hours, the cells were harvested and then subjected to apoptosis and cell-cycle analysis. (b) Representative images of the apoptosis rate of the cells. (c) Quantification of apoptotic cells. (d) Representative images of the cell-cycle distribution profiles of the cells. (e and f) Quantification of the cell-cycle distribution. (g) Representative images of wound-healing assays. HaCaT cells were transiently transfected with siGADD45a or mock and treated with BSA or AGEs. After 72 hours, wound-healing assays were performed, as described in Materials and Methods. (h) Statistical analyses of wound-healing assays. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs control group. Ctrl, control; FITC-A, fluorescein isothiocyanate-area; N.S., not significant. OD, optical density; PE-A, phycoerythrin-area. Figure 2. View largeDownload slide GADD45a affects the biological function of HaCaT cells. (a) Proliferation of BSA- or AGEs-treated cells following the downregulation of GADD45a or transfection of scrambled control siRNA (Mock), examined with cell counting kit-8. (b–f) HaCaT cells in a logarithmic growth phase (∼30% to 40% confluence) were treated with mock or siGADD45a and then exposed to 100 µg/mL AGEs or BSA. After 72 hours, the cells were harvested and then subjected to apoptosis and cell-cycle analysis. (b) Representative images of the apoptosis rate of the cells. (c) Quantification of apoptotic cells. (d) Representative images of the cell-cycle distribution profiles of the cells. (e and f) Quantification of the cell-cycle distribution. (g) Representative images of wound-healing assays. HaCaT cells were transiently transfected with siGADD45a or mock and treated with BSA or AGEs. After 72 hours, wound-healing assays were performed, as described in Materials and Methods. (h) Statistical analyses of wound-healing assays. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs control group. Ctrl, control; FITC-A, fluorescein isothiocyanate-area; N.S., not significant. OD, optical density; PE-A, phycoerythrin-area. Figure 3. View largeDownload slide GADD45a is required for MMP-9 expression. (a–c) Relative levels of the GADD45 family genes and MMP-9 mRNA in HaCaT cells following siRNA-mediated knockdown of GADD45a, GADD45b, or GADD45g. Scrambled siRNA (Mock) was used as a control. RNA levels were normalized to ACTB mRNA. (d–f) Relative protein expression was analyzed by Western blotting. (g) qRT-PCR was used to analyze GADD45a and MMP-9 levels in HaCaT cells after inhibition of the GADD45a gene in the process of AGEs. (h) Relative levels of GADD45a/MMP-9 protein in HaCaT cells following siRNA-mediated knockdown of GADD45a and treatment with 100 µg/ml AGEs (48 hours). (i and j) Relative levels of (i) GADD45a/MMP-9 mRNA and (j) protein in HaCaT cells, 48 hours after infection with GADD45a-adenovirus (Ad-GADD45a). Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the control group. Ctrl, control; qRT-PCR, quantitative reverse transcription PCR. Figure 3. View largeDownload slide GADD45a is required for MMP-9 expression. (a–c) Relative levels of the GADD45 family genes and MMP-9 mRNA in HaCaT cells following siRNA-mediated knockdown of GADD45a, GADD45b, or GADD45g. Scrambled siRNA (Mock) was used as a control. RNA levels were normalized to ACTB mRNA. (d–f) Relative protein expression was analyzed by Western blotting. (g) qRT-PCR was used to analyze GADD45a and MMP-9 levels in HaCaT cells after inhibition of the GADD45a gene in the process of AGEs. (h) Relative levels of GADD45a/MMP-9 protein in HaCaT cells following siRNA-mediated knockdown of GADD45a and treatment with 100 µg/ml AGEs (48 hours). (i and j) Relative levels of (i) GADD45a/MMP-9 mRNA and (j) protein in HaCaT cells, 48 hours after infection with GADD45a-adenovirus (Ad-GADD45a). Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the control group. Ctrl, control; qRT-PCR, quantitative reverse transcription PCR. Knockdown of GADD45a promotes diabetic HaCaT cell migration via the transcriptional downregulation of MMP-9 Our previous experiments have shown that increased MMP-9 expression contributes to delayed wound healing in late-stage diabetic foot ulcers (8). As GADD45a is a powerful nuclear protein and is overexpressed in diabetic skin tissues, we proposed that it might be involved in regulating MMP-9 expression. To investigate further the association between GADD45a levels and MMP-9 expression, we transiently knocked down GADD45a expression in HaCaT cells by RNA interference without AGE treatment. Real-time PCR and Western blot results showed that GADD45a expression was significantly downregulated, as shown in Fig. 3(a) and 3(b). Downregulation of GADD45a resulted in decreased MMP-9 mRNA and protein expression compared with controls; however, downregulation of GADD45b/g did not affect the transcription or post-transcriptional levels of MMP-9 [Fig. 3(c)–3(f)]. Following exposure to AGEs, we found that the mRNA and protein levels of MMP-9 were also inhibited by siRNA GADD45a [Fig. 3(g) and 3(h)]. Then, we generated the recombinant adenovirus (Ad)-GADD45a and transfected HaCaT cells for 48 hours. The mRNA and protein levels of MMP-9 following the upregulation of GADD45a were significantly higher than those in the control group [Fig. 3(i) and 3(j)]. Based on the previous results, we concluded that GADD45a regulates MMP-9 expression. GADD45a activates MMP-9 transcription by inducing the demethylation of the MMP-9 promoter As our previous study showed that wound healing in diabetic skin was related to MMP-9 DNA demethylation (19), we reasoned that GADD45a might be involved in the activation of the transcription of MMP-9 via DNA demethylation. We detected MMP-9 promoter methylation in diabetic rat skin tissues and AGE-treated HaCaT cells. There are 12 CpG sites in the MMP-9 promoter of the rat [Fig. 4(a)]. Because the method of MassARRAY could not detect the −691-bp site, we evaluated the methylation levels of 11 CpG sites. In diabetic rat skin tissues, MMP-9 promoter methylation was significantly decreased at three CpG sites (−329, −97, and −56 bp) compared with nondiabetic skin tissues [Fig. 4(b)]. After treatment of HaCaT cells with AGEs for different durations, we evaluated 10 CpG sites in the MMP-9 gene promoter and observed DNA demethylation in several CpG sites; in particular, after 72 hours of treatment, there was a substantial decrease of methylation (Supplemental Fig. 2). Therefore, the time point of 72 hours may be important for substantial demethylation of the MMP-9 promoter by the demethylation protein complex. Among the 10 CpG sites of MMP-9 in HaCaT cells treated with AGEs for 72 hours, DNA methylation was significantly decreased at three CpG sites (−233, −223, and −36 bp) compared with those in control cells [Fig. 4(c) and 4(d)]. Figure 4. View largeDownload slide GADD45a activates MMP-9 transcription by promoter demethylation. (a) The position of CpG sites in the promoter of rat MMP-9 is indicated by “lollipops.” Arrows refer to the MMP-9 TSS. The location of two pairs of BSP primers is represented. (b) DNA methylation levels around the TSS of MMP-9 in diabetic or nondiabetic rat skin tissues were measured by MassARRAY. (c) The position of CpG sites in the promoter of human MMP-9 is indicated by lollipops. (d) DNA methylation levels around the TSS of MMP-9 in HaCaT cells treated with AGEs or BSA were measured by MassARRAY. (e and f) BSA- or AGEs-exposed HaCaT cells were transfected with siRNA against GADD45a or a scrambled control RNA (Mock). DNA methylation was measured by MassARRAY. (g) Methylation analysis of the MMP-9 promoter of HaCaT cells after treatment with Ad-GADD45a or control adenoviruses (Vector). Methylation levels of 10 CpG sites were measured by MassARRAY. (h) Location of DNA fragments (C1–C6) are indicated in the scheme. (i and j) ChIP assay monitoring the association of RNAP II and p300 at different regions of the MMP-9 promoter in AGE-treated HaCaT cells transfected with siGADD45a or control RNA (Mock). Values represent the enrichment of the bound protein fraction relative to IgG. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and #P < 0.01 vs the corresponding control group. BSP, bisulfite sequencing PCR; ChIP, chromatin immunoprecipitation; N.S., not significant. Figure 4. View largeDownload slide GADD45a activates MMP-9 transcription by promoter demethylation. (a) The position of CpG sites in the promoter of rat MMP-9 is indicated by “lollipops.” Arrows refer to the MMP-9 TSS. The location of two pairs of BSP primers is represented. (b) DNA methylation levels around the TSS of MMP-9 in diabetic or nondiabetic rat skin tissues were measured by MassARRAY. (c) The position of CpG sites in the promoter of human MMP-9 is indicated by lollipops. (d) DNA methylation levels around the TSS of MMP-9 in HaCaT cells treated with AGEs or BSA were measured by MassARRAY. (e and f) BSA- or AGEs-exposed HaCaT cells were transfected with siRNA against GADD45a or a scrambled control RNA (Mock). DNA methylation was measured by MassARRAY. (g) Methylation analysis of the MMP-9 promoter of HaCaT cells after treatment with Ad-GADD45a or control adenoviruses (Vector). Methylation levels of 10 CpG sites were measured by MassARRAY. (h) Location of DNA fragments (C1–C6) are indicated in the scheme. (i and j) ChIP assay monitoring the association of RNAP II and p300 at different regions of the MMP-9 promoter in AGE-treated HaCaT cells transfected with siGADD45a or control RNA (Mock). Values represent the enrichment of the bound protein fraction relative to IgG. Data are presented as the means ± SD of three independent experiments (n = 3). *P < 0.05 and #P < 0.01 vs the corresponding control group. BSP, bisulfite sequencing PCR; ChIP, chromatin immunoprecipitation; N.S., not significant. Knockdown of GADD45a led to increase methylation at the −36-bp CpG site with BSA exposure, thus supporting that GADD45a is required for the maintenance of MMP-9 promoter hypomethylation [Fig. 4(e)]. However, with AGEs treatment, siGADD45a transfection led to a 39.3% increase in DNA methylation around the TSS of MMP-9 [−233, −223, and −36 bp; Fig. 4(f)]. Then, we transfected HaCaT cells with Ad-GADD45a-Flag and detected the DNA methylation level of the MMP-9 promoter after 48 hours. We found that methylation was significantly reduced at three CpG sites [−712, −223, and −36 bp; Fig. 4(g)]. Indeed, knockdown of GADD45a in AGEs-treated HaCaT cells (expressing both GADD45a and MMP-9) led to decreased occupancy of both RNAP II and histone acetyltransferase p300 at the MMP-9 promoter [Fig. 4(h)–4(j)]. Therefore, knockdown of GADD45a inhibited the AGEs-induced demethylation of the MMP-9 promoter and inhibited MMP-9 expression in HaCaT cells. GADD45a mediates MMP-9 promotor demethylation by TDG-dependent base excision repair GADD45a removes methylated cytosine residues through nucleotide excision repair (NER) or base excision repair (BER) machinery (25–27). To detect the mechanism of MMP-9 demethylation, we examined the mRNA expression levels of the NER pathway [xeroderma pigmentosum (XP)A, XPC, XPF, and XPG], as well as the BER pathway (TDG), after AGEs exposure in HaCaT cells at different time points. The results showed that TDG expression was significantly higher in AGEs-treated HaCaT cells than in control cells at 72 hours, whereas the NER pathway mRNA expression levels did not change [Fig. 5(a)]. To clarify the role of TDG, we monitored MMP-9 expression after knockdown of TDG and showed that it was markedly reduced [Fig. 5(b) and 5(c)]; however, the MMP-9 mRNA expression levels were not significantly different after knockdown of NER genes (Supplemental Fig. 3). To show that GADD45a recruits TDG, an essential protein of the BER pathway, to mediate active demethylation, we treated cells with CRT0044876 (a BER inhibitor) and monitored MMP-9 expression. The BER inhibitor prevented the GADD45a-dependent expression of MMP-9, but GADD45a expression was not affected [Fig. 5(d)]. In addition, the GADD45a-mediated expression and DNA demethylation of MMP-9 were inhibited upon knockdown of TDG [Fig. 5(e) and 5(f)], further indicating that TDG cooperates with GADD45a to activate MMP-9 expression. To demonstrate that GADD45a and TDG regulate MMP-9 expression by adjusting the demethylation level of the MMP-9 promoter, we treated cells with 5-aza-2′-deoxycytidine, which causes DNA demethylation, and detected the MMP-9 expression levels. 5-Aza-2′-deoxycytidine promoted MMP-9 expression and prevented the siGADD45a- or siTDG-induced decrease in MMP-9 expression [Fig. 5(g) and 5(h)]. Figure 5. View largeDownload slide GADD45a-mediated MMP-9 promotor demethylation by TDG-dependent BER. (a) Relative mRNA levels of NER genes (XPA, XPC, XPF, and XPG) and the BER gene TDG in HaCaT cells treated with AGEs or BSA. RNA levels were normalized to ACTB mRNA. (b and c) HaCaT cells were transfected with siRNA against TDG or a scrambled control siRNA (Mock), and (b) qRT-PCR and (c) Western blots were used to analyze the level of MMP-9 and TDG expression. (d) Relative levels of MMP-9 and GADD45a mRNA in HaCaT cells treated with AGEs for 24 hours, followed by treatment with the BER inhibitor CRT0044876 (CRT) for 48 hours. DMSO was used as a solvent control, and GADD45a expression was considered a DNA damage-response control. RNA levels were normalized to ACTB mRNAs. (e) Relative levels of MMP-9 mRNA in HaCaT cells infected with Ad-GADD45a or control adenoviruses (Vector). Cells were cotransfected with siRNA against TDG or a control RNA (Mock). MMP-9 mRNA was measured by qPCR. RNA levels were normalized to ACTB mRNAs. (f) Methylation analysis of the MMP-9 promoter in HaCaT cells after treatment with siRNA against TDG (siTDG) or a scrambled control (Mock). Methylation of CpGs 1–10 was measured by MassARRAY. (g and h) Expression levels of GADD45a, TDG, and MMP-9 in HaCaT cells treated with 5-aza-2′-deoxycytidine (5aza) and transfected with siGADD45a or siTDG for 72 hours, as assayed by (g) qRT-PCR and (h) Western blotting. Data are presented as the mean ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs the corresponding control group. Ctrl, control; DMSO, dimethyl sulfoxide; NC, negative control; qRT-PCR, quantitative reverse transcription PCR. Figure 5. View largeDownload slide GADD45a-mediated MMP-9 promotor demethylation by TDG-dependent BER. (a) Relative mRNA levels of NER genes (XPA, XPC, XPF, and XPG) and the BER gene TDG in HaCaT cells treated with AGEs or BSA. RNA levels were normalized to ACTB mRNA. (b and c) HaCaT cells were transfected with siRNA against TDG or a scrambled control siRNA (Mock), and (b) qRT-PCR and (c) Western blots were used to analyze the level of MMP-9 and TDG expression. (d) Relative levels of MMP-9 and GADD45a mRNA in HaCaT cells treated with AGEs for 24 hours, followed by treatment with the BER inhibitor CRT0044876 (CRT) for 48 hours. DMSO was used as a solvent control, and GADD45a expression was considered a DNA damage-response control. RNA levels were normalized to ACTB mRNAs. (e) Relative levels of MMP-9 mRNA in HaCaT cells infected with Ad-GADD45a or control adenoviruses (Vector). Cells were cotransfected with siRNA against TDG or a control RNA (Mock). MMP-9 mRNA was measured by qPCR. RNA levels were normalized to ACTB mRNAs. (f) Methylation analysis of the MMP-9 promoter in HaCaT cells after treatment with siRNA against TDG (siTDG) or a scrambled control (Mock). Methylation of CpGs 1–10 was measured by MassARRAY. (g and h) Expression levels of GADD45a, TDG, and MMP-9 in HaCaT cells treated with 5-aza-2′-deoxycytidine (5aza) and transfected with siGADD45a or siTDG for 72 hours, as assayed by (g) qRT-PCR and (h) Western blotting. Data are presented as the mean ± SD of three independent experiments (n = 3). *P < 0.05 and **P < 0.01 vs the corresponding control group. Ctrl, control; DMSO, dimethyl sulfoxide; NC, negative control; qRT-PCR, quantitative reverse transcription PCR. GADD45a binds to TDG to promote MMP-9 promoter demethylation To monitor the interaction between GADD45a and TDG, we performed Co-IP experiments with HaCaT cells overexpressing Flag-GADD45a and diabetic rat skin tissues and confirmed the association of TDG with GADD45a [Fig. 6(a) and 6(b)]. TDG is involved in a major active DNA demethylation pathway by excising the oxidation products, participating in TET-mediated oxidation of methylcytosines (28, 29). Our previous studies have shown that TET2 binds to the promoter of MMP-9, which expands its demethylation under diabetic conditions (19). To examine further whether GADD45a cooperates with the TET2–TDG demethylation pathway to activate MMP-9 expression, we examined the changes in 5-methylcytosine (5mc) and 5-hydroxymethylcytosine (5hmc) and found no substantial difference between the siGADD45a and mock groups [Fig. 6(c) and 6(d)] or between the Ad-GADD45a and vector groups treated with BSA or AGEs [Fig. 6(e) and 6(f)]; however, the levels of 5mc in the AGEs treatment groups were lower than those in the BSA treatment groups [Fig. 6(c)], which is consistent with our previous report (19). We also determined that there was no association between TET2 and GADD45a (data not shown), confirming that GADD45a did not directly affect the role of TET2 or the overall methylation level in diabetic HaCaT cells. Moreover, the chromatin immunoprecipitation (ChIP) experiment revealed that GADD45a has a substantial enrichment at the MMP-9 promoter in AGEs-treated HaCaT cells but not IgG or TDG [Fig. 6(g) and 6(h)], thus supporting that GADD45a binds to the MMP-9 promoter and recruits TDG to mediate demethylation. Figure 6. View largeDownload slide GADD45a binds to TDG to promote MMP-9 promoter demethylation. (a) FLAG-tagged GADD45a was immunoprecipitated from lysates of transfected HaCaT cells, and coprecipitated TDG was evaluated on immunoblots with antibodies against TDG and FLAG. The input sample was 10% of the lysate that was used for IP. The positions of molecular weight marker proteins are indicated. (b) GADD45a was immunoprecipitated from lysates of diabetic rat skin tissues. (c and d) 5mc and 5hmc levels, which reflect global gene methylation, were measured by the quantification of methylcytosine and hydroxymethylcytosine, respectively. 5mc and 5hmc levels in HaCaT cells, 48 hours after transfection with siGADD45a or control RNA (Mock) and treatment with BSA or AGEs. (e and f) 5mc and 5hmc levels in the MMP-9 promoter were measured using HaCaT cells by quantifying methylcytosine and hydroxymethylcytosine levels after treatment with Ad-GADD45a or control adenoviruses (Vector). (g and h) ChIP assays showing the enrichment of GADD45a and TDG at the MMP-9 promoter in HaCaT cells treated with BSA or AGEs. Coprecipitated DNA was analyzed by qPCR. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the corresponding control group. Figure 6. View largeDownload slide GADD45a binds to TDG to promote MMP-9 promoter demethylation. (a) FLAG-tagged GADD45a was immunoprecipitated from lysates of transfected HaCaT cells, and coprecipitated TDG was evaluated on immunoblots with antibodies against TDG and FLAG. The input sample was 10% of the lysate that was used for IP. The positions of molecular weight marker proteins are indicated. (b) GADD45a was immunoprecipitated from lysates of diabetic rat skin tissues. (c and d) 5mc and 5hmc levels, which reflect global gene methylation, were measured by the quantification of methylcytosine and hydroxymethylcytosine, respectively. 5mc and 5hmc levels in HaCaT cells, 48 hours after transfection with siGADD45a or control RNA (Mock) and treatment with BSA or AGEs. (e and f) 5mc and 5hmc levels in the MMP-9 promoter were measured using HaCaT cells by quantifying methylcytosine and hydroxymethylcytosine levels after treatment with Ad-GADD45a or control adenoviruses (Vector). (g and h) ChIP assays showing the enrichment of GADD45a and TDG at the MMP-9 promoter in HaCaT cells treated with BSA or AGEs. Coprecipitated DNA was analyzed by qPCR. Data are presented as the means ± SD of three independent experiments (n = 3). **P < 0.01 vs the corresponding control group. Discussion Our findings identified GADD45a as an important signature protein in diabetic skin ulcers. GADD45a was shown to be markedly upregulated in the skin tissue of patients with diabetic foot ulcers and in the skin of diabetic Sprague- Dawley rats; moreover, the upregulation of GADD45a was associated with MMP-9 expression. More importantly, our study showed that inhibiting GADD45a increased the migration ability of diabetic skin cells by downregulating the expression of MMP-9 in vitro. The suppression of TDG levels to prevent GADD45a-mediated MMP-9 demethylation may be a mechanism whereby overexpressed MMP-9 degrades the ECM in diabetic skin ulcers. Our study also demonstrated that GADD45a protein interacts with TDG in vitro and in vivo. Based on these findings, we propose the model illustrated in Fig. 7. The targeting of GADD45a, a demethylation-promoting protein, may be a good approach to preventing the amputation of a diabetic foot. Figure 7. View largeDownload slide A model of the role of GADD45a in the regulation of MMP-9 demethylation. (Lower) In normal cells, the promoter of MMP-9 is methylated, and MMP-9 expression levels are low. (Upper) In diabetic cells, GADD45a is overexpressed and binds to the MMP-9 promoter. GADD45a recruits TDG to induce promoter demethylation via BER. Demethylated MMP-9 harbors RNAP II (Pol II) and active chromatin marker p300, thus promoting MMP-9 expression. Figure 7. View largeDownload slide A model of the role of GADD45a in the regulation of MMP-9 demethylation. (Lower) In normal cells, the promoter of MMP-9 is methylated, and MMP-9 expression levels are low. (Upper) In diabetic cells, GADD45a is overexpressed and binds to the MMP-9 promoter. GADD45a recruits TDG to induce promoter demethylation via BER. Demethylated MMP-9 harbors RNAP II (Pol II) and active chromatin marker p300, thus promoting MMP-9 expression. Evidence has revealed that hyperglycemia could induce genomic hypomethylation and aberrant gene expression within the liver of type 1 DM rats (30), the skin of type 1 DM zebrafish (31), and human primary aortic endothelial cells (32, 33). However, the expression level and effect of the GADD45a protein remain poorly understood in diabetic skin tissues. Moreover, it is unknown whether GADD45a participates in regulating MMP-9 promoter methylation in diabetic skin cells. Herein, our data indicate that GADD45a is capable of upregulating MMP-9 expression in diabetic skin cells, but its downregulation leads to aberrant MMP-9 activation without AGEs stimulation, suggesting that GADD45a is an essential factor for MMP-9 expression in HaCaT cells, which warrants investigation in a large number of clinical samples. In our study, GADD45a/b/g expression was different in diabetic skin tissues and cells. These differences revealed that GADD45 family proteins were tissue specific in human tissues and diseases. To explain the roles of GADD45a in MMP-9 expression and promoter demethylation under the diabetic conditions, GADD45a interference RNA or Ad-GADD45a was constructed and transfected into HaCaT cells. We demonstrated that downregulated GADD45a could reverse the pathological effects of MMP-9 on the degradation of ECM proteins by promoting the migration of diabetic cells. Moreover, GADD45a interference RNA did not affect cell proliferation or apoptosis in our study, indicating that the extent of inhibition of GADD45a is appropriate to study the change of the demethylation level of the MMP-9 promoter. The expression of GADD45a leads to an important change in the CpG sites of the MMP-9 promoter, which undergoes up- or downregulation of methylation after the corresponding changes of GADD45a expression. Therefore, GADD45a expression should be inhibited in an appropriate range, which is a more effective reduction of MMP-9 demethylation and more beneficial to the treatment of a diabetic foot ulcer. Regarding the demethylation mechanism of GADD45a, in our study, we indicated that the GADD45a protein bound and interacted with the −668- to −538-bp segments of the MMP-9 promoter. The segments are rich with transcription factor binding sites, such as nuclear factor κB, Sp1 transcription factor, and AP-1 transcription factor; therefore, they may involve the process of GADD45a-mediated demethylation, thereby regulating MMP-9 expression. Actually, the role of DNA demethylation is to affect transcription factor binding (34, 35). Moreover, our results showed that MMP-9 and GADD45a protein levels increased after 24 hours by treatment with AGEs. This phenomenon suggests that GADD45a may be involved in the demethylation of MMP-9 in the early stage, and there may exist another mechanism that adjusts the expression of MMP-9, such as regulation by noncoding RNA, histone modifications, and transcription factor activity. Therefore, the function of this segment in the demethylation of the MMP-9 promoter and regulation of MMP-9 gene expression should be investigated in further studies. Multiple mechanisms underlying active demethylation have been proposed, including C–C bond cleavage on the exocyclic methyl group and BER and NER pathways of replacement of the methylated cytosine (36). Our study showed that GADD45a interacts with TDG and helps mediate the demethylation of the MMP-9 promoter using BER and TDG. TDG can replace 5mc with unmethylated cytosine by excising the interactive oxidation products of DNA methylation, including 5-formylcytosine and 5-carboxylcytosine (28, 37, 38). We observed that TDG was not enriched in the MMP-9 promoter. Therefore, GADD45a cannot stimulate TDG to undergo specific substrate binding; rather, GADD45a may act as a scaffold protein that directs TDG to the promoter of target genes and excises the oxidation products (5-formylcytosine and 5-carboxylcytosine). The interaction of GADD45a and TDG induced the demethylation of gene promoters, which is consistent with a recently proposed model (39, 40). We analyzed the methylation levels in the MMP-9 promoter following AGEs treatment of different durations. The results showed that methylation was reduced, not only after 72 hours but also at 48 and 96 hours (Supplemental Fig. 3). We found that TDG levels significantly increased only after 72 hours and that its levels were decreased after 96 hours; therefore, the location of demethylation proteins seems to be more important than their expression levels in the process of demethylation. Further investigation will be required to reveal the molecular mechanisms by which GADD45 proteins regulate the role of TDG enzymes in specific loci in the chromatin contexts. In addition, GADD45a interacts with TET-mediated oxidative DNA demethylation by directly binding to each other (41); therefore, many auxiliary factors may be involved in modulating the activity of the TET-TDG system. Moreover, we found that knockdown of GADD45a had no effect on the levels of 5mc and 5hmc but did influence MMP-9 promoter hypermethylation and gene downregulation. This is consistent with GADD45-mediated demethylation being restricted to single-copy genes, and global methylation appears to be unaffected by GADD45a expression (42, 43); this phenomenon was also demonstrated in GADD45a knockout mice (44). Our previous studies demonstrated that TET2 contributes to the development of diabetic skin ulcers by promoting MMP-9 promoter demethylation (19, 20). Therefore, we favor a model in which GADD45a promotes active DNA demethylation by recruiting TDG to target loci and further promotes the TET-TDG demethylation pathway. In summary, we propose that GADD45a interacts with target loci in the MMP-9 promoter and promotes HaCaT cells migration by inhibiting MMP-9 gene expression. These findings reveal a role of GADD45a in diabetic skin ulcers and provide a therapeutic target for chronic refractory skin ulcers. The changes of DNA methylation occur frequently in diabetic tissues and cells so the GADD45a-mediated DNA demethylation may not be limited to the MMP-9 promoter but may also be relevant to the global epigenetic changes observed during the occurrence of diabetes and its complications. Abbreviations: 5hmc 5-hydroxymethylcytosine 5mc 5-methylcytosine ACTB actin B Ad adenovirus AGE advanced glycation end product BER base excision repair bp base pair BSA bovine serum albumin ChIP chromatin immunoprecipitation Co-IP coimmunoprecipitation DM diabetes mellitus ECM extracellular matrix GADD45a growth arrest and DNA damage-inducible 45a HaCaT human keratinocyte IgG immunoglobulin G IP immunoprecipitation IRS insulin receptor substrate MMP matrix metalloproteinase NER nucleotide excision repair, qPCR, quantitative polymerase chain reaction RNAP II RNA polymerase II SD standard deviation siGADD45a small interfering growth arrest and DNA damage-inducible 45a siRNA small interfering RNA STZ streptozotocin TDG thymine-DNA glycosylase TET 10–11 translocation protein TSS transcriptional start site XP xeroderma pigmentosum. Acknowledgments We are grateful to Kai Huang of the Department of Vascular Surgery and Gang Zeng of the Department of Orthopaedics at the Sun Yat-sen Memorial Hospital of Sun Yat-sen University for providing human skin tissues. Financial Support: This study was supported by National Natural Science Foundation of China Grants 81370910 (to M.R.) and 81471034 (to L.Y.); Natural Science Foundation of Guangdong Province Grant S2013010016443 (to M.R.); Science and Technology Planning Project of Guangdong Province, China Grant 2014A020212161 (to J.Z.); Grant [2013]163 from the Key Laboratory of Malignant Tumor Molecular Mechanism and Translational Medicine of Guangzhou Bureau of Science and Information Technology; Grant KLB09001 from the Key Laboratory of Malignant Tumor Gene Regulation and Target Therapy of Guangdong Higher Education Institutes; and a grant from the Guangdong Science and Technology Department (2015B050501004). Author Contributions: L.Z. and W.W. contributed to the conception and design of the study; acquisition, analysis, and interpretation of the data; and drafting and writing of the article. C.Y., T.Z., M.H., X.W., N.L., K.S., C.W., and J.Z. participated in the acquisition of data and critical revision of the manuscript, collected data, and revised the article. M.R. and L.Y. analyzed data, planned the experiments, performed the literature search, and revised the article. M.R. and L.Y. are the guarantors of this work and, as such, had full access to all of the data in the study; they take responsibility for the integrity of the data and the accuracy of the data analysis. All of the authors approved the final version to be published. Disclosure Summary: The authors have nothing to disclose. References 1. Shaw JE, Sicree RA, Zimmet PZ. Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract . 2010; 87( 1): 4– 14. Google Scholar CrossRef Search ADS PubMed  2. Hinchliffe RJ, Andros G, Apelqvist J, Bakker K, Friederichs S, Lammer J, Lepantalo M, Mills JL, Reekers J, Shearman CP, Valk G, Zierler RE, Schaper NC. A systematic review of the effectiveness of revascularization of the ulcerated foot in patients with diabetes and peripheral arterial disease [published correction appears in Diabetes Metab Res Rev. 28(4):376]. Diabetes Metab Res Rev . 2012; 28( Suppl 1): 179– 217. Google Scholar CrossRef Search ADS PubMed  3. Lobmann R, Schultz G, Lehnert H. Proteases and the diabetic foot syndrome: mechanisms and therapeutic implications. Diabetes Care . 2005; 28( 2): 461– 471. Google Scholar CrossRef Search ADS PubMed  4. Chao CY, Cheing GL. Microvascular dysfunction in diabetic foot disease and ulceration. Diabetes Metab Res Rev . 2009; 25( 7): 604– 614. Google Scholar CrossRef Search ADS PubMed  5. Kyriakides TR, Wulsin D, Skokos EA, Fleckman P, Pirrone A, Shipley JM, Senior RM, Bornstein P. Mice that lack matrix metalloproteinase-9 display delayed wound healing associated with delayed reepithelization and disordered collagen fibrillogenesis. Matrix Biol . 2009; 28( 2): 65– 73. Google Scholar CrossRef Search ADS PubMed  6. Singh K, Agrawal NK, Gupta SK, Mohan G, Chaturvedi S, Singh K. Differential expression of matrix metalloproteinase-9 gene in wounds of type 2 diabetes mellitus cases with susceptible -1562C>T genotypes and wound severity. Int J Low Extrem Wounds . 2014; 13( 2): 94– 102. Google Scholar CrossRef Search ADS PubMed  7. Zhai J, Wang Y. MDI 301, a synthetic retinoid, depressed levels of matrix metalloproteinases and oxidative stress in diabetic dermal fibroblasts. Oncotarget . 2017; 8( 27): 43889– 43896. Google Scholar PubMed  8. Yang C, Zhu P, Yan L, Chen L, Meng R, Lao G. Dynamic changes in matrix metalloproteinase 9 and tissue inhibitor of metalloproteinase 1 levels during wound healing in diabetic rats. J Am Podiatr Med Assoc . 2009; 99( 6): 489– 496. Google Scholar CrossRef Search ADS PubMed  9. Li N, Luo HC, Yang C, Deng JJ, Ren M, Xie XY, Lin DZ, Yan L, Zhang LM. Cationic star-shaped polymer as an siRNA carrier for reducing MMP-9 expression in skin fibroblast cells and promoting wound healing in diabetic rats. Int J Nanomedicine . 2014; 9: 3377– 3387. Google Scholar CrossRef Search ADS PubMed  10. Li N, Luo HC, Ren M, Zhang LM, Wang W, Pan CL, Yang LQ, Lao GJ, Deng JJ, Mai KJ, Sun K, Yang C, Yan L. Efficiency and safety of β-CD-(D3)7 as siRNA carrier for decreasing matrix metalloproteinase-9 expression and improving wound healing in diabetic rats. ACS Appl Mater Interfaces . 2017; 9( 20): 17417– 17426. Google Scholar CrossRef Search ADS PubMed  11. Labrie M, St-Pierre Y. Epigenetic regulation of mmp-9 gene expression. Cell Mol Life Sci . 2013; 70( 17): 3109– 3124. Google Scholar CrossRef Search ADS PubMed  12. Ling C, Groop L. Epigenetics: a molecular link between environmental factors and type 2 diabetes. Diabetes . 2009; 58( 12): 2718– 2725. Google Scholar CrossRef Search ADS PubMed  13. Villeneuve LM, Reddy MA, Natarajan R. Epigenetics: deciphering its role in diabetes and its chronic complications. Clin Exp Pharmacol Physiol . 2011; 38( 7): 451– 459. Google Scholar CrossRef Search ADS PubMed  14. Volkmar M, Dedeurwaerder S, Cunha DA, Ndlovu MN, Defrance M, Deplus R, Calonne E, Volkmar U, Igoillo-Esteve M, Naamane N, Del Guerra S, Masini M, Bugliani M, Marchetti P, Cnop M, Eizirik DL, Fuks F. DNA methylation profiling identifies epigenetic dysregulation in pancreatic islets from type 2 diabetic patients. EMBO J . 2012; 31( 6): 1405– 1426. Google Scholar CrossRef Search ADS PubMed  15. Kowluru RA, Shan Y, Mishra M. Dynamic DNA methylation of matrix metalloproteinase-9 in the development of diabetic retinopathy. Lab Invest . 2016; 96( 10): 1040– 1049. Google Scholar CrossRef Search ADS PubMed  16. Peppa M, Stavroulakis P, Raptis SA. Advanced glycoxidation products and impaired diabetic wound healing. Wound Repair Regen . 2009; 17( 4): 461– 472. Google Scholar CrossRef Search ADS PubMed  17. Zhu P, Yang C, Chen LH, Ren M, Lao GJ, Yan L. Impairment of human keratinocyte mobility and proliferation by advanced glycation end products-modified BSA. Arch Dermatol Res . 2011; 303( 5): 339– 350. Google Scholar CrossRef Search ADS PubMed  18. Tan Q, Wang W, Yang C, Zhang J, Sun K, Luo HC, Mai LF, Lao Y, Yan L, Ren M. α-ketoglutarate is associated with delayed wound healing in diabetes. Clin Endocrinol (Oxf) . 2016; 85( 1): 54– 61. Google Scholar CrossRef Search ADS PubMed  19. Zhang J, Yang C, Wang C, Liu D, Lao G, Liang Y, Sun K, Luo H, Tan Q, Ren M, Yan L. AGE-induced keratinocyte MMP-9 expression is linked to TET2-mediated CpG demethylation. Wound Repair Regen . 2016; 24( 3): 489– 500. Google Scholar CrossRef Search ADS PubMed  20. Ling L, Ren M, Yang C, Lao G, Chen L, Luo H, Feng Z, Yan L. Role of site-specific DNA demethylation in TNFα-induced MMP9 expression in keratinocytes. J Mol Endocrinol . 2013; 50( 3): 279– 290. Google Scholar CrossRef Search ADS PubMed  21. Niehrs C, Schäfer A. Active DNA demethylation by Gadd45 and DNA repair. Trends Cell Biol . 2012; 22( 4): 220– 227. Google Scholar CrossRef Search ADS PubMed  22. Salvador JM, Brown-Clay JD, Fornace AJ, Jr. Gadd45 in stress signaling, cell cycle control, and apoptosis. Adv Exp Med Biol . 2013; 793: 1– 19. Google Scholar CrossRef Search ADS PubMed  23. Schäfer A. Gadd45 proteins: key players of repair-mediated DNA demethylation. Adv Exp Med Biol . 2013; 793: 35– 50. Google Scholar CrossRef Search ADS PubMed  24. Wang N, Yang C, Xie F, Sun L, Su X, Wang Y, Wei R, Zhang R, Li X, Yang B, Ai J. Gadd45α: a novel diabetes-associated gene potentially linking diabetic cardiomyopathy and baroreflex dysfunction. PLoS One . 2012; 7( 12): e49077. Google Scholar CrossRef Search ADS PubMed  25. Barreto G, Schäfer A, Marhold J, Stach D, Swaminathan SK, Handa V, Döderlein G, Maltry N, Wu W, Lyko F, Niehrs C. Gadd45a promotes epigenetic gene activation by repair-mediated DNA demethylation. Nature . 2007; 445( 7128): 671– 675. Google Scholar CrossRef Search ADS PubMed  26. Cortellino S, Xu J, Sannai M, Moore R, Caretti E, Cigliano A, Le Coz M, Devarajan K, Wessels A, Soprano D, Abramowitz LK, Bartolomei MS, Rambow F, Bassi MR, Bruno T, Fanciulli M, Renner C, Klein-Szanto AJ, Matsumoto Y, Kobi D, Davidson I, Alberti C, Larue L, Bellacosa A. Thymine DNA glycosylase is essential for active DNA demethylation by linked deamination-base excision repair. Cell . 2011; 146( 1): 67– 79. Google Scholar CrossRef Search ADS PubMed  27. Schmitz KM, Schmitt N, Hoffmann-Rohrer U, Schäfer A, Grummt I, Mayer C. TAF12 recruits Gadd45a and the nucleotide excision repair complex to the promoter of rRNA genes leading to active DNA demethylation. Mol Cell . 2009; 33( 3): 344– 353. Google Scholar CrossRef Search ADS PubMed  28. He YF, Li BZ, Li Z, Liu P, Wang Y, Tang Q, Ding J, Jia Y, Chen Z, Li L, Sun Y, Li X, Dai Q, Song CX, Zhang K, He C, Xu GL. Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science . 2011; 333( 6047): 1303– 1307. Google Scholar CrossRef Search ADS PubMed  29. Pastor WA, Aravind L, Rao A. TETonic shift: biological roles of TET proteins in DNA demethylation and transcription. Nat Rev Mol Cell Biol . 2013; 14( 6): 341– 356. Google Scholar CrossRef Search ADS PubMed  30. Williams KT, Garrow TA, Schalinske KL. Type I diabetes leads to tissue-specific DNA hypomethylation in male rats. J Nutr . 2008; 138( 11): 2064– 2069. Google Scholar CrossRef Search ADS PubMed  31. Olsen AS, Sarras MP, Jr, Leontovich A, Intine RV. Heritable transmission of diabetic metabolic memory in zebrafish correlates with DNA hypomethylation and aberrant gene expression. Diabetes . 2012; 61( 2): 485– 491. Google Scholar CrossRef Search ADS PubMed  32. Pirola L, Balcerczyk A, Tothill RW, Haviv I, Kaspi A, Lunke S, Ziemann M, Karagiannis T, Tonna S, Kowalczyk A, Beresford-Smith B, Macintyre G, Kelong M, Hongyu Z, Zhu J, El-Osta A. Genome-wide analysis distinguishes hyperglycemia regulated epigenetic signatures of primary vascular cells. Genome Res . 2011; 21( 10): 1601– 1615. Google Scholar CrossRef Search ADS PubMed  33. Dhliwayo N, Sarras MP, Jr, Luczkowski E, Mason SM, Intine RV. Parp inhibition prevents ten-eleven translocase enzyme activation and hyperglycemia-induced DNA demethylation. Diabetes . 2014; 63( 9): 3069– 3076. Google Scholar CrossRef Search ADS PubMed  34. Al-Romaih K, Sadikovic B, Yoshimoto M, Wang Y, Zielenska M, Squire JA. Decitabine-induced demethylation of 5′ CpG island in GADD45A leads to apoptosis in osteosarcoma cells. Neoplasia . 2008; 10( 5): 471– 480. Google Scholar CrossRef Search ADS PubMed  35. Hashimoto K, Oreffo RO, Gibson MB, Goldring MB, Roach HI. DNA demethylation at specific CpG sites in the IL1B promoter in response to inflammatory cytokines in human articular chondrocytes. Arthritis Rheum . 2009; 60( 11): 3303– 3313. Google Scholar CrossRef Search ADS PubMed  36. Walsh CP, Xu GL. Cytosine methylation and DNA repair. Curr Top Microbiol Immunol . 2006; 301: 283– 315. Google Scholar PubMed  37. Ito S, D’Alessio AC, Taranova OV, Hong K, Sowers LC, Zhang Y. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature . 2010; 466( 7310): 1129– 1133. Google Scholar CrossRef Search ADS PubMed  38. Maiti A, Drohat AC. Thymine DNA glycosylase can rapidly excise 5-formylcytosine and 5-carboxylcytosine: potential implications for active demethylation of CpG sites. J Biol Chem . 2011; 286( 41): 35334– 35338. Google Scholar CrossRef Search ADS PubMed  39. Arab K, Park YJ, Lindroth AM, Schäfer A, Oakes C, Weichenhan D, Lukanova A, Lundin E, Risch A, Meister M, Dienemann H, Dyckhoff G, Herold-Mende C, Grummt I, Niehrs C, Plass C. Long noncoding RNA TARID directs demethylation and activation of the tumor suppressor TCF21 via GADD45A. Mol Cell . 2014; 55( 4): 604– 614. Google Scholar CrossRef Search ADS PubMed  40. Xu GL, Walsh CP. Enzymatic DNA oxidation: mechanisms and biological significance. BMB Rep . 2014; 47( 11): 609– 618. Google Scholar CrossRef Search ADS PubMed  41. Kienhöfer S, Musheev MU, Stapf U, Helm M, Schomacher L, Niehrs C, Schäfer A. GADD45a physically and functionally interacts with TET1. Differentiation . 2015; 90( 1–3): 59– 68. Google Scholar CrossRef Search ADS PubMed  42. Gavin DP, Kusumo H, Sharma RP, Guizzetti M, Guidotti A, Pandey SC. Gadd45b and N-methyl-D-aspartate induced DNA demethylation in postmitotic neurons. Epigenomics . 2015; 7( 4): 567– 579. Google Scholar CrossRef Search ADS PubMed  43. Schäfer A, Schomacher L, Barreto G, Döderlein G, Niehrs C. Gemcitabine functions epigenetically by inhibiting repair mediated DNA demethylation. PLoS One . 2010; 5( 11): e14060. Google Scholar CrossRef Search ADS PubMed  44. Engel N, Tront JS, Erinle T, Nguyen N, Latham KE, Sapienza C, Hoffman B, Liebermann DA. Conserved DNA methylation in Gadd45a(−/−) mice. Epigenetics . 2009; 4( 2): 98– 99. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society

Journal

EndocrinologyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off