Myocardin-Related Transcription Factor A Epigenetically Regulates Renal Fibrosis in Diabetic Nephropathy

Abstract

Diabetic nephropathy (DN) is one of the most common complications associated with diabetes and characterized by renal microvascular injury along with accelerated synthesis of extracellular matrix proteins causing tubulointerstitial fibrosis. Production of type I collagen, the major component of extracellular matrix, is augmented during renal fibrosis after chronic exposure to hyperglycemia. However, the transcriptional modulator responsible for the epigenetic manipulation leading to induction of type I collagen genes is not clearly defined. We show here that tubulointerstitial fibrosis as a result of DN was diminished in myocardin-related transcription factor A (MRTF-A) -deficient mice. In cultured renal tubular epithelial cells and the kidneys of mice with DN, MRTF-A was induced by glucose and synergized with glucose to activate collagen transcription. Notably, MRTF-A silencing led to the disappearance of prominent histone modifications indicative of transcriptional activation, including acetylated histone H3K18/K27 and trimethylated histone H3K4. Detailed analysis revealed that MRTF-A recruited p300, a histone acetyltransferase, and WD repeat-containing protein 5 (WDR5), a key component of the histone H3K4 methyltransferase complex, to the collagen promoters and engaged these proteins in transcriptional activation. Estradiol suppressed collagen production by dampening the expression and binding activity of MRTF-A and interfering with the interaction between p300 and WDR5 in renal epithelial cells. Therefore, targeting the MRTF-A–associated epigenetic machinery might yield interventional strategies against DN-associated renal fibrosis.

Type 2 diabetes engenders a series of microvascular pathologies, including diabetic retinopathy, diabetic polyneuropathy, and diabetic nephropathy, as a consequence of glucose intolerance.¹ Chronic exposure of tissues to high levels of glucose results in accumulation of reactive oxygen species, overproduction of proinflammatory mediators, and interstitial fibrosis.^2–4 Renal tubulointerstitial fibrosis represents a hallmark event in the pathogenesis of diabetic nephropathy, the progression of which is associated with poor diagnosis in patients with ESRD.⁵ Despite decades of vigorous research, the mechanism underlying tubulointerstitial fibrosis and hence, an effective interventional strategy remain elusive.

Type I collagen, a heterotrimer consisting of one α1 (encoded by COL1A1) chain and two α2 (encoded by COL1A2) chains, is the major component of extracellular matrix. Expression of COL1A1/COL1A2 genes is markedly increased in humans⁶ and model animals⁷ with diabetic nephropathy in vivo and renal tubular epithelial cells (RTEs) treated with glucose in vitro.⁸ Mounting evidence has implicated epigenetic regulators in the transactivation of collagen type I genes. For instance, the universally expressed histone acetyltransferase p300 activates COL1A1/COL1A2 cells in a number of different cellular contexts.^9–12 Two independent studies have found that SET7/9, a histone monomethyltransferase, and ASH1L, a component of the mammalian H3K4–H3K36 methyltransferase complex, mediate TGF-β–induced COL1A1/COL1A2 transcription by methylating H3K4 on the proximal COL1A1/COL1A2 promoters.^13,14 It is yet unclear how the various histone/DNA-modifying proteins modulate collagen type I gene transactivation and contribute to tubulointerstitial fibrosis in the pathogenesis of diabetic nephropathy.

Myocardin-related transcription factor A (MRTF-A) was initially identified as a cofactor for SRF in cardiac and smooth muscle cells.¹⁵ Later investigations have revealed a more versatile role for MRTF-A in a wide range of pathophysiologic events, including leukocyte adhesion,¹⁶ fibroblast–myofibroblast transition,¹⁷ and megakaryocyte differentiation/maturation.¹⁸ Newly emerged evidence has suggested that MRTF-A can directly activate collagen type I transcription in cardiac and lung fibroblast cells.^19,20 Our findings, as summarized in this report, link MRTF-A activation to high glucose-induced COL1A1/COL1A2 transcription through an epigenetic pathway in vitro and in vivo. Therefore, targeting MRTF-A may yield novel strategies that alleviate renal fibrosis and prevent the progression of diabetic nephropathy.

Results

MRTF-A Deficiency Attenuates Renal Fibrosis in Mice with Diabetic Nephropathy

To evaluate the involvement of MRTF-A in the pathogenesis of renal fibrosis in the context of diabetic nephropathy, wild-type (WT) or MRTF-A–deficient (KO) mice were fed with a high-fat diet (HFD) for 16 weeks. Compared with WT mice, KO mice displayed no detectable differences in body weight (Supplemental Figure 1A), insulin sensitivity (Supplemental Figure 1, B and C), or gross renal morphology (Supplemental Figure 1E) under physiologic conditions. When fed a HFD, KO mice exhibited a small but significant decrease in terms of weight gain (Supplemental Figure 1A). Metabolic parameters, including glucose tolerance (Supplemental Figure 1B) and insulin tolerance (Figure 1C), were both improved in MRTF-A–deficient mice. More importantly, diabetic nephropathy, as measured by plasma albumin versus creatinine levels (Figure 1A), expression of type IV collagen (Supplemental Figure 1D), expansion of mesangial matrix (Supplemental Figure 1E), and infiltration of immune cells (CD3⁺ lymphocytes, CD45⁺ leukocytes, and F4/80⁺ macrophages) (Supplemental Figure 1F), was alleviated in MRTF-A–null mice. Histologic analyses revealed that there was significantly diminished tubulointerstitial fibrosis in KO mice as opposed to WT mice fed the HFD diet (Figure 1, B–D, Supplemental Figure 1G). In addition, mRNA levels of collagen type I were downregulated in KO mice (Figure 1E), indicative of a direct transcriptional repression of fibrogenesis.

Figure 1.

MRTF-A deficiency attenuates renal fibrosis in mice with diabetic nephropathy. WT or KO mice were induced to develop diabetic nephropathy by HFD. (A) Urinary albumin excretion was measured as described in Concise Methods. n=4 mice for each group. Renal fibrosis was evaluated by (B) picrosirius red and (C) Masson's trichrome staining and quantified by Image Pro. (D) Collagen type I levels were evaluated by immunofluorescence staining and quantified by Image Pro. Scale bar, 50 μM. (E) Expression of type I collagen in the kidneys was examined by quantitative PCR. *P<0.05. AL, ad libitum chow.

To solidify the role of MRTF-A in tubulointerstitial fibrosis, we exploited a second mouse model of diabetic nephropathy, in which mice were injected with streptozotocin (STZ). Insulin sensitivity was augmented in KO mice receiving STZ injection compared with WT mice (Supplemental Figure 2A). In addition, we observed amelioration of diabetic nephropathy in MRTF-A–null mice compared with WT mice, which was evidenced by decreases in urinary albumin secretion (Supplemental Figure 2B), Col4a1 expression (Supplemental Figure 2C), expansion of mesangial matrix (Supplemental Figure 2D), inflammatory infiltrates (Supplemental Figure 2E), and renal fibrosis (Supplemental Figure 2, F–I). Together, these data suggest that MRTF-A deficiency might alleviate renal fibrosis in the context of diabetic nephropathy.

MRTF-A Contributes to Glucose-Induced Collagen Type I Transactivation

Having observed that tubulointerstitial fibrosis was ameliorated in mice in the absence of MRTF-A, we examined whether MRTF-A could mediate glucose-induced collagen type I gene transactivation in RTEs. Overexpression of MRTF-A in NRK-52E (Figure 2A), an immortalized rat RTE line, and HK-2 (Supplemental Figure 3A), an immortalized human RTE line, markedly potentiated glucose (35 mM)-induced activation of COL1A1 and COL1A2 promoters. On the contrary, interference with MRTF-A activity by a dominant negative (DN) mutant blocked activation of collagen type I gene transcription by glucose (Figure 2B). Similarly, glucose-induced collagen gene transactivation was lost in MRTF-A–deficient MEF cells (Supplemental Figure 3B). When MRTF-A was depleted from NRK-52E cells with siRNA (Supplemental Figure 3C), induction of endogenous collagen type I mRNA (Figure 2C) and protein (Figure 2D) levels by glucose was also alleviated. Finally, we isolated primary RTEs from either WT or KO mice. Although WT RTEs responded to the treatment of glucose by upregulating collagen expression, the response was significantly weakened in KO RTEs (Figure 2, E and F). Together, these data suggest that MRTF-A is involved in tubulointerstitial fibrosis by directly activating collagen type I gene transcription in response to high glucose.

Figure 2.

MRTF-A contributes to glucose-induced collagen type I transactivation. (A) Collagen promoter luciferase constructs were transfected into NRK-52E cells with or without MRTF-A followed by treatment with glucose. Data are expressed as relative luciferase units (RLUs). (B) Collagen promoter luciferase constructs were transfected into NRK-52E cells with or without MRTF-A DN followed by treatment with glucose. (C and D) NRK-52E cells were transfected with indicated siRNAs followed by treatment with glucose. Expression of type I collagen was examined by (C) quantitative PCR (qPCR) and (D) Western blot. (E and F) Primary RTEs were isolated from WT or KO mice and treated with glucose. Expression of type I collagen was examined by (E) qPCR and (F) Western blot. *P<0.05. HG, high glucose; LG, low glucose; SCR, scrambled RNA.

To identify new MRTF-A target genes involved in fibrosis, we profiled gene expression in primary RTEs isolated from WT or KO mice exposed to high glucose using a PCR array system. Several profibrogenic genes, including α-smooth muscle actin 2, type III collagen (Col3a1), and tissue inhibitor of metalloproteinase 1, were downregulated in MRTF-A–deficient RTEs treated with high glucose compared with WT RTEs (Supplemental Figure 4). These data were validated in the kidneys in two different mouse models of DN (Supplemental Figure 5, A and B) as well as NRK-52E cells (Supplemental Figure 5C). We then used chromatin immunoprecipitation (ChIP) assays to authenticate direct interaction between MRTF-A and putative target genes. MRTF-A binding on gene promoters but not intronic regions was stronger in mice with DN induced by either HFD diet (Supplemental Figure 6A) or STZ injection (Supplemental Figure 6B). In addition, high glucose enhanced MRTF-A occupancies on the promoters of newly identified genes in NRK-52E cells (Supplemental Figure 6C). Collectively, these data suggest that MRTF-A might exert a broad role in regulating fibrogenesis-related transcription in renal epithelial cells leading to tubulointerstitial fibrosis.

MRTF-A Is Activated by Glucose In Vitro and In Vivo

The finding that MRTF-A was involved in high glucose-induced collagen transactivation raised the possibility that MRTF-A expression and/or activity might be affected directly by high glucose in RTEs. Indeed, glucose increased the message and protein levels of MRTF-A in a time-dependent (Figure 3, A and B) and dose-dependent (Supplemental Figure 7, A and B) manner. We also observed an induction of MRTF-A promoter activity by glucose in NRK-52E, indicating that accelerated transcription of MRTF-A likely contributed to increased MRTF-A levels under the influence of glucose (Supplemental Figure 7C). In addition, glucose exposure stimulated the occupancy of MRTF-A but not that of a control IgG on collagen type I gene promoters (Figure 3C). Finally, we examined the expression and activity of MRTF-A in mice. MRTF-A messages went up by 50% in the kidneys of mice that were fed the HFD to develop diabetic nephropathy compared with mice on a control diet (Figure 3D). MRTF-A binding to the collagen promoters was also enhanced in the kidneys of DN mice (Figure 3E). Similarly, MRTF-A expression (Figure 3F) and binding activity (Figure 3G) went up in the kidneys of mice that were induced to develop DN by STZ injection. Of interest, we observed that induction of MRTF-A expression (Supplemental Figure 8A) and binding activity (Supplemental Figure 8B) in the presence of high glucose was attenuated by 17β-estradiol, which has previously been shown to improve renal function in the context of DN.²¹ Collectively, these data show that MRTF-A expression and activity were responsive to glucose in vitro and in vivo.

Figure 3.

MRTF-A is activated by glucose in vitro and in vivo. (A–C) NRK-52E cells were treated with glucose (35 mM) and harvested at indicated time points. Expression of MRTF-A was measured by (A) quantitative PCR (qPCR) and (B) Western blot. (C) ChIP assays were performed with anti–MRTF-A or IgG. (D and E) C57/BL6 mice were fed an HFD to induce diabetic nephropathy as described in Concise Methods. (D) MRTF-A expression in the kidneys was measured by qPCR. (E) ChIP assays were performed using kidney lysates with anti–MRTF-A. n=3 for each group. (F and G) C57/BL6 mice were injected with STZ to induce diabetic nephropathy as described in Concise Methods. (F) MRTF-A expression in the kidneys was measured by qPCR. n=3 mice for each group. (G) ChIP assays were performed using kidney lysates with anti–MRTF-A. n=3 for each group. *P<0.05. AL, ad libitum chow; Gapdh, glyceraldehyde-3-phosphate dehydrogenase; vec, vehicle.

RTF-A Alters the Chromatin Structure on the Type I Collagen Gene Promoters

To examine the mechanism through which MRTF-A contributes to high glucose-induced collagen type I gene transactivation in renal epithelial cells, we performed ChIP assays with antibodies that recognize active histone modifications. In NRK-52E, glucose treatment stimulated the enrichment of acetylated histone H3 lysine 18 (H3K18Ac) (Figure 4B), H3K27Ac (Figure 4C), and trimethylated H3K4 (H3K4Me3) (Figure 4D) but not H3K9Ac (Figure 4A) on the collagen promoters. When MRTF-A was silenced by siRNA, all of the active histone modifications except for H3K9Ac were erased. We also used primary RTEs to perform similar experiments. In WT RTEs, high glucose invoked the accumulation of H3K18Ac (Figure 4F), H3K27Ac (Figure 4G), and H3K4Me3 (Figure 4H) on the collagen promoters, all of which disappeared in MRTF-A–deficient RTEs. On the contrary, there was no significant change in H3K9Ac levels in either cell type (Figure 4E). ChIP assays using kidney lysates from mice induced to develop DN by either HFD (Supplemental Figure 9) or STZ injection (Supplemental Figure 10) showed similar patterns. We were also able to confirm that MRTF-A orchestrated differential histone modifications on the promoters of newly identified target genes, including Col1a3, α-smooth muscle actin 2, and tissue inhibitor of metalloproteinase 1, in the kidneys of DN mice (Supplemental Figures 11 and 12). Therefore, MRTF-A may participate in glucose-induced collagen transactivation in renal epithelial cells by shaping the chromatin structure.

Figure 4.

MRTF-A influences the chromatin structure on the type I collagen gene promoters. (A–D) NRK-52E cells were transfected with indicated siRNAs followed by treatment with glucose. ChIP assays were performed with (A) anti-AcH3K9, (B) anti-AcH3K18, (C) anti-AcH3K27, and (D) anti-H3K3Me3. (E–H) Primary RTEs were isolated from WT or KO mice and treated with glucose. ChIP assays were performed with (E) anti-AcH3K9, (F) anti-AcH3K18, (G) anti-AcH3K27, and (H) anti-H3K3Me3. *P<0.05. HG, high glucose; LG, low glucose; SCR, scrambled RNA.

MRTF-A Cooperates with p300 to Activate Type I Collagen Transcription

It has been showed that the histone acetyltransferase p300 is responsible for H3K18/H3K27 acetylation in cells, whereas PCAF acetylates H3K9.²² Therefore, we tested the hypothesis that MRTF-A recruits p300 to the collagen promoter to activate transcription in response to glucose treatment. When NRK-52E cells were exposed to high glucose, the occupancy of p300 on the collagen promoters but not the GADPH promoter was increased with time (Figure 5A). In addition, re-ChIP revealed that a p300–MRTF-A complex was detectable on the collagen promoters only when cells were treated with high glucose (Figure 5B). Two lines of evidence suggest that p300 binding on the collagen promoters depends on MRTF-A. First, when siRNA was used to silence MRTF-A in NRK-52E cells, treatment of high glucose was unable to induce the binding of p300 (Figure 5C). Second, RTEs isolated from KO mice exhibited reduced p300 binding on the collagen promoters compared with RTEs isolated from WT mice (Supplemental Figure 13A).

Figure 5.

MRTF-A cooperates with p300 to activate type I collagen transcription. (A) NRK-52E cells were treated with glucose (35 mM) and harvested at indicated time points. ChIP assays were performed with anti-p300 or IgG. (B) NRK-52E cells were treated with or without glucose (35 mM) for 24 hours. Re-ChIP assay was performed with indicated antibodies. (C) NRK-52E cells were transfected with indicated siRNAs followed by treatment with glucose. ChIP assay was performed with anti-p300. (D) Collagen promoter luciferase constructs were transfected into NRK-52E cells with indicated expression constructs. (E) Collagen promoter luciferase constructs were transfected into NRK-52E cells with indicated expression constructs followed by treatment with glucose or mannitol. (F and G) NRK-52E cells were transfected with indicated siRNAs followed by treatment with glucose. Expression of type I collagen was examined by (F) quantitative PCR and (G) Western blot. *P<0.05. HG, high glucose; LG, low glucose; RLU, relative luciferase unit; SCR, scrambled RNA.

When overexpressed, p300 synergized with MRTF-A and glucose to activate the promoter activities of type I collagen genes (Figure 5, D and E). In contrast, PCAF did not alter MRTF-A–induced activation of collagen promoters (Supplemental Figure 13B). However, siRNA-mediated knockdown of p300 (Supplemental Figure 13C) diminished the capacity of MRTF-A to activate the collagen promoters (Supplemental Figure 13D); p300 silencing by siRNA also abrogated the induction of endogenous collagen type I messages and proteins by high-glucose treatment (Figure 5, F and G). Together, these data suggest that p300 is indispensible for high glucose-induced, MRTF-A–mediated collagen type I gene transcription.

MRTF-A Cooperates with WD Repeat-Containing Protein 5 to Activate Type I Collagen Transcription

WD repeat-containing protein 5 (WDR5) is the core component of the mammalian H3K4 trimethyltransferase complex.²³ Therefore, we examined whether WDR5 could be involved in type I collagen transactivation. Exposure to high glucose increased the occupancy of WDR5 and additionally promoted the interaction between WDR5 and MRTF-A on the collagen gene promoters (Figure 6, A and B). Depletion of MRTF-A by siRNA or genetic ablation of MRTF-A prevented the recruitment of WDR5 (Figure 6C, Supplemental Figure 14A). Reciprocally, WDR5 enhanced the ability of MRTF-A and glucose to activate collagen transcription (Figure 6, D and E). Finally, siRNA targeting WDR5 (Supplemental Figure 14B) blocked the induction of collagen expression in renal epithelial cells by glucose (Figure 6, F and G). These data support our hypothesis that the interplay between WDR5 and MRTF-A is essential for glucose-induced collagen transactivation.

Figure 6.

MRTF-A cooperates with WDR5 to activate type I collagen transcription. (A) NRK-52E cells were treated with glucose (35 mM) and harvested at indicated time points. ChIP assays were performed with anti-WDR5 or IgG. (B) NRK-52E cells were treated with or without glucose (35 mM) for 24 hours. Re-ChIP assay was performed with indicated antibodies. (C) NRK-52E cells were transfected with indicated siRNAs followed by treatment with glucose. ChIP assay was performed with anti-WDR5. (D) Collagen promoter luciferase constructs were transfected into NRK-52E cells with indicated expression constructs. (E) Collagen promoter luciferase constructs were transfected into NRK-52E cells with indicated expression constructs followed by treatment with glucose or mannitol. (F and G) NRK-52E cells were transfected with indicated siRNAs followed by treatment with glucose. Expression of type I collagen in the kidneys was examined by (F) quantitative PCR and (G) Western blot. *P<0.05. HG, high glucose; LG, low glucose; RLU, relative luciferase unit; SCR, scrambled RNA.

p300 And WDR5 Synergistically Activate Glucose-Induced Collagen Type I Transactivation

A recent investigation revealed crosstalk between histone acetylation and H3K4 methylation in activating p53-dependent transcription.²⁴ Therefore, we examined the possibility that a similar scheme might exist in glucose-induced collagen transcription. Indeed, exposure to high glucose induced a p300–WDR5 complex on the collagen gene promoters (Figure 7A). Furthermore, coexpression of p300 and WDR5 synergistically activated collagen promoter activities in the presence of glucose (Figure 7B). The interplay between p300 and WDR5 seemed to be bridged by MRTF-A, because the p300–WDR5 complex disappeared from the collagen promoters in primary renal epithelial cells isolated from KO mice (Supplemental Figure 15A). Treatment of estrogen also led to the disruption of the p300–WDR5 complex on the collagen promoters induced by glucose, possibly by its virtue of downregulating MRTF-A expression (Supplemental Figure 15B). Collectively, these data point to a role for MRTF-A as a coordinator of the dialogue between different histone-modifying enzymes.

Figure 7.

p300 and WDR5 synergistically activate glucose-induced collagen type I transactivation. (A) NRK-52E cells were treated with or without glucose (35 mM) for 24 hours. Re-ChIP assay was performed with indicated antibodies. (B) Collagen promoter luciferase constructs were transfected into NRK-52E cells with indicated expression constructs followed by treatment with glucose or mannitol. *P<0.05. RLU, relative luciferase.

Discussion

Chronic exposure to high glucose is intimately associated with various diabetic complications. Recent investigations have established a link between the epigenetic machinery and high glucose-induced cellular dysfunction.^25–27 In this study, we show that the transcriptional modulator MRTF-A could function as a sensor of glucose in renal epithelial cells to promote collagen type I gene expression and contribute to renal tubulointerstitial fibrosis by coordinating histone H3K18/H3K27 acetylation and H3K4 trimethylation.

In response to high glucose, RTEs switch to a proinflammatory and profibrotic phenotype.²⁸ In MRTF-A–deficient mice challenged with diabetic stress, collagen production is appreciably reduced compared with WT littermates, confirming previous reports that MRTF-A is essential for cellular fibrogenesis.^19,20,29 HFD-induced renal tubulointerstitial fibrosis is attenuated in MRTF-A–deficient mice. Consistent with this model, we show here that MRTF-A, after it is activated by glucose, can directly upregulate collagen type I synthesis by recruiting histone acetyltransferases and methyltransferases in both primary RTEs and immortalized tubular epithelial cells. These observations, however, do not rule out the possibility that MRTF-A might influence fibrogenesis in diabetic nephropathy through alternative routes in vivo. MRTF-A is known to promote epithelial-to-mesenchymal transition.^30,31 A reasonable hypothesis would be that decreased synthesis of collagen in glucose-treated MRTF-A–null epithelial cells is a result of decelerated epithelial-to-mesenchymal transition.³² We have previously shown that MRTF-A silencing dampens the interaction between circulating leukocytes and vascular endothelium.¹⁶ Because infiltration of innate immune cells contributes to tissue fibrosis,³³ it is likely that reduced renal fibrosis might be a secondary consequence of a skewed immune response. Additionally, MRTF-A–deficient mice were more resistant to weight gain than WT littermates on an HFD (Supplemental Figure 1), indicating that attenuated renal fibrosis in vivo might be secondary to improved metabolic profile. These uncertainties will have to be addressed in tissue-specific KO animal models in future studies.

Several recent reports have highlighted the role of the epigenetic machinery in mediating cellular response to nutrient surplus (e.g., high glucose). Accumulation of histone acetylation and H3K4 trimethylation on gene promoters after exposure to various nutrients has been noted in jejunum,³⁴ mesangial cells,¹³ and endothelial cells.³⁵ Meanwhile, mice harboring two point mutations in the Mll2 gene, which encodes one of the H3K4 methyltransferases, exhibit spontaneous insulin resistance.³⁶ Of interest, Soetikno et al.³⁷ have shown that curcumin, a nonspecific p300 inhibitor, alleviates diabetic nephropathy in STZ-injected rats. Our data allude to a scenario where high glucose engages p300 and WDR5 in collagen activation and tubulointerstitial fibrosis and reinforce the notion that screening for small molecule compounds that can inhibit the activities of these proteins may yield novel therapeutic solutions.

Crosstalk between different histone-modifying enzymes has recently emerged as a theme in regulating gene expression. Roeder and colleagues²⁴ have shown that p300-mediated histone acetylation serves as a prerequisite for H3K4 trimethylation and transcriptional activation by p53. We show that high glucose promotes the interaction between p300 and WDR5 on the collagen promoter in an MRTF-A–dependent manner. To our knowledge, this is the first demonstration that crosstalk between two different histone modifications can sense the energy input to regulate cellular behavior. More importantly, MRTF-A seemed to be necessary for the p300–WDR5 complex on collagen gene promoters (Supplemental Figure 7). This is consistent with the notion that epigenetic factors (histone enzymes included) rely on transcription factors to be recruited to specific sites for their participation in the regulation of gene expression.³⁸ A recent study using a high-throughput ChIP-sequencing technique revealed that distinctive peaks of H3K9/K14 on both proximal and distal regulatory elements in hyperglycemic endothelial cells can be used to explain gene expression profile in these cells.³⁹ It is of great interest to determine on a genome-wide scale how MRTF-A modulates the occupancies of p300 and WDR5, whether this calibration can respond to energy input, and how these dynamics contribute to changes in cellular function.

In summary, this report portrays MRTF-A as a glucose-responsive regulator of fibrogenesis in renal epithelial cells through coordination of the crosstalk between p300 and WDR5. Pharmaceutical screening of compounds capable of disrupting these interactions may offer novel treatment options for patients with late-stage diabetic renal complications.

Concise Methods

Cell Culture

Immortalized rat (NRK-52E; ATCC) and human (HK-2; ATCC) RTEs were maintained in DMEM supplemented with 10% FBS. Primary murine RTEs were isolated as described previously.⁴⁰

Plasmids, Transfection, and Reporter Assay

Promoter–luciferase fusion constructs for Col1a1 and Col1a2¹⁰ and expression constructs for MRTF-A,⁴¹ p300,⁴² and WDR5⁴³ have been previously described. siRNA sequences were CAGGUGAAUUACCCAAAGGUA for rat MRTF-A, CCCCAUGGAACAGCAU for p300, and CACCUGUUAAGCCAAACUA for WDR5. Transient transfections were performed with Lipofectamine 2000 (Invitrogen). Luciferase activities were assayed using a luciferase reporter assay system (Promega). All experiments were repeated at least three times.

Animals

All animal protocols were approved by the Nanjing Medical University Intramural Ethics Committee on Animal Studies. To induce diabetic nephropathy, 6- to 8-week-old male WT or MRTF-A KO⁴⁴ C57/BL6 mice were fed with an HFD (D12331; Research Diets) for 16 weeks to induce diabetic nephropathy. Alternatively, mice were injected intraperitoneally with STZ (200 mg/kg) as previously reported⁴⁵ and euthanized 16 weeks after injection. Mice were housed in metabolic cages to collect urine samples. Proteinuria was examined using commercially available kits as previously described.⁴⁶ For glucose tolerance tests, mice fasted overnight were injected intraperitoneally with 2 g/kg glucose, and blood samples were taken at the indicated intervals. For insulin tolerance tests, mice fasted overnight were injected intraperitoneally with 0.75 IU/kg soluble insulin. Blood glucose was measured using an Accu-Chek compact glucometer (Roche).

Protein Extraction, Immunoprecipitation, and Western Blot

Whole-cell lysates and nuclear proteins were obtained as previously described.¹⁶ Specific antibodies or preimmune IgGs were added to and incubated with cell lysates overnight before being absorbed by Protein A/G-plus Agarose beads (Santa Cruz Biotechnology). Precipitated immune complex was released by boiling with 1×SDS electrophoresis sample buffer. Western blot analyses were performed with anticollagene type I (Rockland), anti–β-actin (Sigma-Aldrich), anti–MRTF-A, anti-p300, anti-GAPDH (Santa Cruz Biotechnology), and anti-WDR5 (Bethyl Laboratories) antibodies.

RNA Isolation and Real-Time PCR

RNA was extracted with the RNeasy RNA Isolation Kit (Qiagen). Reverse transcription reactions were performed as previously described using a SuperScript First-Strand Synthesis System (Invitrogen).⁴⁷ Primers and Taqman probes used for real-time reactions were purchased from Applied Biosystems.

PCR Array

The fibrosis PCR array was performed using the RT2 Profiler PCR Array System from Qiagen following the manufacturer’s instructions. Briefly, primary mouse RTEs isolated from either WT (control group) or MRTF-A–deficient (group 1) mice were treated with high glucose (35 mM) for 24 hours; 1 μg total RNA extracted from the two groups of cells was reverse transcribed using the RT2 First Strand Kit supplied by the vendor. Then, the cDNA was mixed with 2×RT2 SYBR Green Mastermix, and 25 μl mix was dispensed into the Mouse Fibrosis PCR Array (PAMM-120Z). Quantitative PCR was performed on an Applied Biosystems StepOnePlus System. Cycle threshold values were calculated using StepOne software v2.1 with automatic baseline settings and a threshold of 1.2. The fold change for each gene was calculated using the ΔΔcycle threshold method and normalized by the housekeeping gene supplied by the vendor.

ChIP

ChIP and re-ChIP assays were performed essentially as described before⁴⁸ with anti–MRTF-A, anti-p300 (Santa Cruz Biotechnology), anti-acetyl H3K9, anti-acetyl H3K18, anti-acetyl H3K27, anti-trimethyl H3K4 (EMD Millipore), or anti-WDR5 (Bethyl Laboratories). Precipitated genomic DNA was amplified by real-time PCR with the primers listed in Table 1. Serially diluted genomic DNA extracted from normal cells/tissues was used to generate a standard curve to calculate the amount of DNA being precipitated by a particular antibody. A total of 10% of the starting material was also included as the input. Data were then normalized to the input and expressed as fold changes compared with the control group.

Table 1.

ChIP real-time quantitative PCR primers

Gene Name	Primer Sequences
Mouse Col1a1	Forward: 5′-ATTTGAAGTCCCAGAAAG-3′
	Reverse: 5′-AGAAACTCCCGTCTGCTC-3′
Mouse Col1a2	Forward: 5′-CTTCGTGCATGACTTCAGCTTT-3′
	Reverse: 5′-CGTCCTTTAGCATGGCAAGAC-3′
Mouse Col1a3 #1	Forward: 5′-GACTCTGGCAAAACTCAAAGTATCA-3′
	Reverse: 5′-TAGGAATGTGCTTTGTGATAGCCT-3′
Mouse Col1a3 #2	Forward: 5′-AGACCTTCATTCCCAGCTACTTG-3′
	Reverse: 5′-CTCTCTACCACTGACCTGCATCTC-3′
Mouse Acta2	Forward: 5′-AGCAGAACAGAGGAATGCAGTGGAAGAGAC-3′
	Reverse: 5′-CCTCCCACTCGCCTCCCAAACAAGGAGC-3′
Mouse Timp1 #1	Forward: 5′-AGGACTGTGCATGACGTGGAG-3′
	Reverse: 5′-ACAGTGGAGAATAAATGTCCATGC-3′
Mouse Timp1 #2	Forward: 5′-TGTGGTCAAGCAAAGCATCTG-3′
	Reverse: 5′-TGGGTTTGTAGCTCAATTGTGC-3′
Rat Col1a1	Forward: 5′-ATCCTTCTGATTTGAGGTC-3′
	Reverse: 5′-AGGTGAAACTCCCGTCTG-3′
Rat Col1a2	Forward: 5′-GACATGCTCAAGTGCTGAGTCAC-3′
	Reverse: 5′-AGATTGCACAATGTGACGTCG-3′
Rat Col3a1	Forward: 5′-ATCCTTCTGATTTGAGGTC-3′
	Reverse: 5′-AGGTGAAACTCCCGTCTG-3′
Rat Timp1	Forward: 5′-CTCTGCCACCCCTCACCA-3′
	Reverse: 5′-GGACTGGATGGGCCTCGT-3′
Rat Acta2	Forward: 5′-CATGCACGTGGACTGTACCT-3′
	Reverse: 5′-AAAGATGCTTGGGTCACCTG-3′
Rat Gapdh	Forward: 5′-ATCACTGCCACCCAGAAGACTGTGGA-3′
	Reverse: 5′-CTCATACCAGGAAATGAGCTTGACAAA-3′

Acta2, α-smooth muscle actin2; Gapdh, glyceraldehyde-3-phosphate dehydrogenase; Timp1, tissue inhibitor of metalloproteinase 1.

Histology

Histologic analyses were performed essentially as described before^49,50 Pictures were taken using an Olympus IX-70 microscope and quantified with Image Pro.

Statistical Analyses

One-way ANOVA with post hoc Scheffe analyses were performed using the SPSS package. Unless otherwise specified, P values<0.05 were considered statistically significant.

Disclosures

None.