Biochimica et Biophysica Acta 1849 (2015) 1219–1228

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbagrm

MKL1 is an epigenetic modulator of TGF-β induced fibrogenesis

Zhiwen Fan a, Chenzhi Hao a,MinLia, Xin Dai a, Hao Qin a,JianfeiLia,HuihuiXua, Xiaoyan Wu a, Liping Zhang b,MingmingFanga,c, Bisheng Zhou a, Wenfang Tian a,YongXua,⁎ a Key Laboratory of Cardiovascular Disease, Department of Pathophysiology, Nanjing Medical University, Nanjing, China b Department of Biochemistry, Xinjiang Medical University, Urumqi, China c Department of Nursing, Jiangsu Jiankang Vocational University, Nanjing, China article info abstract

Article history: Transforming growth factor (TGF-β) induced activation of portal fibroblast cells serves as a primary cause for Received 16 May 2015 liver fibrosis following cholestatic injury. The underlying epigenetic mechanism is not clear. We studied the Received in revised form 13 July 2015 role of a transcriptional modulator, megakaryoblastic leukemia 1 (MKL1) in this process. We report here that Accepted 31 July 2015 MKL1 deficiency ameliorated BDL-induced liver fibrosis in mice as assessed by histological stainings and expres- Available online 1 August 2015 sion levels of pro-fibrogenic . MKL1 silencing by small interfering RNA (siRNA) abrogated TGF-β induced transactivation of pro-fibrogenic genes in portal fibroblast cells. TGF-β stimulated the binding of MKL1 on the Keywords: fi Epigenetics promoters of pro- brogenic genes and promoted the interaction between MKL1 and SMAD3. While SMAD3 Transcriptional regulation was necessary for MKL1 occupancy on the promoters, MKL1 depletion impaired SMAD3 binding reciprocal- Portal fibroblast ly. TGF-β treatment induced the accumulation of trimethylated histone H3K4 on the gene promoters by Liver fibrosis recruiting a methyltransferase complex. Knockdown of individual members of this complex significantly weak- TGF-β ened the binding of SMAD3 and down-regulated the activation of portal fibroblast cells. In conclusion, we have Histone methylation identified an epigenetic pathway that dictates TGF-β induced pro-fibrogenic transcription in portal fibroblast thereby providing novel insights for the development of therapeutic solutions to treat liver fibrosis. © 2015 Elsevier B.V. All rights reserved.

1. Introduction primarily derived from portal fibroblasts [6]. A host of growth factors and metabolites have been shown to activate portal fibroblasts, among Fibrogenesis is an adaptive response to liver injury, which can lead which TGF-β is by far the most extensively studied [7–9].Itisunclear, to, if left unattended to, irreversible tissue scar formation, cirrhosis however, how the epigenetic machinery modulates TGF-β signaling and and hepatocellular carcinoma [1]. In the process of liver fibrosis, there pro-fibrogenic transcriptional program in portal fibroblast cells. is accelerated synthesis and deposition of extracellular matrix (ECM) Megakaryoblastic leukemia 1 (MKL1), also termed - such as collagen type I, collagen type III, and fibronectin. related transcription factor A (MRTF-A), is a transcriptional modulator The mechanism underlying transactivation of these so-called “pro- that plays a redundant role in smooth muscle cell differentiation [10]. fibrogenic” genes remains incompletely understood. Independent investigations in our laboratory as well as several It is generally believed that hepatic stellate cells (HSCs), transitioning other laboratories have recently uncovered a role for MKL1 in tissue from a quiescent lipid-storing state to a proliferative ECM-producing fibrogenesis in a number of different settings including ischemia- state, primarily contribute to liver fibrosis following hepatotoxic injuries induced myocardial fibrosis [11], bleomycin-induced pulmonary fibro- (e.g., chemicals, viruses) [2]. On the other hand, there has been much con- sis [12], and diabetes-induced renal fibrosis [13]. Previous investigations troversy regarding the source of cells for liver fibrogenesis as a result of have also revealed that MKL1 regulates transcription by bridging the cholestatic injury (e.g., bile duct obstruction); several different types of epigenetic machinery to various intrinsic and extrinsic stress cues cells, including HSCs and hepatocytes, have been suggested to play a [14–18].Buildingonthesefindings, we sought to clarify the role of role in this process [3–5]. By using a series of delicate lineage tracing MKL1 in TGF-β induced activation of portal fibroblast cells and the experiments, Iwaisako et al have recently demonstrated that the pro- underlying epigenetic mechanism in the present study. fibrogenic myofibroblasts observed in cholestatic liver injury are 2. Materials and methods

Abbreviations: TGF-β, transforming growth factor; MKL1, megakaryoblastic leukemia 2.1. Animals 1; BDL, bile duct ligation; ChIP, chromatin immunoprecipitation; COMPASS, complex proteins associated with SET; TF, transcription factor fi ⁎ Corresponding author at: Nanjing Medical University, Nanjing, Jiangsu 210029, China. MKL1 de cient (KO) mice were obtained from Dr. Steve Morris at E-mail address: [email protected] (Y. Xu). St. Jude Children's Hospital [19]. To induce liver injury, 8-week old, KO

http://dx.doi.org/10.1016/j.bbagrm.2015.07.013 1874-9399/© 2015 Elsevier B.V. All rights reserved. 1220 Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228 male mice and wild type (WT) littermates were anesthetized with Col1a2 promoter, 5′-GACATGCTCAAGTGCTGAGTCAC-3′ and 5′- ketamine. The common bile duct was ligated twice with silk sutures. AGATTGCACAATGTGACGTCG-3′; Acta2 promoter, 5′-CATGCACGTG Bile duct ligation (BDL) and sham-operated mice were sacrificed two GACTGTACCT-3′ and 5′-AAAGATGCTTGGGTCACCTG-3′; Gapdh pro- weeks following the surgical procedure. All animal experiments have moter, 5′-ATCACTGCCACCCAGAAGACTGTGGA-3′ and 5′-C TCATAC been performed following guidelines by the intramural Committee on CAGGAAATGAGCTTGACAAA-3′. All experiments were repeated at Ethical Conduct of Animal Studies. least three times.

2.2. Cell culture and treatment 2.7. Histology

Primary rat portal fibroblast cells were isolated and maintained as Histological analyses were performed essentially as described before previously described [20]. The identity of primary portal fibroblast [27,30]. Briefly, paraffin sections were stained with hematoxylin and cells were verified by immunostaining as previously described [21]. eosin (Sigma), picrosirius red (Sigma), or Masson's trichrome (Sigma) Mouse embryonic fibroblast cells were isolated from wild type and according to standard procedures. Alternatively, the sections were MKL1 deficient mice [19].TGF-β was purchased from R&D. Prior to blocked with 10% normal goat serum for 1 h at room temperature and treatment, cells were starved in .4% DMEM overnight. then incubated with anti-collagen type I or anti-α-SMA antibodies. Staining was visualized by incubation with an appropriate biotinylated 2.3. Plasmids, siRNAs, and transient transfection 2° antibody and developed with a streptavidin-horseradish peroxidase kit (Pierce) for 20 min. Pictures were taken using an Olympus IX-70 mi- Col1a1, Col1a2,andActa2 promoter–luciferase fusion constructs [22] croscope. For quantification, positive stainings were counted using as well as small interfering RNA (siRNA) sequences for rat ASH2, WDR5, Image Pro (Media Cybernetics). At least 3 slides were included for and SET1 [23] have been have been previously described. FLAG-tagged each mouse and for each slide at least 5 fields were counted. Data are full-length MKL1 expression construct in a pCMV backbone was a gift expressed as relative staining compared to the control group which is from Dr. Ron Prywes [24]. Transient transfections were performed arbitrarily set as 1. with Lipofectamine 2000 (Invitrogen). Luciferase activities were assayed 24–48 h after transfection using a luciferase reporter assay 2.8. Statistical analysis system (Promega). For retrovirus production, the FLAG-tagged MKL1 was amplified from its original vector by PCR and sub-cloned into a One-way ANOVA with post-hoc Scheffe analyses were performed pBabe-puro vector [25]. Retroviral stocks were produced in 293FT using an SPSS package. Unless otherwise specified, p values smaller cells as previously described [26] and used to infect MEF cells in the than .05 were considered statistically significant (*). presence of polybrene (Sigma). All experiments were repeated at least three times. 3. Results

2.4. extraction and Western blotting 3.1. MKL1 deficient mice are resistant to bile duct ligation (BDL) induced liver fibrosis Whole-cell lysates were obtained by re-suspending cell pellets in RIPA buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1% Triton X-100) with We first examined the involvement of MKL1 in TGF-β induced liver freshly added protease inhibitor (Roche). Nuclear proteins were ex- fibrogenesis in a murine BDL model, in which TGF-β has been shown to tracted using a NE-PER kit (Thermo Fisher) as previously described play a key role in driving the synthesis of multiple ECM proteins [27]. Western blot analyses were performed with anti-collagen type I [31–33]. BDL-induced liver fibrosis differs from chemical/toxin- (Rockland, 600-401-103), anti-α-SMA (Sigma, A2547), anti-β-actin induced liver fibrosis in that portal fibroblast cell plays a central role (Sigma, A1978), anti-α-tubulin (Sigma, T9026), anti-MKL1 (Santa in the former model whereas hepatic stellate cell plays a dominant Cruz, SC-32909), anti-Lamin B (Santa Cruz, SC-6216), anti-ASH2 (Bethyl role in the latter [6]. BDL procedure triggered liver fibrosis in both Laboratories, A300-489A), anti-WDR5 (Bethyl Laboratories, A300- wild type and MKL1-deficient mice although MKL1 deficiency signifi- 289A), and anti-SET1 (Bethyl Laboratories, A302-429A) antibodies. All cantly down-regulated fibrogenesis as evidenced by picrosirius red, experiments were repeated at least three times. which stains fibrous collagens (e.g., collagen type I and type III) (Fig. 1A) and Masson's trichrome, which stains most ECM components 2.5. RNA isolation and real-time PCR including collagens (Fig. 1B) stainings. Quantitative PCR (qPCR) and Western blotting showed that expression levels of collagen type I RNA was extracted with the RNeasy RNA isolation kit (Qiagen). (encoded by col1a1 and col1a2 genes) and alpha smooth muscle actin Reverse transcriptase reactions were performed as previously described (encoded by acta2 gene) were reduced in MKL1 knockout mice com- using a SuperScript First-strand Synthesis System (Invitrogen) [28]. pared to wild type littermates (Fig. 1C, D). Diminished fibrogenesis in Primers and Taqman probes used for real-time reactions were the liver as a result of MKL1 deficiency was also corroborated by histo- purchased from Applied Biosystems. All experiments were repeated at chemical staining with an anti-type I collagen antibody or an anti-α- least three times. SMA antibody (Fig. S1A, S1B). Of note, suppressed fibrogenesis could not be fully accounted for by the extent of liver injury because MKL1 de- 2.6. Chromatin immunoprecipitation (ChIP) ficiency only modestly alleviated plasma alanine aminotransferase (ALT) levels (Fig. S1C). Together, these data suggest that MKL1 is essen- ChIP and Re-ChIP assays were performed essentially as de- tial for liver fibrosis following cholestatic injury. scribed before [29] using an EZ-Magna ChIP kit (Millipore). ChIP reactions (100 μg/reaction) were performed with anti-MKL1 (Santa 3.2. MKL1 mediates TGF-β induced fibrogenesis in portal fibroblast cells Cruz, SC-32909), anti-SMAD3 (Abcam, Ab28379), anti-trimethyl H3K4 (Millipore, 07-473), anti-ASH2 (Bethyl Laboratories, A300-489A), A recent report using the lineage tracking technique has unequivo- anti-WDR5 (Bethyl Laboratories, A300-289A), and anti-SET1 (Bethyl cally shown that biliary portal fibroblast cells are the major source of Laboratories, A302-429A). Precipitated genomic DNA was amplified fibrogenesis in BDL induced liver injury [6]. Therefore, we verified the by real-time PCR with the following primers: Col1a1 promoter, 5′- role of MKL1 in TGF-β induced fibrogenesis in rat primary portal fibro- ATCCTTCTGATTTGAGGTC-3′ and 5′-AGGTGAAACTCCCGTCTG-3′; blast cells. Depletion of endogenous MKL1 using small interfering RNA Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228 1221

Fig. 1. MKL1 deficient mice are resistant to bile duct ligation (BDL) induced liver fibrosis. 8-week old male wild type (WT) or MKL1 knockout (KO) mice were subject to bile duct ligation (BDL) as described under Materials and methods section. Mice were sacrificed 2 weeks following the procedure. (A, B) Paraffin sections were stained with picrosirius red (A) and Masson's trichrome (B). Magnification: 10×; scale bar, 100 μm. (C, D) Expression of fibrogenic genes was measured by qPCR (C) and Western blotting (D).

(siRNA) down-regulated the induction of col1a1, col1a2,andacta2 undergone the BDL procedure than the sham mice (Fig. S2A). In addi- genes as evidenced by both qPCR (Fig. 2A) and Western blotting tion, TGF-β treatment directly promoted MKL1 enrichment in cultured (Fig. 2B) measurements. To further assess the requirement of MKL1 in portal fibroblast cells (Fig. S2B). Finally, ChIP assay showed that TGF-β TGF-β induced transactivation of pro-fibrogenic genes, we transfected treatment increased the occupancies of MKL1 on the promoter regions reporter constructs for col1a1, col1a2, and acta2 gene promoters into of col1a1, col1a2,andacta2 genes, but not the gapdh gene (Fig. 2E). Col- embryonic fibroblast cells isolated from wild type (WT) and MKL1 lectively, these data demonstrate that MKL1 is essential for TGF-β in- knockout (KO) mice, respectively. As shown in Fig. 2C, while TGF-β duced cellular fibrogenesis. stimulated the promoter activities of col1a1, col1a2,andacta2 genes in WT cells, the transactivation was lost in KO cells; ectopically introduced 3.3. TGF-β promotes the interaction between MKL1 and SMAD3 MKL1, however, was able to restore the responsiveness of KO cells to TGF-β treatment. More importantly, when MKL1 was re-expressed in SMAD3 is the major sequence-specific transcription factor down- KO cells by retrovirus-mediated infection we observed an induction of stream of TGF-β signaling that plays a key role in cellular fibrogenesis endogenous pro-fibrogenic genes by TGF-β (Fig. 2D). [34,35]. Therefore, we probed the potential interaction between MKL1 We also examined the effect of TGF-β on MKL1 nuclear accumula- and SMAD3 in portal fibroblast cells. Re-ChIP assay showed that upon tion. There was more nuclear MKL1 in the livers of mice that had TGF-β stimulation there was enhanced interaction between MKL1 and 1222 Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228

SMAD3 on the promoters of pro-fibrogenic genes, but not on the gapdh knockdown also abrogated MKL1 binding. Reciprocally, MKL1 depletion promoter (Fig. 3A). Knockdown of SMAD3 impaired the binding of led to a significant down-regulation, though not complete loss, of MKL1 on the pro-fibrogenic promoters (Fig. 3B); as a control, MKL1 SMAD3 binding on the pro-fibrogenic promoters (Fig. 3C); by

Fig. 2. MKL1 mediates TGF-β induced fibrogenesis in portal fibroblast cells. (A, B) Primary rat portal fibroblasts were transfected with siRNA targeting MKL1 (siMkl1) or scrambled siRNA (SCR) followed by treatment with TGF-β (2 ng/ml) for 48 h. Expression of pro-fibrogenic genes was examined by qPCR (A) and Western blotting (B). (C) Indicated promoter constructs were transfected into MKL1-deficient MEF cells with or without MKL1 followed by treatment with TGF-β. Luciferase activities were normalized by both protein concentration and GFP fluorescence. (D) MEF cells were infected with retrovirus carrying either an MKL1 expression construct or an empty vector (EV) followed by treatment with TGF-β. Expression of pro- fibrogenic genes was examined by Western blotting. (E) Primary rat portal fibroblasts were with or without TGF-β (2 ng/ml) for 48 h. ChIP assays were performed with anti-MKL1 or anti-SMAD3 or pre-immune IgG. Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228 1223

Fig. 3. TGF-β promotes the interaction between MKL1 and SMAD3. (A) Primary rat portal fibroblasts were with or without TGF-β (2 ng/ml) for 48 h. Re-ChIP assays were performed with indicated antibodies. (B) Primary rat portal fibroblasts were transfected with siRNA targeting SMAD3 (siSmad3) or scrambled siRNA (SCR) followed by treatment with TGF-β (2 ng/ml) for 48 h. ChIP assay was performed with anti-MKL1 and anti-SMAD3. (C) Primary rat portal fibroblasts were transfected with siRNA targeting MKL1 (siMkl1) or scrambled siRNA (SCR) follow- ed by treatment with TGF-β (2 ng/ml) for 48 h. ChIP assay was performed with anti-MKL1 and anti-SMAD3. 1224 Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228 Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228 1225 comparison, SMAD3 depletion erased its own binding altogether. These tissues, targeting MKL1 might prove helpful in combating excessive fi- data suggest that there is a dynamic interplay between SMAD3 and brosis and loss of tissue function. MKL1 during TGF-β induced fibrogenesis in portal fibroblast cells. Our data indicate that MKL1 deficient mice were protected from BDL-induced liver fibrosis. Furthermore, MKL1 depletion prevented ac- fi β 3.4. TGF-β promotes the recruitment of histone H3K4 methyltransferase tivation of primary portal broblast cells in response to TGF- stimula- β complex tion. These data echo previous reports that MKL1 mediate TGF- triggered pathophysiological events in renal epithelial cells [40], pan- fi Transcriptional activation is invariably associated with an enrich- creatic islet endothelial cells [41], skin broblast cells [42],andlungcar- ment of trimethylated histone H3K4 (H3K4Me3), mediated by the cinoma cells [18], suggesting that MKL1 might serve as an integral part β fi COMPASS methyltransferase complex, on proximal promoters [36,37]. of the TGF- signaling circuit; a whole-genome expression pro ling β Indeed, ChIP assay showed that in response to TGF-β stimulation, may help determine the scope of regulation by MKL1 in TGF- induced fi H3K4Me3 levels were increased on the promoter regions of col1a1, cellular brogenesis. Transcriptome analysis has shown that actin cyto- fi col1a2,andacta2 genes, but not the gapdh gene, in portal fibroblast skeleton re-organization is among the top 15 signi cantly changed β cells (Fig. 4A). In contrast, monomethylated H3K4 (H3K4Me) was not KEGG pathways in response to TGF- [43]. Since MKL1 activity can be altered in these settings (Fig. S3). We also found that several compo- regulated by the dynamic interchange between F-actin and G-actin β nents of the COMPASS methyltransferase complex, including ASH2, [44], it is of high interest to determine whether TGF- induced MKL1 ac- WDR5, and SET1, were up-regulated by TGF-β at both mRNA and pro- tivation as observed here might be a result of cytoskeleton remodeling. tein levels in portal fibroblast cells (Fig. 4B). More importantly, TGF-β On the other hand, MKL1 is universally expressed, which raises an fi promoted the occupancies of the COMPASS proteins on the promoters intriguing question as to whether MKL1 in cells other than portal bro- fi of pro-fibrogenic genes (Fig. 4C). blasts might contribute to liver brosis. For instance, we observed a fi We have previously demonstrated that MKL1-dependent recruit- small but signi cant decrease of liver injury in MKL1 KO mice as mea- ment of COMPASS, in response to lipopolysaccharide (LPS), low oxygen sured by ALT levels (Fig. S1C) suggesting that the down-regulation of fi tension, and angiotensin II respectively, contributes to the pathogenesis liver brosis might be attributable, at least in part, to the amelioration of experimental colitis [38], pulmonary hypertension [39], and cardiac of liver injury. In fact, our previous data suggest that MKL1 participates fl hypertrophy [15,16]. Therefore, we probed whether a similar interplay in the regulation of in ammation in macrophages [38],aprocess fi between MKL1 and COMPASS could take place in portal fibroblast pivotal to liver injury and brosis. Alternatively, MKL1 has been cells in response to TGF-β treatment. Re-ChIP assay showed that the in- shown to transactivate the pro-apoptotic genes Bok and Noxa fi teraction between MKL1 and COMPASS proteins was enhanced on the [45], alluding to the scenario wherein MKL1 de ciency might pro- promoter regions of col1a1, col1a2,andacta2 genes, but not the gapdh mote hepatocyte survival and thus complicating data interpretation fi fi gene, following TGF-β treatment in portal fibroblast cells (Fig. S4). In ad- solely based on a PF-speci croleofMKL1indriving brogenesis. This lin- dition, siRNA-mediated knockdown of MKL1 attenuated the recruit- gering issue will be addressed once a tissue-restricted MKL1 knockout ment of COMPASS (Fig. 4C). Taken together, these data suggest that animal model becomes available. TGF-β promotes the recruitment of histone H3K4 methyltransferase We present evidence that a dynamic interplay between MKL1 and fi complex in a MKL1-dependent manner. SMAD3, the major sequence-speci c transcription factor (TF) for TGF- β induced fibrogenesis, accompanied transactivation of pro-fibrogenic genes. These data are reminiscent of a previously reported interaction fi 3.5. Histone H3K4 methyltransferase complex regulates pro- brogenic between MKL1 and NF-κB in programming cellular pro-inflammatory transcription by altering SMAD3 kinetics transcription in macrophages [38]. MKL1 modulates transcription and by this virtue participates in cellular processes by forming complexes β Finally, we evaluated the role of COMPASS proteins in TGF- in- with sequence-specific TFs. Recent ChIP-seq analyses suggest that fi fi duced brogenesis in portal broblast cells. Depletion of individual MKL1 can bind to DNA motifs recognized by several TF involved in COMPASS proteins using siRNA abrogated the induction of col1a1, liver fibrosis including Sp1, AP-1, SRF, and NF-κB in both NIH3T3 cells col1a2,andacta2 expression at both mRNA (Fig. 5A) and protein [46] and macrophages [47]. Interestingly, neither of these studies re- β (Fig. 5B) levels, suggesting that COMPASS is essential for TGF- to pro- vealed a significant enrichment of MKL1 on SMAD consensus motifs. fi gram cellular brogenesis. Next, we asked whether or not COMPASS One obvious explanation is that the interplay between MKL1 and could function to facilitate the binding of SMAD3 to target promoters SMAD3 is signal-dependent, which means without TGF-β MKL1 cannot β fi as means of promoting TGF- induced brogenesis. To this end, we si- be recruited to SMAD sequences and therefore is unable to influence lenced ASH2, WDR5, and SET1 simultaneously (siCOMPASS) and then SMAD3 activity. Alternatively, SMAD3 might not be the sole or predom- examined the kinetics of SMAD3 binding using ChIP assay. As shown inant TF that MKL1 relies on to mediate TGF-β induced fibrogenesis in Fig. 5C, without COMPASS, SMAD3 bound to target promoters with since AP-1, Sp-1, and SRF have all been shown to contribute to TGF-β fi β much less ef ciency in response to TGF- stimulation. Thus, COMPASS signaling [42,48,49]. Therefore, a context-specific evaluation of MKL1 β fi may contribute to TGF- induced brogenesis by stabilizing SMAD3 binding is likely to clarify the interaction between MKL1 and various binding on gene promoters. TFs in TGF-β induced fibrogenesis. We show here that MKL1-dependent recruitment of COMPASS pro- 4. Discussion teins contribute to TGF-β induced activation of portal fibroblast cells by altering SMAD3 kinetics. This is in agreement with a recent report by In this report, we have investigated the role of MKL1 in cholestatic the Evans laboratory that Vitamin D receptor (VDR) activation antago- liver injury-triggered fibrosis by establishing a relationship between nizes liver fibrosis by evicting SMAD3 from the pro-fibrogenic gene MKL1-dependent recruitment of histone modifying enzymes and TGF- promoters in a chromatin-dependent manner: VDR binding to the β induced activation of portal fibroblast cells. In light of recent findings chromatin renders it deprived of active histone modifications and im- that implicate MKL1 in cellular fibrogenesis in a number of different permissible for SMAD3 binding [50]. Our data re-affirm the notion

Fig. 4. TGF-β promotes the recruitment of histone H3K4 methyltransferase complex. (A) Primary rat portal fibroblasts were with or without TGF-β (2 ng/ml) for 48 h. ChIP assay was per- formed with anti-H3K4Me3. (B) Primary rat portal fibroblasts were with or without TGF-β (2 ng/ml) for 48 h. Expression of COMPASS proteins was examined by qPCR and Western blot- ting. (C) Primary rat portal fibroblasts were transfected with indicated siRNAs followed by treatment with TGF-β (2 ng/ml) for 48 h. ChIP assays were performed with anti-ASH2, anti- WDR5, or anti-SET1. 1226 Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228

Fig. 5. Histone H3K4 methyltransferase complex regulates pro-fibrogenic transcription by altering SMAD3 kinetics. (A, B) Primary rat portal fibroblasts were transfected with indicated siRNAs followed by treatment with TGF-β (2 ng/ml) for 48 h. Expression of pro-fibrogenic genes was examined by qPCR (A) and Western blotting (B). (C) Primary rat portal fibroblasts were transfected with SCR or three individual siRNAs targeting ASH2, WDR5, and SET1 (siCOMPASS) followed by treatment with TGF-β (2 ng/ml). ChIP assays were performed with anti- SMAD3 or anti-MKL1. Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228 1227 that proper chromatin structure is a prerequisite for the recruitment of [5] S.C. Hellerbrand, H. Wang, D.A. Tsukamoto, R.A. Brenner, Rippe, Expression of intra- cellular adhesion molecule 1 by activated hepatic stellate cells, Hepatology 24 TFs and hence transcriptional activation. Moreover, MKL1 binding (1996) 670–676. was also attenuated by COMPASS silencing likely due to insufficient [6] K. Iwaisako, C. Jiang, M. Zhang, M. Cong, T.J. Moore-Morris, T.J. Park, X. Liu, J. Xu, P. SMAD3 binding. Of note, MKL1 can also bring other members of the epi- Wang, Y.H. Paik, F. Meng, M. Asagiri, L.A. Murray, A.F. Hofmann, T. Iida, C.K. Glass, D.A. Brenner, T. Kisseleva, Origin of myofibroblasts in the fibrotic liver in mice, genetic machinery, including histone acetyltransferase [13], histone Proc. Natl. Acad. Sci. U. S. A. 111 (2014) E3297–E3305. demethylase [51], and chromatin remodeling proteins [52] to regulate [7] H. Clouzeau-Girard, C. Guyot, C. Combe, V. Moronvalle-Halley, C. Housset, T. transcription. Importantly, MKL1 has been shown to moderate the Lamireau, J. Rosenbaum, A. Desmouliere, Effects of bile acids on biliary epithelial fi crosstalk between different branches of the epigenetic machinery [16, cell proliferation and portal broblast activation using rat liver slices, Lab. Investig. 86 (2006) 275–285. 17]. It has been argued that epigenetic factors (histone methyltransfer- [8] Z. Li, J.A. Dranoff, E.P. Chan, M. Uemura, J. Sevigny, R.G. Wells, Transforming growth ase in this case) show no preference in terms of sequence recognition factor-beta and substrate stiffness regulate portal fibroblast activation in culture, – [53,54]. Instead, they rely on sequence-specific TFs to be recruited to Hepatology 46 (2007) 1246 1256. fi fl [9] E. Borkham-Kamphorst, C. Schaffrath, E. Van de Leur, U. Haas, L. Tihaa, S.K. Meurer, speci c gene regulatory motifs to in uence transcription. Once recruit- Y.A. Nevzorova, C. Liedtke, R. Weiskirchen, The anti-fibrotic effects of CCN1/CYR61 ed, they may reciprocally influence TF binding by fine-tuning the chro- in primary portal myofibroblasts are mediated through induction of reactive oxygen matin structure. Based on this notion and on our own observation, we species resulting in cellular senescence, apoptosis and attenuated TGF-beta signal- – β ing, Biochim. Biophys. Acta 1843 (2014) 902 914. propose that TGF- stimulation prompts SMAD3 binding (likely with [10] D.Z. Wang, S. Li, D. Hockemeyer, L. Sutherland, Z. Wang, G. Schratt, J.A. Richardson, low affinity initially) to its consensus sequence, which is followed by A. Nordheim, E.N. Olson, Potentiation of activity by a family MKL1 recruitment. MKL1 in turn acts a bridge bringing down of myocardin-related transcription factors, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 14855–14860. COMPASS to modify chromatin structure to stabilize SMAD3 binding [11] E.M. Small, J.E. Thatcher, L.B. Sutherland, H. Kinoshita, R.D. Gerard, J.A. Richardson, and hence TGF-β dependent gene expression. Future studies should J.M. Dimaio, H. Sadek, K. Kuwahara, E.N. Olson, Myocardin-related transcription focus on examining, on a genome-wide scale, how MKL1 controls TF factor-a controls myofibroblast activation and fibrosis in response to myocardial infarction, Circ. Res. 107 (2010) 294–304. binding by altering the chromatin structure. [12] L.L. Luchsinger, C.A. Patenaude, B.D. Smith, M.D. Layne, Myocardin-related transcrip- In summary, our data suggest that MKL1 is an epigenetic modulator tion factor-A complexes activate type I collagen expression in lung fibroblasts, J. Biol. of TGF-β induced fibrogenesis. Specifically, our data reveal for the first Chem. 286 (2011) 44116–44125. time that the kinetics of SMAD3 binding on its target promoters and [13] H. Xu, X. Wu, H. Qin, W. Tian, J. Chen, L. Sun, M. Fang, Y. Xu, Myocardin-related transcription factor A epigenetically regulates renal fibrosis in diabetic nephropathy, consequently TGF-β stimulated cellular fibrogenesis is modulated by J. Am. Soc. Nephrol. 26 (2015) 1648–1660. H3K4 methylation in a MKL1-dependent manner. The ability of MKL1 [14] F. Fang, Y. Yang, Z. Yuan, Y. Gao, J. Zhou, Q. Chen, Y. Xu, Myocardin-related transcrip- to modulate liver fibrogenesis seems to be a function of its interaction tion factor A mediates OxLDL-induced endothelial injury, Circ. Res. 108 (2011) 797–807. with the histone methyltransferase complex, which alludes to the [15] X. Weng, L. Yu, P. Liang, D. Chen, X. Cheng, Y. Yang, L. Li, T. Zhang, B. Zhou, X. Wu, H. possibility of using small-molecule compounds to modulate COMPASS Xu, M. Fang, Y. Gao, Q. Chen, Y. Xu, Endothelial MRTF-A mediates angiotensin II activity as means of alleviating liver fibrosis. induced cardiac hypertrophy, J. Mol. Cell. Cardiol. 80 (2015) 23–33. [16] X. Weng, L. Yu, P. Liang, L. Li, X. Dai, B. Zhou, X. Wu, H. Xu, M. Fang, Q. Chen, Y. Xu, A crosstalk between chromatin remodeling and histone H3K4 methyltransferase com- Conflict of interest plexes in endothelial cells regulates angiotensin II-induced cardiac hypertrophy, J. Mol. Cell. Cardiol. 82 (2015) 48–58. [17] Y. Yang, X. Cheng, W. Tian, B. Zhou, X. Wu, H. Xu, F. Fang, M. Fang, Y. Xu, MRTF-A None. steers an epigenetic complex to activate endothelin-induced pro-inflammatory transcription in vascular smooth muscle cells, Nucleic Acids Res. 42 (2014) 10460–10472. Transparency document [18] X. Cheng, Y. Yang, Z. Fan, L. Yu, H. Bai, B. Zhou, X. Wu, H. Xu, M. Fang, A. Shen, Q. Chen, Y. Xu, MKL1 potentiates lung cancer cell migration and invasion by epigenet- The Transparency document associated with this article can be ically activating MMP9 transcription, Oncogene (2015)http://dx.doi.org/10.1038/ onc.2015.14. found, in the online version. [19] Y. Sun, K. Boyd, W. Xu, J. Ma, C.W. Jackson, A. Fu, J.M. Shillingford, G.W. Robinson, L. Hennighausen, J.K. Hitzler, Z. Ma, S.W. Morris, Acute myeloid leukemia-associated Mkl1 (Mrtf-a) is a key regulator of mammary gland function, Mol. Cell. Biol. 26 Acknowledgments (2006) 5809–5826. [20] J.W. Wen, A.L. Olsen, M. Perepelyuk, R.G. Wells, Isolation of rat portal fibroblasts by This work is supported in part by the National Basic Research "973" in situ liver perfusion, J. Vis. Exp. 29 (2012) (pii: 3669). [21] M. Fausther, J.R. Goree, E.G. Lavoie, A.L. Graham, J. Sevigny, J.A. Dranoff, Establish- Program of China (2012CB517503), the Natural Science Foundation of ment and characterization of rat portal myofibroblast cell lines, PLoS One 10 China (31270805, 31200645, 81402550, 91439106), the Natural Sci- (2015) e0121161. ence Foundation of Jiangsu Province (BK20140906, BK20141098), and [22] W. Tian, C. Hao, Z. Fan, X. Weng, H. Qin, X. Wu, M. Fang, Q. Chen, A. Shen, Y. Xu, Myocardin related transcription factor A programs epigenetic activation of hepatic the Program for New Century Excellent Talents in University of China stellate cells, J. Hepatol. 62 (2015) 165–174. (NCET-13-0989). YX is a Fellow at the Innovative Collaboration Center [23] H. Xu, X. Wu, H. Qin, W. Tian, J. Chen, L. Sun, M. Fang, Y. Xu, Myocardin-related for Cardiovascular Translational Medicine of Jiangsu Province. transcription factor A epigenetically regulates renal fibrosis in diabetic nephropathy, J. Am. Soc. Nephrol. 26 (2015) 1648–1660. [24] B. Cen, A. Selvaraj, R.C. Burgess, J.K. Hitzler, Z. Ma, S.W. Morris, R. Prywes, Appendix A. Supplementary data Megakaryoblastic leukemia 1, a potent transcriptional coactivator for serum response factor (SRF), is required for serum induction of SRF target genes, Mol. Cell. Biol. 23 (2003) 6597–6608. Supplementary data to this article can be found online at http://dx. [25] T. Zhang, J.G. Berrocal, K.M. Frizzell, M.J. Gamble, M.E. DuMond, R. Krishnakumar, T. doi.org/10.1016/j.bbagrm.2015.07.013. Yang, A.A. Sauve, W.L. Kraus, Enzymes in the NAD+ salvage pathway regulate SIRT1 activity at target gene promoters, J. Biol. Chem. 284 (2009) 20408–20417. [26] J. Xia, M. Fang, X. Wu, Y. Yang, L. Yu, H. Xu, H. Kong, Q. Tan, H. Wang, W. Xie, Y. Xu, References A2b adenosine signaling represses CIITA transcription via an epigenetic mechanism in vascular smooth muscle cells, Biochim. Biophys. Acta 1849 (2015) 665–676. [1] D.Y. Zhang, S.L. Friedman, Fibrosis-dependent mechanisms of hepatocarcinogenesis, [27] W. Tian, H. Xu, F. Fang, Q. Chen, Y. Xu, A. Shen, Brahma-related gene 1 bridges Hepatology 56 (2012) 769–775. epigenetic regulation of proinflammatory cytokine production to steatohepatitis [2] S.L. Friedman, Stellate cells: a moving target in hepatic fibrogenesis, Hepatology 40 in mice, Hepatology 58 (2013) 576–588. (2004) 1041–1043. [28] L. Sun, H. Li, J. Chen, Y. Iwasaki, T. Kubota, M. Matsuoka, A. Shen, Q. Chen, Y. Xu, [3] B.L. Copple, Hypoxia stimulates hepatocyte epithelial to mesenchymal transition by PIASy mediates hypoxia-induced SIRT1 transcriptional repression and epithelial- hypoxia-inducible factor and transforming growth factor-beta-dependent mecha- to-mesenchymal transition in ovarian cancer cells, J. Cell Sci. 126 (2013) nisms, Liver Int. 30 (2010) 669–682. 3939–3947. [4] A.S. Chu, R. Diaz, J.J. Hui, K. Yanger, Y. Zong, G. Alpini, B.Z. Stanger, R.G. Wells, Lineage [29] F. Fang, D. Chen, L. Yu, X. Dai, Y. Yang, W. Tian, X. Cheng, H. Xu, X. Weng, M. Fang, J. tracing demonstrates no evidence of cholangiocyte epithelial-to-mesenchymal tran- Zhou,Y.Gao,Q.Chen,Y.Xu,Proinflammatory stimuli engage brahma related gene 1 sition in murine models of hepatic fibrosis, Hepatology 53 (2011) 1685–1695. and brahma in endothelial injury, Circ. Res. 113 (2013) 986–996. 1228 Z. Fan et al. / Biochimica et Biophysica Acta 1849 (2015) 1219–1228

[30] L. Sun, H. Li, J. Chen, V. Dehennaut, Y. Zhao, Y. Yang, Y. Iwasaki, B. Kahn-Perles, D. molecule promotes myofibroblast differentiation and wound healing, Proc. Natl. Leprince, Q. Chen, A. Shen, Y. Xu, A SUMOylation-dependent pathway regulates Acad. Sci. U. S. A. 110 (2013) 16850–16855. SIRT1 transcription and lung cancer metastasis, J. Natl. Cancer Inst. 105 (2013) [43] E.P. Brennan, M.J. Morine, D.W. Walsh, S.A. Roxburgh, M.T. Lindenmeyer, D.P. Brazil, 887–898. P.O. Gaora, H.M. Roche, D.M. Sadlier, C.D. Cohen, C. Godson, F. Martin, Next- [31] J. Napoli, D. Prentice, C. Niinami, G.A. Bishop, P. Desmond, G.W. McCaughan, Sequen- generation sequencing identifies TGF-beta1-associated gene expression profiles in tial increases in the intrahepatic expression of epidermal growth factor, basic renal epithelial cells reiterated in human diabetic nephropathy, Biochim. Biophys. fibroblast growth factor, and transforming growth factor beta in a bile duct ligated Acta 1822 (2012) 589–599. rat model of cirrhosis, Hepatology 26 (1997) 624–633. [44] E.N. Olson, A. Nordheim, Linking actin dynamics and gene transcription to drive [32] D.M. Bissell, D. Roulot, J. George, Transforming growth factor beta and the liver, cellular motile functions, Nat. Rev. Mol. Cell Biol. 11 (2010) 353–365. Hepatology 34 (2001) 859–867. [45] D. Shaposhnikov, A. Descot, J. Schilling, G. Posern, Myocardin-related transcription [33] Z. Qi, N. Atsuchi, A. Ooshima, A. Takeshita, H. Ueno, Blockade of type beta factor A regulates expression of Bok and Noxa and is involved in apoptotic transforming growth factor signaling prevents liver fibrosis and dysfunction in the signalling, Cell Cycle 11 (2012) 141–150. rat, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 2345–2349. [46] C. Esnault, A. Stewart, F. Gualdrini, P. East, S. Horswell, N. Matthews, R. Treisman, [34] B. Schnabl, Y.O. Kweon, J.P. Frederick, X.F. Wang, R.A. Rippe, D.A. Brenner, The role of Rho-actin signaling to the MRTF coactivators dominates the immediate transcrip- Smad3 in mediating mouse hepatic stellate cell activation, Hepatology 34 (2001) tional response to serum in fibroblasts, Genes Dev. 28 (2014) 943–958. 89–100. [47] L. Xie, MKL1/2 and ELK4 co-regulate distinct serum response factor (SRF) transcrip- [35] H. Weng, P.R. Mertens, A.M. Gressner, S. Dooley, IFN-gamma abrogates tion programs in macrophages, BMC Genomics 15 (2014) 301. profibrogenic TGF-beta signaling in liver by targeting expression of inhibitory and [48] M. Wygrecka, D. Zakrzewicz, B. Taborski, M. Didiasova, G. Kwapiszewska, K.T. receptor Smads, J. Hepatol. 46 (2007) 295–303. Preissner, P. Markart, TGF-beta1 induces tissue factor expression in human lung [36] H. Santos-Rosa, R. Schneider, A.J. Bannister, J. Sherriff, B.E. Bernstein, N.C. Emre, S.L. fibroblasts in a PI3K/JNK/Akt-dependent and AP-1-dependent manner, Am. J. Schreiber, J. Mellor, T. Kouzarides, Active genes are tri-methylated at K4 of histone Respir. Cell Mol. Biol. 47 (2012) 614–627. H3, Nature 419 (2002) 407–411. [49] G. Cordova, A. Rochard, C. Riquelme-Guzman, C. Cofre, D. Scherman, P. Bigey, E. [37] A. Shilatifard, The COMPASS family of histone H3K4 methylases: mechanisms of Brandan, SMAD3 and SP1/SP3 transcription factors collaborate to regulate connec- regulation in development and disease pathogenesis, Annu. Rev. Biochem. 81 tive tissue growth factor gene expression in myoblasts in response to transforming (2012) 65–95. growth factor beta, J. Cell. Biochem. 116 (2015) 1880–1887. [38] L. Yu, X. Weng, P. Liang, X. Dai, X. Wu, H. Xu, M. Fang, F. Fang, Y. Xu, MRTF-A medi- [50] N. Ding, R.T. Yu, N. Subramaniam, M.H. Sherman, C. Wilson, R. Rao, M. Leblanc, S. ates LPS-induced pro-inflammatory transcription by interacting with the COMPASS Coulter, M. He, C. Scott, S.L. Lau, A.R. Atkins, G.D. Barish, J.E. Gunton, C. Liddle, M. complex, J. Cell Sci. 127 (2014) 4645–4657. Downes, R.M. Evans, A vitamin D receptor/SMAD genomic circuit gates hepatic [39] D. Chen, Y. Yang, X. Cheng, F. Fang, G. Xu, Z. Yuan, J. Xia, H. Kong, W. Xie, H. Wang, M. fibrotic response, Cell 153 (2013) 601–613. Fang, Y. Gao, Y. Xu, Megakaryocytic leukemia 1 directs a histone H3 lysine 4 [51] K. Lockman, J.M. Taylor, C.P. Mack, The histone demethylase, Jmjd1a, interacts with methyltransferase complex to regulate hypoxic pulmonary hypertension, Hyperten- the myocardin factors to regulate SMC differentiation marker gene expression, Circ. sion 65 (2015) 821–833. Res. 101 (2007) e115–e123. [40] T. Morita, T. Mayanagi, K. Sobue, Dual roles of myocardin-related transcription fac- [52] J. Zhou, M. Zhang, H. Fang, O. El-Mounayri, J.M. Rodenberg, A.N. Imbalzano, B.P. tors in epithelial mesenchymal transition via slug induction and actin remodeling, Herring, The SWI/SNF chromatin remodeling complex regulates myocardin- J. Cell Biol. 179 (2007) 1027–1042. induced smooth muscle-specific gene expression, Arterioscler. Thromb. Vasc. Biol. [41] H. Mihira, H.I. Suzuki, Y. Akatsu, Y. Yoshimatsu, T. Igarashi, K. Miyazono, T. Watabe, 29 (2009) 921–928. TGF-beta-induced mesenchymal transition of MS-1 endothelial cells requires Smad- [53] M. Ptashne, On the use of the word ‘epigenetic’, Curr. Biol. 17 (2007) R233–R236. dependent cooperative activation of Rho signals and MRTF-A, J. Biochem. 151 [54] M. Ptashne, Epigenetics: core misconcept, Proc. Natl. Acad. Sci. U. S. A. 110 (2013) (2012) 145–156. 7101–7103. [42] L.S. Velasquez, L.B. Sutherland, Z. Liu, F. Grinnell, K.E. Kamm, J.W. Schneider, E.N. Olson, E.M. Small, Activation of MRTF-A-dependent gene expression with a small