Synergistic Roles of Scleraxis and Smads in the Regulation of Collagen 1Α2 Gene Expression

Home , Scleraxis

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1823 (2012) 1936–1944

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

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

Synergistic roles of scleraxis and Smads in the regulation of collagen 1α2 gene expression

Rushita A. Bagchi, Michael P. Czubryt ⁎

Institute of Cardiovascular Sciences, St. Boniface General Hospital Research Centre, University of Manitoba, Winnipeg, Manitoba, Canada

article info abstract

Article history: Cardiac fibrosis is marked by increased deposition of extracellular matrix components including fibrillar col- Received 15 March 2012 lagens, leading to impaired cardiac contractility and function. We recently demonstrated that the transcrip- Received in revised form 2 July 2012 tion factor scleraxis is expressed in collagen-producing cardiac fibroblasts and myofibroblasts, is up- Accepted 3 July 2012 regulated in the collagen-rich scar following myocardial infarction and is sufficient to transactivate the Available online 13 July 2012 human collagen 1α2 (COL1A2) gene, suggesting a central role in fibrosis. Here we describe the mechanism of scleraxis-mediated regulation of the COL1A2 promoter. Using chromatin immunoprecipitation in primary Keywords: fi Gene expression human cardiac broblasts in combination with luciferase assays, we demonstrate that two E box sequences Collagen within the proximal COL1A2 promoter are required for scleraxis-mediated transactivation. Expression of bHLH transcription factor scleraxis itself was induced by receptor Smad3, an effector of the pro-fibrotic growth factor TGF-β1, and at- fi Cardiac broblast tenuated by inhibitory Smad7. TGF-β1 augmented the effect of scleraxis on COL1A2 transactivation, an effect Fibrosis which was due to synergy between scleraxis and Smad3. Mutation of the COL1A2 Smad-binding element significantly attenuated the ability of scleraxis to transactivate the promoter, while mutation of the scleraxis- interacting E boxes attenuated the effect of Smad3, suggesting that these factors form a common signaling complex at the promoter. COL1A2 promoter transactivation and Col1α2 gene expression in cardiac fibroblasts were completely abrogated by a scleraxis basic domain deletion mutant in a dominant negative fash-

ion, blocking the ability of TGF-β1 to activate collagen synthesis and suggesting that scleraxis–DNA interaction is absolutely required for this process. Scleraxis thus appears to play a key role in the transcriptional regulation of type I collagen synthesis. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Type I collagen is generated from two strands of procollagen 1α1and one strand of 1α2, which are secreted from cardiac fibroblasts and Cardiac fibrosis is the relative increase in extracellular matrix (ECM) myofibroblasts and undergo further processing in the extracellular mi- synthesis that is a common pathological response to a wide variety of lieu. Transcriptional regulation of procollagen mRNA synthesis represents stresses, including hypertension, valve disease, diabetes and myocardial a key point of control for managing matrix generation and fibrosis. A pri- infarction. Fibrosis interferes with both systolic and diastolic functions, mary regulator of this process is TGF-β, which acts via a heterodimeric contributing to cardiac decompensation and failure, and increasing pa- membrane-bound receptor to activate the Smad intracellular signaling tient morbidity and mortality [1,2]. Treatments aimed specifically at re- pathway. Following receptor-mediated phosphorylation, Smads migrate ducing fibrosis thus have the potential to ameliorate the progression of to the nucleus and directly transactivate ECM genes such as collagen failure, but success in this area has been limited. For example, targeting 1α2 [5,6]. Other transcriptional regulators including Sp1 and Sp3 also of the pro-fibrotic TGF-β pathway has failed to yield a tractable anti- contribute to collagen gene activity, and it is likely that still other players fibrotic therapy, despite the central role of this pathway in driving fibro- are involved [7,8]. sis [3]. We recently demonstrated that the basic helix–loop–helix (bHLH) A key component of the ECM is type I collagen, representing over 40% transcription factor scleraxis was sufficient to up-regulate expression of the fibrillar collagen present in both healthy and diseased hearts [4]. of the collagen 1α2 gene in primary cardiac fibroblasts [9]. Scleraxis expression increased in cardiac myofibroblasts – activated forms of fibro- Abbreviations: bHLH, basic helix–loop–helix; ECM, extracellular matrix; EMSA, blasts that synthesize elevated levels of collagen – or in response to

electrophoretic mobility shift assay; qPCR, quantitative real-time polymerase chain TGF-β1 treatment. Scleraxis expression was also elevated in the reaction collagen-rich scar following infarction. Others have reported that ⁎ Corresponding author at: Institute of Cardiovascular Sciences, St. Boniface General scleraxis regulates expression of the collagen 1α1 gene in tenocytes, Hospital Research Centre, University of Manitoba, 351 Tache Avenue, Winnipeg, Manitoba, Canada R2H 2A6. Tel.: +1 204 235 3719; fax: +1 204 231 1151. and scleraxis gene deletion results in defects in the development of E-mail address: [email protected] (M.P. Czubryt). intermuscular and force-transmitting tendons concomitant with type I

0167-4889/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamcr.2012.07.002 R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944 1937 collagen loss [10,11]. Mounting evidence thus supports our hypothesis with 75 μl bovine serum albumin (0.5 mg/ml)/herring sperm DNA that scleraxis is a critical regulator of collagen gene expression in the (200 μg/ml)/protein A-agarose mixture for 1 hour with rotation at heart, either independently or downstream of TGF-β1. In the current 4°Ctoreducenon-specific background. Lysates with soluble chromatin study, we found that scleraxis expression was up-regulated following collected after brief centrifugation were incubated at 4 °C overnight over-expression of pro-fibrotic Smad3, and was down-regulated in re- with 10 μg anti-scleraxis antibody or 10 μg pre-immune serum as a sponse to over-expression of inhibitory Smad7. negative control. The production of this antibody by our laboratory bHLH transcription factors including scleraxis regulate gene expres- was previously described [9]. The DNA-protein immune complexes sion via interaction with defined E box sequences (CANNTG, where N is were precipitated, washed, eluted and subsequently reverse cross- any nucleotide) [12,13]. We previously showed that scleraxis regulates linked, then purified by phenol/chloroform extraction. DNA samples transactivation of the 3.7 kb human collagen 1α2 (COL1A2) gene pro- were subjected to PCR amplification using primers specifictothe moter [9]. Deletion of the scleraxis DNA-binding domain (ScxΔBD) at- scleraxis-binding E box sequences (Table 1) in the human COL1A2 tenuated this transactivation, indicating that scleraxis must physically gene proximal promoter. Primers encompassing a region of the human contact the promoter. Here we report the identification of two critical GAPDH promoter were used as a negative control (Table 1;Upstate E boxes within the COL1A2 proximal promoter that are required for Cell Signaling Solutions EZ ChIP kit). scleraxis-mediated promoter transactivation, determined by luciferase assay and chromatin immunoprecipitation (ChIP). 2.3. Luciferase reporter assays These E boxes proximally flank a Smad binding element (SBE) previously shown to be critical for TGF-β induction of COL1A2 expression, NIH-3T3 fibroblasts were seeded in 6-well dishes 24 h prior to suggesting the possibility of co-regulation [14]. Using luciferase report- transfection to reach ~80% confluency. Samples were co-transfected er assays, we found that Smad3 synergistically up-regulated COL1A2 (Lipofectamine 2000; Invitrogen, USA) with 500 ng reporter plasmid expression with scleraxis, while Smad7 was inhibitory and Smad2 had (COL1A2 full-length 3.7 kb or 0.7 kb proximal promoters, or empty a modest effect. Mutation of the SBE attenuated scleraxis-mediated pGL3 Basic as a control) plus 500 ng expression vectors as needed transactivation of the COL1A2 promoter, while mutation of the two E (pECE-HA-FLAG-Scx, pECE-ScxΔBD, pCMV tag3C-Myc-Smad3, pCMV boxes attenuated the effect of Smad3, suggesting that cooperation be- tag3C-Myc-Smad2, pcDNA3.1(+)-FLAG-Smad7). Mutation of the E tween these factors is required to support full promoter activation. boxes and SBE within the COL1A2 0.7 kb promoter was performed Because the DNA-binding domain of scleraxis was absolutely required using a QuikChange II Site-Directed Mutagenesis kit (Stratagene) and for activity, we hypothesized that the DNA-binding mutant ScxΔBD may nested PCR using primers listed in Table 2. Renilla luciferase expression inhibit collagen 1α2 expression in a dominant negative fashion [9].Lucif- vector (pRL) was used as a transfection control. The luciferase activity of erase assays demonstrated that this mutant does, in fact, attenuate the each sample was assayed using the Dual Luciferase Reporter Assay Sys- ability of intact scleraxis to regulate COL1A2 transactivation. More impor- tem (Promega, USA) on a TD20/20 luminometer (Turner BioSystems, tantly, ScxΔBD inhibited collagen 1α2expressioninbothNIH-3T3 USA). fibroblasts and primary rat cardiac myofibroblasts, and made them resistant to TGF-β1-mediated collagen up-regulation. Our results indicate that scleraxis and Smads downstream of TGF-β1 work together to medi- 2.4. Treatment of cells ate type I collagen expression, and suggest that interference with scleraxis function may provide a new approach to the development of NIH-3T3 fibroblasts were grown to ~70% confluency, then co- anti-fibrotic therapies. transfected with COL1A2 0.7 kb reporter plasmid and pECE-HA-FLAG- Scx expression vector for luciferase assays, or with pECE-HA-FLAG-Scx 2. Materials and methods or pECE-ScxΔBD expression vectors alone for mRNA analysis. Cells

were treated with 10 ng/ml TGF-β1 or vehicle for 24 h post- 2.1. Cell culture transfection where required. Cell lysates were collected for luciferase assay or qRT-PCR. Primary rat cardiac myofibroblasts (passage 2) NIH-3T3 mouse embryonic fibroblasts were cultured in DMEM were grown in 6-well tissue culture dishes to ~70% confluence in com- supplemented with 10% FBS, 1% penicillin/streptomycin and 1% plete medium followed by equilibration for 24 h in serum-free medium. L-glutamine, and were seeded 24 h prior to transfection in 6-well cul- Cells were infected with adenoviruses encoding LacZ (AdLacZ), FLAG- ture plates to reach ~70% confluence. Adult human cardiac fibroblasts Smad3 (AdSmad3), FLAG-Smad7 (AdSmad7), FLAG-scleraxis, FLAG- were obtained commercially (Cell Applications Inc., USA). Primary ScxΔBD or LacZ (control) for 24 h. Serum-free medium was replaced cardiac fibroblasts were isolated from adult rats by Langendorff perfu- and cells were then treated with 10 ng/ml TGF-β1 or vehicle as re- sion of 0.1% collagenase type II (Worthington Biochemicals, USA). quired, and were harvested for protein or RNA 24 h later. Total protein Cells were collected after gentle centrifugation as previously described and total RNA was used for western blotting and qPCR analysis respec- [15] and maintained in DMEM-F12 supplemented with 10% FBS, 1% tively. The scleraxis adenovirus was described previously [9].Togener- penicillin/streptomycin and 1 mM L-ascorbic acid. Preparations of fibro- ate the ScxΔBD adenovirus, we used PCR to amplify and sub-clone the blasts obtained through this method are typically ≥95% as confirmed by ScxΔBD coding region into pSHUTTLE-CMV transfer vector from the cell-specific marker analysis [16]. Animals were cared for in accordance AdEasy Vector System (Qbiogene, Canada). The shuttle vector was with the guidelines of the Canadian Council on Animal Care and the then cloned into BJ5183 bacterial cells to generate replication- University of Manitoba Animal Protocol Management and Review deficient adenoviruses. Committee. Table 1 2.2. Chromatin immunoprecipitation assay (ChIP) ChIP assay PCR primers. Primer Direction Sequence (5′→3′) Adult human cardiac fibroblasts (~8×106 cells) were grown in fi μ E boxes 1 & 2 Forward CTTCCCTCCCTCCTTCTCTG 100 mm dishes and xed by addition of 280 l 37% formaldehyde to Reverse ACTATGGGACACGTCGGAGT 10 ml of culture media for 10 min at 37 °C. The cells were then washed E box 3 Forward CAAAAGTGTCCACGTCCTCA with ice-cold PBS and harvested in SDS lysis buffer. Cell lysates were Reverse CCCGGATCTGCCCTATTTAT sonicated to shear DNA to 200–500 bp fragments. A 1% aliquot of this GAPDH Forward TACTAGCGGTTTTACGGGCG Reverse TCGAACAGGAGGAGCAGAGAGCGA cell lysate was used as an input control. Samples were pre-cleared 1938 R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944

Table 2 COL1α2 proximal promoter cloning and mutagenesis PCR primers.

Primer Direction Sequence (5′→3′)

COL1α2—0.7 kb Forward GCCCTTTCCCGAAGTCATAAGAC Reverse TCGAGCCCGGGCTAG ΔE1 mutagenesis Forward GAGCCCTCCACCCTACCATGGGCCTACAGGGCACAG Reverse CTGTGCCCTGTAGGCCCATGGTAGGGTGGAGGGCTC ΔE2 mutagenesis Forward GGCCTACAGGGCACTGCAGAGGCGGGACTGGA Reverse TCCAGTCCCGCCTCTGCAGTGCCCTGTAGGCC ΔE3 mutagenesis Forward CGCAGGCTCCTCCCCCGGGTGGCTGCCCGGGC Reverse GCCCGGGCAGCCACCCGGGGGAGGAGCCTGCG ΔSBE mutagenesis Forward ATGCGGAGGGCGGAGGTATGCGCACAACGAGTCAGAGTTTCCCCTT Reverse AAGGGGAAACTCTGACTCGTTGTGCGCATACCTCCGCCCTCCGCAT

Underlined sequences denote mutated E boxes or SBE in the COL1A2 proximal promoter.

2.5. Quantitative real‐time-PCR (qRT-PCR) reactions, 1 μg anti-scleraxis antibody was incubated overnight at 4 °C prior to addition of the biotin-labeled probes to the reaction [9]. Total RNA was isolated from cells using commercial kits (Sigma, DNA-protein complexes were then fractionated on a non-denaturing USA) according to manufacturer's instructions. 25 ng RNA was used 6% native polyacrylamide gel at 100 V for 1 h. Chemiluminescent com- for each of the reactions performed using the qScript One-Step SYBR plexes were detected using CL-XPosure clear blue film (Thermo Scientif- Green qRT-PCR Kit for iQ (Quanta Biosciences, USA) on an iQ5 ic, Canada). multicolor real-time PCR thermocycler (Bio-Rad, USA). mRNA abun- −ΔΔCT dance was calculated using the 2 method and was normalized 2.8. Statistical analysis to GAPDH. Specific primers used to amplify sequences of interest are listed in Table 3. Statistical significance of data was determined using Student t-test or one-way analysis of variance (ANOVA) with Student-Newman– 2.6. Western blotting Keuls post-hoc analysis. Results with pb0.05 were considered statistically significant. Total cell protein was isolated using RIPA buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.5% Na-deoxycholate, 1% 3. Results Triton-X 100 and 0.1% SDS) containing complete mini protease inhibitor cocktail (Roche, Canada). Proteins were separated by SDS-PAGE and 3.1. Regulation of scleraxis expression by Smads then transferred to PVDF membranes (Pall, USA). Primary antibodies used for western blotting were rabbit anti-scleraxis antibody [9], We previously reported that scleraxis expression is up-regulated in mouse anti-α-tubulin antibody (DSHB, USA) or rat anti-collagen type cardiac fibroblasts by TGF-β , and in turn regulates the COL1A2 promot- I antibody (Cedarlane, Canada). Antibodies were detected using West- 1 er [9]. Because Smads are the primary mediator of TGF-β signals in fi- ern Blotting Luminol Reagent (Santa Cruz Biotechnology, USA) and 1 brosis, we assayed the effect of pro-fibrotic Smad3 or inhibitory CL-Xposure blue X-ray film (Thermo Scientific, Canada). Quantity One Smad7 on scleraxis expression. Adenoviruses encoding Smad3 or software (Bio-Rad, USA) was used to measure relative band intensity. Smad7 were used to transfect primary rat cardiac fibroblasts. Smad3 The intensity of the scleraxis bands was normalized to those of over-expression significantly up-regulated scleraxis expression, while α-tubulin to control for loading. Smad7 attenuated scleraxis expression (Fig. 1). These changes were observed at the level of both transcription and translation, and demon- 2.7. Electrophoretic mobility shift assay (EMSA) strate that Smad signaling can regulate scleraxis expression. COS7 cells were transfected with 1 μg scleraxis expression vector, and 48 h later nuclear proteins were isolated using NE-PER nuclear and 3.2. Scleraxis transactivation of the proximal COL1α2 gene promoter cytoplasmic extraction reagents (Pierce Biotechnology, USA). 5 μlnucle- ar extract was used in binding reactions as required. Assays were Scleraxis binds specifically to E boxes similar to other bHLH tran- performed using a Lightshift Chemiluminescent EMSA kit (Pierce Bio- scription factors [12,17]. We previously noted the existence of up to technology, USA) as per manufacturer's instructions. Binding reactions 12 putative E boxes in the human 3.7 kb COL1A2 gene promoter [9]. were incubated at room temperature for 30 min in buffer containing Three of these sites are located within the proximal COL1A2 gene pro- β 10 mM Tris pH 7.5, 100 mM KCl, 0.5 mM EDTA, 5% glycerol, 50 ng moter (~700 bp), a region that is highly responsive to TGF- 1 signaling poly(dI-dC), 5 μg bovine serum albumin and 20 fmol biotin end- and rich in binding sites for transcription factors that regulate collagen labeled probe. For competition experiments, a 250-fold molar excess synthesis including Sp1, Sp3 and Smads [7,8]. To test whether this prox- (5 pmol) of unlabeled probe was included in the binding reactions and imal region could be transactivated by scleraxis, we amplified by incubated for 15 min before addition of the labeled probes. For supershift high-fidelity PCR a sequence comprising 706 bp of the COL1A2 proximal promoter, including 54 bp of 5′ untranslated region. This sequence was sub-cloned into the luciferase reporter vector pGL3 Basic for further Table 3 study. This sequence encompasses three putative E boxes, in contrast to qRT-PCR primers [9]. previously-reported COL1A2 proximal promoter constructs which lack ′ Amplicon Direction Sequence (5′→3′) the two most 5 distal E boxes [18]. We examined the relative transactivation by scleraxis of the proxi- Scleraxis Forward AACACGGCCTTCACTGCGCTG Reverse CAGTAGCACGTTGCCCAGGTG mal COL1A2 promoter (COL1A2pr-0.7) compared to the full-length Col1α2 Forward GTCCCCGAGGCAGAGAT 3.7 kb promoter (COL1A2pr-3.7). Both promoters were strongly and Reverse CCTTTGTCAGAATACTGAGCAGC significantly transactivated by scleraxis, but the proximal promoter GAPDH Forward TGCACCACCAACTGCTTAGC exhibited a nearly three-fold higher level of activity compared to the Reverse GGCATGGACTGTGGTCATGAG full-length promoter (Fig. 2A). We also compared transactivation of R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944 1939

A results suggest that the full-length promoter may contain repressor el- 2.5 ements that negatively affect transactivation. These results also suggest that transactivation of the COL1A2 gene by scleraxis may be largely de- Ad-LacZ * termined by sequences within the proximal promoter. 2.0 Ad-Smad 3.3. E box-mediated transactivation of the COL1A2 proximal promoter by 1.5 scleraxis

As noted above, the COL1A2 proximal promoter contains three puta- 1.0 tive scleraxis-binding E boxes (designated E1 to E3; Fig. 3A), including one sequence (E2) that exactly matches a previously-deﬁned scleraxis binding site in the aggrecan gene promoter [19]. Homologous E boxes 0.5 * are located in the mouse and rat Col1α2 promoters in positions closely

Scleraxis mRNA expression (fold) proximal to their locations in the human promoter (Fig. 3A). Chromatin 0 immunoprecipation assay conﬁrmed that scleraxis directly binds to the Smad3 Smad7 promoter regions encompassing the E box sequences present within the COL1A2 promoter in normal human cardiac ﬁbroblasts ex vivo B (Fig. 3B). However, ChIP resolution is determined by the size of the Ad-LacZ Ad-Smad3 immunoprecipitated DNA fragments, typically 200 to 500 bp. The gap between E1 and E2 is only 12 bp, while E1 and E3 are separated by Scleraxis 30 310 bp, thus while it is clear that scleraxis binds to this region, it is

α-tubulin 50 A 180 # Ad-LacZ Ad-Smad7 160 *

140 Scleraxis 30 120

α-tubulin 50 100 80 C 60 * 40

1.6 Luciferase expression (fold) * Ad-LacZ 20 1.4 Ad-Smad 0 pGL3 COL1 2pr-3.7 COL1 2pr-0.7 1.2 B 1.0 # Empty vector * 0.8 * 8 Scleraxis 0.6

0.4 6 0.2 Scleraxis protein expression (fold) 0 4 Smad3 Smad7 * ﬁ Fig. 1. Regulation of scleraxis expression by Smads. A) Primary adult rat cardiac bro- 2

blasts were infected with adenovirus encoding Smad3 (MOI10), Smad7 (MOI100) or Luciferase expression (fold) LacZ (MOI10 or 100) adenoviruses. After treatment for 24 h, cells were lysed and total RNA was isolated for quantitative PCR. Scleraxis mRNA abundance was calculated using the 2−ΔΔCT method and normalized to GAPDH. Results represent three indepen- 0 COL1 2pr-3.7 COL1 2pr-0.7 dent experiments, normalized to AdLacZ (control), and are reported as mean±standard error; *pb0.05 vs. AdLacZ. B) Isolated adult rat cardiac fibroblasts were infected with adenoviruses encoding LacZ or Smads as in A. Twenty-four hours later, total cell Fig. 2. Transactivation of the proximal human COL1A2 promoter by scleraxis. A) NIH- fi fl protein was isolated for western blotting with anti-scleraxis antibodies and 3T3 broblasts at ~70% con uence were co-transfected with luciferase reporter plas- α-tubulin as a loading control. Numbers refer to molecular weight in kilodaltons. Re- mids (3.7 kb full length or 0.7 kb proximal COL1A2 promoter cloned into pGL3 Basic, sults are quantified in C and represent three independent experiments, normalized to or empty pGL3 Basic (pGL3) as control) plus scleraxis expression vector (pECE-HA- α-tubulin expression, and are reported as mean±standard error; *pb0.05 vs. AdLacZ. FLAG-Scx). Renilla luciferase expression vector pRL was co-transfected as a transfection control. The luciferase activity of each sample was assayed 24 h later. Results of three independent experiments were normalized to pGL3 and pRL, and are reported as the full-length and proximal promoters by scleraxis in comparison to an mean±standard error; *pb0.05 vs. pGL3, #pb0.05 vs. full length promoter. B) Lucifer- empty expression vector. Again we observed that both promoters were ase reporter assays were performed as in A with either full length or proximal COL1A2 luciferase reporter vectors, plus empty pECE or scleraxis expression vector. Results significantly transactivated by scleraxis, and that the proximal promot- were normalized to pECE and pRL. Results of three independent experiments are fi er was induced to a statistically signi cantly greater degree than the reported as mean±standard error; *pb0.05 vs. empty vector, #pb0.05 vs. full-length full-length promoter, also by approximately three-fold (Fig. 2B). These promoter. 1940 R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944 difficult to ascertain by ChIP the importance of any single E box. To (Fig. 3C). Intriguingly, mutation of both E1 and E3 completely prevented determine which of these three sites are involved in scleraxis- promoter transactivation by scleraxis. Transactivation was significantly mediated transactivation of the COL1A2 promoter, we performed muta- lower than control, and was significantly decreased further when all tion analysis studies in which individual or combinations of E boxes three E boxes (ΔE1/2/3) were mutated. were mutated to prevent scleraxis binding in luciferase reporter assays. These results indicate that E boxes E1 and E3 are strictly required for Mutation of any single E box (ΔE1, ΔE2 or ΔE3; Fig. 3A) resulted in a sig- transactivation of the COL1A2 proximal promoter by scleraxis. Con- nificant attenuation of promoter transactivation by scleraxis, although versely, E2 appears to have a contributory effect but is not absolutely re- transactivation was still significantly greater than control empty vector quired: although E2 mutation alone reduced transactivation by scleraxis, the double mutations ΔE1/2 and ΔE2/3 did not cause the dra- matic attenuation observed with ΔE1/3, although the triple mutant was fi Δ A E1 E2 E3 signi cantly lower than E1/3 (Fig. 3C). One possibility is that E2, separated from E1 by only 12 nucleotides, may indirectly modulate tran- -652 +54 Hs scription factor binding to E1, perhaps by affecting local DNA Mm topography and thus access by scleraxis to E1. Rr The attenuation of scleraxis-mediated transactivation of the COL1A2 promoter below control in the ΔE1/3 and ΔE1/2/3 experiments may be fi E box Sp1 Sp1/Sp3 SBE due to endogenous scleraxis expression in the NIH-3T3 broblasts used in the reporter assay [9]. Thus, the intact COL1A2-pr0.7 promoter re- E1: CAAGTG E2: CAGGTG porter may exhibit some activity even in the absence of exogenously- ΔE1: CcAtTG ΔE2: CtGcaG added scleraxis. Mutation of the reporter E boxes would eliminate SBE: CAGACA E3: CAGCTG transactivation by endogenous scleraxis, resulting in reporter activity ΔSBE: CgcACA ΔE3: CccggG below control levels. To confirm the ability of scleraxis to bind to E boxes 1 and 3, electro- B Scleraxis Pre-immune phoretic mobility shift assays were performed using scleraxis-containing ChIPInput ChIP Input nuclear extracts and biotin-labeled oligonucleotides corresponding to one or the other E box. A shifted band was observed for both E boxes E1/E2 which was lost with a 250-fold excess of unlabeled oligonucleotide com- petitor (Fig. 3D, arrows). Both bands were supershifted upon inclusion of E3 anti-scleraxis antibody (Fig. 3D, arrowheads), confirming the formation of a scleraxis-DNA complex. GAPDH C 10 3.4. Synergistic regulation of the COL1A2 promoter by scleraxis and * Smads

P<0.001 ﬁ β 8 Multiple pro- brotic TGF- 1-responsive signaling pathways con- verge on the COL1A2 proximal promoter [7,8]. While we have shown

that TGF-β1 also induces scleraxis expression, and that scleraxis in # P<0.001 * turn up-regulates COL1A2 expression, it is unclear whether TGF-β1 6 *# *# and scleraxis exert independent effects on promoter activation [9]. We therefore examined whether TGF-β1 and scleraxis exert additive ef-

# fects on proximal COL1A2 promoter transactivation. While scleraxis 4 * # * alone strongly activated the promoter, the addition of TGF-β1 induced

Fig. 3. Interaction of scleraxis with the proximal COL1A2 promoter. A) The proximal

Luciferase expression (fold) − 2 human COL1A2 gene promoter, spanning a region from 652 to +54 relative to the transcription start site, contains three putative E boxes (black rectangles, denoted E1 P<0.05 to E3) that match the consensus CANNTG E box sequence. For comparison, the locations # α * *# of homologous E boxes within the mouse (Mm) and rat (Rr) Col1 2 gene promoters are 0 depicted in comparison to the human (Hs) promoter. Binding sites for various other Control pr-0.7 ΔE1 ΔE2 ΔE3 ΔE1/2 ΔE2/3 ΔE1/3 ΔE1/2/3 regulatory transcription factors within the human promoter are indicated, including +Scleraxis Sp1 GC boxes (light gray ovals), Sp1/Sp3 T boxes (dark gray ovals), and a Smad Binding Element (SBE; white rectangle). The E box and SBE mutations used in the study are also D indicated. B) Chromatin immunoprecipitation was performed on human cardiac fibro- E-Box1 E-Box3 blasts using anti-scleraxis (Scx) antibody or pre-immune serum. PCR primers were used to amplify the collagen 1α2 promoter sequence flanking E boxes 1 and 2 (E1/ E2) or E box 3 (E3). Genomic DNA was used as a positive input control. Scleraxis binds to both the E1/E2 and E3 sites. A region of the human GAPDH promoter was used as a negative control. C) Point mutations were sequentially introduced to the three E boxes of the COL1A2 proximal promoter individually or in combination as per the sequences shown in panel A. NIH-3T3 fibroblasts were transfected with intact (pr-0.7) or mutated (ΔE1, ΔE2 etc.) proximal COL1A2 promoter reporters plus scleraxis expression vector or empty pECE (Control). pRL was co-transfected as a transfection control. Twenty-four hours after transfection, luciferase assays were performed. Results of three independent experiments were normalized to Control and pRL, and are reported as mean±standard error; *pb0.001 vs. Control, #pb0.001 vs. pr-0.7. D) The binding of scleraxis to E boxes 1 and 3 was confirmed by EMSA. Scleraxis-containing Nuclear extract ++++++ COS7 nuclear extracts were incubated with biotin-labeled oligonucleotide probes Oligo probe ++++ ++++ corresponding to either E box. A 250-fold molar excess of unlabeled oligonucleotide 250x cold + + (250× cold) was used to compete the scleraxis-DNA complex (arrow). Inclusion of an Anti-scleraxis + + anti-scleraxis antibody resulted in a supershift of the complex (arrowhead). R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944 1941 a further statistically significant increase in promoter activation other bHLH proteins while at the same time interfering with binding (Fig. 4A). This result suggests that scleraxis may function together to DNA [21]. We previously generated a scleraxis mutant lacking the with other TGF-β1-responsive signaling pathways such as Smads. DNA-binding basic domain (ScxΔBD) and demonstrated that this mu- To determine whether scleraxis directly interacts with the Smad tant was completely unable to transactivate the COL1α2 promoter [9]. signaling pathway to regulate transcription, luciferase assays were performed with scleraxis in combination with pro-fibrotic Smad3, A Smad2 or inhibitory Smad7. As expected, either scleraxis or Smad3 16 alone significantly transactivated the proximal COL1A2 promoter. *# Smad2 alone had a statistically significant but much more modest ef- 14 fect on COL1A2 transactivation. Intriguingly, scleraxis and Smad3 12 together exhibited synergistic activation of the promoter (~33-fold activation, compared to 10~14-fold activation for scleraxis or 10 * Smad3 alone; Fig. 4B). Conversely, Smad7 completely abrogated pro- 8 moter transactivation by scleraxis. Smad2 had no effect in combination with scleraxis, and only a modest effect when transfected with 6 Smad3 and scleraxis together, in keeping with its greatly diminished 4 ability to regulate gene expression compared to Smad3 [20]. Luciferase expression (fold) 2

3.5. Cross-talk in proximal COL1A2 promoter activation by Smad3 and 0 β scleraxis Control Vehicle TGF- 1 Scleraxis The scleraxis binding sites are in close juxtaposition with those of B other fibrotic regulators: E1 and E3 are a mere 167 and 136 nucleotides, #& 35 Empty vector #& * respectively, away from the SBE (Fig. 3A), suggesting the possibility Expression vector * that these sites contribute to the assembly of a common transcription 30 complex. We tested whether scleraxis-mediated transactivation of the 25 proximal COL1A2 promoter is dependent on Smad3 and vice versa. Scleraxis-mediated promoter transactivation was significantly attenuat- 20 ed by mutation of the SBE (Fig. 4C). Similarly, Smad3-mediated promoter activation was significantly attenuated by mutation of the two critical 15 scleraxis binding sites, E1 and E3 (Fig. 4D).Inbothcases,promoter * 10 * * transactivation was reduced by approximately 50%, similar to the loss – of activity observed when the HLH protein protein interaction domain Luciferase expression (fold) 5 * ofscleraxisisdeleted[9]. Collectively, these results suggest that scleraxis # 0 and Smad3 functionally interact to regulate COL1A2 trans-activation. ScxS2S3 Scx Scx Scx S7 Scx +S2 +S3 +S2 +S7 3.6. Blockade of COL1A2 expression by a scleraxis dominant negative C +S3 mutant 16 14 * bHLH transcription factors contain a basic domain that mediates DNA binding [12,13]. Deletion of this domain may give rise to a protein 12 that acts in a dominant negative fashion to inhibit gene expression. For 10 example, Inhibitor of Differentiation 2 is an HLH transcriptional repressor that functions by binding to ubiquitous bHLH factors like E12 and 8 *# E47, thus preventing these factors from forming heterodimers with 6

Fig. 4. Cross-talk between scleraxis and Smad signaling. A) NIH-3T3 fibroblasts were 4 transfected with empty pECE (Control) or scleraxis expression vectors plus the proximal COL1A2 luciferase reporter vector. Twenty-four hours later, cells were treated with vehicle Luciferase expression (fold) 2 or 10 ng/ml TGF-β as indicated for an additional 24 h prior to luciferase assay. pRL was 1 0 co-transfected as a transfection control. Results of three independent experiments were pGL3 pr-0.7 pr-0.7ΔSBE normalized to Control and pRL, and are reported as mean±standard error; *pb0.05 vs. Control, #pb0.05 vs. scleraxis plus vehicle. B) Luciferase reporter assays were performed D using proximal COL1A2 promoter reporter vector, plus empty pECE or scleraxis expression 16 vector; or pCMV tag3C or pCMV tag3C-Myc-Smad3 or pCMV tag3C-Myc-Smad2; or * pcDNA3.1(+) or pcDNA3.1(+)-FLAG-Smad7. pRL was co-transfected as a transfection con- 14 trol. Results of three independent experiments were normalized to the respective empty vectors and pRL, and are reported as mean±standard error; *pb0.001 vs. empty vector, 12 #pb0.001 vs. scleraxis alone; &pb0.001 vs. Smad3 alone. C) NIH-3T3 fibroblasts were 10 transfected with the intact (pr-0.7) or SBE mutant (pr-0.7ΔSBE) proximal COL1A2 promoter reporter or pGL3 (Control), plus scleraxis expression vector. pRL was co-transfected as a 8 transfection control. Twenty-four hours after transfection, luciferase assays were *# performed. Results of three independent experiments were normalized to empty pGL3 vec- 6 tor and pRL, and are reported as mean±standard error; *pb0.05 vs. pGL3, #pb0.05 vs. pr-0.7. D) NIH-3 T3 fibroblastsweretransfectedwiththeintact(pr-0.7)ordoubleEbox 4 mutant (pr-0.7ΔE1/3) proximal COL1A2 promoter reporter or pGL3 (Control), plus scleraxis expression vector. pRL was co-transfected as a transfection control. Luciferase as- Luciferase expression (fold) 2 says were performed 24 h after transfection. Results of three independent experiments were normalized to empty pGL3 vector and pRL, and are reported as mean±standard 0 error; *pb0.05 vs. pGL3, #pb0.05 vs. pr-0.7. pGL3 pr-0.7 pr-0.7ΔE1/3 1942 R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944

We tested whether this mutant may act as a dominant negative regula- We next examined whether ScxΔBD could interfere with TGF- tor of COL1α2 promoter transactivation via luciferase assays in NIH-3T3 β1-mediated induction of collagen 1α2 gene expression. We con- fibroblasts using ScxΔBD in combination with scleraxis. Scleraxis- firmed that collagen 1α2 mRNA expression was up-regulated to sim- mediated transactivation of the proximal COL1A2 promoter was atten- ilar levels in NIH-3T3 fibroblasts in response to TGF-β1 treatment or uated by ScxΔBD in a dose-dependent manner, with as little as 100 ng scleraxis over-expression (Fig. 5B). Strikingly, ScxΔBD completely ab- transfected ScxΔBD vector sufficient to completely block transactiva- rogated that effect of TGF-β1 on collagen 1α2 expression in these tion (Fig. 5A). The three highest doses of ScxΔBD vector significantly re- cells. Indeed, ScxΔBD also resulted in a significant reduction in basal duced luciferase activity below the level of the empty vector control. As collagen 1α2 mRNA expression compared to vehicle-treated cells in Fig. 3C above, this suggests that basal transactivation of the reporter transfected with an empty expression vector. These results support by endogenous factors was inhibited, and demonstrates that the the suggestion that scleraxis plays a critical role in mediating

ScxΔBD mutant does, indeed, behave as a dominant negative regulator TGF-β1-induced COL1α2 gene expression, and demonstrate that of scleraxis-mediated gene transactivation. ScxΔBD is capable of interfering with TGF-β1-mediated collagen production. We conﬁrmed these results in primary rat passage 2 cardiac myoﬁbroblasts at both the mRNA and protein level (Fig. 5C, D).

A 4. Discussion 14 * 12 Cardiac fibrosis remains a significant health problem, with virtual- ly no approved therapies directly targeting this process in human pa- 10 tients. In light of the contributory nature of fibrosis in cardiac 8 diseases, the identification of novel regulators of this pathological process is thus an important goal for the development of new thera- 6 peutic approaches to improve patient outcomes. The data presented # 4 * here support our hypothesis that scleraxis is an important new target for the development of these anti-fibrotic therapies, and provides an 2

Luciferase expression (fold) example of a novel approach to attenuating collagen 1α2 production # # # 0 * * * using our ScxΔBD mutant [3,9]. Control - Δ Scx BD Herein we have clarified the mechanism by which scleraxis regulates Scleraxis human COL1A2 gene expression by determining the relative importance B of three E boxes within the proximal gene promoter. Mutation analysis 6 reveals that E1 and E3 are critical for scleraxis-mediated transactivation * fi Vehicle of this promoter (Fig. 3C). The nding that scleraxis binds to these E 5 fi TGF-β1 boxes under basal conditions in adult human cardiac broblasts * (Fig. 3B) suggests that scleraxis contributes directly to maintenance of 4 * normal cardiac fibroblast collagen expression. Pro-fibrotic conditions in- 3 duce scleraxis expression, further augmenting collagen gene expression as reported by our laboratory [9]. The potent increase in collagen gene 2 expression mediated by scleraxis (Figs. 2–5), coupled with the induction

2 mRNA expression (fold) β α NS of scleraxis expression by TGF- 1 [9] or its effector Smad3 (Fig. 1) indi- 1 cates that scleraxis represents a novel TGF-β1-inducible mechanism to

Col1 # # * regulate collagen synthesis. 0 * Empty vector Scleraxis ScxΔBD Until now, it has been unclear whether scleraxis represents a wholly novel mechanism for regulating collagen expression independent of

C p<0.01 6 & * Fig. 5. Inhibition of collagen 1α2 expression by a scleraxis dominant negative mutant Vehicle 5 ScxΔBD. A) NIH-3T3 fibroblasts were transfected with the proximal COL1A2 TGF-β1 * promoter reporter plus empty pECE (Control) or intact scleraxis expression vectors. 4 * Some samples were co-transfected with increasing amounts (20, 100, 250 or 500 ng DNA) of ScxΔBD. pRL was co-transfected as a transfection control. Twenty four hours 3 after transfection, luciferase assays were performed. Results of three independent experiments were normalized to pECE and pRL, and are reported as mean±standard error; *pb0.05 vs. Control, #pb0.05 vs. intact scleraxis alone (−). B) NIH-3T3 2 fibroblasts were transfected with empty pECE, intact scleraxis or ScxΔBD expression 2 mRNA expression (fold) vectors. Twenty four hours later, cells were treated with vehicle or 10 ng/ml TGF-β . α NS 1 1 After another 24 h, total RNA was isolated for quantitative RT-PCR. Col1α2 mRNA # # −ΔΔCT Col1 * * abundance was calculated using the 2 method and was normalized to GAPDH. 0 Results of three independent experiments were normalized to samples receiving Ad-LacZ Ad-Scleraxis Δ Ad-Scx BD empty vector plus vehicle, and are reported as mean±standard error; *pb0.05 vs. empty vector+vehicle, #pb0.05 vs. corresponding intact scleraxis-transfected samples. NS, not significant. C) A similar experiment was carried out as in panel B, using D Ad-LacZ Ad-ScxΔBD primary rat passage 2 myofibroblasts and adenoviruses encoding scleraxis or ScxΔBD

(MOI 10). Results of three independent experiments were normalized to samples re-

i i

i ceiving empty vector plus vehicle, and are reported as mean±standard error;

h h

h b b

G G

G *p 0.01 vs. empty vector+vehicle, #p 0.001 vs. corresponding intact scleraxis-

T V

T V infected samples; &pb0.01 vs. empty vector plus TGF‐β1. NS, not signiﬁcant. D) Prima- ﬁ β Collagen I 150 ry rat passage 2 myo broblasts were treated with vehicle or 10 ng/ml TGF- 1 with or without adenovirus encoding the ScxΔBD mutant for 24 h. Total protein lysates were α-tubulin separated by SDS-PAGE, western blotted as in the Materials and methods and visual- 50 ized using anti-collagen type I antibody. α-tubulin was used as a loading control. R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944 1943

existing pathways such as TGF-β1/Smad signaling. The data presented transcriptional complex. It is also possible that these factors both bind here suggests that, while scleraxis alone is sufficient to regulate collagen to a common but as yet unknown partner. 1α2 expression, there is extensive cross-talk between scleraxis and It is likely that scleraxis partners with other proteins for full activity. the TGF-β1/Smad signaling network. We found that scleraxis expression While our dominant negative ScxΔBD DNA-binding mutant potently was significantly up-regulated in response to Smad3 but down- blocked collagen 1α2expression(Fig. 5), deletion of the SBE in the regulated by inhibitory Smad7 in primary rat cardiac fibroblasts COL1A2 promoter only reduced transactivation by scleraxis by ~50%

(Fig. 1). This finding places scleraxis downstream of TGF-β1/Smad signal- (Fig. 4C). If scleraxis was absolutely dependent on Smad3 for formation ing, although it is likely that other independent signals may be capable of of a transcription complex, one would expect complete attenuation of inducing scleraxis expression. scleraxis-mediated transactivation of the SBE mutant promoter. Others Importantly, there appears to be considerable functional interaction have suggested that full scleraxis activity may be dependent on between scleraxis and Smad3. Analysis of the proximal COL1A2 gene heterodimerization with ubiquitous E box-binding proteins such as promoter reveals tight flanking of the SBE by the E boxes that are critical E12 and E47 [11,12,17,19,24]. The very close spacing of the E boxes E1 for scleraxis-mediated regulation (Fig. 3A), suggesting the possibility and E2 within the COL1A2 promoter may facilitate interaction with that scleraxis and Smad3 form a common transcriptional complex. How- just such a heterodimer. This may explain why deletion of E2 reduced ever, co-immunoprecipitation experiments failed to demonstrate an ac- activation of the promoter by scleraxis, but was not as critical as E1 tual physical interaction between scleraxis and Smad3 (results not and E3 (Fig. 3C): scleraxis could bind the promoter alone, without a shown). Smad2 does not appear to play a critical role in this regard, binding partner, but still maintain partial activity consistent with our and thus does not functionally substitute for Smad3: Smad2 exerted sta- scleraxis HLH mutant. However, others have suggested that E protein tistically significant (pb0.01) but very modest effects on COL1A2 pro- heterodimerization with scleraxis is not absolutely required for moter transactivation, either alone or in combination with scleraxis scleraxis function and our own previous data indicates that E12 and and/or Smad3 (Fig. 4B). This is not surprising given a previous report E47 are not regulators of scleraxis-mediated transactivation of the that Smad2 is a much weaker driver of TGF-β-mediated gene COL1A2 promoter, thus the importance of these E proteins is likely to transactivation than Smad3 [20]. be gene- or context-dependent [9,19]. Despite an apparent lack of physical interaction between Smad3 and While protein–protein interaction may not be required for scleraxis scleraxis, there is clear functional interaction between these factors, function, our results suggest that scleraxis itself plays a critical role in since mutation of the SBE significantly attenuated scleraxis-mediated regulating collagen gene expression. ScxΔBD not only attenuated basal promoter transactivation, while mutation of the E boxes had the same collagen 1α2 mRNA expression, it also fully prevented induction by effect on Smad3 (Fig. 4C, D). In both cases, the promoter mutations TGF-β1, a potent inducer of fibrosis (Fig. 5B–D). These findings, when did not eliminate the ability of scleraxis or Smad3 to transactivate the considered in the context of the report that scleraxis knockout results promoter, suggesting that either factor alone can drive transcription, in improper development of force-transmitting tendons lacking type I albeit with reduced activity. This finding thus conceptually agrees collagen, strongly support a key central role for scleraxis in fibrillar col- with our previous study in which deletion of the scleraxis HLH protein– lagen synthesis in multiple tissues [10].OurScxΔBD mutant protein may protein interaction domain reduced but did not eliminate scleraxis be useful for proof-of-concept experiments to demonstrate a central role activity to a similar degree (~50%). We postulated at the time that HLH for scleraxis in fibrosis. We have shown that scleraxis is up-regulated in deletion prevented scleraxis from interacting with a key co-factor or the collagen-rich cardiac infarct scar, and others have demonstrated that partner, thus reducing but not eliminating its activity [9].Wehypothe- scleraxis expression becomes activated in collagen-rich dermal keloids sized that this partner may be Smad3, since scleraxis and Smad3 [9,25]. Both of these disparate pathologies are regulated by TGF-β1, exhibited a synergistic effect on COL1A2 promoter activation compared thus it is possible that interference with scleraxis function may be a use- to either factor alone (Fig. 4B) and since previous studies have demon- fulapproachinanumberoffibrotic diseases marked by over-expression strated binding of Smads (including Smad3) to bHLH transcription of type I collagen including cardiac fibrosis [26,27]. The results presented factors [22,23]. However, since our initial co-immunopreciptation ex- here support a model of COL1A2 gene expression mediated by the syn- periments appear to rule this interaction out, it is necessary to consider ergistic interaction of scleraxis and Smad3 downstream of TGF-β1 the possibility that these factors may simply form part of a common (Fig. 6).

TGF- 1

Smad 3

Smad7

Scleraxis

Scleraxis Smad 3 Scleraxis COL1 2 E1/2SBE E3

Fig. 6. Proposed model for interaction of scleraxis and Smads in the regulation of COL1A2 gene expression. Pro-ﬁbrotic TGF-β1 binds to its plasma membrane-bound receptor complex to activate Smad3, which in turn regulates the COL1A2 gene via direct binding to the SBE located in the promoter. Our data demonstrates that Smad3 also induces scleraxis expression via an unknown mechanism. Both effects of Smad3 are negatively regulated by inhibitory Smad7. Increased scleraxis expression works synergistically with Smad3 to strongly up-regulate COL1A2 expression via scleraxis binding to E boxes within the proximal promoter. 1944 R.A. Bagchi, M.P. Czubryt / Biochimica et Biophysica Acta 1823 (2012) 1936–1944

Acknowledgements [11] V. Lejard, G. Brideau, F. Blais, R. Salingcarnboriboon, G. Wagner, M.H. Roehrl, M. Noda, D. Duprez, P. Houillier, J. Rossert, Scleraxis and NFATc regulate the expression of the pro-alpha1(I) collagen gene in tendon fibroblasts, J. Biol. Chem. 282 The authors thank Dr. Eric Olson (University of Texas Southwestern (2007) 17665–17675. Medical Center, Dallas) for the pECE-HA-FLAG-Scleraxis construct; [12] P. Cserjesi, D. Brown, K.L. Ligon, G.E. Lyons, N.G. Copeland, D.J. Gilbert, N.A. Jenkins, E.N. Olson, Scleraxis: a basic helix-loop-helix protein that prefigures skeletal forma- Dr. Jean-Jacques Lebrun (McGill University, Montreal) for Smad expres- tion during mouse embryogenesis, Development 121 (1995) 1099–1110. sion vectors; Dr. Ian Dixon for Smad-encoding adenoviruses; Mr. [13] J.D. Molkentin, E.N. Olson, Combinatorial control of muscle development by basic Stephen Jones for the full-length COL1A2 luciferase reporter; and Mr. helix–loop–helix and MADS-box transcription factors, Proc. Natl. Acad. Sci. U. S. A. – Leon Espira, Ms. Sari Yakubovich, Ms. Kristin Swan and Ms. Laura 93 (1996) 9366 9373. [14] S.J. Chen, W. Yuan, S. Lo, M. Trojanowska, J. Varga, Interaction of smad3 with a prox- Albak for technical assistance. This work was supported by an operating imal smad-binding element of the human alpha2(I) procollagen gene promoter grant to M.P.C. from the Canadian Institutes for Health Research required for transcriptional activation by TGF-beta, J. Cell. Physiol. 183 (2000) – (MOP-106671). M.P.C. was a McDonald Scholar and New Investigator 381 392. [15] J.E. Raizman, J. Komljenovic, R. Chang, C. Deng, K.M. Bedosky, S.G. Rattan, R.H. of the Heart and Stroke Foundation of Canada. R.A.B. was supported Cunnington, D.H. Freed, I.M. Dixon, The participation of the Na+−Ca2+ ex- by a Manitoba Health Research Council Coordinated Graduate Student- changer in primary cardiac myofibroblast migration, contraction, and prolifera- ship Award. tion, J. Cell. Physiol. 213 (2007) 540–551. [16] H. Ju, J. Hao, S. Zhao, I.M. Dixon, Antiproliferative and antifibrotic effects of mimosine on adult cardiac fibroblasts, Biochim. Biophys. Acta 1448 (1998) 51–60. [17] A.L. Carlberg, R.S. Tuan, D.J. Hall, Regulation of scleraxis function by interaction References with the bHLH protein E47, Mol. Cell Biol. Res. Commun. 3 (2000) 82–86. [18] K. Higashi, D.J. Kouba, Y.J. Song, J. Uitto, A. Mauviel, A proximal element within [1] W.S. Colucci, E. Braunwald, Pathophysiology of heart failure, in: E. Braunwald the human alpha 2(I) collagen (COL1A2) promoter, distinct from the tumor ne- (Ed.), Heart Disease, W.B. Saunders Company, Philadelphia, 1997, pp. 394–420. crosis factor-alpha response element, mediates transcriptional repression by [2] W.B. Kannel, Epidemiology of heart failure in the United States, in: P.A. interferon-gamma, Matrix Biol. 16 (1998) 447–456. Poole-Wilson, W.S. Colucci, B.M. Massie, K. Chatterjee, A.J.S. Coats (Eds.), Heart [19] Y. Liu, H. Watanabe, A. Nifuji, Y. Yamada, E.N. Olson, M. Noda, Overexpression of a single Failure, Churchill Livingstone, New York, 1997, pp. 279–288. helix–loop–helix-type transcription factor, scleraxis, enhances aggrecan gene expression [3] L. Espira, M.P. Czubryt, Emerging concepts in cardiac matrix biology, Can. J. Phys- in osteoblastic osteosarcoma ROS17/2.8 cells, J. Biol. Chem. 272 (1997) 29880–29885. iol. Pharmacol. 87 (2009) 996–1008. [20] R.P. Nagarajan, Y. Chen, Structural basis for the functional difference between [4] B. Swynghedauw, Molecular mechanisms of myocardial remodeling, Physiol. Rev. Smad2 and Smad3 in FAST-2 (forkhead activin signal transducer-2)-mediated 79 (1999) 215–262. transcription, Biochem. J. 350 (Pt 1) (2000) 253–259. [5] B. Wang, J. Hao, S.C. Jones, M.S. Yee, J.C. Roth, I.M. Dixon, Decreased Smad 7 ex- [21] K. Langlands, X. Yin, G. Anand, E.V. Prochownik, Differential interactions of Id proteins pression contributes to cardiac fibrosis in the infarcted rat heart, Am. J. Physiol. with basic-helix–loop–helix transcription factors, J. Biol. Chem. 272 (1997) 19785–19793. Heart Circ. Physiol. 282 (2002) H1685–H1696. [22] S.J. Yoon, A.E. Wills, E. Chuong, R. Gupta, J.C. Baker, HEB and E2A function as [6] M. Schiller, D. Javelaud, A. Mauviel, TGF-beta-induced SMAD signaling and gene SMAD/FOXH1 cofactors, Genes Dev. 25 (2011) 1654–1661. regulation: consequences for extracellular matrix remodeling and wound [23] D. Liu, B.L. Black, R. Derynck, TGF-beta inhibits muscle differentiation through healing, J. Dermatol. Sci. 35 (2004) 83–92. functional repression of myogenic transcription factors by Smad3, Genes Dev. [7] F. Ramirez, S. Tanaka, G. Bou-Gharios, Transcriptional regulation of the human 15 (2001) 2950–2966. alpha2(I) collagen gene (COL1A2), an informative model system to study fibrotic [24] T. Muir, I. Sadler-Riggleman, M.K. Skinner, Role of the basic helix-loop-helix tran- diseases, Matrix Biol. 25 (2006) 365–372. scription factor, scleraxis, in the regulation of Sertoli cell function and differenti- [8] S. Boast, M.W. Su, F. Ramirez, M. Sanchez, E.V. Avvedimento, Functional analysis ation, Mol. Endocrinol. 19 (2005) 2164–2174. of cis-acting DNA sequences controlling transcription of the human type I colla- [25] M. Naitoh, H. Kubota, M. Ikeda, T. Tanaka, H. Shirane, S. Suzuki, K. Nagata, Gene gen genes, J. Biol. Chem. 265 (1990) 13351–13356. expression in human keloids is altered from dermal to chondrocytic and osteo- [9] L. Espira, L. Lamoureux, S.C. Jones, R.D. Gerard, I.M. Dixon, M.P. Czubryt, The basic genic lineage, Genes Cells 10 (2005) 1081–1091. helix-loop-helix transcription factor scleraxis regulates fibroblast collagen syn- [26] D.A. Bettinger, D.R. Yager, R.F. Diegelmann, I.K. Cohen, The effect of TGF-beta on thesis, J. Mol. Cell. Cardiol. 47 (2009) 188–195. keloid fibroblast proliferation and collagen synthesis, Plast. Reconstr. Surg. 98 [10] N.D. Murchison, B.A. Price, D.A. Conner, D.R. Keene, E.N. Olson, C.J. Tabin, R. Schweitzer, (1996) 827–833. Regulation of tendon differentiation by scleraxis distinguishes force-transmitting ten- [27] J. Jagadeesan, A. Bayat, Transforming growth factor beta (TGFbeta) and keloid dis- dons from muscle-anchoring tendons, Development 134 (2007) 2697–2708. ease, Int. J. Surg. 5 (2007) 278–285.