YB-1 Coordinates Vascular Smooth Muscle α-Actin Gene Activation by Transforming Growth Factor β1 and Thrombin during Differentiation of Human Pulmonary Myofibroblasts

INTRODUCTION

Fibrosis is a serious complication of chronic cardiopulmonary diseases and postsurgical complication of heart and lung transplant (Pickering and Boughner, 1990; Armstrong et al., 1997; Howell et al., 2002; Chapman, 2004). Although management of acute allograft rejection is accomplished through the use of immunosuppressive agents that can limit immune cell infiltration, the cause of chronic rejection and failure in accepted allografts is poorly understood thus offering no treatment solution short of retransplant. Leading the formation of scar tissue is a specialized stromal cell referred to as the myofibroblast that contains abundant microfilament networks composed of smooth muscle-specific contractile protein isoforms (Tomasek et al., 2002; Hinz and Gabbiani, 2003; Grotendorst et al., 2004). Chronic accumulation of myofibroblasts is associated with excessive extracellular matrix protein biosynthesis, hypercontractility, and organ-destructive remodeling. Vascular smooth muscle α-actin (SMαA) is one of the major contractile proteins expressed by differentiated myofibroblasts (Darby et al., 1990; Ronnov-Jessen and Petersen, 1996; Hinz et al., 2001; Cogan et al., 2002). Normally, SMαA-enriched myofibroblasts are early, transient participants in stromal wound healing processes. We reason that characterization of the initial molecular events associated with activation of the SMαA gene in stromal cells would help identify novel checkpoints for regulating the extent of myofibroblast involvement in wound healing. Therapeutic interventions aimed at managing potentially rate-limiting steps in myofibroblast differentiation may help protect against uncontrolled tissue remodeling and the so-called “endless healing” phenotype (Tomasek et al., 2002) with attendant hypertrophic scarring and loss of parenchymal cell function.

In recent studies from our laboratory, we have observed that the SMαA gene is transcriptionally activated by the cytokine TGFβ1 via Smad-dependent signaling (Cogan et al., 2002; Subramanian et al., 2004). In the context of wound healing, this potent, profibrotic cytokine is released from a latent state at the site of tissue injury by activators intimately associated with local monocyte/macrophage infiltration and adhesion such as matrix metalloproteinase (MMP)-2 and -9, thrombospondin-1, and αVβ6 integrin (Leask and Abraham, 2004). Less well understood, and especially relevant to the situation of organ transplantation in the immunosuppressed human patient population, is the role of nonimmune based surgical trauma and graft ischemic/reperfusion injury in promoting myofibroblast activation and chronic tissue remodeling. The serine protease thrombin, released during donor organ harvest and subsequent engraftment not only initiates the enzyme cascade responsible for coagulation but also has been reported to activate latent TGFβ1 (Howell et al., 2002) and possess profibrotic and extravascular cellular signal transduction properties that lead to tissue remodeling in both the native and transplanted heart and lung (Erlich et al., 2000; Holschermann et al., 2000; Mackman, 2003). In this regard, our studies of syngeneic murine heart grafts has revealed a form of alloantigen-independent organ remodeling (Armstrong et al., 1997) that evolves slowly but resembles in form the far more robust perivascular and interstitial fibrotic response observed in long-term cardiac allografts in immunosuppressed recipients (Subramanian et al., 2002). In accepted allografts, we hypothesize that persistent, low-grade immune cell infiltration amplifies an underlying peri-transplant wound healing response based on the profibrotic action of TGFβ1 and thrombin released as a consequence of mechanical and/or reperfusion injury during transplant surgery. Amplification of these nonimmune-based wound healing processes by alloantigen or alloantibody augments dysfunctional remodeling leading to premature graft failure.

To more fully understand the molecular mechanism of peri-transplant wound healing, we have examined the effect of thrombin on SMαA gene regulation in human pulmonary fibroblasts (hPFBs) and compared its action to TGFβ1, a well known transcriptional activator of the smooth muscle actin gene. In contrast to regulatory mechanism reported for TGFβ1, thrombin had a significant, potentially supplemental effect on SMαA translation. A common feature of SMαA up-regulation mediated by both thrombin and TGFβ1 was reliance on the cold shock domain protein YB-1 (Matsumoto and Wolffe, 1998; Wolffe, 1998) that represses SMαA gene transcription in stromal fibroblasts yet exhibits additional RNA binding properties that regulate mRNA translational efficiency. YB-1 dissociated from SMαA enhancer DNA and was exported from the nucleus in the presence of TGFβ1 or its Smad coregulatory mediators. On the other hand, thrombin displaced YB-1 from SMαA exon 3 mRNA coding sequences previously shown to be required for translational silencing (Kelm et al., 1999b). Thrombin-facilitated RNA release and nuclear reentry of YB-1 provides compelling evidence for a regulatory loop to control SMαA gene output via coordinate control at both the transcriptional and translational levels. Published evidence indicates that thrombin activates several growth and inflammation-associated genes and that promoter regions flanking these genes contained binding sites for YB-1 (Minami et al., 2004). Besides reducing SMαA gene expression, YB-1 also represses type I collagen α1 and α2 gene transcription (Norman et al., 2001; Higashi et al., 2003). Thus, nuclear YB-1 may temper TGFβ1-dependent, myofibroblast accumulation while stimulating inflammatory cytokines necessary for angiogenesis, restored microperfusion, and infiltration of organ-reconstructive circulating progenitor cells into the wound provisional matrix.

MATERIALS AND METHODS

Cell Culture Methods and Preparation of Protein Extracts

Human pulmonary fibroblasts (hPFB) were established in primary culture from enzyme-dispersed tissue fragments of human neonatal lung tissue obtained at autopsy and were the kind gift of Dr. Daren L. Knoell (Departments of Pharmacy and Internal Medicine, The Ohio State University, Columbus, OH). Pulmonary fibroblasts were maintained in a 1:1 mixture of Ham's F-12 and DMEM (1.0 g/l d-glucose) supplemented with penicillin-streptomycin-Fungizone, Gentamicin (50 μg/ml), and 10% heat-inactivated fetal bovine serum (hiFBS; all culture medium reagents from Invitrogen, Carlsbad, CA). Nonhuman primate COS7 kidney fibroblasts were maintained in DMEM (4.5 g/l d-glucose) supplemented with penicillin-streptomycin and 10% hiFBS. Cell lines were cultivated in a humidified incubator at 37°C at 5% CO₂. Fibroblasts were rendered quiescent by a 48-h exposure to HEPES-buffered DMEM (1.0 g/l d-glucose) containing 0.5% hiFBS and penicillin-streptomycin-Fungizone. Recombinant human TGFβ1 (5 ng/ml, final concentration; R&D Systems, Minneapolis, MN) was added to cultures for varying periods before preparation of protein extracts. Human plasma thrombin (1000 NIH U/mg protein) was obtained from Calbiochem (La Jolla, CA) and used between 1 U and 10 U/ml as noted in the text. The metabolic inhibitors actinomycin D and cycloheximide were used at 5 and 10 μg/ml, respectively, in experiments on thrombin-activated myofibroblasts. Inhibitors of the TGFβ1 type I receptor serine/threonine kinase (SB431542; Sigma-Aldrich, St. Louis, MO) and thrombin serine protease (Thromstop, N-α-NAPAP; American Diagnostica, Stamford, CT) were administered at 1 and 0.1 μM, respectively, for 1 h before exposing hPFBs to medium supplemented with TGFβ1 or thrombin plus fresh aliquots of inhibitors. For protein extract preparations, cell monolayers were washed twice with Dulbecco's phosphate-buffered saline (PBS), scraped into fresh PBS, sedimented at 3000 rpm, washed once more in PBS, and resuspended in eight packed-cell volumes of hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.2 mM phenylmethylsulfonyl fluoride [PMSF], and 0.5 mM dithiothreitol [DTT]). Cells were allowed to swell for 10 min on ice before transfer to a Dounce homogenizer for processing with a type B pestle. Nuclei were collected from ruptured cells by centrifugation for 15 min at 4000 rpm and suspended in one-half packed-pellet volume of ice-cold, low salt buffer (20 mM HEPES, pH 7.9, 25% glycerol, 1.5 mM MgCl₂, 0.02 M KCl, 0.2 mM EDTA, 0.2 mM PMSF, 0.5 mM DTT). High salt buffer (20 mM HEPES, pH 7.9, 25% glycerol, 1.5 mM MgCl₂, 1.2 M KCl, 0.2 mM EDTA, 0.2 mM PMSF, and 0.5 mM DTT) equal to one-half packed-pellet volume was added, and the nuclei were further extracted with gentle rocking for 30 min at 4°C. Nuclei were collected by centrifugation for 30 min at 14,500 rpm and dialyzed against 50 volumes of dialysis buffer containing 20 mM HEPES, pH 7.9, 20% glycerol, 100 mM KCl, 0.2 mM EDTA, 0.2 mM PMSF, and 0.5 mM DTT. After dialysis, supernatants were collected by centrifugation at 14,500 rpm for 20 min for use in biochemical assays. Whole cell extracts were prepared from PBS-rinsed monolayers using radioimmunoprecipitation (RIPA) buffer (1× PBS, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, protease inhibitor cocktail, 0.2 mM PMSF, and 0.5 mM DTT) at 0.6 ml/100-mm culture plate followed by gentle rocking for 15 min at 4°C. Adherent lysed cell remnants were scraped into 0.3 ml of RIPA buffer and combined with the original lysate into a single microcentrifuge tube and the supernatant fraction was collected at 10,000 × g for 10 min at 4°C.

DNA- and RNA-binding Assays

Several synthetic biotinylated oligonucleotide probes were used in this study as follows: 1) a pyrimidine-rich, 30-base segment of single-strand DNA encompassing the reverse strand of the 5′-flanking region of the mouse SMαA gene located between -195 and -164 (also referred to as D_R, 5′-ccctcgtcttgtctccttacgtcaccttctct-3′) and previously shown to bind YB-1 with high affinity; Cogan et al., 1995; Kelm et al., 1999a); 2) the RNA coding element located in exon 3 of SMαA mRNA (CE-RNA, 5′-gggaguaaugguuggaaugggccaaaaaga-3′), previously shown to bind YB-1 and Pur proteins (Kelm et al., 1999b); and 3) its mutant counterpart (CE-RNAmut, 5′-uugaguaaugguuuuccguggccaaccaga-3′). Underlined sequences denote regions changed by mutation. Reaction mixtures containing protein extract (100 μg of protein) and biotinylated oligonucleotides (100 pmol; Integrated DNA Technologies, Coralville, IA) were incubated in a buffer containing poly(dI-dC), 10 mM Tris, pH 7.5, 50 mM NaCl, 0.5 mM DTT, 0.5 mM EDTA, 0.12 mM PMSF, 4% glycerol. Protein-biotinylated DNA complexes were captured on streptavidin-immobilized paramagnetic particles (Promega, Madison, WI; 0.6 ml/reaction, 30-min incubation) as described previously (Cogan et al., 2002; Subramanian et al., 2004). After washing four times with buffer containing 25 mM Tris-HCl, pH 7.5, 1 mM EDTA, and 100 mM NaCl, bound protein was eluted using 2× protein denaturing buffer and analyzed by SDS-PAGE and immunoblotting procedures.

Mammalian Protein Overexpression Plasmids, Cell Transfection, and Reporter Gene Assays

Fibroblasts at 40-50% confluence were transfected with mixtures of the previously described SMαA promoter:reporter fusion plasmids (VSMP4 and VSMP8) plus plasmids encoding various transcriptional regulatory proteins (Cogan et al., 2002; Subramanian et al., 2004). The Mirus (Invitrogen, San Diego, CA) transfection reagent was used with a protocol provided by the manufacturer. VSMP4 contains multiple SMαA promoter regulatory elements including two different transcriptional activation elements referred to as SPUR (Subramanian et al., 2004) and THR (Cogan et al., 2002). VSMP8 has been widely used to generate transgenic mouse lines and contains a 3.6-kb 5′-flanking/first intron promoter fragment from mouse SMαA genomic DNA containing both smooth muscle-specific developmental control and tissue injury response elements (Wang et al., 1997, 1998; Maeda et al., 1999; Lopes et al., 2003; Hassanain et al., 2005). For translation assays, the P4/CE (3′) construct was made by incorporating the exon 3 coding sequence of the mouse SMαA gene into VSMP4 as described previously (Kelm et al., 1999b). The exon 3 coding element previously was shown to repress chloramphenicol acetyl-transferase (CAT) reporter gene mRNA translation, but not transcription, when placed in the 5′-untranslated region of the CAT reporter gene at a position 3′ to the transcriptional start site. Plasmids encoding the human Smad 2, 3, and 4 proteins were kindly provided by Drs. L. Choy and R. Derynck (University of California, San Francisco, San Francisco, CA). Plasmids were purified using QIAGEN preparative resin and a protocol provided by the manufacturer (QIAGEN, Valencia, CA). Forty-eight hours after transfection, cells were washed three times with ice-cold PBS and then lysed using CAT enzyme-linked immunosorbent assay (ELISA) lysis buffer (Roche Diagnostics, Indianapolis, IN). Whole cell extracts were clarified at 14,000 × g for 10 min at 4°C and briefly stored at -20°C before immunoassay. Total protein in extracts was determined by BCA colorimetric assay (Pierce Chemical, Rockford, IL), and reporter gene activity was determined using a commercial CAT ELISA kit (Roche Diagnostics) and expressed on a per microgram protein basis. Transfections were performed in triplicate and repeated three to five times. Data sets were subjected to analysis of variance to assess statistical significance set at p < 0.05.

Immunoblot, Northern Blot, and Immunofluorescence Microscopy Procedures

Proteins (10-μg aliquots) were size-fractionated by SDS-PAGE using 10% polyacrylamide gels and then electrophoretically transferred to nitrocellulose membranes (Whatman Schleicher and Schuell, Keene, NH). After overnight blocking at 4°C in Tris-buffered saline (TBS) (25 mM Tris-HCl, pH 7.5, and 150 mM NaCl) containing 3% (wt/vol) nonfat dry milk and 0.5% bovine serum albumin (BSA), blots were incubated with selected rabbit polyclonal antibodies (1-2 μg/ml) for 90 min at room temperature with gentle rocking. Antibodies specific for Smad proteins were obtained commercially (Santa Cruz Biotechnology, Santa Cruz, CA) and YB-1/MSY1- and Pur α/β-specific rabbit polyclonal antibodies (anti-MSY1 M85-110 or anti-MSY1 M276-302 and anti-Pur P42-69, respectively) have been described previously (Kelm et al., 1999a). Antibodies for detection of p42/p44 Erk proteins and their phosphorylated derivatives were obtained from Upstate Biotechnology (Lake Placid, NY). Blots were washed four times at room temperature over a 20-min period in TBS containing Tween 20 [0.05% (vol/vol)]. Horseradish peroxidase-conjugated, goat anti-rabbit secondary antibody (1:1500) was then applied for 45 min after which time the blots were washed as described above and processed for antibody visualization by chemiluminescence (GE Healthcare, Piscataway, NJ) and imaged onto Biomax film (Eastman Kodak, Rochester, NY). For Northern blots, cells were extracted with TRIzol reagent (Invitrogen), and 10-μg aliquots of total RNA were size fractionated by formaldehyde/agarose denaturing gel electrophoresis, diffusion blotted onto nitrocellulose membrane, and hybridized using a radiolabeled actin cDNA probe that recognizes all actin isoform mRNAs (Strauch and Reeser, 1989). For densitometric evaluation, blots were exposed to a phosphor screen for analysis in a PhosphorImager (GE Healthcare). To examine YB-1 and SMαA localization after thrombin activation (5 U/ml; 5 min), hPFBs were washed in PBS, fixed in 2% paraformaldehyde for 30 min, permeabilized in 0.1% Triton X-100 for 15 min, preblocked with 1.5% goat serum for 30 min, and then incubated for 1 h with primary antibodies diluted in 3% BSA (anti-MSY1 M276-302 at 7 μg/ml or anti-human SMαA mouse monoclonal clone 1A4; [DakoCytomation California, Carpinteria, CA], at 1:100 dilution). After washing in PBS, samples were treated with goat anti-mouse secondary antibody conjugated with Alexa 594 (for SMαA detection) or goat anti-rabbit secondary antibody conjugated with Alexa 568 (for YB-1 detection). Secondary antibodies were obtained from Molecular Probes (San Diego, CA) and used at 1:1000 dilutions. Some samples also were briefly treated with 4,6-diamidino-2-phenylindole (DAPI) (1:1000; Molecular Probes) to visualize cell nuclei. Cells were viewed under epifluorescence illumination optics using an Axioscope microscope equipped with a 40× Neofluar objective (Carl Zeiss, Thornwood, NY).

RESULTS

Differential Action of TGFβ1 and Thrombin on SMαA Expression in Human Pulmonary Myofibroblasts

Thrombosis and fibrosis, mediated by the actions of thrombin and TGFβ1, respectively, are tissue repair processes that minimize blood loss and facilitate construction of a myofibroblast-enriched, provisional extracellular matrix needed to close open wounds. Of particular interest were possible similarities between TGFβ1- and thrombin-based SMαA gene induction required for conversion of stromal fibroblasts into contractile smooth muscle-like myofibroblasts. As a first approach, we compared the effects of TGFβ1 and thrombin on the behavior of single strand-specific, nucleic acid-binding proteins first identified by our laboratories as essential negative regulators of SMαA gene transcription in rodent myofibroblasts and smooth muscle cells (Cogan et al., 1995; Sun et al., 1995; Kelm et al., 1999a). SMαA gene expression in primary cultures of hPFBs exhibited a robust, 9-fold increase in response to TGFβ1 over the same concentration range previously demonstrated to be effective for gene stimulation in mouse embryonic fibroblast and neonatal rat aortic smooth muscle cell lines (Figure 1A). TGFβ1 also stimulated more than a 10-fold increase in transcription from reporter genes containing a 191-base pair mouse SMαA enhancer element harboring TGFβ1-activating elements (VSMP4) as well as a larger 3.6-kb 5′-flanking/first intron promoter fragment (VSMP8) from mouse SMαA genomic DNA (Min et al., 1990) that contains additional smooth muscle-specific developmental control elements (our unpublished data). Similarly, transcription of SMαA mRNA from the native human SMαA gene was markedly enhanced in hPFBs by TGFβ1 administered over a 16-h period (Figure 1B). In contrast, Northern blots prepared from hPFBs exposed to either 2 or 10 U/ml thrombin showed only a very modest increase in nonmuscle β- and γ-actin mRNA levels between 2 and 8 h before settling back to a low baseline level of expression after 16 h (Figure 1B). Despite its rather unremarkable effect on net SMαA mRNA accumulation, however, thrombin administered to hPFB at doses between 2-10 U/ml caused a rapid and significant increase in SMαA protein level (Figure 1C). Evaluation of Western blots prepared from hPFBs treated with 5 U/ml thrombin plus the metabolic inhibitors actinomycin D or cycloheximide indicated a requirement for active SMαA protein biosynthesis rather than de novo SMαA mRNA transcription (Figure 1C). Together, the data indicated that rapidly elevated SMαA protein synthesis in hPFBs after exposure to thrombin seemed to be based largely on a posttranscriptional control mechanism involving highly efficient utilization of a small pool of available SMαA mRNA.

TGFβ1 Dissociates YB-1 Repressor from the Activated SMαA Promoter

The recent literature contains several reports implicating the YB-1 single strand-specific, nucleic acid-binding protein, or its murine counterpart MSY-1, in control of both gene transcription and mRNA translation (Capowski et al., 2001; Esnault and Malter, 2003; Kohno et al., 2003; Skabkin et al., 2004; Matsumoto et al., 2005). With regard to elevated SMαA gene expression observed in human myofibroblasts, we discovered that the transcription- and translation-promoting activities of TGFβ1 and thrombin, respectively, were associated with a marked change in YB-1 subcellular distribution and nucleic acid binding activity. As shown in Figure 2, the physical interaction of hPFB YB-1 with its cognate pyrimidine (Pyr)-rich, single-stranded DNA (ssDNA) binding element from the SMαA enhancer decreased markedly within 30 min after exposure to 5 ng/ml TGFβ1. Further reduction of YB-1 interaction with DNA was evident during the next 3 h followed by a very slight recovery at 6 h. Reduced ssDNA binding occurred during the 60-min interval when little change was noted in the level of available YB-1 protein in unfractionated hPFB nuclear protein extracts, suggesting that loss of YB-1-ssDNA interaction in TGFβ1-activated hPFBs was not due to simple dilution of the available protein pool.

Reduced interaction between YB-1 protein and SMαA enhancer DNA in TGFβ1-activated human myofibroblasts supported our previous findings showing that cold-shock domain proteins function to repress SMαA gene transcription (Kelm et al., 1999a; Carlini et al., 2002). However, our previous studies on mouse embryonic fibroblasts containing the MSY-1 homologue of human YB-1 provided only indirect evidence for repressor function because the experimental approach was designed simply to reveal a general increase in transcriptional activity when point mutations known to block MSY-1 binding to Pyr-rich, ssDNA were placed into the context of SMαA enhancer:reporter fusion genes (Cogan et al., 1995; Sun et al., 1995). The binding of other SMαA gene repressors, Purα and Purβ, also was inhibited in these mutant constructs thus potentially obscuring the specific contribution of MSY-1 to gene repression. To examine the regulatory properties of MSY-1 more directly, nonhuman primate COS7 kidney fibroblasts were transfected with the TGFβ1-responsive mouse SMαA enhancer construct VSMP4 plus a mammalian expression plasmid encoding full-length mouse MSY-1 cDNA. As shown in Figure 3A, overexpression of mouse MSY-1 protein in simian COS7 fibroblasts was sufficient to block baseline transcription from the VSMP4 enhancer consistent with its action as a SMαA gene repressor. Additional results presented in Figure 3A indicate that combined overexpression of Smad proteins 2, 3, and 4 caused more than a twofold increase in SMαA enhancer activity in COS7 fibroblasts as well as significantly offset the inhibitory effect of mouse MSY-1 protein overexpression on enhancer transcriptional activity. These observations were an important extension of our recent studies showing that Smads also can partly relieve repression of the SMαA enhancer by the Purα protein (Subramanian et al., 2004). To examine how Smad protein overexpression influenced the distribution and DNA-binding properties of endogenous YB-1 protein, we examined nuclear, cytosolic, and DNA-bound protein fractions by Western blot analysis. Activation of the SMαA enhancer by overexpression of Smads 2, 3, and 4 during a 48-h period was accompanied by a decrease in the level of YB-1 protein in COS7 nuclei with concomitant accumulation in the cytoplasmic fraction (Figure 3B). Depletion of YB-1 from the nucleus also was accompanied by decreased interaction with its cognate, Pyr-rich ssDNA enhancer-binding element. Interestingly, ssDNA-YB-1 interaction decreased rapidly subsequent to Smad protein overexpression and was substantially reduced in parallel with depletion of YB-1 from the nucleus. Nuclear uptake of Smads 2 and 3 exhibited a notable lag phase and was not significantly increased until 24-48 h after transfection (Figure 3B) when a significant increase in VSMP4 enhancer activity was noted (our unpublished data). The results imply that interruption of YB-1 interaction with single-strand SMαA enhancer DNA and nuclear export were early responses to Smad protein overexpression in transfected fibroblasts.

Thrombin Disrupts Interaction of YB-1 and the Pur α/β Corepressor Proteins with Cytoplasmic SMαA mRNA

Smad 2-, 3-, and 4-dependent activation of SMαA enhancer transcription in fibroblasts was associated with redistribution of YB-1 between the nuclear and cytosolic protein pools. We next examined whether the ability of thrombin to activate SMαA gene expression and myofibroblast differentiation (Bogatkevich et al., 2001) also involved redeployment of YB-1 protein but in a manner that would be functionally consistent with translational rather than transcriptional control. YB-1 and related cold-shock domain family members are known to influence RNA-based processes in several ways, including unfolding secondary structural elements in mRNA as well as recruiting other proteins required for ribonucleoprotein (RNP) packaging, transport, turnover, and/or translation (Kohno et al., 2003). Of initial interest was how thrombin influenced YB-1 subcellular distribution during SMαA gene activation and whether modulation of YB-1 pools by thrombin was mechanistically related to the reduction in nuclear YB-1 level observed in TGFβ1-activated hPFBs. We also were interested in examining how the single strand-specific nucleic acid binding proteins, Purα and Purβ, behaved in the presence of thrombin because not only do these proteins collaborate with YB-1/MSY-1 to repress SMαA gene expression (Kelm et al., 1999a; Carlini et al., 2002), they also have been identified as important linker proteins for tethering mRNA to microtubules and motor proteins as well as transport of 1000S mRNA:protein granules in other cell types (Ohashi et al., 2002; Kanai et al., 2004). For specifically analyzing how thrombin might influence translational control in myofibroblasts, we used a novel VSMP4:CAT reporter fusion construct containing a segment of the SMαA mRNA coding element (CE) located in exon 3 that has structural similarity to the MCAT-THR enhancer within the SMαA promoter including a high degree of purine and pyrimidine asymmetry. The salient feature of this construct is that the CE blocks translation when placed 3′ to the transcriptional start site of the CAT reporter gene (Kelm et al., 1999b). Constitutive transcriptional activity of this construct in fibroblasts is mediated by the VSMP4 component and unaffected by the inserted CE(3′) module. However, translation of the CAT-CE(3′) mRNA transcript is markedly impaired due to high-affinity association of available YB-1/MSY-1, Purα, and Purβ with the CE(3′) motif. In this capacity, the CAT-CE(3′) fusion represents a useful tool for assessing how agents such as thrombin alter mRNA translational efficiency in a YB-1/MSY-1/Purα/Purβ binding-dependent manner. More than a 3.5-fold increase in CAT protein synthesis was observed when CAT-CE(3′)-transfected hPFBs were exposed to 5 U/ml thrombin over a 10-min period (Figure 4A). Posttransfection treatments with thrombin were designed to be of short duration to specifically assess de novo translation of preexisting CAT mRNA. Pulmonary fibroblasts transfected with the CE(3′) reporter gene construct were maintained for 48 h before the 10-min treatment with thrombin to allow for steady-state accumulation of CAT-CE(3′) mRNA. Direct assessment of YB-1 and Pur protein interaction with CE-RNA also was evaluated in both the hPFB nuclear and cytosolic protein pools after short-term exposure to thrombin. As shown in Figure 4B, thrombin specifically disrupted binding between CE-RNA and the YB-1 and Pur proteins in the cytosolic pool. Interestingly, interaction of the CE-RNA with these proteins in the nuclear compartment was markedly enhanced by thrombin over this same time period. Protein interaction with SMαA RNA was sequence specific because transversion mutations introduced into the CE-RNA at positions held in common with SMαA enhancer sites required for ssDNA-specific binding of YB-1 and Pur proteins eliminated all binding in both the nuclear and cytosolic compartments (our unpublished data).

Rapid Deployment of SMαA Thin Filaments and Nuclear Uptake of YB-1 in Myofibroblasts Is Mediated by Thrombin-specific Signaling

Net redistribution of YB-1 from the cytosol to the nucleus occurred rapidly in thrombin-treated hPFBs. Immunofluorescence microscopy was used to examine spatial relationships between the subcellular distribution of YB-1 and assembly of SMαA thin filaments in thrombin-treated hPFBs. Within 5 min after exposure to thrombin, diffusely distributed cytosolic YB-1 was consolidated within myofibroblast nuclei (Figure 5). Brief exposure to thrombin also resulted in a striking increase in deployment of SMαA thin filament networks in hPFBs. Together with the results of biochemical fractionation and SMαA mRNA binding studies shown in Figure 4, the morphological analysis of thrombin-activated myofibroblasts implied that relocalization of YB-1 from the cytosol to nuclear compartment was temporally linked to rapid assembly of the SMαA cytoskeleton in myofibroblasts. Because thrombin has been implicated as one of the several cellular protease capable of activating latent TGFβ1 stores in stromal cells, we performed additional studies to evaluate the relative contribution of TGFβ1- and thrombin-specific receptor signaling in myofibroblast SMαA protein accumulation. As shown in Figure 6, inclusion of an inhibitor of the type I TGFβ1 receptor serine/threonine kinase in the culture medium (SB431542) substantially reduced SMαA protein accumulation in TGFβ1-treated hPFBs, but it had no grossly restrictive effect on induction by thrombin. Further supporting an independent molecular mechanism of action, a thrombin serine protease inhibitor attenuated SMαA accumulation in thrombin-treated hPFBs but did not prevent induction in the presence of TGFβ1 (Figure 6). The ability of thrombin to rapidly augment SMαA expression in hPFBs does not seem to involve amplification of an underlying TGFβ1-dependent process such as activation of latent TGFβ1 or increased mRNA transcription but instead seems to have its basis in a thrombin-specific action such as enhanced efficiency of SMαA protein synthesis and/or thin filament stability. In this regard, phosphorylated extracellular signal-regulated kinase (ERK) proteins have been shown to mediate thrombin PAR-1 receptor-based signal transduction and promote rapid elevation of smooth muscle-specific myosin gene expression in thrombin-treated vascular smooth muscle cells (Reusch et al., 2001). As shown in Figure 7A, the Ras-Raf-(mitogen-activated protein kinase kinase (MEK)-ERK pathway inhibitors U0126 and PD98059 both blocked nuclear uptake of YB-1 and increased the size of the cytosolic YB-1 pool relative to baseline levels seen in control hPFB preparations. MEK1 inhibition by U0126 also partly blocked accumulation of SMαA protein in thrombin-treated hPFBs while causing a notable reduction in the level of phosphorylated p44/p42 ERK signaling intermediary proteins in hPFBs (Figure 7B).

DISCUSSION

In this report, we described a novel role for the YB-1 nucleic acid-binding protein in coordinating SMαA gene expression in human pulmonary myofibroblasts. YB-1 contains a highly conserved cold-shock domain that mediates nucleic acid binding in response to cellular stress signals, including low temperature, drug toxicity, reactive oxygen signaling, and UV damage (Matsumoto and Wolffe, 1998) (Kohno et al., 2003). In addition to SMαA, several other genes important in wound healing and cellular proliferation and survival also are regulated by YB-1, including those encoding α1 and α2 (I) collagen (Norman et al., 2001; Higashi et al., 2003), the B chain isoform of platelet-derived growth factor (PDGF) (Steinina et al., 2000), MMP-2 (Mertens et al., 1997), vascular endothelial growth factor (Coles et al., 2004), granulocyte macrophage-colony stimulating factor (Coles et al., 2000), p21 and p53 (Kohno et al., 2003), and the Fas death receptor (Lasham et al., 2000). In our recent studies of myofibroblasts and arterial smooth muscle cells, we discovered that MSY-1, the YB-1 homologue in rodents, was able to repress SMαA gene transcription by binding the pyrimidine-rich strand of the MCAT enhancer located in an inverted repeat between positions -190 and -150 of the SMαA promoter (Kelm et al., 1999a; Carlini et al., 2002). Alternative base pairing within the inverted repeat produces a thermodynamically stable stem-loop structure. The pyrimidine-rich, reverse-strand component of the loop binds YB-1 thus potentially disrupting promoter interaction with transcriptional activators such as Smad 2, Smad 3, Sp1, and TEF-1 that require duplex DNA proximal to the MCAT enhancer (Cogan et al., 2002; Subramanian et al., 2004). The MCAT enhancer region also exhibits substantial chromatin conformational change in response to TGFβ1 (Becker et al., 2000). Relationships between YB-1, RNA binding, and thrombin signaling also were noted and of interest because thrombin activation of PDGF-B gene transcription in human endothelial cells has been shown to require YB-1 proteolytic processing and nuclear translocation (Steinina et al., 2000).

Results reported here extend findings of TGFβ1-regulated interplay between Sp1, a GC-rich SMαA promoter element called SPUR, and the Pur family of single-strand nucleic acid binding proteins in myofibroblasts (Subramanian et al., 2004). Although YB-1 also exhibits TGFβ1-dependent interplay with the SMαA promoter, its action is centered on the MCAT enhancer located ∼120 base pairs upstream from SPUR. However, the ability of YB-1 and Pur repressor proteins to self-associate and bind diverse transcriptional activators, including Sp1, serum response factor, TEF-1, and Smad proteins (Kelm et al., 1999a, 2003; Subramanian et al., 2004) implies that they do not function in complete physical isolation. Although originally identified as repressors of SMαA gene transcription, YB-1 and Pur proteins also have roles in RNA processing and transport of RNPs along cytoskeletal filaments in the cytoplasm (Ruzanov et al., 1999) (Ohashi et al., 2000, 2002; Stickeler et al., 2001; Kohno et al., 2003; Kanai et al., 2004). The YB-1 protein family member p50 is one of the most abundant components of ribonucleoprotein particles in eukaryotes and has a notable mRNA-dependent affinity for muscle and nonmuscle actin filaments (Ruzanov et al., 1999). Excessive amounts of p50 in RNPs effectively blocked both translation and F-actin interaction, whereas high mRNA:p50 ratios were associated with active polyribosomal RNPs and RNP-actin filament complexes. Titration studies using human recombinant YB-1 and atomic force microscopy revealed that RNPs with a low YB-1: mRNA ratio were unfolded and better configured for efficient polyribosomal binding and translation (Skabkin et al., 2004). Published reports also indicated that C-terminal proteolytic processing (Steinina et al., 2000) may influence YB-1 shuttling dynamics between the nuclear and cytoplasmic compartments thus providing another control point for governing protein function. Pur proteins also have been identified in RNPs where they form links with motor proteins such as kinesin and myosin Va and IIB in neuronal cells (Ohashi et al., 2000, 2002; Kanai et al., 2004). Pur protein-RNPs are thought to couple microtubules with kinesin motor proteins to transport mRNAs from their sites of synthesis in cell bodies to distal polyribosomes located near synaptic terminals.

As RNA-binding proteins, YB-1 and Pur proteins perform unique tasks in the cytosol that complements their more conventional roles as nuclear transcriptional regulatory proteins. In this regard, myofibroblasts rapidly assemble a polarized SMαA thin filament system capable of producing directional forces necessary for reshaping the provisional extracellular matrix and guiding wound contraction. We speculate that myofibroblasts use YB-1 and Pur proteins to coordinate the transcription, transport, and compartmentalization of SMαA mRNA in a manner that facilitates efficient construction and deployment of a polarized smooth muscle-like contractile apparatus at sites of tissue injury. In this scheme, nuclear YB-1 and Pur proteins that become released from SMαA enhancer DNA in TGFβ1-activated myofibroblasts may no longer function as transcriptional repressors but instead assist as chaperones in the packaging, transport, and translation of SMαA mRNA. RNA-based actions of YB-1 and Pur proteins seem to be directed by thrombin signaling that likely coexists with TGFβ1 signaling within the complex microenvironment of the healing wound. Transience is a key feature of SMαA gene expression and thin filament deployment in myofibroblasts and we speculate that thrombin-directed release of YB-1 and Pur proteins from cytosolic RNPs and subsequent nuclear reentry may serve this requirement by rerepressing the SMαA gene and terminating myofibroblast differentiation. Although requiring additional experiments, this hypothetical scheme constitutes a regulatory loop whereby SMαA gene output in TGFβ1-activated myofibroblasts initially is amplified at the translational control level, then quenched by thrombin (Figure 8). The magnitude of SMαA protein output in myofibroblasts subsequent to transcriptional activation by TGFβ1 may be rate-limited by thrombin-dependent shuttling of YB-1 and Pur proteins off cytosolic mRNA and onto enhancer DNA in the nucleus. The Ras-Raf-MEK-ERK signaling pathway stimulated by thrombin may be especially relevant in attenuation of TGFβ1 signaling because ERK-mediated phosphorylation of the Smad 2 or 3 linker region has been shown to prevent nuclear accumulation of these proteins (Kretzschmar et al., 1999). ERK signaling in pulmonary myofibroblasts might also influence ribosomal protein phosphorylation and increase the net rate of protein synthesis during wound healing similar to its role in mechanical force-injured cardiac fibroblasts (MacKenna et al., 1998). On the other hand, thrombin may employ ERK to counteract TGFβ1 action on myofibroblast differentiation via induction of the Egr-1 DNA-binding protein known to activate promoters for proinflammatory tumor necrosis factor-α (Verrecchia and Mauviel, 2004) and interleukin-1β (Tan et al., 2003) that both have well characterized anti-fibrotic, matrix-destructive properties. Related studies on endothelial cells showed that thrombin in some cases can counteract TGFβ1 signals because a PAR-1 peptide agonist caused TGFβ1 receptor internalization and suppressed Smad 2 and 3 nuclear translocation (Tang et al., 2005). Together, the available evidence suggests that thrombin may provide an important physiological rheostat designed to promote rapid construction of the contractile provisional matrix in healing wounds in parallel with hemostasis while preventing maladaptive tissue fibrosis due to chronically unchecked TGFβ1 signaling. A key question to be addressed by future experimentation is to determine whether the thrombin-mediated nuclear reentry of YB-1 and Pur proteins amplifies myofibroblast TGFβ1-dependent differentiation by facilitating the transport and translation of SMαA mRNA, terminates the TGFβ1 signal by enabling YB-1 and Pur proteins to reaccess their cognate repressor sites in the SMαA enhancer, or both. Chronic diseases associated with excessive accumulation of myofibroblasts, faulty SMαA gene expression, and hypertrophic scarring such as idiopathic pulmonary fibrosis, chronic obstructive pulmonary disease, kidney epithelial mesenchymal transition, and transplant arteriosclerosis all may exhibit dysregulation in the coordination of thrombin and TGFβ1 signaling as a common predisposing condition.

Although the mechanistic details will require further study, disruption of YB-1 binding to the MCAT enhancer in the SMαA promoter seems to be an early response to TGFβ1 treatments. SMαA gene activation required Smad binding to multiple CAGA sites distributed within the large 40-base pair segment of DNA that contains the seven-base pair MCAT enhancer (Subramanian et al., 2004). There are three CAGA motifs between -190 and -150, but only two base pairs of one CAGA overlaps with the MCAT site. Thus, although loss of YB-1 binding could improve Smad access to one of its cognate sites in the promoter, the positions of the other two sites may render them less susceptible to YB-1 interference. Indeed, Smad binding is observed in a different region of the -190/-150 segment called the THR (TGFβ1 hyperreactivity region) but not at the MCAT where YB-1 binds (Subramanian et al., 2004). Chromatin conformation analysis before and after TGFβ1 also revealed that the MCAT-THR region unfolds during myofibroblast differentiation (Becker et al., 2000). Although the molecular basis for unfolding is not known, competitive interplay between YB-1 and Smads does not necessarily have to occur on promoter DNA but instead may involve protein complexes assembled in the cytosol that indirectly influence protein subunit availability or function in the nucleus. Data presented in Figure 3 imply that changes in YB-1 interaction with promoter DNA as well as the reduction in nuclear level of YB-1 both precede nuclear accumulation of Smad 2 and 3. In this regard, there is evidence that the anti-TGFβ1 properties of interferon-γ derive in part from its ability to foster binding of YB-1 to Smad 3 (Higashi et al., 2003). Within the complex, YB-1 prevents Smad 3 from linking with its p300 histone acetyl-transferase coactivator thus blocking transcription of the type α2(I) collagen gene in fibroblasts. Understanding dynamic interplay between YB-1, Smads, and MCAT-THR DNA will require more detailed examination of protein complex formation in the cytosol as well as protein-DNA interaction and assessment of DNA secondary structure in the nucleus. SMαA gene activation may involve more sophisticated control than can be provided by simple replacements between repressors and activators on promoter DNA.

In summary, YB-1 represses SMαA gene transcription and binds the MCAT region of the SMαA promoter previously shown to undergo chromatin unfolding in TGFβ1-activated myofibroblasts. Transcriptional activation of the SMαA gene by TGFβ1 was accompanied by loss of YB-1 binding to the MCAT enhancer and nuclear export. In contrast, thrombin stimulated SMαA gene activation at the translational level and caused YB-1 and Pur proteins to dissociate from translational repression sites on mRNA in parallel with rapid SMαA protein synthesis. Because thrombin by itself had no significant activating effect on SMαA mRNA transcription, it will be important in future studies to establish whether thrombin-dependent nuclear reentry of YB-1 is designed to rerepress the SMαA gene and/or destabilize preexisting SMαA mRNA thus contributing to termination of both the TGFβ1 activation signal and myofibroblast differentiation. In a normally healing wound, myofibroblasts are transient participants and active forms of TGFβ1 and thrombin are likely to coexist. The ability of TGFβ1 and thrombin to independently exploit the DNA- and RNA-binding properties of YB-1, respectively, adds a new dynamic perspective to control of gene expression during myofibroblast differentiation that may reveal strategies for therapeutic management of these cells in chronic fibrotic diseases that affect the heart, lung, liver, and kidney.

FOOTNOTES

FOOTNOTES

ACKNOWLEDGMENTS