Mouse hearts subjected to repeated transplant surgery and ischemia-reperfusion injury develop substantial interstitial and perivascular fibrosis that was spatially associated with dysfunctional activation of fetal smooth muscle α-actin (SMαA) gene expression in graft ventricular cardiomyocytes. Compared with cardiac fibroblasts in which nuclear levels of the Sp1 and Smad 2/3 transcriptional-activating proteins increased markedly after transplant injury, the most abundant SMαA gene-activating protein in cardiomyocyte nuclei was serum response factor (SRF). Additionally, cardiac intercalated discs in heart grafts contained substantial deposits of Purα, an mRNA-binding protein and known negative modulator of SRF-activated SMαA gene transcription. Activation of fetal SMαA gene expression in perfusion-isolated adult cardiomyocytes was linked to elevated binding of a novel protein complex consisting of SRF and Purα to a purine-rich DNA element in the SMαA promoter called SPUR, previously shown to be required for induction of SMαA gene transcription in injury-activated myofibroblasts. Increased SRF binding to SPUR DNA plus one of two nearby CArG box consensus elements was observed in SMαA-positive cardiomyocytes in parallel with enhanced Purα:SPUR protein:protein interaction. The data suggest that de novo activation of the normally silent SMαA gene in reprogrammed adult cardiomyocytes is linked to elevated interaction of SRF with fetal-specific CArG and injury-activated SPUR elements in the SMαA promoter as well as the appearance of novel Purα protein complexes in both the nuclear and cytosolic compartments of these cells.
- smooth muscle actin
- cardiac fibrosis
- cardiac transplant
- gene transcription
transition of stromal cardiac fibroblasts into smooth muscle-like myofibroblasts is an early aspect of wound healing (16, 20, 52) and fibrosis during native heart disease and after cardiac transplant (9, 19, 24, 38, 39, 42). Perivascular fibrosis has been described as a significant pathobiologic feature of chronic rejection in heart transplant recipients (2, 3) that may limit long-term graft survival. The accumulation of smooth muscle α-actin (SMαA) cytoskeletal microfilaments in myofibroblasts combined with increased deposition of interstitial type I collagen by these cells could disrupt myocardial perfusion and evoke cardiomyocyte stress responses (21) including activation of fetal contractile protein genes and cellular hypertrophy (1, 9, 28, 46). Shape deformation of cardiac fibroblasts and cardiomyocytes by accumulation of rigid scar tissue further stimulates expression of receptors for the pro-fibrotic agonists angiotensin II and transforming growth factor (TGF)-β1 with downstream activation of signaling intermediaries including Smads, Erk, Jak2, and Akt that can amplify both myofibroblast differentiation and dysfunctional cardiomyocyte responses (43, 45).
We previously showed that de novo activation of SMαA gene transcription was an early indicator of myofibroblast differentiation and cardiomyocyte fetal gene reprogramming in accepted murine heart allografts (47, 48). SMαA gene activation in myofibroblasts involved TGF-β1-regulated interplay of the Pur α/β transcriptional repressors within a 200-bp region of the promoter that contains essential binding sites for the TEF1, Sp1, serum response factor (SRF), and Smad 2/3 transcriptional activators (10, 11, 30, 49, 57). While reactivation of the developmentally silenced SMαA gene in cardiomyocytes has been described in rodent models of heart failure (8), very little is known about its molecular control, particularly in the context of cardiac transplant, fibrosis, and chronic rejection pathobiology. To explore this question, we used a syngeneic mouse heart graft model involving two rounds of transplant surgery and ischemia-reperfusion injury (53) that causes extensive interstitial and perivascular fibrosis and activation of fetal SMαA gene expression in cardiomyocytes. Compared with cardiac fibroblasts where retransplant was linked to increased nuclear levels of Sp1, Smad 2, and Smad 3, the most abundant SMαA gene-activating protein in cardiomyocytes was SRF. In parallel studies on perfusion-isolated, mechanically unloaded adult rat cardiomyocytes, we observed collaborative interaction between cardiac SRF and the Purα SMαA gene repressor protein at the SPUR cis-regulatory promoter element previously shown to mediate TGF-β1 responsiveness in myofibroblasts (49). Moreover, there was notable accumulation of Purα at cardiac intercalated discs that were located proximal to nascent SMαA-enriched thin filaments. Dynamic interplay between the SRF activator and Pur repressor proteins at the CArG and SPUR transcriptional regulatory sites represents a new, possibly rate-limiting step for activation of fetal SMαA gene expression in stressed cardiomyocytes after cardiac transplant.
MATERIALS AND METHODS
Murine syngeneic heart transplantation.
Five-week-old, specific pathogen-free female FVB/N mice were purchased from Harlan Sprague-Dawley (Indianapolis, IN) or from the Jackson Laboratory (Bar Harbor, ME) and were maintained in a barrier facility for use as heart donors and graft recipients. Donor hearts were heterotopically transplanted into the abdomen of recipient mice using an adaptation of the method of Corry et al. (13) as described in detail in our earlier reports (2–4). Three groups of three syngeneic cardiac grafts (isografts) were harvested 15 days after transplantation, retransplanted into another cohort of FVB/N recipients, and maintained for a second 15-day, posttransplant observation period, during which time grafts were explanted on days 3, 7, 11, and 15 for immunohistochemical and biochemical analyses. Controls were nontransplanted donor hearts from age- and sex-matched FVB/N mice or isografts subjected to only a single round of transplant surgery into FVB/N recipients and harvested at either 15 days or 26 days after transplant. To prepare protein extracts for DNA-binding or immunoblot studies, explanted or nontransplanted control hearts were visualized under a dissection microscope and trimmed to remove the atria and coronary vessels. Trimmed ventricles were frozen in liquid nitrogen, ground to powder while submerged under liquid nitrogen using a mortar and pestle, and extracted as described below. To prepare FVB/N donor mice harboring a SMαA promoter:green fluorescent protein (GFP) reporter transgene (VSMP8EGFP), the enhanced GFP gene sequence (BD Biosciences Clonetech, Palo Alto, CA) was ligated to the 1.5-kb mouse VSMP8 promoter (37, 55) and microinjected into mouse embryos using standard protocols used by the Transgenic Mouse Facility in the Neurobiotechnology Center at The Ohio State University. Mice heterozygous for the VSMP8EGFP transgene were identified in litters by PCR analysis of genomic DNA and used as donors for heterotopic cardiac transplant. The investigation was performed using Institutional Laboratory Animal Care and Use Committee-approved protocols and conformed with the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health (NIH Publication No. 85-23, revised 1996).
Antibodies specific for Sp1, SRF, Smad 2/3, and phosphorylated Smad 2 were obtained commercially (Santa Cruz Biotechnology, Santa Cruz, CA, or Upstate Biotechnology, Lake Placid, NY) and the YB-1-, Pur α-, and Purβ-specific rabbit polyclonal antibodies (anti-YB1 M85–110, anti-YB1 M276–302, anti-Purα 291–313, and anti-Purβ 302-324) were developed in the laboratory of R. J. Kelm, Jr. as described previously (30–32). Immunohistochemistry was performed on 4-μm-thick, paraffin-embedded sections from 4% paraformaldehyde-fixed isografts or nontransplanted donor hearts (47, 48). Sections were deparaffinized, rehydrated, blocked with 2% goat serum, 0.1% BSA, and 0.05% Tween 20 in PBS (blocking solution), and incubated for 90 min at 37°C with the primary antibodies at a concentration of 2 μg/ml in blocking solution. After being washed, sections were incubated for 30 min with a horseradish peroxidase (HRP)-conjugated, goat anti-rabbit secondary antibody, and a Vectastain protocol was followed for color development using a DAB peroxidase kit (Vector Laboratories, Burlingame, CA). To detect SMαA, HRP-conjugated anti-human SMαA mouse monoclonal (clone 1A4, Dako Cytomation California, Carpinteria, CA) was used at 1:100 dilution. Dual localization of two antigens was accomplished by incubating tissue sections with a second antibody after the detection protocol was completed for the first antibody and the sections were washed in Tris-buffered saline (TBS), pH 7.4. To distinguish between the two antibody reaction products, a nickel solution was added to the second color development reagent, which produced a blue-black precipitate that contrasted with the red-brown reaction product obtained using the first developing solution without nickel ions. Sections were viewed on a Zeiss Axioscope 40 microscope using ×10 and ×20 brightfield objectives, and color images were digitally recorded with Zeiss MRGrab 1.0 software.
Isolation, culture, and transfection of mechanically unloaded cardiomyocytes.
Highly enriched preparations (>95% pure) of adult rat ventricular cardiomyocytes were isolated from female rats (Sprague-Dawley-Ivanovas, 300–350 g) by retrograde perfusion of hearts with collagenase type 2 (281 U/ml; Worthington Biochemical, Freehold, NJ) according to the method reported by Eppenberger-Eberhardt et al. (18). After perfusion, the tissue was minced and incubated at 37°C for another 10 min with KB medium (26). The enzyme-digested heart tissue suspension was filtered through a single layer of cheesecloth and then centrifuged at 2,500 rpm for 1 min. Cell pellets were resuspended and plated in M199 medium for 4 h at 37°C to allow fibroblast attachment to the substrate and removal. The now enriched, nonadherent, and rhythmically beating cardiomyocytes then were transferred to 1% laminin-coated, 100-mm tissue culture dishes (Upstate) in M199 medium supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin, and 10 μM cytosine arabinoside. Medium was changed every third day over the course of a 15-day observation period. Isolated cardiomyocytes were transfected with a SMαA promoter:chloramphenicol acetyl transferase (CAT) reporter fusion plasmid (VSMP4) plus mammalian transcription factor expression plasmids using the Lipofectamine 2000 (Invitrogen, San Diego, CA) transfection reagent and protocol provided by the manufacturer (11, 12, 49). Plasmid encoding the full-length human SRF protein was kindly provided by Dr. M. Gupta, University of Chicago (15). The His-tagged Pur α and Pur β protein expression plasmids have been described previously (10). Plasmids used for transfections were purified using Qiagen preparative resin. Forty-eight hours after transfection, cells were washed three times with cold PBS and then lysed using CAT ELISA lysis buffer (Roche Applied Science, Indianapolis, IN) and clarified at 14,000 g for 10 min at 4°C. Total lysate protein was determined by BCA colorimetric assay (Pierce Chemical, Rockford, IL), and reporter gene activity was determined using a commercial CAT ELISA kit (Roche Applied Science) and expressed on a per μg 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. Transfection efficiency in isolated cardiomyocytes was assessed using the pmax-GFP plasmid (Amaxa, Gaithersburg, MD). The pmax-GFP (1 μg) was combined with 40 μl Lipofectamine 2000, and after a 20-min incubation period at room temperature, the mixture was added to cardiomyocyte culture preparations. GFP-positive cardiomyocytes were counted using immunofluorescence microscopy, and transfection efficiency was determined to be 40% under the experimental conditions used in our study.
Isolated cardiomyocytes or frozen, pulverized ventricles were extracted for 30 min on ice using 0.6 ml RIPA buffer containing 1× PBS, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, protease inhibitor cocktail, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 0.5 mM dithiothreitol (DTT) and were then centrifuged at 10,000 g for 10 min at 4°C to remove insoluble debris. Aliquots containing 2–4 μg of protein were size-fractionated by SDS-PAGE on 10% polyacrylamide gels and then electrophoretically transferred to nitrocellulose membranes (Schleicher & Schull, Keene, NH) for 90 min at 300 mA in 25 mM Tris·HCl, 192 mM glycine, and 20% (vol/vol) methanol. After overnight blocking at 4°C in (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, blots were incubated with primary antibodies (described above in Immunohistochemical Methods) at 1–2 μg/ml for 90 min at room temperature with gentle rocking. Blots were rinsed four times at room temperature over a 20-min period in TBS containing Tween 20 (0.05% vol/vol; TBST). HRP-conjugated secondary antibody (1:1,500) then was applied for 45 min, and the blots were washed as above and processed for antibody visualization by chemiluminescence (ECL, Amersham Biosciences, Arlington Heights, IL) and imaged onto Biomax film (Eastman Kodak, Rochester, NY). For detection of SMαA, membranes were incubated with HRP-conjugated 1A4 (1:100) in TBST. Following a 60–90 min incubation at ambient temperature with gentle mixing, the membrane was washed 2–3 times in TBST and developed using reagents provided in a DAB-peroxidase Vectastain kit (Vector Laboratories).
Preparation of protein extracts for DNA binding assays.
Cardiomyocytes were washed twice with Dulbecco's PBS, scraped into fresh PBS, sedimented at 3,000 rpm, washed once more in PBS, and resuspended in 8 packed-cell volumes of ice-cold, low-salt buffer (20 mM HEPES, pH 7.9, 25% glycerol, 1.5 mM MgCl2, 20 mM KCl, 0.2 mM EDTA, protease inhibitor cocktail, 0.2 mM PMSF, and 0.5 mM DTT) using a Dounce homogenizer with a type B pestle. Frozen, pulverized heart graft powder was resuspended in the same low-salt buffer and similarly processed using a Dounce homogenizer. 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 the total volume was added, and lysates were further extracted with gentle rocking for 30 min at 4°C. After dialysis against 50 volumes of dialysis buffer containing 10 mM Tris, pH 7.5, 50 mM NaCl, 0.5 mM DTT, 0.5 mM EDTA, 0.2 mM PMSF, and 20% glycerol, supernatants were collected by centrifugation at 14,500 rpm for 20 min, assayed for protein concentration using the BCA colorimetric method, aliquoted, and stored at −80°C for use in DNA-binding assays. Synthetic biotinylated oligonucleotide probes used in the present study correspond to various essential regulatory sequences present in the mouse SMαA promoter: 1) double-stranded SPUR DNA, 2) SPUR variants harboring mutations in either the GC-rich Sp1-binding domain or GGA Pur protein-binding domain, and 3) double-stranded probes harboring the CArG B binding site for SRF in both native and mutant contexts (12, 49, 50). Protein extract (100 μg protein) and biotinylated oligonucleotides (100 pmol; Integrated DNA Technologies, Coralville, IA) were combined in a binding buffer containing poly(dI-dC), 10 mM Tris, pH 7.5, 50 mM NaCl, 0.5 mM DTT, 0.5 mM EDTA, 0.2 mM PMSF, and 4% glycerol, and protein-biotin-DNA complexes were captured on streptavidin-immobilized paramagnetic beads (Promega, Madison, WI; 0.6 ml/reaction, 30 min incubation) as previously described (11, 49). After being washed four times with buffer containing 25 mM Tris·HCl, pH 7.5, 1 mM EDTA, and 100 mM NaCl, bound protein was eluted from the beads using 2× protein denaturing buffer and evaluated by SDS-PAGE immunoblot.
Protein immunoprecipitation methods.
SRF:Pur protein interaction was evaluated in protein extracts prepared from transfected fibroblasts. Non-human primate COS7 fibroblasts were maintained in Dulbecco's modified Eagle's medium (4.5 g/l d-glucose) supplemented with penicillin-streptomycin and 10% fetal bovine serum, cultivated in a humidified incubator at 37°C at 5% carbon dioxide, and transfected upon reaching 40–50% confluence. Nuclear or whole cell protein extracts were prepared 48 h after transfection with mammalian protein expression plasmids (11, 49). For immunoprecipitation of recombinant His epitope-tagged Pur protein, extracts were combined with 2 μg of a commercial His-tag antibody (Sigma Chemical, St. Louis, MO) in 200 μl of solution D (20 mM HEPES, pH 7.9, 100 mM KCl, 0.2 mM EDTA, 20% glycerol, 1 mM DTT, and 0.2 mM PMSF). After a 1-h incubation at 4°C on a rotating-platform mixer, a 20-μl aliquot of protein G-agarose (Sigma), which had been previously washed and suspended in solution D, was added, and the mixture was incubated 16 h at 4°C with rotation. Agarose beads with immobilized protein were collected by centrifugation at 2,500 rpm for 5 min at 4°C, washed 4 times with PBS, suspended in 1× SDS-PAGE sample buffer, and heated at 95°C for 5 min, and released proteins were evaluated by SDS-PAGE and immunoblotting with selected antibodies as noted in the text. Aliquots of whole protein extract were evaluated by immunoblot before immunoprecipitation to verify protein overexpression in transfected fibroblast preparations. To monitor interaction between native forms of SRF and Pur proteins in protein extracts prepared from perfusion-isolated cardiomyocytes, we used a rabbit anti-SRF primary antibody (Santa Cruz) in combination with a secondary antibody detection system and manufacturer's protocol that avoids interference from immunoglobin heavy and light chains that are coprecipitated with the SRF-Pur protein complexes (Rabbit IgG TrueBlot Set, eBioscience, San Diego, CA).
Accepted, major histocompatibility complex (MHC) class I and II mismatched FVB/N (H-2k) or DBA/2 (H-2d) donor hearts in gallium nitrate-immunosuppressed C57BL6 (H-2b) recipients developed chronic rejection with substantial fibroplasia within 60 to 90 days after transplant (Fig. 1A) . Of interest, allograft myofibrils that were in close proximity to regions of interstitial fibrosis in the left ventricle free wall and septum often were strongly reactive with a SMαA-specific antibody (Fig. 1B). Myofibrils in isografts (FVB/N to FVB/N or DBA/2 to DBA/2) did not exhibit SMαA antibody reactivity as described in our previous works (47, 48). The SMαA gene encodes a fetal contractile protein isoform that normally is not expressed in adult mammalian cardiomyocytes (36, 54). Cardiomyocytes in allografts from transgenic FVB/N donor mice harboring the tissue-injury responsive mouse SMαA promoter (VSMP8) ligated to an enhanced GFP reporter gene showed robust fluorescence (Fig. 1D) that was most pronounced in regions of substantial interstitial fibrosis (Fig. 1C). The behavior of the reporter transgene in cardiomyocytes following allotransplant essentially recapitulated the histopathologic response of the native SMαA gene and was not active in donor hearts or nonfibrotic isografts (data not shown). Although very little is known about the mechanism that governs fetal actin gene reactivation in the heart, the reporter gene data shown in Fig. 1D suggested that transplant chronic rejection activated the same SMαA promoter elements previously shown to be required for TGF-β1-dependent myofibroblast differentiation during wound healing (6, 11, 49, 51).
To better define relationships between wound healing, fibroplasia, and fetal SMαA gene reactivation in cardiomyocytes, we used an experimental model first developed by Tullius and Tilney (53) that segregates alloimmune-mediated inflammatory responses from the innate wound healing process. The approach involves subjecting syngeneic cardiac isografts to two cycles of transplant surgery. Evaluation of “two-hit” FVB/N to FVB/N mouse heart isografts between 3 and 15 days after retransplant revealed widespread interstitial fibrosis (Fig. 2A) and increased SMαA protein expression in the ventricular myocardium (Fig. 2B) that was not observed in isografts subjected to a single round of heterotopic transplant [data not shown (47)]. Moreover, expression of the SMαA gene transcriptional activators Sp1, Smads 2 and 3, and phosphorylated Smad 2 previously shown to be highly enriched in TGF-β1-activated myofibroblasts in healing wounds all were substantially higher in cardiac fibroblast nuclei in two-hit hearts compared with nontransplanted donor hearts or isografts subjected to a single round of transplantation (Fig. 3). Smad proteins were noted as early as 3 days after second transplant. In contrast, SRF, an essential SMαA gene activator and muscle differentiation factor, was present in donor hearts at the time of transplant, but staining appeared slightly more intense 15 days after retransplant (Fig. 3). Importantly, SRF appeared most abundant in larger cardiomyocyte nuclei compared with the Sp1 activator that was completely confined to interstitial cardiac fibroblast nuclei.
The observed nuclear colocalization of Sp1 and Smads in cardiac fibroblasts after retransplant injury was of interest because we recently showed that the TGF-β1-regulated Smad 2 and 3 proteins induce SMαA gene transcription during myofibroblast differentiation by altering functional interplay of the Sp1 activator and Purα transcriptional repressor proteins at a cis-regulatory region of SMαA promoter DNA called SPUR (49). However, colocalization of these SMαA gene activators in cardiac fibroblasts did not provide insight about the mechanism of SMαA gene reprogramming in graft cardiomyocytes that seems to be associated with enhanced nuclear localization of SRF rather than Sp1. To examine whether SRF was capable of binding SMαA promoter fragments that could potentially mediate the cardiomyocyte response to retransplantation injury, we performed DNA-binding assays using ventricle extracts prepared at various times after retransplant in combination with a DNA probe containing the essential CC(A/T-rich)6GG consensus motif (CArG B) needed for activation of the fetal SMαA gene in embryonic muscle cells (12, 50). As shown in Fig. 4, SRF was detected on immunoblots prepared from nontransplanted donor hearts, and its level in retransplanted hearts increased somewhat over the 11-day posttransplant observation period. Importantly, SRF binding to the CArG B promoter fragment increased about threefold after retransplant, becoming statistically significant by day 11 (P < 0.01) relative to binding observed in nontransplant control hearts (Fig. 4). In some retransplanted hearts, SRF:CArG B interaction was slower to evolve during the initial 7-day posttransplant recovery period, but significantly higher SRF binding was always evident by day 11. The increase in SRF:CARG B interaction at day 11 corresponded to the time when SMαA protein was maximally expressed in two-hit heart grafts, exceeding the level observed in nontransplanted hearts by about 2.7-fold (Fig. 4). We also studied the accumulation and DNA-binding activity of Pur proteins that have recently been described as mediators of fetal cardiac myosin expression in heart failure patients and negative regulators of SRF-mediated SMαA gene expression in vascular smooth muscle cells (10, 22, 32). While simply expecting that Purα and Purβ levels might be reduced as SMαA accumulated in retransplanted grafts, we instead observed a 2.7-fold increase in total Pur protein availability and a sixfold increase in SPUR DNA-binding activity within 7 days after retransplant relative to levels observed in nontransplanted hearts (Fig. 4). Total Pur protein binding to SPUR generally peaked 3 days after retransplant, with some variable reduction in DNA binding occasionally noted during the subsequent 8-day observation period. However, this decrease was noted in some two-hit grafts despite a sustained increase in the size of the Pur protein pool for the duration of the 11-day study.
Our previous immunohistologic studies revealed Purβ protein localization in the interstitial, fibroblast-enriched zones of DBA/2 cardiac allografts between 30 and 60 days after transplant, but reactivity of the Purα-specific antibody in those particular preparations was only weakly diffuse (47). However, new insight on Purα distribution was obtained in the two-hit isograft model in which expression of this protein was especially pronounced between 7 and 11 days after second transplant and was localized not only in the cardiac interstitium but also more substantially at the intercalated discs of cardiac myofibrils (Fig. 5). Intercalated disc localization appeared specific for Purα because Purβ repressor staining in two-hit isografts was mostly confined to cell nuclei within the cardiac interstitium as noted previously in our studies of chronically rejected cardiac allografts (data not shown). A significant number of Purα-positive intercalated discs were flanked by short SMαA-positive thin filament bundles suggestive of a functional link between Purα protein extranuclear localization and de novo remodeling of the cardiomyocyte actin cytoskeleton (Fig. 5). While both Pur proteins form a TGF-β1-dissociable complex with Sp1 that governs SMαA gene output in myofibroblasts, the paucity of nuclear Sp1 expression in two-hit cardiomyocytes (Fig. 3) implied that Pur proteins might have different regulatory properties and/or protein partners in muscle cells. In this regard, we recently showed that SRF binds Purβ in vitro and neutralizes its ability to repress the SMαA promoter in differentiated rat aortic smooth muscle cells (32, 34). SMαA gene transcription is spontaneously reactivated in collagenase perfusion-isolated, mechanically unloaded adult cardiomyocytes, providing a potentially useful in vitro model system for examining details of Pur and SRF protein interplay (18). Freshly isolated and highly enriched (>95% purity) preparations of rod-shaped adult rat cardiomyocytes continued to beat in serum-containing medium but dedifferentiated over a 14-day period into stellate, neonatal-like cells that reexpressed substantial amounts of fetal SMαA protein (Fig. 6, A and B). Upon transfection, freshly isolated cardiomyocytes were unable to express a CAT reporter gene construct harboring the minimal cis-regulatory control sequences from the SMαA promoter (VSMP4) needed for myofibroblast transcriptional response to TGF-β1 (Fig. 6C). However, the ability of cardiomyocytes to activate VSMP4 increased significantly between days 7 and 10 after isolation, corresponding to the period of peak SMαA protein accumulation in these preparations. In accord with the finding that heart retransplant was not linked to increased nuclear localization of Sp1 in cardiomyocytes, we did not detect this transcriptional activator in freshly isolated cardiomyocytes compared with isolated cardiac fibroblasts where it was abundantly expressed (Fig. 7). Although a trace amount of Sp1 occasionally was detected at about day 7 in culture, cardiac fibroblast contamination is considered unlikely because these adherent cells are efficiently removed from heart perfusion isolates, and a cell proliferation inhibitor was used in the culture medium to prevent growth of any residual nonmyocyte contaminants that may have escaped removal.
Pur protein expression and DNA-binding activity in isolated cardiomyocytes exhibited complex behavior similar to that seen in extracts prepared from two-hit ventricular tissue. While readily observed in heart tissue extracts, Purβ was barely detectable in isolated cardiomyocytes (Fig. 8A). Purα was far more abundant, and although its level of expression was largely unchanged over the 14-day observation period, SPUR-binding activity of this protein actually decreased about fivefold between days 0 and 10 in culture (Fig. 8B). The reduced binding of Purα to SPUR during extended cardiomyocyte culture was of interest not only because it resembled the kinetic trend noted in SPUR-binding assays on intact two-hit hearts between 3 and 11 days after retransplant (refer to Fig. 4) but also corresponded to the period when SMαA protein accumulated in isolated cardiomyocytes (refer to Fig. 6B). In accord with immunohistologic evaluation of ventricular tissue sections showing prominent nuclear localization in cardiomyocytes, SRF was highly expressed by freshly isolated cardiomyocytes as well as reprogrammed myocytes following a 14-day culture period (Fig. 9A). However, a striking difference in SMαA promoter-binding activity was noted for SRF in each myocyte preparation. SRF in freshly isolated cardiomyocytes assayed at day 0 exhibited lower affinity for the functionally important CArG B consensus motif in the SMαA promoter compared with SRF in reprogrammed cardiomyocytes assayed at day 14 (Fig. 9B). These results are important because robust SRF interaction with CArG B also was observed in extracts prepared from late-stage two-hit hearts grafts that showed substantial fibrosis and SMαA protein accumulation (refer to Fig. 4). As noted earlier, SRF in extracts prepared from nontransplanted hearts did not bind with great avidity to CArG B, consistent with the transcriptionally inactive status of the fetal SMαA gene in normal adult cardiac muscle. Interestingly, four molecular size variants of SRF were detected on immunoblots prepared from perfusion-isolated cardiomyocytes (Fig. 9B) corresponding in size to the 67-kDa full-length form of SRF plus 57-kDa, 52-kDa, and 40-kDa alternatively spliced forms of SRF (Δ5, Δ4,5, and Δ3,4,5, respectively) described by others (7, 15). Although the functional significance of this observation is not known, all SRF variants observed in isolated cardiomyocytes were capable of binding the CArG B probe under our DNA-binding assay conditions and required an intact CC(A/T-rich)GG motif for binding (Fig. 9C). Curiously, the three SRF splice variants were not observed in binding assays performed on intact two-hit heart grafts (refer to Fig. 4).
SRF was able to activate the minimal promoter:CAT reporter fusion construct (VSMP4) when overexpressed in day 5–7 cardiomyocytes, whereas both Purα and Purβ functioned as potent transcriptional repressors of VSMP4, much like the behavior of these proteins in transfected fibroblasts and smooth muscle cells (Fig. 10). The relative paucity of Sp1 expression in both native (Fig. 3) and isolated SMαA-positive cardiomyocytes (Fig. 7) led us to examine the hypothesis that the more abundant SRF protein might substitute for fibroblast Sp1 in controlling transcriptional activation of the SMαA gene in cardiac muscle cells. Indeed, SRF:Pur complexes were readily immunoprecipitated from transfected COS7 non-human primate kidney fibroblasts that overexpressed SRF in combination with one or both Pur protein isoforms (Fig. 11A). In these same preparations, SRF blocked Pur protein-mediated repression of the SMαA promoter and was especially potent in neutralizing repression mediated by Purα (Fig. 11B). SPUR DNA provides a Pur protein-binding site that helps regulate SMαA gene transcription in myofibroblasts (49). Because CArG B and SPUR are in close spatial proximity within a 160-bp segment of nucleosomal DNA, we examined whether SPUR could support binding of the SRF:Pur protein complex in cardiomyocytes. SPUR DNA was able to bind SRF from reprogrammed cardiomyocytes, but not freshly isolated cells, and formation of the SPUR:SRF complex was dependent on the presence of an intact GGA Pur-binding site in the SPUR probe (Fig. 12A). Moreover, SRF was able to bind SPUR DNA only when both Pur proteins were co-expressed in transfected COS7 cells (Fig. 12b). The SPUR DNA probe lacks CC(A/T-rich)GG motifs needed to support prototypical SRF binding implying that this activator binds SPUR via a Pur protein bridge. Finally, an anti-Pur protein antibody was able to coimmunoprecipitate full-length SRF, but not the SRF splice variants, from extracts prepared from day 14 cardiomyocytes (Fig. 12C). Taken together, the data suggest that the formation of a novel nucleoprotein complex consisting of full-length SRF, Pur protein, and SPUR DNA may be associated with de-repression of fetal SMαA gene expression in reprogrammed adult cardiomyocytes.
The cardiac isograft retransplantation model used in the present study was designed to examine cellular and molecular aspects of wound healing in the absence of inflammatory cytokines and T cell-mediated chronic rejection. Since class I and II MHC antigens are identical between donor and host in this model, cellular and humoral alloimmune responses do not occur and the histopathology observed in “two-hit” isografts is driven largely by repeated peritransplant surgical trauma and ischemia-reperfusion injury. Whereas all heterotopic hearts grafts are mechanically unloaded, isografts subjected to only a single round of transplant surgery do not exhibit the rapid and substantial development of interstitial and perivascular fibrosis noted in two-hit isografts, nor do they contain reprogrammed cardiomyocytes that express fetal SMαA protein (4, 47, 48). Retransplantation injury specifically provokes heart graft remodeling that seems to first emerge in the cardiac stromal cell compartment. The early deposition of extracellular matrix protein in two-hit isografts is likely to produce abnormal biomechanical stress forces in the myocardium that not only amplifies profibrotic signaling in cardiac fibroblasts but also secondarily stress-activates expression of fetal genes in ventricular cardiomyocytes, leading to further dysfunctional remodeling. Biomechanical stress has been reported to stimulate expression of angiotensin II and TGF-β1 receptors that mediate profibrotic autocrine and paracrine signaling in both cardiac fibroblasts and myocytes (43). However, the molecular mechanisms that control changes in cardiomyocyte gene transcription after transplant are poorly understood. The main goal of work described in the present report was to identify relationships between the expression, localization, and dynamic interplay of transcription factors required for de-repression of the fetal SMαA gene in cardiomyocytes, a leading indicator of myocyte stress and dysfunction during cardiomyopathy.
We presented immunohistologic evidence showing that TGF-β1 receptor-regulated Smad proteins accumulated in cardiac stromal cells within 3 days after isograft retransplant and before peak expression of the SMαA gene. This observation is consistent with a recent report showing accumulation of TGF-β1 mRNA in chronically rejected murine heart grafts and extends those findings by showing peritransplant nuclear translocation of Smads in the cardiac interstitium need not necessarily require an alloimmune stimulus (14). Transcription of the SMαA and type I collagen subunit genes, two major indicators of myofibroblast differentiation and wound healing activity, is coordinately activated by TGF-β1 receptor-regulated Smad protein signaling (11, 17, 25, 40, 49). In studies on isolated fibroblasts, Sp1 and Smad transcriptional activators bind and neutralize the Purα and Purβ repressor proteins at the SPUR transcriptional control site in the SMαA promoter. SPUR DNA contains tandem GC-rich and GGA motifs that mediate Sp1 and Pur protein binding, respectively. Protein:protein interactions at SPUR DNA are regulated by TGF-β1 and appear to be essential for myofibroblast differentiation in vitro (49) and efficient dermal wound healing in the mouse (51). Colocalization of Sp1 and Smads in two-hit cardiac fibroblasts provides new in vivo evidence for the dynamic interplay model for SMαA gene activation during myofibroblast differentiation (49, 57).
As a secondary response to cardiac retransplant, the dormant fetal SMαA gene in cardiomyocytes becomes stress-activated as fibroplasia emerges in the left ventricle. The SMαA gene is one of several fetal-type genes that are coordinately regulated by SRF, a key transcriptional mediator of cardiac hypertrophic responses and progression to heart failure (7, 15, 56). Cardiomyocytes isolated from SRF null mice exhibit impaired expression of contractile protein genes and grossly disrupted sarcomeres and polysomes (5). SRF was the predominant SMαA gene-activating protein detected in cardiomyocyte nuclei from two-hit hearts. In normally developing myogenic cells, SRF binds to a canonical CC(A/T-rich)GG consensus motif present in virtually all contractile protein gene promoters including the SMαA promoter where it appears twice within the first 150 bp of the 5′-flanking region (CArG A and CArG B). The more upstream element (CArG B) is functionally more important for SMαA gene transcription in murine smooth muscle cells and myofibroblasts (12, 50). Importantly, SRF binding to CArG B was acquired only after retransplant and was not observed in nontransplanted adult mouse donor hearts or freshly isolated adult rat cardiomyocytes. However, SRF acquired CArG B-binding activity when cardiomyocytes were maintained in extended culture and thus may represent a new indicator for fetal gene reprogramming in cardiomyocytes during heart graft failure. The relative paucity of Sp1 expression in both two-hit and perfusion-isolated cardiomyocytes further suggested that SRF might play a major role in de novo activation of fetal SMαA gene transcription in these cells. In this regard, we discovered that cardiomyocyte SRF binds to the same activating site in the proximal SMαA promoter, referred to as SPUR, that was shown in myofibroblasts to be occupied by a TGF-β1-regulated complex consisting of the Sp1 activator and Purα/β protein repressors. However, Sp1 was not detected in freshly isolated cardiomyocytes and only modestly upregulated in reprogrammed cardiomyocytes after a 2-wk culture period. In contrast, the level and SPUR-binding activity of both SRF and Purα were highly elevated after transplant, although DNA-binding activity of Purα decreased somewhat after retransplant. We speculate that Pur protein binding to SPUR in cardiomyocytes might be governed by physical interaction with SRF rather than Sp1. Increased SRF binding to CArG B in stressed cardiomyocytes may influence Pur affinity for SPUR since both sites are located approximately on the same side of the DNA helix within nucleosome structural context (37). Coimmunoprecipitation data suggests that SRF may also form off-DNA complexes with Pur proteins that theoretically could modulate their ability to bind and/or repress transcription from SPUR DNA. Native SRF splice variants in isolated cardiomyocytes did not bind Pur proteins but were able to bind both CArG B and SPUR DNA, supportive of their possible role as dominant-negative transcriptional regulators in the heart (7). By itself, native Purα in cardiomyocyte or whole heart extracts does not seem to bind the SMαA gene-specific CArG B motif (X. Liu and A. R. Strauch, unpublished observations), although in vitro synthesized Purβ has been reported to exhibit some binding affinity for certain skeletal muscle actin-type CArG elements (23). SRF has many protein partners that generally fall into two cofactor families (56), the ets domain-containing ternary complex factors (TCF) and the myocardin-related transcription factors (MRTF). Pur proteins, however, are less well studied and are seemingly unique as SRF coactors, although they do bind an ets-like GGA motif favored by TCF cofactors and may indirectly bind actin filaments in a manner resembling MRTF family members (29, 41, 44).
Pur proteins function as SMαA gene repressors in transfected cardiomyocytes, much like their role in myofibroblasts. Curiously, formation of a SRF:Purα protein complex in transfected COS7 cells was associated with enhanced activation of the SMαA promoter. We hypothesize that SRF binds the NH2 or COOH termini of SPUR-immobilized Purα protein, causing the repressor's central DNA-binding domain to alter its position on or affinity for the GGA motif within SPUR. “Assisted removal” of Purα probably does not improve SRF interaction at SPUR since this DNA element lacks a CArG box motif needed to stabilize SRF binding. However, removal of Pur protein may improve physical access of SRF to the CArG B element located nearby on an adjacent nucleosomal DNA loop. Alternatively, the SRF:Purα protein complex may have unique activation properties that could stimulate transcription at SPUR in a manner resembling the SRF:myocardin or SRF:Elk-1 heteromeric complexes that operate CArG-dependent promoters in muscle cells (56). In this regard, Purα has been implicated as a transcriptional activator in certain cell types (27, 33). Increased loading of hypothetical SRF:Purα “superactivation” complexes onto SPUR DNA may be a beneficial response if stimulation of fetal SMαA gene transcription is an absolute requirement for cardiomyocyte survival and function after cardiac transplant injury. On the other hand, the observed increase in Pur protein binding to SPUR in two-hit heart grafts might represent an exuberant compensatory response to injury whereby stressed cardiomyocytes are attempting to silence a fetal gene that has become inappropriately activated. Finally, it is important to consider that the observed increase in the size of the Pur protein pool after retransplant as well as the ability of Purα to bind single-strand DNA and RNA (41) may reflect a novel role for this protein in packaging and transporting newly synthesized mRNA molecules needed for cardiomyocyte adaptation to stress and/or transplant injury (see below). Future studies should be directed toward identifying essential peptide determinants and biochemical signals needed for SRF:Purα and SRF:SPUR:Purα physical interplay during activation of the SMαA promoter in stressed cardiomyocytes and whether this interplay is specifically designed to override developmental restrictions that prevent fetal gene reactivation in the normal heart. We also are planning to examine the kinetics of SRF binding to SPUR DNA in long-term two-hit isografts. As currently implemented, the 15-day posttransplant end point used in the current study mainly captures early aspects of fibroplasia where myofibroblast Sp1 interactions with SPUR DNA may interfere with study of the SRF:SPUR interaction. More abundant cardiac myofibroblast-derived Sp1 in protein extracts prepared from two-hit grafts may dilute or compete with less abundant SRF for SPUR binding because cardiomyocyte remodeling is just emerging at this time and is often highly localized to specific regions of the left ventricle. We speculate that Sp1 may have a competitive advantage over SRF in binding SPUR DNA in pull-down assays because it binds directly to a 10-bp GC-rich consensus sequence, whereas SRF interaction with SPUR is indirect and seems to be mediated by a Pur protein bridge attached to the 3′-end of this site.
During reprogramming of SMαA gene expression in cardiomyocytes, the nuclear SRF activator accumulates at SPUR and CArG B DNA while Pur protein repressors are deployed to actin filament anchorage points at the cardiac intercalated discs. Pur proteins have been identified as important linker proteins for tethering mRNA to microtubules, microfilaments, and motor proteins as well as intracellular transport of ribonucleoprotein particles to sites of mRNA translation (29, 41, 44). The YB-1 binding partner of Pur proteins in fibroblasts and muscle cells (10, 30) is a well-known cardiac stress-response protein that has dual function as a SMαA gene corepressor in the nucleus and the ability to stimulate actin bundle formation in the cytosol (44) as well as bind components of the intercalated disc such as cardiac ankyrin repeat protein (58). In myofibroblasts, Pur proteins also avidly bind to a sequence motif in exon 3 of SMαA mRNA that blocks translation (31). Taken together, these findings suggest that Pur proteins may function posttranscriptionally as RNA chaperones to package and transport mRNAs necessary for expression of wound healing- and/or stress-associated gene products in the injured heart. The localization of Pur proteins at cardiac intercalated discs in two-hit isografts was especially pronounced and may be relevant to reconfiguration of cardiac α-actin thin filament ends to accept addition of newly synthesized SMαA monomers. Delivery of Pur protein:SMαA mRNA ribonucleoprotein complexes to polysomes located near actin thin filament ends anchored at intercalated discs may provide a means for coupling de novo translation of SMαA G-actin subunits with F-actin polymerization in stressed cardiomyocytes. The barbed end of F-actin is proximal to the intercalated disc and represents the fast site for addition of new actin monomers. While quite preliminary, the presence of short SMαA filamentous bundles emerging from Pur protein-enriched intercalated discs in two-hit isografts seems to provide some credence for this hypothesis. The specific involvement of Pur proteins in linking the mRNA- and actin filament-based aspects of posttransplant cardiac remodeling is an area of active investigation in our laboratory. In this regard, Pur protein appears to be less tightly bound to SMαA mRNA in SMαA-positive cultured cardiomyocytes compared with freshly isolated cells that do not express this α-actin polypeptide (A. Zhang, J. J. David, and A. R. Strauch, unpublished observations). These emerging data are consistent with our earlier reports showing that dissociation of Pur protein from SMαA mRNA enhances its overall translational efficiency (31, 57).
In summary, we have shown that posttransplant cardiomyocyte remodeling seems to involve conversion of an adult transcriptional program into one that more closely resembles the scheme used by injury-activated myofibroblasts. SRF may have evolved a unique role in stressed cardiomyocytes to de-repress the fetal SMαA gene and provide new types of actin subunits that are functionally more compatible with wound healing activity in the heart compared with cardiac α-actin that may be better specialized for high efficiency activation of sarcomeric myosin ATPase (35). In this scheme, SRF appears to neutralize Pur protein transcriptional repressors by forming a novel complex that may reduce repressor potency and/or DNA-binding affinity within the SMαA gene promoter. Redirection of Pur protein function after cardiac injury also may provide the means to export newly transcribed SMαA mRNA from the nucleus to sites of protein translation and de novo thin filament remodeling at cardiac intercalated discs.
The study was supported by National Heart, Lung, and Blood Institute Grants HL-70294 and HL-85109 (to A. R. Strauch) and HL-54281 (to R. J. Kelm).
↵† Charles G. Orosz is deceased.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society