In addition to mediating cell-to-cell electrical coupling, gap junctions are important in tissue repair, wound healing, and scar formation. The expression and distribution of connexin43 (Cx43), the major gap junction protein expressed in the heart, are altered substantially after myocardial infarction (MI); however, the effects of Cx43 remodeling on wound healing and the attendant ventricular dysfunction are incompletely understood. Cx43-deficient and wild-type mice were subjected to proximal ligation of the anterior descending coronary artery and followed for 6 days or 4 wk to test the hypothesis that reduced expression of Cx43 influences wound healing, fibrosis, and ventricular remodeling after MI. We quantified the progression of infarct healing by measuring neutrophil expression, collagen content, and myofibroblast expression. We found significantly reduced transformation of fibroblasts to myofibroblasts at 6 days and significantly reduced collagen deposition both in the infarct at 6 days and at 4 wk in the noninfarcted region of Cx43-deficient mice. As expected, transforming growth factor (TGF)-β, a profibrotic cytokine, was dramatically upregulated in MI hearts, but its phosphorylated comediator (pSmad) was significantly downregulated in the nuclei of Cx43-deficient hearts post-MI, suggesting that downstream signaling of TGF-β is diminished substantially in Cx43-deficient hearts. This diminution in profibrotic TGF-β signaling resulted in the attenuation of adverse structural remodeling as assessed by echocardiography. These findings suggest that efforts to enhance the expression of Cx43 to maintain intercellular coupling or reduce susceptibility to arrhythmias should be met with caution until the role of Cx43 in infarct healing is fully understood.
- infarct remodeling
- transforming growth factor-β
ventricular remodeling is a dynamic and time-dependent process that occurs in response to myocardial infarction (MI); it frequently results in cardiac dysfunction, can progress to heart failure, and is a major determinant of morbidity and mortality associated with coronary artery disease. Cell death post-MI initiates a reparative wound healing process that replaces the infarcted myocardium with scar tissue to maintain the structural integrity of the ventricle (57). This process requires the coordination of multiple cell types, including neutrophils and macrophages, which mediate the removal and replacement of injured tissue. Inflammatory cells release cytokines and growth factors required for fibroblast proliferation and differentiation (10, 19). After infarction, fibroblasts are responsible for orchestrating a complex remodeling process that balances the degradation of extracellular matrix components and collagen deposition (35). Fibroblasts are arguably the most important component of the “living scar” and are responsible for myocardial repair, infarct healing, and scar formation (56). When a large area of the myocardium is infarcted, the remodeling process typically results in left ventricular (LV) dilatation, hypertrophy of the noninfarcted myocardium, and myocardial contractile dysfunction (46).
Previously, we (64) have shown that native cardiac fibroblasts are functionally coupled by gap junctions containing both connexin (Cx)43 and Cx45 and that fibroblast proliferation is influenced by the expression level of Cx43. Cx43 is both downregulated and redistributed in the ventricular myocardium after MI (44). Many studies have supported the notion that a reduction in Cx43 likely underlies malignant arrhythmias in infarcted hearts (21, 44, 45). However, little is known about the relationship between the loss of Cx43 and ventricular infarct remodeling after MI, specifically whether reduced Cx43 expression influences the process of infarct healing.
Gap junctions are responsible for the cell-to-cell electrical coupling required for synchronized cardiac contraction. Gap junction channels permit the exchange of ions, small signaling molecules, and metabolites between adjacent cells (22). An increasing amount of evidence has suggested that gap junctions play an important role in the wound healing process, although the sometimes disparate results have underscored the complexities of wound healing and scar formation (8, 9, 18, 40, 49). Ehrlich and Diez (18) showed that inhibition of gap junction communication by heptanol disrupted wound repair, the development of granulation tissue, and scar maturation. Although fibroblast density at the edge of the wound was greater, there was less infiltration of fibroblasts into and a lower density of myofibroblasts within the granulation tissue (at 6 days) with reduced coupling. Similarly, reduced Cx43 expression limited neointima formation after vascular balloon injury by decreasing the inflammatory response and reducing smooth muscle cell infiltration and proliferation (8). However, Qiu et al. (49) showed that transient downregulation of Cx43 by a Cx43 antisense gel applied to incisional and excisional skin lesions not only reduced inflammation but accelerated wound closure. Mori et al. (39) confirmed that the downregulation of Cx43 in antisense oligonucleotide-treated skin wounds results in a reduced inflammatory response. In contrast to the findings reported by Ehrlich and Dietz (18), however, Mori et al. (39) observed increased migration of fibroblasts in the wound, enhanced formation of granulation tissue, and increased differentiation of fibroblasts to myofibroblasts (at 7 days). Differences in responses may be due to the fact that various processes, including inflammation, fibroblast differentiation, and cell proliferation, are likely affected by changes in gap junction channel coupling and Cx43 expression differentially.
One of the most important mechanisms responsible for directly inducing the transformation of myofibroblasts and other profibrotic processes is signaling through the activation of transforming growth factor (TGF)-β (5, 33). TGF-β pathways have pleiotropic effects, the majority of which are mediated via the activation of Smad proteins through phosphorylation (31, 54). Interestingly, Cx43 modulates TGF-β signaling through a competitive interaction with Smads (11). The present study was performed to test the hypothesis that a reduced baseline expression of Cx43 influences wound healing and ventricular remodeling after MI. The reduced level of Cx43 expression exhibited by Cx43-deficient mice in this study is comparable with that observed in patients with diabetes or nonischemic cardiomyopathy, for example, who subsequently could experience a MI (17, 27, 29, 34, 41). We found that although cell proliferation was increased, both the transformation of fibroblasts to myofibroblasts and collagen deposition were reduced in Cx43-deficient hearts. Our data demonstrate that the remodeling of Cx43 that occurs after coronary ligation influences fibroblast function and delays scar formation and cardiac wound healing post-MI. We present data that are consistent with the downregulation of downstream TGF-β signaling in Cx43-deficient hearts. Pharmacological modulation of gap junction function has been explored as a novel therapeutic approach to preventing malignant cardiac arrhythmias (28) and enhancing the efficacy of cell therapy post-MI (47, 50). However, clear understanding of the role of functional expression of Cx43 in ventricular remodeling is essential for developing safe and effective pharmacological agents designed to prevent morbidity and mortality post-MI by modulating gap junction function.
MATERIALS AND METHODS
Animals were handled in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All research presented here was conducted in conformity with the American Physiological Society's “Guiding Principles in the Care and Use of Animals.” All protocols were approved by the Washington University Animal Studies Committee.
Cx43+/+ and Cx43+/− mice.
We maintain a colony of wild-type (WT) mice (Cx43+/+) and mice heterozygous for a Cx43-null allele (Cx43+/−) for breeding in a standard barrier facility. Founder mice (C57BL/6J,129-Gja1tm1Kdr) were originally purchased from Jackson Laboratories (Bar Harbor, ME) and were subsequently backbred into a C57BL/6J background. The genotypes of all mice were determined by PCR as previously described (24).
MI and sham-operated controls.
All surgery was performed by the Mouse Cardiovascular Phenotyping Core of Washington University School of Medicine. The left anterior coronary artery was ligated (30) in Cx43+/+and Cx43+/− mice that were studied 6 days or 4 wk post-MI. Mice (10–20 wk old) were anesthetized with a mixture of ketamine and xylazine (87 and 13 mg/kg, respectively), intubated, and artificially ventilated. A left thoracotomy was performed, and the left anterior descending coronary artery was ligated. The chest was closed, and mice were allowed to recover. Mice that were subjected to the same surgical procedure without ligation of the coronary artery served as sham-operated controls.
All echocardiographic studies were performed by a single observer (A. Kovacs) who was blinded to genotype. Pre-MI echocardiographic examination was not performed. In a previous study (4), we found no differences in cardiac structure and function in Cx43-deficient mice compared with WT mice, even in advanced age. Echocardiographic assessment of post-MI remodeling was quantified by measuring LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), and ejection fraction (EF) on echo images that were acquired from subsets of Cx43+/+ and Cx43+/− mice with MI after 6 days and 4 wk using previously described and validated methods (25).
Histological analysis of infarct size.
Infarct size was measured in Cx43+/+ and Cx43+/− hearts 6 days and 4 wk post-MI as previously described (25). Investigators were blinded to genotype. Excised hearts were fixed in formalin, cut transversely into three roughly equal short-axis ventricular slices (apical, middle, and basal portions), embedded in paraffin, and sectioned at a thickness of 5 μm. Sections from the approximate midpoints of the three short-axis ventricular slices were stained with Masson's trichrome stain for the quantitative analysis of infarct size and structure. Images were analyzed offline using ImageJ software (NIH, http://rsb.info.nih.gov/ij/). The portions of each of the three short-axis ventricular sections occupied by the infarct scar, noninfarcted LV free wall, and entire interventricular septum from the LV to right ventricular chambers were traced, the areas were measured and summed, and infarct size was calculated as the total area occupied by the scar divided by the total ventricular area (infarcted plus noninfarcted areas). The area of unresorbed myocytes in the infarct area was measured using ImageJ and was divided by the infarct area to obtain the area of unresorbed necrotic myocytes as a percentage of the infarct area.
For Cx43 staining, tissue sections were blocked in PBS containing 0.1% Triton X-100 and 3% normal goat serum for 30 min and then incubated in mouse anti-Cx43 antibody (1:400, Chemicon) at 4°C overnight followed by washes and an exposure to Cy3-conjugated goat anti-mouse antibody (1:200, Jackson ImmunoResearch) for 2 h. For phospho-Smad (pSmad)2/3 staining, tissue sections were incubated in mouse anti-pSmad antibody (1:50, Santa Cruz Biotechnology) at 4°C overnight followed by washes and an exposure to Cy3-conjugated goat anti-mouse antibody (1:300, Jackson ImmunoResearch) for 2 h. Neutrophils were labeled with rabbit anti-myeloperoxidase (MPO) antibody (1:500, NeoMarkers, Fremont, CA). Myofibroblasts were labeled with anti-α-smooth muscle actin (α-SMA) antibody with horseradish peroxidase (HRP; used neat, Dako). For MPO and α-SMA immunostaining, tyramide signal amplification (NEN Life Science) was used. Briefly, after endogenous peroxidase activity was quenched with 3% H2O2, cells were incubated in blocking buffer (TNB) for 30 min, and sections were incubated overnight in primary antibody at 4°C. For MPO staining, EnVision+ anti-rabbit with HRP secondary antibody (neat, Dako) was used for 1 h. Sections were incubated in biotinyl tyramide amplification reagent (1:300, tyramide signal amplification, NEN Life Science) followed by Cy3-conjugated streptavidin (1:300, Jackson ImmunoResearch). For α-SMA staining, as stated above, the primary antibody contained HRP; after an overnight incubation in primary antibody, sections were incubated in biotinyl tyramide amplification reagent (1:300, NEN Life Science) followed by Cy2-conjugated streptavidin (1:500, Jackson ImmunoResearch). Sections were stained with Hoechst nuclear stain (1:10,000, Cambrex) for 10 min for the fluorescence visualization of individual nuclei.
pSmad immunostaining was quantified by analyzing the digital images in Adobe Photoshop CS2 using the red color histogram function. All hearts were stained in a single experiment. All images were acquired blinded at one sitting such that the same microscope and camera settings were used to acquire all images for a given experiment. Neutrophils were quantified by counting MPO positive (MPO+) cells per area (in mm2). Myofibroblast expression was quantified by calculating the area of the α-SMA-positive signal using Adobe Photoshop CS2.
One hour before euthanization, each mouse was given an injection of bromodeoxyuridine (BrdU; Sigma, 100 μg/g body wt) to mark proliferating cells as previously described (60). The heart and a segment of the small intestine (a cell proliferation control) were rinsed briefly in PBS. Hearts were fixed in formalin, cut transversely into three short-axis ventricular slices, and embedded in paraffin for sectioning and staining with an anti-BrdU antibody (1:400, Invitrogen) followed by Cy3-conjugated goat anti-mouse antibody (1:300, Jackson ImmunoResearch). Sections were stained with Hoechst nuclear stain (1:10,000, Cambrex) for 10 min for the fluorescence visualization of individual nuclei. BrdU-positive cells were counted and expressed as a percentage of the total number of cells obtained from counting total Hoechst-stained nuclei.
Sirius red staining.
Collagen content was determined by quantitative morphometry of Sirius red-stained sections. Sirius red staining was performed as previously described (63). Briefly, after being deparaffinized, sections were treated with 0.1% Sirius red (Sigma) solution for 30 min, rinsed with distilled water twice, and then treated with 0.1% fast green (Sigma) to stain noncollagenous protein. Multiple digital images were acquired from both the mid and apical slices of each heart (6 days post-MI) to create two composite images (at mid and apical levels) of the entire infarct region, which were then analyzed using Adobe Photoshop CS2. Similarly, multiple digital images were acquired from the basal, noninfarcted slice of each heart (4 wk post-MI) to create a composite of the remote, noninfarcted region, which was then analyzed using Adobe Photoshop. Collagen content was determined from the following equation: collagen content (%) = [Sirius red-positive stained area/(total tissue area − vascular lumenal areas)] × 100.
Cell extracts were prepared by homogenization (20 strokes) using a Duall 20 glass conical pestle (Kimble/Kontes, Vineland, NJ) and sonication (6 × 15 s, Cole-Parmer 8850, Cole-Parmer Instruments, Chicago, IL) in a buffer solution containing (in mM) 1 NaHCO3, 5 EDTA, and 1 EGTA (pH 8.0) with the following protease inhibitors: 1 μM pepstatin, 100 nM aprotinin, 1 mM benzamidine, 1 mM iodoacetamide, 1 μM leupeptin, and 1 mM PMSF. Proteins were resolved by SDS-PAGE (10%) and transferred to nitrocellulose membranes for incubation with rabbit anti-Cx43 (1:5,000, Zymed), mouse anti-αSMA with HRP (1:150, Dako), rabbit anti-TGF-β1 (1:200, Santa Cruz Biotechnology), or mouse anti-GAPDH (1:10,000, Fitzgerald) antibodies. Thirteen or 15% SDS gels were used for TGF-β immunoblots. Goat anti-rabbit and sheep anti-mouse secondary antibodies [1:20,000 (Jackson ImmunoResearch) and 1:10,000 (GE Healthcare/Amersham), respectively] were used except in the case of α-SMA blots, which did not require secondary antibody due to the presence of HRP. Protein bands were visualized using Western Lightning Chemiluminescence Reagent Plus (Perkin-Elmer, Boston, MA). Immunoblots were quantified by densitometric analysis using Adobe Photoshop.
All values are expressed as means ± SD. Two-group comparisons were made using Student's t-test for grouped data. ANOVA and a post hoc multiple-comparisons test were used for comparing three or more groups of data. P values of <0.05 were considered significant.
A total of 45 mice (21 WT and 24 Cx43+/− mice) were subjected to permanent ligation of the left anterior descending coronary artery. Five mice (3 WT and 2 Cx43+/− mice; 11%) died within 2 days of surgery. Of the 40 mice that completed the study, 25 mice were studied 6 days post-MI and 15 mice were studied 4 wk post-MI. There were no differences in the heart-to-body weight ratios of WT (0.52 ± 0.05) and Cx43+/− (0.52 ± 0.05) mice post-MI. The mean heart-to-body weight ratio for the sham-operated mice was 0.43 ± 0.04. The mean infarct size of the relatively large infarcts created in this study was not different in Cx43-deficient mice compared with WT mice 6 days post-MI (46 ± 4% vs. 46 ± 5%, respectively) or 4 wk post-MI (24 ± 3% vs. 25 ± 6%, respectively).
As expected, Cx43 immunostaining was reduced at intercellular junctions in Cx43-deficient hearts compared with WT hearts in either the presence or absence of MI. Both WT and Cx43+/− hearts exhibited remodeling of Cx43 that included downregulation and redistribution post-MI (not shown), as has been reported by numerous laboratories (26, 29, 44).
Scar formation and infarct remodeling are delayed in Cx43-deficient hearts.
Despite the fact that infarct size was not different in Cx43-deficient hearts compared with WT hearts, when examined in detail, significant differences were observed in the architecture and cellular activity of Cx43+/− infarcts. Analysis of trichrome-stained sections showed that the area of the unresorbed myocardium was significantly (P = 0.005) larger in Cx43+/− (40 ± 5%, n = 8) versus WT infarcts (17 ± 8%, n = 8) 6 days post-MI (Fig. 1A), indicating that the fibrotic replacement of the necrotic myocardium was delayed and/or protracted in Cx43-deficient hearts. In addition, we measured the number of MPO+ cells in the infarct region and observed significantly (P = 0.005) more MPO+ cells present in the infarct region of Cx43+/− (136 ± 30 MPO+ cells/mm2, n = 4) compared with WT (68 ± 27 MPO+ cells/mm2, n = 6) hearts 6 days post-MI (Fig. 1B), suggesting that the peak inflammatory response could, indeed, have been delayed in Cx43-deficient hearts. Dead, necrotic myocytes were almost completely replaced with fibrotic scar, and only a few MPO+ cells remained in the nearly fully matured infarct regions in both Cx43+/− and WT hearts 4 wk post-MI and were not quantified.
Consistent with the larger areas of the unresorbed myocardium in Cx43+/− infarcts, we also observed markedly less fibrosis and collagen deposition in Cx43+/− hearts post-MI. Of note, although we did not measure collagen content before MI, our previous study (4) demonstrated that there was no difference in the minimal amount of fibrosis present in Cx43+/− hearts compared with WT hearts at either 17 wk or 2 yr of age. We measured collagen content in the infarct regions of Cx43+/− and WT hearts by Sirius red staining 6 days post-MI and in the remote, noninfarcted regions as well as infarct regions 4 wk post-MI. Collagen deposition within the infarct region was significantly (P = 0.016) reduced in Cx43+/− (36 ± 6%, n = 8) versus WT (50 ± 13%, n = 6) hearts 6 days post-MI (Fig. 2A). As expected, after 4 wk, the infarct scars in both Cx43+/− and WT hearts consisted nearly entirely of collagen and were not different relative to genotype (Supplemental Fig. S1)1, consistent with the virtually identical infarct sizes. However, we postulated that a delay in infarct remodeling that was evident 6 days post-MI might further influence the development of fibrosis differentially in the noninfarcted regions. Accordingly, we quantified collagen deposition in the remote, noninfarcted areas in the hearts of Cx43-deficient and WT mice 4 wk post-MI. Interestingly, collagen deposition as assessed by Sirius red staining was significantly reduced in both the free wall (2.83 ± 0.57%, n = 8, P = 0.003) and septum (1.33 ± 0.07%, n = 3, P < 0.001) of Cx43+/− hearts compared with the corresponding areas (free wall: 6.71 ± 2.66%, n = 5, and septum: 2.69 ± 0.28%, n = 4, respectively) in WT hearts (Fig. 2B).
Cell proliferation is increased but transformation of fibroblasts to myofibroblasts is reduced in Cx43-deficient hearts 6 days post-MI.
After injury such as that occurring with MI, fibroblasts undergo a phenotype switch to myofibroblasts, which are the primary cell type responsible for collagen synthesis and the development of fibrosis in infarcted hearts. To determine whether the level of Cx43 expression might influence fibroblast function post-MI, and thereby be responsible for the altered infarct remodeling we observed, we assessed both cell proliferation after BrdU incorporation and myofibroblast production as reflected by α-SMA staining. Cell proliferation in the infarct region was increased significantly (P = 0.02) in Cx43+/− (6.1 ± 0.5%, n = 4) compared with WT (4.1 ± 0.7%, n = 4) hearts (Fig. 3A). On the other hand, the transformation of fibroblasts to myofibroblasts was reduced significantly (P = 0.002), as indicated by significantly less α-SMA-positive immunostaining in Cx43+/− (32 ± 16%, n = 8) compared with WT (57 ± 5%, n = 8) hearts (Fig. 3B). Thus, despite an increased proliferation of resident cells in Cx43-deficient hearts post-MI, the reduced conversion of fibroblasts to myofibroblasts in these hearts was consistent with reduced collagen deposition in the infarcted region of Cx43+/− hearts during scar maturation 1 wk post-MI and in the noninfarcted areas of Cx43+/− hearts undergoing postinfarct ventricular remodeling 4 wk post-MI (Fig. 2).
Beneficial postinfarct remodeling in Cx43-deficient hearts 4 wk post-MI.
Postinfarct remodeling leading to infarct expansion, alterations in LV chamber dimensions, and hypertrophy of noninfarcted regions occurs in humans post-MI. To determine whether the cellular and tissue alterations observed in Cx43-deficient hearts that we observed post-MI would translate to any structural or functional differences during and beyond the acute phase of inflammation and injury, we performed echocardiography at 6 days and 4 wk post-MI. Although LVEDV was significantly (P < 0.001) larger (indicating deleterious dilation) after MI (77 ± 23 μl, n = 25 WT plus Cx43+/−) compared with sham-operated mice (29 ± 5 μl, n = 5) and EF was significantly (P < 0.001) depressed after MI (22 ± 9%, n = 25 WT plus Cx43+/−) compared with sham-operated mice (64 ± 5%, n = 5), there were no differences in LVEDV, LVESV, or EF between WT and Cx43+/− hearts 6 days post-MI (Table 1). However, after 4 wk, consistent with reduced fibrosis in the remote, noninfarcted regions, adverse dilatation was attenuated (P = 0.054) in Cx43-deficient (91.3 ± 10.3 μl, n = 8) compared with WT (118.5 ± 34.1 μl, n = 5) hearts. Dilatation during systole as reflected by LVESV also tended to be less detrimental in Cx43-deficient (65.0 ± 11.9 μl) compared with WT (85.9 ± 32.0 μl) mice, but this difference did not reach statistical significance (P = 0.116). EF was not significantly different between groups (Table 1).
TGF-β signaling is reduced in Cx43-deficient infarct scars.
TGF-β is upregulated during MI; its signaling pathway plays an important role in infarct healing and ventricular remodeling (5). Because TGF-β is known to regulate fibroblast proliferation and differentiation, it is likely that TGF-β signaling influences not only the inflammatory response to MI but also the development of cardiac fibrosis and repair. Our data from Cx43-deficient hearts subjected to MI are consistent with a downregulation of the profibrotic effects of TGF-β. We hypothesized that either TGF-β protein itself was downregulated or TGF-β downstream signaling was impaired in Cx43-deficient hearts.
To address these possibilities, we first quantified TGF-β cytokine expression in immunoblots prepared from infarct and noninfarcted regions of Cx43+/− and WT hearts 6 days post-MI. The TGF-β ligand was significantly upregulated in the infarct region (P = 0.005–0.009 compared with WT or Cx43+/− noninfarcted regions or sham-operated mice). However, the upregulation of TGF-β was nearly identical in Cx43+/− (3.9 ± 1.5 density units, n = 4) and WT (4.0 ± 1.3 units, n = 3) hearts subjected to MI compared with Cx43+/− noninfarcted regions (1.0 ± 0.4 units, n = 4), WT noninfarcted regions (0.8 ± 0.5 units, n = 3), or sham-operated hearts (1.0 ± 0.1 units, n = 4; Fig. 4).
Dai et al. (11) reported that Cx43 can induce the release of Smad2/3 from binding sites on microtubules and effectively increase TGF-β signaling events. Consistent with this finding, they reported that knockdown of Cx43 resulted in more Smad2/3 binding to microtubules, resulting in less pSmad2 and subsequently less accumulation of Smad2/3 and Smad4 in the nucleus (decreasing TGF-β signaling events). We hypothesized that a similar mechanism related to competition between Cx43 and Smad2/3 for binding to microtubules may be operative in Cx43-deficient hearts. Thus, we measured pSmad immunostaining in the infarct and noninfarcted regions 6 days post-MI. As anticipated, pSmad expression was significantly (P = 0.018) decreased in Cx43+/− (53.6 ± 15.2 intensity units, n = 4) compared with WT (90.8 ± 19.8 units, n = 5) infarcts 6 days post-MI (Fig. 5). Furthermore, pSmad expression was significantly (P = 0.022) decreased in the remote, noninfarcted regions of Cx43+/− (57.3 ± 13.8 units, n = 4) compared with WT (90.0 ± 18.6 units, n = 5) hearts 6 days post-MI (Fig. 5). Together, these data suggest that in the presence of comparable levels of TGF-β protein, downstream signaling in the TGF-β pathway is impaired. Finally, pSmad expression was measured in the noninfarcted region at three levels (base, mid, and apex) and in the infarct region at two levels (mid and apex) 4 wk post-MI. There were no differences (P = 0.986) observed between Cx43-deficient and WT hearts (Supplemental Fig. S2), indicating that the mechanism for postinfarct ventricular remodeling at 4 wk postinfarction is likely to be pSmad independent.
Post-MI remodeling occurs in both the infarcted and noninfarcted myocardium and is responsible for favorable wound healing and compensatory responses to tissue loss on the one hand and the deleterious development of impaired ventricular function and heart failure on the other. Profibrotic processes modify the pathogenesis and prognosis of heart failure by contributing not only to contractile dysfunction but also to dysrhythmia and sudden death. In the heart, Cx43 expression and gap junction intercellular communication are best known for playing critical roles in maintaining electrical coupling and normal cardiac rhythm (58). However, they also play vital roles in inflammatory responses, cell proliferation, and wound repair (9, 18, 39, 49, 61). Substantial attention is currently being focused on efforts to enhance gap junction intercellular communication as a means for reducing susceptibility to ventricular arrhythmia. However, full understanding of how enhanced coupling may also influence ventricular remodeling could be overlooked in the exciting quest for promising new approaches to antiarrhythmic therapy. Attempts to target gap junctions to reduce the risk of arrhythmogenesis may have unanticipated effects. In the present study, we report that infarct remodeling is altered and ventricular dysfunction is attenuated in mice with reduced endogenous expression of Cx43 subjected to subsequent infarction.
Gap junction intercellular communication is involved in various stages of wound healing (2, 8, 9, 18, 39, 40, 49, 61), although differences in injury models, cell and tissue types, and connexin isoforms studied make interpretation of mechanistic contributions difficult. For example, downregulation of Cx43 significantly increases the rate of wound closure in models of skin injury (39, 49, 61). Specifically, in excised skin, downregulation of Cx43 resulted in a reduced inflammatory response and enhanced keratinocyte proliferation, which led to earlier wound closure (39). Wound closure and MI scar maturation are not directly comparable. However, of interest to our present work, Mori et al. (39) showed that downregulation of Cx43 reduced the inflammatory response that was induced after skin injury. Typically, the early inflammatory response, which peaks within the first 2–3 days post-MI, has already begun to wane by the sixth day post-MI. In the present study, however, more MPO+ cells and unresorbed myocardium persisted in the infarct region of Cx43-deficient hearts 6 days post-MI. Although we do not know whether the MPO+ cells represent neutrophils, monocytes, macrophages, or eosinophils, the delayed or sustained inflammatory response in Cx43-deficient infarcts occurred concomitant with a delay in the fibrotic replacement of damaged myocytes. Our data suggest that reduced Cx43 expression may delay the migration of inflammatory cells into the ischemic zone, thus delaying resorption of the necrotic myocardium. Mechanisms for delayed inflammatory cell infiltration may involve reduced leukocyte-leukocyte or leukocyte-endothelial cell contacts (42, 43) but were not the focus of the present study. Ultimately, it is not clear how important a delay in the early inflammatory response is to the final maturation of the scar. Of note, an inflammatory response is not necessary for wound healing in skin (16, 37). In the present study, reduced baseline expression of Cx43 did not influence scar formation at the later, 4-wk time point.
Cardiac fibroblasts play a pivotal role in maintaining and remodeling the extracellular matrix (3, 7, 20, 36). Previously, we (64) found that native murine ventricular fibroblasts express both Cx43 and Cx45 and that the level of Cx43 expression influences intercellular coupling and cell proliferation in culture. Specifically, reduced expression of Cx43 resulted in increased proliferation. In the present study, consistent with our previous result (64), cell proliferation was increased in Cx43+/− hearts, even after MI, compared with WT hearts. However, after MI, fibroblasts undergo a phenotypic switch to myofibroblasts that express α-SMA. Myofibroblasts are responsible for collagen synthesis and promote fibrosis in infarcted hearts (38, 52, 56, 62). Our present data demonstrate that although cell proliferation was enhanced, the transformation of fibroblasts to myofibroblasts was significantly reduced in Cx43-deficient hearts post-MI. Our data also show that a smaller myofibroblast population resulted in reduced cardiac fibrosis compared with that in WT hearts and that the decrease in fibrosis resulted in attenuated structural defects detected by echocardiography. These results suggest that reduced expression of Cx43 is an independent determinant of the cardiac fibrosis that develops after MI and that indiscriminate overexpression of Cx43 may adversely impact post-MI ventricular remodeling induced by persistent coronary occlusion.
TGF-β is upregulated dramatically in experimental models of MI (12, 14, 15) and plays multiple important roles in ventricular remodeling (5), notably through pSmad-mediated signaling pathways (6). TGF-β is a potent inducer of tissue fibrosis; stimulation of TGF-β induces the conversion of fibroblasts to myofibroblasts and enhances extracellular matrix protein synthesis (13, 33). Recent studies have shown that Cx43 competes with Smad2/3 or other proteins for binding to microtubules (11, 53). Specifically, elevated Cx43 can release Smad2/3 from microtubules, resulting in its phosphorylation and subsequent increases in downstream signaling (11). In addition, Cx43 reportedly contributes to TGF-β signaling and regulating the differentiation of cardiac fibroblasts into myofibroblasts (1). In the present study, we showed that the expression level of TGF-β increased significantly in both Cx43+/− and WT hearts post-MI and that the increase was observed in both the scar as well as noninfarcted regions. There was no difference, however, in TGF-β expression between Cx43+/− and WT hearts post-MI, suggesting that the reduced expression of Cx43 did not affect the synthesis or secretion of TGF-β directly. However, nuclear expression of pSmad2/3 was reduced significantly in both the scar and noninfarcted regions in Cx43+/− hearts, suggesting that reduced downstream TGF-β signaling is involved in the effects on fibroblast transformation and collagen deposition. Our data are consistent with the work of Dai et al. (11) and suggest that in the presence of reduced Cx43 expression, with more Smad2/3 bound to microtubules, less Smad2/3 is available for phosphorylation and subsequent pSmad2/3-Smad4 complex formation.
Limitations of the study.
Although fibroblasts were the cells most likely quantified after BrdU injection and staining, we cannot rule out contributions by pericytes or endothelial cells. Likewise, although we used MPO as a marker for neutrophil infiltration, MPO is also expressed by monocytes and macrophages. Our data do not allow us to distinguish between the various inflammatory cells populating the infarct. Furthermore, because an extensive time course was not performed, we do not know whether these inflammatory cells arrived in the infarct later or remained there longer. Finally, although our data are consistent with impaired TGF-β signaling and pSmad-dependent infarct remodeling 6 days post-MI, consistent with a peak in the TGF-β time course post-MI (23), the mechanism(s) responsible for postinfarct ventricular remodeling 4 wk post-MI remains unknown.
In summary, despite the fact that infarct size was the same in Cx43-deficient and WT hearts subjected to MI, and global echocardiographically detected defects were modest, the infarct architecture and post-MI ventricular remodeling of both infarcted and noninfarcted regions were dramatically different in Cx43-deficient hearts compared with WT hearts. Such striking heterogeneities are likely to reveal functionally relevant correlates in animal models and in humans as sophisticated imaging modalities evolve to detect regional abnormalities (32, 48, 51, 55). Our present data point to an underappreciated mechanism whereby Cx43 expression levels can contribute to ventricular remodeling and interstitial fibrosis via TGF-β-mediated signaling post-MI. These findings may have important implications for infarct remodeling in patients who already exhibit reduced Cx43 expression due to diabetes, hypertrophy, or heart failure (17, 27, 29, 34, 41). Gap junction pharmacology is a burgeoning new area that may offer innovative and efficacious antiarrhythmic therapy (28, 59). Our data suggest that it will be essential to evaluate whether new antiarrhythmic peptides influence or modulate ventricular remodeling and fibrosis to insure that novel compounds that improve gap junction intercellular communication and reduce ischemia-induced arrhythmia do not increase cardiac fibrosis and worsen ventricular function post-MI.
Mice were housed in a facility supported by National Institutes of Health (NIH) Grant C06-RR-015502. This work was supported by an American Heart Association Fellowship (to Y. Zhang) and by NIH Grant HL-066350 (to K. A. Yamada).
No conflicts of interest are declared by the author(s).
The authors thank Carla J. Weinheimer (Mouse Cardiovascular Phenotyping Core) for performing all survival surgeries, Bill Coleman and Marlene Scott (Histology and Microscopy Core) for cutting all the tissue sections and performing the trichrome stains, the Mouse Genetics Core for mouse husbandry support, and Dr. Robert Heuckeroth for the use of the fluorescence microscope and camera system.
↵1 Supplemental Material for this article is available online at the American Journal of Physiology-Heart and Circulatory Physiology website.
- Copyright © 2010 the American Physiological Society