Research ArticleCARDIOMYOPATHY

Severe dystrophic cardiomyopathy caused by the enteroviral protease 2A–mediated C-terminal dystrophin cleavage fragment

See allHide authors and affiliations

Science Translational Medicine  01 Jul 2015:
Vol. 7, Issue 294, pp. 294ra106
DOI: 10.1126/scitranslmed.aaa4804

Abstract

Enterovirus infection can cause severe cardiomyopathy in humans. The virus-encoded 2A protease is known to cleave the cytoskeletal protein dystrophin. It is unclear, however, whether cardiomyopathy results from the loss of dystrophin or is due to the emergence of a dominant-negative dystrophin cleavage product. We show for the first time that the 2A protease–mediated carboxyl-terminal dystrophin cleavage fragment (CtermDys) is sufficient to cause marked dystrophic cardiomyopathy. The sarcolemma-localized CtermDys fragment caused myocardial fibrosis, heightened susceptibility to myocardial ischemic injury, and increased mortality during cardiac stress testing in vivo. CtermDys cardiomyopathy was more severe than in hearts completely lacking dystrophin. In vivo titration of CtermDys peptide content revealed an inverse relationship between the decay of membrane-bound CtermDys and the restoration of full-length dystrophin at the sarcolemma, in support of a physiologically relevant loss of dystrophin function in this model. CtermDys gene titration and dystrophin replacement studies further established a target threshold of 50% membrane-bound intact dystrophin necessary to prevent mice from CtermDys cardiomyopathy. Conversely, the NtermDys fragment did not compete with dystrophin and had no pathological effect. Thus, CtermDys must be localized to the sarcolemma, with intact dystrophin <50% of normal levels, to exert dominant-negative peptide-dependent cardiomyopathy. These data support a two-hit dominant-negative disease mechanism where membrane-associated CtermDys severs the link to cortical actin and inhibits both full-length dystrophin and compensatory utrophin from binding at the membrane. Therefore, membrane-bound CtermDys is a new potential translational target for virus-mediated cardiomyopathy.

INTRODUCTION

Cardiotropic enterovirus (EV) infection is associated with marked cardiac dysfunction in humans (1, 2). EV infection is an important cause of acute myocarditis with clinical presentation ranging from mild cardiomyopathy to sudden cardiac death. Five percent to 10% of healthy, asymptomatic individuals show evidence of EV infection in the heart (3, 4). EV has been detected in 20 to 40% of patients who died suddenly after acute myocardial infarction and in 40% of patients with idiopathic dilated cardiomyopathy, indicating that EV infection increases the susceptibility of the heart to injury and dysfunction. However, the direct mechanism by which EV infection compromises cardiac function has not been fully established.

EV infection is known to alter the structural integrity of dystrophin in the heart, a key cytoskeletal protein encoded by the Duchenne muscular dystrophy (DMD) gene. Genetic mutations leading to loss of dystrophin cause DMD—a severe, progressive, and fatal disease of marked cardiac and skeletal muscle deterioration. Dystrophin exists as a subsarcolemmal constituent of the dystrophin-glycoprotein complex (DGC) (5), which is a multimeric, sarcolemma-spanning protein complex that provides a mechanical link between the cytoskeleton and the extracellular matrix. Dystrophin participates in this linkage by binding cortical actin at its N terminus and the DGC protein β-dystroglycan at its C terminus (6). Spanning between these two domains of dystrophin are spectrin-like repeats with interspersed hinge domains constituting the large rod domain region of the peptide. Loss of dystrophin in DMD causes mechanical fragility of the sarcolemma, providing the mechanism of incontrovertible skeletal and cardiac muscle deterioration (79).

Previous studies showed that Coxsackievirus-encoded 2A protease (2Apro) cleaves dystrophin within the hinge 3 domain (10, 11). This led to the provocative hypothesis that cleavage of dystrophin directly contributes to the cardiomyopathy associated with EV infection. Subsequent studies found that EV infection causes truncation of dystrophin into fragments consistent with 2Apro cleavage in vivo, and that transgenic cardiac-directed expression of 2Apro causes cardiomyopathy (11, 12). One complicating issue is that 2Apro cleaves numerous other substrates, including poly-A binding protein, eIF-4G, and serum response factor, which may also contribute to EV infection–associated heart failure (13). On this point, a recent study by Lim and colleagues (14) showed that expression of a 2Apro cleavage–resistant dystrophin peptide can block cardiomyopathy by EV infection/2Apro expression, providing evidence that dystrophin is a key target in viral-mediated cardiomyopathy.

It is currently unknown, however, whether it is the loss of full-length intact dystrophin or the emergence of dystrophin fragments as dominant-negative peptides that causes EV-mediated cardiomyopathy. This is a major open question, because deficiency of full-length dystrophin content is known to be sufficient to cause cardiomyopathy (15, 16) and expression of endogenous truncated isoforms of dystrophin can cause marked dystrophic skeletal myopathy (17, 18). These isoforms, derived from alternative promoters within the DMD gene, retain dystrophin’s C-terminal β-dystroglycan binding domain but lack dystrophin’s actin binding domains. Here, it is notable that the two products of dystrophin cleavage by EV-encoded 2Apro are a 2426–amino acid N-terminal fragment (NtermDys) containing dystrophin’s actin binding domains and a 1251–amino acid C-terminal fragment (CtermDys) containing dystrophin’s β-dystroglycan binding domain (Fig. 1A) (10).

Fig. 1. Generation of transgenic mice and expression of dystrophin and DGC proteins.

(A) (Top) Schematic of dystrophin structure. Hinge domains are indicated by blue boxes. Spectrin-like repeats are indicated by red boxes, with yellow boxes constituting the rod domain actin binding region of dystrophin. Arrow at hinge 3 indicates 2Apro cleavage site. (Bottom) Schematics of the NtermDys and CtermDys transgenes with N-terminal epitope tags. NtermDys contains an N-terminal FLAG epitope tag, the N-terminal domain of dystrophin, and the N-terminal portion of the central rod domain (including spectrin-like repeats 1 to 19, and hinges 1, 2, and part of hinge 3). CtermDys contains an N-terminal MYC epitope tag, dystrophin’s cysteine (Cys)–rich domain, C-terminal domain, and C-terminal portion of the rod domain (containing part of hinge 3, hinge 4, and spectrin-like repeats 20 to 24). (B) Representative Western blot for dystrophin expression in membrane proteins of NTg, CtermDys, and NtermDys transgenic mice. Quantification of dystrophin expression shown as expression relative to loading normalized to NTg. (C) Immunofluorescence detection of CtermDys in a heart section of CtermDys mice. Scale bar, 100 μm. Inset: Merged laminin (green) and CtermDys (red), indicating the basal lamina. (D to H) Quantification of β-dystroglycan (D), γ-sarcoglycan (E), α-sarcoglycan (F), α-dystroglycan (G), and utrophin (H) expression in membranes isolated from NTg and transgenic mice. Quantification of protein expression shown as expression relative to loading normalized to NTg. In (B) and (D) to (H), data are means ± SEM (n = 4 to 7 per group); *P < 0.05 versus NTg, analysis of variance (ANOVA) and Dunnett’s multiple comparison test.

Because the CtermDys cleavage peptide is structurally similar, but not identical to, the previously described dominant-negative isoforms of dystrophin in skeletal muscle, we tested here the hypothesis that the presence of this fragment is sufficient to cause dystrophic cardiomyopathy in the heart. Additionally, because EV infection has been associated with increased risk of death after acute myocardial infarction (3), we hypothesized that the CtermDys fragment resulting from dystrophin cleavage by 2Apro increases susceptibility to ischemic injury and cardiac stress. Insight into dystrophin cleavage fragments as potential poison peptides would help instruct future clinical interventions targeting their decay.

Here, we report that the 2Apro cleavage product CtermDys is stable at the sarcolemma and nucleates key DGC proteins. In otherwise normal mice, CtermDys—but not NtermDys—caused cardiac fibrosis, increased pump dysfunction during ischemia/reperfusion, and increased mortality during cardiac stress testing in vivo. The magnitude of CtermDys-dependent cardiomyopathy was more severe than in mice with complete dystrophin deficiency. Forced overexpression of full-length dystrophin dislodged CtermDys from the membrane and prevented cardiomyopathy. In comparison, the NtermDys fragment had no effect on dystrophin content and did not cause cardiomyopathy. These findings support a mechanism whereby CtermDys is retained at the membrane, severing intact dystrophin’s connection to actin and preventing both intact dystrophin and compensatory utrophin from binding to the sarcolemma. This two-hit model can explain the more severe cardiomyopathy observed for CtermDys compared with the cardiomyopathy of complete dystrophin deficiency. On the basis of these results, membrane-bound CtermDys could serve as a new target for therapeutic remediation of Coxsackievirus-mediated cardiomyopathy.

RESULTS

Dystrophin peptide fragment expression in the heart

In the first series of experiments, transgenic mice were generated expressing either the N-terminal or the C-terminal product of primary dystrophin cleavage site by 2Apro specifically in the heart (Fig. 1). The N-terminal transgene-encoded protein (NtermDys) has a predicted molecular weight of 283 kD and consists of an N-terminal FLAG epitope tag and the first 2426 amino acids of dystrophin. This includes the N-terminal domain of dystrophin and the N-terminal portion of the central rod domain (including spectrin-like repeats 1 to 19 and hinges 1, 2, and part of hinge 3) (Fig. 1A). Thus, this fragment retains both the N-terminal and rod domain actin binding regions of dystrophin (19). Previous reports have shown that inclusion of an N-terminal FLAG epitope tag does not impair dystrophin’s actin binding function (20).

Western blot of total protein extracts from several lines of transgenic mice was performed to assess transgene expression. The NtermDys transgene was detected using an anti-FLAG antibody and an anti-dystrophin antibody (recognizing amino acids 1181 to 1388 of dystrophin, present in both dystrophin and the NtermDys transgene). Four transgenic lines (368, 367, 10, and 13) exhibited a range of transgene expression (fig. S1A). Line 368 was chosen as the primary transgenic line for further study and will hereafter be referred to as NtermDys unless otherwise stated.

The C-terminal transgene (CtermDys) has a predicted molecular weight of 145 kD and consists of an N-terminal MYC tag and the last 1251 amino acids of dystrophin. This includes dystrophin’s cysteine-rich domain, C-terminal domain, and C-terminal portion of the rod domain (containing part of hinge 3 and all of hinge 4, and spectrin-like repeats 20 to 24) (Fig. 1A). Thus, this transgene retains dystrophin’s dystroglycan binding and C-terminal scaffolding regions. For this transgene, two lines of transgenic mice were obtained that bred successfully. Western blot of total protein extracts was carried out to assess transgene expression. The CtermDys protein was detected using an anti-MYC antibody and an anti-dystrophin antibody (recognizing amino acids 3661 to 3677 of dystrophin, present in both dystrophin and the CtermDys transgene). Epitope tagging of truncated dystrophins caused no apparent effects (18). No significant difference was detected in expression of the CtermDys transgene between transgenic lines 1 and 4 (fig. S1B). Line 4 was chosen for further study due to ease of breeding.

Expression of dystrophin and DGC proteins in transgenic mice

Dystrophin and DGC proteins are important to investigate because of their roles in membrane stability in both acquired and inherited forms of cardiomyopathy (21, 22). In total heart muscle protein extracts, expression of endogenous membrane–bound dystrophin was normal in NtermDys transgenic mice but reduced in CtermDys transgenic mice (Fig. 1B and fig. S1). To determine the effects of CtermDys on DGC proteins at the membrane, Western blots were performed on potassium chloride (KCl)–washed microsomes extracted from the myocardium. Expression levels of dystrophin, α-dystroglycan, α-sarcoglycan, β-dystroglycan, utrophin, neuronal nitric oxide synthase (nNOS), syntrophin, caveolin-3, and γ-sarcoglycan in NtermDys transgenic mice were similar in nontransgenic (NTg) mice, but α-dystrobrevin-2 was significantly reduced (Fig. 1, B to H, and fig. S2).

Similar to findings in total protein extracts, dystrophin expression at the membrane was significantly reduced in CtermDys transgenic mice to about 30% of normal levels (Fig. 1B). Expression of α-dystroglycan, α-sarcoglycan, β-dystroglycan, α-dystrobrevin-2, and γ-sarcoglycan was significantly increased in CtermDys transgenic mice compared with that in NTg animals (Fig. 1, D to G, and fig. S2A), whereas expression of syntrophin and caveolin was slightly, but significantly, reduced (fig. S2, B and C). Utrophin and nNOS were unchanged in CtermDys transgenic hearts compared to those in NTg controls (Fig. 1H and fig. S2D).

The subcellular localization of the NtermDys transgene was assessed by immunofluorescence on heart sections of NtermDys transgenic mice. NtermDys transgenic mice were crossed onto the dystrophin-deficient mdx genetic background [NtermDys(mdx)]. The NtermDys transgene localized to the sarcolemma, similar to dystrophin, but was also found more diffusely throughout cardiac myocytes (fig. S3). Expression of the DGC protein β-dystroglycan was unchanged in NtermDys(mdx) hearts (fig. S3A). Some NtermDys was detected in the nucleus of cardiac myocytes, but not in the nuclei of other cells within the heart, such as fibroblasts [dystrophin (mid-rod) in fig. S3]. Subcellular localization of the CtermDys transgene was determined by immunofluorescence on heart sections from CtermDys transgenic mice. The CtermDys transgene localized at the sarcolemma in transgenic animals (fig. S4), similar to NtermDys.

Heart structure and membrane integrity in CtermDys mice

Mdx mice show progressive histopathology within the myocardium as seen by inflammation and fibrosis, although overall mdx heart size is not significantly altered (8, 23, 24). In comparison, CtermDys mice had enlarged hearts and significant myocardial fibrosis, with evidence of inflammation (Fig. 2, A and B). Fibrosis was significantly greater in CtermDys than in mdx hearts, suggesting that CtermDys transgenic animals have more severe pathology than both NTg littermates and mdx mice. Hearts from NtermDys adult mice showed no difference in fibrosis compared to those from NTg animals (Fig. 2B). Evans blue dye (EBD) uptake area was larger than the fibrotic area in both mdx and CtermDys mice (Fig. 2C). Although increased membrane permeability is likely the initiating insult leading to fibrosis, there was not a 1:1 correlation between dye-positive area and fibrotic area.

Fig. 2. Heart morphology, fibrosis, and membrane instability in hearts of CtermDys mice.

(A) Whole-heart images (scale bar, 2 mm) and transverse sections [hematoxylin and eosin (H&E)], and heart-to-tibia length ratio. Data (expressed as ratio of heart weight to tibia length) are means ± SEM (n = 4). P value was determined by t test. The representative H&E-stained heart transverse section indicates inflammation in fibrotic area [scale bars, 1 and 0.5 mm (inset)]. (B) Sirius Red–Fast Green staining of representative NTg, NtermDys, CtermDys, and mdx mouse hearts (age 6 months). Scale bar, 200 μm. Fibrotic area was quantified from the stained images. Data are means ± SEM (n = 8 to 15). P values were determined by ANOVA. (C) Representative mosaic images of EBD uptake (18 hours after dye injection) in hearts of NTg, mdx, NtermDys, CtermDys, NtermDys(mdx), CtermDys(mdx), and DTg mice. Scale bars, 1 mm. The dye uptake percentage out of whole-heart area was quantified from the images. Data are means ± SEM (n = 7 to 11 per group). *P < 0.05; **P < 0.01; ***P < 0.001 versus NTg, ANOVA.

Previous studies have shown that dystrophin deficiency predisposes the sarcolemma to mechanical disruption during stress with adrenergic agonists, including isoproterenol (7, 25). To determine whether expression of the NtermDys or CtermDys transgenes causes mechanical instability of the cardiac sarcolemma, EBD uptake was assessed during isoproterenol stress in vivo. NTg littermates showed minimal dye uptake in cardiac myocytes during this stress, similar to NtermDys transgenic lines 368, 367, 10, and 13 (Fig. 2 and fig. S5). In contrast, CtermDys transgenic mice showed significantly greater dye uptake compared with NTg mice and was similar to dystrophin-deficient mdx mice (Fig. 2C).

Because cleavage of dystrophin by 2Apro results in both N- and C-terminal cleavage products, CtermDys (line 4) mice were crossed with NtermDys (line 368) mice to create double transgenic (DTg) mice. DTg mice showed significantly greater EBD uptake during isoproterenol stress than did NTg mice, but similar to mdx and CtermDys transgenic mice (Fig. 2B). Collectively, these results indicate that expression of CtermDys, but not NtermDys, is sufficient to cause sarcolemmal instability.

Increased susceptibility to ischemic injury and stress-induced mortality in CtermDys transgenic mice

A previous study showed that EV infection may predispose patients to death after acute myocardial infarction (3). The mechanism by which EV infection increases susceptibility to ischemic injury is not known, but given that CtermDys mice show dystrophic cardiomyopathy, we hypothesized that the presence of the CtermDys transgene would increase susceptibility to ischemic injury. To test this hypothesis, hearts of CtermDys and NTg littermate controls were isolated, perfused, and subjected to 20 min of ischemia and 60 min of reperfusion. Before ischemia, NTg and CtermDys transgenic hearts had similar systolic function as measured by left ventricular developed pressure (LVDP) (Fig. 3A). However, recovery of systolic function during reperfusion was significantly impaired in CtermDys hearts compared to that in NTg hearts (Fig. 3, A and B). Additionally, recovery of diastolic function during reperfusion was significantly impaired in CtermDys hearts, as shown by elevated left ventricular end-diastolic pressure (LVEDP), compared to that in NTg hearts (Fig. 3C).

Fig. 3. Ischemic injury in isolated hearts and adrenergic stress testing of transgenic mice in vivo.

(A and B) LVDP and recovery of LVDP in hearts of NTg and CtermDys mice during ischemia/reperfusion injury. *P < 0.05 versus CtermDys by two-way ANOVA. Data are means ± SEM (n = 7 per group). (C) LVEDP of NTg and CtermDys hearts during ischemia/reperfusion injury. Data are means ± SEM (n = 7 per group). P value was determined by t test. (D) Survival study by cardiac stress testing in vivo. Animals were administered with isoproterenol (three times per day, every 8 hours) starting at time 0 (arrows), and survival of NTg, mdx, and CtermDys mice was evaluated over the course of 5 days. n = 10 to 15 animals per group. P values were determined by Mantel-Cox test.

To further examine the cardiomyopathy of CtermDys transgenic mice in an in vivo setting, mice were subjected to cardiac stress by isoproterenol administration for five consecutive days. Most NTg mice survived this cardiac challenge (92%), whereas only 40% of mdx survived (Fig. 3D). All CtermDys mice (n = 10) succumbed within the first 3 days of this stress test (Fig. 3D).

Ectopic full-length dystrophin rescues CtermDys-mediated cardiomyopathy

CtermDys mice showed increased mortality during adrenergic stress and increased cardiac fibrosis compared with mdx mice (Figs. 2 and 3), suggesting novel mechanisms for the cardiomyopathy caused by the CtermDys cleavage fragment. To determine whether these phenotypic differences were founded in aberrant effects on DGC function, or some other direct or nonspecific effect of transgenesis, CtermDys transgenic mice were crossed with two lines of mice transgenically expressing a human intact dystrophin (hDys) transgene. Relative to native dystrophin, the hDys 600 transgene is expressed at 50% of normal dystrophin and the hDys 649 transgene is expressed at levels 10-fold greater than normal dystrophin (Fig. 4A). CtermDys-hDys649(mdx) DTg mice demonstrated a significant reduction in CtermDys protein expression (Fig. 4B) with reduced EBD uptake and complete survival during adrenergic stress similar to NTg mice (Fig. 4, C to E). CtermDys-hDys600(mdx) mice demonstrated no decrease in CtermDys expression and no improvement in EBD uptake (Fig. 4, B and C). However, there was a significant improvement in survival during adrenergic stress testing (Fig. 4D), suggesting that titrating full-length dystrophin levels can confer myocardial protection in CtermDys mice during stress.

Fig. 4. Expression of full-length functional dystrophin corrects dystrophic cardiomyopathy in CtermDys transgenic mice.

(A) Representative Western blots of heart total protein extracts from NTg and transgenic animals. Dystrophin expression was detected using an antibody that detects both mouse and human dystrophin. Desmin served as the loading control. (B) Quantification of CtermDys protein expression in (A) is relative to desmin and normalized to CtermDys mice, and is expressed as means ± SEM (n = 4 to 5). P value was determined by ANOVA. (C) Quantification of EBD uptake in NTg, CtermDys, and DTg(mdx) hearts. Data are means ± SEM (n = 4 to 5 per group). P values were determined by ANOVA. (D) Survival of NTg, CtermDys, CtermDys-hDys649(mdx), and CtermDys-hDys600(mdx) mice during prolonged adrenergic stress. n = 11 per group. P value comparing CtermDys to NTg and CtermDys-hDys649(mdx) was determined by Mantel-Cox test. (E) Representative images of EBD uptake in NTg, CtermDys, and DTg(mdx) hearts. Scale bars, 1 mm.

In vivo titration of the CtermDys peptide

To further validate physiological relevance, and to establish biological dose-response for the cardiotoxic effects of CtermDys, floxed CtermDys mice were crossed with mice expressing a cardiac-specific tamoxifen-inducible Cre recombinase transgene [Mer-Cre-Mer (MCM)] (Fig. 5A) (12). Tamoxifen was administered to DTg CtermDys × MCM mice, resulting in efficient excision of the CtermDys transgene and a time-dependent reduction in cardiac CtermDys protein (Fig. 5B). Concurrent with decay of the CtermDys peptide, there was a reciprocal increase in full-length dystrophin at the membrane (Fig. 5B). This is direct evidence of a stoichiometric balance between CtermDys and full-length dystrophin at the membrane. By assuming fast tamoxifen-based CtermDys gene excision and subsequent fast CtermDys mRNA decay (estimated 24 hours), the half-life of the CtermDys peptide was calculated to be ~3 to 4 days in the mouse heart in vivo. If CtermDys locus excision and mRNA decay were slower, then the CtermDys peptide half-life would be <3 days. In either case, the CtermDys peptide turnover rate was markedly faster than that estimated for full-length dystrophin, which has been calculated as months in mice (26).

Fig. 5. Survival during physiological stress testing and gene-dose analysis by in vivo cardiac titration of the CtermDys peptide.

(A) CtermDys mice were crossed with mice expressing a cardiac-specific tamoxifen-inducible Cre recombinase (MCM). Tamoxifen was administered to the resulting DTg CtermDys × MCM mice to study the decay of the CtermDys protein in the heart after gene excision in vivo. (B) Representative Western blot showing CtermDys expression level decay and full-length endogenous dystrophin increase at the sarcolemma in CtermDys × MCM mice after tamoxifen treatment. Dystrophin and CtermDys were both detected with the same C-terminal antibody. Desmin was used as loading control. Quantification of full-length dystrophin and CtermDys expression in CtermDys × MCM mice after tamoxifen treatment. Data are means ± SEM (n = 6 per time point). (C) Survival curve of transgenic and NTg mice (n = 8 to 12 per group) starting 5 days after tamoxifen treatment, during isoproterenol stress test. (D) Survival curves for transgenic and NTg mice during isoproterenol stress test (no tamoxifen injected) (n = 6 to 8 per group). P value was determined by Mantel-Cox test.

To determine the effect of CtermDys gene excision on susceptibility to adrenergic stress, CtermDys × MCM mice and controls were subjected to cardiac stress testing. Six days after tamoxifen administration, average CtermDys protein levels were reduced by ~75%, and intact full-length dystrophin was increased to 40 to 50% of normal levels at the membrane (Fig. 5B). At this time point, CtermDys-MCM mice showed a significant reduction in mortality compared to CtermDys mice and similar to MCM and NTg mice (Fig. 5C). When not treated with tamoxifen, CtermDys-MCM mice demonstrated lower survival during adrenergic stress than did MCM and NTg control mice, but similar to CtermDys mice (Fig. 5D).

DISCUSSION

EV infection–mediated cardiomyopathy is a vexing clinical disease with outcomes ranging from mild to cardiogenic shock and sudden death in humans (2, 27). Insights into disease etiology stem from studies showing that viral infection mediates cleavage of the cytoskeletal protein dystrophin (1012, 14, 28). Unresolved, however, is the molecular mechanism by which dystrophin cleavage causes disease. Specifically, is it the loss of intact dystrophin or rather the emergence of dominant-negative cleavage fragments that underlies EV infection cardiomyopathy? Here, we discovered that CtermDys, the C-terminal dystrophin cleavage fragment by Coxsackie B virus–encoded 2Apro, is sufficient to directly cause severe cardiomyopathy. This finding has implications for human disease, as discussed below.

Our results provide direct evidence that the CtermDys fragment is stable at the cardiac sarcolemma and efficiently nucleates the DGC in vivo. We propose a two-hit CtermDys-based disease mechanism whereby (i) membrane-tethered CtermDys mechanically uncouples the key actin-DGC linkage and (ii) CtermDys retards full-length dystrophin recovery and compensatory utrophin binding at the membrane to cause marked dystrophic cardiomyopathy (Fig. 6). This model explains the more severe cardiomyopathy resulting from CtermDys—in terms of increased cardiac fibrosis and increased mortality during cardiac stress—than from lack of dystrophin. CtermDys, by blocking compensatory up-regulation of the dystrophin paralog utrophin at the membrane, differs from fully dystrophin-deficient muscle where utrophin can partially compensate for complete lack of dystrophin to lessen disease (2931). Our results, however, do not exclude other downstream deleterious effects of CtermDys. Evidence here of CtermDys as a loss of dystrophin function cleavage fragment provides new mechanistic insight into the clinical presentation of EV infection–based cardiomyopathy and establishes a foundation on which to guide new therapeutic interventions.

Fig. 6. Two-hit model of CtermDys as a loss of dystrophin function peptide.

(A) Normal DGC network in human cardiac muscle. (B) Proposed two-hit model of EV infection–based CtermDys cardiomyopathy showing severed link to actin and, because CtermDys is anchored to the DGC, prevention of utrophin compensatory binding. (C) Dystrophin-deficient myocyte with capacity for utrophin to bind and compensate, in part, for loss of dystrophin.

The physiological relevance of the experimental approaches used here and, by extension, potential clinical insights warrant discussion. Mechanistic insights enabled by direct in vivo titration of the CtermDys fragment provide strong support for the CtermDys loss of dystrophin function model (Fig. 6). The direct documentation of the CtermDys fragment decay rate in the heart in vivo underscores the clinical and physiological relevance of this experimental model. Specifically, our results show a synchronous inverse relationship between the decay of the CtermDys peptide and simultaneous restitution of membrane-bound intact dystrophin. The inverse relationship observed here is taken as direct evidence that CtermDys turnover and intact dystrophin are in stoichiometric balance in these hearts in vivo, substantiating this as a valid experimental model of human cardiomyopathy upon EV infection. We propose this CtermDys experimental model as physiologically relevant in terms of dissecting mechanisms of proteolytic cleavage of dystrophin by EV in wild-type hearts, and therefore reflects what would be seen in patients with Coxsackievirus infection.

Knowlton and colleagues (14, 27) provided evidence that viral infection–mediated dystrophin cleavage causes cardiac muscle membrane destabilization, leakiness, and dysfunction. The earlier proposed mechanistic basis for this effect, stemming from previous in vitro studies, has been that virus-encoded 2Apro cleaves and then displaces dystrophin and several DGC partners from the sarcolemma, suggesting that it is the loss of intact dystrophin at the membrane that causes disease (32). However, quantitative evidence for displacement of membrane-bound dystrophin by 2Apro in hearts in vivo was lacking. This was further complicated by previous findings showing that non–viral-mediated cardiomyopathy causes loss of membrane-bound dystrophin (33). Together, these data raised uncertainties regarding the mechanism of how dystrophin cleavage causes disease. Our data demonstrate that the 2Apro-cleaved CtermDys peptide readily localizes and efficiently nucleates the DGC at the sarcolemma in vivo, thus permitting direct quantitative analysis of dystrophin cleavage fragments in the heart. Our findings gain support from previous studies of expression of CtermDys-like truncated dystrophin isoforms, Dp116 and Dp71, shown to be able to localize to skeletal muscle membranes (17, 18). As discussed below, CtermDys as a defective dystrophin fragment attached to the sarcolemma—as opposed to dystrophin deficiency—has implications for the mechanism of disease and potential therapeutics.

The CtermDys-based loss-of-function mechanism is supported by our data showing that cardiomyopathy can be blocked by cardiac overexpression of intact dystrophin at levels sufficient to displace membrane-bound CtermDys. Further, cardioprotection is conferred by direct CtermDys peptide titration in vivo. Here, we calculate that 30% of normal levels of intact dystrophin at the membrane are insufficient to confer cardioprotection and estimate a therapeutic target of >50% membrane-bound dystrophin required to prevent cardiomyopathy. In principle, as above, a therapeutic dystrophin/CtermDys ratio could be achieved by limiting CtermDys formation by infection/dystrophin cleavage suppression (34) and/or by introduction of full-length wild-type or noncleavable dystrophin (14) by gene therapy strategies (23). In this regard, these findings are interesting in comparison to the therapeutic range of ~20% for dystrophin to prevent DMD in muscle (35). We therefore predict that the titration of CtermDys away from the sarcolemma would have beneficial effects on heart performance. Other therapeutic avenues would involve small molecules or other drugs developed to stabilize full-length dystrophin, by limiting or preventing its cleavage during Coxsackievirus infection. Proof of concept for this approach has been established in transgenic animals with a modified dystrophin resistant to cleavage (14).

Our findings of the deleterious effects of the CtermDys peptide on cardiac muscle, together with the unexpected rapid CtermDys turnover kinetics (~3 days), may provide new insight into understanding the clinical presentation and time course of EV infection–mediated cardiomyopathy. Clinically, EV infection–based cardiomyopathy is challenging to diagnose where disease onset can be sudden and severe, and then rapidly resolve (2). This disease presentation could be explained in part by CtermDys as a loss-of-function peptide arising rapidly at the onset of infection/dystrophin cleavage. In turn, fast decay of the CtermDys peptide may help explain the clinical conundrum of rapid resolution of disease that can occur within days of viral clearance (loss of 2Apro activity) (2, 27). Rapid CtermDys peptide turnover is an unexpected outcome based on estimates of intact dystrophin turnover rate thought to be on the order of months (26). The basis for fast CtermDys fragment turnover is not known. The potent effects of CtermDys may also help explain how fulminant cardiomyopathy and acute heart failure can occur with only apparent regional or focal evidence of infection. Here, CtermDys as a loss-of-function dystrophin fragment could help explain the substantial cardiac deficits, such as cardiogenic shock, associated with this disease. Our findings indicate not only that the CtermDys protein is stable enough to contribute significantly to cardiac dysfunction over the course of days after infection but also that its effects are more severe than those attributed to full loss of intact dystrophin.

In conclusion, CtermDys is a dominant-negative peptide sufficient to cause severe dystrophic cardiomyopathy. The significance of this finding rests in the identification of a 2Apro-mediated dystrophin cleavage product as a novel target for therapeutic intervention in EV infection. Data support a model whereby CtermDys must be tethered to the sarcolemma to exert its dominant loss-of-function effects. Thus, membrane-bound CtermDys content serves as a new target for potential therapeutic interventions. Here, two general approaches for potential translation to human health could derive from this work, including gene-based expression of functional dystrophins to dislodge CtermDys, or using small molecules to facilitate faster decay of CtermDys. If successful, these approaches would be expected to improve overall heart function in Coxsackievirus-mediated cardiomyopathy.

MATERIALS AND METHODS

Study design

Our hypothesis was that viral-mediated cleavage of dystrophin would lead to dominant-negative peptides deleterious to heart structure and function in vivo. We were interested in testing the hypothesis that cardiac stress would significantly unmask potential dominant-negative effects of the dystrophin cleavage fragments. Statistical analysis and sample size justification was derived from our previous animal studies (7, 21, 24) that were sufficiently powered to obtain significant insights into the cardiomyopathy of muscular dystrophy. In general, sample sizes averaged 7 to 10 per group for heart function and structure. For histological and mouse studies, randomization and blinding was carried out by using ear tag or sample ID numbers, which did not indicate genotype. The person carrying out the experiment and subsequent analysis was not aware of the genotypes associated with the ear tag or sample ID until after data collection was complete, at which time the experimenter was unblinded. Further details of sample size/replicates are given in the figure legends.

Generation of transgenic mice

The NtermDys and CtermDys transgenic constructs were polymerase chain reaction (PCR)–cloned from the murine dystrophin complementary DNA containing an N-terminal FLAG tag (gift from J. Ervasti). For the NtermDys transgene, the forward primer 5′-ttttttttgcggccgctacggcaaggtgctgtgcacggatctgccc-3′ and reverse primer 5′-ttttttttgcggccgccctgaccgtgcccctggactgagcactact-3′ were used to generate a 7.3-kbp (kilo–base pair) product with flanking Not I restriction sites. For the CtermDys transgene, the forward primer 5′-ttttttttgcggccgcgagcagaacgtgatctcggaggaggacctgggagcctctgccagtcagactgttactcta-3′ and reverse primer 5′-ttttttttgcggccgcactgaaactaaggactccatcgctctgccc-3′ were used to generate a 3.9-kbp product with flanking Not I restriction sites. PCR mutagenesis was also used to add an N-terminal MYC tag to the CtermDys transgene. These transgenic constructs were then inserted between loxP sites in a transgenic vector with expression dictated by the cardiac-specific α-myosin heavy chain promoter. These constructs were then injected into (C57BL/6 × SJL)F2 mouse eggs, and potential founders were screened for the transgene by PCR.

Animals

Transgenic mice were backcrossed onto the C57BL/10SnJ genetic background for two to three generations (Jackson Laboratory, #000666). We tracked the hypomorphic dysferlin allele inherited from SJL mice, and this was bred out of mice used in this study. Male and female mice of 2 to 3 months of age were analyzed, with the exception of mice used for histological analyses, which were 6 months old. To generate transgenic (mdx) mice, male transgenic mice were bred with mdx females (Jackson Laboratory, #001801). Transgenic male pups from these crosses are on the mdx background and were then used for experiments to characterize the phenotype of CtermDys(mdx) mice.

Total protein extraction

Total cellular protein was extracted using a method whereby hearts were frozen in liquid nitrogen and then pulverized and resuspended in 1% SDS, 5 mM EGTA, and protease inhibitors. Samples were boiled for 2 min and then centrifuged at 14,000g for 2 min, and the supernatant was collected. Protein concentration was determined using a BCA protein assay kit (Thermo Scientific).

Membrane protein isolation

Hearts were prepared as described previously (23). Briefly, hearts were similarly frozen in liquid nitrogen, pulverized and resuspended in a buffer solution lacking detergent, and centrifuged at 14,000g for 25 min. The supernatant was collected and centrifuged at 100,000g for 40 min. The pellet resulting from this spin was then resuspended in buffer containing 0.6 M KCl and washed for 1 hour at 4°C. Samples were then centrifuged at 150,000g for 40 min, and the resultant pellet was resuspended in a tris-sucrose buffer. Protein concentration was determined using a BCA protein assay kit (Thermo Scientific).

Western blotting

Fifty micrograms of protein was loaded per sample in 4 to 20% tris-HCl gels for SDS–polyacrylamide gel electrophoresis (Bio-Rad). Protein was then transferred to nitrocellulose or polyvinylidene difluoride membranes. Membranes were blocked in 5% nonfat dry milk in tris-buffered saline, and primary antibodies were applied for 1 hour at room temperature. The following primary antibodies were used: rabbit anti-dystrophin (C-terminal; Abcam, ab15277, 1:1000), mouse anti-dystrophin (mid-rod; Millipore, mab1692, 1:200), mouse anti–α-dystroglycan (Millipore, 05-593, 1:1000), rabbit anti-desmin (Novus Biologicals, NB120-15200, 1:1000), mouse anti-MYC (Cell Signaling Technology, 2276, 1:1000), mouse anti-FLAG (Sigma, F1804, 1:500), mouse anti–α-sarcoglycan (Vector Laboratories, VP-A105, 1:100), mouse anti–γ-sarcoglycan (Vector Laboratories, VP-G803, 1:100), mouse anti–β-dystroglycan (Vector Laboratories, VP-B205, 1:100), dystrobrevin (BD Transduction, 610766, 1:500), syntrophin (Abcam, ab11425, 1:1000), caveolin-3 (BD Transduction, 610420, 1:1000), nNOS (Invitrogen/Zymed, 61-7000, 1:250), and utrophin (Santa Cruz Biotechnology, 8A4, 1:50). Secondary antibodies were then applied for 1 hour at room temperature, and blots were imaged using an Odyssey infrared scanner (LI-COR).

Immunofluorescence

Hearts of transgenic mice were embedded in optimal cutting temperature (OCT) and sectioned. Heart sections were fixed in 3% formaldehyde for 15 min at room temperature, then washed in phosphate-buffered saline (PBS), and blocked in 5% normal goat serum + 0.3% Triton X-100 in PBS for 1 hour at room temperature. Sections were then incubated with primary antibodies in 1% bovine serum albumin (BSA) + 0.3% Triton X-100 in PBS overnight at 4°C. The following primary antibodies were used: rabbit anti-laminin (Sigma, L9393. 1:500), mouse anti-MYC (Cell Signaling Technology, 2276, 1:500), mouse anti-dystrophin (Millipore, mab1692, 1:50), and mouse anti–β-dystroglycan (Vector Laboratories, VP-B205, 1:50). Sections were then washed and incubated with secondary antibodies in 1% BSA + 0.3% Triton X-100 in PBS for 1 hour at room temperature. Sections were washed in PBS and mounted using Vectashield mounting medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories). Slides were visualized using Zeiss LSM510 META confocal microscope (Carl Zeiss).

EBD uptake

Membrane permeability was assessed using intraperitoneal injections of EBD in PBS at a dose of 200 μg/g. Eighteen hours later, mice were given three injections of isoproterenol at a dose of 500 ng/g at 18, 20, and 22 hours after EBD injection. Mice were then euthanized at 24 hours after EBD injection. Hearts were then cut into three sections across their short axis and frozen in OCT. Percent Evans blue–positive area from thin sections of these samples was quantified using ImageJ (National Institutes of Health). For each heart, the Evans blue–positive area reported is an average of three cross sections. Sections were imaged on a Zeiss Axio Observer Z1 inverted microscope. Mosaic images were created with AxioVision 4.7 (Carl Zeiss).

Histology

Hearts of 6-month-old mice were embedded in OCT or fixed in 10% formalin overnight and embedded in paraffin. Hearts were sectioned and stained with Sirius Red and Fast Green to detect collagen deposition within the myocardium and imaged on a Zeiss Axio Observer Z1 inverted microscope (Carl Zeiss).

Isolated heart preparation

Mice were injected with 300 U of sodium heparin and anesthetized with sodium pentobarbital. The heart and lungs were removed after thoracotomy and placed immediately in ice-cold Krebs-Henseleit buffer (118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO4, 1.2 mM KH2PO4, 10 mM glucose, 25 mM NaHCO3, 2.5 mM CaCl2, 0.5 mM EDTA). The lungs and thymus were trimmed away to expose the aorta, which was then cannulated. Hearts were then perfused at a constant pressure of 80 mmHg with Krebs-Henseleit buffer warmed to 37°C and brought to pH 7.4 by bubbling with 95% O2, 5% CO2. Hearts were paced at 7 Hz, and changes in left ventricular pressure were monitored by insertion of a water-filled balloon with an in-line pressure transducer into the left ventricle. Within the left ventricle, the balloon was inflated to an end-diastolic pressure of 3 to 8 mmHg. After 10 to 15 min of stabilization time, hearts were subjected to global no-flow ischemia for 20 min. Hearts were not paced during ischemia. Hearts were then reperfused for 60 min, and pacing was reinitiated at 8 min after the end of ischemia (36). Data were collected at a sampling rate of 400 Hz and analyzed using Chart 6 software (ADInstruments).

Cre-mediated transgene excision

Mice were treated with a single dose of tamoxifen (40 mg/kg) (Sigma, T5648) by intraperitoneal injection. Tamoxifen was dissolved in peanut oil at a concentration of 10 mg/ml.

Isoproterenol survival study

Mice were administered three doses of isoproterenol or isoprenaline at a dose of 10 mg/kg per day for five consecutive days (30 mg/kg per day) by intraperitoneal injection. Survival was assessed at 8-hour intervals.

Five days after tamoxifen treatment, the mice received isoproterenol (Sigma, I6504) at a dose of 10 mg/kg at 9 a.m., 1 p.m., and 5 p.m. for 5 days (30 mg/kg per day) by intraperitoneal injection (Fig. 5). Isoproterenol was dissolved in normal saline at a concentration of 3 mg/ml. Survival was assessed at each injection time.

Statistics

Comparisons between two groups were made using Student’s two-tailed t test. When more than two groups were being compared, one-way ANOVA was used with a Newman-Keuls multiple comparison posttest. When more than one independent variable was tested, two-way ANOVA with a Bonferroni posttest was used to compare groups. Log-rank (Mantel-Cox) test was used to assess significance in survival studies. Data are shown as means ± SEM. All statistical analyses were carried out using Prism (GraphPad Software).

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/7/294/294ra106/DC1

Fig. S1. Transgene expression in hearts of NtermDys and CtermDys transgenic mice.

Fig. S2. DGC protein expression in the hearts of transgenic mice.

Fig. S3. Subcellular localization of NtermDys protein in the heart.

Fig. S4. Subcellular localization of CtermDys protein in the heart.

Fig. S5. Evidence of normal membrane integrity in the hearts of NtermDys transgenic lines.

REFERENCES AND NOTES

  1. Acknowledgements: We thank J. Martindale, B. Thompson, J. Ervasti, and K. Prins for additional research support and/or comments on the manuscript. Funding: This work was supported by grants from NIH and the Muscular Dystrophy Association (J.M.M.). We acknowledge the support of the Greg Marzolf Jr. Foundation (to M.S.B., Marzolf Fellowship Award) and the Lillehei Heart Institute in aiding these studies. Author contributions: M.S.B. and J.M.M. wrote the manuscript, designed the experiments, and led the manuscript revisions. M.S.B. conducted most of the experiments and analyzed the data, including the statistics. F.V.S. performed the in vivo CtermDys titration analysis. F.B.B. performed the heart morphology analysis. D.T. led development of the hDys mice. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data and materials are available.
View Abstract

Navigate This Article