Proteomic analysis identifies key differences in the cardiac interactomes of dystrophin and micro-dystrophin

Abstract

ΔR4-R23/ΔCT micro-dystrophin (μDys) is a miniaturized version of dystrophin currently evaluated in a Duchenne muscular dystrophy (DMD) gene therapy trial to treat skeletal and cardiac muscle disease. In pre-clinical studies, μDys efficiently rescues cardiac histopathology, but only partially normalizes cardiac function. To gain insights into factors that may impact the cardiac therapeutic efficacy of μDys, we compared by mass spectrometry the composition of purified dystrophin and μDys protein complexes in the mouse heart. We report that compared to dystrophin, μDys has altered associations with α1- and β2-syntrophins, as well as cavins, a group of caveolae-associated signaling proteins. In particular, we found that membrane localization of cavin-1 and cavin-4 in cardiomyocytes requires dystrophin and is profoundly disrupted in the heart of mdx5cv mice, a model of DMD. Following cardiac stress/damage, membrane-associated cavin-4 recruits the signaling molecule ERK to caveolae, which activates key cardio-protective responses. Evaluation of ERK signaling revealed a profound inhibition, below physiological baseline, in the mdx5cv mouse heart. Expression of μDys in mdx5cv mice prevented the development of cardiac histopathology but did not rescue membrane localization of cavins nor did it normalize ERK signaling. Our study provides the first comparative analysis of purified protein complexes assembled in vivo by full-length dystrophin and a therapeutic micro-dystrophin construct. This has revealed disruptions in cavins and ERK signaling that may contribute to DMD cardiomyopathy. This new knowledge is important for ongoing efforts to prevent and treat heart disease in DMD patients.

Figure 1

Expression of μDys in cardiomyocytes rescues expression of DAPC proteins. (A and B) Schematic representation of dystrophin and μDys in relation to each other and to major known protein binding domains. The epitope recognized by the MANEX1011B antibody is marked. DAG: dystroglycans; SGC: sarcoglycans; SYN: syntrophins; DTNA: α-dystrobrevins; MT: microtubules. (C) Immunostaining of heart tissue sections from wild-type (WT), mdx^5cv and μDYS-mdx^5cv (μDYS) mice with the MANEX1011B antibody to visualize dystrophin and μDys (red). Scale bar: 20 μm. (D) Western blot of total heart protein extracts probed with the MANEX1011B antibody. (E) Dystrophin, μDys and DAPC expression levels in lysates from WT (N = 5–7), mdx^5cv (N = 8–13) and μDYS-mdx^5cv (N = 5–6) hearts. Values (mean ± standard deviations) are normalised to wild-type. Corresponding representative western blots are shown in Supplementary Material, Figure S5. ^*P < 0.05, ^***P < 0.001, one-way ANOVA. (F and H) Superplots of immunofluorescence intensity measurements of dystrophin/μDYS and DAPC proteins at the cardiomyocyte membrane. Small symbols are individual immunofluorescence measurements (N = 40) per mouse. Large symbols indicate the mean for each individual mouse (N = 3/group). Lines indicate the grand mean ± standard deviation. ^*P < 0.05, ^**P < 0.01, ^***P < 0.005, ^****P < 0.001, two-way repeated measures ANOVA. (G) Immunostaining of heart tissue sections with antibodies to the indicated DAPC proteins (green). Nuclei are counterstained with DAPI (blue). Scale bar: 15 μm. βDG: β-dystroglycan; βSG: β-sarcoglycan; SNTA1: α1-syntrophin; DTNA: pan-dystrobrevin antibody; DTNA1: α1-dystrobrevin; DTNA2: α2-dystrobrevin; DTNA3: α3-dystrobrevin.

Figure 2

Expression of μDys in cardiomyocytes prevents development of histopathology. (A and B) Representative montages of heart sections from wild-type (WT), mdx^5cv and μDYS-mdx^5cv (μDYS) mice immunostained for collagen I at 6 months (A) and 12 months (B) of age. Scale bar: 400 μm. (C) Quantification of fibrosis based on the percentage of the cardiac section area positive for collagen I. (D) Quantification of cardiomyocyte hypertrophy based on measurements of the minimum Feret diameter. (E and F) Quantification of capillary density normalised to the area (C) or the number of cardiomyocytes (D). Representative images used for quantifications of cardiomyocyte hypertrophy and capillary density are shown in Supplementary Material, Figure S7. Values in graphs are means ± standard deviations. ^*P < 0.05; ^**P < 0.01; ^***P < 0.005, ^****P < 0.001, one-way ANOVA performed within each age group separately. #P < 0.05, Student’s paired t-test comparison between 6 and 12 months of age for each individual mouse genotype.

Figure 3

Analysis of proteins that co-IP with dystrophin and μDys. Western blot analyses of total cardiac lysates from wild-type (WT), mdx^5cv (MDX) and μDYS-mdx^5cv (μDYS) mice, and of IPs performed with the MANEX1011B antibody (Manex IP) or the MW8 control antibody (MW8 IP). (A) β-dystroglycan (βDG), β1-syntrophin (SNTB1), α1-, α2- and α3- dystrobrevins (DTNA1, DTNA2, DTNA3), and Ahnak are detected in cardiac lysates and MANEX1011B IPs from both wild-type and μDYS-mdx^5cv mice. Full-length cavin-1 is detected in heart lysates from both wild-type and μDYS-mdx^5cv mice but only co-purifies with dystrophin in MANEX1011B IPs. No proteins are detected in control MW8 IPs. (B) An antibody to the N-terminus of cavin-1 detects proteolytic fragments of 22 and 28 kDa in all cardiac lysates. The 28 kDa fragment is detected in MANEX1011B IPs (Manex IP) from wild-type (WT) but not mdx^5cv (MDX) or μDYS-mdx^5cv (μDYS) mice. The smaller 21 kDa fragment is obscured by the IgG light chain of the MANEX1011B antibody. (C) Cavin-2, -3, and -4 and caveolin-3 (Cav3) are detected in all cardiac lysates. Cavins co-IP with dystrophin in wild-type mice, but are absent or strongly reduced in IPs from μDYS-mdx^5cv mice. Caveolin-3 does not co-IP with either dystrophin or μDys.

Figure 4

Membrane localization but not expression of cavins is disrupted in mdx^5cv mice and is not rescued by μDys. (A) Quantification of protein expression levels of cavins and caveolin-3 (Cav3) in cardiac lysates from 6 months old wild-type (WT), mdx^5cv and μDYS-mdx^5cv (μDYS) mice. Protein levels were normalized to GAPDH probed on the same membrane. Data (mean ± standard deviations; N = 5 mice/group) are expressed as fold differences relative to expression levels in wild-type lysates. No significant differences were found (one-way ANOVA). (B) Representative western blot used for quantifications shown in A. C–D. Immunostaining of heart tissue sections with antibodies to cavins (green) or to laminin-α2 (red) to visualize the membrane of cardiomyocytes and cavin-1 (C) or cavin-4 (D). Blue = nuclei. In C, arrows point to cardiomyocyte membranes to highlight differences in cavin-1 staining between genotypes, while asterisks mark capillaries sitting outside the laminin-α2 outline that are strongly reactive for cavin-1 in all genotypes. E–F. Superplots of immunofluorescence intensity measurements of cavin-1 (E) and cavin-4 (F) at the cardiomyocyte membrane. Small symbols are individual immunofluorescence measurements (N = 40) per mouse. Large symbols indicate the mean for each individual mouse (N = 3/group). Lines indicate the grand mean ± standard deviation. ^**p < 0.01, ^***p < 0.005, two-way repeated measures ANOVA. (G) Immunostaining of heart sections for caveolin-3 (red). Nuclei are counterstained with DAPI (blue). Scale bars: 10 μm.

Figure 5

Perinuclear cavin-4 localization is increased and ERK signalling impaired in mdx^5cv and μDYS-mdx^5cv mice. (A) Triple staining of heart tissue sections from 6 months old wild-type (WT), mdx^5cv and μDYS-mdx^5cv (μDYS) mice for laminin-α2 (cyan) to visualize the outline of cardiomyocytes, DAPI (red) to visualize nuclei and cavin-4 (green). Arrows indicate nuclei located within cardiomyocytes with perinuclear cavin-4 immunofluorescence. No cavin-4 staining was associated with nuclei from interstitial cells. Scale bar: 25 μm. (B) Quantification of cardiomyocyte nuclei with perinuclear cavin-4 labelling relative to the total number of cardiomyocyte nuclei. A minimum of 200 cardiomyocyte nuclei were counted for each mouse. N = 3 mice/group. (C) Single and double (Merge) fluorescence images of a nitrocellulose membrane double-labelled with antibodies to ERK1/2 (green) and phosphoERK1/2 (pERK1/2; red). The bottom part of the membrane was cut and probed with an antibody to GAPDH to ensure comparable protein loading. (D) Densitometric quantification of levels of phosphorylated ERK1/2 (pERK) relative to total ERK1/2 (ERK). N = 5 mice/group. (E) Quantitative RT-PCR analysis of genes that are regulated by ERK during cardiac remodeling. The ΔΔCt method was used to normalize gene expression to Gapdh. Data was then expressed as a fold-change relative to values in wild-type mice. Data in all graphs are mean ± standard deviation. ^*P < 0.05; ^**P < 0.01, ^***P < 0.005, ^****P < 0.001, one-way ANOVA followed by a Bonferroni test adjusted for multiple comparisons.