Integrin α7 Mutations Are Associated With Adult‐Onset Cardiac Dysfunction in Humans and Mice

Abstract

Background

Integrin α7β1 is a major laminin receptor in skeletal and cardiac muscle. In skeletal muscle, integrin α7β1 plays an important role during muscle development and has been described as an important modifier of skeletal muscle diseases. The integrin α7β1 is also highly expressed in the heart, but its precise role in cardiac function is unknown. Mutations in the integrin α7 gene (ITGA7) have been reported in children with congenital myopathy.

Methods and Results

In this study, we described skeletal and cardiac muscle pathology in Itga7 ^−/− mice and 5 patients from 2 unrelated families with ITGA7 mutations. Proband in family 1 presented a homozygous c.806_818del [p.S269fs] variant, and proband in family 2 was identified with 2 intron variants in the ITGA7 gene. The complete absence of the integrin α7 protein in muscle supports the ITGA7 mutations are pathogenic. We performed electrocardiography, echocardiography, or cardiac magnetic resonance imaging, and histological biopsy analyses in patients with ITGA7 deficiency and Itga7 ^−/− mice. The patients exhibited cardiac dysrhythmia and dysfunction from the third decade of life and late‐onset respiratory insufficiency, but with relatively mild limb muscle involvement. Mice demonstrated corresponding abnormalities in cardiac conduction and contraction as well as diaphragm muscle fibrosis.

Conclusions

Our data suggest that loss of integrin α7 causes a novel form of adult‐onset cardiac dysfunction indicating a critical role for the integrin α7β1 in normal cardiac function and highlights the need for long‐term cardiac monitoring in patients with ITGA7‐related congenital myopathy.

Nonstandard Abbreviations and Acronyms

Cx43

connexin‐43

ITGA7

integrin subunit alpha 7 (Human gene)

Itga7

integrin subunit alpha 7 (Mouse gene)

WT

wild‐type

Clinical Perspective.

What Is New?

• This study identifies 5 new patients from 2 families with integrin α7 deficiency that exhibit a novel form of adult‐onset cardiac dysfunction.

• The adult‐onset cardiac dysfunction is mainly characterized by first‐degree heart block and intraventricular conduction delay.

• To our knowledge, this study characterizes for the first time a mouse model that recapitulates human integrin α7 (ITGA7) gene‐related cardiac disease and supports the use of this preclinical model to investigate the mechanistic role of integrin α7 in cardiac function.

What Are the Clinical Implications?

• The present study highlights the requirement for long‐term cardiac monitoring in patients with ITGA7‐related congenital myopathy.

• Our study supports genetic screening for mutations in the ITGA7 gene for patients who present with myopathy of unknown genetic cause.

Integrins are heterodimeric cell receptors composed of α‐ and β‐subunits that act as cell mechanosensors and mechanotransducers. Integrins serve as a transmembrane linkage system that binds extracellular matrix proteins to the cortical actin inside the cell. ¹ , ² , ³ , ⁴ The integrin α7β1 is a major laminin receptor and regulator of skeletal muscle development and disease. In skeletal muscle, integrin α7β1 is localized throughout the myofiber and enriched at myotendinous and neuromuscular junctions, where it plays a critical role during neuromuscular and myotendinous junctions development, myoblast migration, and myofiber survival. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Integrin α7β1 is also expressed in satellite cells and differentiating myoblasts and serves as an important regulator of muscle regeneration and repair. ⁵ , ⁸ , ⁹ , ¹⁶ , ¹⁷ Several isoforms of integrin α7 are generated by alternative RNA splicing. ¹⁸ , ¹⁹ , ²⁰ Integrin α7A and α7B cytoplasmic variants are the major isoforms in human skeletal muscle, ⁹ whereas the α7B isoform is the only one found in cardiomyocytes. ²¹ Integrin α7X1 and α7X2 extracellular isoforms have been shown to bind with different affinities to various laminin isoforms. ²²

The integrin α7β1 is detected during late fetal mouse cardiac development and in adult cardiac muscle. ⁹ , ²¹ , ²³ , ²⁴ , ²⁵ The integrin α7β1 is a major modifier in muscular dystrophy, ¹⁰ , ¹² and mutations in the integrin α7 (ITGA7) gene, which encodes the integrin α7 protein, cause congenital myopathy. Some studies have described skeletal muscle pathology or dysfunction in patients and mice lacking integrin α7, but these only involved a limited number of patients with ITGA7 deficiency with pediatric clinical presentations. ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ Only 1 study has reported a proband carrying variants in ITGA7 and MYH7 (myosin heavy chain beta) genes with features of cardiomyopathy. ²⁹ Other family members demonstrating isolated cardiomyopathy only carried the MYH7 variant, and therefore the role of integrin α7 in cardiac function remains unknown.

Here, we report mutations in the integrin α7 gene are associated with adult‐onset cardiac dysfunction in humans and mice. This study uncovers an important role for integrin α7 in normal cardiac function and identified integrin α7‐related cardiac arrhythmia as a novel form of cardiac dysfunction.

METHODS

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Next Generation Sequencing

Family 1, II.4 was screened using the SureSelect Focused Exome (Agilent) kit, according to the manufacturer's protocol. Sequencing was performed on a Hiseq 1500 (Illumina, CA). The Illumina fastq sequencing data were mapped to the human reference assembly, hg19 (GRCh37; University of California, Santa Cruz  genome browser) by Novoalign Software (Novocraft Inc). After removal of polymerase chain reaction duplicates (Picard) and reads without a unique mapping location, variants were extracted using the Maq model in SAMtools and outputted by the following criteria: consensus quality >30, single nucleotide polymorphism quality >30, and root mean square mapping quality >30. These variant calls were then annotated using Annovar software. Variants with an allele frequency higher than 1% in ExAC, 1000 Genome, or ESP6500 database were filtered out. Synonymous and deep intronic variants were excluded from the analysis. We screened and analyzed 322 genes related to neuromuscular disease and hereditary cardiomyopathies reported in the GeneTable of Neuromuscular Disorders. ³²

Family 2, II.4 was screened using SureSelectXT Human All Exon 50 Mb Kit V5 (Agilent). Sequencing was performed using a Hiseq 4000 (Illumina). Read mapping and variant calling were conducted using Burrows‐Wheeler Aligner software package BWA (http://bio‐bwa.sourceforge.net/index.shtml) and the genome analysis toolkit GATK (https://software.broadinstitute.org/gatk/; Broad Institute, Cambridge), respectively. Annotation of variants was done using in‐house pipelines and a muscle disease gene panel was applied (Genome Diagnostics, Radboudumc & Maastricht University Medical Center, The Netherlands). Additional screening of genes (n=21) associated with hypertrophic cardiomyopathy was performed. Variants with an allele frequency higher than 1% in gnomAD / ExAC or higher than 1% in the Radboudumc exome database of >50 000 control exomes were filtered out. Deep intronic variants were excluded from the analysis (standard settings of the “variant effect predictor” from the ensemble).

Mice Genotyping

Itga7 ^−/− mice were generated and genotyped as previously described. ³³ Heterozygous Itga7 mice were crossed to obtain homozygous Itga7 ^−/− mutants and C57BL/6J wild‐type (WT) control littermates. Mice were euthanized by CO₂ asphyxiation followed by cervical dislocation under the American Veterinary Medical Association guidelines for euthanasia. For this study, a total of 17 Itga7 ^−/− mice and 15 WT littermates were used.

Isolation of Cardiac Myocytes

Single ventricular myocytes were isolated from adult Hartley guinea pigs as described previously. ³⁴ The methods followed The Guide for the Care and Use of Laboratory Animals, as adopted by the National Institutes of Health and approved by the Institutional Animal Care and Use Committee at the University of Nevada, Reno.

Immunofluorescence

Cryosections from human and murine muscle biopsies were incubated with primary antibodies. Human muscle biopsies were incubated for 1 hour at room temperature with the following primary antibodies: integrin α7 (1:400, Abcam Ab75224), integrin α5 (1:80, Abcam Ab6131), integrin β1D (1:100, Abcam Ab8991), laminin α2‐1 (300kDa) (1:50, Alexis 804‐190‐C100), laminin α2‐2 (80kDa) (1:4000, Millipore MAB1922), and laminin α5 (1:4000, Millipore MAB1924). Murine muscle sections were incubated overnight at 4 °C with the following primary antibodies: integrin α7B, ²⁰ integrin β1D, ²⁰ laminin (pan antibody against skeletal muscle laminin isoforms) (Sigma L9393; 1:400), laminin α2 (Sigma L0663; 1:100), and connexin‐43 ([Cx43] Thermo Fisher Scientific CX‐1B1; 1:200). The sections were then washed using 0.1 M PBS pH 7.2. Human muscle biopsy sections were incubated with biotinylated secondary antibodies (1:200) for 30 minutes at room temperature. After washing with PBS, the human muscle biopsy sections were incubated with streptavidin‐conjugated to Alexa Fluor 594 (1:1000, Molecular Probes) for 15 minutes at room temperature. Following another wash, human muscle sections were mounted using Hydromount mounting medium. In parallel, murine muscle sections were incubated with secondary antibodies conjugated with fluorescein isothiocyanate and tetramethylrhodamine (Jackson Immuno Research, 1:200) for 1 hour at room temperature and mounted with Vectashield Antifade Mounting Medium with 4′,6‐diamidino‐2‐phenylindole. Imaging for human muscle biopsies was performed on a Leica DM6B epifluorescent microscope, whereas images from murine muscle specimens were acquired on an Olympus IX81 microscope. The acquired images were analyzed in Fiji version 1.49.

Histology

Sirius Red staining was performed on murine muscle sections. Sections were hydrated in a series of 100%, 95%, and 80% ethanol gradient incubations of 3 minutes each. The sections were then washed in water and stained with Gills Hematoxylin (S5400‐1D, Fisher Scientific) for 10 minutes. After washing, the sections were incubated in Scott solution for 3 minutes, washed once again, and incubated in 0.1% Sirius Red/Picric Acid (SO‐674, Rowley Biochemical) for 30 minutes. The sections were then washed twice for 5 minutes in acidified water and dehydrated in a series of 80%, 95%, and 100% ethanol gradient incubations for 10 minutes each. The dehydration was followed by a 5‐minute incubation with Xylene and mounting with dibutylphthalate polystyrene xylene DEPEX medium.

Electrocardiography

Electrocardiography was performed as validated. ³⁵ 10‐month‐old Itga7 ^−/− mice (n=8) and wild‐type littermates (C57BL/6J) (n=10) were anesthetized with 2% isoflurane mixed with 0.5 L/min 100% O₂ and kept on a heating pad to maintain constant body temperature. Three shielded lead wire electrodes (AD Instruments, Australia) were inserted subcutaneously. Lead II tracings were recorded using PowerLab 6/26 supplemented with an animal Dual BioAmp and analyzed with LabChart Pro software (AD Instruments). Lead II channel was amplified and sampled at a rate of 4 kHz and 5 mV range of a high‐pass filter setting of 1 Hz. Averaged electrocardiography parameters were obtained from continuous 10‐minute recordings.

Echocardiography

Transthoracic echocardiography was performed on 10‐month‐old Itga7 ^−/− mice (n=12) and WT littermates (C57BL/6J) (n=12) following published guidelines. ³⁶ Mice were anesthetized with 2% isoflurane mixed with 0.5 L/min 100% O₂ and kept on a heating pad to maintain body temperature. Scanning was performed using the Vevo2100 image system with a high‐resolution transducer at a frequency of 40 MHz (MS550) (FUJIFILM VisualSonics, Canada). Two‐dimensional B‐mode images and left ventricular M‐mode tracings were acquired from the parasternal short‐axis view at the papillary muscles level. M‐mode measurements were analyzed using Vevo LAB v3.2.0 (FUJIFILM VisualSonics).

Study Approval

The human study was conducted in accordance with the Declaration of Helsinki. Informed consent for genetic research testing was obtained from all patients where applicable and the study was performed with institutional ethical approval. Family 1 was evaluated at the Queen Square Centre for Neuromuscular Diseases, National Hospital for Neurology and Neurosurgery, and at the Barts Heart Centre, London, UK. Family 2 was assessed at the Department of Neurology, Radboud University Medical Center, Nijmegen, The Netherlands.

All procedures involving mice were performed under the approved protocol 00399 from the Institutional Animal Care and Use Committee of the University of Nevada, Reno.

Statistical Analysis

All data are expressed as mean ± SEM. All measurements followed the normal distribution. Student t‐test was performed on WT versus Itga7 ^−/− values for electrocardiography and echocardiography data using GraphPad Prism software. Representative P values and symbols are described in the figure legends.

RESULTS

We identified 5 individuals from 2 unrelated families (Figure 1A and 1B) with a clinical presentation of adult‐onset cardiac conduction defects, respiratory insufficiency, and mild skeletal myopathy (Table S1). The proband in family 1 is a 52‐year‐old man from Kashmir (Figure 1A; II.4). Next generation sequencing identified a homozygous c.806_818del [p.S269fs] variant in ITGA7 (NM_002206.3) that was confirmed by Sanger sequencing (Figure S1). The variant is extremely rare in the general population (0.000004 in Gnomad v2.1.1) with no homozygous case reported. No alternative or concomitant relevant pathogenic variant in other genes was found. Aside from isolated stridor at birth because of vocal cord paresis, early history was unremarkable. He presented with palpitations, first‐degree heart block (PR interval 200 ms), intraventricular conduction delay, and left axis deviation at age 37 years (data not shown). By the age of 47 years, the PR interval on his ECG had progressed to 222 ms and QRS duration to 136 ms with intraventricular conduction delay (Figure 1Ca). Further investigation revealed paroxysmal atrial flutter that was successfully treated by ablation. Subsequent 24‐hour Holter analysis identified frequent ventricular ectopy (>3000 PVC/24 hours) including runs of nonsustained ventricular tachycardia, as well as intermittent episodes of Mobitz type 1 second‐degree atrioventricular block. Transthoracic echocardiography and cardiac magnetic resonance imaging demonstrated localized left ventricular hypertrophy confined to the basal‐mid inferior wall (maximal wall thickness 15 mm) with preserved left ventricular systolic function (Figure 1Cb and 1Ce). Computed tomography coronary angiography was normal. Given the presence of progressive conduction disease, together with evidence of nonsustained ventricular tachycardia and new symptoms of presyncope, the patient received an implantable cardioverter‐defibrillator. Cardiac assessment of an elder brother (Figure 1A; II.3) revealed similar cardiac pathology with conduction abnormalities, mild left ventricular hypertrophy, and limited fibrosis (see Table S1 and Figure S2). Quadriceps muscle histopathology was consistent with a mild myopathy (Table S1). Respiratory insufficiency with nocturnal hypoventilation and chronic respiratory acidosis requiring bilevel positive airway pressure therapy was evident from his early 50s (Table S1).

Figure 1. Patient family pedigree and respective cardiac investigation.

Pedigree of families 1 and 2 (A and B) with filled blue shapes representing affected family members. In (A), arrow indicates the family 1 proband (A‐II:4), and the black plus signs (+) indicate the identified alleles c.806_818del variant in integrin α7 gene (NM_002206.3). In (B), arrow indicates the family 2 proband (B‐II:4), the black plus signs indicate the alleles c.2357+1G>A and red plus signs indicate the alleles c.2278‐1G>A variants in integrin α7 gene identified in the genetically tested family members. Electrocardiography in family 1 proband (II:4) demonstrated first‐degree heart block, intraventricular conduction delay, left axis deviation, and PR interval 222 ms and QRS duration 136 ms (Ca). Cardiac magnetic resonance imaging, 3 chamber cine, end‐diastole, showing basal septal mild hypertrophy (white arrowhead) and 3 basal inferolateral crypts (white arrows) (Cb). Cardiac magnetic resonance imaging, basal short‐axis late gadolinium enhancement image showing fibrosis in the inferolateral wall (white arrow) (Cc). Cardiac magnetic resonance imaging, basal short‐axis cine, end‐diastole, showing localized hypertrophy in the basal inferior wall (white arrow) (Cd). Cardiac magnetic resonance imaging, 4 chamber cine, end‐diastole, revealing mild basal septal hypertrophy (white arrow) (Ce).

Next generation sequencing in the proband from family 2 (Figure 1B; II.4) identified 2 intron variants in the ITGA7 gene (NM_002206.3) (c.2357+1G>A [r.spl?] and c.2278‐1G>A [r.spl?]) affecting canonical splicing sites. These variants are extremely rare in the general population (MAF 0.000035 and no cases reported, respectively) with no homozygous cases reported. No alternative or concomitant relevant pathogenic variant in other genes was found. The presence of these variants was confirmed by Sanger sequencing (Figure S1), and the complete absence of integrin α7 staining in the muscle supports that both mutations are pathogenic and likely in trans position (Figure 2G). She was hospitalized at age 61 years with a history of relapsing respiratory insufficiency of unknown cause from her early 50s. Medical history was notable for cataract, left bundle branch block, atrial flutter, and high‐grade atrioventricular block necessitating implantation of a dual‐chamber rate‐modulated pacemaker. Aside from a pCO₂ of 10.87 kPa, the neuromuscular examination was unremarkable with only mild bilateral weakness of ankle dorsiflexion (Medical Research Council score 4+). Ascites were present on the systemic examination. Investigations including quadriceps muscle biopsy were indicative of a myopathic disorder (Table S1). Family history revealed that a sister (Figure 1B; II‐7), who had right ventricular failure, pulmonary hypertension, and restrictive pulmonary disease of unknown cause, had died of respiratory insufficiency at age 55 years. Another sister (II‐8) was assessed at the age of 46 for limb‐girdle muscular weakness and respiratory failure postoperatively. Following genetic analysis in the proband, the same ITGA7 variants were identified in her affected siblings (Figure 1B).

Figure 2. Myopathy in patients with integrin α7 gene deficiency.

Quadriceps muscle biopsies from the family 1 proband taken at 37 years (D through F), the family 2 proband (G through I, M through O) taken at 62 years, and an unaffected adult control (A through C, J through L). Immunofluorescent staining of frozen sections is shown for integrin α7 (A, D, and G), integrin β1D (B, E, and H), integrin α5 (C, F, and I), laminin α2 with 2 different antibodies (J, M, K, and N), and laminin α5 (L and O). There is a complete absence of integrin α7 (D and G) and a reduction of integrin β1D (E and H) at the sarcolemma in both patients compared with the control (A and B). Integrin α5 staining is present on endomysial capillaries (F and I) comparable to the control (C). In the family 2 proband, immunostaining for laminin α2 (M and N) is comparable with the control (J and K), and there is also considerable upregulation of laminin α5 (O) at the basal lamina compared with the normal endomysial vascular labeling in the control (L). Scale bar: 100 μm.

To characterize the myopathic disorder in more detail, we reanalyzed quadriceps muscle biopsies from the 2 probands and an unaffected adult control muscle biopsy by immunofluorescence (Figure 2). We detected an absence of sarcolemmal localized integrin α7 in both probands' biopsies (Figure 2D and 2G) compared with the unaffected adult control (Figure 2A) and a reduction of integrin β1D in both probands' biopsies (Figure 2E and 2H) compared with the unaffected adult control (Figure 2B). Integrin α5 expression was seen in the endomysial capillaries of both patients' biopsies (Figure 2F and 2I) similar to the adult control (Figure 2C). Laminin α2 expression in the family 2 proband biopsy (Figure 2M and 2N) was comparable with the adult control (Figure 2J and 2K). Laminin α5 expression was restricted to the endomysial vasculature in the adult control (Figure 2L) but secondarily upregulated on the myofiber basal lamina in the family 2 proband biopsy (Figure 2O).

We next analyzed muscles from the Itga7 ^−/− mouse model (19). Analysis of 8‐month‐old Itga7 ^−/− mice diaphragm muscle revealed a decrease in integrin β1D (Figure S3E), but no change in laminin α2 (Figure S3F) compared with WT controls (Figure S3A and S3B). Laminin α5 was upregulated in the diaphragm of Itga7 ^−/− mice (Figure S3G) compared with WT controls (Figure S3C), but no significant changes in laminin α2 or α5 were observed in the gastrocnemius muscle of Itga7 ^−/− mice (Figure S3M and S3N) compared with WT controls (Figure S3I and S3J), suggesting that changes in laminin α5 protein in the diaphragm muscle may serve as a compensatory mechanism.

We also analyzed the fibrotic content in skeletal muscle of Itga7 ^−/− mice. We detected increased fibrosis in the diaphragm (Figure S3H) and gastrocnemius (Figure S3O and S3P) muscles of Itga7 ^−/− mice compared with WT (Figure S3D, S3K and S3L). Whereas the collagen content was increased throughout the diaphragm (Figure S3H), Itga7 ^−/− gastrocnemius muscle exhibited only a few fibrotic regions (Figure S3O and S3P), indicating the diaphragm is more affected in 8‐month‐old mice.

We next characterized electrical and mechanical cardiac function in Itga7 ^−/− mice. Abnormal electrocardiography patterns and mechanical function were detected at 10 months of age in Itga7 ^−/− hearts (Figures 3 and 4). Lead II electrocardiography analysis showed a significant increase in PR, QRS, and QT intervals in Itga7 ^−/− mice compared with WT (Figure 3F through 3H). We also observed increased P duration in Itga7 ^−/− mice (Figure 3E). In the cardiac cycle, we found a decrease in R amplitude in Itga7 ^−/− murine hearts compared with WT hearts (Figure 3K), while no significant differences in P, Q, S, and T amplitudes, heart rate, and RR interval were detected (Figure 3I, 3J, 3L, 3M, 3C, and 3D).

Figure 3. Electrocardiography analysis in wild‐type and Itga7 ^−/− mice.

Representative traces are shown from a lead II ECG from 10‐month‐old wild‐type (n=10) (A) and Itga7 ^−/− mice (n=8) (B). Quantitative evaluation reveals prolonged P duration (E) and PR interval (F) suggesting alterations in conduction through the atrioventricular node, and prolonged QRS interval (G) and QT interval (H) in Itga7 ^−/− mice indicating ventricular conduction delay. No significant difference was found in heart rate (C), RR interval (D), P amplitude (I), Q amplitude (J), S amplitude (L), and T amplitude (M). In the absence of integrin α7, R amplitude (K) is significantly decreased. Student t‐test analysis represented by statistical significance of mean±SEM.*P<0.05, **P<0.01. Itga7^{−/ −} , indicates integrin α7 knockout; and WT, wild‐type.

Figure 4. Echocardiography analysis in wild‐type and Itga7 ^−/− mice.

Representative echocardiograms (M‐mode view) of 10‐month‐old wild‐type (n=12) (A) and Itga7 ^−/− mice (n=12) (B). Echocardiographic measurements show reduced ejection fraction (D), fractional shortening (E), cardiac output (F), and stroke volume (G) in Itga7 ^−/− mice suggesting systolic dysfunction. No significant differences were found in left ventricular (LV) mass/body weight ratio (C), LV systolic volume (H), LV diastolic volume (I), LV posterior wall thickness at systole (J), LV internal diameter at systole (L), LV internal diameter at diastole (M), and LV anterior wall thickness at diastole (O). Itga7 ^−/− mice presented a significant increase of about 14% in LV posterior wall thickness at diastole (K) and a decrease of about 20% in LV anterior wall thickness at systole (N). Student t‐test analysis represented by statistical significance of mean ± SEM.*P<0.05, **P<0.01. d indicates diastole; Itga7 ^−/−, integrin α7 knockout; LV, left ventricle; LVAW, left ventricular anterior wall; LVID, left ventricular internal diameter; LVPW, left ventricular posterior wall; and WT, wild‐type.

Mechanical cardiac function parameters were significantly reduced in 10‐month‐old Itga7 ^−/− mice compared with age‐matched WT controls. Itga7 ^−/− mice presented reduced ejection fraction, fractional shortening, cardiac output, and stroke volume (Figure 4D through 4G). We also observed a 14% increase in the left ventricular posterior wall thickness at diastole (Figure 4K) and a decrease of about 20% in the left ventricular anterior wall thickness at systole (Figure 4N) in Itga7 ^−/− mice compared with WT controls. No significant differences between the 2 groups were observed in the other parameters analyzed (Figure 4C, 4H, 4I, 4J, 4L, 4M, and 4O).

To further elucidate the role of integrin α7 in cardiac function, WT and Itga7 ^−/− murine cardiac muscle sections were immunolabeled for integrin α7B, α5, α6, β1D, laminin α2, and Cx43 (Figure S4). While integrin α7B is absent in Itga7 ^−/− hearts, as revealed by immunohistochemistry, in the ventricle anterior wall (Figure S4F) and in the compact ventricle myocardium (Figure S4N), we did not detect major differences in integrin β1D expression (Figure S4G) when compared with WT controls (Figure S4B, S4J, and S4C). Integrin α5 was localized to the intercalated discs and integrin α6 was localized to the lateral membrane of Itga7 ^−/− hearts (Figure S4O and S4P) similar to WT hearts (Figure S4K and S4L). The absence of integrin α7 did not affect the expression of laminin‐α2 (Figure S4E and S4I), β‐dystroglycan (Figure S4M and S4Q), talin (Figure S5A and S5C), or vinculin (Figure S5B and S5D).

Cx43 (connexin‐43) is widely expressed in the heart and can be found in the lower brand branches, Purkinje fibers, and atrial and ventricular cardiomyocytes. ³⁷ We confirmed Cx43 protein localization at gap junctions in adult WT mouse cardiac muscle (Figure 5A through 5E) and cardiomyocytes isolated from the guinea pig (Figure S6A through S6D) where it colocalizes with integrin α7 at the intercalated discs of cardiomyocytes. We also detected Cx43 in Itga7 ^−/− Purkinje fibers (Figure S4D and S4H). Interestingly, we identified Cx43 redistribution to the lateral membrane in Itga7 ^−/− cardiomyocytes (Figure 5F through 5J). Additionally, histological qualitative analysis of fibrosis in Itga7 ^−/− mice revealed no difference in the fibrotic content between the hearts of WT (Figure 5K and 5K′) and Itga7 ^−/− mice (Figure 5L and 5L').

Figure 5. Histological analysis of wild‐type and Itga7 ^−/− cardiac muscle.

Transversal view of immunostaining for integrin α7 (green) (A, C, D, F, H, and I) and connexin‐43 (red) (A, B, C, E, F, G, H, and J) in wild‐type (A‐E) and Itga7 ^−/− (F through J) cardiac muscle. Connexin‐43 localization is restricted to the intercalated discs in wild‐type cardiomyocytes (A, B, C, and E), whereas connexin‐43 is redistributed to the lateral membrane of the Itga7 ^−/− (F, G, H, and J) cardiomyocytes. Sirius Red staining for wild‐type (K and K′) and Itga7 ^−/− (L and L') cardiac muscle reveals no significant changes in fibrosis levels in the absence of integrin α7. Yellow arrows indicate positive staining for integrin α7 (D) and connexin‐43 (E and J). The merge images include 4'6‐diamidino‐2‐phenylindole  (blue), integrin α7 (green), and connexin‐43 (red). Cx43 indicates connexin‐43; Itga7 ^−/−, integrin α7 knockout; and WT, wild‐type. Representative images of n=3 for each genotype. Scale bar: 25 μm.

DISCUSSION

We report that ITGA7 mutations that result in loss of integrin α7 protein might cause adult‐onset cardiomyopathy with cardiac conduction defects and mechanical dysfunction, respiratory insufficiency, and myopathic features. Our data show that loss of integrin α7 in mice leads to myopathy with predominant diaphragmatic involvement resembling the clinical features observed in the patients. The lack of integrin α7 also results in cardiomyopathy with cardiac conduction defects and mechanical dysfunction in Itga7 ^−/− mice that resemble the probands who presented first‐degree heart block and intraventricular conduction delay.

We also identified the importance of integrin α7β1 in the localization of Cx43 in murine cardiomyocytes which when altered could contribute to the pathological mechanism of ITGA7‐related cardiac arrhythmias and systolic dysfunction. Cx43 is an important component of gap‐junction channels located at the intercalated discs where it mediates cardiomyocyte contraction by regulating intercellular ion propagation in cardiac muscle. ³⁸ , ³⁹ , ⁴⁰ Interestingly, Cx43 redistribution to the lateral membrane of cardiomyocytes has been shown to contribute to altered cardiac electrical conduction, arrhythmias, and reduced contractile function of the heart. ⁴¹ Cx43 remodeling has also been associated with many cardiac diseases including Duchenne muscular dystrophy‐related arrhythmogenesis, ⁴² , ⁴³ hypertrophic cardiomyopathy, and heart failure. ⁴⁴ Altered electrical conduction and reduced systolic function observed in Itga7 ^−/− mice may be related to the Cx43 remodeling observed in Itga7 ^−/− cardiomyocytes. Taken together, our results suggest that Cx43 redistribution might be associated with the mechanosensing and mechanotransduction function of integrin α7 in cardiac muscle.

Another possible consequence of the loss of ITGA7 may include impaired L‐type calcium channel activity involved in ventricle depolarization. ⁴⁵ This was previously shown to be regulated by laminin and integrin β1 interactions in cat atrial myocytes. ⁴⁶ Kwon et al. have also found that integrin α7β1 regulates voltage‐gated L‐type calcium channels in myogenic cells through laminin‐binding or integrin α7 activating antibodies. ⁴⁷ Moreover, studies have shown integrin α7 interacts with and stabilizes RyR2 (ryanodine receptor 2) to protect cardiomyocytes from ischemia/reperfusion injury. ⁴⁸ Further studies will be needed to elucidate if any of these mechanisms also contribute to the altered cardiac electrical activity observed in patients with ITGA7‐deficiency and Itga7 ^−/− mice.

Together, this study identifies a potential novel role for the integrin α7β1 in normal cardiac function and indicates a potential new genetic cause for idiopathic adult‐onset cardiac dysfunction. Our study also highlights the need for long‐term cardiac monitoring in patients with ITGA7‐related congenital myopathy. Finally, our study supports genetic screening for mutations in the ITGA7 gene for patients who present with myopathy and predominant respiratory involvement.

Sources of Funding

This study was supported by grants from the Muscular Dystrophy Association (MDA238981 and MDA628561) and National Insitutes of Health ‐ National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIH‐NIAMS) (R01AR064338; R01AR053697) to D. Burkin. A. Oliveira‐Santos was supported by a Raymond H. Berner Graduate School Scholarship. T. Fontelonga was supported by a Mick Hitchcock Scholarship. Dr Phadke was supported by National Health Service (NHS) England Highly Specialised Service for Congenital Myopathies and Congenital Muscular Distrophies. Dr Matthews was supported by a Wellcome Clinical Research Career Development Fellowship. Dr Lopes was supported by a Medical Research Council UK Clinical Academic Research Partnership award. Part of this work was undertaken at University College London Hospitals/University College London, which received a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centres funding scheme.

Disclosures

The authors have declared that no conflict of interest exists.

Supporting information

Table S1

Figures S1–S6