Whole-exome sequencing of cases with familial cardiomyopathy

Introduction: Cardiomyopathies are disorders of the myocardium that can lead to heart failure, arrhythmias and sudden death. Heritable forms include dilated, hypertrophic and arrhythmogenic cardiomyopathy (DCM, HCM and ACM respectively). As heterogeneous disorders, over 50 genes have been implicated in these cardiomyopathies to date. However, the yield of genetic testing ranges from less than 40% in idiopathic DCM to over 50% in ACM and HCM, indicating that many causal genes are yet to be identified. This is particularly true in African populations, where the genetics of cardiomyopathy is underexplored. In a review of the role of next-generation sequencing in gene discovery, over 20 new cardiomyopathy genes were found to have been identified through exome sequencing of cardiomyopathy patients. The literature review also highlighted the need for functional validation of newly identified disease genes. Therefore, the aims of this investigation were to utilise exome sequencing to identify disease-causing mutations in South African families with heritable cardiomyopathy, and to establish methods of variant validation through functional modelling in zebrafish. Methods: Five probands and 34 relatives were included in this investigation. The probands and their relatives were clinically examined and diagnosed with DCM, HCM or ACM at Groote Schuur Hospital, Cape Town. Exome sequencing was performed on each of the five probands as well as at least one other family member. Variants of interest were identified by filtering the exome sequencing data by allele frequency, variant quality, variant consequence, predicted deleteriousness, and the potential inheritance patterns as determined by family history analysis. Variants occurring in known cardiomyopathy genes were prioritised, but genes outside the cardiac panel were considered based on literature mining, expression in the heart, and results of prior animal models. Candidate variants were validated by Sanger sequencing and assessed using international criteria for pathogenicity. The candidate ACM gene POLG was investigated in zebrafish larvae using two genetic manipulations. Firstly, zebrafish polg was disrupted using CRISPR/Cas9 in single-cell embryos and, at three days post-fertilisation, the phenotypic effects were compared to uninjected control larvae, as well as larvae in which other known cardiomyopathy genes were disrupted. Secondly, human POLG cDNA was cloned, and the c.2942A>G variant introduced using site-directed mutagenesis; this construct was used to generate variant POLG mRNA that was injected into zebrafish embryos. Larvae were phenotypically examined at four days post-fertilisation and compared to three control groups (unmutated POLG-injected, water-injected, and uninjected embryos). Results: In three families, genotype-phenotype correlations were identified that have not yet been reported in South Africa, although this genetic overlap between cardiomyopathies has been described elsewhere. Family 1: the mutation MYH7 c.4394C>T (p.S1465L) was identified in three siblings with DCM. Although MHY7 is typically associated with HCM, mutations in this region have been reported in DCM patients in other populations. Family 2: the mutation GLA c.774_775del (p.R259Rfs*5) was found in a mother and her son, both of whom had been diagnosed with HCM. The finding of a pathogenic truncating GLA mutation in this family resulted in the genetic rediagnosis of those individuals with Fabry disease, an HCM phenocopy. Family 3: in this large DCM family consisting of three affected brothers and their nephew, no pathogenic variants were identified, but two variants of uncertain significance (VUSs) were found in the genes DSC2 and PKP2. Both variants fulfilled some criteria for pathogenicity, but have not been associated with DCM in South African patients before. In Families 4 and 5, no mutations in known cardiomyopathy-causing genes were identified. Family 4: exome sequencing revealed the variant POLG c.2492A>G (p.Y831C) in this ACM family, with a clinical phenotype consisting of arrhythmia and left ventricular fibrosis. This was a VUS, but in vivo modelling using CRISPR/Cas9 in zebrafish larvae demonstrated that disruption of the gene may impair cardiac development, while expression of the c.2492A>G variant in zebrafish larvae resulted in a significant reduction in heart rate, ventricle size and cardiac output. These results indicate that POLG variation may underly the arrythmia observed in the family, while prior mouse models reported that POLG mutations can induce cardiac fibrosis. Family 5: rare, compound heterozygous missense mutations in ITGB5 were identified as the candidate causative variants in this small family with severe paediatric DCM, possibly affecting adhesion of cardiomyocytes to the extracellular membrane. Conclusion: In total, pathogenic or likely pathogenic mutations were identified in two out of five families studied, while three VUSs with moderate or strong pathogenic potential were identified in two other families. The potential role of POLG in human cardiomyopathy and arrhythmic phenotypes is a finding that should be explored further, as should the putative role of ITGB5 in paediatric cardiomyopathy. This study indicates how exome sequencing, combined with in vivo functional analysis, can identify variants that are likely to contribute to disease in human patients. These techniques may prove useful in bridging the gap in cardiomyopathy knowledge in Africa