Unraveling Missing Genes and Missing Inheritance in Arrhythmogenic Cardiomyopathy

Arrhythmogenic cardiomyopathy (AC) is an enigmatic inherited cardiomyopathy characterized by fibro-fatty replacement of cardiac myocytes and ventricular arrhythmias^1-3. When affecting predominantly the right ventricle, this disease is referred as arrhythmogenic right ventricular cardiomyopathy (ARVC). The prevalence of AC in the general population has been estimated in a range from 1 in 2000 to 1 in 5000 individuals. Clinically, it is characterized by palpitations, syncope and sudden cardiac death (SCD) due to ventricular arrhythmias. This cardiomyopathy is a major cause of SCD in the young people and in athletes^1-3. To date, there is no effective pharmacological therapy for this deadly cardiac disease, which may require implantable defibrillators and heart transplantation at the advanced stage⁵.

AC is a hereditary autosomal dominant disease with reduced penetrance and variable clinical expression. Thus far, 5 causal genes have been identified, all coding for desmosomal proteins⁶. These known causal desmosomal genes are plakoglobin (JUP), desmoplakin (DSP), plakophilin-2 (PKP2), desmoglein-2 (DSG2) and democollin-2 (DSC2). Desmosomal proteins are structural proteins important for cell-cell attachment and communication. PKP2 is the most common causal gene for AC, accounting for up to 43% of the cases^7-9. Most of PKP2 mutations are Loss-of-Function (LoF) mutations resulting in reduced expression of PKP2 (haploinsufficiency ^{10, 11}). Atypical forms of AC are caused by mutations in non-desmosomal genes such as TGF-beta-3 (TGFβ3), cardiac ryanodine receptor (RYR2), transmembrane protein 43 (TMEM43) and lamin A/C (LMNA). Current diagnostic strategies rely on NextGeneration DNA Sequencing (NGS) that has poor sensitivity for genomic deletions and identifies causative variants in only ∼50% of the AC cases, leaving half of cases with unknown genetic cause. To address the possibility of genomic mutations, Pilichou et al. turned to Multiplex Ligation-dependent Probe Amplification (MLPA) to quantify genomic deletions in 160 AC patients who had negative NGS testing for the expected AC genes. ¹⁶

The human genome consists of over 3 billion base pairs of deoxyribonucleic acid (DNA) packaged into two sets of 23 chromosomes of which one set of chromosomes is inherited from each parent. These chromosomes contain DNA sequences encoding for approximately 30,000 genes. In general, each gene is presumed to be present in two copies in a genome. However, recent studies have revealed different copy number variants (CNVs) of large DNA segments ranging in size from kilobases to megabases as the source of genetic diversity in the general population. Since then, several research groups have identified CNVs of large segments of chromosomes that encompass genes and that are associated with diseases^12-15. CNVs can cause diseases through different mechanisms: variation of gene dosage through insertions or deletions (a), unmasking of a recessive allele alter the gene expression through inversions, deletions, or translocations (b), gene expression modifications through interaction with regulatory elements (c), or combination of two or more CNVs can produce a complex disease, whereas individually the changes produce no effect (d) which can explain “missing inheritance” which sometimes occurs. The proband can inherit the disease associated CNVs from an unaffected parent, which highlights the incomplete penetrance of the disease resulting from change in gene dosage. Noteworthy, in this issue, Pilichou et al. also showed that several family members carrying a complete PKP2 deletion CNV remained asymptomatic during a mean 6±5 years of follow up period, further supporting the gene dosage theory¹⁶.

In their study, Pilichou et al. identified 9 heterozygous CNVs in 11 of the 160 AC probands of Italian descents (7% of cases) by using MLPA¹⁶. These patients were negative for pathogenic mutations in 5 known AC desmosomal genes screened by conventional genetic testing approaches. By using the MLPA technology, the authors screened these patients for the presence of CNVs in PKP2, DSP, JUP, DSC2, DSG2, TGFβ3 and RYR2 and reported for the first time the identification of heterozygous partial deletion/duplications of DSG2 and DSC2 genes in clinically affected AC patients. Although the authors identified for the first time a de novo duplication of DSC2 exons 7-9 in a patient with definite ARVC, which did not segregate with the disease, it is unknown if this intragenic CNV caused a gain of function, or a loss of function due to disruption of the reading frame. In addition, the authors also identified partial and complete deletions of PKP2 gene in 9 out of 160 probands in agreement with the fact that the majority of NGS sequence mutations cluster in PKP2. Among the 9 probands, 5 probands were carriers of a complete CNV deletion of the PKP2 gene, and possibly shared a founder variant as suggested by haplotype analysis. Together, these findings implicate the involvement of rare AC genomic rearrangements in a subset of AC patients, with a mechanism of loss-of-function, as found in the majority of mutations identified by NGS.

Interestingly, the genetic analysis of Pilichou et al. revealed the possible involvement of digenic heterozygosity in two probands (H-III;1 and K-III,1). One proband was a digenic heterozygous carrier of a point mutation c.2491C>T (p.Leu831Phe) in DSG2 and a CNV that resulted in deletion of exons 6-11 in PKP2. The c.2491C>T (p.Leu831Phe) point mutation in DSG2 is a reported pathogenic variant that affects a non-conserved amino acid. Another proband from family K was a heterozygous carrier of c.536A>G (p.Asp179Gly) in DSC2, a recessive point mutation, and a deletion of 482kb on chromosome 18q that encompassed both DSG2 and DSC2. Both pathogenic point mutations were reported to derive from the unaffected father's allele. These findings raise the notion that, in some AC cases, the additive effects from two pathogenic variants should be considered in determining the association with the phenotype and severity of the disease. Single nucleotide polymorphisms (SNPs) and CNVs may act in concert on different cellular molecular mechanisms contributing to the development of the phenotype. Hence, integrative analysis by combining both SNPs and CNVs data would uncover the underlying phenotype traits of a disease.

Cascade genetic screening of the family members revealed low disease penetrance of 32% in relatives carrying heterozygous CNVs of DSG2, DSC2 and PKP2, implicating that other genetic and non-genetic factors could be involved in the pathogenesis of the disease. The identification of complete deletion of PKP2 genes from affected probands in this study and by other research groups supports the hypothesis that PKP2 haploinsufficiency is not the sole cause of the pathogenesis of AC. The authors found that several family members carrying the complete PKP2 deletion CNV were phenotypically negative. Although, in this study, the authors have isolated mRNA from patient blood cells and performed reverse transcription on selected fragment to confirm the deletion of 44bp in PKP2 exon 4, allelic asymmetries are tissue-specific and the presence of polymorphism in cis-acting elements DNA sequence and its transcription regulatory factors could contribute to the differences in transcription levels in different cell types¹⁷.

In summary, the identification of rare CNVs by Pilichou et al. highlights that CNVs causing large genomics rearrangements may contribute to the heterogeneous clinical phenotypes in the development and progression of AC. The current study expands the notion that routine clinical genetic testing should be expanded to include CNVs identification using MLPA approach in mutation screening for AC patients, since it provides a high-throughput and cost-effective tool, suitable for clinical laboratories. While Pilichou et al. provided suggestive evidence of CNVs in the pathogenesis of AC, further study on the identified CNVs functionality is required to establish causality. It is clear that there is a need to bridge the gap between CNVs identified in AC patients and their effects on gene expression levels, either directly on the affected gene or indirectly through structural alternations effects on nearby genes; thus disease models (either human cell lines or mouse models) are necessary to characterize the molecular pathways affected by the identified CNVs in order to provide a correct diagnosis and potentially identify therapeutic targets to treat AC patients. Although it is beyond the scope of the current study, Pilichou et al. and other research groups have raised the question of what are the genetic factors contributing to the structural rearrangement of chromosome nearby desmosomal gene loci, which still remains unknown.