Identification of established arrhythmogenic right ventricular cardiomyopathy mutation in a patient with the contrasting phenotype of hypertrophic cardiomyopathy

The patient is a 61-year white (European decent, non-Finish) female who presented with unexplained syncope 4 years ago. She had no prior history of syncope, cardiac arrhythmias, cardiovascular diseases, and no family history of cardiomyopathies. She had a past medical history of depression and mild hypothyroidism, which have been controlled with medical therapy. Her physical examination was unremarkable. A 12-lead ECG was notable for possible left atrial enlargement, q waves in L1 and aVL, a QRS morphology resembling right bundle branch block with an epsilon wave, and non-specific ST and T changes (Fig. 1a and b). Holter monitoring (24 h) showed a run of 5-beat ventricular tachycardia and frequent ventricular premature beats and couplets (Fig. 1c). An echocardiogram showed asymmetric left ventricular hypertrophy with an interventricular septal thickness of 19 mm (normal range: 6–11 mm), a posterior wall thickness of 1.1 cm, a hyperdynamic left ventricle with an ejection fraction of > 60%, a normal right ventricular size and function, and reduced myocardial tissue Doppler velocities (Fig. 1, d–g). She had no overload conditions, such as systemic arterial hypertension or valvular pathology to explain cardiac hypertrophy. The findings were consistent with the diagnosis of HCM. She was treated with a beta-blocker for HCM, which she soon discontinued. She was taking aspirin (81 mg/d), cholecalciferol (1000 units/d), and levothyroxine (75 mg/d), and received monthly intramuscular injection of paliperidone palmitate for depression.

The proband had no family history of HCM or ARVC (Fig. 2). She was never married and had no child. Proband’s mother (I-2 in the Pedigree) had atrial fibrillation with a controlled ventricular rate at 80 bpm and left anterior fascicular block on a 12-lead ECG but no Epsilon wave or evidence of cardiac hypertrophy. Individual I-2 had an echocardiogram at 80 years of age, which was 4 years prior to her death, which showed a normal left ventricular size and function with an interventricular wall thickness of 0.9 cm. She had no evidence of right ventricular dilatation or dysfunction on the echocardiogram. Proband’s father (I-1 in the Pedigree) had a history of coronary artery disease and died in a car accident at an old age. Proband’s brother (II-2) had no history of heart disease and died from septic shock and respiratory failure, as complications of liver cirrhosis caused by hepatitis C. Proband’s half-sister had degenerative aortic valve disease and underwent aortic valve replacement in her 50s.

Whole exome sequencing (WES) was performed on the Illumina 2500 platform as part of clinical genetic testing by a commercial company. Sequencing produced 2 × 100 bp reads, which were aligned to the human reference genome (hg19) using BWA-Mem (v0.7.8) [4]. The BAM files, containing the sequence alignment data, were analyzed to identify the GVs using Platypus (v0.8.1) [5]. CASSANDRA (v2.0) was used to annotate the GVs [6]. The discovered variants were filtered by minor allele frequency <1% (ExAc overall minor allele frequency, http://​exac.​broadinstitute.​org/​) and prioritized providing they met any of the following criteria:

The mean depth of coverage of exons in the entire dataset was 178 times. In the entire data set, 95.3% of the exons had at least 10 sequence reads. Further analysis of the 15 genes known to cause HCM and phenocopy conditions, namely MYBPC3, MYH7, TNNT2, TNNI3, TPM1, ACTC1, MYL2, MYL3, FH11, NEXN, PLN, GLA, LAMP2, and PRKAG2 showed 100% had at least 10 reads and 98% had at least 20 reads (Table 1).

A total of 80 candidate GVs in the WES data, including 29 insertions/deletions, 16 variants a introducing premature stop codon, 3 variants affecting the splicing sites, and 32 deleterious non-synonymous variants met the filtering criteria (Additional file 1: Table S1). No pathogenic variant in any of the established causal genes for HCM was identified. Given the high read coverage (Table 1), the possibility of missing a causal variant in HCM genes was very low. Ten of the candidate pathogenic variants were in genes associated with a phenotype per Online Mendelian Inheritance in Man (OMIM) database [10]. The list included CLDN16, COL6A2, COL6A3, FANCD2, GPR179, LFNG, NOTCH3, PKP2, POF1B, TGIF1 (Table 2).

Among the pathogenic variants identified, the most plausible pathogenic variant for the phenotype in the proband is a rare splice variant (c.2146-1G > C) in the PKP2 gene (NM_001005242.2), which encodes plakophilin 2 (<1 in 20,000 in ExAc database in European – Non-Finnish population). PKP2 is the most common causal gene for ARVC [11]. The variant is predicted to disrupt the canonical splice acceptor site in exon 10 of the PKP2 gene and has been reported in multiple independent probands/families to cause ARVC [11‐14]. The variant was read 183 times in the WES data and was confirmed by Sanger sequencing (Fig. 3). Because the c.2146-1G > C variant is known to cause ARVC [11‐14], the patient was further evaluated for this distinct form of cardiomyopathy. The ECG and echocardiogram were unchanged. CMR showed interventricular septal hypertrophy but no evidence of right ventricular dilatation, dysfunction, aneurysm, or fibro-adipocyte infiltration (Fig. 1, panels f and g, and Additional file 2: Movie S1 and Additional file 3: Movie S2).

Although different mutations in a single gene can cause distinct phenotypes, the observed association of a single mutation with two distinct and contrasting phenotypes, namely HCM and ARVC, is rare, if not unique. The former is characterized by cardiac myocyte hypertrophy, classically in the left ventricle and the latter by myocyte atrophy, apoptosis, and excess fibro-adipocytes, classically in the right ventricle. The patient had a clear clinical diagnosis of HCM, as evidenced by the presence of asymmetric cardiac hypertrophy with a predominant involvement of the interventricular septum, on multiple echocardiograms and CMR, in the absence of a secondary cause, such as systemic arterial hypertension or aortic stenosis. However, the patient did not meet the clinical criteria for the diagnosis of ARVC, despite carrying a well-established mutation in the most common causal gene for ARVC [11]. The QRS pattern resembling an epsilon wave is considered a major, albeit insufficient, diagnostic criterion for the ARVC, evoking possible ARVC [15]. The absence of a full ARVC phenotype in the proband might reflect incomplete penetrance, which is a known feature of the PKP2 variants [14]. The Epsilon wave observed in the proband is not a feature of HCM but is a major diagnostic criterion for ARVC. The electrocardiographic pattern also resembles the pattern observed in Brugada syndrome [16], which is also associated with the PKP2 mutations [17]. However, cardiac hypertrophy is not a feature of the Brugada syndrome.

The published evidence for the pathogenic role of the c.2146-1G > C variant in ARVC is strong [11‐13]. However, its causal role in HCM is uncertain. Rare variants, including the pathogenic variants, are population-specific. Consequently, detection of a pathogenic variant, previously identified as a disease-causing variant, in a single individual is insufficient to imply causality. Indeed, unambiguous ascertainment of genetic causality, regardless of the causal gene and variant, is exceedingly challenging if not impossible. HCM in this patient might be caused by undetected pathogenic variants, structural variants, and variants did not meet the filtering criteria [18]. The exome of this patient also contained a rare variant in the OBSCN gene, which is a candidate gene for DCM but not an established causal gene for HCM [19]. As shown in Table 1, WES had provided an excellent coverage to 15 known HCM and HCM-phenocopy genes. Thus, it is very unlikely that a pathogenic variant in the common HCM gene was undetected. Finally, it is also possible that a copy number variant or a large deletion, not detected by whole exome sequencing, is responsible for HCM in this case, albeit such variants are found in < 1% of HCM cases [20]. Thus, one cannot exclude concomitant presence of HCM and ARVC phenotypes in this case.