Background

Arrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited cardiac disorder, affecting predominantly the right and sometimes the left ventricle and is characterized by progressive fibro-fatty replacement of ventricular myocardial tissue [1,2,3,4,5]. Clinical manifestations include recurrent, typically exercise-related ventricular arrhythmias, syncope, heart failure and sudden cardiac death [6,7,8]. Initial presentation is generally at adolescence or young adulthood [3]. Diagnosis is based on the revised 2010 Task Force Criteria (TFC) by Marcus et al., establishing complex requirements for right ventricular (RV) function and structure, electrocardiographic (ECG) findings and genetic or familial background of the disease [9]. Prevalence in the general population has been estimated to be from 1:5000 to 1:1000 [10, 11].

Mutations in genes encoding cardiac desmosomal proteins are the major determinants of ARVC [12]. These include plakophilin-2 (PKP2), desmoplakin (DSP), desmoglein-2 (DSG2), desmocollin-2 (DSC2) and plakoglobin (JUP) genes [13]. The mode of inheritance is typically autosomal dominant with incomplete penetrance and variable expression [14]. Rare recessive forms may be seen either as sole cardiomyopathy or in conjunction with systemic cutaneous disorders with palmoplantar keratoderma and woolly hair [15,16,17,18,19]. Recent studies have revealed that cardiac desmosomal gene mutations also associate with dilated cardiomyopathy (DCM) phenotype [20, 21]. Mutations in the DSP gene explained 5.5% of DCM cases in a Finnish study cohort [22].

As much as 30–50% of ARVC patients carry at least one variant classified as disease causing [23, 24]. However, molecular genetic diagnostic yield is known to be highly variable, at least partly due to different variant classification practices. Variability in clinical manifestations, reduced penetrance and digenic inheritance, along with disease modifying lifestyle factors such as exercise act together in complicating the determination of conclusive inheritance patterns [25,26,27,28]. End-stage ARVC is also sometimes difficult to distinguish from DCM, with significantly overlapping morphology due to left ventricular (LV) involvement and apparent similarities in molecular genetic etiology [29]. Thus, more research is needed to reveal a group of genes associated with the ARVC phenotype, their inheritance patterns, and genotype-phenotype associations in the disease.

Here, we report two Finnish pedigrees (n = 20) in which a homozygous DSG2 variant c.1003A > G, p.(Thr335Ala) co-segregated with ARVC, indicating autosomal recessive inheritance and complete penetrance. All heterozygous mutation carriers were healthy in terms of cardiac disease, even though this variant has been previously described as disease causing in a heterozygous state.

Case presentation

The index patients of these two families fulfilled the revised diagnostic criteria for ARVC by Marcus et al. [9]. Family history was obtained and pedigrees were drawn. Our clinicians assessed the adult family members at Helsinki University Hospital or Päijät-Häme Central Hospital where possible by physical examination, resting 12-lead ECG, signal-averaged electrocardiography (SAECG), appropriate laboratory tests and transthoracic echocardiography or cardiac magnetic resonance imaging (MRI). From those already deceased, we acquired all available hospital records concerning cardiac diseases. The clinical diagnoses of the family members were also based on the 2010 TFC. All participants are of Finnish ethnicity.

Molecular genetic studies

Blueprint Genetics carried out genetic testing using targeted sequencing panels utilizing oligonucleotide-selective sequencing (OS-Seq) [30]. Family 1 proband’s genetic evaluation was performed using the Pan Cardiomyopathy Panel, covering 103 genes associated with pediatric and adult onset cardiomyopathies and their phenocopies. Family 2 proband was genetically evaluated using the Heart Panel, covering 133 genes associated with cardiomyopathies and hereditary arrhythmias. An additional file describes these next generation sequencing (NGS) based panels in more detail (see Additional file 1). Our variant filtering scheme is outlined in Fig. 1. We studied the presence of Thr335Ala in DSG2 in probands’ relatives by bi-directional Sanger sequencing.

Fig. 1
figure 1

Overview of the variant filtering scheme. Filters used are listed on the left, and the number of variants on each step are depicted on the right. QS = quality score; MAF = minor allele frequency; ExAC = Exome Aggregation Consortium; HGMD = Human Gene Mutation Database; 1000G = 1000 Genomes Project; BpG = Blueprint Genetics; ACMG = American College of Medical Genetics and Genomics

Full size image

Family screening

We found both probands to be homozygous for the missense variant c.1003A > G, p.(Thr335Ala), (rs191564916) in DSG2. In silico predictions for this variant are contradictory, as it is predicted to be damaging by PolyPhen [31] and Mutation Taster [32] but tolerated by SIFT [33]. There is one heterozygous carrier of the variant in the 1000 Genomes Project [34] and a total of 67 carriers in the Exome Aggregation Consortium database (ExAC, n = 60,275) [35]. The p.(Thr335Ala) in DSG2 is fairly common especially among the Finnish population as six of the carriers were from Finland (n(FIN) = 3304). However, there are no homozygote individuals in the ExAC control cohorts.

We present five homozygotes and nine heterozygotes older than 20 years from two families with this variant. The main clinical characteristics of the subjects are shown in Table 1. Pedigrees of families 1 and 2 are demonstrated in Fig. 2.

Table 1 Clinical characteristics of the two families with recessively inherited ARVC associated with Thr335Ala in DSG2
Full size table
Fig. 2
figure 2

Pedigrees of the two families affected with the c.1003A > G, p.(Thr335Ala) mutation in DSG2. Black-filled symbols represent individuals who fulfill the 2010 Task Force Criteria by Marcus et al. [9]. Age of the family members and their genotypes are listed below the symbols. Parents of the affected individuals in family 1 are obligatory carriers of the variant. In family 2, one of them is an obligatory heterozygous carrier, while the other may either be heterozygous or homozygous for the variant

Full size image

Family 1

Proband (II.3) is a 52-year-old male with clinical suspicion of ARVC. He had recurrent episodes of sustained ventricular tachycardia (VT) and a collapse, with the onset of symptoms at age 42. Electrophysiological study (EPS) revealed a large scar in the RV wall, and a hemodynamically unstable VT of right ventricular outflow tract (RVOT) origin was inducible. Cardiac MRI showed no abnormalities at the time of initial investigations. During the follow-up, RV enlargement was noted, and mild LV dilatation occurred without changes in systolic function or wall thickness. RV ejection fraction (EF) was diminished. Patient has an implantable cardioverter defibrillator (ICD), and is treated with beta-blockers.

Proband’s sister (II.5) is a 54-year-old female with an initial presentation of acute-onset VT with angina pain and elevated cardiac markers at age 37. Angiography reveled a coronary artery stenosis which was thought to be the culprit lesion, and percutaneous coronary intervention (PCI) was performed successfully. Thereafter, patient had recurrent events of sustained VT with no associated cardiac marker elevations and ICD was implanted. Patient later suffered an ST-elevation myocardial infarction with another PCI performed. Catheter ablation was done with no reduction in arrhythmic episodes. During the follow up, no change in LV diameter was noted, but the right ventricle underwent marked dilatation. Patient also suffered an event of ventricular fibrillation, and was resuscitated. This arrhythmia was thought to be caused by an old infarct scar. She is treated with beta-blockers and amiodarone.

Proband’s brother (II.4) is a 53-year-old male with no history of arrhythmias or other known cardiovascular disease. In the initial examination, cardiac MRI showed RV dilatation and microaneurysms in the inferior RV wall and near the RVOT, with an akinetic zone present inferiorly in the free wall. Septum was slightly hypokinetic and thickened, and in the apical septum there was signal change suggestive of myocardial fatty infiltration.

Screening of family 1 revealed altogether three homozygotes, six heterozygotes and one wild type individual (Fig. 2). All family members homozygote for the Thr335Ala fulfilled the criteria for definitive ARVC diagnosis, whereas none of the heterozygous carriers had findings relating to ARVC or any other cardiac disease. However, one of the index patient’s children (III.2) has had a collapse of unknown etiology. The wild type homozygous sibling (II.7) has had paroxysmal atrial fibrillation. The index patient’s mother, who was an obligatory heterozygous carrier of the mutation, had congestive heart failure stemming from myocardial infarction and chronic atrial fibrillation.

Family 2

Proband (II.1) is a 56-year-old male with ARVC. He had an event of acute chest pain at the age of 51 with elevated cardiac markers. Patient was initially treated as having a non-ST elevation myocardial infarction. Subsequent angiography showed no signs of major coronary artery disease. Thereafter, patient experienced episodes of sustained monomorphic VT. Cardiac MRI showed RV dilatation and free wall akinesia. LV was also dilated and thinning of LV myocardium and hypokinesia was present. VT was inducible in EPS, but catheter ablation was not performed. Diagnostic changes were not found in endomyocardial biopsy, although immunohistochemical stains showed borderline myocarditis by Dallas criteria. Patient has an ICD, and is treated with beta blockers.

Proband’s brother (II.5) is 59 years old and has a history of atrial fibrillation, hypertension, hypercholesterolemia and three-vessel coronary artery disease with bypass surgery performed. Initially, cardiac MRI showed RV free wall microaneurysms, along with an RVOT aneurysm and apical dyssynchronous contraction (Fig. 3). RV was also slightly dilated and systolic function was diminished. During the follow up in a subsequent MRI, dilatation of the RV progressed markedly and the RVOT aneurysm also enlarged (diameter of 75 mm in axial plane). LV size and function was in the normal range, although it also underwent dilatation during the follow-up. Ventricular fibrillation was inducible in EPS, and ICD has been implanted.

Fig. 3
figure 3

Cardiac magnetic resonance images of a 59-year-old male homozygote for the Thr335Ala in DSG2 (II.5, family 2). Short-axis (a), axial (b) and right ventricular outflow tract (RVOT) (c) views of the heart show aneurysmal dilatation of the RVOT (arrows). Maximum dimension of the aneurysm was 67 mm and right ventricular end-diastolic volume index (RVEDVI) was 109 ml/m2. During follow-up, further dilatation of the aneurysm and the RV was observed. Four-chamber view shows rare microaneurysms in the right ventricular free wall (arrowheads) (d). LV = left ventricle; RV = right ventricle; RVOT = right ventricular outflow tract; RA = right atrium; AO = aorta

Full size image

Screening of family 2 revealed in total two homozygotes and three heterozygotes (Fig. 2). Again, both of the homozygotes fulfilled the criteria for definitive ARVC diagnosis. Heterozygote carriers had no symptoms relating to ARVC or any other cardiac disease. Neither of the index patient’s parents’ medical records showed any signs suggestive of cardiomyopathy. One of them is an obligatory heterozygous carrier of the mutation, while the other may either be heterozygous or homozygous for the variant.

Discussion and conclusions

Based on the families described in this study and the family published by Rasmussen et al. (see below) [19], the p.(Thr335Ala) in DSG2 associates with cardiomyopathy only in a homozygous state, as heterozygotes are considered healthy/not affected. This is consistent with recessive inheritance pattern. These findings highlight the importance of careful interpretation of the desmosomal variants. It seems that truly informative genetic counseling may require extensive clinical and genetic work-up in families with ARVC, especially when a variant is too common in control cohorts to be fully penetrant disease causing as heterozygote.

Rasmussen et al. have identified the same DSG2 variant in a family in which two brothers had severe ARVC [19]. These brothers were homozygous for the variant whereas their healthy parents and two healthy siblings were carriers, suggestive of recessive inheritance. Detailed protein investigations demonstrated that the Thr335Ala variant protein was expressed and incorporated into desmosomes.

In the literature and open databases, the Thr335Ala in DSG2 has been reported in ≥15 patients with ARVC (ClinVar accession SCV000060918) [25, 26, 36,37,38,39,40,41,42,43,44]. Family segregation is rarely studied in these reports. At least six individuals with ARVC carried an additional pathogenic PKP2 variant [36, 37, 40,41,42]. We have previously detected the same DSG2 variant as heterozygous in two patients and as homozygous in one patient. The homozygous patient was in his twenties with DCM, and did not carry any other variants considered significant for the phenotype. One heterozygous carrier was a boy who suffered sudden cardiac death, and the other was a female with clinically suspected DCM but who also carried Arg634His in DSC2 (classified as a variant of unknown significance).

The Thr335Ala has been reported in DCM patients in two studies [20, 45]. In the first study the index patient had DCM, requiring heart transplantation at the age of 45 [20]. His brother was diagnosed with asymptomatic DCM at the age of 52. In addition to the Thr335Ala, both brothers carried p.(Glu1833Val) in DSP which is common in the ExAC reference population and likely has no effect on phenotype. Another brother (49y) and their mother’s sister (76y) were also carriers of these variants and they were both healthy. Two other healthy siblings (47y and 55y) carried only the Thr335Ala variant. Therefore, DCM was diagnosed in 33% of the Thr335Ala carriers in the family. Other genetic factors were recognized that probably had a role in disease penetrance. In the other study, a DCM patient had both the Thr335Ala in DSG2 and Ala897fs*4 in DSC2 [45]. However, the DSC2 variant is unlikely to contribute to the phenotype, as it is also common in the ExAC database and the frameshift leads to a stop codon in the normal position.

There have also been other homozygous DSG2 variants such as the p.(Val55Met) found in a DCM patient [46]. His parents were both carriers of the mutation but only the father had DCM. However, the father’s disease had later onset and a milder course. Immunostaining and electron microscopy of explanted LV wall myocardium from the homozygous proband revealed pale intercalated discs and significantly shorter desmosomes compared to wild-type myocardium. Unfortunately, no heterozygous carrier was studied by immunostaining. Severe disease in a patient with homozygous DSG2 mutation and a milder disease in a carrier may indicate a dosage effect of DSG2 mutations on cardiac function.

Variant interpretation is still challenging as earlier studies form major pitfalls by false classifications related to small reference populations, co-incidental segregations and evaluation of only a small subset of the potentially meaningful genes behind a patient’s phenotype. In diagnostic laboratories and clinics, variant classification practices are largely based on recommendations by the American College of Medical Genetics and Genomics (ACMG) [47]. Without available family segregation data, variants are classified as pathogenic only when the genotype has been identified in a certain number of patients (e.g. 5–20) with the appropriate phenotype, or when the variant has occurred de novo multiple times in association with sporadic disease. Segregation analyses of large families are considered the best approach to define causativity of a variant and assess its penetrance. Large-scale genetic research and variant sharing will eventually bring more consistency to the evaluation of families with inherited cardiac diseases.