Introduction

Arrhythmogenic cardiomyopathy (ACM) is a rare cardiomyopathy with 1 per 5000 prevalence and is defined by ventricular enlargement, myocardial fibro-fatty replacement, conduction defects, inflammation, and apoptosis, leading to ventricular arrhythmias [1]. Arrhythmogenic right ventricular dysplasia (ARVD) and subsequently ACM have been an evolving diagnosis throughout the recent decade with cardiac magnetic resonance imaging (CMRI) integration into the phenotype identification of cardiomyopathies [2]. A major cause of heart failure, ACM can occur due to genetic factors. Genetic ACM is more severe, has an earlier onset, and is heterogeneous; it is caused by mutations in genes encoding cardiac desmosomes, namely cadherins, plakins, and armadillo proteins [3,4,5]. More than 60% of ACM cases have desmosomal gene mutations; however, non-desmosomal gene mutations have also been detected [3, 6, 7]. A genetic variant interpreted as a pathogenic variant fulfills a major criterion set by the Task Force Criteria (TFC) for ACM recognition [8]. ACM can be typically present in an autosomal dominant pattern with incomplete penetrance and variable expression, but autosomal recessive inheritance is also known in a small number of cases [9, 10]. Many ACM patients have no identifiable mutation, with only one-fifth of all ACM cases familial. These patients may have pathogenic variants or causative genes not yet identified. Recently, a meta-analysis indicated that family history among ACM patients with non-desmosomal gene mutations had a lower frequency [11]. This clear gap in our understanding of ACM has led to continuous surveys to identify other causative genes of ACM.

PPP1R13L is a novel causative gene that encodes the inhibitor of apoptosis-stimulating p53 protein (iASPP), which plays regulatory and inflammatory roles in desmosomes. PPP1R13L is located on chromosome 19q13.32 [12, 13]. In 2009, Simpson and colleagues identified a frameshift variant (p.Ser322GlnfsTer4) in 13 animals with cardiomyopathy and woolly haircoat syndrome. They indicated that the bovine PPP1R13L sequence had 91.9% sequence similarity and 88.7% identity with the human PPP1R13L sequence [14]. Falik-Zaccai and coworkers identified a PPP1R13L missense variant (p.Tyr747Ter) in 3 patients with DCM associated with mild skin, teeth, and hair abnormalities (cardio-cutaneous syndrome) [15]. In 2019, in a study by Poloni and colleagues, a novel missense variant (p.Ala620Pro) of PPP1R13L was recognized in an ACM patient [13]. Recently, Robinson and coworkers reported three frameshifts, one nonsense, and one stop-loss in five pediatric dilated cardiomyopathy (DCM)-affected families. These studies provide evidence that PPP1R13L is a causative gene in cardiomyopathies (Table 1). Genetic investigation of patients with cardiomyopathy is significant for possible personalized cure and risk prediction of family members [16]. Recently, cardiomyopathy genetic research has been improved by the emergence of next-generation sequencing (NGS) approaches. Here, we identified a homozygous stop-gain pathogenic variant of PPP1R13L in a DCM/ACM Iranian patient by whole-exome sequencing [17].

Table 1 The reported variants in PPP1R13L and their related phenotypes
Full size table

Methods

Study subject and ethics statement

In this study, a consanguineous Iranian family with three offspring, namely a two-year-old ACM son, a thirteen-year-old healthy son, and a ten-year-old healthy daughter, was referred to the Cardiogenetics Research Center, Rajaie Cardiovascular Medical and Research Center, Tehran, Iran (Fig. 1a). The affected offspring (III-3, Fig. 1a) was born via cesarean section. Based on clinical, laboratory and morphological findings, ACM was diagnosed when he was six months old, three to four weeks after the infant was generally unwell. There was no history of fever in his past medical history. Blood, urine, and metabolic parameters were normal. His growth was acceptable, and he had no other abnormalities up to two years old. Despite medical treatment, however, his ACM progressed to end-stage heart failure, and he underwent heart transplantation at the age of eighteen months. There was not skin or hair abnormalities in the proband. No family history was observed in the pedigree. All the medical histories and the clinical information were collected in Rajaie Cardiovascular Medical and Research Center, Tehran, Iran, in accordance with the Declaration of Helsinki, and the study protocol was approved by the institutional ethics committee (IR.RHC.REC.1399.044). Written informed consent was obtained from all participants.

Fig. 1
figure 1

The image illustrates Family pedigree and genetic analysis of a PPP1R13L c.580C > T variant. a The pedigree of the family is shown herein. b The genotypes of a novel stop-gain pathogenic variant c.580C > T (p.Gln194Ter) were detected in the affected proband as homozygous and in his siblings and parents as heterozygous. The other available individuals had a normal sequence or carried the variant as heterozygous

Full size image

Cardiac magnetic resonance imaging and pathology

Cardiac magnetic resonance was conducted with a 1.5T MAGNETOM Avanto (Siemens Healthcare, Erlangen, Germany). The standard protocol was performed with the acquisition of steady-state free precession cine imaging in long- (four-, two-, and three-chamber) and short- (breath-hold) axis cine pictures plus the right ventricular inflow-outflow view. The short Tau inversion-recovery (STIR) sequence in long- (four-, two-, and three-chamber) and short- (breath-hold) axis cine pictures was taken. Early and late gadolinium enhancements of a short-axis stack of the three long-axis views and the right ventricular inflow-outflow images were collected by using magnitude and phase-sensitive inversion recovery reconstructions after the administration of 0.15 mmol/kg of gadoterate meglumine (gadolinium-DOTA, Dotarem, Guerbet SA, Paris, France).

In addition, the myocardium obtained from the transplanted heart was formalin-fixed and paraffin-embedded for analysis. Full-thickness sections from the right and left free ventricular walls and septa were prepared. The slides were stained with both hematoxylin and eosin (H&E) and Masson trichrome stains.

Whole-exome sequencing and segregation analysis

Genomic DNAs of the family members were extracted from their peripheral blood by using the DNSol Midi kit (Roche: Product No. 50072012). The WES of the proband (III-3, Fig. 1a) was performed by Macrogen (Seoul, South Korea) on an Illumina HiSeq 4000. Raw data (FASTQ format) were analyzed in the Cardiogenetics Research Center, Rajaie Cardiovascular Medical and Research Center, Tehran, Iran. In general, according to our in‐house setup NGS analysis pipeline (i.e., mapping, variant calling, and annotation), point variants, duplications, and deletions/insertions (indels) were detected. Public international databases such as the 1000 Genomes Project (http://www.1000genomes.org/), the Genome Aggregation Database (gnomAD) (https://gnomad.broadinstitute.org/), the Greater Middle East (GME) (http://igm.ucsd.edu/gme/), and Iranome (http://www.iranome.ir) were used to analyze the minor allele frequency (MAF) of the variants. Reported mutations in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar) and Human Gene Mutation Database (HGMD) (http://www.hgmd.cf.ac.uk/ac/index.php) as likely pathogenic/pathogenic were considered the priority. Bioinformatics analysis was carried out on the remaining variants through the application of online tools such as Combined Annotation-Dependent Depletion (CADD) (https://cadd.gs.washington.edu/home), SIFT (https://sift.bii.a-star.edu.sg/), PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), PROVEAN (http://provean.jcvi.org/index.php), and MutationTaster (http://www.mutationtaster.org/) to predict the variant effect of the function/structure of the protein.

Further, ENTPRISE-X (http://cssb2.biology.gatech.edu/entprise-x/) was employed to predict the consensus of nonsense variants. The variants predicted by most of the tools as damaging were selected for segregation analysis. The standards of the American College of Medical Genetics and Genomics (ACMG) were applied for the interpretation of the variants [18].

One primer pair was designed by using Primer3 (v.04.0) (http://bioinfo.ut.ee/primer3-0.4.0/). Forward primer 5′-ACACCAACCCTTCCACTAATG-3′ and reverse primer 5′-GTCAGACTCGTTCCAGGCT -3′ were applied to amplify the variant sequence. Polymerase chain reaction (PCR) was performed on a SimpliAmp™ Thermal Cycler (Thermo Fisher Scientific) with 300 ng of genomic DNA, 1.5 mmol/L of MgCl2, 200 mmol/L of dNTP, 10 pmol/L of primers, and 1 U of Taq DNA polymerase (Amplicon, UK) at 95 °C for 5 min of incubation and amplification (30 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C; 35 cycles). The PCR products were sequenced on an ABI Sequencer 3500XL PE (Applied Biosystems), and the sequences were analyzed by CodonCode Aligner (v.7.1.2) (https://www.codoncode.com/aligner/).

Results

Clinical analysis

Cardiac magnetic resonance showed a severely enlarged right ventricle with severely reduced right ventricular systolic function, concomitant with a mildly enlarged left ventricle with severely reduced left ventricular systolic function (Additional file 1). The right ventricular end-diastolic volume indexed to the body surface area was 265 mL/m2, and the right ventricular ejection fraction was 13%. The left ventricular end-diastolic volume indexed to the body surface area was 96 mL/m2, and the left ventricular ejection fraction was 19%. The right ventricular inflow-out flow cine images (Additional file 2) visualized segmental aneurysmal formations with dyskinesia in the right ventricular outflow tract and the sub-tricuspid region. The STIR sequences showed no inflammation or edema. Late gadolinium enhancement images demonstrated nearly-circumferential subepicardial fibrosis in the basal-to-mid inferior, inferolateral, and mid-anterolateral walls of the left ventricle (Fig. 2a, b). Given the cardiac magnetic resonance features of regional wall motion abnormalities in the right ventricle and biventricular systolic dysfunction and fibrosis pattern in the left ventricle, the phenotype was compatible with ACM with biventricular involvement. In addition, the 12-lead electrocardiograph (ECG) showed RS pattern in the inferior leads and QR pattern in V1 with poor R progression, and persistent prolonged S wave in the precordial leads. (Additional file 3). The clinical surveys of other available members of the pedigree were normal.

Fig. 2
figure 2

The image depicts dilated cardiomyopathy as confirmed by pathology and cardiac magnetic resonance. a and b The late enhancement sequences at the basal and mid-level short-axis views, respectively, show nearly-circumferential subepicardial enhancements in the basal-to-mid inferior, inferolateral, and mid-anterolateral walls of the left ventricle. c Microscopic examination shows cardiac myocytes with nuclear enlargement, anisonucleosis, and vacuolar degeneration (H&E X400). d Trichrome staining of the right ventricular free wall shows replacement fibrosis

Full size image

Macroscopic examination of the transplanted heart showed the enlargement of both ventricular cavities. The maximal thickness of the right and left free ventricular walls was 1 cm and 1.5 cm, respectively, and the interventricular septum had a maximal thickness of 1.4 cm. Histological features revealed nuclear enlargement, anisonucleosis, and vacuolar degeneration. Trichrome staining of the right and left ventricular free walls showed foci of interstitial and replacement fibrosis (Fig. 2c, d).

Molecular analysis

WES detected a novel homozygous stop-gain pathogenic variant of PPP1R13L (NM_006663.4:exon4:c.580C > T: p.Gln194Ter) in the proband. This variant is predicted to make a truncated PPP1R13L protein and was not reported in any databases. The c.C580T:p.Q194X variant with a CADD Phred of 38 was predicted as disease-causing, deleterious, and pathogenic by MutationTaster, ENTPRISE-X, and ACMG, respectively.

PPP1R13L genotype among the family members

In the confirmation and segregation analysis of the family members, the only affected individual, proband (III-3, Fig. 1b), had the c.C580T:p.Q194X variant as homozygous, and his siblings and parents had the variant as heterozygous (II-6, II-7, III-1, and III-2, Fig. 1b). The other available unaffected members of the pedigree had a normal sequence (II-4 and II-9, Fig. 1b) or carried the variant as heterozygous (II-3, Fig. 1b). Clinical examinations of all available members, heterozygous individuals, were also completely healthy.

Discussion

ACM diagnosis has been evolved with the use of advanced cardiac imaging and electrophysiological studies [2, 19]. The phenotype of our case was determined by cardiac MRI features which are as follows: Dilated RV index versus left ventricle, presence of regional wall motion abnormalities in the right ventricle plus reduced left ventricular function and fibrosis pattern which all are more in favor of ARVD with concomitant LV involvement; ACM per definition versus DCM.

Our study highlighted PPP1R13L as an important candidate in ACM. Given that the first symptom of ACM is mostly sudden death, early recognition of at-risk individuals is crucial. WES is a powerful approach to detect novel pathogenic variants of heart diseases. With the aid of this tool, we identified a novel PPP1R13L stop-gain variant (c.580C > T: p.Gln194Ter) in an autosomal recessive ACM pedigree. In recent years, PPP1R13L variants have been identified in patients with cardio-cutaneous syndrome [15, 20], cattle woolly haircoat syndrome [14], DCM [20], and ACM [13] (Table 1). The location of all reported variants was shown in Fig. 3. PPP1R13L encodes the iASPP protein in the intercalated discs of cardiomyocytes [21], and iASPP interacts with desmin and desmoplakin. Indeed, iASPP represses NF-κB transcriptional activity, which is followed by heart failure development [22]. A deficiency of iASPP has been recently related to autosomal recessive cardiomyopathy in Wa3 mice [23] and Poll Hereford calves [14]. Notari and colleagues identified iASPP as a gatekeeper of ACM and a desmosome key regulator. They formed a ternary assembled with desmin and desmoplakin and showed that deficient iASPP in mice led to ventricular tachycardia, arrhythmia, ACM, and sudden death [24]. In our study, the c.580C > T variant, in both PPP1R13L alleles, was predicted to cause truncated iASPP proteins. Djouadi and coworkers indicated the PPARγ pathway activation in iASPP-truncated hearts and its implication in ACM adipogenesis. The exact mechanism of the iASPP effect on the heart is unknown; nonetheless, it seems that deficient iASPP cannot inhibit p53 and its enhanced activity leads to the activation of PPARγ and elevated LPIN 1 levels [25]. The LPIN 1 protein is assembled with PPARγ for gene expression regulation in the fatty acid oxidation process [26]. Deficient iASPP protein does not interact with desmin and desmoplakin and, thus, may render individuals susceptible to ACM.

Fig. 3
figure 3

The image demonstrates the distributions of the reported variants of PPP1R13L. The regions and domains were obtained from UniProt (https://www.uniprot.org/)

Full size image

In light of the results of the present study, we suggest that PPP1R13L could be as a potential ACM candidate gene. Further functional studies are, however, needed to confirm the importance of this gene in ACM and clarify the exact mechanisms with a view to formulating new therapeutic approaches.

Conclusions

In the era of NGS technologies, especially WES, there are opportunities to resurvey diseases with no known genetic reasons. Such approaches in recent years have revealed ACM-related variants in individuals with neither family history nor ACM symptoms. Our study revealed homozygous variant in the PPP1R13L gene as a cause of ACM. Thus, the best way to achieve the most accurate diagnosis is to combine clinical and genetic tests and interpret the results collectively.

Accession Number

The accession number of the variant in ClinVar is as follows:

NM_006663.4 (PPP1R13L): c.580C > T (p.Gln194Ter): VCV001188825.1