What's new?

Most of the DNA mutations leading to hypertrophic cardiomyopathy (HCM) are inherited with an autosomal dominant pattern and are unique or seen in a limited number of families. However, in some ethnic groups, founder mutations have been described. We report the analysis of 404 genes in a cohort of 29 Polish patients affected by HCM. Our results showed that an ultra‑rare alteration in the MYBPC3 gene, absent in databases from large‑scale sequencing projects, was a mutation hotspot in the present Polish cohort and could represent a novel founder pathogenic variant. Moreover, we found that 1 patient was homozygous for a novel truncating mutation in the CSRP3 gene, thus increasing the number of known human CSRP3 knockout cases to 3. This finding also suggests that the autosomal recessive inheritance pattern could be more frequent in HCM than reported thus far.

Introduction

Hypertrophic cardiomyopathy (HCM) is a cardiac disease characterized by left ventricular hypertrophy (thickness ≥15 mm) unexplained by secondary causes, and a nondilated left ventricle with preserved or increased ejection fraction. It is an important cause of sudden cardiac death, and its prevalence has been estimated at approximately 1/500 individuals in the adult population.¹ Hypertrophic cardiomyopathy is considered a disease of the cardiac sarcomere mainly caused by pathogenic variants in over 10 loci, with MYBPC3, MYH7 and TNNT2 accounting for approximately 50% of the HCM families and encoding cardiac myosin‑binding protein C, β-myosin heavy chain, and cardiac troponin T, respectively.² Other inheritable causes of the disease include pathogenic variants in genes encoding proteins involved in calcium handling and for proteins of the Z‑disk.³ Most of the pathogenic variants found in HCM genes are missense and unique or private within families.² These alterations are believed to have a dominant negative effect, acting as a “poison polypeptide” on sarcomere function.⁴ The only exception is MYBPC3, in which about two‑thirds of pathogenic variants are truncating and haploinsufficiency is postulated as a pathogenetic mechanism of the disease.⁵ To date, at least 12 different truncating MYBPC3 pathogenic mutations have been reported as founder variants in populations from Finland, France, Japan, Iceland, Italy, Spain, and the Netherlands.^5-8

In most adolescents and adults, HCM is inherited as an autosomal dominant trait with a clinical outcome characterized by incomplete penetrance and variable expression. The disease phenotype can indeed be modulated by environmental factors, by the genetic context (including polymorphisms of the renin–angiotensin system [RAS]),⁹ and by the occurrence of the causal pathogenic variant or variants in homozygosity, compound heterozygosity, or double heterozygosity.¹⁰ Janin et al¹¹ have recently reported 2 unrelated kindreds with homozygous truncating variants in the cysteine and glycine‑rich protein 3 (CSRP3) gene encoding a member of the CSRP family of muscle LIM protein (MLP).

It is likely that HCM individuals are underdiagnosed,¹² especially in countries where the availability of genetic testing based on next‑generation sequencing (NGS) is limited. Considering the prevalence of HCM, the estimated number of HCM patients in Poland is approximately 78 000. Until now, only single reports describing genetic causes of HCM in Polish patients have been reported.^13-15 Therefore, the aim of the current study was to assess the genetic background of HCM in a cohort of patients from the south‑eastern part of Poland, analyzed by an NGS panel including 404 genes known to harbor alterations affecting cardiovascular system function. Thanks to this approach, we could identify a MYBPC3 truncating alteration that could represent a novel Polish founder MYBPC3 pathogenic variant. Also, we detected a new patient with the homozygous CSRP3 variant; thus, the number of known HCM cases with null homozygous alterations in this gene has increased to 3.

Patients and methods

Statistical assessment of genotype–phenotype correlations

All statistical analyses were performed using the SPSS software package version 20.0 (IBM SPSS Statistics, Milan, Italy). For each patient, the clinical and molecular data were tabulated (Figure 1). Phenotype data were presented as continuous variables obtained from clinical data and instrumental measurements, and they were summarized using means and standard deviations. The mean values of age at diagnosis, maximal wall thickness, and left ventricular mass were compared using the t test for independent data between the subgroups of patients with different genotypes. Also, associations between sex, the genotype of 2 polymorphisms in the genes of the RAS, and clinical features were tested. The penetrance of HCM, according to age at diagnosis, was analyzed by the Kaplan–Meier method, and differences between cumulative hazard were evaluated with the log‑rank test. Due to the relatively low number of patients, no adjustments were planned for multiple testing. Therefore, the analysis is exploratory and the results should be considered as hypothesis generating.

Results and discussion

In total, 12 of the 29 patients (41%) were found to carry at least 1 pathogenetic, likely pathogenetic, or VUS‑favoring pathogenic alteration of the “core HCM genes,” as classified by the 2015 ACMG/AMP guidelines and ClinVar database (Table 1 and Figure 1). Most of the changes were identified in the MYBPC3, MYH7 and TNNT2 genes. The genotype of the rs5186 and rs699 polymorphisms, mapping respectively in the angiotensin receptor type 1 (AGTR1) and angiotensinogen (AGT) genes, belonging to the RAS, was also recorded (Figure 1). Indeed, some studies have shown that rs5186 and rs699 may influence the clinical phenotype of HCM,²² since the RAS regulates cardiac function, blood pressure, and electrolyte homeostasis.²³ However, in the present cohort, no significant correlation with the disease expression was found.

The identified mutational spectrum included 10 distinct substitutions: 6 missense, 3 nonsense, and 1 intronic change. Most of the missense variants were sarcomeric, and, among them, 3 mapped in MYH7, 2 in MYBPC3, and 1 in TNNT2. On the other hand, 2 of the nonsense alterations were found in MYBPC3 and 1 in CSRP3. The GnomAD allele frequency of all the sarcomeric alterations was smaller than 0.01%, that is, the MAF threshold suggested for HCM pathogenic variants.²⁴

A single variant was found in more than 1 patient (Supplementary material, Figure S1A). The MYBPC3 c.2541C>G variant, resulting in the premature insertion of a stop codon in exon 24 (p.Tyr847Ter), was indeed identified in 4 unrelated cases, despite the fact that it is absent in the GnomAD. Carriers of the MYBPC3 c.2541C>G showed a lower age at HCM diagnosis compared with carriers of other DNA variants, with a 50% probability of HCM diagnosis at 38 years for MYBPC3 c.2541C>G carriers, in comparison with the age of 49 years for the carriers of other DNA variants (P = 0.04; Figure 2). The MYBPC3 c.2541C>G mutation was the only identified variant in cases 10, 74, and 3566, while it was associated with another sarcomeric alteration (ie, the MYH7 c.3116A>G,p.Glu1039Gly) in patient 22. The allele frequency of the MYH7 c.3116A>G was very close to the 0.01% threshold, and we infer that this variant could be a modifier allele or a low penetrance variant contributing to the severity of HCM expression in patient 22. Among the 4 carriers of the MYBPC3 c.2541C>G, patient 22 indeed manifested the earliest age at diagnosis, the need for septal myectomy at the age of 41 years, and the presence of a long QTc interval (Figure 1). The segregation of MYBPC3 c.2541C>G and of MYH7 c.3116A>G was tested in 4 relatives of case 22: a 50‑year‑old sister affected by HCM and 3 children aged 9, 12, and 16 years, who are currently asymptomatic. All of them were found to carry MYBPC3 c.2541C>G in heterozygosity, while they had the wild‑type MYH7 c.3116A variant (Supplementary material, Figure S1A).

The MYBPC3 c.2541C>G variant was also detected in heterozygosity in a 19‑year‑old daughter of patient 74 (Supplementary material, Figure S1A). On clinical evaluation, she did not display any chronic cardiac or noncardiac disease. Her echocardiogram showed normal left ventricular (left ventricular end‑diastolic diameter, 43 mm) and atrial size (left atrial appendage, 17 cm²; right atrial appendage, 13.5 cm²), as well as normal thickness of intraventricular septum (end‑diastolic diameter, 10 mm). Also, left ventricular systolic function was normal (70%), and she did not show any valvular abnormalities, signs of systolic anterior motion, or left ventricular outflow obstruction. Therefore, at present, the children of case 22 and the daughter of case 74 are asymptomatic carriers of c.2541C>G and possibly would show later onset of the disease due to an age‑related penetrance.

To investigate whether carriers of the MYBPC3 c.2541C>G variant could share a common ancestor, we performed a markers analysis. A haplotype, spanning about 5 Mb, was found to segregate with the c.2541G variant allele in the family of patient 22. The alleles of the c.2541G‑haplotype were also present in all the MYBPC3 c.2541C>G carriers (Table 2), even if we could not reconstruct their phase. This finding supports the hypothesis that the MYBPC3 c.2541C>G variant could have a founder role in the Polish population. In support of this hypothesis, the MYBPC3 c.2541C>G variant has been recently found in another Polish patient with HCM, not related to any of the cases reported herein.²⁵

Case 58 was found to carry the truncating CSRP3 c.364C>T (p.Arg122Ter) variant in homozygosity (Supplementary material, Figure S1B), as confirmed by SYBR Green–based quantitative PCR on ABI7900 HT Fast Real Time PCR System (ThermoFisher) (Supplementary material, Figure S1C). Patient 58 was a 56‑year‑old woman diagnosed with HCM at the age of 54 due to early fatigue and chest pain. She mainly complained of nonspecific and episodic chest pain (discomfort) on exertion. Echocardiography showed asymmetric interventricular septal hypertrophy (18 mm), posterior wall of 12 mm, ejection fraction of 70%, normal right ventricle, and maximal (provoked) left ventricular outflow tract gradient of 18 mm Hg without the systolic anterior motion of the mitral valve. Also, severe diastolic dysfunction with no signs of pulmonary hypertension was observed. The 24‑hour electrocardiogram highlighted 5 runs of nonsustained ventricular tachycardia (the longest one of 11 beats). Family history was negative for HCM, but the CSRP3 c.364C>T variant was found to segregate in heterozygosity in a 29‑year‑old son (Supplementary material, Figure S1B), who currently does not suffer from any chronic cardiac or noncardiac diseases. He has no symptoms and is physically active. On echocardiography, he has normal left ventricular (left ventricular end‑diastolic diameter, 46 mm) and atrial sizes (left atrial appendage, 17.5 cm²; right atrial appendage, 17 cm²). However, his intraventricular septum was mildly thickened (end‑diastolic diameter, 13 mm), with normal systolic function (ejection fraction, 67%), no valvular abnormalities, and no sign of systolic anterior motion and left ventricular outflow tract obstruction.

The CSRP3 c.364C>T (p.Arg122Ter) mutation is not a common variant since it is absent in GnomAD‑Genomes and GnomAD‑Exomes‑European databases. In total, only 3 CSRP3 c.364C>T alleles are listed in GnomAD‑Exomes (ie, 2 in African and 1 in South Asian populations), but never in a homozygous state. It has not been described in the literature before, but it is listed as likely pathogenic in the ClinVar database because of the following evidence: 1) it is a rare variant; 2) it is predicted to cause the loss of protein function either by protein truncation or nonsense‑mediated mRNA decay; 3) CSRP3-null mice develop cardiomyopathy and heart failure due to a disrupted cardiomyocyte architecture²⁶;and 4) myocardial biopsies of a HCM patient with a heterozygous CSRP3 missense variant showed myocyte disarray and a reduced level of MLP, suggesting that cardiomyopathy may stem from CSRP3 haploinsufficiency.²⁷ To date, 24 carriers of a CSRP3 alteration were reported in the literature or included in the ClinVar database: 18 had HCM and 6 were affected by dilated cardiomyopathy. Among the HCM patients, 9 carried a heterozygous CSRP3 variant (including 7 amino acid substitutions and 2 truncating variants), 2 cases were heterozygous but also carried a second variant in another gene, while 2 patients harbored a truncating alteration in homozygosity (Table 3).¹¹ Therefore, patient 58 described herein is the third CSRP3 human knockout case reported so far. Our finding strengthens the assumption that at least several CSRP3 variants lead to HCM with an autosomal recessive inheritance¹¹ rather than with an autosomal dominant transmission as recorded in the Online Mendelian Inheritance in Man (OMIM) database (OMIM: *600824). The mouse model knock‑in for CSRP3 pathogenic variants also corroborates the hypothesis of autosomal recessive inheritance of the cardiac disease and recalls the findings displayed by patient 58 and his son.²⁸ Indeed, the heterozygous mice for CSRP3 pathogenic variants do not display an overt cardiac phenotype (except for an increase in anterior wall thickness), while the homozygous mutated mouse shows a clear cardiomyopathy phenotype.²⁸

The “non­–core HCM genes” with mutations in the present HCM cohort were GYG1, GUSB, PMM2 and SCO2. Cases 64, 161, 211, and 218 were indeed identified as heterozygous carriers of autosomal recessive alterations associated with PMM2, SCO2, GUSB, and GYG1 deficiency, respectively (Table 4). The GYG1, GUSB, PMM2, and SCO2 variants were prioritized by the TGex software since they were related to the cardiomyopathy phenotype: GYG1 c.304G>C, found in patient 218, has previously been described in homozygosity in cases affected by severe cardiomyopathy without skeletal muscle weakness²⁹; the GUSB and PMM2 genes are responsible for genetic disorders associated with HCM³⁰; the SCO2 c.418G>A variant has been previously reported by the name of “c.1541G>A” in homozygosity or in compound heterozygosity in individuals with SCO2-related clinical features including HCM.^31,32 Therefore, these results can be considered “secondary findings” for which we cannot exclude a possible hypomorphic impact over the patients’ phenotype. Also, the SCO2 p.Glu140Lys variant was previously associated with autosomal dominant high‑grade myopia.³³ In the present patient, presenting low myopia of –2.00 diopters in both eyes, the SCO2 c.418G>A variant exhibited reduced penetrance.

In conclusion, this report expands the mutational spectrum and the inheritance pattern of HCM. The ultra‑rare MYBPC3 c.2541C>G (p.Tyr847Ter) alteration, found in 9 cases (of which 4 were index patients), while absent in databases from large‑scale sequencing projects, acts as a variant hotspot in the present Polish cohort and correlates with a younger age at HCM diagnosis. These findings, if confirmed in a wider population of the same ethnic origin, will increase the number of truncating founder MYBPC3 alterations.

The identification of the novel homozygous null CSRP3 variant leading to HCM suggests that the autosomal recessive inheritance pattern could be more frequent in HCM than reported so far. In heterozygosity, the null CSRP3 allele seems to correlate only with a mild thickening of the intraventricular septum.