Mutation analysis of the phospholamban gene in 315 South Africans with dilated, hypertrophic, peripartum and arrhythmogenic right ventricular cardiomyopathies

Cardiomyopathy is an important cause of heart failure in Sub-Saharan Africa, accounting for up to 30% of adult heart failure hospitalisations. This high prevalence poses a challenge in societies without access to resources and interventions essential for disease management. Over 80 genes have been implicated as a cause of cardiomyopathy. Mutations in the phospholamban (PLN) gene are associated with dilated cardiomyopathy (DCM) and severe heart failure. In Africa, the prevalence of PLN mutations in cardiomyopathy patients is unknown. Our aim was to screen 315 patients with arrhythmogenic right ventricular cardiomyopathy (n = 111), DCM (n = 95), hypertrophic cardiomyopathy (n = 40) and peripartum cardiomyopathy (n = 69) for disease-causing PLN mutations by high resolution melt analysis and DNA sequencing. We detected the previously reported PLN c.25C > T (p.R9C) mutation in a South African family with severe autosomal dominant DCM. Haplotype analysis revealed that this mutation occurred against a different haplotype background to that of the original North American family and was therefore unlikely to have been inherited from a common ancestor. No other mutations in PLN were detected (mutation prevalence = 0.2%). We conclude that PLN is a rare cause of cardiomyopathy in African patients. The PLN p.R9C mutation is not well-tolerated, emphasising the importance of this gene in cardiac function.

PLN in Africans with cardiomyopathy is unknown 11 . A PLN founder mutation, PLN p.R14del, was identified in large European cohorts, including 10-15% of Dutch patients with ARVC or DCM, and was associated with high mortality and poor prognosis 10,12 . As this mutation arose 575-825 years ago, the possibility exists that the PLN p.R14del founder mutation may be present in descendants of Dutch settlers and may cause cardiomyopathy in a subset of individuals of European descent in Southern Africa 13 . The aim of this study was to determine if PLN is a cause of cardiomyopathy in Africans, and explore the possibility of PLN founder mutations common between the African population and individuals with European ancestry.
Detailed investigations into the ancestry of individual DCM 320.1 pointed to an autosomal dominant inheritance pattern within this DCM family (Fig. 1C). The proband (DCM 320.1; II:2) had a heart transplant at the age of 35 years while her younger sister (DCM 320.5; II:3) was likewise affected with DCM and required a heart transplant at the age of 39 years. The proband's son (DCM 320.3; III:1) developed DCM at the age of 24 years and underwent a heart transplant a year later. The echocardiogram showed features of DCM with a left ventricular ejection fraction of 20%. Coronary angiography showed patent epicardial coronary arteries, and left ventriculography displayed a dilated left ventricle with poor systolic function. The proband's daughter (DCM 320.4; III:3) was asymptomatic but the echocardiogram showed borderline dilatation of the left ventricle (left ventricular end diastolic dimension of 5.3 cm) and left ventricular ejection fraction of 47%. Further enquiry also revealed that the proband's mother had died of a heart condition at the age of 36 years. However, we were not able to confirm whether this individual was affected with DCM.
Subsequent mutation screening of PLN for the available members of this family found the c.25C > T (p.R9C) mutation in the proband's affected son (DCM 320.3; Individual III:1) and daughter (DCM 320.4; Individual III:3) Bioinformatic analysis tools MutationTaster, PolyPhen-2 and Align GVGD predicted the PLN c.25C > T (p.R9C) mutation to be disease-causing. Further, transgenic mice bearing the PLN p.R9C mutation developed severe DCM and underwent premature death, confirming the pathogenicity of this mutation 6 . Bioinformatic RNA tools mfold and RNAfold predicted this variant would change mRNA secondary structure whereas the ESEfinder tool predicted that this variant would alter serine/arginine-rich splicing factor (SRSF1 and SRSF5) exonic splice enhancer recognition sites.
As the PLN c.25C > T mutation was previously reported in a family of European descent by Schmitt and colleagues 6 , we wanted to ascertain if these mutations arose independently or if they were inherited from a common ancestor, possibly as early as the 17 th century with the European colonisation of South Africa. Haplotypes were constructed using published microsatellite markers spanning 5.15 Mb across the PLN gene (D6S454, PLN − 200 K, PLN + 200 K and D6S412) 13 . In the DCM 320 family, the PLN c.25C > T mutation was found to be part of a disease-specific haplotype spanning this 5.15 Mb region ( Fig. 2A). The PLN c.25C > T mutation was found to be part of a different haplotype spanning 1.85 Mb in the MDO DCM family described by Schmitt et al. 6 (Fig. 2B).

Discussion
In the initial report of the PLN c.25C > T (p.R9C) mutation by Schmitt and colleagues 6 , the mutation was associated with an autosomal dominant inheritance pattern in DCM which was characterised by increased cardiac chamber dimensions, decreased contractile function at 20-30 years of age and progression to heart failure within 5-10 years after symptom onset 6 , a phenotype remarkably similar to that observed in the DCM 320 family. Similarly, in the recent report by Truszkowska and others, the PLN p.R9C mutation was detected in an individual with acute onset of DCM at the age of 21 years, leading to heart transplantation at 22 years of age 18 .
Functional analysis of the effect of the PLN p.R9C mutation on the protein and its role in the development of DCM revealed significant consequences of this mutation. A transgenic mouse model harbouring the p.R9C mutation was shown to develop DCM 6 . Further evidence has suggested that PLN p.R9C results in decreased responsiveness to β -adrenergic stimulation secondary to reduced phosphorylation by protein kinase A 6,19,20 . This results in altered calcium kinetics and aberrant contractility. The presence of severe DCM in this family with PLN p.R9C reinforces the importance of this mutation in the pathogenesis of DCM.
We have established that the PLN c.25C > T mutation forms part of a disease-specific haplotype spanning 5.15 Mb across PLN in the DCM 320 family that differed from the 1.85 Mb haplotype observed in the MDO DCM family described by Schmitt et al. 6 . It is therefore unlikely that this mutation was inherited from a common ancestor. By contrast, the PLN p.R14del mutation, has been identified as a founder mutation in a large Dutch cohort of patients with ARVC, as well patients from Germany and Spain 13,21 . A mouse model with cardiac-specific overexpression of this mutation presented with DCM associated with arrhythmias, cardiac fibrosis and premature death 8 .
Although a total of 315 South African cardiomyopathy patients (ARVC, DCM, HCM, and PPCM) were screened in this study, only one disease-causing mutation was found in PLN (frequency of 0.2%) illustrating that PLN is not a common cause of cardiomyopathy in South Africa. The failure to detect PLN mutations in patients with PPCM is consistent with the findings of others 22 (Table 1).
While the majority of these PLN mutations have been associated with severe cardiomyopathy, only PLN mutations associated with DCM meet the strict criteria to be called definitively pathogenic. These are nonsynonymous variants that alter a highly conserved amino acid residue across species, are absent from large numbers of healthy controls, demonstrate statistically significant co-segregation with affected members of a family, and are identified by an unbiased genetic approach. Important additional evidence comes from (1) the recapitulation of disease in animals engineered to express the variant, as exists for the PLN p.R9C mutation 6 or (2) the same mutation is identified in unrelated individuals with the disease phenotype, as we have demonstrated herein for the PLN p.R9C mutation.
As far as we are aware, the three PLN mutations detected in patients with HCM are disease-associated and cannot be classified as definitely pathogenic for several reasons. First, the authors who reported the PLN c.116T > G (p.L39X) mutation in HCM provide evidence indicating the association of this mutation with HCM in a single proband 31 . There is no apparent familial segregation of this mutation with disease. The authors also provide no functional evidence demonstrating the role of this mutation in disease pathogenesis. Other authors provide evidence that this mutation is associated with disease in a single proband with HCM. The proband's daughter carried the mutation but was unaffected with HCM, while the 3 year old granddaughter was affected with this disease 28 . The authors suggest that this may be due to incomplete penetrance associated with this mutation, but the fact that the affected granddaughter is so young calls this explanation into question. Also, no functional evidence is provided for the role of this mutation in DCM pathogenesis. Second, the PLN c.1-77A > G has been associated with HCM in a single proband 29 . No evidence for familial segregation or functional evidence suggesting a role of this mutation in HCM pathogenesis is provided. Finally, Medin et al. report the PLN c.1-42C > G variant as associated with disease in a family with HCM 32 . One of the proband's sons was unaffected with HCM even though he carried the mutation, which may be explained by incomplete penetrance. However, no functional evidence suggesting the role of this variant in DCM pathogenesis is provided. For these reason, we classify these three variants as HCM-associated but not definitely pathogenic (Table 1).
In conclusion, we have identified the previously characterised c.25C > T PLN mutation that segregates with severe DCM in a South African kindred following the screening of 315 patients with different types of cardiomyopathy. No PLN mutations were identified in subjects with ARVC, HCM or PPCM. PLN mutations are a rare cause of cardiomyopathy in South Africans which should be added to screening panel for cardiomyopathy in the country.

Methods
The methods were carried out in accordance with the approved guidelines of the University of Cape Town Human Research Ethics Committee, and they are presented in terms of samples, genetic screening, and bioinformatic analysis, and microsatellite analysis below.  www.nature.com/scientificreports/ Blood samples were collected for molecular genetic testing. The population controls were anonymous blood donors from the Western Province Blood Transfusion Service who provided blood samples for DNA isolation. The study was approved by the Human Research Ethics Committee of the University of Cape Town, and informed consent was obtained from all participants. All ARVC cases included in this study had previously been screened for mutations in the desmosomal genes (DSP, PKP2, DSC2, DSG2 and JUP) known to cause ARVC but were found not to harbour any pathogenic mutations. In our screen we included 315 ARVC, DCM, HCM and PPCM cases with unidentified pathogenic mutations.

Samples. South
Genetic screening. Mutation screening of PLN included 111 ARVC, 95 DCM, 40 HCM, and 69 PPCM patients and was performed by high resolution melt analysis (HRM) and Sanger sequencing using the following primers: Forward -5′ -CCAGGCTACCTAAAAGAAGAC-3′ ; Reverse -5′ -TTCCTGTCTGCATGGGATG-3′ . HRM is proven mutation screening method with a high sensitivity and specificity 34  PCR products were run on a 2% agarose gel for verification of reaction success and product size and specificity. These products were then analysed using capillary electrophoresis with the ABI PRISM ® 3130 × l Genetic Analyser (Applied Biosystems, Foster City, CA, USA) at the Division of Human Genetics, University of Cape Town. Results were analysed using GeneMapper ® v4.1 software (Applied Biosystems) and haplotypes were constructed using Cyrillic v2.0 (Cyrillic Software, United Kingdom).