From diagnostic testing to precision medicine: The evolving role of genomics in cardiac channelopathies and cardiomyopathies in children

Abstract

Pediatric sudden cardiac death (SCD) is the sudden unexpected death of a child or adolescent due to a presumed cardiac etiology. Heritable causes of pediatric SCD are predominantly cardiomyopathies and cardiac ion channelopathies. This review illustrates recent advances in determining the genetic cause of established and emerging channelopathies and cardiomyopathies, and how broader genomic sequencing is uncovering complex interactions between genetic architecture and disease manifestation. We discuss innovative models and experimental platforms for resolving the variant of uncertain significance as both the variants and genes associated with disease continue to evolve. Finally, we highlight the growing problem of incidentally identified variants in cardiovascular disease-causing genes and review innovative methods to determining whether these variants may ultimately result in penetrant disease. Overall, we seek to illustrate both the promise, and inherent challenges, in bridging the traditional role for genetics in diagnosing cardiomyopathies and channelopathies to one of true risk-predictive precision medicine.

Introduction

Disease-associated variants in genes that encode key cardiac proteins can cause molecular defects that affect cardiac function and predispose to sudden cardiac death (SCD). SCD is the abrupt and unexpected death of an individual due to a presumed cardiac cause, affecting up to 250,000 individuals in the United States each year but is particularly devasting to families and communities when it happens to children and adolescents[1]. Pediatric SCD, defined as SCD of a person <18 years of age, affects about ~2 individuals/100,000 life-years and is most prevalent in adolescents and young adults from 12-21 years[2, 3]. Diseases which predispose to pediatric SCD are typically either cardiac channelopathies, molecular defects in cardiac ion channels which lead to arrythmia predisposition, or cardiomyopathies, primary myocardial disease resulting from molecular defects in proteins such as the cardiac sarcomere, both of which are heritable. Traditionally, this heritability has translated to diagnostic genetic testing in pediatric patients with suspicion for channelopathies, cardiomyopathies, or aborted SCD. Recently, genomic sequencing has uncovered a complex genetic architecture for these diseases which holds tremendous promise for predicting phenotypic expression of disease and prognosis, if the variants that cause disease can be definitively identified.

The genetic basis of established and emerging cardiac channelopathies

Cardiac channelopathies are pathologic alterations of cardiac-specific ion channels, or their interacting/regulatory proteins, that lead to channel dysfunction and result in increased arrhythmia susceptibility (Figure 1). Channelopathies encompass several arrhythmia syndromes including long QT syndrome (LQTS), Brugada syndrome (BrS), catecholaminergic polymorphic ventricular tachycardia (CPVT), and short QT syndrome (SQTS). Patients with LQTS have a prolonged corrected QT (QTc) interval on electrocardiogram (ECG) and can present with syncope and can experience SCD as the initial presenting symptom[4]. It is estimated that 65-75% patients with LQTS host variants that affect the K⁺ channels KCNQ1-encoded Kv7.1, KCNH2-encoded Kv11.1, and the SCN5A-encoded Nav1.5 sodium channel [5-8]. CPVT is a lethal channelopathy with a mortality of 30-50% by age 40 when untreated[9]. It is characterized by adrenergically mediated polymorphic ventricular tachycardia that can result in SCD[10]. Variants in genes that encode the sarcoplasmic calcium release channel (RYR2-encoded RyR2) or, rarely, calcium-handling proteins (CASQ2, CALM1, CALM2, and CALM3) are associated with CPVT[11]. Pathologic CPVT-associated variants in RYR2 result in increased Ca²⁺ leak of stored calcium from the sarcoplasmic reticulum which triggers arrhythmia. Patients with BrS are often asymptomatic but can present with right bundle branch block and ST elevation in precordial leads on ECG and these findings can be unmasked with fever or Nav1.5 pharmacologic blockade[12, 13]. Loss-of-function SCN5A variants are the most common cause of BrS, while several other genes have been associated with rare cases of BrS[14-17]. SQTS patients have an abnormally fast ventricular repolarization time which can predispose to ventricular arrhythmias and SCD. Genetic variants that cause gain-of-function in K⁺ channel genes such as KCNH2 and KCNQ1, among other genetic mechanisms, contribute to the short QT-interval found on the ECG[18, 19].

Figure 1:

Association of cardiac channelopathies with genetic variants in genes that encoded cardiac ion channels. Gain-of-function indicates disease-association with variants that cause a gain in the biophysical properties of a given channel, such as peak or late ion current, while loss-of-function indicates disease association with variants that cause a biophysical loss.

A relatively new cardiac channelopathy is multifocal ectopic Purkinje-related premature contraction (MEPPC) which is a rare channelopathy that manifests as frequent premature ventricular complexes which originate from the fascicular-Purkinje system and can trigger ventricular arrhythmias and SCD[20]. While the genetic cause of MEPPC is still emerging, gain-of-function variants in SCN5A have been linked to the disease and there is evidence of genetic and phenotype overlap between MEPPC and SCN5A-mediated LQTS[21]. In addition, new channelopathy syndrome, RyR2 Ca²⁺-release deficiency syndrome (CRDS, OMIM #115000), has been recently identified. CRDS results from loss-of-function variants in RYR2[22], and while the precise biophysical mechanism is still emerging, CRDS-associated variants reduce expression of RyR2, reduce sensitivity of RyR2 to caffeine-mediated Ca²⁺ release, and raises the threshold of store-operated calcium release[23-25]. While studies to date are largely in vitro, these findings suggest that the 5diminished responsiveness of RyR2 to elevated store Ca²⁺ in the sarcoplasmic reticulum may lead to pro-arrhythmic Ca²⁺ release events in specific electrophysiologic conditions. In addition to missense variants which reduce RyR2 function, homozygous duplication of the RYR2 promoter, 5’UTR, and 5’ exons of RYR2 have also been reported, leading to reduced RyR2 expression and development of CRDS[26]. Finally, a new potential genetic cause of SQTS and sudden death has been described in variants in ATP1A3-encoded sodium-potassium ATPase alpha 3 subunit. Disease-associated variants have been linked with sudden unexplained death in epilepsy[27, 28], and recent work as linked the D801N missense variant in ATP1A3 with short QT interval and predisposition to SCD[29]. While more work is needed to determine the mechanism of this disease in the heart, and while ATP1A3 may represent a “stand along” SCD disease locus, these early studies are suggestive of a novel modulator of cardiac repolarization.

Post-mortem genetic evaluation following an unexplained sudden death

While done on a primarily research basis for many years, the clinical utility of post-mortem genetic evaluation following SCD has been recently codified into clinical practice. A so-called molecular autopsy is defined as the investigation of molecular defects that could have contributed or caused death in an individual, particularly the identification of genetic variants in SCD-associated genes. Cardiac ion channelopathy and cardiomyopathy genes are a key component of a post-mortem gene testing as up to 20% of sudden deaths that remain unresolved after a traditional autopsy will have a disease-associated variant in one of these genes[30, 31]. Further, given the significant proportion of individuals with heritable channelopathies or cardiomyopathies who remain genotype negative, clinical evaluation of family members of decedent is important. This is highlighted in recent consensus guidelines from the Heart Rhythm Society and the Asia Pacific Heart Rhythm society on the evaluation of those who experience sudden unexplained death or are resuscitated from sudden cardiac arrest[32]. These guidelines highlight the need for genetic testing in those who are resuscitated from cardiac arrest and post-mortem genetic evaluation of decedents for SCD-associated gene variants. These guidelines are inclusive of any age group, including pediatric patients.

Phenotypic expansion of genetic cardiomyopathies

Cardiomyopathies are due to pathologic alterations in proteins that are involved in cardiac myocyte contractility and development, thereby disrupting cardiac contractility, relaxation, and/or myocyte viability. Pediatric primary cardiomyopathy can have multiple etiologies and has an incidence rate of 1 per 100,000 patient-years[33, 34]. Three classes of cardiomyopathy have strong genetic association: hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), and arrhythmogenic right ventricular cardiomyopathy (ARVC). HCM is one of the most common causes of SCD in the pediatric population and the most common cause of SCD in young athletes[35]. Patients with HCM have an abnormally thickened/hypertrophied ventricle with impaired relaxation and may develop left ventricular outflow tract obstruction and arrhythmias[36]. Variants associated with HCM are traditionally in sarcomeric genes, principally MYH7, MYBPC3, and TNNT2, among others[37, 38]. DCM is characterized by dilation of ventricles with loss of systolic function which can result in heart failure and SCD. Approximately 45% of DCM in children is heritable[39]. However, DCM has variable expressivity in families and over 40 genes are associated with DCM, including many sarcomeric genes. Truncating variants in TTN-encoded titin, which are predicted to lead to truncation of the protein, contribute to 25% of inherited DCM while all other genes are minor causes[40, 41].

ARVC is classically marked by fibrofatty replacement of cardiomyocytes and can present with ventricular arrhythmias[42]. Disease-causing variants localize to desmosome protein encoding genes such as PKP2, DSP, DSG2, DSC2, and JUP, and transmembrane protein encoding gene TMEM43[43]. There is recent evidence that DSP-positive individuals might have a distinct form of cardiomyopathy. Classically, ARVC affects the right ventricular myocardium; however, DSP-positive individuals can have left ventricular involvement with an inflammatory component to the development and progression of disease[44, 45]. Recent evidence suggests that over 95% variants that are associated with the disease are in desmosomal genes[46]. Further, DSP-positive individuals, including pediatric cases, can present with a cardiocutaneous syndrome of cardiomyopathy, curly hair, and palmoplantar keratoderma in childhood[47]. Pediatric cases have been shown to have wide-range of non-cardiac findings including epidermolysis bullosa, reflecting the role of DSP in maintaining cell-to-cell adhesion in both keratinocytes and cardiac myocytes[48, 49]. This expansion of phenotype is reflective of cardiomyopathies, and there has been increasing appreciation that cardiomyopathies lie along a spectrum of diverse disease phenotypes, rather than discrete disease entities. For example, recent consensus statements have sought to redefine cardiomyopathies classically associated with arrhythmic forms of DCM and ARVC into a new class of arrhythmic cardiomyopathy[50]. As exome sequencing allows for genetic diagnoses of atypical disease presentations, the phenotypic spectrum of cardiomyopathic disease is likely to increase.

Resolving diagnostic uncertainty in genetic testing

Given the strong genetic basis of disease, pediatric and adult individuals clinically suspected to have a cardiac channelopathy or cardiomyopathy undergo diagnostic genetic testing to identify a gene and variant which may be responsible for their disease[51, 52]. While some genetic variants are clearly disease-associated, and others are clearly not, many variants have an unclear association with disease and are labeled variants of uncertain significance (VUS). The American College of Medical Genetics and Genomics (ACMG) classifies variants as pathogenic (P), likely pathogenic (LP), variant of uncertain significance (VUS), benign (B), or likely benign (LB), reflecting a probabilistic assignment of disease-association[53]. Clarifying the diagnostic significance of a VUS, shifting it to either LP/P or LB/B, is key for full implementation of a precision medicine-based approach to care. While there are several ways to resolve whether a VUS may be disease-associated, advances in genome editing through CRISPR, and induced pluripotent stem cells (iPSC) models of cardiac disease, have opened the door for functional validation of VUSs (Figure 2)[54]. CRISPR-based genome editing can create genomic knock-in mice hosting patient-identified VUSs to evaluate for evidence of disease, while similar knock-ins of human iPSCs (or derived directly from the patient in question) can be differentiated into cardiac myocytes (iPSC-CMs) to model disease, or lack thereof. Although differentiation methods to optimize the maturity of iPSC-CMs are continually reported[55, 56], both mouse models and iPSC-CMs have been used to resolve VUSs in both cardiomyopathies and channelopathies[57-59]. While these studies provide insight to diagnostic uncertainty, high-throughput methods are necessary to adequately validate the many variants that are continually reported. Methods such as high-throughput patch clamp[60] and machine learning/artificial intelligence[61] are promising alternatives to circumvent these challenges.

Figure 2:

Approaches to resolving a variant with diagnostic uncertainty. State-of-the-art functional/experimental validation of variants of uncertain significance (VUS) can employ genome-edited knock-in mouse models and induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) from either the patient who hosts the VUS in question, or knock-in of the VUS into a wild-type line. Functional studies then determine the likelihood that the VUS imparts a disease phenotype. Alternative promising approaches such as high throughput patch-clamp can achieve a higher throughput.

Interpretation of genetic variants changes over time

Further complicating the diagnostic uncertainty of VUSs is the evolution of both gene and variant disease-association overtime. Recent work by ClinGen[62] and ClinVar[63], and other research groups[64], have sought to objectively assess disease associations based on the strength and rigor of literature-based evidence. For example, ClinGen re-evaluation of LQTS found that of 17 genes previously reported to cause disease, 9 genes were disputed including AKAP9, ANK2, KCNE2, KCNJ5, SCN4B, and SNTA1. This re-evaluation is particularly salient in BrS, where ClinGen re-evaluation labeled all genes, aside from SCN5A, as disputed[65]. Among cardiomyopathy genes, similar studies have called into question a number of genes, leading to ClinGen re-evaluation of all genes in HCM, DCM, and ARVC. For example, in a recent study evaluating 51 DCM-associated genes, 32 had limited evidence, were disputed, or had no disease association[66]. A second curation study reduced DCM associated genes from 48 to 14[67]. Interestingly, resolving gene-disease associations reduced VUS burden, suggesting that clarification of which genes are truly associated with disease will ultimately help resolve diagnostic uncertainty associated with VUS[67]. Similarly, recent work in HCM has categorized only 8 out of 33 genes as definitively associated to disease and 22 had little evidence for association[68].

Likewise, genetic variants are increasingly appreciated to change over time. Periodic re-assessment of the role of distinct variants in disease can inform diagnosis as additional evidence is reported over time[69]. For example, ion channel recordings from high-throughput patch-clamp screening of SCN5A variants have influenced variant association to disease[60]. Similarly, family-based studies of patients with VUS can lead to reclassification of these variants as either LP or LB/B[70]. Between 2015-2019, a fifth of all likely pathogenic variants in ClinVar were re-classified as VUS, and over half of all VUS were re-classified as likely benign[71]. Recent work has also shown that variants in cardiac channelopathy genes have a ~1% chance/year/variant of changing interpretation in a clinically meaningful way (i.e. LP/P to VUS) and most often change toward uncertainty (i.e. VUS)[72]. These studies highlight the need to periodically reassess the pathogenicity of cardiovascular disease-associated variants found in patients.

Polygenic risk and genetic modifiers

Improved genome sequencing technology has facilitated the identification of new genetic modulators of cardiac physiology and disease (Figure 3)[73]. In one architype, a single Mendelian disease variant with a large effect can be modified by other genetic changes which, in isolation, have a smaller physiologic effect[74]. In another, multiple genetic variants with a relatively small physiologic effect integrate physiologically to cause phenotype variability and, in extreme cases, disease development. Careful characterization of these genetic modifiers can allow for quantification risk ascribed to a given variant, and when combined with other variants, opens the door to a polygenic risk score[75]. Genome-wide association studies (GWAS) are key to determining the influence of common genetic variants across the genome, conferring over 71,000 variant-trait associations since 2005[76]. In channelopathies, GWAS-derived polygenic risk scores have highlighted the influence of genetic architecture in determining phenotypic variability. In BrS, recent findings showed that multiple single nucleotide polymorphisms (SNPs) can cause BrS in the absence of bonefide SCN5A pathogenic variant, highlighting polygenic risk in developing disease among SCN5A genotype-negative individuals[77, 78]. GWAS can also describe variation in a population. For example, polygenic risk contribution to QTc in patients with different forms of LQTS overlap with variants that determine QTc in the general population. However, this risk contributes less to QTc variability within the general population, and more to genotype negative LQTS[79, 80]. In cardiomyopathies, GWAS have elucidated the complex genotype-phenotype association in both disease-associated and “protective” alleles[81]. Recent GWAS performed in patients with DCM have identified loci associated with canonical DCM genes while suggesting that novel areas of the genome may have influence over disease development[82]. Interestingly, common variants associated with HCM development can also be associated with DCM, with opposing physiologic effects[83, 84].

Figure 3:

Assessment of polygenic disease using genome wide association studies. Population studies using genome sequencing can determine the physiologic impact of common genetic variants on a phenotype, including disease. Utilizing modeling, such as machine learning, these variants can be used to predict penetrance or phenotype of cardiovascular diseases with a genetic etiology.

GWAS have shed light on polygenic risk and have tremendous potential in achieving a true precision medicine-based approach to cardiovascular genetic disease. However, inclusion of diverse populations is paramount to generalization of these associations given that most GWAS studies are conducted in populations with an overwhelming representation of European descent[85-87]. International organizations and researchers are calling for inclusion of ancestry-diverse groups to inform polygenic risk, as this will make findings applicable to more patients[88, 89].

Incidental findings in cardiac channelopathy and cardiomyopathy-associated genes

Given marked improvements in the ease and cost of DNA sequencing, exome (ES) and genome sequencing (GS) is increasingly used in the evaluation of infants and children. In prenatal diagnostic testing for congenital abnormalities, ES has demonstrated a 10% diagnostic yield[90]. In critically ill infants, the actionability of diagnostic ES is particularly great, altering care in 1 out of every 5 sick infants by affording higher diagnostic rates at an earlier age and in shorter hospital stays[91-93], with over $1 million in cost reduction[94]. Despite the diagnostic promise of ES and GS, incidental findings from these tests remain a challenge to implementation of ES and GS in predictive genetic testing. This is particularly true when these variants localize to cardiac channelopathy and cardiomyopathy genes. To address this, ACMG has outlined genes in which incidental findings are clinically actionable and should thus be reported as secondary findings[95]. Multiple studies suggest that ~1% of patients undergoing ES or GS will have a LP/P variant in a cardiovascular gene[96, 97]. While potentially informative of undiagnosed disease, this rate is much higher than the prevalence of the heritable cardiovascular diseases in the population[98]. In pediatric patients who underwent ES for non-cardiac indications, the rate of incidental LP/P variants found in CPVT-associated genes was 20-fold higher than the disease prevalence and 10-fold higher in LQTS-associated genes[99, 100]. This poses a diagnostic challenge in discerning which individuals are at higher risk for developing cardiovascular disease.

One approach to the evaluation of incidental variants is a Bayesian, probabilistic approach (Figure 4). A “pre-test” probability of clinical disease is established by a comprehensive evaluation that is individualized to the patient, personal/family history, and the variant in question. This includes a history, family history involving at least 3 generations, a physical exam, and appropriate diagnostic testing. After determination of the pretest probability, the strength of association of the gene and variant is incorporated to determine a “post-testing” probability that the patient has a disease-associated variant. However, onset of disease phenotype due to variants in disease associated genes can manifest at a later age, thus limiting the conclusiveness of physical exam and cardiovascular testing at the time of sequencing. Longitudinal follow-up of the patient can inform post-test probability and reassessment of the variant over time. The strength of association of the variant with the disease in question should be comprehensive as variant calling by clinical genetic testing companies can be inconsistent based on limited clinical data to evaluate findings[101]. Further, this variant reevaluation typically occurs in the context of the ACMG variant classification framework and can aid in assigning pathogenicity to variants in disease associated genes[53]. Recent work has shown that incorporation of specific clinical and genetic findings into this probabilistic model can lead to improved diagnostic clarity[102]. Additionally, signal-to-noise analysis at the amino acid level can discern pathogenicity using a pathogenic variant relative risk to identify pathogenic hotspots[103]. This approach has been shown to increase the diagnostic accuracy of incidentally identified variants when incorporated into ACMG criteria[100]. These approaches, and others, are needed to balance the promise of diagnostic ES and GS testing with clear discrimination of the likelihood that incidentally identified variants in cardiovascular disease-genes will lead to disease expression.

Figure 4:

Evaluation of patients with incidental variants in cardiovascular genes. Patients with incidentally identified variant in a cardiac channelopathy or cardiomyopathy-associated gene found on diagnostic exome or genome sequencing should be evaluated to determine the likelihood that the identified variant is truly disease associated. One way to accomplish this is through a structured framework consisting of a comprehensive clinical evaluation of the patient, to determine a “pre-test” probability that the patient, or their family, might have cardiac disease, followed by re-evaluation of the genetic finding to determine a post-test probability of disease-association.

Conclusion

Advances in genetic sequencing are leading to identification of new cardiac channelopathies and cardiomyopathies, broadening the known phenotypes of disease, and offer the promise of true disease risk-prediction. Future advances in GWAS using diverse populations will expand the use of polygenic risk and gene modifiers in patient care, and the low cost and high diagnostic yield of genome sequencing will continue to promote the use of GS and ES in pediatric patients. This predicted rise in the use of GS and ES in other clinical scenarios in a genetic association landscape that is increasingly complex will inevitably lead to a precision medicine-based approach to prevent disease if challenges in incidental variant interpretation can be overcome.