Issues and Challenges in Diagnostic Sequencing for Inherited Cardiac Conditions

Transcription

1 Papers in Press. Published November 22, 2016 as doi: /clinchem The latest version is at Clinical Chemistry 63: (2017) Reviews Issues and Challenges in Diagnostic Sequencing for Inherited Cardiac Conditions Roddy Walsh 1,2 and Stuart A. Cook 2,3,4,5* BACKGROUND: Inherited cardiac conditions are a relatively common group of Mendelian diseases associated with ill health and death, often in the young. Research into the genetic causes of these conditions has enabled confirmatory and predictive diagnostic sequencing to become an integral part of the clinical management of inherited cardiomyopathies, arrhythmias, aortopathies, and dyslipidemias. CONTENT: Currently, the principle benefit of clinical genetic testing is the cascade screening of family members of patients with a pathogenic variant, enabling targeted follow up of presymptomatic genotype-positive individuals and discharge of genotype-negative individuals to health. For the affected proband, diagnostic sequencing can also be useful in discriminating inherited disease from alternative diagnoses, directing treatment, and for molecular autopsy in cases of sudden unexplained death. Advances in sequencing technology have expanded testing panels for inherited cardiac conditions and driven down costs, further improving the cost-effectiveness of genetic testing. However, this expanded testing requires great rigor in the identification of pathogenic variants, with domain-specific knowledge required for variant interpretation. SUMMARY: Diagnostic sequencing has the potential to become an integral part of the clinical management of patients with inherited cardiac conditions. However, to move beyond just confirmatory and predictive testing, a much greater understanding is needed of the genetic basis of these conditions, the role of the environment, and the underlying disease mechanisms. With this additional information it is likely that genetic testing will increasingly 1 NIHR Cardiovascular Biomedical Research Unit, Royal Brompton and Harefield NHS Foundation Trust and Imperial College London, London, UK; 2 National Heart and Lung Institute, Imperial College London, London, UK; 3 National Heart Research Institute Singapore, National Heart Centre Singapore, Singapore; 4 MRC Clinical Sciences Centre, Imperial College London, London, UK; 5 Division of Cardiovascular & Metabolic Disorders, Duke-National University of Singapore, Singapore. * Address correspondence to this author at: Division of Cardiovascular & Metabolic Disorders, Duke-National University of Singapore, 8 College Rd., Singapore , Singapore. stuart.cook@duke-nus.edu.sg. Received June 1, 2016; accepted October 24, Previously published online at DOI: /clinchem American Association for Clinical Chemistry be used for stratified and preventative strategies in the era of genomic medicine American Association for Clinical Chemistry Inherited cardiac conditions (ICCs) 6 comprise a broad range of syndromes affecting the heart and major blood vessels, including cardiomyopathies, arrhythmias, aortopathies, familial hypercholesterolemia (FH), and congenital structural heart defects. Collectively, these disorders are relatively common and are associated with adverse clinical outcomes such as sudden cardiac death (SCD), heart failure, and premature coronary heart disease (CHD). ICCs are typically monogenic (each case caused by a single variant in 1 gene) autosomal dominant conditions, but genetically and allelically heterogeneous (there are many genes in which variation can cause disease). ICC genes feature prominently in the American College of Medical Genetics and Genomics list of proposed genes to be routinely analyzed and interpreted in all exome or genome sequencing (1). Research into the genetic factors underlying ICCs has made substantial progress in recent years and has ensured that genetic testing for several conditions has made the transition from research to clinical diagnostic laboratories. However, considerable challenges in fully understanding the genetics of these diseases, and explaining how observed variation relates to clinical phenotype, will need to be addressed to maximize the benefit of diagnostic sequencing in ICCs. Major ICCS While ICCs comprise a wide range of conditions, from relatively common to extremely rare, this review focuses on some of the conditions for which the genetic etiology has been most studied, and therefore diagnostic sequenc- 6 Nonstandard abbreviations: ICC, inherited cardiac condition; FH, familial hypercholesterolemia; SCD, sudden cardiac death; CHD, coronary heart disease; HCM, hypertrophic cardiomyopathy; LV, left ventricular; DCM, dilated cardiomyopathy; LV, left ventricular; ARVC, arrhythmogenic right ventricular cardiomyopathy; RV, right ventricular; LQTS, long QT syndrome; CPVT, catecholaminergic polymorphic ventricular tachycardia; FM, familial hypercholesterolaemia; LDL-C, LDL cholesterol; QTc, corrected QT; dhplc, denaturing HPLC; NGS, next generation sequencing; WES, whole exome sequencing; WGS, whole genome sequencing; VUS, variant of unknown significance; ExAC, exome aggregation consortium; SNP, single nucleotide polymorphism. 1 Copyright (C) 2016 by The American Association for Clinical Chemistry

2 Reviews ing most implemented, including cardiomyopathies, arrhythmias, and FH. The contribution of the main genes associated with these conditions are summarized in Table 1. HYPERTROPHIC CARDIOMYOPATHY Hypertrophic cardiomyopathy (HCM) is defined by the presence of primary left ventricular (LV) hypertrophy above a defined threshold in the absence of other cardiac or systemic conditions, and is characterized by myocyte disarray and myocardial fibrosis (2). It affects approximately 1 in 500 people and is the most frequent cause of sudden death in young people and athletes, although the majority of patients have a normal life expectancy. The genetic basis of HCM was first elucidated in the 1990s, with variants in sarcomeric genes identified through linkage studies in large family pedigrees (3). Although variants in over 50 genes have since been claimed to be associated with HCM, 8 sarcomeric genes still account for the vast majority of pathogenic variants, with expanded panels offering little increased clinical sensitivity (4). 7 Human genes: TTN, titin; PKP2, plakophilin 2; DSC2, desmocollin 2; DSG2, desmoglein 2; DSP, desmoplakin; JUP, junction plakoglobin; KCNQ1, potassium voltage-gated channel subfamily Q member 1; KCNH2, potassium voltage-gated channel subfamily H member 2; SCN5A, sodium voltage-gated channel alpha subunit 5; RYR2, ryanodine receptor 2 (cardiac); CASQ2, calsequestrin 2 (cardiac muscle); TRDN, triadin; CALM1, calmodulin 1 (phosphorylase kinase, delta); KCNJ2, potassium voltage-gated channel subfamily J member 2; LDLR, low density lipoprotein receptor; TNNT2, troponin T2, cardiac type; TNNI3, troponin I3, cardiac type; TPM1, tropomyosin 1 (alpha); ACTC1, actin, alpha, cardiac muscle 1; MYBPC3, myosin binding protein C, cardiac; MYH7, myosin, heavy chain 7; LMNA, lamin A/C; NOS1AP, nitric oxide synthase 1 adaptor protein; APOB, apolipoprotein B; PCSK9, proprotein convertase subtilisin/kexin type 9; MYL2, myosin, light chain 2; MYL3, myosin, light chain 3; BAG3, BCL2 associated athanogene 3; RBM20, RNA binding motif protein 20. DILATED CARDIOMYOPATHY Dilated cardiomyopathy (DCM) is currently defined by the presence of left ventricular (LV) or biventricular dilation and systolic dysfunction in the absence of abnormal loading conditions (hypertension, valve disease) or coronary artery disease sufficient to cause global systolic impairment. The causes of DCM can be classified as genetic or nongenetic, but there are circumstances in which genetic predisposition interacts with extrinsic or environmental factors (5). DCM is a frequent cause of heart failure and the most common indication for heart transplantation and is estimated to affect at least 1 in 500 people. DCM is a genetically heterogeneous disease, although only titin (TTN) 7 has a substantial ( 5%) yield in genetic testing, with truncating variants (frameshift, nonsense, and essential splice site variants that disrupt the open reading frame) in this gene accounting for 15% 25% of patients with DCM (6, 7). Although DCM gene panels have expanded to incorporate novel gene associations, the yield of clinical testing still does not exceed 40% (8). ARRHYTHMOGENIC RIGHT VENTRICULAR CARDIOMYOPATHY Arrhythmogenic right ventricular cardiomyopathy (ARVC) is a progressive cardiomyopathy and is estimated to affect 1 in people. It most frequently affects the right ventricular myocardium, although LV involvement may also be observed. Most variants associated with ARVC occur in genes encoding desmosomal proteins [plakophilin 2 (PKP2), desmocollin 2 (DSC2), desmoglein 2 (DSG2), desmoplakin (DSP), junction plakoglobin (JUP)] (9). Diagnostic testing in ARVC is hampered by the relatively high rate of rare missense variants in these genes in the general population, which limits interpretability and questions many published associations that were not adequately controlled (10, 11). However truncating variants in these genes are likely to be actionable, especially in PKP2 where they occur in approximately 25% of ARVC patients (12). LONG QT SYNDROME Long QT syndrome (LQTS) is an inherited arrhythmia that causes cardiac syncope and SCD. It is diagnosed by a characteristic QT prolongation on electrocardiogram and is estimated to affect 1 in 2500 people. It occurs more commonly in the autosomal dominant form (Romano- Ward syndrome) and rarely as the autosomal recessive Jervell and Lange-Nielsen syndrome (associated with sensorineural hearing loss). Variants in 3 ion channel genes account for over 70% of patients with a definitive diagnosis of LQTS (13, 14) (although the yield is substantially less for all referral cases (15): loss of function variants in the potassium ion channel genes potassium voltage-gated channel subfamily Q member 1 (KCNQ1) and potassium voltage-gated channel subfamily H member 2(KCNH2; LQT1 and LQT2 respectively) and gain of function variants in the sodium ion channel gene sodium voltage-gated channel alpha subunit 5 (SCN5A; LQT3). More than 10 other ion channels and associated proteins are more rarely associated with LQTS. BRUGADA SYNDROME Brugada syndrome is an inherited cardiac arrhythmia characterized by distinct type ST segment elevation on electrocardiogram examination and is associated with an increased risk of SCD in young adults. The prevalence of the disease varies significantly based on location and ethnicity and is highest in Southeast Asia, where it affects approximately 1 in 2000 people and is known in the local vernacular in Thailand as Lai Tai (death in sleep). Loss of function variants in SCN5A, both truncating and missense, occur in approximately 20% of Brugada patients (16). Variants in over 20 other genes have also been 2 Clinical Chemistry 63:1 (2017)

3 Diagnostic Sequencing for ICCs Reviews Table 1. Summary of the main genes associated with major ICCs described in this review. ICC Gene Gene name Proportion of cases References HCM MYBPC3 Myosin binding protein C, cardiac 15% 20% Alfares et al. (4); Walsh et al. (12) MYH7 Myosin, heavy chain 7 10% 15% TNNI3 Troponin I3, cardiac type 2% TNNT2 Troponin T2, cardiac type 2% TPM1 Tropomyosin 1 (alpha) 1.5% MYL2 Myosin, light chain 2 1% MYL3 Myosin, light chain 3 1% ACTC1 Actin, alpha, cardiac muscle 1 0.5% DCM TTN Titin 13% 27% Herman et al. (6); Roberts et al. (7) MYH7 Myosin, heavy chain 7 5% Pugh et al. (8); Walsh et al. (12) LMNA Lamin A/C 5% DSP Desmoplakin 5% TNNT2 Troponin T2, cardiac type 2% 3% TPM1 Tropomyosin 1 (alpha) 2% TNNI3 Troponin I3, cardiac type <1% BAG3 BCL2 associated athanogene 3 N/A a Norton et al. (61); Villerd et al. (62) RBM20 RNA binding motif protein 20 N/A Brauch et al. (63); Lietal.(64) ARVC PKP2 Plakophilin 2 25% 30% Walsh et al. (12) DSP Desmoplakin 9% DSG2 Desmoglein 2 3.5% DSC2 Desmocollin 2 <2% JUP Junction plakoglobin N/A LQTS KCNQ1 Potassium voltage-gated channel subfamily Q member 1 KCNH2 SCN5A Potassium voltage-gated channel subfamily H member 2 Sodium voltage-gated channel alpha subunit 5 Brugada SCN5A Sodium voltage-gated channel alpha subunit 5 40% 55% Hedley et al. (65) 35% 45% 2% 8% Kapplinger et al. (16) CPVT RYR2 Ryanodine receptor 2 (cardiac) 50% Priori et al. (17); Laitinen et al. (18); Medeiros-Domingo et al. (19) CASQ2 Calsequestrin 2 (cardiac muscle) N/A TRDN Triadin N/A CALM1 Calmodulin 1 (phosphorylase kinase, delta) N/A KCNJ2 Potassium voltage-gated channel subfamily N/A J member 2 FH LDLR Low density lipoprotein receptor 70% 80% Henderson et al. (66); Marduel et al. (67); Bertolini et al. (68) APOB Apolipoprotein B 2% 7% PCSK9 Proprotein convertase subtilisin/kexin type 9 <1% a N/A, not applicable. Clinical Chemistry 63:1 (2017) 3

4 Reviews associated with the disease, but their prevalence is uncertain, with no large, broadly sequenced cohorts yet published. CATECHOLAMINERGIC POLYMORPHIC VENTRICULAR TACHYCARDIA Catecholaminergic polymorphic ventricular tachycardia (CPVT) is an inherited cardiac arrhythmia characterized by polymorphic ventricular tachycardia induced by adrenergic stress and has an estimated prevalence of 1 in people. Missense variants in the cardiac ryanodine receptor encoding the ryanodine receptor 2 (RYR2) gene are detected in approximately 50% of CPVT probands (17 19). Although variant interpretation in this gene is complicated by the relatively high background frequency of rare variants, a tiered strategy has been proposed that relies on the clustering of pathogenic variants into 3 distinct clusters in RYR2 (19). A small number of CPVT cases have been associated with variants in the calsequestrin 2 (CASQ2; autosomal recessive), triadin (TRDN), calmodulin 1 (CALM1), and potassium voltage-gated channel subfamily J member 2 (KCNJ2) genes. FAMILIAL HYPERCHOLESTEROLAEMIA Familial hypercholesterolemia (FH) leads to premature CHD due to raised plasma levels of LDL cholesterol (LDL-C). It is the most common monogenic disorder, affecting up to 1 in 200 people, though most remain undiagnosed (20). Genetic testing for FH has a high yield, with variants in the low density lipoprotein receptor (LDLR) gene detected in approximately 80% of cases. The availability of highly effective statin therapy for patients with FH ensures that diagnosis of this disease through genetic testing in a proband, and follow on cascade screening of the family, can have substantial health benefits, which has led to national efforts in some countries. Benefits of Genetic Testing The practical benefits of incorporating genetic testing into the diagnostic screening process for ICCs encompass several factors. CASCADE SCREENING One of the main drivers for genetic testing in ICCs is to enable cascade screening of family members of the index patient. If a definitive pathogenic variant is identified in the proband, relatives can be genotyped to determine who is at risk of developing the disease, which may be beneficial for presymptomatic and younger family members. Genetic testing has twin advantages: (a) those without the causative variant can be discharged and exempted from regular follow-up clinical screening, offering both economic savings to healthcare systems and psychological benefits to individuals (lifelong surveillance is standard practice for many ICCs, especially cardiomyopathies, due to age-related penetrance and many conditions can never be confidently ruled out without a genetic test), and (b) those carrying the variant can be offered detailed clinical investigations, follow-up, and hopefully increasingly, prophylactic treatments. The latter is particularly critical for assessing the risk of SCD in patients with arrhythmias and cardiomyopathies and for employing statin therapy to treat CHD in young patients with pathogenic FH variants. DIFFERENTIAL DIAGNOSIS AND THERAPY GUIDANCE Currently, genetic testing of a clinically affected proband provides limited impact into clinical management of the proband. In most cases, the phenotype of the patient will be defined through clinical diagnostic approaches (such as imaging) that provide the main basis for therapeutic decisions. An exception to this is the diagnosis of metabolic cardiomyopathies, which can present with LV hypertrophy and therefore be initially diagnosed as HCM. Genetic testing in these cases can reveal pathogenic variants in the causative genes of the following conditions Danon disease, Fabry disease, Wolff Parkinson White syndrome, and glycogen storage disease II/Pompe disease that can directly impact treatment options, e.g., enzyme replacement therapy for Fabry disease. Detecting or confirming the subtype of LQTS through genetic testing can also play a role in guiding therapy through lifestyle modifications and drug treatment. Based on known triggering stimuli, patients with LQT1 are advised to avoid strenuous exercise (particularly swimming and competitive sports), patients with LQT2 should try to avoid unexpected auditory stimuli, whereas patients with LQT3 may be advised to be monitored or accompanied while sleeping (21). Treatment with blockers such as atenolol or nadolol is recommended for all patients with LQTS but may be most effective for LQT1 (22), in particular in patients with specific classes of KCNQ1 variants (23). In contrast, LQT3 (caused by gain of function variants in SCN5A) may be treated with sodium channel blockers such as mexilitine, flecainide, or ranolazine (24). RISK ASSESSMENT AND GENOTYPE PHENOTYPE CORRELATION Initial hopes that genetic testing for ICCs could be useful for risk assessment, i.e., prediction of the disease severity/ outcomes and the specific therapy that warranted, have been confounded by the limited correlation between genotype and phenotype in many of these conditions. ICCs are generally characterized by variable expressivity and penetrance, even between family members carrying the same pathogenic variant, and there is currently sparse 4 Clinical Chemistry 63:1 (2017)

5 Diagnostic Sequencing for ICCs Reviews evidence that different classes of variants have different outcomes. In several ICCs, patients with multiple pathogenic variants (either compound in the same gene or digenic) have more severe phenotypes than patients with single variants. In LQTS, patients with 2 pathogenic variants had longer corrected QT (QTc) intervals and were more likely to experience cardiac events and cardiac arrest than patients with single variants (25). This gene dosage effect has also been observed in HCM, in which multiple pathogenic variants have been associated with more adverse outcomes in a number of studies (26 28). In addition, patients with HCM who have a sarcomeric gene variant have been shown to have an increased risk of cardiovascular events and heart failure compared to patients with no identified pathogenic variant (29, 30). Differentiating phenotype based on the presence of variants in specific genes has thus far had limited success. In HCM, there have been conflicting reports as to whether variants in thin filament genes [troponin T type 2, cardiac type (TNNT2), troponin I3, cardiac type (TNNI3), tropomyosin 1 (alpha) (TPM1), actin, alpha, cardiac muscle 1 (ACTC1)] confer distinct clinical phenotypes to variants in the major thick filament genes [myosin binding protein C, cardiac (MYBPC3), myosin heavy chain 7 (MYH7)], likely due to limited study sizes given the small proportion of patients with HCM who have thin filament variants. However, a recent multicenter study concluded that thin filament variants were associated with an increased risk of advanced LV dysfunction and heart failure, though with milder LV hypertrophy and outflow tract obstruction and similar rates of arrhythmia and SCD (31). No difference in clinical outcomes have been observed between patients with HCM harboring the 2 major thick filament variants (30). Limited data exist for genotype-phenotype correlation in DCM although it is believed that patients with lamin A/C (LMNA) variants have higher risks of cardiac events, particularly malignant ventricular arrhythmias (32). COST-EFFECTIVENESS OF GENETIC TESTING A number of studies have been published assessing the cost-effectiveness of genetic testing in patients with HCM and their families in the UK (33), Australia (34), and the US (4). These studies examined economic decision models comparing genetic cascade screening with the standard clinical approach for identifying and treating individuals at risk, as well as the potential savings associated with detecting and discharging genotypenegative relatives. The UK and Australian studies concluded that genetic cascade screening is more costeffective for HCM than clinical diagnosis alone, with the incremental cost per life year saved/gained of in the UK (33) and 9509 in Australia (34), well within the recommended thresholds of (UK National Institute for Health and Care Excellence guidelines) and (Australian Pharmaceutical Benefits Advisory Committee). The savings associated with discharging genotype-negative relatives were estimated as 8615 for an individual over his/her lifetime in the UK study (33) and $0.7 million for a cohort of 2912 patients with HCM in the US study, covering 691 discharged family members (4). Given the conservative estimates used in these models, the benefits associated with expanding cascade screening beyond first degree relatives together with the recent dramatic reductions in the costs of genetic sequencing, the cost-effectiveness of comprehensive genetic testing in HCM is likely to be substantially greater. Because the diagnostic yield in DCM is approximately 20%, the savings will be less than for HCM for discharges, but may be greater for prophylactic approaches in genotype positive individuals because DCM has substantial costs associated with heart failure treatments, devices, and transplantation. New cost-effectiveness studies for DCM (in the era of TTN sequencing) and other ICCs are required to accurately gauge the economic benefits of genetic testing in these conditions, particularly to incorporate the ongoing reductions in sequencing costs. Sequencing Technologies Advances in DNA sequencing technologies have had a dramatic effect on the use of genetic testing for clinical use. Diagnostic testing initially depended on techniques such as denaturing HPLC (dhplc), Sanger sequencing, or array-based technologies (35). These methods were variously expensive (Sanger), time consuming (dhplc, Sanger), lacking in analytical sensitivity (dhplc), and difficult to scale up (all, especially for large genes such as TTN). Although genetic testing using these techniques became established for a number of ICCs, these limitations prevented their wide scale adoption as a routine part of clinical management and it will be some time before clinical workflows take advantages of the new opportunities now apparent. The development of next generation sequencing (NGS) technologies since 2005 has addressed many of these shortcomings. NGS approaches have dramatically reduced the costs and timescales compared to traditional techniques, while offering an analytically sensitive and accurate sequencing for variant detection, although many clinical laboratories still, likely unnecessarily, confirm findings with an alternative technique (usually Sanger sequencing). NGS has also enabled expanded testing to account for the increasing number of genes associated with many ICCs, though this can present substantial challenges in variant interpretation and domainspecific knowledge is increasingly important. Clinical Chemistry 63:1 (2017) 5

6 Reviews Fig. 1. Sequencing coverage metrics for ICC genes on the TruSight Cardio panel using MiSeq and NextSeq and comparison with coverage for these genes using WES, Deep WES (i.e. 6 samples combined to obtain sequencing depth equivalent to TruSight Cardio panel), and WGS, as well as approximate targeted enrichment and sequencing cost per sample. Figure generated from data in Table 2 of (36). Currently, NGS strategies for diagnostic testing include predefined gene panels, whole exome sequencing (WES) and whole genome sequencing (WGS), each of which offers distinct advantages and disadvantages. Gene panels can offer comprehensive sequencing of most or all genes that have been associated with the disease in question, allowing both the sequencing and data analysis to focus on these validated genes. Pan-ICC panels such as Illumina s TruSight Cardio panel (36) offer a single laboratory assay for a wide range of cardiac diseases, with the possibility to restrict data analysis to just those genes or conditions in question. Gene panels have been demonstrated to obtain excellent coverage of target regions and analytical sensitivity in variant detection, outperforming both WES and WGS (36) (Fig. 1). However, panels do not allow for variant detection in novel genes, require periodic review to account for newly discovered gene associations and targeted sequencing is poor at detecting certain variants classes such as copy number variants. WES involves the targeted sequencing of all coding regions of the genome, comprising approximately 30 6 Clinical Chemistry 63:1 (2017)

7 Diagnostic Sequencing for ICCs Reviews million bases across over genes. WES enables comprehensive analysis of the protein-altering variation in the genome (which accounts for the vast majority of known pathogenic variants in ICCs), both for genes with a known role in the disease of interest and for the discovery of novel genetic associations. However, identifying causative variants in novel genes is extremely difficult if not impossible in individual cases, given the large number of very rare or private variants found in every exome. WES is more successful at identifying putative pathogenic variants in trios and small nuclear families in which a known ICC gene is not involved, particularly in cases of recessive or de novo inheritance (e.g., an affected child with 2 unaffected parents, unusual for ICCs) (37, 38). WES can also be used as a comprehensive gene panel test, with analysis restricted to those genes associated with the disease (an exome slice ), with the possibility of reanalysis when novel genetic links are proposed. However, limitations in sequencing coverage associated with WES, particularly for key disease genes and especially for difficult-to-capture but clinically important exons (e.g., exon 1 of KCNQ1), should be taken into account before pursuing this strategy. WGS involves the complete sequencing of every base pair in the genome. It is the most comprehensive sequencing strategy available, provides more uniform coverage than WES, is more suited to detecting large structural variants, and is increasingly an affordable option with reduced sequencing costs. However, the practical advantages of WGS for diagnostic testing of individual patients are currently questionable because prioritizing the thousands of rare variants observed in each genome, and identifying actionable putative pathogenic variants, is an extremely difficult proposition, and even more challenging than the analysis of novel coding variants identified by WES. There are also major issues relating to data processing, storage, and indeed ownership with WGS. It is true that the data may be reinterrogated for other variants in the case of a second genetic condition (though it is questionable how often this is really needed), but if the data are in one hospital in one format and kept in deep storage, there are large logistic, manpower, and cost implications with accessing these data and sharing them with a second healthcare provider. Variant Analysis and Interpretation The accurate interpretation of genetic variants, in particular the identification of pathogenic variants that can impact the clinical management of patients and their families, is the single greatest challenge in diagnostic genetic testing for ICCs (and other Mendelian disorders). Unlike research findings, the consequences of a falsepositive or false-negative result in a clinical diagnostic setting can be devastating. However, as a result of this, clinical laboratories may be too conservative in assigning pathogenicity to variants and reporting back variants of unknown significance (VUS) that could be classified as at least likely pathogenic with more discriminating analysis (12). Multiple lines of evidence can be assessed to help classify variants (Table 2). These include the population frequency of a variant [derived from databases such as 1000 Genomes, the Exome Sequencing Project, and the Exome Aggregation Consortium (ExAC)](39), prior knowledge of the clinical and functional effect of a variant, segregation of a variant in affected family members (for the patient in question or previously studied or published pedigrees), and computational algorithms that predict the effect of a variant on protein function. Recently, guidelines such as those produced by the American College of Medical Genetics and Genomics (40) have attempted to standardize how this evidence is utilized to classify variants, by determining the weight that should be applied to each line of evidence and the requirements for categorizing variants into 5 recommended classes pathogenic, likely pathogenic ( 90% chance), VUS, likely benign ( 90% chance), and benign. Although the American College of Medical Genetics and Genomics guidelines provide a framework for the analysis of variants associated with Mendelian disease, their implementation in clinical diagnostic laboratories requires addressing the ambiguities associated with some lines of evidence. In particular, defining the evidence required for a variant to be classified as known pathogenic is challenging. The advent of population databases revealed that many variants that had been published as disease-causing (and listed in resources such as the Human Gene Mutation Database) were in fact present at far too high a frequency to be disease-causing (10, 41), indicating that these studies were poorly controlled or relied on insufficient evidence for defining pathogenicity. The ClinVar database was recently established to enable clinical laboratories to share information on the relationships between human variations and phenotypes (42). However, the standard of evidence associated with ClinVar entries is highly variable, obliging users to manually assess the data supporting each claim. To assist these efforts, well established cardiac diagnostic laboratories including the Laboratory of Molecular Medicine (4, 8) and the Oxford Molecular Genetics Laboratory (12) have recently published compendiums of their experience of over 10 years of cardiomyopathy genetic testing. We have created an online resource, the Atlas of Cardiac Genetic Variation ( to enable users to easily access these data. Currently the evidence required to classify a variant as pathogenic relies heavily on observed segregation of Clinical Chemistry 63:1 (2017) 7

8 Reviews Table 2. Classes of evidence used in the interpretation of variants. Evidence Pros Cons Segregation Can provide very strong evidence of pathogenicity if observed in sufficient affected family members, especially including more distant relatives. Lack of segregation, i.e., absence of the variant in clearly affected family members, provides strong evidence against pathogenicity. Data from other family pedigrees with the same variant (published and previous lab experience) can also be used as evidence. Known Established evidence that a variant is pathogenic pathogenicity for a disease provides a quick and clear result. Functional evidence Population frequency Computational prediction Pathogenic hotspots Can provide powerful data on the functional effect of variants, particularly in well validated assays. Allows for quick filtering of most variants identified in broad genetic tests. Data are easily calculated for all missense and splice region variants. A multitude of algorithms are available, with consensus predictions offering increased accuracy. Variants occurring in regions enriched in disease cohorts and/or depleted in controls have increased likelihood of pathogenicity. Large datasets can generate a prior probability of pathogenicity for variants found in enriched regions. Large family pedigrees are rarely available for genetic testing in the clinic. Incomplete penetrance of many ICCs means only affected family members can be used in segregation analysis. Segregation may imply that the locus is linked to the disease rather than the specific variant. Negates one of main benefits of genetic testing, i.e., cascade screening of family members to assess disease status. Defining the amount and type of evidence that is sufficient to unambiguously classify a variants as pathogenic can be difficult and subjective. Many variants in published studies have been incorrectly labelled as diseasecausing. Gathering evidence on known variants from the literature and online databases can be time-consuming. Lack of standardization of functional assays for many ICCs. Translation of cellular and tissue assays and animal models to the clinical phenotype is often uncertain. Rarity in population databases does not imply that a variant is pathogenic. As population databases can include affected individuals, determining a frequency that is compatible with pathogenicity can be difficult. The patient s ethnicity should match the population studied, data for certain population groups can be limited or unavailable. Low positive predictive values mean these can only be used as supporting evidence for pathogenicity. Algorithms assess effect on protein function rather than pathogenicity in disease. Not relevant for all genes associated with ICCs. the variant with the phenotype in affected families, a significant excess of the variant in disease cohorts over control groups (it should be absent or at very low frequencies, e.g., 1:10000 alleles for HCM, in population databases), and characterization with wellestablished functional assays, preferably in vivo animal models. The exception is truncating variants in genes with a known loss of function mechanism for the disease, such as MYBPC3 in HCM, where the probability of pathogenicity is high enough to negate the need for additional evidence. For other variants, either novel or for which limited data already exists, it can be difficult to definitively assess their effect on phenotype, given that segregation analysis and functional characterization are 8 Clinical Chemistry 63:1 (2017)

9 Diagnostic Sequencing for ICCs Reviews Fig. 2. Cumulative increase in the number of genes reported to be associated with major ICCs. Data are derived from the earliest published link for each gene:disease pair in the HGMD database. not routinely available in diagnostic laboratories. To address this, recent studies have focused on improving the interpretability of novel and rare variants found in key cardiac genes, by using large disease and control cohorts to identify regions of ion channels (associated with LQTS) (14, 43), MYH7 (missense variants in HCM) (12) and TTN (truncating variants in DCM) (7) that are enriched with variants in patients. A prior probability of pathogenicity for rare variants found in these regions can be calculated from these analyses (described as the etiological fraction (12) or estimated predictive value (14), which if sufficiently high (e.g., 97% in the HCM cluster of MYH7 (12), can provide strong evidence that a variant is disease-causing. An increasing issue facing clinical genetics laboratories is the reinterpretation of previously identified variants in the light of new evidence, particularly if this leads to the reclassification of variants that have been reported to patients. New entries to resources like ClinVar or analysis techniques as described above may enable some VUS to be reclassified as pathogenic and initiate cascade screening in the affected patients families. More seriously, variants that have previously been reported as disease causing have been reassessed with improved, population-specific controls such as provided by ExAC. Inadequate control cohorts were recently shown to cause misdiagnosis in HCM patients of African ancestry (44). Challenges and Future Directions in Genetic Testing MISSING HERITABILITY AND INACCURATE GENETIC ASSOCIATIONS Despite decades of research into the genetics of Mendelian cardiac disease, the genetic test yield for several ICCs has remained stubbornly low, particularly for important conditions like HCM (approximately 40%), DCM (approximately 30% at best) and Brugada syndrome (approximately 20% at best). Although not all cases of these conditions will have a genetic basis, the substantial proportion of familial cases without identified causative variants suggests that additional genetic factors remain to be discovered. These factors may include large structural deletions (currently largely invisible using existing techniques) and copy number variations, variants affecting regulatory regions of genes or exon splicing, mosaicism (whole body or tissue level), or polygenic models of disease for which several variants and the environment may contribute to the phenotype. To date, most studies have focused on trying to identify protein-altering variation in candidate disease genes with known cardiac function and a hypothetical role in the ICC in question, or using broad panels of genes associated with a range of ICCs (45, 46). These studies have substantially increased the number of genes putatively associated with a number of ICCs (Fig. 2). Rare Clinical Chemistry 63:1 (2017) 9

10 Reviews Fig. 3. Proportion of individuals with rare variants (ExAC allele frequency < ) in HCM and DCM clinical cohorts (red = pathogenic, orange = likely pathogenic, yellow = VUS) compared to ExAC (grey columns). Many genes on clinical testing panels show no excess over ExAC, casting doubt on their involvement in cardiomyopathy. ˆ, genes analyzed in fewer than 200 cases. Reproduced from (12) with permission. variants (typically defined as a population allele frequency 0.001) in these genes that are identified through sequencing of large disease cohorts are classified as putative pathogenic variants, supported variously by computational algorithms, functional assays, or segregation of the variant in small family groups. However, this approach fails to recognize that although these variants may be individually rare, they can be collectively quite common when assessed across the entire gene. Without sequencing a similarly sized control cohort, or assessing the level of rare variation in genes in the available population databases such as ExAC, it cannot be demonstrated that there is an excess burden of rare variants in cases for these putatively associated genes. Indeed, many genes that have been linked to cardiomyopathies and are regularly sequenced in clinical diagnostics have no case excess compared to the ExAC reference population (12) (Fig. 3). Because most of these genes lack any variants with detailed linkage or segregation data for the diseases in question, it is likely they have been erroneously associated with these conditions. Far from addressing the missing heritability, these studies have clouded our understanding of the genetic architecture of many ICCs, leading to an increase in the number of VUSs, or potentially even false positive results. Because diagnostic sequencing can now easily accommodate large gene panels or even WES/WGS approaches, it is critical that the analysis of detected variants focuses on genes with strong evidence for association with the query condition. To assist diagnostic laboratories in this task, the Clinical Genome Resource is developing frameworks to assess the evidence for associations between genes and particular diseases (47). A clear understanding of the roles of genes in Mendelian diseases is an essential step for the analysis and interpretation of genetic variants identified in patients. MODIFIER GENES AND VARIANTS Most ICCs are characterized by variable expressivity, i.e., the severity of disease can diverge strongly among genotype positive variant carriers. Although this is partly explained by the varying effects of different classes of variants and the impact of multiple variants as described above, phenotypic diversity is also observed within families amongst individuals who share the same primary pathogenic variant. Environmental factors are likely to be responsible for some of this effect but it is believed that additional genetic factors may also play a role. These modifying variants can add to the genetic burden and exacerbate the phenotype or may act as protective factors to limit the expression of disease. Some initial studies have begun to explore the role of modifying variants in ICCs, particularly focusing on LQTS. Common single nucleotide polymorphisms (SNPs) in the nitric oxide synthase 1 adaptor protein (NOS1AP) gene have been shown through genome wide association studies to be associated with QTc interval (48 51). 10 Clinical Chemistry 63:1 (2017)

11 Diagnostic Sequencing for ICCs Reviews Genotype-phenotype studies on patients with LQTS have shown that SNPs in this gene are strongly associated with QTc interval ( 12 ms/allele) (52) and an increased risk of cardiac events (52, 53). Because these SNPs can have a cumulative effect, genetic risk scores have been developed to analyze the overall effect of the modifier variants on phenotype and clinical outcomes (52). Another study assessed how SNPs in the 3 untranslated region of the KCNQ1 gene can repress the expression of their associated KCNQ1 allele (54); patients with SNPs on the mutated KCNQ1 allele had shorter QTc intervals and fewer symptoms, whereas those with SNPs on the normal allele had longer QTc intervals and more symptoms. The cumulative burden of SNPs associated with small phenotypic effects has also been studied in FH. Using 12 SNPs identified by metaanalysis of several genome wide association studies that are associated with raised LDL-C, a weighted gene score was calculated and compared in FH variant-positive, FH variant-negative, and control subjects (55). The score was significantly higher in both FH cohorts compared to controls, suggesting that the raised LDL-C concentrations in variantnegative patients may have a polygenic etiology, and that the LDL-C-associated SNPs may also contribute to the genetic burden of variant-positive patients, explaining some of the variable penetrance observed in this condition. POPULATION SCREENING Genetic sequencing in ICCs has thus far been limited to testing probands and, where a causative variant has been identified, cascade screening of family members. With the ever decreasing costs associated with DNA sequencing, the possibility of universal genetic testing (as applied in new-born screening for cystic fibrosis) or targeted testing of particular at risk groups (e.g., athletes at risk of SCD) has begun to be explored. However, because the frequency of rare variation in many key cardiac genes has recently been revealed to be substantially higher than previously recognized (10, 12, 56), the prior probability that a variant detected in a phenotype-negative individual will be disease-causing is low, given the rarity of these genetic conditions. Additionally, the limited therapeutic options available for cardiomyopathies and arrhythmias limits the benefits that could be derived from identifying at risk individuals. One condition for which a universal or targeted screening approach could prove attractive is FH, which has the advantages of a relatively high genetic yield and effective clinical management with statin therapy. One of the strategies that has been proposed is reverse cascade screening, whereby infants with increased total cholesterol concentrations are genetically screened and, if a causative variant is identified, cascade screening applied to their parents (57). Another approach targets patients with a clinical history that suggests an increased risk of FH. A report of patients under age 60 years with a history of coronary artery disease found that 14.3% were estimated to have FH (58). However, a recent study that sequenced the 3 genes LDLR, apolipoprotein B (APOB), and proprotein convertase subtilisin/kexin type 9 (PCSK9) in individuals from several coronary artery disease, control and prospective cohorts identified FH mutations in 2% of participants with LDL-C 190 mg/dl, though this proportion could be increased by assessing structural variants, additional missense variants not classified as pathogenic, and polygenic SNPs. The presence of FH mutations did however confer a substantial increased risk of coronary artery disease across all LDL-C categories (59). If more widely adopted, approaches such as these have the potential to significantly increase the diagnosis of FH in the population, with prophylactic treatment offering substantial clinical and economic benefits (60). Conclusion Genetic testing in ICCs has made a rapid transition over the last 30 years from discovery research studies to evidence-based genetic sequencing of patients and their families. Recent advances in both sequencing technologies, which have driven down associated costs, and in the databases and analysis techniques used for variant interpretation, which improve the accuracy and analytical and clinical sensitivity of testing, should ensure that genetic testing becomes a standardized element of clinical management for many of these conditions. This promises to considerably increase the diagnosis of these diseases, particularly through cascade screening of patients families, and reduce their clinical burden through appropriate clinical management and treatment. However, to fully unleash the potential of genetic testing in cardiac disease, further progress will be required in our understanding of the genetic etiology of these conditions and the impact of variants on clinical outcomes in presymptomatic individuals and in those with existing disease. Efficient data sharing between clinical laboratories (using resources like ClinVar) and the application of novel research findings will be essential to improve the interpretation of rare variation and hence the accuracy and yield of diagnostic tests. Addressing the issue of missing heritability, particularly by exploring genetic variation in nonprotein coding parts of the genome, will be important to maximize the benefits of genetic sequencing for ICCs. Finally, investigating the basis of the phenotypic variability observed in many ICCs, and the role that modifier variants may have in explaining Clinical Chemistry 63:1 (2017) 11

12 Reviews this, may enable genetic testing to play a greater role in directing clinical care and treatment. Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. References Authors Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest: Employment or Leadership: None declared. Consultant or Advisory Role: None declared. Stock Ownership: None declared. Honoraria: S. Cook, Illumina with regard to TrueSight Cardio kit. Research Funding: None declared. Expert Testimony: None declared. Patents: None declared. 1. Green RC, Berg JS, Grody WW, Kalia SS, Korf BR, Martin CL, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med 2013;15: Authors/Task Force members, Elliott PM, Anastasakis A, Borger MA, Borggrefe M, Cecchi F, et al ESC Guidelines on diagnosis and management of hypertrophic cardiomyopathy: the Task Force for the Diagnosis and Management of Hypertrophic Cardiomyopathy of the European Society of Cardiology (ESC). Eur Heart J 2014;35: Jarcho JA, McKenna W, Pare JA, Solomon SD, Holcombe RF, Dickie S, et al. Mapping a gene for familial hypertrophic cardiomyopathy to chromosome 14q1. N Engl J Med 1989;321: Alfares AA, Kelly MA, McDermott G, Funke BH, Lebo MS, Baxter SB, et al. Results of clinical genetic testing of 2,912 probands with hypertrophic cardiomyopathy: expanded panels offer limited additional sensitivity. Genet Med 2015;17: Pinto YM, Elliott PM, Arbustini E, Adler Y, Anastasakis A, Böhm M, et al. Proposal for a revised definition of dilated cardiomyopathy, hypokinetic non-dilated cardiomyopathy, and its implications for clinical practice: a position statement of the ESC working group on myocardial and pericardial diseases. Eur Heart J 2016;37: Herman DS, Lam L, Taylor MRG, Wang L, Teekakirikul P, Christodoulou D, et al. Truncations of titin causing dilated cardiomyopathy. N Engl J Med 2012;366: Roberts AM, Ware JS, Herman DS, Schafer S, Baksi J, Bick AG, et al. Integrated allelic, transcriptional, and phenomic dissection of the cardiac effects of titin truncations in health and disease. Sci Transl Med 2015;7: 270ra6. 8. Pugh TJ, Kelly MA, Gowrisankar S, Hynes E, Seidman MA, Baxter SM, et al. The landscape of genetic variation in dilated cardiomyopathy as surveyed by clinical DNA sequencing. Genet Med 2014;16: Marcus FI, Edson S, Towbin JA. Genetics of arrhythmogenic right ventricular cardiomyopathy: a practical guide for physicians. J Am Coll Cardiol 2013;61: Andreasen C, Nielsen JB, Refsgaard L, Holst AG, Christensen AH, Andreasen L, et al. New population-based exome data are questioning the pathogenicity of previously cardiomyopathy-associated genetic variants. Eur J Hum Genet 2013;21: Kapplinger JD, Landstrom AP, Salisbury BA, Callis TE, Pollevick GD, Tester DJ, et al. Distinguishing arrhythmogenic right ventricular cardiomyopathy/dysplasiaassociated mutations from background genetic noise. J Am Coll Cardiol 2011;57: Walsh R, Thomson KL, Ware JS, Funke BH, Woodley J, McGuire KJ, et al. Reassessment of Mendelian gene pathogenicity using 7,855 cardiomyopathy cases and 60,706 reference samples. Genet Med [Epub ahead of print 2016 Aug 17]. 13. Napolitano C, Priori SG, Schwartz PJ, Bloise R, Ronchetti E, Nastoli J, et al. Genetic testing in the long QT syndrome: development and validation of an efficient approach to genotyping in clinical practice. JAMA 2005;294: Kapa S, Tester DJ, Salisbury BA, Harris-Kerr C, Pungliya MS, Alders M, et al. Genetic testing for long-qt syndrome: distinguishing pathogenic mutations from benign variants. Circulation 2009;120: Kapplinger JD, Tester DJ, Salisbury BA, Carr JL, Harris- Kerr C, Pollevick GD, et al. Spectrum and prevalence of mutations from the first 2,500 consecutive unrelated patients referred for the FAMILION long QT syndrome genetic test. Heart Rhythm 2009;6: Kapplinger JD, Tester DJ, Alders M, Benito B, Berthet M, Brugada J, et al. An international compendium of mutations in the SCN5A-encoded cardiac sodium channel in patients referred for Brugada syndrome genetic testing. Heart Rhythm 2010;7: Priori SG, Napolitano C, Tiso N, Memmi M, Vignati G, Bloise R, et al. Mutations in the cardiac ryanodine receptor gene (hryr2) underlie catecholaminergic polymorphic ventricular tachycardia. Circulation 2001;103: Laitinen PJ, Brown KM, Piippo K, Swan H, Devaney JM, Brahmbhatt B, et al. Mutations of the cardiac ryanodine receptor (RyR2) gene in familial polymorphic ventricular tachycardia. Circulation 2001;103: Medeiros-Domingo A, Bhuiyan ZA, Tester DJ, Hofman N, Bikker H, van Tintelen JP, et al. The RYR2-encoded ryanodine receptor/calcium release channel in patients diagnosed previously with either catecholaminergic polymorphic ventricular tachycardia or genotype negative, exercise-induced long QT syndrome: a comprehensive open reading frame muta. J Am Coll Cardiol 2009;54: Nordestgaard BG, Chapman MJ, Humphries SE, Ginsberg HN, Masana L, Descamps OS, et al. Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society. Eur Heart J 2013;34: a. 21. Priori SG, Wilde AA, Horie M, Cho Y, Behr ER, Berul C, et al. HRS/EHRA/APHRS expert consensus statement on the diagnosis and management of patients with inherited primary arrhythmia syndromes: document endorsed by HRS, EHRA, and APHRS in May 2013 and by ACCF, AHA, PACES, and AEPC in June Heart Rhythm 2013;10: Priori SG, Napolitano C, Schwartz PJ, Grillo M, Bloise R, Ronchetti E, et al. Association of long QT syndrome loci and cardiac events among patients treated with betablockers. JAMA 2004;292: Barsheshet A, Goldenberg I, O-Uchi J, Moss AJ, Jons C, Shimizu W, et al. Mutations in cytoplasmic loops of the KCNQ1 channel and the risk of life-threatening events: implications for mutation-specific response to -blocker therapy in type 1 long-qt syndrome. Circulation 2012;125: Barsheshet A, Dotsenko O, Goldenberg I. Congenital long QT syndromes: prevalence, pathophysiology and management. Paediatr Drugs 2014;16: Westenskow P, Splawski I, Timothy KW, Keating MT, Sanguinetti MC. Compound mutations: a common cause of severe long-qt syndrome. Circulation 2004; 109: Richard P, Charron P, Carrier L, Ledeuil C, Cheav T, Pichereau C, et al. Hypertrophic cardiomyopathy: distribution of disease genes, spectrum of mutations, and implications for a molecular diagnosis strategy. Circulation 2003;107: Ingles J, Doolan A, Chiu C, Seidman J, Seidman C, Semsarian C. Compound and double mutations in patients with hypertrophic cardiomyopathy: implications for genetic testing and counselling. J Med Genet 2005;42:e Girolami F, Ho CY, Semsarian C, Baldi M, Will ML, Baldini K, et al. Clinical features and outcome of hypertrophic cardiomyopathy associated with triple sarcomere protein gene mutations. J Am Coll Cardiol 2010;55: Olivotto I, Girolami F, Ackerman MJ, Nistri S, Bos JM, Zachara E, et al. Myofilament protein gene mutation screening and outcome of patients with hypertrophic cardiomyopathy. Mayo Clin Proc 2008;83: Li Q, Gruner C, Chan RH, Care M, Siminovitch K, Williams L, et al. Genotype-positive status in patients with hypertrophic cardiomyopathy is associated with higher rates of heart failure events. Circ Cardiovasc Genet 2014;7: Coppini R, Ho CY, Ashley E, Day S, Ferrantini C, Girolami F, et al. Clinical phenotype and outcome of hypertrophic cardiomyopathy associated with thin-filament gene mutations. J Am Coll Cardiol 2014;64: van Rijsingen IA, Arbustini E, Elliott PM, Mogensen J, Hermans-van Ast JF, van der Kooi AJ, et al. Risk factors for malignant ventricular arrhythmias in lamin a/c mutation carriers a European cohort study. J Am Coll Cardiol 2012;59: Wordsworth S, Leal J, Blair E, Legood R, Thomson K, Seller A, et al. DNA testing for hypertrophic cardiomyopathy: a cost-effectiveness model. Eur Heart J 2010;31: Ingles J, McGaughran J, Scuffham PA, Atherton J, Semsarian C. A cost-effectiveness model of genetic testing 12 Clinical Chemistry 63:1 (2017)