Gene-Targeted Analysis of Clinically Diagnosed Long QT Russian Families

Gene-Targeted Analysis of Clinically Diagnosed Long QT Russian Families Paolo Enrico Maltese,1 PhD, Nina Orlova,2 MD, Eugenia Krasikova,3 MD, Elena Emelyanchik,3 PhD, Anna Cheremisina,4 MD, Alina Kuscaeva,5 MD, Alla Salmina,2 PhD, Roberta Miotto,1 BSc, Alice Bonizzato,1 MLT, Giulia Guerri,1 PhD, Monia Zuntini,1 BSc, Svetlana Nicoulina,2 PhD, and Matteo Bertelli,1 PhD

Summary Long QT syndrome (LQTS) has great genetic heterogeneity: more than 500 mutations have been described in several genes. Despite many advances, a genetic diagnosis still cannot be established in 25-30% of patients. The aim of the present study was to perform genetic evaluation in 9 Russian families with LQTS; here we report the results of 4 positive probands and their relatives (a total of 16 individuals). All subjects underwent clinical examination, 12-lead ECG, and Holter monitoring. Genetic analysis of the 14 genes mainly involved in LQTS was performed using a next-generation sequencing approach. We identified two new mutations (KCNQ1 gene) and 6 known mutations (AKAP9, ANK2, KCNE1 and KCNJ2 genes) in 4 out of 9 probands, some of which have already been described in association with LQTS. Segre- gation studies suggest a possible causative role for KCNQ1 p.(Leu342Pro), AKAP9 p.(Arg1609Lys), KCNE1 p.(As- p85Asn), and KCNJ2 p.(Arg82Gln) variations. Our study confirmed the high genetic heterogeneity of this disease and highlights the difficulties to reveal clear pathogenic genotypes also in large pedigrees. To the best of our knowledge, this is the first genetic study of LQTS patients from Russian families. (Int Heart J 2017; 58: 81-87)

Key words: LQTS, Cardiology, Channels, Channelopathy, Next generation sequencing

common cause of severe ventricular arrhythmia is the cal features, and prognosis may vary significantly in individu- syndrome of long QT interval, in which violation of als with different genotypes,6) depending on homozygous/het- A potassium or sodium channels due to genetic muta- erozygous mutations, combinations of different mutations and tion slows repolarization of the myocardium, lengthening the polymorphisms, as well as environment, which may influence QT interval on the electrocardiogram. Symptoms occur sud- the clinical manifestations of existing genotypes. denly and often without warning.1) Nine Russian families with LQTS were recruited for the Congenital long QT syndrome (henceforth LQTS) is a genetic evaluation; genetically determined disease occurring in one case per 3000- here we present the results of clinical and genetic charac- 5000 members of the general population.2) The disease is he- terization of 4 Russian probands with LQTS in which we reditary in approximately 85% of cases, while about 15% of found variations probably related to the phenotype. We identi- cases are a result of new spontaneous mutations.1) LQTS is fied two new genetic variations and 6 known mutations, some usually inherited in an autosomal dominant manner;3) most of which have already been described in association with mutations have been found in genes encoding the alpha or beta LQTS. We also performed a family segregation study and new subunits of plasma membrane ion channels crucial for heart mutations were further characterized for their putative patho- beat regulation. Although more than 500 mutations have been genic potential by in-silico evaluation. described in 14 genes (Table I), a genetic diagnosis cannot be established in 25-30% of patients. Presentation of the disease is mainly monogenic and pen- Methods etrance, and ranges from 25 to 90%.4) Approximately 10% of patients with LQTS bear at least two mutations associated with Study subjects: Nine probands and their families were exam- the condition;5) polygenic or composite varieties usually have a ined in clinics of Krasnoyarsk city. All patients were enrolled more severe phenotype. after genetic counseling to explain the risks and benefits of ge- The length of the QT interval, ECG characteristics, clini- netic testing. They were invited to sign specific consent forms

From the 1 MAGI Non-Profit Human Medical Genetics Institute, Rovereto, Italy, Departments of 2 Internal Diseases I and 3 Pediatrics, Krasnoyarsk State Medical Uni- versity, 4 Federal Centre of Cardiovascular Surgery in Krasnoyarsk, and 5 Department of Internal Diseases II, Krasnoyarsk State Medical University, Krasnoyarsk, Russia. This work was supported by grant from the Autonoma Provincia di Trento (grant number S503/2016/238507). Address for correspondence: Paolo Enrico Maltese, PhD, MAGI Non-Profit Human Medical Genetics Institute, Via Delle Maioliche 57/D, 38068 Rovereto, TN, Italy. E-mail: [email protected] Received for publication May 25, 2016. Revised and accepted August 1, 2016. Released in advance online on J-STAGE December 21, 2016. All rights reserved by the International Heart Journal Association. 81 Int Heart J 82 MALTESE, ET AL January 2017

Table I. Genes Involved in LQTS Type Gene RefSeq number Gene Name Prevalence (OMIM ID:) (OMIM ID:) LQTS1 KCNQ1 Potassium voltage-gated channel subfamily Q NM_000218, NM_181798 40-50% (192500) (607542) member 1

LQTS2 KCNH2 NM_172056, NM_172057, NM_001204798, Potassium voltage-gated channel subfamily H 35-45% (613688) (152427) NM_000238 member 2

LQTS3 SCN5A NM_001099405, NM_001160161, NM_001160160, Sodium voltage-gated channel alpha subunit 5 2-8% (603830) (600163) NM_001099404, NM_000335, NM_198056

LQTS4 ANK2 NM_001148, NM_001127493, NM_020977 Ankyrin 2 < 1% (600919) (106410)

NM_000219, NM_001127669, NM_001127670, LQTS5 KCNE1 Potassium voltage-gated channel subfamily E NM_001127668, NM_001270405, NM_001270403, < 1% (613695) (176261) regulatory subunit 1 NM_001270402, NM_001270404

LQTS6 KCNE2 Potassium voltage-gated channel subfamily E NM_172201 < 1% (613693) (603796) regulatory subunit 2

LQTS7 KCNJ2 Potassium voltage-gated channel subfamily J NM_000891 < 1% (170390) (600681) member 2

NM_001167625, NM_001129844, NM_001129834, NM_001129838, NM_000719, NM_001167624, NM_001129839, NM_001167623, NM_001129841, LQTS8 CACNA1C NM_001129840, NM_001129827, NM_001129836, Calcium voltage-gated channel subunit alpha 1 < 1% (601005) (114205) NM_001129846, NM_001129835, NM_001129830, C NM_001129843, NM_199460, NM_001129833, NM_001129831, NM_001129842, NM_001129837, NM_001129829, NM_001129832

LQTS9 CAV3 NM_001234, NM_033337 Caveolin 3 < 1% (611818) (601253)

LQTS10 SCN4B NM_174934, NM_001142349, NM_001142348 Sodium voltage-gated channel beta subunit 4 < 0.1% (611819) (608256)

LQTS11 AKAP9 NM_005751, NM_147185 A-kinase anchoring protein 9 < 0.1% (611820) (604001)

LQTS12 SNTA1 NM_003098 Syntrophin alpha 1 < 0.2% (612955) (601017)

LQTS13 KCNJ5 Potassium voltage-gated channel subfamily J NM_000890 < 0.1% (613485) (600734) member 5

LQTS14 CALM1 NM_006888 Calmodulin 1 Very Rare (616247) (114180)

to make their clinical and genetic data available for research the QT interval using the Bazetta formula against the standards and publication. of the European Agency for the Evaluation of Medical Prod- The work described in this article was carried out in ac- ucts. T-wave morphology was also observed as an indicator of cordance with The Declaration of Helsinki and was approved syndrome type. The clinical probability of disease was scored by the Krasnoyarsk State Medical University Ethics Commit- in points on the basis of Schwartz criteria.7) tee (protocol No 49/2012). Demographic details and information on personal and Probands and relatives underwent clinical examination, family medical history were recorded and used to describe 12-lead ECG, Holter monitoring, and blood sampling at the families, to determine the inheritance pattern, and to draw ped- Federal Cardiological and Vascular Center (Krasnoyarsk) and igrees. the Department of Internal Medicine I and Department of Molecular genetic analysis using NGS: Genetic testing was Pediatrics, Krasnoyarsk State Medical University “V.F.Voino- performed at MAGI Laboratory (MAGI Non-Profit Human Yasenetsky”. Medical Genetics Institute, Rovereto, Italy). A diagnosis of LQTS involved analyzing the length of Targeted resequencing for 14 genes known to be involved Vol 58 No 1 GENETICS OF LONG QT SYNDROME IN RUSSIAN FAMILIES 83 in LQTS (Table I) was performed by NGS using an Illumina dbSNP137, EVS and 1000 Genome Project with allelic fre- commercial kit “TruSight One sequencing panel” on the Illu- quency not more than 0.03. mina MiSeq platform (http://www.illumina.com/ products/ The following criteria were applied to evaluate the patho- trusight-one-sequencing-panel.ilmn). genic nature of the variant set selected: 1) known mutation; In-solution target enrichment was performed according to nonsense, frameshift, essential splice site (affecting conserved the manufacturer’s protocol for TruSight One sequencing panel consensus motif), start and stop loss variants were considered kit (Illumina) starting from 5 ng of genomic DNA for each as most likely to be disease causing; 2) missense variants hav- sample. ing an allelic frequency less than 0.01 in dbSNP138 and with Each 3-plex sample library pool was sequenced using a predicted deleterious effects were considered potentially path- 150 bp paired-end reads protocol on the MiSeq sequencer (Il- ogenic variants. lumina, San Diego, CA) according to the manufacturer’s pro- Sanger validation and segregation analysis: All mutations tocol. were validated by Sanger sequencing in probands and then the Data analysis: Raw read data in fastq format, generated by genotype-phenotype correlation of each mutation was evaluat- Ilumina MiSeq reporter software (version 2.5), were analyzed ed by family segregation study using a Beckman Coulter to generate the final set of sequence variants using an in-house CEQTM 8000 Genetic Analysis System (Beckman Coulter, Mi- pipeline with the following modules: mapping, duplicate read lano, Italy). removal, indel realignment, quality calibration, coverage anal- Electropherograms of amplified fragments were analyzed ysis, variant calling, and annotation. In brief, the sequencing using ChromasPro 1.5 (Technelysium Pty. Ltd., Australia) and reads were mapped to the genome build hg19 using the Bur- Sequencher 5.0 (Gene Codes®; Ann Arbor, MI, USA) software rows-Wheeler Aligner (BWA)8) with default settings (version and compared to GenBank reference sequences with the Basic 0.7.5a-r405). Next, duplicate fragments were marked and Local Alignment Search Tool (BLAST; http://blast.ncbi.nlm. eliminated with the MarkDuplicates GATK Tool (version v2.5- nih.gov). 2-gf57256b).9,10) The BAM alignment files generated were re- fined by local realignment and base quality score recalibration using RealignerTargetCreator and IndelRealigner GATK Results Tools. Statistical and coverage analysis of final BAM files were performed using SAMTools (Sequence Alignment/Map Genetic analysis was performed on 9 unrelated probands Tools) and BEDTools.11) Reads aligned to the designed target with a clinical diagnosis of LQTS. In this study, we performed regions (coding exons and 15 bp flanking of gene-disease sub- NGS analysis for the 14 known LQTS-related genes (Table I) panel, Table I) were collected for variant calling and subse- using a TruSight One sequencing panel (Illumina). The aver- quent analysis. The following data per sample was generated age number of mappable reads per sample was 10 M. On aver- by coverage analysis: average read depth, low coverage target age, 96% of the target bases of gene-disease subpanels were regions (< 10X); % of target bases with coverage ≥ 10X. Se- covered at least 10×, with a mean coverage of 117× per sam- quence variant calling was performed using three SNP and ple. With this approach we identified 8 heterozygous muta- genotype calling tools: GATK UnifiedGenotyper, Varscan tions, two of which are novel mutations, in 4 affected subjects; (version v2.3),12) and BCFTools of SAMTools (version 0.1.19- we discovered mutations in KCNQ1 (NM_000218), KCNH2 44428cd). The output data from the three variant callers was (NM_000238), ANK2 (NM_001148), KCNE1 (NM_000219), united/pooled and converted into a standard vcf file using a KCNJ2 (NM_000891), and AKAP9 (NM_005751) genes (Ta- custom script. Called variants were annotated using Annovar ble II). Overall, a mutation was found in 44% (4/9) of patients. software.13) Segregation study was then performed by extending the genet- Target region coverage less than 10 reads was additional- ic analysis to a total of 16 relatives. ly analyzed by Sanger sequencing. Features of genetic variations and results from pedigree Variant filtering and prioritization: To identify mutations pre- analysis of patients are shown in Table II and the Figure. viously reported as pathogenic, we consulted the Human Gene Sanger sequencing chromatograms are shown in the Supple- Mutation Database (HGMD) (http://www.biobase-internation- mental Figure. al.com/product/hgmd); allele frequencies were checked in the Family 1: The female proband, 20 years of age with onset at ExAC database (exac.broadinstitute.org/) and Exome Variant 16 years, complained of syncope and premature ventricular Server (EVS) database (http://evs.gs.washington.edu/EVS/). beats. Examination showed a heart-rate-corrected QT interval To exclude polymorphisms, we also searched the public data- (QTc) > 450 msec and QTcmax of 530 msec. The patient un- base dbSNP (http://www.ncbi.nlm.nih.gov/SNP/). New nucle- derwent implantation of a cardioverter-defibrillator (ICD). The otide variations were assessed for pathogenicity using the sister is healthy, without clinical manifestations and normal PolyPhen 2 algorithm (Polymorphism Phenotyping v2; http:// QTc. The mother has no symptoms of the disease, but a 12- genetics.bwh.harvard.edu/pph2)14) considering the HumVar- lead ECG showed QTc > 460 msec and scored an intermediate trained model, SIFT algorithm (Sorting Intolerant From Toler- probability of LQTS by the Schwartz-Moss system. ant; http://sift.bii.a-star.edu.sg/)15) and MutationTaster (http:// Family 2: The proband first suffered from syncope at 7 years www.mutationtaster.org).16) When possible, wild-type amino of age. A 12-lead ECG showed a QTc > 450 msec and QTc- acid properties were compared with the variations (http://www. max of 489 msec. Examination of the parents showed that the russelllab.org/aas/aas.html).17) mother is healthy, has no symptoms and a normal QTc, where- Variants were selected for subsequent study on the basis as the father has no symptoms of the disease, but a 12-lead of the following criteria: 1) variants in target region; 2) previ- ECG showed QTc > 450 msec. ously reported in HGMD and HumsVar database; 3) present in Family 3: The 14-year-old female proband sought medical Int Heart J 84 MALTESE, ET AL January 2017

Table II. Features of Genetic Variations Found in Four LQTS Families Nucleotide Amino acid Mutation dbSNP MAF MAF MAF Gene Genotype SIFT Polyphen change change Taster accession N° dbSNP ExAC EVS KCNQ1 c.1025T>C p.(Leu342Pro) Het D PrD DC Novel NA NA NA KCNQ1 c.1011C>G p.(Ile337Met) Het D PrD DC Novel NA NA NA KCNH2 c.526C>T p.(Arg176Trp) Het D PoD P rs36210422 NA NA NA ANK2 c.9854T>C p.(Ile3285Thr) Het D PrD DC rs36210417 0.004 0.0082 0.0068 KCNE1 c.253G>A p.(Asp85Asn) Het DBP rs1805128 0.38 0.0091 0.0087 KCNJ2 c.245G>A p.(Arg82Gln) Het D PrD DC rs199473653 NA NA NA AKAP9 c.139C>T p.(His47Tyr) Het TBP rs35669569 0.0022 0.0081 0.0096 AKAP9 c.4826G>A p.(Arg1609Lys) Het T PoD P rs148146011 0.0006 0.0004 0.0001 In-silico evaluation of the 8 genetic variations identified in 4 probands and their relatives. Het indicates genotype heterozygous; WT, wild-type; T, SIFT score system: tolerated; D, SIFT score system: damaging; B, Polyphen 2 score system: benign; PoD, Polyphen 2 score system: possibly damaging less confident prediction, PrD, Polyphen 2 score system: probably damaging more confident prediction; P, Mutation Taster score system: polymorphism; DC, Mutation Taster score system: disease-causing; MAF, Mutation Taster score system: minor allele frequency; and NA, not available.

Figure. Family pedigree illustrating cosegregation of gene mutations and LQTS phenotypes. * Documented clinical evaluation; E+ and E-, positive and negative to genetic test respectively; ● Obligate carrier [AKAP9 p.(His47Tyr)]; | Asymptomatic/presymptomatic carrier. Schwartz diagnostic scoring system: ≤ 1 point, low probability of LQTS; 2 to 3 points, intermediate probability of LQTS; ≥ 4 points, high probability of LQTS.

advice about syncope and premature ventricular contractions (> 4 points according to the Schwartz-Moss scoring system) of during physical activity. A 12-lead ECG showed a QTc > 470 developing the disease. The brother of the mother (II:2) is in msec and QTcmax of 515 msec. Her brother also has syncope the same situation, with a QTcmax of 480 msec. The proband’s and a QTcmax of 474 msec; he too was diagnosed with LQTS. cousin (III:2) is at intermediate risk with a QTcmax of 469 The proband’s mother (II:3) has no symptoms of the disease, msec. The grandmother (I:2) and grandfather (I:1) of the but her QTcmax was 513 msec and she is therefore at high risk proband from the mother’s side were also studied. The grand- Vol 58 No 1 GENETICS OF LONG QT SYNDROME IN RUSSIAN FAMILIES 85 mother is healthy with a normal QTc. The grandfather has out a pathogenic role also for this rare missense variation. syncope, tachycardia, and QTc > 450 msec. The proband has been on non-selective beta-blocker ther- Family 4: A female infant of healthy parents first manifested apy for 1 year. This controlled the syncope, but myopia mani- syncope at 10 years of age. A 12-lead ECG showed a QTc > fested after beginning therapy and deteriorated to -4 D. Thera- 450 msec and QTcmax of 475 msec, ie, LQTS. Both parents py was therefore suspended. The patient is currently stable, are healthy and without symptoms; a 12-lead ECG showed without syncope. normal QTc. In family 3, the proband bears two mutations in genes encoding potassium channel α and β-subunits (KCNE1 and KCNH2, respectively); mutations in these genes are associated Discussion with reduced outgoing K+ current in phase 3 of the action potential and are expected to express a severe phenotype. The ef- Long QT Syndrome is considered a very heterogeneous fect of this combination was shown to cause a dominant-nega- hereditary condition. In the present genetic study, in 9 LQTS tive effect that reduced IKr by Nof, et al and Yoshikane, et patients we investigated the presence of variations in 14 genes al,28,29) who however considered the p.(Asp85Asn) variation in known to be associated with the condition. In 4 out of 9 pro KCNE1 combined with different KCNH2 mutations. This situ- bands, we identified two new genetic variants in the KCNQ1 ation could explain the proband’s LQTS phenotype but fails to gene and another 6 variants already reported in AKAP9, explain the overt disease in his brother. KCNJ2, KCNH2, KCNE1, and ANK2 genes (Table II). A third mutation in the ANK2 gene was found in the In family 1 we found p.(Arg82Gln) in the KCNJ2 gene proband, his affected brother, and the healthy father. Mutations that encodes the potassium channel Kir2.1. Mutations in this in this gene have been proposed as modulators of human ar- gene are involved in Andersen-Tawil Syndrome (ATS)20) which rhythmias, but p.(Ile3285Thr) is described as “likely benign” is characterized by different phenotypes involving multiple in dbSNP and therefore not associated with the condition. Ac- body districts and organs. In our study, the proband only had cordingly, results from this family seem to reject any possible severe heart abnormalities that included cardiac arrhythmias. association of this variant with the disease. The proband’s mother carries the same mutation but she only Variable expression was also found in symptom-free fam- showed a subclinical form of the disease with an intermediate ily members in whom 12-lead ECG and Holter monitoring probability of LQTS according to the Schwartz-Moss system. showed a subclinical phenotype. The incomplete penetrance of this mutation was also noted by Although it is described as having a role as LQTS modi- Davies, et al 21) in the UK, but in contrast to our findings they fier,28-30) we showed a perfect segregation in our family tree for found ATS phenotype without cardiac involvement in their the p.(Asp85Asn) variation in the KCNE1 gene. proband. It should be noted that these differences may be ex- Nevertheless, we cannot exclude the presence of impor- plained by the different populations studied in our report and tant variations in other genes not evaluated in this study. previous reports. They also demonstrated by functional study Wolff-Parkinson-White syndrome (syncope + ECG with that the mutation resulted in the expression of nonfunctional long QT-interval) was suspected in the proband. As the first channels. line treatment for this syndrome, radiofrequency catheter abla- All family members were found to harbor a second mutation was performed 31) but syncope and the ECG-pattern per- tion in the KCNH2 gene (p.Arg176Trp), that we also found in sisted. A defibrillator was implanted with good clinical effect. two members of family 3 and that had already been associated Other members of the family with LQT respond well to beta- with LQTS 2 in the Finnish population.22) Contrasting evidence blockers. of the pathogenicity of this mutation can be found in the litera- The proband of family 4 showed two new variations of ture and many authors report clinical variability and low pene- uncertain significance (VUS), namely p.(Leu342Pro) and p. trance.23-25) Since we identified the p.(Arg176Trp) in two (Ile337Met) in the KCNQ1 gene and the “benign” p.(Hy- healthy individuals from two different families, our results s47Tyr) variant in the AKAP9 gene. Mutations in KCNQ1 are seem to support the hypothesis that this variation is likely be- associated with the common LQTS type 1. Both missense vari- nign. The proband is on beta-blocker therapy with good clini- ants were predicted as potentially pathogenic variants by 3/3 cal results. Her younger sister and mother (both heterozygotes pathogenicity prediction tools (Table II). However, both VUS and symptom-free) have been invited for further investigation. were also found in the proband’s mother, who did not have any In family 2, only one variation, p.(Arg1609Lys) in the clinical sign of the disease. This situation cannot be explained AKAP9 gene, was found in the proband and her father. Al- by the concomitant presence of the p.(Hys47Tyr) variant in the though listed in dbSNP (rs148146011, MAF 0.0006), at the AKAP9 gene, since the protein is known to possess the N-ter- time of this report no one has yet shown any association with minus binding sites for KCNQ1 located on its residues 29- disease phenotypes. Although the condition was subclinical in 46;26) residue 47 is not conserved among species, thus confirm- the father and only detected by ECG, segregation study in our ing that it is not important for the protein’s function. The family suggests that it can cause LQTS. AKAP9 was shown to mutation p.(Leu342Phe) in the same codon has already been possess two binding sites for KCNQ1 in its C-terminus (resi- described in LQTS patients and in a case of syncope before 10 dues 1574-1643); of note, the p.(Arg1609Lys) variant is locat- years of age.18,19) The importance of this new VUS is suggested ed in this domain.26) Only one mutation in this site, namely by the fact that neighbour codon 341 in the same protein re- p.(Ser1570Leu), has been reported to be associated with LQTS gion, ie, transmembrane segment S6, is commonly mutated in and its pathogenicity was supported by functional studies con- LQTS 1 patients and mutant patients have a high risk of cardi- firming that the variant acts by reducing the interaction with ac events; in a study of 244 patients with KCNQ1-A341V gen- KCNQ1.27) Taken together, these observations seem to point otype, about 70% experienced at least one cardiac event before Int Heart J 86 MALTESE, ET AL January 2017

40 years of age, 30% of whom had a fatal or near-fatal event;32) DNA sequencing data. Genome Res 2010; 20: 1297-303. the other 30% were asymptomatic in the first 4 decades, like 10. DePristo M, Banks E, Poplin R, et al. A framework for variation the proband’s mother who is only 27. In her case, we cannot discovery and genotyping using next-generation DNA sequencing data. Nat Genet 2011; 43: 491-8. exclude a first cardiac event before the age of 40 years. The 11. Quinlan AR, Hall IM. BEDTools: a flexible suite of utilities for proband is on beta-blocker therapy which controls LQT and is comparing genomic features Bioinformatics 2010; 26: 841-2. well tolerated. 12. Koboldt DC, Chen K, Wylie T, et al. VarScan: variant detection in Conclusions: Several studies associate LQTS with genetics massively parallel sequencing of individual and pooled samples. but these associations are mainly demonstrated in a population- Bioinformatics 2009; 25: 2283-5. dependent fashion; to our knowledge this is the first study re- 13. Wang K, Li M, Hakonarson H. ANNOVAR: Functional annota- porting clinical and genetic evaluation of LQTS patients from tion of genetic variants from next-generation sequencing data. Nu- cleic Acids Res 2010; 38: e164. Russian families. It is also one of the few reports considering 14. Adzhubei IA, Schmidt S, Peshkin L, et al. A method and server all known genes associated with the disease until now. Despite for predicting damaging missense mutations. Nat Methods 2010; this, the results remain controversial and further studies are 7: 248-9. therefore required to fully understand the underlying causative 15. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non- mechanisms. synonymous variants on protein function using the SIFT algo- Interestingly, in 3 out of 4 families, probands were found rithm. Nat Protoc 2009; 4: 1073-81. 16. Schwarz JM, Cooper DN, Schuelke M, Seelow D. Mutation- positive to at least two mutations in different genes. We cannot Taster2: mutation prediction for the deep-sequencing age. Nat say anything about the links between digenic mutations and Methods 2014; 11: 361-2. more severe phenotypes because of our small statistical sam- 17. Betts MJ, Russell RB. Amino acid properties and consequences of ple, but some reports in the literature highlight this possibili- substitutions. In: Barnes MR, Gray IC eds. Bioinformatics for Ge- ty.33) neticists. Chichester, UK: John Wiley & Sons, Ltd; 2003: 289- Approximately 70-75% of LQTS patients have mutations 316. in one of the 3 major LQTS-causing genes (KCNQ1, KCNH2 18. Donger C, Denjoy I, Berthet M, et al. KVLQT1 C-terminal missense mutation causes a forme fruste long-QT syndrome. Circula- and SCN5A), the result of which is that a genetic diagnosis 34) tion 1997; 96: 2778-81. cannot be established in 25-30% of patients. In our study, the 19. Kapplinger JD, Tester DJ, Salisbury BA, et al. Spectrum and prev- mutation detection rate was 44% (4/9 probands) even with ap- alence of mutations from the first 2,500 consecutive unrelated pa- plication of NGS for 14 LQTS-related genes; this is unlikely to tients referred for the FAMILION long QT syndrome genetic test. reflect the real genetic variant distribution in Russian patients Heart Rhythm 2009; 6: 1297-303. because of the small number of subjects. 20. Andelfinger G, Tapper AR, Welch RC, Vanoye CG, George AL Jr, Benson DW. KCNJ2 mutation results in Andersen syndrome with In conclusion, our study confirms the genetic heterogene- sex-specific cardiac and skeletal muscle phenotypes. Am J Hum ity of this disease, while the presence of 5 patients negative to Genet 2002; 71: 663-8. genetic testing suggests that other unknown genes are involved. 21. Davies NP, Imbrici P, Fialho D, et al. Andersen-Tawil syndrome: This next generation sequencing approach confirmed to new potassium channel mutations and possible phenotypic varia- be of invaluable help to study diseases characterized by tion. Neurology 2005; 65: 1083-9. marked genetic heterogeneity as LQTS, though the patho- 22. Swan H, Viitasalo M, Piippo K, Laitinen P, Kontula K, Toivonen L. genicity of mutations could be difficult to demonstrate in large Sinus node function and ventricular repolarization during exercise stress test in long QT syndrome patients with KvLQT1 and HERG pedigrees also. potassium channel defects. J Am Coll Cardiol 1999; 34: 823-9. 23. Fodstad H, Swan H, Laitinen P, et al. Four potassium channel mutations account for 73% of the genetic spectrum underlying long- References QT syndrome (LQTS) and provide evidence for a strong founder effect in Finland. Ann Med 2004; 36: 53-63. 1. Moss AJ, Schwartz PJ. 25th anniversary of the International Long- 24. Fodstad H, Bendahhou S, Rougier JS, et al. Molecular characteri- QT Syndrome Registry: an ongoing quest to uncover the secrets zation of two founder mutations causing long QT syndrome and of long-QT syndrome. Circulation 2005; 111: 1199-201. (Review) identification of compound heterozygous patients. Ann Med 2006; 2. Moss AJ. Long QT Syndrome. JAMA 2003; 289: 2041-4. (Re- 38: 294-304. view) 25. Marjamaa A, Salomaa V, Newton-Cheh C, et al. High prevalence 3. Bokil NJ, Baisden JM, Radford DJ, Summers KM. Molecular ge- of four long QT syndrome founder mutations in the Finnish popu- netics of long QT syndrome. Mol Genet Metab 2010; 101: 1-8. lation. Ann Med 2009; 41: 234-40. (Review) 26. Li Y, Chen L, Kass RS, Dessauer CW. The A-kinase anchoring 4. Medeiros-Domingo A, Iturralde-Torres P, Ackerman MJ. Clinical protein Yotiao facilitates complex formation between adenylyl cy- and genetic characteristics of long QT syndrome. Rev Esp Cardiol clase type 9 and the IKs potassium channel in heart. J Biol Chem 2007; 60: 739-52. (Review, Spanish) 2012; 287: 29815-24. 5. Goldenberg I, Zareba W, Moss AJ. Long QT Syndrome. Curr 27. Chen L, Marquardt ML, Tester DJ, Sampson KJ, Ackerman MJ, Probl Cardiol 2008; 33: 629-94. (Review) Kass RS. Mutation of an A-kinase-anchoring protein causes long- 6. Moss AJ, Goldenberg I. Importance of knowing the genotype and QT syndrome. Proc Natl Acad Sci U S A 2007; 104: 20990-5. the specific mutation when managing patients with long QT syn- 28. Nof E, Barajas-Martinez H, Eldar M, et al. LQT5 masquerading drome. Circ Arrhythm Electrophysiol 2008; 1: 213-26. as LQT2: a dominant negative effect of KCNE1-D85N rare poly- 7. Schwartz PJ, Moss AJ, Vincent GM, Crampton RS. Diagnostic morphism on KCNH2 current. Europace 2011; 13: 1478-83. criteria for the long QT syndrome. An update. Circulation 1993; 29. Yoshikane Y, Yoshinaga M, Hamamoto K, Hirose S. A case of 88: 782-4. (Review) long QT syndrome with triple gene abnormalities: digenic muta- 8. Li H. and Durbin R. Fast and accurate short read alignment with tions in KCNH2 and SCN5A and gene variant in KCNE1. Heart Burrows-Wheeler Transform. Bioinformatics 2009; 25: 1754-60. Rhythm 2013; 10: 600-3. 9. McKenna A, Hanna M, Banks E, et al. The Genome Analysis 30. Nishio Y, Makiyama T, Itoh H, et al. D85N, a KCNE1 polymor- Toolkit: a MapReduce framework for analyzing next-generation phism, is a disease-causing gene variant in long QT syndrome. J Vol 58 No 1 GENETICS OF LONG QT SYNDROME IN RUSSIAN FAMILIES 87

Am Coll Cardiol 2009; 54: 812-9. Chin Med J (Engl) 2014; 127: 1482-6. 31. Cakmak N, Cakmak M, Akyol A, et al. Effect of radiofrequency 34. Chang YS, Yang YW, Lin YN, Lin KH, Chang KC, Chang JG. catheter ablation on Doppler echocardiographic parameters in pa- Mutation analysis of KCNQ1, KCNH2 and SCN5A genes in Tai- tients with Wolff-Parkinson-White syndrome. Int Heart J 2007; wanese long QT syndrome patients. Int Heart J 2015; 56: 450-3. 48: 165-75. 32. Crotti L, Spazzolini C, Schwartz PJ, et al. The common long-QT syndrome mutation KCNQ1/A341V causes unusually severe clinical manifestations in patients with different ethnic backgrounds: Supplemental Files toward a mutation-specific risk stratification. Circulation 2007; 116: 2366-75. Supplemental Figure 33. Jimmy JJ, Chen CY, Yeh HM, et al. Clinical characteristics of pa- Please see supplemental files; tients with congenital long QT syndrome and bigenic mutations. https://www.jstage.jst.co.jp/article/ihj/58/1/58_16-133/_article/supplement