Use of Clinical Exome Sequencing in Isolated Congenital Heart Disease.

Abstract

HomeCirculation: Cardiovascular GeneticsVol. 10, No. 3Use of Clinical Exome Sequencing in Isolated Congenital Heart Disease Free AccessCase ReportPDF/EPUBAboutView PDFSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessCase ReportPDF/EPUBUse of Clinical Exome Sequencing in Isolated Congenital Heart Disease Laura Zahavich, MSc, Sarah Bowdin, MD and Seema Mital, MD Laura ZahavichLaura Zahavich From the Division of Cardiology, Department of Pediatrics, University of Toronto, Hospital for Sick Children, Toronto, Ontario, Canada. Search for more papers by this author , Sarah BowdinSarah Bowdin From the Division of Cardiology, Department of Pediatrics, University of Toronto, Hospital for Sick Children, Toronto, Ontario, Canada. Search for more papers by this author and Seema MitalSeema Mital From the Division of Cardiology, Department of Pediatrics, University of Toronto, Hospital for Sick Children, Toronto, Ontario, Canada. Search for more papers by this author Originally published4 May 2017https://doi.org/10.1161/CIRCGENETICS.116.001581Circulation: Cardiovascular Genetics. 2017;10:e001581Clinical CaseA newborn boy was transferred to our hospital with antenatal diagnosis of complex congenital heart disease (CHD) that included hypoplastic left heart syndrome with mitral atresia, aortic stenosis, double-outlet right ventricle, and normally related great arteries. Antenatal screening including maternal integrated prenatal screening and anatomy scan at 19 weeks was normal. Diagnosis was confirmed at a fetal echocardiogram performed at 23-weeks gestation because of a paternal family history of CHD. This included a history of ventricular septal defect and pulmonary stenosis in 33-year-old father who had cardiac surgery at 4 years of age, tetralogy of Fallot in the paternal uncle, and unspecified valve surgery in the paternal grandmother. The 33-year-old mother had a history of uterine fibroids, no history of CHD, and no previous pregnancy losses. The parents were nonconsanguineous. Figure 1 shows the family pedigree.Download figureDownload PowerPointFigure 1. Family pedigree showing affected members (black color). HLHS indicates hypoplastic left heart syndrome; PS, pulmonary stenosis; TOF, tetralogy of Fallot; and VSD, ventricular septal defect.Following delivery at 39+3 weeks, the patient underwent staged palliation including a Norwood procedure at 1 day of age and a superior cavopulmonary connection at 6 months of age. At age 7 months, he had an episode of duskiness during physical therapy at home followed by bradycardia with poor perfusion, which required urgent hospitalization and institution of extracorporeal membrane oxygenation. The baby was transitioned to a right ventricular assist device on day 5 of extracorporeal membrane oxygenation. Unfortunately, on day 10 after device implantation, he had an extensive proximal left middle cerebral artery infarct with significant cortical and white matter loss. Mechanical circulatory support was withdrawn, and the patient passed away.Genetic Test ResultsGenetic evaluation was performed during the first hospitalization. No dysmorphic features were identified in the patient or his father, and chromosomal microarray showed a normal 46, XY male chromosome complement. DNA was sent for sequencing for 3 CHD-associated genes (GDF1, NKX2.5, and TBX5) for which clinical testing was available at the time. This did not identify any pathogenic variants. After obtaining parental consent, clinical whole-exome sequencing (WES) was performed in the patient through a Clinical Laboratory Improvement Amendments–certified laboratory. WES identified a pathogenic heterozygous truncating variant, c.5767del (p.Q1923fs), in the NOTCH1 gene, which was verified in patient DNA using Sanger sequencing. This variant had not been previously reported and was not identified in the Exome Aggregation Consortium database containing sequencing data of 60 706 individuals.1 On the basis of the expected and observed frequency of variants in the database, NOTCH1 is felt to be extremely intolerant to loss-of-function variants. The NOTCH1 variant was inherited from the affected father (mother was variant negative) and therefore, based on segregation within affected family members, was determined to be the CHD-causing variant in our patient. We reconfirmed pathogenicity based on the American College of Medical Genetics classification for the purpose of this vignette.The parents were counseled about recurrence risk in subsequent pregnancies based on autosomal dominant inheritance pattern of the NOTCH1 mutation. The severity of CHD observed in this child coupled with the strong family history of CHD influenced parental decision to pursue preimplantation genetic testing for their next pregnancy to increase their odds of having a healthy unaffected child. This resulted in a subsequent unaffected pregnancy.Besides the NOTCH1 mutation, WES identified 3 additional variants of uncertain significance (VUS), one each in the OBSCN, c.2123C>T (p.P708L) and MED13L, c.1808G>T (p.G603) genes, which were maternally inherited, and one in FOXL1, c.842T>C (p.L281P), which was paternally inherited. A variant was also identified in the DSP gene, c.88G>A (p.V30M), which is reported to be associated with arrhythmogenic cardiomyopathy. Although the DSP variant has been reported previously as a likely benign variant or as a VUS, it was reported by the testing laboratory as pathogenic in our patient. We reconfirmed variant interpretation using American College of Medical Genetics classification for this report. Because the parents had consented to return of all medically actionable genetic test results, information about the potentially pathogenic variant in DSP found in the child was provided to the parents. The parents were offered genetic testing for the DSP gene variant and clinical screening for arrhythmogenic cardiomyopathy. The variant was again found to be paternally inherited, but the father did not have clinical features of arrhythmogenic cardiomyopathy at the time of evaluation.DiscussionGenetics of CHDCHD occurs in 0.8% to 1% of all live births, with 20% being inherited.2,3 There are >200 genes known to be associated with cardiac morphogenesis. However, no one gene accounts for >5% of CHD cases.4–15 An important challenge in genetic testing in CHD is that most mutations in patients or families with isolated CHD are extremely rare or private mutations in a large gene set, making targeted sequencing of a panel of CHD genes impractical. Further small kindred sizes preclude linkage analysis.16,17 Therefore, clinical exome sequencing was offered to the family after candidate gene sequencing was negative. Table shows a list of genes in which point mutations have been associated with both syndromic and isolated CHD.Table. Single Gene Mutations Associated With Congenital Heart DiseaseGeneCardiac LesionsSyndromeBRAFPS, ASD, HCMCFC syndrome; Noonan syndromeCBLPS, ASD, HCMNoonan syndromeCHD7ASD, VSD, TOFCHARGE associationCITED2ASD, VSD, TAPVDETS1Complex CHDJacobsen syndromeGATA4TOF, AVSD, ASD, VSDGDFTGA, DORV, TOF,IAAHRASPS, HCM, conduction abnormalitiesCostello syndromeHRASPS, ASD, HCMNoonan syndromeJAG1PS, TOF, PPSAlagille syndromeKDM6AASD, VSD, TOF, CoA, PDA, TGAKabuki syndromeKMT2DASD, VSD, TOF, CoA, PDA, TGAKabuki syndromeKRASPS, ASD, HCMCFC syndrome; Noonan syndromeMAP2K1PS, ASD, HCMCardiofaciocutaneous syndromeMAP2K2PS, ASD, HCMCardiofaciocutaneous syndromeMYH6ASDNF1PS, ASD, HCMNoonan syndromeNKX2.5ASD, TOFNOTCH1LHLNOTCH2PS, TOF, PPSAlagille syndromeNR2F2AVSDNRASPS, ASD, HCMNoonan syndromePTPN11PS, ASD, HCMNoonan syndromeRAF1PS, ASD, HCMNoonan syndromeSEMA3EASD, VSD, TOFCHARGE associationSHOC2PS, ASD, HCMNoonan syndromeSOS1PS, ASD, HCMNoonan syndromeTBX1Conotruncal defectsVelocardiofacial syndromeTBX5ASD, VSD, AVSD, conduction abnormalitiesHolt–Oram syndromeTFAP2bPDAChar syndromeASD indicates atrial septal defect; AVSD, atrioventricular septal defect; CHD, congenital heart disease; CoA, coarctation; DORV, double-outlet right ventricle; HCM, hypertrophic cardiomyopathy; IAA, interrupted aortic arch; LHL, left-sided heart lesions; PDA, patent ductus arteriosus; PS, pulmonary stenosis; PPS, peripheral pulmonary stenosis; TGA, transposition of great arteries; TAPVD, total anomalous pulmonary venous drainage; TOF, tetralogy of Fallot; and VSD, ventricular septal defect.Application of WESWES allows simultaneous exploration of rare variants in a large number of potentially causal genes and has been successful in identifying rare variants in isolated nonsyndromic CHD in several research studies.4,12,18,19 The availability of Exome Aggregation Consortium, one of the largest aggregated exome databases of >60 000 individuals, and the development of guidelines by the American College of Medical Genetics on standardized variant interpretation for pathogenicity have improved accuracy of variant classification.20 The approach includes ascribing pathogenicity to ultrarare variants (ie, variants not previously reported in reference genomes or seen with minor allele frequency of <0.01%), to variants that are predicted damaging by multiple in silico prediction programs and on functional validation studies, to variants previously reported with human disease phenotypes, and to variants that are in biologically plausible genes.21–23 In this regard, the NOTCH-signaling pathway is central to cardiogenesis required for establishing the left–right axis and looping of the heart tube.24 Mutations in NOTCH1 have been associated with left-sided heart lesions like bicuspid aortic valve, aortic stenosis, coarctation, and hypoplastic left heart syndrome, which made the finding of a pathogenic variant in this gene in our patient biologically plausible.25 However, variant classification can be discordant from laboratory to laboratory based on their own interpretation guidelines with variant interpretation changing over time.26 Therefore, access to other affected and unaffected family members, in particular, family trios for sequencing is extremely important when offering WES to a proband to allow segregation analysis. In our family, segregation of the pathogenic variant with the affected proband and affected father but not unaffected mother facilitated variant classification as pathogenic. The location of the pathogenic variant in NOTCH1 (red) overlaid onto variants found in Exome Aggregation Consortium v0.3, shown in gray, is shown in Figure 2A.27,28Figure 2B shows the variant location in red on the 3-dimensional protein structure of the ankyrin repeat domain of NOTCH1, chain A.29,30Download figureDownload PowerPointFigure 2. NOTCH1 variant location. A, The p.Q1923fs variant, shown in red, was mapped to the NOTCH1 amino acid sequence (NM_017617) and overlaid onto variants found in Exome Aggregation Consortium v0.3, shown in gray using R v3.3.2 and the lattice package. B, The 2F8Y.pdf file was used, which depicts the ankyrin repeat domain of NOTCH1 with the p.Q1923fs variant shown in red mapped onto the chain A structure by uploading the file to VMD v1.9.2.An important observation was the discordance of cardiac phenotypes among affected family members, which is being recognized with increasing frequency and which complicates the ability to perform targeted gene sequencing based on phenotype.31 Although the affected proband had a primarily left-sided heart lesion typical for NOTCH1 mutations,25,32 the parents and grandparents manifested right-sided heart lesions, suggesting that other factors may contribute to heterogeneity of phenotype. In this regard, it is tempting to speculate that one or more of the VUS found in other cardiac developmental genes could have contributed to the diverse phenotypes within the same family. For example, the proband inherited not only the NOTCH1 causal variant but also a VUS in FOXL1 and in DSP from the affected father and in OBSCN and MED13L from the unaffected mother. Studies have identified FOXL1 variants in hypoplastic left heart syndrome, de novo mutations in OBSCN in heterotaxy and conotruncal defects, and MED13L haploinsufficiency in conotruncal defects.12,32–34 Also, although DSP mutations have been characteristically associated with arrhythmogenic cardiomyopathy, a potential role of DSP in cardiac development cannot be excluded because there is overlap among some cardiomyopathy and CHD-associated genes (eg, OBSCN).35 As the oligogenic origins of CHD are being increasingly recognized, genome-wide exploration rather than a targeted search for only known cardiac developmental genes is, therefore, gaining traction for complex disorders like CHD.19However, WES has limitations. Although WES is currently cheaper than whole-genome sequencing, it does not identify copy-number variants or variants in noncoding regions. Also, detection of secondary variants adds additional burden on the testing laboratory with the requirement to report back clinically actionable findings in genes unrelated to the primary disorder to the family.36 A higher detection of VUS results in uncertainty about the diagnosis, which can be difficult for families. Recent experience with whole-genome sequencing suggests that it can provide a diagnosis in up to a third of pediatric patients presenting with congenital malformations and neurodevelopmental disorders compared with conventional testing.37 Whole-genome sequencing has the advantage of detecting both copy-number variants and single-nucleotide variants, thus providing a virtual chromosome microarray and a virtual WES. In the future, interrogation of the intergenic sequence could increase diagnostic yield further as the entire genome is better annotated.Clinical Use of Positive Genetic Test ResultEstablishing a genetic cause for CHD is important. It allows the medical team to tailor care toward the known features of the diagnosed condition in the affected individual and provides the family with a more accurate prediction of recurrence in a future pregnancy (for the proband and the parents’ future offspring).38 Identification of a genetic cause also informs genetic screening of other first- and second-degree relatives, allowing them to benefit from accurate advice about their own health and risk to their children. In this regard, identification of a genetic cause in this family allowed us to predict a more specific recurrence risk of 50% in subsequent offspring albeit with the caveat of possibility of incomplete penetrance that can result in milder phenotypes in some mutation carriers. The family was counseled accordingly before offering preimplantation genetic testing for their subsequent pregnancy, which resulted in an unaffected pregnancy. Preimplantation genetic diagnosis provides an option for genetic testing of gene-positive families before establishing a pregnancy, which reduces the likelihood of pregnancy termination should the fetus be affected.39 In the United States, some health insurance companies cover preimplantation genetic diagnosis but only for infertility, and others cover preimplantation genetic diagnosis but not in vitro fertilization, which is part of the process.Secondary Findings in WESAn important component of WES is the detection of medically actionable genetic variants that may be unrelated to the primary phenotype (also known as secondary findings). The finding of a DSP gene variant linked to arrhythmogenic cardiomyopathy in our proband is such an example and resulted in clinical and genetic screening of parents for this condition. This highlights the importance of obtaining informed parental consent for return of secondary findings before offering next-generation sequencing in the proband. The American College of Medical Genetics generated a list of 56 genes that should be interrogated by clinical laboratories independent of the primary disease for the purpose of return of secondary findings in these genes.36 Most laboratories offer the opportunity for patients or parents to opt out of receiving these results. The return of results to a minor for risk of adult-onset disorders remains controversial.Costs of WESThe use of WES for clinical testing typically raises questions about cost effectiveness of the approach. There are >10 clinical laboratories in North America that offer WES with costs ranging from US$4800 to $9000 depending on whether proband or trio whole-exome sequencing is ordered. Most of the cost is related to manpower needs for clinical interpretation because costs of sequencing itself are <US$1500. The interpretation costs are expected to decline with time as variant interpretation pipelines become more automated and accurate. Currently, because many genetic tests are not covered by health insurance, the burden of payment or copayment is often a barrier to genetic testing. For our patient, the cost of chromosomal microarray and of CHD gene panel testing was US$2440, whereas the cost of WES of proband and Sanger sequencing of the NOTCH1 mutation and FOXL1, OBSCN, and MED13L variants in the parents was US$7000. Despite an overall higher cost, WES allowed confirmation of a genetic diagnosis in the family, which allowed informed reproductive decision making and the ability to have a healthy second infant. Besides obvious health benefits, it resulted in significant healthcare-related cost savings because average costs of caring for an infant with complex CHD is estimated at US$150 000 per annum.40ConclusionsIn summary, WES is an emerging option for genetic testing in patients with isolated CHD. However, given the challenges of accurate variant interpretation for pathogenicity and the burden of consenting for and reporting secondary variants, WES should only be offered in the context of adequate pre- and post-test genetic counseling and availability of other affected and unaffected family members who can be tested for segregation analysis.AcknowledgmentsWe thank Michelle Chan Seng Yue for her assistance with variant mapping on to the 3D protein structure.DisclosuresNone.FootnotesCorrespondence to Seema Mital, MD, Hospital for Sick Children, 555 University Ave, Room 4510, Toronto, Ontario M5G 1X8, Canada. E-mail [email protected]