Familial Ebstein AnomalyCLINICAL PERSPECTIVE
Whole Exome Sequencing Identifies Novel Phenotype Associated With FLNA
Background—Familial Ebstein anomaly is a rare form of congenital heart disease. We report 7 individuals among 2 generations of 1 family with Ebstein anomaly. This family was first reported in 1991 by Balaji et al in which family members were also reported to have a mild skeletal phenotype. The most likely mechanism of inheritance was concluded to be autosomal dominant. We sought to identify the genetic pathogenesis in this family using a next generation sequencing approach.
Methods and Results—Whole exome sequencing was performed in 2 cousins in this family using the Agilent SureSelect Human all Exon 51 Mb version 5 capture kit. Data were processed through an analytic in-house pipeline. Whole exome sequencing identified a missense mutation in FLNA (Filamin A), an actin-binding protein located at Xq28, mutations in which are associated with the skeletal phenotypes Frontometaphyseal dysplasia, Otopalatodigital, and Melnick–Needles syndrome, with X-linked periventricular nodular heterotopia and FG syndrome (Omim, 305450). Review of the phenotypes of those with the mutation in this family shows increased severity of the cardiac phenotype and associated skeletal features in affected males, consistent with X-linked inheritance.
Conclusions—Although congenital heart disease is reported in families with mutations in FLNA, this is the first report of individuals being affected by Ebstein anomaly because of a mutation in this gene and details the concurrent skeletal phenotype observed in this family.
Ebstein anomaly (EA) is a rare form of congenital heart disease (CHD) that occurs in ≈1 per 200 000 births.1 In those with EA, there is malformation of the tricuspid valve with adherence of the septal and posterior valve leaflets to the myocardium, whereas the anterior leaflet is redundant, perforated, or tethered. This is associated with apical displacement of the valve annulus and concomitant atrialization and enlargement of the right ventricle.1 An associated interatrial communication can also be identified in the majority of cases with this anomaly.
See Editorial by Winlaw et al
Genetic factors are recognized in the pathogenesis of this condition as it is more common in those with a family history of (any) CHD,2 but families in which there are multiple cases of EA occurring in a Mendelian pattern of inheritance are rare.1 The only genes implicated to date are NKX2-5, identified in association with a variety of structural congenital heart defects with phenotypes including EA3 and the sarcomeric gene β-myosin heavy chain (MYH7). Mutations in MYH7 have been found in association with EA, initially in 4 individuals within the same family with this lesion and in 6% of a second cohort of unrelated individuals with this type of CHD.4
Over the last decade, next generation DNA sequencing has become widely available. The techniques used allow capture of the whole exome, and massively parallel DNA sequencing offers a valuable means to identify genes underlying Mendelian disorders.5 Next generation DNA sequencing has been widely applied to study familial and sporadic forms of CHD and helped to identify several genes underlying these conditions.6,7 Although these findings have helped understanding of CHD, they do not resolve the cause of most CHD cases.7 Although the majority of CHD is of unknown pathogenesis, next generation sequencing has led to the discovery of NR2F2, a pleiotropic transcription factor, mutations in which are associated with nonsyndromic atrioventricular septal defects.8
Here, we report a familial case of EA(Figure 1)9 manifesting across at least 2 generations, caused by a novel missense variant identified through whole exome sequencing (WES) of an affected cousin pair. The clinical manifestation of this family was reported previously but the cause was not identified.9
Subjects and Methods
The family was ascertained through the Wessex Clinical Genetics Service, having presented for genetic counseling about the likely recurrence risk of EA after the death of the male proband. Twenty-five years after their original presentation, we obtained consent for detailed genetic investigations in surviving members of the family. The proband, male III (1) was severely affected; he was born at 33 weeks of gestation and died at 10 days of age. Unfortunately, no genetic material was available from him. WES was conducted on DNA samples from individuals III (3) and III (5). Six affected surviving members of the family were examined by a clinical geneticist and their echocardiograms were reviewed by 2 cardiologists. Detailed phenotyping was completed after the identification of a likely causal mutation. Targeted characterization of phenotypic features associated previously with the likely causal gene was undertaken.
Genomic DNA was extracted from peripheral venous blood samples collected in EDTA. DNA concentration was estimated using the Qubit 2.0 Fluorometer and A 260:280 ratio calculated using a nanodrop spectrophotometer. The average DNA yield obtained was 150 µg/mL, and ≈20 μg of DNA was used for next generation sequencing for each patient.
WES Data Generation and Data Analysis
WES was performed using the Agilent SureSelect Human all Exon 51 Mb version 5 capture kit. As described previously,10,11 default parameters were applied: fastq raw data generated from Illumina paired-end sequencing were aligned against the human reference genome (hg19) using Novoalign (novoalign/2.08.02). SAMtools12 (samtools/0.1.19) was used to call variation and ANNOVAR (annovar/2013Feb21)13 was applied for variant annotation against a database of RefSeq transcripts. A bespoke script was used to assign individual variants as novel if they were not reported previously in the dbSNP137 databases,14 1000 Genomes Project,15 the Exome Variant Server of European Americans of the NHLI-ESP project with 6500 exomes (http://evs.gs.washington.edu/EVS/), in 46 unrelated human subjects sequenced by Complete Genomics,16or in the Southampton database of reference exomes (n=329).
Resultant variant files for each individual were subjected to further in-house quality control tests to detect DNA sample contamination and ensure sex concordance by assessing autosomal and X-chromosome heterozygosity. Variant sharing between all pairs of individuals was assessed to confirm sample relationships. Sample provenance was confirmed by independent genotyping of a validated single nucleotide polymorphism panel, developed specifically for exome data.17
MYH7 and NKX2-5 are the 2 candidate genes known to be involved in EA. These 2 genes were tested previously by Sanger sequencing within the family with negative findings. Nevertheless, these genes were selected as top priority to confirm negative results from this alternative sequencing resource.
The Human Gene Mutation Database Professional 2013.4 (BIOBASE Biological Databases)18 was interrogated using the inclusive term heart disease to retrieve an extensive list of causal genes known to be involved in heart conditions. The query returned 338 genes from which 219 were selected as associated with clinical phenotypes similar to congenital heart defects using the following terms: for example, congenital heart defects, heart valve defect, atrial septal defects and postaxial hexadactyly, and cardiomyopathy and hypertrophic (Table I in the Data Supplement for complete list).
Using WES on the affected cousin pair, their exomes were compared to identify all genetic variants shared between them. All shared variants were removed that were observed (in various states of zygosity) within the local Southampton control cohort of exomes (n=329). Of the remaining variants, synonymous variants with low likelihood to impact protein function, splicing variants with a MaxEnt score of <3, and variants with low conservation across species (PhyloP <0.99) were removed. The known phenotypic effects of the remaining variants were examined for relevance to cardiac anomalies.
This study meets the standards expected by the governance structure of the National Health Service and a clear and informed consent was taken from the family about the potential benefits and limitations of such work.
Quality Control Analysis
Exome data were high quality defined by 80% of mappable bases of the Gencode defined exome represented by coverage of at least 20 reads. The average depth of coverage was 60.52 and 54.81 for patients III.3 and III.5 respectively, Table II in the Data Supplement. The 2 samples from the affected cousins selected for exome analysis exhibited expected variant sharing for third degree relatives and no excess of sharing with any other sample on same dispatch DNA plate. Autosomal and X-chromosome heterozygosity were consistent with sex and did not indicate any sample contamination. VerifyBamID19 did not indicate any presence of contamination, and the application of a single nucleotide polymorphism tracking panel17 confirmed sample provenance.
In the Tier 1, no coding variation was found in MYH7 and NKX2-5 consistent with previous Sanger sequencing results.
Tier 2, the analysis was extended to the 219 genes prioritized from Human Gene Mutation Database. A total of 429 variants were called in either affected cousin across the 219 genes. Given the apparent dominant mode of inheritance evidenced by the pedigree segregation and further supported by literature findings,20,21 401 mutations seen in heterozygous form in a local reference database of 329 individual without heart defects were excluded. Of the remaining 28 variants, a further 27 variants were excluded from downstream analysis as they were not observed in both affected cousins and occurred in 1 individual only. A single variant on the X-chromosome in the FLNA(Filamin A) gene remained (NM_001110556, c.G4660A, p.G1554R). This variant was novel by our assessment and consistently predicted to be deleterious by several in silico software annotation tools (SIFT, Polyphen2, Mutation Taster, and Gerp=5.67).
Finally, for the Tier 3 analysis, we identified 16 543 variants shared between the cousins. These were filtered to yield a final list of 9 variants (Table III in the Data Supplement). Comprehensive literature review excluded 8 variants within genes functionally irrelevant to phenotype (Table III in the Data Supplement). The novel nonsynonymous FLNA variant (c.G4660A, p.G1554R) represented the only outstanding candidate variant. The variant was confirmed using Sanger sequencing. We confirmed the segregation of the genotypic variant assumed to be causal to be consistent with the inheritance pattern of the family on the 6 living affected blood relatives diagnosed with EA. The variant was confirmed as carried in heterozygous state in all affected females and in hemizygous state in all affected males. Patient IV(1) was not tested for the variant.
Mutation Within FLNA
The p.G1554R mutation segregating in our family with EA sits within the 14th repeated rod-domain (Figure 2). The mutation replaces a nonpolar amino acid with a polar amino acid and is likely to impact tertiary structure as indicated by PolyPhen2 and GERP scores.
A summary of the cardiac, musculoskeletal, and facial phenotypes of the family members available for clinical examination is shown in the Table.
Medical notes were not available but there was a history, given by surviving relatives of this individual having had a mitral valve replacement in middle age.
Patient I (2)
Limited information was available but was reported by her daughter, II(2), to have no symptoms of cardiac disease at the age of 93 years. She also had a son and a daughter with a different partner and neither child, or in turn, any of their 9 children have presented with cardiac disease.
Patient II (1)
The echocardiogram demonstrated mild apical displacement of the tricuspid valve with a large antero-superior leaflet. There was no mitral valve anomaly and no mitral stenosis or regurgitation; the aortic valve was normal. The left atrium was mildly dilated, consistent with the age of the patient.
Patient II (2)
In this description, the following findings also relate to those found at surgery (repair of EA and placement of a 32 mm tricuspid annuloplasty ring). There was a large anterior leaflet of the tricuspid valve that was partially tethered. There was a partially tethered posterior leaflet and a fully Ebsteinized septal leaflet of the tricuspid valve.
Patient III (1)
The infant was born at 33 weeks gestation with hydrops fetalis. Echocardiography showed severe EA with absent flow into the pulmonary arteries. Despite optimum medical treatment, including with ventilation, dopamine, and prostaglandin infusions, he died at 10 days of age.
Patient III (2)
The findings were similar to III(4) above with mild apical displacement of the septal leaflet of the tricuspid valve (1.3 cm) and mild associated tricuspid regurgitation.
Patient III (3)
The following findings relate to those found at surgery (repair of EA and replacement of the mitral valve) and on preoperative 3-dimensional echocardiography. The tricuspid valve was typical of EA with a large mobile anterior-superior leaflet fused to the posterior leaflet that was partially tethered. There were some secondary cords to the back of both antero-superior and posterior leaflets, with complete tethering of the final portion of the posterior leaflet and Ebsteinization of the whole of the septal leaflet. There was a funnel-like opening into the right ventricle with associated tricuspid stenosis and regurgitation. The mitral valve was stenotic with a fixed orifice (1 cm2) and heavy calcification of the posterior leaflet and a parachute arrangement of the mitral chordae that were fibrosed and fused.
Patient III (4)
There was isolated mild apical displacement of the tricuspid valve (1.3 cm) with mild associated tricuspid regurgitation.
Patient III (5)
On examination of echocardiogram findings, the tricuspid valve was dysplastic with apical displacement of the septal leaflet (2 cm), mild tricuspid regurgitation, and mild tricuspid stenosis (peak gradient 7, mean 5 mm Hg) with E/A reversal. The mitral valve was also dysplastic with a large antero-lateral papillary muscle and 3 small postero-medial papillary muscles instead of 1 well-formed structure. Mild mitral regurgitation and mild mitral stenosis (peak gradient 14, mean 4 mm Hg) were seen with a dilated left atrium (length 8 cm, volume 34 mL, and area 18.1 cm2). There was a small amount of calcification on the anterior leaflet. An effectively bicuspid aortic valve was seen with 2 low commissures and a small right coronary cusp. There was mild aortic regurgitation but no aortic stenosis; there was no pulmonary regurgitation. A moderate secundum atrial septal defect with an associated left to right shunt was evident.
Patient IV (1)
Neonatal clinical examination, including detailed cardiovascular examination was normal. At the time of assessment of other family members, patient IV(1)’s parents declined an echocardiogram, genetic testing, or further clinical assessment.
Although certain mutations in FLNA are associated with a neurological phenotype (periventricular nodular heterotopia), there was no history of seizures or unsteadiness in any members of the family. Cerebellar examination was normal in all. There was no history of learning difficulties in any family member examined. Systematic enquiry about urologic diagnoses and deafness was also negative. One family member III(4) had a diagnosis of irritable bowel syndrome but no other abnormalities of bowel motility were reported. About the skeletal phenotype, all family members examined gave a history of stiff joints with both surviving males having fixed flexion of the knees and ankles. A history of joint stiffness was also reported by the family to have been present in family member I(1); anecdotally his gait was also reported to be the same as that observed by the family in both III(3) and III(5). All patients examined had proximally placed and externally rotated fifth toes; male III(5) also had short thumbs with hypoplastic distal phalanges. Both males and 2 of the females had limited supination of the elbows. A dental phenotype was observed in III(5) who had oligodontia with incomplete eruption of secondary dentition. Keloid scarring, noted in all members of the family who were examined was also reported to have been present in I(1; Table for summary).
Although not examined as part of this study, individual IV(1) was described by the family as having no joint limitation, clearly demonstrated to them by her enthusiasm for exercise and general flexibility. The ease of movement in IV(1) was described as a direct contrast to the joint restriction seen in members of generation III during childhood. Patient I(1) was deceased at the time of the study; he died from bowel cancer at the age of 75 years.
EA is a rare condition and even less frequently observed segregating in families. We describe a family with evidence for this rare cardiac defect over at least 2 generations. This family was described previously, but genetic pathogenesis remained elusive.
In this study, we have applied contemporary sequencing technology and effective filtering of a cousin pair and identified a single variant within an X-linked gene known to cause cardiac developmental defects but not associated previously with EA. We have confirmed the mutation using traditional Sanger techniques and demonstrated heterozygous and hemizygous carrier status for all female and male affected family members respectively.
The single base substitution (G>A) causes a missense mutation in which a nonpolar Glycine amino acid residue is replaced with a positively charged Arginine. The variant was within a highly conserved (Polyphen=0.999) residue region of the protein that encodes for the 14th Ig-like domain. The tertiary structure of the Ig-like repeated domain consists in 7 antiparallel β sheets and therefore it might be expected that the p.G1554R mutation causes a conformational change in the β strands.
We have found no previous evidence of this variant in our local or public repository databases. However, mutations within the Ig-Rod domain are causal of a variety of congenital heart defects.
The phenotypic spectrum of disease associated with mutations in FLNA is broad and includes several conditions with a predominantly skeletal phenotype: FG syndrome, Frontometaphyseal dysplasia, Melnick–Needles syndrome,27 Otopalatodigital syndrome, types I and II,27 and terminal osseous dysplasia.28 It also includes individuals with a predominantly neurological phenotype (periventricular nodular heterotopia)29 and those with cardiac manifestations of disease (X-linked cardiac valvular dysplasia).30 Its causality is further underscored by the missense and loss of function variants presented in ExAC.31 The family identified were examined for any features of the above conditions. Macrocephaly, seen in FG syndrome and was found in both of the males and 1 female, however, this condition also includes mental retardation, not present in this family. Survival to adulthood noted in this family is not typical in males affected by Melnick–Needles syndrome or Otopalatodigital syndrome type II. Skeletal features including joint contractures, prominent supraorbital ridges, and proptosis observed in members of the family described are typically associated with fibromuscular dysplasia;27 however, typical of this condition is also conductive and sensorineural hearing loss, neither of which were present in this family. In addition, while cardiac anomalies are described in fibromuscular dysplasia, EA is not.
The data available for the 6 cases described above (based on a combination of the echocardiogram and magnetic resonance imaging findings and the operative findings in the 2 patients who underwent surgery) clearly identify varying degrees of severity of EA in these cases. The phenotype varied between mild septal displacement of the septal leaflet of the tricuspid valve with mild associated tricuspid regurgitation, to severe EA with a combination of tricuspid stenosis and regurgitation. In 3 of the 4 females (II(1), III(4), and III(2)), the anomaly was subtle with isolated mild displacement of the tricuspid valve and only mild tricuspid regurgitation. Only in female patient II (2) was the anomaly more severe and required surgery. This more severe phenotype could be explained by a hypothesized skewed pattern of X-chromosome inactivation that in many other X-linked disorders has been considered as the cause of the phenotype in female carriers.
The males III (3) and III (5) had more severe cardiac presentation with both tricuspid and mitral valves involvement requiring surgery. It is unusual for patients with EA to have mitral valve disease: this is congenital in nature but has also been associated with calcification that was progressive in at least 1 of them. The patient III (3) developed severe mitral valve stenosis and this was an important factor in the timing of surgery. The other, III (5), had a dysplastic mitral valve with only mild calcification of the anterior leaflet and only mild stenosis and regurgitation. Only 1 patient had an atrial septal defect and 1 patient (operated) had a patent foramen ovale. However, it is not possible to exclude small patent foramen ovales in the other patients as they have not had a contrast echocardiogram.
The finding of mitral valve disease in addition to EA and abnormalities of the pulmonary and aortic valves in members of this family illustrates a phenotypic overlap with X-linked valvular dysplasia.23 The variability of the phenotype comparing affected males with each other (and similarly, comparing the affected females with each other) is consistent with that observed among individuals who harbor pathogenic variants in other genes that predispose to the manifestation of a cardiac phenotype, such as GATA4.32
In conclusion, the absence of a neurological phenotype, survival of some males to adulthood, and the presence of EA in multiple members of this family mean that the FLNA mutation described is associated with the unique phenotype seen in this family and furthermore, it represents a novel cause for familial EA. However, because of the presence of additional features segregating with EA in this family, there is little evidence that mutations in FLNA will be a common cause of sporadic, nonsyndromic EA.
We thank all the patients and their families for their contribution to this study.
Sources of Funding
Dr Andreoletti is supported by The Crohn’s in Childhood Research Association and The Gerald Kerkut Charitable Trust.
The Data Supplement is available at http://circgenetics.ahajournals.org/lookup/suppl/doi:10.1161/CIRCGENETICS.116.001683/-/DC1.
- Received December 2, 2016.
- Accepted September 20, 2017.
- © 2017 American Heart Association, Inc.
- Attenhofer Jost CH,
- Connolly HM,
- Dearani JA,
- Edwards WD,
- Danielson GK
- Postma AV,
- van Engelen K,
- van de Meerakker J,
- Rahman T,
- Probst S,
- Baars MJ,
- et al
- Arrington CB,
- Bleyl SB,
- Matsunami N,
- Bonnell GD,
- Otterud BE,
- Nielsen DC,
- et al
- Balaji S,
- Dennis NR,
- Keeton BR
- Christodoulou K,
- Wiskin AE,
- Gibson J,
- Tapper W,
- Willis C,
- Afzal NA,
- et al
- Drmanac R,
- Sparks AB,
- Callow MJ,
- Halpern AL,
- Burns NL,
- Kermani BG,
- et al
- de Castro E,
- Sigrist CJ,
- Gattiker A,
- Bulliard V,
- Langendijk-Genevaux PS,
- Gasteiger E,
- et al
- Robertson SP,
- Twigg SR,
- Sutherland-Smith AJ,
- Biancalana V,
- Gorlin RJ,
- Horn D,
- et al
- Kyndt F,
- Gueffet JP,
- Probst V,
- Jaafar P,
- Legendre A,
- Le Bouffant F,
- et al
This study details the finding of a novel missense variant in FLNA gene (filamin A) identified through whole exome sequencing in a case of familial Ebstein anomaly (EA). These results provide evidence for a highly penetrant identifiable genetic cause for this condition in the 6 living affected blood relatives diagnosed with EA. Though rare, familial EA has been reported previously, including in association with mutations in MYH7 and NKX2.5, both of which are inherited in an autosomal dominant way. FLNA is located on the X-chromosome and the X-linked inheritance of this novel variant is consistent with the phenotypes observed in this family. An X-linked pattern of inheritance for EA was not reported previously in the literature. Appreciation of the different inheritance patterns of familial EA is important to give accurate counseling on the recurrence risks for this condition. In addition, the clinical details in this report expand the phenotypic spectrum associated with mutations in FLNA. The recognition of these additional features will allow the clinician to conduct comprehensive clinical evaluation when assessing individuals with pathogenic variants in this gene. Such features may also act as diagnostic clues when assessing those with EA and a possible underlying genetic cause.