Familial transposition of the great arteries caused by multiple mutations in laterality genes
- Alessandro De Luca1,
- Anna Sarkozy1,6,
- Federica Consoli1,
- Rosangela Ferese1,
- Valentina Guida1,
- Maria Lisa Dentici1,
- Rita Mingarelli1,
- Emanuele Bellacchio1,
- Giulia Tuo2,
- Giuseppe Limongelli3,
- Maria Cristina Digilio4,
- Bruno Marino5,
- Bruno Dallapiccola1
- 1Mendel Laboratory, Casa Sollievo della Sofferenza Hospital, IRCCS, San Giovanni Rotondo, Italy
- 2Department of Pediatric Cardiology, G Gaslini Hospital, Genoa, Italy
- 3Department of Pediatric Cardiology, Monaldi Hospital, Naples, Italy
- 4Department of Medical Genetics, Bambino Gesù Hospital, Rome, Italy
- 5Division of Pediatric Cardiology, Department of Pediatrics, La Sapienza University, Rome, Italy
- 6Current address: Northern Genetic Service, Institute of Human Genetics, Newcastle University, International Centre for Life, Newcastle upon Tyne, UK
- Correspondence to Dr A De Luca, Istituto CSS-Mendel, Viale Regina Margherita 261, Roma 00198, Italy;
- Accepted 3 November 2009
- Published Online First 20 November 2009
Background The pathogenesis of transposition of the great arteries (TGA) is still largely unknown. In general, TGA is not associated with the more common genetic disorders nor with extracardiac anomalies, whereas it can be found in individuals with lateralisation defects, heterotaxy and asplenia syndrome (right isomerism).
Objective To analyse genes previously associated with heterotaxy in order to assess mutations in familial TGA unassociated with other features of laterality defects.
Methods Probands of seven families with isolated TGA and a family history of concordant or discordant congenital heart disease were screened for mutations in the ZIC3, ACVR2B, LEFTYA, CFC1, NODAL, FOXH1, GDF1, CRELD1, GATA4 and NKX2.5 genes.
Results Mutation analysis allowed the identification of three sequence variations in two out of seven TGA probands. A FOXH1 (Pro21Ser) missense variant was found in a proband who was also heterozogous for an amino acid substitution (Gly17Cys) in the ZIC3 gene. This ZIC3 variant was also found in another family member with a second sequence variation (Val150Ile) in the NKX2.5 gene homeodomain who was affected by multiple ventricular septal defects. A second proband was found to harbour a splice site variant (IVS2-1G→C) in the NODAL gene.
Conclusions The present study provides evidence that some cases of familial TGA are caused by mutations in laterality genes and therefore are part of the same disease spectrum of heterotaxy syndrome, and argues for an oligogenic or complex mode of inheritance in these pedigrees.
- Congenital heart disease
- molecular genetics
- transportation of the great arteries
- transposition of great vessels
Congenital heart disease (CHD) is the most common major birth defect, occurring in four to eight per 1000 live births. Transposition of the great arteries (TGA) accounts for 5–7% of all CHD, and is the most common cyanotic disorder and the most frequent CHD diagnosed in the neonatal period, with a prevalence of 0.2 per 1000 live births. In general, TGA is considered a conotruncal defect or a heart outflow defect.1 2 The most common form of TGA is the dextro-looped type, which consists of a discordant ventriculo-arterial connection so that the aorta incorrectly arises anterior and right-sided from the right ventricle, whereas the pulmonary artery incorrectly arises posterior and left-sided from the left ventricle. In contrast to the normal heart in which both outflow tracts and great vessels show a dextral (right-handed) spiralisation, in TGA the great vessels present a parallel course lacking normal spiralisation.
The pathogenesis of TGA is largely unknown. The genetic contribution versus environmental factors has been controversially discussed. Most patients are sporadic, but a few TGA families have been reported with monogenic or oligogenic inheritance.3 From the embryological point of view, two major hypotheses for TGA development have been suggested: an anomalous infundibular rotation and an aorticopulmonary septum anomaly. TGA is difficult to reproduce with animal experiments, although some interesting results have been obtained using retinoid acid in pregnant rats, in which TGA or a heterotaxy phenotype was developed using either retinoic acid or antagonists of retinoic acid, such as teratogens.4 Although TGA is rarely associated with genetic syndromes and with additional extracardiac anomalies,2 a few cases of TGA have been reported in patients with DiGeorge syndrome and deletion of chromosome 22q11.5 In contrast, TGA is often associated with other cardiac and extracardiac anomalies in children with lateralisation defects, heterotaxy and asplenia syndrome (right isomerism).6 Several families have also been reported in which some members had heterotaxy syndrome whereas others displayed isolated cardiovascular defects, including TGA.7 TGA with or without right isomerism of the lungs has been reported in mice mutated in two genes involved in the lateralisation process, Smad2 and Nodal.8 In humans, a few genes associated with heterotaxy syndrome, namely ZIC3,9 10 CFC111 and NODAL,12 were found to be mutated in isolated TGA,10 12 13 suggesting that some cases of TGA might relate to laterality defects restricted to the heart, in the absence of other heterotaxy features.
In order to assess whether isolated TGA shares a common genetic aetiology with heterotaxy, we screened seven genes known to be responsible for a subset of laterality defects, including ZIC3,9 ACVR2B,14 LEFTYA,15 CFC1,11 NODAL,12 NKX2.516 and CRELD117 in a group of familial TGA unassociated with other situs anomalies. Most of these genes either participate or cooperate in the Nodal signalling pathway, which plays an essential role in early embryonic development, mesoderm and endoderm formation and left–right axis patterning. In addition, we screened GATA4, a gene previously associated with dextrocardia,18 GDF119 and FOXH1,20 two genes mutated in a variety of CHD including TGA.
Materials and methods
The study group consisted of seven families in which the proband had complete TGA and the other affected family members presented concordant or discordant CHD.3 In all the subjects, the cardiac defects were confirmed by one or more of the following: echocardiography, cardiac catheterisation, surgical intervention and/or autopsy. All individuals had situs solitus of the atria, levocardia, D-loop of the ventricle and concordant atrioventricular connections (patent atrioventricular valves). Patients with any type of single ventricle and/or classic findings of heterotaxy were excluded. The presence of any syndromic disorder was excluded by clinical examination at the time of enrollment into the study. Information about family history was obtained by an interview with the parents of the index cases. First, mutation analysis of candidate genes was performed in the probands. Once a mutation was identified, the analysis was extended to the other available family members. Informed consent was obtained from all individuals in accordance with the protocol approved by the institutional review boards of the participating institutions. Family pedigrees and anatomical features of the patients are shown in figure 1.
Mutation analysis was performed by screening the entire coding sequence and flanking intronic portions of each candidate gene by denaturing high-performance liquid chromatography (DHPLC), using a 3500HT WAVE DNA fragment analysis system (Transgenomic, Omaha, Nebraska, USA). PCR settings, amplicon lengths and resolution temperatures for DHPLC analysis are available upon request. Bidirectional sequencing for the characterisation of DHPLC-positive cases was performed using the ABI BigDye Terminator Sequencing Kit v.1.1 (Applied Biosystems, Foster City, California, USA) and an ABI 3130 Genetic Analyzer (Applied Biosystems). We used the ClustalW program (http://www.ebi.ac.uk/clustalw/) to analyse the level of conservation of sequence variants in orthologous genes, SSF software (http://www.umd.be/SSF) to perform in-silico predictions of the effects of splice site mutations on RNA splicing and the PolyPhen (Polymorphism Phenotyping) program (http://genetics.bwh.harvard.edu/pph/) to calculate the effects of missense mutations on protein structure. As PDB structure 1NK2, corresponding to Drosophila melanogaster homeobox protein Vnd in complex with DNA, was available as the template, homology modelling with MODELLER version 9v621 was used to asses the putative consequences of the Val150Ile missense mutation on NKX2.5 protein functioning. To determine the allele frequency of each genetic variant in the unaffected population, 300 Italian subjects (600 chromosomes) of white origin were genotyped by direct sequencing for each gene investigated. Numbering of DNA sequences starts from the A of the ATG initiator codon (nucleotide +1). For protein sequences the codon for the initiator methionine is codon 1.
DHPLC analysis followed by bidirectional sequencing disclosed no pathogenic mutation in the ACVR2B, LEFTYA, GDF1, CRELD1 and GATA4 genes. A C→T heterozygous transition at nucleotide 61 in the FOXH1 gene was found in the proband of family 5 (patient HD649). This nucleotide change causes the substitution of the small amino acid proline to the polar amino acid serine at codon 21. This patient also harboured a G→T transversion at nucleotide 49 of the ZIC3 gene, causing the substitution of the evolutionarily conserved small non-polar amino acid glycine at codon 17 to the polar amino acid cysteine. PolyPhen software predicted both FOXH1 Pro21Cys and ZIC3 Gly17Cys mutations to be pathogenetic, with a position-specific independent count (PSIC) score difference of 2.1 and 1.9, respectively. The ZIC3 variant was also detected in patient HD653, who was affected by multiple ventricular septal defects. This individual was also heterozygous for a G to A change at nucleotide 448, substituting a valine with an isoleucine at codon 150 in a conserved region of the homeobox domain of the NKX2.5 gene. By homology modelling we determined that Val150 is involved in van der Waals contact with residues Phe 145, Gln 147, Leu 153, forming a dense packing of atoms that contributes to the tertiary structure of the DNA binding region of NKX2.5 (see supplemental figure 1, available online only). Mutation analysis and the pedigree of family 5 are reported in figure 2a,b, respectively. A G to C transversion at an invariant splice site residue (IVS2-1G→C) was detected in the NODAL gene in the proband of family 6 (patient HD273). This change was predicted to abolish the canonical 3′-splice site in intron 2 by SSF software (consensus values: wild type 93.46, mutant 64.51, variation (%) −30.97). In addition to this variant, patient HD273 harboured two non-synonymous changes (61A→C; Asn21His and 140G→A; Arg47Gln) in cis on the same allele on the CFC1 gene. Bioinformatics analysis by PolyPhen software predicted that Arg47Gln missense change could be a possible damaging mutation, with a PSIC score difference of 1.6 (values >1.7 are considered damaging), whereas the Asn21His variant was considered benign, with a PSIC score difference of 1.1. For family 6, no affected family member other than the proband was available for DNA testing. Mutation analysis and the pedigree of family 6 are reported in figure 3a,b, respectively. Excepting for CFC1 Asn21His and Arg47Gln, which were detected in two out of 300 controls (2/600 chromosomes, allele frequency 0.3%), all sequence variations were found to be absent in 300 normal individuals. Partial alignments of human genes with orthologues in the vicinity of the mutated amino acids are shown in supplemental figure 2 (available online only).
In summary, mutational screening of the entire coding sequence of the ZIC3, ACVR2B, LEFTYA, CFC1, NODAL, FOXH1, GDF1, CRELD1, GATA4 and NKX2.5 genes in seven TGA probands with a family history of concordant or discordant CHD allowed the identification of three sequence variations, Pro21Ser (FOXH1), Gly17Cys (ZIC3) and IVS2-1G>C (NODAL), in two subjects.
According to the pathogenetic classification proposed by Clark1 and recently by Botto et al,22 complete TGA is classified as a conotruncal heart defect, resulting from abnormalities of ectomesenchymal tissue migration, often associated with chromosomal deletion 22q11.2 in the setting of DiGeorge/velocardiofacial syndrome. However, only a few cases of TGA and deletion 22q11.2 have been reported,5 arguing that the morphogenesis of this cardiac defect is probably different. On the other hand, Ferencz et al2 classified TGA within the group of conotruncal anomalies, but considered this malformation distinct from the cases with ‘normally related great arteries’, that is, tetralogy of Fallot and interrupted aortic arch. Based on animal experiments, on the co-occurrence of TGA in patients with heterotaxy,7 on the presence of mutations in laterality genes in patients with isolated TGA13 and on the familial aggregation of complete and congenitally corrected TGA,3 we suggested that the pathogenetic group of situs and loop abnormalities might include some cases of complete TGA.6 In the present study, we have demonstrated that a number of families with TGA could harbour mutations in laterality genes. FOXH1 and NODAL are essential components of the NODAL signalling pathway. In vertebrates, this pathway specifies the embryo's left-sidedness and reversal of this pathway causes situs inversus.23 According to its central role in development, orthologues of NODAL have been described in lower organisms such as ascidians, sea urchins and recently snails.24 Interestingly, in snails NODAL was proved to be involved in body chirality,24 providing evidence of a previously suggested biological link between the spiral pattern of ventricular infundibula and great arteries6 and shell chirality in snails.24
A total of three heterozygous sequence alterations was identified in the present familial TGA patients. The FOXH1 missense change was found in a proband harbouring a second sequence variant, specifically a missense change in the ZIC3 gene. The effects of these mutations on protein function were not investigated. Nevertheless, analysis of orthologous genes, in-silico predictions of the effects of FOXH1 Pro21Ser and ZIC3 Gly17Cys variants on protein structure and screening of large numbers of control chromosomes suggest that these changes affect protein functioning. FOXH1 residue Pro21 is located N-terminally to the forkhead domain of FOXH1 (amino acids 33–134), necessary for binding DNA. Although the lack of experimental data does not allow a role to be assigned to the Pro21Ser substitution, it is worth notice that Pro21 is adjacent to a group of positively charged and conserved residues KRRKKR (amino acids 22–27) in the forkhead domain of FOXH1. We can therefore speculate that one of the possible changes associated with the Pro21Ser mutation might be an altered interaction of FOXH1 with DNA. Of note, in addition to the proband, the Gly17Cys ZIC3 mutation was detected in a family member affected by multiple ventricular septal defects, who also harboured a second sequence variation in the NKX2.5 gene homeodomain. NKX2.5 mutations have previously been identified in patients with CHD, including septal defects, conotruncal abnormalities, cardiomyopathy and hypoplastic left heart syndrome, while they are very rare in TGA.25 Therefore, the cardiac anomaly seen in this patient might be caused either by ZIC3 or NKX2.5 mutations, or could result from the additive effect of both genes. Based on computer prediction, the NODAL IVS2-1G→C variant significantly reduces the efficiency of the splice site acceptor of exon 3, which is the one containing the NODAL gene termination codon. Mutations altering the splice acceptor of the last exon of a gene generally result in the activation of a cryptic splice site, rather than in exon skipping,26 and may lead to a truncated or an abnormal protein. The NODAL IVS2-1G→C splice site change was found in a proband also harbouring two CFC1 non-synonymous sequence variations, Asn21His and Arg47Gln. These changes have previously been associated with laterality defects.27 28 Nevertheless, they have also been reported at significant frequencies in unaffected African-American individuals. Within unaffected individuals of Italian origin, the allelic frequency was 0.3% for Asn21His and 0.3% for Arg47Gln. Further investigation is necessary to comprehend whether these CFC1 polymorphic variants act as susceptibility alleles in conjunction with other genes and/or with environmental factors.
The pattern of penetrance observed in family 5, in which the patients harbour more than one mutation, while their unaffected parents are heterozygotes for a single mutation passed to their children, could be explained by assuming a ‘two’ or ‘multiple’ hit model, with one or more independently inherited genes influencing the phenotype. A ‘multiple’ hit model has been proposed for holoprosencephaly, the most common structural malformation of the embryonic brain, in which the mutations detectable in patients represent only one of the several and discrete steps needed to produce a pathological disturbance.29 Rare examples of mutations in more than one gene have previously been described and also support a similar model in some cardiac defects. For example, evidence for a combinatorial role between NODAL and ZIC3 was suggested in a woman with situs ambiguous (incomplete left/right reversal).9 In addition, Bassi et al30 reported on an individual with heterotaxy and variants in both NODAL and FOXA2. Bamford et al11 described an individual affected by dextrocardia, TGA and right isomerism bearing both a CFC1 mutation and a NODAL missense variant. Roessler et al20 reported CHD patients with more than one mutation in NODAL signalling pathway genes and reduced NODAL signalling strength.
The present data provide evidence that some families with TGA: (1) harbour mutations in laterality genes and therefore are part of the same disease spectrum including heterotaxy syndrome; (2) follow a ‘two’ or ‘multiple’ hits genetic model, with independently inherited genes influencing the cardiac phenotype; and (3) carry mutations in genes involved in NODAL signalling, supporting a previously suggested biological link between the spiral pattern of the ventricular infundibula and great arteries6 and shell chirality in snails.24 However, mutations in laterality genes were identified in only a subset of families. It remains to be clarified whether the other TGA causative genes are involved in the NODAL signalling pathway or participate in other networks of cardiac development. The observation that a number of TGA patients harbour more than one mutation could have important implications for molecular testing. In these TGA families, molecular analysis of entire pathways would be more robust than testing individual genes, presaging future studies based on the analysis of the entire genome. Furthermore, the observation that two individuals in the same family displayed discordant phenotype and different mutations prompts the molecular analysis of all affected family members rather than single probands in families with discordant CHD phenotypes.
The authors wish to thank the patients who participated in this research. They would like to extend special thanks to Arnaldo Morena and Roberto Cespa for their helpful information technology support.
Linked articles 188938.
Funding This research was supported by the Italian Ministry of Health grant RF2009 and RC2009, and Italy–USA Program on Rare Diseases (Istituto Superiore di Sanità).
Competing interests None.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the Casa Sollievo della Sofferenza Hospital, San Giovanni Rotondo.
Provenance and peer review Not commissioned; externally peer reviewed.