Author + information
- Received December 11, 2002
- Revision received March 18, 2003
- Accepted May 30, 2003
- Published online November 5, 2003.
- Doff B. McElhinney, MD*,
- Elizabeth Geiger, MS*,
- Joshua Blinder, BS*,
- D. Woodrow Benson, MD, PhD† and
- Elizabeth Goldmuntz, MD, FACC*,* ()
- ↵*Reprint requests and correspondence:
Dr. Elizabeth Goldmuntz, Division of Cardiology, The Children's Hospital of Philadelphia, Abramson Research Center 702A, 3516 Civic Center Blvd., Philadelphia, Pennsylvania 19104-4318, USA.
Objectives The purpose of this study was to estimate the frequency of NKX2.5mutations in specific cardiovascular anomalies and investigate genotype-phenotype correlations in individuals with NKX2.5mutations.
Background Recent reports have implicated mutations in the transcription factor NKX2.5as a cause of various congenital heart defects (CHD).
Methods We tested genomic deoxyribonucleic acid from 608 prospectively recruited patients with conotruncal anomalies (n = 370), left-sided lesions (n = 160), secundum atrial septal defect (ASD) (n = 71), and Ebstein's malformation (n = 7) for NKX2.5mutations.
Results Twelve distinct mutations in the NKX2.5coding region were identified in 18 of 608 patients (3%), including 9 of 201 (4%) with tetralogy of Fallot, 3 of 71 (4%) with a secundum ASD, one each with truncus arteriosus, double-outlet right ventricle, L-transposition of the great arteries, interrupted aortic arch, hypoplastic left heart syndrome, and aortic coarctation, but in no patients with D-transposition of the great arteries (n = 86) or valvar aortic stenosis (n = 21). Eleven of the mutations were amino acid-altering missense nucleotide substitutions or deletions, and one was predicted to cause premature termination of translation. None of the mutations were in the homeodomain. Sixteen of the 18 individuals with NKX2.5mutations in this study had no family history of congenital cardiovascular anomalies, and one had first-degree atrioventricular (AV) block.
Conclusions NKX2.5mutations occur in a small percentage of patients with various CHD. Most of the mutations identified in this study were missense, outside the homeodomain, and not associated with AV block. These findings suggest that NKX2.5mutations in non-homeodomain regions may be important in the development of human structural cardiac defects.
Genetics in context
Focus on gene mutations in congenital heart disease
Although congenital heart defects (CHD) are the leading cause of morbidity and mortality in infants, the etiology of most CHD remains unknown, with the influence of genetics still a topic of debate. The increased recurrence risk for parents with an affected child suggests the potential for inherited predisposing mutations that are incompletely penetrant.
Heterozygous mutations in the transcription factor, NKX2-5, were among the first evidence of a genetic cause for congenital heart disease and were initially found in pedigrees with autosomal dominant transmission of cardiac septal defects. Most reported NKX2-5mutations were in the homeodomain, a critical part of the protein that interacts with deoxyribonucleic acid (DNA), and were typically associated with cardiac conduction anomalies. However, the impact of NKX2-5mutations on “sporadic” forms of isolated CHD in a broad population was unknown. In this study by McElhinney et al., 608 children with a variety of non-familial CHD, but normal conduction, were studied for mutations in NKX2-5. Three percent of the entire cohort, a relatively large number for a single gene, had mutations that did not appear to be polymorphisms. Interestingly, none were in the critical homeodomain, suggesting that other protein domains likely affect specific developmental processes without disrupting development of the conduction system.
It remains intriguing that disruption of a single gene can cause the diverse forms of CHD described here. This could be attributed to specific roles of distinct regions of the NKX2-5 protein or simply the presence of other stochastic events. Finally, the notion of CHD as an inherited disorder with incomplete penetrance is supported by the common identification of an NKX2-5point mutation in a normal parent. Similar analyses for other critical genes may together begin to establish a genetic basis for most “sporadic” forms of CHD. This effort will lay the foundation for more accurate genetic counseling, clinical interventions, and most importantly, new approaches for prevention of CHD in those who are genetically susceptible.
Deepak Srivastava, University of Texas Southwestern Medical Center, Dallas, Texas
Mutations of NKX2.5, primarily in the homeodomain, have been identified in at least 11 kindreds with atrioventricular (AV) conduction block and concurrent congenital heart malformations, including secundum atrial septal defect (ASD) and tetralogy of Fallot (Table 1) (1–3). For this reason, we genotyped prospectively recruited patients with tetralogy of Fallot for NKX2.5and identified four missense mutations in six of the 114 patients (5%) (4). All of these mutations were outside of the homeodomain and in individuals without AV conduction abnormalities, as was the case in one previously reported patient with tetralogy of Fallot and an NKX2.5mutation (2). To investigate further the genotype-phenotype correlation in patients with congenital heart disease and an NKX2.5mutation, and to estimate the frequency of NKX2.5mutations in individuals with specific congenital cardiovascular anomalies, we genotyped 608 prospectively recruited patients with congenital heart disease by conformation-sensitive gel electrophoresis (CSGE) and/or direct sequencing. In contrast to earlier studies, which ascertained subjects on the basis of familial congenital heart disease (1–3), the subjects in this report were enrolled on the basis of specific cardiovascular phenotypes without regard for family history or the presence of AV conduction abnormalities.
Genotype data from a total of 608 subjects are summarized for this report. Genomic deoxyribonucleic acid (DNA) from 474 new subjects with conotruncal cardiovascular anomalies (n = 236), left-sided lesions (n = 160), secundum ASD (n = 71), and Ebstein's malformation (n = 7) were tested for mutations in the coding region of NKX2.5. Conotruncal anomalies included new patients with tetralogy of Fallot (n = 67), D-transposition of the great arteries (n = 86), double-outlet right ventricle (n = 31), interrupted aortic arch (n = 23), truncus arteriosus (n = 22), and L-transposition of the great arteries (n = 7). Left-sided lesions included hypoplastic left heart syndrome (n = 80), coarctation of the aorta (n = 59), and valvar aortic stenosis (n = 21). To estimate frequency, we included the 134 patients with tetralogy of Fallot previously studied in our laboratories and reported earlier, given that they were enrolled in the same protocols as the other subjects (2,4). Therefore, the total number of patients described in this report is 608.
Subjects were drawn from a cohort of patients that were recruited prospectively into studies on the genetic basis of congenital heart disease from 1991 to 2001. Informed consent was obtained from study participants in accordance with protocols approved by the Institutional Review Board for Human Research at the Children's Hospital of Philadelphia or the Medical University of South Carolina. All subjects with conotruncal malformations were tested for a chromosome 22q11 deletion as previously described (5). Some patients underwent other cytogenetic testing if clinically indicated. Patients were invited to enroll in the aforementioned studies after admission to the hospital, exclusively on the basis of the CHD present in the proband, with no selection for or against familial congenital heart disease or syndromic associations. However, patients with a chromosome 22q11 deletion, trisomy 21, or other identified chromosomal anomaly were excluded from the present analysis. Cardiovascular diagnoses were confirmed by attending pediatric cardiologists (E.G. and D.W.B.), who reviewed echocardiograms and/or echocardiogram reports, cardiac catheterization reports, and operative notes if applicable. Electrocardiograms were reviewed for abnormalities of AV conduction in all patients with secundum ASD and Ebstein's malformation, as well as those with other anomalies who were found to have an NKX2.5mutation. Family history of congenital heart disease was ascertained by report (i.e., clinical testing of parents was not performed for the purposes of this study), and an effort was made to obtain parental DNA for mutation analysis in all patients found to have an NKX2.5mutation. Medical records of patients with an NKX2.5mutation were reviewed to determine whether any non-cardiac congenital malformations or recognized genetic syndrome was present.
Deoxyribonucleic acid was extracted from either whole blood or a lymphoblastoid cell line using standard techniques. The coding region of the NKX2.5gene, including exon/intron boundaries, was amplified from genomic DNA by polymerase chain reaction (PCR) using standard techniques, as described previously (4). The PCR products were screened for sequence alterations either by CSGE, a modification of heteroduplex analysis (6), as described previously (4), or by direct sequencing. The PCR products were directly sequenced in certain cohorts (e.g., all patients with tetralogy of Fallot, Ebstein's malformation, or an ASD) and in selected patients with other diagnoses. The remaining patients were evaluated initially by CSGE, and in those with an aberrant band on the CSGE gel, the shifted region of the gene was subsequently sequenced with an automated cycle sequencer (ABI Prism 373 or 377 Sequencer, ABI BigDye Terminator Cycle Sequencing Kit, Applied Biosystems, Foster City, California), as described previously (4). Among the 608 patients, the NKX2.5coding region was sequenced in its entirety in 316 (52%) and partially in 80 (13%). Sequence alterations were examined in the context of the open reading frame to determine whether they would alter the corresponding amino acid.
Twelve distinct mutations in the coding region of NKX2.5were identified in 18 of the 608 (3%) probands evaluated. Four mutations in seven of these subjects were previously reported (2,4)but are included in this report to allow for estimation of frequency, as previously noted. The cardiovascular anomalies in the individuals tested and affected are summarized in Table 2. The details of the specific mutations are summarized in Table 3.
Among patients with conotruncal anomalies, NKX2.5mutations were detected in 13 patients, including 9 of 201 patients with tetralogy of Fallot (4%), and one each with truncus arteriosus, double-outlet right ventricle, interrupted aortic arch, and L-transposition of the great arteries. Notably, mutations were found among the 86 patients with D-transposition of the great arteries. Among patients with conotruncal anomalies and an NKX2.5mutation, the sidedness or branching pattern of the aortic arch was abnormal in 6 of the 11 in whom the aortic arch anatomy was defined, including 5 of the 9 mutation-positive patients with tetralogy of Fallot and the mutation-positive patient with interrupted aortic arch. Of 160 patients with left-sided lesions, two were found to have mutations, including one with hypoplastic left heart syndrome and one with coarctation of the aorta and an intact ventricular septum. Mutations were detected in 3 of 71 patients with a secundum ASD (4%), and none of 7 with Ebstein's malformation. None of the mutation-positive patients had extracardiac anomalies or syndromic features.
Eleven of the mutations were missense nucleotide substitutions (n = 10) or in-frame deletions (n = 1) predicted to result in alteration or deletion of a single amino acid residue, and one was a five-nucleotide insertion predicted to cause premature termination of translation (Table 3). Four of the mutations were located in (n = 1) or immediately 3′ to (n = 3) the conserved amino-terminal TN domain; one was midway between the TN domain and the homeodomain; one was located just 5′ to the homeodomain; three were located in (n = 2) or immediately 3′ to (n = 1) the conserved carboxy-terminal NK2 domain; and three were located in the distal 3′ coding region (Fig. 1). None of the mutations were in the homeodomain. Abnormal AV conduction was identified in only one mutation-positive patient, who had first-degree AV block with a secundum ASD and an insertion mutation in the 3′ coding region predicted to truncate the protein.
Sixteen of the 18 individuals with an NKX2.5mutation had no known family history of congenital heart disease (although parents were not assessed for the presence of occult cardiovascular anomalies in this study). The other two were known to have at least one relative with a cardiovascular defect, including a bileaflet aortic valve in the grandfather of the patient with an Ala127Glu mutation and secundum ASD, and a ventricular septal defect in the father of a patient with an Arg25Cys mutation and tetralogy of Fallot. The respective NKX2.5mutations were carried by both affected relatives. Of the 16 patients with no family history of congenital heart disease, parental DNA was available and evaluated for the corresponding mutation by direct sequencing in six (Lys15Ile, Glu21Gln, Arg25Cys n = 2, Ala219Val, D235AFSter, del291Asn). In all six of these cases, one parent carried the mutation, consistent with decreased penetrance. Parental DNA was not available for the remaining 10 patients with mutations.
The NKX2.5gene was sequenced from the genomic DNA of 50 random control subjects who were not known to have congenital cardiovascular disease. None of the 12 mutations detected in our study patients were found in the 100 control chromosomes. Among the mutations identified in our study population, the Arg25Cys mutation was found in multiple subjects, whereas the other mutations were unique. Five of the seven patients with an Arg25Cys mutation in this series were of African-American ethnicity, one was Hispanic, and one was Caucasian. This mutation was identified in 2 of 43 (5%) African-American control subjects, though it was not present in 100 chromosomes from random control subjects, as previously described (4).
NKX2.5is the earliest known marker of myocardial progenitor cells in all species in which it has been studied (7–10). Just as mutation of the Drosophila NKX2.5homolog, tinman, results in failure of heart formation (11,12), homozygous inactivation of NKX2.5in mammals results in impaired cardiac looping and embryonic lethality (13,14). Mice with a heterozygous cardiac-specific deletion of NKX2.5exhibit a more subtle phenotype, which consists of atrial septal dysmorphogenesis, occasional abnormalities of the aortic valve, and, in females, prolongation of the PR interval (15). Numerous studies have investigated the mechanisms of NKX2.5regulation and its interaction with other transcription factors thought to be important in early cardiac development (14,16–21). Through these and other studies, it has become clear that NKX2.5plays a central role in the determination of myocardial cell fate, and there is evidence that it may be important during postnatal life as well (22). Accordingly, there has been considerable interest in the potential role of NKX2.5in human cardiac development and disease.
Based on our identification of NKX2.5mutations in individuals with isolated tetralogy of Fallot (4), we extended our investigation to other conotruncal lesions. The independent finding of an Arg25Cys mutation in an individual with hypoplastic left heart syndrome provided the rationale to extend the scope of our study to encompass prospectively enrolled patients with left-sided lesions as well. The association of NKX2.5mutations with secundum ASD and tricuspid valve anomalies (2)prompted our prospective evaluation of patients with sporadic secundum ASDs and Ebstein's malformation. In this report, we have detailed our findings in these prospectively recruited cohorts of individuals with primarily sporadic congenital cardiovascular anomalies. Among 608 patients genotyped, 18 (3%) were found to have an NKX2.5mutation. Although NKX2.5mutations were found in individuals with cardiovascular anomalies of different types and categories (secundum ASD, various conotruncal anomalies, and two left-sided lesions), the most notable findings of this study are the frequency of mutations in patients with secundum ASD (4%) and tetralogy of Fallot (4%), and the absence of mutations in a relatively large cohort of patients with D-transposition of the great arteries.
Eleven of the 12 mutations found in this study were unique. Only the Arg25Cys mutation was found in multiple study patients (n = 7). The significance of this mutation is not clear, as has been discussed in a previous report (4). Functional analysis of this mutation reported by Kasahara et al. (23)showed normal nuclear localization and normal binding of the mutant protein to monomeric DNA binding sites, but a threefold higher concentration of mutant protein was required for equally distributed binding to monomeric and dimeric DNA binding sites, suggesting a slight impairment of dimerization by the mutant NKX2.5. The effect of this mutation on NKX2.5function will require further evaluation.
In the majority of previously reported patients with an NKX2.5mutation in conjunction with familial AV block and a secundum ASD, the NKX2.5mutation alters coding sense in the homeodomain and/or is predicted to result in a truncated protein (1–3). In contrast, all of the mutations that we detected were in the 5′ or 3′ coding region of the gene, with no homeodomain mutations. Only one patient with an NKX2.5mutation (the only patient with a truncating mutation in this series) had AV block. Experimental studies have shown that the carboxy-terminal portion of NKX2.5is important in NKX2.5function distinct from the role of the homeodomain (23–25). Kasahara et al. found this portion of the gene to be critical for the cooperative homodimerization and heterodimerization (with other transcription factors such as GATA4) of NKX2.5protein on dimeric DNA binding sites (24). In functional studies of the previously reported NKX2.5mutations, these authors demonstrated decreased binding of mutant NKX2.5proteins with mutations outside the homeodomain to dimeric DNA binding sites despite normal binding to monomeric sites (23).
In addition, Schinke et al. (25)examined the effects of cardiac-specific deletion of the NK2 domain of NKX2.5in mice. Null mutants died in utero at E14.5 and exhibited a variety of structural cardiovascular anomalies, including ventricular septal defect, double-outlet right ventricle, and AV septal defect, as well as downregulation of the ventricular markers Irx4 and MLC-2v, specifically in the right ventricle. Heterozygous NK2 mutant mice, which survived to term, also developed various structural cardiovascular defects. Binding of the NK2 mutant protein to DNA was not impaired, and mutant mice did not demonstrate AV conduction abnormalities. These findings suggest that the NK2 domain plays an important role in cardiovascular development independent of the homeodomain.
Our finding of 12 distinct mutations in the 5′ and 3′ coding regions of NKX2.5in 18 of 608 patients (3%) with congenital cardiovascular anomalies, along with the aforementioned experimental findings, suggests that NKX2.5mutations outside of the homeodomain, either alone or in conjunction with as-of-yet unidentified modifying factors, may cause structural congenital cardiovascular anomalies without affecting AV conduction. In 8 of the 16 patients with an NKX2.5mutation, a parental carrier was identified, two of whom were known to have a congenital cardiovascular anomaly. The finding of reduced phenotypic penetrance associated with the NKX2.5mutations identified in this study is consistent with patterns frequently observed in complex developmental anomalies and may indicate that these mutations are associated with an increased risk of congenital heart disease. We do not believe that the mutations we have detected are random changes, insofar as they were not detected among control subjects or patients with certain diagnoses (e.g., D-transposition of the great arteries). Further analysis of NKX2.5-mediated regulation of cardiovascular development, as well as functional studies of the mutations that have been identified, is necessary to determine the mechanisms by which non-homeodomain NKX2.5mutations lead to congenital cardiovascular anomalies.
We thank Seigo Izumo, MD, for valuable discussion and commentary on the manuscript.
☆ Supported by the National Institutes of Health (P50 HL62177 [E.G.] and P50 HL61006 [D.W.B.]), The Ethel Brown Foerderer Fund for Excellence (E.G.), The University of Pennsylvania Research Foundation (E.G.), and the General Clinical Research Center of The Children's Hospital of Philadelphia (M01-RR00240 [E.G.]). Present address for Dr. McElhinney: Department of Cardiology, Children's Hospital, Boston, Massachusetts.
- atrial septal defect
- congenital heart defect
- conformation-sensitive gel electrophoresis
- deoxyribonucleic acid
- polymerase chain reaction
- Received December 11, 2002.
- Revision received March 18, 2003.
- Accepted May 30, 2003.
- American College of Cardiology Foundation
- Schott J.J.,
- Benson D.W.,
- Basson C.T.,
- et al.
- Goldmuntz E.,
- Geiger E.,
- Benson D.W.
- Goldmuntz E.,
- Clark B.J.,
- Mitchell L.E.,
- et al.
- Ganguly A.,
- Rock M.J.,
- Prockop D.J.
- Komuro I.,
- Izumo S.
- Azpiazu N.,
- Frasch M.
- Bodmer R.
- Lyons I.,
- Parsons L.M.,
- Hartley L.,
- et al.
- Tanaka M.,
- Chen Z.,
- Bartunkova S.,
- Yamasaki N.,
- Izumo S.
- Biben C.,
- Weber R.,
- Kesteven S.,
- et al.
- Schwartz R.J.,
- Olson E.N.
- Sepulveda J.L.,
- Vlahopoulos S.,
- Iyer D.,
- Belaguli N.,
- Schwartz R.J.
- Thompson J.T.,
- Rackley M.S.,
- O'Brien T.X.
- Kasahara H.,
- Usheva A.,
- Ueyama T.,
- Aoki H.,
- Horikoshi N.,
- Izumo S.
- Schinke M.,
- Litovsky S.,
- Usheva A.,
- Kasahara H.,
- Aoki H.,
- Izumo S.