CHD are the most common major birth defects with an estimated incidence of 10/1000 live births.Reference Payne, Chang, Koenig, Zinn and Garg 1 – Reference Reller, Strickland, Riehle-Colarusso, Mahle and Correa 3 Despite great advances in medical care for these newborns, considerable morbidity and mortality continue to be associated with CHD as it is the leading cause of infant death and illness associated with birth defects.Reference Yang, Chen, Correa, Devine, Mathews and Honein 4 Of the total number of cases of CHD, 20% are due to aneuploidy and other chromosomal syndromes, with the remaining 80% classified as “sporadic” CHD.Reference Bittel, Butler and Kibiryeva 5 Most patients with CHD do not have affected family members and the low recurrence rate suggests polygenic inheritance, but there may also be a high rate of de novo mutations.Reference Erdogan, Larsen and Zhang 6 Other than single-gene analysis, chromosomal microarray testing has been shown to have a high diagnostic yield by detecting the genetic basis of disease due to pathogenic genomic copy number variants. Copy number variants are duplications and/or deletions that cause a change in the gene dosage. If they are found in >1% of the general population, they are considered polymorphisms; however, if present in <1% of individuals, they are more likely to be disease causing.Reference Richards and Garg 7
Array comparative genomic hybridisation (aCGH) is a DNA microarray-based technology that detects submicroscopic copy number variants in the genome in the kilobase range making it a more sensitive modality when compared with traditional karyotype analysis, which has a resolution of only 5–10 Mb.Reference Richards and Garg 7 The increased resolution over conventional karyotyping is therefore at least fivefold and is the major advantage of this molecular technique.Reference Breckpot, Thienpont and Peeters 8 aCGH is widely used in clinical practice and recent studies have demonstrated its ability to detect pathogenic copy number variants in 10–15% of patients with developmental delay, intellectual disability, and multiple congenital anomalies.Reference Menten, Maas and Thienpont 9 In addition, Reddy et alReference Reddy, Page and Saade 10 showed an increase of 41.9% in diagnosis of clinically important deletions and duplications by aCGH in stillbirths.
Previous studies have indicated a high rate of chromosomal abnormalities occuring in children with CHD. Bachman et alReference Bachman, Deward, Chrysostomou, Munoz and Madan-Khetarpal 11 examined 45 CHD patients by karyotype analysis and aCGH and identified an additional 22.2% of patients with copy number variants, suggesting aCGH should be used as a first-tier test for neonates with CHD. Warburton et alReference Warburton, Ronemus and Kline 12 analysed 223 families with at least one child affected by conotruncal or hypoplastic left heart disease and found a substantially higher rate of de novo copy number variants in probands with CHD than in control families (9 versus 2%). Further work by Carey et al compared long-term outcomes in patients with single ventricles with and without copy number variants. Children with CHD had 10% more rare copy number variants who in turn had worse neurocognitive and growth outcomes at 14 months of age.Reference Carey, Liang and Edwards 13
Even with increasing evidence of improved detection of chromosomal imbalances with aCGH, many questions remain unanswered. Of the individuals evaluated by aCGH, are those with CHD at higher risk of abnormal results? And do specific copy number variants correspond to specific heart defects? The purpose of this study was to identify aCGH abnormalities in patients with CHD and compare them with a population without CHD. We also wished to determine whether there were common copy number variants in CHD, which may suggest involvement of novel genes in heart disease or development.
Methods
This was a retrospective cohort study of all patients who underwent aCGH between January, 2009 and March, 2011 at the University of Alabama at Birmingham. The Institutional Review Board at The University of Alabama at Birmingham approved the study. The University of Alabama at Birmingham is a regional referral centre and the only paediatric cardiac surgery centre in the state of Alabama. There was an estimated 60,000 live births per year in Alabama during the time of the study of which ∼2.3/1000 required some sort of cardiac intervention in the 1st year. Therefore, it is estimated that each year 138 children in the state of Alabama would require cardiac intervention within the 1st year of life. We collected aCGH results from a database maintained by the University of Alabama at Birmingham Department of Genetics and clinical diagnostic information regarding their heart disease and other conditions from electronic medical records. Informed consent was waived as no personal identifying information was collected.
All individuals who underwent an aCGH for any reason during the time of the study were evaluated. Individuals with CHD were compared with individuals without heart disease. The majority of aCGH analyses were performed on patients who also had neurodevelopmental problems and/or congenital anomalies. These additional diagnoses are listed in Table 1 along with cardiac diagnosis and results of the aCGH. The diagnosis of CHD was confirmed when it was made by a cardiologist with echocardiographic confirmation. As this is a retrospective study, the diagnoses were obtained from medical records and were made by referring physicians during clinical visits. Patients were only excluded if test results or medical information was not available. Patients with CHD were placed into one of five categories based on the nature of structural anomalies, including left-sided heart disease – aortic valve anomalies, hypoplastic left heart syndrome, and coarctation of the aorta; endocardial – atrial septal defect, ventricular septal defect, and atrioventricular canal defect; conotruncal – tetralogy of Fallot, transposition of the great arteries, double outlet right ventricle, interrupted aortic arch, and truncus arteriosus; patent ductus arteriosus; and complex/other heart disease – pulmonic stenosis/atresia, total anomalous pulmonary venous return, and more complex defects such as complex atrioventricular canal and transposition of the great arteries, and complex double outlet right ventricle and coarctation of the aorta.Reference Warburton, Ronemus and Kline 12
Table 1 Cardiac patients who had abnormal array results with their specific abnormality and CHD.
aCGH=array comparative genomic hybridisation; ASD=atrial septal defect; AV=atrioventricular; MAPCAS=major aortopulmonary collaterals; MPA=main pulmonary artery; PDA=patent ductus arteriosus; PFO=patent foramen ovale; SVC=superior vena cava; VSD=ventricular septal defect
High-resolution whole-genome aCGH analysis was performed using the 4×44k Agilent oligo-array (Agilent Technologies, Santa Clara, California, United States of America). This is a custom-designed array that is based on the International Standards for Cytogenomic Arrays consortium design. DNA was extracted from the patients’ peripheral blood using the Qiagen blood mini kit (Qiagen, Valencia, California, United States of America). DNA labelling, slide hybridisation, washing, and scanning were performed following the manufacturer’s protocol. The arrays were scanned using the Agilent high-resolution microarray scanner (Agilent Technologies). The scanned arrays were analysed using the “Feature Extraction v9.5” and “DNA Analytics v4.0” software (Agilent Technologies). All genomic breakpoints were mapped using the UCSC genome browser using human genome build 36 (NCBI36/hg18). Interpretation of abnormal aCGH results was carried out according to the American College of Medical Genetics standards and guidelines for interpretation and reporting of postnatal constitutional copy number variants. Copy number variants were classified into benign, variant of uncertain clinical significance, likely benign, variant of uncertain significance, likely pathogenic, and pathogenic.Reference Kearney, Thorland, Brown, Quintero-Rivera and South 14
Results
A total of 1309 patients underwent aCGH analyses between January, 2009 and March, 2011. A total of 57 patients were excluded from the study because of inadequate records or inability to match to medical records. Of the remaining 1252 patients who were analysed, there were 173 patients who had a diagnosis of CHD. The diagnosis of CHD ranged from simple defects such as a ventricular septal defect or valvular stenosis to commonly described defects such as tetralogy of Fallot or transposition of the great vessels. Some patients had more complex and mixed structural CHD.
Many patients with CHD were diagnosed with known syndromes. Of the patients, eight had well-described cytogenetic syndromes: three patients with Williams syndrome, two patients with CHARGE syndrome, one patient with DiGeorge syndrome, one patient with trisomy 13, and one patient with Wolf–Hirschhorn syndrome. In addition, several patients were ultimately diagnosed with other syndromes or single-gene disorders such as campomelic dysplasia, Smith–Lemli–Opitz, Sotos, Noonan, Alagille, and Beckwith–Wiedemann syndromes. Of the 173 patients with identified CHD, 151 had other clinical problems including developmental delays, dysmorphisms, or congenital malformations, as described in Table 1. Therefore, only 22 patients could be considered as having isolated CHD. Of those 22 patients, only three had abnormal array results and they were all duplications: one patient with supravalvular aortic stenosis had a duplication at 5p14.1; a male infant with aortic hypoplasia, ventricular septal defect, and an anomalous right subclavian artery had a duplication at Xp27.1; and an infant with d-transposition, ventricular septal defect, and coarctation of the aorta had a duplication at 6q11.1.
Of the 1252 total patients included in the study, 335 (26.76%) had abnormal aCGH results. Of the 173 patients with identified CHD, 50 (28.9%) had abnormal aCGH results. Of the 50 cardiac patients with abnormal aCGH, 34 had also been examined by traditional karyotype analysis, and 12 of the 34 karyotypes were abnormal including six translocations, one ring chromosome, two deletions, one duplication, one marker chromosome, and one trisomy (Table 1). Although it is well documented in the previous literature that aCGH is more sensitive than traditional karyotype in identifying copy number variants, our cohort did not have enough karyotypes performed to have statistical evidence to support this. Of the 1079 patients without a diagnosis of CHD, 284 (26.3%) had abnormal aCGH results, not statistically different from those with CHD.
The cardiac patients were analysed by category with the highest percentage of abnormal aCGH results being in the left heart disease group with 14/31 (45.16%), and lower percentages in other categories (p=0.03 by χ2 compared with all other categories combined; Table 2).
Table 2 Diagnostic yield of aCGH analysis in patients with various categories of CHD.
aCGH=array comparative genomic hybridisation
To determine whether any of the aCGH abnormalities spanned regions known to be involved in cardiac development, we evaluated chromosomal regions that span genes known to be involved in cardiac development or disease. There were three patients in this study who were noted to have copy number variants involving two of these genes. TBX5 (12q24.2) encodes a protein called T-box 5 that plays an important role during embryonic development. T-box 5 is a transcription factor that activates genes involved in the normal development of the upper limbs and the heart, specifically the formation of the ventricular septum and the electrical system.Reference Stennard and Harvey 15 This area was duplicated in a patient with a right ventricular dominant atrioventricular canal, a small left ventricle and hypoplastic transverse aorta, and transposed great vessels with anterior main pulmonary artery. She also had 11 rib pairs with the fusion of T5–T6 and dysplastic vertebrae. There were two patients who had abnormalities at 8p23.1 that span the GATA4 gene. GATA4 is a zinc finger transcription factor thought to regulate genes involved in embryogenesis and in myocardial differentiation and function. Mutations in this gene have been associated with cardiac septal defects.Reference Wang, Sun and Qiao 16 A patient had an 8p23.1 duplication that spans GATA4 with multiple aCGH abnormalities: arr 8p23.3p23.1(181,530-6,867,773)x3, 8p23.1(10,856,762-12,448,433)x3, and 18q23(73,323,973-76,111,164)x1, with the phenotype of pulmonic stenosis, a hypoplastic right ventricle, and aortic stenosis. Another patient with a partial atrioventricular canal and left superior vena cava had an 8p23.1 deletion: arr 8p23.1(11,623,512-11,653,694)x1 that does not include GATA4. She also had a congenital diaphragmatic hernia, frontal bossing, high nasal bridge, and mild hypotonia.
Other genes implicated in cardiac disease were reviewed. ZIC3 is a gene that encodes a putative zinc finger transcription factor whose cytogenetic location is at Xq26.2 and has been implicated in heterotaxy and isolated CHD.Reference Cowan, Tariq and Ware 17 NK Homeobox 5 (NKX2-5) located at 5p35.1 is essential in cardiac development, and mutations cause various congenital heart malformations.Reference McElhinney, Geiger, Blinder, Benson and Goldmuntz 18 GATA6 (cytogenetic location 18q11.2) encodes for a transcription factor, which is broadly expressed in the developing heart and is critical for normal cardiac morphogenesis.Reference Maitra, Koenig, Srivastava and Garg 19 LEFTY located at 1q42.12 is implicated in left/right axis malformation and has been demonstrated in mouse models.Reference Kosaki, Bassi and Kosaki 20 CRELD1 is located at 3p25.3 and has been implicated in atrioventricular septal defects.Reference Ghosh, Bhaumik and Ghosh 21 NODAL is also involved in left- to right-axis malformation and heterotaxy.Reference Mohapatra, Casey and Li 22 TBX20 located at 7p14.2 is involved in morphology of the heart and has been implicated in atrial septal defects and tetralogy of Fallot.Reference Kirk, Sunde and Costa 23 No cardiac patients in this study had array abnormalities that correspond to these cytogenetic loci.
Patients with abnormal aCGH results but no CHD are listed in Table 3. Common microdeletions and microduplications such as 22q11 and 15q11 were observed in this population as well. However, the non-cardiac group had many results involving the sex chromosomes X and Y. There was no significant overlap in abnormal results between the cardiac and non-cardiac groups. A non-cardiac patient with preaxial polysyndactyly of the great toes, frontal bossing, and developmental delay had a deletion in the cytogenetic location of TBX20.
Table 3 Non-cardiac patients who had abnormal array results.
aCGH=array comparative genomic hybridisation; ADD=attention deficit disorder; ADHD=attention deficit hyperactivity disorder; ARPKD=autosomal recessive polycystic kidney disease; BPD=bronchopulmonary dysplasia; CCAM=congenital cystic adenomatoid malformation; CLCP=cleft lip and cleft palate; CP=cleft palate; FISH=fluorescence in situ hybridisation; FTT=failure to thrive; GH=growth hormone; HIE=hypoxic-ischemic encephalopathy; IUGR=intrauterine growth restriction; MPLA=multiplex ligation-dependent probe amplification; MR=mental retardation; OCD=obsessive-compulsive disorder; OTC=ornithine transcarbamylase; PDD=pervasive developmental disorder; SGA=small for gestational age; VCFS=velocardiofacial syndrome; WGA=whole-genome amplification
Discussion
Our study demonstrates the utility of aCGH in CHD in a reasonably large single-centre cohort. In our cohort, although patients with CHD had no more copy number variants than patients without CHD, patients with left-sided heart disease had an increased rate of microdeletions and microduplications. We also identified specific aCGH anomalies associated with CHD in this cohort that may enable future mechanistic studies of abnormal cardiac development.
aCGH is capable of identifying submicroscopic chromosomal abnormalities that may be missed by traditional karyotype. It has been increasingly used to detect abnormalities in individuals with developmental delay, intellectual disability, autism spectrum disorders, and/or multiple congenital anomalies.Reference Miller, Adam and Aradhya 24 The G-banded karyotype is the classic test for cytogenetic abnormalities, but it may miss imbalances in the 5–10 Mb range. It is also limited because of interpersonal and interlaboratory variation in detection rates. aCGH is more sensitive at identifying pathologic copy number variants owing to the higher resolution. The increased resolution allows for increased detection of disease-causing genes. In a sample of 532 stillbirths, aCGH provided better detection of genetic abnormalities (8.3% by aCGH versus 5.8% by karyotype; p=0.007).Reference Reddy, Page and Saade 10
aCGH may permit the identification of abnormalities in genes known to be involved in cardiac development, and help in identifying possible novel genes involved in CHDs. In our study, of the 50 patients with CHD and abnormal aCGH results, three had abnormalities involving genes known to be involved in cardiac development (TBX5 and GATA4). Several novel genes that may possibly be involved in CHD were identified. A patient with tetralogy of Fallot, developmental delay, seizure disorder, and mandibular crowding had the following array abnormality: arr 16p12.1(21,744,793-22,315,573)x1. When examining the deleted region with the USCS Browser (http://www.genome.ucsc.edu/cgi-bin/hgGateway), multiple genes, such as OTOA, UQCRC2, VWA3A, EEF2K, and POLR3E, were identified in the deleted area, but none of them have found to be associated with CHD.Reference Kent, Sugnet and Furey 25
Another patient with a partial atrioventricular canal, left superior vena cava, congenital diaphragmatic hernia, frontal bossing, high nasal bridge, and mild hypotonia had the following deletion: arr 8p23.1(11,623,512-11,653,694)x1. This location corresponds to the NEIL2 gene (Homo sapiens nei endonuclease VIII-like 2), which belongs to a class of DNA glycosylases homologous to the bacterial Fpg/Nei family (MIM 608933). These glycosylases initiate the first step in base excision repair by cleaving bases damaged by reactive oxygen species and introducing a DNA strand break via the associated lyase reaction. This gene has been found to be important in predisposition to certain cancers.Reference Bandaru, Sunkara, Wallace and Bond 26 There were five copy number variants at this locus, four deletions and one duplication, also described by Soemedi et al.Reference Soemedi, Wilson and Bentham 27
Soemedi et alReference Soemedi, Wilson and Bentham 27 described an association between recurrent 15q11.2 deletions and those with CHD, but no genes in the region have been previously associated with CHD. Our population with CHD had only one patient with 15q11 deletion, but three with a duplication of the locus. Greenway et alReference Greenway, Pereira and Lin 28 described recurrent copy number variants in patients with tetralogy of Fallot, including copy number variants at 1q21.1, 3p25.1, 7p21.3, and 22q11.2. In our study, no patients were identified with copy number variants involving 3p25.1 or 7p21.3. There was one patient with a deletion of 1q21.1 that had an unbalanced atrioventricular canal and heterotaxy. There were three patients with copy number variants involving 22q11.2, of which two were duplications and one was a deletion. The patient with a deletion had non-syndromic tetralogy of Fallot. The one with duplication had an atrial septal defect, ventricular septal defect, left micropthalmia, and colobomas. The other had coarctation of the aorta along with developmental delay and a diagnosis of autism.
The utility of routine aCGH analysis in all patients with CHD is currently unclear. Richards and Garg performed aCGH on 40 individuals, 20 of whom had CHD and other anomalies or developmental delay (syndromic CHD), and the others had isolated CHD. The risk of having disease-causing copy number variants was 45% in syndromic CHD, compared with none with isolated CHD.Reference Richards and Garg 7 Similarly, Richards et al recruited 20 children with CHD and additional birth defects and compared them with 20 children with isolated CHD. They detected a 25% rate of copy number variants in the population with additional defects – mostly in those with neurologic defects – but none in the population with isolated CHD. This study advocates for screening of copy number variants in children with CHD and a neurologic abnormality.Reference Richards, Santos and Nichols 29 In addition, Erdogan et alReference Erdogan, Larsen and Zhang 6 described copy number variants in 105 individuals with isolated CHD. They only detected 18 rare copy number variants and the majority were duplications. Therefore, although rare copy number variants may play a role in isolated CHD, aCGH may not be clinically indicated.Reference Erdogan, Larsen and Zhang 6 , Reference Richards and Garg 7
In contrast, Thienpont et al examined 60 patients with CHD and either a second major malformation and/or mental handicap who could not be diagnosed genetically with clinical exam and karyotyping. Of the patients, 30% carried imbalances that are not described in normal individuals.Reference Thienpont, Mertens and de Ravel 30 Breckpot expanded this study to an additional 90 patients and found that 31% had potentially significant imbalances.Reference Breckpot, Thienpont and Peeters 8
In our study, only 22 patients could be described as having isolated CHD, and of those only three had an abnormality detected by aCGH. Similar to the study conducted by Ergogan, all three patients had duplications.Reference Erdogan, Larsen and Zhang 6
Even with greater detection of variants, many are of unclear significance and thus require additional investigation and provide limited clinical information.Reference Bachman, Deward, Chrysostomou, Munoz and Madan-Khetarpal 11 Hitz et al compared families with CHD with controls. They identified 73 unique copy number variants in 54 individuals in the left-sided CHD cohort, suggesting that unique copy number variants contribute significantly to left-sided CHD.Reference Hitz, Lemieux-Perreault and Marshall 31 Our finding that patients with left-sided heart disease had a higher rate of aCGH abnormalities is consistent with this study, and suggests that aCGH may have a higher diagnostic yield in patients with left-sided CHD, and perhaps of little benefit in isolated CHD patients.
Our study is unique in the large number of congenital heart patients who were evaluated, a total of 173 individuals. This number is, however, relatively small considering the wide range of anomalies that can be observed, both in the clinical phenotype and in the aCGH array results. A multi-centre analysis of aCGH data may help identify microdeletions and microduplications containing important genes in cardiac development. A limitation is that there were no real controls because none of the patients were “normal”. All patients, either with or without CHD, had a clinical indication for an aCGH. In addition, there were no familial samples as this was a retrospective study. In addition, there was considerable selection bias. Only patients who underwent an aCGH were included. Therefore, patients who may have been diagnosed by other means, such as fluorescence in situ hybridisation, were not included. This may have removed patients with deletions of 22q11 locus from the study. This is important because 22q11 is recognised as being a frequent genetic cause of CHD. There are also potential systemic biases. All patients with cardiac disease who are admitted to the University of Alabama at Birmingham regional neonatal intensive care unit have an aCGH ordered routinely. However, patients transferred in from other hospitals would be missed by the study if they had an aCGH sent to another institution. Moreover, if infants were diagnosed with CHD after the newborn period, they may not have had an aCGH ordered unless a geneticist saw them as an outpatient.
Even with greater detection of genetic abnormalities in CHD, 70% of the patients remain without any identifiable genetic abnormalities. These may be the result of an undiscovered single-gene defect or combination of multiple gene defects, or the result of epigenetic alterations due to environmental effects. Next-generation sequencing may be an alternative modality for individuals with CHD and/or other anomalies. One study tested 250 probands, of which 62 carried 86 mutant alleles that satisfied criteria for a molecular diagnosis giving an overall rate of positive molecular diagnosis of 25%. This higher diagnostic yield supports the use of whole-exome sequencing as a diagnostic test. Questions about cost-effectiveness, accuracy, yield, and integration into clinical care, however, need to be addressed in future studies.Reference Yang, Muzny and Reid 32
Conclusion
aCGH analysis may have a role as a first-tier test for individuals with left-sided CHD, or CHD with neurodevelopmental problems and/or multiple congenital anomalies. However, our study shows that it does not have a significant diagnostic yield in patients with isolated CHD. A prospective study of aCGH in patients with CHD that is not confounded by selection bias may clarify the role of copy number variants in patients with isolated CHD. In addition, other modalities such as whole-exome sequencing may have value in identification of candidate genes associated with CHD.
Acknowledgements
None.
Financial Support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Conflicts of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethical standards of National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research and with the Helsinki Declaration of 1975, as revised in 2008, and has been approved by the institutional review board at the University of Alabama at Birmingham.
