Conceived and designed the experiments: SBG PJG. Performed the experiments: KNS CEK AG. Analyzed the data: KNS VM SBG PJG. Contributed reagents/materials/analysis tools: HH CEK PAD BPCK SM JTG JC VM PJG. Wrote the paper: KNS SBG PJG. Recruited patients: JR. Originated and designed the project, analyzed the results: PJG. Co-wrote the manuscript: PJG KNS SBG. Provided guidance, design, and support throughout the project: KNS SBG. At the Center for Applied Genomics performed the genotyping and control populations providing selected data to PJG and SBG: HH CEK JTG. Supervised the data analysis: JTG VM SBG. Performed all of the sequencing, custom genotyping and analytic programming: KNS. Coordinated the clinical collection of samples: PJG JR. Was involved in preliminary experiments: AG. Provided samples from Toronto: SM JGC. Provided samples from the Netherlands: PAD BPCK.
The authors have declared that no competing interests exist.
Congenital heart disease (CHD) is the most common birth abnormality and the etiology is unknown in the overwhelming majority of cases. ISLET1 (ISL1) is a transcription factor that marks cardiac progenitor cells and generates diverse multipotent cardiovascular cell lineages. The fundamental role of ISL1 in cardiac morphogenesis makes this an exceptional candidate gene to consider as a cause of complex congenital heart disease. We evaluated whether genetic variation in
CHD affects 1 in 20 live births, 1 in 100 of which require an intervention
During vertebrate cardiac development, the 3-dimensional structure of the heart is formed from the differentiation and interaction of multiple tissue derivatives, or fields
We conducted a two-stage candidate gene study to test the hypothesis that germline common genetic variants in
We analyzed 30 SNPs spanning a 237 kb region around
a) Analysis of SNP data within and surrounding
Three SNPs analyzed in stage 1 (rs3762977, IVS1+17C>T, rs1017) were located within the
Genotypes | Controls [n (%)] | Cases [n (%)] | OR [95% CI] | |
A/A | 329 (75.3) | 65 (79.3) | 1.00 | |
A/G | 102 (23.3) | 15 (18.3) | 0.74 (0.41/1.36) | 0.338 |
G/G | 6 (1.4) | 2 (2.4) | 1.68 (0.33/8.55) | 0.527 |
log-additive: | 0.87 (0.52/1.47) | 0.607 | ||
C/C | 402 (93.1) | 70 (85.4) | 1.00 | |
C/T | 30 (6.9) | 12 (14.6) | 2.30 (1.12/4.70) | 0.023 |
A/A | 182 (42.8) | 21 (25.3) | 1.00 | |
A/T | 192 (45.2) | 43 (51.8) | 1.94 (1.11/3.40) | 0.020 |
T/T | 51 (12.0) | 19 (22.9) | 3.23 (1.61/6.46) | 0.0009 |
log-additive: | 1.81 (1.29/2.54) | 0.0007 | ||
A/A | 1128 (18.1) | 202 (76.2) | 1.00 | |
A/G | 289 (20.0) | 58 (21.9) | 1.12 (0.82/1.54) | 0.481 |
G/G | 28 (1.9) | 5 (1.9) | 1.00 (0.38/2.62) | 0.997 |
log-additive: | 1.08 (0.82/1.42) | 0.571 | ||
C/C | 1334 (92.3) | 246 (92.8) | 1.00 | |
C/T | 111 (7.0) | 19 (7.2) | 0.93 (0.56/1.54) | 0.773 |
A/A | 591 (40.9) | 91 (34.3) | 1.00 | |
A/T | 672 (46.5) | 128 (48.3) | 1.24 (0.93/1.66) | 0.148 |
T/T | 182 (12.6) | 46 (17.3) | 1.64 (1.12/2.43) | 0.013 |
log-additive: | 1.27 (1.05/1.54) | 0.013 |
We then delineated the patterns of risk in these populations by using the expectation maximization (EM) method to estimate haplotypes and risk of CHD from the 6
To independently validate our findings, we studied
Haplotype analysis in the stage 2 white population confirmed that the A-C-T haplotype is significantly associated with risk among whites (Global
a) The A-C-T risk haplotype in white stage 1 (US) and stage 2 (US, Canada, Netherlands) populations. Odds ratios (95% CIs) for each stage are denoted by black boxes (gray lines). Summary OR estimates are represented by black diamonds, where diamond width corresponds to 95% CI bounds. Box and diamond heights are inversely proportional to precision of the OR estimate. b) The G-C-T risk haplotype in black/African American stage 1 (US) and stage 2 (US) populations. Odds ratios (95% CIs) are denoted as in 2a.
Haplotypes | rs3762977 | IVS1+17C>T | rs1017 | Frequency (%) | OR [95% CI] | |
1 | A | C | A | 0.645 | 1.00 | |
2 | A | C | T | 0.192 | 1.27 (1.09/1.48) | 0.0018 |
3 | G | C | T | 0.098 | 1.07 (0.88/1.30) | 0.5068 |
4 | A | T | T | 0.034 | 1.04 (0.75/1.44) | 0.8216 |
5 | G | C | A | 0.022 | 1.10 (0.78/1.53) | 0.5928 |
Global haplotype association | 0.000004 |
Rare estimated haplotypes (cumulative frequency = 0.0099) not shown.
The precise distribution of CHD diagnoses was different between stage 1 and stage 2 populations (
To understand the role of
To independently validate our findings in blacks/African Americans, we analyzed a distinct set of 49 US black/African American cases and 1,845 US black/African American controls. In this stage 2 population, the relative risk for rs3762977 was consistent with that seen in stage 1 blacks/African Americans (OR = 1.20, 95% CI 0.74–1.95,
Our results demonstrate that two different
The biologic rationale is compelling: ISL1 is a transcription factor that marks cardiac progenitor cells and controls secondary heart field differentiation, and new evidence suggests that purified populations of ISL1+ progenitor cells are capable of self-renewal and expansion into cardiomyocytes, smooth muscle, and endothelial lineages
United States cases and controls were recruited from the Children's Hospital of Philadelphia (CHOP) between 12/12/2003 and 08/25/2008 on a protocol approved by the Institutional Review Boards of CHOP and the University of Michigan, and parents provided written informed consent. 31.6% (613/1939) of all eligible cases seen at the CHOP cardiac center in this time period participated in this study. Cases and controls from Toronto and the Netherlands were recruited on institution-specific protocols, and were also approved by the IRBs of CHOP and the University of Michigan. Cases were children with complex congenital heart disease requiring surgical repair. Controls were patients without congenital heart disease recruited through either the CHOP Health Care Network by CHOP clinicians and nursing staff or through UMC Utrecht. The controls were screened by counselors who evaluated by patient history for a absence of structural heart disease. All patients had been evaluated by a medical doctor. Ethnicity for cases was determined by self-report and principal components analysis, and ethnicity for controls was determined by principal components analysis. This analysis was performed for all stage 2 subjects of unknown ethnicity as well as a representative subset of stage 1 cases and controls of known ethnicity (
Stage 1 and stage 2 genotypes were requested for 27
Genotypes for the 3 SNPs within
Genotyping accuracy by sequencing was assessed with repeat sequencing of a subset of genotypes. Two measures were used to assess imputation error for the three
Single SNP analyses were conducted using unconditional logistic regression to calculate odds ratios as implemented in SAS (version 9.1). Haplotypes were estimated and tested for association with CHD using the Haplo.stats package in R (
We performed additional analyses to determine the sensitivity of our previously described method of defining ethnicity by principal components analysis in stage 2. The alternative classification method we employed was the ANCESTRYMAP
Diagnosis distribution in stage 1 and stage 2 case-control studies. Cases were chosen a priori to represent a wide variety of developmental phenotypes that include developmental structures aberrantly formed as derivatives of the secondary heart field. These diagnostic choices were informed from lineage tracing analyses of Isl1+ progenitor cells in rodents. ALCAPA- anomalous left coronary artery from the pulmonary artery, ASD- atrial septal defect, AVSD- atrioventricular septal defect/AV canal, CCTGA- congenitally corrected transposition of the great arteries, L-TGA- L-transposition of the great arteries, D-TGA- D- transposition of the great arteries, DORV- double outlet right ventricle, HLHS- hypoplastic left heart syndrome, AS-aortic stenosis, AA- aortic atresia, MS- mitral stenosis, MA- mitral atresia, DILV- double inlet left ventricle, TAPVC- total anomalous pulmonary venous connection, TOF- tetralogy of Fallot, VSD- ventricular septal defect.
(0.50 MB EPS)
Chromosome 5 variation in the ISL1 region. The location of ISL1 on chromosome 5 (Build 36) is depicted, where exons of the ISL1 gene are depicted as shaded boxes, the 5′ UTR and 3′ UTR are depicted as white boxes, and introns are represented as black lines. The three SNPs within ISL1 studied in stage 1 and stage 2 are depicted with respect to their location in the gene. The six SNPs flanking ISL1 identified as significantly associated with risk of CHD in stage 1 are indicated along chromosome 5.
(0.42 MB EPS)
Ethnic distribution of cases and controls by cluster analysis. The first two principal components from a principal components analysis utilizing all SNPs on chromosome 5 that are contained within the Illumina HumanHap550 array are plotted for a) stage 1 cases of known ethnicity, where PC1≤0.025 captures white cases and PC1>0.025 captures black/African American cases; b) stage 2 cases of unknown ethnicity, where PC1≤0.025 defines white cases and PC1>0.025 defines black/African American cases; c) stage 1 controls of known ethnicity, where PC1≤0.0059 captures white controls and PC1>0.0059 captures black/African American controls; b) stage 2 cases of unknown ethnicity, where PC1≤0.0059 defines white controls and PC1>0.0059 defines black/African American controls.
(2.84 MB EPS)
ANCESTRYMAP admixture estimation using 26 Ancestral Informative Markers. The distribution of estimated percent European ancestry for a) all stage 1 US subjects of known ethnicity (n = 650), b) stage 1 US whites (n = 251), c) stage 2 US blacks/African Americans (n = 399), and d) stage 2 US subjects of unknown ethnicity (n = 3610). 65% cutoff is represented by a red line. Individuals above 65% European ancestry were defined as white in stage 2 US subjects and below 65% were defined as black/African American.
(0.53 MB EPS)
PCR primers & conditions for ISL1 sequencing.
(0.05 MB DOC)
Stage 1: ISL1 variation identified by sequencing.
(0.08 MB DOC)
Ancestral informative markers available for stage 2 subjects.
(0.05 MB DOC)
ISL1 associations with risk of HLHS and TGA in white populations.
(0.03 MB DOC)
CHD diagnoses among whites with ACT haplotype.
(0.07 MB DOC)
CHD diagnoses among blacks/African Americans with GCT haplotype.
(0.05 MB DOC)
We would like to thank Thomas L. Spray for his support of this project; N. Burnham, R. Chiavacci, C. Hamatake, D. Hartman, and J. Hufford for management of patient samples. Additional control samples from the Netherlands were provided by R.A. Ophoff.