|Home | About | Journals | Submit | Contact Us | Français|
The results of 4 successful genome-wide association studies have been reported recently, yielding 4 distinct loci, ORMDL3, CHI3L1, PDE4D, and DENND1B,1–4 with the disparity in the results likely caused by differences in the phenotypes studied. In an attempt to dissect further the genetic network in complex asthmatic phenotypes, we investigated allergic asthma specifically defined in children at age 6 years. Cases were patients with allergic asthma with early-onset persistent asthma defined as having physician diagnosed asthma and wheeze at ages 1 and 6 years and asthma medication use and allergies as reported by the mother; controls were defined as those having no diagnosis of asthma or reactive airway disease and no symptoms at ages 1 or 6 years and no asthma medication use and no allergies (see Definition of Cases and Controls section in the Online Repository at www.jacionline.org for additional recruitment details and Table E1 for demographic and phenotype data). Sixty-six cases and 42 controls were successfully genotyped and analyzed by Affymetrix Genome-Wide Human Single Nucleotide Polymorphism (SNP) Array 5.0 (Affymetrix, Santa Clara, Calif) after whole-genome amplification. To test for potential population stratification, we calculated a genomic inflation factor for this study of 1.03 by using EIGENSTRAT,5 indicating only minimal background stratification (see Statistical Analysis section in the Online Repository at www.jacionline.org).
No SNP achieved genome-wide significance using strict Bonferroni criteria for multiple testing (0.05/396,207 = 1.26 × 10−7) as the threshold of significance (see this article’s Fig E1 in the Online Repository at www.jacionline.org). Overall, there were 39 SNPs with P values <1 × 10−4 for either the allelic or genotypic test, with 11 of these 39 SNPs forming 3 clusters (2 or more SNPs) in 3 discrete coding regions (Table I; see this article’s Table E2 in the Online Repository at www.jacionline.org). Of the 5 SNPs selected, 4 of them were successfully genotyped in the larger collection of case (n = 104) and control (n = 503) samples (All Perinatal Risk of Asthma in Infants with Asthmatic Mothers [All PRAM]) that included the original case (n = 66) and control (n = 42) samples. The P values for these SNPs did not change substantially from their initial significance levels, with all P values within an order of magnitude from their initial value from the genome-wide association study (GWAS; Tables I and E2).
Notably, 1 SNP, rs11684634, achieved a P value of 8.9 × 10−7 at the first stage GWAS under the allelic association test (by Fisher exact test), which was reduced to 3.82 × 10−5 on regeno-typing of this SNP on the same subjects by using unamplified DNA via a TaqMan assay. Several inconsistent genotype calls were observed between the 2 processes (see TaqMan Genotyping section in the Online Repository at www.jacionline.org). Nevertheless, judging from the result in the extended cohort (P = 3.4 × 10−2; All PRAM minus the samples in the original GWAS), we regarded this SNP as potentially interesting. SNP rs11684634 is located on chromosome 2 in an intron within the PDE11A gene (Entrez Gene ID 50940). We resequenced all unique exons of the 4 isoforms of PDE11A in a subset of the GWAS cases (n = 50) on the basis of their rs11684634 genotype and all controls (n = 42). We identified 123 variants (114 SNPs and 9 insertions/deletions; see this article’s Table E3 in the Online Repository at www.jacionline.org) in the 48 fragments that were sequenced to cover PDE11A. Four of the variants, located in 2 regions of PDE11A, were potentially in a high degree of linkage disequilibrium with rs11684634. Two of the 4 variants were then successfully genotyped in the extended cohort (Table I).
To attempt to replicate the association to PDE11A, we took a gene-centric approach6 and queried 5 existing GWASs of asthma phenotypes for evidence for association with SNPs within PDE11A. Three of the 5 datasets had at least 1 SNP with a P value <.05 within PDE11A (Table II), with only the MRC-A/UK-C dataset not yielding a nominally significant SNP. When rs11684634 was imputed for the British 1958 Birth Cohort (B58C) dataset, a nominally significant result (P = .046; Table II) was obtained for this cohort among cases with asthma with onset ≥17 years of age and had the same direction of association as in the PRAM cohort. SNP rs11685634 was not significant among childhood-onset patients with asthma in the B58C cohort (Table II). The combined P value for rs11684634 for All PRAM and B58C adult-onset cases by using the Fisher method was 6.0 × 10−5.
The report of an association with asthma to PDE4D3 lends credence to our finding of an association within PDE11A, another member of the phosphodiesterase superfamily of genes. Furthermore, the same cohorts exhibiting an association to PDE4D also show nominal associations to SNPs within PDE11A, suggesting that multiple loci throughout this superfamily of genes may contribute to the susceptibility of asthma. Several phosphodiesterase (PDE)–4 inhibitors have been found to be effective in suppressing inflammation and have been developed as potential therapies for chronic obstructive pulmonary disease and asthma.7 PDE3 and PDE7 are also expressed in inflammatory cells, and PDE5 was initially used to treat patients with exercise-induced asthma and showed bronchodilator effects.8,9
One major limitation of the current study is the fact that the initial discovery sample used for the GWAS is underpowered to detect loci with small effect size. We have attempted to address this issue by validating the PDE11A association in a larger set of cases and controls, albeit from the same study population, as well as compare our case allele frequencies to additional sets of publicly available control samples (see Table E4 in the Online Repository at www.jacionline.org), all of which validated the original findings. Furthermore, we examined the evidence for association to SNPs within PDE11A in 5 independent case/control populations, of which 3 had at least 1 SNP with a nominally significant P value. The finding of a nominal association in the B58C cohort with rs11684634 after imputation is encouraging that this SNP is of interest, and further work needs to be done to replicate this association in additional datasets. However, these results must be interpreted with caution because the association was observed only among adult-onset patients with asthma and not among the childhood-onset cases in the B58C cohort. Further, the observation that 2 additional datasets each had at least 1 SNP within PDE11A with a nominal association suggests the possibility of multiple variants throughout the PDE11A that may be contributing to the susceptibility to asthma, potentially a number of rare variants that we were underpowered to detect in this study. This is also supported by the fact that the SNPs in the replication datasets were not in linkage disequilibrium with rs11684634 (see this article’s Fig E2 in the Online Repository at www.jacionline.org). The lack of overwhelming significance in the replication samples may also be a result of phenotypic heterogeneity among these populations, with both Framingham Heart Study and B58C enrolling adult-onset patients with asthma and none of the replication populations limiting their case populations to the strict phenotypic definition of childhood allergic asthma used in the PRAM study. Polymorphisms within PDE11A may be specific for allergic asthma, but until this specific phenotype is studied in additional populations, this conclusion cannot be drawn. In spite of this, the PDE11A finding is a compelling candidate gene for asthma and worth further investigation.
Supported by grant AI41040 from the National Institutes of Health. The Children’s Health Study is supported by P01 ES011627 and P30 ES007048 from the National Institute of Environmental Health Sciences and R01 HL087680 from the NHLBI. We acknowledge use of genotype data from the British 1958 Birth Cohort DNA collection, funded by the UK Medical Research Council, grant G0000934, and The Wellcome Trust, grant 068545/Z/02. We also acknowledge the use of the Yale University Biomedical High Performance Computing Center and NIH grant RR19895, which funded the instrumentation.
Data generated by the genome-wide association studies of bipolar and schizophrenia from the Genetic Association Information Network (GAIN) were used for the analyses described in this article. The data were obtained from the database of Genotype and Phenotype (dbGaP) found at http://www.ncbi.nlm.nih.gov/gap through dbGaP accession numbers phs000017.v3.p1 and phs000021.v2.p1 for bipolar and schizophrenia, respectively.
Disclosure of potential conflict of interest: The authors have declared that they have no conflict of interest.