A Genome-Wide Association Study of Post-bronchodilator Lung Function in Children with Asthma.

To the Editor: Three meta-analyses of genome-wide association studies (GWAS) in healthy subjects have revealed multiple loci for FEV1 and FEV1/FVC (1–3), but only one study included even a small proportion of children (∼7%) (3). A separate GWAS of lung function in children and young adults (including subjects with and without respiratory diseases) did not yield significant results but replicated prior findings in adults (4), and a study in Hutterites identified variants in the 15q23 locus related to lung function (5). The only GWAS of lung function in subjects with asthma (aged 12 years and older, mostly adults) did not show novel results but replicated findings in healthy subjects (6). No prior GWAS has examined post-bronchodilator (post-BD) FEV1 and FEV1/FVC, which may best reflect true lung function (and not solely disease severity) in subjects with asthma. Even though Puerto Ricans share a disproportionate burden from asthma in the United States (7), there has been no GWAS of lung function in this ethnic group. GWAS of minority populations at risk are important from a public health perspective and can identify novel susceptibility variants for well-characterized phenotypes (8). Moreover, loci identified in patients without asthma likely influence lung function in patients with asthma (9), but unidentified variants may also influence lung function in subjects with asthma (10). We conducted a GWAS of understudied lung function phenotypes (post-BD FEV1 and FEV1/FVC) in a unique high-risk population (Puerto Rican children with asthma) and attempted to replicate our top findings in independent cohorts. We also examined whether previous findings for GWAS of pre-BD lung function were replicated for post-BD lung function. From September 2003 to June 2010, 618 children with asthma (physician-diagnosed asthma and one or more episode of wheeze in the previous year) were recruited in Hartford, Connecticut (n = 267), and San Juan, Puerto Rico (n = 351), as described elsewhere (11). At both study sites, eligibility criteria included age 6–14 years and having four Puerto Rican grandparents. All participants completed a protocol that included questionnaires, blood sample collection, and spirometry (conducted with an EasyOne spirometer; NDD Medical Technologies, Andover, MA), following American Thoracic Society recommendations for children (12). After baseline spirometry, subjects were given 200 μg of an albuterol metered-dose inhaler, and spirometry was repeated after 15 minutes. The best FEV1 and FVC from each test were selected for analysis. Of the 618 participants, 560 had blood samples and sufficient DNA for genome-wide genotyping, conducted using the HumanOmni2.5 BeadChip platform (Illumina Inc., San Diego, CA), as previously described (11). After excluding subjects with low marker call rate (n = 31), sex mismatch (n = 6), or missing post-BD phenotypes (n = 76), 447 subjects remained in the analysis. Written parental consent was obtained for participating children, from whom written assent was also obtained. The study was approved by the institutional review boards of the University of Puerto Rico (San Juan, PR), Brigham and Women’s Hospital (Boston, MA), and the University of Pittsburgh (Pittsburgh, PA). Replication was attempted in three cohorts of children with asthma: the Genetics of Asthma in Costa Rica Study (13), the Childhood Asthma Management Program (14), and the Genes-environments and Admixture in Latino Americans study (15). The Genotype-Tissue Expression Biorepository (16) was then used to assess the association between our top-replicated single nucleotide polymorphisms (SNPs) and expression of nearby genes in human tissues. Bronchial epithelial brushings were obtained from healthy nonsmokers: three males and one female (median age, 27 yr; range, 24–40 yr). Expression data were generated using Illumina TrueSeq RNA sequencing. RNAseq data were mapped to the human genome (build 37.2), and relative expression level was assessed by fragments per kilobase of transcript per million mapped reads (17). For the GWAS, we built linear regression models under an additive genetic model, adjusting for age, sex, principal components, and height. The top 20 SNPs were then tested using the same genetic model in the replication cohorts. The results for the replication cohorts were combined in a meta-analysis with weights proportional to the inverse variance of the beta. The main characteristics of study participants are shown in Table 1. Subjects at both study sites had similar age and lung function, but there was a slightly higher proportion of female subjects in Hartford than in San Juan. The GWAS results for both phenotypes did not show significant evidence of population stratification based on genomic inflation factors (see Figure E1 in the online supplement). Eight SNPs were associated with post-BD FEV1 at P < 9 × 10−7 in Puerto Ricans (Table E1 and Figure E2). Of these eight SNPs, two (rs7946574 and rs2984842) were associated with post-BD FEV1, at P < 5 × 10−8 (Figure E3). Results for rs7946574 were not significantly replicated, but findings for rs2984842 were replicated in the same direction of effect in all cohorts (combined P for the analysis in replication cohorts = 0.005) (Figure 1). This SNP, an intronic variant of the gene for fibroblast growth factor 14 (FGF14), is associated with increased expression of FGF14 in subcutaneous adipose tissue (P = 0.01) and lung (P = 0.10) (16). FGF14 was expressed in bronchial brushes from healthy nonsmokers at a level ∼2.5-fold greater than that for the cystic fibrosis transmembrane conductance regulator gene (CFTR) and 7-fold greater than that for the gene for hypoxanthine phosphoribosyltransferase 1 (HPRT1). Table 1. Baseline Characteristics of Participating Children Figure 1. Forest plot of effect sizes on post-bronchodilator FEV1 (in milliliters) and 95% confidence intervals for rs2984842 in discovery and replication cohorts. CAMP = Childhood Asthma Management Program; GACRS = Genetics of Asthma in Costa Rica Study; GALA ... Five SNPs had genome-wide significant (P < 5 × 10−8) associations with post-BD FEV1/FVC in Puerto Ricans (Table E2, and Figures E2 and E4) but were not replicated in other cohorts. We then tested whether SNPs identified by GWAS for pre-BD FEV1 are associated with post-BD FEV1 in Puerto Rican children. We identified four SNPs (tagging eight published SNPs for pre-BD FEV1) that were associated with post-BD FEV1 at P < 0.05 in the 4q24 locus (Table E3). This region contains multiple genes, including GSTCD, INTS12, and NPNT. Our top replicated association for post-BD FEV1 (for rs2984842) is in FGF14, a gene in the fibroblast growth factor family that is most highly expressed in brain tissue; it is also expressed in lung, pituitary, and adipose tissues. Although FGF14 has not been associated with lung phenotypes, it was highly expressed in bronchial brushes from healthy adults, and other members of the FGF family (with which it shows homology) have been associated with lung development or disease, including FGF10 (18) and FGF7 (19). We also replicated previously published findings for pre-BD FEV1 in the 4q24 locus, using the post-BD FEV1 phenotype. Lack of replication of some previously identified variants may reflect lack of statistical power because of the small size of our discovery cohort, but it may also be a result of differences in linkage disequilibrium resulting from racial ancestry. To our knowledge, this is the first GWAS of post-BD lung function and the first GWAS of lung function in Puerto Ricans. We have identified a novel gene associated with post-BD FEV1 in children with asthma. Further studies are needed to identify functional mechanisms.