A genome-wide association study of bronchodilator response in Latinos implicates rare variants

Abstract

Background: The primary rescue medication to treat acute asthma exacerbation is the short-acting β₂-adrenergic receptor agonist; however, there is variation in how well a patient responds to treatment. Although these differences might be due to environmental factors, there is mounting evidence for a genetic contribution to variability in bronchodilator response (BDR).

Objective: To identify genetic variation associated with bronchodilator drug response in Latino children with asthma.

Methods: We performed a genome-wide association study (GWAS) for BDR in 1782 Latino children with asthma using standard linear regression, adjusting for genetic ancestry and ethnicity, and performed replication studies in an additional 531 Latinos. We also performed admixture mapping across the genome by testing for an association between local European, African, and Native American ancestry and BDR, adjusting for genomic ancestry and ethnicity.

Results: We identified 7 genetic variants associated with BDR at a genome-wide significant threshold (P < 5 × 10(-8)), all of which had frequencies of less than 5%. Furthermore, we observed an excess of small P values driven by rare variants (frequency, <5%) and by variants in the proximity of solute carrier (SLC) genes. Admixture mapping identified 5 significant peaks; fine mapping within these peaks identified 2 rare variants in SLC22A15 as being associated with increased BDR in Mexicans. Quantitative PCR and immunohistochemistry identified SLC22A15 as being expressed in the lung and bronchial epithelial cells.

Conclusion: Our results suggest that rare variation contributes to individual differences in response to albuterol in Latinos, notably in SLC genes that include membrane transport proteins involved in the transport of endogenous metabolites and xenobiotics. Resequencing in larger, multiethnic population samples and additional functional studies are required to further understand the role of rare variation in BDR.

Keywords: BDR; Bronchodilator response; GALA I; GALA II; GWAS; Genes-Environments & Admixture in Latino Americans; Genetics of Asthma in Latino Americans; Genome-wide association study; IGF; Insulin-like growth factor; LD; Latinos; Linkage disequilibrium; MAF; Minor allele frequency; QC; Quality control; SABA; SLC; SNP; Short-acting β(2)-adrenergic receptor agonist; Single nucleotide polymorphism; Solute carrier; admixture mapping; asthma; genome-wide association study; rare variants; β(2)-Adrenergic receptor; β(2)AR.

Figure 1. QQ plot for genome-wide allelic associations with bronchodilator response

The expected distribution of p-values is based on a uniform distribution. Black circles indicate all SNPs (inflation factor, (λ=1.005), whereas blue circles include only common SNPs with a MAF >= 5% (λ=1.004). The shaded region represents the 95% concentration band. QQ plots for the subset of Puerto Ricans and Mexicans are shown in Supplementary Figure 2.

Figure 2. Manhattan plot of genome-wide association studies with bronchodilator response

Association testing for BDR was performed using linear regression including ethnicity, local, and global African and Native American ancestry as covariates. SNPs meeting a genome-wide significance threshold of p < 5×10⁻⁸ are colored in red. Manhattan plots for the subset of Puerto Ricans and Mexicans are shown in Supplementary Figure 4.

Figure 3. Admixture mapping in GALA II Puerto Ricans for (A) African, (B) Native American, and (C) European ancestry

Ancestry association testing for BDR was performed using linear regression including ethnicity, global African, and global Native American ancestry as covariates. Peaks obtaining statistical significance based on permutations are indicated in red. Results for Mexicans are shown in Supplementary Figures 5.

Figure 4. Top signals of allelic association under the admixture mapping peak at 1p3 in Mexicans

Plotted are the –log10(p-values) for tests of allelic association with BDR using linear regression adjusting for ethnicity, local, and genomic African and Native American ancestry. Significant associations were observed at two rare variants in the intron and 3′UTR of SLC22A15 in Mexicans after correcting for a reduced number of tests.

Figure 5. Expression levels of SLC22A15 in the lung

(A) Relative transcript abundance across human tissues assessed using quantitative rtPCR; SLC22A15 mRNA was detected at levels over 50,000x higher in the lung as compared to the liver. (B) Immunohistochemistry staining of lung tissue adapted from the Human Protein Atlas (http://www.proteinatlas.org/ENSG00000163393/normal/bronchus), showing the transcription of SLC22A15 (brown) in the cytoplasm and membrane of bronchial epithelial cells (indicated by a black arrow).