Mapping the 17q12-21.1 Locus for Variants Associated with Early-Onset Asthma in African Americans

Abstract

Rationale: The 17q12-21.1 locus is one of the most highly replicated genetic associations with asthma. Individuals of African descent have lower linkage disequilibrium in this region, which could facilitate identifying causal variants.Objectives: To identify functional variants at 17q12-21.1 associated with early-onset asthma among African American individuals.Methods: We evaluated African American participants from SAPPHIRE (Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race-Ethnicity) (n = 1,940), SAGE II (Study of African Americans, Asthma, Genes and Environment) (n = 885), and GCPD-A (Study of the Genetic Causes of Complex Pediatric Disorders-Asthma) (n = 2,805). Associations with asthma onset at ages under 5 years were meta-analyzed across cohorts. The lead signal was reevaluated considering haplotypes informed by genetic ancestry (i.e., African vs. European). Both an expression-quantitative trait locus analysis and a phenome-wide association study were performed on the lead variant.Measurements and Main Results: The meta-analyzed results from SAPPHIRE, SAGE II, and the GCPD-A identified rs11078928 as the top association for early-onset asthma. A haplotype analysis suggested that the asthma association partitioned most closely with the rs11078928 genotype. Genetic ancestry did not appear to influence the effect of this variant. In the expression-quantitative trait locus analysis, rs11078928 was related to alternative splicing of GSDMB (gasdermin-B) transcripts. The phenome-wide association study of rs11078928 suggested that this variant was predominantly associated with asthma and asthma-associated symptoms.Conclusions: A splice-acceptor polymorphism appears to be a causal variant for asthma at the 17q12-21.1 locus. This variant appears to have the same magnitude of effect in individuals of African and European descent.

Keywords: African Americans; GSDMB; ORMDL3; asthma; chromosome 17.

Figure 1.

Locus zoom plots of variants in the chromosome 17q12–21.1 region and the association with asthma onset at age <5 years versus control subjects among African American participants from the (A) three-cohort meta-analysis and (B) after conditioning on the rs11078928 genotype. chr17 = chromosome 17; GSDMB = gasdermin-B; ZPBP2 = zona pellucida–binding protein 2.

Figure 2.

Linkage disequilibrium (LD) between variants at the 17q12–21.1 locus among African American SAPPHIRE (Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race–Ethnicity) participants without asthma but (A) homozygous for European ancestry and (B) homozygous for African ancestry. The haplotype block containing the most significant association for early-onset asthma is noted (black circle) and (C) shown in greater detail. The degree of pairwise LD (r² × 100) between two markers is shown in each square; darker shaded squares and higher numbers (range, 0–100) denote higher LD.

Figure 3.

Shown are the various GSDMB (gasdermin-B) transcript isoforms and their relationship to their mapped genomic position on chromosome 17. GSDMB is transcribed from the reverse strand; hence, the exons (labeled at the bottom of the figure) are numbered in order from right to left. The thicker boxes denote regions that are transcribed: yellow signifies the untranslated regions, and blue signifies the translated regions. The bottom arrow shows the location of the start codon and the direction of translation. Dashed lines are intronic regions not included in the final transcript isoform. The footnote for each transcript isoform denotes the resulting product encoded: *protein with 403 amino acids (GSDMB-1), ^†protein with 416 amino acids (GSDMB-3), ^‡protein with 394 amino acids (GSDMB-2), ^§no protein because of retained intron, ^||nonsense-mediated decay, and ^¶protein with 407 amino acids (GSDMB-4).

Figure 4.

The relationship between rs11078928 genotype and GSDMB (gasdermin-B) isoforms. The ENST identification numbers correspond to the alternatively spliced gene isoforms shown in Figure 3. The P value (Kruskal-Wallis one-way ANOVA) for the univariable relationship between genotype and transcript expression is shown at the top of each plot. Solid circles denote transcript read abundance within individuals; the boxes denote the interquartile range of transcript expression; and the middle line within each box is the median level of transcript expression. ENST = Ensembl-spiced transcript.