Fine mapping and positional candidate studies identify HLA-G as an asthma susceptibility gene on chromosome 6p21

Abstract

Asthma affects nearly 14 million people worldwide and has been steadily increasing in frequency for the past 50 years. Although environmental factors clearly influence the onset, progression, and severity of this disease, family and twin studies indicate that genetic variation also influences susceptibility. Linkage of asthma and related phenotypes to chromosome 6p21 has been reported in seven genome screens, making it the most replicated region of the genome. However, because many genes with individually small effects are likely to contribute to risk, identification of asthma susceptibility loci has been challenging. In this study, we present evidence from four independent samples in support of HLA-G as a novel asthma and bronchial hyperresponsiveness susceptibility gene in the human leukocyte antigen region on chromosome 6p21, and we speculate that this gene might contribute to risk for other inflammatory diseases that show linkage to this region.

Figures

Figure 1

Linkage to 6p21 in the Chicago families. A, The dashed line shows the results of the initial genome screen with framework markers. The solid line shows the results of five additional STRPs between framework markers D6S1281 and D6S1019. B, 1-Mb region from D6S258 to D6S265 with positional candidate genes. (All known genes [blue]—but not all pseudogenes [brown], STRPs [green], and intragenic SNPs [purple]—are included.) Circles = SNPs; triangles = in/dels; squares = STRPs; rectangle = HLA-A genotype, which is comprised of multiple SNPs. See appendix A (online only) for allele frequencies and results of association studies.

Figure 2

LD block structure in the extended class I region. A, Graph of LD map, showing LDUs on the Y-axis and distance on the X-axis (Zhang et al. 2002a). Shaded boxes show the five blocks in this region. B, Pairwise TDT of variants within each block. Results for Chicago families are shown in the lower half, and results for Chicago trios are in the upper half. P values were derived by simulations that were conditioned on the evidence of linkage but were not corrected for multiple comparisons.

Figure 3

A, Control using an irrelevant IgG antibody. B–F, Sections labeled with the anti-HLA-G5 antibody. Epithelial cell labeling may be increased in basal epithelial cells immediately above the basement membrane (arrows in panels B, C, and E) but also may label columnar cells (arrowheads in panel B). In areas of damage and focal denudation (D), labeling may be found in remaining epithelial cells. Labeling is also seen in mucosal and submucosal gland cells (F) in the airway, which are of epithelial origin. Panels A and E are from individuals without asthma; panel C is from an individual without asthma but with a 30-year smoking history; panels B, D, and F are from individuals with asthma. mAB 1-2C3 detects soluble HLA-G5 in human bronchial epithelial cells. Human donor lungs that could not be used for transplantation were obtained under an institutional review board–approved protocol from the Regional Organ Bank of Illinois. Diagnoses were extracted from medical records. Bronchi were dissected and frozen in OCT. Five-micron sections were stained for antibodies against HLA-G1, HLA-G2, HLA-G5, and HLA-G6 isoforms, as described elsewhere (Morales et al. 2003). Original magnification was ×400 for all images. Antibodies specific for HLA-G1, HLA-G2, and HLA-G6 were negative in lung sections from patients with asthma as well as those without asthma (not shown).