A genome-wide cross-trait analysis identifies genomic correlation, pleiotropic loci, and causal relationship between sex hormone-binding globulin and rheumatoid arthritis

Abstract

Background Our study aims to investigate an intrinsic link underlying sex hormone-binding globulin (SHBG) and rheumatoid arthritis (RA), which remains inconclusive in observational settings. Methods Summary statistics were collected from the largest GWAS(s) on SHBG adjusted for BMI (SHBGadjBMI; Noverall = 368,929; Nmen = 180,094; Nwomen = 188,908), crude SHBG (Noverall = 370,125; Nmen = 180,726; Nwomen = 189,473), and RA (Ncase = 22,350; Ncontrol = 74,823). A genome-wide cross-trait design was performed to quantify global and local genetic correlation, identify pleiotropic loci, and infer a causal relationship. Results Among the overall population, a significant global genetic correlation was observed for SHBGadjBMI and RA ($$r_{{\text{g}}}$$ r g = 0.11, P = 1.0 × 10−4) which was further supported by local signal (1q25.2). A total of 18 independent pleiotropic SNPs were identified, of which three were highly likely causal variants and four were found to have effects on both traits through gene expression mediation. A putative causal association of SHBGadjBMI on RA was demonstrated (OR = 1.20, 95% CI = 1.01–1.43) without evidence of reverse causality (OR = 0.999, 95% CI = 0.997–1.000). Sex-specific analyses revealed distinct shared genetic regions (men: 1q32.1-q32.2 and 5p13.1; women: 1q25.2 and 22q11.21-q11.22) and diverse pleiotropic SNPs (16 in men and 18 in women, nearly half were sex-specific) underlying SHBGadjBMI and RA, demonstrating biological disparities between sexes. Replacing SHBGadjBMI with crude SHBG, a largely similar yet less significant pattern of results was observed. Conclusion Our cross-trait analysis suggests an intrinsic, as well as a sex-specific, link underlying SHBG and RA, providing novel insights into disease etiology.