Genome-wide Mendelian randomization mapping the influence of plasma proteome on major depressive disorder

Abstract

Plasma proteins play critical roles in a series of biological processes and represent a major source of translational biomarkers and drug targets. In this study, we performed Mendelian randomization (MR) to explore potential causal associations of protein quantitative trait loci (pQTL, n = 54,219) with major depressive disorder (MDD) using summary statistics from the PGC (n = 143,265) and further replicated in FinnGen cohort (n = 406,986). Subsequently, gene expression quantitative trait loci (eQTL) of identified proteins were leveraged to validate the primary findings in both PGC and FinnGen cohorts. We implemented reverse causality detection using bidirectional MR analysis, Steiger test, Bayesian co-localization and phenotype scanning to further strengthen the MR findings. In primary analyses, MR analysis revealed 2 plasma protein significantly associated with MDD risk at Bonferroni correction (P < 3.720 × 10–5), including butyrophilin subfamily 2 member A1 (BTN2A1, OR = 0.860; 95 % CI, 0.825–0.895; P = 1.79 × 10–5) and butyrophilin subfamily 3 member A2 (BTN3A2, OR = 1.071; 95 % CI, 1.056–1.086; P = 3.89 × 10–6). Both the identified proteins had no reverse causality. Bayesian co-localization indicated that BTN2A1 (coloc.abf-PPH4 = 0.620) and BTN3A2 (coloc.abf-PPH4 = 0.872) exhibited a shared variant with MDD, a finding that was subsequently validated by HEIDI test. In the replication stage, BTN2A1 and BTN3A2 were successfully validated in the FinnGen cohort. This study genetically determined BTN2A1 and BTN3A2 were associated with MDD and these findings may have clinical implications for MDD prevention.