Integrative analysis of immune-cell eQTLs and Mendelian randomization identifies genetically supported genes associated with breast cancer susceptibility.

Abstract

Breast cancer susceptibility loci identified by genome-wide association studies (GWAS) have revealed extensive genetic contribution to disease risk, yet the genes and immune regulatory mechanisms underlying these associations remain largely unresolved. Here, we integrated immune-cell-specific cis-eQTL data with breast cancer GWAS using a two-sample Mendelian randomization framework. Genetic instruments from five immune-cell resources (DICE, eQTLGen, OneK1K, Soskic, and TenK10K) were evaluated against FinnGen R12 breast cancer summary statistics (24,270 cases and 222,078 controls of European ancestry). Candidate associations were subsequently assessed using Bayesian colocalization, SuSiE fine-mapping, and SMR across independent eQTL datasets and GTEx tissues. TenK10K-based MR identified 516 FDR-significant gene-cell-type associations (131 unique genes), of which 433 associations (106 genes) were supported by convergent genetic evidence (colocalization or LD-based). The APOBEC3A/APOBEC3B locus showed the strongest evidence, where higher immune-cell expression was associated with reduced breast cancer risk. SMR analyses validated 42 of these genes, with YBEY, KCNN4, and ATG10 consistently supported across four independent datasets. In breast tissue, 10 genes showed FDR-significant SMR associations with concordant effect directions. Functional analyses implicated antigen presentation, monocyte differentiation, and APOBEC-mediated DNA editing, with protein-interaction networks converging on the hormone-related genes ESR1 and FGFR2. These findings identify immune-cell-specific regulatory genes that may contribute to breast cancer susceptibility and provide genetically informed candidates for future functional studies.