Abstract
BACKGROUND: There may be potential associations between various pathogens, antibody immune responses, and breast cancer (BC), but the specific mechanisms and causal relationships remain unclear. METHODS: First, multiple Mendelian randomization (MR) methods were used for univariable MR analysis to explore potential causal relationships between 34 antibody immune responses (related to 12 pathogens), 46 antibody immune responses (related to 13 pathogens), antibody responses post-COVID-19 vaccination, 731 immune cell types, and various BC subtypes (including overall BC, ER-positive, ER-negative, Luminal A, Luminal B, Luminal B HER2-negative, HER2-positive, and triple-negative BC). The primary results were then subjected to reverse MR analysis, heterogeneity testing using Cochran's Q, and horizontal pleiotropy testing. Robust findings were further used to design mediation pathways involving antibody immune responses, immune cells, and BC. After adjusting the effect estimates using multivariable MR (MVMR), a two-step mediation analysis was conducted to explore mediation pathways and mediation proportions. Finally, linkage disequilibrium score regression (LDSC) was applied to analyze the genetic correlation between phenotypes along mediation pathways, and cross-phenotype association analysis (CPASSOC) was performed to identify pleiotropic SNPs among three phenotypes along these pathways. Bayesian colocalization tests were conducted on pleiotropic SNPs using the multiple-trait-coloc (moloc). RESULTS: We identified potential causal relationships between 15 antibody immune responses to 8 pathogens (Hepatitis B virus, Herpes Simplex Virus 2, Human Herpesvirus 6, Polyomavirus 2, BK polyomavirus, Cytomegalovirus, Helicobacter pylori, Chlamydia trachomatis), 250 immune cell phenotypes, and various BC subtypes. MVMR-adjusted mediation analysis revealed four potential mediation pathways. LDSC results showed no significant genetic correlation between phenotypes pairwise. CPASSOC analysis identified two potential mediation pathways with common pleiotropic SNPs (rs12121677, rs281378, rs2894250). However, none of these SNPs passed the Bayesian colocalization test by moloc. These results excluded horizontal pleiotropy, stabilizing MR analysis results. CONCLUSION: This study utilized MR methods to analyze potential causal relationships between various antibody immune responses, immune cell types, and BC subtypes, identifying four potential regulatory mediation pathways. The findings of this study offer potential targets and research directions for virus-related and immunotherapy-related studies, providing a certain level of theoretical support. However, limitations such as GWAS sample size constraints and unclear specific pathophysiological mechanisms need further improvement and validation in future studies.