Exploring the causal role of the human gut microbiome in breast cancer risk

The relationship between the human gut microbiome and breast cancer (BC) risk is uncertain. We used Mendelian randomization (MR) and multivariable linear regression to investigate the causal relationship between the human gut microbiome and BC risk (overall and subtype-specific) and interrogate the role of circulating metabolites as potential mediators. Summary-level data from genome-wide association studies of gut microbial traits (N=3,890), circulating metabolites (N=7,824), overall (N=247,175) and subtype-specific (N=200,898) BC risk, including triple negative BC (TNBC; N=120,493), alongside individual-level data from the Flemish Gut Flora Project (N=2,242). Sensitivity analyses were applied to examine the robustness of causal estimates. Initial two-sample MR analyses suggested that the presence of bacteria within the Ruminococcus genus increased overall BC risk, that higher abundances of bacteria within the Parabacteroides genus decreased and presence of unclassified bacteria within the Bacteroidales order increased TNBC risk, and that several circulating metabolites altered BC risk (e.g., homostachydrine with overall BC and Luminal A BC, and betaine with TNBC). Whilst, in multivariable linear regression analyses, several of these metabolites were also associated with the microbial traits that influenced BC risk, importantly, given directional inconsistency of associations across analyses of microbial traits, metabolites and BC coupled with sensitivity analyses implying violations of MR assumptions, associations were unlikely to reflect causality nor a causal pathway between the gut microbiome, metabolites, and BC risk.