Supplementary Material for: The Causal Role of Gut Microbiota and Immune Cell Mediation in Gastroenteropancreatic Neuroendocrine Neoplasms: A Mendelian Randomization Study

Background: Gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) are a diverse group of tumors arising from neuroendocrine cells, often characterized by slow growth and subtle symptoms, leading to advanced-stage diagnoses. The gut microbiota (GM) has been implicated in various gastrointestinal malignancies, but its relationship with GEP-NENs remains unclear.
Objective: This study aimed to elucidate the causal relationship between specific GM traits and GEP-NENs using Mendelian Randomization (MR) analysis, and to explore the potential mediating role of immune cells in this relationship.
Methods: A two-sample MR analysis was employed using genome-wide association study (GWAS) data to explore causal associations between GM and GEP-NENs. Genetic variants significantly associated with GM traits were selected as instrumental variables (IVs). Forward MR analysis identified significant GM traits associated with GEP-NENs, followed by reverse analysis to exclude reverse causation. Bayesian weighted Mendelian randomization (BWMR) was employed for verification. Mediation analysis was conducted using a two-step approach to evaluate the mediating role of immune cell traits in these causal relationships. Sensitivity analyses, including heterogeneity and pleiotropy tests, were conducted to ensure the robustness of the findings.
Results: Our analysis identified significant causal associations between 45 GM traits and GEP-NENs, involving 25 bacterial taxa and 20 metabolic pathways. Specifically, s_Paraprevotella_xylaniphila showed the strongest association with pancreatic NEN (P-NEN, OR: 0.473, P: 0.005); metabolic pathway PWY.6588 (pyruvate fermentation to acetone) exhibited the strongest association with gastric NEN (G-NEN, OR: 2.706, P: 0.003); pathway PWY0.845 (superpathway of pyridoxal 5'-phosphate biosynthesis and salvage) had the strongest association with small intestine NEN (SI-NEN, OR: 1.898, P: 0.001); and g_Roseburia displayed the strongest association with colorectal NEN (C-NEN, OR: 0.504, P: 0.007). Furthermore, mediation analyses revealed that specific immune cell traits significantly mediated these associations, with CD3⁻ lymphocyte absolute count showing a notable mediation effect (12.187%, P: 0.019) between g_Clostridium and G-NEN, highlighting the crucial role of immune modulation in GEP-NEN pathogenesis.
Conclusions: In conclusion, our MR analysis provides robust causal evidence that specific GM taxa significantly influence the risk of GEP-NENs, and that this effect is partially mediated through discrete subsets of circulating immune cells. These findings underscore the GM–immune axis as a novel preventive and therapeutic target in GEP-NENs.