Mediating Effects of Plasma Metabolites in Inflammatory Protein-Lymphoma Causality: A Mendelian Randomization Study.

Introduction: Diffuse large B-cell lymphoma (DLBCL) pathogenesis is poorly understood, with limited causal evidence linking circulating inflammatory proteins (CIPs) and metabolites to disease risk. Observational studies face challenges from confounding and reverse causation, while existing Mendelian randomization (MR) analyses lack bidirectional designs and multi-omics integration.

Methods: A bidirectional two-sample MR design was applied using inverse-variance weighting (IVW). Genetic instruments for 91 CIPs derived from Olink proteomic data (14,824 participants). DLBCL genetic associations (1,050 cases; 314,193 controls) were obtained from FinnGen (R10 release). Data for 1,091 blood metabolites and 309 metabolite ratios were sourced from the GWAS Catalog.

Results: Ten CIPs exhibited causal effects on DLBCL. Risk-increasing proteins included: IL-10 (OR=1.46, 95%CI=1.05-2.03), TSLP (1.37,1.01-1.84), IL-17C (1.34,1.05-1.72), NRTN (1.30,1.02-1.66), OPG (1.29,1.01-1.66), and MCP1 (1.26,1.04-1.52). Protective proteins included: CD40 (0.82,0.67-1.00), CXCL9 (0.78,0.61-0.98), CD5 (0.77,0.61-0.97), and MCP3 (0.76,0.58-0.99). Reverse causation was absent for 7 proteins. Mediation analysis revealed 17.2% (p=0.048) of CD5's protective effect was mediated by 1-methylhistidine.

Discussion: These findings establish CIPs as causal factors in DLBCL pathogenesis and identify metabolite-mediated pathways as novel mechanistic links. The bidirectional design and multi-omics integration overcome key limitations of prior research, though statistical power for some mediation tests was limited by metabolite GWAS sample sizes.

Conclusion: Plasma inflammatory proteins causally influence DLBCL risk, partially mediated by metabolites. This underscores metabolite pathways as potential targets for therapeutic intervention.