Evaluating the Impacts of Inflammatory Proteins on the Risk of Alzheimer's Disease

Accumulating evidence suggests that inflammation contributes to the pathogenesis of Alzheimer's disease (AD). However, the relationship between circulating inflammatory proteins and AD remains incompletely understood. Two-sample Mendelian randomization (MR) was used to assess the associations of 91 circulating inflammatory proteins with AD. The discovery stage protein GWAS included 14,824 participants, and the validation stage analysis used UK Biobank data from 54,219 participants. The AD dataset comprised 85,934 cases and 401,577 controls. The inverse variance weighted (IVW) method was the primary approach, complemented by weighted median, MR-Egger, heterogeneity, and pleiotropy analyses. MR estimates from both stages were subsequently combined by meta-analysis, followed by drug-gene interaction analyses. In the discovery stage, four proteins showed nominal associations with AD: AXIN1 (OR = 1.082, P = 0.026), NGF (OR = 1.061, P = 0.048), CXCL11 (OR = 0.952, P = 0.017), and TNFRSF9 (OR = 0.946, P = 0.048). However, none survived FDR correction (FDR = 0.799). In the validation stage, only TNFRSF9 reached nominal significance (OR = 0.870, P = 0.031), which did not survive FDR correction (FDR = 0.125); AXIN1 showed an opposite direction. Meta-analysis provided suggestive pooled estimates for CXCL11 (OR = 0.956, P = 9.34 × 10⁻3), TNFRSF9 (OR = 0.926, P = 0.034), and NGF (OR = 1.061, P = 0.034), but these should not be regarded as independent confirmation. DGIdb analysis identified interactions between the genes of interest and various drugs or compounds. These findings provide exploratory genetic evidence linking circulating inflammatory proteins to AD, but robust replication was not established after multiple testing correction and requires further validation.