Exploring the Protein–Metabolite Interplay to Discover Novel Drug Targets for Alzheimer's Disease

ABSTRACT Background The protein–metabolite interactions play a pivotal role in the pathogenesis of Alzheimer's disease ( AD ); however, the causal significance of these interactions in AD pathogenesis remains elusive. Methods We integrated datasets of protein quantitative trait loci (pQTL), metabolite quantitative trait loci (mQTL), and genome‐wide association study (GWAS) for AD . Initially, we performed MR analysis to identify proteins and metabolites causally associated with AD risk. To guarantee the robustness of these findings, Bayesian colocalization analysis was employed to confirm shared genetic signals. Meanwhile, Steiger filtering was utilized to verify the directionality of causal effects. Sensitivity analyses were further carried out to assess potential pleiotropy and bias. Moreover, we employed MR analysis to assess the causal relationships between proteins and metabolites, identifying protein–metabolite pairs that are associated with AD . Finally, we evaluated the druggability of these causal proteins as potential therapeutic targets. Results Proteomic and metabolomic MR analyses identified 14 proteins and 50 metabolites causally associated with AD . Among these, five proteins (BIN1, C1R, SHPS1, LILRB1, and TMEM106B) passed Bayesian colocalization analysis (PPH4 ≥ 80%). Notably, LILRB1 exhibited a protective effect against AD , while BIN1, C1R, SHPS1, and TMEM106B were implicated as risk factors for the disease. Of the 50 metabolites, 38 passed sensitivity and Steiger filtering analyses and were utilized for further protein‐metabolite pair analysis. Subsequent MR analysis revealed 13 protein–metabolite pairs associated with AD , involving four proteins (BIN1, C1R, SHPS1, and LILRB1) and 12 metabolites. Notably, drugs targeting BIN1, including M344, L‐type calcium channel blockers, and BIN1 monoclonal antibodies, have been reported as potential treatments for AD . For the remaining three drug targets, medications used for other diseases were identified as potentially applicable. Conclusions Our findings highlight four proteins as potential drug targets for AD , and suggest their corresponding protein–metabolite pairs should be prioritized for analysis of potential mechanistic pathways.