Cross-Tissue Proteomic Mendelian Randomization Identifies Therapeutic Targets for Vascular Dementia

Vascular dementia (VD) is the second most prevalent form of dementia, yet no disease-modifying treatments are available. Proteins in the brain, cerebrospinal fluid (CSF), and plasma with strong genetic associations represent promising therapeutic targets. However, a comprehensive, proteome-wide screening for VD-associated proteins has not been conducted. We employed a cross-tissue proteomic Mendelian randomization (MR) framework, integrating pQTL data from brain, CSF, and plasma with VD GWAS. We identified causal VD-associated proteins, assessed their consistency across VD subtypes, explored metabolite-mediated effects via metabolomic MR, determined cell-type specificity using single-nucleus RNA sequencing (snRNA-seq), and prioritized therapeutics through drug-target interaction analysis and molecular modeling. MR analysis identified APOE as the strongest VD-associated protein in both CSF and plasma, indicating systemic relevance. Several proteins, including the 14-3-3 family, AREG, SMOC1, and UBE2G2, were exclusively associated in CSF, suggesting CNS-specific roles. Metabolomic MR revealed key APOE-mediated metabolites linked to disease progression. snRNA-seq showed APOE upregulation in excitatory neurons of VD patients. Drug screening highlighted benserazide and puromycin as top candidates, validated by molecular simulations. This study systematically identifies genetically validated proteomic targets for VD, establishing APOE as a central molecular driver. By integrating proteomics, genomics, metabolomics, and drug discovery, we propose a comprehensive framework for targeted therapeutic development, supporting precision medicine in VD treatment.