Zhaocheng Li, Zifei Liu, Yuhang Zhang, Yulin Lei, Yanqing Wang
Visual impairment has been proposed as a potentially modifiable risk factor for dementia, including Alzheimer's disease (AD), but the biological mechanisms linking visual disturbance to AD remain unclear. It is unknown whether AD and visual disturbance share specific genetic determinants, how such effects are distributed across brain regions, and whether they converge on discrete molecular pathways that could be targeted for prevention or treatment. Multi-trait analysis of GWAS was applied to summary statistics from large-scale GWAS of AD and a visual disturbance (VD) phenotype, yielding VD-informed AD association statistics (VDAD). Gene-level associations were then mapped across brain tissues by integrating single-tissue and cross-tissue transcriptome-wide association studies, brain proteome-wide association analyses and gene-based tests. A small set of genes (GRN, PVR and RAB29) emerged as consistently associated with both AD and VDAD. A broader panel of genes showed partially overlapping but also phenotype-specific architectures, with PM20D1 and LRRC37A2 more prominent in AD and SLC41A1 and MTCH2 more strongly implicated in VDAD. These signals were anchored to frontal cortical, hippocampal, cerebellar and striatal tissues, suggesting involvement of circuits that integrate visual input with memory, motor control and balance. Network and pathway analyses highlighted a GRN-centered lysosomal axis and metabolic pathways linked to PM20D1 and SLC41A1 as candidate mechanisms bridging visual disturbance and AD risk. These findings support the view that visual disturbance-related AD risk is shaped by a limited set of genes and pathways that are supported across complementary molecular analyses and mapped to distributed brain/CNS reference tissues. These findings generate testable hypotheses for future studies with detailed visual phenotyping.