A multi-omics framework integrating gut microbiota, blood metabolites, and immune cells to elucidate the pathogenesis of Alzheimer’s disease

Abstract

Background Alzheimer’s disease (AD) develops through complex interactions between the central nervous system and peripheral systems. The microbiota-metabolite-immune axis has emerged as an important focus of AD research. However, the coordinated mechanisms that regulate this axis remain poorly understood. Methods We used a multi-stage, multi-omics strategy to systematically investigate peripheral–central interactions in AD. The analytical framework integrated Mendelian randomization (MR), summary-data-based Mendelian randomization (SMR), differential expression analysis, machine learning, single-cell and spatial transcriptomics, and quantitative real-time polymerase chain reaction (qPCR) trend confirmation. Results Exploratory MR analyses identified multiple microbial taxa, metabolites, and immune cell phenotypes showing associations consistent with potential causal effects on AD. Integrating the SMR and MR findings with differential expression analysis led to the identification of 31 core genetically associated genes. A five-gene predictive model comprising ATF7IP2, TWSG1, PTPRN2, ASCC3 and IGF1R was then developed using machine learning. The diagnostic potential of the individual feature genes was further evaluated in an external validation dataset. Spatial transcriptomic analyses revealed clear cell type-specific expression patterns in brain tissue, with IGF1R, ASCC3and TWSG1 showing potential co-localization in oligodendrocytes. qPCR trend confirmation in pooled samples produced expression trends consistent with the directions inferred from eQTL-based MR. Conclusions This study mapped a regulatory network underlying the AD microbiota–metabolite-immune-brain axis and identified core genes with potential diagnostic and therapeutic value. The spatial transcriptomic findings, while primarily based on in situ co-localization analysis, highlight a biologically plausible but provisional working hypothesis regarding an active role for oligodendrocytes in AD pathology. Overall, this study supports a systems-level view of AD that may inform precision medicine strategies.