A multi-omics Mendelian randomization study identifies putatively causal genes and pathways for Alzheimer's disease associated with reactive astrocytes.

Abstract

BackgroundNeuroinflammation, particularly involving reactive astrocytes, plays a pivotal role in Alzheimer's disease (AD) progression.ObjectiveHowever, the causal genes and regulatory pathways linking astrocyte reactivity to AD risk remain unclear.MethodsWe performed a multi-omics summary-data-based Mendelian randomization (SMR) analysis, integrating large-scale AD GWAS data with methylation (mQTLs), expression (eQTLs), and protein (pQTLs) data for 514 reactive astrocyte-related genes. Causal inference was strengthened using colocalization and tissue-specific validation, and pathway enrichment.ResultsOur multi-omics SMR analysis identified a core set of high-confidence genes related to reactive astrocytes with putatively causal roles in AD. Based on mQTL-eQTL analysis, SPARC (cg08331313), RPS6KA2 (12 CpG sites), and LEP (4 CpG sites) demonstrated significant regulatory cascades, where methylation changes modulated gene expression and subsequently influenced AD risk. The eQTL-pQTL analysis revealed GCDH as genes with strong expression-protein correlations. Cross-tissue validation between blood and brain pinpointed a robust set of seven genes, including MAPK3, HLA-DQB1, and CSF3, that consistently associate with AD risk, highlighting systemic effects. Functional and network analyses of these causal candidates revealed that they converge upon the MAPK signaling pathway as a central mechanistic hub in both peripheral blood and brain tissue. Furthermore, network analysis identified seven hub genes, including MAPK3, LEP, and STAT1, as critical regulators within this astrocyte-centered AD network.ConclusionsOur study systematically identifies genetically predicted reactive astrocyte-related genes in AD pathogenesis through multi-omics Mendelian randomization, highlighting MAPK signaling as a putatively causal mechanistic hub.