Abstract

Psychiatric, neurodevelopmental, and neurodegenerative disorders, including Alzheimer’s disease (AD), attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), bipolar disorder (BIP), major depressive disorder (MDD), and schizophrenia (SCZ), exhibit complex etiologies driven by immune and metabolic dysregulation. While distinct in their clinical onset, these conditions share overlapping molecular vulnerabilities. This study pioneers an integrative multi-omics framework, combining multi-tissue TWAS, cross-disorder pleiotropy analyses, Mendelian Randomization (MR), predictive machine learning, and BV2 microglial profiling. Crucially, our analysis uncovered a robust “Dual-Axis” etiological architecture. First, a systemic metabolic axis emerged as a primary driver, particularly involving FADS2-mediated lipid dysregulation and gut-brain axis interactions. This axis shares mechanisms between bipolar disorder and schizophrenia, with Multi-tissue TWAS revealing peripheral contributions (e.g., liver, colon) to CNS pathology. Second, MHC Class II-mediated pathways, driven by HLA-DRA , HLA-DRB1 , HLA-DQB1 , and HLA-DQA1 , emerged as a transdiagnostic neuroinflammatory nexus across AD, BIP, MDD, and SCZ, orchestrating antigen presentation to CD4 + T-helper cells. To bridge these genomic findings with cellular function, BV2 microglial profiling was performed to provide a cellular-context reference specifically for the identified immune risk component. This cellular model confirmed that the immunogenetic risk burden maps to a specific proinflammatory activation state characterized by upregulated neurotoxins ( Lcn2 , Nos2 , Ccl2 ) and suppressed lipid transport/phagocytosis ( Apoe , Cd68 ). Machine learning models leveraging these signatures achieved robust predictive performance, particularly for BIP and MDD. MR analyses uncovered causal roles of immune, lipid, and microbial pathways, with shared metabolic signatures (e.g., N-acetylarginine) across