Multi-Omics Integration and Causal Inference Identify HSD17B1 as a Potential Nobiletin Target Linking Neurosteroid Metabolism to Alzheimer’s Disease

Alzheimer’s disease (AD) is characterized not only by neuronal dysfunction but also by profound remodeling of the brain microenvironment, including immune–glial activation and metabolic dysregulation. Increasing evidence also implicates neurosteroid-related pathways in AD and dementia. Nobiletin has shown neuroprotective effects in AD-related models, but its upstream human targets and mechanism-based translational relevance remain insufficiently defined. Here, we integrated multi-omics analyses, interpretable machine learning, causal inference, structural modeling, and experimental validation to identify candidate nobiletin-associated molecular nodes in AD. HSD17B1 consistently emerged as a central AD-associated candidate across multiple analytical layers and showed reproducible discriminatory performance in independent validation cohorts. SHAP analysis further identified HSD17B1 as a major contributor to the optimal predictive model, while Mendelian randomization supported a protective association between genetically increased HSD17B1 expression and AD risk. Immune infiltration, single-cell, and spatial transcriptomic analyses linked HSD17B1 to glia-associated remodeling and regionally heterogeneous expression patterns in AD. Molecular docking and molecular dynamics simulations supported the structural feasibility of nobiletin binding to HSD17B1, and in an Aβ1–42-induced SH-SY5Y cell model, nobiletin increased HSD17B1 expression at both the mRNA and protein levels. Together, these findings support HSD17B1 as an AD-associated and nobiletin-responsive candidate molecular node, highlight a potential connection between nobiletin and neurosteroid-related regulation, and provide an integrated framework for target prioritization and validation in AD.