Implications of autolysosome- astrocyte-associated signature in the pathogenesis of Alzheimer's disease: evidence from artificial intelligence and multi-omics and clinical validation.

Background: Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by amyloid-beta plaques and neurofibrillary tangles. Dysfunctional cellular clearance mechanisms, particularly autophagy-lysosomal pathways, and reactive astrocytosis are prominent pathological features, yet their interrelationship remains poorly defined. Objective: This study aimed to decipher a novel co-expression molecular signature linking autolysosomal dysfunction and astrocyte reactivity in AD pathogenesis. Methods: We performed Limma, WGCNA and Xcell algorithms in AD patient hippocampus bulk profiles for enrichment of astrocyte and autolysosome (AA)-associated DEGs. Next, explainable machine learning and consensus clustering enables the identification of AA-associated diagnostic model and molecular subgroups for AD patients at bulk level. Besides, AA-associated central pathogenic factor was identified, and its corresponding biological implications for AD were assessed at AD patient hippocampus single-cell level in temporal and spatial manners. Next deep learning algorithm (Drugreflector) and molecular docking enriched natural compounds for the treatment of AD by targeting AA-associated hub gene. Finally, AD clinical peripheral blood samples were collected for estimation of hub gene expression patterns. Results: 5 AA-associated shared DEGs can elaborate diagnostic and patient stratification capacity for AD patients. HMGCR can be considered as astrocyte-distributed central pathogenic and Berberine-oriented therapeutic target for AD patients. Conclusion: Our findings unveil AA-associated diagnostic model and molecular subgroups coupled with HMGCR center pathogenic and druggable role in AD, which represents an actionable clinical target for AD patients.