Identification and validation of mitophagy and astrocyte-related molecular signature in the pathogenesis of Alzheimer's disease: evidence from ensemble learning-driven multi-omics and clinical validation

BackgroundAlzheimer's disease (AD) is a progressive neurodegenerative disorder with limited diagnostic tools and therapeutic options. Dysregulated mitophagy in astrocytes plays a pivotal role in AD pathogenesis. This study aims to identify a mitophagy and astrocyte (MA)-associated molecular signature for AD diagnosis and therapeutic targeting.MethodsLimma, WGCNA, xCell, PPI network and integrated machine learning pipeline coupled with SHAP were deployed on AD patient hippocampal bulk profiles (GSE28146, GSE36980, GSE29378, GSE48350) for identification of MA-associated predictive model and hub gene. Next, astrocyte patten and MA-associated hub gene molecular performance were estimated in hippocampal single-cell profile of AD patients (GSE163577) via advanced analytical frameworks. In addition, active learning framework and molecular docking was deployed in GSE29378 for identification of therapeutic candidate for AD patients by targeting MA-associated hub gene. Furthermore, AD hippocampal tissues were collected, and then MA-associated hub gene expression was estimated.ResultsA core 8-gene MA signature (ITSN1, VLDLR, CYP7A1, SREBF2, RASL12, TPMT, CYP4X1, ARHGEF) was identified, which can guide the molecular subgroup identification and predictive model construction for AD patients. ITSN1 can be considered as the MA-associated hub gene in AD pathogenesis, which was up-regulated and predominantly expressed in astrocytes. Drug repositioning identified BRD-K10008415 as the potential compound predicted to reverse the AD signature by targeting ITSN1.ConclusionsThis study identified ITSN1 as a MA-associated critical hub potential connecting mitophagy dysregulation and astrocyte dysfunction in AD. We also identified MA-associated molecular signatures that can potentially elaborate predictive effects on AD pathogenesis. BRD-K10008415 can be considered as potential candidate for AD treatment by targeting ITSN1.