AI-Driven Scientific Prompting and Sequential Discovery Pipeline for GSK-3β-Targeted Predictive Modeling and Therapeutic Insights in Alzheimer's disease Using Apigenin from Wheatgrass

Abstract

Alzheimer’s disease (AD) is the leading neurodegenerative cause of dementia, characterized by amyloid-β (Aβ) plaque accumulation, tau hyperphosphorylation, neuroinflammation, and synaptic failure. Glycogen synthase kinase-3β (GSK-3β) centrally mediates these pathologies by driving tau phosphorylation, influencing APP processing, and modulating inflammatory signaling, making it a compelling therapeutic target. Apigenin (4′,5,7-trihydroxyflavone), a flavonoid abundant in wheatgrass and medicinal plants, exhibits neuroprotective, antioxidant, and anti-inflammatory properties in preclinical models. Emerging evidence indicates apigenin attenuates AD-relevant pathology via modulation of the PI3K/Akt/GSK-3β axis, reducing tau phosphorylation, lowering Aβ production, and enhancing neuronal survival and cognitive outcomes in vitro and in vivo. To accelerate translation, we conducted a systematic literature analysis (2010–2025) across major databases to evaluate links between apigenin, GSK-3β modulation, and AD endpoints. We also propose a reproducible, AI-driven sequential discovery pipeline combining literature mining, molecular docking, machine-learning ADMET prediction, and large language model guided experimental design to prioritize apigenin derivatives with improved bloodbrain barrier permeability and pharmacokinetics. Integrating natural product pharmacology with AI-enabled predictive modeling can streamline lead optimization and experimental planning. Collectively, current preclinical data support apigenin as a modulatory agent of GSK-3β-related pathways in AD, and AI-guided workflows offer a promising route to refine apigenin analogs and accelerate development of next-generation GSK-3β-targeted therapeutics for Alzheimer’s disease.