Machine Learning and Molecular Modeling Strategy for the Identification of CNS-Active Acetylcholinesterase Inhibitors

Background: Acetylcholinesterase (AChE) is a key therapeutic target in neurological disorders, and the discovery of novel inhibitors with improved efficacy and pharmacokinetic properties remains a significant challenge. Methods: In this study, an integrated computational and experimental approach was employed to identify potential AChE inhibitors. A machine learning-based model was developed to predict bioactive compounds from large chemical libraries, followed by Blood-Brain Barrier (BBB) permeability screening to ensure Central Nervous System (CNS) suitability. The shortlisted compounds were further evaluated using molecular docking, molecular dynamics simulations, and MM-PBSA binding free-energy calculations to assess their interaction profiles and stability. Selected top-ranked compounds were subjected to in vitro biological evaluation for AChE inhibitory activity. Results: The results demonstrated that several screened compounds, including Z1498348710 and Z1824281875, exhibited notable inhibition with AChE activity reduced to approximately 75% and 72%, respectively, compared to the control. Other compounds such as Z29542160, Z105150208, Z94570687, and Z94570675 showed moderate inhibitory effects, maintaining AChE activity in the range of 82-86%. In comparison, the reference inhibitors Donepezil and Neostigmine bromide displayed significantly stronger inhibition, reducing AChE activity to approximately 20% and 18%, respectively. Conclusions: Overall, the identified compounds demonstrated moderate AChE inhibitory activity while exhibiting favorable predicted physicochemical and computational profiles. These findings suggest that they represent promising starting points for medicinal chemistry optimization and future development as CNS-active AChE inhibitors.