Identification of phosphodiesterase 10 A modulators for neurodegenerative and psychiatric disorders: Combination of physics-based virtual screening and machine learning approaches

Abstract

Phosphodiesterase (PDE) is a crucial enzyme that regulates intracellular signal transduction by breaking down cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) into inactive forms. Among the 11 PDE families, PDE10A has gained attention as a potential therapeutic target for neurodegenerative and psychiatric disorders. This study aimed to identify potent inhibitors targeting the active site of PDE10A. A ligand-guided virtual screening method was used to find potential modulators from the ZINCPharmer database. The ligand library was subjected to grid-based molecular docking using AutoDock Vina (ADV) and PLANTS tools. Absolute binding affinity was predicted and refined with KDEEP. The docking protocol was validated by evaluating ADMET properties of sorted compounds using ADMET-AI. Protein-ligand interactions were analyzed with ProteinPlus. The final four compounds ZINC09233950, ZINC19374064, ZINC33686121, and ZINC58090432 showed binding affinities of -9.1, -9.3, -9.7, and -9.3 kcal/mol, respectively. Molecular dynamics (MD) simulations were conducted over 100 ns to assess the stability of the protein-ligand complexes within a cubic water box. The binding free energies of selected compounds were evaluated using the MM-GBSA method, confirming their potential as PDE10A inhibitors. The study identified potential inhibitors and highlighted the value of a ligand-guided drug discovery approach in enhancing specificity and efficacy.

Keywords: Ligand-guided virtual screening; MM-GBSA method; Machine learning; Molecular docking; Phosphodiesterase 10A.