In silico design, 2D-QSAR study, pharmacophore modeling and molecular docking of novel Glutaminyl Cyclase inhibitors

Tóm tắt/Abstract

This study aimed to design novel Glutaminyl Cyclase (QC) inhibitors based on 2D-QSAR study, ligand-based pharmacophore modeling and molecular docking. A 2D-QSAR model was developed from a dataset of 1681 QC inhibitors by using Support Vector Regression (SVR) algorithm with 256-bit Morgan fingerprints. Internal (10-fold cross-validation) and external validation were performed with good results: R²_train = 0.992, Q² = 0.804, and R²_pred = 0.826. Next, a set of 29 QC inhibitors was selected to generate pharmacophore models, which were then validated by an external test set with 405 active QC inhibitors and 15487 decoys. The result was potential with EF = 30.43 and GH = 0.75. Then 7 series of novel QC inhibitors were designed with scaffold-based approach by utilizing the library ScaffoldGraph and 2D-QSAR model. Among these, 78 compounds from series 5, bearing 5,6-dimethoxy-1-isoindolinone hybrid 5-methylimidazole core structure, showed good results with predicted pIC₅₀ ranging from 7.00 to 7.36 and satisfied the pharmacophore model. Moreover, the molecular docking study showed that these compounds exhibited good docking scores ranging from -7.31 to -8.68 kcal.mol^-1, favorable binding modes and important interactions with the protein, suggesting that this scaffold is promising for the discovery and development of novel anti-Alzheimer’s Disease (AD) agents inhibiting QC.