Design of BBB permeable BACE-1 inhibitor as potential drug candidate for Alzheimer disease: 2D-QSAR, molecular docking, ADMET, molecular dynamics, MMGBSA

Abstract

BACE-1 is a prime therapeutic target for treatment of Alzheimer disease as it cleaves the β-site of APP leading to formation of amyloid plaques. A dataset of 229 benzo-fused heterocyclic compounds reported as BACE-1 inhibitors was utilized to develop various QSAR models (regression and classification) utilizing Monte Carlo algorithm. The dataset was randomly split into different sets for generation of models. The IIC and CCC were calculated to increase the predictive ability of generated models. Among various models, split-1 of Model-1 demonstrated the highest robustness and predictive accuracy for pIC₅₀ values. Internal and external validation was performed which further confirmed the reliability of this model. Structural features responsible for enhancing or reducing pIC₅₀ values were identified and were utilized to design library of 255 compounds. Compounds having pIC₅₀ > 5.0 were further screened on the basis of BBB permeability predicted via ADMET lab 3.0. Total nineteen compounds were found to be BBB permeable which were then docked into PDB: 2WJO. Finally, four compounds with high docking scores were identified and compared with existing BACE-1 inhibitor. MD simulations and MMGBSA analysis were performed and results demonstrated minimal fluctuations throughout the simulation of 100 ns with good binding affinity. This study highlights development of robust QSAR model which assists to design new compounds and predicts them for anti β-secretase activity. Design of novel four molecules were proposed which exhibits good potency, BBB permeability, excellent binding affinity and stable conformations with BACE-1 making them promising candidates for further development.