Molecular Modeling and Virtual Screening: Application to Computer-Aided Design of Anticancer Inhibitors of Human Dihydroorotate Dehydrogenase

Abstract

Abnormal and uncontrolled cell growth leading to dysfunction of cellular metabolism causes a serious disease called cancer. In 2022, 20 million people had cancer and 9.7 million died. Cancer is a real health problem given the number of deaths caused by this disease. This article presents our work aimed at designing new inhibitors of dihydroorotate dehydrogenase (DHODH), an enzyme involved in the process of tumor development and evolution. Using the computer-aided rational design method, we constructed a QSAR model that translates the linear correlation between the Gibbs free energy and the experimental inhibition constant from a training set containing 32 4-quinoline carboxylic acid derivatives (QCADx). The excellent predictive power of the model in the solvent is attested by the values of the coefficient of determination (R2= 0.88). The active conformations of the ligands of the training set that allowed the QSAR model to be carried out were used to build the DHODH inhibition pharmacophore model ( . A detailed analysis of DHODH catalytic residues was performed to identify key enzyme-inhibitor interactions, guiding the design of a virtual combinatorial library of 59,265 4-quinoline carboxylic acid derivatives (QCADs). Using the PH4 model for virtual library screening, 111 analogues were identified and the most active has a predictive activity 100 times higher than that of the brequinar (most ligand of the training set). This study ends with a verification of the good stability of the DHODH-QCADx complex and the flexibility of the active conformation of the inhibitor for some QCAD analogues with help of the molecular dynamics simulations. The combination of molecular modeling and in silico screening with the PH4 model of the VCL has enabled the identification of new potential candidates for anticancer agents with favorable pharmacokinetic profiles.