Identification of Novel PI3Kδ Selective Inhibitors by SVM-Based Multistage Virtual Screening and Molecular Dynamics Simulations

Abstract

Phosphoinositide 3 kinase delta (PI3Kδ) is a lipid kinase that has been implicated in a variety of immune mediated disorders. The research on isoform selectivity was crucial for reducing side effects. In the current study, an optimized hierarchical multistage virtual screening method was utilized for screening the PI3Kδ selective inhibitors. The method sequentially applied a support vector machine (SVM), a protein ligand interaction fingerprint (PLIF) pharmacophore, and a molecular docking approach. The evaluation of the validation set showed a high hit rate and a high enrichment factor of 75.1% and 301.66, respectively. This multistage virtual screening method was then utilized to screen the NCI database. From the final hit list, Compound 10 has great potential as the PI3Kδ inhibitor with micromolar inhibition in the PI3Kδ kinase activity assay. This compound also shows selectivity against PI3Kδ kinase. The method combining SVM, pharmacophore, and docking was capable of screening out the compounds with potential PI3Kδ selective inhibitors. Moreover, structural modification of Compound 10 will contribute to investigating the novel scaffold and designing novel PI3Kδ inhibitors.