Discovery of new JNK3 inhibitory chemotypes via QSAR-Guided selection of docking-based pharmacophores and comparison with other structure-based pharmacophore modeling methods

Abstract

The kinase c-Jun N-terminal Kinase 3 (JNK3) plays an important role in neurodegenerative diseases. JNK3 inhibitors have shown promising results in treating Alzheimer's and Parkinson's diseases. This prompted us to model this interesting target via three established structure-based computational workflows; namely, docking-based Comparative Intermolecular Contacts Analysis (db-CICA), pharmacophore modeling via molecular-dynamics based Ligand-Receptor Contact Analysis (md-LRCA), and QSAR-guided selection of crystallographic pharmacophores. Moreover, we compared the performances of resulting pharmacophores against binding models generated via a newly introduced technique, namely, QSAR-guided selection of docking-based pharmacophores. The resulting pharmacophores were validated by receiver operating characteristic (ROC) curve analysis and used as virtual search queries to screen the National Cancer Institute (NCI) database for promising anti-JNK3 hits of novel chemotypes. Eleven nanomolar and low micromolar hits were identified, three of which were captured by QSAR-guided docking-based pharmacophores.