Intermediate-based virtual screening of c-Kit kinase inhibitors as potential anti-tumor agents via ab inito folding, molecular dynamics simulation, and molecular docking

Uncontrolled activation of c-Kit is closely related to the pathogenesis and progression of leukemia, gastrointestinal cancer, and other malignant diseases. Although there are several inhibitors available, due to the limitation of selectivity and the unfavorable side effects, designing and discovering highly selective inhibitors targeting c-Kit kinase, especially the gain of function mutation (for example c-Kit D816V), is still necessary. To identify novel c-Kit inhibitors, a metastable state-based virtual screening approach, which was successfully implemented in other kinase inhibitors, was employed in the current study. The results from our current study demonstrated the residues adjacent to the DFG motif within the activation loop could fold into short alpha-helices aside from the random coil, which was commonly found in the crystal structure. By expanding the conformation pool of the activation loop via PyRosetta-based ab initio folding protocol, we constructed a series of structural models of the c-Kit kinase intermediate between the inactive and active states. After evaluation of the thermal stability of the metastable state with molecular dynamics simulation, one structural model showed higher stability of alpha-helix, and the activation loop was retained. Considering the wild-type and D816V mutated KIT kinase shared similar metastable states during the kinase activation process, we developed a hypothesis that the identified intermediate might hold the potential to identify inhibitors targeting D816V mutations from the compound database. As expected, the intermediate structure showed higher selectivity to KIT D816V selective inhibitors, such as bezuclastinib, avapritinib, BLU-263, and elenestinib, than imatinib or masitinib. The virtual screening of the available KIT kinase inhibitor database further identified vorolanib, semaxanib, henatinib, and pexmetinib may possess potential inhibitory effects against wild type, as well as the mutated c-Kit kinase. The results from our current study not only proposed a novel structural model that could be used for the identification of selective c-Kit D816V inhibitors but also identified several potential inhibitors from available kinase inhibitors, which might shed new light on the design of new therapeutic approaches for c-Kit mutation-driven malignant diseases.