Combining AI and traditional screening for discovery of a potent ROCK2 inhibitor against lymphoma

Abstract

The AGC family serine/threonine kinase, ROCKs, is essential in various biological processes, such as cell contraction, migration, and morphological changes. The ROCKs family, consisting of two main members, ROCK1 and ROCK2, shares similarities but possesses distinct functions. In recent years, the therapeutic potential of ROCK2 in the anti-tumor field, especially in the field of leukemia, has received widespread attention from researchers. In this study, a combination of traditional methods and advanced AI-driven techniques, including the AI-based Deepdock algorithm following conventional screening, was employed for molecule discovery to enhance the efficacy of virtual screening. This combined approach led to the identification of several potential ROCK2 inhibitor molecules. One of these, compound 5, demonstrated a notable inhibitory effect, with an IC50 of 35.66 nM in the kinase ADP-GLO assay. Subsequent tests on lymphoma cell lines validated its cytotoxic efficacy, especially in the Z138 human lymphoma cell line, where a significant killing effect at 19 nM was observed. This research underscores the potential of combining AI with traditional screening methods in drug discovery and highlights a promising ROCK2 inhibitor for lymphoma treatment.