Structural and dynamic determinants for highly selective RET kinase inhibition reveal cryptic druggability.

Abstract

The structural and dynamic determinants for highly selective RET kinase inhibition are poorly understood. Here we demonstrate by applying an integrated structural, computational and biochemical approach that the druggability landscape of the RET active site is determined by the conformational setting of the ATP-binding (P-) loop and its coordination with the αC helix. Open and intermediate P-loop structures display additional druggable vulnerabilities within the active site that were not exploited by first generation RET inhibitors. We identify a cryptic pocket adjacent to the catalytic lysine formed by K758, L760, E768 and L772, that we name the post-lysine pocket, with higher druggability potential than the adenine-binding site and with important implications in the regulation of phospho-tyrosine kinase activity. Crystal structure and simulation data show that the binding mode of highly-selective RET kinase inhibitors LOXO-292 and BLU-667 is controlled by a synchronous open P-loop and αC-in configuration that allows accessibility to the post-lysine pocket. Molecular dynamics simulation show that these inhibitors efficiently occupy the post-lysine pocket with high stability through the simulation time-scale (300 ns), with both inhibitors forming hydrophobic contacts in the pocket further stabilized by pi-cation interactions with the catalytic K758. Engineered mutants targeting the post-lysine pocket impact on inhibitor binding and sensitivity, as well as RET tyrosine kinase activity. The identification of the post-lysine pocket as a cryptic druggable vulnerability in the RET kinase and its exploitation by second generation RET inhibitors has important implications for future drug design and the development of personalized therapies for patients with RET-driven cancers.