Abstract 5616: Crystal structure of drug-resistant ROS1 G2032R in complex with zidesamtinib, a clinical-stage ROS1 inhibitor with best-in-class potential

Abstract Introduction. ROS1 fusions are recurrent oncogenic drivers in non-small cell lung cancer (NSCLC). Tyrosine kinase inhibitors (TKIs) crizotinib, entrectinib, and repotrectinib are FDA-approved for treatment of ROS1-positive NSCLC. Approximately 40% of patients progressing on crizotinib harbor the ROS1 G2032R resistance mutation, which is thought to sterically clash with and reduce the activity of many ROS1 TKIs. To address this and other medical needs, zidesamtinib was designed to 1) maintain activity for ROS1 and ROS1 resistant mutants, including G2032R; 2) be brain-penetrant, given the high incidence of brain metastases in ROS1-positive NSCLC; and 3) avoid TRK-family kinases, including TRKB, whose inhibition is associated with neurological toxicity. Here, we disclose the first known crystal structure of ROS1 G2032R, solved in complex with zidesamtinib. Analysis of this structure, combined with computational simulations, provides a molecular explanation for zidesamtinib’s high affinity and selectivity for this drug-resistant ROS1 mutant. Methods. ROS1 G2032R was expressed in Sf9/Sf21 cells, purified via standard nickel chromatography, and crystallized in the presence of zidesamtinib via sitting-drop diffusion methods. Using our and other crystal structures, free-energy perturbation (FEP) calculations were conducted with the Schrodinger-Maestro software to predict affinity changes caused by the ROS1 G2032R mutation for various ROS1 TKIs. Results. We solved a 2.2 Å crystal structure of ROS1 G2032R in complex with zidesamtinib. The aminopyridine moiety of zidesamtinib forms key hinge interactions with the E2027 - M2029 backbone. There is no apparent steric clash between zidesamtinib and R2032—a feature thought to underlie affinity loss for many prior ROS1 TKIs. FEP calculations predicted affinity changes caused by the G-R mutation for various ROS1 TKIs that correlate well with biochemical inhibition data, with zidesamtinib showing the smallest experimental potency shift between ROS1 and ROS1 G2032R. Our modeling also suggests a steric clash between zidesamtinib and TRKA Y591, providing a rationale for selectivity for ROS1 over TRK-family kinases. Conclusion. Analysis of our ROS1 G2032R crystal structure and FEP modeling provide a structural explanation for how ROS1 G2032R impairs the binding of many ROS1 TKIs while maintaining high sensitivity to zidesamtinib. The data presented here, together with previously disclosed preclinical and clinical data, support our design of zidesamtinib as a differentiated ROS1 TKI with best-in-class potential for the treatment of ROS1-positive NSCLC. Citation Format: Joseph M. Magrino, Anupong Tangpeerachaikul, Henry Pelish, Scot Mente. Crystal structure of drug-resistant ROS1 G2032R in complex with zidesamtinib, a clinical-stage ROS1 inhibitor with best-in-class potential [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5616.