Mechanisms of resistance to active state selective tri-complex RAS inhibitors
Tri-complex inhibitors (TCIs) act as molecular glues to recruit cyclophilin A (CYPA) to the active (GTP-bound or ON) conformation of RAS, which in turn prevents the activation of downstream effector proteins like RAF and PI3K. Emerging data demonstrate clinical activity, including tumor regressions, in patients with RAS driven cancers. Despite being promising therapeutic interventions, the mechanisms of resistance in patients treated with these inhibitors remain unknown. Here we studied matched baseline and post-progression specimens from patients treated with the RAS(ON) multi-selective inhibitor daraxonrasib (RMC-6236). Tissue or cell-free DNA specimens were collected from 40 patients with RAS-mutant non-small cell lung, colorectal, or other cancers. Eighteen patients (45%) were found to have acquired alterations in RAS signaling intermediates, including recurrent alterations in KRAS, BRAF, RAF1, MAP2K1/2 and PIK3CA. Preclinical resistance models mirrored the alterations observed in patients. We found that secondary KRAS Y64X mutations caused resistance by disrupting an important pi-pi interaction between KRAS and the indole ring of daraxonrasib, which lowers the affinity of the daraxonrasib:CYPA binary complex for active KRAS. We also identified kinase-dead and low-activity BRAF mutations in samples with acquired resistance. This is puzzling, because TCIs are expected to prevent the interaction between RAS and BRAF, which is needed for hypoactive mutants to dimerize and signal. We now show that RAF dimers are harder to displace from active RAS, as compared to their monomeric forms. Indeed, enhanced RAF dimerization attenuated the ability of TCIs to recruit CYPA to active RAS, resulting in diminished inhibition of oncogenic signaling and tumor growth. Thus, several clinical resistance alterations converge at attenuating the formation of the RAS:daraxonrasib:CYPA tri-complex, either by preventing daraxonrasib binding or by inducing RAF dimers.