Although KRAS G12C inhibitors (G12Ci) produce promising initial clinical responses, their efficacy is frequently limited by the rapid emergence of resistance. Developing strategies to overcome or delay resistance is therefore essential, but the diversity of resistance mechanisms complicates the design of effective targeted combinations. We hypothesise that a more broadly applicable approach would be to exploit the ability of KRAS G12C inhibition to partially reverse the KRAS driven immunosuppressive tumour microenvironment (TME) and promote immune-mediated elimination of drug-resistant cancer cells.

We developed a preclinical model to mimic development of resistance to G12Ci by subcutaneously co-engrafting reporter-traced isogenic G12Ci-responsive cancer cells with a minor subpopulation of G12Ci-resistant cells. As resistant models, we used either intrinsically insensitive cells (KRAS G12D) or cells that had acquired resistance following in vivo treatment with G12Ci. Growth of resistant cells was monitored using in vivo luciferase imaging and end point flow cytometry.

KRAS G12C inhibition as monotherapy using the RAS(ON) G12C-selective inhibitor RMC-4998 led to rapid expansion of the drug-resistant subpopulation. Notably, combining G12Ci with immune-enhancing therapies, including PD-1 blockade or SHP2 inhibition, resulted in complete elimination of the resistant cells, despite treatments having no effect on tumours composed exclusively of resistant cells. This bystander elimination and the resulting complete responses depended on adaptive immunity, as resistant cells persisted in immune-deficient mice. Mechanistically, these therapies reshaped the TME, promoting accumulation of pro-inflammatory macrophages and increased infiltration of NK and T lymphocytes, including CD8+ T cells recognising the shared tumour antigen between G12Ci-sensitive and resistant cells. In parallel, resistant cells exhibited increased interferon response signatures, likely induced by changes occurring in the drug-sensitive tumour compartment. Loss of IFNγ receptor reduced the susceptibility of resistant cells to immune-mediated clearance, indicating that tumour-intrinsic IFNγ responsiveness contributes to their elimination.

Our preclinical data demonstrate that rationally designed combination strategies can stimulate anti-tumour immune responses capable of promoting immune-mediated clearance of G12Ci-resistant subpopulations. These preclinical findings provide a foundation for the development of therapeutic combinations aimed at overcoming resistance to G12Ci and improving the durability of treatment responses.

This work was supported by the Francis Crick Institute, the ERC Advanced Grant RASImmune and Revolution Medicines, Inc. under a collaborative research agreement.