MEK1/2 inhibition transiently alters the tumor immune microenvironment to enhance immunotherapy efficacy against head and neck cancer

Background Although the mitogen-activated protein kinases (MAPK) pathway is hyperactive in head and neck cancer (HNC), inhibition of MEK1/2 in HNC patients has not shown clinically meaningful activity. Therefore, we aimed to characterize the effect of MEK1/2 inhibition on the tumor microenvironment (TME) of MAPK-driven HNC, elucidate tumor-host interaction mechanisms facilitating immune escape on treatment, and apply rationale-based therapy combination immunotherapy and MEK1/2 inhibitor to induce tumor clearance. Methods Mouse syngeneic tumors and xenografts experiments were used to analyze tumor growth in vivo. Single-cell cytometry by time of flight, flow cytometry, and tissue stainings were used to profile the TME in response to trametinib (MEK1/2 inhibitor). Co-culture of myeloid-derived suppressor cells (MDSC) with CD8(+) T cells was used to measure immune suppression. Overexpression of colony-stimulating factor-1 (CSF-1) in tumor cells was used to show the effect of tumor-derived CSF-1 on sensitivity to trametinib and anti-programmed death- 1 (alpha PD-1) in mice. In HNC patients, the ratio between CSF-1 and CD8A was measured to test the association with clinical benefit to alpha PD-1 and alpha PD-L1 treatment. Results Using preclinical HNC models, we demonstrated that treatment with trametinib delays HNC initiation and progression by reducing tumor cell proliferation and enhancing the antitumor immunity of CD8(+) T cells. Activation of CD8(+) T cells by supplementation with alpha PD-1 antibody eliminated tumors and induced an immune memory in the cured mice. Mechanistically, an early response to trametinib treatment sensitized tumors to alpha PD-1-supplementation by attenuating the expression of tumor-derived CSF-1, which reduced the abundance of two CSF-1R(+)CD11c(+) MDSC populations in the TME. In contrast, prolonged treatment with trametinib abolished the antitumor activity of alpha PD-1, because tumor cells undergoing the epithelial to mesenchymal transition in response to trametinib restored CSF-1 expression and recreated an immune-suppressive TME. Conclusion Our findings provide the rationale for testing the trametinib/alpha PD-1 combination in HNC and highlight the importance of sensitizing tumors to alpha PD-1 by using MEK1/2 to interfere with the tumor-host interaction. Moreover, we describe the concept that treatment of cancer with a targeted therapy transiently induces an immune-active microenvironment, and supplementation of immunotherapy during this time further activates the antitumor machinery to cause tumor elimination.