Expanding immunotherapy options in cholangiocarcinoma: Role of CD27 agonist in combination with PD-L1 and MEK inhibition on antitumor effect and CD8+ T cells.

Abstract

518 Background: Despite addition of PD-L1 therapy to unresectable cholangiocarcinoma (CCA) standard of care, there remains an unmet need for novel therapies. Our recent phase II trial of dual PD-L1 inhibition (PD-L1i) and MEK inhibition (MEKi) significantly improved progression-free survival (PFS). While MEKi enhances tumor antigen presentation and T cell tumor infiltration, it also impairs T cell priming and activation with notable CD27 upregulation. Our work reveals that CD27 agonists (CD27Ag) foster T cell stem-like and resident memory phenotypes, negating inhibitory effect of MEKi on CD8+ T cells. We hypothesized that addition of a CD27Ag to PD-L1i/MEKi combination therapy would rescue T cell activation and improve antitumor efficacy. Methods: Immune cells from pmel-1 mice were isolated, activated with gp100, and cultured in vitro with 100 IU/mL IL-2 and single, dual, or triple therapy with CD27Ag, cobimetinib (MEKi) and/or αPD-L1. Day 3 flow cytometry characterized CD8+ T cell phenotype and cytokine expression. Additionally, syngeneic murine CCA cell line URCCA4.3 (KrasG21D, tp53-/-) was injected into C57BL/6 mice, subcutaneously or intrahepatic, followed by ~3 weeks of treatment with mono, dual, or triple therapy. For CD8 depletion studies, CD8-depleting antibody was given on the day prior to and following surgery and every 3-4 days afterwards. Tumor and tumor-draining lymph nodes were harvested at study endpoint. Lymphocytes were characterized by flow cytometry. Results: MEKi, alone or with αPD-L1, reduced frequency of activated T cells in vitro (dose-dependent effect on CD69+). CD27Ag combined with MEKi maintained prominent T cell activation. Triple therapy PD-L1i/MEKi/CD27Ag significantly reduced tumor growth rate in mice versus mono or dual therapies (p<0.05). Tumor-infiltrating lymphocytes comprised a stem-like (Tcf-1+), effector memory (CD44+CD62L-), and tissue resident memory (CD103+CD69+) phenotype in triple therapy versus single or dual therapy (p<0.01, p<0.0001, p<0.0001 respectively). All groups had similar numbers of CD8+ stem-like cells (Tcf-1+Tim-3-) in draining lymph nodes, but these cells were only present in tumor from the triple therapy arm (p<0.05). Cell depletion studies showed that CD8+ T cells were required for efficacy of triple therapy in vivo. Finally, an aggressive orthotopic liver tumor model showed improved median survival in triple therapy mice versus control (37.5 days versus 29.5 days, p<0.01). Conclusions: CD27 agonism prevents the impairment of T cell activation that was mediated by MEKi. Moreover, this CD27Ag/MEKi/PDL1i triple therapy bolstered trafficking of T cells with stem-like, effector, or resident memory properties, in turn enhancing the antitumor response. This therapy is being evaluated in metastatic CCA patients in an ongoing Phase II clinical trial.