Oncolytic measles viruses to support BiTE and CAR T cell re-direction immunotherapies of solid tumors

Abstract

Bispecific T cell engagers (BiTEs), artificial antibodies designed to cross-link T cells to tumor cells, and chimeric antigen receptors (CARs), expressed on the surface of modified T cells and transferring activating signals via intracellular domains, can re-direct T cells to tumor surface antigens for efficient treatment of hematological malignancies. In solid tumors, however, these novel immunotherapies face fundamental challenges: Physical exclusion of T cells and an immunosuppressive microenvironment prevent efficacy while systemic administration is associated with severe toxicities. Oncolytic viruses have emerged as ideal agents for combination immunotherapies, as lytic replication in tumors does not only induce tumor debulking, but also immunogenic cancer cell death and local inflammation. We therefore hypothesized that oncolytic virotherapy, by eliminating physical barriers and reversing local immunosuppression, can promote anti-tumor T cell responses and thus provide a potent treatment option for solid tumors in combination with T cell re-direction. We furthermore hypothesized that tumor-targeted expression of a virus-encoded BiTE could achieve local BiTE activity without systemic side effects. To test these hypotheses, we engineered oncolytic measles viruses (MV) encoding BiTEs targeting CD20 or carcinoembryonic antigen, respectively, as model antigens (MV-BiTE). Kinetics of viral replication and virus-mediated cytotoxicity showed minor differences compared to unmodified virus, and functional BiTEs could be obtained from the supernatant of virus-infected cells. A newly established syngeneic solid tumor model of B16-CD20-CD46 murine melanoma cells stably expressing human antigens CD20 and CD46 as BiTE target antigen and measles virus entry receptor, respectively, engrafted in immunocompetent C57BL/6 mice when implanted subcutaneously. In this model, treatment with MV-BiTE targeting CD20 significantly prolonged survival compared to control treatments with unmodified MV, MV encoding CEA-specific control BiTE, purified BiTE, or carrier fluid only, respectively. UV irradiation of MV-BiTE completely inhibited viral replication but did not abrogate efficacy in this model. Treatment efficacy was furthermore not impaired in mice previously immunized with MV. Increased T cell infiltration into tumors and an effector T cell phenotype characterized by high intratumoral CD8+ T cell levels and low relative abundance of regulatory T cells was observed upon MV-BiTE treatment. Targeted transcriptome analysis revealed upregulation of genes associated with T cell activation, proliferation, and differentiation, but also with inhibition and exhaustion, providing a rationale for combination with checkpoint inhibitors. Combinations of MV, BiTE, and CAR T cell therapies were investigated in a pancreatic cancer model. Cytotoxic potential of each treatment alone towards human pancreatic adenocarcinoma cells was observed in vitro. However, no significant benefit of combinations was observed in a pilot experiment in immunodeficient mice, and T cell persistence was limited. For future in vivo monitoring of T cells by magnetic resonance imaging, labeling with iron oxide nanoparticles was established. In addition, we describe a mathematical model for in silico predictions of treatment outcome to optimize scheduling of combination therapies. Taken together, this study shows for the first time efficacy of a BiTE-encoding oncolytic virus in an immunocompetent mouse model. This warrants further investigations towards future translation. We furthermore introduced a mathematical model of combination immunovirotherapy to further support the development of novel treatment options for cancer patients.