Membrane localisation and checkpoint blockade enhance xenoantigen delivery to redirect pre-existing immunity against tumours

Cancer immunotherapies often rely on the recognition of tumour antigens, which strongly limits their efficacy upon heterogeneous antigen expression or downregulation. A strategy to overcome this limitation is to redirect pre-existing antiviral immunity against tumours through the delivery of xenoantigens. While many studies have addressed this by repurposing licensed vaccines, we here investigated the underlying mechanisms of immune redirection via the delivery of non-adjuvanted xenoantigen proteins, thereby avoiding confounding adjuvant- or pathogen-specific effects. Using B16F10 melanoma cells engineered to express the model antigen OVA, we found that tumour rejection in pre-immunised mice depends on the subcellular localisation of the xenoantigen, with membrane-bound antigens eliciting stronger rejection than dose-matched soluble cytoplasmic antigens. Enhanced rejection of membrane-bound OVA expressing tumours was associated with stronger CD4+ T cell responses. In addition, pre-immunisation also increased recruitment of inflammatory monocytes and macrophages at the tumour site. To translate this concept therapeutically, we developed a membrane-targeting OVA fusion protein which, upon intratumoural delivery, redirected pre-existing immunity and made tumours responsive to anti-PD-1 therapy. Importantly, these findings were further validated using the clinically relevant varicella zoster virus (VZV) glycoprotein E (gE) antigen and the licensed varicella vaccine Varivax. Our approach provides a mechanistic and translational perspective for treating poorly immunogenic tumours, leveraging widespread pathogen-specific immune memory in combination with anti-PD1 therapy in cancer patients.