PD-L1 directed bispecific Vδ2-T cell engager combines lysis of PD-L1 expressing tumor cells with PD-1 immune checkpoint inhibition and modulation of the tumor immune microenvironment.

BACKGROUND: Despite the notable success of programmed death ligand 1 (PD-L1)/programmed death 1 (PD-1) immune checkpoint blockade in cancer, resistance remains a substantial challenge. Combining immune checkpoint blockade with direct targeting of effector T cells to tumor cells might improve outcome for a broader spectrum of patients with cancer. Phosphoantigen-responsive Vγ9Vδ2-T cells are potent immune cells that play a pivotal role in tumor immunosurveillance and can coordinate downstream immune activity. Here, we explored whether a Vδ2 bispecific T cell engager (bsTCE) directed against PD-L1 could combine lysis of PD-L1+ tumor cells with PD-L1 immune checkpoint blockade. METHODS: PD-L1 specific single domain antibodies (VHHs) were tested for binding to PD-L1 and their ability to interfere with PD-1 binding and function. One PD-L1 VHH was selected for fusion to a Vδ2-T cell receptor specific VHH (PD-L1xVδ2 bsTCE) and tested for its ability to activate Vγ9Vδ2-T cells and lyse melanoma cell lines, as well as patient-derived renal cell carcinoma (RCC) and metastatic melanoma cells. These patient-derived tumor suspensions were also used to explore effects on CD4+ and CD8+ T cells and myeloid cells. Infiltration of Vγ9Vδ2-T cells and tumor kill was tested in a three-dimensional tumor spheroid melanoma model. RESULTS: A PD-L1xVδ2 bsTCE was generated and shown to block PD-1 binding resulting in the release of PD-1+ cells from PD-L1 mediated inhibition. The PD-L1xVδ2 bsTCE also mediated robust Vγ9Vδ2-T cell activation, efficient lysis of RCC and melanoma cell lines and/or patient-derived tumor cells, and infiltration of Vγ9Vδ2-T cells into a three-dimensional melanoma spheroid model. Of interest, exposure of co-cultures of Vγ9Vδ2-T cells and patient-derived tumor suspensions to the PD-L1xVδ2 bsTCE resulted in upregulation of activation markers on tumor-infiltrated CD4+ and CD8+ T cells and lysis of PD-L1+ myeloid cells with a shift in the myeloid compartment from macrophage-like cells to more mature dendritic cells with costimulatory molecule expression. CONCLUSIONS: A PD-L1xVδ2 bsTCE was generated that acts as PD-1/PD-L1 immune checkpoint inhibitor, enhances Vγ9Vδ2-T cell activation, infiltration and tumor lysis and reshapes the tumor microenvironment towards a more proinflammatory state. By targeting both PD-L1 expressing tumor and myeloid cells, it addresses key challenges of current therapies and thereby offers a promising novel therapeutic strategy.