PD-L2 Provides a Unique Downstream Inhibitory Signature in T cells, Distinct from That of PD-L1, Making It a Worthwhile Immunotherapy Target
Introduction
Checkpoint blockade therapies, especially those targeting the PD-1/PD-L1 axis, have revolutionized cancer treatments, leading to unprecedented results in progression free survival (PFS) and overall survival (OS). Despite these advances, numerous patients fail to respond or relapse following an initial period of response. Poor detection reagents and lack of equivalence between mouse and human function has fostered a lack of appreciation for the significant role of PD-L2 in tumor immune suppression. Given its unique expression pattern relative to PD-L1, we hypothesize that PD-L2 engages PD-1 in a distinct manner which propagates a unique downstream inhibitory signature in T cells.
Methods
Jurkat T cells with a Nuclear factor of activated T cells (NFAT) luciferase reporter were cocultured with Chinese hamster ovary (CHO) cells transduced to overexpress huPD-L1 or huPD-L2. We interrogated the difference in downstream PD-1 inhibitory signaling for PD-L1 and PD-L2 from a proteomic, transcriptomic, apoptotic, mitotic, and functional standpoint. A bioluminescent readout in the Jurkat assay demonstrated the effects of PD-L1 and PD-L2 dampening T cell activation. Western blots allowed for the temporal analysis of phosphoproteomic signatures, and microarrays along with gene set enrichment analysis (GSEA) helped elucidate the ensuing transcriptomic effects. Similar experiments were performed on human T cells isolated from peripheral blood mononuclear cells (PBMCs) as well as therapeutic chimeric antigen receptor (CAR) T cells. Recovery assays in the form of anti-CD3/CD28 activation post PD-L1/PD-L2 inhibition also allowed for the assessment of marked differences in functional cytokine production (IL-2, IFNγ, and TNFα).
Results
Unlike murine PD-L2, the human ortholog confers a purely inhibitory signature in T cells explaining its significant overexpression in human tumors. When compared head-to-head, we discovered that PD-L2 provides a distinct, and more readily reversible inhibitory signal compared to PD-L1. While both PD-Ligands equally inhibited NFAT activation and cytokine production through PD-1, PD-L1 and PD-L2 differed in the temporal dephosphorylation of membrane proximal proteins in the CD3ζ chain signaling axis. GSEA provided valuable information with regards to each inhibitory signature, with PD-L1 having a greater effect on the mTORC pathway and PD-L2 negatively affecting microtubule spindle formation. Propidium iodide studies showed that PD-L2 preferentially accumulates T cells in the S phase of the cell cycle, while PD-L1 arrests T cells in the G1/G0 phase. PD-L2 inhibited T cells could respond to CD3/CD28 restimulation by restoring high-level IL-2 production, whereas those inhibited with PD-L1 could not.
Conclusion
Gene duplication events allow for sub-functionalization and, given its unique expression pattern, we demonstrate that PD-L2 confers an inhibitory signature through PD-1 distinct from that engaged by PD-L1. Like PD-L1, PD-L2 drives T cells to an inhibited state; however, PD-L2 inhibited T cells are capable of recovering partial effector function following restimulation while those exposed to PD-L1 are not. We speculate that during primary T cell responses, induction of PD-L2 may set aside a “ready reserve” of T cells that can be reactivated if needed if pro-inflammatory stimulations persist.