Structural, functional, and evolutionary differences between PD-L1 and PD-L2

Abstract

Abstract The PD-1/PD-L1/PD-L2 axis is a critical immune checkpoint that tips immune responses towards tolerance. This axis has been heavily targeted in cancer immunotherapy with monoclonal antibodies, which block receptor-ligand interactions. PD-L1 is constitutively expressed at low levels and is induced on nearly all tissues upon IFN-gamma signaling. In contrast, PD-L2 expression is restricted mainly to antigen presenting cells (APC)s, such as dendritic cells and macrophages. PD-L2 has a 4-fold stronger affinity for PD-1 than does PD-L1. It has been proposed that W110 of PD-L2 accounts for this affinity difference; the corresponding PD-L1 residue is A121. To test this hypothesis, we produced, in a mammalian expression system, a set of PD-L1 and PD-L2 proteins with the W and A swapped, and studied their binding to PD-1 using surface plasmon resonance. Surprisingly, rather than decreasing affinity as expected, a W110A substitution improved affinity by decreasing the off-rate. We propose a novel structural mechanism for PD-L2’s affinity advantage, one that interestingly emerged upon the evolution of placental mammals. We examined the functional effect that these structural alterations have on the PD-Ligands ability to inhibit T cell activation and found a correlation. Additionally, we discovered that PD-Ligands homo- and hetero-dimerize in a manner similar to other B7-family members, with PD-L2 displaying the strongest homo-dimerization. Collectively, these data support our hypothesis that PD-L2 expression on APCs provides an inhibitory advantage, one that may heighten the tolerogenic hurdle that an immune response must overcome during the priming phase.