2D kinetic study of of PD-1 interaction and its inhibition of T-cell antigen recognition

Abstract

Programmed death-1 (PD-1) is an immune-checkpoint receptor with its primary function to maintain peripheral tolerance of the adaptive immune responses. The importance of PD-1 is evidenced by its deficiency leading to autoimmune disorders, its central role in the identification and restoration of the exhausted phenotypes of antigen-specific T cells, and the great success in targeting this pathway for cancer immunotherapy. To better understand the fundamental question as how PD-1 achieves the potent but well-controlled inhibition, we applied kinetic approaches focusing on its in situ ligand binding characteristics, and the early impact on antigen recognition by the T cell receptor (TCR) and coreceptor CD8. Different from the weak three-dimensional (3D) affinities measured in solution using purified PD-1 and ligands, the two-dimensional (2D) affinities of ligand binding to mouse and human PD-1 expressed on cell membrane span a range from middle to strong, whereas PD-L1–B7-1 binding is much weaker. Comparison of 2D and 3D affinities of PD-1 with B7-1–CD28 and B7-1–cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) as well as others reveals distinct kinetic mechanisms underlying the inhibition of PD-1 and CTLA-4, and differential enhancement of in situ ligand binding for various receptors by the cellular environment. By integrating the 2D kinetic analysis of PD-1 with TCR and CD8, we probed an apparent “negative cooperativity” between these two axis, manifested as reduced molecular bond number and bond lifetime when respective ligands were co-presented. Examination with force spectroscopy suggested the “negative cooperativity” to be the net outcome of suppressed “positive cooperativity” between TCR and CD8. Moreover, the dependence of this suppression on Src homology region 2 domain-containing phosphatase-2 (SHP-2) and lymphocyte-specific protein tyrosine kinase (Lck) further identified it as a “binding-signaling-binding” feedback mechanism representing fine-tuning of antigen recognition by costimulatory/coinhibitory receptors via targeting the TCR–CD8 machinery. In situ kinetic analysis also indicated the existence of a novel binding partner for human PD-L1, which was identified and validated to be CD222. The hPD-L1–CD222 interaction consists of both protein-protein and lectin-carbohydrate binding components, and is stronger than hPD-L1–PD-1 according to its higher 3D and 2D affinity/avidity. Most importantly, CD222 is upregulated on the plasma membrane of activated T cells and competes with PD-1 for hPD-L1, suggesting potentially significant functions on T cells at least in part by perturbing the hPD-L1–PD-1 interaction. Overall, our results provide an in depth understanding of the in situ interaction and function of PD-1, and uncover a novel interaction of hPD-L1–CD222, highlighting the complexity and significance of costimulatory/coinhibitory molecules in modulating T cell responses.