N-glycosylation of PD-L1 modulates the efficacy of immune checkpoint blockades targeting PD-L1 and PD-1

The PD-L1/PD-1 pathway is crucial for immune regulation and has become a target in cancer immunotherapy. However, in order to improve patient selection for immune checkpoint blockade (ICB) therapies, better selection criteria are needed. This study explores how the N-glycosylation of PD-L1 affects its interaction with PD-1 and ICB efficacy, focusing on its four N-linked glycosylation sites: N35, N192, N200, and N219.

Human PD-L1 glycosylation mutants-at each individual site or at all four sites together (Nx4)-were tested for their functional interaction with PD-1 using an artificial immune checkpoint reporter assay (IcAR-PD1). The blocking efficacy of anti-PD-L1 and anti-PD-1 antibodies was evaluated using human breast cancer cell lines (MDA-MB231 and MCF7), as well as A375 melanoma and A549 lung carcinoma cells expressing the glycosylation mutants. Results were validated through ex vivo activation and cytotoxicity assays using human CD8+ T cells.

The binding of the PD-L1_N35A mutant to PD-1 was not effectively blocked by anti-PD-L1 and anti-PD-1 ICBs. In contrast, high blocking efficacy of PD-L1 binding to PD-1 was obtained at minimal ICB concentrations when PD-L1 did not express any glycosylation site (PD-L1_Nx4 mutant). The PD-L1_N35A mutant produced elevated levels of PD-L1 as a soluble (sPD-L1) and extracellular vesicles (EV)-bound molecule; in contrast, the PD-L1_Nx4 mutant had lower sPD-L1 and EV levels. PD-L1 glycosylation status influenced the ability of PD-L1 interactions with PD-1 to down-regulate T-cell activation and cytotoxicity, with the PD-L1_N35A mutant showing the lowest levels of T cell functions and the PD-L1_Nx4 mutant the highest.

The N-glycosylation of PD-L1 at all four sites interferes with the ability of anti-PD-L1 and anti-PD-1 ICBs to block PD-L1 interactions with PD-1; in contrast, glycosylation at the N35 site enhances ICB blocking efficacy. These effects are connected to the ability of sPD-L1 to compete with ICB binding to PD-L1 or PD-1. Thus, assessing PD-L1 glycosylation, beyond expression levels, could improve patient stratification and outcomes.