Effect of aberrant N-glycosylation of PD-L1 on the potency of anti-PD-1 treatments.

e14599 Background: Immune checkpoint inhibitors (ICIs) targeting the PD1/PDL1 axis have revolutionized the way we treat cancer patient. Yet, response rate to these treatments remains low. Post translational modifications of PD-L1 in the form of N-glycosylation play a crucial role in both receptor-ligand interaction and in the potency of PD1/PDL1 ICIs. Understanding the effect of each N-glycosylation site is crucial for patient stratification for treatment groups and may help to increase response rate to anti-PD1/PDL1 ICIs. We recently developed the Immune-Checkpoint Artificial Reporter (IcAR) technology to assess specific ICI potency for target tumors including clinical samples. Using IcAR expressing recombinant human PD1 receptor we can analyze clinical response of patients to PD1-PDL1 blocking drugs and analyze the glycan-modified PDL1 expressing cancer cells. Methods: We have created two cell lines (MDA-MB231 and MCF7) overexpressing PD-L1 in its wildtype form, single mutated glycosylation sites (N35A, N192A, N200A, N219A) and all sites mutated together (Nx4). Using IcAR-PD1 we measured the functional binding of the different mutated cell lines, cell derived xenografts (CDX) and fixed cell samples to evaluate baseline binding. Later we have incubated IcAR-PD1 and mutated cell liens together with multiple clinically used PD1/PDL1 ICIs to evaluate the inhibition of interaction of each ICI. Results: While staining and functional interaction of PDL1 and single mutants is similar, removing all four N-glycosylation sites (Nx4) decreased both the staining of PDL1 and functional binding. Interestingly, blocking capacity of different ICIs was impaired by N35A PDL1 mutation, and was absolute through all concentration (2.5 to 40 ug/mL) when employed on the Nx4 cell lines. While anti-PDL1 ICIs were managed to block PD-1/PD-L1 interaction trough all cell lines and mutants, the inhibitory capacity of anti-PD1 ICIs was impaired by glycosylation, especially on N35. Conclusions: Aberrant glycosylation of PD-L1 impair the potency of anti-PD-1 treatment in cancer cell lines. These results might explain why patients with negative PD-L1 staining (Nx4), might respond to ICIs, and suggest an explanation of mechanism in which patients with high PD-L1 staining (i.e. N35), might not respond to anti-PD-1 treatments. Our results shed a light on the mechanism of ICI resistance and response – using this knowledge patient will be stratify better into treatment groups, getting oncology a step closer to precision immunology.