The inhibitory receptor CD94/NKG2A on CD8+ tumor-infiltrating lymphocytes in colorectal cancer: a promising new druggable immune checkpoint in the context of HLAE/ß2m overexpression.

We previously demonstrated that HLA-E/ß2m overexpression by tumor cells in colorectal cancers is associated with an unfavorable prognosis. However, the expression of its specific receptor CD94/NKG2 by intraepithelial tumor-infiltrating lymphocytes, their exact phenotype and function, as well as the relation with the molecular status of colorectal cancer and prognosis remain unknown. Based on a retrospective cohort of 234 colorectal cancer patients, we assessed the expression of HLA-E, ß2m, CD94, CD8, and NKp46 by immunohistochemistry on tissue microarray. The expression profile of HLA-E/ß2m on tumor cells and the density of tumor-infiltrating lymphocytes were correlated to the clinicopathological and molecular features (Microsatellite status, BRAF and RAS mutations). Then, from the primary tumors of 27 prospective colorectal cancers, we characterized by multiparameter flow cytometry the nature (T and/or NK cells) and the co-expression of the inhibitory NKG2A or activating NKG2C chain of ex vivo isolated CD94+ tumor-infiltrating lymphocytes. Their biological function was determined using an in vitro redirected cytolytic activity assay. Our results showed that HLA-E/ß2m was preferentially overexpressed in microsatellite instable tumors compared with microsatellite stable ones (45% vs. 19%, respectively, p = 0.0001), irrespective of the RAS or BRAF mutational status. However, HLA-E/ß2m+ colorectal cancers were significantly enriched in CD94+ intraepithelial tumor-infiltrating lymphocytes in microsatellite instable as well as in microsatellite stable tumors. Those CD94+ tumor-infiltrating lymphocytes mostly corresponded to CD8+ αß T cells, and  to a lesser extent to NK cells, and mainly co-expressed a functional inhibitory NKG2A chain. Finally, a high number of CD94+ intraepithelial tumor-infiltrating lymphocytes in close contact with tumor cells was independently associated with a worse overall survival. In conclusion, these findings strongly suggest that HLA-E/ß2m-CD94/NKG2A represents a new druggable inhibitory immune checkpoint, preferentially expressed in microsatellite instable tumors, but also in a subgroup of microsatellite stable tumors, leading to a new opportunity in colorectal cancer immunotherapies.

IL-12-dependent inducible expression of the CD94/NKG2A inhibitory receptor regulates CD94/NKG2C+ NK cell function.

The inhibitory CD94/NKG2A and activating CD94/NKG2C killer lectin-like receptors specific for HLA-E have been reported to be selectively expressed by discrete NK and T cell subsets. In the present study, minor proportions of NK and T cells coexpressing both CD94/NKG2A and CD94/NKG2C were found in fresh peripheral blood from adult blood donors. Moreover, CD94/NKG2A surface expression was transiently detected upon in vitro stimulation of CD94/NKG2C+ NK cells in the presence of irradiated allogeneic PBMC or rIL-12. A similar effect was observed upon coculture of NKG2C+ NK clones with human CMV-infected autologous dendritic cell cultures, and it was prevented by an anti-IL-12 mAb. NKG2A inhibited the cytolytic activity of NKG2C+ NK clones upon engagement either by a specific mAb or upon interaction with a transfectant of the HLA class I-deficient 721.221 cell line expressing HLA-E. These data indicate that beyond its constitutive expression by an NK cell subset, NKG2A may be also transiently displayed by CD94/NKG2C+ NK cells under the influence of IL-12, providing a potential negative regulatory feedback mechanism.

NKG2A inhibits invariant NKT cell activation in hepatic injury.

Activation of invariant NKT (iNKT) cells in the liver is generally regarded as the critical step for Con A-induced hepatitis, and the role of NK cell receptors for iNKT cell activation is still controversial. In this study we show that blockade of the NKG2A-mediated inhibitory signal with antagonistic anti-NKG2A/C/E mAb (20d5) aggravated Con A-induced hepatitis in wild-type, Fas ligand (FasL)-mutant gld, and IL-4-deficient mice even with NK cell and CD8 T cell depletion, but not in perforin-, IFN-gamma-, or IFN-gamma- and perforin-deficient mice. Consistently, 20d5 pretreatment augmented serum IFN-gamma levels and perforin-dependent cytotoxicity of liver mononuclear cells following Con A injection, but not their FasL/Fas-dependent cytotoxicity. However, blockade of NKG2A-mediated signals during the cytotoxicity effector phase did not augment cytotoxic activity. Activated iNKT cells promptly disappeared after Con A injection, whereas NK1(-) iNKT cells, which preferentially expressed CD94/NKG2A, predominantly remained in the liver. Pretreatment with 20d5 appeared to facilitate disappearance of iNKT cells, particularly NK1(-) iNKT cells. Moreover, Con A-induced and alpha-galactosylceramide-induced hepatic injury was very severe in CD94/NKG2A-deficient DBA/2J mice compared with CD94/NKG2A-intact DBA/2JJcl mice. Overall, these results indicated that a NKG2A-mediated signal negatively regulates iNKT cell activation and hepatic injury.