Role of HLA-G in shaping Natural Killer cell biology in cancer

Abstract

The human leukocyte antigen-G (HLA-G) is a non-classic MHC class I molecule. It is called “non-classic” because it differs from classic MHC class I molecules due to its very low amount of polymorphism and highly tissue-restricted expression (Figure 1). HLA-G have a role in this tolerogenic immune response that takes place at the maternal-fetal interface, by inhibiting cytotoxicity and cytokine production by NK and T cells. Moreover, apart from its physiological expression, HLA-G can also be expressed during pathological conditions such as cancer, where it also downregulates the immune response and allows the tumor progression. The natural killer (NK) cells are innate immune effectors that can spontaneously kill tumor cells through the secretion of cytolytic molecules such as granzymes and perforins. Moreover, they can be subdivided due to the cell surface density of CD56 in CD56bright and CD56dim. The CD56dim subset is the most abundant in peripheral blood (approximately 90%) and expresses high levels of CD16 whereas the minority of NKs are CD56bright CD16dim/negative. The CD56bright subset is described to have a regulatory role, whereas the CD56 dim is more cytotoxic (Figure 2). The aim of this work was to characterize the role of HLA-G in the regulation of the IS focusing in the NK cells and to study the biological implications in the context of cancer. Moreover, we explored the possibility of targeting the HLA-G molecule. We designed and performed an in vitro co-culture (CC) system to study the consequences of the presence or absence of HLA-G in the target cells (JEG3) in the immune effector cells (NK92). We observed an increase of the CD56 marker when the NK92 cells were co-cultured in the presence of HLA-G. We also developed another in vitro culture model to check the effect of the soluble-HLA-G in the NK92 cells. With this aim, we generated conditioned media (CM) from JEG3 and HuP-T3 cells and characterized the level of secreted-HLA-G present by ELISA. We used this CM as a physiological source of HLA-G and cultured the NK cells with it. We observed as well as in the CC experiments an increase of the CD56 marker in presence of the HLA-G (NK92+ CM form JEG3 WT). Importantly, we also observed the down-regulation of IFN-γ, granzyme and perforin secretion. We confirmed these results using recombinant-HLA-G protein. Also, the granzyme B and perforin secretion was reverted, however, we did not observe a reversion of the IFNγ secretion pattern. Furthermore, we tried to block the HLA-G effect by the addition of a monoclonal anti-HLA-G antibody (Thermofisher, clone 87G). However, due to the lack of effect that we obtained with this antibody and given cross-reactivity previously described in bibliography, we decided to design our own anti-HLA-G blocking antibody. We targeted a region in the alpha-3 domain described to be responsible for the ILT-2 receptor binding in order to functionally block the effect of HLA-G. The antibody we obtained, worked in a dose-dependent manner in NK92 reverting the effects produced by the soluble HLA-G. However, we did not have the same results with the pb-NK cells. In this work we have shown the implication of HLA-G in the regulation of the NK cell maturation and function. This is very important given the role of the NK cells in the context of the immune-modulation in cancer. Particularly, we have described for the first time, the role of HLA-G in the modulation of the CD56 surface marker. Interestingly, HLA-G induces a NK cell phenotype that recapitulates the one present at the decidua in the pregnancy context. This parallelism between cancer and pregnancy demonstrate us that once again cancer hijacks a preexisting mechanism for its own benefit. Moreover, we have explored the possibility of reverting this effect produced in NK cells using an anti-HLA-G antibody. However we described only partial effects, highlighting the fact that further work is required to evaluate HLA-G as a potential therapeutic target.