NK cell education and adhesion molecules : implications for immunotherapy

Sammanfattning: NK cells have a key role in immune surveillance. They detect aberrant expression of stressinduced molecules and MHC class I molecules with an array of activating and inhibitory receptors. Inhibitory receptors are not only important during the effector phase to guarantee tolerance towards self cells, but also provide the means to learn what self is, during a process termed NK cell education. This mechanism ensures that only NK cells expressing inhibitory receptors for self MHC-I become functionally responsive to react to down-regulation of MHC-I in otherwise normal cells. Recent evidence suggests that education is tunable and changes according to the net signaling input an individual NK cell receives. In paper I, we studied such retuning effects of NK cell education in different settings relevant to immunotherapy and asked whether they would impair anti-tumor responses. We used two mouse models of clinically relevant approaches to interfere with inhibitory receptor signaling, antibody mediated checkpoint blockade of inhibitory receptors, and allogeneic adoptive transfer. Despite adaptation of NK cell responsiveness resulting in tolerance to healthy MHC-I- cells, we could still detect efficient anti-tumor responses. To date, there is no known marker that distinguishes educated from hypo-responsive NK cells. In paper II we investigated the relationship between the adhesion molecule and activating receptor DNAM-1 and the education state of mouse NK cells. DNAM-1 expression appeared early in development prior to inhibitory receptors for MHC-I. We observed in several experimental situations that DNAM-1 levels are higher on NK cells that have Ly49 inhibitory receptors for self MHC-I, and thus are considered to be educated. Furthermore, we found an MHC-I independent, tight correlation between DNAM-1 expression and the inhibitory receptor NKG2A. Based on the data, we propose a model for how DNAM-1 levels may influence NK cell functions in different developmental and functional stages. While NK cell education ensures strong tolerance to most normal cells, one exception is the interaction with DCs. Cross-talk between NK cells and DCs can lead to functional maturation of both cell types, but may also result in NK cell-mediated killing of the DC under certain conditions. In paper III, we analyzed the impact of an inflammatory environment on this interaction. We report that non-inflammatory DCs are relatively resistant to NK cell killing, while under inflammatory conditions, NK cells kill DCs via recognition of CD155 and ICAM-1 by DNAM-1 and LFA-1. In paper IV we explored the possibility of NK cell activation by vaccination with alphagalactosyl-ceramide (αGalCer) loaded exosomes directed to stimulate iNKT cells. We observed a preferential proliferation of educated NK cells in response to iNKT cell activation, as opposed to published results in virus models where preferential proliferation of uneducated cells is observed. This results in increased missing self killing of normal cells, and increased killing of both MHC-I+ and MHC-I- tumor cells. Together, these findings contribute to illustrate the role of NK cells and NK cell education in immune surveillance against cancer cells. The thesis discusses these results in light of NK cells and their application in immunotherapy.