This suggested that cross-linking of NKG2D was sufficient for rejection of ligand-expressing tumor cell lines. Ab blocking of NKG2D inhibited cytotoxicity against NKG2D-L-expressing tumor cells indicating a direct activating rather than a costimulating function of NKG2D 18. However, a possible role of other ligands could BTK animal study not be excluded in these studies. Direct evidence for a role of NKG2D receptors in tumor surveillance was provided by a recent study where onset of spontaneous malignancies

was accelerated when mice were devoid of NKG2D expression 19. Likewise, it has not been clearly defined if MHC class I-mediated signals are necessary or sufficient for NK-cell activity. In primary leukemias, lack of inhibition was not sufficient to confer cytotoxicity 20, but cell lines were rendered NK-resistant by HLA-C transfection, thus indicating a requirement of MHC class I down-regulation for NK-cell activity 21. On the other hand, cells displaying normal levels of MHC class I were susceptible to NK-cell lysis if effector cells became otherwise activated 22, 23. A clue to an understanding of these data might be a two-signal

requirement of NK-cell activation. In resting but not pre-activated NK cells, NKG2D was identified as a coactivation signal that needed coengagement of other receptors, such as 2B4 and natural cytotoxicity Temsirolimus receptors (NCR) 24. In another study, NK-dependent lysis of some tumors was only dependent

on NCR, whereas in other tumors, synergistic Erastin molecular weight effects of NCR and NKG2D were found 25. Recently, a sequential NK-cell activation process was proposed 26. In this model, activation of resting NK cells required a priming signal that was provided by IL-2 or by unknown ligands of tumor cells independently of IL-2, and a subsequent triggering event that was mediated by CD69. MHC class I down-regulation was not needed for tumor-induced NK activity in this study 26. Resistance of tumors might either arise through a lack of priming of NK cells (type 1 evasion) or by the inability of the tumor to deliver triggering signals to already primed NK cells (type 2 evasion) 26. Reports suggesting a two-signal requirement for NK-cell activation were only based on in vitro studies, and the role of NKG2D that was described as an NCR in earlier studies 27, 28 was not addressed in the context of the two-stage model 26. We were therefore interested in the mechanisms of NK-cell activation in tumor surveillance in vivo and we specifically investigated the role of “missing self” and of NKG2D/ligand interactions as well as the mechanisms underlying tumor escape. We previously showed that missing self can induce strong and protective NK-cell responses in a tumor transplantation model 6, but this may not reflect the situation in endogenous tumors.