Furthermore, the iTreg cell induction protocol was modified and the cell line Colo699 was used instead of PCI-13. In all conditions, iTreg cells significantly enhanced NK cell degranulation (Fig. 3C). We also added the supernatant of anti-CD3-activated iTreg cells to NK cells to evaluate an iTreg cells derived soluble factor responsible for iTreg cell–NK cell interaction but could not detect significant effects on NK cell degranulation. Further, iTreg cells were tested negative for surface expression of potentially NK

activating NKG2D ligands, ULBP1, ULBP2, ULBP3, MICA, and MICB (data not shown). Next, we sought to identify the cytolytic DMXAA effector mechanism responsible for the increased cytotoxicity of NK cells when co-cultured with iTreg cells. To exclude that iTreg cells themselves exhibit cytotoxicity PD98059 concentration on tumor cells, we tested iTreg cells for the expression of perforin and FasL as well as their capacity to lyse tumor cells. iTreg cells

neither expressed perforin nor FasL nor induced tumor cell lysis when co-cultured with tumor cells (data not shown). To investigate if the observed enhanced cytotoxicity of NK cells is mediated by soluble factors, we added the supernatant of iTreg cells/NK cells co-cultures to the tumor cells but could not detect any tumor cell lysis (data not shown). These observations suggested that tumor cell lysis is mediated by a direct cell–cell interaction between NK cells and target cells; thus, we used concanamycin A (CMA) and inhibitory antibodies to block perforin-, FasL-, and TRAIL-mediated

cytotoxicity, respectively. As depicted in Fig. 4, tumoricidal activity of non-stimulated Lck NK cells in the absence of iTreg cells was predominantly mediated by perforin. This is illustrated by the reduction of tumor cell lysis by CMA (Fig. 4A), while inhibitory antibodies, which blocked FasL and TRAIL had no effect (Fig. 4B and C). Consistent with our data in Fig. 3, NK cells showed a significantly higher cytotoxicity towards tumor cells when they were co-cultured with iTreg cells overnight prior to the addition of 51-Cr-labeled target cells (triangles in Fig. 4A–C). This effect was significantly reduced if NK cells were pretreated with CMA or inhibitory antibodies to FasL, while anti-TRAIL antibodies had a minor or no effect (Fig. 4A–C). In summary, natural cytolytic activity of NK cells is mainly mediated by perforin, while death receptor pathways like FasL and TRAIL play a minor role. In contrast, iTreg cell-induced cytotoxicity of NK cells is mediated by perforin and FasL-associated pathways. In the next series of experiments, we wanted to further characterize the phenotype of NK cells after co-culture with iTreg cells to potentially explain increased NK cell activity.