The Journal of Immunology

Enhanced In Vivo Growth of Lymphoma Tumors in the Absence of the NK-Activating Receptor NKp46/NCR11

Gili G. Halfteck, Moran Elboim, Chamutal Gur, Hagit Achdout, Hormas Ghadially, and Ofer Mandelboim2

The in vitro elimination of virus-infected and tumor cells by NK cells is regulated by a balance between signals conveyed via specific inhibitory and activating receptors. Whether NK cells and specifically the NK-activating receptor NKp46 (NCR1 in mice) are directly involved in tumor eradication in vivo is still largely unknown. Since the NKp46/NCR1 tumor ligands have not been identified yet, we use a screening technique to identify functional ligands for NKp46/NCR1 which is based on a cell reporter assay and discover a NCR1 ligand in the PD1.6 lymphoma line. To study whether NKp46/NCR1 is important for the eradication of PD1.6 lymphoma in vivo, we used the Ncr1 knockout Ncr1gfp/gfp mice generated by our group. Strikingly, all Ncr1 knockout mice developed growing PD1.6 tumors, whereas initial tumor growth was observed in the wild-type mice and tumors were completely rejected as time progressed. The growth of other lymphoma cell lines such as B10 and EL4 was equivalent between the Ncr1 knockout and wild-type mice. Finally, we show that PD1.6 lymphoma cells are less killed both in vitro and in vivo in the absence of NKp46/NCR1. Our results therefore reveal a crucial role for NKp46/NCR1 in the in vivo eradication of some lymphoma cells. The Journal of Immunology, 2009, 182: 2221–2230.

t was hypothesized that the immune system surveys the body reduced resistance to transplanted tumor cell lines (20–22). for nascent malignancies, eliminating some tumors and slow- Subsequent studies demonstrated that mice carrying the ho- I ing the growth of others (1). This hypothesis is supported by mozygous beige mutation, which are defective in granule exo- the findings that humans with congenital or acquired immunode- cytosis and have a profound defect in NK cell cytolytic func- ficiency have a significantly higher incidence of malignancies (2, tion, develop virus-induced and carcinogen-induced tumors at a 3). Consistent with such a role, various components of the immune higher frequency (23–25). However, these studies did not prove system, such as CTLs, NK cells, and Abs can exert potent activity a role for NK cells in tumor immunity, since the defects caused against different types of tumors in vitro (4–6). by the beige mutation are not limited to NK cells. Other studies NK cells are bone marrow-derived lymphocytes of the innate demonstrated an increased rate of carcinogen-induced sarcomas immune system comprising ϳ5–15% of PBLs. NK cells play an (and of tumors induced by oncogenic viruses) in mice lacking important role in the early elimination of virus-infected cells, bac- perforin, the pore-forming granule involved in cytotox- teria, intracellular parasites, and tumor cells (7–14). In addition, icity by NK cells and CTLs (26). In contrast to perforin-defi- NK cells have been shown to also perform several noncytotoxic cient mice, mice genetically deficient for CD8ϩ T cells did not functions, including: Ag presentation (15), activation of dendritic show significant defects in the control of carcinogen-induced cells (16), an important regulatory function in the maternal-fetal tumor growth, suggesting a possible role of NK cells in tumor interface (17), and secretion of immunomodulatory cytokines such surveillance (26). Importantly, perforin-deficient mice also de- ␥ ␣ as IFN- , TNF- , and chemokines causing T cells to shift to a TH1 veloped spontaneous disseminated lymphomas at a much higher phenotype (18). frequency than immunocompetent control mice (27). Whether The antitumor cell activity of NK cells was the hallmark of the perforin dependence of tumor resistance in normal mice their original discovery (19). However, only a limited number reflects the activity of NK cells or T cells, or both, was not of in vivo studies investigated whether NK cells in general and determined in this system. NK receptors in particular are involved in tumor eradication. NK cell’s activation is governed by numerous receptors Studies involving depletion of NK cells in mice often showed (some of which are activating while others are inhibitory) that recognize various classes of cell surface ligands (14, 28). The balance between these countervailing signals tightly regulates Lautenberg Center for General and Tumor Immunology, The Hebrew University, Hadassah Medical School, Jerusalem, Israel NK cell activation (14, 28). Many of the inhibitory receptors Received for publication June 11, 2008. Accepted for publication December 10, 2008. expressed by NK cells are specific for MHC class I molecules as predicted by the “missing self-recognition” hypothesis (29) The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance (although other, non-MHC-restricted receptors were also iden- with 18 U.S.C. Section 1734 solely to indicate this fact. tified (30–32)). In contrast, activating NK cell receptors recog- 1 This work was supported by the US–Israel Bi-national Science Foundation, the nize self, stress-inducible, pathogen-derived, and tumor ligands, Israeli Cancer Research Foundation, the Israeli Science Foundation, the European Consortium (Grants MRTN-CT-2005 and LSCH-CT-2005-518178), and the AICR. whose expression is induced or up-regulated in target cells (14, O.M. is an Edward Crown Professor of Molecular Immunology. 28). In that respect, a significant breakthrough in understanding 2 Address correspondence and reprint requests to Prof. Ofer Mandelboim, The He- NK cell activation was made following the identification of brew University, Hadassah Medical School, P.O.B. 12272 Jerusalem, Israel. E-mail three novel, NK-specific activating receptors, including NKp30, address: [email protected] NKp44, and NKp46 and its mouse ortholog NCR1 (33–35), Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 which are collectively known as natural cytotoxicity receptors www.jimmunol.org/cgi/doi/10.4049/jimmunol.0801878 2222 ENHANCED TUMOR GROWTH IN THE ABSENCE OF Ncr1

(NCRs).3 Among the NCRs, NKp46 and its mouse ortholog Fusion , Abs, and flow cytometry NCR1 are constitutively expressed by NK cells and, in fact, are The NCR1-Ig, NKp46-Ig, NKp46 D1-Ig, and NKG2D-Ig fusion proteins the best marker which distinguishes NK cells from other im- were generated in COS-7 cells and purified by affinity chromatography mune cells in all tested mammals, including humans (35), mice using a protein G column, as previously described (44). The staining of cell (36, 37), monkeys (38), rats (39), and bovines (40). lines was visualized using a secondary PE-conjugated goat anti-human Ab Several in vitro studies have demonstrated that NKp46 in hu- (Jackson ImmunoResearch Laboratories). Abs used in this work included: a hybridoma-producing mAb 12CA5 (anti-HA tag), a polyclonal Ab mans and NCR1 in mice are major killer receptors involved in the against mouse NKp46 AF2225 (R&D Systems), staining of both Abs was recognition and killing of tumor cells (5, 41). In keeping, a low visualized using fluorescein-conjugated goat anti-mouse Ab and fluores- concentration of NKp46 has been observed in several types of cein-conjugated donkey anti-goat Ab (Jackson ImmunoResearch Labora- malignancies, including acute myeloid leukemia (42). We have tories), respectively, PE-conjugated anti-CD3 Ab, and negative control mouse IgG1/R-PE (DakoCytomation). Additionally, a hybridoma-produc- previously demonstrated that viral hemagglutinins (HA) of various ing mAb PK136 (anti-NK1.1) was used for the depletion of NK cells influenza viruses are recognized by NKp46 and NCR1 (14, 37, 43, in vivo. 44); however, a cellular tumor ligand for NKp46 or NCR1 still BW assay remains to be identified. To study the in vivo function of NKp46, we have previously For measurements of IL-2 production resulting from the interaction be- generated a knockout mouse in which a GFP cassette was tween NCR1 and ligands expressed by target tumor cell lines, 50,000 BW ␨ “knocked in” into the Ncr1 locus, thereby rendering Ncr1 non- or BW/NCR1- cells were coincubated with 50,000 irradiated (3000 rad) gfp/gfp cells of various mouse cell lines for 48 h at 37°C and 5% CO2. Superna- functional (Ncr1 ) (37). As we have previously reported, tants were collected and the level of IL-2 was quantified by using com- influenza infection was lethal in Ncr1gfp/gfp mice in both 129/Sv mercially available anti-mouse IL-2 mAbs (BD Pharmingen) and standard and C57BL/6 backgrounds, demonstrating that the recognition ELISA. of HA by NKp46/NCR1 is essential for influenza virus eradi- NK cell depletion cation in vivo (37). That being said however, the direct role of Fifty micrograms of anti-NK1.1 Ab in a volume of 300 ␮l of sterile PBS NKp46/NCR1 in the elimination of tumor cells in vivo has not ϩ was injected i.v. into the tail vein of C57BL/6 Ncr1gfp/ mice. Sterile PBS been clarified. The results obtained in these mice regarding tu- was injected as a control. Depletion was verified 2 days after injection: mor growth were much less conclusive, as we have observed mice were bled; RBCs were lysed (using ACK buffer), and GFP-positive reduced clearance of the RMAS tumor cells in vivo in Ncr1gfp/gfp NK cells were detected in peripheral blood cells by flow cytometry. NK mice in the 129/Sv background but not in the C57BL/6 back- depletion was further verified every 2 days during the experiments. ground (37). In vitro cytotoxicity assays In this study, we identified an unknown ligand for NKp46/ Heterozygous and Ncr1-knockout C57BL/6 mice were injected i.p. with NCR1 expressed on the PD1.6 lymphoma line and show, both in 200 ␮g of poly(I):poly(C) (Sigma-Aldrich). Spleens were removed 18 h vitro and in vivo, that eradication of PD1.6 lymphoma is impaired later and NK cells were isolated from extracted splenocytes using a in the absence of NCR1. Finally, we demonstrate that the mech- mouse NK isolation (Miltenyi Biotec) and an AutoMACS instru- 5 anism by which NK cells, and NKp46/NCR1 specifically, are in- ment according to the manufacturer’s instruction. NK cells (5 ϫ 10 ) cells were coincubated with the indicated target cells in triplicate at a volved in the eradication of the PD1.6 lymphoma cells is direct ratio of 1:1 in the presence of 0.1 ␮g of a allophycocyanin-conjugated cytotoxicity rather than secretion of IFN-␥. anti-CD107a Ab (1D4B; Southern Biotechnology Associates). Follow-

ing a 2-h incubation at 37°C and 5% CO2, cells were washed and an- alyzed by flow cytometry. Materials and Methods In vivo lung clearance assay Mice and in vivo tumor experiments For the in vivo cytotoxicity assay, we used a published fluorescence la- All experiments were performed using mice 6–8 wk old of the C57BL/6 beling method with a modification to avoid possible influence of intrinsic background. The generation of NKp46/NCR1 knockout mice in the GFP present in Ncr1gfp/gfp mice (37, 45). Cells were labeled with Vybrant C57BL/6 background was previously described (37). All of the ex- cell labeling solution (Molecular Probes). Subsequently, HeLa cells were periments reported herein were performed in a specific, pathogen-free labeled with Vybrant DiD; PD1.6 cells were labeled with Vybrant Dil. unit of the Hebrew University Medical School (Ein-Kerem, Jerusalem) Cells (2 ϫ 106) of each cell type were mixed in 400 ␮l of sterile PBS and in accordance with the guidelines of the ethics committee. For in vivo injected into the tail vein of C57BL/6 Ncr1ϩ/ϩ or Ncr1gfp/gfp mice. Lungs, tumor growth monitoring, mice were injected s.c. with 1 ϫ 106 tumor ␮ spleen, liver, lymph nodes, and kidneys were explanted either immediately cells in a volume of 200 l of sterile PBS. Tumor size was measured after injection or 3 h later. Single-cell suspensions were obtained (by using every 1–2 days after injection using a micrometer. Experiments were cell strainers) and fluorescence was analyzed by flow cytometry. The ratio repeated three to four times. For the detection of NK cells and T cells of the tested PD1.6 target cells to the internal control HeLa cells was ϫ 6 in tumors, mice were injected s.c. with 1 10 tumor cells and sacri- calculated. ficed 11 days later when tumors were removed and analyzed. Experi- ments were repeated twice. Cytokine secretion assay Heterozygous Ncr1ϩ/gfp and NCR1-deficient Ncr1gfp/gfp C57BL/6 mice Cells were injected i.p with 200 ␮g of poly(I):poly(C) (Sigma-Aldrich). Spleens were removed 18 h later and NK cells were isolated from extracted spleno- The cell lines used in this study were: for the in vivo study, we used the cytes using a mouse NK isolation kit (Miltenyi Biotec) and an AutoMACS C57BL/6-derived murine radiation leukemia virus-induced thymic lym- instrument according to the manufacturer’s instruction. Fifty thousand NK phomas PD1.6 and B10, the murine thymoma EL4. In addition, we used cells were coincubated with the indicated target cells at a ratio of 1:1 for the murine mastocytoma P815, the human epithelial carcinoma HeLa, the ␥ 48 h at 37°C and 5% CO2. Supernatants were collected and IFN- levels murine thymoma BW, the RMAS lymphoma, and the murine lymphoma were determined with a commercially available anti-mouse IFN-␥ mAb YAC-1. The P815, BW, PD1.6, EL4, B10, and YAC-1 cell lines were all (BD Pharmingen) and standard ELISA. grown in complete RPMI medium that was supplemented with 10% FCS and 0.05 mM ␤ME. The HeLa cell line was grown in complete DMEM medium that was supplemented with 10% FCS. Results Mouse lymphoma cell lines express ligand(s) for NKp46/NCR1 The importance of NK cells in general and that of NKp46 in par- 3 Abbreviations used in this paper: NCR, natural cytotoxicity receptor; HA, ticular in the in vitro killing of many tumors is well established hemagglutinin. (28, 46, 47). We have previously demonstrated that NCR1, the The Journal of Immunology 2223

FIGURE 1. Expression of NKp46/NCR1 ligands on mouse tumor cell lines. A–C, Binding of NCR1-Ig (A), NKp46-Ig (B), or NKp46 D1-Ig (C) fusion proteins to various mouse tumor cell lines. Gray shading, background staining with secondary Ab only; dark lines, specific staining with the indicated fusion protein. The various tumor cell lines are indicated. Data from one of four independent experiments are shown. D, Flow cytometry analysis of BW cells expressing HA-tagged NCR1␨ chimeric protein using anti-HA tag and anti-NCR1 mAbs. Gray shading, background staining with secondary Ab only; dark line, specific staining with the indicated Ab. Data from one of three independent experiments are shown. E, IL-2 secretion by BW or BW/NCR1-␨ cells following a 48-h incubation with several mouse tumor cell lines. Values are mean Ϯ SD for triplicate samples. Data from one of three independent experiments are shown. F, IL-2 secretion from BW or BW/NCR1-␨ cells following a 48-h incubation with target cell lines in the presence or absence of NCR1-Ig fusion protein at a concentration of 5 ␮g/sample. Values are mean Ϯ SD for triplicate samples. Data from one of two independent experiments are shown. 2224 ENHANCED TUMOR GROWTH IN THE ABSENCE OF Ncr1

FIGURE 2. Enhanced PD1.6 tumor growth in Ncr1gfp/gfp mice. Tumor size was monitored every 1–2 days after s.c. injection of 1 ϫ 106 PD1.6 cells in six heterozygous Ncr1gfp/ϩ mice (f) and six NCR1-deficient Ncr1gfp/gfp mice (u). Summary of the average tumor size is shown. Data from one of four independent experiments are shown. mouse ortholog of NKp46, plays a critical function in the in vivo eradication of influenza virus (37). However, its function in the in vivo killing of tumor cells has remained unclear (37). Since the cellular ligands for NKp46 are yet unknown, we studied whether NKp46 or its mouse ortholog (NCR1) recognizes ligand(s) ex- pressed by mouse tumor cell lines using Ig fusion proteins. For this purpose, the extracellular portion of NKp46 receptor or the extra- cellular portion of NCR1 receptor was fused to the Fc portion of human IgG1, as previously described (37, 44). As a negative con- trol, we used the truncated extracellular portion of NKp46 missing the receptor binding domain NKp46 D1-Ig (47). The binding of the various fusion proteins to various mouse tumor cells lines was tested in flow cytometry (Fig. 1, A–C). Several mouse lymphoma cell lines, including PD1.6, B10, and EL4, were recognized by NCR1-Ig fusion protein (Fig. 1A), suggesting that they express ligand(s) for NCR1. Other cell lines tested, such as BCL1, AML, FIGURE 3. Enhanced PD1.6 tumor growth in NK-depleted mice. A, 66.3, and LB, were not recognized (data not shown). Furthermore, Flow cytometry analysis of peripheral blood cells derived from mice in- all cell lines that were shown to express ligands for NCR1 were jected with anti-NK1.1 Ab (right plot) or with sterile PBS only (left plot) also stained by NKp46-Ig fusion protein (Fig. 1B). The fact that to detect GFP-positive NK cells. Percentages of GFP-positive cells are there is cross-reactivity between human and mouse NKp46 recep- indicated in the plots. B, Tumor size was monitored every 2 days following tor and that mouse tumor cell lines are recognized also by s.c. injection of 1 ϫ 106 PD1.6 cells in five NK-depleted heterozygous NKp46-Ig is not surprising because it was previously shown Ncr1gfp/ϩ mice (f) and five untreated heterozygous Ncr1gfp/ϩ mice (u). that NKp46 and NCR1 ligands share great homology (14, 37). The Summary of average tumor size is shown. NKp46-Ig and NCR1-Ig staining was specific, as no staining was observed with the control fusion protein NKp46 D1-Ig (Fig. 1C). Additionally, the tested cell lines were stained using NKG2D-Ig pression on the cell surface. As can be seen in Fig. 1D, expression fusion protein to test whether they express NKG2D ligands and, if of the chimeric protein on BW cells was verified by both anti- so, whether there is a difference in the expression level of these NCR1 and anti-HA tag Abs. In this system, the triggering of the ligands between the cell lines. The result obtained showed low receptor fused to the ␨-chain by its ligand will lead to the secretion expression of NKG2D ligands on all tested cell lines, with no of IL-2. Indeed, since BW cells express an unknown ligand for marked difference in their expression level (data not shown). NCR1 (Fig. 1, A and B), secretion of IL-2 from BW/NCR1-␨ cells was observed even without any target cells, whereas little or no NCR1 is efficiently activated by PD1.6 IL-2 self-secretion was observed from BW cells alone (Fig. 1E). We next wanted to evaluate which of the cell lines that were shown Next, we wondered which of the tumor cell lines, recognized by to express ligands for NCR1/NKp46 will efficiently activate these NCR1-Ig and NKp46-Ig (Fig. 1, A and B), will be able to induce receptors. To that end, we have used the BW reporter system. BW IL-2 secretion from BW/NCR1-␨ that could overcome the effect cells are a mouse thymoma cell line lacking the expression of the mediated by the endogenous NCR1 ligand expressed by BW cells. TCR ␣- and ␤-chains, which can be induced to secrete IL-2. The BW or BW/NCR1-␨ cells were incubated with the various mouse BW cells also express an unknown ligand for NCR1, as demon- tumor cell lines that were shown to express ligand(s) for NCR1 strated by flow cytometry analysis of BW cells stained with the (Fig. 1A) and IL-2 level was quantified in cell supernatants by NCR1-Ig, NKp46-Ig, and the negative control NKp46 D1-Ig fu- ELISA 48 h later. Ligation of the NCR1-␨ with anti-HA tag Ab sion proteins (Fig. 1, A–C). An NCR1-␨ chimeric protein, com- induced a strong IL-2 production which exceeded the basal IL-2 posed of the extracellular portion of the NCR1 receptor fused to secretion from BW/NCR1-␨ (Fig. 1E), indicating that this assay the mouse ␨-chain, was constructed and stably expressed in BW could be used to detect functional ligands for NCR1. Such a ligand cells. The NCR1-␨ construct also contains a HA-tag to detect ex- was shown to exist on PD1.6 cells as a significant 2-fold increase The Journal of Immunology 2225

FIGURE 4. NK and T cells are detected within tumors. A, Flow cytometry analysis of PD1.6 cells using PE-conjugated anti-CD3 mAb. Gray shading, background staining with secondary Ab only; dark line, specific staining. Data from one of three independent experiments are shown. B and C, Detection of immune cells within tumors. NK and T cells were detected within tumors derived from either three Ncr1gfp/ϩ (B) or three Ncr1gfp/gfp (C) mice at day 11 after injection of PD1.6 tumor cells. Data from one of two independent experiments are shown.

in the secretion level of IL-2 was observed when these cells were to appear from day 7 onward, reached a maximum average size of 65 coincubated with BW/NCR1-␨ cells (Fig. 1E), whereas incubation mm2 on day 11, then began to diminish and were not visible through- of PD1.6 cells with untransfected BW cells resulted in little or no out the end point of the experiment (day 24; Fig. 2). In marked con- IL-2 secretion (Fig. 1E). The increased production of IL-2 was due trast, in all Ncr1gfp/gfp mice, tumor growth was gradual, reaching to the interaction of the unknown ligand, expressed by PD1.6 cells, ϳ400 mm2 by day 24, which was when mice were sacrificed due to with the NCR1 receptor, as it was blocked by NCR1-Ig fusion ethical reasons (Fig. 2). These results demonstrate that PD1.6 tumor protein (Fig. 1F). Taken together, these results suggest that PD1.6 growth in vivo is NKp46/NCR1 dependent. cells were the only cells tested that express a potent functional ligand for NCR1. NK cells are responsible for initial control of tumor growth A role for NK cells has been shown not only in the direct killing Enhanced in vivo PD1.6 tumor growth in the absence of NCR1 of tumor cells during the initial phase of the innate antitumor To test the in vivo function of NK cells in general and of NKp46/ response, but also in the activation of T cells and the regulation of NCR1 in particular in tumor eradication, we used the NKp46/NCR1- the adaptive immune response that follows. Therefore, the differ- knockout Ncr1gfp/gfp mice that we have generated (37). Heterozygous ence between tumor growth pattern observed in heterozygous Ncr1gfp/ϩ or NCR1-deficient Ncr1gfp/gfp mice in the C57BL/6 back- Ncr1gfp/ϩ and NCR1-deficient Ncr1gfp/gfp mice could be due to ground were injected s.c. with 1 ϫ 106 PD1.6 cells and tumor growth differences in either T cell responses or to a direct NK cell re- was monitored every 1–2 days. Remarkably, in the absence of sponse. To further establish the role played by NK cells in the NKp46/NCR1, tumor eradication was impaired. In both the Ncr1gfp/ϩ PD1.6 killing, PD1.6 s.c. tumor growth was monitored in NK- mouse group and the Ncr1gfp/gfp mouse group, tumor growth was depleted and nondepleted mice. Heterozygous Ncr1gfp/ϩ mice visible beneath the skin by day 5 and measurable around day 7 fol- were i.v. injected with 50 ␮g of anti-NK1.1 Ab into the tail vein. lowing injection (Fig. 2). In all Ncr1gfp/ϩ mice tested, tumors started Sterile PBS was injected as a negative control. We have chosen to 2226 ENHANCED TUMOR GROWTH IN THE ABSENCE OF Ncr1 use heterozygous mice for this experiment because all of their NK cells specifically express a GFP reporter (37) that can easily be visualized by flow cytometry and, importantly, they have been shown to be normal and to behave as wild-type mice (37). To verify NK cell depletion, mice were bled 2 days after injection and cells were analyzed by flow cytometry to detect GFP-expressing NK cells. Blood from a wild-type Ncr1ϩ/ϩ mouse that does not carry a GFP reporter was used as a reference. As can be seen in Fig. 3A, ϳ1% of GFP-positive cells were detected in blood derived from untreated heterozygous mice, accounting for peripheral blood NK cells, whereas no GFP-positive cells were detected in blood derived from NK-depleted mice. Next, 1 ϫ 106 PD1.6 tumor cells were injected s.c. into NK- depleted and nondepleted mice, and tumor growth was monitored every 2 days for a period of 12 days. In agreement with the above results, in the NK-depleted mouse group, tumors were visible be- neath the skin by day 3 and were measurable by day 4, whereas tumors were visible only by day 7 in the nondepleted mouse group. In both mouse groups, tumors were gradually growing until day 10 when they started to decrease and the experiment was terminated due to loss of NK depletion (depletion was verified every 2 days during the experiment and on day 10 NK cells emerged in the depleted mice). Over the entire period of the experiment, tumors were larger in the NK-depleted mouse group (Fig. 3B). Thus, NK cells in general and NCR1 in particular play an important role in the in vivo killing of PD1.6. Detection of NK cells within tumors The above experiments show that the function of NK cells in gen- FIGURE 5. NCR1 does not play a role in the in vivo B10 and EL4 eral and NKp46/NCR1 in particular is critical between days 5 and tumor growth. Tumor size was monitored every 2 days after s.c. injection ϫ 6 11, the time window in which tumors start growing. To test of 1 10 B10 tumor cells (A) or EL4 tumor cells (B) in 6 wild type Ncr1ϩ/ϩ mice and 6 NCR1-deficient Ncr1gfp/gfp mice. Data from one of whether NK cells and T cells are present within tumors at these three independent experiments are shown. time points and to determine whether the “tumor-enhancing” effect observed in the NKp46/NCR1 knockout mice was associated with a decreased migration of NK cells or T cells, we used the gfp/gfp gfp/ϩ Ncr1 and the heterozygous Ncr1 mice. Mice were s.c. gfp/gfp gfp/ϩ injected with 1 ϫ 106 PD1.6 tumor cells and sacrificed at day 11 between NCR1-deficient Ncr1 and Ncr1 mice was ob- after injection (the same day in which tumor growth reached its served indicates that NCR1 does not play a role in the trafficking maximum in the heterozygous Ncr1ϩ/gfp mice (Fig. 2). Tumors of NK cells in tumors in vivo. Similar results were observed with were removed and lymphocytes in the tumors were quantified by influenza infection (37). flow cytometry. NK cells were visualized using the GFP they spe- cifically express and T cells were identified by staining with anti- B10 and EL4 in vivo tumor growth is not altered in the absence CD3 mAb. Importantly, PD1.6 cells are a thymoma cells line that lacks expression of CD3 on their cell surface in vitro (Fig. 4A). To of NCR1 verify that PD1.6 cells do not express CD3 even after in vivo To test whether the BW screening assay could indeed select for growth, that would complicate the detection of migrating T cells, tumors expressing functional ligands for NCR1, we also examined gfp/ϩ gfp/gfp heterozygous Ncr1 and Ncr1 mice were s.c. injected the in vivo tumor growth of other lymphoma cell lines such as B10 with 1 ϫ 106 PD1.6 tumor cells. Mice were sacrificed at day 9 and EL4 that were recognized by NCR1 and NKp46-Ig (Fig. 1, A after injection (when tumors were large enough to allow their re- and B) but were not able to efficiently activate the BW reporter moval), tumors were removed, and tumor content was analyzed by ϩ ϩ system (Fig. 1E). Wild-type Ncr1 / or NCR1-deficient flow cytometry for the binding of NCR1-Ig (to detect the PD1.6 Ncr1gfp/gfp mice were injected s.c. with 1 ϫ 106 B10 or EL4 tumor tumors) and for CD3 expression. When gating on NCR1-Ig-posi- cells and tumor growth was monitored every 1–2 days. Similar to tive cells, no CD3 expression was seen, indicating that PD1.6 do not express CD3 even after in vivo growth (data not shown). NK PD1.6, both in B10 (Fig. 5A) and in EL4 (Fig. 5B) tumor growth cells were detected within PD1.6 tumors derived from both was visible beneath the skin by day 6 and measurable by day 8 ϩ/ϩ gfp/gfp Ncr1gfp/gfp and Ncr1gfp/ϩ mice, comprising ϳ2% of tumor content after injection in both Ncr1 mice and Ncr1 mice. How- with no appreciable difference between tumors derived from either ever, in marked contrast to PD1.6 tumor growth, no significant Ncr1gfp/gfp or Ncr1gfp/ϩ mice (Fig. 4, B and C). T cells were also difference in the in vivo growth of either B10 or EL4 tumors was present in the tumors, reaching around 60% of cells in the tumors observed in the absence or presence of NCR1. In the case of B10 with no substantial differences observed between Ncr1gfp/gfp and tumor growth (Fig. 5A), tumor size peaked at day 8 after injection, Ncr1gfp/ϩ mice (Fig. 4, B and C). These results indicate that NK diminished thereafter, and was not visible by day 11 after injection cells and T cells enter tumors during growth and are both present in both groups of mice. EL4 tumors (Fig. 5B); on the other hand, by day 11 after injection of cells in equal numbers in both mouse were detectable until day 14 after injection, peaking at day 12, and strains. Moreover, the fact that no difference in NK cell percentage reached bigger sizes than B10 tumors did. The Journal of Immunology 2227

FIGURE 6. Direct killing of PD1.6 lymphoma cells is mediated by the NKp46/NCR1 receptor. A, Poly(I):.poly(C)-activated NK cells were isolated from heterozygous Ncr1ϩ/gfp and NCR1-deficient Ncr1gfp/gfp using the AutoMACS instrument and incubated for 2 h with various tumor cell lines (indicated in the x-axis) and anti-CD107a Ab. CD107a levels on GFP-marked NK cells were quantified using flow cytometry analysis. Values are mean Ϯ SD for duplicate samples. B and C, Reduced in vivo killing of PD1.6 cells in Ncrgfp/gfp mice. Vybrant Dil-labeled PD1.6 cells and Vybrant DiD-labeled HeLa cells were injected into the tail vein of three wild-type Ncr1ϩ/ϩ mice and three NCR1-deficient Ncr1gfp/gfp mice. Cells in the lungs (B) were quantified 3 h after injection using flow cytometric analysis. Data from one of three independent experiments are shown. C, PD1.6 cells were labeled as above and cells in various organs (indicated in the x-axis) were monitored either immediately after the injection or 3 h later. D, IFN-␥ secretion of poly(I):poly(C)-activated NK cells from mice of different genotypes (key) in response to a 48-h incubation with different target cells. Values are mean Ϯ SD for duplicate samples. Data from one of three independent experiments are shown.

Direct killing of PD1.6 lymphoma cells is mediated by the positive control YAC-1 lymphoma cells. A marked decrease in NKp46/NCR1 receptor CD107a expression was observed when PD1.6 tumor cells were in- gfp/gfp Activation of NK cells by target cells can result in direct cytotoxicity cubated with NCR1-deficient NK cells derived from Ncr1 mice, and/or in production of cytokines such as IFN-␥ and TNF-␣ (18). indicating that the interaction between PD1.6 cells and NK cells via Therefore, we next examined the functional involvement of NK cells NCR1 in vitro leads to direct cytotoxicity. In contrast, much less kill- in the immune-mediated control of tumor growth. To test whether ing was obtained when NK cells were incubated with B10 and EL4 NCR1 is directly involved in the direct killing of PD1.6 tumor cells, tumor cells, as percentages of CD107a-positive NK cells were almost we used an in vitro cytotoxicity assay in which NK cells from het- as low as that of the negative control and no difference was seen gfp/gfp ϩ/gfp erozygous Ncr1ϩ/gfp mice and NCR1-deficient Ncr1gfp/gfp mice were between Ncr1 - and Ncr1 -derived NK cells. Thus, NCR1 is stained for CD107a (LAMP-1), a cytoplasmic granules release marker not critically involved in the killing of B10 and EL4 and the general and a direct indicator of cytotoxicity. Mice where preactivated by ability of NK cells to kill these two targets is poor. The reduced killing poly(I):poly(C) 18 h before the experiment since the killing of targets of PD1.6 tumors was due to the absence of NCR1 and not due to a by NK cells was minimal in the absence of such activation (data not general reduction in NK cytotoxicity of the NCR1-deficient Ncr1gfp/gfp shown). NK cells were isolated from mice splenocytes using an mice because similar levels of YAC-1 killing were observed when ϩ AutoMACS instrument, incubated for 2 h with the various tumor cell either Ncr1gfp/gfp-orNcr1 /gfp-derived NK cells were used. lines, and analyzed by flow cytometry for the expression of CD107a. These combined in vivo and in vitro results indicate that the As can be seen in Fig. 6A, a high percentage of NK cells derived from killing of B10 and EL4 tumor cells is not NCR1 dependant be- heterozygous Ncr1ϩ/gfp mice were CD107a positive upon incubation cause these lymphoma cell lines were efficiently rejected by both with PD1.6 tumor cells, reaching almost the percentage observed with wild-type and NCR1-deficient mice, were less killed in vitro by 2228 ENHANCED TUMOR GROWTH IN THE ABSENCE OF Ncr1

NK cells, and were less recognized by the NCR1 using the BW compared with the RMAS cell line. Taken together, these results reporter assay. Thus, immune cells other than NK cells probably demonstrate that the interaction between NK cells and PD1.6 lym- play a more significant role in the rejection of these two cell lines. phoma cells, via NCR1, does not lead to IFN-␥ secretion rather Indeed, several works have shown that both B10 and EL4 lym- than to direct cytotoxicity. phomas are subject to T cell-mediated antitumor immune re- sponses (48, 49). To further reinforce the results obtained in vitro, we used an in Discussion vivo cytotoxicity assay, a lung tumor clearance assay with minor In recent years, the immune surveillance hypothesis of tumor im- modifications owing to the presence of GFP in the mice (37, 45). munity, once dismissed, is on the rebound. Data from induced In this assay, fluorescence-labeled tumor cells are injected i.v. into mutant mouse models lacking distinct immune cell populations or the tail vein of mice, and cells that reach the lungs get trapped in effector molecules clearly demonstrate an increased incidence of the lung’s capillary blood vessels where they are subject to im- spontaneous tumors as well as a higher susceptibility to induced mune cell-mediated attack. Taken into account the short-term du- and transplanted tumors (2, 3, 23, 24, 26, 27). ration of the experiment along with no previous exposure of the NK cells represent a highly specialized lymphoid population mice to the injected tumor cells, such an experiment probably as- characterized by a potent cytolytic activity against tumor and vi- says in vivo the involvement of innate immune cells such as NK rally infected cells (1, 7, 8, 14). A role for NK cells in the rejection cells in the killing of the injected cells. Because injection of the of tumors had been proposed shortly after their discovery as a cells might vary between different mice and because we aimed to unique lymphocyte subset (4, 50) and has been established in many compare the ability between the Ncr1gfp/gfp and wild-type Ncr1ϩ/ϩ in vitro and in few in vivo studies since then. mice to kill tumor cells in vivo, each injection was controlled using In the work presented here, we have decided to focus on the role internal control HeLa cells, which are not killed by NK cells (Fig. 6A). of the NKp46/NCR1 receptor, a main activating NK cell receptor, HeLa cells were labeled using Vybrant DiD, PD1.6 target cells were in tumor recognition in vivo. labeled using Vybrant Dil, and 2 ϫ 106 cells of each cell type were To assay for the presence of potent unknown ligands for NCR1, mixed and injected together into the tail veins of Ncr1-deficient we have devised a BW-NCR1/␨ reporter system that is based on Ncr1gfp/gfp and wild-type Ncr1ϩ/ϩ mice. Lungs were explanted 3 h the fact that a self unknown ligand for NCR1 exists on BW cells later and the amount of labeled cells was quantified by flow cytom- causing continuous secretion of IL-2 due to constant triggering of etry. We observed that PD1.6 cells were less cleared from the lungs the NCR1-␨ fusion receptor. Our hypothesis was that tumor cells in Ncr1gfp/gfp mice compared with their lung clearance in Ncr1ϩ/ϩ expressing a significant amount of ligand(s) for NCR1 will induce mice: the ratio of PD1.6:HeLa was ϳ0.5 in Ncr1ϩ/ϩ mice but was an IL-2 secretion that would overcome the self-secretion of BW- ϳ0.95 in Ncr1gfp/gfp mice, meaning that the in vivo killing of PD1.6 NCR1/␨ cells. The PD1.6 lymphoma cell line was the only cell line cells was impaired in Ncr1gfp/gfp mice compared with their killing in tested that showed such an effect in the BW-NCR1/␨ assay. Ncr1ϩ/ϩ mice (Fig. 6B). A discrepancy was observed between the BW reporter assay To test whether tumor cells undergo clearance in other organs, (Fig. 1E) and the Ig fusion-binding assay (Fig. 1, A–C). Although 2 ϫ 106 Vybrant Dil-labeled PD1.6 cells were i.v. injected into the all tumors (B10, EL4, and PD1.6) bind both NCR1-Ig and tail vein of wild-type mice. Lungs, spleens, livers, inguinal lymph NKp46-Ig to the same extent, only PD1.6 was able to induce a nodes, and kidneys were harvested either immediately after injec- strong IL-2 secretion from the BW-NCR1/␨ cells. Few explana- tion or 3 h later, and the amount of labeled cells in these organs tions may account for this phenomenon: was quantified by flow cytometry. The number of cells reaching 1. Differences in the glycosylation of a bound (BW assay) and the lymph nodes or kidney was very small (Fig. 6C). Tumor cell soluble receptor (Ig fusion protein) could account for the differ- clearance was seen in the lungs and liver and was more pro- ential binding capacity of putative ligands. If that is the case, using nounced in the lungs. Thus, the i.v. injected tumor cells rapidly Ig fusion proteins might be problematic in some situations since appear in various organs and are subjected to killing by innate cells both in vitro and in vivo assays demonstrated that indeed the kill- in these organs. Interestingly, in the spleen, there was an increase ing of PD1.6 is NCR1 dependent and the killing of EL4 and B10 in tumor cell counts along the course of the experiment (Fig. 6C). is not. These results suggests that the BW assay is superior and One possible explanation for such an increase is that the spleen more reliable than the Ig-fusion proteins. directly captures Ags from the blood and is highly populated with 2. It is also possible that PD1.6 stimulates more effectively T cells (ϳ200 ϫ 106). The amount of NK cells in the spleen is only the BW-NCR1/␨ than B10 or EL4 because of the expression of ϳ0.5%. Thus, it might be possible that the PD1.6 tumor cells in the molecules other than NKp46 ligands, e.g., adhesion molecules spleen are inaccessible for NK cells due to the huge excess of the or ligands for other receptors. However, since we show that T cells. PD1.6 tumor growth in vivo is accelerated in the absence of NK cells may also have an indirect effect on the control of tumor NCR1, whereas the in vivo growth of B10 and EL4 is similar in growth by cytokine secretion and/or regulation of the adaptive im- the knockout and wild-type mice, we concluded that the BW mune response. To test whether the interaction of NK cells with system is indeed a more reliable assay as compared with the Ig PD1.6 tumor cells can cause IFN-␥ secretion, NK cells were ex- fusion proteins to predict the in vivo and the in vitro outcome tracted from poly(I):poly(C)-activated splenocytes of both wild- of killing of cells by a particular NK receptor whose ligands are type and NKp46/NCR1 knockout mice. NK cells were incubated unknown. with PD1.6 cells or control mouse tumor cell lines. Supernatants The nature of the cellular ligands for NKp46/NCR1 remains were collected 48 h later and examined for IFN-␥ levels by ELISA. elusive, despite extensive research. We speculate that one of the The RMAS tumor cell line, which has been previously shown to reasons for that might be the inability to select for tumors which express a ligand for NCR1 (37), and the HeLa cell line, which has express sufficient amounts of ligands, maybe due to the use of been shown not to express such a ligand (Fig. 6A), were used as Ig fusion proteins, which, as we show here, might be inefficient positive and negative controls, respectively. As demonstrated in in some cases. We hope that the system described here will Fig. 6D, no significant IFN-␥ secretion above basal levels (no tar- allow the identification of the tumor ligand for NCR1 in the gets) was observed in response to incubation with PD1.6 cells, as future. The Journal of Immunology 2229

The only ligands for NKp46/NCR1 identified to date are viral properties of human NK cells directly regulate T cell activation. J. Clin. Invest. HA (44). We have previously demonstrated that the sialilation of 114: 1612–1623. 16. Moretta, L., G. Ferlazzo, C. Bottino, M. Vitale, D. Pende, M. C. Mingari, and NKp46/NCR1 is required for its interaction with HA (14, 44). A. Moretta. 2006. Effector and regulatory events during natural killer-dendritic Interestingly, the PD1.6 lymphoma cell line was generated by in- cell interactions. Immunol. Rev. 214: 219–228. trathymic inoculation of the radiation leukemia virus (51) How- 17. Hanna, J., D. Goldman-Wohl, Y. Hamani, I. Avraham, C. Greenfield, S. Natanson-Yaron, D. Prus, L. Cohen-Daniel, T. I. Arnon, I. Manaster, et al. ever, our unpublished data showed that the sialilation of NKp46/ 2006. Decidual NK cells regulate key developmental processes at the human NCR1 is not needed for the recognition of PD1.6 cells. These fetal-maternal interface. Nat. Med. 12: 1065–1074. 18. Long, E. O. 2002. Tumor cell recognition by natural killer cells. Semin. Cancer results are in agreement with our previous work which demon- Biol. 12: 57–61. strated that tumor recognition is sialic acid independent (47). 19. Kiessling, R., E. Klein, and H. Wigzell. 1975. “Natural” killer cells in the mouse: We demonstrate that NCR1 is directly involved in the in vivo I. Cytotoxic cells with specificity for mouse Moloney leukemia cells: specificity and distribution according to genotype. Eur. J. Immunol. 5: 112–117. killing of PD1.6. Although we cannot completely exclude the pos- 20. Ljunggren, H. G., and K. Karre. 1985. Host resistance directed selectively against sibility that NKp46/NCR1 is also indirectly involved in the rec- H-2-deficient lymphoma variants: analysis of the mechanism. J. Exp. Med. 162: ognition of PD1.6 cells in vivo, we think that this option is un- 1745–1759. 21. Seaman, W. E., M. Sleisenger, E. Eriksson, and G. C. Koo. 1987. Depletion of likely. Our results show that PD1.6 tumors are less cleared, natural killer cells in mice by monoclonal antibody to NK-1.1: reduction in host probably by NK cells, in vivo in the absence of NCR1. In addition, defense against malignancy without loss of cellular or humoral immunity. J. Im- we show in vitro that the killing of PD1.6 by NK cells is dependent munol. 138: 4539–4544. ␥ 22. Kim, S., K. Iizuka, H. L. Aguila, I. L. Weissman, and W. M. Yokoyama. 2000. on NCR1. In contrast, no difference in the secretion of IFN- be- In vivo activities revealed by natural killer cell-deficient mice. tween wild-type and NCR1-deficient NK cells upon incubation Proc. Natl. Acad. Sci. USA 97: 2731–2736. with the tumor cells was observed. Thus, we concluded that NCR1 23. Karre, K., G. O. Klein, R. Kiessling, G. Klein, and J. C. Roder. 1980. Low natural in vivo resistance to syngeneic leukaemias in natural killer-deficient mice. Nature is directly involved in the killing of PD1.6. 284: 624–626. We demonstrate a window of opportunities for NK cell ac- 24. Karre, K., G. O. Klein, R. Kiessling, G. Klein, and J. C. Roder. 1980. In vitro tivity. Between days 7 and 13, NK cells in general and NCR1 NK-activity and in vivo resistance to leukemia: studies of beige, beige//nude and wild-type hosts on C57BL background. Int. J. Cancer 26: 789–797. in particular are crucial for fighting tumor growth. In the ab- 25. Talmadge, J. E., K. M. Meyers, D. J. Prieur, and J. R. Starkey. 1980. Role of NK sence of NCR1, beyond this time window, tumors developed cells in tumour growth and metastasis in beige mice. Nature 284: 622–624. even in the presence of a fully competent T cell activity. Since 26. van den Broek, M. E., D. Kagi, F. Ossendorp, R. Toes, S. Vamvakas, W. K. Lutz, C. J. Melief, R. M. Zinkernagel, and H. Hengartner. 1996. Decreased tumor the NCR1 knockout mice are healthy and all T cell parameters surveillance in perforin-deficient mice. J. Exp. Med. 184: 1781–1790. tested so far are intact (37), we suggest that the main function 27. Smyth, M. J., K. Y. Thia, S. E. Street, D. MacGregor, D. I. Godfrey, and of NCR1/NKp46 is in the control of initial tumor growth; if J. A. Trapani. 2000. Perforin-mediated cytotoxicity is critical for surveillance of spontaneous lymphoma. J. Exp. Med. 192: 755–760. tumor size exceeds beyond a threshold point, it can no longer be 28. Moretta, L., and A. Moretta. 2004. Unravelling natural killer cell function: trig- controlled even by T cells. gering and inhibitory human NK receptors. EMBO J. 23: 255–259. 29. Ljunggren, H. G., and K. Karre. 1990. In search of the “missing self”: MHC molecules and NK cell recognition. Immunol. Today 11: 237–244. Disclosures 30. Lebbink, R. J., and L. Meyaard. 2007. Non-MHC ligands for inhibitory immune The authors have no financial conflict of interest. receptors: novel insights and implications for immune regulation. Mol. Immunol. 44: 2153–2164. 31. Stern, N., G. Markel, T. I. Arnon, R. Gruda, H. Wong, S. D. Gray-Owen, and References O. Mandelboim. 2005. (CEA) inhibits NK killing via 1. Burnet, F. M. 1967. Immunological aspects of malignant disease. Lancet. 1: interaction with CEA-related cell adhesion molecule 1. J. Immunol. 174: 1171–1174. 6692–6701. 2. Gatti, R. A., and R. A. Good. 1971. Occurrence of malignancy in immunodefi- 32. Markel, G., N. Lieberman, G. Katz, T. I. Arnon, M. Lotem, O. Drize, ciency diseases: a literature review. Cancer 28: 89–98. R. S. Blumberg, E. Bar-Haim, R. Mader, L. Eisenbach, and O. Mandelboim. 3. Smyth, M. J., D. I. Godfrey, and J. A. Trapani. 2001. A fresh look at tumor 2002. CD66a interactions between human melanoma and NK cells: a novel class immunosurveillance and immunotherapy. Nat. Immunol. 2: 293–299. I MHC-independent inhibitory mechanism of cytotoxicity. J. Immunol. 168: 4. Herberman, R. B., M. E. Nunn, and D. H. Lavrin. 1975. Natural cytotoxic reac- 2803–2810. tivity of mouse lymphoid cells against syngeneic acid allogeneic tumors: I. Dis- 33. Pende, D., S. Parolini, A. Pessino, S. Sivori, R. Augugliaro, L. Morelli, tribution of reactivity and specificity. Intl. J. Cancer 16: 216–229. E. Marcenaro, L. Accame, A. Malaspina, R. Biassoni, et al. 1999. Identification 5. Sivori, S., S. Parolini, E. Marcenaro, R. Castriconi, D. Pende, R. Millo, and and molecular characterization of NKp30, a novel triggering receptor involved in A. Moretta. 2000. Involvement of natural cytotoxicity receptors in human natural natural cytotoxicity mediated by human natural killer cells. J. Exp. Med. 190: killer cell-mediated lysis of neuroblastoma and glioblastoma cell lines. J. Neu- 1505–1516. roimmunol. 107: 220–225. 34. Vitale, M., C. Bottino, S. Sivori, L. Sanseverino, R. Castriconi, E. Marcenaro, 6. Weiss, L., S. Reich, O. Mandelboim, and S. Slavin. 2004. Murine B-cell leukemia R. Augugliaro, L. Moretta, and A. Moretta. 1998. NKp44, a novel triggering lymphoma (BCL1) cells as a target for NK cell-mediated immunotherapy. Bone surface molecule specifically expressed by activated natural killer cells, is in- Marrow Transplant. 33: 1137–1141. volved in non-major histocompatibility complex-restricted tumor cell lysis. 7. Lanier, L. L. 2005. NK cell recognition. Annu. Rev. Immunol. 23: 225–274. J. Exp. Med. 187: 2065–2072. 8. Biron, C. A., K. B. Nguyen, G. C. Pien, L. P. Cousens, and T. P. Salazar-Mather. 35. Pessino, A., S. Sivori, C. Bottino, A. Malaspina, L. Morelli, L. Moretta, 1999. Natural killer cells in antiviral defense: function and regulation by innate R. Biassoni, and A. Moretta. 1998. Molecular cloning of NKp46: a novel member cytokines. Annu. Rev. Immunol. 17: 189–220. of the immunoglobulin superfamily involved in triggering of natural cytotoxicity. 9. Unanue, E. R. 1997. Studies in listeriosis show the strong symbiosis between the J. Exp. Med. 188: 953–960. innate cellular system and the T-cell response. Immunol. Rev. 158: 11–25. 36. Biassoni, R., A. Pessino, C. Bottino, D. Pende, L. Moretta, and A. Moretta. 1999. 10. Laskay, T., M. Rollinghoff, and W. Solbach. 1993. Natural killer cells participate The murine homologue of the human NKp46, a triggering receptor involved in in the early defense against Leishmania major infection in mice. Eur. J. Immunol. the induction of natural cytotoxicity. Eur. J. Immunol. 29: 1014–1020. 23: 2237–2241. 37. Gazit, R., R. Gruda, M. Elboim, T. I. Arnon, G. Katz, H. Achdout, J. Hanna, 11. Warfield, K. L., J. G. Perkins, D. L. Swenson, E. M. Deal, C. M. Bosio, U. Qimron, G. Landau, E. Greenbaum, et al. 2006. Lethal influenza infection in M. J. Aman, W. M. Yokoyama, H. A. Young, and S. Bavari. 2004. Role of the absence of the natural killer cell receptor Ncr1. Nat. Immunol. 7: natural killer cells in innate protection against lethal Ebola virus infection. J. Exp. 517–523. Med. 200: 169–179. 38. De Maria, A., R. Biassoni, M. Fogli, M. Rizzi, C. Cantoni, P. Costa, R. Conte, 12. Yokoyama, W. M., S. Kim, and A. R. French. 2004. The dynamic life of natural D. Mavilio, B. Ensoli, A. Cafaro, et al. 2001. Identification, molecular cloning killer cells. Annu. Rev. Immunol. 22: 405–429. and functional characterization of NKp46 and NKp30 natural cytotoxicity recep- 13. Vankayalapati, R., B. Wizel, S. E. Weis, H. Safi, D. L. Lakey, O. Mandelboim, tors in Macaca fascicularis NK cells. Eur. J. Immunol. 31: 3546–3556. B. Samten, A. Porgador, and P. F. Barnes. 2002. The NKp46 receptor contributes 39. Falco, M., C. Cantoni, C. Bottino, A. Moretta, and R. Biassoni. 1999. Identifi- to NK cell lysis of mononuclear phagocytes infected with an intracellular bac- cation of the rat homologue of the human NKp46 triggering receptor. Immunol. terium. J. Immunol. 168: 3451–3457. Lett. 68: 411–414. 14. Arnon, T. I., G. Markel, and O. Mandelboim. 2006. Tumor and viral recognition 40. Storset, A. K., I. O. Slettedal, J. L. Williams, A. Law, and E. Dissen. 2003. by natural killer cells receptors. Semin. Cancer Biol. 16: 348–358. Natural killer cell receptors in cattle: a bovine killer cell immunoglobulin-like 15. Hanna, J., T. Gonen-Gross, J. Fitchett, T. Rowe, M. Daniels, T. I. Arnon, receptor multigene family contains members with divergent signaling motifs. R. Gazit, A. Joseph, K. W. Schjetne, A. Steinle, et al. 2004. Novel APC-like Eur. J. Immunol. 33: 980–990. 2230 ENHANCED TUMOR GROWTH IN THE ABSENCE OF Ncr1

41. Sivori, S., D. Pende, C. Bottino, E. Marcenaro, A. Pessino, R. Biassoni, in human natural killer cell-mediated cytolysis. Annu. Rev. Immunol. 19: L. Moretta, and A. Moretta. 1999. NKp46 is the major triggering receptor in- 197–223. volved in the natural cytotoxicity of fresh or cultured human NK cells: correlation 47. Arnon, T. I., H. Achdout, N. Lieberman, R. Gazit, T. Gonen-Gross, G. Katz, between surface density of NKp46 and natural cytotoxicity against autologous, A. Bar-Ilan, N. Bloushtain, M. Lev, A. Joseph, et al. 2004. The mechanisms allogeneic or xenogeneic target cells. Eur. J. Immunol. 29: 1656–1666. controlling the recognition of tumor- and virus-infected cells by NKp46. Blood 42. Costello, R. T., S. Sivori, E. Marcenaro, M. Lafage-Pochitaloff, M. J. Mozziconacci, 103: 664–672. D. Reviron, J. A. Gastaut, D. Pende, D. Olive, and A. Moretta. 2002. Defective expression and function of natural killer cell-triggering receptors in patients with 48. Li, J., L. Theofanous, S. Stickel, H. Bouton-Verville, K. E. Burgin, S. Jakubchak, acute myeloid leukemia. Blood 99: 3661–3667. T. E. Wagner, and Y. Wei. 2007. Transfer of in vitro expanded T lymphocytes 43. Arnon, T. I., M. Lev, G. Katz, Y. Chernobrov, A. Porgador, and O. Mandelboim. after activation with dendritomas prolonged survival of mice challenged with 2001. Recognition of viral hemagglutinins by NKp44 but not by NKp30. Eur. EL4 tumor cells. Int. J. Oncol. 31: 193–197. J. Immunol. 31: 2680–2689. 49. Kamisaku, H., S. Aizawa, M. Kitagawa, Y. Ikarashi, and T. Sado. 1997. Limiting 44. Mandelboim, O., N. Lieberman, M. Lev, L. Paul, T. I. Arnon, Y. Bushkin, dilution analysis of T-cell progenitors in the bone marrow of thymic lymphoma- D. M. Davis, J. L. Strominger, J. W. Yewdell, and A. Porgador. 2001. Recog- susceptible B10 and -resistant C3H mice after fractionated whole-body X-irra- nition of haemagglutinins on virus-infected cells by NKp46 activates lysis by diation. Int. J. Radiat. Biol. 72: 191–199. human NK cells. Nature 409: 1055–1060. 50. Herberman, R. B., M. E. Nunn, H. T. Holden, and D. H. Lavrin. 1975. Natural 45. Oberg, L., S. Johansson, J. Michaelsson, E. Tomasello, E. Vivier, K. Karre, and P. Hoglund. 2004. Loss or mismatch of MHC class I is sufficient to trigger NK cytotoxic reactivity of mouse lymphoid cells against syngeneic and allogeneic cell-mediated rejection of resting lymphocytes in vivo: role of KARAP/DAP12- tumors: II. Characterization of effector cells. Int. J. Cancer. 16: 230–239. dependent and -independent pathways. Eur. J. Immunol. 34: 1646–1653. 51. Ben David, Y., M. Kotler, and E. Yefenof. 1987. A highly leukemogenic radi- 46. Moretta, A., C. Bottino, M. Vitale, D. Pende, C. Cantoni, M. C. Mingari, ation leukemia virus isolate is a thymotropic, immunosuppressive retrovirus with R. Biassoni, and L. Moretta. 2001. Activating receptors and coreceptors involved a unique RNA structure. Int. J. Cancer 39: 492–497.