Tumor Immunoediting by NKp46 Moran Elboim, Roi Gazit, Chamutal Gur, Hormas Ghadially, Gili Betser-Cohen and Ofer Mandelboim This information is current as J Immunol 2010; 184:5637-5644; Prepublished online 19 of October 7, 2021. April 2010; doi: 10.4049/jimmunol.0901644 http://www.jimmunol.org/content/184/10/5637 Downloaded from References This article cites 37 articles, 13 of which you can access for free at: http://www.jimmunol.org/content/184/10/5637.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/ • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 7, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2010 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Tumor Immunoediting by NKp46

Moran Elboim,* Roi Gazit,† Chamutal Gur,* Hormas Ghadially,* Gili Betser-Cohen,* and Ofer Mandelboim*

NK cells interact with a wide variety of hazardous cells including pathogen-infected and tumor cells. NKp46 is a specific NK killer receptor that recognizes various influenza hemagglutinins and unknown tumor ligands. It was recently shown that NKp46 plays a significant role in the in vivo eradication of tumor cells; however, the role played by NKp46 in vivo with regard to tumor de- velopment is still unclear. In this study, we used the 3-methylcholanthrene (MCA)-induced fibrosarcoma model in NKp46-deficient mice to test the NKp46 recognition of carcinogen-induced tumors. We show that although the rate of MCA-induced tumor formation was similar in the presence and in the absence of NKp46, the expression of its unknown ligands was NKp46 dependent. The unknown NKp46 ligands were nearly absent in tumors that originated in wild-type mice, whereas they were detected in tumors that originated in the NKp46-deficient mice. We demonstrate that the interactions between NKp46 and its MCA tumor-derived ligands lead to the secretion of IFN-g but not to the elimination of the MCA-derived tumor cells. In addition, we show that the in vivo growth of Downloaded from MCA-derived tumor cells expressing high levels of the NKp46 ligands is NKp46 and IFN-g dependent. Thus, we present in this study a novel NKp46-mediated mechanism of tumor editing. The Journal of Immunology, 2010, 184: 5637–5644.

atural killer cells are innate lymphocytes able to quickly NCR1 receptor in the KO mice is impaired (7). We have demonstrated respond to a wide range of transformed and pathogen- that the NCR1 KO mice are sensitive to influenza infection, thus infected cells (1–3). The activity of NK cells is controlled confirming in vivo the importance of the NKp46/NCR1-HA inter- N http://www.jimmunol.org/ by a balance of signals delivered by inhibitory and activating re- actions. Although the function of NK cells against viral infections was ceptors (1, 4). The main activating receptors include NKp46, studied in detail (1, 15, 16), only limited number of studies in- NKp44, and NKp30 (collectively known as the natural cytotoxicity vestigated the in vivo function of NKp46 against tumors, mainly by receptors [NCRs]), NKp80, 2B4, and NKG2D (5, 6). In mice, using tumor cell lines (17, 18).Thus, whether NKp46 is alsoimportant NKG2D and 2B4 are present; however, the only functional member for the growth of carcinogen-induced tumors and what would be the of the NCRs is the mouse ortholog of the human NKp46, named effect of the NKp46 on tumor immunoediting was not investigated. NCR1 (5, 7). Only a few ligands were identified for the activating The immune response against developing tumors was studied in NK receptors. These ligands are either tumor derived, such as B7- detail using the MCA model (19, 20). Smyth and colleagues (20)

H6, an NKp30 ligand, stress induced, such as the ligands for define three stages of tumor immunoediting: elimination, equilib- by guest on October 7, 2021 NKG2D, or self-ligands, such as AICL for NKp80 and BAT3 for rium, and escape. They demonstrated that the major immune pop- NKp30 (6, 8–11). With regard to the NCRs, we have demonstrated ulations involved in tumor immunoediting are T cells. The in- that the viral hemagglutinin (HA) proteins of various influenza volvement of NK cells in immunoediting has so far been studied strains interact with the human NKp44 and NKp46 and with the only in relation to NKG2D and DNAM1 (19, 21, 22). It was shown mouse NCR1 and that these interactions lead to increased killing of that the MCA-induced tumors were more frequently observed influenza-infected cells (7, 12, 13). In addition, we showed that in mice treated with blocking anti-NKG2D Abs (22). In contrast, NKp30 interacts with the HCMV’s pp65 protein, leading to the another report failed to observe an increased incidence of MCA- inhibition of killing by NK cells (14). induced sarcomas in NKG2D-deficient mice (19). Moreover, be- Our group knocked out the Ncr1 gene and inserted the GFP re- cause NKG2D is also expressed by several T cell subsets (10), porter gene instead (7). Almost all NK cells in the knockout (KO) the effect of a specific NK cell receptor such as NKp46 on the de- and heterozygous mice are labeled green, and the function of the velopment of carcinogen-induced tumors and on tumor immuno- editing remains obscure. By using the MCA-induced carcinogenesis model in Ncr1-deficient mice, we demonstrate in this study a tumor *Lautenberg Center for General and Tumor Immunology, Institute for Medical immunoediting mechanism aimed at reducing the NCR1-dependent Research Israel Canada, Hebrew University-Hadassah Medical School, Jerusalem, † IFN-g secretion during tumor development. Israel; and Immune Disease Institute, Harvard Stem Cell Institute, Harvard Medical School, Boston, MA 02115 Received for publication May 27, 2009. Accepted for publication March 14, 2010. Materials and Methods This work was supported by grants from the Israeli Science Foundation, the Israeli Mice Science Foundation (Morasha), a Croatia Israel research grant, a MOST-DKFZ re- All experiments were performed using 8- to 12-wk-old C57BL/6 mice. The search grant, the European Consortium (Grant MRTN-CT-2005), the Rosetrees Trust, generation of the Ncr1gfp/gfp mice was described previously (7). IFN-g– the Israel Cancer Association (Grant 20100003), and the Association for International tm1Ts Cancer Research (all to O.M). O.M is a Crown Professor of Molecular Immunology. deficient mice (B6.129S7-Ifng /J) were obtained from The Jackson Laboratory (Bar Harbor, ME). All experiments were performed in the Address correspondence and reprint requests to Prof. Ofer Mandelboim, Lautenberg specific pathogen-free unit of the Hadassah Medical School (Ein Kerem, Center for General and Tumor Immunology, Institute for Medical Research Israel Canada, Hebrew University-Hadassah Medical School, 91120 Jerusalem, Israel. Jerusalem), according to the guidelines of the ethical committee. E-mail address: [email protected] Carcinogenesis Abbreviations used in this paper: HA, hemagglutinin; NCR, natural cytotoxicity re- ceptor; poly(I):poly(C), polyinosinic:polycytidylic acid. Mice were injected i.m. with 50, 25, or 5 mg MCA (Sigma-Aldrich, St. Louis, MO) emulsified in olive oil by heating in boiling water. Tumor Copyright Ó 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00 growth was examined weekly for a period of at least 200 d. www.jimmunol.org/cgi/doi/10.4049/jimmunol.0901644 5638 TUMOR IMMUNOEDITING BY NKp46

Generation of MCA-induced cell lines cells were labeled with Vybrant DiD. A total of 2 3 106 cells of each of the cell type were mixed in 400 ml PBS and injected into the tail vein of C57BL/6 Two weeks after tumor appearance, the tumors were removed, and part of Ncr1+/+ or Ncr1gfp/gfp mice.Lungswereexplanted3hlater.Single-cell the tumor was taken for digestion by trypsin for 5 min. Following this suspensions were obtained by using a cell strainer, and fluorescence was trypsin treatment, tumors were mashed in a culture dish, and RPMI 1640 analyzed by flow cytometry. The ratio of the tested MCA-induced cell lines medium was added. After an overnight incubation in 37˚C and 5% CO2, target cells to the internal control HeLa cells was calculated. tumor clumps were removed, and the remaining cells were used as tumor- derived cell lines that were grown in RPMI 1640 medium supplemented CD107a assays with 10% FCS and 1% sodium pyruvate, glutamine, nonessential amino acids, and penicillin-streptomycin solution. Mice were injected i.p with 200 mg poly(I):poly(C), and spleens were removed 18 h later. NK cells were isolated from splenocytes using the Fusion proteins and flow cytometry mouse NK isolation kit (Miltenyi Biotec, Auburn, CA) and an autoMACS instrument according to the manufacturer’s instruction. NK cells (5 3 105) The NCR1-Ig, NKp46-Ig, and NKG2D-Ig fusion proteins were generated in were coincubated with the indicated target cells in a ratio of 1:1 in the COS-7 cells and purified by affinity chromatography using a protein G presence of 0.1 mg of an allophycocyanin-conjugated CD107a Ab (1D4B; column, as described previously (13). The staining of cell lines was vi- Southern Biotechnology Associates, Birmingham, AL) for 2 h. Afterward, sualized using a secondary PE-conjugated goat anti-human Abs (Jackson cells were washed and analyzed by flow cytometry. To analyze the in- ImmunoResearch Laboratories, West Grove, PA). volvement of NKG2D, saturating amounts of blocking Abs to NKG2D (C7; a gift from W. Yokoyama, Department of Medicine, Washington BW reporter assay University School of Medicine, St. Louis, MO) were added to the ex- A total of 50,000 BW or BW-NKp46-z cells were coincubated at 1:1 ratio periments before the start of the coculture. with 50,000 irradiated (3000 rad) specified cells of various cell lines for Cytokine secretion assay 48 h at 37˚C and 5% CO2. Supernatants were collected and the levels of

IL-2 were quantified by using commercially available anti-mouse IL-2 Mice were injected i.p. with 200 mg poly(I):poly(C), and spleens were Downloaded from mAbs (BD Pharmingen, San Diego, CA) and standard ELISA.2332 removed 18 h later, splenocytes were extracted, and NK cells were isolated as described above. NK cells (50,000) were coincubated with the indicated Growth of MCA tumor cell lines target cells at a ratio of 1:1 for 48 h at 37˚C and 5% CO2. Supernatants All experiments were performed by using 8- to 12-wk-old C57BL/6 mice. were collected and IFN-g levels were determined using a commercially Groups of 12 Ncr1+/+, Ncr1gfp/gfp, or IFN-g–deficient mice were injected available ELISA assay (BD Pharmingen). s.c. with 5 3 105 of the various MCA-induced tumor cells, and tumor size was measured every 2–3 d after injection using a micrometer. Results http://www.jimmunol.org/ NK cell depletion NCR1 plays no significant role in the development of MCA-induced fibrosarcomas Mice were injected i.v with 200 mg anti-NK1.1 Ab (PK136) in 200 ml sterile PBS. Ab was injected into the tail vein every 7 d during the experiment. It was previously demonstrated that NCR1 is involved in the in vivo Sterile PBS was injected as a control. Depletion was verified every 3 d after killing of some tumor cell lines (7, 17, 18). To address the role played injection. by NCR1 in carcinogen-induced tumors, we used the MCA-induced In vitro cytotoxicity assays fibrocarcinoma model. NCR1 KO-Ncr1gfp/gfp and wild-type (WT)- Ncr1+/+ littermates were injected with MCA, and fibrosarcomas For the in vitro assay, 200 mg polyinosinic:polycytidylic acid [poly(I):poly

(C)] (Sigma-Aldrich) were injected i.p., and splenocytes were removed arose at the site of injection after a long latency period. Fig. 1 by guest on October 7, 2021 18 h later. Target cells were labeled with [35S]methionine and incubated demonstrates the timeline of tumor progression in KO and WT mice. for 5 h at 37˚C with effector splenocytes at various E:T ratios. Specific In MCA-injected mice (50 mg/mouse), tumors began to appear in the killing was calculated as described previously (23). Ncr1gfp/gfp mice 78 d postinjection, whereas in the Ncr1+/+ mice In vivo lung clearance assay tumors appeared ∼20 d later (Fig. 1A). When lower doses of MCA were injected (25 mg/mouse), tumors started to appear in the For the in vivo clearance assay, we used a published fluorescence labeling gfp/gfp ∼ +/+ method with a modification to avoid possible influence of intrinsic GFP Ncr1 mice at 88 d postinjection, whereas in the Ncr1 mice, present in the Ncr1gfp/gfp mice. Target cells were labeled with Vybrant Dil tumors were observed ∼10 d later (Fig. 1B). Finally, early initial cell labeling solution (Molecular Probes, Carlsbad, CA), and control HeLa tumor growth was also observed in the Ncr1gfp/gfp injected with

FIGURE 1. The absence of NCR1 has no effect on MCA tumor growth. Mice were injected with 50 mg MCA (Ncr1+/+, n = 11; Ncr1gfp/gfp, n = 12) (A), with 25 mg MCA (Ncr1+/+, n = 12; Ncr1gfp/gfp, n = 12) (B), or with 5 mg MCA (Ncr1+/+, n = 16; Ncr1gfp/gfp, n = 22) (C). Figure shows per- centage of tumor-positive mice following injection. One representative experiment is shown out of three performed. The Journal of Immunology 5639

5 mg/mouse MCA (Fig. 1C). The differences observed between proteininwhichthetransmembraneandintracellulartailofthemouse the Ncr1gfp/gfp and the Ncr1+/+ were noted only in one to two mice, CD3z-chain is fused to the extracellular portion of NKp46 protein. and in most cases, there were no significant differences in the In this system, triggering the chimeric protein by its ligand will lead tumor growth rates between the Ncr1gfp/gfp and Ncr1+/+ mice (Fig. to the secretion of IL-2 (thus, the unknown NCR1 ligands, if present, 1). Thus, we concluded that NCR1 plays no significant role in should induce IL-2 production). Because we observed that the hu- preventing the development of the MCA-induced fibrosarcomas. man NKp46-Ig recognizes the MCA tumor ligands similarly as NCR1-Ig (examples are shown in Fig. 3A and data not shown), we NKp46-mediated immunoediting used the BW cells expressing the human NKp46 protein fused to To further investigate the role played by NCR1 in MCA-induced z-chain to address the functional relevance of the NCR1 ligands. tumor cell recognition, we generated cell lines from the primary Importantly, and in agreement with the NCR1-Ig binding results gfp/gfp +/+ tumors of both Ncr1 and Ncr1 genotypes. Cell lines with (Fig. 2), only the tumors that originated in the NCR1-deficient mice +/+ gfp/gfp a WT or KO suffix were originated from Ncr1 or Ncr1 were able to stimulate IL-2 production and secretion to a level mice, respectively. Because the NCR1 ligands are unknown, we similar to that observed after cross-linking the NKp46-z chimeric studied whether NCR1 recognizes ligands on the MCA-induced receptor by a mAb (Fig. 3B). In contrast, all tumors that were iso- cell lines by using NCR1-Ig and NKp46-Ig fusion proteins. For lated from WT mice failed to induce IL-2 secretion (Fig. 3B). These this purpose, the extracellular portion of each receptor was fused observations imply on the existence of NCR1-mediaed tumor im- to the Fc of human IgG1, as described previously (13). As can be munoediting process. As expected, because of tumor heterogeneity, seen in Fig. 2A, all cell lines were recognized by NCR1-Ig fusion not all tumors that originated in the Ncr1gfp/gfp mice induced NKp46- protein. Remarkably, we noticed that all cell lines that originated mediated IL-2 secretion. Importantly, a direct correlation was ob- in the WT mice expressed low levels of NCR1 ligands, whereas served between the expression of the NCR1 ligands detected by the Downloaded from cell lines that originated in NCR1-deficient mice expressed, in NCR1-Ig fusion protein (Fig. 2A) and the stimulation of the IL-2 general, higher levels of NCR1 ligands (Fig. 2A). Several KO cell production. The KO cell lines, which demonstrated low levels of lines such as 19KO, 70KO, 79KO, and 99KO expressed lower NCR1 ligands expression (19KO, 70KO, 79KO, and 99KO; Fig. levels of NCR1 ligands compared with the other KO cell lines. 2A), did not stimulate IL-2 secretion whereas others, which express However, on average, the levels of the NCR1 ligands were higher high levels of the NCR1 ligands, did (Fig. 3B). Thus, in the presence gfp/gfp in the Ncr1 mice as compared with their WT littermates (Fig. of NCR1, the developing tumors undergo an immunoediting process http://www.jimmunol.org/ 2A). Furthermore, all cell lines expressed low levels of NKG2D aiming at reducing NKp46 ligands expression. ligands, irrespective of whether they were generated in the WT or gfp/gfp in the Ncr1gfp/gfp mice (Fig. 2B and data not shown). Increased immunogenicity of the Ncr1 -derived cell lines To further investigate whether indeed in the presence of NCR1 the BecausetheNCR1ligandsaredownregulatedwhentumorsgrowinthe emerging tumors express less NCR1 ligands, we used an additional WT mice facing NCR1 pressure, we assumed that these ligands might assay using reporter cell lines. BW cells are mouse thymoma tumor influence tumor growth. We therefore injected MCA-induced tumor cells that had been transfected with constructs encoding a chimeric celllinesthatexpresseitherhighorlowlevelsoftheNCR1ligandsinto by guest on October 7, 2021

FIGURE 2. Expression of NCR1 and NKG2D ligands on the MCA-induced tumors. A, Binding of NCR1-Ig to the MCA-induced cell lines and average of all cell lines originated in Ncr1+/+ versus all cell lines originated in Ncr1gfp/gfp. pp , 0.005; t test; the differences are statistically significant. B, Binding of NKG2D-Ig to the MCA-induced 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. 5640 TUMOR IMMUNOEDITING BY NKp46 Downloaded from

FIGURE 3. NKp46-mediated immunoediting. A, Binding of NKp46-Ig and NCR1-Ig to the MCA-induced cell lines. Gray shading, background staining http://www.jimmunol.org/ with secondary Ab only; dark lines, specific staining with the indicated fusion protein. The various tumor cell lines are indicated. Data from one of three independent experiments are shown. B, IL-2 secretion from BW-NKp46 cell lines incubated with the various targets indicated in the x-axis. Anti-NKp46 mAb (461.G1 clone) was used as positive control. pp , 0.05; t test; the differences are statistically significant.

Ncr1gfp/gfp and Ncr1+/+ mice. Tumor sizes were measured for a period MCA-induced tumor cells were killed by activated NK cells, but of 3 wk until the tumors had grown too big and the mice had to be no differences were observed in the percentages of specific killing sacrificed because of ethical reasons. among the various NK cells derived either from Ncr1gfp/gfp, from by guest on October 7, 2021 In accordance with our prediction, the tumors that originated Ncr1+/+, or from the Ncr1gfp/+ mice (data not shown). in the absence of NCR1 (such as 60KO, 20KO, and 82KO) and We next assumed that it might be that the poly(I):poly(C) acti- expressed the putative cellular NCR1 ligands, grew faster and vation of NK cells, which results in increased killing, abrogate the reached a bigger mass in Ncr1gfp/gfp mice as compared with the differences that exists between the Ncr1gfp/gfp and the Ncr1+/+ mice. Ncr1+/+ mice (Fig. 4A). In contrast, tumors that originated in the We therefore used the in vivo lung clearance assay in which NK Ncr1+/+ mice facing the NCR1-dependent immunoediting pro- cells are not manipulated. Fluorescence-labeled tumors cells were cess, such as cell lines 2WT, 3WT, and 40WT, demonstrated injected into the tail vein of the various mice. Because the injection similar growth kinetics in WT and in the NCR1-deficient mice of cells might vary between different mice and as we aimed to (Fig. 4A). Importantly, a direct correlation was observed between compare between the ability of the Ncr1gfp/gfp and Ncr1+/+ mice to the NCR1-Ig staining (Fig. 2), the BW reporter assays (Fig. 3) and kill tumor cells in vivo, each injection was controlled by reference the in vivo growth of the different tumors. No differences in tumor cells such as HeLa cells. As can be seen in Fig. 5B, no differences growth rates were observed when low levels of NCR1 ligands were observed between the Ncr1gfp/gfp and Ncr1+/+ mice, even in were present (even with the 19KO, which expresses low levels of the in vivo lung clearance killing assays. Moreover, the cell line NCR1 ligands), whereas differences were noticed when tumors 2WT, which expresses low levels of the NCR1 ligands, was killed expressed high levels of the unknown ligands (Fig. 4A). The more efficiently in vivo, as compared with line 82KO (Fig. 5B), growth rate differences were NCR1 dependent and NK cell de- further indicating that NCR1 plays no significant role in the killing pendent as when NK cells were depleted by injection of anti- of the fibrosarcoma tumors. The PD1.6 tumor cell line, which is NK1.1 Abs, the tumor growth rates of the WT and Ncr1gfp/gfp mice killed in vitro and in vivo by NK cells in an NCR1-dependent were similar (Fig. 4B). manner (17), was used as a positive control and, indeed, it exhibited increased killing by the Ncr1+/+ NK cells. Thus, we concluded that The killing of all MCA-induced tumors is NKG2D but not NCR1 NCR1 plays no role in the MCA tumor killing. dependent It was previously shown that NKG2D is involved in the killing of Next we sought to determine the mechanism responsible for the MCA-induced tumors (22). To test the role of NKG2D in the differences in the tumor growth. We first performed in vitro killing presence and in the absence of NCR1, we stained the NK cells from assays using splenocytes isolated from either Ncr1gfp/gfp or Ncr1+/+ Ncr1gfp/gfp and heterozygous-Ncr1gfp/+ mice for CD107a (LAMP-1, mice and incubated them with the MCA-induced tumor cell lines. a marker for cytoplasmic granules release), after incubating the NK Ncr1+/+ and Ncr1gfp/gfp mice were preactivated by poly(I):poly(C) cells with the various MCA-induced tumors cell lines. We used the before performing the experiments, because in the absence of such Ncr1gfp/gfp and the Ncr1+/gfp mice, because in these mice, NK cells stimulation, the basal killing of target cells by NK cells was can be visualized specifically and accurately, as a GFP reporter gene minimal (data not shown). As can be seen in Fig. 5A, all of the was introduced into the Ncr1 locus (7). The Ncr1+/gfp mice are The Journal of Immunology 5641

FIGURE 4. Increased growth of the MCA tumors originated in the NCR1- deficient mice. The indicated MCA- induced tumor cell lines were injected s.c. into Ncr1gfp/gfp and Ncr1+/+ mice (500,000 cells/mouse). Tumor size was measured during a 3-wk period. A, Tumor growth in mice injected with cell lines originated in Ncr1gfp/gfp mice, and tumor growth in mice injected with cell lines originated in Ncr1+/+ mice. B, NK cells were depleted 2 d before the injection of tumor cell lines. Upper panel, Tumor growth in mice injected with the 60KO line. Lower panel, Tumor growth in mice injected with line 2WT. Shown is one represen- tative experiment of three performed. Downloaded from

normal and behave similarly to the Ncr1+/+ mice (7, 24, 25). As Thus, NKG2D and not NCR1 is involved in the killing of the MCA- above, no significant difference in the surface expression of induced tumor by NK cells. http://www.jimmunol.org/ CD107a was observed between the NK cells derived either from the Ncr1+/gfp or from the Ncr1gfp/gfp mice (Fig. 5C). Importantly, the NCR1 ligands trigger IFN-g secretion killing of all MCA-induced cell lines, irrespective of whether they Activation of NK cells could result in direct cytotoxicity and/or in were obtained from the Ncr1gfp/gfp or Ncr1+/+ mice, was NKG2D cytokine secretion, mainly IFN-g. Because we observed that dependent as it was blocked by an anti-NKG2D mAb (Fig. 5C). NCR1 plays no significant role in the direct killing of the MCA- by guest on October 7, 2021

FIGURE 5. The killing of all MCA-induced tumor lines is NKG2D dependent. A, Target cells were radioactively labeled and then incubated with NK cells from Ncr1+/+ and Ncr1gfp/gfp mice for 5 h at an E:T. ratio of 100:1. Data from one of three independent experiments are shown. Values are mean 6 SD for triplicate samples. B, In vivo cytotoxicity assay. Vybrant Dil-labeled MCA-induced tumor cell lines (2WT mice and 82KO) or the PD1.6 cell line was injected together with a control Vybrant DiD-labeled HeLa cells into the tail vein of Ncr1+/+ and Ncr1gfp/gfp mice. Cells in the lungs were quantified 3 h after injection by using the flow cytometry. Data from one of three independent experiments are shown. Values are mean 6 SD for triplicate samples. C, Ncr1gfp/gfp and Ncr1+/gfp mice were incubated for 2 h with MCA-induced tumor cell lines and with YAC-1 or HeLa cells, in the presence or absence of a blocking Ab for NKG2D. The CD107a levels on NK cells were quantified using the flow cytometry. Data from one of three independent experiments are shown. pp , 0.05; t test; the differences are statistically significant. 5642 TUMOR IMMUNOEDITING BY NKp46

FIGURE 6. IFN-g secretion is in- duced by MCA tumor cell lines origi- nated in the absence of NCR1. IFN-g secretion of poly(I):poly(C)-activated NK cells from mice of different geno- types in response to a 48-h incubation with different target cells. Values are mean 6 SD for duplicate samples. Data from one of three independent experi- ments are shown. pp , 0.05; t test; the differences between the Ncr1gfp/gfp and Ncr1+/gfp mice, which are statistically significant.

induced tumors (Fig. 5), our next goal was to determine whether the tumor sizes were similar between the Ncr1gfp/gfp and the IFN-g– the NCR1 ligands are involved in IFN-g secretion by NK cells. deficient mice in all MCA-induced cell lines (Fig. 7). As above, NK cells were extracted from poly(I):poly(C)-activated spleno- reduced tumor growth rates were observed in the WT mice when Downloaded from cytes of both Ncr1+/gfp and Ncr1gfp/gfp mice and were incubated MCA-induced tumors expressing high levels of NCR1 ligands were with MCA-induced tumor cell lines or control tumor cells lines. used (Fig. 7). Supernatants were collected 48 h later and tested for the presence Taken together, our results demonstrate a novel NCR1-de- of IFN-g by ELISA. Significantly higher levels of IFN-g were pendent tumor immunoediting process of MCA-induced tumors, observed when NK cells derived from the Ncr1+/gfp mice were which probably evolved to diminished IFN-g secretion. incubated with target cells that originated in the Ncr1gfp/gfp mice http://www.jimmunol.org/ (Fig. 6). IFN-g secretion was not enhanced above the background Discussion level when NK cells were incubated with tumor cell lines that The MCA-induced fibrosarcoma model has been used to establish originated in the Ncr1+/+ mice or with cell lines that originated in principles of tumor immunosurveillance and immunoediting (19, 26, the Ncr1gfp/gfp mice but express low levels of the NCR1 ligands 27). It was previously reported that MCA- induced tumors were (Fig. 6). When HeLa cells, which do not express the murine more frequent in mice treated for a long period of time with blocking NKp46 ligands, were incubated with the NK cells no IFN-g se- NKG2D Abs (22). In contrast, another report failed to observe an cretion was observed. increased incidence of MCA-induced sarcomas in NKG2D-deficient The Ncr1gfp/gfp NK cells were not defective in their IFN-g pro- mice (19). The reasons for these discrepancies are not completely by guest on October 7, 2021 duction, because equal amounts of IFN-g were detected when YAC- understood; however, because NKG2D is not expressed exclusively 1 cells were incubated with NK cells derived from both genotypes. by NK cells, it is still unclear whether the absence of an NK specific Because IFN-g secretion was not detected when NCR1 ligands killer receptor will affect MCA-induced tumor growth. were absent (Fig. 6), we concluded that the MCA-induced tumors By using the MCA model in the Ncr1gfp/gfp gene-targeted mice, are killed mainly in an NKG2D-dependent manner (Fig. 5), we were able to address two important issues. The first is whether whereas the production of cytokines is mainly NCR1 dependent. NCR1 is crucial for immune surveillance of MCA-induced fi- To demonstrate in vivo that the unknown NCR1 ligands are in- brosarcomas in vivo and the second is whether there is a selective volved in the secretion of IFN-g by NK cells, we injected the various pressure that will favor the downregulation of the unknown NCR1 MCA-induced tumor cell lines into IFN-g–deficient mice ligands (i.e., immunoediting). In this paper, we report that al- (B6.129S7-Ifngtm1Ts/J). As predicted, in the IFN-g–deficient mice, though NCR1 does not play a significant role in preventing the

FIGURE 7. Tumor growth in the IFN-g–deficient mice. The indicated MCA-induced tumor cell lines were injected s.c. into Ncr1gfp/gfp, Ncr1+/+, and IFN-g–deficient mice (500,000 cells/mouse). Tumor size was mea- sured during a 3-wk period. Shown is one representative experiment of three performed. The Journal of Immunology 5643 formation of MCA-induced fibrosarcomas, the developing tumors, References facing the NCR1 pressure, downregulate its unknown ligands. 1. Arnon, T. I., G. Markel, and O. Mandelboim. 2006. Tumor and viral recognition Many different functions are assigned to NK cells (28), and it by natural killer cells receptors. Semin. Cancer Biol. 16: 348–358. 2. Lodoen, M. B., and L. L. Lanier. 2006. Natural killer cells as an initial defense was even demonstrated that NK cells possess some adoptive im- against pathogens. Curr. Opin. Immunol. 18: 391–398. munity features (29–31). In addition to their cytolytic ability, a well- 3. Newman, K. C., and E. M. Riley. 2007. Whatever turns you on: accessory-cell– established function of NK cells is cytokine secretion (32). IFN-g is dependent activation of NK cells by pathogens. Nat. Rev. Immunol. 7: 279–291. 4. Achdout, H., I. Manaster, and O. Mandelboim. 2008. Influenza virus infection a prototypic NK cell cytokine that was shown to be essential for the augments NK cell inhibition through reorganization of major histocompatibility rejection of transplanted tumor cells and for the decrease in MCA- complex class I proteins. J. Virol. 82: 8030–8037. 5. Moretta, A., C. Bottino, M. Vitale, D. Pende, C. Cantoni, M. C. Mingari, induced tumor growth (33). Importantly, we show that in the MCA- R. Biassoni, and L. Moretta. 2001. Activating receptors and coreceptors involved model, IFN-g secretion by NK cells is NCR1 dependent, because in human natural killer cell-mediated cytolysis. Annu. Rev. Immunol. 19: 197–223. cell lines that originated in Ncr1+/+ mice and express low levels of 6. Welte, S., S. Kuttruff, I. Waldhauer, and A. Steinle. 2006. Mutual activation of natural killer cells and monocytes mediated by NKp80-AICL interaction. Nat. the NCR1 ligands could not trigger the IFN-g secretion by NK cells. Immunol. 7: 1334–1342. As expected, because of tumor heterogeneity, not all cell lines that 7. Gazit, R., R. Gruda, M. Elboim, T. I. Arnon, G. Katz, H. Achdout, J. Hanna, originated in the Ncr1gfp/gfp expressed high levels of the NCR1 li- U. Qimron, G. Landau, E. Greenbaum, et al. 2006. Lethal influenza infection in the absence of the natural killer cell receptor gene Ncr1. Nat. Immunol. 7: 517–523. gands. In general, however, we observed higher levels of NCR1 li- 8. Bauer, S., V. Groh, J. Wu, A. Steinle, J. H. Phillips, L. L. Lanier, and T. Spies. gands in tumors developed in vivo in the absence of NCR1, and in all 1999. Activation of NK cells and T cells by NKG2D, a receptor for stress- inducible MICA. Science 285: 727–729. cases, a direct correlation was observed between the expression of 9. Brandt, C. S., M. Baratin, E. C. Yi, J. Kennedy, Z. Gao, B. Fox, B. Haldeman, the NCR1 ligands (as detected by NCR1-Ig staining or BW reporter C. D. Ostrander, T. Kaifu, C. Chabannon, et al. 2009. The B7 family member assays), IFN-g secretion, and the in vivo growth of tumors. Thus, in B7-H6 is a tumor cell ligand for the activating natural killer cell receptor NKp30 in humans. J. Exp. Med. 206: 1495–1503. Downloaded from the presence of NCR1, developing tumor undergo an immunoediting 10. Ljunggren, H. G. 2008. Cancer immunosurveillance: NKG2D breaks cover. process to reduce the NCR1 ligands expression, probably to affect Immunity 28: 492–494. IFN-g secreted by NK cells. 11. Pogge von Strandmann, E., V. R. Simhadri, B. von Tresckow, S. Sasse, K. S. Reiners, H. P. Hansen, A. Rothe, B. Bo¨ll, V. L. Simhadri, P. Borchmann, The unknown NCR1 ligands seem to have a dual role, either et al. 2007. Human leukocyte antigen-B-associated transcript 3 is released from inducing cytokine secretion (as shown in this paper for the MCA- tumor cells and engages the NKp30 receptor on natural killer cells. Immunity 27: induced tumors) or promoting cytolytic activity by NK cells (as 965–974. 12. Arnon, T. I., M. Lev, G. Katz, Y. Chernobrov, A. Porgador, and O. Mandelboim. http://www.jimmunol.org/ shown for some lymphomas [17]). Several reasons could account for 2001. Recognition of viral hemagglutinins by NKp44 but not by NKp30. Eur. J. this observation. It is possible that various ligands exist for NCR1, Immunol. 31: 2680–2689. 13. Mandelboim, O., N. Lieberman, M. Lev, L. Paul, T. I. Arnon, Y. Bushkin, some of which activate cytokine secretion, some are involved in NK D. M. Davis, J. L. Strominger, J. W. Yewdell, and A. Porgador. 2001. Recog- cytotoxicity, and others might promote both activities. Alternatively, nition of haemagglutinins on virus-infected cells by NKp46 activates lysis by it is possible that the same ligands are involved in NK cytotoxicity human NK cells. Nature 409: 1055–1060. 14. Arnon, T. I., H. Achdout, O. Levi, G. Markel, N. Saleh, G. Katz, R. Gazit, and cytokine production, and the outcome of the NK cell response T. Gonen-Gross, J. Hanna, E. Nahari, et al. 2005. Inhibition of the NKp30 ac- and the magnitude of the effect are determined by additional factors, tivating receptor by pp65 of human cytomegalovirus. Nat. Immunol. 6: 515–523. which might be cell specific. It has indeed been demonstrated that 15. Jonjic´, S., M. Babic´, B. Polic´, and A. Krmpotic´. 2008. Immune evasion of natural killer cells by viruses. Curr. Opin. Immunol. 20: 30–38. the activation of NKp46 could not induce cytotoxicity of freshly 16. Lanier, L. L. 2008. Evolutionary struggles between NK cells and viruses. Nat. by guest on October 7, 2021 isolated NK cells but rather requires the ligation of additional re- Rev. Immunol. 8: 259–268. 17. Halfteck, G. G., M. Elboim, C. Gur, H. Achdout, H. Ghadially, and O. Mandelboim. ceptors to exert the NK activity. However, in activated NK cells, the 2009. Enhanced in vivo growth of lymphoma tumors in the absence of the NK- engagement of NKp46 alone triggers cytotoxicity (34). The only activating receptor NKp46/NCR1. J. Immunol. 182: 2221–2230. ligands identified so far for NKp46/NCR1 are various viral he- 18. Lakshmikanth, T., S. Burke, T. H. Ali, S. Kimpfler, F. Ursini, L. Ruggeri, M. Capanni, V. Umansky, A. Paschen, A. Sucker, et al. 2009. NCRs and DNAM-1 magglutinins (HAs) (7, 13). It was shown that NK cells kill virus- mediate NK cell recognition and lysis of human and mouse melanoma cell lines infected melanoma cells such as 1106mel through the interactions in vitro and in vivo. J. Clin. Invest. 119: 1251–1263. between the NKp46 receptor and the viral HA (13). In contrast, when 19. Guerra, N., Y. X. Tan, N. T. Joncker, A. Choy, F. Gallardo, N. Xiong, S. Knoblaugh, D. Cado, N. M. Greenberg, N. R. Greenberg, and D. H. Raulet. dendritic cells where infected with the influenza virus, the in- 2008. NKG2D-deficient mice are defective in tumor surveillance in models of teraction between NKp46 and HA led to IFN-g secretion and not to spontaneous malignancy. [Published erratum appears in 2008 Immunity 28: 723.] Immunity 28: 571–580. cytotoxicity (35). Finally, we recently demonstrated that the inter- 20. Koebel, C. M., W. Vermi, J. B. Swann, N. Zerafa, S. J. Rodig, L. J. Old, actions between NKp46 and the HA protein H1 could directly lead M. J. Smyth, and R. D. Schreiber. 2007. Adaptive immunity maintains occult to cytotoxicity, whereas when NKp46 interacts with the H5 HA, cancer in an equilibrium state. Nature 450: 903–907. 21. Iguchi-Manaka, A., H. Kai, Y. Yamashita, K. Shibata, S. Tahara-Hanaoka, additional receptors were needed to enhance the killing (36). Thus, S. Honda, T. Yasui, H. Kikutani, K. Shibuya, and A. Shibuya. 2008. Accelerated even when NKp46 is triggered by the same HA ligand, the functional tumor growth in mice deficient in DNAM-1 receptor. J. Exp. Med. 205: 2959–2964. outcome of such activation might be different. 22. Smyth, M. J., J. Swann, E. Cretney, N. Zerafa, W. M. Yokoyama, and Y. Hayakawa. 2005. NKG2D function protects the host from tumor initiation. J. We show in this paper that in the MCA-induced model, the Exp. Med. 202: 583–588. NKp46 ligands are involved in IFN-g secretion. We and other 23. Mandelboim, O., H. T. Reyburn, M. Vale´s-Go´mez, L. Pazmany, M. Colonna, G. Borsellino, and J. L. Strominger. 1996. Protection from lysis by natural killer showed that NKp46 is important for melanoma and lymphoma cells of group 1 and 2 specificity is mediated by residue 80 in human histo- eradication (17, 18), and recently, we demonstrated that the compatibility leukocyte antigen C alleles and also occurs with empty major NKp46 ligands are also involved in type 1 diabetes development histocompatibility complex molecules. J. Exp. Med. 184: 913–922. 24. Beuneu, H., J. Deguine, B. Breart, O. Mandelboim, J. P. Di Santo, and P. Bousso. (37). Clearly, the identification of the unknown NKp46 ligand is of 2009. Dynamic behavior of NK cells during activation in lymph nodes. Blood special importance. We think that the MCA-induced tumor cell 114: 3227–3234. lines generated in this study, which express high and low levels of 25. Satoh-Takayama, N., C. A. Vosshenrich, S. Lesjean-Pottier, S. Sawa, M. Lochner, F. Rattis, J. J. Mention, K. Thiam, N. Cerf-Bensussan, O. Mandelboim, et al. 2008. the unknown NKp46 ligands, are not only a good example of Microbial flora drives interleukin 22 production in intestinal NKp46+ cells that tumor immunoediting but also that these cell lines might also be provide innate mucosal immune defense. Immunity 29: 958–970. 26. Sozmen, M., R. Tunca, and S. Dag Erginsoy. 2008. Cyclin A expression is as- a valuable tool for the NKp46 ligand identification. sociated with apoptosis and mitosis in murine 3-methylcholanthrene-induced fibrosarcomas. Exp. Toxicol. Pathol. 61: 41–49. 27. Swann, J. B., M. D. Vesely, A. Silva, J. Sharkey, S. Akira, R. D. Schreiber, and M. J. Smyth. 2008. Demonstration of inflammation-induced cancer and cancer Disclosures immunoediting during primary tumorigenesis. Proc. Natl. Acad. Sci. USA 105: The authors have no financial conflicts of interest. 652–656. 5644 TUMOR IMMUNOEDITING BY NKp46

28. Di Santo, J. P. 2008. Natural killer cells: diversity in search of a niche. Nat. 34. Bryceson, Y. T., M. E. March, H. G. Ljunggren, and E. O. Long. 2006. Synergy Immunol. 9: 473–475. among receptors on resting NK cells for the activation of natural cytotoxicity and 29. Cooper, M. A., J. M. Elliott, P. A. Keyel, L. Yang, J. A. Carrero, and cytokine secretion. Blood 107: 159–166. W. M. Yokoyama. 2009. Cytokine-induced memory-like natural killer cells. 35. Draghi, M., A. Pashine, B. Sanjanwala, K. Gendzekhadze, C. Cantoni, Proc. Natl. Acad. Sci. USA 106: 1915–1919. D. Cosman, A. Moretta, N. M. Valiante, and P. Parham. 2007. NKp46 and 30. O’Leary, J. G., M. Goodarzi, D. L. Drayton, and U. H. von Andrian. 2006. NKG2D recognition of infected dendritic cells is necessary for NK cell activa- T cell- and B cell-independent adaptive immunity mediated by natural killer tion in the human response to influenza infection. J. Immunol. 178: 2688–2698. cells. Nat. Immunol. 7: 507–516. 36. Achdout, H., T. Meningher, S. Hirsh, A. Glasner, Y. Bar-On, C. Gur, 31. Sun, J. C., J. N. Beilke, and L. L. Lanier. 2009. Adaptive immune features of natural killer cells. Nature 457: 557–561. A. Porgador, M. Mendelson, M. Mandelboim, and O. Mandelboim. 2010. 32. Bryceson, Y. T., and E. O. Long. 2008. Line of attack: NK cell specificity and Killing of avian and swine influenza by natural killer cells. J. Virol. 84: 3993– integration of signals. Curr. Opin. Immunol. 20: 344–352. 4001. 33. Blankenstein, T., and Z. Qin. 2003. The role of IFN-g in tumor transplantation 37. Gur, C., A. Porgador, M. Elboim, R. Gazit, S. Mizrahi, N. Stern-Ginossar, immunity and inhibition of chemical carcinogenesis. Curr. Opin. Immunol. 15: H. Achdout, H. Ghadially, Y.Dor, T. Nir, et al. 2010. The activating receptor NKp46 148–154. is essential for the development of type 1 diabetes. Nat. Immunol. 11: 121–128. Downloaded from http://www.jimmunol.org/ by guest on October 7, 2021