Immunity Expansion and Is not Required for Tumor T Cell + ...

The Journal of Immunology

Endogenous IL-21 Restricts CD8ⴙ T Cell Expansion and Is not Required for Tumor Immunity1 Henrik Søndergaard,*‡ Jonathan M. Coquet,† Adam P. Uldrich,* Nicole McLaughlin,* Dale I. Godfrey,† Pallavur V. Sivakumar,§ Kresten Skak,‡ and Mark J. Smyth2* IL-21 has antitumor activity through actions on NK cells and CD8ⴙ T cells, and is currently in clinical development for the treatment of cancer. However, no studies have addressed the role of endogenous IL-21 in tumor immunity. In this study, we have studied both primary and secondary immune responses in IL-21ⴚ/ⴚ and IL-21Rⴚ/ⴚ mice against several experimental tumors. We found intact immune surveillance toward methylcholanthrene-induced sarcomas in IL-21ⴚ/ⴚ and IL-21Rⴚ/ⴚ mice compared with wild-type mice and B16 melanomas showed equal growth kinetics and development of lung metastases. IL-21Rⴚ/ⴚ mice showed competent NK cell-mediated rejection of NKG2D ligand (Rae1␤) expressing H-2bⴚ RMAS lymphomas and sustained transition to CD8ⴙ T cell-dependent memory against H-2bⴙ RMA lymphomas. ␣-Galactosylceramide stimulation showed equal expansion and activation of NKT and NK cells and mounted a powerful antitumor response in the absence of IL-21 signaling, despite reduced expression of granzyme B in NKT, NK, and CD8ⴙ T cells. Surprisingly, host IL-21 significantly restricted the expansion of Ag-specific CD8ⴙ T cells and inhibited primary CD8ⴙ T cell immunity against OVA-expressing EG7 lymphomas, as well as the secondary expansion of memory CD8ⴙ T cells. However, host IL-21 did not alter the growth of less immunogenic MC38 colon carcinomas with dim OVA expression. Overall, our results show that endogenous IL-21/IL-21R is not required for NK, NKT, and CD8ⴙ T cell-mediated tumor immunity, but restricts Ag-specific CD8ⴙ T cell expansion and rejection of immunogenic tumors, indicating novel immunosuppressive actions of this cytokine. The Journal of Immunology, 2009, 183: 7326 –7336.

T

he novel class I cytokine IL-21 is a member of the common ␥-chain receptor family. IL-21 is primarily produced by activated CD4⫹ T cells and NKT cells (1, 2) and signals through its unique IL-21R. IL-21R is expressed by most leukocytes including B, T, NK, NKT, macrophages, and dendritic cells (DCs),3 giving IL-21 substantial effects in both humoral and cellular immune responses; these effects include costimulation of B cell proliferation, differentiation, and isotype switching, Th17 cell differentiation of CD4⫹ T cells, increased CD8⫹ T cell expansion and effector function, and activation of NK and NKT cells (3). These profound immunomodulatory effects of IL-21 regulates immune responses in a variety of diseases, including infections, autoimmunity, and cancer (3–5). The antitumor effects of IL-21 have been extensively investigated in mouse tumor models using a range of different sources of IL-21 delivery such as, IL-21-transfected tumor cell lines, IL-21*Cancer Immunology Program, Peter MacCallum Cancer Centre, East Melbourne, † Department of Microbiology and Immunology, The University of Melbourne, Parkville, Victoria, Australia; ‡Immunopharmacology, Novo Nordisk A/S, Måløv, Denmark; and §Zymogenetics, Seattle, WA 98102 Received for publication August 17, 2009. Accepted for publication October 6, 2009. 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 with 18 U.S.C. Section 1734 solely to indicate this fact. 1

This work was supported by Grant 454569 from the National Health and Medical Research Council of Australia Program (to M.J.S., and D.I.G.), by a Cancer Research Institute Postgraduate Scholarship (to J.M.C.), a Doherty Fellowship (to A.P.U.), and by National Health and Medical Research Council Research Fellowships (to M.J.S. and D.I.G.). D.I.G. and M.J.S. have received research support from Novo Nordisk A/S.

2

Address correspondence and reprint requests to Dr. Mark J. Smyth, Peter MacCallum Cancer Centre, St Andrews Place, East Melbourne, 3002, Victoria, Australia. E-mail address: [email protected]

3

Abbreviations used in this paper: DC, dendritic cell; MCA, methylcholanthrene; ␣GC, alpha-galactosylceramide; MSCV, murine stem cell virus; WT, wild type. Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.0902697

expressing plasmids and recombinant mouse IL-21. In this study, antitumor effects have been demonstrated both on established s.c. tumors, on lung and liver metastases from i.v. injected tumors, and on disseminated tumors (4). In addition to its antitumor effects as monotherapy, IL-21 shows additional efficacy when used in combination with several other therapies (5). The antitumor activity of IL-21 is mainly mediated by NK cells (6 – 8) and CD8⫹ T cells (9 –11) with requirement of IFN-␥, perforin, or both (7, 8, 12–14) depending on specific model conditions. Consistently, IL-21 has been found to sustain Ag-specific CD8⫹ T cell responses (9), augmenting both naive and memory CD8⫹ T cell expansion (11, 15) and boosting both CD8⫹ T cell and NK cell cytotoxicity in vitro (11, 16). In a few reports, B cells have also been suggested to be involved in IL-21 antitumor responses with increased production of tumor-specific IgG (17, 18). Moreover, IL-21 has been found to modulate the activity and cytokine production of NKT cells and enhance the antitumor effects mediated by NKT cell stimulation (2, 8, 19). Based on these data, IL-21 is currently in clinical trials for the treatment of metastatic melanoma and renal cell carcinoma where it has shown an acceptable safety profile with reports of responding patients along with several indications of in vivo immune activity (20 –22). The majority of studies on IL-21 in tumor immunology have used exogenous sources of IL-21. However, IL-21 also plays an important role in the pathogenesis of several autoimmune diseases and here the role of endogenous IL-21/IL-21R has been widely studied. Particularly, crossing of IL-21R-deficient (IL-21R⫺/⫺) mice onto diabetes prone NOD mice, spontaneous arthritis susceptible K/XbN mice, and BXSB-Yaa mice that develop a systemic lupus erythematosus-like syndrome all showed significantly reduced disease activity (23–25). Furthermore, collagen-induced arthritis were reduced in DBA/1 mice treated with IL-21R.Fc (26) and IL-21⫺/⫺ mice showed resistance to DSS- and TNBS-induced colitis (27). To this end, blockade of the host IL-21/IL-21R axis

The Journal of Immunology has been suggested as a potential therapeutic opportunity in several of these autoimmune disorders. To date, no studies have investigated the role of endogenous IL-21/IL-21R signaling in the control of tumor initiation, growth and metastasis. Based on the evident antitumor effect of exogenous IL-21, we wanted to explore the contribution of host signaling via the endogenous IL-21/IL-21R axis in both primary and secondary tumor immune responses. In fully backcrossed IL-21 and IL-21R gene-targeted mice, we investigated the response to several different experimental tumors controlled by different effector cells or with responsiveness to IL-21 therapy. We show in this study that endogenous IL-21 is not required for tumor immunity but reveal a novel suppressive effect of IL-21 on CD8⫹ T cell immunity.

Materials and Methods Mice C57BL/6 (B6) wild-type (WT) mice were purchased from the Walter and Eliza Hall Institute of Medical Research (Melbourne, Victoria, Australia), or bred in-house at the Peter MacCallum Cancer Centre Animal Facility (Melbourne, Victoria, Australia). B6 IL-21-deficient (IL-21⫺/⫺) mice, originally provided by Dr. P. Sivakumar (Zymogenetics, Seattle, WA), and B6 IL-21R⫺/⫺ mice, originally provided Dr. W. J. Leonard (National Heart, Blood and Lung Institute, Bethesda, MD) were backcrossed to B6 for 8 –12 generations at the Peter MacCallum Cancer Centre Animal Facility (Melbourne, Victoria, Australia). B6 TCR.J␣18⫺/⫺ mice obtained from Dr. M. Taniguchi (Chiba, Japan) were backcrossed for 12 generations to B6 at the Peter MacCallum Cancer Centre Animal Facility (Melbourne, Victoria, Australia). Mice age 6 –14 wk were used in all experiments in accord with animal ethics guidelines of the Peter MacCallum Cancer Centre.

Cell lines and culture conditions B16 (F10) melanoma cells (American Type Culture Collection (CRL6322; ATCC), RMA (H-2b⫹), RMAS (H-2b⫺), RMAS murine stem cell virus (MSCV) (empty vector transfected) and stable transfectant RMASRae1␤, and MC38 OVAdim (28) were all maintained in RPMI 1640 supplemented with 10% heat-inactivated FCS, 100 U/ml penicillin, 100 ␮g/ml streptomycin, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, 2 mM glutamax, and 15 mM HEPES buffer (all from Invitrogen). Chicken OVA transfected EL-4 T cell lymphoma cells EG7 (CRL-2113; ATCC) and a more immunogenic variant of EG7 were grown in similar medium supplemented with 0.4 ␮g/ml geneticin G418 (Invitrogen).

Reagents and Abs ␣-Galactosylceramide (␣GC), a marine sponge glycolipid that activates CD1d-restricted NKT cells (29), was provided by the Pharmaceutical Research Laboratories (Kirin Brewery), and working concentrations were prepared in PBS. Methylcholanthrene (MCA) was purchased from SigmaAldrich. For flow cytometric analysis, the following anti-mouse Abs were used: FITC-conjugated, FoxP3 (FJK-16s) and CD44 (IM7) (eBioscience); PE-conjugated CD25 (PC61.5) and TCR-␤ (H57-597) (eBioscience); PECy5.5-conjugated TCR-␤ (H57-597) (eBioscience); PE-Cy7-conjugated, CD3 (145-2C11), CD8 (53-6.7), and NK1.1 (PK136) (BD Pharmingen); allophycocyanin-conjugated TCR-␤ (H57-597), CD62L (MEL-14), CD4 (GK1.5), B220 (RA3-6B2) (eBioscience), allophycocyanin-Cy7-conjugated CD3 (145-2C11), CD8 (53-6.7) and B220 (RA3-6B2) (BD Pharmingen); and Pacific blue-conjugated CD4 (RM4-5) (eBioscience). PE-conjugated SIINFEKL-loaded MHC class I tetramer from Dr. A. Brooks (University of Melbourne, Parkville, Australia) and ␣GC-loaded CD1d tetramer produced in house by K. Kyparissoudis (University of Melbourne, Parkville, Australia), using a construct originally provided by M. Kronenberg (La Jolla Institute for Allergy and Immunology, La Jolla, CA) were used to detect OVA-specific CD8⫹ T cells and NKT cells, respectively. Furthermore, Fc receptor block (clone 2.4G2; grown in-house) was used in all experiments to prevent nonspecific binding of Abs. Cell suspensions were stained in FACS tubes or 96-well U-bottom plates for 30 min at 4°C in the dark and washed between incubations. Flow cytometric acquisition was performed on a FACSCanto-II or LSR-II (BD Biosciences). Nonviable lymphocytes were excluded on the basis of staining with 7-aminoactinomycin D (eBioscience) or hydroxystilbamidine methanesulfonate/FluoroGold (FG) (Molecular Probes). Analysis was performed using FlowJo software (Tree Star) and FACSDiva software (BD Biosciences).

7327 MCA-induced fibrosarcoma model Male B6 WT, IL-21⫺/⫺, or IL-21R⫺/⫺ mice were injected s.c. with 100 ␮l of corn oil containing MCA (100 ␮g or 25 ␮g) on the left hind flank and were monitored for the onset and progression of tumors on a weekly basis until 50 wk of age (30). Survival was plotted using a Kaplan-Meier curve in which the time when tumors exceeded 150 mm2 was used as endpoint.

Other experimental tumor models B6 WT, IL-21⫺/⫺, or IL-21R⫺/⫺ mice were inoculated s.c. in the right flank with 105 B16F10 melanoma, 5 ⫻ 106 RMAS (H-2b-negative), 5 ⫻ 106 RMA-S MSCV (empty vector transfected), 5 ⫻ 106 RMAS-Rae1␤, 5 ⫻ 105 MC38 OVAdim, or 3 ⫻ 106 immunogenic EG7 cells. Rechallenges were made more than 40 days post tumor regression with 106 RMA or 3 ⫻ 106 EG7 cells. In these experiments, tumor size was calculated as a product of two perpendicular diameters measured with a digital caliper approximately three times per week. The termination criterion was a tumor volume of 150 mm2, which was used as a surrogate survival endpoint in KaplanMeier analysis. In other experiments, B6 WT, IL-21⫺/⫺, or IL-21R⫺/⫺ mice were injected i.v. in the tail vein on day 0 with 2 ⫻ 105 B16 melanoma cells either unpulsed or pulsed for 2 days in vitro with 500 ng/ml ␣GC. On day 14 postinoculation, mice were sacrificed, lungs were harvested, and the number of surface metastasis per lung was counted using a dissecting microscope.

In vivo stimulation of NKT cells B6 WT, IL-21⫺/⫺, IL-21R⫺/⫺, or TCR.J␣18⫺/⫺ mice were injected i.p. with 2 ␮g of ␣GC prepared in PBS at 200-␮l doses. After 2, 24, 72, or 144 h postinjection, livers and spleens were harvested to assay for NK, NKT, and CD8⫹ T cell activation by flow cytometry. NKT cells were identified as TCR␤⫹␣GC/CD1dtetramer⫹ cells, NK cells as NK.1.1⫹TCR␤⫺␣GC/CD1dtetramer⫺ and CD8⫹ T cells as TCR␤⫹CD8⫹ cells. For intracellular staining, cells were cultured in GolgiStop (BD Biosciences) for 2– 4 h before being surface stained and subsequently fixed and permeabilized using the BD Cytofix/Cytoperm Plus Fixation/Permeabilization kit (BD Biosciences). In experiments in which ␣GC was used in conjunction with chicken OVA, B6 WT, IL-21⫺/⫺, or TCR.J␣18⫺/⫺ were injected with OVA (400 ␮g/mouse) and ␣GC (1 ␮g/mouse) i.v. in 200 ␮l of PBS. On days 7 and 35 postinjection, livers and spleens were harvested and assayed for CD8⫹ T cell activation by staining with SIIN FEKL-loaded MHC class I tetramers. For intracellular IFN-␥ staining, cells were restimulated with SIINFEKL peptide for 12 h in vitro in the presence of GolgiStop for the last 4 h, and subsequently stained for CD8 and intracellular IFN-␥ and analyzed by flow cytometry.

Statistics Student’s t test (two-tailed, assuming equal variance), two-tailed MannWhitney U test or Kruskal-Wallis test with Dunn’s posttest was used for statistical evaluations of differences between WT B6 mice and IL-21⫺/⫺/ IL-21R⫺/⫺ mice as indicated in each experiment. Mantel Cox Log Rank test was used to evaluate statistical differences in Kaplan-Meier analyses. Data are generally shown as individual observations or as mean ⫾ SEM unless otherwise noted, and p ⬍ 0.05 was considered statistically significant.

Results Endogenous IL-21 does not protect from MCA-induced carcinogenesis To date no studies have evaluated the effect of endogenous IL-21/ IL-21R signaling in tumor immunity or tumor immune surveillance. Initially we wanted to study the involvement of the IL-21/ IL-21R axis in tumor immune surveillance and here we used the chemical carcinogen 3-MCA, which is known for its ability to promote fibrosarcoma carcinogenesis (31) and illustrate natural antitumor immunity (32). Groups of B6 WT, IL-21⫺/⫺, and IL21R⫺/⫺ mice were inoculated s.c. with 25 or 100 ␮g of MCA and observed for fibrosarcoma development over a period of 50 wk (Fig. 1). At 100 ␮g of MCA, ⬎80% of WT mice were sacrificed over the course of the experiment, whereas ⬃50% were sacrificed at the 25-␮g dose. However, at the doses examined in this study, neither the onset nor incidence of MCA-induced fibrosarcomas were significantly different between the IL-21⫺/⫺ and IL-21R⫺/⫺

7328

HOST IL-21 IN TUMOR IMMUNITY

FIGURE 1. Endogenous IL-21 does not protect against MCA-induced carcinogenesis. Cohorts of mice were challenged with 25 or 100 ␮g of the chemical carcinogen 3⬘ MCA s.c. and were monitored for sarcoma incidence. Data depict survival of B6 WT, IL-21⫺/⫺, and IL21R⫺/⫺ mice. Survival was defined as when tumor size ⬎150 mm2 for n ⫽ 16 –21 mice per group. Results are representative of two independent experiments.

mice compared with WT, suggesting that natural immune surveillance against chemically induced fibrosarcomas does not require IL-21 signaling. IL-21/IL-21R signaling does not protect against B16F10 tumor growth or lung metastasis In previous studies of IL-21 antitumor activity, we and others have found that the B16 melanoma model was sensitive to treatments with IL-21 alone or in different combinations (6, 10, 11, 14, 33). In this study, we used the B16 melanoma model to examine whether host-derived IL-21 or IL-21R signaling might play a role in the control of B16 tumor growth. WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice were injected with B16F10 melanoma cells s.c. and monitored for tumor growth and survival (Fig. 2A) or i.v. and evaluated 14 days later for development of lung metastases (Fig. 2B). The results of the s.c. experiment showed equal growth kinetics of B16F10 in both WT, IL-21⫺/⫺, and IL-21R⫺/⫺ with similar survival measured as the individual time when tumor size exceeded 150 mm2. Furthermore, after i.v. injection of B16F10, the number of metastases observed in the lungs 14 days later was also equivalent between WT and IL-21⫺/⫺ or IL-21R⫺/⫺ mice. These results suggest that endogenous IL-21/IL-21R signaling does not protect against B16 melanoma growth or metastasis development. Primary tumor rejection by NK cells via NKG2D and transition to CD8⫹ T cell immunity does not require IL-21R signaling IL-21 has been suggested to mediate tumor rejection through NK cells and more specifically via the activating receptor NKG2D (6, 7). To specifically investigate NK cell-mediated tumor immunity and particularly NKG2D rejection mechanisms, we used the TAPdeficient T cell lymphoma cell line RMAS, which expresses very

FIGURE 2. B16 melanoma growth and lung metastases are not controlled by IL-21/IL-21R signaling. B6 WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice were injected with B16F10 melanoma cells either s.c. with 1 ⫻ 105 cells on the right flank and monitored for tumor growth and survival defined as time when tumor size ⬎150 mm2 (A) or 2 ⫻ 105 cells i.v. and harvesting of lungs 14 days later for evaluation of lung metastases (B). Tumor growth curves represent mean ⫾ SEM and Kaplan-Meier survival curves depict survival defined as time when tumor size ⬎150 mm2 for n ⫽ 9 –12 mice in A. Data show mean ⫾ SEM for n ⫽ 6 –7 mice in B.

low levels of H-2b and a variant transfected to express the NKG2D ligand retinoic acid inducible-1␤ (Rae1␤). RMAS lymphomas are well known for their NK cell sensitivity both in vitro and in vivo (34) and Rae1␤-transfected variants have been shown to enhance the NK cell-mediated rejection (35). In this study, we inoculated IL-21R⫺/⫺ and WT mice s.c. with parental RMAS cells, RMAS MSCV (RMAS cells transfected with an empty vector), and RMAS-Rae1␤ (Fig. 3A). WT and IL-21R⫺/⫺ showed equal outgrowth of the RMAS and RMAS MSCV, whereas RMAS-Rae1␤ showed equal rejection in both strains (9 of 10 tumors were rejected in WT and 9 of 11 tumors were rejected in IL-21R⫺/⫺). These results suggest competent NKG2D-dependent NK cell-mediated tumor rejection despite IL-21R deficiency. To further explore the functional properties of this NK cellmediated rejection we next addressed whether or not endogenous IL-21/IL-21R signaling plays a role in the transition from innate to adaptive immunity, which has previously been suggested (16). In this experiment, we rechallenged WT and IL-21R⫺/⫺ mice that initially had rejected RMAS-Rae1␤ over 40 days postregression with Ag-processing competent and H-2b-positive cells RMA (Fig. 3B). As a control, RMA tumors were inoculated in naive WT mice and these tumors all grew out. The RMAS-Rae1␤ immunized


7329 NKT cell activation primes NK cells and mediates antitumor response in the absence of IL-21/IL-21R function

FIGURE 3. Primary NKG2D-dependent tumor rejection by NK cells and transition to CD8⫹ T cell memory is intact in IL-21R⫺/⫺ mice. A, B6 WT and IL-21R⫺/⫺ mice were challenged with 5 ⫻ 106 RMAS, RMAS MSCV (empty vector), or RMAS-Rae1␤ s.c. on the right flank and monitored for tumor growth. B, Naive B6 WT mice and WT and IL-21R⫺/⫺ mice that rejected their primary RMAS-Rae1␤ tumor challenge were rechallenged ⬎40 days postrejection with 106 RMA cells s.c. on the contralateral flank and monitored for tumor growth. Data in A depict mean ⫾ SEM for n ⫽ 8 –11 mice per group. Data in B show individual tumor growth curves for n ⫽ 6 mice (naive WT), n ⫽ 10 mice (immune WT), and n ⫽ 9 mice (immune IL-21R⫺/⫺).

mice, however, showed potent memory responses regardless of the lack of IL-21R signaling with initial tumor growth followed by complete rejection already within 10 days post-rechallenge in both strains. These results suggest that IL-21R signaling is redundant in the transition from an initial NK cell-mediated immunization to a successful CD8⫹ T cell-mediated response.

Recently, our lab showed that NKT cells produce IL-21 following ␣GC stimulation and show activation in response to IL-21 stimulation (2). In this experiment, we used ␣GC stimulation as a way to induce IL-21 production and to investigate whether the activation and antitumor response of NKT cells could in part be mediated by IL-21. We injected WT, IL-21⫺/⫺, and IL-21R⫺/⫺ with 2 ␮g of ␣GC i.p. and quantified the expansion of NKT cells after 2, 24, 72, and 144 h in liver and spleen using flow cytometry where NKT cells were identified as TCR␤⫹␣GC/CD1dtetramer⫹ lymphocytes (Fig. 4A). Two hours after ␣GC stimulation the detection of NKT cells decreased compared with unstimulated mice and almost disappeared after 24 h due to down-regulation of their TCR. However, after 72 h a substantial expansion (4- to 7-fold in both liver and spleen) of the NKT cell population was observed followed by a contraction of the NKT cell population by 144 h (Fig. 4A). NKT cell expansion was found to be similar between WT and IL-21⫺/⫺/IL-21R⫺/⫺ mice, suggesting a normal proliferative response of NKT cells without IL-21 signaling. In addition, NKT cells were stained for CD4 and intracellular IFN-␥ expression after ␣GC stimulation (Fig. 4B). These data showed a rapid induction of, and increased proportion of, IFN-␥ expressing NKT cells after 2 and 24 h in both CD4⫹/CD4⫺ subsets, which declined again at 72 h and almost returned to baseline by 144 h. FACS plots are only shown for liver NKT cells but similar results were obtained in the spleen (data not shown). No difference was seen between WT and IL-21⫺/⫺ or IL-21R⫺/⫺ NKT cell IFN-␥ responses after ␣GC stimulation, suggesting that cytokine production from NKT cells in response to ␣GC does not require endogenous IL-21. Because ␣GC also stimulates potent IFN-␥ production by NK cells (36), NK cell activation was also examined by gating on NK cells (␣GC/CD1d tetramer⫺NK1.1⫹␤TCR⫺) and detecting their intracellular IFN-␥ expression (Fig. 4, C and D). Representative FACS plots of gated liver NK cells are shown, but similar results were found in spleen NK cells (data not shown). The results show that NK cell proportions and intracellular IFN-␥ expression was comparable between NK cells from WT, IL-21⫺/⫺ and IL-21R⫺/⫺ mice following ␣GC stimulation at all time points (Fig. 4, C and D). Thus, IL-21 did not appear to be important for NK cell IFN-␥ production following ␣GC stimulation. The granularity/cytotoxicity of NK cells (33), CD8⫹ T cells (11) and NKT cells (2) can also be augmented by IL-21. In this experiment, we analyzed the expression of intracellular granzyme B on gated NKT, NK, and CD8⫹ T cells following ␣GC administration (Fig. 4E). Intracellular granzyme B expression was most strikingly up-regulated in NK cells within 24 h of stimulation. NKT, NK, and CD8⫹ T cells from WT mice expressed slightly more intracellular granzyme B than IL-21⫺/⫺ mice at the 72 h time point. IL-21 also appeared to be involved in the maintenance of NKT cell granzyme B expression at 144 h as well as NK cell granzyme B expression at 24 h. These results suggest that endogenous IL-21 is produced after ␣GC stimulation and may play a role in ␣GC-mediated granzyme B up-regulation and possibly cytotoxicity of NKT, NK, and CD8⫹ T cells. Therefore, to explore the capacity of IL-21⫺/⫺ NKT cell-mediated tumor immune responses, we injected WT and IL21⫺/⫺ mice i.v. with B16F10 melanoma cells either unpulsed or pulsed in vitro for 2 days with 500 ng/ml ␣GC and 14 days later lungs were harvested to count metastases (Fig. 4F). This experimental protocol, based on a previously published approach (37), affected neither the in vitro growth nor viability of the tumor cells

7330


FIGURE 4. Activated NKT cells prime NK cells and mediate antitumor response in the absence of IL-21 or IL-21R. B6 WT (n ⫽ 4 mice), IL-21⫺/⫺ (n ⫽ 4 mice), and IL-21R⫺/⫺ (n ⫽ 2– 4 mice) were administered 2 ␮g of ␣GC i.p. Liver and spleen specimens were harvested 2, 24, 72, or 144 h later to assay for NKT and NK cell activation by flow cytometry. A, NKT cells were gated as TCR␤⫹␣GC/CD1dtetramer⫹ and the absolute number of NKT cells in the liver (left) and spleen (right) of mice after administration of ␣GC are shown. B, Representative profiles of CD4 vs intracellular IFN-␥ expression of gated NKT cells from the liver following ␣GC. Number represents the percentage of cells in that quadrant. C, Representative density profiles of TCR␤ vs NK1.1 expression on non-NKT cells in the liver following ␣GC administration where gated region depicts NK cells (␣GC/ CD1dtetramer⫺NK1.1⫹TCR␤⫺). The number represents the percentage of cells in that gate. D, Representative histograms of intracellular IFN-␥ expression by gated NK cells from C. WT and IL-21⫺/⫺ mice were administered 2 ␮g of ␣GC i.p. and liver and spleen specimens were harvested at the specified time points and cells were stained for TCR-␤, ␣GC/CD1dtetramer, NK1.1, CD8, and intracellular granzyme B. E, Representative histograms of n ⫽ 4 mice at each time point of granzyme B expression on gated NKT cells (␣GC/CD1d tetramer⫹TCR␤⫹), NK cells (␣GC/CD1d tetramer⫺NK1.1⫹␤TCR⫺) and CD8⫹ T cells (␣GC/CD1d tetramer⫺TCR␤⫹CD8⫹) from WT and IL-21⫺/⫺ mice after ␣GC administration. F, B6 WT and IL-21⫺/⫺ mice were injected i.v. with 2 ⫻ 105 B16F10 melanomas either unpulsed or pulsed in vitro for 2 days with 500 ng/ml ␣GC, 14 days later lungs were harvested and lung metastasis were counted. Data depict mean ⫾ SEM for n ⫽ 7– 8 mice. Representative experiment is of three separate experiments, and Kruskal-Wallis test with Dunn’s posttest was used to compare B16 ⫾ ␣GC in WT and IL-21⫺/⫺. ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001.

(data not shown). The results showed a significant reduction in the number of lung metastases in mice that had received the ␣GCpulsed B16F10 cells compared with mice receiving unpulsed cells, but this reduction was similar in WT and IL-21⫺/⫺, suggesting that IL-21 did not contribute to the NKT cell-mediated antitumor response in this model. Similar results, but with increasing number of lung metastases, were obtained when cells were pulsed with lower concentrations of ␣GC (data not shown). CD8⫹ T cell expansion, but not IFN-␥ production is enhanced in IL-21⫺/⫺ mice Stimulation of NKT cells with ␣GC have also been shown to enhance CD8⫹ T cell responses (38 – 40). To test whether IL-21 production from NKT cells might contribute to the generation of CD8⫹ T cell-based immune responses, we injected WT, IL-21⫺/⫺,

or TCR.J␣18⫺/⫺ mice with chicken OVA protein i.p. ⫾ 1 ␮g of ␣GC and detected the Ag-specific CD8⫹ T cell response after 7 or 35 days in spleens and livers via SIINFEKL-loaded MHC class I tetramers (Fig. 5A). Immunization with OVA alone did not produce a very strong Ag-specific response, whereas addition of ␣GC produced a marked increase in SIINFEKL-specific CD8⫹ T cells in both WT and IL-21⫺/⫺ on day 7. Intriguingly, CD8⫹ T cells expanded to a greater proportion in livers of IL-21⫺/⫺ mice ( p ⬍ 0.05), suggesting that IL-21 inhibited the expansion of Ag-specific CD8⫹ T cells. As a control NKT cell-deficient TCR.J␣18⫺/⫺ mice showed no increased response to OVA plus ␣GC. On day 35, the level of OVA-specific CD8⫹ T cells were reduced to the levels seen with OVA immunization alone in both strains. For a more functional readout, splenocytes from WT or IL-21⫺/⫺ mice were also re-stimulated on day 7 or 35 with SIINFEKL peptide for 12 h


7331

FIGURE 6. OVAdim expressing MC38 tumor growth is similar in WT and IL-21⫺/⫺ mice. B6 WT and IL-21⫺/⫺ mice were challenged with 5 ⫻ 105 MC38-OVAdim cells s.c. on the right flank and monitored for tumor growth. Survival was defined as time when tumors ⬎150 mm2. Tumor growth curves represent mean ⫾ SEM and Kaplan-Meier survival curves depict survival defined as time when tumor size ⬎150 mm2 for n ⫽ 12 mice.

FIGURE 5. CD8⫹ T cell expansion, but not IFN-␥ response is enhanced in IL-21⫺/⫺ mice. B6 WT, TCR.J␣18⫺/⫺, or IL-21⫺/⫺ mice were injected i.v. with either 400 ␮g of OVA alone or 400 ␮g of OVA plus 1 ␮g/mouse of ␣GC i.p. A, On days 7 or 35, spleens and livers of mice were harvested to detect the Ag-specific CD8⫹ T cell response by SIINFEKL-loaded MHC class I tetramer staining. B, On day 7 or day 35 after OVA plus ␣GC injection, splenocytes from WT or IL-21⫺/⫺ mice were restimulated with SIINFEKL peptide for 12 h in vitro, stained for CD8 and intracellular IFN-␥, and the percentages of CD8⫹IFN-␥⫹ were obtained by flow cytometry. Each data point represents individual mice and Mann-Whitney U test was used to compare WT and IL-21⫺/⫺ groups. ⴱ, p ⬍ 0.05.

in vitro, stained for CD8 and intracellular IFN-␥ expression and analyzed by flow cytometry (Fig. 5B). On day 7 a detectable amount of IFN-␥ was found (0.1%– 0.2% of lymphocytes), but with similar levels in both WT and IL-21⫺/⫺ and on day 35 the levels of IFN-␥ had dropped to very low levels. Taken together, these results suggest that an Ag-specific CD8⫹ T cell response can be mounted by NKT cell stimulation without IL-21 production, but that endogenous IL-21 has suppressive effects on Ag-specific CD8⫹ T cell expansion at least in liver. Primary antitumor response against OVAdim expressing MC38 colon carcinoma is normal in IL-21⫺/⫺ mice Seeing that both NK cell- and NKT cell-mediated tumor immune responses were largely unchanged by the lack of endogenous IL-21 and IL-21R, but that a CD8⫹ T cell response increased in IL-21⫺/⫺ mice, we next wanted to explore the functionality of CD8⫹ T cell-dependent antitumor immune responses without host IL-21 signaling. Initially, we used a MC38 colon carcinoma cell line transfected to express low levels of OVA (MC38 OVAdim). This cell line was recently established in our laboratory and found to give 100% outgrowth in WT mice, but with a period of tumor

growth inhibition (our unpublished observations). Another MC38 variant established with high expression of OVA (MC38 OVAbright) showed complete CD8⫹ T cell-dependent rejection in WT (28), suggesting that MC38 OVAdim tumors generate an insufficient immune response for tumor rejection with the potential for modulation in IL-21⫺/⫺ mice. We injected WT and IL-21⫺/⫺ mice with MC38 OVAdim cells s.c. and monitored tumor growth and survival, defined as time when tumor size ⬎150 mm2 (Fig. 6). The results showed equal tumor growth and survival in WT and IL-21⫺/⫺ mice with a period of tumor growth delay observed in both strains. These results suggest that MC38 OVAdim tumor growth is not controlled by IL-21-dependent mechanisms. IL-21⫺/⫺ and IL-21R⫺/⫺ mice show potent CD8⫹ T cell mediated antitumor response against immunogenic EG7 tumors To investigate the CD8⫹ T cell response in another and more immunogenic tumor model we used an immunogenic variant of the OVA-expressing EL-4 lymphoma cell line EG7 grown in our laboratory. EG7 lymphomas have previously been found to respond to IL-21 therapy (9). Our EG7 variant gives a minor fraction of spontaneous rejections in WT mice (our unpublished observations), so this model has an active CD8⫹ T cell driven tumor immune response that possibly could be modulated by IL-21/IL-21R deficiency. WT and IL-21⫺/⫺ mice (Fig. 7A) or IL-21R⫺/⫺ mice (Fig. 7B) were inoculated with EG7 cells s.c. and monitored for tumor growth and survival defined as the time when individual tumor size exceeded 150 mm2. The results showed, as previously observed, spontaneous complete rejection in WT mice of ⬃18% (3/17) and 33% (5/15) in the two separate experiments shown. Interestingly, IL-21⫺/⫺ and IL-21R⫺/⫺ mice both showed increased rejection rates of 53% (9/17) and 94% (15/16), respectively. In KaplanMeier survival analysis this difference was significant for both IL21⫺/⫺ and IL-21R⫺/⫺ when compared with WT ( p ⬍ 0.01). Furthermore, we detected the OVA-specific CD8⫹ T cell response in blood using SIINFEKL-loaded MHC class I tetramers on day 13 postinjection of tumor in the IL-21⫺/⫺ experiment (Fig. 7C) and

7332


FIGURE 7. IL-21⫺/⫺ and IL-21R⫺/⫺ mice show powerful immune response against immunogenic EG7 tumors. In two separate experiments, B6 WT and IL-21⫺/⫺ (A) or IL-21R⫺/⫺ (B) mice were challenged with 3 ⫻ 106 EG7 cells s.c. on the right flank and monitored for tumor growth and survival, which was defined as time when tumors ⬎150 mm2. On day 13 postinjection of tumor, blood was drawn from WT and IL-21⫺/⫺ mice and stained with SIINFEKL-loaded MHC class I tetramers (C) and similarly on day 10 postinjection tumor in WT and IL-21R⫺/⫺ mice (D). Data are individual tumor growth curves showing number of mice in which tumors have grown out or regressed out of total mice in A and B. Kaplan-Meier survival curves depict time when individual tumors exceeded 150 mm2, Mantel Cox Log Rank test was used to evaluate statistical differences in survival between WT and IL-21⫺/⫺ or IL-21R⫺/⫺. ⴱⴱ, p ⬍ 0.01. Dot plots represent the percentage of SIINFEKL-specific cells of total CD8⫹ T cells in blood from individual mice. MannWhitney U test was used to compare differences between WT and IL-21⫺/⫺ or IL-21R⫺/⫺. ⴱ, p ⬍ 0.05.

day 10 postinjection of tumor in the IL-21R⫺/⫺ experiment (Fig. 7D). In general, more than 90% of the detected OVA-specific CD8⫹ T cells had the phenotype CD62L⫺CD44⫹ (data not shown), indicating that they were activated effector-memory cells. In the IL-21⫺/⫺ experiment the results on day 13 showed a robust OVA-specific CD8⫹ T cell response with a significant increase in the percentage of OVA-specific CD8⫹ T cells in IL-21⫺/⫺ mice compared with WT mice with a mean of 9.6% vs 2.6% ( p ⬍ 0.05), respectively. Similar results were found in IL-21R⫺/⫺ mice, where OVA-specific CD8⫹ T cells were increased already on day 10 postinjection of tumor with generally lower percentages as expected at this earlier time point (mean of 1.1% in IL-21R⫺/⫺ mice vs 0.7% in WT mice, ( p ⬍ 0.05). Furthermore, we found that the level of OVA-specific CD8⫹ T cells showed significant inverse correlation with the tumor size at the time of sampling and it was also significantly predictive of complete rejection (data not shown). Together, these observations suggest a suppressive role of endogenous IL-21 and IL-21R signaling in CD8⫹ T cell-dependent immunity toward immunogenic tumors, restricting the expansion of Ag-specific CD8⫹ T cells. CD8⫹ T cell-mediated memory responses are intact in IL-21⫺/⫺ and IL-21R⫺/⫺ mice Seeing the suppressive effect of IL-21 on primary tumor immune responses we next explored whether IL-21⫺/⫺ or IL-21R⫺/⫺ mice might have altered long-term CD8⫹ T cell-based memory responses. To test this, cohorts of WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice that initially had rejected their primary tumor challenge with the immunogenic EG7 tumors were rechallenged with conventional EG7 tumor cells. In naive WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice, these tumors all grew out (data not shown). The rechallenged mice were injected between 40 days and up to 6 mo after their primary tumor rejection and monitored for tumor growth (Fig. 8A). Initial tumor growth was observed in most mice followed by complete rejection of all tumors in both WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice, within 10 –15 days post injection, suggesting that IL-21⫺/⫺ and IL-21R⫺/⫺ mice have competent CD8⫹ T cell-mediated mem-

ory responses. To evaluate the magnitude of the adaptive memory response in the absence of IL-21, blood samples were collected pre-rechallenge and on day 3 post-rechallenge to detect the level of OVA-specific CD8⫹ T cells using SIINFEKL-loaded MHC class I tetramers (Fig. 8B). Before rechallenge all mice displayed a significant level of circulating SIINFEKL-specific CD8⫹ T cells but with no difference between strains (WT: 1.4%, IL-21⫺/⫺: 1.5%, IL-21R⫺/⫺: 2.3%, mean % of total CD8⫹ T cells), suggesting that IL-21/IL-21R deficiency does not alter the circulating effectormemory CD8⫹ T cell pool. On day 3 post-rechallenge the level of OVA-specific CD8⫹ T cells generally increased in all three strains compared with pre-rechallenge (WT: 1.8%, IL-21⫺/⫺: 2.7%, IL21R⫺/⫺: 3.7%, mean % of total CD8⫹ T cells). For comparison, naive WT and IL-21⫺/⫺ or IL-21R⫺/⫺ mice challenged with EG7 showed on average ⬍0.25% OVA-specific CD8⫹ T cells, indicating that the level of de novo generated OVA-specific CD8⫹ T cells was very low on day 3 postchallenge of tumor as expected. The average absolute changes of OVA-specific CD8⫹ T cells on day 3 relative to pre-rechallenge was 0.4% in WT mice, but was increased to 1.2% in IL-21⫺/⫺ mice and 1.3% in IL-21R⫺/⫺ mice. This difference was statistically significant in IL-21⫺/⫺ mice compared with WT mice ( p ⬍ 0.05), but not for IL-21R⫺/⫺ mice. Overall, these data suggest that CD8⫹ T cell memory is intact in mice deficient for IL-21 and IL-21R, but that IL-21 also suppresses the secondary expansion of Ag-specific CD8⫹ T cells.

Discussion In this study, we present the first set of data examining both primary and secondary tumor immune responses in mice deficient of IL-21 or IL-21R. Our results show that these mice have intact tumor immune surveillance, primary and secondary tumor immunity, but, in contrast to the literature, reveal a suppressive role for endogenous IL-21 during Ag-specific CD8⫹ T cell expansion and reactivity to immunogenic tumors. Exogenous IL-21 has shown significant antitumor activity in numerous experimental mouse tumor studies, either secreted from artificial tumor cells, expressed by plasmids, or injected as recombinant protein (4). Studies of


7333

FIGURE 8. CD8⫹ T cell-mediated memory responses are intact in IL-21⫺/⫺ and IL-21R⫺/⫺ mice. A, Groups of B6 WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice that had initially rejected a primary tumor challenge with 3 ⫻ 106 EG7 (immunogenic) cells s.c. on the right flank were rechallenged ⬎40 days and up to 6 mo postrejection with 3 ⫻ 106 EG7 (conventional) s.c. on the flank contralateral to the primary challenge and monitored for tumor growth. B, Pre-rechallenge and on day 3 post-rechallenge blood was drawn and stained with SIINFEKL-loaded MHC-I tetramers. Data in A depict individual tumor growth curves, for n ⫽ 8 (WT), n ⫽ 13 (IL-21⫺/⫺), and n ⫽ 14 (IL-21R⫺/⫺) mice and are representative of three independent experiments. B, Dot plots represent percentage of SIINFEKL-specific cells of total CD8⫹ T cells in blood from individual mice pre-rechallenge and day 3 post-rechallenge and the absolute percentage of change from pre-rechallenge to day 3 post-rechallenge. Kruskal-Wallis test with Dunn’s posttest was used to compare the absolute percentage of change in WT, IL-21⫺/⫺, and IL-21R⫺/⫺ mice ⴱ, p ⬍ 0.05 compared with WT.

endogenous IL-21 have established that the cytokine and its signaling pathway play a significant role in the pathogenesis of several experimental autoimmune diseases, including colitis (27), diabetes (25), arthritis (24), and systemic lupus erythematosus (23), with experimental autoimmune encephalomyelitis being debated (41, 42). These results highlight that IL-21 is primarily a proinflammatory cytokine. However, our data now extend these results and propose that endogenous IL-21 might also have potentially immunosuppressive effects. IL-21 therapy has shown antitumor activity through effects on NK cells and CD8⫹ T cells (4), and augmented antitumor responses by NKT cells (8). Based on these data we anticipated that IL-21⫺/⫺ and IL-21R⫺/⫺ mice might have reduced immunity to tumors particularly mediated by these effector cells. We found intact tumor immune surveillance toward MCA-induced sarcomas despite the lack of functional IL-21 or IL-21R. In contrast to transplantable tumor models, the MCA-induced sarcoma model reflects a more biologically relevant carcinogenesis process, which is subjected to natural tumor immune surveillance mainly conducted by NKT cells and NK cells in a perforin- and IFN-␥-dependent manner (43, 44). So, during the course of this model, activation of NK and NKT cells must be occurring. This seems to be a suitable model to test for potential IL-21 involvement as IL-21 is reported to increase IFN-␥ production and perforin-dependent antitumor responses by NK cells as well as stimulate granularity and IFN-␥ production by NKT cells, which can also produce IL-21 upon activation (2, 33). We used the B16 melanoma model and H-2b⫺ RMAS-Rae1␤ tumors to detect potential IL-21 involvement in primary tumor

responses with different degrees of immune activity. B16 melanomas showed equal s.c. growth kinetics and development of lung metastases and RMAS-Rae1␤ tumors were equally rejected in the absence of IL-21 signaling. Recombinant IL-21 has previously shown antitumor responses in both tumor models; in B16 melanoma through actions on either NK cells or CD8⫹ T cells (6, 10, 33), and in RMAS-Rae1␤ tumors through NK cells in an NKG2Ddependent manner (7). B16 melanomas are well known for being aggressive and poorly immunogenic tumors, which might not generate a sufficient host immune response for endogenous IL-21 to play a role and the level of endogenous IL-21 production could in this study be insignificant or otherwise redundant. RMAS-Rae1␤ tumors clearly show an active immune response as seen previously (35), but in this study it might be that NKG2D stimulation alone is sufficient for tumor rejection and because this response is mainly NK cell-mediated it could also be that IL-21-producing cells are not properly engaged. However, when we rechallenged mice postrejection of RMAS-Rae1␤ tumor with H-2b⫹ RMA cells, which are controlled by T cells but not NK cells, we found powerful but equal rejection compared with WT. Although these results suggest involvement of T cells, under these conditions IL-21 signaling was not essential for the successful transition from innate NK cellmediated immunity to adaptive T cell-mediated immunity, which has previously been suggested (16). A critical factor in these experiments is whether IL-21 is produced to any significant amount. To address this proposal, we used the CD1d-restricted ligand ␣GC, which induce IL-21 production from NKT cells as recently described (2). NKT cells secondarily enhance IFN-␥ production and antitumor activity of NK cells (45),

7334 IL-21 can activate NKT cells (2), and similarly enhance IFN-␥ production and antitumor activity of NK cells (16, 33). So, it could be that IL-21 was a potential link in the direct effects of ␣GC on NKT cells and in the sequential effects on NK cells. However, we found that NKT cell expansion and IFN-␥ production in response to ␣GC stimulation were not influenced by host IL-21 nor was the concomitant IFN-␥ production from NK cells. These data suggest that IL-21 does not have a major autocrine role during ␣GC activation of NKT cells or in paracrine stimulation of NK cells. Our finding that IL-21 was required to some extent for increased granzyme B expression in NKT, NK and CD8⫹ T cells following ␣GC stimulation indicates that IL-21 is produced following ␣GC stimulation and possibly plays a role in the optimal development of effector functions in these cells. This finding is supported by several studies in both mice and humans showing that IL-21 up-regulates granzyme B in NK cells, CD8⫹ T cells and B cells (2, 21, 22, 46, 47). However, our finding that ␣GC pulsed B16 melanomas were also strongly rejected in the absence of IL-21 suggests that host IL-21 is not a critical factor in ␣GC-mediated NKT celldependent tumor rejection. In contrast, we have previously shown that the sequential activation of NKT cells and NK cells with ␣GC followed by recombinant IL-21 stimulation was a very effective treatment of experimental tumors (8). Interestingly, this study showed that the additional effects of IL-21 were seen only if the treatment was given 3 days after ␣GC administration and not if IL-21 treatment was started together with or 6 days after ␣GC administration, showing that the timing or context of IL-21 is essential for the antitumor response. So, although the IL-21 production we achieve following ␣GC stimulation might be insignificant to add to the antitumor effect, another explanation might be that IL-21 is produced at a suboptimal time point to add to the antitumor effect in the models tested. IL-21 mediates antitumor effects through stimulation of CD8⫹ T cells, and is a potent costimulant of Ag-specific CD8⫹ T cell expansion and cytotoxicity (9 –11, 48), suggesting that such responses might be impaired without functional IL-21 signaling. In contrast to what is presented in the literature with exogenous IL21, we found that endogenous IL-21 has suppressive activity on CD8⫹ T cell responses, restricting Ag-specific CD8⫹ T cell expansion and antitumor activity against immunogenic tumors. These results suggest that IL-21 is produced during T cell sensitive tumor immune responses, but unexpectedly has suppressive rather than stimulatory effects. In support of these findings, we have previously reported that IL-21⫺/⫺ and IL-21R⫺/⫺ showed exacerbated responses to experimental autoimmune encephalomyelitis (41). Our results are also supported by studies showing that IL-21 can maintain DCs in an immature state inhibiting Ag presentation and T cell activation (49, 50). Furthermore, they are supported by results showing that the antitumor activity of IL-21 is lost by treating mice around the time of tumor inoculation as compared with treatment started just 2 days posttumor inoculation (9). A recent trio of studies suggests that IL-21 signaling is essential for maintaining CD8⫹ T cell responses during chronic viral infections (51– 53). Interestingly, they also show that host IL-21 limits Ag-specific CD8⫹ T cell expansion during the acute phase of infections, whereas IL-21 is strictly needed to sustain virus-specific CD8⫹ T cells during chronic infections (52). Together with our results, these findings support our notion that the timing or context of IL-21 is critical for the outcome of an immune response; whereas endogenous IL-21 produced during the priming phase might limit expansion of Ag-specific CD8⫹ T cells and antitumor activity, IL-21 delivered after the initial priming works to sustain or boost CD8⫹ T cell activity against both viruses and tumors.

HOST IL-21 IN TUMOR IMMUNITY We mainly found increased CD8⫹ T cell expansion in the liver in response to OVA and ␣GC immunization, which is most likely due to the increased proportion of NKT cells in this organ compared with the spleen. However, we only observed an increase in OVA-specific CD8⫹ T cells on day 7 after ␣GC and OVA immunization, whereas similar levels were found on day 35 postimmunization. And, the circulating level of effector/memory CD8⫹ T cells more than 40 days postrejection of EG7 was similar between WT, IL-21⫺/⫺ and IL-21R⫺/⫺ mice. These results indicate that host IL-21 is not involved in the long-term maintenance of effector/memory CD8⫹ T cells. This suggestion seems to contrast findings with recombinant IL-21 (9), however, in this study tumors were not completely rejected as in our model, but instead showed chronic growth inhibition maintaining the presence of Ag. Furthermore, in studies of viral infections IL-21 was not needed to maintain circulating effector/memory CD8⫹ T cells after a resolved infection, only for maintaining effector CD8⫹ T cells during chronic infections (53). Taken together, these results indicate that IL-21 is not required to maintain circulating memory CD8⫹ T cells in the absence of Ag, but can sustain CD8⫹ T cell responses during Ag persistency. This is also consistent with the findings that TCR stimulation is required for IL-21 production (1, 2). In addition to its effects during primary CD8⫹ T cell-mediated tumor immunity, IL-21 also has a putative role in secondary CD8⫹ T cell memory responses (9, 11, 15, 54). However, we found that secondary CD8⫹ T cell memory responses were normal or even increased in IL-21⫺/⫺ and IL-21R⫺/⫺ mice, indicating that endogenous IL-21 might also limit secondary effector CD8⫹ T cell expansion. Consistently, IL-21 was not required for viral recall responses (53). However, in this study IL-21 did not limit secondary Ag-specific T cell expansion and although differences might have been missed due to small group sizes, more work is warranted to confirm the role of endogenous IL-21 during secondary CD8⫹ T cell expansion. At present, the mechanism behind the suppressive effect of IL-21 is unknown. Recently, it was found that TCR priming in the presence of IL-21 results in IL-10-producing immunosuppressive T cells (55). So, the production of IL-21 in response to Ag could mediate suppression either through inhibition of DCs as stated above or by induction of IL-10-producing immunosuppressive T cells. In contrast with this scenario, we did not see any suppressive effects of host IL-21 on the growth of less immunogenic MC38-OVAdim tumors. This may be due to an insufficient Ag-induced immune response or other redundant immune suppressive mechanisms by this tumor. In this study, we anticipated that endogenous IL-21 would to some degree be involved in tumor immunity. However, based on the data presented in this study, we conclude that endogenous IL-21 and IL-21R is not required for immune surveillance or tumor immunity mediated by NK, NKT, or CD8⫹ T cells, at least in the range of tumor models tested. We have not investigated the role of endogenous IL-21 in the development of B cell lymphomas or in B cell-dependent tumor rejection, in which IL-21 could have a role (18, 56). However, our findings suggest that anti-IL-21 treatment, which has been proposed as a potential treatment strategy for many autoimmune diseases (57), would not compromise tumor immune surveillance. Indeed, given the enhanced CD8⫹ T cell response, short-term blockade of host IL-21 signaling during the early priming phase followed by stimulation with recombinant IL-21 therapy might even enhance tumor immunity. We have highlighted that the actions of IL-21 are potentially very context-dependent, but it is still unknown which cells produce IL-21 during a tumor immune response, in which anatomical context, and when does this happen. Injection of recombinant IL-21 into mice results in high nanogram per milliliter serum concentrations that are maintained for

The Journal of Immunology many hours (10), and generally endogenous serum cytokine concentrations are only in the picogram range even after stimulation. Thus it is perhaps not surprising that our study highlights the great difference between studying the function of an administered recombinant cytokine and the endogenous cytokine by gene targeting. CD4⫹ T cells are the most likely producers of IL-21 in our tumor model and the need for TCR stimulation suggests that secondary lymphoid organs are the primary location for IL-21 production. This suggests that IL-21 could be produced during the early priming phase, which could be the context of its suppressive actions. However, we believe the generation of reporter mice coexpressing, e.g., GFP along with IL-21 will be an essential tool to formally answer these questions. In summary, we have demonstrated that endogenous IL-21 signaling is not required for tumor immune surveillance, NK, NKT, and CD8⫹ T cell-dependent primary tumor immunity, or for CD8⫹ T cell-dependent secondary memory responses. However, we found that endogenous IL-21 restricts CD8⫹ T cell expansion and immunity toward immunogenic tumors. These results reveal an unexpected suppressive role for IL-21 in Ag-specific immunity in contrast to its general perception as a proinflammatory cytokine in both cancer immunotherapy and autoimmunity.

Acknowledgments We thank Michelle Stirling for the breeding and maintenance of the mice in these studies and Charlene Guan and Suzanne Medwell for the genotyping of mice. We also thank Ralph Rossi for flow cytometry support and Konstantinos Kyparissoudis for providing ␣GC-loaded CD1d tetramer.

Disclosures H.S. and K.S. are employed by Novo Nordisk, and P.V.S. is employed by Zymogenetics. D.I.G. and M.J.S. have received research support from Novo Nordisk. The remaining authors have no financial conflict of interest.

References 1. Parrish-Novak, J., S. R. Dillon, A. Nelson, A. Hammond, C. Sprecher, J. A. Gross, J. Johnston, K. Madden, W. Xu, J. West, et al. 2000. Interleukin 21 and its receptor are involved in NK cell expansion and regulation of lymphocyte function. Nature 408: 57– 63. 2. Coquet, J. M., K. Kyparissoudis, D. G. Pellicci, G. Besra, S. P. Berzins, M. J. Smyth, and D. I. Godfrey. 2007. IL-21 is produced by NKT cells and modulates NKT cell activation and cytokine production. J. Immunol. 178: 2827–2834. 3. Spolski, R., and W. J. Leonard. 2008. Interleukin-21: basic biology and implications for cancer and autoimmunity. Annu. Rev. Immunol. 26: 57–79. 4. Leonard, W. J., and R. Spolski. 2005. Interleukin-21: a modulator of lymphoid proliferation, apoptosis and differentiation. Nat. Rev. Immunol. 5: 688 – 698. 5. Skak, K., M. Kragh, D. Hausman, M. J. Smyth, and P. V. Sivakumar. 2008. Interleukin 21: combination strategies for cancer therapy. Nat. Rev. Drug Discov. 7: 231–240. 6. Wang, G., M. Tschoi, R. Spolski, Y. Lou, K. Ozaki, C. Feng, G. Kim, W. J. Leonard, and P. Hwu. 2003. In vivo antitumor activity of interleukin 21 mediated by natural killer cells. Cancer Res. 63: 9016 –9022. 7. Takaki, R., Y. Hayakawa, A. Nelson, P. V. Sivakumar, S. Hughes, M. J. Smyth, and L. L. Lanier. 2005. IL-21 enhances tumor rejection through a NKG2D-dependent mechanism. J. Immunol. 175: 2167–2173. 8. Smyth, M. J., M. E. Wallace, S. L. Nutt, H. Yagita, D. I. Godfrey, and Y. Hayakawa. 2005. Sequential activation of NKT cells and NK cells provides effective innate immunotherapy of cancer. J. Exp. Med. 201: 1973–1985. 9. Moroz, A., C. Eppolito, Q. Li, J. Tao, C. H. Clegg, and P. A. Shrikant. 2004. IL-21 enhances and sustains CD8⫹ T cell responses to achieve durable tumor immunity: comparative evaluation of IL-2, IL-15, and IL-21. J. Immunol. 173: 900 –909. 10. Sondergaard, H., K. S. Frederiksen, P. Thygesen, E. D. Galsgaard, K. Skak, P. E. Kristjansen, N. Odum, and M. Kragh. 2007. Interleukin 21 therapy increases the density of tumor infiltrating CD8⫹ T cells and inhibits the growth of syngeneic tumors. Cancer Immunol. Immunother. 56: 1417–1428. 11. Zeng, R., R. Spolski, S. E. Finkelstein, S. Oh, P. E. Kovanen, C. S. Hinrichs, C. A. Pise-Masison, M. F. Radonovich, J. N. Brady, N. P. Restifo, et al. 2005. Synergy of IL-21 and IL-15 in regulating CD8⫹ T cell expansion and function. J. Exp. Med. 201: 139 –148. 12. Comes, A., O. Rosso, A. M. Orengo, E. Di Carlo, C. Sorrentino, R. Meazza, T. Piazza, B. Valzasina, P. Nanni, M. P. Colombo, and S. Ferrini. 2006. CD25⫹ regulatory T cell depletion augments immunotherapy of micrometastases by an IL-21-secreting cellular vaccine. J. Immunol. 176: 1750 –1758.

7335 13. Di Carlo, E., A. Comes, A. M. Orengo, O. Rosso, R. Meazza, P. Musiani, M. P. Colombo, and S. Ferrini. 2004. IL-21 induces tumor rejection by specific CTL and IFN-␥-dependent CXC chemokines in syngeneic mice. J. Immunol. 172: 1540 –1547. 14. Ma, H. L., M. J. Whitters, R. F. Konz, M. Senices, D. A. Young, M. J. Grusby, M. Collins, and K. Dunussi-Joannopoulos. 2003. IL-21 activates both innate and adaptive immunity to generate potent antitumor responses that require perforin but are independent of IFN-␥. J. Immunol. 171: 608 – 615. 15. Allard, E. L., M. P. Hardy, J. Leignadier, M. Marquis, J. Rooney, D. Lehoux, and N. Labrecque. 2007. Overexpression of IL-21 promotes massive CD8⫹ memory T cell accumulation. Eur. J. Immunol. 37: 3069 –3077. 16. Kasaian, M. T., M. J. Whitters, L. L. Carter, L. D. Lowe, J. M. Jussif, B. Deng, K. A. Johnson, J. S. Witek, M. Senices, R. F. Konz, et al. 2002. IL-21 limits NK cell responses and promotes antigen-specific T cell activation: a mediator of the transition from innate to adaptive immunity. Immunity 16: 559 –569. 17. Nakano, H., T. Kishida, H. Asada, M. Shin-Ya, T. Shinomiya, J. Imanishi, T. Shimada, S. Nakai, M. Takeuchi, Y. Hisa, and O. Mazda. 2006. Interleukin-21 triggers both cellular and humoral immune responses leading to therapeutic antitumor effects against head and neck squamous cell carcinoma. J. Gene Med. 8: 90 –99. 18. Daga, A., A. M. Orengo, R. M. Gangemi, D. Marubbi, M. Perera, A. Comes, S. Ferrini, and G. Corte. 2007. Glioma immunotherapy by IL-21 gene-modified cells or by recombinant IL-21 involves antibody responses. Int. J. Cancer 121: 1756 –1763. 19. Maeda, M., Y. Yanagawa, K. Iwabuchi, K. Minami, Y. Nakamaru, D. Takagi, S. Fukuda, and K. Onoe. 2007. IL-21 enhances dendritic cell ability to induce interferon-␥ production by natural killer T cells. Immunobiology 212: 537–547. 20. Davis, I. D., B. K. Skrumsager, J. Cebon, T. Nicholaou, J. W. Barlow, N. P. Moller, K. Skak, D. Lundsgaard, K. S. Frederiksen, P. Thygesen, and G. A. McArthur. 2007. An open-label, two-arm, phase I trial of recombinant human interleukin-21 in patients with metastatic melanoma. Clin. Cancer Res. 13: 3630 –3636. 21. Davis, I. D., B. Brady, R. F. Kefford, M. Millward, J. Cebon, B. K. Skrumsager, U. Mouritzen, L. T. Hansen, K. Skak, D. Lundsgaard, et al. 2009. Clinical and biological efficacy of recombinant human interleukin-21 in patients with stage IV malignant melanoma without prior treatment: a phase IIa trial. Clin. Cancer Res. 15: 2123–2129. 22. Frederiksen, K. S., D. Lundsgaard, J. A. Freeman, S. D. Hughes, T. L. Holm, B. K. Skrumsager, A. Petri, L. T. Hansen, G. A. McArthur, I. D. Davis, and K. Skak. 2008. IL-21 induces in vivo immune activation of NK cells and CD8⫹ T cells in patients with metastatic melanoma and renal cell carcinoma. Cancer Immunol. Immunother. 57: 1439 –1449. 23. Bubier, J. A., T. J. Sproule, O. Foreman, R. Spolski, D. J. Shaffer, H. C. Morse, III, W. J. Leonard, and D. C. Roopenian. 2009. A critical role for IL-21 receptor signaling in the pathogenesis of systemic lupus erythematosus in BXSB-Yaa mice. Proc. Natl. Acad. Sci. USA 106: 1518 –1523. 24. Jang, E., S. H. Cho, H. Park, D. J. Paik, J. M. Kim, and J. Youn. 2009. A positive feedback loop of IL-21 signaling provoked by homeostatic CD4⫹. J. Immunol. 182: 4649 – 4656. 25. Spolski, R., M. Kashyap, C. Robinson, Z. Yu, and W. J. Leonard. 2008. IL-21 signaling is critical for the development of type I diabetes in the NOD mouse. Proc. Natl. Acad. Sci. USA 105: 14028 –14033. 26. Young, D. A., M. Hegen, H. L. Ma, M. J. Whitters, L. M. Albert, L. Lowe, M. Senices, P. W. Wu, B. Sibley, Y. Leathurby, T. P. Brown, C. Nickerson-Nutter, J. C. Keith, Jr., and M. Collins. 2007. Blockade of the interleukin-21/interleukin-21 receptor pathway ameliorates disease in animal models of rheumatoid arthritis. Arthritis Rheum. 56: 1152–1163. 27. Fina, D., M. Sarra, M. C. Fantini, A. Rizzo, R. Caruso, F. Caprioli, C. Stolfi, I. Cardolini, M. Dottori, M. Boirivant, F. Pallone, T. T. MacDonald, and G. Monteleone. 2008. Regulation of gut inflammation and th17 cell response by interleukin-21. Gastroenterology 134: 1038 –1048. 28. Gilfillan, S., C. J. Chan, M. Cella, N. M. Haynes, A. S. Rapaport, K. S. Boles, D. M. Andrews, M. J. Smyth, and M. Colonna. 2008. DNAM-1 promotes activation of cytotoxic lymphocytes by nonprofessional antigen-presenting cells and tumors. J. Exp. Med. 205: 2965–2973. 29. Smyth, M. J., N. Y. Crowe, D. G. Pellicci, K. Kyparissoudis, J. M. Kelly, K. Takeda, H. Yagita, and D. I. Godfrey. 2002. Sequential production of interferon-␥ by NK1.1⫹ T cells and natural killer cells is essential for the antimetastatic effect of ␣-galactosylceramide. Blood 99: 1259 –1266. 30. 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 immunoediting during primary tumorigenesis. Proc. Natl. Acad. Sci. USA 105: 652– 656. 31. Crowe, N. Y., M. J. Smyth, and D. I. Godfrey. 2002. A critical role for natural killer T cells in immunosurveillance of methylcholanthrene-induced sarcomas. J. Exp. Med. 196: 119 –127. 32. Swann, J. B., and M. J. Smyth. 2007. Immune surveillance of tumors. J. Clin. Invest. 117: 1137–1146. 33. Brady, J., Y. Hayakawa, M. J. Smyth, and S. L. Nutt. 2004. IL-21 induces the functional maturation of murine NK cells. J. Immunol. 172: 2048 –2058. 34. Karre, K., H. G. Ljunggren, G. Piontek, and R. Kiessling. 1986. Selective rejection of H-2-deficient lymphoma variants suggests alternative immune defence strategy. Nature 319: 675– 678. 35. Hayakawa, Y., J. M. Kelly, J. A. Westwood, P. K. Darcy, A. Diefenbach, D. Raulet, and M. J. Smyth. 2002. Cutting edge: tumor rejection mediated by

7336

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

NKG2D receptor-ligand interaction is dependent upon perforin. J. Immunol. 169: 5377–5381. Carnaud, C., D. Lee, O. Donnars, S. H. Park, A. Beavis, Y. Koezuka, and A. Bendelac. 1999. Cutting edge: cross-talk between cells of the innate immune system: NKT cells rapidly activate NK cells. J. Immunol. 163: 4647– 4650. Shimizu, K., A. Goto, M. Fukui, M. Taniguchi, and S. Fujii. 2007. Tumor cells loaded with ␣-galactosylceramide induce innate NKT and NK cell-dependent resistance to tumor implantation in mice. J. Immunol. 178: 2853–2861. Nishimura, T., H. Kitamura, K. Iwakabe, T. Yahata, A. Ohta, M. Sato, K. Takeda, K. Okumura, K. L. Van, T. Kawano, et al. 2000. The interface between innate and acquired immunity: glycolipid antigen presentation by CD1d-expressing dendritic cells to NKT cells induces the differentiation of antigen-specific cytotoxic T lymphocytes. Int. Immunol. 12: 987–994. Hermans, I. F., J. D. Silk, U. Gileadi, M. Salio, B. Mathew, G. Ritter, R. Schmidt, A. L. Harris, L. Old, and V. Cerundolo. 2003. NKT cells enhance CD4⫹ and CD8⫹ T cell responses to soluble antigen in vivo through direct interaction with dendritic cells. J. Immunol. 171: 5140 –5147. Stober, D., I. Jomantaite, R. Schirmbeck, and J. Reimann. 2003. NKT cells provide help for dendritic cell-dependent priming of MHC class I-restricted CD8⫹ T cells in vivo. J. Immunol. 170: 2540 –2548. Coquet, J. M., S. Chakravarti, M. J. Smyth, and D. I. Godfrey. 2008. Cutting edge: IL-21 is not essential for Th17 differentiation or experimental autoimmune encephalomyelitis. J. Immunol. 180: 7097–7101. Nurieva, R., X. O. Yang, G. Martinez, Y. Zhang, A. D. Panopoulos, L. Ma, K. Schluns, Q. Tian, S. S. Watowich, A. M. Jetten, and C. Dong. 2007. Essential autocrine regulation by IL-21 in the generation of inflammatory T cells. Nature 448: 480 – 483. Street, S. E., E. Cretney, and M. J. Smyth. 2001. Perforin and interferon-␥ activities independently control tumor initiation, growth, and metastasis. Blood 97: 192–197. Smyth, M. J., N. Y. Crowe, and D. I. Godfrey. 2001. NK cells and NKT cells collaborate in host protection from methylcholanthrene-induced fibrosarcoma. Int. Immunol. 13: 459 – 463. Hayakawa, Y., K. Takeda, H. Yagita, S. Kakuta, Y. Iwakura, K. L. Van, I. Saiki, and K. Okumura. 2001. Critical contribution of IFN-␥ and NK cells, but not


46. 47.

48.

49.

50. 51. 52. 53.

54.

55.

56.

57.

perforin-mediated cytotoxicity, to anti-metastatic effect of ␣-galactosylceramide. Eur. J. Immunol. 31: 1720 –1727. Casey, K. A., and M. F. Mescher. 2007. IL-21 promotes differentiation of naive CD8 T cells to a unique effector phenotype. J. Immunol. 178: 7640 –7648. Skak, K., K. S. Frederiksen, and D. Lundsgaard. 2008. Interleukin-21 activates human natural killer cells and modulates their surface receptor expression. Immunology 123: 575–583. Liu, S., G. Lizee, Y. Lou, C. Liu, W. W. Overwijk, G. Wang, and P. Hwu. 2007. IL-21 synergizes with IL-7 to augment expansion and anti-tumor function of cytotoxic T cells. Int. Immunol. 19: 1213–1221. Brandt, K., S. Bulfone-Paus, A. Jenckel, D. C. Foster, R. Paus, and R. Ruckert. 2003. Interleukin-21 inhibits dendritic cell-mediated T cell activation and induction of contact hypersensitivity in vivo. J. Invest. Dermatol. 121: 1379 –1382. Brandt, K., S. Bulfone-Paus, D. C. Foster, and R. Ruckert. 2003. Interleukin-21 inhibits dendritic cell activation and maturation. Blood 102: 4090 – 4098. Yi, J. S., M. Du, and A. J. Zajac. 2009. A vital role for interleukin-21 in the control of a chronic viral infection. Science 324: 1572–1576. Elsaesser, H., K. Sauer, and D. G. Brooks. 2009. IL-21 is required to control chronic viral infection. Science 324: 1569 –1572. Frohlich, A., J. Kisielow, I. Schmitz, S. Freigang, A. T. Shamshiev, J. Weber, B. J. Marsland, A. Oxenius, and M. Kopf. 2009. IL-21R on T cells is critical for sustained functionality and control of chronic viral infection. Science 324: 1576 –1580. Rodrigues, L., S. Nandakumar, C. Bonorino, B. T. Rouse, and U. Kumaraguru. 2009. IL-21 and IL-15 cytokine DNA augments HSV specific effector and memory CD8⫹ T cell response. Mol. Immunol. 46: 1494 –1504. Spolski, R., H. P. Kim, W. Zhu, D. E. Levy, and W. J. Leonard. 2009. IL-21 mediates suppressive effects via its induction of IL-10. J. Immunol. 182: 2859 –2867. Gelebart, P., Z. Zak, M. Anand, J. en-Bard, H. M. Amin, and R. Lai. 2009. Interleukin-21 effectively induces apoptosis in mantle cell lymphoma through a STAT1-dependent mechanism. Leukemia 23: 1836 –1846. Ettinger, R., S. Kuchen, and P. E. Lipsky. 2008. Interleukin 21 as a target of intervention in autoimmune disease. Ann. Rheum. Dis. 67 (Suppl. 3): iii83–iii86.