Identification of an Indispensable Role for Tyrosine Kinase 2 in CTL-Mediated Tumor Surveillance

Research Article

Olivia Simma,1 Eva Zebedin,1 Nina Neugebauer,1 Carola Schellack,1 Andreas Pilz,3 Souyet Chang-Rodriguez,4 Karen Lingnau,5 Eva Weisz,1,6 Eva Maria Putz,1 Winfried F. Pickl,2 Thomas Felzmann,4 Mathias Mu¨ller,7,8 Thomas Decker,3 Veronika Sexl,1 and Dagmar Stoiber1,6

1Institute of Pharmacology and 2Institute of Immunology, Medical University of Vienna; 3Max F. PerutzLaboratories, Department of Microbiology and Immunobiology, University of Vienna; 4Children’s Cancer Research Institute, St. Anna Kinderkrebsforschung; 5Intercell AG, Campus Vienna Biocenter; 6Ludwig Boltzmann Institute for Cancer Research; 7Institute of Animal Breeding and Genetics, Veterinary University of Vienna; and 8University Center for Biomodels Austria VUW, Vienna, Austria

Abstract role of Tyk2 for cancer progression. We have previously shown that À/À Tyk2À/À Tyk2 mice are prone to develop B-cell leukemia in an Abelson- We showed previously that natural killer cells lack the À/À ability to lyse leukemic cells. As a consequence, the animals are induced leukemia model. In this study, Tyk2 mice were found to leukemia prone. Here, we show that the impaired tumor have impaired tumor surveillance as a result of decreased natural À/À killer (NK)-cell function (18). Search for cancer-associated genetic surveillance extends to T cells. Challenging Tyk2 mice with EL4 thymoma significantly decreased disease latency. The alterations has led to the identification of different germ-line crucial role of Tyk2 for CTLfunction was further characterized mutations and sequence variants of Tyk2 in various tumors À/À using the ovalbumin-expressing EG7 cells. Tyk2 OT-1 mice (19–21). Tomasson and colleagues describe a potential association developed EG7-induced tumors significantly faster compared of Tyk2 germ-line sequence changes to acute myelogenous with wild-type (wt) controls. In vivo assays confirmed the defect leukemia pathogenesis. Elevated Tyk2 expression levels detected in CD8+ cytotoxicity on Tyk2 deficiency and clearly linked it to in breast cancer cell lines prompted the authors to propose Tyk2 as type I IFN signaling. An impaired CTLactivity was only a potential biomarker for breast cancer diagnosis (22). Moreover, À/À observed in IFNAR1 animals but not on IFN; or IL12p35 loss of Tyk2 expression and signaling has been shown to inhibit the deficiency. Accordingly, EG7-induced tumors grew faster in invasiveness of human prostate cancer cells (23) and of leukemic B À/À À/À À/À À/À cells in a murine model system of human Burkitt’s lymphoma IFNAR1 and Tyk2 but not in IFNg or IL12p35 mice. Adoptive transfer experiments defined a key role of Tyk2 in (24). Complete loss of Tyk2 in humans results in enhanced allergic CTL-mediated tumor surveillance. In contrast to wt OT-1 cells, and impaired antimicrobial responses as recently described by À/À Tyk2 OT-1 Tcells were incapable of controlling EG7-induced Minegishi and colleagues (25). tumor growth. [Cancer Res 2009;69(1):203–11] NK cells and CTLs are regarded as the major mediators of the natural host response against developing tumors. It is nowadays widely accepted that the immune system cannot only distinguish Introduction self from nonself but is additionally able to identify and Signal transduction via the Janus kinase/signal transducer and subsequently eliminate cancerous self. This process is referred activator of transcription (Jak/Stat) pathway regulates cell to as tumor immune surveillance and is one of three phases proliferation, differentiation, and survival of hematopoietic cells that make up the concept of immunoediting (26). The elimination (1–3). Aberrant activation of Jak/Stat signaling was observed in and immunoediting processes involve different cell types, ah multiple tumors and leukemia, and great effort has been made most importantly T and NK cells and cytokines such as IFNs and IL12. in the past years to elucidate and define the role of the À/À Jak/Stat pathway in cancer formation and tumor immune Because IFN and IL12 signaling is compromised in Tyk2 surveillance (4–9). mice, we postulated that not only the NK-cell–mediated antitumor Tyrosine kinase 2 (Tyk2), one of the Jaks, is activated on binding response but also tumor surveillance by CTLs is impaired in these mice. Indeed, our study proves that Tyk2 is a central player in of type I IFN, interleukin (IL) 6, IL10, IL12, and IL13 to the À À À À / + respective receptors (10). Mice deficient for Tyk2 (Tyk2 / mice) tumor surveillance mediated by cytotoxic T cells. Tyk2 CD8 are viable and fertile but were described to have defects in T cells exhibited a severe defect in the antigen-specific cytotoxicity in vitro and in vivo. We postulate a link between this functional type I IFN, IL12, and also type II IFN (IFNg) signaling (11, 12). À À À À / Tyk2 / mice show an increased susceptibility to various defect and the compromised type I IFN signaling in Tyk2 pathogens (11, 13–15) but are resistant against lipopolysaccharide animals, whereas type II IFN and IL12 seem to be of minor (LPS)–induced endotoxin shock (16). In addition, Tyk2 is a key relevance. regulator for Th1/Th2 balance (17). Despite the broad interest for Tyk2 in infection, only sparse information is available on the Materials and Methods Mice À À À À À À À À Requests for reprints: Dagmar Stoiber, Institute of Pharmacology, Medical Wild-type (wt), Rag2 / (27), Tyk2 / (11), IFNc / (28), IFNAR1 / À À University of Vienna, Waehringerstrasse 13A, A-1090 Vienna, Austria. Phone: (29), IL12p35 / (30), and OT-1 (31) mice were on C57BL/6 background 43-1-4277-64130; Fax: 43-1-4277-9641; E-mail: [email protected]. I2009 American Association for Cancer Research. and were maintained under specific pathogen-free conditions at the doi:10.1158/0008-5472.CAN-08-1705 Biomedical Research Institute, Medical University of Vienna (Vienna, www.aacrjournals.org 203 Cancer Res 2009; 69: (1).January 1, 2009

Austria). Experiments were carried out with gender- and age-matched 6- to Tumor Models À À 12-wk-old mice. All experimental procedures were carried out according to Systemic EL4 tumor model. Tyk2 / and wt recipient mice received Austrian Law (TierversuchsgesetzBGBl Nr. 501/1989). 5 Â 103 EL4 cells in 200 AL sterile PBS via tail vein injection. Mice were monitored for disease onset. Sick mice were sacrificed and examined. Livers Culture of Tumor Cell Lines were isolated, weighed, and fixed in 3.7% formaldehyde and analyzed. EL4 and EG7 cells were maintained in DMEM (Sigma-Aldrich) containing S.c. EG7 tumor formation. EG7 cells (106) in 100 AL PBS were injected 10% heat-inactivated FCS (PAA Laboratories), 100 units/mL penicillin- s.c. into the shaved right flanks of recipient mice. Mice were checked every streptomycin (Life Technologies), 2 mmol/L L-glutamine (Life Technologies), other day for the development of tumors. After 10 d, mice were sacrificed A h nonessential amino acids (PAA Laboratories), and 5 mol/L -mercaptoe- and tumors and spleens were isolated and weighed. Tumors and spleens thanol (Sigma-Aldrich). were homogenized and subjected to FACS analysis. Generation and Cultivation of Bone Marrow–Derived CD8+ T-Cell Depletion Dendritic Cells CD8+ T cells were depleted by i.p. injections of 100 Ag 53-6.72 anti–CD8+- Bone marrow was isolated from femora and tibiae of 6- to 8-wk-old mice. specific antibody (BioXCell) twice a week, starting 2 d before tumor cell The immature dendritic cell (DC) population was enriched by panning injection. over human IgG (Sigma-Aldrich)–coated plates (Optilux, Falcon, Becton Dickinson Labware). Cells were cultivated in complete DMEM supple- Adoptive Cell Transfer À À mented with 10% (v/v) granulocyte macrophage colony-stimulating factor CD8+ T cells were isolated by MACS from wt OT-1 and Tyk2 / OT-1 (GM-CSF) supernatant derived from the GM-CSF–producing hybridoma spleens. Two days after s.c. EG7 tumor cell injection, tumor-bearing mice À À line X6310. Culture medium was exchanged every other day. On day 6, received 106 wt OT-1 or Tyk2 / OT-1 CD8+ cells in 100 AL PBS i.v. Tumor immature bone marrow–derived DCs (BMDC) were activated with LPS width (L) and length (l) were measured with a caliper and tumor size was (100 ng/mL; Escherichia coli O111:B4; Calbiochem) or CpG-ODN 1668 calculated using the following formula: l Â L Â (l + L)/2. The experiment (1 Amol/L; MWG-Biotech AG) and loaded with ovalbumin (OVA; 10 Ag/mL, was terminated after 20 d. Tumors, lymph nodes, and spleens were excised, b grade V; Sigma-Aldrich), OVA257-264 (SIINFEKL, 10 Ag/mL, H2-K , derived one part fixed in 3.7% formaldehyde, and the other part homogenized and from chicken OVA; Bachem), or left untreated. After 24 h, cells were subjected to FACS analysis. harvested and used for subsequent experiments. The supernatants were In vitro collected for ELISA. Cytotoxicity Assay A standard 51Cr-release assay was performed to measure the in vitro À À Measurement of Cytokine Production cytotoxicity of wt versus Tyk2 / OT-1 T cells as described (32) with slight À À Cytokine production of BMDCs was determined by measuring IL12p70 modifications. Splenocytes from wt and Tyk2 / OT-1 mice were cocultured (Quantikine Mouse IL-12p70 Immunoassay, R&D Systems) and IL10 with irradiated (30 Gy), peptide-loaded (10 Ag/mL, 1 h at 37jC) naive (OptEIA Mouse IL10 ELISA set, BD Biosciences) in cell culture supernatants splenocytes in the presence of 100 units/mL IL2 (Fresenius Kabi) for 5 d. of 24-h stimulated BMDCs. Effectors were harvested and analyzed for %CD8+Va2+ cells by FACS and adjusted. Targets (104 per well) were incubated with effectors for + CD8 T Lymphocyte Isolation 5 h at 37jC. Percent specific lysis was determined with following formula: + À/À Naive CD8 cells were isolated from spleens of wt OT-1 and Tyk2 [CTL-induced release (cpm) À spontaneous release (cpm)]/[maximum + OT-1 mice using the CD8a Cell Isolation kit (Miltenyi Biotech). Purity of release (cpm) À spontaneous release (cpm)] Â 100. the enriched CD8a+ T cells was evaluated by fluorescence-activated cell sorting (FACS). Isolated cells were usually >90% CD8+ Va2+. In vivo CTLAssay Recipient mice were immunized via footpad injection (50 AL/footpad) Antigen-Specific T-Cell Proliferation and Activation with SIINFEKL (0.1 mg/mouse) alone, SIINFEKL in combination with À À Tyk2 / OT-1 and wt OT-1 splenocytes were stained with 2.5 Amol/L adjuvant (IC31 or 50 Ag CpG-ODN 1668), or PBS. Seven days later, 6-carboxyfluoresein diacetate succinimidyl ester (CFSE; Molecular Probes). immunized mice received differentially (0.025, 0.25, and 2.5 Amol/L) CFSE- CD8+ cells (106) were then cocultivated with mature peptide-loaded BMDCs labeled syngeneic target cells. The CFSElo population was unpulsed, the mi (T-cell to BMDC ratio = 10: 1). Cell samples were taken on 3 subsequent CFSE population was pulsed with an irrelevant peptide (mTRP-2181-188, days and analyzed by FACS. Remaining T cells were harvested for RNA VYDFFVWL; H2-Kb, derived from murine tyrosine-related protein-2; isolation. Bachem), and the CFSEhi population was pulsed with the relevant peptide (SIINFEKL, 10 Ag/mL). Target cells were mixed at a 1:1:1 ratio and 3 Â 107 Flow Cytometric Analysis (FACS) cells were injected via tail vein into immunized mice. Eighteen to 24 h after a The following antibodies were used: FITC-anti-V 2, allophycocyanin the adoptive transfer, mice were sacrificed and peripheral blood, popliteal (APC)-anti-CD8a, APC-anti-CD11c, R-phycoerythrin (R-PE)-anti-CD11b, lymph nodes, and spleens were isolated and single-cell suspensions were FITC-anti-CD40, biotin-anti-CD80, FITC-anti-MHCII, biotin-anti-CD86, analyzed by FACS. Specific killing was calculated as follows: % specific FITC-anti-CD107a, streptavidin-PerCP, R-PE-anti-CD44, and FITC-anti- killing = [1 À (% CFSEhi / % CFSElo)] Â 100. Results are expressed as the CD69 (all BD Pharmingen). Samples were analyzed using a FACScan or mean percentage of cell lysis observed for each experimental group F SD. FACScanto II (Becton Dickinson). Statistical Analysis RNA Isolation and Semiquantitative Reverse Transcription- Differences between mean values were calculated using an unpaired PCR Student’s t test, a Mann-Whitney U test, or a one-way ANOVA test, which RNA was isolated with Trizol (TRI Reagent, Sigma-Aldrich) and was corrected by Dunnett’s multiple comparison correction, as applicable. subsequently DNase I treated (DNase I recombinant, Roche Diagnostics). Differences in Kaplan-Meier plots were analyzed for statistical significance The reverse transcription reaction was performed with the GeneAmp RNA using the log-rank test. P values of <0.05 were considered statistically PCR Core kit (Roche Diagnostics). Primer sequences are available on significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001). request. The cDNA concentration was normalized to the amount of h-actin.

Histology Results Spleens and livers were fixed with 3.7% formaldehyde for 24 h, washed in PBS, and embedded in paraffin. Sections (5 Am) were stained with H&E Tyk2 deficiency facilitates systemic EL4-induced disease using standard protocols. Confocal images (magnification, Â100) were progression. We have previously shown that, due to a NK-cell À À taken using a Zeiss Axio Imager.Z1 microscope. defect, Tyk2 / mice are highly susceptible to Abelson-induced B

Cancer Res 2009; 69: (1).January 1, 2009 204 www.aacrjournals.org

Figure 1. Tyk2À/À mice are susceptible to EL4 lymphoma. Tyk2À/À and wt control mice received 5 Â 103 EL4 cells via tail vein injection. A, i, the median survival time of wt mice was 38.5 d (n =6). Tyk2À/À mice died significantly earlier with a median survival time of 24 d (n =6; P =0.0011). ii, H&E-stained histologic sections show comparably dense tumor cell infiltrations in livers of Tyk2À/À (right) and wt mice (middle). Left, healthy liver tissue. B, where indicated, CD8+ T cells were depleted by i.p. injections of 53-6.72 antibody. CD8+ T-cell–depleted wt mice (black dotted line; median survival, 31 d; n =4) diseased significantly earlier compared with nondepleted wt controls (black line; median survival, 38 d; n =5; P < 0.05). Disease progression in CD8+ T-cell–depleted mice was similar to that in Tyk2À/À nondepleted (gray line; median survival, 27 d; n =6) and Tyk2À/À CD8+ T-cell–depleted (gray dotted line; median survival, 25 d; n =4) animals. Points, mean; bars, SE. Both experiments were repeated twice with similar results.

lymphoid leukemia formation (18). In the current study, we signals provided by antigen-presenting cells (APC). In case of naive questioned whether Tyk2 is also required for CD8+ T-cell–mediated T cells, DCs serve as APCs, which interact via MHC class I and À À tumor surveillance. Therefore, we challenged Tyk2 / recipient costimulatory receptors with the T cell. The diminished tumor À À mice with the thymoma cell line EL4. The transplantation of EL4 surveillance in Tyk2 / mice might therefore result from impaired cells and of their OVA-transfected derivatives, EG7 cells, is a DC function. To analyze the interaction between DCs and T cells À À frequently used tumor model for CD8+ T-cell–mediated cytotoxicity after a defined stimulus, we intercrossed Tyk2 / mice with (33–35). I.v. injection of EL4 cells into mice induces a rapidly transgenic OT-1 animals. OT-1 mice are transgenic for a T-cell À/À progressing leukemia/lymphoma. Tyk2 recipient mice injected receptor that recognizes the immunodominant OVA257-264 peptide with EL4 cells developed first signs of the disease significantly SIINFEKL presented by H2-Kb. earlier than wt recipients (Fig. 1Ai). However, in spite of the BMDCs were prepared and maturation was induced by exposure significant difference in disease progression, the disease pattern to LPS or CpG-ODN. This procedure induces the up-regulation of À À was comparable: wt and Tyk2 / animals had dramatically CD40, CD80, CD86, and MHC II that are characteristic markers for enlarged livers with macroscopic tumor nodules as illustrated in mature BMDCs. No differences were observed between wt and À À Fig. 1Aii. The importance of cytotoxic T cells for tumor surveillance Tyk2 / BMDCs irrespective of the stimulus used (Fig. 2A; data not of EL4 cells was confirmed using CD8+ T-cell–depleted animals shown). Mature antigen-presenting DCs can activate an antigen- (Fig. 1B). Whereas CD8+ T-cell depletion exerted only a minor effect specific T-cell clone and induce T-cell proliferation. As depicted in À À in Tyk2 / animals, a significantly accelerated disease progression Fig. 2B, LPS-stimulated or CpG-ODN–stimulated and SIINFEKL- À À was observed in wt CD8+ T-cell–depleted mice when compared loaded wt and Tyk2 / BMDCs induced proliferation of purified with control mice (P < 0.05). This finding supports our hypothesis OT-1 transgenic CD8+ cells equally well (Fig. 2B, top). Hence, of a key role for Tyk2 in CD8+-mediated tumor surveillance BMDCs do not require Tyk2 to induce the expansion of an antigen- À À resulting in the significantly shortened disease latency in Tyk2 / specific T-cell population. To rule out that antigen processing is À À animals. impaired, BMDCs were loaded with undigested OVA. Tyk2 / and Tyk2 is dispensable for the maturation of BMDCs in vitro. wt BMDCs loaded with OVA could equally well induce antigen- An efficient antitumor CD8+ CTL response is initiated by activating specific T-cell proliferation (data not shown). Mature BMDCs also www.aacrjournals.org 205 Cancer Res 2009; 69: (1).January 1, 2009

À À secrete cytokines (IL10 and IL12) and thereby shift the balance to indicates that Tyk2 / CD8+ T cells are equally well equipped either the Th1 or Th2 side of the immune response. Again, no to use either killing pathway. The exocytotic process itself significant alterations in LPS-stimulated or CpG-ODN–stimulated (degranulation) can be quantified by analyzing surface expression À À cytokine secretion were detectable between wt and Tyk2 / of CD107a, also known as lysosome-associated membrane protein BMDCs (Fig. 2C and D). 1. After degranulation, CD107a appears on the cellular surface. As À/À Tyk2 OT-1 CD8+ T cells display an impaired in vitro illustrated in Fig. 3C, CD107a surface expression was comparable À À cytotoxicity despite regular antigen-specific expansion. Be- on wt OT-1 and Tyk2 / OT-1 T cells on activation by mature cause our experiments excluded severe defects in BMDC function, peptide-loaded BMDCs. we next focused on CD8+ T-effector cells themselves. CD8+ T cells However, when OT-1 T cells were coincubated with EG7 cells and À À were purified from Tyk2 / OT-1 and wt OT-1 animals and in vitro cytotoxicity was analyzed, we realized a significant activated by coculture with LPS-stimulated or CpG-ODN–stimu- difference: whereas wt OT-1 effector cells efficiently killed EG7 À À lated and peptide-loaded BMDCs. In this experimental setting, cells, Tyk2 / OT-1 T cells displayed a highly significant reduction CD8+ cells expanded equally well irrespective of the genotype of the in specific lysis at all E:T ratios measured (Fig. 3D). À/À BMDCs or T cells (Fig. 2B). Accordingly, the expression of the T-cell Enhanced tumor formation in Tyk2 OT-1 animals is activation markers CD44 and CD69 was induced to the same extent linked to impaired CTLkilling. Simplified in vitro settings may (Fig. 3A). not sufficiently mimic the complex in vivo situation. Therefore, À À Activated CD8+ CTLs use two different mechanisms to kill their we challenged Tyk2 / OT-1 T-cell function in vivo by s.c. À À target cells: the granule-dependent (perforin/granzyme) or granule- injection of the EG7 cell line. Again, Tyk2 / OT-1 animals were independent (death receptor–induced cell death; e.g., Fas-FasL) significantly less capable to suppress the proliferation of the À À pathway. As depicted in Fig. 3B, Tyk2 / CD8+ cells expressed injected EG7 cells (Fig. 4A). To unequivocally link this important comparable levels of FasL, granzyme B, and perforin mRNA. This defect to the T-cell compartment, we adoptively transferred either

Figure 2. Bone marrow–derived Tyk2À/À DCs have no obvious defects in vitro. A, DCs were generated from wt and Tyk2À/À bone marrow in the presence of GM-CSF. After 6 d of culture, BMDCs were stimulated with LPS for 24 h and analyzed for surface expression of CD40, CD80, CD86, and MHC II on CD11b+CD11c+ gated BMDCs for each genotype. Costimulatory molecule expression on stimulated BMDCs (black curve) was compared with expression levels on immature BMDCs (gray filled curve). No differences in the expression patterns of the surface molecules could be detected between wt (top row) and Tyk2À/À (bottom row) BMDCs. B, mature SIINFEKL peptide-loaded BMDCs were incubated with CFSE-labeled naive OT-1 splenocytes and T-cell proliferation was determined by flow cytometry on 3 subsequent days. Representative histogram plots of wt OT-1 T cells cocultivated with either wt BMDCs (top left)orTyk2À/À BMDCs (top right)and Tyk2À/À OT-1 T cells incubated with either wt BMDCs (bottom left)orTyk2À/À BMDCs (bottom right) are shown. Proliferation rates were comparable in all T-cell plus BMDC combinations tested (ratio, 10:1). The experiment was repeated four times with similar results. C and D, IL12 and IL10 concentration in the supernatants of LPS-stimulated BMDCs was measured by ELISA. Tyk2À/À BMDCs were found to produce equal amounts of IL12 (C) and IL10 (D) compared with wt BMDCs. BMDCs incubated with medium only served to determine basal secretion levels. Cumulative data from at least two independent experiments were summarized. Fold increase after LPS stimulation in comparison with basal levels is shown; basal levels were comparable (IL12 wt: 5.77 F 0.50 pg/mL, Tyk2À/À: 5.98 F 0.43 pg/mL; IL10 wt: 160.5 F 27.89 pg/mL, Tyk2À/À :165.1F 24.14 pg/mL).

Cancer Res 2009; 69: (1).January 1, 2009 206 www.aacrjournals.org

Figure 3. Tyk2À/À CD8+ T cells are impaired in cytotoxic target cell killing. Apart from that, Tyk2À/À CD8+ T cells were phenotypically indistinguishable from wt CD8+ T cells on activation in vitro. Splenocytes from Tyk2À/À OT-1 and wt OT-1 mice were cocultured for 3 d with Tyk2À/À and wt mature peptide-loaded BMDCs, respectively. A, expression of activation markers CD44 and CD69 on OT-1 CD8+ T cells was measured by flow cytometry on all 3 d (n, wt BMDCs with wt OT-1 T cells; E, wt BMDCs with Tyk2À/À OT-1 T cells; 5, Tyk2À/À BMDCs with wt OT-1 T cells; D, Tyk2À/À BMDCs with Tyk2À/À OT-1 T cells). The percentages of positively stained CD8+ cells are depicted. B, semiquantitative reverse transcription-PCR analysis revealed no differences in mRNA levels of FasL, perforin, and granzyme B in Tyk2À/À compared with wt OT-1 CD8+ T cells after cocultivation with mature peptide-loaded BMDCs. C, after 3 d of incubation, degranulation of OT-1 CD8+ T cells was determined by detection of lysosome-associated membrane protein CD107a by flow cytometry. Gray filled curve, cultures with immature unloaded BMDCs served as negative controls. Representative data of three independent experiments are depicted. D, the in vitro cytotoxicity of wt OT-1 (black curve) and Tyk2À/À OT-1 (gray curve) CD8+ T cells was measured in a standard 51Cr-release assay using EG7 cells as targets. EL4 cells served as negative controls (data not shown). The specific lysis at different E:T cell ratios (100:1, 33:1, 11:1, 3.7:1, and 1.2:1) is shown (wt: 50.96 F 1.55%, 45.70 F 1.71%, 33.89 F 0.77%, 19.17 F 1.27%, 9.92 F 0.65%; Tyk2À/À: 35.03 F 0.43%, 27.01 F 0.07%, 16.80 F 0.15%, 7.51 F 1.17%, 2.63 F 1.00%). Points, mean from triplicate wells; bars, SD.

À À À À wt OT-1 or Tyk2 / OT-1 CD8+ T cells into Tyk2 / mice 2 days The antigen-specific CTLfunction is compromised in the after EG7 tumor inoculation. As depicted in Fig. 4B (left), wt OT-1 absence of Tyk2/IFNAR1 signaling. Thus far, our observations cells significantly repressed EG7 tumor growth. In contrast, the revealed a severe defect in CTL-mediated tumor surveillance in À À À À injection of Tyk2 / OT-1 cells failed to inhibit tumor progression Tyk2 / mice. This defect is attributed to an impaired T-cell À À À À (Fig. 4B, middle). Tumors under the control of Tyk2 / OT-1 function and is recapitulated in Tyk2 / OT-1 animals. Previous CD8+ cells had a similar growth pattern when compared with experiments have shown that Tyk2 is required for IFN and IL12 PBS-injected controls (Fig. 4B, right). This observation was signaling, which is also discussed to contribute to CTL-mediated confirmed by the comparison of tumor weights 17 days after cytotoxicity. To test whether one of these factors is indeed essential the initial injection (Fig. 4C). The quantification of OT-1 cells for CTL-mediated effector function, we performed in vivo CTL À À À À À À À À (CD8+ Va2+) in the spleens of the recipient animals revealed assays with wt, Tyk2 / ,IFNc / , IFNAR1 / , and IL12p35 / mice. higher numbers of OT-1 TcellsinwtOT-1–treated mice Wt mice that had been immunized with peptide and adjuvant compared with control mice that, however, did not meet the showed a major cytotoxic effect (Fig. 5A, arrow). In contrast, in À À À À criteria of statistical significance (wt OT-1: 0.98 F 0.21%, Tyk2 / Tyk2 / mice, the antigen-specific cytotoxicity was significantly OT-1: 0.47 F 0.13%, PBS: 0.50 F 0.15%). In a similar experiment reduced (Fig. 5A and Bi). This was accompanied by reduced À À but this time using CD8+ T cells from previously tumor cell– numbers of IFNg-secreting Tyk2 / splenocytes (data not shown). À À immunized wt and Tyk2 / mice, comparable observations were Under the same assay conditions, antigen-specific cytotoxicity in À À À À made. Tumor development was delayed on adoptive transfer of wt IFNc / and IL12p35 / mice was unaltered (Fig. 5Bii and Biii). À À CD8+ T cells but not on adoptive transfer of Tyk2 / CD8+ However, the CTL response was severely and significantly impaired À À T cells (data not shown). These experiments underscore and in IFNAR1 / mice (Fig. 5Biv). À/À confirm the important function of Tyk2 in CTL-mediated tumor S.c. EG7 tumor growth is accelerated in Tyk2 and À/À cell rejection. IFNAR1 mice. Our findings indicated that the severe reduction www.aacrjournals.org 207 Cancer Res 2009; 69: (1).January 1, 2009

Figure 4. Tyk2À/À CD8+ T cells are impaired in controlling EG7 tumor growth. A, 106 EG7 cells were injected s.c. into recipient mice and tumor weight was determined after 10 d. Tumors were significantly bigger in Tyk2À/À OT-1 mice compared with wt OT-1 mice (0.11 F 0.04 g, n =9, versus 0.01 F 0.01 g, n =8; P =0.036). B, Tyk2À/À recipient mice were injected s.c. with 106 EG7 tumor cells on day À2. On day 0, tumor-bearing mice received 106 wt OT-1 CD8+ T cells (left; n =6) or Tyk2À/À OT-1 CD8+ T cells (middle; n =6) via tail vein injection. Right, control mice were injected with PBS only (n =6). In PBS-treated mice, tumors grew progressively, whereas in mice treated with wt OT-1 CD8+ T cells tumor growth was significantly slower. In mice treated with Tyk2À/À OT-1 CD8+ T cells, tumors grew equally fast as in PBS-treated controls. C, the tumor weight was determined at the end of the experiment (day 17). Mice that had received wt OT-1 T cells had significantly smaller tumors than the PBS-treated control mice (1.12 F 0.40 g, n =6, versus 4.49 F 0.40 g, n =6; P =0.0022). Tumors isolated from Tyk2À/À OT-1 T-cell–treated mice were comparable in weight to the controls (3.54 F 0.75 g, n =6). Columns, mean tumor weight; bars, SE.

À À in CD8+ CTL function in Tyk2 / animals is unrelated to an this result and again proves that Tyk2 deficiency renders animals impaired IL12 and IFNg signaling but rather associated with the highly susceptible for the development of malignant disease. The inability to react to type I IFN. To test whether these findings are enhanced susceptibility for EL4-induced tumor formation has to relevant for CTL-mediated tumor formation, 106 EG7 cells were be accredited, at least for the most part, to defects of CTL- À À injected s.c. into wt and Tyk2 / animals. Tumor formation was mediated cytotoxicity. The depletion of CD8+ T cells significantly analyzed 10 days later. As expected, tumor growth was significantly accelerated tumor growth in wt animals but had almost no effect À À À À increased in Tyk2 / animals. Depletion of CD8+ cells in wt mice in Tyk2 / mice. The significant contribution of Tyk2 for CTL- significantly enhanced tumor growth but did not result in altered mediated cytotoxicity was further supported by several other À À À À tumor formation in Tyk2 / mice, confirming the CTL-mediated findings: Tyk2 / mice immunized with antigen and adjuvant tumor surveillance for s.c. EG7 tumor rejection. No significant show a severe defect in CTL-mediated cytotoxicity and reduced À À differences were observed when wt and IFNc / mice were IFNg production. Moreover, we showed the detrimental effect of À À compared (Fig. 6B). We next compared EG7 tumor growth in this defect on tumor surveillance in vivo: Tyk2 / OT-1 mice À À À À IL12p35 / and IFNAR1 / mice. Interestingly and in accordance develop significantly larger EG7 cell–initiated tumors than wt with our previous findings, EG7 cell–induced tumors were OT-1 control mice. À À significantly bigger in IFNAR1 / mice, whereas IL12p35 deficiency Tyk2 has been implicated in IFN and IL12 signaling, both of was irrelevant for the outcome of this experiment (Fig. 6C). Hence, which are considered important for CD8+ T-cell effector functions these experiments define type I IFN as key component in this (36, 37). Interestingly, IL12 and IFNg were dispensable for target process. cell killing in the in vivo CTL assay. In line with this, IL12 or IFNg deficiency did not confer a growth advantage on EG7-induced tumors, confirming the negligible role of these cytokines in CTL- Discussion mediated tumor surveillance in this experimental system. However, À À A role of Tyk2 in NK-cell–mediated tumor surveillance against in IFNAR1 / mice, the CTL-mediated cytotoxicity was severely À À À À hematopoietic malignancies has recently been established (18). In impaired. IFNAR1 / and Tyk2 / animals displayed similar our current study, we extend these findings and show that Tyk2 is phenotypes indicating that both type I IFN and Tyk2 are key also a critical component of CD8+-mediated tumor surveillance components of CTL-mediated cytotoxicity. Recent observations downstream of type I IFN signaling. made in vaccination studies further support our hypothesis that À À Tumors develop significantly faster in Tyk2 / animals. The i.v. type I IFN is of utmost importance for the generation of tumor- À À injection of the thymoma cell line EL4 in Tyk2 / mice confirmed specific CTLs (38).

Cancer Res 2009; 69: (1).January 1, 2009 208 www.aacrjournals.org

Although initially IFNg was considered the key IFN involved in Our work now emphasizes that beside these well-known tumor surveillance, the importance of type I IFN has soon been effects, type I IFNs are important components of CTL effector recognized. Type I IFN exerts immune modulatory effects on function in antitumor immunity and that these effects are several cell types, including NK cells, DCs, and effector cells of the mediated via Tyk2. One might speculate that type I IFN- adaptive immune response. These activities can, in part, be dependent Tyk2 activation is a limiting factor for CTL-dependent attributed to an increased expression of MHC I and costimulatory cytotoxicity, whereas Tyk2 activation downstream of IFNg and molecules, thus an enhancement of antigen presentation and T-cell IL12 contributes to immune surveillance of tumors exerted by the activation. In addition, type I IFN was described to be critical innate immune system. for the antitumor function of NK cells as well as T lymphocytes One way by which type I IFN signaling might impair CTL (26, 39). activation is via DC-mediated antigen presentation. We and others

Figure 5. Antigen-specific cytotoxicity is impaired in Tyk2À/À mice. For the in vivo CTL assay, Tyk2À/À and wt mice were immunized with SIINFEKL alone or SIINFEKL in combination with adjuvant via footpad injection. One week after immunization, mice received a mixed target cell population consisting of unloaded cells labeled with 0.025 mmol/L CFSE, cells loaded with an irrelevant antigen (mTRP-2181-188) labeled with 0.25 mmol/L CFSE, and cells loaded with the relevant antigen (SIINFEKL) labeled with 2.5 mmol/L CFSE. After 18 to 24 h, the draining (popliteal) lymph nodes, spleens, and peripheral blood of wt and Tyk2À/À mice were isolated and subjected to flow cytometric analysis. No antigen-specific cytotoxicity was detected in naive mice (A, top) or mice immunized with peptide alone (A, middle). Only in lymph nodes from wt mice immunized with peptide and adjuvant, the cell population loaded with the relevant peptide had disappeared, indicating antigen-specific killing (A, bottom left, arrow). In case of Tyk2À/À mice, the target cell population was still detectable. For each group, a representative histogram plot is shown. B, i, the antigen-specific killing in Tyk2À/À mice was significantly impaired (wt: 76.30 F 12.48%, n =4, versus Tyk2À/À : 26.66 F 1.96%, n =3; P =0.0205). In vivo CTL assays were performed in IFNcÀ/À, IL12p35À/À, and IFNAR1À/À mice. The antigen- specific CTL function in IFNcÀ/À mice (ii) and in IL12p35À/À (iii) was not reduced compared with wt mice. iv, antigen-specific killing in IFNAR1À/À mice was significantly lower than in wt mice (wt: 75.33 F 5.76%, n =4, versus IFNAR1À/À : 31.60 F 2.56%, n =5; P =0.0001). Data depicted refer to lymph nodes; data from peripheral blood and spleen were comparable (data not shown). Experiments were repeated twice with similar results.

www.aacrjournals.org 209 Cancer Res 2009; 69: (1).January 1, 2009

À À have failed to show any defects in Tyk2 / DC maturation and their ability to stimulate T-cell proliferation in vitro (15). Nevertheless, we cannot yet fully exclude a contribution of Tyk2 for DC functions in vivo. This issue will ultimately be settled by the use of conditionally gene targeted animals. Based on our experiments, we propose that type I IFN and Tyk2 critically contribute to the functional maturation of CTLs. À À The adoptive transfer of Tyk2 / OT-1 CD8+ T cells did not suppress tumor growth in vivo. In contrast, the adoptive transfer of wt OT-1 CD8+ cells efficiently suppressed tumor growth. In this experimental setting, the adoptively transferred CTLs themselves determine the experimental outcome because all other experimental conditions and the immune environment are superimposable. Even under the premises that DC activation is of minor importance within the OT-1 system, the defect in the CTLs still remains a matter of fact. Thus, the adoptive transfer experiments point at an intrinsic defect within CD8+ T cells on Tyk2 deficiency. In view of our findings, it seems unlikely that the proliferation of CD8+ cells is impaired À À on Tyk2 deficiency. Tyk2 / CD8+ cells proliferated with identical kinetics in vitro. These findings strengthen the concept that Tyk2 inhibition is a powerful therapeutic option for patients suffering from diseases linked to the hyperactivation of the immune system. Tyk2 inhibition is supposed to efficiently suppress NK- and CTL- mediated killing and this renders Tyk2 inhibitors highly suitable for the treatment of autoimmune disorders or for patients undergoing transplantation. Similarly, Jak3 has been discussed as novel drug target for the treatment of rheumatoid arthritis, psoriasis, and transplant rejection (40). Clinical studies are ongoing to prove the usefulness of the Jak3 inhibitors (NCT00483756, NCT00263328, NCT00413699, and NCT00550446).9 We suggest that Tyk2 might be an even more promising candidate. Tyk2 inhibitors should stand out with higher selectivity and lower probability of adverse side effects because fewer signaling pathways are affected by the loss of Tyk2.

9 http://clinicaltrials.gov

Disclosure of Potential Conflicts of Interest

Figure 6. Control of EG7 tumor growth is dependent on Tyk2 and IFNAR1. No potential conflicts of interest were disclosed. Recipient mice were s.c. injected with 106 EG7 cells. After 10 d, mice were sacrificed and tumors were isolated. A, tumor formation in wt, wt CD8+ T-cell–depleted, Tyk2À/À, and Tyk2À/À CD8+ T-cell–depleted mice was À À compared. Tumors of Tyk2 / animals (1.10 F 0.12 g, n =12) were significantly Acknowledgments heavier (P < 0.001) than tumors of wt mice (wt: 0.33 F 0.10 g, n =9). + Tumors of CD8 T-cell–depleted wt animals (0.87 F 0.08 g, n =6) were similar Received 5/6/2008; revised 10/13/2008; accepted 10/18/2008. À/À + À/À in weight to tumors of Tyk2 and CD8 T-cell–depleted Tyk2 animals Grant support: Austrian Academy of Sciences at the Institute of Pharmacology (1.16 F 0.19 g, n =8) but significantly ( P < 0.05) bigger than tumors of DOC-FFORTE fellowship (O. Simma), Austrian National Bank project no. 9875 and À/À À/À nondepleted wt mice. B, tumor growth in wt, Tyk2 , and IFNc mice was FWF grant P19534-B13 (D. Stoiber), Austrian Federal Ministry of Science and Research À/À compared. Tumors of Tyk2 animals were significantly (P < 0.01) bigger than grant BM_WFa GZ200.112/1-VI/1/2004 and FWF grant SFB F2808 (M. Mu¨ller), and À/À tumors of wt animals (Tyk2 : 0.76 F 0.13 g, n =7, versus wt: 0.32 F 0.06 g, FWF grants P15033 and SFB F2810 (V. Sexl). À/À n =10). Tumor growth in IFNc mice was similar to that in wt mice (0.40 F The costs of publication of this article were defrayed in part by the payment of page À/À À/À À/À 0.08 g, n =8). C, tumor formation in wt, Tyk2 , IL12p35 , and IFNAR1 charges. This article must therefore be hereby marked advertisement in accordance À/À mice was compared. In IL12p35 mice, tumor growth was comparable with 18 U.S.C. Section 1734 solely to indicate this fact. À/À with tumor growth in wt mice (IL12p35 : 0.70 F 0.13 g, n =8 versus wt: We thank Udo Losert and the staff of the Biomedical Research Institute for taking À/À À/À 0.72 F 0.05 g, n =25). Similar to Tyk2 mice (Tyk2 : 1.09 F 0.16 g, n =8; good care of our mice; Marina Karaghiosoff, Birgit Strobl, and Richard Moriggl for À/À P < 0.05), tumors in IFNAR1 mice grew significantly faster than in wt mice helpful discussions and support; and the Department of Neuro-Immunology, Center (1.05 F 0.13 g, n =8; P < 0.05). Columns, mean tumor weight; bars, SE. for Brain Research, Medical University of Vienna for help with histopathology.

Cancer Res 2009; 69: (1).January 1, 2009 210 www.aacrjournals.org

References Tyk2 in APC function of dendritic cells for priming 29. Muller U, Steinhoff U, Reis LF, et al. Functional role of CD8+ T cells producing IFN-g. Eur J Immunol 2006; type I and type II interferons in antiviral defense. 1. Ihle JN. The Stat family in cytokine signaling. Curr 36:3060–70. Science (New York) 1994;264:1918–21. Opin Cell Biol 2001;13:211–7. 16. Karaghiosoff M, Steinborn R, Kovarik P, et al. Central 30. Mattner F, Magram J, Ferrante J, et al. Genetically 2. Yamaoka K, Saharinen P, Pesu M, Holt VE III, role for type I interferons and Tyk2 in lipopolysaccharide- resistant mice lacking interleukin-12 are susceptible to Silvennoinen O, O’Shea JJ. The Janus kinases (Jaks). induced endotoxin shock. Nat Immunol 2003;4:471–7. infection with Leishmania major and mount a polarized Genome Biol 2004;5:253. 17. Seto Y, Nakajima H, Suto A, et al. Enhanced Th2 cell- Th2 cell response. Eur J Immunol 1996;26:1553–9. 3. O’Shea JJ, Gadina M, Schreiber RD. Cytokine signaling mediated allergic inflammation in Tyk2-deficient mice. 31. Hogquist KA, Jameson SC, Heath WR, Howard JL, in 2002: new surprises in the Jak/Stat pathway. Cell 2002; J Immunol 2003;170:1077–83. Bevan MJ, Carbone FR. T cell receptor antagonist 109 Suppl:S121–31. 18. Stoiber D, Kovacic B, Schuster C, et al. TYK2 is a key peptides induce positive selection. Cell 1994;76:17–27. 4. Yu H, Jove R. The STATs of cancer—new molecular regulator of the surveillance of B lymphoid tumors. 32. Derdak SV, Kueng HJ, Leb VM, et al. Direct targets come of age. Nat Rev 2004;4:97–105. J Clin Invest 2004;114:1650–8. stimulation of T lymphocytes by immunosomes: virus- 5. Garcia R, Bowman TL, Niu G, et al. Constitutive 19. Ruhe JE, Streit S, Hart S, et al. Genetic alterations in like particles decorated with T cell receptor/CD3 ligands activation of Stat3 by the Src and JAK tyrosine kinases the tyrosine kinase transcriptome of human cancer cell plus costimulatory molecules. Proc Natl Acad Sci U S A participates in growth regulation of human breast lines. Cancer Res 2007;67:11368–76. 2006;103:13144–9. carcinoma cells. Oncogene 2001;20:2499–513. 20. Kaminker JS, Zhang Y, Waugh A, et al. Distinguishing 33. Hollenbaugh JA, Dutton RW. IFN-g regulates donor 6. Gao B, Shen X, Kunos G, et al. Constitutive activation cancer-associated missense mutations from common CD8 T cell expansion, migration, and leads to apoptosis of JAK-STAT3 signaling by BRCA1 in human prostate polymorphisms. Cancer Res 2007;67:465–73. of cells of a solid tumor. J Immunol 2006;177:3004–11. cancer cells. FEBS Lett 2001;488:179–84. 21. Tomasson MH, Xiang Z, Walgren R, et al. Somatic 34. Chung SW, Cohen EP, Kim TS. Generation of tumor- 7. Frank DA, Mahajan S, RitzJ. B lymphocytes from mutations and germline sequence variants in the specific cytotoxic T lymphocyte and prolongation of the patients with chronic lymphocytic leukemia contain expressed tyrosine kinase genes of patients with de novo survival of tumor-bearing mice using interleukin-18- signal transducer and activator of transcription (STAT) acute myeloid leukemia. Blood 2008;111:4797–808. secreting fibroblasts loaded with an epitope peptide. 1 and STAT3 constitutively phosphorylated on serine 22. Song XC, Fu G, Yang X, Jiang Z, Wang Y, Zhou GW. Vaccine 2004;22:2547–57. residues. J Clin Invest 1997;100:3140–8. Protein expression profiling of breast cancer cells by 35. Chamoto K, Wakita D, Narita Y, et al. An essential 8. Coffer PJ, Koenderman L, de Groot RP. The role of dissociable antibody microarray (DAMA) staining. Mol role of antigen-presenting cell/T-helper type 1 cell-cell STATs in myeloid differentiation and leukemia. Cell Proteomics 2008;7:163–9. interactions in draining lymph node during complete Oncogene 2000;19:2511–22. 23. Ide H, Nakagawa T, Terado Y, Kamiyama Y, Muto S, eradication of class II-negative tumor tissue by T-helper 9. Gouilleux-Gruart V, Debierre-Grockiego F, Gouilleux F, Horie S. Tyk2 expression and its signaling enhances the type 1 cell therapy. Cancer Res 2006;66:1809–17. et al. Activated Stat related transcription factors in invasiveness of prostate cancer cells. Biochem Biophys 36. Ikeda H, Old LJ, Schreiber RD. The roles of IFNg in acute leukemia. Leuk Lymphoma 1997;28:83–8. Res Commun 2008;369:292–6. protection against tumor development and cancer 10. Schindler C, Levy DE, Decker T. JAK-STAT signaling: 24. Schuster C, Muller M, Freissmuth M, Sexl V, Stoiber immunoediting. Cytokine Growth Factor Rev 2002;13: from interferons to cytokines. J Biol Chem 2007;282: D. Commentary on H. Ide et al.,‘‘Tyk2 expression and its 95–109. 20059–63. signaling enhances the invasiveness of prostate cancer 37. Trinchieri G. Interleukin-12 and the regulation of 11. Karaghiosoff M, Neubauer H, Lassnig C, et al. Partial cells’’. Biochem Biophys Res Commun 2008;366:869–70. innate resistance and adaptive immunity. Nat Rev impairment of cytokine responses in Tyk2-deficient 25. Minegishi Y, Saito M, Morio T, et al. Human tyrosine Immunol 2003;3:133–46. mice. Immunity 2000;13:549–60. kinase 2 deficiency reveals its requisite roles in multiple 38. Wakita D, Chamoto K, Zhang Y, et al. An indispens- 12. Shimoda K, Kato K, Aoki K, et al. Tyk2 plays a cytokine signals involved in innate and acquired able role of type-1 IFNs for inducing CTL-mediated restricted role in IFNa signaling, although it is required immunity. Immunity 2006;25:745–55. complete eradication of established tumor tissue by for IL-12-mediated T cell function. Immunity 2000;13: 26. Swann JB, Smyth MJ. Immune surveillance of tumors. CpG-liposome co-encapsulated with model tumor 561–71. J Clin Invest 2007;117:1137–46. antigen. Int Immunol 2006;18:425–34. 13. Strobl B, Bubic I, Bruns U, et al. Novel functions of 27. Shinkai Y, Rathbun G, Lam KP, et al. RAG-2-deficient 39. Dunn GP, Bruce AT, Sheehan KC, et al. A critical tyrosine kinase 2 in the antiviral defense against murine mice lack mature lymphocytes owing to inability to function for type I interferons in cancer immunoediting. cytomegalovirus. J Immunol 2005;175:4000–8. initiate V(D)J rearrangement. Cell 1992;68:855–67. Nat Immunol 2005;6:722–9. 14. Schleicher U, Mattner J, Blos M, et al. Control of 28. Dalton DK, Pitts-Meek S, Keshav S, Figari IS, Bradley 40. Changelian PS, Flanagan ME, Ball DJ, et al. Leishmania major in the absence of Tyk2 kinase. Eur J A, Stewart TA. Multiple defects of immune cell function Prevention of organ allograft rejection by a specific Immunol 2004;34:519–29. in mice with disrupted interferon-g genes. Science (New Janus kinase 3 inhibitor. Science (New York) 2003;302: 15. Aizu K, Li W, Yajima T, et al. An important role of York) 1993;259:1739–42. 875–8.

www.aacrjournals.org 211 Cancer Res 2009; 69: (1).January 1, 2009

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Identification of an Indispensable Role for Tyrosine Kinase 2 in CTL-Mediated Tumor Surveillance

Olivia Simma, Eva Zebedin, Nina Neugebauer, et al.

Cancer Res 2009;69:203-211.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/69/1/203

Cited articles This article cites 40 articles, 10 of which you can access for free at: http://cancerres.aacrjournals.org/content/69/1/203.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/69/1/203.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/69/1/203. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.