| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Internal Medicine,
Department of Surgery,
Department of Pathology,
Center of Clinical Transfusion Medicine, Eberhard Karls-University, Tübingen, Germany;
¶ Department of Internal Medicine, Johannes Gutenberg-University, Mainz, Germany; and
|| Department of Immunology, Eberhard Karls-University, Tübingen, Germany
2Correspondence: Department of Internal Medicine, Eberhard-Karls University, Otfried-Mueller-Str. 10, D-72076 Tuebingen, Germany. E-mail: helmut.salih{at}med.uni-tuebingen.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
production of NK cells demonstrating that constitutive expression of GITRL by tumor cells diminishes NK cell antitumor immunity. GITRL-Ig fusion protein or cell surface-expressed GITRL did not induce apoptosis in NK cells, but diminished nuclear localized c-Rel and RelB, indicating that GITR might negatively modulate NK cell NF-
B activity. Taken together, our data indicate that tumor-expressed GITRL mediates immunosubversion in humans.—Baltz, K. M., Krusch, M., Bringmann, A., Brossart, P., Mayer, F., Kloss, M., Baessler, T., Kumbier, I., Peterfi, A., Kupka, S., Kroeber, S., Menzel, D., Radsak, M. P., Rammensee, H-G., Salih, H. R. Cancer immunoediting by GITR (glucocorticoid-induced TNF-related protein) ligand in humans: NK cell/tumor cell interactions.
Key Words: tumor immunity • TNF family • TNFRSF18 • immune escape
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Both adaptive and innate immune effector cells participate in antitumor immunity, and on the tumor side this process is influenced by a variety of factors like cellular origin, stromal responses, cytokine production profile, and inherent immunogenicity (1)
. Differentiation, proliferation, activation, and death of both immune and tumor cells are substantially influenced by various members of the TNF/TNF receptor (TNFR) superfamily, and thus these molecules markedly influence cancer immunoediting (2)
.
The glucocorticoid-induced TNF-related protein (GITR, also known as TNFRSF18) came to the attention of immunologists because of its purported role in reversing immunosuppressive effects of regulatory T cells (Treg) in mice (reviewed in ref. 3
). Human GITR and its ligand (GITRL) have been identified independently by two groups in 1999 (4
, 5)
. Two years earlier GITR had been identified in mice (6)
. GITR is constitutively expressed at low levels on CD4+ and CD8+ responder T cells and is up-regulated following activation and on Treg (7
, 8)
. In mice, expression of GITR has also been detected on B cells, NKT cells, and macrophages (8
, 9)
. The cognate ligand of GITR, GITRL, has in humans been found on mRNA level in various healthy nonlymphoid tissues, while GITRL protein has been detected on endothelial cells and cells of the eye, and can be up-regulated on the latter by proinflammatory cytokines (4
, 5
, 10
, 11)
. In mice, GITRL has been cloned more recently, and murine GITRL protein has been detected on dendritic cells (DC), monocytes, macrophages, B cells, endothelial cells, and osteoclasts (5
, 6
, 11
12
13
14)
.
Until now, studies analyzing the function of the GITR-GITRL system focused on T cell biology (3)
. GITR was shown to costimulate T cells and to abrogate suppression of Treg (6
7
8
, 11
12
13)
. In mouse models, GITR has been implicated in the development of autoimmune diseases, graft-vs.-host disease, and in the immune response against infectious pathogens (reviewed in ref. 3
). Furthermore, application of monoclonal antibodies (mAb) against mouse GITR or injection of adenovirus expressing recombinant GITRL into palpable tumors has been reported to increase animal survival and even lead to the eradication of tumors, which has been attributed to T cell activation (15
16
17
18
19)
. The functional relevance of GITR and GITRL for non-T cells, however, has yet to be fully explored. Signaling via murine GITRL has been shown to influence activation and survival of macrophages, cytokine production of DC, and differentiation of osteoclasts (11
, 20
21
22
23)
. Recently it was shown that GITRL on human plasmacytoid DC (pDC) is involved in the modulation of NK cell responses (24)
. Since GITR stimulation has been suggested as a possible tool for future immunotherapeutic strategies (17
18
19)
and because it has been reported that suppression of human Treg is not inhibited by GITR, indicating that GITR plays different roles in mice and humans (25)
, we set out to investigate the role of GITRL in human tumor biology. We report here for the first time that GITRL is constitutively expressed by human tumors and directly modulates their immunogenicity, cytokine release, and interaction with NK cells found to express GITR.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Antibodies and cytokines
The mAb used in this study were anti-GITR (clone110416), anti-GITRL (clone 109101 and 109114), anti-CD3-FITC, and anti-CD56-PE-Cy5 from R&D Systems(Wiesbaden, Germany); anti-human PE and anti-mouse PE from Jackson Immunoresearch (West Grove, PA, USA); anti-c-Rel and anti-RelB from Santa Cruz Biotechnology (Santa Cruz, CA, USA); anti-CD40, anti-CD54, anti HLA class I and II and anti-EpCAM were obtained internally at Eberhard Karls-University Tuebingen, (Tuebingen, Germany). Biotin-labeled rabbit-anti-mouse immunoglobulin and streptavidin-HRP were from DAKO (Glostrup, Denmark). Human IL-2 and IL-15 were from Proleukin, Chiron (Ratingen, Germany). All other reagents were obtained from Sigma (St. Louis, MO, USA).
Fusion proteins
The GITR and GITRL fusion proteins (human GITR or GITRL, respectively, with human IgG1 tail) were generated by COS cell transfection as described previously (26)
. cDNA fragments encoding the extracellular regions of human GITR and GITRL were obtained by PCR using the following primers: 5-CAAGCTTGATGGCACAGCACGGGGCG-3 and 5-CGGGATCCCGTCGCGCGCAGGCAAG-3 for GITR, 5-ACTAGTGGAGATGAATTGGGGATTTGC-3, and 5-GGATCCCAATTAGGACTGCTAAGGAg-3 for GITRL. The GITR and GITRL extracellular domain PCR products were digested with BamHI and HindIII or SpeI and BamHI, respectively, prior to ligation into a pCDM8 vector containing a human IgG1-tail cut with BamHI/HindIII or SpeI/BamHI, respectively.
Transfectants and cell lines
C1R cells were subcloned and the same clone was transfected by electroporation (250 V/500 µF) using 15 µg of the vector RSV.5neo or RSV.5neo containing the GITRL open reading frame. Stable transfectants (C1R-neo and C1R-GITRL) were selected in IMDM supplemented with 1.8 mg/ml G418 (PAA Laboratories, Linz, Austria) as described previously (27)
.
The human cell lines K562 (leukemia), U373MG, U87MG, U251MG (glioma), NCCIT, NT2, 2102Ep (germ cell), Colo205, HCT116 (colon), SKMel, (melanoma), LX1 (lung), PC3 (prostate), and MCF-7 (breast) were obtained internally at Eberhard Karls-University Tuebingen.
Preparation of NK cells
NK cells were isolated from peripheral blood by negative selection using NK cell Isolation Kit II and MACS columns (Miltenyi Biotech, Bergisch Gladbach, Germany), yielding a cell population of >90% NK cells as determined by flow cytometry. Polyclonal NK cells were generated by resuspension of nonplastic-adherent PBMC in medium containing 10% FCS, 2 mM L-glutamine, and 50 U/ml IL-2 and incubation with irradiated RPMI8866 feeder cells (ratio 4:1). After 10 days, purity of NK cells was determined by flow cytometry. Experiments were performed when purity was >90%.
Flow cytometry
NK cells were incubated with anti-GITR or mouse IgG1; tumor cells were stained with anti-GITRL, anti-CD40, anti-CD54, anti-EpCAM, anti-HLA class I and II, or isotype control (all 10 µg/ml), followed by anti-mouse-PE (1:100). Where indicated, after incubation with anti-GITR, anti-GITRL, or mouse IgG1 (20 µg/ml), cells were incubated with GITR-Ig, GITRL-Ig, or human IgG1 as control (all 10 µg/ml), followed by a mouse-adsorbed anti-human-PE (1:100). Analysis was performed on a FACScan® (Becton Dickinson, Franklin Lakes, NJ, USA). Specific fluorescence indices (SFI) of GITR stainings were calculated by dividing median fluorescences obtained with GITR mAb by median fluorescences obtained with the isotype control.
Immunohistochemistry
Tumor paraffin sections of 3 µm were mounted on adhesive slides, deparaffinized, rehydrated, and incubated with 5% BSA-PBS to block unspecific protein-protein interactions. Anti-GITRL clone 109114 (1 µg/ml) was incubated at 4°C in a humid chamber overnight. Biotin-labeled rabbit-anti-mouse immunoglobulins and a horseradish peroxidase-streptavidin complex were subsequently applied for 30 min at room temperature. Diamino-benzidin was used as chromogen.
Reverse transcriptase PCR
RT-PCR was performed as described previously (26)
. GITRL primers were 5-GCTGTGGCTTTTTGCTCA-3 and 5-ACCCCAGTATGTATTATTT-3. Primers for 18SrRNA were 5-CGGCTACCACATCCAAGGAA-3 and 5-GCTGGAATTACCGCGGCT-3. The PCR products (expected size 365 bp) for human GITRL or 186 bp for human 18SrRNA, respectively) were separated by electrophoresis on agarose gels and visualized by staining with ethidium bromide
Real-time PCR
Real-time PCR analysis was performed as described previously (27)
. Samples were normalized to 18SrRNA to account for the variability in the initial concentration of the total RNA and conversion efficiency of the reverse transcription reaction. Primers for 18SrRNA were 5-CGGCTACCACATCCAAGGAA-3 and 5-GCTGGAATTACCGCGGCT-3; primers for GITR were 5-GCGCTTTCGGGCCCTGTGCGGCCTG-3 and 5-GACCTGTGTGTACGTCAG-3.
Cytotoxicity assay
Cytotoxicity was analyzed by a standard chromium release assay. Target cells were labeled with 1.85 MBq of Na251CrO4 (Amersham, Freiburg, Germany) for 1 h at 37°C and washed three times. In blocking experiments, mAb were added at 10 µg/ml 30 min prior to the assay. Cells were washed and effector cells were titrated on the target cells and incubated for 4 h at 37°C. Maximum release was determined from target cells lysed in 1% Triton X-100. Percentage of lysis was calculated as follows: 100 x (experimental release–spontaneous release):(maximum release–spontaneous release).
Determination of cytokines
IFN- production of NK cells was analyzed using OptEIATM sets from PharMingen (San Diego, CA, USA); TGF-ß production of tumor cells was determined using the DuoSet ELISA development system from R&D Systems according to the manufacturer’s instructions.
Determination of apoptosis in NK cells
5 x 105 NK cells were cultured alone or on immobilized GITRL-Ig or isotype control for the indicated times. Then apoptosis was determined by resuspension of NK cells in 100 µl binding buffer containing 2 µl Annexin-V-FITC and propidium iodide (PI) and incubation on ice for 15 min, followed by FACS. Measurement of apoptosis was assessed as described using the Nicoletti method (28)
. Briefly, cells were harvested by centrifugation and permeabilized with buffer containing sodium citrate (0.3%), triton-X100 (0.01%), and PI (50 µg/ml) for FACS analysis of the percentage of dead cells. To exclude that potential apoptotic cells had disappeared during the culture, cells were counted at the end of the apoptosis assays.
Determination of NF-B modulation
Nuclear extracts were prepared as described previously (29)
. Protein concentrations of nuclear extracts were determined using a bicinchoninic acid assay (Pierce, Perbio Science, Bonn, Germany). Nuclear extracts (20 µg) were separated by SDS-PAGE and transferred onto nitrocellulose membrane (Schleicher & Schuell, Kassel, Germany). The blots were probed with c-Rel or RelB antibody. Bands were visualized by enhanced chemiluminescence staining (Amersham Bioscience Europe, Freiburg, Germany).
Statistical analysis
Where indicated, results were compared using the Student’s t test. A P value of <0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A). Similar results were found when the cells were analyzed using a GITR-Ig fusion protein (Fig. 1B
and data not shown). Binding of GITR-Ig to GITRL-expressing tumor cells was not reduced by the presence of the two anti-GITRL mAb (clone109101 and 109114), suggesting that these anti-GITRL mAb do not block GITR-GITRL interaction (Fig. 1B
). None of the tumor cell lines used in this study expressed GITR or substantial levels of the Fc receptors CD16, CD32, or CD64 as determined by FACS (data not shown). Expression of GITRL by tumor cell lines was further determined by RT-PCR analysis. In line with the results of the FACS analysis, tumor cell lines with GITRL surface expression expressed mRNA for GITRL, as revealed by the expected products of 365 bp, whereas no GITRL mRNA was detected in C1R-neo, K562, and SKMel cells (Fig. 1C
). To exclude that expression of GITRL was a consequence of long-term in vitro culture of cell lines, we next analyzed the expression of GITRL mRNA in primary gastrointestinal tumors. RT-PCR analysis of patient tumor samples revealed that one of three colon cancers, two of three rectum cancers, and one of three stomach cancers expressed substantial levels of GITRL mRNA (Fig. 1D
). The expression of GITRL on primary tumor tissues was further confirmed by Immunohistochemistry (Fig. 1E
), which revealed that primary human tumors can in fact express GITRL protein.
|
GITRL modulates surface protein expression and TGF-ß release of tumor cells
To determine whether the tumor-expressed GITRL was functionally active, we cultured GITRL-expressing tumor cells for 24 h alone, on immobilized GITR-Ig or human IgG1 as control. While no substantial changes in apoptosis were observed (data not shown), FACS analysis revealed that GITRL stimulation markedly reduced expression of the immunostimulatory molecules CD40 and CD54/ICAM1 and the adhesion molecule EpCAM on the tumor cell surface; HLA class I expression remained unchanged (Fig. 2
A, B). In contrast, with none of the indicated GITRL-negative tumor cells did we observe an effect of the GITR-Ig fusion protein on expression of CD40, CD54, or EpCam. This demonstrates that the effects observed with GITRL-positive cell lines are in fact due to stimulation of GITRL and not due to a potential unspecific effect of the GITR-Ig protein. It is noteworthy that the mock transfectants C1R-neo expressed CD40, CD54, and low levels of HLA class II (but not EpCam or HLA class I; data not shown), while no relevant expression of these molecules was detected on the C1R-GITRL cells (Fig. 2B
).
|
Next, we wanted to evaluate whether stimulation of GITRL also altered release of the immunosuppressive cytokine TGF-ß by tumor cells. We cultured the indicated GITRL-positive and -negative tumor cells on immobilized GITR-Ig or human IgG1 as control for 24 h and analyzed the culture supernatants by ELISA. With the GITRL-positive cell lines (HCT116, MCF7, NCCIT, LX1, and C1R-GITRL transfectants), we observed a substantial and significant (all P<0.05) induction of TGF-ß production by stimulation of GITRL on the tumor cells, while no effect of the GITR-Ig was observed with the GITRL-negative SKMel, PC3, K562 cells, or C1R-neo mock transfectants (Fig. 2C
). This confirmed that the effects of the GITR-Ig fusion protein observed with GITRL-positive cell lines were specifically due to signaling via the GITRL. Thus, GITRL signaling alters the expression of regulatory surface molecules and stimulates production of the immunosuppressive cytokine TGF-ß in tumor cells.
GITR is expressed on NK cells and up-regulated following activation
NK cells were cultured for the indicated times in the presence or absence of 10 ng/ml IL-2 or IL-15, respectively, prior to analysis of GITR mRNA and surface expression. FACS analysis revealed that NK cells constitutively express low levels of GITR at the cell surface, and expression is increased by cytokine treatment. Treatment with IL-15 induced a slightly more pronounced effect compared with IL-2 (SFI 20.1 vs. 18.1 after 24 h and SFI 25.9 vs. 21.1 after 48 h, respectively; Fig. 3
A). Expression peaked after 48 h of incubation with both cytokines. GITR expression could also be detected by staining NK cells with GITRL-Ig and human IgG1 as control, and binding of GITRL-Ig was markedly reduced after preincubation of NK cells with the anti-GITR mAb but not by mouse IgG1 as control (Fig. 3B
). This suggests that the anti-GITR mAb blocks GITR receptor-ligand interaction. Real-time PCR analysis of mRNA induction after cytokine treatment revealed that GITR mRNA increased after 4 h of stimulation with the cytokines (Fig. 3C
). IL-15 induced a more pronounced increase in mRNA expression than IL-2, and mRNA peaked at 
24 h, which is in line with the data obtained from FACS analysis.
|
Tumor-expressed GITRL diminishes NK cell cytotoxicity
Next we analyzed whether GITRL expression by tumor cells influenced the reactivity of human NK cells identified to express GITR. Thus, we performed coculture assays with GITRL-expressing tumor cell lines and both unstimulated and IL-15-activated NK cells, and analyzed NK cell cytotoxicity in chromium release assays. Blocking GITR-GITRL interaction by addition of anti-GITR mAb (10 µg/ml) caused a significant (all P<0.05) increase in NK cell cytotoxicity against the GITRL-expressing tumor cell lines LX1, 2102Ep, and HCT116 as well as the C1R-GITRL transfectants by up to 50%, whereas isotype control had no effect. Although IL-15 pretreatment augmented NK cell cytotoxicity, the inhibitory effect of GITRL was observed both with untreated and IL-15-treated NK cells (Fig. 4
A).
|
To ascertain that the increase of NK cell cytotoxicity in the presence of anti-GITR mAb was in fact due to blocking GITR-GITRL interaction and not due to a potential activation of NK cells by the mAb, we performed cytotoxicity assays with the GITRL-negative K562, SKMel, and PC3 cells as well as the C1R-neo mock transfectants. NK cell cytotoxicity against the GITRL-negative tumor cells was not altered by the presence of anti-GITR mAb, excluding that the anti-GITR mAb agonistically stimulated NK cells (Fig. 4B
). Thus, GITRL expression by tumor cells mediates resistance to NK cell cytotoxicity.
Tumor-expressed GITRL inhibits IFN- release of NK cells
Release of IFN- is a second major mechanism by which NK cells participate in antitumor immunity. To analyze whether tumor-expressed GITRL altered IFN- production of NK cells, we cultured untreated and IL-15-treated NK cells for 24 h alone or in the presence of the GITRL-expressing tumor cells NCCIT, LX1, HCT116, and the C1R-GITRL transfectants. Where indicated, 10 µg/ml of the anti-GITR mAb or isotype control were added, then culture supernatants were analyzed by ELISA (Fig. 5
A). In the absence of tumor cells, untreated NK cells produced very little IFN- whereas prestimulation of NK cells with IL-15 induced release of low levels of IFN-. Addition of the anti-GITR mAb did not induce IFN- production of untreated and IL-15-stimulated NK cells, which indicates that this mAb does not activate NK cells. The presence of tumor cells markedly increased the levels of IFN- both in cocultures with untreated and IL-15-stimulated NK cells. It is noteworthy that the IFN- production of NK cells in response to different tumor cell lines varied substantially both in the presence and absence of IL-15. In assays with IL-15-treated NK cells, production of IFN- was significantly (all P<0.05) increased by addition of anti-GITR mAb, while isotype control had no effect. In contrast, only a minor effect of the presence of anti-GITR mAb on the IFN- production was observed in cocultures with untreated NK cell (Fig. 5A
).
|
To verify that tumor-expressed GITRL reduced NK cell IFN- production and to further exclude a potential agonistic effect of the anti-GITR mAb in the cocultures, untreated or IL-15-stimulated NK cells were incubated with the GITRL-negative tumor cell lines SKMel, PC3, and K562 as well as C1R-neo mock transfectants (Fig. 5B
). Addition of anti-GITR mAb did not alter the IFN- production in cocultures with these GITRL-negative cells, which further ascertained that the mAb does not activate NK cells (Fig. 5B
). Thus, in addition to reducing cellular cytotoxicity, tumor-expressed GITRL also diminishes IFN- production of NK cells.
GITRL does not induce apoptosis, but reduces NF-B activity in NK cells
Since proapoptotic effects of GITR stimulation have been reported (30)
, we investigated whether induction of apoptosis after stimulation of GITR on NK cells was responsible for reduced NK cell reactivity. Untreated and IL-15-stimulated NK cells were incubated alone, on immobilized GITRL-Ig or isotype control for 6 h, 12 h, 24 h, and 48 h, then the percentage of apoptotic NK cells was determined. No relevant differences in the percentage of apoptotic NK cells were observed after stimulation with GITRL-Ig fusion protein (Fig. 6
A, B and data not shown). In addition, NK cells were counted at the end of the assay. At neither time point did we observe relevant differences in the total cell number thus excluding that apoptotic cells had disappeared during the culture (data not shown). Furthermore, we performed cocultures of untreated and IL-15-stimulated NK cells with C1R-GITRL transfectants and C1R neo cells as control to analyze the potential effect of cell surface expressed GITRL on NK cell apoptosis. After cocultures for 12 h, 24 h, and 48 h, we again determined NK cell apoptosis by FACS analysis of Annexin V and PI and by using the Nicoletti assay. Selection of NK cells was performed by CD56 double staining. We observed no difference in the percentage of apoptotic NK cells in cocultures with C1R-GITRL compared with C1R-neo (Fig. 6A, B
and data not shown). This indicates that the inhibitory effects of GITRL on NK cell cytokine production and cellular cytotoxicity are not a consequence of induction of apoptosis in GITR-expressing NK cells.
|
Recently it was reported that GITR stimuli alter activation of the transcription factor NF-B in T cells (31
, 32)
. Therefore, we investigated whether GITR stimulation modulated NF-B activation in NK cells. Untreated and IL-15-activated NK cells were incubated for 24 h alone, on immobilized GITRL-Ig or isotype control. Western blot analysis for c-Rel was performed in nuclear extracts, with activated DC serving as positive control. In IL-15-activated NK cells, we observed a marked reduction of nuclear localized c-Rel by triggering GITR on NK cells (Fig. 6C
). We again analyzed the effect of cell surface-expressed GITRL on NF-B activity in cocultures of untreated and IL-15-activated NK cells with C1R-neo and C1R-GITRL cells. The latter were also used in the presence or absence of blocking anti-GITR mAb. We noted reduced NF-B activity in the presence of CIR-GITRL compared with C1R-neo cells. Addition of anti-GITR, but not isotype control to cocultures of NK cells and C1R-GITRL cells partially restored NF-B activity as determined by Western blot for nuclear localized RelB. These additional results support our findings obtained with immobilized fusion protein and further indicate that the inhibitory effect of GITR on NK cell function might be mediated at least in part by inhibition of signals that activate NF-B.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It is assumed that the development of clinically apparent cancer after a cell-intrinsic oncogenic event is dependent on the interaction of tumor cells with the immune system, a reciprocal process that substantially influences whether transformed cells are eliminated or progress to a life-threatening malignancy. In fact, the induction of tolerance of innate and adaptive immune effector cells to the tumor might be a required factor in tumorigenesis (reviewed in refs. 1
, 33
). Among others, many members of the TNF family have been shown to play an important role in the reciprocal interaction of tumor cells with the immune system (2)
. In this study we addressed the yet unknown role of the TNF family member GITRL and its counterpart GITR in cancer immunoediting in humans. GITRL expression has been detected in mice on DC, monocytes, macrophages, B cells, and osteoclasts and in humans in certain healthy tissues (4
, 5
, 8
, 10
, 11
, 13
, 14
, 20
21
22
, 24)
. The fact that GITRL is constitutively expressed on resting antigen-presenting cells distinguishes it from most other TNF family members, which are not detectable in the resting state and are up-regulated after activation (3)
. We demonstrate for the first time that GITRL is constitutively expressed on human tumor cells. Analysis by FACS revealed expression of GITRL at different densities on several cell lines of epithelial, neuroectodermal, and germ line origin, and GITRL expression was confirmed by RT-PCR analysis of mRNA from these tumor cell lines. To ascertain that GITRL expression on tumor cell lines was not an artifact due to long-term culture of tumor cells in vitro, we investigated primary tumors from patients with gastrointestinal malignancies by RT-PCR and immunohistochemistry, which confirmed that primary human tumors can express GITRL.
Bidirectional signaling has been reported for many ligands of the TNF family, and in mice reverse signaling through GITRL stimulates macrophages and osteoclasts, alters cytokine production of DC, and influences apoptosis (20
21
22
23
, 26)
. We reasoned that the constitutive expression of GITRL on tumor cells could stimulate an as yet unrecognized pathway modulating tumor growth, survival, or tumor cell immunogenicity. In fact, we observed that GITRL signaling mediated a marked down-regulation of the immunoregulatory surface molecules CD40 and CD54 and of the adhesion molecule EpCAM on tumor cells. C1R-GITRL transfectants, in contrast to the mock transfectants C1R-neo, lacked relevant expression of CD40, CD54, and HLA class II even in the absence of GITR-Ig stimulation. This observation and the fact that none of the tumor cell lines used in this study, including C1R-neo and C1R-GITRL, expressed GITR (data not shown) suggest that overexpression of GITRL per se mediates effects, as reported for its receptor and other immunoregulatory molecules like DR3, CD40, NGFR, IL-3 receptor, and ErbB-2 (reviewed in ref. 30
). Since CD40, CD54, and HLA class II mediate activation of immune cells (e.g., ref. 34
,35
), expression and signaling via GITRL might enable tumor cells to avert immune responses and to attenuate the protective antitumor functions of immune effector cells. Expression of EpCAM on tumor cells is assumed to inhibit metastasis (36)
. Therefore, reduction of EpCAM expression due to constitutive GITRL signaling or under selective pressure by GITR-expressing immune cells might enable the tumor cells to spread and hide in niches where they might escape the immune response (33)
. While no induction of apoptosis after stimulation of tumor-expressed GITRL was observed (data not shown), we found that stimulation of GITRL substantially induced the production of TGF-ß by tumor cells. This cytokine has been proposed to function as an autocrine and paracrine factor stimulating tumor cell motility, invasiveness, and metastasis. Furthermore, TGF-ß potently mediates immunosuppression by reducing antigen presentation by DC, inhibiting the activity of IFN-, reducing T cell proliferation, and suppressing NK cell cytotoxicity and IFN- production (reviewed in ref. 33
). There is still controversy as to whether the ligands of the TNF family can signal, especially since surprisingly few data are available regarding the underlying biochemical transduction processes. However, a broad array of experimental data make a convincing case for the existence of reverse signaling via different members of the TNF family, including GITRL (reviewed in ref. 37
). None of the tumor cells used in our study expressed substantial levels of Fc receptors or GITR (data not shown), suggesting that the observed effects of the GITR-Ig were in fact due to reverse signaling via the GITRL, especially since no effect of the GITR-Ig fusion protein was observed with GITRL-negative tumor cell lines. Thus, GITRL expression seems to affect the interaction of tumor cells with the immune system by influencing tumor cell immunogenicity and metastasis and creating an immunosuppressive cytokine microenvironment.
Several recent studies in mice reported that the GITR-GITRL system plays an important role in antitumor immunity, which was attributed to modulation of T cell activity (15
16
17
18
19)
. However, little is known regarding the functional relevance of GITR and GITRL for non-T cells. Furthermore, GITR seems to mediate different effects in mice and humans, since suppression of human Treg, in contrast to their murine counterparts, is not inhibited by GITR (25)
. This led us to investigate the effect of tumor-expressed GITRL on human NK cell functions, since NK cells substantially contribute to cancer immunosurveillance (38
, 39)
. Activation of NK cells requires stimulation of activating surface molecules like, for example, NKG2D, KLRB1, 2B4, CD16, CD96, or natural cytotoxicity receptors whereas inhibitory signals are mediated mainly after recognition of MHC class I molecules by NK cell receptors such as KIR, CD94/NKG2A/C, and LILR, but also by various other inhibitory receptors (e.g., LAIR-1, MAFA, gp49B1, CD66a) (38
, 39)
. Among the proteins that modulate NK cell functions are various members of the TNF/TNFR family (39)
. We demonstrate that GITR is constitutively expressed on NK cells and up-regulated after activation by IL-2 and IL-15. Expression of GITR seems not to be associated with certain NK cell subpopulations, since we observed no differences of GITR expression in CD56brightCD16–, CD56dimCD16+, or CD56+CD8+ NK cells (data not shown). It has been demonstrated that GITR is constitutively expressed on responder T cells and up-regulated after activation, with peak expression after 24–72 h, which is in line with our data regarding GITR expression on NK cells (3)
. Thus, in both NK cells and T cells GITR is constitutively expressed and up-regulated after activation, whereas most other NK receptors are expressed on T cells only on activation. Our findings are in line with the results of a recent study that, while this manuscript was in preparation, reported on GITR expression on NK cells: Hanabuchi and co-workers also observed constitutive expression of GITR on NK cells and up-regulation of expression after cytokine stimulation. As in our study, incubation with IL-15 induced slightly higher levels of GITR expression compared with IL-2 (24)
. However, this difference is marginal and not likely to be functionally relevant. In contrast to the work of Hanabuchi and co-workers, who analyzed the role of GITRL expressed by pDC on NK cell activation, we directly addressed the effect of tumor-expressed GITRL on the reactivity of NK cells. To eradicate developing tumors, NK cells must perform two critical tasks—namely, cytotoxicity and production of IFN-—the latter participating in cancer elimination by inhibiting cellular proliferation and angiogenesis, promoting apoptosis, and stimulating the adaptive immune system (40)
. First, we analyzed the cytotoxicity of NK cells against GITRL-expressing tumor cell lines of different histogenesis. Surprisingly, we observed that blocking GITR-GITRL interaction using an anti-GITR mAb mediated an 50% increase in cytotoxicity, and this was observed with both untreated and cytokine-activated NK cells. In addition, blocking GITR-GITRL interaction also increased INF- production of NK cells in cocultures with GITRL-expressing tumor cells. The levels of IFN- produced by NK cells varied substantially depending on the respective cell line, and it is noteworthy that a substantial increase of IFN- production by GITR-blockade was observed solely in the presence of IL-15, whereas the IFN- levels produced by NK cells in the absence of IL-15 were low and not markedly altered by GITR blockade. It has previously been reported that regulation of NK cell functions differs in resting and activated NK cells according to specific combinations of receptors (39)
. Why the increase of IFN- production due to GITR blockade, in contrast to the effect on cytotoxicity, was dependent on cytokine activation of NK cells remains to be elucidated.
The finding that GITRL expression on tumor cells attenuates NK cell reactivity is seemingly in contrast to the recently reported data for GITR/GITRL in the interaction of pDC and NK cells (24)
. The discrepancy with the findings of Hanabuchi and co-workers could be due in part to differing experimental conditions and reagents. In our experiments, the anti-GITR mAb, but not the anti-GITRL mAb used by Hanabuchi and co-workers, blocked binding of GITRL to its receptor GITR. Addition of anti-GITR mAb to NK cells alone did not induce IFN- production, and cytotoxicity and IFN- production of NK cells in cocultures with GITRL-negative tumor cells were not altered by addition of the anti-GITR mAb, which further ascertained that the anti-GITR mAb had no agonistic function. Together, these results demonstrate that the increase of NK cell reactivity against GITRL-expressing tumor cells mediated by the anti-GITR mAb was in fact due to blocking GITR-GITRL interaction. It cannot be excluded that coculture of GITRL-expressing cells like pDC with GITR-expressing NK cells or with potentially agonistic GITRL mAb for longer periods, as in the work of Hanabuchi et al., might stimulate the GITRL-expressing cells, which in turn could alter NK cell activity.
In T cells, both increased and reduced proliferation as well as pro- and antiapoptotic effects have been reported after GITR stimulation due to activation of TNF receptor-associated factors (TRAF), the proapoptotic protein Siva, and/or the CD95/Fas pathway (reviewed in refs. 30
, 31
). To ascertain that GITR transduced inhibitory signals into NK cells, we directly stimulated the GITR receptor on untreated or IL-15-activated NK cells using our GITRL-Ig fusion protein with human IgG1 as control. In addition, we performed cocultures of NK cells with GITRL-expressing C1R cells or mock transfectants. Subsequently, we determined the percentage of apoptotic NK cells. The fact that we observed no induction of apoptosis in NK cells due to the presence of GITRL suggests that inhibitory rather than proapoptotic signals were responsible for the reduced NK cell reactivity. GITR has been shown to modulate five of the six mammalian TRAF proteins identified so far: TRAF1, 2, 3, 4, and 5 (31
, 32)
. Whereas TRAF1, 2, 4, and 5 activate NF-B, TRAF3 has been shown to antagonize the effects of TRAF2 in NF-B activation. Moreover, GITR can both positively and negatively modulate NF-B, since recent data suggest that GITR-induced NF-B activation is inhibited by TRAF2 (31
, 32)
. This revealed a novel regulatory role for TRAF2 and documented intriguing differences between TRAF2-mediated signaling triggered by GITR and other members of the TNFR superfamily. To analyze whether GITR negatively regulates the activity of NF-B in NK cells, we stimulated GITR on NK cells using immobilized GITRL-Ig protein and, in addition, performed cocultures of NK cells and C1R-GITRL or C1R-neo cells, respectively, followed by analysis of c-Rel and RelB by Western blot. In fact, we found that NF-B activity of NK cells was markedly reduced after GITR stimulation. NF-B activity in cocultures with C1R-GITRL transfectants could be partially restored by addition of blocking GITR mAb whereas isotype control had no effect. Thus, the reduction of NK cell cytotoxicity and IFN- production after GITR triggering might be caused at least in part by diminished NF-B activity, possibly due to association of GITR with inhibitory TRAF molecules.
Our data obtained in the human system are in contrast to available data from mouse models, where GITR activation by agonistic mAb or injection of adenovirus expressing recombinant GITRL into tumors potently stimulated antitumor immunity. The effects of GITR stimulation have in these studies been attributed to enhanced T cell activity, possibly dependent on IFN- and CD178, which has led to the suggestion that agonistic anti-GITR mAb might be instrumental in treating advanced cancers (15
16
17
18
19)
. However, GITR stimulation by the antibody DTA1 used in these studies or using recombinant ligand might not reflect the consequences of GITR interaction with its natural, tumor-expressed ligand in vivo. In addition, the beneficial effects of DTA1 treatment have been suggested to be due at least in part to depletion of Treg (3)
. Furthermore, GITR and GITRL are expressed by multiple cell types, and studies evaluating immune responses in GITR–/– mice so far have not led to a clear picture of the role of GITR in normal physiology. While a costimulatory role for GITR in combination with anti-CD3 has been reported, other investigators describe that T cells from GITR–/– mice are hyperresponsive to immobilized anti-CD3 and that GITR abrogates suppression of murine Treg (e.g., refs. 7
, 8
, 12
). This supports the notion that both activating and inhibitory signals can be mediated by GITR. Similar findings have been reported with another molecule involved in NK cell activation: the immunoreceptor 2B4/CD244 (39)
. It is now assumed that 2B4 causes differential functional outcomes via negative or positive alternative adaptors acting in its pathway (41)
; likewise, GITR may mediate differential effects, possibly due to association with different TRAF molecules.
Our data indicate that in the reciprocal interaction between human tumor cells and NK cells, GITRL expression provides an advantage for tumors since GITRL, in addition to reducing tumor cell immunogenicity, adhesion molecule expression, and inducing production of TGF-ß, directly diminishes cytotoxicity and IFN- production of NK cells by transducing signals via GITR, which may negatively modulate NF-B. Therefore, even though GITR-triggering can costimulate T lymphocytes, and in mice can induce rejection of tumors, our data implicate that, in humans, tumor-expressed GITRL is a mediator of immunosubversion. Additional studies are required to elucidate possible differences between GITRL effects in mice and humans and to delineate the factors that determine whether GITRL signaling is immunostimulatory or immunosuppressive due to its effects (e.g., in T cells vs. tumors and NK cells) before further steps regarding a potential use of GITR/GITRL-modulating reagents in tumor therapy are undertaken.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication November 27, 2006. Accepted for publication February 15, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. E. C. Lutsiak, Y. Tagaya, A. J. Adams, J. Schlom, and H. Sabzevari Tumor-Induced Impairment of TCR Signaling Results in Compromised Functionality of Tumor-Infiltrating Regulatory T Cells J. Immunol., May 1, 2008; 180(9): 5871 - 5881. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Liu, Z. Li, S. P. Mahesh, S. Pantanelli, F. S. Hwang, W. O. Siu, and R. B. Nussenblatt Glucocorticoid-induced Tumor Necrosis Factor Receptor Negatively Regulates Activation of Human Primary Natural Killer (NK) Cells by Blocking Proliferative Signals and Increasing NK Cell Apoptosis J. Biol. Chem., March 28, 2008; 283(13): 8202 - 8210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 |
