Dendritic cell (DC) maturation and function are influenced by the surrounding cytokine milieu. We demonstrate tumor-associated suppression of DCs in stimulating allogeneic and tumor-specific CTL and type 1 (IFN-γ-producing) responses in both CD4- and CD8-positive T cells. DCs from MB49-bearing female mice fail to stimulate proliferative and IFN-γ-producing responses in allogeneic mixed lymphocyte cultures. MB49 also inhibited DC function in stimulating type 1 responses against our tumor-specific antigen, the male antigen, HY. DCs from MB49-bearing male mice were unable to restimulate effective HY-specific CTLs or IFN-γ. Tumor-induced interleukin (IL) 10 was found to be specifically responsible for DC dysfunction in response to antigenic driven maturation. This was demonstrated by restoration of DC function in splenic DCs from MB49-bearing female IL-10 knockout mice (HY disparity), whereas not in MB49-bearing male IL-10 knockout mice (no HY disparity). Finally, any tumor-induced systemic inhibitory effect on bone marrow precursors could be overcome by generation of bone marrow-derived DCs ex vivo. These bone marrow-derived DCs derived from MB49-bearing B6 mice were capable of inducing control levels of proliferation in allogeneic mixed lymphocyte reactions and a type 1 (IFN-γ) cytokine profile. The BM-DCs were also capable of restimulating HY-specific CTL and IFN-γ production. These studies reveal the tumor-associated in vivo effects of IL-10 inhibition on DC function in eliciting a type 1 immune response in both allogeneic and tumor-specific responses.
DCs3 are professional APCs that induce and modulate adaptive immune responses (1). In a normal functioning immune environment, DCs acquire, process, and present antigen in the context of costimulation resulting in the generation of antigen-specific responses, as well as in some cases tolerance (1). Antigen presentation by functional DCs leads to the generation of both humoral and cell-mediated responses, the latter associated particularly with antitumor activity (2, 3). Tumors may directly or indirectly influence the activity of DCs as a means of escaping antitumor immune responses. Tumor-induced cytokines have been found to be involved in suppressing effective antitumor DC function. VEGF has been shown to impair the maturation of DCs, and PGE2 has been shown to bias DC development toward a type 2 phenotype (4, 5, 6). Finally, IL-10, a type 2 cytokine, has also been shown to influence DC recruitment and function in both tumor and nontumor systems (7, 8, 9, 10, 11).
Whereas IL-10 has been associated with both immune stimulation and suppression in tumor models, we and others have shown IL-10 to be involved in tumor immune escape. We were the first to show that IL-10 was produced in biopsies of human melanoma and bladder carcinoma (12). In our murine MB49 bladder cancer model, we have shown that IL-10 is present at the tumor site and inhibits type 1 immune responses against a tumor antigen and nontumor antigen presented at the tumor site (13). Whereas tumor-associated IL-10 has been implicated in the inhibition of type 1 immune responses, the effects of tumor-associated IL-10 on DC function have yet to be defined. In nontumor systems, in vitro IL-10 exposure has been found to inhibit DC generation, maturation, and prevent stimulation of type 1 (IFN-γ) responses to keyhole limpet hemocyanin (8, 14). Transgenic mouse models overexpressing IL-10 have also implicated the involvement of IL-10 in DC dysfunction, whereas retroviral transduction of DCs with IL-10 promoted CTL generation and inhibited tumor progression (10, 15, 16).
Given the known effects of IL-10 on DCs, we have examined whether tumor-associated IL-10 affects DC function in our MB49 model. By using MB49, a tumor model that induces IL-10 secretion, we provide in vivo evidence of tumor-induced IL-10 inhibition on DC function. DCs from tumor-bearing mice are deficient in stimulating both CD4 and CD8 T-cell responses to allogeneic and tumor-specific antigens. Specifically, cell-mediated responses driven by DCs from tumor-bearing mice were inhibited. Furthermore, we have determined that tumor-induced IL-10 is directly responsible for suppressing function of antigenically activated DCs and not DCs from MB49-bearing male IL-10 KO mice, which have no HY antigenic disparity. This is consistent with in vitro data indicating the ability of IL-10 to specifically suppress immature DC maturation with no effect on mature DCs (8). Finally, all of the tumor-induced deficiencies are overcome by ex vivo generation of bone-marrow-derived DCs from tumor-bearing mice.
MATERIALS AND METHODS
Animals and Tumor.
C57BL/6J (B6), C57BL/6-IL10tm1cgn mice (IL-10 KO), and BALB/c (4–6 weeks old) mice were obtained from Jackson Labs (Bar Harbor, ME) and maintained in a HEPA filtered cage system for at least 1 week before use. The MB49 tumor cell line, 12-dimethylbenz(a)anthracene-induced in male C57BL/6 bladder epithelial cells, was provided by Dr. Timothy Ratliff when at Washington University (St. Louis, MO). It was carried in vitro and in vivo in our laboratory (17). MB49 was maintained in complete medium (TCM) composed of RPMI 1640 (Life Technologies, Inc., Rockville, MD) supplemented with 10% FCS, 2 mm l-glutamine, 1 mm Na-pyruvate, 50 IU/ml penicillin/streptomycin, 0.5× MEM amino acids solution, and 100 μm MEM Nonessential Amino Acids Solution (Life Technologies, Inc.).
DC Purification from Spleens.
DCs were isolated as described previously (18). Briefly, spleens were harvested from control and tumor-bearing mice (24 days after s.c. injection of 106 MB49 cells), injected with 1 ml of collagenase D (Boehringer Mannheim, Indianapolis, IN) with a 25-gauge needle, and cut into small pieces, which were incubated at 37°C for 25 min and then passed through a 40 mesh CELLECTOR screen (Thomas Scientific, Swedesboro, NJ). Cells were centrifuged at 200 × g for 10 min, and the cell pellet was resuspended in PBS-0.5% (w/v) BSA at 2.5 × 108 cells/ml. Cells were incubated with 100 μl MACS CD11c Microbeads (Miltenyi Biotec, Auburn, CA) per 1 × 108 cells for 20 min at 4°C, then washed with PBS-0.5% BSA. Enrichment yielded an average of 1.25 × 106 CD11c-labeled cells per spleen after magnetic separation on a LS+ Column (Miltenyi Biotec).
Cells were resuspended in PBS/5% FCS supplemented with 0.1% w/v sodium azide and stained with CD11c, CD80, CD86 (BD PharMingen, San Diego, CA), H-2Db Class I (MHC I), I-Ab (MHC Class II; Beckman Coulter, Miami, FL), or biotinylated DEC-205 (ammonium sulfate-purified HB-290 ascites) specific antibodies. Flow cytometry was conducted on a FACScan and analyzed using CELLQuest software (Becton Dickinson, San Diego, CA). CD11c staining revealed a similar DC enrichment in both MB49-bearing female B6 mice and nontumor-bearing B6 mice. Expression of B7.1, B7.2, and MHC class I and II markers were similar in control and tumor-bearing mice (Fig. 1).
Generation of BM-DCs in Vitro.
DC generation culture was established as described (19). Briefly, bone marrow was flushed from femurs and tibias of either MB49-bearing or nontumor-bearing B6 mice. RBCs were lysed by ammonium chloride lysing buffer, and remaining cells were resuspended at 1 × 106 cells/ml in TCM supplemented with 10 ng/ml GM-CSF and 10 ng/ml IL-4 (Sigma, St. Louis, MO) in a 75-cm2 flask. After 3 days, floating cells were gently removed, and fresh medium was added. On day 7, nonadherent and loosely adherent cells were harvested and enriched on a 14.5% metrizamide-TCM gradient solution.
ELISA/ELISPOT for IFN-γ.
Splenic responder cells (1 × 107) were plated in a 24-well plate with DCs (2 × 105) in a total of 2 ml of TCM +50 μm β-mercaptoethanol. After 3 days, supernatant was removed and assayed for IFN-γ in a ELISA using paired anti-IFN-γ antibodies (BD PharMingen).
ELISPOT analysis was performed as described previously (20). Briefly, MultiScreen 96-well plates (Millipore, Bedford, MA) were coated with 20 μg/ml anti-IFN-γ (BD PharMingen). Plates were washed with PBS-0.25% Tween then blocked with PBS-5% FCS. BALB/c responder cells and irradiated DCs (2 × 104) were added. After 18 h, plates were washed with PBS-0.25% Tween and a final wash of distilled H2O. Biotinylated anti-IFN-γ antibody was added to the plates at a concentration of 4 μg/ml. Plates were incubated for 2 h at room temperature and washed with PBS-0.25% Tween. Streptavidin-alkaline phosphatase (Boehringer Mannheim) was added in 50 μl of PBS-5% FCS and incubated for 2 h at room temperature. The plates were washed with PBS-0.25% Tween with a final wash of PBS and developed with 50 μl of 5-bromo-4-chloro-3-indolyl-phosphate/nitroblue tetrazolium chloride (Boehringer Mannheim). Spots were enumerated using a Stereomaster stereo microscope (Fisher Scientific, Pittsburgh, PA).
[3H]Thymidine Proliferation Assay.
Splenic responders cells harvested as above (2 × 105) were plated (five replicates) in a round-bottomed 96-well plate with irradiated (2500 R) DCs at indicated responder:stimulator ratios in a total of 200 μl TCM +50 μm β-mercaptoethanol. After 4 days of incubation at 37°C, 5% CO2, the cells were pulsed with 1 μCi/well of [3H]thymidine (NEN, Boston, MA) for 18–24 h. Cells were harvested in a Skatron cell harvester (Sunnyvale, CA) onto a glass filtermat and counted in a β-scintillation counter.
51Cr Release Assay.
Female B6 mice were immunized i.p. with 5 × 107 male splenocytes. After 2 weeks, female splenocytes (7 × 106) harvested from the primed mice were restimulated with 1.4 × 105 irradiated DCs in a 24-well plate in a total of 2 ml of TCM +50 μm β-mercaptoethanol. Cultures were maintained for 5 days at 37°C, 5% CO2, after which nonadherent effector cells were harvested. MB49 cells were labeled in 100 μCi of 51Cr (NEN) for 1 h at 37°C, and then washed three times with warm TCM. 51Cr-labeled MB49 target cells (1 × 104) and effector cells were added at known E:T ratios in a total of 200 μl of TCM to a 96-well round-bottomed plate. Plates were incubated for 4 h at 37°C, 5% CO2, and then 100 μl of supernatant was removed, and 51Cr release measured with a gamma counter (Packard Bioscience, Meriden, CT). Percentage of specific lysis was calculated from the formula (experimental release − spontaneous release) × 100/(maximal release in 5% TX-100-spontaneous release).
Our results were expressed as the mean ± SE. Comparisons were done by Student’s t test. Comparisons with a difference of P < 0.05 were considered significant.
DCs from MB49-bearing Mice Manifest Reduced Stimulatory Activity in Primary Allogeneic Mixed Lymphocyte Reaction.
As an initial assessment of the effects of MB49 on DC function, we examined the ability of splenic DCs from control and MB49-bearing mice to stimulate in standard allo-mixed lymphocyte reaction. Female B6 mice were injected s.c. with MB49, and after 24 days, DCs were harvested as described in “Materials and Methods” and used as stimulators with BALB/c spleen cells as responders. DCs enriched from spleens as opposed to DCs grown ex vivo were compared to assess the in vivo effects of MB49 on DC maturation. When used to stimulate proliferation as measured using [H3]thymidine incorporation, DCs from MB49-bearing mice manifested significantly less stimulatory activity (Fig. 2,A). Similarly, DCs from MB49-bearing mice stimulated decreased IFN-γ production by BALB/c splenocytes when compared with control DCs from nontumor-bearing mice as measured in bulk culture (Fig. 2,B) and using IFN-γ ELISPOT analysis (Fig. 2 C). The decreased IFN-γ production indicates a suppression of the normal type 1 response seen in allo-MLC stimulated by DCs (21).
Subset analysis was performed to determine whether reduced DC activity was preferentially associated with CD4 or CD8 T-cell stimulation. Steinbrink et al. (8) has shown that IL-10-exposed DCs induce tolerance in CD4 cells in a nontumor system, whereas Macatonia et al. (14) has shown IL-10 to affect both CD4 and CD8 IFN-γ production when stimulated by DCs. In our study, IFN-γ production was decreased in both CD4 and CD8 T cells when stimulated by DCs from tumor-bearing mice (Fig. 2 B).
MB49 Affects DC Function in a HY-specific Mixed Lymphocyte Reaction.
To assess MB49 impacts on the ability of DCs to stimulate a response to tumor-associated antigen, we examined the ability of DCs from MB49-bearing mice to stimulate responses to the tumor expressed male antigen, HY (reviewed in Ref. 22). We have shown previously that MB49, which was derived from male urothelial cells, expresses HY by reverse transcription-PCR and the susceptibility of MB49 to HY-specific CTLs (13).
Tumor effects on HY-specific DC functional activity were assessed by determining whether DCs from male MB49-bearing mice were capable of restimulating splenocytes from male primed female B6 mice. Female mice were primed in vivo with male splenocytes. After 2 weeks, the female splenocytes were harvested and restimulated in vitro with DCs from either MB49-bearing male B6 mice or normal male control mice. Resultant CTL activity was decreased in effector cells restimulated using DCs from tumor-bearing mice (Fig. 3,A). This clearly correlates with our data indicating decreased HY-specific CTLs in MB49-bearing mice.4 There was also a decrease in HY-specific IFN-γ production with stimulation by DCs from tumor-bearing mice (P < 0.04; Fig. 3 B), which additionally supports the inability of tumor-exposed DCs to generate CTLs and also other type 1-specific immune mechanisms. This indicates a tumor antigen-specific DC dysfunction in generating a type 1 response against MB49.
We next determined whether DCs from MB49-bearing mice were capable of priming an antimale HY response resulting in the generation of HY-CTL. The in vivo functional activity of DCs from control and MB49-bearing male mice was assessed by priming female B6 mice s.c. with 5 × 105 DCs from either male B6 mice bearing MB49 tumors or control male B6 mice. After 2 weeks, splenocytes from the primed females were harvested and restimulated with male splenocytes in mixed lymphocyte reactions. IFN-γ production determined by ELISA and CTL generation measured through a 51Cr release indicated that equivalent antimale responses were generated from both DC populations. Furthermore, both female C57BL/6 mice primed with DCs from tumor-bearing male C57BL/6 and nontumor-bearing control mice were capable of rejecting subsequent MB49 challenge (4 of 4 from each group rejected MB49; data not shown). Whereas this may speak against a defect, cross-priming by responder female APCs, presenting HY derived from the injected male DCs, may be responsible for the inability to distinguish differences in in vivo functional activity of the tumor-bearing mice-derived DCs (23). This will be additionally discussed in the discussion.
DC Alloreactivity Is Restored in MB49-bearing Female IL-10 KO Mice.
We have previously shown MB49 to use an IL-10-mediated tumor immune escape mechanism. Female IL-10 KO B6 mice generate a type 1 response to the tumor-expressed HY antigen and reject MB49 challenge, whereas wild-type female B6 mice succumb to the tumor (13). Having demonstrated a tumor-associated impact on DC function, we set to determine whether tumor-associated IL-10 was responsible for the DC dysfunction. Two populations of DCs have been found recently to migrate to lymphoid tissue (24). There is a steady state migration of nonantigenically stimulated DCs into the lymphoid tissue and a population of DCs that migrates after antigenic stimulation (25). Determination of whether specifically antigen stimulated DCs are impacted is especially relevant to assessment of antitumor responses. The HY antigenic difference between the male and female IL-10 KO system afforded us the opportunity to stringently define in vivo whether MB49-associated IL-10 affected nonantigenically challenged DCs or DCs exposed to HY-mediated maturation.
Initial studies on DC function from MB49-bearing female IL-10 KO mice were conducted to distinguish the impact of tumor-induced IL-10 on DCs exposed to HY antigenic stimulation. To determine the role of IL-10 in the DC dysfunction we repeated the allo-MLC studies described above using splenic DCs harvested from tumor-bearing and nontumor-bearing control female IL-10 KO mice. The use of the IL-10 KO mice in our studies is made possible by our prior demonstration that the tumor-associated IL-10 seen in MB49-bearing mice is host-derived and not produced by MB49 itself (13). DCs were similarly harvested from tumor-bearing or nontumor-bearing controls and used as stimulators in allo-MLC with BALB/c responder splenocytes. The results indicated that DCs from MB49-exposed female IL-10 KO B6 mice were equivalent to DCs from nontumor-bearing female B6 IL-10 KO mice in stimulating proliferation (Fig. 4,A) and IFN-γ production (Fig. 4 B), thus implicating IL-10 as contributing to dysfunction of DC maturation and function.
DC Functional Activity Is Deficient in MB49-bearing Male IL-10 KO Mice.
Assessing the DC function of MB49-bearing male IL-10 KO mice provides the opportunity to reveal the nonantigen-specific effects of IL-10 on naïve DCs. We hypothesize that rejection of MB49 by female B6 mice is mediated via a response to the HY antigen. Similar to female B6 mice, splenocytes from MB49-bearing male B6 mice have been found to generate IL-10. Male IL-10 KO mice offer the opportunity to assess the effects of the MB49 tumor (cytokine) environment on DC function in a system where there is no IL-10 but also no antigenic disparity (HY), thus no apparent stimulation of tumor-specific T-cell responses that could lead to alterations and maturations in DC function via cytokine production by antigen-specific T cells (26). With this in mind, male IL-10 KO mice were injected s.c. with MB49. After 24 days, DCs were purified from either tumor-bearing or nontumor-bearing control mice and used as stimulators against BALB/c responders in an allogeneic mixed lymphocyte reaction as above. DCs from MB49-bearing male IL-10 KO mice stimulated similarly reduced levels of proliferation in BALB/c responders as IL-10 producing cells (Fig. 5,A). Tumor inhibition of DC function was additionally characterized by assessing the DC capability in generating antigen-specific (HY) responses. DCs purified from MB49-bearing male IL-10 KO mice were used to restimulate splenocytes from female B6 mice primed previously in vivo with male splenocytes. Decreased IFN-γ production (Fig. 5,B; although not as dramatic as seen with wild-type MB49-bearing male DC stimulation as seen in Fig. 3,B) and decreased CTL ability (Fig. 5 C) confirmed that in male IL-10 KO mice, there is a similar MB49-mediated dysfunction of DCs as controls. In our model, tumor-associated DC dysfunction is seemingly dependent on IL-10 only in female challenged mice and not in male mice. These results may reflect the specific effects of IL-10 on antigen-dependent DC maturation. Female B6 mice bearing MB49 are exposed to the potential influence of HY-specific T cells and their products, whereas male B6 mice challenged with MB49 have no obvious antigenic stimulus. In the presence or absence of antigenic stimulation, DCs migrate to lymphoid tissues (reviewed in Ref. 27), but maturation and function of these differing DC populations have shown to be dependent on an inflammatory stimuli (26, 28). Therefore, with our results evaluating DCs under different antigenic exposures, we would conclude that IL-10 may be responsible for inhibiting the function of DCs exposed to antigen-specific T-cell responses to tumor antigen, whereas DCs from MB49 challenged male mice, which have not been stimulated, are not affected by IL-10. The DC dysfunction in MB49-challenged male mice may be because of other DC inhibitors. The effect of these DC inhibitors including VEGF and PGE2 (which MB49 does produce)4 may be revealed when the dramatic suppressive effects of IL-10 are excluded (4, 5, 6). Whereas the studies presented here were not designed to definitively assess the effects of IL-10 and the variety of other known and unknown tumor-associated products on DC function, they additionally support the presence of overlapping influences on DC function in a variety of tumor and nontumor systems. This will be additionally discussed in the discussion.
BM-DCs from MB49-bearing Mice Have Normal Stimulatory Activity.
The above-described tumor-associated effects on DC function have all been based on the utilization of splenic DCs, which have matured under the influence of tumor and its associated regulatory factors. Indeed, we have focused on splenic DCs for the purpose of assessing the impact of tumor on function. However, given the potential of DCs in a variety of immunotherapy settings, we have expanded our studies to include examining DCs derived in vitro from bone marrow precursors harvested from tumor-bearing mice. If such ex vivo propagation of DCs overcomes the negative influences of tumor on DC function demonstrated above, it will provide strong support for using ex vivo propagated DCs as opposed to in vivo elicited populations. To determine whether the negative effects of tumor on DC function could be overcome by generating DCs from BM-DCs, bone marrow was extracted from MB49-bearing C57BL/6 mice and cultured in 10 ng/ml GM-CSF and 10 ng/ml IL-4 to generate DCs. These BM-DCs were then assessed for functional activity as above. BM-DCs derived from MB49-bearing female mice manifested control levels of stimulatory capacity as assessed by T-cell proliferation in an allogeneic mixed lymphocyte reaction (Fig. 6,A) and IFN-γ production (Fig. 6,B) demonstrating that, at least in our model, tumor-associated DC dysfunction can be overcome by generating BM-DCs outside of the tumor milieu. BM-DCs from MB49-bearing male mice were assessed for their ability to restimulate HY-specific CTLs. Male BM-DCs were harvested and used to restimulate male specific effectors cells generated by priming female B6 with 5 × 107 male splenocytes i.p. Contrary to what we have described above for splenic DCs, the results indicated that BM-DCs from male tumor-bearing B6 mice were equally efficient in stimulating male-specific CTLs as compared with nontumor-bearing male control BM-DCs (Fig. 6 C) and, thus, whatever single or multiple negative regulation on DC functions are present in tumor-bearing mice, the DC precursors derived from tumor-bearing mice are not effected.
We describe the effects of the growth of MB49 on DC function in stimulating both CD4 and CD8 T-cell responses to allogeneic and tumor-specific antigens. Purified splenic DCs from MB49-bearing mice were dysfunctional in stimulating T-cell proliferation and a type 1 (IFN-γ) response from CD4 and CD8 T cells. This decrease of IFN-γ production was attributed to a decrease in frequency of IFN-γ producing responder T cells after stimulation with allogeneic DCs. MB49 was then shown to decrease HY-specific DC function in generating type 1 (IFN-γ) responses. The tumor influence on DC activity prevented an antigen-specific cell-mediated response against MB49. These tumor-associated effects were mediated through IL-10 in an antigen-dependent manner. DCs from MB49-challenged IL-10 KO female mice were fully competent in stimulating type 1 responses. Finally, whereas DCs maturing in vivo under the influence of the tumor were dysfunctional, bone marrow precursors harvested from tumor-bearing mice and cultured with GM-CSF and IL-4 matured to DCs with normal function demonstrating that the negative tumor influence could be overcome by ex vivo propagation.
The generation of long-lived adaptive immunity to tumors like any antigen is the product of the interaction between APCs and responder T-cell populations, and is dependent on the presence of an appropriate recognizable tumor antigen. Tumors have developed escape mechanisms active at both the induction and effector phases of the response (13, 29, 30). As examples, we have demonstrated using our MB49 model that IL-10, which we and others have shown to be present in a number of tumor types in humans including melanoma and bladder cancer, can significantly suppress antitumor responses shown here to be because of disordered DC antigen presentation function (12, 13). In addition to IL-10, transforming growth factor β has been shown to inhibit IL-2-dependent proliferation of T cells and to down-regulate generation of antitumor immunity (29). VEGF has also been implicated in immunosuppression by suppression of DC function (30). It is our working hypothesis that the understanding of how tumors inhibit the induction of an immune response may be crucial to the design of an effective antitumor immunotherapy. Current strategies for active immunotherapy include DNA plasmid immunization, viral vaccinations, and DC therapy (reviewed in Ref. 31). All of these strategies are dependent on properly functioning APCs. Therefore, disordered APC function associated with tumor growth may have a major limiting effect on the success of a variety of vaccine-based strategies for immunotherapy of tumors (4, 5, 6).
DCs as professional APCs have the capacity to stimulate potent immune responses yet may also induce anergy against tumor antigens (32). Here, we describe splenic DCs, inhibited by tumor-associated IL-10, which are dysfunctional in stimulating proliferation and IFN-γ production in allogeneic and tumor antigen-specific responses. In nontumor systems, similar results of decreased IFN-γ production have been shown with DCs cultured in IL-10 (14). The functional differences of DCs cultured in IL-10 have been attributed to changes in surface molecules. IL-10 was found to inhibit lipopolysaccharide up-regulation of B7.2 in human DCs (33), whereas other studies indicated IL-10 pretreatment of Langerhans cells did not affect MHC class II expression or B7 (34). Elegant experiments by Chakraborty et al. (35) suggest the presence of both a stimulatory and inhibitory population of DCs in humans, whereby autocrine levels of IL-10 may bias toward development of functionally inhibitory DCs. This is paralleled in murine studies, where autocrine levels of IL-10 prevent maturation and functional activity, and where DCs secreting high levels of IL-10 induce tolerance (36, 37). Our laboratory has also compared the DC production of IL-10 from both control and tumor-bearing mice by ELISA and reverse transcription-PCR, and found no differences in IL-10 production.4 The source of MB49-associated IL-10 production from host cells is being investigated and may be similar to that described in clinical systems where the tumor induces host cells to produce IL-10 mediated by PGE2, tumor necrosis factor α, or transforming growth factor β (38, 39, 40). Furthermore, the IL-10-mediated immune suppression only seen in the presence of antigen stimulation may suggest regulatory T cell (Tr1) activity. IL-10 production by CD4 Tr1 cells has been determined to suppress CTL generation and antitumor immunity (41). In contrast, Segal et al. (42) described an IL-10-producing CD4+ T-cell population mediating tumor rejection in a tumor-induced IL-10 glioma model. Regardless of the cellular source, our studies demonstrate signification dysregulation of DC function associated with in vivo tumor growth, and support such IL-10 production as having a major role in tumor immune escape.
A given in the field of immunotherapy is that an appropriate recognizable tumor antigen be available for both the induction and effector arms of the response. We use the male antigen (HY) as our surrogate tumor antigen. HY is a well-studied transplantation antigen capable of mediating the rejection of male tissue in female hosts. It is expressed ubiquitously in male cells with a nonimmunogenic female counterpart (22). Furthermore, priming female B6 mice with male splenocytes is associated with in vivo MB49 rejection and CTL generation.5 Interestingly, the notable loss of a type 1 response demonstrated when restimulating HY-specific CTL and IFN-γ-producing responses with DCs from tumor-bearing male mice was not seen in our priming studies using DCs from MB49-bearing male mice. However, the ability of male DCs from MB49-bearing mice to prime an antimale response in naïve female B6 may be multifactorial and likely complicated by presentation of male DC-derived HY by female (host) APCs, which may account for the generation of antimale CTL and protection against MB49 challenge (23, 43).
In the absence of IL-10, we demonstrated previously that female IL-10 KO mice challenged with MB49 were capable of mounting antigen-specific type 1 immune responses and rejecting the MB49 tumor. We now demonstrate that tumor-induced IL-10 is responsible for the DC inhibition of a type 1 immune response. The in vivo effects of IL-10 on DC function, which we report, are consistent with findings of Sharma et al. (10) using an elegant transgenic model in which IL-10 is overexpressed behind the IL-2 promoter. Ours is the first demonstration of in vivo IL-10-based DC dysfunction secondary to tumor growth. In our model, the IL-10 effects on DC dysfunction are associated with tumor growth, and we believe models the likely effect of IL-10 in human melanoma and bladder cancer (12, 13). The extent of tumor suppression on DC function extends beyond the local tumor environment. This was determined by assessing systemic APC function from splenic DCs. We also found that any tumor-mediated defects on bone marrow precursors could be overcome by ex vivo generation of BM-DCs. These BM-DCs derived from tumor-bearing mice were capable of generating a strong type 1 response and generating CTL capable of lysing MB49. Other groups have also demonstrated BM-DCs from tumor-bearing mice to be effective in tumor immunization, and we hypothesize that our BM-DCs from tumor-bearing mice will also be efficacious even in the presence of tumor-induced inhibitory products (44). In contrast, Shurin et al. (45) demonstrated recently that ex vivo generated BM-DCs from colon adenocarcinoma (producing IL-10)-bearing mice had decreased CD40 expression and IL-12 production. These differences may be because of differing sources of IL-10 and their associated effects on different compartments including the bone marrow.
Our findings described here made in studying B6 female mice bearing the antigen-disparate MB49 tumor (expressing the male HY antigen) clearly demonstrate that the presence of the tumor and, thus, products of the tumor-host interaction result in an IL-10-dependent suppression of DC function. In addition, we show using male IL-10 KO mice that the IL-10-mediated suppression requires the presence of antigenic stimulation to be manifested. Our studies demonstrate that when MB49 is grown in male B6 mice compared with male IL-10 KO, the absence of IL-10 fails to restore DC function. This may be accounted for by the absence in the male setting of antigenic stimulation by HY and its downstream manifestations in generation of help. The male DCs have no obvious antigenic disparity with which to generate an immune response. These male DCs may reflect a population of “nonactivated” DCs or rather a “quiescent” (termed by Shortman and Liu; Ref. 46) population. While DCs from both antigen-challenged and nonantigen-challenged mice migrate to lymphoid tissues, they dramatically differ in maturation and function depending on exposure to antigen and inflammatory stimuli (26, 27, 28, 46). This indicates that tumor-associated IL-10 may more greatly impact “activation” or antigenic maturation of DCs with less impact on unstimulated or quiescent DCs. The non-IL-10-dependent deficiencies in male DCs seen when we use DCs from tumor-bearing male IL-10 KO B6 mice may be accounted for by alternate DC inhibitors such as VEGF and PGE2 (4, 5, 6). VEGF has been demonstrated to impair DC maturation with no demonstration of antigen dependency (47). Whereas these nonspecific DC inhibitors may also be present in tumor-bearing female IL-10 KO mice, IL-10 is clearly the dominant tumor-specific DC suppressive factor. We show that in the absence of IL-10, female DCs from MB49-bearing mice stimulate type 1 responses comparable with nontumor-bearing controls and that the lack of IL-10 results in tumor rejection. We have determined MB49 to inhibit the function of both quiescent and activated DCs. Our evidence that tumor-induced IL-10 selectively inhibits activation of DCs responding to antigenic stimuli and not quiescent DCs reflects the major role of IL-10 in suppressing DC function critical to tumor-specific immune responses. The current understanding of DC function indicates the necessity of activated DCs in generating strong type 1 responses; therefore, the ablation of tumor-induced IL-10 may supercede that of other tumor-induced DC inhibitors (48).
We should note that whereas we and others have demonstrated that the presence of IL-10 can have profound immune-suppressive activity in multiple tumor models as it does in response to defined antigen systems, studies predominantly using IL-10-transfected preclinical models have shown significant enhancement in antitumor responses (49, 50, 51, 52, 53, 54). In fact, when we overexpress IL-10 in MB49 we also demonstrate enhanced antitumor responses (55). One possible explanation for this apparent discrepancy in the actions of IL-10 may be attributed to the high concentration of IL-10 production generated by transfection versus lower physiological levels associated with nontransfected tumors such as studied here. In fact, we see at least a 1–2 log difference in IL-10 level at the tumor site of mice bearing MB40 versus IL-10-transfected MB49.4
Taken together, the above studies demonstrate the pleiotropic effects of IL-10 on the generation and/or regulation of antitumor immune responses. We conclude, supported by our previous published data and that presented here, that the “physiological” production of IL-10 seen in a variety of human tumors and in nontransfection-based models has the potential of significantly impairing the generation of T-cell immune responses to tumor (13). With a better understanding of a mechanism with which tumors modulate the immune system, we can assess the limitations of current immunotherapeutic strategies and provide mechanistic support for next generation approaches.
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.
Supported by USPHS Grant CA-42908 (to E. C. L.) and fellowship 02-2012-CCR-50 from the New Jersey State Commission on Cancer Research (to A. S. Y.).
The abbreviations used are: DC, dendritic cell; VEGF, vascular endothelial growth factor; PGE2, prostaglandin E2; IL, interleukin; KO, knockout; BM, bone marrow-derived; GM-CSF, granulocyte macrophage colony-stimulating factor; ELISPOT, enzyme-linked immunospot; APC, antigen-presenting cell.
B. Halak, unpublished observations.
Manuscript in preparation.