Abstract
Evaluation of immune dysfunction during the tumor-bearing state is a critical issue in combating cancer. In this study, we initially found that IL-6, one of the cachectic factors, suppressed CD4+ T cell–mediated immunity through downregulation of MHC class II by enhanced arginase activity of dendritic cells (DC) in tumor-bearing mice. We demonstrated that administration of Ab against IL-6R (anti–IL-6R mAb) greatly enhanced T cell responses and inhibited the growth of tumor in vivo. We also found that IL-6 upregulated the expression of arginase-1 and arginase activity of DC in vitro. Tumor-infiltrating CD11c+ DC exhibited upregulated mRNA expression of arginase-1 but reduced expression of MHC class II in parallel with the increase in serum IL-6 levels at the late stage in tumor-bearing hosts. However, the administration of anti–IL-6R mAb into tumor-bearing mice inhibited both the downmodulation of MHC class II and the upregulation of arginase-1 mRNA levels in DC. Furthermore, we noted that Nω-hydroxy-L-arginine or L-arginine, an arginase-1 inhibitor, blocked the reduction in MHC class II levels on CD11c+ DC during the tumor-bearing state. Finally, we demonstrated that the administration of Nω-hydroxy-L-arginine at the peritumor site significantly enhanced CD4+ T cell responses and inhibited tumor growth. Thus, IL-6–mediated arginase activation and the subsequent reduction in MHC class II expression on DC appeared to be critical mechanisms for inducing dysfunction of the immune system in the tumor-bearing state. Blockade of the IL-6–arginase cascade is a promising tool to overcome the dysfunction of antitumor immunity in tumor-bearing hosts.
Introduction
It is well accepted that the tumor microenvironment promotes the growth of tumor and inhibits antitumor immune responses by various immunosuppressive tumor-derived factors (1, 2). For instance, the serum levels of IL-10 and TGF-β were highly elevated in the tumor-bearing state, and such strong immunosuppressive cytokines severely block the function of antitumor effector T cells, including CTL (3, 4). Tumor-derived TGF-β is also involved in the generation of Foxp3+CD4+ regulatory T cells, which suppress the antitumor immune responses (5, 6). In addition, it was reported that myeloid-derived suppressor cells (MDSC) were generated in the tumor-bearing state, resulting in the dysfunction of antitumor immunity (7, 8). Therefore, it is of critical importance to elucidate the mechanisms underlying the dysfunction of the immune system in the tumor-bearing state to develop novel strategies to treat tumors.
APC play pivotal roles in immunity protection against tumor or microbes by cooperating with the subsequent activation of CD4+ and CD8+ T cells. Dendritic cells (DC) are the most powerful professional APC involved in T cell–mediated immunity (9). Both CD4+ and CD8+ T cells recognize Ags via TCR by interaction with MHC class II– or MHC class I–bound Ag peptides, respectively, on DC. It was also reported that costimulatory molecules and cytokines are important to strengthen the T cell/DC interaction and to trigger T cell immune responses (9, 10). Therefore, proper regulation of DC function is crucial for inducing subsequent activation of effector T cells in tumor immunity (11, 12).
It was reported that several immunosuppressive factors were induced in APC, including DC and macrophages, during the tumor-bearing state. Induction of IDO in DC inhibited T cell responses by depleting tryptophan and enhanced the production of kynurenine, which was toxic to lymphocytes (13). Macrophages, granulocytes, or MDSC also suppressed T cell–mediated immune responses by regulation of L-arginine metabolism via enzymatic mechanisms involving arginase-1 and NO synthase (8, 14).
Recently, it was reported that IL-6, a pleiotropic cytokine that regulates the growth, differentiation, and survival of a variety of cells (15), significantly inhibited DC maturation via STAT3 activation both in vitro and in vivo (16, 17). It was also demonstrated that some tumor cells produced significant levels of IL-6 (18, 19). In fact, many investigators indicated that serum levels of IL-6 occasionally increased in the tumor-bearing state and in cancer patients (20–23). However, the negative effect of IL-6 on DC function in the tumor-bearing state has remained unclear.
In the present work, we established a tumor-bearing mouse model with CMS-G4 fibrosarcoma cells, which spontaneously produced large amounts of IL-6. Using this IL-6–overproducing tumor-bearing model, we found that IL-6 induced upregulation of arginase-1 expression in DC. In addition, the arginase-1 activation caused the downregulation of MHC class II expression on DC, which resulted in decreased CD4+ T cell–mediated antitumor immunity. Thus, in the present study, we propose a novel mechanism of DC-dependent CD4+ T cell immune dysfunction by IL-6 overproduction during the tumor-bearing state. Our findings also indicate that the IL-6–arginase cascade is a promising target for developing innovative strategies of cancer immunotherapy.
Materials and Methods
Mice and cells
Female wild-type BALB/c mice were obtained from Charles River Japan (Kanagawa, Japan). OVA323–339-specific I-Ad–restricted TCR-transgenic (Tg) mice (DO11.10) were kindly donated by Dr. K.M. Murphy (Washington University School of Medicine, St. Louis, MO). Il-6−/− mice were donated by Dr. Y. Iwakura (University of Tokyo, Tokyo, Japan). Rag2−/− mice were donated by Dr. M. Ito (Central Institute for Experimental Animals, Kawasaki, Japan). All mice were maintained in specific pathogen–free conditions according to the guidelines of the animal department at Hokkaido University and were used at 6–8 wk of age. GM-CSF–producing CHO cells were a gift from Dr. T. Sudo (Toray Industries, Tokyo, Japan). The fibrosarcoma cell line CMS-G4 was generated by intradermal injection of methylcholanthrene into wild-type mice (24). CMC-6KO was generated by the same method from Il-6−/− mice. CMC–6KO–Mock and CMC–6KO–IL-6 cells were prepared by retrovirus infection system with the pMX–IRES–GFP vector, which was kindly donated by Dr. T. Kitamura (The University of Tokyo). IL-6 cDNA was constructed by reverse transcription of total RNA from LPS-stimulated bone marrow–derived DC (BMDC) generated from wild-type mice and inserted into pMX–IRES–GFP. Transduction of pMX–IRES–GFP (CMC–6KO–Mock) or pMX–IL-6–IRES–GFP (CMC–6KO–IL-6) into CMC-6KO cells was performed using the same method described previously (16). GFP+ cells were sorted by FACSAria (BD Biosciences, San Diego, CA) and used in the experiments.
Cytokines, Abs, and chemicals
Recombinant murine GM-CSF and IFN-γ were purchased from PeproTech (London, U.K.). Fluorescence-conjugated anti-CD11c (HL3), anti–IL-4 (11B11), anti–IFN-γ (XMG1.2), anti–H-2Kd (AF6-88.5), and anti–I-Ad (AF6-120.1) mAbs, as well as purified anti-CD3ε (145-2C11), anti-CD28 (37.51), anti–I-A/I-E (M5/114), and purified anti–arginase-1 (clone 19) mAbs, were purchased from BD Biosciences. Purified rat anti-mouse IL-6R mAb (MR16-1) and isotype-control rat IgG1 Ab were kindly provided by Chugai Pharmaceutical (Shizuoka, Japan). Nω-hydroxy-L-arginine (nor-NOHA) was purchased from Calbiochem (San Diego, CA), and L-arginine was purchased from Sigma-Aldrich (Tokyo, Japan).
Generation of BMDC
Bone marrow cells were prepared from wild-type mice in the presence of murine GM-CSF (20 ng/ml) or culture supernatant from GM-CSF–producing CHO cells for 6 d, as described previously (16, 25). Loosely adherent clustering cells were harvested on days 6–8, and CD11c+ DCs were isolated using an IMag Cell Separation System with anti-CD11c mAb–bound beads or FACSAria (BD Biosciences). The purity of CD11c+ DC was >95%, and they were used in the experiments.
Tumor-bearing mice model
CMS-G4 cells (2 × 106) were inoculated intradermally (i.d.) into wild-type mice, and the tumor size was measured using micrometer calipers. In some experiments, CMS-G4 cells (2 × 106) were inoculated i.d. into RAG2−/− mice. The volume of the tumor was calculated using the following formula: volume (mm3) = 0.2 × (length [mm] × width [mm]) × (height [mm])2. In the therapeutic experiments, wild-type mice were inoculated i.d. with CMS-G4 cells and treated with anti–IL-6R mAb or isotype control rat IgG1 (250 μg) every 3 d or with nor-NOHA (20 mg/kg) or PBS every 2 d. In the i.p. tumor model, CMS-G4 cells (2 × 106) were inoculated i.p. into wild-type mice (day 0), and nor-NOHA (20 mg/kg) or L-arginine (500 mg/kg) was injected i.p. at days −1 and 0 and every 2 d thereafter. CMC-6KO, CMC–6KO–Mock, or CMC–6KO–IL-6 cells were used in some experiments.
Flow cytometry
CMS-G4 cells (2 × 106) were injected i.p. into wild-type mice, and peritoneal exudate cells were collected on day 7. Peritoneal exudate cells or BMDC were stained with fluorescence-conjugated mAbs after purified anti-FcγR mAb treatment. Data were acquired using a FACSCalibur (BD Biosciences) and analyzed with CellQuest software (BD Biosciences).
Cell sorting
For CD11c+ DC isolation, CMS-G4 cells (2 × 106) were injected i.p. into wild-type BALB/c mice, and peritoneal exudate cells were collected on day 7. Peritoneal exudate cells were stained with fluorescence-conjugated anti-CD11c mAb after treatment with purified anti-FcγR mAb. CD11c+ DC were isolated by FACSAria (BD Biosciences). For CD4+ and CD8+ T cell isolation, CMS-G4 cells (2 × 106) were inoculated i.d. into wild-type mice, and spleen cells were collected on day 28. CD4+ and CD8+ T cells were isolated from spleen cells using the IMag system (BD Biosciences) with anti-CD4 or anti-CD8 mAb-bound beads, respectively. Purity of the sorted cells was >95%.
ELISA
Serum IL-6, IL-4, and IL-10 levels were determined using OptEIA Mouse IL-6, IL-4, and IL-10 Kits (BD Biosciences), respectively, according to manufacturer’s protocol.
For evaluation of IFN-γ production by activated T cells, BMDC and CD4+ T cells isolated from DO11.10 mice were cocultured or not with OVA323–339 peptide (0.01 μg/ml) for 48 h. In other experiments, spleen cells were stimulated with soluble anti-CD3ε mAb (2 μg/ml) or Con A (2.5 μg/ml) for 24 h, and CD4+ and CD8+ T cells were stimulated with plate-bound anti-CD3ε mAb and anti-CD28 mAb for 48 h. IFN-γ levels in culture supernatants were determined using the OptEIA Mouse IFN-γ Kit (BD Biosciences).
Determination of arginase activity
Arginase activity was measured as described by Corraliza et al. (26). Briefly, cells (1 × 105) were lysed with lysis buffer for 30 min, and MnCl2 was added. After heat activation of the arginase, L-arginine was hydrolyzed by incubating the lysates at 37°C for 60 min. The reaction was stopped with H2SO4/H3PO4/H2O (1:3:7). The urea concentration was measured after the addition of α-isonitrosopropiophenone, followed by heating at 100°C for 20–40 min. Data are given as milliunit of arginase/106 cells, where 1 U arginase is defined as the amount of enzyme that catalyzes the formation of 1 μg urea/min.
PCR
Total RNA was extracted from cells using an ISOGEN RNA extraction kit (Nippongene, Toyama, Japan), according to the manufacturer’s instructions. First-strand cDNAs were synthesized using Superscript II RNaseH-Reverse Transcriptase (Invitrogen, Carlsbad, CA) and amplified with a thermal cycler system (Perkin Elmer) using gene-specific primer pairs as follows: mouse β-actin, 5′-GTGATGGTGGGAATGGGTCAG-3′ and 5′-TTTGATGTCACGCACGATTTCC-3′ and mouse arginase-1, 5′-CAGAAGAATGGAAGAGTCAG-3′ and 5′-CAGATATGCAGGGAGTCACC-3′. After amplification, PCR products with appropriate size were separated by electrophoresis on 1% agarose gels containing ethidium bromide and visualized by UV light illumination. In some experiments, genes for murine MHC class I, MHC class II, and β-actin were amplified using a thermal cycler (LightCycler; Roche, Indianapolis, IN). The following primer sequences were used in this study: MHC class I (forward: 5′-CCGCAGAACTCAGAAGTCG-3′, reverse: 5′-GAAATACCTCAGCGAGTGTGG-3′, Roche Universal probe: #18), MHC class II (forward: 5′-TGGAGGTGAAGACGACATTG-3′, reverse: 5′-CTCATCACCATCAAATTCAAATG-3′, Roche Universal probe: #80), and β-actin (forward: 5′-AGCCATGTACGTAGCCATCCA-3′, reverse: 5′-TCTCCGGAGTCCATCACAATG-3′, and probe: 5′-TGTCCCTGTATGCCTCTGGTCGTACCA-3′). Sample signals were normalized to the housekeeping gene β-actin according to the ΔΔCt method: ΔCt = ΔCtsample − ΔCtreference. Relative mRNA expression levels against the control sample were then calculated for each sample.
Western blotting
BMDC were stimulated with IL-6 (50 ng/ml) for 48 h, and equal numbers of cells were lysed with lysis buffer containing 1% Triton X-100 in the presence of protease inhibitors. The cell lysates were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore). Membranes were blocked with Tween 20–PBS containing 5% skim milk and probed with arginase-1 and β-actin Abs. Membranes were washed and incubated with a secondary Ab conjugated with peroxidase. The protein levels were detected using an LAS-1000 system (GE Healthcare) with ECL (Perkin Elmer). In some experiments, SDS-stable MHC class II αβ-dimer levels of BMDC generated under normal conditions, arginine-free conditions, or arginine-free conditions with 2.5 mM L-arginine were evaluated, as described previously (16).
Statistical analyses
In vitro experiments were repeated at least three to five times. In vivo experiments, consisting of 5–10 mice in each group, were performed independently two or three times. Single representative experiments are shown in the figures. Mean values and SD were calculated for each experiment's data and are shown in the figures. Significant differences in the results were determined using the two-sided Student t test, and p < 0.05 was considered statistically significant.
Results
Increased serum IL-6 in tumor-bearing mice inhibits T cell responses and promotes tumor growth
IL-6 is reported to be one of the major cachectic factors during the tumor-bearing state (27). To investigate the function of IL-6 in the tumor-bearing host, we established a mouse model with CMS-G4 fibrosarcoma cells, which spontaneously produce large amounts of IL-6, and confirmed that the serum IL-6 levels increased significantly (24–30-fold) in the tumor-bearing mice for 1–2 wk (Fig. 1A). The serum IL-10 and IL-4 levels did not increase in the studied mice (Fig. 1B, 1C). To address the effect of the IL-6–signaling cascade on immune responses, we injected the tumor-inoculated mice with i.v. mAb against IL-6R (anti–IL-6R mAb). After stimulation with anti-CD3ε mAb or Con A, T cell–mediated immune responses were reduced significantly (67.5 and 55%, respectively) in the tumor-bearing mice compared with normal mice (Fig. 1D). To further investigate the effect of IL-6 on DC function, CD11c+ DC were isolated from the CMS-G4–bearing mice treated with control IgG or anti–IL-6R mAb and then cocultured with Ag-specific CD4+ T cells in the presence of Ag peptide. As a result, DC from anti–IL-6R mAb–injected mice significantly increased IFN-γ production by CD4+ T cells (3.1- or 2.2-fold) compared with DC from control mice after anti-CD3 or Con A stimulation, indicating that blockade of the IL-6–signaling pathway significantly increased the ability of DC to activate Ag-specific CD4+ T cell responses in the tumor-bearing host (Fig. 1E). We further examined the effect of anti–IL-6R mAb injection in vivo on the growth of tumor. The blockade of IL-6 signaling with anti–IL-6R mAb significantly reduced (by 60%) the growth of tumor at 32 d after tumor inoculation (Fig. 1F). CMS-G4 fibrosarcoma cells did not express IL-6R on their cell surface (data not shown), indicating that anti–IL-6R mAb might not directly affect the tumor cells. These findings indicated that IL-6 was actually involved in tumor-escape mechanisms through the downmodulation of type 1 T cell–immune responses because of defects in DC function.
Increased serum IL-6 in tumor-bearing mice inhibits T cell responses and promotes tumor growth. (A–C) CMS-G4 fibrosarcoma cells (2 × 106) were inoculated i.d. into wild-type BALB/c mice. One and two weeks after the inoculation, serum was collected from the mice to determine IL-6, IL-4, and IL-10 levels by ELISA. Data are mean ± SD (n = 5). (D) Anti–IL-6R mAb or rat IgG1 isotype control (250 μg) was injected i.v. every 3 d after the tumor inoculation. After 28 d, spleen cells (5 × 105) were stimulated with anti-CD3ε mAb or Con A for 36 h in vitro. IFN-γ levels in the culture supernatants were determined by ELISA. Data are mean ± SD (n = 5). (E) CMS-G4 cells (2 × 106) were inoculated i.p. into wild-type mice, and mice were treated with anti–IL-6R mAb or rat IgG1 isotype control (250 μg/mouse) on days 0, 2, 4, and 6 after tumor inoculation. After 7 d, CD11c+ DC (1 × 104) collected from the peritoneal cavity of tumor-bearing mice were cocultured with CD4+ T cells (1 × 105) from DO11.10 Tg mice in the presence of OVA class II peptide (0, 0.25, or 1 μM) for 24 h in vitro. IFN-γ levels in the culture supernatants were determined by ELISA. Data are mean ± SD (n = 5). (F) CMS-G4 cells (2 × 106) were inoculated i.d. into wild-type mice, and mice were treated with anti–IL-6R mAb or rat IgG1 isotype control (250 μg/mouse) every 3 d. The tumor growth was evaluated by measuring its volume. Data are mean ± SD (n = 5). *p < 0.05, Student t test.
IL-6 enhances the expression and activity of arginase-1 in DC in vitro
We next asked what molecular mechanisms were involved in developing IL-6–mediated immunosuppressive tumor escape. Previous studies reported that the IL-6–STAT3–signaling pathway inhibited the maturation and activation of DC both in vitro and in vivo (16, 17). We investigated the effect of IL-6 on the gene expression of immunosuppressive factors in DC. As a result, we found that the arginase-1 mRNA level, but not the IL-10 or TGF-β mRNA level, was upregulated at 3 h after IL-6 treatment (Fig. 2A, data not shown). We also confirmed that arginase-1 protein levels increased remarkably in DC at 48 h after stimulation with IL-6 (Fig. 2B). Furthermore, arginase enzymatic activity was significantly enhanced (2.2-fold) in the IL-6–treated DC compared with untreated DC (Fig. 2C). These results clearly show that the IL-6–signaling cascade enhances arginase-1 expression and enzymatic activity in DC, suggesting that IL-6–induced upregulation of arginase-1 in DC might be one of the mechanisms for tumor immune evasion. This hypothesis was supported by previous findings that L-arginine consumption resulted in immunosuppression (8).
IL-6 enhances the expression and activity of arginase-1 in DC in vitro. (A) BMDC were cultured or not with IL-6 (50 ng/ml) for 3 h. Arginase-1 mRNA expression levels were determined by RT-PCR. (B) BMDC were cultured or not with IL-6 (50 ng/ml) for 48 h. Protein levels of arginase-1 were determined by Western blot analysis. (C) BMDC were cultured or not with IL-6 (50 ng/ml) for 48 h. Arginase-1 activity was determined as described in Materials and Methods. Three independent experiments were performed, and representative results are shown. *p < 0.05, Student t test.
IL-6–arginase–signaling cascade regulates MHC class II expression levels of DC in tumor-bearing mice
To evaluate the critical role of IL-6–induced arginase-1 in DC in the tumor-bearing host, we examined whether upregulation of arginase-1 in DC was actually induced by IL-6 during tumor growth in vivo. As shown in Fig. 3A, CD11c+ DC infiltrated into the tumor-growing peritoneal cavity expressed high levels of arginase-1. However, their enhanced expression of arginase-1 was greatly inhibited by blocking IL-6 signaling with anti–IL-6R mAb in vivo. We also showed that injection of CMS-G4 cells into Il-6−/− mice induced arginase-1 expression, and the expression was significantly blocked (by 68.5%) after treatment with anti–IL-6R mAb (Fig. 3B). Moreover, to precisely evaluate whether tumor-derived IL-6 increases arginase-1 expression, we prepared Mock- or IL-6–transduced CMC-6KO cells, IL-6–deficient squamous cell carcinoma cells, and demonstrated that arginase-1 expression levels on CD11c+ DC were higher in the mice injected with IL-6–transduced CMC-6KO cells than Mock–CMC–6KO cells (Fig. 3C, 3D). These data clearly indicated that tumor-derived IL-6 was involved in inducing arginase-1 expression on DC. In contrast, we confirmed that arginase-1 expression levels were reduced in CMS-G4–bearing Il-6−/− mice compared with wild-type mice (Fig. 3E). Taken together, we demonstrated that IL-6 derived from both tumor cells and host cells contributed to arginase-1 expression on DC in the tumor microenvironment.
The inhibition of arginase-1 induces the restoration of MHC class II expression on DC in the tumor-bearing state. Wild-type (A) or Il-6−/− (B) BALB/c mice were inoculated i.p. with CMS-G4 cells (2 × 106) and treated with anti–IL-6R mAb or rat IgG1 isotype control (250 μg/mouse) every 3 d. After 1 wk, peritoneal exudate cells were collected. CD11c+ cells were isolated from the peritoneal exudate cells, and their expression of arginase-1 mRNA was determined by RT-PCR. Three independent experiments were performed, and representative results are shown. (C) CMC–6KO–Mock and CMC–6KO–IL-6 cells (5 × 104) were cultured in 96-well plates for 24 h, and IL-6 production levels in the culture supernatants were measured by ELISA. Mean ± SD (n = 3) are shown. (D) Il-6−/− mice were inoculated i.p. with CMC–6KO–Mock or CMC–6KO–IL-6 cells (2 × 106) and treated with anti–IL-6R mAb or rat IgG1 isotype control (250 μg/mouse) every 3 d. After 1 wk, peritoneal exudate cells were collected. CD11c+ cells were isolated from the peritoneal exudate cells, and their expression levels of arginase-1 mRNA were determined by RT-PCR. Three independent experiments were performed, and representative results are shown. (E) CMS-G4 cells (2 × 106) were inoculated i.p. into wild-type or Il-6−/− mice. After 1 wk, CD11c+ cells were isolated from the peritoneal exudate cells, and their expression levels of arginase-1 mRNA were determined by RT-PCR. Three independent experiments were performed, and representative results are shown. (F) The expression levels of MHC class II on CD11c+ cells from the peritoneal exudate were determined by flow cytometry. Results are mean fluorescence intensity (MFI); representative data (mean ± SD; n = 5) are shown. (G and H) Wild-type mice were inoculated i.p. with CMS-G4 cells (2 × 106) and treated with PBS, nor-NOHA (G: 20 mg/kg/2 d), or L-arginine (H: 500 mg/kg/2 d). After 1 wk, the expression levels of MHC class II on peritoneal exudate CD11c+ cells were determined by flow cytometry. Results are MFI; representative data (mean ± SD; n = 5) are shown. *p < 0.01, Student t test.
We also investigated the effect of anti–IL-6 mAb on the expression levels of MHC molecules on CD11c+ DC infiltrated into the tumor-growing peritoneal cavity. MHC class II expression levels were reduced significantly (by 31.4%) on CD11c+ DC of tumor-bearing mice compared with those from normal mice (Fig. 3F). The reduction was restored by the administration of anti–IL-6R mAb. We also confirmed the IL-6–dependent arginase-1 expression and MHC II downregulation in the mice bearing other IL-6–producing tumor cells (Supplemental Fig. 1A–D). Moreover, the reduction in MHC class II expression levels was restored by the injection of arginase inhibitor, nor-NOHA (Fig. 3G), or L-arginine (Fig. 3H) into the peritoneal cavity of tumor-bearing mice. MHC class I expression levels on CD11c+ DC were not altered in tumor-bearing mice, even when they injected with anti–IL-6R mAb, nor-NOHA, or L-arginine (data not shown). These findings suggested that IL-6 signaling induced arginase-1 expression in CD11c+ DC, which caused the downregulation of MHC class II expression on DC in the tumor-bearing host.
L-arginine is required for MHC class II expression on DC and DC-dependent T cell stimulation
We next investigated the effect of L-arginine on DC function in vitro. Flow cytometric analysis revealed that surface MHC class I levels were not affected, whereas MHC class II levels were reduced significantly (by 70.8%) in the absence of L-arginine (Fig. 4A). We confirmed that CD11c expression levels on DC were not altered, even in the absence of L-arginine, indicating that this reduction in MHC class II was not the result of a trivial toxic effect (data not shown). We further performed immunoblot analysis to clarify the influence of arginine on the expression levels of MHC class II molecules. As a result, protein levels of MHC class II αβ-dimers of DC decreased markedly in the L-arginine-free condition (Fig. 4B). In addition, quantitative RT-PCR analysis revealed that MHC class II expression levels were reduced significantly (by 95%) in DC cultured in arginine-free medium compared with 1 or 10 mM of arginine-containing medium, whereas MHC class I expression on DC was not affected (Fig. 4C).
L-arginine starvation induces downregulation of MHC class II, but not MHC class I, on DC. (A) BMDC were generated in three media: complete RPMI 1640 (Normal), RPMI 1640 medium without L-arginine (Arg free), or RPMI 1640 medium without L-arginine + L-arginine (2.5 mM) (L-Arg). The expression levels of MHC class II and MHC class I on CD11c+ cells were determined by flow cytometry. Results are mean fluorescence intensity (MFI); representative data (means ± SD; n = 5) are shown. (B) BMDC were generated in normal Arg-free RPMI with or without 2.5 mM L-arginine. MHC class II αβ dimer levels of each BMDC group were determined by Western blot analysis. Representative results of three independent experiments are shown. (C) BMDC were generated in no L-arginine–RPMI 1640 with 0, 1, 10 mM L-arginine. mRNA expression levels of MHC class II and MHC class I were determined by quantitative PCR. Representative results from three independent experiments are shown. (D) CD4+ T cells prepared from DO11.10 Tg mice were cocultured with BMDC for 48 h in the presence or absence of OVA class II peptide (0.01 μg/ml). IFN-γ levels in the culture supernatants were measured by ELISA. Mean ± SD (n = 5) are shown. *p < 0.01, Student t test.
To determine the Ag-presentation ability of DC under the L-arginine–free condition, DC were cultured with CD4+ T cells, which were obtained from OVA323–339 peptide–specific TCR-Tg (DO11.10) mice. IFN-γ production by DO11.10 CD4+ T cells with OVA peptide was significantly reduced in DC cultured under L-arginine–free conditions. However, OVA peptide–induced IFN-γ production by DO11.10 CD4+ T cells was enhanced significantly (2.9- or 5.4-fold) in the presence of DC cultured with 1 or 10 mM of L-arginine, respectively (Fig. 4D). In contrast, downregulation of MHC class I–dependent T cell responses was not observed (data not shown). These data suggested that L-arginine appeared to be required for MHC class II expression on DC, although the concentration of arginine (1–10 mM) used in the experiments was 10–100 times higher than the physiologic concentration of arginine found in serum. In addition, under L-arginine–deficient conditions, DC revealed less Ag-presenting ability for CD4+ T cells.
Arginase inhibitor restored CD4+ T cell–mediated immune responses through upregulation of MHC class II expression on DC in vivo
Finally, we examined the effect of the arginase inhibitor nor-NOHA on T cell–immune responses in tumor-bearing mice and on tumor growth. We confirmed that the expression of arginase-1 mRNA and protein increased in tumor-infiltrating leukocytes (TIL) (data not shown). The peritumoral injection of nor-NOHA significantly reduced tumor growth (by 72.4%) on day 30 after tumor inoculation (Fig. 5A). We also found that MHC class II expression levels were enhanced significantly (2-fold) on CD11c+ DC of TIL in nor-NOHA–treated tumor-bearing mice compared with those in untreated tumor-bearing mice (Fig. 5B). However, no significant changes were observed with regard to the expression of MHC class I. We further examined the effect of in vivo nor-NOHA injection on T cell–mediated immune responses. After stimulation with anti-CD3ε mAb, IFN-γ production by spleen cells increased significantly (2.5-fold) in nor-NOHA–treated tumor-bearing mice compared with untreated control tumor-bearing mice (Fig. 5C). Moreover, both CD4+ T cells and CD8+ T cells, especially CD4+ T cells isolated from spleen cells of nor-NOHA–treated tumor-bearing mice, exhibited higher IFN-γ production compared with those from untreated control tumor-bearing mice (Fig. 5D, 5E). To investigate the involvement of T cell responses in nor-NOHA–induced antitumor effects, we evaluated tumor growth in Rag2−/− mice. We found that nor-NOHA treatment did not inhibit tumor growth in these mice (Fig. 5F). In addition, we confirmed that IFN-γ treatment inhibited the growth of CMS-G4 tumor cells and increased the expression levels of MHC class I, but not MHC class II, in vitro (Fig. 5G, 5H). Therefore, IFN-γ production by T cells induced after nor-NOHA treatment is involved in antitumor effects through direct tumor growth inhibition or MHC class I upregulation of tumor cells, which facilitate tumor susceptibility to CTL-mediated cytotoxicity. Thus, we concluded that the arginase inhibitor nor-NOHA prevented tumor growth by recovering downmodulated MHC class II expression on DC in tumor-bearing mice, which might induce the preferable CD4+ T cell–mediated immune responses, even in tumor-bearing hosts.
Administration of the arginase-1 inhibitor nor-NOHA in vivo rescues T cell dysfunction and inhibits tumor growth. (A) Wild-type BALB/c mice were inoculated i.p. with CMS-G4 cells (2 × 106) and treated with nor-NOHA (20 mg/kg) at the peritumor site every 2 d. Tumor growth was evaluated by measuring the volume. Mean ± SD (n = 5) are shown. (B) TIL were collected 4 wk after tumor inoculation. mRNA expression levels of MHC class II on CD11c+ cells in TIL were determined by flow cytometry. Results are mean fluorescence intensity (MFI); representative data (mean ± SD; n = 5) are shown. (C) Four weeks after tumor inoculation, spleen cells were stimulated with anti-CD3ε mAb for 24 h. IFN-γ levels in the culture supernatants were measured by ELISA. Mean ± SD (n = 3) are shown. Four weeks after the tumor inoculation, CD4+ T cells (D) or CD8+ T cells (E) were isolated from spleen cells and stimulated with plate-bound anti-CD3ε mAb and anti-CD28 mAb for 48 h. IFN-γ levels in the culture supernatants were measured by ELISA. Mean ± SD (n = 3) are shown. (F) Rag2−/− BALB/c mice were inoculated i.p. with CMS-G4 cells (2 × 106) and treated with nor-NOHA (20 mg/kg) at the peritumor site every 2 d. Tumor growth was evaluated by measuring the volume. Mean ± SD (n = 5) are shown. (G and H) CMS-G4 cells were cultured in the presence of IFN-γ (0, 0.1, 1, or 10 μM) for 24 h. (G) Cell growth was evaluated by MTT assay. Mean ± SD (n = 3) are shown. (H) Surface expression levels of MHC class I and MHC class II were evaluated by flow cytometry. Representative results of three independent experiments are shown. *p < 0.01, Student t test.
Discussion
We demonstrate in this study that blockade of IL-6 signaling by anti–IL-6R mAb exerts an antitumor effect through the downmodulation of arginase-1 activity and upregulation of MHC class II expression on tumor-associated CD11c+ DC. It is considered essential to understand a strong immunosuppressive mechanism in the tumor microenvironment in order to develop a novel tumor immunotherapy (1). Our findings clearly point to a novel tumor immune-evasion mechanism in tumor-bearing hosts. Namely, we propose that the IL-6–arginase–signaling cascade is involved in the dysfunction of DC-dependent CD4+ Th cell activation in tumor-bearing mice. Moreover, a powerful tool for IL-6–signaling blockade, anti–IL-6R mAb that was developed by Kishimoto and colleagues (28–30), has the ability to improve the immune dysfunction of DC and MHC class II–dependent Th cell responses in tumor-bearing mice, which results in the inhibition of tumor growth (Fig. 1).
DC are the most powerful professional APC bridging innate and acquired immunity (9, 10). Ag-processed DC play a pivotal role in triggering the activation of CD4+ T cells or CD8+ T cells through antigenic peptide-bound MHC class II/TCR or Ag-bound MHC class I/TCR interactions. In addition to type 1 cytokines derived from DC1 and Th1 cells (31), it is well accepted that DC/CD4+ T cell interactions via CD40/CD40L, rather than DC/CD8+ T cell interactions, are required to induce full activation of CD8+ CTL (32–34). Therefore, the potentiation of DC-dependent CD4+ T cell activation is important to induce protective immunity in hosts experiencing strong immunosuppression as the result of infectious disease or tumor burden (35, 36). Indeed, we previously proposed that the introduction of Th1-dominant immunity is critical for inducing fully activated CTL and maintaining memory CTL, which are essential for inducing the complete cure of tumor-bearing mice (37–39).
It was reported in both mouse and human systems that hosts bearing various types of cancer show increased levels of inflammatory cytokines, such as IL-1β, GM-CSF, TGF-β, TNF-α, and IL-6 (40–43). IL-6 was recently demonstrated to be a candidate responsible for STAT3-dependent negative modulation of antitumor immunity (16, 44). Therefore, neutralization of IL-6 activity may become a novel strategy for improving the suppressed immune system in tumor-bearing hosts, as well as for enhancing antitumor immunity. Kishimoto’s group (28–30) developed anti–IL-6R mAb, which has had a great impact in the therapy of rheumatoid arthritis, Castleman’s disease, and multiple myeloma. In this study, we show that administration of anti–IL-6R mAb can significantly overcome the immune-suppressive condition in tumor-bearing hosts (i.e., downregulation or loss of expression of MHC class II molecules on DC) (Fig. 3). The anti–IL-6R mAb appears to alter the tumor microenvironment, which might permit activation of tumor-specific T cells by direct contact with tumor-associated DCs highly expressing MHC class II.
Previous studies showed a critical role for IL-6–STAT3 pathway–suppressed DC maturation including surface expression of MHC class II and attenuated CD4+ T cell and CD8+ T cell responses both in vivo and in vitro (17). Furthermore, signaling from various tumor-derived factors may converge on STAT3, and hyperactivation of STAT3 results in reduced production of mature DC (44). Therefore, inhibition of the IL-6–STAT3 pathway may be an attractive therapeutic approach to improve the function of DC in the tumor-bearing host. In this study, we provide evidence that the IL-6–arginase-1 pathway decreases the expression of MHC class II on DC mediated by L-arginine consumption in vivo. These phenomena may explain, in part, how IL-6 signaling decreases the intracellular MHC class II αβ dimer, Ii, and H2-DM levels by enhancing lysosome protease activity in DC in vitro (16)
It was shown that type-2 cytokines, including IL-4, IL-13, and IL-10, produced by tumor cells induced arginase-1 in murine peritoneal and bone marrow cell–derived macrophages and MDSC in vitro (8, 45). However, it remains unclear which tumor-derived factor is the cytokine responsible for initiating the production of arginase-1 in the tumor-bearing host. In this study, we demonstrate that both tumor- and host cell–derived IL-6 induce arginase-1 activity in tumor-associated CD11c+ DC, as well as peritoneal macrophages, in vivo. Interestingly, although serum IL-6 levels increased significantly in the tumor-bearing mice (Fig. 1A), serum IL-4 and IL-10 were not detected using CMS-G4 cells (Fig. 1B, 1C). In addition, IL-13 expression was not detected in the DC from the tumor-bearing mice (data not shown), indicating that these type-2 cytokines, with the exception of IL-6, are not involved in arginase-1 expression in this model. Moreover, previous studies (46, 47) demonstrated that PGE2 and IL-8 induced arginase expression in mouse DC or human tumor cells. We examined mRNA expression levels of Cox-2, Ptges (PGE2-related factors), and KC (murine IL-8 homolog) in the tumor cells used in the present experiments. As shown in Supplemental Fig. 1E, all tumor cell lines expressed these factors, suggesting that PGE2 and IL-8, as well as IL-6, might be involved in arginase-1 expression in the tumor microenvironment.
L-arginine, a nonessential amino acid, plays an important role in several biological systems, including the immune system (8). It was demonstrated that the lack of L-arginine blocks T cells proliferation via regulation of T cell cycle progression both in vitro and in vivo (45). However, it remains unresolved whether L-arginine starvation affects the function of APC, such as DC. In this study, we find that IL-6 induced arginase-1 expression in DC, and the starvation of L-arginine caused the downmodulation of the expression of MHC class II molecules but not MHC class I molecules. In addition, we confirmed that administration of nor-NOHA into tumor-bearing mice restored the downregulation, suggesting that the arginine levels controlled by arginase-1 contribute to MHC class II expression of DC in vivo. Moreover, selective blocking of the IL-6–arginase cascade with anti–IL-6R mAb or the arginase inhibitor nor-NOHA significantly improved DC and MHC class II–dependent CD4+ T cell activation and prevented the increase in immunosuppressive arginase activity in DC in the tumor-bearing host. It is also notable that such blockade of the IL-6–arginase cascade inhibited tumor growth in vivo. It was reported that tumor-infiltrating MDSC expressed arginase-1 and inhibited immune responses in the tumor microenvironment (48). Therefore, injection of nor-NOHA would inhibit arginase activity of both DC and MDSC.
We recently demonstrated that administration of anti–IL-6R mAb increased antitumor activity and found that the antitumor effects were abolished in mice depleted of CD8+ T cells (49). In addition, we confirmed that IFN-γ treatment inhibited the growth of CMS-G4 tumor cells and increased the expression levels of MHC class I, but not MHC class II, in vitro, suggesting an increased susceptibility of tumor cells to antitumor CTL (Fig. 5G, 5H). Therefore, both IFN-γ–producing CD4+ T cells and CD8+ T cells would be involved in antitumor effects in anti–IL-6 mAb or nor-NOHA treatment.
Thus, our novel finding proposes that the IL-6–arginase–signaling cascade may be a potentially useful target for developing a new strategy in cancer immunotherapy. It was reported that patients with lung, colon, renal, prostate, or breast cancer had significantly increased serum IL-6 levels, which were closely correlated with malignancies (50–54). Therefore, we speculate that anti–IL-6 mAb and nor-NOHA might be potent therapeutic agents for a range of cancer types. We are currently investigating tumor treatment using a combination of anti–IL-6R mAb and chemotherapy, which may improve immunosuppression mechanisms during the tumor-bearing state.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Dr. Y. Iwakura for providing IL-6−/− mice, Dr. M. Ito for providing RAG2−/− mice, Dr. T. Kitamura for providing pMX vector, and Dr. T. Sudo for providing GM-CSF–producing CHO cells. We also thank Chugai Pharmaceutical (Shizuoka, Japan) for their kind gifts of rIL-6 and anti–IL-6R Ab MR16-1.
Footnotes
This work was supported in part by a Grant-in-Aid for a National Project “Knowledge Cluster Initiative” (2nd stage, “Sapporo Biocluster Bio-S”), a Grant-in Aid for Scientific Research (B), and a Grant-in Aid for Young Scientists (B) (22700894 to D.W., 22790370 to H.K., and 22300331 to T.N.) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, as well as by the Joint Research Program of the Institute for Genetic Medicine, Hokkaido University.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- BMDC
- bone marrow–derived dendritic cell
- DC
- dendritic cell
- i.d.
- intradermally
- MDSC
- myeloid-derived suppressor cell
- nor-NOHA
- Nω-hydroxy-L-arginine
- Tg
- transgenic
- TIL
- tumor-infiltrating leukocyte.
- Received December 27, 2011.
- Accepted November 8, 2012.
- Copyright © 2013 by The American Association of Immunologists, Inc.