Tumor-Educated CD11b Ia Regulatory Dendritic Cells Suppress T Cell ...

Download PDF
2 downloads 0 Views 821KB Size Report
Comment
Arginase I played a major role in the suppression of T cell response by ... 3 Abbreviations used in this paper: DC, dendritic cell; DCreg, regulatory DC; mDC,.
The Journal of Immunology

Tumor-Educated CD11bhighIalow Regulatory Dendritic Cells Suppress T Cell Response through Arginase I1 Qiuyan Liu,* Chaoxiong Zhang,* Anna Sun,† Yuanyuan Zheng,* Li Wang,† and Xuetao Cao2*† Tumors can induce generation and accumulation of the immunosuppressive cells such as regulatory T cells in the tumor microenvironment, contributing to tumor escape from immunological attack. Although dendritic cell (DC)-based cancer vaccine can initiate antitumor immune response, regulatory DC subsets involved in the tolerance induction attracted much attention recently. Our previous studies demonstrate that the stromal microenvironment of the spleen, lung, and liver can program generation of CD11clowCD11bhighIalow DCs with regulatory function (CD11bhighIalow regulatory DCs). However, whether and how the tumor microenvironment can program generation of CD11bhighIalow regulatory DCs remain to be investigated. In this study, we used the freshly isolated tumor cells to mimic tumor microenvironment to coculture DCs and found that the freshly isolated tumor cells could drive DCs to differentiate into regulatory DCs with a CD11clowCD11bhigh Ialow phenotype and high expression of IL-10, NO, vascular endothelial growth factor, and arginase I. Tumor-educated CD11bhighIalow regulatory DCs inhibited CD4ⴙ T cell proliferation both in vitro and in vivo. 3LL lung cancer-derived TGF-␤ and PGE2 were responsible for the generation of regulatory DCs. PGE2 was the main inducer of arginase I in regulatory DCs. Arginase I played a major role in the suppression of T cell response by regulatory DCs induced by 3LL lung cancer. A natural counterpart of CD11bhighIalow DCs was identified in tumor tissue, and CD11bhighIalow DCs sorted from 3LL lung cancer tissue expressed arginase I and inhibited T cell response. Therefore, tumors can educate DCs to differentiate into a regulatory DC subset, which contributes to constitution of the immunosuppressive tumor microenvironment and promotes tumor immune escape. The Journal of Immunology, 2009, 182: 6207– 6216.

D

endritic cells (DC)3 play important roles in initiating innate and adaptive immune response, which is critical for the antitumor immune response (1– 4). However, increasing evidence shows that DCs can also induce immune tolerance or down-regulate immune response (5). It has been shown that imDC can induce T cell anergy, generate regulatory T (Treg) cells, and promote alloantigen-specific tolerance (6, 7). Also, in vivo delivery of Ag to nonactivated or resident DCs, which phenotypically resemble imDCs, results in the induction of tolerance of CD8⫹ T cells (8). It is now well accepted that the ability of DCs to initiate immune responses or induce tolerance is strictly dependent on their maturation state or subsets. In recent years, more attention has been paid for the subsets of DCs with regulatory

*National Key Laboratory of Medical Immunology and Institute of Immunology, Second Military Medical University, Shanghai, People’s Republic of China; and †Institute of Immunology, Zhejiang University School of Medicine, Hangzhou, People’s Republic of China Received for publication November 21, 2008. Accepted for publication March 11, 2009. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1

This work was supported by grants from the National Natural Science Foundation of China (Grants 30771984, 30672386, and 30721091) and the National Key Research Program of China (Grant 2007CB512403).

2

Address correspondence and reprint requests to Dr. Xuetao Cao, National Key Laboratory of Medical Immunology and Institute of Immunology, Second Military Medical University, Shanghai 200433, People’s Republic of China. E-mail address: [email protected]

3

Abbreviations used in this paper: DC, dendritic cell; DCreg, regulatory DC; mDC, mature DC; imDC, immature DC; TIDC, tumor-infiltrating DC; Treg, regulatory T cell; VEGF, vascular endothelial growth factor; MDSC, myeloid-derived suppressor cells; 7-AAD, 7-aminoactinomycin D; 1,4-PBIT, 1,4-phenylene-bis(1,2-ethanediyl)bis-isothiourea dihydrobromide; nor-NOHA, N-hydroxy-nor-L-arginine.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.0803926

function, and these regulatory DCs (DCreg) have been shown to be able to inhibit T cell response and inflammation (9 –11). Ex vivo generation of regulatory DCs has been proposed as a new experimental therapeutic approach for the autoimmune/inflammatory diseases (12–14). To date, many kinds of DCreg with different phenotypes have been reported; however, most of them were induced in vitro by culturing DCs in the immunosuppressive cytokines or drugs. Our previous studies show that the stromal microenvironment of the spleen, lung, and liver can drive DCs and hemopoietic progenitors to differentiate to regulatory DCs with CD11clowCD11bhighIalow phenotype and high secretion of IL-10, NO, and IP-10 but less IL-12 (15–18). The CD11clowCD11bhighIalow DCregs that we identified can inhibit T cell proliferation and also induce Treg cell generation, thus suppressing T cell-mediated inflammation and autoimmune diseases (17, 18). Interestingly, CD11bhighIalow DCregs derived from different organ stromal microenvironments are generated through different mechanisms; for example, liver stromaderived M-CSF is responsible for the generation of CD11bhigh Ialow DCregs (17), but splenic stroma-derived TGF-␤ induces the generation of the DCregs (15). In addition, CD11bhighIalow DCregs derived from different organ stromal microenvironments may exhibit their immunosuppressive function via different mechanisms; for example, splenic stroma-educated CD11bhighIalow DCregs inhibit T cells via secretion of NO (15); however, pulmonary stromaeducated CD11bhighIalow DCregs inhibit T cells via secretion of PGE2 (18). Tumor microenvironment is well known to be immunosuppressive (19 –21). Tumor cells consistently release many kinds of immunosuppressive and proinflammatory factors such as vascular endothelial growth factor (VEGF), TGF-␤, IL-10, PGE2, M-CSF, and IL-6 which facilitate tumor immune escape and tumor growth (22–24). Many kinds of immunosuppressive cells are present in the tumor microenvironment, including myeloid-derived suppressor

6208 cells (MDSC) and Treg cells. Also, DCs have been detected in the leukocyte infiltrate of various tumor tissues including breast, colon, stomach, lung, bladder, and pancreatic carcinoma. Tumor-infiltrating DCs (TIDC) with an immature phenotype were described 10 years ago to be related to the tumor immunosuppression (25, 26). Aspord et al. (27) reported that TIDCs could prime CD4⫹ T cells to produce IL-13 which in turn promoted early tumor development. Tumors were shown to promote differentiation of IL-10 and/or TGF-␤-secreting DCs that in turn expanded CD4⫹CD25⫹ Treg cells (28). IL-6 secreted by breast cancer cells can switch the differentiation of monocytes from DCs to macrophages, resulting skewing Ag presentation toward Ag degradation (29). As we know, tumor microenvironment can actively program immune dysfunction for the tumor cell escape of immunological attack (20), we wonder whether tumor microenvironment can program the generation of CD11bhighIalow regulatory DCs which in turn promote tumor escape of immune control. If so, what are the mechanisms for the generation of CD11bhighIalow regulatory DCs in the tumor microenvironment? And what are the mechanisms for their immunosuppressive function? In this study, we freshly isolated tumor cells from tumor tissue and then used these freshly isolated tumor cells to mimic tumor microenvironment of different origin (3LL lung cancer, CT26 colon cancer, B16 melanoma) to coculture DCs. We found that the 3LL lung cancer microenvironment could drive DCs to differentiate into CD11bhighIalow regulatory DCs via TGF-␤ and PGE2, and CD11bhighIalow regulatory DCs could inhibit T cell response via arginase I. More importantly, the natural counterpart of CD11bhighIalow DCs with similar regulatory function was identified in tumor tissue. Our results suggest that the tumor microenvironment can educate DCs to differentiate to regulatory DCs which in turn suppress T cell response, thus providing a new mechanistic explanation for tumor immune escape.

Materials and Methods Mice and cell line C57BL/6 mice and BALB/c mice were obtained from Joint Ventures Sipper BK Experimental Animal. DO11.10 OVA323–339-specific TCR-transgenic mice, EGFP-transgenic mice (ActbEGFP) with C57BL/6 background were obtained from The Jackson Laboratory. (C57BL/6 ⫻ DO11.10)F1 mice were prepared by crossing C57BL/6 mice with DO11.10 mice. Smad3⫺/⫺ mice (H-2kb) were provided by Professor Xiao Yang (Beijing Institute of Biotechnology, Beijing, China; Ref. 30). All mice were maintained under specific pathogen-free conditions and used at 6 – 8 wk of age. The 3LL Lewis lung cancer cell line derived from C57BL/6 origin was obtained from the American Type Culture Collection and maintained in RPMI 1640 complete medium (PAA Laboratories) supplemented with 10% FCS (PAA Laboratories).

Reagents Recombinant mouse GM-CSF, IL-4, and ELISA kits for murine IL-2, IFN-␥, TGF-␤, VEGF, IL-12, IL-10, and PGE2 were purchased from R&D Systems. Fluorescein-conjugated mAbs to CD4, CD8, CD11b, CD11c, Iab, CD40, CD80, CD86, and isotype control mAbs were purchased from BD Pharmingen. Microbead-conjugated mAbs to CD4, CD11b, and CD11c were purchased from Miltenyi Biotec and fluorescein-conjugated mAbs to F4/80, Gr-1, CD25, Foxp3, B7H1, B7DC, and 4-1BBL were obtained from eBioscience. Specific Ab for arginase and actin were obtained from Santa Cruz Biotechnology. Neutralizing Abs to IL-10, TGF-␤, M-CSF, or VEGF and isotype controls were purchased from R&D Systems. Neutralizing Abs to PGE2 were purchased from Cayman Chemical. 7-Aminoactinomycin D (7AAD), LPS, saponin, brefeldin A, BSA, 1-methyltryptophan, and 1,4phenylene-bis(1,2-ethanediyl)bis-isothiourea dihydrobromide (1,4-PBIT), were from Sigma-Aldrich. The arginase I-specific inhibitor N-hydroxy-nor-Larginine (nor-NOHA) was obtained from Calbiochem.

TUMOR PROGRAM GENERATION OF DCregs Cell preparation, coculture of freshly isolated tumor cells and DCs 3LL tumor cells were freshly isolated from the tumor tissue 2 wk after s.c. inoculation with 3LL cells into the C57BL/6 mice. The tumor tissue was digested by collagenase (100 ␮g/ml) for 1 h and then passed through the 40 ␮m pore size mesh to obtain the single cells. Then the tumor cells were cultured in 6-well plates in RPMI 1640 supplemented with 10% FCS for 1 wk and finally collected as the freshly isolated tumor cells to coculture with DCs. Mouse bone marrow-derived DCs were generated by culturing in GMCSF and IL-4 as described previously (15–18). On day 5, the dendritic proliferating clusters were collected and purified by anti-CD11c microbeads as imDCs. Purified imDCs were stimulated with LPS (100 ng/ml) for another 2 days and then collected as mature DCs (mDC). The purity of DCs was ⬎95% as confirmed by FACS. Once the freshly isolated tumor cells reached 50% confluence, DCs were seeded onto a 3LL cell monolayer at a density of 2.0 ⫻ 106/ml/well in a 6-well plate, with replacement of RPMI 1640 supplemented with 5% FCS. DCs were cocultured with the tumor cells for at least 60 h and then washed off the layer with 0.1% trypsin and 5 mM EDTA and purified with CD11c magnetic microbeads. In some experiments, anti-TGF-␤ mAb, antiVEGF, anti-GM-CSF mAb, anti-G-CSF mAb, or anti-GM-CSF mAb, cyclo-oxygenase 2 inhibitor, or arginase inhibitor were added, respectively, at the beginning of the coculture.

Flow cytometry The phenotype of the cells and phagocytosis of FITC-conjugated OVA were analyzed by LSR II flow cytometry (BD Biosciences) as described previously (15–18).

Assays for Ag-specific CD4⫹ T cell response Assays for Ag-specific CD4⫹ T cell response were performed as described previously (15–18). Briefly, for assay of Ag-specific T cell proliferation, splenic CD4⫹ T cells from DO11.10 OVA323–339-specific TCR-transgenic ⫻ C57BL/6)F1 hybrid mice were positively selected with anti-CD4coated microbeads (Miltenyi Biotec) by MACS and then cocultured with DCs treated as indicated in the presence of OVA323–339 peptide at a ratio of 1:10 (DC:T) in round-bottom 96-well plates (1 ⫻ 105 T cells/200 ␮l/ well) for 5 days. For intracellular staining, brefeldin A (Sigma-Aldrich) was added 6 h before the end of coculture. After staining for CD4, cells were fixed and permeabilized using Cytofix/Cytoperm (BD Biosciences), followed by intracellular staining of IFN-␥ or IL-2. Proliferation of T cells was analyzed by double staining with anti-CD4⫹FITC and 7-AAD resuspended in exactly 300 ␮l of PBS, and the number of CD4⫹ and 7-AAD⫺ cells were counted by FACS. In some experiments, 1-methytryptophan, indomethacin, 1,4-PBIT, NS389, nor-NOHA, or L-arginine were added, respectively, at the beginning of the coculture. For assay of Ag-specific T cell proliferation triggered by DCs in vivo, naive DO11.10 T cells (6 ⫻ 106) were transferred i.p. into BALB/c mice. DCs (6 ⫻ 106) were incubated with 1 ␮M OVA323–339 peptide at 37°C for 6 h, washed, and injected i.p. into the mice transferred with DO11.10 T cells. After 5 days, the cells from mesenteric lymph nodes and spleen were double stained with FITC-conjugated anti-CD4 mAb, PE-conjugated antiOVA322–339 peptide-specific TCR mAb KJ1-26, and counted by FACS as described previously (15–18).

Assay for cytokines and NO Cytokines in the supernatant of the DC-T coculture system were assayed with ELISA kits. NO production was assayed by measurement of the nitrite concentration with the Griess assay as described previously (15–18).

Assays for arginase expression and activity Cell extracts were obtained and the expression of arginase was detected by Western blot (31). Cytoplasmic extracts were electrophoresed in 12% Trisglycine gels (Novex), transferred to polyvinylidene difluoride membranes, and immunoblotted with the appropriate Abs. Proteins were visualized using SuperSignal West Femto Maximum Sensitivity Substrate, as instructed by the manufacturer (Pierce). Arginase activity was tested by measuring the production of L-ornithine. Briefly, Cell lysates (5 ␮g) were added to 25 ␮l of Tris-HCl (50 mM; pH 7.5) containing 10 mM MnCl2. This mixture was heated at 55– 60°C for 10 min to activate arginase. Then, a solution containing 150 ␮l of carbonate buffer (100 mM; Sigma-Aldrich) and 50 ␮l of L-arginine (100 mM) was added and incubated at 37°C for 20 min. The hydrolysis reaction from L-arginine to L-ornithine was identified by a colorimetric assay after the addition of ninhydrin solution and incubation at 95°C for 1 h.

The Journal of Immunology

6209

FIGURE 1. Tumor drives differentiation of DCs into CD11bhighIalow DCregs in vitro. A, imDCs were cocultured with freshly isolated 3LL tumor cells for 60 h (we named them as DCregs) and then labeled with Ab to Iab, CD40, CD80, CD86, CD11b, and CD11c, respectively, for phenotypic analysis by flow cytometry. Dotted lines represent isotype control. Numbers in histograms indicate geometric mean fluorescence intensity. B, Phagocytic ability of DCregs. Phagocytic ability was assessed for OVA-FITC phagocytosis by flow cytometry. Numbers in histograms indicated geometric mean fluorescence of test samples. Ctrl, control (cells incubated with OVA-FITC at 4°C). Data represent one of at least three independent experiments with similar results. C, NO expression and cytokine profile of DCregs stimulated with or without LPS (1 ␮g/ml) for 24 h, as measured by ELISA and Griess assays, respectively. Results were the mean ⫾ SD of triplicate wells. ⴱⴱ, p ⬍ 0.01. D, The expression and activity of arginase I in tumor-educated DCreg. DCreg were prepared as mentioned above, the protein expression of arginase I was detected by Western blot. The activity of arginase I was tested by measuring the production of L-ornithine as described in Materials and Methods. E, Analysis of F4/80 or Gr-1 expression in tumor-educated DCregs. DCregs were prepared as mentioned above; the expression of F4/80 or Gr-1 was analyzed by flow cytometry. Dotted lines represent isotype control.

Phenotypic and functional identification of DCregs in tumor tissue Tumor-infiltrating mononuclear cells were isolated from tumor tissue by Percoll density gradient centrifugation. The cells were labeled with anti-CD11b-allophycocyanin, anti-CD11c-PE, and anti-Ia-FITC and then analyzed by LSR II flow cytometry (BD Biosciences). The CD11clowCD11bhighIalow cells, sorted from tumor-infiltrating mononuclear cells by FACSDiva (BD Biosciences), were investigated for their effect on the mDC-initiated Ag-specific CD4 T cell proliferation as described previously (15–18).

Statistical analysis Comparisons between experimental groups and relevant controls were performed by Student’s t test. A value of p ⬍ 0.05 was considered a statistically significant difference.

Results Freshly isolated tumor cells educate DCs to differentiate to DCs with a CD11bhighIalow phenotype and high expression of immunosuppressive factors Bone marrow-derived immature CD11c⫹ DCs were cocultured with the freshly isolated 3LL lung cancer cells in vitro for 60 h and

then washed off the layer with 0.1% trypsin and 5 mM EDTA and purified with CD11c magnetic microbeads. As shown in Fig. 1A, after coculture with the freshly isolated 3LL lung tumor cells, DCs showed a distinct phenotype with high expression of CD11b, low expression of CD11c, Iab, and CD86, and unchanged expression of CD40 and CD80. Just like imDCs, such DCs remained higher phagocytic ability than mDCs (Fig. 1B). Even stimulated with LPS, the unique phenotype and potent phagocytic ability of such DCs remained unchanged, indicating that such DCs exhibit a stable state. As compared with the cytokine profile of the imDCs before coculture with the freshly isolated tumor cells, these tumoreducated DCs secreted higher levels of NO, IL-10, and VEGF but secreted low levels of IL-12 (Fig. 1C), with or without LPS stimulation. Also, expression and activity of arginase I was higher in those DCs compared with that in control DCs (Fig. 1D). In addition, we found that the expression of membrane molecules such as B7-H1, B7-DC, 4-1BBL, and TRAIL on those DCs remained unchanged as compared with that of control DCs (data not shown). To distinguish those CD11bhigh DCs from the macrophages or MDSC with high CD11b expression, we further detected the F4/80

6210 and Gr-1 expression in those tumor-educated CD11bhighIalow DCs. We found that those CD11bhighIalow DCs did not express F4/80 or Gr-1 (Fig. 1E), further confirming that those tumor-educated CD11bhighIalow DCs are different from the CD11bhigh macrophages and MDSC. Taken together with the following functional experiments of those DCs, the similar phenotype (CD11clow CD11bhighIalow) and cytokine profile (high IL-10 and NO) of the CD11bhighIalow DCregs that we identified in the spleen, lung, and liver before (15–18) suggest that the tumor may drive differentiation of imDCs to DCregs, and those tumor-educated DCs may be involved in the suppression of immune response. Thus, we termed those tumor-educated DCs DCregs. Additionally, we further demonstrated that the freshly isolated tumor cells could educate mDCs to differentiate into CD11bhighIalow DCregs after coculture for 60 h (data not shown). Furthermore, we showed that freshly isolated CT26 colon cancer cells or B16 melanoma cells could induce generation of DCregs after coculture with DCs. Altogether, tumor microenvironment may educate DCs to differentiate to the subset of DCs with distinct phenotype (CD11clowCD11bhighIalow) and cytokine profile (high IL-10, NO, VEGF, and arginase I). Tumor-educated CD11bhighIalow DCs inhibit T cell response both in vitro and in vivo Next, we investigated whether the tumor-educated CD11bhighIalow DCs exerted immunoregulatory function, just like the DCreg we identified in the lymph organs and nonlymph organs before. As shown in Fig. 2A, mDCs could effectively prime proliferation of OVA323–339 peptide-specific CD4⫹ T cells, whereas, the tumoreducated CD11bhighIalow DCs alone could not. When tumor-educated CD11bhighIalow DCs were added to the mDCs/CD4T coculture system, the mDC-initiated CD4 T cell proliferation in vitro was suppressed profoundly. Furthermore, once adoptively transferred together with OVA323–339-loaded mDCs into the mice preinjected with DO11.10 T cells, the tumor-educated CD11bhighIalow DCs significantly inhibited the Ag-specific CD4 T cell proliferation in vivo, because the number of Ag peptide-specific DO11.10 T cells in spleen and lymph node was reduced markedly 5 days later as compared with that in the mice without adoptive transfer of the tumor-educated CD11bhighIalow DCs (Fig. 2B). As we know, CD4⫹CD25⫹Foxp3⫹ Tregs increased markedly in the tumorbearing mice and cancer patients both systemically (in the lymph organs and peripheral blood) and locally (in tumor tissue), which contribute to the tumor-induced immunosuppression or immune escape. Furthermore, DCregs have been shown to be able to induce generation of Treg cells. Therefore, we wondered whether tumoreducated CD11bhighIalow DCs could promote generation of Treg cells in vivo. As shown in Fig. 2C, the percentage of CD4⫹ CD25⫹Foxp3⫹ Treg cells in the spleen and lymph nodes increased significantly after adoptive transfer of tumor-educated CD11bhigh Ialow DCs into the mice, indicating that the systemic antitumor immunity could be impaired by tumor-educated CD11bhighIalow DCs. We went further to inoculate 3LL tumor cells together with CD11bhighIalow DCs into the mice s.c. and then found that DCregs could promote s.c. tumor growth and shorten the survival of tumor-bearing mice (supplemental Fig. 1, A and B).4 We found that coculture with 3LL tumor cell line in vitro could also induce the generation of CD11bhighIalow DCregs, but the efficacy of such CD11bhighIalow DCregs was less than that of CD11bhighIalow DCregs generated by coculture with the freshly isolated tumor cells in the suppression of CD4⫹ T cell proliferation (Fig. 2D). Additionally, DCregs derived from coculture with other tumor 4

The online version of this article contains supplemental material.

TUMOR PROGRAM GENERATION OF DCregs cells such as CT26 colon cancer cells or B16 melanoma cells could also significantly inhibit mDC-initiated Ag-specific CD4 T cell proliferation (Fig. 2E). Collectively, these results demonstrate that, similar to the DCregs phenotypically and functionally that we identified before, the tumor-educated CD11bhighIalow DCs can inhibit T cell response and exert regulatory function. Therefore, those tumor-educated CD11bhighIalow DCs are DCregs. Tumor-derived TGF-␤ and PGE2 are responsible for the differentiation of CD11bhighIalow regulatory DCs To elucidate the mechanism by which tumor microenvironment induce DCs to differentiate to DCregs, we first determined whether cell-cell contact is required for the process by using a transwell system in the 60-h coculture of DCs and freshly isolated 3LL tumor cells. As shown in Fig. 3A, the DCregs could be generated with same regulatory function in transwell system, indicting that tumor-DC cell-cell contact is not required, and soluble factors secreted by tumor cells were enough for the differentiation of regulatory DCs. Our previous study once showed that 3LL tumor cells could secrete many immunosuppressive cytokines including TGF-␤, M-CSF, G-CSF, VEGF, GM-CSF, PGE2, and chemokines such as MIP-3␣ which could chemoattract more imDCs to tumor tissue (32). To determine which factor was involved in differentiation of DCregs, the neutralizing Abs against TGF-␤, M-CSF, G-CSF, VEGF, GM-CSF, and PGE2 were added to the coculture system, respectively, at the beginning. After 60 h of coculture, the DCs were collected for the assay of their regulatory function. We found that blockade of TGF-␤ or PGE2 in the coculture system led to partial loss of inhibitory functions of DCreg (Fig. 3, B and D). Simultaneous blockade of TGF-␤ and PGE2 reversed the inhibitory functions of DCreg more significantly (Fig. 3D) than blockade of TGF-␤ or PGE2 alone. However, blockade of M-CSF, G-CSF, VEGF, and GM-CSF could not significantly reduce the inhibitory function of DCregs (data not shown). Similarly, mouse CT26 colon cancer cells or B16 melanoma cells also produced TGF-␤ and PGE2, and coculture with both CT26 cells and B16 cells could induce the generation of CD11bhighIalow DCs. Additionally, blockade of TGF-␤ and PGE2 in the coculture system of DCs with CT26 or B16 tumor cells could partially reduce the inhibitory functions of CD11bhighIalow DCs (Fig. 3E), indicating that TGF-␤ and PGE2 were partially involved in the induction of DCregs by colon cancer or melanoma but other more important factors need to be investigated. The data indicate that TGF-␤ and PGE2 may be the main factors involved in the induction of DCregs by lung cancer, but other factors may play important roles in the induction of DCregs by colon cancer or melanoma. These data suggest that different kinds of tumor cells may have different mechanisms to induce DCregs with similar phenotype and suppressive function. Furthermore, we used DCs prepared from Smad3-deficient mice (Smad3⫺/⫺) to coculture with the freshly isolated 3LL tumor cells, and we found that the inhibitory function of Smad3⫺/⫺ DCs was significantly more reduced than those derived from wild-type DCs cocultured with freshly isolated 3LL tumor cells (Fig. 3C). These results demonstrate that 3LL lung cancer-derived TGF-␤ and PGE2 play important roles in the differentiation of DCs into DCregs driven by tumor microenvironment. CD11bhighIalow DCregs-derived arginase I is responsible for the suppression of T cell response What is the mechanism by which tumor-educated DCregs inhibit T cell response? As mentioned above, CD11bhighIalow DCregs secreted high levels of IL-10, VEGF, NO and expressed arginase I (Fig. 1, C and D). IL-10 and VEGF are well known immunosuppressive factors. NO has been shown to be involved in the

The Journal of Immunology

6211

FIGURE 2. Tumor-educated regulatory DCs inhibit mature DC-inititated CD4⫹ T cell proliferation. A, CD4⫹ T cells from DO11.10 OVA323–339specific (TCR-transgenic ⫻ C57BL/6)F1 hybrid mice were cocultured with DCregs and/or mDC in the present of OVA peptides, 5 days later, the total number of viable CD4⫹ T (CD4⫹7-AAD⫺) cells in each well was measured by flow cytomerty. B, The in vivo inhibitory function of DCregs. DO11.10 T cells and OVA-pulsed DCregs and/or mDCs were transferred together i.p. into recipient mice. After 5 days, the cells from mesenteric lymph nodes (LN) and spleens were double stained with FITC-CD4 and PE-KJI-26 and counted by FACS. Numbers in CD4-gated plots indicate percentage of DO11.10 cells (KJI-26⫹) among the total CD4⫹ T cells. C, DCregs were prepared as mentioned above and transferred into C57BL/6 mice via tail vein (5 ⫻ 106/mouse). Three days later, the percentages of CD4⫹CD25⫹Foxp3⫹ Treg cells in spleen and lymph nodes were measured by flow cytometry. CD4⫹ cells were first gated and further analyzed using CD25 and Foxp3 markers. D, Suppression of CD4⫹ T cell proliferation by CD11bhighIalow DCregs derived from the coculture with 3LL cell lines (DCreg-L) or the freshly isolated tumor cells (DCreg-T). E, DCregs derived from coculture with freshly isolated CT26 or B16 tumor cells inhibited T cell proliferation in vitro. Data represent one of at least three experiments with similar results. ⴱ, p ⬍ 0.05, ⴱⴱ, p ⬍ 0.01.

inhibition of T cell proliferation and also to be responsible for the immunosuppressive function of DCregs we identified before (15). L-Arginine is a nonessential amino acid that plays roles in several biological systems including the immune response (33). L-Arginine is metabolized by arginase, and tumor-associated MDSC-derived arginase could impair T cell proliferation (34, 35). Thus, we screened for which factor might be involved. Neutralizing Abs against IL-10 or VEGF did not reverse the suppression of T cell proliferation by DCregs (data not shown), indicating that IL-10, VEGF were not involved in the regulatory function of the tumoreducated DCreg. However, inhibitor of inducible NO synthase (1,4-PBIT or NG-monomethyl-L-arginine monoacetate) could partially reverse suppression of T cell proliferation by DCregs (Fig.

4A), suggesting that NO is partially involved in DCreg-medicated suppression of T cell function. The data also indicated that there are other factor(s) involved in regulatory function of tumor-educated DCregs. As mentioned above, DCregs also expressed high levels of arginase I. Next, we added the arginase inhibitor norNOHA or added exogenous L-arginine to a DCreg/maDC/CD4T coculture system, and we found that arginase inhibitor (norNOHA) and exogenous L-arginine significantly reversed the suppression of T cell proliferation by DCregs (Fig. 4B), showing the major role of arginase I in DCreg-medicated suppression of T cell function. Furthermore, arginase inhibitor (nor-NOHA) also partially reduced the inhibitory function of DCregs generated from the coculture system with the CT26 or B16 tumor cells (Fig. 4C),

6212

TUMOR PROGRAM GENERATION OF DCregs

FIGURE 3. Tumor-derived TGF-␤ and PGE2 are involved in the generation of tumor-educated DCregs. A, Tumor-derived soluble factors are involved in DCreg-medicated suppressive function of CD4⫹ T cell proliferation. Using a transwell system in 3LL/DC coculture experiments, DCs were cultured in the upper chamber of culture inserts with 0.4 ␮m pore size, 3LL tumor cells were added to the lower chamber and incubated for 60 h, and then DCregs (transwell-DCreg) were harvested and further cocultured with CD4⫹ T cells from DO11.10 OVA323–339-specific (TCR-transgenic ⫻ C57BL/6)F1 mice for 5 days in the presence of OVA, Finally, the number of viable CD4⫹ T cells (CD4⫹7-AAD⫺) were detected by flow cytometry. B, Neutralizing Ab against TGF-␤ partially reversed the DCreg-medicated suppression of CD4⫹ T cell proliferation. In the 3LL/DC coculture system, different concentrations of neutralizing Ab against TGF-␤ (5 ␮g/ml, 10 ␮g/ml) were added at the beginning of coculture. Isotype-matched Ab was included as a negative control. After 60 h, DCregs were harvested and further cocultured with CD4⫹ T cells from DO11.10 OVA323–339-specific (TCR-transgenic ⫻ C57BL/6)F1 mice for 5 days in the presence of OVA. Finally, the number of viable CD4⫹ T cells (CD4⫹7-AAD⫺) was detected by flow cytometry. C, DCreg-derived Smad3⫺/⫺ mice significantly lost the suppressive function to inhibit CD4 ⫹ T cell proliferation. DCs derived form Smad3⫺/⫺ mice were cocultured with the freshly isolated 3LL tumor cells as described above, DCreg-derived Smad3⫺/⫺ mice were harvested and cocultured with CD4⫹ T cells from DO11.10 OVA323–339-specific (TCR-transgenic ⫻ C57BL/6)F1 mice for 5 days in the presence of OVA. Finally, the number of viable CD4⫹ T cells (CD4⫹7-AAD⫺) was detected by flow cytometry. D, Simultaneous blockade of PGE2 and TGF-␤ reversed DCreg-medicated suppression of CD4 ⫹ T cell proliferation more significantly than blockade of PGE2 or TGF-␤ alone. Just like the blockade of TGF-␤ by neutralizing Ab, PGE2 was blocked by different concentrations of neutralizing Ab against PGE2 (5 ␮g/ml, 10 ␮g/ml) added at the beginning of coculture. Isotype-matched Abs were included as a negative control. E, Simultaneous blockade of PGE2 and TGF-␤ partially reduced suppression of CD4⫹ T cell proliferation by DCregs that derived from coculture with freshly isolated CT26 or B16 tumor cells. Results were means ⫾ SD of triplicate wells. ⴱⴱ, p ⬍ 0.01.

indicating that other factors but not arginase I may play important roles in the suppressive function of DCregS derived by coculture with CT26 or B16 tumor cells. To know which factor(s) derived from tumor cells are responsible for arginase I expression in DCregs, we stimulated DCs with PGE2 (1 ␮g/ml) or/and TGF-␤ (50 ng/ml) for 24 h; then arginase I expression was detected by Western blot As shown in Fig. 4D, PGE2, but not TGF-␤, significantly up-regulated arginase I expression in DCs, and also we found that blockade of TGF-␤ could reduce NO production of DCreg (Fig. 4E). Together, our data suggested that 3LL lung cancer-derived PGE2 is a main inducer of arginase I of the DCregs and

that 3LL lung cancer-derived TGF-␤ is the main inducer of NO of the DCreg. In addition to the partial role of NO in the immunosuppressive function of DCregs, arginase I derived from the 3LL lung cancer-educated DCreg is the major factor for their regulatory function. Identification of natural counterpart of CD11bhighIalow DCregs in tumor tissue With DC/tumor cell coculture system in vitro, we found that freshly isolated tumor cells could educate DCs into CD11bhighIalow regulatory DC subset. To exclude the possibility that this is the artificial

The Journal of Immunology

6213

FIGURE 4. Arginase I is involved in DCreg-medicated suppression of CD4⫹ T cell proliferation. CD4⫹ T cells from (DO11.10 ⫻ C57BL/6)F1 mice were cocultured with DCregs and/or mDC in the present of OVAs. NO inhibitor (1,4PBIT, 5 ␮g/ml or NG-monomethyl-L-arginine monoacetate (NMMA), 1 ␮M) or arginase inhibitor (nor-N, 50 ␮M) or L-arginine (20 mM) were added at the beginning of the coculture, respectively. Five days later, the number of viable CD4⫹ T cells (CD4⫹7-AAD⫺) was detected by flow cytometry. A, NO inhibitor partially reversed DCreg-medicated suppression of CD4 ⫹ T cell proliferation. B, Arginase inhibitor significantly reversed DCreg-medicated suppression of CD4 ⫹ T cell proliferation. C, Arginase inhibitor partially reversed suppression of CD4 ⫹ T cell proliferation by DCregs that derived from coculture with freshly isolated CT26 or B16 tumor cells. D, PGE2 was the main inducer of arginase I expression in DCs. DCs were stimulated by TGF-␤ or/and PGE2 for 24 h, then arginase I expression was detected by Western blot. E, TGF-␤ neutralizing Ab reduced NO production of DCregs. Neutralizing Ab against TGF-␤ was added to 3LL/DC coculture system, and then NO was measured by Griess assay. Data represent one of at least three experiments with similar results. ⴱ, p ⬍ 0.05, ⴱⴱ, p ⬍ 0.01.

phenomenon observed in vitro, we went further to confirm whether CD11clowCD11bhighIalow regulatory DCs do exist in tumor tissue. Tumor-infiltrating mononuclear cells were isolated from 3LL tumor tissue, and then the cell subpopulations were analyzed by FACS. As shown in Fig. 5A, CD11clowCD11bhighIalow DCs did exist in the tumor-infiltrating mononuclear cells and constituted ⬃20% of CD11bhigh tumor-infiltrating mononuclear cells. Then, these cells were sorted from the tumor-infiltrating mononuclear cells by FACSDiva. We found that the sorted CD11clowCD11bhigh Ialow cells derived from tumor tissue could significantly inhibit Ag peptide-specific CD4 T cell proliferation (Fig. 5B). Furthermore, the sorted CD11clowCD11bhighIalow cells expressed high level arginase I (Fig. 5C). When the arginase I inhibitor was used to pretreat the sorted CD11clowCD11bhighIalow cells, and then these CD11clowCD11bhighIalow cells were observed for their suppressive function. As shown in Fig. 5D, the arginase I inhibitor could significantly reduce the ability of the sorted CD11clow CD11bhighIalow cells to suppress the CD4 T cell proliferation.

To know whether the CD11clowCD11bhighIalow DCs are more frequent in tumor tissues than those in healthy organs, we analyzed the frequency of the CD11bhighIalow DCreg in tumor tissues and healthy organs. We found that the CD11clowCD11bhigh Ialow DCreg accounted for ⬃1% of total spleen, ⬃1% of total lymph nodes, and ⬃1% of the total liver mononuclear cells in healthy mice. However, in tumor-bearing mice, the CD11clow CD11bhighIalow DCreg accounted for ⬃1.5% of total spleen, ⬃1.1% of total lymph nodes, ⬃1.3% of the total liver mononuclear cells, ⬃8.9% of tumor-infiltrating mononuclear cells (Fig. 5E). We also found that the CD11clowCD11bhighIalow DCs were more frequent in tissues of CT26 colon cancer and B16 melanoma as compared with that in healthy organs (supplemental Fig. 2, A and B). Taken together, the results indicate that the sorted CD11clowCD11bhighIalow cells from tumor tissue have functional characteristics to those of the tumor-educated DCregs that we identified in vitro, and they may represent as natural counterpart of DCreg in tumor tissue.

6214

TUMOR PROGRAM GENERATION OF DCregs

Discussion

FIGURE 5. Isolation and identification of CD11clowCD11bhighIalow regulatory DCs from tumor tissue. 3LL tumor cells were subcutaneously inoculated into C57BL/6 mice. After 2 wk, tumor tissue was isolated and digested by collagenase (100 ␮g/ml) for 1 h, and then tumor-infiltrating mononuclear cells were isolated by Percoll density gradient centrifugation. A, Tumor-infiltrating mononuclear cells were labeled with anti-CD11b-allophycocyanin, CD11cFE, and Ia-FITC and analyzed by flow cytometry. CD11bhigh cells were first gated and further analyzed using CD11c and Ia marker. B, CD11clowCD11bhigh Ialow DCs were sorted by FACSDiva from tumor tissue and then cocultured with peptide-specific CD4 T cells or CD4 T/mDC for 5 days. The relative cell number of viable CD4 T cells was assayed by FACS. ⴱⴱ, p ⬍ 0.01. C, Expression of arginase I in CD11clowCD11bhighIalow DCs sorted from tumor tissue was detected by Western blot. D, The arginase I inhibitor (nor-N, 50 ␮M) was used to pretreat the CD11clowCD11bhighIalow DCs sorted from tumor tissue for 1 h, and then these DCregs were cocultured with mDC/CD4⫹ T cells for 5 days in the presence of OVA. Finally, the relative cell number of viable CD4 T cells was assayed by FACS. E, The CD11clowCD11bhighIalow DCs are more frequent in tumor tissues. Freshly isolated splenocytes, lymph node (LN) cells, liver mononuclear cells, and tumor-infiltrating mononuclear cells from 3LL tumor-bearing mice (day14) or normal mice were labeled with antiCD11b-allophycocyanin, CD11c-FE, and Ia-FITC, and then the CD11clow CD11bhighIalow DCs were analyzed by flow cytometry. ⴱⴱ, p ⬍ 0.01.

There are at least four possibilities that tumor-associated DCs may be involved in tumor escape from immunological attack. First, tumor or/and tumor microenvironment may inhibit mature DCs into the tumor tissue, and tumor or/and tumor microenvironment can secrete some chemokines such as MIP-3␣ which may selectively chemoattract imDCs to tumor tissue. Secondly, tumor can inhibit maturation of tumor-infiltrating imDC, which in turn may induce T cell tolerance/dysfunction. Thirdly, the capability of phagocytosis and Ag processing of DCs may be inhibited by tumor. Fourth, the ability of DCs migrating out of the tumor and into lymph nodes may be impaired in the tumor microenvironment (20, 36 – 40). Recently, DCreg subsets, which are involved in the induction of tolerance or down-regulation of immune response, have been regarded as another potential mechanism for tumor immune escape. The tumor microenvironment is well known to be immunosuppressive (19 –21). Several kinds of regulatory or tolerogenic DCs have been reported recently. Aspord et al. (27) found that TIDC prime CD4⫹ T cells to produce IL-13 which in turn promotes early tumor development, Ghiringhelli et al. (41) reported that tumor cells convert immature myeloid DC into TGF-␤-secreting cells which can induce CD4⫹CD25⫹ Treg cell generation, and Kuang et al. (42) demonstrated that tumor microenvironments educate DCs to adopt a semimature phenotype, which in turn aids tumor immune escape by causing defects in the CD3-TCR complex and deletion of T cells. In this study, we have demonstrated that tumor microenvironment can educate DCs to differentiate to CD11clow CD11bhighIalow DCreg with immunosuppressive function, which can inhibit T cell response. Furthermore, those regulatory DCs did not express F4/80 or Gr-1, thus distinguishing them from the CD11bhigh macrophages and MDSC. As to the pathological roles of tumor-educated DCregs in the tumor immune escape, we found that in vivo adoptive transfer of DCregs could promote tumor growth in vivo and shorten the survival of tumor-bearing mice, when these DCregs were coinjected s.c. with 3LL tumor cells into the mice. Therefore, we propose another possible explanation for the mechanisms of tumor immune escape. Tumor-derived factors such as TGF-␤, VEGF, IL-10, and PGE2 and other factors such as gangliosides and lactate have been verified to be able to regulate differentiation and function of DCs and T cells (22–24). For example, M-CSF and IL-6 inhibit the differentiation of DC, and IL-10 inhibits the differentiation, maturation, and function of DCs, shifting differentiation toward a macrophage phenotype. TGF-␤ inhibits the maturation, IL-12 production, and Ag-presenting capacity of LPS-stimulated DCs. VEGF inhibits DC development from CD34⫹ precursor cells. VEGF inhibits DC development leading to an increase in immature Gr1⫹ myeloid cells, and neutralizing Ab against VEGF could revert the inhibitory effect in vivo. Which factors derived from tumor microenvironment are involved in the differentiation and regulatory function of tumor-educated DCreg? 3LL lung cancer cells secrete high level of TGF-␤, VEGF, M-CSF, and PGE2 but not IL-6 and IL-10, as demonstrated by us before (32). In our tumor-DC coculture system, blockade of TGF-␤ and PGE2 can significantly reverse DCreg-medicated suppression of CD4 T cell proliferation, and PGE2 is responsible for arginase I expression and TGF-␤ for NO production. The data suggest that TGF-␤ and PGE2 derived from 3LL lung cancer cells may be the key factors involved in generation of suppressive DCreg. Importantly, we found that CD11bhigh Ialow DCregs are more frequent in tumor tissues than that in healthy organs such as spleen, lymph nodes and liver. As to whether and how other tumor metabolites such as gangliosides and

The Journal of Immunology lantate in the tumor microenvironment may be involved in the tumor-driven generation of DCregs needs to be further investigated. Just as stromal cells from different healthy tissues induce DCreg via different factors, different kinds of tumor cells may have different mechanisms to induce DCregs. For the induction of DCreg by lung cancer, the lung cancer-derived TGF-␤ and PGE2 played important roles. Whereas colon cancer- or melanoma-derived TGF-␤ and PGE2 were partially involved in the induction of DCregs by colon cancer and melanoma, other more important factors involved in the induction of DCregs by colon cancer and melanoma need to be identified in the future. Tumor-educated DCregs secrete several kinds of immunosuppressive factors. Which factors derived from DCregs are involved in the suppression of mDC-initiated Ag-specific CD4 T cell proliferation? Increasing evidence shows that NO, IDO, arginase, and IL-10 were involved in suppression of T cell functions. CD11clow CD45RB⫹ DCs down-regulate T cell response mainly by inducing Treg cells (10, 44), and CD11bhighIalow splenic DCregs inhibit T cell proliferation via NO (15), also other DCregs can exert their regulatory function through IDO (9, 44). The mechanisms of tumor-derived DCregs to inhibit T cell proliferation are different from that reported before. We found that the NO inhibitor could partially reverse DCreg-medicated T cell suppression, suggesting there were other factors playing major roles in the regulatory function of tumor-educated DCregs. Gabrilovich et al. (45) and Bronte et al. (46) reported that the CD11b⫹ myeloid cells isolated from the spleens of tumor-bearing mice could suppress T cell functions in vitro; additionally, they could be sorted into three subpopulations, CD11b⫹Gr-1⫺, CD11b⫹Gr-1⫹, and CD11b⫺Gr-1⫹. High arginase activity was found only in the mature CD11b⫹Gr-1⫺ subpopulation but not in the immature CD11b⫹Gr-1⫹ myeloid cells or the CD11b⫺Gr-1⫹ cells. In our coculture system, 3LL lung cancer-educated DCregs express high levels of CD11b and arginase and inhibit T cell response mainly through arginase I. Arginase I was partially involved in the suppression of T cell response by DCregs induced by CT26 colon cancer and B16 melanoma, indicating that other factors may play more important roles in the suppressive function of DCregs induced by colon cancer and melanoma, and these more important factors need to be investigated in the future. Our data suggest that DCregs induced by different tumors may exhibit their immunosuppressive function via different mechanisms. Also, the data indicate that there are many kinds of immunosuppressive myeloid cells present in the tumor microenvironment, which may inhibit immune response against cancer through secretion of soluble factors. These regulatory myeloid cells and factors constitute the components of immunosuppressive tumor microenvironment, contributing to tumor immune escape and tumor growth. In conclusion, we demonstrate that 3LL lung cancer-derived TGF-␤ and PGE2 could drive DCs to differentiate into CD11clow CD11bhighIalow DCregs that suppress T cell response via arginase I. One population of CD11clowCD11bhighIalow cells have been verified in the tumor tissue, and the CD11clowCD11bhighIalow cells freshly sorted from tumor tissue also express arginase I and inhibit T cell response. Therefore, the tumor-educated CD11bhighIalow DCregs may play important role in the tumor immune escape.

Acknowledgments We sincerely appreciate Dandan Zhang and Miao Chen for excellent technical assistance.

Disclosures The authors have no financial conflict of interest.

6215

References 1. Steinman, R. M. 1991. The dendritic cell system and its role in immunogenicity. Annu. Rev. Immunol. 9: 271–296. 2. Steinman, R. M., and J. Banchereau. 2007. Taking dendritic cells into medicine. Nature 449: 419 – 426. 3. Melief, C. J. 2008. Cancer immunotherapy by dendritic cells. Immunity 29: 372–383. 4. Vulink, A., K. J. Radford, C. Melief, and D. N. Hart. 2008. Dendritic cells in cancer immunotherapy. Adv Cancer Res. 99: 363– 407. 5. Steinman, R. M., D. Hawiger, and M. C. Nussenzweig. 2003. Tolerogenic dendritic cells. Annu. Rev. Immunol. 21: 685–711. 6. Charbonnier, L. M., L. M. van Duivenvoorde, F. Apparailly, C. Cantos, W. G. Han, D. Noe¨l, C. Duperray, T. W. Huizinga, R. E. Toes, C. Jorgensen, and P. Louis-Plence. 2006. Immature dendritic cells suppress collagen-induced arthritis by in vivo expansion of CD49b⫹ regulatory T cells. J. Immunol. 177: 3806 –3813. 7. Cools, N., V. F. Van Tendeloo, E. L. Smits, M. Lenjou, G. Nijs, D. R. Van Bockstaele, Z. N. Berneman, and P. Ponsaerts. 2008. Immunosuppression induced by immature dendritic cells is mediated by TGF-␤/IL-10 double-positive CD4⫹ regulatory T cells. J. Cell. Mol. Med. 12: 690 –700. 8. Bonifaz, D., K. Bonnyay, M. Mahnke, M. C. Rivera, M. C. Nussenzweig, and R. M. Steinman. 2002. Efficient targeting of protein antigen to the dendritic cell receptor DEC-205 in the steady state leads to antigen presentation on major histocompatibility complex class I products and peripheral CD8⫹ T cell tolerance. J. Exp. Med. 196: 1627–1638. 9. Munn, D. H., M. D. Sharma, J. R. Lee, K. G. Jhaver, T. S. Johnson, D. B. Keskin, B. Marshall, P. Chandler, S. J. Antonia, R. Burgess, Jr., C. L. Slingluff, and A. L. Mellor. 2002. Potential regulatory function of human dendritic cells expressing indoleamine 2,3-dioxygenase. Science 297: 1867–1870. 10. Wakkach, A., N. Fournier, V. Brun, J. P. Breittmayer, F. Cottrez, and H. Groux. 2003. Characterization of dendritic cells that induce tolerance and T regulatory 1 cell differentiation in vivo. Immunity 18: 605– 617. 11. Chorny, A., E. Gonzalez-Rey, A. Fernandez-Martin, D. Pozo, D. Ganea, and M. Delgado. 2005. Vasoactive intestinal peptide induces regulatory dendritic cells with therapeutic effects on autoimmune disorders. Proc. Natl. Acad. Sci. USA 102: 13562–13567. 12. Sato, K., N. Yamashita, N. Yamashita, M. Baba, and T. Matsuyama. 2003. Regulatory dendritic cells protect mice from murine acute graft-versus-host disease and leukemia relapse. Immunity 18: 367–379. 13. Adorini, L. 2003. Tolerogenic dendritic cells induced by vitamin D receptor ligands enhance regulatory T cells inhibiting autoimmune diabetes. Ann. NY Acad. Sci. 987: 258 –261. 14. Li, H., G. X. Zhang, Y. Chen, H. Xu, D. C. Fitzgerald, Z. Zhao, and A. Rostami. 2008. CD11c⫹CD11b⫹ dendritic cells play an important role in intravenous tolerance and the suppression of experimental autoimmune encephalomyelitis. J. Immunol. 181: 2483–2493. 15. Zhang, M., H. Tang, Z. Guo, H. An, X. Zhu, W. Song, J. Guo, X. Huang, T. Chen, J. Wang, and X. Cao. 2004. Splenic stroma drives mature dendritic cells to differentiate into regulatory dendritic cells. Nat. Immunol. 5: 1124 –1133. 16. Tang, H., Z. Guo, M. Zhang, J. Wang, G. Chen, and X. Cao. 2006. Endothelial stroma programs hematopoietic stem cells to differentiate into regulatory dendritic cells through IL-10. Blood 108: 1189 –1197. 17. Xia, S., Z. Guo, X. Xu, H. Yi, Q. Wang, and X. Cao. 2008. Hepatic microenvironment programs hematopoietic progenitor differentiation into regulatory dendritic cells maintaining liver tolerance. Blood 112: 3175–3185. 18. Li, Q., Z. Guo, X. Xu, S. Xia, and X. Cao. 2008. Pulmonary stromal cells induce the generation of regulatory DC attenuating T-cell-mediated lung inflammation. Eur. J. Immunol. 38: 2751–2761. 19. Kim, R., M. Emi, K. Tanabe, and K. Arihiro. 2006. Tumor-driven evolution of immunosuppressive networks during malignant progression. Cancer Res. 66: 5527–5536. 20. Rabinovich, G. A., D. Gabrilovich, and E. M. Sotomayor. 2007. Immunosuppressive strategies that are mediated by tumor cells. Annu. Rev. Immunol. 25: 267–296. 21. Zou, W. 2005. Immunosuppressive networks in the tumour environment and their therapeutic relevance. Nat. Rev. Cancer 5: 263–274. 22. Shurin, M. R., G. V. Shurin, L. A. Okshin, Z. R. Yurkovetsky, D. W. Gutkin, G. Chatta, H. Zhong, B. Han, and R. L. Ferris. 2006. Intratumoral cytokines/ chemokines/growth factors and tumor infiltrating dendritic cells: friends or enemies? Cancer Metastasis Rev. 25: 333–356. 23. Kusmartsev, S., and D. I. Gabrilovich. 2006. Effect of tumor-derived cytokines and growth factors on differentiation and immune suppressive features of myeloid cells in cancer. Cancer Metastasis Rev. 25: 323–331. 24. Lin, W. W., and M. A. Karin. 2007. Cytokine-mediated link between innate immunity, inflammation, and cancer. J. Clin. Invest. 117: 1175–1183. 25. Todryk, S., A. A. Melcher, N. Hardwick, E. Linardakis, A. Bateman, M. P. Colombo, A. Stoppacciaro, and R. G. Vile. 1999. Heat shock protein 70 induced during tumor cell killing induces Th1 cytokines and targets immature dendritic cell precursors to enhance antigen uptake. J. Immunol. 63: 1398 –1408. 26. Bell, D., P. Chomarat, D. Broyles, G. Netto, G. M. Harb, S. Lebecque, J. Valladeau, J. Davoust, K. A. Palucka, and J. Banchereau. 1999. In breast carcinoma tissue, immature dendritic cells reside within the tumor, whereas mature dendritic cells are located in peritumoral areas. J. Exp. Med. 190: 1417–1426. 27. Aspord, C., A. Pedroza-Gonzalez, M. Gallegos, S. Tindle, E. C. Burton, D. Su, F. Marches, J. Banchereau, and A. K. Palucka. 2007. Breast cancer instructs dendritic cells to prime interleukin 13-secreting CD4⫹ T cells that facilitate tumor development. J. Exp. Med. 204: 1037–1047.

6216 28. Ghiringhelli, F., P. E. Puig, S. Roux, A. Parcellier, E. Schmitt, E. Solary, G. Kroemer, F. Martin, B. Chauffert, and L. Zitvogel. 2005. Tumor cells convert immature myeloid dendritic cells into TGF-␤-secreting cells inducing CD4⫹ CD25⫹ regulatory T cell proliferation. J. Exp. Med. 202: 919 –929. 29. Chomarat, P., J. Banchereau, J. Avoust, and A. K. Palucka. 2000. IL-6 switches the differentiation of monocytes from dendritic cells to macrophages. Nat. Immunol. 1: 510 –514. 30. Yang, X., J. J. Letterio, R. J. Lechleider, L. Chen, R. Hayman, H. Gu, A. B. Roberts, and C. Deng. 1999 Targeted disruption of SMAD3 results in impaired mucosal immunity and diminished T cell responsiveness to TGF-␤. EMBO J. 18: 1280 –1291. 31. Guo, Z., M. Zhang, H. An, W. Chen, S. Liu, J. Guo, Y. Yu, and X. Cao. 2003. Fas ligation induces IL-1␤-dependent maturation and IL-1␤-independent survival of dendritic cells: different roles of ERK and NF-␬B signaling pathways. Blood 102: 4441– 4447. 32. Wang, L., Q. Liu, Q. Sun, C. Zhang, T. Chen, and X. Cao. 2008. TLR4 signaling in cancer cells promotes chemoattraction of immature dendritic cells via autocrine CCL20. Biochem. Biophys. Res. Commun. 366: 852– 856. 33. Munder, M., H. Schneider, C. Luckner, T. Giese, C. D. Langhans, J. M. Fuentes, P. Kropf, I. Mueller, A. Kolb, M. Modolell, and A. D. Ho. 2006. Suppression of T-cell functions by human granulocyte arginase. Blood 108: 1627–1634. 34. Bronte, V., P. Serafini, A. Mazzoni, D. M. Segal, and P. Zanovello. 2003. Larginine metabolism in myeloid cells controls T-lymphocyte functions. Trends Immunol. 24: 302–306. 35. Makarenkova, V. P., V. Bansal, B. M. Matta, L. A. Perez, and J. B. Ochoa. 2006. CD11b⫹/Gr-1⫹ myeloid suppressor cells cause T cell dysfunction after traumatic stress. J. Immunol. 176: 2085–2094. 36. Ferrone, S., and T. L. Whiteside. 2007. Tumor microenvironment and immune escape. Surg. Oncol. Clin. North Am. 16: 755–774.

TUMOR PROGRAM GENERATION OF DCregs 37. Fricke, I., and D. I. Gabrilovich. 2006. Dendritic cells and tumor microenvironment: a dangerous liaison. Immunol. Invest. 35: 459 – 483. 38. Drake, C. G., E. Jaffee, and D. M. Pardoll. 2006. Mechanisms of immune evasion by tumors. Adv. Immunol. 90: 51– 81. 39. Gajewski, T. F., Y. Meng, and H. Harlin. 2006. Immune suppression in the tumor microenvironment. J. Immunother. 29: 233–240. 40. Dhodapkar, M. V., K. M. Dhodapkar, and A. K. Palucka. 2008. Interactions of tumor cells with dendritic cells: balancing immunity and tolerance. Cell Death Differ. 15: 39 –50. 41. Ghiringhelli, F., P. E. Puig, S. Roux, A. Parcellier, E. Schmitt, E. Solary, G. Kroemer, F. Martin, B. Chauffert, and L. Zitvogel. 2005. Tumor cells convert immature myeloid dendritic cells into TGF-␤-secreting cells inducing CD4⫹CD25⫹ regulatory T cell proliferation. J. Exp. Med. 202: 919 –929. 42. Kuang, D. M., Q. Zhao, J. Xu, J. P. Yun, C. Wu, and L. Zheng. 2008. Tumoreducated tolerogenic dendritic cells induce CD3⫺ down-regulation and apoptosis of T cells through oxygen-dependent pathways. J. Immunol. 181: 3089 –3098. 43. Svensson, M., A. Maroof, M. Ato, and P. M. Kaye. 2004. Stromal cells direct local differentiation of regulatory dendritic cells. Immunity 21: 805– 816. 44. Popov, A., and J. L. Schultze. 2008. IDO-expressing regulatory dendritic cells in cancer and chronic infection. J. Mol. Med. 86: 145–160. 45. Kusmartsev, S., and D. I. Gabrilovich. 2006. Role of immature myeloid cells in mechanisms of immune evasion in cancer. Cancer Immunol. Immunother. 55: 237–245. 46. Bronte, V., E. Apolloni, A. Cabrelle, R. Ronca, P. Serafini, P. Zamboni, N. P. Restifo, and P. Zanovello. 2000. Identification of a CD11b(⫹)/Gr-1(⫹)/ CD31(⫹) myeloid progenitor capable of activating or suppressing CD8(⫹) T cells. Blood 96: 3838 –3846.