The suppressor of cytokine signaling (SOCS) family of negative regulatory proteins is up-regulated in response to several cytokines and pathogen-associated molecular patterns (PAMP) and suppresses cellular signaling responses by binding receptor phosphotyrosine residues. Exposure of bone marrow–derived dendritic cells (BMDC) to 1D8 cells, a murine model of ovarian carcinoma, suppresses their ability to express CD40 and stimulate antigen-specific responses in response to PAMPs and, in particular, to polyinosinic acid:poly-CMP (polyI:C) with the up-regulated SOCS3 transcript and protein levels. The ectopic expression of SOCS3 in both the macrophage cell line RAW264.7 and BMDCs decreased signaling in response to both polyI:C and IFNα. Further, knockdown of SOCS3 transcripts significantly enhanced the responses of RAW264.7 and BMDCs to both polyI:C and IFNα. Immunoprecipitation and pull-down studies show that SOCS3 binds to the IFNα receptor tyrosine kinase 2 (TYK2). Because polyI:C triggers autocrine IFNα signaling, binding of SOCS3 to TYK2 may thereby suppress the activation of BMDCs by polyI:C and IFNα. Thus, elevated levels of SOCS3 in tumor-associated DCs may potentially resist the signals induced by Toll-like receptor 3 ligands and type I IFN to decrease DC activation via binding with IFNα receptor TYK2. [Cancer Res 2008;68(13):5397–404]
Suppressor of cytokine signaling (SOCS) proteins are rapidly induced by cytokines and represent classic feedback inhibitors (1–3). The eight members of the SOCS protein family are SOCS1 to SOCS7 and cytokine-inducible Src homology domain 2 (SH2)-containing protein. Each has a central SH2 domain that targets the SOCS protein to phosphorylated tyrosine residues on activated cytokine receptors or the associated Janus-activated kinase (JAK), and thereby down-regulates signaling. SOCS3 is transiently induced by pathogen-associated molecular patterns and by a wide variety of cytokines, including IFNγ, interleukin (IL)-2, IL-3, IL-6, and IL-10. The rapid induction of SOCS3 by IL-10 is signal transducer and activator of transcription (STAT) dependent, especially in the presence of lipopolysaccharide (LPS; refs. 1, 2, 4). SOCS3 contributes to the inhibition of both macrophage activation and inflammatory cytokine production (1). In vitro studies have also proposed roles for SOCS3 in growth hormone function, IL-2–dependent T-cell proliferation, as well as leptin, IL-10–dependent signaling (1). SOCS3 inhibits IL-1 signaling by targeting the TRAF-6/TAK1 complex (5), whereas SOCS3-mediated inhibition of Epo receptor (EpoR), gp130, and leptin receptor (LeptinR) signaling occurs via competitive inhibition of SHP-2 binding to Y759 of gp130 (6), Y895 of LeptinR (7), and Y401 of EpoR (8), resulting in the blockade of SHP-2–mediated extracellular signal-regulated kinase activation. The high-affinity interaction of SOCS3 with gp130 through its SH2 domain and with JAKs through its kinase inhibitory region provides specificity to inhibition of gp130 signaling (9). Thus, these proteins are critical regulators of cytokine signaling and JAK/STAT pathway.
Dendritic cells (DC) play a central role in innate and adaptive immune responses and are subject to both negative and positive regulation. Toll-like receptor ligands (TLRL) activate DCs to secrete a variety of inflammatory cytokines and promote the expression of costimulatory molecules. Each TLR mediates distinctive responses in association with a different combination of five TIR domain-containing adapters (MyD88, TIRAP/MAL, TRIF, TRAM, and SARM) through the homophilic interaction of TIR domains (10, 11). Typically, the TLRL polyinosinic acid:poly-CMP (polyI:C) signals through TLR3 and TRIF to produce high levels of type I IFN (IFNα; ref. 12). The TLR signaling pathways are negatively regulated by IL-1 receptor (IL-1R)-associated kinase M, SOCS1, MyD88 short, single immunoglobulin IL-1R–related molecule, and ST2 (13). SOCS1 has been linked to the negative regulation of TLR signaling by mediating the degradation of Mal (14). In contrast to TLRLs, tumor and/or tumor-derived factors can decrease the function of DCs. Tumor cells prevent mouse DC maturation induced by TLR ligands (15). Because immune suppression is associated with ovarian cancer, we explored the effect of the murine model of ovarian cancer on the response of murine bone marrow–derived DCs (BMDC) to TLRL. Here, we find that exposure to ovarian tumor cells elevates SOCS3 expression in BMDCs. These tumor-associated DCs also exhibit blunted TLR3 signaling that is associated with binding of SOCS3 to the IFNα receptor tyrosine kinase 2 (TYK2).
Materials and Methods
Mice. Four- to 6-wk-old male C57BL/6 mice (Beijing Animal Center) were maintained in a pathogen-free animal facility at least 1 wk before use. Experiments were performed in accordance with Institutional Guidelines.
Cells, reagents, antibodies, and flow cytometric analysis. Mouse bone marrow–derived DCs (BMDCs) were prepared as described previously (16). Briefly, BMDCs were prepared from bone marrow cells collected by removing the femur bones of mice, cutting off each end, and flushing out the bone marrow with RPMI 1640 using a syringe. The pooled cells were harvested by centrifugation at 600 × g for 10 min and resuspended in 2 mL ACK buffer for 5 min at room temperature to lyse the RBCs. These cells were washed in RPMI 1640 supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin and the DCs were cultured in medium containing 500 units/mL granulocyte macrophage colony-stimulating factor (R&D Systems) for 7 d. The monocyte/macrophage cell line RAW264.7 (American Type Culture Collection) was grown in RPMI 1640 supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. The mouse ovarian surface epithelial cell line 1D8 provided by Katherine F. Roby (University of Kansas Medical Center, Kansas City, KS; ref. 17) was described in our previous article (18). Expression of the costimulatory molecules CD40, CD80, CD86, CD22.2, CD72, and CTLA-4 or the surface antigens CD5, B220, CD11b, and CD11c was not detected on 1D8 (data not shown). Multiple tumors formed within 60 d after i.p. injection of 1 × 107 1D8 ovarian carcinoma cells (data not shown) in syngeneic C57BL/6 mice. For the generation of tumor-associated DCs, BMDCs were first cultured alone for 5 d and then with mouse ovarian surface epithelial tumor cell line 1D8 for further 4 d by staining with FITC-labeled anti-CD80 antibody and sorted by flow cytometry (FACScan, Becton Dickson).
Cells were treated with 25 μg/mL polyI:C (19), 0.1 μg/mL LPS (20), and 100 units/mL IFNα for 2 d and then analyzed by flow cytometry or used as stimulators after loading human papillomavirus (HPV) 16 virus-like particles (VLP). These doses of TLRLs were used in every experiment unless otherwise stated. For each analysis, an isotype-matched control monoclonal antibody (mAb) was used as a negative control. polyI:C, LPS, and IFNα were purchased from Invivogen and PBL Biomedical Laboratories. To assess surface marker phenotypes of tumor-associated DCs and control DCs, the cells were collected in ice-cold PBS and incubated with the following antibodies: FITC-conjugated anti-mouse CD86 (GL1), anti-mouse CD80 (16-10A1), and anti-mouse CD40 (3/23). All of these antibodies were purchased from PharMingen. Cells were then washed twice and resuspended in PBS with 1% paraformaldehyde and 1% FCS and then kept at 4°C before flow cytometric analysis.
Preparation of HPV16 VLP-specific splenocytes and assay of antigen-specific T-cell responses. To determine the VLP-specific T-cell responses, C57BL/6 mice were immunized in the tail vein with HPV16 VLP (10 μg/mouse) or PBS, respectively, on days 0, 7, and 14. On day 24, the pooled splenic cells were collected and deposited as responders. BMDCs from different treatments were used as stimulators after loading with HPV16 VLPs to induce the production of IFNγ (16). VLP-specific splenocytes (2 × 106) were stimulated by HPV16 VLP-loaded DC. The ratio of DC to VLP-specific splenocytes was 1:20. After 3 d, the IFNγ in the supernatant was assayed by capture ELISA.
Construction of pcDNASOCS3 plasmids, design and synthesis of SOCS3 small interfering RNA, and transfection. pcDNASOCS3 plasmids were constructed by cloning mice SOCS3 gene using the pcDNATOPOV5/HIS kit according to the manufacturer's protocol1234
Reverse transcription-PCR analysis. Total cellular RNA was prepared using Trizol reagent (Invitrogen) as recommended by the manufacturer. Reverse transcription-PCR (RT-PCR) was performed by SuperScript one-step RT-PCR with Platinum Taq according to the protocol provided (Invitrogen). Amplification conditions were as follows: cDNA synthesis and predenaturation: perform 1 cycle (50°C for 15 min, 94°C for 2 min); PCR amplification: perform 40 cycles (denature, 94°C for 15 s; anneal, 55°C for 30 s; extend, 72°C for 1 min/kb); and final extension: 1 cycle (72°C for 10 min). The primers used include the following: mSOCS3, 5′-ATGGTCACCCACAGCAAGTTTCC and 5′-AAGTGGAGCATCATAAACTCCC; glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 5′-ATGGTGAAGGTCGGTGTGAACGGATTTGGC and 3′-CATCGAAGGTGGAAGAGTGGGAGTTGCTGT.
ELISA. Commercial sandwich ELISA kits were used for quantitation of IFNγ (Pierce Endogen) and IFNα (PBL Biomedical Laboratories). The absorbance of each of the sample was measured at 450 nm using a SpectraMax 190 ELISA plate reader. Cytokine levels were quantified from two to three titrations using standard curves and expressed in pg/mL.
Immunoprecipitation and Western blot analysis. V5-tagged SOCS3 vector–transfected or only plasmid vector–transfected cells were extracted in hypotonic lysis buffer as previously described. The nuclei were removed by centrifugation at 12,000 × g for 15 min at 4°C, and then the lysate was precleared with 100 μL of protein G-Sepharose and 20 μg/mL isotypic control antibody for 2 h at 4°C. Immunoprecipitation was performed with anti-V5 antibody and protein G-Sepharose for 16 h at 4°C with slow agitation. The immunoprecipitates were washed six times with ice-cold lysis buffer and resolved by 10% SDS-PAGE. Immunoprecipitates then were analyzed by Western blotting using antibodies to anti-JAK1 or TYK2 (Ciba Corning). The membranes were developed using the enhanced chemiluminescence detection system (Amersham Biosciences).
Preparation of glutathione S-transferase-SOCS3-green fluorescent protein fusion protein and pull-down experiments. pGEX-6P-GFP was generated by ligation of pGEX-6P-1 with green fluorescent protein (GFP; Pharmacia) from pGL-2 (Clontech) between EcoRI and SalI. The SOCS3 gene cloned from mouse monocyte cell line RAW264.7 was subcloned into plasmid pGEX-6P-GFP to generate glutathione S-transferase (GST)-SOCS3-GFP or GST-GFP fusion protein using our previously described methods (21, 22). GST-SOCS3-GFP was bound to a GST Trap FF column, and detergent extracts of 2 × 106 DCs in buffer A were passed over the column. After extensive washing, the column was eluted with 50 mmol/L Tris-HCl and 10 mmol/L reduced glutathione at pH 8.0. Eluates were analyzed using antibodies to anti-JAK1 or anti-TYK2 or anti-phospho-TYK2 (Tyr1054/1055) and anti-phospho-JAK1 (Tyr1022/1023) polyclonal antibodies (Cell Signaling Technology).
Murine BMDCs exhibit reduced TLR3 signaling and elevated SOCS3 on exposure to model ovarian cancer. Numerous studies indicate that the presence of tumor cells alters the phenotype and functional potential of antigen-presenting cells. Indeed, coculture of BMDCs with 1D8 ovarian carcinoma cells caused loss of the typical morphologic character of DCs manifested as fewer dendrites and rounding (Fig. 1A). Coculture with 1D8 tumor cells also affects the expression of costimulatory molecules and generation of cytokines (18) and decreased antigen-presenting function of BMDCs (Fig. 1B). The exposure of BMDCs to polyI:C or LPS promoted antigen-specific T-cell activation (Fig. 1B) and surface CD40 expression (Fig. 1C). However, these responses were blunted in each case when the BMDCs had been cocultured with 1D8 tumor cells. This phenomenon was most marked for stimulation with polyI:C, a TLR3 ligand (Fig. 1B and C). Thus, we hypothesized that exposure of BMDCs to 1D8 tumor cell–derived factors caused expression of gene products that blunt TLR intracellular signaling in general and TLR3 signaling in particular. To explore this possibility, we used Affymetrix microarray to compare specific transcript levels in BMDCs that were cocultured with 1D8 tumor cells (ratio of tumor cells to DC was 1:5) or exposed to 25% (v/v) 1D8 tumor cell–conditioned medium versus untreated BMDCs. We examined the microarray data set based on biological process and intracellular signaling cascade and searched for genes whose expression increased more on coculture with 1D8 (live or irradiated) and with 1D8 conditioned medium. Several intracellular signaling molecules fitted these criteria [Supplementary Table S1 and microarray data at Gene Expression Omnibus (GEO);53), including TLR signaling (23, 24). Therefore, we first independently validated the increased SOCS3 expression by BMDCs after exposure to 1D8 tumor cells and conditioned medium by RT-PCR for SOCS3 transcripts and then by Western blotting analysis for SOCS3 protein (Fig. 1D). Both transcription and expression of SOCS3 were up-regulated (Fig. 1D). In vivo evidence showed also that tumor-bearing mice indeed up-regulated SOCS3 in primary DCs from spleen or tumor-draining lymph nodes. Similar to BMDCs on exposure to tumor, these DCs isolated from spleens of tumor-bearing mice or tumor-draining lymph nodes also had decreased antigen-presenting ability and remarkably reduced responses to polyI:C (Supplementary Fig. S1). Notably, overexpression of SOCS1, a family member also up-regulated in tumor-associated DCs and known to suppress TLR3 signaling (25), was not observed (microarray data at GEO; GEO accession number: GSE10741; data not shown), suggesting differential regulation of SOCS family members.
Ectopic SOCS3 expression suppresses TLR3 signaling. To investigate the relevance of SOCS3 up-regulation in TLR signaling, we generated stable transfectants of the monocyte/macrophage RAW264.7 cell line with an expression vector encoding V5-tagged SOCS3 or control vector. Stable expression of V5-SOCS3 was shown by immunofluorescence (Fig. 2A) with a tag-specific mAb. Stimulation of RAW264.7 cells, stably transfected with either vector alone or V5-SOCS3 expression vector, with LPS significantly up-regulated surface CD40 expression (Fig. 2B,3 and B4). However, ectopic V5-SOCS3 expression suppressed the induction of CD40 by polyI:C (Fig. 2B 2). Similar data were obtained for CD80 and CD86, but these effects were weaker than for CD40 (data not shown).
To further characterize the suppressor function of SOCS3, we investigated the effect of SOCS3 transcript knockdown with siRNA on the expression of CD40 by RAW264.7 cells in response to stimulation by polyI:C. Transfection of RAW264.7 cells with SOCS3-targeted but not control siRNA reduced the levels of SOCS3 transcripts and protein without affecting GAPDH transcript or actin levels (Fig. 3A). SOCS3 knockdown promoted the expression of CD40 (Fig. 3B) after stimulation with polyI:C.
Inhibitory effect of SOCS3 on the expression of CD40 is dependent on IFNα. The negative regulation RAW264.7 cell activation by TLRLs was particularly marked in the case of polyI:C by comparison with LPS. Because TLR3 activation results in much higher levels of type I IFN than signaling via TLR2, TLR4, or TLR9 (only TLR4 was shown), we investigate the possible role of type I IFN in inhibitory effect of SOCS3 on the expression of CD40. As shown in Fig. 4A, the control RAW264.7 and SOCS3-transfected RAW264.7 cells produced similar levels of IFNα in response to polyI:C, suggesting that SOCS3 does not suppress TLR3 signaling by reducing IFNα production. Therefore, we sought to determine the influence of elevated or diminished levels of SOCS3 on the induction of surface CD40 after IFNα stimulation of RAW264.7 cells. Ectopic V5-SOCS3 overexpression reduced surface expression of CD40 after stimulation of RAW264.7 cells with IFNα. Conversely, knockdown of endogenous SOCS3 markedly enhanced their levels (Fig. 4). These findings suggest that SOCS3 acts on autocrine/paracrine IFNα signaling induced by polyI:C stimulation to suppress the expression of CD40.
SOCS3 binds to TYK2. The negative regulation of IFNα signaling in RAW264.7 cells and the presence of an SH2-binding domain in SOCS3 imply that SOCS3 may potentially interact with component(s) of the type I IFN pathway to suppress signaling. Although type I IFN comprises several subtypes of IFNα, a single gene encoding IFNβ as well as IFNτ and IFNλ each binds to the type I IFN receptor (IFNR). The type I IFNR is a heterodimer of IFNα receptor 1 subunit (IFNαR1) and IFNαR2 that are associated with the tyrosine kinases TYK2 and JAK1, respectively, to promulgate signaling. To further investigate the mechanism of immunoregulation by SOCS3, we performed immunoprecipitation and Western blotting analysis. The lysates of V5-SOCS3 and empty vector–transduced RAW264.7 cells were immunoprecipitated using V5-specific or isotype-matched mAb (Fig. 5A). Bands corresponding to V5-SOCS3 (∼30 kDa) and TYK2 (∼135 kDa) were observed in silver-stained anti-V5 immunoprecipitates of V5-SOCS3–transfected cells but absent in isotype control immunoprecipitates and V5 immunoprecipitates from vector-transfected RAW264.7 cells. The presence of TYK2 in the anti-V5 but not isotype control antibody immunoprecipitates of V5-SOCS3–expressing cells was confirmed by immunoblotting using TYK2 antibody (Fig. 5B). JAK1 was absent from the anti-V5 immunoprecipitates of V5-SOCS3–expressing cells (Fig. 5B). In pull-down studies, lysate of BMDCs was incubated with GST-SOCS3-GFP or GST-GFP bound on glutathione-Sepharose. TYK2 [but not JAK1 (data not shown)] was detected bound to GST-SOCS3-GFP but not to GST-GFP (Fig. 5C).
To determine whether ligation of the type I IFNR and TYK2 phosphorylation enhances binding to SOCS3, RAW264.7 cells were stimulated by IFNα or as a control IFNγ. IFNα promoted the binding of SOCS3 to phospho-TYK2, whereas IFNγ did not at 5 min after stimulation (Fig. 5D). Interestingly, binding of SOCS3 to phospho-TYK2 promoted by IFNα followed the typical time course of TYK2 phosphorylation (Fig. 5), indicating the binding of SOCS3 to phospho-TYK2. Thus, binding of SOCS3 with phospho-TYK2 might explain suppression of TLR3 and IFNα-induced expression of CD40.
SOCS3 inhibits activation of DCs by polyI:C and IFNα. To investigate the action of SOCS3 on antigen presentation, we raised or lowered the levels of SOCS3 in BMDCs using a nucleofector method by ectopic V5-SOCS3 overexpression or siRNA-mediated knockdown of endogenous SOCS3 transcripts, respectively. These BMDCs were stimulated with either polyI:C or IFNα and cocultured with antigen-specific splenocytes after loading antigen (VLP). T-cell activation was assessed by measuring the production of IFNγ. BMDCs transfected with the V5-SOCS3 expression construct induced low level of production of IFNγ by VLP-specific spleen cells after loading with VLP and stimulation with either polyI:C or IFNα, whereas pcDNA vector only–transfected DC produced levels of IFNγ similar to control BMDC (Fig. 6). Knockdown of SOCS3 with specific but not irrelevant siRNA dramatically enhanced the activation of BMDCs induced by both polyI:C and IFNα (Fig. 6). These findings suggest that SOCS3 is involved in the negative regulation of DC in activation and antigen-presenting function of DC induced by polyI:C and IFNα.
In this study, we found that expression of SOCS3 is up-regulated on exposure to murine ovarian carcinoma 1D8 cells. Although it is unclear what 1D8 tumor-derived factor or factors trigger SOCS3 up-regulation in DCs, several reports describe high-level production of IL-6 and IL-10 in ovarian cancer (26), which are known to induce the expression of SOCS3 (4). The elevated levels of SOCS3 in tumor-associated DCs inhibit the signals induced by TLR3 ligands and type I IFN to decrease DC activation via binding with IFNα receptor TYK2. Thus, our findings suggest that one particular member of the JAK kinase family, TYK2, is an important target of SOCS3-mediated suppression in antigen-presenting cells.
The JAK kinase family comprises at least four different tyrosine kinases (TYK2, JAK1, JAK2, and JAK3) that share significant structural homology with each other, and therefore, it will be interesting to determine the reason for the specificity of SOCS3 for TYK2 as this is in contrast to direct inhibition of all four JAK kinases by SOCS1 (2, 27). SOCS molecules control cytokine receptor signaling by several mechanisms (28), including indirect inhibition of JAKs via binding of SOCS to the cytoplasmic receptor domains (thereby sterically hindering the constitutive JAK binding to the receptor chains), competitive inhibition of STAT binding to receptor chains, and inhibition of downstream signaling pathways. SOCS proteins can also interact with receptor phosphotyrosine through their SH2 domains (8, 29, 30). SOCS3 can bind to the catalytic site of JAK family kinases and inhibit their activity (8, 31). We here found that the interaction of SOCS3 with TYK2 reduced both the polyI:C- and IFNα-induced activation of DCs. Thus, SOCS3 induced by ovarian carcinoma cells and tumor supernatants in antigen-presenting cells may suppress multiple pathways critical to the induction of tumor-specific immunity, including antigen presentation and activation by TLRLs and type I IFN.
Type I IFNs have been widely used for the treatment of chronic infections, including hepatitis C and papillomavirus, and some cancers, such as melanoma. The effects of type I IFNs on DC function are contradictory, as some studies show that they promote the differentiation/activation of DCs (32, 33), whereas others find inhibitory effects (34–37). Perhaps, inhibition of type I IFN signaling by SOCS3 could offer an explanation for some of these contradictions. Type I IFNs have been shown to act as a maturation factor for DCs (36, 37) when SOCS3 is presumably at a low level. Notably, the IFNα-induced DC activation and maturation was enhanced in the presence of anti-IL-10 antibody (38). Because IL-10 is an effective inducer of SOCS3, enhanced maturation in the presence of anti-IL-10 may be the result of lower SOCS3 levels.
SOCS3 is a potent negative regulator of multiple proinflammatory pathways (39). Macrophage cytokine production and osteoclast generation were markedly enhanced in the absence of SOCS3. Enhanced T lymphocyte activation in the absence of SOCS3 may drive disease and also promote osteoclastogenesis. SOCS3 in tumor-associated DCs inhibits the signals induced by TLR3 ligands and type I IFN to decrease DC activation, suggesting that SOCS3 may play an important role in tumor-mediating immune suppression also.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Grant support: Nankai University, National Science Foundation of China grant 30771967, The Ministry of Science and Technology grant 06C26211200695, 863 grant 2006AA020502, Tianjin grant 07JCZDJC03300, and Tianjin Creative grant 06ZHCXSH04800.
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.
We thank Dr. RongZhe Che for his help with microscopy and Dr. Conover Talbot, Jr. (Johns Hopkins University School of Medicine) for his help in microarray data.