In this study we show that activation of STAT pathways is developmentally regulated and plays a role in dendritic cell (DC) differentiation and maturation. The STAT6 signaling pathway is constitutively activated in immature DC (iDC) and declines as iDCs differentiate into mature DCs (mDCs). However, down-regulation of this pathway during DC differentiation is accompanied by dramatic induction of suppressors of cytokine signaling 1 (SOCS1), SOCS2, SOCS3, and cytokine-induced Src homology 2-containing protein expression, suggesting that inhibition of STAT6 signaling may be required for DC maturation. In contrast, STAT1 signaling is most robust in mDCs and is not inhibited by the up-regulated SOCS proteins, indicating that STAT1 and STAT6 pathways are distinctly regulated in maturing DC. Furthermore, optimal activation of STAT1 during DC maturation requires both IL-4 and GM-CSF, suggesting that synergistic effects of both cytokines may in part provide the requisite STAT1 signaling intensity for DC maturation. Analyses of STAT1−/− DCs reveal a role for STAT1 in repressing CD86 expression in precursor DCs and up-regulating CD40, CD11c, and SOCS1 expression in mDCs. We further show that SOCS proteins are differentially induced by IL-4 and GM-CSF in DCs. SOCS1 is primarily induced by IL-4 through a STAT1-dependent mechanism, whereas SOCS3 is induced mainly by GM-CSF. Taken together, these results suggest that cytokine-induced maturation of DCs is under feedback regulation by SOCS proteins and that the switch from constitutive activation of the STAT6 pathway in iDCs to predominant use of STAT1 signals in mDC is mediated in part by STAT1-induced SOCS expression.
Dendritic cells (DCs)3 play critical roles in initiating and modulating immune responses and are implicated in mechanisms of peripheral tolerance (1, 2). In most tissues, DCs are present in the immature DC (iDC) state. Although the iDC is adept at capturing and processing Ags, it does not stimulate naive T cells because it lacks the requisite level of accessory signals for T cell activation (1, 2). After the capture of Ag, iDC undergo extensive transformations, down-regulate their Ag-capturing capabilities, and up-regulate the expression of chemokine receptors that allows them to home to regional lymphoid organs (2). Terminally differentiated or mature DCs (mDCs) express high levels of costimulatory molecules (CD40, CD80, and CD86) and MHC class I and II molecules and secrete diverse immune modulatory molecules that stimulate naive Th cells to differentiate into Th1 or Th2 effector cells (1, 3). In addition to their role in orchestrating essential aspects of the adaptive immune response, DCs respond to microbes by producing proinflammatory cytokines (IL-12, IFNαβ, and IFN-γ) that activate innate lymphocytes (CD8+ and NKT cells) to directly kill infected cells. Thus, DCs provide a link between innate and adaptive immune systems (3).
Major interest in DC biology stems from the expanding use of DCs as adjuvant to enhance host immunity (4, 5, 6, 7). Injection of Ag-loaded mDCs has been shown to rapidly enhance Ag-specific CD4+ and CD8+ T cell immunity in humans, making mDCs attractive vectors for boosting cellular immunity against infectious or neoplastic diseases (8, 9, 10). In contrast, injection of healthy human volunteers with Ag-pulsed iDC induces Ag-specific, IL-10-producing, immunoregulatory cells and inhibits Ag-specific CD8+ T cell effector function in freshly isolated T cells, leading to significant interest in iDC as vectors for delivering immunoregulatory peptides to attenuate immunopathogenic responses of allergic and autoimmune diseases (6, 11). Although these studies clearly demonstrate the feasibility of using DC vaccines in the treatment of diverse human diseases, they also urge caution in the purity of the DC preparations used, as the presence of even low numbers of contaminating mDCs or iDC may provoke undesirable inflammatory or tolerogenic responses, respectively. DCs used in gene therapy protocols are derived by stimulating monocytes or bone marrow-derived hemopoietic progenitor cells (HPCs) with GM-CSF plus IL-4, and functional maturation of the cells in vitro is readily induced by a variety of microbial products, including LPS or cytokines such as IL-1, GM-CSF, and TNF-α (12, 13, 14). However, IL-10 blocks the maturation process, and addition of IFN-γ to IL-4- plus GM-CSF-stimulated monocytes switches their differentiation from DCs to macrophages, underscoring the plasticity of the differentiation process and the critical role played by cytokines in myeloid cell lineage commitment (15). Thus, the generation of large quantities of highly enriched mDC or iDC populations for use in DC vaccines requires in-depth understanding of how cytokine-induced developmental signals control checkpoints of DC maturation.
A number of cytokines involved in DC maturation transduce their extracellular signals to the nucleus through activated STAT proteins (16, 17) and the duration or intensity of the cytokine-induced signal is under feedback regulation by a newly described eight-member family of intracellular proteins called suppressors of cytokine signaling (SOCS) (18, 19). SOCS proteins are characterized by the presence of an Src homology 2 domain and a C-terminal conserved domain called the SOCS box, and their inhibitory effects derive from direct interaction with cytokine receptors and/or Janus kinases (JAKs), thereby preventing recruitment of STATs to the signaling complex (18, 19). SOCS proteins have recently been shown to modulate macrophage effector functions and negatively regulate LPS-induced macrophage activation (20, 21, 22), suggesting that they may also be involved in the regulation of DC differentiation and/or maturation.
In this report we generated DCs at three distinct stages of differentiation and have used them to investigate the roles of SOCS and JAK/STAT pathways in DC differentiation and maturation. We show that the cytokines, IL-4 and GM-CSF, activate distinct and overlapping STAT pathways in differentiating and mature DCs. Activation of STAT6 signaling is not detected in freshly isolated precursor DCs (pDCs), but is constitutively activated in primary iDC and progressively declines as the cells differentiate into mDCs. In addition, IL-4 was found to synergize with GM-CSF to induce significant activation of STAT1, and analysis of STAT1-null mice reveals that the STAT1 signal transduction pathway is essential for complete maturation of DC. We further show that IL-4- and GM-CSF-induced activation of JAK/STAT signal transduction pathways in DCs is accompanied by significant up-regulation of SOCS1, SOCS2, SOCS3, and cytokine-induced SH2 protein (CIS), suggesting that the differentiation and maturation of DC may be regulated in part by SOCS proteins.
Materials and Methods
Generation of DCs
Bone marrow (BM)-derived DCs were generated as previously described (13). Briefly, BM was obtained from tibias and femurs of 8- to 12-wk-old C57BL/6 wild-type (WT) or STAT1−/− (gift from D. Levy, New York University, New York, NY) mice. ScaI+ BM cells were selected using ScaI MultiSort microbeads and Midi-MACS separation columns (Miltenyi Biotec, Auburn, CA). ScaI+ cells were lineage depleted by labeling with FITC-conjugated anti-CD4, anti-CD8, anti-CD11b, anti-Gr-1, and anti-B220 (BD PharMingen, San Diego, CA); binding to anti-FITC microbeads (Miltenyi Biotec), and passing them through MiniMACS separation columns (purity, 95 ± 2% Lin−ScaI+). Lin−ScaI+ cells were grown as suspension cultures in Teflon jars (37°C, CO2) with IMDM (Life Technologies, Grand Island, NY), supplemented with 10% FBS (HyClone Laboratories, Logan, UT) and the following recombinant murine cytokines: 100 ng/ml stem cell factor, 20 ng/ml IL-3, 50 ng/ml M-CSF, 5 ng/ml GM-CSF, and 25 ng/ml FMS-related tyrosine kinase-3 (FLT3) ligand (all from R&D Systems, Minneapolis, MN). DCs derived after 9 days in this medium are referred to as pDCs. Immature DCs were generated by switching pDCs to a medium containing GM-CSF (100 ng/ml) plus IL-4 (1500 U/ml) and propagating the cells for 3 additional days. Mature DCs were obtained after overnight stimulation of iDCs with LPS (LPS; Escherichia coli 026:B6; used at 2 μg/ml). For FACS analysis, cells were preincubated with purified anti-mouse CD16/CD32 Fc Block (BD PharMingen) before staining with the designated mAb or isotype control for 30 min, then washed with buffered normal saline. For staining, labeled CD11b, CD11c, I-Ab, CD86, or CD40 and their corresponding isotype control Abs (BD PharMingen) were used. Analysis was performed on the FACSort (BD Biosciences, San Jose, CA) using CellQuest software (BD Biosciences).
Quantitative RT-PCR analysis
RNA (10 μg), SuperScript II reverse transcriptase (Life Technologies, Gaithersburg, MD), and oligo(dT)12–16 were used for first-strand cDNA synthesis as previously described (23, 24). A negative control reaction without reverse transcriptase was performed for each RNA sample. RNA samples were normalized to 18S rRNA using the TaqMan Ribosomal RNA Control Reagents kit (Applied Biosystems, Foster City, CA). Real-time 5′-nuclease fluorogenic RT-PCR analysis was performed on an ABI 7700 (Applied Biosystems) or ICycler iQ Real-Time PCR (Bio-Rad, Hercules, CA) Sequence Detection System with the following primers: SOCS1, 5′-ACCTTCTTGGTGCGCGAC-3′ and 5′-AAGCCATCTTCACGCTGAGC-3′; SOCS2,5′-GGTTGCCGGAGGAACAG TC-3′ and 5′-GAGCCTCTTTTAATTTCTCTTTGGC-3′; SOCS3, 5′-CCTTCAGCTCCAAAAGC GAG-3′ and 5′-GCTCTCCTGCAGCTTGCG-3′; CIS, 5′-CCAGCCATGCAGCCCTTA-3′ and 5′-C GTCTTGGCTATGCACAGCA-3′; and β-actin, 5′-CAAGTCATCACTATTGGCAACGA-3′ and 5′-CCCAAGAAGGAAGGCTGGA-3′. The hybridization probes are: SOCS1, 6FAM-TCGCC AACGGAACTGCTTCTTCG-TAMRA; SOCS2, 6FAM-CGCGTCTGGCGAAAGCCCTG-TAMRA; SOCS3, 6FAM-CCAGCTGGTGGTGAACGCCGT-TAMRA; CIS, 6FAM-CCCAG AGGAAGTGACAGAGGGAGACCCC-TAMRA; and β-actin, 6FAM-CGGTTCCGATGCCCTGA GGCTC-TAMRA. PCR parameters are as recommended for the TaqMan Universal PCR Master Mix kit (Applied Biosystems). SOCS copies per cell were calculated by extrapolation from standard curves generated using SOCS plasmid cDNAs as previously described (23). SOCS-specificcDNAs were provided by Drs. D. Hilton (Walter and Eliza Hall Institute, Melbourne, Australia) and H. Young (National Institutes of Health, Bethesda, MD). SOCS1, SOCS2, SOCS3, or CIS plasmid cDNA was purified by two sequential cesium chloride bandings, and triplicate samples of 10-fold serial dilutions of cDNA were assayed and used to construct the standard curves. It should be emphasized that standard curves generated from the SOCS cDNA dilution series show excellent linearity, indicating the precise quantitative relationship between cDNA copy number and fluorescence signal intensity within the dynamic range of the assay (data not shown).
Northern blot analysis
Northern blot analysis was performed with 20 μg of RNA as previously described (24). The integrity and comparability of RNA preparations used for analysis were verified by agarose-formaldehyde gel electrophoresis; comparable amounts of 18S and 28S ribosomal RNAs were detected for all the RNA preparations used. Mouse SOCS1, SOCS2, SOCS3, and CIS cDNAs were used as hybridization probes.
Western blot analysis
Primary DC isolated from 6- to 8-wk-old C57BL/6 mouse bone marrows were solubilized in 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 2 μM leupeptin, 2 μM pepstatin, 0.1 μM aprotinin, 1 mM [4-(2-aminoethyl)----benzenesulfonyl-fluoride, hydrochloride] (AEBSF), 0.5 mM PMSF, and 1 μM E-64 on ice. Protein levels were determined by bicinchoninic acid assay as recommended by manufacturer (Pierce, Rockford, IL). Samples were heated for 10 min at 95°C in 1× sample buffer and electrophoresed in 4–20% SDS-PAGE. Gels were electroblotted onto polyvinylidene fluoride membranes, blocked with 5% nonfat milk, and probed with Abs specific for STAT1, STAT6, and β-actin, (Santa Cruz Biotechnology, Santa Cruz, CA); SOCS1, SOCS3, and pSTAT5 (Zymed Laboratories, San Francisco, CA); or phosphorylated STAT1 (pSTAT1), pSTAT6 (Cellular Signaling Technology, Beverly, MA). Preimmune serum was also used in parallel as a control, and signals were detected with HRP-conjugated secondary F(ab′)2 Abs (Zymed Laboratories) using the ECL system (Amersham Pharmacia Biotech, Arlington Heights, IL).
Nuclear extracts were prepared using buffer containing the following protease inhibitors: 2 μM leupeptin, 2 μM pepstatin, 0.1 μM aprotinin, 1 mM AEBSF, 0.5 mM PMSF, and 1 μM E-64 as previously described (25). Protein levels were determined by the bicinchoninic acid method as recommended, and extracts were stored at −70°C until use. Nuclear extract (5 μg) in binding buffer (20 mM HEPES (pH 7.9), 50 mM KCl, 10% glycerol, 0.5 mM DTT, and 0.1 mM EDTA) containing 0.14 μg/μl poly(dI-dC) was incubated on ice for 20 min. Labeled dsDNA probe was then added and incubated for an additional 15 min at room temperature. The probes used are the IFN-γ activation site (GAS) motif from the FcγRI promoter (5′-AGCTTGTATTTCCCAGAAAAGGGATC-3′) (26, 27), and the GAS motif from the human Cε regulatory region (5′-AGTCAAGACCTTTTCCCAAG AAATCTATC-3′) (28, 29). The double-stranded oligonucleotides were labeled by fill-in reaction using Klenow polymerase (Invitrogen, Carlsbad, CA) with either [α-32P]dATP or [α-32P]dGTP (3000 Ci/mmol; Amersham Pharmacia Biotech). Samples were electrophoresed in 5% polyacrylamide gel in 0.25× Tris-borate-EDTA buffer. For supershift assays, the indicated Ab was added to the binding buffer containing nuclear extract mixture and preincubated on ice for 10 min. 32P-labeled probe was then added, and the entire mixture was incubated for an additional 20 min on ice before electrophoresis. Gel-shift grade anti-mouse STAT1, STAT3, STAT4, STAT5a, STAT5b, or STAT6 polyclonal Abs (Santa Cruz Biotechnology) were used.
Myeloid pDCs and iDCs differentially respond to LPS
The pDCs were generated from BM-derived Lin−ScaI+ HPCs as previously described (13). Phenotypic characterization of pDCs by FACS analyses shows that the cells are CD11bbright and CD11cmed, features associated with myeloid cells of the DC lineage (Fig. 1) (30). The CD40low, CD86low, surface MHC IIlow, cytoplasmic MHC IIhigh immunophenotype of the cells further indicates that the pDC are akin to immature DC (Fig. 1) (30). Precursor DCs remain arrested at the pDC developmental stage, and the expression of costimulatory molecules such as CD40 and CD86 is not induced by LPS, IL-4, and low concentrations of GM-CSF (Fig. 1, thick lines) (13). However, pDC are readily induced to differentiate into iDCs by switching the cells to medium containing high concentrations of GM-CSF plus IL-4 and culturing the cells for 3 additional days; the resulting iDCs do up-regulate the expression of costimulatory molecules (CD40, CD86) and surface MHC II molecules in response to LPS stimulation. It is of note that switching the pDCs to a medium containing high concentration of GM-CSF alone also allows differentiation into cells that are immunophenotypically similar to iDC. However, the latter cells are marginally responsive to maturation with LPS (data not shown), suggesting that IL-4 is required for DC maturation in our experimental system. These results indicate that the pDC described in this study is phenotypically and functionally distinct from the iDC reported by others and represents a much earlier stage of DC differentiation.
STAT signaling pathways are differentially activated in differentiating and mDC
In view of the requirement for both IL-4 and GM-CSF for differentiation and functional maturation of DC, we examined the signal transduction pathways that underlie their effects. Binding of IL-4 or GM-CSF to its cognate receptor on DCs leads to the recruitment of one or more members of the STAT family to the receptor where they are activated by phosphorylation of a critical tyrosine residue in their C-terminal region (16). The activated STATs translocate into the nucleus where they interact with GAS elements in the promoter of target genes and activate or repress transcription of the gene (16). IL-4 mainly activates STAT6, whereas GM-CSF can activate a number of STAT proteins, including STAT1, STAT5, and STAT6, in mouse mDCs (12), and the particular STAT member used for transducing the GM-CSF signal can vary depending on the cell type and its physiological state. To examine the STAT member(s) that mediates GM-CSF or IL-4 activities in differentiating and mature DCs, we performed Western blot analysis using whole-cell extracts from freshly generated pDC, iDC, or mDC and Abs specific for tyrosine-phosphorylated STAT1 (pSTAT1), pSTAT5, or pSTAT6 proteins. Fig. 2 shows the results of a representative experiment in which we detected constitutive activation of STAT5 in DCs at all three developmental stages. However, similar analysis revealed a differential pattern of STAT1 and STAT6 activation by these cells; whereas the STAT1 signaling pathway is constitutively activated in the terminally differentiated mDC but in neither pDC nor iDC, constitutive activation of STAT6 is detected only in iDC (Fig. 2). In addition, Western blot analysis using STAT1 or STAT6 Abs shows that STAT1 protein is less in pDCs, suggesting that the expression of STAT1 may increase during the maturation process (Fig. 2). In contrast, the STAT6 level is comparable among pDCs, iDCs, and mDCs, suggesting that detection of pSTAT6 in iDCs and not in pDCs or mDCs derives from specific activation of STAT6 in iDCs. This mutually exclusive pattern of STAT1 or STAT6 use thus suggests that these signaling pathways may play a role in regulating genes that control checkpoints of DC maturation.
STAT6 is preferentially activated in differentiating DCs and declines as the cells mature
To determine whether the constitutive activation of STAT6 in iDCs is induced by IL-4 or GM-CSF, we analyzed the level of pSTAT6 in pDC or iDC after stimulation with either cytokine. It should be noted that unlike the experiments described in Fig. 2 where the freshly isolated cells are analyzed directly, in these experiments the freshly isolated pDCs, iDCs, and mDCs were washed to remove any exogenous cytokines and then starved for 2 h in serum-free medium before restimulation with cytokines; this reduces background due to autophosphorylation by serum proteins or other stimuli in culture and allows for better examination of the response to cytokine signaling. In Fig. 3,A, we show that activation of STAT6 occurs only after stimulation with IL-4, but not in cells treated with GM-CSF or in control cells. We also characterized the intensity and duration of STAT6 signals during the course of DC differentiation by analyzing interaction of the activated STAT6 with its cognate DNA motif in the nucleus of cytokine-activated DCs. Consistent with previous reports (31, 32), we show by EMSA that activated STAT6 binds to the N4 palindromic Cε-GAS motif (see the C1 arrow in Fig. 3,B). In line with our Western blot results (Fig. 3,A), the retarded DNA-protein complex is observed only in cells treated with IL-4 and not in control cells or cells stimulated with GM-CSF. We confirm by supershift analysis that the retarded band is induced by STAT6 binding (Fig. 3 C). In some experiments the cells were stimulated with both GM-CSF and IL-4 to assess possible synergistic or antagonistic effects of two cytokines on STAT6 signaling during DC maturation. Costimulation of the cells with GM-CSF and IL-4 had very little or no effect, although a slight synergistic effect on STAT6 activation was observed in mDC. These results suggest involvement of STAT6 signaling pathways in differentiation of myeloid DC progenitor to iDC. Diminution of STAT6 signaling as the cells differentiate from pDC to mDC further suggests that DC may be less reliant on STAT6 signaling pathways as they mature or that inhibition of this signaling pathway is essential for maturation of DC.
STAT1 activation is mediated by synergistic interaction between IL-4 and GM-CSF
To determine whether STAT1 is involved in transducing GM-CSF and/or IL-4 signals in DCs, we analyzed the levels of pSTAT1 in whole-cell extracts derived from cytokine-stimulated iDC or mDC. In contrast to iDC, IL-4 or GM-CSF induces detectable activation of STAT1 in mDC (Fig. 4,A). Nuclear extracts from control or cytokine-treated pDC, iDC, or mDC were subsequently analyzed by the relatively sensitive EMSA to determine whether our inability to detect activated STAT1 in pDC or iDC is due to the low sensitivity of the Western blot assay. As shown in the gel-shift analysis using the oligonucleotide probe containing the FcγRI GAS element, IL-4 or GM-CSF induces a retarded DNA-protein complex in all three phenotypic DC stages (Fig. 4,B). The results of supershift analysis on the extracts further reveal that the retarded band consists of STAT1-DNA complexes (Fig. 4,C). Although GM-CSF and IL-4 appear to be equivalent in their ability to activate STAT1 signals in pDC and iDC, in three independent experiments, IL-4 alone was found to be less effective in mDC (Fig. 4 B). This observation is consistent with diminution of STAT6 signaling in mDC and underscores the possibility that terminally differentiated DCs are less reliant on IL-4-induced signals. In addition, the data reveal that both IL-4 and GM-CSF are required for optimal activation of STAT1, particularly in early differentiating DCs.
Functional maturation of DCs requires STAT1
The constitutive activation of the STAT1 signaling pathway in mDC (Fig. 2) and its induction by GM-CSF and IL-4 at all stages of DC development (Fig. 4,B) suggest that this signal transducer that regulates growth and differentiation in lymphocytes (33, 34) may also be an important regulator of DC differentiation and/or maturation. To assess the role of STAT1, if any, in DCs, we isolated HPCs from the BM of WT and STAT1-null mice under the conditions described above for generating pDC, iDC, and mDC. DCs from STAT1-null mice are similar to WT cells in morphology and expression of surface markers (CD11b, CD11c) characteristic of myeloid cells (Fig. 5). However, pDCs from the STAT1-null mice have a significantly higher level of constitutive CD86 expression than those from WT cells, suggesting that this costimulatory is negatively regulated by STAT1 at early stages of DC development. After stimulation of iDC with LPS, a significant increase in CD86 expression was observed in both WT and STAT1−/− iDCs, suggesting that STAT1 signaling is not required for CD86 expression and that one of the effects of LPS on maturing DCs is to override the inhibitory effects of STAT1 on CD86 expression. Comparisons of WT with STAT1−/− iDC further reveal that STAT1 is required for optimal expression of CD40 and that the effects of the LPS maturation signal is directed at iDCs. In contrast to either CD86 or CD40 expression, regulation of MHC class II expression appears to be independent of STAT1.
DC maturation is accompanied by transcriptional up-regulation of SOCS genes
Cytokine signaling pathways in hemopoietic cells are under feedback regulation by SOCS and have recently been shown to inhibit LPS-induced macrophage activation and cytokine-induced CD40 expression in macrophage (20, 21, 35). To determine whether STAT pathways of DCs are also under feedback regulation by SOCS, we examined whether SOCS expression increases during the course of DC differentiation. We show that SOCS1, SOCS3, and CIS genes are constitutively expressed at relatively low levels in pDCs and that the levels of these SOCS members are dramatically up-regulated as the cells differentiate into mDCs (Fig. 6,A). The preferential expression of SOCS2 in iDC is particularly intriguing and may provide a molecular marker of the iDC phenotype. This pattern of SOCS expression has been detected in three independent real-time quantitative RT-PCR assays and by Northern blot analysis (data not shown). The observed changes in SOCS expression are also reflected by changes in SOCS protein levels as DCs differentiate from pDCs to mDCs (Fig. 6 B). These results suggest that signal transduction pathways induced by cytokines during DC maturation are under feedback regulation by SOCS.
SOCS1 and SOCS3 expression is differentially regulated by IL-4 and GM-CSF
To characterize the effects of IL-4 or GM-CSF signaling on induction of SOCS expression in DCs, we analyzed SOCS mRNA levels in pDC, iDC, or mDC stimulated with either cytokine. IL-4 was found to be a more potent inducer of SOCS1 and SOCS2 expression, whereas GM-CSF was more important for SOCS3 induction (Fig. 7,A). In contrast, IL-4 or GM-CSF induced CIS expression to a similar extent, and optimal expression of CIS required both cytokines (Fig. 7,A). As a control for these analyses, we analyzed another SOCS member that is normally not inducible by cytokines. As indicated in Fig. 7,B, SOCS5 is constitutively expressed in DCs, and its expression is not induced by cytokines, such as IL-4 and GM-CSF, that promote DC differentiation. Because SOCS1 is dramatically induced during the course of DC differentiation, and the STAT1 pathway is essential for complete maturation of DC, we examined whether STAT1 is involved in regulating SOCS1 expression in DC. Whole-cell extracts from WT or STAT1−/− iDCs stimulated with IL-4 or GM-CSF were therefore analyzed for SOCS1 protein expression by Western blotting. In concert with our real-time RT-PCR results (Fig. 7,A), SOCS1 is induced in WT iDCs by IL-4, but not GM-CSF (Fig. 7,C). As described above, STAT6 signaling decreases as DCs differentiate to functionally mature DCs, and this coincides with a concomitant increase in STAT1 activation. We therefore examined the effects on STAT6 activation in STAT1-null mDCs. In contrast to WT cells (Fig. 2), STAT6 is constitutively activated in mDCs from STAT1−/− mice (Fig. 7,D). Moreover, the STAT1−/− mDCs exhibit a high sensitivity to IL-4 stimulation, as evidenced by the significant expression of induced expression of pSTAT6 (Fig. 7 D).
In this study we have used DCs at three distinct stages of differentiation to characterize the role of cytokine-induced signal transduction pathways during maturation of DC. The pDCs, iDCs, or mDCs analyzed are grown in suspension (13), making them ideally suited for studying signaling mechanisms; nonspecific signals normally induced by adherence of DCs to standard tissue culture plastic flasks or resulting from application of mechanical stress to dislodge the cells from the bottom of the flasks are eliminated and do not confound analysis of cytokine-dependent maturation stimuli. In this experimental system both IL-4 and GM-CSF are required for functional maturation of DCs and mediate their effects through activation of the JAK/STAT signal transduction pathway. Whereas IL-4 mediates its effects primarily through STAT6, GM-CSF activates STAT1 and STAT5, and the pattern of STAT use varies during the course of DC differentiation. For example, the STAT5 signal is constitutively activated at all stages of DC maturation, whereas STAT1 and STAT6 pathways are preferentially used at different stages of DC differentiation and maturation (Fig. 2). It is interesting that most cytokine mixtures used in generating DCs in vitro include GM-CSF and/or FLT3 ligand, two factors that mediate their effects through activation of STAT5 (36, 37, 38). For example, one or more members of the γC family cytokines (IL-2, IL-7, IL-13, or IL-15) have been used in combination with GM-CSF, and these cytokines predominantly signal through STAT5 and STAT3 (39, 40). In contrast, IL-6, a gp130 family cytokine that activates STAT3 is able to induce DC differentiation from BM cells in conjunction with FLT3 ligand in the absence of GM-CSF (41). The results of these studies, taken together with our current data showing that STAT6 complements STAT5 and STAT1 in inducing DC differentiation and maturation, underscore the central role played by STAT signaling pathways in DCs. An unresolved issue in DC development is whether several myeloid and lymphoid DC subtypes described in the literature represent distinct lineages or result from the different cytokine mixtures used in their derivation. In the context of our current data, it may well be that different cytokine mixtures induce distinct patterns of STAT activation and that the DC phenotype is in part determined by its repertoire of STAT signaling pathways and how these pathways are integrated into the DC developmental program.
We show that STAT6 is constitutively activated in iDC, but not in pDC or mDC (Fig. 2), and in contrast to a previous report (12), our Western blot (Fig. 3,A) and gel-shift (Fig. 3, B and C) data indicate that IL-4 is the primary inducer of STAT6 activation, not GM-CSF. Detection of STAT6 phosphorylation in the other study may be attributed to the FCS present in the medium during stimulation, because in three independent experiments we could not detect activated STAT6 activation in cells stimulated with GM-CSF in serum-free medium. The absence of constitutive STAT6 use in freshly isolated mDCs is particularly surprising considering that mDCs are also cultured in medium containing both IL-4 and GM-CSF. This suggests that one of the effects of pathogens and bacterial extracts, such as LPS, that promote functional maturation of DCs is to down-regulate the STAT6 signaling pathway in terminally differentiated DCs. Inhibition of the STAT6 signaling pathway during the course of DC differentiation is further supported by EMSA results showing progressive diminution of STAT6-induced transcriptional activities as DCs mature (Fig. 3). These results suggest that the IL-4-induced STAT6 pathway may be involved in differentiation from the pDC to the iDC stage but may not be essential for differentiation of iDCs into mDC.
In contrast, STAT1 activation is readily detected by EMSA at all stages of DC development. However, we could detect the tyrosine-phosphorylated STAT1 protein mainly in mDCs, suggesting that a more robust STAT1 signaling pathway is required in functionally mature DCs. The relatively lower level of STAT1 activation in pDC and iDCs may derive from fundamentally different mechanisms by which STAT1 signaling is regulated by cytokines in differentiating and terminally differentiated DCs; although activation of STAT1 in mDCs is mediated primarily by GM-CSF in combination with LPS, optimal activation of this pathway in pDC requires both IL-4 and GM-CSF (Fig. 4,B). Thus, a threshold signal may be required to initiate differentiation into mDCs, and cross-talk between IL-4 and GM-CSF may provide the requisite level of STAT1 signaling intensity to immature DCs. Although the mechanism that sustains high levels of STAT1 activation in mDCs is not clear at this time, we cannot rule out a mechanistic role for autocrine DCs activators such as types I/II IFNs, IL-12, and IL-23 (42, 43). For example, pathogenic organisms, LPS, lipoteichoic acid, peptidoglycan, or bacteria (CpG) DNA induce DCs to secrete IL-12 via Toll-like receptor signaling, and this leads to induction of T-bet and IFN-γ (44, 45, 46). IL-12 is positively regulated by IFN-γ, and it, in turn, participates in a positive feedback loop that promotes IFN-γ secretion (44). Thus, an autocrine loop involving IL-12/STAT4, T-bet, and IFN-γ may contribute to the sustained STAT1 activation in mDCs. The results of analyses of freshly isolated HPCs from STAT1−/− mice underscore the functional relevance of STAT1 signaling in DC differentiation and reveal that its regulatory effects are complex and depend on the developmental stage of the DC. Although STAT1 inhibits CD86 expression in cells at the pDC stage, its inhibitory effects are overridden in mDCs, presumably by maturation signals, such as LPS (Fig. 5,B). In contrast, the STAT1 signaling pathway appears to have no effect on CD40 expression at the pDC stage, but is required for optimal expression of CD40 in mDCs (Fig. 5 C). Contrary to the regulatory effects on CD86 and CD40 expression, STAT1 has no effect on MHC class II expression.
SOCS proteins are differentially expressed in Th1 or Th2 cells and have been implicated in Th cell lineage commitment and maintenance (23, 24). A potential role for SOCS in DC differentiation is suggested by results showing that pDCs express low levels of SOCS mRNAs/proteins and dramatically up-regulate the expression of SOCS1, SOCS2, SOCS3, and CIS as they differentiate into mDCs (Fig. 6). Although the target of SOCS repression in DCs is unknown, SOCS1 has been shown to inhibit cytokine induction of CD40 expression in macrophages, suggesting that this critical costimulatory molecule required for Ag presentation by APCs is under feedback regulation by SOCS (35). However, our data do not support a direct role of SOCS proteins in CD40 regulation in DCs, because constitutive expression of high levels of SOCS1, SOCS3, and CIS mRNA does not prevent expression of CD40 in mDCs (Figs. 5 and 6). Instead, our data suggest that the target of SOCS inhibition in DC is the STAT6 signaling pathway because 1) constitutive levels of SOCS mRNAs or proteins inversely correlate with the ability to activate STAT6 signaling pathway; and 2) the diminution of STAT6 signaling during DC maturation is associated with dramatic up-regulation of SOCS expression. It is also striking that differentiating DCs (pDC and iDC) that robustly activate the STAT6 pathway do not express CD40, whereas down-regulation of STAT6 signaling in mDC (Fig. 3) correlates with up-regulation of CD40 expression. The switch from constitutive activation of the STAT6 pathway in iDCs to predominant use of STAT1 signals in mDC is particularly intriguing and suggests a possible role for STAT1 in the induction of SOCS expression. This possibility is suggested by our results showing that DCs from STAT1−/− mice are compromised in their ability to up-regulate SOCS1 expression in response to cytokine stimulation. We further show that the inducer of SOCS1 expression in this DC model is IL-4 and not GM-CSF, and this is consistent with IL-4-induced STAT6 signaling in DCs as target of feedback regulation by SOCS.
The common use of the STAT signaling pathways by most factors that influence DC development coupled with mounting evidence that IL-12, IL-4, GM-CSF, IL-6, IL-23, and LPS induce SOCS expression suggest that in addition to their role in DC maturation, SOCS proteins may impact on other aspects of innate and adaptive immune responses. For example, IL-12/STAT4 signaling is under feedback regulation by SOCS in hemopoietic cells, and sustained production of IL-12 by DCs up-regulates the expression of T-bet, IL-12Rβ, and IFN-γ and promotes the development of Th1 responses (42). In contrast, activation of the IL-6/STAT3 pathway switches the differentiation of monocytes from DCs to macrophage (47), represses tolerogenic functions of CD8α+ DCs (48, 49), and inhibits Th1 responses in lymphocytes by SOCS1- and/or SOCS3-dependent mechanisms (50, 51). In this study we have shown that the lowest level of SOCS gene expression occurs in pDC, whereas the highest level is observed in mDC. This implies that mature and immature DCs may differ in their sensitivity to cytokines regulated by SOCS. High constitutive expression of SOCS in mDCs further suggests that responsiveness to stimulation by certain cytokines may be lost after maturation. Understanding how activation and cross-talk between the many STAT pathways that converge on differentiating DCs are regulated and how feedback regulation of these pathways by SOCS family proteins contributes to DC flexibility of presentation programs (Th1 vs Th2; immunity vs tolerance) will be an area of great interest in the near future.
In summary, we have shown that the pattern of STAT use changes during the course of DC maturation. The STAT6 pathway is constitutively activated in iDCs and progressively declines as the cells differentiate into functionally mature DCs. However, a decline in STAT6 activation is not observed in mDCs from STAT1 knockout mice. In contrast, STAT1 signaling is detected at all stages of DC differentiation, but is most robust in mDCs, and whereas it inhibits CD86 in pDCs, it is required for up-regulation of CD40, CD11c, and SOCS expression in mDCs. Differential activation or repression of STAT1 and STAT6 signaling pathways, respectively, correlates with the tremendous increase in SOCS expression, suggesting that these two critical and antagonistic signaling pathways differ in their sensitivity to SOCS-mediated negative feedback regulation during DC maturation. Finally, signaling pathways and proteins that are differentially activated, repressed, or expressed in DCs at different stages of maturation constitute potential therapeutic targets for modulating DC activities.
Abbreviations used in this paper: DC, dendritic cell; BM, bone marrow; CIS, cytokine-induced Src homology 2-containing protein; GAS, IFN-γ activation site; HPC, hemopoietic progenitor cell; iDC, immature DC; JAK, Janus kinase; mDC, mature DC; pDC, precursor DC; SOCS, suppressor of cytokine signaling; WT, wild type; FLT3, FMS-related tyrosine kinase-3.