The antagonism between the cytokines IFN-γ and IL-4 is well documented, but the mechanism by which IL-4 inhibits IFN-γ-induced gene expression is not clearly understood. CD40 is a type I transmembrane protein that is critical for proper functioning of the immune system. We have previously shown that IFN-γ is the most potent inducer of CD40 expression by macrophages and microglia. In this report, we describe the molecular mechanisms by which IL-4 inhibits IFN-γ-induced CD40 expression. IL-4 suppresses IFN-γ-induced CD40 gene expression in both macrophages and microglia, and such inhibition is dependent on the activation of STAT-6. Nuclear run-on and transfection studies indicate that IL-4-mediated repression is at the transcriptional level. Furthermore, IL-4 inhibition of IFN-γ-induced CD40 expression is specific, since IL-4 does not inhibit IFN-γ-induced IFN-responsive factor-1 gene expression. Site-directed mutagenesis studies demonstrate that two STAT binding sites, named proximal and distal IFN-γ-activated sequences, in the human CD40 promoter are important for IL-4 inhibition of IFN-γ-induced CD40 promoter activity. Moreover, EMSAs indicate that IL-4-activated STAT-6 binds to these two STAT binding sites. These results suggest that IL-4 inhibition of IFN-γ-induced CD40 gene expression is mediated by direct STAT-6 binding to the CD40 promoter.
Regulation of immune responses is in part accomplished by a complex network of soluble molecules, including cytokines (for review, see Refs. 1 and 2), which can function in a synergistic and/or antagonistic manner. The two arms of the immune system are controlled by two compartments of Th cells that secrete distinct subsets of cytokines (3). IFN-γ is a cytokine that, together with IL-12, promotes Th1-mediated inflammatory reactions, while Th2-derived cytokines such as IL-4, IL-10, and IL-13 promote humoral immunity and oppose Th1-dependent activities (for review, see Ref. 4). An example of antagonism between Th1 and Th2 cytokines is the opposing effects of IFN-γ and IL-4 in their modulation of gene expression in the mononuclear phagocyte, a cell type that participates in both Th1- and Th2-mediated immunity (for review, see Ref. 5). IFN-γ negatively regulates IL-4-induced gene expression (e.g., IL-4Rα, FcεR, and the mannose receptor) (6, 7, 8), while IL-4 inhibits IFN-γ induction of cytokines (TNF-α, IL-1β, and IL-12) (9, 10, 11), chemokines (IFN-γ-induced protein-10 and monokine induced by IFN-γ (MIG)3) (9, 12), and cell surface molecules (ICAM-1, FcγR, and IL-12R) (13, 14, 15).
IL-4 activates at least four distinct signaling pathways, each of which has the potential to influence gene expression (for review, see Ref. 16). First, IL-4 activates phosphatidylinositol-3 kinase by tyrosine phosphorylation of c-Fes, a protein tyrosine kinase (17). Second, the IL-4R α-chain (IL-4Rα) activates a signaling pathway upon phosphorylation of insulin receptor substrate 1 and 2 (18, 19). The third signal transduction pathway involves activation of the Ras/mitogen-activated protein kinase cascade by the adaptor proteins Src homology 2 sequence containing protein and insulin/IL-4R-interacting protein (20, 21). The fourth signaling pathway used by IL-4 for transcriptional activation is the Janus kinase (JAK)/STAT pathway (22, 23).
STAT factors are latent cytoplasmic proteins that are activated upon tyrosine phosphorylation (for review, see Ref. 24). Both IFN-γ and IL-4 can initiate the JAK/STAT signaling pathway. IFN-γ binds to its heterodimeric receptor (α- and β-chains) and induces tyrosine phosphorylation of STAT-1α via JAK1 and JAK2 (for review, see Ref. 25). Similarly, the ligation of IL-4 to its receptor, composed of the IL-4Rα and a common γc-chain, activates STAT-6 through JAK1 and JAK3 (for review, see Ref. 16). Tyrosine-phosphorylated STATs dimerize, translocate to the nucleus, and bind Stat-binding elements (SBE) with the palindromic sequence of 5′-TCC-N2–4-GAA-3′ to activate gene transcription (for review, see Ref. 26). IL-4 inhibition of IFN-γ-induced gene expression can be accomplished by different mechanisms (for review, see Ref. 5). Posttranscriptional destabilization of IFN-γ-induced TNF-α mRNA by IL-4 has been described; however, the molecular mechanism was not elucidated (9). Several models have been proposed to explain the mechanisms by which IL-4 inhibits IFN-γ-induced gene transcription. First, IL-4-activated STAT-6 can compete with STAT-1 for a limited amount of the endogenous coactivator CREB-binding protein (CBP) (12). Second, STAT-6 competes with STAT-1 for binding to the same SBE (27). Third, it has been suggested that IL-4-activated STAT-6 induces an inhibitory factor(s) that mediates inhibition of IFN-γ-induced expression of the IFN-responsive factor (IRF) 1 gene (28). Last, a STAT-6-independent mechanism(s) has been proposed to be involved in IL-4 suppression of IFN-γ-induced IL-12 and TNF-α production (29).
CD40 is a member of the TNF receptor superfamily that is expressed by many different cells types (for review, see Ref. 30). The ligand for CD40 (CD154, glycoprotein 39) is expressed mainly and transiently on activated T cells. The interaction of CD40, which is expressed on APCs, and CD154 is critical for a productive immune response. Humans that fail to express CD154 present with a disease, X-linked hyper-IgM syndrome, with symptoms such as elevated levels of IgM and low or virtual absence of other Ab isotypes, susceptibility to recurrent bacterial and viral infections, and defects in T cell-mediated immunity (for review, see Ref. 31). Defects in CD40 signaling have also been described for X-linked hyper-IgM syndrome patients (32). CD40- or CD154-deficient mice display similar manifestations as the human counterpart (33, 34). Cytokine up-regulation of CD40 expression appears to be cell specific, and most CD40 inducers, such as IFN-γ, TNF-α, IL-1β, and GM-CSF, have proinflammatory properties (35, 36, 37, 38, 39, 40, 41). We have shown previously that IFN-γ is the most potent inducer of CD40 expression in macrophages and microglia, the resident macrophage of the brain (39, 41).
IL-4 has been shown to exert a beneficial effect in various inflammatory autoimmune diseases (42, 43, 44, 45). CD40 ligation on macrophages/microglia leads to the induction of numerous proinflammatory cytokines/chemokines, such as TNF-α, IL-12, and macrophage inflammatory protein-1α, as well as the expression of inflammation-associated molecules, such as ICAM-1, CD40, and class II MHC (for review, see Ref. 33). We hypothesize that one mechanism by which IL-4 mediates its anti-inflammatory effects is by suppression of IFN-γ-induced CD40 expression. In this study we demonstrate that IL-4 suppresses IFN-γ-induced CD40 gene expression at the transcriptional level in macrophages/microglia, and that this inhibition requires IL-4-activated STAT-6. Using transient transfection and EMSA, we demonstrate that IL-4-activated STAT-6 binds to two different SBEs in the human CD40 promoter to inhibit STAT-1α-mediated CD40 promoter activity. Collectively, our data indicate that IL-4 uses a novel STAT-6-dependent mechanism to inhibit IFN-γ-induced CD40 expression in macrophages/microglia.
Materials and Methods
Recombinant proteins and reagents
Recombinant murine IL-4 and IFN-γ were purchased from Genzyme (Boston, MA). Rat IgG2a-κ anti-mouse CD40 Ab (clone 3/23), biotinylated mouse anti-rat IgG2a, and PE-conjugated strepavidin were purchased from PharMingen (San Diego, CA). Polyclonal STAT-1α antisera and monoclonal anti-phosphotyrosine Ab (4G10) were obtained from Upstate Biotechnology (Lake Placid, NY), and mouse IRF-1, STAT-3, and STAT-6 antisera were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
The microglial cell line EOC13 was derived from C3H/HeJ CH-2k mice using a nonviral immortalization procedure as previously described (46). This CSF-1-dependent line is B7.1+, Mac-1+, CD45+, and class I MHC+ as well as phagocytic. The EOC13 cell line was maintained in DMEM complete medium as described previously (39). The murine macrophage cell line RAW264.7 (TIB-71) was purchased from American Type Culture Collection (Manassas, VA) and maintained as recommended. Primary microglia from wild-type BALB/c and STAT-6-deficient mice were prepared as described previously (39, 47, 48). STAT-6-deficient mice were purchased from The Jackson Laboratory (Bar Harbor, ME).
Cytokine treatment and quantitative analysis of CD40 protein expression by immunofluorescence flow cytometry
Cells were plated at 2 × 105 cells/well into 12-well plates (Costar, Cambridge, MA), and duplicate wells were treated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 40–48 h as previously described (39). The cells were scraped, incubated with 100 μl of 2.4G2 hybridoma supernatant (which contains rat anti-mouse FcγR Ab) supplemented with 10% normal mouse serum for 30 min at 4°C, washed, incubated with 10 μg/ml anti-CD40 Ab for 30 min at 4°C, incubated with 10 μg/ml biotinylated anti-rat IgG2a for 30 min at 4°C, then washed, and incubated with 10 μg/ml PE-conjugated strepavidin for 30 min at 4°C. The cells were then washed and fixed in a final volume of 200 μl of 1% paraformaldehyde and analyzed on a FACScan (Becton Dickinson, Mountain View, CA). Negative controls were incubated with isotype-matched Ab. Ten thousand cells were analyzed for each sample.
RNA isolation, riboprobes, and RNase protection assay (RPA)
Total cellular RNA was isolated from confluent monolayers of EOC13, RAW265.7 or murine primary microglia stimulated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 20 h (EOC13) or 8 h (RAW264.7 and primary microglia). The RNA isolation procedure and RPA were conducted as described previously (39).
Nuclear extracts and EMSAs
Cells were incubated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), and IFN-γ plus IL-4 for 30 min or were treated with IL-4 for 2 h before the addition of IFN-γ for 30 min, then nuclear extracts were prepared as described previously (39). The binding reaction was performed in a total volume of 15 μl (5 μg of nuclear extract, 20,000 cpm of probe, 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5% glycerol, 1 mM EDTA, 1 mM DTT, 1.5 μg of poly(dI · dC)) at 25°C for 15 min, then run on 6% polyacrylamide gel in 0.5× TBE buffer (50 mM Tris HCl (pH 8.0), 45 mM borate, 0.5 mM EDTA) for 1.5 h. For supershift analysis, 1 μg of the indicated Ab was added, or for competition analysis, a 50-fold molar excess of the indicated cold oligonucleotides was added to the nuclear extracts and incubated on ice for 30 min, followed by an additional incubation for 15 min with the labeled probe. The sequences of the probes used in this study are as follows: proximal IFN-γ-activated sequence (GAS) (pGAS), 5′-TTAGACTTGTGGGGAATGTTCTGGGGAAACTCCTGC-3′; medial GAS (mGAS), 5′-GGAAACTCTTCCTTGAAACGCCTCC-3′; and distal GAS (dGAS), 5′-GAGGGAATTTCCTTTGAAAGAGAGCG-3′. The underlining indicates the SBE core consensus sequence of the oligonucleotides.
Transient transfection and analysis
One microgram of the human CD40 promoter constructs (wild-type and mutant constructs) was cotransfected with 0.2 μg of the pCMV-β-galactosidase (β-gal) construct into 4 × 105 RAW264.7 cells in 6-well plates using the Lipofectamine Plus (Life Technologies, Rockville, MD) method as previously described (41). The promoterless pGL3-Basic was used as a negative (background) control in all experiments. Relative luciferase activity (RLA) was calculated as the ratio of luciferase activity to that of β-gal activity of the same sample. Fold induction was calculated as the ratio of RLA between cytokine-treated and medium-treated samples that were transfected with the same construct. Transient transfection experiments were not performed in either the EOC13 cells or primary microglia, as these cells are not amenable to various transfection protocols (data not shown).
Nuclear run-on assay
This assay was performed as described previously (41). Briefly, EOC13 cells were treated with medium, IFN-γ, IL-4, or IFN-γ plus IL-4 for 6 h. This period of cytokine exposure was previously determined to be maximal for IFN-γ-induced CD40 transcription (data not shown). Nuclei were harvested and resuspended in storage buffer, then frozen at −80°C until labeling. Transcripts that were initiated in the cells were allowed to continue in the presence of [α-32P]UTP at 30°C for 30 min. Labeled RNA was extracted and hybridized to membrane cross-linked plasmid DNA of pGEM-4Z as an empty vector control, murine CD40, and murine GAPDH for reference at 42°C for 48 h. The membranes were washed, then exposed to x-ray film. Quantification of bound labeled RNA was performed by scanning with the PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Values for CD40 expression were normalized to GAPDH levels for each experimental condition.
Immunoprecipitation and Western blot
Cells were treated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for the indicated times, and cell lysates were prepared as described previously (39). For immunoprecipitation of STAT-1α, 0.5 mg of total protein was precleared with normal rabbit serum before incubation with polyclonal antisera against STAT-1α (5 μl). Protein G-agarose (50 μl) was added for 2 h at 4°C, the immunoprecipitates washed three to five times with lysis buffer, eluted from the agarose beads by boiling in 2× SDS sample buffer, and subjected to 6% SDS-PAGE. Proteins were then transferred to nitrocellulose and probed with monoclonal anti-phosphotyrosine Ab 4G10 (1 μg/ml). Membranes were stripped at 50°C in buffer containing 100 mM 2-ME, 2% SDS, and 62.5 mM Tris-HCl (pH 6.7) with occasional shaking and reprobed for STAT-1α protein. For detection of IRF-1 proteins, 50 μg of total protein was boiled in sample buffer, separated on 10% SDS-PAGE, and then transferred to nitrocellulose membrane and probed with anti-IRF-1 Ab. Enhanced chemiluminescence was used for detection of bound Ab. Quantification was performed on the Bio-Rad Gel Doc 1000 using the Molecular Analyst Program (Bio-Rad, Hercules, CA).
Levels of significance for comparison between samples were determined by Student’s t test distribution.
IL-4 inhibits IFN-γ-induced CD40 expression
We have previously shown that IFN-γ is the most potent inducer of CD40 expression in macrophages/microglia (39, 41). Since IL-4 has been shown to antagonize IFN-γ-induced gene expression (9, 12, 27, 28), we studied the effect of IL-4 on IFN-γ-induced CD40 expression in macrophages/microglia. The murine microglial cell line EOC13, the macrophage cell line RAW264.7, and primary murine microglia were incubated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 48 h, then assessed for surface expression of CD40 (Fig. 1). Optimal expression of CD40 protein is detected after a 48-h incubation with IFN-γ (41). EOC13 and primary microglia expressed low constitutive levels of CD40, while the macrophage cell line RAW264.7 expressed moderate constitutive levels of CD40. In all three cell types, IL-4 alone marginally increased surface expression of CD40, while IFN-γ strongly enhanced the expression of CD40 (Fig. 1). A simultaneous treatment with IL-4 and IFN-γ led to suppression of IFN-γ-induced CD40 protein expression on all three cell types (Fig. 1). The decrease in the mean fluorescent intensity was ∼62% (EOC13), ∼52% (RAW264.7), and ∼48% (primary microglia). Dose-response experiments were conducted using IL-4 at concentrations of 0.01–20 ng/ml. Optimal inhibition was observed with 5–10 ng/ml of IL-4 (data not shown).
Next, we wished to determine whether IL-4 affected steady-state levels of IFN-γ-induced CD40 mRNA. EOC13 cells, RAW264.7 cells, and murine primary microglia were incubated with medium, IFN-γ, IL-4, or IFN-γ plus IL-4 for 20 h (EOC13) or 8 h (RAW264.7 and primary microglia), then total RNA was harvested and analyzed for CD40 and GAPDH mRNA expression using RPA (Fig. 2). Low levels of CD40 mRNA were expressed constitutively (lanes 1, 5, and 9), and IFN-γ enhanced the accumulation of CD40 mRNA by about 48-fold (EOC13), about 54-fold (RAW264.7), and about 37-fold (primary microglia; lanes 2, 6, and 10). Treatment with IL-4 alone had no effect on CD40 mRNA expression; however, IL-4 suppressed IFN-γ-induced CD40 mRNA expression in all three cell types by ∼77% (EOC13), ∼60% (RAW264.7), and ∼57% (primary microglia; lanes 4, 8, and 12). These data indicate that IL-4 inhibits IFN-γ-induced CD40 expression by reducing steady state levels of CD40 mRNA in macrophages/microglia. We performed experiments in the EOC13 cell line to determine the kinetics of IL-4 inhibition. Inhibition was optimal when IL-4 and IFN-γ were added simultaneously; however, a strong inhibitory effect was also observed when IL-4 was added 30 min pre- or post-IFN-γ treatment (data not shown).
IL-4 does not affect IFN-γ-induced CD40 mRNA stability
To ascertain the level of inhibition of IFN-γ-induced CD40 mRNA expression by IL-4, we examined the effect of IL-4 on the stability of IFN-γ-induced CD40 mRNA. EOC13 cells were treated with IFN-γ in the absence or the presence of IL-4 for 20 h, then the transcriptional inhibitor actinomycin D (5 μg/ml) was added for an additional 1–8 h. RNA was isolated at these different time points and analyzed for CD40 and GAPDH mRNA levels (Fig. 3,A). The inclusion of IL-4 inhibited IFN-γ-induced CD40 mRNA levels by ∼65% (compare lanes 2 and 7). However, IL-4 did not have any effect on the decay rate of IFN-γ-induced CD40 mRNA in EOC13 cells (Fig. 3,B). The half-life of IFN-γ-induced CD40 mRNA in the absence or the presence of IL-4 was >8 h. Similar results were obtained using RAW264.7 cells (Fig. 3 C). Collectively, these results suggest that IL-4 does not affect the stability of IFN-γ-induced CD40 mRNA.
IL-4 does not affect IFN-γ activation of the JAK/STAT signal transduction pathway
One possible mechanism by which IL-4 may inhibit IFN-γ-induced gene expression is to interfere with IFN-γ activation of the JAK/STAT signaling pathway (49, 50). To investigate this possibility, we first tested the effect of IL-4 on IFN-γ-induced tyrosine phosphorylation of STAT-1α. EOC13 cells were incubated with medium, IFN-γ, IL-4, or IFN-γ plus IL-4 for 30 min. Cell lysates were immunoprecipitated with polyclonal antisera against STAT-1α, then analyzed by Western blotting using anti-phosphotyrosine Ab (4G10). As shown in Fig. 4,A (top panel), IFN-γ treatment induced tyrosine phosphorylation of STAT-1α (lane 2), and IL-4 did not affect this response (lane 4). The blot was stripped and reprobed for STAT-1α to determine the amount of STAT-1α protein present in each lane (Fig. 4 A, bottom panel). Identical results were obtained when cells were pretreated with IL-4 for 2 h, then exposed to IFN-γ for 30 min (data not shown). Similar results were also obtained with RAW264.7 cells (data not shown). These results indicate that IL-4 does not interfere with IFN-γ induction of STAT-1α tyrosine phosphorylation.
Next, we determined whether IL-4 affects the DNA binding ability of tyrosine-phosphorylated STAT-1α. EOC13 or RAW264.7 cells were stimulated with medium, IFN-γ, IL-4, or IFN-γ plus IL-4 for 30 min or were treated with IL-4 for 2 h before the addition of IFN-γ for 30 min. Nuclear extracts were prepared and analyzed by EMSA using the oligonucleotide containing the mGAS sequence from the human CD40 promoter. We previously determined that the mGAS element is critical for IFN-γ-induced CD40 promoter activity, and that STAT-1α binds to this element (41). There was no complex formation with extracts from medium-treated cells (Fig. 4 B, lanes 1 and 10), while a complex was detected from IFN-γ-treated cells (lanes 2 and 11). The IFN-γ-induced complex was verified to contain STAT-1α by supershift analysis (data not shown), in agreement with our previous findings (41). Complex formation was not observed using nuclear extracts from IL-4-treated cells (lanes 3 and 12). Simultaneous treatment of IL-4 plus IFN-γ or a 2-h pretreatment with IL-4, then exposure to IFN-γ, did not alter the mobility or pattern of the IFN-γ-induced complex (lanes 4, 5, 13, and 14). The complex formed in the presence of IFN-γ plus IL-4 was supershifted with STAT-1α Ab, but not with normal rabbit serum, STAT-3, or STAT-6 Abs (lanes 6–9). These data suggest that IL-4 does not influence the binding of tyrosine-phosphorylated STAT-1α to the mGAS element of the CD40 promoter. Furthermore, these results indicate that STAT-6 does not bind to the human CD40 mGAS element.
To further investigate the influence of IL-4 on IFN-γ signaling, we examined whether IL-4 could modulate IFN-γ induction of the IRF-1 transcription factor in EOC13 and RAW264.7 cells. IRF-1 is a member of the IRF family of proteins that is inducible upon IFN-γ stimulation (51). Cells were incubated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 2 h, then cell lysates were analyzed by Western blotting for IRF-1 protein expression. As demonstrated in Fig. 4 C, IFN-γ-induced expression of IRF-1 in EOC13 and RAW264.7 cells (lanes 2 and 6), and the inclusion of IL-4 did not affect IFN-γ induction of IRF-1 (lanes 4 and 8). Similar results were obtained when the cells were pretreated with IL-4 for 2 h, then exposed to IFN-γ for an additional 2 h (data not shown). These results collectively demonstrate that IL-4 does not indiscriminately inhibit IFN-γ-induced gene expression in macrophages/microglia.
IL-4 inhibits IFN-γ-induced CD40 transcription
We have shown that IL-4 suppresses IFN-γ-induced CD40 gene expression by reducing the steady state levels of CD40 mRNA (Fig. 2). We further demonstrated that IL-4 does not affect the stability of IFN-γ-induced CD40 mRNA or interfere with IFN-γ-activated JAK/STAT signaling (Figs. 3 and 4). Thus, we next examined whether IL-4 inhibits IFN-γ-induced CD40 transcription. EOC13 cells were treated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 6 h, then nuclei were harvested and subjected to nuclear run-on analysis. This period of treatment was previously determined to be optimal for IFN-γ-induced CD40 transcription (data not shown). Labeled RNA was then hybridized to membrane-anchored plasmids containing CD40 cDNA, pGEM-4Z as empty vector control, and GAPDH cDNA as reference (Fig. 5,A). IFN-γ treatment resulted in an ∼18-fold enhancement of CD40 transcription compared with the medium-treated sample (panels 1 and 2). Treatment with IL-4 alone modestly increased CD40 transcription (∼2-fold; panel 3). However, IL-4 suppressed IFN-γ-induced CD40 transcription by ∼65% (panel 4), indicating that IL-4 inhibits IFN-γ-induced CD40 expression at the transcriptional level. To further confirm these results in another cell type, RAW264.7 cells were transiently transfected with the human CD40p0.7 construct, which contains 711 bp of the 5′-flanking sequence of the human CD40 gene, and a β-gal construct for normalization (41). The transfected cells were incubated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 12 h, then luciferase and β-gal activities were analyzed (Fig. 5 B). Treatment of the transfected RAW264.7 cells with IFN-γ resulted in an ∼16-fold enhancement of promoter activity compared with that in the medium-treated sample. IL-4 alone enhanced CD40 promoter activity by about 2-fold; however, IL-4 suppressed IFN-γ-induced CD40 by ∼50%. These results collectively indicate that IL-4 inhibits IFN-γ-induced CD40 transcription.
STAT-6 is required for IL-4 inhibition of IFN-γ-induced CD40 expression
IL-4 inhibition of IFN-γ-induced MIG and iNOS expression is dependent on the transcription factor STAT-6 (12, 52). We studied the role of STAT-6 in IL-4 inhibition of IFN-γ-induced CD40 expression. Microglia isolated from STAT-6-deficient mice (47, 48) were incubated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 48 h, then assessed for CD40 surface expression by flow cytometry (Fig. 6). IFN-γ-induced CD40 expression in STAT-6-deficient cells; however, IL-4 suppression of IFN-γ-induced CD40 expression was abrogated in STAT-6-deficient microglia. These findings indicate that IL-4 inhibition of IFN-γ-induced CD40 expression requires the STAT-6 transcription factor.
The pGAS and dGAS elements of the human CD40 promoter are important for IL-4 inhibition of IFN-γ-induced CD40 promoter activity
Within the human CD40 promoter, there are three SBEs that we have designated pGAS, mGAS, and dGAS. We previously identified the mGAS and dGAS elements as important for IFN-γ-induced CD40 promoter activity (41) and wanted to ascertain the potential role of each of the SBEs in IL-4-mediated inhibition of IFN-γ-induced CD40 promoter activity. The wild-type CD40 construct and constructs with mutations in the dGAS (mdGAS) and pGAS (mpGAS) elements were cotransfected with a β-gal construct into RAW264.7 cells, which were then treated with medium, IFN-γ (75 U/ml), IL-4 (10 ng/ml), or IFN-γ plus IL-4 for 12 h and analyzed for luciferase and β-gal activities (Fig. 7). The construct with a mutation in the mGAS element was not analyzed because this mutation abolishes IFN-γ induction of promoter activity (41). IL-4 inhibited IFN-γ induction of the wild-type construct by ∼50%. Using the construct with a mutation in the pGAS element (mpGAS), IFN-γ inducibility was comparable to that of the wild-type construct; however, the IL-4 inhibitory activity was abrogated (Fig. 7). Similar to our previous report (41), mutation of the dGAS element reduced IFN-γ-induced CD40 promoter activity by ∼50% (mdGAS construct), and IL-4 did not inhibit this response further. These results suggest that the two SBEs (the pGAS and dGAS elements) play important roles in IL-4 inhibition of IFN-γ-induced CD40 promoter activity.
STAT-6 binds to the pGAS and dGAS elements of the human CD40 promoter
IL-4 activates the JAK/STAT pathway, specifically, the activation of STAT-6 (22, 47). STAT-6 has been shown to antagonize STAT-1α-mediated IFN-γ induction of IRF-1, MIG, and NOS genes (12, 27, 28, 52). Since we have shown that the pGAS and dGAS elements are important for IL-4 inhibition of IFN-γ-induced CD40 promoter activity (Fig. 7), we next determined whether STAT-6 binds to either of these SBEs in the human CD40 promoter. Using an oligonucleotide containing the pGAS sequence as a probe and nuclear extracts from medium- or cytokine-treated EOC13 and RAW264.7 cells, EMSA was performed (Fig. 8). Nuclear extracts from untreated or IFN-γ-treated cells did not form complexes on the pGAS probe (lanes 1, 2, 12, and 13). However, in the presence of IL-4, a complex was observed (lanes 3 and 14), which was not affected by the inclusion of IFN-γ (lanes 5 and 16). This IL-4-induced complex was competed away with a 50-fold molar excess of the unlabeled pGAS and dGAS oligonucleotide, but not with mGAS-containing oligonucleotides (lanes 6–8), demonstrating that the binding of the complex is specific. The identity of the IL-4-induced complex was verified by supershifting with anti-STAT-6 Ab, but not anti-STAT-1 or anti-STAT-3 Abs (lanes 9–11). When nuclear extracts from cells pretreated with IL-4 for 2 h then exposed to IFN-γ for 30 min were tested, the intensity of the STAT-6 complex was diminished (lanes 4 and 15). In fact, the STAT-6 complex was not detectable after an 8-h treatment with IL-4 (data not shown). Similar results were obtained when the dGAS probe was used (Fig. 8 B). These data demonstrate that IL-4-activated STAT-6 binds to the pGAS and the dGAS sequences of the CD40 promoter, and treatment with IFN-γ does not modify the pattern of complex formation.
In this study, we report on the mechanism by which IL-4 inhibits IFN-γ-induced CD40 expression in macrophages and microglia. IL-4 suppresses IFN-γ-induced CD40 protein and mRNA expression in both cell types (Figs. 1 and 2) and does so through the activation of STAT-6, because IL-4 treatment of STAT-6-deficient microglia failed to attenuate IFN-γ-induced CD40 expression (Fig. 6). Although IL-4 posttranscriptionally destabilizes IFN-γ-induced TNF-α mRNA (9), our data support an inhibitory effect of IL-4 on IFN-γ-induced CD40 transcription (Fig. 5), while ruling out possible posttranscriptional mechanisms (Fig. 3). Similar to other studies (27), we found that IL-4 does not interfere with IFN-γ activation of the JAK/STAT pathway (Fig. 4). The IL-4 inhibition of STAT-1α-dependent gene transcription is not a global phenomenon, because IL-4 does not inhibit IFN-γ induced IRF-1 expression (Fig. 4 C), indicating a selective inhibitory effect of IL-4 on IFN-γ-induced CD40 gene expression.
Competitive binding of STAT-6 and STAT-1 to the same SBE has been suggested for the IRF-1 and MIG promoters (12, 27). Unlike the promoters of these genes, which contain one functional SBE, the CD40 promoter contains at least three SBEs (the dGAS, mGAS, and pGAS elements), two of which (the dGAS and mGAS elements) are required for maximal IFN-γ induction of CD40 expression (41). The mGAS element is critical for IFN-γ activation of the CD40 promoter, such that mutation of this site abrogates IFN-γ-induced promoter activity. Therefore, we could not study the functional effect of IL-4 on this mGAS element. However, it is unlikely that IL-4 makes use of this GAS element to suppress IFN-γ-induced CD40 promoter activity, because STAT-6 does not bind to it (Fig. 4,B). In contrast, IL-4-activated STAT-6 strongly binds to the dGAS (an N4 SBE) element (Fig. 8,B). The dGAS element is functionally important for IFN-γ induction of CD40 promoter activity because mutation of this element reduced IFN-γ-induced promoter activity by ∼50% (Fig. 7) (41); however, with the binding conditions used in this study, STAT-1α binding was not observed (Fig. 8 B). To demonstrate STAT-1α binding to the dGAS element, a binding buffer that has been optimized for STAT-1 binding must be used (39), and only upon overexposure do we observe a faint STAT-1α-containing complex (data not shown). One explanation for this disparity is that other factors may stabilize STAT-1α binding to the dGAS element in vivo. This differential affinity of the dGAS element for STAT-1α and STAT-6 may play a pivotal role in the regulation of CD40 gene expression by IFN-γ and IL-4. Binding of STAT-6 to the dGAS element may prevent STAT-1α from binding to this site, thus decreasing IFN-γ-induced CD40 promoter activity.
STAT-6 binding to the dGAS element is not sufficient for inhibition of STAT-1α-induced CD40 promoter activity. Preservation of another N4 SBE, the pGAS element located at −129 bp, is also necessary for IL-4 to suppress IFN-γ-induced CD40 promoter activity. We have shown previously and in this study that the pGAS element does not participate in IFN-γ-induced CD40 promoter activity, and STAT-1α does not bind to this element (Figs. 7 and 8,A) (41). However, IL-4-activated STAT-6 does bind to the pGAS element (Fig. 8 A). Together, these data indicate that STAT-6 binding to the pGAS element is also necessary for IL-4 repression of IFN-γ-induced CD40 promoter activity.
One of the interesting properties of STAT-6 is that its trans-activating potential varies depending on the SBE to which it binds (12, 27, 53, 54). In the case of CD40, STAT-6 has modest trans-activating potential because treatment of cells with IL-4 alone increased CD40 expression only ∼2-fold above that in medium-treated controls (Figs. 1 and 5). The modest induction of CD40 promoter activity by IL-4 alone required preservation of the pGAS element, but not the dGAS element (data not shown). However, intact pGAS and dGAS elements are essential for IL-4 inhibition of IFN-γ-induced CD40 expression. It is attractive to speculate that competitive binding of STAT-6 to the dGAS element prevents STAT-1α from binding the same GAS element. If STAT hindrance or competition at the dGAS element is the mechanism, then preservation of the pGAS element would not be necessary for IL-4 inhibition of IFN-γ-induced CD40 promoter activity. The observed requirement for both GAS elements suggests that cooperation between STAT-6 binding to two different SBE may be necessary for the inhibition of STAT-1α-dependent CD40 transcription. STAT-1α dimers have been shown to interact by their N-terminal domain when they bind to adjacent GAS elements (55, 56); however, similar interactions between STAT-6 dimers have not been demonstrated. Thus, STAT-6 cooperation may be accomplished by interacting with an as yet unidentified integrator or coactivator.
It has also been proposed that STAT-6 competes with STAT-1α for binding to the coactivator CBP because it was observed that both STAT-6 and STAT-1α coimmunoprecipitate with CBP (12). However, overexpression of CBP does not reverse the inhibitory effect of STAT-6 on STAT-1α-activated IRF-1 promoter activity (28). Furthermore, STAT-6 and STAT-1α interact with CBP at different domains (57, 58), and such interactions are independent of IL-4 or IFN-γ treatment (12), suggesting that sequestration of CBP by STAT-6 is an unlikely mechanism. Since CBP/p300 is required for STAT-1α-mediated gene transcription (57, 59), the interaction of STAT-6 may alter the conformation of the CBP/p300 and STAT-1α complex, thereby decreasing the trans-activating potential of CBP/p300. The involvement of CBP/p300 in IFN-γ-induced CD40 gene expression and the effect of STAT-6 on CBP/p300:STAT-1α are currently under investigation.
Suppression of IFN-γ-induced IRF-1 gene expression by IL-4 occurred only when low doses of IFN-γ (1–10 U/ml) were used; IL-4 had no inhibitory effect at higher concentrations of IFN-γ (27). Moreover, this inhibition required supraphysiologic levels of STAT-6 (28). In our system IL-4 is able to inhibit IFN-γ-induced CD40 expression regardless of the IFN-γ concentration used (1–1000 U/ml; data not shown). Furthermore, endogenous levels of STAT-6 in the RAW264.7, EOC13, and primary microglial cells are sufficient for IL-4 to inhibit STAT-1α-dependent CD40 transcription. IL-4 inhibition of IFN-γ-induced IRF-1 gene expression required high levels of STAT-6 and low doses of IFN-γ (27, 28), leading to the speculation that a critical threshold of activated STAT-6 relative to STAT-1α is required for STAT-6-mediated inhibition of STAT-1α-induced IRF-1 gene transcription (28). However, this is not the case for CD40, as IL-4 inhibits IFN-γ induced CD40 expression in a dose-dependent manner (data not shown).
Other possible mechanisms exist by which IL-4 may inhibit IFN-γ-induced CD40 expression. IL-4 may prevent serine phosphorylation of STAT-1α, which is necessary for maximal STAT-1α activity (60). Moreover, we have demonstrated that the Ets proteins PU.1 and/or Spi-B are required for IFN-γ-induced CD40 promoter activity (41), and the activity of PU.1 can be modulated by serine phosphorylation (61). Although treatment with IL-4 does not affect the binding of PU.1 and/or Spi-B to the Ets sites on the CD40 promoter (data not shown), IL-4 may regulate the activity of PU.1 and/or Spi-B by modulating the phosphorylation status of these Ets proteins.
The recently discovered family of the suppressors of cytokine signaling (SOCS) proteins, which can attenuate the JAK/STAT signaling pathway, have been proposed to negatively regulate cytokine signaling pathways (for review, see Ref. 62). Among the SOCS proteins, SOCS-1 has been shown to interact with activated JAK2, preventing downstream tyrosine phosphorylation of STAT-1α and subsequent STAT-1α activity (49, 63, 64, 65). Indeed, IL-4 is able to induce appreciable levels of SOCS-1 mRNA in EOC13, RAW264.7, and primary microglial cells (data not shown); however, treatment with IL-4 does not affect the ability of IFN-γ to activate STAT-1α or to induce IRF-1 expression in these cells (Fig. 4). As a result, we have excluded IL-4-induced SOCS-1 as a negative regulator of IFN-γ-induced CD40 expression in macrophages/microglia.
In conclusion, this study demonstrates a novel STAT-6-mediated mechanism that inhibits IFN-γ-induced CD40 expression. Besides binding to the same GAS element (dGAS) as STAT-1α does, IL-4-activated STAT-6 also binds to another distinct SBE (pGAS) to suppress STAT-α-dependent transcription of CD40. The binding of STAT-6 to both the dGAS and pGAS elements is essential for IL-4 to suppress STAT-1α-mediated CD40 gene expression. Negative regulation of IFN-γ-induced CD40 in macrophages/microglia by a Th2 cytokine may be one of the mechanisms by which IL-4 antagonizes the development of Th1 cells and dampens inflammatory responses. It will be of interest to determine whether this mechanism of negative regulation of CD40 expression occurs in other APCs, such as dendritic cells or B cells.
This work was supported by National Multiple Sclerosis Society Grant RG2205B9 (to E.N.B.) and National Institutes of Health Grant NS36765 (to E.N.B.). V.T.N. is supported by the National Institutes of Health Predoctoral Fellowship T32AI07493. We acknowledge the support of the University of Alabama at Birmingham Flow Cytometry Core Facility (AM20614).
Abbreviations used in this paper: MIG, monokine induced by IFN-γ; CBP, CREB-binding protein; GAS, IFN-γ-activated sequence; pGAS, proximal GAS; mGAS, medial GAS; dGAS, distal GAS; IRF, IFN-responsive factor; JAK, Janus kinase; RPA, RNase protection assay; RLA, relative luciferase activity; β-gal, β-galactosidase; SBE, STAT-binding element; SOCS, suppressors of cytokine signaling.