Promoter Analysis Reveals Critical Roles for SMAD-3 and ATF-2 in Expression of IL-23 p19 in Macrophages¹ Free

Cytokines of the IL-12 family produced by macrophage and dendritic cell lineages are heterodimers. Therefore, each of these cytokines is encoded by different genes, which must be expressed simultaneously, in order for an active form of the cytokine to be secreted (1). IL-23, a member of the IL-12 cytokine family, consists of a p40 subunit coupled to a p19 subunit (2). In contrast, IL-12 consists of the same p40 subunit coupled to a p35 subunit (3). Despite the facts that p19 is a homologue of p35 and also dimerizes with p40, IL-23 and IL-12 have distinct functions in development of effector and memory CD4⁺ T cell subsets. IL-12 stimulates the development of the Th1 IFN-γ-secreting CD4⁺ T cell subset, whereas IL-23 stimulates development of the CD4⁺ Th17 subset (4), which secretes IL-17, IL-17F, TNF-α, IL-6, and IL-22 during the adaptive immune response (2, 5, 6, 7).

IL-12 has a role in the immune response to intracellular infection (8), antiviral immunity (9), and anticancer immunity (10), whereas IL-17 has a role in antifungal immunity (11). Due to the presence of p40, IL-12 was also suspected to play a role in the development of several T cell-mediated autoimmune diseases, such as diabetes, inflammatory bowel disease, multiple sclerosis (MS)³/experimental autoimmune encephalomyelitis, and collagen-induced arthritis (12, 13). However, because of the discovery that p40 is also part of IL-23, IL-12 has been exonerated, and it is now accepted that IL-23 has the essential role in the development of T cell-mediated autoimmune diseases (2, 14). IL-23 appears to play a role in these autoimmune diseases by inducing development of Th17 (5, 15, 16). Therefore, the mechanisms to control expression of the p19 subunit of IL-23 must be elucidated.

Very little is known about the trigger for expression of IL-23 p19 during autoimmune diseases such as MS. One hypothesis on the development of T cell-mediated autoimmune diseases asserts that certain viral epitopes that mimic self epitopes stimulate autoimmune CD4 T cell responses (17). Our corollary to that hypothesis is that the virus must also induce production of IL-23. We have shown that Theiler’s murine encephalomyelitis virus (TMEV), which infects macrophages and induces an MS-like disease in SJL/J mice (18), triggers expression of both IL-23 subunits from the mouse macrophage cell line, RAW264.7 (19, 20). However, our understanding of the mechanisms for expression of IL-23 is not complete.

Macrophage responses to viruses occur in part through TLR pathways (21). TLRs have extracellular domains consisting of multiple leucine-rich repeat elements and a cytoplasmic domain that belong to the IL-1/Toll receptor family (22). Although TLRs that recognize molecular structures unique to bacterial and fungal cells (TLR1, TLR2, TLR4, TLR5, and TLR6) are localized to the plasma membrane, TLRs that recognize viral and bacterial nucleic acids (TLR3, TLR7, and TLR9) are localized at endosomal membranes (23). We have shown that TLR3 and TLR7 contribute to IL-23 p19 expression during challenge of macrophages with TMEV (19, 20). More recent promoter analyses have revealed, in both macrophages and dendritic cells, that p19 expression is dependent on binding of c-Rel and RelA NF-κB to the proximal p19 promoter (24, 25). However, cytokine gene expression usually involves multiple transcription factors. We have shown that p35 gene expression is controlled by NF-κB and IFN response factor (IRF)-1 (26). Therefore, we hypothesize that additional transcription factors are required for IL-23 p19 expression. The present investigation shows that besides a site for NF-κB, regulatory elements for IRF-3, activating transcription factor (ATF)-2, and Sma- and Mad-related protein (SMAD)-3 at the p19 promoter are essential for promoter activity. However, activation of SMAD-3 and ATF-2, but not IRF-3, is essential to p19 expression during the response to TMEV or TLR3 pathway activation.

Female SJL/J mice were purchased from Harlan Sprague-Dawley. The mouse macrophage cell line, RAW264.7 (American Type Culture Collection), was grown in DMEM cell culture medium (Invitrogen) containing 10% FBS (Invitrogen) and 50 μg/ml gentamicin (Invitrogen). The DA strain of TMEV was obtained from K. Drescher (Department of Medical Microbiology and Immunology, Creighton University, Omaha, NE). TMEV was grown in BHK-21 cells to produce stocks with 1 × 10⁷ PFU/ml. Macrophages were challenged with TMEV using a multiplicity of infection of 1. Loxoribine, an agonist of TLR7, and polyinosine-polycytidylic acid (poly(I:C)), an agonist of TLR3, were obtained from InvivoGen. SP 600125 (10 μM final concentration), an inhibitor of JNK MAPK; SB 203580 (10 μM), an inhibitor of p38 MAPK; and U0126 (20 μM), an inhibitor of ERK MAPK, were obtained from Promega.

Spleens were extracted from female SJL/J mice and placed into cold RPMI 1640 culture medium supplemented with glutamine, sodium pyruvate, 0.05 mM 2-ME, and 10% FBS. Cells were dispersed using 70-μm mesh screens; washed in Dulbecco’s PBS; treated with erythrocyte-lysing reagent containing 0.15 M NH₄Cl, 1.0 M KHCO₃, and 0.1 mM Na₂EDTA; washed; and resuspended in cell culture medium. Cells were counted with a hemacytometer using trypan blue.

Femurs were extracted from female SJL/J mice, severed at both ends, and using a 25 G needle, 5 ml of cold RPMI 1640 culture medium supplemented with glutamine, sodium pyruvate, 0.05 mM 2-ME, and 10% FBS was flushed through the shaft. Cells were washed in Dulbecco’s PBS, treated with erythrocyte-lysing reagent, washed, and resuspended in RPMI 1640 cell culture medium. Cells were counted with a hemacytometer using trypan blue.

Splenic mononuclear cell population has been reported to contain ∼10% macrophages (27), located primarily in the red pulp, and ∼1% dendritic cells, located primarily in the white pulp (28), both of which are adherent. Therefore, enriched splenic macrophages (SPM) were derived by adding 5 × 10⁵ splenic mononuclear cells to individual wells of a 96-well tissue culture plate, incubating at 37°C for 48 h, and then washing away nonadherent cells. Bone marrow-derived macrophages (BMM) were obtained by incubating 5 × 10⁴ bone marrow cells/well of 96-well plates with 20 ng/ml mouse rGM-CSF (Invitrogen) added on days 1 and 4. On day 7, nonadherent cells were removed and fresh RPMI 1640 culture medium was added. Adherent SPM and BMM were then challenged with TMEV in the presence or absence of neutralizing anti-TGF-β1 (clone 9016; R&D Systems). RAW264.7 cells were seeded into six-well plates at 1 × 10⁶/ml culture medium. After 24 h, nonadherent cells were removed, and 1 ml of culture medium was added. The adherent RAW264.7 cells were either untreated (control), stimulated with poly(I:C) (50 μg/ml), loxoribine (200 μM), LPS (500 ng/ml), poly(I:C) plus loxoribine, or challenged with TMEV. In one set of experiments, cells were untreated or pretreated for 30 min before infection with SP 600125 (10 μM), SB 203580 (10 μM), and U0126 (at 20 μM) or 1 μl of DMSO carrier.

The p19 promoter (based upon accession number NT 039502) was amplified from genomic murine DNA by PCR using an upstream primer containing a SstI restriction site, 5′-CGAGCTCGAGGTTCTTAGCCAGCATTC-3′, and a downstream primer containing an XhoI restriction site, 5′-CCGCTCGAGCTTGTTCCCTGCTTCTCAGA-3′, and cloned into the pGL3-basic vector (p19_prompGL3) (19). Sequential 5′ deletions to remove potential promotor transcriptional regulatory elements (Fig. 2A) were also generated by PCR using an upstream primer containing a SstI restriction site and a downstream primer containing an XhoI restriction site. Using a QuikChange II site-directed mutagenesis kit (Stratagene), nucleotides in the IRF-3 site at bp −734 to −731 were mutated from ATTT to CCAC; in the SMAD-3 site at bp −584 to −581 they were mutated from CAGAC to ACAG; in the ATF-2 site at bp −571 to −568 they were mutated from TGAG to GACT; in the IRF-7 site at bp −533 to −532 they were mutated from TT to AG, and at bp −526 to −525 they were mutated from TT to CG; and in the NF- κB site at bp −215 to −214 they were mutated from GG to TC, and at bp −210 to −209 they were mutated from CC to AA. The sequence of each insert was verified at the core facility of the Beadle Center for Biotechnology, University of Nebraska. All plasmids were isolated using a Qiagen endo-free plasmid kit.

Plasmids were transfected into the nucleus of RAW264.7 cells using the Cell Line Nucleofector Kit (Amaxa Biosystem), according to manufacturer’s instructions. Cells were transfected with 2 μg of the pGL3 reporter constructs along with 0.02 μg of a pRL-SV40 reference construct that constitutively expresses Renilla luciferase. We routinely obtained 50–60% transfection efficiency, with little impact on cell viability. Following transfection, cells were seeded at 8.3 × 10⁴ onto 96-well plates. After overnight culture, transfected cells were challenged with TMEV or stimulated with 50 μg/ml poly(I:C) or 200 μΜ loxoribine. After 24 h, cells were lysed and luciferase activity was measured with Dual-Luciferase Reporter Assay System (Promega) using a Polarstar Optima luminometer (BMG Labtech). Luciferase activity from pGL3 was normalized to luciferase activity from pRL-SV40.

After challenge with TMEV or stimulation with poly(I:C) and loxoribine, RNA was extracted using the RNAeasy kit of Qiagen, according to the manufacturer’s instructions, as we have done previously (20). cDNAs were prepared from 1 μg of RNA, as previously described (20). The sense/antisense primers used for quantitative PCR analyses were as follows: 5′-TACTGCCGCTTCTGCTCCCACT-3′/5′-GATGGCTTCGATGCGCTTC CGT-3′ for TGF-β1 to yield a 124-bp product, and 5′-TTGTCAGCA ATGCATCCTGCAC-3′/5′-ACAGCTTTCCAGAGGGGCCATC-3′ for GAPDH to yield a 149-bp product. Quantitative real-time PCR was performed with the Platinum-SYBR Green I-UDG-quantitative PCR SuperMix (Invitrogen) using an ABI Prism 7000 thermal cycler in which 1 μl of cDNA was incubated at 50°C for 2 min and 95°C for 2 min, followed by 40 cycles of 95°C for 15 s and 60°C for 30 s. Cycle thresholds (Ct) were normalized to Ct of GAPDH for each cDNA, and was expressed by fold increase using 2^−ΔΔCt (29). Representative samples of quantitative real-time PCR products were run electrophoretically on an agarose gel to confirm that a single PCR product was obtained.

IL-23 and TGF-β1 ELISA kits were obtained from eBiosciences. Ninety-six-well plates were coated with anti-mouse IL-23 p19 (clone G23-8) or 1.0 μg/ml anti-mouse/human TGF-β1 (clone eBioTB2F) in coating buffer at 4°C overnight. After five washes with PBS/0.05% Tween 20, each plate was blocked before addition of various concentrations of rIL-23, TGF-β1 standards, samples, or acid-activated samples (for TGF-β1 assay). The plates were incubated at room temperature for 3 h. After five washes with PBS/Tween 20, each plate was incubated with biotinylated anti-mouse IL-12/23 p40 (clone C17.8) or biotinylated anti-mouse/human TGF-β1 (clone eBio16TFB) at room temperature for 1 h. After five washes, the plates were incubated with avidin-HRP at room temperature for 30 min. The plates were washed seven times and incubated with 3,3′,5,5′ tetramethylbenzidine substrate/hydrogen peroxide solution. After adding 2 N H₂SO₄ stop solution, IL-23 and TGF-β1 were measured by determining ODs at 450 nm, with reference at 570 nm using an ELISA spectrophotometric plate reader (PolarStar Optima).

Expression vectors producing short-hairpin (sh) RNA against murine TLR3 (shTLR3) and TLR7 (shTLR7) were purchased from InvivoGen. Expression vectors producing shRNA against murine IRF-3 (shIRF-3), SMAD-3 (shSMAD-3), ATF-2 (shATF-2), or irrelevant scrambled sequence (shSCR) were constructed by inserting shDNA sequences (predicted to produce functional small interfering RNA from the SiRNAWizard program of InvivoGen) into the psiRNA-h7SK G₁ expression vector obtained from InvivoGen (Table I). Plasmids containing inserts were isolated using a Qiagen endo-free plasmid kit. RAW264.7 cells were transfected with 2 μg of shRNA expression vectors using the Cell Line Nucleofector Kit (Amaxa). Transfected cells were seeded at 1 × 10⁶ or 8.3 × 10⁴ onto 6- or 96-well plates, respectively, and then stimulated with poly(I:C) or loxoribine or challenged with TMEV before Western blot analysis.

To evaluate activation of IRF-3, SMAD-3, and ATF-2, RAW264.7 cells after challenge with TMEV or stimulation with poly(I:C) or loxoribine cells were lysed with protein-lysing reagent (Cell Signaling Technology), as we have done previously. To measure intracellular p19 protein, RAW264.7 cells were treated with brefeldin (10 μg/ml) to block protein secretion after 8 h of TMEV challenge or poly(I:C) and loxoribine stimulation, and then lysed at 24 h. The cell extracts were added to a sample buffer, and protein concentrations were determined using the Bio-Rad protein assay kit. Each sample containing equal amounts of protein was run through a 10% SDS, Tris-glycine-polyacrylamide gel, and transferred to a nitrocellulose membrane, as previously described (20). The primary Abs used for immunoblotting include anti-IL-23 p19 Ab (R&D Systems), anti-IRF-3 (Zymed Laboratories/Invitrogen), anti-phospho-IRF-3 (Ser³⁹⁶; Upstate Biotechnology), anti-ATF-2, anti-phospho-ATF-2 (Thr^69/71), anti-SMAD-3, anti-phospho-SMAD-3 (Ser^423/425; Cell Signaling Technology), and anti-β-tubulin E7 (Developmental Studies Hybridoma Bank, University of Iowa, Department of Biological Sciences, Iowa City, IA). The membranes were then incubated with IRDye800-labeled (Invitrogen) or Alexa Fluor680-labeled (Rockland Immunochemicals) secondary Abs in blocking buffer for 1 h. The membrane was washed three times, and infra-red light emissions were detected with a Li-Cor Odyssey system.

RAW264.7 cells were seeded in triplicate at 8.3 × 10³ cells/well in a 96-well plate. SPM and BMM were seeded at 5.0 × 10⁴ cells/well in a 96-well plate. Twenty-four hours after plating, cells were challenged with TMEV or stimulated, as described above, for the indicated time points. Cells were fixed in 3.7% formaldehyde for 20 min at room temperature and permeabilized by four 5-min washes with 0.1% Triton X-100, and nonspecific reactivity was blocked with Li-Cor Odyssey blocking buffer for 1 h. Anti-phospho-IRF-3 (Ser³⁹⁶; Upstate Biotechnology) or anti-phospho-SMAD-3 (Ser^423/425; Cell Signaling Technology) diluted in blocking buffer (1/100) was incubated with the cells overnight. Plates were then rinsed three times with PBS-0.1% Tween 20 and incubated for 1 h with anti-rabbit IRDye800 (Rockland; 1:100) and Sapphire700 (1:1000), a nonspecific cell stain that accumulates in both the nucleus and cytoplasm of fixed cells. Plates were washed three times with PBS-0.1% Tween 20, and then scanned with the Li-Cor Odyssey system. Fluorescence intensity at the 800-nm channel was normalized to the fluorescence intensity at the 700 channel.

ChIP assays were performed using the ChIP-It enzymatic digestion kit of Active Motif, according to the manufacturer’s recommendation. Briefly, RAW264.7 cells and splenic cells (2 × 10⁷), unstimulated or challenged for 6 h with TMEV, were cross-linked with 1% formaldehyde. Nuclei were isolated into buffer containing protease inhibitors and PMSF using a Dounce homogenizer and subjected to enzymatic digestion to yield to 300- to 1000-bp DNA fragments. Nuclei were immunoprecipitated with 3 μg of specific rabbit anti-IRF-3, anti-SMAD-3, or anti-ATF-2 (Cell Signaling Technology) overnight at 4°C. Protein-DNA cross-links of both input and precipitated DNA were reversed at 94°C for 15 min. Samples were treated with proteinase K before PCR analysis. Input and precipitated DNA were amplified by primers encompassing the IRF-3, SMAD-3, and ATF-2 sites in the promoter of p19: sense, 5′-ACCCGGGGAATGCCCTTACTTACTATTTCT-3′, and antisense, 5′-TCAAGGTTTATTCTTACCCAACCCCAGTC-3′. They were also amplified by primers designed away from the p19 promoter within the p19 open reading frame (ORF): sense, 5′-GCTGGATTGCAGAGCAGTAATA-3′, and antisense, 5′-GCATGCAGAGATTCCGAGAGAG-3′, as a negative control in a 32-cycle PCR. Input DNA was used as a positive control. The PCR products were analyzed by 1.8% agarose electrophoresis.

Previously, we reported that p19 expression in RAW264.7 cells responding to TMEV challenge is dependent on TLR3 and TLR7 pathways (20). To confirm this requirement, RAW264.7 cells were transfected with the p19 promoter reporter vector (p19_prompGL3) that we previously described (19) and then challenged with TMEV, stimulated with the TLR3 agonist, poly(I:C) (50 μg/ml), or stimulated with the TLR7 agonist, loxoribine (200 μM). TMEV challenge or poly(I:C), but not loxoribine stimulation, induced significant p19 promoter activity by 24 h (Fig. 1, A and B). To evaluate further the requirement for TLR3 and TLR7, shTLR3 and shTLR7 expression vectors were transfected into RAW264.7 cells along with p19_prompGL3. With a transfection efficiency of ∼50%, shTLR3 and shTLR7 significantly decreased expression of TLR3 and TLR7 in RAW264.7 cells during TMEV challenge (20). Likewise, transfection of either shTLR3 or shTLR7 significantly reduced murine p19 promoter activity following TMEV challenge of RAW264.7 cells (Fig. 1 C). These data confirm that challenge of RAW264.7 cells with TMEV induces p19 expression through TLR3 and TLR7, but stimulation through TLR7 alone will not induce p19.

Mise-Omata et al. (25) have shown that a putative site for NF-κB located at nt −215 in the murine p19 promoter is functional. Because there is significant homology between the murine and human p19 promoters (Fig. 2,A), we used the MatInspector transcription factor search program (30, 31) to identify additional transcription factor binding sites in the murine p19 promoter. MatInspector analysis found potential sites for IRF-3 (nt −734 to −731), IRF-7 (nt −533 to −525), ATF-2 (nt −571 to −568), SMAD-3 (nt −584 to −581), and NF-κB (nt −215 to −209) in the murine p19 promoter (Fig. 2,A). There was 100% homology between the murine and human NF-κB sites (Fig. 2,A), 41% homology in the region of the IRF-3 site, 78% homology at the SMAD-3 site, 57% homology at the ATF-2 site, and 90% homology at the IRF-7 site. Therefore, 5′ deletion mutations and site mutations at each of these sites were created in p19_prompGL3. The 5′ deletions were made upstream of and including the IRF-3, SMAD-3, ATF-2, IRF-7, and NF-κB sites using PCR, and these PCR products were reinserted into pGL3. RAW264.7 cells transiently transfected with each of these promoter reporter constructs were challenged with TMEV or stimulated with poly(I:C) or loxoribine. The responsiveness of the IL-23 p19 promotor to TMEV or poly(I:C) declined significantly upon elimination of the region upstream from nt −729, a region that contains the putative IRF-3 binding site (Fig. 2, B and C). Elimination of regions upstream from and including the SMAD-3, ATF-2, IRF-7, and NF-κB binding sites did not restore responsiveness. These results suggest that the IRF-3 binding site in the p19 promoter is a positive regulatory element for p19 expression. However, the roles of the SMAD-3, ATF-2, and IRF-7 binding sites were not clear. RAW264.7 cells were then transfected with the wild-type p19_prompGL3 or p19_prompGL3 constructs in which the IRF-3, SMAD-3, ATF-2, or IRF-7 sites were mutated. These cells were then challenged with TMEV or stimulated with poly(I:C) or loxoribine for 24 h. Mutations at the IRF-3, SMAD-3, ATF-2, or NF-κB binding sites significantly decreased the p19 promoter response to TMEV or poly(I:C) (Fig. 2, D and E). In contrast, mutation at the IRF-7 binding site did not affect the responsiveness of the p19 promoter. These results suggest that IRF-3, SMAD-3, ATF-2, and NF-κB are essential transcription factors for expression of p19.

The results to date suggest that TMEV stimulation of TLR3 and TLR7 pathways leads to activation of IRF-3, SMAD-3, and ATF-2 for p19 expression. IRF-3, SMAD-3, and ATF-2 are transcription factors that are each activated by hyperphosphorylation (9, 32, 33, 34). We have shown previously that TMEV challenge of RAW264.7 cells results in activation of ATF-2 (19). To determine whether IRF-3 and SMAD-3 are also activated by TMEV challenge or stimulation with poly(I:C), the phosphorylation status of each transcription factor was determined by immunoblot analysis. Challenge of RAW264.7 cells with TMEV or stimulation with poly(I:C) plus loxoribine triggered phosphorylation of IRF-3 within 90 min and phosphorylation of SMAD-3 and ATF-2 within 30 min (Fig. 3). To confirm our results, we quantified phosphorylation of IRF-3 and SMAD-3 by measuring the intensity of intracellular immunofluorescence and phosphorylation of ATF-2 by measuring intensity of Western blot phospho-ATF-2 images. The results confirm that TMEV challenge and poly(I:C)/loxoribine stimulation triggered significant intracellular phosphorylation of IRF-3 and SMAD-3 and stimulated ATF-2 phosphorylation, as measured by digital intensity of immunoblots compared with unstimulated controls (Fig. 3).

To examine further the role of IRF-3, SMAD-3, and ATF-2 in IL-23 p19 promotor activity, we constructed plasmid vectors that express shRNA, which are predicted to significantly reduce expression of each of these transcription factors. Transfection of shIRF-3, shSMAD-3, or shATF-2 vectors into RAW264.7 cells reduced total and activated IRF-3, SMAD-3, and ATF-2 (Fig. 4) during challenge with TMEV compared with transfection of a SCR shRNA vector. RAW264.7 cells transfected with shSMAD-3 or shATF-2 exhibited decreased p19 promoter activity 24 h after challenge with TMEV (Fig. 5,A) or stimulation with poly(I:C) (Fig. 5,B) compared with cells transfected with shSCR. Despite a reduction in IRF-3 activation in shIRF-3-transfected RAW264.7 cells challenged with TMEV, p19 promoter activity was not significantly different from in TMEV-challenged cells transfected with control plasmid. In contrast, transfection of shIRF-3 reduced IL-23 p19 promoter activity after stimulation with poly(I:C)/loxoribine, but not enough to achieve significance (Fig. 5,B). To confirm these data, intracellular p19 expression was measured by Western immunoblot in RAW264.7 cells that were transfected with shIRF-3, shSMAD-3, shATF-2, or shSCR, and then challenged with TMEV. TMEV-challenged RAW264.7 cells transfected with shSMAD-3 or shATF-2, but not shIRF-3, exhibited a decrease in p19 production compared with cells transfected with shSCR (Fig. 5 C). Because TMEV stimulation of RAW264.7 cells requires both TLR3 and TLR7 (20), we decided to stimulate RAW264.7 cells with poly(I:C) and loxoribine together to more closely mimic the TMEV challenge. Much like the TMEV challenge, poly(I:C)/loxoribine stimulated IL-23 p19 expression, whereas shSMAD-3 and shATF-2 significantly reduced the poly(I:C)/loxoribine-induced IL-23 p19 expression. Altogether, these data show that activation of SMAD-3 and ATF-2 is critical for expression of p19.

The data to date indicate that TMEV challenge of RAW264.7 cells results in activation of ATF-2, IRF-3, and SMAD-3, of which ATF-2 and SMAD-3 are required for p19 expression. To determine whether these transcription factors are present at the endogenous p19 promoter following challenge of RAW264.7 cells with TMEV, we used ChIP assays. As shown in Fig. 6 A, using primers (Primer A) specific to the p19 promoter region flanking the IRF-3, SMAD-3, and ATF-2 sites, challenge of RAW264.7 cells with TMEV induced association of SMAD-3 and ATF-2 with the p19 promoter. Interestingly, IRF-3 was constitutively present at the p19 promoter in RAW264.7 cells, whereas TMEV challenge increased this association somewhat. Using primers specific to the p19 ORF as a negative control (Primer B) did not yield any detectable PCR products.

SJL/J mice are susceptible to TMEV-induced demyelinating autoimmune disease in part because their macrophages can become persistently infected with TEMV (35). We hypothesize that these TMEV-infected macrophages from SJL/J mice produce IL-23, thus contributing to the demyelinating autoimmune disease. To determine whether these transcription factors are present at the endogenous p19 promoter of primary SJL/J macrophages, we conducted ChIP assays on SPM from SJL/J mice that were challenged with TMEV. Similar to RAW264.7 cells, TMEV challenge of SPM induced association of SMAD-3 and ATF-2 with the endogenous p19 promoter in SJL/J macrophages (Fig. 6,B). Previously, we showed that IRF-3 is constitutively active and localized in the nucleus in SJL/J macrophages (36). Interestingly, in agreement with that finding, we show in this study that IRF-3 is strongly associated with the p19 promoter in unchallenged SJL/J macrophage (Fig. 6 B). However, by 6 h after TMEV challenge, IRF-3 could no longer be found associated with the p19 promoter. Altogether, these data confirm that activation of SMAD-3 and ATF-2 induces their association with the IL-23p19 promoter, whereas IRF-3 is constitutively present at that promoter.

Although it is clear that ATF-2 and NF-κB are activated following viral infection of cells through TLR pathways, SMAD-3 is activated only through the TGF-β family of cytokines (37). Therefore, TGF-β1 expression by RAW264.7 cells and primary macrophages was evaluated by RT-PCR and ELISA following TMEV challenge, poly(I:C), or loxoribine stimulation. Within 6 h, significant TGF-β1 mRNA expression was induced in RAW264.7 cells challenged with TMEV (Fig. 7,A) or stimulated with poly(I:C) (Fig. 7,B). Unchallenged RAW264.7 cells and especially BMM produced measurable TGF-β1 (Fig. 7 C). In contrast, unchallenged SPM did not produce measurable TGF-β1. Nevertheless, challenge of BMM, SPM, and RAW264.7 cells with TMEV induced significant TGF-β1 protein secretion compared with unchallenged cells.

Because SMAD-3 is phosphorylated within 30 min after TMEV challenge of RAW264.7 cells, but detectable TGF-β1 mRNA is induced between 3 and 6 h, we challenged SPM, RAW264.7 cells, and BMM with TMEV in the presence or absence of neutralizing Ab to TGF-β1 and then measured secreted IL-23 protein by ELISA, activation of SMAD-3 by intracellular immunofluorescence, and intracellular p19 protein by Western immunoblot. All three cell types produced low, but detectable levels of IL-23 constitutively. However, IL-23 production from SPM and BMM of SJL/J mice, as well as RAW264.7 cells, was induced in response to TMEV challenge, whereas neutralizing anti-TGF-β1 caused a significant reduction in TMEV-induced IL-23 in SPM and BMM, and nearly a significant reduction in RAW264.7 cells (p = 0.06) (Fig. 8,A). TMEV challenge of BMM and SPM increased SMAD-3 phosphorylation within 30 min (Fig. 8,B), whereas neutralizing anti-TGF-β1 prevented TMEV-induced SMAD-3 activation. Likewise, TMEV challenge of RAW264.7 cells induced p19 protein expression, whereas neutralizing anti-TGF-β1 prevented this expression of p19 (Fig. 8 C). Therefore, TGF-β1 is most likely responsible for SMAD-3 activation and IL-23 expression in TMEV-challenged macrophages.

Previously, we showed that ATF-2 is activated following TMEV challenge of RAW264.7 cells. ATF-2 is activated by phosphorylation through either the JNK, ERK, or p38 MAPK pathways (38), all of which are activated following TMEV infection of RAW264.7 cells or SJL/J macrophages (36, 39, 40). To determine which of the MAPKs are responsible for the ATF-2 activation that is required for p19 expression, RAW264.7 cells were pretreated with SB203580 (10 μM), which inhibits activation of components downstream of the p38 pathway; U0126 (20 μM), which inhibits activation of ERK; SP 600125 (10 μM), which inhibits JNK MAPKs; or 1 μl of DMSO, which was used to dissolve the inhibitors. Inhibition of either the ERK or p38 MAPK had no effect on ATF-2 activation in response to TMEV challenge of RAW264.7 cells (Fig. 9,A). In contrast, inhibition of the JNK MAPK pathway reduced activation of ATF-2 in response to TMEV. Previously, we have seen that ERK is important in p19 promoter activity (19). We next determined the effect of the MAPK inhibitors on p19 expression using Western immunoblot. TMEV induced expression of p19, and 1 μl of DMSO increased that induction (Fig. 9,B). Inhibition of either the ERK or JNK, but not the p38, MAPK pathways reduced expression of p19 in response to TMEV to below that found with TMEV alone or TMEV with DMSO (Fig. 9 B). Therefore, ATF-2 activation for p19 expression in response to TMEV challenge of RAW264.7 cells occurs through the JNK MAPK pathway, but both the ERK and JNK inhibitors decreased TMEV-induced and DMSO-induced p19 expression.

Two recent reports have shown that following its activation, NF-κB binds to the proximal p19 promoter to participate in p19 expression in response to TLR activation (24, 25). The results of the present investigation clearly show that besides NF-κB, activation of SMAD-3 and ATF-2 transcription factors is also involved in p19 expression by macrophages challenged with TMEV. The interaction of multiple transcription factors has been documented for other genes. IFN-β expression requires coordinated binding of the transcription factors ATF-2/c-Jun, IRF-3, IRF-7 (41), and NF-κB to the well-studied IFN-β virus-inducible enhancer (34, 42). Likewise, we have shown that expression of the promoter activity for the p35 subunit of IL-12 requires both IRF-1 and NF-κB for optimal expression (26, 43). In the present study, we show that the transcription from the p19 promoter involves SMAD-3, ATF-2, and NF-κB.

However, activation of SMAD-3 in the macrophages used in this study is not directly related to TMEV challenge or poly(I:C) stimulation. SMAD transcription factors are activated by TGF-β and other members of the TGF-β family (44, 45). TGF-β binding to the TGF-β type II receptor induces formation of a complex with the type I receptor, which then leads to phosphorylation of SMAD-2 and -3 and their association with SMAD-4. The activated SMAD-2/3/4 complex translocates to the nucleus and directly participates in TGF-β-dependent transcriptional activation by binding to SBEs, such as the one located within the p19 promoter (46, 47). We have shown in this study that as early as 6 h following TMEV challenge or poly(I:C) stimulation, TGF-β1 expression is induced. Nevertheless, it appears that RAW264.7 cells and BMM produce some TGF-β1 constitutively. We show in this work that neutralization of TGF-β1 blocks SMAD-3 phosphorylation and prevents TMEV-induced IL-23 production by macrophages.

However, SMADs are also phosphorylated by MAPKs. ERK MAPKs, which are activated by macrophages challenged with TMEV, appear to have a negative impact upon nuclear localization of SMADs by additional phosphorylation of SMADs (48). In contrast, p38 and JNK MAPK phosphorylation of SMAD-3 enhances its nuclear localization (49). Therefore, we can conclude that TMEV-challenged macrophages express substantially more TGF-β1, exhibit more activated SMAD-3, and express activated MAPKs, which regulate SMAD-2/3/4 in positive and negative manners.

Like SMAD-3, ATF-2 is activated directly by TMEV challenge by way of TLR3 and TLR7 pathway stimulation of MAPKs (50). ATF-2 is a member of the ATF/CREB family transcription factor and binds to the cAMP response elements of promoters. MAPKs, p38, ERK, and JNK can phosphorylate and thus activate ATF-2 (51, 52). On ATF-2, activated JNK phosphorylates Thr⁶⁹/Thr⁷⁰/Ser⁹⁰, activated p38 phosphorylates Thr⁶⁹/Thr⁷⁰, whereas activated ERK phosphorylates Thr⁷⁰ (38). Our results show that the most important MAPK activation of ATF-2 leading to p19 expression in response to TMEV is JNK MAPK. However, p38 MAPKs are also activated by TGF-β through the TGF-β receptor and TGF-β-activated kinase-1 (53). In addition, ATF-2 expression and activation are induced by TGF-β (53). Therefore, TGF-β receptor signaling could have led to ATF-2 activation in response to TMEV challenge.

At the enhancesome of the IFN-β promoter, activated ATF-2 associates with c-Jun (54). However, ATF-2 has also been reported to bind directly to SMAD-3/4 complexes and is phosphorylated by TGF-β signaling through TGF-β-activated kinase-1 and p38 (53). This indicates that ATF-2 is a common nuclear target of the TGF-β/SMAD signaling and can form stable complexes with SMAD-3. It is interesting that the SMAD- and ATF-2 (cAMP response element)-binding elements are separated by 8 nt in the p19 promoter, suggesting that SMAD-3 and ATF-2 may be part of an enhancer complex for p19 expression. The present results show for the first time that IL-23 p19 gene expression requires such cooperation between ATF-2 and SMAD-3.

To our surprise, down-regulation of IRF-3 did not affect promoter activity or expression of IL-23 p19 protein, even though the putative IRF-3 site was important to p19 expression and IRF-3 is bound to this site. IRF-3 is activated by way of the TLR3 pathway by phosphorylation of Ser³⁹⁶ (55). However, additional phosphorylations through the PI3K/Akt pathway are required for complete activation, binding to the CBP protein, and association with promoter regions (55). We have shown previously that IRF-3 is constitutively active and localized to the nucleus in SJL/J macrophages, which are susceptible to TMEV-induced demyelinating disease, but not in macrophages from B10.S mice, which are resistant to TMEV-induced demyelinating disease (36). ChIP assays presented in this study agree with that observation, in that IRF-3 exhibits strong constitutive association with the endogenous p19 promoter in SJL/J macrophages and to some extent RAW264.7 cells. Interestingly, IRF-3 association with the endogenous p19 promoter was absent 6 h after TMEV challenge in SJL/J macrophages. We have seen unstimulated SJL/J macrophages express a high background of p19 (our unpublished data) and IFN-β (40) mRNA. These results suggest that because of constitutive activation of IRF-3, SJL/J macrophages may be primed for a quick burst of p19 mRNA expression compared with other strains of mice, therefore rendering SJL/J mice susceptible to demyelinating disease.

The constitutive association of IRF-3 with the p19 promoter may be one reason that knockdown of IRF-3 expression had little effect on p19 promoter activity. However, IRF-3 is only one member of an IRF family of transcription factors, IRF-1 to IRF-9 (reviewed in Ref. 56). IRFs (57) share a highly conserved N-terminal DNA binding domain, which binds directly to IFN-stimulated response elements or IFN response elements, such as that found in the p19 promoter (58) (reviewed in Ref. 57). Several IRF family proteins play essential roles in induction of type I IFNs (34) and proinflammatory cytokines (59). It is plausible that in the knockdown of IRF-3 due to shIRF-3, other IRFs compensated for IRF-3. Therefore, the IRF-3 binding site is required for p19 promoter activity, but other IRFs may possibly bind to this site in lieu of IRF-3.

In addition to ATF-2 and SMAD-3, NF-κB cRel and RelA subunits have been previously shown to be essential for p19 expression (24, 25). The NF-κB family members, RelA (p65), RelB, cRel, p50/p105, and p52/p100, form homo- and heterodimers (60, 61). Dimers of NF-κB p50 with p65 or cRel, or of p52 with RelB are located in the cytoplasm in an inactive complex with IκB. TLR stimulation leads to phosphorylation of IκBs, release of NF-κB, and IκB ubiquitination and proteasomal degradation (62, 63). Both reports have shown that a mutation of proximal κB site abolishes IL-23 p19 promotor activity in dendritic cells and macrophages. The present study confirms these findings in RAW264.7 cells and extends these findings by showing the involvement of other transcription factors, ATF-2 and SMAD-3, in IL-23 p19 expression.

TMEV infection of macrophages in certain strains of mice, such as SJL/J, leads to a demyelinating disease much like MS. Given the facts that many studies point to the involvement of macrophages and IL-23 in the pathology of MS, establishing that TMEV-induced cellular signaling leads to IL-23 expression by macrophages is crucial to understanding the mechanism by which virus infections could possibly lead to MS. The goal of the present study was to determine some of the mechanisms by which TMEV-infected macrophages induce expression of IL-23 at the transcriptional level. Because the RAW264.7 macrophage cell line expresses most TLRs and becomes chronically infected with TMEV (20, 64), these cells are ideal to investigate innate antiviral cellular signaling for expression of IL-23. Although TLRs play an important role in the innate immune responses against RNA viruses, it is possible that other innate antiviral response pathways are involved. Retinoic acid-inducible gene I and melanoma differentiation-associated gene 5 pathways, which are cytoplasmic, are also activated during RNA virus infection of macrophages. Activation of retinoic acid-inducible gene I (65) and melanoma differentiation-associated gene 5 (66, 67) pathways, which detect cytoplasmic viral RNA, triggers activation of NF-κB and IRF-3/7, which cooperate in induction of type I IFN. It is unknown whether these cytoplasmic pathways also lead to expression of IL-23.

The authors have no financial conflict of interest.