IL-27, a multifunctional cytokine produced by APCs, antagonizes inflammation by affecting conventional dendritic cells (cDC), inducing IL-10, and promoting development of regulatory Tr1 cells. Although the mechanisms involved in IL-27 induction are well studied, much less is known about the factors that negatively impact IL-27 expression. PGE2, a major immunomodulatory prostanoid, acts as a proinflammatory agent in several models of inflammatory/autoimmune disease, promoting primarily Th17 development and function. In this study, we report on a novel mechanism that promotes the proinflammatory function of PGE2. We showed previously that PGE2 inhibits IL-27 production in murine bone marrow–derived DCs. In this study, we show that, in addition to bone marrow–derived DCs, PGE2 inhibits IL-27 production in macrophages and in splenic cDC, and we identify a novel pathway consisting of signaling through EP2/EP4→induction of cAMP→downregulation of IFN regulatory factor 1 expression and binding to the p28 IFN-stimulated response element site. The inhibitory effect of PGE2 on p28 and irf1 expression does not involve endogenous IFN-β, STAT1, or STAT2, and inhibition of IL-27 does not appear to be mediated through PKA, exchange protein activated by cAMP, PI3K, or MAPKs. We observed similar inhibition of il27p28 expression in vivo in splenic DC following administration of dimethyl PGE2 in conjunction with LPS. Based on the anti-inflammatory role of IL-27 in cDC and through the generation of Tr1 cells, we propose that the PGE2-induced inhibition of IL-27 in activated cDC represents an important additional mechanism for its in vivo proinflammatory functions.
Interleukin-27, a cytokine that belongs to the IL-12 family, is a p28/EBI3 heterodimer that signals through the IL-27Rα(WSX-1)/gp130 receptor (1, 2). The inclusion of IL-27 within the IL-12 cytokine family is based on the IL-27 subunits EBI3 and p28 being related to IL-12p40 and IL-12p35, respectively. The homology extends to the IL-27/IL-12 receptors with IL-27Rα/WSX-1 being homologous to IL12Rβ2 (3). Myeloid cells including dendritic cells (DC) are major IL-27 producers following activation through TLRs, CD40, and IFN type I and II receptors (3, 4). In a model using conditional IL27p28-deficient mice, CD11c+ DC were identified as the critical source of IL-27 (5). Expression of IL-27Rα is limited to immune cells, especially T and NK cells (2), with IL-27Rα increasing following T cell activation (6). Recently, conventional DC (cDCs) were reported to express IL-27Rα at higher levels than plasmacytoid DC (7).
Although initial studies suggested a proinflammatory role for IL-27, models of infectious and autoimmune diseases in Il27ra–deficient mice identified IL-27 as a negative regulator of T cell functions (3, 8). Both cDC and CD4 T cells are targets for IL-27–mediated negative regulation. Mascanfroni et al. (7) reported recently that IL-27 signaling in cDC limits Th1 and Th17 differentiation and function. In addition, IL-27 was also reported to induce the expansion and differentiation of Tr1 cells, one of the two major types of regulatory CD4 T cells (8, 9). The Foxp3-IL-10+IFN-γ+ Tr1 cells secrete high levels of IL-10, and both Tr1 differentiation and production of IL-10 are IL-27 dependent (10–13).
Although the induction of IL-27 production has been studied extensively, much less is known about the mechanisms that limit IL-27 expression. Presently, only extracellular ATP acting on purinergic receptors and C5a acting through C5aR were reported to inhibit IL-27 production (14, 15).
PGE2, a major prostanoid derived from arachidonic acid, is produced by immune cells including cDC in response to proinflammatory stimuli and signals through four EP receptors (EP1–4), which vary in affinity, signal duration, and signaling pathway (16–19). The role of PGE2 in DC maturation and function has been studied extensively. In immature DC, PGE2 accelerates maturation and promotes a proinflammatory phenotype (20–22). In terms of DC cytokines and chemokines, PGE2 was reported to reduce IL-12, TNF-α, CCL3, and CCL4 (23–25), while upregulating IL-23 and promoting Th17 function (26–30). The effects of PGE2 on Th1 differentiation depend on the experimental system. In the presence of APCs, PGE2 suppresses Th1 differentiation (23), whereas when purified naive CD4 T cells are differentiated in Th1 polarizing conditions, PGE2 promotes Th1 differentiation (31). This apparent discrepancy can be explained by the fact that in the first system, PGE2 inhibits the APC release of IL-12, a required Th1 differentiation factor, whereas in the second system, which includes exogenous IL-12, PGE2 upregulates the expression of IL-12Rβ2 in differentiating Th1 cells (31). The proinflammatory role of PGE2 has been confirmed in vivo in models of contact hypersensitivity, inflammatory bowel disease, rheumatoid arthritis, and experimental autoimmune encephalomyelitis (EAE), and shown to be mostly associated with increases in Th1/Th17 differentiation and/or function (18, 19, 32, 33).
In this study we report on a novel proinflammatory function of PGE2, mediated through the inhibition of IL-27 production by cDCs. At present, little is known about the interaction between PGE2 and IL-27. Two publications, including one from our laboratory, reported previously that PGE2 inhibited LPS-induced IL-27p28 expression in bone marrow–derived DC (BMDC) (34) and THP-1 cells (35). In the current study we report that PGE2 inhibits IL-27 expression in BMDC and bone marrow–derived macrophages (BMDM), and in splenic DCs in vitro and in vivo. The inhibitory effect of PGE2 is mediated through EP2/EP4 receptors, cAMP induction, and inhibition of IFN regulatory factor (IRF) 1 expression and binding to the p28 promoter IFN-stimulated response element (ISRE) site.
Materials and Methods
C57BL/6 (6–10 wk old) and B6.129S2-Irf1tm1Mak (Irf1−/−) mice were purchased from The Jackson Laboratory. 129S6/SvEv-Stat1tm1Rds (Stat1−/−) and wild type (WT) mice (129S6/SvEv) were from Taconic Farms. C57BL/6 Stat2−/− and corresponding WT mice were provided by Dr. A.M. Gamero (Temple University, Philadelphia, PA). Mice were housed and maintained in accordance with protocols approved by the Institutional Animal Care and Use Committee at Temple University.
PGE2, LPS (Escherichia coli O55:B5), peptidoglycan (Staphylococcus aureus), polyI:C, and DNase were purchased from Sigma-Aldrich. GM-CSF, M-CSF, and IFN-γ were from Peprotech. Dimethyl-PGE2, butaprost, sulprostone, misoprostol, the specific activator of the exchange protein activated by cAMP (EPAC), 8-pCPT-2′-O-Me-cAMP (8-CPT), and the PI3K inhibitor LY294002 were purchased from Cayman Chemical. PKI (5–24) was purchased from Santa Cruz Biotechnology. PKI (6–22), U0126, JNK inhibitor II, EP receptor antagonists PF-04418948 and ONO-AE3-208, and dibutyryl-cAMP were from Calbiochem. Capture and biotinylated anti-mouse IL-27p28 Ab and recombinant mouse IL-27p28 were from R&D Systems. Capture and biotinylated anti-mouse IFN-β Abs were from BioLegend. Streptavidin-HRP was purchased from BioLegend. The tetramethylbenzidine substrate reagent set was from BD Biosciences. Anti-IRF1 and IgG control Abs for chromatin immunoprecipitation (ChIP) were from Santa Cruz Biotechnology. Primary anti-mouse Abs for Western blot were from Cell Signaling Technology (IRF1), Abcam (GAPDH), and BD Biosciences (β-actin). IFN-β was purchased from PBL Assay Science. APC-conjugated anti-mouse CD11c and PE-Cy5-conjugated anti-mouse MHC class II (MHCII) Abs were purchased from eBioscience. FITC-conjugated anti-mouse CD80, CD86, and CD40 were purchased from BD Biosciences.
DC and macrophage culture
Primary DCs were derived from bone marrow cells as previously described (36). Macrophages were derived from bone marrow cells cultured in the presence of 10 ng/ml M-CSF, and adherent cells were collected on day 7. Spleen DCs were isolated from spleens dissociated in the presence of 0.5 mg/ml liberase TL (Roche Diagnostics) and DNase, and magnetically sorted using CD11c MicroBeads per the manufacturer’s instructions (Miltenyi Biotec). Purity was determined by FACS analysis (>93% CD11c+).
Gene expression analysis of il-10, Il12a(p35), Il12b(p40), Il23a(p19), ifnb, p28, irf1, β-actin, and Gapdh was performed using SYBR-green based quantitative RT-PCR (qRT-PCR). The following primer sets were used: Il10 sense 5′-CCTGGTAGAAGTGATGCCCC-3′ and antisense 5′-TCCTTGATTTCTGGGCCATG-3′; Il12a(p35) sense 5′-GAGGACTTGAAGATGTACAG-3′ and antisense 5′-TTCTATCTGTGTGAGGAGGGC-3′; Il12b(p40) sense 5′-GACCCTGCCGATTGAACTGGC-3′ and antisense 5′-CAACGTTGCATCCTAGGATCG-3′; Il23a(p19) sense 5′-TGCTGGATTGCAGAGCAGTAA-3′ and antisense 5′-ATGCAGAGATTCCGAGAGA-3′; ifnb sense 5′-CCCTATGGAGATGACGGAGA-3′ and antisense 5′-ACCCAGTGCTGGAGAAATTG-3′, p28 (in vitro experiments) sense 5′-TCTGGTACAAGCTGGTTCCTGG-3′ and antisense 5′-TAGCCCTGAACCTCAGAGAGCA-3′; p28 (in vivo studies) (37) forward 5′-ATCTCGATTGCCAGGAGTGA-3′ and antisense 5′-GTGGTAGCGAGGAAGCAGAGT-3′; irf1 sense 5′-CCCACAGAAGAGCATAGCAC-3′ and antisense 5′-AGCAGTTCTTTGGGAATAGG-3′; β-actin sense 5′-AGCTTCTTTGCAGCTCCTTCGTTGC-3′ and antisense 5′-ACCAGCGCAGCGATATCGTCA-3′; Gapdh sense 5′-GGAGCGAGACCCCACTAA-3′ and antisense 5′-ACATACTCAGCACCGGCCTC-3′. Expression levels of each gene were calculated using relative standard curves and normalized to either β-actin or Gapdh.
Focused PCR array
BMDCs were treated with LPS (1 μg) with or without PGE2 (10−6 M) for 90 min. mRNA was extracted and reverse transcribed using the RT2 First Strand kit. Expression levels of transcription factors were determined using the RT2 Profiler PCR Array Mouse Transcription Factors.
Cells were treated as indicated and incubated at 4°C for 30 min with anti-mouse CD11c, CD80, CD86, CD40, MCHII, or corresponding isotype controls. Cells were washed three times and collected using BD FACSCalibur. For all experiments, stained cell preparations were compared with isotype controls to identify and gate on positive cell populations. Data were analyzed using FlowJo (Ashland, OR) and GraphPad Prism 5.0.
Cytokine levels in cell culture supernatants were quantified by sandwich ELISA. Cells were plated in 12-well culture plates (1–2 × 106/ml) and stimulated for 24 h. Detection limit was 5 pg/ml for IL-27 and 15 pg/ml for IFN-β.
Western blot analysis
BMDCs (2–3 × 106) were serum starved for 4 h prior to treatment with LPS and PGE2. Samples were lysed, denatured (95°C, 5 min), and loaded on 10% Bis-Tris gels. Following electrophoresis, the protein was transferred to polyvinylidene difluoride membranes (Bio-Rad) and probed overnight with primary Abs. Membranes were washed, incubated with HRP goat anti-rabbit Ig (BD Biosciences), then detected using Immobilon Western chemiluminescent HRP substrate (Millipore). Membranes were then stripped and reprobed for β-actin.
BMDCs (1 × 107) were stimulated for 3–4 h then fixed with 1% formaldehyde for 15 min. Cells were treated with 125 mM glycine for 5 min to stop fixation, washed twice with PBS containing protease inhibitors, lysed, sonicated using a sonic Dismembrator (Fisher Scientific), and processed as previously described (38). Precipitated DNA and 10% input was subjected to quantitative PCR with primers specific to the ISRE site within the p28 promoter: sense 5′-GCTGAAAGTACAAGTAGGACAGAA-3′ and antisense 5′-AGCCATCTCCTGG GTAGG-3′. Results were analyzed using the ΔΔ cycle threshold method.
In vivo experiment
C57BL/6 mice (n = 4–8 per group) were injected i.p. with vehicle (0.4% DMSO in PBS) or dmPGE2 (200 μg/kg) followed by a second injection 4 h later with vehicle or dmPGE2 in addition to LPS (10 mg/kg). CD11c+ cells were purified from spleens 4 h after the second set of injections. mRNA was extracted and subjected to qRT-PCR for p28 expression.
Results represent mean ± SD. Comparisons between groups were performed using unpaired t test or one-way ANOVA followed by Bonferroni correction. Statistical significance was determined with p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001). Graphs were generated and statistical analysis performed using GraphPad Prism 5.0.
PGE2 inhibits IL-27 production in CD11c+ DC and macrophages
The effect of PGE2 on IL-27 production was first tested in BMDCs stimulated with various concentrations of LPS (0.1, 0.5, and 1 μg/ml). We observed a similar inhibition pattern with more than 50% inhibition of IL-27 by PGE2 10−6–10−8 M. In subsequent experiments we used LPS at a concentration of 1 μg/ml unless otherwise noted. The PGE2 concentrations were chosen in agreement with the concentrations reported in vivo in inflammatory conditions (10−6–10−7M) (39–41). PGE2 reduced IL-27 production in a dose-dependent manner in both LPS- and LPS+IFN-γ–stimulated BMDCs (Fig. 1A). A similar inhibitory pattern was observed in BMDCs stimulated with peptidoglycan or polyI:C (Fig. 1B, 1C), indicating that PGE2 affects both MyD88 and TRIF signaling pathways. The inhibition of IL-27 by PGE2 is not restricted to BMDCs. We observed a similar inhibitory pattern in splenic DCs (Fig. 1D) and in BMDM stimulated with LPS, LPS+IFN-γ, and polyI:C (Fig. 1E, 1F).
Because various TLR ligands induce endogenous PGE2, we tested the effect of endogenous PGE2 on IL-27 production in BMDCs treated with LPS in the presence or absence of the cyclooxygenase 2 (COX2) inhibitor NS-398. Although there was some variability from experiment to experiment, the levels of IL-27 were increased in the presence of NS-398 concentrations, which abolished the generation of endogenous PGE2 (Fig. 1G). These results support a role for endogenous PGE2 in the reduction of IL-27 production.
PGE2 reduced mRNA expression levels of p28, but not ebi3, in both LPS− and LPS+IFN-γ–treated DCs (Fig. 2A) and in LPS-treated BMDM (Fig. 2B), indicating that the reduction in IL-27 is transcriptional through inhibition of p28 expression.
To determine if the effect of PGE2 on IL-27 was associated with a general reduction in DC activation, we assessed the effects of PGE2 on MHCII, costimulatory molecules, and cytokine expression. PGE2 had minimal effects on surface MHCII, CD80, and CD86 expression in DCs treated with LPS or LPS+IFN-γ, with the only statistically significant decrease in the percentage of CD40+ cells (Supplemental Fig. 1A) and CD40 median fluorescent intensity (230 with PGE2 versus 390 without PGE2). As expected in terms of cytokine expression, PGE2 increased il-10 and il23a, and had a pronounced inhibitory effect on il12a and il12b expression (Supplemental Fig. 1B).
PGE2 signals through EP2/EP4 and cAMP to reduce IL-27
To examine the role of EP receptors, BMDCs were treated with LPS in the presence of the receptor agonists butaprost (EP2), misoprostol (EP3, EP4 > EP1), PGE1OH (EP4 > EP3), and sulprostone (EP3 > EP1). Butaprost, misoprostol, and PGE1OH, but not sulprostone, inhibited IL-27 production (Fig. 3A), suggesting that EP1 and EP3 were not involved in the inhibition of IL-27. To confirm the role of EP2/EP4, DC were pretreated for 30 min with the selective receptor antagonists ONO-AE3-208 (EP4) and/or PF-04418948 (EP2), followed by treatment with LPS in the presence of PGE2. Addition of either antagonist had slight effects on the inhibition of IL-27, whereas addition of both EP2 and EP4 antagonists completely reversed the inhibitory effect of PGE2 (Fig. 3B), confirming the role of both EP2 and EP4.
Signaling through EP2 and EP4 is known to activate adenylate cyclase leading to increased levels of cAMP (42). To determine the involvement of cAMP in the inhibition of IL-27, we treated DCs with LPS in the presence of the stable, cell-permeable cAMP analog dibutyryl-cAMP (dbcAMP). We found that dbcAMP mimicked the effects of PGE2, inhibiting IL-27 production in a dose-dependent manner (Fig. 3C). Furthermore, addition of the adenylate cyclase activator forskolin also inhibited IL-27 production (Fig. 3D). These results indicate that PGE2 signals through cAMP to inhibit IL-27 production in LPS-stimulated DCs.
Increased intracellular cAMP levels led to activation of a number of downstream signaling pathways, primarily through activation of EPAC and PKA (43, 44). To assess the involvement of EPAC, we treated DCs with LPS in the presence of the selective EPAC activator, 8-CPT. In contrast to PGE2, 8-CPT had no effect on LPS-induced production of IL-27 (Fig. 4A). To determine the role of PKA signaling, we pretreated DCs with the PKA inhibitors PKI (6–22), PKI (5–24), or H89 followed by LPS with or without PGE2. The PKA inhibitors did not affect IL-27 production in LPS-stimulated DCs, and did not reverse the inhibitory effect of PGE2 (Fig. 4B, 4C). These results strongly suggest that the inhibitory effect of PGE2 on IL-27, although cAMP dependent, is not mediated through EPAC or PKA activation. We showed previously that PI3K activation mediated the inhibitory effect of PGE2 on CCL3/4 expression in DCs (45). However, PI3K does not appear to be involved in the inhibition of IL-27, because the PI3K inhibitor LY294002 alone or in combination with PKI (6–22) did not reverse the inhibitory effect of PGE2 (Fig. 4D). In terms of MAPK, none of the three major pathways (i.e., ERK1/2, JNK, and p38 MAPK) appear to be involved, because the use of ERK and JNK inhibitors did not reverse the PGE2 inhibitory effect on p28 expression (Fig. 4E), and we did not observe changes in p38 MAPK phosphorylation upon treatment with PGE2 (results not shown).
Inhibition of IL-27 by PGE2 is mediated through effects on IRF1
Molle et al. (37) reported that IL-27p28 expression following TLR4 signaling is regulated at the transcriptional level in two stages. The first stage (2–4 h) is responsible for the initial p28 expression and involves MyD88-dependent IRF1 and IRF3 activation. The second stage, responsible for sustained p28 expression (6–12 h), depends on IFN-α/β receptor signaling leading to STAT1-dependent sustained IRF1 activation and formation of the IFN-stimulated gene factor 3 (ISGF3) (STAT1/STAT2/IRF9) complex. Prior studies identified a proximal ISRE site within both human and mouse p28 promoters, and established that, in addition to the ISGF3 complex, both IRF1 and IRF3 act as essential p28 transactivating factors by binding to the ISRE site (Fig. 5A) (37, 46–50).
To identify transcription factors affected by PGE2 in LPS-stimulated DCs, we used an RT2 profiler PCR assay and compared early expression (90 min) of various transcription factors in LPS-stimulated DCs treated with or without PGE2. Seven genes were differentially expressed (at levels higher than 2-fold difference), including Irf1, which was downregulated by PGE2 (Supplemental Table I). Irf1 downregulation by PGE2 was confirmed by qRT-PCR (Fig. 5B), Western blot (Fig. 5C) and by flow cytometry (reduction from 71% IRF1+ cells in LPS-treated DCs to 53% following PGE2 treatment, data not shown). In addition, dbcAMP had a similar inhibitory effect as PGE2 on IRF1 mRNA and protein expression (Fig. 5B, 5C).
We further investigated the possible involvement of IRF1 in the PGE2 inhibition of IL-27 by comparing WT and irf1−/− DCs. As expected from the reported role of IRF1 in IL-27p28 expression, irf1−/− DCs expressed much lower levels of p28 mRNA than WT DCs. However, in contrast to WT DCs, the inhibitory effect of PGE2 was lost in irf1−/− DCs (Fig. 5D). Next, we used ChIP assays to determine the effect of PGE2 on IRF1 binding to the ISRE site within the mouse p28 promoter. Stimulation with LPS led to IRF1 binding to the ISRE site, and PGE2 and dbcAMP significantly reduced IRF1 binding (Fig. 5E). These results support IRF1 as a major mediator in the PGE2 inhibition of LPS-induced IL-27 expression. The fact that both PGE2 and dbcAMP have similar inhibitory effects on IL-27 production and on IRF1 binding to the p28 ISRE site supports the role of cAMP as mediator of the inhibitory effect of PGE2 on IL-27 production.
Inhibition of IL-27 and IRF1 by PGE2 is not mediated through effects on STAT1 or STAT2
In DCs, p28 expression has been shown to require the sequential recruitment of IRF1, IRF3, and ISGF3 to the endogenous promoter region (37). STAT1 and STAT2 are essential components of ISGF3 and participate in IRF1 expression (51–53). Therefore, we investigated whether the inhibitory effect of PGE2 on IL-27 is mediated through effects on STAT1 and/or STAT2. We first assessed the effects of PGE2 on STAT1 Tyr701 phosphorylation in LPS-treated DCs by Western blotting and flow cytometry. The results were inconsistent, with slight or no reductions in phosphorylated STAT1. Next, we addressed the role of STAT1/STAT2 by testing the effects of PGE2 in STAT1- and STAT2-deficient DCs. Although WT DCs produced IL-27 in response to LPS and significantly increased IL-27 production upon LPS+IFN-γ treatment, Stat1−/− DCs produced much lower levels of IL-27 (more than 90% reduction) and lost the ability to respond to IFN-γ. However, PGE2 retained its inhibitory activity for IL-27 (Fig. 6A). In terms of irf1 expression, Stat1−/− DCs were unable to increase irf1 expression in response to IFN-γ, but similar to the IL-27 data, PGE2 reduced irf1 in both WT and Stat1−/− DCs (Fig. 6B). These results strongly suggest that PGE2 inhibition of IL-27 and downregulation of irf1 expression are STAT1 independent.
Stat2−/− DCs maintained their capacity to respond to IFN-γ, although IL-27 production was reduced for both LPS and LPS+IFN-γ treatment as compared with WT DCs (Fig. 6C). PGE2 retained its inhibitory capacity in the Stat2−/− DCs (Fig. 6C), which was associated with a reduction in irf1 expression (Fig. 6D), suggesting that, similar to STAT1, the effects of PGE2 on IL-27 and IRF1 are STAT2 independent.
Reduction in endogenous IFN-β by PGE2 plays a minor role in the inhibition of IL-27
Upon stimulation with LPS, DCs produce both IL-27 and IFN-β. Subsequently, IFN-β signaling through IFN-α/β receptor results in the formation of the ISGF3 complex and sustained p28 expression (37). PGE2 suppresses IFN-β production in macrophages (54). In DC, LPS led to expression of both ifnb and p28 with peaks at 1 and 3h, respectively. PGE2 completely reduced p28 expression and IL-27 production, whereas ifnb expression and protein production were only partially reduced (Fig. 7A). To ascertain whether inhibition of IL-27 by PGE2 results from a reduction in DC-derived IFN-β, we supplemented the cultures with exogenous IFN-β. PGE2 retained most of its inhibitory activity (>80% for PGE 10−6 M, and 50% for 10−8 M) (Fig. 7B). We observed a similar pattern in irf1 expression, with PGE2 still reducing irf1 expression in DC treated with LPS in the presence of exogenous IFN-β (Fig. 7C).
PGE2 reduces in vivo IL-27 expression in splenic DC
To determine whether PGE2 affects in vivo IL-27 production in splenic CD11c+ DC, we analyzed p28 expression in mice inoculated with LPS and PGE2. The control group was injected i.p. with vehicle, whereas experimental groups received dmPGE2 (stable PGE2 analog), LPS, or LPS+dmPGE2 (Fig. 8A). Splenic CD11c+ cells were collected four hours later and p28 expression was analyzed by qRT-PCR. DCs from vehicle- and dmPGE2-treated mice did not express p28 mRNA. LPS administration resulted in an increase in p28 expression compared with the vehicle control, and coadministration of dmPGE2 reduced p28 expression (Fig. 8B).
Recently, the regulatory function of IL-27 in both innate and adaptive immunity and its crucial role in the generation of Tr1 cells has received a lot of attention (3, 8, 55). Although the mechanisms involved in the induction of IL-27 in cDCs and macrophages have been studied extensively, little is known about the factors and mechanisms that negatively impact IL-27 expression. To our knowledge, we were the first to report that PGE2 inhibited IL-27 production by BMDCs (56), and a later publication reported on a similar finding in THP-1 cells (35). In this study we report that PGE2 inhibits IL-27 production and p28 expression through a novel mechanism involving EP2/EP4-mediated cAMP increase, subsequent inhibition of irf1 expression and reduced IRF1 binding to the p28 promoter ISRE site. We also observed a similar reduction in p28 expression in vivo in splenic CD11c+ DCs following inoculation of dmPGE2 in conjunction with LPS.
In vivo, PGE2 has been described as a proinflammatory agent in models of contact hypersensitivity, inflammatory bowel disease, rheumatoid arthritis, and EAE, primarily through EP4 signaling leading to Th1/Th17 differentiation (27, 31, 34, 56–59). In vitro proinflammatory effects of PGE2 include DC and CD4 T cells. PGE2 contributes to DC maturation and promotes migration to draining lymph nodes (20, 36, 60). PGE2 also stimulates IL-23 production in DCs and upregulates IL-23R in T cells, promoting the generation of pathogenic Th17 cells (22, 26–29, 61). Although previous reports indicated that PGE2 favors Th2 at the expense of Th1 differentiation primarily through its inhibitory effect on IL-12 production in APCs (23, 62), PGE2 was also reported to facilitate Th1 differentiation in polarizing conditions, which include addition of exogenous IL-12 through the upregulation of IL-12Rβ2 and IFN-γR1 in the differentiating naive T cells (31).
In contrast to the effects on Th1, Th2, and Th17 differentiation, the PGE2 effects on regulatory T cells are less studied, with reports suggesting differences between tumor versus non-tumor microenvironments. Tumor-derived COX2/PGE2 was shown to promote Foxp3+Treg and Foxp3−Tr1 cells in cancer (63–67). In contrast, PGE2 inhibits the differentiation of non-tumor murine Foxp3+Treg and of CD46-induced Tr1 cells (68, 69).
IL-27, expressed and secreted by activated cDCs, was identified as the essential Tr1 differentiation and growth factor, and was shown recently to also affect cDC function (7, 70–72). In DCs, IL-27 reduced costimulatory molecules, IL-12, IL-6, and IL-23, and inhibited DC capacity to generate Th1/Th17 cells (7). In addition, adoptive transfer of IL-27–treated DC reduced EAE and suppressed recall responses of Th1 and Th17 cells (7). In CD4 T cells, IL-27 promoted both Tr1 differentiation and IL-10 production through STAT1/STAT3-mediated upregulation of c-Maf, AhR, and Egr2→Blimp1 (73, 74).
Previously, we reported that PGE2 inhibited IL-27 expression in BMDCs (34), and Zhu et al. (35) reported similar results in THP-1 cells. In this study we report that PGE2 inhibition of IL-27 is similar in BMDCs, BMDMs, and splenic DCs, and occurs upon activation of various TLRs. Both EP2 and EP4 receptors, previously identified in DCs and shown to mediate most PGE2 effects in immune cells, were involved (25, 75, 76). The role of cAMP was confirmed by similar inhibitory effects of PGE2, dbcAMP, and forskolin. However, neither EPAC nor PKA appeared to be involved. The lack of effect of PKA, PI3K, MAPK, and JNK inhibitors and the lack of effect of PGE2 on p38 phosphorylation, suggests that none of these signaling pathways play an essential role in the PGE2-induced inhibition of IL-27.
Upon analyzing mRNA expression of IL-27 subunits we found that PGE2 inhibited p28, but not Ebi3. Expression of mouse p28 is partially dependent on c-Rel binding to the distal NF-κB site. However, binding of IRF1 to the proximal ISRE site represents the essential event for both human and mouse p28 expression (46–49). A temporal analysis of LPS-induced p28 expression showed that the initial burst in p28 transcription requires IRF1 and IRF3, with rapid nuclear translocation of IRF1 being MyD88 dependent (37). This is followed by TRIF/IRF3-mediated production of IFN-β, which then upregulates and sustains STAT1-dependent IRF1 expression, and activates the ISGF3 complex (STAT1/STAT2/IRF9). The latter events serve to sustain p28 expression (37).
We investigated whether PGE2 inhibits IL-27 through effects on IFN-β, IRF1, IRF3, STAT1, and/or STAT2. The effect of PGE2 does not appear to be mediated through reduction in IFN-β. Although we observed PGE2 inhibition of endogenous IFN-β, it was only partial (∼50%) as compared with >90% inhibition of IL-27. Moreover, the inhibitory effect of PGE2 was maintained in the presence of exogenous IFN-β (at concentrations equivalent to, or higher than the endogenous levels observed in the absence of PGE2).
We identified IRF1 as the most important contributor to the effect of PGE2 on p28 expression. In DCs stimulated with LPS and treated with PGE2, PGE2 inhibited irf1 expression as early as 2 h. A significant reduction in intracellular IRF1 protein was also observed by Western blot. In agreement with the crucial role of IRF1 in p28 expression, LPS-treated irf1−/− DC expressed low levels of p28. Moreover, PGE2 lost its suppressive activity for p28 expression in irf1−/− DCs, suggesting that the inhibitory effect was mediated primarily through reduction in IRF1. This was also supported by ChIP analysis, where PGE2 abolished LPS-induced IRF1 binding to the p28 ISRE site. Although STAT1 represents a major factor in the induction of IRF1, it does not play a role in the inhibitory effect of PGE2. In Stat1−/− DCs although IL-27 production was reduced, PGE2 still inhibited IL-27 and irf1 expression. Stat2−/− DCs expressed irf1 at higher levels than WT DCs, presumably due to an increase in STAT1 dimers, but again PGE2 was able to inhibit both IL-27 production and irf1 expression. These results suggest that the PGE2 inhibition of IRF1 and subsequently of p28 expression is not mediated through STAT1 or STAT2.
Most of the effects of PGE2 on immune cells are mediated through induction of cAMP (77, 78). Not surprisingly, we found that, similar to PGE2, dbcAMP and forskolin inhibited IL-27 production in cDCs. What was surprising, however, was the lack of involvement of either PKA or EPAC, the classical signaling molecules downstream of cAMP. We established a link between cAMP and IRF1 by showing that similar to PGE2, dbcAMP inhibits irf1 expression, reduces the levels of intracellular IRF1 protein, and inhibits IRF1 binding to the ISRE site in the p28 promoter. The involvement of cAMP in IRF1 inhibition has been previously reported for two other cAMP inducing agents, i.e., cholera toxin and Bordetella pertussis, without identifying the intermediary signaling molecules (79, 80). We were able to eliminate a number of possible intermediaries such as PKA, EPAC, PI3K, MAPKs, STAT1, and STAT2. However, at present, the link between cAMP and IRF1 in the inhibition of IL-27 by PGE2 in cDC remains to be identified.
The relevance of our findings is reinforced by the fact that we observed similar inhibition of IL-27 in splenic DCs in vitro and in vivo following treatment with exogenous PGE2. In terms of the role of endogenous PGE2, the increase in IL-27 production in cDC stimulated with LPS in the presence of a selective COX2 inhibitor points to a role similar to the exogenous PGE2. This remains to be confirmed in vivo. Because the microsomal PGE2 synthase mPGES1 is the major contributor for the specific generation of inducible PGE2, mPGES1-deficient mice (58) are the most attractive target for investigating the role of endogenous PGE2 in modulating the IL-27 response. However, a more recent report cautions against the assumption that the genetic global deletion of mPGES1 would affect solely PGE2, because mPGES1−/− DC exhibit preferential shunting toward PGD2 production (81). Therefore, the development and use of a conditional temporal mPGES1 knockout might be a better option.
In conclusion, our results indicate that PGE2 inhibits IL-27 production in TLR-activated cDC by reducing p28 expression through the EP2/EP4→cAMP→IRF1 pathway. The inhibition occurs also in the presence of type I and II IFNs, suggesting that the reduction in endogenous IFN-β is not a major contributing factor. In contrast, inhibition of IRF1 mRNA and protein expression and of IRF1 binding to the p28 ISRE site plays a major role in the PGE2 inhibition of IL-27. In an inflammatory milieu, early release of PGE2 due to the activation of inducible COX2 and mPGES1 could be a determining factor in promoting acute inflammation. Among the variety of PGE2-related physiological effects, inhibition of IL-27 in APCs could contribute to a reduction in the differentiation and possibly activation of Tr1 cells, and act as an additional proinflammatory mechanism.
We thank Dr. Italo Tempera and Dr. Kayla Martin (Lewis Katz School of Medicine, Temple University) for assistance with ChIP studies.
This work was supported by National Institutes of Health Grant R01AI047325 (to D.G.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
bone marrow–derived DC
bone marrow–derived macrophage
8-pCPT-2′-O-Me-cAMP (EPAC activator)
experimental autoimmune encephalomyelitis
exchange protein activated by cAMP
IFN regulatory factor
IFN-stimulated gene factor 3
IFN-stimulated response element
MHC class II
The authors have no financial conflicts of interest.