Macrophages activate the production of cytokines and chemokines in response to LPS through signaling cascades downstream from TLR4. Lipid mediators such as PGE2, which are produced during inflammatory responses, have been shown to suppress MyD88-dependent gene expression upon TLR4 activation in macrophages. The study reported here investigated the effect of PGE2 on TLR3- and TLR4-dependent, MyD88-independent gene expression in murine J774A.1 macrophages, as well as the molecular mechanism underlying such an effect. We demonstrate that PGE2 strongly suppresses LPS-induced IFN-β production at the mRNA and protein levels. Poly (I:C)-induced IFN-β and LPS-induced CCL5 production were also suppressed by PGE2. The inhibitory effect of PGE2 on LPS-induced IFN-β expression is mediated through PGE2 receptor subtypes EP2 and EP4, and mimicked by the cAMP analog 8-Br-cAMP as well as by the adenylyl cyclase activator forskolin. The downstream effector molecule responsible for the cAMP-induced suppressive effect is exchange protein directly activated by cAMP (Epac) but not protein kinase A. Moreover, data demonstrate that Epac-mediated signaling proceeds through PI3K, Akt, and GSK3β. In contrast, PGE2 inhibits LPS-induced TNF-α production in these cells through a distinct pathway requiring protein kinase A activity and independent of Epac/PI3K/Akt. In vivo, administration of a cyclooxygenase inhibitor before LPS injection resulted in enhanced serum IFN-β concentration in mice. Collectively, data demonstrate that PGE2 is a negative regulator for IFN-β production in activated macrophages and during endotoxemia.

Prostaglandin E2, produced by macrophages and other cells in response to inflammatory stimuli, has been shown to modulate macrophage activation in part by suppressing the release of cytokines and/or chemokines (1, 2, 3). Several lines of evidence suggest that the production of PGE2 during inflammation constitutes a negative feedback mechanism which limits the production of, among other mediators, TNF-α, IL-1, and CCL4 in immune cells (2, 4, 5).

The production of cytokines and chemokines by macrophages can be initiated through the engagement of the pattern recognition receptors (i.e., TLRs) expressed on these cells. The cytoplasmic domain of TLRs transmits signals downstream via interactions with Toll/IL-1R homology domain-containing adaptor molecules. Among them, MyD88 plays a central role in TLR signaling as it is shared by almost all TLRs. Further advances in the understanding of TLR signaling have identified genes whose induction is independent of MyD88 (6). In this regard, 71% of the LPS-responsive genes in macrophages were shown to be modulated independently of MyD88 (7).

The purpose of the present study was to characterize the effect of PGE2 on LPS-induced, MyD88-independent gene expression and to elucidate the molecular mechanism responsible for such regulation. Experiments focused on IFN-β, the only type I IFN produced by macrophages upon TLR4 activation, and a prototypical MyD88-independent gene. Results show that PGE2, at concentrations found in acute inflammatory sites in vivo, imposes a strong suppression on LPS-induced IFN-β production. In addition, PGE2 suppresses another MyD88-independent gene upon TLR4 stimulation, namely CCL5, and TLR3-mediated, poly (I:C)-induced IFN-β production by J774A.1 cells. Furthermore, findings demonstrate the divergent regulation of PGE2-mediated signaling components on MyD88-dependent and -independent cascades downstream from TLR4 activation in macrophages. Finally, blocking cyclooxygenase (COX)3 activity in vivo results in higher post-LPS serum IFN-β concentration.

The murine macrophage-like J774A.1 cell line was obtained from the American Type Culture Collection. Cells were grown in DMEM (Invitrogen Life Technologies), supplemented with 10% FBS (HyClone), penicillin (100 U/ml), and streptomycin (100 U/ml). Cells were cultured at 105 cells/well in 0.2 ml of culture medium in 96-well plates (BD Biosciences) for supernatant harvesting and at 2 × 106 cells/well in 2 ml of culture medium in 6-well plates (BD Labware) for RNA or protein extraction. Specific cell treatments in the different experiments are described in the figure legends and in Results. Cell viability was determined using Neutral Red uptake at the end of all experiments. None of the treatments affected cell viability.

ELISA was used to measure IFN-β protein accumulation in supernatants harvested from macrophages as described by Weinstein et al. (8). The level of CCL5 was measured with a commercially available ELISA kit, according to the manufacturer’s instruction (R&D Systems). TNF-α production was measured by ELISA, with capture and detection Abs purchased from BD Biosciences and Pierce, respectively.

Total RNA was isolated using the RNeasy Mini extraction kit (Qiagen). cDNA was synthesized from 1 μg of total RNA using the First-Strand cDNA Synthesis kit (GE Healthcare) according to manufacturer’s instructions. Quantitative PCR was performed with SYBR Green quantitative PCR SuperMix (Stratagene) and the Mx4000P QPCR system (Stratagene). PCR primer pairs (Table I) were obtained from Invitrogen Life Technologies. The following cycling conditions were used for the amplification of IFN-β and β-actin: 10 min at 95°C as the initial denaturation step; 15 s at 95°C (1 min for β-actin), 45 s at 59°C and 30 s at 72°C as the amplification step; and a final cooling step down to 4°C. The melting point curve for primer specificity was run for 30 s at 55°C. Primer specificity was confirmed by melting curve analysis and agarose gel electrophoresis. No nonspecific products were observed. Serial dilutions of plasmids containing the cloned PCR products were used to generate standard curves. All the gene expression data presented in Results were normalized to β-actin.

Table I.

Sequences of primer pairs used in RT-PCR

PrimerSequences (5′ → 3′)
IFN-β forward TCCAAGAAAGGACGAACATTCG 
IFN-β reverse TGAGGACATCTCCCACGTCAA 
TNF-α forward CACGCTCTTCTGTCTACTGA 
TNF-α reverse CACTTGGTGGTTTGCTACGA 
EP1 forward CCAACAGGCGATAATGGCAC 
EP1 reverse TGGCGACGAACAACAGGAAG 
EP2 forward TTCATATTCAAGAAACCAGACCCTGGTGGC 
EP2 reverse AGGGAAGAGGTTTCATCCATGTAGGCAAAG 
EP4 forward GACTGGACCACCAACGTAACGGCCTACGCC 
EP4 reverse ATGTCCTCCGACTCTCTGAGCAGTGCTGGG 
β-actin forward TGTGATGGTGGGAATGGGTCAG 
β-actin reverse TTTGATGTCACGCACGATTTCC 
PrimerSequences (5′ → 3′)
IFN-β forward TCCAAGAAAGGACGAACATTCG 
IFN-β reverse TGAGGACATCTCCCACGTCAA 
TNF-α forward CACGCTCTTCTGTCTACTGA 
TNF-α reverse CACTTGGTGGTTTGCTACGA 
EP1 forward CCAACAGGCGATAATGGCAC 
EP1 reverse TGGCGACGAACAACAGGAAG 
EP2 forward TTCATATTCAAGAAACCAGACCCTGGTGGC 
EP2 reverse AGGGAAGAGGTTTCATCCATGTAGGCAAAG 
EP4 forward GACTGGACCACCAACGTAACGGCCTACGCC 
EP4 reverse ATGTCCTCCGACTCTCTGAGCAGTGCTGGG 
β-actin forward TGTGATGGTGGGAATGGGTCAG 
β-actin reverse TTTGATGTCACGCACGATTTCC 

The expression of E prostanoid receptor (EP) subtypes was analyzed by conventional PCR. The cycling conditions included: 3 min at 94°C as the initial denaturation step; 30 s at 94°C, 45 s at 55°C for EP1 (58.5°C for EP2 and 65°C for EP4) and 1 min at 72°C as the amplification step; and a final cooling step down to 4°C.

Macrophages were harvested in cold lysis radio immune precipitation assay buffer (50 mM Tris-HCl (pH 7.4); 150 mM NaCl, 0.25% deoxycholic acid, 1% Nonidet P40, and 1 mM EDTA), together with protease inhibitor mixture (Roche Applied Sciences). Protein concentrations were determined by BCA assay (Pierce). For Western blot analysis, total protein (20 μg) was fractionated by SDS-polyacrylamide gels and was transferred to nitrocellulose membranes (Bio-Rad). The membranes were blocked, washed, and incubated with Akt and phospho-Akt Abs (1/1000 dilution) followed by HRP-conjugated secondary Abs (1/3000 dilution). Signals were detected with ECL Western Blotting Detection Reagents (GE Health) according to manufacturer’s instructions.

Intracellular Ca2+ flux was assessed in real time with the fluorescent probe fluo-4/AM dye (Invitrogen Life Technologies), monitored with a Nikon TE2000U inverted fluorescent microscope (Nikon Instruments). Briefly, 300,000 J774A.1 macrophages were loaded with 3 ng/ml fluo-4/AM dye in HBSS (Invitrogen Life Technologies) supplemented with 10 mM HEPES (pH 7.4) for 30 min at 25°C. Just before use, cells were washed with HBSS plus 10 mM HEPES to remove excess fluo-4/AM. Fluorescent images (ex 489/em 522 nm) were acquired every 5 s along with corresponding bright field images every 30 s, for 40 min at 25°C. Regions were drawn around the cells and total cellular fluorescent intensity was measured and plotted over time.

Intracellular cAMP level was assayed using an enzyme immunoassay kit from Cayman Chemical. In brief, J774A.1 cells were seeded into 12-well plates at 1.5 × 106cells/well and incubated with the phosphodiesterase inhibitor IBMX (2 mM) for 30 min at 37°C. The reaction was started by the addition of PGE2, butaprost, or ONO-AE1–329 for 10 min at 37°C. The reaction was terminated by aspirating the supernatant, and cells were immediately harvested in 0.1 N HCl by scraping. The cell suspension was centrifuged for 10 min at 1000 × g, 4°C. The supernatants were assayed using the cAMP assay kit according to manufacturer’s instructions.

Six- to 8-wk-old Swiss Webster male mice were purchased from Taconic Farms. Experimental protocols were approved by the Institutional Animal Care and Use Committee at Rhode Island Hospital.

Mice were injected with ketorolac (20 mg/kg; Cayman Chemical) or saline i.p. (n = 6/group). One hour later, animals were challenged with LPS (1 mg/kg; i.p.). Blood was harvested 2 h later by cardiac puncture and used for the measurement of IFN-β. Preliminary experiments demonstrated peak serum IFN-β occurred 2 h after LPS exposure.

LPS, PGE2, butaprost, forskolin, 8-bromo-cAMP, LiCl, SB216763, IBMX, and HEPES were obtained from Sigma-Aldrich. Poly (I:C) was from In vivoGen. Protein kinase A (PKA) inhibitors H-89 and KT5720, 8-CPT-2′-O-Me-cAMP, and wortmannin were obtained from Calbiochem. The protein molecular mass markers, as well as penicillin and streptomycin, were obtained from Invitrogen Life Technologies. Anti-phospho-Akt and -Akt Abs were purchased from Upstate Signaling. The HRP-conjugated anti-rabbit Ab was obtained from Cell Signaling. EP1 agonist 17-phenyl trinor PGE2 was purchased from Cayman Chemical and EP4 agonist ONO-AE1–329 was provided by Ono Pharmaceuticals. H2SO4 was obtained from Fisher Scientific.

All experiments were performed at least three times. Cell culture data are means ± SD from quadruplicate samples in a representative experiment. Statistical analysis was by ANOVA, with Dunnett’s or Student-Newman-Keuls post-hoc tests (cell culture experiments) or the Mann-Whitney U test (in vivo experiments). A value of p < 0.05 was considered statistically significant.

PGE2 dose-dependently suppressed LPS (100 ng/ml)-induced IFN-β production (Fig. 1). Endogenous PGE2 did not contribute to the suppressive effect since the addition of the COX inhibitor indomethacin (10 μM) did not alter IFN-β release (data not shown).

FIGURE 1.

PGE2 inhibits LPS-induced IFN-β production in murine J774A.1 cells. Murine J774A.1 macrophages were incubated with PGE2 for 1 h, followed by LPS (100 ng/ml) for 16 h. Supernatants were harvested and IFN-β production was measured by ELISA. Unstimulated J774A.1 cells were used as a negative control and not included in the statistical analysis; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s.

FIGURE 1.

PGE2 inhibits LPS-induced IFN-β production in murine J774A.1 cells. Murine J774A.1 macrophages were incubated with PGE2 for 1 h, followed by LPS (100 ng/ml) for 16 h. Supernatants were harvested and IFN-β production was measured by ELISA. Unstimulated J774A.1 cells were used as a negative control and not included in the statistical analysis; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s.

Close modal

To determine whether the suppressive effect of PGE2 was unique to IFN-β among MyD88-independent genes induced by LPS, the production of chemokine CCL5 was measured in the presence or absence of PGE2. PGE2 was found to dose-dependently reduce LPS-induced CCL5 production (Fig. 2 A).

FIGURE 2.

PGE2 dose-dependently inhibits LPS-induced CCL5 secretion as well as poly (I:C)-induced IFN-β production. A, Cells stimulated with LPS (100 ng/ml) were pretreated or not for 1 h with PGE2. Secreted CCL5 was measured by ELISA after 16 h; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s. B, Cells stimulated with the TLR3 ligand poly (I:C) (50 ng/ml) were pretreated or not for 1 h with PGE2. IFN-β production was measured by ELISA after 16 h; ∗, p < 0.05 vs poly (I:C) alone, ANOVA/Dunnett’s.

FIGURE 2.

PGE2 dose-dependently inhibits LPS-induced CCL5 secretion as well as poly (I:C)-induced IFN-β production. A, Cells stimulated with LPS (100 ng/ml) were pretreated or not for 1 h with PGE2. Secreted CCL5 was measured by ELISA after 16 h; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s. B, Cells stimulated with the TLR3 ligand poly (I:C) (50 ng/ml) were pretreated or not for 1 h with PGE2. IFN-β production was measured by ELISA after 16 h; ∗, p < 0.05 vs poly (I:C) alone, ANOVA/Dunnett’s.

Close modal

Activation of TLR3 can also trigger IFN-β release via a MyD88-independent pathway. A robust IFN-β production was detected by ELISA in cells stimulated with the TLR3 synthetic ligand poly (I:C) (Fig. 2 B) and PGE2 treatment before poly (I:C) stimulation suppressed IFN-β production.

The regulation of IFN-β gene expression occurs mainly at the transcriptional level (9, 10). As shown in Fig. 3, LPS stimulation led to a rapid increase in IFN-β mRNA, with maximum level detected at 4 h. PGE2 pretreatment reduced LPS-induced IFN-β mRNA by 4- and 6-fold at 2 and 4 h, respectively (Fig. 3). Moreover, the suppressive effect of PGE2 on LPS-induced IFN-β mRNA expression was maintained when PGE2 was added simultaneously with or 30 min after LPS (data not shown).

FIGURE 3.

PGE2 suppresses LPS-induced IFN-β gene expression. Total RNA was isolated from J774A.1 macrophages treated or not with 50 ng/ml PGE2 for 1 h, followed by LPS (100 ng/ml) for the times indicated (□, LPS only; ν, ▪ PGE2 + LPS). Real-time quantitative PCR was used to analyze IFN-β and β-actin using primers described in Table I. In LPS-stimulated groups, ∗ indicates p < 0.05 vs 0 h, ANOVA/Dunnett’s. InLPS + PGE2-treated groups, † indicates p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls.

FIGURE 3.

PGE2 suppresses LPS-induced IFN-β gene expression. Total RNA was isolated from J774A.1 macrophages treated or not with 50 ng/ml PGE2 for 1 h, followed by LPS (100 ng/ml) for the times indicated (□, LPS only; ν, ▪ PGE2 + LPS). Real-time quantitative PCR was used to analyze IFN-β and β-actin using primers described in Table I. In LPS-stimulated groups, ∗ indicates p < 0.05 vs 0 h, ANOVA/Dunnett’s. InLPS + PGE2-treated groups, † indicates p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls.

Close modal

Conventional RT-PCR was conducted to characterize the PGE2 receptor expression patterns in J774A.1 cells. As shown in Fig. 4, EP1, EP2, and EP4 mRNAs were detected in unstimulated cells. As reported by Sugimoto et al. (11), LPS stimulation resulted in increased EP2 mRNA expression (Fig. 4). EP3 mRNA was not detected in the cells (data not shown).

FIGURE 4.

PG receptor subtypes EP1, EP2, and EP4 mRNA expression in J774A.1 cells. Total RNA was extracted from J774A.1 cells treated or not with 100 ng/ml LPS for 2 h. Conventional RT-PCR was performed to analyze the expression of EP1, EP2, and EP4 using the primers described in Table I. β-actin was included as a loading control.

FIGURE 4.

PG receptor subtypes EP1, EP2, and EP4 mRNA expression in J774A.1 cells. Total RNA was extracted from J774A.1 cells treated or not with 100 ng/ml LPS for 2 h. Conventional RT-PCR was performed to analyze the expression of EP1, EP2, and EP4 using the primers described in Table I. β-actin was included as a loading control.

Close modal

To test whether EP mRNA expression correlated with receptor function, cells were treated with specific EP agonists. EP1 is known to activate Ca2+ signaling (reviewed in Ref. 12). Treatment of macrophages with the EP1 agonist 17-phenyl trinor PGE2 did not result in significant intracellular Ca2+ influx (Fig. 5, left panel).

FIGURE 5.

EP2 and EP4, but not EP1, are functional in J774A.1 cells. Left panel, Macrophages were loaded with fluo-4/AM as described in Materials and Methods. The effect of EP1 agonist 17-phenyl trinor PGE2 (100 nM, a) was assessed as changes in cytosolic Ca2+ levels, measured over 20 min. Ionomycin (20 μM, b) and KCl (100 μM, c) were given at the end of experiment to produce a Ca2+ spike (positive control). The y-axis represents total cellular fluorescence intensity (TCFL). Right panel, Macrophages (1.5 × 106) were preincubated with phosphodiesterase inhibitor IBMX for 30 min, followed by 10-min treatment with PGE2 (50 ng/ml), butaprost (100 nM; designated as EP2 in the graph), or ONO-AE1–329 (100 nM; designated as EP4 in the graph). Untreated cells were included as a negative control and are shown in the first column. cAMP level was measured using an enzyme immunoassay as described in Materials and Methods; ∗, p < 0.05 vs untreated group, ANOVA/Dunnett’s.

FIGURE 5.

EP2 and EP4, but not EP1, are functional in J774A.1 cells. Left panel, Macrophages were loaded with fluo-4/AM as described in Materials and Methods. The effect of EP1 agonist 17-phenyl trinor PGE2 (100 nM, a) was assessed as changes in cytosolic Ca2+ levels, measured over 20 min. Ionomycin (20 μM, b) and KCl (100 μM, c) were given at the end of experiment to produce a Ca2+ spike (positive control). The y-axis represents total cellular fluorescence intensity (TCFL). Right panel, Macrophages (1.5 × 106) were preincubated with phosphodiesterase inhibitor IBMX for 30 min, followed by 10-min treatment with PGE2 (50 ng/ml), butaprost (100 nM; designated as EP2 in the graph), or ONO-AE1–329 (100 nM; designated as EP4 in the graph). Untreated cells were included as a negative control and are shown in the first column. cAMP level was measured using an enzyme immunoassay as described in Materials and Methods; ∗, p < 0.05 vs untreated group, ANOVA/Dunnett’s.

Close modal

In contrast to EP1, EP2 and EP4 receptors stimulate adenylyl cyclase activity. There was an 8-fold increase in the intracellular cAMP level when macrophages were treated with PGE2 (50 ng/ml) as compared with untreated cells for 10 min (6.1 ± 0.6 vs 44 ± 4.3 pM/1.5 × 106 cells; Fig. 5, right panel). Both butaprost, an EP2 agonist, and ONO-AE1–329, an EP4 agonist (at 100 nM for 10 min) elevated intracellular cAMP, demonstrating the functionality of the receptors.

The EP2- and EP4-specific agonists mentioned above were used to examine the roles of these receptors in mediating the inhibition of LPS-induced IFN-β production by PGE2. Similar to the effect of PGE2, either agonist dose-dependently suppressed LPS-induced IFN-β production (Fig. 6). Furthermore, the relative magnitude of the inhibition of LPS-induced IFN-β production imposed by EP2- and EP4-specific agonists correlated with the levels of intracellular cAMP produced (Fig. 5). In contrast, EP1-specific agonist 17-phenyl trinor PGE2 (100 nM) failed to alter LPS-induced IFN-β mRNA expression when added alone or together with EP2- or EP4-specific agonists (data not shown).

FIGURE 6.

The inhibitory effect of PGE2 on LPS-induced IFN-β production is mediated by PG receptor subtypes EP2 and EP4. Cells were treated with LPS, or with PGE2 (50 ng/ml), the EP2-specific agonist butaprost, or the EP4-specific agonist ONO-AE1–329 for 1 h, followed by LPS. IFN-β production was measured by ELISA on supernatants collected after overnight LPS exposure; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s.

FIGURE 6.

The inhibitory effect of PGE2 on LPS-induced IFN-β production is mediated by PG receptor subtypes EP2 and EP4. Cells were treated with LPS, or with PGE2 (50 ng/ml), the EP2-specific agonist butaprost, or the EP4-specific agonist ONO-AE1–329 for 1 h, followed by LPS. IFN-β production was measured by ELISA on supernatants collected after overnight LPS exposure; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s.

Close modal

A cell membrane-permeable cAMP analog, 8-Br-cAMP, was used to test the hypothesis that an elevation of intracellular cAMP level is necessary for the effect of PGE2 on IFN-β production. The real-time quantitative RT-PCR data shown in Fig. 7,A illustrate that forskolin, an activator of adenylyl cyclase, could mimic the effect of PGE2 by suppressing LPS-induced IFN-β production in a dose-dependent manner. cAMP analog 8-Br-cAMP had a similar effect (Fig. 7 B).

FIGURE 7.

Both the cAMP analog 8-bromo-cAMP and forskolin mimic the inhibitory effect of PGE2 on LPS-induced IFN-β gene expression and production. A, Cells were treated with the adenylyl cyclase activator forskolin for 1 h, followed by LPS (100 ng/ml) for 2 h. Total RNA was isolated and reverse transcribed. IFN-β gene expression was measured by real-time PCR and normalized to β-actin; ∗, p < 0.05 vs untreated group, ANOVA/Dunnett’s. B, J774A.1 macrophages were incubated with LPS (100 ng/ml) alone, or with PGE2 (50 ng/ml) or cAMP analog 8-bromo-cAMP at the indicated concentrations, followed by LPS (100 ng/ml). Supernatants were harvested after 16 h and IFN-β was determined by ELISA; ∗, p < 0.05 vs untreated group, ANOVA/Dunnett’s.

FIGURE 7.

Both the cAMP analog 8-bromo-cAMP and forskolin mimic the inhibitory effect of PGE2 on LPS-induced IFN-β gene expression and production. A, Cells were treated with the adenylyl cyclase activator forskolin for 1 h, followed by LPS (100 ng/ml) for 2 h. Total RNA was isolated and reverse transcribed. IFN-β gene expression was measured by real-time PCR and normalized to β-actin; ∗, p < 0.05 vs untreated group, ANOVA/Dunnett’s. B, J774A.1 macrophages were incubated with LPS (100 ng/ml) alone, or with PGE2 (50 ng/ml) or cAMP analog 8-bromo-cAMP at the indicated concentrations, followed by LPS (100 ng/ml). Supernatants were harvested after 16 h and IFN-β was determined by ELISA; ∗, p < 0.05 vs untreated group, ANOVA/Dunnett’s.

Close modal

A well-characterized signaling target downstream from cAMP is cAMP-responsive PKA. To investigate whether PKA was involved in the PGE2-mediated suppression of LPS-induced IFN-β production, cells were treated with PKA inhibitors H89 or KT-5720 before LPS. Neither treatment could reverse the inhibitory effect of PGE2 on LPS-induced IFN-β gene expression or protein release (Fig. 8 A; KT-5720 data not shown).

FIGURE 8.

The inhibitory effect of PGE2 on LPS-induced IFN-β gene expression involves Epac but not PKA. A, Cells were pretreated or not with the PKA inhibitor H89 (10 μM) for 30 min, followed by PGE2 (50 ng/ml, 1 h), and then LPS for 2 h. Control groups included unstimulated cells, and cells stimulated with LPS alone. Total RNA was isolated and subjected to real-time quantitative RT-PCR using primers targeting IFN-β and β-actin; ∗, p < 0.05 vs control group; †, p < 0.05 vs LPS-alone group, ANOVA/Student-Newman-Keuls. B, Cells were treated with Epac-specific activator 8-CPT-2′-O-Me-cAMP for 1 h followed by 2 h of LPS. Control groups included LPS alone, or 1 h of PGE2 followed by 2 h LPS. Total RNA was isolated, reverse transcribed, and subjected to real-time quantitative PCR using primers targeting IFN-β and β-actin; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s.

FIGURE 8.

The inhibitory effect of PGE2 on LPS-induced IFN-β gene expression involves Epac but not PKA. A, Cells were pretreated or not with the PKA inhibitor H89 (10 μM) for 30 min, followed by PGE2 (50 ng/ml, 1 h), and then LPS for 2 h. Control groups included unstimulated cells, and cells stimulated with LPS alone. Total RNA was isolated and subjected to real-time quantitative RT-PCR using primers targeting IFN-β and β-actin; ∗, p < 0.05 vs control group; †, p < 0.05 vs LPS-alone group, ANOVA/Student-Newman-Keuls. B, Cells were treated with Epac-specific activator 8-CPT-2′-O-Me-cAMP for 1 h followed by 2 h of LPS. Control groups included LPS alone, or 1 h of PGE2 followed by 2 h LPS. Total RNA was isolated, reverse transcribed, and subjected to real-time quantitative PCR using primers targeting IFN-β and β-actin; ∗, p < 0.05 vs LPS alone, ANOVA/Dunnett’s.

Close modal

cAMP-mediated signaling also occurs through Epac (13). The expression of Epac in murine macrophages has been documented (14). Real-time quantitative RT-PCR demonstrated that the Epac-specific activator 8-CPT-2′-O-Me-cAMP dose-dependently suppressed LPS-induced IFN-β gene expression (Fig. 8 B). Because 8-CPT-2′-O-Me-cAMP does not activate PKA, the data indicate that Epac represents the downstream effector molecule mediating the inhibitory effect of PGE2 on LPS-induced IFN-β production.

PI3K is considered a negative regulator of TLR signaling (15). Moreover, there is evidence that activation of Epac by cAMP can signal through the PI3K/Akt pathway (16). The effect of PGE2 on PI3K activation was assessed by Western blot analysis using a phospho-specific Ab against residue serine 473 of Akt. Results revealed that treatment of macrophages with PGE2 increased phosphorylation of Akt in a time-dependent manner (Fig. 9,A). The ability of PGE2 to phosphorylate/activate Akt has been reported elsewhere with similar kinetics to those shown in Fig. 8,A (17). To discern the EP receptor responsible for Akt activation, cells were stimulated with either butaprost or ONO-AE1–329, and the two agonists induced Akt activation with similar kinetics (Fig. 9 A). Although the activation of PI3K by EP4 agonist has been reported (18), results described here show that such response can also be elicited by an EP2 agonist.

FIGURE 9.

The role of PI3K/Akt/GSK3β in the suppression of IFN-β in macrophages. A, Cells were treated with PGE2 (50 ng/ml), EP2-specific agonist butaprost (100 nM; EP2 in the graph), EP4-specific agonist ONO-AE1–329 (100 nM; EP4 in the graph), or 8-Br-cAMP (100 μM) for the indicated times. Whole cell lysates were harvested, separated by SDS-PAGE, and subjected to Western blot analysis using an Ab recognizing phospho-Akt (Ser473). The same blot was stripped and reblotted with anti-Akt Ab. B, Cells were stimulated with the Epac-specific activator 8-CPT-2′-O-Me-cAMP (250 μM) following pretreatment with the PI3K inhibitor wortmannin (1 μM) for 30 min. Whole cell lysates were harvested for phospho-Akt Western blot. Loading control was obtained by stripping and reblotting for total Akt. C, Cells were untreated, incubated with LPS, or with PGE2 for 1 h followed by LPS for the indicated time points. Western blot analysis was conducted using phospho-Akt and total Akt Abs.

FIGURE 9.

The role of PI3K/Akt/GSK3β in the suppression of IFN-β in macrophages. A, Cells were treated with PGE2 (50 ng/ml), EP2-specific agonist butaprost (100 nM; EP2 in the graph), EP4-specific agonist ONO-AE1–329 (100 nM; EP4 in the graph), or 8-Br-cAMP (100 μM) for the indicated times. Whole cell lysates were harvested, separated by SDS-PAGE, and subjected to Western blot analysis using an Ab recognizing phospho-Akt (Ser473). The same blot was stripped and reblotted with anti-Akt Ab. B, Cells were stimulated with the Epac-specific activator 8-CPT-2′-O-Me-cAMP (250 μM) following pretreatment with the PI3K inhibitor wortmannin (1 μM) for 30 min. Whole cell lysates were harvested for phospho-Akt Western blot. Loading control was obtained by stripping and reblotting for total Akt. C, Cells were untreated, incubated with LPS, or with PGE2 for 1 h followed by LPS for the indicated time points. Western blot analysis was conducted using phospho-Akt and total Akt Abs.

Close modal

To investigate the downstream signaling components responsible for the activation of Akt, the cAMP analog 8-Br-cAMP as well as the Epac-specific activator 8-CPT-2′-O-Me-cAMP were used to stimulate J774A.1 cells. Akt activation/phosphorylation was induced by both compounds (Fig. 9, A and B). Akt phosphorylation was mediated exclusively through PI3K because wortmannin (a PI3K-specific inhibitor) abolished Akt phosphorylation in response to PGE2, butaprost, ONO-AE1–329, 8-Br-cAMP, or 8-CPT-2′-O-Me-cAMP (data shown for 8-CPT-2′-O-Me-cAMP, Fig. 9,B). Fig. 9 C shows that LPS triggered maximal level of Akt phosphorylation/activation at 15 min, and that PGE2 prolonged LPS-induced Akt activation in the cells.

To directly test the involvement of PI3K/Akt on LPS-induced IFN-β production, cells were pretreated with the PI3K inhibitor wortmannin. Real-time quantitative RT-PCR results indicated that wortmannin completely reversed the inhibitory effect of PGE2 on LPS-induced IFN-β gene expression (Fig. 10 A). Furthermore, the normalized level of IFN-β mRNA was twice as high in wortmannin-treated cells than in LPS-treated controls (12.2 ± 3.0 vs 6.6 ± 1.3).

FIGURE 10.

PI3K activity suppresses LPS-induced IFN-β gene expression and GSK3β is an integral component of such suppression. Cells were incubated or not with PGE2 following pretreatment with the PI3K inhibitor wortmannin (1 μM) (A) or GSK3β inhibitors LiCl or SB216763 (B) for 30 min, and then stimulated with LPS (100 ng/ml). Total RNA was harvested 2 h later. Real-time quantitative RT-PCR was used to analyze IFN-β and β-actin as described in Materials and Methods; ∗, p < 0.05 vs untreated group; †, p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls.

FIGURE 10.

PI3K activity suppresses LPS-induced IFN-β gene expression and GSK3β is an integral component of such suppression. Cells were incubated or not with PGE2 following pretreatment with the PI3K inhibitor wortmannin (1 μM) (A) or GSK3β inhibitors LiCl or SB216763 (B) for 30 min, and then stimulated with LPS (100 ng/ml). Total RNA was harvested 2 h later. Real-time quantitative RT-PCR was used to analyze IFN-β and β-actin as described in Materials and Methods; ∗, p < 0.05 vs untreated group; †, p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls.

Close modal

GSK3β is a constitutively active protein kinase which becomes inactive upon Akt phosphorylation (19). To mimic Akt-induced inactivation, GSK3β inhibitors LiCl or SB216763 were used to pretreat cells followed by LPS stimulation. Real-time quantitative RT-PCR showed that inactivating GSK3β reduced LPS-induced IFN-β gene expression, an effect similar to that of PGE2 (Fig. 10 B).

TNF-α is a prototypical cytokine induced via the MyD88-dependent signaling pathway downstream from TLR4. PGE2 has been reported to suppress LPS-induced TNF-α expression and production (2). Real-time quantitative RT-PCR revealed that, similar to the regulation of IFN-β, both EP2 and EP4 receptors were involved in the PGE2-mediated suppression of TNF-α gene expression (Fig. 11,A). Moreover, 8-Br-cAMP mimicked the inhibitory effect of PGE2 (Fig. 11,B). In contrast with findings on IFN-β, pretreatment with the PKA inhibitor H89 reversed the inhibitory effect of PGE2 on LPS-induced TNF-α expression, whereas the Epac activator did not have an effect. Furthermore, treatment with a PI3K inhibitor before LPS stimulation did not alter TNF-α gene expression (Fig. 11 B).

FIGURE 11.

LPS-induced TNF-α gene expression is differentially regulated in response to PGE2. Cells were treated with PGE2 (50 ng/ml), EP2- or EP4-specific agonists (100 nM butaprost and 100 nM ONO-AE1–329) or with 8-Br-cAMP (100 μM), PKA inhibitor H89 (10 μM), Epac activator 8-CPT-2′-O-Me-cAMP (250 μM), or PI3K inhibitor wortmannin (1 μM; B) for 1 h, followed by LPS (100 ng/ml) stimulation for 2 h. Total RNA was extracted and subjected to reverse transcription. Real-time quantitative PCR was conducted to analyze TNF-α gene expression, normalized to β-actin as described in Materials and Methods. In A, ∗, p < 0.05 vs untreated group; †, p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls test. In B, ∗, p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls test.

FIGURE 11.

LPS-induced TNF-α gene expression is differentially regulated in response to PGE2. Cells were treated with PGE2 (50 ng/ml), EP2- or EP4-specific agonists (100 nM butaprost and 100 nM ONO-AE1–329) or with 8-Br-cAMP (100 μM), PKA inhibitor H89 (10 μM), Epac activator 8-CPT-2′-O-Me-cAMP (250 μM), or PI3K inhibitor wortmannin (1 μM; B) for 1 h, followed by LPS (100 ng/ml) stimulation for 2 h. Total RNA was extracted and subjected to reverse transcription. Real-time quantitative PCR was conducted to analyze TNF-α gene expression, normalized to β-actin as described in Materials and Methods. In A, ∗, p < 0.05 vs untreated group; †, p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls test. In B, ∗, p < 0.05 vs LPS alone, ANOVA/Student-Newman-Keuls test.

Close modal

The biological relevance of in vitro results was tested in vivo. Animals were given the nonselective COX inhibitor ketorolac (20 mg/kg) before LPS challenge (1 mg/kg) and the serum concentration of IFN-β was measured 2 h later. ELISA analysis showed that IFN-β level was higher (184 ± 41 IU/ml) in ketorolac-treated animals than in saline-injected controls (104 ± 49 IU/ml) (p < 0.05, Mann-Whitney U test). The serum IFN-β level in naive animals was below 5 IU/ml (data not shown).

Lipid mediators such as PGE2 can regulate immune and inflammatory responses by modulating the production of cytokines and chemokines. PGE2 has previously been shown to suppress MyD88-dependent proinflammatory gene expression, including TNF-α, IL-1, and CCL4, by macrophages (2, 4, 5). In this study, we report that PGE2 modulates TLR3- and TLR4-dependent, MyD88-independent gene expression in these cells.

PGE2 was found to exhibit a strong inhibitory effect on LPS-induced IFN-β mRNA and protein production in cultured murine J774A.1 macrophages (Fig. 1). The suppressive effect of PGE2 was dose-dependent and occurred in a concentration range that is physiologically relevant (fluids collected from sterile wounds in mice contain 54 ± 14 pg/ml PGE2, our unpublished observation). The finding that PGE2 could suppress LPS-induced CCL5 production, as well as TLR3-dependent IFN-β production indicates that the inhibitory effect of PGE2 is not restricted to IFN-β or to TLR4 ligand, but extends to other genes regulated via MyD88-independent signaling cascade, as well as to TLR3 activation (Fig. 2).

PGE2 exerts its biological actions by binding to EPs located mainly on the plasma membrane (20). Although EP1 mRNA expression was detected in J774A.1 cells, 17-phenyl trinor PGE2 (EP1 agonist) failed to trigger intracellular Ca2+ influx, suggesting that EP1 may not be a functional receptor in J774A.1 cells. Moreover, 17-phenyl trinor PGE2 cannot duplicate the effect of PGE2 on LPS-induced IFN-β production. EP2 and EP4 receptors are coupled to the stimulation of adenylyl cyclase activity via Gs protein, leading to elevations of intracellular cAMP (21, 22). The presence of PGE2 receptor subtypes 2 and 4 (EP2 and EP4) in J774A.1 macrophages has been reported elsewhere (23) and was confirmed here (Fig. 4). Stimulation with either EP2- or EP4-specific agonists increased intracellular cAMP in the cells and had a suppressive effect on LPS-triggered IFN-β production (Fig. 5). Half-maximal inhibition of LPS-induced IFN-β production was obtained with 100 nM butaprost and with <10 nM ONO-AE1–329 (Fig. 6). Both binding affinity (12) and expression level of EP4 mRNA could explain the stronger potency of ONO-AE1–329, a conclusion supported by the higher intracellular cAMP formation after EP4 stimulation.

The activation of EP2 and EP4 receptors leads to increases in intracellular cAMP. That both the adenylyl cyclase activator forskolin and the cAMP analog 8-Br-cAMP could mimic the inhibitory effect of PGE2 on LPS-induced IFN-β gene expression and protein production (Fig. 7) indicated that cAMP is required for the inhibitory effect of PGE2. cAMP signals through the recruitment of intracellular protein targets, including at least PKA and Epac (13, 24). The use of H89, a specific inhibitor of PKA, did not lead to reversal of the inhibitory effect of PGE2 on LPS-induced IFN-β expression or protein production (Fig. 8,A). The lack of H89 effect was not due to the absence of PKA in these cells because PKA was found to be essential in the regulation of LPS-induced TNF-α gene expression by PGE2 (Fig. 11). In contrast, the Epac activator, 8-CPT-2′-O-Me-cAMP suppressed LPS-induced IFN-β production in a dose-dependent manner (Fig. 8 B), similar to the effects brought about by PGE2 itself and by a cAMP analog. Taken together, the results indicate that cAMP mediates the suppressive effect of PGE2 on IFN-β via an Epac-dependent, PKA-independent pathway in J774A.1 cells.

cAMP has been reported to mediate PI3K activation through Epac (25). Fig. 9 demonstrates that PGE2, EP2/EP4 agonists, a cAMP analog, or an Epac-specific activator can induce PI3K/Akt activation. Paradoxically, LPS can also trigger Akt phosphorylation (Fig. 9,C). It is, therefore, unclear as how both LPS and PGE2 can activate PI3K/Akt activities, yet the two treatments (LPS vs PGE2 plus LPS) produced opposite effects on IFN-β production. The duration and/or magnitude of Akt activation may be critical in determining the level of IFN-β production. Furthermore, additional regulatory pathway(s) could be activated or suppressed in response to LPS and/or PGE2, leading to different effects on IFN-β production. That blocking PI3K activity with wortmannin resulted in enhanced IFN-β mRNA expression provides direct evidence that PI3K is a negative regulator for IFN-β production in macrophages (Fig. 10 A).

GSK3β is a serine/threonine kinase whose activity is inhibited by Akt-dependent phosphorylation. We found that inhibiting GSK3β activity could mimic the effect of PGE2 (Fig. 10 B), supporting the hypothesis that PGE2 inhibits LPS-induced IFN-β gene expression through a PI3K/Akt/GSK3β-signaling pathway. GSK3β has been implicated in the control of p65/NFκB transcriptional activity in the context of TNF-α signaling (26). However, the role of GSK3β in the regulation of IFN-β expression in macrophages is currently undefined.

The regulation of signaling components downstream from PGE2 and EPs on LPS-induced, MyD88-dependent TNF-α was also analyzed. Data presented in Fig. 11 showed a pattern of differential regulation of TNF-α vs IFN-β, where the point of divergence occurs downstream from cAMP. In contrast to findings on IFN-β, blocking PKA activity with H89 completely reversed the inhibitory effect of PGE2 on LPS-induced TNF-α gene expression, whereas the Epac activator had no effect. Moreover, PI3K and its downstream signaling components Akt and GSK3β were not involved in the suppression of LPS-induced TNF-α production by PGE2 (Fig. 11). The divergent regulation of type I IFN and TNF-α expression is indicative of the tight regulation under which immune cells function. Although activated by a common upstream second messenger cAMP, differential modulation allows PKA and Epac to exert different effects on their downstream targets. The involvement of distinct intracellular pathways resulting in the regulation of MyD88-dependent and -independent genes by PGE2 provides potential targets for therapies directed toward the regulation of inflammatory responses. In conclusion, the present study demonstrates that PGE2 negatively regulates the production of type I IFN (IFN-β) through EP2 and EP4 in murine macrophages, and in vivo in LPS-injected mice. These findings confirm a substantial role for PGE2 in modulating the magnitude of inflammatory responses.

We thank Nicole Morin for her assistance in performing Ca2+ influx experiment and Dr. Jean M. Daley for critical review of the manuscript.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by National Institutes of Health Grant GM-42859 (to J.E.A).

3

Abbreviations used in this paper: COX, cyclooxygenase; PKA, protein kinase A; Epac, exchange protein directly activated by cAMP; EP, E prostanoid receptor.

1
Kunkel, S. L., M. Spengler, M. A. May, R. Spengler, J. Larrick, D. Remick.
1988
. Prostaglandin E2 regulates macrophage derived tumor necrosis factor gene expression.
J. Biol. Chem.
263
:
5380
-5384.
2
Kunkel, S. L., R. C. Wiggins, S. W. Chensue, J. Larrick.
1986
. Regulation of macrophage tumor necrosis factor production by prostaglandin E2.
Biochem. Biophys. Res. Commun.
137
:
404
-410.
3
Bogdan, C., Y. Vodovotz, C. F. Nathan.
1991
. Macrophage deactivation by interleukin 10.
J. Exp. Med.
174
:
1549
-1555.
4
Brandwein, S. R..
1986
. Regulation of interleukin 1 production by mouse peritoneal macrophages.
J. Biol. Chem.
261
:
8624
-8632.
5
Jing, H., E. Vassiliou, D. Ganea.
2003
. Prostaglandin E2 inhibits production of the inflammatory chemokines CCL3 and CCL4 in dendritic cells.
J. Leukocyte Biol.
74
:
868
-879.
6
Kawai, T., O. Adachi, T. Ogawa, K. Takeda, S. Akira.
1999
. Unresponsiveness of MyD88-deficient mice to endotoxin.
Immunity
11
:
115
-122.
7
Bjorkbacka, H., K. A. Fitzgerald, F. Huet, X. Li, J. A. Gregory, M. A. Lee, C. M. Ordija, N. E. Dowley, D. T. Golenbock, M. W. Freeman.
2004
. The induction of macrophage gene expression by LPS predominantly uses Myd88-independent signaling cascades.
Physiol Genomics
19
:
319
-330.
8
Weinstein, S. L., A. J. Finn, S. H. Dave, F. Meng, C. A. Lowell, J. S. Sanghera, A. L. DeFranco.
2000
. Phosphatidylinositol 3-kinase and mTOR mediate lipopolysaccharide-stimulated nitric oxide production in macrophages via interferon-β.
J. Leukocyte Biol.
67
:
405
-414.
9
Raj, N. B., P. M. Pitha.
1983
. Two levels of regulation of β-interferon gene expression in human cells.
Proc. Natl. Acad. Sci. USA
80
:
3923
-3927.
10
Whittemore, L. M., T. Maniatis.
1990
. Postinduction turnoff of β-interferon gene expression.
Mol. Cell. Biol.
64
:
1329
-1337.
11
Sugimoto, Y., S. Narumiya, A. Ichikawa.
2000
. Distribution and function of prostanoid receptors: studies from knockout mice.
Prog. Lipid Res.
39
:
289
-314.
12
Tsuboi, K., Y. Sugimoto, A. Ichikawa.
2002
. Prostanoid receptor subtypes.
Prostaglandins Other Lipid Mediat.
68–69
:
535
-556.
13
de Rooij, J., F. J. T. Zwartkruis, M. H. G. Verheijen, R. H. Cool, S. M. B. Nijman, A. Wittinghofer, J. L. Bos.
1998
. Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP.
Nature
396
:
474
-477.
14
Aronoff, D. M., C. Canetti, C. H. Serezani, M. Luo, M. Peters-Golden.
2005
. Macrophage inhibition by cyclic AMP (cAMP): differential roles of protein kinase A and exchange protein directly activated by cAMP-1.
J. Immunol.
174
:
595
-599.
15
Fukao, T. K., S. Koyasu.
2003
. PI3K and negative regulation of TLR signaling.
Trends Immunol.
24
:
358
-363.
16
Jing, H., J.-H. Yen, D. Ganea.
2004
. A novel signaling pathway mediates the inhibition of CCL3/4 expression by prostaglandin E2.
J. Biol. Chem.
279
:
55176
-55186.
17
Li, J., S. Yang, T. R. Billiar.
2000
. Cyclic nucleotides suppress tumor necrosis factor α-mediated apoptosis by inhibiting caspase activation and cytochrome c release in primary hepatocytes via a mechanism independent of Akt activation.
J. Biol. Chem.
275
:
13026
-13034.
18
Fujino, H., K. A. West, J. W. Regan.
2002
. Phosphorylation of glycogen synthase kinse-3 and stimulation of T-cell factor signaling following activation of EP2 and EP4 prostanoid receptors by prostaglandin E2.
J. Biol. Chem.
277
:
2614
-2619.
19
Cross, D. A. E., D. R. Alessi, P. Cohen, M. Andjelkovich, B. A. Hemmings.
1995
. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B.
Nature
378
:
785
-789.
20
Narumiya, S..
1994
. Prostanoid receptors: structure, function and distribution.
Ann. NY Acad. Sci.
744
:
126
-138.
21
Nishigaki, N., M. Negishi, A. Honda, Y. Sugimoto, T. Namba, S. Narumiya, A. Ichikawa.
1995
. Identification of prostaglandin E receptor “EP2” cloned from mastocytoma cells as EP4 subtype.
FEBS Lett.
364
:
339
-341.
22
Honda, A., Y. Sugimoto, T. Namba, A. Watabe, A. Irie, M. Negishi, S. Narumiya, A. Ichikawa.
1993
. Cloning and expression of a cDNA for mouse prostaglandin E receptor EP2 subtype.
J. Biol. Chem.
268
:
7759
-7762.
23
Katsuyama, M., R. Ikegami, H. Karahashi, F. Amano, Y. Sugimoto, A. Ichikawa.
1998
. Characterization of the LPS-stimulated expression of EP2 and EP4 prostaglandin E receptors in mouse macrophage-like cell line, J774.1.
Biochem. Biophys. Res. Commun.
251
:
727
-731.
24
Daniel, P. B., W. H. Walker, J. F. Habener.
1998
. Cyclic AMP signaling and gene regulation.
Annu. Rev. Nutr.
18
:
353
-383.
25
Mei, F. C., J. Qiao, O. M. Tsygankova, J. L. Meinkoth, L. A. Quilliam, X. Cheng.
2002
. Differerntial signaling of cyclic AMP.
J. Biol. Chem.
277
:
11497
-11504.
26
Hoeflich, K. P., J. Luo, E. A. Rubie, M. S. Tsao, O. Jin, J. R. Woodgett.
2000
. Requirement for glycogen synthase kinase-3β in cell survival and NF-κB activation.
Nature
406
:
86
-90.