EP2 and EP4 receptors on muscularis resident macrophages mediate LPS-induced intestinal dysmotility via iNOS upregulation through cAMP/ERK signals.

Intestinal resident macrophages play an important role in gastrointestinal dysmotility by producing prostaglandins (PGs) and nitric oxide (NO) in inflammatory conditions. The causal correlation between PGs and NO in gastrointestinal inflammation has not been elucidated. In this study, we examined the possible role of PGE(2) in the LPS-inducible inducible NO synthase (iNOS) gene expression in murine distal ileal tissue and macrophages. Treatment of ileal tissue with LPS increased the iNOS and cyclooxygenase (COX)-2 gene expression, which lead to intestinal dysmotility. However, LPS did not induce the expression of iNOS and COX-2 in tissue from macrophage colony-stimulating factor-deficient op/op mice, indicating that these genes are expressed in intestinal resident macrophages. iNOS and COX-2 protein were also expressed in dextran-phagocytized macrophages in the muscle layer. CAY10404, a COX-2 inhibitor, diminished LPS-dependent iNOS gene upregulation in wild-type mouse ileal tissue and also in RAW264.7 macrophages, indicating that PGs upregulate iNOS gene expression. EP(2) and EP(4) agonists upregulated iNOS gene expression in ileal tissue and isolated resident macrophages. iNOS mRNA induction mediated by LPS was decreased in the ileum isolated from EP(2) or EP(4) knockout mice. In addition, LPS failed to decrease the motility of EP(2) and EP(4) knockout mice ileum. EP(2)- or EP(4)-mediated iNOS expression was attenuated by KT-5720, a PKA inhibitor and PD-98059, an ERK inhibitor. Forskolin or dibutyryl-cAMP mimics upregulation of iNOS gene expression in macrophages. In conclusion, COX-2-derived PGE(2) induces iNOS expression through cAMP/ERK pathways by activating EP(2) and EP(4) receptors in muscularis macrophages. NO produced in muscularis macrophages induces dysmotility during gastrointestinal inflammation.

inflammatory cytokines, such as interleukins and tumor necrosis factor-α, and chemical inflammatory mediators, such as bacterial wall components, prostaglandins (PGs), and nitric oxide (NO), induce gastrointestinal inflammation. Lipopolysaccharide (LPS) from gram-negative bacteria is a major causative factor of gastrointestinal inflammation (5). LPS stimulation activates nuclear factor κB (NF-κB) via a toll-like receptor-4 (TLR-4)-mediated signaling cascade that induces inflammatory-related substances, such as tumor necrosis factor-α, interleukin-1β, monocyte chemotactic protein-1, PGs, and inducible NO synthase (iNOS) (19, 22). Monocytes/macrophages are one of the most LPS-sensitive types of inflammatory cells.

In cases of peritonitis and postoperative ileus, the intestinal lumen is invaded by inflammatory cells, such as intestinal muscularis macrophages, from the outside of the intestinal wall. The intestinal muscle layer contains a dense network of ramified resident macrophages in the serosa, myenteric plexus, and interior muscle region (29, 31, 38). In tissue culture studies of the small intestinal muscle layer, which rule out the influence of infiltrating cells, LPS stimulation upregulates the expression of the cyclooxygenase (COX)-2 and iNOS genes in resident muscularis macrophages, which can impair intestinal motility by releasing NO (4, 13, 49). In the sepsis-induced ileus model, not only muscularis resident macrophages, but also infiltrating monocyte-derived macrophages and neutrophils induced motility disorder through PGs and NO (7, 37). In addition to direct exposure of endotoxin in the case of sepsis-induced ileus, these muscularis macrophages induced muscularis inflammation that resulted in motility disorder in an intestinal manipulation-mediated postoperative ileus model (20, 21, 40, 41, 50) and a chemical-induced colitis model (14, 25, 45). Therefore, muscularis macrophages are thought to play a major role in the development of intestinal motility disorders caused by various types of intestinal inflammation (1, 36).

LPS stimulation upregulates the COX-2 gene expression in macrophages (16). COX-2 is a highly inducible enzyme that catalyzes the production of PGH2 from arachidonic acid. PGH2 is converted into a series of different PGs, dependent on the profile of specific PG synthases (34, 35). PGs have short half-lives and can, therefore, affect only cells that reside locally within the vicinity of their release. Therefore, PGs produced within the intestinal muscle layer are likely to affect only cells within that layer, such as smooth muscle cells, resident macrophages, interstitial cells of Cajal, or myenteric neurons. LPS also induces iNOS gene expression via NF-κB activation through TLR-4/myeloid differentiation factor 88 signaling in macrophages (32, 54). Many reports have indicated the existence of cross talk between PG-mediated signaling and iNOS-induction signaling in different types of cells (3, 15, 48). Previously, our laboratory reported that, in rat intestinal muscularis resident macrophages, LPS-induced iNOS gene expression was almost completely abolished by indomethacin; meanwhile, LPS-induced COX-2 gene expression was not abolished by Nω-nitro-l-arginine methyl ester (l-NAME), suggesting that COX-2 derived PGs are critical for LPS-induced iNOS gene expression (13). However, the details of the COX-2/PGs pathway related to iNOS induction in intestinal macrophages during inflammation remain unclear.

In the present study, we aimed to clarify the types of PGs and PG receptors that contribute to iNOS induction and the subsequent dysmotility in LPS-stimulated mouse ileum. Our findings provide new evidence to indicate that PGE2 produced by LPS stimulus can activate iNOS gene expression through the EP2 and EP4 receptors in intestinal muscularis macrophages.

MATERIALS AND METHODS

Animals.

Male C57BL6/J mice (8 wk old) were purchased from Charles River Japan. The op/op mice were purchased from Jackson Laboratories. EP2- or EP4-receptor knockout (KO) mice were established as previously reported (12, 18, 42). Protocols for the animal experiments and care were approved by the Institutional Review Board of the University of Tokyo (approval code: P07–084). All experiments were performed in strict compliance with the Guide to Animal Use and Care from the University of Tokyo and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Preparation of RAW264.7 cells.

The RAW264.7 murine macrophage cell line (TIB-71, American Type Culture Collection, Manassas, VA) was maintained at 37°C and 5% CO2 in 10-cm dishes with DMEM (Gibco, Grand Island, NY), supplemented with antibiotics and 10% FBS. Cells in passages 5–8 were used in the experiments. Before the start of each experiment, the cells were incubated overnight in culture medium with 1% FBS.

Preparation of peritoneal macrophages.

To collect peritoneal macrophages, we injected 2 ml of 10% protease peptone (BD Life Sciences, Sparks, MD) intraperitoneally into the C57BL/6J mice or the EP2 KO mice. After 48 h, the peritoneal cavity was washed with 5 ml of ice-cold phosphate-buffered saline, and the macrophages were collected. The macrophages were centrifuged, suspended in RPMI-1640 medium (Gibco), seeded onto 6-cm dishes, and allowed to adhere for 2 h. Floating cells were then washed out, and the adherent cells were used in the experiments. Before the start of each experiment, the cells were incubated overnight in culture medium with 1% FBS.

Preparation of ileal tissues.

Ileal tissues were prepared as previously described (13, 14). Briefly, the ileum from each mouse was then dissected into 2- to 3-cm-long segments and cut open along the mesenteric attachment, and the mucosa and submucosa were removed. The remaining muscle layers were incubated with physiological salt solution (PSS) containing (in mM) 136.9 NaCl, 5.4 KCl, 1.0 MgCl2, 23.8 NaHCO3, 1.5 CaCl2, and 5.5 glucose, aerated with 95% O2-5% CO2 to adjust pH to 7.3 at 37°C. LPS exposure was performed in the organ bath for 4 h, followed by pretreatment of drugs for 30 min in each condition.

Cell suspension preparation and FACS analysis for resident macrophages.

Ileal tissues were washed in PBS and cut into small pieces before digestion in 400 U/ml collagenase D for 45 min at 37°C. After digestion, the tissue was disrupted and filtered through a 70-μm cell strainer. The cell suspension was washed two times in cold Hanks' balanced salt solution (Gibco) with 2% FBS (Gibco) and stained in FACS buffer (2% PBS, 2% FBS, 0.01% sodium azide). Before staining, cells were incubated with Fc Block (anti-CD16/32; BD Biosciences) for 15 min at 4°C. Anti-CD11b antibody (anti-Mac-1, MAb M1/70, rat IgG2b-PE; eBioscience, San Diego, CA) was added for 20 min at 4°C. Appropriate isotype controls were included in all experiments. Cells were analyzed by using a FACSAria cell sorter and Diva software (BD Biosciences).

Real-time RT-PCR analysis.

Before the quantitative RT-PCR (qRT-PCR) analysis of mouse iNOS, COX-2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA, RNA isolation was performed as previously described (14). Briefly, total RNA was subsequently extracted from the tissues or the cells with TRIzol (Invitrogen Japan, Tokyo, Japan), then the cells were precipitated with isopropanol, and suspended to a concentration of 1 μg/μl in RNase-free distilled water. RT-PCR was performed for evaluating the expression of DP and EP receptors. Briefly, first-strand cDNA was synthesized with random 9-mer oligonucleotide primers and avian myeloblastosis virus reverse transcriptase XL at 30°C for 10 min, 55°C for 30 min, 99°C for 5 min, and 4°C for 5 min. PCR amplification using the “hot start” method with Taq-Gold (Perkin-Elmer, Branchburg, NJ) was conducted in the presence of the oligonucleotide primers listed in Table 1. The PCR samples were denatured initially at 95°C for 10 min, amplified at 32 cycles at 94°C for 40 s, 55°C for 1 min, and 72°C for 1 min with a thermal cycler (Takara PCR Thermal Cycler MP, Takara Biomedicals, Tokyo, Japan). The PCR products in each cycle were separated electrophoretically on a 2% agarose gel containing 0.1% ethidium bromide. To avert the risk of contaminating the DNA, we performed PCR amplification by using total RNA in the absence of the reverse transcription step as a negative control. A model FAS-III ultraviolet trans-illuminator (Toyobo, Tokyo, Japan) was used for visualizing the fluorescent bands.

Table 1.

PCR primer sequences and optimal product sizes

GAPDH (gene accession no. M 32599)
For semiquantitative RT-PCR
Forward (5′ to 3′)	TGTTCCTACCCCCAATGTGA
Reversal (5′ to 3′)	CCCTGTTGCTGTAGCCGTAT
Size, bp	269
For real-time quantitative RT-PCR
Forward (5′ to 3′)	AAAATGGTGAAGGTCGGTGTG
Reversal (5′ to 3′)	AATGAAGGGGTCGTTGATGG
Size, bp	111
iNOS (gene accession no. NM 010927.1)
For semiquantitative RT-PCR
Forward (5′ to 3′)	GTGGTGACAAGCACATTTGG
Reversal (5′ to 3′)	GGCTGGACTTTTCACTCTGC
Size, bp	487
For real-time quantitative RT-PCR
Forward (5′ to 3′)	TCAGCCAAGCCCTCACCTAC
Reversal (5′ to 3′)	CCAATCTCTGCCTATCCGTCTC
Size, bp	108
COX-2 (gene accession no. NM 009962.2)
For semiquantitative RT-PCR
Forward (5′ to 3′)	CCCCCACAGTCAAAGACACT
Reversal (5′ to 3′)	CCCCAAAGATAGCATCTGGA
Size, bp	770

Real-time qRT-PCR was performed for evaluating time-dependent changes in the expressions of iNOS and COX-2 mRNA, as previously described (46). ABI PRISM 7000 instrument (Applied Biosystems, Forester City, CA) was used for qRT-PCR amplification and detection. qRT-PCR samples were prepared in triplicate, with each sample comprising a 25-μl reaction mixture in a MicroAmp optical 96-well reaction plate sealed with an optical adhesive cover (Applied Biosystems). Each reaction well was treated with 2.5 μl of template DNA, 12.5 μl of platinum SYBR Green qPCR SuperMix-UDG (Invitrogen Japan, Tokyo, Japan), 10 pmol each of forward and reverse primers, and 500 nM carboxy-X-rhodamine reference dye. Plasmid and genomic DNA were serially diluted 10-fold, and this step was performed in triplicate for establishing the standard calibration curves, which are constructed by plotting the threshold cycle (Ct) vs. the log concentration. For any unknown total DNA sample, the absolute quantity of both plasmid and genomic DNA was obtained by interpolating the Ct value from the sample against the standard calibration curves. A negative control was set up by substituting the template with double-distilled H2O. Repeatedly, this resulted in a high Ct value, which was taken to be the nadir, or lowest detectable range.

Immunohistochemistry.

Mice received intraperitoneal injections of Dextran-Texas Red Beads (molecular weight 7000, 2 mg/mouse, Invitrogen, Tokyo, Japan). Twelve hours later, mice were killed, and ileum muscle strips were collected. The muscle strips were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for detecting iNOS or with Zamboni solution for detecting COX-2, and then they were processed for whole mount preparations. Samples were incubated overnight at 4°C with anti-iNOS antibody (1:500; Transduction Laboratories, Lexington, KY) or anti-COX-2 antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, CA). The primary antibodies were detected with Alexa 568-conjugated anti-mouse IgG (1:1,000; Invitrogen, Tokyo, Japan) or FITC-conjugated anti-mouse IgG (1:200; Sigma, St. Louis, MO), respectively. Colocalization was analyzed by using a confocal laser scanning microscope (LSM510; Zeiss, Tokyo, Japan). Dextran-positive cells were considered to be intestinal resident macrophages (38).

Measurement of muscle tension.

Longitudinal muscle tension was measured as previously described (13, 14). Briefly, muscle tension was isometrically recorded with a force displacement transducer. Muscle strips were cut into 3 × 5-mm pieces that were attached to a holder under a resting tension of 10 mN. After equilibration for 15 min in a bath, each strip was repeatedly exposed to high-K+ (72.7 mM) solution until stable responses were observed.

NO measurement.

NO released into the culture medium was measured by monitoring the fluorescent compounds created by 3-diaminonaphthalene exposed to NO2/NO3. After FBS starvation, RAW264.7 macrophages (3 × 105 cells/well) were incubated with 1% FBS DMEM without phenol red, then the cells were stimulated with LPS (1 μg/ml) for 4–24 h. Medium from each cell-culture experiment was immediately frozen in liquid nitrogen and kept at −80°C until assayed. After the medium was thawed on ice, the amounts of NO2/NO3 were measured with the appropriate assay kits, according to the manufacturer's instructions (Dojindo Laboratories, Kumamoto, Japan).

Western blot analysis.

RAW264.7 cells stimulated with LPS (1 μg/ml) for 0–24 h were homogenized in homogenizing buffer to extract protein. The homogenizing buffer contained 50 mM Tris·HCl, 150 mM NaCl, 5 mM EDTA, 1 mM Na3VO4, 50 mM NaF, 5 mM sodium pyrophosphate, and 1% Triton X-100. Pefabloc protease inhibitor cocktail (Roche, Indianapolis, IL) was added to the homogenizing solution. Twenty micrograms of total protein were loaded into each lane to detect iNOS and heat shock protein (HSP)-90. Proteins were separated by electrophoresis and then transferred to a polyvinylidene difluoride membrane. The membrane was blocked by incubation with PBS containing 5% powdered milk for 30 min at room temperature. Membranes were incubated overnight at 4°C in blocking buffer that contained anti-iNOS antibody (1:1,000, Transduction Laboratories) or anti-COX-2 antibody (1:2,000 dilutions, Cayman) or anti-HSP-90 antibody (1:1,000, Santa Cruz) as the primary antibody. Then membranes were incubated for 1 h at room temperature with biotinylated anti-mouse IgG (for iNOS, 1:1,000, Vector Laboratories) or horseradish peroxidase anti-rabbit IgG (for COX2 and HSP-90, 1:1,000, Vector Laboratories) as the secondary antibody. Then, to detect iNOS, membranes were incubated with horseradish peroxidase-streptavidin (1:1,000, Zymed Laboratories) for 1 h at room temperature. Targeted proteins were visualized with an enhanced chemiluminescence plus Western blotting detection system (Amersham Biosciences). Bands were visualized by using an LAS-1000 mini luminescence imager (Fuji Film, Tokyo, Japan).

Statistical analyses.

Results are expressed as means ± SE. The control and test groups were compared by using one-way ANOVA, followed by Dunnett's multiple-comparison test. P values < 0.05 were considered statistically significant.

RESULTS

LPS induces iNOS and COX-2 expression in intestinal resident macrophages and inhibits smooth muscle contraction.

At first, we tried to identify intestinal resident macrophages that phagocytosed Texas-Red-conjugated dextran beads (molecular weight 70,000) in the ileal muscle tissue from C57BL/6J mice (wild-type) and op/op mice. We confirmed the existence of intestinal resident macrophages in the wild-type mice, but not in the op/op mice (Fig. 1).

Fig. 1.

Effect of LPS on cyclooxygenase (COX)-2 and inducible nitric oxide synthase (iNOS) protein expression in small intestine isolated from wild-type (WT) and op/op mice. A: fluorescence micrograph of intestinal macrophages that phagocytosed Texas Red-conjugated dextran beads in the myenteric plexus region of small intestine isolated from WT (C57BL/6J) and op/op mice. B: immunohistochemistry of iNOS and COX-2 in the myenteric plexus region of small intestine isolated from WT mice treated with Texas Red-conjugated dextran injection. The isolated tissue was exposed to LPS (100 μg/ml, 4 h) in a organ bath. Results are typical of 4 experiments.

In the ileal muscle layer from C57BL/6J mice, LPS exposure (100 μg/ml, 4 h) upregulated iNOS and COX-2 proteins in resident macrophages in the myenteric plexus region (Fig. 1). In wild-type mice intestine, LPS upregulated the iNOS and COX-2 mRNA expression levels (0.39 ± 0.21 to 1.13 ± 0.19, 0.93 ± 0.14 to 1.32 ± 0.2, respectively; P < 0.05, n = 4). In contrast, neither iNOS nor COX-2 mRNA expression levels were increased by LPS exposure to the ileum of op/op mice lacking intestinal resident macrophages (0.64 ± 0.24 to 0.76 ± 0.18, 0.81 ± 0.2 to 0.91 ± 0.13, respectively; n = 4) (Fig. 2, A and B, respectively). Thus LPS activates muscularis resident macrophages, resulting in the induction of COX-2 and iNOS. Figure 3, A and B, shows carbachol-induced contraction in the presence or absence of LPS treatment in the ileal tissues isolated from wild-type mice and op/op mice. LPS treatment significantly diminished the carbachol-induced contraction in wild-type mice. A NOS inhibitor, l-NAME (300 μM), recovered the muscle contractility near to the control level. In contrast, LPS did not induce dysmotility in the ileal tissues isolated from op/op mice (Fig. 3, C and D). Furthermore, pretreatments (30 min) with a selective iNOS inhibitor 1400W (1 mM) and a selective COX-2 CAY10404 attenuated LPS-induced dysmotility. In contrast, short-time treatment with CAY10404, after LPS exposure and just before the carbachol application, failed to recover the LPS-induced dysmotility (Fig. 3), indicating that continuous PG signaling is critical for LPS-induced dysmotility.

Fig. 2.

Effect of LPS on COX-2 and iNOS mRNA expression in small intestine isolated from WT and op/op mice. Changes are shown in the expression levels of the iNOS (A) and COX-2 (B) genes 4 h after LPS (100 μg/ml) stimulation of the ileal tissue isolated from WT (C57BL/6J) and op/op mice. Each bar indicates mean ± SE (n = 4 each) of semiquantitative RT-PCR analysis. *P < 0.05 compared with the control.

Fig. 3.

Effects of LPS-derived COX-2 and iNOS products on smooth muscle contraction of small intestine isolated from WT and op/op mice. A–D: representative traces (A and B) and summary (C and D) of carbachol (1 nM to 1 μM)-induced contraction and the effect of a NOS inhibitor, Nω-nitro-l-arginine methyl ester (l-NAME; 300 μM), in ileal circular smooth muscle treated with or without LPS (100 μg/ml, 4 h) in the tissue from C57BL/6J mice (A and C; WT) and op/op mice (B and D). Controls were incubated with physiological salt solution (PSS; refer to materials and methods) for 4 h. E: effects of a nonselective NOS inhibitor, l-NAME (300 μM), a selective iNOS inhibitor 1400W (1 mM), and a selective COX-2 inhibitor CAY10404 (1 μM) on carbachol (1 μM)-induced contraction in ileal circular smooth muscle treated with or without LPS (100 μg/ml, 4 h) in the tissue from C57BL/6J mice. Controls were incubated with PSS for 4 h. pre, 30 min pretreatment before the LPS administration; post, application of CAY at ## min before carbachol stimuli in the LPS-treated muscle strips for 4 h. Each symbol indicates mean ± SE of 3–4 experiments. *P < 0.05 compared with the control. #P < 0.05 compared with the LPS-treated tissue (n = 4 each).

COX-2-derived PGE2 induces iNOS expression via EP2 and EP4 receptors and attenuates ileal motility.

CAY10404 (pretreatment) attenuated the LPS-induced iNOS gene expression in ileal tissue isolated from wild-type mice (48.2 ± 12.9%; n = 4) (Fig. 4), indicating that COX-2-derived PGs participate in iNOS induction by LPS stimulus. To identify the PG species that are involved in the iNOS mRNA expression, we tested the effects of PGD2, PGE2, PGF2α, and thromboxane A2 (1 μM, 4 h) on ileal tissues. We found that PGE2 was the only PG that increased iNOS mRNA expression (0.13 ± 0.06 to 1.08 ± 0.03; n = 4) (Fig. 4; other than PGs, data not shown). We further tested selective agonists of PGE2 receptors for their effects on iNOS induction, and we found that the EP2-selective agonist ONO-AE1–248 (1 μM) and the EP4-selective agonist ONO-AE1–329 (1 μM) induced the expression of iNOS in the ileum tissue (Fig. 4).

Fig. 4.

Effects of COX-2-derived prostaglandin (PG) E2 and EP2/EP4 receptors on LPS-induced iNOS gene expression and LPS-induced dysmotility in isolated ileal tissue. A: change in iNOS mRNA expression 4 h after incubation with PSS alone (control), LPS (100 μg/ml), or PG agonists stimulation in ileal tissue. CAY10404 (1 μM) was applied 30 min before the LPS stimulation. The concentration of each PG agonist (1 μM) was selected to induce submaximum responses of biological reactions. mRNA expressions were detected by real-time RT-PCR. EP1, ONO-DI-004; EP2, ONO-AE1–259; EP3, ONO-AE-248; and EP4, ONO-AE1–329. Each bar indicates mean ± SE (n = 4 each). *P < 0.05 compared with the control. #P < 0.05 compared with the LPS-treated tissue. B: change in iNOS gene induction 4 h after LPS (100 mg/ml) stimulation in ileal tissue isolated from WT, EP2 knockout (KO), and EP4 KO mice. Each bar indicates mean ± SE of 4 independent semiquantitative RT-PCR analyses. *P < 0.05 compared with LPS-untreated tissue. #P < 0.05 compared with LPS-treated tissue isolated from WT mice. C and D: carbachol-induced contraction of ileal tissue isolated from EP2 KO (C) and EP4 KO (D) mice, with or without l-NAME, in the presence or absence of LPS treatment (100 μg/ml, 4 h). Each symbol indicates mean ± SE of 3–4 experiments.

We next investigated the effects of LPS on iNOS mRNA expression and intestinal dysmotility in ileal tissue from EP2 or EP4 KO mice. In intestine from wild-type mice, LPS stimulation significantly increased iNOS mRNA. In contrast, in the ileal tissue from EP2 or EP4 KO mice, the LPS-induced upregulation of iNOS gene expression was greatly reduced (Fig. 4). In ileal tissue isolated from EP2 or EP4 KO mice, there was no effect of LPS on carbachol-induced ileal contractions (Fig. 4, C and D), which contrasted with findings in WT mice (Fig. 3).

Estimation of mechanisms of LPS-mediated iNOS induction through PGE2/EP2 and EP4 signaling in macrophages.

We next investigated possible mechanisms of LPS-mediated iNOS induction via PGE2/EP2 and EP4 signaling by using RAW264.7 macrophages, ileal resident macrophages, and peritoneal macrophages. In RAW264.7 macrophages, LPS upregulated iNOS mRNA and protein expression, which was almost completely inhibited by COX-2 inhibitor CAY10404 (Fig. 5, A and B). COX-2 and iNOS protein expressions and nitrate accumulation in the culture medium occurred in a time-dependent manner (Fig. 5, B and C).

Fig. 5.

Time-dependent changes in iNOS mRNA, protein, and nitric oxide (NO) production in RAW264.7 macrophages stimulated with LPS. LPS (1 μg/ml) was added, and time-dependent change in iNOS mRNA (A), iNOS and COX-2 proteins (B), and NO (C) were measured. A: NOS mRNA induction by LPS stimulus (1 μg/ml) in RAW264.7 macrophages. Each symbol shows mean ± SE of 4 independent real-time RT-PCR analyses. CAY10404 (1 μM) was added 30 min before the LPS stimulus. B: Western blot analysis of iNOS and COX-2 protein in RAW264.7 macrophages stimulated with LPS (1 μg/ml). Typical results (top) and histogram (bottom) of 3 experiments are shown. Heat shock protein-90 (HSP-90) is considered as a housekeeping protein during LPS stimulation. CAY10404 was added 30 min before the LPS stimulation. C: accumulated NO detected as nitrates (NOx) in the culture medium was measured in RAW264.7 macrophages stimulated with LPS (1 μg/ml).

To strengthen our hypothesis that LPS induces NO in ileum macrophages via EP2/EP4 receptors, ileal resident macrophages were isolated by FACS analysis against CD11b, a major cell-surface marker of macrophages (Fig. 6A). In isolated ileal resident macrophages, LPS also upregulated iNOS mRNA expression, which was almost completely inhibited by COX-2 inhibitor CAY10404 (Fig. 6). PGE2 significantly upregulated iNOS mRNA expression. EP2-selective agonist (ONO-AE1–259, 1 μM) and EP4-selective agonist (ONO-AE1–329, 1 μM), but not EP1- and EP3-selective agonists (ONO-DI-004 and ONO-AE-248, 1 μM) enhanced iNOS mRNA expression (Fig. 6). Conversely, in RAW264.7 cells, BW245C [DP (DP1) selective agonist, 1 μM], DK-PGD2 [CRTH2 (DP2) selective agonist, 1 μM], and PGF2α (FP selective agonist, 1 μM) did not increase iNOS gene expression (data not shown). Peritoneal macrophages purified from EP2 KO mice only weakly upregulated iNOS gene expression upon stimulation with LPS compared with wild-type peritoneal macrophages (Fig. 6). These results indicate that LPS induces iNOS in ileal resident macrophages produced via PGE2/EP2 signaling.

Fig. 6.

PG signaling for iNOS mRNA induction in ileal muscularis resident macrophages. A: fluorescence-activated cell sorting (FACS) detection of CD11b-positive resident macrophages in an enzymatically digested ileum cell suspension from C57BL/6J. CD11bHigh population (in the circle) was isolated. SSC, side scatter; PE, phycoerythrin. B: effect of LPS and EP agonists on iNOS mRNA expression in isolated resident macrophages. LPS (1 μg/ml) stimulation for 4 h was used as a positive control. *P < 0.05 compared with resting macrophages. We used the concentration of each EP agonist (1 μM) that induces submaximum responses in biological reactions, as determined from published references. EP1, ONO-DI-004; EP2, ONO-AE1–259; EP3, ONO-AE-248; and EP4, ONO-AE1–329. C: effect of LPS stimulation (1 μg/ml, 4 h) in the peritoneal macrophages isolated from EP2 KO mice. Each bar shows mean ± SE of 4 real-time RT-PCR analyses.

As EP2 and EP4 receptors have been reported to couple with Gs signaling, we next investigated the effects of modulators of cAMP/PKA signaling on iNOS gene expression using RAW264.7 macrophages. A membrane-permeable cAMP derivative, dibutyryl-cAMP, and an adenylate cyclase activator, forskolin, dose-dependently upregulated iNOS gene expression (Fig. 7). In addition, a selective PKA inhibitor, KT5720, inhibited iNOS gene expression in the LPS-stimulated macrophages (Fig. 7).

Fig. 7.

PKA signaling for iNOS mRNA induction in LPS-stimulated macrophages. All experiments were analyzed from semiquantitative RT-PCR. A: effect of dibutyryl (db)-cAMP (1 and 10 μM) and forskolin (1 and 10 μM) on iNOS mRNA expression in RAW264.7 macrophages at 4 h after the stimulation. Macrophages stimulated with LPS (1 μg/ml, 4 h) were used as a positive control. B: effect of protein kinase A inhibitor, KT5720 (0.01–10 μM), on the 1 μg/ml LPS-induced iNOS mRNA expression in RAW264.7 macrophages. KT5720 was applied 30 min before the LPS stimulation. C: effects of various protein kinase inhibitors on EP2 and EP4 selective agonist (1 μM)-induced iNOS mRNA expressions. An ERK inhibitor, PD-98059 (1 μM), and a phosphatidylinositol 3-kinase inhibitor, wortmannin (WM; 100 nM), were added 30 min before the EP2 and EP4 agonist, ONO-AE1–259 and ONO-AE1–329, respectively. Each bar shows the mean ± SE of 4 experiments. *P < 0.05 compared with resting macrophages. #P < 0.05 compared with each agonist stimulation (1 mg/ml, 4 h).

Finally, we examined the effects of KT5720, ERK inhibitor (PD-98059), and phosphatidylinositol 3-kinase (PI3K) inhibitor (wortmannin) on EP2 and EP4 agonist-induced iNOS gene expression (Fig. 7). We found that KT5720 and PD-98059, but not wortmannin, significantly inhibited iNOS gene expression stimulated by EP2 and EP4 agonists.

DISCUSSION

Intestinal inflammation is associated with dysfunction from the mucosal barrier, which can occur in intestinal bowel diseases, allergic diarrhea, and infectious enteritis, or conditions that affect muscularis or outer layers of the bowel wall, such as peritonitis and postoperative ileus. In the latter case, intestinal muscularis macrophages play an important role in inducing intestinal dysmotility during inflammation by producing PGs and NO (7, 20, 40, 41, 50). However, the roles of PGs in iNOS induction and the subsequent intestinal dysmotility are not well understood. In the present study, we demonstrate that PGE2 produced due to LPS stimulation activates the muscularis resident macrophages via EP2 and EP4 receptors to express the iNOS gene, which, in turn, induces intestinal dysmotility via the production of NO. PGE2 induces iNOS gene expression through the cAMP/PKA/MAPK signal transduction pathways in the intestinal muscularis macrophages.

Intestinal muscularis macrophages play an important role in inducing intestinal dysmotility during inflammation by producing PGs and NO (7, 20, 40, 41, 50). Indeed, LPS upregulates the gene expression of COX-2 and iNOS in the intestinal muscle layer. Our laboratory's previous work also showed that LPS stimulation upregulated COX-2 and iNOS protein in ED2-positive resident macrophages in rat (8, 13). In the present study, immunohistochemical analysis revealed that the COX-2 and iNOS genes were expressed in intestinal resident macrophages that are ramified shape, able to phagocytose dextran. In ileal tissue isolated from macrophage colony-stimulating factor-deficient mice (op/op mice), LPS could not induce the expression of the COX-2 and iNOS genes. As muscularis resident macrophages do not develop in the intestinal muscle layer of op/op mice (30), we concluded that LPS can directly activate muscularis resident macrophages to induce COX-2 and iNOS gene expression in mouse small intestine, which is consistent with a previous report (51). A limitation of the present study is that the anti-F4/80 antibody was not available for isolation of ileal muscularis resident macrophages, because the antibody was not efficient without paraformaldehyde fixation. At present, cell biological characters of ileal muscularis macrophages are still unclear, whereas CD11b (47), CD11c (9), and CD163 (52) are also well known as a marker of resident macrophages. Our results indicate that LPS stimulation enhanced iNOS mRNA expression in a COX-2/PGE2-dependent manner in the isolated CD11bHigh cells. These evidences lead us to conclude that isolated CD11bHigh cells are ileal muscular resident macrophages.

Although the activation of TLR-4 by LPS can directly activate NF-κB via myeloid differentiation factor 88/IL-1 receptor-associated kinase-1/tumor necrosis factor receptor-associated factor signaling (19, 28) that induces iNOS gene expressions in various immune cells, iNOS induction by LPS in intestinal muscularis macrophages may be mediated indirectly through the production of PGs, because the LPS-induced iNOS gene expression was almost completely inhibited by a COX-2 selective inhibitor, CAY10404. It has been reported that a COX-2 inhibitor ameliorated intestinal inflammation and associated intestinal dysmotility, resulting in iNOS suppression in an LPS-induced sepsis model (1, 26, 40). However, the type of PG and the receptor that contribute to iNOS gene expression in the muscularis macrophages during LPS stimulus was not previously determined. Our pharmacological analysis indicated that only PGE2 can activate the macrophages via EP2 and EP4 receptors to upregulate the expression of the iNOS mRNA (Fig. 4, A and B). Our laboratory's previous study provided evidence that LPS induces PGE2 in rat ileal muscle tissue (13). Surgical manipulation also markedly increased the PGE2 level in the peritoneal cavity (26). This study examined the effects of LPS on iNOS gene expression in intestines of EP2 or EP4 KO mice, and we found that LPS-activated iNOS gene induction was less than that of wild-type mice. Although we could not study EP2 and EP4 double-KO mice, it is highly possible that LPS indirectly induces iNOS gene expression both via PGE2/EP2 and PGE2/EP4 pathways in the intestinal macrophages.

It has been reported that NO itself is able to enhance PG production via COX nitrosylation and/or COX upregulation (33). However, in our laboratory's previous report (13), NG-monomethyl-l-arginine did not interfere with LPS-induced COX-2 expression or iNOS expression in rat ileal tissue. Taken together, in ileal smooth muscle tissue, these findings suggest NO may not affect PGs production, and that other components of LPS-induced intestinal inflammatory cascade induce PGs.

To clarify the intracellular signaling underlying iNOS induction via PGE2, we further determined the effect of LPS on iNOS induction using RAW264.7 macrophages and purified peritoneal macrophages. In RAW264.7 macrophages, we confirmed that LPS induced iNOS mRNA and protein expression in a time-dependent manner, and that this induction of iNOS was abolished by COX-2 inhibitor. We also measured the levels of nitrates in the culture medium of macrophages stimulated with LPS and found that nitrates accumulated in the medium. In peritoneal macrophages isolated from EP2 KO mice, LPS-induced iNOS gene expression was significantly inhibited in a manner similar to the inhibition in mouse intestinal tissue. These results indicate that RAW264.7 macrophages and intestinal muscularis macrophages use the same EP receptor signaling pathway to induce iNOS mRNA in response to LPS.

The iNOS upregulation due to LPS, EP2 agonist, and EP4 agonist were all attenuated by PKA inhibitor KT-5720. In addition, a membrane-permeable cAMP analog, dibutyryl-cAMP, and an adenylate cyclase activator, forskolin, increased iNOS gene expression in a concentration-dependent manner. These results suggest that cAMP/PKA signaling predominantly contributes to the LPS-induced iNOS expression through the EP2 and EP4 receptors. In fact, EP2 and EP4 receptors bind to Gs protein in a human embryonic kidney expression system (35, 39, 44). Some studies have examined the downstream signaling cascades from PKA activation to iNOS expression. In human oral squamous cell carcinoma SCC-9 cells, PGE2/EP2 signaling enhanced iNOS expression via transactivation with the EGF receptor through Src activation (6). In RAW264.7 macrophages, lipoteichoic acid also upregulated iNOS through PGE2/EP2 signaling following p38 MAPK and ERK activation, but not JNK activation (3, 17). On the other hand, the downstream signaling of EP2 is mediated mostly by PKA, while the downstream signaling of EP4 is mediated equally by PKA and PI3K (10, 11, 39). Thus iNOS expression can be induced via PKA and PI3K. LY-294002, a PI3K inhibitor, has been shown to inhibit LPS-induced NO production in RAW264.7 cells (23, 24). In the present study, we found that ERK inhibitor, but not PI3K inhibitor, partially inhibited EP2- and EP4-induced iNOS gene expression, suggesting that at least the ERK pathway may contribute to the PGE2/EP2 or EP4/cAMP/PKA signaling cascade. Further studies are required to establish which cascade is most effective for PGE2-mediated iNOS induction in the intestine.

The present study clearly indicates that LPS exposure activates intestinal muscularis resident macrophages to produce PGE2, which, in turn, upregulates iNOS expression in autocrine and paracrine manners via EP2/EP4 receptors to induce intestinal dysmotility. In the clinical condition, COX-2-mediated PGE2 and NO produced by iNOS from infiltrated monocyte-derived macrophages and leukocytes may also be important factors in the induction of intestinal dysmotility. In fact, intestinal manipulation-mediated muscularis inflammation is reduced in COX-2-deficient mice compared with wild-type mice, and intestinal dysmotility is ameliorated by COX-2 inhibitor (1, 26). These results indicate that nonsteroidal anti-inflammatory drugs (NSAIDs) could ameliorate intestinal dysfunction. However, NSAIDs cause intestinal damage, such as hemorrhagic lesions, because of the depletion of endogenous physiologically functional PGE2 (2, 53). Therefore, selective COX-2 inhibitors or selective PGE2 receptor antagonists may be effective treatments for intestinal inflammation with motility dysfunction. Kunikata et al. (27) studied the effects of indomethacin in a small intestinal ulceration animal model and found that intestinal hemorrhagic lesions induced by NSAIDs are caused by the inhibition of EP3/EP4 receptors. On the other hand, EP3 receptor appears to play an important role in the growth suppression of colon carcinogenesis (43). These results indicate that EP2 receptor blockade may be a useful therapeutic strategy for the treatment of intestinal dysmotility due to muscularis inflammation.

In summary, we clarified that PGE2 produced by LPS stimulus activates the muscularis resident macrophages via EP2 and EP4 receptors to express the iNOS gene, which, in turn, induces intestinal dysmotility via the production of NO. PGE2 induces iNOS gene expression through the cAMP/PKA/MAPK signal transduction pathways in the intestinal muscularis macrophages.

GRANTS

This work was supported by Grants-in-Aid for Scientific Research from the Japanese Ministry of Education (to M. Hori, nos. 18380173 and 21380178; to H. Ozaki, no. 20228005; and to T. Murata, no. 19688014), the Yakult Bioscience Foundation (H. Ozaki), and the 2008 Strategic Research Base Development Program for Private Universities from the Ministry of Education, Culture, Sports, Science and Technology of Japan (T. Tajima).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

T.T., T. Murata, K.A., M.M., and T. Matsuoka performed experiments; T.T., T. Murata, K.A., Y.U., M.M., T. Matsuoka, and S.N. analyzed data; T.T., T. Murata, K.A., Y.U., M.M., T. Matsuoka, S.N., H.O., and M.H. interpreted results of experiments; T.T. and M.M. prepared figures; T.T. drafted manuscript; T.T. and M.H. edited and revised manuscript; T.T. and M.H. approved final version of manuscript; T. Murata, K.A., Y.U., T. Matsuoka, S.N., H.O., and M.H. conception and design of research.