Lipopolysaccharide-Induced Up-Regulation of Triggering Receptor Expressed on Myeloid Cells-1 Expression on Macrophages Is Regulated by Endogenous Prostaglandin E₂¹

Triggering receptor expressed on myeloid cells-1 (TREM-1)³ is a recently discovered cell surface molecule that has been identified on neutrophils and monocytes (1, 2). The soluble form of TREM-1 (sTREM-1) is detected in bronchoalveolar lavage fluid from patients with microbial infection and has been demonstrated to act as an inhibitor of TREM-1 (3, 4, 5, 6). TREM-1 is a 30-kDa glycoprotein belonging to the Ig superfamily and its expression is up-regulated by various ligands for TLRs (7, 8, 9). Activation of TREM-1 expressed on neutrophils and monocytes by an agonistic mAb has been shown to stimulate the expression of various proinflammatory cytokines, chemokines, and cell surface molecules (1, 7, 8, 9). Furthermore, LPS causes synergistic enhancement of cytokine production by monocytes in response to the agonistic mAb, indicating that TREM-1 amplifies inflammatory responses initiated by TLRs (1, 7, 8, 9). Although the natural ligands for TREM-1 have not been identified, its essential role in acute inflammatory responses has been demonstrated in murine models of septic shock, because blocking of TREM-1 by a sTREM-1 improves the survival of mice with bacterial sepsis (6, 9). Thus, activation of TREM-1 may play a crucial role in the inflammatory response to microbes.

PGs are multipotent mediators that modulate a number of pathophysiological responses. PGs are produced by metabolism of arachidonic acid through activation of cyclooxygenase (COX). COX has two isoforms, which are COX-1 and COX-2 (10). COX-1 is constitutively expressed, whereas COX-2 is expressed at low level by most normal resting cells. COX-2 expression is induced by various TLR ligands (11, 12) and release of PGs is significantly increased in various animal models of endotoxemia or sepsis (13, 14). In particular, PGE₂ has been shown to function as a mediator of sepsis-induced immunosuppression, an inhibitor of proinflammatory cytokine production by macrophages, and an inducer of IL-10 production (15). In contrast, PGE₂ has several detrimental effects in sepsis, including vasodilation and increased vascular permeability (16). Several previous studies have shown that COX inhibitors can improve the survival of mice after burn infection or administration of a lethal dose of LPS (17, 18, 19). These findings indicate that PGs play an important role in microbial inflammation, including sepsis or endotoxemia. However, the precise mechanisms by which PGs (particularly PGE₂) have a regulatory effect on microbial inflammation have not been determined.

Although TREM-1 is clearly induced by LPS, little is known regarding the biological influence of PGE₂ on TREM-1 during microbial inflammation. Therefore, we conducted this study to investigate the biological effects of PGE₂ on the expression and action of TREM-1.

DI-004 (an EP1 agonist), AE1–259-01 (an EP2 agonist), AE-248 (an EP3 agonist), and AE1-329 (an EP4 agonist) were provided by Ono Pharmaceuticals. A monoclonal rat anti-mouse TREM-1 Ab and a monoclonal mouse anti-human TREM-1 Ab, as well as control mouse IgG1 and a polyclonal anti-actin Ab, were obtained from R&D Systems and Santa Cruz Biotechnology, respectively. HRP-conjugated rabbit anti-mouse IgG and HRP-conjugated rabbit anti-rat IgG were purchased from DakoCytomation. Specific ELISAs for human TNF-α and human IL-8 were obtained from BioSource International. 8-Bromoadenosine 3′, 5′ cyclic monophosphate (8-Br-cAMP), LPS, the MEK (MAPKK) inhibitor PD98059, the p38 MAPK inhibitor SB203580, and the PI3K inhibitor LY294002 were purchased from Sigma-Aldrich, while the protein kinase A (PKA) inhibitor H-89 was obtained from Seikagaku. PGD₂, PGE₂, 1S-[1α, 2α(Z),3β(1E,3S),4α]-7-[3-[3-hydroxy-4-(4-iodophenoxy)-1-butenyl]-7-oxabicyclo[2,2,1]hept-2-yl]-5-heptenoic acid (I-BOP), a COX-1 inhibitor (SC560), a COX-2 inhibitor (NS398), and a PGE₂ EIA kit were obtained from Cayman Chemical.

Resident peritoneal macrophages (RPM) were isolated from male ICR mice (6–8 wk old) as reported elsewhere (15). Heparinized peripheral blood was obtained from healthy volunteers and human PBMC were isolated by density-gradient centrifugation with Ficoll-Paque. After washing with PBS, the RPM or PBMC were suspended in RPMI 1640 medium (Sigma-Aldrich) supplemented with 5% heat-inactivated FCS (HyClone), 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen Life Technologies) for culture under a 5% CO₂ atmosphere at 37°C.

RPM (1 × 10⁶ cells) or PBMC (2 × 10⁶ cells) were incubated for the indicated periods with or without various concentrations of PGs (PGD₂, PGE₂, I-BOP), 8-Br-cAMP, LPS, or E-series of prostaglandin (EP EP1–4) agonists. Then the expression of murine TREM-1 (mTREM-1), human TREM-1 (hTREM-1), and soluble hTREM-1 (hsTREM-1) was investigated.

RPM were incubated in the presence or absence of SC560 and/or NS398 for 1 h to inhibit endogenous COX activity, and then the cells were incubated in the presence of LPS for the indicated periods. To block protein kinase activity, RPM were incubated in the presence or absence of various inhibitors such as SB203580, PD98059, LY294002, or H89 for 30 min, after which the cells were incubated with LPS or PGE₂ for the indicated periods.

At the termination of incubation, cells and culture supernatants were obtained by centrifugation. Total RNA and protein were isolated by using RLT lysis buffer (Qiagen). Samples of cell lysate and culture medium were stored at −80°C until use.

Total RNA was extracted from cell lysates using an RNeasy Mini kit (Qiagen). The RNA was treated with DNase I (Qiagen) and cDNA was synthesized from 2 μg of random-primed total RNA in a volume of 20 μl using Omniscript reverse transcriptase (Qiagen). mTREM-1, GAPDH, and COX-2 were assessed by quantitative real-time PCR (SYBR) using specific oligonucleotide primers. hTREM-1, hsTREM-1, and rRNA were assessed by quantitative real-time PCR (TaqMan) using specific oligonucleotide primers and probe. hsTREM-1 was identified as a splice variant of hTREM-1 with a 193-base deletion (exon 3) from bases 471 to 663 (GenBank accession no. AF287008). To avoid amplification of hsTREM-1 mRNA, the forward primer for hTREM-1 was designed to fit exon 3 (the deletion site). To amplify only hsTREM-1, the backward primer was designed to hybridize to the 3′ end of exon 2 as well as the 5′ end of exon 4. These primers could specifically amplify hTREM-1 and hsTREM-1, respectively. The sequences of the primers and probes are listed in Table I. The rRNA primers and probe were purchased from Applied Biosystems. The SYBR PCR was performed in duplicate using a 25-μl reaction mixture containing 1 μl of cDNA, QuantiTect SYBR Green PCR (Qiagen), and 300 nM each of the sense and antisense primers. The PCR mixture was incubated for 15 min at 95°C, and then amplification was performed for 45 cycles, consisting of denaturation at 94°C for 15 s, annealing at 58°C for 30 s, and extension at 72°C for 30 s. TaqMan PCR was performed in duplicate with a 25-μl reaction mixture containing 1 μl of cDNA, 12.5 μl of QuantiTect Probe PCR (Qiagen), 400 nM each of the sense and antisense primers, and 200 nM of the probe. The PCR mixture was incubated for 15 min at 95°C to activate HotStarTaq DNA polymerase. Subsequently, amplification was performed for 45 cycles, consisting of denaturation at 94°C for 15 s and combined annealing extension at 59°C for 1 min. During the extension step, the ABI Prism 7700 Sequence Detection System monitored PCR amplification in real time by quantitative analysis of the emitted fluorescence. The amount of each sample mRNA was evaluated relative to the control sample, which was assigned a value of 1 arbitrary unit.

Culture medium or RPM (1 × 10⁶ cells) was dissolved in sample buffer (350 mM Tris (pH 6.8), 10% SDS, 30% glycerol, 600 mM DTT, and 0.05% bromphenol blue), loaded onto 10% SDS-PAGE gel, and run at 20 mA for 1.5 h. Proteins in the supernatant were transferred to a polyvinylidene difluoride membrane (Roche Diagnostics) for 1.5 h at 200 mA by semidry blotting. The membrane was then blocked with 5% skim milk in PBS containing 0.05% Tween 20 for 1 h at 37°C, washed with PBS containing 0.1% Tween 20, and incubated overnight at 4°C with a monoclonal anti-human TREM-1 Ab (1 μg/ml). The blots were washed four times with TBS and incubated for 30 min with HRP-conjugated rabbit anti-mouse IgG. Immunoreactive bands were developed using a chemiluminescent substrate (ECL plus; Amersham Biosciences).

Flat-bottom plates were precoated with 5 μg/ml of a monoclonal anti-human TREM-1 Ab or an isotype-matched control Ab (mouse IgG1) overnight at 4°C. After washing with PBS, PBMC (1 × 10⁵ cells) were preincubated with or without PGE₂ (1 μM) for 5 h. Then the PBMC were added to the Ab-coated wells, and briefly spun in a centrifuge at 1200 rpm to bind TREM-1. After incubation for 24 h, culture medium was obtained by centrifugation and stored at −20°C until the levels of TNF-α and IL-8 in the supernatant were determined by specific ELISAs.

Concentration of PGE₂ in the culture supernatant was determined by using a PGE₂ EIA kit according to the manufacturer’s instructions.

Results are expressed as the mean ± SD. Statistical analysis was performed using the paired Student t test and p < 0.05 was considered to indicate significance.

PGE₂ is a mediator with a wide variety of biological effects in the process of microbial inflammation. To determine whether PGE₂ could influence the expression and action of TREM-1 in macrophages, RPM were pretreated with PGE₂ at a concentration of 1 μM for 1 h and then the cells were subsequently incubated in the presence or absence of LPS (100 ng/ml) for 1 h. Expression of mTREM-1 was determined by quantitative real-time PCR. LPS significantly increased expression of the TREM-1 gene (Fig. 1), as previously reported. PGE₂ also caused significant induction of TREM-1 expression and the magnitude of gene expression was significantly higher in PGE₂-treated cells than in LPS-treated cells. Furthermore, a combination of PGE₂ and LPS caused additive enhancement of mTREM-1 expression by RPM.

To investigate the time course of PGE₂-induced expression of mTREM-1, RPM were incubated with 1 μM PGE₂ for the indicated periods. Induction of gene expression occurred quite rapidly and was observed as early as 1 h after stimulation, following declined for 12 h (Fig. 2,A). RPM were incubated with varying concentrations of PGE₂ for 1 h to determine whether physiological levels of PGE₂ enhanced mTREM-1 expression. It was shown that PGE₂ increased mTREM-1 expression in a concentration-dependent manner. PGE₂ at a concentration as low as 10⁻¹⁰ M significantly induced mTREM-1 expression and maximal expression occurred after stimulation with 10⁻⁶–10⁻⁷ M PGE₂ (Fig. 2 B).

It has been demonstrated that monocytes and macrophages express various receptors for arachidonic acid metabolites, which are referred to EP, thromboxane-like prostanoid (TP), and CRTH2 (20). Therefore, we investigated the effects of specific ligands for these receptors on mTREM-1 expression by RPM. The cells were incubated for 1 h with PGE₂ (an EP receptor ligand), I-BOP (a TP receptor ligand), or PGD₂ (a CRTH2 ligand), and TREM-1 expression was evaluated by quantitative real-time PCR. PGE₂ induced TREM-1 expression, while neither I-BOP nor PGD₂ up-regulated mTREM-1 expression, indicating that PGE₂ was a specific inducer of mTREM-1 expression among these PGs (Fig. 2 C).

Western blot analysis using a specific anti-mouse TREM-1 mAb was performed to evaluate mTREM-1 protein expression by RPM after incubation with or without PGE₂ for 5 h. mTREM-1 protein was detected faintly when RPM were incubated with the vehicle alone, whereas increased expression of mTREM-1 was clearly seen when RPM were incubated with PGE₂ (10⁻⁶ M) for 5 h (Fig. 2 D).

It has been demonstrated that LPS induces TREM-1 expression as well as the release of PGE₂ by macrophages (7, 8, 9, 21). To evaluate the possible influence of endogenous PGE₂ on LPS-induced mTREM-1 expression, RPM were stimulated with LPS (100 ng/ml) for the indicated periods, after which PGE₂ production and mTREM-1 gene expression were determined by EIA and quantitative real-time PCR, respectively. PGE₂ synthesis gradually increased up to 4 h, and then the maximum level was maintained until 11 h after stimulation (Fig. 3). In contrast, gene expression of mTREM-1 occurred quite rapidly and was seen as early as 0.5 h after stimulation when PGE₂ production was not detected. Maximum induction of mTREM-1 was observed at 2–4 h after stimulation and gene expression returned to the basal level by 8 h. These findings indicated that PGE₂ synthesis did not precede the induction of mTREM-1 gene expression.

mTREM-1 expression was significantly induced by a physiological concentration of PGE₂. Therefore, the regulatory roles of PGE₂ on TREM-1 expression was investigated. Because PGE₂ synthesis is regulated by COX-1 and COX-2, RPM were incubated for 1 h in the presence or absence of SC560 (a selective COX-1 inhibitor) or NS398 (a selective COX-2 inhibitor), and then the cells were stimulated with LPS for 2 h. mTREM-1 expression and PGE₂ synthesis was determined by quantitative real-time PCR and EIA, respectively. Both inhibitors for COX-1 and COX-2 partially, but significantly, inhibited LPS-induced expression of mTREM-1 (Fig. 4,A). When the effects of COX inhibitors on PGE₂ synthesis by LPS-stimulated RPM were investigated, these inhibitors also suppressed PGE₂ synthesis (Fig. 4 B). Vehicle (DMSO) did not affect on LPS-induced PGE₂ synthesis and mTREM-1 expression (data not shown).

To investigate the effect of PGE₂ on LPS-induced mTREM-1 mRNA expression at early time points, RPM were stimulated with LPS in the presence or absence of COX inhibitors for the indicated periods. Both COX-1 and COX-2 inhibitors failed to inhibit mTREM-1 expression at 0.5 h after LPS stimulation, whereas mTREM-1 expression at 1 h was partially inhibited, and that at 2 and 4 h was significantly abolished by COX inhibitors (Fig. 4 C). These findings indicated that the effect of PGE₂ on LPS-induced mTREM-1 expression was predominant at late time points (>2 h after stimulation) but not at early time points (0.5 and 1 h after stimulation).

After RPM were incubated with LPS in the presence or absence of COX inhibitors for various times, the expression of mTREM-1 protein was determined by Western blot analysis. Both inhibitors reduced LPS-induced TREM-1 expression at 4 and 8 h but not at 2 h after stimulation (Fig. 4 D). Actin as the internal control was similarly detected in all samples. These findings indicated that LPS-induced expression of mTREM-1 on RPM was at least partly promoted by endogenous PGE₂.

It has been shown that the biological functions of PGE₂ are mediated by four specific receptors, which are coupled to G-protein and are referred to as EP1 to EP4 (20). To determine which EP receptors mediated PGE₂-induced mTREM-1 expression, RPM were incubated with four synthetic agonists specific for each of the EP receptors (each at a concentration of 1 μM), and then mTREM-1 expression was evaluated by quantitative real-time PCR. The EP4 agonist significantly up-regulated TREM-1 expression in RPM, whereas the EP1, EP2, and EP3 agonists failed to enhance TREM-1 expression (Fig. 5 A).

Activation of the EP4 receptor enhances intracellular accumulation of cAMP via adenylate cyclase. Therefore, we examined the influence of 8-Br-cAMP (a stable cAMP analog) on mTREM-1 expression by RPM. Treatment of RPM with 8-Br-cAMP at a concentration of 5 × 10⁻⁴ M for 1 h significantly enhanced expression of the mTREM-1 gene by RPM (Fig. 5 B). EP4 agonist and 8-Br-cAMP also induced TREM-1 mRNA expression in J774.1 and PBMC (data not shown). These finding clearly suggested that PGE₂-induced TREM-1 expression on RPM was related to EP4 receptor- and cAMP-mediated signaling.

Intracellular cAMP is a major regulator of PKA (22) and cAMP also activates the PI3K-, p38 MAPK-, and ERK-signaling pathways (23, 24, 25). Therefore, we investigated the signaling pathways involved in PGE₂-induced expression of TREM-1 by using synthetic inhibitors of these kinases. A PKA inhibitor (H89), a p38 MAPK inhibitor (SB203580), and a PI3K inhibitor (LY294002) significantly suppressed PGE₂-induced TREM-1 expression, whereas a MAPKK inhibitor (PD98059) failed to influence TREM-1 expression (Fig. 6). Inhibitory effects of these inhibitors were observed in a dose- or time-dependent manner (data not shown). These results suggested that PGE₂-induced TREM-1 expression was mediated via the PKA, PI3K, and p38 MAPK pathways.

An agonistic anti-TREM-1 mAb has been shown to stimulate the production of proinflammatory cytokines by monocytes (1, 7, 9). It was difficult to transfer PGE₂-treated RPM to Ab-coated wells, because the cells tightly adhere to the culture dishes. Therefore, PBMC were used to determine whether TREM-1 could enhance cytokine production by PGE₂-treated monocytes. Cells were incubated in the presence or absence of PGE₂ (10⁻⁶ M) for 5 h, and then harvested for incubation in agonistic anti-TREM-1 mAb-coated wells for 24 h. Then the levels of TNF-α and IL-8 in the culture supernatant were determined by specific ELISAs. The agonistic anti-TREM-1 mAb caused a significant increase of TNF-α production by PGE₂-pretreated PBMC (Fig. 7,A). Production of TNF-α by PGE₂-treated cells was 6-fold higher than that by untreated cells. The agonistic anti-TREM-1 mAb also increased IL-8 production by PGE₂-pretreated PBMC and the magnitude of this enhancement was 4.6-fold (Fig. 7 B). These results indicated that TREM-1 induced by PGE₂ was functional and enhanced the production of proinflammatory cytokines by PBMC.

It has been demonstrated that human monocytes are capable of expressing sTREM-1 as well as the cell surface form (3, 4, 5). Although sTREM-1 has also been identified in mice, the precise structure and function of soluble mTREM-1 are not yet known. Therefore, we investigated the expression of hTREM-1 and hsTREM-1 by PGE₂-treated PBMC to evaluate which type was predominantly expressed. After PBMC were incubated with PGE₂, hTREM-1 and hsTREM-1 gene expression were separately determined by quantitative real-time PCR.

Induction of both hTREM-1 and hsTRM-1 gene expression occurred quite rapidly and was seen as early as 1 h after stimulation, a subsequent declined until 9 h (Fig. 8,A). To determine whether physiological concentrations of PGE₂ could induce the expression of hTREM-1 and hsTREM-1, PBMC were incubated with various concentrations of PGE₂ for 1 h and gene expression was determined. It was found that PGE₂ promoted both hTREM-1 and hsTREM-1 expression in a concentration-dependent manner, with maximal expression occurring at 10⁻⁶ or 10⁻⁷ M (Fig. 8 B).

Western blot analysis using a specific anti-human TREM-1 mAb was performed to detect sTREM-1 protein. PBMC were incubated with or without PGE₂ for 5 h, and then the sTREM-1 protein level in the culture supernatant was determined by Western blotting. sTREM-1 was detected at very low levels when PBMC were incubated with the vehicle alone, whereas sTREM-1 expression was increased when PBMC were incubated with PGE₂ for 5 h (Fig. 8 C). Taken together, these findings showed that PGE₂ up-regulated the expression of hTREM-1 as well as hsTREM-1 by monocytes.

The present study provided evidence that PGE₂ up-regulates mTREM-1 expression by RPM, as well as hTREM-1 and hsTREM-1 expression by human PBMC. LPS-induced TREM-1 expression is at least partly regulated by endogenous PGE₂, because COX inhibitors significantly reduced TREM-1 expression. PGE₂-induced TREM-1 expression was mediated by EP-4, cAMP, and various kinases such as PKA, PI3K, and p38 MAPK. PGE₂-induced TREM-1 was functional, because agonistic anti-TREM-1 mAb promoted a significant increase in the production of TNF-α and IL-8.

It is known that TREM-1 is specifically up-regulated by microbial products such as LPS, lipoteichoic acid (LTA), or zymosan (1, 7, 9, 26). However, the present study provided the first demonstration that PGE₂ could induce TREM-1 expression by both RPM and PBMC. Induction of TREM-1 expression by PGE₂ was also observed in a human monocyte cell line (U937) and a murine macrophage cell line (J774.1) (data not shown), indicating that PGE₂ is an inducer of TREM-1 expression by both monocytes and macrophages. PGE₂ was a specific regulator of mTREM-1 expression, because specific ligands for the CRTH2 and TP receptors (which are expressed on macrophages) failed to induce mTREM-1 expression. Biological effect of endogenous PGE₂ on TREM-1 expression was predominant in the late phase of LPS-induced TREM-1 expression. This is based on the findings that COX inhibitors abrogated mTREM-1 expression after 2 h and also reduced mTREM-1 protein expression after 4–8 h following LPS stimulation.

In the present study, both COX-1 and COX-2 inhibitors suppressed PGE₂ synthesis and TREM-1 induction. It has been documented that LPS promoted PGE₂ production through the induction and activation of the COX-2, but not COX-1 (21, 27). However, Rouzer et al. (21) also reported that SC560 (COX-1 inhibitor) inhibits PG synthesis through inhibition of COX-2 as well as COX-1 in LPS-stimulated RPM. Because this cross-inhibition was not observed in other cells, it appeared to be specific in RPM. TNF-α is also an inducer of PGE₂ synthesis, but a previous study demonstrated that TNF-α had a limited effect on TREM-1 expression (7, 28). The reasons for this difference are not known, but it might be related to different mechanisms of action on monocytes and macrophages.

EP4, one of the receptors for PGE₂, increases intracellular cAMP levels via activation of adenylate cyclase and promotes activation of the PKA, PI3K, p38 MAPK, and MAPKK pathways (22, 23, 24, 25). The present study demonstrated that a specific EP4 agonist and 8-Br-cAMP both enhanced mTREM-1 gene expression, while inhibitors of PKA, p38 MAPK, and PI3K blocked the PGE₂-induced increase of mTREM-1 expression. These findings suggested that PGE₂-induced up-regulation of mTREM-1 expression was mediated by the binding of PGE₂ to EP4, which was followed by accumulation of cAMP and activation of various kinases, including PKA, p38 MAPK, and PI3K. This is consistent with the findings of previous studies demonstrating that PGE₂ potentially activate various kinases such as PKA, PI3K, and p38 MAPK independently (29, 30). COX inhibitors failed to completely suppress LPS-induced up-regulation of mTREM-1 expression by RPM, and these inhibitors abolished TREM-1 expression only in the late phase, but not early phase of LPS stimulation. These indicate that other pathways might also be involved in the induction of TREM-1 expression by LPS. Knapp et al. (28) recently demonstrated that a PI3K-dependent pathway played a central role, while MAPK also played a limited role, in LPS-induced up-regulation of TREM-1 expression by monocytes. Several signaling pathways might be involved in LPS-induced TREM-1 expression, and the endogenous PGE₂-mediated pathway seems to be one of the mechanisms of LPS-induced TREM-1 expression on monocytes and macrophages.

TLR and TREM-1 cooperate to induce an inflammatory response, because activation of TREM-1 causes a marked increment in the production of proinflammatory cytokines by macrophages when LPS is used as the costimulus (1, 8, 9). TREM-1 activates a downstream signaling pathway through DAP12, which involves tyrosine phosphorylation, activation of mitogen-activated protein kinases, and mobilization of Ca²⁺. In contrast, TLRs directly recognize certain microbial products and components, such as LPS, LTA, and bacterial DNA. MyD88, IRAK, TRAF6, and IKK are essentially involved in the TLR-signaling pathway. These kinases can potentially induce the production of proinflammatory cytokines via the activation of NF-κB (31). Natural ligands for TREM-1 remain to be identified. If specific ligands for TREM-1 are located at the foci of microbe-induced inflammation, interactions between TREM-1 and TLRs can synergistically induce inflammatory responses. In this case, cooperation between TLRs and TREM-1 could occur at several levels during the process of LPS-induced inflammation. The present study showed that an LPS-induced increase in the production of PGE₂ promoted TREM-1 expression, and activation of TREM-1 on PGE₂-treated PBMC enhanced the production of proinflammatory cytokines. Based on these findings, we hypothesized that PGE₂-induced up-regulation of TREM-1 expression may play an important role in enhancing the TLR-mediated response of macrophages to LPS stimulation.

Several line of evidence indicated that decoy receptors can modulate inflammatory responses by blocking the action by agonists (32, 33). sTREM-1 is a natural decoy receptor that could potentially inhibit TREM-1-mediated activation of cells through competition with natural ligand(s) for receptor binding. Synthetic sTREM-1 has been shown to inhibit LPS-induced cytokine production by monocytes in vitro (6). Furthermore, a recombinant sTREM-1 fusion protein and synthetic soluble TREM-1 have been shown to protect mice against lethal LPS challenge or bacterial sepsis by suppressing inflammatory cytokine production (6, 9). In contrast, it has been demonstrated that PGE₂ can suppress the production of various cytokines (such as TNF-α, IL-8, MCP-1, IFN-γ-inducible protein-10, and MIP-1β) by LPS-stimulated macrophages through EP2- and/or EP4-mediated pathways (34, 35). PGE₂ also induces the production of IL-10, which can have an anti-inflammatory effect (36). Present study demonstrated that PGE₂ induced the release of sTREM-1 by PBMC. Therefore, PGE₂ might suppress inflammation not only by inhibiting the production of proinflammatory cytokines, but also by inducing expression of the decoy receptor sTREM-1 and increasing the production of IL-10. However, activation of TREM-1 on PGE₂-treated PBMC enhanced the production of proinflammatory cytokines, indicating that PGE₂ may exert bidirectional effects on monocytes and macrophages to modulate inflammation through altering the expression of TREM-1 and sTREM-1.

Blocking of PGs has been shown to increase LPS-induced cytokine production both in vitro and in vivo (36, 37, 38, 39). This is consistent with the previous finding that PGE₂ and EP4 agonists attenuated LPS-induced cytokine production in mice (40). However, a number of studies have provided evidence that COX inhibitors can improve survival after the onset of endotoxic shock and COX-deficient mice are resistant to endotoxin-induced inflammation and death (17, 19, 41). Thus, the precise pathophysiological role of PGE₂ in microbe infection still remains undefined. Further investigations should be directed toward the in vivo effects of PGE₂-induced TREM-1 and sTREM-1 in sepsis models.

Increased expression of TREM-1 has been observed at sites of inflammation caused by microbial pathogens (9). However, we recently demonstrated that monosodium urate monohydrate (MSU) crystals induced mTREM-1 expression in monocytes and macrophages in vitro and in vivo (42), indicating that TREM-1 might be involved in the development of acute gouty arthritis. We also observed that MSU crystal-induced mTREM-1 expression is regulated, at least in part, by endogenous produced PGE₂ (our unpublished data). These findings suggest the possibility that PGE₂ might enhance TREM-1 expression in nonmicrobial inflammatory diseases including acute gouty arthritis. If a natural TREM-1 ligand is also induced in nonmicrobial inflammation, it could enhance inflammatory responses by activating PGE₂-induced TREM-1. Furthermore, nonmicrobial products such as heat shock protein 60, which are induced in various inflammatory diseases, have been shown to stimulate TLRs (43). Thus, it is presumed that activation of PGE₂-induced TREM-1 and TLRs by specific ligands might cooperatively increase the inflammatory responses in patients with nonmicrobial inflammatory diseases.

The present study provided a first evidence that PGE₂ induces the expression of both TREM-1 and sTREM-1 by macrophages. This finding sheds new light on the role of PGE₂ as a regulator of the inflammatory response to microbial infection. Further investigations should be directed toward the assessment of pathophysiological roles of TREM-1 and sTREM-1 in various inflammatory diseases. Such studies may help to elucidate the precise role of PGE₂-induced TREM-1 expression in inflammation and could possibly provide evidence leading to new strategies for the treatment of inflammatory diseases.

We thank Rie Hasegawa and Mitsuko Mizuno for their technical support.

The authors have no financial conflict of interest.