Gestational Age-Dependent Up-Regulation of Prostaglandin D Synthase (PGDS) and Production of PGDS-Derived Antiinflammatory Prostaglandins in Human Placenta

Context: The importance of prostaglandin (PG) signaling pathways to the maintenance of pregnancy and initiation of labor is well recognized. However, the complexity of these pathways and the mechanism(s) of their coordinated regulation in physiological and pathological conditions are only now being appreciated.

Objectives: In this report we provide new evidence of a complete pathway for the biosynthesis and actions of PGD₂ and its metabolites within human gestational tissues.

Materials and Methods: Using immunohistochemistry and Northern and Western blotting, we demonstrate the dynamic regulation of H-type PGD synthase (PGDS) in placenta during gestation; in contrast, L-type PGDS and its PG products were detected in amniotic fluid, with increased amounts associated with labor.

Results: Placental tissues were shown to express both forms of the PGD₂ receptor identified to date, D prostanoid₁ (DP₁) and DP₂/chemotactic receptor on type 2 helper T cells, with a distribution consistent with the villous placenta being a major target, as well as source, of PGD₂. In vitro, placental PGD₂ production was shown to be stimulated upon inflammatory activation, whereas PGD₂ and its J series metabolites exerted potent inhibitory effects on placental cytokine production.

Conclusions: These findings suggest that PGDS-derived prostanoids play important physiological roles in the placenta, such as immunoregulation and feto-placental communication, while potentially having a regulatory role in the processes of parturition.

PROSTAGLANDINS (PGs) ARE pluripotent lipid mediators derived from membrane glycerophospholipid metabolism. As signaling molecules, PGs have been shown to play key roles in almost all normal physiological and pathophysiological processes (1, 2) and to modulate rates of cell proliferation, differentiation, and apoptosis in vitro (3–6). They are synthesized via a multienzyme cascade involving the actions of phospholipases and cyclooxygenases. Finally, terminal PG synthases convert the common prostanoid substrate, PGH₂, into biologically active PGs such as PGD₂, PGE₂, PGF_2α, PGI₂ (prostacyclin), and thromboxane (TXA₂) (7–10).

PGs, particularly of the E and F series, have well-established roles in pregnancy and labor (1, 2). PGE₂, for example, is a potent uterotonic agent that is produced within the amniotic cavity in increased amounts before and during labor and plays a central role in the cervical ripening process. In contrast, much less is known about the actions of PGD₂ despite the fact that it was first described as a major product of human gestational tissues approximately 20 yr ago (11). Studies of nonreproductive tissues have identified roles for PGD₂ in diverse arenas, such as sleep induction (12), control of smooth muscle tone (13, 14), modulation of intraocular pressure (15, 16), and initiation and resolution of inflammatory responses (17–19).

In addition to its actions mediated via the D prostanoid (DP) receptor, PGD₂ has also been identified as a specific high affinity ligand for a novel G protein-coupled receptor known as the chemotactic receptor on type 2 helper T cells (CRTH2; also termed DP₂) (17, 20–22). Secondly, PGD₂ has been shown to be converted to additional biologically active metabolites, including 9α,11β-PGF₂ and the so-called cyclopentenone PGs of the J series (23–25). The PGJ₂ hydration product, 15-deoxy-Δ^12,14PGJ₂ (15d-PGJ₂), has been of particular interest after its identification as a ligand for nuclear peroxisome proliferator-activated receptors (PPARs) and as a potent inflammatory inhibitor and inducer of apoptosis (26–28). Within placental tissues, 15d-PGJ₂ has been shown to induce apoptosis in amnion and trophoblast cell lines (5, 30, 31) and to inhibit PGE₂ and cytokine production (32).

Two distinct PGD₂-synthesizing enzymes have been characterized and cloned (33, 34). They differ considerably in terms of structure, localization, and biological function, and both have been the subject of intense investigation into their properties and functional relevance to a variety of normal and pathological conditions (35, 36). The first, lipocalin-type PGDS (L-PGDS) is identical with the β-trace protein, a major secreted component present in human cerebral spinal fluid (CSF) (34). Recently, measurement of changes in L-PGDS levels in human cervicovaginal fluid has been suggested as a possible indicator of membrane rupture during pregnancy (37). The second enzyme, hemopoietic PGDS (H-PGDS), is a cytosolic protein that has an absolute requirement for glutathione for catalytic activity. Its tissue distribution differs considerably among species, but it is expressed widely in many tissues where it is found in antigen-presenting cells, mast cells, megakaryocytes, and type 2 helper T lymphocytes (35, 38).

Given the high concentration of PGD₂ produced by the human placenta (11), the importance of inflammatory reactions in term and preterm labor (39, 40), and the immunoregulatory actions of PGD₂ and its derivatives (17, 41, 42), we hypothesized that PGD₂ may play important roles in regulating immunological processes critical for the maintenance of pregnancy and in the pathology of preterm labor. Therefore, we tested the hypothesis that within the placental and extraplacental membranes there is a local/paracrine system for the biosynthesis and actions of PGD₂, and that identifiable parts of the system change specifically with gestation and labor.

Subjects and Methods

Collection of tissues

Maternal consent for the collection of placental tissues was obtained according to the guidelines of the local human ethics committees. Tissues were obtained from pregnancies delivered at term before the onset of labor after elective cesarean section [term no labor (TNL); n = 15], after term spontaneous labor (TSL) and uncomplicated vaginal delivery (n = 15), and after preterm delivery (PTD; n = 31). The PTD samples were further subcategorized according to presence of PTL either with (PTL+) or without (PTL−) evidence of intrauterine infection. Leukocyte infiltration of the PTD gestational membranes was evaluated immunohistochemically on paraffin-embedded, full-thickness gestational membranes using an antibody to leukocyte common antigen (CD45; DakoCytomation, Cupertino, CA) as described previously (39, 43). Full details of the collection, processing, and categorization of tissues have been described previously (43–45).

Amniotic fluid (AF) collection

AF samples were collected by transabdominal amniocentesis. All women provided informed consent before the collection of AF. The collection and use of AF were approved by the human investigation committees of participating institutions [Wayne State University, Hutzel Hospital (Detroit, MI), and Sotero del Rio Hospital (Puente Alto, Chile)] and were approved for research purposes by the internal review board of the National Institutes of Child Health and Human Development. Several studies using these fluids have been published recently, which include full details of their collection and clinical categorization (46, 47). Four groups were analyzed: TNL (n = 23), TSL (n = 32), preterm not in labor (n = 32) and PTL (n = 34)

Explant and cell culture

The production of PGs by placental tissues was studied in vitro using previously described explant models of amnion, choriodecidua, and villous placenta (48–50). All placentas used in these studies were obtained after cesarean section at term before the onset of labor with maternal consent. After an initial 24-h equilibration period in Ham’s F-12/DMEM (Irvine Scientific, Santa Ana, CA) containing 10% heat-inactivated fetal calf serum (Invitrogen Life Technologies, Inc., Grand Island, NY) plus antibiotics, explant cultures were treated in serum-free medium containing 0.1% bovine γ-globulin (Sigma-Aldrich Corp., St. Louis, MO) with the indicated test agents. After 24 h, media were removed for assay, and the wet weight of tissue in each well was determined so production rates could be normalized to tissue content. Concentrations of three stimuli, lipopolysaccharide (LPS; 5 μg/ml; Sigma-Aldrich Corp.), IL-1β (1 ng/ml; Immunex Corp., Seattle, WA), and TNF-α (10 ng/ml; John Fraser, University of Auckland, New Zealand), were chosen based on previous studies (48, 49).

The effects of PGD₂ and its metabolites on placental cytokine production were investigated using placental trophoblast cultures prepared using dispase digestion of term villous placental tissue, followed by purification over Percoll as described previously (51). Trophoblast preparations (>80% purity, determined by immunocytochemistry) were cultured in 24-well plates in medium 199 supplemented with 10% fetal calf serum for 72 h to promote syncytialization. Exposure to PGs or ethanol vehicle control was carried out in quadruplicate wells in serum-free medium 199 containing 0.1% bovine γ-globulin. Cytokine production rates were normalized to cellular protein (52). Experiments were performed three times on different placentas, and the results pooled and analyzed collectively.

Immunohistochemistry

Paraformaldehyde-fixed, paraffin-embedded tissue blocks were sectioned at 4 μm, dewaxed, and rehydrated before antigen retrieval [microwaved for 5 min in 0.5 m Tris buffer (pH 10) for PGDS immunolabeling or for 10 min in 10 mm citric acid (pH 6) for DP receptor immunolabeling]. Endogenous peroxidase activity was blocked with 3% H₂O₂ in 50% methanol. The primary antibodies used were rabbit anti-L-PGDS (Cayman Chemical Co., Ann Arbor, MI), rabbit anti-H-PGDS (Cayman Chemical Co.), and rabbit anti-DP1 receptor (Novus Biologicals, Inc., Littleton, CO). The specificity of labeling was determined by preadsorption of antibodies with a 10-fold molar excess of their blocking peptides and by comparison with control rabbit IgG (Sigma-Aldrich Corp.) at the same concentration. Immunoperoxidase staining was developed using Sigma Fast-DAB (Sigma-Aldrich Corp.).

Western blot analysis

Total cellular protein was extracted from tissue samples (0.5–1 g) as previously described (44), and protein concentrations were determined by the Lowry method calibrated against BSA (DC Protein Assay Kit, Bio-Rad Laboratories, Auckland, New Zealand). Fifteen micrograms of total protein lysates from each sample were separated on 4–12% gradient NuPAGE bis-Tris precast gels (Invitrogen, Grand Island, NY) and transferred to a Hybond ECL nitrocellulose membrane (Amersham Biosciences, Aylesbury, UK). Immunodetection was achieved by incubation of blots with rabbit polyclonal antibodies to either H-PGDS or L-PGDS and visualized by exposure to x-ray film (Amersham Biosciences). The specificity of labeling was determined by preadsorption of antibodies with a 10-fold molar excess of their blocking peptides. The sequences of the H-PGDS- and L-PGDS-blocking peptides have been reported previously (33, 53). Positive controls for H-PGDS or L-PGDS were human placenta and rat CSF, respectively. After scanning the image, the intensity of the protein signal was determined using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

Quantitative real-time PCR

Total RNA was isolated by the guanidinium thiocyanate-acid phenol method (54) as previously described (43). Extracted RNA was treated with ribonuclease-free deoxyribonuclease I (Invitrogen New Zealand Ltd., Auckland, New Zealand). Synthesis of cDNA was carried out using the SuperScript First Strand Synthesis Kit for RT-PCR (Invitrogen New Zealand Ltd.) with 1.0 μg deoxyribonuclease-treated RNA as template and 0.5 μg oligo(deoxythymidine)_12–18 as primer. TaqMan probes and primers for L-PGDS, H-PGDS, DP₁, DP₂/CRTH2, and 18S rRNA from Assay-on-Demand gene expression kits (Applied Biosystems, Foster City, CA) were used. The relative expressions of H-PGDS, L-PGDS, DP₁, and DP₂/CRTH2 mRNA were normalized to the amount of 18S rRNA in the same cDNA sample according to the manufacturer’s instructions (Applied Biosystems User Bulletin No. 2 and Chemistry Guide P/N 4330019).

RIA

PGE₂ was measured by RIA using previously described methods (49). PGD₂ was measured by RIA using tritiated PGD₂ tracer (Amersham Biosciences) and anti-PGD₂ antiserum (11) according to standard protocols (11, 49, 55).

15d-PGJ₂ enzyme immunoassay

15d-PGJ₂ concentrations were measured using a competitive enzyme immunoassay kit (Correlate-EIA, Assay Designs, Inc., Ann Arbor, MI) according to the manufacturer’s instructions. The limit of detection was 37 pg/ml, and the intraassay precision (coefficient of variation) was less than 7.5% according to the manufacturer’s specifications.

Cytokine immunoassays

IL-6 and IL-8 were assayed by two-site ELISA using commercially available capture and detection antisera (R&D Systems, Inc., Minneapolis, MN) as previously described (49).

Statistical analysis and presentation of data

Production rates of cytokines and PGs in explant culture were expressed as a percentage of the control value. Statistical significance was assessed by ANOVA, followed by Dunnett’s test or Bonferroni t test post hoc. Differences in AF concentrations of PGs and in mRNA expression, as determined by real-time PCR, were assessed by Mann-Whitney or Kruskal-Wallis test for nonparametric data. P < 0.05 was considered statistically significant.

Results

The two isoforms of PGDS (L-PGDS and H-PGDS) immunolocalized to distinct regions of human term and preterm gestational tissues (Figs. 1 and 2²). In placental villous tissues, specific labeling for L-PGDS was identified in the syncytiotrophoblast layer of both preterm and term placenta, with prominent staining of the apical membrane (Fig. 1, A and B). At term, cytoplasmic syncytial staining was readily apparent in the syncytium, giving rise to a characteristic beaded appearance (Fig. 1B). Similarly, strong H-PGDS immunolabeling was observed in the syncytial layer of the preterm placenta (Fig. 1C). However, cellular localization of H-PGDS appeared to change with gestational age, and by term, immunoreactive H-PGDS was mainly localized to cells lining the villous capillaries, with little or no labeling observed in the syncytium (Fig. 1D). Labeling was completely absent in the corresponding negative controls in which primary antibody was preincubated with a 10-fold excess of blocking peptide or was omitted completely (data not shown).

In gestational membranes collected from both preterm and term pregnancies, strong immunolabeling for L-PGDS was observed in the cells of all tissues, including amnion epithelial, reticular, chorionic trophoblast, and decidual cells (Fig. 2, A–C). In contrast, only very weak labeling of the gestational membranes was observed for H-PGDS (Fig. 2, D–F). No labeling was observed in the corresponding negative controls (data not shown). No significant staining of infiltrating leukocytes was apparent, although on some slides occasional cells positive for L- or H-PGDS were visualized among maternal blood cells (not shown).

The effects of labor and intrauterine infection on the expression of H-PGDS in amnion, choriodecidual, and villous placental samples are presented in Fig. 3A. Considerable variability existed in the level of H-PGDS protein expression in the amnion and choriodecidual samples, whereas H-PGDS protein expression in the villous placenta was much more consistent. Densitometric analysis of H-PGDS in gestational tissues revealed that the level of expression of H-PGDS protein was significantly higher in the villous placenta from term vs. preterm deliveries, whereas the level of expression in amnion and choriodecidua was unchanged between preterm and term samples. There was no statistical difference in the net amount of H-PGDS in any tissue as a result of the onset of labor at term or the presence of intrauterine infection preterm (Fig. 3B). Immunoreactive L-PGDS was undetectable by this method in gestational tissues (data not shown).

Pooled amniotic fluid samples were obtained at term or preterm (with and without labor) and subjected to Western blot analysis (Fig. 4). Both anti-L-PGDS and anti-H-PGDS antibodies were used. A specific band of the expected size for L-PGDS (∼32 kDa) was detected in rat CSF (positive control; not shown) and in all pooled amniotic fluid samples regardless of gestational age and state of labor (Fig. 4). Labeling of a 24-kDa band, specific for H-PGDS, was observed in villous placenta (positive control), but was not detected in any of the pooled amniotic fluid samples (Fig. 4). In each case, the specificity of antibody labeling was confirmed by the elimination of signal with preincubation of the primary antibody with a 10-fold excess of its blocking peptide (data not shown).

Changes in the expression of mRNA for both enzymes are presented in Fig. 5. L-PGDS mRNA was detectable in all three tissues and in all patients sampled. There was no significant effect of preterm intrauterine infection or term labor on the level of expression of L-PGDS mRNA. Therefore, data were combined to determine the relative level of expression for each tissue (Fig. 5A). L-PGDS mRNA levels were significantly higher in the choriodecidua and placenta than in the amnion (P < 0.001). There was no significant difference in the level of expression between the choriodecidua and placenta. The expression of H-PGDS mRNA was detectable in all three tissues, with no significant effect of preterm intrauterine infection or labor at term (Fig. 5B). Analyzing the combined data for each tissue revealed that the relative expression of H-PGDS mRNA was lowest in the amnion and choriodecidua, with a significantly higher level of expression in the villous placenta (P < 0.001; Fig. 5B).

Concentrations of PGD₂ and its active metabolite, 15d-PGJ₂, in amniotic fluid were determined by specific RIA and commercial ELISA, respectively (Fig. 6). PGD₂ was present in amniotic fluid samples at concentrations ranging from 0–1200 pg/ml. Amniotic fluid PGD₂ concentrations increased significantly with labor at term (P < 0.05, by two-way ANOVA/least squares analysis) and was also independently associated with both gestational age and labor (general linear model; Fig. 6). The active PGD₂ metabolite, 15d-PGJ₂, was measured in pooled amniotic fluid samples at concentrations ranging from 80–1100 pg/ml (∼0.3–3 nm); there was a notable trend toward higher concentrations at term than in preterm samples (Fig. 6), similar to the PGD₂ pattern, but statistical analysis could not be performed because of the pooled nature of the samples.

Functional regulation of PGDS was investigated by determining PGD₂ production in vitro by explant cultures of gestational tissues, both basally (Table 1) and in response to stimulation with inflammatory mediators (Fig. 7). All tissue explants (amnion, choriodecidua, and villous placenta) produced PGD₂ basally, with production rates ranging from 2.5 pg/mg·24 h for amnion to 95.8 pg/mg·24 h for placental explants (Table 1). Basal PGD₂ production was significantly greater in placenta than in amnion or choriodecidual explants (P < 0.05). Basal production of PGE₂ was greater than PGD₂ in all tissues, ranging from 10.8 pg/mg·24 h in choriodecidua to 700.4 pg/mg·24 h in placenta (Table 1). The proinflammatory mediator, IL-1β, at concentrations of 0.4 ng/ml or more induced a concentration-dependent increase in PGE₂ production in both amnion and choriodecidual explants (Fig. 7A). Similarly, IL-1β induced an increase in PGD₂ production from both tissues, with significant effects observed at 2 and 10 ng/ml (Fig. 7A). In contrast, TNF-α exerted differential effects on the production of the two PGs. Although PGE₂ production was significantly stimulated in the amnion by concentrations of 4 ng/ml or more and in the choriodecidua by concentrations of 20 ng/ml or more, the effect of TNF-α on PGD₂ production was more complex (Fig. 7A). In the amnion, TNF-α, at all concentrations tested, failed to increase PGD₂ production significantly. However, in choriodecidual explants, TNF-α induced a significant stimulation of PGD₂ production at concentrations of 20 ng/ml or more (Fig. 7A). In explant cultures of villous placenta, LPS (at 5 μg/ml) induced a significant (4- to 5-fold) increase in the production of both PGE₂ and PGD₂ (Fig. 7B).

The effects of PGD₂ and PGJ₂ series PGs on the production of IL-6 and IL-8 by villous trophoblast cultures are presented in Fig. 8. The production of IL-6 was significantly decreased by approximately 50–90% of control values by all PGs tested. This decrease was statistically significant (P < 0.05) for PGD₂, PGJ₂, and 15d-PGJ₂. In contrast, IL-8 production was significantly decreased only by 12d-PGJ₂ and 15d-PGJ₂; PGD₂ had no significant effect on IL-8 production, whereas PGJ₂ caused a modest, but nonsignificant, decrease in IL-8 production (Fig. 8).

To determine whether gestational tissues are targets for PGD₂-related actions, we studied placental DP receptor expression. DP₁ receptor mRNA was detectable in amnion, choriodecidua, and placenta, although the levels in amnion were lower than those in the other two tissue types (below the limit of detection in eight of 24 individual amnion samples; data not shown). In choriodecidual and placental samples, there were no significant differences in DP₁ mRNA expression between patient groups; therefore, the data were combined to give the overall relative expression of DP₁ mRNA per tissue (Fig. 9A). The relative expression of DP₁ mRNA was not significantly different between the choriodecidual and placental samples, but there was less individual variability in the placental expression levels. In contrast, the expression of DP₂/CRTH2 mRNA showed a much more restricted tissue distribution than that observed for DP₁ mRNA. None of the 24 amnion samples (preterm and term) analyzed had detectable DP₂/CRTH2 mRNA levels. Furthermore, in choriodecidua from preterm deliveries, only three of 12 samples had detectable DP₂/CRTH2 mRNA levels; interestingly, each of these three positive samples was collected from individuals with histological chorioamnionitis (data not shown). At term, seven of 12 choriodecidual samples analyzed had detectable DP₂/CRTH2 mRNA levels (data not shown). In contrast, all preterm and term placental villous samples had detectable levels of DP₂/CRTH2 mRNA (Fig. 9B). However, relative expression was extremely variable and was not significantly different between preterm and term placenta.

DP₁ receptor cellular localization was studied by immunohistochemistry. DP₁ receptor protein was immunolocalized to cells within the chorion and decidual layers of the gestational membranes, with either no labeling or very weak labeling in the amnion epithelium (Fig. 10A). Strong immunolabeling for DP₁ receptor protein was identified in the syncytiotrophoblast layer of the villous placenta. No specific labeling was observed to the fetal vessels in these placental samples (Fig. 10B). No specific labeling was observed in control sections, where the primary antibody was omitted or replaced with an equivalent concentration of rabbit IgG (data not shown).

Discussion

The results presented here provide evidence of a complete pathway for PGD₂ synthesis and actions within human gestational tissues. We have demonstrated for the first time the expression of PGD synthases and DP receptors to placental tissues and localized their cellular sites of expression. We have also documented dynamic regulation of components of this system with gestational age. In vitro, we have described the regulation of PGD₂ production by human gestational tissues and, furthermore, demonstrate a potent antiinflammatory role of this PG and its metabolites in the placenta.

The importance of PG signaling pathways to the maintenance of pregnancy and the initiation of labor has been recognized for many years (4, 56, 57). Until recently, the focus has been on the regulation of cyclooxygenase-1 and -2 as the presumed rate-limiting step in the PG biosynthetic pathway (14). It is now apparent, however, that there are multiple points of regulatory control within these pathways. In particular, specific synthases provide the mechanism for fine tuning of PG signaling in a temporal and tissue- and cell-specific manner.

The two known PGD synthases, L- and H-PGDS, are present in human gestational tissues, as demonstrated by both immunohistochemistry/immunoblotting and real-time PCR. The two enzymes, however, have nonidentical tissue distribution patterns; L-PGDS appears to be fairly ubiquitous and is present in cells of the amnion, chorion, decidua, and villous placental syncytial trophoblasts. In contrast, H-PGDS expression is restricted mainly to the villous placenta and, more interestingly, has a dynamic localization, changing from syncytiotrophoblastic in preterm tissues to peri- and intravascular labeling at term. The reason for this dynamic change in the localization and abundance of H-PGDS enzyme is presently unknown, but we speculate that it may reflect a requirement for increased local increase in PGD₂ synthesis in close proximity to the fetal vessels. The possibility thus exists that placental PGD₂ is released into the fetal circulation near term, where it may act on fetal tissues. For instance, in the fetus, PGD₂ is a potent pulmonary vasodilator and can cause marked increases in cardiac output (58)

Real-time PCR analysis confirmed the placenta as the major site of H-PGDS mRNA expression. However, there is no evidence that mRNA levels for either enzyme change according to stage of pregnancy, onset of labor, or presence of intrauterine infection, in contrast to the protein studies. This either reflects a sensitivity issue with this particular assay or points to possible posttranscriptional regulation of protein synthesis or protein stability for H-PGDS. The endogenous signal(s) responsible for the dynamic change in H-PGDS localization and the level of expression in the term placenta remains to be determined.

We found that basal levels of PGD₂ production were greatest in the placenta, corresponding to the high levels of both L-PGDS and H-PGDS expression in this tissue, suggesting that one or both of these enzymes are constitutively active in the term placenta. Human gestational membranes produce a number of proinflammatory cytokines (TNF-α, IL-1β, IL-6, and IL-8) and PGs, the production of which is enhanced in response to inflammatory stimuli and bacterial cell wall products such as LPS (59–61). However, their effects on PGD₂ production have not been examined previously.

Of the two PGD synthases, only L-PGDS was detected in amniotic fluid. L-PGDS is a secreted enzyme, so this observation is not surprising. The source of this secreted protein is likely to be amnion epithelial cells, although contributions from the placenta and fetal tissues cannot be discounted. Our findings support those of Shiki et al. (37), who reported recently that the concentrations of L-PGDS in amniotic fluid were higher during the second half of pregnancy, suggesting a role for PGD₂ in the amniotic cavity at term. Together with our present results, this would suggest that both forms of PGDS are up-regulated during the later stages of pregnancy, but that this occurs in different intrauterine compartments: amniotic fluid for L-PGDS and placenta for H-PGDS. This is consistent with our finding of increases in PGD₂ concentrations in amniotic fluid with gestational age and an additional increase with the onset of labor. Whether this is the result of an increase in placental H-PGDS, amniotic fluid L-PGDS, or a combination of both remains to be determined.

We have also demonstrated, for the first time, the presence in amniotic fluid of the active PGD₂ metabolite, 15d-PGJ₂. Levels of this PG were lower than those of PGD₂, as would be expected, and concentrations increased after labor at term in parallel to PGD₂. The presence of 15d-PGJ₂ is likely to be the result of nonenzymatic dehydration of PGD₂, which is known to occur in biological fluids (25), although this requires experimental confirmation. There is debate as to the relevance of 15-d-PGJ₂ in biological systems based on the requirement for micomolar concentrations to exert its effects, in contrast to its low nanomolar concentration determined in biological fluids. The present findings of antiinflammatory effects at 10-μm concentrations suffer from the same interpretive difficulties. However, a recent study from our laboratory demonstrated that 15d-PGJ₂ could exert effects at nanomolar concentrations (62). It should also be noted that PPARs are nuclear receptors, so the concentrations of 15d-PGJ₂ measured in biological fluids may grossly underestimate the intracellular concentrations.

Previous studies have reported actions of PGD₂ on myometrial contractility (63–67), but little is known about the direct effects of this PG and its metabolites on the placenta (68). The present findings suggest an antiinflammatory action of PGD₂ and its J series metabolites on placental production of inflammatory cytokines. PGD₂ was a potent inhibitor of placental IL-6 production, as were PGJ₂ and 15d-PGJ₂. In contrast, PGD₂ did not inhibit the production of IL-8, although a significant inhibition was observed with both PGJ₂ and 15d-PGJ₂. Placental IL-8 production has been reported to be more constitutive than that of IL-6 (69).

We have previously shown that 15d-PGJ₂ can activate nuclear PPARs (61, 70, 71), and that the placenta, amnion, and choriodecidua all express mRNA for the three α-, β-, and γ-PPAR subtypes (70, 71). In the present study we demonstrate for the first time the expression of both currently known G protein-coupled receptors for PGD₂, namely, DP₁ and DP₂/CRTH2, in gestational tissues. In contrast to DP₁, the expression of mRNA for DP₂/CRTH2 appears to be largely restricted to the placenta, with only three choriodecidual samples from patients with intrauterine infection having detectable mRNA. It is probable that the positive expression of DP₂/CRTH2 in the placenta is due to the presence of resident and circulating immune cells (72–74), whereas isolated choriodecidual samples express DP₂/CRTH2 as a reflection of white blood cell infiltration associated with intrauterine infection.

To conclude, we have described a complete pathway for PGD₂ biosynthesis and signaling in human gestational tissues. The cellular localization of H-PGDS is dynamically regulated in the placenta at term, whereas the expression of H-PGDS and the production of PGD₂ increase with gestation and with labor. Synthesis of PGD₂ is up-regulated by inflammatory mediators, whereas PGD₂ and its J series metabolites have direct antiinflammatory actions in the human term placenta. It is highly likely that PGD₂ has additional potent regulatory actions on other aspects of placental and pregnancy physiology, including specific receptor-mediated effects on gestational membranes and the fetus itself.

Acknowledgments

We thank the surgeons and staff at the National Women’s Hospital (Auckland, New Zealand) for their assistance with the collection of tissues.

This work was supported by the Auckland Medical Research Foundation and the Health Research Council of New Zealand.

Current address of R.J.A.H.: Department of Anatomy with Radiology Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand.

Abbreviations:

 
• AF,

  Amniotic fluid;

 
• CRTH2,

  chemotactic receptor on type 2 helper T cells;

 
• CSF,

  cerebral spinal fluid;

 
• DP,

  D prostanoid;

 
• 15d-PGJ₂,

  15-deoxy-Δ^12,14-prostaglandin J₂;

 
• H-PGDS,

  hemopoietic prostaglandin D synthase;

 
• L-PGDS,

  lipocalin-type prostaglandin D synthase;

 
• LPS,

  lipopolysaccharide;

 
• PG,

  prostaglandin;

 
• PGDS,

  prostaglandin D synthase;

 
• PPAR,

  peroxisome proliferator-activated receptor;

 
• PTD,

  preterm delivery;

 
• TNL,

  term no labor;

 
• TSL,

  term spontaneous labor;

 
• TXA₂,

  thromboxane.

References