The Role of the Prostaglandin D₂ Receptor, DP, in Eosinophil Trafficking¹

Prostaglandin (PG)³ D₂ is released by activated mast cells during the allergic response (1) and substantial evidence has accumulated that PGD₂ might be crucially involved in the initiation and perpetuation of allergic inflammation. A significant contribution of PGD₂ to the late phase allergic reaction is suggested by enhanced eosinophilic lung inflammation and cytokine release in transgenic mice over-expressing PGD₂ synthase (2). The biological effects of PGD₂ are principally mediated by two distinct receptors, the D-type prostanoid (DP) receptor and the chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2) (3). Moreover, at higher concentrations PGD₂ is a ligand for the thromboxane receptor (TP), which mediates the bronchoconstricting effect of PGD₂ (4).

CRTH2 is expressed on Th2-type T cells, eosinophils and basophils (5) and mediates their chemotaxis to PGD₂ (6). In addition, CRTH2 is activated by several PGD₂ metabolites, including 13,14-dihydro-15-keto-(DK)-PGD₂, PGJ₂, Δ¹²PGJ₂, and 15-deoxy-PGJ₂ (7, 8) and a thromboxane (TX) metabolite, 11-dehydro-TXB₂ (9). Moreover, CRTH2 mediates the respiratory burst and degranulation of eosinophils (8, 10), induces the production of proinflammatory cytokines in Th2 cells (11), and enhances the release of histamine from basophils (12). In humans, CRTH2 is the most reliable marker for Th2 cells (13), and in animal models, CRTH2 mediates eosinophil infiltration into the lungs and skin and aggravates the pathology of allergic responses (14, 15, 16). Therefore, CRTH2 antagonists are being considered as a potentially useful approach for the treatment of asthma and allergic disease. The cyclooxygenase inhibitor indomethacin was also found to be a potent CRTH2 agonist (17, 18) and has thus provided a useful pharmacophore for small molecule antagonists to CRTH2 (19).

The alternate PGD₂ receptor, DP, is expressed more widely, including platelets, several types of leukocytes, the vasculature, the CNS, retina, nasal mucosa, lungs, and intestine (6, 20, 21, 22, 23). Functionally, DP-mediated responses include inhibition of platelet aggregation, induction of vasorelaxation, mucin secretion, and lowering intraocular pressure (23, 24, 25). DP agonists have been suggested to inhibit neutrophil, basophil, and dendritic cell function (26, 27, 28, 29). Eosinophils have been found to express the mRNA for the DP receptor and to show delayed apoptosis after treatment with supramaximal concentrations (1 μM) of the selective DP agonist BW245c (10). In vivo, a DP antagonist blocks Ag-induced rhinitis, conjunctivitis, and pulmonary inflammation in the guinea pig (30), and DP-deficient mice exhibit reduced pulmonary inflammation in response to allergen (31). These in vivo studies hence point to a proinflammatory role of DP but are difficult to explain based on the known functional responses of DP receptor activation.

We have previously shown that Δ¹²PGJ₂ and PGD₂ markedly prime eosinophils for chemotaxis toward other chemoattractants, such as the CCR3 ligand eotaxin, and that Δ¹²PGJ₂ is capable of mediating the rapid mobilization of eosinophils from the isolated hind limb of guinea pigs (8). These observations suggested novel mechanisms by which PGD₂ or its metabolites generated at sites of allergic inflammation could act as systemic signals for the supply of eosinophils from the bone marrow and enhance their extravasation to the tissue. Hence we set out to investigate the differential roles of CRTH2 and DP in the mobilization of eosinophils from the bone marrow and chemotaxis. Our data demonstrate that the DP receptor plays an important modulator role in the recruitment of eosinophils into the tissue. Thus, DP antagonists might be useful to treat conditions associated with eosinophilic inflammation such as asthma or other allergic diseases.

Adult Dunkin-Harley guinea pigs (either sex, 350–450 g body weight) were obtained from the Research Institute for Laboratory Animal Breeding (Medical University of Vienna, Himberg, Austria).

All laboratory reagents were obtained from Sigma-Aldrich, unless specified. Assay buffer as used in the shape change and chemotaxis experiments was made from Dulbecco’s modified PBS (with 0.9 mM Ca²⁺ and 0.5 mM Mg²⁺; Invitrogen Life Technologies), 0.1% BSA, 10 mM HEPES, and 10 mM glucose (pH 7.4). For the perfusion of the guinea pig hind limb, modified Krebs-Ringer bicarbonate buffer was prepared with 10 mM d-glucose, 3.33 mM CaCl₂, 0.49 mM MgCl₂·6H₂O, 4.56 mM KCl, 120 mM NaCl, 0.7 mM Na₂HPO₄, 1.5 mM NaH₂PO4, and 24 mM NaHCO₃, and supplemented with 0.1% BSA. Recombinant human eotaxin and synthetic guinea pig eotaxin were purchased from PeproTech and Cell Sciences, respectively. PGD₂, DK-PGD₂, BW245c, LTB₄, ramatroban, SQ29458, and BWA868c (for receptor selectivity and references, see Table I) were purchased from Cayman Chemical. Drugs were dissolved in ethanol or DMSO and further diluted in assay buffer to give a final concentration of the solvents <0.1%. [³H]PGD₂ was obtained from PerkinElmer Life and Analytical Sciences.

The sequence of guinea pig DP (32) was amplified by PCR from guinea pig lung cDNA and inserted via 5′ HindIII and 3′ BamHI into pcDNA3.1⁺. The sequence identity of the inserts was verified by restriction endonuclease digestion and sequencing in both directions. HEK293 or 1321N1 cells expressing guinea pig DP were cultivated in DMEM containing 10% FBS, 100 IU/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 400 μg/ml G418.

HEK293 cells expressing guinea pig DP receptors were seeded into 96-well plates at a density of 30,000 cells/well. Competition binding experiments on whole cells were then performed ∼18–24 h later using [³H]PGD₂ (PerkinElmer; 172 Ci/mmol) as a radiotracer in a binding buffer consisting of HBSS and 10 mM HEPES (pH 7.5); 0.6 nM [³H]PGD₂ was used for guinea pig DP equilibrium competition binding. Competing ligands were diluted in DMSO that was kept constant at 1% (v/v) of the final incubation volume. Total and nonspecific binding were determined in the absence and presence of 10 μM PGD₂. Binding reactions were routinely conducted for 3 h at 4°C to exclude receptor internalization and terminated by two washes (100 μl each) with ice-cold binding buffer. Radioactivity was determined by liquid scintillation counting in a Topcount (Packard Instruments) following overnight incubation in Microscint 20.

1321N1 cells expressing guinea pig DP were metabolically labeled with 2 μCi of [³H]adenine (Amersham; TRK311) in 24-well plates for 18–24 h at 37°C. They were then washed twice with PBS and stimulated with increasing concentrations of PGD₂ in HEPES-buffered saline supplemented with 1 mM isobutylmethylxanthine for 30 min at 37°C. The reaction was stopped by adding 5% (w/v) ice-cold trichloroacetic acid supplemented with 0.1 mM cAMP and 0.1 mM ATP. [³H]cAMP was separated from the remaining nucleotides using anion exchange chromatography, and radioactivity was counted after addition of HiSafe3 scintillation fluid (PerkinElmer Life and Analytical Sciences).

This study was approved by the Ethics Committee of the Medical University of Graz. Eosinophil shape change was recorded using human blood sampled from healthy volunteers or guinea pig blood obtained by cardiac puncture after an overdose of pentobarbital sodium (8). Ninety-microliter aliquots of citrated whole blood were stimulated with 10 μl of agonists for 4 min at 37°C. The samples were then transferred to ice and fixed with 250 μl of fixative solution followed by NH₄Cl-induced lysis of RBC (33). Cells were then washed and resuspended in 250 μl of fixative solution. Samples were immediately analyzed on a FACSCalibur flow cytometer (BD Biosciences). Eosinophils were distinguished from other cells by means of their autofluorescence in the FL-1 and FL-2 fluorescence channels (8). Shape change was determined as the increase of forward scatter compared with vehicle stimulation. The forward scatter values are shown as dimensionless arbitrary units.

Preparations of polymorphonuclear leukocytes (PMNL) were prepared by dextran sedimentation and Histopaque gradients, and eosinophils were further purified by negative magnetic selection using an Ab mixture directed against CD2, CD14, CD16, CD19, CD56, glycophorin A from StemCell Technologies. The purity of the eosinophil preparations was usually above 97%, the contaminating cells being mononuclear cells. The viability of the cells was >98%. Thirty microliters of assay buffer or agonists were placed into the bottom wells of a 48-well micro-Boyden chemotaxis chamber (NeuroProbe). Eosinophils were suspended in assay buffer at 2 × 10⁶/ml, and 50 μl of the suspension were placed into the top wells of the plate, which was separated from the bottom wells by a 5-μm pore size polyvinylpyrrolidone-free polycarbonate filter. Baseline migration was determined in wells containing only assay buffer. The chamber was incubated at 37°C in a humidified CO₂ incubator for 1 h and the membrane was carefully removed. Cells that had migrated to the bottom wells were enumerated by flow cytometry (FACSCalibur, BD Biosciences), and contaminating cells were gated out by side scatter and autofluorescence as previously described (8).

The femoral bone cavity was flushed, erythrocytes were lysed by hypotonic shock, and PMNLs were prepared by Histopaque gradients. The cells were resuspended in RPMI 1640 (1% FCS) at 10 × 10⁶/ml, and 100-μl aliquots of the suspension were placed into Transwell inserts with 5-μm pore size polyvinylpyrrolidone-free polycarbonate filters (Corning). The cells were allowed to migrate toward 600 μl of assay buffer or agonists in the bottom wells of 24-well tissue culture plates for 1 h at 37°C in a humidified CO₂ incubator. Baseline migration was determined in wells containing only assay buffer. Cells that had migrated to the bottom wells were enumerated by flow cytometry (FACSCalibur), and eosinophils were distinguished from neutrophils by side scatter and autofluorescence (34). To compare the migration of eosinophils and neutrophils, responses were expressed as the chemotactic index (i.e., relative to baseline migration).

The guinea pig hind limb was perfused as previously described (8, 34). The external iliac artery and vein were exposed and the caudal abdominal artery, superficial iliac circumflex artery, and pudendoepigastric trunk along with their satellite veins were ligated. Polyethylene cannulas (0.8 mm outside diameter) were inserted into the external iliac artery and vein. Modified Krebs-Ringer bicarbonate buffer (gassed with 95% O₂ and 5% CO₂, 37°C) was infused at 4 ml/min via the arterial cannula and removed from the venous cannula using a peristaltic pump. After an equilibration period of 20 min, the perfusate fractions were collected every 10 min and centrifuged at 300 × g for 10 min. The cell pellet was resuspended in Kimura’s stain and nucleated leukocytes and Kimura-positive eosinophils were counted in a Neubauer hemacytometer.

The DP and CRTH2 receptors were visualized on human peripheral blood eosinophils by immunofluorescence following a double antibody staining protocol. For visualization of the CRTH2 receptor, a suspension of unfixed 1 × 10⁷ PMNL in PBS was blocked with human IgG (Sigma-Aldrich) for 30 min at 4°C. Cells were washed once with PBS by centrifugation at 400 × g for 5 min and then incubated with rat mAb against CRTH2 (clone BM16; BD Biosciences) diluted 1/25, or the isotype control rat IgG2a (BD Biosciences) in PBS for 30 min at 4°C. Following a washing step, cells were incubated for 30 min in the dark at 4°C with the secondary Ab (Cy3-conjungated goat anti-rat secondary Ab; Chemicon; dilution 1/250). PMNL were washed twice with PBS, and cells were cyto-spinned on cover slips. Subsequently, PMNL were fixed with 3.7% formaldehyde in PBS at 4°C over night. After washing, the cells were blocked with powdered milk and human IgG (Sigma-Aldrich) in PBS for 60 min at room temperature. Cells were washed twice with PBS and then incubated with rabbit polyclonal Ab against the DP receptor (Sigma-Aldrich) diluted 1/125 or rabbit IgG (Sigma-Aldrich) as negative control in PBS with powdered milk. After incubation with primary Ab, PMNL were washed six times with PBS and then incubated for 30 min in the dark at room temperature with secondary Ab (AlexaFluor 488 goat anti-rabbit IgG; Invitrogen Life Technologies; dilution 1/500). Finally, after six washes with PBS, the nuclei were stained with 4′,6′-diamidino-2-phenylindole (DAPI; Invitrogen Life Technologies). The fluorescent signal was recorded with an Olympus IX70 fluorescence microscope and an Olympus UPlanApo - 60×/1.20 lens, using Olympus DP50-CU digital camera and Olympus CellP̂ software. Images were further processed also with CellP̂ for contrast and brightness adjustments.

The expression of DP and CRTH2 receptors was investigated in human bone marrow. Sections (3 μm) were prepared from EDTA decalcified material, deparaffinized by xylene, rehydrated in graded alcohols, and submitted to microwave epitope retrieval (10 min at 750 W using pH 9.0 Ag retrieval buffer S2367 obtained from Dako). After treatment in 3% H₂O₂ for 10 min, sections were incubated for 30 min using goat polyclonal Ab against eosinophil peroxidase (EPX; Santa Cruz Biotechnology) diluted 1/25, rabbit polyclonal Ab against the DP receptor (Sigma-Aldrich) diluted 1/200 and rat mAb against CRTH2 (clone BM16; BD Biosciences) diluted 1/10 in Dako Ab diluent. After enhancement using MultiLink (Dako E0453) diluted 1/100 and biotinylated StreptABComplex (Dako K5001) or alkaline phosphatase red (Dako K5005) diluted 1/100 for 30 min each, detection was performed using peroxidase substrate chromogen ready-to-use (Dako K5001) or alkaline phosphatase (fast red) under microscopic control. After each development procedure, blocking was performed for 5 min using 3% H₂O₂ with extensive rinsing using Tris-HCl buffer (pH 7.6) in between. Images were taken on a Nikon Eclipse E600 microscope with a Nikon Plan 40×/0.65 lens using Nikon DS-5M-U1 digital camera and Nikon Digital Sight DS-U1 software. Images were further processed with Adobe Photoshop CS, used for additional white balance, contrast, and brightness adjustments.

IC₅₀ and EC₅₀ values were determined by nonlinear regression analysis using the Prism 3.0 software (GraphPad Software). Antagonistic potencies in functional assays are given as pIC₅₀ values that were obtained by competing with increasing concentrations of inhibitor compound for an agonist concentration required to elicit ∼75–80% of the maximal agonist efficacy. IC₅₀ values generated by this procedure are very closely matching the true affinity of antagonists determined by the analysis according to Arunlakshana and Schild (35).

Data are shown as the mean ± SEM except where otherwise stated. Statistical comparisons of groups were performed using two-way ANOVA for repeated measurements or the Mann-Whitney U test. Probability values of p < 0.05 were considered statistically significant.

While BWA868c has been investigated before with respect to its affinity and selectivity toward human DP and CRTH2 receptors (36), guinea pig DP has been characterized so far only in functional assays. Thus we expressed the guinea pig DP receptor in HEK293 and 1321N1 cells for binding experiments and assay of cAMP accumulation, respectively. As expected, BWA868c and the DP-selective agonist BW245c exhibited high affinity for the guinea pig DP receptor, with log K_i values of −7.95 ± 0.08 and −8.95 ± 0.06, respectively, and displaced PGD₂ from the receptor (Fig. 1,A). In assays of cAMP accumulation, BW245c was a full agonist of DP comparable with PGD₂, with pEC₅₀ values of −8.89 ± 0.11 and −8.48 ± 0.09, respectively. In contrast, BWA868c caused a minute degree of cAMP accumulation at higher concentrations (Fig. 1 B), which was compatible with partial agonism already reported for human DP (20).

The selectivity of BWA868c against guinea pig CRTH2 was investigated in eosinophils, which naturally express CRTH2, using the shape change assay. Stimulation with chemoattractants results in immediate responses of granulocytes including reorganization of the cytoskeleton and shape change, which can be detected by flow cytometry (8). PGD₂ was a potent and highly effective inducer of eosinophil shape change of guinea pig eosinophils, although with a 3- to 10-fold lower potency than in human whole blood (Fig. 2). Ramatroban inhibited the PGD₂-induced shape change of human and guinea pig eosinophils (Fig. 2), but had no effect on eotaxin-induced shape change (data not shown, n = 4). In contrast, BWA868c had no effect on shape change responses of guinea pig and human eosinophils, irrespectively of the stimulant used (PGD₂, Fig. 2; eotaxin, data not shown, n = 4). Since the DP agonist BW245c did not cause shape change (data not shown, n = 4), it seems that eosinophil shape change responses to PGD₂ are exclusively mediated by CRTH2.

Using BWA868c and ramatroban as potent and selective antagonists for DP and CRTH2, respectively, we investigated which of the PGD₂ receptors was mediating the mobilization of mature eosinophils from bone marrow in the in situ perfused guinea pig hind limb preparation. PGD₂ caused an immediate increase of eosinophils in the perfusate when added during the period of 20–40 min after the start of the experiments, whereas the vehicle of the prostanoid alone had no effect (Fig. 3,A); 10 nM of PGD₂ caused a similar degree of eosinophil release as 1 nM eotaxin. The release of noneosinophilic cells from the bone marrow was not increased by PGD₂ or eotaxin, whereas LTB₄ nonselectively induced the release of both eosinophils and other nucleated cells (Fig. 3 A).

To determine the role of CRTH2, DP, and TP receptors in eosinophil mobilization from bone marrow, PGD₂ antagonists or their vehicle were infused throughout the experiment, and PGD₂ was added to the perfusate for the period of 20–40 min after the start of the perfusion. The ability of PGD₂ to increase the release of eosinophils from bone marrow was not altered by the TP-selective antagonist SQ29548 (100 nM), but it was inhibited by the CRTH2 antagonist ramatroban (1 μM) by 70–80%, and also by the DP antagonist BWA868c (100 nM) to a similar extent (Fig. 3,B). In contrast, the DP antagonist BWA868c did not inhibit the response to DK-PGD₂ (Fig. 3 B). Therefore, these data show that PGD₂-induced release of eosinophils is mediated both by CRTH2 and DP receptors, while in agreement with its known selectivity for CRTH2, DK-PGD₂ mobilized eosinophils solely via CRTH2.

Unexpectedly, the DP agonists BW245c or ZK110841 did not induce eosinophil release by themselves and the addition of BW245c to the infusion of DK-PGD₂ did not enhance the response of bone marrow eosinophils when compared with DK-PGD₂ alone (Fig. 4,A). Hence, these findings raised the question whether BW245c was inactive on DP receptors in this preparation or might even behave as a DP antagonist. When infused throughout the experiment, BW245c (100 nM) in fact reduced the ability of PGD₂ to release eosinophils from the bone marrow, whereas the response to DK-PGD₂ was not affected (Fig. 4 B). From these observations it may be inferred that BW245c behaves as an antagonist of DP receptors in guinea pig bone marrow.

To further investigate the mechanisms by which CRTH2 and DP activation induce the release of eosinophils from bone marrow, we harvested the leukocytes from the cavity of the guinea pig femoral bone to determine their migration. Bone marrow eosinophils showed a chemotactic response to PGD₂ (3–300 nM) similar to human peripheral blood eosinophils, whereas neutrophil chemotaxis was not stimulated by PGD₂ (Fig. 5,A). Migration of bone marrow eosinophils was significantly inhibited by the CRTH2 antagonist ramatroban and, to a larger extent, also by the DP antagonist BWA868c (Fig. 5,B). Human eosinophils purified from peripheral blood likewise migrated toward PGD₂, and this response was inhibited by ramatroban but not SQ29548 (Fig. 6,A). BWA868c (100 nM) also attenuated the migration of human eosinophils by causing a 3- to 6-fold shift of the PGD₂ concentration-response curve to the right (Fig. 6,A). The inhibitory effect of BWA868c was unrelated to nonspecific attenuation of eosinophil responsiveness, because BWA868c pretreatment did not alter the migration toward eotaxin (Fig. 6,A) or LTB₄ (n = 5, data not shown). The DP-selective agonists BW245c (3–300 nM) or ZK110841 (10–30 nM) did not stimulate the chemotaxis of eosinophils, neither alone (Fig. 6,B) nor in combination with DK-PGD₂ or eotaxin (data not shown). However, pretreatment of eosinophils with BW245c (100 nM) significantly attenuated the migratory response toward PGD₂ (Fig. 6 C). This suggests again that, in contrast to PGD₂, conventional DP agonist do not activate chemotactic pathways in eosinophils, but might even act as antagonist at DP-mediated eosinophil responses.

To further substantiate our findings of DP being involved in eosinophil release from bone marrow and eosinophil chemotaxis, we attempted to visualize the expression of DP and CRTH2 in human peripheral blood eosinophils and eosinophils in bone marrow biopsy specimens. Using immunohistochemistry, we identified DP and CRTH2 in normal bone marrow specimens of three different individuals (data not shown) and one patient suffering from hypereosinophilia of the bone marrow. In these samples, DP was located in mature eosinophils with typical bi-lobed nuclei that also stained positive for eosinophil peroxidase (Fig. 7,A). CRTH2 could be observed in mature eosinophils and also in cells with larger, nonsegmented nuclei consistent with eosinophil precursors (Fig. 7,B). Using double-staining immunofluorescence microscopy, we observed that eosinophils from peripheral blood expressed both DP and CRTH2, whereas neutrophils expressed DP only, although at higher levels than eosinophils (Fig. 8). In contrast, cells incubated with the respective isotype-matched primary control Abs followed by the secondary Abs did not exhibit positive staining.

In the current study, we show for the first time that the DP receptor plays an important role in eosinophil trafficking, i.e., chemotaxis and mobilization of eosinophils from the bone marrow. Hence, our data provide a rationale for the use of DP receptor antagonists or mixed CRTH2/DP antagonists in the treatment of allergic disease. In this study, we observed that PGD₂ induced the release of mature eosinophils from the bone marrow and that this effect was inhibited by either the CRTH2 antagonist ramatroban or the DP antagonist BWA868c, but not the TP antagonist SQ29548. The CRTH2-selective agonist DK-PGD₂ also mobilized eosinophils from the bone marrow, but BWA868c did not inhibit this response, which demonstrates the DP selectivity of this antagonist. The mobilization of eosinophils from the bone marrow was apparently linked with the chemotactic activity of PGD₂ as the prostanoid induced the in vitro migration of eosinophils isolated from the guinea pig femoral cavity. One possible mechanism underlying the inhibitory effect of BWA868c on eosinophil mobilization from bone marrow might be the blockade of DP receptors on endothelial cells, the activation of which by PGD₂ may support the release of bone marrow cells to the blood stream when the eosinophils are stimulated through CRTH2. However, both DP and CRTH2 seem to be involved in the migration of bone marrow eosinophils at the cellular level, because either BWA868c or ramatroban inhibited the in vitro chemotaxis of isolated bone marrow eosinophils. This was unexpected because the DP receptor has not been implicated in the chemotactic response of eosinophils to PGD₂ before. Therefore we further investigated the role of DP in the migration of human eosinophils purified from peripheral blood. As expected, the CRTH2 antagonist markedly attenuated the PGD₂-induced chemotaxis of human eosinophils purified from peripheral blood, whereas the TP-selective antagonist SQ29548 had no effect. However, the DP antagonist BWA868c also significantly reduced the migration toward PGD₂, by inducing a 6-fold rightward shift of the concentration response curve to PGD₂ without altering the maximal response.

Our data suggested for the first time that, in addition to CRTH2, the alternate PGD₂ receptor DP might also be involved in eosinophil chemotaxis and mobilization of eosinophils from the bone marrow. This conclusion was based on the observation that BWA868c inhibited the chemotaxis and mobilization from bone marrow of eosinophils that had been stimulated with PGD₂. BWA868c has previously been characterized as a highly selective antagonist at human DP receptors with virtually no affinity for human CRTH2 (36). The same seems to hold true for the guinea pig, since BWA868c potently inhibited the binding of PGD₂ to recombinant guinea pig DP, but it did not inhibit the PGD₂-induced shape change of guinea pig eosinophils, a response solely mediated by CRTH2. Moreover, BWA868c did not inhibit the chemotaxis of human eosinophils toward eotaxin and LTB₄. Finally, the selectivity of the DP antagonist was also demonstrated by its lack of effect when eosinophil mobilization from bone marrow was induced by the selective CRTH2 agonist, DK-PGD₂.

Whether eosinophils express the corresponding DP receptor protein has not been investigated before. Using immunohistochemistry we found that eosinophils in human bone marrow and in peripheral blood express both CRTH2 and DP receptors, which further emphasizes the importance of DP in eosinophil trafficking. BW245c and ZK110841 have previously been described as selective DP agonists, which can potently mimic those PGD₂ effects that are mediated by the DP receptor, such as inhibition of platelet aggregation, vasodilation and ocular hypotension (24, 37, 38). BW245c has been reported to delay eosinophil apoptosis at supramaximal (1 μM) concentration (10). In this study, we found that BW245c and ZK110841 were not chemotactic for eosinophils, and none of the DP agonists induced the mobilization of eosinophils from in the in situ perfused guinea pig hind limb preparation. On the contrary, BW245c actually inhibited the responses to PGD₂, i.e., the chemotaxis of peripheral blood eosinophils and eosinophil release from bone marrow. However, eosinophil mobilization induced by DK-PGD₂ was not attenuated by BW245c.

These observations suggest that the DP receptor, that is involved in eosinophil locomotion, behaves differently than the DP receptors in the vasculature or platelets: although BWA868c is a DP antagonist in all these cell types (38, 39), BW245c is a DP agonist in platelets and vascular smooth muscle (39, 40), but it does not activate chemotactic pathways in eosinophils and even might act as a DP antagonist in eosinophil responses. Similar disparities had been observed in the conjunctiva, where PGD₂ increased conjunctival microvascular permeability in a BWA868C-sensitive manner, despite the fact that BW245c failed to evoke a plasma exudation response (25). These findings might be explained by unique coupling of the DP receptor in eosinophils: previous studies by others and our own data suggest that activation of CRTH2 in eosinophils leads to shape change through G_q G-proteins, phospholipase C, Ca²⁺ mobilization from intracellular stores, MAP kinases, and actin polymerization (17, 41), while the CRTH2-mediated chemotaxis of eosinophils depends on G_i/o (36). By contrast, DP is coupled to G_s and activation of adenylyl cyclase (3, 41). From our data it appears that the events leading to PGD₂-stimulated shape change are solely mediated by CRTH2 without an involvement of DP. Monneret et al. (41) have shown that both PGD₂ and BW245c stimulate cAMP formation in eosinophils in a BWA868c-sensitive manner. However, an increase of cAMP is likely to attenuate eosinophil chemotaxis (42) and blockade of DP with BWA868c amplifies the PGD₂-induced up-regulation of CD11b (41). Therefore, DP must be linked to additional yet unknown mechanisms that positively control eosinophil locomotion. Our data with chemotaxis of human eosinophils and release of bone marrow eosinophils suggest that this additional mechanism is not activated by BW245c. But since BW245c has the same binding site to the DP receptor as PGD₂ (Fig. 1), it acts as an antagonist with respect to this putative novel mechanism. This is reminiscent of a recent observation made by our group showing that some CRTH2 antagonists can selectively block PGD₂-induced arrestin translocation but not the activation of heterotrimeric G-proteins (43). Gervais et al. (10) have suggested that PGD₂ can delay the onset of apoptosis in cultured eosinophils through activation of DP. However, during the course of the chemotactic response (1 h of incubation at 37°C) we did not observe an anti-apoptotic effect of PGD₂ nor a proapoptotic or toxic effect of the DP antagonist BWA868c as investigated using annexin-V/propidium iodide staining (n = 6, data not shown). This demonstrated that the reduced migratory response toward PGD₂ after pretreatment with the DP antagonist cannot be attributed to reduced cell viability. Thus, further studies are needed to define the mechanisms by which DP activation positively modulates the PGD₂-induced locomotion of eosinophils.

Eosinophil influx to sites of allergic reactions is associated with tissue injury and airway hyperresponsiveness (44). In animal models of allergy, mice genetically deficient in eosinophils show reduced tissue damage and airway hyperresponsiveness (45, 46). Asthmatic patients that had been medicated with respect to eosinophil counts in sputum have significantly fewer severe asthma exacerbations than patients receiving conventional therapy (47). The magnitude of eosinophil infiltration at inflammatory sites is determined both by the supply of eosinophils from the bone marrow and the rate of cell extravasation into the tissue. Our current data suggest that CRTH2 is involved in the rapid mobilization of eosinophils from the bone marrow. In addition, previous in vivo studies have also suggested a proinflammatory role for the second PGD₂ receptor, DP; antagonists of DP were found to block Ag-induced inflammation in the guinea pig (30), and DP-deficient mice responded to allergen challenge with reduced pulmonary inflammation (31). However, the pharmacological mechanisms that underlie these observations have remained unclear. Here we have demonstrated for the first time that eosinophils in the bone marrow and in peripheral blood express the DP receptor and that DP is substantially involved in eosinophil recruitment. This notion applies both to eosinophil release from the bone marrow and chemotaxis. Our current data provide novel insights into the roles of PGD₂ in eosinophil function and hence add further evidence for the potential usefulness of DP antagonists in the treatment of allergic disease.

The authors have no financial conflict of interest.