Prostaglandin (PG)D2, which has long been implicated in allergic diseases, is currently considered to elicit its biological actions through the DP receptor (DP). Involvement of DP in the formation of allergic asthma was recently demonstrated with DP-deficient mice. However, proinflammatory functions of PGD2 cannot be explained by DP alone. We show here that a seven-transmembrane receptor, CRTH2, which is preferentially expressed in T helper type 2 (Th2) cells, eosinophils, and basophils in humans, serves as the novel receptor for PGD2. In response to PGD2, CRTH2 induces intracellular Ca2+ mobilization and chemotaxis in Th2 cells in a Gαi-dependent manner. In addition, CRTH2, but not DP, mediates PGD2-dependent cell migration of blood eosinophils and basophils. Thus, PGD2 is likely involved in multiple aspects of allergic inflammation through its dual receptor systems, DP and CRTH2.
Leukocytes migrate from the peripheral blood to sites of inflammation through interaction with various cellular and humoral components such as selectins, integrins, and locally produced chemoattractants 1,2. Allergic inflammation is characterized by selective accumulation of Th2 cells, eosinophils, and basophils, and is often triggered by antigen-stimulated mast cells 3,4, although the mechanisms supporting these phenomena are largely unclear.
Recently, we cloned a novel putative chemoattractant receptor named CRTH2, which is a seven-transmembrane G protein–coupled receptor structurally related to members of the N-formyl peptide receptor (FPR) subfamily 5. CRTH2 is intriguing in that it is selectively expressed in Th2 cells, T cytotoxic type 2 cells, eosinophils, and basophils 5,6,7. Furthermore, CRTH2 can mediate intracellular Ca2+ mobilization in response to a factor(s) released from activated mast cells, suggesting that CRTH2 may be closely involved in mast cell–mediated allergic inflammation 6.
In this study, we show that the mast cell–derived CRTH2 agonist is prostaglandin (PG)D2. This result was somewhat surprising because PGD2 is generally considered to exert its action via the DP receptor (DP; references 8,9,10) and because CRTH2 has no significant homology in amino acid sequence with DP or all other known prostanoid receptors 10. However, in view of the structure, it may be notable that, like CRTH2, some other lipid mediator receptors such as lipoxin (LX)A4 receptor FPRL1 and the new leukotriene (LT)B4 receptor BLT2 are also more closely akin to FPR than the prostanoid receptors 11,12. PGD2 is the major prostanoid that is produced by activated mast cells and thereby has long been implicated in allergic diseases 10,13. Most recently, involvement of the DP-mediated activities of PGD2 in the formation of allergic asthma has been demonstrated with DP-deficient mice 14. However, contribution of DP to the direct effects of PGD2 on leukocytes, such as induction of eosinophil migration 15, has not yet been established, raising new questions such as how these receptors are being used.
Our results show that although CRTH2 functions as a selective PGD2 receptor as well as DP, it sharply differs from DP in signaling pathways. Furthermore, CRTH2 but not DP induces migration of Th2 cells, eosinophils, and basophils in response to PGD2. Our findings elucidated a novel function of PGD2 on T cell responses, and CRTH2 is the first receptor that directly and selectively connects a major lipid mediator of activated mast cells with Th2 cells, eosinophils, and basophils.
Materials And Methods
Human DP cDNA was generated by reverse-transcription (RT)-PCR using a small intestine poly(A)+ RNA (CLONTECH Laboratories, Inc.) as a template and the following primers: 5′-CTTCCGAAGCCTTCACTCCAGCCCTCTGCTCCCG (sense) and 5′-GTTCTTTTCTAGAAAATGTGACATATTCCTCAGCTTACC (antisense). The DP cDNA was cloned into the HindIII-XbaI sites of pRc/CMV (Invitrogen), yielding an expression plasmid pRc/DP.
Mast cell supernatants were prepared as reported previously 6. K562 and Jurkat lines were transfected with pRc/B19 (a CRTH2 expression vector; references 5, 6), pRc/DP, or pRc/CMV and cloned as described 5. Human Th1 and Th2 lines were generated from the PBMCs of healthy adults by stimulation with a purified protein derivative of Mycobacterium tuberculosis in the presence of IL-12 and IFN-γ (Th1 condition) or with an extract of Dermatophagoides in the presence of IL-4 and neutralizing anti–IFN-γ and anti–IL-12 mAbs (Th2 condition), then expanded by IL-2 and used at days 7–14 for examinations as described previously 5. In some chemotaxis assays, Th2 cells highly expressing CRTH2 were enriched by sorting with anti-CRTH2 mAb BM16 5 a few days before use. PBMCs and granulocytes were isolated as described 6.
Ca2+ mobilization assays were performed as described 6. LTC4, LTB4, LXA4, prostanoids, and their synthetic analogues were purchased from Cayman Chemical except for fluprostenol (Biomol) and PGD2 methyl ester (Sigma-Aldrich). Thrombin and monocyte chemotactic protein 1 (MCP-1) were obtained from Sigma-Aldrich and PeproTech, respectively. For inactivation of G proteins, cells (2.5 × 106 cells/ml) were incubated with pertussis toxin (PTX; Sigma-Aldrich) or cholera toxin (CTX; Sigma-Aldrich) at 1 μg/ml for 2 h at 37°C.
3H-PGD2 Binding Analysis.
Cells washed and suspended in HBSS containing 10 mM Hepes, pH 7.3 (5 × 106 cells/100 μl) were incubated with 1 nM 3H-PGD2 (Amersham Pharmacia Biotech) for 60 min on ice. Cells were then separated from free 3H-PGD2 by centrifugation on RPMI 1640 containing 1 M sucrose, and radioactivity bound to the cells was measured on a liquid scintillation counter. Nonspecific binding was determined in the presence of 10 μM cold PGD2.
Basophils and eosinophils were purified as described 6. Neutrophils (99% pure as determined by May-Grünwald Giemsa staining) were purified from granulocytes with anti-CD16–coated magnetic beads (Miltenyi Biotec). Monocytes (98% pure), T cells (92% pure), and B cells (91% pure) were isolated from PBMCs with anti-CD14–, anti-CD3–, and anti-CD19–coupled magnetic beads (Miltenyi Biotec), respectively. RT-PCR was performed as described previously 5,6. Primers for DP were 5′-GCAACCTCTATGCGATGCA (sense) and 5′-CAAGGCTCGGAGGTCTTCT (antisense). PCR conditions were 30 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min.
This was performed on a 96-well microchemotaxis chamber with a 5-μm-pore filter (Neuroprobe) according to the manufacturer's instructions. In brief, cells (1–2 × 105 cells/50 μl) and test samples (29 μl) prepared in RPMI 1640 containing 10% FCS and 10 mM Hepes, pH 7.3, were applied to top and bottom wells, respectively, of the chamber. After incubation at 37°C for 1 h, cells in the bottom wells were counted by flow cytometry. In the case of blood leukocytes, cell types were determined by Wright's staining.
CRTH2 Is a Novel Receptor for PGD2.
We previously found that mast cells produce a CRTH2-specific agonist(s) that can induce Ca2+ mobilization in CRTH2-transfected K562 (K562/CRTH2) but not mock-transfected K562 (K562/neo) (Fig. 1 A; reference 6). In a screening of known mast cell products 13,16, we found that only PGD2 can induce Ca2+ mobilization in K562/CRTH2 but not K562/neo at nanomolar concentrations (Fig. 1 B). Other mast cell–related and –unrelated inflammatory mediators such as histamine, serotonin, platelet-activating factor, LTB4, LTC4, regulated upon activation normal T cell expressed and secreted (RANTES), MCP-1, eotaxin, IL-8, LXA4, fMLP, C5a, and angiotensin II did not induce any significant CRTH2-mediated Ca2+ mobilization (5; and data not shown). Treatment of mast cells with aspirin, a PG synthesis inhibitor, completely impaired production of the active component. The active component of the mast cell supernatants showed an elution profile identical to that of PGD2 on a reverse phase chromatography using a μRPC C2/C18 SC 2.1/10 column (Amersham Pharmacia Biotech) (data not shown). Having once responded to PGD2, K562/CRTH2 did not show any appreciable Ca2+ mobilization by the mast cell active fraction, and vice versa (cross-desensitization; Fig. 1 B). These observations demonstrate that PGD2 is the CRTH2 agonist that is produced by activated mast cells.
The above finding was supported by the experiments using K562/CRTH2 and DP-transfected K562 (K562/DP; Fig. 1 A). Both lines exhibited specific binding of 3H-PGD2 with similar numbers of binding sites per cell and Kd values of 31.3 and 37.7 nM, respectively, whereas 3H-PGD2 did not bind to K562/neo (Fig. 1C and Fig. D). A neutralizing CRTH2-specific mAb BM7 selectively inhibited the binding of 3H-PGD2 to K562/CRTH2 but not K562/DP, whereas control rat IgG had no effect on this assay (Fig. 1 D). To establish ligand specificity, binding affinities to the CRTH2 and DP of PGD2 analogues, major prostanoids, a thromboxane (TX)A2 receptor agonist U46619 8, a synthetic DP agonist BW245C 8, LTB4, and LXA4 were determined by competitive binding assay (Table). The results demonstrate that PGD2 binds to CRTH2 as well as DP with the highest affinity among various prostanoids. Interestingly, one of the PGD2 metabolites, 13,14-dihydro-15-keto-PGD2 (DK-PGD2), showed an affinity to CRTH2 as high as PGD2 to CRTH2, whereas DP showed at least 600-fold lower affinity to DK-PGD2 than PGD2 as estimated by the competition assay. In contrast, BW245C was highly selective for DP.
Consistently, Ca2+ mobilization was induced in both K562/CRTH2 and K562/DP by PGD2, whereas DK-PGD2 and BW245C were selectively effective in K562/CRTH2 and K562/DP, respectively, in Ca2+ mobilization assay (Fig. 1 E). No significant Ca2+ mobilization was induced in K562/CRTH2 by up to 250 nM PGA2, PGI2, and TXB2 (data not shown). From these findings, we have concluded that CRTH2 serves as the novel selective receptor for PGD2. Furthermore, DK-PGD2 and BW245C are useful reagents that specifically induce CRTH2- and DP-mediated responses, respectively.
Differences in Signaling Pathways Associated with CRTH2 and DP.
To gain insight into functional differences between CRTH2 and DP, signaling pathways associated with the two receptors were examined. DP induced cAMP generation in response to PGD2 as reported previously (Fig. 2 A; reference 9), and DP-mediated Ca2+ mobilization was blocked by CTX but not by PTX as expected (Fig. 2 B). In contrast, PTX but not CTX inhibited CRTH2-mediated Ca2+ mobilization (Fig. 2 B), and no production of cAMP was seen by interaction of PGD2 with CRTH2, as expected (Fig. 2 A). These results indicate that CRTH2 is coupled with Gαi-type G protein and DP with Gαs-type protein.
Since many seven-transmembrane Gαi-coupled receptors are involved in chemotaxis of leukocytes 1, we then addressed the issue of whether CRTH2 was responsible for chemotaxis. CRTH2-transfected Jurkat (Jurkat/CRTH2) was indeed attracted by PGD2 and DK-PGD2 but not by BW245C (Fig. 2 C). In sharp contrast, DP-transfected Jurkat (Jurkat/DP) was suppressed even in its spontaneous migration by PGD2 or BW245C. Thus, differential effects of PGD2 between CRTH2- and DP-expressing cells were clearly observed, presumably resulting from the difference in signaling molecules associated with the receptors.
Functional Properties of CRTH2 in Normal Leukocytes.
We next confirmed the above findings with normal leukocytes that naturally express CRTH2. Cultured Th2 but not Th1 cells showed specific binding of 3H-PGD2, which was blocked by anti-CRTH2 mAb BM7 (Fig. 3 A, and data not shown). In response to PGD2 and DK-PGD2, intracellular Ca2+ levels in Th2 cells increased (Fig. 3 B) and chemotactic migration was induced (Fig. 3 C). PGD2 had no effect on Ca2+ mobilization and chemotactic migration of Th1 cells (Fig. 3B and Fig. C). DP agonist BW245C had no effect on Ca2+ mobilization and chemotactic migration of Th2 cells (Fig. 3B and Fig. C). The neutralizing anti-CRTH2 mAb BM7 nearly completely inhibited PGD2- and DK-PGD2–dependent Ca2+ mobilization and migration in Th2 cells without affecting their responses to an irrelevant chemoattractant, MCP-1 (Fig. 3B and Fig. D). Similar results were obtained in all of five pairs of Th1 and Th2 lines from five different adults.
In RT-PCR analysis, CRTH2 was expressed highly in basophils and eosinophils as well as in Th2 lines and weakly in monocytes, whereas DP was expressed mainly in basophils (Fig. 3 E). When total peripheral blood leukocytes were tested for chemotactic migration, only basophils and eosinophils were significantly attracted by PGD2 (Fig. 3 F). When partially purified (neutrophil-depleted) leukocytes were tested for the assay, both basophils and eosinophils more clearly exhibited their migratory responses towards PGD2 and DK-PGD2 (Fig. 3 G). BW245C, a selective agonist for DP, showed no chemotactic effect on either cell type (Fig. 3 G). Anti-CRTH2 mAb BM7 but not control rat IgG blocked the migratory responses to PGD2 of basophils and eosinophils (Fig. 3 H). Circulating CRTH2+CD4+ lymphocytes also showed a tendency to migrate to PGD2 or DK-PGD2 at ∼1.5–3-fold higher frequencies than to vehicle or BW245C (data not shown).
The migratory responses of the above three normal cell types to PGD2 were completely abolished by PTX. A checkerboard analysis demonstrated that the effect of PGD2 on Th2 cells was mostly chemotactic, whereas on basophils and eosinophils there was a partly chemokinetic effect as reported on eosinophils 15. PGD2 is also known to weakly cross-react with receptors for PGE2, PGF2α, and TXA2 18. However, involvement of these receptors in chemotaxis was ruled out because PGE2, fluprostenol (a PGF2α receptor–selective agonist; reference 8), and U46619 (a TXA2 receptor–selective agonist; reference 8) had no significant chemotactic effect on these three cell types at nanomolar ranges. Taken together, these results indicate that CRTH2 mediates chemotactic/chemokinetic signal of PGD2 in a Gαi-dependent manner in Th2 cells, eosinophils, and basophils.
Prostanoid receptors have been classified into TX receptor (TP), PGE receptor (EP1, EP2, EP3, and EP4), PGF receptor (FP), PGI receptor (IP), and PGD receptor (DP; references 8, 10). In view of the structure, CRTH2 appears to be unique among members of prostanoid receptors in that it lacks any consensus amino acid sequence motifs that are shared by all other known prostanoid receptors 10. Indeed, a phylogenetic analysis shows that CRTH2 is more akin in the overall structure to the FPR subfamily members and receptors for LTs than known prostanoid receptors (Fig. 4). Furthermore, the chromosomal location of CRTH2 (11q12.2) (sequence data available from GenBank/EMBL/DDBJ under accession no. AC004126) differs from those of other receptors listed in Fig. 4, suggesting that CRTH2 might form a novel receptor gene subfamily. The mechanisms by which CRTH2 binds PGD2 with an affinity as high as DP remain to be clarified.
PGD2 is currently considered to elicit its biological actions through DP 8,10. In fact, well-established PGD2-induced activities such as vasodilation, relaxation of other various smooth muscles, and inhibition of platelet aggregation are apparently mediated by DP because they are also induced by a DP-specific agonist BW245C 18,19. However, several in vivo and ex vivo effects of PGD2 such as an increase in microvascular permeability, eosinophil infiltration, and goblet cell depletion are not mimicked by BW245C 20. Thus, questions on the existence of a BW245C-insensitive PGD2 receptor subtype(s) has been repeatedly proposed 20,21,22. In addition, although DK-PGD2 is generally thought to be biologically inactive in many systems, several investigators actually observed its effects on some PGD2-sensitive tissues 23. Our finding that CRTH2 is the novel PGD2 receptor functioning differently from DP may lead to the resolution of such long-standing questions.
A notable difference between CRTH2 and DP is the signaling molecules: Gαi is used for CRTH2 whereas Gαs is for DP. As shown by many cell types with Gαi-coupled receptors, CRTH2 leads Th2 cell, eosinophil, and basophil induction of chemotaxis/chemokinesis in response to PGD2. Consequently, these types of cells may accumulate at the sites of allergic inflammation 3,4. Indeed, it was demonstrated that topical application of PGD2 causes significant accumulation of eosinophils in the guinea pig conjunctiva and the dog trachea 20,24. In contrast, DP-mediated PGD2 signals caused reduction in spontaneous cell migration in DP-transfected Jurkat cells and had no effect on the migration of DP-expressing basophils in our study. These results suggest that DP-mediated signals may not, in nature, lead to chemotactic or chemokinetic behavior in leukocytes. Although PGD2 has been reported to modulate various activities of neutrophils, we could not detect any substantial expression of CRTH2 or DP in blood neutrophils 25. However, further studies are required to finally establish the actual involvement of these receptors in the PGD2-induced modulation of neutrophils.
Although the mechanisms by which DP contributes to the formation of allergic inflammation remain to be clarified, our results suggest that DP and CRTH2 may be cooperatively involved in allergic inflammation through different processes. A plausible scenario could be that PGD2 is largely produced by mast cells upon antigen stimulation, inducing local vasodilation via DP, which enhances extravasation of blood leukocytes 26, followed by chemotactic migration of Th2 cells, eosinophils, and basophils via CRTH2 in cooperation with other chemotactic mediators such as CC chemokines TARC and eotaxin 27,28. Thus, CRTH2 may be a new favorable target for allergic disease therapies. Generation of selective inhibitors for CRTH2 or CRTH2-deficient mice should help to elucidate the physiological and pathophysiological roles of the CRTH2/PGD2 system and its relative importance in the host defense mechanisms.
We thank Drs. K. Katayama and F.B. Hoshino (BML) for drawing the phylogenetic tree.