Molecular Human Reproduction Vol.8, No.2 pp. 181–187, 2002

Accumulation of CRTH2-positive T-helper 2 andT-cytotoxic 2 cells at implantation sites of human deciduain a prostaglandin D2-mediated manner

Toshihiko Michimata1, Hiroshi Tsuda1, Masatoshi Sakai1, Masaki Fujimura1,Kinya Nagata2, Masataka Nakamura3 and Shigeru Saito1,4

1Department of Obstetrics and Gynecology, Toyama Medical and Pharmaceutical University, Sugitani, Toyama 930-0194, 2R&DCenter, BML, Kawagoe, Saitama 350-1101 and 3Human Gene Sciences Center, Tokyo Medical and Dental University, Bunkyo-ku,Tokyo 113-8510, Japan

4To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, Toyama Medical and PharmaceuticalUniversity, 2630 Sugitani Toyama-shi, Toyama 930-0194, Japan. E-mail: s30saito@ms.toyama-mpu.ac.jp

T-helper (Th) 2-type cytokines predominate in decidua, plausibly accounting for protection of a semiallograft, theembryo and placenta, from attack by the maternal immune system. However, localization of Th2 and T-cytotoxic(Tc) 2 cells in decidua has not been reported, presumably because of the difficulty in detecting intracellular cytokinesin tissues. Here, by staining tissues for a novel surface maker of Th2/Tc2, the chemoattractant receptor-homologousmolecule CRTH2, which is expressed on Th2 cells, we show that CRTH2� Th2 cells and CRTH2� Tc2 cells aresignificantly increased at the materno–fetal interface (implantation site) in decidua. We also show that trophoblast,uterine epithelium and endometrial glands all express haematopoietic-type prostaglandin (PG) D2 synthase (hPGDS).Since CRTH2 is a chemoattractant receptor for PGD2 and mediates PGD2-dependent migration of blood Th2 cells,our findings suggest that Th2 and Tc2 cells may be recruited to the materno–fetal interface, at least in part in aPGD2-mediated manner.

Key words: CRTH2/decidua/PGD2/pregnancy/Th2

IntroductionCD4� helper T cells are classified as T-helper (Th) 1 cellsor Th2 cells according to patterns of cytokine production(Mosmann and Coffman, 1989; Romagnani, 1994; Mosmannand Sad, 1996). Th1 cells produce interleukin (IL)-2, interferon(IFN)-γ and tumour necrosis factor (TNF)-β. These mediatorsinduce activation of cytotoxic T (Tc) cells, presumably leadingto spontaneous abortion (Chaouat et al., 1995; Hill et al.,1995; Kirishnan et al., 1996; Piccinni et al., 1998) and pre-eclampsia (Saito et al., 1999a,b). Th2 cells produce IL-4, IL-5,IL-6, IL-10, IL-13 and leukaemia inhibitory factor (LIF),which inhibit the Th1 cell switch from Th0 cells. CD8�

cytotoxic T cells have similarly been shown to be divided intoTc1 cells synthesizing IL-2, IFN-γ and TNF-β, and Tc2 cellssynthesizing IL-4, IL-5, IL-6, IL-10 and IL-13 (Mosmann andSad, 1996; Adkis et al., 1999). The physiological protectionfrom fetal rejection is believed to be due to a Th2-type responseat the maternal–fetal interface (Lin et al., 1993; Wegmannet al., 1993; Russell et al., 1997; Piccinni et al., 1998; Saito,2000), although IL-4 and IL-10 double deficient mice showneither maternal nor feto–placental deficiency under very cleanconditions (Svensson et al., 2001).

© European Society of Human Reproduction and Embryology 181

We have found no significant differences in the sub-populations of Th2/Tc2 cells and Th1/Tc1 cells in peripheralblood T cells between non-pregnant women and women inearly pregnancy (Saito et al., 1999b,c). On the other hand, wehave shown that the proportion of Th2/Tc2 cells in decidua issignificantly higher than in peripheral blood, while the numberof Th1/Tc1 cells in decidua is significantly lower than inperipheral blood (Saito et al., 1999c).

Recruitment of Th2 and Tc2 cells into the endometrium mayoccur during endometrial decidualization in early pregnancy.Chemokine receptors are differentially expressed on Th1 andTh2 cells, resulting in different distributions of these cells(Sallusto et al., 1997, 1998a,b; Bonecchi et al., 1998; Loetscheret al., 1998; Zingoni et al., 1998; Annunziato et al., 1999;Imai et al., 1999).

We have developed a novel method to detect Th2 cellsby staining a chemoattractant receptor-homologous moleculeexpressed on Th2 cells (CRTH2). Compared with intracyto-plasmic cytokine molecules, detection of CRTH2 is easier byimmunohistochemistry. CRTH2 is selectively expressed onTh2 cells, Tc2 cells, eosinophils and basophils (Nagata et al.,

T.Michimata et al.

1999a,b; Cosmi et al., 2000; Tsuda et al., 2001), and has beenshown to be a second receptor for prostaglandin D2 (PGD2),additional to the PGD receptor (DP) (Hirai et al., 2001).In addition, CRTH2, but not DP receptor, mediates PGD2-dependent cell migration of Th2 cells (Hirai et al., 2001).Thus, PGD2-dependent migration of Th2 and Tc2 cells in earlypregnancy decidua may be mediated by CRTH2. Two PGD2

Table I. The numbers and ratios of lymphocyte subpopulations in decidua

Decidua P value

Implantation Area remote fromsite implantation site

CD45� (cells) 967.60 � 277.79 834.13 � 251.55a 0.0894CD56� (cells) 663.17 � 198.52 534.01 � 183.13 0.0487CD56�/CD45� (%) 69.10 � 11.03 63.76 � 8.97 0.0943CD3� (cells) 114.92 � 54.77 57.58 � 38.24 0.0012CD3�/CD45� (%) 11.60 � 2.61 6.42 � 2.06 �0.0001CD8� (cells) 61.54 � 34.04 30.47 � 23.06 0.0034CD8�/CD45� (%) 6.08 � 1.70 3.33 � 1.32 �0.0001CD3�/CD8– (cells)b 53.15 � 22.46 27.33 � 16.03 0.0006CD3�/CD8–/CD45� (%) 5.48 � 1.35 3.11 � 0.90 �0.0001

aPer 5000 cells.bCD3�CD8– � number of CD3� cells minus number of CD8� cells.

Figure 1. Localization of CRTH2�CD3� and CRTH2�CD8� cells in early pregnancy decidua. CRTH2� T cells were detected by doublefluorescence staining of CRTH2 and CD3 or CD8. Immunofluorescence staining for CD3 (A green), CD8 (D green), and CRTH2 (B and Ered) is shown. Yellow staining (arrows) in C and F represents the overlap of green and red, identifying CRTH2�CD3� (C) andCRTH2�CD8� (F) cells. Original magnification �400.

182

synthases (PGDS) have been characterized: (i) lipocalin type,or brain type, PGDS (Boie et al., 1995) and (ii) haematopoieticPGDS (hPGDS), which is expressed in placenta, Fallopiantube, lung and fetal liver (Kanaoka et al., 1997).

In this report, we show the localization of Th2 and Tc2cells in human decidua and the expression of hPGDS at theimplantation site by immunohistochemistry.

Materials and methods

Specimens

Human decidual tissues (6 and 10 weeks of gestation) wereobtained from induced abortion cases (n � 15). Informed consentwas obtained for all cases. All of the sampling and use of thetissues for this study were approved by the Toyama Medical andPharmaceutical University Ethics Committee. Tissues were fixed in10% neutral buffered formalin for 48 h and embedded in paraffin.

Monoclonal antibodies

CD45 (specific for leukocytes; 1:200 dilution; Dako, High Wycombe,Buckingham, UK), CD56-1B6 [a marker for natural killer (NK) cells;1:200 dilution; Novocastra, Newcastle upon Tyne, UK], CD3-PS1(CD3ε, a marker for T cells; 1:200 dilution; Novocastra), CD8 (amarker for cytotoxic/suppresser T cells; 1:20 dilution; Dako) and

CRTH2 at implantation sites

Figure 2. Localization of CRTH2�CD3� cells in early pregnancy decidua. CRTH2�CD3� cells were detected around extravilloustrophoblasts (A), decidual vessels and endometrial gland cells (B) and at distance from the implantation area (C) by immunofluorescencestaining of CRTH2 and CD3. Immunofluorescence staining for CRTH2–CD3� cells (green, arrows) and CRTH2�CD3� cells (yellow,arrowheads) is shown. Non-specific immune serum staining was used for a control (D). ET � extravillous trophoblast; V � blood vessel;EG � endometrial gland cell. Original magnification �400.

Figure 3. Localization of haematopoietic prostaglandin D2 synthase (PGDS) in the early pregnancy placenta and decidua. HaematopoieticPGDS was detected in uterine epithelial cells (A), syncytiotrophoblasts and cytotrophoblasts (B arrow head) and extravillous trophoblasts(B arrow), and endometrial gland cell (C) by immunofluorescence staining. Non-immune serum was used for a control (D). Originalmagnification �100.

183

T.Michimata et al.

Table II. Numbers of CRTH2�CD8� and CRTH2�CD4� T cells

Decidua P-value

Implantation Area remote fromsite implantation site

CD3� (cells/HPF) 20.28 � 7.51 8.21 � 4.07 �0.0001CRTH2�CD3� (cells/HPF) 4.11 � 1.69 1.14 � 0.37 �0.0001CRTH2�CD3�/CD3� (%) 22.05 � 9.84 16.23 � 7.44 0.0389CD8� (cells/HPF) 11.95 � 4.50 4.89 � 2.09 �0.0001CRTH2�CD8� (cells/HPF) 2.23 � 0.79 0.71 � 0.29 �0.0001CRTH2�CD8�/CD8� (%) 20.31 � 7.76 15.53 � 5.71 0.0323CD4� (cells/HPF)a 8.33 � 3.98 3.21 � 2.44 0.0001CRTH2�CD4� (cells/HPF) 1.85 � 0.56 0.40 � 0.15 �0.0001CRTH2� CD4�CD4� (%)a 25.81 � 18.18 16.64 � 6.84 0.0390

aCD4� � number of CD3� cells minus number of CD8� cells.CRTH2� cells among human decidual T cells were localized by double immunofluorescence staining ofCRTH2 and CD3 or CD8. The numbers of positive cells per five high-power field (HPF) at the implantationsite and at a distance from the implantation site are shown (n � 15).

cytokeratin-AE1/AE3 (specific for epithelial cells; 1:800 dilution;Boehringer Mannheim Biochemica, Mannheim, Germany) were allused for immunohistochemical staining.

Additionally, we used the recently developed rat monoclonalantibody to CRTH2, designated BM-16 (5 µg/ml) (Nagata et al.,1999a), and mouse monoclonal antibody to hPGDS, termed EBC45(5 µg/ml) (Tanaka et al., 2000).

Immunohistochemical staining

Paraffin-embedded tissues sectioned at a 5 µm thickness weredeparaffinized in xylene and rehydrated in a graded ethanol series tophosphate-buffered saline (PBS, pH 7.2). Antigen retrieval of thesections was performed by autoclaving in 10 mmol/l citrate buffer(pH 6.0) at 120°C for 10 min. The embedded sections were stainedimmunohistochemically using anti-CD45, anti-CD3 and anti-CD8antibodies in conjunction with a streptavidin-biotin-peroxidase kit(Nichirei, Tokyo, Japan).

Briefly, the sections were pretreated for 15 min in 3% hydrogenperoxide in methanol to quench endogenous peroxidase activity.Then, after a brief wash in a large amount of water, the sections wereincubated for 10 min in 10% normal goat serum to block non-specificbinding sites prior to the application of primary antibodies. Sectionswere then incubated overnight at 4°C with the primary monoclonalantibodies (mAbs). Biotinylated rabbit anti-mouse immunoglobulins(Ig) and then streptavidin-biotinylated peroxidase reagent were appliedaccording to the manufacturer’s instructions. After each incubationstep, sections were washed briefly in PBS. Antibody staining wascarried out with diaminobenzidine (DAB; Nichirei) for 10 min.Finally, sections were counterstained with Mayer’s haematoxylin andcoverslipped using mounting medium. Serial sections were stainedwith anti-cytokeratin antibody to detect extravillous trophoblasts.

Decidual T cells were also stained for CRTH2 along with CD3 orCD8. Briefly, after deparaffinization, antigen retrieval was performedby autoclaving at 120°C for 10 min. Sections were washed in PBSand incubated for 10 min in 10% normal goat serum prior toapplication of primary antibodies. For double staining, sections wereincubated overnight at 4°C with a combination of rat monoclonalanti-human CRTH2 antibody (5 µg/ml) and mouse monoclonal anti-human CD3 antibody (diluted 1:200), or anti-human CRTH2 antibody(5 µg/ml) and mouse monoclonal anti-human CD8 antibody (diluted1:20). Sections were washed in PBS and incubated for 30 min withAlexa Fluor 594-labelled goat anti-rat IgG (diluted 1:100; MolecularProbes, OR, USA) and Alexa Fluor 488-labelled goat anti-mouseIgG (diluted 1:100; Molecular Probes). Sections were mounted with

184

SlowFade antifade kits (Molecular Probes) and examined under aconfocal laser scanning microscope (LSM510; Karl Zeiss, Tokyo,Japan). All images were processed with LSM510 version 2.02software for image analysis.

Haematopoietic PGDS was detected by immunofluorescence stain-ing using the method described above with mouse monoclonalantibody against hPGDS (as a first antibody 5 µg/ml) and AlexaFluor 488-conjugated goat anti-mouse IgG (diluted 1:100) as asecond antibody.

Quantification and data analysis

The implantation sites were identified by cytokeratin staining to detecttrophoblasts and epithelial cells in decidua and the areas examinedwere classified as representing implantation site decidua (near theextravillous trophoblastic area), or non-implantation site (deciduadistant from the extravillous trophoblast area). CD45-, CD56-, CD3-and CD8-positive cells per 5000 cells were counted in these areas,and the ratios were calculated. CRTH2�CD3� and CRTH2�CD8�

cells per five different fields were separately counted under a confocallaser scanning microscope, and ratios were calculated.

Statistical analysis

Data are expressed as the mean � SD. Statistical differences wereevaluated by analysis with Student’s t-test. Values of P � 0.05 wereaccepted as indicating significance.

Results

Cell numbers of the lymphocyte subpopulation in decidua

To clarify the distribution of lymphocyte subsets at thematerno–fetal interface (implantation site), tissue specimenswere stained and the numbers of cells positive for CD56 (amarker for NK cells), CD3 (a marker for all T cells), CD8[a marker for cytotoxic T (Tc) cells] and CD45 (a marker forall leukocytes) were counted. The numbers of CD45� cellswere slightly increased at the materno–fetal interface (implanta-tion site) compared with those in the area remote from theimplantation site (Table I). There was a particularly denseinfiltration of CD56� cells in the decidua. The numbers ofCD56� cells at the materno–fetal interface (implantation site)were significantly higher than at the distant site, but nosignificant difference was noted in the CD56�/CD45� ratio

CRTH2 at implantation sites

between these locations. On the other hand, CD3� and CD8�

cells in decidua at the materno–fetal interface (implantationsite) were significantly more numerous compared with thosein decidua remote from the materno–fetal interface. The CD3�/CD45� and CD8�/CD45� ratios at the materno–fetal interfacewere significantly higher than those far from the implantationsite (Table I). Unfortunately, the marker for Th cells, thecommercially available CD4 monoclonal antibody (Dako andNovocastra) was not effective in staining paraffin-embeddedtissue samples (by any antigen-retrieval treatment such asautoclaving, microwaving or trypsin digestion); thus, thenumbers of Th cells (CD4� cells) were calculated by subtractingthe CD8� cell count from the CD3� cell count. The numbersof CD3�CD8– cells and the ratio of CD3�CD8– to CD45�

cells at the materno–fetal interface (implantation site) werealso significantly higher than those in decidua remote fromthe implantation site similar to CD8� cells (Table I). Theseresults clearly indicate that both the numbers and ratios of Thand Tc cells increase at the materno–fetal interface.

Localization of the CRTH2-positive T cell subpopulation

CD3� populations in decidua were further characterized bystaining with the antibody specific to CRTH2. CRTH2- (red,Figure 1B,E), CD3- or CD8-positive cells (green, Figure1A,D), and cells expressing both antigens (yellow, Figure1C,F, arrow) were separately counted. At the materno–fetalinterface (implantation site), many CD3� cells were surroundedby extravillous trophoblasts (Figure 2A), endometrial glandcells (Figure 2B) and decidual vessels (Figure 2B). On theother hand, few CD3� cells, limited to a perivascular distribu-tion, were found in sites remote from the materno–fetalinterface (Figure 2C). No staining was observed when non-immune mouse IgG was used in lieu of the first antibody(Figure 2D). The numbers of CRTH2� CD3� and CRTH2�

CD8� cells were greater at the materno–fetal interface than atthe distant site (Table II). The proportion of CRTH2� cellsamong CD3� cells at the materno–fetal interface was signific-antly higher than that at the remote site (Table II). A similarincrease in ratio of CRTH2� cells among CD8� cells wasobserved. The calculated numbers of CRTH2� CD3�CD8–

and the ratio of CRTH2�CD3�CD8– cells to CD3�CD8– cellsat the materno–fetal interface were also significantly higherthan those at the remote site (Table II). Our results indicatethat Tc2 cells, in addition to Th2 cells, predominate at thematerno–fetal interface.

Localization of hPGDS

To gain insight into a mechanism which may induce thepredomination of Th2 and Tc2 cells at the implantationsite, the localization of cells producing hPGDS, an enzymesynthesizing PGD2, was determined by immunofluorescence.Uterine epithelial cells (Figure 3A),villous trophoblasts (Figure3B, arrow head) and extravillous trophoblasts (Figure 3B,arrow) were shown to express hPGDS, as did the cells of theendometrial glands (Figure 3C). Decidual stromal cells andstromal cells in the chorion did not express hPGDS (Figure3A,B,C). No staining was recognized when non-immune mouseIgG was used as the first antibody (Figure 3D). These results

185

suggest that PGD2 is produced by uterine epithelial cells,endometrial gland cells and trophoblast.

DiscussionBecause T cells are scattered throughout the decidua basalisand decidua parietalis (Kabawat et al., 1985; Bulmer 1992;Vassiliadou and Bulmer, 1996, 1998; King et al., 1998a; King2000), few reports have compared the distribution of T cellsubsets at the materno–fetal interface, i.e. the implantationsite, with that in parietal decidua. In the present study, weshow that CD8� T cells and CD3�CD8– T cells (i.e. CD4�

cells) are more numerous at the implantation site than remotefrom it. Not only T cells, but also CD56� NK cells, increasedat the materno–fetal interface in terms of total lymphocytenumbers. However, only the T cell ratio against the totalnumber of lymphocytes increased at the materno–fetal inter-face, suggesting that T cells are selectively recruited to thesite of implantation.

Decidual CD16–CD56bright NK cells express weak surfaceCD8 and intracytoplasmic CD3ε, γ, and δ (Nishikawa et al.,1991; King et al., 1998a,b). Thus, cells positive for CD8and CD3 include CD8� T cells and CD16–CD56bright

CD8dim NK cells, and CD3� T cells and CD16–CD56bright NKcells respectively. In this study, when CD8bright and CD3bright

cells were considered as CD3� cells, the percentage of CD8�

cells and CD3� cells in CD45� cells was 3–6% and 6–11%respectively. In contrast, the CD56�/CD45� ratio was ~70%,indicating that CD8bright cells and CD3bright cells were probablyT cells, not CD16–CD56bright NK cells.

Here, we have demonstrated increases in numbers ofCRTH2�CD3�CD8– and CRTH2�CD8� cells as well asincreases in the ratios of CRTH2�CD8� and CRTH2�

CD3�CD8– cells against total T cells at the materno–fetalinterface compared with those far from the implantationsite. These results imply that Th2 and Tc2 cells selectivelyaccumulate at the site of implantation. This assumption maybe supported by the observation that CRTH2� T cells wereseen around decidual blood vessels, endometrial gland cellsand extravillous trophoblasts, while few cells positive forCRTH2 were present around blood vessels in decidua awayfrom the implantation site. Thus, it is likely that Th2 and Tc2cells migrate into the materno–fetal interface by attraction ofchemotactic factor(s) specific for Th2 and Tc2 cells. Suchfactors may be produced by trophoblast and endometrial glandcells at the implantation site. Drake et al. have reported thatcytotrophoblasts can attract monocytes and CD56bright NK cellsby producing monocyte inflammatory protein (MIP) 1α (Drakeet al., 2001). They have also reported that cytotrophoblast-conditioned medium contains a chemotactic factor for T cells,though they did not identify this substance. Dang and Heybornehave reported that uterine NKT cells recognize a class I/classI-like molecule other than CD1 and the fetal class I moleculecould expand the number of uterine NKT cells (Dang andHeyborne, 2001). These data suggest that the fetus or fetaltrophoblasts can regulate the maternal immune system.

Our data demonstrate that hPGDS is expressed not only inmaternal endometrial gland cells and endometrial epithelial

T.Michimata et al.

cells, but also in fetal trophoblast, presumably resulting in thesecretion, at the implantation site, of PGD2 that functions as achemoattractant for CRTH2� T cells. Indeed, decidual endo-metrial gland cells and chorionic tissues have been shown tosecrete enough PGD2 (Mitchell et al., 1982; Norwitz andWilson, 2000) and we observed an overlap in the localizationsof hPGDS-expressing cells and CRTH2� T cells. Collectively,Th2 and Tc2 cells may be recruited from peripheral blood intothe implantation site, at least in part in a PGD2-mediatedmanner.

Expression of hPGDS has been found to be high in theFallopian tube, suggesting that PGD2 may foster physiologicalor maturation functions in the embryo (Kanaoka et al.,1997). Luteotropic effects of PGD2 have been reported inhumans (Bennegard et al., 1990). Progesterone has beenreported to induce the conversion of Th0 cells into Th2 cells(Piccinni et al., 1995; Lim et al., 1998). Thus, progesteroneand PGD2 may interact with each other within decidual tissues,resulting in Th2- and Tc2-predominant immune conditions.Furthermore, endometrial PGD2 release increases during themid-luteal phase, when blastocyst implantation occurs (Reesand Kelly, 1986). PGD2 secretion by monocytes and separatedglandular cells of human decidua has also been reported(Norwitz et al., 1992; Norwitz and Wilson, 2000). The degrada-tion product of PGD2, 15-deoxy-∆12, 14-PGJ2

, stimulatesperoxisome proliferator-activated receptor-γ (PPARγ) activityin the trophoblast, and activated PPARγ enhances trophoblastdifferentiation (Schaiff et al., 2000). PGD2 is therefore thoughtto have essential functions in reproduction.

Interestingly, PGD2, but not other major eicosanoids, pro-duced by parasites, specifically impedes the TNFα-triggeredmigration of epidermal Langerhans cells through the adenylatecyclase-coupled PGD2 receptor (DP receptor) (Angeli et al.,2001). In response to stimulation occurring during infection,Langerhans cell are activated, and a proportion of them migratevia afferent lymphatics to regional lymph nodes where theyaccumulate as immunostimulatory dendritic cells. Upon arrivalin the lymph nodes, mature dendritic cells translate the tissue-derived information into the language of T cells, providingthem with an antigen-specific signal. So, the inhibition ofLangerhans cells migration could represent a stratagem for theparasites to escape the host immune system. Human deciduacontains potent immunostimulatory dendritic cells (Kammereret al., 2000). PGD2 in the decidua may inhibit the dendriticcell migration towards draining lymph nodes, and as a resultmaternal T cells would not attack fetal cells. Thus, PGD2 maycontribute to the maintenance of pregnancy by suppressingantigen presentation by dendritic cells and by controlling theTh1/Th2 balance through its dual receptor systems, and, assuggested by this study, CRTH2.

ReferencesAkdis, M., Simon, H.U., Weigl, L., Kreyden, O., Blaser, K. and Akdis, C.A.

(1999) Skin homing (cutaneous lymphocyte-associated antigen-positive)CD8� T cells respond to superantigen and contribute to eosinophilia andIgE production in atopic dermatitis. J. Immunol., 163, 466–475.

Angeli, B., Faveeuw, C., Roye, O., Fontanine, J., Teissier, E., Capron, A.,Wolowczwk, I., Capron, M. and Trottein, F. (2001) Role of the parasite-

186

derived prostaglandin D2 in the inhibition of epidermal Langerhans cellmigration during Schistosomiasis infection. J. Exp. Med., 193, 1135–1147.

Annunziato, F., Cosmi, L., Galli, G., Beltrame, C., Romagnani, P., Manetti,R., Romagnani, S. and Maggi, E. (1999) Assessment of chemokine receptorexpression by human Th1 and Th2 cells in vitro and in vivo. J. Leukoc.Biol., 65, 691–699.

Bennegard, B., Hahlin, M. and Hamberger, L. (1990) Luteotropic effects ofprostaglandins I2 and D2 on isolated human corpora luteum. Fertil. Steril.,54, 459–464.

Boie, Y., Sawyer, N., Slipetz, D.M., Metters, K.M. and Abramovitz, M. (1995)Molecular cloning and characterization of the human prostanoid DPreceptor. J. Biol. Chem., 270, 18910–18916.

Bonecchi, R., Bianchi, G., Bordignon, P.P., D’Ambrosio, D., Lang, R., Borsatti,A., Sozzani, S., Allavena, P., Gray, P.A., Mantovani, A. et al. (1998)Differential expression of chemokine receptors and chemotacticresponsiveness of type 1 T helper cells (Th1s) and Th2s. J. Exp. Med.,187, 129–134.

Bulmer, J.N. (1992) Immune aspects of pathology of the placental bedcontributing to pregnancy pathology. Bailliere’s Clin. Obstet. Gynaecol., 6,461–488.

Chaouat, G., Meliani, A.A., Martal, J., Raghupathy, R., Elliot, J., Mosmann,T. and Wegmann, T.G. (1995) IL-10 prevents naturally occurring fetal lossin the CBA�DBA/2 mating combination, and local defect in IL-10production in this abortion-prone combination is corrected by in vivoinjection of IFN-τ. J. Immunol., 154, 4261–4268.

Cosmi,L., Annunziato, F., Iwasaki, M., Galli, G., Manetti, R., Maggi, E.,Nagata, K. and Romagnani, S. (2000) CRTH2 is the most reliable markerfor the detection of circulating human type 2 Th and type 2 T cytotoxiccells in health and disease. Eur. J. Immunol., 30, 2972–2979.

Dang, Y. and Heyborne, K.D. (2001) Regulation of uterine NKT cells by afetal class I molecule other than CD1. J. Immunol., 166, 3641–3644.

Drake, P.M., Gunn, M.D., Charo, I.F., Tsou, C.L., Zhou, Y., Huang, L. andFisher, S.J. (2001) Human placental cytotrophoblast attract monocytes andCD56bright natural killer cells via the actions of monocyte inflammatoryprotein 1α. J. Exp. Med., 193, 1199–1212.

Hill, J.A., Polgar, K. and Anderson, D.J. (1995) T-Helper 1-type immunity totrophoblast in women with recurrent spontaneous abortion. J. Am. Med.Assoc., 273, 1933–1936.

Hirai, H., Tanaka, K. Yoshie, O., Ogawa, K., Kenmotsu, K., Takamori, Y.,Ichimasa, M., Sugamura, K., Nakamura, M., Takano, S. et al. (2001)Prostaglandin D2 selectively induces chemotaxis in T helper type 2 cells,eosinophils, and basophils via seven-transmembrane receptor CRTH2. J.Exp. Med., 193, 255–261.

Imai, T., Nagira, M., Takagi, S., Kakizaki, M., Nishimura, M., Wang, J., Gray,P.W., Matsushima, K. and Yoshie, O. (1999) Selective recruitment ofCCR4-bearing Th2 cells toward antigen-presenting cells by the CCchemokines thymus and activation-regulated chemokine and macrophage-derived chemokine. Int. Immunol., 11, 81–88.

Kabawat, S.E., Mostoufi-Zadeh, M., Driscoll, S.G. and Bhan, A.K. (1985)Implantation site in normal pregnancy. A study with monoclonal antibodies.Am. J. Pathol., 118, 76–84.

Kanaoka, Y., Ago, H., Inagaki, E., Nakayama, T., Miyano, M., Kikuno, R.,Fujii, Y., Eguchi, N., Toh, H., Urade, Y. et al., (1997) Cloning and crystalstructure of hematopoietic prostaglandin D synthase. Cell, 90, 1085–1095.

King, A. (2000) Uterine leukocytes and decidualization. Hum. Reprod. Update,6, 28–36.

King, A., Burrows, T., Verma, S., Hiby, S. and Loke, Y.W. (1998a) Humanuterine lymphocytes. Hum. Reprod. Update, 4, 480–485.

King, A., Gardner, L., Sharkey, A. and Loke, Y.W. (1998b) Expression ofCD3ε, CD3ζ, and RAG-1/RAG-2 in decidual CD56� NK cells. Cell.Immunol., 183, 99–105.

Kammerer, U., Schoppet, M., McLellan, A.D., Kapp, M., Huppertz, H.I.,Kampgen, E. and Dietl, J. (2000) Human decidua contains potentimmunostimulatory CD83� dendritic cells. Am. J. Pathol., 157, 159–169.

Krishnan, L., Guilbert, L.J., Wegmann, T.G., Belosevic, M. and Mosmann,T.R. (1996) T helper 1 response against Leishmania major in pregnantC57BL/6 mice increases implantation failure and fetal resorptions. J.Immunol., 156, 653–662.

Lim, K.J., Odukoya, O.A., Ajjan, R.A., Li, T.C., Weetman, A.P. and Cooke,I.D. (1998) Profile of cytokine mRNA expression in peri-implantationhuman endometrium. Mol. Hum. Reprod., 4, 77–81.

Lin, H., Mosmann, T.R.,Guilbert, L., Tuntipopipat, S. and Wegmann, T.G.(1993) Synthesis of T helper 2-type cytokines at the maternal–fetal interface.J. Immunol., 151, 4562–4573.

CRTH2 at implantation sites

Loetscher, P., Uguccioni, M., Bordoli, L., Baggiolini, M. and Moser, B. (1998)CCR5 is characteristic of Th1 lymphocytes. Nature, 391, 344–345.

Mitchell, M.D., Kraemer, D.L. and Strickland, D.M. (1982) The humanplacenta: a major source of prostaglandin D2. Prostaglandins Leukot. Med.,8, 383–387.

Mosmann, T.R. and Coffman, R.L. (1989) Th1 and Th2 cells: differentpatterns of lymphokine secretion lead to different functional properties.Ann. Rev. Immunol., 7, 145–173.

Mosmann, T.R. and Sad, S. (1996) The expanding universe of T-cell subsets:Th1, Th2 and more. Immunol. Today, 17, 138–146.

Nagata, K., Tanaka, K., Ogawa, K., Kenmotsu, K., Imai, T., Yoshie, O., Abe,H., Tada, K., Nakamura, M., Sugamura, K. et al. (1999a) Selectiveexpression of a novel surface molecule by human Th2 cells in vivo. J.Immunol., 162, 1278–1286.

Nagata, K., Hirai, H., Tanaka, K., Ogawa, K., Aso, T., Sugamura, K.,Nakamura, M. and Takano, S. (1999b) CRTH2, an orphan receptor of T-helper-2-cells, is expressed on basophils and eosinophils and responds tomast cell-derived factor(s). FEBS Letters, 459, 195–199.

Nishikawa, K., Saito, S., Morii, T., Hamada, K., Ako, H., Narita, N., Ichijo,M., Kurahayashi, M. and Sugamura, K. (1991) Accumulation of CD16–

CD56� natural killer cells with high affinity interleukin 2 receptors inhuman early pregnancy decidua. Int. Immunol., 3, 743–750.

Norwitz, E.R. and Wilson, T.W. (2000) Secretory component: a potentialregulator of endometrial-decidual prostaglandin production in early humanpregnancy. Am. J. Obstet. Gynecol., 183, 108–117.

Norwitz, E.R., Starkey, P.M. and Bernal, A.L. (1992) Prostaglandin D2production by term human decidua: cellular origins defined using flowcytometry. Obstet. Gynecol., 80, 440–445.

Piccinni, M.P., Guidizi, M.G., Biagiotti, R., Beloni, L., Giannarini, L.,Sanpognaro, S., Parronchi, P., Manetti, R., Annunziato, F. and Livi, C.(1995) Progesterone favors the development of human T helper cellsproducing Th-2 type cytokines and promotes both IL-4 production andmembrane CD30 expression in established Th1 cells clones. J. Immunol.,155, 128–133.

Piccinni, M.P., Beloni, L., Livi, C., Maggi, E., Scarselli, G. and Romagnani,S. (1998) Defective production of both leukemia inhibitory factor and type2 T-helper cytokines by decidual T cells in unexplained recurrent abortions.Nature Med., 4, 1020–1024.

Rees, M.C. and Kelly, R.W. (1986) Prostaglandin D2 release by endometriumand myometrium. Br. J. Obstet. Gynaecol., 93, 1078–1082.

Romagnani, S. (1994) Lymphokine production by human T cells in diseasestates. Ann. Rev. Immunol., 12, 227–257.

Russell, A.S., Johnston, C., Chew, C. and Maksymowych, W.P. (1997)Evidence to reduced Th1 function in normal pregnancy: a hypothesis forthe remission of rheumatoid arthritis. J. Rheumatol., 24, 1045–1050.

Saito, S. (2000) Cytokine network at the feto-maternal interface. J. Reprod.Immunol., 47, 87–103.

Saito, S., Umekage, H., Sakamoto, Y., Sakai, M., Tanebe, K., Sasaki, Y. andMorikawa, H. (1999a) Increased T-helper-1-type immunity and decreased

187

T-helper-2-type immunity in patients with preeclampsia. Am. J. Reprod.Immunol., 41, 297–306.

Saito, S., Sakai, M., Sasaki, Y., Tanebe, K., Tsuda, H. and Michimata, T.(1999b) Quantitative analysis of peripheral blood Th0, Th1, Th2 and theTh1:Th2 cell ratio during normal human pregnancy and preeclampsia. Clin.Exp. Immunol., 117, 550–555.

Saito, S., Tsukaguti, N., Hasegawa, T., Michimata, T., Tsuda, H. and Narita,N. (1999c) Distribution of Th1, Th2, and Th0 and the Th1/Th2 cell ratiosin human peripheral and endometrial T cells. Am. J. Reprod. Immunol., 42,240–245.

Sallusto, F., Mackay, C.R. and Lanzavecchia, A. (1997) Selective expressionof the eotaxin receptor CCR3 by human T helper 2 cells. Science, 277,2005–2007.

Sallusto, F., Lanzavecchia, A. and Mackay, C.R. (1998a) Chemokines andchemokine receptors in T-cell priming and Th1/Th2-mediated responses.Immunol. Today, 19, 568–574.

Sallusto, F., Lenig, D., Mackay, C.R. and Lanzavecchia, A. (1998b) Flexibleprograms of chemokine receptor expression on human polarized T helper1 and 2 lymphocytes. J. Exp. Med., 187, 875–883.

Svensson, L., Arvola, M., Sallstrom, M.A., Holmdahl, R. and Mattsson, R.(2001) The Th2 cytokines IL-4 and IL-10 are not crucial for the completionof allogeneic pregnancy in mice. J. Reprod. Immunol., 51, 3–7.

Schaiff, W.T., Carlson, M.G., Smith, S.D., Levy, R., Nelson, D.M. andSadovsky, Y. (2000) Peroxisome proliferator-activated receptor-γ modulatesdifferentiation of human trophoblast in a ligand-specific manner. J. Clin.Endocrinol. Metab., 85, 3874–3881.

Tanaka, K., Ogawa, K., Sugamura, K., Nakamura, M., Takano, S. and Nagata,K. (2000) Differential production of prostaglandin D2 by human helper Tcell subsets. J. Immunol., 164, 2277–2280.

Tsuda, H., Michimata, T., Sakai,M., Nagata, K., Nakamura, M. and Saito, S.(2001) A novel surface molecule of Th2- and Tc2-type cells, CRTH2expression on human peripheral and decidual CD4� and CD8� T cellsduring the early stage of pregnancy. Clin. Exp. Immunol., 123, 105–111.

Vassiliadou, N. and Bulmer, J.N. (1996) Quantitative analysis of T lymphocytesubsets in pregnant and nonpregnant human endometrium. Biol. Reprod.,55, 1017–1022.

Vassiliadou, N. and Bulmer, J.N. (1998) Characterization of endometrial Tlymphocyte subpopulations in spontaneous early pregnancy loss. Hum.Reprod., 13, 44–47.

Wegmann, T.G., Lin, H., Guilbert, L. and Mosmann, T.R. (1993) Bidirectionalcytokine interactions in the maternal–fetal relationship: is successfulpregnancy a TH2 phenomenon? Immunol. Today, 14, 353–356.

Zingoni, A., Soto, H., Hedrick, J.A., Stoppacciaro, A., Storlazzi, C.T.,Sinigaglia, F., D’Ambrosio, D., O’Garra, A., Robinson, D., Rocchi, M.et al. (1998) The chemokine receptor CCR8 is preferentially expressed inTh2 but not Th1 cells. J. Immunol., 161, 547–551.

Received on May 21, 2001; resubmitted on August 29, 2001; accepted onNovember 2, 2001