reproduction...ppard without signiﬁcant binding to ptgir (seimandi et al. 2005) was used. since...

R
EPRODUCTIONRESEARCH
Expression patterns and role of prostaglandin-endoperoxidesynthases, prostaglandin E synthases, prostacyclin synthase,prostacyclin receptor, peroxisome proliferator-activatedreceptor delta and retinoid x receptor alpha in ratendometrium during artificially-induced decidualization

Carolina Gillio-Meina1, Sen Han Phang1, James P Mather1, Brian S Knight1

and Thomas G Kennedy1,2

Departments of 1Physiology and Pharmacology and 2Obstetrics and Gynecology, The University of Western Ontario,London, Ontario, Canada N6A 5C1

Correspondence should be addressed to T G Kennedy; Email: [email protected]

Abstract

To determine if changes in endometrial expression of the enzymes and receptors involved in prostaglandin (PG) synthesis and action

might provide insights into the PGs involved in the initiation of decidualization, ovariectomized steroid-treated rats at the equivalent of

day 5 of pseudopregnancy were given a deciduogenic stimulus and killed at various times up to 32 h thereafter. The expression of

PG-endoperoxide synthases (PTGS1 and PTGS2), microsomal PGE synthases (PTGES and PTGES2), cytosolic PGE synthase (PTGES3),

prostacyclin synthase (PTGIS), prostacyclin receptor, peroxisome proliferator-activated receptor d (PPARD) and retinoid x receptor a

(RXRA) in endometrium was assessed by semiquantitative RT-PCR, western blot analyses and immunohistochemistry. In addition, to

determine which PG is involved in mediating decidualization, we compared the ability of PGE2, stable analogues of PGI2, L165041

(an agonist of PPARD), and docasahexanoic acid (an agonist of RXRA) to increase endometrial vascular permeability (EVP, an early event

in decidualization), and decidualization when infused into the uterine horns of rats sensitized for the decidual cell reaction (DCR). EVP

was assessed by uterine concentrations of Evans blue 10 h after initiation of infusions. DCR was assessed by the uterine mass 5 days after

the initiation of the infusions. Because enzymes associated with the synthesis of PGE2, including PTGS2, are up-regulated in response to a

deciduogenic stimulus and because PGE2 was more effective than the PGI2 analogues and PPARD and RXRA agonists in increasing EVP

and inducing decidualization, we suggest that PGE2 is most likely the PG involved in the initiation of decidualization in the rat.

Reproduction (2009) 137 537–552

Introduction

An increase in the endometrial vascular permeability(EVP) at the sites of blastocyst apposition is one of theearliest macroscopically identifiable signs of blastocystimplantation (Psychoyos 1973). In rodents, this step isfollowed by proliferation and differentiation of endo-metrial stromal cells into decidual cells, a processtermed decidualization (reviewed by Kennedy et al.2007). In addition, the endometrium of rodents respondsto non-specific stimuli with increased EVP and sub-sequent decidualization provided that the stimulus isapplied at the appropriate time (Psychoyos 1973).

Previous studies conducted in various species haveshown that during pregnancy, the levels of prostaglan-dins (PGs) are elevated at implantation sites, suggesting arole of PGs in implantation (reviewed by Kennedy et al.2007). Moreover, numerous studies have shown that

q 2009 Society for Reproduction and Fertility

ISSN 1470–1626 (paper) 1741–7899 (online)

administration of inhibitors of PG synthesis and PGantagonists can inhibit or delay the initiation ofimplantation (reviewed by Kennedy et al. 2007). Inaddition, indomethacin, an inhibitor of PG biosynthesis,reduces the artificially induced decidual cell reaction(DCR), suggesting a role for PGs in the process ofdecidualization (reviewed by Kennedy et al. 2007). Theeffects of this inhibitor can be reversed, at least partially,by exogenous PGs (reviewed by Kennedy et al. 2007).

PG production requires the liberation of arachidonicacid (AA) from membrane phospholipids by phospho-lipases. AA is then converted into PG endoperoxide H2

(PGH2) by PG-endoperoxide synthases (PTGS), alsotermed cyclooxygenases. Two isoforms of PTGS, knownas PTGS1 and PTGS2, are involved in the process. OncePGH2 is formed, it is the substrate for the synthesis of fivedifferent prostanoids, PGE2, PGI2, PTGDS, PGF2a and

DOI: 10.1530/REP-08-0294

Online version via www.reproduction-online.org

Downloaded from Bioscientifica.com at 08/21/2021 09:14:26AMvia free access

http://dx.doi.org/10.1530/REP-08-0294

538 C Gillio-Meina and others

thromboxane A2 via specific synthases (Wang & Dey2005). Previous studies have suggested that of theseprostanoids only PGE2 and PGI2 are involved inimplantation. PGE synthases (PTGES) catalyze thesynthesis of PGE2 from PGH2. Several forms of PTGESincluding microsomal PTGES (PTGES), microsomalPTGES 2 (PTGES2), and cytosolic PTGES (PTGES3)have been characterized and are coupled with PTGSs.Prostaglandin I2 (PGI2 or prostacyclin) synthesis occurswhen PGH2 is converted to PGI2 by PGI synthase(PTGIS). Once PGs are produced, they exert their actionsvia specific cell surface G-protein coupled receptors andnuclear receptors (Helliwell et al. 2004). PGE2 acts viathe PGE family of receptors (PTGER1-PTGER4), whereasPGI2 acts via prostacyclin receptor (PTGIR). Recently,based on the ability of some PGI2 analogues to act asligands for the nuclear receptor peroxisome proliferator-activated receptor d (PPARD) that then heterodimerizesto retinoid x receptor a (RXRA), it has been suggestedthat the effects of PGI2 may be mediated by a PPARD–RXRA complex (Lim & Dey 2002).

It has been proposed that PGI2, acting via PPARD, isthe primary PG mediating implantation and decidualiza-tion in mice (Lim et al. 1999). This proposal was basedon studies in PTGS2 null mice where analogues ofPGI2 that induce DNA-binding and transcriptionalactivation by PPARD and PPARD ligands were moreeffective than PGE2 at inducing implantation anddecidualization. Moreover, in mice, it has been shownthat PTGS2, PTGIS, and the receptor PPARD coexist atthe implantation site in mouse endometrium during theperi-implantation period (Lim et al. 1999). Theseobservations have been interpreted to indicate that inthe mouse the PTGS2/PTGIS system generates PGI2 andmediates embryo implantation via PPARD (Lim et al.1999). By contrast, in the hamster, it has been reportedthat PGE2 but not PGI2 is the major PG produced, and isa product of induced PTGS2 and PTGES at theimplantation site (Wang et al. 2004). Based primarilyon the responses to exogenously administered PGs, it hasbeen proposed that PGE2 is the primary PG involved inrats (reviewed by Kennedy et al. 2007). Moreover, theexpression of mRNA for three subtypes of the PGEreceptors (Ptger2, Ptger3, and Ptger4) is correlated withendometrial preparation for decidualization in the ratand consistent for a role for PGE2 in this process (Papay &Kennedy 2000). Whether the mouse or the hamster isrepresentative of mammalian species in general isunknown. Thus, to determine if changes in endometrialexpression of the enzymes and receptors involved in PGsynthesis and action could provide insight into the PGsinvolved in the initiation of decidualization in the rat,we assessed levels of transcripts and proteins for PTGS1and PTGS2, PTGES, PTGES2, PTGES3, PTGIS, PTGIR,PPARD, and RXRA. We chose to study artificiallyinduced decidualization rather than pregnancy becausethe timing of endometrial stimulation can be readily

Reproduction (2009) 137 537–552

controlled in this model. The early endometrialresponses to artificial stimuli and implanting blastocystsare thought to be similar. Furthermore, to determinewhich PG has a primary role in initiation of decidualiza-tion in the rat endometrium, we compared theeffectiveness of PGE2 and various analogues of PGI2 toincrease EVP and induce decidualization when infusedinto the uterine lumen of rats sensitized for the DCR.PGI2 analogues were selected based on their differentaffinities for PTGIR and PPARD. Those tested werecarbaprostacyclin (cPGI2), a ligand for both PTGIR andPPARD (Forman et al. 1997); cicaprost, a ligand forPTGIR but not PPARD (Forman et al. 1997); and AFP-07a high-affinity ligand for PTGIR (Chang et al. 1997) butno information available for binding to PPARD. Inaddition, L165041, a synthetic high-affinity ligand forPPARD without significant binding to PTGIR (Seimandiet al. 2005) was used. Since PPARD heterodimerizeswith RXRA upon activation, docasahexanoic acid(DHA), a ligand for RXRA (Germain et al. 2006) wasalso included.

Results

Pattern of expression of PTGS1, PTGS2, PTGES,PTGES2, PTGES3, PTGIS, PTGIR, PPARD, and RXRAin the rat endometrium

Semiquantitative RT-PCR indicated that Ptgs1 transcriptlevels were significantly lower (P!0.05) in theendometrium at 4, 8, 16, and 32 h after the applicationof a deciduogenic stimulus compared with 0 and 2 h(Fig. 1). Because no differences were detected betweenthe stimulated and non-stimulated endometrium at 8and 32 h (PO0.05), the results suggest that the decreasewas due to the time and not to the stimulus. Although,levels of Ptgs1 mRNA changed, levels of PTGS1 proteindid not change significantly (PO0.05) at any of thetimes studied (Fig. 1).

Ptgs2 transcript levels significantly increased at 2 and4 h after the deciduogenic stimulus was given (P!0.05),and the levels returned to prestimulated levels at 8, 16,and 32 h (Fig. 2). In contrast to PTGS1 protein levels,PTGS2 protein was significantly higher at 8 h andremained elevated at 16 and 32 h (P!0.05). PTGS2protein levels were significantly (P!0.05) higher instimulated endometrium at 8 and 32 h than in non-stimulated endometrium, indicating that the increasewas a consequence of the deciduogenic stimulus ratherthan the passage of time (Fig. 2).

Ptges mRNA levels increased significantly at 16 h andremained elevated at 32 h (P!0.05; Fig. 3). The increaseat 32 h was the result of the deciduogenic stimulus sincetranscript levels in the stimulated endometrium werehigher (P!0.05) than in the non-stimulated endometriumat this time. There were no significant differences(PO0.05) in PTGES protein levels in the stimulated

www.reproduction-online.org


Figure 1 Expression of PTGS1 in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8,and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.

Prostaglandins in rat decidualization 539

endometrium at any time studied. However, PTGESprotein levels were higher at 32 h in the stimulatedendometrium than the non-stimulated at the sametime (P!0.05). This was a consequence of a decrease inPTGES protein levels in the non-stimulated endometrium.

No changes in Ptges2 mRNA levels were detected inthe stimulated endometrium at any time studied (Fig. 4).However, the transcript levels were significantly(P!0.05) higher in the non-stimulated than the stimu-lated endometrium at 8 and 32 h, suggesting that thestimulus decreased the levels. The PTGES2 protein levelswere significantly higher at 2 and 4 h when comparedwith all the other time points (P!0.05; Fig. 4).


No changes in mRNA levels were detected for Ptges3 atany time studied (Fig. 5). By contrast, western blot analysisindicated that PTGES3 protein levels decreased (P!0.05)with time in stimulated endometrium, being lowest at 16and 32 h. However, this decrease was not attributable tothe deciduogenic stimulus since a similar decrease wasseen in non-stimulated endometrium at 8 and 32 h.

There were no statistical differences (PO0.05) of Ptgistranscript levels at any time following stimulation orbetween stimulated and non-stimulated endometrium at8 and 32 h (Fig. 6). When western blot analyses wereconducted using antibodies against PTGIS protein, twoimmunoreactive bands were obtained; one at a molecu-lar mass of 56–57 kDa and one at w 75 kDa (Fig. 6).When the primary antibody was incubated in thepresence of its own blocking peptide, all the bandsobserved disappeared (data not shown), suggesting thatthe two bands could represent different isoforms of thesame protein that are present specifically in ratendometrium. To investigate changes in the expressionof PTGIS protein during initiation of decidualization, wequantified only the 56–57 kDa band for PTGIS becausethis was the band reported by the supplier andcorresponds to the one previously reported in humanfallopian tubes (Huang et al. 2002) and uterinemyometrium (Giannoulias et al. 2002), and rat liver(Suhara et al. 2002). In our study, no significant changes(PO0.05) were detected for PTGIS protein levels at anytime studied (Fig. 6).

Semiquantitative RT-PCR and western blot analyseswere also conducted to study the expression of the PGI2receptors in the rat endometrium. These receptorsincluded PTGIR, PPARD, and RXRA. Ptgir mRNA levelswere significantly higher (P!0.05) in the stimulatedendometrium at all time points when compared with 0 h(Fig. 7). However, no significant differences were foundin the transcript levels between stimulated and non-stimulated at 8 and 32 h. The expression of Ptgir mRNAis very low in the rat endometrium compared with theother genes studied. Moreover, we were unable to detecta quantifiable band in endometrium by western blotanalysis under conditions that demonstrated thepresence of PTGIR protein in other rat tissues includingheart, lung, kidney, and spleen (data not shown).

There were no statistical differences (PO0.05) forPpard or Rxra mRNA levels at any time followingstimulation or in transcript levels when comparingstimulated and non-stimulated endometrium at 8 and32 h (Figs 8 and 9 respectively). No significant changeswere detected for PPARD protein at any time studied(Fig. 8), although there was a trend (0.05!P!0.1) forPPARD protein expression levels to increase with time.When western blot analyses were conducted usingantibodies against RXRA protein, we detected threeimmunoreactive bands: one at w 80–85 kDa, one at54 kDa and one at w 40–45 kDa (Fig. 9). There are atleast two major physiological forms of the RXRA protein

Reproduction (2009) 137 537–552


Figure 2 Expression of PTGS2 in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.

Figure 3 Expression of PTGES in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.


in human and mouse livers; a full-length form with amolecular weight of 54 kDa and a truncated form thatlacks a portion of the amino-terminal A/B domain with amolecular weight of 44 kDa (Matsushima-Nishiwakiet al. 1996). In normal human prostatic epithelial celllines and prostatic adenocarcinoma cells, there are threeforms of the proteins; the full-length form with amolecular weight of 54 kDa and two truncated formsof 47 and 44 kDa (Zhong et al. 2003). In addition, in rattestis, three RXRA isoforms were detected in germ cells,whereas only one major isoform was observed inextracts from Sertoli cells and testes (Dufour & Kim

Reproduction (2009) 137 537–552

1999). During our study, the three bands observed forRXRA disappeared when the primary antibody wasincubated in the presence of its own blocking peptide(data not shown) suggesting that the bands couldrepresent different isoforms from the same protein thatare present specifically in rat endometrium. To investi-gate changes in the expression of RXRA protein duringinitiation of decidualization, we quantified only the54 kDa band for RXRA for several reasons; this is theband reported by the supplier; it is the only isoformfound it in the human placenta (Tarrade et al. 2001);according to the literature, it is the only full-length



Figure 4 Expression of PTGES2 in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.

Figure 5 Expression of PTGES3 in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.


form; and is the most abundant form of the proteinthat is present in most tissues and cell lines studied(Matsushima-Nishiwaki et al. 1996, Dufour & Kim 1999,Zhong et al. 2003). RXRA protein levels significantlyincreased (P!0.05) at 16 h and remained high at 32 h ascompared with 0 and 2 h (Fig. 9). Moreover, PPARD andRXRA protein levels significantly increased (P!0.05) at32 h in stimulated as compared with non-stimulatedendometrium, suggesting that the increase was due to thestimulus and not to the time (Figs 8 and 9 respectively).


Protein localization of PTGS1, PTGS2, PTGES, PTGES2,PTGES3, PTGIS, PTGIR, PPARD, and RXRA in the ratendometrium

To localize the enzymes involved in PGE2 and PGI2synthesis, immunohistochemistry was conducted onsections of rat uteri at different times after a deciduogenicstimulus. Results of the immunolocalization are shownin Fig. 10. PTGS1 was localized in the nucleus as well asin the cytoplasm of the luminal epithelial cells, glandularepithelial cells, and with less intensity, in the stromal

Reproduction (2009) 137 537–552


Figure 6 Expression of PTGIS in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.

Figure 7 Expression of Ptgir in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) for one setof endometrium is shown. Histograms represent the mean opticaldensity ratios GS.E.M. of cDNAs of interest normalized to Gapdh cDNA(B) from four independent isolations.


cells at all times studied. The intensity of the signalappeared to be similar at all times.

At all times studied, PTGS2 was localized only tothe cytoplasm of the stromal cells. There were low levelsof PTGS2 staining at 0, 2, 4, and 8 h confined to theantimesometrial stromal cells. After 8 h, the signalstarted to spread throughout the stroma and waslocalized to the mesometrial as well as the antimesome-trial stromal cells. At 32 h, all stromal cells in the sectionstained positive for PTGS2. In concordance with the

Reproduction (2009) 137 537–552

results obtained in the western blot analysis, there wasincreased intensity in staining and more PTGS2 positivestromal cells at 32 h in stimulated endometrium than atthe same time in non-stimulated endometrium.

PTGES was localized to the nucleus and to thecytoplasm of the glandular epithelium, luminal epi-thelium and in the stroma at all times studied. Theintensity of the signal appeared to be similar at all times.PTGES2 and PTGES3 were localized to the nucleus aswell as in the cytoplasm of the stromal cells, glandularand luminal epithelium at all times studied. The stainingfor both proteins was very intense in the luminalepithelium and glands at 0, 2, 4, and 8 h and appearedto decrease at 16 and 32 h.

PTGIS was localized mainly to the nucleus andcytoplasm of the glandular epithelium and the stainingdid not change over time.

PPARD staining was extremely weak; however, it waspossible to identify some staining mainly in thecytoplasm of glandular and luminal epithelium and thesignal did not change at any time.

Finally, we studied the localization of RXRA in the ratendometrium. RXRA was localized mainly to thecytoplasm of the glands, luminal epithelium, and withless intensity in the stroma at all time points studied.The staining appeared intense at 8 and 16 h anddecreased at 32 h.



Figure 8 Expression of PPARD in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.

Figure 9 Expression of RXRA in rat uterine endometrium following adeciduogenic stimulus. Samples from animals at 2, 4, 8, 16, 32 hfollowing stimulation with sesame oil or non-stimulated (NS) at 0, 8and 32 h are shown. A photograph of a representative gel (A) or blot(C) for one set of endometrium is shown. Histograms represent themean optical density ratios GS.E.M. of cDNAs of interest normalized toGapdh cDNA (B) or proteins of interest normalized to ACTB protein(D) from four independent isolations.


Effects of cPGI2 and PGE2 infusion

All data were analyzed as a 2-factor mixed modelANOVA. Factor 1 was a between-animal comparison,and was the effect of the type of compound infused.Factor 2 was a within-subject comparison, being theeffect of infusion (i.e. infused horn versus non-infusedhorn). For uterine Evans blue concentrations, ANOVAindicated a highly significant interaction (Fig. 11A,P!0.001) between the effects of infusion and compoundinfused. Uterine dye concentrations did not differ(Fig. 11A, PO0.05) in non-infused horns, whereasinfusion of PGE2 resulted in markedly increased


(Fig. 11A, P!0.001) dye concentrations above those ofhorns infused with vehicle or cPGI2. The infusion ofcPGI2 resulted in uterine Evans blue concentrations notdifferent from those of the vehicle.

For uterine mass, statistical analyses of logarithmicallytransformed data revealed that there was a highlysignificant interaction (Fig. 11B, P!0.001) betweeneffects of infusion and of compound. In all groups, theinfused horns were significantly heavier (Fig. 11B,P!0.05) than non-infused horns. The infused horns ofboth the PGE2 (geometric mean of 867 mg) and cPGI2

Reproduction (2009) 137 537–552


Figure 10 Localization of PTGS1, PTGS2, PTGES,PTGES2, PTGES3, PTGIS, PPARD and RXRA proteinsin rat uteri following a deciduogenic stimulus. Uterinesections from animals stimulated with sesame oil attimes where maximal signal was detected (PTGS1, 2 h;PTGS2, 16 h; PTGES, 32 h; PTGES2, 4 h; PTGES3, 2 h;PTGIS, 4 h; PPARD, 32 h; and RXRA, 32 h) are shown.Uterine sections from non-stimulated animals (NS) at 0and 32 h are also shown. Immunohistochemistry wasperformed as described in Materials and Methods.Positive staining is shown in brown. In each slide, bar:100 micrometer. Negative controls included sub-stitution of primary antibodies with the same concen-tration of normal rabbit IgG.


(geometric mean of 551 mg) groups differed significantly(Fig. 1B, P!0.05) from the vehicle control (geometricmean of 325 mg). In addition, PGE2 infused hornswere significantly heavier (Fig. 11B, P!0.05) than thecPGI2 infused horns. The non-infused horns of eachgroup were not significantly different (Fig. 11B, PO0.05).

Effects of cicaprost, AFP-07, and PGE2 infusion

For uterine Evans blue concentrations, statisticalanalyses showed a highly significant interaction(Fig. 12A, P!0.001) between the effects of infusionand of compound. Evans blue concentrations in PGE2

infused horns were significantly greater (Fig. 12A,P!0.05) than those of the cicaprost and AFP-07 infusedhorns. In addition, concentrations of Evans blue in theAFP-07 infused horns were significantly higher (Fig. 12A,

Reproduction (2009) 137 537–552

P!0.05) than cicaprost infused horns. These resultsimply that PGE2 elicited the largest increase in vascularpermeability, followed by AFP-07, then cicaprost. Fornon-infused horns, Evans blue concentrations weresignificantly (Fig. 12A, P!0.05) higher in the AFP-07group with no differences observed between the othergroups, suggesting systemic effects of AFP-07.

For the DCR, ANOVA of the logarithmically trans-formed data revealed a highly significant interaction(Fig. 12B, P!0.001) between infusion and effect ofcompound. For all groups, the masses of infused uterinehorns were significantly (Fig. 12B, P!0.05) greater thantheir respective non-infused horns. For infused uterinehorns, those infused with PGE2 were the heaviest(geometric mean of 1010 mg) but not significantly(Fig. 12B, PO0.05) different that those infused withAFP-07 (geometric mean of 650 mg); both compounds



Figure 11 Effects of PGE2 and cPGI2 on the EVP response (A), measuredby uterine concentrations of Evans blue dye (meanGS.E.M., nZ6 to 7)and DCR (B), assessed by uterine weight (meanGS.E.M., nZ6 to 7).Geometric means for infused (grey bars) and non-infused horns(white bars) are shown for the DCR experiment (B). PGE2 infusionmarkedly increased (A, P!0.001) uterine Evans blue concentrations,while cPGI2 infused horns were not different from the vehicle. For theDCR, both cPGI2 and PGE2 infused horns were significantly (B,P!0.05) heavier than the vehicle infused horns. In addition, PGE2

horns were heavier (B, P!0.05) than cPGI2 infused horns. For bothexperiments, non-infused horns were not significantly (PO0.05)different between groups.

Figure 12 Effects of PGE2, cicaprost and AFP-07 on the EVP response(A), measured by uterine concentrations of Evans blue dye (meanGS.E.M., nZ7) and DCR (B), assessed by uterine weight (meanGS.E.M.,nZ7). Geometric means for infused (grey bars) and non-infused horns(white bars) are shown for the DCR experiment (B). Infused horns ofPGE2, cicaprost, and AFP-07 resulted in significantly (A, P!0.05)higher Evans blue concentrations than vehicle infused horns. Inaddition, Evans blue concentration in PGE2 infused horns weresignificantly greater (A, P!0.05) than both cicaprost and AFP-07infused horns; dye concentrations were higher (A, P!0.05) in AFP-07than cicaprost infused horns. Only AFP-07 non-infused horns showedan increase in Evans blue concentrations, suggesting systemic effects ofthis compound. For the DCR, infused horns of all groups weresignificantly heavier (B, P!0.05) than their respective non-infusedhorns. PGE2 infused horns were significantly heavier (B, P!0.05) thanvehicle and cicaprost, but not different (B, P!0.05) from AFP-07,infused horns. Cicaprost infusion produced uterine horns that weresimilar to the vehicle and AFP-07 infused horns. No statisticaldifferences were observed in non-infused horns between groups.


resulted in uterine horns significantly (Fig. 12B, P!0.05)heavier than the vehicle infused horns (geometric meanof 310 mg). Cicaprost infused horns (geometric mean of471 mg) were not significantly (Fig. 12B, PO0.05)different in mass from either vehicle infused or AFP-07infused horns, but were significantly (Fig. 12B, P!0.05)lighter than PGE2 infused horns. There were no statisticaldifferences (Fig. 12B, PO0.05) between treatmentgroups when comparing uterine masses of the non-infused horns.

Effects of L165041, DHA, and PGE2 infusion

For uterine Evans blue concentrations, there was a highlysignificant interaction (Fig. 13A, P!0.001) between theeffects of infusion and compound. A significant increase(Fig. 13A, P!0.05) in uterine Evans blue concentrationwas observed only in the PGE2 infused horns. Evans blueconcentrations in the infused horns of L165041, DHA,and combination groups were not different from thoseof the vehicle infused horns (Fig. 13A, PO0.05) or totheir respective non-infused horns (Fig. 13A, PO0.05).No significant changes were observed between non-infused horns.

For the DCR, ANOVA of logarithmically transformeddata revealed a highly significant interaction (Fig. 13B,


P!0.001) between the effects of infusion and ofcompound. All groups, with the exception of thecombination treatment, showed differences (Fig. 13B,P!0.05) in mass between infused and non-infusedhorns. A marked increase in uterine mass was observedin the PGE2 infused horns (geometric mean of 1324 mg)when compared with the vehicle infused horn (geo-metric mean of 235 mg) (Fig. 13B, P!0.05). Statisticalanalyses also revealed a significant interaction (Fig. 13B,P!0.005) between the effects of L165041 and DHA.This effect was apparent as the independent infusions ofL165041 (geometric mean of 359 mg) or DHA (geo-metric mean of 166 mg) resulted in a significant(Fig. 13B, P!0.05) increase or decrease, respectively,in uterine mass, while the infused horns of thecombination treatment (geometric mean of 209 mg)did not differ from the control (Fig. 13B, PO0.05). Themasses of non-infused horns in all groups except thecombination group, were significantly different(Fig. 13B, P!0.05) from the non-infused horns of thevehicle group.

Reproduction (2009) 137 537–552


Figure 13 Effects of PGE2, L165041, DHA, and combination ofL165041 and DHA on the EVP response (A), measured by uterineconcentrations of Evans blue dye (meanGS.E.M., nZ6 to 7) and DCR (B),assessed by uterine weight (meanGS.E.M., nZ5). Geometric means forinfused (grey bars) and non-infused horns (white bars) are shown for theDCR experiment (B). Only PGE2 infusion caused a significant increase(A, P!0.05) in uterine Evans blue concentrations with no otherdifferences observed between groups. For DCR, PGE2 infusion caused amarked increase in uterine mass. Although, there were effects from theinfusion of L165041 alone or DHA alone, these responses were marginalin comparison to PGE2 infusion.


Discussion

In rats, an increase in EVP is first detectable about 4 hafter a deciduogenic stimulus is given, whereas differ-entiation of stromal cells to decidual cells starts afterabout 20 h (Psychoyos 1973). During pregnancy, thesetwo processes are sequential steps that vary in timedepending upon the species. In rats, the increase in EVPoccurs late in the afternoon of day 5, whereas inhamsters and mice, it occurs in the afternoon and theevening of day 4 respectively. Previous studies in miceand hamsters have suggested a role of PTGS2, but not ofPTGS1, in implantation (Chakraborty et al. 1996, Limet al. 1999, Wang et al. 2004, Pakrasi & Jain 2008) andalso in decidualization and placentation in mice (Limet al. 1997, Wang et al. 2007, Pakrasi & Jain 2008). Thepresence of Ptgs1 transcripts in the rat endometriumprior to implantation is in agreement with the hamsterstudy (Wang et al. 2004), where it was detected as earlyas days 1 and 3, before the initiation of implantationoccurs. However, down-regulation of Ptgs1 mRNA at 4 hafter the deciduogenic stimulus in the rat endometriumappears to be different from the results reported in mice.In mice, no changes in Ptgs1 transcripts were detected atany time after a deciduogenic stimulus (Lim et al. 1997),and down-regulation of Ptgs1 transcripts was onlydetected in pregnant animals at the time of blastocystattachment (Chakraborty et al. 1996). At this time, our

Reproduction (2009) 137 537–552

study is the only one available reporting the expressionof PTGS1 protein, a better indication of function thantranscript levels, during early stages of the initiation ofdecidualization. The levels of PTGS1 protein did notchange at any time after the deciduogenic stimulus,indicating that is not regulated during the increase in EVPor the initiation of decidualization in rat endometrium.A recent report conducted in pregnant rats indicated thatPTGS1 protein is localized in the luminal epithelium atthe implantation sites on day 4 and day 5, prior toimplantation (Cong et al. 2006). On day 6 of pregnancy,after the increase in EVPand initiation of decidualizationhave already occurred, PTGS1 protein was localized tothe luminal epithelium and at a basal level in thesubluminal stroma (Cong et al. 2006). In our study,PTGS1 protein was localized to the luminal epithelialcells, glandular epithelial and the stromal cells before,during and after the increase in EVP and the initiation ofdecidualization, suggesting that it could have a role inboth of these events in the rat endometrium.

In agreement with studies in mice (Lim et al. 1997,1999), our study indicates that in the rat endometriumPtgs2 transcripts were significantly higher at 2 and 4 hfollowed by a decrease at 8 h after the deciduogenicstimulus was given. Because increased levels of PTGS2protein were not detected until 8 h after stimulation, itseems that in the rat endometrium increased expressionof PTGS2 is not required for the increase in EVP, aspreviously suggested in mice, but may be required forsubsequent decidual transformation. In hamsters, endo-metrial PTGS2 protein was detected at day 1 and themorning of day 4 of pregnancy, prior to whenimplantation occurs (Wang et al. 2004). In mice,PTGS2 protein is present in pregnant mice on day 5,after EVP and initiation of decidualization have occurred(Lim et al. 1997). Our immunolocalization studiesindicated that before and during the increase in EVP,PTGS2 protein is specifically localized in the stromalcells at the antimesometrial area where embryoattachment occurs during implantation and wheredecidual transformation is first seen. At the time whendifferentiation of stromal cells to decidual cells starts, thesignal for PTGS2 spreads throughout the endometrialstroma and is also localized in the stromal cells of themesometrial area. No signal of PTGS2 was detected inthe luminal epithelium at any time. These results agreewith those reported by Cong et al. (2006) where inpregnant rats, PTGS2 protein was localized in thestromal cells on the morning of day 5, before theinitiation of implantation, and continued to be expressedin the stroma at both implantation sites and inter-implantation sites through to day 6, after initiation ofimplantation had occurred.

There appear to be species differences in localizationof PTGS2 in the endometrium in the implantationperiod. While our results in rat showed that PTGS2protein is localized in stromal cells before and after the




initiation of decidualization, in hamsters, PTGS2 proteinis localized to the luminal epithelial cells on the morningof day 4, prior to implantation. On the afternoon of day 4and the morning of day 5, after implantation has beeninitiated in hamsters, PTGS2 is detected in epithelial andin several layers of stromal cells adjacent to theblastocysts (Wang et al. 2004). The results in rats alsodiffer from those in mice, where PTGS2 was localizedto the luminal epithelium, underlying stroma and inglandular epithelium 24 h after a deciduogenic stimuluswas given (at the time when decidualization would havestarted) and then dramatically decrease at 48 and 120 h(Pakrasi & Jain 2008). In our study, as well as in those ofWang et al. (2004) and Cong et al. (2006), there was agradual increase in the number of PTGS2 positivestromal cells as shown by immunostaining. Theseresults agree also with what is reported in pregnantmice, where PTGS2 protein gradually localizes instromal cells and the primary decidual zone surroundingthe implanting embryos after implantation has beeninitiated (Lim et al. 1997).

The constitutive expression and localization of PTGS1in the luminal epithelium at all times as well as thegradual increase in positive stromal cells for PTGS2 andthe up-regulation of the expression detected by westernblot analysis at later times suggest that PTGS1 andPTGS2 have distinct roles in the initiation of implanta-tion/decidualization in the rat endometrium. Forexample, PTGS1 may be involved in producing PGsthat increase EVP, whereas PGs produced by up-regula-tion of PTGS2 may be involved in the proliferation anddifferentiation of stromal cells to decidual cells duringthe initiation of decidualization.

Here, we also determined the expression of all thedifferent PGE synthases (PTGESs) including PTGES,PTGES2, and PTGES3. We detected a significantincrease in the expression of PTGES2 protein at 2 and4 h and high levels of PTGES3 protein at 0 and 2 h. Theseresults suggest a role for these two enzymes in theincrease of PGE2 levels to initiate the increase in EVP.

There is some evidence to suggest that PTGS1 andPTGS2 may be coupled to a particular PTGES. Inprevious in vitro studies of early stages of theinflammatory process and PGE2 production, PTGES3was found to be coupled to PTGS1 (Tanioka et al. 2000),whereas in delayed inflammatory responses, PTGES wasthe only enzyme associated with PTGS2 (Murakamiet al. 2000). Moreover, a recent study in rat endometriumduring the peri-implantation period suggested thatbecause there is a co-localization of PTGS2 and PTGESin the subluminal stromal cells and also a co-localizationof PTGS1, PTGES3, and PTGES2 in the luminalepithelium surrounding the implantating blastocyst,PGE2 could be produced through PTGS2/PTGES,PTGS1/PTGES3, and/or PTGS1/PTGES2 pathways(Cong et al. 2006). However, it is still unclear whichspecific PTGS is interacting with which PTGES in the rat


endometrium to increase PGE2 production during theinitiation of the decidualization. Our immunohisto-chemistry study shows that both PTGES and PTGS2 arepresent in stromal cells and were highly expressedduring later times after the deciduogenic stimulus wasgiven, suggesting that in the rat endometrium PGE2

could be produced through the PTGS2/PTGES pathwayat the later stages of the initiation of decidualization toregulate proliferation and differentiation of stromalcells. Because PTGES3 and PTGES2 protein levelsincrease only at early times and these enzymes appearto co-localize with PTGS1 in luminal epithelium andglandular cells at those times, we suggest that the PTGS1/PTGES3 or PTGS1/PTGES2 pathway could be involvedin producing PGE2 at early stages to regulate EVP.

It has been previously suggested that PGI2 is generatedvia the PTGS2/PTGIS pathway during implantation in themouse endometrium (Lim et al. 1999). In mice, PTGISprotein was co-localized with PTGS2 in stromal cellssurrounding the implantating embryo (Lim et al. 1999).In our study, no significant changes were found in theexpression of PTGIS transcripts and protein at any timeafter the deciduogenic stimulus. Moreover, no co-loca-lization of PTGIS with PTGS2 in the stromal cells wasfound, suggesting that the PTGS2/PTGIS pathway isunlikely to be involved in EVP or initiation ofdecidualization in the rat endometrium.

Based on the studies in Ptgs2 null mice, it has beenrecently suggested that PGI2 acts as an endogenousligand of PPARD, which then heterodimerizes withRXRA to activate decidualization (Lim et al. 1999).There were no significant changes in the levels of PPARDtranscripts or proteins at any time after the deciduogenicstimulus in the rat endometrium. These results aresupported by those obtained with Ppard knockoutmice, where it was shown that the animals havesuccessful, although delayed implantation and defectswere only in placental development and parturition(Wang et al. 2007). In our study, we did find a significantincrease for RXRA protein and a trend for increasedPPARD protein at 16 and 32 h after the deciduogenicstimulus. Our results are in agreement with Pakrasi & Jain(2008) where in mice PPARD protein expression issignificantly higher 24 h after deciduogenic stimuluswas given. However, it is important to highlight thatby that time, the increase in EVP and the firstdifferentiation of decidual cells would have occurred,suggesting that these PPARD and RXRA receptors, ifthey have a role in rodents, may only be in the laterstages of decidualization.

The intrauterine infusion of PGE2 induced deciduali-zation as indicated by an increase in EVP and uterinemass. By contrast, PTGIR, PPARD, and RXRA agonistsproduced varying results. These results together with theexpression data suggest that these three receptors do notplay an important role in the early stages of decidualiza-tion. Cicaprost and AFP-07, compounds that are highly

Reproduction (2009) 137 537–552



selective for PTGIR (Chang et al. 1997, Jones et al. 1997),elicited an EVP response, although not to the same extentas PGE2; cPGI2, which binds to both PTGIR and PPARD(Kiriyama et al. 1997), did not. cPGI2 and AFP-07induced decidualization while cicaprost infusion pro-duced uterine weights that were similar to the vehicle. Inaddition, L165041 and DHA infused separately ortogether were unable to evoke increased EVP orappreciably induce decidualization. Although therewere some effects of these synthetic ligands, they didnot produce the same magnitude of response in EVP andDCR when compared with PGE2. These findings are notconsistent with studies by Lim et al. (1997, 1999) wherecPGI2 was able to partially restore decidualization inPtgs2 null mice. It is possible that the main PG mediatingimplantation is species specific (reviewed by Kennedyet al. 2007). In addition, Lim et al. (1997) injected PGsinto the uterine lumen. However, sustained and elevateduterine levels of PGs may be necessary to produce amaximal response in decidualization (reviewed Kennedyet al. 2007). It was also shown that AFP-07 elicited agreater increased EVP response than cicaprost. Further-more, cicaprost and AFP-07 evoked an EVP responseand significantly increased uterine mass when comparedwith the control while L165041 did not. This suggeststhat PTGIR may play a greater role in decidualizationthan PPARD that is surprising for several reasons: 1) ratPtgir transcripts levels were low as compared with theother genes studied, 2) rat PTGIR protein levels were notdetectable, and 3) Ptgir null mice showed normalimplantation (Murata et al. 1997).

Previous studies used 9-cis retinoic acid (9cRA) toactivate RXRA (Lim et al. 1999) but 9cRA can pan-activate other retinoid-X-receptor subtypes (i.e. RXRB,RXRG) (Allenby et al. 1993). Since RXRB was detected inthe peri-implantation mouse uterus (Lim et al. 1999), anRXRB and PPARD heterodimer could be involved inimplantation associated events. To focus on the RXRAand PPARD heterodimer more precisely, we used DHA,which is more specific for RXRA than 9cRA (Lengqvistet al. 2004) and also suggested to bind only to the RXRAsubtype (reviewed by Germain et al. 2006). Neither theincrease in EVP nor the DCR was observed when DHAwas infused alone, indicating that RXRA may not beessential for these processes. Furthermore, a synergisticincrease in EVP or uterine mass was not observed wheninfusing both PPARD and RXRA agonists concurrently,which is not consistent with in vitro synergistic responsesreported in previous investigations (Lim et al. 1999).Studies have shown that activating PPARD initiates cellproliferation in keratinocytes (Romanowska et al. 2008)while RXRA activation induces a caspase-dependentapoptotic pathway (Lee et al. 2005). This may explainthe absence of a synergistic response when L165041 andDHA were infused together as countervailingmechanisms may have been activated. It is also possiblethat an RXR homodimer (Germain et al. 2006) pathway

Reproduction (2009) 137 537–552

was activated which induced signaling events indepen-dent of the PPARD/RXRA system. It cannot be ruled outthat the simultaneous presence of L165041 and DHAinterferes with the ability of these ligands to bind to theirtargets. Interestingly, non-infused horns of L165041 andDHA groups showed a significant, but marginal increaseand decrease respectively, in uterine mass, whencompared with the vehicle group suggesting systemiceffects of these compounds.

In conclusion, the present study reports the expressionof the enzymes involved in the PGE2 and PGI2biosynthetic pathways at the levels of mRNA andproteins at different stages during the initiation ofdecidualization in rat endometrium. Up-regulation ofPTGES2 and PTGES3 proteins and co-localization withPTGS1 protein in luminal epithelium and glands wasobserved at the time when EVP is increased anddecidualization is initiated, suggesting that this is themost likely pathway by which PGE2 production isincreased in response to a deciduogenic stimulus.Since PTGS2 protein is present at 0 h, it could contributeto PGE2 production early in the response to adeciduogenic stimulus. Differential cellular sites of PGproduction by PTGS1 and PTGS2 and PTGESs mayaccount for distinct roles of these isoforms in variouscellular functions and in differences between species.Since stable agonists of PTGIR were unable to eliciteither an increase in EVP or decidualization to the sameextent as PGE2 and because we were unable to detect theprotein in the rat endometrium, it seems unlikely thatPTGIR is the main pathway mediating implantation anddecidualization. Furthermore, the selective agonists forPPARD and RXRA were also unable to elicit an increasein EVP or appreciably induce decidualization. Inaddition, stable analogues of PGI2 that are agonists forPPARD were less effective than PGE2. This stronglysuggests that PGI2 acting on the previously proposedPPARD/RXRA system may not be the molecular focalpoint in the processes of decidualization and implan-tation in the rat. Furthermore, it should be noted thatPGI2 itself has never been shown to be a ligand forPPARD, only that some stable analogues of PGI2 areligands. Ppard and Rxra may ultimately mediate adifferent aspect of pregnancy such as placental develop-ment, as indicated in Ppard and Rxra deficient micestudies (Sapin et al. 1997, Barak et al. 2002, Wang et al.2007). Overall, the data suggest that PGE2 is the primaryPG involved in decidualization in the rat.

Materials and Methods

Reagents

Unless otherwise indicated, all general chemicals includingprogesterone, oestradiol and sesame oil were purchased fromSigma. First-strand cDNA synthesis kit and all reagents forwestern blot analyses were purchased from GE Healthcare




Bio-Sciences Inc. (Baie d’Urfe, QC, Canada). DNase I andPlatinum Taq DNA polymerase were purchased from Invitro-gen. Secondary antibodies were purchased from Sigma.Primary antibodies, unless otherwise specified, were pur-chased from Cayman Chemical Company (Ann Arbor, MI,USA). Normal rabbit IgG and anti-PTGS1 primary antibodieswere purchased from Upstate Biotechnology (Lake Placid, NY,USA). Anti-RXRA primary antibody was purchased from SantaCruz Biotechnology Inc. (Santa Cruz, CA, USA). Anti-ACTB(b-actin) primary antibody was purchased from ChemiconInternational (Temecula, CA, USA). The same primary andsecondary antibodies were used for the western blot andimmunohistochemistry studies.

Animal and tissue preparation

Female Sprague–Dawley rats (200–225 g body mass; CharlesRiver, St Constant, QC, Canada) were housed in temperature-and light-controlled conditions (lights on from 05:00 to19:00 h) with free access to food and water. Animals wereovariectomized under isoflurane anesthesia (Abbot Labora-tories, Saint-Laurent, QC, Canada) and allowed at least 5 daysto recover. Ovariectomized rats were treated with oestradioland progesterone s.c. to mimic pseudopregnancy as previouslydescribed (Kennedy & Ross 1997). To induce decidualization,animals received a bilateral injection of 0.1 ml sesame oil intothe uterine lumen around noon on the equivalent of day 5 ofpseudopregnancy and were referred to as stimulated asopposed to non-stimulated (NS). Rats were killed bydecapitation at 2, 4, 8, 16, 32 h after the deciduogenicstimulus was given or at 0, 8, and 32 h for the non-stimulatedanimals (NS). The non-stimulated animals killed at the lattertwo time points were included in order to determine if anychanges observed in the endometrium of stimulated animalswere a consequence of the deciduogenic stimulus or of theprogression of pseudopregnancy. Endometrium was separatedfrom myometrium and placed in GTC solution (4 M guanidi-nium thiocyanate, 0.5% sarkosyl, 25 mM sodium citrate, 0.1%v/v antifoam A, 5% b-mercaptoethanol) or RIPA buffer (50 mMTris–HCl, 150 mM NaCl, 1% Triton X-100, 1% sodiumdeoxycholate, 0.1% SDS, 100 mM Na3VO3, and COMPLETE,Mini, EDTA-free protease inhibitors cocktail (Roche Diagnos-tics)) for subsequent RNA and protein isolation respectively.Endometrium from three animals for each treatment group waspooled and four sets were used for each treatment. Portions ofuteri were fixed in 4% paraformaldehyde prior to immunohis-tochemistry studies.

All procedures involving animals were performed inaccordance with the guidelines of the Canadian Council onAnimal Care and The University Council on Animal Care at theUniversity of Western Ontario.

Semiquantitative RT-PCR

Semiquantitative RT-PCR (RT-PCR) was performed on Ptgs1,Ptgs2, Ptges, Ptges2, Ptges3, Ptgis, Ptgir, Ppard, and Rxratranscripts to detect differences in mRNA levels following theapplication of the deciduogenic stimulus. First-strand cDNAsyntheses were performed on DNase-treated total RNA from


rat endometrium by reverse transcription using a First-strandcDNA Synthesis Kit. Samples were then subjected to PCR in aGene Amp PCR System 2400 (PerkinElmer, Woodbridge, ON,Canada) using Platinum Taq DNA polymerase. The oligonu-cleotide primers were designed using Primer 3 program(Whitehead Institute for Biomedical Research, Cambridge,MA, USA). Downstream and upstream primers, GenBankaccession for cDNA sequence and amplification conditionsfor each gene studied are shown in Table 1. The cycle numberfor each gene studied was determined to be in the linear rangefor each primer set in preliminary experiments. A watercontrol and no RT product were carried through all reactions.All PCR products were confirmed by sequencing. PCRsamples were electrophoretically separated on 1.5% agarosegels. Bands were quantified by Image Master VDS software.Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) wasused as the internal control. Optical density units of thecDNA bands of all genes studied were normalized with thecorresponding Gapdh cDNA. Downstream and upstreamprimers, GenBank accession for cDNA sequence andamplification conditions for these two internal controls arealso shown in Table 1.

Western blot analyses

To determine changes in endometrial protein levels followingendometrial stimulation, western blot analyseswere conducted.Endometrial extracts were prepared by homogenizing tissue inseven volumes of ice-cold RIPA buffer using 10 strokes of aPotter–Elvehjem homogenizer. Extracts were incubated on icefor 30 min and then centrifuged at 15 700 g for 15 min at 4 8C topellet cellular debris. Supernatants were retained and stored atK70 8C. Protein concentrations were determined using Bio-Raddye reagent (BioRad Laboratories). Equivalent amounts ofprotein (40 or 80 mg) were separated in 12 or 15% SDS-PAGEgels and electrotransferred to 0.2 or 0.45 micrometer nitro-cellulose membranes (GE Healthcare Bio-Sciences Inc). Afterblocking in 5% non-fat milk, Tris-buffered saline containing 1%Tween-20, membranes were incubated with rabbit polyclonalantibodies against PTGS1 (1 mg/ml; cat. no. 06-970; Millipore,Billerica, MA, USA), PTGS2 (4 mg/ml; cat. no. 160107; CaymanChemical Company), PTGES (4 mg/ml; cat. no. 160140; CaymanChemical Company), PTGES2 (4 mg/ml; cat. no. 160145;Cayman Chemical Company), PTGES3 (1.34 mg/ml; cat. no.160150; Cayman Chemical Company), PTGIS (2.5 mg/ml; cat.no. 100023; Cayman Chemical Company), PPARD (8 mg/ml;cat. no. 101720; Cayman Chemical Company), or RXRA(2 mg/ml; cat. no. 55350; Santa Cruz Biotechnology) overnightat 4 8C. Secondary antibodies included HRP-goat anti-rabbitIgG (1:10 000; cat. no. A6154; Sigma) or HRP-rabbit anti-mouseIgG (1:10 000; cat. no. A9044; Sigma). Immunoreactive bandswere detected by enhanced chemiluminescence (GE Health-care Bio-Sciences Inc). As a negative control normal rabbit IgG(cat. no. 12-370; Millipore) was used instead of the primaryantibody. To assess protein loading, membranes were incubatedwith mouse monoclonal ACTB (1:10000; cat. no. mab1501;Millipore) antibody. Protein products of all the genes studiedwere normalized using ACTB.

Reproduction (2009) 137 537–552


Table 1 Primers, GenBank accession numbers, and amplification conditions for each gene studied and each internal control.

Target genes Primer sequence (5 0/3 0) Sense PositionGenBankaccession no. Amplification conditions

Ptgs1 TGCATGTGGCTGTGGATGTCATCAA US 1392–1416 S67721 94 8C-45 sCACTAAGACAGACCCGTCATCTCCA DS 1816–1840 63 8C-1 min 35 Cycles

72 8C-1 minPtgs2 AACCCACCCCAAACACAG US 337–354 NM_000963 94 8C-1 min

CTGGCCCTCGCTTATGATCT DS 728–747 60 8C-1.5 min 33 Cycles72 8C-1 min

Ptges ATCAAGATGTACGCGGTGGCT US 115–135 AF280967 94 8C-45 sCACTTCCCAGAGGATCTGTA DS 467–486 63 8C-1 min 35 Cycles

72 8C-1 minPtges2 GAAGGACCGAGATTAAATTC US 421–440 XM_001078154 94 8C-45 s

GCCTTCATGGGTGGGTAATA DS 575–594 57 8C-1 min 35 Cycles72 8C-1 min

Ptges3 ACCATGCAGCCTGCTTCTGC US 328–347 AY281130 94 8C-45 sCATGACTGGCCGGATTCTCC DS 568–587 63 8C-1 min 33 Cycles

72 8C-1 minPtgis CTGGACCCACACTCTTAC US 241–258 U53855 94 8C-1 min

CACTCCATACAGGGTCAG DS 541–558 60 8C-1.5 min 35 Cycles72 8C-1 min

Ptgir CTGGAGAAGACGGAAACAAA US 1–20 D28966 94 8C-1 minTGTCACACAGCATCGTCCCA DS 430–449 60 8C-1.5 min 35 Cycles

72 8C-1 minPpard CGGGAAGAGGAGAAAGAGG US 282–300 U75918 94 8C-45 s

AGCGGATAGCGTTGTGG DS 662–678 58 8C-1 min 33 Cycles72 8C-30 s

Rxra GGACACCAAACATTTCCTGCC US 3–23 AM392401 94 8C-45 sGATGTGCTTGGTGAAGGA DS 400–417 58 8C-1 min 35 Cycles

72 8C-30 sGapdh ATGGGAAGCTGGTCATCAAC US 261–280 NM_017008 95 8C-1 min

GGATGCAGGGATGATGTTCT DS 681–700 58 8C-1.2 min 35 Cycles72 8C-1 min

US and DS indicate upstream and dowstream primers respectively. During the PCR protocol, initial denaturation was at 94 8C for 5 min and the finalextension was at 72 8C for 30 min for all genes studied. The size of the predicted products were 449 bp for Ptgs1, 411 bp for Ptgs2, 372 bp for Ptges,174 bp for Ptges2, 260 bp for Ptges3, 318 bp for Ptgis, 449 bp for Ptgir, 397 bp for Ppard, 415 bp for Rxra, and 440 bp for Gapdh cDNAs.


Immunohistochemistry

After fixation in 4% paraformaldehyde, the uterine horns wereembedded in paraffin and w 6 micrometer serial sectionswere prepared and placed on SuperfrostPlus microscopeslides (VWR International Ltd., Mississauga, ON, Canada).Sections were deparaffinized in xylene, rehydrated through aseries of ethanol washes and rinsed in water. Endogenousperoxidase activity was blocked by incubating sections in0.3% H2O2 in methanol for 40 min at room temperature.Slides were blocked for 1 h in PBS supplemented with 10%normal goat serum. Localization of PTGS1, PTGS2, PTGES,PTGES2, PTGES3, PTGIS, PPARD, and RXRA proteins wasperformed by incubating sections of rat uteri with either rabbitpolyclonal antibodies against PTGS1 (3.3 mg/ml), PTGS2(6.7 mg/ml), PTGES (16 mg/ml), PTGES2 (5 mg/ml), PTGES3(2 mg/ml), PTGIS (7.14 mg/ml), PPARD (20 mg/ml) or RXRA(5 mg/ml) overnight at 4 8C. Negative controls includedsubstitution of the primary antibodies with the sameconcentration of normal rabbit IgG. Sections were incubatedwith 1:3000 HRP-conjugated goat anti-rabbit IgG in 10% goatserum for 1 h at room temperature. To visualize boundantibody, the 3,3 0-diaminobenzidine tetrahydrochloride-pluskit (Zymed Laboratories Inc., San Francisco, CA, USA) wasutilized according to the manufacturer’s instructions. Sectionswere then briefly counterstained (10 s) with hematoxylin

Reproduction (2009) 137 537–552

solution (Gill no. 3, Sigma), and examined using a Nikonmicroscope. Images were captured using a Nikon microscope,camera and ACT-1 software (Nikon Inc., Mississauga, ON,Canada). Final images were generated using Photoshop 5.5.For each gene studied, the immunohistochemical staining wasrepeated twice at each time point with sections obtained fromfour different rats.

Intrauterine infusions

On the equivalent of day 5 of pseudopregnancy, the day ofmaximal uterine sensitization for the DCR (Kennedy & Ross1997) the rats were randomized to the appropriate number oftreatment groups for that experiment to receive unilateralintrauterine infusions of PGE2 (Cayman Chemical Company) orvarious agonists. Because of differing solubility properties of thecompounds, it was not possible to compare the effects of all ofthem in a single experiment. All compounds were infused at therate of 1.42 nmol/h using Alzet osmotic mini-pumps, model2001 (pumping rate 1 ml/h) (DURECT Corporation, Cupertino,CA, USA). cPGI2 was obtained from Cayman ChemicalCompany cicaprost was a generous gift of Dr Fiona McDonald,Schering AG, Berlin, Germany; AFP-07 was generouslyprovided by Dr Yasushi Matsamura, (Asahi Glass Co., Ltd.,Yokohama, Japan); L165041 was obtained from Calbiochem,




San Diego, CA; and DHA (cis-4, 7, 10, 13, 16, 19-docosa-hexanoic acid) from Sigma-Aldrich. cPGI2, cicaprost, and AFP-07 were infused in 1.5% ethanol, 3 mg/ml b-cyclodextran(Schering), 10 mM Tris (Sigma-Aldrich), pH 8.3. L165041 andDHA were infused in 50% (v/v) dimethyl sulphoxide (CaledonLaboratories Ltd, Georgetown, ON, USA), 0.5 mM indometha-cin (Sigma-Aldrich) in PBS, pH 7.4. PGE2 was infused in thesame vehicle as the compounds to which it was beingcompared. Preparation of the pumps and the technique forinserting them have been described previously (Kennedy &Lukash 1982). In summary, one uterine horn of each animal wasinfused from the uterotubal end towards the cervical end. Thecontra-lateral non-infused horn served as a control thatprovided an opportunity to detect possible systemic effects ofthe infused compounds. Two to 3 h prior to pump insertion, theanimals received 2 mg indomethacin in 0.4 ml 20% ethanol insesame oil s.c. to inhibit PG synthesis in response to theunavoidable trauma of pump insertion. This combination of s.c.and intrauterine administration of indomethacin has beenshown to reduce endogenous PG production during decid-ualization (Kennedy 1985).

Endometrial vascular permeability

Changes in EVP were assessed 10 h after pump insertion byusing Evans blue (GURR, Searle Diagnostics, High Wycombe,England), a macromolecular dye which, in the circulation, bindsavidly to albumin (Psychoyos 1971). Nine hours and 45 minafter pump insertion, the rats were given Evans blue (60 mg/kg)via a lateral tail vein under isoflurane anesthesia. The animalswere decapitated 15 min later and uterine Evans blue concen-trations in infused and non-infused horns were determined byspectrophotometry as described by Udaka et al. (1970).

Decidualization

The extent of uterine decidualization was evaluated 5 daysafter the insertion of the pumps by separately weighing theinfused and non-infused horns (Finn & Keen 1963, Yochim &De Feo 1963).

Statistical analyses

Expression of all the RT-PCR products was normalized toGapdh and the expression of all the protein products to ACTB.Data from RT-PCR studies and western blot analyses weresubjected to ANOVA to determine statistical significance oftreatments. The effects of time after stimulation weredetermined by one-way ANOVA, and the effects of stimulationby two-way ANOVA. Differences were considered statisticallysignificant when P!0.05. ANOVAs were also performed todetermine statistical significance of treatment effects on EVPand decidualization, with the variance being partitioned on abetween- and within-animal basis where appropriate. Whereappropriate, data were logarithmically transformed prior toANOVA to remove or substantially reduce heterogeneity ofvariance, as indicated by Bartlett’s test (Snedecor 1956). Inaddition, Duncan’s tests were performed for between-animalgroup comparisons, and paired t-tests for within-animal


comparisons when significant interactions were observed.Practical Statistics Version 3.0 (Canadian Technology 1990)program was used for all the analyses.

Declaration of interest

The authors declare that there is no conflict of interest thatcould be perceived as prejudicing the impartiality of theresearch reported.

Funding

This project was supported by CIHR grant MT-10414. Thisstudy was also funded by the Canadian Institute of HealthResearch (CIHR) Master’s Trainee Award to Sen Han Phang.

Acknowledgements

We thank Elizabeth Ross for her valuable help with the surgery,hormone injections and tissue collections and Gerald Barbe forhis technical assistance in the design of the primers for theRT-PCR technique. The authors also thank Elizabeth Ross forher valuable assistance in preparing the manuscript.

References

Allenby G, Bocquel MT, Saunders M, Kazmer S, Speck J, Rosenberger M,Lovey A, Kastner P, Grippo JF, Chambon P & Levin AA 1993 Retinoicacid receptors and retinoid X receptors: interactions with endogenousretinoic acids. PNAS 90 30–34.

Barak Y, Liao D, He W, Ong ES, Nelson MC, Olefsky JM, Boland R &Evans RM 2002 Effects of peroxisome proliferator-activated receptor d onplacentation, adiposity, and colorectal cancer. PNAS 99 303–308.

Chakraborty I, Das SK, Wang J & Dey SK 1996 Developmental expressionof the cyclo-oxygenase-1 and cyclo-oxygenase-2 genes in the peri-implantation mouse uterus and their differential regulation by theblastocyst and ovarian steroids. Journal of Molecular Endocrinology 16107–112.

Chang CS, Negishi M, Nakano T, Morizawa Y, Matsumura Y & Ichikawa A1997 7,7-Difluoroprostacyclin derivative, AFP-07, a highly selective andpotent agonist for the prostacyclin receptor. Prostaglandins 53 83–90.

Cong J, Diao HL, Zhao YC, Ni H, Yan YQ & Yang ZM 2006 Differentialexpression and regulation of cylooxygenases, prostaglandin E synthasesand prostacyclin synthase in rat uterus during the peri-implantationperiod. Reproduction 131 139 –151.

Dufour JM & Kim KH 1999 Cellular and subcellular localization of sixretinoid receptors in rat testis during postnatal development: identifi-cation of potential heterodimeric receptors. Biology of Reproduction 611300–1308.

Finn CA & Keen PM 1963 The induction of deciduomata in the rat.Journal of Embryology and Experimental Morphology 11 673–682.

FormanBM,Chen J&EvansRM1997 Hypolipidemic drugs, polyunsaturatedfatty acids, and eicosanoids are ligands for peroxisome proliferator-activated receptors a and d. PNAS 94 4312–4317.

Germain P, Chambon P, Eichele G, Evans RM, Lazar MA, Leid M,De Lera AR, LotanR,MangelsdorfDJ&GronemeyerH2006 Internationalunion of pharmacology, LXIII. retinoid X receptors. PharmacologicalReviews 58 760–772.

Giannoulias D, Alfaidy N, Holloway AC, Gibb W, Sun M, Lye SJ &Challis JRG 2002 Expression of prostaglandin I2 synthase, but notprostaglandin E synthase, changes in myometrium of women at termpregnancy. Journal of Clinical Endocrinology and Metabolism 875274–5282.

Reproduction (2009) 137 537–552



Helliwell RJ, Berry EBE, O’Caroll SJ & Mitchell MD 2004 Nuclearprostaglandin receptors: role in pregnancy and parturition? Prostaglandins,Leukotrienes and Essential Fatty Acids 70 149–165.

Huang JC, Arbab F, Tumbusch KJ, Goldsby JS, Matijevic-Aleksic N &Wu KK2002 Human fallopian tubes express prostacyclin (PGI ) synthase andcyclooxygenases and synthesize abundant PGI. Journal of ClinicalEndocrinology and Metabolism 87 4361–4368.

Jones RL, Quian YM, Wise H, Wong HNC, LamWL, Chan HW, Yim APC &Ho JKS 1997 Relaxant actions of nonprostanoid prostacyclin mimeticson human pulmonary artery. Journal of Cardiovascular Pharmacology 29525–535.

Kennedy TG 1985 Evidence for the involvement of prostaglandinsthroughout the decidual cell reaction in the rat. Biology of Reproduction33 140–146.

Kennedy TG & Lukash LA 1982 Induction of decidualization in rats by theintrauterine infusion of prostaglandins. Biology of Reproduction 27253–260.

Kennedy TG & Ross HE 1997 Temporal- and hormone-dependent changesin uterine sensitization for the decidual cell reaction and decidualizationin vitro of rat endometrial stromal cells. Journal of Reproduction andFertility 109 129–136.

Kennedy TG, Gillio-Meina C & Phang SH 2007 Prostaglandins and theinitiation of blastocyst implantation and decidualization. Reproduction134 635–643.

Kiriyama M, Ushikubi F, Kobayashi T, Hirata M, Sugimoto Y & Narumiya S1997 Ligand binding specificities of the eight types and subtypes of themouse prostanoid receptors expressed in Chinese hamster ovary cells.British Journal of Pharmacology 112 217–224.

Lee KW, Ma L, Yan X, Liu B, Zhang XK & Cohen P 2005 Rapid apoptosisinduction by IGFBP-3 involves an insulin-like growth factor-independentnucleomitochondrial translocation of RXRa/Nur77. Journal of BiologicalChemistry 280 16942–16948.

Lengqvist J, de Urquiza AM, Bergman AC, Willson TM, Sjovall J, Perlman T& Griffiths WJ 2004 Polyunsaturated fatty acids including docosahex-anoic and arachadonic acid bind to the retionoid x receptor a ligand-binding domain. Molecular and Cellular Proteomics 3 692–703.

Lim H & Dey SK 2002 A novel pathway of prostacyclin signaling-hangingout with nuclear receptors. Endocrinology 143 3207–3210.

Lim H, Paria BC, Das SK, Dinchuk JE, Langenbach R, Trzaskos JM & Dey SK1997 Multiple female reproductive failures in cyclooxygenase2-deficient mice. Cell 91 197–208.

Lim H, Gupta RA, Ma WG, Paria BC, Moller DE, Morrow JD, DuBois RN,Trzaskos JM & Dey SK 1999 Cyclo-oxygenase-2-derived prostacyclinmediates embryo implantation in the mouse via PPARd. Genes andDevelopment 13 1561–1574.

Matsushima-Nishiwaki R, Shidoji Y, Nishiwaki S, Yamada T, Moriwaki H &Muto Y 1996 Aberrant metabolism of retinoid X receptor proteins inhuman hepatocellular carcinoma. Molecular and Cellular Endocrinology121 179–190.

Murakami M, Naraba H, Tanioka T, Semmyo N, Nakatani Y, Kojima F,Ikeda T, Fueki M, Ueno A, Oh-ishi S & Kudo I 2000 Regulation ofprostaglandin E2 biosynthesis by inducible membrane-associatedprostaglandin E2 synthase that acts in concert with cyclooxygenase-2.Journal of Biological Chemistry 275 32783–32792.

Murata T, Ushikubi F, Matsuoka T, Hirata M, Yamasaki A, Sugimoto Y,Ichikawa A, Aze Y, Tanaka T, Yoshida N, et al. 1997 Altered painperception and inflammatory response in mice lacking prostacyclinreceptor. Nature 388 678–682.

Pakrasi PL & Jain AK 2008 Cyclooxygenase-2 derived PGE2 and PGI2 playan important role via EP2 and PPARd receptors in early steps of oilinduced decidualization in mice. Placenta 29 523–530.

Papay KD& Kennedy TG 2000 Characterization of temporal and cell-specificchanges in transcripts for prostaglandin E2 receptors in pseudopregnantrat endometrium. Biology of Reproduction 62 1515–1525.

Reproduction (2009) 137 537–552

Psychoyos A 1971 Methods for studying changes in capillary permea-bility of the rat endometrium. In Methods in Mammalian Embryology,pp 334–338. Ed. JC Daniel. San Francisco: W. H. Freeman andCompany.

Psychoyos A 1973 In Handbook of Physiology, pp 187–215. Eds RO Greep,EB Astwood & SR Geiger. Bethesda: American Physiological Society(Section 7, part 2).

Romanowska M, al Yacoub N, Seidel H, Donandt S, Gerken H, Phillip S,Haritonova N, Artuc M, Schweiger S, Sterry W, et al. 2008 PPARdenhances keratinocyte proliferation in psoriasis and induces heparin-binding EGF-like growth factor. Journal of Investigative Dermatology 128110–124.

Sapin V, Dolle P, Hindelang C, Kastner P & Chambon P 1997 Defects of thechorioallontoic placenta in mouse RXRa null fetuses. DevelopmentalBiology 191 29–41.

Seimandi M, Lemaire G, Pillon A, Perrin A, Carlavan I, Voegel JJ, Vignon F,Nicolas JC & Balaguer P 2005 Differential responses of PPARa, PPARd,and PPARg reporter cell lines to selective PPAR synthetic ligands.Analytical Biochemistry 344 8–15.

Snedecor GW 1956 Statistical Methods, 5th edn. Ames, Iowa: The IowaState University Press.

Suhara H, Sawa Y, Fukushima N, Kagisaki K, Yokoyama C, Tanabe T,Ohtake S & Matsuda H 2002 Gene transfer of human prostacyclinsynthase into the liver is effective for the treatment of pulmonaryhypertension in rats. Journal of Thoracic and Cardiovascular Surgery 123855–861.

Tanioka T, Nakatani Y, Semmyo N, Murakami M & Kudo I 2000 Molecularidentification of cytosolic prostaglandin E2 synthase that is functionallycoupled with cyclooxygenase-1 in immediate prostaglandin E2 biosyn-thesis. Journal of Biological Chemistry 275 32775–32782.

Tarrade A, Schoonjans K, Pavan L, Auwerx J, Rochette-Egly C, Evain-Brion D & Fournier T 2001 PPARg/RXRa heterodimers control humantrophoblast invasion. Journal of Clinical Endocrinology and Metabolism86 5017–5024.

Udaka K, Takeuchi Y & Movat HZ 1970 Simple method for quantitation ofenhanced vascular permeability. Proceedings of the Society forExperimental Biology and Medicine 133 1384–1387.

Wang H & Dey SK 2005 Lipid signaling in embryo implantation.Prostaglandins & Other Lipid Mediators 77 84–102.

Wang X, Su Y, Deb K, Raposo M, Morrow JD, Reese J & Paria BC 2004Prostaglandin E2 is a product of induced prostaglandin-endoperoxidesynthase 2 and microsomal-type prostaglandin E synthase at theimplantation site of the hamster. Journal of Biological Chemistry 27930579–30587.

Wang H, Xie H, Sun X, Tranguch S, Zhang H, Jia X, Wang D, Das SK,Desvergne B, Wahli W, et al. 2007 Stage-specific integration of maternaland embryonic peroxisome proliferator-activated receptor d signaling iscritical to pregnancy success. Journal of Biological Chemistry 28237770–37782.

Yochim JM & De Feo VJ 1963 Hormonal control of the onset, magnitudeand duration of uterine sensitivity in the rat by steroid hormones of theovary. Endocrinology 72 317–326.

Zhong C, Yang S, Huang J, Cohen MB & Roy-Burman P 2003 Aberration inthe expression of the retinoid receptor, RXRa, in prostate cancer. CancerBiology & Therapy 2 179–184.

Received 8 July 2008

First decision 19 August 2008

Revised manuscript received 24 November 2008

Accepted 4 December 2008



reproduction...ppard without signiﬁcant binding to ptgir (seimandi et al. 2005) was used. since...

Documents