Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Prostanoids are potent autocoids that mediate numerous activities in the mammalian body such as smooth muscle contraction/relaxation, induction of pain, lipolysis and platelet aggregation (Andersen and Ramwell, 1974), intraocular pressure control (Wang et al., 1990), iris sphincter contraction (Goh and Kishino, 1994), immunoregulation and luteolysis (see Coleman et al., 1994 for review). These diverse physiological and pharmacological effects of endogenous and synthetic prostaglandins are transduced by cell surface receptors which exhibit different degrees of selectivity for the natural PGs, PGD₂, PGE₂, PGF₂, PGI₂ (prostacyclin), and TXA₂ (Coleman et al., 1994). PG receptor nomenclature and classification, as determined by pharmacological (i.e., using selective agonists/antagonists) and molecular cloning techniques, defines the following major receptor subtypes present in the mammalian body: DP, EP (with further subtypes EP₁^, EP₂^, EP₃^, EP₄), FP, IP and TP (Coleman et al., 1994). Alternative genomic splicing results in further subtypes of the EP₃ receptor namely, EP_3A, EP_3B, EP_3C and EP_3D (Coleman et al., 1994).

PG receptor-effector coupling can be summarized as follows: FP, TP and EP₁ receptors preferentially couple to G_q/G_q/11 and their activation results in the formation of inositol trisphosphate and diacylglycerol and mobilization of intracellular Ca⁺⁺ (Abramovitz et al., 1994); the DP, EP₂, EP₄ and IP receptors preferentially couple to G_s and activation of these receptors activates adenylyl cyclase to produce intracellular cAMP (Sugimoto et al., 1994; Coleman et al., 1994). Several subtypes of EP receptors have been identified, including EP₁, EP₂, EP₃ and EP₄, which couple to various G-proteins (see Negishi et al., 1993; Coleman et al., 1994 for reviews). Furthermore, numerous splice variants of the EP₃ receptor have been described which couple to a multitude of G-proteins and second messenger systems (Coleman et al., 1994; Kotani et al., 1995).

The corpus luteum is formed from the Graffian follicle after ovulation and is composed of small and large luteal cells (and numerous nonluteal cells) which secrete various hormones during and up till mid-cycle, after which the whole structure regresses unless pregnancy is successful (Niswender and Nett, 1988). Preliminary studies reported some years ago demonstrated the presence of binding sites for [³H]PGE₁ (Kimball and Lauderdale, 1975; Rao, 1976), [³H]PGE₂ (Rao, 1976) and [³H]PGF₂ (Rao, 1976; Powell et al., 1976) in homogenates and/or purified membranes of BCLM. Additionally, [³H]PGF₂ binding sites were detected in human (Rao et al., 1977), rat (Wright et al., 1979) and ovine (Balapure et al., 1989) corpus luteum. However, the detailed pharmacological analysis of these binding sites was hampered and inconclusive at the time because of the scarcity at the time of PGs with sufficiently high affinity and selectivity. With the recent availability of a whole range of potent and selective PG agonists and antagonists, we decided to determine the pharmacological specificity of [³H]PGE₁, [³H]PGE₂ and [³H]PGF₂ binding in BCLM homogenates to classify the PG receptor-subtypes present in this tissue and to correlate the receptor binding affinities of the FP and EP₃ ligands used in the binding studies with their functional potencies and efficacies in a number of functional assays (our studies and those from the literature). We have also employed emulsion-coated film autoradiographic techniques to visualize the binding sites for these radioligands in thin sections of BCLM.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

[³H]PGE₁ and [³H]PGE₂ binding assays. Total particulate BCLM homogenates were prepared by standard homogenization (tissue disruptor setting 5 for 4 min; 15 g/ml Krebs buffer, pH 7.4) and centrifugation (30,000 × g for 20 min/4°C) procedures. The supernatants were discarded and the tissue pellets washed by two resuspension/centrifugation steps as above. The final washed tissue homogenates (total particulate preparation) (20-60 mg/ml) were incubated with [³H]PGE₁ or [³H]PGE₂ (both at 0.9-2 nM final concentration) in Krebs buffer (pH 7.4) in a total volume of 0.5 ml for 1 hr at 23°C. The nonspecific binding was defined with 1 µM unlabeled PGE₂. The assays were terminated by rapid vacuum filtration (using Whatman GF/B glass fiber filter previously soaked in 0.3% polyethylenimine) and the receptor-bound radioactivity determined by liquid scintillation spectrometry at 50% efficiency.

[³H]PGF₂ binding assays. Washed BCLM total particulate homogenates (20 mg/ml in Krebs buffer; pH 7.4; see above) were prepared as described above and were incubated with [³H]PGF₂ (0.9-1.5 nM final) and increasing concentrations (in duplicate) of the test compound for 2 hr at 23°C in a total volume of 0.5 ml as previously described (Powell et al., 1976). The nonspecific binding was defined with 1 to 10 µM unlabeled PGF₂ or cloprostenol. The assays were terminated by rapid vacuum filtration, using Whatman GF/B glass fiber filters previously soaked in 0.3% polyethylenimine, and the receptor-bound radioactivity determined by liquid scintillation spectrometry at 50% efficiency.

EP₃ functional assays. Receptor selection and amplification technology (R-SAT) kits containing the human EP₃ receptor transfected in mammalian cells (cotransfected with genes for -galactosidase) were used to determine the relative potencies and efficacies of various PGs to stimulate cell proliferation. The agonist-induced responses were quantified using a colorimetric assay of -galactosidase in a 96-well format as recommended by the manufacturer (Receptor Technologies Inc., Wanooski, VT). Potent ligands of the expressed receptor on these cells stimulate cell proliferation in a concentration-dependent manner. After a 4-day incubation of cells with varying concentrations of the ligand, the -galactosidase activity (as a measure of relative cell number) was assayed by adding a substrate of the enzyme (o-nitrophenyl--D-galactopyranoside) that is hydrolyzed to a 405-nm absorbing product. The 96-well assay plates were routinely read on a microplate reader at two time points (6 and 24 hr) after adding the enzyme substrate. The signal (absorbance) increased with time but the signal-to-noise ratio was typically no more than 2 at both time points and equivalent results were obtained at 6- and 24-hr absorbance determinations.

FP phosphoinositide turnover assays. [³H]-IPs produced by agonist-mediated activation of phospholipase C were quantified by previously published procedures (Sharif et al., 1996; Griffin et al., 1997). In brief, confluent Swiss 3T3 cells were exposed (triplicate determinations) to 1.0 to 1.5 µCi of [³H]-myo-inositol in 0.5 ml DMEM for 24 to 30 hr at 37°C. Then cells were rinsed once with DMEM/F-12 containing 10 mM LiCl, and the agonist stimulation experiment was performed in 0.5 ml of the same medium to facilitate accumulation of [³H]-IPs (Berridge et al., 1982). Cells were exposed to the agonist or solvent for 60 min at 37°C, followed by aspiration of the medium and immediate addition of 1 ml of ice-cold 0.1 M formic acid. The cell lysates (0.9 ml) were loaded on columns containing 1 ml AG 1-X8 anion exchange resin in the formate form. Unincorporated [³H]-myo-inositol was removed by washing with 10 ml of distilled water and discarded. The [³H]-IPs were collected into scintillation vials by washing the columns with 8 ml of 50 mM ammonium formate and 4 ml of 1.2 M ammonium formate with 0.1 M formic acid. The radioactivity associated with the total [³H]-IPs was determined by scintillation counting on a beta-counter.

Autoradiographic studies. Frozen bovine corpus lutea were obtained from Pel-Freez (Rogers, AR) and rapidly frozen onto microtome chucks in Tissue-Tek O.C.T plastic embedding material (Miles Inc., Elkhart, IN). Sagittal tissue sections (20 µm) were cut at -17°C on freezing microtome and collected on gelatinized glass microscope slides (Sharif and Hughes, 1989; Sharif, 1996). The sections were preincubated in 550 ml of ice-cold 50 mM Tris.HCl (pH 7.4) containing 100 mM NaCl, 3 mM CaCl₂ and 5% bovine serum albumin (fraction V) for 1 hr. The slides were then laid flat on metal rods and covered with 1 ml of the different prepared solutions containing 10 nM [³H]PGF₂ in the assay buffer (see above) in the absence or presence of 100 µM unlabeled PGF₂ or fluprostenol to define total and nonspecific binding for the FP receptors, respectively; and 5 nM [³H]PGE₂ in the absence or presence of 100 µM unlabeled PGE₂ to label the EP receptors, respectively. After an incubation at 23°C for 60 min to achieve equilibrium, the solutions were poured off the slides and the latter rinsed in 550 ml of ice-cold buffer (see above; containing 1% instead of 5% bovine serum albumin) on a rotary mixer for 40 min. The slides were then dried in a stream of cool air and placed in a vacuum desiccator overnight at room temperature. Autoradiograms from the sections and [³H]Microscale radiation standards were generated over a 6-mo period with the films being subsequently developed, fixed, photographed and quantified by image analysis (Blue et al., 1995; Sharif, 1996).

Data analyses. The original data (dpm) from the different ligand binding experiments were analyzed using a nonlinear, iterative curve-fitting computer program (logistic function) (Sharif et al., 1996; 1997). Additional analyses were performed using the "EBDA" suite of computer programs (McPherson, 1983). The inhibition constants (K_is) were calculated from IC₅₀ values as previously described (Sharif et al., 1996, 1997). The PI turnover and cell proliferation (receptor selection amplification technology) functional data were analyzed by the sigmoidal fit function of the Origin Scientific Graphics software (Microcal Software, Northampton, MA) to determine agonist potency (EC₅₀ value) and efficacy. Autoradiographs were analyzed using an Agfa Horizon Ultra scanner and an Optimas 5.2 Software package.

Materials. [³H]PGE₂ (171 Ci/mmol); [³H]PGE₁ (52-56 Ci/mmol) and [³H]PGF₂ (150-175 Ci/mmol) were purchased from Du Pont-NEN (Boston, MA); [³H]Microscales, the radiation-sensitive Hyperfilm for autoradiographic studies, and [³H]-myo-inositol (18.3 Ci/mmol) were purchased from Amersham Corp. (Arlington Heights, IL). Kodak D19 and Kodafix were purchased from a local photography shop. Swiss albino mouse 3T3 fibroblasts (CCL-92, passage 116) were purchased from the American Type Culture Collection (Rockville, MD). Tissue culture and other reagents were purchased from Life Technologies (Grand Island, NY) included: Dulbecco's minimal essential medium, Dulbecco's minimal essential medium/F12 mixture, glutamine, gentamicin, trypsin/EDTA, balanced salt solution, phosphate-buffered saline without Ca⁺⁺ or Mg⁺⁺, Hanks' balanced salt solution and N-2-hydroxyethylpiperazine-N'-2-ethane sulfonic acid. Fetal bovine serum (HyClone, Logan, UT) was heat-inactivated at 56°C for 30 min and stored at -20°C. Frozen bovine corpus lutea from numerous cows of unknown menstrual/hormonal status were obtained from Pel-Freez. Ethylene diamine tetra acetic acid (di-sodium salt), Tris base, bovine serum albumin, formic acid, ammonium formate, LiCl and polyethylenimine were purchased from Sigma Chemical Co. (St. Louis, MO); AG 1-X8 anion exchange resin was a product of Bio-Rad (Hercules, CA). Ecolume scintillation fluid was supplied by ICN Biomedicals (Costa Mesa, CA). The following PGs were generous gifts from various companies: misoprostol, SC19220 and SC46275 from G.D. Searle & Co. (Skokie, IL), GR63799 from GlaxoWellcome (Stevenage, UK), ZK118182 from Schering AG (Berlin and Bergkamen, Germany) and S-1033 from Shionogi & Co., Ltd. (Osaka, Japan). Enprostil, sulprostone, butaprost, UF-021 and BWA868C were synthesized under contract or synthesized by our colleagues in the Research Chemistry department. The rest of the PGs used in our studies were purchased from Cayman Chemical Co. (Ann Arbor, MI). All other standard reagents, chemicals and buffers were purchased from Sigma. Optimas Software package was purchased from Optimas Corp. (Bothell, WA) and the AgfaHorizon Ultra scanner was purchased from Agfa-Gevaert AG (Boston, MA).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Specific binding of [³H]PGE₁ to washed total particulate homogenates of BCLM comprised 76 ± 7% of the total binding (e.g., total binding dpm = 1701 and nonspecific binding dpm = 469 at 2 nM). [³H]PGE₂ at 1 nM exhibited 82 ± 4% specific binding (e.g., 1893 total binding dpm and 345 dpm for nonspecific binding), although [³H]PGF₂ (10 nM) binding comprised 60 ± 2% of the total binding (1246 dpm total and 499 dpm nonspecific). Specific binding of all three radioligands was linearly related to tissue concentration and attained equilibrium within 1 to 2 hr at 23°C (data not shown).

Scatchard analysis of competition data revealed the presence of an apparent single population of binding sites for [³H]PGE₁ and [³H]PGE₂ with dissociation constants (K_ds) of 2.76 - 3.39 nM and apparent receptor density (B_max) of 1.5 to 1.56 pmol/g wet weight (n = 3-4) (fig. 1a and b). However, [³H]PGF₂ binding was best fitted to interaction with two classes of binding sites of differing affinities and densities (K_d1 = 6.51 ± 0.65 nM, B_max1 = 2.33 ± 0.26 pmol/g wet weight; K_d2 = 986 ± 269 nM; B_max2 = 44.8 ± 11.3 pmol/g wet weight, n = 11) (e.g., fig. 1c). Specific [³H]PGE₁ and [³H]PGE₂ binding was concentration-dependently inhibited by a diverse group of PGs but with different affinities. The pharmacological specificity of [³H]PGE₁ and [³H]PGE₂ binding was very similar and PGs selective for the EP₃ receptor subtype (e.g., GR63799, sulprostone, enprostil, misoprostol, SC-46275) exhibited the highest (nanomolar) affinity competing for sites occupied by these radioligands (fig. 2a, b,c; table 1). [³H]PGE₁ and [³H]PGE₂ binding was weakly (K_is > 1 µM) influenced by EP₁-selective (SC-19220), FP-selective (fluprostenol, PHXA85, cloprostenol), DP-selective (BWA868C; ZK118182), IP-selective (iloprost) and TP-selective (U46619) PGs (fig. 2a, b, c; table 1). The pharmacological specificity of [³H]PGE₁ and [³H]PGE₂ binding to BCLM membranes was very similar (fig. 2c). Furthermore, the pharmacology of [³H]PGE₂ binding to BCLM and the recombinant EP₃ receptor from the mouse was well correlated (fig. 2d).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Scatchard plots of specific [3H]PGE1 (a), [3H]PGE2 (b) and [3H]PGF2 (c) binding to BCLM membranes. The plots shown for [3H]PGE1 and [3H]PGE2 are from one representative experiment, while the data for [3H]PGF2 are from 11 experiments; the mean ± S.E.M. of the Kd and Bmax data from 3 to 11 experiments are also shown for these radioligands.

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Competition curves for various PGs displacing specific [3H]PGE2 (a and b) binding and correlation of pharmacology of [3H]PGE1 and [3H]PGE2 binding (c) to BCLM membranes (r = 0.973; P < 0.001). d depicts a correlation plot of [3H]PGE2 binding to BCLM and [3H]PGE2 binding to the cloned EP3 receptor from the mouse (Kiriyama et al, 1995) (r = 0.798, P < 0.0099). Data are mean ± S.E.M. from three to six experiments for the BCLM.

In our studies, only the high-affinity [³H]PGF₂ binding site was pharmacologically characterized. The identity and characterization of the low-affinity binding site/state of the FP receptor remains to be determined. Specific [³H]PGF₂ binding was potently displaced from the high-affinity sites by FP-selective agents like fluprostenol, PHXA85 and cloprostenol with nanomolar affinities, but weakly (K_is > 1 µM) by other PGs showing high selectivity for other PG receptor subtypes mentioned above (fig. 3; table 2).

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Competition curves for various PGs displacing specific [3H]PGF2 binding to BCLM membranes. Data are mean ± S.E.M. from 3 to 25 independent experiments.

View this table: [in this window] [in a new window]   TABLE 2 Affinities of selected PGs competing for specific [3H]PGF2 binding to BCLM membranes

The relative specificity and potency of EP₃-selective and FP-selective PGs tested in the binding assays were confirmed using functional assays in a cell-line expressing a human EP₃ receptor (fig. 4; table 3) and in Swiss 3T3 cells expressing an FP receptor (fig. 5; table 3), respectively. The receptor binding affinities and functional potencies of PGs tested in the current studies correlated well (r = 0.89-0.96; fig. 6a and b) indicating a pharmacological similarity of the receptor binding site and the receptor effector mechanisms for the EP₃ and FP receptors. It should be noted that to lend further credence to the results from the present studies, relevant ligand binding affinity data and functional potency data from the literature were also added to the correlation plots (figs. 6a and b) and also added to tables 1 to 3 for comparison.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Agonist-stimulated R-SAT functional responses at the recombinant human EP3 receptor. Data are mean ± S.E.M. from 3 to 11 independent experiments.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Agonist-stimulated PI turnover functional responses at the FP receptor in Swiss 3T3 cells. Data shown are from a representative experiment (error bars are from triplicate determinations) but the mean ± S.E.M. of results from up to 105 independent experiments are shown in table 3.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Correlation plots of receptor binding data and functional data. Results obtained from current studies using the human EP3 receptor in the R-SAT mode and those from Negishi et al (1994) (only PGE2 and misoprostol) for the inhibition of cAMP production in cells transfected with the recombinant mouse EP3 receptor correlated with EP3-binding to BCLM membranes are shown in a. The correlation plot shown in b depicts data from our FP-binding studies in the BCLM and functional studies in Swiss 3T3 cells () and also functional data from the cat iris sphincter contraction assay (Goh and Kishino, 1994; ) which contain FP receptors. The correlation coefficients were: EP3 binding and functional data, r = 0.96; FP binding and functional data, r = 0.89. Additionally, there was a good correlation between the FP binding data for BCLM from our studies and the cat iris contraction data (r = 0.83).

To localize the FP and EP₃ receptor binding sites in the bovine corpus luteum, autoradiographic techniques using emulsion-coated films were used. As can be seen in figure 7, the granulosa cells of this tissue expressed the highest density of FP receptors (total binding = 3763 ± 182 attomol/mg tissue; nonspecific binding = 86 ± 3 attomol/mg tissue; specific binding = 3677 ± 156 attomol/mg tissue; n = 40 readings from four sections). Very low binding was associated with connective tissue and blood vessels in the BCLM. EP₃ receptor sites were also localized to the granulosa cells but the density was much lower than the FP receptor sites detected under the present experimental conditions (data not shown).

View larger version (66K): [in this window] [in a new window]   Fig. 7.   Autoradiographic localization of FP receptors in BCLM sections. The top figure depicts total binding of [3H]PGF2 and the lower figure depicts nonspecific binding of [3H]PGF2 defined in the presence of 100 µM unlabeled fluprostenol. Similar degree of non-specific binding was obtained when 1-100 µM unlabeled fluprostenol or PGF2 were used. Note the high specific binding and high density of [3H]PGF2 binding sites associated with the granulosa cells but the very low density on connective tissue and blood vessels. CT, Connective tissue; GC, granulosa cells; BV, blood vessel. Magnification, ×4 of original.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Our studies have demonstrated the presence of high-affinity and specific binding sites for [³H]PGE₁, [³H]PGE₂ and [³H]PGF₂ in washed total particulate BCLM homogenates. Scatchard analyses indicated that [³H]PGE₁ and [³H]PGE₂ bound to a single population of nanomolar affinity receptor sites (figs. 1a, 1b). Whilst the latter observations confirmed the initial findings of Kimball and Lauderdale (1975) for [³H]PGE₁ binding, we have now provided additional data which demonstrated that both [³H]PGE₁ and [³H]PGE₂ labeled the same population of EP-receptor sites in this tissue because the K_d and B_max values obtained for both radioligands were very similar (figs. 1a and b). The affinity values for both [³H]PGE₂ and [³H]PGE₁ binding to the washed BCLM homogenates were also similar to K_ds obtained in recombinant EP-class (EP₃) receptors (K_ds = 0.3-6 nM) (Negishi et al., 1994; Kotani et al., 1995). Furthermore, we have shown that the pharmacological specificity (using 25 different prostanoids) of both radioligands in the BCLM was essentially identical (fig. 2c; table 1).

[³H]PGF₂ labeled an high-affinity (nanomolar K_d) and an apparent low-affinity (micromolar K_d) site in the washed BCLM homogenates as determined from several experiments (e.g., fig. 1c). These findings appeared to correlate well with the observation of high- and low-affinity states of the [³H]PGF₂-labeled sites in the rat (Wright et al., 1979; K_ds of 5 nM and 0.4 µM) and ovine (Balapure et al., 1989; K_ds of 17 nM and 0.5 µM) corpus lutea. Although the physiological relevance of this heterogeneity initially appeared to be unclear, apparent functional correlates for these affinity states have recently been demonstrated in ovine corpus luteum in vivo (Custer et al., 1995). Thus, the latter authors showed that occupancy of the high-affinity state of the FP receptor was responsible for the initial release of oxytocin, and that when these became desensitized the low-affinity state of the receptor mediated both the subsequent oxytocin release and the suppression of progesterone secretion during the luteolytic process (Harrison et al., 1987). The low-affinity state of the [³H]PGF₂-labeled sites in luteal cells may thus be an FP receptor, or it may represent a receptor for another eicosanoid which the corpora lutea secrete but which also binds PGF₂. This aspect remains to be further investigated.

Our major aim was to define the PG receptor subtype(s) binding sites present in the BCLM using the latest pharmacological tools. Accordingly, the current studies have demonstrated that both [³H]PGE₂ and [³H]PGE₁ selectively label EP₃ receptors in the BCLM since potent and highly selective EP₃-receptor PGs, such as GR-63799, SC-46275, sulprostone and enprostil competed for specific [³H]PGE₂ and [³H]PGE₁ binding with nanomolar affinities, while PGs selective for other PG receptor subtypes (such as fluprostenol and PHXA85 for the FP, BWA868C and ZK-118182 for DP, SC-19220 for EP₁, iloprost for IP, butaprost for EP₂ and U46619 for TP) were weak competitors (figs. 2a, b, c; table 1). The receptor binding affinities obtained for nine key prostanoids (some selective for the EP₃-receptor such as sulprostone, GR63799 and misprostol) against the cloned mouse EP₃-receptor (Kiriyama et al., 1997) correlated well with those obtained in the BCLM in our studies (table 1; fig. 2d). The binding affinities of these key PGs also correlated well with their potencies for stimulating functional responses at the cloned human EP₃-receptor in the R-SAT format (our studies) and for the most part with the cloned EP₃-receptor from the mouse (Negishi et al., 1994) (fig. 4; table 3). An interesting observation pertains to cloprostenol in the EP₃ binding and functional assays. Cloprostenol has been traditionally regarded as a selective FP-agonist much like fluprostenol (see Coleman et al., 1994 for review), except that cloprostenol in some in vitro tissue contraction assays had tended to exhibit some minor EP₃-like activity. Clearly, our binding and functional studies show that cloprostenol, unlike other FP-agonists like fluprostenol and PHXA85, has appreciable EP₃-receptor affinity and considerable agonistic activity at the EP₃ receptor (tables 1-3).

Our studies do not indicate which of the four EP₃-splice variants bind [³H]PGE₂ and [³H]PGE₁ in the BCLM preparations. Literature reports indicate that PGE₂ stimulates cAMP and progesterone synthesis in BCLM (Marsh, 1970), marmoset (Michael et al., 1993) and human luteal cells (Hahlin et al., 1988) and that the EP₃-selective PG, sulprostone, induces luteolysis in late pregnancy (Sander et al., 1982). Because EP_3B and EP_3C subtypes are coupled positively to adenylyl cyclase via G_s (Namba et al., 1993; Coleman et al., 1994) the latter appear to be the best candidates, but this requires further investigation and confirmation. The fact that EP₃ receptors are present in the BCLM (Tsai et al., 1996) and in human ovary (Kotani et al., 1995) has been recently demonstrated using molecular biological techniques. Unfortunately neither Tsai et al. (1996) nor Kotani et al. (1995) probed which specific EP₃ receptor splice variants may be present in the BCLM and human luteal cells and thus this aspect requires further work in the future.

The pharmacology of the [³H]PGF₂-labeled high-affinity receptor sites in the BCLM clearly indicates the identification of an FP receptor binding site. Supportive pharmacological evidence for this conclusion is that traditional prototypic FP-selective PGs such as fluprostenol, PHXA85 (Latanoprost acid), cloprostenol and 17-phenyl-PGF₂, were potent competitors of specific [³H]PGF₂ binding to the washed BCLM homogenates exhibiting nanomolar affinities (table 2), although PGs selective for DP (BWA868C, ZK-118182), EP-receptors (enprostil, sulprostone,), IP (PGI₂) and TP (U46619) had micromolar affinities (table 2). Similar results for a few of these compounds have recently been reported by Goh and Kishino (1994). The same profile also held true in the specific FP-receptor-mediated PI turnover response mechanism in Swiss 3T3 cells (figs. 5 and 6b; table 3) resulting in a good correlation between the FP-receptor binding in BCLM membranes and FP-receptor-mediated functional studies. A similar strong correlation was also observed between the FP binding affinities in the BCLM and the ability of various prostanoids to contract the cat iris sphincter muscle in vitro (Goh and Kishino, 1994; fig. 6b). It was noticeable, however, that although the "classic" FP-agonists (fluprostenol, PHXA85 and cloprostenol; Coleman et al., 1994; Stjernschantz et al., 1995) possessed high-affinities and intrinsic activities coupled with high functional potencies (fig. 5; tables 2 and 3), the recently described FP-like PG derivatives, S-1033 (Goh and Kishino, 1994) and UF-021 (Sakurai et al., 1992), exhibited relatively low FP-receptor affinities and potencies (tables 2 and 3). As to the functional relevance of the FP-receptor binding sites detected in the BCLM in our studies, it is known that PGF₂ in vitro stimulates PI turnover and Ca⁺⁺-mobilization in BCLM luteal cells (Davis et al., 1987) via the high-affinity receptor sites, and that the in vivo activation of FP receptors by PGF₂ and cloprostenol results in oxytocin release from small and large luteal cells followed by a suppression of progesterone secretion in the corpus luteum leading ultimately to luteal regression (Harrison et al., 1987; Custer et al., 1995).

Emulsion autoradiographic studies by Chegini et al. (1991) indicated that the binding sites labeled by [³H]PGF₂ and [³H]PGE₂ in the BCLM were apparently located on both small and large luteal cells and that the relative density of the FP and EP receptors, under their experimental conditions, were approximately the same. However, because the section thickness and rinsing procedures at the end of the receptor labeling in our studies were significantly different than those used by the later authors, we found a lower level of [³H]PGE₂-labeled sites than those labeled by [³H]PGF₂ in the BCLM sections. Furthermore, although Chegini et al. (1991) fixed the tissue sections prior to the emulsion autoradiography, but no fixation was performed for our film-based autoradiography, it is possible that the [³H]PGE₂-labeled sites, perhaps being more labile than the [³H]PGF₂-labeled sites, were diminished by the more extensive rinsing procedures during our studies. Another factor that could strongly influence the final results include the time at which the corpora lutea are harvested since the receptor mRNAs (Tsai et al., 1996) and the density of FP and EP receptor proteins vary significantly during the luteal cycle, as does the relative density on the small and large luteal cells (Niswender and Nett, 1988; Chegini et al., 1991).

In conclusion, the detailed pharmacological analyses of [³H]PGE₂/[³H]PGE₁ and [³H]PGF₂ binding to BCLM preparations using numerous PG receptor-selective compounds strongly supported the identification of EP₃ (perhaps the EP_3B/C splice variants) and FP receptors in the BCLM. These conclusions were supported further by the confirmatory functional classification of the numerous EP₃ and FP ligands employed currently using functional assays conducted in cell-lines containing the cloned EP₃ and constitutively expressed FP receptors. Further studies involving molecular biological techniques are suggested by the current results in order to define the presence of the different splice variants of the EP₃ receptor in the BCLM and corpus lutea of other species including humans.