Abstract
A detailed pharmacological characterization of the prostaglandin (PG) receptor coupled to phosphoinositide (PI) turnover and intracellular calcium mobilization in Swiss 3T3 mouse fibroblast cells was undertaken. The pharmacological profile of this functional receptor was compared with the pharmacological profile of specific [3H]PGF2 α binding to bovine corpus luteum membranes, which are known to contain a bona fide FP receptor. PGs that were potent stimulators and full agonists in the PI turnover assay in the 3T3 cells were the following (for all, n = 3–45): 16-phenoxy-PGF2 α (EC50 = 0.61 ± 0.1 nM), cloprostenol (EC50 = 0.73 ± 0.04 nM), 17-phenyl-PGF2 α (EC50 = 2.71 ± 0.35 nM), fluprostenol (EC50 = 3.67 ± 0.61 nM), PhXA85 (EC50 = 27.3 ± 5.63 nM) and PGF2 α (EC50 = 28.5 ± 5.26 nM). However, PGD2 (EC50 = 155 ± 29.9 nM;E max = 49% of cloprostenol), PGE2 (EC50 = 2570 ± 566 nM;E max = 59%) and U46619 (EC50 = 1060 ± 310 nM; E max = 63%) were less potent and were partial agonists, and iloprost and BW245C were inactive. Although the PGs tested exhibited lower affinities in the [3H]PGF2 α binding assay than their functional potencies in the PI turnover assay, the rank orders of potencies and affinities were well correlated (r = 0.94; n = 15 compounds). However, the PI turnover assay was more sensitive than the calcium mobilization assay for rank ordering PG agonists. In conclusion, the Swiss 3T3 cells express an FP receptor coupled to PI turnover and intracellular Ca++mobilization signal transduction pathways. The pharmacological profile of this receptor was similar to that of the FP receptor found in the bovine corpus luteum, a tissue previously used to clone the first pharmacologically defined FP receptor.
The physiological and pharmacological effects of endogenous and synthetic prostanoids have been characterized in many in vitro and in vivomodels and in more limited studies in humans (Coleman et al., 1981, 1994; Beckmann et al., 1988; Mitchellet al., 1994). These studies have established that PGs produce diverse and complex physiological effects, which are mediated by membrane-bound receptors that exhibit some degree of selectivity for the natural PGs, namely PGD2, PGE2, PGF2 α, PGI2 (prostacyclin) and thromboxane A2 (Coleman et al., 1990, 1994;DiMarzo, 1995). The current nomenclature for PG receptors (Colemanet al., 1994) defines the following receptor subtypes, which are believed to exist in the mammalian body: DP, EP (with further subtypes EP1, EP2, EP3 and EP4), FP, IP and TP (table 1). Due to alternative splicing at the genomic level, further subtypes of the EP3 receptor have been proposed (Coleman et al., 1994). The major determinants of this nomenclature have been the relatively recent availability of some selective agonists and, to a lesser degree, a few selective antagonists. The genes coding for all major classes of known PG receptors, from animal and human tissues, have now been cloned and expressed in various host cells (Narumiya, 1994; Coleman et al., 1994; Abramovitz et al., 1994; Boie et al., 1995) and the resultant receptor proteins biochemically and pharmacologically verified. These data, together with amino acid sequence data, have confirmed that all of the PG receptor proteins belong to the superfamily of G-protein-coupled receptors having seven transmembrane domains (Narumiya, 1994). The degree of homology among PG receptors of different classes can range up to 40% for the full-length sequence (approximately 400 amino acids) but is generally higher among the transmembrane domains of closely related proteins of this superfamily (Narumiya, 1994; Lake et al., 1994; Boie et al., 1995).
With regard to coupling of these PG receptors to distinct G-proteins and enzymes mediating the production of intracellular second messengers, the following information is available: FP, TP and EP1 receptors belong to one subfamily of PG receptors that couple to Gq or Gq/11, resulting in formation of inositol trisphosphate and mobilization of intracellular Ca++ (Narumiya, 1994; Coleman et al., 1994); the DP and IP receptors couple to Gs, with consequent activation of adenylyl cyclase and production of cAMP (Sugama et al., 1989; Namba et al., 1994; Boie et al., 1995). At present, several subtypes of EP receptors have been identified, which couple to various G-proteins (Narumiya, 1994; Negishiet al., 1995). Molecular biology techniques have revealed that the EP3 receptor subtypes are highly homologous except in regions of the cytoplasmic peptide loops, which bind to different G-proteins, consistent with the distinctive molecular pharmacology of the several EP3 receptor subtypes/splice variants. Because knowledge of the structural relationships among the different PG receptors has increased rapidly, new avenues of research into the functional roles and relationships among these important proteins are being identified.
The FP receptor, which mediates the biological effects of PGF2 α, has been characterized by functional pharmacological methods (Coleman, 1987; Chen et al., 1995), by receptor binding assays (Powell et al., 1976; Woodwardet al., 1995) and recently by molecular cloning techniques (Abramovitz et al., 1994; Sugimoto et al., 1994;Sakamoto et al., 1994; Lake et al., 1994). In the absence of selective antagonists, potent and selective agonists, such as fluprostenol and cloprostenol, have been used to classify the FP receptor (Coleman et al., 1994). As with other classes of PGs, there appear to be marked species and tissue differences in the physiological effects of PGF2 α. PGF2 α causes luteal regression in many species (Coleman et al., 1990) and has been reported to contract smooth muscle in various tissues, including the gastrointestinal tract, blood vessels and uterus (Coleman et al., 1990, 1994), from which the human FP receptor cDNA was isolated (Abramovitz et al., 1994; Lake et al., 1994). In the eye, PGF2 α and analogs have been shown to lower intraocular pressure in humans and primates, but this effect is not observed in all species (Bito et al., 1983; Wang et al., 1990). The potential of PGF2 α analogs as new therapeutic agents to lower elevated intraocular pressure has led to increased interest in the FP receptor as a target for drug discovery.
Preliminary studies have demonstrated the presence of a PG receptor linked to calcium mobilization in Swiss 3T3 cells (Woodward and Lawrence, 1994; Woodward et al., 1995). However, limited pharmacological studies were performed and no data were presented for coupling of these receptors to the PI turnover signaling mechanism or the pharmacological characteristics of the latter system. The aims of the present study, therefore, were to characterize the pharmacological properties of the PG receptor on the Swiss 3T3 cell line using PI turnover and calcium mobilization bioassays and to correlate these parameters with the pharmacology of [3H]PGF2 α receptor binding to bovine corpus luteum membranes, a tissue highly enriched in FP receptors (Powell et al., 1976; Coleman et al., 1994) and from which the FP receptor was first successfully cloned (Sakamoto et al., 1994). A preliminary account of the present studies has been recently presented (Griffin et al., 1995).
Materials and Methods
Cell culture.
Swiss albino mouse 3T3 fibroblasts were grown in DMEM containing 4.5 g/liter glucose and 110 mg/liter sodium pyruvate, which was supplemented with 2 mM l-glutamine, 10 μg/ml gentamicin sulfate and 10% fetal bovine serum. Cells were passaged, before reaching confluence, by treatment with 0.05% trpysin/0.53 mM EDTA and were seeded at high dilution for maintenance of the contact-inhibited phenotype (Takuwa et al., 1989). For agonist stimulation experiments, cells were grown to confluence in 24-well uncoated plastic plates.
Receptor binding experiments.
The competitive FP receptor binding assay was performed with a bovine corpus luteum membrane preparation (20 mg/ml) incubated with [3H]PGF2 α (0.9–1.5 nM; 150–175 Ci/mmol) and increasing concentrations (in duplicate) of the test compound for 2 hr at 23°C. The nonspecific binding was defined with 1 to 10 μM unlabeled PGF2 α. The assays were terminated by rapid vacuum filtration, using Whatman GF/B glass fiber filters that had been previously soaked in 0.3% polyethylenimine, and the receptor-bound radioactivity was determined by liquid scintillation counting at 50% efficiency. The data were analyzed by a nonlinear, iterative, curve-fitting program (Michel and Whiting, 1984; Sharif et al., 1991).
PI turnover experiments.
[3H]IPs produced by agonist-mediated activation of PLC were quantified by previously published procedures (Sharif et al., 1994, 1996). Briefly, confluent cells were exposed to 1.0 to 1.5 μCimyo-[3H]inositol (18.3 Ci/mmol) in 0.5 ml of DMEM for 24 to 30 hr at 37°C. Then cells were rinsed once with DMEM/F-12 medium 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; Sharif et al., in press). Cells were exposed to agonist or solvent for 60 min at 37°C (triplicate determinations), followed by aspiration of the medium and immediate addition of 1 ml of 0.1 M formic acid (held at 4°C). The plates were kept cold and then frozen. Samples frozen up to 1 week were thawed before chromatographic separation of radiolabeled components. The cell lysates (0.9 ml) were loaded on columns packed with approximately 1 ml of AG 1-X8 anion-exchange resin. The elution procedure consisted of washes with 10 ml of water, then 8 ml of 50 mM ammonium formate and finally 4 ml of 1.2 M ammonium formate with 0.1 M formic acid, which was collected in a scintillation vial. To this eluate was added 15 ml of scintillation fluid, and the total [3H]IPs were determined by scintillation counting in a beta-counter. 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, relative to the standard cloprostenol.
Studies of intracellular calcium mobilization.
The agonist-stimulated mobilization of intracellular Ca++ in Swiss mouse 3T3 cells was investigated by the fura-2 fluorescent Ca++ chelator method (Grynkiewicz et al., 1985), using stirred suspensions of cells in cuvettes. Some modifications of published procedures (Yamaguchi et al., 1988) were made, as described. To avoid trypsin degradation of the membrane receptors, cells grown in 175-cm2 flasks were detached by room temperature incubation with 0.05% EDTA in phosphate-buffered saline without Mg++ and Ca++, containing 0.1% glucose, for 35 to 45 min. The cell suspension was centrifuged briefly, and the pellet was immediately resuspended in Hanks’ BSS containing 1.3 mM Ca++. Cells were subsequently washed twice in Hanks’ BSS containing 10 mMN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid buffer, pH 7.4, and 0.1% BSA (designated as BSS-BSA buffer) and were then loaded with 2 μM fura-2/AM in the BSS-BSA buffer at room temperature in the dark for 60 min. Fura-2/AM was removed by washing the cells three times with the BSS-BSA buffer. Cell counts and cell viability (by trypan blue exclusion) were determined, and a concentrated stock suspension of the fura-2-loaded cells in the BSS-BSA buffer was stored on ice in the dark.
For the agonist stimulation experiments, cells were diluted to a concentration of 0.5 to 1.0 × 106/ml in a total volume of the BSS-BSA buffer sufficient for 10 to 12 experiments. After gentle thorough dispersion of this cell suspension, 1.5-ml aliquots were transferred to matched quartz cuvettes. This cell suspension was equilibrated at room temperature (23°C) for 1.0 to 1.5 hr and gently dispersed before the agonist stimulation experiment. Agonist stimulation experiments were routinely performed in the BSS-BSA buffer used by other laboratories (Yamaguchi et al., 1988; Nakadaet al., 1990). Control experiments established that the potency of cloprostenol was not altered by omitting BSA from the buffer. Fluorescence changes were measured with a Perkin-Elmer MPF-66 fluorescence spectrophotometer in cell suspensions maintained at room temperature and stirred with an overhead paddle-type stirrer. Excitation wavelength was 340 nm, with 1-nm excitation slit width; emission was monitored at 510 nm, with the emission slit width adjusted to optimize the net fluorescence change established by the calibration procedure. Agonists were added as ethanol solutions of the acid species, in a total volume no greater than 1% of the cuvette volume; solvent blanks were typically no greater than 10% of the maximal agonist-dependent response. After the agonist-stimulated response had decayed, the fluorescence signal was calibrated by adding digitonin (50 μg/ml) and finally either 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (4 mM) or ethylene glycol-O,O′-bis(2-aminoethyl)-N,N,N′,N′-tetraacetic acid (6.6 mM) plus Tris base (5 mM) to the final cuvette concentrations shown in parentheses. Based on the maximal and minimal responses determined by the calibration procedure, the agonist-dependent increases in intracellular fluorescence were converted to intracellular Ca++ concentrations by standard equations, assuming aK d value of 224 nM (Grynkiewicz et al., 1985). This K d value was reproduced using Ca++/EDTA calibration buffers. The properties of the PG-stimulated Ca++ response of this cell type were determined to be similar to those reported (Woodward et al., 1990; Woodward and Lawrence, 1994). Because of desensitization of the response, only one agonist stimulation experiment was performed with each sample (cuvette) of cells. Data were analyzed by the Origin software, as described above, to determine agonist potency (EC50 value).
Materials.
Swiss albino mouse 3T3 fibroblasts (CCL-92, passage 116) were purchased from the American Type Culture Collection (Rockville, MD). Tissue culture reagents and other reagents purchased from Life Technologies (Grand Island, NY) included DMEM, DMEM/F-12 medium, glutamine, gentamicin, trypsin/EDTA, BSS, phosphate-buffered saline without Ca++ or Mg++, Hanks’ BSS andN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid. Fetal bovine serum (Hyclone, Logan, UT) was heat-inactivated at 56°C for 30 min and stored at −20°C. EDTA (disodium salt), Tris base, BSA, digitonin, formic acid, ammonium formate, LiCl and polyethylenimine were supplied by Sigma Chemical Co. (St. Louis, MO). Ethylene glycol-O,O′-bis(2-aminoethyl)-N,N,N′,N′-tetraacetic acid was a product of Fluka BioChemika (Buchs, Germany); Molecular Probes, (Eugene, OR) was the source of fura-2, fura-2/AM, the nonfluorescent Ca++ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid and Ca++/EDTA calibration buffers. Amersham Corp. (Arlington Heights, IL) was the source ofmyo-[3H]inositol, and [3H]PGF2 α was purchased from DuPont NEN (Boston, MA). 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). U73122 and U73343 were purchased from Biomol Research Laboratories (Plymouth Meeting, PA). All PGs were purchased from Cayman Chemical Co. (Ann Arbor, MI) or synthesized at Alcon by published methods, except as follows: SC-46275 and SC-19220 were kindly provided by G.D. Searle (Skokie, IL), and AH6809 was purchased from Tocris-Cookson (St. Louis, MO).
Results
PI turnover and receptor binding experiments.
The cloprostenol-stimulated [3H]IPs accumulation in Swiss 3T3 cells at 37°C was linear for at least 1 hr over a range of agonist concentrations (data not shown), and the maximum response did not decrease significantly even at cloprostenol concentrations as high as 100 μM. Therefore, all subsequent studies were conducted at 37°C for 1 hr. The accumulation of [3H]IPs in these cells in response to stimulation by six PGs known to have some degree of selectivity for distinct PG receptors is shown in figure1. The potencies of these compounds were as follows (for all, n = 3–45): 16-phenoxy-PGF2 α (EC50 = 0.61 ± 0.1 nM), cloprostenol (EC50 = 0.73 ± 0.04 nM), 17-phenyl-PGF2 α (EC50 = 2.71 ± 0.35 nM), fluprostenol (EC50 = 3.67 ± 0.61 nM), PhXA85 (EC50 = 27.3 ± 5.63 nM) and PGF2 α (EC50 = 28.5 ± 5.26 nM) (table 2; fig. 1). In contrast, PGD2(EC50 = 155 ± 29.9 nM; E max = 49% of cloprostenol), PGE2 (EC50 = 2570 ± 566 nM; E max = 59% of cloprostenol) and U46619 (EC50 = 1060 ± 310 nM;E max = 63% of cloprostenol) were less potent and were partial agonists (table 2; fig. 1). Based on these results, cloprostenol was selected as the “standard control agonist” included in all experiments.
The concentration-dependent displacement of [3H]PGF2 α from the FP receptor-containing membrane preparation by this group of PGs is shown in figure 2. The rank orders of affinities of these compounds in the binding assay and potencies in the functional assay were similar (fig. 3; see below). Additional PGs with well-characterized activities at various PG receptors (table 1) were evaluated in both test systems to confirm the identity of this receptor and to establish the degree of correlation of the receptor potency and binding affinity data; these data are summarized in table 2, along with published data for a small subset of PGs evaluated in various FP receptor-containing preparations used by other researchers. As shown in figure 3, there is an excellent linear correlation of log EC50 in the functional assay with the log IC50in the binding assay (r = 0.94; n = 15 compounds). These results demonstrated that PGF2 α and some well-characterized FP receptor agonists are the most potent compounds in both assays and also the most efficacious agonists in the functional assay. The inactive PGs and analogs included the IP receptor agonist iloprost, the potent and selective DP receptor agonist BW245C, butaprost (a selective EP2 receptor agonist), misoprostol (an agonist at the EP2 and EP3 receptors), SC-46275 (an EP3 agonist), anandamide (an endogenous cannabinoid derived from arachidonic acid) (DiMarzo et al., 1994) and 8-isoprostane, an isomer of PGF2 α resulting from lipid peroxidation reactions in vitro and in vivo (Morrow et al., 1990) (table 2). A few compounds reported to have some degree of selectivity at other PG receptor classes (table 1) (Coleman et al., 1994) (e.g.,17-phenyl-PGE2, an EP1 receptor agonist, and enprostil, reported to be an EP3 receptor agonist) had measurable activity in both test systems but exhibited <1% of the potency of cloprostenol in the 3T3 cell system (table 2).
Two EP1 receptor antagonists, AH6809 and SC-19220 (table1), had no effect on the PI turnover response to 17-phenyl-PGE2 (data not shown). Also, the response to PGD2 was not antagonized by the selective DP receptor antagonist BWA868C (table 1) (Coleman et al., 1994); a similar negative result was obtained when the TP receptor antagonist SQ 29,458 was evaluated as an inhibitor of the potent TP receptor agonist U-46619 (table 1) (Coleman et al., 1994) (data not shown). The fluprostenol-stimulated [3H]IPs response was inhibited by the known PLC inhibitor U73122 (Bleasdale et al., 1990), with an IC50 of 1.24 ± 0.21 μM (n = 4), but was not altered by high concentrations of U73343, an inactive analog of U73122 (fig. 4).
The effects of other classes of PLC-linked agonists in Swiss 3T3 cells were evaluated with the same experimental protocol. We confirmed other reports (Takuwa et al., 1989) that endothelin-1 is a potent agonist coupled to PLC activation in this cell type (EC50 = 1.32 ± 0.44 nM; n = 6). In addition, we demonstrated that the response to endothelin-1 was inhibited by the potent ETA antagonist BQ-610 (data not shown). Nonprostanoid agonists that were inactive included bradykinin, histamine, angiotensin II, Arg-vasopressin, platelet-activating factor, bombesin, neurokinin A, substance P, carbachol and peptide YY. These data provide additional evidence for the receptor selectivity of the PG-stimulated response.
Intracellular calcium mobilization studies.
Studies of agonist-stimulated mobilization of intracellular Ca++ in 3T3 cells were undertaken with a limited number of compounds. The intracellular Ca++ concentration signal produced by the active agonists was concentration-dependent (fig. 5). The rank order of potency and efficacy of PGs as stimuli of intracellular Ca++ mobilization (table 3) was similar to that established in the [3H]IPs production assay (table 2), with FP agonists being more potent and efficacious than agents selective for DP and EP receptors (table 3). However, the most potent agonists that could be discriminated and thus readily rank ordered in the PI turnover assay, i.e., cloprostenol, fluprostenol, 17-phenyl-PGF2 α, 16-phenoxy-PGF2 α and PhXA85, had similar potencies (60–70 nM) in the intracellular Ca++mobilization assay.
Discussion
We have used the techniques of PI turnover and intracellular Ca++ mobilization to pharmacologically define the functionally coupled PG receptors present on Swiss mouse 3T3 fibroblast cells. Because Powell et al. (1976) previously demonstrated the enrichment of FP receptors in the bovine corpus luteum, and because the first FP receptor cloning was successfully accomplished using this tissue (Sakamoto et al., 1994), we compared the pharmacological profile of the 3T3 cell PG receptor with that of the bovine corpus luteum, using receptor binding techniques. The results of all our studies strongly indicate that the sole receptor mediating PLC-induced PI turnover and intracellular Ca++ mobilization in the 3T3 cells is the FP receptor. Supportive data for this conclusion include the findings that, whereas prototypic FP agonists (such as cloprostenol, fluprostenol, PHXA85 and 17-phenyl-PGF2 α) (table 1) had nanomolar potency and were highly efficacious agonists in both functional assays and receptor binding assays, PGs selective for DP receptors (e.g., BW246C, BWA868C and PGD2) (table 1), EP receptors (e.g., enprostil, sulprostone, misoprostil, butaprost, 17-phenyl-PGE2 and 11-deoxy-16,16-dimethyl-PGE2) (table 1), IP receptors (e.g., iloprost) and TP receptors (e.g., U46619) were of much lower potency and showed significantly lower affinities than the FP ligands. Moreover, the failure of two EP1receptor antagonists (SC-19220 and AH6809) (table 1) and a DP antagonist (BWA868C) to inhibit the [3H]IPs response to 17-phenyl-PGE2 and PGD2, respectively, indicated that these PGs were also activating the FP receptor in the Swiss 3T3 cells. To further corroborate these conclusions, we recently showed the presence of a strong signal for an FP receptor mRNA transcript in Swiss 3T3 cell lysates, using RT-PCR techniques, and only a very faint transcript band for TP receptor mRNA (N. A. Sharif, M. Senchyna, D. J. Crankshaw and B. W. Griffin, unpublished observations).
Although initial intracellular Ca++ mobilization studies in 3T3 cells suggested the potential presence of FP receptors in these cells (Woodward et al., 1990; Woodward and Lawrence, 1994), we have now provided extensive pharmacological evidence, using two different bioassays, to corroborate these earlier findings. Our studies have also indicated that the resolution of agonist potencies and efficacies, and thus the ability to rank order compounds, was much more clearly defined using the [3H]IPs accumulation assay, compared with the intracellular Ca++ mobilization assay. Moreover, the role of PLC was conclusively demonstrated by inhibition of the response by U73122, a selective PLC inhibitor (Bleasdaleet al., 1990). The inhibition constant of U73122 (1.24 μM) was in the range of published values for various PLC-coupled receptors (Bleasdale et al., 1990). The aforementioned combined information was not previously available in the literature and is an important finding for future use of the Swiss 3T3 cells for pharmacological evaluation of PGs. As mentioned above, although there are limited data demonstrating that PGF2 αproduces the expected IP response in both NIH 3T3 and Swiss 3T3 cells (Corps et al., 1989; Nakao et al., 1993), studies of PLC activation in either cell line by other PGs, including known selective FP receptor agonists, have not been reported. Hence, our PI turnover data for the Swiss 3T3 cells represents novel information that is important for the classification of the PG receptor expressed by these cells.
Other studies with “3T3” cells of various types have been reported.Corps et al. (1989) studied the Ca++mobilization responses of single, attached “Swiss 3T3” cells to PGF2 α, vasopressin and other peptides. These agonists, at a single high concentration, also stimulated formation of IPs in fura-2-loaded cells. The cellular responses of “NIH 3T3” cells to PGF2 α, and binding of [3H]PGF2 α to membranes from these cells, were characterized by Nakao et al. (1993), who attributed some discrepancies between their results and earlier data to clonal variations in the NIH 3T3 cell line. In a study of PGF2 α-induced signal transduction in “3T3-L1” fibroblasts, it was reported that vasopressin, bradykinin and bombesin produced no response (Nakada et al., 1990), consistent with our negative results with these agonists in the Swiss 3T3 cells. Although some differences in the particular cells used, as well as the methodology, could account for differences in PGF2 α potency in the different studies, our EC50 values of 28.5 nM and 102 nM for the [3H]IPs accumulation and Ca++ responses, respectively, agree more closely with the values reported by Nakaoet al. (1993) (EC50 values of 46 nM for [3H]IPs and 75 nM for Ca++) than those determined by Nakada et al. (1990) for the 3T3-L1 fibroblasts (250 nM for both responses). Similar potency values (EC50 = 36–45 nM) for PGF2 α were also reported for bovine corpus luteal cell PI turnover and intracellular Ca++ mobilization (Davis et al., 1987). Quantitative potency data were not reported in recent publications describing the Ca++ mobilization responses of Swiss 3T3 cells in stirred cell suspensions (Woodward et al., 1990; Woodward and Lawrence, 1994), but EC50values estimated from their graphical data in those publications appear to be generally consistent with our results for PGF2 α, fluprostenol, PGD2 and PGE2. In the same studies, sulprostone and U-46619 increased intracellular Ca++ with potencies estimated to be in the range of 1 to 5 μM and BW245C was inactive, consistent with the quantitative PI turnover data reported in this work.
An excellent correlation between FP receptor functional potency in the Swiss 3T3 cells and binding affinity of several PGs for the bovine corpus luteal FP receptors was observed (fig. 3). The receptor affinity and rank order data for six key PGs, obtained from inhibition of [3H]17-phenyl-PGF2 α binding to Swiss 3T3 cells membranes (Woodward et al., 1995), also correlated well with our functional data in the 3T3 cells and the [3H]PGF2 α binding data from the corpus luteum. Furthermore, our functional and ligand binding data were also well correlated with the rank order of potency of fluprostenol, cloprostenol, PGF2 α, PGD2 and PGE2, as determined in dog and cat iris contraction assays (Coleman et al., 1990). All of these results are consistent with recent gene sequencing data suggesting that the FP receptor probably exists as a single isoform with a high degree of homology among species (Lake et al., 1994). However, the study ofLake et al. (1994) revealed variability in size of the mouse FP receptor transcripts (2–6 kilobases) and also low levels of a larger transcript (6.5 kilobases) in those human tissues (e.g., ovary) with highest amounts of the 6-kilobase transcript. Based on these results, the possibility of minor subtypes of the FP receptor could not be eliminated. However, the profiles of the concentration-response and concentration-inhibition plots of our functional and receptor binding data indicated the presence of a single FP receptor exhibiting high affinity for PGF2 α analogs in both preparations. Even though there appears to be considerable homology in the FP receptor of various species, there are some notable differences in pharmacological characteristics of this receptor expressed in transfected cells (typically COS cells). As evidence for receptor identity, the ligand binding properties of each of these receptors have been determined with membranes from the transfected cells; these results have demonstrated that PGF2 α binds more strongly than other classes of PGs (Abramovitz et al., 1994; Sugimoto et al., 1994; Sakamoto et al., 1994; Lake et al., 1994). However, the IC50 value of PGF2 α for these various FP receptors ranged from a low of about 2.5 nM (Sugimoto et al., 1994;Abramovitz et al., 1994) to 40 nM (Graves et al., 1995). The COS cell-expressed human FP receptor has comparable affinity for PGF2 α (2.8 nM) and PGD2 (7.0 nM) and only about 30-fold lower affinity for PGE2 (85 nM), compared with PGF2 α (Abramovitz et al., 1994). In contrast, the FP receptors of cow, sheep, mouse and rat, expressed and characterized by very similar techniques, have IC50 values of 300 to 500 nM for PGD2 and 0.4 to 2.0 μM for PGE2 (Sugimoto et al., 1994;Sakamoto et al., 1994; Lake et al., 1994; Graveset al., 1995). Based on these limited data, the PG binding selectivity of the FP receptor gene product from the several animal sources appears to be similar to that of the bovine corpus luteum FP receptor used in our studies. The significant differences in absolute IC50 values of PGF2 α for the expressed FP receptor genes of different species are unexpected, because the ligand-binding region of this receptor should be among the most highly conserved parts of the protein. Differences in the experimental conditions used to transfect, select and grow the transfected cells could influence the membrane environment of the FP receptor and, consequently, the specific and nonspecific binding characteristics of the FP receptor preparation. Also, the specific activity of the radiolabeled ligand probe determines the minimal detectable signal and the limits of the assay. Although differences in methodology could account for the observed species differences in binding of the FP receptor gene product to PGF2 α and other classes of PGs, other explanations for these results, such as distinct FP receptor subtypes, cannot be excluded at this stage.
In our study, the IC50 value for binding of PGF2 α to the FP receptor of bovine corpus luteum was about 3-fold larger than the largest value reported for the expressed FP receptor gene. As the data in table 1 and figure 3indicate, the IC50 values determined from the binding assay were about 1 log unit larger than the EC50 values for agonist-dependent PI hydrolysis by Swiss 3T3 cells. Several factors, in addition to those mentioned above, may account for the differences in absolute potency or binding affinity (IC50 value) measured for a given compound in various assays. The sensitivity of each assay is determined by the minimal number of receptors required for an acceptable signal-to-noise ratio (reflecting the inherent “signal” of the radiolabeled probe or fluorescent probe, in the case of the Ca++ assay), the degree of amplification of the response by physiological means (activation of an enzyme) or other methods (such as the use of LiCl to inhibit metabolism of IP species) (Berridge et al., 1982), the sources and types of background signals in the different assays and other experimental conditions that may be unique to the particular assay. For example, the PI turnover assay has some unique characteristics, including enzymatic amplification, LiCl-dependent accumulation of the radiolabeled “second messengers” and, as discussed, no apparent contribution from other functional receptors on Swiss 3T3 cells that might bind PGs. Consequently, the signal-to-noise ratio for this assay was excellent, and the responses produced by very low concentrations of very potent agonists could be easily detected with small numbers of cells. It is more difficult to directly compare the potency values for the [3H]IPs accumulation and Ca++ responses measured in 3T3 cells, because of the limited number of compounds evaluated in the latter assay. To the best of our knowledge, there are few published articles with extensive comparative potency data generated in the two types of assays. There are multiple physiological controls over the kinetics, magnitude and duration of the intracellular Ca++ response that cannot be easily varied by the experimenter. Also, the use of detached cells (in contrast to the attached cell monolayers used in the PI turnover assay), reliance on a chelator-ion association phenomenon and inherent limitations of quantifying fluorescence signals in a biological milieu impose additional constraints on absolute potency values that can be measured by the fura-2 method. Although each assay of receptor “activity” used in this study appears to have somewhat different discriminatory power for the most potent compounds, the rank order of potency of compounds with a wide range of potencies was similar in all assays, again reinforcing the tenets of the pharmacological process to classify the PG receptors.
The quantitative potency and efficacy data obtained in this study confirm that the FP receptor, like other PG receptors, is not absolutely selective for FP agonists and responds to relatively high concentrations of other classes of endogenous PGs and synthetic analogs. PGs are considered to serve mainly as locally produced autocrine and paracrine signals. Consequently, their production by cyclooxgenase 1 and 2 (Laneuville et al., 1994; Mitchellet al., 1994) and their subsequent metabolism are tightly controlled, to ensure that their concentrations are maintained within physiological limits. Also, tissue-specific control over the particular PGs synthesized from arachidonic acid and the identity, density and occupancy of specific PG receptors contribute to the selectivity of prostanoid-mediated physiological effects. Other factors also influence the functional selectivity of the PG receptors, such as the type of G-protein involved in the ensuing signal transduction cascade. The possibility that interaction of the receptor with its G-protein could be a mechanism of control of ligand binding affinity, and thus selectivity, has received experimental support for certain PG receptors (Negishi et al., 1993). As PG receptors and their associated G-proteins are characterized more thoroughly, and as the molecular functions of the cyclooxygenase enzymes are elucidated, our understanding of the mechanisms controlling functional selectivity of different PG classes will greatly increase.
In conclusion, the data in this report have increased our understanding of the in vitro pharmacological properties of the FP receptor and have provided additional evidence that the PG-responsive receptor in Swiss 3T3 cells, which is coupled to PLC activation and intracellular Ca++ mobilization, is indeed the FP receptor. Our data demonstrated that an assay for FP receptor function that uses the physiological signal amplification process coupled to receptor activation, e.g., [3H]IPs formation, provided by an intact cellular system is a valuable method to generate quantitative potency and efficacy data for agonists with a broad range of potencies and efficacies. However, as described in this report, ligand binding techniques used in combination with Ca++mobilization and PI turnover studies provided a more thorough characterization of the physiological and pharmacological properties of the FP receptor, and thus such a multidisciplinary approach represents a powerful means to study and classify receptors.
Acknowledgments
We express appreciation to our colleagues in the Research Chemistry group for synthesizing some of the PGs used in our studies. Terry Davis is thanked for expert technical assistance in some binding experiments. Drs. T. R. Dean, M. R. Hellberg, V. Sallee and L. M. DeSantis are thanked for valuable discussions during these studies. The support and encouragement of Dr. B. York during the course of these studies are also gratefully acknowledged.
Footnotes
-
Send reprint requests to: Naj Sharif, Ph.D., Molecular Pharmacology Unit, Alcon Laboratories, Inc. (R2–19), 6201 South Freeway, Fort Worth, TX 76134-2099.
- Abbreviations:
- AM
- acetoxymethyl ester
- BSA
- bovine serum albumin
- BSS
- balanced salt solution
- DMEM
- Dulbecco’s modified Eagle medium
- IPs
- inositol phosphate species
- PG
- prostaglandin
- PI
- phosphoinositide
- PLC
- phospholipase C
- Received September 20, 1996.
- Accepted January 10, 1997.
- The American Society for Pharmacology and Experimental Therapeutics