Prostaglandin F_2α Receptor-Dependent Regulation of Prostaglandin Transport

Experimental Procedures

Materials.

All the cell culture media and G418 were purchased from Life Technologies Inc. (Gaithersburg, MD). Zeocin was purchased from Invitrogen (Carlsbad, CA). The anion exchange resin AG 1-X8 (formate form, 200–400 mesh) and 30% acrylamide/bisacrylamide solution were purchased from Bio-Rad (Hercules, CA). [³H]PGF_2α (212 Ci/mmol), [³H]PGE₂ (159 Ci/mmol), myo-[2-³H]inositol (18.0 Ci/mmol), cyclic AMP radioimmunoassay, and enhanced chemiluminescence kits were obtained from Amersham Pharmacia Biotech (Piscataway, NJ). PGF_2α was obtained from Cayman Chemicals (Ann Arbor, MI). Cholera toxin and pertussis toxin were obtained from List Biological Laboratories Inc. (Campbell, CA).N-[2-([p-bromocinnamyl]amino)ethyl]-5-isoquinolinesulfonamide (H89 dihydrochloride) and dibutyryl cyclic AMP (dBcAMP) were obtained from Calbiochem (La Jolla, CA). 3-Isobutyl-1-methylxanthine (IBMX) was obtained from Sigma (St. Louis, MO) and 4,4′-diisothiocyanato-2,2′-stilbenedisulfonic acid, disodium salt hydrate (DIDS) was purchased from Aldrich (Milwaukee, WI). Iloprost was purchased from Schering-Plough (Berlin, Germany). Anti Gα_s, Gα_i, Gα_q/11, Gα₁₂, and Gα₁₃ antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Cell Culture and Transfection.

The human FP was cloned in our laboratory (Kunapuli et al., 1997). A nine-amino-acid hemagglutinin epitope (HA) (YPYDVPDYA) was inserted between the N-terminal initiator methionine and the second amino acid, as described previously for the HA-IP (Smyth et al., 1996). The HA-FP cDNA was subcloned into theBamHI/EcoRI sites of the mammalian expression vector pcDNA3.1 (Invitrogen) and used for stable transfection of HEK293 cells (HA-FP cells). The cDNA encoding hPGT, kindly donated by Victor L. Schuster of the Albert Einstein College of Medicine (Bronx, NY), was subcloned into the HindIII-NotI sites of pcDNA3.1 or pcDNA3.1/Zeo (Invitrogen). The hPGT in pcDNA3.1 was used for stable transfection of HEK293 cells (hPGT cells), whereas the hPGT in pcDNA3.1/Zeo was used for transfection of HA-FP cells. Subcloning into pcDNA3.1/Zeo allowed us to use zeocin as a second selection marker. Cells overexpressing both the HA-FP and the hPGT are indicated as HA-FP/hPGT cells. A Gα_s dominant negative cDNA (A³⁶⁶S/G²²⁶A/E²⁶⁸A) in pcDNAI was obtained from American Type Culture Collection, Manassas, VA). The minigene constructs in pcDNA3.1 were kindly donated by Drs. Annette Gilchrist and Heidi E. Hamm (Northwestern University, Chicago, IL).

HEK293 cells (American Type Culture Collection) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 50 units/ml penicillin, 100 μg/ml streptomycin, 25 mM HEPES, and 2 mM glutamine under 5% CO₂ at 37°C. For transfection, cells were seeded at 1.5 × 10⁶ cells/100-mm dish and, the next day, transfected with 10 μg/dish cDNA by liposome-mediated transfer [N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl sulfate; Roche Molecular Biochemicals, Mannheim, Germany] according to the manufacturer's instructions. Eight hours after the beginning of transfection, the medium was changed and selection of the clones was performed with 0.75 mg/ml G418 (for HA-FP and hPGT cells) or with 0.75 mg/ml G418 plus 0.3 mg/ml zeocin (for HA-FP/hPGT cells). Binding of [³H]PGF_2α to membranes from HA-FP/hPGT cells was saturable and revealed aB_max value of 0.52 pmol/mg protein and aK_d value of 17.6 nM. Nonspecific binding never exceeded 12% of total binding.

For transient transfection, cells were seeded at 3 × 10⁵ cells/well in six-well plates and transfected 1 to 2 days later. When cells were transfected with two plasmids, we used 0.5 μg of each plasmid DNA mixed with 3 μl of FuGENE6 (Roche Molecular Biochemicals) per well in 2 ml of medium, according to the manufacturer's instructions. In the experiments with Gα_s dominant negative or minigene constructs, three plasmids were used simultaneously. In these cases, we transfected each well with 0.4 μg of HA-FP or hPGT cDNA, and 1.2 μg of Gα_s dominant negative or minigene constructs. Plasmid DNA was mixed with 4.5 μl of FuGENE6 per well in 2 ml of medium. Experiments were performed 48 h after transfection. NIH3T3 cells (American Type Culture Collection) were cultured in DMEM supplemented with 10% fetal bovine serum, 50 units/ml penicillin, 100 μg/ml streptomycin, and 2 mM glutamine under 10% CO₂ at 37°C.

Inositol Phosphate Formation.

Inositol phosphate formation was measured in HA-FP/hPGT, hPGT, and NIH3T3 cells. Cells were plated in six-well plates at 2.5 × 10⁵ cells/well and, the following day, incubated with 2 μCi/ml ofmyo-[2-³H]inositol for 20 to 24 h in inositol-free DMEM. The day of the experiment, the medium was replaced with the same medium containing 20 mM LiCl, and the cells were stimulated with different concentrations of PGF_2α. HA-FP/hPGT cells were stimulated for 10 min, whereas hPGT and NIH3T3 cells were stimulated for 30 min. Reactions were terminated by aspiration of the medium and inositol phosphates were extracted with 750 μl of 10 mM formic acid for 30 min, as described previously (Vezza et al., 1996; Habib et al., 1997). Agonist-stimulated inositol phosphate formation is expressed as a percentage of the vehicle-treated sample.

Uptake of [³H]PGF_2α

Uptake of [³H]PGF_2α or [³H]PGE₂ was carried out as described previously (Lu et al., 1996). HA-FP/hPGT and NIH3T3 cells were plated in six-well plates at 2.5 × 10⁵ cells/well 1 day before the experiment. The hPGT cells did not adhere firmly to the wells. Thus, they were plated 3 to 4 days before the experiment, to allow them to adhere and spread.

Cells were washed with DMEM and incubated with 0.6 nM [³H]PGF_2α or 0.7 nM [³H]PGE₂ in the same medium at room temperature for different time intervals. Uptake was terminated by addition of ice-cold DMEM containing 5% bovine serum albumin. Cells were washed three times with DMEM, scraped in phosphate-buffered saline, and counted by liquid scintillation. In selected experiments, NIH3T3 cells were incubated for 15 min at 37°C with the anion transporter inhibitor DIDS (1 mM). Uptake of [³H]PGF_2α was measured in the presence of DIDS, as described previously (Chan et al., 1998).

To assess whether stimulation of the FP alters hPGT activity, HA-FP/hPGT or hPGT cells were stimulated with different concentrations of PGF_2α, or its vehicle, for 10 min at 37°C. When transiently transfected HEK293 cells were used, PGF_2α or iloprost was incubated for 30 min before measurement of [³H]PGF_2α uptake. Cells were then washed and incubated with 0.6 nM [³H]PGF_2α for 10 min at room temperature and uptake experiments were performed as described above. Results are expressed as a percentage of the vehicle-treated sample or as disintegrations per minute per microgram of protein. In the latter case, an aliquot was removed from each sample before addition of scintillation fluid. Protein concentration was determined using the Bradford assay with bovine serum albumin as a standard.

The cyclic AMP-dependent protein kinase inhibitor H89 and the cyclic AMP anolog dBcAMP (10 μM) were incubated for 30 min at 37°C before the addition of PGF_2α and the assessment of transporter activity.

Cholera or pertussis toxin was used at 250 ng/ml, unless otherwise indicated in the text, and was incubated for 20 to 24 h at 37°C. Cells were then washed, stimulated with PGF_2αor its vehicle for 10 min at 37°C, washed again, and incubated with [³H]PGF_2α to assess transporter activity.

Cyclic AMP Measurement.

HA-FP/hPGT cells were plated in six-well plates at 2.5 × 10⁵ cells/well. Cells were incubated with IBMX (0.5 mM) for 5 min and then stimulated with different concentrations of PGF_2α, or its vehicle, for 10 min at 37°C. HA-FP cells were seeded at 1.2 × 10⁵ cells/ml in 12-well plates and incubated, or not, with pertussis toxin 250 ng/ml for 20 to 24 h. Cells were then stimulated with different concentrations of PGF_2α, or its vehicle, for 10 min in the presence of IBMX (0.5 mM). HA-IP cells (Smyth et al., 1996) were plated in 12-well plates at 1.2 × 10⁵ cells/ml and incubated with cholera toxin for 20 to 24 h and stimulated with different concentrations of iloprost (10 pM-10 nM) for 10 min at 37°C.

Reactions were terminated by aspiration of the medium. Cyclic AMP was extracted with 600 μl of ice-cold 65% ethanol for 30 min and quantitated by radioimmunoassay, as described previously (Smyth et al., 1996).

Immunoblotting.

A polyclonal peptide antibody was raised in rabbits to a 17 amino acid sequence of the carboxyl-terminal tail of the hPGT (H-RVKKNKEYNVQKAAGLI-OH) (Research Genetics Inc., Huntsville, AL). Immunoblotting was performed as described previously (Vezza et al., 1996). Briefly, blots were incubated for 1 h with the anti-hPGT antibody diluted 1:2000 in Tris-buffered saline (50 mM Tris-HCl, 250 mM NaCl, pH 7.4) containing 0.1% Tween 20 and 5% milk. Anti-Gα_s, Gα_i, Gα_q/11, Gα₁₂, and Gα₁₃ antibodies were diluted 1:200 in the same buffer. A peroxidase-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch, West Grove, PA) was used as secondary antibody and was diluted 1:5000 in the same buffer. Positive bands were revealed by enhanced chemiluminescence.

Statistical Analysis.

All results are presented as mean ± S.E.M. Statistical analysis was performed by Student's ttest. A p value of 0.05 was considered to be statistically significant.

Results

We carried out functional experiments in NIH3T3 cells to determine whether the FP and the PGT might be coexpressed in the same cells. PGF_2α (1 nM-3 μM) induced a dose-dependent increase of inositol phosphate with an EC₅₀ value of 52.5 ± 0.15 nM. The maximal response (4.09 ±2.4-fold above the vehicle-treated sample, n = 6 experiments in duplicate), was attained at 1 μM PGF_2α, demonstrating FP expression in NIH3T3 cells, consistent with previous data (Kunapuli et al., 1997). Expression of PGT in NIH3T3 cells was demonstrated measuring uptake of [³H]PGF_2α. Uptake was time-dependent and was inhibited by the anion transporter inhibitor DIDS (Fig. 1).

We developed HEK293 cells stably expressing the hemagglutinin-tagged FP (HA-FP) together with the hPGT (HA-FP/hPGT cells) to study the regulation of hPGT by FP. We also developed cell lines expressing the hPGT alone (hPGT cells) or the HA-FP alone (HA-FP cells). hPGT expression in several clones of HA-FP/hPGT and hPGT cells was verified by immunoblotting. Two clones were selected for further experiments and uptake of [³H]PGF_2α in these clones confirmed hPGT expression. [³H]PGF_2α uptake after 10-min incubation was 159.7 ± 13.0 dpm/μg of protein in HA-FP/hPGT cells, and 27.1 ± 2.4 dpm/μg of protein in hPGT cells; absolute dpm values ranged between 8,000 and 30,000 dpm/ml in HA-FP/hPGT cells, and between 700 and 1700 dpm/ml in hPGT cells.

Untransfected HEK293 and HA-FP cells did not take up PGs. In fact, the low level of radioactivity associated with untransfected cells or with HA-FP cells incubated with [³H]PGE₂ was not diminished by incubation with DIDS. In addition, the radioactivity associated with HA-FP cells was similar to that associated with untransfected HEK293 cells. For example, the radioactivity measured in HEK293 after 20-min incubation with [³H]PGE₂ was 0.57 and 0.45 dpm/μg of protein in the absence and in the presence of DIDS, respectively. In HA-FP cells, we measured 0.66 and 0.71 dpm/μg of protein in the absence and in the presence of DIDS, respectively. In contrast, [³H]PGE₂ uptake at 10 min was inhibited by DIDS 1 mM by 93 to 96% in HA-FP/hPGT cells.

We checked FP expression in HA-FP/hPGT and in hPGT cells. As expected, HA-FP/hPGT cells express the FP, as demonstrated by an increase of inositol phosphate production upon stimulation with PGF_2α for 10 min (Fig.2). By contrast, this response was absent in hPGT cells stimulated for up to 30 min with PGF_2α (Fig. 2).

We measured [³H]PGF_2αuptake in HA-FP/hPGT cells to determine whether stimulation of the FP modifies hPGT activity. Cells were stimulated with different concentrations of PGF_2α, washed, and incubated with [³H]PGF_2α to measure uptake of this prostaglandin. Preincubation of the cells with PGF_2α caused a dose-dependent inhibition of [³H]PGF_2α uptake (Fig.3A). Similar results were obtained when using [³H]PGE₂ in place of [³H]PGF_2α to assess transporter activity (data not shown). Because residual nonradioactive PGF_2α could influence the uptake results, we measured the levels of PGF_2α in the washings by mass spectrometry. Surprisingly, we found residual, detectable PGF_2α even after washing the cells 10 times (7 and 3 nM after incubation of the cells with PGF_2α 100 nM and 1 μM, respectively). Although a dilutional effect by cold PGF_2αmight have confounded our results and cannot be completely excluded, the high levels of PGF_2α after the first wash (15 and 202 nM after incubation with PGF_2α 100 nM and 1 μM, respectively) do not parallel the relatively small inhibition of uptake. In addition, the inhibitory effects of added PGF_2α on transport were similar after washing the cells once versus 10 times, despite a substantial difference in the residual concentrations of this compound at these time points.

We also evaluated uptake of [³H]PGF_2α in HA-FP/hPGT cells pretreated with PGE₂. PGE₂ activates the FP with a potency roughly 50% lower than PGF_2α, as demonstrated by measurement of inositol phosphate production in HA-FP/hPGT cells (4.4-fold increase over basal with 1 μM PGF_2αand 2.3-fold increase with 1 μM PGE₂). On the other hand, PGE₂ is imported by the hPGT with an affinity comparable with that of PGF_2α (Lu et al., 1996). Preincubation of HA-FP/hPGT cells with PGE₂ 1 μM inhibited [³H]PGF_2α uptake less efficiently than the same concentration of PGF_2α([³H]PGF_2α uptake: 75.5 ± 2.3% of vehicle-treated sample with PGE₂ versus 51.4 ± 1.3% with PGF_2α, n = 3 experiments performed in duplicate), indicating that the regulation of the hPGT is receptor-mediated.

Inhibition of [³H]PGF_2αuptake after stimulation of the FP with PGF_2αwas also evident in NIH3T3 cells. These results were poorly reproducible (data not shown), possibly due to the low PGT expression. Indeed, we were able to detect PGT expression in NIH3T3 cells only by ribonuclease protection assay, but not by less sensitive techniques, such as Northern or Western blotting (data not shown).

Because the FP is a GPCR, we investigated the possibility of G protein-dependent regulation of the hPGT by the FP. Immunoblotting revealed that both HEK293 and NIH3T3 cells express at least Gα_s, Gα_i, Gα_q/11, Gα₁₂, and Gα₁₃ (data not shown). To assess whether Gα_s or Gα_i might be involved in hPGT regulation, we repeated uptake experiments in HA-FP/hPGT cells preincubated for 20 h with 250 ng/ml cholera or pertussis toxin. Cholera toxin completely abolished the inhibitory effect of PGF_2α whereas pertussis toxin, in parallel experiments, did not have any effect on the inhibition of transporter activity induced by stimulation of the cells with PGF_2α (Fig. 3A). The effect of cholera toxin on regulation of transporter activity was dose-dependent (Table1) and concentrations higher than 5 ng/ml were necessary to counteract the FP-mediated inhibition of prostaglandin uptake. For example, dose-response curves to PGF_2α in the absence or in the presence of 1 ng/ml cholera toxin were superimposable (data not shown). To test the hypothesis that the effect of cholera toxin under our experimental conditions is due to Gα_s down-regulation (Mochly-Rosen et al., 1988; Chang and Bourne, 1989; Boehm et al., 1996), we measured cyclic AMP levels in cells stably overexpressing the hemagglutinin-tagged PGI₂ receptor (HA-IP cells), which activates adenylate cyclase through Gα_s. Cyclic AMP levels increased from 0.01 ± 0.003 pmol/10⁴ cells (n = 4) to 2.26 ± 0.14 pmol/10⁴ cells (n = 4) in cells stimulated with 10 nM iloprost and not treated with cholera toxin. By contrast, cyclic AMP was not increased above the basal levels in cells pretreated with 250 ng/ml cholera toxin (0.01 ± 0.0006 pmol/10⁴ unstimulated cells versus 0.02 ± 0.003 pmol/10⁴ cells when stimulated with 10 nM iloprost, n = 4). We also measured cyclic AMP levels in HA-IP cells treated with different concentrations of cholera toxin (1–100 ng/ml) and stimulated with 0.1 nM iloprost. Even at 1 and 5 ng/ml, cholera toxin markedly inhibited the agonist-stimulated increase in cyclic AMP (from 69.1 ± 5.5-fold over basal in the absence of cholera toxin to 13.0 ± 0.7- and 10.0 ± 0.02-fold over basal after pretreatment with 1 and 5 ng/ml cholera toxin, respectively).

To determine whether ligation of the FP activates Gα_s, we measured cyclic AMP formation in HA-FP or HA-FP/hPGT cells stimulated for 10 min with different concentrations of PGF_2α. Pretreatment of the cells for 5 min with 0.5 mM IBMX was necessary to detect a cyclic AMP increase under these experimental conditions. PGF_2α, at concentrations that inhibit hPGT-mediated [³H]PGF_2α uptake, increased cyclic AMP levels, irrespective of whether the cells had been pretreated with pertussis toxin. PGF_2αincreased cyclic AMP levels 2.94 ± 0.43- and 3.21 ± 0.72-fold over basal (n = 7) at 100 nM and 1 μM, respectively, in HA-FP cells not pretreated with pertussis toxin (basal = 0.07 ± 0.006 pmol of cyclic AMP/10⁴ cells). The increase in cyclic AMP levels was 3.31 ± 0.28 and 5.45 ± 0.62 fold over basal (n = 9) with PGF_2α 100 nM and 1 μM, respectively, in cells preincubated with 250 ng/ml pertussis toxin (basal = 0.06 ± 0.006 pmol of cyclic AMP/10⁴ cells). An increase in cyclic AMP, even in the presence of pertussis toxin, which inhibits Gα_i, implies that PGF_2αactivates Gα_s to stimulate adenylate cyclase. Enhancement of receptor-mediated stimulation of cyclic AMP in pertussis toxin-treated cells or membrane preparations has already been reported (Hazeki and Ui, 1981; Katada et al., 1982).

We then investigated whether stimulation of the FP with its cognate ligand regulates hPGT through cyclic AMP-dependent protein kinase activation. HA-FP/hPGT cells were preincubated for 30 min with the stable cyclic AMP analog, dBcAMP, or with the cyclic AMP-dependent protein kinase inhibitor, H89. Neither stimulation nor kinase inhibition had any major effect on PGF_2α-mediated inhibition of hPGT activity (Table 2). We verified the effectiveness of H89 as a cyclic AMP-dependent protein kinase inhibitor in in vitro phosphorylation experiments. At 10 μM, H89 completely suppressed histone H1 phosphorylation by cyclic AMP-dependent protein kinase (data not shown). In addition, we did not observe phosphorylation of hPGT in HA-FP/hPGT cells treated with dBcAMP or stimulated with PGF_2α (data not shown) indicating that hPGT regulation is not accompanied by cyclic AMP-dependent protein kinase-mediated phosphorylation and that this kinase does not play an important role in hPGT regulation.

In addition to cells expressing both HA-FP and hPGT, we developed a cell line expressing only the hPGT (hPGT cells) that was used as a negative control. Indeed, hPGT cells do not express the FP to a detectable extent, as reflected by a lack of stimulation of inositol phosphate synthesis by PGF_2α (Fig. 2). Thus, as expected, stimulation of hPGT cells with PGF_2αdid not have any effect on hPGT activity, irrespective of pretreatment with cholera or pertussis toxin (Fig. 3B). In addition, neither H89 nor dBcAMP had any effect on [³H]PGF_2α uptake in hPGT cells (data not shown).

We investigated whether FP stimulation with the isoprostane 8,12-iso-iPF_2α-III would have effects similar to those of PGF_2α on hPGT regulation. This compound, previously known as 12-iso-PGF_2α (Rokach et al., 1997), also activates the FP (Kunapuli et al., 1997). 8,12-iso-iPF_2α-III influenced [³H]PGF_2α uptake only at 100 μM or above (Fig. 4), consistent with its lower potency as an FP ligand than PGF_2α (Kunapuli et al., 1997). 8,12-iso-iPF_2α-III caused a modest increase in cyclic AMP production in HA-FP/hPGT cells at 1 and 10 μM (up to 2.35 ± 0.18 and 1.23 ± 0.17 pmol/well with 10 and 1 μM 8,12-iso-iPF_2α-III, respectively, versus a basal level of 0.55 ± 0.17 pmol/well,n = 3) after 30-min incubation in the presence of IBMX. In parallel experiments, 1 μM PGF_2α increased cyclic AMP up to 3.03 ± 0.4 pmol/well (n = 3), again consistent with the isoprostane acting as a weak FP agonist. Measurement of inositol phosphate production in HA-FP/hPGT cells stimulated for 10 min with 8,12-iso-iPF_2α-III confirmed that this isoprostane is roughly 100 times less potent than PGF_2α in activating the FP. Indeed, inositol phosphate production increased 2.7-fold over basal in cells stimulated with 1 μM PGF_2α, and 2.1- and 2.6-fold over basal in cells stimulated with 10 and 100 μM 8,12-iso-iPF_2α-III, respectively.

We studied HEK293 cells transiently transfected with hPGT and HA-FP to demonstrate that FP-mediated inhibition of hPGT function occurs in cells other than the particular HA-FP/hPGT clone that we selected for the majority of the experiments. Cells were stimulated with PGF_2α for 30 min before measurement of [³H]PGF_2α uptake and PGF_2α-dependent inhibition of the hPGT was again demonstrable (Fig. 5). In parallel experiments, we transiently transfected HEK293 cells with hPGT and HA-IP (Smyth et al., 1996). PGI₂, unlike PGF_2α, is not transported via the PGT, although its receptor also couples to Gs activation (Coleman et al., 1994). HEK293 cells transfected with hPGT and HA-IP were stimulated with the PGI₂ analog iloprost, incubated for 30 min before measurement of [³H]PGF_2α uptake. Iloprost did not have any effect on hPGT activity under these conditions (Fig. 5), despite causing a marked increase in cyclic AMP levels (up to 354.2 ± 47.4 and 235.0 ± 12.5 pmol/well with 100 and 1 nM iloprost, respectively, versus a basal level of 3.77 ± 0.29 pmol/well, n= 3).

We used a Gα_s dominant negative construct (Iiri et al., 1999) and a Gα_s minigene construct (Rasenick et al., 1994; Gilchrist et al., 1999) to implicate further Gα_s in FP-mediated hPGT regulation. The former has been shown to inhibit human chorionic gonadotropin-stimulated cyclic AMP accumulation in transfected COS7 cells (Iiri et al., 1999), whereas the latter inhibits β-adrenergic activation of Gα_s (Rasenick et al., 1994). HEK293 cells were transiently transfected with HA-FP, hPGT, and the construct of interest. Cells were stimulated 2 days later with PGF_2α and incubated with [³H]PGF_2α to evaluate transporter activity.

Although the Gs dominant negative was not expressed at high levels, as determined by immunoblotting, transfection of cells with this construct (Fig. 6) or the Gα_s minigene (Fig.7) blunted the inhibition of transporter activity after FP stimulation. On the other hand, expression of Gα_i or Gα_i random order minigene constructs failed to influence FP-mediated inhibition of transporter activity (Fig. 7).

Discussion

Prostaglandins are charged organic anions at physiological pH, thus they may transverse biological membranes inefficiently (Bito and Baroody, 1975; Kanai et al., 1995; Lu et al., 1996). Although initial hydrolysis and subsequent oxidation rapidly inactivate both PGI2 and thromboxane A₂, PGE₂, PGD₂, and PGF_2α are metabolized intracellularly. Thus, it is assumed that termination of the activity of these PGs in vivo is due to uptake followed by intracellular oxidation (Schuster, 1998). Although a carrier-mediated transport for PGs has been hypothesized since the 1960s and multidrug resistance proteins that are transporters for leukotrienes have been characterized (Cole and Deeley, 1998), a PGT has been cloned only recently (Kanai et al., 1995; Lu et al., 1996). The PGT mediates both the uptake (Lu et al., 1996) and the efflux (Chan et al., 1998) of PGs, thus it may be important both in termination of PG effects, and in extracellular release of newly synthesized PGs for ligation of membrane GPCRs.

Untransfected HEK293 cells and cells transfected only with the HA-FP failed to take up [³H]PGF_2α. This indicates that expression of the FP per se does not influence [³H]PGF_2α uptake, and that possible internalization of the ligand-bound receptor does not contribute significantly to the intracellular levels of [³H]PGF_2α. On the other hand, engagement of the receptor with its cognate ligand dose-dependently regulated import of both PGF_2αand PGE₂ when the hPGT was present.

Given that cold PGF_2α is not completely removed by washing the cells before the incubation with [³H]PGF_2α, we cannot completely exclude a dilutional effect of the residual cold ligand. On the other hand, the amount of residual PGF_2αdoes not correlate with inhibition of uptake, suggesting that a receptor-mediated effect on transporter activity does indeed exist. This hypothesis is strengthened by the following observations: PGF_2α did not influence transporter activity in hPGT cells, which do not express the HA-FP; and PGE₂ regulates the hPGT in HA-FP/hPGT cells with a potency lower than PGF_2α, consistent with the lower potency of PGE₂ as an FP agonist.

Activation of the FP by a structurally distinct ligand, the isoprostane 8,12-iso-iPF_2α-III, also regulated transporter function. Isoprostanes are free radical-catalyzed products of arachidonic acid and may act as incidental ligands at membrane receptors for prostanoids in vivo (Audoly et al., 2000). Although the isoprostane is a less potent FP agonist, both PGF_2α and 8,12-iso-iPF_2α-III activate phospholipase C and adenylate cyclase, presumably via Gq and Gs, respectively. The two ligands can also activate divergent signaling pathways, at least in cardiac myocytes (Kunapuli et al., 1998).

To address the role of G proteins in receptor-transporter interactions, we first examined the effects of pertussis and cholera toxin, probes for Gi and Gs, respectively. Pertussis toxin ADP-ribosylates the α subunit of G proteins of the Gi family when the α subunit is bound to GDP. Thus, pertussis toxin stabilizes Gα_i in the inactive conformation. For this reason, pertussis toxin is widely used as an inhibitor of Gα_i (Simon et al., 1991). Regulation of Gα_s by cholera toxin is complex. ADP ribosylation of Gα_s by cholera toxin stabilizes the GTP-bound conformation of Gα_s and decreases its intrinsic GTPase activity (Casey and Gilman, 1988; Freissmuth et al., 1989). This leads to a persistent activation of this G protein, thereby enhancing adenylate cyclase stimulation and production of intracellular cyclic AMP. Prolonged incubation of the cells with cholera toxin, on the other hand, results in down-regulation of Gα_s with loss of this G protein from the plasma membrane (Mochly-Rosen et al., 1988; Chang and Bourne, 1989; Boehm et al., 1996). When used to probe FP-dependent pathways, cholera toxin, but not pertussis toxin, blocked the effect of FP activation by PGF_2α on transporter function in a dose-dependent manner. These results suggest involvement of Gs in the interaction of the FP with the transporter. Prolonged incubation of cells with cholera toxin has been described to cause Gα_s down-regulation (Mochly-Rosen et al., 1988; Chang and Bourne, 1989; Boehm et al., 1996) and, as a consequence, reduced hormone-stimulated cyclic AMP production (Mochly-Rosen et al., 1988). Consistent with this, we observed that the PGI₂ analog iloprost is less efficacious in stimulating an increase in cyclic AMP in HA-IP cells treated for 20 to 24 h with cholera toxin. Although cholera toxin, at concentrations as low as 1 ng/ml, markedly decreased iloprost-stimulated cyclic AMP, doses of the toxin higher than 5 ng/ml were necessary to counteract the effect of PGF_2α on [³H]PGF_2α uptake in HA-FP/hPGT cells. Although cholera toxin may have effects unrelated to Gα_s, these results distinguish dose-dependent regulation of adenylate cyclase from that of the hPGT. Indeed, interaction of the FP with the transporter seems to be largely independent of its ability to catalyze adenylate cyclase activation. Neither pharmacological activation nor inhibition of cyclic AMP-dependent protein kinase modified FP-mediated regulation of transporter function. In addition, marked activation of the cyclase via Gs coupled to the PGI₂ receptor also failed to regulate transporter function. Finally, ligation of the FP with PGF_2α fails to result in phosphorylation of the transporter, despite regulation of its function.

Although these observations do not support a role for Gs-mediated cyclase activation in the regulation of FP-dependent inhibition of transporter function, additional experiments do support the importance of Gs in this phenomenon. Others have shown that cholera toxin causes persistent activation of Gα_s by inhibiting its intrinsic GTPase activity (Casey and Gilman, 1988; Freissmuth et al., 1989); thereafter, it causes a loss of Gα_s from the plasma membrane (Mochly-Rosen et al., 1988; Chang and Bourne, 1989;Boehm et al., 1996). Thus, the effects of the toxin might be attributable to adenylate cyclase activation, excluded for reasons mentioned above, or be due to Gα_sdown-regulation. Results obtained by specific inhibition of Gα_s with dominant negative constructs support the latter contention. By contrast, a minigene construct directed against Gα_i failed to modify FP-regulated hPGT function, as did pertussis toxin. Although the mechanism through which FP-mediated Gs activation regulates hPGT functions is presently unknown, the kinetics of the reaction suggest that a protein-protein interaction might be responsible for hPGT regulation. Additional experiments would be required to determine whether Gs interacts directly with hPGT, or whether other proteins mediate this interaction.

In conclusion, engagement of the FP by its cognate ligand PGF_2α and, to a much lesser extent, by the isoprostane 8,12-iso-iPF_2α-III dose-dependently inhibits import of PGF_2α, potentially for metabolic inactivation. Should this system operate in vivo, it might serve initially to amplify rapidly the effects of high local concentrations of PGF_2α, which would subsequently result in receptor desensitization. Prostanoids have been shown to augment more gradually their formation by induction of cyclooxygenase-2 (Barry et al., 1997). Although the details by which receptor activation regulates transporter function remain to be elucidated, our data suggest that interaction of the receptor with the transporter is dependent on Gα_s, apparently functioning in a role independent of its capacity to catalyze activation of adenylate cyclase.