Prostaglandin E2 (PGE2) triggers a vast array of biological signals and physiological events. The prostaglandin transporter (PGT) controls PGE2 influx and is rate-limiting for PGE2 metabolism and signaling termination. PGT global knockout mice die on postnatal day 1 from patent ductus arteriosus. A high-affinity PGT inhibitor would thus be a powerful tool for studying PGT function in adult animals. Moreover, such an inhibitor could be potentially developed into a therapeutic drug targeting PGT. Based on structure-activity relationship studies that built on recently identified inhibitors of PGT, we obtained N-(2-(2-(2-azidoethoxy)ethoxy)ethyl)-4-((4-((2-(2-(2-benzamidoethoxy)ethoxy)ethyl)amino)-6-((4-hydroxyphenyl)amino)-1,3,5-triazin-2-yl)amino)benzamide (T26A), a competitive inhibitor of PGT, with a Ki of 378 nM. T26A seems to be highly selective for PGT, because it neither interacts with a PGT homolog in the organic anion transporter family nor affects PGE2 synthesis. In Madin-Darby canine kidney cells stably transfected with PGT, T26A blocked PGE2 metabolism, resulting in retention of PGE2 in the extracellular compartment and the negligible appearance of PGE2 metabolites in the intracellular compartment. Compared with vehicle, T26A injected intravenously into rats effectively doubled the amount of endogenous PGE2 in the circulation and reduced the level of circulating endogenous PGE2 metabolites to 50%. Intravenous T26A was also able to slow the metabolism of exogenously injected PGE2. These results confirm that PGT directly regulates PGE2 metabolism and demonstrate that a high-affinity inhibitor of PGT can effectively prevent PGE2 metabolism and prolong the half-life of circulating PGE2.
Prostaglandins such as PGE2 are autocrine and paracrine lipid mediators that trigger a wide variety of signals (Miller, 2006). These signals are terminated by the metabolism of PGE2 (Ferreira and Vane, 1967; Hamberg and Samuelsson, 1971; Schuster, 1998). Oxidative inactivation of PGE2 is catalyzed by 15-hydroxyprostaglandin dehydrogenase (15PGDH; E.C. 22.214.171.124), which is localized in the cytoplasm (Casey et al., 1980, 1982; Kobayashi et al., 1992). Thus, PGE2 has to be translocated from the outside to the inside of cells for it to be metabolized to PGE2 metabolites (PGE2-Ms), including 15-keto PGE2. The prostaglandin transporter (PGT), a member of the organic anion transporter (OATP) family (Kanai et al., 1995), mediates the energetically active influx of PGE2 (Chi et al., 2006), accounting for almost 100% of PGE2 internalization (Kanai et al., 1995; Chi and Schuster, 2010). In cell culture, we have demonstrated that PGT-mediated PGE2 influx is required and is rate-limiting for PGE2 metabolism (Nomura et al., 2004). In accord with these findings, we have reported that global deletion of PGT in mice results in elevated systemic PGE2 and reduced plasma PGE2-M levels (Chang et al., 2010). Taken together, these data indicate that PGT plays a central role in PGE2 metabolism and signal termination.
As an experimental tool, however, such a genetic approach has limitations, because PGT global knockout mice cannot survive past postnatal day 1 because of patent ductus arteriosus unless they are rescued by administration of indomethacin to the mother before parturition (Chang et al., 2010). A chemical agent that specifically inhibits PGT would be advantageous, both to rapidly assess PGT functions in a variety of different physiological and pathophysiological models and further investigate the role of PGT in controlling PGE2 metabolism and thus signaling.
Previously, we reported a new class of PGT inhibitors obtained by screening a small molecular library (Chi et al., 2006). The best hit of that screening was T34 (Fig. 1) with a binding constant (Ki) of 3.7 μM. Although the Ki of T34 on PG uptake is almost identical to that of bromcresol green (Bito and Salvador, 1976, Kanai et al., 1995), N-(2-(2-(2-azidoethoxy)ethoxy)ethyl)-4-((4-((2-(2-(2-benzamidoethoxy)ethoxy)ethyl)amino)-6-(phenethylamino)-1,3,5-triazin-2-yl)amino)benzamide (T34) has considerable potential to be optimized into a high-affinity inhibitor by virtue of its three side chains, R1, R2, and R3 (Fig. 1). Accordingly, we conducted extensive structure-activity relationship (SAR) studies starting with the T34 platform, which allowed us to obtain N-(2-(2-(2-azidoethoxy)ethoxy)ethyl)-4-((4-((2-(2-(2-benzamidoethoxy)ethoxy)ethyl)amino)-6-((4-hydroxyphenyl)amino)-1,3,5-triazin-2-yl)amino)benzamide (T26A) (Fig. 1), a competitive inhibitor of PGT, with a Ki of 378 nM (10-fold more potent than T34). It is noteworthy that we show that T26A is specific for PGT and in whole-animal studies document the ability of T26A to systemically inhibit PGE2 metabolism.
Materials and Methods
The cell lines used in this study were wild-type MDCK (WT-MDCK) cells that do not express endogenous PGT and MDCK cells stably transfected with the green fluorescent protein-tagged rat PGT (PGT-MDCK) as generated in the laboratory of author V.L.S. (Endo et al., 2002). Tritium-labeled PGE2 ([3H]PGE2) was purchased from PerkinElmer Life and Analytical Sciences (Waltham, MA). Unlabeled PGE2, ovine cyclooxygenase (COX) 1, ovine COX2, human recombinant COX2, and human recombinant microsomal PGE2 synthase (m-PGES) were obtained from Cayman Chemical (Ann Arbor, MI).
Synthesis of Small-Molecule Compounds.
The small-molecule triazine compounds were synthesized following reported procedures (Min et al., 2007).
Inhibitory Effects of Small Molecules on PGT Activity.
These procedures were as described previously (Chi et al., 2006). In brief, PGT-MDCK cells were seeded on 24-well plates. Seventy two hours later, cells were washed twice with chilled Waymouth buffer. Then 25 μl of Waymouth buffer containing either vehicle (DMSO) or small organic compounds was added to each well; this was immediately followed by the addition of 175 μl of Waymouth buffer containing [3H]PGE2. In each well, the total volume of uptake buffer was 200 μl. Organic compounds were first dissolved in DMSO and then diluted in Waymouth buffer. The final concentration of small molecules and [3H]PGE2 in the uptake buffer were 5 μM and 1 nM, respectively. At 10 min, the uptake of [3H]PGE2 was stopped by aspiration of uptake buffer; this was followed by immediate washing twice with 500 μl of chilled Waymouth buffer. Cells were then lysed with 100 μl of lysis buffer containing 0.25% SDS and 0.05 N NaOH. Scintillation solution (1.5 ml) was added to each well, and intracellular [3H]PGE2 was counted by MicroBeta Counter (PerkinElmer Life and Analytical Sciences). The percentage inhibition of [3H]PGE2 uptake by compounds was calculated as [(uptakevehicle − uptakeinhibitor) ÷ (uptakevehicle)] × 100.
Ki Values of Inhibitors.
The procedures were as described previously (Chi et al., 2006). We first measured IC50 of an inhibitor. We then obtained the initial velocities of PGE2 uptake at various initial extracellular concentrations of PGE2 in the presence of various concentrations of the inhibitor. The concentrations of the inhibitor were equal to 1- to 2-fold and 2-to 3-fold the IC50. Ki values were obtained by curve-fitting the reciprocal of initial velocities of PGE2 uptake versus the reciprocal of extracellular PGE2 concentrations at various concentrations of the inhibitors. At low PGE2 concentrations, the extracellular concentrations were taken as 3H-labeled PGE2, which has a specific activity of 500 μCi/mol. At high concentrations of PGE2, we made a mixture of 3H-labeled and unlabeled PGE2 to a final specific activity of 25 μCi/mol.
Ki of T26A to OATPc.
HeLa cells stably transfected with pMEP4-OATPc were obtained from Dr. Allan Wolkoff's laboratory at the Albert Einstein College of Medicine (Bronx, NY). Culture of these cells and induction of OATPc were conducted as described previously (Wang et al., 2003). The initial rates of [35S]sulfobromophthalein (a gift from Dr. Wolkoff's laboratory) uptake by OATPc at various concentrations of [35S]sulfobromophthalein in the presence of various concentrations of T26A were determined by methods similar to those described previously (Chi et al., 2006), and the Ki of T26A to OAPTc was calculated accordingly (Chi et al., 2006).
T26A Effects on Cyclooxygenase Activity.
The effects of T26A on the activities of ovine COX1, ovine COX2, and human recombinant COX2 enzymes were measured by using a kit from Cayman Chemical.
T26A Effects on m-PGES Activity.
DMSO or various concentrations of T26A dissolved in DMSO (2.5 μl) and 5 μl of m-PGES (1 unit/μl) were added to 99.5 μl of reaction buffer containing 100 mM NaH2PO4 at pH 7.2, 0.1% Triton X-100, 1 mM EDTA, and 2.5 mM glutathione. The reaction mixture was incubated at room temperature for 15 min. Eighteen microliters of 283.7 μM (at final concentration of 40 μM) prostaglandin H2 was added to the reaction mixture to initiate the reaction. At 30 s, 5 μl of reaction mixture was taken to 45 μl of quench buffer containing 50 mg/ml stannous chloride in 0.1 N HCl. All of these procedures were conducted at room temperature. Synthesized PGE2 in the reaction buffer was analyzed by using the PGE2 EIA kit from Cayman Chemical.
T26A Effects on PGE2 Synthesis.
Wild-type MDCK cells that do not express endogenous PGT were seeded onto 24-well plates at 30% confluence. Three days later, they were treated with 1 μM arachidonic acid (AA) in the presence of either vehicle (DMSO) or 5 μM T26A at 37°C for various durations. Cell media were collected at various time points. PGE2 concentrations in the media were assayed by an EIA kit from Cayman Chemical.
Extracellular PGE2 and Intracellular PGE2-M.
MDCK cells were seeded onto six-well plates at 30% confluence. Three days later, they were treated with 10 μM bradykinin (to increase endogenous PGE2 synthesis) in the presence of either vehicle (DMSO) or 5 μM T26A at 37°C for various durations. In separate experiments, exogenous PGE2 was added to the medium at the same time as either vehicle or 5 μM T26A was added, and cells were treated for various durations. At various time points, media were collected for measurements of extracellular concentrations of PGE2. Cells were washed with phosphate-buffered saline twice, lysed with 250 μl of phosphate-buffered saline containing 0.1 M HCl and 0.1% Triton X-100 at room temperature for 15 min, and scraped off the plates. Cell suspensions were pipetted up and down for several times to ensure thorough lysing. Cell lysates were collected and centrifuged at 10,000g, 4°C for 10 min. Supernatants were collected. PGE2-M in the supernatants was measured using an EIA kit from Cayman Chemical, which measures the sum of PGE2 metabolites, including 15-keto PGE2 and 13,14-dihydro-15-keto PGE2.
Acute Effects of T26A on PGE2 Metabolism In Vivo.
Sprague-Dawley male rats (Charles River Laboratories, Inc., Wilmington, MA) weighing 300 to 350 g were anesthetized with xylazine (10 mg/kg)-ketamine (50 mg/kg) and were then administered 2000 U heparin (Sigma-Aldrich, St. Louis, MO). After stable anesthesia was obtained, a polyethylene catheter (PE 50; 0.97 mm OK, 0.58 mm i.d.) was advanced into the right ventricle via the right jugular vein for the administration of compounds. A Millar catheter (SPR-249; Millar Instruments, Inc., Houston, TX) was inserted into the right carotid artery and connected to a syringe for blood withdrawal. Thereafter, 200 μl of vehicle (4% DMSO + 4% cremophor) or 25 mM T26A were injected into the jugular vein, 13 min later 1 ml of blood was withdrawn from the carotid artery, and then 100 ng of PGE2 in 200 μl of saline was immediately injected intravenously. The time at which PGE2 was injected was considered 0 min in Fig. 8, C and D. Thereafter, 1 ml of arterial blood was withdrawn at 1, 3, 5, and 10 min. Indomethacin (10 μM) was added to blood immediately after withdrawal to block further PG synthesis, and the blood was immediately centrifuged at 15,000 rpm and 4°C for 15 min. Plasma was collected and kept at −80°C for PGE2 and PGE2-M measurements. PGE2 and PGE2-M were measured using PGE2 and PGE2-M EIA kits from Cayman Chemical.
Structure-Activity Relationship Studies of T34.
As shown in Fig. 1, T34 has three side-chain moieties, R1, R2, and R3, that could be modified. We designed and synthesized the second generation of this class of compounds by varying the R1, R2, and R3 chain length, charge, and aromaticity/aliphaticity. Modification of R1 and R3 did not significantly improve affinity (data not shown). However, changes in R2 profoundly affected the inhibitory activities of the compounds. We therefore focused on further modification of R2. The structure-activity relationships of selected second-generation PGT inhibitors are summarized in Table 1. The percentage inhibition of tracer PGE2 uptake at 5 μM concentration was calculated as [(uptakevehicle − uptakeinhibitor)/uptakevehicle] × 100. Overall, the compounds containing an aromatic benzene ring showed better inhibitory activity than the compounds containing an aliphatic chain in the R2 moiety. The charge of the group on the aromatic ring had significant impact on inhibitory activity (-OH or -H or -F > -NH2), indicating that positive charge decreases inhibition.
Figure 2 shows the effect of chain length on inhibition. We varied the number of carbons from 0 to 10 for the aliphatic R2 and from 0 to 4 for the aromatic R2. When R2 is an aliphatic straight chain ending with a -CH3, the inhibition increases as n increases and reaches the highest level at n of 7; thereafter, the inhibition decreases as n increases further (Fig. 2A). When R2 ends with an aromatic benzene, only a benzyl type of substituent, which has one carbon between the benzene ring and the main scaffold, reduces the inhibitory potency dramatically (Fig. 2B). The inhibitions with n = 0 and 2 carbons between the benzene ring and the main scaffold are almost equal and are higher than all of the others (Fig. 2B). When R2 ends with a phenol group, the inhibition is higher when n = 0 than n = 2 carbons between the benzene ring and the main scaffold (Fig. 2C), implying that -OH enhances inhibition and that rigidity in the R2 group is in favor of inhibition. These results show that the inhibitory activity is sensitive to the structure of the R2 moiety, suggesting that R2 is critical for binding to PGT.
Kinetics of PGT Inhibitors.
We next determined the binding constants of the top three inhibitors listed in Table 1, namely, T26A, N-(2-(2-(2-azidoethoxy)ethoxy)ethyl)-4-((4-((2-(2-(2-benzamidoethoxy)ethoxy)ethyl)amino)-6-(phenylamino)-1,3,5-triazin-2-yl)amino)benzamide (T28A), and N-(2-(2-(2-azidoethoxy)ethoxy)ethyl)-4-((4-((2-(2-(2-benzamidoethoxy)ethoxy)ethyl)amino)-6-((4-fluorophenyl)amino)-1,3,5-triazin-2-yl)amino)benzamide (T25A), by measuring the initial velocities of PGE2 uptake by PGT at various concentrations of PGE2 in the presence of various concentrations of one of these inhibitors. We generated Ki values and determined the type of inhibition of these inhibitors by fitting Lineweaver-Burk curves, as shown in Fig. 3. The resulting Ki values of these three inhibitors are listed in Table 2. They all are competitive inhibitors. T26A has the highest affinity with a Ki of 378 nM, a value reasonably close to the binding constant of the substrate PGE2 to PGT (90 nM) (Kanai et al., 1995).
Modification of T34 as above had revealed that aromaticity was in favor of inhibition. We therefore increased aromaticity of T26A by replacing the phenol with a naphthol group. As shown in Table 3, the Ki of T26A-1 is slightly higher than that of T26A, suggesting that either the interaction between PGT and the inhibitor does not require aromaticity higher than that possessed by the phenyl ring or that the potential higher affinity caused by higher aromaticity of the naphthalene ring was cancelled by the bulkiness of the naphthalene ring.
Comparison of the best three compounds in Table 2 indicates that -OH on the benzene ring enhances binding affinity. To determine whether the position of -OH mattered, we moved -OH from the para to the meta position (T26A-2). Table 3 shows that this reduced inhibition.
The -OH of T26A can be either a donor or a receiver of protons. To address this issue, we replaced it with -OCH3 (T26A-3). The methoxy group of T26A-3 slightly reduced the inhibitory activity compared with T26A, suggesting that this -OH is possibly a proton donor, and therefore the amino acid of PGT that interacts with T26A is possibly a proton receiver.
The native substrates of PGT contain a -COOH at the end of the carbon chain, and PGT functions as an anion transporter (Chan et al., 1998, 2002). Thus, we hypothesized that adding a -COOH could increase binding to PGT substrate sites as well as increasing water solubility. However, a -COOH did not increase affinity (Table 3), nor did it affect solubility (data not shown).
Specificity of T26A.
For any chemical modulator, specificity is always an issue. PGT is an organic anion transporter and belongs to the larger OATP family (Hagenbuch and Meier, 2004). Because it is not feasible to test T26A specificity against all of members of this transporter superfamily, we chose OATPc, which has high homology to PGT (Kanai et al., 1995). We measured the Ki of T26A against this transporter using its substrate sulfobromophthalein (Wang et al., 2003). The Ki of T26A for OATPc is approximately 10 μM, i.e., almost 30-fold higher than that for PGT, suggesting that T26A has high specificity in favor of PGT compared with other organic anion transporters.
T26A on COX Activity.
Another potential concern is whether T26A interacts with components of the PGE2 signaling pathway involved in PGE2 production, such as COX1/2. To directly determine whether T26A had any effect on COX activity, we obtained commercially available ovine COX1, ovine COX2, and human recombinant COX2 enzymes and measured PG synthesis upon addition of AA in the absence and presence of various concentrations of T26A. Figure 4 shows that T26A did not have any significant effects on the activities of those three enzymes, either at a concentration close to the Ki of T26A to PGT or at a concentration 100-fold higher than the Ki of T26A to PGT.
T26A on PGES Activity.
PGE2 synthesis also relies on the activity of PGE2 specific synthases, such as m-PGES. We further determined that T26A did not have significant effect on m-PGES activity (Fig. 5).
T26A on Overall PGE2 Synthesis in MDCK Cells in the Absence of PGT.
As an integrative test of whether T26A affects PGE2 synthesis, and to confirm the results shown in Figs. 4 and 5, we measured PGE2 synthesis/release in response to AA addition in WT-MDCK cells (which lack PGT expression), in both the presence and absence of T26A. As shown in Fig. 6, T26A did not affect PGE2 synthesis and release at any time point, arguing against any additional T26A effects on m-PGES-2 and/or cytosolic PGES.
Together, these results indicate that T26A is a highly selective inhibitor of PGT, which is the influx step for PGE2 that is required for PGE2 intracellular metabolism.
Effects of PGT Inhibition on PGE2 Metabolism in Cell Culture.
We next determined whether PGT inhibition with T26A affects the metabolism of PGE2 generated endogenously in response to bradykinin. We treated either WT-MDCK cells or PGT-MDCK cells with bradykinin and either vehicle or T26A and measured extracellular PGE2 and intracellular PGE2-M. In WT-MDCK cells, the absence of PGT resulted in almost all of the released PGE2 remaining in the extracellular compartment (Fig. 7A). Extracellular PGE2 accumulated as time went on, whereas the amount of PGE2-M in the intracellular compartment remained low (Fig. 7B). In these WT-MDCK cells, T26A did not affect PGE2 metabolism (Fig. 7, A and B), a result consistent with the lack of a PGE2 uptake mechanism in these wild-type cells. In contrast, the expression of PGT at the plasma membrane enabled PGT-MDCK cells to transport newly released medium PGE2 from the extracellular compartment back into the intracellular compartment and deliver it to 15PGDH for subsequent oxidation to PGE2-M (Fig. 7, A and B). Thus, extracellular PGE2 was 2- to 3-fold lower in PGT-MDCK cells than in WT-MDCK cells, whereas intracellular PGE2-M was 2- to 3-fold higher in PGT-MDCK cells than in WT-MDCK cells (Fig. 7, A and B). These results are similar to those previously reported from the laboratory of author V.L.S. (Nomura et al., 2004). It is noteworthy that T26A almost completely blocked PGE2 influx and subsequent oxidation, such that PGE2 was retained in the extracellular compartment at the levels similar to those in WT-MDCK cells (Fig. 7, A and B).
To determine whether PGT is capable of regulating the metabolism of exogenous PGE2, we applied 10 nM exogenous PGE2 to cells, together with vehicle or T26A. As was the case with T26A on endogenous PGE2 metabolism, in WT-MDCK cells almost all of the exogenously added PGE2 remained in the extracellular compartment as intact PGE2 over 3 h (Fig. 7, C and D). In contrast, PGT-MDCK cells eliminated 95% of the added PGE2 from the medium; this resulted in a rise of intracellular PGE2-M (Fig. 7, C and D). With time, extracellular PGE2 gradually increased, whereas intracellular PGE2-M gradually decreased, a result of the pump leak-mediated overshoot of PGE2 influx that we have described previously (Chan et al., 1998; Chi et al., 2006). It is noteworthy that T26A was able to abrogate the PGE2 metabolism caused by PGT (Fig. 7, C and D). These results demonstrate that extracellular PGE2, even at concentrations much higher than endogenous levels, is rapidly internalized by PGT and is subsequently metabolized to PGE2-M, and that this PGE2 metabolism can be effectively prevented by inhibiting PGT with T26A.
T26A on PGE2 Metabolism In Vivo.
Endogenous and exogenous PGE2 are nearly completely metabolized in a single passage through the lung (Ferreira and Vane, 1967; Piper et al., 1970; Dawson et al., 1975; Schuster, 1998), and PGT is very strongly expressed in the lung (Kanai et al., 1995; Lu et al., 1996; Pucci et al., 1999). To assess the ability of T26A to block PGE2 metabolism in vivo, we injected vehicle/T26A into the jugular vein of rats and withdrew blood via the carotid artery. Figure 8A shows that the arterial concentration of endogenous PGE2 in anesthetized rats injected intravenously with T26A peaked at more than 200% that of rats injected with vehicle. Conversely, the concentration of endogenous arterial PGE2-M was reduced 50% by intravenous T26A (Fig. 8B). T26A also slowed the elimination of exogenously added PGE2. As shown in Fig. 8C, compared with rats previously injected with vehicle alone, those injected with T26A achieved a 3- to 4-fold higher arterial PGE2 concentration after an intravenous PGE2 injection. In accord with these results, the concentration of arterial PGE2-M was 3- to 4-fold lower in T26A-treated rats compared with that of vehicle-treated controls.
The present studies describe the development, using structure-function analysis, of a high-affinity inhibitor, T26A, of the PGT. In cultured MDCK cells, T26A blocks PGE2 uptake and its subsequent metabolism to PGE2 metabolites including 15-keto-PGE2. When injected intravenously into rats, T26A raises the plasma level of endogenous PGE2 and reduces the plasma level of the PGE2 metabolites. Rats preinjected with T26A demonstrate impaired metabolism of exogenously administered PGE2. T26A does not affect the activities of purified COX1, COX2, and m-PGES, nor does it affect overall PGE2 synthesis in cultured cells. Thus, T26A increases PGE2 via inhibition of metabolism rather than activation of synthesis.
PGE2 triggers a vast variety of signals including inflammation, vasodilation, and angiogenesis (Weeks, 1972; Clyman et al., 1978; Tsujii et al., 1998). The extent of PG signaling depends, to a large degree, on its concentration at cell-surface receptors, which is determined, in turn, by the relative rates of synthesis and metabolism. Although prostanoid investigators historically focused on synthetic pathways, rapid metabolic clearance of PGs was reported as early as the 1960s (Ferreira and Vane, 1967). Subsequent work demonstrated that PGE2 added to the perfusate of the isolated, perfused lung or kidney disappears within 5 min and reappears as the 15-keto-PGE2 metabolite in the perfusion collections (Bito and Baroody, 1975; Dawson et al., 1975; Anderson and Eling, 1976; Bito, 1976; Bito et al., 1976, 1977; Eling and Anderson, 1976; Eling et al., 1977; Hawkins et al., 1977, 1978). These groundbreaking early studies further revealed that when bromcresol green, a nonspecific organic anion transport inhibitor, was applied to the perfusion system PGE2 appeared intact in the perfusion collections, suggesting that PGE2 metabolism depended on its transport. These earlier studies are consistent with the present data as presented in Fig. 7, C and D, namely, MDCK cells [which express 15PGDH (Nomura et al., 2005)] that have been engineered to also express a PGE2 uptake carrier (PGT) convert PGE2 to the 15-keto metabolite.
Despite this extensive prior literature, the molecular mechanism of PG transport and its role in regulating PG signal termination have been appreciated only recently. We molecularly identified the PGT (Kanai et al., 1995) and postulated that it is responsible for PGE2 uptake before enzymatic oxidation (Schuster, 1998, 2002). We have previously shown in culture cells that PGT is critical for PGE2 metabolism (Nomura et al., 2004), because PGT-mediated PGE2 reuptake is the main pathway for PGE2 influx (Chi and Schuster, 2010) and is a prerequisite for PGE2 intracellular metabolism (Nomura et al., 2004). We have also reported that adult mice rendered null at the PGT locus have elevated systemic (urinary) PGE2 levels and lowered plasma PGE2 metabolite levels compared with wild-type mice (Chang et al., 2010), indicating that PGT plays an important role in regulating PG levels in the circulation. That said, because in those genetic studies PGT was knocked out from the single-cell stage, those studies did not allow us to determine whether the increased PGE2 in plasma was a direct result of PGT deletion or rather was caused by one or more biological modulations that resulted indirectly from PGT deletion. The fact that pharmacological inhibition of PGT with T26A in the present experiments increased plasma PGE2 within several minutes (Fig. 8) indicates that PGT does indeed directly modulate extracellular PGE2 levels. The present results also suggest that pharmacological PGT inhibition is a powerful tool for studying the biological roles of PGT.
The SAR studies described in this article allowed us to obtain an inhibitor, T26A, that is 10-fold more potent than it precursor (Chi et al., 2006). There is some structural resemblance between T26A and PGE2 (Fig. 1, B and C). R2 of T26A resembles the five-member ring in PGE2. The -OH on the ring of PGE2 is important for binding, and it is probably a proton donor, rather than a receiver (Chi et al., 2010). Likewise, the -OH of R2 is important for binding and is probably a proton donor because T26A binds to PGT more strongly than T26A-3 does (Table 3). R1 and R3 possibly resemble the two acyl chains of PGE2. In this regard, replacing the -OH group of R2 with a -COOH did not improve affinity (Table 3), suggesting that a -COOH in the end of either R1 or R3 could possibly improve solubility without adversely affecting affinity.
The results of SAR studies of T34 and T26A provide a basis for speculating about the binding site of PGT. T34 has three moieties, only one of which, R2, is sensitive to inhibition, suggesting that PGT possibly has only one binding site for its substrates. The present studies suggest that the amino acids of PGT that binds to the substrates is probably positively charged, because a negative charge on the inhibitor site increases affinity (Table 1). These results are in accord with our previous studies demonstrating that PGT is inhibited by disulfonic stilbenes, niflumic acid, and the thiol reactive anion MTSES [Na(2-sulfonatoethyl)methanethiosulfonate] (Chan et al.,1998); by our cysteine-scanning mutagenesis and molecular modeling of putative transmembrane 10, which indicated that the substrate binding of PGT is formed among its membrane-spanning segments, with four residues along the cytoplasmic end of helix 10 contributing to one surface of the binding site (Chan et al., 1999); and by our site-directed mutagenesis studies indicating a critical role for arginine 560 in substrate translocation (Chan et al., 2002).
In addition, we hypothesize that the docking pocket for T26A probably has limited space to accommodate a substrate, because replacing phenol with naphthol did not increase the affinity (Table 3), even though naphthol has higher aromaticity than phenol, which is presumably in favor of binding (Table 1).
All small-molecule inhibitors suffer from possible lack of specificity. We have addressed the issue of specificity by examining the affinity of T26A for a closely related gene family member, OATPc (SLCO1B1) that has high homology with PGT (Kanai et al., 1995). SLCO1B1 is a particularly relevant homolog for this study because it mediates the hepatic uptake of various drugs, including most statins and statin acids (König et al., 2006), and because genetic variations in the SLCO1B1 gene have been associated with both statin-associated myopathy and reductions in the lipid-lowering effect of statins (SEARCH Collaborative Group, 2008). The 10-fold weaker binding of T26A to OATPc compared with PGT suggests that T26A is a selective inhibitor of PGT within the OATP gene family.
In addition, we determined whether T26A directly interacted with synthases of PGE2. Insignificant effects of T26A on the activities of COX1, COX2, and m-PGES (Figs. 4 and 5) eliminated the concern that increased PGE2 by T26A could be caused by activation of those synthases. This was further confirmed by the fact that T26A did not affect overall PGE2 production in cells that did not express PGT (Fig. 6). That T26A does not affect PGE2 synthesis is also supported by the in vivo data showing that endogenous serum PGE2 levels were more than doubled within short period of time (15 min) after T26A injection (Fig. 8A).
In addition to being a powerful basic research tool for investigating the fundamental role of PGT in PG metabolism and signal termination, a potent PGT inhibitor such as T26A provides a starting point for the development of therapeutic agents targeting PGT. Because PGE2 and other PGs trigger a vast array of beneficial physiological events, and PGT regulates the metabolism of PGs, a specific inhibitor of PGT could potentially be developed for clinical applications, such as open-angle glaucoma (Parrish et al., 2003) and pulmonary hypertension (Gomberg-Maitland and Olschewski, 2008), among others.
Participated in research design: Chi and Schuster.
Conducted experiments: Chi and Jasmin.
Contributed new reagents or analytic tools: Chi, Min, Lisanti, and Chang.
Performed data analysis: Chi, Jasmin, and Schuster.
Wrote or contributed to the writing of the manuscript: Chi, Min, and Schuster.
This work was supported by the National Institutes of Health National Institute of Diabetes and Digestive and Kidney Diseases [Grant 5R01-DK049688]; and the American Heart Association [Grant 0830336N].
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- prostaglandin E2
- PGE2 metabolite
- microsomal PGE2 synthase
- 15-hydroxyprostaglandin dehydrogenase
- prostaglandin transporter
- organic anion transporter
- structure-activity relationship
- Madin-Darby canine kidney
- wild type
- arachidonic acid
- dimethyl sulfoxide
- enzyme immunoassay
- Received March 8, 2011.
- Accepted August 16, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics