The identification of prostanoid receptors in tissues is greatly facilitated by the use of competitive receptor antagonists. However, while potent and selective blockers for prostaglandin D (DP), prostaglandin E subtype-1 (EP1), and thromboxane (TP) receptors have been developed, 1 antagonists for the other prostanoid receptors including the prostacyclin (IP) receptor remain elusive. A partial agonist for the IP receptor could be a potential lead to an IP antagonist and the current paper addresses whether the prostacyclin analogue taprostene 2,3 is such an agent (Fig. 1).
Prostacyclin partial agonists are of further interest since they may offer a clinical profile different to IP full agonists (eg, prostacyclin and iloprost) 4 currently used in the management of thrombotic disorders and pulmonary hypertension. Inhibition of platelet activation and reduction of pulmonary vascular resistance are two mechanisms potentially underlying these therapeutic effects. However, a major drawback has been hypotension and syncope arising from systemic vasodilatation. Inhibition of smooth muscle proliferation is a third mechanism that may be important in the management of pulmonary hypertension. 5 In this context, Clapp et al 6 found that the prostacyclin analogues, cicaprost, iloprost, beraprost, and UT-15, had different maxima for both inhibition of proliferation of human pulmonary artery smooth muscle cells and the associated activation of adenylate cyclase. Whether these profiles resulted from different agonist efficacies at the IP receptor was not investigated.
Recently, we reported that taprostene does not completely relax ring preparations of pig carotid artery and saphenous vein and rabbit mesenteric artery and saphenous vein contracted with phenylephrine. 7 However, it was difficult to assign partial agonist status to taprostene from its interaction with an IP full agonist (AFP-07, 8 TEI-9063, 9 or cicaprost 10 (Fig. 1)) since the latter agents also activated sensitive prostanoid EP4 (relaxant) and/or EP3 (contractile) systems in the preparations. Vascular EP4 systems, as opposed to EP2 systems, are noted for their high agonist sensitivity. 11 Moreover, vascular EP3 systems strongly synergize with α1-adrenoceptor and TP receptor systems, 12 thereby enhancing the actions of phenylephrine and U-46619, respectively, when they are used to induce tone.
In the current study, we have re-investigated the potential IP partial agonism of taprostene. Our strategy had 3 elements. Firstly, vascular preparations with low complements of EP3 and EP4 receptors were sought, although this was not entirely successful. Secondly, we hoped to exploit possible differences in the agonist activities of prostacyclin analogues on EP4 receptors from different species. In recombinant systems, for example, cicaprost binds with moderate affinity to the human EP4 receptor (Ki = 44 nM), 13 but has minimal affinity for the mouse EP4 receptor (Ki >10,000 nM). 14 Also, AFP-07, the most potent IP agonist reported so far, binds well to the mouse EP4 receptor (Ki ~20 nM). 8 Thus, on guinea-pig saphenous vein, the first of our new preparations, we found that the EP4 antagonist AH 23848 did not block the relaxant actions of cicaprost and AFP-07, even though a sensitive EP4 system was present. AH 23848 must be used cautiously, since its pA2 for the EP4 receptor is only 5.4, 11 and its selectivity may be low based on radioligand binding studies with recombinant prostanoid receptors. 13 Thirdly, the specificity of taprostene for the IP receptor was determined through its interaction with other relaxant agonists. The study therefore consisted of investigations of the activity of taprostene on 3 preparations, guinea-pig saphenous vein, rat tail artery, and mouse thoracic aorta, utilizing 3 IP full agonists, AFP-07, TEI-9063, and cicaprost, and as appropriate PGE2 as an EP4 agonist, ONO-AE1-259 (Fig. 1) as a selective EP2 agonist, 15,16 isoprenaline as a β-adrenoceptor agonist, and acetylcholine as a muscarinic agonist.
EP3 receptors were identified in the vascular preparations from the agonist potency ranking SC-46275 > sulprostone > 17-phenyl-ω-trinor PGE2 (17-phenyl PGE2) 17; SC-46275 is highly selective for the EP3 receptor, 18 while 17- phenyl PGE2 has highest potency on the EP1 receptor. 19 In addition, we have investigated whether taprostene activates prostanoid DP receptors, since its ω-chain is similar to that of some DP agonists with vasodilator properties, for example BW 245C (Fig. 1) 20,21 and several 5-thia-prostacyclin analogues. 22 BW 245C and the selective DP antagonist BW A868C, which has a pA2 of 8.7, 23 were used for this purpose.
Experimental procedures involving live animals were performed under license issued by the Government of the Hong Kong SAR and endorsed by the Animal Experimental Ethics Committee of the Chinese University of Hong Kong. Male Dunkin-Hartley guinea-pigs (400–500 g), male Sprague-Dawley rats (200–250 g), and male Balb/c mice (25–30 g) were killed by cervical dislocation and exsanguination.
Isolated Smooth Muscle Preparations
The lateral saphenous vein from guinea-pig, proximal tail artery from rat, and descending thoracic aorta from mouse were excised and placed in Krebs-Henseleit solution: NaCl 118, KCl 4.7, CaCl2 2.5, MgSO4 1.2, KH2PO4 1.18, NaHCO3 25, glucose 10 mM. Adherent connective tissue was removed. Four or eight vessel rings (3 mm long) were cut and suspended using 50 μm steel wire in 1 or 2 Danish Myo Technology Model 619M myographs for tension recording. The tissue baths (8.0 mL) contained Krebs-Henseleit solution at 37°C, aerated with 95% O2/5% CO2 to maintain pH at 7.4; washout was by draining and replacement of bathing solution from a syringe. Tension signals were relayed to a MacLab 4 amplifier and saved to a Macintosh PowerMac computer system (sampling rate 40/min). In all experiments, the bathing solution contained 1 μM indomethacin to suppress endogenous prostaglandin production.
During an equilibration period of 1 hour, the vessel rings were stretched until the resting tension held steady at 0.4 g for saphenous vein and 1.0 g for both tail artery and aorta. The preparations were then checked for reproducible contractile responses to 40 mM KCl; the rare preparation that showed weak responses was discarded. The basic protocol involved 2 dosing sequences (S1 and S2), each lasting 90 minutes and consisting of addition of common antagonist/channel blocker, followed by tone-inducing agent, followed by cumulative dosing (1, +3, +10 ….) with prostanoid or non-prostanoid agonist. During S2, the antagonist of interest and its vehicle were applied 15 minutes before the tone-inducer to test and control preparations, respectively. In the taprostene interaction experiments, taprostene (3 μM) and its vehicle were applied after the tone inducer in S2 to test and control preparations, respectively. Tone was induced by 3 μM phenylephrine on guinea-pig saphenous vein, 300 nM phenylephrine, or 100 nM U-46619 on rat tail artery, and either 300 nM phenylephrine or a combination of 50 nM phenylephrine and 300 nM sulprostone on mouse aorta. For guinea-pig saphenous vein only, S1 was preceded by a priming sequence of PGE2 (0.01–0.44 nM); these data were not used in dose-ratio calculations. In each experiment on rat tail artery and mouse aorta, cumulative doses of saline were applied to obtain time-control data, and the corresponding mean responses for each set of experiments were used to correct control and test agonist responses. Time-control sequences were not obtained on guinea-pig saphenous vein since tone induced by phenylephrine was highly stable.
In each sequence, responses were measured as a percentage of the generated tone and expressed as mean ± SEM, where n is the number of preparations used (each from a different animal). GraphPad Prism software was used to fit sigmoidal curves to both individual and mean log concentration-response data according to the equation:
in which A and IC50 are the molar concentrations of agonist eliciting a given response and a 50% maximal response, respectively and n is the slope factor (Hill factor). Some of the ONO-AE1-259 concentration-response data were fitted with Prism's 2-site competition equation. The low concentration-asymptote was constrained to 100%, except for control agonist / fixed taprostene data, when it was set to the response to taprostene. A quoted pIC50 value is the negative logarithm of the IC50 value obtained by curve-fitting, while a pIC value at a particular relaxation level was read from fitted or interpolated lines using the on-screen measuring device in Prism. When control agonist and agonist / fixed taprostene curves had similar maxima (Student paired t test), the method of Pöch and Holzmann 24 was used to calculate a theoretical addition curve for the latter. The response to 3 μM taprostene was converted into a concentration of interacting agonist (A) using the control curve, the concentration of the interacting agonist in the presence of taprostene (A') was then added, and the addition response corresponding to A + A' on the control curve was obtained and plotted against A'. For control agonist and agonist/fixed taprostene curves with different maxima, a suitable model for calculating an addition curve, and particularly the maximum response, was not obvious. A theoretical addition curve for the taprostene interaction was derived from the mean responses using the functional synergism equation of Ariens et al 25, which employs fractions of the maximum tissue response (100% relaxation). In each case, experimental and derived curves were similar indicating additivity of action. However, these derived curves have not been plotted on the figures, since we had doubts on the validity of using 100% relaxation in the formula. Instead, a paired t test was applied to pIC50 values for control agonist and agonist / fixed taprostene curves; P > 0.05 equates to “no antagonism” in Table 2.
Descriptive statistics and t tests were performed using GraphPad Prism software; ± error terms represent SEM. A dose-ratio (geometric mean) was considered statistically significant when its 95% confidence interval (95% CI) excluded 1.0. SuperANOVA software (Abacus Concepts Inc., USA) was used to perform ANOVA coupled with pre-planned contrasts of cell means. 26 All tests were two-tailed and the significance limit was set at P = 0.05.
Drugs and Solutions
The following compounds were gifts: taprostene from Grünenthal GmbH, Germany; AFP-07 (7,7-difluoro-16S,20-dimethyl-18,19-didehydro PGI2) from Asahi Glass Co., Japan; TEI-9063 (17S,20-dimethyl-Δ6,6a-6a-carba PGI1) from Teijin Co., Japan; cicaprost and sulprostone from Schering AG, Germany; ONO-AE1-259 (16R-9β-chloro-19,20-didehydro-9,15-dideoxy-16-hydroxy-17,17-trimethylene PGE2) from ONO Pharmaceuticals, Japan; AH 23848 (rac-[1α(Z),2β, 5α]-7-[5-([1,1′-biphenyl]-4-ylmethoxy)-2-(4-morpholynyl)-3-oxocyclopentyl]-4-heptenoic acid), GR 32191 (9α-(biphenyl)-methoxy-11β-hydroxy-12β-(N-piperidinyl)-ω-octanor-prost-4Z-enoic acid) and BW A868C ((±)-3-benzyl-5-(6-carboxyl)-1-(2-cyclohexyl-2-hydroxyethylamino)-hydantoin) from GlaxoSmithKline Research & Development, UK; SC-46275 (methyl 7-[2β-[6-(1-cyclopenten-1-yl)-4R-hydroxy-4-methyl-1E,5E-hexadienyl]-3α-hydroxy-5-oxo-1R,1α-cyclopentyl]-4Z-heptenoate) from GD Searle, USA. The following compounds were purchased: BW 245C (5-(6-carboxyhexyl)-1-(3-cyclohexyl-3-hydroxypropyl)hydantoin), PGE2, fluprostenol, 17-phenyl-ω-trinor PGE2, U-46619 (11,9-epoxymethano PGH2), and SC-51322 (8-chlorodibenz[b,f] [1,4]oxazepine-10(11H)-carboxylic acid, 2-[3-[furanyl-methyl)-thio]-1-oxopropyl]hydrazide) from Cayman Chemical Co., USA; acetylcholine chloride, indomethacin, nifedipine, and phenylephrine hydrochloride from Sigma-Aldrich, USA.
Primary stock solutions (concentrations for free acid/base) were prepared as follows: prostanoid agonists (5–10 mM), in-domethacin (20 mM), BW A868C (10 mM), SC-51322 (10 mM), and nifedipine (10 mM) in absolute ethanol; phenylephrine (100 mM) in distilled water; GR 32191 (1 mM), isoprenaline (10 mM), and acetylcholine (1 mM) in 0.9% sodium chloride solution (saline). The stock solution of AH 23848 (2 mM) was prepared by dissolving the calcium salt, Ca(acid)2, in 10% NaHCO3 solution and then diluting 10-fold with saline; aliquots were stored at −20°C and a new aliquot was used for each day's experiments. All sub-stocks were prepared with saline, except for SC-51322, which was diluted with ethanol since it has a propensity to crystallize in aqueous solution.
Prostanoid receptors mediating contraction or relaxation of ring preparations of guinea-pig saphenous vein, rat tail artery, and mouse thoracic aorta were first characterized using a range of agonists, the EP1 antagonist SC-51322, 27 the EP4 antagonist AH 23848, and the DP antagonist BW A868C. (Table 1 shows dose-ratios for the latter two agents.) The interactions of taprostene with selected relaxant agents were then examined on each preparation (Table 2). Although the prostanoids investigated are known to have minimal agonist activity at TP receptors, it was important to exclude this action completely. The potent and selective TP receptor antagonist GR 32191 28 was therefore used routinely on guinea-pig saphenous vein and mouse thoracic aorta preparations at concentrations of 200 and 500 nM, respectively, and on rat tail artery at 500 nM except when tone was induced by the TP agonist U-46619. These GR 32191 concentrations abolished ~70% maximal responses induced by 100 nM U-46619, but had no effect on matching responses to phenylephrine.
Guinea-Pig Saphenous Vein
Identification of Prostanoid Receptors
AFP-07, TE1-9063, and cicaprost induced ≥ 90% relaxation of the phenylephrine-contracted saphenous vein, whereas the maximum response of taprostene was lower (26–61%, n = 6) (Fig. 2A). AH 23848 (30 μM) produced small right-shifts of the log concentration-response curves of all 4 agonists, which were statistically significant (95% CI of the dose-ratio excludes unity); the vehicle did not produce significant shifts (Table 1). Figure 2B shows log concentration-response curves in the presence of AH 23848; AFP-07 (pIC50 = 8.65) was 3.0 and 17 times more potent than TEI-9063 and cicaprost, respectively, consistent with previously reported data for IP receptor systems in vascular smooth muscle. 7 Taprostene was the least potent of the prostacyclin analogues (pIC50 = 7.05).
EP2 and EP4 receptors also mediate relaxation of the saphenous vein based on the following findings. Firstly, relaxation induced by the selective EP2 agonist ONO-AE1-259 (pIC range for 40% relaxation = 7.2–8.2 for S1) was not blocked by 30 μM AH 23848, whereas the log concentration-response curve for PGE2 (pIC50 = 9.75 for S1) underwent a parallel right-shift of nearly one log unit in the presence of 30 μM AH 23848 (Table 1). Application of the Schild equation to the PGE2 data gave a pA2 value of 5.4 for AH 23848, which is identical to the value obtained on piglet saphenous vein, the archetypal EP4 preparation. 7,11 Secondly, sensitivity to PGE2 remained constant throughout the experimental period (for 40% relaxation, pIC = 9.68 ± 0.18, 9.74 ± 0.17, and 9.71 ± 0.17 for priming sequence, S1 and S2, respectively, n = 8), whereas responses to ONO-AE1-259 became smaller on repeated exposure (S1/S2 dose-ratio = 2.22, BW A868C control in Table 1). In separate experiments, exposure of the saphenous vein to 5 μM ONO-AE1-259 for 30 minutes during S1 resulted in marked desensitization to cumulative application of the same agonist in S2, while curves for PGE2 on matching preparations were unaffected (Fig. 2C). A standard PGE2 sequence was applied in S1 to both PGE2 control and ONO-AE1-259 control preparations. Some of the control and all of the desensitization curves for ONO-AE1-259 appeared to be biphasic, whereas monophasic curves were always seen in the presence of 30 μM AH 23848 (S1) (Fig. 3C). Thus, it is likely that EP2 and EP4 receptors in the saphenous vein are activated by ONO-AE1-259 at concentrations > 1 and > 500 nM, respectively.
BW A868C at 10 μM marginally antagonized relaxation induced by both PGE2 and the selective DP agonist BW 245C (Table 1), the latter having a low potency (for 40% relaxation, pIC = 6.33; 71–92% relaxation at 5.5 μM, n = 7). Relaxations induced by taprostene and ONO-AE1-259 were not blocked by 10 μM BW A868C (Table 1). The authenticity of our sample of BW A868C was confirmed by its ability at 0.1 μM to inhibit markedly relaxation of rabbit saphenous vein induced by BW 245C (pIC50 = 7.42, n = 3, data not shown), in agreement with the results of Lydford et al 29.
Evidence for EP1, EP3 or FP contractile systems in guinea-pig saphenous vein was not found. Under phenyleph-rine-induced tone, both 17-phenyl-ω-trinor PGE2 (1–1000 nM), a moderately selective EP1 agonist, and SC-46275 (10–5000 nM), a highly selective EP3 agonist, induced relaxation, with log concentration-response curves parallel to that of PGE2. PGE2 was 390 and 510 (n = 2) and 1520, 1570, and 2630 (n = 3) times more potent, respectively, similar to values found for these analogues on sensitive EP4 systems in other blood vessel preparations. 7 Sulprostone (EP3 > EP1) was also a low potency relaxant agent, with a shallow, almost linear log concentration-response curve: threshold responses were seen at 10 nM and 37 to 65% relaxation at 5.5 μM (n = 4). Fluprostenol, a potent and selective FP receptor agonist, behaved similarly to sulprostone, producing threshold responses at 10 nM and 29 and 42% relaxation at 5.5 μM (n = 2).
Interactions of Taprostene with Other Relaxant Agents
In S1, control and test preparations were exposed to cumulative doses of AFP-07, TEI-9063, cicaprost, PGE2, or acetylcholine. A standard PGE2 sequence was substituted for ONO-AE1-259 in S1 to minimize desensitization of the EP2 receptor system. In S2, control and test preparations were treated with vehicle and 3 μM taprostene, respectively, followed by the appropriate cumulative relaxant. AH 23848 (30 μM) was present during both sequences. Taprostene caused significant right-shifts of the log concentration-response curves for AFP-07, TEI-9063 (Fig. 3A), and cicaprost in parallel to the corresponding theoretical addition curves (see Methods, Table 2). The log dose-ratios for the 3 prostacyclin analogues were not significantly different (P > 0.05 for each pairwise comparison, 1-factor ANOVA). In contrast, taprostene did not oppose the relaxant action of PGE2 (Fig. 3B); the dose-ratio of 0.54 (95% CI = 0.29–1.01) is close to being significantly different from unity, and further experiments are required to determine if there is any synergism between taprostene and PGE2. The maximum response for the ONO-AE1-259 / 3 μM taprostene combination was significantly greater than the control ONO-AE1-259 maximum (87 ± 2.4 versus 77 ± 3.1%; P < 0.01, paired t test), whereas there was no significant change in pIC50 (7.76 ± 0.14 versus 7.89 ± 0.07, P > 0.05) (Fig. 3C). Although theoretical addition curves were not plotted (see Methods for an explanation), it was clear that taprostene did not oppose the relaxant action of ONO-AE1-259. The same profile was seen with acetylcholine / 3 μM taprostene: maximum relaxation = 95 ± 0.8 versus 86 ± 3.0% (P < 0.05); pIC50 = 7.56 ± 0.06 versus 7.44 ± 0.16 (P > 0.05) (taprostene response = 50 ± 5.3%, n = 6).
Using similar protocols, AFP-07 and a fixed concentration of 0.1 nM PGE2 (12 ± 1.3% relaxation, n = 6) showed slight synergism (95% CI for dose-ratio = 0.66–0.85), while AFP-07 and 0.3 nM PGE2 (33 ± 1.4% relaxation, n = 6) behaved additively (95% CI for dose ratio = 0.73–1.39; data not shown). Also, the maximum response to ONO-AE1-259 in the presence of a fixed concentration of 1.25 nM AFP-07 (relaxation = 20 ± 1.5%, n = 5) was greater than its control maximum (81 ± 4.1 versus 74 ± 1.6%, P < 0.05), with no significant change in pIC50 (7.86 ± 0.09 versus 7.68 ± 0.04, P > 0.05, data not shown).
Rat Tail Artery
Identification of Prostanoid Receptors
In initial experiments involving tone induced by either phenylephrine (100 nM) or U-46619 (50 nM), it was often difficult to quantitate relaxant responses due to the appearance of slow oscillations as the tail artery preparation relaxed. Nifedipine at 100 nM abolished these oscillations with some reduction in the tonic contraction. It was decided to generate tone with 100 nM U-46619 in the presence of 100 nM nifedipine, since there was less fade than with the phenylephrine/nifedipine combination. As shown in Figure 4A, TEI-9063 induced greater maximum relaxation than AFP-07 and cicaprost (96–111% versus 78–100 and 70–95%, respectively). A similar profile was also found in a later series of experiments (n = 6). Weak relaxation was induced by both taprostene (pIC50 = 6.59, maximum relaxation = 27–34%, n = 6) and PGE2 (pIC50 = 8.28, maximum relaxation = 32–70%, n = 6), while ONO-AE1-259 (10 nM–10 μM) induced minimal relaxation. Antagonism by AH 23848 was not studied owing to its ability to block TP receptors and hence the tone induced by U-46619. 17-Phenyl PGE2 (1.0–444 nM), sulprostone (10–440 nM), SC-46275 (1–1440 nM), and fluprostenol (10–1444 nM) only elicited small transient contractions at the highest concentration in the cumulative sequence indicating that EP1, EP3, and FP receptors are unlikely to be present. In addition, 17-phenyl PGE2 (3–333 nM) elicited no further relaxation in preparations partially relaxed with 50 nM AFP-07.
In view of the different maxima shown by the prostacyclin analogues and because we wished to block EP4 receptors with AH 23848, we again attempted to use phenylephrine (300 nM) as the tone-inducer (in the presence of 100 nM nifedipine and 500 nM GR 32191); the fade was found to be acceptable in these experiments (~10% over 40 minutes). In S1, AFP-07, TEI-9063, and cicaprost induced similar maximum relaxations (71–80, 74–82, and 74–81%, respectively, n = 4); in S2, treatment with 30 μM AH 23848 caused small but significant shifts of their log concentration-response curves (Table 1) and slightly reduced their maximum relaxations (by 5, 6, and 11%, respectively). In the presence of AH 23848, AFP-07 (pIC50 = 7.99) was 3.4 and 13.5 times more potent than TEI-9063 and cicaprost, respectively (Fig. 4B), again indicating the presence of an IP receptor system. Relaxation to 3 μM taprostene ranged from 18 to 31% (n = 6). PGE2 again induced weak relaxation, and its log concentration-response curve was shifted to the right in a parallel manner (~1 log unit) by 30 μM AH 23848 (Table 1). The pA2 value for AH 23848 was calculated to be 5.7, consistent with the presence of an EP4 receptor. ONO-AE1-259 (10–3000 nM) and BW 245C (10–3000 nM) induced less than 10% relaxation, indicating that EP2 and DP relaxation systems are unlikely to be present.
Interactions of Taprostene with Other Relaxant Agents
Under the U-46619 / nifedipine protocol, 3 μM taprostene antagonized relaxation induced by AFP-07, TEI-9063, and cicaprost (Table 2) (Fig. 4). However, log dose-ratios for AFP-07 and cicaprost were significantly larger than that for TEI-9063 (P < 0.01 for both pairwise comparisons, 1-factor ANOVA). The AFP-07 / taprostene curve was located parallel to the addition curve (Fig. 4C), whereas the cicaprost / taprostene curve appeared to approach a higher maximum (Fig. 4E). The maximum response for the PGE2 / 3 μM taprostene combination was significantly greater than the control PGE2 maximum (55 ± 4 versus 44 ± 6%, P < 0.05), with no significant change in pIC50 (8.31 ± 0.06 versus 8.27 ± 0.07, P > 0.05, n = 6;Fig. 4F), indicating no antagonism by taprostene. Isoprena-line had a bell-shaped curve (0–10, 15–58, and −10–50% relaxation at 12.5, 550, and 1800 nM, respectively, n = 6), and also appeared to interact additively with taprostene (data not shown).
Under the phenylephrine / nifedipine / GR 32191 / AH 23848 protocol, the log concentration-response curve for AFP-07 was shifted to the right in parallel with the addition curve; the dose-ratio was similar to the corresponding U-46619 / nifedipine value (Table 2). Again, the maximum response for the PGE2 / 3 μM taprostene combination was significantly greater than the control PGE2 maximum (52 ± 2 versus 43 ± 2%, P < 0.01), with no significant change in pIC50 (7.62 ± 0.21 versus 7.59 ± 0.16, P > 0.05, n = 6), indicating no antagonism by taprostene. At this time, our stock of AH 23848 was almost exhausted, and consequently the interactions of taprostene with TEI-9063 and cicaprost were not examined.
Mouse Thoracic Aorta
Identification of Prostanoid Receptors
On mouse aorta preparations contracted with 300 nM phenylephrine in the presence of 500 nM GR 32191, PGE2 (12.5–550 nM, S1 or S2) invariably induced further contraction, as did the PGE analogues, SC-46275, sulprostone, and 17-phenyl PGE2 (Fig. 5A). The SC-46275 curve was constructed from single concentrations applied to separate preparations since responses were slow to reach a stable level. The high potency of SC-46275 and the potency ranking SC-46275 > sulprostone > 17-phenyl PGE2 are indicative of an EP3 as opposed to an EP1 receptor system, as described in the Introduction. The absence of an EP1 system is also supported by the inability of the EP1 antagonist SC-51322 to block contractile responses to sulprostone and I7-phenyl PGE2. Mean responses at each sulprostone concentration (0.125, 1.375, 13.88, 139 nM, S2) under vehicle and 1.25 μM SC-51322 treatments were not significantly different (P> 0.05 for pairwise comparisons at each concentration, repeated-measures 2-factor ANOVA, n = 5). Similar results were obtained for 17-phenyl PGE2 (3.75, 16.25, 53.75, 179 nM, S2), except for highest concentration (554 nM), where the SC-51322 mean was significantly greater than the vehicle mean (P< 0.01, n = 6). Sulprostone (1–1000 nM) had no effect on the resting tone of the aorta, indicating a silent EP3 receptor system that exhibits strong synergism with the α1-adrenoceptor system. ONO-AE1-259 (12.5–5550 nM) (Fig. 5A) induced only weak contraction (<10%).
As shown in Figure 5A, relaxation curves for cicaprost and taprostene were simple sigmoids, whereas AFP-07, and TEI-9063 had bell-shaped curves. We supposed that the latter profiles were due to EP3 agonism opposing IP agonism as the AFP-07 / TEI-9063 concentration increased. In an attempt to moderate this EP3 agonist activity, tone was generated with a low concentration of phenylephrine (50 nM) and a near-maximal concentration of sulprostone (300 nM) (in the presence of 500 nM GR 32191). The log concentration-response curves for AFP-07, TEI-9063, and cicaprost now became simple sigmoids with similar maxima (57–80, 51–75, and 55–78%, respectively, all n = 8); AFP-07 (pIC50 = 7.81) was 2.5 and 8.5 times more potent than TEI-9063 and cicaprost, respectively. This strategy also uncovered a relaxant action of both PGE2 (Fig. 5C) and ONO-AE1-259, although the latter action was only observed at high concentrations (1.8 and 5.55 μM; 12 ± 4 and 22 ± 5%, respectively, n = 6). SC-51322 (1.25 μM) had no significant effect on the PGE2 relaxation curve (n = 5, P> 0.05 for pairwise comparisons at each agonist concentration, repeated-measures 2-factor ANOVA, data not shown). The selective FP agonist fluprostenol (12.5–5550 nM) produced weak contraction (~10%) at the higher concentrations only.
Interactions of Taprostene with Relaxant Agonists
Under phenylephrine-induced tone, 3 μM taprostene insurmountably antagonized relaxation induced by AFP-07, TEI-9063, and cicaprost, with dose-ratios usually exceeding 30 (data not shown). In contrast, surmountable antagonism was seen under the phenylephrine / sulprostone protocol (eg, AFP-07 in Figure 5B). Log dose-ratios for the 3 full agonists were not significantly different (P> 0.05 for each pairwise comparison, 1-factor ANOVA), although the variation in the cicaprost data was high (Table 2). TEI-9063 is a potent EP1 agonist, unlike cicaprost. 30 However, 1.25 μM SC-51322 had no effect on either the relaxant activity of TEI-9063 (SC-51322 treatment: IC50 = 7.31, maximum relaxation = 65–76%, n = 6; vehicle treatment: IC50 = 7.45, maximum relaxation = 59–76%, n = 6) or its antagonism by taprostene (SC-51322 treatment: dose-ratio = 11.9, 95% CI = 4.7–29.8, n = 6; vehicle treatment: dose-ratio = 10.2, 95% CI = 5.4–19.5, n = 6, Table 2). The maximum response for the PGE2 / 3 μM taprostene combination was not significantly different from the control PGE2 maximum (P > 0.05) (Fig. 5C). However, a dose-ratio could be satisfactorily derived in only 4 of the 8 experiments. Nevertheless, the location of the PGE2 / 3 μM taprostene curve to the left of the control PGE2 curve suggests that the relaxant action of PGE2 is unopposed by taprostene. Isoprenaline relaxed the aorta to maximum of 79 to 96% (for 40% relaxation, pIC = 6.65. n = 8); a small reversal of contraction was seen at the highest concentration tested (5.55 μM), probably due to its α1-agonist action opposing its β -agonist action. 31 There was a slight synergism between the relaxant actions of isoprenaline and taprostene (Table 2).
Prostanoid Receptor Profiles of Vascular Preparations
In addition to a relaxant IP receptor system, each preparation appears to contain a relaxant EP4 system, while the guinea-pig saphenous vein also contains a relaxant EP2 system and the mouse aorta a contractile EP3 system (Fig. 6). In relation to the EP4 systems, the evidence for mouse aorta is based mainly on the greater potency of PGE2 over the selective EP2 agonist ONO-AE1-259 (~100-fold) since the effect of AH 23848 was not studied. DP receptor systems appear to be absent from the 3 preparations. The selective DP agonist BW 245C had only low relaxant potency, and on the guinea-pig saphenous vein, where it was most potent, its action (and that of PGE2) was only marginally inhibited by a high concentration of the potent and selective DP antagonist BW A868C. By analogy with previous studies on DP and EP4 receptors in rabbit saphenous vein, 32 our results are readily explained by the low agonist potency of BW 245C on the EP4 receptor and the low-affinity block of this receptor by BW A868C (pA2 = 5.1).
Taprostene as Partial Agonist at Prostacyclin Receptors
Ideally, for taprostene to be classified as an IP partial agonist, it must (1) induce a lower maximum response than the IP receptor system is capable of, (2) be blocked by an IP receptor antagonist, and (3) surmountably antagonize an IP full agonist while interacting additively (or even synergistically) with other relaxant agonists. Condition (1) was clearly met on the 3 vascular preparations. Condition (2) eluded us since IP antagonists are not yet available; indeed, they represent an eventual goal of our investigations. Condition (3), the main focus of our experiments, was also largely met. The log-concentration-response curves for the 3 putative IP full agonists, AFP-07, TEI-9063, and cicaprost, were consistently right-shifted by 3 μM taprostene, whereas the curves for the other agonists were unchanged or marginally left-shifted. However, if the 3 IP full agonists act solely on the IP system in a particular preparation, we would expect identical dose-ratios for their interactions with a fixed concentration of taprostene, and this was clearly not always the case. Each preparation will be dealt with in sequence.
On guinea-pig saphenous vein, where agreement between dose-ratios was good, interference with the interaction of a full and a partial agonist for the IP receptor is most likely to stem from the highly sensitive EP4 relaxant system present in this preparation. In our earlier experiments on piglet saphenous vein, 7 30 μM AH 23848 afforded mean dose-ratios of 3.8, 3.7, and 4.9 against relaxation induced by AFP-07, TEI-9063 and cicaprost, respectively and minimal block of taprostene-induced relaxation (dose-ratio < 2.0). We attributed these profiles to simultaneous activation of EP4 and IP receptors by the IP full agonists. In the current experiments on guinea-pig saphenous vein, 30 μM AH 23848 afforded dose-ratios for AFP-07, TEI-9063, and cicaprost of approximately 1.7, 1.2, and 1.6, respectively (corrected for shifts in control preparations) and a similar value for taprostene (1.6) (Table 1). These smaller shifts could easily be due to low-affinity block of IP receptors by AH 23848 (required pA2 ~4.3), or to another mechanism given the rather high concentration of AH 23848 used. Thus, we have no evidence that the 3 IP full agonists activate EP4 receptors in the guinea-pig saphenous vein. However, we cannot entirely exclude contributions from EP4 receptors when higher concentrations of AFP-07, TEI-9063, or cicaprost were used to overcome the taprostene blockade, since 30 μM AH 23848 produces limited antagonism of the EP4 system. It should be possible to obtain much greater EP4 receptor block using a recently reported antagonist GW 627368X, which from a preliminary report 33 has a pA2 of 9.2 on piglet saphenous vein (PGE2 as agonist).
Taprostene did not oppose relaxation of guinea-pig saphenous vein induced by PGE2, ONO-AE1-259, and acetylcholine, pointing to its specificity for the IP receptor. Radioligand binding experiments with recombinant prostanoid receptors have shown ONO-AE1-259 to be highly selective for the EP2 receptor: Ki = >10,000, 3.0, >10,000, and 6000 for mouse EP1 EP2, EP3, and EP4 receptors, respectively, and >10,000 for the human IP receptor. 15,34 Corresponding values for PGE2 are 5.0, 38, 3.1, 18, and >10,000 nM. ONO-AE1-259 had to be used with great care in the guinea-pig saphenous vein experiments since it desensitized its own (EP2) relaxant action; EP4 responses induced by PGE2 were not affected by previous exposure to a high concentration of ONO-AE1-259. This profile contrasts with a previous study involving mouse recombinant receptor/adenylate cyclase systems, in which the EP4 receptor was found to be more susceptible than the EP2 receptor to short-term agonist-induced desensitization. 35
On rat tail artery contracted with phenylephrine and under TP and EP4 receptor blockade, AFP-07, TEI-9063, and cicaprost each produced approximately 70% maximum relaxation, and taprostene selectively opposed the relaxant action of AFP-07. These actions are likely to involve only the IP system, which is incapable of completely inhibiting the generated tone. In contrast, with a similar level of tone induced by U-46619 and no prostanoid receptor blockade, TEI-9063 induced complete relaxation, AFP-07 and cicaprost induced approximately 87 and 80% relaxation, respectively, and taprostene's antagonism of TEI-9063 was less than that of AFP-07 or cicaprost. Simultaneous activation of EP4 and IP receptors by TEI-9063 would explain its profile. Although the EP4 system in the tail artery is not particularly sensitive or strong (PGE2 IC50 = 5 nM, maximum relaxation = 45%), inspection of Figure 4A suggests that TEI-9063 would require only one-hundredth of the EP4 potency of PGE2 to summate significantly with its IP agonist activity. An alternative explanation is that (in addition to its IP agonism) TEI-9063 specifically inhibits the action of the TP agonist. In this context, TEI-9063 is known to be a potent EP1 agonist, 30 and Walsh and Kinsella 36 have shown that EP1 receptor activation by either 17-phenyl PGE2 or prostacyclin inhibits TP receptor-mediated mobilization of internal calcium in human embryonic kidney (HEK-293) cells by a PKC-dependent, PKA-independent mechanism. However, EP1 receptor-mediated functional antagonism of this nature seems unlikely given that 17-phenyl PGE2 did not inhibit U-46619-induced contraction of the tail artery in the absence or the presence of relaxation induced by AFP-07.
A preliminary report by Karibe et al 37 indicated that the mouse thoracic aorta has a contractile TP system and a relaxant IP system. High concentrations of PGE2 (≥ 1 μM) caused contraction, probably through activation of TP receptors since this response was absent in the aorta of the TP receptor gene-deleted mouse. Based on this apparent lack of EP receptors, we decided to use the aorta to investigate the activity of taprostene. However, our demonstration of a potency ranking for contraction of SC-46175 > sulprostone >17-phenyl PGE2 in the presence of TP receptor blockade indicates the presence of an EP3 receptor system. A contribution to contraction from EP1 receptors is unlikely, given the lack of effect of the EP1 antagonist SC-51322 at 1.25 μM; its pA2 is 8.1 in guinea-pig ileum 27 and 8.8 in a human recombinant EP1 receptor / reporter gene assay. 38 The EP3 system in the mouse aorta is silent in the resting condition, but synergizes strongly with the α1-adrenoceptor system. A profile of this nature has not been reported previously for a prostanoid receptor in vascular smooth muscle. In other blood vessels, EP3 systems typically generate low maximum responses and synergism is associated with a contractile response to the EP3 agonist, however small. 12,39 Our strategy of downgrading any EP3 agonism of the relaxant prostanoids by using the EP3 agonist sulprostone to induce tone appears to have been successful. AFP-07, TEI-9063, and cicaprost surmounted the taprostene blockade on the phenylephrine / sulprostone-contracted aorta, but not on the phenylephrine-contracted aorta. In this context, moderate affinity for the mouse recombinant EP3 receptor has been reported for both AFP-07 (IC50 ~300 nM) 8 and cicaprost (Ki = 170 nM). 14 We have also uncovered a relaxant action of PGE2 on the mouse aorta under the phenylephrine / sulprostone protocol, which may be due to activation of EP4 receptors. We do not know whether this medium-sensitivity EP4 system interfered with the interactions of taprostene with AFP-07, TEI-9063, and cicaprost on the IP system.
Global View of Taprostene's Actions
Figure 6 shows log concentration-response relationships for AFP-07 and taprostene on 7 vascular preparations including the 3 reported in this paper. The trend from high to low sensitivity of the IP receptor system is associated with transition from full to partial agonism for taprostene and with transition from complete to incomplete relaxation for AFP-07. In a second series experiments on rabbit saphenous vein, AFP-07 and taprostene were more potent (pIC50 values move leftwards by 0.9 and 0.7 units, respectively, in Figure 6) and taprostene induced complete relaxation at 500 nM. 7 This within-tissue trend is also compatible with taprostene being an affinity-driven agonist. 40 Overall, taprostene appears to have reduced efficacy at IP receptors in vascular preparations from 5 species: pig, rabbit, guinea-pig, rat, and mouse. In the case of the human IP receptor, taprostene is a full agonist on both human pulmonary artery 30 and human platelets. 41 Both of these preparations are highly sensitive to IP agonist action, and it would be of interest to conduct further experiments to see whether increasing the strength of the excitatory input (contractile tone in the blood vessel ring) 42 results in a lower maximum response for taprostene relative to AFP-07.
Other Reports of Partial Agonists for Prostacyclin Receptors
Carbacyclin (6a-carba PGI2) and its geometric isomer 5Z-carbacyclin have been claimed to be full and partial agonists, respectively, on IP receptors in rabbit saphenous vein myocytes (cyclic AMP assay). 43 However, the concentrations of carbacyclin used were some 100-fold higher than those effective in relaxing ring preparations of rabbit mesenteric artery in our study. 7 Moreover, very high concentrations (200–1000 μM) of the 5Z-isomer were used to demonstrate antagonism of cyclic AMP elevation induced by prostacyclin, carbacyclin, and PGE1. Given that the isolated myocytes may contain both EP3 contractile and EP2 and EP4 relaxant systems (by analogy with findings for the vessel ring), we feel that there is considerable doubt about the assignment of 5Z-carbacyclin as an IP partial agonist.
In the non-prostanoid prostacyclin mimetic series, IP partial agonism has been reported for EP 157 on pig platelets, 44 BMY 45778 on human platelets, 45 and octimibate on human coronary artery. 46 However, these highly lipophilic agents also appear to suppress phospholipase C activation by a mechanism not involving IP receptors 47; the action of the aggregation- or tone-inducer could thus be affected, resulting in interference with the full/partial agonist interaction at IP receptors. Clearly, the detection and elimination of this type of problem requires careful design of experimental protocols.
We have presented evidence that taprostene is a partial agonist at IP receptors in several vascular smooth muscle preparations. Taprostene has a benzene ring inserted between C1 and C5 (Fig. 1), which conjugates the 1-carboxylate and 5(6a)-vinyl ether groups thereby contributing to the high acid-resistance of the vinyl ether. 2 It is possible that this modification, which confers considerable rigidity on the α-chain, is also responsible for the loss of efficacy at IP receptors. By analogy with previous work on the TP receptor, 48 it may be possible to modify the ω-chain of taprostene to reduce efficacy and enhance affinity, leading to a potent and specific IP receptor antagonist. Such an agent could be used to distinguish the classic IP receptor (present in the vascular preparations studied by us) and the novel subtype of IP receptor reported in the central nervous system, which tends to have excitatory actions and different agonist structure-activity relationships to the classic receptor. 49 Both PGE2 and prostacyclin and are potential mediators of inflammatory pain and their biosynthesis is inhibited by cyclooxygenase inhibitors such as indomethacin. 50,51 An IP receptor antagonist should assist in defining the relative contributions of these prostanoids to the inflammatory state.
Gifts of prostanoids from pharmaceutical companies mentioned in Materials and Methods are gratefully acknowledged. The assistance of Ms. Becky Kwan with the preparation of the manuscript is much appreciated.
1. Wise H, Jones RL. An introduction to prostacyclin and its receptors. In:Prostacyclin and its Receptors
. New York: Kluwer Academic; 2000:1–27.
2. Flohé L, Bohlke H, Frankus E, et al. Designing prostacyclin analogues. Drug Res.
3. Schneider J, Friderichs E, Ko ¨gel B, et al. Taprostene sodium. Cardiovasc Drug Rev.
4. Wise H, Jones RL. IP-receptors in the vasculature. In:Prostacyclin and its Receptors
. New York: Kluwer Academic; 2000:137–188.
5. Sitbon O, Humbert M, Simmoneau G. Primary pulmonary hypertension: Current therapy. Prog Cardiovasc Dis
6. Clapp LH, Finney P, Turcato S, et al. Differential effects of stable prostacyclin analogs on smooth muscle proliferation and cyclic AMP generation in human pulmonary artery. Am J Respir Cell Mol Biol
7. Jones RL, Chan KM. Distinction between relaxations induced via EP4
and IP1 receptors in pig and rabbit blood vessels. Br J Pharmacol
8. Chang CS, Negishi M, Nakano T, et al. 7,7-Difluoroprostacyclin derivative, AFP-07, a highly selective and potent agonist for the prostacyclin receptor. Prostaglandins
9. Negishi M, Hashimoto H, Yatsunami K, et al. TEI-9063, a stable and highly specific prostacyclin analogue for the prostacyclin receptor in mastocytoma P-815 cells. Prostaglandins
10. Sturzebecher S, Haberey M, Muller B, et al. Pharmacological profile of a novel carbacyclin derivative with high metabolic stability and oral activity in the rat. Prostaglandins
11. Coleman RA, Grix SP, Head SA, et al. A novel inhibitory prostanoid receptor in piglet saphenous vein. Prostaglandins
12. Jones RL, Shum WWC, Gurney AM. Synergism between prostanoids and other vasoactive agents. J Card Surg
13. Abramovitz M, Adam M, Boie Y, et al. The utilization of recombinant prostanoid receptors to determine the affinities and selectivities of prostaglandins and related analogs. Biochim Biophys Acta
14. Kiriyama M, Ushikubi F, Kobayashi T, et al. Ligand binding specificities of the eight types and subtypes of the mouse prostanoid receptors expressed in Chinese hamster ovary cells. Br J Pharmacol
15. Suzawa T, Miyaura C, Inada M, et al. The role of prostaglandin E receptor subtypes (EP1
, and EP4
) in bone resorption: an analysis using specific agonists for the respective EPs. Endocrinology
16. Cao J, Shayibuzhati M, Tajima T, et al. In vitro pharmacological characterization of the prostanoid receptor population in the non-pregnant porcine myometrium. Eur J Pharmacol
17. Jones RL, Qian YM, Chan KM, et al. Characterization of a prostanoid EP3
-receptor in guinea-pig aorta: partial agonist action of the non-prostanoid ONO-AP-324. Br J Pharmacol
18. Savage MA, Moummi C, Karabatsos PJ, et al. SC-46275: a potent and highly selective agonist at the EP3
receptor. Prostaglandins Leukot Essent Fatty Acids
19. Lawrence RA, Jones RL, Wilson NH. Characterization of receptors involved in the direct and indirect actions of prostaglandins E and I on the guinea-pig ileum. Br J Pharmacol
20. Whittle BJ, Moncada S, Mullane K, et al. Platelet and cardiovascular activity of the hydantoin BW245C, a potent prostaglandin analogue. Prostaglandins
21. Town MH, Casals-Stenzel J, Schillinger E. Pharmacological and cardiovascular properties of a hydantoin derivative, BW 245 C, with high affinity and selectivity for PGD2 receptors. Prostaglandins
22. Leff P, Giles H. Classification of platelet and vascular prostaglandin D2
(DP) receptors: estimation of affinities and relative efficacies for a series of novel bicyclic ligands. With an appendix on goodness-of-fit analyses. Br J Pharmacol
23. Giles H, Leff P, Bolofo ML, et al. The classification of prostaglandin DP-receptors in platelets and vasculature using BW A868C, a novel, selective and potent competitive antagonist. Br J Pharmacol
24. Pöch G, Holzmann S. Quantitative estimation of overadditive and under-additive drug effects by means of theoretical, additive dose-response curves . J Pharmacol Methods
25. Ariens EJ, Van Rossum JM, Simonis AM. A theoretical basis of molecular pharmacology. Part III: interaction of one or two compounds with two independent receptor systems. ArzneimForsch.
26. Glass GV, Hopkins KD. Multiple comparisons and trend analysis. In:Statistical Methods in Education and Psychology (3 rd Edition
). Boston: Allyn and Bacon; 1995:444–481.
27. Hallinan EA, Stapelfeld A, Savage MA, et al. 8-Chlorodibenz [b.f] [1.4]oxazepine-10(11H)-carboxylic acid, 2-[3-[2-(furanylmethyl)thio]-1-oxopropyl]hydrazide (SC-51322): a potent PGE2
antagonist and analgesic. Bioorg Med Chem Lett
28. Lumley P, White BP, Humphrey PPA. GR 32191, a highly potent and specific thromboxane A2
blocking drug on platelets and vascular and air-ways smooth muscle in vitro. Br J Pharmacol
29. Lydford SJ, McKechnie KCW, Leff P. Interaction of BW A868C, a prostanoid DP-receptor antagonist, with two receptor subtypes in the rabbit isolated saphenous vein. Prostaglandins
30. Jones RL, Qian YM, Wise H, et al. Relaxant actions of nonprostanoid prostacyclin mimetics on human pulmonary artery. J Cardiovasc Pharmacol
31. Russell A, Watts S. Vascular reactivity of isolated thoracic aorta of the C57BL/6J mouse. J Pharmacol Exp Ther
32. Lydford SJ, McKechnie KC, Dougall IG. Pharmacological studies on prostanoid receptors in the rabbit isolated saphenous vein: a comparison with the rabbit isolated ear artery. Br J Pharmacol
33. Wilson RJ, Giblin G, Foord S, et al. GW627368X: a novel, potent and selective EP4
antagonist. Br J Pharmacol
34. Narumiya S, FitzGerald GA. Genetic and pharmacological analysis of prostanoid receptor function. J Clin Invest
35. Nishigaki N, Negishi M, Ichikawa A. Two Gs-coupled prostaglandin E receptor subtypes, EP2
, differ in desensitization and sensitivity to the metabolic inactivation of the agonist. Mol Pharmacol
36. Walsh MT, Kinsella BT. Regulation of the human prostanoid TPα and TPβ receptor isoforms mediated through activation of the EP1
and IP receptors. Br J Pharmacol
37. Karibe H, Hara A, Yuhki K-O, et al. Characterization of the prostanoid receptors participating in contraction or relaxation of the aorta using mice lacking the prostanoid receptors. Pharmacologist.
38. Durocher Y, Perret S, Thibaudeau E, et al. A reporter gene assay for high-throughput screening of G-protein-coupled receptors stably or transiently expressed in HEK293 EBNA cells grown in suspension culture. Anal Biochem
39. Jones RL, Hung HYG, Lam FYF. Synergistic actions of prostanoid EP3
-receptor agonists on rat femoral artery. Pharmacologist
. 2002;44(Suppl 1):A81.
40. Kenakin T. Efficacy. In:Pharmacologic Analysis of Drug-Receptor Interaction (3 rd Edition)
. Philadelphia: Lippincott-Raven Publishers; 1997:289–330.
41. Michel G, Seipp U. In vitro studies with the stabilized epoprostenol analogue taprostene. Effect on platelets and erythrocytes. Drug Res.
42. Lew MJ. Extended concentration-response curves used to reflect full agonist efficacies and receptor-occupancy-response coupling ranges. Br J Pharmacol
43. Corsini A, Folco GC, Fumagalli R, et al. (5Z)-carbacyclin discriminates between prostacyclin receptors coupled to adenylate cyclase in vascular smooth muscle and platelets. Br J Pharmacol
44. Armstrong RA, Lawrence RA, Jones RL, et al. Functional and ligand binding studies suggest heterogeneity of platelet prostacyclin receptors. Br J Pharmacol
45. Seiler S, Brassard CL, Arnold AJ, et al. Octimibate inhibition of platelet aggregation: stimulation of adenylate cyclase through prostacyclin receptor activation. J Pharmacol Exp Ther
46. Merritt JE, Brown AM, Bund S, et al. Primate vascular responses to octimibate, a non-prostanoid agonist at the prostacyclin receptor. Br J Pharmacol
47. Chow KB, Wong YH, Wise H. Prostacyclin receptor-independent inhibition of phospholipase C activity by non-prostanoid prostacyclin mimetics. Br J Pharmacol
48. Wilson NH, Jones RL. Prostaglandin endoperoxide and thromboxane A2
analogues. In: Pike JE, Morton DR, eds. Advances in Prostaglandin, Thromboxane and Leukotriene Research
. New York: Raven Press; 1985:393–425.
49. Watanabe Y, Matsumura K, Takechi H, et al. A novel subtype of prostacyclin receptor in the central nervous system. J Neurochem
50. Vane JR. Inhibition of prostaglandin synthesis as mechanism of action for aspirin-like drugs. Nature
51. Bley KR, Hunter JC, Eglen RM, et al. The role of IP prostanoid receptors in inflammatory pain. Trends Pharmacol Sci