A New Family of Thromboxane Receptor Antagonists with Secondary Thromboxane Synthase Inhibition

Home , Dazoxiben, U46619

0022-3565/02/3012-618–624$7.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 301, No. 2 Copyright © 2002 by The American Society for Pharmacology and Experimental Therapeutics 4765/979886 JPET 301:618–624, 2002 Printed in U.S.A.

CECIL R. PACE-ASCIAK,1 DENIS REYNAUD, PETER DEMIN, RUKSHANA ASLAM, and ANDREA SUN Programme in Integrative Biology, Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada Received November 16, 2001; accepted January 18, 2002 This article is available online at http://jpet.aspetjournals.org

ABSTRACT ϭ ϫ We report herein a novel class of thromboxane receptor (TP formation and aggregation evoked by collagen with an IC50 8 receptor) antagonists modeled on unstable natural lipids that we 10Ϫ7 M. Other PBT (hepoxilin cyclopropane) analogs so far tested identified several years ago, the hepoxilins. These antagonists were 5- to 10-fold less active, and the native hepoxilins were have been rendered chemically and biologically more stable than about 500-fold less active. Neither PBT-3 nor the other analogs the natural compounds through structural modification by chem- inhibited 12-lipoxygenase, phospholipase A2, or cyclooxygenase ical synthesis. We demonstrate that the analogs inhibit the aggre- 1 or 2, and weakly stimulated adenyl cyclase (threshold stimula- gation of human platelets in vitro evoked by the thromboxane tion at 10Ϫ7 M and little selectivity for each of the PBT com- receptor agonists, I-BOP ([1S-[1␣,2␣(Z),3␤(1E,3S*),4␣]]-7-[3-[3- pounds). TP antagonism by PBT-3 was further demonstrated in 125 hydroxy-4-(4-iodophenoxy)-1-butenyl]-7-oxabi-cyclo[2.2.1]hept- receptor binding studies through use of I-BOP, where the IC50 2-yl]5-heptenoic acid) and U46619 (9,11-dideoxy-9␣,11␣- for PBT-3 was 8 ϫ 10Ϫ9 M, approximately 16-fold less than for methanoepoxy-prosta-5Z,13E-dien-1-oic acid). The most potent I-BOP itself. These findings identify a new mode of action of of the analogs described, PBT-3 [10(S)-hydroxy-11,12-cyclopro- PBT-3 and other related analogs as primarily TP antagonists. pyl-eicosa-5Z,8Z,14Z-trienoic acid methyl ester], has an IC50 ver- These studies identify a new family of compounds useful in further sus aggregation by I-BOP ϭ 0.6 ϫ 10Ϫ7 M and versus U46619 ϭ development as novel therapeutics for thromboxane-mediated 7 ϫ 10Ϫ7 M, representing one of the most potent anti-aggregating diseases. substances so far described. PBT-3 also inhibits thromboxane

Platelet aggregation is an important component of the diate, prostaglandin endoperoxide, common to all the hemostatic mechanism that prevents undesired bleeding, in prostaglandins and thromboxane. On the other hand extreme which a platelet plug forms at the site of injury to the blood cases of aggregation can lead to serious outcome as in septic vessel, leading to cessation of bleeding. Several aggregation shock and thrombosis (Parellada and Planas, 1977; Randall pathways have been described (Packham, 1993), one of which and Wilding, 1982; Fiedler et al., 1989; Silver et al., 1995; is the thromboxane pathway (Hamberg et al., 1974, 1975; Wolkow et al., 1997; Zaitsu et al., 1999). By comparison, Diczfalusy and Hammarstrom, 1979; Hammarstrom and Dic- NSAIDs such as aspirin reduce platelet aggregation through zfalusy, 1980). The active mediator in this pathway is throm- inhibition at the early stage of cyclooxygenase within the boxane A2, a powerful unstable pro-aggregating substance platelet but, also, in other cells/tissues, such as the blood and a vasoconstrictor of blood vessels. It is formed in plate- vessel wall, prevent the formation of all prostaglandins, some lets from arachidonic acid, a fatty acid present in mem- of which are anti-aggregatory and therefore beneficial, e.g., branes, by the enzyme, cyclooxygenase, through an interme- prostacyclin and prostaglandin E2 (Vane, 1978; Willis, 1978; Bunting et al., 1983; Harada et al., 1998). Therefore, proper This study was supported in part by a grant (MT-4181) to C.R.P.-A. from the Canadian Institutes of Health Research and from The Hospital for Sick Chil- management of thromboxane-mediated disease is desirable dren. at the level of thromboxane (synthesis and/or receptor ac- 1Associated with the Department of Pharmacology, University of Toronto, Toronto, ON M5S 1A8, Canada. tion), as this leaves the beneficial prostaglandins in place

ABBREVIATIONS: NSAIDs, nonsteroidal anti-inflammatory drugs; TP receptor, thromboxane A2 receptor; ADAM, 9-anthryldiazomethane; TSI, thromboxane synthase inhibitor; TRA, thromboxane receptor antagonist; TxB2, thromboxane B2; AA, arachidonic acid; HPLC, high-performance liquid chromatography; 12-HETE, 12(S)-hydroxyeicosa-5Z,8Z,10E,14Z-trienoic acid; HHT, 12-hydroxy-heptadeca-5Z,8E,10E-trienoic acid; PBT, hepoxilin cyclopropane analog; PBT-1, 8(S)-hydroxy-11,12-cyclopropyl-eicosa-5Z,9E,14Z-trienoic acid methyl ester; PBT-2, 8(R)-hydroxy-11,12-cyclopropyl- eicosa-5Z,9E,14Z-trienoic acid methyl ester; PBT-3, 10(S)-hydroxy-11,12-cyclopropyl-eicosa-5Z,8Z,14Z-trienoic acid methyl ester; PBT-4, 10(R)- hydroxy-11,12-cyclopropyl-eicosa-5Z,8Z,14Z-trienoic acid methyl ester; 12-LOX, 12(S)-lipoxygenase; COX, cyclooxygenase; Plase A2, phospholipase ␣ ␣ ␤ ␣ A2; PG, prostaglandin; I-BOP, [1S-[1 ,2 (Z),3 (1E,S*),4 ]]-7-[3-[3-hydroxy-4-(4-iodophenoxy)-1-butenyl]-7-oxabicyclo-[2.2.1]hept-2-yl]-5-heptenoic acid; U46619, 9,11-dideoxy-9␣,11␣-methanoepoxy-prosta-5Z,13E-dien-1-oic acid. 618 Novel Family of Thromboxane Receptor Antagonists 619

(Bunting et al., 1983). Recently, this approach has identified Binding of 125I-BOP to Platelets. Washed platelets were pre- a new class of compounds, oxazolecarboxamide-substituted pared as described above, except that the platelet suspension was alkenoic acids with dual TSI/TRA activities (Takeuchi et al., made up at a concentration of 10 ϫ 106 cells/0.5 ml (Dorn, 1991) in a 1998). clear medium (140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, During the course of our studies on the metabolic conver- 10 mM sodium-free HEPES, and 10 mM glucose, pH 7.3). The binding assay involved the addition of radioligand (30,000 cpm 125I-BOP) sion of arachidonic acid, we discovered a novel pathway with to all tubes in triplicates, containing either various concentrations of products, hepoxilins, that showed biological actions on a va- unlabeled I-BOP (10Ϫ9-10Ϫ7 M) or PBT-3 (10Ϫ9-10Ϫ7 M) or U46619 riety of systems (Pace-Asciak et al., 1983; Pace-Asciak and (10Ϫ9-10Ϫ7 M) in 1 ␮l ethanol. Additional tubes containing excess Martin, 1984; Pace-Asciak, 1994). Since these compounds unlabeled I-BOP were included to assess the extent of nonspecific were unstable biologically, we prepared by total chemical binding. Binding was allowed to take place during 30 min at 37°C; synthesis a family of analogs, PBTs, that were both chemi- free radioligand was removed by rapid vacuum filtration through cally and biologically suitable for in vivo studies (Demin and Whatman (Maidstone, UK) GF/B glass fiber filters prewashed with Pace-Asciak, 1993). Indeed, we found that different com- clear medium. The tubes and the filters were rapidly washed with pounds within this family acted on different in vivo systems, ice-cold clear medium (three times with 3 ml). The radioactivity on e.g., insulin secretion (Pace-Asciak et al., 1999), decrease in the filters containing the ligand-receptor complexes was counted in plasma glucose (unpublished observations), or inhibition of an LKB (Uppsala, Sweden) Compugamma CS counter. Measurement of Adenyl Cyclase Activity. Human washed lung fibrosis (Pace-Asciak et al., 2000). During more compre- platelets (350 ϫ 106 cells) in 1 ml of assay buffer were preincubated hensive screening for in vitro biological actions, we discov- during 10 min at 37°C. Dimethyl sulfoxide (1 ␮l) alone (control) or ered that one of these analogs, PBT-3, potently inhibited containing various concentrations of PBT-1, -2, -3, and -4 (2.8 ϫ collagen-evoked aggregation of human platelets through se- 10Ϫ7-10Ϫ5 M final concentration) was added, and the cells were lective blockade of the thromboxane synthetic pathway (Rey- incubated for a further 5 min. The reaction was stopped by the naud et al., 2001). The present results demonstrate a far addition of 12% trichloroacetic acid, and the samples were sonicated more important and powerful action of this compound as an (four times for 3 sec). The samples were left on ice for 60 min to antagonist of the TP receptor revealing a more likely mech- extract cAMP. After centrifugation at 2500g for 15 min, the super- anism for its inhibitory actions on platelet aggregation. natants were transferred and extracted five times with 5 ml of water-saturated diethyl ether to remove the trichloroacetic acid. The aqueous phase was transferred and lyophilized. cAMP was measured 125 Experimental Procedures using specific double antibody radioimmunoassay kits with I- labeled cAMP according to the instructions by the manufacturer Materials. The hepoxilin analogs, PBT-1, -2, -3, and -4, were (Amersham Biosciences, Piscataway, NJ). Results are expressed in prepared as previously described (Demin and Pace-Asciak, 1993). picomoles/350 million cells. Experiments were performed in tripli- Collagen (Chrono-Par) was purchased from Chronolog Corp. (Hav- cate for each point and repeated twice. ertown, PA). ADAM reagent (9-anthryldiazomethane) was from Re- Measurement of COX-1 and COX-2 Activity. COX-1 and search Organics Inc. (Cleveland, OH). U46619, I-BOP, and 125I-BOP COX-2 enzyme preparations were purchased from Cayman Chemical were from Cayman Chemical (Ann Arbor, MI). (Ann Arbor, MI). Preliminary studies established that 40 U of COX-1 Isolation of Human Platelets. Healthy human subjects who or 20 U of COX-2 could convert about 70% of 14C-AA [specific activity had not taken NSAIDs for at least 2 weeks were used. Blood was 55 mCi/mmol; Ontario Isotopes, Flamborough, ON, Canada; 100,000 drawn into plastic syringes containing citric acid-sodium citrate- cpm were diluted with 0.5 ␮g unlabeled arachidonic acid (Cayman dextrose (9:1, v/v). It was immediately centrifuged at 23°Cat200g Chemicals)/1 ml assay] into products in vitro. Different amounts of for 15 min. The platelet-rich plasma was transferred into fresh PBT-3 (10 ␮g and 20 ␮g) were added and the conversion of AA into plastic tubes and centrifuged at 400g for 5 min. The supernatant was products was assessed during a 10-min reaction in 1 ml phosphate discarded and the platelet pellet was resuspended in fresh medium buffer at 37°C. After extraction, products were assessed by thin layer containing 137 mM NaCl, 1 mM KCl, 0.4 mM NaH2PO4, 5.5 mM chromatography (Silica Gel G, ethyl acetate/acetic acid, 99:1, v/v). glucose, 20 mM HEPES, and 1 mM CaCl2, pH 7.4, and allowed to After development, the plates were scanned for radioactive products stand at room temperature for 30 min. The platelet count was ad- with a Berthold thin layer chromatography radiochromatogram justed to 350 ϫ 106 cells with medium to make 0.5 ml/assay/cuvette scanner (PerkinElmer Instruments, Norwalk, CT), and the radioac- for each measurement. tivity was quantified by scraping zones of silica gel, placing in scin- Measurement of Platelet Aggregation. Appropriate calibra- tillation vials, elution with 1 ml of methanol/water (1:1, v/v), and tion of the platelet aggregation profiler (model PAP-4C, Bio/Data addition of scintillation medium. Radioactivity was determined with Corp., Horsham, PA) for 0% and 100% transmission was carried out conventional counting in a beta scintillation counter (Beckman LS with a sample of platelet suspension and cell-free medium, respec- 3800; Beckman Coulter, Inc., Fullerton, CA). tively. A total of 0.5 ml of platelet suspension was added to sili- Measurements of Platelet-Derived Eicosanoids. Measure- conized glass tubes (four samples at a time) and heated with mag- ment of eicosanoids formed by platelets during treatment with col- netic stirring (900 rpm) to 37°C for 1 min in the aggregometer. Either lagen or the TP receptor agonists in the presence or absence of PBT-3 vehicle alone (1 ␮l ethanol) or PBT analog at various concentrations was carried out by HPLC after appropriate derivatization with a in ethanol (1 ␮l) was added, followed by agonist 2 min later (either fluorescent tag (ADAM), which forms a fluorescent ester (Demin et collagen at 2 ␮g/0.5 ml or the thromboxane receptor agonist, U46619, al., 1995). The method was adapted to measure the following com- at 10 ng/0.5 ml, or I-BOP at 2 ng/0.5 ml). The response was recorded pounds: TxB2, HHT, 12-HETE, and AA. The platelet suspension at for the next 5 min. In experiments addressing whether levels of the end of the experiment was mixed with ethyl acetate, 100 ng of endogenously produced thromboxane A2 play a role in the inhibition prostaglandin B1 was added as internal standard, and the mixture of aggregation by PBT-3, platelets were treated with aspirin (20 was acidified to pH 3 with 0.1 N HCl. After centrifugation, the ␮g/0.5 ml), followed either by collagen (in which aggregation was organic layer was separated, washed twice with water to neutrality, inhibited) or the thromboxane agonists, I-BOP or U46619 (which and evaporated to dryness. The residue was resuspended in ethyl resulted in aggregation); PBT-3 was added after aspirin but before acetate and half of the sample was taken for derivatization. It was I-BOP or U46619 at the above-mentioned doses, resulting in inhibi- diluted to 0.2 ml with ethyl acetate containing 20 ␮g ADAM reagent tion of aggregation. and was left in the dark for 2 h. The solvent was then evaporated and 620 Pace-Asciak et al. the residue was acetylated with a solution of pyridine/acetic anhy- dride (3:1, v/v) for 16 h at 23°C. The reagents were evaporated to dryness and the residue was resuspended in acetonitrile. One-tenth of the sample was used for HPLC analysis. Dose-response curves for varying amounts of test compounds were generated and the data were expressed as percentage of inhibition of control representing agonist-induced platelet aggregation. Each point was investigated three times and statistical analysis of the data was carried out (see below). Chromatography. Analysis of the anthryl (ADAM)-acetate derivatives of TxB2, HHT, and 12-HETE (AA only forms an ADAM derivative) in the extracted platelet samples was carried out on a Hewlett Packard (Palo Alto, CA) (1100 series) HPLC to which was attached a Shimadzu (Kyoto, Japan) fluorescent detector (RF- 10AXL). The detector was operated with excitation at 364 nm, emis- sion at 411 nm. Chromatographic separation of the compound was carried out on a Waters (Milford, MA) C18 Novapak column (3.9 ϫ 300 mm) using acetonitrile/water (80:20) at injection and after 10 min programmed with a linear gradient to 100% acetonitrile during 20 min. Statistical Analysis. Values stated are the mean Ϯ S.D. of the number of observations (n) indicated. Analysis of statistical signifi- cance was performed using Student’s t test involving the Macintosh StatView software program. Inhibition data (Figs. 3 and 4) were fitted to a line of best fit through a Michaelis-Menten-like hyperbolic treatment with a Kaleidagraph statistical software package.

Results PBT Analogs Inhibit Aggregation Evoked by TP Re- ceptor Activation in Washed Human Platelets in Vitro. Dose-related aggregation curves for I-BOP, a potent thromboxane receptor agonist, indicated that at a concentration of 2 ng/0.5 ml, it caused approximately 70% aggregation (data not shown). This dose was therefore chosen because it repre- sented a point at which inhibition curves for the test compounds could be most sensitive. We tested four related PBT analogs (PBT-1 to -4) on the I-BOP-evoked aggregation of Fig. 1. A, aggregometer curves showing the effects of four PBT analogs human platelets. Figure 1A shows aggregometer curves for (50 ng each/0.5 ml of platelet suspension containing 350 ϫ 106 cells) on the aggregation of human washed platelet suspensions evoked by I-BOP all four compounds at 50 ng each. These data show that (2 ng). PBTs were added 2 min before I-BOP. B, inhibition by PBTs at two PBT-3 was clearly more active than the other three analogs concentrations (20 and 50 ng/0.5 ml) of the aggregation of human washed in inhibiting aggregation. Figure 1B shows a comparison of platelets evoked by I-BOP (2 ng). two concentrations (20 and 50 ng/0.5 ml) of the four compounds, clearly resolving PBT-3 as the more active of the 2001). Analysis of the thromboxane formed in these experi- compounds, although all four compounds appear to inhibit ments indicated that collagen-evoked formation of thrombox- ϫ Ϫ7 aggregation evoked by I-BOP. PBT-3 is about 5-fold more ane was blocked by PBT-3 with an IC50 of 4 10 M active than the other analogs and is about 500 times more (Reynaud et al., 2001). In separate studies we showed that active than the native hepoxilins (data not shown). I-BOP-evoked or U46619-evoked aggregation was not accom- Dose-Related Inhibition by PBT-3 of I-BOP- and panied by thromboxane formation (data not shown); hence U-46619-Evoked Aggregation in Washed Human Plate- PBT-3 inhibition of the action of these two thromboxane lets in Vitro. Fig. 2A shows aggregation responses in human mimetics is due to direct inhibition at the TP receptor level washed platelet suspensions evoked by I-BOP and the inhi- and is not dependent on endogenous formation (or blockade) bition of this response by different doses of PBT-3 added 2 of thromboxane A2. These experiments clearly show that the min before the addition of I-BOP. Figure 2B shows aggrega- primary action of PBT-3 is at the level of the TP receptor, tion responses of human washed platelets challenged with although at higher concentrations thromboxane formation is U46619 and their inhibition by different amounts of PBT-3. also inhibited (Reynaud et al., 2001). Quantitative data for these experiments are shown in Fig. 3, Competition of I-BOP Binding to Platelets by PBT-3. demonstrating an IC50 for inhibition of aggregation evoked Additional confirmation that inhibition of aggregation by by I-BOP of 0.6 ϫ 10Ϫ7 M PBT-3. PBT-3 also inhibited the PBT-3 occurs at the level of the TP receptor was obtained 125 aggregation evoked by the agonist, U46619, with an IC50 of through competition binding studies with I-BOP as ligand. 7 ϫ 10Ϫ7 M (Fig. 3). Collagen evokes the aggregation of This reagent has been used as a specific agonist for the TP human platelets through the formation of thromboxane A2. receptor. If PBT-3 antagonizes the aggregation of platelets PBT-3 dose dependently prevented collagen-evoked aggrega- caused by I-BOP (as shown in Figs. 1–3), we thought that it ϫ Ϫ7 tion of platelets with an IC50 of 8 10 M (Reynaud et al., must compete for the TP receptor. Figure 4 shows that PBT-3 Novel Family of Thromboxane Receptor Antagonists 621

file of such an experiment. Both collagen and U46619 cause platelets to aggregate (see Fig. 5, lines 1 and 2, respectively). Aspirin greatly reduced collagen-evoked aggregation (Fig. 5, early part of line 3). We chose a dose of 20 ␮g aspirin for this study from earlier dose-response studies (not shown), because this dose blocked collagen effects almost completely, demonstrating that collagen-induced aggregation is mediated through the formation of endogenous thromboxane. Conversely, aspirin at this dose did not block U46619-evoked aggregation (Fig. 5, later part of line 3), demonstrating that aspirin did not interfere with the TP receptor or the cascade of events initiated by U46619. Addition of PBT-3 to aspirin- treated platelets before the addition of U46619 resulted in a blockade of the aggregation induced by U46619 (Fig. 5, line 4; compare with line 3). This finding, together with the binding data of Fig. 4, confirms that PBT-3 caused inhibition of the action of U46619 at the TP receptor level and is independent of the formation of endogenous thromboxane. Effects of PBT Analogs on Platelet Cyclic AMP Lev- els. Figure 6 shows the effects of the four PBT analogs on platelet cyclic AMP levels. All four compounds caused a stimulation of adenyl cyclase activity, but this occurred at concentrations much larger than those required for antagonism of I-BOP binding to the TP receptor; in addition, there was little discrimination between the four PBT analogs in stim- ulating cyclic AMP formation. This suggests that the mode of action of the PBT analogs, but especially of PBT-3, in inhibiting platelet aggregation could only partly be ascribed to stimulation of cyclic AMP formation. Antagonism of the TP receptor is a more likely and effective mechanism of action of PBT-3 in preventing aggregation of platelets evoked by TP receptor activation. Lack of Inhibition by PBT-3 of COX-1, COX-2, 12-

LOX, and Plase A2. Several experiments were carried out to investigate whether PBT-3 affected several enzymes involved in the generation of various eicosanoids or whether it Fig. 2. Aggregation curves showing the pro-aggregatory actions of the acted to block selectively thromboxane formation subsequent thromboxane receptor agonists, I-BOP (A) and U46619 (B) in washed to the actions of PBT-3 at the TP receptor. The data are human platelet suspensions and the inhibition of aggregation by various summarized in Table 1. Neither COX-1, COX-2, 12-LOX, nor amounts of PBT-3 added 2 min before the addition of the agonist. Values shown are typical of three separate experiments; quantitative data are Plase A2 is inhibited by relatively large amounts of PBT-3, shown in Fig. 3. i.e., about 3 to 4 log doses greater than that required to antagonize the TP receptor. In contrast, PBT-3 inhibited 125 competes for the binding of I-BOP in a dose-dependent significantly TxB2 formation in platelets (Table 1, last col- ϭ way. Competition curves are shown for I-BOP itself (IC50 umn). ϫ Ϫ9 ϭ ϫ Ϫ9 0.5 10 M), PBT-3 (IC50 8.1 10 M), and U46619 (IC ϭ 4.1 ϫ 10Ϫ9 M), another known TP receptor agonist. 50 Discussion PBT-3 is about 16-fold less active in competing with I-BOP 125 for I-BOP binding to the platelet TP receptor, but about Because thromboxane A2 is a powerful aggregating sub- equal to U46619. This study demonstrates that PBT-3 an- stance and a powerful constrictor of blood vessels, the control tagonizes I-BOP binding to the TP receptor, and together of its formation in the body is an important requirement with its inhibition of aggregation evoked by the two TP re- when its formation/action goes astray as in episodes of ceptor agonists, I-BOP and U46619, provides evidence that thrombosis or in states of septic shock in which thromboxane PBT-3 acts as a TP receptor antagonist. formation is believed to be exacerbated. Normal hemostasis Inhibition of Platelet Aggregation by PBT-3 Is Inde- requires a balance between thromboxane, formed by plate- pendent of Endogenous Thromboxane Synthesis. To lets, and prostacyclin, a powerful anti-aggregating substance establish whether endogenous formation of thromboxane A2 formed by endothelial cells in the wall of the blood vessel. plays a role in the action of PBT-3 in inhibiting the aggrega- Indeed, in disease states in which prostacyclin levels are tion process rather than the action of thromboxane, i.e., at abrogated, the tendency to thrombosis is a serious matter to the TP receptor, we carried out experiments in which plate- be reckoned with. The most common way to control these lets were treated with aspirin (to block endogenous throm- events is with high doses of NSAIDs, which inhibit the for- boxane formation) followed by the agonist, U46619 (which mation of thromboxane and other prostaglandins by blocking still causes aggregation). Figure 5 shows an aggregation pro- the precursor, PGH2, from being formed by the enzyme cy- 622 Pace-Asciak et al.

Fig. 3. Dose-response curves showing the inhibition by PBT-3 of the action of the thromboxane receptor agonists, I-BOP and U46619, and of collagen on the in vitro aggregation of washed human platelets. Values represent data from three separate experiments Ϯ S.D.

Fig. 4. Antagonism of 125I-BOP binding to washed human platelets by PBT-3, I-BOP, and U46619 (n ϭ 3 separate experiments; values represent the mean Ϯ S.D.). Note the similarity of potency of PBT-3 and U46619, a known TP receptor agonist.

Fig. 5. Platelet aggregation curves (typical of three separate experiments) demonstrating actions of PBT-3 at the TP receptor, independent of endogenous thromboxane formation. The study was carried out with normal platelets to which acetylsalicylic acid (ASA) was added (see above) to block endogenous formation of thromboxane and challenging with collagen (much reduced aggregation showing inhibitory effects of acetylsalicylic acid) or the TP receptor agonist, U46619 (pos- itive aggregation because this agonist bypasses the requirement for thromboxane formation). PBT-3 inhibited the U46619-evoked aggregation in the presence of acetylsalicylic acid, indicating an action on the TP receptor. Arrows indicate the time of addition of the various substances as reflected in the legend. Note curve 1, which shows the integrity of platelets to respond to collagen in the absence of acetylsalicylic acid. Novel Family of Thromboxane Receptor Antagonists 623

Scheme 1. Scheme describing TSI/TRA actions of PBT-3 reported in this study. Note that the main effect of PBT-3 is at the TP receptor as shown in this study with minor actions on TSI and cAMP formation, resulting in an overall inhibition of platelet aggregation. Arrow line width indicates level of activity. (ϩ), activation; (Ϫ), inhibition; X, no effect (see Table 1).

The results reported herein show novel anti-aggregating effects of a new family of eicosanoids which selectively inhibit the action of thromboxane at the TP receptor level, with secondary actions on thromboxane synthase at higher concentrations (Scheme 1). Human platelets appear to predom- Fig. 6. Dose-related effects of four PBT compounds on platelet cyclic AMP ␣ levels. Washed human platelets were incubated with the compounds inantly express the TP form of the two TP isoforms, sug- shown for 5 min, and the cells were extracted and measured by radioim- gesting that PBT-3 may be binding to this isoform, resulting munoassay for cyclic AMP as described under Experimental Procedures. in the blockade of both I-BOP- and U46619-evoked aggregation (Walsh et al., 2000). Interestingly, whereas 8-epi-pros- clooxygenase. Aspirin has been a drug of choice because of its taglandin F2␣ competes for I-BOP binding in human platelets popular use, and because it is cheap to produce and is gen- with 1000-fold less effectiveness than I-BOP itself (Kinsella erally well tolerated by most people. However, aspirin sensi- et al., 1997), PBT-3 is only 16-fold less active than I-BOP (see tivity has been noted, leading to asthma (Szczeklik et al., Fig. 4), making it far more effective than 8-epi-prostaglandin ␣ 1977, 2001), Reye’s syndrome (Baldwin 2000), and Down’s F2␣ as an antagonist to the TP receptor in platelets. The syndrome (Ebadi and Kugel, 1970), as well as gastric ulcers PBT compounds do not block thromboxane formation at the (Konturek et al., 1981; Brzozowski et al., 2001), a common level of COX-1 or COX-2 (Table 1), hence allowing the forma- side effect. For the latter reason “super aspirins” have been tion of PGH2 and its redirection into the beneficial prosta- designed which reduce the incidence of gastric lesions. Two glandins, PGE2 and PGI2. Additional studies have shown such drugs are Celebrex (celecoxib, 4-[5-(4-methylphenyl)- that PBT-3 increased cAMP formation by washed human 3-(trifluoromethyl)-1H-pyrazol-1-yl]benzenesulfonamide) platelets but required a concentration greater than 0.5 ϫ and VIOXX (rofecoxib, 4-[4-(methylsulfonyl)phenyl]-3- 10Ϫ6 M (Fig. 6), several log concentrations greater than the phenyl-2(5H)-furanone) (Everts et al., 2000; Urban, 2000). concentration required to inhibit I-BOP binding or action in These two drugs appear to selectively inhibit COX-2 platelets, suggesting that activation of adenyl cyclase repre- (COX-1 is present in platelets), an enzyme that produces sents a minor pathway of PBT-3 action. In vivo studies have the thromboxane precursor, PGH2; however, these drugs do shown that the PBT analogs reported herein are active, safe, not affect formation of thromboxane (formed via COX-1) in and well tolerated in rodent models of diabetes and in inflam- platelets. Hence, the super aspirins do not inhibit platelet- matory models of lung fibrosis (Pace-Asciak et al., 2000). derived thromboxane, the compound involved in the aggre- Although interest in the development of combined TSI/ gation process. TRA drugs is not new, the many drugs that have been devel-

TABLE 1 Effect of PBT-3 on various enzyme systems related to arachidonic acid metabolism

Enzyme activity (n ϭ 3) (% of control values) Concentration of PBT-3 a a b b b COX-1 COX-2 Plase A2 12-LOX TX-synthase ␮M Mean Ϯ S.D. 30.0 116.8 Ϯ 11NS 107.1 Ϯ 14NS ND ND ND 3.00 119.6 Ϯ 14NS 109.2 Ϯ 8.0NS 120.8 Ϯ 18.3NS 102.5 Ϯ 5.7NS 29.9 Ϯ 6.7*** 0.60 ND ND 89.5 Ϯ 15.6NS 73.0 Ϯ 4.9** 41.0 Ϯ 4.7*** 0.30 ND ND 91.9 Ϯ 5.6NS 85.2 Ϯ 7.4NS 65.7 Ϯ 6.0** 0.05 ND ND 100.0 Ϯ 6.9NS 79.4 Ϯ 7.6NS 94.2 Ϯ 15.2NS a Calculated from the percentage of 14C-AA remaining from TLC analysis of incubation of 40 U of COX-1 or 20 U of COX-2 (10 min, 37°C) with 14C-AA (see Experimental Procedures for details). Under these conditions, in control experiments (without PBT-3) the amount of AA remaining at the end of the experiment was approximately 30%, the rest being converted into polar products including HETEs and prostaglandins. b Calculated from HPLC analysis of the amount of AA (Plase A2), 12-HETE (12-LOX), TxB2, and HHT (TX-synthase) formed during incubation of washed human platelets treated with collagen Ϯ PBT-3 (6 min, 37°C) and expressed as percentage of collagen-stimulated controls (in the absence of PBT-3) (see Experimental Procedures for details). * ϭ 0.01 Ͻ p Ͻ 0.025; ** ϭ 0.0005 Ͻ p Ͻ 0.01; *** ϭ p Ͻ 0.0005; NS, p Ͼ 0.025; ND, not determined. 624 Pace-Asciak et al. oped, ranging in effectiveness within the concentrations of receptor alpha isoform (TPalpha) functionally couples to the G proteins Gq and Ϫ4 Ϫ8 G11 in vivo and is activated by the isoprostane 8-epi-prostaglandin F2alpha. 10 to 10 M, have not proven clinically successful (Boehm J Pharmacol Exp Ther 281:957–964. et al., 1996; Moncada et al., 1977; Needleman et al., 1977). A Konturek SJ, Radecki T, Brzozowski T, Piastucki I, Zmuda A, and Dembinska-Kiec A (1981) Aspirin induced gastric ulcers in cats—prevention by prostacyclin. Dig recent addition to this group of drugs has been made with the Dis Sci 26:1003–1012. development of substituted ␻-phenyl-␻-(3-pyridyl) alkenoic Moncada S, Bunting S, Mullane K, Thorogood P, and Vane JR (1977) Imidazole: a selective inhibitor of thromboxane synthetase. Prostaglandins 13:611–618. acids (Takeuchi et al., 1998). These drugs were shown to be Needleman P, Raz A, Ferrendelli J-A, and Minkes M (1977) Application of imidazole ϫ Ϫ9 as effective as TRAs with an IC50 of 55 10 M. The as a selective inhibitor of thromboxane synthetase in human platelets. Proc Natl compounds were found active in washed platelet suspensions Acad Sci USA 74:1716–1720. Pace-Asciak CR (1994) Hepoxilins: a review of their cellular actions. Biochim Bio- and in ex vivo models, but were less active on platelet-rich phys Acta 1215:1–8. plasma, possibly due to plasma binding or metabolism. Our Pace-Asciak CR, Demin PM, Estrada M, and Liu G-Y (1999) Hepoxilins raise circu- ϫ Ϫ9 lating insulin levels in vivo. FEBS Lett 461:165–168. compounds display similar potency (IC50 8 10 M), yet we Pace-Asciak CR, Granstro¨m E, and Samuelsson B (1983) Arachidonic acid epoxides: have found them to be active in vivo in models of inflamma- isolation and structure of 2 hydroxy epoxide intermediates in the formation of 8,11,12-trihydroxy eicosatrienoic acid and 10,11,12-trihydroxy eicosatrienoic acid. tion and diabetes, suggesting that although the PBTs may J Biol Chem 258:6835–6840. bind to albumin, they are still effective in those models. We Pace-Asciak CR, Luo X, Demin P, and Tanswell K (2000) Hepoxilin analogs inhibit bleomycin-induced lung fibrosis in the mouse. Adv Prostaglandin Thromboxane believe that the mechanism of action of the PBT compounds Leukot Res, 11th International Conference, Florence, Italy, June 4–8, Abstr pp 18. described in our studies may be through their ability to Pace-Asciak CR and Martin JM (1984) Hepoxilin, a new family of insulin secreta- gogues formed by intact rat pancreatic islets. Prostaglandins Leukot Med 16:173– potently antagonize the TP receptor. 180. Packham MA (1993) Role of platelets in thrombosis and hemostasis. Can J Physiol References Pharmacol 72:278–284. Baldwin GS (2000) Do NSAIDs contribute to acute fatty liver of pregnancy? Med Parellada PP and Planas JM (1977) Action of selective inhibitor of thromboxane Hypotheses 54:846–849. synthetase on experimental thrombosis induced by arachidonic acid in rabbits. Boehm JC, Smietana JM, Sorenson ME, Garigipati RS, Gallagher TF, Sheldrake PL, Lancet ii:40. Bradbeer J, Badger AM, Laydon JT, Lee JC, et al. (1996) 1-substituted 4-aryl-5- Randall MJ and Wilding RIR (1982) Acute arterial thrombosis in rabbits: reduced pyridinylimidazoles: a new class of cytokine suppressive drugs with low 5-lipoxy- platelet accumulation after treatment with thromboxane synthetase inhibitor genase and cyclooxygenase inhibitory potency. J Med Chem 39:3929–3937. dazoxiben hydrochloride (UK-37,248-01). Thromb Res 28:607–616. Brzozowski T, Konturek PC, Konturek SJ, Sliwowski Z, Pajdo R, Drozdowicz D, Ptak Reynaud D, Sun A, Demin P, and Pace-Asciak CR (2001) Novel platelet anti- A, and Hahn EG (2001) Classic NSAID and selective cyclooxygenase (COX)-1 and aggregating substances. Biochem Biophys Res Commun 284:580–582. COX-2 inhibitors in healing of chronic gastric ulcers. Microsc Res Tech 53:343–353. Silver RM, Edwin SS, Trautman MS, Simmons DL, Branch DW, Dudley DJ, and Bunting S, Moncada S, and Vane JR (1983) The prostacyclin-thromboxane A2 Mitchell MD (1995) Bacterial lipopolysaccharide-mediated fetal death. Production balance: pathophysiological and therapeutic implications. Br Med Bull 39:271– of a newly recognized form of inducible cyclooxygenase (COX-2) in murine decidua 276. in response to lipopolysaccharide. J Clin Investig 95:725–731. Demin P, Reynaud D, and Pace-Asciak CR (1995) High-performance liquid chro- Szczeklik A, Gryclewskik RJ, and Czerniawska-Mysik G (1977) Participation of matographic separation of fluorescent esters of hepoxilin enantiomers on a chiral prostaglandins in pathogenesis of aspirin-sensitive asthma. Arch Pharmacol 297: stationary phase. J Chromatogr B 672:282–289. S99–S101. Demin PM and Pace-Asciak CR (1993) Synthesis of racemic 11,12-cyclopropyl ana- Szczeklik A, Nizankowska E, Bochenek G, Nagraba K, Mejza F, and Swierczynska M (2001) Safety of a specific COX-2 inhibitor in aspirin-induced asthma. Clin Exp logs of hepoxilins A3 and B3. Tetrahedron Lett 34:4305–4308. Diczfalusy U and Hammarstrom S (1979) A structural requirement for the conver- Allergy 31:219–225. sion of prostaglandin endoperoxides to thromboxanes. FEBS Lett 105:291–295. Takeuchi K, Kohn TJ, True TA, Mais DE, Wikel JH, Utterback BG, Wyss VL, and Dorn GW II (1991) Mechanism for homologous downregulation of thromboxane A2 Jakubowski JA (1998) Development of dual-acting agents for thromboxane recep- receptors in cultured human chronic myelogenous leukemia (K562) cells. J Phar- tor antagonism and thromboxane synthase inhibition. 3. Synthesis and biological macol Exp Ther 259:228–234. activities of oxazolecarboxamide-substituted ␻-phenyl-␻-(3-pyridyl)alkenoic acid Ebadi MS and Kugel RB (1970) Alteration in metabolism of acetylsalicylic acid in derivatives and related compounds. J Med Chem 41:5362–5374. children with Down’s syndrome: decreased plasma binding and formation of sali- Urban M (2000) COX-2 specific inhibitors offer improved advantages over traditional cyluric acid. Pediatr Res 4:187–193. NSAIDs. Orthopedics 23:S761–S764. Everts B, Wahrborg P, and Hedner T (2000) COX-2-specific inhibitors—the emer- Vane JR (1978) The mode of action of aspirin-like drugs. Agents Actions 8:430. gence of a new class of analgesic and anti-inflammatory drugs. Clin Rheumatol Walsh M-T, Foley JF, and Kinsella BT (2000) The ␣, but not the ␤, isoform of the 19:331–343. human thromboxane A2 receptor is a target for prostacyclin-mediated desensiti- Fiedler VB, Perzborn E, Seuter F, Rosentreter U, and Bo¨shagen H (1989) Reduction zation. J Biol Chem 275:20412–20423. of in vivo coronary artery thrombosis by the novel thromboxane antagonist (3R)- Willis AL (1978) Platelet aggregation mechanisms and their implications in haemo- 3-(4-fluorophenylsulfonamido)-1,2,3,4-tetrahydro-9-carbazolepropan oic acid. Arz- stasis and inflammatory disease, in Inflammation. Handbook of Experimental neim Forsch 39:1527–1530. Pharmacology (Ferreira SH and Vane JR, eds), Springer-Verlag, Berlin, pp 138– Hamberg M, Svensson J, and Samuelsson B (1975) Thromboxanes: a new group of 205. biologically active compounds derived from prostaglandin endoperoxides. Proc Wolkow PP, Bartus JB, and Gryglewski RJ (1997) Pneumotoxicity of lipopolysac- Natl Acad Sci USA 72:2994–2998. charide in nitric oxide-deficient rats is limited by a thromboxane synthase inhib- Hamberg M, Svensson J, Wakabayashi T, and Samuelsson B (1974) Isolation and itor. J Physiol Pharmacol 48:645–653. structure of two prostaglandin endoperoxides that cause platelet aggregation. Proc Zaitsu M, Hamasaki Y, Matsuo M, Miyazaki M, Hayasaki R, Muro E, Yamamoto S, Natl Acad Sci USA 71:345–349. Kobayashi I, Ichimaru T, and Miyazaki S (1999) Induction of cytosolic phospho- Hammarstrom S and Diczfalusy U (1980) Biosynthesis of thromboxanes. Adv Pros- lipase A2 and prostaglandin H2 synthase-2 by lipopolysaccharide in human poly- taglandin Thromboxane Res 6:267–274. morphonuclear leukocytes. Eur J Haematol 63:94–102. Harada Y, Kawamura M, Hatanaka K, Saito M, Ogino M, Ohno T, Ogino K, and Yang QS (1998) Differing profiles of prostaglandin formation inhibition between Address correspondence to: Prof. C. R. Pace-Asciak, Research Institute, selective prostaglandin H synthase-2 inhibitors and conventional NSAIDs in in- The Hospital for Sick Children, 555 University Avenue, Toronto, ON, Canada flammatory and non-inflammatory sites of the rat. Prostaglandins 55:345–358. M5G 1X8. E-mail: [email protected] Kinsella BT, O’Mahony DJ, and Fitzgerald GA (1997) The human thromboxane A2