Platelets enzymatically convert prostaglandin H 3 (PGH 3 ) into thromboxane A 3 . Both PGH 2 and thromboxane A 2 aggregate human platelet-rich plasma. In contrast, PGH 3 and thromboxane A 3 do not. PGH 3 and thromboxane A 3 increase platelet cyclic AMP in platelet-rich plasma and thereby: ( i ) inhibit aggregation by other agonists, ( ii ) block the ADP-induced release reaction, and ( iii ) suppress platelet phospholipase-A 2 activity or events leading to its activation. PGI 3 (Δ 17 -prostacyclin; synthesized from PGH 3 by blood vessel enzyme) and PGI 2 (prostacyclin) exert similar effects. Both compounds are potent coronary relaxants that also inhibit aggregation in human platelet-rich plasma and increase platelet adenylate cyclase activity. Radioactive eicosapentaenoate and arachidonate are readily and comparably acylated into platelet phospholipids. In addition, stimulation of prelabeled platelets with thrombin releases comparable amounts of eicosapentaenoate and arachidonate, respectively. Although eicosapentaenoic acid is a relatively poor substrate for platelet cyclooxygenase, it appears to have a high binding affinity and thereby inhibits arachidonic acid conversion by platelet cyclooxygenase and lipoxygenase. It is therefore possible that the triene prostaglandins are potential antithrombotic agents because their precursor fatty acids, as well as their transformation products, PGH 3 , thromboxane A 3 , and PGI 3 , are capable of interfering with aggregation of platelets in platelet-rich plasma.

The prostaglandin endoperoxide ring structure alone does not establish suitability as a substrate for thromboxane synthetase, but the degree of unsaturation and carbon chain length are also essential features. Thus, human platelet microsomes can synthesize thromboxane A2, thromboxane A3, but not thromboxane A1 from their respective endoperoxides. The potent vasoconstrictor property of thromboxanes can be dissociated from its capacity to produce platelet aggregation. Furthermore, thromboxane formation is not an essential process in platelet aggregation. The observations indicate the remarkable structural specificity of both the synthetic enzymes, cyclooxygenase and thromboxane synthetase, as well as the vascular and platelet receptor sites.

Prostaglandin release from various isolated perfused organs is characteristically dependent on the stimulus; thus, certain stimuli caused prostaglandin release in some organs and not in others. Adenosine triphosphate and adenosine diphosphate were potent stimulators of prostaglandin biosynthesis in the kidney, spleen, spleen fat pad, heart, liver and lung; adenosine monophosphate and adenosine were inactive. Epinephrine caused prostaglandin release from the kidney, spleen, and liver, whereas, angiotensin was agonistic in the kidney, spleen, spleen fat pad, and liver. Indomethacin abolished the release (biosynthesis) of prostaglandins from all organs by each agonist.

A B S T R A C T When thrombin is added to washed human platelets, one of its actions results in activation of a phospholipase that hydrolyzes arachidonic acid from phospholipids. The arachidonate is converted to the cyclic endoperoxides (prostaglandin G2 and prostaglandin H2) by fatty acid cyclo-oxygenase. These compounds are then converted to thromboxane A2, also called rabbit aorta-contracting substance, by thromboxane synthetase. These labile, pharmacologically active compounds then break down to inactive products including thromboxane B2 and malonaldehyde. Incubation of platelets with either dibutyryl cyclic adenosine 3',5'-monophosphate (dBcAMP) or prostaglandin E, (PGE,) before thrombin addition blocks the subsequent formation of oxygenated products of arachidonic acid including thromboxane A2, thromboxane B2, and malonaldehyde. In contrast, when arachidonic acid is added directly to platelets, prior incubation with dBcAMP or PGE, does not inhibit production of the prostaglandins or their metabolites. Thrombin treatment of platelets also blocks the acetylation of cyclo-oxygenase by aspirin since the hydrolyzed arachidonic acid competes with aspirin for the active site on cyclo-oxygenase. Prior treatment of platelets with dBcAMP or PGE, reverses the thrombin inhibition of the acetylation of cyclo-oxygenase. We conclude that agents which elevate platelet cAMP levels inhibit the hydrolysis of arachidonic acid from platelet phospholipids. We also find that prostaglandin synthesis can be dissociated, in part, from platelet aggregation and release, and that cAMP has separate actions on these processes. Higher thrombin concentrations are required to stimu-

Human platelet suspensions release a rabbitaorta-contracting substance (previously identified as thromboxane A2) during aggregation produced by arachidonic acid, prostaglandin endoperoxide, thrombin, and collagen. Incubation of platelets with imidazole did not interfere with the aggregation produced by these agonists but markedly reduced the generation of the rabbit-aorta-contracting substance. We find that imidazole inhibited the conversion of exogenous or endogenous prostaglandin endoperoxide into thromboxane A2. Imidazole selectively inhibits thromboxane synthetase in intact human platelets, because this agent blocks the conversion of ['4Clarachidonate into [14C thromboxane B2 but does not inhibit the conversion of [I4C arachidonate into [14C prostaglandin E2. The inhibition of thromboxane synthetase by imidazole is not the result of an alteration in platelet 3':5'-cyclic AMP levels. These results illustrate the utility of imidazole as a harmacological tool and demonstrate the two unique and dissociable properties of the endoperoxides themselves-their ability to aggregate platelets and their enzymatic conversion to the potent vasoconstrictor thromboxane.Numerous recent investigations have established the diversity of the pathways of arachidonic acid metabolism. Prostaglandins (PGs) E2, F2a, and D2, once considered the primary prostaglandin products, have now been joined by such novel compounds as: the cyclic endoperoxide intermediates (PGG2 and PGH2) (1, 2); the thromboxanes A2 and A3 (3-5); and 6-ketoPGFia (6-10) and its newly discovered labile intermediate (9,11,12). The enzymatic conversion of arachidonate to the endoperoxides is the common initial step in the biosynthesis of the biologically active products. The endoperoxides appear to be the substrates for several enzymatic conversions (i.e., PGH2 PGE2; PGH2 -thromboxane A2; PGH2 -"X" --6-ketoPGFia). The remarkable differences in the biological activities of the end products generated from PGH2 indicate that inhibition of the initial cyclo-oxygenase reaction using aspirin or indomethacin may lead to unpredictable physiologic effects. It would obviously be useful to have agents that selectively inhibit the synthesis of the various products originating from PGH2. We demonstrated that the anti-inflammatory agent benzydamine inhibits the enzymatic conversion of PGH2 into thromboxane A2. However, this compound is not selective and also inhibits cyclo-oxygenase (13). Similarily, a phloretin phosphate derivative, N-0164, was reported to inhibit thromboxane synthetase (14), but this compound lacks specificity in that it is also a prostaglandin receptor antagonist in isolated smooth muscle preparations (15). We now report that the commonly available reagent imidazole selectively inhibits thromboxane synthetase but does not block cyclo-oxygenase. Using imidazole as a pharmacological tool with human platelet suspensions, we show that exogenous or endogenous PG endoperoxides can induce aggregation without being converted into thromboxane A2. A preliminary r...