Activation of phospholipase C is dissociated from arachidonate metabolism during platelet shape change induced by thrombin or platelet-activating factor. Epinephrine does not induce phospholipase C activation or platelet shape change.

The present study compares the molecular mechanism by which thrombin, platelet-activating factor, and epinephrine induce platelet activation. Thrombin and platelet-activating factor induce an initial activation of phospholipase C, as measured by formation of 1,2-diacylglycerol and phosphatidic acid, during platelet shape change which is independent of and dissociated from metabolism of arachidonic acid. Phospholipase C activation and shape change are independent of extracellular Ca2+ and Mg2+. Formation of cyclooxygenase products occurs subsequent to the initial activation of phospholipase C and those metabolites are associated with platelet aggregation and further activation of phospholipase C. On the other hand, epinephrine is an unique platelet stimulus since it requires extracellular divalent cations and does not induce platelet shape change or activation of phospholipase C. Our results indicate that activation of phospholipase C may be a mechanism by which physiological agonists can activate platelets independently of extracellular divalent cations.

The present study compares the molecular mechanism by which thrombin, platelet-activating factor, and epinephrine induce platelet activation. Thrombin and platelet-activating factor induce an initial activation of phospholipase C, as measured by formation of 1,2-diacylglycerol and phosphatidic acid, during platelet shape change which is independent of and dissociated from metabolism of arachidonic acid. Phospholipase C activation and shape change are independent of extracellular Ca2+ and Mg2+. Formation of cyclooxygenase products occurs subsequent to the initial activation of phospholipase C and those metabolites are associated with platelet aggregation and further activation of phospholipase C. On the other hand, epinephrine is an unique platelet stimulus since it requires extracellular divalent cations and does not induce platelet shape change or activation of phospholipase C. Our results indicate that activation of phospholipase C may be a mechanism by which physiological agonists can activate platelets independently of extracellular divalent cations. ~~ In various cell types, the binding of hormones to their specific receptors induces distinct changes in the membrane phospholipids. In platelets, both the stimulation of the degradation of the inositol phospholipids and the release of arachidonic acid from phospholipids occur following platelet stimulation with various agonists such as thrombin, collagen, ADP, and platelet-activating factor (1)(2)(3)(4)(5)(6)(7)(8). Degradation of inositol-containingphospholipids by phospholipase C sequentially leads to the formation of 1,2-diacylglycerol and its phosphorylatedproduct phosphatidic acid (9,lO). These products remain inside the cell and might mediate platelet activation, since 1,2-diacylglycerol can activate protein kinase C (11,12) and phosphatidic acid can act as a Ca2+ ionophore and a fusogen at low concentrations of calcium (13)(14)(15)(16). Intracellular accumulation of phosphatidic acid could also activate phospholipase A2 (17) leading to the liberation of * This work was supported by the Deutsche Forschungsgemeinschaft. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
j To whom reprint requests should be addressed. arachidonic acid from various phospholipids (5). Arachidonic acid is metabolized by cyclooxygenase and thromboxane synthase to the biologically active endoperoxides and thromboxane A2 which can act inside the platelet or are released to the outside where they can activate other platelets (18). The significance of these changes in lipid metabolism for platelet function is, however, not clearly defined. Experiments in which platelets were stimulated in citrated platelet-rich plasma by ADP, platelet-activating factor, or I-epinephrine showed that cyclooxygenase products (endoperoxides and thromboxane AZ) are involved in later physiological platelet responses such as release reaction and second wave of aggregation, whereas they are not involved in primary aggregation Platelet shape change is the first measurable physiological platelet response preceding other responses such as platelet aggregation and release reaction (25). The present study indicates that platelet shape change induced by thrombin or platelet-activating factor is closely related to the activation of phospholipase C, but independent of the liberation and metabolism of arachidonic acid. In contrast, epinephrine does not induce phospholipase C activation or platelet shape change. (19-24).

EXPERIMENTAL PROCEDURES
Materials-Human thrombin, aspirin, trifluoperazine, quinacrine (mepacrine), arachidonic acid, 1,2-diolein, phosphatidic acid, prostaglandin El, bovine fibrinogen (Type I-S, lot F8630), creatine phosphate, creatine phosphokinase, hirudin, and potato apyrase (Grade I, lot A6132) were all obtained from Sigma. Platelet-activating factor was purchased from Calbiochem (Frankfurt, FRG). l-Epinephrine was purchased from Serva (Heidelberg, FRG), dissolved as a 10 mM solution in tartaric acid (10 mM), and stored at -20 "C. [  the previous 4 weeks. Blood was anticoagulated with 0.15 volume of ACD buffer (85 mM trisodium citrate, 111 mM dextrose, 71 mM citric acid, pH 5.5), and platelet-rich plasma (pH 6.8) was obtained by centrifugation at 200 X g for 20 min. Platelet-rich plasma (40-80 ml) was incubated with 200 pCi of [3H]arachidonic acid at 37 "C for 2 h in the presence of prostaglandin DP or prostaglandin E, (1 pg/ml) to prevent platelet activation. Platelets were then separated from plasma after addition of prostacyclin (300 ng/ml) and washed twice with a modified Ca2+-and M$+-free Tyrode-Hepes buffer containing 1 mM EGTA and prostaglandin Iz (300 ng/ml) to prevent platelet activation as detailed previously (26,27). Polypropylene material was used throughout the preparation procedure. In experiments in which the role of extracellular Ca2+ and Mg2+ for platelet lipid metabolism stimulated by platelet-activating factor was studied, the washing and resuspending buffer contained 1.5 mM CaC12 and 1 mM MgCb instead of 1 mM EGTA. Platelets were washed once in that buffer to which heparin (25 units/ml), prostacyclin (300 ng/ml), and EDTA/Tris buffer (aH 8.65) were added to a final concentration of EDTA/Tris of 5 mM/2.5 mM.
Platelet PreDaration for Stimulation with EDineDhrint-A different washing procedure was developed for platelet stimulation with epinephrine since the effect of epinephrine is dependent on the presence of extracellular Ca2+ and M$+, and epinephrine can act synergistically with other agonists such as ADP or thrombin which could be present in trace amounts in the platelet suspension (28)(29)(30)(31)(32)(33)(34)(35). Platelets were prelabeled with i3H]arachidonic acid as described above, and aspirin (1 mM final concentration) was added 15 min prior to the centrifugation of the platelets. Platelets from 60-80 ml of plateletrich plasma were then washed once in 30 ml of the Tyrode-Hepes buffer which contained 0.1 mM CaCI,, 1 mM MgC12, potato apyrase (200 pg/ml), heparin (25 units/ml), and prostacyclin (300 ng/rnl). After centrifugation, platelets were resuspended in the same buffer without heparin or prostacyclin. The temperature of the washing and resuspension buffer was kept at 37 'C, and platelets were stored at this temperature (35). Hirudin (2.5 units/rnl) and creatine phosphate/ creatine phosphokinase (2 mM/20 units/ml) were added to the platelet suspension 2 min before addition of epinephrine in order to avoid possible synergisms of epinephrine with trace amounts of endogenous thrombin or ADP. Thrombin (<0.25 units/ml) or ADP (<1 p~) added to those platelet suspensions did not induce platelet aggregation. Fibrinogen (50 pg/ml) was added 1 min before epinephrine. Epinephrine (100 p~) consistently induced platelet aggregation under those conditions. In some experiments, apyrase was omitted from the washing and resuspension buffer, and the buffer contained 1.5 mM CaCI, and 1 mM MgCl,. In these experiments the platelets lost their discoid shape and sometimes spotaneously aggregated.
Platelet Shape Change and Aggregation-Suspensions of washed platelets were adjusted to 4-6 X 10' platelets per ml and the experiments carried out between 30 and 90 min after final resuspension. Responsiveness of the platelets to the various agonists decreased slowly during that time period. Platelet suspensions (0.5 ml) were placed into aggregometer tubes, stirred (1100 rpm) for 2 min at 37 "C in the aggregometer, and then exposed to thrombin, platelet-activating factor, or epinephrine for various times. Shape change and aggregation of platelets were recorded (22, 36). The use of two 2-channel aggregometers allowed the simultaneous study of control and stimulated platelet suspensions. In some experiments, platelet shape change was examined by scanning electron microscopy (37). In studies in which inhibitors were used, platelet suspensions were preincubated for 2 min at 37 "C with indomethacin(disso1ved in ethanol), trifluoperazine (dissolved in ethanol/water, 50/50), or quinacrine (dissolved in HZO). The final ethanol concentration was 0.2% and did not affect platelet responses. Aspirinized platelets were prepared as described above.
Platelet Lipid Metabolism-Incubations (0.5 ml) were stopped by addition of 0.5 ml of methanol to the samples which were then transferred into 1.4 ml of chloroform/methanol (1:1.2). Unlabeled standards of phosphatidic acid and 1,2-diolein were added, and samples were then partitioned after addition of 0.62 ml of chloroform and 0.62 ml of 0.2% formic acid. The lower organic phases were evaporated under Nz, split into two halves, and lipids were separated on thin layer chromatography with two different solvent systems. Benzene/ diethyl ether/ethanol/NH3 (10080:4:0.2) was used to separate 1,2diacylglycerol and arachidonic acid (9). The upper phase of ethyl acetate/2,2,4-trimethylpentane/acetic acid/H,O (90:5020100) was used to separate phosphatidic acid, thromboxane B2, HHT, HETE, and arachidonic acid (1). Substances were localized by co-chromatog-raphy with unlabeled standards which were visualized by iodine vapor. After evaporation of iodine, the specific areas were scraped, transferred into scintillation vials, and 3H radioactivity was measured by liquid scintillation counting. Samples were counted in a Beckrnan 0scintillation counter (type LS 330; efficiency 60%) for 10 to 20 min to reach a counting error lower than 3%.
Data Presentation-Assays were done in duplicates or triplicates (for 1,2-diacylglycerol and arachidonic acid) and compared to unstimulated control samples which were set to 100%. In the control samples the range of radioactivity (counts per min of 3H) for the different lipids was as followed 500 to 1500 for phosphatidic acid, 100 to 350 for 1,2-diacylglycerol, 200 to 500 for arachidonic acid, and 80 to 180 for thromboxane Bz, HHT, or HETE. These variations were due to differences of incorporation of [3H]arachidonic acid into platelet phospholipids (range 20-5596). Data are presented as mean k S.E. of individual experiments from different blood donors (unless otherwise indicated). Statistical significance was calculated by the paired student t test using the data of the single assays of the individual experiments.

Platelet Stimulation with Thrombin and Platelet-activating
Factor-Human platelets were prelabeled with [3Hjarachidonic acid and washed in the presence of prostacyclin to prevent any platelet activation and to obtain the platelets in unstimulated discoid shape (26, 27, 37). The platelets were resuspended in buffer containing 1 mM EGTA and exposed to low concentrations of thrombin or platelet-activating factor, and platelet shape change and alterations of 3H-labeled platelet lipids were monitored simultaneously. Thrombin at low concentrations (0.05-0.075 unit/ml) induces only shape change ( Fig. la) without subsequent platelet aggregation and an increase in the formation of 1,2-diacylglycerol and phosphatidic acid (Fig. 2). Free arachidonic acid is not significantly increased up to 60 s, and metabolism of arachidonic acid via platelet cyclooxygenase or lipoxygenase is not detectable. per ml) prelabeled with [3H]arachidonic acid were placed into aggregometer tubes and preincubated for 2 min at 37 "C with or without trifluoperazine (40 pM) or indomethacin (5 pM) before exposure to thrombin or platelet-activating factor. Aspirin-pretreated platelets were prepared as described under "Experimental Procedures." a, thrombin (0.06 unit/ml); b, trifluoperazine plus thrombin (0.06 unit/ ml); c, thrombin (0.1 unit/ml); d, indomethacin plus thrombin (0.1 unit/ml); e, platelet-activating factor ( i O Y M); f, aspirin plus plateletactivating factor (IO" M). . In addition, aspirin does not affect platelet shape change or the formation of 1,2-diacylglycerol and phosphatidic acid induced by thrombin (data not shown). A small increase in the concentration of thrombin (0.09-0.12 unitlml) results in platelet aggregation which occurs after platelet shape change (Fig. IC). The formation of 1,2-diacylglycerol and phosphatidic acid shows a characteristic biphasic pattern, i.e. an initial small increase of 1,2-diacylglycerol and phosphatidic acid in parallel to platelet shape change, and a subsequent sharp increase of 1,2-diacylglycerol and phosphatidic acidobservedafter 30 s that coincides withplatelet aggregation and formation of arachidonate metabolites (Fig. 3). Indomethacin or aspirin reduces the second increase of 1,2diacylglycerol and phosphatidic acid, and also platelet aggregation (Figs. 3 and Id).
The differences between the changes in 3H-labeled lipids occurring during the selective platelet shape change induced by the lower thrombin concentrations (Fig. 2) and the shape change which precedes aggregation induced by the higher thrombin concentrations (Fig. 3) are summarized in Table I. First, it indicates that formation of 1,2-diacylglycerol and phosphatidic acid occurs more rapidly and is higher during the platelet shape change that precedes aggregation as compared to the platelet shape change induced by the lower thrombin concentrations. Secondly, free arachidonic acid only increases during the platelet shape change that precedes aggregation.
The lipid changes induced by platelet-activating factor during platelet shape change are similar to those induced by thrombin. Platelet-activating factor (0.1-1 p M ) stimulates the formation of 1,2-diacylglycerol and phosphatidic acid during platelet shape change (see Figs. 4 and 5). It was noticed that the increase of free arachidonic acid during platelet shape change occurs in some of the experiments but not in others (Fig. 4, a and b). However, in all experiments metabolism of arachidonic acid by platelet cyclooxygenase or lipoxygenase is not detectable. In experiments with platelet-activating factor in which free arachidonic acid increases, pretreatment with trifluoperazine or mepacrine prevents the increase of free arachidonic acid without affecting the formation of phosphatidic acid or platelet shape change (Table 11). Similarly, aspirin does not affect the formation of phosphatidic acid or platelet shape change (Fig. 1, e and f, Table 11). In some experiments, platelet shape change induced by thrombin or platelet-activating factor was analyzed by scanning electron microscopy and found to be comparable to previously reported results (37) (data not shown). The presence of extracellular Ca2+ and M$+ does not significantly affect the formation of l,Z-diacylglycerol, phosphatidic acid, or arachidonic acid induced by platelet-activating factor (1 p~) ; activation of phospholipase C and release of arachidonic acid induced by platelet-activating factor are comparable whether the platelet buffer contains either Ca2+ plus Mg2+ or EDTA (see Fig. 5). Again, release of arachidonic acid occurred in some experiments but not in others. Formation of arachidonic acid metabolites was found neither in the presence nor in the absence of Ca2+ and Mg2f (data not shown). Platelet-activating factor induces rapid platelet aggregation in the presence of Ca2+ and M$+, whereas in the presence of EDTA only shape change occurs.
Platelet Stimulation with Epinephrine-The platelet-stimulating effect of epinephrine is critically dependent on the presence of extracellular divalent cations and especially on the ratio of Ca2+ to M$+ (28-34). We developed a system to study the direct effect of epinephrine on platelets. Possible synergistic effects of trace amounts of ADP or thrombin with epinephrine were precluded by the addition of ADP scavengers and hirudin. In addition, platelets were aspirinized. As shown in Fig. 6, epinephrine induces a slow aggregation response without a preceding platelet shape change. In contrast, platelet-activating factor rapidly induces platelet shape change and reversible aggregation. Doses of platelet-activating factor (10 or 100 nM) and epinephrine (100 p~) that induce a similar degree of platelet aggregation were chosen to compare the effects of these two agonists. Epinephrine has no effects on platelet lipid metabolism; formation of phosphatidic acid is not significantly enhanced, and production of 1,2diacylglycerol, release of arachidonic acid, and formation of HETE are not detectable. In contrast, platelet-activating factor stimulates the formation of 1,Z-diacylglycerol and phosphatidic acid (Fig. 7). The platelet-aggregating effect of epinephrine and platelet-activating factor is dependent on extracellular divalent cations; the addition of 5 mM EDTA (after buffering with citrate to avoid shape change induced by EDTA (38, 39)) abolishes platelet aggregation induced by epinephrine and platelet-activating factor, but does not affect platelet shape change induced by platelet-activating factor (Fig. 6). Epinephrine, in the presence of EDTA, induces only a continuous slow decrease of light transmission without changing the amplitude of the oscillations of the platelet suspension.
We further found that epinephrine could induce formation of phosphatidic acid and stimulate arachidonic acid metabolism, if the platelet suspension did not contain ADP scavengers and if the platelet cyclooxygenase was not blocked by aspirin. Epinephrine could then induce considerable production of thromboxane B, and phosphatidic acid (Table 111). Indomethacin suppressed the formation of thromboxane B2, and in parallel it reduced (but not abolished) the amount of phosphatidic acid formed. Scavengers of ADP, however, completely blocked the effect of epinephrine on formation of phosphatidic acid (Table 111).

DISCUSSION
Platelet shape change is the first measurable platelet response. Our study shows that low concentrations of thrombin or platelet-activating factor which only induce platelet shape change increase the formation of 1,2-diacylglycerol and phosphatidic acid indicating the activation of the phosphodiesteratic cleavage of the inositol phospholipids (phosphatidylinositol, phosphatidylinositol 4-monophosphate, and phosphatidylinositol 4,5-bisphosphate; Refs. 1, 2, 4, 6-9, 40-46). Epinephrine which does not induce platelet shape chal.ge does not activate phospholipase C. Metabolism of arachidonic acid by platelet cyclooxygenase or lipoxygenase during platelet shape change does not occur. Agents which block the release of arachidonic acid from platelet phospholipids (trifluoperazine, mepacrine) or inhibit platelet cyclooxygenase (indomethacin, aspirin) do not affect the formation of 1,2-diacylglycerol and phosphatidic acid and the shape change of platelets induced by thrombin or platelet-activating factor. The results indicate that physiological agonists such as thrombin and platelet-activating factor induce initial platelet responses, i.e. platelet shape change and activation of phospholipase C without the participation of arachidonate metabolism.  Although the increase in the formation of 1,2-diacylglycerol and phosphatidic acid during platelet shape change is small (Figs. 2-4, Refs. 27, 37 and 47), the actual increase of those substances could occur in specific compartments of the cell resulting in high concentrations at critical sites. 1,2-Diacylglycerol may activate protein kinase C which phosphorylates a 40,000-dalton protein in platelets (11,12). Phosphorylation of that protein during platelet shape change has been observed (37,47,48); its function is, however, not yet known. Another protein which is phosphorylated during platelet shape change

Phospholipase C and Platelet
Shape Change 8291 is the myosin light chain which is directly involved in the contractile events of the cytoskeleton occurring during platelet shape change (48-50). The myosin light chain kinase is under control of Ca2+ which could be mobilized by phosphatidic acid, or, as recent observations suggest, by inositol 1,4,5trisphosphate (51), one of the products of phospholipase C attack on phosphatidylinositol 4,5-bisphosphate (46).
The present results strengthen the hypothesis of the sequential stimulation of phospholipases C and AZ (1,17). It is possible to dissociate two levels of platelet activation by using different concentrations of thrombin; low concentrations of thrombin (0.05-0.075 unit/ml) which induce platelet shape change without subsequent aggregation increase the formation of a small amount of 1,2-diacylglycerol and phosphatidic acid without an increase of free arachidonic acid. Higher concentrations of thrombin (0.09-0.12 unit/ml) leading sequentially to shape change and platelet aggregation induce during platelet shape change a more effective formation of 1,2-diacylglycerol and phosphatidic acid, and, in addition, an increase of free arachidonic acid. Cyclooxygenase activity appears to play a triggering role for the initiation of platelet aggregation induced by these concentrations of thrombin (0.09 to 0.12 unit/ml, see Figs. Id and 3). Cyclooxygenase products such as endoperoxides and thromboxane Az may be responsible for the sharp increase of 1,2-diacylglycerol and phosphatidic acid during aggregation, since inhibition of cyclooxygenase by indomethacin or aspirin reduces this second increase of 1,2-diacylglycerol and phosphatidic acid. Thus, the cyclooxygenase products formed at the onset of aggregation might act as positive feedback promoters for stimulation of phospholipase C (27,47). The increase of free arachidonic acid and the formation of the lipoxygenase product HETE which are observed during platelet aggregation are also reduced by inhibition of cyclooxygenase, but only partly. One would have expected that inhibition of cyclooxygenase would lead to an increased formation of arachidonic acid and HETE; the contrary is, however, observed. It thus appears that the inhibition of cyclooxygenase has also removed a stimulating effect of endogenous endoperoxides/thromboxane Az on the release of arachidonic acid. This effect of active cyclooxygenase products on the release of arachidonic acid may not be mediated directly, but via activation of phospholipase C, since it has been previously shown that phosphatidic acid may trigger the release of arachidonic acid (17). These observations seem to indicate that the release of arachidonic acid which is observed during platelet aggregation induced by low concentrations of thrombin is somehow dependent on the activity of phospholipase C.
It is important to note that liberation of arachidonic acid can already be observed during platelet shape change. The liberated arachidonic acid is, however, not metabolized. An increase of free arachidonic acid during shape change was consistently found with the higher concentrations of thrombin (0.09-0.12 unit/ml) and in some experiments with plateletactivating factor (Table I, Figs. 3 and 4). Extracellular Ca2+ and M F did not influence the release of arachidonic acid induced by platelet-activating factor (Fig. 5 ) . At present we cannot explain why the arachidonic acid released during shape change is not metabolized. The findings imply that some kind of coupling mechanism between the liberation and metabolism of arachidonic acid must exist which is not readily operative during platelet shape change. Recently we obtained evidence that the polymerization of actin might be involved in the coupling of arachidonic acid release to arachidonic acid metabolism (52).
Platelet aggregation induced by epinephrine is not preceded by platelet shape change, and epinephrine does not stimulate phospholipase C and metabolism of arachidonic acid. Epinephrine is known to bind to a2 receptors on platelets, to inhibit adenylate cyclase, and to increase the Ca2+ uptake of intact human platelets (30, 53, 54). It can directly induce exposure of fibrinogen receptors to which fibrinogen binds in the presence of Caz+ or Mg2+ leading to platelet aggregation (33). Since under those conditions formation of phosphatidic acid or 1,2-diacylglycerol was not observed, it seems that exposure of fibrinogen receptors and aggregation can occur in the absence of phospholipase C activation. The results further indicate that a2 receptors (in contrast to a1 receptors) are not coupled to the metabolism of the inositol phospholipids, and inhibition of adenylate cyclase might not be related to phospholipase C activation. Under certain conditions, if trace amounts of ADP were present in the platelet suspensions and platelet cyclooxygenase was not inhibited, epinephrine could, however, induce formation of phosphatidic acid. The effect was not due to a direct action of epinephrine, since the formation of phosphatidic acid was partly inhibited by cyclooxygenase inhibition and completely abolished by scavengers of ADP. These data indicate that epinephrine can act synergistically with ADP to induce formation of phosphatidic acid which is then partly mediated by cylooxygenase products. Epinephrine and ADP have synergistic effects on exposure of fibrinogen receptors, platelet aggregation, and release reaction (30-34). A possible source for trace amounts of extracellular ADP could be the contamination of the platelet suspension with red cells which become leaky during the washing and resuspending procedure (35).
We have previously observed that the level of phospholipase C activation induced by various agonists correlates with the degree of the physiological platelet response (27,47,49). The present study shows that platelet shape change and activation of phospholipase C are induced in the absence of extracellular Ca2+ or M$+. Addition of Ca2+ and M 2 + does not affect platelet shape change or phospholipase C activation; it only allows binding of fibrinogen to their receptors and crosslinking of fibrinogen molecules, thereby leading to platelet aggregation (55). It is not clear if platelet shape change requires mobilization of intracellular Ca2+ or not (56-58). The present results strengthen the significance of phospholipase C activation for platelet shape change. Epinephrine is an unique platelet stimulus as it requires extracellular divalent ions and does not activate phospholipase C. Activation of phospholipase C could, therefore, be a mechanism by which agonists can rapidly stimulate platelets independently of extracellular divalent cations.

Phospholipase C and Platelet
Shape Change