Platelet Agonists Enhance the Import of Phosphatidylethanolamine into Human Platelets*

It is unknown whether the endocytosis-independent transfer of phospholipids from lipoproteins to platelets is regulated by platelet agonists such as thrombin. The movements of the choline phospholipids phosphatidylcholine and sphingomyelin (labeled with either 14C or the fluorescent pyrenedecanoic acid) between low density lipoproteins and platelets were unaffected by thrombin (0.5 unit/ml). In contrast, thrombin accelerated the import of diacyl phosphatidylethanolamine (PE) and alkenylacyl phosphatidylethanolamine into platelets by about 4-fold. Similarly, thrombin receptor-activating peptide (15 μm), collagen (10 μg/ml), and ADP (10 μm) enhanced PE uptake. High density lipoprotein particles and egg phosphatidylcholine vesicles were also donors for stimulation of platelet PE import. Part of the [14C]arachidonic acid-labeled PE transferred from low density lipoprotein to platelets activated by thrombin and collagen was metabolized to 14C-eicosanoids. Inhibitors of protein kinase C partially prevented thrombin-induced [14C]PE uptake, while direct activators of protein kinase C increased incorporation of [14C]PE into platelets. Proteinaceous factor(s) recovered in the extracellular medium from ADP- and thrombin-activated platelet suspensions were found to accelerate the transfer of pyrenedecanoic acid-labeled PE between donor and acceptor lipid vesicles. The stimulation of import of ethanolamine phospholipids led to a 2-fold enhancement of the prothrombinase activity of thrombin-activated platelets. Our study demonstrates that physiological platelet stimuli increase specifically the transfer of ethanolamine phospholipids from lipoproteins to platelets through a secretion-dependent mechanism. This might contribute to the increase of procoagulant activity of stimulated platelets.

The processes of phospholipid transfer between different plasma lipoproteins (1), between the two leaflets of biological membranes (2), as well as between different intracellular membranes (3) have been elucidated in considerable detail in recent years. Much less is known about the characteristics of phospholipid transfer between plasma lipoproteins and cells. Lipoproteins may supply their phospholipids by endocytosis of whole lipoprotein particles or by selective transfer of phospholipids to the acceptor cells. Within the plasma compartment, phospholipids (and cholesterol) carry major portions of highly polyunsaturated fatty acids with four and more double bonds such as, e.g., arachidonic acid. Only very small quantities of these fatty acids are present in the free form in plasma (4). Previous work indicates that arachidonic acid esterified to cholesterol within LDL 1 is taken up by endocytosis of the lipoproteins and further processed to bioactive eicosanoids in growth factor-activated fibroblasts (5). Apart from the potential role of phospholipid transfer in supplying cells with polyunsaturated fatty acids, the newly imported phospholipids may serve other functions in intracellular signaling mechanisms. Second messengers such as diacylglycerol and ceramide are generated in part by phospholipase C-and sphingomyelinase-mediated hydrolysis of membrane phosphatidylcholine (PC) and sphingomyelin (SM), respectively. Changes in import of precursor phospholipids may well modulate the intracellular signaling pathways regulated by these messenger molecules.
Phospholipids of the plasma membranes of cells are also known to be indispensable elements of the coagulation cascade. Within this process, interaction of extracellular coagulation factors with binding proteins on the plasma membrane of platelets is greatly facilitated by distinct phospholipids. In particular, the appearance of the negatively charged phosphatidylserine in the outer leaflet of the plasma membrane of platelets appears to be an essential cofactor for some steps in the coagulation cascade such as the formation of prothrombin (6,7). Additionally, anticoagulant activities mediated by proteins of endothelial cell membranes are modulated by specific phospholipids (8,9).
We recently observed that plasma low and high density lipoproteins (LDL, HDL) rapidly delivered PC, phosphatidylethanolamine (PE), and SM to human platelets (10). Import of these phospholipids into platelets was independent of endocytosis. Apparently, phospholipids of the monolayer that surrounds lipoproteins are transferred to the outer leaflet of the plasma membrane of platelets. It is possible that such an alteration of the phospholipid composition of the platelet plasma membrane phospholipid uptake could modulate certain steps of the coagulation process. Activation of platelets is an initial event in blood coagulation. It is accompanied by rapid changes in platelet phospholipid homeostasis (11,12). In order to evaluate whether activation of platelets was associated with changes in phospholipid transfer from lipoproteins to platelets, we analyzed the effect of platelet agonists on the import of individual phospholipids.
Enrichment of LDL and HDL with Pyrene-labeled Phospholipids-Pyrene-labeled phospholipids were incorporated into lipoproteins essentially as described previously (10). Briefly, 1 mol of py-PC, py-PPC, py-PE, py-PPE, or py-SM were dissolved together with 3 mol of egg PC in 20 l of ethanol. The phospholipids were directly added to 3 ml of human plasma at 37°C under argon by very slow injection with gentle stirring. 0.74 mM diethyl-p-nitrophenylphosphate and 3 mM NaN 3 were added and the suspension incubated for 24 h at 37°C under argon. Pyrene-labeled lipoproteins were isolated by ultracentrifugation at 4°C (14). Low density and high density lipoproteins were recovered at densities (g/ml) of 1.019 -1.063 and 1.069 -1.210, respectively.
Before starting the incubations with the platelets, the lipoproteins were extensively dialyzed at 4°C under argon against a buffer containing 10 mM Tris-HCl, 150 mM NaCl, 0.3 mM EDTA (pH 7.4). The protein contents of the labeled lipoproteins were determined (15). Pyrene-labeled lipoproteins thus obtained migrated at the same height as native lipoproteins as evidenced by agarose gel electrophoresis. Lipids from LDL and HDL labeled with pyrene-containing phospholipids were extracted (16) and phospholipids separated by TLC as described previously (10). The plates were viewed under UV light. Fluorescence was detected exclusively in the phospholipid fraction originally labeled in the vesicles.
Incorporation of 14 C-Labeled Lipoproteins into LDL-14 C-Labeled phospholipids were incorporated into LDL particles according to a previously published protocol (10). Briefly, [ 14 C]PC, [ 14 C]PE, and [ 14 C]SM were incorporated into egg PC vesicles and, subsequently, incubated with fresh plasma for 24 h at 37°C under argon. 14 C-Labeled LDL was isolated by ultracentrifugation at 4°C (14). The specific activities thus obtained were 0.6 -9.1 ϫ 10 4 cpm/nmol of respective phospholipid. Before the start of incubations with platelets, the lipoproteins were extensively dialyzed at 4°C under argon against a buffer containing 10 mM Tris-HCl, 150 mM NaCl, 0.3 mM EDTA (pH 7.4). 14 C-Labeled lipoproteins migrated at the same height as the native lipoproteins as determined by agarose gel electrophoresis.
Incubation of Platelets with Labeled Lipoproteins and Vesicles-Blood (anticoagulated by 0.38% citrate) from healthy volunteers was centrifuged at 180 ϫ g for 20 min and the supernatant recovered. Platelet-rich plasma was subsequently centrifuged at 3,000 ϫ g for 3 min and the pellet washed two times at room temperature with a buffer consisting of (mM): 138 NaCl, 3 KCl, 1 MgCl 2 , 15 Hepes, 9 citrate, 5 EDTA, 5 glucose, containing, in addition, 350 mg/100 ml albumin and 3 units/ml apyrase (pH 6.3, "washing buffer"). After the second washing, the supernatant was removed and the cells suspended in the incubation buffer. The suspension was slowly warmed up to 37°C and, in most cases, incubated for 20 min with 2 mM EGTA before addition of thrombin (or other substances) and phospholipid donors.
Subsequently, platelets were suspended with either 14 C-labeled LDL, pyrene-labeled LDL, pyrene-labeled HDL, or vesicle suspensions containing pyrene phospholipids in incubation buffer at 37°C. For the preparation of phospholipid vesicles, 250 g of egg phosphatidylcholine and 5 g of py-PC, py-PPC, py-PE, py-PPE, or py-SM were dissolved in 30 l of ethanol and thereafter slowly injected under argon into 2 ml of a buffer containing (in mM): 138 NaCl, 3 KCl, 1 MgCl 2 , 15 Hepes, 9 citrate, 5 EDTA, 5 glucose (pH 7.4, "incubation buffer"). 30 l of the vesicle suspension was added to 1.97 ml of incubation buffer for the incubation with washed platelets.
In the case of pyrene-labeled donors, the fluorescence was monitored directly in the suspensions every 30 s (on-line conditions). After incubation with 14 C-labeled LDL, platelets were separated from the donors by centrifugation, washed once, and the platelet-associated radioactivity determined. In some cases, following incubation of platelets with pyrene-labeled donors, fluorescence intensities were also determined after separation of donors and acceptors by centrifugation (off-line experiments). Monomer and excimer fluorescence of the suspensions were determined at emission wavelengths of 380 nm and 480 nm, respectively, with excitation at 340 nm (excitation and emission slits of 5 and 10 nm). Incorporation of pyrene-labeled phospholipids into the platelets was followed by the increase in monomer intensity after addition of platelets to the donors. Fluorescence measurements were carried out using a Shimadzu RF-5001-PC spectrofluorometer (Shimadzu Germany, Duisburg, Germany).
Estimation of Phospholipid Transfer between Donor and Acceptor Vesicles-The transfer of py-PE between donor and acceptor vesicles was determined essentially as described (17). For the preparation of donor vesicles, 250 g of egg PC, 45 g of phosphatidic acid, and 7.5 g of py-PE were dissolved in ethanol and dispersed by slow injection into 300 l of a buffer containing 10 mM Tris, 0.02% NaN 3 (pH 7.4; "Tris buffer"). Acceptor vesicles were prepared by dissolving 10 mg of egg PC in ethanol dispersed in 1 ml of Tris buffer. Subsequently, 50 l of acceptor vesicle solution was mixed with 10 l of donor vesicle solution and 40 l of Tris buffer. To this suspension, 200 l of either Tris buffer or extracellular medium obtained from suspensions of untreated or activated platelets was added. The latter media were prepared by activating the platelets for 5 min at 37°C with either ADP (10 M) or thrombin (0.5 unit/ml) and a subsequent centrifugation step (1,500 ϫ g, 15 min). After 0 and 30 min, 100 l of the mixtures (consisting of the extracellular media plus the two types of vesicles) was loaded onto a small anion exchange column (DEAE-Sepharose C6-LB, 1 ml of column resin; pretreated with 100 l of the acceptor vesicle solution) and the column eluted with 1.5 ml of Tris buffer. Under these conditions, the negatively charged donor vesicles are retained while the uncharged acceptor vesicles are eluted from the column. The acceptor vesicles were solubilized by a 10-min treatment with 2% Triton X-100 at 37°C and their pyrene monomer fluorescence measured as described above. The transfer of py-PE into the acceptor vesicles was calculated as difference in the monomer fluorescence of the acceptor vesicles between 0 and 30 min of incubation.
Determination of Arachidonic Acid Metabolites-14 C-Labeled arachidonic acid and its metabolites were quantified as described in Ref. 18. After the end of incubation of platelets with [ 14 C]PE-LDL, 3 volumes of diethyl ether/methanol/0.2 M citric acid (30:4:1; v/v), which had been precooled to Ϫ20°C, were added to the suspensions. After vortexing the mixture was centrifuged, the organic layer evaporated and redissolved in ethyl acetate. The samples and standards were applied to TLC plates (20 ϫ 20 cm, Merck, Darmstadt), which had been pretreated for 60 min at 80°C. Separation was performed by one-dimensional TLC using the organic phase of ethyl acetate/isooctane/acetic acid/water (110:50:20: 100; v/v) as solvent. In this system, TXB 2 , HETE, and arachidonic acid migrated at R F values of 0.19, 0.64, and 0.75, respectively. The radioactivity in the spots corresponding to these R F values was determined in a liquid scintillation counter.
Assessment of Platelet Procoagulant Activity-Procoagulant activity in the presence of activated platelets was determined essentially as described previously (19). Platelets (5 ϫ 10 6 /0.5 ml) were incubated without or with the indicated concentrations of thrombin and collagen plus either LDL or phospholipid vesicles for 5 min at 37°C in a buffer composed of (in mM): 145 NaCl, 10 Hepes, 5 KCl, 1 MgCl 2 , 5 glucose (pH 7.4, "coagulation buffer"). 60 l of this suspension were added to 540 l of coagulation buffer containing, in addition, 5 mM CaCl 2 , 1.1 nM factor Va, and 0.52 nM factor Xa (final concentrations). Thereafter, factor II (suspended in coagulation buffer, final concentration 0.43 M) and the chromogenic substrate S-2238 were added (final concentration 0.26 mM). The increase in absorption at 405 nm was measured and compared with standard curves obtained with different concentrations of thrombin.
Miscellaneous Procedures-For assessment of changes in platelet morphology, samples of platelet suspensions were transferred into aggregometer cuvettes and incubated at 37°C while stirring (1,100 rpm). Platelet shape change and aggregation were monitored by measuring light transmission in an aggregometer (Fresenius, Bad Homburg, Germany). All mean values are given Ϯ S. D.  Fig. 2 (upper panel). The presence of thrombin augmented platelet-associated 14 C by 2.5-, 2.7-, and 2.6-fold after 30, 60, and 120 s, respectively. Iloprost pretreatment nearly completely prevented the increase in platelet-associated 14 C as induced by thrombin within the 120-s period (upper panel of Fig. 2). Transfer of phospholipids from plasma lipoproteins to platelets can be directly monitored by using donor lipoproteins containing py-labeled phospholipids (10). LDL particles were enriched in py-labeled choline phospholipids (PC, its plasmalogen analog plasmenylcholine (PPC), or SM) as well as in py-labeled ethanolamine phospholipids (py-PE, py-PPE). Platelets were incubated with LDL containing py-labeled phospholipids. For all phospholipids tested, the intensity of pyrene monomer fluorescence continuously increased within this time interval as measured under on-line conditions. In the same time period, excimer fluorescence decreased, thus indicating incorporation of py-phospholipids into the acceptor membrane. Thrombin (0.5 unit/ml) did not significantly alter the incorporation of the pyrene-labeled choline phospholipids PC, PPC, and SM (data not shown). In contrast, transfer of py-PE from LDL to platelets was increased by thrombin (Fig. 2, lower panel). A maximal 4.1-fold increase was observed after 1 min. Comparable results were obtained when the transfer of py-PE occurring within 1 min was determined under off-line conditions (from 2.9 Ϯ 0.7 (Ϫ thrombin) to 12.4 Ϯ 3.6 (ϩ thrombin) ng of pyrene incorporated/3 ϫ 10 8 platelets, n ϭ 4). Thrombin also enhanced incorporation of py-PPE by 3.8-fold, as assessed after a 1-min incubation of py-PPE-labeled LDL with platelets (from 4.8 Ϯ 0.6 to 18.2 Ϯ 3.8 ng of pyrene incorporated/3 ϫ 10 8 platelets, n ϭ 4). In platelets pretreated for 5 min with 100 nM iloprost thrombin only slightly increased import of py-PE (Fig. 2, lower panel).

Specific Enhancement of Import of Ethanolamine
Changes in platelet morphology induced by thrombin were evaluated under identical conditions as for the experiments with suspensions of py-labeled lipoproteins. To this purpose, light transmission was measured in an aggregometer while In some experiments, after incubation of platelets with lipoproteins, the suspensions were subsequently treated for 45 min at 37°C with a 40-fold excess of unlabeled LDL. In order to determine the localization of 14 C within the donor lipoproteins before the start of incubation with platelets, lipids were extracted from the lipoproteins (16). The lower (organic) phase was resuspended in CHCl 3 /CH 3 OH (2:1), dried and subjected to 1-dimensional TLC in CHCl 3 /CH 3 OH/H 2 O/NH 3 (90/54/5.5/ 5.5). More than 95% of total 14 C was present in the phospholipid fraction originally labeled. After incubation of platelets with 14 C-LDL (see above), platelets were also subjected to lipid extraction and neutral lipids as well as phospholipids separated by TLC as described (10). The percentages of platelet-associated 14 C present in the same phospholipid fraction as originally labeled was Ͼ92% (absence of thrombin) and Ͼ85% (presence of thrombin). Empty columns, without thrombin. Filled columns, with thrombin. ϩ, addition of unlabeled LDL. stirring the platelet suspension (1,100 rpm). After 10 -20 s, thrombin elicited a maximal reduction in light transmission indicating shape change. In contrast, in platelets pretreated for 5 min with 100 nM iloprost, the maximal decrease in light transmission induced by thrombin was reduced by 76% and 83% as compared with untreated platelets (values of two experiments). Thus, iloprost pretreatment inhibits both stimulation of uptake of ethanolamine phospholipids and platelet shape change.
In further experiments, it was investigated whether the increase in PE import elicited by thrombin was restricted to LDL as donor. HDL particles containing either py-PE or py-PPE were incubated for 1 min at 37°C with platelets. Thrombin increased the transfer of py-PE and py-PPE from HDL to platelets by 3.6-and 3.8-fold, respectively (Table I). Transfer of the two ethanolamine phospholipids was also enhanced by thrombin when lipid vesicles containing egg PC and either py-PE or py-PPE were employed as donors (Table I). Thrombin stimulated transfer of py-PE and py-PPE from vesicles to platelets by 4.2-and 4.4-fold, respectively. When platelets were incubated with HDL or lipid vesicles enriched in py-PE and subsequently platelet-associated pyrene contents assessed (offline experiments), the results obtained were essentially similar to those obtained in on-line experiments (data not shown). The effect of the thrombin inhibitors hirudin (2.5 units/ml) and dansylarginine (0.1 M) on transfer of py-labeled ethanolamine phospholipids from LDL or vesicles to platelets was investigated. Hirudin binds to thrombin, thereby impeding interaction of thrombin with the thrombin receptor (20), while dansylarginine blocks the proteolytic activity of thrombin (21). Both agents rather completely prevented the stimulation of py-PE and py-PPE transfer to platelets as elicited by thrombin (Table I).
In order to analyze whether other platelet agonists also affected transfer of py-PE from lipid vesicles to platelets, platelets were activated with either collagen (10 g/ml) or ADP (10 M). The incorporation of py-PC and py-SM into platelets was not influenced by the two agonists (Fig. 3). However, transfer of py-PE was stimulated by 2.5-and 2.8-fold upon addition of collagen and ADP, respectively. When collagen plus thrombin were present together, the stimulation increased from 3.9-fold (thrombin alone) to 6.3-fold (thrombin ϩ collagen). In the presence of ADP plus thrombin, the activation of py-PE import was 6.1-fold (thrombin ϩ ADP) (Fig. 3). Accordingly, both ADP and collagen additively activate the thrombin-induced transfer of py-PE from lipid vesicles to platelets.
As thrombin thus emerged as most potent stimulator of platelet PE uptake, the effect of thrombin was analyzed in more detail. Fig. 4 shows data on the concentration dependence of the action of thrombin on platelet-associated 14 C or pyrene as assessed after incubation of platelets with either LDL particles (labeled with 14 C-phospholipids) or egg PC vesicles (containing py-phospholipids). In platelets incubated with [ 14 C]PE-labeled LDL, platelet-associated [ 14 C]PE steeply increased when elevating the thrombin concentration from 0.005 to 0.1 unit/ml (Fig. 4, upper panel). At concentrations Ͼ0.1 unit/ml, no further enhancement was noted. Half-maximal stimulation was observed between 0.01 and 0.05 unit/ml. A comparable concentration dependence was observed when the effect of thrombin on transfer of py-PE from vesicles to platelets was analyzed (Fig. 4, lower panel). Half-maximal stimulation was noted at 0.02 unit of thrombin/ml. At all concentrations investigated, thrombin did not affect platelet [ 14 C]PC levels in platelets as determined after incubation with LDL labeled in [ 14 C]PC (Fig.  4, upper panel). The platelet agonist did also not affect transfer of py-PC (Fig. 4, lower panel).
Platelet Activation Mechanisms Involved in the Enhanced Import of Ethanolamine Phospholipids-In order to evaluate whether the platelet thrombin receptor mediated the thrombin-induced stimulation of import of ethanolamine phospholipids, the effect of TRAP was investigated. TRAP directly activates the thrombin receptor without displaying the proteolytic activity of thrombin (12). TRAP (15 M) increased transfer of py-PE and py-PPE from vesicles to platelets by 3.6-and 4.2fold, respectively. In platelets preincubated for 5 min with 100 nM iloprost, TRAP did not enhance the import of py-labeled ethanolamine phospholipids into platelets. Furthermore, incorporation of py-labeled choline phospholipids (PC, PPC, and SM) was unaffected by TRAP (Fig. 5, upper panel). With LDL as donor (30 g of LDL protein containing 0.7 g of py-PE), TRAP (15 M) increased the import of py-PE into platelets from 2.7 Ϯ 0.6 (without TRAP) to 10.5 Ϯ 3.4 ng of py-PE incorporated/3 ϫ TRAP thus mimicked the stimulation of transfer of ethanolamine phospholipids from different donors to platelets as induced by thrombin. This indicates that the effect of the latter agonist is mediated by the thrombin receptor.
When platelet PKC activity was directly stimulated with 12-O-tetradecanoylphorbol-13-acetate, platelet-associated 14 C was increased by 2.3-fold following incubation of platelets with [ 14 C]PE-labeled LDL (Table II). OAG, another direct activator of PKC, induced a 2.0-fold elevation of platelet-associated 14 C. The effects of the two stimulators were prevented by preincubation with staurosporine (Table II). The increase in plateletlinked 14 C induced by 12-O-tetradecanoylphorbol-13-acetate and OAG persisted when bound 14 C-LDL was removed by excess of unlabeled LDL (data not shown). Thus, enhancement of PE import by thrombin is partially blocked by PKC inhibitors and PKC activators stimulate the incorporation of PE suggesting involvement of PKC in the effect of the agonist on phospholipid uptake. Earlier work indicates that the inhibition of cyclooxygenase induced by weak platelet agonists such as ADP and epinephrine prevents the secretory response of the platelets while their capacity to induce shape change and aggregation is retained (22,23). Accordingly, the effect of ADP on the platelet uptake of py-labeled ethanolamine phospholipids was analyzed after pretreatment of the platelets with aspirin. The transfer was registered in the presence of apyrase (0.6 unit/ml, in order to avoid desensitization of the platelet ADP receptor by ADP; Ref. 24). In untreated platelets, ADP stimulated the uptake of py-PE and py-PPE by 3.2-and 4.3-fold, respectively (Table III). In aspirin-pretreated platelets, ADP only marginally increased platelet phospholipid uptake. Control experiments using the TABLE I Effect of hirudin and dansylarginine on thrombin-induced transfer of pyrene-labeled ethanolamine phospholipids from LDL, HDL, and lipid vesicles to platelets Hirudin (2.5 units/ml) or dansylarginine (0.1 M, both dissolved in incubation buffer) were given to the platelets (3 ϫ 10 8 /2 ml) 1 min before addition of either LDL (30 g of protein), HDL (80 g of protein), or vesicles of egg PC (3.8 g of phospholipid) enriched in pyrene-labeled phospholipids. The suspensions were incubated for 1 min at 37°C in the absence or presence of thrombin (0.5 unit/ml). The quantities of pyrene present in the lipoproteins and in the vesicles (determined as described in legend to Fig. 2) were 0.5 and 0.08 g, respectively. The increase in monomer fluorescence between 0 and 1 min of incubation of platelets with labeled particles as measured under on-line conditions was taken as incorporation of py-labeled ethanolamine phospholipids into platelets. Results are from experiments on platelets of four to six different donors. ND, not determined.  aggregometer indicated that, under the same experimental conditions, ADP was still able to induce platelet shape change and aggregation.
On the basis of these experiments, we tested the hypothesis whether the extracellular medium recovered from suspensions of activated platelets affected the transfer of py-PE between two types of lipid vesicles in vitro. Platelets were activated for 5 min with either ADP (10 M) or thrombin (0.5 unit/ml), the extracellular medium isolated by centrifugation and added to a suspension consisting of donor vesicles (containing py-PE) and acceptor vesicles. The transfer of py-PE to the acceptor vesicles was unaffected by the presence of the extracellular medium obtained from untreated platelets (Fig. 6). The supernatant recovered from ADP-and thrombin-activated platelets, in contrast, stimulated py-PE transfer by 10.4-and 11.0-fold, respectively. The increase in py-PE transfer elicited by the supernatant isolated from ADP-activated platelets was reduced by 85% when the platelets had been pretreated with aspirin (Fig. 6). In order to remove microvesicles shed from the activated platelets, the extracellular medium from thrombin-stimulated platelets was isolated and thereafter centrifuged at 12,000 ϫ g for 20 min. The supernatant thereby obtained stimulated py-PE transfer to an extent similar to that for the original extracellular medium (data not shown).
In further experiments, we evaluated whether proteins were involved in the stimulating effects of the extracellular media on py-PE transfer. After dialyzing the supernatants recovered

Influence of pretreatment with aspirin on ADP-induced stimulation of platelet uptake of py-labeled ethanolamine phospholipids
To the platelet-rich plasma 1 mM aspirin was added and the suspension incubated for 15 min at 37°C. Platelets (3 ϫ 10 8 /2 ml) were suspended in the incubation buffer (supplemented with 0.6 unit of apyrase/ml) and incubated at 37°C for 3 min with vesicles containing 3.8 g of egg PC and 0.08 g of either py-PE or py-PPE. The increase in monomer fluorescence during this time interval was measured under on-line conditions. Results are from experiments on platelets of four different donors and two donors (where no S.D. is given). from suspensions of platelets activated with either ADP or thrombin, the transfer of py-PE between donor and acceptor vesicles was still enhanced (98% and 82% of the stimulation obtained without dialysis; Fig. 6). The presence of either thrombin (0.5 unit/ml) or of apyrase alone (0.6 unit/ml) did not affect py-PE transfer between the two types of lipid vesicles (data not shown). When the supernatants obtained from ADP-or thrombin-activated platelets were briefly boiled, the stimulatory capacity was lowered by more than 80% (Fig. 6). In addition, after precipitation of the supernatant proteins with ammonium sulfate (3.3 M), the accelerating effect of the extracellular medium recovered from ADP-activated platelets on py-PE transfer was reduced by 74%. The extracellular medium from thrombin-activated platelets was isolated, dialyzed (in order to remove ADP) and thereafter added to a suspension consisting of untreated platelets (3 ϫ 10 8 /2 ml) and egg PC vesicles (3.8 g of phospholipid) enriched in py-PE (0.08 g). In order to prevent stimulation of platelet py-PE uptake by thrombin itself, dansylarginine (0.1 M; cf. Table I) was added. After a 3-min incubation at 37°C, platelet import of py-PE was measured under on-line conditions. Under these conditions, platelet py-PE uptake was increased by 6.4fold compared with the py-PE incorporation determined in the presence of an extracellular medium isolated from untreated platelets (0.46 Ϯ 0.18 (supernatant from untreated platelets) versus 2.93 Ϯ 0.55 ng of pyrene/3 ϫ 10 8 platelets (supernatant from thrombin-activated platelets), n ϭ 3). The enhancing effect was strongly reduced when the (dialyzed) extracellular medium recovered from the thrombin-activated platelets was briefly boiled or its proteins removed by precipitation with ammonium sulfate (data not shown). Together, these data suggest that proteinaceous products secreted by the platelets mediate the increased platelet uptake of py-PE after activation with ADP or thrombin.
Eicosanoid Production and Platelet Procoagulant Activity during Stimulation of PE Import-Upon activation with thrombin, stimulation of phospholipases A 2 induces liberation of arachidonic acid from platelet membrane phospholipids which, in turn, can be further metabolized to eicosanoids such as thromboxane A 2 (TXA 2 ) or 12-hydroxyeicosatetraenoic acid (HETE). After incubation of platelets with LDL particles con-taining 14 C-20:4-PE, the amount of 14 C associated with arachidonic acid, [ 14 C]TXB 2 (the stable hydrolysis product of TXA 2 ), and [ 14 C]HETE was determined (Table IV). As compared with control platelets, the amount of 14 C present in these substances increased by 4 -7-fold in platelets activated by thrombin (0.5 unit/ml) and by 6 -11-fold after stimulation with thrombin plus collagen (10 g/ml). The proportion of 14 C associated with TXB 2 and HETE was higher in activated platelets while the percentage of arachidonic acid was lower (Table IV). In separate experiments, LDL was enriched with 14 C-18:2-PE (instead of 14 C-20:4-PE) and incubated with untreated and thrombin-activated platelets under the same experimental conditions as those described in Table IV. [ 14 C]Linoleic acid was determined after the end of incubation by the same TLC separation as for [ 14 C]arachidonic acid (see "Experimental Procedures"). The amount of 14 C-18:2 determined was found to be increased by 1.3-fold in thrombin-activated compared with control platelets (mean of two separate experiments). Together, the data indicate that arachidonic acid-containing PE species imported into activated platelets are partially utilized for liberation of arachidonic acid and subsequent production of eicosanoids.
The procoagulant activity of platelets is supposed to be mainly due to an increase in phosphatidylserine (PS) content of the outer monolayer of the platelet plasma membrane (6, 7). However, recent data indicate that the presence of PE may also play an important role (25). It was thus tested whether the presence of a lipoprotein donor containing PE (such as LDL) affected platelet prothrombinase activity. Control experiments indicated that LDL particles did not contain any measurable PS (data not shown). Platelets were incubated for 5 min with thrombin (0.5 unit/ml), a time-interval yielding maximal stimulation of the platelet procoagulant activity under the experimental conditions applied (data not shown). Thrombin as well as thrombin plus collagen (10 mg/ml) increased the platelet procoagulant activity by 4.5-and 10.4-fold (Fig. 7). The presence of LDL further enhanced the stimulation induced by the platelet agonists resulting in a 10.5-fold (thrombin) and 18.3fold elevation (thrombin ϩ collagen) of the platelet prothrombinase activity as compared with untreated platelets.
In order to assess whether the phospholipid components of LDL could be responsible for the effect of the particles on platelet procoagulant activity, thrombin-activated platelets were coincubated with pure PC vesicles. The prothrombinase activity of thrombin-activated platelets remained unaffected (Fig. 7). When 4.8 mol % of the PC vesicles was substituted by either PE or PPE, the prothrombinase activity of thrombinactivated platelets was increased by 2.5-and 3.0-fold, respectively. The procoagulant activity of platelets stimulated with thrombin plus collagen was increased by 2.3-fold with PC/PE   (Fig. 7). The presence of either LDL (50 g of LDL protein/ml) or PC vesicles (63 nmol/ml) enriched with 4.8 mol % ethanolamine phospholipids) did not affect the prothrombinase activity of untreated platelets (data not shown). The concentration dependence of the effect of thrombin on platelet procoagulant activity in the presence of PC/PPE vesicles is shown in the lower panel of Fig. 7. Half-maximal stimulation of the prothrombinase activity was observed at 0.05 unit of thrombin/ ml. A maximal effect was seen with 0.1 and 0.5 unit of thrombin/ml. Taken together, the data indicate that the augmented PE transfer from LDL and lipid vesicles to platelets elicited by thrombin is associated with an increase in the procoagulant activity of platelets. DISCUSSION We observed in the present study that the platelet stimuli thrombin, collagen, and ADP specifically enhance the import of ethanolamine phospholipids into platelets. Platelets possess specific binding sites for LDL on their plasma membranes, which include glycoproteins IIb/IIIa (26). Import of phospholipids into platelets may in principle be facilitated by increased binding of lipoprotein particles, as this might enhance the physical interaction between donor and acceptor monolayers. However, if so, increased binding of LDL to platelets would also be expected to stimulate incorporation of choline phospholipids into platelets (PC, PPC, and SM). This was not the case ( Figs.  1 and 4). Furthermore, thrombin barely affected binding of 125 I-LDL to the platelets (Fig. 1). In addition, as in both untreated and thrombin-activated platelets nearly 90% of bound 125 I-apo B could be removed by an excess of unlabeled LDL, it is unlikely that under the experimental conditions employed the platelet agonist induced endocytosis of LDL particles. Thus, the stimulatory action of thrombin on transfer of ethanolamine phospholipids to platelets is apparently not mediated by increased platelet binding of LDL or by endocytosis of the lipoproteins.
In order to gain insight into the mechanism(s) responsible for the stimulatory effect of thrombin on transfer of ethanolamine phospholipids, the involvement of donor lipoproteins was evaluated. Transfer of py-labeled ethanolamine phospholipids was accelerated to a similar extent using HDL or egg PC vesicles compared with LDL as donors (Table I). The concentration dependence of the thrombin effect was also comparable using either LDL (labeled with [ 14 C]PE) or lipid vesicles (labeled with py-PE) as donors (Fig. 4). We conclude that LDL or HDL apoproteins are not involved in the thrombin-induced stimulation of PE transfer to platelets.
In a second set of experiments, the role of the platelet acceptors for the effect of thrombin on PE transfer was evaluated. Thrombin cleaves the platelet thrombin receptor, thereby creating a new N-terminal portion of the receptor that functions as a tethered ligand to activate the receptor (12). Both proteolysis and binding to the receptor are necessary to activate intracellular signaling pathways in platelets. Inhibition of the proteolytic function of thrombin by dansylarginine prevented the stimulatory action of thrombin on PE import (Table I). On the other hand, addition of TRAP, which activates the thrombin receptor, induced a 4-fold stimulation of import of ethanolamine phospholipids (Fig. 5). Inhibitors of PKC partially reduced the thrombin-induced stimulation of transfer of labeled PE into platelets while direct activators of PKC enhanced PE import (Table II). PKC is part of the signaling cascade coupled downstream to the platelet thrombin receptor. Together, the data indicate that the effect of thrombin on incorporation of ethanolamine phospholipids is mediated by activation of the thrombin receptor and requires signaling through PKC.
ADP and collagen were also found to stimulate platelet uptake of ethanolamine phospholipids (Fig. 3). When the secretory response of ADP-activated platelets was prevented by aspirin pretreatment (22), the stimulation of uptake of pylabeled ethanolamine phospholipids was reduced (Table III). We therefore tested the hypothesis that the enhanced uptake of the ethanolamine phospholipids by activated platelets might be mediated by platelet secretory products. The extracellular media obtained from suspensions of either ADP-or thrombinactivated platelets markedly accelerated the transfer of py-PE between two types of lipid vesicles (Fig. 6). In the same in vitro system of lipid vesicles, we attempted to gain some insight into the nature of the factors(s) involved in the enhancement of py-PE transfer. Different experimental approaches indicated that the products secreted by the platelets responsible for the augmented transfer of the ethanolamine phospholipids were likely to be proteins. In future work, we plan to isolate and identify these proteins.
The ethanolamine phospholipids of plasma lipoproteins are particularly enriched in arachidonic acid (27). The results of the present study indicate that the transfer of PE-bound [ 14 C]arachidonic acid from lipoproteins to platelets is increased by thrombin. Previous studies imply that the arachidonic acid esterified to platelet membrane ethanolamine phospholipids can be released by stimulation of phospholipase A 2 activity as elicited by thrombin (28 -30). We observed increased production of [ 14 C]arachidonic acid and its metabolites [ 14 C]thromboxane B 2 and [ 14 C]HETE in thrombin-(and collagen)-activated platelets after incubation with LDL containing 14 C-20:4-PE (see "Results"). The free arachidonic acid is most probably generated by hydrolysis of platelet 14 C-20:4-PE mediated by platelet cytosolic phospholipase A 2 (cPLA 2 ). Subsequently, the polyunsaturated fatty acid is metabolized to the eicosanoids. Secretory phospholipase A 2 (sPLA 2 ), which is released by activated platelets, could additionally liberate the arachidonic acid FIG. 7. Procoagulant activity of platelets in the presence of LDL and phospholipid vesicles. Platelets (1 ϫ 10 7 /ml) were incubated without or with thrombin (0.5 unit/ml) for 5 min at 37°C. In some samples either LDL (50 g/ml) or vesicles containing egg PC (63 nmol/ ml) or egg PC (60 nmol/ml) plus ethanolamine phospholipids (PE and PPE; 3 nmol/ml) were present (as indicated). Subsequently, the suspension was first incubated with 1.1 nM factor Va and 0.52 nM factor Xa (in the presence of CaCl 2 , see "Experimental Procedures") and thereafter incubated with factor II (0.43 M). Prothrombinase activity was measured by using the chromogenic substrate S-2238 (0.26 mM). Upper panel, Co, no addition; T, presence of thrombin (0.5 unit/ml); C, presence of collagen (10 g/ml). Columns with horizontal bars indicate the presence of donors with ethanolamine phospholipids. Lower panel, concentration dependence of the effect of thrombin on prothrombinase activity in the presence of PC/PPE vesicles. from extracellular 14 C-20:4-PE present in LDL (or microvesicles shed from platelets). However, EDTA, which inactivates sPLA 2 but not cPLA 2 (31), was present in the incubation buffer, thus arguing against involvement of sPLA 2 . Furthermore, platelet cPLA 2 preferentially hydrolyzes phospholipid species with arachidonic acid at sn 2 while sPLA 2 shows no fatty acid specificity for the sn 2 position (31). Enhanced import of 14 C-20:4-PE into activated platelets was accompanied by a severalfold increase in release of 14 C-20:4 (Table IV). A comparable elevation in the incorporation of 14 C-18:2-PE ( Fig. 1) barely altered production of 14 C-18:2 (see "Results"). This preferential liberation of 20:4 over 18:2 is again in support of the view that platelet cPLA 2 is most likely responsible for the release of [ 14 C]arachidonic acid under the experimental conditions employed.
The appearance of PS in the outer monolayer of platelet plasma membranes is thought to be responsible for the procoagulatory effect of activated platelets. The magnitude of increase in aminophospholipid content depends on the strength and type of stimulation. The platelet agonists thrombin or collagen, when present alone, elicit only minor increases (7). Recent data indicate that, apart from PS, PE may also play a decisive role (25). Under physiological conditions, platelets are surrounded by HDL and LDL. When shape change, the earliest functional response on platelet activation, was nearly completely prevented by preincubation with iloprost, stimulation of import of ethanolamine phospholipids into platelets by thrombin was greatly reduced (Figs. 2 and 5). Interestingly, a previous study indicates that in platelets suspended in platelet-rich plasma (containing the lipoproteins), prostacyclin inhibits platelet procoagulant activity (32), in line with the above mentioned hypothesis.
We observed in the present investigation that LDL enhanced the procoagulant activity of platelets stimulated with thrombin alone or with thrombin plus collagen (Fig. 7). The effect elicited by LDL was mimicked by egg PC vesicles containing ethanolamine phospholipids. Pure PC vesicles were without effect. Therefore, under the experimental conditions employed, enhancement of platelet procoagulant activity in platelets activated by thrombin and collagen requires donors containing ethanolamine phospholipids. The concentration dependence of the thrombin-elicited stimulation of py-PE import (Fig. 4) and the one of thrombin-induced procoagulant activity in the presence of PC/PPE vesicles (Fig. 7) were comparable. Together, our data indicate that the acceleration of import of ethanolamine phospholipids is associated with an increased procoagulatory activity of platelets.
In conclusion, by activating the platelet thrombin receptor and signaling through PKC thrombin specifically stimulates the transfer of ethanolamine phospholipids from donor lipopro-teins to platelet acceptors. The enhanced transfer of ethanolamine phospholipids is independent of the apoprotein component of lipoproteins and does not involve the endocytosis of LDL. The platelets stimuli ADP and collagen mimic the effect of thrombin. The stimulation of platelet import of ethanolamine phospholipids is apparently mediated by proteins secreted from the activated platelets. Thus, during platelet activation, import of phospholipids from extracellular donors to platelets is specifically regulated.