The transfer of VLDL-associated phospholipids to activated platelets depends upon cytosolic phospholipase A2 activity.

We previously reported that VLDL could transfer phospholipids (PLs) to activated platelets. To identify the metabolic pathway involved in this process, the transfer of radiolabeled PLs from VLDL (200 μM PL) to platelets (2 × 108/ml) was measured after incubations of 1 h at 37°C, with or without thrombin (0.1 U/ml) or LPL (500 ng/ml), in the presence of various inhibitors, including aspirin, a cyclooxygenase inhibitor (300 μM); esculetin, a 12-lipoxygenase inhibitor (20 μM); methyl-arachidonyl-fluorophosphonate (MAFP), a phospholipase A2 (PLA2) inhibitor (100 μM); 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis (acetoxymethyl) ester (BAPTA-AM), a Ca2+ chelator (20 μM); bromoenol lactone (BEL), a Ca2+- independent phospholipase A2 (iPLA2) inhibitor (100 nM); and 1-[6-[[17β-3-methoxyestra-1,3,5(10)-trien-17-yl-]amino]hexyl]1H-pyrrole-2,5-dione (U73122), a phospholipase C (PLC) inhibitor (20 μM). Aspirin and esculetin had no effect, showing that PL transfer was not dependent upon cyclooxygenase or lipoxygenase pathways. The transfer of PL was inhibited by MAFP, U73122, and BAPTA-AM. Although MAFP inhibited both cytosolic phospholipase A2 (cPLA2) and iPLA2, only cPLA2 is a calcium-dependent enzyme. Because calcium mobilization is favored by PLC and inhibited by BAPTA-AM, the transfer of PL from VLDL to platelets appeared to result from a cPLA2-dependent process. The inhibition of iPLA2 by BEL had no effect on PL transfers.

More recently, we considered the possibility that VLDLassociated PLs also could be transferred to platelets (9). Albeit less abundant than LDL and HDL in fasting condi-tions, VLDL may have increased postprandial concentrations. VLDLs are secreted by the liver in the circulation, where they undergo hydrolysis of their core triglyceride content through the successive actions of LPL and hepatic lipase, which ultimately results in the formation of LDL (10,11). During this process, the excess of VLDL surface components, including apolipoproteins, cholesterol, and PLs, is released from the particles. Although a large part of cholesterol and PL is transferred to HDL (12,13), our work demonstrated that VLDL-associated PLs can also be transferred to platelets and that these transfers are favored by LPL and platelet activation (9). This effect of LPL results from two different actions. First, the LPL-mediated lipolysis of VLDL destabilizes the particle surface, thereby favoring the release of PLs. Second, the fatty acids released during lipolysis favor platelet activation, as judged from their increased thromboxane production. That the transfer of PL from VLDL to platelets depended upon platelet activation was confirmed by our observation that thrombin stimulates both platelet thromboxane production and PL transfer.
Although platelets are able to import PL from various lipoprotein fractions, the underlying mechanisms of these transfers appeared to be complex. They are independent of lipoprotein binding and internalization (7). In agreement with this concept, the scavenger receptor class B type 1, which can mediate the specific import of PL into various cells, was shown to be absent in platelets (8). However, major differences emerge when comparing the PL transfers obtained from the different lipoprotein fractions. The transfer of LDL-or HDL-derived PE into platelets, but not that of PC or sphingomyelin, was stimulated by platelet activators, including thrombin, collagen, and ADP, and was dependent upon the secretion of an unidentified cellular protein factor (14). In contrast, there was no apparent specificity of the PL species transferred from VLDL to platelets (9). Both LPL and thrombin stimulated the import by platelets of VLDL-derived palmitoyl-arachidonyl-PC, palmitoyl-arachidonyl-PE, and dipalmitoyl-PC with similar efficiencies. Thus, the regulation of PL uptake by platelets appears to dramatically depend upon the lipoprotein used as the donor. Although LDL might preferentially transfer certain PL species, namely PE, by a specific mechanism, VLDL could supply all types of PLs to platelets without consideration of their nature. This concept prompted us to further characterize the metabolic pathway governing the transfer of PL from VLDL to platelets. In this work, using a variety of metabolic inhibitors, we present in vitro evidence showing that this transfer results from a cPLA 2 -dependent process. ethane-N,N,N ¶,N ¶-tetraacetic acid tetrakis (acetoxymethyl) ester (BAPTA-AM), and polyoxyethylen-9-laurylether were obtained from Sigma Chemical (St. Louis, MO). PLs were assayed using enzymatic kits from Wako Chemicals GmbH (Neuss, Germany). Thromboxane B 2 (TXB 2 ) concentrations were determined using the enzyme immunoassay Biotrak system from Amersham Biosciences (Orsay, France). Esculetin, methyl arachidonyl fluorophosphonate (MAFP), 1-[6-[[17b-3-methoxyestra-1,3,5(10)trien-17-yl-]amino]hexyl]1H-pyrrole-2,5-dione (U73122), and bromoenol lactone (BEL) were from Biomol (Plymouth Meeting, PA).

Materials
Isolation and labeling of lipoproteins VLDL (d , 1.006 g/ml) and lipoprotein-deprived plasma (d . 1.21 g/ml) were isolated from human plasma by preparative ultracentrifugation (15). Depending upon the volume of plasma, the ultracentrifugation was performed either in a Beckman LE 80K using a 50.2 fixed-angle rotor or in a Beckman TL-100 tabletop ultracentrifuge using a TLA 100.3 fixed-angle rotor. The resulting preparations were then dialyzed extensively against a buffer containing 150 mM NaCl, 1 mM EDTA, 10 mM Tris-HCl, and 10 mM NaN 3 , pH 7.4. VLDLs were labeled with [ 14 C]PAPC (1 mCi/10 mmol lipoprotein PL) as reported previously (16). The desired amount of radioactive label was dried under nitrogen, solubilized in ethanol, and added to VLDL under vortexing. To avoid destruction of the lipoprotein structure, the final proportion of ethanol in the samples was maintained at ,1% (v/v). The samples were then incubated for 3 h at 37jC, and the labeled VLDLs were reisolated by ultracentrifugation.

Platelet isolation
Fresh blood was collected at the local blood bank (Etablissement Français du Sang) from healthy volunteers. Blood was drawn into a one-seventh volume of a solution containing 19.6 mM citric acid, 89.4 mM sodium citrate, 16.1 mM NaH 2 PO 4 , and 128.7 mM dextrose, pH 5.6. The platelet isolation procedure was essentially based on that described previously (17). Briefly, plateletrich plasma was obtained after blood centrifugation at 200 g for 17 min at 20jC and acidified to pH 6.4 with 0.15 M citric acid. Platelets were immediately pelleted by centrifugation at 900 g for 12 min and washed in acidified lipoprotein-deprived plasma. After repelleting, the platelets were finally washed and resuspended in a Tyrode-HEPES buffer solution containing 137 mM NaCl, 2.7 mM KCl, 0.41 mM NaH 2 PO 4 , 11.9 mM NaHCO 3 , 1 mM MgCl 2 , 5.5 mM glucose, and 5 mM HEPES, pH 7.35. Platelet suspensions were left for 1 h at room temperature before experiments were started.

Platelet aggregation
To avoid abnormal experimental data that could result from unknown medical treatment taken by blood donors, each platelet preparation was controlled for its functional ability to aggregate before being used in our studies. Aggregations were induced by arachidonic acid and performed in a Chronolog dual-channel aggregometer (Coulter, Margency, France) according to the turbidimetric method of Born (18).

Transfers of labeled PLs from VLDL to platelets
Labeled VLDLs (200 nmol/ml PL) were incubated with platelets (3 3 10 8 cells) in a final volume of 1.5 ml at 37jC for 1 h (9). For each experiment, the VLDLs and platelets were each isolated from a single donor. At the end of the incubations, platelets were separated from the medium by centrifugation. The pelleted platelets were first washed in plasma to remove nonspecifically adsorbed labeled VLDL and then in Tyrode-HEPES buffer. The final pellets were dissolved by overnight incubation in 0.25 ml of 0.4% polyoxyethylen-9-laurylether and counted for radioactivity. The results were finally corrected for nonspecifically adsorbed radioactivity at time 0 (nonincubated platelets).

Determination of TXB 2 production
The production of TXB 2 by platelets was determined at the beginning and end of incubations carried out under various conditions as described above. However, platelets were not separated from their media before the TXB 2 assay, thereby permitting the measurement of total TXB 2 .

Determination of the [1-14 C]arachidonic acid content of platelet PLs
PL hydrolysis was assayed in platelets prelabeled with [1-14 C] arachidonic acid (20 nCi/ml) for 1 h at 37jC (27). After separation from the medium by centrifugation, platelets were washed in Tyrode-HEPES buffer and incubated for 5 min in the absence or presence of thrombin or cPLA 2 inhibitors. For these particular experiments, no VLDL was added to the incubation medium to avoid PL transfers. At the end of the incubations, platelets were washed and resuspended in 1 ml of buffer. Lipids were then extracted and separated by thin-layer chromatography in a solvent composed of hexane-diethyl ether-acetic acid (80:20:1). The PL spots were finally scraped off and counted for radioactivity.

Determination of plasma phospholipid transfer protein activity
Phospholipid transfer protein (PLTP) activity was determined using an in vitro assay in which we measured the transfer of radio-labeled PL from VLDL to HDL using delipidated plasma as the source of PLTP, as described previously (16). Basal as well as PLTP-facilitated PL transfers were measured in the absence or presence of cPLA 2 inhibitors at the same concentrations as those used with platelets. The results are expressed as percentages of PL transferred during 2 h incubations.

RESULTS
To explore the dependence of VLDL-associated PL transfer to platelets upon platelet activation, we performed a first series of experiments in which various inhibitors were used to block different metabolic pathways involved in platelet activation. The transfers of [ 14 C]PAPC from VLDLs to platelets were measured after incubations of 1 h at 37jC. They were stimulated by ?2.5and 3.5-fold when platelets were activated by either thrombin (0.1 U/ml) or LPL (500 ng/ml), respectively (Fig. 1). When the experiments were performed in the presence of aspirin, a cyclooxygenase inhibitor, at a concentration (300 mM) known to block the formation of thromboxane, no change was observed in the transfers of PL whether stimulated or not by either thrombin or LPL. Similarly, the transfers of PL remained unchanged when the 12-lipoxygenase-dependent metabolic pathway was inhibited by esculetin (20 mM). In contrast, when the incubations were carried out in the presence of MAFP, a PLA 2 inhibitor, the stimulating effects of both thrombin and LPL on PL transfers were abolished, resulting in values comparable to those of controls.
Three different PLA 2 enzymes are present in human platelets: cPLA 2 , iPLA 2 , and secretory PLA 2 . Only the two former may be affected by MAFP. However, their activities can be discriminated on the basis of their Ca 21 dependence, because cPLA 2 but not iPLA 2 is a Ca 21 -dependent enzyme. This prompted us to study the thrombin-or LPLstimulated transfers of [ 14 C]PAPC from VLDL to platelets in the presence of either BAPTA-AM, a Ca 21 chelator, or U73122, a PLC inhibitor (Fig. 2). The results showed that both BAPTA-AM and U73122 inhibited the stimulating effects of thrombin and LPL, clearly suggesting that PL import into platelets was controlled by a cPLA 2 -dependent process rather than an iPLA 2 -dependent process. This was confirmed by the results obtained when the incubations were performed in the presence of 100 nM BEL. The latter, which at this concentration inhibits iPLA 2 but not cPLA 2 (28), had no effect on PL transfers, whether stimulated or not by either thrombin or LPL. To control that metabolic inhibitors did not directly modify the ability of VLDL to transfer PLs, in particular by inhibiting the activity of PLTP that could be present at the lipoprotein surface, we performed in vitro assays in which we measured the transfers of radiolabeled PLs from VLDL to HDL ( Table 1). None of MAFP, BAPTA-AM, or U73122 had any effect on basal or PLTP-facilitated PL transfers.
Finally, we determined the metabolic effects of BAPTA-AM and U73122 on platelet PL hydrolysis and TXB 2 pro-duction. In the absence of inhibitors, we observed a decrease of the arachidonic content of PLs that was stimulated by thrombin, clearly showing a major hydrolysis ( Table 2). In the presence of U73122, which inhibits both cPLA 2 and PLC, PL hydrolysis was strongly inhibited, whereas it was only partially decreased in the presence of BAPTA-AM, which inhibits only cPLA 2 . In addition, clear stimulating effects of thrombin and LPL emerged when the platelet production of TXB 2 was studied (Fig. 3). The thrombin-stimulated production of TXB 2 was totally inhibited by BAPTA-AM or U73122, whereas they only partially decreased the stimulating effect of LPL. As expected, BEL, which specifically inhibits iPLA 2 , had no effect on TXB 2 production.

DISCUSSION
We recently showed that platelets are able to import PL from VLDL and that this process is stimulated when platelets are activated through the action of thrombin or LPL (9). The present work was intended to characterize the  Values shown are means 6 SEM. Platelets prelabeled with [1-14 C] arachidonic acid were incubated for 5 min at 37jC, in a total volume of 1.5 ml, in the presence or absence of thrombin (0.1 U/ml) and metabolic inhibitors. After lipid extraction and separation by thin-layer chromatography, PLs were counted for radioactivity as described in Materials and Methods. Data are expressed as percentages of radiolabeled arachidonic acid content at time 0. relationship between the transfer of PL from VLDL to platelets and platelet activation. The latter results from the combination of complex metabolic pathways that can be summarized as follows. Plasma membrane PL may be hydrolyzed by cPLA 2 , resulting in the release of fatty acids. Among these, arachidonic acid can be metabolized through two different pathways (29)(30)(31). The first, which depends upon cyclooxygenase activity, leads to the formation of prostaglandins and thromboxane A 2 , a very potent platelet activator. The second pathway is governed by the action of 12-lipoxygenase, which transforms arachidonic acid in 12-hydroperoxy-eicosatetraenoic acid, a metabolite also able to favor platelet activation. In addition, phosphatidylinositol diphosphate may also be hydrolyzed by PLC, leading to the formation of diacylglycerols and inositol triphosphate. Although diacylglycerols stimulate metabolic cascades resulting in protein phosphorylation, granule secretion, and the release of fatty acids by diacylglycerol and monoacylglycerol lipases, inositol triphosphate favors the increase of Ca 21 intracellular concentration (3,4).
To study the putative involvement of these metabolic pathways, we measured the transfer of PL from VLDL to platelets in the presence of aspirin or esculetin, which inhibit cyclooxygenase and 12-lipoxygenase, respectively (20,21). We observed no changes in the transfers of PL to platelets, whether stimulated or not by thrombin or LPL, indicating that neither cyclooxygenase-nor 12lipoxygenase-dependent processes were directly responsible for the magnitude of PL transfers. In contrast, these PL transfers were decreased when the incubations were performed in the presence of MAFP, an inhibitor of PLA 2 . PLA 2 is a superfamily of enzymes consisting of secretory and intracellular species. The latter comprise cPLA 2 and iPLA 2 , which are both inhibited by MAFP (22,23). To discriminate between these two enzymes, we took advantage of the differences in their mechanism of action. cPLA 2 but not iPLA 2 is a Ca 21 -dependent enzyme (28). The activation of cPLA 2 depends upon two synergistic processes: the catalytic activity of the enzyme requires a previous phosphorylation, and its translocation to the membrane necessitates the mobilization of cytosolic Ca 21 (32,33). Because the latter may be regulated by the PLC pathway, we measured the thrombin-and LPL-stimulated transfers of PL from VLDL to platelets in the presence of U73122, a PLC inhibitor (19). The PL transfers were clearly inhibited. However, as mentioned above, in addition to inositol triphosphate-stimulated Ca 21 mobilization, the action of PLC generates diacylglycerols that stimulate different metabolic cascades (3,4). Thus, to distinguish between these various effects, we performed experiments in which U73122 was substituted for BAPTA-AM, a Ca 21 chelator (25,26). The PL transfers were decreased similarly, showing that the effect of PLC was attributable to its ability to mobilize cytosolic Ca 21 . Overall consideration of these results clearly suggests that the transfer of PL from VLDL to platelets depends upon a cPLA 2 -dependent process rather than an iPLA 2dependent process. This concept was finally confirmed by our observation that PL transfers remained unaffected when iPLA 2 was specifically inhibited in the presence of BEL (24).
In addition to PL transfers, the importance of cPLA 2 in platelet metabolism was assessed by measuring the arachidonic acid content of PLs in platelets and their TXB 2 production. A strong PL hydrolysis was observed during platelet incubations, which was in part dependent upon cPLA 2 activity. As shown previously (9), both thrombin and LPL stimulated platelet TXB 2 production. The stimulating effect of thrombin was totally inhibited by BAPTA-AM or U73122, whereas that of LPL was only partially decreased. This was expected, because LPLstimulated platelet activation is initially attributable to the uptake by platelets of the fatty acids released during LPL-mediated lipolysis of VLDL (9).
Although the question of how the transfers of PL are precisely affected by cPLA 2 activity has not been addressed directly in this work, a likely mechanism may be considered. Because the cPLA 2 -stimulated hydrolysis of PL occurs at the inner leaflet of the plasma membrane, the enzyme activity necessarily results in disequilibrium of the PL concentrations between the membrane inner and outer leaflets. Several physiological processes are known to lead to comparable disequilibrium, which is compensated by flip-flop mechanisms (34,35). Thus, on the basis of a similar mechanism, we speculate that, in our case, the net translocation of PL from the outer to the inner leaflet of the membrane would cause a PL deficit in the outer leaflet that might be compensated by the import of PL from VLDL. Fig. 3. Effects of cPLA 2 and iPLA 2 inhibitors on the production of thromboxane B 2 (TXB 2 ) by platelets. Platelets were incubated with VLDL (200 nmol PL/ml) for 1 h at 37jC, in a total volume of 1.5 ml, in the absence (white bars) or presence of U73122 (20 mM), a PLC inhibitor (black bars); BAPTA-AM (20 mM), a Ca 21 chelator (hatched bars); or BEL (100 nM), an iPLA 2 inhibitor (gray bars). The thrombin-stimulated (0.1 U/ml) or LPL-stimulated (500 ng/ml) productions of TXB 2 are expressed as percentages of those obtained in basal conditions. TXB 2 concentrations were measured using a commercial kit, as described in Materials and Methods. Values shown are means 6 SEM from three independent experiments.