Eur. J. Biochem. 236, (1996) 0 FEBS 1996

Save this PDF as:

Size: px
Start display at page:

Download "Eur. J. Biochem. 236, (1996) 0 FEBS 1996"


1 Eur. J. Biochem. 236, (1996) 0 FEBS 1996 Phagocytic activation induces formation of platelet-activating factor in human monocyte-derived macrophages and in macrophage-derived foam cells Relevance to the inflammatory reaction in atherogenesis Christine DENTAN, Philippe LESNIK, M. John CHAPMAN and Ewa NINIO Institut National de la SantC et de la Recherche MCdicale U-321, Unit6 de recherches sur Les LipoprotCines et l AthCrogCnkse, H8pital de la Pitit, Paris, France (Received 9 November 1995) - EJB 95 1$58/1 Monocyte-derived macrophages and macrophage-derived foam cells in arterial tissue may undergo phagocytic activation and thereby contribute to an inflammatory reaction. We have investigated the effect of phagocytic activation on the formation of platelet-activating factor (l-o-alkyl-2-acetyl-sn-glycero-3- phosphocholine, PAF-acether, PAF), a proinflammatory phospholipid, in human monocyte-derived macrophages (macrophages) and in cholesterol-loaded macrophage foam cells (foam cells). Adherent human monocyte-derived macrophages were transformed into foam cells upon incubation with acetylated lowdensity lipoproteins (Ac-LDL). Such foam cells characteristically displayed a markedly increased content of cholesteryl esters compared with macrophages ( pglpg DNA and pg/pg DNA, n = 5, respectively). After phagocytic stimulation with serum-opsonized zymosan (OPZ), both macrophages and foam cells synthesized PAF transiently with maximal production ( pmol PAF/pg DNA, n = 5, corresponding to pmol PAF/106 cells, as assessed by bioassay) occurring approximately 15 min after stimulation. A major fraction of the synthesized PAF remained cell-associated; such PAF was composed mainly of the hexadecyl (16 : 0 PAF, -75 %) and the octadecenyl (18 : 1 PAF) species and of trace amounts of octadecyl (18:O PAF), as assessed by reverse-phase liquid chromatography. Addition of exogenous 16:O lyso-paf alone triggered PAF formation ( pmol PAF/pg DNA, after 15 min of cellular stimulation) ; simultaneous cellular stimulation with OPZ and 16 : 0 lyso-paf increased PAF formation in an additive manner. Acetyltransferase, the enzyme which acetylates the precursor lyso-paf and transforms it into PAF, displayed elevated activity both in macrophages and in foam cells, attaining pmol PAF formed. min-. mg DNA (n = 4); such elevated activity was not increased by OPZ-stimulation. The activity of acetylhydrolase, the PAF-degrading enzyme, was similar in macrophages and in foam cells, and varied between 120 pmol and 320 pmol PAF degraded. min-. mg DNA- (n = 5). Cell-associated acetylhydrolase activity was increased significantly by 40? 15 % (P< 0.003, n = 5) after min of activation with OPZ compared with non-stimulated cells and may account for the rapid decrease in cellular PAF content observed approximately 30 min after stimulation. These studies have established that metabolism of PAF in foam cells closely resembles that in macrophages, and thus PAF metabolism is largely independent of cellular cholesterol content. Moreover our data are consistent with the hypothesis that both macrophages and macrophage-derived foam cells upon phagocytic-activation constitute a significant transient source of PAF at inflammatory sites in the arterial intima where this phospholipidic mediator may exert potent proatherogenic and prothrombotic effects. Keywords : platelet-activating factor ; acetylhydrolase ; macrophage foam cells ; cholesterol ; atherogenesis. Platelet-activating factor (1-O-alkyl-2-acetyl-sn-glycero-3- of PAF synthesis is the remodelling route; this pathway inkolves phosphocholine, PAF-acether, PAF) is a potent inflammatory the activation of acetyltransferase, which acetylates lyso-paf mediator which is synthesized by several types of cells upon (l-o-alkyl-sn-glycero-3-phosphocholine), converting it to PAF activation, including monocytes, macrophages, endothelial cells and platelets (reviewed in references [l, 21). The major pathway Correspondence to E. Ninio, INSERM U-321, Unit6 de recherches sur Les LipoprotCines et 1 AthCrogCnese, Pavillon Benjamin Delessert, HBpital de la PitiC, 83 Bd. de I HGpital, F Paris Cedex 13, France Abbreviations. GroPCho, sn-glycero-3-phosphocholine; PAF, PAF- acether, platelet-activating factor, 1-0-alkyl-2-acetyl-GroPCho; lyso- PAF, 1-0-alkyl-GroPCho; LDL, low-density lipoprotein; HDL, highdensity lipoprotein; OPZ, serum-opsonised zymosan : Ac-LDL, acetylated LDL; 16:O PAF, 1-0-hexadecyl-2-acetyl-GroPCho; 18:O PAF, 1-0-octadecyl-2-acetyl-GroPCho; 18 : 1 PAF, 1-0-octadecenyl-2-acetyl- GroPCho. Ei1zq me.r.. Acetyltransferase (EC ); acetylhydrolase (EC ). (reviewed in reference [3]). A significant fraction of cellular PAF is secreted and transported in the circulation in association with albumin, although a minor fraction is associated with plasma lipoproteins [4]. The effects of PAF on target cells are mediated at extremely low concentrations ( M) via a specific receptor, which has been detected on a variety of cell types and cloned from guinea pig lung tissue [5], and from human leukocytes [6]. Circulating human monocytes express a functional PAF receptor at their surface and contain the corresponding mrna [7]. The levels of intracellular and secreted PAF are controlled by acetylhydrolase, a calcium-independent enzyme, which hydrolyzes PAF, converting it to lyso-paf (reviewed in reference

2 Dentan et al. (EUK J. Biochem. 236) 49 [S]). Acetylhydrolase is secreted by human macrophages [9], platelets [lo] and hepatocytes [I I], and is principally transported in the circulation in association with low-density and high-density lipoproteins (LDL and HDL, respectively) 112, 131. Acetylhydrolase appears to be a key enzyme in the pathophysiologic regulation of PAF-mediated effects 1141 and has been recently cloned from human monocyte-derived macrophages [ 141. It is now well established that PAF plays a major role in several pathologies, such as anaphylactic shock and ischemia (reviewed in reference [ 11). Moreover, substantial evidence suggests that PAF exerts proatherogenic effects in the microenvironment of the arterial intima and that in addition this phospholipidic mediator plays a major role in subsequent thrombotic complications (reviewed in reference [ 151). Furthermore, PAF has been identified as a component of the atheromatous plaque [16], and animal studies have demonstrated the protective effect of PAF antagonists against atherosclerosis [17]. PAF induces release of several substances which may exert proatherogenic effects, including elastase [ 181, which degrades components of extracellular matrix of the intima and media, and active oxygen species [19, 201, which induce tissue damage and contribute to oxidation of LDL [21]. PAF is therefore implicated in inflammatory events in atherogenesis which are intimately associated with the activation of monocytes, macrophages, endothelial cells and platelets (reviewed in reference [21]). Such activated cells are known to synthesize PAF (reviewed in references [I, 2, 151). It was therefore of considerable interest to study the synthesis of PAF by cholesterol-loaded foam cells, which are characteristic of atherosclerotic lesions [21]. Foam cells originate in large part from monocyte-derived macrophages, accumulating large amounts of cholesteryl esters as a result of their unregulated uptake of modified LDL [22]. Oxidized and modified LDL have been identified in human atherosclerotic plaques [21] and aggregated forms of these lipoproteins are internalized by a scavenger-receptor-independent route involving phagocytosis via Fc receptors [23]. Crystals of cholesterol, a typical feature of lipid-rich atherosclerotic plaques, may be internalised by macrophages via a phagocytic mechanism [24]. Furthermore, macrophages may phagocytose cellular fragments that result from apoptosis and necrosis in plaque tissue [25]. Thus in the present study, we have employed serum-opsonized zymosan (OPZ) as a well-established model of phagocytic activation involving Fc receptors [ 181. Using this approach, we investigated the effect of phagocytic activation upon the synthesis of PAF in human monocyte-derived macrophages (denoted macrophages) and cholesterol-loaded macrophage foam cells (denoted foam cells). Macrophages were transformed into foam cells upon uptake of acetylated LDL (Ac-LDL) via scavenger receptors [22, 241. Our studies show that phagocytic activation of macrophages and foam cells with OPZ results in synthesis of PAF, apparently via the remodelling pathway ; acetyltransferase, whose activity was elevated in such cells, but which was insensitive to OPZ stimulation, is a key regulator of this pathway. By contrast, acetylhydrolase activity increased significantly in OPZ-stimulated foam cells in parallel with PAF biosynthesis, suggesting that this enzyme is implicated in regulation of cellular PAF content. These results provide new evidence that PAF is produced transiently by macrophages and cholesterol-loaded macrophage foam cells activated by phagocytosis. PAF of macrophage origin may therefore be bioactive in the microenvironment of the atheromatous plaque where it can exert potent proinflammatory, proatherogenic and prothrombotic effects. EXPERIMENTAL PROCEDURES Materials. Zymosan, acetyl-coenzyme A (acetyl-coa), 1- O-stearoyl-2-arachidonoyl-sn-glycero-3-phosphocholine, fatty acid-free BSA and DNA from calf thymus, were purchased from Sigma (Chemical Co.). Lipase A1 from Rhizopus arrhizus, Nutridoma HU, lactate dehydrogenase assay were supplied by Boehringer-Mannheim, and bicinchoninic acid protein reagent by Pierce. The PAF antagonist BN (ginkgolide, terpbne) purified from a natural extract and the synthetic PAF antagonist, WEB 2086 (triazolobenzodiazepine), were the gifts from Institut Henri Beaufour (Le Plessis Robinson, France) and Boehringer- Mannheim, respectively. RPMI 1640, with or without phenol red, culture medium was obtained from Eurobio, and gentamicin from Schering-Plough. Anti-CD68 (clone KP1 ), anti-cd3 (clone T3-4B5) and goat anti-mouse-fitc-conjugated antibodies were from Dako. Chromogenic Limulus amebocyte lysate assay was purchased from Biogenic. 1-0-Hexadecyl-2-acetyl- GroPCho (16: 0 PAF), 1-0-hexadecyl-GroPCho (16 : 0 lyso- PAF), 1-0-octadecenyl-GroPCho (18 : 1 lyso-paf) and 1-0-octadecyl-GroPCho (18: 0 lyso-paf) were supplied by Novabiochem. Standards used in normal-phase HPLC were dipalmitoyl- GroP [N-methyl-'H]choline (30-60 Ci/mmol), [I-alkyl-'Hllyso- PAF (30-60 Ci/mmol) and [2-~cetyl-~H]16:0 PAF (10-30 Ci/ mmol; Du Pont-New England Nuclear). Standards used in reverse-phase-hplc were [2-acetyl-'H]16: 0 PAF (10-30 Ci/ mmol), [l-octudecyl-'h]paf (18:O PAF, Ci/mmol, Amersham International) and [1-oct~decenyl-~H]PAF (18: 1 PAF), which was prepared by enzymatic acetylation of 18 : 1 lyso-paf in the presence of [~cetyl-~h]coa(du Pont-New England Nuclear) [26]. Purification and chemical modification of LDL. LDL was isolated in the density interval g/ml by sequentjal preparative ultracentrifugation from normolipidemic human plasma [27) to which gentamicin (50 pg/ml), EDTA (final concentration, 0.03 mm), and a serine-protease inhibitor phenylmethylsulfonyl fluoride (final concentration, 1 mm) were added. The isolated LDL was dialyzed extensively against 0.01 M NaCI/P, at ph 7.4 (139 mm NaCI, 5 mm Na,HPO,. 12H,O, 5 mm NaH,PO,. H,O) at 4 C and the purity of each LDL preparation was evaluated as described earlier [27]. The contents of conjugated dienes, thiobarbituric acid-reactive substances and lipoperoxides in freshly isolated LDL were assessed as described earlier and were characteristic of native LDL Acetylated LDL (Ac-LDL) was prepared according to the procedure of Basu et al. [29]. The elevated net electrical charge on Ac-LDL, compared with that of the native LDL from which it was derived, was estimated by electrophoresis in agarose gel (Corning) [30]. Ac-LDL was extensively dialyzed at 4"C, firstly for 24 h against NaCUP, at ph 7.4 and secondly for 4 h against RPMI Ac-LDL was subsequently filtered through a 0.22-pm filter (Costar) and the protein content of lipoproteins was determined by the bicinchoninic acid using BSA as the standard. Preparation of serum-opsonized zymosan, acetyl-coa and 16:O lyso-paf. OPZ was prepared according to Nakagawara et al. [31]. A stock solution (20 mg/ml) in 150 mm NaCl was stored in aliquots at -80 C. Acetyl-CoA (10 mm) was dissolved in Hepes buffer (4 mm, ph 5) and stored in small aliquots at -80 C. 16:0, 18: 1 and 18:O lyso-paf (1 mm) were dissolved separately in 150 mm NaCl containing 0.25 % fatty-acid-free BSA and stored at -20 C. Isolation, culture and cholesterol loading of human monocyte-derived macrophages. Monocytes were isolated from the blood of healthy, normolipidemic volunteers (thrombopheresis residues), as previously described [18]. The cells were cultured and grown in 35 mmx 10 mm plastic tissue culture dishes (Primaria, Falcon, Becton Dickinson) with RPMI 1640 medium containing 10 % heat-inactivated pooled human serum and

3 50 Dentan et al. ( Em J. Biochem. 236) 40 pg/ml of gentamicin At day 8 of culture, cells were washed three times with serum-free RPMI 1640 medium containing 40 pg/ml gentamicin. They were differentiated into adherent monocyte-derived macrophages (hereafter denoted as macrophages) and were free of lymphocytes. Cholesterol-loaded macrophages were obtained by incubation of macrophages for 48 h in RPMI 1640 medium supplemented with 40 pgfml gentamicin, 1 % Nutridoma HU and 100 pg Ac-LDL proteinlml 118, 241 ; these cells will hereafter be denoted as macrophage-derived foam cells, or foam cells. Macrophages were washed and incubated in similar medium but in which Ac-LDL was omitted. Determination of intracellular cholesterol and DNA content, cell viability and endotoxin content of OPZ and Ac- LDL. Intracellular cholesterol was extracted as described by Gamble et al. [321. Total and free cellular cholesterol contents were determined using an enzymatic kit (BiomCrieux). The cellular mass of esterified cholesterol was estimated as (total cholesterol mass-free cholesterol mass) X 1.67 ; 1.67 is the factor representing the ratio of the average molecular mass of cholesteryl ester to free cholesterol [27]. The viability of cells after incubation for 48 h with Ac-LDL and of macrophages was assessed by measuring the release of lactate. dehydrogenase, a cytosolic enzyme, into the extracellular medium, with a commercial kit. The total cellular content of DNA in each pellet of cellular ethanolic extracts was estimated by the method of Burton et al. [33], using DNA from calf thymus as the standard. Results for formed PAF are expressed as pmol PAF/pg cellular DNA; acetylhydrolase and acetyltransferase activities are expressed/pg cellular DNA calculated as the mean value for several wells. The number of cells at day 10 of culture varied from to 0.750X10h cells/well, as deduced from their DNA content. We assumed that lo6 cells correspond to 8 pg DNA. The endotoxin content of OPZ and Ac-LDL preparations was measured prior to addition to the cellular medium using the chromogenic Lirnulus amebocyte lysate assay. Stimulation of cells. At day 10 of culture, macrophages and foam cells were washed three times with RPMI 1640 without phenol red but containing o/o fatty acid-free BSA; the latter binds PAF secreted by activated cells. Cells were warmed for 15 min at 37 C in the latter medium prior to cellular stimulation which was performed in duplicate wells for various periods of time with OPZ (1 mg/ml, final concentration). In some experiments, acetyl-coa (200 pm, final concentration) or 16 : 0 lyso- PAF, 18: 1 lyso-paf or 18:O lyso-paf (40 pm, final concentration) were added to the medium either alone or in the presence of OPZ. Control dishes, in which OPZ, acetyl-coa and lyso- PAF were omitted, were incubated under similar conditions. At the end of the incubation period, the reaction was stopped either by extraction of extracellular and cellular PAF with ethanol (80% final content, by vol.) for PAF quantification, or by cooling (4 C) for assay of cellular acetyltransferase, or by cooling and addition of EDTA (4 mm, final concentration) for cellular and extracellular acetylhydrolase assay. PAF assay. PAF secreted into the incubation medium and cell-associated PAF were separately extracted with ethanol (final concentration 80%, by vol.) for at least 1 h at room temperature. Ethanolic extracts of cellular supernatants and of cells scraped off the plastic dishes were centrifuged (500Xg, for 15 min) and brought to dryness under vacuum. Samples were kept at 4 C for PAF bioassay and for further purification and analysis [34]. Dry cellular and extracellular ethanolic extracts were redissolved in a small volume of ethanol (60%, by vol.) for quantitation of PAF by the thromboxane-a2-independent and ADP-independent aggregation of washed rabbit platelets as previously described [35]. The aggregating activity of the samples was mea- sured over the linear portion of the calibration curve performed with 5-50 pg synthetic 18:0 PAF. The results are expressed as equivalent pmol of PAF/pg cellular DNA. Total PAF was calculated as the sum of supernatant and cell-associated PAF. The specific PAF antagonists BN (60 pm) [36] and WEB 2086 (4 pm) [371 totally inhibited platelet aggregation induced by all extracts. Various samples were treated with 1000 U/ml lipase A1 from Rhizopus arrhizus [38] and plaielet aggregation was again measured for estimation of the percentage of platelet aggregation induced by PAF and by the snl Ester analog of PAF. Lipase A1 from R. urrhizus exclusively hydrolyzes the snl ester bond of the ester analog of PAF, but not the snl ether bond of PAF. Analysis of PAF molecular species. In the next step, ccllular and extracellular ethanolic extracts obtained under similar conditions of stimulation were pooled and brought to dryness. The dry residues were suspended in 300 p1 of HPLC mobile phase (dichloromethane/methanol/water, 400: 333 : 33 by vol.) and purified on a Microporasil column (Waters Ass.) at a tlow rate of 1 ml/min [39]. 'H-labelled dipalmitoyl-gropcho, lyso- PAF and PAF standards were used as markers. Fractions were collected, dried and assayed for PAF content. The tubes containing PAF activity were pooled to separate the molecular species of PAF on a reverse-phase Spherisorb C, HPLC column (Touzart et Matignon) using ammonium acetate (10 mm)/acetonitrile/ methanol (120:140:40, by vol.) as the mobile phase at a tlow rate of 1 ml/min. The retention times of PAF molecular species were determined using 'H-labelled 16:0, 18:0 and 1S:l PAF standards as described in [26]. Fractions were collected, dried and assayed for PAF content. The yield of platelet-aggregating activity of PAF upon separation by normal-pha~e and re\ mephase HPLC varied in the range 70-80%. Acetyltransferase assay. Cells were scraped off in NaWP,, centrifuged (500Xg 10min, 4"C), resuspended in a UaCl solution (150 mm) and disrupted by sonication at 4 C for 15 s (Sonics and Materials Inc., Danbury). The acetyltransferax activity in cell lysates was immediately measured as previously described [40] in the presence of 40 pm 16:O lyso-pai- and 200 pm [acetyl-'hiaccoa (0.3 pci/loo nmol) for 15 min at 37 C. After subtraction of controls incubated in the 'I 'b\ence of lyso-paf, the results are expressed as pinol PAF forned. min-' pg DNA-'. Acetylhydrolase assay. The cellular supernatants were immediately separated from the cellular pellets and kept at 4 C for acetylhydrolase assay. Cells were scraped off into the buffer used in the acetylhydrolase assay and disrupted by sonication at 4 C for 15 s (Sonics and Materials Inc.). Acetylhydrolase iictivity was measured in cell supernatants and in cell lysates according to Palmantier et al. [41] with some modifications as described earlier [28] in the presence of 50 pm (final concentration) l-o-hexadecyl-2-[~h]acetyl-.rn-glycero-3-phosphocholine ([ucetyz-'h]paf, 0.15 pcu2.5 nmol) for 10 min at 37 C. After subtraction of controls containing the heat-denatured enzyme from human serum, results are expressed as pmol PAF degraded min-'. pg DNA '. The specificity of acetylhydrolase towards PAF was verified in various samples, suggesting that the activity that we measured was not classical calcium-dependent phospholipase A?. On the one hand, cold l-o-stearoyl-2-arachidonoyl-sn-glycero-3-phosphocholine (50 pm), the substrate for phospholipase A,, was added to the assay mixture together with [ncetyl-'hipaf (50 pm) and acetylhydrolase activity was totally recovered (data not shown). On the other hand, when EDTA (2 mm) was replaced by CaC1, (10 mm) in the acetylhydrolase reaction buffer, (~cetyl-~h]paf degradation was not modified (data not shown).

4 Dentan et al. ( Em J. Biochem. 236) 51 Table 1. PAF content in OPZ-stimulated human foam cells and OPZstimulated macrophages in the presence of exogenous 16: 0 lyso-paf. Human foam cells and macrophages were incubated at 37 C for 15 min in the presence of OPZ (1 mg/ml) and/or 16:0 lyso-paf (40 pm) PAF was extracted and quantitated as described in the legend to Fig. 1. Treatment PAF content of foam cells macrophages (n = 2) (n = 3) Time (min) Fig. 1. Effect of phagocytic stimulation with OPZ on the PAF content in human macrophage-derived foam cells and macrophages. Foam cells (0) and macrophages (0) were incubated at 37 C for 7-60 min with OPZ (1 mg/ml, final concentration). Secreted and cell-associated PAF were extracted with ethanol and bioassayed using washed rabbit platelets as described in Experimental Procedures. Total PAF was calculated as the sum of secreted and cell-associated PAE The specific PAF antagonists, BN (60 pm) and WEB 2086 (4 pm), totally inhibited platelet aggregation induced by all cell extracts. The results are means L SD of duplicate determinations (coefficient of variation < 20%) of five experiments using cells from separate donors. PAF contents in OPZstimulated foam cells and in OPZ-stimulated macrophages were not statistically different (P> 0.05). Statistical analysis. Results are expressed as mean values t SD. Mean values were compared by the Student's t test with significance defined as P < RESULTS Characterization of foam cells. Human macrophages differentiated for 8 days from monc,luclear cell cultures were exclusively CD68 positive and CD3 negative as visualized by the indirect immunostaining method. In agreement with our previous studies [24], exposure of such macrophages to Ac-LDL (1 00 yg protein/ml) for 48 h at 37 "C resulted in substantial elevation in the cellular content of cholesterol and particularly of cholesteryl esters, transforming them into foam cells. Such foam cells showed a 3.3 t0.8(n = 5)-fold increase in total cellular cholesterol content compared with macrophages ; such foam cells contained 3.2i0.7 pg free cholesterol/pg DNA (n = S), cholesteryl esters ( pg/pg DNA, n = 5) representing more than SO % of total cellular cholesterol. Cholesteryl ester content increased approximately 20-fold in comparison with macrophages, in which sterol esters were barely detectable (0.2? 0.3 pg/pg DNA, n = 5). Macrophages typically possessed a low level of free cholesterol ( pg/pg DNA, n = 5). Finally, incubation of Ac-LDL with macrophages for 48 h was non-toxic to the cells, as there was no significant increase in the release of the cytosolic enzyme lactate dehydrogenase into the extracellular medium (1654% and 15?3%, n = 3, for foam cells and macrophages, respectively). Morever, the endotoxin content of Ac-LDL was less than SO pg/loo pg Ac-LDL protein; such levels of endotoxin are lower than those required for alteration of PAF metabolism 142, 431. PAF formation in OPZ-stimulated human foam cells and OPZ-stimulated macrophages. The formation of PAF by foam cells in response to OPZ was compared with that of macrophages stimulated under the same conditions, and data are presented in Fig. 1. The endotoxin content of OPZ was less than 3 np/mg; such levels of endotoxin are lower than those required OPZ 16 : 0 lyso-paf : 0 Iyso-PAF + OPZ pmol/pg DNA ?: ] " 0.6 for alteration of PAF metabolism [42, 431. PAF was quantitated by the thromboxane-a2-independent and ADP-independent aggregation of washed rabbit platelets. Large interassay variations were observed in the quantity and the time-course of PAF production by foam cells and macrophages from different donors, as already observed in human cells [9]. However, in each experiment, the time course of PAF formation by OPZ-stimulated foam cells and OPZ-stimulated macrophages showed a similar profile (P > 0.05). PAF formation started immediately after addition of OPZ to the cellular medium and reached a maximal value 7-30 min after OPZ-induced stimulation (Fig. 1). Cell-associated PAF content decreased rapidly after attaining a peak level. In response to OPZ, foam cells synthesized similar quantities of PAF compared with macrophages (P>O.O5). The maximal production of PAF formed by both OPZ-stimulated foam cells and OPZ-stimulated macrophages varied in the range l.l(o.86i0.21) pmol/pg DNA (n = S), corresponding to ( ) pmol/loh cells. Only a small fraction of the total PAF formed ( , IZ = 3) was secreted into the medium by stimulated cells. No PAF activity was recovered from non-stimulated control cells (foam cells and macrophages). In three experiments, the time-course of PAF formation induced by OPZ in the presence or absence of acetyl-coa (200 ym) showed no significant difference (data not shown). In contrast, addition of 16: 0 lyso-paf (40 pm) alone induced PAF formation in both macrophages and foam cells with similar magnitude to that upon OPZ-triggered stimulation, PAF formation varying in the range pmol/pg DNA (n = 3) after IS min of stimulation. Simultaneous stimulation of cells with OPZ and 16:O lyso-paf increased PAF formation in an additive manner (Table 1). Similarly, when 18:l lyso-paf was added to the cells, PAF was formed, however to a lesser extent compared with 16:0 lyso-paf supplementation. Addition of 18 : 0 lyso-paf was not accompanied by PAF formation (data not shown). Analysis of PAF molecular species. PAF synthesized by OPZstimulated cells was first purified by normal-phase HPLC. The major portion of PAF activity was recovered in a single peak which eluted with the [acetyl-'hipaf external standard. Trace amounts of aggregating activity eluted in fractions corresponding to the elution time of phosphatidylcholine and could correspond to the peak X already described by Benveniste et al. [4]. Analysis of PAF molecular species purified by normal-phase

5 52 Dentan et al. (Eul: J. Biochem. 236) 10- =' v E c* d :l 1x:o xo 701 I T t Time (rnin) Fig. 2. Reverse-phase HPLC analysis of the alkyl chains of PAF from OPZ-stimulated human foam cells and OPZ-stimulated macrophages. Foam cells (0) and macrophages (0) were incubated at 37 C for 15 min with OPZ (1 mg/ml, findl concentration). Secreted and cellassociated PAF were extracted with ethanol and bioassayed. Samples from several experiments were pooled and PAF was purified by normalphase HPLC using a Micropordsil column. The separation of PAF molecular species was performed by reverse-phase HPLC using a Spherisorb c, column as described in Experimental Procedures. The arrows indicate the retention time of external 'H-radiolabelled standards. PAF was bioassayed in 1-ml fractions after evaporation of the mobile phase, using washed rabbit platelets. The percentage of each PAF molecular species is expressed as a percentage of total PAF aggregating activity calculated as the sum of the aggregating activities of 16:0, 18:O and 18:l PAF of stimulated macrophages and stimulated foam cells. HPLC was achieved by reverse-phase HPLC. Three peaks of PAF activity were recovered when ethanolic extracts of OPZstimulated-foam cells and OPZ-stimulated-macrophages were analyzed (Fig. 2). The first major peak, eluting at min, corresponded to 16:O PAF; the second peak, eluting at min, corresponded to 18: 1 PAF; and the third, minor peak, eluting around 46 min, corresponded to 1 8 : 0 PAF. The relative proportions of the three molecular species of PAF were similar in extracts of OPZ-stimulated-foam cells and OPZ-stimulatedmacrophages (Fig. 2). Similar elution patterns were observed in separations of PAF by normal-phase and reverse-phase HPLC in experiments in which acetyl-coa was added to OPZ for cellular stimulation (data not shown). Lipase A1 analysis of PAF and the ester-analog of PAF. Pooled cellular ethanolic extracts were treated in two experiments with lipase A1 from R. urrhizus. In OPZ-stimulated-foam cells and OPZ-stimulated-macrophages, 80% and 82% of the aggregating activity were recovered, respectively, after enzymatic treatment compared with control samples which had not been treated with lipase A1. Recovered aggregating activity corresponded to PAF activity since it was totally abolished by the specific PAF antagonists (BN and WEB 2086). Under the same conditions of lipase A1 treatment, the aggregating activity of the synthetic ester analog of PAF was totally lost, whereas that of synthetic PAF was completely recovered. These results show that both OPZ-stimulated foam cells and OPZ-stimulated macrophages synthesized a significant amount of the ester analog of PAF, together with PAF. Acetyltransferase activity in OPZ-stimulated human foam cells and OPZ-stimulated macrophages. The time-course of acetyltransferase activity was investigated for 3-60 min in foam cells and macrophages both challenged with OPZ. The acetyltransferase activity in cell lysates was quantified by the incorporation of [3H]acetate from [ucetyl-'h]coa into synthetic lyso- PAF. The basal level of acetyltransferase activity was similar in foam cells and macrophages (P > 0.05) and was elevated com Time (min) Fig. 3. Effect of phagocytic stimulation with OPZ on acetylhydrolase activity in human macrophage-derived foam cells and macrophages. Foam cells (0) and macrophages (0) were incubated at 37 C for 7-60 min with OPZ (1 mg/ml, final concentration). In control dishes, i'oam cells (A) and macrophages (A) were incubated under similar conditions but in the absence of OPZ; both non-stimulated cell types exhihired a similar constant basal level of total acetylhydrolase activity (P > 0.05) which varied in the range pmol PAF degraded. min-'. pg DNA-'. Total acetylhydrolase activity was calculated as the sum of secreted and cell-associated acetylhydrolase activities which were determined by the measurement of free ['Hlacetate released from 51) pm [ucetyl-'hipaf after incubation for 10 min. The results are means?sd of duplicate determinations (coefficient of variation < 20%) of fiw experiments using cells from separate donors. The time-course of the evolution of acetylhydrolase activity in OPZ-stimulated foam cells and in OPZ-stimulated macrophages was not statistically different (P > (1.05). (*), P < as compared with non-stimulated cells (A, A). pared with the enzymatic activity of other PAF-forming cells measured in our previous studies [44]. Such high activity, which varied in the range ( ) pmol PAF formed. min-'. pg DNA-' (n = 4) corresponded to ( ) pmol PAF formed. min-'. 10F cells. The level of acetyltransferase activity was not increased in OPZ-activated macrophages and in OPZ-activated foam cells relative to nonstimulated cells (controls). Acetyltransferase activities in cells stimulated by OPZ were not significantly increased (1.05-fold?0.14, 1.07-fold?0.07 and 1.04-fold?0.10 at 3, 7, 15 nun of stimulation, respectively, relative to control cells ; P > 0.05 ). Acetylhydrolase activity in OPZ-stimulated human foam cells and OPZ-stimulated macrophages. The time-course of acetylhydrolase activity was investigated in foam cells and macrophages both challenged with OPZ. Acetylhydrolase activity was determined in supernatants and sonicated cells by the measurement of free ['Hlacetate released from 50 pm [mx~yl- 3H]PAF. Both non-stimulated control cell types exhibited a constant basal level of acetylhydrolase activity when incubated for up to 60 min in fresh medium. This level was similar in foam cells and in macrophages (P>O.OS) and varied in the range (216292)pmol PAF degraded. min-'. pg DVA ' (n = S), corresponding to (1728t-736) pmol PAF degraded. min. 10.' cells. Both types of cells secreted only a small fraction of enzyme activity (11.5 t 4.6 % of total acetylhydrolase activity, n = 5). The time-course of total acetylhydrolase activity in OPZstimulated cells is shown in Fig. 3. In each experiment, the tiniecourse of total acetylhydrolase activity by OPZ-stimulated foam cells and OPZ-stimulated macrophages showed a similar profile (P > 0.05). Total acetylhydrolase activity increased rapidly after addition of OPZ to the cellular medium and reached a plateau after approximately 20 min (Fig. 3), corresponding to a signifi-

6 Dentan et al. (EM J. Biochern. 236) 53 cant increase of 40 -C 15 % (P < 0.003, n = 5) both in foam cells and macrophages upon OPZ stimulation. In three experiments, comparison of the time-course of cell-associated and secreted acetylhydrolase from OPZ-stimulated cells in the presence or absence of acetyl-coa showed no significant difference (data not shown). DISCUSSION In this study, we have demonstrated that human monocytederived macrophages and macrophage-derived foam cells synthesize the inflammatory mediator, PAF, upon stimulation of their phagocytic activity by OPZ. We used a well-established model of human cholesterol-loaded macrophage foam cells that implicates an unregulated uptake of Ac-LDL via scavenger receptors [22, 241. The phagocytic activation of both macrophages and macrophage-derived foam cells appears to be an important feature in the development of the atherosclerotic plaque Our results strongly suggest therefore that the phagocytic activation of both cell types results in PAF production in fatty streak and advanced lesions characterised by the presence of numerous foam cells. Thus, a part of the potential proatherogenic and prothrombotic effects of PAF present in atheromatous plaques [161 may be attributed to PAF both of macrophage and of macrophage-derived foam cell origin. Synthesis of PAF in OPZ-stimulated, cholesterol-loaded macrophage foam cells closely resembled that in macrophages upon OPZ stimulation as shown by the time-course of PAF formation, by the relative proportions of cell-associated and secreted PAF, by the composition of the various molecular species of PAF and by the synthesis of the ester analog of PAF. Moreover, marked similarities in acetyltransferase activity and in the time-course profile of the increase in acetylhydrolase activity in OPZ-stimulated foam cells and in their non-cholesterolloaded counterparts were also observed. Thus, the increase in cellular cholesterol content of foam cells and especially of cholesteryl esters, compared with macrophages, does not appear to affect the metabolism of PAF to a significant degree. This finding is especially interesting, as the expression of several proteins appears to be upregulated in cholesterol-loaded macrophage foam cells; these include that of apo-e [45] and tissue factor [24] and elastase-like activity [18]. By contrast, the expression of both acetyltransferase and acetylhydrolase appears to be insensitive to cellular cholesterol content. PAF synthesis rapidly reached peak levels (at 7-30 min) in both OPZ-stimulated macrophages and OPZ-stimulated foam cells, thereby resembling the response typical of other PAFforming cells. We assume that such synthesis occurs via the remodelling pathway, which is sensitive to activation by OPZ [3]. However, we were unable to demonstrate activation of acetyltransferase upon OPZ-induced stimulation as usually observed in other cellular models : a 3-5-fold increase of acetyltransferase activity is typically observed in OPZ-stimulated human monocytes, neutrophils and murine macrophages (reviewed in reference [3]). The basal activity of acetyltransferase, which attained 1280? 540 pmol PAF formed. min-'. 10-' cells in both foam cells and macrophages, was elevated compared with the enzymatic activity detected in our previous studies under the same conditions in human monocytes (13? 6 pmol PAF formed. min-'. cells) and in neutrophils (60213 pmol PAF formed. min-l. cells) Such elevated acetyltransferase activity, which has not as yet been reported in the literature, may be due to permanent activation of this enzyme in human macrophages and in foam cells. In both non-stimulated and OPZ-activated cell types (foam cells and macrophages), the ad- dition of acetyl-coa to the cellular medium did not affect the metabolism of PAF and more especially did not increase the level of PAF formation in contrast to in vivo activated murine macrophages 126, 461. However, addition of 16:0 lyso-paf triggered PAF formation alone and also increased OPZ-stimulated formation twofold, suggesting that the basal activity of acetyltransferase is sufficient to produce PAF even in the absence of a phagocytic stimulus. In murine mast cells, PAF synthesis is triggered by addition of exogenous lyso-paf by two mechanisms: either by direct acetylation of added lyso-paf or by the CoAindependent transacylase reaction that generates lyso-paf from internal stores [47]. Both mechanisms require the presence of high acetyltransferase activity. In human neutrophils stimulated by phagocytosis, lyso-paf is also a limiting substrate, but when added exogenously together with OPZ, it enhances PAF formation twofold [44]. The synthesis of PAF by stimulated macrophages and stimulated foam cells was transient, probably reflecting the rapid increase in PAF-degrading acetylhydrolase activity which occurred during the 15-min period immediately after OPZ stimulation. Both PAF and acetylhydrolase were primarily cell associated. Acetylhydrolase may therefore act to control cellular levels of PAF and thus may be responsible for the rapid decrease observed in cellular PAF content which commenced approximately 20 min after OPZ stimulation of PAF synthesis (Fig. 1). A parallel increase in cellular acetylhydrolase activity and PAF formation has also been observed in activated human neutrophils and eosinophils [48]. Moreover, when monocytes mature to adherent macrophages, increase of basal cellular and extracellular acetylhydrolase activities parallels a decrease in the capacity of the cells to synthesize and secrete PAF upon stimulation 19, 411. Finally, we observed that the basal cellular acetylhydrolase activity of foam cells and of macrophages was alike, suggesting that Ac-LDL-associated acetylhydrolase is probably catabolized by lysosomal enzymes, as occurs for other constituents of Ac- LDL [22]. As in many PAF-forming cells (reviewed in references 11, 2, 1 S]), with the exception of circulating monocytes [44] and neutrophils [44, 341, most of the synthesized PAF remains cell associated in macrophages and in foam cells. Several functions of cell-associated PAF have been documented in cells. Membrane-bound PAF was shown to act as an adhesion molecule in human neutrophils for platelets [49]. Moreover, PAF bound to endothelial cells activates secretion of monocyte chemotactic protein-i in P-selectin adherent human monocytes via activation of nuclear factor KB [SO]. Equally, PAF which is synthesized by human endothelial cells and neutrophils stimulates production of superoxide anions by the same cells [20]. In a similar manner, in vivo foam cell-associated PAF may act not only as a chemotactic factor for monocytes but also as an activator of the synthesis and secretion of atherogenic substances (see above). Our reverse-phase HPLC analyses revealed that a major part of macrophage-associated and foam cell-associated PAF corresponded to the 16:O species. 16:O PAF has been identified as the major molecular species in many cell types, such as human neutrophils [34, 511 and in vivo activated murine macrophages [26] ; 18 : 1 and 18 : 0 PAF are also formed in these cells [26, 34, :0 PAF was demonstrated to be the most potent molecular species of PAF in several biological models (reviewed in reference [52]) as well as in platelet aggregation [531. Eventual differences in the potency of the various molecular species of PAF is potentially of importance relative to its proinflammatory role in atherogenesis. In conclusion, it has previously been shown that activated human endothelial cells, monocytes, neutrophils, platelets synthesize PAF. Our present studies demonstrate that stimulated hu-

7 54 Dentan et al. (Eur: J. Biochem. 236) man monocyte-derived macrophages and cholesterol-loaded macrophage foam cells represent a major source of PAF. Furthermore, our data indicate that PAF is transiently present in an intact form for a period of time sufficient to exert proinflammatory, proaggregatory and proatherogenic effects in arterial intima. Indeed, the presence of PAF in atheromatous plaques from canine and human coronary arteries has been documented 1161, and a close link between tumor necrosis factor-induced angiogenesis and the in situ formation of PAF has been described Furthemore, PAF contributes to the increased permeability of the endothelial monolayer and is involved in the release of tissue-type plasminogen activator from vessel walls Recent studies showed that PAF plays a pivotal role in the lymphocyte-mediated expression of tissue factor by endothelial cells [57] and thus PAF participates in thrombus formation. Morever, PAF stimulates transcription of heparin-binding epidermalgrowth factor in monocytes through an enhanced nuclear factor KB activity and therefore PAF is equally implicated in smoothmuscle cell proliferation [SS]. Taken together, the potent actions of PAF on platelet aggregation [l], superoxide [19, 201 and elastase [18] release, and the initiation of eicosanoid synthesis via arachidonate release from phospholipids [.59], are consistent with a key role of this mediator in atherogenesis. The presence of PAF in the arterial intima is probably controlled by the level of acetylhydrolase transported in lipoproteins 113, 141. In this context, it is relevant that we 1281 and others [60] have shown recently that acetylhydrolase is progressively inactivated upon oxidation of LDL, which thus loses its ability to protect against the proinflammatory actions of PAF. These findings lead us to propose that PAF, potentially in concert with additional proinflammatory mediators and factors expressed by macrophages and foam cells [18, 19, 21, 241, plays a significant, transient role in the initiation and progression of atheromatous plaques and in their thromboembolic complications. C. D. and P. L. were the recipients of a Research Fellowship from the French Ministry of Research and Technology. These studies were supported by INSERM; they were initiated at INSERM U200 (Clamart, France) and we thank Dr J. Benveniste for his kind advice and Ms J. Bidault for technical assistance. We thank Dr P. Dutartre and Laboratoires Fournier (Contrat de Valorisation INSERM, no ) for partial financial support. We are indebted to Ms C. Debets-Albertini (Centre Dipartemental de Transfusion Sanguine, Crkteil, France) for the generous gift of thrombopheresis residues, Ms M. Antonucci for technical assistance and to Ms V. Soulier for help in the preparation of the manuscript. REFERENCES 1. Benveniste, J. (1988) Paf-acether: an ether phospholipid with biological activity, frog. Clin. Bid. Res. 282, pp Bussolino, F. & Camussi, G. (1995) Platelet-activating factor produced by endothelial cells - A molecule with autocrine and paracrine properties, Eur: J. Bioclzem. 229, Ninio, E. & Joly, F. (1991) Transmembrane signalling and pafacether biosynthesis, Lipids 26, Benveniste, J., Nunez, D., Duriez, P., Korth, R., Bidault, I. & Fruchart, J. C. (1988) Preformed paf-acether and lyso paf-acether are bound to blood lipoproteins, FEBS Lett. 226, Honda, Z. I., Nakamura, M., Miki, I., Minami. M., Watanabe, T., Seyama, Y., Okado, H., Toh, H., Ito, K., Miyamoto, T. & Shirnizu, T. (1 991 j Cloning by functional expression of platelet-activating factor receptor from guinea-pig lung, Nature 349, Nakamura, M., Honda, Z. I., Izumi, T., Sakanaka, C., Mutoh, H., Minami, M., Bito, H., Seyama, Y., Matsumoto, T., Noma, M. & Shirnizu, T. (1991) Molecular cloning and expression of plateletactivating factor receptor from human leukocytes, J. Biol. Cheun. 266, Thivierge, M., Alarni, N., Muller, E., de Brum-Fernandes, A..I. & Rola-Pleszczynski, M. (1993) Transcriptional modulatior of platelet-activating factor receptor gene expression by cyclic AMP, J. Biol. Chern. 268, Snyder, F. (1995) Platelet-activating factor and its analogs: nietabolic pathways and related intracellular processes, Biochim. Biophys. Acta 1254, Elstad, M. R., Stafforini, D. M., Mclntyre, T. M., Prescott, S. hl. & Zimmerman, G. A. (1989) Platelet-activating factor acet! Ihydrolase increase during macrophage differentiation: a novel mechanism that regulates accumulation of p1atelet-activ;iting factor, J. Biol. Chem. 264, Korth, R., Bidault, J., Palmantier, K., Benveniste, J. & Ninio, E. (1993) Human platelets release a paf-acether : acetylhydrolase similar to that in plasma, Lipids 28, Tarbet, E. B., Stafforini, D. M., Elstad, M. R., Zimmerman, C;. A, McIntyre, T. M. & Prescott, S. M. (1991) Liver cells secrete the plasma form of platelet-activating factor acetylhydrolase, J. Biol. Chem. 266, Stafforini, D. M., Prescott, S. M. & McIntyre, T. M. (1987) HLiinan plasma platelet-activating factor acetylhydrolase: purification and properties, J. Biol. Chem. 262, Tselepis, T. A., Dentan, C., Karabina, S. A,, Chapman, M. J. & Ninio, E. (1995) PAF-degrading acetylhydrolase is preferenrially associated with dense LDL and VHDL-1 in human plasma: catalytic characteristics and relation to the monocyte-derived enzyme, Arterioscler: Thromb. Vase. Biol. 1.5, Tjoelker, L. W., Wilder, C., Eberhardt, C., Stafforini, D. M., Dietsch, G., Schimpf, B., Hooper, S., Letrong, H., Cousens, L. S., Zinimerman, G. A., Yamada, Y., Mcintyre, T. M., Prescott, S. M. & Gray. P. W. (199.5) Anti-inflammatory properties of a platelet-actil ating factor acetylhydrolase, Nature 374, Evangelou, A. M. (1994) Platelet-activating factor (PAF)- implications for coronary heart and vascular diseases, Prostngltriitliri.~ Leukotrienes Essent. Fatty Acids 50, Mueller, H. W., Haught, C. A,, Mcnatt, J. M., Cui, K. X.. Ga,,kell, S. J., Johnston, D. A. & Willerson, J. T. (1995) Measurement of platelet-activating factor in a canine model of coronary thromhosis and in endarterectomy samples from patients with advmced coronary artery disease, Circ. Res. 77, Feliste, R., Perret, B., Braquet, P. & Chap, H. (1989) Protcctive effect of BN 52021, a specific antagonist of platelet-actixating factor (PAF-acether) against diet-induced cholesteryl ester drposition in rabbit aorta, Atherosclerosis 78, Rouis, M., Nigon, F., Lafuma, C., Hornebeck, W. & Chapninx M. J. (1990) Expression of elastase activity by human nionocytemacrophages is modulated by cellular cholesterol content, inflammatory mediators, and phorbol myristate acetate. Arferioc/cro.Fis 10, Rouis, M., Nigon, F. & Chapman, J. (1988) Platelet-activating factor is a potent stimulant of the production of active oxygen species by human monocyte-derived macrophages, Biocl7em. Biophyx Kes. Commun. 156, Stewart, A. G., Dubbin, P. N., Harris, T. & Dusting, G. J. (1990) Platelet-activating factor may act as a second messenger in the release of icosanoids and superoxide anions from leukocytes and endothelial cells, Proc. Natl Acrid. Sci. USA 87, ,). 21. Ross, R. (1993) The pathogenesis of atherosclerosis: a perspective for the 1990s, Nature 362, Goldstein, J. L., Ho, Y. K., Basu, S. K. & Brown, M. S. I 1979) Binding site on macrophages that mediates uptake and degradation of acetylated low density lipoprotein, producing massiw cholesterol deposition, Proc. Nut/ Acad. Sci. USA 76, Hoff, H. F., Zyromski, N., Armstrong. D. & O Neil, J. (1993) Aggregation as well as chemical modification of LDL during oxidation is responsible for poor processing in macrophages, J. Lipid Rex 34, Lesnik, P., Rouis, M., Skarlatos, S., Kruth, H. S. ;G Chapman. M. J. (1992) Uptake of exogenous free cholesterol induced upregulation of tissue factor expression in human monocyte-derived niacrophages, Proc. Natl Acnd. Sci. USA 89, Falk, E., Shah, P. K. & Fuster, V. (1995) Coronary plaque disruption. Circulation 92,

8 26. Ninio, E.. Maiza, H. & Bidault, J. (1993) Autocrine amplification of PAF-acether formation in immunologically activated murine macrophages, J. Leukocyte Biol. 54, Chapman, M. J., Laplaud, P. M,, Luc, G., Forgez, P., Bruckert, E., Goulinet, S. & Lagrange, D. (1988) Further resolution of the lowdensity lipoprotein spectrum in normal human plasma: physicochemical characteristics of discrete subspecies separated by density gradient ultracentrifugation, J. Lipid Res. 29, Dentan, C., Lesnik, P., Chapman, M. J. & Ninio, E. (1994) PAFacether-degrading acetylhydrolase in plasma LDL is inactivated by copper- and cell-mediated oxidation, Arterioscler: Thromh. 14, Basu, S. K., Goldstein, J. L., Anderson, R. G. W. & Brown, M. S. (1976) Degradation of cationized low-density lipoprotein and regulation of cholesterol me.tabolism in homozygous familial hypercholesterolemia fibroblasts, Proc. Natl Acad. Sci. USA 73, Noble, R. P. (1968) Electrophoretic separation of plasma lipoproteins in agarose gel, J. Lipid Res. 9, Nakagawara, A., Nathan, C. F. & Cohn, Z. A. (1981) Hydrogen peroxide metabolism in human monocytes during differentiation in vitro, J. Clin. Invest. 68, Gamble, W., Vaughan, M., Kruth, H. S. & Avigan, J. (1978) Procedure for determination of free and total cholesterol in micro or nanogram amounts suitable for studies with cultured cells, J. Lipid Res. 19, Burton, K. (1956) A study of the conditions and mecanism of the diphenylamine reaction for the colorimetric estimation of deoxyribonucleic acid, Biochem. J. 62, Leyravaud, S., Bossant, M. J., Joly, F., Bessou, G., Benveniste, J. & Ninio, E. (1989) Biosynthesis of paf-acether. X. Phorbol myristate acetate-induced paf-acether biosynthesis and acetyltransferase activation in human neutrophils. J. Immunol. 143, Bossant, M. J., Ninio, E., Delautier, D. & Benveniste, J. (1990) Bioassay of platelet-activating factor acether by rabbit platelet aggregation, Methods Enzymol. 187, Nunez, D., Chignard, M., Korth, R., Le Couedic, J. P., Norel, X., Spinnewyn, B., Braquet, P. & Benveniste, J. (1986) Specific inhibition of paf-acether induced platelet activation by BN and comparison with the paf-acether inhibitors kadsurenone and CV 3988, Eur: J. Pharmucol. 123, Cazals-Stenzel, J. & Weber, K. (1987) Triazolodiazepines: dissociation of their PAF (platelet-activating factor) antagonistic and CNS activity, Br: J. Pharmacol. 90, Benveniste, J., Le Couedic, J. P., Polonsky, J. & TencC, M. (1977) Structural analysis of purified platelet-activating factor by lipases, Nuture 269, TencC, M., Polonsky, J., Le Couedic, J. P. & Benveniste, J. (1980) Release, purification and characterization of platelet-activating factor (PAF), Biochimie (Paris) 62, Ninio, E., Mencia-Huerta, J.-M. & Benveniste, J. (1983) Biosynthesis of platelet-activating factor (paf-acether). V. Enhancement of acetyltransferase activity in murine peritoneal cells by calcium ionophore A23187, Biochim. Biophys. Acta 751, Palmantier, R., Dulioust, A,, Mai za, H., Benveniste, J. & Ninio, E. (1989) Biosynthesis of paf-acether. Paf-acether output in murine peritoneal macrophages is regulated by the level of acetylhydrolase, Biochem. Biophys. Res. Cornmun. 162, Worthen. G. S., Seccombe, J. F., Clay, K. L., Guthrie, L. A. &Johnston, R. B. (1988) The priming of neutrophils by lipopolysaccharide for production of intracellular platelet-activating factor. Potential role in mediation of enhanced superoxide secretion, J. Imtnunol. 140, Narahara, H., Frenkel, R. A. &Johnston, J. M. (1993) Secretion of platelet-activating factor acetylhydrolase following phorbol esterstimulated differentiation of HL-60 cells, Arch. Biochem. Biophys. 301, Jouvin-Marche, E., Ninio, E., Beaurain, G., Tence, M., Niaudet, P. & Benveniste, J. (1984) Biosynthesis of paf-acether (platelet-activa- Dentan et al. (Eur: J. Biochem. 236) 55 ting factor). VII. Precursors of paf-acether and acetyl-transferase activity in human leukocytes, J. Immunol. 133, Rouis. M., Nigon. F., Eggerman, T. L., Brewer, H. B. & Chapman, M. J. (1990) Apoljpoprotein E expression by human-monocytederived macrophages. Modulation by opsonised zymosan and cholesterol, Eur: J. Biockem. 189, Roubin, R., Dulioust, A,, Haye-Legrand, I., Ninio, E. & Benveniste, J. (1986) Biosynthesis of paf-acether. VIII. Impairment of pafacether production in activated macrophages does not depend upon acetyltransferase activity, J. Immunol. 136, Colard, O., Bidault, J., Breton, M. & Ninio, E. (1993) Biosynthesis of platelet-activating factor in cultured mast cells. Involvement of the CoA-independent transacylase demonstrated by analysis of the molecular species of platelet-activating factor, Eur: J. Biochem. 216, Lee, T. C., Malone, B., Wasserman, S. I., Fitzgerald, V. & Snyder, E (1982) Activities of enzymes that metabolize platelet-activating factor (l-alkyl-2-acetyl-sn-glycero-3-phosphocholine) neutrophils and eosinophils from humans and the effect of a calcium ionophore, Biochern. Biophys. Res. Commun. 105, Ninio, E., Leyravaud, S., Bidauk, J., Jurgens, P. & Benveniste, J. (1991) Cell adhesion by membrane-bound-paf-acether, Int. Immunot. 3, Weyrich, A. S., McIntyre, T. M., McEver, R. P., Prescott, S. M. & Zimmermann, G. A. (1995) Monocyte tethering by P-selectin regulates monocyte chemotactic protein-1 and tumor necrosis factora secretion. Signal integration and NF-KB translocation, J. Clin. Invest. 95, Mueller, H. W., O Flaherty, J. T. & Wykle, R. L. (1984) The molecular species distribution of platelet-activating factor synthesized by rabbit and human neutrophils, J. Bid. Chem. 259, Mac Manus, L. M., Woodard, D. S., Deavers, S. I. & Pinckard, R. N. (1993) Biology of disease. Paf molecular heterogeneity: pathobiological implications, Lab. Invest. 69, TencC, M., Coeffier, E., Heymans, F., Polonsky, J., Codfroid, J. J. & Benveniste, J. (1981) Structural analogs of platelet-activating factor (PAF-acether), Biochimie (Paris) 63, Montrucchio, G., Lupia, E., Battagha, E., Passerini, G., Bussolino, F,, Emanuelli, G. & Camussi, G. (1994) Tumor necrosis factor a-induced angiogenesis depends on in situ platelet-activating Sactor biosynthesis, J. Exp. Med. 180, Handley, D. A., Arbeeny, C. M., Lee, M. L., Van Wen, R. G. & Saunders, R. N. (1984) Effect of platelet-activating factor on endothelial permeability to plasma macromolecules, lmmunoplzarmacology 8, Emeis, J. J. & Kluft, C. (1985) PAF-acether-induced release of tissue-type plasminogen activator from vessel walls, Blood Schmid, E., Muller, T. H., Budzinski, R. M., Pfizenmaier, K. & Binder, K. (1995) Lymphocyte adhesion to human endothelial cells induces tissue factor expression via a juxtacrine pathway, Thromb. Haemostasis 73, Pan, Z. X., Kravchenko, V. V. & Ye, R. D. (1995) Platelet-activating factor stimulates transcription of the heparin-binding epidermal growth factor-like growth factor in monocytes - Correlation with an increased kappa B binding activity, J. Bid. Chem. 270, Voelker, N. F., Worthen, S., Reeves, J. T. & Henson, H. M. (1982) Nonimmunological production of leukotrienes induced by platelet-activating factor, Science 218, Liapikos, T. A., Smaragda, A., Karabina, S. A,, Tsoukatos, D. C., Ikmopoulos, C. A. & Tselepis, A. D. (1994) Platelet-activating factor formation during oxidative modification of low-density li- poprotein when PAF-acetylhydrolase has been inactivated, Biochirn. Bio~pk,v.s. Acta 1212,