macrophage-like U937 cells

Chemistry and Physics of •

ClENCE IRELAND

Chemistry and Physics of Lipids 70 (1994) 109--119

LIPIDS

Lipoprotein-associated paf (LA-paf) was found in washed human platelets and monocyte/macrophage-like U937 cells Ruth Korth *a, Klaus Zimmermann b, Werner Oskar Richter b aForschung in der Allgemeinmedizin FIDA bei dem 1NSERM U200, Palestrinastrasse 7A, 80639 Miinchen, Germany bZweite Medizinische Klinik On Klinikum Grosshadern der Ludwig-Maximilians Universitiit Miinchen, Marchioninistrasse 15, Miinchen, Germany

(Received 27 July 1992; revision received 29 July 1993; accepted 18 September 1993)

Abstract

Beyond cholesterol, inflammatory ether phospholipids such as platelet-activating factor (paf) may play a role in atherogenesis. (1) We detected a pal-like compound ('LA-paf') associated with human serum lipoproteins, mainly in LDL but not with the lipoprotein-poor fraction. (2) LA-paf was also found in washed human platelets, from where it was partially released during platelet aggregation in response to pal (50 nM) or thrombin (1 U). In addition, resident monocyte/macrophage-like U937 cells carried huge amounts of LA-paf (41 ng per 107 cells) and metabolized added [3H]paf to a labelled compound co-eluting with the retention time of LA-paf in standard HPLC. (3) Functionally, LA-paf had a comparable potency to synthetic paf, because LA-paf aggregated washed aspirin-treated platelets in a concentration-dependent manner. The specific pal receptor antagonist WEB2086 inhibited the platelet aggregation induced by three distinct LA-paf preparations as compared with synthetic pal with similar inhibitory concentrations (ICso: 35.6 4- 12.8, 24.0 4- 4.0, 38.0 4- 15.8 nM for LA-paf, and 43.6 4- 6.5 nM for synthetic paf), indicating that LApal interacted with pal receptors. (4) However, LA-paf had a distinct retention time using high-pressure liquid chromatography (HPLC) as compared with synthetic pat'. LA-paf cluted at 9-15 min and synthetic pal at 21-24 rain. In addition, total and non-specific [3H]paf binding to intact washed human platelets was affected differently by the two unlabelled agonists: while LA-paf increased total and non-specific (but not specific) binding in a significant manner (P <0.002 and P <0.007) as LDL did (P <0.006 and P <0.03), synthetic pal decreased total binding (P <0.03). Similarly, low-density lipoproteins (LDL) increased significantly the total [3H]paf binding. In contrast, pal did not affect specific [125I]LDL binding to human fibroblasts. Our results show the presence of LA-paf in lipoproteins,

* Corresponding author. Abbreviations: ACD, acid citrate dextrose; AAGPC, alkyl-acyi-(Iong-chain)-sn-glycero-3-phosphocholine;BSA, fatty acid-free bovine

serum albumin; CP, creatinephosphate; CPK, creatinephosphokinase; FCS, fetal calf serum: HDL, high-density lipoprotein; HPLC, high-pressure liquid chromatography; ICs0, inhibitory 50% values: LA-paf, lipoprotein-associated platelet activating factor; LDL, low-density lipoprotein; lyso paf, l-O-octadecyl-sn-glycero.3-phosphocholine: PAGE, polyacryl gel electrophoresis: PtdCho, phosphatidylcholine; SDS, sodium dodecyl sulfate; synthetic pal', synthetic platelet activating factor (I-O-octadecyl-2-acetyl-sn-glycero-3 phosphocholine); TCA, trichloracetate; VLDL, very low-density lipoprotein; [3H]paf, radiolabelled synthetic paf in position 1 of the molecule ( l-O-[ 3H]-octadecyl-2-acetyl-sn-glycero-3-phosphocholine). 0009-3084/94/$07.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved. SSDI 0009-3084(93)02275-V

!10

R. Korth et aL / Chem. Phys. Lipids 70 (1994) 109-119

washed human platelets and monocyte/macrophage-like cells. As LDL and LA-paf purified from the same LDL particles increased significantly the total [3H]paf binding to intact human platelets, it might modulate platelet adherence to vascular endothelial cells. Key words: Lipoproteins; LA-paf; Paf receptors; Human platelets; Monocytes

1. Introduction

Beyond cholesterol, ether phospholipids may play an initiating role in the development of atherosclerotic lesions (for reviews see Refs, 1, 2). These ether phospholipids require for their biological activity at least an ether group in the sn-1 position of the glycerol with a short acyl chain in position 2 [3-7]. Paf-acether (paf, originally platelet-activating factor) is a potent inflammatory mediator [8], chemically 1-O-alkyl-2-acetyl-snglycero-3-phosphocholine with an acetyl group in position 2 [3,4] and saturated or non-satui'ated alkyl groups in position 1 of the molecule [9-11]. Only ether phospholipids esterified in position 2 with maximal four carbon fatty acid derivates such as 1-O-alkyl-2-butyroyl-sn-glycero- 3-phosphocholine exhibit similar biological potency as compared with paf [4]. In contrast, oxidized compounds with an ester group in position 1 of the molecule are more than a thousand times less potent as compared with paf [12]. Besides various inflammatory activities, paf plays an important role in vascular diseases (for review see Ref. 1). In addition, paf seems to be involved in the development of atherosclerotic lesions. This theory is based on the following observations. (1) Cells involved in atherosclerosis, such as endothelial cells, monocyte/macrophagelike cells and platelets, synthesize paf in response to various stimuli (for review see Ref. 13) and are activated via specific paf binding sites [14-17] (for review see Ref. 1). (2) The atherogenetic lowdensity lipoprotein (LDL) increases the number of paf receptors on the surface of monocyte/macrophage-like cells [171. (3) Phospholipid transfer proteins have been isolated from various cells (for review see Ref. 18). (4) Paf receptor antagonists, such as chemically defined ginkgolides [14,15], prevent diet-induced cholesteryl ester deposition in rabbit aortas [191.

Specific paf receptor antagonists such as the hetrazepine WEB2086 [16,20] (for review see Ref. 1) share similar pharmacological properties with chemically defined extracts of Ginkgo biloba, a plant used in traditional Chinese medicine as treatment against lung and heart diseases. The paf antagonist WEB2086 interacts specifically with paf receptors as the antagonist IC50 values correlated closely for paf binding. The paf-mediated platelet aggregation (or Ca 2+ flux) with a parallel rightward shift of the paf dose response curve reaches maximal values with inefficiency of WEB2086 against platelet aggregation in response to aspirin or ADP [16]. Paf receptors from guinea pig lungs have been demonstrated on the molecular level [21]. Differences between paf receptors on the surface of intact blood cells and adherent endothelial cells [14-17] (for review see Ref. 1) and/or their corresponding signal transduction pathway [22-24] have been demonstrated before. The specific acetylhydrolase degrades paf to lyso paf in human serum lipoproteins [25] and is released by various cells, for example activated human platelets [26]. Lyso paf is further reacylated intracellularly to the inactive 1-O-alkyl-2-acyl (long-chain)-sn-glycero-3-phosphocholine (for review see Refs. 1, 13). These ether compounds (AAGPC) are not hydrolized by the cellular acetylhydrolase, as this enzyme is highly specific and does not catabolize substrates with long acyl chains [25,27]. Monocyte/macrophage-like cells metabolize added paf to AAGPC with an intermediate of lyso paf, and their long-time incubation with LDL enriches the intracellular acetylhydrolase [171. In the light of a previously observed affinity of paf-like compounds to proteins such as serum albumin [14], it was feasible to test a putative paf distribution in plasma and cellular lipoproteins. Since paf interacts with specific receptors on plasma membranes of intact platelets, inducing plate-

111

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

let activation [14-16], it is possible that internal substances modify this paf binding (for review see Ref. 1). In this context, fatty acid-free serum albumin is a natural paf antagonist, because it inhibits the activity and binding of paf in a non-specific manner [14]. Analogous to paf, phorbolesters or LDL activate the protein kinase C by phosphorylation of a 45-kDA protein band as substrate to trigger paf binding to monocyte/macrophage-like U937 cells [17,28]. Pafor LDL at submaximal concentration activate the protein kinase C also in platelets with immediate shape change [23,29]. Finally, a receptor-mediated uptake of LDL by platelets and monocyte/macrophage-like ceils has been demonstrated [29,30]. Taken these mechanisms together, various stimuli of the protein kinase C regulate paf receptors. As an interaction of paf with LDL could have fundamental influences, we studied the presence of a paf-like activity associated with LDL ('LA-paf') and the effect of LDL on platelet paf receptors. Finally, platelets and monocyte/macrophage-like U937 cells carried huge amounts of LA-paf, and monocyte/macrophage-like cell lines metabolized added [3H]paf to compounds co-eluting with LApaf in standard-phase HPLC. 2. Methods 2.1. Lipoproteins

Low-density lipoproteins (LDL), very lowdensity lipoproteins (VLDL) and high-density lipoproteins (HDL) from three informed healthy male volunteers who had fasted for at least 8-12 h were prepared using gradient ultracentrifugation as described [31] (VLDL, d < 1.017 g/ml; LDL, d 1.023-1.061 g/ml; HDL, d 1.070-1.115 g/ml). LDL (d < 1.006-1.063 g/ml) and lipoprotein-poor serum were also prepared using sequential preparative ultracentrifugation as described [32,33]. Each fraction was dialyzed against 0.15 M NaCI/I mM EDTA, pH 7.5. LDL was labelled ([125iodinel-LDL, [125I]LDL) using the iodinemonochloride method, as described [34]. Total apolipoprotein content was measured in the different lipoprotein fractions using the method of Lowry [35]. Lipoprotein fractions were stored at 4°C and used within 1 week.

Integrity and purity as well as exclusion of peroxides of the lipoproteins were tested using gel electrophoresis and autoradiography as described [36]. The presence of lipoproteins in sonicated platelets (5 x l09) was investigated using a sodium dodecyl sulfate (SDS) polyacryl gel electrophoresis (PAGE) as described [37], comparing platelet proteins with commercial standards of purified apoprotein B in parallel with low and high molecular weight marker proteins. 2.2. Phospholipid extraction

Phospholipid extraction was performed using the method of Bligh and Dyer [38] from human lipoproteins and washed cells after sonication (six pulses of l0 s at 4 ° C, Bronson Sonic Power Company, Danbury, CT). Briefly, dichloromethane/ methanol (1/2, v/v) was added to sonicated lipoproteins or cells overnight at 4 °C before adding dichloromethane/water (l/l, v/v) containing 2% acetic acid. Organic phases were collected and water phases were washed three times with 1 volume dichloromethane. Samples were applied to a high-pressure liquid chromatography (HPLC, Microporasil column 3.9 mm ID x 300 mm length, Waters Associates, Milford, MA), which was then eluted with dichloromethane/methanol/water (5:4.:l,v/v) at a flow rate of 1 ml/min, as described [3]. LA-paf was identified by comparison with radioactive standards of phosphatidylcholine (PtdCho), sphingomyelin, synthetic paf and lysophosphatidylcholine. LA-paf eluted in the HPLC fractions at 9-15 min, pal at 21-24 min and lyso paf (not shown) at 28-33 min. Phospholipids were stored at 4° C under nitrogen after dichloromethane evaporation and suspended freshly with ethanol (0.1% v/v, final concentrations). The LA-paf activity was measured within 1 week by aggregation of rabbit platelets. Rabbit platelets (3 x l08 platelets per ml) were used to quantify LA-paf because of their high paf sensitivity [8]. When rabbit platelets were washed here as described above and before for human platelets [15] they exhibit similar paf binding kinetics as compared with human platelets (our data, not shown). Lysine aspirin (100 #M) was incubated for 30 min prior to the last wash and CP (1 mM)/CPK

112

(10 U per ml) were added directly before [8,15] aggregation measurements [39]. The aggregating activity of LA-paf was measured over the linear portion of a calibration curve of synthetic paf. The amount of LA-paf was then expressed as ng paf equivalent activity as compared with standards of synthetic paf and calculated in nM on the basis of the molecular weight of synthetic paf (551.8 g/l). The inhibitory WEB2086 effect on paf-(or LApaf)-mediated platelet aggregation was tested after 1 min preincubation at 37° C using the agonist concentration leading to 80% (ECs0) of the maximal cell response. The IC50 values were then calculated as 50% inhibitory WEB2086 concentrations from the graphs testing three distinct LA-paf preparations with three distinct platelet preparations as compared with synthetic paf.

2.3. [SHlpaf binding studies [3H]paf binding studies to washed human platelets were performed as described [11,14]. Briefly, washed human platelets were incubated with [3H]paf and unlabelled paf (0.1 to 50 nM) in the absence and presence of LDL (30 #g/ml, 30 min, 20°C, 0.25% BSA, w/v). Simultaneous platelet incubation with labelled and unlabelled compounds was reached by adding LDL (30 #g/ml, final concentrations), LA-paf (1 nM), unlabelled synthetic paf(l nM) or vehicles to the cell-free buffer containing [3H]paf (0.065 and 0.65 nM) with and without WEB2086 (1 #M). Binding experiments were started by the addition of platelet suspension (50 #1) to 450 #1 buffer. In negative control experiments (not shown), [3H]paf (0.65 nM) was added with and without unlabelled lyso paf (50 nM). Platelets were separated from supernatants after 30 min incubation at 20°C using vacuum filtration. Filter-bound radioactivity without cells was subtracted from that in the presence of platelets. The labelled [3H]paf binding was calculated as fmol per 5 x 107 platelets, and values are means ± 1 S.D. from three different experiments. Statistics were performed using the Mann-Whitney test (n = 3).

2.4. LA-paf in platelets Platelets from informed healthy male volunteers were washed as described [15]. Briefly, human

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

blood was taken from the cubital vein in acid citrate dextrose (ACD 7:l,v/v) and centrifuged (100 x g, 15 min) to obtain platelet-rich plasma (PRP). The platelets were then washed three times by centrifugation (900 x g, 10 min) in Tyrode's buffer containing ACD 9:1 (v/v, pH 6.4) and incubated with lysine-aspirin (100 #M) for 30 min prior to the last wash. Platelets were then suspended in Tyrode's buffer containing 0.25% BSA without ACD and Ca 2+ (109 platelets per ml), pH 6.4. The final platelet concentration was diluted to 1 × 109 or 5 × 109 platelets per ml with Tyrode's buffer containing 0.25% BSA and 1.3 mM Ca 2+, pH 7.4. Platelet aggregation was performed for 1 and 5 min with paf (50 nM) or thrombin (1 U) under stirring at 37°C in the presence of human fibrinogen (0.16 mg/ml). Platelets were sonicated (six pulses of 10 s at 4 ° C, Bronson Sonic Power Company, Danbury, CT) after 1 min aggregation with paf (50 nM) or thrombin (1 U) as compared with vehicle for testing LA-paf after phospholipid extraction as described above. The putative presence of lipoproteins in sonicated washed human platelets (5 × 109) was examined in parallel using SDSPAGE [37]. In these experiments non-treated human platelets were compared with stirred control platelets and platelet aggregation with paf (50 nM) or thrombin (1 U).

2.5. LA-paf in monocyte/macrophage-like cells and metabolism of labelled paf

U937

Next, LA-paf in monocyte/macrophage-like U937 cells was tested [40]. U937 cells were grown in stationary suspension culture in RPMI 1640 containing 10% FCS and 2 mM L-glutamine at 37°C in a humidified atmosphere of 5% CO2 and 95% air. The cells were diluted with fresh medium (1/10, v/v) twice a week. After 48 h in culture they were washed three times by centrifugation (1000 x g for 10 min) in Tyrode's buffer containing BSA (0.25% w/v, pH 7.4) [14]. Washed monocyte/macrophage-like U937 cells were sonicated and phospholipid extraction was performed as described above for platelets or intact cells were incubated with [3H]paf (2.8 nM per 2.5 x 106 cells/ml) for 1 h at 37° C as compared with control buffer. Phospholipid extraction was performed

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

and the organic phase was collected to perform HPLC analysis as described above.

2.6. [1251]LDL binding on human dermal fibroblasts Fibroblasts were incubated in a humidified 5% atmosphere in stock flasks (75 cm 2) containing alpha-MEM with 20% fetal calf serum (FCS). They were used at the fourth passage. For experiments, confluent monolayers of cells were detached with 0.9% NaC1 solution (w/v) containing trypsin and EDTA (0.05% and 0.02%, w/v, final concentrations). Fibroblasts (6 x l04) were seeded into 60-mm dishes for a 5-day incubation period in a 20% FCS-containing alpha-MEM (3 ml), then fresh alpha-MEM containing 10% human lipoprotein-deficient serum (2 ml) was added for a further 2-day incubation period as described [33]. Finally, [125I]LDL and increasing concentrations of unlabelled LDL, in the absence and presence of unlabelled synthetic paf (l 5 nM), were added to fibroblasts kept in alpha-MEM containing 10% human lipoprotein-poor serum for an incubation period of 3.5 h at 37°C. Adherent fibroblasts were washed in cold NaCl solutions (4°C, 0.15 M NaCI, 0.05 M Tris-HCl, pH 7.4), first with 2 mg/ml BSA, then without BSA, as described [33] with some modification, as the cell layers were detached using a HBSS buffer which contained trypsin and EDTA (0.05% and 0.01%, w/v, final concentrations) for 20 min at room temperature. They were washed and separated from supernatants by centrifugation (20 min, 1500 x g at 4°C) before supernatants were aspirated for the counting of radioactivity of the cell pellets in a Gamma counter using a standard procedure. CO 2

2. 7. Exclusion of cholesterol in LA-paf The presence of cholesterol in LA-paf was excluded by the addition of LA-paf (3.1/~g/10/~l) or vehicle (ethanol) to a cholesterol standard (200 td), which was then tested as described by Siedel et al. [41] using a Boehringer Mannheim automated analysis for a BM/Hitachi System 717 in Laborgemeinschaft Miinchner ~rzte by Dr. T. Becker (Munich).

113

2.8. Materials Tyrode's buffer was composed of (in mM) NaC1, 137; KCI, 2.68; NaHCO3, 11.9; MgCI2, 1.0; NaH2PO4, 0.41; dextrose, 0.5; HEPES 5.0. The following reagents were used: acid citrate dextrose (ACD) composed of citric acid (0.8%), trisodic citrate (2.2%) and glucose (2.45%); citric acid (0.15 M) (all from Merck-Darmstadt, Germany). Aspirin as lysine salt (Aspegic, Egic laboratory, Amilly, France). Low and high molecular weight marker proteins were from Bio-Rad and apoprotein B, fatty acid-free bovine serum albumin (BSA), CP/CPK or trypsin were from Sigma Chemicals (USA). RadiolabeUed synthetic paf in position 1 of the molecule ([3Hlpaf, 80 Ci/mmol) and labelled lyso-paf ([3Hllyso pat', 80 Ci/mmol) as well as PCS and OCS scintillation fluid were from Amersham (Amersham, UK) and were dissolved in pure ethanol. Unlabelled synthetic paf was from Bacbem (Bubendorf, Switzerland) and solubilized in ethanol. WEB2086 was from Boehringer-Ingelheim (Germany) and solubilized in water with 0.1 N HCL ultrasonically every day. Whatman GF/C filters were from Ferri~re, France, and the Millipore vacuum system from Molsbeim (Germany). RPMI-1640 culture medium, the delipidated fetal calf serum CPSR 1, alpha-MEM and FCS were from Gibeo (Eggenstein, Germany). Dulbecco's phosphate-buffered saline and HBSS buffer were from Biochem AG (Germany). U937 cells from ATCC (USA) were described before

[421. 3.Results

3.1. LA-paf in LDL, platelet aggregation with LApaf and inhibition by a specific paf receptor antagonist A paf-like activity ('LA-paf') was found in human lipoproteins. The LA-paf from LDL coeluted in the HPLC fractions (at 9-15 min) with the functionally inactive phosphatidylcholine. This retention time was clearly different as compared with synthetic paf (at 21-24 rain) (Fig. 1). The activity of LA-paf was calculated per mg lipoprotein from corresponding LDL fractions. Calculating the ability of LA-paf to stimulate

114

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119 LA -paf

12.

paf

10 8 0

6

"

2

.,J

o

,L

o

. . . . . . .

~o

s

l's

~ .

2o

----

--

25

Retention

fl

3o

time (rain)

Fig. I. HPLC elution profile of LA-paf as compared with synthetic pal standards. LDL phospholipids (O, O) obtained from two informed healthy male volunteers by sequential preparative ultracentrifugation [31] were extracted using the method of Biigh and Dyer [38] and the total biological activity was quantified using an aggregation of aspirinated, CP/CPKtreated rabbit platelets [8]. Values are then calculated as ng per mg of the corresponding lipoprotein, as shown in the Results section.

a

b

c

platelet aggregation in comparison with synthetic p a l standards, 4.6 4- 1.3 ng per mg lipoprotein was found in L D L with respect to V L D L 3.4 4- 0.6 ng and to H D L 2.5 4- 1.3 ng per mg lipoprotein was measured. This was compared with 0.2 4- 0.2 ng per mg protein in lipoproteinpoor serum (means 4- S.D., n = 3). L A - p a f extracted from L D L showed a similar concentration-dependent and time-dependent aggregatory response to synthetic p a f when testing aspirin and CP/CPK-treated rabbit platelets (Fig. 2A,B). The specific p a f receptor antagonist WEB2086 inhibited platelet aggregation response to either L A - p a f or synthetic p a f in a highly reproducible manner (Fig. 3). The ICs0 values o f the specific p a f receptor antagonist WEB2086 were 35.6 4- 12.8 nM for the first, 24.0 4- 4.0 nM for the second and 38.0 4- 5.8 n M for the third LAp a l preparation, compared with 43.6 4- 6.5 n M for synthetic pat', indicating a similar agonist potency and at least one site o f L A - p a f interacting with platelet p a l receptors. The values are means 4- 1 S.D. from three distinct aggregation assays rising three distinct L A - p a f preparations as

®

100 IO0 50

0 £:100

'i

75

.s

25

i

a

b i

c

id

i

min

o W m

~

'-

so

o

112

112

J••,,,.

1"

1'rain

Fig. 2. Dose response curve of rabbit platelet aggregation in response to different concentrations of synthetic pal (A) and LApal (B). Aggregating activity was measured over defined concentrations of synthetic pal(A: a = 25 pM, b = 100 pM, c = 130 pM) as compared with LA-paf (B: a = 32 pM, b = 68 pM, c = 83 pM, d = 109 pM, final concentrations). Values are taken from one experiment and are representative of three.

0

0.5

1 2 10g WEB-2086 (nM)

3

Fig. 3. Dose response curve of WEB2086 inhibition of rabbit platelet aggregation in response to LA-paf (O) and paf (e). Agonists were used at concentrations that induce 80% of the maximal aggregatory response (ECs0). Values are pooled from three LA-paf preparations quantified with three aggregation assays [8] and thus represent means ~- I S.D. of nine LA-paf and three paf experiments.

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

115 [3H]paf (0.065 and 0.65 riM). L D L (30/~g per ml) increased both the total as well as the non-specific [3H]paf binding, assessed either with unlabelled paf (Fig. 4) or the specific p a f antagonist WEB2086 (Table l). The changes o f both total and non-specific binding were significant, as shown in Table 1 (P < 0.006 and P <0.03), with no significant modulation o f the specific [3H]paf binding.

150

100

50 t,I

3.3. The effect of LA-paf on [3Hlpaf binding to washed human platelets I

I

I

0 0.1 0.25

!

at

•

05 1 Unlabelled paf

I.

"

50

(nM)

Fig. 4. Increase of total [3H]paf binding to washed human platelets with LDL as compared with vehicle. Intact washed human platelets were incubated with [3H]paf (0.065 nM, 30 min, 20°C) and increasing concentrations of unlabelled paf (0, 0.1-50 nM) in a BSA buffer (0.25%, w/v) with [I] or without [ O] LDL (30 #g/ml). Platelets were separated by vacuum filtration. Values are means 4- 1 S.D. of three experiments.

compared with three distinct aggregation assays using synthetic paf.

3.2. The effect of L D L on [3H]paf binding to washed human platelets Next we investigated whether the L D L preparations modulate [3H]paf binding to platelets. Indeed, a simultaneous incubation o f L D L with platelets enhanced significantly the binding o f

Analogous to corresponding L D L preparations, L A - p a f (1 riM) enhanced significantly the total and non-specific binding o f [3H]paf to intact human platelets (P < 0.002 and P < 0.007) without significant increase o f the specific [3H]paf binding (Table 1). In contrast to unlabelled LApaf, unlabelled synthetic p a f (l nM) inhibited significantly the total [3H]paf binding (P < 0.03), with a significant decrease o f the specific [3H]paf binding (P < 0.002).

3.4. LA-paf in washed human platelets Next we verified the presence o f L A - p a f in platelets. Indeed, when extracting washed platelets from healthy male volunteers 10 9 platelets contained 3.1 ± 0.5 ng LA-paf. L A - p a f was partly (not fully) released during l rain aggregation o f washed human platelets in response to exogenous paf (50 riM) or thrombin (1 U) as the L A - p a f concentration o f washed platelets decreased to

Table 1 Labelled paf binding in the presence of LDL or LA-paf [3H]paf binding (fmol)

Vehicle

LA-paf (1 nM)

LDL (30 #g/ml)

paf (1 nM)

Total Non-specific Specific

178.5 ± 27.4 108.5 ± 30.8 70.2 ± 8.5

259.8 ± 16.4" 185.7 ± 22.0* 85.3 ± 23.6

243.9 ± 23.4* 160.6 ± 13.9" 74.5 ± 11.0

138.5 ± 8.3* 106.0 ± 16.6 31 + 15.3"

Labelled paf binding to intact washed human platelets in the presence of LA-paf (1 nM), synthetic paf (I nM), LDL (30 #g/ml) or vehicle. The unlabelled compounds were added with [3H]paf (0.65 nM) to the BSA buffer (0.25%, w/v) before the binding experiment was started by the addition of platelets. Cells were separated by vacuum filtration after 30 min incubation at 20° C. The specific [3H]paf binding was assessed with the specific paf receptor antagonist WEB2086 (l #M). Values are expressed as fmol [3H]paf bound to l0 s platelets and are the means ± l S.D. from three different experiments. *P < 0.05 as compared with vehicle.

116

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

m

.J

I W

D

4

1

2

3

4

5

7

8

9

10 11

12 13

14

15

Fig. 5. SDS-PAGE from sonicated washed human platelets. Platelet proteins from healthy male volunteers were compared with commercial standards of low (lanes 1, 8) and high weight marker proteins (lanes 9, 15) and purified apoprotein B (lanes 2, 13) indicated with arrows. Resident platelets (5 x 109) (lanes 3-7) were compared with platelets after 5 min aggregation in response to paf (50 nM, lanes 11, 14) or thrombin (1 U, lane 12) or stirred control platelets (lane 10).

2.2 ± 0.2 and 2.1 ± 0.1 ng after 1 min aggregation. Platelets remained intact during aggregation, since no LDH release was detected (6.5 ± 1.5%, means ± 1 S . D . , n = 3). A detectable amount of proteins with molecular weights similar to proteins in commercial standards of low (Fig. 51,8) and high (Fig. 59,15)molecular weight marker proteins and purified

apoprotein B (Fig. 52,13 ) w a s found in sonicated platelets from healthy male volunteers (Fig. 53_7, n = 4), arguing for an uptake rather than synthesis of LA-paf by platelets. These putative lipoproteins in SDS-PAGE remained present in stirred control platelets after 5 min aggregation with 50 nM paf (Fig. 510,11,14) and decreased after 5 min aggregation with thrombin (Fig. 5~2).

Table 2 Metabolism of [3H]paf by monocyte/macrophage-like U937 cells Retention time:

Void volume

9-15 min

18-21 min paf

28-31 min Lyso paf

Percentage: Experiment 1 Experiment 2

1.8 4.5

20 32

78 61

0.2 2.9

The [3Hlpaf metabolism of monocyte/macrophage-like U 937 cells was investigated. [3Hlpaf (2,5 nM) was added for a 1-h incubation period at 37 ° C to washed U937 cells (2.5 x 106 per ml) before phospholipid analysis was performed with HPLC as described [14]. U937 cells were cultured for at least 48 h in t0% FCS medium at 37 ° C. [3H]paf remained intact in the control buffer. Values are expressed as a percentage of the total label (166 931 dpm) after subtraction of the background values.

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

3.5. Cellular LA-paf and [31-IlPaf metabolism in monocyte/macrophage-like U937 cells Next we investigated the presence of LA-paf in washed monocyte/macrophage-like U937 cells. Resident monoeyte/macrophage-like U937 cells carried huge amounts of LA-paf (41 ng LA-PAF per 107 cells). They metabolized exogenously added [3H]paf to a labelled compound which coeluted from standard-phase HPLC with LA-paf as the retention time was 9-15 min (Table 2). [3H]paf remained intact in the control buffer during the incubation period (1 h at 37°C).

3.6. Lack of paf effect on [1251]LDL binding to human fibroblasts Paf did not interfere with the receptor binding of LDL to fibroblasts, as neither the total nor the specific binding of [n25I]LDL to human fibroblasts in culture were modified by exogenous added paf (Fig. 6).

0 0

2O

40

60

80

100

117

....

120

L DL (pg/ml)

Fig. 6. The lack of a paf effect on [ n251]LDL receptor binding to human fibroblasts in culture. The receptor-dependent [1251]LDL uptake into adherent fibroblasts (6 x l04 per dish) was evidenced with increasing concentrations of unlabelled LDL with ( • ) or without (&) paf (15 nM). Dermal fibroblasts from informed healthy volunteers were cultured before and detached after binding procedure, as described in Material and Methods and in Ref. 33. Values are means 4- 1 S.E.M. using fibroblasts from three different healthy volunteers.

3. 7. Exclusion of cholesterol in LA-paf A putative cholesterol contamination of LA-paf was excluded, as the addition of LA-paf to a standard value of cholesterol did not increase the standard value as compared with vehicle (204.5 ± 2.5 vs. 204.0 ± 2.6 mg/dl, means ± 1 S.D., n = 3). 4. Discussion

Lipoproteins, platelets and monocyte/macrophage-like U937 cells carried a paf-like activity (lipoprotein-associated pal, 'LA-paf'). LA-paf was distinguished from synthetic paf because both substances eluted at different retention times using standard HPLC, suggesting differences of LA-paf as compared with synthetic pal. In addition, LApal increased significantly the total and nonspecific [3H]paf binding to intact washed human platelets, whereas unlabelled synthetic pal competed with labelled paf on the receptor, level. However, the aggregation .of aspirinated and CP/CPK-treated platelets in response to LA-paf and synthetic paf (3 x 10-II M) were indistinguishable, indicating that LA-paf contains at least an ether group in the sn-1 position of glycerol with a short acyl group in position 2. The specificity of the assays has been evidenced before with various synthetic phospholipids using platelet aggregation [4-8] or radioligand binding assays (for review see Ref. 1). In addition, the agonist potency of LA-paf and paf was similar because the highly specific pal receptor antagonist WEB2086 [16] inhibited here the aggregation of aspirinated and CP/CPKtreated platelets in response to both compounds with similar antagonist ICs0 values (43.6 ± 6 nM). The clear inhibitory dose response curves of WEB2086 using LA-paf from three distinct LDL preparations argue strongly for at least one site in LA-paf bound to platelet pal receptors. A phospholipid with biological activity characteristic of pal has been extracted from human plasma. Whereas one group proposed an acetyl group in position 2 of glycerol characteristic of the ether phospholipid pal [43], another group suggested a diacyl ester compound derived from lacyl-2-arachidonoyl-GPC [44]. One has to consider that the biological activity was tested with

118

aggregation of platelets and was defined as an equivalent of a particular exogenous concentration of synthetic paf. As platelet aggregation in response to paf correlates closely with the specific [3HlPaf binding [15,16], changes in non-specific binding cannot be detected using the platelet aggregation assay, and other putative functional changes cannot formally be excluded. A putative LDL oxidation [45] was excluded here because LA-paf was used from LDL preparations whose integrity and purity as well as absence of peroxides were tested. Solvent was evaporated and LA-paf was stored under nitrogen at 4° C, diluted freshly and used within 1 week. The low agonist and antagonist concentrations further exclude the possibility that LA-paf could be an oxidized phospholipid with an ester group in position 1 of the molecule, because a more than 1000-fold lower agonist potency of these compounds was described before [12]. A contamination of LA-paf with cholesterol that could account for changes in binding of [3Hlpaf (our unpublished data) was excluded here. It was of high clinical importance that paf did not modulate [125I]LDL binding to human fibroblasts. The total and non-specific [3H]paf binding to intact human platelets increased significantly in the presence of LDL or LA-paf (Table 1). As LDL has been shown to induce an immediate shape change of platelets with an intermediate of protein kinase C [17,23,28,29], a significant modulation of the total and non-specific [3H]paf binding cannot be simply regarded as an obscure phenomenon. This might be an effect of LDL lipids, but the LDL recognition sites on platelets could also interfere here with the shown LA-paf in LDL preparations. Platelets carried LA-paf and they partly released LA-paf. Recent reports of abundant LDL particles in an extremely dilated open cannicular system [29] and the demonstration of LA-paf in platelets argue here for LDL-mediated uptake rather than synthesis of LA-paf by platelets. Resident monocyte/macrophage-like U937 cells metabolized exogenously added [3Hlpaf to a compound eluting at 9-15 min using standard HPLC, and these resident monocyte/macrophage-like U937 cells carried huge amounts of biologically active LA-paf (co-eluting with these labelled paf

R. Korth et al./ Chem. Phys. Lipids 70 (1994) 109-119

metabolites). LA-paf was obviously not recognized by the cellular acetylhydrolase as LA-paf accumulated in resident monocyte/macrophagelike cells. These U937 cells have been shown to degrade synthetic paf with similar enzyme kinetics as compared with plasma and human platelets [17,25,26]. Thus our findings might suggest, for example, an ether phospholipid with four carbon fatty acid derivates in position 2, as these compounds are hardly recognized by the specific acetylhydrolase and they still aggregate aspirinated and CP/CPK-treated rabbit platelets [4,25]. Short-chain fatty acid derivates in position 2 are produced during oxidation of long-chain fatty acid derivates, for example of the cellular pal' metabolite AAGPC [44,45]. Next, the labelled paf metabolites from monocyte/macrophage-like cells will be purified to complete the structural characterization of cellular LA-paf [40]. In this context, R. Korth will perform further studies to compare LA-paf with synthetic phospholipids. Clinically these findings may have a physiological significance, as atherogenetic effects of LDL could be prevented by paf antagonists. LDL triggered here paf receptors on the surface of intact human platelets, and recent reports show an LDL-mediated expression of paf receptors on the surface of monocyte/macrophage-like cells [17]. Thus, hyperlipidemia might mediate increased platelet adherence to endothelial cells [46,47] with putative subendothelial recruitment of monocytes carrying LA-paf into the subendothelial tissue.

5. Acknowledgments We gratefully thank Professor Paltauf (Technische Universit~it, Graz, Austria) for helpful discussions. We also thank J. Benveniste (INSERM U 200, France) for corrections to this article and for his permission to publish results obtained by R. Korth in the INSERM U 200 laboratory in 1986 (except LDL binding assays). We further acknowledge the expert technical assistance of J. Bidault (INSERM U 200) and thank M. Middeke M.D. (Ludwigs-Maximilians Universit,it, Miinchen) for performing the statistical analysis and T. Becker M.D. (Laborgemeinschaft Miinchner .~rzte) for measuring cholesterol.

R. Korth et al./ Chem. Phys. Lipids 70 C1994) 109-119

6. References I 2 3

4 5 6 7

8 9 10 I1 12

13 14 15 16

17 18

19 20 21

22 23

C. Meade, H. Heuer and R. Kempe (1991) Biochem. Pharmacol. 41,657-668. D. Steinberg, S. Parthasarathy, T.E. Carew, J.C. Khoo and J.L. Witztum (1989) N. Engl. J. Med. 320, 915-924. J. Benveniste, M. Tenc6, P. Varenne, J. Bidault, C. Boullet and J. Polonsky (1979) C.R. Acad. Sci. (Paris) 289, 1037-1040. C.A. Demopoulos, R.N. Pinckard and D.J. Hanahan (1979) J. Biol. Chem. 254, 9355-9358. R. Korth, I. Hillmar, T. Muramatsu and N. Z611ner (1983) Chem. Phys. Lipids 33, 47-53. C. Lalau Keraly, E. Coeffier, M. Tenc6, M.C. Borrel and J. Benveniste (1983) Br. J. Haematol. 53, 513-520. W.E. Berdel, R. Korth, A. Reicbert, W.J. Houlihan, U. Bicker, H. Nomura, W.R. Vogler, J. Benveniste and J. Rastetter (1987) Anticancer Res. 7, 1181-1187. J. Benveniste, P. Henson and C,G. Cochrane (1972) J. Exp. Med. 13, 1356-1377. T. Muramatsu, N. Totani and H.K. Mangold (1981) Chem. Phys. Lipids 29, 121-127. R. Korth, H.K. Mangold, T. Muramatsu and N. Z611ner (1985) Chem. Phys. Lipids 36, 209-214. R. Korth, H. Riess, G. Brehm and E. Hiller (1986) Thrombos. Res. 41,699-706. P.L. Smiley, K.E. Stremler, S.M. Prescott, G.A. Zimmerman and T.M. Mclntyre (1991) J. Biochem. Chem. 266, 11104-11110 S.M. Prescott., T.M. Mclntyre and G.A. Zimmerman (1990) Thrombos. Haemostas. 64 (1), 99-103. R. Korth and J. Benveniste (1987) Eur. J. Pharmacol. 142, 331-341. R. Korth, D. Nunez, J. Bidault and J. Benveniste (1988) Eur. J. Pharmacoi. 152, 101-110. R. Korth, M. Hirafuji, C. Lalau-Keraly, D. Delautier, J. Bidault and J. Benveniste (1989) Br. J. Pharmacol. 98, 653-661. R. Korth and M. Middeke (1991) Chem. Phys. Lipids 59, 207-213. F. Paltauf and G. Daum (1990) in: H. Hilderson (Ed.), Subcellular Biochemistry, Vol. 16, Plenum Press, New York, pp. 279-296. R. Feliste, B. Pesset, P. Braquet and H. Chap (1989) Atherosclerosis 78, 151-158. J. Casals-Stenzel and K.H. Weber (1986) Br. J. Pharmacol. 90, 139-146. Z.I. Honda, M. Nakamura, I. Miki, M. Minami, T. Watanabe, Y. Seyama, H. Okado, H. Toh, K. Ito, T. Miyamoto and T. Shimizu (1991) Nature 349, 342-346. J.T. O'Flaherty, D.P. Jacobson, and J.F. Redman (1988) J. Biol. Chem. 264, 6836-6843. E.G. Lapetina and F.L. Siegel (1983) J. Biol. Chem. 258, 7241-7244.

119 24 25 26 27 28 29

30

31 32 33

34 35 36 37 38 39 40 41 42 43 44 45 46

47

S.B. Hwang, M.H. Lam and S.S. Pong (1986) J. Biol. Chem. 261,532-537. D.M. Stafforini, S.M. Prescott and T.M. Mclntyre (1987) J. Biol. Chem. 9, 4223-4230. R. Korth, J. Bidault, R. Palmantier, J. Benveniste and E. Ninio (1993) Lipids 28, 193-199. U.P. Steinbrecher and P.H. Pritchard (1989) J. Lipid Res. 30, 305-315. R.C. Ong, T.J. Yoo and T.M. Chiang (1991) Cell. lmmunol. 137, 283-291. L.H. Block, M. Knorr, E. Vogt, R. Locber, W. Vetter, B. Groscurth, B.Y. Qiao, D. Pometta, R. James, M. Regenass and A. Pletscher (1988) Proc. Natl. Acad. Sci. 85, 885-859. M. Esfahani, L. Secrbo, S. Lund-Katz, D.M. DePace, R. Maniglia, J.K. Alexander and M.C. Phillips (1986) Biochim. Biophys. Acta 889, 287-300. R.J. Havel, H.A. Eder and J.J. Bragdon (1955) J. Clin. Invest. 34, 1345-1353. M.J. Chapman, S. Goldstein, D. Lagrange and P.M. Laplaud (1981) J. Lipid Res. 22, 339-358. J.L. Goldstein, K.L. Bsau and M.S. Brown (Eds.) (1983) Methods in Enzymatology, Vol. 98, Academic Press, New York, pp. 241. A.S. McFarlane (1958) Nature 181, 53-60. O.H. Lowry, N.J. Rosebrough, A.L. Farr and R.J. Randall (1951) J. Biol Chem. 193, 265-275. R.P. Noble (1968) J. Lipid Res. 19, 693-700. U.K. Laemmli (1970) Nature 227, 680-684. E.G. Bligh and W.J. Dyer (1959) J. Biocbem. Physiol. 37, 911-917. G.V. Born and M.J. Cross (1963) J. Physiol. 168, 178-195. R. Korth (1990) European Patent Application, Bulletin 91/49, Publication no. 0 459 432 AI. J. Siedel, J.O. H/igele, J. Ziegenhorn and A.W. Wahlefeld (1983) Clin. Chem. 29, 1075-1080. C. Sundstrom and K. Nilsson (1976) Int. J. Cancer 17, 565-577. J. Benveniste, D. Nunez, P. Duriez, R. Korth, J. Bidault and J.C. Fruchart (1988) FEBS Lett. 226, 371-376. B. Banjeri, P.V. Subbaiah, R.E. Gregg and J.E. Bagdade (1989) J. Lipid Res. 30, 1907-1916. M.T. Quinn, S. Parthasarathy and D. Steinberg (1988) Proc. Natl. Acad. Sci. USA 85, 2805-2809. S. Moncada, R.M.J. Palmer and E.A. Higgs (1989) in: J.D. Catravas, C.N. Gillis and U.S. Ryan (Eds.), Proceedings of NATO Advanced Study lnsititute on Vascular Endothelium: Receptors and Transduction Mechanisms, Plenum Press, New York, pp. 217-223. R. Korth (1987) European Patent, Patentbtatt 92/05, Publication no. 0 312 913 BI.