Platelet activating factor-induced neuronal apoptosis is initiated independently of its G-protein coupled PAF receptor and is inhibited by the benzoate orsellinic acid

The bioactive lipid mediator platelet activating factor (PAF) is recognized as a key effecter of neuronal apoptosis, yet it is not clear whether its G-protein coupled receptor (PAFR) initiates or prevents PAF neurotoxicity. Using PAFR-/- and congenic wild-type mice, we show that PAF triggers caspase-3/7 activity and neuronal death in PAFR-/- but not PAFR+/+ cerebellar granule neurons. Restoring receptor expression by recombinant adenoviral infection protected cells from PAF challenge. Neuronal death was not mediated by nitric oxide or N-methyl-d-aspartate receptor signaling given that N-nitro-l-arginine methyl ester and MK-801 did not inhibit PAF-induced neuronal loss in PAFR-/- neurons. To intervene in PAFR-independent neurotoxicity, the anti-apoptotic actions of three structurally distinct PAF antagonists were compared to a panel of plant and fungal benzoic acid derivatives. We found that the PAF antagonist BN 52021 but not FR 49175 or CV 3988 inhibited PAFR-independent neurotoxicity. Orsellinic acid, a fungal-derived benzoic acid, blocked PAF-mediated neuronal apoptosis without affecting PAFR-mediated neuroprotection. These findings demonstrate that PAF can transduce apoptotic death in primary neurons independently of its G-protein coupled receptor, that PAFR activation is neuroprotective, and that orsellinic acid effectively attenuates PAFR-independent neuronal apoptosis.     

Identification of functional platelet-activating factor receptors in Raji lymphoblasts

The binding and metabolism of platelet-activating factor (PAF) were characterized in Raji, a human Burkitt's lymphoma-derived cell line. Raji lymphoblasts readily metabolized PAF by deacetylation-reacylation at 37 degrees C, but not at 4 degrees C. Binding studies conducted at 4 degrees C demonstrated specific binding that reached saturation within 80 min. This binding was only partially reversible. Scatchard analysis of PAF binding data revealed a single class of PAF binding sites (17,800 +/- 3,600/cell) with a K of 2.3 +/- 0.3 nM. These high-affinity PAF binding sites were shown to be functional receptors, as 100 pM to 1 microM PAF increased free intracellular calcium in a dose-dependent manner. The dose of PAF necessary to achieve half maximal calcium mobilization response was 6.3 nM, which was in the range of the K for the receptor calculated from the binding studies. The structurally dissimilar PAF receptor antagonists CV-3988 and BN52021 inhibited the PAF-induced calcium changes at doses that competed with PAF binding. These studies provide the first evidence for a functional PAF receptor expressed on a lymphocyte cell line.     

Antagonism of PAF-induced death in mice

The ability of three platelet activating factor (PAF) antagonists, BN52021, L652, 731 and 48740RP, and the leukotriene antagonist FPL55712 to block iv PAF-induced death was tested in mice. PAF-induced sudden death was been previously characterized as a model of systemic anaphylaxis and circulatory shock related its hypotensive actions. Of the drugs, BN52021 and L652, 731 provided dose-dependent protection against PAF toxicity, whereas the others had no effect. 48740RP was, however active against PAF-induced rabbit platelet aggregation. BN52021 was inactive in three other mouse sudden death models in which arachidonic acid, U46619 or collagen combined with epinephrine is injected iv to provoke a thrombotic/ischemic sudden death. In contrast, the TXA₂ antagonist SQ29548 inhibited the acute toxicity of two of these latter challenges (arachidonic acid and thromboxane agonist U46619), but was inactive against PAF lethality.These results suggest that PAF toxicity in mice is a specific model for PAF agonism, and is not mediated by TXA₂ or peptido-leukotrienes. Further, PAF-induced mortality should be a simple and useful technique for testing potential PAF antagonists for activity by various routes of administration.     

Pioglitazone induces plasma platelet activating factor-acetylhydrolase and inhibits platelet activating factor-mediated cytoskeletal reorganization in macrophage

Platelet activating factor (PAF) is a key molecule for inflammation. To examine a role of peroxisome proliferator-activated receptor gamma (PPARgamma) in inflammatory reactions of atherosclerosis, we investigated the effects of 15-deoxy-(Delta12,14)-Prostaglandin J2 (15d-PGJ2) and pioglitazone, PPARgamma ligands, on plasma PAF-acetylhydrolase (PAF-AH) expression in THP-1 macrophages. PAF-AH mRNA and protein were up-regulated by the PPARgamma ligands. Prostaglandin F2alpha (PGF2alpha), a PARgamma inhibitor, abrogated the up-regulation of PAF-AH mRNA by pioglitazone, suggesting that PPARgamma activation is involved in the induction of PAF-AH by pioglitazone. As PAF promotes the cell motility with cytoskeletal reorganization, we investigated the effect of pioglitazone on PAF-mediated morphological changes in THP-1 macrophages. In the absence of pioglitazone, PAF promoted the elongation of actin cytoskeleton, which was inhibited by pretreatment with pioglitazone. In contrast, pioglitazone was not able to inhibit the morphological changes induced by C-PAF, a non-hydrolyzable PAF agonist. Thus, it is suggested that PAF-induced morphological changes could be inhibited by pioglitazone through PAF-AH, which rapidly hydrolyzed PAF. These data propose that PPARgamma/PAF-AH pathway is a clinical target for the prevention against atherosclerosis.     

Aromatic hydrocarbon 1,2-diacetylbenzene cross-links neurofilament and other proteins

Here we explored the mechanisms of secretory phospholipase A2 (sPLA2) and glutamate (glu) in neuronal signalling and cell damage. Rats or primary neuronal cultures were treated with MK-801 and injected with/exposed to sPLA2 or glu. MK-801 partially inhibited sPLA2- and glu-induced neuronal death as well as [3H]arachidonic acid release. The involvement of cytosolic PLA2 (cPLA2) and plateletactivating factor (PAF) in sPLA2 or glu signalling was explored by treating cells with the selective cPLA2 inhibitor, AACOCF3, PAF-acetyl hydrolase (PAF-AH) or the presynaptic PAF-receptor antagonist, BN52021. AACOCF3 blocked sPLA2- and glu-induced neuronal death by 26 and 77%, respectively. PAF-AH ameliorated sPLA2 as well as glu neurotoxicity by 31 and 47%, whereas BN52021 inhibited sPLA2 induced neurotoxicity by 11% but did not significantly protect against glu-induced neurotoxicity. Expression in neurons of early response genes in response to sPLA2 or glu was further examined. An up-regulation of COX-2, c-fos, and c-jun, but not COX-1, was observed at earlier time points after rat striatal injection of glu as compared to sPLA2 injection. Moreover we treated neuronal cells with COX-2 inhibitors and found that neuronal cell death after sPLA2 and glu exposure was inhibited by 35 and 33%, respectively. Thus sPLA2 activates a neuronal signalling cascade that includes activation of cPLA2, AA-release, production of PAF and induction of COX-2. Hence sPLA2 and glu signalling are overlapping, but not identical. Cytosolic PLA2 may primarily drive glutamatergic neurotransmission, whereas PAF plays a more crucial role in sPLA2 neuronal signalling.
Acknowledgements:   Supported by EPSCoR grant NSF/LEQSF(2001-04)-RII-01 from the National Science Foundation.     

Involvement of platelet-activating factor (PAF) in cerebral post-ischemic phase in Mongolian gerbils

Platelet-activating factor (PAF) is a potent mediator of anaphylaxis and shock. In addition, evidence for PAF participation in gastric, intestinal and heart post-ischemic phase has been recently demonstrated. Ginkgo biloba extracts improve cerebral metabolism and protect brain against hypoxic damage in various models of cerebral ischemia. Potent and specific antagonists of PAF have been found in Ginkgo biloba and termed Ginkgolides: BN 52020, BN 52021, BN 52022, BN 52024. We therefore undertook the investigation of the role of Ginkgolides in cerebral ischemia obtained by bilateral ligature of the common carotid for 10 min and 6 h of recirculation in male Mongolian adult gerbils. Given preventively (one week treatment 10 mg/kg/day orally) or at the time of clamping, BN 52021 and related Ginkgolides dose-dependently antagonize morbidity assessed by the stroke-index. Similarly the mitochondrial respiration evaluated by the respiratory control ratio is significantly improved. In both determinations, the range of activity: BN 52021 greater than, BN 52020 greater than BN 52022 greater than BN 52024 shows that the effect of Ginkgolides in cerebral ischemia are correlated with their PAF antagonistic properties. Given curatively, 1 h after declamping, BN 52021 is able to reverse the cerebral impairment trend. Kadsurenone and brotizolam, two other chemically unrelated PAF antagonists led to similar recovery. Therefore PAF appears to play an important role in the post-ischemic phase after bilateral carotid ligation in Mongolian gerbils.     

Identification of platelet-activating factor receptors in P388D1 murine macrophages

Platelet-activating factor (PAF) binding and metabolism by eight murine and human cell lines was analyzed. Only the murine P388D1 macrophage line had specific, high affinity PAF binding sites. PAF binding reached saturation within 10 min at room temperature and was irreversible. Minimal PAF metabolism was observed at the time binding saturation was achieved. Scatchard analysis of PAF binding revealed a single class of PAF receptors (7872 +/- 1310/cell) which had a dissociation constant of 0.08 +/- 0.01 nM (mean +/- SEM, eta = 6). The dissociation constant was confirmed independently by quantifying the kinetics of initial specific PAF binding. PAF binding was stereospecific, required an sn-2 acetyl substituent, and was inhibited by structurally diverse PAF antagonists including kadsurenone, BN 52021, triazolam, and CV3988. The fact that the receptors are functionally active was shown by the observation that 1 to 100 pM PAF increased free intracellular calcium in P388D1 cells in a dose-related manner. These studies demonstrate that P388D1 macrophages have functional PAF receptors whose affinity and structural specificities are similar to PAF receptors in other cells. The availability of a stable cell line that binds but does not metabolize PAF will greatly facilitate studies of the PAF receptor.     

The accumulation of platelet activating factor in the injured cornea may be interrelated with the synthesis of lipoxygenase products

The rabbit cornea accumulates platelet activating factor (PAF) three hrs after alkali burn. PAF was isolated by HPLC and assayed by platelet aggregation. This bioactivity was blocked by the PAF receptor antagonists BN 52021 and alprazolam. Added PAF increases the chemiluminescence response of the cornea in vitro and BN 52021 inhibits this effect. In vivo experiments show that the synthesis of 5-HETE and 12-HETE is inhibited by the PAF antagonist BN 52021. It is concluded that a metabolic interrelationship may exist between the PAF cycle and the lipoxygenation of arachidonic acid, and that drugs that affect these lipid mediators may modulate the inflammatory response of the anterior segment of the eye.     

Effects of platelet activating factor antagonists on arachidonic acid-induced platelet aggregation and TXA2 formation

The effects of the PAF receptor antagonists WEB 2086, WEB 2170, BN 50739 and BN 52021 on AA-induced platelet aggregation (PA) and TXA2 formation were investigated in comparison with the TXA2 synthetase inhibitor HOE 944 and the TXA2 receptor antagonist BM 13.177. All PAF antagonists tested were weak inhibitors of AA-induced PA and TXA2 formation (IC50 values between 80 and 2,737 mumol/l). HOE 944 was effective in concentrations 2-3 orders of magnitude lower than PAF antagonists in inhibiting TXA2 generation. These results imply that the inhibition of TXA2 formation is of minor relevance for the actions of the investigated PAF antagonists in AA-induced PA.     

Platelet-activating factor acetylhydrolase isoforms I and II in human uterus. Comparisons with pregnant uterus and myoma

The concentrations of platelet-activating factor (PAF) that possesses the ability to stimulate myometrial contraction are partially regulated by intracellular type of platelet-activating factor acetylhydrolase (PAF-AH) in many tissues. Tissue cytosol contains at least two intracellular PAF-AH, isoforms I and II. To examine the relationship between the activity and isoforms of intracellular PAF-AH in human uterine myometrium and myoma, we assayed the PAF-AH activity and identified the PAF-AH isoforms I and II by Western blot analysis. The intense bands of the alpha2 and ss subunits of PAF-AH isoform I were detected in nonpregnant uterus; however, the specific bands of the alpha1 subunit of PAF-AH isoform I and the PAF-AH isoform II were extremely weak. The levels of the alpha2 and ss subunits and PAF-AH activity in pregnant uterus (37-39 wk gestation) were significantly lower than those in nonpregnant uterus. On the other hand, the level of ss subunit and the PAF-AH activity in myoma were significantly higher than those in nonpregnant uterus. No significant difference was found in the expression of the PAF-AH isoform II among three tissues. These results indicate that the change in the PAF-AH activity observed in pregnant uterus and myoma are due to the lower or higher protein expression of the PAF-AH isoform I, especially the alpha2 and/or ss subunits. The decrease of the uterine PAF-AH activity in the late stage of pregnancy may facilitate the action of PAF to stimulate myometrial contraction.     

Interactions between platelet activating factor and eicosanoids during endotoxic shock in anaesthetized pigs

The effects of platelet activating factor (PAF) on eicosanoid release during endotoxic shock was investigated in anaesthetized pigs receiving 5 mug kg(-1) Escherichia coli endotoxin (LPS) into the superior mesenteric artery over a 60 min period, by measuring plasma levels of a variety of mediators. Fifteen of the 31 animals infused with LPS and not treated with BN 52021, a PAF receptor antagonist, died within 30 min after the commencement of LPS infusion (non-survivors), while the other 16 survived the experimental period of 3 h, though in a state of shock (survivors). No alterations were observed in plasma concentrations of eicosanoids in the non-survivors. A significant, though transient, increase in eicosanoid concentrations occurred only in the survivors. Treatment with BN 52021 (4 mg kg-1, i.v.) injected 5 min prior to LPS infusion, failed to exert any effect on the survival rate. However, pretreatment with BN 52021 prevented circulatory collapse in the survivors and reduced the concentration of cyclooxygenase enzyme products, without affecting LTB(4) release. Exogenous administration of PAF (0.01 mug kg(-1)) caused hypotension and increased TXB(2) levels although 6-keto PGF(1alpha) and LTB(4) concentrations were unchanged. The data suggest that prostanoid formation may be secondary to PAF release in circulatory collapse evoked by LPS infusion in survivors, and give further support to the suggestion that PAF prostanoid interaction is important during endotoxic shock. However, their role in early death seems to be negligible, indicating the importance of other mediators.     

Platelet activating factor-induced apoptosis is inhibited by ectopic expression of the platelet activating factor G-protein coupled receptor

The pro-inflammatory lipid mediator platelet activating factor (PAF: 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine) accumulates in ischemia, epilepsy, and human immunodeficiency virus-1-associated dementia and is implicated in neuronal loss. The present study was undertaken to establish a role for its G-protein coupled receptor in regulating neurotoxicity. PC12 cells do not express PAF receptor mRNA as demonstrated by northern analysis and RT-PCR. In the absence of the G-protein coupled receptor, PAF (0.1-1 micro m) triggered chromatin condensation, DNA strand breaks, oligonucleosomal fragmentation, and nuclear disintegration characteristic of apoptosis. Lyso-PAF (0.001-1 micro m), the immediate metabolite of PAF, did not elicit apoptotic death. Concentrations of PAF or lyso-PAF that exceeded critical micelle concentration had physicochemical effects on plasma membrane resulting in necrosis. Apoptosis but not necrosis was inhibited by the PAF antagonist BN52021 (1-100 micro m) but not CV3988 (0.2-20 micro m). Ectopic PAF receptor expression protected PC12 transfectants from ligand-induced apoptosis. PAF receptor-mediated protection was inhibited by CV3988 (1 micro m). These data provide empirical evidence that: (i) PAF can initiate apoptosis independently of its G-protein coupled receptor; (ii) PAF signaling initiated by its G-protein coupled receptor is cytoprotective to PC12 cells; (iii) the pro- and anti-apoptotic effects of PAF on PC12 cells can be pharmacologically distinguished using two different PAF antagonists.     

Effects of platelet activating factor (PAF) and its receptor antagonist BN 52021 on isolated perfused guinea-pig heart

The cardiac effects of PAF and its antagonist BN 52021 have been investigated on the isolated perfused guinea-pig heart maintained at a constant hydrostatic perfusion pressure of 80 cm water. In this model, PAF (1 x 10(-11) to 1 x 10(-7) moles) induced a dose-dependent coronary vasoconstriction, a decrease in heart rate and a fall in contractile force. BN 52021 (1 x 10(-6) to 2 x 10(-4) M) dose-dependently inhibited the vasospasm induced by PAF (1 x 10(-10) moles). BN 52021 also antagonized the decrease in coronary flow and heart rate, but not that of contractile force induced by a high dose of PAF (1 x 10(-7) moles). This dose of PAF also significantly (p less than 0.001) provoked a marked release of TxB2 but did not alter the generation of 6 Keto PGF1 alpha, PGE2 or LTC4. The PAF-induced increase in TxB2 release was completely abolished by BN 52021.     

The effect of platelet-activating factor and its antagonist--BN 52021--on immune reactions in vivo in mice and in vitro

The effect produced by the injection of platelet activation factor (PAF) and its antagonist BN 52021 on the intensity of humoral immune response in (CBA x C57BL)F1 mice was studied. PAF was found to stimulate the formation of antibodies to sheep red blood cells. In addition PAF stimulated the phagocytic activity of mouse peritoneal macrophages. The stimulation of immune response under the action of PAF may be attributed to an increase in the phagocytic activity of macrophages. The stimulating effect of PAF on immune response in vivo was abolished by the injection of BN 52021, the antagonist of PAF. At the same time the dose-dependent decrease of immune response was observed after the injection of BN 52021. Indomethacin, an inhibitor of prostaglandin synthesis, when administered to mice treated with BN 52021, abolished the BN 52021-induced suppression of humoral immune response. Mouse peritoneal macrophages, treated in vitro with BN 52021, were found to produce significantly more prostaglandin E than control macrophages. Thus, BN 52021 induced the suppression of humoral immune response in vivo; this suppression was probably due to the action of prostaglandin E2, a messenger of the second order. Besides, the PAF antagonist BN 52021 significantly decreased leukotriene B4 production by macrophages in vitro. BN 52021 may be supposed to switch over the synthesis and/or secretion of arachidonic acid from the lipoxygenase pathway to the cycloxygenase one.     

Platelet-activating factor acetylhydrolase (PAF-AH)

Platelet-activating factor (PAF) is one of the most potent lipid messengers involved in a variety of physiological events. The acetyl group at the sn-2 position of its glycerol backbone is essential for its biological activity, and its deacetylation induces loss of activity. The deacetylation reaction is catalyzed by PAF-acetylhydrolase (PAF-AH). A series of biochemical and enzymological evaluations revealed that at least three types of PAF-AH exist in mammals, namely the intracellular types I and II and a plasma type. Type I PAF-AH is a G-protein-like complex consisting of two catalytic subunits (alpha1 and alpha2) and a regulatory beta subunit. The beta subunit is a product of the LIS1 gene, mutations of which cause type I lissencephaly. Recent studies indicate that LIS1/beta is important in cellular functions such as induction of nuclear movement and control of microtubule organization. Although substantial evidence is accumulating supporting the idea that the catalytic subunits are also involved in microtubule function, it is still unknown what role PAF plays in the process and whether PAF is an endogenous substrate of this enzyme. Type II PAF-AH is a single polypeptide and shows significant sequence homology with plasma PAF-AH. Type II PAF-AH is myristoylated at the N-terminus and like other N-myristoylated proteins is distributed in both the cytosol and membranes. Plasma PAF-AH is also a single polypeptide and exists in association with plasma lipoproteins. Type II PAF-AH as well as plasma PAF-AH may play a role as a scavenger of oxidized phospholipids which are thought to be involved in diverse pathological processes, including disorganization of membrane structure and PAF-like proinflammatory action. In this review, we will focus on the structures and possible biological functions of intracellular PAF-AHs.     

Poster Sessions CP11: Neurotoxicity and Neuroprotection. Secretory phospholipase A2 signalling partly overlaps glutamate‐mediated events

Here we explored the mechanisms of secretory phospholipase A2 (sPLA2) and glutamate (glu) in neuronal signalling and cell damage. Rats or primary neuronal cultures were treated with MK‐801 and injected with/exposed to sPLA2 or glu. MK‐801 partially inhibited sPLA2‐ and glu‐induced neuronal death as well as [3H]arachidonic acid release. The involvement of cytosolic PLA2 (cPLA2) and plateletactivating factor (PAF) in sPLA2 or glu signalling was explored by treating cells with the selective cPLA2 inhibitor, AACOCF3, PAF‐acetyl hydrolase (PAF‐AH) or the presynaptic PAF‐receptor antagonist, BN52021. AACOCF3 blocked sPLA2‐ and glu‐induced neuronal death by 26 and 77%, respectively. PAF‐AH ameliorated sPLA2 as well as glu neurotoxicity by 31 and 47%, whereas BN52021 inhibited sPLA2 induced neurotoxicity by 11% but did not significantly protect against glu‐induced neurotoxicity. Expression in neurons of early response genes in response to sPLA2 or glu was further examined. An up‐regulation of COX‐2, c‐fos, and c‐jun, but not COX‐1, was observed at earlier time points after rat striatal injection of glu as compared to sPLA2 injection. Moreover we treated neuronal cells with COX‐2 inhibitors and found that neuronal cell death after sPLA2 and glu exposure was inhibited by 35 and 33%, respectively. Thus sPLA2 activates a neuronal signalling cascade that includes activation of cPLA2, AA‐release, production of PAF and induction of COX‐2. Hence sPLA2 and glu signalling are overlapping, but not identical. Cytosolic PLA2 may primarily drive glutamatergic neurotransmission, whereas PAF plays a more crucial role in sPLA2 neuronal signalling. Acknowledgements: Supported by EPSCoR grant NSF/LEQSF(2001‐04)‐RII‐01 from the National Science Foundation.     

Failure of platelet-activating factor (PAF-acether) to induce decidualization in mice and failure of antagonists of PAF to inhibit implantation

The possible role of platelet-activating factor (PAF) in the uterine responses associated with implantation was investigated. Attempts to trigger a decidual cell response in the uteri of hormonally sensitized, ovariectomized mice by instilling PAF-acether (1-1000 ng) intraluminally were unsuccessful. The effect of PAF antagonists on implantation was investigated in females ovariectomized on Day 3 of pregnancy and treated with progesterone. Implantation was induced in these females by injection of 10 ng oestradiol-17 beta on Day 8. Hourly intraperitoneal injections of three PAF antagonists (WEB 2086, CV 3988 and BN 52021 at doses of 1.2-1.4 mg/kg) given over a 24-h period starting 1 h before the injection of oestradiol-17 beta had no significant effect on the occurrence of implantation sites. Intraluminal injection of WEB 2086 (15 micrograms) or BN 52021 (5 micrograms) either 3 h before or 6 h after the nidatory oestradiol also had no significant inhibitory effect on implantation. SRI 63-441 given once daily over the first 4 days of pregnancy at a dose of 40 micrograms/30 g body weight had no inhibitory effect on the establishment of pregnancy. These results are not consistent with a critical role for PAF in implantation in mice.     

Human endothelial cells are target for platelet-activating factor. II. Platelet-activating factor induces platelet-activating factor synthesis in human umbilical vein endothelial cells.

Platelet-activating factor (PAF), a phospholipid mediator with broad and potent biologic activities, is synthesized by several inflammatory cells including endothelial cells (EC). PAF is also an effective stimulating agent for EC leading to increased cell permeability and adhesivity. We examined the synthesis of PAF in human umbilical cord vein EC after stimulation of EC with PAF or with its nonmetabolizable analog 1-O-alkyl-2-N-methyl-carbamyl-sn-glycero-3-phosphocholine (C-PAF). PAF (1 to 100 nM) induced a dose- and time-dependent increase of PAF synthesis as detected by [3H]acetate incorporation into PAF fraction. Stimulation of PAF synthesis occurred via activation of the "remodeling pathway" as the 1-O-alkyl-2-lyso-sn-glycero-3-phosphocholine (lyso-PAF):acetyl-CoA acetyltransferase was dose-dependently increased after PAF treatment. The de novo pathway of PAF synthesis was not activated under these conditions. C-PAF was able to mimic the effect of authentic PAF on [3H] acetate incorporation. The inactive metabolite lyso-PAF (100 nM) had no influence on PAF synthesis in EC. CV-3988, BN 52021, and WEB 2086, potent and specific antagonists of PAF suppressed PAF effects on the remodeling pathway completely. The PAF- and C-PAF-induced [3H]PAF remained 93% cell-associated and was not degraded up to 10 min after stimulation. Characterization of the [3H]acetate-labeled material co-migrating with authentic PAF revealed that a significant proportion (approximately 57%) was actually 1-acyl-2-acetyl-sn-glycero-3-phosphocholine. PAF-induced PAF synthesis might be an important mechanism for amplifying original PAF signals and potentiating adhesive interactions of circulating cells with the endothelium.