Selective Release Of Arachidonic Acid Research Articles

Activation of cytosolic phospholipase A2 and fatty acid transacylase is essential but not sufficient for thrombin-induced smooth muscle cell proliferation

Thrombin is a potent stimulant of smooth muscle cell (SMC) proliferation in inflammatory conditions, leading to pathological thickening of vascular walls in atherosclerosis and airway remodeling in asthma. Cell proliferation requires the formation and remodeling of cell membrane phospholipids (PLs), involving the activation of PL-metabolizing enzymes. Yet, the role of specific PL-metabolizing enzymes in SMC proliferation has hardly been studied. To bridge this gap, in the present study, we investigated the role of key enzymes involved in PL metabolism, the PL-hydrolyzing enzyme phospholipase A2 (PLA2) and the PL-synthesizing enzyme lysophosphatidic acid-fatty acid transacylase (LPAAT), in thrombin-induced proliferation of bovine aortic SMCs (BASMCs). Concomitantly with the induction of BASMC proliferation, thrombin activated cytosolic PLA2 (cPLA2-alpha), expressed by selective release of arachidonic acid and mRNA expression, as well as LPAAT, expressed by nonselective incorporation of fatty acid and mRNA expression. Specific inhibitors of these enzymes, arachidonyl-trifluoromethyl-ketone for cPLA2 and thimerosal for LPAAT, suppressed their activities, concomitantly with suppression of BASMC proliferation, suggesting a mandatory requirement for cPLA2 and LPAAT activation in thrombin-induced SMC proliferation. Thrombin acts through the protease-activated receptor (PAR-1), and, accordingly, we found that thrombin-induced BASMC proliferation was suppressed by the PAR-1 inhibitor SCH-79797. However, the PAR-1 inhibitor did not prevent thrombin-induced mRNA expression of cPLA2 and LPAAT, implying that the activation of cPLA2 and LPAAT is essential but not sufficient for thrombin-induced proliferation of BASMCs.

Secretory phospholipase A2 receptor-mediated activation of cytosolic phospholipase A2 in murine bone marrow-derived mast cells.

The current study examined the signal transduction steps involved in the selective release of arachidonic acid (AA) induced by the addition of secretory phospholipase A2 (sPLA2) isotypes to bone marrow-derived mast cells (BMMC). Overexpression of sPLA2 receptors caused a marked increase in AA and PGD2 release after stimulation of BMMC, implicating sPLA2 receptors in this process. The hypothesis that the release of AA by sPLA2 involved activation of cytosolic PLA2 (cPLA2) was next tested. Addition of group IB PLA2 to BMMC caused a transient increase in cPLA2 activity and translocation of this activity to membrane fractions. Western analyses revealed that these changes in cPLA2 were accompanied by a time-dependent gel shift of cPLA2 induced by phosphorylation of cPLA2 at various sites. A noncatalytic ligand of the sPLA2 receptor, p-amino-phenyl-alpha-D-mannopyranoside BSA, also induced an increase in cPLA2 activity in BMMC. sPLA2 receptor ligands induced the phosphorylation of p44/p42 mitogen-activated protein kinase. Additionally, an inhibitor of p44/p42 mitogen-activated protein kinase (PD98059) significantly inhibited sPLA2-induced cPLA2 activation and AA release. sPLA2 receptor ligands also increased Ras activation while an inhibitor of tyrosine phosphorylation (herbimycin) inhibited the increase in cPLA2 activation and AA release. Addition of partially purified sPLA2 from BMMC enhanced cPLA2 activity and AA release. Similarly, overexpression of mouse groups IIA or V PLA2 in BMMC induced an increase in AA release. These data suggest that sPLA2 mediate the selective release of AA by binding to cell surface receptors and then inducing signal transduction events that lead to cPLA2 activation.

Competitive, Reversible Inhibition of Cytosolic Phospholipase A2 at the Lipid-Water Interface by Choline Derivatives That Partially Partition into the Phospholipid Bilayer

Cytosolic phospholipase A2 (cPLA2) catalyzes the selective release of arachidonic acid from the sn-2 position of phospholipids and is believed to play a key cellular role in the generation of arachidonic acid. When assaying the human recombinant cPLA2 using membranes isolated from [3H]arachidonate-labeled U937 cells as substrate, 2-(2'-benzyl-4-chlorophenoxy)ethyl-dimethyl-n-octadecyl-ammonium chloride (compound 1) was found to inhibit the enzyme in a dose-dependent manner (IC50 = 5 microM). It was over 70 times more selective for the cPLA2 as compared with the human nonpancreatic secreted phospholipase A2, and it did not inhibit other phospholipases. Additionally, it inhibited arachidonate production in N-formyl-methionyl-leucyl-phenylalanine-stimulated U937 cells. To further characterize the mechanism of inhibition, an assay in which the enzyme is bound to vesicles of 1,2-dimyristoyl-sn -glycero-3-phosphomethanol containing 6-10 mol % of 1-palmitoyl-2-[1-14C]arachidonoyl-sn-glycero-3-phosphocholine was employed. With this substrate system, the dose-dependent inhibition could be defined by kinetic equations describing competitive inhibition at the lipid-water interface. The apparent equilibrium dissociation constant for the inhibitor bound to the enzyme at the interface (KI*app) was determined to be 0.097 +/- 0.032 mol % versus an apparent dissociation constant for the arachidonate-containing phospholipid of 0.3 +/- 0.1 mol %. Thus, compound 1 represents a novel structural class of inhibitor of cPLA2 that partitions into the phospholipid bilayer and competes with the phospholipid substrate for the active site. Shorter n-alkyl-chained (C-4, C-6, C-8) derivatives of compound 1 were shown to have even smaller KI*app values. However, these short-chained analogs were less potent in terms of bulk inhibitor concentration needed for inhibition when using the [3H]arachidonate-labeled U937 membranes as substrate. This discrepancy was reconciled by showing that these shorter-chained analogs did not partition into the [3H]arachidonate-labeled U937 membranes as effectively as compound 1. The implications for in vivo efficacy that result from these findings are discussed.

Cytosolic phospholipase A2 (cPLA2) distribution in murine brain and functional studies indicate that cPLA2 does not participate in muscarinic receptor-mediated signaling in neurons

Cytosolic phospholipase A2 (cPLA2) catalyzes the selective release of arachidonic acid from the sn-2 position of membrane phospholipids and has been suggested as an effector in the receptor-mediated release of arachidonic acid in signal transduction. The potential role of cPLA2 as an effector in muscarinic acetylcholine receptor signaling was investigated through ectopic expression of either the m1 or m5 receptor in combination with cPLA2 in COS-1, CHO and U-373 MG cell lines. U-373 MG and COS-1 cells express undetectable or very low levels of cPLA2. CHO cell extracts are characterized by a significant endogenous PLA2 activity that was increased over 20-fold following transient expression with cPLA2 cDNA. However, in none of the cells lines did the co-expression of muscarinic receptor and cPLA2 result in a significant increase in muscarinic receptor-mediated arachidonic acid release over cells expressing muscarinic receptor alone. The distribution of cPLA2 mRNA and cPLA2 immunoreactivity in murine brain were determined in order to investigate a potential role for cPLA2 in neurotransmission. cPLA2 mRNA was expressed in white matter, including cells contained within linear arrays characteristic of interfascicular oligodendrocytes. cPLA2 immunoreactivity in white matter was evident throughout the processes of fibrous astrocytes. cPLA2 expression in gray matter was confined to astrocytes at the pial surface of the brain. cPLA2 mRNA was detected in pia mater, both at the brain surface and inner core of the choroid plexus. cPLA2 may not be directly linked to neurotransmission since enzyme expression, mRNA, and cPLA2 immunoreactivity were undetectable in neurons of murine brain. Support or regulation of neurotransmission may be provided through the activity of cPLA2 in glial cells.

Mechanisms that account for the selective release of arachidonic acid from intact cells by secretory phospholipase A 2

The current study examined mechanisms that account for the selective release of arachidonic acid (AA) from cells by secretory phospholipase A 2 (sPLA 2). Initial studies demonstrated that low concentrations of group I and group III PLA 2 isotypes and an sPLA 2-enriched extract from bone marrow-derived mast cells (BMMC) selectively released AA from mast cells. Much higher concentrations of group II PLA 2 were required to release comparable quantities of AA. Group I PLA 2 also selectively released AA from another mast cell line (CFTL-15) and a monocytic cell line (THP-1). In contrast, high concentrations of group I PLA 2 were required to release fatty acids from a promyelocytic cell line (HL-60) and this release was not selective for AA. Binding studies revealed that cell types (BMMC, CFTL-15 and THP-1) which selectively released AA also had the capacity to specifically bind group I PLA 2. However, group II PLA 2, which did not selectively release AA from cells, also did not specifically bind to these same cell types. Additional studies revealed that sPLA 2 binding to the mast cell receptor was attenuated after stimulation with antigen or ionophore A23187. Reverse transcriptase–polymerase chain reaction analyses indicated the presence of mRNA for the sPLA 2 receptor in BMMC, CFTL-15 and THP-1 and the absence of this mRNA in HL-60. Final studies demonstrated that p-aminophenyl-α- d-mannopyranoside BSA, a known ligand of the sPLA 2 receptor, also selectively released AA from mast cells but not from HL-60 cells. These experiments indicated that receptor occupancy alone (without PLA 2 activity) is sufficient to induce the release of AA from mast cells. Together, these data reveal that specific isotypes of sPLA 2 have the capacity to selectively release AA from certain cells by their capacity to bind to sPLA 2 receptors on the cell surface.

A beta-lactam inhibitor of cytosolic phospholipase A2 which acts in a competitive, reversible manner at the lipid/water interface.

Cytosolic phospholipase A2 (cPLA2) catalyzes the selective release of arachidonic acid from the sn-2 position of phospholipids and is believed to play a key cellular role in the generation of arachidonic acid. When assaying the human recombinant cPLA2 using membranes isolated from [3H]arachidonate-labeled U937 cells as substrate, 3,3-Dimethyl-6-(3-lauroylureido)-7-oxo-4-thia-1-azabicyclo[3,2,0] heptane-2-carboxylic acid (1) was found to inhibit the enzyme in a dose-dependent manner (IC50 = 72 microM). This beta-lactam did not inhibit other phospholipases, including the human nonpancreatic secreted phospholipase A2. The inhibition of cPLA2 was found not to be time-dependent. This, along with the observation that the degradation of the inhibitor was not catalyzed by the enzyme, demonstrates that the inhibition does not result from the formation of an acyl-enzyme intermediate with the active site serine residue. Moreover, the ring-opened form of 1 is also able to inhibit cPLA2 with near-equal potency. To further characterize the mechanism of inhibition, an assay in which the enzyme is bound to vesicles of 1,2-dimyristoyl-sn-glycero-3-phosphomethanol containing 6-10 mole percent of 1-palmitoyl-2-[1-14C]-arachidonoyl-sn-glycero-3-phosphocholine was employed. With this substrate system, the dose-dependent inhibition was defined by kinetic equations describing competitive inhibition at the lipid/water interface. The apparent dissociation constant for the inhibitor bound to the enzyme at the interface (KI*app) was determined to be 0.5 +/- 0.1 mole% versus an apparent dissociation constant for the arachidonate-containing phospholipid of 0.4 +/- 0.1 mole%. Thus, 1 represents a novel structural class of inhibitors of cPLA2 which partitions into the phospholipid bilayer and competes with the phospholipid substrate for the active site.

Presence of Glycerol Masks the Effects of Phosphorylation on the Catalytic Efficiency of Cytosolic Phospholipase A2

Cytosolic phospholipase A2catalyzes the selective release of arachidonic acid from the sn-2 position of phospholipids and is believed to play a key cellular role in the generation of arachidonic acid. The enzymatic activity of cPLA2is affected by several mechanisms, including substrate presentation and the phosphorylation state of the enzyme. Using covesicles of 1-palmitoyl-2-arachidonoyl-[arachidonoyl-1-14C]-sn-glycero-3-phosphocholine and 1,2-dimyristoyl-phosphatidylmethanol as substrate, the effects of phosphorylation on the interfacial binding and catalytic constants were investigated. Phosphorylated and dephosphorylated enzyme forms were shown to have identical values of 2.6 μmforKappM, an equilibrium dissociation constant which consists of the intrinsic dissociation constant from the lipid/water interface (KS) and the dissociation constant for phospholipid from the active site (KM*). Moreover, the values ofKM*for phosphorylated and dephosphorylated enzyme did not differ significantly (0.4 ± 0.1 and 0.2 ± 0.1, respectively). However, dephosphorylation of the enzyme reduced the value ofkcatby 39%. The phosphorylation state of the enzyme had no effect on either the cooperativity shown by this enzyme or the thermal stability of the enzyme. Surprisingly, the presence of glycerol (4m) masks the effect of phosphorylation onkcat. Instead, glycerol increased the value ofkcatby 440% for the phosphorylated enzyme and by 760% for the dephosphorylated form. Moreover, addition of glycerol had only small effects onKappM. The increase in thekcatupon addition of glycerol results from a substantial decrease in the activation energy from 29.4 to 14.8 kcal· mol−1. To determine whether the effects of phosphorylation of the enzyme or addition of glycerol are unique to this artificial substrate, membranes from U937 cells were isolated and used as substrate. With these membranes, the dephosphorylated enzyme was only 21% less active than the phosphorylated enzyme. In the presence of glycerol, there was no detectable difference the two enzyme forms, and the rate of hydrolysis was increased by 300–390% over that measured in the absence of glycerol. These results suggest that the catalytic efficiency of the phosphorylated enzyme is not particularly relevant to its activationin vivo.Moreover, it may be that glycerol is mimicking the effect of some unidentified factor which greatly enhances the catalytic efficiency of the enzyme.

Phospholipase D is involved in cytosolic phospholipase A 2-dependent selective release of arachidonic acid by fMLP-stimulated rat neutrophils

When rat polymorphonuclear neutrophils (PMN) were treated with N-formyl-Met-Leu-Phe (fMLP), the release of arachidonic acid in preference to other fatty acids was observed. Levels of arachidonic acid reached a plateau within 5 min, and were accompanied by an ∼ 4-fold increase in in vitro phospholipase (PL) A 2 and PLD activities in PMN lysates. Treatment of PMN with ethanol (an inhibitor of PLD-mediated phosphatidic acid formation), propranolol (a phosphatidic acid phosphatase inhibitor), or 4-bromophenacylbromide (a PLA 2 inhibitor), each suppressed fMLP-stimulated arachidonate release. Treatment with RHC-80267 (a diacylglycerol lipase inhibitor), however, had no such effect. The cytosolic PLA 2 (cPLA 2) inhibitor, arachidonoyl trifluoromethyl ketone, suppressed PLA 2 activity in PMN homogenates and arachidonate release by fMLP-treated PMN. These results suggest that fMLP-elicited arachidonate release is mediated by cPLA 2 but not diacylglycerol lipase, and that the activation of cPLA 2 is downstream of the PLD-dependent signaling pathway.

Survey shows unprotected sex is a common behaviour in bisexual men

Cytosolic phospholipase A₂ (cPLA₂) catalyzes the selective release of arachidonic acid from the <i>sn</i>-2 position of phospholipids and is believed to play a key cellular role in the generation of arachidonic acid. BMS-229724 (4-[4-[2-[2-[bis(4-chlorophenyl)methoxy]ethyl-sulfonyl]ethoxy]phenyl]-1,1,1-trifluoro-2-butanone) was found to be a selective inhibitor of cPLA₂(IC₅₀ = 2.8 μM) in that it did not inhibit secreted phospholipase A₂ in vitro<i>,</i> nor phospholipase C and phospholipase D in cells. The compound was active in inhibiting arachidonate and eicosanoid production in U937 cells, neutrophils, platelets, monocytes, and mast cells. With a synthetic covesicle substrate system, the dose-dependent inhibition could be defined by kinetic equations describing competitive inhibition at the lipid/water interface. The apparent equilibrium dissociation constant for the inhibitor bound to the enzyme at the interface (<i>K</i>_I*^app) was determined to be 1 · 10⁻⁵ mol% versus an apparent dissociation constant for the arachidonate-containing phospholipid of 0.35 mol%. The unit of concentration in the interface is mole fraction (or mol%), which is related to the surface concentration of substrate, rather than bulk concentration that has units of molarity. Thus, BMS-229724 represents a novel inhibitor of cPLA₂, which partitions into the phospholipid bilayer and competes with phospholipid substrate for the active site. This potent inhibition of the enzyme translated into anti-inflammatory activity when applied topically (5%, w/v) to a phorbol ester-induced chronic inflammation model in mouse ears, inhibiting edema and neutrophil infiltration, as well as prostaglandin and leukotriene levels in the skin. In hairless guinea pigs, BMS-229724 was active orally (10 mg/kg) in a UVB-induced skin erythema model in hairless guinea pigs.

Zymosan induces selective release of arachidonic acid from rabbit alveolar macrophages via stimulation of a β-glucan receptor

Zymosan, which is composed primarily of α-mannan and β-glucan polymers, is a well recognized activator of macrophages. The type receptor by which unopsonized zymosan induces arachidonic acid release was investigated. It was found that particulate β-glucan and zymosan stimulated an identical dose-dependent release of arachidonic acid. This release of arachidonic acid by zymosan was blocked by soluble β-glucans whereas soluble mannan had no effect. This inhibition was not due to a general toxic effect of the soluble β-glucans as they had no effect on calcium ionophore-induced release of arachidonic acid. β-glucan-induced fatty acid release from these cells was shown to be fairly specific for arachidonic acid. These data reveal that zymosan stimulates the specific release of arachidonic acid from rabbit alveolar macrophages, at least in part, via a β-glucan receptor.

Cytosolic phospholipase A2 is coupled to hormonally regulated release of arachidonic acid.

Cytosolic phospholipase A2 (cPLA2) binds to natural membrane vesicles in a Ca(2+)-dependent fashion, resulting in the selective release of arachidonic acid, thus implicating cPLA2 in the hormonally regulated production of eicosanoids. Here we report that the treatment of Chinese hamster ovary (CHO) cells overexpressing cPLA2 with ATP or thrombin resulted in an increased release of arachidonic acid as compared with parental CHO cells, demonstrating the hormonal coupling of cPLA2. In contrast, CHO cells overexpressing a secreted form of mammalian PLA2 (sPLA2-II) failed to show any increased hormonal responsiveness. Interestingly, we have noted that the activation of cPLA2 with a wide variety of agents stimulates the phosphorylation of cPLA2 on serine residues. Pretreatment of cells with staurosporin blocked the ATP-mediated phosphorylation of cPLA2 and strongly inhibited the activation of the enzyme. Increased cPLA2 activity was also observed in lysates prepared from ATP-treated cells and was sensitive to phosphatase treatment. These results suggest that in addition to Ca2+, the phosphorylation of cPLA2 plays an important role in the agonist-induced activation of cPLA2.

Turnover of eicosanoid precursor fatty acids among phospholipid classes and subclasses of cultured human umbilical vein endothelial cells

Using cultured human umbilical vein endothelial cells, in which phosphatidylcholine (PC) is equally pulse-labelled by various eicosanoid precursor fatty acids (EPFAs), we have studied the remodelling of EPFAs among the phospholipid classes and subclasses with and without activation, and the relationship of this remodelling process to the selective release of arachidonic acid (AA) by phospholipase A2-mediated cell stimulation. When endothelial cells are pulse-incubated with radiolabelled EPFA for 15 min, greater than 80% of cell-associated radioactivity is present in phospholipids, among which greater than 60% is found in 1,2-diacyl-sn-glycero-3-phosphocholine (diacyl PC). After removing unincorporated radioactivity, reincubation of the pulse-labelled cells for up to 6 h results in progressive decrease in EPFA-labelled diacyl PC, increase in AA- or eicosapentaenoic acid (EPA)-labelled 1-O-alk-1-enyl-2-acyl-sn-glycero-3-phosphoethanolamine (plasmalogen PE) and increase only in AA-labelled 1-O-alkyl-2-acyl-sn-glycero-3-phosphocholine (alkyl PC). This redistribution of radiolabelled phospholipids is not altered by the presence of excess non-radiolabelled EPFAs. When aspirin-treated EPFA-labelled endothelial cells are stimulated with ionophore A23187, a very selective release of AA is noted in comparison with eicosatrienoate (ETA) or EPA, accompanied by an equivalent decrease in AA-labelled diacyl PC and specific increase in AA-labelled plasmalogen PE and alkyl PC. These selective changes in AA radioactivity induced by A23187 are enhanced 2-fold by pretreating the AA-labelled cells with phorbol 12-myristate 13-acetate, which by itself induces no changes. The changes in radioactivity induced by A23187 without and with phorbol ester among the released AA, the diacyl PC and the plasmalogen PE are significantly correlated with each other. These results indicate that human endothelial cells incorporate EPFAs (AA, ETA, EPA) equally into diacyl PC but selectively release AA esterified into diacyl PC with specific remodelling into plasmalogen PE and alkyl PC.

Biochemical and subcellular distribution of arachidonic acid in rat myocardium.

Selective release of arachidonic acid from prelabeled phospholipid pools has been observed following exposure of neonatal rat cardiac myocytes to metabolic inhibitors in vitro and has been correlated temporally with the development of irreversible sarcolemmal damage. Hydrolysis of phospholipids with release of arachidonic acid may be an important mechanism in the pathogenesis of sarcolemmal damage induced by ischemia. To elucidate potential subcellular loci of arachidonic acid release in ischemic myocardium, we characterized the phospholipid composition of adult rat myocardial sarcolemma and delineated the biochemical and subcellular distribution of radiolabeled arachidonic acid in neonatal rat myocytes incubated with [3H]-arachidonic acid for selected intervals. Although arachidonic acid comprised 13 +/- 3% of total aliphatic moieties in combined choline and ethanolamine glycerophospholipids isolated from purified adult rat myocardial sarcolemma, radiolabeled arachidonic acid could not be detected in sarcolemma of cultured neonatal rat myocytes incubated with 5 nM [3H]arachidonic acid for 24 h. Radioactivity was located almost exclusively in mitochondria and internal cytoplasmic membranes (primarily sarcoplasmic reticulum), which collectively contained 90% of myocyte radioactivity. These results indicate that radiolabeled arachidonic acid released from prelabeled phospholipid pools on exposure of neonatal rat myocytes to oxidative inhibitors is derived from mitochondria and internal cell membranes. The diminutive labeling of the sarcolemma suggests that turnover of arachidonoyl phospholipids is slower in the sarcolemma than in other membranous organelles.

Ionophore-induced metabolism of phospholipids and eicosanoid production in porcine aortic endothelial cells: selective release of arachidonic acid from diacyl and ether phospholipids

Confluent cultures of porcine aortic endothelial cells were prelabeled with 1 μM [ 14C]arachidonic acid complexed to 1 μM bovine serum albumin. After washing, the cells were stimulated with 1 μM A23187 for time intervals between 30 s and 30 min. Cellular lipids were extracted and separated into major lipid classes and phospholipid subclasses. The external medium was analyzed for released radioactive eicosanoids. The time-course of total release of 14C radioactivity demonstrated a biphasic nature of A23187-induced changes in endothelial cell lipids. Early, from 30 s to 5 min, substantial losses of [ 14C]arachidonic acid from diacylphosphatidylethanolamine and phosphatidylinositol, as well as an abrupt increase in diacylphosphati-dylcholine-associated radioactivity were observed. These initial changes coincided with the release of 14C-labeled cyclooxygenase products. Later changes (5–30 min) included a sustained progressive loss of 14C radioactivity from alkenyl (alk-1-enyl) acylphosphatidylethanolamine and diacylphosphatidylcholine. These later changes coincided with the elaboration of 14C-labeled lipoxygenase products. Although unequivocal assignments cannot be made, the data suggest that specific pools of arachidonic acid provide precursors for individual classes of eicosanoids.

Prostaglandin biosynthesis and lipolysis in subcellular fractions from rabbit kidney medulla.

Three separate prostaglandin-generating activities are associated with plasma membranes, mitochondria and microsomal fractions from rabbit kidney medulla. In the plasma membranes and mitochondria, but not in microsomal fractions, Ca2+ ions stimulate the activity of phospholipase A2, yielding selective release of arachidonic acid and linoleic acid and concomitant increase in prostaglandin E2 formation.

Arachidonic acid in amniotic fluid after extra-amniotic instillation of rivanol for midtrimester abortion

Free arachidonic acid was measured by gas liquid chromatography in amniotic fluid in 13 women during induction of midtrimester abortion. The abortions were induced by extraamniotic instillation of 0.1% Rivanol. Serial sampling of amniotic fluid were performed through a trans-abdominal catheter up to 22 hrs. Free, total arachidonic acid showed a significant increase from 26±8 ng/ml to 293±102 ng/ml at 22 hrs. The percentage of free arachidonic acid in total free fatty acids increased significantly from 2.2±0.5 to 6.1±1.6% during the same time. The C results suggest a selective release of arachidonic acid during Rivanol-induced abortion.