Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Ionophore toxins belong to a large group of proteins of bacterial, plant and animal origin that alter plasma membrane permeability by forming small lesions (Harvey, 1990). The importance of this group of toxins is evident from the finding that of the several hundred toxins studied to date, a large group are ionophores. Although the pore/channel activity of these toxins was intensively investigated (Bernheimer and Rudy, 1986), their mode of interference with the cellular signal transduction pathways is largely unknown (Bloch-Shilderman et al., 1996). According to pore size and mechanism of cell membrane injury, ionophore toxins have been classified into three major groups (Thelestam and Mollby, 1979; Bloch-Shilderman et al., 1997): 1) thiol-activated toxins that form large pores, 2) toxins with surfactant activity on the plasma membrane and 3) toxins that form voltage-dependent ionic channels. Recent studies suggest that thiol-activated toxins such as streptolysin S (Abu-Raya et al., 1993a), surfactant toxins including melittin and mastoparan (Choi et al., 1992) and other toxins such as maitotoxin (Choi et al., 1990) and canatoxin (Barja-Fidalgo et al., 1991) activate the AA cascade in rat pheochromocytoma PC12 cells, rat brain synaptosomes and rabbit platelets.

Over the past ten years we have isolated and characterized a new ionophore toxin, Pardaxin, derived from the fish Pardachirus marmoratus (Lazarovici et al., 1986, 1988; Shai et al., 1988). Pardaxin is an acidic, amphipathic and hydrophobic polypeptide composed of 33 amino acids (Lazarovici et al., 1986) that forms voltage-dependent pores in liposomes (Loew et al., 1985) and planar lipid bilayers (Lazarovici et al., 1992; Shi et al., 1995). The toxin exhibits a neurotoxic, excitatory activity toward neurons (Lazarovici, 1994), induces contraction of guinea pig ileum (Primor, 1986), hemolysis and reduction of blood pressure and is lethal to rats (Primor and Lazarovici, 1981). Pardaxin was proposed as a pharmacological tool for studying neurotransmitter release (Lazarovici and Lelkes, 1992) because it stimulates exocytosis in a variety of neuronal preparations by both calcium-dependent and calcium-independent mechanisms (Lazarovici and Lelkes, 1992; Arribas et al., 1993). The major aim of the present study was to examine the ability of Pardaxin to stimulate the AA cascade in PC12 cells, in the presence or absence of [Ca]_o.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. The following drugs and chemicals were used in this study: [³H]AA (210 Ci/mmole), [³H]PGE₂, [³H]TXB₂, [³H]6-keto-PGF₁ (120-200 Ci/mmole) and [³H]myoinositol (10-25 Ci/mmole) (New England Nuclear, Boston, MA), [³H]Dex (35-60 Ci/mmole) and ortho[³²P]phosphate (10 mCi/ml) (Amersham, Buckinghamshire, UK), antibodies directed against PGE₂, TXB₂ and 6-keto-PGF₁ (BioYeda, Rehovot, Israel), fura 2-AM (Molecular Probes Inc., Junction City, OR), indomethacin, EGTA, dextran, Dex, carbachol, charcoal, A23187, NDGA, PGE₂, TXB₂, 6-keto-PGF₁, esculetin, pBPB-4 and mepacrine (Sigma Chemical Co., St. Louis, MO), RU-28318 and RU-38486 (UCLAF-Roussel, Romanville, France) and RHC-80267, BEL and MAFP (Biomol, Plymouth Meeting, PA).

Toxins. Pardaxin, a natural toxin secreted from the glands of the fish Pardachirus marmoratus, was isolated by the liquid chromatography method as described (Lazarovici et al., 1986). The native toxin used in the present study was purified by HPLC and found to be homogeneous by SDS-gel electrophoresis and by amino acid analysis (Lazarovici et al., 1986). The primary structure of Pardaxin: NH₂-GFFALIPKIISSPLFKTLLSAVGSALSSSGGQE-COOH was determined by sequencing. Pardaxins were synthesized by the solid-phase method on (phenylacetamido) methyl-amino acid resin, as previously described (Shai and Oren, 1996). Coupling was carried out using freshly prepared hydroxybenzotriazole active esters of Boc-amino acids. The resin-bound peptides were then transaminated with 30% ethylenediamine in dimethylformamide for 3 days, followed by filtration of the resin, precipitation of the protected peptides with ether and removal of the protecting groups with hydrofluoric acid. The synthetic peptides, like the native toxin, were purified to >95% homogeneity by reverse-phase HPLC, on a C₁₈ column, using a linear gradient of 25% to 80% acetonitrile in 0.1% trifluoro-acetic acid for 40 min and then subjected to amino acid analysis to confirm their composition (Shai and Oren, 1996).

Cell viability. In the present study, subcytotoxic concentrations of Pardaxin were used. Pardaxin concentration was considered subcytotoxic when less than 10% cell death was obtained, as determined by trypan blue exclusion (Abu-Raya et al., 1993b).

PC12 cultures. PC12 cells were grown in DMEM supplemented with 7% fetal calf serum, 7% horse serum, 100 µg/ml streptomycin and 100 units/ml penicillin. The cell cultures were maintained in an incubator at 37°C in an atmosphere of 6% CO₂. Medium was changed twice weekly, and the cultures were split at a 1:6 ratio once a week. Experiments were performed with PC12 cells grown on 6-well or small Petri dishes coated with equal parts of collagen (0.01 mg/ml collagen in 0.1 M acetic acid) and poly L-lysine (0.01 mg/ml) (Abu-Raya et al., 1993a).

Measurement of PLA₂ activity. 1. AA release: PC12 cells were grown in 6-well dishes in serum-containing DMEM for 24 hr at 37°C. The growth medium was then removed and replaced with serum-free DMEM to which [³H]AA (0.5 µCi/ml) was added, and the plates were incubated for 4 hr. The medium (containing the nonincorporated isotope) was removed, and the cells were washed three times with PBS (138 mM NaCl, 8 mM Na₂HPO₄, 0.5 mM MgCl₂ and 0.9 mM CaCl₂, pH 7.4). The rinsed cells were then incubated with 1 ml of PBS supplemented with 20 mM glucose and 1 mg/ml fatty acid free-BSA for 10 min at 37°C (Fink and Guroff, 1990). Release assays were initiated by adding the tested compound in the presence or absence of different inhibitors, and the cultures were further incubated at 37°C for various intervals. At the indicated times, 200 µl of incubation medium was collected from each well and centrifugated for 10 min (1000 × g); then the supernatant was collected and the volume measured. Release of [³H]AA was measured in 100-µl aliquots in a liquid scintillation counter. The amount of protein in each well was estimated according to Lowry (Lowry et al., 1951). The amount of AA released was expressed as counts per minute per milligram of protein, or as percentage of the control. 2. Changes in phospholipids content: PC12 cells were grown to confluency on small Petri dishes (2 × 10⁶ cells/dish) in serum-containing DMEM for 24 hr at 37°C. The growth medium was removed and replaced with 1% delipidated serum containing [³²P]orthophosphate (1 µCi/ml), and the cultures were incubated for 18 hr at 37°C. Thereafter, the medium containing nonincorporated isotope was removed, and the cultures were washed three times with PBS. The rinsed cells were incubated in PBS with 5 µM Pardaxin for 15 min at 37°C. After extensive washing, the lipid were extracted with cold n-propanol for 4 hr, evaporated and separated by two-directional thin-layer chromatography on precoated silica-gel plates (Merck, Darmstadt). ³²P-labeled phospholipids were exposed to autoradiography (Kodak, XRP-54 film), followed by counting of the radioactive spots as previously described (Yavin and Zutra, 1977). Because the level of PE was not changed after treatment with Pardaxin, the levels of the major phospholipids are presented relative to the PE level.

Eicosanoid RIA. The cultures were exposed at 37°C for various intervals to different concentrations of Pardaxin. At the end of the experiment, the medium was removed and centrifuged at 4°C for 10 min at 1000 × g, and aliquots were removed for RIAs of the different eicosanoids. (Abu-Raya et al., 1993a,b). Standard curves were generated with the respective eicosanoids. After incubation of samples (or standard) for 18 to 24 hr with the appropriate antiserum and radioligands, free and bound compounds were separated by dextran coated with activated charcoal. Radioactivity was counted in a -scintillation counter (LKB, Wallac Oy, Finland).

Measurement of [Ca]_i levels. The concentration of [Ca]_i was measured using the fluorescent calcium chelator Fura 2, as previously described (Lazarovici and Lelkes, 1992). PC12 cells were collected and incubated at 20°C for 60 min in the dark in culture medium containing 5 µM Fura 2-AM. After washing, the cells were resuspended in fresh medium at a concentration of 2 × 10⁶ cells/ml, in the presence or absence of Pardaxin, for 15 min in a SPEX Fluorolog 2 spectrofluorometer. [Ca]_i was measured by the fluorescence obtained by excitation at 340 nm and emission at 510 nm. A 435-nm cut-off filter was used to reduce light scattering. After an initial equilibration of the fluorescent signal for 3 to 5 min, the base line remained stable over the duration of the experiment (15 min) (Lazarovici and Lelkes, 1992).

Measurement of PLC activity. PC12 cells were pretreated with [³H]myoinositol, and assays were performed as previously described (Fink et al., 1989). The [³H]inositol phosphates in the cultures were extracted with perchloric acid and analyzed by anion-exchange chromatography.

[³H]Dex binding. PC12 cultures grown to confluency on small Petri dishes (2 × 10⁶ cells/dish) were washed with serum-free DMEM and incubated with different concentrations of [³H]Dex for 2 hr at 37°C in 1 ml of serum-free medium. To measure nonspecific binding, the same experiments were performed in the presence of 1 µM unlabeled Dex. In the competition experiments, cells were incubated with [³H]Dex (20 nM) and the respective antagonist (1 µM). At the end of the incubation, the cultures were cooled for 30 min at 4°C and washed three times with cold DMEM. After sonication, aliquots were counted in a scintillation counter.

Statistics. All results are presented as mean ± S.E.M. of three experiments (n = 3-6 in each experiment). Determination of statistically significant differences between experimental groups was performed using ANOVA analysis, and differences were considered significant when P values less than .05 were obtained.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Stimulation of eicosanoid release in PC12 cells by Pardaxin. The effect of Pardaxin on PGE₂ release in the presence or absence of [Ca]_o is presented in figure 1. Treatment of the cultures with 5 µM and 15 µM Pardaxin for 30 min in calcium-containing medium stimulated PGE₂ release by 5- and 13-fold, respectively, over that of the control (P < .01). In the absence of [Ca]_o, PGE₂ release in the presence of Pardaxin (5-15 µM) was between 50% and 70% of that obtained in the presence of [Ca]_o (fig. 1). Pardaxin-induced PGE₂ release was detected as early as 5 min after addition of the toxin and reached a maximum at 30 min (data not shown). Much like native Pardaxin, synthetic Pardaxin (5 µM) stimulated PGE₂ release about 3-fold over that of the control (table 1). To verify the selective effect of Pardaxin on PGE₂ release, three Pardaxin structural analogs without -helical structure (Shai and Oren, 1996) were tested. They proved to be ineffective in stimulating PGE₂ release (table 1).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Dose-dependent stimulation of PGE2 release from PC12 cells by Pardaxin in the presence (+Ca) or absence (Ca) of [Ca]o (+ 1 mM EGTA). PC12 cells were incubated with Pardaxin in the appropriate concentrations at 37°C for 30 min, and PGE2 was determined in the culture medium. The values are the mean ± S.E.M. of three experiments (n = 3).

1

1

Stimulation of AA release from PC12 cells by Pardaxin. Because the triggering of eicosanoid release by Pardaxin could be due to phospholipases stimulation, we explored this possibility by treating PC12 cultures labeled with [³H]AA with several concentrations of Pardaxin, in the presence or absence of [Ca]_o. In the presence of [Ca]_o, 1 µM and 10 µM Pardaxin stimulated AA release 2.3-fold (P < .05) and 10-fold, respectively (P < .01), vs. the control (fig. 2). In the absence of [Ca]_o, 1 µM and 10 µM Pardaxin stimulated AA release 1.6-fold (P < .05) and 7.1-fold, respectively (P < .05), vs. the control (fig. 2). Stimulation of AA release by Pardaxin (5 µM) was detected after 5 min of incubation, and maximal stimulation was measured after 30 min of incubation. The same time course of AA release was observed in the presence or absence of [Ca]_o (fig. 3). A careful comparison of the kinetics of AA release observed in the presence and absence of [Ca]_o indicates that upon short-term (5-min) incubation, the bulk (about 80%) of the Pardaxin-induced AA release was independent of [Ca]_o, whereas after 30 min, about 50% of the Pardaxin-induced AA release was independent of [Ca]_o (fig. 3).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Dose-dependent stimulation of [3H]AA release from PC12 cells by Pardaxin in the presence (+Ca) or absence (Ca) of [Ca]o (+ 1 mM EGTA). Cells labeled with [3H]AA were incubated with the toxin for 30 min. The results are the mean ± S.E.M. of three experiments (n = 3).

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Time course of Pardaxin-induced [3H]AA release from PC12 cells in the presence (+Ca) or absence (Ca) of [Ca]o (+ 1 mM EGTA). Cells prelabeled with [3H]AA were incubated with Pardaxin (PX, 5 µM). The results are the mean ± S.E.M. of three experiments (n = 3).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   The effect of PLA2 inhibitors on AA release by Pardaxin. [3H]AA-labeled cells were preincubated for 24 hr with Dex (1 µM) or for 15 min with mepacrine (Mep, 100 µM) or pBPB (50 µM), before stimulation with Pardaxin (5 µM), for 15 min. The values are the mean ± S.E.M. of three experiments (n = 3). The basal release value of [3H]AA (1204 ± 123 cpm/mg protein) was subtracted from all experimental values. Release of AA in all experimental groups was lower (P < .01) than in the control (Pardaxin alone).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   The effect of [Ca]o on [Ca]i in response to Pardaxin. Fura 2-AM-loaded PC12 cells were treated with KCl (50 mM), ionomycin (20 µM), or Pardaxin (5 µM) in medium containing 2 mM CaCl2 (+Ca++) (panels A and B) or calcium-free medium containing 1 mM EGTA (Ca++) (panel B). Arrows indicate time of addition of the compound.

The effect of Dex on Pardaxin-stimulated PGE₂ release. Under conditions identical to those described in table 2, PC12 cultures were treated with Dex, a known inhibitor of eicosanoid synthesis (Glaser et al., 1993), to examine its effect on Pardaxin-stimulated PGE₂ release. As table 6 shows, treatment of the cells with Dex (100 nM and 1 µM) for 4 hr inhibited PGE₂ release by 20% to 30%. Treatment of the cells with Dex (1 µM) for 24 hr inhibited Pardaxin-induced PGE₂ release by 50% to 60% (data not shown). To clarify whether the partial inhibition of Pardaxin-induced AA and PGE₂ release by Dex was due to insufficient concentration of steroid, we investigated corticosteroid receptor binding in PC12 cells (fig. 6). The concentrations of Dex used in the experiments, 100 nM and 1 µM, were higher than those required to saturate the corticosteroid receptors, as measured by radioreceptor assay, with [³H]Dex (fig. 6A). Competition experiments among Dex (20 nM), RU-28318 (a corticosteroid type-I receptor antagonist, 1 µM) and RU-38486 (a corticosteroid type-II receptor antagonist, 1 µM) (McEwen et al., 1986; De Kloet, 1991) characterized the [³H]Dex binding sites in PC12 cells as type-II receptors (fig. 6B). These results were further supported by the ability of RU-38486 (table 6), but not of RU-28318 (data not shown), to abolish Dex inhibition of Pardaxin-induced PGE₂ release.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Characterization of [3H]Dex-specific binding by PC12 cells. A) [3H]Dex specific binding at different concentrations. B) Competition between [3H]Dex (DEX, 20 nM) and corticosteroid receptor antagonists (1 µM). PC12 cultures (2 × 106 cells/dish) were incubated with [3H]Dex for 2 hr at 37°C in the presence (nonspecific binding) or absence (total binding) of unlabeled Dex (1 µM). The specific binding data represent the mean ± S.E.M. of three experiments (n = 3).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In the present study, we demonstrate that both native and synthetic Pardaxin trigger the AA cascade in PC12 cells. This stimulation is mediated by activation of PLA₂, as reflected in the massive release of AA and various eicosanoids. Stimulation of the AA cascade in PC12 cells by Pardaxin was blocked by a variety of AA cascade inhibitors.

The activation of cellular phospholipases is thought to mediate signal transduction of certain receptors, and it has been related to a variety of cellular events (Serhan et al., 1996). A moderate increase in AA release is observed after receptor-mediated PLA₂ activation by different ligands such as nerve growth factor (Fink and Guroff, 1990), ACh (Felder et al., 1990a), serotonin (Felder et al., 1990b), histamine (Murayama et al., 1990), bradykinin (Allen et al., 1992) and epidermal growth factor (Schalkwijk et al., 1995). In the present study, we found that Pardaxin-induced AA release in PC12 cells was up to 10-fold greater than that in the control. Such strong stimulation of PLA₂ activity is typical of other cytolytic toxins (Abu-Raya et al., 1993a).

The effect of cytolytic toxins on PLA₂ stimulation may be explained by assuming a drastic change in plasma membrane phospholipid organization due to insertion of the -helical structures in the phospholipid bilayer. Indeed, Pardaxin, an -helical toxin, induced aggregation of PS vesicles (Lelkes and Lazarovici, 1988); melittin, another -helical toxin, changed the phase-transition properties of the phospholipid bilayer (Mollay, 1976). It is well known that PLA₂ activity is very sensitive to the molecular interactions between the plasma membrane constituents, lateral surface pressure and the surface potential of the phospholipid interface (Jain and Berg, 1989; Bell and Biltonen, 1989). Therefore, it is tempting to propose that structural changes in the plasma membrane produced by the binding of small amphipathic polypeptides (Maher and Singer, 1984) such as Pardaxin increase the accessibility of the phospholipid to the hydrolytic action of PLA₂, resulting in enhanced AA release. Lacking a significant -helical structure, Pardaxin analogs with D-amino acids (diasteromers) (Shai and Oren, 1996) cannot insert into the plasma membrane, so there is no activation of PLA₂ and no eicosanoid release, as shown in table 1. Our study indicates that the addition of Pardaxin to PC12 cells resulted in a preferential reduction in PC level that was accompanied by an increase in LPC formation (table 4). This suggests that AA is mobilized mainly from PC during Pardaxin treatment.

Pardaxin, as well as other -helical cytolysins, aggregate in the plasma membrane to form ionic channels (Lear et al., 1988; Bloch-Shilderman et al., 1997), followed by calcium influx (fig. 5, Nikodijevic et al., 1992; Lazarovici and Lelkes, 1992). The latter phenomenon appears to be partially responsible for the stimulatory effect of Pardaxin on neurotransmitter release (Lazarovici and Lelkes, 1992; Lazarovici, 1994). In the present study, Pardaxin-induced AA release due to PLA₂ stimulation was observed in the presence or absence of [Ca]_o (figs. 2 and 3). As a matter of fact, the calcium-independent activation of PLA₂ by Pardaxin occurred earlier and was the major contributor to AA release. PLA₂, the rate-limiting enzyme in the AA cascade, may be divided into three main categories: calcium-dependent secretory PLA₂ (sPLA₂); calcium-dependent cytosolic PLA₂ (cPLA₂) and calcium-independent PLA₂ (iPLA₂) (Balsinde and Dennis, 1997). In calcium-containing medium, it would appear reasonable that Pardaxin induces an increase in [Ca]_i because of calcium influx from the extracellular medium (fig. 5), thereby stimulating calcium-dependent PLA₂. However, Pardaxin also stimulated PLA₂ activity in the absence of [Ca]_o, (figs. 2 and 3). This stimulation cannot be explained by PLC activation or release of calcium from intracellular stores, because no increase in [³H]IP₁ and no decrease in ³²P-PI level (table 4) were observed after Pardaxin treatment. Furthermore, no change in [Ca]_i occurred in calcium-free medium (fig. 5). Therefore, under these conditions, Pardaxin-induced AA release is probably due to stimulation of iPLA₂. The participation of iPLA₂ in Pardaxin-induced AA release is further supported by the finding that the iPLA₂ inhibitors BEL (5 µM) and MAFP (25 µM) markedly suppressed this process (table 5). BEL selectively inhibits iPLA₂ (Balsinde and Dennis, 1996). Recently, it was reported that BEL also blocks PA phosphohydrolase, resulting in inhibition of DAG production (Balsinde and Dennis, 1996). It is unlikely that this would affect Pardaxin-induced AA release, because Pardaxin did not cause any change in the level of PA (table 4), and the DAG lipase inhibitor RHC-80267 did not influence Pardaxin stimulation of PLA₂ (table 3). The other PLA₂ inhibitor we tested, MAFP, although it does not discriminate between cPLA₂ and iPLA₂ (Lio et al., 1996) also inhibited Pardaxin-induced AA release. Further identification of the PC12 PLA₂ enzymes stimulated by Pardaxin is now under way in our laboratory.

Dex, a known inhibitor of eicosanoid synthesis in different cellular systems (Glaser et al., 1993), only partially blocked Pardaxin-induced AA release (fig. 4) and PGE₂ production (table 6) at receptor-saturating concentrations (fig. 6A). Using corticosteroid receptor antagonists (McEwen et al., 1986; De Kloet, 1991), we characterized, for the first time, the corticosteroid receptors in PC12 cells as type II receptors (fig. 6B). This finding is also supported by the ability of the corticosteroid type II receptor antagonist RU-38486 to abolish Dex-induced inhibition of PGE₂ release in response to Pardaxin (table 6). Because PC12 cells are derived from a tumor of the adrenal medulla (Tischler and Greene, 1978), in which the chromaffin cells are exposed to high levels of glucocorticoids, it is not surprising that the low-affinity corticosteroid type II receptors that we found in the PC12 cells are similar to those observed in chromaffin cells (Betito et al., 1992).

The AA release after Pardaxin stimulation of PLA₂ results in the synthesis of the cyclooxygenase products in PC12 cells: mainly PGE₂ and to a lesser degree TXB₂ and 6-keto-PGF₁. Similar results were obtained with PC12 cultures under ischemic conditions (Abu-Raya et al., 1993b). In the present experiments, pretreatment of PC12 cells with indomethacin completely abolished the release of these eicosanoids (table 2). Our results point to a direct correlation between Pardaxin-stimulated AA release and PGE₂ production: 1) in the presence of [Ca]_o, Pardaxin caused greater amounts of AA (figs. 2 and 3) and PGE₂ (fig. 1) to be released than in calcium-free medium, and 2) inhibition of AA release in response to PLA₂ inhibitors (fig. 4) corresponded to a parallel reduction in Pardaxin-induced PGE₂ release (table 6 and data not shown).

Stimulation of the AA cascade and the subsequent formation of eicosanoids such as prostacyclin (PGI₂) in endothelial cells (Suttorp et al., 1985), leukotriene B₄ in polymorphonuclear leukocytes (Suttorp et al., 1987) or platelet-activating factor (PAF) in pulmonary artery endothelial cells (Suttorp et al., 1992) was reported for Staphylococcus aureus -toxin. Other cytolysins, such as streptolysin S, Stoichatus helianthus toxin, parcelsin and cobra direct lytic factor, also stimulated AA and eicosanoid release from PC12 cells (Abu-Raya et al., 1993a). The generation of eicosanoids by cytolysins might explain the wide spectrum of physiological and pathological effects caused by these toxins, such as neurotransmitter release, pulmonary hypotension, blood coagulation, endotoxic shock, edema and inflammation (Seeger et al., 1984; Bhakdi et al., 1988; Snyder, 1990) observed upon exposure to such toxins.

As we have noted, Pardaxin, in contrast to maitotoxin (Choi et al., 1990), melittin and mastoparan (Choi et al., 1992) did not stimulate PLC. In addition, Pardaxin treatment did not affect PLD activity as measured by PA formation. The present study indicates that the AA release by this toxin was due mainly to PLA₂ stimulation. Therefore, Pardaxin represents a more selective pharmacological tool for investigating the role of the AA cascade in cellular processes, the identification of PLA₂ enzymes and the development of PLA₂ inhibitors.