Effect of Arachidonic Acid Reacylation on Leukotriene Biosynthesis in Human Neutrophils Stimulated with Granulocyte-macrophage Colony-stimulating Factor and Formyl-methionyl-leucyl-phenylalanine*

Priming of human neutrophils with granulocyte-macrophage colony-stimulating factor (GM-CSF) followed by treatment with formyl-methionyl-leucyl-phenylalanine (fMLP) stimulates cells in a physiologically relevant manner with modest 5-lipoxygenase activation and formation of leukotrienes. However, pretreatment of neutrophils with thimerosal, an organomercury thiosalicylic acid derivative, led to a dramatic increase (>50-fold) in the production of leukotriene B4 and 5-hydroxyeicosatetraenoic acid, significantly higher than that observed after stimulation with calcium ionophore A23187. Little or no effect was observed with thimerosal alone or in combination with either GM-CSF or fMLP. Elevation of [Ca2+]i induced by thimerosal in neutrophils stimulated with GM-CSF/fMLP was similar but more sustained compared with samples where thimerosal was absent. However, [Ca2+]i was significantly lower compared with calcium ionophore-treated cells, suggesting that a sustained calcium rise was necessary but not sufficient to explain the effects of this compound on the GM-CSF/fMLP-stimulated neutrophil. Thimerosal was found to directly inhibit neutrophil lysophospholipid:acyl-CoA acyltransferase activity at the doses that stimulate leukotriene production, and analysis of lysates from neutrophil preparations stimulated in the presence of thimerosal showed a marked increase in free arachidonic acid, supporting the inhibition of the reincorporation of this fatty acid into the membrane phospholipids as a mechanism of action for this compound. The dramatic increase in production of leukotrienes by neutrophils when a physiological stimulus such as GM-CSF/fMLP is employed in the presence of thimerosal suggests a critical regulatory role of arachidonate reacylation that limits leukotriene biosynthesis in concert with 5-lipoxygenase and cytosolic phospholipase A2α activation.

Neutrophils play an important role in natural immunity and function to eliminate microbes and dead tissues. They are also a major cell involved in the acute inflammatory response, since they respond rapidly to chemotactic stimuli and can be activated by cytokines produced pri-marily by macrophages to perform degradative and phagocytic functions. Upon activation at the site of inflammation, they release lysosomal enzymes and neutral proteinases, reactive oxygen species, including H 2 O 2 and O 2 . (1,2), and metabolites of arachidonic acid (AA), 2 in particular leukotriene B 4 (LTB 4 ) (3). Biosynthesis of leukotrienes involves the release of AA from membrane phospholipids by Ca 2ϩ -dependent cytosolic phospholipase A 2 ␣ (cPLA 2 ␣) (4) and its conversion into 5-hydroperoxyeicosatetraenoic acid and subsequently leukotriene A 4 (LTA 4 ), catalyzed by 5-lipoxygenase (5-LO) (5). LTA 4 is then enzymatically converted to LTB 4 by LTA 4 -hydrolase (6) or to leukotriene C 4 (LTC 4 ) by the addition of a molecule of glutathione through the action of LTC 4 -synthase (7). The site for most of these biochemical events is the perinuclear region (8).
Regulation of leukotriene production can occur at different levels; the increase in Ca 2ϩ and release of free arachidonate by cPLA 2 ␣ are fundamental for 5-LO full activation. Specifically, Ca 2ϩ binds to a calciumdependent phospholipid binding domain (C2) on cPLA 2 ␣ that triggers its translocation from the cytosol to the membrane of the Golgi apparatus, nuclear envelope, and endoplasmic reticulum, where it can access its phospholipid substrate (9). 5-LO activity is enhanced by ATP-dependent phosphorylation, and micromolar concentrations of Ca 2ϩ are required for its catalytic activity and translocation from the cytosol to the nuclear membrane, where it interacts with the 5-LO-activating protein, which facilitates the conversion of AA into leukotrienes by "presenting" the substrate to 5-LO (10).
Leukotriene synthesis is characterized by a high cellular specificity, with neutrophils synthesizing exclusively LTB 4 , since they contain only 5-LO and LTA 4 -hydrolase. In these cells, LTB 4 is also rapidly metabolized into biologically inactive -oxidation products (20-OH-LTB 4 and 20-COOH-LTB 4 ) by a specific cytochrome P-450 enzyme (11).
A common model for neutrophil activation in vitro has been the exposure to formyl-methionyl-leucyl-phenylalanine (fMLP), a bacterial chemotactic tripeptide that interacts with a specific membrane receptor (12), but the stimulation with this peptide alone does not lead to formation of detectable amounts of leukotrienes (13). The pretreatment of neutro-phils with granulocyte-macrophage colony-stimulating factor (GM-CSF), a glycoprotein that induces proliferation and differentiation of myeloid progenitor cells (14), results in a "priming" effect on these cells, with consequent increase of the leukotriene production in response to the stimulation with physiological agonists, such as fMLP and platelet-activating factor (15). The exact mechanisms through which GM-CSF exerts its priming effects are not entirely known, but there is evidence that this cytokine activates protein kinase C and cPLA 2 ␣ phosphorylation without increasing intracellular Ca 2ϩ and causes an enhanced release of intracellular AA (16).
The levels of free AA in cells are regulated by not only the cPLA 2 ␣mediated release from phospholipids but also the deacylation-reacylation cycle that results in reincorporation of free arachidonate into phospholipids by sequential action of acyl-CoA ligase and lysophosphatidylcholine:acyl-CoA acyltransferase (LAT) (EC 2.3.1.23) (17, 18). The initial reacylation of arachidonoyl-CoA (AA-CoA) into neutrophil phospholipids is thought to take place with 1-acyl-2-lyso-phospholipids followed by a transfer of arachidonate by a CoA-independent transacylation to plasmalogen lysoglycerophosphoethanolamine lipids (19).
Thimerosal is a water-soluble thiosalicylic acid derivative, mainly used as a preservative in vaccines and various chemicals, which is characterized by the presence of an atom of mercury in its structure (20) and has been used to enhance eicosanoid production in cells. It is well known as a sulfhydryl reagent, and it has been hypothesized that thimerosal covalently interacts with sulfhydryl groups present in the active site of target enzymes, including LAT (21), leading to enzyme inactivation. This organomercurial has been shown to induce intracellular Ca 2ϩ spikes in different cell types (22,23) and to increase prostaglandin synthesis in murine macrophages (24,25) and leukotriene production in neutrophils stimulated with fMLP (26) but also to have an inhibitory effect on 5-LO and the -oxidation system of LTB 4 in neutrophils stimulated by A23187 (27).
The work described here suggests that regulation of leukotriene production in the neutrophil involves not only the activation of cPLA 2 ␣ and 5-LO but also the absolute level of free arachidonate maintained during the activation event. Reacylation of free arachidonate is dependent upon formation of AA-CoA and lysophospholipid substrates and upon LAT activity. Regulation of this reacylation pathway represents a key factor in the absolute level of eicosanoid production in the human neutrophil.
Neutrophils were then centrifuged and resuspended in calcium/magnesium-free Hanks' balanced salt solution (Invitrogen) at a concentration of 10 ϫ 10 6 cells/ml and primed with either GM-CSF (1 nM), tumor necrosis factor ␣ (10 ng/ml), or Kdo 2 -lipid A (100 ng/ml) for 30 min at 37°C. After the addition of CaCl 2 (final concentration 2 mM) and MgCl 2 (final concentration 0.5 mM), the cells were stimulated with fMLP (10 Ϫ7 M) for 10 min at 37°C. Thimerosal at different concentrations (5-100 M) was added for 2 min at 37°C before fMLP. Final incubation volume was 1 ml.
Alternatively, cells were stimulated with calcium ionophore A23187 Metabolite Separation and Analysis by Reversed Phase HPLC and Electrospray Ionization-Mass Spectrometry-An aliquot of each sample (35 l) was injected into an HPLC system, and separation of the different metabolites was conducted using a C18 column (Columbus 150 ϫ 1 mm, 5 m; Phenomenex) eluted at a flow rate of 50 l/min with a linear gradient from 25 to 100% of mobile phase B. Solvent B was increased from 25 to 85% in 24 min, increased to 100% in 2 min, and held at 100% for a further 12 min. The HPLC system was directly interfaced into the electrospray source of a triple quadrupole mass spectrometer (Sciex API 2000, PE-Sciex, Thornhill, Canada) where mass spectrometric analyses were performed in the negative ion mode using multiple reaction monitoring of the specific transitions m/z 319 3 115 for 5-HETE, m/z 335 3 195 for LTB 4 4 . Quantitation was performed using a standard isotope dilution curve as previously described (29).
Calcium Assay-Neutrophils (10 ϫ 10 6 /ml in Hanks' balanced salt solution) were loaded with 5 M Indo-1 AM and 0.02% Pluronic F-127 for 30 min at 37°C. Cells were then washed and incubated with or without GM-CSF (1 nM) for an additional 30 min at 37°C. After the addition of a 1 mM concentration each of CaCl 2 and MgCl 2 , primed cells were pretreated with thimerosal (50 M) or vehicle 2 min prior to stimulation with fMLP (10 Ϫ7 M). Cells that had not been treated with GM-CSF were stimulated with the calcium ionophore ionomycin (5 M). Intracellular calcium changes were monitored by measuring the ratio of the fluorescence emissions at 405 and 485 nm (excitation wavelength 355 nm) in a Fluorolog-3 spectrofluorometer (Horiba Jobin Yvon, Edison, NJ).
Gas Chromatography (GC)/MS and Quantitation of Free Fatty Acids-One aliquot (corresponding to 1 ϫ 10 6 cells) of neutrophil lysate (1: 8 ]arachidonic acid were added. After mixing, samples were extracted with 1 ml of iso-octane. The extract was dried under a stream of N 2 and then derivatized by adding 25 l each of 1% pentafluorobenzyl bromide and 1% diisopropylethylamine. After mixing, the vials were incubated at room temperature for 20 min, dried under a stream of N 2 , and reconstituted in 100 l of iso-octane. Analysis of the samples was performed by negative ion chemical ionization GC/MS (30) on a Finnigan DSQ GC/MS system (Thermo Finnigan, Thousand Oaks, CA), using a ZB-l column (15 m ϫ 0.25-mm inner diameter ϫ 0.10-m film thickness; Phenomenex). The GC was programmed to increase the temperature from 150 to 250°C at 20°C/ min, 280°C at 5°C/min, and finally 310°C at 30°C/min and held at 310°C for 1 min. The mass spectrometer was operated in the negative ion chemical ionization mode using methane as reagent gas. Data were acquired by selected ion monitoring of the following fatty acids: myristic (m/z 227), palmitic (m/z 255), stearic (m/z 283), linolenic (m/z 277), linoleic (m/z 279), oleic (m/z 281), eicosapentaenoic (m/z 301), and arachidonic (m/z 303). The ions at m/z 259, 286, 283, and 311 were monitored for [ 13 8 ]arachidonic acids, respectively. Peak areas of each analyte were measured, and the ratio of the area from the fatty acidderived ion to that from the internal standard was calculated for each. These ratios were then compared with calibration curves for each analyte prepared from commercially available standards to calculate fatty acid amounts.
Preparation of Microsomes from Human Neutrophils-The isolation of microsomes from human neutrophils was achieved following a previously described protocol for the extraction of LAT from pig spleen membranes (31) with modifications. Briefly, neutrophils, either resting or stimulated, were suspended in LAT homogenization buffer (250 mM sucrose, 50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 20% w/v glycerol) containing a protease inhibitor mixture (Complete TM ; 1 tablet dissolved in 20 ml of buffer) at a concentration of 100 ϫ 10 6 cells/ml. The cells were disrupted by sonication (4 ϫ 15 s at 4°C) using a high intensity ultrasonic processor (Vibracell VCX 600; Sonics and Materials Inc., Newtown, CT) equipped with a microtip (3-mm diameter) set at an amplitude of 23% and then centrifuged at 15,000 ϫ g for 15 min at 4°C to pellet the unbroken cells. The supernatant was centrifuged at 100,000 ϫ g for 1 h at 4°C, and the microsome pellet was resuspended in LAT assay buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA) at a final protein concentration of 10 mg/ml. Protein content was determined by the bicinchoninic acid assay (Pierce), using bovine serum albumin dissolved in 0.9% NaCl as a standard.
LAT Activity Assay-Neutrophil microsomes (20 g) were incubated in LAT assay buffer (200 l final volume) containing bovine serum albumin (12.5 M) with AA-CoA (30 M) and with or without thimerosal (50 M) for 2 min at 37°C. A mixture of 16:0a/OH-GPCho and 19:0a/OH-GPCho (50 M each) was added, and the reaction was allowed to proceed for 30 min at 37°C. Samples were then extracted using chloroform/methanol according to the method of Bligh and Dyer

RESULTS
Leukotriene Production and Thimerosal-Leukotriene synthesis in the human neutrophil can be initiated through different stimuli. In our study, neutrophils (10 ϫ 10 6 /ml) were pretreated with 1 nM GM-CSF for 30 min at 37°C and then stimulated with 10 Ϫ7 M fMLP for 10 min at 37°C. Metabolite production was assessed by liquid chromatography followed by mass spectrometry through measurement of the ion abundances for the specific collision-induced transformations for 20-OH-LTB 4 , LTB 4 and ⌬ 6 -trans-LTB 4 , LTC 4 , 5,6-di-HETEs, 5-HETE, and AA and the internal standards [d 4 ]LTB 4 and [d 8 ]5-HETE.
The stimulation with GM-CSF and fMLP resulted in a modest production of total LTB 4 (0.24 Ϯ 0.09 ng/10 6 cells) and 5-HETE (0.22 Ϯ 0.08 ng/10 6 cells) (Fig. 1A, inset), with total LTB 4 corresponding to the sum of LTB 4 and the -oxidized metabolite 20-OH-LTB 4 . The small amount of LTC 4 detected was probably due to the presence of eosinophils in the cell preparation, since neutrophils do not express LTC 4synthase, the enzyme responsible for the conversion of LTA 4 into LTC 4 . The combination of GM-CSF and fMLP was necessary to achieve a detectable amount of 5-LO metabolites, since the analysis of lysates obtained from cells stimulated with either GM-CSF or fMLP alone did not show any measurable quantity of LTB 4 or 5-HETE (Fig. 2). Pretreatment of GM-CSF-primed cells with the organomercury compound thimerosal (50 M) resulted in a dramatic increase (50-fold) in the synthesis of both total LTB 4 and 5-HETE (12.22 Ϯ 0.35 and 10.12 Ϯ 2.52 ng/10 6 cells, respectively) after stimulation with fMLP ( Figs. 1B and 2).
Thimerosal alone or with GM-CSF did not induce the synthesis of detectable amounts of 5-LO metabolites, whereas the combination of thimerosal and fMLP yielded an increase in synthesis compared with samples from cells stimulated with GM-CSF and fMLP (0.81 Ϯ 0.13 and 2.4 Ϯ 0.3 ng/10 6 cells for LTB 4 and 5-HETE, respectively) (Fig. 2), in agreement with previously published data (26). The dramatic effects on leukotriene synthesis observed when neutrophils were primed with GM-CSF and stimulated with fMLP in the presence of thimerosal sug-gested that full cell stimulation and activation of cPLA 2 ␣ and 5-LO were required for this organomercury compound to exert its activity. The translocation of cytosolic 5-LO to nuclear membranes was also assessed in GM-CSF/fMLP-challenged neutrophils in the presence and absence of thimerosal (50 M) by SDS-PAGE and immunoblotting. Thimerosal treatment led to a substantial increase in membrane associated 5-LO, comparable with A23187-treated cells (see supplemental Fig. 1).
The production of 5-LO metabolites in the presence of GM-CSF, fMLP, and thimerosal was induced very rapidly, with a maximal production of eicosanoids observed already at 3 min following the addition of fMLP (data not shown), and this level of total LTB 4 did not change after 30 min. In subsequent experiments, cells were challenged for 10 min to be able to make comparisons with other stimuli, such as ionomycin, that require longer incubation times. Surprisingly, the absolute quantity of leukotriene production in thimerosal-treated neutrophils was higher than that observed when these cells were challenged with the calcium ionophore A23187 (8.08 Ϯ 0.74 ng/10 6 cells).
The same increase in leukotriene production by thimerosal was observed when neutrophils were primed with tumor necrosis factor ␣ or Kdo 2 -lipid A (bacterial endotoxin) and challenged with fMLP (Table  1). This effect of increasing LTB 4 production in activated neutrophils was also observed in the presence of other sulfhydryl-modifying reagents such as p-chloromercuribenzoic acid and Alexa Fluor phenylmercury (data not shown).
When different concentrations of thimerosal (5-100 M) were tested on neutrophil preparations stimulated with a standard GM-CSF/fMLP protocol, a significant increase in total LTB 4 production was observed even at the lowest dose of thimerosal tested (5 M) (1.81 Ϯ 0.11 ng/10 6 cells), with a maximal effect at ϳ75 M thimerosal (13.6 Ϯ 0.63 ng/10 6 cells) (Table 1). When 100 M thimerosal was added to the neutrophil samples, LTB 4 synthesis dropped to values lower (1.65 Ϯ 0.3 ng/10 6 cells) than those observed with 5 M thimerosal, suggesting that at higher concentrations in the dose-response curve alternative effects of thimerosal could be observed that decreased 5-LO biosynthesis of LTA 4 from arachidonate. At these higher concentrations, thimerosal (Ͼ75 M) has been reported to be quite cytotoxic to cells (33). The initiation of events at thimerosal concentrations below 75 M appeared to be very fast and irreversible, since the same effects on leukotriene synthesis could be observed when thimerosal 50 M was added to the cells 30 min before (together with GM-CSF) or at the same time as fMLP and also in experiments where the compound was removed by washing thimerosal-pretreated neutrophils with buffer to remove nonbound drug before stimulation with fMLP (data not shown). When neutrophils were stimulated with GM-CSF/fMLP in the absence of calcium ion (1 mM EGTA), there was no leukotriene production either in the presence or the absence of thimerosal (data not shown).
Thimerosal and Intracellular Ca 2ϩ -The effects of thimerosal on [Ca 2ϩ ] i in neutrophils was studied in cells loaded with Indo-1 AM. In all experiments, primed neutrophils responded to fMLP (10 Ϫ7 M) with a rapid rise in [Ca 2ϩ ] i (Fig. 3). The calcium ionophore A23187 was not used in these experiments due to its intrinsic autofluorescence, but rather ionophore stimulation was carried out using ionomycin. When fMLP was added in the presence of 50 M thimerosal, an increase in [Ca 2ϩ ] i similar to that obtained with fMLP alone was observed, but the Ca 2ϩ level was consistently more sustained. Neutrophils stimulated with ionomycin (5 M) showed a higher increase in [Ca 2ϩ ] i compared with GM-CSF/fMLP-stimulated cells in the presence of thimerosal (Fig.  3). Since leukotriene production following ionomycin stimulation of human neutrophils was very modest and below the levels observed with even the lowest concentration of thimerosal tested in GM-CSF/fMLPtreated cells, 0.52 Ϯ 0.06 ng/10 6 cells compared with 1.81 Ϯ 0.11 ng/10 6 cells for LTB 4 (Table 1), it can be concluded that the sustained calcium response in the presence of thimerosal would be necessary, but it is not sufficient to explain the effects observed on eicosanoid synthesis.
Adenosine Release and Thimerosal-Adenosine is a known regulator of leukotriene biosynthesis (34), and the effect of thimerosal on adenosine present in the extracellular media of neutrophil incubations was separately investigated. A slight increase in released adenosine was observed from unstimulated neutrophils (23.3 ng/ml) relative to GM-CSF/fMLP-treated cells (31.6 ng/ml). A similar elevation of released adenosine was observed when thimerosal was present in the control incubation (29.8 ng/ml) as well as in stimulated neutrophils in the presence of thimerosal (34.9 ng/ml).
Free Fatty Acid Levels and Thimerosal-Gas chromatography coupled to mass spectrometry was utilized to determine the exact quantity

. LTB 4 and 5-HETE production in human neutrophils. Total LTB 4 (LTB 4 ϩ 20-OH-LTB 4 ) and
5-HETE production in human neutrophils (10 ϫ 10 6 / ml) stimulated with GM-CSF (1 nM) and fMLP (10 Ϫ7 M) in the presence and the absence of thimerosal (50 M) is indicated. 5-LO metabolites were analyzed by reversed phase HPLC followed by electrospray ionization-mass spectrometry, and absolute quantities were calculated through relative abundances of the specific ion transition relative to deuterium-labeled internal standard. Values are expressed as mean Ϯ S.E. (n ϭ 5), with each sample run in duplicate n.d., not detected.

TABLE 1 LTB 4 synthesis in human neutrophils
Human neutrophils (10 ϫ 10 6 /ml) were stimulated with fMLP (10 Ϫ7 M, 10 min at 37°C) after priming with GM-CSF (1 nM), tumor necrosis factor ␣ (10 ng/ml), or Kdo 2 -lipid A (100 ng/ml) (30 min at 37°C) in the presence or absence of thimerosal (5-100 M). Thimerosal was added 2 min before stimulation with fMLP. Alternatively, cells were stimulated with A23187 or ionomycin (5 M, 10 min at 37°C) after 2 min of thermal equilibration at 37°C. After the addition of the internal standards ͓d 4 ͔LTB 4 and ͓d 8 ͔5-HETE, samples were extracted using a C18 solid phase extraction cartridge, and metabolites were analyzed by reversed phase HPLC followed by electrospray ionization-mass spectrometry in the MRM mode. Values are expressed as mean Ϯ S.E. (n ϭ 5), with each sample run in duplicate. ND, not detected. In separate experiments, exogenous arachidonate (0.3-100 M) was added to GM-CSF-primed neutrophil suspension followed by treatment with fMLP. A dose-dependent increase in leukotriene product was observed, and at the highest arachidonate concentrations employed (100 M), leukotriene biosynthesis was found to be 13.4 Ϯ 0.5 (ng/10 6 cells), comparable with that observed when neutrophils were primed and activated when thimerosal (50 M) was present.
LAT Activity-The levels of free AA in cells are maintained by a balance between hydrolysis from membrane phospholipids by phospholipases A 2 and reincorporation into phospholipids catalyzed by LAT (17,19). Thimerosal had been shown to inhibit the activity of this enzyme in tumor cells (33). To specifically evaluate the effects of thimerosal on human neutrophil LAT, neutrophil microsomes were incubated with AA-CoA in the presence or the absence of thimerosal and with a mixture of 16:0a/OH-GPCho and 19:0a/OH-GPCho substrates for 30 min. The formation of the corresponding phospholipids was assessed by liquid chromatography-mass spectrometry. Neutrophil microsomal fractions by themselves did not contain detectable amounts of these lysophosphatidylcholines but did contain very small quantities of 16:0a/20:4-GPCho and 18:0a/20:4-GPCho. There was no endogenous 19:0a/20:4-GPCho, as monitored in the assay (Fig. 4A). AA-CoA was not incorporated into the standard lysophospholipids in the absence of neutrophil microsomes (Fig. 4B). When the microsomes were incubated with AA-CoA and the mixture of standard lysophospholipids, the corresponding diacylphospholipids were formed and detected through their specific tandem mass spectrometric ion transformation at corresponding retention times (Fig. 4C). LAT enzyme activity in the neutrophil microsomes was expressed as a ratio between the ion abundance areas of the diacylphosphatidylcholines to their corresponding lysophosphatidylcholine precursors. The addition of thimerosal (50 M) 2 min prior to the addition of the lysophospholipid mixture resulted in almost a complete inhibition of the formation of phospholipids (Fig. 4D). The inhibitory effect of thimerosal on the membrane LAT enzyme was irreversible, as observed in experiments where thimerosal was removed from the incubation buffer before the addition of the lysophosphocholines (data not shown). Neutrophil microsomes appeared to be more efficient in converting 16:0a/OH-GPCho rather than 19:0a/OH-GPCho into the corresponding diacylphospholipid, but in both cases, pretreatment with thimerosal (50 M) profoundly inhibited the reincorporation of the AA-CoA into the lysophospholipids, almost 90% (Table 3). Very similar inhibitory effects on LAT in vitro activity were obtained with 50 M p-chloromercuribenzoic acid or 50 M Alexa Fluor phenylmercury (data not shown).
Using the LAT activity assay, the dependence of thimerosal concentration on diacylphosphatidylcholine synthesis from lysophosphatidylcholine and AA-CoA was measured (Fig. 5). The production of diacylphosphatidylcholines decreased exponentially with ϳ50% inhibition at 20 M thimerosal. This concentration of thimerosal was very close to the estimated concentration of thimerosal which stimulated LTB 4 production to the half-maximal level (Table 1).
Since thimerosal was found to elevate intracellular Ca 2ϩ levels in a prolonged fashion (Fig. 3), the activity of LAT was measured at different Ca 2ϩ concentrations. Since EDTA was present in the assay buffer, CaCl 2 was added in amounts calculated to yield the concentrations of free calcium needed (35). LAT activity appeared to be calcium-independent, since no significant changes in activity were observed when free Ca 2ϩ concentration was changed in a stepwise manner from 0 to 5 M (data not shown).
The activity of LAT was also determined in microsomes isolated from neutrophils treated with and without thimerosal (50 M) under the same conditions used to stimulate leukotriene production (see above) in intact cells. Noncovalently attached thimerosal was washed away as a consequence of the microsome preparation. LAT was significantly inhibited in microsomes from thimerosal exposed neutrophils compared with microsomes from untreated neutrophils (Fig. 6), supporting the assumption that thimerosal covalently binds to LAT to irreversibly inhibit its activity.

DISCUSSION
The biosynthesis of leukotrienes in the neutrophil is known to be the result of a complex series of reactions. Considerable focus has centered on the release of AA from membrane phospholipids by cPLA 2 ␣ following translocation as well as translocation of 5-LO to the perinuclear environment. It is there that 5-LO-activating protein assembles with the other two enzymes in a manner that permits AA to enter the active site of 5-LO and be converted into LTA 4 (36). The initiation of these events through activation of transmembrane receptors by cytokines such as GM-CSF and by chemoattractants such as fMLP is known to trigger the

TABLE 2 GC/MS analysis of free fatty acids in the human neutrophil
Lysates from human neutrophils (1 ϫ 10 6 ) stimulated as indicated were analyzed by GC/MS operated under negative ion chemical ionization conditions. Each fatty acid was detected as the carboxylate anion, and quantitation was achieved by stable isotope dilution of added internal standards. Values (ng/1 ϫ 10 6 cells) are expressed as mean Ϯ S.E. (n ϭ 3), with each sample run in duplicate.

Fatty acid
Resting GM-CSF fMLP GM-CSF/fMLP/thimerosal A23187 generation of detectable quantities of LTB 4 in the isolated human neutrophil. The elevation of intracellular calcium is critical for the translocation events as well as for the activation of 5-LO (4, 10). The central importance of cPLA 2 ␣ was clearly established through studies by the groups of Shimizu (37) and Bonventre (38), who independently observed that cells derived from animals that did not express cPLA 2 ␣ failed to generate leukotrienes as well as other metabolites of AA and platelet-activating factor. The results of the experiments described here suggest that other control mechanisms are in place that limit the extent to which AA is converted into LTA 4 within the neutrophil. Previous papers have investigated the effects of thimerosal on eicosanoid production in different cell types and by different stimuli with alternative hypotheses proposed to explain the mechanism of action of this organomercury compound. However, the dramatic effects of this compound on leukotriene production and polyunsaturated fatty acid release in neutrophils stimulated by the combination of GM-CSF and fMLP have not been reported. Previous reports suggest that thimerosal elevated the production of prostaglandins and leukotrienes by increasing intracellular calcium (39) or by

TABLE 3 Effect of thimerosal on LAT activity in neutrophil microsomes
Neutrophil microsomes (20 g) were incubated for 30 min at 37°C with a mixture of 16:0a/OH-GPCho and 19:0a/OH-GPCho (50 M each) and AA-CoA (30 M) and in the presence or the absence of thimerosal (50 M). Bligh-Dyer extracts of membrane samples were analyzed by reversed phase HPLC followed by electrospray mass spectrometry in the MRM mode. LAT activity is expressed as a ratio between the intensity peak areas of the product diacylphospholipid to the corresponding lysophospholipid precursor. inhibiting reacylation of AA into cellular phospholipids (33). Hecker et al. (39) speculated that thimerosal only slightly increased liberated AA and that its mechanism of action in increasing leukotriene biosynthesis in the neutrophil was the result of further activation of cPLA 2 ␣. Kaever et al. (33) suggested that the increased production of PGE 2 in P388D1 cells was a consequence of interfering with the fatty acid deacylation-reacylation cycle that would result in elevated concentrations of AA. Such an effect was thought to be a result of direct inhibition of LAT, as evidenced by the decrease in [ 14 C]AA incorporated into the different phospholipid classes, most predominantly glycerophosphocholine (33). In our studies, an induction of LTB 4 production was achieved by priming neutrophils with GM-CSF and stimulating them with fMLP. This protocol has been used by many investigators to generate leukotrienes in a physiologically relevant manner, as opposed to incubation with calcium ionophores. One of the surprising findings was that when thimerosal was present in GM-CSF-and fMLP-stimulated neutrophils, there was a greater than 50-fold increase in LTB 4 production as well as in AA release and that these increases were significantly higher than those achieved through stimulation with the calcium ionophore A23187. These studies demonstrated that the level of leukotriene production was parallel to the level of free AA present in these cells. Thus, the concentration of free arachidonic acid within a cell (Table 2) can drive 5-lipoxygenase formation of LTA 4 . This, as well as the observed increase in membrane associated 5-LO measured by immunoblotting in thimerosal treated cells, was consistent with the recent observation that free arachidonate can increase 5-LO translocation to the nuclear envelope (40). However, the level of intracellular calcium ion did not follow this pattern. Whereas thimerosal increased the duration of the [Ca 2ϩ ] i spike initiated by fMLP, it was not to the extent observed for ionomycin. However, the quantity of leukotrienes produced in the presence of ionomycin or A23187 was substantially lower than that with GM-CSF/fMLP treated neutrophils in the presence of thimerosal. Together with the EGTA experiments, these results suggested that whereas the elevation of [Ca 2ϩ ] i is critically involved in the generation of leukotrienes as previously reported (5), it is not sufficient to account for the extent to which LTB 4 biosynthesis was elevated when thimerosal was incubated with neutrophils.
Since the most likely explanation for these effects was the failure of thimerosal-treated cells to reacylate AA, a direct assay to follow AA reacylation into lysophospholipids was investigated in our experimental paradigm. It was clear that the quantity of free AA present within the neutrophil was directly related to the quantity of the LTB 4 produced, no matter what stimulus was employed to release AA, elevate [Ca 2ϩ ] i , and activate 5-LO. This would be the expected result if substrate availability for 5-LO were a major component driving the absolute level of leukotriene production. Since the reacylation pathway catalyzed by LAT reduces the absolute level of free AA by converting lysophosphatidylcholine into arachidonate-containing phospholipids, the reacylation pathway could be a fundamental control mechanism regulating the total quantity of leukotrienes produced within the neutrophil stimulated by receptor occupation or phagocytosis. Thimerosal, in the concentration range used for intact neutrophil studies, caused a dose-dependent inhibition of LAT activity in neutrophil microsome preparations. Moreover, the increased leukotriene synthesis observed in thimerosal-treated neutrophils correlated with the irreversible LAT inhibition.
The fact that thimerosal and other organomercury compounds studied had such a profound effect would also suggest that the oxidation of sulfhydryl groups within the active site of critical enzymes, such as LAT, may significantly alter the overall quantity of leukotrienes that result from neutrophil activation. This level of regulation of leukotriene biosynthesis has been suggested but has not been widely recognized. Rather, the activation of cPLA 2 ␣ and translocation of 5-LO are generally thought to be the critical events in the regulation of this biosynthetic pathway. Reactive oxygen species generated as a result of neutrophil activation of the NADPH oxidase respiratory burst could have an effect on the sulfhydryl-dependent reacylation pathway. For some time, it has been recognized that the reacylation pathway is sensitive to the presence of reduced sulfhydryl groups for full activity (41). In fact, when H 2 O 2 was added to alveolar macrophages, a dose-dependent decrease in arachidonate reacylation could be measured (42). This effect was not due to inhibition of LAT but rather to a decrease in ATP, a cofactor required for AA-CoA generation. GM-CSF is known to induce a respiratory burst with generation of reactive oxygen species (43), and this may decrease the availability of AA-CoA, which is a substrate for LAT. There are additional mechanisms that could reduce AA-CoA generation such as mutations of AA-CoA ligase, FACL4, which leads to mental retardation (44). However, there have not been reports of aberrant eicosanoid production in such individuals with this genetic defect.  In summary, thimerosal irreversibly interacts with LAT to inhibit reacylation of AA into phospholipids. When neutrophils were primed with GM-CSF followed by stimulation with fMLP, there was a reproducible release of AA as well as production of LTB 4 . However, when thimerosal was present, there was a significant increase in the measured amounts of free AA and LTB 4 . Whereas there was a prolonged increase in [Ca 2ϩ ] i , this effect alone could not explain the enhanced synthesis of leukotrienes mediated by thimerosal. These experiments support the hypothesis of a regulatory role of the reacylation pathway on leukotriene biosynthesis and indicate that alterations in this pathway, which could involve AA-CoA ligase as well as LAT, play a central role in leukotriene production in the human neutrophil. Another feature that may explain differences in the extent to which leukotrienes are made in individuals could be genetic diversity in the reacylation pathway either through AA-CoA ligase or LAT.