Release of leukotriene A4 versus leukotriene B4 from human polymorphonuclear leukocytes.

The reactive intermediate formed by 5-lipoxygenase metabolism of arachidonic acid, leukotriene A4, is known to be released from cells and subsequently taken up by other cells for biochemical processing. The objective of this study was to determine the relative amount of leukotriene A4 synthesized by human polymorphonuclear leukocytes (PMNL) that is available for transcellular biosynthetic processes. This was accomplished by diluting cell suspensions and measuring the relative amounts of enzymatic versus nonenzymatic leukotriene A4-derived metabolites after challenge with the Ca2+ ionophore A23187. Nonenzymatic leukotriene A4-derived metabolites were used as a quantitative index of the amount of leukotriene A4 released into the extracellular milieu. The results obtained demonstrated that in human PMNL, the relative amounts of nonenzymatic versus enzymatic leukotriene A4-derived metabolites increased with decreasing cell concentrations. After a 20-fold dilution of PMNL in cell preparations, a doubling in the amount of nonenzymatic leukotriene A4-derived metabolites was observed following challenge (from 53.9 ± 1.3 to 110.4 ± 8.9 pmol/106 PMNL, p < 0.01). Reduction of possible cell-cell interactions by dilution suggested that over 50% of leukotriene A4 synthesized is released from the PMNL. These data provide evidence that, in human PMNL preparations, transfer of leukotriene A4 to neighboring PMNL is taking place, resulting in additional formation of leukotriene B4 and its ω-oxidized metabolites 20-hydroxy- and 20-carboxy-leukotriene B4. Neutrophil reuptake of extracellular leukotriene A4 leads to an underestimation of the fraction of leukotriene A4 that is in fact available for transcellular metabolism when tight cell-cell interactions occur, such as during PMNL adhesion to the microvascular endothelium and diapedesis.

Arachidonic acid oxidation, catalyzed by cyclo-oxygenase or 5-lipoxygenase, leads to potent biologically active molecules such as thromboxane, prostacyclin, and leukotrienes (1,2). While most biochemical studies have focused on cells that possess cyclooxygenase or 5-lipoxygenase (5-LO), 1 it is now clear that the formation of eicosanoids is not strictly limited to those cells which have these primary oxidative enzymes. The discovery of reactive intermediate transfer in eicosanoid biosynthesis was made by Marcus (3) who showed that plateletderived endoperoxides could be transformed into prostacyclin by adjacent endothelial cells. More recently, conversion of LTA 4 to LTB 4 and cysteinyl leukotrienes LTC 4 , LTD 4 , and LTE 4 has been shown in cells that do not possess 5-LO activity, such as red blood cells (4), platelets (5), endothelial cells (6,7), and smooth muscle cells (8).
Polymorphonuclear leukocytes (PMNL) possess relatively large amounts of 5-lipoxygenase, the enzyme catalyzing the sequential conversion of arachidonic acid to 5-hydroperoxyeicosatetraenoic acid (5-HPETE) and LTA 4 (9). Upon cell activation, significant amounts of LTB 4 and its -oxidized metabolites 20-hydroxy-and 20-carboxy-LTB 4 are released into the extracellular milieu together with nonenzymatic breakdown products of LTA 4 , namely ⌬ 6 -trans-LTB 4 isomers and 5,6-dihydroxyeicosatetraenoic acids (5,6-diHETEs) (10 -12). Recent studies in complex organ systems (13)(14)(15)(16)(17)(18) showed that perfusion of PMNL in the isolated lung or heart of the rabbit only caused a significant increase in the production of cysteinyl leukotrienes when PMNL were activated during the perfusion process. These data suggest that transcellular biosynthesis of cysteinyl leukotrienes might indeed be of physiopathological relevance when tight cell-cell interactions occur, such as during adhesion and diapedesis of PMNL through the microvascular endothelium of a functioning organ system.
In light of these observations it was of interest to assess the relative amount of LTA 4 released from PMNL and therefore available for transcellular biosynthesis of cysteinyl leukotrienes, with respect to total LTA 4 synthesized. The release of LTA 4 into the extracellular milieu would remove this intermediate from intracellular LTA 4 hydrolase that catalyzes conversion of LTA 4 into LTB 4 . Intracellular LTB 4 can be further metabolized by a specific cytochrome P-450 to 20-hydroxy-LTB 4 and 20-carboxy-LTB 4 (19). The extracellular (released) LTA 4 will react with water with a half-life lower than 30 s (20) to yield the nonenzymatic products, ⌬ 6 -trans-LTB 4 , ⌬ 6 -trans-12-epi-LTB 4 , and 5,6-dihydroxyeicosatetraenoic acid isomers. But PMNL are able to take up exogenously added LTA 4 (21) and metabolize it into LTB 4 , thus reducing the fraction of released LTA 4 that is actually available for transcellular metabolism (or nonenzymatic hydrolysis).
In the present study experiments were designed to test the effect of dilution on the quantitative profile of LTA 4 metabolites produced after challenge with the Ca 2ϩ ionophore A23187. The hypothesis that in diluted cell preparations LTA 4 would have less chance of being reabsorbed and metabolized by vicinal PMNL has been tested. In a previous study, Cluzel et al. (22) showed that the use of diluted cell suspensions provided important information concerning the amount of platelet activating factor and LTB 4 released by PMNL. Using a similar approach, we provide evidence that significant transcellular metabolism of LTA 4 does indeed take place in purified human PMNL preparations.

EXPERIMENTAL PROCEDURES
Chemicals and Reagents-All chemicals used were reagent-grade and obtained from commercial sources. Eicosanoids were purchased from Cayman Chemical Co. (Ann Arbor, MI). 12-O-Methyl,all-trans-LTB 4 derivatives were prepared by reacting LTA 4 (20 nmol) with methanol (1 ml) acidified with HCl and purified by RP-HPLC. HPLC-grade solvents were obtained from Merck (Darmstadt). Type I "plus" water was obtained using a MilliQ Plus water purifier (Millipore, Molsheim), fed with double distilled water.
Preparation of Human Polymorphonuclear Leukocytes-Human polymorphonuclear leukocytes (PMNL) were obtained from blood withdrawn from healthy donors that had not taken medications for at least 1 week; blood was collected into a 50-ml polypropylene centrifuge tube containing 5.7 ml of ACD (41 mM citric acid ϫ H 2 O, 100 mM sodium citrate ϫ 2H 2 O, 136 mM glucose) and carefully mixed. After centrifugation for 15 min at room temperature (RT) and 280 ϫ g, platelet-rich plasma was removed, and residual blood was combined with an equal volume of saline and 0.5 volume of dextran T-500 (6%, w/v, in saline), followed by thorough mixing, and allowed to stand at RT for 30 min. The leukocyte-enriched upper phase was centrifuged for 15 min at RT and 280 ϫ g. The pelleted cells were then subjected to erythrocyte lysis by gentle resuspension in 1 volume of a 0.2%, w/v, NaCl solution and further dilution with 1 volume of a solution of the following composition, 3.98 g of NaCl ϩ 0.5 g of sucrose in 250 ml of distilled water at ϩ4°C. Mononuclear cells were separated by centrifugation on Ficoll cushions (density 1.077 g/ml) for 30 min at RT and 400 ϫ g. The pellet was then resuspended and the obtained PMNL washed twice with 10 -15 ml of phosphate-buffered saline without Ca 2ϩ and Mg 2ϩ (PBS 2Ϫ ). Cells were finally resuspended at a final concentration of approximately 20 -30 ϫ 10 6 cells ml Ϫ1 in PBS 2Ϫ and kept on ice until used. This preparation contained more than 95% PMNL as assessed by differential count on May-Grunwald/Giemsa-stained cytocentrifugates.
Cell Incubations-Challenge of PMNL samples at different concentrations was carried out with the calcium ionophore A23187 (Calbiochem, 5 M) for 2 or 10 min at 37°C, after addition of Ca 2ϩ (2 mM) and Mg 2ϩ (0.5 mM) and 5 min of thermal equilibration. In selected experiments, human serum albumin (Sigma) was added to a final concentration of 10 mg ml Ϫ1 ; in order to achieve 5-LO activation, concentration of A23187 was raised to 50 M. LTA 4 free acid was obtained through base hydrolysis of LTA 4 methyl ester using ice-cold acetone/NaOH 0.25 M (4:1, v/v) at room temperature for 60 min. LTA 4 free acid was added either to increasing amounts of human PMNL (1-20 ϫ 10 6 ) at a final concentration of 0.4 M or to a fixed amount of 20 ϫ 10 6 PMNL ml Ϫ1 in increasing concentrations (0.1-10 M). Metabolism of exogenous LTA 4 was allowed to proceed for 10 min at 37°C.
Incubations were terminated with 2 volumes of ice-cold methanol containing the HPLC internal standard PGB 2 (25 ng) and samples analyzed by RP-HPLC.
RP-HPLC Analysis-Samples were diluted with water to a final methanol concentration lower than 20%, and extraction was quickly carried out using a solid phase cartridge (Supelclean LC-18, Supelco, Bellafonte, PA). The retained material was eluted using 90% aqueous MeOH. After evaporation, the dried extract was reconstituted in 600 l of HPLC mobile phase A (methanol/acetonitrile/water/acetic acid, 10: 10:80:0.02, v/v/v/v, pH 5.5, with ammonium hydroxide) and injected into an HPLC gradient pump system (model 126, Beckman Analytical, Palo Alto, CA) connected to a diode array UV detector (model 168, Beckman Analytical) using a microprocessor-controlled autosampler (Jasco 851-AS, Tokio, J), with sample kept at 4°C. UV absorbance was monitored at 280 nm, and full UV spectra (210 -340 nm) were acquired at a rate of 0.5 Hz.
Positive identification of enzymatic and nonenzymatic LTA 4 metabolites was obtained through UV spectral analysis of chromatographic peaks eluting at characteristic retention times. Quantitation was carried out on positively identified peaks only, using their HPLC peak areas relative to that of PGB 2 at 280 nm and calculated from the responses of standard compounds. The ratio (enzymatic-LTA 4 metabo-lites)/(nonenzymatic-LTA 4 metabolites) was calculated from the HPLC data. Enzymatic-LTA 4 metabolites was used as a collective name for LTC 4 , LTB 4 , 20-hydroxy-LTB 4 , and 20-carboxy-LTB 4 ; nonenzymatic-LTA 4 metabolites was used as a collective name for ⌬ 6 -trans-LTB 4 isomers, 5,6-dihydroxyeicosatetraenoic acids, and 12-O-methyl-⌬ 6trans-LTB 4 isomers (12).
Normalized data were obtained expressing as 100% the total amount of LTA 4 -derived metabolites observed in a given sample.
Data Analysis-Comparison of enzymatic-and nonenzymatic-LTA 4 metabolites in different cell concentration groups was carried out by analysis of variance and post hoc analysis performed with the Hsu MCB test to assess whether means were lower than the unknown maximum (or greater than the unknown minimum). Comparison of LTA 4 metabolites at 2 and 10 min were carried out by Student's t test.
Analysis of variance and regression analysis were used to examine the relationship between the cell concentration and different parameters studied. Values were expressed as mean Ϯ standard error of the mean (S.E.) of 3-5 different PMNL preparations. A value of p Ͻ 0.05 was considered to be of statistical significance.

RESULTS
Human Polymorphonuclear Leukocyte Cell Incubations-Administration of synthetic LTA 4 free acid to human PMNL resulted in a cell number-dependent formation of enzymatic-LTA 4 metabolites (data not shown).
Decreasing the concentration of PMNL caused a marked increase of LTB 4 with respect to 20-OH-and 20-COOH-LTB 4 ( Figs. 1 and 4), in agreement with previous data (22,26) and suggesting that -oxidation of LTB 4 is mainly carried out after reuptake of released LTB 4 .
Reducing the incubation time for A23187 challenge from 10 to 2 min resulted in the appearance of methyl trapping metab-olites of LTA 4 (10,27), indicating the presence of intact LTA 4 (4.5 Ϯ 1.1% of total LTA 4 metabolites in diluted and 11.3 Ϯ 0.6% in concentrated PMNL preparations) at this shorter incubation time.
The relative amounts of nonenzymatic-LTA 4 metabolites represented 41.1 Ϯ 1.6 and 50.3 Ϯ 1.5% at 10 and 2 min, respectively, in diluted PMNL incubations (p Ͻ 0.01). A similar shift toward nonenzymatic-LTA 4 metabolites was observed in concentrated PMNL incubations (Table I), where nonenzymatic metabolites represented 23.1 Ϯ 1.3% at 10 min and 37.7 Ϯ 1.4% at 2 min after challenge (p Ͻ 0.001). Total amounts of LTA 4 metabolites per million cells were also significantly lower at the shorter incubation time studied (Table I).
Human serum albumin, at a final concentration of 10 mg ml Ϫ1 , totally inhibited the production of 5-LO metabolites after challenge with A23187 5 M for 2 min (22), possibly due to binding to albumin itself (28). Increasing the concentration of A23187 to 50 M restored a well detectable production of LTA 4derived metabolites (Table I). Interestingly, the quota of nonenzymatic metabolites was significantly decreased in diluted cell incubations (43.4 Ϯ 1.8 versus 50.2 Ϯ 1.5%, p Ͻ 0.05 versus without albumin) but increased in concentrated PMNL preparations (43.5 Ϯ 2 versus 37.7 Ϯ 1.4%, p Ͻ 0.05 versus without albumin) if compared with samples without albumin. In agreement with the stabilizing effect of albumin, intact LTA 4 , represented by 12-O-methyl-all-trans-LTB 4 -derivatives, was 18.4 Ϯ 0.8 and 26.5 Ϯ 1.8% of total LTA 4 -derived metabolites, in diluted and concentrated PMNL preparation, respectively. DISCUSSION Over the last 5 years an increasing body of evidence has indicated the importance of transcellular metabolism of leukotriene A 4 in complex organ systems. Grimminger and co-workers (13)(14)(15) showed that perfusion and activation of PMNL in the isolated lung of the rabbit resulted in the production of significantly increased amounts of cysteinyl leukotrienes, with respect to activation of PMNL alone. Similar results were obtained in our laboratory using a PMNL-perfused isolated rabbit heart (16,18). These reports have raised the issue of determining how much of the LTA 4 synthesized by PMNL can be made available to adhering cells (namely endothelial or smooth muscle cells). The overall potential for LTA 4 transfer from PMNL (donor cell) to acceptor cells (29) has usually been quantitated by evaluating the production of nonenzymatic-LTA 4 metabolites in purified PMNL incubations (13,14,16). Nevertheless, given the capacity of PMNL to actively take up LTA 4 from the extracellular milieu and convert it to LTB 4 (21), such a calculation would still underestimate the amount of LTA 4 provided by PMNL to neighboring cells. The evidence presented in this report clearly demonstrates that LTA 4 represents the major metabolite released from PMNL to the extracellular milieu. Diluted cell suspensions have been used as an approach to limit the quota of LTA 4 that, once released into the extracellular milieu, is available for further enzymatic metabolism by neighboring PMNL. Challenge of sequentially diluted PMNL preparations with the calcium ionophore A23187 resulted in increased amounts of nonenzymatic-LTA 4 metabolites (namely ⌬ 6 -trans-LTB 4 isomers and 5,6-diHETEs) with respect to enzymatic-LTA 4 metabolites (LTB 4 and its -oxidized metabolites). 5-Keto-(7E,9E,11Z,14Z)-eicosatetraenoic acid (25), a reported nonenzymatic-LTA 4 metabolite, was not observed in A23187challenged PMNL preparations, whereas it was present in detectable amounts when synthetic LTA 4 was used at concentrations higher than 1 M. The decrease in the ratio of (enzymatic metabolites)/(nonenzymatic metabolites) was well correlated with the decreased concentration of PMNL, although regression analysis suggested that a progressive saturation of enzymatic metabolism was observed at the higher PMNL concentrations used. Administration of increasing concentrations of synthetic LTA 4 to PMNL at a concentration of 20 ϫ 10 6 ml Ϫ1 supported this hypothesis, in agreement with previous data indicating that the LTA 4 hydrolase, and not 5-lipoxygenase, is the limiting factor in the synthesis of LTB 4 in human leukocytes (30). Concentrations of LTA 4 upon challenge with A23187, as estimated from the total amount of LTA 4 -derived metabolites observed, resulted in approximately 1.13, 2.3, and 5 M in PMNL preparations at 5, 10, and 20 ϫ 10 6 PMNL ml Ϫ1 , respectively. These LTA 4 concentrations are very compatible with saturation of the enzymatic metabolite formation observed with exogenous LTA 4 .
The O-methyl trapping products of LTA 4 were observed in incubations terminated 2 min after A23187 challenge, indicating the presence of intact LTA 4 even after this time period. The total amounts of LTA 4 -derived metabolites was significantly lower if compared with amounts observed after 10 min, both in diluted and in concentrated PMNL preparations. However, shortening the incubation time after challenge led to a significant shift toward nonenzymatic-LTA 4 metabolites at both concentrations studied. This would be consistent with a time-dependent reuptake of LTA 4 into the PMNL and subsequent conversion into LTB 4 . It is known that albumin is able to stabilize LTA 4 , increasing its half-life at physiological pH from a few seconds to over 20 min (20). The effect of human serum albumin, at a concentration of 10 mg ml Ϫ1 , was studied in diluted and concentrated PMNL preparations, after challenge with A23187 for 2 min. The results obtained showed that in diluted cell preparations, stabilization of LTA 4 by albumin was able to partially revert the effect of dilution, allowing intact LTA 4 to travel to distant PMNL and be enzymatically transformed into LTB 4 . On the other end, in the presence of higher concentrations of LTA 4 , such as in concentrated cell preparations, a favorable competition by the bound LTA 4 versus that LTA 4 which can be taken up from the albumin complex by the human PMNL exists, resulting in the trapping of intact LTA 4 in the extracellular milieu.
In addition to affecting LTA 4 metabolism, dilution of PMNL preparations influenced the amounts of LTB 4 relative to the -oxidized metabolites 20-hydroxy-LTB 4 and 20-carboxy-LTB 4 observed upon A23187 challenge. Decreasing the PMNL concentration, linked with the increase in nonenzymatic-LTA 4 metabolites, also resulted in a 4-fold increase in LTB 4 , with a complementary decrease in 20-COOH-and 20-OH-LTB 4 . These results, in agreement with previous studies (22,26), indicate that LTB 4 synthesized in purified PMNL preparations is first released and then -oxidative metabolism occurs after reuptake by the cells.
The data presented indicate that the amount of nonenzymatic-LTA 4 metabolites observed in PMNL challenged by calcium ionophore A23187 in vitro, at the commonly used concentrations of 10 -20 ϫ 10 6 cell/ml, results in a substantial underestimation of the fraction of LTA 4 that is indeed available for transcellular metabolism. In the past, the fact that PMNL themselves act as acceptor cells for released LTA 4 , as well as cells that convert released LTB 4 to -oxidized metabolites, was evidently overlooked (Fig. 5). The use of diluted human PMNL preparations permitted the estimate that over 50% of the LTA 4 synthesized through activation of 5-lipoxygenase by the use of the calcium ionophore A23187 is actually released into the extracellular milieu. Preliminary data, using a different ap-  proach, suggested that the fraction of LTA 4 secreted by PMNL could represent up to 80% of the total (31).
At variance with what is generally accepted, the data reported here indicate that the majority of LTA 4 is released before being metabolized to LTB 4 and is therefore available for transcellular biosynthesis to cysteinyl leukotrienes by proximal cells. Inhibition of LTC 4 formation arising from the interaction of human PMNL and glomerular endothelial cells, by antibodies against CD18 and L-selectin, has recently been reported (32). Adhesion of PMNL to potential LTA 4 acceptor cells would therefore appear to be a key step toward an efficient transfer of LTA 4 resulting in the formation of cysteinyl leukotrienes. Close interaction may cause direct transfer of LTA 4 from the PMNL to the LTC 4 synthase carrying endothelial cells, resulting in substantial changes in the metabolic profile of 5-lipoxygenase-derived products. It has been shown that LTB 4 may enhance its own biosynthesis in an autocrine fashion (33); similarly, cysteinyl leukotrienes may amplify their own biosynthetic mechanisms, inducing endothelial cell-dependent neutrophil adhesion and subsequent transcellular metabolism of LTC 4 (34). Metabolism of LTA 4 to cysteinyl leukotrienes within the microvasculature may have considerable pathophysiological consequences, in light of the ability of LTC 4 and LTD 4 to induce profound modification of vascular permeability leading to edema formation (35).
The data presented in this work indicate that LTA 4 , the main 5-LO-derived metabolite released by PMNL, can therefore be considered as a lipid mediator itself, not as much for its intrinsic biological activity as for its ability to promote the production of bioactive compounds in cells other than those in which it is synthesized.