Human Group V Phospholipase A2 Induces Group IVA Phospholipase A2-independent Cysteinyl Leukotriene Synthesis in Human Eosinophils*

We previously reported that exogenously added human group V phospholipase A2 (hVPLA2) could elicit leukotriene B4 biosynthesis in human neutrophils through the activation of group IVA phospholipase A2 (cPLA2) (Kim, Y. J., Kim, K. P., Han, S. K., Munoz, N. M., Zhu, X., Sano, H., Leff, A. R., and Cho, W. (2002) J. Biol. Chem. 277, 36479-36488). In this study, we determined the functional significance and mechanism of the exogenous hVPLA2-induced arachidonic acid (AA) release and leukotriene C4 (LTC4) synthesis in isolated human peripheral blood eosinophils. As low a concentration as 10 nm exogenous hVPLA2 was able to elicit the significant release of AA and LTC4 from unstimulated eosinophils, which depended on its ability to act on phosphatidylcholine membranes. hVPLA2 also augmented the release of AA and LTC4 from eosinophils activated with formyl-Met-Leu-Phe + cytochalasin B. A cellular fluorescent PLA2 assay showed that hVPLA2 had a lipolytic action first on the outer plasma membrane and then on the perinuclear region. hVPLA2 also caused the translocation of 5-lipoxygenase from the cytosol to the nuclear membrane and a 2-fold increase in 5-lipoxygenase activity. However, hVPLA2 induced neither the increase in intracellular calcium concentration nor cPLA2 phosphorylation; consequently, cPLA2 activity was not affected by hVPLA2. Pharmacological inhibition of cPLA2 and the hVPLA2-induced activation of eosinophils derived from the cPLA2-deficient mouse corroborated that hVPLA2 mediates the release of AA and leukotriene in a cPLA2-independent manner. As such, this study represents a unique example in which a secretory phospholipase induces the eicosanoid formation in inflammatory cells, completely independent of cPLA2 activation.

Phospholipase A 2 (PLA 2 ) 1 catalyzes the hydrolysis of the sn-2 ester bond of membrane phospholipids, the products of which can be transformed into potent inflammatory lipid mediators, including eicosanoids (i.e. prostaglandins, leukotrienes, and thromboxanes) and platelet-activating factor. Multiple forms of PLA 2 s, including a number of secretory PLA 2 s (sPLA 2 ) (1) and several intracellular enzymes (2)(3)(4), have been identified from mammalian tissues. The involvement of group IVA cytosolic phospholipase A 2 (cPLA 2 ) in cellular eicosanoid formation has been well documented (2,5,6). However, physiological roles of most of sPLA 2 isoforms have not been fully defined. Accumulating evidence has indicated that some sPLA 2 s work in concert with cPLA 2 to induce eicosanoid formation in different mammalian cells (7)(8)(9)(10)(11)(12). For example, it was recently reported that exogenously added human group V PLA 2 (hVPLA 2 ) could elicit the release of arachidonic acid (AA) and leukotriene B 4 (LTB 4 ) from human neutrophils by activating cPLA 2 via an increase in cellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) and cPLA 2 phosphorylation (13,14). As part of our continuing effort to understand the physiological functions and regulation of sPLA 2 s, we investigated the effects of exogenously added hVPLA 2 on the activation of human eosinophils.
Expression and Purification of sPLA 2 -Recombinant human group IIa PLA 2 (hIIaPLA 2 ) was prepared as described (19). Recombinant hVPLA 2 and mutants were expressed in Escherichia coli, refolded, and purified as described previously (20,21). The purity of enzymes assessed by sodium dodecyl sulfate-polyacrylamide electrophoresis was consistently higher than 90%.
Isolation of Human Peripheral Blood Eosinophils-Mildly atopic, nonsmoking donors were recruited for eosinophil donation. Atopy was defined by criteria used in the University of Chicago Asthma Research Center for the National Heart, Lung, and Blood Institute Human Cooperative Asthma Genetics Project (22). Donors demonstrating Ͼ1.5% peripheral blood eosinophils (PBE) were used in this study. PBE were isolated by the negative immunomagnetic selection technique as previously described (15,18). The purity of cells was assessed by differential cell counts from a Wright-Giemsa air-dried smear, and viability of cells was confirmed by trypan blue exclusion analysis.
Isolation of Eosinophils from Murine Bone Marrow-Bone marrow cells from femurs of mice were cultured in RPMI medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, 0.1 mg/ml streptomycin, 10 ng/ml interleukin-5, and 5 ng/ml granulocyte macrophagecolony stimulating factor. Refeeding was performed by the addition of the same medium in the absence of granulocyte macrophage colonystimulating factor at days 3, 6, and 8. Matured eosinophils were harvested at day 13, and cultured cells were activated with buffer control, 100 nM N-formyl-Met-Leu-Phe (fMLP) plus 5 g/ml cytochalasin B (CB), or 100 nM hVPLA 2 .
Measurement of sPLA 2 by Sandwich Enzyme-linked Immunosorbent Assay-Purified hVPLA 2 was used as the standard protein. Microplate wells were coated with equivalent concentrations of MCL-1B7 and MCL-2A5 (2.5 g/ml each) (23,24) as capture antibodies in 50 nM carbonate buffer, pH 9.6, overnight at 4°C. The samples were added to the microplates and incubated overnight at 4°C. After washing with the phosphate-buffered saline (PBS), pH 7.4, 0.5 g/ml of biotinylated MCL-3G1 (23,24) was added to the wells and further incubated for 60 min. After washing with deionized water and three times with PBS with 0.5% Triton X-100, extravidin was added onto the wells and incubated for additional 60 min. Again, the treated wells were washed with PBS with 0.5% Tween 20 followed by the addition of 50 l of p-nitrophenylphosphate as substrate. Absorbance at 405 nm was measured using a Thermomax microplate spectrophotometer (Molecular Devices, Menlo Park, CA). Protein quantitation was performed as described (23). All of the assays were performed in duplicate, and the amounts of protein are expressed in pg/10 6 cells.
Measurement of AA, LTB 4 , and LTC 4 Release-Isolated PBE (10 6 cells/group) were labeled with 0.5 Ci of [ 3 H]AA overnight at 37°C. The unincorporated AA was washed three times with the Hanks' balanced salt solution (HBSS) containing 1.2 mM CaCl 2 and 0.2% bovine serum albumin. Radiolabeled cells (10 6 ) were resuspended in 90 l of the same buffer, preincubated with a selected inhibitor for 30 min at 37°C if necessary (and washed three times with the same buffer to remove the residual inhibitor), and then stimulated with hVPLA 2 , hVPLA 2 /W31A, hVPLA 2 /R100E/K101E, or hIIaPLA 2 . As a positive control, PBE were also activated with 100 nM fMLP and 5 g/ml CB. The reaction was quenched by centrifugation, and the radioactivity in the cell pellet and the supernatant was separately measured by a two-channel liquid scintillation counter. LTB 4 and LTC 4 levels (pg/10 6 cells) in the supernatant were determined using enzyme-linked immunoassay kits from Cayman Chemical Company (Ann Arbor, MI).
Measurement of Phosphorylation of ERK1/2 MAP Kinase and cPLA 2 by Immunoblotting Analysis-To determine whether the hVPLA 2 -induced LTC 4 secretion is mediated via ERK1/2 MAP kinase and subsequent cPLA 2 phosphorylation, PBE were stimulated with 100 nM hVPLA 2 for different periods, and the phosphorylation was monitored as follows. Treated cells were lysed in a cell disruption buffer (20 mM Tris-HCl, 30 mM Na 4 P 2 O 7 , 50 mM NaF, 40 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, protease inhibitors tablet) and centrifuged at 400 ϫ g for 10 min to remove the nuclear and cellular debris (15). The prepared samples were loaded onto a 10% acrylamide gel for ERK1/2 MAP kinase and a 7.5% gel for cPLA 2 , and the electrophoresis was run under reducing conditions as described previously (15). The Western blot was probed with the antibodies specific for ERK1/2, phosphorylated ERK1/2, and Ser 505 -phosphorylated cPLA 2 , respectively, and visualized using an enhanced chemiluminescence system (Amersham Biosciences).
Measurement of cPLA 2 Activity-PBE (2 ϫ 10 6 cells/group) were incubated with the HBSS control, 100 nM hVPLA 2 , or 100 nM fMLP with 5 g/ml CB for 20 min at 37°C. The stimulation was quenched by adding 1 ml of ice-cold water, and the cell mixture was centrifuged. The pellet was resuspended in 70 l of lysis buffer (20 mM Tris-HCl, pH 8.0, containing 2.5 mM EDTA, 10 g/ml leupeptin, 5 g/ml aprotinin, 1 mM of phenylmethylsulfonyl fluoride, 2 mM of Na 3 VO 4 , 50 mM NaF, and 5 g/ml pepstatin) and sonicated briefly. The resulting cell lysate was pretreated with 10 l of dithiothreitol (final concentration, 10 mM) on ice for 5 min to inactivate sPLA 2 , and 10 l of 1 mM CaCl 2 (final concentration, 0.1 mM) was then added to each sample. The cPLA 2 substrate solution was prepared by drying chloroform solution of [ 14 C]1strearoyl-2-arachidonoyl-sn-glycero-3-phosphocholine solution under a stream of N 2 and suspending the film in 100 l of 10% aqueous ethanol by vortexing. The reaction was initiated by adding a 10-l portion of the substrate solution (final concentration, 9 M) to each cell lysate. The reaction was carried out for 30 min at 37°C and was quenched by adding 560 l of Dole's reagent (heptane, 2-propanol, and 1 N H 2 SO 4 (400:390:10, v/v/v)), followed by the addition of 110 l of H 2 O, and the mixture was vortexed for 20 s and then centrifuged at 13,000 ϫ g. The 180 l of upper layer was transferred to 800 l of hexane mixed with 25 mg of silica gel. The mixture was vortexed and centrifuged, 800 l of the supernatant was mixed with 2 ml of scintillation fluids, and then the radioactivity was counted in a liquid scintillation counter. The cPLA 2 activity was expressed in terms of pmol AA/min/10 6 cells.
Measurement of 5-LO Activity-The 5-LO activity in PBE lysate (from 2 ϫ 10 6 cells; see above) was measured as described previously (25,26). The relative 5-LO activity was then expressed in terms of the percentage of increase over the control (unstimulated) value.

Measurement of [Ca 2ϩ
] i was performed with a Zeiss LSM 510 laser scanning confocal microscope using Fluo-4 acetoxymethyl ester as indicator. PBE (10 7 cells/ml) were incubated in HBSS containing 1.2 mM Ca 2ϩ , 1% bovine serum albumin, and 2 M Fluo-4 for 30 min at 37°C. Labeled cells were seeded into each of eight wells on a sterile Nunc Lak-TeKII TM chambered cover glass filled with 400 l of HBSS containing 1.2 mM Ca 2ϩ and incubated at 37°C with 5% CO 2 for 10 min. After washing once with HBSS containing 1.2 mM Ca 2ϩ , 100 nM hV-PLA 2 (or 10 M ionomycin as a positive control and the buffer as a negative control) was added, and the fluorescence intensity of Fluo-4 was monitored with a 488-nm Argon/Krypton laser and a 530-nm line pass filter. A 63ϫ (1.2 numerical aperture) water immersion objective was used for all of the experiments. The images were analyzed using the analysis tools provided in the Zeiss biophysical software package.
In Vivo Assay of hVPLA 2 Activity-The cellular PLA 2 assay was performed as described previously (28). Briefly, PBE (10 6 cells/ml) were seeded into each of eight wells on a sterile Nunc Lak-TeKII TM chambered cover glass filled with 400 l of HBSS and incubated at 37°C with 5% CO 2 for 10 min. After the cells were washed once with HBSS, they were overlaid with 10 l of PED6 vesicle solution (0.75 mM 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoserine/cholesterol/1-palmitoyl-2-oleoylsn-glycero-3-phosphoglycerol/PED6 (107:31:20:1 in mole ratio) mixed vesicles in HBSS) and incubated for 50 min at 37°C with 5% CO 2 . After rinsing the labeled cells six times with HBSS containing 1.2 mM Ca 2ϩ , hVPLA 2 was added to cells. Imaging was done with a Zeiss LSM510 laser scanning confocal microscope with the detector gain adjusted to eliminate the background autofluorescence. The BODIPY TM signal from the hydrolyzed PED6 was visualized with a 488-nm argon/krypton laser and a 530-nm band pass filter. A 63ϫ (1.2 numerical aperture) water immersion objective was used for all of the experiments. The images were analyzed using the analysis tools provided in the Zeiss biophysical software package.
Immunocytostaining-Cytoslides of treated cells were prepared and fixed with 2% paraformaldehyde for 30 min in PBS. After three washes with PBS, the fixed cells were blocked sequentially with 2% bovine serum albumin and permeabilized with 1% saponin in PBS for 1 h. For 5-LO, the 5-LO antibody was added to samples, incubated for 60 min at 37°C, and then further incubated with fluorescein goat anti-rabbit immunoglobin G for another 60 min. After six washes with PBS, the treated cells were covered with glass slips using Perma Fluor, and the 5-LO signal was visualized using a Zeiss LSM510 laser scanning confocal microscopy. For cPLA 2 , the prepared slides were stained with the cPLA 2 -specific antibody and processed using the Vecstatin ABC kit (VECTOR Laboratories, Burlingame, CA). The color reaction was allowed to proceed for 30 min at room temperature. The slides were examined by light microscopy to determine the localization of cPLA 2 .
Data Analysis-The Data are expressed as the means Ϯ S.E. for each group. Individual statistical comparisons of paired data were assessed by Student's t test, and p Ͻ 0.05 was considered to be statistically significant. Where multiple comparisons were made, differences among the populations were evaluated by analysis of variance followed by Bonferroni correction. 4 by Exogenous hVPLA 2 in Eosinophils-It has been shown that human eosinophils contain cPLA 2 (29) and hIIaPLA 2 (30). It is not known, however, whether other forms of sPLA 2 s are present in eosinophils. We thus measured the cellular content of endogenous hVPLA 2 in PBE activated with fMLP (100 nM)/CB (5 g/ml) by the sandwich enzyme-linked immunosorbent assay for hVPLA 2 (23). hVPLA 2 was detected neither in resting PBE nor in fMLP/CBactivated PBE for 2 h after stimulation. The lack of hVPLA 2 was not due to proteolytic degradation of hVPLA 2 , because the amount of exogenous hVPLA 2 added to the cell lysates of eosinophils (data not shown) and other cells (23) could be accurately determined. Estimated from the sensitivity limit of our sandwich enzyme-linked immunosorbent assay (23), hVPLA 2 should exist in less than 2 ng/10 6 PBE cells.

Release of AA and LTC
Because PBEs lack endogenous hVPLA 2 , we measured the effect of exogenously added hVPLA 2 on PBE. First, we measured the release of [ 3 H]AA from intact PBE treated with hVPLA 2 . The exogenous hVPLA 2 caused time-dependent (Fig.  1A) and concentration-dependent (Fig. 1B) release of AA. The AA release caused by 100 nM hVPLA 2 was twice greater than the background (p Ͻ 0.02) at 15 min and gradually decreased (Fig. 1A). The AA release was significantly greater than the background with Ն10 nM hVPLA 2 (Fig. 1B). This AA releasing activity was dependent on the activity of hVPLA 2 on phosphatidylcholine (PC) membranes, because hIIaPLA 2 (20) and the W31A mutant of hVPLA 2 (21), which have much lower activity on and lower affinity for PC vesicles, respectively, did not significantly induce the AA release. Furthermore, the R100E/ K101E mutant of hVPLA 2 , which has wild type-like activity on PC membranes but has much reduced affinity for cell surface heparan sulfate proteoglycans (HSPG) (13), caused considerably less AA release than the wild type (ϳ60%) under the same conditions. This suggests that both activity toward PC and affinity for HSPG are important for the AA releasing activity of hVPLA 2 on unstimulated PBE. As was the case with human neutrophils (13) and HEK 293 cells (28), HSPG binding of hVPLA 2 was required for its internalization to PBE, because the immunoblotting analysis of the cell lysate showed that treatment of PBE with 1 unit/ml of heparinase I for 30 min at 37°C abrogated the internalization of hVPLA 2 (data not shown). This, in conjunction with the reduced AA release by R100E/K101E, implies that a significant portion of AA released by the wild type hVPLA 2 derives from intracellular membranes.
We then measured the effect of exogenous hVPLA 2 on the LTC 4 release from intact PBE. As was the case with the AA release, hVPLA 2 induced the LTC 4 release from PBE in a concentration-dependent manner (Fig. 2). The LTC 4 release caused by 100 nM hVPLA 2 was comparable with that caused by 100 nM fMLP with 5 g/ml CB. Again, hIIaPLA 2 and hVPLA 2 / W31A (data not shown) did not significantly induce LTC 4 release. Interestingly, hVPLA 2 /R100E/K101E produced LTC 4 at a basal level, suggesting that AA released at the outer plasma membrane by this enzyme is not conducive to LTC 4 synthesis. We also measured the effect of exogenous hVPLA 2 on PBE activated with 100 nM fMLP with 5 g/ml CB. Incubation of fMLP/CB-treated PBE with 100 nM hVPLA 2 greatly augmented the release of AA (Fig. 3A) and LTC 4 (Fig. 3B) from these cells.
To determine whether cPLA 2 is involved in the hVPLA 2induced AA and LTC 4 release, we treated PBE with AACOCF 3 , which inhibits both cPLA 2 and group VI calcium-independent PLA 2 (31), prior to the addition of hVPLA 2 . AACOCF 3 up to 30 M had a negligible effect on the release of AA (Fig. 4A) and LTC 4 (Fig. 4B) from intact PBE. By contrast, pretreatment of PBE with 10 M of a sPLA 2 inhibitor, LY311727 (15), attenuated the release of AA and LTC 4 to a nearly base-line level. Taken together, these results indicate that hVPLA 2 can induce the release of AA and LTC 4 from both intact and activated PBE and that this activity depends not on the activities of cPLA 2 but on the catalytic activity of sPLA 2 , most likely exogenously added hVPLA 2 .
Effects of hVPLA 2 on cPLA 2 and 5-LO Activities-To understand how hVPLA 2 stimulates PBE without activating cPLA 2 , we measured the effect of exogenously added hVPLA 2 on [Ca 2ϩ ] i , cPLA 2 activity, cPLA 2 phosphorylation, and cPLA 2 translocation. We previously reported that the lipolytic action of hVPLA 2 on the outer plasma membrane of human neutrophils resulted in an increase in [Ca 2ϩ ] i and cPLA 2 phosphorylation, both of which led to cPLA 2 activation (14). Fig. 5A illustrates the time lapse change in [Ca 2ϩ ] i after PBE cells were treated with hVPLA 2 . Evidently, 100 nM hVPLA 2 did not induce the rise in [Ca 2ϩ ] i for 10 min until 10 M ionomycin was added as a positive control. In fact, no increase in [Ca 2ϩ ] i was seen for more than a hour with 100 nM hVPLA 2 (data not shown). hVPLA 2 did not cause the phosphorylation of either cPLA 2 (Fig. 5C) or ERK1/2 MAP kinase (Fig. 5B), which has been shown to be involved in cPLA 2 phosphorylation. Accordingly, the in vitro cPLA 2 assay of the cell lysates revealed no increase in cPLA 2 activity after the hVPLA 2 treatment, whereas treatment with 100 nM fMLP and 5 g/ml CB caused a 3-fold increase in cPLA 2 activity (Fig. 6A). Furthermore, the immunocytostaining of cPLA 2 in PBE showed that cPLA 2 was distributed in the cytoplasm before and after the hVPLA 2 treatment, whereas it was localized in the perinuclear region after fMLP/CB treatment (Fig. 6B).
We also measured the effect of exogenous hVPLA 2 on 5-LO activity and translocation. First, we measured the 5-LO activity of the cell lysates after cells were treated with hVPLA 2 or fMLP/CB. Fig. 6C shows that 100 nM hVPLA 2 caused a 2-fold increase in 5-LO activity over the buffer control, and this activation was even greater than the effect by 100 nM fMLP and 5 g/ml CB. Because this in vitro assay was performed in the buffer solution containing all activators of 5-LO, including ATP and Ca 2ϩ , the activation should be due to an enzyme modification, most likely 5-LO phosphorylation. However, an attempt to detect the 5-LO phosphorylation by a gel shift assay was not successful, presumably because of a small difference in the electrophoretic mobility between the phosphorylated and nonphosphorylated forms. We then determined the localization of 5-LO before and after treatment with 100 nM hVPLA 2 by immunocytostaining. As shown in Fig. 6D, 5-LO was found predominantly in the cytosol in unstimulated cells, but it translocated to the nuclear membrane and the cytoplasmic granules upon exposure to hVPLA 2 . This translocation pattern was similar to that seen with cells treated with a calcium ionophore A23187 (1 M). As described above, however, hVPLA 2 -induced membrane translocation of 5-LO was not attributed to the increase in [Ca 2ϩ ] i , because the preincubation of cells with 0.1 mM EDTA in the growth medium before hVPLA 2 treatment did not affect the 5-LO translocation (data not shown).
The Site of hVPLA 2 Action in Eosinophils-We previously reported an in vivo assay that allows the real time monitoring of the PLA 2 activity in the living cell using a fluorescent phospholipid PED6 (28). In this assay, the release of the fatty acid from PED6 gives an enhanced fluorescent signal caused by the relief of fluorescence quenching. To determine the sites of lipolytic action of hVPLA 2 , we labeled all membranes of PBE with PED6 and performed time lapse imaging of PED6 hydrolysis. Fig. 7A illustrates that PED6 hydrolysis occurred first at the plasma membrane, most likely in the outer leaflet, and the perinuclear region. The pretreatment of cells with 30 M AA-COCF 3 did not change this pattern, underscoring that cPLA 2 is not involved in this process (Fig. 7B). Furthermore, the pre- ] i was seen only after 10 M ionomycin was added. B, PBE (10 7 cells/ml) were treated with 100 nM of hVPLA 2 for a given period at 37°C, and aliquots of cell lysates containing the same number of cells were subjected to Western blot analysis and probed with the antibodies specific for the phosphorylated ERK1/2. Equal loading of samples is confirmed by staining the corresponding samples with the ERK1/2 antibodies. Untreated cells and cells treated with 100 nM fMLP ϩ 5 g/ml CB were used as negative (CTR) and positive (fMLP) controls, respectively. C, PBE were treated with the HBSS buffer, 100 nM fMLP with 5 g/ml CB, or 100 nM of hVPLA 2 for 20 min, and Western blotting was performed with the anti-Ser 505 -phosphorylated cPLA 2 antibodies. The positive control (ϩCTR) and negative control (ϪCTR) were the Ser 505 -phosphorylated cPLA 2 expressed and purified from Sf9 cells and the cell lysate of untreated cells, respectively. Comparable blotting patterns were observed for PBE from four different donors. treatment with LY311727 for 30 min abrogated the signal at the perinuclear region but not at the plasma membrane (Fig.  7C). It was found that the 30-min incubation of human neutrophils and eosinophils with 10 M LY311727 fully suppressed the sPLA 2 activity of the cell lysates (data not shown), indicating that the inhibitor permeabilized the cells under our experimental conditions. Thus, Fig. 7C corroborates that hVPLA 2 is responsible for the PED6 hydrolysis at the perinuclear region. Furthermore, because the residual LY311727 was extensively washed from the cells prior to the addition of hVPLA 2 to the cells, the signal at the plasma membrane should originate from the outer leaflet. Neither hIIaPLA 2 nor hVPLA 2 /W31A yielded a detectable fluorescence signal under the same conditions (data not shown). Furthermore, hVPLA 2 /R100E/K101E induced the fluorescence signal only at the plasma membrane (Fig. 7D). In conjunction with data shown in Figs. 1 and 2, these results indicate that although hVPLA 2 liberates fatty acids, including AA, from both the outer plasma membrane and perinuclear membranes, a main site of AA release that is conducive to LTC 4 biosynthesis is the perinuclear membranes.
Effect of hVPLA 2 on Eosinophils Derived from the cPLA 2deficient Mouse-To corroborate the notion that hVPLA 2 activates eosinophils in a cPLA 2 -independent manner, we derived eosinophils from the mouse whose cPLA 2 gene is disrupted (cpla 2 Ϫ/Ϫ ) (5) and compared their properties with those derived from the wild type mouse (cpla 2 ϩ/ϩ ). These mice also naturally lack hIIaPLA 2 (hIIapla 2 Ϫ/Ϫ ), and thus the mutant has the double gene knockout (hIIapla 2 Ϫ/Ϫ /cpla 2 Ϫ/Ϫ ) (5). Murine bone marrow-derived eosinophils exhibited the characteristics of human eosinophils, including the expression of cell surface adhesion molecules (e.g. CD11b and CCR3 receptors) and the release of eosinophil peroxidase. 2 As expected from the lack of cPLA 2 , 100 nM fMLP and 5 g/ml CB failed to induce the AA release from hIIapla 2 Ϫ/Ϫ /cpla 2 Ϫ/Ϫ cells, whereas they elicited the significant AA release from hIIapla 2 Ϫ/Ϫ /cpla 2 ϩ/ϩ cells (Fig.  8A). In contrast, 100 nM hVPLA 2 was able to produce significant amounts of AA from both types of cells. Again, this activity was not affected by 30 M AACOCF 3 (data not shown). Because murine and guinea pig eosinophils preferentially produce LTB 4 over LTC 4 (32), we measured the LTB 4 release from wild type and mutant cells. As shown Fig. 8B, 100 nM fMLP and 5 g/ml CB caused the LTB 4 release from hIIapla 2 Ϫ/Ϫ /cpla 2 ϩ/ϩ cells but not from hIIapla 2 Ϫ/Ϫ /cpla 2 Ϫ/Ϫ cells, whereas 100 nM hVPLA 2 elicited the LTB 4 release from both types of cells. These data thus confirm that hVPLA 2 can induce the leukotriene biosynthesis in eosinophils in a cPLA 2 -independent manner. The full activity of hVPLA 2 on hIIapla 2 Ϫ/Ϫ /cpla 2 ϩ/ϩ cells also indicates that the endogenous hIIaPLA 2 of human eosinophils does not contribute significantly to LTC 4 biosynthesis during the hVPLA 2 -induced cell activation. DISCUSSION Our previous studies have shown that hVPLA 2 can elicit the AA release and the eicosanoid biosynthesis in different mammalian cells, including human neutrophils (13,14,21,28,33). In neutrophils that contain cPLA 2 and several forms of endogenous sPLA 2 s including hVPLA 2 (34), the exogenous hVPLA 2 acts on the outer plasma membrane and thereby induces the release of AA and LTB 4 mainly through the activation of cPLA 2 (14). The enzyme is eventually internalized via HSPG binding and degraded (13). The present work shows that in the case of human PBE, hVPLA 2 acts not only on the outer plasma membrane but also on the nuclear membrane via the HSPG-mediated internalization. It has been reported that human eosinophils contain cPLA 2 (29) and hIIaPLA 2 (30). Our sandwich enzyme-linked immunosorbent assay experiment indicated that human PBE do not contain a detectable amount of hVPLA 2 , and studies using hIIapla 2 Ϫ/Ϫ mouse eosinophils indicated that endogenous hIIaPLA 2 does not participate in the hVPLA 2 -induced cell activation. Although we could not preclude the possibility that other sPLA 2 s are present in eosino-phils and participate in eosinophil activation, this at least allowed us to investigate the effect of exogenous hVPLA 2 without having to sort out the contribution from endogenous hVPLA 2 and hIIaPLA 2 . As was the case with human neutrophils, hVPLA 2 was able to induce the release of AA and LTC 4 from both intact and activated eosinophils. This activity depends on the PC activity of the enzyme, as evidenced by the loss of activity by the W31A mutation that was shown to dramatically reduce the affinity of hVPLA 2 for PC membranes and consequently lower its activity on PC membranes (21). However, that is where the similarity ends. In the case of hVPLA 2induced activation of neutrophils, both the AA and the LTB 4 release showed the biphasic time dependence largely caused by the biphasic activation of cPLA 2 via a rise in [Ca 2ϩ ] i and cPLA 2 phosphorylation (14). Our results indicate that none of these phenomena occur during the activation of human eosinophils by hVPLA 2 . Both the pharmacological inhibition of cPLA 2 and the activation of eosinophils derived from the cPLA 2 -deficient mouse show that the hVPLA 2 -induced activation of eosinophils requires neither the presence nor the activation of cPLA 2 . As a matter of fact, our results show that hVPLA 2 can induce the synthesis and release of LTC 4 from PBE whether or not cPLA 2 is at resting state or activated by its potent agonist, fMLP/CB. This in turn suggests that exogenous hVPLA 2 can independently act as an effective activator of human eosinophils for LTC 4 production under different physiological conditions, possibly including conditions of airway inflammation and hyperresponsiveness. It should be noted that the 10 -100 nM of hVPLA 2 that is necessary for eosinophil activation might be higher than its physiological concentration. However, a significantly lower tissue or exudate concentration of hVPLA 2 might be required for the activation of eosinophils under pathophysiological conditions in which eosinophils and other hVPLA 2releasing cells might work in close proximity, which would in turn result in higher local concentration of hVPLA 2 on eosinophil surfaces. It has been reported that group V PLA 2 is expressed in macrophages (35)(36)(37), neutrophils (34), mast cells (38), and epithelial cells (23). Thus, the potential cell-cell interaction between eosinophils and other activated cells may turn out to be a key step in the progress of airway inflammation and hyperresponsiveness. Further investigation is needed to address this important issue.
This study not only represents the first observation that the LTC 4 synthesis in granulocytes may not require activation of cPLA 2 but also provides new insights into the unique mechanism by which hVPLA 2 achieves this feat. In the case of neutrophils, two products of hVPLA 2 -catalyzed hydrolysis of the outer plasma membrane PC, lysophosphatidylcholine in particular, and LTB 4 were shown to cause an increase in [Ca 2ϩ ] i that in turn induces the translocation of cPLA 2 and 5-LO to the FIG. 8. Effect of hVPLA 2 on the AA and LTB 4 release from eosinophils derived from cPLA 2 -deficient mouse. Eosinophils derived from the wild type mouse (cpla 2 ϩ/ϩ ; black bars) and the cPLA 2 -deficient mouse (hIIapla 2 Ϫ/Ϫ / cpla 2 Ϫ/Ϫ ; open bars) were labeled with [ 3 H]AA, treated with either 100 nM fMLP with 5 g/ml CB, or 100 nM hVPLA 2 for 20 min at 37°C, and the release of AA (A) and LTB 4 (B) was measured. nuclear membrane (14). Then these enzymes work hand in hand to generate AA and to convert it into LTB 4 . Our data show that hVPLA 2 also drives 5-LO to the nuclear membrane; however, this translocation is caused by a calcium-independent mechanism, presumably via 5-LO phosphorylation. It has been reported that the activity and localization of 5-LO can be modulated in a calcium-independent manner through its phosphorylation (39,40); however, the mechanism underlying these effects is poorly understood. Although our gel shift assay could not detect the mobility shift of 5-LO after hVPLA 2 treatment, this should not preclude the 5-LO phosphorylation. As a matter of fact, the in vivo phosphorylation of 5-LO has not been experimentally demonstrated. Undoubtedly, further studies are needed to address this complex issue, which is beyond the scope of this investigation. Intuitively, hVPLA 2 and 5-LO should be located in the vicinity for effective transformation of AA into LTC 4 . The results from our in vivo fluorescence PLA 2 assay and the AA/LTC 4 release assays using the hVPLA 2 wild type and its R100E/K101E mutant support the notion that the majority of AA converted to LTC 4 originates from intracellular membranes, most notably nuclear membrane. It has been shown that HSPG binding is essential for the internalization of most sPLA 2 s (10,11,41,42). In addition, hydrolysis products of PC were shown to be required for the entry of hIIaPLA 2 and hVPLA 2 into unstimulated mammalian cells (28). It would therefore seem that in the case of unstimulated PBE, the PC hydrolysis at the outer plasma membrane is primarily responsible for the internalization hVPLA 2 and does not directly provide AA for 5-LO. Once internalized and delivered to the nuclear membrane, hVPLA 2 would provide 5-LO with AA in proximity. It is unclear at present how exactly hVPLA 2 is delivered to the nuclear membrane and how hVPLA 2 shows the lipolytic activity at the nuclear membrane in the presence of submicromolar calcium. Although these questions require further investigation and are thus beyond the scope of this study, our previous studies indicated that hVPLA 2 could indeed show the significant lipolytic activity at the nuclear membrane of different mammalian cells under physiological conditions (28).
In summary, this study demonstrates that a sPLA 2 isoform, hVPLA 2 , can cause the synthesis of LTC 4 in human PBE by a mechanism that does not utilize cPLA 2 . As such, the study illustrates that sPLA 2 s can induce the eicosanoid biosynthesis in mammalian cells by diverse mechanisms depending on the cell type and the physiological conditions.