Extranuclear lipid bodies, elicited by CCR3-mediated signaling pathways, are the sites of chemokine-enhanced leukotriene C4 production in eosinophils and basophils.

Eosinophils and basophils, when activated, become major sources of cysteinyl leukotrienes, eicosanoid mediators pertinent to allergic inflammation. We show that the C-C chemokines, eotaxin and RANTES (regulated upon activation normal T cell expressed and secreted), activate eosinophils and basophils for enhanced leukotriene C(4) (LTC(4)) generation by distinct signaling and compartmentalization mechanisms involving the induced formation of new cytoplasmic lipid body organelles. Chemokine-induced lipid body formation and enhanced LTC(4) release were both mediated by CCR3 receptor G protein-linked downstream signaling involving activation of phosphoinositide 3-kinase, extracellular signal-regulated kinases 1 and 2, and p38 mitogen-activated protein kinases. Chemokine-elicited lipid body numbers correlated with increased calcium ionophore-stimulated LTC(4) production; and as demonstrated by intracellular immunofluorescent localization of newly formed eicosanoid, lipid bodies were the predominant sites of LTC(4) synthesis in both chemokine-stimulated eosinophils and chemokine-primed and ionophore-activated eosinophils. Eotaxin and RANTES initiated signaling via phosphoinositide 3-kinase and mitogen-activated protein kinases both elicits the formation of lipid body domains and promotes LTC(4) formation at these specific extranuclear sites.

prolong eosinophil survival (8). Eosinophils and basophils contain the single LTC 4 -synthesizing enzyme, LTC 4 synthase, and eosinophils are the predominant cellular source of this enzyme among resident and recruited cells in the bronchial tissues of asthmatics (9,10). In all cells, the highly regulated generation of LTs is dependent on activation of specific phospholipases and LT-synthesizing enzymes and involves small molecules (e.g. Ca 2ϩ ) and activation-dependent redistribution of 5-LO to specific membranous compartments within cells (11). One major candidate mechanism potentially involved in regulating LT formation is the translocation of 5-LO to the nuclear envelope (10,11), but in eosinophils this nuclear translocation has been associated with both increased and decreased LTC 4 formation (10,12).
Although mechanisms that activate specific leukocytes to generate LTs currently focus on the perinuclear envelope as the site of regulated LT formation (11), eosinophils and other leukocytes associated with inflammatory reactions in vivo characteristically contain increased numbers of extranuclear lipid-rich domains in the form of cytoplasmic lipid bodies (13). These enigmatic organelles, often overlooked if their defining lipid content is lost during cell staining, have long been noted to be prominent in many cell types associated with inflammation (14, 15). Although lipid bodies lack a delimiting membrane, their lipid content overlies a poorly understood honeycomb membranous matrix (16). Neither the genesis nor function of these organelles is well defined. Although eicosanoid-forming enzymes have been localized to lipid bodies, including 5-LO and LTC 4 synthase in eosinophil lipid bodies (16,17), to date there is no direct evidence that lipid bodies are sites of eicosanoid synthesis.
Both C-C chemokines, eotaxin and RANTES (regulated upon activation normal T cell expressed and secreted), signaling via CCR3 receptors expressed on eosinophils (18) and basophils (19), are active in recruiting these leukocytes to sites of allergic inflammation (20). Since other chemoattractants can enhance LT formation (21), we investigated the intracellular pathways by which these chemokines may activate leukocytes to enhance their regulated formation of cysteinyl LTs. Notably, chemokine engagement of CCR3 receptors initiated G protein-linked downstream signaling involving phosphoinositide 3-kinase (PI3K) and the mitogen-activated protein (MAP) kinases, extracellular signal-regulated kinases (ERK) 1/2 and p38, to induce the formation of new lipid body organelles and to enhance eosinophil LTC 4 release. Moreover, these cytokines elicited and promoted LTC 4 formation via PI3K and MAP kinase signaling not at the perinuclear envelope but rather specifically at these extranuclear lipid body domains.

EXPERIMENTAL PROCEDURES
Eosinophil and Basophil Purification-Peripheral blood was obtained with informed consent from 12 normal donors, and eosinophils * This work was supported by National Institutes of Health Grants AI20241, AI22571, AI41995 and HL56386. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
5-LO Immunolocalization-Basophil suspensions (10 6 /ml) were incubated (37°C) with eotaxin (100 ng/ml) for 1 h and then cytocentrifuged (500 rpm, 5 min) onto glass slides. Cytospin preparations were then fixed in 2% paraformaldehyde for 5 min. After three 10-min washes with 0.05% saponin (Sigma) and 1% human serum (Pierce) in HBSS, slides were incubated for 30 min with 200 l of polyclonal antiserum anti-5-LO (1:100 dilution) (Cayman) or with the isotype-matched rabbit IgG control. After washes with HBSS containing 0.05% saponin and 1% goat serum, slides were incubated for 30 min with Alexa546-labeled goat anti-rabbit IgG antibody (1:100 dilution) (Molecular Probes), washed with 0.1% saponin, washed with HBSS, and an aqueous mounting medium (Polysciences Inc., Warrington, PA) was applied to each slide before coverslip attachment. Slides were viewed with a 100ϫ objective by both phase-contrast and fluorescence microscopy using the Eclipse TE300 Nikon (Tokyo, Japan) fluorescence microscope. Electronic photography was performed by spot cooled color digital camera (model 1.3.0; Diagnostic Instruments Inc., Sterling Heights, MI) in conjunction with the image editing program Photoshop (Adobe Systems Inc., San Jose, CA).
Immunodetection of LTC 4 at Its Sites of Production in Eosinophils-To immunolocalize LTC 4 at its formation sites within eosinophils, viable eosinophils were embedded in an agarose matrix enabling their morphology and generated products to be microscopically localized (22). With modifications of prior techniques (24, 25), water-soluble 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide (EDAC) (Sigma) was used to cross-link eicosanoid carboxyl groups to amines in adjacent proteins. Briefly, to prepare the agarose matrix, 2.5% agarose (24°C gelling point) (Promega, Madison, WI) in sterile distilled H 2 O was melted at 70°C; and while liquid at 37°C, 9 volumes of agarose were mixed with 1 volume of 10ϫ concentrated RPMI 1640 medium. One volume of this medium-supplemented agarose was mixed at 37°C with 3 volumes of eosinophils at 15 ϫ 10 6 cells/ml in RPMI 1640 medium containing 1% fatty acid free-human albumin and 1 volume of RPMI 1640 medium containing 2% fatty acid free-albumin. As indicated, potential agonists, eotaxin (100 ng/ml), RANTES (100 ng/ml), or IL-8 (500 ng/ml), were added in 0.1 volumes to agarose/eosinophil mixtures. Immediately thereafter, 20-l samples were gently spread onto microscope slides and covered with CoverWell chambers (Grace Bio-Labs, Bend, OR). Each slide was overlaid with RPMI 1640 medium containing 1% albumin and an identical concentration of the stimulus present in the agarose/eosinophil mixture. Slides were incubated (37°C, humidified 5% CO 2 ) for 1 h. Overlying medium was removed and replaced with RPMI 1640, 1% albumin medium with or without 0.1 M A23187, and incubated for 15 min (37°C; 5% CO 2 ). Incubations were stopped by removing the chambers and fixing and permeabilizing the cells with 0.5% EDAC in HBSS for 30 min. In control experiments, eosinophils were alternatively fixed with paraformaldehyde (2%, 5 min) with and without 0.1% saponin permeabilization. After three washes (5 min each) with HBSS, the Alexa488-labeled rat anti-cysteinyl LT detection mAb (clone 6E7; Sigma) (Alexa488 protein labeling using a kit from Molecular Probes, Eugene, OR) was added (400 l of 10 g/ml) for 1 h. Slides were washed with HBSS, and an aqueous mounting medium (Polysciences) was applied to each slide before coverslip attachment. Slides were viewed by both phase-contrast and fluorescence microscopy, and photography and image analysis were performed as above. Two hundred eosinophils were scored, and the percentages of those exhibiting green staining for intracellular immunoreactive LTC 4 were calculated.
As a specificity control for the immunolocalization of LTC 4 , Al-exa488-labeled rat IgG1 (Sigma) was routinely included as a nonimmune isotype control for the anti-cysteinyl LT detection antibody. In addition, four other control conditions were evaluated: (i) substituting paraformaldehyde fixation with and without saponin permeabilization for the EDAC cross-linking and fixation step; (ii) using neutrophils rather than eosinophils embedded in the gel-matrix; (iii) pre-treating eosinophils in suspension for 30 min with the 5-LO activating protein inhibitor MK886 (10 M) prior to chemokine stimulation, and (iv) pretreating eosinophils with PI3K inhibitor, wortmannin (1 M), for 30 min prior to chemokine stimulation.
Analysis of Intracellular Distribution of Lipid Bodies in Agarose Matrix-embedded Eosinophils-To monitor changes in cell morphology, eosinophils were embedded in the agarose matrix, as described above, with or without chemokines. Following incubation of 1 h, cells were fixed with 2% paraformaldehyde (5 min) and stained with OsO 4 as described above. To obtain a measure of polarized redistribution of lipid bodies, cells were divided approximately into two halves, one of which contained the nucleus. Lipid bodies were enumerated in the two halves, and those cells found to have Ͼ75% of their lipid bodies in the nuclear half of the cell were scored as exhibiting polarized lipid body distribution. In four experiments, 25-50 cells were analyzed.
Statistical Analysis-Data were expressed as mean Ϯ S.D. Statistical comparisons were done by analysis of variance followed by Newman-Keuls Student's test. Differences were considered significant when p Ͻ 0.05. Correlation coefficients were determined by linear regression with significance (F test) at p Ͻ 0.05.

Eotaxin and RANTES Increase and Redistribute Lipid
Bodies-Resting eosinophils exhibited a spherical shape and a normal content of osmiophilic lipid bodies (9.2 Ϯ 1.0 lipid bodies/cell, mean Ϯ S.D., n ϭ 12) distributed generally throughout the cytoplasm (Fig. 1A). As fully expected from prior fluid phase assays (18), eotaxin (Fig. 1B) and RANTES (Fig. 1C), but not the C-X-C chemokine, IL-8 ( Fig. 1D), elicited shape changes in eosinophils, including cytoplasmic veiling, increased polarization of their overall shape and nuclear location, and prominent cytoplasmic projections (lamellipodia and uropodia). Not anticipated by prior results, eotaxin and RAN-TES also elicited changes in intracellular lipid bodies within eosinophils. Both chemokines dose-dependently induced new lipid body formation ( Fig. 2A). The magnitude of lipid body induction with each C-C chemokine (at 100 ng/ml) was comparable to that with 1 M PAF, a recognized stimulus for leukocyte lipid body formation (16,23,26). In contrast, IL-8 failed to elicit lipid body formation. In eotaxin-and RANTES-stimulated eosinophils, lipid bodies were located in cytoplasmic projections and especially in the perinuclear area ( Fig. 1, B and D).
In C-C chemokine polarized cells, many lipid bodies became aligned in groups proximate to the nuclear membrane and in the margin of the trailing uropod. Eosinophils in which Ͼ75% of lipid bodies were in the perinuclear half of the cell were 56 Ϯ 9% with eotaxin (100 ng/ml) (p Ͻ 0.05 versus unstimulated, n ϭ 4) and 59 Ϯ 5% with RANTES (100 ng/ml) (p Ͻ 0.05 versus unstimulated, n ϭ 3) in contrast to only 16 Ϯ 7% with unstimulated eosinophils (mean Ϯ S.D., n ϭ 4) and 19 Ϯ 6% with IL-8 (500 ng/ml) (n ϭ 3). Thus, eotaxin and RANTES induced both the formation of new lipid bodies and their polarized redistribution within eosinophils.
The capacity of the two chemokines to elicit new formation of intracellular lipid body organelles was not restricted to eosinophils. Basophils share with eosinophils several functional features including CCR3 expression and the ability to produce LTC 4 (19). Unstimulated basophils contained 3.1 Ϯ 1.0 lipid bodies/cell (mean Ϯ S.D., n ϭ 4). Both eotaxin and RANTES, but not IL-8, stimulated new lipid body formation in basophils with magnitudes comparable to that with 1 M PAF (Fig. 3, A  and B).
The induction of lipid body formation in eosinophils and basophils by eotaxin and RANTES was mediated through the G protein-linked CCR3 chemokine receptor. Pre-treatment of eosinophils or basophils with either pertussis toxin or a blocking anti-CCR3 mAb (but not an isotype control antibody) significantly inhibited lipid body formation induced by both C-C chemokines (Table I, Fig. 3B). CCR3-initiated downstream signaling pathways active in lipid body formation included mobilization of intracellular pools of Ca 2ϩ and PI3K and ERK 1/2 and p38 MAP kinases. Eotaxin-induced lipid body formation occurred in a Ca 2ϩ -free medium and was blocked (by 82 Ϯ 1%; n ϭ 3, p Ͻ 0.05) by pre-treatment with the cell permeable Ca 2ϩ chelator, BAPTA-AM (25 77), but not by its impermeable analog BAPTA free acid (8 Ϯ 8%). Wortmannin and LY294002, the PI3K inhibitors, PD98059 and U0126, two specific inhibitors of the ERK1/2 activating kinase (MAP ERK kinase), and SB203586, a p38 MAP kinase inhibitor, each inhibited eotaxinstimulated lipid body formation (Fig. 4A). Thus, eotaxin stim-  Both exogenous PAF (16,23) and PAF formed endogenously in eosinophils in response to IL-5 or Fc␥ receptor engagement (26) also signal via pertussis toxin-inhibitable G protein-linked PAF receptors to elicit lipid body formation. For PAF-induced lipid body formation, post-PAF receptor downstream signaling is obligately dependent on 5-LO activation to form 5-(S)-hydroxyeicosatetraenoic acid (23). We, therefore, evaluated whether CCR3 downstream signaling to form lipid bodies was mediated either by autocrine PAF formation or dependent on endogenous 5-LO activation. Concentrations of the PAF receptor antagonist CV6209, the 5-LO inhibitor AA861, and the 5-LO activating protein inhibitor MK886 that blocked PAFinduced lipid body formation in eosinophils did not inhibit lipid body formation induced by the C-C chemokines (Table I). Other downstream signaling for lipid body formation also differed for PAF-and CCR3-mediated responses. Although PAF-elicited lipid body formation is inhibitable by the protein kinase C inhibitors, chelerythrine and calphostin C (27,28), these inhibitors did not block CCR3-mediated lipid body formation in eosinophils (Fig. 4A). The tyrosine kinase inhibitors, herbimycin and genistein, inhibited neither CCR3-elicited (Fig. 4A) nor PAF-elicited (27) lipid body formation. Thus, the intracellular signaling pathways leading to lipid body induction differ between the PAF and CCR3 chemokine receptors, although both are G protein-linked, seven-transmembrane-spanning receptors that are expressed on eosinophils and basophils and can initiate lipid body formation.
Eotaxin and RANTES Enhance Eosinophil and Basophil LTC 4 Production-Both eotaxin and RANTES very effectively primed eosinophils for increased LTC 4 release in response to a submaximal 0.1 M concentration of calcium ionophore A23187 (Fig. 2B). Pre-stimulation of eosinophils for 1 h with eotaxin or RANTES, but not with IL-8, dose-dependently evoked increases in A23187-induced LTC 4 production. At 100 ng/ml, eotaxin-and RANTES-pre-stimulated eosinophils released about 2.2-and 3.2-fold as much LTC 4 as did eosinophils challenged with A23187 alone. The increased quantities of LTC 4 generated by eosinophils primed with increasing concentrations of the C-C chemokines correlated highly with the increased numbers of elicited lipid bodies (r ϭ 0.91 (p Ͻ 0.05) and r ϭ 0.93 (p Ͻ 0.05), for eotaxin and RANTES, respectively) (Fig.  2, A and B). Moreover, inhibition of PI3K with wortmannin and LY294002, ERK1/2 with PD98059 or U0126, and p38 with SB203580, at concentrations that inhibited eotaxin-induced lipid body formation (Fig. 4A), also inhibited LTC 4 production by eosinophils (Fig. 4B). Again, inhibitors of protein kinase C or tyrosine kinase, which did not inhibit eotaxin-induced lipid body formation (Fig. 4A), also failed to affect the consequent LTC 4 production (Fig. 4B).
Although eosinophils stimulated for 1 h with 100 ng/ml eotaxin or RANTES formed more lipid bodies, quantities of LTC 4 released extracellularly in supernatants from eosinophils not activated by the calcium ionophore were not sufficient to be detectable by enzyme immunoassays (EIA) (data not shown). Supernatants of eotaxin-or RANTES-stimulated basophils also did not contain sufficient levels of LTC 4 to be detectable by EIA, but both C-C chemokines effectively primed basophils for increased LTC 4 release in response to A23187. At 100 ng/ml, eotaxin-and RANTES-pre-stimulated basophils released 8.1and 9.3-fold, respectively, as much LTC 4 as did basophils challenged with A23187 alone. Moreover, pertussis toxin and anti-CCR3 neutralizing antibody blocked both lipid body induction (Fig. 3B) and enhanced LTC 4 production by basophils stimulated with eotaxin and RANTES (data not shown).
Lipid Bodies Are Sites for Eotaxin-and RANTES-enhanced LTC 4 Synthesis in Eosinophils-Since eotaxin-and RANTESinitiated signaling led to correlative quantitative increases in both lipid body formation and enhanced LTC 4 formation, we employed a new strategy for direct in situ immunolocalization of intracellular LTC 4 to ascertain the intracellular compartmentalization of cysteinyl LT synthesis. Unstimulated eosino- phils exhibited no immunofluorescent staining for LTC 4 (Table  II, Fig. 5A), demonstrating no background staining or preformed LTC 4 , as expected in the absence of cell stimulation. In contrast, eosinophils activated for 15 min with 0.1 M calcium ionophore A23187 yielded intense and localized immunofluorescent staining for LTC 4 in virtually all eosinophils (Table II, Fig. 6A). Much of the staining was in a linear perinuclear pattern, although sites distant from the nuclei also exhibited LTC 4 labeling. The specificity of this immunofluorescent staining for LTC 4 was ascertained. First, the detected LTC 4 should be a product of an active 5-LO pathway in eosinophils; accordingly MK886, which blocks 5-LO pathway activity, completely abolished immunofluorescent staining for LTC 4 in A23187stimulated eosinophils (Table II, Fig. 6A). Second, specificity of the immunofluorescence for cysteinyl LTs was supported by the absence of immunostaining when a fluorochrome-labeled isotype control antibody replaced the anti-cysteinyl LT mAb (Table II). Third, A23187-stimulated neutrophils (which do not form cysteinyl LTs) exhibited no immunofluorescent labeling with the fluorochrome-labeled anti-cysteinyl LT mAb (Table  II). Fourth, substitution of paraformaldehyde fixation with and without saponin permeabilization for EDAC treatment abolished immunofluorescent detection of LTC 4 in A23187-stimulated eosinophils, indicating that the cell permeabilization and eicosanoid lipid-protein cross-linking effected by EDAC were essential for intracellular LTC 4 immunodetection (data not shown). The detecting fluorochrome-labeled mAb recognizes LTC 4 , LTD 4 , and LTE 4 , but the latter two are formed extracellularly, whereas only LTC 4 is generated intracellularly (6). These findings validate the specificity for detecting LTC 4 formed at sites within stimulated eosinophils.
About 25% of eosinophils stimulated with eotaxin or RAN-TES exhibited staining for immunoreactive LTC 4 (Fig. 5, B and  C), that was fully inhibitable by the 5-LO inhibitor, MK886, and the PI3K inhibitor, wortmannin (Table II). In contrast, eosinophils incubated with IL-8 exhibited no anti-LTC 4 immunofluorescence (Table II, Fig. 5D). Although occasional perinuclear LTC 4 localization was observed in eotaxin-stimulated (Fig. 5B, arrowhead) and RANTES-stimulated eosinophils, most LTC 4 was localized at punctate locations (Fig. 5, B and C, arrows) distant from the nucleus. Thus, in the absence of calcium ionophore stimulation, the two C-C chemokines were directly stimulating low level LTC 4 formation within eosinophils, although extracellular levels remained beneath the sensitivity of LTC 4 EIA assays of supernatant fluids (as noted above). This chemokine-stimulated LTC 4 formation occurred predominantly at discrete extranuclear lipid body sites. Moreover, LTC 4 formation at lipid bodies was dependent on PI3K activation.
Further evidence of the roles of lipid bodies as sites of LTC 4 formation was found in eosinophils that were first primed with eotaxin or RANTES and then stimulated with calcium ionophore to activate the 5-LO pathway. Whereas unprimed and non-polarized eosinophils stimulated with A23187 exhibited substantial linear, perinuclear rim anti-LTC 4 staining, consistent with localization at the nuclear envelope (Fig. 6A, arrowheads), and some focal staining at extranuclear sites, the pattern of LTC 4 localization was different in eotaxin-and RAN-TES-primed eosinophils. In these C-C chemokine-primed and A23187-stimulated eosinophils, only occasional LTC 4 immunofluorescence was detected with a linear perinuclear rim pattern (Fig. 6C, arrowhead). More extensive anti-LTC 4 staining exhibited a punctate pattern, with very distinct focal staining proximate to, but separate from, the nucleus (Fig. 6B, asterisk) and within the uropodia and lamellipodia (Fig. 6, B and C, arrows) of polarized eosinophils, fully consistent in size, form, number, and distribution with cytoplasmic lipid bodies (Fig. 1,  B and C). Again, pre-treatment of eosinophils with the 5-LO pathway inhibitor, MK886, completely abolished all LTC 4 immunostaining (Fig. 6, B and C). The PI3K inhibitor, wortmannin, which inhibited chemokine enhanced release of LTC 4 from eosinophils (Fig. 4), inhibited chemokine-elicited LTC 4 production only at all (both preformed and chemokine-elicited) lipid body sites, with no inhibitory effect on the immunoreactive LTC 4 generated at the perinuclear membrane of eosinophils (Fig. 6). Notably, in A23187-activated eosinophils, wortmannin inhibited lipid body LTC 4 formation, but inhibited neither LTC 4 production at the perinuclear membrane (Fig. 6) nor overall LTC 4 generation, as assessed by immunofluorescent microscopy (Table II) and ELISA assays of released LTC 4 (data not shown).
In basophils, LTC 4 localization at lipid bodies was not possible since these cells were destroyed during EDAC cross-linking and fixation, but basophil lipid bodies were sites of 5-LO localization (Fig. 3C). DISCUSSION Our results elucidate mechanisms whereby chemoattractrant chemokines may activate a major functional response in eosinophils and basophils, specifically their capacity to generate cysteinyl LTs, and also provide novel findings pertinent to the regulated intracellular compartmentalization of eicosanoid formation by multiple cell types. The C-C chemokines, eotaxin and RANTES, acting via CCR3 receptors on eosinophils and basophils, may not only recruit these leukocytes to sites of allergic inflammation but also activate distinct intracellular signaling and compartmentalization mechanisms within these recruited cells to enhance their capacity to form cysteinyl LTs. The means whereby these chemokines enhance eosinophil and basophil LT formation were suggested by the finding that these two C-C chemokines, but not IL-8, stimulated the formation of new cytoplasmic lipid bodies, organelles previously implicated in eicosanoid synthesis (17). This induction of lipid body formation was mediated via G protein-linked CCR3 chemokine receptors and was dependent on downstream activation of PI3K and the ERK1/2 and p38 MAP kinases. Activation of ERK and p38 MAP kinases in eosinophils in response to chemoattractants (21) and of PI3K in human basophils in response to IgE-mediated stimulation (29) has been shown to participate in the regulated activation of LTC 4 formation, but the means for such regulation has been undefined. Inhibitors of PI3K, ERK1/2, and p38 MAP kinases inhibited both CCR3-mediated lipid body formation and CCR3-mediated priming for enhanced eosinophil LTC 4 release. CCR3-mediated downstream signaling, in contrast to IL-5-and Fc␥ receptor-mediated stimulation of eosinophil lipid body formation (26), did not require endogenous PAF formation and, distinct from PAF, which signals through its own heterotrimeric G-protein-linked, seven-transmembrane-spanning receptor, was not dependent on intermediate 5-LO activity or PKC activation (23).  4. Signaling via activation of PI3K and ERK1/2 and p38 MAP kinases mediates lipid body formation (A) and priming for LTC 4 production (B) by C-C chemokine-stimulated human eosinophils. Eosinophils were pretreated for 30 min with kinase inhibitors (as indicated) and then incubated with eotaxin for 1 h for induction of lipid bodies. Thereafter, eosinophils were incubated with 0.1 M A23187 for 15 min for LTC 4 production. To normalize for differences in base-line LTC 4 release between different donors, results are presented as percentage of increase in the LTC 4 released by A23187-stimulated cells. Results are means Ϯ S.D. from four independent assays. **, p Ͻ 0.01 compared with eotaxin.
That the formation of new lipid body organelles elicited by CCR3-mediated signaling provided a distinct intracellular compartment for the regulated generation of LTC 4 was indicated by several findings. The numbers of eotaxin-and RAN-TES-elicited lipid bodies correlated with the magnitudes of the priming responses for increased extracellular release of LTC 4 elicited by prior exposure of eosinophils to eotaxin and RAN-TES. Inhibitors of PI3K and ERK1/2 and p38 MAP kinases inhibited both CCR3-mediated lipid body formation and enhanced LTC 4 release by eosinophils. More directly, newly formed LTC 4 was localized almost exclusively to lipid bodies in eotaxin-and RANTES-stimulated eosinophils, in contrast to perinuclear localization of LTC 4 in eosinophils activated solely with calcium ionophore A23187. In ionophore-activated eosinophils, the PI3K inhibitor, wortmannin, failed to inhibit the overall eosinophil LTC 4 generation that was shown to occur at perinuclear membranes. In contrast, wortmannin, which inhibited the chemokine-enhanced extracellular release of LTC 4 , blocked chemokine-stimulated LTC 4 formation at both preformed and chemokine-elicited lipid bodies, but not at the perinuclear membrane. These findings indicated that chemokine-elicited activation of PI3K was involved in both the induction of new lipid body formation and the regulated activation of LTC 4 formation at all lipid body domains, providing additional evidence that chemokine-elicited lipid bodies were the principal sites of enhanced LTC 4 synthesis.
Heretofore, the intracellular sites of eicosanoid formation in any cell have not been directly demonstrated, but rather have been inferred based on the immunolocalization of specific eicosanoid-forming enzymes. Hence, the translocation of 5-LO from the cytosol to the nucleus in eosinophils in response to calcium ionophore has suggested that enhanced LTC 4 is formed at perinuclear sites (12) (consistent with our findings in eosinophils solely activated by ionophore). Indeed, in many cells, major sites of synthesis of both 5-LO-and cyclooxygenase pathway-derived eicosanoids are believed to be the perinuclear membranes (11,30,31). Based on calcium ionophore-elicited translocation of 5-LO from either the nucleus or the cytosol, redistribution of 5-LO to the nuclear envelope has been found in neutrophils, eosinophils, alveolar macrophages, blood monocytes, mast cells, and the rat basophilic leukemia mast cell-like cell line (11). A role for extranuclear sites for 5-LO catalyzed LT formation, however, would be compatible with the earlier finding that enucleate neutrophil cytoplasts generate LTB 4 in response to A23187 (32). More recently, we established that enucleate eosinophil cytoplasts formed lipid bodies in response to PAF stimulation and that the numbers of lipid bodies in cytoplasts correlated with levels of primed LTC 4 and PGE 2 released by cytoplasts following submaximal A23187 challenge (16). Moreover, lipid bodies in enucleate cytoplasts were sites of immunolocalized 5-LO, cyclooxygenase, and LTC 4 synthase proteins (16). Our findings extend knowledge of the regulated intracellular compartmentalization of eicosanoid formation. Although perinuclear membranes may be sites of eicosanoid formation in cells singularly activated with calcium ionophore A23187 (as we confirmed), in leukocytes first stimulated with specific chemokines or other leukocyte agonists, e.g. PAF, lipid bodies (as present in vivo in inflammation-associated leukocytes) are the predominant sites of enhanced eicosanoid formation.
Since the immunolocalization of eicosanoid-forming proteins need not reflect the regulated activity of these enzymes, we utilized a more direct approach to detect the intracellular sites of 5-LO-and LTC 4 synthase-mediated LTC 4 formation in eosinophils. Eosinophils incubated in a gel matrix were fixed and permeabilized with a water soluble cross-linker EDAC, enabling: 1) the covalent cross-linking of eicosanoid carboxyl groups to adjacent amines in proteins at the sites of eicosanoid formation, 2) the penetration of detecting fluorochrome-conjugated anti-cysteinyl LT mAb into eosinophils, and, importantly, 3) the relative preservation of lipid body domains (which dissipate with air drying or commonly used alcohol fixation). Eosinophils were especially amenable to this strategy for localizing LTC 4 , their dominant 5-LO product. In eosinophils, 5-LO and LTC 4 synthase enzymatic activities are tightly coupled, as indicated by the paucity of nonenzymatic breakdown products of intermediate LTA 4 (33); both enzymes have been localized at eosinophil lipid bodies (16). In basophils, 5-LO was also localized at lipid bodies; but basophils, unlike eosinophils, were not durable to the LTC 4 fixation process. In contrast, in calcium ionophore-activated neutrophils, 5-LO-catalyzed formation of LTA 4 is not tightly coupled to subsequent LTB 4 formation, which may be mediated by cytosolic leukotriene A 4 hydrolase or may occur substantially following extracellular release of LTA 4 (34). Immunofluorescent localization of newly formed LTC 4 in eosinophils activated with calcium ionophore exhibited a predominantly perinuclear staining, remarkably similar to the immunolocalization of 5-LO protein reported in ionophore-activated eosinophils (12). In contrast, eosinophils exposed to eotaxin or RANTES prior to ionophore activation showed pre- Lipid bodies are complex and as yet poorly understood organelles. They are not unique to eosinophils and basophils and are found in a diversity of cells ranging from fibroblasts to endothelial cells to leukocytes (35). Lipid body numbers characteristically increase in cells associated with various forms of inflammation (14, 15,36). Lipid body formation is rapidly inducible by specific stimuli, including hypoxia in endothelial cells (37), cis-unsaturated fatty acids in neutrophils (38), and receptor-mediated ligands, including PAF (23) and, as shown here, specific chemokines for eosinophils and basophils. Lipid bodies contain esterified arachidonic acid, ill defined membranous structures, and in many cell types, including eosinophils, basophils, neutrophils, alveolar macrophages, mast cells, one or more enzymes involved in eicosanoid synthesis, including cyclooxygenase, 5-LO, 15-LO, and/or LTC 4 synthase (16, 39 -41). Lipid bodies also contain cytosolic phospholipase A 2 and several signal transducing kinases, including PI3K, ERK1, ERK2, p85, and p38 MAP kinases (42,43). Although these findings have suggested that lipid bodies might be sites of regulated eicosanoid formation in several cell types, direct evidence for such had been lacking. Our direct demonstration that eosinophil lipid bodies are specific sites of LTC 4 synthesis has implications for many other cells types, especially those involved in diverse inflammatory responses, in which lipid bodies might likewise serve as sites of regulated eicosanoid formation.
Although eosinophils stimulated with eotaxin or RANTES did not release levels of LTC 4 adequate for detection in supernatants by EIA, the heightened sensitivity of the immunofluo-FIG. 5. C-C chemokines directly stimulate LTC 4 production in eosinophils. Phase-contrast (left panels) and fluorescent (right panels) microscopy of identical fields of eosinophils incubated for 1 h, and then fixed with EDAC and stained with Alexa488-labeled anti-cysteinyl LT mAb. To facilitate intracellular localization, anti-LTC 4 immunoreactive sites (green staining) were overlaid on phase-contrast images and white lines were drawn delineating the nuclear perimeter. Unstimulated (A) and IL-8-stimulated (D) eosinophils display no fluorescent LTC 4 immunostaining. Representative eosinophils stimulated with eotaxin (100 ng/ml) (B) and RANTES (100 ng/ml) (C) exhibit perinuclear (arrowheads) and punctate cytoplasmic (arrows) immunoreactive LTC 4 . Bar, 5 m.
FIG. 6. Perinuclear and extranuclear production of LTC 4 in human eosinophils. Phase-contrast and fluorescent microscopy of identical fields of eosinophils incubated for 1 h with the C-C chemokines (as indicated) and activated for 15 min with 0.1 M A23187. Eosinophils were fixed with EDAC and stained with Alexa488-labeled anti-cysteinyl LT mAb. To facilitate intracellular localization, anti-LTC 4 immunoreactive sites (green staining) were overlaid on phase-contrast images and white lines were drawn delineating the nuclear perimeter. Eosinophils activated with A23187 exhibit fluorescent anti-LTC 4 staining especially at the perinuclear envelope (arrowheads) as well as at some extranuclear sites (arrow). Eosinophils stimulated with eotaxin (100 ng/ml) (B) or RANTES (100 ng/ml) (C) prior to A23187 activation exhibited some perinuclear anti-LTC 4 staining (arrowheads) and abundant punctate LTC 4 immunoreactive staining both proximate to the nucleus (*) and at more distant cytoplasmic sites (arrows). Eosinophils pre-treated with the 5-LO inhibitor, MK 886 (10 M), had no anti-LTC 4 immunofluorescent staining (right panels), whereas cells pre-treated with the PI3K inhibitor, wortmannin, showed only perinuclear staining for LTC 4 . Bar, 5 m. rescent detection of LTC 4 formation enabled the demonstration that about a quarter of eosinophils stimulated solely with these C-C chemokines, and not with IL-8, synthesized low levels of LTC 4 . Since LTC 4 is being recognized to have autocrine effects on eosinophils, including prolonging their longevity (8) and regulating the vesicular transport-mediated release of preformed eosinophil granule-derived cytokines (e.g. IL-4), 2 the capacity of the two chemokines to stimulate even low level LTC 4 synthesis intracellularly at focal extranuclear sites may augment eosinophil effector functioning. Moreover, the stimulation of eosinophils for increased LTC 4 formation by eotaxin and RANTES, based on their receptor-mediated signaling to induce and activate lipid body organelles that are sites for regulated LTC 4 synthesis, promotes cysteinyl LT generation and release by recruited eosinophils and further contributes to the pathogenesis of allergic inflammation.