Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Lipoxins belong to the eicosanoid family of bioactive lipid mediators that carry trihydroxytetraene structures as distinguishing features. They are generated in mammals predominantly by transcellular biosynthetic routes during cell-cell interactions, which is now recognized as an important means of both amplifying and generating new lipid-derived mediators (reviewed in Serhan, 1997). Two major transcellular routes of LX biosynthesis by LO interactions in human cell types are established. LXs generated by these two pathways carry their C-15 hydroxyl group mainly in the 15S-configuration, which is inserted by lipoxygenase-based mechanisms. LXs not only are formed in vitro in isolated cells but also are generated in humans and in experimental animals. Situations that lead to cell-cell adherence and inflammation in vivo can produce LXs within local microenvironments (reviewed in Serhan, 1997). At the nanomolar level, LXs display vasodilatory and immunoregulatory roles in in vitro and in vivo models. LXs inhibit, for example, both neutrophil and eosinophil chemotaxis (Lee et al., 1989; Soyombo et al., 1994). In coincubation systems, LXA₄ inhibits PMN transmigration across both endothelial and epithelial cells (Colgan et al., 1993). Thus their immunoregulatory actions implicate them as endogenous "stop signals" acting on human leukocytes (Serhan, 1997).

Aspirin is the lead nonsteroidal anti-inflammatory drug (Weissmann, 1991). Besides its antithrombotic and anti-inflammatory action, low doses of aspirin have several newly recognized beneficial actions that include prevention of cardiovascular diseases (Savage et al., 1995) and decreasing incidence of lung, colon and breast cancer (Giovannucci et al., 1995; reviewed in Levy, 1997; Schreinemachers and Everson, 1994). They may also be relevant in treatment of human immunodeficiency virus (Macilwain, 1993). Although inhibition of PGHS could account for many of ASA's pharmacological properties (Samuelsson, 1982), it is difficult to ascribe the newly recognized therapeutic effects of ASA solely to the inhibition of prostaglandin formation (Weissmann, 1991). Along these lines, a novel biosynthetic pathway was recently uncovered that is triggered by ASA treatment and costimulation of human leukocytes and either vascular endothelial cells or mucosal epithelial cells (Clària et al., 1996; Clària and Serhan, 1995). Briefly, ASA acetylation of PGHS-2 in endothelial or epithelial cells switches the enzyme's catalytic activity from a prostaglandin synthase to an R-lipoxygenase also found with isolated enzymes (Lecomte et al., 1994; Mancini et al., 1994; Xiao et al., 1997). Hence, in either activated endothelial or epithelial cells, acetylated PGHS-2 generates 15R-HETE from endogenous stores of AA and leads to production of 15-epi-LXs via transcellular biosynthesis in PMN (reviewed in Serhan, 1997). The 15-epi-LXs (ATL, aspirin-triggered LXs) have their C-15 hydroxyl group in the R configuration that is retained from the precursor 15R-HETE (Clàra et al., 1996; Clària and Serhan, 1995; Serhan, 1997).

The newly uncovered ATL have been examined in several experimental settings. It has already been established that 15-epi-LXA₄ is approximately twice as potent as native LXA₄ in inhibiting neutrophil adhesion to endothelial cells (Clària and Serhan, 1995) and that 15-epi-LXB₄ inhibits cell proliferation in vitro (Clària et al., 1996). Recently, we found that stable analogs of LXA₄ and 15-epi-LXA₄specifically, analogs that resist rapid inactivationapplied topically to mouse ears dramatically inhibit leukocyte infiltration (Takano et al., 1997). These analogs also inhibit leukocyte rolling and adherence in vivo in the rat mesenteric microvasculature (Scalia et al., 1997). Given that ATL are generated via ASA-triggered transcellular biosynthesis and possess potent actions in events of interest in inflammation and proliferative disorders, they may contribute to some of the beneficial actions of ASA. Therefore, it is critical to establish cellular and animal models to test this hypothesis and develop methods with appropriate sensitivity and selectivity to elucidate biological settings involved in the generation and roles of ATL.

The antibody-based assays (radioimmunoassay or ELISA), which can detect picogram amounts of materials, have proved valuable tools in detecting eicosanoids (reviewed in Granström et al., 1987). Along these lines, we developed a LXA₄ ELISA that detects LXA₄ generated in vitro and in vivo (Levy et al., 1993). When the aspirin-triggered 15-epi-LX pathway was uncovered, we found that the LXA₄ antiserum, which is selective when compared with other eicosanoids and LXA₄ isomers (Levy et al., 1993), did not distinguish the chirality at carbon 15 between LXA₄ and 15-epi-LXA₄ (Clària and Serhan, 1995). Thus this LXA₄ antiserum made possible the preliminary assessment of 15-epi-LXA₄ in ASA-dependent fashion in rat kidney (Badr et al., 1996). Here, we report the development of a new 15-epi-LXA₄-selective ELISA with high sensitivity and stereoselectivity that distinguishes 15-epi-LXA₄ from native LXA₄, and we demonstrate its utility for detecting 15-epi-LXA₄ generation by both cells and animal models.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. The following drugs and chemicals were kindly provided by or obtained from the sources indicated: THP-1 (human acute monocytic leukemia cell line) and Sf9 (Spodoptera frugiperda) cells (ATCC, Rockville, MD), cell culture media and reagents (Biowhittaker, Walkersville, MD), goat anti-rabbit IgG (Zymed, San Francisco, CA), 96-well polystyrene plates for ELISA (Costar, Cambridge, MA), anti-human PGHS-2 polyclonal antibody (Oxford Biomedical Research, Oxford, MI), extract-clean C18 cartridges (500 mg) (Alltech, Deerfield, IL), 3,3',5,5'-tetramethyl benzidine (TMB), protein A beads, LPS (derived from Escherichia coli, serotype 055:B5), ionophore A23187, indomethacin, phenylmethyl sulfonyl fluoride (PMSF), TritonX-100, ethylenediaminetetraacetic acid (EDTA), leupeptin and aprotinin (Sigma Chemical Co., St. Louis, MO), AA (Cayman Chemical Co., Ann Arbor, MI), 15R-HETE, 15S-HETE, 5S-HETE, 12S-HETE, LXA₄ and LXB₄ (Cascade Biochem, Reading, Berkshire, England), 15-epi-LXA₄-methyl ester, 15-(R/S)-methyl-LXA₄ and 15-epi-LXB₄-methyl ester prepared by total organic synthesis using described procedures (Nicolaou et al., 1991; Serhan et al., 1995), 17-ODYA (Biomol, Plymouth Meeting, PA), methyl formate (Eastman Kodak, Rochester, NY), aspirin (Spectrum Chemical, Cardena, CA), LXA₄ ELISA kit (ELISA Technology Inc., Lexington, KY), and chemiluminescence substrate kits (Boehringer Mannheim, Indianapolis, IN). Recombinant baculovirus containing human PGHS-2 cDNA was a generous gift from Dr. Robert A. Copeland of DuPont Merck (Wilmington, DE).

Cell culture and isolation. THP-1 cells were grown in RPMI supplemented with 10% fetal bovine serum and antibiotics in a 37°C incubator with 5% CO₂. Human PMN were isolated from fresh peripheral blood of healthy donors who had denied taking aspirin or other medication for at least one week. PMNs were isolated by Ficoll-Hypaque gradient centrifugation and dextran sedimentation (Böyum, 1968) as described in Levy et al. (1993). Cells were suspended in PBS⁺⁺ and contained 96 ± 3% PMN (n = 3), as enumerated by light microscopy.

Preparation of 15-epi-LXA₄ antibody. 15-epi-LXA₄ was coupled to KLH through the succinimide ester of 15-epi-LXA₄ (Tai and Yuan, 1978) and used as the antigen to produce polyclonal antibody. 15-epi-LXA₄ was also labeled with HRP through the succinimide ester of 15-epi-LXA₄, which was then diluted (1:1000) and used to develop an ELISA for 15-epi-LXA₄. Rabbits were immunized initially with 1 mg of 15-epi-LXA₄-KLH conjugate through s.c. injections. They were then boosted twice monthly with 0.5 mg of the conjugate and bled via their ear veins one week after each boost. The blood was collected, and 1/10 volume of 3.8% sodium citrate was added to prevent clotting. It was then centrifuged at 2000 × g for 10 min. Clear supernatants were collected as crude sera and used to prepare the ELISA.

15-epi-LXA₄ ELISA. Polystyrene 96-well plates were precoated with about 1 µg of affinity-purified goat anti-rabbit IgG per well in 100 µl of coating buffer (0.1 M NaHCO₃/Na₂CO₃, pH 9.6) for 2 hr at room temperature. Each plate was then blocked with 200 µl of 5% nonfat milk, 5% sucrose in EIA buffer (100 mM K₂HPO₄/KH₂PO₄, pH 7.4, containing 150 mM NaCl, 0.5% BSA and 0.025% proclin 300) for 1 hr at room temperature. After three washings with buffer consisting of 10 mM K₂HPO₄/KH₂PO₄, pH 7.4, containing 0.05% Tween 20 (i.e., wash buffer), the assay was initiated by adding diluted antisera (1:10,000 dilution, 50 µl), HRP-labeled 15-epi-LXA₄ (100 µl) and serial dilutions of synthetic 15-epi-LXA₄ or samples (50 µl) into each well. Each plate was gently shaken for 1 hr at room temperature. After three washings with wash buffer, the enzyme reaction was carried out by adding 200 µl of substrate (1.25 mM TMB and 6 mM H₂O₂ in 25 mM sodium citrate buffer, pH 3.5) for 30 min at room temperature. The absorbance changes at 650 nm were monitored using an ELISA Plate Reader (Bio-Tek Instruments, Inc., Winooski, VT). All samples were assayed in duplicate.

Immunoprecipitation of synthetic 15-epi-LXA₄. Immobilized 15-epi-LXA₄ IgG was prepared by incubation of 0.5 ml of protein A beads with 15-epi-LXA₄ antiserum in ELISA buffer (100 µl for 18 hr). The beads were washed and suspended in 0.5 ml of ELISA buffer. The synthetic 15-epi-LXA₄ was then incubated with immobilized IgG for 1 hr at room temperature. The beads were rapidly pelleted, suspended in ELISA buffer (0.5 ml) and added to two volumes of cold methanol (1°C). The samples were then extracted using solid-phase C18 cartridges (Serhan, 1990) and taken to LC/MS/MS for further analysis.

LC/MS/MS of synthetic LXA₄ and 15-epi-LXA₄. LC/MS/MS was performed on an LCQ (Finnigan MAT, San Jose, CA) quadrupole ion trap mass spectrometer system equipped with an electrospray atmospheric pressure ionization probe. Compounds were dissolved in methanol and injected into the HPLC component, which consisted of a SpectaSYSTEM P4000 quaternary gradient pump (Thermo Separation Products, San Jose, CA), a LUNA C18-2 (150 × 2 mm, 5 µm) column (Phenomenex, Torrance, CA), and a rapid spectra scanning SpectraSYSTEM UV2000 UV/VIS absorbance detector (Thermo Separation Products). This column was eluted isocratically with methanol/water/acetic acid (65:35:0.01, v/v/v) at 0.2 ml/min into the electrospray probe. The spray voltage was set to 5 kV and the heated capillary to 250°C. Over a 2-sec scan cycle, full-scan mass spectra (MS) were acquired by scanning between m/z 95-410 in the negative ion mode, followed by the acquisition of product ion mass spectra (MS/MS) of the most intense molecular anions (e.g., [M-H] = m/z 351 for LXA₄ and 15-epi-LXA₄).

Cell Incubations. For PMN incubations, freshly isolated PMN (20 × 10⁶/ml) were incubated in Dulbecco's phosphate-buffered saline (PBS⁺⁺) with 5 µM of A23187 for 30 min at 37°C in the presence of 15R-HETE (10 µM) or vehicle (ethanol) alone (Clària and Serhan, 1995). All incubations were stopped at the indicated time intervals by addition of two volumes of cold methanol (1°C). For coincubation of THP-1 and PMN, THP-1 was incubated for 16 hr with LPS (1 µg/ml) at 37°C (Endo et al., 1996; Fu et al., 1990). The cells were then pelleted, suspended (10 × 10⁶/ml) in PBS⁺⁺ and then treated with ASA (300 µM) for 20 min at 37°C and coincubated with freshly isolated PMN at different cell ratios for 5 min at 37°C, followed by addition of 5 µM of A23187 and 20 µM of AA for 30 min at 37°C. These concentrations were selected on the basis of reported values. For example, ASA treatment switches human PGHS-2 oxygenase activity to an R-lipoxygenase activity that peaks at 20 min (Xiao et al., 1997). 15-epi-LXA₄ generation in coincubation of cell types without P450 appears to be solely dependent on ASA acetylation of PGHS-2. This notion is further supported by the finding that statistically significant differences in 15-epi-LXA₄ levels were not observed with ASA concentrations from 100 to 500 µM (Clària and Serhan, 1995), which suggests that acetylation of PGHS-2 is the limiting event in this system and that acetylation of endothelial cell PGHS-2 in intact cells was saturated at these concentrations. Also, it was established that 15R-HETE production by acetylated PGHS-2 is both AA- and time-dependent. For example, 20 µM of AA gave maximum production of 15R-HETE at 3 min, which remained at equivalent levels for 5 to 30 min with recombinant PGHS-2 (Mancini et al., 1997). Thus these conditions [i.e., ASA (300 µM for 20 min) and AA (20 µM for 30 min)] were utilized for THP-1/PMN coincubation experiments to optimize 15-epi-LXA₄ generation. Incubations were terminated and samples were then extracted with solid-phase C18 cartridges (Serhan, 1990). The methyl formate eluants from the solid-phase extraction were taken to dryness with a stream of N₂, and the samples were next suspended in methanol/water (2:1; 100 µl) and taken for ELISA analysis. Serial dilutions of the samples were incubated with standard amounts of HRP-labeled 15-epi-LXA₄ and 15-epi-LXA₄ antiserum in 96-well plates precoated with goat anti-rabbit IgG.

Mouse peritoneal lavage. Balb/c mice with an average weight of 20 g were individually injected within the peritoneum with LPS (1.25 mg/kg b.wt.). After 16 hr, the mice were treated with either 5 mg of indomethacin or vehicle (ethanol), again by i.p. injection, for 30 min before treatment with ASA (125 mg/kg b.wt.) and 0.7 mg of 17-ODYA or ASA alone via i.p. injection for 30 min. In vivo ASA treatment at a dose of 150 mg/kg s.c. or 30 to 300 mg/kg p.o. showed almost complete inhibition of TXB₂ production in mouse (Anton et al., 1990; Molinari et al., 1987). At an i.p. dose of 25 to 100 mg/kg, ASA was also shown to inhibit photochemically induced platelet aggregation in pial microvessels of mouse in vivo (el-Sabban and Radwan, 1997). 17-ODYA is a potent inhibitor of P450 eicosanoid metabolism that does not inhibit either cyclooxygenase or LO activity (Muerhoff et al., 1989). Addition of 17-ODYA (5 µM, 20 min) to A549 cells resulted in about 50% reduction in 15-HETE generation in vitro (Clària et al., 1996). In addition, 17-ODYA was shown to inhibit the metabolism of AA by cytochrome P450 in renal cortical microsomes of rats in a concentration-dependent fashion with IC₅₀ < 100 nM. Even at very high concentrations (5 and 25 µM), 17-ODYA did not inhibit the production of either PGE₂ or PGF₂ (Zou et al., 1994). The results of these experiments guided the amounts of inhibitors and parameters selected for the in vivo experiments. To this end, peritonitis was induced by i.p. injection of 2 ml of 2% casein, which was prepared as described in Yamaki and Oh-ishi (1990). The mice were sacrificed 4 hr later, and peritoneal lavages were collected. For some experiments, the lavage was added directly to two volumes of cold methanol (1°C). In other experiments, the lavages were treated further with indomethacin (300 µM) or vehicle (DMSO) for 5 min and with ASA (300 µM) and 17-ODYA (5 µM) or ASA alone for 5 min before the addition of ionophore A23187 (5 µM) for 30 min at 37°C. The incubations were individually added to two volumes of cold methanol (1°C) and allowed to precipitate at 4°C for 30 to 60 min. The samples were then extracted with solid-phase C18 cartridges, vide supra.

Immunoblots for PGHS-2. THP-1 cells were exposed to LPS (1 µg/ml) for 16 hr at 37°C. Balb/c mice with an average weight of about 20 g were exposed to LPS (1.25 mg/kg b.wt.) for 16 hr by i.p. injection. Cell pellets from THP-1 incubations as well as mouse peritoneal lavage samples were obtained after centrifugation at 1000 rpm for 10 min at room temperature. Individual cell pellets were washed with PBS⁺⁺, freeze-thawed three times in liquid nitrogen and suspended in lysis buffer consisting of 50 mM Tris/HCl, pH 7.5, containing 10 mM EDTA, 1% Triton X-100, 1 mM PMSF, 20 µM leupeptin and 10 µg/ml aprotinin. Equal amounts of protein were subjected to SDS-PAGE and transferred to a polyvinylidene fluoride (PVDF) microporous membrane by electroblotting. Membranes were blocked in 5% nonfat milk in TBST (0.9% NaCl and 0.05% Tween-20 in 20 mM Tris/HCl, pH 7.4) and probed with an anti-human PGHS-2 polyclonal antibody (1:100 dilution) for 1 hr. After washing three times with TBST, membranes were incubated with HRP-linked goat anti-rabbit IgG (1:10,000 dilution) for 1 hr, and the immunoreactive bands were developed by incubating with chemiluminescence substrates and visualized by exposure to an X-ray film.

Statistical analysis. Results were expressed as the mean ± S.E.M. from n = 3. Statistical evaluation of the results was carried out using Student's t test, and P values < .05 were considered statistically significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

ELISA sensitivity and selectivity. To develop a sensitive and selective ELISA for 15-epi-LXA₄, rabbits (denoted numbers 5347 and 5348) were immunized with KLH-linked 15-epi-LXA₄, and antisera were collected and tested for their ability to recognize 15-epi-LXA₄, which was prepared by total organic synthesis. After conditions for ELISA were optimized (see "Materials and Methods"), concentration dependence was evaluated using serial dilutions of unlabeled 15-epi-LXA₄ as a competitor as in figure 1. Fifty percent inhibition of maximum binding with HRP-linked 15-epi-LXA₄ (IC₅₀) was ~0.14 pmol (50 pg) and ~0.21 pmol (75 pg) of 15-epi-LXA₄ for antisera 5347 and 5348, respectively. The minimum amounts detectable (~80% of maximum binding) were ~10 fmol (3.5 pg) and ~21 fmol (7.5 pg) of 15-epi-LXA₄ for 5347 and 5348, respectively (not shown). Because antiserum 5347 displayed higher affinity toward 15-epi-LXA₄, it was selected for further experiments.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Specificity of 15-epi-LXA4 antiserum and direct comparison with LXA4 antiserum. 15-epi-LXA4 (1-1,000 pg), several lipoxin analogs (panel A) or individually related HETEs (panel B) were incubated with standard amounts of HRP-labeled 15-epi-LXA4 and 15-epi-LXA4 antiserum in 96-well plates precoated with goat anti-rabbit IgG as described in "Materials and Methods." On the vertical axis, % B/B0 represents the bound enzyme activity in the presence (B) of unlabeled compound relative to its absence (B0). All data are expressed as the mean ± S.E.M. from n = 3.

LC/MS/MS of synthetic and immunoprecipitated 15-epi-LXA₄. Because the anti-LXA₄ antisera recognized both 15-epi-LXA₄ and LXA₄, with a preference for 15-epi-LXA₄, it was essential also to establish and confirm the physical properties of the synthetic materials used to generate the antisera. Liquid chromatography/mass spectrometry analysis gave UV chromatograms when monitored at 300 nm with single major peaks with retention times of 14.7 and 15.7 min for LXA₄ and 15-epi-LXA₄, respectively, using a LUNA C18-2 column (see "Materials and Methods"). The two epimers do not resolve as free acids with some HPLC columns. The present column and mobile phase made possible about 1 min of separation. The major anion in the MS spectra for these retention times was m/z 351 (fig. 2A and C, cleavage site a), which represented [M-H] for both LXA₄ and 15-epi-LXA₄. Each MS/MS spectrum of LXA₄ and 15-epi-LXA₄ revealed essentially the same fragmentation patterns (fig. 2, B and D). Prominent daughter ions were observed for both at m/z 333 [351-H₂O]; 315 [351-H₂O, -H₂O]; 307 [351-CO₂] (see fig. 2, A and C, cleavage site b); 289 [351-H₂O, -CO₂]; 271 [351-H₂O, -H₂O, -CO₂]; 251 [351-CHO(CH₂)₄CH₃] (see fig. 2, A and C, cleavage site d); 235 [351-CHO(CH₂)₃COOH] (see fig. 2, A and C, cleavage site c'); 233 [d-H₂O]; 215 [d-H₂O, -H₂O]; 207 [d-CO₂]; 189 [d-H₂O, -CO₂]; 135 [d-CHO(CH₂)₃COOH] and 115 [CHO(CH₂)₃COO] (see fig. 2, A and C, cleavage site c). These ions are consistent with those recently reported for LXA₄ analyzed by electrospray/collision-induced dissociation mass spectrometry (Griffiths et al., 1996). Successive MS/MS analysis revealed statistically significant differences in the intensities of certain daughter ions in these two epimers, particularly for m/z 333, 251, 215 and 115 (table 2). Conformational differences between the molecules, arising from opposite stereochemistries at carbon 15, are likely to influence the relative probabilities for each fragmentation, and these properties may therefore serve as an additional means for identification of these compounds (determination of UV absorbance, retention times and intensities of selective MS/MS fragments; see table 2).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   LC/MS/MS of synthetic LXA4 and 15-epi-LXA4. Mass spectra of synthetic LXA4 (panel A) and 15-epi-LXA4 (panel C) and MS/MS analysis of LXA4 (panel B) and 15-epi-LXA4 (panel D) were performed with a Finnigan LCQ LC/MS/MS as described in "Materials and Methods."

Generation of 15-epi-LXA₄ by activated human PMN. To examine the ability of this ELISA to quantitate 15-epi-LXA₄ generated by cellular sources and to qualify its potential uses, we stimulated isolated suspensions of human peripheral blood PMN (10⁷ cells) in the presence or absence of 15R-HETE. Upon activation, approximately 40 ng of 15epi-LXA₄ was generated by 10⁷ PMN exposed to 15R-HETE (fig. 3), whereas only less than ~1 ng of 15-epi-LXA₄ was found in the absence of the exogenous precursor 15R-HETE with PMN from peripheral blood of healthy individuals. These values (i.e., ~40 ng/10⁷ PMN) were in agreement with recently reported values (Clària et al., 1996; Clària and Serhan, 1995) obtained with isolated peripheral blood PMN from healthy volunteers, where 15-epi-LXA₄ production was quantitated by combined RP-HPLC and ELISA analyses. Given the sensitivity of 15-epi-LXA₄ ELISA, we could detect 15-epi-LXA₄ from less than 10⁷ PMN, and it proved to be linear in this range (i.e., dependent on cell number). This cell incubation number, 10⁷ PMN, was selected for these experiments to ensure statistically significant values that facilitated quantitation and established the time course of 15-epi-LXA₄ generation, which minimized the impact of individual donor variations with lower cell numbers. Thus the present ELISA proved useful in detecting 15-epi-LXA₄ generated by cellular sources.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Generation of 15-epi-LXA4 by activated PMN. Human PMN (10 × 106) were incubated with 5 µM of A23187 at 37°C for 30 min in the presence of 15R-HETE (10 µM) or vehicle alone. For the kinetics of 15-epi-LXA4 production (inset), incubations were collected at indicated time-points, and cells were separated from supernatant by rapid microcentrifugation. Incubations were terminated by adding two volumes of ice-cold methanol, and the samples were purified by extraction with solid-phase C18 cartridges. Serial dilutions of the samples were incubated with standard amounts of HRP-labeled 15-epi-LXA4 and 15-epi-LXA4 antiserum in 96-well plates for ELISA analysis as described in "Materials and Methods." The values are expressed as the mean ± S.E.M. from n = 3. Statistical difference was obtained (*P < .01) by comparing values from incubations with stimulated PMN alone and incubations with stimulated PMN plus 15R-HETE.

Generation of 15-epi-LXA₄ by coincubation of PMN and THP-1 cells. Both LX and ATL are predominantly generated via transcellular biosynthesis during cell-cell interaction in mammalian tissues (reviewed in Serhan, 1997). Only a few individual cell types are known to generate lipoxins as a product of a single cell origin, such as activated macrophages isolated from rainbow trout (Pettitt et al., 1991) or endogenously primed PMN isolated from asthmatic patients (Chavis et al., 1996). It was therefore of interest to determine whether single cell types, which possess both PGHS-2 and 5-LO, have the ability to generate 15-epi-LXA₄. Along these lines, the production of 15-epi-LXA₄ by THP-1 cells was evaluated using this new ELISA (fig. 4). PGHS-2 was up-regulated in the THP-1 cells by exposure to LPS (1 µg/ml) for 16 hr (figs. 4 and 5). LPS elicits a marked increase in PGHS-2 mRNA as well as protein levels in macrophages (Morham et al., 1995). Of interest, significant amounts of 15-epi-LXA₄ were not detected in LPS-treated THP-1 cells incubated in the presence of ASA and exogenous AA when compared with values obtained with reagents alone (ASA, AA and A23187). It was recently shown that expression of 5-LO can be up-regulated by GM-CSF in human monocytes (Ring et al., 1996). However, significant increases in 15-epi-LXA₄ production were not observed in THP-1 cells treated with GM-CSF (1 ng/ml) for 48 hr and LPS (1 µg/ml) for 16 hr, although 5-LO transcript levels were increased after this treatment (data not shown). Thus the presence of both 5-LO and PGHS-2 within a single cell type was not sufficient to generate 15-epi-LXA₄.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   15-Epi-LXA4 generation by costimulation of human PMN and THP-1 cells. THP-1 cells (10 × 106) were incubated with LPS (1 µg/ml) for 16 hr and treated with ASA (300 µM) for 20 min before coincubation with or without PMN at a cell ratio of 1:6 (THP-1/PMN) for 30 min at 37°C in the presence of A23187 (5 µM) and AA (20 µM). The coincubations were performed at different cell ratios in the presence or absence of AA (20 µM) or A23187 (5 µM) (inset). The incubations were terminated, and the samples were extracted for ELISA analysis as in figure 1. The values are expressed as the mean ± S.E.M. from n = 3. Statistical differences were obtained when compared to values from incubations at a cell ratio of 1:0 (*P < .01, **P = .01) and compared with values from incubations at a cell ratio of 1:6 with vehicle alone (+P < .01, ++P = .04).

View larger version (47K): [in this window] [in a new window]   Fig. 5.   PGHS-2 is expressed by THP-1 cells and mouse peritoneal lavage cells: Western blot analysis. Mice were injected with LPS (1.25 mg/kg b.wt.). THP-1 cells were treated with LPS (1 µg/ml) at 37°C. Sixteen hours later, the THP-1 cells and cells from peritoneal lavage were collected and suspended in lysis buffer. The whole-cell lysates (20 µg/lane) were subjected to SDS-PAGE and transferred to PVDF membrane by electroblotting. The blot was probed using an anti-human PGHS-2 polyclonal antibody (1100 dilution) and developed as described in "Materials and Methods." Lane 1 contains cell lysates from untreated peritoneal lavage, Lane 2 the cell lysates from LPS-treated peritoneal lavage. Lane 3 contains the untreated THP-1 cells, Lane 4 the LPS-treated THP-1 cells. Molecular weight markers are indicated by arrows.

Generation of 15-epi-LXA₄ by mouse peritoneal exudates. To determine whether 15-epi-LXA₄ can be detected with animal models, experiments were carried out using a mouse peritonitis model. In this model, PGHS-2 protein levels were shown to be up-regulated by i.p. injection of LPS. Peritonitis was induced by i.p. injection of casein (as in Yamaki and Oh-ishi, 1990). In these experiments, up-regulation of PGHS-2 was also demonstrated by Western blot analysis, and the immunoreactive bands were observed at ~70 kDa in peritoneal lavage samples from LPS-treated mice (fig. 5, lanes 1 and 2). The molecular sizes of the PGHS-2 proteins were found to be slightly different between mouse peritoneal cell lysates and human THP-1 cells (fig. 5, lane 4). Because the size of PGHS-2 protein in THP-1 cells was the same as that in Sf9 cells overexpressing human PGHS-2 (data not shown), it is possible that PGHS-2 from human and mouse were glycosylated to a different extent. Along these lines, it has been demonstrated that mouse PGHS-2 has four N-glycosylation sites and that their molecular sizes varies from 65 to 74 kDa as a result of N-glycosylation heterogeneity (Otto et al., 1993).

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Peritoneal inflammatory exudates from ASA-treated mice generate 15-epi-LXA4. Mice were injected with LPS (1.25 mg/kg b.wt.) to induce PGHS-2, and 16 hr later they were treated with vehicle or (panel a) one bolus dose of ASA (0.125 g/kg b.wt.) by i.p. injection for 30 min before casein-initiated induction of neutrophil infiltration for 4 hr. b) Consecutive doses of ASA were given in a separate group of animals that received ASA (0.125 g/kg b.wt.) by i.p. injection 30 min before casein injection and before sacrificing the mice. The lavage exudates were collected and stimulated with 5 µM A23187 (30 min, 37°C). The values are expressed as the mean ± S.E.M. from three independent experiments. The values from both treatment groups were statistically different (*P = .05 for group a, **P = .03 for group b) from those obtained with animals that did not receive ASA (vehicle). The samples were extracted as described in "Materials and Methods." b.wt., body weight; IP, i.p. injection.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Indomethacin but not 17-ODYA inhibits 15-epi-LXA4 generation by mouse peritoneal exudates. Mice were injected with LPS (1.25 mg/kg b.wt.). Sixteen hours later, they were treated with ASA (0.125 g/kg b.wt.) and with 5 mg indomethacin, ASA (0.125 g/kg b.wt.) and 0.7 mg 17-ODYA or with ASA alone before induction of PMN infiltration by i.p. injection of casein. The peritoneal lavage samples (18 × 106 PMN/mouse) were incubated further with 300 µM ASA and indomethacin, 300 µM ASA and 5 µM 17-ODYA or ASA alone and stimulated with A23187 (5 µM) as described in "Materials and Methods." The reactions were stopped and the samples extracted. The values are expressed as the mean ± S.E.M. from three independent experiments. Values for indomethacin plus ASA treatment were statistically different from ASA treatment alone (*P < .01) and ASA plus ODYA (+P = .02). b.wt., body weight; IP, i.p. injection.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Earlier, we developed a LXA₄ ELISA that was both sensitive and selective for LXA₄ when directly compared with other eicosanoids including cross-reactivity with LXB₄ (Levy et al., 1993). Of particular interest, this antibody was later found to recognize both the recently identified 15-epi-LXA₄ and native LXA₄ (data not shown), which have been found to be generated by different biosynthetic routes, 15-epi-LXA₄ giving greater biopotencies in several systems. The physical properties of these biologically derived epimers are very similar. Indeed, MS/MS analysis of LXA₄ and 15-epi-LXA₄ indicated that both give essentially identical ions upon fragmentation (fig. 2B and D), differing only in the relative intensities of certain daughter ions (table 2). Therefore, we developed a new ELISA using a 15-epi-LXA₄ antibody that recognizes the newly identified 15-epi-LXA₄ (IC₅₀  50 pg, minimum detection = 3.5 pg) in the picogram range (see fig. 1). This antibody also proved to be stereoselective toward 15-epi-LXA₄ in that it recognizes the carbon 15 alcohol in the R configuration and distinguishes between LXA₄ and 15-epi-LXA₄.

Antisera used in the present experiments were raised with synthetic compounds of absolute stereochemistry. The specificity of anti-15-epi-LXA₄ antiserum was clearly demonstrated with eicosanoids that lacked either a C-15 alcohol group (i.e., 5S-HETE and 12S-HETE) or the tetraene structure (i.e., 15R-HETE and 15S-HETE). Also, antibody recognition was reduced when a methyl group was introduced at C-15 to replace hydrogen (i.e., 15-R/S-methyl-LXA₄) at carbon 15. Eicosanoids carrying a 15S-alcohol group (i.e., LXA₄, LXB₄ and 15S-HETE) were not recognized by this antibody (fig. 1, A and B). In earlier reports, 15-epi-LXA₄ was identified by RP-HPLC combined with gas chromatography analyses of derivatized materials, which distinguish 15-epi-LXA₄ from native LXA₄ by their different retention times in each chromatography system (Clària et al., 1996; Clària and Serhan, 1995). In order to achieve base-line separation, methyl esters of both LXA₄ and 15-epi-LXA₄ and derivatization were required. Derivatization could lead to quantity losses of the compounds of interest during sample work-up. The present results indicated that this ELISA is at least 200 times more sensitive than UV-monitored RP-HPLC (minimum UV detection ~ 1 ng). Thus the present ELISA proved to be highly sensitive and stereoselective and is a rapid and relatively inexpensive procedure for monitoring 15-epi-LXA₄.

Results from kinetic experiments performed to evaluate 15-epi-LXA₄ production by PMN suggest that the activation of 5-LO in PMN displays a similar time course in converting either AA (Krump and Borgeat, 1994; Pouliot et al., 1994) or 15R-HETE (fig. 3). PMN stimulated with the divalent cation ionophore A23187 gives a sustained release of both arachidonate and leukotriene B₄ formation evident within 1 min that plateaus by 3 min with washed PMN in suspension (Krump and Borgeat, 1994). This time course is similar in the case of 15-epi-LXA₄ generation by PMN (fig. 3 inset).

Generation of 15-epi-LXA₄ by THP-1 cells and leukocyte-leukocyte interactions was also investigated. Because THP-1 cells possess both PGHS-2 and 5-LO, it was possible in theory (see fig. 8) that this cell type could generate 15-epi-LXA₄ without requiring cell-cell interactions and transcellular biosynthesis. THP-1 cells incubated alone did not, however, produce detectable amounts of 15-epi-LXA₄ (fig. 4). In this particular cell line, it is possible that functional 5-LO is not located at appropriate intracellular sites within the THP-1 cells to interact with PGHS-2-derived endogenous 15R-HETE. In sharp contrast, in the presence of human PMN that possess functional 5-LO, 15-epi-LXA₄ generation was increased by more than 10-fold after ASA treatment (fig. 4). This value for 15-epi-lipoxin generation is in the range of, and comparable to, those generated during either PMN-vascular endothelial or PMN-airway epithelial cell interactions (Clària et al., 1996; Clària and Serhan, 1995).

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Proposed scheme for generating 15-epi-LXA4 by murine inflammatory exudates. Arachidonic acid is converted to 15R-HETE predominantly by ASA-acetylated PGHS-2 in macrophage and epithelial cells. Cytochrome P450 might make only a minor contribution to 15R-HETE generation in the mouse. 15R-HETE then undergoes transcellular conversion by 5-LO in PMN to 15-epi-LXA4 via a 15-epi-5(6)-epoxytetraene intermediate.

15-epi-LXA₄ was also generated from endogenous sources of AA (11-14 nM) within levels required to evoke bioactivity (Clària et al., 1996; Clària and Serhan, 1995; Serhan et al., 1995), because 15-epi-LXA₄ inhibits PMN adhesion to vascular endothelial cells at levels as low as 0.1 nM in vivo (Clària and Serhan, 1995; also see Scalia et al., 1997). Together, these are the first results demonstrating that 15-epi-LXA₄ not only is generated during endothelial or epithelial cell-leukocyte (neutrophil) interactions, but also is generated during leukocyte-leukocyte interactions, as exemplified here by results obtained with THP-1 cells (a monocytic leukemia cell line) activated with human PMN.

It is of interest that 15-epi-LXA₄ was generated by exudate inflammatory cells taken from a casein-induced murine peritonitis model (fig. 6). In this complex inflammatory milieu, 15-epi-LXA₄ was formed and was probably the product of PMN carrying 5-LO activity (PMN represent 73% of total lavage cells) interacting with acetylated PGHS-2 of macrophages. LPS induces PGHS-2 activity in mouse peritoneal macrophages (Reddy and Herschman, 1994). It also induces TNF- gene expression (Goldfeld et al., 1990), which could have contributed to the up-regulation of PGHS-2. The macrophages represented ~10% of the total cells within the inflammatory exudates. In theory, it is also possible that epithelial or other cell types of the peritoneal cavity could have contributed 15R-HETE to produce 15-epi-LXA₄. Upon stimulation of exudate cells from ASA-treated mice, 15-epi-LXA₄ was clearly generated in an ASA-dependent fashion (fig. 6). Here, small amounts of 15-epi-LXA₄ were detected without ASA treatment of the mice; they could have been the product of PGHS-2 acetylated by potential endogenous acetylating agents (see fig. 8). It is well known that acetylation of histone, as a part of post-translational modification, is important for transcription, assembly of chromatin and maintenance of cell differentiation (Wade et al., 1997). Thus other PGHS-2 endogenous acetylating routes may exist within the mouse peritoneal cavity, or, of equal likelihood, small amounts of 15R-HETE could be produced by origins other than PGHS-2 or P450 that remain to be identified (vide infra).

LPS was used in THP-1/PMN coincubations and in a mouse peritonitis model to induce PGHS-2 (fig. 5, lanes 2 and 4). LPS was shown to induce PGHS-2 activity in human monocytes, reaching maximum values within 12 to 24 hr after treatment (Endo et al., 1996; Fu et al., 1990), and a similar induction time course was observed in mouse peritoneal macrophages (Reddy and Herschman, 1994). PGHS-2 can also be up-regulated by other stimuli, such as growth hormones, cytokines, serum and phorbol esters (Herschman, 1996), in a wide range of cell types. It was shown that, in HUVEC/PMN and A549/PMN coincubations, formation of 15-epi-LXA₄ was enhanced by IL-1, which induced PGHS-2 in HUVEC and A549 cells (Clària et al., 1996; Clària and Serhan, 1995). In addition, IL-1 and TNF- were shown to induce PGHS-2 in HT-29 C1.19A cells (human enterocytes), which, in the presence of ASA, generated 15R-HETE (a precursor of 15-epi-LXA₄ biosynthesis) (Gronert et al., 1998). Therefore, it appears that other specific cytokines that can regulate PGHS-2 appearance in vivo (i.e., TNF- and IL-1) should also trigger 15-epi-LXA₄ generation in THP-1/PMN coincubations as well as in murine peritonitis induced by cytokines when ASA is administered, because acetylation of PGHS-2 is a required step in this pathway for 15R-HETE formation.

In the mouse, several other origins of substrate may contribute to 15-epi-LXA₄. For example, it is also possible that racemic 15-HETE was generated either by epithelial cytochrome P450 (Capdevila et al., 1986) or via nonenzymatic pathways involving reactive oxygen species released by activated PMN (Fridovich and Porter, 1981). Without stimulation of the inflammatory exudates, the amounts of 15-epi-LXA₄ generated by ASA-treated mice were in a relatively low range (0.3-0.6 ng, 0.17-0.34 nM in each individual mouse) (data not shown). These values could reflect the long exposure of the mouse peritoneum to high levels of LPS in vivo that could induce the production of proinflammatory cytokines or other factors that might uncouple and/or metabolically inactivate 15-epi-LXA₄ transcellular biosynthesis. Eicosanoids are local mediators and are well known to be rapidly generated and degraded by further transformation. Nevertheless, 15-epi-LXA₄ was detected within peritonitis samples, yet with additional ex vivo treatment of exudates with ASA, the amounts of 15-epi-LXA₄ were increased to ~10 ng per 5 ml of the peritoneal lavage from each mouse (fig. 7). Because further ASA treatment enhanced the amounts of 15-epi-LXA₄ generated, the levels found in vivo may reflect incomplete PGHS-2 acetylation by ASA and/or degradation of 15-epi-LXA₄ by further conversion and metabolism or, which is more likely, absorption of the compound by the surrounding peritoneal tissues during the time course of the experimental peritonitis. It should also be noted that ASA undergoes rapid hydrolysis and deacetylation in aqueous medium and may have done so, to some extent, in the present experiment when ASA was injected into the mouse peritoneal cavity. Given these considerations, 15-epi-LXA₄ was produced and generated in an ASA-dependent fashion from endogenous sources within a range commensurate with its currently known bioactions.

We also tested the relative contributions of ASA-acetylated PGHS-2 and/or P450 to the production of 15-epi-LXA₄ in the mouse inflammatory exudates (fig. 7) for comparison with the pathways observed with human tissues (Clària et al., 1996; Clària and Serhan, 1995). Oxygenation of AA by cytochrome P450 results in ~40% R and 60% S configuration of 15-HETE (Capdevila et al., 1986). In addition, cytochrome P450 contributes to 15-HETE generation by human adenocarcinoma-derived A549 cells (Clària et al., 1996) and results in 15-epi-LXA₄ generation during A549-PMN costimulation. In contrast, in this mouse model, 15-epi-LXA₄ formation was not inhibited by the P450 inhibitor (17-ODYA). The present results also show that indomethacin, which specifically blocks PGHS-2 acetylation in vitro (Mancini et al., 1997), inhibited 15-epi-LXA₄ production by mouse tissues, which suggests that 15-epi-LXA₄ formation in this scenario occurs mainly via cyclooxygenase-dependent mechanisms. Taken together, our results suggest that 15-epi-LXA₄ formation by mouse peritoneal inflammatory exudates occurs predominantly via an ASA- and PGHS-2-dependent route (figs. 7 and 8). This is the first demonstration of 15-epi-LXA₄ generation in a model of murine peritonitis.

It was previously shown that lipoxins are generated in several species of experimental animals. For example, LXA₄ is generated in mouse kidney with glomerulonephritis in a P selectin-dependent fashion (Mayadas et al., 1996). Also, LXA₄ and LXB₄ are both formed in ischemic rat brain (Kim and Tominaga, 1989). LXA₄ was recently found to regulate LTB₄-mediated delayed hypersensitive reactions in guinea pig (Feng et al., 1996) and to inhibit infiltration of neutrophils to glomerulonephritic kidney in rats (Papayianni et al., 1995). For the newly identified 15-epi-LXA₄, we have reported in preliminary experiments that 15-epi-LXA₄ was generated in rat kidney in ASA-dependent fashion (Badr et al., 1996). Together with the findings of the present studies, namely that 15-epi-LXA₄ was generated in a model of murine peritonitis, it is likely that 15-epi-LXA₄ can also be generated upon ASA treatment in other experimental animals in vivo where cell-cell interactions are accelerated and PGHS-2 is up-regulated (e.g., during the inflammation process). These results also raise the possibility that other PGHS-2 endogenous acetylating routes may exist in the mouse peritoneal cavity that give rise to 15-epi-LXA₄ (see fig. 8).

In summary, we developed a new ELISA for 15-epi-LXA₄ that proved highly sensitive as well as stereoselective, compared with its epimer LXA₄ at the level required to interact selectively with 15-epi-LXA₄. Utilizing this ELISA, we established that 15-epi-LXA₄ generation proceeds via transcellular biosynthesis during heterotypic leukocyte-leukocyte interactions. Furthermore, this is the first evidence for 15-epi-LXA₄ generation by inflammatory exudate cells from a murine model with LPS and casein-induced peritonitis. Together, these findings document useful means to investigate 15-epi-LXA₄ generation further.