Structural Requirements for Activation of the 5-Oxo-6E,8Z, 11Z,14Z-eicosatetraenoic Acid (5-Oxo-ETE) Receptor: Identification of a Mead Acid Metabolite with Potent Agonist Activity

Metabolism of arachidonic acid by the 5-lipoxygenase (5-LO) pathway leads to the formation of leukotriene (LT) B₄, LTC₄, LTD₄, and 5-HETE (Funk, 2001). LTB₄, acting through the BLT₁ receptor, is a potent activator of neutrophils and lymphocytes. LTD₄ interacts with the cysteinyl-LT₁ and cysteinyl-LT₂ receptors to stimulate smooth muscle contraction, cytokine release from leukocytes, and various other responses. Although 5-HETE itself has only relatively weak biological activities, it is oxidized to the potent granulocyte chemoattractant 5-oxo-ETE by the action of 5-hydroxyeicosanoid dehydrogenase (5-HEDH) (Powell and Rokach, 2005). 5-HEDH is a microsomal enzyme that is dependent on NADP⁺ as an electron acceptor. This limits its activity because NADP⁺, in contrast to its reduced form NADPH, is only present at low concentrations within cells. Thus, 5-oxo-ETE synthesis from arachidonic acid requires both activation of 5-LO in conjunction with increased NADP⁺ levels, which can be induced by oxidative stress and, in phagocytic cells, by activation of the respiratory burst. 5-HEDH is present in most types of leukocytes (Powell and Rokach, 2005) as well as endothelial (Erlemann et al., 2006) and epithelial cells (Erlemann et al., 2007).

5-Oxo-ETE induces a variety of rapid responses in neutrophils and eosinophils, including calcium mobilization, actin polymerization, CD11b expression, and L-selectin shedding (Powell and Rokach, 2005). It also stimulates superoxide production and degranulation in neutrophils primed with tumor necrosis factor-α (O'Flaherty et al., 1994) or granulocyte macrophage colony-stimulating factor (O'Flaherty et al., 1996). It is a potent chemoattractant for both neutrophils (Powell et al., 1993) and eosinophils (Powell et al., 1995a) and induces transendothelial migration of eosinophils (Dallaire et al., 2003). It also has modest chemoattractant effects on monocytes and stimulates actin polymerization (but not calcium mobilization) (Sozzani et al., 1996) and granulocyte macrophage colony-stimulating factor release (Stamatiou et al., 2004) in these cells. When administered in vivo to Brown Norway rats by intratracheal instillation, 5-oxo-ETE elicits pulmonary eosinophilia (Stamatiou et al., 1998). In humans, it induces the infiltration of both eosinophils and neutrophils into the skin (Almishri et al., 2005). In addition to its effects on inflammatory cells, 5-oxo-ETE has been reported to block the induction of apoptosis in prostate tumor cells by MK886 (Ghosh and Myers, 1998) and selenium (Ghosh, 2004).

The biological effects of 5-oxo-ETE are mediated by the G_i protein-coupled oxoeicosanoid (OXE) receptor (Hosoi et al., 2005). This receptor is expressed on eosinophils, neutrophils, and monocytes (Hosoi et al., 2002; Jones et al., 2003) and has also been reported to be present on prostate tumor cells but not on normal prostate epithelial cells (Sundaram and Ghosh, 2006). Metabolism of 5-oxo-ETE by a variety of pathways, including reduction back to 5-HETE by 5-HEDH, reduction of the 6,7-double bond by a Δ⁶-reductase, conversion to 5-oxo-12-HETE by 12-lipoxygenase, and ω-oxidation by LTB₄ 20-hydroxylase, results in dramatic reductions in OXE-receptor agonist activity (Powell and Rokach, 2005). However, the relationship of biological activity to carbon chain length and the presence or absence of the Δ⁸, Δ¹¹, and Δ¹⁴ double bonds is unknown. Because such knowledge is critical for the design of potent OXE-receptor agonists and antagonists, we prepared a series of 5-oxo-ETE analogs with different chain lengths and numbers of double bonds and investigated their potencies in eliciting biological responses in neutrophils and eosinophils.

Materials and Methods

Oxygenated Fatty Acids. 5-Oxo-ETE was prepared by chemical synthesis as described previously (Khanapure et al., 1998). 5-Oxo-20:5 was prepared by incubation of the corresponding 5-hydroxy compound (5-HEPE; obtained from Cayman Chemical, Ann Arbor, MI) with a microsomal fraction from human neutrophils in the presence of NADP⁺ as described previously (Powell et al., 1995b). All of the other 5-oxo fatty acids shown in Fig. 1 as well as 5-hydroxy-20:3 were prepared by total chemical synthesis using methods that will be published separately. The structures were confirmed by nuclear magnetic resonance and mass spectrometry. All compounds were purified by reversed-phase high-performance liquid chromatography (RP-HPLC) within 1 month of use. 13-HODE was prepared by oxidation of linoleic acid with soybean lipoxygenase type 1B (Sigma-Aldrich, St. Louis, MO) (Hamberg and Samuelsson, 1967).

Other Reagents. Phorbol 12-myristate 13-acetate (PMA) was purchased from Sigma-Aldrich, whereas NADP⁺ and A23187 were obtained from Roche Diagnostics (Laval, QC, Canada) and Calbiochem (La Jolla, CA), respectively. Dimethyl sulfoxide was purchased from Fisher Scientific (Nepean, ON, Canada).

Preparation of Neutrophils. Neutrophils were purified from whole blood as described previously using dextran 500 (from Leuconostoc; Sigma-Aldrich) to remove red blood cells followed by centrifugation over Ficoll-Paque (GE Healthcare, Baie d'Urfe, QC Canada) to remove mononuclear cells and hypotonic lysis of any remaining red blood cells (Powell et al., 1992). The neutrophils were suspended in phosphate-buffered saline (PBS).

Measurement of Calcium Mobilization in Neutrophils. Intracellular calcium was measured in neutrophils loaded with indo-1 acetoxymethyl ester (Invitrogen, Carlsbad, CA) as described previously (Powell et al., 1996). The cells were washed and resuspended in PBS. Five minutes before commencing data acquisition, Ca²⁺ and Mg²⁺ were added to give final concentrations of 1.8 and 1 mM, respectively. After stabilization of the baseline, fluorescence was measured using a Deltascan 4000 spectrofluorometer (Photon Technology International, Birmingham, NJ) with a temperature-controlled cuvette holder equipped with a magnetic stirrer.

Evaluation of CD11b Expression in Neutrophils. CD11b expression was evaluated by flow cytometry (Powell et al., 1997) using a FACSCalibur instrument (BD Biosciences, San Jose, CA). Unfractionated leukocytes, prepared by treatment of whole blood with dextran followed by hypotonic lysis as described above, were incubated for 10 min with various concentrations of 5-oxo fatty acids. The cells were then stained with fluorescein isothiocyanate-labeled anti-CD11b (Beckman Coulter, Fullerton, CA), and neutrophils were identified on the basis of their forward- and side-scatter properties.

Measurement of Actin Polymerization in Eosinophils. F-actin was measured using unfractionated leukocytes prepared as described above. The leukocytes were first treated with PC5-labeled anti-CD16 (Beckman Coulter) to label neutrophils and then incubated with 5-oxo fatty acids for 20 s. The cells were then fixed with formaldehyde and stained by treatment with a mixture of lysophosphatidylcholine and NBD-phallacidin (Invitrogen) as described previously (Monneret et al., 2002). F-actin was then measured by flow cytometry in eosinophils, which were identified as a population of cells with high side scatter and low expression of CD16.

Neutrophil Migration. Neutrophil migration was measured using 48-well microchemotaxis chambers (Neuro Probe Inc., Cabin John, MD) and Sartorius cellulose nitrate filters (8-μm pore size, 140-μm thickness; Neuro Probe Inc.) as described previously (Powell et al., 1996). 5-Oxo fatty acids were added to the bottom wells, and neutrophils were added to the top wells. After 2 h, the filters were removed and stained, and the numbers of cells on the bottom surfaces were counted in five different fields at a magnification of 400× for each incubation, each of which was performed in triplicate.

Analysis of Metabolites of Mead Acid. A suspension of neutrophils (2 × 10⁶ cells/ml) in PBS containing 1.8 mM Ca²⁺ and 1 mM Mg²⁺ was incubated with Mead acid (50 μM; Cayman Chemical) in the presence of A23187 (5 μM) and PMA (50 nM). After various times, the incubations were terminated by addition of methanol (0.65 ml) containing 0.15% trifluoroacetic acid and cooling to 0°C. After addition of 13-HODE (100 ng) as an internal standard, the products were analyzed by automated precolumn extraction coupled to RP-HPLC (Powell, 1987) using a Waters Alliance system (Waters, Milford, MA). Products were quantitated by comparing the areas of their peaks of UV absorbance at their λ_max with that of 13-HODE. The conditions for preparative experiments were similar, with the exception that 10⁸ neutrophils in 20 ml were used, and the products were extracted manually on a C₁₈ Sep-Pak (Waters) (Powell, 1980). A Novapak C18 column (3.9 × 150 mm; Waters) was used as the stationary phase, whereas the mobile phase was a linear gradient over 40 min between water/acetonitrile/methanol/acetic acid (60:30: 10:0.02) and water/acetonitrile/methanol/acetic acid (10:38:52:0.02) with a flow rate of 1 ml/min.

Mass Spectrometry. Mead acid metabolites were identified by Fourier transform mass spectrometry (FTMS) as described previously (Erlemann et al., 2007) using an IonSpec 7.0 Tesla instrument. MS/MS experiments were performed after isolating the monoisotopic parent ion using 500 ms of sustained off-resonance (1 kHz) irradiation collision-induced dissociation and a 90-ms nitrogen gas pulse.

Data Analysis. With the exception of EC₂₀ and IC₅₀ values, the values are expressed as the means ± S.E. of data from n independent experiments, as indicated in the figure legends. The concentration-response data are normalized and are expressed as percentages of the response to 1 μM 5-oxo-ETE for each individual experiment. “EC₂₀” values are the concentrations of 5-oxo fatty acids that induced a response equal to 20% of the response to 1 μM 5-oxo-ETE (which was the maximal response to 5-oxo-ETE in the case of actin polymerization). IC₅₀ values are the concentrations of 5-oxo fatty acids that inhibited the response to 5-oxo-ETE (10 nM) by 50%. Both EC₂₀ and IC₅₀ values are presented as geometric means with the 95% confidence intervals shown in brackets. The statistical significance of differences in EC₂₀ and IC₅₀ values and maximal responses was determined by one-way repeated-measured analysis of variance using the Bonferroni test as a multiple comparison method. Although it was not possible to test all of the compounds in the same experiment, a complete concentration-response curve to 5-oxo-ETE was done in each case. For the purposes of statistical analysis, the results for each of the 5-oxo fatty acids were compared with those for 5-oxo-ETE only from the same experiments.

Results

Effects of 5-Oxo-ETE Analogs on Calcium Mobilization in Neutrophils. We investigated the effects of the 5-oxo-ETE analogs illustrated in Fig. 1 on cytosolic calcium levels in indo-1-loaded human neutrophils. Various concentrations of 5-oxo-ETE analogs were first added, followed 90 s later by 5-oxo-ETE (10 nM) to determine whether the initially added compound induced homologous desensitization or antagonized the response to 5-oxo-ETE. Digitonin was then added to determine the maximal fluorescence response in the presence of a saturating concentration of calcium (Fig. 2). When 5-oxo-ETE was added 90 s after addition of vehicle, a strong calcium response was observed (Fig. 2A). This was diminished only slightly by prior addition of 5-oxo-14:2 (10 μM), which did not itself induce any detectable calcium transient (Fig. 2B). In contrast, 5-oxo-16:2 (10 μM) elicited a modest calcium transient and completely abolished the response to subsequent addition of 5-oxo-ETE (Fig. 2C). One thousand-fold lower concentrations of 5-oxo-20:3 (Fig. 2D) and 5-oxo-20:4 (i.e., 5-oxo-ETE; Fig. 2E) induced nearly identical strong calcium responses and dramatically reduced the response to 5-oxo-ETE when it was added 90 s later.

The concentration-response curves for the effects of 5-oxo-ETE and other 5-oxo fatty acids on calcium mobilization are shown in Fig. 3. All of the compounds tested had 5-oxo substituents, but, for simplicity, only the carbon chain lengths and numbers of double bonds are indicated by the labels. The response to 5-oxo-ETE itself seems to be biphasic, tending toward a plateau at 1 μM but then increasing substantially at the highest concentration tested (10 μM) (Fig. 3A). A number of the other analogs tested also exhibited this type of concentration-response relationship, in which a plateau did not seem to have been reached by a concentration of 10 μM. For this reason, it was not possible to calculate precise EC₅₀ values. Instead, EC₂₀ values were calculated, but there was still considerable variability for some compounds that had low maximal responses (Table 1).

Figure 3A shows the effects on calcium mobilization of a series of 5-oxo-6-trans-8-cis dienoic acids between 12 and 20 carbons compared with that of 5-oxo-ETE. Among these dienoic acids, 5-oxo-18:2 gave the strongest response and was approximately 20 times more potent than 5-oxo-20:2. 5-Oxo-16:2 was much less active and seemed to have a considerably reduced maximal response compared with the longer-chain compounds. Neither 5-oxo-14:2 nor 5-oxo-12:2 induced a detectable calcium response. In Fig. 3B, the effects of 20-carbon 5-oxo fatty acids containing different numbers of double bonds are shown. 5-Oxo-20:3 is nearly as active as 5-oxo-ETE, but the response drops off considerably as the number of double bonds is reduced, with 5-oxo-20:1 having only weak agonist activity, and 5-oxo-20:0 displaying no detectable activity at the highest concentration tested (10 μM). The effects of three different 5-oxo-20:3 isomers and two 5-oxo-monoenoic fatty acids are compared in Fig. 3C. The responses to 5-oxo-20:3 (Δ^6E,8Z,11Z) and 8-trans-5-oxo-20:3 (Δ^6E,8E,11Z) are quite similar, with the exception that the 8-cis isomer tends to elicit a stronger response at lower concentrations, whereas the 8-trans isomer seems to induce a slightly higher maximal response. As observed for the dienoic acid series, 5-oxo-18:1 induces a stronger response than its C₂₀ homolog.

Desensitization of 5-Oxo-ETE-Induced Calcium Mobilization by 5-Oxo Fatty Acids. The ability of each of the 5-oxo fatty acids to desensitize neutrophils to subsequent addition of 5-oxo-ETE (10 nM) was investigated. The rank order of potency for desensitization was similar to that for agonist activity. However, in this case, all of the analogs induced complete desensitization to 5-oxo-ETE, with the exception of 5-oxo-12:2, 5-oxo-14:2, and 5-oxo-20:0. Parallel concentration-response curves were obtained with well defined, maximal responses, permitting calculation of IC₅₀ values (see Fig. 6). 5-Oxo-ETE, 5-oxo-20:3, 8-trans-5-oxo-20:3, and 5-oxo-18:2 all had very similar concentration-response curves (Fig. 3, D–F) with IC₅₀ values of approximately 3 nM. Δ^6E,8Z,14Z-5-Oxo-20:3 and 5-oxo-20:2 were approximately five to eight times less potent than 5-oxo-ETE in inducing desensitization to 5-oxo-ETE, whereas 5-oxo-16:2, 5-oxo-18:1, and 5-oxo-20:1 were approximately 100 to 150 times less potent. Only very modest responses were observed for 5-oxo-14:2 and 5-oxo-20:0 at the highest concentration tested (10 μM), whereas 5-oxo-12:2 was without any detectable activity (Fig. 3, D and E).

Up-Regulation of CD11b by 5-Oxo-ETE Analogs. The effects of 5-oxo fatty acids on the surface expression of CD11b by neutrophils were also examined. In general, the results were similar to those obtained for calcium mobilization, with 5-oxo-ETE, 5-oxo-20:3, 8-trans-5-oxo-20:3, and 5-oxo-18:2 all being potent inducers of CD11b expression (Fig. 4, A–C) with EC₂₀ values between approximately 2 and 5 nM (Table 1). The response to 5-oxo-ETE tended toward a plateau at 1 μM but then increased sharply at 10 μM (Fig. 4A). This pattern was observed to an even greater extent for 8-trans-5-oxo-20:3 (Fig. 4C) but not for 5-oxo-20:2 (Fig. 4A). 5-Oxo-20:2 was a better inducer of CD11b expression (EC₂₀, 10 nM) than calcium mobilization (EC₂₀, 38 nM) (Fig. 4A). However, it induced a maximal response that was clearly lower than that to 5-oxo-ETE. We also investigated the effect of 5-oxo-20:5 on CD11b expression and found it to be slightly less potent than 5-oxo-ETE (Fig. 4B). Unlike all of the other 5-oxo fatty acids, we prepared 5-oxo-20:5 biosynthetically by incubating 5-HEPE with neutrophil microsomes, and we did not have sufficient quantities to test concentrations higher than 1 μM. 5-Oxo-18:1 was nearly 10 times more potent than 5-oxo-20:1 in inducing CD11b expression (Fig. 4C).

Effects of 5-Oxo Fatty Acids on Actin Polymerization in Eosinophils. The potencies of 5-oxo fatty acids in inducing actin polymerization in eosinophils were determined using flow cytometry. These cells were identified in unfractionated leukocyte preparations on the basis of minimal labeling with anti-CD16 and high side scatter. The use of unfractionated leukocytes has the advantage of reducing the number of manipulations to which the cells are subjected, thus lessening the chances of activation during cell preparation. Although it is theoretically possible that the responses of eosinophils could have been mediated by the release of stimulating factors from neutrophils, this would seem unlikely because of the very rapid actin polymerization response, which was determined after incubation with agonists for only 20 s.

In contrast to their effects on calcium mobilization and CD11b expression, the responses to most of the compounds tested reached plateaus. However, the maximal responses to some of the analogs were considerably lower than that to 5-oxo-ETE, which increased the levels of F-actin (i.e., polymerized actin) to 75 ± 5% above the basal level. In contrast, 5-oxo-16:2 only increased F-actin levels to 23 ± 3% above baseline (p < 0.001; Fig. 5A), whereas the maximal responses observed for 5-oxo-20:1 and 5-oxo-18:1 were 33 ± 3% (p < 0.001) and 48 ± 2% (p < 0.001) above baseline, respectively (Fig. 5C). The maximal responses to Δ^6E,8Z,14Z-5-oxo-20:3 (43 ± 18% above control; p = 0.001) and 5-oxo-20:2 (37 ± 5% above control; p < 0.001) were also lower than that to 5-oxo-ETE by approximately 40% (Fig. 5, C and A, respectively).

Because of the variability in the maximal responses among the different analogs, EC₂₀ rather than EC₅₀ values were determined (Table 1). The effects of carbon chain length on potency are shown in Fig. 6A. The corresponding IC₅₀ values for inhibition of 5-oxo-ETE-induced calcium mobilization (cf. Fig. 3D) are shown for comparison. Among the 5-oxo-dienoic acids tested, 5-oxo-18:2 was the most potent, having an EC₂₀ (1.9 nM) significantly lower than that of its C₂₀ homolog 5-oxo-20:2 (EC₂₀, 10 nM; p < 0.02) and slightly lower than that of 5-oxo-ETE (EC₂₀, 3.1 nM; N.S.) (Fig. 6A). Similar results were obtained for the desensitization by 5-oxo fatty acids of 5-oxo-ETE-induced calcium mobilization in neutrophils, with 5-oxo-ETE and 5-oxo-18:2 being more potent than 5-oxo-20:2. Of the two monoenoic acids tested, the C₁₈ fatty acid (5-oxo-18:1; EC₂₀, 132 nM) was also more potent in stimulating actin polymerization than the corresponding C₂₀ compound (5-oxo-20:1; EC₂₀, 289 nM; p < 0.05) (Fig. 6A).

The relationship between the number of olefinic double bonds and agonist potency is shown in Fig. 6B. At least one double bond is required for appreciable agonist activity, because 5-oxo-20:0 has only a very modest effect at the highest concentration tested (10 μM) (Fig. 5B). In this series, the most potent substance was 5-oxo-20:3, which had an EC₂₀ value (1.8 nM) significantly lower than that of 5-oxo-ETE (EC₂₀, 3.1 nM) (p < 0.05). In general, the IC₅₀ values for calcium desensitization paralleled the EC₂₀ values for actin polymerization, with the exception that 5-oxo-20:3 and 5-oxo-ETE had similar potencies (Fig. 6B).

Effects of 5-Oxo-20:3 and Related Compounds on Neutrophil Migration. Because of the high potency of 5-oxo-20:3 in eliciting various responses in neutrophils and eosinophils, we investigated its chemotactic effects on neutrophils. Both 5-oxo-20:3 and 8-trans-5-oxo-20:3 strongly stimulated neutrophil chemotaxis, with concentration-response curves that were virtually indistinguishable from that of 5-oxo-ETE (Fig. 7) and EC₅₀ values between 30 and 50 nM. The maximal response for 5-oxo-20:2 was approximately 25% lower than that for 5-oxo-ETE (p < 0.05).

Formation of 5-Oxo-20:3 from Mead Acid by Neutrophils. Because of the potent effects of 5-oxo-20:3 on neutrophils and eosinophils, we investigated the possibility that it could be formed from Mead acid (i.e., 5Z,8Z,11Z-eicosatrienoic acid) by neutrophils. Mead acid was incubated with neutrophils in the presence of the calcium ionophore A23187 and PMA, and the products were analyzed by RP-HPLC (Fig. 8A). The major product had a retention time (t_R) of 32.9 min and a λ_max at 235 nm (Fig. 8B), and cochromatographed with authentic chemically synthesized 5-hydroxy-6,8,11-eicosatrienoic acid (5-hydroxy-20:3). Two less polar products absorbing at 237 and 235 nm (t_R 26.2 and 29.8 min, respectively), probably other monohydroxy metabolites, were also formed, but their identities were not determined. Three major products absorbing at 280 nm were also detected (Fig. 8A). The least polar of these (t_R, 35.3 min) had a λ_max at 280 nm (Fig. 8B) and cochromatographed with authentic 5-oxo-20:3. Two more polar products with t_Rs of 13.7 and 14.8 min had identical UV spectra typical of conjugated trienes with λ_max values at 258, 269, and 279 nm (Fig. 8B). These products are presumably identical to 6-trans-LTB₃ and 12-epi-6-trans-LTB₃, which have previously been reported to be formed from Mead acid by neutrophils (Stenson et al., 1984).

The time course for the formation of the four products discussed above is shown in Fig. 8C. All of these products were formed very rapidly and reached maximal levels by approximately 2 min. 5-Hydroxy-20:3 was the major product at all time points, followed by 5-oxo-20:3 and the two 6-trans isomers of LTB₃.

To provide more conclusive evidence for the identities of the two major products as 5-hydroxy-20:3 and 5-oxo-20:3, these substances were purified by RP-HPLC after incubation of Mead acid with neutrophils and analyzed by mass spectrometry. Analysis of the material in the peak attributed to 5-hydroxy-20:3 by high-resolution electrospray ionization-FTMS in mass spectrometry mode gave an intense M-1 ion at m/z 321.2435 ( requires 321.2435). Collision-induced dissociation of this ion gave a series of product ions (Fig. 9A) with m/z 303.2328 (M-1-H₂O; requires 303.2330), 285.2227 (M-2×H₂O; C₂₀H₂₉O^–1 requires 285.2224), 277.2531 (M-1-CO₂; C₁₉H₃₃O^–1 requires 277.2537), 259.2429 (M-1-CO₂-H₂O; requires 259.2431), 205.1959 (C₆-C₂₀; requires 205.1962), and 115.0398 (C₁-C₅-H; requires 115.0401). The ions at m/z 115 and 205 are formed by cleavage of the bond between carbons 5 and 6 and are indicative of the presence of a hydroxyl group at C₅. This product ion spectrum is virtually identical to that obtained for authentic chemically synthesized 5-hydroxy-20:3 (data not shown).

FTMS analysis of the material in the HPLC peak attributed to 5-oxo-20:3 gave an intense M-1 ion at m/z 319.2279 ( requires 319.2279), which, when subjected to collision-induced dissociation (Fig. 9B), gave a series of product ions at m/z 301.2174 (M-1-H₂O; requires 301.2173), 283.2065 (M-2×H₂O; C₂₀H₂₇O^–1 requires 283.2067), 275.2379 (M-1-CO₂; C₁₉H₃₁O^–1 requires 275.2380), 257.2272 (M-1-CO₂-H₂O; requires 257.2275), 247.2064 (C₄-C₁₈; C₁₇H₂₇O^–1 requires 247.2067), 245.1909 (C₄-C₁₈-H₂; C₁₇H₂₅O^–1 requires 245.1911), 205.1959 (C₆-C₂₀; requires 205.1962), 129.0554 (C₁-C₆+2H; requires 129.0557), 113.0242 (C₁-C₅-H; requires 113.0244), and 111.0449 (129-H₂O; requires 111.0452). The ions at m/z 113 and 205 formed by cleavage between carbons 5 and 6 are consistent with the presence of an oxo group at carbon 5. This mass spectrum is nearly identical to that obtained for authentic 5-oxo-20:3 (data not shown).

Discussion

The results of the present study clearly show that the OXE receptor is selective for 5-oxo fatty acids containing 18 or 20 carbons and double bonds in the 6 and 8 positions. Comparison of a series of 5-oxo-6E,8Z-dienoic acids with carbon chain lengths between 12 and 20 revealed, somewhat surprisingly, that the C₁₈ compound 5-oxo-18:2 was the most potent among this group (Fig. 6A). Of the two 5-oxo-monoenoic acids investigated, the C₁₈ fatty acid 5-oxo-18:1 was also more potent than its C₂₀ counterpart 5-oxo-20:1. Of the C₂₀ compounds tested, 5-oxo-20:3 and its 8-trans isomer were the most potent in stimulating actin polymerization in eosinophils (Fig. 6B). These compounds were equipotent with 5-oxo-ETE in stimulating chemotaxis and desensitization of the OXE receptor in neutrophils, and they were slightly less potent in inducing surface expression of CD11b and calcium mobilization in these cells. The optimum number of double bonds is thus 3 or 4, with greater (e.g., 5-oxo-20:5) or fewer (e.g., 5-oxo-20:2) double bonds resulting in reduced potency. Furthermore, the Δ¹¹ double bond is clearly much more important than the Δ¹⁴ double bond, because 5-oxo-20:3 (Δ^6E,8Z,11Z) is considerably more potent than Δ^6E,8Z,14Z-5-oxo-20:3 in inducing all of the responses investigated. Loss of the Δ¹⁴ double bond likewise has little impact on the effects of leukotrienes on their receptors. LTD₃ was reported to be equipotent with LTD₄ in stimulating contraction of guinea pig ileum (Hammarström, 1981), whereas LTB₃ was nearly as potent as LTB₄ in activating neutrophils (Evans et al., 1985a).

The dramatically reduced potency caused by shortening the carbon chain to 16 carbons (5-oxo-16:2) suggests that the hydrophobic ω-end of the molecule is important for recognition by the OXE receptor. This is in agreement with our previous finding that introduction of a polar hydroxyl group at C₂₀ (5-oxo-20-HETE) results in a reduction in potency of nearly 100-fold (Powell et al., 1996). It is interesting to note that 5-oxo-16:2 seems to be a partial agonist, because it induced maximal responses that were only 34 ± 3% (actin polymerization) and 20 ± 8% (calcium mobilization) of those to 5-oxo-ETE. In contrast, 5-oxo-16:2 completely blocked 5-oxo-ETE-induced calcium mobilization. It is possible that further modification of the structure of this compound could result in a compound with antagonist properties.

The lack of activity of short-chain analogs of 5-oxo-ETE suggests that short-chain lipid peroxidation products such as 5-oxovaleric acid, which can be released from peroxidized phospholipids by platelet-activating factor (PAF) acetyl hydrolase (Stremler et al., 1989), would not activate the OXE receptor. In contrast, peroxidized phospholipids containing short-chain aldehydes have been shown to activate the PAF receptor (Smiley et al., 1991).

The presence of double bonds in the 6 and 8 positions of 5-oxo fatty acids is clearly important for biological activity. In the absence of any carbon-carbon double bonds (5-oxo-20:0), there is only a small degree of activity at the highest concentration tested (10 μM). The importance of the Δ⁶ double bond is consistent with the lack of activity of 6,7-dihydro-5-oxo-ETE (i.e., 5-oxo-8Z,11Z,14Z-eicosatrienoic acid), which we previously reported to be a metabolite of 5-oxo-ETE formed by a calcium/calmodulin-dependent olefin reductase in neutrophils (Berhane et al., 1998). Although the presence of a single double bond in the 6 position (i.e., 5-oxo-20:1) clearly increased activity, this compound was still approximately 100 times less potent than 5-oxo-ETE. The further addition of a Δ⁸ double bond resulted in markedly increased biological activity. However, the actual configuration of this double bond seems to be of much lesser importance, because there was little difference between the potencies of 5-oxo-20:3 (Δ^6E,8Z,11Z) and 8-trans-5-oxo-20:3 (Δ^6E,8E,11Z). This is a little surprising, because alteration of the configuration of this double bond would be expected to affect the orientation of the distal, hydrophobic portion of the molecule, which, as noted above, is important for a strong interaction with the OXE receptor. This suggests that there is sufficient flexibility in the ligand to permit the terminal part of the molecule (∼C₁₈– C₂₀) to interact appropriately with the binding site on the receptor. We previously found that alteration of the configuration of the Δ⁸ double bond of 5-oxo-ETE (i.e., 8-trans-5-oxo-ETE) reduced potency by approximately 6-fold compared with 5-oxo-ETE. The greater impact of this modification in the case of 5-oxo-ETE may be due to the presence of the Δ¹⁴ double bond, which, although not required for agonist activity, might restrict the flexibility of this part of the molecule, reducing its ability to compensate for the lack of a Δ⁸cis double bond.

The concentration-response curves for the effects of 5-oxo-ETE on calcium mobilization and CD11b expression seem to be biphasic, tending toward maxima at ∼1 μM and then increasing markedly at 10 μM. We conducted a limited number of experiments at higher concentrations of 5-oxo-ETE (up to 100 μM), which resulted in substantially greater, but more variable, responses, but we have not pursued this further. A similar pattern was observed for 5-oxo-18:2 and 8-trans-5-oxo-20:3. On the other hand, the responses to some of the other analogs, such as Δ^6E,8Z,14Z-5-oxo-20:3 and 5-oxo-20:2, seemed to reach maxima, at least in the case of CD11b expression. In contrast, the concentration-response curves for the effects of most of the 5-oxo fatty acids on actin polymerization in eosinophils reached maxima within the concentration range tested. Similar results were obtained with neutrophils (data not shown). The most likely explanation for the biphasic response is that 5-oxo-ETE is a weak activator of another unrelated receptor at very high, presumably nonphysiologic, concentrations. Activation of this second receptor would then result in increased calcium mobilization and CD11b expression, without having an appreciable effect on actin polymerization. We previously found that the ability of agonists to stimulate calcium mobilization and CD11b expression in eosinophils did not necessarily parallel their ability to promote actin polymerization, with 5-oxo-ETE eliciting relatively stronger actin responses compared with other agonists such as PAF (Powell et al., 1999) and eotaxin (Powell et al., 2001).

Although it would seem conceivable that the responses to 5-oxo fatty acids that we observed are mediated principally by the OXE receptor, it is possible that interaction with other receptors could have contributed to some extent, as discussed above. Moreover, the degree of stimulation of OXE and nonOXE receptor-mediated responses may not have been the same for all of the compounds tested. The contribution of responses mediated by other receptors should be reduced by focusing more on the responses to lower concentrations of agonists by using EC₂₀ rather than EC₅₀ values. Furthermore, the results obtained for desensitization of 5-oxo-ETE-induced calcium mobilization used a low concentration of 5-oxo-ETE (10 nM), which would minimize activation of receptors other than the OXE receptor. Thus, inhibition of this response by 5-oxo fatty acids is probably due to their effects on the OXE receptor. The actin polymerization response to these compounds in eosinophils is also probably a good indicator of OXE receptor activation, because responses to all of the more potent agonists reached plateaus by approximately 1 μM. In general, there is very good agreement in the relative potencies of the analogs tested with respect to the different responses (Table 1).

Because of the potency of 5-oxo-20:3 in activating neutrophils and eosinophils, we were interested in determining whether it could be synthesized from a polyunsaturated fatty acid (PUFA) precursor, namely the ω9 PUFA 5,8,11-eicosatrienoic acid, otherwise known as Mead acid. In contrast to ω6 and ω3 PUFA, which are essential fatty acids that must be supplied by the diet, ω9 fatty acids are formed de novo by mammalian cells. In conditions of essential fatty acid deficiency, Mead acid accumulates in tissue lipids (Mead, 1956). Although it cannot be converted to prostanoids by cyclooxygenase because it lacks the 14,15-double bond, it can be converted to mono- and dihydroxyeicosatrienoic acids by this enzyme (Elliott et al., 1986). In contrast, Mead acid can be converted to leukotrienes (Hammarström, 1981) and 5-hydroxy-20:3 (Wei et al., 1985) by 5-LO. Although both arachidonic acid and Mead acid are transformed to LTA₄ and LTA₃, respectively, by 5-LO, these two substances are metabolized differently from one another. Both LTA₄ and LTA₃ are good substrates for LTC synthase (Hammarström, 1981), whereas only LTA₄ is extensively metabolized by LTA hydrolase, and LTA₃ inhibits this enzyme (Evans et al., 1985b). Thus, instead of being metabolized to LTB₃ by neutrophils, Mead acid-derived LTA₃ is nonenzymatically converted principally to 6-trans-LTB₃ and 12-epi-6-trans-LTB₃ by these cells.

The major product formed from Mead acid by neutrophils in the present study was 5-hydroxy-20:3, the identity of which was confirmed by mass spectrometry. Substantial amounts of the two 6-trans isomers of LTB₃ were also detected. In addition, 5-oxo-20:3 was identified by comparison of its properties and mass spectrum with those of the authentic chemically synthesized compound. This is the first report of this compound from either a biological or chemical source, and the first documentation of its potent biological effects. Because neutrophils are unable to synthesize significant amounts of LTB₃ or LTC₃ from Mead acid, it would seem that the major biologically active 5-LO product formed from this PUFA by these cells is 5-oxo-20:3.

In conclusion, C₁₈ and C₂₀ 5-oxo-Δ^6,8 fatty acids are potent activators of the OXE receptor-mediated activation of human neutrophils and eosinophils. The most potent compounds tested were 5-oxo-ETE, 5-oxo-20:3, 8-trans-5-oxo-20:3, and 5-oxo-18:2. Neutrophils convert the ω9 PUFA Mead acid to one of these products, 5-oxo-20:3, which could potentially be formed and act as a proinflammatory mediator in situations of dietary essential fatty acid deficiency.