5-Oxo-eicosanoids and hematopoietic cytokines cooperate in stimulating neutrophil function and the mitogen-activated protein kinase pathway.

The newly defined eicosatetraenoates (ETEs), 5-oxoETE and 5-oxo-15(OH)-ETE, share structural motifs, synthetic origins, and bioactions with leukotriene B4 (LTB4). All three eicosanoids stimulate Ca2+ transients and chemotaxis in human neutrophils (PMN). However, unlike LTB4, 5-oxoETE and 5-oxo-15(OH)-ETE alone cause little degranulation and no superoxide anion production. However, we show herein that, in PMN pretreated with granulocyte-macrophage or granulocyte colony-stimulating factor (GM-CSF or G-CSF), the oxoETEs become potent activators of the last responses. The oxoETEs also induce translocation of secretory vesicles from the cytosol to the plasmalemma, an effect not requiring cytokine priming. To study the mechanism of PMN activation in response to the eicosanoids, we examined the activation of mitogen-activated protein kinase (MAPK) and cytosolic phospholipase A2 (cPLA2). PMN expressed three proteins (40, 42, and 44 kDa) that reacted with anti-MAPK antibodies. The oxoETEs, LTB4, GM-CSF, and G-CSF all stimulated PMN to activate the MAPKs and cPLA2, as defined by shifts in these proteins' electrophoretic mobility and tyrosine phosphorylation of the MAPKs. However, the speed and duration of the MAPK response varied markedly depending on the stimulus. 5-OxoETE caused a very rapid and transient activation of MAPK. In contrast, the response to the cytokines was rather slow and persistent. PMN pretreated with GM-CSF demonstrated a dramatic increase in the extent of MAPK tyrosine phosphorylation and electrophoretic mobility shift in response to 5-oxoETE. Similarly, 5-oxoETE induced PMN to release some preincorporated [14C]arachidonic acid, while GM-CSF greatly enhanced the extent of this release. Thus, the synergism exhibited by these agents is prominent at the level of MAPK stimulation and phospholipid deacylation. Pertussis toxin, but not Ca2+ depletion, inhibited MAPK responses to 5-oxoETE and LTB4, indicating that responses to both agents are coupled through G proteins but not dependent upon Ca2+ transients. 15-OxoETE and 15(OH)-ETE were inactive while 5-oxo-15(OH)-ETE and 5(OH)-ETE had 3- and 10-fold less potency than 5-oxoETE, indicating a rather strict structural specificity for the 5-keto group. LY 255283, a LTB4 antagonist, blocked the responses to LTB4 but not to 5-oxoETE. Therefore, the oxoETEs do not appear to operate through the LTB4 receptor. In summary, the oxoETEs are potent activators of PMN that share some but not all activities with LTB4. The response to the oxoETEs is greatly enhanced by pretreatment with cytokines, indicating that combinations of these mediators may be very important in the pathogenesis of inflammation.

Cells and Bioassays-We isolated PMN (Ͼ95% PMN, Ͻ 5 platelets/ 100 PMN, no red blood cells) from normal human blood (9). To assay degranulation, 1.3 ϫ 10 6 PMN in 0.5 ml of Hanks' buffer were incubated at 37°C for 20 min, exposed to BSA Ϯ cytokine for 0 -40 min, treated with 5 g/ml cytochalasin B for 3 min, challenged for 5 min, placed on ice, and centrifuged (12,000 ϫ g; 0.2 min; 4°C) to obtain supernatants that were assayed for LDH, lysozyme, and ␤-glucuronidase (10). Results are reported as net enzyme release, i.e. the percentage of total cell enzyme released by stimulated PMN minus that released by unstimulated but otherwise identically treated (including cytokine exposure) PMN. None of the stimuli or reagents used here caused net release of cytosolic LDH. Hence, observed enzyme release reflected degranulation, not cell lysis. To assay O 2 Ϫ production, 1 ϫ 10 7 PMN in 1 ml of Hanks' buffer containing 50 nM cytochrome c were incubated at 37°C for 20 min, exposed to GM-CSF for 17 min and 5 g cytochalasin B for 3 min, and stimulated while being monitored (550 nm) with a splitbeam spectrophotometer. Data were corrected for the responses of PMN stimulated in the presence of 50 g/ml superoxide dismutase and reported as superoxide dismutase-inhibitable changes in optical density (13). We assayed secretory vesicle mobilization by the method of Borregaard et al. (42). 2-4 ϫ 10 8 PMN in 5 ml of Hanks' buffer were challenged at 37°C for 20 min, transferred to 4°C cavitation buffer (5 ml), and further processed at 4°C. Cells were subjected to N 2 cavitation, freed of nuclei by low speed centrifugation, normalized for protein and LDH, and centrifuged through Percoll discontinuous gradients to obtain 19 fractions that were enriched with markers for cytosol (LDH, fractions 1-3), endoplasmic reticulum (NADPH-dependent cytochrome c reductase, fractions 1-7), plasmalemma (surface alkaline phosphatase and [ 14 C]concanavalin A-labeled cell surface glycoprotein, fractions 4 -6), secretory vesicles (total minus surface alkaline phosphatase, peaks at fractions 5-7 and 10 -13), light Golgi (UDP-[ 3 H]galactose:Nacetylglucosamine galactosyltransferase, fractions 4 -6), secondary granules (vitamin B 12 -binding protein, fractions [12][13][14], heavy Golgi (UDP-[ 3 H]galactose:N-acetylglucosamine galactosyltransferase, fractions [12][13][14], and primary granules (␤-glucuronidase, fractions 16 -18), as detailed elsewhere (43). Alkaline phosphatase was assayed by incubating (30 min; 37°C) 60 l of a Percoll gradient fraction in 640 l of a pH 10 solution of 50 mM 2-amino-2-methylpropanol, 0.4% Triton X-100, 14 mM MgCl 2 , and 1.6 mg of p-nitrophenol phosphate. Reactions were centrifuged (12,000 ϫ g; for 5 min; 20°C) to remove Percoll, supernatants were quantitated for p-nitrophenol by measuring absorbance at 410 nm, and data were converted to rates of p-nitrophenol formed (nmol/30 min). Surface alkaline phosphatase was assayed identically except the reaction mixtures did not contain Triton X-100. Latent alkaline phosphatase is the total minus surface alkaline phosphatase. During secretory vesicle mobilization, latent alkaline phosphatase falls while surface alkaline phosphatase rises (42). To assay secretory vesicle mobilization more easily, 5 ϫ 10 7 PMN in 2.5 ml of Hanks' buffer were incubated with stimuli at 37°C for 20 min, resuspended in 4°C cavitation buffer, subjected to N 2 cavitation, and centrifuged (200 ϫ g; 4 min; 4°C) to remove nuclear debris. Postnuclear supernatants were assayed for total, surface, and latent alkaline phosphatase activity. Results are given as the percentage of the total alkaline phosphatase activity that is latent (42,44).
PMN Ca 2ϩ Depletion and Pertussis Toxin Treatment-PMN, 6 ϫ 10 6 /ml, were incubated at 37°C for 90 min in Ca 2ϩ -free Hanks' buffer containing 1 M EGTA and 1 M Fura-2 AM, washed twice, incubated (6 ϫ 10 6 cells/ml) in Hanks' buffer containing 0 or 1.4 mM Ca 2ϩ at 37°C for 20 min and then challenged. These Ca 2ϩ -depleted PMN, if incubated with no Ca 2ϩ , fail to alter cytosolic Ca 2ϩ upon challenge by Ca 2ϩmobilizing agonists, but if incubated with 1.4 mM Ca 2ϩ , mount full Ca 2ϩ transient responses (45). For pertussis toxin studies, PMN (1.8 ϫ 10 7 / ml) in Ca 2ϩ -free buffer were incubated with 4 g/ml pertussis toxin or BSA at 37°C for 120 min, treated with 1.4 mM Ca 2ϩ for 20 min, and challenged for 1-5 min. These PMN have 60% lower Ca 2ϩ transient and G protein activation responses to 5-HETE, perhaps because of the extensive incubation period (9, 10). Following challenge, PMN suspensions were centrifuged (12,000 ϫ g; for 10 s; 20°C) and processed as described for MAPK.
Arachidonic Acid Release-PMN (2 ϫ 10 7 /ml) in Ca 2ϩ -free Hanks' buffer were incubated with 300,000 dpm/ml [ 14 C]arachidonic acid for 30 min (37°C), washed in Ca 2ϩ -free Hanks buffer (2.5 mg/ml BSA) and then Ca 2ϩ -free Hanks' buffer, and resuspended in Hanks' buffer (1.4 mM Ca 2ϩ ). Suspensions (500 l) were incubated (37°C) for 15 min, treated with 0 -2000 pM GM-CSF for 20 min, stimulated for 0.25 min, and treated with 10 mg/ml BSA (final concentration) to remove free arachidonic acid from cells (46). After 4.75 min, suspensions were centrifuged (12,000 ϫ g; 5 s; 20°C) to isolate 400 l of supernatant fluid. Pellets plus the lower 100 l of supernatants were treated with 100 l of 0.1% Triton X-100; reaction tubes were washed with 100 l of 0.1% Triton X-100, and suspensions and washing fluids were pooled. The latter pools as well as the upper supernatants were separately counted for radioactivity. Results are expressed as the percentage of total cellular radiolabel released by stimulated cells minus that released by cells that were treated identically (including GM-CSF exposure) but incubated with BSA instead of 5-oxoETE.
MAPKs-Anti-ERK-1 antibody revealed three immunoreactive proteins in PMN, here designated as p40, p42, and p44 according to their electrophoretic mobility (Fig. 4A). To identify these bands further, we used four other antibodies. Polyclonal anti-ERK-1 K23 antibody reacted with all three bands. Monoclonal anti-ERK-2 antibody reacted with p42 exclusively (Fig.  4A). Polyclonal anti-ERK-2 C-14 antibody reacted with p44 and p42 but only to a relatively minor extent with p40 (data not shown). An antibody to p38 identified a single band, which did not comigrate with p40. The band reacting with p38 antibody did not show a mobility shift in cells stimulated with 5-oxoETE, FMLP, or other stimuli under conditions were all three ERK-1 antibody-reactive bands exhibited mobility shifts (see below). Based on the specificities of these antibodies (47)(48)(49)(50)(51)(52)(53)(54), these data identify p44 and p42 as ERK-1 and -2, respectively, and p40 as an ERK-1-like protein. We found that our ability to detect p40 was lost very quickly (3-5 min) if the samples were left in the solution of protease and phosphatase inhibitors without boiling. 2 This may explain why previous investigators have not observed this protein. All three proteins were converted to slower migrating forms in PMN stimulated by G-CSF, GM-CSF, LTB 4 , or 5-oxo/OH-ETEs but not by cells incubated with 15-HETE or 15-oxoETE (Fig. 4, A and B). Such mobility shifts characteristically occur when MAPKs are activated by phosphorylation (46,47). Indeed, we observed that PMN stimulated with 5-oxoETE showed new bands at p44, p42, and p40 that were detected with anti-ERK-1-Y(PO 4 ) antibody (Fig. 7). MAPK gel-shift responses to GM-CSF started at 5 min, peaked by 20 min, and slowly declined while those to 5-oxoETE and LTB 4 began by 0.25 min, maximized at 1 min, and disappeared in less that 5 min (Fig. 5A). In studies not shown, LY 255283 (1-4 M) blocked responses to 0.1-10 nM LTB 4 yet did not alter those to 5-500 nM 5-oxoETE. Finally, no stimulus, even at levels 50-fold higher than those causing half-optimal effects (ED 50 levels were ϳ100 pM GM-CSF, 200 pM G-CSF, 500 pM LTB 4 , 10 nM 5-oxoETE, 50 nM 5-oxo-15(OH)-ETE, and 100 nM 5-HETE), shifted more than 50% of any band to its slower form. Primed PMN achieved far greater responses. PMN treated with 200 or 20 pM GM-CSF phosphorylated essentially all of p40 on exposure to 100 nM 5-oxoETE (Fig. 6). Similar results were observed using anti-ERK-1-Y(PO 4 ) antibody. Unstimulated PMN did not contain phosphorylated MAPK. However, anti-ERK-1-Y(PO 4 ) reactive bands (p40, p42, and p44) were observed at 0.25 and 1 min after stimulation (Fig. 7). The reactive bands were transient and were not observed at 4 min. PMN, primed with either 20 or 200 pM GM-CSF for 20 min, showed greatly enhanced reactivity in all three bands upon challenge with 5-oxoETE at all times tested (Fig. 7). cPLA 2 -In resting PMN, antibody to cPLA 2 detected a protein at ϳ85 kDa plus a variable amount of a slower migrating species. Since these two bands correspond to the non-activated and activated forms of cPLA 2 (43, 48), some cPLA 2 may have been activated by our cell manipulations. In any case, resting PMN had by far the most dominant form of cPLA 2 in the faster 2 A. B. Nixon, unpublished data.

FIG. 4. Identification of the MAPKs in human PMNs
. PMN were stimulated with G-CSF or GM-CSF for 5 min or LTB 4 , 15-HETE, or 5-oxoETE for 1 min and lysed. Lysates were resolved by 12% polyacrylamide SDS gel electrophoresis, transferred to nitrocellulose membranes, reacted with the indicated antibody, and visualized (see "Experimental Procedures"). While the stimuli caused p40 and p42 to split, p44 is shown only to broaden. However, analyses using 10% polyacrylamide gels indicated that p44 clearly split into two bands in PMN exposed to 5-oxoETE, LTB 4 , G-CSF, or GM-CSF. The results presented are representative of seven experiments.
Arachidonic Acid Release-PMN incubated with [ 14 C]arachidonic acid incorporated Ͼ 90% of added radioactivity. When the prelabeled PMNЈs were incubated with BSA Ϯ GM-CSF there was no increase in [ 14 C]arachidonic acid release compared to control cells (ϳ1% of total cell label). PMN exposed to Ն 500 nM 5-oxoETE released small amounts of radiolabel. GM-CSF (Ն 20 pM), while not altering basal arachidonic acid release, greatly increased the potency of 5-oxoETE in eliciting this response (Fig. 8). In the presence of 2000 pM GM-CSF, 5 nM 5-oxoETE was a thousandfold more active (Fig. 8). By comparison, 500 and 5000 nM 15-HETE were totally ineffective in stimulating arachidonic acid radiolabel release in unprimed or GM-CSF (20 -2000 pM)-primed PMN (data not shown). The arachidonic acid release in response to FMLP (1 M) was also enhanced by pretreatment with GM-CSF. Hence, the positive cooperation between 5-oxoETE and GM-CSF in activating MAPK (Figs. 6 and 7) was associated with a similar co-operativity in mobilizing arachidonic acid (Fig. 8). , 500 or 5000 nM) or FMLP (1000 nM) for 0.25 min, treated with excess BSA (10 g/ml) for 4.75 min, and separately analyzed for the amount of radioactivity released. Data (means Ϯ S.E., n ϭ 5-6) are given as the percentage of total cellular radiolabel released after correcting for the release found in cells that were treated with 0 -2000 pM GM-CSF and then challenged with BSA instead of 5-oxoETE or FMLP. Asterisks indicate values significantly (p Յ 0.05, Student's paired t test) above that seen in identically stimulated but unprimed PMN. 15-HETE (500 and 5000 nM) failed to stimulate arachidonic acid release in unprimed or GM-CSF-primed cells (data not shown). and resuspended in Ca 2ϩ -free buffer had normal MAPK mobility shifts in response to 5-oxoETE and LTB 4 , as compared to control cells (data not shown) or Ca 2ϩ -depleted PMN suspended with 1.4 mM Ca 2ϩ (Fig. 9A). On the other hand, pertussis toxin-treated PMN had reduced MAPK responses to 5-ox-oETE and LTB 4 (Fig. 9B). We also used the anti-ERK-1-Y(PO 4 ) antibody to confirm that extracellular Ca 2ϩ was not required for tyrosine phosphorylation of MAPK and that tyrosine phosphorylation in response to 5-oxoETE was sensitive to pertussis toxin (data not shown). These PMN responded fully to 200 pM GM-CSF (data not shown). Hence, the effects of 5-oxoETE and LTB 4 on MAPK require pertussis toxin-sensitive G proteins but not Ca 2ϩ transients. DISCUSSION GM-CSF and G-CSF made PMN as much as 10-fold more sensitive to LTB 4 , PAF, and FMLP. They had an even greater impact on 5-oxo/OH-ETEs, rendering these otherwise ineffective agents potent inducers of degranulation and O 2 Ϫ production (Fig. 1). Thus, GM-CSF and G-CSF cooperate with 5-oxo/ OH-ETEs in vitro and may do so in inflamed tissue where cytokines, arachidonic acid metabolites, and PMN concentrate (1-7). The 5-oxo/OH-ETEs also stimulated PMN secretory vesicles to translocate to plasmalemma, an effect not requiring priming (Figs. 2 and 3). These vesicles contain integrins, which, when moved to the cell surface, promote PMN attachment to blood vessel walls (44). It has been proposed that chemotactic factors cause circulating PMN to extravasate by mobilizing secretory vesicles and that cytokines propagate inflammation by up-regulating degranulation and O 2 Ϫ production induced by chemoattractants. 5-Oxo/OH-ETEs may mimic chemotactic factors in these regards, perhaps with one interesting qualification. Granule enzymes and O 2 Ϫ mediate host defense yet also injure tissues. The nearly absolute cytokine requirement for 5-oxo/OH-ETEs to elicit the latter responses may serve to control their toxicity. Relevant to such toxicity, 5-oxo/OH-ETEs act independently of LTB 4 receptors and are unlikely to be influenced by the LTB 4 receptor antagonists now under clinical study. Drugs interfering with 5-lipoxygenase disrupt 5-HETE as well as LTB 4 synthesis and might be better suited than LTB 4 receptor antagonist to reduce inflammation initiated by arachidonic acid metabolites. On the other hand, specific 5-oxo/ OH-ETE antagonists would be useful for defining these eicosanoids' physiology and pathology, particularly in relation to LTB 4 .
During cell stimulation, MAPKs are phosphorylated on tyrosine and threonine to become competent in phosphorylating and activating various effector enzymes (46 -49). Cells commonly express two MAPKs, most often p44 and p42 isoforms. However, at least some cell types also contain: (a) high molecular weight MAPKs (p46, p54, p63; Refs. 46, 47, and 50 -52); (b) MAPKs migrating with, but differing in structure from, p44 or p42 MAPK (46,51); (c) low molecular weight MAPKs (p38 and p36; Refs. 46, 51, and 52); or (d) MAPK-like proteins (p38 Mpk2, p56 Jnk, and p45 Jnk) that do not react with any anti-MAPK antibody tested to date (49,(53)(54)(55). We find that PMN have p44, p42, and p40 bands that react with two different anti-ERK-1 antibodies. Although they seem analogous to the p45, p43, and p41 bands detected by this antibody in rat PC12 (46) or RBL-2H-3 (47) cells, p42 reacted with monospecific anti-ERK-2 antibody (Fig. 4A) and therefore corresponds to the p41 band in PC12 and RBL-2H-3 cells. Furthermore, we detected a p38 antibody-reactive band that was clearly ahead of the p42 band. This band did not show gel shifts in PMN stimulated with 5-oxoETE, FMLP, or 12-phorbol 13-myristate acetate (data not shown). PMN thus resemble PC12 and RBL-2H-3 cells in containing ERK-1 at ϳ44 kDa and ERK-2 at ϳ42 kDa but differ from them in having ERK-1 antigen at 40 rather than 45 kDa. The p40 reacted minimally with polyclonal anti-ERK-2 C14 antibody, which is directed toward the amino terminus of p42. Therefore, the p40 that we observe may be more related to p44 than to p42 and appears unrelated to previously described p39-p41 MAPKs. Our results also contrast with findings that PMN have only one (p44 (Refs. 31  MAPKs. Since PMN studies often used different MAPK antibodies from those used here, comparisons are difficult. Nonetheless, three observations seem particularly pertinent. First, Torres et al. (40) reported that PMN had p42 and p40 but no p44 MAPK, as defined by several anti-ERK-Abs. We and Torres et al. may have employed the same anti-ERK-2 antibody but different anti-ERK-1 antibody batches. This would explain why we and others (31,32,38,39) find p44 ERK-1 in PMN. The p40, not p44, has evaded earlier detection. Second, Waterman and Sha'afi (59) observed a p40 protein kinase to be tyrosine-phosphorylated in PMN exposed to tumor necrosis factor-␣ but not GM-CSF. The protein did not react with an anti-ERK-1 antibody of specificity differing from those used in our studies. Third, Yuo et al. (60) and Rollet et al. (58) observed that bands at p36 -p40 were phosphorylated on tyrosine in PMN exposed to GM-CSF or tumor necrosis factor-␣ but did not characterize these bands further. Our results clarify these and other conflicting findings by revealing that PMN have p42 ERK-2 , p44 ERK-1 , and a p40 ERK-1 -like protein, all of which are regulated by a diversity of extracellular signals. The cells also have a p38, which differs from p40, p42, and p44. Gel shifts were not observed with p38 under conditions where shifts of p40, p42, and p44 were observed. Our data do not exclude the possibility that p38 is activated, since conditions for revealing p38 gel shifts have not been reported.
GM-CSF (33)(34)(35)(36)(37)(38)(39), FMLP (31,33,(35)(36)(37)40), and PAF (35, 36) induce PMN to activate MAPK, as defined in mobility shift, tyrosine phosphorylation, or activity assays. G-CSF may (33, 60) or may not (34, 35, 61) have such effects. We confirm these results with GM-CSF and extend them to LTB 4 . Both stimuli caused mobility shifts in all three proteins detected by anti-ERK-1 antibody. Responses to GM-CSF developed slowly and endured for Ͼ40 min. The response to LTB 4 peaked at 1 min and ended by 5 min. G-CSF also induced p44 ERK-1 , p42 ERK-2 , and p40 mobility shifts, although its effects were modest. This agrees with findings that G-CSF weakly stimulates PMN to increase MAPK phosphorylating activity and to tyrosine phosphorylate 40 -44-kDa bands (33,60). We further found that 5-oxo/OH-ETEs caused rapid (0.25 min) and transient (2.5 min) mobility shifts in p44 ERK-1 , p42 ERK-2 , and p40 (Fig. 5). Hence, activation of all three proteins seems a constant and possibly central event in PMN activation. These data also indicate that p40 is an extracellular signal-regulated protein, possibly related to the 36 -40-kDa proteins tyrosine-phosphorylated in PMN exposed to GM-CSF or tumor necrosis factor-␣ (58 -60). Further studies will be required to determine if p40 is a new MAPK, or a splicing or post-translationally modified variant of ERK-1 or related to JNKs (53)(54)(55). However, the p40 detected in PMN reacted with a series of antibodies that are directed toward peptide sequences that are not found in JNKs or p38 (56,62) and it seems unlikely that the p40 ERK-1 -like protein seen here is related to the latter kinases. cPLA 2 mobility shifts (Fig. 5B) correlated closely with changes in MAPK mobility although the former persisted much longer than the latter. MAPKs phosphorylate cPLA 2 on serine 505, thereby slowing its migration and doubling its activity (48). In separate work (61), we determined that LTB 4 , 5-ox-oETE, or 5-HETE each stimulated cPLA 2 activity and concomitant phosphorylation of the enzyme. Together, these results make it clear that LTB 4 and 5-oxo/OH-ETEs cause PMN to phosphorylate and activate MAPKs, one or more of which then phosphorylate and activate cPLA 2 . It is worth emphasizing that such events may be necessary but are not sufficient for many PMN functions. LTB 4 and 5-oxoETE have quite similar effects on MAPK and cPLA 2 yet very different impacts on degranulation and O 2 Ϫ production. Clearly, signals other than those of the MAPK and cPLA 2 pathways may make critical contributions not only to PMN function but also phospholipid deacylation, arachidonic acid release, and PAF synthesis. It is worth noting that Ն500 nM 5-oxoETE stimulated small amounts of arachidonic acid release, this effect was comparable to that achieved by 1 M FMLP, and that GM-CSF enhanced the arachidonic acid release in response to 5-oxoETE (Fig. 8). Hence, the cytokine-primed PMN have not only an increased MAPK activation (Figs. 6 and 7) but also an enhanced arachidonic acid release in response to 5-oxoETE. This suggests that priming occurs at a step before MAPK activation and stimulates cPLA 2 activation and arachidonic acid release. In primed PMN, 5-oxoETE may prove to be similarly potent in influencing PAF synthesis as well as other downstream events regulated by MAPKs.
In conclusion, we stress four findings. First, structure-activity relations for 5-oxoETE, 5-oxo-15(OH)-ETE, 5-HETE, 15-oxoETE, and 15-HETE in activating MAPK and cPLA 2 , in translocating secretory vesicle, and in degranulating PMN paralleled each other as well as those previously established for eliciting Ca 2ϩ transients, chemokinesis, and aggregation responses (8 -13). This suggests that all stimulatory effects of 5-oxo/OH-ETEs use a single recognition system. Second, pertussis toxin blocked MAPK responses to LTB 4 and 5-oxo/OH-ETEs, whereas LY 255283 inhibited diverse responses to LTB 4 without influencing those to 5-oxo/OH-ETEs (Fig. 7, Table I). The 5-oxo/OH-ETE recognition system likely involves a G protein-and Ca 2ϩ -linked receptor distinct from LTB 4 receptors. Third, Ca 2ϩ -depleted PMN responded to 5-oxo/OH-ETEs with MAPK mobility shifts (Fig. 9). These data are the first to indicate that 5-oxo/OH-ETEs act on Ca 2ϩ -independent as well as Ca 2ϩ -dependent pathways, both of which are initiated by G proteins. Fourth, GM-CSF greatly enhanced the effects of 5-oxo/OH-ETEs on degranulation, O 2 Ϫ production, and MAPK activation and arachidonic acid release (Figs. 1 and 6 -8). 5-Oxo/OH-ETEs thus co-operate with hematopoietic cytokines to stimulate MAPK, cPLA 2 , and functional responses.