Regulation of Delayed Prostaglandin Production in Activated P388D1 Macrophages by Group IV Cytosolic and Group V Secretory Phospholipase A2s*

Group V secretory phospholipase A2 (sPLA2) rather than Group IIA sPLA2 is involved in short term, immediate arachidonic acid mobilization and prostaglandin E2 (PGE2) production in the macrophage-like cell line P388D1. When a new clone of these cells, P388D1/MAB, selected on the basis of high responsivity to lipopolysaccharide plus platelet-activating factor, was studied, delayed PGE2 production (6–24 h) in response to lipopolysaccharide alone occurred in parallel with the induction of Group V sPLA2 and cyclooxygenase-2 (COX-2). No changes in the level of cytosolic phospholipase A2(cPLA2) or COX-1 were observed, and Group IIA sPLA2 was not detectable. Use of a potent and selective sPLA2 inhibitor, 3-(3-acetamide 1-benzyl-2-ethylindolyl-5-oxy)propanesulfonic acid (LY311727), and an antisense oligonucleotide specific for Group V sPLA2revealed that delayed PGE2 was largely dependent on the induction of Group V sPLA2. Also, COX-2, not COX-1, was found to mediate delayed PGE2 production because the response was completely blocked by the specific COX-2 inhibitor NS-398. Delayed PGE2 production and Group V sPLA2expression were also found to be blunted by the inhibitor methylarachidonyl fluorophosphonate. Because inhibition of Ca2+-independent PLA2 by an antisense technique did not have any effect on the arachidonic acid release, the data using methylarachidonyl fluorophosphonate suggest a key role for the cPLA2 in the response as well. Collectively, the results suggest a model whereby cPLA2 activation regulates Group V sPLA2 expression, which in turn is responsible for delayed PGE2 production via COX-2.

Arachidonic acid (AA) 1 mobilization and the generation of prostaglandins by major immunoinflammatory cells such as macrophages and mast cells usually occur in two phases. The immediate phase, which takes minutes and is elicited by Ca 2ϩmobilizing agonists such as platelet-activating factor (PAF), is characterized by a burst of AA liberation. In some cells such as P388D 1 macrophages (1, 2) and MMC-34 mast cells (3), this burst is mainly produced by a secretory phospholipase A 2 (sPLA 2 ) but is strikingly regulated by the cytosolic Group IV phospholipase A 2 (cPLA 2 ).
The delayed phase of prostaglandin production is accompanied by the continuous supply of AA over long incubation periods spanning several hours. There is some discrepancy about the identity of the PLA 2 isoform(s) involved in the delayed phase. Despite this phase being independent of a Ca 2ϩ increase, the cPLA 2 has often been suggested to be critically involved (3)(4)(5). However, other studies have suggested the quantitatively more important role of the sPLA 2 , an enzyme that is dramatically induced during long term incubation of the cells with a variety of stimuli (4 -6). There is, however, agreement that COX-2, another enzyme whose expression is augmented dramatically after long term stimulation, is absolutely required for long term PGE 2 production, irrespective of the constitutive presence of COX-1 (7)(8)(9).
Using a new clone of the P388D 1 macrophage-like cells termed P388D 1 /MAB, we provide herein evidence for the involvement of Group V sPLA 2 in delayed PGE 2 production. Furthermore, our results suggest that Group V sPLA 2 expression is dependent upon the activation of Group IV cPLA 2 .
Cell Culture and Labeling Conditions-P388D 1 cells were maintained at 37°C in a humidified atmosphere at 90% air and 10% CO 2 in Iscove's modified Dulbecco's medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and nonessential amino acids. P388D 1 cells were plated at 10 6 /well, allowed to adhere overnight, and used for experiments the following day. All experiments were conducted in serum-free Iscove's modified Dulbecco's medium. When required, radiolabeling of the P388D 1 cells with [ 3 H]AA was achieved by including 0.5 Ci/ml [ 3 H]AA during the overnight adherence period (20 h). Labeled AA that had not been incorporated into cellular lipids was removed by washing the cells four times with serum-free medium containing 1 mg/ml albumin.

Measurement of PGE 2 Production and Extracellular [ 3 H]AA
Release-The cells were placed in serum-free medium for 30 min before the addition of LPS for different periods of time. Afterward, the supernatants were removed and cleared of detached cells by centrifugation, and PGE 2 was quantitated using a specific radioimmunoassay (Per-sPective Biosystems, Framingham, MA). For [ 3 H]AA release experiments, cells labeled with [ 3 H]AA were used, and the incubations were performed in the presence of 0.5 mg/ml bovine serum albumin. The supernatants were removed, cleared of detached cells by centrifugation, and assayed for radioactivity by liquid scintillation counting. The standard LPS/PAF stimulation protocol for immediate responses has been described previously (1). Briefly, the cells were incubated for 1 h with 200 ng/ml LPS followed by a 10-min incubation with 100 nM PAF.
Northern Blot Analyses-Total RNA was isolated from unstimulated or LPS-stimulated cells by the TriZOL reagent method (Life Technologies, Inc.), exactly as indicated by the manufacturer. Fifteen g of RNA were electrophoresed in a 1% formaldehyde/agarose gel and transferred to nylon filters (Hybond, Amersham Pharmacia Biotech) in 10ϫ SSC buffer. Hybridizations were performed in QuickHyb solution (Stratagene) following the manufacturer's instructions. 32 P-Labeled probes for Group IIA or glyceraldehyde-3-phosphate dehydrogenase were coincubated with the filters for 1 h at 66°C followed by three washes with 2ϫ SSC containing 0.1% SDS at room temperature for 30 min. A final wash was carried out at 60°C for 30 min with 1ϫ SSC containing 0.1% SDS. For Group V sPLA 2 , hybridizations were performed in Ex-pressHyb solution (CLONTECH) following the manufacturer's instructions. The 32 P-labeled probes were co-incubated with the filters for 1 h at 66°C followed by two washes with 2ϫ SSC containing 0.05% SDS for 15 min: the first at room temperature and the second at 40°C. Afterward the filters were washed twice more with 0.1ϫ SSC containing 0.1% SDS for 15 min at room temperature. Bands were visualized by autoradiography.
For Group VI iPLA 2 antisense experiments, a protocol identical to that reported elsewhere was followed (12).
Data Presentation-Assays were carried out in duplicate or triplicate. Each set of experiments was repeated three times with similar results. Unless otherwise indicated, the data presented are from representative experiments.

AA Release in a Novel P388D 1 Cell Clone (MAB)-Stimula-
tion of murine P388D 1 macrophages with nanomolar amounts of the receptor agonist PAF results in very little AA mobilization. However, preincubation of the cells with LPS prior to stimulation with PAF increases the release of AA by these cells well above unstimulated levels, the relative magnitude of the response being dependent on the cell batch (13,14). We have now selected by limit dilution a clone of the P388D 1 cells termed MAB, which shows a remarkably higher [ 3 H]AA release response to LPS/PAF when compared with the ATCC batch of P388D 1 cells from which the MAB clone was obtained (Fig. 1). More interestingly, in addition to an immediate response to LPS/PAF ( Fig. 2A), cells from the MAB clone also exhibited a delayed [ 3 H]AA release response, spanning several hours, to LPS alone ( Fig. 2A). The dose response of the effect of LPS on long term [ 3 H]AA release is shown in Fig. 2B. The maximal effect was observed at a LPS dose as low as 10 ng/ml.
Prostaglandin Production by P388D 1 /MAB Cells- Fig. 2C shows the time course of PGE 2 production by LPS in these cells as measured by radioimmunoassay, which corresponded well with the [ 3 H]AA mobilization response. Thus, LPS-induced PGE 2 barely increased above controls within the first 3 h of treatment, rising afterward, and reaching a plateau after 12 h.
The effect of LPS on the protein levels of the two COX isoenzymes these cells express (2) was assessed by immunoblot. Expression of COX-1 did not change along the time course of LPS activation (data not shown), whereas COX-2 levels noticeably increased with maximal expression between 12 and 18 h (Fig. 3A). Interestingly, LPS-induced COX-2 expression almost parallels PGE 2 generation (cf. Figs. 2C and 3A), suggesting that COX-2 is the enzyme responsible for LPS-induced PGE 2 synthesis. Indeed, the COX-2-specific inhibitor NS-398 (15) completely blocked LPS-induced PGE 2 production (Fig. 3B). Therefore, LPS-delayed PGE 2 generation depends exclusively on COX-2, irrespective of the continued presence of COX-1.
cPLA 2 Involvement in LPS-induced Long Term Responses-Expression of the Group IV cPLA 2 protein in P388D 1 /MAB cells was constitutive and did not change after exposure to LPS. To address the possible involvement of cPLA 2 in LPS-induced AA mobilization in P388D 1 /MAB cells, experiments were conducted with MAFP, an inhibitor that has previously been shown to block the immediate, cPLA 2 -dependent [ 3 H]AA release in LPS/PAF-treated macrophages (1). As shown in Fig. 4, MAFP strongly blocked the LPS-induced long term [ 3 H]AA release response. MAFP has recently been observed to inhibit the Group VI iPLA 2 in addition to the cPLA 2 (10). Therefore, it could be possible that part of the MAFP effects reported herein resulted from inhibition of the iPLA 2 in addition to any effect on the cPLA 2 . We have recently described the inhibition of iPLA 2 expression in P388D 1 cells by antisense RNA oligonucleotides (12). Using this technique, we have been able to significantly inhibit iPLA 2 expression, assayed both by protein content by immunoblot and activity by a specific in vitro assay (12). Antisense RNA inhibition of the iPLA 2 under identical conditions as those shown previously (12) showed no reduction in AA release in response to LPS (not shown). Therefore these data make it likely that the above reported effects of MAFP on the response are because of the inhibition of the cPLA 2 .
Role of sPLA 2 -LPS-induced long term [ 3 H]AA release was also noticeably blocked by the selective sPLA 2 inhibitor LY311727 (17), indicating that in addition to the cPLA 2 , a sPLA 2 is also involved in the process (Fig. 4). PGE 2 production by LPS was also inhibited by LY311727 by about 90%. Although originally described as a selective Group II sPLA 2 inhibitor (17), we have recently shown that this compound is also a potent Group V sPLA 2 inhibitor (18).
Our previous work (11) has demonstrated that P388D 1 cells express measurable message levels for Group V sPLA 2 , both under unstimulated and LPS/PAF-stimulated conditions. However, message levels for Groups IIA sPLA 2 or IIC sPLA 2 were undetectable even by reverse transcriptase-polymerase chain reaction (11). As shown in Fig. 5A, an antisense oligonucleotide specific for Group V sPLA 2 (ASGV-2) strongly blocked PGE 2  production in LPS-treated cells, whereas its sense control (SGV-2) had no effect. Moreover, mRNA analyses by Northern blot at long times of stimulation with LPS confirmed the presence of mRNA for Group V sPLA 2 (Fig. 5, B and C) but not for Group IIA sPLA 2 (data not shown). The signal for Group V sPLA 2 markedly increased after LPS stimulation, reaching a peak at approximately 10 h. PLA 2 activity measurements in the supernatants of LPSstimulated cells revealed a time-dependent increase in activity (Fig. 6), which correlated well with the time course of Group V sPLA 2 mRNA induction (cf. Figs. 5B and 6). Extracellular PLA 2 activity was decreased if the experiments were conducted in the presence of the RNA synthesis inhibitor actinomycin D (Fig. 7A). This increased extracellular activity was found to correspond to that of Group V sPLA 2 by the following criteria: (i) it was completely inhibited by the sPLA 2 inhibitor LY311727 (Fig. 7B) and (ii) it was decreased in supernatants from cells treated with an antisense RNA oligonucleotide specific for Group V sPLA 2 , ASGV-2 (11) (Fig. 7C).
Role of cPLA 2 in sPLA 2 Activation-Our previous studies have indicated that the immediate AA release triggered by LPS/PAF in these cells involves the sequential action of both cPLA 2 and sPLA 2 , with the activity of the latter being dependent on previous activation of the former (1, 2). Thus we sought to investigate if a similar cross-talk exists between the two enzymes during long term stimulation conditions. We found that no increased PLA 2 activity beyond what was observed in the basal state could be found in supernatants from cells treated with MAFP (Fig. 6). In addition, the cPLA 2 inhibitor markedly decreased the LPS-induced expression of Group V sPLA 2 mRNA (Fig. 8). DISCUSSION A striking hallmark of the immunoinflammatory response is the generation of oxygenated derivatives of AA such as the prostaglandins. The response of major prostaglandin-secreting cells such as macrophages and mast cells to proinflammatory stimuli is generally biphasic (4). The first phase is completed within minutes after the addition of the stimulus, whereas the second phase usually takes several hours (4). Using the murine macrophage-like cell line P388D 1 , we have been studying the molecular mechanisms responsible for AA mobilization and prostaglandin production in response to LPS/PAF. When primed by LPS, these cells will respond to Ca 2ϩ -mobilizing stimuli such as PAF by generating a rapid burst of free AA, part of which is converted to prostaglandins such as PGE 2 . Strikingly, this process is completed within a few minutes after the addition of PAF (19). No free AA or prostaglandins are produced after the immediate phase is completed, not even after several hours of cell exposure to LPS/PAF (13). Such a behavior, which is abnormal for a macrophage cell, has prevented us from studying the molecular mechanisms responsible for delayed prostaglandin production in macrophages. In an attempt to overcome this problem, we subcloned the P388D 1 cells by limit dilution, and selecting on the basis of high responsivity to LPS/PAF, we obtained a clone termed MAB, which shows enhanced sensitivity to LPS/PAF in the immediate phase (min) and exhibits a delayed response (h) to LPS alone.
Using the MAB clone, we have characterized the LPS-induced delayed prostaglandin production in terms of the role played by distinct PLA 2 enzymes and their coupling with down-stream COX enzymes during LPS signaling. Our previous work on the immediate response of the cells to LPS/PAF highlighted the very important role played by the novel Group V sPLA 2 as the provider of most of the free AA directed to PGE 2 biosynthesis (11). Herein, several lines of evidence suggest that Group V sPLA 2 also behaves as a major provider of AA for the delayed phase of PGE 2 production in LPS-treated cells. First, delayed [ 3 H]AA release and PGE 2 production correspond with the induction of Group V sPLA 2 mRNA and enhanced secretion of a sPLA 2 -like activity to the supernatants, with no change in the constitutive levels of cPLA 2 and no detectable induction of Group IIA sPLA 2 . Second, delayed PGE 2 production is strongly blunted by LY311727, a selective sPLA 2 inhibitor. Third, an antisense oligonucleotide specific for Group V sPLA 2 (11) suppresses Group V sPLA 2 activity and inhibits delayed PGE 2 production. Our conclusions in this regard fully agree with recent works by Kudo and co-workers (20, 21) that were published while this manuscript was under review. By using transfection techniques, Kudo and co-workers (20,21) have also documented the importance of Group V sPLA 2 in delayed AA release and PGE 2 production.
Our data have also implicated the cPLA 2 as an important step in LPS signaling by enabling the subsequent action of the sPLA 2 . Thus the cPLA 2 inhibitor MAFP (1) markedly blocked both long term [ 3 H]AA release and Group V sPLA 2 mRNA induction. Collectively, these results suggest an intriguing cross-talk between the cPLA 2 and the Group V sPLA 2 for the delayed phase of prostaglandin production in macrophages. This is a very interesting concept because cross-talk appears to exist as well between these two enzymes during the immediate phase of prostaglandin production (1,2). In the immediate phase, cPLA 2 activation generates a rapid and early burst of free AA inside the cell that enables sPLA 2 activation by an as yet unidentified mechanism (1,2). In the delayed phase, cPLA 2 activation influences sPLA 2 apparently by regulating sPLA 2 mRNA levels.
Cross-talk between cPLA 2 and sPLA 2 in the immediate phase of prostaglandin production was also found to take place in mast cells (3) when the same protocol originally used in macrophages (1) was employed. Furthermore, a recent study by Kuwata et al. (22) about fibroblasts suggests that cross-talk between cPLA 2 and sPLA 2 in the delayed phase could also constitute a general mechanism of activation. Using a different cPLA 2 inhibitor, arachidonyl trifluoromethyl ketone, Kuwata et al. (22) found that cPLA 2 inhibition blocked sPLA 2 expression in fibroblasts, leading to reduced PGE 2 generation. The study by Kuwata et al. (22) is interesting not only because it supports the possible universality of cross-talk between cPLA 2 and sPLA 2 but because the sPLA 2 expressed by rat fibroblasts is a Group IIA enzyme, not Group V. This lends further support to the emerging notion that Group IIA and Group V sPLA 2 may be functionally redundant (23). In addition, Kuwata et al. (22) were able to show that overcoming cPLA 2 inhibition by exogenous AA partially restored the Group IIA sPLA 2 expression. These results suggest that the AA mobilized by cPLA 2 is responsible for cross-talk between cPLA 2 and sPLA 2 (22). This is again reminiscent of what happens in the immediate phase of activation, wherein the cPLA 2 -derived AA is also responsible for cross-talk between cPLA 2 and sPLA 2 , albeit by different mechanisms (1,2). Unfortunately, inhibition by MAFP of Group V sPLA 2 expression and activity could not be reversed in our macrophage studies with LPS alone by supplementing the medium with exogenous AA (up to 100 M). This was not unexpected because P388D 1 cells manifest an extraordinarily high capacity to import free AA from exogenous sources and incorporate it into membranes (19,24,25), which is much higher than that of most other cells (26). Thus, the half-life of the free AA in the cell would be too short to adequately mimic the low but continued production of AA-derived cPLA 2 upon long term LPS exposure.
A model has recently emerged suggesting differential actions of COX-1 and COX-2 by virtue of differential coupling to distinct PLA 2 s (2, 3,6,8,20,21,27). Thus, depending on whether cPLA 2 or sPLA 2 is the provider of free AA, either COX-1 or COX-2 would be responsible for PGE 2 release. However, which PLA 2 form couples to which COX isoform appears to depend strongly on cell type. We have recently demonstrated that the immediate, PAF receptor-mediated phase of PGE 2 production in LPS-primed macrophages involves sPLA 2 coupling to COX-2 (2). The current results support a similar kind of coupling for the delayed PGE 2 production in LPS-treated cells. Identical coupling has been suggested by Arm and co-workers (6) for the delayed phase of PGE 2 generation in mast cells. These results raise another interesting concept regarding the regulation of PGE 2 during both phases of activation. As is the case for AA release ( Fig. 2A), we have observed that the amount of PGE 2 generated during the Ca 2ϩ -dependent short term stimulation is comparable to the amount produced in the late phase. It follows from this comparison that although the effector enzymes involved in the response are the same (i.e. cPLA 2 , sPLA 2 , COX-2), the regulatory mechanisms differ. Thus, in the short phase at low levels of COX-2, it appears that the dramatic burst in AA release is what determines the amount of PGE 2 produced. In contrast, in the delayed phase at comparably lower AA availability, it appears that both the induction of large amounts of COX-2 protein and of the AA provider, Group V sPLA 2 , determine the amount of PGE 2 produced.
It is important to note, however, that our results have not excluded that a minor fraction of the long term PGE 2 produced in response to LPS could arise from the AA generated by the cPLA 2 . Should this be the case, some cPLA 2 /COX-2 coupling may exist as well, similar to what has been suggested by Reddy and Herschman (3) for delayed PGD 2 production in mast cells and by Murakami et al. (5) in cells derived from Group IIAdeficient mice. The striking feature of the current work is that although COX-1 is present in active form in the P388D 1 cells (2), it appears to be spared from the process of long term PGE 2 production. This finding remains unexplained but has recently been recognized in other cell types as well (6,8,22). Recent work by Spencer et al. (16) showed no differences in the distribution of COX-1 versus COX-2 among subcellular fractions in a variety of cells. Thus subcellular compartmentalization may not be the cause for COX-1 not being utilized during LPS signaling. Other putative explanations may include the existence of COX-selective regulatory components, selective coupling to terminal PG synthases, or kinetic differences in AA utilization by the two isoforms.
In summary, we have established a subclone of P388D 1 cells, MAB, that displays long term responsiveness to LPS in terms of PGE 2 generation. We have confirmed (11) that these cells express Group V sPLA 2 , not Group IIA sPLA 2 , and found that (i) Group V sPLA 2 is a key enzyme in long term AA mobilization as well and (ii) Group V sPLA 2 is functionally coupled to COX-2. Furthermore, our results have suggested that cPLA 2 plays a key role in long term AA mobilization, at least partly by regulating the expression of Group V sPLA 2 .