Cytosolic Phospholipase A2 Group IVα but Not Secreted Phospholipase A2 Group IIA, V, or X Induces Interleukin-8 and Cyclooxygenase-2 Gene and Protein Expression through Peroxisome Proliferator-activated Receptors γ 1 and 2 in Human Lung Cells*

It has been reported that interleukin-8 (IL-8) and cyclooxygenase-2 (COX-2) expression is regulated by peroxisome proliferator-activated receptor (PPAR)-γ synthetic ligands. We have shown previously that cytosolic phospholipase A2 (cPLA2) is able to activate gene expression through PPAR-γ response elements (Pawliczak, R., Han, C., Huang, X. L., Demetris, A. J., Shelhamer, J. H., and Wu, T. (2002) J. Biol. Chem. 277, 33153–33163). In this study we investigated the influence of cPLA2 and secreted phospholipase A2 (sPLA2) Group IIA, Group V, and Group X on IL-8 and COX-2 expression in human lung epithelial cells (A549 cells). We also studied the results of cPLA2 activation by epidermal growth factor (EGF) and calcium ionophore (A23187) on IL-8 and COX-2 reporter gene activity, mRNA level, and protein synthesis. cPLA2 overexpression and activation increased both IL-8 and COX-2 reporter gene activity. Overexpression and activation of Group IIA, Group V, or Group X sPLA2s did not increase IL-8 and COX-2 reporter gene activity. Methyl arachidonyl fluorophosphate, a cPLA2 inhibitor, inhibited the effect of A23187 and of EGF on both IL-8 and COX-2 reporter gene activity, steady state levels of IL-8 and COX-2 mRNA, and IL-8 and COX-2 protein expression. Small inhibitory RNAs directed against PPAR-γ1 and -γ2 blunted the effect of A23187 and of EGF on IL-8 and COX-2 protein expression. Moreover small inhibitory RNAs directed against cPLA2 decreased the effect of A23187 and EGF on IL-8 and COX-2 protein expression. These results demonstrate that cPLA2 has an influence on IL-8 and COX 2 gene and protein expression at least in part through PPAR-γ.

85-kDa cytosolic phospholipase A 2 (cPLA 2 ) 1 is a cytoplasmic enzyme that metabolizes phospholipids to release arachidonic acid. Upon activation by various stimuli (including but not limited to calcium ionophore (A23187), IL-1␤, tumor necrosis factor-␣, and interferon-␥) cPLA 2 is translocated to the cellular membranes, releasing arachidonic acid from membrane phospholipids (1)(2)(3)(4). Stimuli such as epidermal growth factor (EGF), oxidative stress, or IL-1␤ may also cause cPLA 2 activation through its phosphorylation (5,6). Previously we have reported that cPLA 2 is able to alter gene expression through PPAR-␥ activation and binding to peroxisome proliferator response elements (4). The calcium ionophore A23187 activates cPLA 2 by increasing intracellular calcium levels, causing cPLA 2 translocation to the nuclear envelope (4,6,7). EGF is also thought to increase cytosolic phospholipase A 2 activity through activation of the mitogen-activated protein kinase pathway and phosphorylation of cPLA 2 at serine 505 (and possibly other serine residues) (5). Both stimuli are known to increase cPLA 2 enzymatic activity and arachidonate release. PLA 2 enzymes catalyze the hydrolysis of membrane glycerophospholipids to liberate arachidonic acid and lysophospholipids. So far, more than 20 enzymes that possess PLA 2 activity have been identified and cloned in mammals (for reviews, see Refs. 8 -15). PLA 2 s have been classified according to their molecular weight, homology, and calcium influence on enzyme activity. cPLA 2 appears to play a role in intracellular arachidonate release, whereas low molecular weight PLA 2 s (such as Groups IA, IB, IIA, IIC, and V) may be involved in extracellular arachidonic acid release due to the fact that they are secreted into extracellular milieu upon cell stimulation. Thus, there may be distinct roles or cross-talk between PLA 2 s in cell signaling (16,17). On the other hand a group of calcium-independent phospholipases A 2 seems to have a broader substrate specificity. Several lines of evidence (for a review, see Ref. 18) suggest that the Group VI iPLA 2 may be responsible for phospholipid fatty acid remodeling in resting cells. The role of iPLA 2 in intracellular cell signaling remains to be clarified.
IL-8 is a chemokine that is produced and secreted by human lung cells (19,20). It has strong chemotactic properties for neutrophils and eosinophils. Moreover IL-8 is an important proinflammatory cytokine that plays a role in allergic inflammation.
The IL-8 promoter contains a PPAR-␥ response element localized to Ϫ1060 relative to the transcription start site, suggesting that PPAR-␥ may play a role in regulation of IL-8 transcription. It has been reported that several PPAR-␥ agonists such as troglitazone, rosiglitazone, and others activate IL-8 transcription and enhanced IL-8 secretion in many cell systems including but not limited to lung cells (21)(22)(23)(24)(25).
Cyclooxygenase is a key enzyme in prostaglandin synthesis. Cyclooxygenase exists in two isoforms (26). COX-1 is a housekeeping gene constitutively expressed in most human cells. COX-2 is a highly inducible cyclooxygenase isoform. Various proinflammatory stimuli such as IL-1␤, tumor necrosis factor-␣, and interferon-␥ have been reported to increase COX-2 expression in many biological systems including bronchial cells (27)(28)(29). The COX-2 promoter contains a PPRE, thus PPAR-␥ agonists including anti-inflammatory drugs may influence COX-2 transcription and expression. 15-⌬ 12,14 -Prostaglandin J 2 , thiazolidinediones, and non-steroidal anti-inflammatory drugs have been reported to alter COX-2 expression (22,24,30,31). As mentioned cPLA 2 activation might increase expression of genes containing PPRE in promoter regions. This hypothesis has been proved using an artificial reporter gene as described elsewhere (4). The purpose of this study was to investigate whether Group IVA cPLA 2 and other phospholipases (such as secreted phospholipase A 2 Group IIA, V, or X) might influence IL-8 and COX-2 gene and protein expression in human lung cells.
Transient Transfection Assay-The cPLA 2 overexpression plasmid was obtained from Drs. J. D. Clark and J. L. Knopf at the Genetics Institute, Boston, MA (22). The sPLA 2 Group IIA expression vector in pcDNA3.1 (Invitrogen) was obtained as described previously (4). The sPLA 2 Group V and Group X vectors were obtained from Drs. D. A. Bass and Michael Seeds (23). The IL-8 luciferase reporter gene was obtained from Dr. Robert L. Danner. The COX-2 reporter gene was a kind gift from Dr. Stephen Prescott (24). A plasmid with ␤-galactosidase gene driven by a CMV promoter was obtained from Clontech. A549 cells were transfected with 1.5 g of reporter gene (IL-8 or COX-2) and 1.5 g of expression vector. 0.2 g of ␤-galactosidase/CMV plasmid was added as a control for transfection efficiency. Cells were transfected in 6-well dishes (PGS Scientific, Bethesda, MD) using LipofectAMINE Plus reagent (Invitrogen) for 4 h in serum-free Ham's F-12K medium (containing 2 mM L-glutamine). After transfection, medium was replaced with standard Ham's F-12K medium containing 10% fetal bovine serum and cells were maintained for 16 h. After exposure to cPLA 2 or sPLA 2 activators as described below, cells were washed three times in ice-cold phosphate-buffered saline and lysed using Passive Lysis Buffer (Promega, Madison, WI). Cell lysate was frozen at Ϫ80°C. Luciferase activity was measured using a luciferase assay system (Promega) with a Turner TD20 luminometer (Promega). ␤-Galactosidase was measured using a Beta-Gal enzyme-linked immunosorbent assay kit (Roche Applied Science).
Immunoblotting-A549 cells were grown on 6-well dishes and treated with 1 M calcium ionophore A23187 (Calbiochem) for 8 h. In experiments involving cPLA 2 or sPLA 2 inhibitors, the inhibitors were added to medium (at the specified concentration) for 2 h before the experiments and maintained throughout the time of exposure to calcium ionophore or EGF. The vehicle, Me 2 SO or methyl acetate, was added to control cultures. Cells were harvested with trypsin (E-PET, Biofluids, Rockville, MD), scraped, collected, and washed three times with cold 1ϫ phosphate-buffered saline. Cells were transferred to 0.5 ml of homogenization buffer containing 50 mM Hepes (pH 8.0), 1 mM EDTA, 1 mM EGTA, 100 M leupeptin, 1 mM dithiothreitol, 10 mM phenylmethylsulfonyl fluoride, 0.5 mM soybean trypsin inhibitor, 15 mM aprotinin, and 0.25% Triton X-100. Cells were sonicated three times for 15 s and centrifuged at 1,000 ϫ g for 5 min. Total protein of the cell lysate was assayed by BCA reagent (Pierce). Samples containing 20 g of crude cell lysate protein were separated on 8 or 11% Tris-glycine gels (Invitrogen) using 1ϫ Tris-glycine SDS running buffer. The separated proteins were electrophoretically transferred onto nitrocellulose membranes (Invitrogen). The membranes were then blocked using 5% nonfat dry milk with 0.1% Tween 20 for 2 h at room temperature. Protein expression was detected by using a 1:1000 dilution of rabbit anti-human PPAR-␥1 or -␥2 antibody or a 1:200 dilution of rabbit anti-human COX-2 antibody (Cayman Chemical) and a 1:1000 dilution of horseradish peroxidase-conjugated mouse anti-rabbit IgG as the second antibody (Jackson ImmunoResearch Laboratory, Inc., West Grove, PA). The blot was developed using the ECL Western blotting detection system (Amersham Biosciences) and exposed to Eastman Kodak MR radiographic film for 1 min.
Arachidonic Acid Release from A549 Cells-Whole cell arachidonic acid release was performed as described previously (4). Briefly cells grown in 6-well dishes were transfected with expression vector or empty vector as described above. After 4 h, medium was changed, and cells were labeled for 16 h with 1 Ci/ml [5,6,8,9,11,12,14, H]arachidonic acid ([ 3 H]arachidonic acid) (214 Ci/mmol) (Amersham Biosciences) in Ham's F-12K medium with 10% fetal bovine serum. Subsequently cells were washed and treated with or without IL-1␤. At the end of the 4-h treatment period, medium was harvested and centrifuged at 1000 ϫ g for 5 min. An aliquot of medium was transferred to scintillation vials containing 10 ml of Bio-Safe II scintillation fluid (Research International Products Inc., Mount Prospect, IL) and counted in an LS 6500 scintillation counter (Beckman).
Gene Expression Measurements Using a Real Time Polymerase Chain Reaction-Cells were grown as described above and exposed to EGF, A23187, or MAFP. Total RNA was isolated using an RNeasy kit (Qiagen, Valencia, CA). IL-8 and COX-2 expression was measured using a real time PCR mRNA quantification using a TaqMan system (Applied Biosystems, Foster City, CA). COX-2 probe and primer sets were obtained from Synthegen (Houston, TX). To measure COX-2 mRNA expression, the following primer sequences were used based on mRNA sequence (GenBank TM accession number U04636): forward primer, 5Ј-GCTCAAACATGATGTTTGCATTC-3Ј; reverse primer, 5Ј-GCTGGCC-CTCGCTTATGA-3Ј; probe sequence, 5Ј-TGCCCAGCACTTCACGCAT-CAGTT-3Ј. Commercially available probe and primer sets from Applied Biosystems were used to measure IL-8 and RNase P1 expression. 1 g of total RNA was reverse transcribed using a reverse transcription kit from Applied Biosystems. Real time polymerase chain reaction was conducted using an Applied Biosystems kit and run on a 7900HT instrument (Applied Biosystems) according to the manufacturer's manual using RNase P1 gene as a standard. Relative gene expression is presented as a -fold induction as compared with control Ϯ S.E.
Transfection of A549 Cells with Small Inhibitory RNA (siRNA) Directed against PPAR-␥-siRNAs targeting bases 4 -23 of the PPAR-␥1 and 4 -24 of the PPAR-␥2 coding sequences were obtained from Integrated DNA Technologies (Coralville, IA). Untemplated TTs were added to the 3Ј-end of each strand. The siRNA sequences were 5Ј-GU-UGACACAGGAUGCCAUUTT-3Ј (PPAR-␥1) and 5Ј-GGUGAAACUCU-GGGAGAUUCTT-3Ј (PPAR-␥2). The single-stranded siRNAs were annealed by incubating a 100 mM concentration of each single strand in annealing buffer (100 mM potassium acetate, 30 mM Hepes, pH 7.4, 2 mM magnesium acetate) for 2 min at 90°C and slowly cooled down to room temperature. Cells grown in 6-well plates were transfected with 100 nM siRNA duplexes using LipofectAMINE reagent (Invitrogen) (5 ml in 1 ml of culture medium) for 5 h. After transfection, medium was changed, and cells were maintained in medium with fetal bovine serum for 16 h. The effect of siRNA on PPAR-␥1 and -␥2 protein expression was assessed by immunoblotting as described above. Cells were then treated with or without A23187 or EGF as specified below. The effect of treatment of cells with siRNA duplexes on IL-8 protein levels was determined by enzyme-linked immunosorbent assay of cellular supernatants and for COX-2 protein expression by immunoblotting of cell lysates.
Transfection of A549 Cells with siRNA Directed against cPLA 2 -RNA-DNA chimeras were synthesized by Integrated DNA Technologies. The sequence used to generate 100 nmol of siRNA duplex was: 5Ј-AAC UCU AGG GAC AGC AAC AUU TT; the complementary sequence was 5Ј-AAU GUU GCU GUC CCU AGA GUU TT, which corresponds to bases 299 -319 of the cPLA 2 coding sequence (GenBank TM accession number M68874). The annealing procedure was performed as described above. Medium (Ham's F-12 with glutamine) was incubated with LipofectAMINE (5 l/ml) (Invitrogen) with or without siRNA duplex (final concentration, 20 nM) for 20 min at room temperature. Cells were treated for 5 h at 37°C. Medium was then changed to Ham's F-12 with glutamine and fetal bovine serum, and the cells were incu-bated an additional 72 h. Cell lysate was collected for cPLA 2 Western blots. The remaining cells were incubated for 8 h with medium alone, A23187(10 Ϫ6 M), or EGF (20 ng/ml). Medium was collected for IL-8 assay by enzyme-linked immunosorbent assay, and cell lysate was collected for COX-2 and cPLA 2 Western blots.
Electrophoretic Mobility Shift Assay-PPRE probes were synthesized by Keystone Laboratories (Camarillo, CA) corresponding to PPRE sequences (underlined) present in the COX-2 (5Ј-GAGGCGACAGGTCAT-AACCCTACT-3Ј) and IL-8 promoters (5Ј-GGGTCCTCAGAGGTCAGA-CTTGGTGT-3Ј). The inverted PPRE sequence in the COX-2 promoter is at Ϫ3599 to Ϫ3573 relative to the transcription start site and is present as bases of Ϫ3542 to Ϫ3565 in GenBank TM accession number AF044206. The IL-8 PPRE sequence represents bases Ϫ1070 to Ϫ1045 relative to the transcription start site and is present as bases 412-437 of GenBank TM accession number M28130. Single-stranded nucleotides were reannealed by heating to 95°C for 5 min and cooled down slowly to room temperature. A549 cells were incubated with and without A23187 (10 Ϫ6 M) or EGF (10 ng/ml) for 30 min, 1 h, and 2 h prior to harvest at the 2-h time point. Nuclear extracts were prepared using a nuclear extraction kit according to the manufacturer's directions (Sigma). DNA binding was performed by incubating 3 g of nuclear protein in a total volume of 10 l of binding buffer (50 mM NaCl, 10 mM Tris-HCl, pH 7.5, 0.5 mM dithiothreitol, 0.5 mM EDTA, 1 mM MgCl 2 , 0.05 g/l poly(dI-dC)⅐poly(dI-dC), 15% glycerol) and 10,000 cpm 32 Plabeled double-stranded PPRE probes for 20 min at room temperature. The specificity of protein binding to labeled probe was assessed by competition with unlabeled probe or with PPRE consensus sequence. For the competition experiments, a 100ϫ excess of unlabeled probe or PPRE consensus sequence (CAAAACTAGGTCAAAGGTCA) (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the electrophoretic mobility shift assay mixture. Nuclear protein derived from cells exposed to A23187 (1 M) for 60 min was utilized for these experiments. Protein-DNA complexes were resolved on a 6% DNA retardation gel (Invitrogen) in 0.5ϫ Tris-borate-EDTA buffer at 200 V for 30 min. The dried gel was exposed to x-ray film (Kodak) with an intensifying screen at Ϫ70°C overnight or until adequate signal was developed. An Amersham Biosciences 301 computing densitometer was used to digitize images.
Statistical Analysis-Statistical analysis was performed using Microsoft Excel 2000 (Redmond, WA) software run on an iMAC computer (Apple, Cupertino, CA). Comparisons were performed using two-tailed unpaired Student's t tests. Values of p Ͻ 0.05 were considered statistically significant.
FIG. 1. Effect of the calcium ionophore A23187 or EGF treatment of A549 cells on steady state levels of IL-8 and COX-2 mRNA. These effects were at least in part inhibited by MAFP. Subconfluent A549 cell were exposed to A23187 (10 Ϫ6 M) (A), EGF (20 ng/ml) (B), Me 2 SO, or medium for specified times. RNA was isolated, and IL-8 and COX-2 mRNA levels were assayed by real time PCR as described under "Experimental Procedures." Data are expressed as a -fold increase from control Ϯ S.E. (n ϭ 3-4). *, p Ͻ 0.05 as compared with control. In C and D cells were pretreated with MAFP (10 mM) for 2 h and coincubated with A23187 (10 Ϫ6 M) (C) or EGF (20 ng/ml) (D) for 4 h. RNA was isolated, and IL-8 (closed bars) and COX-2 (open bars) mRNA levels were assayed by real time PCR as described under "Experimental Procedures." Data are expressed as a -fold increase from control Ϯ S.E. (n ϭ 3-4). *, p Ͻ 0.05 as compared with control. **, p Ͻ 0.05 as compared with cells exposed to A23187 or EGF and methyl acetate (a MAFP vehicle) only.

RESULTS
The Influence of cPLA 2 Activation on IL-8 and COX-2 mRNA Levels-Real time polymerase chain reaction was used to study change in steady state levels of IL-8 and COX-2 mRNA after treatment. Calcium ionophore A23187 and EGF increased both IL-8 and COX-2 mRNA levels as shown in Fig. 1, A and B. Treatment with the calcium ionophore A23187 resulted in an increase in IL-8 and COX-2 mRNA levels. The increase in the IL-8 transcript was present even after 24 h. By this time, the COX-2 mRNA levels had returned to control levels (Fig. 1A). EGF treatment of A549 cells also resulted in an increase in steady state mRNA levels for both IL-8 and COX-2. This effect was transient for COX-2 and disappeared after 2 h. IL-8 mRNA levels were still increased at the 24-h time point (Fig. 1B). These effects were in part decreased by MAFP as shown in Fig.  1, C and D.
The Influence of cPLA 2 Activation on IL-8 Reporter Gene Activity: the Effect of MAFP, a cPLA 2 Inhibitor-To test whether cPLA 2 overexpression might induce IL-8 transcrip-tion, a reporter gene assay was used. Cells were transfected with an IL-8 reporter gene and cotransfected with a cPLA 2 expression vector or empty vector. Twenty-four hours after transfection, cells were exposed for 4 h to culture medium with Me 2 SO, A23187 (10 Ϫ6 M), or EGF (20 ng/ml) to activate cPLA 2 . cPLA 2 overexpression induced an increase in IL-8 reporter gene activity as shown on Fig. 2A. When transfected cells were stimulated by A23187 (10 Ϫ6 M) for 4 h, IL-8 reporter gene activity was increased as compared with cells treated with Me 2 SO (A23187 vehicle). These results suggest that both cPLA 2 activation and cPLA 2 overexpression and activation induce IL-8 reporter gene activity. Further cPLA 2 activation induced by EGF caused a similar effect on IL-8 reporter gene activity (Fig. 2B). The effect of cPLA 2 activation (either by A23187 or EGF) was in part inhibited by preincubation with a specific cPLA 2 inhibitor, MAFP (10 M), as shown in Fig. 2, A  and B.
The Influence of cPLA 2 Activation on COX-2 Reporter Gene Activity: the Effect of MAFP, a cPLA 2 Inhibitor-Similar exper-FIG. 2. The influence of cPLA 2 activation on IL-8 (A and B) and COX-2 (C and D) reporter gene activity. Subconfluent A549 cells were transfected with 1.5 g of IL-8 or COX-2 reporter gene and 1.5 g of cPLA 2 overexpression vector or empty vector, respectively. Cells were cotransfected with 0.2 g of ␤-galactosidase/CMV vector to normalize for transfection efficiency. Some cultures were preincubated (for 2 h) and coincubated with 10 M MAFP or vehicle (methyl acetate). After preincubation, cells were treated with or without A23187 (10 Ϫ6 M) (A and C) or EGF (20 ng/ml) (B and D) for 4 h. Cell lysates were prepared and assayed for luciferase and ␤-galactosidase activity as described under "Experimental Procedures." Data shown are mean Ϯ S.E. (n ϭ 4 -6). *, p Ͻ 0.05 as compared with cells transfected with empty vector (VC) and treated with A23187 or EGF, respectively. **, p Ͻ 0.05 as compared with cells transfected with cPLA 2 expression vector and exposed to A23187 or EGF, respectively.
iments were performed to test the hypothesis that cPLA 2 overexpression might induce COX-2 transcription. cPLA 2 overexpression induced an increase of COX-2 reporter gene activity as shown on Fig. 2C. When transfected cells were stimulated with A23187 (10 Ϫ6 M) for 4 h, COX-2 reporter gene activity was increased compared with cells treated with Me 2 SO (A23187 vehicle). These data suggest that both cPLA 2 activation and cPLA 2 overexpression and activation induce COX-2 reporter gene activity. Furthermore treatment of cells with EGF produced a similar effect on COX-2 reporter gene activity. The effect of EGF was also in part inhibited by preincubation with a specific cPLA 2 inhibitor, MAFP (10 M), as shown on Fig. 2D.
The Influence of sPLA 2 Group IIA, sPLA 2 Group V, and sPLA 2 Group X Activation on IL-8 and COX-2 Reporter Gene Activity-cPLA 2 is not the only phospholipase A 2 expressed in human lung cells. Therefore, we used vectors overexpressing sPLA 2 Group IIA, Group V, and Group X proteins and tested whether overexpression of these enzymes might influence IL-8 and COX-2 reporter gene activity. IL-1␤ is known to induce the release of sPLA 2 from the cells to the medium. This process is associated with an increase in sPLA 2 enzyme activity in the extracellular space, activated by the calcium levels present in the medium. Previously we have shown that transfection of cells with an sPLA 2 Group IIA expression vector induces an increase in arachidonic acid release (4). In Fig. 3, A and B, we present evidence that transfection of cells with sPLA 2 Group V or Group X vectors, respectively, results in increased arachidonate release compared with cells transfected with empty vectors. These data suggest that these vectors produce functionally active proteins enhancing arachidonate release. Fig. 4, A, B, and C, demonstrate the lack of activation of IL-8 reporter gene in cells transfected with expression vectors encoding sPLA 2 Group IIA, V, or X. In all three cases, an increase in sPLA 2 activity is associated with a decrease in IL-8 reporter gene activity suggesting the possibility of an inhibitory effect of secreted phospholipase products on IL-8 transcription. Fig. 5, A, B, and C, demonstrate the influence of sPLA 2 activation on COX-2 reporter gene activity. Transfection of cells with expression vectors encoding for Group IIA, V, or X isoforms of secreted phospholipases A 2 did not activate COX-2 reporter gene expression. Activation of sPLA 2 activity seems to be associated with a decrease in COX-2 reporter gene activity suggesting the possibility of an inhibitory effect of secreted phospholipase products on transcription of this gene.
Effect of Inhibitors of sPLA 2 , iPLA 2 , or cPLA 2 on Spontaneous IL-8 and COX-2 Reporter Gene Activity in A549 Cells-To confirm the role of cPLA 2 in the regulation of IL-8 and COX-2 expression, cells transfected with IL-8 or COX-2 reporter genes were exposed for 12 h to sPLA 2 , iPLA 2 , or cPLA 2 inhibitors. Neither the sPLA 2 inhibitor thioetheramide-PC nor the iPLA 2 inhibitor BEL influenced IL-8 or COX-2 reporter gene activity suggesting no role for these enzymes in the regulation of IL-8 and COX-2 transcription de novo. MAFP, a cPLA 2 inhibitor, decreased both IL-8 and COX-2 reporter gene activity suggesting that cPLA 2 might regulate even basal IL-8 and COX-2 transcriptional activity (Fig. 6, A and B).
Effect of Inhibitors of sPLA 2 , iPLA 2 , or cPLA 2 on Calcium Ionophore and EGF-stimulated IL-8 and COX-2 Reporter Gene Activity in A549 Cells-To establish the role of cPLA 2 activation by A23187 and EGF in the regulation of IL-8 and COX-2 de novo transcription, cells transfected with IL-8 or COX-2 reporter genes were preincubated for 2 h with sPLA 2 , iPLA 2 , or cPLA 2 inhibitors. Afterward some cultures were exposed to A23187 (1 M) or EGF (20 ng/ml) for 4 h (with inhibitors present in the medium) or exposed to Me 2 SO, an A23187 vehicle, or medium, respectively. Neither the sPLA 2 inhibitor thio-etheramide-PC nor the iPLA 2 inhibitor BEL abolished the A23187 or EGF effect on IL-8 or COX-2 reporter gene activity suggesting no role for these enzymes in regulation of IL-8 and COX-2 transcription de novo upon calcium-dependent or EGF cell activation. MAFP, a cPLA 2 inhibitor, decreased calcium ionophore-stimulated IL-8 and COX-2 reporter gene activity suggesting that cPLA 2 might regulate IL-8 and COX-2 transcriptional activity (Fig. 6, C and D, respectively). Similarly MAFP decreased EGF-stimulated IL-8 and COX-2 reporter gene activity again suggesting that cPLA 2 might regulate IL-8 and COX-2 transcriptional activity (Fig. 6, E and F).
The Influence of cPLA 2 Activation on IL-8 Protein Expression in A549 Cells-The observation that cPLA 2 activation increases IL-8 reporter gene activity and IL-8 mRNA levels was supported by the measurements of IL-8 protein secreted by A549 cells. Exposure of A549 cells to A23187 resulted in a time-dependent increase in secretion of IL-8 as shown in Fig.  7A. A significant increase in IL-8 in the medium was present at 4 h and beyond. This effect was in part inhibited by MAFP suggesting that this effect might be cPLA 2 -dependent as shown  in Fig. 7B. A similar effect was obtained when A549 cells were exposed to EGF (20 ng/ml) as shown in Fig. 7C. The EGF effect was present at the 2-h time point. The effect of EGF was in part inhibited by MAFP as shown in Fig. 7D.
The Influence of cPLA 2 Activation on COX-2 Protein Expression in A549 Cells-The observation that cPLA 2 activation increases COX-2 reporter gene activity and COX-2 mRNA levels was supported by the assessment of COX-2 protein levels in  8 (A, C, and E) and COX-2 (B, D, and F) F). Subconfluent A549 cells were transfected with 1 g of IL-8 or COX-2 reporter gene, respectively. Cells were cotransfected with 0.2 g of ␤-galactosidase/CMV vector to normalize for transfection efficiency. After 4 h the medium was changed, and cells were exposed to BEL (1 M), an iPLA 2 inhibitor, MAFP (10 M), a cPLA 2 inhibitor, or thioetheramide-PC (10 M), an sPLA 2 inhibitor, for 12 h or medium with vehicle, which served as a control (A and B). Data are shown as mean Ϯ S.E. (n ϭ 4 -6). *, p Ͻ 0.05 as compared with control (cells exposed to Me 2 SO/MeAc, vehicles for PLA 2 inhibitors). Some cells were preincubated with BEL (1 M), an iPLA 2 inhibitor, MAFP (10 M), a cPLA 2 inhibitor, thioetheramide-PC (10 M), a sPLA 2 inhibitor, or vehicle for 2 h and then treated with A23187 (1 M) or Me 2 SO (an A23187 vehicle) for 4 h (C and D). Similarly, instead of A23187, EGF was utilized as an alternative cPLA 2 activator in E and F; cells exposed to medium alone were treated as a control. Cell lysate was prepared and assayed for luciferase and ␤-galactosidase activity as described under "Experimental Procedures." Data are shown as mean Ϯ S.E. (n ϭ 3-4). *, p Ͻ 0.05 as compared with control (cells exposed to Me 2 SO or medium, vehicles for A23187 or EGF, respectively). A549 cells by performing immunoblotting. Treatment of A549 cells with A23187 resulted in a time-dependent increase in cellular COX-2 protein levels as shown in Fig. 8A, left panel. This effect was in part inhibited by MAFP suggesting that this effect might be cPLA 2 -dependent as shown in Fig. 8A, right panel. When A549 cells were treated with EGF, increases in cellular levels of COX-2 protein were also noted (Fig. 8B, left  panel). This increase in COX-2 protein levels in response to EGF treatment was, in part, inhibited by pretreatment of the cells with MAFP (Fig. 8B, right panel).

reporter gene activity in quiescent cells (A and B) or after stimulation with A23187 (C and D) or with EGF (E and
The Effect of cPLA 2 Activation on IL-8 and COX-2 Protein Expression Is Decreased by siRNAs for PPAR-␥1 and PPAR-␥2-To confirm that the effect of cPLA 2 activation on IL-8 and COX-2 expression is indeed via PPRE activation, we used cells transiently transfected with siRNAs for both PPAR-␥ genes. C, subconfluent A549 cells were treated with EGF (20 ng/ml) for the specified times. D, A549 cells were preincubated with MAFP (10 M) for 2 h and incubated with EGF (20 ng/ml) for 8 h. Culture supernatants were collected, and IL-8 concentrations were measured by an immunoassay as described under "Experimental Procedures." Data are expressed as mean Ϯ S.E. (n ϭ 3-6). *, p Ͻ 0.05 as compared with control. **, p Ͻ 0.05 as compared with cells exposed to A23187 or EGF, respectively. A23187 or EGF. As shown in Fig. 9, B and C, transfection of A549 cells with siRNAs directed against PPAR-␥1 and PPAR-␥2 resulted in a diminished response to A23187 or to EGF in IL-8 and COX-2 protein expression suggesting that this effect is at least in part mediated through PPAR-␥1, PPAR-␥2, or both.
The Effect of Calcium Ionophore and EGF on IL-8 and COX-2 Protein Expression Is Decreased by siRNAs for cPLA 2 -To reconfirm that cPLA 2 activation may influence IL-8 and COX-2 expression we utilized cells transiently transfected with siR-NAs for the cPLA 2 gene. Figs. 10A and 11A present immuno-blots for cPLA 2 revealing its expression in control cells and a decrease in cPLA 2 expression in cells transiently transfected with siRNAs. These results suggest that the use of siRNA might be a useful tool in limiting cPLA 2 expression. These cells were used to study the influence of A32187 or EGF on IL-8 and COX-2 protein expression. As shown in Figs. 10B and 11B, transfection of A549 cells with siRNAs directed against cPLA 2 resulted in a diminished response to A23187 or to EGF in COX-2 and IL-8 protein expression suggesting that this effect is at least in part mediated through cPLA 2 . cPLA 2 Activation Results in an Increase in PPAR Binding to PPRE Sequences Derived from IL-8 and COX-2 Promoters-To confirm that as a result of cPLA 2 activation arachidonate or its metabolites might bind to PPAR and activate PPAR response elements, an electrophoretic mobility shift assay was performed. Probes representing PPAR response elements derived from the IL-8 or COX-2 promoters were used. Nuclear proteins were derived from cells treated with A23187 or EGF. Exposure of A549 cells to EGF increased nuclear protein binding to the probe derived from the IL-8 promoter with a maximum effect at 60 min (Fig. 12A, upper panel). Similarly nuclear protein isolated from cells exposed to A23187 exhibited an increased binding to the probe. Maximum binding was also obtained when cells were exposed to A23187 for 60 min (Fig. 12A, lower  panel). Exposure of A549 cells to EGF increased nuclear protein binding to the probe derived from the COX-2 promoter with maximum effect at 60 min (Fig. 12B, upper panel). Similarly nuclear protein isolated from cells exposed to A23187 exhibited an increased binding to the probe. Maximum binding was obtained when cells were exposed to A23187 for 60 min (Fig. 12B, lower panel). These data suggest that activation of transcription of IL-8 and COX-2 may be related to an increased binding of activated PPARs to PPREs in the promoter of these genes. To confirm the specificity of the binding to the PPRE in the IL-8 and COX-2 promoters, an inhibition experiment was performed showing that a 100ϫ excess of unlabeled probe or unlabeled PPRE consensus oligonucleotide significantly decreased the binding (Fig. 12, C and D). Both unlabeled oligonucleotide and PPRE consensus oligonucleotide markedly de- FIG. 9. The effect of cPLA 2 activation on IL-8 and COX-2 expression is mediated through PPAR-␥. Subconfluent A549 cells were transfected with siRNA against PPAR-␥1 and -␥2 as described under "Experimental Procedures." After 24 h cells were collected, and 10 g of cell lysate was immunoblotted as described under "Experimental Procedures." A represents an immunoblot developed with anti-PPAR-␥ antibody, which recognizes PPAR-␥1 and PPAR-␥2 protein.
The immunoblot shown is representative of three with similar results. After 24 h some cultures were exposed to A23187 (10 Ϫ6 M), EGF (20 ng/ml), or medium for 8 h. Supernatants were collected and assayed for IL-8. Cells were lysed and processed as described under "Experimental Procedures," and 20 g of cell lysate was subjected to immunoblotting. B, IL-8 protein levels in cellular supernatants measured by an immunoassay. Data are expressed as mean Ϯ S.E. (n ϭ 3-6). *, p Ͻ 0.05 as compared with control cells (cells treated with LipofectAMINE only). C, COX-2 protein production measured by immunoblotting. Cells were lysed and processed as described under "Experimental Procedures." 20 g of cellular lysate was subjected to immunoblotting. Blots shown are representative of three with similar results. Ctrl, control.
FIG. 10. The effect of A23187 and EGF on COX-2 expression is mediated through cPLA 2 . Subconfluent A549 cells were transfected with siRNA against cPLA 2 as described under "Experimental Procedures." After 72 h cells were collected, and 10 g of cell lysate was immunoblotted as described under "Experimental Procedures." A represents an immunoblot developed with anti-cPLA 2 antibody. The immunoblot shown is representative of three with similar results. After 24 h some cultures were exposed to A23187 (10 Ϫ6 M), EGF (20 ng/ml), or medium for 8 h. Cells were lysed and processed as described under "Experimental Procedures," and 20 g of cell lysate was subjected to immunoblotting. B shows COX-2 protein production measured by immunoblotting. Cells were lysed and processed as described under "Experimental Procedures." Blots shown are representative of three with similar results. creased labeled probe binding suggesting that these effects may be specific. DISCUSSION PPAR-␥ is one of the low affinity dietary lipid receptors reviewed by Rosen and Spiegelman (32). It is widely expressed in various tissues including liver and lung cells and cell lines (4,26,27,(33)(34)(35)(36). After binding a ligand (e.g. thiazolidinediones, arachidonate, 15-hydroxyeicosatetraenoic acid, or others), it forms a heterodimer with cis-retinoic acid retinoid X receptor, binds to a PPAR response element, and activates transcription of selected genes. It has been well documented that PPAR-␥ is able to activate genes such as glucose transporter 4, COX-2, IL-8, and others (22,37,38). COX-2 is one of two main isoforms of cyclooxygenase, an enzyme in the arachidonate metabolism cascade that leads to prostaglandin synthesis. Expression of COX-2 is increased by bacterial endotoxin, IL-1␤, tumor necrosis factor-␣, or interferon-␥. This is in contrast to COX-1, which is a housekeeping gene. The COX-2 transcription start site has been identified, and its promoter has been cloned. The promoter contains at least one PPAR response element, which is located at Ϫ3599 relative to the transcription start site. COX-2 reporter gene activity and protein expression is stimulated by PPAR-␥ agonists such as 15-⌬ 12,14 -prostaglandin J 2 , arachidonate, and thiazolidinedione (37,39,40). It also might be controlled by some non-steroidal anti-inflammatory drugs (37).
Here we report for the first time that cPLA 2 activation induces COX-2 and IL-8 expression. Five lines of evidence support this conclusion. First, treatment of A549 cells with activators of cPLA 2 , A23187 or EGF, increased activity of both COX-2 and IL-8 reporter genes. Furthermore overexpression of cPLA 2 and activation of cPLA 2 increased transcriptional activity of both COX-2 and IL-8 reporter genes. These effects were inhibited by coincubation with a cPLA 2 inhibitor, MAFP. Second, in cells transfected with the IL-8 reporter gene or the COX-2 reporter gene, coincubation the cPLA 2 inhibitor MAFP inhibited reporter gene expression. This effect was not noted with coincubation with the iPLA 2 inhibitor bromoenol lactone or with the sPLA 2 inhibitor thioetheramide-PC. Third, treatment of cells with A23187 or with EGF increased steady state levels of IL-8 mRNA and COX-2 mRNA. These effects were inhibited by coincubation with MAFP. Fourth, incubation of A549 cells with A23187 or with EGF increased release of IL-8 protein and increased cellular COX-2 protein levels. These effects were in part inhibited by treatment with MAFP. Finally the linkage of cPLA 2 activation and IL-8 or COX-2 expression to PPAR␥ activation was studied in three ways. First, electrophoretic mobility shift assays using PPRE sequence from the IL-8 and the COX-2 promoters demonstrated increased shift in both sequences in response to treatment of cells with A23187 or EGF suggesting that these sequences do bind PPARs after stimulation. Second, A549 cells were treated with siRNAs directed at PPAR-␥1 and -␥2. This treatment reduced the protein levels of PPAR-␥1 and -␥2. This treatment also reduced production of IL-8 protein and COX-2 protein in response to A23187 or to EGF compared with cells treated with transfecting reagent alone. Third, we used cells with diminished cPLA 2 expression (using siRNA) to confirm that the effect of A23187 and EGF on IL-8 and COX-2 protein expression is at least in part mediated through cPLA 2 activation. This again suggests that cytosolic phospholipase A 2 might play a role in regulation of IL-8 and COX-2 expression.
Interestingly there are no data available so far suggesting a role of other phospholipases A 2 such as Groups IIA, V, and X or iPLA 2 in this process. Previously we have shown that sPLA 2 Group IIA does not appear to be involved in activation gene transcription driven through PPARs (4). In this report we suggest that sPLA 2 types IIA, V, and X are not involved in activation of COX-2 and IL-8 transcription in response to A23187 or EGF. Two lines of evidence support this hypothesis. First, cotransfection of IL-8 or COX-2 reporter gene together with sPLA 2 Group IIA, V, or X did not increase reporter gene activity even after stimulation with IL-1␤. Second, only MAFP, a cPLA 2 inhibitor, decreased reporter gene activity when transfected cells were exposed to various PLA 2 inhibitors in both quiescent and A23187-or EGF-stimulated cells. All of these data taken together suggest that IL-8 and COX-2 transcription is not influenced by sPLA 2 Group IIA, V, or X activation. One of the functional differences between cPLA 2 and sPLA 2 is that upon activation cPLA 2 is translocated from the cytoplasm to various cell membranes (including the nuclear envelope but excluding the outer cell membrane), whereas sPLA 2 is actively secreted into intercellular space where it is activated. These data taken together with our previous observations might suggest that the translocation allows cPLA 2 to deliver arachidonate and its metabolites to the nucleus and therefore to regulate gene expression.
Studies involving cPLA 2 ␣ null mice have provided convincing evidence of the role of cPLA 2 in the inflammatory process in several diseases including but not limited to airway inflammation, arthritis, and acute lung injury (41)(42)(43). So far, its role in FIG. 11. The effect of A23187 and EGF on IL-8 expression is mediated through cPLA 2 . Subconfluent A549 cells were transfected with siRNA against cPLA 2 as described under "Experimental Procedures." After 72 h cells were collected, and 10 g of cell lysate was immunoblotted as described under "Experimental Procedures." A represents an immunoblot developed with anti-cPLA 2 antibody. The immunoblot shown is representative of four with similar results. After 72 h some cultures were exposed to A23187 (10 Ϫ6 M), EGF (20 ng/ml), or medium for 8 h. Medium was collected, and IL-8 concentrations were measured by enzyme-linked immunosorbent assay. Cells were lysed, and total cellular protein was measured as described under "Experimental Procedures." B, data expressed as mean Ϯ S.E. in picograms of IL-8 standarized for 1 g of cellular protein. *, p Ͻ 0.05 as compared with control cells.
these events was presumed to be related to prostaglandin and leukotriene synthesis. In this report, we suggest that the role of cPLA 2 in inflammation also might be related to a direct effect of its products on gene expression at least in part through PPAR-␥1 and -␥2 pathways.
Both stimuli (EGF and A23187) may alter gene expression FIG. 12. Cytosolic phospholipase A 2 activation increases nuclear protein binding to PPRE sequences derived from the IL-8 promoter (A) and from the COX-2 promoter (B). Subconfluent A549 cells were exposed to EGF (20 ng/ml, upper panels) or A23187 (1 M, lower panels) for the specified time intervals. To confirm specificity of the binding to PPRE sequences derived from the IL-8 promoter (C) and from the COX-2 promoter (D) a 100ϫ excess of the cold probe or PPRE consensus oligonucleotide, respectively, was added to the electrophoretic mobility shift assay mixture utilizing nuclear protein derived from cells exposed to A23187 (1 M). Nuclear proteins were isolated as described under "Experimental Procedures." Electrophoretic mobility gel shift assay was performed. After drying, the gel was exposed to Kodak MR film for 16 h or until adequate signal appeared. The autoradiograph shown is representative of three different experiments, each with similar results. The arrows indicate DNA-nuclear protein complexes. through other pathways. EGF acting through the EGF receptor is able to activate the mitogen-activated protein kinase cascade leading to transcription factor phosphorylation. Similarly A23187 increases intracellular calcium levels, which may activate various pathways including protein kinase A. These pathways might contribute to regulation of IL-8 and COX-2 expression in a cPLA 2 -independent way (33-35, 44 -48). Utilization of cells with decreased cPLA 2 expression provided the evidence that cPLA 2 or downstream products of this pathway might play a role in EGF-and calcium-dependent regulation of IL-8 and COX-2 expression in human lung cells.
Recently the role of sPLA 2 Group IIA in regulating sPLA 2 expression has been linked to cPLA 2 activation by secreted sPLA 2 in rat mesangial cells. This autocrine loop might utilize PPAR-␣ (36). As we have previously shown A549 cells do not express PPAR-␣ (4). This may explain why sPLA 2 overexpression did not increase IL-8 or cyclooxygenase-2 expression in A549 cells. Further studies investigating the role of enzymes such as COX-1, COX-2, 12-lipooxygenase, and 15-lipooxygenase and other lipid products of cPLA 2 activation are needed to clarify the role of cPLA 2 in the regulation of gene expression.