Prostaglandin E2 Induces Interleukin-8 Gene Transcription by Activating C/EBP Homologous Protein in Human T Lymphocytes*

The effect of prostaglandin E2 (PGE2) in regulating the synthesis of the pro-inflammatory chemokine inter-leukin-8 (IL-8) in T lymphocytes is not yet defined, even though it may reduce or enhance IL-8 synthesis in other cell types. Here, we demonstrate that, in human T cells, PGE2 induced IL-8 mRNA transcription through prostaglandin E2 receptors 1- and 4-dependent signal transduction pathways leading to the activation of the transcription factor C/EBP homologous protein (CHOP), never before implicated in IL-8 transcription. Several kinases, including protein kinase C, Src family tyrosine kinases, phosphatidylinositol 3-kinase, and p38 MAPK, were involved in PGE2-induced CHOP activation and IL-8 production. The transactivation of the IL-8 promoter by CHOP was NF-κB-independent. Our data suggest that PGE2 acts as a potent pro-inflammatory mediator by inducing IL-8 gene transcription in activated T cells through different signal transduction pathways leading to CHOP activation. These findings show the complexity with which PGE2 regulates IL-8 synthesis by inhibiting or enhancing its production depending on the cell types and environmental conditions.

Prostaglandin (PG) 1 E 2 , a product of the cyclooxygenation of arachidonic acid, is a potent mediator of immune responses (1)(2)(3) and inflammation (4 -6). PGE 2 is thought to down-regulate cell-mediated immunity by inhibiting the functional activities of macrophages (7)(8)(9)(10) and T cell proliferation and differentiation, expression of membrane receptors, secretion of diverse cytokines, and other effector functions in cellular immune reactions (11). On the other hand, it has been shown that PGE 2 is also able to stimulate some cellular immune activities, such as the expression and activation of matrix metalloproteases in HSB.2 human leukemic T cells (12)(13), the escape of some T cells from activation-induced apoptosis (14), and the macrophage production of interleukin (IL)-6 and IL-10 (15)(16)(17). Thus, it has been postulated that the heterogeneity of PGE 2 immunoregulatory effects may depend on distinctive signaling mechanisms of different PGE 2 receptors. PGE 2 also displays a complex regulatory function in IL-8 gene expression depending on the concentrations employed and the cell specificity. IL-8 is a potent chemoattractant for neutrophils in local inflammatory sites and is produced by a wide variety of cells in response to different stimuli (18). At physiological as well as pathological concentrations up to 100 M, PGE 2 is able to up-regulate endogenous IL-8 expression in human intestinal epithelial cells (19 -20) and to enhance IL-8 production in human synovial fibroblasts stimulated with IL-1␤ (21). In contrast, PGE 2 has no effect on neutrophil-derived IL-8 induced by lipopolysaccharide (22); down-regulates IL-8 in response to lipopolysaccharide in human alveolar macrophages and blood monocytes (23); and suppresses the production of chemokines, including IL-8, in human macrophages (24). The production of IL-8 in lymphocytes, viz. T cells, is crucial for the ability of the immune system to control infection and inflammation. However, despite the well known activity of PGE 2 as a mediator of immune and phlogistic responses, the role of this molecule in T cell production of IL-8 has not been well documented. In this study, we observed that PGE 2 significantly increased the amount of IL-8 by inducing gene transcription through a novel signaling mechanism mediated by C/EBP homologous protein (CHOP). We demonstrate that the signaling pathways triggered by EP1 and EP4 receptors and p38 MAPK are involved in this process. Therefore, we postulate that PGE 2 may enhance the phlogistic response by inducing the release of the chemokine IL-8 from activated T lymphocytes.
Cells and Culture Conditions-Peripheral T lymphocytes were isolated from human venous blood obtained from health adult volunteers as described previously (25). Briefly, diluted whole blood (100 ml) was layered in centrifuge tubes onto 120 ml of Histopaque-1077 gradient (Sigma). T lymphocytes were magnetically separated from B cells using colloidal paramagnetic microbeads conjugated to mouse anti-CD19 surface antigen monoclonal antibody expressed on B lineage cells using negative selection columns (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). T cell expression of activation markers such as CD25 and CD71 was tested using cytofluorometric analysis. Only CD25 Ϫ and CD71 Ϫ lymphocytes were processed. Jurkat T cells (clone E6.1) were obtained from the European Collection of Cell Cultures (Sigma) and maintained in a humidified atmosphere of 5% CO 2 and 95% air (37°C) in RPMI 1640 medium (Euroclone, Milan) supplemented with 50 IU/ml penicillin, 50 g/ml streptomycin (Euroclone), and 20% heat-inactivated fetal calf serum (Euroclone). Cells were routinely tested for mycoplasma infection, and cultures were renewed from frozen stocks every 2 months.
Cytokine Production Assay-T cells were cultured in round-bottom 24-well plates (1 ϫ 10 6 cells/well) in the presence of anti-CD3 (1 g/ml) plus anti-CD28 (250 ng/ml) antibodies (26). Cells were then immediately stimulated with PGE 2 , or PGI 2 , at different concentrations (stock solutions (100 mM) were prepared in ethanol and stored at Ϫ20°C) and cultured for 24 h. In some experiments, cells were treated with various protein kinase inhibitors and EPR1/2 antagonists 1 h prior to the addition of PGE 2 . In other experiments, cells were treated with EPR1-4 agonists only and cultured for 24 h. After stimulation, supernatants were collected and analyzed for IL-8 using an enzyme-linked immunosorbent assay (ELISA) kit (BioSource International, Nivelles, Belgium) according to the manufacturer's instructions.
RNA Isolation and Reverse Transcription-PCR-T cells were cultured in 35-mm well plates (5 ϫ 10 6 cells/well) in the absence or presence of anti-CD3 (1 g/ml) plus anti-CD28 (250 ng/ml) antibodies, and cells were then immediately stimulated for the indicated times with PGE 2 at final concentrations of 10 and 100 M. Total RNA was extracted with TRIzol (Invitrogen, Milan) according to the manufacturer's instructions. For reverse transcription-PCR, 1 g of total RNA was reverse-transcribed in a total volume of 20 l with an IMProm-II TM reverse transcriptase kit (Promega, Milan) according to the manufacture's instructions. 20 l of reverse transcription products were brought to a volume of 100 l containing 2 mM MgCl 2 , 0.2 mM PCR nucleotide mixture, a 1 M concentration of both the upstream and downstream PCR primers (Sigma), 5 units of Taq DNA polymerase (Transgenomic, Inc., Bergamo, Italy), and 10ϫ PCR buffer (Transgenomic, Inc.). Two pairs of primers were used in this study. The primer sequences were as follows: IL-8, 5Ј-ATG ACT TCCAAG CTG GCC GTG GCT-3Ј (sense) and 5Ј-TCT CAG CCC TCT TCA AAA ACT TCT C-3Ј (antisense); and ␤-actin, 5Ј-TGA CGG GGT CTA CCC ACA CTG TGC CCC ATC TA-3Ј (sense) and 5Ј-CTA GAA GCA TTG CGC TGG ACG ATG GAG GG-3Ј (antisense). Amplification was carried in a DNA thermal cycler (Applied Biosystems, Milan) after an initial denaturation at 94°C for 4 min. This was followed by 37 cycles of PCR using the following temperature and time profile: denaturation at 94°C for 3 min, primer annealing for 50 s at 58°C for IL-8 and at 62°C for ␤-actin, primer extension at 72°C for 1 min, and a final extension at 72°C for 7 min. The PCR products were visualized by electrophoresis on 1% agarose gel in 1ϫ buffer containing 89 mM Tris borate and 2 mM EDTA (pH 8.3) after staining with 0.5 g/ml ethidium bromide. The UV light-illuminated gels were photographed, and the relative sum intensity was calculated by normalizing the sum intensity of the IL-8 product to the ␤-actin mRNA control.
Protein Kinase C (PKC) Assay-T cells were grown in 10-cm dishes and, where indicated, transfected with PTGER1 small interfering RNA (siRNA). The cells were then incubated with the appropriate stimuli for 15 min and homogenized in 50 mM Tris-HCl (pH 7.5) containing 0.3% (w/v) ␤-mercaptoethanol, 5 mM EDTA, 10 mM EGTA, 50 g/ml phenylmethylsulfonyl fluoride, and 10 mM benzamidine. PKC activity in the lysates was determined using the protein kinase assay from Amersham Biosciences (Milan) according to the manufacturer's protocol.
Western Blot Analysis-Cells were cultured in 35-mm well plates at a concentration of 5 ϫ 10 6 cells/well. After stimulation, cells were lysed at 4°C in 150 l of M-PER lysis buffer (Pierce), 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, and 1 mg/ml leupeptin. Cell lysates were centrifuged at 14,000 ϫ g for 15 min at 4°C, and the supernatants were collected. Protein concentrations of extracts were determined using a standard Bradford protein assay (Bio-Rad, Milan). Cell lysates were dissolved in 4ϫ SDS-PAGE sample buffer containing 0.5 M Tris-HCl (pH 6.8) 10% SDS, 10% glycerol, 1% mercaptoethanol, and bromphenol blue and boiled for 3 min. After centrifugation at 14,000 ϫ g for 3 min, cell lysates were analyzed on 9% SDS-polyacrylamide gels and transferred to nitrocellulose membranes (Amersham Biosciences). The membranes were blocked in Tris-buffered saline/ Tween supplemented with 1% bovine serum albumin for 1 h. The blots were subsequently incubated with primary rabbit polyclonal antibodies (1:200 dilution) to the EP1 and EP4 receptors (Cayman Chemical) or primary phospho-specific rabbit IgG (1:500 dilution) recognizing p38 MAPK phosphorylated at Thr 180 and Tyr 182 or activating transcription factor-2 (ATF-2) phosphorylated at Thr 71 (Cell Signaling Technology, Milan). Primary anti-␤-actin monoclonal antibody (1:1000 dilution; Sigma) was used to normalize protein loading to that of specific proteins in each lane. After overnight incubation at 4°C, membranes were washed with Tris-buffered saline/Tween and incubated with horserad-ish peroxidase-conjugated donkey anti-rabbit (Amersham Biosciences) or goat anti-mouse (Cell Signaling Technology) IgG as secondary antibody at a dilution of 1:10,000 for 1 h at room temperature. Immunoreactive bands were detected by autoradiography using the SuperSignal West Pico chemiluminescent substrate system (Pierce) according to the manufacturer's instructions.
Transient Transfection-Liposome-mediated transient gene transfer was carried out with DMRIE-C (Invitrogen) as recommended by manufacturer. Briefly, cells were seeded at 2 ϫ 10 6 cells in 35-mm plates and transfected with a total of 2 g of plasmid DNA using 6 g of DMRIE-C/ml of RPMI 1640 medium without fetal calf serum. Transfected DNAs included 1 g of pFR-Luc, a luciferase reporter gene including multimerized Gal4 upstream activating sequences upstream of the minimal promoter, the transactivator Gal-CHOP (50 ng; Stratagene, La Jolla, CA), pSV-nlsLacZ DNA, a ␤-galactosidase expression vector (0.5 g), and an empty plasmid DNA (pBSM) at a final concentration of 2 g/plate. Jurkat cells (clone E6.1) were also transiently cotransfected with the pIL8-Luc plasmid (1 g; kindly provided by Dr. Hector R. Wong, Cincinnati Children's Hospital Medical Center, Cincinnati, OH), pSV-nlsLacZ DNA, and an expression plasmid for CHOP (0.2 g; kindly provided by Dr. Hidetoshi Hayashi, Nagoya City University, Nagoya, Japan) as described in the figure legends. In all experiments, transfections were stopped after 6 h by the addition of an equal volume of RPMI 1640 medium containing 20% fetal calf serum. 24 h after transfection, cells were treated with 10 or 100 M PGE 2 . After an additional 6 h, cells were harvested, and protein extracts were prepared for the luciferase activity assay using luciferin (Promega) as the substrate. Luciferase activity was normalized to ␤-galactosidase activity produced by cotransfected plasmid pnlsLacZ.
siRNA and Transfection-Double-stranded siRNA sequences targeting PTGER1 (GenBank TM /EBI accession number NM000955) and PTGER4 (accession number NM000958) mRNAs corresponding to the coding regions of residues 1009 -1027 (5Ј-CCAGCTTGTCGGTAT-CATG-3Ј) and 954 -972 (5Ј-TCAACCATGCCTATTTCTA-3Ј), respectively, were obtained from Dharmacon (Lafayette, CO). A nonspecific duplex (5Ј-CAGUGGAGAUCAACGUGCAAGUU-3Ј; Dharmacon) was used as a control, which did not significantly affect PTGER1 and PTGER4 mRNA and protein levels relative to the untransfected controls. Concentrations of siRNA and time of incubation were tested. EP1 and EP4 receptor knockdown reached the maximum at a concentration of 100 nM and at 24 -48 h post-transfection. After 72 h, there was a modest increase in EP1 and EP4 receptor levels compared with 48 h post-transfection. To assess gene silencing, the protein levels of the EP1 and EP4 receptors were determined by immunoblotting. Jurkat T cells were plated in 35-mm well plates in RPMI 1640 medium without fetal bovine serum and transfected with double-stranded siRNA using the DMRIE-C reagent according to the manufacturer's instructions.
24 h after the siRNA transfections, the cells were incubated in the absence or presence of anti-CD3 (1 g/ml) plus anti-CD28 (250 ng/ml) antibodies with PGE 2 at final concentrations of 10 and 100 M for an additional 24 h. All cells were harvested for total protein extraction, and supernatants were processed for IL-8 production by ELISA. In another set of experiments, transfection with the Gal-CHOP chimeric transcription factor was followed by siRNA (PTGER1 and PTGER4) transfection 4 h later.
Densitometry and Statistical Analysis-The relative intensities of protein and nucleic acid bands were analyzed using the Digital Sciences 1D program from Kodak Scientific Imaging Systems (New Haven, CT). Standard curves were run, and the data that were obtained were in the linear range of the curve. In addition, all values were normalized to their respective lane loading controls. Data are expressed as means Ϯ S.E. of n determinations. Results were analyzed by two-tailed Student's t test. p values Ͻ0.05 were considered significant.

IL-8 Protein Production after Stimulation with
Prostaglandins-To determine whether PGE 2 could enhance the synthesis of IL-8 protein, anti-CD3/CD28 antibody-stimulated T cells were cultured with different concentrations of PGE 2 , PGF 2␣ , or PGI2 , and IL-8 levels were measured in the supernatant by ELISA. Fig. 1 (A and B) shows that human peripheral T lymphocytes and Jurkat T cells produced significant elevated levels of IL-8 only when treated with PGE 2 at 10 or 100 M, with a higher significance at 100 M. Therefore, stimulation with PGE 2 leads to a dose-dependent release of IL-8 in T cells.  Fig. 2A and B, respectively). The increase in IL-8 mRNA was not observed at 8 h after PGE 2 treatment. The values of IL-8 mRNA were 2-4-fold over the basal expression of untreated cells when 10 M PGE 2 was employed, whereas they were 3-5-fold higher in cells treated with 100 M PGE 2 . These results suggest that PGE 2 increases IL-8 production at the transcriptional level in a time-and dose-dependent fashion. PGE 2 Stimulates IL-8 Production via the EP1 and EP4 Receptors-To determine the EP receptors involved in IL-8 pro-duction by PGE 2 , cells were stimulated with the EP1 agonist PT-PGE 2 (27), the EP2 receptor agonist butaprost (27), the EP3 receptor agonist sulprostone (27), and the EP4 agonist 11deoxy-PGE 1 (28). PT-PGE 2 and 11-deoxy-PGE 1 induced a significant increase in IL-8 release, whereas butaprost and sulprostone were not able to induce chemokine synthesis (Fig. 3A).
In parallel experiments, both the selective EP1 antagonist SC-19220 (28) and the EP1/2 antagonist AH 6809 (29) significantly inhibited the production of IL-8 induced by either 10 or 100 M PGE 2 (Fig. 3B). The effect of AH 6809 was most probably due to inhibition of EP1, but not EP2, since butaprost had no effect on IL-8. To further verify whether the effects of PGE 2 on IL-8 production are mediated through the EP1 and EP4 receptors, we performed post-transcriptional gene silencing experiments for both receptors using synthetic siRNAs and evaluated IL-8 production. Under optimal conditions (100 nM siRNA, 24 h post-transfection), we achieved 70 and 85% knockdown of the EP1 and EP4 receptors at the protein level, respectively (Fig. 3,  C and D, upper). The EP1 and EP4 receptor silencing by specific siRNAs reduced the IL-8 production in PGE 2 -treated cells by ϳ35 and 45%, respectively (Fig. 3, C and D, lower). In contrast, the nonspecific control duplex did not affect IL-8 synthesis by PGE 2 . Therefore, these data indicate that the pro-inflammatory action of PGE 2 is exerted via the EP1 and EP4 receptors.
Inhibition of PKC, Src Family Kinases, and Phosphatidylinositol 3-Kinase (PI3K), but Not Protein Kinase A (PKA), Blocks the PGE 2 -induced Synthesis of IL-8 -The EP1 receptor induces changes in intracellular calcium levels that are believed to be mediated either by the phospholipase C signaling pathway, most likely via G␣ q , or via direct activation of the calcium channel (28). The signal transduction pathway activated by EP1 is represented mainly by PKC (30,31) and Src family kinases (32). To further determine the signaling mechanisms of PGE 2 , cells were pretreated with the PKC inhibitors bisindolylmaleimide I, staurosporine, and H-89 or with the specific Src tyrosine kinase inhibitor PP2, and complete dose responses were evaluated (data not shown). At the concentrations reported in Fig. 4A, all compounds were able to completely suppress IL-8 synthesis induced by 10 or 100 M PGE 2 .
It has been shown that the EP4 receptor primarily utilizes a PI3K-dependent pathway and is less efficiently coupled to adenylyl cyclase and/or has additional pathways not involving cAMP/PKA signaling (33). Thus, we examined the role of these signaling pathways in the enhancement of IL-8 release by PGE 2 using LY 294002, an inhibitor of PI3K, at a concentration of 20 M, as resulted from a dose-response curve (data not shown), and H-89 in a nanomolar range, which inhibits PKA only (34). As shown in Fig. 4B, LY 294002 inhibited IL-8 production by 40%, whereas H-89 treatment did not interfere with the increase in IL-8 synthesis induced by the two concentrations of PGE 2 employed. These results suggest that PKC, Src family kinases, and PI3K, but not PKA, participate in IL-8 synthesis.
PGE 2 Induces PKC Activation in T Cells-To support the evidence of the involvement of PKC in PGE 2 induction of IL-8 synthesis, a PKC activity assay was performed as described under "Materials and Methods." Fig. 5 shows that PGE 2 (10 or 100 M) induced a significant increase in basal PKC activity. The same effect was observed in anti-CD3/CD28 antibodystimulated T cells (data not shown). Pretreatment with the PKC inhibitors completely suppressed the PGE 2 -induced PKC activity enhancement. To examine whether the modulation of PKC is mediated through the EP1 receptor, the cells were treated with EP1 agonist (PT-PGE 2 ) or PGE 2 and EP1 antagonist (SC-19220). PT-PGE 2 caused a significant increase in PKC, whereas SC-19220 completely reverted the stimulatory effect of PGE 2 on PKC activity. Similar results were obtained in cells transfected with PTGER1 siRNA and treated with PGE 2 (data not shown).
Induction of IL-8 Synthesis by PGE 2 Is Not Mediated by NF-B-IL-8 is an important NF-B-regulated gene (35). To verify whether NF-B activation is essential in the production of PGE 2 -induced IL-8 in activated human T cells, Jurkat T cells were treated with the NF-B inhibitor caffeic acid phenethyl ester (1-20 M) 1 h prior to stimulation with PGE 2 (10 or 100 M). None of the caffeic acid phenethyl ester concentrations employed was able to modify the effects of PGE 2 on IL-8 production (data not shown), thus suggesting that the induction of IL-8 by PGE 2 in human T cells is NF-B-independent.

IL-8 Induction by PGE 2 in Human T Cells Occurs via p38 MAPK-
The mechanism of IL-8 induction has been reported to be mediated through the p38 MAPK pathway (36,37). To verify that PGE 2 enhances IL-8 synthesis from stimulated T cells via this MAPK pathway, the p38 MAPK inhibitor SB 203580 (38) was added to cultures 1 h before treatment with PGE 2 (10 or 100 M). Fig. 6 shows that this inhibitor significantly reduced PGE 2induced IL-8 production in T cells in a dose-dependent manner. PGE 2 Activates p38 MAPK in T Cells-As inhibition of p38 MAPK prevented PGE 2 -induced IL-8 synthesis, we investigated the effect of PGE 2 on the phosphorylation of p38 MAPK. After treating cells with PGE 2 for different times (15-480 min), total cell proteins were extracted and subjected to Western blotting with an antibody specific for the phosphorylated form of p38 MAPK. As shown in Fig. 7A, the phosphorylation of p38 MAPK occurred upon treatment of cells with either 10 or 100 M PGE 2 after 60 min and persisted for up to 360 min. The phosphorylation of p38 MAPK induced by PGE 2 resulted in the activation of the kinase since ATF-2, a substrate of p38 MAPK, was also phosphorylated with the same pattern. The specific inhibitor SB 203580 was able to prevent the phosphorylation as well as the activation of p38 MAPK induced by PGE 2 (Fig. 7B).
Activation of p38 MAPK by PGE 2 Is Not Dependent on the PKC, Src Tyrosine Kinase, PI3K, and PKA Pathways-To investigate whether PKC, Src family kinases, PI3K, and PKA are involved in the phosphorylation of p38 MAPK induced by PGE 2 , Jurkat T cells were treated with the specific PKC inhibitor bisindolylmaleimide I, the specific Src tyrosine kinase inhibitor PP2, the PI3K/Akt inhibitor LY 294002, or the PKA inhibitor H-89 1 h prior to treatment with PGE 2 and examined for the activated form of p38 MAPK by Western blot analysis. Treatment with bisindolylmaleimide I, PP2, LY 294002, or H-89 for 1 h did not affect the phosphorylation of p38 MAPK induced by PGE 2 (Fig. 8). Similar results were obtained after treatment with 10 M PGE 2 (data not shown).

PGE 2 -induced IL-8 Synthesis Is Regulated by the Transcription
Factor CHOP-To determine whether PGE 2 regulates IL-8 gene expression, experiments were performed by transfecting resting and activated Jurkat T cells with the wild-type IL-8 promoter, and after treatment with PGE 2 , luciferase activity in cell lysates was measured. Fig. 9 shows that PGE 2 slightly activated the IL-8 promoter in resting T lymphocytes, whereas it was able to further increase anti-CD3/CD28 antibody-induced gene expression. The cotransfection of CHOP up-regulated the luciferase response with respect to anti-CD3/CD28 antibody-stimulated cells. Treatment of cotransfected stimulated cells with PGE 2 further increased the luciferase activity, indicating that PGE 2 induces transcription of the IL-8 gene by activating the transactivation of CHOP. A very high significance was obtained with either 10 or 100 M PGE 2 . PGE 2 Induces the Transactivation Potential of CHOP-Since we demonstrated that the induction of IL-8 synthesis by PGE 2 occurs through p38 MAPK activation, we investigated whether PGE 2 induces the potential transactivation of CHOP, the main substrate of p38 MAPK. Jurkat T cells were transiently co-transfected with the p38 MAPK-responsive Gal-CHOP chimeric transcription factor, which consists of the DNA-binding domain of yeast Gal4 and the transactivation domain of CHOP, together with a luciferase reporter plasmid (pFR-Luc) and a bacterial ␤-galactosidase expression vector (pSV-nlsLacZ). Cells were then treated with 10 or 100 M PGE 2 at different time intervals. Control cells were treated with ethanol only. As shown in Fig. 10, the luciferase activity of Gal-CHOP exhibited a 3-or 4-fold increase over the base-line levels of ethanoltreated cells at 6 h after treatment with 10 or 100 M PGE 2 , respectively. A less but also significant increase occurred at 4 h with either 10 or 100 M PGE 2 . These effects required the CHOP transactivation domain since treatment with PGE 2 failed to induce reporter expression in cells transfected with a plasmid encoding the Gal4 DNA-binding domain only (residues 1-147) and lacking the transactivation domain of CHOP (data not shown). PGE 2 Activates the Transcriptional Activity of CHOP via the EP1 and EP4 Receptors-Since our data demonstrated an involvement of the EP1 and EP4 receptors in IL-8 production, we investigated whether these receptors are also involved in CHOP activation. The EP1 agonist PT-PGE 2 and the EP4 agonist 11-deoxy-PGE 1 significantly increased the transcriptional activity of CHOP at a level 2-2.5-fold higher than in ethanol-treated cells. In contrast, the EP2 agonist butaprost and the EP3 agonist sulprostone did not induce the activation of CHOP-mediated transcription (Fig. 11A). Cotransfection experiments with Gal-CHOP and PTGER1 and PTGER4 siRNAs confirmed the exclusive EP1 and EP4 receptor involvement in the PGE 2 -induced activation of CHOP (Fig. 11B).
PKC, Src Family Kinases, PI3K, and p38 MAPK, but Not PKA, Are Involved in the PGE 2 -induced Transcriptional Activation of CHOP-To investigate the kinase pathways involved in the PGE 2 -induced transcriptional activation of CHOP, transfection experiments were performed with inhibitors of PKC, Src tyrosine kinases, PI3K, p38 MAPK, and cAMP/PKA. Fig. 12 shows that the Gal-CHOP activity stimulated by 10 or 100 M PGE 2 was completely blocked by both staurosporine and bisindolylmaleimide I. A significant inhibition of CHOP activation was also caused by treatment with PP2, LY 294002, and SB 203580. H-89, at a nanomolar concentration inhibiting PKA (34), did not interfere with the transcriptional activation of CHOP induced by PGE 2 .

DISCUSSION
In this study, we have demonstrated that PGE 2 at physiological as well as pathological concentrations, usually reached in several neoplastic, phlogistic, and immunological processes (5, 6, 20, 39), induced IL-8 production in peripheral as well as Jurkat T cells. These results are of particular interest since the FIG. 6. IL-8 protein synthesis induced by PGE 2 is mediated by the activation of p38 MAPK. Jurkat T cells stimulated with anti-CD3 (1 g/ml) plus anti-CD28 (250 ng/ml) antibodies were pretreated with the specific p38 MAPK inhibitor SB 203580 (SB) 1 h prior to incubation with PGE 2 (10 or 100 M) for 24 h. IL-8 protein was measured in cell supernatants by ELISA. Data are depicted as means Ϯ S.E. of five independent experiments. * and **, p Ͻ 0.05 and 0.01, respectively, versus control untreated cells; E and EE, p Ͻ 0.05 and 0.01, respectively, versus 10 M PGE 2treated cells; ϩ and ϩϩ, p Ͻ 0.05 and 0.01, respectively, versus 100 M PGE 2treated cells (based on two-tailed Student's t test).
other PGs employed (PGF 2␣ and PGI 2 ) were ineffective. PGE 2 is ubiquitously synthesized and reaches high levels in inflammatory processes, characterized by infiltrates of neutrophils, macrophages, lymphocytes, and plasma cells (40), following the induced expression of cyclooxygenase-2. Evidence has been reported that 15-deoxy-⌬ 12,14 -PGJ 2 (an active metabolite of PGD 2 ) increases the synthesis of IL-8 in peripheral human T lymphocytes and several human T cell lines (26). In contrast, PGE 2 was tested only on CCRF-CEM cells (acute human lymphoblastic leukemia) and was ineffective also at the concentrations we showed to be able to induce IL-8 synthesis by peripheral and Jurkat T cells. The results obtained by us therefore highlight a new important role for PGE 2 in regulating IL-8 production by T lymphocytes, confirming the pro-inflammatory activity of this prostaglandin.
A variety of transcription factors such as NF-B, NF-IL6, activator protein-1, and octamer-1 have been shown to regulate IL-8 gene transcription (41)(42)(43)(44). To our knowledge, the data reported here describe for the first time that CHOP functions as a transcription factor that regulates IL-8 mRNA transcription. In fact, a regulatory role of CHOP has been shown only in IL-6 gene transcription in human melanoma cells (45,46). Although inflammatory stimuli are known to activate IL-8 gene transcription through NF-B, our study shows that PGE 2 induces IL-8 synthesis through an NF-B-independent pathway.
Using agonists and antagonists of different EP receptor subtypes and specific PTGER1 and PTGER4 siRNAs, we have shown that the EP1 and EP4 receptors, but not the EP2 and EP3 receptors, are involved in PGE 2 -induced IL-8 synthesis and CHOP activation in T cells. Under optimal conditions (100 nM siRNA, 24 or 48 h post-transfection), we achieved an average of 80% knockdown of the EP1 and EP4 receptors at the protein level. Recent data suggest that a complete knockdown is hard to be reached in Jurkat T cells (maximum of 70 -80%) because the silencing is less effective in these cells than in others (47). We demonstrated that EP1 triggering by PGE 2 leads to the activation of PKC, most likely via G␣ q (48,49), which then increases IL-8 gene transcription. Although it was previously believed that the interaction of PGE 2 with the EP4 receptor transiently increases intracellular cAMP and activates PKA, which then phosphorylates downstream effector proteins such as CREB (50). A recent report (34) indicates that EP4 binding induces mainly the activation of PI3K. Our results with the PKA and PI3K inhibitors clearly show that PKA is not involved in PGE 2 -induced CHOP activation and IL-8 synthesis and that PI3K participates in augmenting the transactivation activity of CHOP in the IL-8 promoter. Fiebich et al. (29) hypothesized that the activation of PKC by PGE 2 may be mediated by a possible EP4-like receptor in astroglial cells. Our data rule out the involvement of an EP4-like receptor and lead to the hypothesis that PGE 2 induces IL-8 synthesis in T lymphocytes through the canonic EP4 receptor and EP1-mediated PKC and Src kinase activation.
We have shown that PGE 2 activates CHOP and IL-8 gene transcription not only through the activation of PKC and Src family tyrosine kinases triggered by EP1 and PI3K/Akt triggered by EP4, but also through the activation of p38 MAPK.
Interestingly, p38 MAPK led to the activation of CHOP independently of the other kinases since the PKC inhibitors employed did not block the phosphorylation of p38 MAPK induced by PGE 2 , even though they were able to completely suppress CHOP activation and IL-8 production. In agreement with our results, the activation of p38 MAPK by PGE 2 has been reported to be independent of the activation of PKC in the enhancement of IL-6 synthesis in astrocytes (29). Since, in other cell systems, the activation of PKC by stimuli other than PGE 2 is able to activate p38 MAPK, it may be hypothesized that not only the cell type, but also the stimulus employed has a pivotal role in regulating the activity of PKC. In fact, in the human lung adenocarcinoma cell line H441, hepatocyte growth factor, but not phorbol 12-myristate 13-acetate, simultaneously activates PKC and p42/44 and p38 MAPKs through parallel and nonintersecting pathways (51). Recently, mechanisms linking G protein-coupled receptors to the MAPK cascade have been investigated in different cell types (51), but few data are available regarding the role of tyrosine kinases in activating the p38 MAPK pathway. In human embryonic kidney 293 cells and in neutrophils, the activation of p38 MAPK by G␣ q/11 is mediated by the Src family kinase-dependent pathway (52,53). Our studies employing a protein tyrosine kinase inhibitor (PP2) showed that the PGE 2 -induced activation of p38 MAPK is independent of Src tyrosine kinases, which are instead involved in CHOP activation and IL-8 production. Therefore, similar to PKC, PGE 2 -activated Src tyrosine kinases may also exert a different role in the activation of p38 MAPK, which may depend on cell specificity. However, additional experiments are needed to clarify the signaling pathway that links the EP receptors and the phosphorylation and thus the activation of p38 MAPK. Nevertheless, the involvement of PKC and Src tyrosine kinases appears to be necessary in CHOP activation and IL-8 synthesis since their inhibition completely blocked the effects of PGE 2 . In contrast, the role of p38 MAPK and PI3K appears to be important but not essential in that the p38 MAPK inhibitor SB 203580 and the PI3K inhibitor LY 294002 were able to inhibit only partially the PGE 2 -induced transcriptional activity of CHOP and the synthesis of IL-8. Since none of the kinases activated by EP1 and EP4 binding to PGE 2 induced the phosphorylation of p38 MAPK, it may be hypothesized that these two receptors are not involved in p38 MAPK activation. Therefore, we propose that, in stimulated human T cells, PGE 2 induces IL-8 mRNA transcription by the activation of several signal transduction pathways, including the p38 MAPK, EPR1triggered PKC and Src family tyrosine kinase, and EPR4triggered PI3K pathways (Fig. 13).
In conclusion, the findings described in this study show that PGE 2 induces activated human T lymphocytes to produce the chemokine IL-8. This IL-8 production is through the PKC, Src, and p38 MAPK pathways and is NF-B independent. Our  results indicate that PGE 2 plays a key role in enhancing and sustaining inflammation via stimulation of IL-8 synthesis by activated T cells. Therefore, PGE 2 has profoundly different effects on the production of this chemokine that are dependent on the cell type and environment. The characterization of the pathways stimulated in T cells by cyclooxygenase-2-dependent PGE 2 may have practical clinical impact in addition to pathophysiological significance. Pro-inflammatory stimuli increase the levels of PGE 2 , which induces IL-8 transcription through Src family kinases. PKC is activated via the EP1 receptor, and PI3K is activated via the EP4 receptor and p38 MAPK. These pathways are independently involved in the phosphorylation and activation of CHOP, which transactivates the IL-8 promoter and induces IL-8 protein synthesis. IL-8 release induced by PGE 2 is independent of the activation of PKA and the EP2 and EP3 receptor-dependent signaling mechanisms. TNF-alpha, tumor necrosis factor-␣; LPS, lipopolysaccharide.