Cyclooxygenase-2 inhibitor (SC-236) suppresses activator protein-1 through c-Jun NH2-terminal kinase

GASTROENTEROLOGY 2004;126:136 –147

Cyclooxygenase-2 Inhibitor (SC-236) Suppresses Activator Protein-1 Through c-Jun NH2-Terminal Kinase BENJAMIN CHUN–YU WONG,* XIAO HUA JIANG,*,‡ MARIE C. M. LIN,§ SHUI PING TU,*,‡ JIAN TAO CUI,* SHI HU JIANG,‡ WAI MAN WONG,* MAN FUNG YUEN,* SHIU KUM LAM,* and HSIANG FU KUNG§ *Department of Medicine, Queen Mary Hospital, University of Hong Kong, Hong Kong; ‡Department of Gastroenterology, Rui-jin Hospital, Shanghai, Peoples Republic of China; and §Institute of Molecular Biology, the University of Hong Kong, Hong Kong

Background & Aims: Aspirin exerts antitumor effect partly through blocking tumor promoter-induced activator protein-1 (AP-1) activation. The aim of this study is to determine how specific COX-2 inhibitor SC-236 mediates antitumor effect by modulation of AP-1-signaling pathway. Methods: AP-1 transcriptional activity and DNA-binding activity were detected by luciferase reporter assay and gel shift assay, separately. Mitogenactivated protein kinase (MAPK) activation was determined by Western blot and in vitro kinase assay. Antisense oligonucleotide against c-Jun-N-terminal kinase (JNK) was used to suppress JNK expression. Results: We showed that SC-236 inhibited 12-O-tetradecanoylphorbol-13-acetate (PMA)-induced cell transformation in a dose-dependent manner in JB6 cells. At a dose range (12.5–50 ␮mol/L) that inhibited cell transformation, SC-236 also inhibited anchorage-independent cell growth and AP-1-activation in 3 gastric cancer cells, independent of COX-prostaglandin synthesis. SC236 down-regulated c-Jun-NH2-terminal kinase phosphorylation and activity. Suppression of JNK activity reversed the inhibitory effect on AP-1 activity by SC-236 and suppressed gastric cancer cell growth, indicating that the inhibitory effect of SC-236 on AP-1 activation and cell growth was through interaction with JNK. Conclusions: The inhibitory effect on JNK-c-Jun/AP-1 activation contributes to the antitumor effect of COX-2specific inhibitor, and inhibition of JNK activation may have a therapeutic benefit against gastric cancer.

spirin, nonsteroidal anti-inflammatory drugs (NSAIDs), and cyclooxygenase-2 (COX-2) inhibitors have received great attention because of their protective effects against cancer.1– 4 COX inhibition may explain part of the antitumor activity of NSAIDs. However, NSAIDs also modulate cyclooxygenase-independent signal transduction pathways.5–7 Mechanisms such as induction of apoptosis8 –12 and inhibition of angiogenesis13,14 are observed only at high concentrations of the respective NSAIDs, which are 100- to 1000-fold higher than those

A

needed to inhibit prostaglandin synthesis, in all cell culture studies. Furthermore, a recent clinical study on the inhibitory effect of celecoxib in colonic adenomas in patients suffering from familial adenomatous polyposis showed that a high dose is needed.15 These data indicate that cyclooxygenase-independent mechanisms are important for the antitumor effect of NSAIDs at high doses. Activator protein 1 (AP-1) is an inducible eukaryotic transcription factor comprised of the products of the Jun and Fos oncogene families.16 –18 Several reports have established the role of AP-1 activation in cellular transformation and tumor promotion.19 –21 In JB6 mouse epidermal cell lines, 12-O-tetradecanoylphorbol-13-acetate (PMA) and epithelial growth factor (EGF) induce AP-1 transcriptional activity in promotion-sensitive (P⫹) phenotypes but not in promotion-resistant (P⫺) phenotypes. In contrast, blocking AP-1 induction causes P⫹ cells to revert to the P⫺ phenotype, indicating the unique requirement for AP-1 activity in cell transformation.21 Mitogen-activated protein kinase (MAPK)-mediated phosphorylation is important for the expression and posttranslational modification of AP-1 complex.22,23 As with other MAP kinase cascades, the JNK/c-Jun pathway has been shown to regulate cell growth and tumorigenesis in many different malignancies.24 –26 It is established that Ras-induced transformation requires c-Jun27 and that Ras induces c-Jun phosphorylation on sites that are phosphorylated by JNK.28 In addition, it has been reported that JNK is constitutively activated in several tumor cell lines and that the transforming actions of several oncogenes have been reported to be JNK depenAbbreviations used in this paper: AP-1, activator protein 1; COX, cyclooxygenase; ERK, extracellular signal-regulated protein kinase; JNK, c-Jun-N-terminal kinase; MAPK, mitogen-activated protein kinase; NSAIDs, nonsteroidal anti-inflammatory drugs; PGE2 , prostaglandins E2; PMA, 12-O-tetradecanoylphorbol-13-acetate. © 2004 by the American Gastroenterological Association 0016-5085/04/$30.00 doi:10.1053/j.gastro.2003.10.063

January 2004

dent.29 These data strongly support the hypothesis that JNK is relevant to cancer. A variety of NSAIDs such as aspirin, sodium salicylate, and sulindac have been shown to inhibit cell transformation and tumor promotion by blocking tumor promoter-induced AP-1 activation in both mouse JB6 cells and human tumor cells.20,30 –33 However, whether inhibition of AP-1 activity contributes to the antitumor effect of specific COX-2 inhibitors is still not clear. In the present study, we show that the specific COX-2 inhibitor SC-236 inhibits anchorage-independent cell growth and AP-1 activation in gastric cancer cells. The inhibition of AP-1 activation and cell growth occurs independent of COX-2-prostaglandin synthesis but through suppression of stress-activated protein kinase/ Jun-terminal kinase JNK cascade. To our knowledge, this is the first report showing that specific COX-2 inhibitor suppresses tumor growth through inhibition of AP-1 activity and that down-regulation of SAPK/JNK is essential for inhibition of AP-1 activation in cancer by NSAIDs.

Materials and Methods Cell Culture and Drug Treatments Three gastric cancer cell lines were used in this study. AGS was purchased from the American Type Culture Collection (ATCC, Rockville, MD). MKN-28 and MKN-45 were purchased from RIKEN (The Institute of Physical and Chemical Research), Cell Bank, Japan. Cells were maintained in RPMI-1640 containing 10% fetal bovine serum (FBS), 100 U/mL⫺1 penicillin, 100 ␮g/mL⫺1 streptomycin (Gibco BRL, Life Technologies, NY). Mouse epidermal JB6 P⫹ with AP-1 luciferase reporter stable transfectant cells were kindly provided by Dr. Li JJ (National Cancer Institute, Frederick, MD) and grown in Eagle’s minimal essential medium supplemented with 200 ␮g/mL gentamicin.21 SC-236 (a COX-2 specific inhibitor) was obtained from Searle, IL. Aspirin, NS-398, Nimesulide, and Sunlindac Sulfone were purchased from Sigma (St Louis, MO). PGE2 , PGH1 , and PGH2 were purchased from Cayman Chemicals (Ann Arbor, MI). MEK1/2 inhibitor U0126 was from New England Biolabs (Boston, MA). JNK inhibitor SP600125 was from Calbiochem (La Jolla, CA).

MTT Assay About 5000 cells per well were grown in 96-well microtiter plates and incubated overnight in 100 ␮L of culture medium. Cells were then treated with different concentrations of SC-236 for fixed time intervals. Ten microliters MTT (Fluka, Buchs, Switzerland) labeling reagent (final concentration 0.5 mg/mL) was added into each well, and the cells were incubated for another 4 hours at 37°C. The supernatant was removed, and 100 ␮L of 0.04 mol/L hydrochloric acid in

COX-2 INHIBITOR SUPPRESSES AP-1

137

isopropanol was added to each well. A micro ELISA reader (Bio-Rad, Hercules, CA) measured the absorbency at a wavelength of 595 nm.

Oligonucleotide Treatment The phosphorothioate oligonucleotides used in this study were synthesized and purified by Genset Singapore Biotech Ltd. (Singapore). The sequences of the oligonucleotides used are as follows: JNK1AS, 5⬘-CTC TCT GTA GGC CCG CTT GG-3⬘; JNK2AS, 5⬘-GTC CGG GCC AGG CCA AAG TC-3⬘; JNK1Scr, 5⬘-CTT TCC GTT GGA CCC CTG GG-3⬘; and JNK2Scr, 5⬘-GTG CGC GCG AGC CCG AAA TC-3⬘ as described previously.34,35 Cells growing in log phase were treated with 0.4 ␮mol/L oligonucleotides in the presence of 6 ␮L Lipofectamine 2000 reagent (Invitrogen) per milliliter. Forty-eight hours after treatment, the cells were collected for Western blot analysis or JNK activity assay.

Proliferation Assay Cells were seeded in 24-well tissue culture plates in the presence of RPMI1640 supplemented with 10% FBS. One day later, they were transfected with 0.4 ␮mol/L JNKAS, and JNK scrambled oligonucleotide as described above. After 4 hours of lipofection, the cells were transferred to medium supplemented with 2% FBS. Five days later, the cells were counted with a Coulter counter. For each experiment and all conditions, triplicate wells were counted.

In Vitro JNK Activity Assay The kinase activity assay was carried out using a KinaseSTAR JNK activity assay kit (Biovision). Briefly, subconfluent monolayer cells (in 100-mm plates) were washed with PBS and resuspended in cold JNK lysis buffer supplied in the kit for 10 minutes. Cell debris was removed by centrifugation at 14,000 rpm at 4°C for 15 minutes. Two microliters of JNK antibody was added to 200 ␮L cell lysate and incubated for 1 hour at room temperature. Resuspended Protein A sepharose was added and incubated for another 1 hour. After wash, 50 ␮L of kinase assay buffer and 2 ␮L c-Jun Protein/ATP mixture were added to each immunoprecipitation sample and incubated for 4 hours at 30°C. Protein A bead was spun down and supernatant was collected. Finally, we performed Western blotting using the rabbit anti-phospho-c-Jun antibody at 1:1000 dilutions. A 35-kilodalton band corresponding to the phosphorylated c-Jun protein was detected.

Clonogenic Cell Survival Assay Gastric cancer cells and JB6 P⫹ cells were exposed to different concentrations of SC-236 with or without PMA in 1.5 mL of 0.33% agar medium over 3 mL of 0.5% agar medium as described previously.21,33 Colonies were scored at 14 days, fixed with 70% ethanol, stained with Coomassie Blue, and counted under a dissection microscope. Only those colonies containing at least 50 cells were considered to be viable survivors.

138

WONG ET AL.

Prostaglandins E2 Production Cells were plated at 5 ⫻ 104/well into 24-well plates. Twenty-four hours later, the cultures were exposed to 100 nmol/L PMA with or without SC-236 for another 24 hours. PGE2 secreted into the medium was determined by competitive ELISA using a kit from Cayman Chemicals (Ann Arbor, MI) according to the manufacturer’s instructions. Briefly, medium from the cultured cells were added to 96-well plates coated with an anti-mouse antibody, mixed with a PG/acetylcholinesterase tracer and a monoclonal antibody against prostaglandins, and incubated at 4°C overnight. Unbound PG/ acetylcholinesterase was removed and washed extensively. Bound acetylcholinesterase was detected by Ellman’s reagent and measured at 410 nm.

COX-2 Activity Assay Cells were plated at 5 ⫻ 104/well into 24-well plates. Twenty-four hours later, the cultures were exposed to 100 nmol/L PMA with or without SC-236 for another 24 hours. Cells were collected and homogenized in cold buffer (0.1 mol/L Tris-HCl, pH 7.8, 1 mmol/L EDTA, 250 mmol/L mannitol, and 0.3 mmol/L diethyldithiocarbamic acid). COX activity was assayed colorimetrically by monitoring the appearance of oxidized N,N,N⬘,N⬘-tetramethyl-␳-phenylenediamine (TMPD) at 590 nmol/L using a kit from Cayman Chemicals according to the manufacturer’s instructions. The reaction was initiated by adding 20 ␮L arachidonic acid solution and incubated for 5 minutes at 25°C. COX-1 activity was inhibited by SC560.

Transfection and Luciferase Assay of AP-1 Activity AP-1 reporter plasmid was constructed by inserting collagenase promoter region (⫺73 to ⫹67 containing 1 AP-1 binding site) into luciferase reporter vector pGL-3-basic (Promega Corp., Madison, WI).36 Signal transduction pathway transreporting system containing pFR-Luc plasmid, pFA2cJun plasmid, pFC2-dbd plasmid, and pFC-MEKK plasmid were purchased from Stratagene. For transient transfection experiments, cells were seeded in 12-well plates to 70%– 80% confluence. The cells were transfected with 0.8 ␮g/well AP-1 reporter plasmid using lipofectamine 2000. PRL-CMV vector (0.01 ␮g/well, Promega) was cotransfected as internal control. After transfection for 4 hours, cells were changed to normal medium and allowed to recover overnight. Cells were first treated with different concentrations of SC-236 for 2 hours and then incubated in media in the absence or presence of 100 nmol/L PMA for an additional 24 hours. For the inhibitor experiments, the cells were pretreated with inhibitors 1 hour before SC-236 treatment. Transfected cells were collected and lysed, and the firefly and renilla luciferase activities were measured using the Dual-Luciferase Reporter assay system (Promega, Madison, WI) with a model TD-20/20 Luminometer.

GASTROENTEROLOGY Vol. 126, No. 1

Preparation of Cytoplasmic and Nuclear Extract Nuclear and cytoplasmic extracts were prepared as described by Dignam et al.37 Confluent cells in 10-cm dishes were treated for various times with the indicated effectors. Cells were resuspended in 400 ␮L buffer A (containing 10 mmol/L HEPES, pH 7.9, 1.5 mmol/L MgCl2 , 10 mmol/L KCl, 0.5 mmol/L DTT, 0.5 mmol/L PMSF, 1 ␮g/mL leupeptin, 1 ␮g/mL aprotinin, and 1 ␮g/mL pepstatin A), kept on ice for 15 minutes, lysed gently with 12.5 ␮L of 10% Nonide P-40, and centrifuged at 2000g for 10 minutes at 4°C. The supernatant was collected and used as the cytoplasmic extracts. The nuclei pellet was resuspended in 40 ␮L buffer C (20 mmol/L HEPES, pH 7.9, containing 1.5 mmol/L MgCl2 , 450 mmol/L NaCl, 25% glycerol, 0.2 mmol/L EDTA, 0.5 mmol/L DTT, 0.5 mmol/L PMSF, 1 ␮g/mL leupeptin, 1 ␮g/mL aprotinin, 1 ␮g/mL pepstatin A) and agitated for 30 minutes at 4°C, and the nuclear debris was spun down at 20,000g for 15 minutes. The supernatant (nuclear extract) was collected and stored at ⫺80°C until ready for analysis.

Western Blotting Twenty micrograms of protein were resolved and separated by electrophoresis on a 10% denaturing SDS gel. Proteins were electroblotted onto nitrocellulose membranes. Detection was conducted by immunostaining using specific primary antibodies and horseradish peroxidase-conjugated anti-IgG antibody. The protein bands were visualized by the enhanced chemiluminescence (ECL) assay (Amersham Pharmacia Biotech) following manufacturer’s instructions. Phospho-MAPK family antibody sampler kit and the PhosphoPlus c-Jun Kit were purchased from New England Biolabs (Boston, MA). c-fos Antibody and horseradish peroxidase-conjugated anti-rabbit IgG were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Electrophoretic Mobility Shift Assay Eight micrograms of nuclear proteins were incubated with 1 ␮g each of poly (dI.dC) in the presence of 30 fmol of Digoxin (DIG)-labeled, double-stranded AP-1 probe (5⬘-CGC TTG ATG ACT CAG CCG GAA-3⬘; Santa Cruz) for 15 minutes at room temperature in a total volume of 20 ␮L using DIG gel shift kit (Roche Diagnostics GmbH, Mannheim, Germany). Oligonucleotide competition experiments were performed in the presence of 50-fold excess of unlabeled AP-1 oligonucleotides. DNA complexes were resolved from free probe with 4% nondenaturing polyacrylamide gels in 0.5X Tris-borate-EDTA (pH 8.3) and visualized by fluorography.

Statistical Analysis The data shown were mean values of at least 3 different experiments and expressed as means ⫾ SD. The Student t test was used for comparison. A P value ⬍0.05 is considered statistically significant.

January 2004

COX-2 INHIBITOR SUPPRESSES AP-1

139

Results Specific COX-2 Inhibitor Suppressed Anchorage-Independent Growth of Gastric Cancer Cells To clarify whether specific COX-2 inhibitor SC236 suppresses the transformed phenotype of human gastric epithelial cells, the colony-forming abilities of these cells in soft agar were examined. As shown in Figure 1A, SC-236 inhibited anchorage-independent cell growth in a concentration-dependent manner in all of the 3 gastric cancer cell lines tested. However, the inhibitory effect was most profound in AGS cells with IC50 at about 20 –25 ␮mol/L. The mouse epidermal JB6 cell system is a well-developed model for studying tumor promotion. We, therefore, used this model to further test the antitumor effect of SC-236. As shown in Figure 1B, SC-236 inhibited PMA-induced transformation of JB6 cells in a concentration-dependent manner. The effective inhibitory concentration ranged from 12.5 to 50 ␮mol/L in which no cytotoxic effect on JB6 cells were observed by MTT assay (Figure 1C). We also compared the effect of aspirin on cell transformation in JB6 cells. As anticipated, aspirin inhibited cell transformation in a dosedependent manner (Figure 1B), with IC50 approximately equal to 1 mmol/L. This IC50 value is approximately 40-fold higher than that of SC-236, indicating that SC-236 is a more potent inhibitor of cell transformation in JB6 cells. SC-236 Inhibited PMA-Induced AP-1 Activation in Gastric Cancer Cells Activation of AP-1 by growth factor or tumor promoters such as tumor necrosis factor (TNF)-␣ or PMA has been implicated in cell transformation and cell growth.16 –18 NSAIDs, such as aspirin and sodium salicylate, have been shown to inhibit cell transformation through suppression of PMA or UV-stimulated AP-1 activation.19,33 These findings prompted us to determine whether SC-236 is an inhibitor of AP-1 activation. Three human gastric cell lines were transiently transfected with AP-1 luciferase reporter gene and treated with different concentrations of SC-236 for 2 hours then incubated in the presence of 100 nmol/L PMA for an additional 24 hours. As shown in Figure 2A, PMA caused a remarkable increase in AP-1 transcriptional activity in 3 gastric cell lines, whereas SC-236 inhibited PMA-stimulated AP-1 transcriptional activity in a dose-dependent manner. The effect of SC-236 was most profound in AGS cells with IC50 approximately equaled to 25 ␮mol/L, which was consistent with the cell growth data. In addition, we

Figure 1. Effects of SC-236 on anchorage-independent cell growth and PMA-induced cell transformation. (A) 104 Gastric cancer cells were exposed to different concentrations of SC-236 in 0.33% agar for 14 days and scored for colonies at the end of the experiment. (B) 104 JB6 cells were exposed to different concentrations of SC-236 or aspirin in the presence or absence of 100 nmol/L PMA in 0.33% agar for 14 days and scored for colonies at the end of the experiment. (C) 5 ⫻ 103 JB6 cells were plated on 96-well plates and treated with different concentrations of SC-236 for 24 hours. MTT assay was performed. All results were expressed as the mean of 3 independent experiments ⫾ standard error.

140

WONG ET AL.

GASTROENTEROLOGY Vol. 126, No. 1

CMV or SV40 promoter (data not shown), indicating that the inhibition of AP-1 activation was specific. To further elucidate the mechanism of SC-236 –mediated inhibition of AP-1 activation, AP-1 DNA-binding activity was measured by electrophoretic mobilityshift assay after pretreatment of AGS cells with SC-236. Our findings demonstrated that SC-236 (50 ␮mol/L) slightly suppressed AP-1 DNA binding (Figure 2B, lane 3). Treatment with PMA (100 nmol/L) resulted in a significant increase in AP-1 DNA binding (Figure 2B, lane 2). This increased binding could be eliminated by adding a 50-fold excess of unlabeled AP-1 oligonucleotide, confirming that the electrophoretic mobility-shift band was specific for AP-1 binding (Figure 2B, lane 5). SC-236 treatment markedly inhibited AP-1 DNA binding induced by PMA (Figure 2B, lane 4). Thus, SC-236 inhibited both AP-1 DNA-binding activity and transcriptional activity in gastric cancer cells. SC-236 Inhibited PMA-Induced Phosphorylation of JNK

Figure 2. SC-236 suppressed AP-1 transcriptional and binding activity. (A) Cells transiently expressing AP-1 luciferase reporter gene construct were treated with different concentrations of SC-236 for 2 hours then followed with 100 nmol/L PMA for another 24 hours. Cells were then harvested for analysis of luciferase activity. The firefly luciferase reading was normalized to renilla luciferase reading. Results were expressed as the means of 3 independent experiments ⫾ standard error. (B) AGS cells were treated with 50 ␮mol/L SC-236 alone for 2 hours or pretreated with SC-236 for 1 hour, followed by 100 nmol/L PMA for another 1 hour. Nuclear extracts were prepared and analyzed in an electrophoretic mobility shift assay with a DIG-labeled AP-1 probe. Equal amounts (6 ␮g) of nuclear protein were loaded in each lane. In lane 5, a 50-fold excess of unlabeled oligonucleotide was added before the addition of DIGlabeled probe.

also showed that SC-236 significantly inhibited TNF-␣induced activation of AP-1 transcription in gastric cancer cells (data not shown), which indicated that the inhibitory effect on AP-1 activation by SC-236 was not exclusive to PMA stimulation. Also, SC-236 did not significantly affect the transcriptional activity of the control

Because the maximal inhibitory effect on cell growth and AP-1 activity was found in AGS cells, we chose to use AGS cells in the following study. Extracellular signals, including growth factors, phorbol ester, and UV irradiation stimulate phosphorylation of c-Jun at Ser-63/-73 by JNK and activate AP-1-dependent transcription. To clarify whether the JNK-c-Jun pathway is involved in the inhibition of AP-1 activation by SC-236, we determined both phosphorylation of JNK and c-Jun by Western blot assay in AGS cells. We showed that SC-236 inhibited PMA-induced phosphorylation of cJun and JNK, whereas Western blotting with anti-JNK and anti-c-Jun antibodies showed that the recovery of JNK or c-Jun from cell lysates was not altered by treatment with SC-236 (Figure 3A and 3B). As shown in Figure 3B, PMA led to marked induction of JNK2 and JNK1 phosphorylation by 4.2- and 2.8-fold, respectively. The difference in favor of JNK2 may be due to a higher basal level of activity exhibited by JNK1 compared with JNK2. However, treatment with 50 ␮mol/L SC-236 led to comparable suppression of JNK2 and JNK1 phosphorylation by 60% and 55%. To substantiate the result from Western blot analysis, we further performed JNK activity assay. We showed that SC-236 treatment significantly inhibited PMA-stimulated JNK activity, which was consistent with our Western blot results (Figure 3C). To further determine whether MAPKs other than JNK have been involved in the inhibitory effect of SC-236, we tested phosphorylated protein expression of ERK and p38 after SC-236 treat-

January 2004

COX-2 INHIBITOR SUPPRESSES AP-1

141

Figure 3. SC-236 inhibited PMA-induced JNK phosphorylation and activity. AGS cells were pretreated with different concentrations of SC-236 for 2 hours then exposed to PMA for another 1 hour. Total and phosphorylated proteins were detected with Western blot assay. (A) SC-236 inhibited c-Jun phosphorylation. (B) SC-236 inhibited JNK phosphorylation. Quantification of phosphorylation of c-Jun, ERK, and JNK was performed by scanning densitometry of the bands, and the values shown are the means ⫾ SE (n ⫽ 3) and are expressed as percentages of the maximal increase above the control values. (C) The effect SC-236 and aspirin on JNK kinase activity. JNK activity was determined by an in vitro kinase assay. The JNK activity kit utilized a JNK-specific antibody to immunoprecipitate JNK from cell lysates. Activity of the JNK was then determined in a kinase reaction using recombinant c-Jun as substrate. Phosphorylation of the c-Jun was analyzed by Western blot analysis using a phospho-c-Jun specific antibody. The right panel showed the values (mean ⫾ SE, n ⫽ 3) of the level of JNK activation by in vitro kinase assay obtained from scanning densitometry expressed as a percentage of the maximum increase. *P ⬍ 0.05.

ment. We showed that only 50 ␮mol/L SC-236 significantly inhibited PMA-induced phosphorylation of ERK1/ERK2, whereas phosphorylation of p38 remained unchanged with SC-236 treatment in the range of concentration tested (Figure 3B). The findings presented in Figure 3 indicated that the MAPK pathway, particularly the JNK/c-Jun pathway, might play a critical role in the effect of SC-236-mediated suppression of AP-1 activation. As a control, we also tested the effect of aspirin on JNK activity. Our result showed that aspirin did not have any effect on PMA-induced JNK activation as shown in Figure 3C, indicating that aspirin might inhibit AP-1 activation through a different pathway.

SC-236 Inhibited AP-1 Activation Through JNK-c-Jun Pathway To further elucidate the role of JNK/c-Jun pathway in SC-236 mediated AP-1 suppression, we took advantage of Stratagene’s PathDetect in vivo signal transduction transreporting system (Stratagene). This system includes a unique fusion transactivator plasmid, which consists of c-Jun transcriptional activator fused with the yeast GAL4 DNA-binding domain, and a pFRLuc reporter plasmid, which contains a synthetic promoter with 5 tandem repeats of the yeast GAL4-binding sites that control expression of the Photinus pyralis lu-

142

WONG ET AL.

ciferase gene. Phosphorylation of the transcription activation domain of the fusion c-Jun protein by JNK will activate transcription of the luciferase gene from the reporter plasmid, and their activity reflects the in vivo activation of JNK and the corresponding specific signal

GASTROENTEROLOGY Vol. 126, No. 1

transduction pathways. In this study, after transfection with pFR-Luc and pFA2-cJun into AGS cells, we treated the cells with PMA for 12 hours. As shown in Figure 4A, PMA simulated the pFR-Luc reporter activity to about 1.8-fold compared with the control, whereas pretreatment with 25 ␮mol/L SC-236 significantly inhibited this activation. Furthermore, we cotransfected pFCMEKK plasmid, which was the positive control for pFRLuc activation into AGS cells, and treated with SC-236. Our results showed that SC-236 also significantly inhibited MEKK-stimulated reporter activation, indicating that SC-236 specifically inhibited MEKK-JNK-c-Jun pathway. To substantiate the result that JNK mediates the inhibitory effect of SC-236, we utilized a highly selective JNK inhibitor, SP600125, to observe its effect on SC236-mediated inhibition of AP-1 activation. SP600125 inhibits JNK with more than a 20-fold selectivity vs. other kinases such as extracellular signal-regulated kinases, p38 kinase, or protein kinase C, and others.38 Our results showed that SP600125 completely reversed the inhibitory effect on AP-1 activity by SC-236 (Figure 4B), indicating that JNK is required for the inhibitory effect of SC-236 on AP-1 activation. We also suppressed the activation of ERK by its specific inhibitor U0126. However, pretreatment of AGS cells with U0126 had no significant effect on the inhibitory effect on AP-1 activity by SC-236 (Figure 4C). JNK Inhibition Resulted in Cell Growth Suppression in AGS Cells Although PMA has been shown to stimulate gastric cell growth,39,40 there is no data concerning the role

Š Figure 4. SC-236 inhibited JNK-c-Jun pathway. (A) AGS cells were transiently cotransfected with pFR-Luc and pFA2-c-Jun plasmid in the presence or absence of pFC-MEKK and treated with 25 ␮mol/L SC236 for 2 hours followed with 100 nmol/L PMA for another 12 hours. Cells were then harvested for analysis of luciferase activity. Results were expressed as the means of 3 independent experiments ⫾ standard error. (B). Top panel: Western blot analysis of phospho-JNK. Cells were pretreated with or without 2.5 ␮mol/L SP600125 for 1 hour, followed with PMA for another 1 hour. Bottom panel: AP-1 transcriptional activity assay. Cells were pretreated with or without 2.5 ␮mol/L SP600125 for 1 hour and then treated with SC-236 for 2 hours, followed with PMA for another 24 hours. At the end of the experiments, cells were harvested for analysis of luciferase activity. (C) Top panel: Western blot analysis of phospho-ERK. Cells were pretreated with or without 10 ␮mol/L U0126 for 1 hour, followed with PMA for another 1 hour. Bottom panel: AP-1 transcriptional activity assay. Cells were pretreated with or without 10 ␮mol/L U0126 for 1 hour and then treated with SC-236 for 2 hours, followed with PMA for another 24 hours. At the end of the experiments, cells were harvested for analysis of luciferase activity. *P ⬍ 0.01, #P ⬎ 0.05, compared with PMA treatment without SC-236, separately.

January 2004

Figure 5. Suppression of JNK activity inhibited gastric cancer cell growth. (A) Total protein was extracted from cells 48 hours following treatment with 0.4 ␮mol/L JNK1AS, 0.4 ␮mol/L JNK2AS, 0.2 ␮mol/L JNK1 AS ⫹ 0.2 ␮mol/L JNK2AS, or control oligonucleotides (JNKScr). Protein samples were analyzed by Western blot analysis using specific JNK antibody. (B) Total Jun kinase activity was determined 1 hour following exposure to PMA by an in vitro kinase assay. Forty-eight hours before PMA treatment, cells were transfected with JNK1 AS, JNK2AS, combinations of JNK1AS plus JNK2AS (0.4 ␮mol/L each), or JNKScr. (C) JNKAS inhibited cell growth in AGS cells. Proliferation assays were performed as described in the Materials and Methods section. The cells were maintained in 2% FBS during the experiment. Five days after the treatment, the cells were counted with a Coulter counter. The proliferation data shown here were the average of 2 identical and independent experiments, each carried out in triplicate. *P ⬍ 0.05.

of JNK in the regulation of gastric cancer cell growth. To further establish a clear link between gastric cancer growth and JNK activity, we performed cell growth analysis with antisense oligonucleotides of both JNK1 and JNK2. First, we investigated the effect of JNK1AS and JNK2AS on JNK expression in AGS cells. Cells were treated either with 0.4 ␮mol/L JNK1AS or JNK2AS individually or with JNK1AS and JNK2AS in combination (0.2 ␮mol/L each). Scrambled-sequence oligonucleotides (JNK1Scr and JNK2Scr) were used at the same concentrations and served as controls. Both AS treatments led to ⬎60% reduction in the corresponding protein levels (Figure 5A), whereas neither mock lipo-

COX-2 INHIBITOR SUPPRESSES AP-1

143

fectamine nor treatment with scrambled oligonucleotides had any effect on JNK1 or JNK2 protein expression. To substantiate the results obtained with antisense oligonucleotides, an immunocomplex kinase assay was used to examine basal and PMA-induced JNK activity in mock-, JNKAS-, and JNKScr-treated cells (Figure 5B). Consistent with the reduction in JNK protein levels, PMAinduced JNK activation was markedly reduced in AGS cells treated with JNK1AS and JNK2AS. Neither mock lipofectamine nor treatment with scrambled oligonucleotides affected the JNK kinase activity. Taken together, the experiments described above demonstrated that, using JNKAS, we effectively achieved a significant reduction in JNK expression and therefore directly investigated the roles of JNK1 and JNK2 in regulating the growth of AGS cells. All cells were counted 5 days after transfection with antisense oligonucleotides or scrambled oligonucleotides or Lipofectamine. We showed in this study that inhibition of JNK activity by antisense strategy markedly suppressed gastric cancer cell growth. As shown in Figure 5C, AGS cells displayed a significant reduction in cell viability following treatment with JNKAS; the cell growth was reduced to 70% with JNKAS1 and to 72.4% with JNKAS2. Furthermore, the cell growth suppression by both JNK1AS and JNK2AS was most pronounced, where viability was reduced to 51.7% by 5 days following treatment with the oligonucleotide. Altogether, results from our study indicate that JNK has a predominant role in gastric cancer cell growth. The Inhibitory Effects of SC-236 on AP-1 and JNK Are Independent of COX-2Prostaglandin Inhibition To further explore whether the inhibitory effects on AP-1 and JNK activation by SC-236 occur through COX-prostaglandins inhibition and whether these effects are general to all COX-2 inhibitors, we used other COX-2 inhibitors and sunlindac sulfone in this study as well. First, we determined COX-2 enzyme activity and PGE2 production after SC-236 treatment. As shown in Figure 6A and 6B, 1 ␮mol/L SC-236 already suppressed PMA-induced COX-2 activity and PGE2 production to about 20% and 50%, respectively, of the control. With higher concentrations of SC-236, PGE2 production remained under the detection level (⬍5 pg/105 cells). Next, we investigated the effects of other NSAIDs including 2 COX-2 inhibitors, NS-398 and nimesulide, and sunlindac sulfone on AP-1 and JNK activation in AGS cells. Our results showed that both nimesulide and sunlindac sulfone significantly inhibited PMA-induced AP-1 activation, whereas NS-398 did not (Figure 7A).

144

WONG ET AL.

Figure 6. The effect of SC-236 on COX-2 activity and PGE2 production. (A) Cells were plated on 24-well plates and exposed to PMA with or without SC-236 for 24 hours. COX-2 activity was determined as described in the Materials and Methods section. Results are mean ⫾ SD of triplicate assays and 2 experiments. (B) Cells were plated on 24-well plates and exposed to PMA with or without SC-236 for 24 hours. PGE2 content of the conditioned media was determined by ELISA. Results are mean ⫾ SD of triplicate assays and 2 experiments. *P ⬍ 0.05.

Furthermore, neither NS-398 nor sulindac sulfone inhibited JNK or c-Jun phosphorylation stimulated by PMA as shown in Figure 7B. To clarify this issue further, we determined the effect of exogenous prostaglandins in the induction of anchorage independent cell growth and AP-1 activity in AGS cells. As shown in Figure 7C and 7D, exogenous prostaglandins had no effect on anchorage-independent cell growth and PMA-stimulated AP-1 activation in the presence of SC-236.

Discussion SC-236 inhibited tumor growth through different mechanisms such as induction of apoptosis,12 suppression of cell proliferation,41 and inhibition of angio-

GASTROENTEROLOGY Vol. 126, No. 1

genesis.5,13 Our group previously demonstrated that high doses of SC-236 (12.5–100 ␮mol/L) suppressed gastric cancer cell growth. However, the mechanism(s) remains unknown. AP-1 is considered a mediator of carcinogenesis by its ability to alter gene expression in response to tumor promoters, such as epidermal growth factor, PMA, or UV irradiation.18 Blocking of tumor promoterinduced AP-1 activity by some NSAIDs, such as aspirin and sodium salicylate, partly explains their antitumor effects.30 –33 Thus, we hypothesized that SC-236 might exert its antitumor effect through modulation of AP-1 signaling pathway in gastric cancer. In the present study, we demonstrated that SC-236 inhibited anchorage-independent cell growth in 3 gastric cancer cell lines with the concentration ranging from 12.5 ␮mol/L to 50 ␮mol/L. To substantiate the antitumor effect of SC-236, we used the well-established cell transformation model and showed that SC-236 suppressed PMA-induced JB6 cell transformation with IC50 approximately equal to 25 ␮mol/L. Based on the findings that the same range of SC-236 (12.5–50 ␮mol/L) suppressed PMA-induced AP-1 DNA binding and transactivation in gastric cancer cells, we suggest that the antitumor effects of SC-236 are due, at least in part, to the suppression of AP-1 activity. Of note, most studies demonstrating effects on COX-independent pathways utilized higher concentrations of NSAIDs (100 –1000 ␮mol/L).42 In our study, SC-236 inhibited cell transformation and AP-1 activity dose dependently within the achievable physiologic range of between 12.5 and 50 ␮mol/L. AP-1, consisting of Jun/Fos dimers, is a downstream target of MAP kinase family members including extracellular signal regulated kinases (ERK-1 and -2; p42/p44 MAPK), Jun kinases (JNK), and p38 MAPK. Aspirin and sodium salicylate suppressed AP-1 activation through different mechanisms in both mouse JB6 cells and normal human cells.20,30 –33 Aspirin inhibited UVBinduced AP-1 activity through blocking UVB-induced activation of ERK and JNK as well as p38 kinase.33 Interestingly, it blocked TPA- or epidermal growth factor-induced signaling through a MAP kinase-independent pathway.20 These findings agreed with Schwenger et al., who reported that the same dose of sodium salicylate inhibited ERK/JNK activation induced by TNF but not by other growth factors such as EGF, PDGF, and IL-1 under the same condition in normal human F4S fibroblasts.31,32 These discrepant results suggest that the regulatory mechanisms of NSAIDs on MAPK family members may depend on the cellular context and different external stimuli.

January 2004

In this study, we showed that inhibition of JNK phosphorylation and activation by SC-236 contributed to its effect on AP-1 activation and cell growth. Our conclusion is based on the following points. First, we demonstrated that SC-236 inhibited PMA-induced JNK/cJun phosphorylation as well as JNK activity by both Western blotting analysis and in vitro kinase activity assay. Second, using the specific signal transreporting

COX-2 INHIBITOR SUPPRESSES AP-1

145

system, we found that pretreatment with SC-236 inhibited both PMA- and MEKK-stimulated reporter activation, which indicated that the specific MEKK-JNK-cJun pathway was inhibited by SC-236. These results suggested that SC-236 interacted with JNK, directly or indirectly, and subsequently inhibited the phosphorylation and activation of JNK by PMA and MEKK. The involvement of JNK signaling was further confirmed by showing that pretreatment with JNK inhibitor completely reversed the inhibitory effect of SC-236 on AP-1 activation. In this context, we hypothesized that JNK inhibitor blocked interaction of SC-236 with JNK, thereby preventing SC-236 to inhibit JNK and releasing the inhibitory effect of SC-236 on AP-1 activation. Altogether, these results suggested that JNK signaling was required for the inhibitory effect on AP-1 activation by SC-236. This is the first report showing that NSAIDs inhibited tumor promoter-induced AP-1 activation through suppression of MAPK pathway in human cancer. On the other hand, although 50 ␮mol/L SC-236 showed some degree of inhibition of ERK activation, this concentration is about 2 times the amount of IC50 for inhibition of cell growth and AP-1 activation in AGS cells. Blockage of ERK activity did not have any effect on the SC-236-mediated inhibition of AP-1 activation. Thus, we concluded that SC-236 might have some effect on ERK/MAPK phosphorylation, but the ERK/MAPK pathway did not have a predominant role in the inhibitory effect of SC-236 on AP-1 activation. The role of activated JNK in cell survival in response to extracellular stimuli has been well reported.34,35,43 However, evidence supportive of a proapoptotic function for the JNK pathway has also been documented.29,44 One plausible explanation for these seemingly disparate effects is that JNK serves different functions in diverse cell types and under different stimuli. Using high-affinity and high-specificity phosphorothioate antisense oligonuŠ Figure 7. Exogenous prostaglandins PGE2 , PGH1 , and PGH2 had no effect on cell growth and AP-1 activation. (A) AGS cells were plated on 24-well plates and exposed to 100 ␮mol/L NS-398 or 100 ␮mol/L Nimesulide or 100 ␮mol/L sunlindac sulfone for 2 hours, followed with PMA for another 12 hours. The AP-1 luciferase enzyme activity was measured using the luminometer. (B) AGS cells were pretreated with 100 ␮mol/L sulindac sulfone or 25 ␮mol/L SC-236 or 100 ␮mol/L NS-398 followed with PMA for another 1 hour. Total and phosphorylated proteins were detected with Western blot assay. (C) 104 AGS cells were exposed to 2 ␮g/mL of different prostaglandins in 0.33% agar for 14 days and scored for colonies at the end of the experiment. (D) AGS cells were plated on 24-well plates and exposed to different concentrations of prostaglandins with SC-236 for 2 hours, followed with 100 nmol/L PMA for another 12 hours. The AP-1 luciferase enzyme activity was measured using the luminometer. *P ⬎ 0.05.

146

WONG ET AL.

cleotides targeting JNK1 and JNK2,34,35 we demonstrated that inhibition of JNK1 and JNK2 in AGS cells resulted in marked suppression of cell growth, indicating that JNK was necessary for gastric cancer cell growth. Others have shown that JNK was required for epidermal growth factor-stimulated growth of A549 cells in soft agar,34 and JNK antisense treatment inhibited cell growth and induced apoptosis in glioblastoma.35 To clarify whether the inhibitory effects on AP-1 and JNK activation by SC-236 was related to COX-2 enzyme activity, we examined the effects of other NSAIDs. We demonstrated that sunlindac sulfone, an NSAID lacking antiprostaglandins synthetase activity, inhibited AP-1 activation significantly, whereas NS-398, a selective COX-2 inhibitor, failed to inhibit AP-1 activity. Furthermore, we showed that both sunlindac sulfone and NS-398 did not have any effect on JNK and c-Jun phosphorylation. These results indicated that the inhibitory effect on AP-1 and JNK activation was not necessarily related to the inhibition of COX-2 enzyme activity and was not universal for all COX-2 inhibitors. Several studies have suggested that PGE2 promoted tumorigenesis through the activation of AP-1.45– 47 Consequently, we considered the possibility that inhibition of COX-2prostaglandins synthesis executed autoregulatory effect on AP-1 activity and cell growth. Therefore, we tested the effect of exogenous prostaglandins (PGE2 , PGH1 , PGH2) and showed that they had no effect on cell growth and PMA-induced AP-1 transcription in the presence of SC-236. Therefore, the inhibition of prostaglandin synthesis was not the mechanism responsible for the inhibition of AP-1 transactivation and neoplastic transformation in gastric cancer. In conclusion, we showed that suppression of AP-1 activation contributed to the antitumor effect of a specific COX-2 inhibitor in gastric cancer. Moreover, the inhibitory effect on JNK/c-Jun activation was essential for the suppression of AP-1 activation. Our study provides a novel molecular mechanism by which a specific COX-2 inhibitor exerts its antitumor effect. Our findings highlight the importance of JNK for carcinogenesis and suggest that inhibition of JNK activation may have a therapeutic benefit against gastric cancer.

References 1. Gupta RA, DuBois RN. Aspirin, NSAIDs, and colon cancer prevention: mechanisms? Gastroenterology 1998;114:1095–1098. 2. Kawamori T, Rao CV, Seibert K, Reddy BS. Chemopreventive activity of celecoxib, a specific cyclooxygenase-2 inhibitor, against colon carcinogenesis. Cancer Res 1998;58:409 – 412. 3. Reddy BS, Rao CV, Seibert K. Evaluation of cyclooxygenase-2 inhibitor for potential chemopreventive properties in colon carcinogenesis. Cancer Res 1996;56:4566 – 4569. 4. Sawaoka H, Kawano S, Tsuji S, Tsuji M, Gunawan ES, Takei Y,

GASTROENTEROLOGY Vol. 126, No. 1

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

20.

Nagano Kand Hori M. Cyclooxygenase-2 inhibitors suppress the growth of gastric cancer xenografts via induction of apoptosis in nude mice. Am J Physiol 1998;274:G1061–1065. Piazza GA, Rahm AK, Finn TS, Fryer BH, Li H, Stoumen AL, Pamukcu R, Ahnen FJ. Apoptosis primarily accounts for the growth-inhibitory properties of sulindac metabolites and involves a mechanism that is independent of cyclooxygenase inhibition, cell cycle arrest, and p53 induction. Cancer Res 1997;57:2452– 2459. Hanif R, Pittas A, Feng Y, Koutsos MI, Qiao L, Staiano-Caico L, Shiff SI, Rigas B. Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer by a prostaglandins-independent pathway. Biochem Pharmacol 1996;52:237–245. Elder DJ, Halton DE, Hague A, Paraskeva C. Induction of apoptotic cell death in human colorectal carcinoma cell lines by a cyclooxygenase-2 (COX-2)-selective nonsteroidal anti-inflammatory drug: independence from COX-2 protein expression. Clin Cancer Res 1997;3:1679 –1683. Shiff SJ, Qiao L, Tsai LL, Rigas B. Sulindac sulfide, an aspirin-like compound, inhibits proliferation, causes cell cycle quiescence, and induces apoptosis in HT-29 colon adenocarcinoma cell. J Clin Invest 1995;96:491–503. Goldberg Y, Nassif II, Pittas A, Tsai LL, Dynlacht BD, Rigas B, Shiff SJ. The anti-proliferative effect of sulindac and sunlindac sulfide on HY-29 colon cancer cells: alterations in tumor suppressor and cell cycle-regulatory proteins. Oncogene 1996;12:893–901. Barnes CJ, Cameron IL, Hardman WE, Lee M. Non-steroidal antiinflammatory drug effect on crypt cell proliferation and apoptosis during initiation of rat colon carcinogenesis. Br J Cancer 1998; 77:573–580. Zhu GH, Wong BC, Ching CK, Lai KC, Lam SK. Differential apoptosis by indomethacin in gastric epithelial cells through the constitutive expression of wild-type p53 and/or up-regulation of c-myc. Biochem Pharmacol 1999;58:193–200. Jiang XH, Lam SK, Lin MC, Jiang SH, Kung HF, Slosberg ED, Soh JW, Weinstein IB, Wong BC. Novel target for induction of apoptosis by cyclo-oxygenase-2 inhibitor SC-236 through a protein kinase C-␤ (1)-dependent pathway. Oncogene 2002;21:6113– 6122. Tsuji M, Kawano S, Tsuji S, Sawaoka H, Hori M, DuBois RN. Cyclooxygenase regulates angiogenesis induced by colon cancer cells. Cell 1998;93:705–716. Jones MK, Wang H, Peskar BM, Levin E, Itani RM, Sarfeh IJ, Tarnawske AS. Inhibition of angiogenesis by nonsteroidal antiinflammatory drugs: insight into mechanisms and implication for cancer growth and ulcer healing. Nat Med 1999;5:1418 –1423. Steinbach G, Lynch PM, Phillips RK, Wallace MH, Hawk E, Gordon GB. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000;342:1946 – 1952. Angel P, Karin M. The role of Jun, Fos, and the AP-1 complex in cell-proliferation and transformation. Biochim Biophys Acta 1991;1072:129 –157. Curran T, Franza BR Jr. Fos and Jun: the AP-1 connection. Cell 1988;55:395–397. Angel P, Imagawa M, Chiu R, Stein B, Imbra RJ, Rahmsdorf HJ, Jonat C, Herrlich P, Karin M. Phorbol ester-inducible genes contain a common cis element recognized by a PMA-modulated trans-acting factor. Cell 1987;49:729 –739. Dong Z, Birrer MJ, Watts RG, Matrisian LM, Colburn NH. Blocking of tumor promoter-induced AP-1 activity inhibits induced transformation in JB6 mouse epidermal cells. Proc Natl Acad Sci U S A 1994;91:609 – 613. Dong Z, Huang C, Brown RE, Ma WY. Inhibition of activator protein 1 activity and neoplastic transformation by aspirin. J Biol Chem 1997;272:9962–9970.

January 2004

21. Li JJ, Westergaad D, Ghosh P, Colburn NH. Inhibitors of both nuclear factor-␬B and activator protein-1 activation block the neoplastic transformation response. Cancer Res 1997;57:3569– 3576. 22. Whitmarsh AJ, Davis RJ. Transcription factor AP-1 regulation by mitogen-activated protein kinase signal transduction pathways. J Mol Med 1996;74:589 – 607. 23. Karin M. The regulation of AP-1 activity by mitogen-activated protein kinases. J Biol Chem 1995;270:16483–16486. 24. Antonyak MA, Kenyon LC, Godwin AK, James DC, Emlet DR, Okamoto I, Tnani M, Holgado-Madruga M, Moscatello DK, Wong AJ. Elevated JNK activation contributes to the pathogenesis of human brain tumors. Oncogene 2002;21:5038 –5046. 25. Chen N, Nomura M, She QB, Ma WY, Bode AM, Wang L, Flavell RA, Dong Z. Suppression of skin tumorigenesis in c-Jun NH (2)-terminal kinase-2-deficient mice. Cancer Res 2001;61:3908– 3912. 26. Gupta S, Plattner R, Der CJ, Stanbridge EJ. Dissection of Rasdependent signaling pathways controlling aggressive tumor growth of human fibrosarcoma cells: evidence for a potential novel pathway. Mol Cell Biol 2000;20:9294 –9306. 27. Johnson R, Spiegelman B, Hanahan D, Wisdom R. Cellular transformation and malignancy induced by ras require c-jun. Mol Cell Biol 1996;16:4504 – 4511. 28. Pulverer BJ, Kyriakis JM, Avruch J, Nikolakaki E, Woodgett JR. Phosphorylation of c-jun mediated by MAP kinases. Nature 1991; 353:670 – 674. 29. Ip YT, Davis RJ. Signal transduction by the c-Jun N-terminal kinase (JNK)-from inflammation to development. Curr Opin Cell Biol 1998;10:205–219. 30. Murono S, Yoshizaki T, Sato H, Takeshita H, Furukawa M, Pagano JS. Aspirin inhibits tumor cell invasiveness induced by EpsteinBarr virus latent membrane protein 1 through suppression of matrix metalloproteinase-9 expression. Cancer Res 2000;60: 2555–2561. 31. Schwenger P, Skolnik EY, Vilcek J. Inhibition of tumor necrosis factor-induced p42/p44 mitogen-activated protein kinase activation by sodium salicylate. J Biol Chem 1996;271:8089 – 8094. 32. Schwenger P, Bellosta P, Vietor I, Basilico C, Skolnik EY, Vilcek J. Sodium salicylate induces apoptosis via p38 mitogen-activated protein kinase but inhibits tumor necrosis factor-induced c-Jun N-terminal kinase/stress-activated protein kinase activation. Proc Natl Acad Sci U S A 1997;94:2869 –2873. 33. Huang C, Ma WY, Dong Z. Ultraviolet B-induced activated protein-1 activation does not require epidermal growth factor receptor but is blocked by a dominant negative PKC ␭/␫. Mol Cell Biochem 1996;16:6427– 6435. 34. Bost F, McKay R, Bost M, Potapova O, Dean NM, Mercola D. The Jun kinase 2 isoform is preferentially required for epidermal growth factor-induced transformation of human A549 lung carcinoma cells. Mol Cell Biol 1999;19:1938 –1949. 35. Potapova O, Gorospe M, Bost F, Dean NM, Gaarde WA, Mercola D, Holbrook NJ. c-Jun N-terminal kinase is essential for growth of human T98G glioblastoma cells. J Biol Chem 2000;275:24767– 24775. 36. Angel P, Baumann I, Stein B, Delius H, Rahmsdorg HJ, Herrlich P.

COX-2 INHIBITOR SUPPRESSES AP-1

37.

38.

39.

40.

41.

42. 43.

44. 45.

46.

47.

147

12-O-tetradecanoyl-phorbol-13-acetate induction of the human collagenase gene is mediated by an inducible enhancer element located in the 5-flanking region. Mol Cell Biol 1987;7:2256 – 2266. Dignam JD, Lebovitz RM, Roeder RG. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res 1983;11:1475–1489. Bennett BL, Sasaki DT, Murray BW, O’Leary EC, Sakata ST, Xu W, Leisten JC, Motiwala A, Pierce S, Satoh Y, Bhagwat SS, Manning AM, Anderson DW. SP600125, an anthrapyrazolone inhibitor of Jun N-terminal kinase. Proc Natl Acad Sci U S A 2001;98:13681– 13686. Ochiai A, Yasui W, Kameda T, Takanashi A, Takekura N, Tahara E. The effect of phorbol esters on cell growth and epidermal growth factor receptor modulation in a human gastric carcinoma cell line TMK-1. Hiroshima J Med Sci 1989;38:191–195. Kanamori M, Tajima M, Satoh Y, Hoshikawa Y, Miyazawa Y, Okinaga K, Kurata T, Sairenji T. Differential effect of TPA on cell growth and Epstein-Barr virus reactivation in epithelial cell lines derived from gastric tissues and B cell line Raji. Virus Genes 2000;20:117–125. Richter M, Weiss M, Weinberger I, Fu¨ rstenberger G, Marian B. Growth inhibition and induction of apoptosis in colorectal tumor cells by cyclooxygenase inhibitors. Carcinogenesis 2001;22:17– 25. Turini ME, DuBois RN. Cyclooxygenase-2: a therapeutic target. Annu Rev Med 2002;53:35–57. Potapova O, Haghighi A, Bost F, Liu C, Birrer MJ, Gjerset R, Mercola D. The Jun kinase/stress-activated protein kinase pathway functions to regulate DNA repair and inhibition of the pathway sensitizes tumor cells to cisplatin. J Biol Chem 1997;272: 14041–14044. Karin M, Liu Z, Zandi E. AP-1 function and regulation. Curr Opin Cell Biol 1997;9:240 –246. Piazza GA, Kulchak Rahm AL, Krutzch M, Sperl G, Paranka NS, Gross PH, Brendel K, Gross PH, Sperl G, Ritchie J, Burt RW, Ellsworth L, Ahnen DJ, Pamukcu R. Sunlindac sulfone inhibits azoxymethane-induced colon carcinogenesis in rat without reducing prostaglandin levels. Cancer Res 1995;57:2909 –2915. Simonson MS, Herman WH, Dunn MJ. PGE2 induces c-fos expression by a cAMP-independent mechanism in glomerular mesangial cells. Exp Cell Res 1995;215:137–144. Won JS, Suh HW, Kim YH, Song DK, Huh SO, Lee JK, Lee KJ. Prostaglandins E2 increases proenkephalin mRNA level in rat astrocyte-enriched culture. Mol Brain Res 1998;60:203–214.

Received October 29, 2002. Accepted September 18, 2003. Address requests for reprints to: Benjamin C.-Y. Wong, M.D., Department of Medicine, University of Hong Kong, Queen Mary Hospital, Hong Kong. e-mail: [email protected]; fax: (852) 2872-5828. Supported by Research Grant Council earmarked grant HKU 7309/ 01M of the Hong Kong Special Administration Region, the Simon KY Lee Gastroenterology Research Fund, Queen Mary Hospital and Gastroenterological Research Fund, University of Hong Kong.