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

CP

BJH*S

BptdapDpamAnRdmdfdc2pvataCtsh

AcatadasN

GASTROENTEROLOGY 2004;126:136–147

yclooxygenase-2 Inhibitor (SC-236) Suppresses Activatorrotein-1 Through c-Jun NH2-Terminal Kinase

ENJAMIN CHUN–YU WONG,* XIAO HUA JIANG,*,‡ MARIE C. M. LIN,§ SHUI PING TU,*,‡

IAN TAO CUI,* SHI HU JIANG,‡ WAI MAN WONG,* MAN FUNG YUEN,* SHIU KUM LAM,* andSIANG FU KUNG§

Department of Medicine, Queen Mary Hospital, University of Hong Kong, Hong Kong; ‡Department of Gastroenterology, Rui-jin Hospital,hanghai, Peoples Republic of China; and §Institute of Molecular Biology, the University of Hong Kong, Hong Kong

nctpstt

tatmd(tnIrqMptobmRRppts

cJNE

ackground & Aims: Aspirin exerts antitumor effectartly through blocking tumor promoter-induced activa-or protein-1 (AP-1) activation. The aim of this study is toetermine how specific COX-2 inhibitor SC-236 medi-tes antitumor effect by modulation of AP-1-signalingathway. Methods: AP-1 transcriptional activity andNA-binding activity were detected by luciferase re-orter assay and gel shift assay, separately. Mitogen-ctivated protein kinase (MAPK) activation was deter-ined by Western blot and in vitro kinase assay.ntisense oligonucleotide against c-Jun-N-terminal ki-ase (JNK) was used to suppress JNK expression.esults: We showed that SC-236 inhibited 12-O-tetra-ecanoylphorbol-13-acetate (PMA)-induced cell transfor-ation in a dose-dependent manner in JB6 cells. At aose range (12.5–50 �mol/L) that inhibited cell trans-ormation, SC-236 also inhibited anchorage-indepen-ent cell growth and AP-1-activation in 3 gastric cancerells, independent of COX-prostaglandin synthesis. SC-36 down-regulated c-Jun-NH2-terminal kinase phos-horylation and activity. Suppression of JNK activity re-ersed the inhibitory effect on AP-1 activity by SC-236nd suppressed gastric cancer cell growth, indicatinghat the inhibitory effect of SC-236 on AP-1 activationnd cell growth was through interaction with JNK.onclusions: The inhibitory effect on JNK-c-Jun/AP-1 ac-ivation contributes to the antitumor effect of COX-2-pecific inhibitor, and inhibition of JNK activation mayave a therapeutic benefit against gastric cancer.

spirin, nonsteroidal anti-inflammatory drugs (NSAIDs),and cyclooxygenase-2 (COX-2) inhibitors have re-

eived great attention because of their protective effectsgainst cancer.1–4 COX inhibition may explain part ofhe antitumor activity of NSAIDs. However, NSAIDslso modulate cyclooxygenase-independent signal trans-uction pathways.5–7 Mechanisms such as induction ofpoptosis8–12 and inhibition of angiogenesis13,14 are ob-erved only at high concentrations of the respectiveSAIDs, which are 100- to 1000-fold higher than those

eeded to inhibit prostaglandin synthesis, in all cellulture studies. Furthermore, a recent clinical study onhe inhibitory effect of celecoxib in colonic adenomas inatients suffering from familial adenomatous polyposishowed that a high dose is needed.15 These data indicatehat cyclooxygenase-independent mechanisms are impor-ant for the antitumor effect of NSAIDs at high doses.

Activator protein 1 (AP-1) is an inducible eukaryoticranscription factor comprised of the products of the Junnd Fos oncogene families.16–18 Several reports have es-ablished the role of AP-1 activation in cellular transfor-ation and tumor promotion.19–21 In JB6 mouse epi-

ermal cell lines, 12-O-tetradecanoylphorbol-13-acetatePMA) and epithelial growth factor (EGF) induce AP-1ranscriptional activity in promotion-sensitive (P�) phe-otypes but not in promotion-resistant (P�) phenotypes.n contrast, blocking AP-1 induction causes P� cells toevert to the P� phenotype, indicating the unique re-uirement for AP-1 activity in cell transformation.21

itogen-activated protein kinase (MAPK)-mediatedhosphorylation is important for the expression and post-ranslational modification of AP-1 complex.22,23 As withther MAP kinase cascades, the JNK/c-Jun pathway haseen shown to regulate cell growth and tumorigenesis inany different malignancies.24–26 It is established thatas-induced transformation requires c-Jun27 and thatas induces c-Jun phosphorylation on sites that arehosphorylated by JNK.28 In addition, it has been re-orted that JNK is constitutively activated in severalumor cell lines and that the transforming actions ofeveral oncogenes have been reported to be JNK depen-

Abbreviations used in this paper: AP-1, activator protein 1; COX,yclooxygenase; ERK, extracellular signal-regulated protein kinase;NK, c-Jun-N-terminal kinase; MAPK, mitogen-activated protein kinase;SAIDs, nonsteroidal anti-inflammatory drugs; PGE2, prostaglandins2; PMA, 12-O-tetradecanoylphorbol-13-acetate.

© 2004 by the American Gastroenterological Association0016-5085/04/$30.00

doi:10.1053/j.gastro.2003.10.063

dJ

lfmabetigiitJtiAeN

AtpiRULlvawiNSciMJ

mmo(tir

i(l

sBtCAGTwotf

plJhscc

nflPkaahw�wbsb1p

d1m1ans

January 2004 COX-2 INHIBITOR SUPPRESSES AP-1 137

ent.29 These data strongly support the hypothesis thatNK is relevant to cancer.

A variety of NSAIDs such as aspirin, sodium salicy-ate, and sulindac have been shown to inhibit cell trans-ormation and tumor promotion by blocking tumor pro-oter-induced AP-1 activation in both mouse JB6 cells

nd human tumor cells.20,30–33 However, whether inhi-ition of AP-1 activity contributes to the antitumorffect of specific COX-2 inhibitors is still not clear. Inhe present study, we show that the specific COX-2nhibitor SC-236 inhibits anchorage-independent cellrowth and AP-1 activation in gastric cancer cells. Thenhibition of AP-1 activation and cell growth occursndependent of COX-2-prostaglandin synthesis buthrough suppression of stress-activated protein kinase/un-terminal kinase JNK cascade. To our knowledge,his is the first report showing that specific COX-2nhibitor suppresses tumor growth through inhibition ofP-1 activity and that down-regulation of SAPK/JNK is

ssential for inhibition of AP-1 activation in cancer bySAIDs.

Materials and MethodsCell Culture and Drug Treatments

Three gastric cancer cell lines were used in this study.GS was purchased from the American Type Culture Collec-

ion (ATCC, Rockville, MD). MKN-28 and MKN-45 wereurchased from RIKEN (The Institute of Physical and Chem-cal Research), Cell Bank, Japan. Cells were maintained inPMI-1640 containing 10% fetal bovine serum (FBS), 100/mL�1 penicillin, 100 �g/mL�1 streptomycin (Gibco BRL,ife Technologies, NY). Mouse epidermal JB6 P� with AP-1uciferase reporter stable transfectant cells were kindly pro-ided by Dr. Li JJ (National Cancer Institute, Frederick, MD)nd grown in Eagle’s minimal essential medium supplementedith 200 �g/mL gentamicin.21 SC-236 (a COX-2 specific

nhibitor) was obtained from Searle, IL. Aspirin, NS-398,imesulide, and Sunlindac Sulfone were purchased from

igma (St Louis, MO). PGE2, PGH1, and PGH2 were pur-hased from Cayman Chemicals (Ann Arbor, MI). MEK1/2nhibitor U0126 was from New England Biolabs (Boston,

A). JNK inhibitor SP600125 was from Calbiochem (Laolla, CA).

MTT Assay

About 5000 cells per well were grown in 96-wellicrotiter plates and incubated overnight in 100 �L of cultureedium. Cells were then treated with different concentrations

f SC-236 for fixed time intervals. Ten microliters MTTFluka, Buchs, Switzerland) labeling reagent (final concentra-ion 0.5 mg/mL) was added into each well, and the cells werencubated for another 4 hours at 37°C. The supernatant wasemoved, and 100 �L of 0.04 mol/L hydrochloric acid in

sopropanol was added to each well. A micro ELISA readerBio-Rad, Hercules, CA) measured the absorbency at a wave-ength of 595 nm.

Oligonucleotide Treatment

The phosphorothioate oligonucleotides used in thistudy were synthesized and purified by Genset Singaporeiotech Ltd. (Singapore). The sequences of the oligonucleo-

ides used are as follows: JNK1AS, 5�-CTC TCT GTA GGCCG CTT GG-3�; JNK2AS, 5�-GTC CGG GCC AGG CCAAG TC-3�; JNK1Scr, 5�-CTT TCC GTT GGA CCC CTGG-3�; and JNK2Scr, 5�-GTG CGC GCG AGC CCG AAAC-3� as described previously.34,35 Cells growing in log phaseere treated with 0.4 �mol/L oligonucleotides in the presencef 6 �L Lipofectamine 2000 reagent (Invitrogen) per millili-er. Forty-eight hours after treatment, the cells were collectedor Western blot analysis or JNK activity assay.

Proliferation Assay

Cells were seeded in 24-well tissue culture plates in theresence of RPMI1640 supplemented with 10% FBS. One dayater, they were transfected with 0.4 �mol/L JNKAS, andNK scrambled oligonucleotide as described above. After 4ours of lipofection, the cells were transferred to mediumupplemented with 2% FBS. Five days later, the cells wereounted with a Coulter counter. For each experiment and allonditions, triplicate wells were counted.

In Vitro JNK Activity Assay

The kinase activity assay was carried out using a Ki-aseSTAR JNK activity assay kit (Biovision). Briefly, subcon-uent monolayer cells (in 100-mm plates) were washed withBS and resuspended in cold JNK lysis buffer supplied in theit for 10 minutes. Cell debris was removed by centrifugationt 14,000 rpm at 4°C for 15 minutes. Two microliters of JNKntibody was added to 200 �L cell lysate and incubated for 1our at room temperature. Resuspended Protein A sepharoseas added and incubated for another 1 hour. After wash, 50L of kinase assay buffer and 2 �L c-Jun Protein/ATP mixtureere added to each immunoprecipitation sample and incu-ated for 4 hours at 30°C. Protein A bead was spun down andupernatant was collected. Finally, we performed Westernlotting using the rabbit anti-phospho-c-Jun antibody at:1000 dilutions. A 35-kilodalton band corresponding to thehosphorylated c-Jun protein was detected.

Clonogenic Cell Survival Assay

Gastric cancer cells and JB6 P� cells were exposed toifferent concentrations of SC-236 with or without PMA in.5 mL of 0.33% agar medium over 3 mL of 0.5% agaredium as described previously.21,33 Colonies were scored at

4 days, fixed with 70% ethanol, stained with Coomassie Blue,nd counted under a dissection microscope. Only those colo-ies containing at least 50 cells were considered to be viableurvivors.

TnPtMdcctaBa

TnCTawoatbfS

cbmtcwecr(Amttneblm(e

dwCmKtfPsTmmDt4ms

rwcapcukEpC

wDTmDGpopT

ews

138 WONG ET AL. GASTROENTEROLOGY Vol. 126, No. 1

Prostaglandins E2 Production

Cells were plated at 5 � 104/well into 24-well plates.wenty-four hours later, the cultures were exposed to 100mol/L PMA with or without SC-236 for another 24 hours.GE2 secreted into the medium was determined by competi-ive ELISA using a kit from Cayman Chemicals (Ann Arbor,I) according to the manufacturer’s instructions. Briefly, me-

ium from the cultured cells were added to 96-well platesoated with an anti-mouse antibody, mixed with a PG/acetyl-holinesterase tracer and a monoclonal antibody against pros-aglandins, and incubated at 4°C overnight. Unbound PG/cetylcholinesterase was removed and washed extensively.ound acetylcholinesterase was detected by Ellman’s reagentnd measured at 410 nm.

COX-2 Activity Assay

Cells were plated at 5 � 104/well into 24-well plates.wenty-four hours later, the cultures were exposed to 100mol/L PMA with or without SC-236 for another 24 hours.ells were collected and homogenized in cold buffer (0.1 mol/Lris-HCl, pH 7.8, 1 mmol/L EDTA, 250 mmol/L mannitol,nd 0.3 mmol/L diethyldithiocarbamic acid). COX activityas assayed colorimetrically by monitoring the appearance ofxidized N,N,N�,N�-tetramethyl-�-phenylenediamine (TMPD)t 590 nmol/L using a kit from Cayman Chemicals accordingo the manufacturer’s instructions. The reaction was initiatedy adding 20 �L arachidonic acid solution and incubatedor 5 minutes at 25°C. COX-1 activity was inhibited byC560.

Transfection and Luciferase Assayof AP-1 Activity

AP-1 reporter plasmid was constructed by insertingollagenase promoter region (�73 to �67 containing 1 AP-1inding site) into luciferase reporter vector pGL-3-basic (Pro-ega Corp., Madison, WI).36 Signal transduction pathway

ransreporting system containing pFR-Luc plasmid, pFA2-Jun plasmid, pFC2-dbd plasmid, and pFC-MEKK plasmidere purchased from Stratagene. For transient transfection

xperiments, cells were seeded in 12-well plates to 70%–80%onfluence. The cells were transfected with 0.8 �g/well AP-1eporter plasmid using lipofectamine 2000. PRL-CMV vector0.01 �g/well, Promega) was cotransfected as internal control.fter transfection for 4 hours, cells were changed to normaledium and allowed to recover overnight. Cells were first

reated with different concentrations of SC-236 for 2 hours andhen incubated in media in the absence or presence of 100mol/L PMA for an additional 24 hours. For the inhibitorxperiments, the cells were pretreated with inhibitors 1 hourefore SC-236 treatment. Transfected cells were collected andysed, and the firefly and renilla luciferase activities were

easured using the Dual-Luciferase Reporter assay systemPromega, Madison, WI) with a model TD-20/20 Luminom-ter.

Preparation of Cytoplasmicand Nuclear Extract

Nuclear and cytoplasmic extracts were prepared asescribed by Dignam et al.37 Confluent cells in 10-cm dishesere treated for various times with the indicated effectors.ells were resuspended in 400 �L buffer A (containing 10mol/L HEPES, pH 7.9, 1.5 mmol/L MgCl2 , 10 mmol/LCl, 0.5 mmol/L DTT, 0.5 mmol/L PMSF, 1 �g/mL leupep-

in, 1 �g/mL aprotinin, and 1 �g/mL pepstatin A), kept on iceor 15 minutes, lysed gently with 12.5 �L of 10% Nonide-40, and centrifuged at 2000g for 10 minutes at 4°C. Theupernatant was collected and used as the cytoplasmic extracts.he nuclei pellet was resuspended in 40 �L buffer C (20mol/L HEPES, pH 7.9, containing 1.5 mmol/L MgCl2 , 450mol/L NaCl, 25% glycerol, 0.2 mmol/L EDTA, 0.5 mmol/LTT, 0.5 mmol/L PMSF, 1 �g/mL leupeptin, 1 �g/mL apro-

inin, 1 �g/mL pepstatin A) and agitated for 30 minutes at°C, and the nuclear debris was spun down at 20,000g for 15inutes. The supernatant (nuclear extract) was collected and

tored at �80°C until ready for analysis.

Western Blotting

Twenty micrograms of protein were resolved and sepa-ated by electrophoresis on a 10% denaturing SDS gel. Proteinsere electroblotted onto nitrocellulose membranes. Detection was

onducted by immunostaining using specific primary antibodiesnd horseradish peroxidase-conjugated anti-IgG antibody. Therotein bands were visualized by the enhanced chemilumines-ence (ECL) assay (Amersham Pharmacia Biotech) following man-facturer’s instructions. Phospho-MAPK family antibody samplerit and the PhosphoPlus c-Jun Kit were purchased from Newngland Biolabs (Boston, MA). c-fos Antibody and horseradisheroxidase-conjugated anti-rabbit IgG were obtained from Santaruz Biotechnology (Santa Cruz, CA).

Electrophoretic Mobility Shift Assay

Eight micrograms of nuclear proteins were incubatedith 1 �g each of poly (dI.dC) in the presence of 30 fmol ofigoxin (DIG)-labeled, double-stranded AP-1 probe (5�-CGCTG ATG ACT CAG CCG GAA-3�; Santa Cruz) for 15inutes at room temperature in a total volume of 20 �L usingIG gel shift kit (Roche Diagnostics GmbH, Mannheim,ermany). Oligonucleotide competition experiments wereerformed in the presence of 50-fold excess of unlabeled AP-1ligonucleotides. DNA complexes were resolved from freerobe with 4% nondenaturing polyacrylamide gels in 0.5Xris-borate-EDTA (pH 8.3) and visualized by fluorography.

Statistical Analysis

The data shown were mean values of at least 3 differentxperiments and expressed as means � SD. The Student t testas used for comparison. A P value �0.05 is considered

tatistically significant.

2gtFg3e2aWtictwMapde4Si

pPgytawhActhilteIc

Faw1Jaf5dpexperiments � standard error.


Results

Specific COX-2 Inhibitor SuppressedAnchorage-Independent Growth of GastricCancer Cells

To clarify whether specific COX-2 inhibitor SC-36 suppresses the transformed phenotype of humanastric epithelial cells, the colony-forming abilities ofhese cells in soft agar were examined. As shown inigure 1A, SC-236 inhibited anchorage-independent cellrowth in a concentration-dependent manner in all of thegastric cancer cell lines tested. However, the inhibitory

ffect was most profound in AGS cells with IC50 at about0–25 �mol/L. The mouse epidermal JB6 cell system iswell-developed model for studying tumor promotion.e, therefore, used this model to further test the anti-

umor effect of SC-236. As shown in Figure 1B, SC-236nhibited PMA-induced transformation of JB6 cells in aoncentration-dependent manner. The effective inhibi-ory concentration ranged from 12.5 to 50 �mol/L inhich no cytotoxic effect on JB6 cells were observed byTT assay (Figure 1C). We also compared the effect of

spirin on cell transformation in JB6 cells. As antici-ated, aspirin inhibited cell transformation in a dose-ependent manner (Figure 1B), with IC50 approximatelyqual to 1 mmol/L. This IC50 value is approximately0-fold higher than that of SC-236, indicating thatC-236 is a more potent inhibitor of cell transformationn JB6 cells.

SC-236 Inhibited PMA-Induced AP-1Activation in Gastric Cancer Cells

Activation of AP-1 by growth factor or tumorromoters such as tumor necrosis factor (TNF)- orMA has been implicated in cell transformation and cellrowth.16–18 NSAIDs, such as aspirin and sodium salic-late, have been shown to inhibit cell transformationhrough suppression of PMA or UV-stimulated AP-1ctivation.19,33 These findings prompted us to determinehether SC-236 is an inhibitor of AP-1 activation. Threeuman gastric cell lines were transiently transfected withP-1 luciferase reporter gene and treated with different

oncentrations of SC-236 for 2 hours then incubated inhe presence of 100 nmol/L PMA for an additional 24ours. As shown in Figure 2A, PMA caused a remarkablencrease in AP-1 transcriptional activity in 3 gastric cellines, whereas SC-236 inhibited PMA-stimulated AP-1ranscriptional activity in a dose-dependent manner. Theffect of SC-236 was most profound in AGS cells withC50 approximately equaled to 25 �mol/L, which wasonsistent with the cell growth data. In addition, we

igure 1. Effects of SC-236 on anchorage-independent cell growthnd PMA-induced cell transformation. (A) 104 Gastric cancer cellsere exposed to different concentrations of SC-236 in 0.33% agar for4 days and scored for colonies at the end of the experiment. (B) 104

B6 cells were exposed to different concentrations of SC-236 orspirin in the presence or absence of 100 nmol/L PMA in 0.33% agaror 14 days and scored for colonies at the end of the experiment. (C)

� 103 JB6 cells were plated on 96-well plates and treated withifferent concentrations of SC-236 for 24 hours. MTT assay waserformed. All results were expressed as the mean of 3 independent

aictsc

Ct

disOs3slatbSiis

gclaSsiwbSJaJmFJthpSJaptarMip

Fichwlssabpwpul


lso showed that SC-236 significantly inhibited TNF--nduced activation of AP-1 transcription in gastric cancerells (data not shown), which indicated that the inhibi-ory effect on AP-1 activation by SC-236 was not exclu-ive to PMA stimulation. Also, SC-236 did not signifi-antly affect the transcriptional activity of the control

igure 2. SC-236 suppressed AP-1 transcriptional and binding activ-ty. (A) Cells transiently expressing AP-1 luciferase reporter geneonstruct were treated with different concentrations of SC-236 for 2ours then followed with 100 nmol/L PMA for another 24 hours. Cellsere then harvested for analysis of luciferase activity. The firefly

uciferase reading was normalized to renilla luciferase reading. Re-ults were expressed as the means of 3 independent experiments �tandard error. (B) AGS cells were treated with 50 �mol/L SC-236lone for 2 hours or pretreated with SC-236 for 1 hour, followedy 100 nmol/L PMA for another 1 hour. Nuclear extracts wererepared and analyzed in an electrophoretic mobility shift assayith a DIG-labeled AP-1 probe. Equal amounts (6 �g) of nuclearrotein were loaded in each lane. In lane 5, a 50-fold excess ofnlabeled oligonucleotide was added before the addition of DIG-

abeled probe.

MV or SV40 promoter (data not shown), indicatinghat the inhibition of AP-1 activation was specific.

To further elucidate the mechanism of SC-236–me-iated inhibition of AP-1 activation, AP-1 DNA-bind-ng activity was measured by electrophoretic mobility-hift assay after pretreatment of AGS cells with SC-236.ur findings demonstrated that SC-236 (50 �mol/L)

lightly suppressed AP-1 DNA binding (Figure 2B, lane). Treatment with PMA (100 nmol/L) resulted in aignificant increase in AP-1 DNA binding (Figure 2B,ane 2). This increased binding could be eliminated bydding a 50-fold excess of unlabeled AP-1 oligonucleo-ide, confirming that the electrophoretic mobility-shiftand was specific for AP-1 binding (Figure 2B, lane 5).C-236 treatment markedly inhibited AP-1 DNA bind-ng induced by PMA (Figure 2B, lane 4). Thus, SC-236nhibited both AP-1 DNA-binding activity and tran-criptional activity in gastric cancer cells.

SC-236 Inhibited PMA-InducedPhosphorylation of JNK

Because the maximal inhibitory effect on cellrowth and AP-1 activity was found in AGS cells, wehose to use AGS cells in the following study. Extracel-ular signals, including growth factors, phorbol ester,nd UV irradiation stimulate phosphorylation of c-Jun ater-63/-73 by JNK and activate AP-1-dependent tran-cription. To clarify whether the JNK-c-Jun pathway isnvolved in the inhibition of AP-1 activation by SC-236,e determined both phosphorylation of JNK and c-Juny Western blot assay in AGS cells. We showed thatC-236 inhibited PMA-induced phosphorylation of c-un and JNK, whereas Western blotting with anti-JNKnd anti-c-Jun antibodies showed that the recovery ofNK or c-Jun from cell lysates was not altered by treat-ent with SC-236 (Figure 3A and 3B). As shown in

igure 3B, PMA led to marked induction of JNK2 andNK1 phosphorylation by 4.2- and 2.8-fold, respec-ively. The difference in favor of JNK2 may be due to aigher basal level of activity exhibited by JNK1 com-ared with JNK2. However, treatment with 50 �mol/LC-236 led to comparable suppression of JNK2 andNK1 phosphorylation by 60% and 55%. To substanti-te the result from Western blot analysis, we furthererformed JNK activity assay. We showed that SC-236reatment significantly inhibited PMA-stimulated JNKctivity, which was consistent with our Western blotesults (Figure 3C). To further determine whether

APKs other than JNK have been involved in thenhibitory effect of SC-236, we tested phosphorylatedrotein expression of ERK and p38 after SC-236 treat-

miEucFtetJhsh

watswwLms

F2csiaipo


ent. We showed that only 50 �mol/L SC-236 signif-cantly inhibited PMA-induced phosphorylation ofRK1/ERK2, whereas phosphorylation of p38 remainednchanged with SC-236 treatment in the range of con-entration tested (Figure 3B). The findings presented inigure 3 indicated that the MAPK pathway, particularlyhe JNK/c-Jun pathway, might play a critical role in theffect of SC-236-mediated suppression of AP-1 activa-ion. As a control, we also tested the effect of aspirin onNK activity. Our result showed that aspirin did notave any effect on PMA-induced JNK activation ashown in Figure 3C, indicating that aspirin might in-ibit AP-1 activation through a different pathway.

igure 3. SC-236 inhibited PMA-induced JNK phosphorylation and acthours then exposed to PMA for another 1 hour. Total and phosphory

-Jun phosphorylation. (B) SC-236 inhibited JNK phosphorylation. Qucanning densitometry of the bands, and the values shown are thencrease above the control values. (C) The effect SC-236 and aspirinssay. The JNK activity kit utilized a JNK-specific antibody to immuno

n a kinase reaction using recombinant c-Jun as substrate. Phosphospho-c-Jun specific antibody. The right panel showed the values (btained from scanning densitometry expressed as a percentage of

SC-236 Inhibited AP-1 Activation ThroughJNK-c-Jun Pathway

To further elucidate the role of JNK/c-Jun path-ay in SC-236 mediated AP-1 suppression, we took

dvantage of Stratagene’s PathDetect in vivo signalransduction transreporting system (Stratagene). Thisystem includes a unique fusion transactivator plasmid,hich consists of c-Jun transcriptional activator fusedith the yeast GAL4 DNA-binding domain, and a pFR-uc reporter plasmid, which contains a synthetic pro-oter with 5 tandem repeats of the yeast GAL4-binding

ites that control expression of the Photinus pyralis lu-

AGS cells were pretreated with different concentrations of SC-236 forproteins were detected with Western blot assay. (A) SC-236 inhibitedcation of phosphorylation of c-Jun, ERK, and JNK was performed bys � SE (n 3) and are expressed as percentages of the maximalNK kinase activity. JNK activity was determined by an in vitro kinaseitate JNK from cell lysates. Activity of the JNK was then determinedtion of the c-Jun was analyzed by Western blot analysis using a� SE, n 3) of the level of JNK activation by in vitro kinase assayaximum increase. *P � 0.05.

ivity.latedantifimeanon J

preciphorylameanthe m

cvara

twtP1mtMLOitp

iJ2ionri4etHsb

t

Š

Ftp2CwsCht2he(pPahafw


iferase gene. Phosphorylation of the transcription acti-ation domain of the fusion c-Jun protein by JNK willctivate transcription of the luciferase gene from theeporter plasmid, and their activity reflects the in vivoctivation of JNK and the corresponding specific signal

ransduction pathways. In this study, after transfectionith pFR-Luc and pFA2-cJun into AGS cells, we treated

he cells with PMA for 12 hours. As shown in Figure 4A,MA simulated the pFR-Luc reporter activity to about.8-fold compared with the control, whereas pretreat-ent with 25 �mol/L SC-236 significantly inhibited

his activation. Furthermore, we cotransfected pFC-EKK plasmid, which was the positive control for pFR-

uc activation into AGS cells, and treated with SC-236.ur results showed that SC-236 also significantly inhib-

ted MEKK-stimulated reporter activation, indicatinghat SC-236 specifically inhibited MEKK-JNK-c-Junathway.To substantiate the result that JNK mediates the

nhibitory effect of SC-236, we utilized a highly selectiveNK inhibitor, SP600125, to observe its effect on SC-36-mediated inhibition of AP-1 activation. SP600125nhibits JNK with more than a 20-fold selectivity vs.ther kinases such as extracellular signal-regulated ki-ases, p38 kinase, or protein kinase C, and others.38 Ouresults showed that SP600125 completely reversed thenhibitory effect on AP-1 activity by SC-236 (FigureB), indicating that JNK is required for the inhibitoryffect of SC-236 on AP-1 activation. We also suppressedhe activation of ERK by its specific inhibitor U0126.owever, pretreatment of AGS cells with U0126 had no

ignificant effect on the inhibitory effect on AP-1 activityy SC-236 (Figure 4C).

JNK Inhibition Resulted in Cell GrowthSuppression in AGS Cells

Although PMA has been shown to stimulate gas-ric cell growth,39,40 there is no data concerning the role

igure 4. SC-236 inhibited JNK-c-Jun pathway. (A) AGS cells wereransiently cotransfected with pFR-Luc and pFA2-c-Jun plasmid in theresence or absence of pFC-MEKK and treated with 25 �mol/L SC-36 for 2 hours followed with 100 nmol/L PMA for another 12 hours.ells were then harvested for analysis of luciferase activity. Resultsere expressed as the means of 3 independent experiments �tandard error. (B). Top panel: Western blot analysis of phospho-JNK.ells were pretreated with or without 2.5 �mol/L SP600125 for 1our, followed with PMA for another 1 hour. Bottom panel: AP-1ranscriptional activity assay. Cells were pretreated with or without.5 �mol/L SP600125 for 1 hour and then treated with SC-236 for 2ours, followed with PMA for another 24 hours. At the end of thexperiments, cells were harvested for analysis of luciferase activity.C) Top panel: Western blot analysis of phospho-ERK. Cells wereretreated with or without 10 �mol/L U0126 for 1 hour, followed withMA for another 1 hour. Bottom panel: AP-1 transcriptional activityssay. Cells were pretreated with or without 10 �mol/L U0126 for 1our and then treated with SC-236 for 2 hours, followed with PMA fornother 24 hours. At the end of the experiments, cells were harvestedor analysis of luciferase activity. *P � 0.01, #P � 0.05, comparedith PMA treatment without SC-236, separately.

oTgaaawJcgstp

fhsceJticlo

datJwofJgcfrJbwisn

oCaCwPFPahmNcaAsA

FgtJPJfhJJasFci*


f JNK in the regulation of gastric cancer cell growth.o further establish a clear link between gastric cancerrowth and JNK activity, we performed cell growthnalysis with antisense oligonucleotides of both JNK1nd JNK2. First, we investigated the effect of JNK1ASnd JNK2AS on JNK expression in AGS cells. Cellsere treated either with 0.4 �mol/L JNK1AS or

NK2AS individually or with JNK1AS and JNK2AS inombination (0.2 �mol/L each). Scrambled-sequence oli-onucleotides (JNK1Scr and JNK2Scr) were used at theame concentrations and served as controls. Both ASreatments led to �60% reduction in the correspondingrotein levels (Figure 5A), whereas neither mock lipo-

igure 5. Suppression of JNK activity inhibited gastric cancer cellrowth. (A) Total protein was extracted from cells 48 hours followingreatment with 0.4 �mol/L JNK1AS, 0.4 �mol/L JNK2AS, 0.2 �mol/LNK1 AS � 0.2 �mol/L JNK2AS, or control oligonucleotides (JNKScr).rotein samples were analyzed by Western blot analysis using specificNK antibody. (B) Total Jun kinase activity was determined 1 hourollowing exposure to PMA by an in vitro kinase assay. Forty-eightours before PMA treatment, cells were transfected with JNK1 AS,NK2AS, combinations of JNK1AS plus JNK2AS (0.4 �mol/L each), orNKScr. (C) JNKAS inhibited cell growth in AGS cells. Proliferationssays were performed as described in the Materials and Methodsection. The cells were maintained in 2% FBS during the experiment.ive days after the treatment, the cells were counted with a Coulterounter. The proliferation data shown here were the average of 2dentical and independent experiments, each carried out in triplicate.P � 0.05.

ectamine nor treatment with scrambled oligonucleotidesad any effect on JNK1 or JNK2 protein expression. Toubstantiate the results obtained with antisense oligonu-leotides, an immunocomplex kinase assay was used toxamine basal and PMA-induced JNK activity in mock-,NKAS-, and JNKScr-treated cells (Figure 5B). Consis-ent with the reduction in JNK protein levels, PMA-nduced JNK activation was markedly reduced in AGSells treated with JNK1AS and JNK2AS. Neither mockipofectamine nor treatment with scrambled oligonucle-tides affected the JNK kinase activity.

Taken together, the experiments described aboveemonstrated that, using JNKAS, we effectivelychieved a significant reduction in JNK expression andherefore directly investigated the roles of JNK1 andNK2 in regulating the growth of AGS cells. All cellsere counted 5 days after transfection with antisenseligonucleotides or scrambled oligonucleotides or Lipo-ectamine. We showed in this study that inhibition ofNK activity by antisense strategy markedly suppressedastric cancer cell growth. As shown in Figure 5C, AGSells displayed a significant reduction in cell viabilityollowing treatment with JNKAS; the cell growth waseduced to 70% with JNKAS1 and to 72.4% withNKAS2. Furthermore, the cell growth suppression byoth JNK1AS and JNK2AS was most pronounced,here viability was reduced to 51.7% by 5 days follow-

ng treatment with the oligonucleotide. Altogether, re-ults from our study indicate that JNK has a predomi-ant role in gastric cancer cell growth.

The Inhibitory Effects of SC-236 on AP-1and JNK Are Independent of COX-2-Prostaglandin Inhibition

To further explore whether the inhibitory effectsn AP-1 and JNK activation by SC-236 occur throughOX-prostaglandins inhibition and whether these effectsre general to all COX-2 inhibitors, we used otherOX-2 inhibitors and sunlindac sulfone in this study asell. First, we determined COX-2 enzyme activity andGE2 production after SC-236 treatment. As shown inigure 6A and 6B, 1 �mol/L SC-236 already suppressedMA-induced COX-2 activity and PGE2 production tobout 20% and 50%, respectively, of the control. Withigher concentrations of SC-236, PGE2 production re-ained under the detection level (�5 pg/105 cells).ext, we investigated the effects of other NSAIDs in-

luding 2 COX-2 inhibitors, NS-398 and nimesulide,nd sunlindac sulfone on AP-1 and JNK activation inGS cells. Our results showed that both nimesulide and

unlindac sulfone significantly inhibited PMA-inducedP-1 activation, whereas NS-398 did not (Figure 7A).

FiadiA7aa

ep

gdcuetPiaees

ic1mtpatptatdu�ma�

tcMataiaItdeybIfirme

F(wsS2hEm


urthermore, neither NS-398 nor sulindac sulfone inhib-ted JNK or c-Jun phosphorylation stimulated by PMAs shown in Figure 7B. To clarify this issue further, weetermined the effect of exogenous prostaglandins in thenduction of anchorage independent cell growth andP-1 activity in AGS cells. As shown in Figure 7C andD, exogenous prostaglandins had no effect on anchor-ge-independent cell growth and PMA-stimulated AP-1ctivation in the presence of SC-236.

Discussion

SC-236 inhibited tumor growth through differ-nt mechanisms such as induction of apoptosis,12 sup-ression of cell proliferation,41 and inhibition of angio-

igure 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 orithout SC-236 for 24 hours. COX-2 activity was determined as de-cribed in the Materials and Methods section. Results are mean �D of triplicate assays and 2 experiments. (B) Cells were plated on4-well plates and exposed to PMA with or without SC-236 for 24ours. PGE2 content of the conditioned media was determined byLISA. Results are mean � SD of triplicate assays and 2 experi-ents. *P � 0.05.

enesis.5,13 Our group previously demonstrated that highoses of SC-236 (12.5–100 �mol/L) suppressed gastricancer cell growth. However, the mechanism(s) remainsnknown. AP-1 is considered a mediator of carcinogen-sis by its ability to alter gene expression in response toumor promoters, such as epidermal growth factor,MA, or UV irradiation.18 Blocking of tumor promoter-

nduced AP-1 activity by some NSAIDs, such as aspirinnd sodium salicylate, partly explains their antitumorffects.30–33 Thus, we hypothesized that SC-236 mightxert its antitumor effect through modulation of AP-1ignaling pathway in gastric cancer.

In the present study, we demonstrated that SC-236nhibited anchorage-independent cell growth in 3 gastricancer cell lines with the concentration ranging from2.5 �mol/L to 50 �mol/L. To substantiate the antitu-or effect of SC-236, we used the well-established cell

ransformation model and showed that SC-236 sup-ressed PMA-induced JB6 cell transformation with IC50

pproximately equal to 25 �mol/L. Based on the findingshat the same range of SC-236 (12.5–50 �mol/L) sup-ressed PMA-induced AP-1 DNA binding and transac-ivation in gastric cancer cells, we suggest that thentitumor effects of SC-236 are due, at least in part, tohe suppression of AP-1 activity. Of note, most studiesemonstrating effects on COX-independent pathwaystilized higher concentrations of NSAIDs (100–1000mol/L).42 In our study, SC-236 inhibited cell transfor-ation and AP-1 activity dose dependently within the

chievable physiologic range of between 12.5 and 50mol/L.AP-1, consisting of Jun/Fos dimers, is a downstream

arget of MAP kinase family members including extra-ellular signal regulated kinases (ERK-1 and -2; p42/p44APK), Jun kinases (JNK), and p38 MAPK. Aspirin

nd sodium salicylate suppressed AP-1 activationhrough different mechanisms in both mouse JB6 cellsnd normal human cells.20,30–33 Aspirin inhibited UVB-nduced AP-1 activity through blocking UVB-inducedctivation of ERK and JNK as well as p38 kinase.33

nterestingly, it blocked TPA- or epidermal growth fac-or-induced signaling through a MAP kinase-indepen-ent pathway.20 These findings agreed with Schwengert al., who reported that the same dose of sodium salic-late inhibited ERK/JNK activation induced by TNFut not by other growth factors such as EGF, PDGF, andL-1 under the same condition in normal human F4Sbroblasts.31,32 These discrepant results suggest that theegulatory mechanisms of NSAIDs on MAPK familyembers may depend on the cellular context and differ-

nt external stimuli.

picoJWa

sitJsitispaitttrSitcscictTopt

tHfpfta

Š

Fe2Nwww�p10ephl*


In this study, we showed that inhibition of JNKhosphorylation and activation by SC-236 contributed tots effect on AP-1 activation and cell growth. Our con-lusion is based on the following points. First, we dem-nstrated that SC-236 inhibited PMA-induced JNK/c-un phosphorylation as well as JNK activity by both

estern blotting analysis and in vitro kinase activityssay. Second, using the specific signal transreporting

ystem, we found that pretreatment with SC-236 inhib-ted both PMA- and MEKK-stimulated reporter activa-ion, which indicated that the specific MEKK-JNK-c-un pathway was inhibited by SC-236. These resultsuggested that SC-236 interacted with JNK, directly orndirectly, and subsequently inhibited the phosphoryla-ion and activation of JNK by PMA and MEKK. Thenvolvement of JNK signaling was further confirmed byhowing that pretreatment with JNK inhibitor com-letely reversed the inhibitory effect of SC-236 on AP-1ctivation. In this context, we hypothesized that JNKnhibitor blocked interaction of SC-236 with JNK,hereby preventing SC-236 to inhibit JNK and releasinghe inhibitory effect of SC-236 on AP-1 activation. Al-ogether, these results suggested that JNK signaling wasequired for the inhibitory effect on AP-1 activation byC-236. This is the first report showing that NSAIDsnhibited tumor promoter-induced AP-1 activationhrough suppression of MAPK pathway in human can-er. On the other hand, although 50 �mol/L SC-236howed some degree of inhibition of ERK activation, thisoncentration is about 2 times the amount of IC50 fornhibition of cell growth and AP-1 activation in AGSells. Blockage of ERK activity did not have any effect onhe SC-236-mediated inhibition of AP-1 activation.hus, we concluded that SC-236 might have some effectn ERK/MAPK phosphorylation, but the ERK/MAPKathway did not have a predominant role in the inhibi-ory effect of SC-236 on AP-1 activation.

The role of activated JNK in cell survival in responseo extracellular stimuli has been well reported.34,35,43

owever, evidence supportive of a proapoptotic functionor the JNK pathway has also been documented.29,44 Onelausible explanation for these seemingly disparate ef-ects is that JNK serves different functions in diverse cellypes and under different stimuli. Using high-affinitynd high-specificity phosphorothioate antisense oligonu-

igure 7. Exogenous prostaglandins PGE2, PGH1, and PGH2 had noffect on cell growth and AP-1 activation. (A) AGS cells were plated on4-well plates and exposed to 100 �mol/L NS-398 or 100 �mol/Limesulide or 100 �mol/L sunlindac sulfone for 2 hours, followedith PMA for another 12 hours. The AP-1 luciferase enzyme activityas measured using the luminometer. (B) AGS cells were pretreatedith 100 �mol/L sulindac sulfone or 25 �mol/L SC-236 or 100mol/L NS-398 followed with PMA for another 1 hour. Total andhosphorylated proteins were detected with Western blot assay. (C)04 AGS cells were exposed to 2 �g/mL of different prostaglandins in.33% agar for 14 days and scored for colonies at the end of thexperiment. (D) AGS cells were plated on 24-well plates and ex-osed to different concentrations of prostaglandins with SC-236 for 2ours, followed with 100 nmol/L PMA for another 12 hours. The AP-1uciferase enzyme activity was measured using the luminometer.P � 0.05.

csrtOgag

JadaaCtNpisasewpotPaStbm

acifvCiaa

1

1

1

1

1

1

1

1

1

1

2


leotides targeting JNK1 and JNK2,34,35 we demon-trated that inhibition of JNK1 and JNK2 in AGS cellsesulted in marked suppression of cell growth, indicatinghat JNK was necessary for gastric cancer cell growth.thers have shown that JNK was required for epidermalrowth factor-stimulated growth of A549 cells in softgar,34 and JNK antisense treatment inhibited cellrowth and induced apoptosis in glioblastoma.35

To clarify whether the inhibitory effects on AP-1 andNK activation by SC-236 was related to COX-2 enzymectivity, we examined the effects of other NSAIDs. Weemonstrated that sunlindac sulfone, an NSAID lackingntiprostaglandins synthetase activity, inhibited AP-1ctivation significantly, whereas NS-398, a selectiveOX-2 inhibitor, failed to inhibit AP-1 activity. Fur-

hermore, we showed that both sunlindac sulfone andS-398 did not have any effect on JNK and c-Jun

hosphorylation. These results indicated that the inhib-tory effect on AP-1 and JNK activation was not neces-arily related to the inhibition of COX-2 enzyme activitynd was not universal for all COX-2 inhibitors. Severaltudies have suggested that PGE2 promoted tumorigen-sis through the activation of AP-1.45–47 Consequently,e considered the possibility that inhibition of COX-2-rostaglandins synthesis executed autoregulatory effectn AP-1 activity and cell growth. Therefore, we testedhe effect of exogenous prostaglandins (PGE2, PGH1,GH2) and showed that they had no effect on cell growthnd PMA-induced AP-1 transcription in the presence ofC-236. Therefore, the inhibition of prostaglandin syn-hesis was not the mechanism responsible for the inhi-ition of AP-1 transactivation and neoplastic transfor-ation in gastric cancer.In conclusion, we showed that suppression of AP-1

ctivation contributed to the antitumor effect of a spe-ific COX-2 inhibitor in gastric cancer. Moreover, thenhibitory effect on JNK/c-Jun activation was essentialor the suppression of AP-1 activation. Our study pro-ides a novel molecular mechanism by which a specificOX-2 inhibitor exerts its antitumor effect. Our find-

ngs highlight the importance of JNK for carcinogenesisnd suggest that inhibition of JNK activation may havetherapeutic benefit against gastric cancer.

References1. Gupta RA, DuBois RN. Aspirin, NSAIDs, and colon cancer preven-

tion: 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-2inhibitor for potential chemopreventive properties in colon carci-nogenesis. Cancer Res 1996;56:4566–4569.

4. Sawaoka H, Kawano S, Tsuji S, Tsuji M, Gunawan ES, Takei Y,

Nagano Kand Hori M. Cyclooxygenase-2 inhibitors suppress thegrowth of gastric cancer xenografts via induction of apoptosis innude mice. Am J Physiol 1998;274:G1061–1065.

5. Piazza GA, Rahm AK, Finn TS, Fryer BH, Li H, Stoumen AL,Pamukcu R, Ahnen FJ. Apoptosis primarily accounts for thegrowth-inhibitory properties of sulindac metabolites and involvesa mechanism that is independent of cyclooxygenase inhibition,cell cycle arrest, and p53 induction. Cancer Res 1997;57:2452–2459.

6. Hanif R, Pittas A, Feng Y, Koutsos MI, Qiao L, Staiano-Caico L,Shiff SI, Rigas B. Effects of nonsteroidal anti-inflammatory drugson proliferation and on induction of apoptosis in colon cancer bya prostaglandins-independent pathway. Biochem Pharmacol1996;52:237–245.

7. Elder DJ, Halton DE, Hague A, Paraskeva C. Induction of apoptoticcell death in human colorectal carcinoma cell lines by a cycloox-ygenase-2 (COX-2)-selective nonsteroidal anti-inflammatory drug:independence from COX-2 protein expression. Clin Cancer Res1997;3:1679–1683.

8. Shiff SJ, Qiao L, Tsai LL, Rigas B. Sulindac sulfide, an aspirin-likecompound, inhibits proliferation, causes cell cycle quiescence,and induces apoptosis in HT-29 colon adenocarcinoma cell. J ClinInvest 1995;96:491–503.

9. Goldberg Y, Nassif II, Pittas A, Tsai LL, Dynlacht BD, Rigas B, ShiffSJ. The anti-proliferative effect of sulindac and sunlindac sulfideon HY-29 colon cancer cells: alterations in tumor suppressor andcell cycle-regulatory proteins. Oncogene 1996;12:893–901.

0. Barnes CJ, Cameron IL, Hardman WE, Lee M. Non-steroidal anti-inflammatory drug effect on crypt cell proliferation and apoptosisduring initiation of rat colon carcinogenesis. Br J Cancer 1998;77:573–580.

1. Zhu GH, Wong BC, Ching CK, Lai KC, Lam SK. Differential ap-optosis by indomethacin in gastric epithelial cells through theconstitutive expression of wild-type p53 and/or up-regulation ofc-myc. Biochem Pharmacol 1999;58:193–200.

2. Jiang XH, Lam SK, Lin MC, Jiang SH, Kung HF, Slosberg ED, SohJW, Weinstein IB, Wong BC. Novel target for induction of apopto-sis by cyclo-oxygenase-2 inhibitor SC-236 through a protein ki-nase C-� (1)-dependent pathway. Oncogene 2002;21:6113–6122.

3. Tsuji M, Kawano S, Tsuji S, Sawaoka H, Hori M, DuBois RN.Cyclooxygenase regulates angiogenesis induced by colon cancercells. Cell 1998;93:705–716.

4. Jones MK, Wang H, Peskar BM, Levin E, Itani RM, Sarfeh IJ,Tarnawske AS. Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanisms and implication forcancer growth and ulcer healing. Nat Med 1999;5:1418–1423.

5. Steinbach G, Lynch PM, Phillips RK, Wallace MH, Hawk E, GordonGB. The effect of celecoxib, a cyclooxygenase-2 inhibitor, infamilial adenomatous polyposis. N Engl J Med 2000;342:1946–1952.

6. Angel P, Karin M. The role of Jun, Fos, and the AP-1 complex incell-proliferation and transformation. Biochim Biophys Acta1991;1072:129–157.

7. Curran T, Franza BR Jr. Fos and Jun: the AP-1 connection. Cell1988;55:395–397.

8. Angel P, Imagawa M, Chiu R, Stein B, Imbra RJ, Rahmsdorf HJ,Jonat C, Herrlich P, Karin M. Phorbol ester-inducible genes con-tain a common cis element recognized by a PMA-modulatedtrans-acting factor. Cell 1987;49:729–739.

9. Dong Z, Birrer MJ, Watts RG, Matrisian LM, Colburn NH. Blockingof tumor promoter-induced AP-1 activity inhibits induced transfor-mation in JB6 mouse epidermal cells. Proc Natl Acad Sci U S A1994;91:609–613.

0. Dong Z, Huang C, Brown RE, Ma WY. Inhibition of activatorprotein 1 activity and neoplastic transformation by aspirin. J BiolChem 1997;272:9962–9970.

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

mK

0L


1. Li JJ, Westergaad D, Ghosh P, Colburn NH. Inhibitors of bothnuclear factor- B and activator protein-1 activation block theneoplastic transformation response. Cancer Res 1997;57:3569–3576.

2. Whitmarsh AJ, Davis RJ. Transcription factor AP-1 regulation bymitogen-activated protein kinase signal transduction pathways. JMol Med 1996;74:589–607.

3. Karin M. The regulation of AP-1 activity by mitogen-activatedprotein kinases. J Biol Chem 1995;270:16483–16486.

4. Antonyak MA, Kenyon LC, Godwin AK, James DC, Emlet DR,Okamoto I, Tnani M, Holgado-Madruga M, Moscatello DK, WongAJ. Elevated JNK activation contributes to the pathogenesis ofhuman brain tumors. Oncogene 2002;21:5038–5046.

5. Chen N, Nomura M, She QB, Ma WY, Bode AM, Wang L, FlavellRA, Dong Z. Suppression of skin tumorigenesis in c-Jun NH(2)-terminal kinase-2-deficient mice. Cancer Res 2001;61:3908–3912.

6. Gupta S, Plattner R, Der CJ, Stanbridge EJ. Dissection of Ras-dependent signaling pathways controlling aggressive tumorgrowth of human fibrosarcoma cells: evidence for a potentialnovel pathway. Mol Cell Biol 2000;20:9294–9306.

7. Johnson R, Spiegelman B, Hanahan D, Wisdom R. Cellular trans-formation and malignancy induced by ras require c-jun. Mol CellBiol 1996;16:4504–4511.

8. Pulverer BJ, Kyriakis JM, Avruch J, Nikolakaki E, Woodgett JR.Phosphorylation of c-jun mediated by MAP kinases. Nature 1991;353:670–674.

9. Ip YT, Davis RJ. Signal transduction by the c-Jun N-terminalkinase (JNK)-from inflammation to development. Curr Opin CellBiol 1998;10:205–219.

0. Murono S, Yoshizaki T, Sato H, Takeshita H, Furukawa M, PaganoJS. Aspirin inhibits tumor cell invasiveness induced by Epstein-Barr virus latent membrane protein 1 through suppression ofmatrix metalloproteinase-9 expression. Cancer Res 2000;60:2555–2561.

1. Schwenger P, Skolnik EY, Vilcek J. Inhibition of tumor necrosisfactor-induced p42/p44 mitogen-activated protein kinase activa-tion by sodium salicylate. J Biol Chem 1996;271:8089–8094.

2. Schwenger P, Bellosta P, Vietor I, Basilico C, Skolnik EY, Vilcek J.Sodium salicylate induces apoptosis via p38 mitogen-activatedprotein kinase but inhibits tumor necrosis factor-induced c-JunN-terminal kinase/stress-activated protein kinase activation.Proc Natl Acad Sci U S A 1997;94:2869–2873.

3. Huang C, Ma WY, Dong Z. Ultraviolet B-induced activated pro-tein-1 activation does not require epidermal growth factor recep-tor but is blocked by a dominant negative PKC �/�. Mol CellBiochem 1996;16:6427–6435.

4. Bost F, McKay R, Bost M, Potapova O, Dean NM, Mercola D. TheJun kinase 2 isoform is preferentially required for epidermalgrowth factor-induced transformation of human A549 lung carci-noma cells. Mol Cell Biol 1999;19:1938–1949.

5. Potapova O, Gorospe M, Bost F, Dean NM, Gaarde WA, MercolaD, Holbrook NJ. c-Jun N-terminal kinase is essential for growth ofhuman T98G glioblastoma cells. J Biol Chem 2000;275:24767–24775.

6. Angel P, Baumann I, Stein B, Delius H, Rahmsdorg HJ, Herrlich P. t

12-O-tetradecanoyl-phorbol-13-acetate induction of the humancollagenase gene is mediated by an inducible enhancer elementlocated in the 5-flanking region. Mol Cell Biol 1987;7:2256–2266.

7. Dignam JD, Lebovitz RM, Roeder RG. Accurate transcription ini-tiation by RNA polymerase II in a soluble extract from isolatedmammalian nuclei. Nucleic Acids Res 1983;11:1475–1489.

8. Bennett BL, Sasaki DT, Murray BW, O’Leary EC, Sakata ST, Xu W,Leisten JC, Motiwala A, Pierce S, Satoh Y, Bhagwat SS, ManningAM, Anderson DW. SP600125, an anthrapyrazolone inhibitor ofJun N-terminal kinase. Proc Natl Acad Sci U S A 2001;98:13681–13686.

9. Ochiai A, Yasui W, Kameda T, Takanashi A, Takekura N, TaharaE. The effect of phorbol esters on cell growth and epidermalgrowth factor receptor modulation in a human gastric carcinomacell line TMK-1. Hiroshima J Med Sci 1989;38:191–195.

0. Kanamori M, Tajima M, Satoh Y, Hoshikawa Y, Miyazawa Y,Okinaga K, Kurata T, Sairenji T. Differential effect of TPA on cellgrowth and Epstein-Barr virus reactivation in epithelial cell linesderived from gastric tissues and B cell line Raji. Virus Genes2000;20:117–125.

1. Richter M, Weiss M, Weinberger I, Furstenberger G, Marian B.Growth inhibition and induction of apoptosis in colorectal tumorcells by cyclooxygenase inhibitors. Carcinogenesis 2001;22:17–25.

2. Turini ME, DuBois RN. Cyclooxygenase-2: a therapeutic target.Annu Rev Med 2002;53:35–57.

3. Potapova O, Haghighi A, Bost F, Liu C, Birrer MJ, Gjerset R,Mercola D. The Jun kinase/stress-activated protein kinase path-way functions to regulate DNA repair and inhibition of the path-way sensitizes tumor cells to cisplatin. J Biol Chem 1997;272:14041–14044.

4. Karin M, Liu Z, Zandi E. AP-1 function and regulation. Curr OpinCell Biol 1997;9:240–246.

5. 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 inhibitsazoxymethane-induced colon carcinogenesis in rat without reduc-ing prostaglandin levels. Cancer Res 1995;57:2909–2915.

6. Simonson MS, Herman WH, Dunn MJ. PGE2 induces c-fos ex-pression by a cAMP-independent mechanism in glomerular mes-angial cells. Exp Cell Res 1995;215:137–144.

7. Won JS, Suh HW, Kim YH, Song DK, Huh SO, Lee JK, Lee KJ.Prostaglandins E2 increases proenkephalin mRNA level in ratastrocyte-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., Depart-ent of Medicine, University of Hong Kong, Queen Mary Hospital, Hongong. e-mail: [email protected]; fax: (852) 2872-5828.Supported by Research Grant Council earmarked grant HKU 7309/

1M of the Hong Kong Special Administration Region, the Simon KYee Gastroenterology Research Fund, Queen Mary Hospital and Gas-
roenterological Research Fund, University of Hong Kong.

Documents

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