Research Article Cancer Prevention Research Colon Tumor ...€¦ · Research Article Colon Tumor Cell Growth–Inhibitory Activity of Sulindac Sulfide and Other Nonsteroidal Anti-Inflammatory

Rese

ColSulfIs A

HeathYulia

Abst

Intro

Nocycloopromrectalshownreducpopulprescrtraditizymesfromprosta

AuthorUniverDevelopand 3De

Corres2000 9Fax: 20

doi: 10

©2010

www.a

Do

Published OnlineFirst September 28, 2010; DOI: 10.1158/1940-6207.CAPR-10-0030

Cancer
eventionesearch
arch Article

on Tumor Cell Growth–Inhibitory Activity of Sulindacide and Other Nonsteroidal Anti-Inflammatory Drugs

PrR

ssociated with Phosphodiesterase 5 Inhibition

er N. Tinsley1, Bernard D. Gary2, Jose Thaiparambil2, Nan Li3, Wenyan Lu2, Yonghe Li1,2,
Y. Maxuitenko2, Adam B. Keeton2, and Gary A. Piazza1,2
ractNon

resultisuggewell dmonothat inhibitegrowtinhibiisozymevaluaand Mity toptosisdisplacolon

arachidoglandins

s' Affiliatiosity of Alament Divisipartment o

ponding Ath Avenue S5-581-2093

.1158/1940-

American A

acrjourna

wnloade

steroidal anti-inflammatory drugs (NSAID) display promising antineoplastic activity, but toxicityng from cyclooxygenase (COX) inhibition limits their clinical use for chemoprevention. Studiesst that the mechanism may be COX independent, although alternative targets have not beenefined. Here, we show that the NSAID sulindac sulfide (SS) inhibits cyclic guanosine 3′,5′-phosphate (cGMP) phosphodiesterase (PDE) activity in colon tumor cell lysates at concentrationshibit colon tumor cell growth in vitro and in vivo. A series of chemically diverse NSAIDs also in-d cGMP hydrolysis at concentrations that correlate with their potency to inhibit colon tumor cellh, whereas no correlation was observed with COX-2 inhibition. Consistent with its selectivity forting cGMP hydrolysis compared with cyclic AMP hydrolysis, SS inhibited the cGMP-specific PDE5e and increased cGMP levels in colon tumor cells. Of numerous PDE isozyme–specific inhibitorsted, only the PDE5-selective inhibitor MY5445 inhibited colon tumor cell growth. The effects of SSY5445 on cell growth were associated with inhibition of β-catenin–mediated transcriptional activ-suppress the synthesis of cyclin D and survivin, which regulate tumor cell proliferation and apo-, respectively. SS had minimal effects on cGMP PDE activity in normal colonocytes, whichyed reduced sensitivity to SS and did not express PDE5. PDE5 was found to be overexpressed intumor cell lines as well as in colon adenomas and adenocarcinomas compared with normal co-mucosa. These results suggest that PDE5 inhibition, cGMP elevation, and inhibition of β-catenin
lonic
transcriptional activity may contribute to the chemopreventive properties of certain NSAIDs. Cancer PrevRes; 3(10); 1303–13. ©2010 AACR.

side eand inplay rwith itiallyor COtend tod ofThe

chemto COsugge

duction

nsteroidal anti-inflammatory drugs (NSAID) andxygenase-2 (COX-2)–selective inhibitors display

ising chemopreventive efficacy, especially for colo-cancer. Epidemiologic studies, for example, havethat long-term use of NSAIDs is associated with

ed incidence of colorectal cancer in the generalation by about 30% to 50% (1, 2). Commonlyibed for the treatment of inflammatory conditions,onal NSAIDs inhibit both COX-1 and COX-2 en-, which catalyze the production of prostaglandins

nic acid. Unfortunately, the depletion ofby NSAIDs is associated with significant

tributactivitnon–sulindinduchibitsmodetaglansulindin pat(11) b

ns: 1Department of Pharmacology and Toxicology,bama at Birmingham and 2Drug Discovery andon, Southern Research Institute, Birmingham, Alabama,f Biochemistry, University of Alabama at Birmingham

uthor: Gary A. Piazza, Southern Research Institute,outh, Birmingham, AL 35205. Phone: 205-581-2731;; E-mail: [email protected].

6207.CAPR-10-0030

ssociation for Cancer Research.

ls.org

for Cancer Researchon Acancerpreventionresearch.aacrjournals.org d from

ffects, including gastrointestinal ulcers, bleeding,testinal perforation. COX-2–selective inhibitors dis-educed gastrointestinal toxicity but are associatedncreased risk of myocardial infarction. These poten-fatal adverse effects outweigh the benefits of NSAIDsX-2 inhibitors for preventing colorectal cancer, whicho require high dosages administered over a long peri-time.mechanism of action that is responsible for the

opreventive activity of NSAIDs is widely attributedX-2 inhibition, although numerous reports havested that a COX-independent mechanism may con-e to or be fully responsible for their antineoplasticy (3–8). For example, studies have shown that theCOX-inhibitory sulfone metabolite of the NSAIDac can inhibit human colon tumor cell growth ande apoptosis in vitro (7, 8). Sulindac sulfone also in-tumorigenesis in the azoxymethane-induced rat

l of colon carcinogenesis without suppressing pros-din levels in colonic mucosa (7–9). In clinical trials,ac sulfone (exisulind) caused regression of adenomas
ients with either familial (10) or sporadic polyposisut did not receive Food and Drug Administration
1303

. ugust 16, 2021. © 2010 American Association

http://cancerpreventionresearch.aacrjournals.org/

approshownanalosphatcentrainhib(12–1for itsidentito thePDE

drolyzcleoticyclicsengerinhibidationThe Pinto 1withinconsein thifamilcGMP(PDE1Bothliferat16), ainhibiinhibicancerSul

NSAIDand ithas bbeensize apolypa prodinflamreducinhibsulfonof eitAlthotumorsulfonPDE ireportPDE5activigrowtstudieof PDHere,SS andcell grSS is

inhibin col

Mate

DrugMY

RecomBPS Bfromperoxtaineusedwisechase

CellsThe

HCT1colonCultucondiCollectrypantures d

GrowTiss

a dengrowtof tre(Promtent. Tspecif

ProlifThe

by a ra radiment,timesciencfor anPlatesontoan auamou

ApopCel

densitcells wfor 15breakThe aificati

Tinsley et al.

Cance1304

Do


val due to hepatotoxicity. Mechanistic studies havethat sulindac sulfone and several structurally related

gues can inhibit cyclic guanosine 3′,5′-monopho-e (cGMP) phosphodiesterase (PDE) activity at con-tions that are comparable with those required toit colon tumor cell growth and induce apoptosis4). However, the specific PDE isozyme(s) responsibletumor cell growth–inhibitory activity has not beenfied, nor has this effect been well studied with regardchemopreventive efficacy of conventional NSAIDs.is a metallophosphohydrolase that specifically hy-es the 3′,5′-cyclic phosphate moiety on cyclic nu-des to a 5′-monophosphate, thereby deactivatingAMP (cAMP) or cGMP. PDE terminates second mes-signaling by degrading cyclic nucleotides, whereastion of PDE activity blocks cyclic nucleotide degra-to mimic or amplify cyclic nucleotide signaling.

DE superfamily consists of 20 distinct genes divided1 protein families (15). Despite the heterogeneitythe PDE superfamily, the catalytic domain is highly

rved across family members, yet minor changes with-s domain determine the specificity of each isozymey member for cAMP (PDE4, PDE7, and PDE8),(PDE5, PDE6, and PDE9), or both cAMP and cGMP, PDE2, PDE3, PDE10, and PDE11) as substrates.

cAMP and cGMP have been shown to have antipro-ive and proapoptotic effects in many cell types (15,nd numerous studies have described the growth-tory and apoptosis-inducing effects of various PDEtors on certain tumor cell types, including colorectalcells (17–20).

indac is perhaps the most effective among thes with regard to colon cancer chemoprevention,

s efficacy in numerous preclinical and clinical studieseen well documented. Most notably, sulindac hasshown to cause 60% to 70% reduction of adenomand number in patients with familial adenomatousosis (21). As reviewed previously (22), sulindac isrug that requires metabolic activation for its anti-matory activity. The sulfoxide moiety is reversiblyed by colonic bacteria to form the active COX-itory sulfide metabolite. The non–COX-inhibitorye metabolite is irreversibly generated by oxidationher the sulfide or the sulfoxide form in the liver.ugh both metabolites have been shown to inhibitcell growth and induce apoptosis (7, 8), only thee metabolite has been studied with regard to cGMPnhibition in colon tumor cells (12). Recently, weed that sulindac sulfide (SS) can selectively inhibitactivity in human breast tumor cells and that thisty is closely associated with its ability to inhibith and induce apoptosis (23). Additionally, others using PDE5 antisense have shown the importanceE5 for colon tumor cell growth and survival (20, 24).we characterize the cGMP PDE–inhibitory activity ofother NSAIDs and provide evidence that the tumor

owth–inhibitory and apoptosis-inducing activity ofthe result of cGMP elevation caused by selective

transfpositi

r Prev Res; 3(10) October 2010

for Cancer Researchon Acancerpreventionresearch.aacrjournals.org wnloaded from

ition of the PDE5 isozyme, which is overexpressedon tumor cells compared with normal colonocytes.

rials and Methods

s and reagents5445 and NOR-3 were purchased from BioMol.binant PDE enzymes were purchased fromiosciences. Anti-PDE antibodies were purchasedGeneTex. Anti-rabbit and anti-goat horseradishidase–conjugated secondary antibodies were ob-d from Cell Signaling Technology. DMSO wasas the vehicle for all compounds unless other-specified. All other drugs and reagents were pur-d from Sigma-Aldrich.

and cell culturehuman colon cancer cell lines HT-29, SW480, and16 and the primary culture of normal fetal humanocytes (FHC) were obtained from the American Typere Collection and grown under standard cell culturetions as recommended by the American Type Culturetion. Cell counts and viability were determined byblue exclusion using a hemocytometer. Only cul-isplaying >95% viability were used for experiments.

th assaysue culture microtiter 96-well plates were seeded atsity of 5,000 cells per well. The inhibition of cellh caused by treatment was determined after 72 hoursatment using the luminescent CellTiter-Glo Assayega), which measures viable cells based on ATP con-he assay was done according to the manufacturer'sications.

eration assaysantiproliferative activity of NSAIDs was determinededuction of DNA synthesis, which was measured byoactive thymidine incorporation assay. After treat-cells were incubated at 37°C for 56 hours, at whicha 1 μCi aliquot of [3H]thymidine (Amersham Bio-es) was added to each well. Cells were incubatedother 16 hours for a total of 72 hours of treatment.were harvested with a semiautomatic cell harvesterfilters that bind the DNA. Samples were counted intomated scintillation counter to determine thent of [3H]thymidine uptake.

tosis assayls were seeded in 10-cm tissue culture dishes at ay of 1 × 106 per dish. After 48 hours of treatment,ere collected and fixed with 4% formalin on iceminutes. Samples were stained for DNA strand

s using the APO-BrdU TUNEL Assay kit (Invitrogen).ssay was done according to the manufacturer's spec-ons. The percentage of terminal deoxynucleotidyl
erase–mediated dUTP nick end labeling (TUNEL)–ve cells was quantified using a Guava EasyCyte Plus
Cancer Prevention Research



flow ced forcompsoftwa

Cell lCel

[20 m1 mmand pby centein cproteiificatiscrapecolon

TissuHT-

mice.in 101 mmand awas cat 4°Coninispecif

PDE acGM

tillaticondi7.5),[3H]cGin a vceed fscintilproduer. PDcenceassayfluoresubstrcAMP

cGMPCel

sue cuwere lDirectmanu

WesteCel

in a 1transfwere btainin

marymanu

LucifeCel

lucifeexprescells wactivitmaliz

TolerThe

determSulind100,carbowatermeasu

In vivThe

the hu5- towas a50 mg80 incontrgroupthe tuservedweighday omicewateras pe(IACUby So

ImmuFor

wereDr. AchemantisesynthaffiniPDE5Peptifrom

ExperDru

suredwhichwhen

Sulindac Inhibits PDE5

www.a

Do


ytometer. A minimum of 10,000 events was collect-each treatment group with use of minimal electronicensation. Data were analyzed using CytoSoft 5.0re (Guava Technologies).

ysisls were harvested, vortexed in ice-cold lysis buffermol/L Tris-acetate, 5 mmol/L magnesium acetate,ol/L EGTA, 0.8% Triton X-100, 50 mmol/L NaF,rotease inhibitor cocktail (pH 7.4)], and clarifiedtrifugation at 10,000 × g for 10 minutes at 4°C. Pro-ontent was determined using the bicinchoninic acidn assay (Pierce) following the manufacturer's spec-ons. The colon of normal rats was excised andd with a glass microscope slide to obtain normalocytes, which were then lysed as described.

e homogenization29 tumor cells were grown s.c. in athymic NCr-nu/nuTumors weighing 1 to 2 g each were homogenized0 mmol/L Tris-HCl (pH 7.5), 10 mmol/L NaCl,ol/L EGTA, 1 mmol/L DTT, 0.1% Triton X-100,protease inhibitor cocktail (Sigma). The extract

larified by centrifugation at 20,000 × g for 1 hour. Protein content was determined using the bicinch-c acid protein assay following the manufacturer'sications.

ssaysP PDE activity was measured using [3H]cGMP scin-

on proximity enzyme assay kits (Amersham). Thetions of the assays were 50 mmol/L Tris-HCl (pH8.3 mmol/L MgCl2, 1.7 mmol/L EGTA, 20 μmol/LMP, 2% DMSO, and 0.2 mg/mL of HT-29 lysate

olume of 100 μL. The reaction was allowed to pro-or 15 minutes at 30°C and terminated by addinglation proximity beads, which bind the radiolabeledct. The beads were counted on a scintillation count-E activity was also measured using the IMAP fluores-polarization PDE assay (Molecular Devices). Thewas modified, as described previously (23), to usescein-cAMP and tetramethylrhodamine-cGMP asates, allowing for simultaneous measurement ofand cGMP hydrolysis.

assayls were seeded at a density of 1 × 106 per 10-cm tis-lture dish. After 30 minutes of drug treatment, cellsysed and assayed for cGMP content using the cGMPBiotrak EIA kit (GE Biosciences) according to thefacturer's specifications.

rn blotsl lysates (30 μg protein) were separated by SDS-PAGE2% polyacrylamide gel followed by electrophoreticer to a nitrocellulose membrane. The membranes
locked with 5% bovine serum albumin in TBS con-g 0.05% Tween 20. Incubation of membranes in pri-
says, teight

acrjournals.org


and secondary antibodies was done according to thefacturers' specifications.

rase reporter assaysls were transiently transfected with the TOPFlashrase construct (MilliPROBE) and β-galactosidase–sing vector (Promega). After 24 hours of treatment,ere lysed and both luciferase and β-galactosidaseies were measured. The luciferase activity was nor-ed to the β-galactosidase activity.

ance studiesmaximum tolerated dose (MTD) of sulindac wasined using 5- to 6-week-old male NCr-nu/nu mice.ac was administered by oral gavage at dosages of 50,200, or 400 mg/kg/d as a suspension in 0.5%xymethylcellulose (CMC) and 0.25% Tween 80 in. Mice were observed daily and body weights werered twice weekly.

o antitumor efficacyantitumor efficacy of sulindac was evaluated usingman HT-29 colon tumor xenograft implanted s.c. in6-week-old male NCr-nu/nu athymic mice. Sulindacdministered orally once daily for 40 days at a dose of/kg as a suspension in 0.5% CMC and 0.25% Tweenwater. The experiment consisted of a vehicle-treatedol group and a sulindac-treated group, with eachcontaining 10 mice. Treatment was initiated whenmors reached a size of 75 to 200 mg. Mice were ob-daily for mortality. Tumor dimensions and body

ts were measured twice weekly starting on the firstf treatment. For all animal studies described above,were housed in microisolator cages, provided tapand sterile pelleted diet ad libitum, and euthanizedr Institutional Animal Care and Use CommitteeC) guidelines. All animal protocols were revieweduthern Research IACUC before experimentation.

nohistochemistrymalin-fixed, paraffin-embedded clinical specimensobtained from the Fox Chase Cancer Center byndre Klein-Szanto and processed for immunohisto-istry by avidin-biotin peroxidase detection. PDE5rum was generated in sheep against a 16–amino acidetic peptide sequence contained within the high-ty cGMP binding domain of the enzyme. Anti-IgG was purified using an antigen affinity column.de, antibody, and affinity column were obtainedBethyl Laboratories.

imental design and data analysisg effects on cell growth and PDE activity were mea-, and the potency was expressed as an IC50 value,is the concentration resulting in 50% inhibitioncompared with the vehicle control. For growth as-
he IC50 value was determined by testing a range ofconcentrations with a minimum of four replicates
Cancer Prev Res; 3(10) October 2010 1305



per doDose-warefour-ppeatedity ofbetweand vSEM.the spusing

Resu

NSAIDcorreTo d

COX-activitsulfonand Ctive Cinhibtheirenzymcoloncells. Aoxidecells wdespitconceited psulfoxHowetionsCOX-of sulthat winhib

colonpoteninhibispecifcolonresultnot thassocicy amA se

the acNSAIDecoxibPDE,thesewas obitionwith awas ofor Cpotensalicycarbotazonlow paffectshowhibitwith t

SS setumoSS w

drolysHT-29FHCscell li

Tabsellysa

Com

SSSuliSuliRofeIndo

NOTassaAnti

Tinsley et al.

Cance1306

Do


se. Each assay used a minimum of two replicates.response curves were constructed using Prism5 soft-(GraphPad), which calculates IC50 values using aarameter logistic equation. All experiments were re-a minimum of twice to determine the reproducibil-the results. All values represent a comparisonen drug treatment at the specified concentrationehicle-treated controls. All error bars representCalculation of P values was done by comparingecified treatment group to vehicle-treated controlsa Student's t test.

lts

inhibition of colon tumor cell growthlates with cGMP PDE inhibitionetermine whether cGMP PDE represents a potential

independent target responsible for the antineoplasticy of sulindac, we compared the non–COX-inhibitorye and sulfoxide forms of the drug with the COX-1–OX-2–inhibitory sulfide metabolite, the highly selec-OX-2 inhibitor rofecoxib, and the COX-1–selectiveitor indomethacin. These drugs were evaluated forpotency to inhibit purified COX-1 and COX-2es, cGMP hydrolysis in lysates from HT-29 humantumor cells, and proliferation of HT-29 colon tumors summarized in Table 1, sulindac sulfone and sulf-inhibited the proliferation of HT-29 colon tumorith IC50 values of 89 and 224 μmol/L, respectively,e lacking COX-1– or COX-2–inhibitory activity atntrations up to 300 μmol/L. SS more potently inhib-roliferation of the colon tumor cells than either theide or sulfone forms with an IC50 value of 34 μmol/L.ver, this effect required appreciably higher concentra-compared with those needed to inhibit COX-1 or2. In contrast with COX inhibition, all three formsindac inhibited cGMP PDE activity with IC50 values
ere comparable with their respective IC50 values foriti
growt

leec omethacin tote mor cell proli

p

COX-1 COX-2

1.8 6.3nd >300 >300

proliferative activity was measured after 72 h of drug treatment of HT-2



tumor cell growth and cGMP PDE but with reducedcy compared with SS despite its superior potency tot COX-1 or COX-2. Moreover, the highly potent andic COX-2 inhibitor rofecoxib was unable to inhibittumor cell proliferation or cGMP PDE activity. Theses show that the potency for cGMP PDE inhibition, bute potency for COX-1 or COX-2 inhibition, is closelyated with colon tumor cell growth–inhibitory poten-ong sulindac and its metabolites.ries of chemically diverse NSAIDs representative ofetic acid, fenamate, and proprionic acid families ofs, as well as another COX-2–selective inhibitor, cel-, were compared for potency to inhibit COX-2, cGMPand HT-29 colon tumor cell growth (Fig. 1A). Amonganti-inflammatory drugs, a strong positive correlationbserved between potency for HT-29 cell growth inhi-and potency for inhibition of cGMP PDE activityn r2 value of 0.933 (Fig. 1B). However, no correlationbserved by comparing potencies of these NSAIDsOX-2 inhibition as previously reported (25) withcies for HT-29 cell growth inhibition (Fig. 1C). Thelate family, which contains aspirin as well as non-xylate NSAIDs from the pyrazole (e.g., phenylbu-e), and oxicam (e.g., piroxicam) families displayedotency for HT-29 growth inhibition and did notcGMP PDE activity (data not shown). These datathat a variety of chemically diverse NSAIDs can in-cGMP PDE activity and that this effect is associatedheir in vitro growth-inhibitory potency.

lectively inhibits cGMP hydrolysis in colonr cellsas further studied for its ability to inhibit cGMP hy-is and growth of the human colon cancer cell lines, SW480, and HCT116 and were compared with, which are representative of a nontumorigenic colonne. As shown in Fig. 2A, SS selectively inhibited
h of the human colon tumor cell lines with IC50
ng proliferation. Indomethacin also inhibited values ranging from 73 to 85 μmol/L compared with the

1. Potency of the sulindac metabolites, the COX-2–selective inhibitor rofecoxib, and the COX-1–
tive inhibitor inds, and colon tu
inhibit COX-1, COX-2, cGMP hydrolysis in HT-29 colon tumor cellferation

C50 (μmol/L)

cGMP

1995

9 cells using a trit

. ugust 16, 2021

Prolife

3389

iated thymidine assay.

Cancer Prevention R

. © 2010 American Associa

ound
I
PDE
ration
.7
.9 ac sulfone .9 .4 ac sulfoxide .7 24 nd >300 >300 119 2
coxib >300 2.7 >300 >500methacin 0.017 1.0 24 136

E: Purified ovine COX-1 and COX-2 were used for the COX assays. cGMP hydrolysis was measured by the scintillation proximityy using cell lysates obtained from HT-29 human colon tumor cells, which were used as the source of cGMP PDE isozymes.

esearch

tion


IC50 vof normor cby DNment,increapositiUsi

opedlysis, tcolon

inhib74 μmconceAdditcGMPreducTo

SS inintact

Fig. 1.association between the potencies of HT-29 growth inhibition and inhibition of cGMP hydrolysis in HT-29 cell lysates. C, no association between thepotenc


www.a

Do


alue of 150 μmol/L necessary for growth inhibitionmal colonocytes. In addition to inhibiting colon tu-ell growth, SS also induced apoptosis as measuredA fragmentation (Fig. 2B). After 48 hours of treat-100 and 200 μmol/L SS resulted in 6- and 15-foldses, respectively, in the percentage of TUNEL-ve HT-29 cells.ng a fluorescence polarization assay that we devel-to simultaneously measure cAMP and cGMP hydro-

ies for HT-29 growth inhibition and inhibition of COX-2.

he effects of SS on PDE activity in lysates from HT-29cancer cells and FHC were evaluated (Fig. 2C). SS

ter 30say (F

acrjournals.org


ited cGMP hydrolysis with an IC50 value ofol/L in the HT-29 cell lysate, but appreciably higherntrations were necessary to inhibit cAMP hydrolysis.ionally, FHC displayed reduced sensitivity to thePDE–inhibitory activity of SS, which paralleled its

ed sensitivity to growth suppression by SS.determine the relevance of cGMP PDE inhibition bycolon tumor cell lysates to the effects of the drug incells, intracellular levels of cGMP were measured af-

Colon tumor cell growth, cGMP, and COX-2 inhibitory activity of a group of NSAIDs and COX-2 inhibitors. A, structures of compounds. B, positive

minutes of SS treatment using a cGMP immunoas-ig. 2D). SS treatment of HT-29 and SW480 colon




tumorcGMPwhichcolon

cGMPgrowtTo

colongrowtcyclasthe Hgrowt59 toactivaany o(Fig. 3eventsA p

vinporoliprfor grHCT1MY54

with IMY54HT-29substato SS,tosis ohighltadalaHoweconcebe atttumo

SS seTo

mor cisozyscreensubstPDE5drolysreduc

Fig. 2.B, apopactivitycells af

Tinsley et al.

Cance1308

Do


cells resulted in a significant increase in intracellularlevels at concentrations of 100 and 150 μmol/L,is consistent with concentrations required to inhibittumor cell growth and induce apoptosis.

elevation causes colon tumor cellh inhibitiondetermine if elevated levels of intracellular cGMP intumor cells are sufficient to inhibit colon tumor cellh, the effect of the nitric oxide donor and guanylyle activator NOR-3 on cell growth was evaluated inT-29, SW480, and HCT116 lines. NOR-3 inhibitedh in all three cell lines with IC50 values ranging from74 μmol/L (Fig. 3A). Conversely, the adenylyl cyclasetor forskolin did not significantly affect growth off the cell lines at concentrations up to 100 μmol/LB), suggesting that cGMP, but not cAMP, can signalleading to the suppression of colon tumor cell growth.anel of PDE isozyme–specific inhibitors, includingcetine (PDE1), EHNA (PDE2), milrinone (PDE3),am (PDE4), and MY5445 (PDE5), was evaluatedowth-inhibitory activity in HT-29, SW480, and
16 colon tumor cell lines. Among these, only45 was found to inhibit colon tumor cell growth
shownof PD

of SS in HT-29 cell lysates (solid lines) compared with FHC lysates (dashed lineter 30 min of SS treatment. *, P < 0.05 versus vehicle control.



C50 values ranging from 8 to 18 μmol/L (Fig. 3C).45 also selectively inhibited cGMP PDE activity incell lysates with an IC50 value of 4 μmol/L withoutntially affecting cAMP hydrolysis (Fig. 3D). SimilarMY5445 inhibited proliferation and induced apop-f colon tumor cells (data not shown). In addition,y potent and selective PDE5 inhibitors such asfil and vardenafil suppressed tumor cell growth.ver, high concentrations were required relative tontrations that inhibited purified PDE5, which mayributed to low uptake or high efflux from culturedr cells.

lectively inhibits PDE5determine which PDE isozyme is responsible for tu-ell growth–inhibitory activity of SS, recombinantmes from all 11 isozyme family members wereed for sensitivity to SS using cGMP or cAMP as arate. As previously reported, the cGMP-selectiveisozyme was the most sensitive, although cGMP hy-is by PDE2 and PDE3 was also inhibited, albeit withed sensitivity (23). The difference in sensitivity is
in Fig. 4A with concentration-dependent inhibition
E5 by SS, resulting in an IC50 value of 38 μmol/L. By

Colon tumor cell growth– and cGMP PDE–inhibitory activity of SS. A, dose-dependent growth-inhibitory activity of SS after 72 h of treatment.tosis induction of HT-29 cells by SS after 48 h of treatment. C, dose-dependent cAMP PDE–inhibitory (▴and○) and cGMP PDE–inhibitory (• and⧫)

s). D, increased intracellular cGMP levels in HT-29 and SW480




compPDE3Consicantlynot sPDE2cells,studiebitorscells wwith tin Figdid nPDE3pressicell li(dataisozymcolonthe Pinhibidenafwithting, mity inwherebitiontumorselectiothercell gcolon

ColonovereThe

enomimensPDE5ing nobelingmusc11 ofRepreadeno

TumoassoctransPre

fone cwhichproteaof SScells tgrowtpoundlymphtion aHCT1the lebetweMY54

Fig. 3.elevatinof humA, grownitric ocyclase72 h ofgrowthadenylyforskoliC, growPDE5-sD, cAMcGMP •of MY5445 in HT-29 cell lysates.


www.a

Do


arison, SS inhibited cGMP hydrolysis by PDE2 andwith IC50 values of 97 and 84 μmol/L, respectively.stent with its cGMP selectivity, SS did not signifi-inhibit cAMP hydrolysis by PDE2 or PDE3 (data

hown). To assess the potential contribution of, PDE3, and PDE5 to cGMP degradation in wholetheir activity and expression in colon cell lines wasd using PDE isozyme–specific antibodies and inhi-. First, the levels of protein expression in HT-29ere determined by Western blotting and comparedhe levels of protein expression in FHC. As shown. 4B, HT-29 cells expressed PDE3 and PDE5, butot express PDE2. Conversely, FHC expressedat levels higher than HT-29 cells, but lacked ex-

on of PDE2 and PDE5. Other human colon cancernes, HCT116 and SW480, also expressed PDE5not shown). To determine the importance of thesees to the overall cGMP hydrolytic capacity of thetumor cells, the cGMP PDE–inhibitory activity ofDE2-selective inhibitor EHNA, the PDE3-selectivetor milrinone, and the PDE5-selective inhibitor sil-il were analyzed in HT-29 cell lysates. Consistentthe expression levels determined by Western blot-ilrinone and sildenafil inhibited cGMP PDE activ-the HT-29 lysate by 31% and 42%, respectively,as EHNA had only a modest effect of <12% inhi-(Fig. 4C). The overexpression of PDE5 in coloncells compared with FHC is consistent with the

vity of SS for inhibition of PDE5 compared withcGMP PDE isozymes and for inhibition of tumor
rowth and cGMP PDE compared with normalocytes.
exprestransc

acrjournals.org


adenomas and adenocarcinomasxpress PDE5expression of PDE5 in normal colonic mucosa, ad-as, and adenocarcinomas from archival clinical spec-was determined by immunohistochemistry using a-specific antibody. Analysis of 11 specimens contain-rmal colonic mucosa showed consistent negative la-in the colonic epithelium, despite labeling of the

ularis mucosa. By contrast, 5 of 5 adenomas and13 adenocarcinomas displayed positive labeling.sentative micrographs from normal colon mucosa,mas, and adenocarcinomas are shown in Fig. 4D.

r cell growth–inhibitory activity of SS isiated with inhibition of β-catenincriptional activityvious studies reported that sulindac sulfide and sul-an activate cGMP-dependent protein kinase (PKG),can then phosphorylate β-catenin to promote itssomal degradation (12, 23). We evaluated the effectsand MY5445 on β-catenin signaling in colon cancero determine if this mechanism is involved in theh-inhibitory activity of these compounds. Both com-s resulted in a significant reduction of T-cell factor/oid enhancer factor (Tcf/Lef)–mediated transcrip-s measured by a luciferase reporter assay in HT-29,16, and SW480 colon cancer cells (Fig. 4E). Althoughngth of treatment necessary to cause an effect differeden compounds and between proteins, SS and45 treatment resulted in a significant reduction in

Effects of cyclic nucleotide–g agents on growthan colon cancer cells.th-inhibitory activity of thexide donor and guanylylactivator NOR-3 aftertreatment. B, lack of-inhibitory activity of thel cyclase activatorn after 72 h of treatment.th-inhibitory activity of theelective inhibitor MY5445.P PDE–inhibitory (▪) andPDE–inhibitory ( ) activity

sion of the proteins cyclin D1 and survivin, whoseription is promoted by β-catenin signaling (Fig. 4F).




Fig. 4.of PDE2of PDEthe PDadenocSS (top(blue coafter tre

Tinsley et al.

Cance1310

Do


PDE isozyme expression and activity. A, inhibition of cGMP hydrolytic activity of recombinant PDE2, PDE3, and PDE5 by SS. B, relative levels, PDE3, and PDE5 isozyme expression in HT-29 and FHC as measured by Western blotting with PDE family–specific antibodies. C, relative activities2, PDE3, and PDE5 families in HT-29 cell lysates as measured by overall inhibition of cGMP hydrolysis by the PDE2-selective inhibitor EHNA,E3-selective inhibitor milrinone, and the PDE5-selective inhibitor sildenafil. D, PDE5 expression in normal colon mucosa, adenoma, andarcinomas from human clinical specimens as determined by immunohistochemistry following avidin-biotin peroxidase labeling. E, effects ofgraph) or MY5445 (bottom graph) on Tcf/Lef transcriptional activity as measured by a luciferase reporter assay in HT-29 (black columns), HCT116lumns), and SW480 (red columns) colon cancer cells. F, time-dependent decrease in the expression of cyclin D1 and survivin in SW480 cells
atment with SS or MY5445, with β-actin as a loading control.
r Prev Res; 3(10) October 2010 Cancer Prevention Research

for Cancer Research. on August 16, 2021. © 2010 American Associationcancerpreventionresearch.aacrjournals.org wnloaded from


Thesephorysupprtumor

SS secolonTo

suredHT-29MY54fromculturly inhHT-29mucoturedtencywith tdisplahibitecompin culcell grcreasecells c

SulinTo

above

of sulxenoged bafor 20∼50weighsulindwith fwith vdac atment∼55%with v(datametabfollowdac gthe su5.5 μshowand thefficacble wtumo

Discu

Epi

Fig. 5.sulindainhibitiohumanhomogwith hocolonicB, seleinhibitiocolon tline) coof normline). CsulindaD, inhibtumor gcompadosage of 50 mg/kg/d.


www.a

Do


data are consistent with cGMP-PKG–mediated phos-lation of β-catenin to induce degradation, therebyessing the transcription of key genes that regulatecell proliferation and apoptosis.

lectively inhibits cGMP hydrolysis intumors

further evaluate the tumor selectivity of SS, we mea-the sensitivity of cGMP PDE activity in lysates fromtumors established in athymic mice to SS and

45 and compared the sensitivity to lysates obtainednormal rat colonic mucosa. Similar to the effects ined HT-29 tumor cells, both SS and MY5445 selective-ibited cGMP PDE activity in lysates obtained from thetumors. However, cGMP hydrolysis in normal colonsa displayed reduced sensitivity, comparable with cul-FHC (Fig. 5A and B). SS displayed 4-fold greater po-for inhibition of the tumor cGMP PDE comparedhat present in the normal tissue, whereas MY5445yed >10-fold greater potency. Both compounds in-d cGMP PDE in the tumor lysate with IC50 valuesarable with those necessary for cGMP PDE inhibitiontured tumor cell lysates and for in vitro colon tumorowth inhibition. These results are consistent with in-d expression and activity of PDE5 in colon tumorompared with normal colonocytes.

dac inhibits colon tumor cell growth in vivo
assess the in vivo pharmacologic relevance of theobservations, the tumor growth-inhibitory activity
showinhibi

acrjournals.org


indac was determined using the HT-29 colon tumorraft mouse model. Dosages of sulindac were select-sed on a tolerance study following daily oral gavagedays of treatment. The MTD was determined to be

mg/kg, where no deaths (Fig. 5A) or significantt loss occurred. To determine antitumor efficacy ofac, two groups of 10 mice each were inoculated s.c.ragments of HT-29 tumors; one group was treatedehicle (CMC) and the other was treated with sulin-50 mg/kg once daily by oral gavage. Sulindac treat-caused a significant reduction of tumor growth ofby the end of treatment compared with mice treatedehicle only (Fig. 5D) without affecting body weightnot shown). In addition, levels of sulindac and itsolites were measured in plasma collected 6 hoursing a single oral dose of sulindac at 50 mg/kg. Sulin-enerated plasma levels of 36.7 ± 10.2 μmol/L forlfide, 8.9 ± 1.2 μmol/L for the sulfone, and 23.1 ±mol/L for the parent sulfoxide. These experimentsthat SS is the predominant metabolite in plasmaat dosages of sulindac required for in vivo antitumory can generate plasma levels of SS that are compara-ith the concentrations required to inhibit colonr cell proliferation and PDE5 activity in vitro.

ssion

demiologic, preclinical, and clinical studies have

In vivo antitumor efficacy ofc. A, selectivity of SS forn of cGMP hydrolysis inHT-29 colon tumorenates (solid line) comparedmogenates of normal ratmucosa (dashed line).ctivity of MY5445 forn of cGMP hydrolysis inumor homogenates (solidmpared with homogenatesal colonic mucosa (dashed, MTD determination ofc in male athymic mice.ition of human HT-29 colonrowth in mice treated withred with sulindac at a

n convincing evidence that NSAIDs and COX-2tors have cancer chemopreventive efficacy, especially




with rtestinfromcancercolonNSAIDFor inmor chibitiothan ainhibioxidegrowtcorrelcally dtumoapparselecticell grothergrowtdata sindepcell gCOX-the dchemoBec

growtulatin(s) reover,inhibwhichnecesslysatein HTmunocGMPhibitecGMPactivMY54cGMPcell gshoweby cominhibithat isesis. TcGMPtranscin vitinducAlth

inhibiof thisto dev

of roletentiaselectipresseto seleNontuPDE5The climmumens,mas amucoselectiparedfor boPDE alon cethat ucells tand suAlth

colonrabletent Ptadalahibitrequirtial exity oflimitezymecancesentecGMPgrowtand pWe

ly inhomoally, ScGMPPDE,more,greatetivityhibiteobservthat sdosagpablenecessmentstion oan imsulindpossi

Tinsley et al.

Cance1312

Do


egard to colorectal cancer. Unfortunately, gastroin-al, renal, and cardiovascular toxicities that resultCOX-1 and/or COX-2 inhibition limit their use forchemoprevention. Here, we show that the in vitrotumor cell growth–inhibitory activity of a group ofs is closely associated with cGMP PDE inhibition.stance, the potency of SS to inhibit HT-29 colon tu-ell growth was nearly identical to the potency for in-n of cGMP PDE in tumor cell lysates, yet was moren order of magnitude greater than the potency fortion of COX-1 or COX-2. Sulindac sulfone and sulf-did not inhibit COX but did inhibit tumor cellh and cGMP PDE. Additionally, a strong positiveation existed between the potencies of other chemi-iverse NSAIDs for cGMP PDE inhibition and colonr cell growth inhibition, yet no correlation wasent for COX-2 inhibition. Interestingly, the COX-2–ve inhibitor rofecoxib, which did not inhibit tumorowth, also did not inhibit cGMP PDE, whereas an-COX-2 inhibitor, celecoxib, did inhibit tumor cellh at concentrations that inhibited cGMP PDE. Theseuggest that cGMP PDE inhibition represents a COX-endent mechanism responsible for the in vitro tumorrowth–inhibitory activity of certain NSAIDs and2 inhibitors that potentially could be targeted foriscovery of safer and more efficacious drugs forprevention.ause of its high potency to inhibit colon tumor cellh, SS was used to investigate the role of cGMP in reg-g tumor cell growth and to identify the PDE isozymesponsible for its growth-inhibitory activity. More-SS-induced growth inhibition was associated withition of proliferation and induction of apoptosis,occurred at concentrations comparable with thoseary for inhibition of cGMP hydrolysis in HT-29 cells. The biological relevance of cGMP PDE inhibition-29 cell lysates was confirmed using a cell-based im-assay, which showed the ability of SS to increaselevels in colon tumor cells at concentrations that in-d growth and induced apoptosis. The ability of othersignaling activators, including the guanylyl cyclase

ator NOR-3 and the PDE5-specific inhibitor45, to inhibit growth suggests that activation ofsignaling is sufficient for inhibition of colon tumor

rowth. Consistent with previous findings (12), wed that activation of the cGMP signaling pathwaypounds such as SS and MY5445 is associated with

tion of β-catenin transcriptional activity, a pathwaywell documented to be associated with tumorigen-herefore, inhibition of cGMP PDE, activation ofsignaling, and inhibition of β-catenin–mediated

ription by certain NSAIDs may fully account for theirro tumor cell growth–inhibitory and apoptosis-ing activity and contribute to their anticancer efficacy.ough activation of cGMP signaling is sufficient tot colon tumor cell growth, nonselective activation
pathway is not a plausible pharmacologic approachelop new chemopreventive agents due to the myriad
shownplasm



s that cGMP plays in normal physiology and the po-l for adverse effects. However, we hypothesized thatve inhibition of one or more of the PDE isozymes ex-d in tumor cells may provide amore effective strategyctively induce cGMP levels and kill neoplastic cells.morigenic colon cells were not found to express theisozyme that was expressed in colon tumor cell lines.inical relevance of this observation was confirmed bynohistochemical analysis of human colon speci-which showed increased PDE5 expression in adeno-nd adenocarcinomas compared with normal colonicsa. Consistent with these observations, SS displayedvity for inhibition of cGMP hydrolysis by PDE5 com-with all other PDE isozymes and improved potencyth inhibition of colon tumor cell growth and cGMPctivity when compared with the non-tumorigenic co-lls. These results are consistent with previous studiessed PDE5 antisense in HT-29 human colon tumoro show that PDE5 activity is necessary for growthrvival of colon tumor cells (24).ough the known PDE5 inhibitor MY5445 inhibitedtumor cell growth at concentrations that are compa-to concentrations that can inhibit PDE5, highly po-DE5 inhibitors used for erectile dysfunction such asfil and vardenafil required micromolar levels to in-tumor cell growth compared with nanomolar levelsed to inhibit purified PDE5. There are several poten-planations for this discrepancy, including an inabil-the drugs to access PDE5 within cancer cells due tod uptake, confined subcellular localization of the en-in cancer cells, or an active efflux of the drugs fromr cells. Whatever the reason for this, the data pre-d here strongly suggest that PDE5 inhibition andelevation are responsible for the tumor cell

h–inhibitory and apoptosis-inducing activity of SSossibly other NSAIDs.also show that SS inhibited cGMP hydrolysis not on-human HT-29 colon tumor cell lysates but also ingenates from colon tumors grown in vivo. Addition-S displayed selectivity for inhibition of tumor tissuePDE compared with normal colon tissue cGMP

mirroring the observations in cell cultures. Further-the PDE5-selective inhibitor MY5445 exhibited ar degree of selectivity for inhibition of cGMP PDE ac-in tumor tissue homogenates compared with that ex-d by SS. The pharmacological relevance of theseations are supported by other experiments showingulindac can suppress colon tumor growth in mice ates that can generate plasma levels of SS that are ca-of inhibiting PDE5 in vivo. Although more work isary to confirm the direct effects of such in vivo treat-on cGMP signaling, these data suggest that inhibi-f PDE5 and activation of cGMP signaling may playportant role in the in vivo antineoplastic activity ofac and potentially other NSAIDs. In support of thisbility, pharmacokinetic studies in humans have
that clinically used dosages of sulindac can achieve
a concentrations of the sulfide that are comparable




with ttion aunlikeCOX-lack Cinhibicy andWe

vationsignalchemoNSAIDtweenCOX-for chrantedas a m

Discl

No p

Ackn

We thistopaand Dr

Grant

NIHThe

paymenadvertiindicate

Refe1. Th

drucal

2. Chco

3. Albcanpro

4. Eldde(COfro

5. HainfinBio

6. KaBio

7. Piathea mcyc

8. PiadaCa

9. Piaazoing

10. Storegpo

11. ArbSpbures

12. ThsisaseCa

13. PiaapRe


www.a

Do


hose required for PDE5 inhibition and cGMP eleva-s we report here (21, 22). However, such dosages arely to be sufficiently safe for chemoprevention due todependent toxicities. New derivatives of sulindac thatOX inhibitory activity, but have more potent growthtory activity hold more promise for improved effica-reduced toxicity (26).conclude that PDE5 inhibition and subsequent ele-of intracellular cGMP levels and activation of cGMPing are important mechanisms responsible for thepreventive efficacy of sulindac and potentially others. The differential expression of PDE5 isozymes be-normal and tumorigenic colon cells provides a

independent strategy to identify new drug candidatesemoprevention. As such, further studies are war-
to determine the role of PDE5 and cGMP signaling Rece
zza GA, Thompson WJ, Pamukcu R, et al. Exisulind, a novel pro-optotic drug, inhibits rat urinary bladder tumorigenesis. Cancers 2001;61:3961–8.

14. WhinmeMo

15. Betio

16. Socle

17. Drphgro

18. WhphAm

19. Zh5 rAm

20. Zhesca

21. GiarecN E

22. Dady

23. TintivtumGM33

24. Zhcyinh

25. WaNothatox75

26. PiathagroRe

acrjournals.org


osure of Potential Conflicts of Interest

otential conflicts of interest were disclosed.

owledgments

hank Dr. Andre Klein-Szanto for providing clinical specimens andthologic support for analysis of PDE5 expression in tissue sections. William Thompson for providing the PDE5 antibody.

Support

grants R01 CA131378, R21 NS59509, and R03 CA128021.costs of publication of this article were defrayed in part by thet of page charges. This article must therefore be hereby markedsement in accordance with 18 U.S.C. Section 1734 solely tothis fact.

ived 02/11/2010; revised 04/23/2010; accepted 05/13/2010;
olecular target and pathway for chemoprevention. published OnlineFirst 09/28/2010.
rencesun MJ, Henley SJ, Patrono C. Nonsteroidal anti-inflammatorygs as anticancer agents: mechanistic, pharmacologic, and clini-issues. J Natl Cancer Inst 2002;94:252–66.an TA. Nonsteroidal anti-inflammatory drugs, apoptosis, andlon-cancer chemoprevention. Lancet Oncol 2002;3:166–74.erts DS, Hixson L, Ahnen D, et al. Do NSAIDs exert their coloncer chemoprevention activities through the inhibition of mucosalstaglandin synthetase? J Cell Biochem Suppl 1995;22:18–23.er DJ, HaltonDE, HagueA, ParaskevaC. Induction of apoptotic cellath in human colorectal carcinoma cell lines by a cyclooxygenase-2X-2)-selective nonsteroidal anti-inflammatory drug: independence

m COX-2 protein expression. Clin Cancer Res 1997;3:1679–83.nif R, Pittas A, Feng Y, et al. Effects of nonsteroidal anti-lammatory drugs on proliferation and on induction of apoptosiscolon cancer cells by a prostaglandin-independent pathway.chem Pharmacol 1996;52:237–45.shfi K, Rigas B. Is COX-2 a ‘collateral’ target in cancer prevention?chem Soc Trans 2005;33:724–7.zza GA, Rahm AK, Finn TS, et al. Apoptosis primarily accounts forgrowth-inhibitory properties of sulindac metabolites and involvesechanism that is independent of cyclooxygenase inhibition, cellle arrest, and p53 induction. Cancer Res 1997;57:2452–9.zza GA, Rahm AL, Krutzsch M, et al. Antineoplastic drugs sulin-c sulfide and sulfone inhibit cell growth by inducing apoptosis.ncer Res 1995;55:3110–6.zza GA, Alberts DS, Hixson LJ, et al. Sulindac sulfone inhibitsxymethane-induced colon carcinogenesis in rats without reduc-prostaglandin levels. Cancer Res 1997;57:2909–15.ner GD, Budd GT, Ganapathi R, et al. Sulindac sulfone inducedression of rectal polyps in patients with familial adenomatouslyposis. Adv Exp Med Biol 1999;470:45–53.er N, Kuwada S, Leshno M, Sjodahl R, Hultcrantz R, Rex D.oradic adenomatous polyp regression with exisulind is effectivet toxic: a randomised, double blind, placebo controlled, dose-ponse study. Gut 2006;55:367–73.ompson WJ, Piazza GA, Li H, et al. Exisulind induction of apopto-involves guanosine 3′,5′-cyclic monophosphate phosphodiester-inhibition, protein kinase G activation, and attenuated β-catenin.

ncer Res 2000;60:3338–42.

itehead CM, Earle KA, Fetter J, et al. Exisulind-induced apoptosisa non-small cell lung cancer orthotopic lung tumor model aug-nts docetaxel treatment and contributes to increased survival.l Cancer Ther 2003;2:479–88.avo JA. Cyclic nucleotide phosphodiesterases: functional implica-ns of multiple isoforms. Physiol Rev 1995;75:725–48.nnenburg WK, Beavo JA. Cyclic GMP and regulation of cyclic nu-otide hydrolysis. Adv Pharmacol 1994;26:87–114.ees M, Zimmermann R, Eisenbrand G. 3′,5′-Cyclic nucleotideosphodiesterase in tumor cells as potential target for tumorwth inhibition. Cancer Res 1993;53:3058–61.arton J, Strange JW, Moller GM, et al. Antiproliferative effects ofosphodiesterase type 5 inhibition in human pulmonary artery cells.J Respir Crit Care Med 2005;172:105–13.

u B, Strada S, Stevens T. Cyclic GMP-specific phosphodiesteraseegulates growth and apoptosis in pulmonary endothelial cells.J Physiol Lung Cell Mol Physiol 2005;289:L196–206.

u B, Strada SJ. The novel functions of cGMP-specific phosphodi-terase 5 and its inhibitors in carcinoma cells and pulmonary/rdiovascular vessels. Curr Top Med Chem 2007;7:437–54.rdiello FM, Hamilton SR, Krush AJ, et al. Treatment of colonic andtal adenomas with sulindac in familial adenomatous polyposis.ngl J Med 1993;328:1313–6.vies NM, Watson MS. Clinical pharmacokinetics of sulindac. Anamic old drug. Clin Pharmacokinet 1997;32:437–59.sley HN, Gary BD, Keeton AB, et al. Sulindac sulfide selec-ely inhibits growth and induces apoptosis of human breastor cells by phosphodiesterase 5 inhibition, elevation of cyclicP, activation of protein kinase G. Mol Cancer Ther 2009;8:

31–40.u B, Vemavarapu L, Thompson WJ, Strada SJ. Suppression ofclic GMP-specific phosphodiesterase 5 promotes apoptosis andibits growth in HT29 cells. J Cell Biochem 2005;94:336–50.rner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR.nsteroid drug selectivities for cyclo-oxygenase-1 rathern cyclo-oxygenase-2 are associated with human gastrointestinalicity: a full in vitro analysis. Proc Natl Acad Sci U S A 1999;96:63–8.zza GA, Keeton AB, Tinsley HN, et al. A novel sulindac derivativet does not inhibit cyclooxygenases, but inhibits colon tumor cellwth and induces apoptosis with antitumor activity. Cancer Prev
s 2009;2:572–80.



2010;3:1303-1313. Published OnlineFirst September 28, 2010.Cancer Prev Res Heather N. Tinsley, Bernard D. Gary, Jose Thaiparambil, et al. Associated with Phosphodiesterase 5 InhibitionSulfide and Other Nonsteroidal Anti-Inflammatory Drugs Is

Inhibitory Activity of Sulindac−Colon Tumor Cell Growth

Updated version

10.1158/1940-6207.CAPR-10-0030doi:

Access the most recent version of this article at:

Cited articles

http://cancerpreventionresearch.aacrjournals.org/content/3/10/1303.full#ref-list-1

This article cites 26 articles, 12 of which you can access for free at:

Citing articles

http://cancerpreventionresearch.aacrjournals.org/content/3/10/1303.full#related-urls

This article has been cited by 8 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerpreventionresearch.aacrjournals.org/content/3/10/1303To request permission to re-use all or part of this article, use this link

for Cancer Research. on August 16, 2021. © 2010 American Associationcancerpreventionresearch.aacrjournals.org Downloaded from


http://cancerpreventionresearch.aacrjournals.org/lookup/doi/10.1158/1940-6207.CAPR-10-0030

http://cancerpreventionresearch.aacrjournals.org/content/3/10/1303.full#ref-list-1

http://cancerpreventionresearch.aacrjournals.org/content/3/10/1303.full#related-urls

http://cancerpreventionresearch.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerpreventionresearch.aacrjournals.org/content/3/10/1303


Documents

Research Article Cancer Prevention Research Colon Tumor ...€¦ · Research Article Colon Tumor Cell Growth–Inhibitory Activity of Sulindac Sulfide and Other Nonsteroidal Anti-Inflammatory