Diclofenac, a selective COX-2 inhibitor, inhibits DMH-induced colon tumorigenesis through suppression of MCP-1, MIP-1α and VEGF

MOLECULAR CARCINOGENESIS 50:707–718 (2011)

Diclofenac, A Selective COX-2 Inhibitor, InhibitsDMH-Induced Colon Tumorigenesis ThroughSuppression of MCP-1, MIP-1a and VEGF

Jasmeet Kaur, S.N. Sanyal*

Department of Biophysics, Panjab University, Chandigarh, India

Angiogenesis is a physiological process involving growth of new blood vessels from pre-existing ones; however, it alsoplays a critical role in tumor progression. It favors the transition fromhyperplasia to neoplasia, that is, from a state of cellularmultiplication to uncontrolled proliferation. Therefore targeting angiogenesis will be profitable as a mechanism to inhibittumor’s lifeline. Further, it is important tounderstand the cross-communication between vascular endothelial growth factor

(VEGF)—master switch in angiogenesis and othermolecules in the neoplastic and pro-inflammatorymilieu.We studied therole of two important chemokines [monocyte chemoattractant protein (MCP)-1 and macrophage inflammatory protein(MIP)-la] alongwith VEGF and matrix metalloproteinases (MMPs) in non-steroidal anti-inflammatory drugs (NSAIDs)-

induced chemopreventive effect in experimental colon cancer in rat. 1,2-Dimethylhydrazine (DMH, 30 mg/kg bodyweight,subcutaneously (s.c.) once-a-week) for 18 wk was used as pro-carcinogen and diclofenac (8 mg/kg body weight, orallydaily) as the preferential cyclooxygenase-2 (COX-2) inhibitor. Expression of COX-2 and VEGF was found to be significantly

elevated in theDMH-treatedgroupas compared to the control,whichwas lowerednotably byDiclofenac co-administrationwith DMH. Gelatin zymography showed prominent MMP-9 activity in the DMH-treated rats, while the activity was nearlyabsent in all the other groups. Expression of MCP-1 was found to be markedly increased whereas MIP-1a expression was

found to be decreased in colonicmucosa fromDMH-treated rats, whichwas reversed in the DMH þ Diclofenac group. Ourresults indicate potential role of chemokines alongwith VEGF in angiogenesis in DMH-induced cancer and its chemo-prevention with diclofenac. � 2011 Wiley-Liss, Inc.

Key words: colon cancer; 1,2-dimethylhydrazine; VEGF; MMP-9; MCP-1; MIP-1a

INTRODUCTION

Angiogenesis, the formation of new blood vesselsfrom the pre-existing ones, is a fundamental processoccurring during embryogenesis, development andhomeostasis of adult tissues, and is actively involvedin a wide variety of pathological conditions, such aswound healing, diabetic retinopathy, musculardegeneration, rheumatoid arthritis, psoriasis, andtumor progression [1]. Blood vessels stand as aprimary gate for extravasation of leukocytes andfor the exchange of various biochemical substances[2]. Furthermore, these vessels per se synthesize vaso-active substances [3]. Thus, neovascularization is acritical requirement for tumorgrowthandmetastasisformation [4].The field of angiogenesis research began in the

1960s as an inquiry into how new blood vesselssupport solid tumor growth [5]. Folkman [5] dem-onstrated, in his ground-breaking studies, that solidtumors establish their ownblood supplies by encour-aging the growthof newbloodvessels into the tumortissue to extend beyond the small size of approxi-mately 1 mm3. Thus any significant increase intumor mass must be preceded by an increase inthe vascular supply to deliver nutrients and oxygento the tumor. Once the tumor vasculature is estab-lished, solid tumors acquire the oxygen and

nutrients that enable them to expand exponentiallyand ultimately to metastasize [6]. The endotheliumcomprising these proliferating capillaries secreteparacrine factors that promote survival of adjacentcells by impeding apoptosis, or programmed celldeath [7]. Because of rapid proliferation and insuffi-cient blood supply, progressive tumors easilybecome hypoxic and necrotic [8]. In addition toangiogenic vasculature, it is widely recognized thattumor cells themselves potentially form blood pas-sages [9].Alongwith vascular endothelial growth factor

(VEGF), chemokines have also been shown toregulate angiogenesis directly, as a consequence ofleukocyte infiltration, or growth factor expression[10]. Chemokines are inflammatory cytokines that

Abbreviations: VEGF, vascular endothelial growth factor; MCP-1,monocyte chemoattractant protein-1; MIP, macrophage inflamma-tory protein; COX-2, cyclooxygenase-2; DMH, 1,2-dimethylhydra-zine; PBS, phosphate buffered saline; MPLs, multiple plaquelesions; MMP, matrix metalloproteinase.

*Correspondence to: Department of Biophysics, Panjab University,Chandigarh 160 014, India.

Received 14 August 2010; Revised 4 December 2010; Accepted 13December 2010

DOI 10.1002/mc.20736

Published online 25 January 2011 in Wiley Online Library(wileyonlinelibrary.com).

� 2011 WILEY-LISS, INC.

play critical roles in leukocyte recruitment duringseveral pathological processes [11]. Chemokinesuperfamily is composed of small (8–11 kDa), struc-turally related proteins that induce leukocyte che-motaxis in vitro, which has been divided into twosubfamilies, denotedC–X–Cora andC–Corb, basedon whether or not an intervening amino acid issituated between the first two of four conserved cys-teine residues [12].

Monocyte chemoattractant protein-1 (MCP-1) isthe first C–C chemokine reported to play a direct rolein tumor angiogenesis [13], attributablemostly to itschemotactic effect onmonocytes. MCP-1 is encodedby a single gene, which is well conserved in severalspecies, includinghuman,mouse, and rat [14].MCP-1 is a member of the C–C chemokines (cysteine–cysteine motif) that accumulate and influencemacrophages/monocytes and lymphocytes, andare involved in tumor progression and metastasis[15]. The expression of MCP-1 has been reportedin several tumor types including humanmelanoma,ovarian cancer, and esophageal cancer [16]. How-ever, only a few studies demonstrating MCP-1expression and its biological role in colorectal carci-noma are available to date [17]. Similarly, macro-phage inflammatory protein (MIP)-la is also amember of the C–C subfamily, which includesMIP-1b and MCP-1.

Thus inhibition of angiogenesis presents a vitalmechanism to prevent tumor growth and even tocause tumor regression. However, many aspects ofthe molecular mechanisms underlying the regula-tion of tumor angiogenesis remain still elusive.Understanding such mechanisms will contributenot only to the characterization of the angiogenicprocess itself, but also help to identify moleculartargets for the development of novel chemopreven-tive as well as therapeutic approaches. Therefore inthe present study, we elucidate the moleculesinvolved in promoting angiogenesis in experimentalcolon cancer and its inhibition by diclofenac, a pref-erential cyclooxygenase-2 (COX-2) inhibitor. Ear-lier, we have established the chemopreventivepotential of diclofenac in an experimental coloncancer model induced by 1,2-dimethylhydrazine(DMH) (30 mg/kg body weight) in rat [18–24].Apoptosis or programmed cell death was found tobe the dominating mechanism of cell killings in theanti-cancer end-effect of diclofenac.

MATERIALS AND METHODS

Chemicals

DMH dihydrochloride was purchased fromSigma–Aldrich (St. Louis, MO). Primary antibodiesagainst COX-2, VEGF, MCP-1, MIP-1a and anti-mouse b-actin were purchased from Santa CruzBiotechnology, Inc. (CA), alkaline phosphatase(AP)-conjugated secondary antibodies and BCIP-

NBT (5-Bromo-4-chloro-3-indolylphosphate diso-dium salt/Nitro-blue tetrazolium) from Genei (Ban-galore, India). Phosphate buffered saline (PBS,calcium and magnesium free) and morpholino pro-pane sulfonic acid (MOPS) were from Himedia(Mumbai, India).Diclofenacwas agenerous gift fromRanbaxy Pharmaceuticals Ltd (Gurgaon, India). Allother chemicals and reagents used in the presentstudy were of analytical grade and procured fromthe reputed Indian manufacturers.

Animal Procurement

Male adult Sprague–Dawley rats of body weightbetween 100 and 120 g were obtained from theinbred population of the Central Animal House,Panjab University, Chandigarh. These were accli-matized to the control diet (rodent chow) and waterad libitum for atleast 1 wk.Animalsweremaintainedas per the principles and guidelines of the EthicsCommittee of Animal Care of Panjab University andin general, according to the NIH guidelines (Ruleno 23–85, as revised in 1985). The animals werehoused three per cage in polypropylene cageswith a wire mesh top and a hygienic bed of husk(regularly changed) in awell ventilated animal roomtill the end of the experimental period. The animalrooms were provided with electric heater and watercooler during the winter and summer months,respectively. The animals were also maintainedunder a 12 hr photoperiod of light and darkness,respectively.

Treatment Schedule

Animals were assorted into the following groupsfor 18 wk of the treatment:

Vehicle treated: Animals were administered thevehicle (1 mM EDTA-saline subcutaneously, s.c.)once-a-week and 0.5% carboxymethyl cellulose(CMC) per oral (p.o.) daily.

DMH treated: Animals were administered DMH (pre-pared fresh at 30 mg/kg body weight in 1 mMEDTA-saline, pH 7.0) s.c. once-a-week [22].

DMH þ diclofenac treated: Animals were adminis-tered DMH (prepared fresh at 30 mg/kg bodyweight in 1 mM EDTA-saline, pH 7.0) s.c. once-a-week; and received Diclofenac (8 mg/kg bodyweight in 0.5% CMC) orally (by canulla) daily.The anti-inflammatory dose of diclofenac inthe rat was determined earlier in the laboratoryin a carrageenan induced oedema in the hind paw[22].

Diclofenac treated: Animals were administered Diclo-fenac (8 mg/kg body weight) orally (by canulla)daily.

Animals from each groupwere sacrificed as per thedesigned plan to study various parameters determin-ing the progression/regression of tumors.

708 KAUR AND SANYAL

Molecular Carcinogenesis

Tissue Pathology

All the animals were sacrificed under light etheranesthesia followed by cervical dislocation. Colonswere removed, flushed with chilled physiologicalsaline (NaCl solution, 9 g/L) and opened longitudi-nally. These were thoroughly examinedmacroscopi-cally for the presence of any tumors. Differentregions of colon were carefully examined under ahand held lens for counting the tumors. Chemopre-ventive response was assessed on the basis of tumorincidence, burden and multiplicity that were calcu-lated as follows:

Tumor incidence: percentage of animals havingtumors.

Tumor burden: total number of tumors counted/totalnumber of rats.

Tumor multiplicity: total number of tumors counted/number of tumor bearing rats.

Similarly, the incidence ofmacroscopic neoplasticlesions/plaques called the multiple plaque lesions(MPLs) were also calculated as:

MPL incidence: percentage of animals having MPLs.MPL burden: total number of MPLs counted/totalnumber of rats.

MPL multiplicity: total number of MPLs counted/number of MPLs bearing rats.

Histopathological Analysis

The colonic segments were divided into proximal,middle, and distal regions, fixed immediately in 10%buffered formalin and processed routinely. Paraffinembedded sections were stained with Hematoxylinand Eosin (H&E) for histopathological examin-ations. Presence of carcinoma, hyperplasia, and dys-plasia, aberrant crypt foci associated with aggregatesof lymphoid tissues andmucosal inflammationwereobserved in the stained sections of colon under thelight microscope.For histological studies, small pieces of colonic

segment were taken, washed with ice-cold 0.9%saline and fixed in 10% buffered formalin for24 hr. After fixation, the tissues were processed care-fully for paraffin wax (58–608C). Paraffin sectionswere cut at 5 mmthickness in amicrotome, stretchedon an egg albumin coated slide and then subjected tothe histological technique, where sections werestained with Delafield hematoxylin–eosin. Sectionswere dewaxed in xylene, downgraded (hydrated) indecreasing percentage of alcohols and brought towater and stained with hematoxylin for 20 sec,washed in tapwater, till the appearance of blue color,rinsed in ammonia water, again washed with water,treatedwith acid water (if over stained). The sectionswere upgraded (dehydrated) in alcohol till 70%,stained with 1% alcoholic eosin for 30 sec,

differentiated in 90% alcohol, cleared in xyleneand finally mounted in distyrene plasticizer xylene(DPX).

Western Blot Analysis

Protein extracts (100 mg) from each treatmentgroup were separated on 10% SDS–PAGE. The separ-ated proteins were electrophoretically transferred tonitrocellulose membrane (Genei). Immunoblot wasprepared using primary antibodies (COX-2—1:1000,VEGF—1:1000, MCP-1—1:1000, MIP-1a—1:1000,and b-actin—1:10 000) from Santa Cruz Biotechnol-ogy, Inc. and AP-conjugated respective secondaryantibodies at a dilution of 1:10 000 (Genei). BCIP-NBT detection system was used to develop the blot.Bands obtained were densitometricaly analyzedusing Image J software and the density was normal-ized with b-actin and expressed as gray values indensitometric units. b-actin had been used in theblot as the protein loading control.For preparation of protein extracts, colons were

removed and rinsed from the different treatmentgroups after completion of 18 wk. Total lysates wereprepared in fresh ice-coldprotein lysis buffer [10 mMTris, 100 mM Sodium chloride (NaCl), 5 mM EDTA,1% Triton X-100, 1 mM phenyl methyl sulfonylfluoride (PMSF) and 2 mM dithiothreitol (DTT),pH 8]. The extracts were cleared by centrifugationat 10 000g for 10 min at 48C. The supernatants werecollected as the total lysate. Protein concentrationwas determined by the method of Bradford [25].

Immunohistochemical Analysis

Five micrometers thick paraffin sections of ratcolon were incubated at 608C in an oven for30 min for antigen retrieval and deparaffinized inxylene for 30 min. The sections were then graduallyhydrated in descending series of alcohol (100%,90%, 70%, 50%, and 30%) and brought to water.The non-specific binding was blocked by incubatingthe sections with 2% bovine serum albumin (BSA) inPBS (10 mM,pH7.2) for 30 minat378C.The sectionswere then incubated with the primary antibodies(COX-2—1:1000 and VEGF—1:1000) in a moistchamber for 2 hr at 378C.For negative control only 1% BSA was added

without antibody. After incubation, washings weregiven with PBS, PBS Tween (PBS with 0.05% Tween-20) and PBS, successively, for 5 min each. Thesections were then incubated with the respectivesecondary antibodies at a dilution of 1: 10 000 for2 hr at 378C. Sections were washed again in thesame manner as described above and the reactiveproduct was developed using BCIP/NBT under darkconditions.Reactionwas terminated bywashingwith distilled

water. Sections were counterstained with methylgreen, mounted in DPX and observed under a lightmicroscope.

CHEMOKINES AND VEGF IN CANCER CHEMOPREVENTION 709


Microscopy and Data Analysis

Sections were viewed and photographed with aLeica Optiphot microscope to which was attacheda Leica Digital camera.

Gelatin Zymography for Detection of MatrixMetalloproteinase (MMP) Activity

Degradation of extracellular matrix is crucial formalignant tumor growth, invasion, metastasis, andangiogenesis. MMPs are capable of degrading essen-tially all the matrix components.

For the study of MMP activity, a 10% homogenateof colonic tissue was prepared in PBS followed bycentrifugation at 1000g at 48C to remove the cellulardebris. The resultant supernatant was again centri-fuged at10 000g at 48Cand the resultant supernatantwas subjected to gelatin zymography after esti-mation of protein by Bradford [25] method. Poly-acrylamide minigels (12%) were cast containing0.1% gelatin. Gelatin solution was made up to 2%stock in distilled water and dissolved by heating.Equal amounts of samples (100 mg protein) wereloaded in each lane in standard SDS loading buffercontaining 0.1% SDS without b-mercaptoethanol.Boiling was avoided before loading of the samples[26]. The gel was run at constant current of 15 mA,until the dye front reached the end of the gel. Thegels were then soaked in 200 ml of 2.5% (v/v) TritonX-100 in distilled water in a shaker for 1 hr, with onechange after 30 min at 208C to remove SDS. The gelswere incubated in the incubation assay buffer(50 mM Tris, 200 mM NaCl, 10 mM CaCl2, 0.05%Nonidet-P40, pH 7.8) for 12 hr at 378C and stainedwith Coomassie brilliant blue-R 250 in 50% meth-anol and 10% acetic acid, followed by washing withdistilled water for 1 min. Gelatinolytic activities ofMMPwere detected as the clear zones of lysis againsta blue background. Proteinmolecular weightmarkerwas also run simultaneously with the test samples.The gel was then destained with destaining solution[26]. The gelwas photographedwith aGelDoc (Foto-dyne, Hartland, WI) and analyzed using the Image Jsoftware (NIH, Bethesda, MD).

Statistical Analysis

Statistical analysis of the data was performed bythe one-way analysis of variances (one-way ANOVA)following the Post hoc test using the least significantdifference (LSD)method with the help of a SPSS v 14computer software. A value of P < 0.05 was con-sidered significant in the present study.

RESULTS

Tissue Pathology

Figure 1a shows the profile of mean body weightduring the 18 wk of treatment schedule. All thegroups showed normal body growth and gained

weight when compared to their respective initialbody weight. Figure 1b represents the comparativegross morphology of colonic segments from thedifferent treatment groups. Control and Diclofenacalone treated animals showed the normal colonicmucosa, devoid of any tumor/lesion growth. How-ever, an incidence of 100% tumor was recorded inthe DMH treated group (Table 1). Macroscopictumors were noted in all the regions of the colon.The tumor incidence was seen to be reduced to 25%in the DMH þ Diclofenac co-treated group. Sim-ilarly, maximum tumor burden and tumor multi-plicity were also noted in the DMH treated groupwhile these values were minimally observed in theDMH þ Diclofenac group (Table 1).

Histopathological Analysis

Figure 2 shows the histopathological observationsof colonic segments from the different treatmentgroups. Control (Figure 2a) and diclofenac(Figure 2d) alone treated animals exhibit the normalmucosal architecture. Epithelia from DMH treatedrats showed the occurrence of severe dysplasia andadenocarcinoma (tubular as well as villous)(Figure 2bi and ii) in all the animals. The tumors alsoshowed signet ring cell carcinoma (Figure 2biii),highly invasive carcinoma with transmural infiltra-tions (Figure 2biv) and invasive dysplastic ACFsassociated with the lymphoid follicles (Figure 2bv).In DMH þ Diclofenac group, no sign of carcinomawas noted, while the epithelia showed the occur-rence of hyperplasia and early changes of aberrantcrypt foci only (Figure 2c). These results of histopa-thological observations are summarized in theTable 2.

Expression of COX-2

Figure 3a–c shows the protein as well as immuno-histochemical expression of COX-2 in the colonictissue after 18 wk of treatment schedule, respect-ively. Colonic mucosa from the DMH administeredrats showed an increased protein as well as immu-nohistochemical expression of COX-2 as comparedto the control group. Immunohistochemical expres-sion of COX-2 showed a very high cytoplasmicexpression both in the epithelial and adjacent stro-mal cells in this group. Diclofenac significantlyinhibited the expression of COX-2, when adminis-tered simultaneously with DMH. Control and diclo-fenac only groups, however presented a low level ofCOX-2 expression both in the colonic mucosa andtissue sections.

Expression of VEGF

In the present study, expression of VEGF wasexamined by Western blot as well as immunohisto-chemistry after 18 wk of DMH treatment. A markedincrease in the protein expression of VEGF was seenin the colonic mucosa from DMH treated rats

710 KAUR AND SANYAL


(Figure 4a and b) as compared to the control group.Diclofenac co-treatmentwithDMHwas able to bringdown the expression level of VEGF significantly.Diclofenac alone treated group presented a lowexpression level ofVEGF similar to that of the controlgroup.Immunohistochemical analysis for VEGF also

showed the highest expression in the tissue sectionsfrom DMH treated rats (Figure 4c). The expressionwas identifiable both in the epithelial aswell as in the

stromal cells. Diclofenac co-administration withDMH significantly inhibited the expression of VEGF,stressing its anti-angiogenic potential in the presentstudy.Control anddiclofenac alonegroups,howeverpresented a low level of VEGF expression both in thecolonic mucosa and tissue sections.

Gelatin Zymography

Quantitative gelatin zymography allows the sim-ultaneous measurement of the latent and active

Figure 1. (a) Alterations in the mean body weight (g) of Control, DMH, DMH þ Diclofenac, and Diclofenac onlygroups during 18 wk of treatment schedule. (b) Comparative gross morphology of the colonic segments from theControl, DMH, DMH þ Diclofenac, and Diclofenac only groups examined after 18 wk of treatment schedule.Arrows show occurrence of tumors of different size and shape in both intact and longitudinally opened colons fromDMH treated group. The arrows also show angiogenesis as evident by blood vessels and redness etc. All the othersgroups showed normal colonic mucosa with no sign of malignancy. [Color figure can be viewed in the online issue,which is available at wileyonlinelibrary.com.]



forms of the gelatinases MMP-2 (gelatinase A) andMMP-9 (gelatinase B). The metalloproteinases arecapable of degrading gelatin; therefore, by incorpo-rating gelatin into the polyacrylamide gel a clearzone indicates the presence of a matrix degradingenzyme. It is a representation of the levels of theactive and latent states in both free and complexedform in the various tissues studied. Figure 5a shows azymogram obtained from the protein extracts of allthe treatment groups after 18 wk of treatment.Clearly the extracts contained the proteinase speciesof varying molecular weight exhibiting the gelati-nase activity. Comparisonwith the SDS proteinmol-ecularweightmarker showed themarkedpresence ofactiveMMP-9 at 92 kDa in theDMH treated group ascompared to control group.

The lytic bands of higher molecular weight form,found at approximately 220 kDa may representMMP-9 heterodimer (Figure 5a). There was a signifi-cantly higher level of both activeMMP-9 andMMP-9heterodimer lytic bands in the tumor tissue fromDMHtreated group as compared to thenormal tissuefrom the control group (Figure 5b). DMH þDiclofenac treated group presented significantlylower levels of MMP-9 heterodimer lytic activitywhile active MMP-9 band was found to be nearlyabsent as compared to theDMHalone treated group,demonstrating that diclofenac could alter theincreased invasive properties of the cells by inhibit-ing the activity of MMPs.

Expression of MCP-1 and MIP-1�

In the present study, MCP-1 and MIP-1aexpressions were studied in the colonic mucosa byWestern blot after 18 wk. The protein expression ofMCP-1 was found to be significantly elevated in theDMH treated rats as compared to the control group(Figure 6a and b). DMH þ Diclofenac co-treatedgroup however expressed a significantly lowerprotein level of MCP-1 when compared to theDMH alone treated group. Result from diclofenacalone treated groupwas similar to that of the controlgroup.

The protein expression of MIP-1a was found to besignificantly decreased in the DMH treated rats as

compared to the control group after 18 wk of treat-ment (Figure 6a and b). The expression was howeversignificant in the mucosa from control as well asdiclofenac alone treated groups. Diclofenac whenadministered simultaneously with DMH increasedthe expression of MIP-1a appreciably.

DISCUSSION

In the present study, despite extensive samplingthere was no evidence of carcinoma in situ or Frankmalignancy in any of the control and diclofenacalone treated animals. The neoplastic transform-ations were found to be regressed where DMH wasgiven along with diclofenac, while clearly thecharacteristic neoplastic changes like hyperplasia,dysplasia and the adenoma formations were evidentin the DMH treated animals. Invasive crypts werefound to be associated with the lymphoid follicles.Theoccurrence of dilated crypts recognized ashyper-plasia is an early indication of neoplastic changeswhile the dysplasia is considered as a positive sign ofadenoma and usually characterized by the enlarge-ment of epithelial nuclei, loss of mucin, and deeplystained nuclei [27]. In the advance stages the dys-plastic crypt cells expand and invade the neighbor-ing crypts inducing crypt fission leading to theformation of carcinoma [28].Macroscopic and histologic events associatedwith

the promotion stage of carcinogenesis examinedafter 18 wk of treatment showed the occurrence ofadenocarcinoma in theDMHtreated group. Infiltrat-ing carcinoma andmucin containing signet ring cellcarcinoma were also evident. In addition, all thetumors and mucosal lesions of DMH treated groupwere found to be malignant. The invasiveness of theadenocarcinomas, as found in the DMH treatedgroupof the present study, closely resembles to thosereported earlier [29] and considered as a promotionstage of colon carcinogenesis. DMH þ Diclofenacgroup reported the absence of carcinoma in any ofthe animals. The percentage of tumor incidence wasalso found to be regressed in the DMH þ Diclofenactreated group. These findings, taken together there-fore, imply that non-steroidal anti-inflammatory

Table 1. DMH Induced Carcinogenesis and Regressive Response of Diclofenac Examined After 18 wk of TreatmentSchedule (n ¼ 8 Per Group)

Animal groupsTumor incidence

(%)Tumorburden

Tumormultiplicity

Lesion incidence(%)

Lesionburden

Lesionmultiplicity

Control — — — — — —DMH 100 2 2 100 3 3DMH þ Diclofenac — — — 25 0.25 1Diclofenac — — — — — —

Tumor/lesion incidence ¼ percentage of animals having tumors/lesions.Tumor/lesion burden ¼ total number of tumors/lesions counted/total number of rats.Tumor/lesion multiplicity ¼ total number of tumors/lesions counted/number of tumor/lesions bearing rats.

712 KAUR AND SANYAL


drugs (NSAIDs) (e.g., diclofenac) do start inducingcancer preventing effects much before the onset ofthe occurrence of the carcinogenic events in thepresent model of colon cancer.Overexpression of COX-2 not only inhibits apop-

tosis, but also increases the invasive potential of themalignant cells [30]. Previous works reported theCOX-2 activation, induced in epithelial cells, tobe involved in the regulation of tumor angiogenesis

and cellular proliferation [31]. Diclofenac co-treat-ment with DMH inhibited the expression of COX-2significantly. Control and diclofenac only groupspresented a low expression of COX-2 protein. Itcan be suggested that COX-2 derived prostaglandins(PGs)may be acting on themalignant epithelial cells(an autonomous effect) or on the surroundingstroma (a landscaping effect) to promote tumordevelopment and invasion [32].

Figure 2. Photomicrographs of histologic (H&E) cross sections of the colons from the Control, DMH, DMH þDiclofenac and Diclofenac only groups examined after 18 wk of treatment schedule. (A) Control sections showingthe normal mucosal details of the epithelial surface (100�). (B) (i) Tubular adenocarcinoma (100�), inset showsbudding and branching of malignant glands in submucosa (200�). (ii) Villous adenocarcinoma with pools of mucin(M) (100�) and tumor growing into a lymphatic (L) (inset, 200�). (iii) Signet ring-cell carcinoma (100�). Insets showtypical signet ring cells with nuclei displayed towards one end (200�, 400�). (iv) Invasive tubulovillous adenocarci-noma with few mucin producing goblet cells (100�). Inset shows a budding gland (200�), M—mucin pools.(v) Invasive carcinoma (100�) with budding and branching dysplastic glands (400�) and desmoplasia (arrows).C: Section fromDMH þ Diclofenac group showingmild hyperplasia (�100). (D) Section fromdiclofenac alone groupshowing the normal histology of mucosal surface (100�). [Color figure can be viewed in the online issue, which isavailable at wileyonlinelibrary.com.]



Tumor stromal cells in vivo are known to produceVEGF, implying a co-operation between stromal andcancer cells in the induction of tumor angiogenesis.Increased VEGF level after 18 wk of DMH treatmentsuggests that the angiogenic switch is turned on tofavor vascular growth for tumor expansion [33] veryearly in the process of tumor development. Indeed,

the angiogenic switch clearly correlates with tumorprogression in mouse models of multi-stepcarcinogenesis.Tumor vesselsmight preferentially take advantage

of special signaling molecules under pro-inflamma-tory microenvironment in vivo [34]. For example,COX-2 is known to modulate angiogenesis by

Table 2. Histopathological (H&E Staining) Evaluation of Colonic Tissue From All the Groups After 18 wk of RespectiveTreatments

Animal groups Carcinoma Dysplasia/adenoma HyperplasiaACF associated aggregates

of lymphoid follicle

Control — — — —DMH þþþ þþþ þþþ þþþDMH þ Diclofenac — — þþ —Diclofenac — — — —

þþþ/þþ/þ indicates high occurrence/moderate occurrence/low occurrence, respectively.

Figure 3. (a) Western blot analysis of COX-2 from each group after 18 wk of treatment schedule. b-actin wasused as the protein loading control. (b) Histogram showing densitometric analysis of COX-2 expression from eachgroup after 18 wk of treatment. (c) Photomicrographs (100�) showing the immunohistochemical expression(arrows) of COX-2 from each group after 18 wk of treatment. BCIP/NBT gave dark blue color (arrows) while thecounterstain methyl green stained the sections light green or cyan. The values are mean � SD of four animals.Significantly different at bP < 0.01 and cP < 0.001 in comparison to control and at fP < 0.001 in comparison toDMH by one way ANOVA. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

714 KAUR AND SANYAL


increasing the production of angiogenic factor,VEGF [31]. The role of COX-2 in angiogenesis duringtumor development has been refined by Changet al. [35] who reported that COX-2-derivedprostaglandin E2 (PGE2) induces tumor-associatedangiogenesis, which is required for the initiationand/or progression of mammary cancer in MMTV-COX-2 mice. They observed that PGE2 inducedangiogenesis at the earliest stage of tumor develop-ment, even before PGE2-induced mammary glandhyperplasia at terminal ductal lobular unit, a precur-sor lesion formammary tumorigenesis. This providesa new understanding of the role of angiogenesis inthis process, as angiogenesis was previously thoughtto play a major role in the later stages of tumordevelopment.Moreover, Chang et al. [35] found thatindomethacin inhibited both angiogenesis andtumor growth, suggesting that NSAIDs suppresstumor development by blocking angiogenesis at amuch earlier stage than was previously appreciated.Our results are thus also in the fair general agreement

that VEGF is one of the molecules that can be tar-geted to stop tumor growth and metastasis [36].In cancer, degradation of the extracellular matrix

in tissue surrounding the tumor is a key event intumor cell invasion and metastasis. Matrix degra-dation is believed to be caused by the action ofproteolytic enzymes, including several types ofMMPs [37]. MMPs facilitate the entry of tumor cellsinto thebloodstream, angiogenesis, tumor cell estab-lishment, and growth [38]. In colon cancer, up-regu-lation of mRNA in tumor cells and/or tumor stromahas been reported for MMP-9 [39]. Thus, colorectalcancer offers a goodmodel todefine the involvementof the type IV collagenase in tumor progression.Tsujii et al. [40] demonstrated that COX-2 activitymay be required to maintain the altered phenotypeof increased invasiveness. Thus the activation ofMMP-9 afterDMHtreatment couldhavebeenmodu-lated by increased COX-2 in the DMH treated groupand treatment with a COX inhibitor, diclofenacreversed the increased invasiveness after DMH

Figure 4. (a)Western blot analysis of VEGF in protein extract from each group after 18 wk of treatment schedule.b-actin was used as the protein loading control. (b) Histogram showing the densitometric analysis of proteinexpression of VEGF after 18 wk treatment from each group. (c) Photomicrographs (�100) showing the immuno-histochemical localization (arrows) of VEGF from each group after 18 wk treatment. BCIP/NBT gave dark blue color(arrows) while the counterstain methyl green stained the sections light green or cyan. The values are mean � SD offour animals. Significantly different at bP < 0.01, cP < 0.001 in comparison to control and at fP < 0.001 incomparison to DMH by one way ANOVA. [Color figure can be viewed in the online issue, which is available atwileyonlinelibrary.com.]



treatment by inhibiting the activation of MMP-9.The results emphasize that the proteinase activityof individual MMPs are impending targets in orderto develop suitable intervention strategies.

MCP-1 is not expressed in normal endothelium,but is induced by diverse stimuli such as cytokinesand oxidized adducts, and secreted by endothelialcells, smoothmuscle andmacrophages. Concerningcolon carcinoma, it has been revealed that MCP-1expression contributes to angiogenesis and tumorgrowth leading to a shorter survival [41]. Yoshidomeet al. [42] showed that the MCP-1 expression of

primary colorectal cancer increased as the clinicalstage advanced, suggesting thatMCP-1 appears to berelated to metastasis potency of colorectal cancer.Our results suggest that MCP-1 could be an essentialfactor contributing to angiogenesis as well as tumorinvasion in theDMH treated rats after 18 wk of treat-ment. The data is in agreement with an early reportby Tanaka et al. [43] which suggests that MCP-1 incolorectal adenoma epithelial cells might beinvolved in macrophage migration and COX-2expression, leading to the subsequent developmentof colonic adenoma. This phenomenon involved theenhancement of capillary sprouting induced byeither Monocyte accumulation or as an unknowndirect effect of MCP-1 on endothelial cells. Theresults are also parallel with a previous report byNakashima et al. [44] in which transfection ofMCP-1 into amurine colon adenocarcinoma cell lineincreased lung metastases by augmentation of neo-vascularization. Diclofenac was able to inhibit MCP-1 and this may be contributing significantly to itsanti-angiogenic and anti-neoplastic effects againstDMH-induced tumors. DFU, a new selective COX-2 inhibitor exerted tumor regression activity in aWalker256 tumor model by suppressing MCP-1 pro-duction in tumor tissues as well as in the circulation[45]. Thus the anti-inflammatory activity of diclofe-nac might be a key machinery to down-regulateMCP-1 expression in the present study.MIP-la expression can be induced in a variety of

cell types, including monocytes, macrophage celllines, mast cell lines, Langerhans cells, fibroblasts,and T lymphocytes. This wide variety of activities invitro suggests that MIP-la may have an importantrole in inflammation. MIP-1a gene transfer to anadenocarcinoma cell line has been shown to reducetumorigenicity and induction of protective immun-ity in immunocompetentmice [46]. Thus, inhibitionof MIP-1a by DMH may reduce anti-tumor immun-ity, leading to an increased risk of tumor develop-ment and shorter life span of animals. Diclofenac co-administration abrogated the inhibitory effect ofDMH on MIP-1a production. Thus our study sup-ports pleiotropic effects ofDiclofenac,whichhas alsobeen shown to attenuate Wnt/b-catenin signalingpathway in colon cancer cells by the activation ofNF-kB [47].Our study presents the intimate correlation of

angiogenesis with carcinogenesis as a key approachto inhibit growth and development of tumors byinhibiting angiogenesis with a preferentially selec-tive COX-2 inhibitor. Similar studies on the vascu-lature of tumor-bearing animals have elucidated thepreventive as well as therapeutic limitations of tar-geting only VEGF pathway towards this approach.Thus targeting COX-2, MMP-9, MCP-1, and MIP-1aat once by single drug or combination thereofprovides a better understanding of the molec-ular and cellular cross-talks, pro-inflammatory

Figure 5. (a) Gelatin zymogram of colonic samples from eachgroup after 18 wk of treatment schedule, showing the lytic bandsagainst the dark background corresponding to MMP-9 heterodimer(�220 kDa) as well as active MMP-9 (�92 kDa) on SDS–PAGE gel. ASDS–PAGE molecular weight marker (in kDa) was also run simul-taneously. (b) Histogram showing the densitometric analysis of lyticbands of both MMP-9 heterodimer and active MMP-9 from eachgroup after 18 wk treatment. The values are mean � SD of fouranimals. Significantly different at bP < 0.01, cP < 0.001 in compari-son to control and at eP < 0.01, fP < 0.001 in comparison to DMHbyone way ANOVA. [Color figure can be viewed in the online issue,which is available at wileyonlinelibrary.com.]

716 KAUR AND SANYAL


microenvironment and the effects of anti-angio-genic molecules on local lesion as well as on pre-metastatic sites.

ACKNOWLEDGMENTS

Financial assistance from the Council of Scientificand Industrial Research (CSIR), Govt. of India(37(1308)/07/EMR-II) is gratefully acknowledged.

REFERENCES

1. Folkman J.Angiogenesis in cancer, vascular, rheumatoid andother disease. Nat Med 1995;1:27–31.

2. Mitchell RN, Cotran RS. Hemodynamic disorders, thrombo-sis, and shock. In: Cotran RS, editor. Pathologic basis ofdisease. 6th edition. Philadelphia: WB Saunders Co; 1999.pp. 113–138.

3. Cotran RS, Kumar V, Collins T. Acute and chronic inflam-mation. In: Cotran RS, editor. Pathologic basis of disease. 6thedition. Philadelphia: WB Saunders Co.; 1999. pp. 50–88.

4. Folkman J. What is the evidence that tumors are angio-genesis dependent? J Natl Cancer Inst 1990;82:4–6.

5. Folkman J. Tumor angiogenesis: Therapeutic implications. NEngl J Med 1971;285:1182–1186.

6. Gimbrone M, Leapman S, Cotrans R, Folkman J. Tumordormancy in vitro by prevention of neovascularization. JExp Med 1972;136:261–276.

7. O’Connor DS, Schechner JS, Adida C, et al. Control ofapoptosis during angiogenesis by survivin expression inendothelial cells. Am J Pathol 2000;156:393–398.

8. Pugh CW, Ratcliffe PJ. Regulation of angiogenesis byhypoxia: Role of the HIF system. Nat Med 2003;9:677–684.

9. Hendrix MJ, Seftor EA, Hess AR, Seftor RE. Vasculogenicmimicry and tumor-cell plasticity: Lessons from melanoma.Nat Rev Cancer 2003;3:411–421.

10. BernardiniG, Ribatti D, SpinettiG, et al. Analysis of the role ofchemokines in angiogenesis. J Immunol Methods 2002;273:83–101.

11. Rollins BJ. Chemokines. Blood 1997;90:909–928.12. Brown KD, Zurawski SM,Mossmann TR, Zurwski C. A family

of small inducible proteins secreted by leukocytes is mem-bers of a new superfamily that include leukocyte and fibro-blast-derived inflammatory agents, growth factors andindicators of various activation processes. J Immunol1989;142:679–687.

13. Ohta M, Kitadai Y, Tanaka S, et al. Monocyte chemoattrac-tant protein-1 expression correlates with macrophage infil-tration and tumor vascularity in human gastric carcinomas.Int J Oncol 2003;22:773–778.

14. Timmers HT, Pronk GJ, Bos JL, van-der-Eb AJ. Analysis of therat JE gene promoter identifies an AP-1 binding site essentialfor basal expression but not for TPA induction. Nucleic AcidsRes 1990;18:23–34.

15. Oppenheim JJ, Zachariae CO, Mukaida N, Matsushima K.Properties of the novel pro-inflammatory supergene ‘inter-crine’ cytokine family. Annu Rev Immunol 1991;9:617–648.

Figure 6. (a) Western blot analysis of MCP-1 and MIP-1a in protein extract from each group after 18 wk oftreatment schedule. b-actin was used as the protein loading control. (b) Histogram showing the densitometricanalysis of protein expression of MCP-1 and MIP-1a after 18 wk treatment from each group. The values aremean � SD of four animals. Significantly different at bP < 0.01, cP < 0.001 in comparison to control and atfP < 0.001 in comparison to DMH by one way ANOVA. [Color figure can be viewed in the online issue, whichis available at wileyonlinelibrary.com.]



16. Koide N, Nishio A, Sato T, Sugiyama A, Miyagawa S. Sig-nificance of macrophage chemoattractant protein-1 expres-sion andmacrophage infiltration in squamous cell carcinomaof the esophagus. Am J Gastroenterol 2004;99:1667–1674.

17. Cheadle EJ, Riyad K, Subar D, et al. Eotaxin-2 and colorectalcancer: A potential target for immune therapy. Clin CancerRes 2007;13:5719–5728.

18. Kaur J, Sanyal SN. PI3-kinase/Wnt signaling associationmediates COX-2/PGE2 pathway to inhibit apoptosis in earlystages of colon carcinogenesis and its chemoprevention bydiclofenac. Tumor Biol 2010;31:623–631.

19. Kaur J, Sanyal SN. Modulation of inflammatory changes inearly stages of colon cancer through activation of PPARg bydiclofenac. Eur J Cancer Prev 2010;19:319–327.

20. Kaur J, Sanyal SN. Intrinsic mitochondrial membrane poten-tial change and associated events mediate apoptosis in che-mopreventive effect of diclofenac in colon cancer. Oncol Res2010;18:481–492.

21. Kaur J, Sanyal SN. Alterations in membrane fluidity anddynamics during initial stages of colon carcinogenesis andits chemoprevention by diclofenac. Mol Cell Biochem2010;341:99–108.

22. Kaur J, Sanyal SN. Inductionof apoptosis as potential chemo-preventive effect of dual cyclooxygenase inhibitor, diclofe-nac in early colon carcinogenesis. J Environ Pathol ToxicolOncol 2010;29:41–53.

23. Kaur J, Sanyal SN. Oxidative stress and stress-signaling inchemoprevention of early colon cancer by diclofenac. Am JBiomed Sci 2010;2:63–78.

24. Kaur J, Sanyal SN. Association of PI3-kinase andWnt signal-ing in non-steroidal anti-inflammatory drug-induced apop-tosis in experimental colon cancer. Am J Biomed Sci2009;1:395–405.

25. Bradford MM. A rapid and sensitive method for the quan-titation of microgram quantities of protein utilizing the prin-ciple of protein-dye binding. Anal Biochem 1976;72:248–254.

26. Billings PC, Habres JM, Liao DC, Tuttle SW. Human fibro-blasts contain a proteolytic activity which is inhibited by theBowman–Birk protease inhibitor. Cancer Res 1991;51:5539–5543.

27. Brittan M, Wright NA. Stem cell in gastrointestinal structureand neoplastic development. Gut 2004;53:899–910.

28. van Leeuwen IMM, Edwards CM, Ilyas M, Byrne HM.Towards a multiscale model of colorectal cancer. World JGastroenterol 2007;13:1399–1407.

29. Sengottuvelan M, Senthilkumar R, Nalini N. Modulatoryinfluence of dietary resveratrol during different phases of1,2-dimethylhydrazine induced mucosal lipid-peroxidation,antioxidant status and aberrant crypt foci development in ratcolon carcinogenesis. Biochim Biophys Acta 2006;1760:1175–1183.

30. Tsujii M, DuBois RN. Alterations in cellular adhesion andapoptosis in epithelial cells over-expressing prostaglandinendoperoxide synthase 2. Cell 1995;83:493–501.

31. Tsujii M, Kawano S, Tsuji S, Sawaoka H, Hori M, DuBois RN.Cyclooxygenase regulates angiogenesis induced by coloncancer cells. Cell 1998;93:705–716.

32. Janakiram NB, Rao CV. Role of lipoxins and resolvins as anti-inflammatory and proresolving mediators in colon cancer.Curr Mol Med 2009;9:565–579.

33. Hanahan D, Folkman J. Patterns and emerging mechanismof the angiogenic switch during tumorigenesis. Cell1996;86:353–364.

34. Furuya M, Nishiyama M, Kimura S, et al. Expression ofregulator of G protein signaling protein 5 (RGS5) in thetumor vasculature of human renal cell carcinoma. J Pathol2004;203:551–558.

35. Chang SH, Liu CH, Conway R, et al. Role of prostaglandin E2-dependent angiogenic switch in cyclooxygenase 2-inducedbreast cancer progression. Proc Natl Acad Sci USA2004;101:591–596.

36. MelyykO, ShumanMA, KimKJ. Vascular endothelial growthfactor promotes tumor dissemination by a mechanism dis-tinct from its effect on primary tumor growth. Cancer Res1996;56:921–924.

37. Matrisan LM. Metalloproteinases and their inhibitors inmatrix remodelling. Trends Genet 1990;6:121–125.

38. Nelson AR, Fingleton B, Rothenberg ML, Matrisian LM.Matrix metalloproteinases: Biologic activity and clinicalimplications. J Clin Oncol 2000;18:1135–1149.

39. Zeng ZS, Huang Y, Cohen AM, Guillem JG. Prediction ofcolorectal cancer relapse and survival via tissue RNA levels ofmatrix metalloproteinase-9. J Clin Oncol 1996;14:3133–3140.

40. Tsujii M, Kawano S, Dubois RN. Cyclooxygenase-2 expres-sion in human colon cancer cells increases metastaticpotential. Proc Natl Acad Sci USA 1997;94:3336–3340.

41. Ogawa M, Yamamoto H, Nagano H, et al. Hepatic expres-sion of Ang-2 RNA in metastatic colorectal cancer. Hepatol-ogy 2004;39:528–539.

42. YoshidomeH, KohnoH, Shida T, et al. Significance ofmono-cyte chemoattractant protein-1 in angiogenesis and survivalin colorectal liver metastases. Int J Oncol 2009;34:923–930.

43. Tanaka S, Tatsuguchi A, Futagami S, et al.Monocyte chemo-attractant protein 1 and macrophage cyclooxygenase 2expression in colonic adenoma. Gut 2006;55:54–61.

44. Nakashima E, Mukaida N, Kubota Y, et al. Human MCAFgene transfer enhances the metastatic capacity of a mousecachectic adenocarcinoma cell line in vivo. Pharm Res1995;12:1598–1604.

45. Muta M, Matsumoto G, Nakashima E, Toi M. Mechanicalanalysis of tumor growth regression by the cyclooxygenase-2 inhibitor, DFU, in a walker 256 rat tumor model: Import-ance of monocyte chemoattractant protein-1 modulation.Clin Cancer Res 2006;12:264–272.

46. Nakashima E, Oya A, Kubota Y, et al. A candidate for cancergene therapy: MIP-1a gene transfer to an adenocarcinomacell line reduced tumorigenicity and induced protectiveimmunity in immunocompetent mice. Pharm Res 1996;13:1896–1901.

47. Cho M, Gwaka J, Parka S, et al. Diclofenac attenuates Wnt/b-catenin signaling in colon cancer cells by activation of NF-kB. FEBS Lett 2005;579:4213–4218.

718 KAUR AND SANYAL


Documents

Diclofenac, a selective COX-2 inhibitor, inhibits DMH-induced colon tumorigenesis through suppression of MCP-1, MIP-1α and VEGF