Download Article

13 Pages·2011·0.7 MB·English

Experimental Gastric Carcinogenesis in Cebusapella
Nonhuman Primates
Joana de Fa´tima Ferreira Borges da Costa1., Mariana Ferreira Leal2*., Tanielly Cristina Raiol Silva1,
Edilson Ferreira Andrade Junior1, Alexandre Pingarilho Rezende1, Jose´ Augusto Pereira Carneiro
Muniz3,AntonioCarlosCunhaLacretaJunior4,PauloPimentelAssumpc¸a˜o5,DanielleQueirozCalcagno2,
Samia Demachki5, Silvia Helena Barem Rabenhorst6, Mar´ılia de Arruda Cardoso Smith2, Rommel
Rodriguez Burbano1
1Laborato´rio de Citogene´tica Humana, Instituto de Cieˆncias Biolo´gicas, Universidade Federal do Para´, Bele´m, Brazil, 2Disciplina de Gene´tica, Departamento de
MorfologiaeGene´tica,UniversidadeFederaldeSa˜oPaulo,Sa˜oPaulo,Brazil,3CentroNacionaldePrimatas,Ministe´riodaSau´de,Ananindeua,Brazil,4Departamentode
Medicina Veterina´ria, Universidade Federal de Lavras, Lavras, Brazil, 5Hospital Universita´rio Joa˜o de Barros Barreto, Universidade Federal do Para´, Bele´m, Brazil,
6Laborato´riodeGene´ticaMolecular,DepartamentodePatologiaeMedicinaForense,EscoladeMedicina,UniversidadeFederaldoCeara´,Fortaleza,Brazil
Abstract
Theevolutionofgastriccarcinogenesisremainslargelyunknown.WeestablishedtwogastriccarcinogenesismodelsinNew-
World nonhuman primates. In the first model, ACP03 gastric cancer cell line was inoculated in 18 animals. In the second
model,wetreated6animalswithN-methyl-nitrosourea(MNU).AnimalswithgastriccancerwerealsotreatedwithCanova
immunomodulator. Clinical, hematologic, and biochemical, including C-reactive protein, folic acid, and homocysteine,
analyseswereperformedinthisstudy.MYCexpressionandcopynumberwasalsoevaluated.Weobservedthatallanimals
inoculatedwithACP03developedgastriccanceronthe9thdaythoughonthe14thdaypresentedtotaltumorremission.In
thesecondmodel,allanimalsdevelopedpre-neoplasticlesionsandfivediedofdrugintoxicationbeforethedevelopment
ofcancer.ThelastsurvivingMNU-treatedanimaldevelopedintestinal-typegastricadenocarcinomaobservedbyendoscopy
onthe940thday.ThelevelofC-reactiveproteinlevelandhomocysteineconcentrationincreasedwhiletheleveloffolicacid
decreased with the presence of tumors in ACP03-inoculated animals and MNU treatment. ACP03 inoculation also led to
anemia and leukocytosis. The hematologic and biochemical results corroborate those observed in patients with gastric
cancer,supportingthatourinvivomodelsarepotentiallyusefultostudythisneoplasia.Incelllineinoculatedanimals,we
detectedMYCimmunoreactivity,mRNAoverexpression,andamplification,aspreviouslyobservedinvitro.InMNU-treated
animals, mRNA expression and MYC copy number increased during the sequential steps of intestinal-type gastric
carcinogenesisandimmunoreactivitywasonlyobservedinintestinalmetaplasiaandgastriccancer.Thus,MYCderegulation
supports the gastric carcinogenesis process. Canova immunomodulator restored several hematologic measurements and
therefore,can beappliedduring/after chemotherapy toincrease the tolerability andduration ofanticancer treatments.
Citation:BorgesdaCostaJdFF,LealMF,SilvaTCR,AndradeJuniorEF,RezendeAP,etal.(2011)ExperimentalGastricCarcinogenesisinCebusapellaNonhuman
Primates.PLoSONE6(7):e21988.doi:10.1371/journal.pone.0021988
Editor:YingXu,UniversityofGeorgia,UnitedStatesofAmerica
ReceivedFebruary16,2011;AcceptedJune15,2011;PublishedJuly21,2011
Copyright:(cid:2)2011BorgesdaCostaetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermits
unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.
Funding:ThisstudywassupportedbytheConselhoNacionaldeDesenvolvimentoCient´ıficoeTecnolo´gico(www.cnpq.br)grantnumber550885/2007-2,
302774/2009-2toRRBand301609/2007-1toMdACSandbytheFundac¸a˜odeAmparoa`PesquisadoEstadodeSa˜oPaulo(www.fapesp.br)grantnumber2007/
02470-3toMFLand2010/11174-1toDQC.Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthe
manuscript.
CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.
*E-mail:[email protected]
.Theseauthorscontributedequallytothiswork.
Introduction examination, blood sample collection and biopsy, on the same
animal over a long period of time [3]. Although nonhuman
Gastric cancer isthefourth mostfrequent cancertype and the primate models are not common and are expensive compared to
second highest cause of cancer mortality worldwide. Gastric rodentmodels,thelonglifespanobservedinnonhumanprimates
cancerprevalenceisinfluencedbygeographic,ethnic,andcultural allows forlong-term carcinogenicstudies.
factors [1]. In addition, adenocarcinoma is the most common Chemicalcarcinogenscausegeneticandepigeneticchangesthat
digestive tractneoplasia [2]. leadtoneoplastictransformation.N-methyl-nitrosourea(MNU)is
Nonhuman primates offer a useful model for carcinogenesis a well-known direct carcinogen, which does not need metabolic
studies.Nonhumanprimatespresentclosephylogenicrelationship activation toexertcarcinogenicity. MNUleads totheproduction
to humans and greater similarities with regard to anatomy, of O6-methylguanine adducts, resulting in premutagenic lesions
physiology, biochemistry, and organ systems, as compared to and DNA strand breaks. MNU is a nitrosation product of
rodents. They also present a relatively large organ size which creatininemetabolismthatisformedinthepresenceofnitritesin
enables repeated diagnostic procedures, such as endoscopic the acidic gastric environment. MNU production is associated
PLoSONE | www.plosone.org 1 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
withtheingestionofmeatproducts,curedmeats,andseafood[4]. 2.3 Cell line inoculation
Moreover,itispossiblethatmanyspecies,includinghumans,are One week before the cell line inoculation, the C. apellas of CL,
exposedtocarcinogenicMNU,generatedintheiralimentarytract CLCA1 andCLCA2 groupswere immunosuppressed byasingle
[5].Thus,tumorigenesisinducedbyMNUisaninterestingmodel doseof 50mg/kgof cyclophosphamide.
tostudy gastriccancer. Four gastric cancer cell lines were tested: ACP02, ACP03,
Canovamaybeapotentialanticancertreatmentinpatientswith AGP01andPG100.Thefirst3celllineswereestablishedbyour
gastriccarcinoma.Itisacomplexhomeopathicimmunomodulator researchgroupfromtumorsamplesofindividualsfromNorthern
indicatedforpatientswhoseimmunesystemisdepressed.Canova Brazil [8]. The PG100 a cell line established from a primary
activatesmacrophagesbothinvivoandinvitroandindirectlyinduces gastric adenocarcinoma was obtained from Rio de Janeiro Cell
lymphocyte proliferation [6]. Since innate and adaptive immune Bank,Brazil(BCRJ).OnlytheACP03cellline,thatwasestablish
responses play a role in tumor surveillance and clearance [7], from an intestinal-type gastric cancer, was inoculated in the
enhancing the ability to trigger a specific immunologic response animals included in the present study, since it was the only one
againstmalignantcellsisanimportantanticancerapproach. that was abletostarta tumorigenesis process inC.apella.
In the present study, we aimed to establish a gastric
Oneweekafterimmunosupression,animalsofCL,CLCA1and
carcinogenesis model in Cebus apella, a nonhuman primate. We CLCA2groupsreceivedpercutaneousinoculationof1010cellsof
inducedstomachtumorsbygastriccancercelllineinoculationas ACP03at the85thpassage betweenthemucosalandsubmucosal
well as MNU treatment for the duration of approximately 2.5
layers of antral stomach region. Ultrasonography was used to
years.Weevaluatedbodyweight,serumbiochemistryvaluesand
visualize thestomachtissues during cellline inoculation.
hematological parameters, as well as MYC proto-oncogene
expression and copy number, in these in vivo models. In these
2.4 Canova treatment
models,wealsoassessedifCanovaimmunomodulatorthroughthe
Canova is standardized and authorized by competent agencies
enhancement of immunity can contribute to a reduction in
for medicinal application. Experiments were performed with
adverse effects ofanticancer treatment.
commercial Canova donated by ‘Canova do Brasil’, a Brazilian
company, which holds the international patent of this medicine
Methods
(www.canovadobrasil.com.br).
2.1 Nonhuman Primates Animals of CA, CLCA1 and CLCA2 groups, as well as one
animalofMNUgroup,weretreatedwithCanova.Theseanimals
36adultCebusapella(6–7yearsold)wereevaluated(2.7–3.6kg).
Animals were identified with microchips and were individually received 7 ml/g of Canova daily. Canova concentration was
determined according the study of Sato et al. in mice model [9].
housed in Centro Nacional de Primatas, Para´ State, Brazil. The
Dose distributions were calculated at the time of treatment.
animalswerefedahealthybalanceddietnotenrichedwithsodium
chloride and were weighed daily. In this study, the details of Canova solution was succussed before treatment and injected by
animal welfare and steps taken to ameliorate suffering were in slowinfusionintherightfemoralveinofC.apellainasingledose.
accordance with the recommendations of the Weatherall report,
‘‘The use of non-human primates in research’’. This study was 2.5 MNU treatment
approved by the Ethics Committee of Universidade Federal do The animals of the second model received oral fresh doses of
Para´ (PARECERMED002-10). MNU(N1517Sigma-Aldrich,USA)dailyfor940daysatadosage
According to a basic veterinary examination, all animals were of16mg/kgbodyweight.Theanimalsalsoreceiveddrinkwater
consideredhealthyatthetimeoffirstbloodsampling,endoscopy, containing MNU in light-shielded bottles daily. Water was
and ultrasound. This was confirmed by the animals’ behavior as restricted during MNU treatment and given ad libitum during
judged bytheveterinary check. Canovatreatment.
2.2 Experimental Design 2.6 Animal evaluation
36animalswererandomlyseparatedinsixgroupsandincluded Blood samples of all animals were collected for the determina-
in 2studied models: tion of hematimetric and leukocytic parameters, evaluation of
1o model: cell line inoculation. Negative Control (NC): 6 hepaticandrenalfunctions,andserummeasurementofC-reactive
controlC.apellasthatreceivedsalinesolutioninjectionsinsteadof protein (CRP), folic acid, and homocysteine on day 0 (baseline).
Canovaor celllineinoculation. During the treatment periods, the animals were inspected daily
and their clinical symptoms were recorded. Body weight was
– Canova group (CA): 6 C. apellas treated with 7ml/g of Canova determined and peripheral blood from the left femoral vein was
during14days.Theseanimalsdidnotreceivecelllineinoculation. collected forallserum analyses.
– Celllinegroup(CL):6C.apellasinoculatedwithgastriccancer Chemistry analysis included testing of levels of glucose, urea
cell line and that received saline solution injections instead of nitrogen, creatinine, total protein, albumin, globulin, total
Canova bilirubin, cholesterol, triglyceride, alanine aminotransferase,
– Cell line plus Canova during 10 days (CLCA1): 6 C. apellas aspartate aminotransferase, c-glutamyl transpeptidase, lactate
inoculated with gastric cancer cell line and after 5 days were dehydrogenase, creatine kinase, amylase, calcium, inorganic
treated with7 ml/g of Canovaduring 10days. phosphorus,sodium,potassium,andchloride.Clinicalhematology
included red blood cell count, hemoglobin, hematocrit, platelet
– CelllineplusCanovaduring14days(CLCA2):6C.apellasthat
count, white blood cell (WBC) and differential (segmented
received gastric cancer cell line inoculation and 7 ml/g of
neutrophil, lymphocyte, monocyte, eosinophil and basophile)
Canovaduring 14days(since day0).
counts.Methodsandreferencevaluesformaleadultanimalwere
2omodel:MNUtreatment. 6C.apellas treatedwithMNU. previously described[10].
After tumorigenesis, one animal received Canova treatment CRP was measured by turbidimetric assay as previously
(MNU group). described by Price et al. [11]. Serum level of folic acid was
PLoSONE | www.plosone.org 2 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
measured using chemiluminescent microparticle immunoassay Immunohistochemical analyses for MYC protein were per-
(CMIA) (Abbott System, USA). Serum homocysteine levels were formed on formalin-fixed, paraffin-embedded sections. Immuno-
measured with high pressure liquid chromatography (HPLC) histochemical staining was performed on the paraffin sections
(Betamed, Agilent 1100series,Chromosystems Reagent Kit). accordingtoCalcagnoetal.[15]withprimarymousemonoclonal
Forthefirststudiedmodel,ourresultsfocusmainlyintheblood antibodyagainstMYC(dilution1:50;DBS,USA).Positiveprotein
analyses of days 0 and 14. The analysis of folic acid and of expression was defined as clear nuclear imunostaining in more
homocysteineconcentrationwasalsopresentedonthe9thday.For than 10%of thecells.
the second carcinogenesis model, our results focus mainly in the
blood analyses ofthedays 0,90,120, 300,940and960. 2.9 Data Analysis
Inthefirstmodel(celllineinoculation),wefirstevaluatedthe
2.7 Tissue samples and gastric mucosa examination normaldistributionofalldatausingtheShapiro-Wilknormality
Biopsy samples of gastric normal and non-normal (e.g. non- test to determine subsequent use of appropriate tests for
atrophic and atrophic gastritis, metaplasia, neoplasia) gastric statistical comparison. Data that were not normally distributed
mucosa were collected by endoscopy. Lymphadenectomy was weretransformed(z-scoretransformation)foranalysissuchthat
performed tocollect axillary andinguinal lymphonodesamples. they followed a normal distribution. Analysis of variance in
Gastric mucosa alterations and tumor growth was followed by body weight, serum biochemistry values, hematological param-
endoscopy examination and ultrasonography. A pachymeter was eters, MYC expression, and copy number were performed by
used tomeasurethetumor biopsies. univariateGeneralLinearModel(GLM)followedbyBonferroni
Histologic analysis of gastric mucosa and axillary and inguinal post-hoc test. The effect size for GLM analyses was based on
lymphonodebiopsiesofCebusapellawereembeddedintoparaffin, EtaSquared(g2),inwhich0.15andbelowwasdeterminedasa
cut in5 mm sections,andstained byhematoxylin andeosin. smalleffectsize;0.16–0.40,mediumeffectsize;andabove0.40,
large effect size. Chi-square test was performed to compare
MYC immunostaining among groups. In these analyses, the
2.8 MYC expression and copy number analyses
confidence interval was 95% and p values less than 0.05 were
The MYC proto-oncogene has been described as a key in the
considered significant.
gastriccarcinogenicprocess[12]and,thus,itwasselecttoconfirm
In the second model (MNU treatment), only non-parametric
the presence of a gastric carcinogenesis process. For the first
tests torepeated measures were used due to the small number of
studied model, gastric biopsies of tumorfactions observed on the
samples. The Friedman test followed by Wilcoxon analysis with
9thdayaftercelllineinoculationwereusedtoevaluatetheMYC
Bonferroni’sadjustmentwereperformedtoanalysisofvariancein
expression and copy number, because only fibrotic lesions were
body weight, serum biochemistry values, hematological parame-
observed in the studied animals on the 14th day. For the second
ters, MYC expression, and copy number at days 0 (baseline), 90
studiedmodel,C.apellagastricsamplesatdays0,90,120,300,940
and120.Inthesestatisticalanalyses,oneanimalwasexcludeddue
and 960 were used to evaluate the MYC expression and copy
to its early death. In these analyses, p,0.016 was considered
number.
statisticallysignificant.Also,duetosmallnumbersofanimals,the
Fluorescent in situ Hybridization (FISH) was performed to
statistical data analysis concerning MNU treatment among days
determine MYC gene copy number according to the protocol of
300–940 and Canova treatment among days 940–960, was not
Pinkeletal.[13]withmodificationsintroducedfromCalcagnoet
possible andresultsare presented in adescriptive format.
al. [14,15]. Cells were hybridized with digoxigenin-labeled probe
(ONPON0824, Bioagency Biotechnology, Brazil) for MYC gene
Results
region(8q24)andnucleiwerecounterstainedwith49,6-diamidino-
2-phenylindole antifade. Positive MYC gene signals appeared as 3.1 First model – cell line inoculation
redspotsinnucleiandwerescoredusingthecriteriaofHopman Before this study, several methods of cell line inoculation were
etal. [16]. tested including intraperitoneal, subcutaneous, gavage, orthopic
Quantitative TaqMan Copy Number Variation (CNV) assays implantationandpercutaneous (datanotshown).Thepercutane-
(AppliedBiosystems,USA)usingreal-timequantitativePCR(RT- ousinoculationwastheonlyviathatresultedinthedevelopment
qPCR) were applied as a confirmation to FISH analysis. RT- of a tumorigenic process. We also previously detected that 1010
qPCR was performed using the FAM/MGB-labeled TaqMan cells was the cell number needed to increase the total tumorous
probeforMYCgene(Hs01764918_cn)andVIC/TAMRA-labeled percentage compared to small number of cells (108 e 109).
TaqManCNVRNAseP(#4403326)fortheinternalcontrol.RT- However, the high number of cells did not increase the total
qPCR reactions were performed in quadruplicate with genomic tumorous percentage (datanotshown).
DNA (gDNA) according to the manufacturer’s protocol (Applied In the first model, eighteen C. apella received percutaneous
Biosystems, USA). A known human gDNA (Promega, USA) was inoculation of 1010 of ACP03 cells. On the 9th day after cell line
used forcalibration. inoculation, these animals presented tumorfactions in the antral
MYC mRNA expression was evaluated RT-qPCR. First, region of the stomach (Figure 1A). The tumor volumes of CL,
complementaryDNAwassynthesizedusingHigh-CapacitycDNA CLCA1, and CLCA2 groups were similar and all animals were
Archive kit according to the manufacturer’s protocol (Applied abletoeliminatethetumors.Onthe14thday,theypresentedonly
Biosystems, Poland). All RT-qPCR reactions were performed in an inflammatory zone andfibroticlesions intheregion.
triplicate for both target gene (MYC -Hs00153408_m1) and Among the cell inoculation methods, we also evaluated the
internalcontrol(GAPDH-NM_002046.3).Relativequantification intraperitoneal inoculation. The intraperitoneal ACP03 inocula-
(RQ)ofthegeneexpressionwascalculatedaccordingtoLivakand tion induced lymphatic congestion in C. apella. After 48h of cell
Schmittgen [17]. In the present study, the NC group was line inoculation, the animals presented auxiliary and inguinal
designatedasacalibratorofthefirstmodelwhereasthebaseline lymphnodeenlargement.Lymphadenectomywasperformedand
values (from day 0) of the animals were used to calibrate the the histopathologic analysis showed only reactive lymphoid
second study model. hyperplasia.
PLoSONE | www.plosone.org 3 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
PLoSONE | www.plosone.org 4 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
Figure1.Ultrasonography,immunohistochemistryandFISHanalysisinC.apellagastriccarcinogenesismodels.A)ultrasoundimage
showinga‘‘space’’betweenthestomachwallwhereACP03celllinewasinoculatedanddevelopedatumor(2.5cm);B)ultrasoundimageshowinga
tumormassinaMNU-treatedanimalonthe940thday(5cm);C);MYCimmunoreactivityinatumorsampleofaCLCA1animal(4006);D);lackofMYC
immunoreactivityinnon-atrophicgastritissampleinaMNU-treatedanimal(4006);E)MYCimmunoreactivityinintestinalmetaplasiasampleofa
MNU-treatedanimal(4006);F)MYCimmunoreactivityinatumorsampleofaMNU-treatedanimal(4006);G)lymphocytesofahealthyC.apella
showingtwosignalsforMYCprobe(10006);H)normalgastricmucosacellsofNCanimalpresentingtwoMYCsignals(10006);I)neoplasticgastric
mucosaofCLanimalshowingMYCamplification(10006);J)intestinalmetaplasiasampleofMNU-treatedanimalpresenting1,2and3MYCsignals
(10006); K) tumor sample of MNU-treated animal showing MYC amplification (10006). Arrow indicates the space with gastric cancer cell line;
arrowheadindicatesanormalgastricwallthickness;circleindicatestheproliferativeprocess.
doi:10.1371/journal.pone.0021988.g001
During this study, the animals in the NC group presented andCA(p=0.003andp=0.011,respectively).TheCLgroupalso
normal levels of the biochemical and hematologic evaluated presentedhigherhomocysteinelevelsthananimalsfromCAgroup
parameters according to Riviello et al. study in C. apella [10]. To (p=0.028).However,onthe14thday,onlyanimalsfromCLCA1
our knowledge, there are no CRP, homocysteine, and folic acid grouppresentedhigherhomocysteinelevelsthananimalsfromNC
reference ranges determined for C. apella. However, the non- (p=0.026) and CLCA1 and CLCA2 groups presented higher
treated animals presented homocysteine and folic acid levels levelsthanCA(p=0.013andp=0.29,respectively)group,which
similar tothose described forhealthyMacaca fascicularis [18]. suggests some effects of cell line inoculation and of Canova
One week before thecell line inoculation, the C. apellas of CL, treatment(Figure2E,TableS1).Inthepresentstudy,weobserved
CLCA1andCLCA2wereimmunosuppressedbyasingledoseof that therange ofhomocysteinewas between 2.5–5.21mmol/Lin
50mg/kgofcyclophosphamide.However,onthedayofcellline healthyC.apella(n=36,onday0).Inaddition,weobservedthat2
inoculation, no significant difference was observed among these animals of CL, 4 of CLCA1 and 3 of CLCA2 presented
groups, nor between NC and CA groups, regarding the animal’s homocysteine levels higher than 5.21 mmol/L on the 9th and
baseline weight, biochemical, hematologic, folic acid, and 14th days.
homocysteine measurements. Concerning hematologic analyses, we observed a significant
Concerning the biochemical analysis, significant changes in alteration in leukocyte (F =10.506, p,0.001, by GLM test;
u(ttFhrri4eeg,a2lsy5tcu=nedir2tiired0ode.g1sge3(nr5Fo,4u,(p2Fp5,4s=,20o53.n0=30t5h31.3e6;.9g515342,7t=hp,,0dp.a70,y6.030(0F).01ilg0e,uv1bre;yelsGg2wA2Le=–MrCe0,.to8eTbs4sta3;eb)rglve2aend=Sd1a0)m..C9To8Rn2hP)ge, (e(gFFr24y,=t2h50r==o.6c412y.27t7.e)90,82(,8lFy,4m,2pp5,ph==0o6.c00.y4.0t00e105;5,4(g;F,2p425,2==5g00=2..60=5505081.2.);41c3g3o4,2u),=npt,0s.05aa.n0m06d0o)1,n;ghhgatahe2eme=mos0atgu.tl8ood9cbi8eriind)t,
4,25
Bonferroni post-hoc analyses demonstrated a significant increase groups on the 14th day (Figure 2F–J, Table S1). The Bonferroni
oftriglycerides,ureanitrogen,andCRPlevelinCL,CLCA1and
post-hocanalysesdemonstratedasignificantincreaseofleukocyte
CLCA2 compared to NC and CA groups (p,0.001, for all pair
andlymphocytecountinCA(p,0.001andp,0.001,respective-
wise comparisons). No significant difference in biochemical
ly), CL (p=0.017 and p,0.001, respectively), CLCA1 (p,0.001
measurements was observed between NC and CA groups. The
and p,0.001, respectively) and CLCA2 (p,0.001 and p,0.001,
increaseoftriglyceridelevelinthecell-lineinoculatedanimalswas
respectively) as compared to the NC group, which suggests that
insidethenormalreferencelevelaccordingtoRivielloetal.[10].
Canovaandcelllineinoculationaffectleukocyteandlymphocyte
On the other hand, abnormal levels of the other biochemical
levels.AlthoughCAgrouppresentedabnormallyhighlymphocyte
parameter were observed inthecell-line inoculated animals.
countaccordingtoRivielloetal.[10],theseanimalswereclinically
Toourknowledge,nopreviousstudyreportedthenormallevel
healthy.
ofCRPinhealthyC.apella.Inthepresentstudy,weobservedthat
We also detected that CL group presented a significant
the range of CRP levels were between 0.34–0.99mg/dL (n=36,
reduction of erythrocyte (p=0.006 and p=0.001, respectively
on day 0). The serum CRP level increased 5.7–13.6 folds due to
by Bonferroni analyses) and haemoglobin count (p=0.011 and
cell lineinoculation.
p=0.007,respectively)ascomparedtoNCandCAgroups,which
We observed that the levels of folic acid changed significantly
suggests a cell line inoculation effect that is improved by Canova
among the studied groups on the 9th (F =7.446, p,0.001, by
4,25 treatment.AccordingtoRivielloetal.[10],allanimalsoftheCL
GLM test; g2=0.544) and on the 14th day (F =4.056,
p=0.011;g2=0.394).Bonferronipost-hocanalysesde4m,25onstrated group were anemic. We also observed a significant reduction of
haematocritcountinCLandCLCA2groupsascomparedtoNC
asignificantreductionoffolicacidinCLandCLCA2groupthan
group (p,0.001 and p=0.022, respectively). The CL group also
NC (p=0.008 and p=0.003, respectively) and CA (p=0.03 and
p=0.009,respectively)onthe9thdayandinanimalsfromCLCA2 presented reduced haematocrit count as compared to the CA
group compared toNCgroup(p=0.031)on the 14thday, which group(p,0.001).AccordingtoRivielloetal.[10],5animalsofthe
CL and 1 of the CLCA1 group presented abnormal haematocrit
suggestsaneffectofcelllineinoculation(Figure2D,TableS1).In
level. The hematocrit count difference among groups again
thepresentstudy,weobservedthattherangeoffolicacidlevelwas
suggests a cell line inoculation effect that is in part improved by
13.29–18.84 nmol/dL in healthy C. apella (n=36, on day 0).
However,onthe9thday,4animalsofCL,4ofCLCA2and1of Canovatreatment.
CLCA1presentedlowerlevelsoffolicacid(lessthan13nmol/L)
aswellas3animalsofCL,3ofCLCA2,and1ofCLCA1onthe 3.2 Second model – MNU treatment
14th day. We treated six C. apella with MNU for a duration of
We observed that homocysteine levels changed significantly approximately 2.5 years. All animals developed pre-neoplastic
amongthestudiedgroupsonthe9thday(F =7.887,p,0.001, lesions andfive diedofdrug intoxicationbefore thedevelopment
4,25
by GLM test; g2=0.558) and on the 14th day (F =5.6, of gastric cancer. All animals presented non-atrophic gastritis on
4,25
p=0.002; g2=0.473). On the 9th day, CLCA1 and CLCA2 the 90th day. On the 110th day, one animal died from drug
groups presented higher homocysteine levels than animals from intoxicationandtheotherfiveanimaspresentedatrophicgastristis
NC(p=0.007andp=0.028,respectively,byBonferronianalyses) onthe120thday.Onthe134th,140thand290thdays,threemore
PLoSONE | www.plosone.org 5 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
PLoSONE | www.plosone.org 6 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
Figure2.Abnormalbiochemicalandhematologicmeasurementsinanimalsofthefirstcarcinogenesismodel.A)triglycerides;B)urea
nitrogen; C) C-reactive protein; D) leukocyte; E) lymphocyte; F) erythrocyte; G) haemoglobin; H) haematocrit; I) folic acid; J) homocysteine. NC:
negativecontrol;CA:Canovagroup;CL:animalsinoculatedwithACP03cellline;CLCA1:animalsinoculatedwithACP03celllineandtreatedwith
Canovaduring10days;CLCA2:animalsinoculatedwithACP03celllineandtreatedwithCanovaduring14days.N=6/group.*Significantlydifferent
fromNCgroup(p,0.05)onthe14thday.**SignificantlydifferentfromNCandCAgroups(p,0.05)onthe14thday.1SignificantlydifferentfromNC
group(p,0.05)onthe14thday.{SignificantlydifferentfromCAgroup(p,0.05)onthe9thday.{{SignificantlydifferentfromNCandCAgroups
(p,0.05)onthe9thday.
doi:10.1371/journal.pone.0021988.g002
animalsdiedduetodrugintoxication.Onthe300thday,thetwo On the 960th day, the survinving C. apella was submitted for
surviving C. apella presented intestinal metaplasia in gastric surgical removal of the tumor. This animal was clinically
mucosa. One animal died with symptoms of drug intoxication monitored for one year after the end of the experiment and he
on the 520th day and the last surviving animal developed did not show anycomplications resulting fromthetreatments.
intestinal-type adenocarcinoma in the antral region of stomach.
This tumor was observed by ultrasonography (Figure 1B) and 3.3 MYC copy number
endoscopy on the 940th day and this finding was confirmed by The MYC probe for FISH analysis was first tested in the
histopathologicanalysis.Noneoftheanimalsdevelopedanyother lymphocytes of a healthy C. apella. In C. apella (Figure 1G), the
tumor types. MYCprobehadsimilarefficiencyasthatobservedinourprevious
The deceased animals showed the typical symptoms of resultswith human cells [19].
intoxication: mydriasis, confusion, sleepiness, giddiness, loss of Table 1 shows the mean and standard deviation of the MYC
balance,tremor,hyperthermia,lowfoodconsumption,nonspecific copy number by FISH and qRT-PCR of the groups included
gastrointestinal symptoms (diarrhea and vomiting), urinary in the first carcinogenesis model. By FISH assay, the number of
retention, cutaneous eruptions, and caustic and ulcerative oral cells presenting 2 (F =2578.912, p,0.001, by GLM test;
4,25
lesions.Theyalsopresentedrenal,hepaticandrespiratoryfailure, g2=0.998), 3 (F =150.51, p,0.001; g2=0.960), 4
4,25
hypokalemia, chroniccholecystitis andsteatosis. (F =590.872,p,0.001;g2=0.99)and5ormoreMYCsignals
4,25
In the second model of carcinogenesis, we observed that the (F =117.013, p,0.001; g2=0.949) as well as high MYC
4,25
MNU treatment led to significant changes in triglycerides amplification (F =22.973, p,0.001; g2=0.786) was signifi-
4,25
(x2=7.6, df=2, p=0.0224, by Friedman test), urea nitrogen cantly different among the studied groups (Figure 1H–I). By
(x2=10, df=2, p=0.0067), phosphorus (x2=10, df=2, qRT-PCR, we also observed that the number of MYC copies
p=0.0067), alanine aminotransferase (x2=10, df=2, were significantly different among the studied groups
p=0.0067), total bilirubin (x2=8.4, df=2, p=0.015), creatinine (F =95.986, p,0.001, by GLM test; g2=0.939). The
4,25
(x2=8.4,df=2,p=0.015),CRP(x2=8.4,df=2,p=0.015),folic Bonferroni post-hoc analyses of FISH results also showed that
acid(x2=8.4,df=2,p=0.015)andhomocysteine(x2=10,df=2, thenumber of cellspresenting 3, 4, 5 or moreMYCsignals and
p=0.0067)levels(Figure3A–I,TableS2).However,nosignificant high amplification was significantly higher in CL, CLCA1 and
difference was confirmed by Wilcoxon test with Bonferroni CLCA2groupsthaninNCandCAgroups(p,0.001,forallpair
correction. The largest fold chance was observed in the CRP wise comparisons), confirming RT-qPCR results. These MYC
levels.After90daysofMNUtreatment,theCRPlevelsincreased signalnumberalterationswereobservedinthebiopsiesofallCL,
5.3–13 folds compared to baseline level. The folic acid CLCA1 and CLCA2 animals. No significant difference was
concentration was reduced more than 2-fold change and the observed between CL and CLCA groups, suggesting no Canova
homocysteine concentration increased almost 5-fold on the 120th effect in the MYC copy number.
day as comparedtobaseline levels (Figure 3). Table2showsthemedianandinterquartilerangeofMYCcopy
Concerning hematologic analyses, we observed a significant numberbyFISHandqRT-PCRinsamplesfromanimalstreated
alteration of leukocyte (x2=8.4, df=2, p=0.015), lymphocyte withMNU(secondmodelofcarcinogenesis).UsingtheFISHand
(x2=10, df=2, p=0.0067), neutrophil (x2=10, df=2, RT-qPCR assays, no significant difference was observed among
p=0.0067),erythrocyte(x2=8.4,df=2,p=0.015),haemoglobin biopsiesofday0,90and120afterWilcoxontestwithBonferroni
(x2=10, df=2, p=0.0067) and haematocrit (x2=10, df=2, correction.Onthe300thday,intestinalmetaplasiawasobservedin
p=0.0067) counts during MNU treatment (Figure 3J–O, Table the two surviving animals. They had approximately 30% of cells
S2).AlthoughnosignificantdifferencewasconfirmedbyWilcoxon with3MYCsignalsandabout9.5%ofcellswith4MYCsignalsas
test with Bonferroni correction, all animals reached abnormal determinedbyFISHandalmost3copiesbyqRT-PCR(Figure1J).
levelsofthesehematologicparametersfollowingMNUtreatment, TheanimalthatsurvivedtheendoftheMNUtreatmentshoweda
consistent withRiviello etal.[10]. continuousincreaseinthenumberofcellswithMYCamplification
TheanalysisofCanovatreatmentwasbasedontheobservation during gastric carcinogenesis. On the 940th day, the animal
ofitseffectsinonlyoneanimalthatdevelopedgastriccancer.After developedintestinal-typegastriccancerandhad46%ofcellswith
20 days of Canova treatment, the tumor volume did not change 3 or more MYC copies, including 5% of cells with high
(about 1 cm3). In addition, no change (less than 1.2 fold-change) amplification as determined by FISH and 3 MYC copies by
was observed between the 940th and 960th day concerning the qRT-PCR (Figure 1K). In the second carcinogenesis model,
biochemical,includingfolicacidandhomocysteine,measurements Canova treatment during 20 days did not appear to change the
inthesurvivinganimal.However,weobservedthatCanovaacted number of MYCcopies.
mainly on the hematologic measurements. The surviving animal
presentedmorethan2-foldincreaseinleukocyte,lymphocyte,and 3.4 MYC expression
erythrocyte counts after Canova treatment as well as a 1.4-fold Table1showsthemeanandstandarddeviationofMYCmRNA
increase in the neutrophil count. Canova treatment restored expressionanditsimmunoreactivityinbiopsysamplesofanimals
normal count of leukocyte, lymphocyte and neutrophil according of the first carcinogenesis model. All normal gastric samples (NC
Riviello etal.[10]. and CA groups) showed standard staining for MYC protein
PLoSONE | www.plosone.org 7 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
PLoSONE | www.plosone.org 8 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
Figure3.AbnormalbiochemicalandhematologicmeasurementsinanimalsMNU-treatedandCanova-treated.A)triglycerides;B)urea
nitrogen;C)phosphorus;D)alanineaminotransferase;E)totalbilirubin;F)creatinine;G)C-reactiveprotein;H)leukocyte;I)lymphocyte;J)neutrophil;K)
erythrocyte;L)haemoglobin;M)haematocrit;N)folicacid;O)homocysteine;P)weight.N=6onthe0–90thdays(non-atrophicgastritis);N=5onthe120th
day(atrophicgastritis);N=2onthe300thday(intestinalmetaplasia);N=1onthe940th(gastriccancerdevelopment)and960th(Canovatreatmenteffect).
doi:10.1371/journal.pone.0021988.g003
expression. We also observed an association between MYC primatesofferausefulmodelforcancerresearchandotherbasic
immunoreactivity and cell line inoculation (x2=12, df=1, research into genetic and immunopathogenesis mechanisms as
p=0.001). All the tumor biopsies of animals inoculated with well as for the development and validation of new therapies for
ACP03(CL,CLCA1andCLCA2groups)presentedMYCprotein severaldiseases.Cebusapella,aNewWorldmonkey,isaconvenient
overexpression(Figure1C).WedidnotobserveanincreaseinMYC modelforbiomedicalstudiesbecausetheycanbeeasilyhousedin
mRNAexpressioningastricmucosaofCAgroupcomparedtoNC Primate Research Centers due to their flexibility, opportunism,
group.TheMYCexpressiondifferedamongCA,CL,CLCA1and adaptability, and small size. To our knowledge, this is the first
CLCA2groups(F =885.646,p,0.001,byGLMtest;g2=0.993). studythatestablishedagastriccarcinogenesismodelinCebusapella.
3,20
CL,CLCA1andCLCA2groupspresentedahigherMYCexpression Gastric adenocarcinoma is divided mainly into intestinal and
than the CA group (p,0.001). A six-fold increase in mRNA diffuse types according to Laure´n classification [20]. The
expression was observed in tumor samples of CL, CLCA1 and intestinal-type gastric cancer progresses through a number of
CLCA2 groups relative to the NC group. No difference in MYC sequential steps beginning with atrophic gastritis followed by
expression was observed among CL, CLCA1 and CLCA2, which intestinal metaplasia, intraepithelial neoplasia, and carcinoma
suggestsCanovadoesnoteffectMYCexpression. [21].Ontheotherhand,diffuse-typegastriccancergenerallydoes
Table 2 shows the median and interquartile range of MYC not evolve from precancerous lesions [2]. Here, we induced
mRNA expression and its immunoreactivity in samples from intestinal-type gastric adenocarcinoma in C. apella by treatment
animals treated with MNU (second model of carcinogenesis). By with MNU carcinogen and by human cancer cell line (ACP03)
qRT-PCR,weobservedacontinuousincreaseinMYCexpression inoculation.
during MNU-induced gastric carcinogenesis. Although negative Celllinesderivedfromhumancancersareusefultounderstand
MYC immunoreactivity was observed (Figure 1D), the mRNA thechromosomalalterationsandothermolecularalterationsinthe
expressionwasabout2-foldhigheron120thday(atrophicgastritis) carcinogenesisprocess.Celllinesarealsoanimportanttoolforthe
comparedtobaseline.Onthe300thday,weobservedanabout3- study of anticancer treatments in in vitro and in animal xenograft
fold increase ofMYCmRNAexpression inthemetaplasia lesions models.Fortheinductionofgastriccancerbycelllineinoculation
of the two surviving animals in this period relatively to their inC.apella,weinitiallytestedfourdifferentgastriccancercelllines.
baseline level. On the 940th day, the surviving animal presented ACP02,AGP01andPG100celllinesdidnotinducetumorsinC.
about 5-fold increase in MYC mRNA expression relative to its apelladespitethemethodofinoculation.OnlytheACP03cellline
baselinelevel.MYCnuclearimmunoreactivitywasobservedonly wasabletoinducegastriccancerintheanimals.Thetumorigenic
intheintestinalmetaplasiaandgastriccancerbiopsies(Figure1E– potential of a cell line is attributed to the presence of a subset of
F). No change in MYC expression was observed in the tumor cellscalledcancerstemcells,whichhavecapabilitytorecapitulate
biopsy ofthisanimalafter 20daysof Canovatreatment. thedevelopmentoftheoriginaltumorsinvivo.Thestudyofhuman
cancer stem cells largely relies on models of xenograft transplan-
Discussion tationintoimmunodeficientmice[22].Probably,onlyACP03cell
linehas cancer stemcellproprieties.
4.1 C. apella gastric carcinogenesis One week before ACP03 inoculation, animals of CL, CLCA1
Spontaneous tumors have been reported in nonhuman and CLCA2 groups received a single dose of 50mg/kg of
primates, usually due to the aging process [3]. Nonhuman cyclophosphamide. Our group previously observed that white
Table1.Immunohistochemistry,relativequantitationofmRNAMYCexpressionandMYCgenecopynumbervariationbyTaqman
andfluorescence insituhybridization intumor biopsies ofanimalsincluded inthe firstcarcinogenesis modelon the9th day of
treatments.
CNV
mRNAexpression (number
Group IHQ (mean±SD) [mean±SD]) NucleiexhibitingMYCsignals(mean±SD)
1signal 2signals 3signals 4signals $5signals HA
NC Negative - 2(2.0360.04) 3.1761.17 196.8361.17 - - - -
CA Negative 20.5160.22 2(2.0460.56) 3.3361.21 196.3361.37 0.3360.52 - - -
CL Positivea 6.4560.24b 5(4.6560.43)a 5.8361.94 19.8365.98a 49.17610.57a 69.0064.29a 37.3366.35a 18.3367.78a
CLCA1 Positivea 6.2460.29b 5(4.660.34)a 3.561.76 17.1766.34a 45.3364.32a 70.1766.911a 39.1767.68a 24.6766.4a
CLCA2 Positivea 6.0160.35b 4(4.2460.52)a 4.3362.34 17.6765.65a 47.561.64a 72.6763.14a 38.8363.82a 1968.69a
IHC:immunohistochemistry;CNV:copynumbervariation;HA:highamplification;NC:negativecontrol;CA:Canovagroup;CL:animalsinoculatedwithACP03cellline;
CLCA1:animalsinoculatedwithACP03celllineandtreatedwithCanovaduring5days;CLCA2:animalsinoculatedwithACP03celllineandtreatedwithCanovaduring
9days.
aSignificantlydifferentfromNCandCAgroups(p,0.05).
bSignificantlydifferentfromCAgroup(p,0.05).
doi:10.1371/journal.pone.0021988.t001
PLoSONE | www.plosone.org 9 July2011 | Volume 6 | Issue 7 | e21988
GastricCarcinogenesisModelinNonhumanPrimates
Table2.Immunohistochemistry,relativequantitationofmRNAMYCexpressionandMYCgenecopynumbervariationbyTaqman
andfluorescence insituhybridization inbiopsies ofMNU-treated animals.
mRNAexpression CNV
(median± (number[median±
Treatment IHQ interquartilerange) interquartilerange]) NucleiexhibitingMYCsignals(median±interquartilerange)
1signal 2signals 3signals 4signals $5signals HA
Baseline Negative - 2(2.0460.32) 361.5 19663 161 - - -
MNU/90thday Negative 1.5261.06 2(1.8760.55) 361.5 19463 2.561.75 - - -
MNU/120thdaya Negative 2.2461.73 2(1.9260.74) 460.5 18865 564 262 - -
MNU/300thdayb Positive 3.3560.09 3(2.8560.11) 3.560.5 115.560.5 59.561.5 1963 2.562.5 -
MNU/940thdayc Positive 4.75 3(3.12) 3 105 61 16 5 10
Canova/960thdayc Positive 5.04 3(3.04) 4 109 52 18 9 8
IHC:immunohistochemistry;CNV:copynumbervariation;HA:highamplification.
aFiveanimals;
bTwoanimals;
cOneanimal.
doi:10.1371/journal.pone.0021988.t002
blood cell count decrease after 4 days of cyclophosphamide antral stomach region of one animal. MNU induced intestinal-
treatment (unpublished observation). However, total tumor type gastric carcinogenesis in C. apella, presented the sequential
remission was observed after 14 days of ACP03 inoculation steps similar to those described for humans [21]. Therefore, this
probably due to immune system activation. We did not give a modelallowsforthestudyoftheevolutionofintestinal-typegastric
second cyclophosphamide dose, because in a previous study we cancer,theidentificationofgenesevolvedintheearlystepsofthe
observed that it was lethal to 50% of C. apella (unpublished carcinogenesisprocess,andthedeterminationofspecifictargetsof
observation). Furthermore, the ACP03 cell line may have low gastric neoplastic transformation.
tumorigenic potential and therefore, did not present the same The multistep process of intestinal-type carcinogenesis was also
proliferationandmetastaticpatternsoftheoriginalhumantumor supportedbythedetectionofanincreasedofmRNAexpressionand
intheseanimals.Thismodelcanbeappliedtoothergastriccancer MYCcopynumberduringthesequentialsteps,whichbeginswith
cell lines, especially with high tumorigenic potential, in C. apellas atrophic gastritis and is followed by intestinal metaplasia and
after their immunosupression. This may be useful for the carcinoma. FISH assay and CNV analysis by qRT-PCR showed
identification of hematological and biochemical markers that can that normal mucosa, non-atrophic gastritis, and atrophic gastritis
helptoguide immunotherapy cancer treatments. samples of MNU-treated C. apellas presented mainly cells with 2
The presence of gastric tumor due to ACP03 inoculation was MYCcopies.Intestinalmetaplasiapresentedcellswith3MYCcopies
confirmed by MYC deregulation. The MYC proto-oncogene has asaclonalalteration(30%ofcells)andthisgeneamplificationwas
been described as a key in the gastric carcinogenic process [12]. observedinabout40%ofcells.Ingastriccancersamplesinducedby
Groups of genes involved in cell cycle regulation, metabolism, MNU,almost50%ofcellspresentedMYCamplification,including
ribosomebiogenesis,proteinsynthesis,andmitochondrialfunction 5%ofcellswithMYChighamplification.ThefindingsinC.apella
are over-represented in the MYCtarget genenetwork. MYCalso gastric cancer corroborate our observations in human gastric
consistentlyrepressesgenesinvolvedincellgrowtharrestandcell carcinogenesis, in which the presence of MYC amplification,
adhesion and also has a direct role in the control of DNA includinghighamplification,wasdetectedinallhumanintestinal-
replication [23]. MYC amplification has been observed in gastric typegastriccancer[14,15,25,26,27,28]andasignificantincreaseof
cancer cell lines and primary stomach tumors [8,14,15,19,24, MYC copy number was seen with the evolution of human
25,26,27,28,29,30]. carcinogenesis process: normal mucosa, intestinal metaplasia, and
In the first carcinogenesis model, we observed MYC immuno- gastriccancer[27].
reactivity, amplification (more than 3 MYC copies) and mRNA MYC immunoreactivity was only observed in the intestinal
overexpression in the tumor biopsies. Four MYC copies was the metaplasia and cancer samples, corroborating our previous study
most frequent copy number alteration in the tumor biopsies with human samples [27]. MYC protein overexpression was
supporting our FISH findings in the ACP03 cells culture at the previously detected in intestinal metaplasia and neoplastic tissue
85th passage [31]. from all patients with intestinal type gastric cancer [15,27,28].
Inthesecondstudiedmodel,weinducedgastriccarcinogenesis Immunohistochemistry results demonstrated that clonal MYC
byMNUtreatment.MNUaccumulationleadstothedevelopment amplification isnecessary toinduce itsprotein immunoreactivity.
ofseveraltypesoftumorsinthedigestivetract,i.g.inoralcavity, To our knowledge, this is the third study to describe gastric
larynx,pharynxandmainlyesophagusandstomachofnonhuman adenocarcinomainexperimentalmodelinmonkeys.Takayamaet
primates[3,32,33].MNUinducespre-neoplasticlesionsbeforethe al. reported 2 nonhuman primates of the Old Word with gastric
development of intestinal type gastric adenocarcinoma [34], adenocarcinoma after 10 years of continuous MNU treatment
usually in antral stomach region of treated animals [35]. In the (10 mg/kg)[3].ThesmallsizeofC.apella,aNewWorldmonkey,
presentstudy,allMNU-treatedC.apellaspresentedpre-neoplastic may contribute to the faster development of gastric adenocarci-
lesions: non-atrophic gastritis (6 animals), atrophic gastritis (5 noma compared to the Old-World monkeys treated with MNU.
animals) and intestinal metaplasia (2 animals). We also observed Another previous study in literature reported intraepithelial
the development of intestinal-type gastric adenocarcinoma in the neoplasia induced by ethyl-nitro-nitrosoguanidine treatment in
PLoSONE | www.plosone.org 10 July2011 | Volume 6 | Issue 7 | e21988