iMedPub Journalshttp://www.imedpub.com/

Journal of Rare Disorders: Diagnosis & Therapy ISSN 2380-7245

2016Vol. 2 No. 2: 14

1

Research Article

DOI: 10.21767/2380-7245.100043

© Under License of Creative Commons Attribution 3.0 License | This article is available from: //www.raredisorders.imedpub.com/

Monica L Calicchio1, Naren R Ramakrishna2, Tucker Collins1,3, Keith L Ligon1,3-5 and Ali G Saad1,3,4

1 Department of Pathology, Children’s Hospital Boston, Boston, USA

2 RadiationOncology,BrighamandWomen’s Hospital, Boston, USA

3 HarvardMedicalSchool,Boston,USA4 Department of Pathology, Brigham and

Women’s Hospital, Boston, USA5 CenterforMolecularOncologic

Pathology,Dana-FarberCancerInstitute,Boston, USA

Corresponding author: Ali G Saad

ali.saad@bmcjax.com alisaad2007@hotmail.com

DepartmentofPathology,BaptistHealth/MDAnderson,CancerCenterandWolfsonChildren’sHospital,Jacksonville,FL32202,USA.

Tel: 9042022966

Citation: CalicchioML,RamakrishnaNR, Collins T, et al. Genome-Wide TranscriptionalProfilingofEpendymoma:InsighttoPrognosticIndicators J Rare Dis DiagnTher.2016,2:2.

Genome-Wide Transcriptional Profiling of Ependymoma: Insight to Prognostic Indicators

AbstractPediatric ependymomas are characterized by unpredictable biological behaviorandprognosis.ThepurposeofthisstudyistoidentifygenesthatareselectivelyexpressedinWHOgradeIIIependymomas,relativetogradeIIependymomas.Wehypothesizethatthesegenesplayaroleintheirbiologicprogressionandmayhelpinpredictingtheirclinicalbehavior.Paraffin-embeddedtissuefromthreegradeIIandthreegradeIIIependymomaswereisolatedbyLaserCaptureMicro-dissection.RNAwasextracted,amplifiedandhybridizedtoAffymetrixHumanX3PGeneChipArrayscontainingprobesetsthatdefine47,000humangenes.Statisticalt-testandz-scoreanalysisofdifferentiallyexpressedgeneswasperformedtoidentifygroupsoffunctionallyrelatedgenes.Genome-widetranscriptionalprofilingandz-scoreanalysisrevealedincreasedexpressioninasetofgenesinvolvedinDNAdamagerepair,mitosisandcellcyclecontrolingradeIII,relativetogradeIIependymomas.These genes included components of the G2M DNA damage checkpoint (e.g.CDC2andcyclinB1(CCNB1))andthemitoticspindlecheckpoint(e.g.,BUB1andsecurin(PTTG1)),kinetochoreandcentromereassociatedproteins (e.g.survivin(BIRC5)),andothergenesrecentlyidentifiedinoncogenicsignalinginglioblastoma(e.g.abnormal spindle-likemicrocephalyassociated (ASP)).This study identifiessubsetsofgeneshighlyexpressedingradeIIIependymomascomparedtogradeIIependymomasaspotentialprognosticindicators.

Keywords: Ependymoma; Prognostic indicators; Gene expression; Grade IIependymomas;GradeIIIependymomas

Received: April15,2016; Accepted: May03,2016;Published: May 10, 2016

IntroductionPediatric ependymomas are a heterogeneous group ofneoplasmswith clinical and biological behavior that is difficultto predict. Ependymomas represent the 3rd most frequentintracranialbraintumorinchildren(afterpilocyticastrocytomasandmedulloblastomas), accounting for 6% to 12%of pediatricbraintumors[1].Nonetheless,theyarerelativelyrareandfewerthan150casesareseenannuallyintheUnitedStatesinchildrenlessthan19yearsofage[2].Despiteadvancesinmanagementand treatment, the prognosis of ependymomas remains poor. The five-year progression-free survival ranges between 30%and60%withapoorerprognosisforthosewithresidualdiseaseaftersurgery,anaplastichistologicfeaturesorageyoungerthanthree years. Various investigators have reported that 80%-90%ofependymomadeathswereattributedtotumorprogressionattheprimarysite[3].

Thereisalackoflargeprospectiveclinicaltrialsaimedatbetter

definingreliablefactorsofprognosticsignificance.Retrospectivestudiesaddressingfactorssuchasage,location,gender,extentofresection,andhistologicgradeasprognosticfactorshaveyieldedcontroversialresults.Manyauthorshavedisputedtheusefulnessofgradingependymomas[3-8].Invirtuallyallstudies,theextentof tumor resection (complete vs. incomplete) significantlyinfluenced the clinical outcome [9-16]. However; in others,location appeared to represent themost important prognosticfactor, regardless of the histologic grade [15, 16]. Given thiscontroversy,itappearsthatidentifyingmarkersthatcanpredicttheclinicalbehaviorofependymomasinthepediatricagegroupisnecessary.

Ependymomasarecentralnervoussystemtumorsthatarisefromtheependymalcellsliningthecerebralventriclesandthecentral

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245

2016Vol. 2 No. 2: 14

2 This article is available from: //www.raredisorders.imedpub.com/

canalof thespinalcord [17-19].Recentevidencesuggests thatependymomasariseasaconsequenceofaberrantdevelopmentin which the bulk of the malignant cells are maintained by arare fractionof cancer stemcells [20]. Thesecancer stemcellsarephenotypicallysimilartonormalradialglialcells,butexhibitdysfunctional patterns of self-renewal and differentiation.Although about half of ependymomas appear karyotypicallynormal, a variety of chromosomal abnormalities have beendetected in the tumors [21]. However, molecular studiesanalyzingthegeneticalterationsinependymomashavefailedtoidentifytheclinicallyrelevantgeneticanomalies.

To expand existing information on the molecular features ofependymoma,weusedLaserCaptureMicro-dissection(LCM)todefineandisolatesmallpopulationsofependymomacellsfrombothWHOgradeIIandgradeIIIlesions.Thisallowsustoefficientlyenrichthe“signal”fromhighlypuretumorcellareasanddiminish“noise” generated by the bulk of the cells in a heterogeneoustumor.BycomparingglobalpatternsofgeneexpressioninWHOgrade II and grade III lesions from various anatomic locations,wesoughttofindgenespreferentiallyexpressedbyallgradeIIIlesions,relativetogradeIIlesions.Forasubsetofthegenes,thedatawereconfirmedontheproteinlevelbyimmunoperoxidaseanalysis. This study identifies a module of related genes thataremorehighlyexpressed ingrade III ependymomas,manyofwhichinteracttoregulatethemetaphasetoanaphasetransition.Thefindingssupportapossiblemutatorroleforthesegenesinependymoma development, suggesting that such tumors aredriven in part by genomic instability associated with variantexpressionofkinetochoreandcentromereassociatedgenes.

Materials and MethodSelection of casesFormalin-fixedparaffin-embedded(FFPE)humantissuesamplesfromnewlydiagnosedependymomas(3grades IIand3gradesIII)wereobtainedfromthefilesoftheDepartmentofPathologyatBostonChildren’sHospitalfollowinganIRB-approvedprotocol.CasesweregradedaccordingtoWHOcriteria.ExamplesofgradeIIandgradeIIItumorsareshowninFigures 1A and 2Arespectively.

RNA integrity assessmentInaninitialqualitycontrolstep,FFPEtissueswerescreenedforRNAquantityandqualityusingaquantitativereal-timePCRassaywithprimersdesignedtotwoampliconsonthebeta-actingene,followingthemanufacturer’srecommendedprotocol(MolecularDevices,MountainView,CA).The3’ampliconisdesignedfromaregionroughly100ntandthe5’ampliconroughly400ntfromthepolyAtail,respectively.Theassumptioninthisscreenisthatthe 3’ amplicon represents an indication of RNA quantity anda ratio of 3’:5’ amplicons represents a qualitative indicator oftranscript length and “transcribability”. The sample representsthe average status (i.e., length and “transcribability”) of otherRNA molecules in the same block. This assay was performedon an Mx3000p real time PCR thermal cycler (Stratagene,LaJolla, CA) using the Brilliant SYBR Green QPCR Master Mix(Stratagene). Dilutions of Universal Human Reference RNA(Stratagene) are run from which a standard curve is used to

estimatetheamountofRNAinthesample.Theassaymeasuresthe average beta-actin cDNA length by quantification of thePCRproductyieldfromthe3’endandcomparesthisyieldtoarelatively5’sequence.Thefollowingprimersequencesareused:3’ primers: HBAC1650: 5’-TCCCCCAACTTGAGATGTATGAAG-3’,HBAC1717: 5’-AACTGGTCTCAAGTCAGTGTACAGG-3’; 5’ primers:HBAC1355: 5’-ATCCCCCAAAGTTCACAATG-3’ and HBAC1472:5’-GTGGCTTTAGGATGGCAAG-3’.cDNAgeneratedfromtheuRNAwasseriallydilutedwithpolyI(Sigma-Aldrich,St.Louis,MO)andstandard curves (log uRNA amount vs. Ct) were generated foreachof theprimers sets. The standard curve consistedof fourstandards:100ng,10ng,1ngand0.1ng.cDNAgeneratedduringthe 1ststrandsynthesisreactionservedasthe100ng standard.

Figure 1 ValidationofgenecandidatesbyimmunohistochemistryinarepresentativeWHOGradeIIependymoma.Fromtop,leftpaneltobottom,rightpanelare:(A)H&E(B)Mib1(C)TOPOII(D)EZH2(E)CyclinB1(F)Securin(G)Mitosin(H)Survivin.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245v

2016Vol. 2 No. 2: 14

3© Under License of Creative Commons Attribution 3.0 License

RNAquantityderived from the standard curveof the 3’primerset (HBAC1650, HBAC1717)was used to estimate the quantityofRNAinthesample.TheratiooftheRNAyieldobtainedfromboth setsofPCRprimers is the3’/5’ ratioandwasusedasanindicationofRNAquality.Forexample,ifthecDNAcontainsboththe3’and5’targetsequences,the3’/5’ratiowouldbeabout1.Cross-linkedormodifiedRNAswouldhavea ratiogreater than1.Basedonthis ratio,anestimationof thequalityofRNAwasmade.RNAsyieldingacceptable3’to5’ratioswereselectedformicrodissection.

Laser capture micro-dissectionThree WHO grade II and three WHO grade III ependymomaswithuseableRNAwere identified inthequalitycontrolscreen.MultipleFFPEsectionswerecutoftheblockscontaininglesional

tissueat7μm thickness,deparaffinized, stainedwith toluidineblueanddehydratedfollowingmanufacturer’sprotocol(ParadiseReagent System Kit, Molecular Devices). Tumoral tissue wasmicrodissected from PENmembrane glass slides using the UVcuttinglaserandIRcapturelaserontheVeritasMicrodissectionInstrument(MolecularDevices).

RNA isolation, biotin-labeled cRNA transcription and microarray hybridizationFollowingmicro-dissection,RNAwasextractedfromthecapturedcellsbyincubatingdissectedtissueintheRNAextractionbuffer(Paradise Reagent System Kit,Molecular Devices) overnight at50ºC. RNA isolation was performed following manufacturer’sprotocol using theMiraCol Purification Column as part of theParadiseReagentSystemKitandsampleswereelutedinto12μlelutionbuffer.RNAwasamplifiedaccordingtothemanufacturer’ssuggestedprotocol(ParadiseReagentSystemKit).

Samples were labeled with biotinylated probes using theBioarray High Yield transcription kit following manufacturer’sprotocol (Enzo Biochemical, New York, NY). The concentrationof the biotin-labeled cRNAwas determined byUV absorbanceutilizing a Bio-Tek UV Plate Reader (Bio-Tek Instruments,Winooski, VT). In all cases, 20 μg of each biotinylated cRNApreparation was fragmented, assessed by gel electrophoresisand placed in a hybridization cocktail containing hybridizationcontrols as recommended by themanufacturer. Samples werethen hybridized to the Affymetrix Human X3P GeneChip Arrayat 450C for 24 h. Microarrays were washed and stained usingthe appropriate protocol for the Human X3P GeneChip Arrayon a Model 450 Fluidics station (www.affymetrix.com). TheFluidicsstationprocessiscontrolledbytheAffymetrixGeneChipOperatingSoftware(GCOS).

TheAffymetrixHumanX3PGeneChipArrayisdesignedforwhole-genomeexpressionprofilingofRNAfromformalin-fixedparaffinembedded samples. The target sequences on the X3P Arrayare identical to those used for designing the Human GenomeU133Plus2.0GeneChipArray, fora totalof47,000 transcriptswith61,000probe sets, although theprobeson the two typesof arrays are significantly different. The probe selection regionon theX3PArray is restricted to the300bpat themost3’endof the transcripts. In contrast, the standard Affymetrix designselectsprobesetswithintheregion600basesproximal to the 3’ ends(www.affymetrix.com).TheGeneChip©X3PArrayisusedforwhole-genome expression profiling of formalin-fixed, paraffin-embedded samples.

Data analysis and statisticsImagesfromthescannedchipswereprocessedusinganAffymetrixModel7000scannerwithautoloader(www.affymetrix.com).TheAffymetrix GCOS v1.3 operating system (www.affymetrix.com)controlstheModel7000scanneranddataacquisitionfunctions.Imagefilesweredownloaded,importedandanalyzedusingtheGeneSifterAnalysisEdition(www.genesifter.net).Statisticalt-testanalysiswasperformedinwhichapairwisecomparisonwasmadebetween the ependymoma, WHO grade II vs. ependymoma, WHOgradeIII.Datawasnormalizedtothemean.Resultswere

Figure 2 ValidationofgenecandidatesbyimmunohistochemistryinarepresentativeWHOGradeIIIependymoma.Fromtop,leftpaneltobottom,rightpanelare:(A)H&E(B)Mib1(C)TOPOII(D)EZH2(E)CyclinB1(F)Securin(G)Mitosin(H)Survivin.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245

2016Vol. 2 No. 2: 14

4 This article is available from: //www.raredisorders.imedpub.com/

filteredmorestringentlybyimposingathresholdcut-offof3.0orgreaterfold-changeinexpressionandaqualitycallof1(P)inallreplicatesofatleastonegroup.GeneSifterusesGeneOntology(GO)reportsandz-scorestosummarizethebiologicalprocesses,molecularfunctionsorcellularcomponents,aswellastheKEGG(Kyto Encyclopedia of Genes and Genomes) pathways [22]associatedwithagenelist.Thez-scoreiscalculatedbysubtractingtheexpectednumberofgenesinaGOtermmeetingthecriterionfromtheobservednumberofgenesanddividingbythestandarddeviation of the observed number of genes [23]. Z-scores canthenbeusedtoidentifyGOtermsthataresignificantlyoverorunderrepresentedinagenelist.Toassessthevariabilityingeneexpressionintheblocksofformalinfixedandparaffinembeddedlesions, independent LCM procedures, RNA isolations andmicroarrayhybridizationswerecompletedfromthegradeIIandgradeIIItumorsandtheresultswerecompared.Byplottingtherelative intensitiesof eachof thegenes fromboth samplesonthesamegraph,itispossibletodirectlycomparethedata(Figure S1).ThePearsonregressionanalysissuggeststhattheprocedureisreproducible.

Immunohistochemistry on tumor sectionsFFPE tissue sections were mounted on microscope slides.Immunohistochemical staining was optimized using a VentanaDiscoveryXTautomatedimmunohistochemistryslideprocessingplatform following the manufacturer’s instructions. Followingthe Closed Loop Assay Development (CLAD) protocol (VentanaMedicalSystems,Tucson,AZ),antibodieswereoptimizedusingeithertheOmniMapDABanti-Mouse(HRP)oranti-Rabbit(HRP)detectionkit(VentanaMedicalSystems).Antibodiesusedinthestudy are provided in the Online Data Supplement (Table S1). For each antibody, standard quality control procedures wereundertaken to optimize antigen retrieval, primary antibodydilution,secondaryantibodydetectionandotherfactorsforboth“signalandnoise”.Immunohistochemistryslideswereevaluatedby two neuropathologists (AGS and KLL). For each antibody,a labeling index (LI) was rendered by dividing the number ofpositive tumor cells by the total number of tumor cells. Theareawiththehighestpositivetumorcellswasselectedfor thispurpose.GraphPadsoftware(GraphPadSoftware,SanDiego,CA)wasusedtocalculatethepvalues.

Results and DiscussionBy immunohistochemistry, MIB-1, survivin, mitosin, securin,EZH2,TOPOIIandCdc2labelingindiceswereallhigheringradeIIIependymomasthaningradeIIependymomas.Thisdifferencewas statistically significant and validatedmicroarrayfindings. PvaluesarepresentedinTable 1.

Differentially expressed genes in ependymoma Pairwise analysis was used to identify differentially expressedgenesinthreeWHOgradeIIependymomas,relativetothreeWHOgrade III ependymomas. This analysis combines a fold-changecutoffandcomparisonstatisticstogeneratealistofdifferentiallyexpressed genes. Utilizing this approach, 271 genes (84 up-regulatedand187down-regulated)showsignificantdifferentialregulation,definedashavingelevated(>3fold)ordiminished(<3

fold)expressionintheWHOgradeIIItumorsrelativetogradeIItumors.ConsiderTable S2 forthecomplete listofdifferentiallyexpressedgenes.Bycalculatingz-scores, it ispossibletoobtaingeneontologylistingsforbiologicalprocess,molecularfunctionand cellular component. These ontology listings depict genesthatareeitherover-orunder-represented in theoverall listofdifferentiallyexpressedgenes(z-scoregreaterthan2orlessthan-2,respectively)[23].Thez-scoreanalysisofthebiologicalprocesscategories of differentially expressed genes was striking andrevealedacollectionof19BiologicalProcessGeneOntology(GO)groups involved inmitosis and theDNA replication checkpoint(Table 2).Similarly,ninestatisticallysignificantCellularComponentGOgroups focusedon thepericentric regionof the chromosomeandthekinetochore(Table 2).ThesetofgenescontainedwithintheseGOgroupsoverlapandformamoduleofgenes involvedinmitotic cell division and cellular proliferation (Table 3). The focusesofourstudyweregeneswithinthismoduleofmitosis/cellcyclethathadincreasedexpressioningradeIIIependymoma.

Cell proliferationGrade III ependymomas selectively express genes involved incellularproliferation.Notunexpectedly,gradeIIIependymomasexpress the gene for the antigen identified by themonoclonalantibodyKi67(MKI67)(Table 3 and Figures 1B and 2B).Expressionof this gene is increased 3.84 fold in grade III tumors, relativeto levels found inthegrade II tumors (Figure 3).Thisfinding isconsistentwithreportsbymanyothersandhelpstovalidatethegeneexpressionanalysisfromLCMsamplesofformalinfixedtissue.Expressionofthymidylatesynthase(TYMS),anenzymethat iskey

Antibody Tissue Median ± SE Mean Range P-value

MIB-1(Ki67)Normal brain 0 0 0

0.02GradeII 7 ± 2.5 8.1 1-25GradeIII 15 ± 1.1 15.4 11-20

SurvivinNormal brain 0 0 0

0.04GradeII 3 ± 0.7 3.3 1-8GradeIII 11 ± 2.5 9.7 2-20

MitosinNormal brain 0 0 0

0.008GradeII 2 ± 0.4 2.2 1-5GradeIII 10 ± 2.4 12 5-25

SecurinNormal brain 0 0 0

0.03GradeII 2 ± 0.6 2.7 1-7GradeIII 10 ± 2.4 9.5 3-20

EZH2Normal brain 0 0 0

0.04GradeII 7 ± 2.9 10.8 3-27GradeIII 25 ± 5.1 24.8 9-40

CyclinB1Normal brain 0 0 0

0.052GradeII 1 ± 0.4 1.3 0-4GradeIII 3±0.8 3.5 0-7

TOPOIINormal brain 0 ± 1 1 0-3

0.024GradeII 3 ± 0.6 3 1-7GradeIII 9 ± 1.7 8.1 3-15

Cdc2Normal brain 20 ± 4.4 21.6 15-30

0.03GradeII 3 ± 4.6 7.8 1-45GradeIII 20±8 28.3 10-65

Table 1 Validationof gene candidatesby IHCand their correspondingP-value.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245v

2016Vol. 2 No. 2: 14

5© Under License of Creative Commons Attribution 3.0 License

Biological Process GO group z-score Cellular Component GO group z-scoremitosis 13.64 chromosome,peri-centricregion 14.1Mphaseofmitoticcellcycle 13.54 outerkinetochoreofcondensedchromosome 10.43chromosomesegregation 11.34 DNAtopoisomerasecomplex(ATP-hydrolyzing) 9.09spindlecheckpoint 9.83 condensedchromosomekinetochore 7.98Chromosomecondensation 9.48 spindle 6.57Cytokinesismitosis 8.82 midbody 6.19mitoticchromosomecondensation 8.79 kinetochore 4.86cellcycle 8.29 condensincomplex 4.38mitoticsisterchromatidsegregation 8.19 spindlemidzone 4.38mitoticcheckpoint 7.71establishmentofchromosomelocalization 7.01establishmentofspindlelocalization 6.15mitoticspindleelongation 6.15regulationofexitfrommitosis 5.29apoptoticchromosomecondensation 4.96kinetochoreassembly 4.96G2/Mtransitionofmitoticcellcycle 4.55DNAreplicationcheckpoint 3.74G2/MtransitionDNAdamagecheckpoint 3.37

Table 2 Interestinggeneontologiesbyz-score.

Genes upregulated in grade III tumorsGene Symbol Gene Description Cluster ID Grade II SEM Grade III SEM Fold change

BUB1 Buddinguninhibitedbybenzimidazoles1homolog Hs.469649 0.0474 +/-0.0206 0.3536 +/-0.0183 +7.46

SGOL1 Shugoshin-like1 Hs.105153 0.0389 +/-0.0232 0.2874 +/-0.0291 +7.39CDC2 Celldivisioncycle2,G1toSandG2toM Hs.334562 1.6157 +/-0.2020 11.1596 +/-1.2407 +6.91KNTC2 Kinetochoreassociated2 Hs.414407 0.1563 +/-0.0323 0.937 +/-0.2451 +5.99KIF14 Kinesin family member 14 Hs.3104 0.1688 +/-0.0919 1.0099 +/-0.1489 +5.98CCNB1 CyclinB1 Hs.23960 0.3204 +/-0.0579 1.7126 +/-0.4269 +5.35CCNB2 CyclinB2 Hs.194698 0.965 +/-0.3774 3.7888 +/-0.3885 +3.93PBK PDZbindingprotein Hs.104741 0.1609 +/-0.0246 0.8565 +/-0.1756 +5.32CENPA Centromere protein A Hs.1594 0.1218 +/-0.0262 0.6306 +/-0.0512 +5.18CEP55 Centrosomalprotein55kDa Hs.14559 0.164 +/-0.0269 0.8381 +/-0.0888 +5.11

ASPM Abnormalspindle-likemicrocephalyassociated Hs.104741 0.1035 +/-0084 0.5283 +/-0.0750 +5.10

ZWINT ZW10interactor Hs.591363 1.9926 +/-0.3575 9.8301 +/-0.9727 +4.93TYMS Thymidylate synthetase Hs.592338 1.5112 +/-0.3952 7.0513 +/-0.7508 +4.67CENPF CentromereproteinF,350/400ka Hs.497741 1.0762 +/-0.1126 4.8477 +/-0.8647 +4.50PRC1 Proteinregulatorofcytokinesis1 Hs.567385 1.7588 +/-0.2101 7.7424 +/-1.6503 +4.40IL1B Interleukin1,beta Hs.126256 0.1758 +/-0.0514 0.7302 +/-0.0891 +4.15CDKN3 Cyclin-dependentkinaseinhibitor3 Hs.84113 0.2333 +/-0.0299 0.9527 +/-0.2537 +4.08NUSAP1 Nucleolusandspindleassociatedprotein1 Hs.615092 3.1304 +/-0.7955 12.4848 +/-1.5184 +3.99ANLN Anillin,actinbindingprotein Hs.62180 0.4846 +/-0.0192 1.9148 +/-0.3799 +3.95SPBC25 Spindlepolebodycomponent25homolog Hs.421956 0.1178 +/-0.0400 0.4645 +/-0.1049 +3.94SMC4 Structuralmaintenanceofchromosomes4 Hs.58992 1.7954 +/-0.2603 7.0696 +/-0.8005 +3.94CENPN Centromere protein N Hs.55028 0.1288 +/-0.0463 0.5031 +/-0.1198 +3.91ORC1L Originrecognitioncomplex,subunit1-like Hs.17908 0.063 +/-0.0292 0.2446 +/-0.0306 +3.88ORC1L Originrecognitioncomplex,subunit1-like Hs.17908 0.0916 +/-0.0209 0.3179 +/-0.0393 +3.47MKI67 Ki-67 Hs.80976 0.1976 +/-0.0712 0.7592 +/-0.1441 +3.84CDCKN2C Cyclin-dependentkinaseinhibitor2C Hs.525324 1.0801 +/-0.7249 4.0869 +/-0.3091 +3.78

NUF2 NUF2,NDC80kinetochorecomplexcomponent,homolog Hs.651950 0.159 +/-0.0350 0.6003 +/-0.1017 +3.78

TOP2A Topisomerase(DNA)IIalpha170kDa Hs.156346 0.5733 +/-0.2563 2.1615 +/-0.0933 +3.77

Table 3 Genesinvolvedinmitoticcelldivisionandcellularproliferation.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245

2016Vol. 2 No. 2: 14

6 This article is available from: //www.raredisorders.imedpub.com/

forDNAreplicationandrepair,ispreferentiallyincreased(4.67fold)ingradeIIItumors(Table 3). Increasedexpressionofthymidylatesynthasemaybeanindicationofresistanceto5-fluorouracilandbeassociatedwithpoorprognosis inependymomas, as it is inothertypesofcancer.Additionally,expressionoftopoisomerase II,whichmaintainsthepropertopologyofchromatidDNAandisessentialforchromosomesegregation,isalsoincreasedingradeIIIependymomas(3.77fold)(Table 3 and Figures 1G and 2G). The originrecognitioncomplex(ORC)isaconservedproteincomplexthat is essential for the initiation of DNA replication. ORC1Lis the largest subunitof thiscomplexand levelsof thisproteinvary during the cell cycle in amanner controlled by ubiquitin-mediatedproteolysis.ExpressionofORC1Lisincreased3.88foldingradeIIItumors(Table 3).Enhancerofzestehomolog2(EZH2)is a known repressor of transcription and has been shown topromoteproliferationandinvasionofprostatecancercells[24].ExpressionofEZH2isincreased3.2foldingradeIItumors(Table 3 and Figures 1F and 2F).

G2M DNA damage checkpointThe G2/M DNA damage checkpoint serves to prevent the cellfrom entering mitosis if the genome is damaged. To this end, the activityofthecdc2-cyclinBcomplexplaysakeyroleinregulatingthis transition. As cells approach M phase, the phosphatasecdc25B isactivatedbyphosphorylation,which in turnactivatescdc2. In grade III ependymomas, levels of cdc2 (CDC2) andcyclinB1 (CCNB1) transcripts are increased6.91and5.35 fold,respectivelyrelativetogradeIItumors(Table 3 and Figures 1H and 2H). Similarly, expression of cdc25B (CDC25b) is increased3.04 fold (Table 3 and Figure 4).Levelsofthecheckpointkinases(CHK1andCHK2),thatphosphorylateand inactivatecdc25,arenot changed in grade III tumors. The increased expression of

cdc25binependymomasisconsistentwiththeoverexpressionofcdc25bseeninothermalignancies, includinggliomas,wheremultivariant analysis suggested that increased expression ofcdc25wasanunfavorableprognosticfactor[25].However,overexpressionofcdc25boverridesradiation-inducedG2-Marrestandresultsinincreasedapoptosisinothertypesofcancercells[26].

Mitotic spindle checkpointThespindlecheckpointensuresfaithfulchromosomesegregationbylinkingtheonsetofanaphasetotheestablishmentofbipolarkinetochore-microtubule attachment. This checkpoint is acell cycle surveillance mechanism that ensures the fidelity ofchromosome segregation during mitosis. Our results showedincreasedexpressionofmitoticcheckpointcomponentsincludingBub1 and ZWINT (Figures 5A and 5B). Bub1 (BUB1) directlyinhibits theubiquitin ligaseactivityof theanaphasepromotingcomplexor cyclosome (APC/C)byphosphorylating its activatorCdc20.BUB1depletionleadstotheaccumulationofmisalignedchromatids in which both sister kinetochores are linked tomicrotubulesinanabnormalmanner[27].Increasedexpressionof BUB1 in grade III ependymomas could represent cellularcompensationfordefectsinothermolecularcomponentsofthemitoticspindledamagecheckpointandincreasedexpressionofBUB1might be amarker of ependymomaswith chromosomalinstability.TheexpressionofBUB1ismostdramatically(7.46fold)increasedingradeIIIrelativetogradeIIependymomas.ZWINT(Zwint 10 interactor) is a gene that is required for themitoticcheckpointsignaling[28].ZWINTinteractswithMis12[29]whichis required forkinetochore localizationofZW10.ZW10, in turnanditspartnerRodarerequiredfortherecruitmentofthemotorproteindynein/dynactintothekinetochore.ExpressionofZwint10interactorwasincreased4.93foldingradeIIIependymoma.

NCAPH Non-SMCcondensinIcomplex,subunitH Hs.308045 0.2752 +/-0.0630 1.0333 +/-0.1329 +3.75NCAPG Non-SMCcondensinIcomplex,subunitG Hs.567567 0.4423 +/-0.1776 1.6327 +/-0.1173 +3.69

BIRC5 BaculoviralIAPrepeat-containing5(survivin) Hs.514527 0.4676 +/-0.0641 1.6384 +/-0.2738 +3.50

MKI67 Ki-67 Hs.80976 1.0415 +/-0.5172 4.1638 +/-0.3932 +3.40CKS2 CDC28proteinkinaseregulatorysubunit2 Hs.83758 1.3003 +/-0.0721 4.4016 +/-0.8936 +3.39CENPM Centromere protein M Hs.208912 0.2569 +/-0.0225 0.852 +/-0.1905 +3.32

BIRC5 BaculoviralIAPrepeat-containing5(survivin) Hs.514527 0.1844 +/-0.0750 0.6069 +/-0.0333 +3.29

PTTG1 Pituitarytumor-transforming1 Hs.350966 3.315 +/-0.2738 10.6963 +/-1.1197 +3.23POLE2 Polymerase(DNAdirected),episilon2 Hs.162777 0.2915 +/-0.0282 0.9535 +/-0.0191 +3.21EZH2 Enhancerofzestehomolog2 Hs.444082 1.1795 +/-0.2933 3.7865 +/-0.7287 +3.21CDC25b Celldivisioncycle25homologB Hs.153752 6.263 +/-0.3384 19.0309 +/-2.3736 +3.04CDC45L CDC45celldivisioncycle45-like Hs.474217 0.2088 +/-0.0517 0.6304 +/-0.1348 +3.02GenesdownregulatedingradeIIItumorsGene Symbol GeneDescription ClusterID GradeII SEM GradeIII SEM Foldchange

DYNC1I1 Dynein,cytoplasmic1,intermediatechain1 Hs.440364 3.9184 +/-0.5431 0.1369 +/-0.0805 -28.62

VEGFC VascularendothelialgrowthfactorC Hs.435215 0.6955 +/-0.1632 0.06 +/-0.0370 -11.59TGFA Transforminggrowthfactor,alpha Hs.170009 0.6327 +/-0.1606 0.1027 +/-0.0589 -6.16BTC Betacellulin Hs.591704 0.245 +/-0.0614 0.0518 +/-0.0198 -4.73

LOH11CR2A Lossofheterozygosity,11,chromosomalregion 1, gene A Hs.152944 10.0959 +/-1.6467 2.1902 +/-1.0865 -4.61

PPP3CA Protein phosphatase 3 Hs.435512 1.3463 +/-0.1937 0.4198 +/-0.1367 -3.21

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245v

2016Vol. 2 No. 2: 14

7© Under License of Creative Commons Attribution 3.0 License

Interestingly, levels of dynein were dramatically decreased ingrade III ependymomas, relative to the less aggressivegrade IItumors(Figure 5C).

Downstream effectors of the mitotic checkpointThe anaphase-promoting complex or cyclosome (APC/C)complex is thedownstreameffectorof themitotic checkpoint.Unattached kinetochores activate the BubR1/Mad3-dependentcheckpoint pathway that inhibits Cdc20-mediated activation ofthe APC/C. When all the chromosomes align, the checkpoint

00.5

11.5

22.5

33.5

44.5

5

MKI67

Inte

nsity GradeII

GradeIII

Figure 3 IncreasedexpressionofKi-67inWHOgradeIIIependymoma.

A

B

0

2

4

6

8

10

12

14

CDC2

Inten

sity

GradeIIGradeIII

0

5

10

15

20

25

CDC25b

Inte

nsity GradeII

GradeIII

Figure 4 IncreasedexpressionofCDC2andCDC25BinWHOgradeIIIependymoma.

A

B

C

0

0.5

1

1.5

2

2.5

3

3.5

BUB1B

Inte

nsity GradeII

GradeIII

0

2

4

6

8

10

12

ZWINT

Inte

nsity GradeII

GradeIII

00.5

11.5

22.5

33.5

44.5

5

Dynein

Inte

nsi

ty

GradeIIGradeIII

D

0

2

4

6

8

10

12

14

PTTG1

Inte

nsi

ty

GradeIIGradeIII

Figure 5 Differentialexpressionsofmitoticcheckpointgenesinependymoma.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245

2016Vol. 2 No. 2: 14

8 This article is available from: //www.raredisorders.imedpub.com/

signal is extinguished allowing Cdc20 to activate the APC/Ccomplex. Securin, which was also identified as the pituitarytumor-transforminggene1(PTTG1)isachromosomesegregationregulator and an anaphase inhibitor that binds and inhibitsseparase. APC/Cmediated ubiquitination of securin and cyclinB1 and subsequent degradation by the proteosome triggersanaphase entry. Separase then cleaves cohesin, allowing thesister chromatids to separate. Here we find that both securin(PTTG1)(Figures 1D, 2D and 5D)andcyclinB1(CCNB1)expressionare increased 3.23 and 5.35 fold, respectively, in grade IIIependymomas.Notably,over-orunder-expressionofsecurincangiverisetomitoticcatastrophe,possiblybedrivingchromosomalrearrangement[30].

Kinetochore and centromere associated proteins (Figure 6)Kinetochore associated2 (KNTC2,which is also knownasHec)is part of the Hec1/Ndc80 complex. Expression of KNTC2 isincreasedabout6 fold ingrade III ependymomas.Additionally,a component of the Hec/Ndc80 complex, NUF2, is increased3.78 fold in the grade III ependymomas.Centromereprotein F(CENP-Formitosin) isacentromere-outerkinetochoreprotein.Itlocalizestothespindlemidzoneandtheintracellularbridgeinlate anaphase and telophase, respectively and is subsequentlydegraded. Expression of CENP-F is increased 4.5 fold in gradeIII ependymomas (Figures 1E and 2E). Centromere protein A (CENP-A)encodesacentromereproteinthatcontainsahistonefold domain that is required for targeting to the centromere.CENP-A is proposed to be a component of a modifiednucleosome-likestructureinwhichitreplacesoneorbothcopiesof conventional histone H3. Expression of CENP-A is increased5.18 fold in grade III ependymomas. Shugoshin-like 1 (SGOL1)expressionwasincreased7.39foldingradeIIIrelativetogradeII ependymomas. Survivin (BIRC5, or baculoviral IAP repeat-containing5)(Figures 1C and 2C).Attheonsetofmitosis,survivinregulatesmicrotubululedynamicsat innerkinetochoresviathechromosomalpassengercomplexandmitoticspindleformationviaanindependentpoolnotassociatedwithpassengerproteins.In thechromosomalpassengercomplex,survivin interactswiththeinnercentromereprotein(INCENP),Borealin/DasraBandthemitotickinase,AuroraB.Theseinteractionsarerequiredtotargetthecomplextokinetochores,correctmisalignedchromosomes,properly form the central spindle and complete cytokinesis.Duringmitosis,survivinisachromosomalpassengerproteinthatlocalizes tokinetochoresatmetaphase, transfers to thecentralspindlemidzoneatanaphaseandaccumulates inmidbodiesattelophase.Disruptionofsurvivin-microtubuleinteractionsresultsinlossofsurvivin’santi-apoptosisfunction,amechanisminvolvedin cell death duringmitosis. The overexpression of survivin inependymomas may overcome this apoptosis checkpoint andfavoraberrantprogressionoftransformedcellsthroughmitosis.

Other genes whose products are associated with the regulation of mitosis (Figure 7) CEP55 is centrosomal in interphase cells and is lost from thecentrosomeduringentry intomitosis.Atanaphase,theproteinaccumulatesdiffuselyatthespindlemidzoneandlaterappears

A

0

0.2

0.4

0.6

0.8

1

1.2

1.4

NDC80

Inte

nsity GradeII

GradeIII

B

C

D

0

1

2

3

4

5

6

CENPF

Inte

nsity GradeIIGradeIII

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

SGOL1

Inte

nsity GradeII

GradeIII

0

0.5

1

1.5

2

2.5

BIRC5

Inte

nsity GradeII

GradeIII

Figure 6 Increased expression of kinetochore and centromereassociatedproteinsingradeIIIependymoma.

atthemidbodyandplaysaroleincytokinesis[31].The5.11foldincreasedexpressionofCEP55 in ependymomas, suggests thatthehighlevelofexpressionoftheproteininthetumormayimpairitsabilitytorelocatetothemidbodyandtofunctioninmitoticexit and cytokinesis. Protein regulator of cytokinesis (PRC1)

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245v

2016Vol. 2 No. 2: 14

9© Under License of Creative Commons Attribution 3.0 License

ExpressionofPRC1ingradeIIIependymomasisincreased4.4fold,relativetograde II tumors.Abnormalspindle-likemicrocephalyassociated, ASP. Expression of ASP in grade III ependymomasis increased5.10fold,relativetogradeIIependymomas. It isacriticalregulatorofbrainsize,likelythroughitroleinpromoting

neuroblastsproliferationandsymmetricdivision.ASPhasrecentlybeen shown to be increased in expression in a glioblastomas[32].Condensinsaremulti-subunitproteincomplexes thatplayacentralroleinmitoticchromosomeassemblyandsegregation[33]. Vertebrate cells have two different condensin complexes,condensins I and II, each containingaunique setof regulatorysubunits.CondensinIgainsaccesstochromosomesonlyafterthenuclearenvelopebreaksdownandcollaborateswithcondensinII to assemble metaphase chromosomes with fully resolvedsister chromatids. Condensin II participates in an early stateof chromosome condensation within the prophase nucleus.Structural maintenance of chromosomes (SMC) proteins arefound in condensing complexes and are chromosomal ATPasesthat play fundamental roles in many aspects of higher-orderchromosome organization and dynamics. In eukaryotes, SMC1andSMC3actasthecoreofthecohesincomplexthatmediatesister chromatid cohesion, whereas SMC2 and SMC4 functionas the core of the condensin complexes that are essential forchromosomeassemblyandsegregation.ExpressionofSMC4 ingradeIIIependymomasisincreased3.94fold,ingradeIIIrelativeto relative to grade II ependymoma. The SMC complexes formunique ring- orV-shaped structureswith long coiled-coil arms,andfunctionasATP-modulated,dynamicmolecularlinkersofthegenome [33,34].Aswith SMC4,expressionof the condensin Iregulatory components subunitsG andH (NCAPGandNCAPH)is increased. Levels ofNCAPGandNCAPHare increased about3.7 fold in grade III ependymomas. Since condensins regulateproper assembly of centromeric heterochromatin and therebycontributetotheback-to-backorientationofsisterkinetochores,altered expression of the condensin regulatory componentsmay lead to abnormal interactions between kinetochores andmicrotubules.Similarly,sincecondensinsplaycriticalrolesintheresolutionofsisterchromatidsinmetaphase,alteredexpressionof these regulatory components in progressing ependymomasmayleadtochromosomalalterations,suchastheformationofchromosomebridgesinthesubsequentanaphase.

Recentworkemphasizestheimportanceofidentifyinggenomicdata sets for gene co-expressionmodules that play a key rolein tumor genesis. Here we utilize z-score analysis to identifya module of about 40 genes that consists of mitosis and cellcycle genes that are preferentially expressed in the grade IIIependymoma. In previous studies to identify gene modulesexpressedinglioblastoma,aweightedgeneexpressionnetworkwasconstructedbasedonpairwisePearsoncorrelationsbetweentheexpressionprofiles[32].Unsupervisedhierarchicalclusteringwasusedtodetectmodulesofhighlyco-expressedgenes.Inthatanalysis,amitosis/cellcyclegeneexpressionmoduleof185geneswas identified in glioblastoma that was also present in breastcancer and significantly overlaps with the “metasignature” forundifferentiatedcancer[32].Ofthesegenes,manyareknowntointeractphysicallyand/orfunctionallytoregulatethemetaphasetoanaphasetransition.Thereissignificantoverlapbetweenthegenesfoundintheependymomaandtheglioblastomacellcycle/mitosismodules.Notably,someofthemostconnectedhubgenesin the glioblastoma network were found in the ependymomamodule(e.g.,ASP,BUB1,CCNBP1,CDC2,EZH2,PTTG1,PRC1,andTOP2A). These results suggest that expression of these genes

A

B

C

00.10.20.30.40.50.60.70.80.9

1

CEP55

Inte

nsity GradeII

GradeIII

0123456789

10

PRC1

Inte

nsity GradeII

GradeIII

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

ASP

Inte

nsity GradeII

GradeIII

D

00.20.40.60.8

11.21.41.61.8

2

NCAPG

Inte

nsity GradeIIGradeIII

Figure 7 Increased expression of genes regulating mitosis ingradeIIIependymoma.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245

2016Vol. 2 No. 2: 14

10 This article is available from: //www.raredisorders.imedpub.com/

Marker FunctionIntensity

Grade II Grade III

MKI67 Cell proliferation: theproteinisexpressedduringallactivephasesofthecellcyclebutabsentfromrestingcells 1.04 4.16

CDC2 Downstream effector of the mitotic checkpoint: roleinG2MDNAdamagecheckpoint. Servestopreventthecellfromenteringmitosisifthegenomeisdamaged. 1.16 11.16

CDC25b G2M DNA damage checkpoint:activatedbyphosphorylationandinturnsactivatesCDC2. 6.26 19.03

BUB1B Mitotic checkpoint component: akinasethatplaysakeyroleintheassemblyofthekinetochoresignalingcomplex. 0.87 3.05

ZWINT(Zwint10interactor) Mitotic spindle checkpoint:Requiredformitoticcheckpointsignaling. 1.99 9.83

Dynein Mitotic checkpoint component:isamotorproteinwhichconvertsthechemicalEnergycontainedinATPintothemechanicalenergyofmovement. 3.92 0.14

PTTG1(Securin) Downstream effector of the mitotic checkpoint: achromosomesegregationregulatorandan anaphase inhibitor that binds and inhibits separase. 3.31 10.69

NDC80(Kinetochoreassociated2)

Mitotic spindle checkpoint:Requiredtorecruitasubsetofcheckpointproteinstothekinetochore. 0.15 0.94

CENP-F(mitosin) Mitotic spindle checkpoint: inlateG2,itisrecruitedbyBub1tothekinetochoreandmaintainsthisassociationthroughearlyanaphase. 1.07 4.85

SGOL1(Shugoshin-like1) Mitotic spindle checkpoint:protectscohesioncomplexesatcentromeres. 0.04 0.28

CENP-A Mitotic spindle checkpoint: componentofamodifiednucleosome-likestructureinwhichitreplacesoneorbothcopiesofconventionalhistoneH3. 0.12 0.63

BIRC5(survivin) Mitotic spindle checkpoint: expressedintheG2/Mphaseofthecellcycleinacycleregulatedfashion. 0.46 1.64

CEP55(centrosomalprotein55)

Regulation of mitosis:involvedincentrosome-dependentcellularfunctionssuchascentrosomeduplicationand/orcellcycleprogression,orintheregulationofcytokinesis. 0.16 0.84

PRC1(Proteinregulatorofcytokinesis)

Regulation of mitosis: expressedathighlevelsduringSandG2/ManddropsaftertheCellsexitmitosisandentertheG1phaseofthecellcycle.Itbecomesassociatedwithmitoticspindlesduringmitosisandlocalizestothecellmid-bodyduringcytokinesis.

1.76 7.74

ASP Regulation of mitosis:thehumanorthologofaDrosophilamitoticspindleproteinwhichisessentialfornormalmitoticspindlefunctioninneuroblasts. 0.1 0.53

NCAP-G(CondensinsubunitG)

Mitotic spindle checkpoint:alteredexpressionmayleadstochromosomalalterationssuchastheformationofchromosomesbridgesinthesubsequentanaphase. 0.44 1.63

Table 4 Summaryofgenesinthediscussionsectionalongwiththeirmostimportantbiologicalfunction.

maybemorepredictiveofclinicaloutcomethanthetraditionalmarkersofproliferation.Additionally,thegenesidentifiedinthemodulemaybeup-regulatedbykeymolecular lesions, suchasBRAC1 in breast cancer, that confer a proliferative advantage,thus raising thepossibility that therapies targeting thismodulemay useful in patients with brain tumors. A summary of theresultsispresentedinTables 3 and 4.

Mitoticcellsfacethechallengingtaskoflinkingkinetochorestoshorteningandlengtheningmicrotubulesandactivelyregulatingthese dynamic attachments to produce accurate chromosomesegregation.Themicrotubule-bindinginterfaceofthekinetochoreisofkeyimportanceinchromosomesegregation.

It is tempting to speculate that aneuploidy plays a key role inependymoma progression. The cells found in WHO grade II

ependymomasappear tohavea relativelynormalcomplementof cell division proteins. In the transition from WHO grade IIto grade III, the increasedexpressionof amoduleof cell cycleand cell division genes could disrupt processes responsible formitoticcontrolandDNAmaintenance.Additional chromosomebreakage,structuralrearrangementsandduplicationerrorsarise.These progressing tumor cells rapidly become aneuploid as aresultofaberrantmitoticdivisions.Thesechangesresult inthetumor cell proliferation and the atypical appearance of tumorcellsfoundinthegradeIIIependymoma.

Insummary,wedemonstratedtheexistenceofapanelofgenesthatarehighlyexpressedingradeIIIependymomascomparedtogradeIIependymomas.Thesegenesmaybeuseful insheddinglight on the pathogenesis of ependymomas and their progression into higher grade. As the clinical behavior of ependymomas is

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245v

2016Vol. 2 No. 2: 14

11© Under License of Creative Commons Attribution 3.0 License

stillunpredictable,clinical identificationofthesegenesmayaidindiagnosisandprognosis.ThisisparticularlytrueinasubsetofependymomasinwhichassigningaWHOgradeischallenging,i.e.,thoseependymomaswith focalnecrosis, vascularproliferation,borderline increase in mitoses and atypical cytologic featuresthat fall shortofbeinganaplastic.Thisgroupofependymomasoften represents a diagnostic and management problems forpathologists and neuro-oncologists, respectively. Identificationof a panel of immunostains that can identify this group ofependymomasisdesperatelyneeded.Duetotheiravailabilityforimmunohistochemistrytesting,MIB1,survivin,securin,mitosin,EZH2, TOPOII and cyclin B1 appear to represent promisingmarkers of clinical behavior. We realize the small number ofcases included in this study; however, our study represents animportant substrate for additional studies in order to identifymarkershelpfulinpredictingthebehaviorofependymomasandtherefore establish novelmanagement guidelines. As indicated

above,itappearsthatthereissomeoverlapbetweencertaingenesin glioblastoma and ependymoma. This particular observationseems very promising in identifying those ependymomaswithclinical behavior that defies their histomorphological features(ependymomasthatbehaveaggressivelywhiledisplayinglimitedatypicalhistologicalfeatures).Thisobservationiscurrentlyunderadditionalinvestigationsbyourgroup.

AcknowledgementThis work was supported in part by the National Institutes ofHealth (grant RO1HL35716) and the Harvard Medical School(fundsfromtheWolbachChair).

WeregrettheuntimelydeathofDr.TuckerCollins,M.D.,Ph.D.,onJune8,2007.Hewasagreatmentorandfriend.Hisenthusiasmandsupportwillbegreatlymissed.

Journal of Rare Disorders: Diagnosis & TherapyISSN 2380-7245

2016Vol. 2 No. 2: 14

12 This article is available from: //www.raredisorders.imedpub.com/

References1 BouffetE,PerilongoG,CaneteA,MassiminoM(1998) Intracranial

ependymomasinchildren:acriticalreviewofprognosticfactorsandapleaforcooperation.MedPediatrOncol30:319-329.

2 Ries L, Hankey B (1994) SEER cancer statistics review, 1973-1991:tablesandgraphs.NationalCancerInstitute,Bethesda.

3 HoDM,HsuCY,WongTT,ChiangH(2001)Aclinicopathologicstudyof81patientswithependymomasandproposalofdiagnosticcriteriaforanaplasticependymoma.JNeurooncol54:77-85.

4 Akyuz C, Emir S, AkalanN, Soylemezoglu F, Kutluk T, et al. (2000)Intracranialependymomasinchildhood-aretrospectivereviewofsixty-twochildren.ActaOncol39:97-100.

5 KorshunovA,GolanovA,SychevaR,TimirgazV(2004)Thehistologicgrade is a main prognostic factor for patients with intracranialependymomastreatedinthemicroneurosurgicalera:ananalysisof258patients.Cancer100:1230-1237.

6 Merchant TE, Jenkins JJ, Burger PC, Sanford RA, Sherwood SH, etal. (2002) Influence of tumor grade on time to progression afterirradiationforlocalizedependymomainchildren.IntJRadiatOncolBiolPhys53:52-57.

7 MorkSJ, LokenAC (1977)Ependymoma:a follow-up studyof101cases.Cancer40:907-915.

8 GerdesJ,BeckerMH,KeyG,CattorettiG(1992)Immunohistologicaldetectionoftumourgrowthfraction(Ki-67antigen)informalin-fixedandroutinelyprocessedtissues.JPathol168:85-86.

9 DuffnerPK,KrischerJP,SanfordRA,HorowitzME,BurgerPC,etal.(1998) Prognostic factors in infants and very young children withintracranialependymomas.PediatrNeurosurg28:215-222.

10 Figarella-Branger D, Civatte M, Bouvier-Labit C, Gouvernet J,Gambarelli D, et al. (2000) Prognostic factors in intracranialependymomasinchildren.JNeurosurg93:605-613.

11 Horn B, Heideman R, Geyer R, Pollack I, Packer R, et al. (1999) Amulti-institutional retrospectivestudyof intracranialependymomainchildren:identificationofriskfactors.JPediatrHematolOncol21:203-211.

12 PaulinoAC,WenBC,BuattiJM,HusseyDH,ZhenWK,etal. (2002)Intracranial ependymomas: an analysis of prognostic factors andpatternsoffailure.AmJClinOncol25:117-122.

13 Pollack IF,GersztenPC,MartinezAJ, LoKH, ShultzB, et al. (1995)Intracranial ependymomas of childhood: long-term outcome andprognosticfactors.Neurosurgery37:655-666.

14 RobertsonPL,ZeltzerPM,BoyettJM,RorkeLB,AllenJC,etal. (1998)Survival and prognostic factors following radiation therapy andchemotherapyforependymomasinchildren:areportoftheChildren'sCancerGroup.JNeurosurg88:695-703.

15 GuyotatJ,SignorelliF,DesmeS,FrappazD,MadarassyG,etal.(2002)Intracranialependymomasinadultpatients:analysesofprognosticfactors.JNeurooncol60:255-268.

16 SchildSE,NisiK,ScheithauerBW,WongWW,LyonsMK,etal.(1998)Theresultsofradiotherapyforependymomas:theMayoClinicexperience.IntJRadiatOncolBiolPhys42:953-958.

17 MarksJE,AdlerSJ(1982)Acomparativestudyofependymomasbysiteoforigin.IntJRadiatOncolBiolPhys8:37-43.

18 KovalicJJ,FlarisN,GrigsbyPW,PirkowskiM,SimpsonJR,etal.(1993)Intracranialependymomalongtermoutcome,patternsoffailure.JNeurooncol15:125-131.

19 PeschelRE,KappDS,CardinaleF,ManuelidisEE(1983)Ependymomasofthespinalcord.IntJRadiatOncolBiolPhys9:1093-1096.

20 TaylorMD,PoppletonH,FullerC,SuX,LiuY,etal.(2005)Radialgliacellsarecandidatestemcellsofependymoma.CancerCell8:323-335.

21 Modena P, Lualdi E, FacchinettiF, Veltman J, Reid JF, et al. (2006)Identificationof tumor-specificmolecularsignatures in intracranialependymoma and association with clinical characteristics. J ClinOncol24:5223-5233.

22 KanehisaM,GotoS,KawashimaS,OkunoY,HattoriM (2004)TheKEGG resource fordeciphering thegenome.NucleicAcidsRes32:D277-280.

23 DonigerSW,SalomonisN,DahlquistKD,VranizanK,LawlorSC,etal.(2003)MAPPFinder:UsingGeneOntologyandGenMAPPtocreateaglobalgene-expressionprofilefrommicroarraydata.GenomeBiol4:R7.

24 ChoKS,OhHY,LeeEJ,HongSJ(2007)Identificationofenhancerofzestehomolog2expression inperipheralcirculatingtumorcells inmetastaticprostatecancerpatients:apreliminarystudy.YonseiMedJ48:1009-1014.

25 NakabayashiH,HaraM,ShimizuK(2006)PrognosticsignificanceofCDC25Bexpressioningliomas.JClinPathol59:725-728.

26 MiyataH,Doki Y, YamamotoH, Kishi K, TakemotoH, et al. (2001)OverexpressionofCDC25Boverridesradiation-inducedG2-Marrestandresultsinincreasedapoptosisinesophagealcancercells.CancerRes61:3188-3193.

27 Meraldi P, Sorger PK (2005) A dual role for Bub1 in the spindlecheckpointandchromosomecongression.EmboJ24:1621-1633.

28 WangH,HuX,DingX,DouZ,YangZ,etal. (2004)HumanZwint-1specifies localization of Zeste White 10 to kinetochores and isessential formitoticcheckpointsignaling. JBiolChem279:54590-54598.

29 ObuseC,IwasakiO,KiyomitsuT,GoshimaG,ToyodaY,etal.(2004)AconservedMis12centromerecomplexislinkedtoheterochromaticHP1andouter kinetochoreprotein Zwint-1.NatCell Biol 6: 1135-1141.

30 Zou H, McGarry TJ, Bernal T, Kirschner MW (1999) Identificationof a vertebrate sister-chromatid separation inhibitor involved intransformationandtumorigenesis.Science285:418-422.

31 FabbroM,ZhouBB,TakahashiM,SarcevicB,LalP,etal.(2005)Cdk1/Erk2-andPlk1-dependentphosphorylationofacentrosomeprotein,Cep55, is required for its recruitment tomidbodyandcytokinesis.DevCell9:477-488.

32 HorvathS,ZhangB,CarlsonM,LuKV,ZhuS,etal.(2006)Analysisofoncogenicsignalingnetworks inglioblastoma identifiesASPMasamoleculartarget.ProcNatlAcadSciUSA103:17402-17407.

33 HiranoT(2005)Condensins:organizingandsegregatingthegenome.CurrBiol15:R265-275.

34 LosadaA,HiranoT(2005)Dynamicmolecularlinkersofthegenome:thefirstdecadeofSMCproteins.GenesDev19:1269-1287.