Thisarticleisprotectedbycopyright.Allrightsreserved1
OriginalArticle
TET3mediatesalterationsintheepigeneticmarker5hmCandAktpathwayin
steroid-associatedosteonecrosis?
JieZhao1,2,MD,Xin-longMa1,2,,MD,Jian-xiongMa2,,PhD,LeiSun2,BS,BinLu2,BS,Ying
Wang2,MS,Guo-shengXing3,MS,YanWang2,BS,Ben-chaoDong2,MS,Li-yanXu1,2,MS,
Ming-JieKuang1,2,MS,LinFu1,2,MS,Hao-haoBai2,MS,YueMa3,4,MS,Wei-linJin3,4,MD&PhD
1.TianjinMedicalUniversityGeneralHospital,154,AnshanStreet,HepingDistrict,Tianjin
300052,China
2.InstituteofOrthopedics,TianjinHospital,122,MuNanStreet,HepingDistrict,Tianjin300050,
China
3.DepartmentofInstrumentScienceandEngineering,KeyLab.forThinFilmand
MicrofabricationTechnologyofMinistryofEducation,SchoolofElectronicInformationand
ElectronicEngineering,ShanghaiJiaoTongUniversity,Shanghai200240,China
4.NationalCentersforTranslationalMedicine,ShanghaiJiaoTongUniversity,Shanghai200240,
China
Correspondenceto:Xin-longMa,TianjinMedicalUniversityGeneralHospital,154,Anshan
Street,HepingDistrict,Tianjin300052,China.Phone:86.23197199;Email:
xinlongma2016@163.com;
Jian-xiongMa,InstituteofOrthopedics,TianjinHospital,122,MuNanStreet,HepingDistrict,
Tianjin300050,China.Phone:86.23197122;Email:mjx969@163.com
Runningtitle：TET3-5hmC-AktmediatesGC-inducedosteocyteapoptosisandSAON
?Thisarticlehasbeenacceptedforpublicationandundergonefullpeerreviewbuthasnotbeen
throughthecopyediting,typesetting,paginationandproofreadingprocess,whichmayleadto
differencesbetweenthisversionandtheVersionofRecord.Pleasecitethisarticleasdoi:
[10.1002/jbmr.2992]
AdditionalSupportingInformationmaybefoundintheonlineversionofthisarticle.
InitialDateSubmittedMay10,2016;DateRevisionSubmittedSeptember1,2016;DateFinalDispositionSetSeptember9,2016
JournalofBoneandMineralResearch
Thisarticleisprotectedbycopyright.Allrightsreserved
DOI10.1002/jbmr.2992
Thisarticleisprotectedbycopyright.Allrightsreserved2
Abstract
Steroid-associatedosteonecrosis(SAON)isoneofthecommoncomplicationsofclinical
glucocorticoid(GC)administration,withosteocyteapoptosisappearingastheprimary
histopathologicallesion.However,theprecisemechanismunderlyingSAONremainsunknown.
EpigeneticmodificationmaybeamajorcauseofSAON.Recently,cumulativeresearchrevealed
thatTen-ElevenTranslocation(TET)proteinscancatalyzetheconversionof5-methylcytosine
(5mC)to5-hydroxymethylcytosine(5hmC)andthenaltertheepigeneticstateofDNA.Here,we
reportthatTET3-5hmCwasup-regulatedinthefemoralheadtissuesofSAONpatientsand
MLO-Y4cellswithdexamethasone(Dex)treatment.KnockdownofTET3inMLO-Y4cells
decreased5hmCenrichmentandrescuedDex-inducedapoptosis.Meanwhile,thelocal
intramedullaryinjectionofTET3siRNAinSprague-DawleyratsabrogatedGC-inducedosteocyte
apoptosis,histopathologicalchanges,abnormalMRIsignalsandbonemicrostructuredeclinesinthe
femoralheadinvivo.Moreover,ahydroxymethylatedDNAimmunoprecipitation(hMeDIP)-chip
analysisofDex-treatedosteocytesrevealed456different5hmC-enrichedgenes.TheAktpathway
wasfoundtomediatethefunctionaleffectofDex-induceddynamic5hmCchange;thiswasfurther
verifiedinclinicalsamples.ThelossofTET3inMLO-Y4cellsabrogatedDex-inducedAkt
signalingpathwayinhibition.Therefore,ourdataforthefirsttimeidentifytheeffectofTET3-5hmC
ontheAktpathwayandthenecessityofthissignalingcascadeinSAON,identifyinganew
potentialtherapeutictarget.Thisarticleisprotectedbycopyright.Allrightsreserved
Keywords:osteonecrosis,epigenetics,DNAdemethylation,TET3,apoptosis
Thisarticleisprotectedbycopyright.Allrightsreserved3
Introduction
SAONisconsideredamultifactorialdiseasethatleadstosubchondralcollapsesandtotaljoint
replacementduringlaterstages(1,2).SAONranksfirstamongtheknownriskfactorsof
nontraumaticfemoralheadnecrosis,whichisassociatedwiththewideuseofglucocorticoids(GCs)
totreatmanydiseasesintheclinic(e.g.,rheumatoidarthritisandsystemiclupuserythematosus)(3).
ThereareseveralalternativemechanismsunderlyingSAON,suchasfatembolization,
intramedullarypressurechanges,modifiedarteryconstriction,circulatoryimpairment,coagulation
disordersandcelldysfunction(4,5).However,osteocyteapoptosishasrecentlybeenshowntobethe
primaryhistopathologicalchangeinosteonecrosis(1,6-8).GCsinduceosteocyteapoptosis,which
disruptsthemechanosensoryfunctionoftheosteocytelacunar–canalicularsystemandultimately
leadstothecollapseofthefemoralhead(1),butthemolecularmechanismofosteocyteapoptosis
remainselusive.
Epigenetics,aneffectivemethodforstudyingtheinterplaybetweenenvironmentalsignalsand
thegenome,hasreceivedagreatdealofattentionrecently(9).Itisbecomingapparentthat
methylationmodificationplaysanimportantroleinSAON.AberrantmethylationoftheABCB1
genehasbeenshowntoberesponsibleforthepathogenesisofSAON(4),andicariinmaybenefitthe
mesenchymalstemcellsofpatientswithSAONthroughABCB1promoterdemethylation(10).
However,inthepastfewyears,ithasbecomeapparentthatDNAmethylationisnotastatic
epigeneticmarkbutishighlydynamicandisgovernedbyaprecisemolecularnetworkof
regulators(11).5mCcanbefurtheroxidizedto5hmC,5-formylcytosines(5fC)and
5-carboxylcytosines(5caC)bytheTETproteinfamilyincludingTET1,TET2andTET3,resulting
inactivedemethylation(12,13).Thedysregulationof5hmClevelsmayleadtoneurologicaldiseases,
Thisarticleisprotectedbycopyright.Allrightsreserved4
cancer,arthritisandsomeotherdiseases(14).
InviewoftherecentadvancesintheunderstandingofDNAdemethylationpathwaysand
increasingevidenceofthefunctionofstable5hmCanditsderivativesintheepigenome,wesought
toinvestigatetheeffectofTETand5hmCinSAON.Inthisstudy,wethereforefirstmeasuredthe
expressionofTET1,TET2,TET3and5hmCinbonetissuesfromSAONandfemoralneckfracture
patients.Then,weexaminedthechangesin5hmCandtheTETfamilyinDex-treatedMLO-Y4
cells.AhMeDIP-chipanalysiswasperformedtoexplorethefunctionaleffectsofdynamic5hmC
change,whichwerefurtherstudiedinclinicalsamples.KnockdownofTETinMLO-Y4cellsorin
vivosiRNAtechnologywereusedtoverifytheeffectofTET-5hmConDex-inducedosteocyte
apoptosisandSAON.WeaimedtodeterminetheeffectsofGCsontheepigenomeofosteocyteand
definetheroleofTET-5hmCinaSAONmodel.
Materialsandmethods
1.SAONandfemoralneckfracturebonetissuepreparation
HumanSAONandfemoralneckfracturesampleswereobtainedfromsurgicalproceduresinthe
DepartmentofJointSurgeryandDepartmentofTraumaticOrthopedicsatTianjinHospital,
respectively.Allproceduresregardingobtainingpatientsampleswereapprovedbytheethics
committeeoftheTianjinHospitalInstitutionalReviewBoardandcompliedwiththeWorldMedical
AssociationDeclarationofHelsinki.SixSAONand6femoralneckfracturepatientswererecruited.
TheseverityofSAONandfemoralneckfracturewasevaluatedusingtheFicatandGardenstaging
system,respectively(15,16).SAONoperativeindicationsincludedacollapsednecroticfemoralhead
withintractablepain.Patientswithademonstrablehistoryofdirecttraumaorwiththepossibilityof
acombinationofcauseswereexcluded(17).Duringsurgery,thefemoralheadwasextractedandthen
thenecroticbonearoundthecenteroffemoralheadwasharvestedforstudy.Bonespecimensfrom
Thisarticleisprotectedbycopyright.Allrightsreserved5
6patientswhounderwenthiparthroplastyforfemoralneckfractureswereharvestedfromthesame
placeandusedascontrols.Noage-matchedcontrolscouldbeobtainedbecauseoftheabsenceofa
surgicalindicationforhiparthroplastyinyoungpatientswithfemoralneckfractures.
2.Cellcultures,Dextreatmentandtransfection
MLO-Y4osteocyte-likecells(agiftfromDr.LyndaBonewald,UniversityofMissouri-KansasCity)
wereculturedaspreviouslydescribed(18).Fordose-dependentexperiments,cellsweretreatedwith
varyingconcentrations(10-8Mto10-5M)ofDex(Sigma-Aldrich,USA).Fortime-dependent
experiments,cellsweretreatedwith10-6MDexforvariousdurationsoftime(19,20).Toactivatethe
Aktsignalingpathway,cellswereexposedtothePTENinhibitorbpV(phen)(10-6M,
Sigma-Aldrich)for30min.Afterwards,thecellswereincubatedwith10-6MDexforupto12h.
ForknockdownofTET,TET1specificsiRNA(TET1-1GCAGATGGCCGTGACACAAAT,
TET1-2GCAGCTAGCTATAGAGTATAG),TET2specificsiRNA(TET2-1
CTCAGGGATGTCCTATTGCTAAA,TET2-2GGATGTAAGTTTGCCAGAAGC),TET3specific
siRNA(TET3-1GCTCCAACGAGAAGCTATTTG,TET3-2AAGCGCAACCTATTCTTGGAA)
andcontrolscramblesiRNA(ACGUGACACGUUCGGAGAATT)weresynthesizedbyBiotend
(BiotendBiotechnologyCo.,Ltd,China)basedonpublishedstudies(21).Transfectionswere
performedusingLipofectamine2000TransfectionReagent(Invitrogen,USA)accordingtothe
manufacturer’sprotocol.Twenty-fourhoursthereafter,cellsweretreatedwith10-6MDexforan
additional12h.
3.hMeDIP-chipanddataprocessing
MLO-Y4cellsweretreatedwith10-6MDexfor0h,4hand12h.Thereafter,genomicDNA
(gDNA)wasextractedandsonicatedtorandomfragmentsof200-1000bp.Onemicrogramof
sonicatedgDNAwasimmunoprecipitatedwith1μlof5hmCantibodyovernightat4°Cwith
Thisarticleisprotectedbycopyright.Allrightsreserved6
rockingagitation.Then,antibody-DNAcomplexeswerecapturedwithproteinA/Gbeads
(ThermoFisherScienti?c,USA).Then,5hmC-containingDNAfragmentswerepurifiedusing
QiagenMinElutecolumns(Qiagen,Germany).ForDNAlabeling,theNimbleGenDual-ColorDNA
LabelingKitwasusedaccordingtothemanufacturer’sguidelines(Nimblegen,USA),andthe
labeledDNAwaspurifiedusingisopropanol/ethanolprecipitation.AmouseRefSeqPromoterArray
(Arraystarlnc,Rockville,USA)washybridizedwiththelabeledDNAat42°Cfor16to20hina
hybridizationchamber(Nimblegen)(22).
ThehMeDIPchipdatawereanalyzedwithNimbleScanv2.5(NimbleGen).Anormalized
peakMvalue(peakMvalue=log2hMeDIP/Input)wascreatedforeachsampletoquantifytherelative
valueof5hmC,whichisproportionalto5hmCexpression.ThethresholdsetofpeakMvalue
changes≥0representedthedifferent5hmC-enrichedgenes(DEG).KEGGpathwayandGOanalysis
ofDEGwereperformedusingthedatabaseforannotation,visualizationandintegrateddiscovery
(DAVID)website(23).The-log10(pValue)wasusedasameasureofthesignificanceofthepathway
associatedwiththe5hmCchange.AheatmapandVenndiagramshowingthedynamicchangein
5hmCwereconstructedwithMultiExperimentViewerv4.9andVenny2.1website,respectively.
4.EstablishmentoftheSAONmodelandknockdownofTET3invivo
TheanimalprocedureswereconductedaccordingtotheGuidefortheCareandUseofLaboratory
Animals:EighthEdition,andthestudyprotocolwasapprovedbytheEthicalCommitteeofthe
TianjinMedicalUniversityGeneralHospitalandTianjinHospital.
Twenty-four3-month-oldadultmaleweight-matchedSpragueDawley(SD)rats(203±14g)were
randomlydividedinto4experimentalgroupsof6ratseachandkeptunderthesamestandard
conditions;waterandfoodwereavailableadlibitum.
TheSAONmodelwasestablishedbasedonaprotocolreportedpreviously(24);
Thisarticleisprotectedbycopyright.Allrightsreserved7
methylprednisolonesodium(MPS,Pharmacia&Upjohn,Peapack,USA)wassubcutaneously
injectedatadoseof21mg/kgperdayfor4weeks.ThesiRNAtransfectioncomplexwasprepared
accordingtothemanufacturer''sinstructions.Briefly,5nmolsiRNA(Biotend),40μlEntranster-in
vivotransfectionreagent(Engreen,China),and40?μlof10%glucoseweremixedfor15minsat
roomtemperature.TET3orscramblesiRNA-Entranster-invivocomplexwasinjectedinto
medullarycavityofbilateraldistalfemurtwiceoverthecourseof4weeks(2,25,26).Euthanasiawas
performedatweek4.TheleftfemoralheadswereusedforMicro-CTscan(n=6/group),histological
evaluation(n=6/group)andTUNELanalysis(n=6/group)(27,28).Therightfemoralheadswere
separatedintotwopartsforwesternblot(n=4/group)anddotblotanalysis(n=4/group).
5.Micro-CT-basedtrabeculararchitectureassessment
ThefemoralheadswerescannedwithanInveonmicroPET/CTmanufacturedbySiemens(Berlin,
Germany)atavoltageof80kVandacurrentof500μA,withanentirescanlengthof20mmfrom
thetopofthefemoralheadtothefemoralshaftinaspatialresolutionof10μm.3-Dstructureswere
reconstructedusingtheInveonanalysisworkstation.TheROIwasdeterminedasanirregular
anatomiccontouradjacenttotheendocorticalsurfaceandepiphyseallineinproximalepiphysis.
Thecorticalboneandspongybonewereseparatedmanuallybyautotrace;later,thetrabeculaeand
thebonemarrowwereseparatedusingthethresholdfunction(29).Thebonevolume/totalvolume
(BV/TV),bonesurfacearea/bonevolume(BS/BV),trabecularthickness,trabecularnumber,
trabecularseparationwerecalculated(30).
6.Statisticalanalysis
Osteonecrosiswasde?nedaspreviouslydescribedandtheincidenceofSAONinratswasdefined
asthenumberoffemoralheadswithpredominantemptylacunaedividedby6femoralheadsin
eachgroup(27,28).Numberdatawereexpressedasthemean±SDwithANOVA(Fisher''sleast
Thisarticleisprotectedbycopyright.Allrightsreserved8
significantdifference,LSD)methodtoassesssignificantdifferencesbetweengroups.Analysisof
covariance(ANCOVA)wasusedtocorrecttheagebiasbetweenSAONandfracturepatients.All
experimentswererepeatedatleastthreetimes.StatisticalanalysiswasperformedusingSPSS20.0
software(Chicago,USA).p<0.05wasconsideredsignificant.
Results
TET3-5hmClevelwasup-regulatedinSAONfemoralheads.
ThepatientcharacteristicsarelistedinSupplementaltables1and2.IntheSAONgroup,6SAON
patientsrangingfrom59to77yearsold(66.2±6.6years)wereenrolled.Inthecontrolgroup,6
patientsrangingfrom68to85yearsold(76.2±5.5years)wereincluded.Therepresentative
radiographsshowedprominentfemoralheaderosionandcollapseintheSAONgroupandevident
femoralneckfractureinthecontrolgroup(Figure1A).TUNELanalysiswasperformedtomeasure
cellapoptosisinthebonetissues.IntheSAONgroup,osteocytesembeddedinthetrabecula
displayedstrongTUNELstaining,whichisconsistentwithpreviousstudies(Figure1B,1C)(31).It
hasbeenreportedthatTET3mediatesGC-inducedharmfulneurodevelopmentaleffects(32).To
exploretheroleofTETinthepathogenesisofSAON,TET1,TET2andTET3levelswereanalyzed.
Real-timePCRshowedthatthelevelsofTET1andTET2wereunchanged,whiletheexpressionof
TET3wasincreasedinSAONtissues(Figure1D-1F),whichwasconfirmedbywesternblot
(Figure1G,1H).TET3mediates5mCconversionto5hmC,5fC,and5caC.Asthesteady-statelevel
of5hmCisconsideredmuchhigherthanthepoolof5fCand5caC(33),weevaluatedonlythe5hmC
level.DNAdotblotindicatedstrongenrichmentof5hmCinSAON(Figure1I,1J).
TheTET3-5hmCincreasewasinvolvedinDex-inducedosteocyteapoptosis.
TodeterminetheeffectofGCtreatmentonosteocyteviability,wefirstperformedaninvitro
analysisinMLO-Y4cells.ATUNELassayindicatedthatMLO-Y4cellstreatedwithDex
Thisarticleisprotectedbycopyright.Allrightsreserved9
demonstratedadose-dependentincreaseinapoptosis(Figure2A,2B).Furthermore,MLO-Y4cells
treatedwithatimecourseof10-6MDex(arangefrom2h,4h,8hto12h)showedincreased
apoptosis(Figure2C,2D).ToexplorethemechanismofDex-inducedapoptosis,Bcl-2,Baxand
cleavedcaspase-3proteinlevelswerequantified.Thedatashowedthatpro-apoptoticBaxand
cleavedcaspase-3wereup-regulated,whiletheanti-apoptoticBcl-2proteinlevelwaslowerat8h
and12h(Figure2F).IntheCCK-8andtrypanblueexclusionassay,thenumberofMLO-Y4
normalcellswasfoundtobereducedbyDextreatmentinbothadose-dependentanda
time-dependentmanner(SupplementalFigure1A,1B,2A,2B).Inaddition,anethynyl
deoxyuridine(EdU)incorporationassayindicatedthatDNAreplicationandcellproliferationwere
suppressedbyDex(SupplementalFigure1C,1D,2C,2D).
TostudywhetherTETand5hmCwereinvolvedinDex-inducedapoptosisinvitro,genome-wide
5hmClevelsandtheexpressionofTET1,TET2andTET3werequantified.IFanddotblotshowed
thatglobal5hmCwasup-regulatedbyDexinatime-dependentmanner(Figure2E,2G).Real-time
PCRrevealedthatTET3increasedafterDextreatmentwhileTET2remainedunchanged.Although
TET1increasedslightly,itwasnotassignificantasTET3(Figure2H-2J),whichwassimilartothe
resultsoftheproteinexpressionanalysis(Figure2K).Theup-regulationwasobservedinbothtotal
andnuclearTET3,asrevealedbysubcellularextractionanalysis(Figure2L).Thesedataindicated
thatTET3maylocalizeinnucleusafterDextreatmenttoinduce5hmCenrichmentandinitiate
osteocyteapoptosis.
TheAktpathwaymediatedthefunctionaleffectofDex-induceddynamicchangesin5hmC.
Tounderstandthefunctionaleffectofthe5hmCchange,weperformedahMeDIP-chipassayto
examinethe5hmCepigenomeinMLO-Y4cellsoverthecourseofDextreatment.Followingthe
hMeDIP-chipassayandpurificationof5hmC-enrichedDNAfragments,theArrayStarMouse
Thisarticleisprotectedbycopyright.Allrightsreserved10
RefSeqPromoterArraywasusedtocharacterizethedistributionof5hmCwithinpromoters.Aheat
mapandVenndiagramswereusedtodepictthedynamicchangeinDex-induced5hmCexpression.
Fourhundredfifty-sixgenespredictedtohavedifferentenrichmentlevelsof5hmCwereselected.
Ofthesegenes,92geneswere5hmC-upregulatedand98geneswere5hmC-downregulatedinall
Dex-treatedMLO-Y4cells(4hand12h).Moreover,1225hmC-upregulatedgenesand144
5hmC-downregulatedgenesoccurringinDextreatmentfor12hwerealsoidentified(Figure3A).
Then,weusedVenndiagramstofurthershowthedynamic5hmCchangeindifferenttypesofgenes.
ThegeneswereclassifiedashighCpGgenes(HCG),intermediateCpGgenes(ICG)andlowCpG
genes(LCG)basedontheCpGcontentinthepromoter.Thenumberofgeneswithdifferential
5hmCenrichmentislistedinFigure3B.GOanalysisrevealedsignificantaccumulationofDEG
involvedinseveralimportantcellularprocesses,suchasdevelopment,localization,metabolismand
differentiation(Figure3C).
KEGGpathwayanalysisshowedthatsignalingpathways,includingthePI3K-Aktpathway,
Notchpathway,apoptosispathwayandWntpathway,exhibiteddynamicchangesin5hmClevels
(Figure4A,4B).ThePI3K-AktandNotcharesignalingpathwaysshowedthemostsignificant
5hmCup-regulationanddown-regulation,respectively.ElevengenesassociatedwiththePI3K-Akt
signalingpathwayandthecorrespondingpeakMvaluesarelistedintable1,whichshowsthe5hmC
expressioninthesegenepromoters.Then,weusedreal-timePCRtoquantifythemRNAexpression
ofthesegenes.TheresultsshowedthatPTEN,Ppp2r5dandJAK3wereup-regulatedwithdifferent
patternsinDex-treatedcellsrelativetocontrolcells,whilePIK3r5,PIK3cd,TCL1andPDK2were
down-regulated(Figure4C-4M).ThemRNAlevelsofPIK3c2b,Itga1,Itga11andTlr4remained
almostunchanged,thoughthehMeDIP-chipassayrevealeddifferentenrichmentwith5hmC(Figure
4C-4M).Accordingtothewesternblotassay,PTENwasup-regulatedandp-Aktwas
Thisarticleisprotectedbycopyright.Allrightsreserved11
down-regulatedinallDex-treatedcells.Meanwhile,thetotalAktleveldidnotchangesignificantly
(Figure4N).TofurtherexplorewhethertheAktpathwaymediatesthefunctionaleffectof
Dex-induced5hmCchange,bpV(phen)wasusedtoactivatetheAktpathway.Wefoundthatthe
osteocyteapoptosisinducedbyDexwasreversedsignificantlybyactivatingAktsignaling,and
westernblotrevealedthesamepattern(Figure5A-5C).Significantly,thelevelsofAktand
apoptosiswerealsochangedinSAONtissues.TheexpressionofPTENwasincreasedinSAON,
whilep-Aktandp-PI3Kdecreased.Meanwhile,Bcl-2decreasedattheproteinlevelinSAON,but
Baxandactivecaspase-3increased(Figure5D).ThesedataverifiedthatthePTEN-Aktpathway
mediatesthefunctionaleffectofDex-inducedTET-5hmCchangeinosteocyteandiscriticalfor
osteocyteapoptosisandSAON.
Additionally,wealsofoundchangesinNotch4(notNotch2)expressioninSAONtissues
(SupplementalFigure3).Thus,Notchsignalingmayalsopartiallymediatethefunctionaleffectof
5hmCchangeandcouldbefurtherexploredinthefuture.
KnockdownofTET3abrogatedDex-inducedosteocyteapoptosisandAktpathwayinhibition.
TofurtherconfirmtheeffectofTET-5hmConDex-inducedosteocyteapoptosis,weperformed
siRNA-mediatedknockdownofTET1-3expressioninMLO-Y4cells.Wefirsttestedtheefficacyof
severalreportedTET1-3sequencesinknockingdownTETexpression.After24hoftransfection,
real-timePCRdatashowedthatfivesequences(siTET1-1,siTET2-1,siTET2-2,siTET3-1and
siTET3-2)decreasedTETmRNAexpression(Figure6A).ThecellstransfectedwithsiTET1-1,
siTET2-2andsiTET3-1wereusedinthesubsequentexperiments.Significantly,onlyknockdownof
TET3reversedDex-inducedapoptosis,whileknockdownofTET1orTET2hadnoeffect(Figure
6B).TwelvehoursAfterDextreatment,30.00±1.91%oftransfectedcontrolcells,30.05±2.47%of
siTET1-transfectedcells,and30.75±3.46%ofsiTET2-transfectedcellswereTUNELpositive.In
Thisarticleisprotectedbycopyright.Allrightsreserved12
contrast,thenumberofapoptoticcellsdecreasedto17.14±1.06%insiTET3-transfectedcells
(Figure6C).WesternblotindicatedthatonlyknockdownofTET3inhibitedcaspase-3activation,
althoughknockdownofTET1andTET2hadaslighteffectonBcl-2(Figure6D).Inaddition,
transfectionofsiTET3reversedDex-inducedTET3up-regulationattheproteinlevel.Subcellular
extractionanalysisindicatedthatnuclearTET3wasalsodown-regulated(Figure6E,Supplemental
Figure4A).Meanwhile,dotblotandIFshowedthattheglobalenrichmentof5hmCwasabrogated
afterknockdownofTET3(Figure6F,6G).TheseresultsfurtherverifiedthatTET3isspecificand
indispensableforDex-induced5hmCenrichmentandosteocyteapoptosis.
WefurtherexploredtheactivityoftheAktsignalingpathway.ComparedwiththeDex-treated
group,knockdownofTET3reversedthephosphorylationofAktanddown-regulatedPTENlevels;
thus,thesuppressionoftheAktsignalingpathwaydecreased(Figure6H).
ThedatashowedthatthenumberofnormalMLO-Y4cellsinthesiTET3groupwaslargerthan
thatintheDex-treatedgroup(SupplementalFigure4B,4C).FurtherEdUincorporationassays
showedthatthenumberofEdUpositivecellsincreasedsignificantlyafterTET3knockdown
comparedwithDextreatment(SupplementalFigure4D,4E).
KnockdownofTET3alleviatedSAONinvivo
TET3siRNAtransfectioninvivowasperformedasillustratedinFigure7A.Westernblotshowed
thatsiTET3treatment,whichdidnotalterthelevelsofTET1andTET2,abrogatedMPS-induced
TET3expressioninfemoralheadbonetissues(Figure7B)andfurtherreversedMPSinductionof
global5hmCaccumulation(Figure7C).
WenextexploredthetherapeuticeffectofTET3knockdownonSAONbyperformingHE
staining,TUNELassay,Micro-CTandMRIscansinrats.HEstainingshowedobviousemptybone
lacunaeandadipocytesinthebonemarrowoftheSAONgroup;theseeffectswererescuedby
Thisarticleisprotectedbycopyright.Allrightsreserved13
siTET3treatment.Theincidenceofosteonecrosisatweek4was83.3%(5/6)intheSAONgroup
and0(0/6)inthesiTET3groupbythestandardofHEstainedhistologicalexamination(Figure7D).
TUNELanalysisshowedthattheosteocytesembeddedinthetrabeculaeattheproximalepiphysis
ofthefemoralheaddisplayedintensestainingintheSAONgroup,whilefewosteocytesdisplayed
positivestainingwithsiTET3treatment(Figure7E,7F).
Micro-CTshowedthatMPStreatmentdecreasedtheBV/TV,trabecularnumberandtrabecular
thicknessandincreasedtrabecularseparation,indicatingapparenttrabecularbonelossinSAON
femoralheads;thischangewassignificantlyrescuedbysiTET3treatment(Figure7G,7H).
Moreover,MRIscansshowedheterogeneousT1W1signalsandhigh-intensityT2W1signalsinthe
femoralheadintheSAONandscramblegroups.Incontrast,thesiTET3groupshoweda
homogeneousT1W1signalandalow-intensityT2W1signal,similartotheresultsobservedina
normalfemoralhead(Figure7I).Collectively,thesedatasuggestthatknockdownofTET3
alleviatedosteocyteapoptosisandSAONinvivo.
Discussion
OurresultsuncoveredadirectimpactoftheTET3-5hmC-Aktpathwayonthepathogenesisof
SAON(Figure8).Thisdeductionisbasedonthefollowing.First,weobservedthataglobal
enrichmentof5hmCandTET3increaseinSAONbonespecimens.Second,wefoundthat
Dex-inducedTET3-5hmCup-regulationinMLO-Y4osteocytes.Third,ahMeDIP-chipanalysis
exhibiteddynamic5hmCchangeinthepromoterofgenesassociatedwiththeAktpathway,and
pharmacologicalactivationoftheAktpathwayrescuedDex-inducedosteocyteapoptosis.Fourth,
AktandapoptosissignalswerealteredinSAONsamples.Fifth,knockdownofTET3withspecific
siRNAreversedtheGC-inducedinhibitionofAktsignalingandosteocyteapoptosisinvitroand
SAONinvivo.AnearlierstudyreportedthatGCselicitstrongandpersistenteffectsonDNA
Thisarticleisprotectedbycopyright.Allrightsreserved14
hydroxymethylationinneuralstemcells,withTET3beingakeyfactorforharmful
neurodevelopmentaleffects(32),therebyimplicatingalterationsintheTET3-5hmCpathwayin
SAON.
ThepatternofDNAmethylationatcytosinebasesinthegenomeistightlylinkedtogene
expression,andDNAmethylationabnormalitiesareoftenobservedindiseases.TheTETproteins
promotethereversalofDNAmethylation(33).TET1isindispensableformaintainingembryonic
stemcellpluripotencyandexertstumorsuppressingfunctionsingastriccancer(34,35).TET2was
positivelycorrelatedwiththeregulationofneuronsurvival(36).TET3isrequiredforthenormal
survival,proliferationanddifferentiationofneuralprogenitorcells(32).Nevertheless,theexpression
patternandexactfunctionofthesethreeimportantproteinshavenotbeenreportedinbonetissues
previously.Inourstudy,wefirstfoundtheexpressionofTET1,TET2andTET3inbonetissues.
TET3functionappearstobedistinctfromthatofTET1andTET2inSAONpathogenesis,asour
datashowedthatonlyTET3expressionwassignificantlyup-regulatedandaccompaniedby5hmC
accumulationinSAON.TET1wasalsoslightlyincreasedbyDextreatmentinvitro,butnotas
obviouslyasTET3.Inaddition,onlyknockdownofTET3couldreverseDex-inducedMLO-Y4cell
apoptosis;thus,TET3isspecificallyinvolvedinDex-inducedosteocyteapoptosisandSAON.
ApartfromaconservedcorecatalyticregionintheCterminusofeachTETprotein,TET1and
TET3haveanN-terminalCXXCzincfingerdomainthatcanbindDNAandfacilitaterecruitment
totargetgenesinthegenome(37,38).Thus,GC-inducedTET,especiallyTET3expression,mayaid
inthedemethylationoftargetgenes.Inadditiontoenzymaticroles,TETproteinsserveas
transcriptionalco-activators/co-repressorsbyinteractingwithsometranscriptionalregulatorsand
scaffoldingproteins(39).Furtherstudiesareneededtoclarifytheeffectsofsuchnon-enzymatic
activityofTETproteinsinducedbyGCs.
Thisarticleisprotectedbycopyright.Allrightsreserved15
GCs,includingcortisolasthepredominantGCinhuman,regulatetheexpressionofawidearray
oftargetgenesthroughbindingtoglucocorticoidreceptorα(GRα)andGRβ.Botharemembersof
thenuclearreceptorsuperfamilyandpresentinosteocytes(40,41).GCsleadtomultiplechangesof
osteocytes;however,studiesregardingtherolesofGRsinmediatingtheseeffectsarescarce(41).It
hasbeenreportedthatGRsmediateGCpromotionofosteocyteapoptosisbyactivatingPyk2and
JNK,followedbyinside-outsignalingthatleadstoanoikis(42).OurdemonstrationthatGCsinduce
osteocyteapoptosisisconsistentwithpreviousstudies(18,43-45).WefoundthatGCsinducedTET3
up-regulationandlocalizationtothenucleus,aswellas5hmCenrichmentinatime-dependent
manner,inosteocytes.Thus,itissuspectedthatGRshavethepotentialtomediatetheexpressionof
TET3,initiategenomichydroxymethylationandosteocyteapoptosis.However,howTET3is
modulatedbyGRsandrecruitedintothenucleustomodifymethylationneedstobefurtherstudied.
AhMeDIP-chipassaywasperformedtounderstandthefunctionaleffectof5hmCenrichment.
Thedynamicchangesin5hmCdistributionwithDextreatmentaresimilartothe5hmCdistribution
inprogenitordifferentiationduringchondrogenesisandneurogenesis(12,21).KEGGpathwayanalysis
indicatedthatthePI3K-AktsignalingpathwaycouldmediatethefunctionaleffectofDex-induced
5hmCchange,afindingthatwasverifiedwhenthepharmacologicalactivationoftheAktpathway
abrogatedDex-inducedosteocyteapoptosis.Significantly,wefurtherfoundalterationsinAkt
signalinginclinicalSAONsamples.ThemodulationofAktbyGCsisacellautonomous
mechanismofWnt/β-cateninantagonismthatcontributestotheadverseeffectsofexcessGCs.Akt
attenuationpromotesthebindingofβ-catenintoFoxOsandinhibitionofβ-catenin/TCFactivity(43).
PTHcounteractstheadverseeffectofGCsbyabrogatingthesuppressiveeffectofGCsonAkt
phosphorylationandtheWntpathway(46).Moreover,inhibitionofAkt-mTORC1signalingbyGCs
inducesautophagy,resultinginconnexin43degradationandsubsequentlyimpairingosteocyte
Thisarticleisprotectedbycopyright.Allrightsreserved16
cell-cellcommunication(20).Interestingly,knockdownofTET3invitrocanblocktheAktactivity
inhibitioninducedbyDex.Thus,ourdataforthefirsttimeindicatethatDex-inducedAktpathway
suppressionismediatedbyTET3.
InadditiontoAktpathway-associatedgenes,GC-inducedTET3-5hmCchangesregulatethe
expressionofalargenumberofgenes,suchasNotch,Dkk1andDkk4(datanotshown).Notch
signalingplaysacriticalroleinosteoblastcellfateandfunction,andendothelialNotchactivity
promotesangiogenesisandosteogenesis(47).Dkk1isinvolvedintheinhibitionoftheWntsignaling
pathway,anditsexpressionisassociatedwithosteocyteapoptosisinSAONpatients(48).Boseetal.
reportedthatTET3mediatesstableGC-inducedalterationsinDNAmethylationandDkk1
expressioninneuralprogenitors(32).ThisindicatesthatnotonlytheAktpathwaybutalsoother
targetgenesmaymediatethefunctionaleffectofGC-inducedTET3-5hmCchangeandcouldbe
furtherexplored.
Then,wefoundthatknockdownofTET3withspecificsiRNAtransfection,whichhasbeenused
inmanystudies,couldabrogateSAONinvivo(2,25).InanotherSAONrabbitmodel,theblockade
ofSrcwithspecificsiRNAwasusedasanoveltherapeuticstrategytopreventdestructiverepair(2).
Inourstudy,theeffectiveknockdownofTET3invitroconfirmedthespecificmouseTET3siRNA
sequence.ConsideringthefactthatweusedratstoestablishanSAONmodel,aBLASTsearchwas
performedinGenBanktovalidatethatthesequencewasconservedandalsospecificforrats.
Meanwhile,theefficiencyofTET3knockdowninvivowasconfirmedwithawesternblotassayof
theentirefemoralheadbonetissue.Furthermore,withtheaimofexploringwhetherknockdownof
TET3couldpreventthedevelopmentofSAON,weperformedTET3siRNAinjectiononthefirst
andfifteenthdayofSAONinduction.Interestingly,knockdownofTET3invivoprevented
GC-inducedosteocyteapoptosis,histopathologychanges,trabeculaedeclineandhigh-intensity
Thisarticleisprotectedbycopyright.Allrightsreserved17
T2W1MRIsignalsinratfemoralheads.Unfortunately,ourinvivotransfectionmethodcouldnot
ensurethespecificknockdownofTET3inosteocytes.Liangetal.developedCH6aptamer-
functionalizedlipidnanoparticles(LNPs)toencapsulatesiRNAandboostedinvivo
osteoblast-specificgenesilencing(49).TounderstandtheexactfunctionofTET3inosteocytes,
furtherosteocyte-specificdeletionofTET3withacorrespondingaptamerdeliverysystemis
warranted.
Inconclusion,thecurrentresultsindicatethatTET3-5hmC-AktsignalingiscriticalforGC-induced
osteocyteapoptosisandSAONanduncoveranewmechanismofSAON.Thisfindingindicatesthat
inhibitionofTET3expressioninosteocytesmightbeanewstrategytotreatSAONearly.
Disclosures
Allauthorsstatethattheyhavenoconflictsofinterest.
Acknowledgments
WearegratefultoProf.LyndaBonewaldforthekindgiftofMLO-Y4cells.Thisworkwas
supportedbygrantsfromtheNationalNaturalScienceFoundationofChina(No.81572154),the
ScientificandTechnologicalProjectofTianjinPublicHealthBureau(No.2014KY31)andthe
HealthCareKeyProjectofTianjin(No.14KG123).
Authors’roles:Studydesign:XLMa,JXMa,JZhaoandWLJin.Studyconduct:JZhao,LSun,B
Lu,LYXu,MJKuang,LFu,YMa.Datacollection:JZhao,BLu,LYXu,MJKuang,YWang,BC
Dong,LFu.Dataanalysis:JZhao,YWang,GSXing,HHBai,YMa,WLJin.Draftingmanuscript:
JZhao,JXMa.Revisingmanuscriptcontent:JZhao,JXMa,XLMa,WLJin.Approvingfinal
versionofthemanuscript:All.JZhao,JXMaandXLMaareresponsiblefortheintegrityofthe
dataanalysis.
Thisarticleisprotectedbycopyright.Allrightsreserved18
References
1.WeinsteinRS.Glucocorticoid-inducedosteonecrosis.Endocrine.2012;41(2):183-90.
2.ZhengLZ,CaoHJ,ChenSH,TangT,FuWM,HuangL,etal.BlockageofSrcbySpecific
siRNAasaNovelTherapeuticStrategytoPreventDestructiveRepairinSteroid-Associated
OsteonecrosisinRabbits.JournalofBoneandMineralResearch.2015;30(11):2044-57.
3.StrehlC,BijlsmaJW,deWitM,BoersM,CaeyersN,CutoloM,etal.Definingconditions
wherelong-termglucocorticoidtreatmenthasanacceptablylowlevelofharmtofacilitate
implementationofexistingrecommendations:viewpointsfromanEULARtaskforce.Annalsofthe
rheumaticdiseases.2016;75(6):952-7.
4.SunZ,YangS,YeS,ZhangY,XuW,ZhangB,etal.AberrantCpGislands''hypermethylation
ofABCB1inmesenchymalstemcellsofpatientswithsteroid-associatedosteonecrosis.J
Rheumatol.2013;40(11):1913-20.
5.LeeYK,HaYC,ChoYJ,SuhKT,KimSY,WonYY,etal.DoesZoledronatePreventFemoral
HeadCollapsefromOsteonecrosis?AProspective,Randomized,Open-Label,MulticenterStudy.
TheJournalofboneandjointsurgeryAmericanvolume.2015;97(14):1142-8.
6.WeinsteinRS.Glucocorticoid-inducedosteoporosisandosteonecrosis.EndocrinolMetabClin
NorthAm.2012;41(3):595-611.
7.KerachianMA,SeguinC,HarveyEJ.Glucocorticoidsinosteonecrosisofthefemoralhead:a
newunderstandingofthemechanismsofaction.JSteroidBiochemMolBiol.2009;114(3-5):121-8.
8.WeinsteinRS.Clinicalpractice.Glucocorticoid-inducedbonedisease.NEnglJMed.
2011;365(1):62-70.
9.MinorEA,CourtBL,YoungJI,WangG.Ascorbateinducesten-eleventranslocation(Tet)
Thisarticleisprotectedbycopyright.Allrightsreserved19
methylcytosinedioxygenase-mediatedgenerationof5-hydroxymethylcytosine.TheJournalof
biologicalchemistry.2013;288(19):13669-74.
10.SunZB,WangJW,XiaoH,ZhangQS,KanWS,MoFB,etal.Icariinmaybenefitthe
mesenchymalstemcellsofpatientswithsteroid-associatedosteonecrosisbyABCB1-promoter
demethylation:apreliminarystudy.Osteoporosisinternational.2015;26(1):187-97.
11.TahilianiM,KohKP,ShenY,PastorWA,BandukwalaH,BrudnoY,etal.Conversionof
5-methylcytosineto5-hydroxymethylcytosineinmammalianDNAbyMLLpartnerTET1.Science.
2009;324(5929):930-5.
12.TaylorSE,LiYH,SmeriglioP,RathM,WongWH,BhutaniN.Stable
5-Hydroxymethylcytosine(5hmC)AcquisitionMarksGeneActivationDuringChondrogenic
Differentiation.JournalofBoneandMineralResearch.2016;31(3):524-34.
13.SongCX,SzulwachKE,DaiQ,FuY,MaoSQ,LinL,etal.Genome-wideprofilingof
5-formylcytosinerevealsitsrolesinepigeneticpriming.Cell.2013;153(3):678-91.
14.SunW,ZangL,ShuQ,LiX.Fromdevelopmenttodiseases:theroleof5hmCinbrain.
Genomics.2014;104(5):347-51.
15.FrandsenPA,AndersenE,MadsenF,SkjodtT.Garden''sclassificationoffemoralneckfractures.
Anassessmentofinter-observervariation.TheJournalofboneandjointsurgeryBritishvolume.
1988;70(4):588-90.
16.FicatRP.Idiopathicbonenecrosisofthefemoralhead.Earlydiagnosisandtreatment.The
JournalofboneandjointsurgeryBritishvolume.1985;67(1):3-9.
17.KimT,HongJM,LeeJ,OhB,ParkEK,LeeC,etal.Promoterpolymorphismsofthevascular
endothelialgrowthfactorgeneisassociatedwithanosteonecrosisofthefemoralheadintheKorean
population.OsteoarthritisCartilage.2008;16(3):287-91.
Thisarticleisprotectedbycopyright.Allrightsreserved20
18.KitaseY,BarraganL,QingH,KondohS,JiangJX,JohnsonML,etal.Mechanicalinductionof
PGE2inosteocytesblocksglucocorticoid-inducedapoptosisthroughboththebeta-cateninandPKA
pathways.JournalofBoneandMineralResearch.2010;25(12):2657-68.
19.SatoAY,TuX,McAndrewsKA,PlotkinLI,BellidoT.Preventionofglucocorticoid
induced-apoptosisofosteoblastsandosteocytesbyprotectingagainstendoplasmicreticulum(ER)
stressinvitroandinvivoinfemalemice.Bone.2014;73C:60-8.
20.GaoJ,ChengTS,QinA,PavlosNJ,WangT,SongK,etal.Glucocorticoidimpairscell-cell
communicationbyautophagy-mediateddegradationofconnexin43inosteocytes.Oncotarget.
2016;7(19):26966-78.
21.ZhaoX,DaiJ,MaY,MiY,CuiD,JuG,etal.Dynamicsoften-eleventranslocation
hydroxylasefamilyproteinsand5-hydroxymethylcytosineinoligodendrocytedifferentiation.Glia.
2014;62(6):914-26.
22.WangJH,ChenWL,LiJM,WuSF,ChenTL,ZhuYM,etal.Prognosticsignificanceof
2-hydroxyglutaratelevelsinacutemyeloidleukemiainChina.ProceedingsoftheNational
AcademyofSciencesoftheUnitedStatesofAmerica.2013;110(42):17017-22.
23.HuangdaW,ShermanBT,LempickiRA.Systematicandintegrativeanalysisoflargegenelists
usingDAVIDbioinformaticsresources.Natureprotocols.2009;4(1):44-57.
24.HanN,YanZ,GuoCA,ShenF,LiuJ,ShiY,etal.Effectsofp-glycoproteinonsteroid-induced
osteonecrosisofthefemoralhead.CalcifTissueInt.2010;87(3):246-53.
25.LiCJ,ChengP,LiangMK,ChenYS,LuQ,WangJY,etal.MicroRNA-188regulates
age-relatedswitchbetweenosteoblastandadipocytedifferentiation.TheJournalofclinical
investigation.2015;125(4):1509-22.
26.YanW,FangZ,YangQ,DongH,LuY,LeiC,etal.SirT1mediateshyperbaricoxygen
Thisarticleisprotectedbycopyright.Allrightsreserved21
preconditioning-inducedischemictoleranceinratbrain.Journalofcerebralbloodflowand
metabolism:officialjournaloftheInternationalSocietyofCerebralBloodFlowandMetabolism.
2013;33(3):396-406.
27.YangL,BoydK,KasteSC,KamdemKamdemL,RahijaRJ,RellingMV.Amousemodelfor
glucocorticoid-inducedosteonecrosis:effectofasteroidholiday.JOrthopRes.2009;27(2):169-75.
28.LiuC,JankeLJ,KawediaJD,RamseyLB,CaiX,MattanoLA,Jr.,etal.Asparaginase
PotentiatesGlucocorticoid-InducedOsteonecrosisinaMouseModel.PloSone.
2016;11(3):e0151433.
29.DongY,LiY,HuangC,GaoK,WengX.Systemicapplicationofteriparatideforsteroid
inducedosteonecrosisinaratmodel.BMCMusculoskeletDisord.2015;16:163.
30.BouxseinML,BoydSK,ChristiansenBA,GuldbergRE,JepsenKJ,MullerR.Guidelinesfor
assessmentofbonemicrostructureinrodentsusingmicro-computedtomography.JournalofBone
andMineralResearch.2010;25(7):1468-86.
31.JilkaRL,NobleB,WeinsteinRS.Osteocyteapoptosis.Bone.2013;54(2):264-71.
32.BoseR,SpulberS,KilianP,HeldringN,LonnerbergP,JohnssonA,etal.Tet3mediatesstable
glucocorticoid-inducedalterationsinDNAmethylationandDnmt3a/Dkk1expressioninneural
progenitors.CellDeathDis.2015;6:e1793.
33.RasmussenKD,HelinK.RoleofTETenzymesinDNAmethylation,development,andcancer.
GenesDev.2016;30(7):733-50.
34.ItoS,D''AlessioAC,TaranovaOV,HongK,SowersLC,ZhangY.RoleofTetproteinsin5mC
to5hmCconversion,ES-cellself-renewalandinnercellmassspecification.Nature.
2010;466(7310):1129-33.
35.FuHL,MaY,LuLG,HouP,LiBJ,JinWL,etal.TET1exertsitstumorsuppressorfunctionby
Thisarticleisprotectedbycopyright.Allrightsreserved22
interactingwithp53-EZH2pathwayingastriccancer.Journalofbiomedicalnanotechnology.
2014;10(7):1217-30.
36.MiY,GaoX,DaiJ,MaY,XuL,JinW.ANovelFunctionofTET2inCNS:Sustaining
NeuronalSurvival.IntJMolSci.2015;16(9):21846-57.
37.XuY,XuC,KatoA,TempelW,AbreuJG,BianC,etal.Tet3CXXCdomainanddioxygenase
activitycooperativelyregulatekeygenesforXenopuseyeandneuraldevelopment.Cell.
2012;151(6):1200-13.
38.HuL,LuJ,ChengJ,RaoQ,LiZ,HouH,etal.Structuralinsightintosubstratepreferencefor
TET-mediatedoxidation.Nature.2015;527(7576):118-22.
39.GaoJ,MaY,FuHL,LuoQ,WangZ,XiaoYH,etal.Non-catalyticrolesforTET1protein
negativelyregulatingneuronaldifferentiationthroughsrGAP3inneuroblastomacells.Protein&
cell.2016;7(5):351-61.
40.SurjitM,GantiKP,MukherjiA,YeT,HuaG,MetzgerD,etal.Widespreadnegativeresponse
elementsmediatedirectrepressionbyagonist-ligandedglucocorticoidreceptor.Cell.
2011;145(2):224-41.
41.MoutsatsouP,KassiE,PapavassiliouAG.Glucocorticoidreceptorsignalinginbonecells.
Trendsinmolecularmedicine.2012;18(6):348-59.
42.PlotkinLI,ManolagasSC,BellidoT.Glucocorticoidsinduceosteocyteapoptosisbyblocking
focaladhesionkinase-mediatedsurvival.Evidenceforinside-outsignalingleadingtoanoikis.The
Journalofbiologicalchemistry.2007;282(33):24120-30.
43.AlmeidaM,HanL,AmbroginiE,WeinsteinRS,ManolagasSC.Glucocorticoidsandtumor
necrosisfactoralphaincreaseoxidativestressandsuppressWntproteinsignalinginosteoblasts.
TheJournalofbiologicalchemistry.2011;286(52):44326-35.
Thisarticleisprotectedbycopyright.Allrightsreserved23
44.PlotkinLI,MathovI,AguirreJI,ParfittAM,ManolagasSC,BellidoT.Mechanicalstimulation
preventsosteocyteapoptosis:requirementofintegrins,Srckinases,andERKs.AmJPhysiolCell
Physiol.2005;289(3):C633-43.
45.WangZ,GuoJ.MechanicalinductionofBMP-7inosteocyteblocksglucocorticoid-induced
apoptosisthroughPI3K/AKT/GSK3betapathway.CellBiochemBiophys.2013;67(2):567-74.
46.WeinsteinRS,JilkaRL,AlmeidaM,RobersonPK,ManolagasSC.Intermittentparathyroid
hormoneadministrationcounteractstheadverseeffectsofglucocorticoidsonosteoblastand
osteocyteviability,boneformation,andstrengthinmice.Endocrinology.2010;151(6):2641-9.
47.RamasamySK,KusumbeAP,WangL,AdamsRH.EndothelialNotchactivitypromotes
angiogenesisandosteogenesisinbone.Nature.2014;507(7492):376-80.
48.KoJY,WangFS,WangCJ,WongT,ChouWY,TsengSL.IncreasedDickkopf-1expression
acceleratesbonecellapoptosisinfemoralheadosteonecrosis.Bone.2010;46(3):584-91.
49.LiangC,GuoB,WuH,ShaoN,LiD,LiuJ,etal.Aptamer-functionalizedlipidnanoparticles
targetingosteoblastsasanovelRNAinterference-basedboneanabolicstrategy.Naturemedicine.
2015;21(3):288-94.
Thisarticleisprotectedbycopyright.Allrightsreserved24
Figurelegends
Figure1.TET3-5hmCisup-regulatedinthefemoralheadtissuesofpatientswithSAON.
A.RepresentativeradiographofSAONandfemoralneckfracturepatients.B,C.TUNELanalysis
ofosteocyteapoptosisinSAONandfemoralneckfracture.TUNELpositiveosteocytesappearred.
Nuclei(blue)arecounterstainedwithDAPI.Themergedimageisshowninthebottompanel.Scale
bar=30μm.TheTUNELresultsareexpressedasthepercentageofapoptoticcellsrelativetototal
cells.Dataareexpressedasthemean±SDfromthreeindependentbiologicalreplicates.D-F.
QuantificationofrelativemRNAlevelsofTET1,TET2andTET3inSAONandfemoralneck
fracture.Dataarerepresentedasthemean±SDfromthreeindependentbiologicalreplicates.Inall
graphs,theFracturecontrolgroupwassetto1,andSAONisexpressedasthefoldchangerelative
toFracture.G,H.WesternblotanalysisofTET3inSAONandfemoralneckfracture;atotalof20
μgproteinwasloaded(β-actinwasusedasacontrolforloading).Therelativeexpressionlevelof
TET3wascalculatedasTET3/β-actin.I,J.Quantificationofglobal5hmCinSAONandfemoral
neckfractureswithdotblot;atotalof500ngDNAwasloaded(DNAwasstainedwithmethylene
blueasacontrolforloading).Therelativelevelof5hmCwascalculatedas5hmC/DNA.p<0.05,
p<0.01.
Figure2.DexinducesTET3-5hmCup-regulationinatime-dependentmanner.Alsosee
SupplementalFigure1andsupplementalFigure2.
A,B.TUNELanalysisofMLO-Y4celltreatedwithvaryingconcentrations(10-8Mto10-5M)of
Dexfor12h.C,D.TUNELanalysisofMLO-Y4celloverthecourseofDex(10-6M)treatment.
Dataarerepresentedasthemean±SDfromthreeindependentbiologicalreplicates.Scalebar=30
Thisarticleisprotectedbycopyright.Allrightsreserved25
μm.E.Immunostainingof5hmCinMLO-Y4cellswithDextreatment(green).Nuclei(blue)are
counterstainedwithDAPI;themergedimageisshowninthebottompanel.Scalebar=30μm.F.
WesternblotanalysisofBcl-2,Baxandcleavedcaspase-3inDex-treatedMLO-Y4cells.Inall
graphs,thecontrolissetto1,andtreatmentsarereferredtoasthefoldchangerelativetocontrol.G.
Dotblotanalysisofglobal5hmCinDex-treatedMLO-Y4cells.Leftpanel:dotblotof5hmC,right
panel:methylenebluestaining.H-J.RelativegeneexpressionlevelsofTET1,TET2andTET3in
MLO-Y4cellswithDextreatment.Dataarerepresentedasthemean±SDfromthreeindependent
biologicalreplicates.Inallgraphs,thecontrolissetto1,andtreatmentsarereferredtoasthefold
changerelativetocontrol.K.WesternblotanalysisofTET1,TET2andTET3inDex-treated
MLO-Y4cells;atotalof50μgproteinwasloaded.L.SubcellularanalysisofnuclearTET3with
Dextreatment(histoneH3servedasaloadingcontrol).Twentymicrogramsofnuclearproteinwere
loaded.p<0.05,p<0.01.
Figure3.Dynamicchangesinglobal5hmCwithDextreatment
A.Aheatmapof456geneswithdifferential5hmCenrichmentinMLO-Y4cellswithDex
treatmentatdifferenttimes.Redindicatesupregulationof5hmC;greenindicatesareduction.①
5hmCup-regulationat4hand12h,②5hmCup-regulationat12h,③5hmCdown-regulationat4
hand12h,④5hmCdown-regulationat12h.B.Venndiagramsshowingdynamic5hmCchangein
HCG,ICG,LCGandtotalgenes.Thenumberofgeneswith5hmCup-regulationislistedonthe
upperpanel.Thenumberofgeneswith5hmCdown-regulationislistedonthelowerpanel.①
5hmCup-regulationat12hbutnot4h,②5hmCup-regulationat4hbutnot12h,③5hmC
up-regulationatboth12hand4h,④5hmCdown-regulationat12hbutnot4h,⑤5hmC
down-regulationat4hbutnot12h,⑥5hmCdown-regulationatboth12hand4h.C.
Thisarticleisprotectedbycopyright.Allrightsreserved26
RepresentationofGOtermresultsfromDEGforbiologicalprocess.Upperpanel:5hmC
up-regulatedgenes,lowerpanel:5hmCdown-regulatedgenes.
Figure4.TheAktsignalingpathwayismodifiedbyDex-inducedTET3-5hmCchange.
A,B.KEGGpathwayanalysisofDEGduringDextreatment.Leftpanel:5hmCup-regulatedgenes,
rightpanel:5hmCdown-regulatedgenes.Corresponding–log10(pValue)areshown.Dataanalysis
wasperformedusingDAVID(https://david.ncifcrf.gov/).C-M.Relativegeneexpressionlevelsof
PIK3r5,PIK3cd,PIK3c2b,Ppp2r5d,PTEN,Itga1,Itga11,TCL1,Tlr4,PDK2andJAK3in
Dex-treatedMLO-Y4cells.Dataarerepresentedasthemean±SDfromthreeindependent
biologicalreplicates.Inallgraphs,thecontrolissetto1,andtreatmentsareexpressedasthefold
changerelativetocontrol.N.WesternblotanalysisofPTEN,p-AktandAktinMLO-Y4cellswith
Dextreatment.p<0.05,p<0.01.
Figure5.TheAktsignalingpathwaymediatesthefunctionaleffectofDex-induced
TET3-5hmCchange.
A,B.TUNELanalysisofMLO-Y4cellspretreatedwith10-6MbpV(phen)for30minfollowedby
10-6MDexfor12hours.Dataarerepresentedasthemean±SDfromthreeindependentbiological
replicates.Scalebar=30μm.C.WesternblotanalysisofBcl-2,Baxandcleavedcaspase-3in
MLO-Y4cells.D.Westernblotanalysisofp-Akt,Akt,PTEN,p-PI3K,PI3K,Bcl-2,Baxand
cleavedcaspase-3infemoralneckfractureandSAONbonetissues.p<0.05,p<0.01.
Figure6.KnockdownofTET3abrogatesDex-inducedosteocyteapoptosisandAktpathway
inhibition.AlsoseeSupplementalFigure4.
Thisarticleisprotectedbycopyright.Allrightsreserved27
A.RelativeexpressionlevelofTET1-3inMLO-Y4cellswithspecificsiRNAtreatment.Dataare
representedasthemean±SDfromthreeindependentbiologicalreplicates.Inallgraphs,scramble
issetto1,andtreatmentsareexpressedasthefoldchangerelativetoscramble.B,C.TUNEL
analysisofMLO-Y4cellswithknockdownofTET1-3.Dataarerepresentedasthemean±SDfrom
threeindependentbiologicalreplicates.Scalebar=30μm.D.WesternblotanalysisofBcl-2,Bax
andcleavedcaspase-3inMLO-Y4cellswithknockdownofTET1-3.E.Westernblotand
subcellularanalysisofTET3inMLO-Y4cellswithknockdownofTET3.Fiftymicrogramsoftotal
proteinand20μgnuclearproteinwereloaded.F.Quantificationof5hmClevelsinMLO-Y4cells
withknockdownofTET3bydotblot.Leftpanel:dotblotof5hmC,rightpanel:methyleneblue
staining.G.Immunostainingof5hmCinMLO-Y4cellswithknockdownofTET3(green).Nuclei
(blue)arecounterstainedwithDAPI.Themergedimageisshowninthebottompanel.Scalebar=30
μm.H.WesternblotanalysisofPTEN,p-AktandAkt.p<0.05,p<0.01.
Figure7.KnockdownofTET3alleviatesSAONinrats.
A.SAONwasinducedinSDrats,andinvivoTET3siRNAtransfectionwasperformedasindicated.
B.WesternblotanalysisofTET1,TET2andTET3inthefemoralheadtissuesofrats.Atotalof50
μgproteinwasloaded.C.Dotblotanalysisof5hmCinthefemoralheadtissuesofrats.Atotalof
500ngDNAwasloaded.Leftpanel:dotblotof5hmC,rightpanel:methylenebluestaining.D.
RepresentativeHEstainingoftrabeculaeattheproximalepiphysisofthefemoralhead.Leftpanel:
scalebar=500μm,rightpanel:magnificationofboxedareaintheleftpanel,scalebar=50μm.E,F.
TUNELanalysisofosteocyteapoptosisintrabeculaeattheproximalepiphysisofthefemoralhead
(Red).Dataarerepresentedasthemean±SDfromthreeindependentbiologicalreplicates.Scale
bar=30μm.G.RepresentativeMicro-CTscanofthefemoralhead.H.Quantificationoftrabecular
Thisarticleisprotectedbycopyright.Allrightsreserved28
bonestructurewithintheROIinFigure7G.TrabecularBV/TV,BS/BV,trabecularthickness,
trabecularnumber,andtrabecularseparationwerecalculated.I.RepresentativeMRIT1W1and
T2W1scansofthefemoralhead.TheboxedareasindicatedifferentsignalinT1W1andT2W1scan.
p<0.05,p<0.01.
Figure8.SchematicillustrationoftheeffectoftheTET3-5hmC-AktpathwayonSAON
pathogenesis
TheseobservationsledustohypothesizethatGC-inducedTET3-5hmCchangecouldup-regulate
PTENexpressionanddown-regulatePI3Kexpressioninadditiontoregulatingothertargetgenesto
inhibittheAktpathwayandpromoteosteocyteapoptosisinSAON.
Thisarticleisprotectedbycopyright.Allrightsreserved29
Table1.ThepeakMvalueof11genesassociatedwithAktpathwayatindicatedtimepointwas
listed.
GeneControl4h12h
PIK3r50.6762391.1287061.036246
PIK3cd1.0436551.3563830.977372
PIK3c2b1.3076171.2236271.313212
Ppp2r5d1.6128711.1335851.649453
PTEN1.2731411.2838060.988902
Itga12.0824741.9839411.605688
Itga111.3976811.5865481.605688
TCL11.0716841.1996591.231157
Tlr41.3348231.2708011.635225
PDK20.9479821.2355551.283706
JAK31.2920341.4071951.448056
ThepeakMvaluerevealtherelativequantificationof5hmC.
Thisarticleisprotectedbycopyright.Allrightsreserved30
Figure1
Thisarticleisprotectedbycopyright.Allrightsreserved31
Figure2
Thisarticleisprotectedbycopyright.Allrightsreserved32
Figure3
Thisarticleisprotectedbycopyright.Allrightsreserved33
Figure4
Thisarticleisprotectedbycopyright.Allrightsreserved34
Figure5
Thisarticleisprotectedbycopyright.Allrightsreserved35
Figure6
Thisarticleisprotectedbycopyright.Allrightsreserved36
Figure7
Thisarticleisprotectedbycopyright.Allrightsreserved37
Figure8
Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.
|
|
