Neuropathy of haematopoietic stem cell niche is essential fo

LETTER

doi:10.1038/nature13383

Neuropathyofhaematopoieticstemcellnicheisessentialformyeloproliferativeneoplasms

′nchez-Aguilera1,DanielMart?′n-Pe′rez1,JoanIsern1,XavierLanga1,AlexandarTzankov2,LorenaArranz1,AbelSa

′n3,Yi-ShiuanTzeng4,Dar-MingLai4,Ju¨rgSchwaller2,RadekC.Skoda2PontusLundberg2,SandraMuntio

′nMe′ndez-Ferrer1&Simo

Myeloproliferativeneoplasms(MPNs)arediseasescausedbymuta-tionsinthehaematopoieticstemcell(HSC)compartment.MostMPNpatientshaveacommonacquiredmutationofJanuskinase2(JAK2)geneinHSCs1–4thatrendersthiskinaseconstitutivelyactive,leadingtouncontrolledcellexpansion.Thebonemarrowmicroenvironmentmightcontributetotheclinicaloutcomesofthiscommonevent.Wepreviouslyshowedthatbonemarrownestin1mesenchymalstemcells(MSCs)innervatedbysympatheticnervefibresregulatenormalHSCs5,6.HerewedemonstratethatabrogationofthisregulatorycircuitisessentialforMPNpathogenesis.Sympatheticnervefibres,supportingSchwanncellsandnestin1MSCsareconsistentlyreducedinthebonemarrowofMPNpatientsandmiceexpressingthehumanJAK2(V617F)mutationinHSCs.Unexpectedly,MSCreductionisnotduetodifferentiationbutiscausedbybonemarrowneuraldamageandSchwanncelldeathtrig-geredbyinterleukin-1bproducedbymutantHSCs.Inturn,invivodepletionofnestin1cellsortheirproductionofCXCL12expandedmutantHSCnumberandacceleratedMPNprogression.Incontrast,

a

ControlhumanControl human

MPN

b

10 mRNA (ratio)86420

NES

(human)

c

20 mRNA (ratio)151050

Nes(mouse)

**

*

CJAK2(V617F)

d

CD34C MPNCMPNNes-GFP

administrationofneuroprotectiveorsympathomimeticdrugspre-ventedmutantHSCexpansion.Treatmentwithb3-adrenergicagonists

thatrestoredthesympatheticregulationofnestin1MSCs5,6preventedthelossofthesecellsandblockedMPNprogressionbyindirectlyreducingthenumberofleukaemicstemcells.Ourresultsdemon-stratethatmutant-HSC-drivennichedamagecriticallycontributestodiseasemanifestationinMPNandidentifyniche-formingMSCsandtheirneuralregulationaspromisingtherapeutictargets.

Thestemcellnichehasrecentlyemergedasanoncogenicunitandanimportantelementinregulatingcancerstemcells,includingHSCs7–11.MostMPNpatientswhodonotcarrytheBCR-ABLfusionhaveanacquiredmutationinJanuskinase2(JAK2(V617F))inHSCsthatresultsincon-stitutivekinaseactivity,leadingtouncontrolledexpansionofHSCsanderythroid,megakaryocyticandmyeloidprogenitors1–4.Somaticmutationsinthrombopoietinreceptor12,13orcalreticulin14,15genesarefoundinsomeMPNpatientsandadditionalHSCmutationsalsoaffectdiseaseprogres-sion16,17.ChangesintheHSCmicroenvironmentmightalsocontributetoMPNdevelopment,andexpansionofbonemarrow(BM)fibroblastsandbone-formingcellssuggeststheparticipationofMSCs.

WepreviouslyreportedthatmouseBMnestin1MSCsarerequiredtomaintainHSCs5andthathumanBMnestin1cellscanexpandHSCs18.Herewefoundthat,despiteelevatedBMblood-vesseldensityinMPNpatients,nestin1cellnumberandnestinmessengerRNAexpressionweremarkedlyreduced(Fig.1a,b).ThiswasreproducedinMx1-cre;JAK2(V617F)

Figure1|ApoptosisofBMnestin1HSCnichecellscontributestoMPNprogression.a,BMsectionsofcontrols(left)andMPNpatients(right)immunostainedwithnestin(brown,allpanels)andCD34(red,lowerpanels;magnification3200).Numbersofnestin1nichespermm2(mean6s.d.)were1.1560.3(control)and0.1760.18(MPN;n540;P51026,Mann–WhitneyUtest).b,c,NestinmRNAexpressioninBMcellsfrom

controls(n52),MPNpatients(n511)andmice(n54).d,Nes-GFP1cellsinskullBMofcontrol(left)andMPNmice(right;n510).e–g,CD452CD312Ter1192Nes-GFP1cells(e),clonalself-renewing

mesenchymalsphere(mesensphere)-formingcells(f)andfibroblasticcolony-formingunits(CFU-F,g)inBMcellsfromwild-typemice30weeksaftertransplantationwithcontrol(n57)orMPNBMcells(n512).h,i,Lineage-tracingstudiesofnestin1cells.Femoralsectionsoftamoxifen-treated

Nes-creERT2;RCE:loxPmice28weeksaftertransplantationwithcontrol(h)orMPNBMcells(i),showingfluorescentsignalsfromGFPandnuclei

counterstainedwithDAPI(n54).j,Fractionoflive,earlyandlateapoptoticBMstromalNes-GFP1cellsfromcontrol(n57)orMPNmice(n55).k–m,Bloodcounts(k),femoraltrichromic(l)andspleenhaematoxylinandeosinstainings(m)ofNes-creERT2;iDTAandcontrolmice20weeksaftertransplantationofMPNBMcells(n53).n,Frequencyoflin2sca-11c-kit1(LSK)haematopoieticprogenitorsinBMnucleatedcells(BMNC)of

Nes-creERT2;Cxcl12fl/fl(n55)andcontrollittermates(n57)30weeksaftertransplantationwithMPNBMcellsand24weeksaftertamoxifen

treatment.c–e,j,6-8weeksafterpIpCtreatment.Scalebars(d,h),200mm.Magnification(l,m)3100.Mean6s.e.m.*P,0.05;**P,0.01(unpairedtwo-tailedttest).BM,bonemarrow.C,control(disease-free)mice.

NestinNestinBM mesenssphere-forming cells e

BM Nes-GFP+ cells2,5002,0001,5001,0005000

fg

5,000

500

3,0002,0001,000

0

ControlJAK2(V617F)hNes-creERT2;RCE

i**

BM CFU-F4,0004003002001000

*

**

CJAK2(V617F)

ControlJAK2(V617F)ERT2ET2;iDTA Nes-creERiDTAT

C JAK2

(V617F)

Control

CJAK2(V617F)

j

BM Nes-GFP+ cells (%)6040200

MxCre;JAK2(V617F)

l

Control iDTAT

*

Alive

iDTAT6420

Late apop. Early apop.Late apop.

ERT2ET2;iDTANes-creERT

k

Neutrophils per ml of blood (×106) m

Control iDTATERT2ET2;iDTA Nes-creERiDTAT

n

1.51.00.50.0Haematocrit (%) **

6040200

*

LSK cells (% BMNC)Cxcl12?/?

Nes-creERT2;Cxcl12 *1StemCellNichePathophysiologyGroup,CentroNacionaldeInvestigacionesCardiovasculares(CNIC),28029Madrid,Spain.2UniversityHospitalBasel,CH-4031Basel,Switzerland.3DepartmentofHaematology,IBSAL-HospitalUniversitariodeSalamanca,37007Salamanca,Spain.4NationalTaiwanUniversity,Taipei10002,Taiwan.00MONTH2014|VOL000|NATURE|1

?2014Macmillan Publishers Limited. All rights reservedRESEARCHLETTER

a

302010Principal component 2 0–10–20–30

–80

–60

–40P1 BM

Nes-GFP–Pdgfrα– E12.5

Schwann cellprecursors

Pdgfrα+ Sca1+

P1 BM Nes-GFP– Pdgfrα+

ControlNes-GFP+ P1 BM

Nes-GFP+ Pdgfrα+P0 Schwann cells

b

0mRNA (fold)–200–4000–500–1,000–1,500–2,0000–100,000–200,000–300,0000–1,000–2,000–3,000–4,000–5,000108642Lifrc

Nes-GFP

Sympatheticfibres

Merge

d

Schwann cells

JAK2(V617F)Nes-GFP+

Pdgfrα– Sca1–

–600–2,500–400,000Cxcl12Angpt1Kitl100mRNA (fold)80604020040302010Mbp0Mobp806040202520151052015105ControlControlP1 BM Nes-GFP+ Pdgfrα–

–20

0

20

40

Principal component 10000Plekhb1S100bPlp1Pkp4JAK2(V617F)JAK2(V617F)Nes GFP+ BM stromal cells Apoptotic (%)Sympathetic nerve fibresBM area (% control) e

0.25Th+ area (% BM)0.200.150.100.050.00f

0.6mRNA (ratio)0.40.20.0GFAP1.00.80.60.40.2C MPNGfapg

Schwann cellsBM area (% control) 150100500P = 0.06248CD105+ cells(% BMNC) 3020100241.51.00.5TUNEL+ cells (fold)200*C MPN**35030025020015010050040h

2.0****i

32***ControlLSK*8JAK2(V617F)124****Time (weeks)

0.0CJAK2(V617F)

Time (weeks)Time (weeks)

0.0JAK2(V617F)–

IL1ra–

–++–

0+JAK2(V617F)–IL1ra–+

+–––+–++

IL1raJAK2(V617F)DAPI TUNEL

MSCSchwann

Figure2|BMSchwanncelldeathtriggeredbyHSPC-derivedinterleukin-1bprecedesnestin1MSCapoptosisinMPN.a,PrincipalcomponentanalysisofcontrolandMPNBMCD452CD312Ter1192Nes-GFP1cellscomparedwithmesenchymalstromal(adultBMPdgfra1Sca11andneonatalBMPdgfra1Nes-GFP1/2cells;purpleshadedarea)andSchwanncells(embryonicday12.5(E12.5)SchwanncellprecursorsandneonatalSchwanncells;greenshadedarea;seeMethods).b,qPCRvalidation(n53).c,Nes-gfpskullBM5weeksaftertransplantationwithcontrolorMPNBMcells;fluorescentsignalsofGFP(green)andsympatheticnervefibresdetectedwithanti-tyrosinehydroxylase(Th)antibodies(red).d,Immunostainingofglialfibrillaryacidicprotein(Gfap)tovisualizeSchwanncellsincontrolandMPNBM.Scalebar,200mm(c),100mm(d).e,QuantificationofTh1fibresinBMfromcontrols(n52)andMPN

Neutrophilsper ml of blood (x106)patients(n516).f,GFAPmRNAexpressioninBMcellsofcontrols(n52),MPNpatients(n511)andmice(n54).g,Time-courseanalysesofSchwanncells(Gfap),sympatheticfibres(Th)andNes-GFP1cellapoptosisinNes-gfp;Mx1-cre;JAK2(V617F)1andcontrolmice2,4(n55)and8(n57)weeksafterpIpCtreatment.h,FrequencyofBMCD452CD312Ter1192CD1051cellsafter18-weekinterleukin-1receptorantagonist(IL1ra)treatment(n55).

i,ApoptoticrateofMSCsandSchwanncellsinvitroderivedfromneonatalBMNes-GFP1cellsandco-culturedfor24hwithcontrolorMPNadultBMlin2sca-11c-kit1(LSK)cells(6200ngml21IL1ra)(n53experiments).TUNELstaining(pink)ofSchwanncells;nucleiwerecounterstainedwithDAPI(blue).Scalebar,100mm.Mean6s.e.m.*P,0.05;**P,0.01;***P,0.001

(unpairedtwo-tailedttest).HSPC,haematopoieticstemandprogenitorcells.

a Schwann cells

Control Vehicle

4-methylcatechol b

2.01.51.00.5****Erythrocytesper ml of blood (×109)Neutrophilsper ml of blood (×106)Plateletsper ml of blood (×109)c

2520151050806040200P = 0.096420P = 0.063210*d WT0.0JAK2(V617F)–+–4MC–

BRL37344––

β3-AR KO

++–+–+

micethatdevelopedMPNuponpIpCtreatment19,20(Fig.1c).MPNmicewereintercrossedwithaNes-gfplinetolabelMSCs5withgreenfluorescentprotein(GFP).BothcompoundmutantmiceandNes-gfpanimalstrans-plantedwithmutantBMcellsdevelopedMPNandhadfewerBMMSCsdefinedbyGFP,surfacemarkerexpressionandfunctionalassays(Fig.1d–gandExtendedDataFig.1a–f).BecauseMSClosswascon-comitantwithincipientBMfibrosis,weconductedlong-terminvivolineage-tracingstudiestodeterminewhethernestin1MSCsdifferenti-ateintofibroblastsorosteoblastsinMPN,therebycontributingtothestromalchangesinthesemice19,20(ExtendedDataFig.1g–j).Unexpect-edly,thevascularpatternsofGFP1cellsweresimilartothoseinunaffected

Figure3|Treatmentwithb3-adrenergicagonistorneuroprotectivedrugblocksMPNprogression.a,b,NeurotrophicrescueofBMSchwanncellsblocksMPNprogression.Wild-typemicetransplantedwithMPNBMcellsweretreatedoveramonthwithBRL37344,theneuroprotectivedrug4-methylcatechol(4MC)orvehicle.a,SkullBMimmunostainingofglialfibrillaryacidicprotein(red)tovisualizeSchwanncells(n55);scalebar,100mm.b,Circulatingneutrophils(control,JAK2(V617F)4MC,JAK2(V617F)BRL37344,n55;JAK2(V617F)vehicle,n59).c,Circulatingerythrocytes,neutrophilsandplatelets16weeksaftertransplantationofMPNBMcellsintob3-adrenergicreceptor-deficient(KO,n58)andwild-type(WT)mice(n57).d,VanGiesonstainingsoffemoralBMofmiceinc.e–g,CompensationofBMsympatheticdamagebyselectivesympathomimeticdrugsblocksMPNprogressionandpreventsfibrosis.e,BloodcountsofWTmicetransplantedwithMPNBMcellsandchronicallytreatedwithselectiveb3-adrenergicagonist(BRL37344)orvehicle(n55).f,ReticulinandVanGiesonstainingsoffemoralBMofmiceine(magnification,3200).g,Invivodepletionofnestin1cellsreducesthe

therapeuticeffectofBRL37344.BloodcountsofiDTAcontrol(n55)andNes-creERT2;iDTAmicetreatedwithvehicle(n54)orBRL37344(n55)for6weeks.Drugtreatmentsina,b,e,fstarted4weeksaftertransplantation;treatmentingstarted2weeksaftertransplantationcombinedwithtamoxifen(4weeks).Mean6s.e.m.*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

Plateletsper ml of blood (×109)e

Neutrophilsper ml of blood (×106)WTKOWTKO6,0004,0002,0000WTKOf

JAK2(V617F) vehicleReticulin fibresVan Gieson staining

***********Haemoglobin per dl of blood (g)20100**Haematocrit (%)481216Time (weeks)JAK2(V617F) Vehicle30481216Time (weeks)

JAK2(V617F) BRL37344 80*6040200481216Time (weeks)

Leukocytesper ml of blood (x106)2015105Neutrophils per ml of blood (x106)g

**43210+–

*0Nestin+ cells+BRL37344–

–––+–––+

2|NATURE|VOL000|00MONTH2014

JAK2(V617F) BRL37344 481216Time (weeks)

?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

Nes-gfpmice(Fig.1h,i).Thus,aswasrecentlyreportedinBCR-ABL1MPN21,Nes-GFP2cellsmightgenerateexcessivefibroblastsandoste-oblasts.Nestin1MSCreductionwasinsteadexplainedbyathreefoldincreasedapoptoticrateinmutantmice(Fig.1j),andwasnotpreventedbytheJAKinhibitorruxolitinib(ExtendedDataFig.2).

Todeterminewhethernestin1MSCdeathcouldinturnstimulateMPNprogression,weselectivelydepletednestin1cellsinvivo.ThisdepletiondidnotaffectmatureBMSchwanncells,reportedtoexpressnestin22,butreducedMSCs,associatedwithincreasedwhiteandredbloodcells(Fig.1kandExtendedDataFig.3a–e).BMsectionsrevealedexcessivefibroblastsandboneformation,notyetdetectableincontrolmice(Fig.1l),consistentwithnestin1cellsnotgeneratingfibroblastsorbonecellsinMPN.Diseaseaccelerationfollowingnestin1celldepletionalsomanifestedasseverespleeninfiltration,stillabsentincontrolmice(Fig.1m).Atanearlydiseasestage,mostprimitiveHSCsshowedhighestexpan-sion,leadingtoincreasedhaematopoieticprogenitorsinBM,peripheralbloodandspleen.ThechemokineCxcl12regulatesHSCmigrationandquiescence23,24andishighlyexpressedbynestin1MSCs5.EarlyHSCmobilisationcorrelatedwithdecreasedBMCxcl12,consistentwithMSCreduction.Inaddition,Cxcl12expressiondropped70-foldinMPNBMNes-GFP1cells(ExtendedDataFig.3f–k).DeletionofCxcl12(ref.24)innestin1cellsinvivoincreasedBMhaematopoieticprogenitorsandcir-culatingplatelets(Fig.1nandExtendedDataFig.3l).MSC-derivedCxcl12canthusnegativelyregulateJAK2(V617F)1HSCexpansion.TobetterunderstandBMnestin1cellalterations,genome-wideexpress-ionwasprofiledbynext-generationsequencing.ExpressionofMSCandHSC-relatedgeneswaslowerinMPNNes-GFP1cells,whichinsteadshowedenrichmentinSchwanncellgenesandneural-relatedfunctionalcategor-ies(ExtendedDataFigs4a–fand10).PrincipalcomponentanalysesofindependentbiologicalsamplescomparedwithpubliclyavailabledatashowedthatcontrolNes-GFP1cellswereclosesttomesenchymalpro-genitors,whereasMPNNes-GFP1clusteredawayfromthemandclosetoSchwanncells(Fig.2a).Thesechanges,confirmedbyquantitativePCR(Fig.2b),suggestedanalteredHSCniche’sneuralcomponentinMPN.SympatheticnervefibresandensheathingSchwanncells,adjacenttodistinctiveNes-GFP1cells,andGFAPmRNAexpressionweremarkedlyreducedinBMofMPNpatientsandmice(Fig.2c–fandExtendedDataFig.4g–i).Time-courseanalysisshowedthatBMneuraldamageprecedesNes-GFP1cellapoptosis(Fig.2g),indicatingthatsympatheticneuro-pathycouldsensitisenestin1cellstocelldeathtriggeredbymutantcells.MultiplexELISAdetectedearlyincreasedinterleukin-1binMPNBM(ExtendedDataFig.5a).Thiscytokineanditsactivatingenzymecaspase-1wereexpressedbymonocytes,aspreviouslyreported25,butalsobyhaematopoieticprogenitors(ExtendedDataFig.5b–d).ComparedwithBMNes-GFP2stromalcells,mRNAlevelsofinterleukin-1receptoranditsantagonistwere10-and1,000-foldhigherinNes-GFP1cells,respectively,andspecificallyupregulatedinNes-GFP1cellsinMPN(ExtendedDataFig.5e,f).Thereforewechronicallytreatedmicewithanantagonistofinterleukin-1receptor.ThistreatmentreducedplateletcountsandincreasedBMMSCfrequency,associatedwithreducedcaspase-1mRNAexpressioninhaematopoieticprogenitors(Fig.2handExtendedDataFig.5g–j).WestudiedwhetherJAK2(V617F)1HSCsmightdirectlycauseBMSchwanncelldeath.UnlikeMSCs,BM-derivedSchwanncellsco-culturedwithJAK2(V617F)1haematopoieticprogenitorsshowedthreefoldhigherapoptoticrate,whichwasblockedbyinterleukin-1receptorantagonist(Fig.2i).Together,thesedatasuggestthatHSC-derivedinterleukin-1bcontributestoneuroglialdamage,whichcompromisesMSCsurvival.WethereforeinvestigatedwhethersympatheticneuropathymightunderlieHSCnichealterationsandthusrepresentatherapeutictargetinMPN.WetreatedsymptomaticMPNmicewiththeneuroprotectiveagent4-methylcatechol,whichcanprotectBMsympatheticnervefibresdur-ingchemotherapy26.Schwanncellswerepreservedin4-methylcatechol-treatedmice,associatedwithpreventedneutrophilia(Fig.3a,b).SympatheticnervefibresregulateBMHSCtrafficviab3-adrenergicreceptoractivationinnestin1MSCs5,6.ThisreceptorisnotexpressedinnormalorJAK2(V617F)1haematopoieticcells(ExtendedDataFig.6a).Diseasedevelopmentwas

Neutrophils per ml of blood (×106) Platelets per ml of blood (×106) ?2015105011IL-1β (pg per mg of protein) ?Erythrocytes per ml of blood (×109) a

12108642025???8,0006,0004,0002,0000?????b

4003002001000*******?*11????????**????1315Time (weeks)1315Time (weeks)

JAK2(V617F), veh

111315Time (weeks)JAK2(V617F), BRL37344

*Control, veh Control, BRL37344

c d

Gfap (average size) 806040200e

IL-1β (pg per mg of protein) 806040200*****JAK2(V617F), vehicle, vehicle

CXCL12 (ng per 2 femurs)JAK2(V617F), BRL37344

CFU-C per ml of blood f h g j i

BM Nes-GFP cells 2,0001,5001,0005000Control, veh CFU-C (% BMNC) *CFU-C (% BMNC) 2,500*0.30.20.10.0***LSK cells (% BMNC) 1.51.00.50.020,00015,00010,0005,00001.20.80.40.00.80.60.40.20.0****Control, BRL3734476543210JAK2(V617F), veh

2.0ST-HSCs (×106) JAK2(V617F), 4-methylcatechol JAK2(V617F), BRL37344

Negative mice (%) k

LSK cells (×106) 2.52.01.51.00.50.0***LT-HSCs (×105) ****l ****1001 in 192,641 10**1.51.00.50.0*BRL37344veh1 in 42,167 **1010203040Bone marrow nucleated cells (×104)

mBone marrowSchwann cellHealthNA SNS β3-AR Normal HSC CXCL12 Nestin+MSC CXCL12 SNS compensation (β3 adrenergic agonists) Cytokine stormNestin+ MSC apoptosis β3-AR agonistsβ3-AR No fibrosis or osteosclerosisSNS rescue/protection (4-methylcatechol) Glial and neural damageFibrosis &osteosclerosisMPN IL-1βMutant HSC CXCL12 CXCL12 Figure4|CompensationofBMneuropathyrescuesMSCsandpreventsmutantHSCexpansioninMPN.a–d,EfficacyofBRL37344treatmentinadvancedMPN;WTmicetransplantedwithcontrolorMPNBMcellsweretreatedwithBRL37344orvehicle(veh)uponthrombocytosis(8696233106and1,96862643106plateletspermlofblood,respectively;n55).a,Bloodcounts;{P,0.05,{{P,0.01,{{{P,0.001vsvehicle-treatedcontrolmice.b,IL-1bcontentinBMsupernatant.c,VanGiesonstainingoffemoralBMsections(magnification,3100).d,QuantificationofBMglialfibrillaryacidicprotein(Gfap)immunostaining.e–l,SympathomimeticdrugrestoresBMCxcl12levelsandpreventsHSCexpansionandmobilisation.WTmicetransplantedwithcontrolorMPNBMcellsweretreated4weekswithBRL37344,4-methylcatecholorvehicleover4(e,h),8(f)or16weeks(g,i,j).e,IL-1bcontentinBMsupernatant(controlvehicle,n510;

JAK2(V617F)vehicle,n511;JAK2(V617F)4-methylcatechol,JAK2(V617F)BRL37344,n55).f,BMCD452CD312Ter1192Nes-GFP1cellsinvehicle(n58)andBRL-treated(n510)MPNmice.g,Cxcl12contentinBM

supernatant(n55).h–j,Frequencyofcolony-formingunits(CFU-C)inBMnucleatedcells(BMNC)(h,n55;j,n53)andblood(i),andBMfractionoflin2sca-11c-kit1(LSK)cells(n53)(j).k,LSKcells,long-term(LT-)andshort-term(ST-)HSCs(n55)inBMfrommiceina–d.l,ReductionofleukaemicstemcellsinBRL37344-treatedmice.CD45.2WTmiceweretransplantedwithBMcellsfromCD45.1micetogetherwithBMcellsfromMPNmicetreatedwithvehicleorBRL37344(n55).Frequencyofmicewith,50%donorLSKcellchimaerism16weeksaftertransplantationisplottedagainsttestedcellnumber.MPN-initiatingcellfrequencyisindicated.*P,0.05,Pearsonchi-squaredttest.m,ModelillustratingHSCnichealterationsandrescueinMPN.MPN,myeloproliferativeneoplasm;HSC,haematopoieticstemcell;SNS,sympatheticnervoussystem;MSC,

mesenchymalstemcell;NA,noradrenaline;AR,adrenergicreceptor;C,control(disease-freemice).a–k,Mean6s.e.m.*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

00MONTH2014|VOL000|NATURE|3

?2014Macmillan Publishers Limited. All rights reservedRESEARCHLETTER

acceleratedinmicelackingtheb3-adrenergicreceptor(Fig.3c,d),unco-veringaprotectiveroleofthisreceptorinMPN.Symptomaticmicewerechronicallytreatedwithaselectiveb3-adrenergicagonist(BRL37344)tocompensatefordeficientsympatheticstimulationofnestin1MSCs.Notably,BRL37344treatmentpreventedMPN-associatedneutrophiliaandthrombocytosis,anddelayedredbloodcellreduction,butdidnotaffectbloodcountsinwild-typemice(Fig.3b,eandExtendedDataFig.6b–d).Contrastingwiththeseverefibrosisinvehicle-injectedmice,theBMofBRL37344-treatedanimalswasvirtuallydevoidofexcess-iveboneandfibroblastictissue(Fig.3f).TheseeffectswereHSCniche-dependent,becauseneither4-methylcatecholnorBRL37344affectedthegrowthofculturedhaematopoieticprogenitorsandleukocytosiswasnotrescuedbyBRL37344inmicewithnestin1celldepletion(Fig.3gandExtendedDataFig.7a).Similarly,severalMPNmarkerswereimprovedbytreatmentwiththeclinically-approvedb3-adrenergicagonistmir-abegron(ExtendedDataFig.7b),albeittoalowerextent,probablyowingtoitspoorsolubilityandrelativelylowaffinityforthemurinereceptor.Toinvestigatethepotentialtherapeuticbenefitwhenadmi-nisteredatmoreadvancedstages,thrombocytoticandcontrolmiceweretreatedwithBRL37344.Thistreatmentreducedneutrophilia,erythrocytosis,thrombocytosis,BMinterleukin-1b,fibrosisandosteo-sclerosis,rescuedBMSchwanncells(Fig.4a–d)andblockedSchwanncellprogramactivationinBMnestin1cells(ExtendedDataFig.7c).MPNprogressioncanthusbeblockedbyprotectionorrescueofBMneurogliaandbycompensationofdeficientsympatheticstimulationofnestin1MSCsbyb3-adrenergicagonists.

WenextaskedwhetherMPNblockadecouldbemediatedbypre-servationofMSCsandtheirHSCregulatoryfunction.BRL37344reducedIL-1b,restoredNes-GFP1cellnumberandincreasedCxcl12levelsinBM(Fig.4e–g).EarlyBRL37344or4-methylcatecholtreatmentpreventedmutanthaematopoieticprogenitorexpansion(Fig.4handExtendedDataFig.8).Long-termBRL37344treatmentdidnotcompromisenor-malHSCsbutefficientlydecreasedmutanthaematopoieticprogenitors(Fig.4i,jandExtendedDataFig.9),evenwhenadministeredatthrom-bocytosisstage(Fig.4k).Moreover,BRL37344-treatedMPNmiceshowed4.5-foldreductioninleukaemicstemcells(Fig.4l).

OurfindingspointtomutantHSCsasthecauseofBMneuroglialdamagethatcompromisesMSCsurvivalandfunction,criticallycon-tributingtoMPNpathogenesis(Fig.4m).ThestudyshowsthatthenichedamagetriggeredbythemutantHSCisessentialforthedevelopmentofahaematopoieticmalignancypreviouslyconsideredtobecausedbytheHSCalone.TargetingHSCniche-formingMSCsandtheirneuralregu-lationmaypavethewaytomoreefficienttherapeuticstrategiesinMPN.

6.7.8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.′ndez-Ferrer,S.,Lucas,D.,Battista,M.&Frenette,P.S.HaematopoieticstemcellMereleaseisregulatedbycircadianoscillations.Nature452,442–447(2008).Raaijmakers,M.H.etal.Boneprogenitordysfunctioninducesmyelodysplasiaandsecondaryleukaemia.Nature464,852–857(2010).Walkley,C.R.etal.Amicroenvironment-inducedmyeloproliferativesyndromecausedbyretinoicacidreceptorgammadeficiency.Cell129,1097–1110(2007).Walkley,C.R.,Shea,J.M.,Sims,N.A.,Purton,L.E.&Orkin,S.H.Rbregulatesinteractionsbetweenhematopoieticstemcellsandtheirbonemarrowmicroenvironment.Cell129,1081–1095(2007).Colmone,A.etal.Leukemiccellscreatebonemarrownichesthatdisruptthebehaviorofnormalhematopoieticprogenitorcells.Science322,1861–1865(2008).Iriuchishima,H.etal.Neovascularnicheforhumanmyelomacellsinimmunodeficientmousebone.PLoSONE7,e30557(2012).Pikman,Y.etal.MPLW515Lisanovelsomaticactivatingmutationinmyelofibrosiswithmyeloidmetaplasia.PLoSMed.3,e270(2006).Pardanani,A.D.etal.MPL515mutationsinmyeloproliferativeandothermyeloiddisorders:astudyof1182patients.Blood108,3472–3476(2006).Klampfl,T.etal.Somaticmutationsofcalreticulininmyeloproliferativeneoplasms.N.Engl.J.Med.369,2379–2390(2013).Nangalia,J.etal.SomaticCALRmutationsinmyeloproliferativeneoplasmswithnonmutatedJAK2.N.Engl.J.Med.369,2391–2405(2013).Shih,A.H.,Abdel-Wahab,O.,Patel,J.P.&Levine,R.L.Theroleofmutationsinepigeneticregulatorsinmyeloidmalignancies.NatureRev.Cancer12,599–612(2012).Vannucchi,A.M.etal.Mutationsandprognosisinprimarymyelofibrosis.Leukemia27,1861–1869(2013).Isern,J.etal.Self-renewinghumanbonemarrowmesenspherespromotehematopoieticstemcellexpansion.CellRep.3,1714–1724(2013).Tiedt,R.etal.RatioofmutantJAK2-V617Ftowild-typeJak2determinestheMPDphenotypesintransgenicmice.Blood111,3931–3940(2008).Kubovcakova,L.etal.DifferentialeffectsofhydroxyureaandINC424onmutantalleleburdenandmyeloproliferativephenotypeinaJAK2-V617Fpolycythemiaveramousemodel.Blood121,1188–1199(2013).Schepers,K.etal.Myeloproliferativeneoplasiaremodelstheendostealbonemarrownicheintoaself-reinforcingleukemicniche.CellStemCell13,285–299(2013).Yamazaki,S.etal.NonmyelinatingSchwanncellsmaintainhematopoieticstemcellhibernationinthebonemarrowniche.Cell147,1146–1158(2011).Nagasawa,T.etal.DefectsofB-celllymphopoiesisandbone-marrowmyelopoiesisinmicelackingtheCXCchemokinePBSF/SDF-1.Nature382,635–638(1996).Tzeng,Y.S.etal.LossofCxcl12/Sdf-1inadultmicedecreasesthequiescentstateofhematopoieticstem/progenitorcellsandaltersthepatternofhematopoieticregenerationaftermyelosuppression.Blood117,429–439(2011).Rameshwar,P.,Denny,T.N.,Stein,D.&Gascon,P.Monocyteadhesioninpatientswithbonemarrowfibrosisisrequiredfortheproductionoffibrogeniccytokines.Potentialroleforinterleukin-1andTGF-beta.J.Immunol.153,2819–2830(1994).Lucas,D.etal.Chemotherapy-inducedbonemarrownerveinjuryimpairshematopoieticregeneration.NatureMed.19,695–703(2013).Mignone,J.L.,Kukekov,V.,Chiang,A.S.,Steindler,D.&Enikolopov,G.Neuralstemandprogenitorcellsinnestin-GFPtransgenicmice.J.Comp.Neurol.469,311–324(2004).Balordi,F.&Fishell,G.MosaicremovalofhedgehogsignalingintheadultSVZrevealsthattheresidualwild-typestemcellshavealimitedcapacityforself-renewal.J.Neurosci.27,14248–14259(2007).Sousa,V.H.,Miyoshi,G.,Hjerling-Leffler,J.,Karayannis,T.&Fishell,G.CharacterizationofNkx6-2-derivedneocorticalinterneuronlineages.Cereb.Cortex19(Suppl.1),i1–i10(2009).Brockschnieder,D.,Pechmann,Y.,Sonnenberg-Riethmacher,E.&Riethmacher,D.AnimprovedmouselineforCre-inducedcellablationduetodiphtheriatoxinA,expressedfromtheRosa26locus.Genesis44,322–327(2006).METHODSSUMMARY

Mx1-cre;JAK2(V617F)19,Nes-gfp27,Nes-creERT2(REF28),RCE-loxP29,iDTA30,Cxcl12-floxed24,Adrb3tm1Lowl,CD45.1andCD45.2C57BL/6Jmice(JacksonLaboratories)werehousedinspecific-pathogen-freefacilities.ProtocolswereapprovedbytheAni-malWelfareEthicalCommittee.Invivotreatments,cellextraction,culture,FACS,histologicalanalyses,ELISA,qPCR,genomicandstatisticalanalysesaredetailedintheMethods.

OnlineContentMethods,alongwithanyadditionalExtendedDatadisplayitemsandSourceData,areavailableintheonlineversionofthepaper;referencesuniquetothesesectionsappearonlyintheonlinepaper.Received7July2013;accepted15April2014.Publishedonline22June2014.1.2.3.4.5.James,C.etal.AuniqueclonalJAK2mutationleadingtoconstitutivesignallingcausespolycythaemiavera.Nature434,1144–1148(2005).Kralovics,R.etal.Again-of-functionmutationofJAK2inmyeloproliferativedisorders.N.Engl.J.Med.352,1779–1790(2005).Levine,R.L.etal.ActivatingmutationinthetyrosinekinaseJAK2inpolycythemiavera,essentialthrombocythemia,andmyeloidmetaplasiawithmyelofibrosis.CancerCell7,387–397(2005).Baxter,E.J.etal.AcquiredmutationofthetyrosinekinaseJAK2inhumanmyeloproliferativedisorders.Lancet365,1054–1061(2005).′ndez-Ferrer,S.etal.MesenchymalandhaematopoieticstemcellsformaMeuniquebonemarrowniche.Nature466,829–834(2010).′n-Salamanca,A.M.Mart?′n,A.B.Ricote,AcknowledgementsWethankS.Mart?J.M.Ligos,S.BartlettandtheCNICComparativeMedicine,Genomicsand′n?ezgroupsandBioinformaticsUnitsforassistance;membersofS.M.-F.andB.Iba′ndezfordatadiscussion;G.Enikolopov,G.FishellandD.Riethmacher′a-FernaM.Garc?′nCNIC,SpanishMinistryofforprovidingmice.ThisworkwassupportedbyFundacioEconomyandCompetitiveness(TerCel,SpanishCellTherapyNetwork;PlanNacionalSAF-2011-30308)andConSEPOC-ComunidaddeMadridS2010/BMD-2542grants′nyCajalProgramgrantsRYC-2009-04703/2011-09726andMarietoS.M.-F.;RamoCuriegrantsFP7-PEOPLE-2011-RG-294262/294096toA.S.-A.andS.M.-F.;SwissNationalScienceFoundation310000-120724/1,32003BB_135712/1andSwissCancerLeagueKLS-02398-02-2009grantstoR.C.S.;A.S.-A.receivedaResearchFellowshipfromtheEuropeanHematologyAssociation.S.M.-F.issupportedinpartbyanInternationalEarlyCareerScientistgrantoftheHowardHughesMedicalInstitute.AuthorContributionsL.A.designedandperformedmostexperimentsandanalyses,preparedfiguresandwrotethemanuscript.A.S.-A.performedinvivotransplantationassays.D.M.-P.performedqPCR,genome-wideexpressionanalysesandtogetherwithX.L.providedtechnicalassistance.J.I.performedneonatalcellisolationandculture.X.L.andA.T.performedhistologicalanalyses.S.M.providedhumansamples.Y.-S.T.andD.-M.L.providedCxcl12-floxedmice.P.L.,J.S.andR.C.S.providedsamples,designedexperimentsandhumanstudies.S.M.-F.designedtheoverallstudy,supervisedtheexperimentsandwrotethemanuscript.AuthorInformationThegeneexpressiondatahavebeendepositedintheGeneExpressionOmnibus(GEO)databank(http://www.ncbi.nlm.nih.gov/geo)undertheaccessionnumberGSE55802.Reprintsandpermissionsinformationisavailableatwww.nature.com/reprints.Theauthorsdeclarenocompetingfinancialinterests.Readersarewelcometocommentontheonlineversionofthepaper.CorrespondenceandrequestsformaterialsshouldbeaddressedtoS.M.-F.(smendez@cnic.es).4|NATURE|VOL000|00MONTH2014

?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

METHODS

Humanstudy.Thestudywasapprovedbyinstitutionalreviewboards.WritteninformedconsentwasobtainedfromallpatientsinaccordancewiththeDeclarationofHelsinki.ThediagnosisofMPNwasestablishedaccordingtotherevisedcriteriaoftheWorldHealthOrganization.

Invivopharmacologicaltreatments.Theminimalrequirednumberofanimalswasused.Micewererandomizedfortreatments.ProtocolswereapprovedbytheAnimalWelfareEthicalCommittee.InMx1-cre;JAK2(V617F)19double-transgenicmice,expressionofthehumanJAK2(V617F)mutationisdrivenbytheendogenousJak2promoterandcanbeconditionallyexpressedinhaematopoieticcellsuponMyxovirusresistance-1(Mx1)-drivenCrerecombinaseactivationbypolyinosine-polycytosine(pIpC).pIpC-inducedtransgenicmiceandwild-typemicetransplantedwithBMcellsfromthesemicedevelopprogressivesymptomsofPV19,20.

Age-matched,femalewild-typeC57BL/6JorNes-gfp27micewereusedasrecipi-entsinbonemarrow(BM)transplantationassaysandforinvivopharmacologicaltreatments.Lethallyirradiated(12Gy)recipientmiceweretransplantedwith23106BMcellsfromMx-Cre;JAK2(V617F)miceinducedwithpIpC8weeksbefore,orfromCre-negativemiceasdisease-freecontrols.Theselectiveb3-adrenergicagonistBRL37344(Sigma,St.Louis,MO)wasadministeredat2mgkg21throughintraper-itoneal(i.p.)injectiontwiceperday(every10–12h).Vehicle(salinesolution)dailyinjectionswereperformedinthesameway.Unlessindicated,treatmentwasinitiated4weekspost-transplant,whenanimalsevidencedsignsofMPNinperipheralblood.Micewererandomlydistributedbeforetreatmentinitiation,anddiseasedevelop-mentwasmonitoredovertimeinperipheralbloodsamples,usinganautomatedbloodcounter.Asimilartreatmentprotocolwasperformedwiththeselectiveb3-adrenergicagonistmirabegron(2mgkg21,i.p.,twoinjectionsperdayseparated10–12h).Theneuroprotectivedrug4-methylcatechol(10mgkg21,i.p.)wasinjectedoncedaily.TheJAKinhibitorINCB018424(S-ruxolitinib,AbmoleBioscience)wasadministeredbyoralgavage(30mgkg21,twiceperdayseparated10–12h)in0.5%hydroxypropylmethylcellulose(Sigma)aftersolubilisationinDMSO.TheIL-1recep-torantagonist(Kineret,Sobi,Stockholm,Sweden)wasadministeredbysubcutaneousosmoticpumps(Alzet)infusingacontinuousdosingof40mgkg21perday.Micewereeuthanizedatdifferenttimepointsandsubjectedtocompletenecropsy;BMandspleenwereanalysedblindlybyhistologyandflowcytometry;femoraandskullwereusedforimmunostainings,andhaematopoieticandstromalcellswereaddi-tionallyusedforfunctionalassaysexvivo.

ForquantificationofHSCs,BMcellsfromtreatedmicewereusedforcompet-itiverepopulationassaysusinglimitingcelldilutions.Briefly,CD45.11competitorBMcells(23106)weretransplantedintolethallyirradiatedCD45.21recipientmice,aftermixwith43105,43104and43103donorBMcells.Peripheralbloodchimaerismwasassessedbyflowcytometryevery2–4weeks.Micewerekilled16weeksafterthetransplantandtheBMLSKcellchimaerismwasassessedbyflowcytometry.Aminimumof5í45.21chimaerismintheLSKcellcompartmentwasconsideredaspositiveinrecipientsofcontrolBMcells.InrecipientsofmutantBM,micewereconsideredaspositive(‘leukaemic’)above50í45.21chimaer-ism.ThesoftwareL-Calc(StemCellTechnologies)wasusedforquantificationofHSCsandMPN-initiatingcellsincontrolandmutantdonorBM,respectively.Nes-creERT2(ref.28)micewereinducedwithtamoxifen(Sigma)fortheindicatedtimeperiods.Controlandexperimentalmicewereinjectedi.p.with140mgkg21oftamoxifen(14mgml21solutionincornoil),3timesonalternatedays,andweresimultaneouslyfedwithphytoestrogen-freedietforoneweek,followedbytamox-ifen-containingdietTM400(Harlan).Nes-creERT2;RCE:loxP29micewereusedforinvivolineage-tracingstudies.Toselectivelydepletenestin1cells,Nes-creERT2micewerecrossedwithaCrerecombinase-induciblediphtheriatoxinmouseline(iDTA30).Nes-creERT2micewereadditionallycrossedwithconditionalCxcl12-deficientmice24toselectivelydeleteCxcl12innestin1cellsupontamoxifeninduction.

BMcellextraction,flowcytometryandfluorescence-activatedcellsorting.Forhaematopoieticcellrecovery,boneswerecrushedinamortar,filteredthrougha40-mmmeshtoobtainsinglecellsuspensions,anddepletedofredbloodcellsbylysisin0.15MNH4Clfor10minat4uC.Spleensampleswerehomogenizedandfilteredbeforelysis;bloodsamplesweredirectlylysed.Cells(13106–23106cellspersample)wereincubatedwiththeappropriatedilution(2–5mgml21)offluorescentanti-bodyconjugatesand49,6-diamidino-2-phenylindole(DAPI)fordeadcellexclusion,andanalysedonLSRFortessaflowcytometer(BDBiosciences,FranklinLakes,NJ)equippedwithFACSDivaSoftware(BDBiosciences).Thefollowingantibodieswereused:fluorescentanti-CD45.1(A20),anti-CD45.2(104),anti-B220(RA3-6B2),anti-CD11b(M1/70),anti-CD3e(145-2C11),anti-Ly-6G(1A8),anti-Sca1(E13-161.7),anti-CD34(RAM34),anti-CD135/Flt3(A2F10.1),andbiotinylatedlineageantibod-ies(anti-CD11b,anti-Gr-1,anti-Ter119,anti-B220,anti-CD3e),allfromBDBio-sciences;anti-c-kit(2B8)fromeBioscience(SanDiego,CA).Biotinylatedantibodiesweredetectedwithfluorochrome-conjugatedstreptavidin(BDBiosciences).PhenotypicpopulationsofHSCweredefinedaslong-termhaematopoieticstemcells(HSCs)

(lin2Sca11c-kit1(LSK)CD342Flt32),short-termHSCs(LSKCD341Flt32)andmultipotentprogenitors(MPP)(LSKCD341Flt31).

Forisolationofnestin1cells,boneswerecleanedfromsurroundingtissue,crushedinamortarwithapestle,andcollagenase-digested(cataloguenumber07902,StemCellTechnologies)inashakingwaterbathat37uCfor45min.Cellswerefilteredthrougha40-mmmeshanderythrocyteswerelysedaspreviouslydescribed.Theresultingbonemarrow-enrichedcellsuspensionswerecentrifuged,washedandsuspendedinPBSbuffercontaining2ttalcalfserum(FCS)forfurtheranalyses.Forcellsorting,cellswereenrichedbyimmunomagneticdepletionusingbiotinylatedanti-CD45(104),anti-CD31(MEC13.3),andanti-Ter119antibodiesfollowedbyadditionofstreptavidinmagneticbeads(BDBiosciences),accordingtothemanufacturer’srecommendations.BMstromalCD452CD312Ter1192cellswerefurtherpurifiedaccordingtoGFPfluorescenceusingaFACSAriacellsorter(BDBioscience).Forthedeterminationofapoptoticcells,sampleswerewashedwithPBSaftersurfaceantibodystainingandsubsequentlystainedwithAnnexinV-PacificBlueandSYTOXAADvanced(Invitrogen,LifeTechnologies,Paisley,UK).Forfunctionalassays,cellswereenrichedbyimmunomagneticdepletionusingbio-tinylatedCD45andTer119.Forimmunophenotypiccharacterization,totalBMsampleswerestudiedbyflowcytometryusingthefollowingadditionalantibodies:anti-CD63(NVG-2),anti-CD105(MJ7/18),anti-CD140a(APA5),anti-Vcam,anti-CD51(RMV-7),allfromBiolegend;andanti-CD90.2fromBDBiosciences.Cellculture.Colony-formingunitsinculture(CFU-C)assaywasperformedaspreviouslydescribed31.BRL37344and4-methylcathecolwereaddedtothemethyl-celluloseatthespecifiedconcentrations.Coloniesofmorethan50cellswerescoredafter7daysofincubationat37uC,5%CO2,20%O2inawater-jacketedincubator.ForCFU-Fassays,BMCD452Ter1192cellswereplatedinto6-welldishesandculturedinmaintenancemedium(a-MEM,15üSwithantibiotics).After10–12daysinculture,adherentcellswerefixedwith100%methanolandstainedwithGiemsastain(Sigma)torevealfibroblasticclusters.Colonieswithmorethan50cellswerescoredasCFU-F.

Forclonalself-renewingmesenchymalsphereformation,cellswereplatedinultra-lowadherence35-mmdishes(StemCellTechnologies).Thegrowthmediumcontained15%chickenembryoextract,preparedasdescribed32,33;0.1mMb-mercaptoethanol;1%non-essentialaminoacids(Sigma);1%N2and227supplements(Invitrogen);recombinanthumanfibroblastgrowthfactor(FGF)-basic,recombinanthumanepidermalgrowthfactor(EGF),recombinanthumanplatelet-derivedgrowthfactor(PDGF-AB),recombinanthumanoncostatinM(227aminoacidsOSM)(20ngml21)andrecombinanthumaninsulin-likegrowthfactor-1(IGF-1;40ngml21)(Peprotech)inDMEM/F12(1:1)/humanendothelial(1:2)serum-freemedium(Invitrogen).Thecultureswerekeptat37uCwith5%CO2,20%O2inawater-jacketedincubator.One-halfmediumchangeswereperformedweekly.Mesensphereswerescoredatday10.

Forco-cultureofBM-derivedSchwanncellsandMSCswithBMLSKcellsfromcontrolandmutantmice,neonatalBMCD452CD312Ter1192Nes-GFP1Pdgfra2/1cellscontainingSchwanncellprecursorsandMSCs,respectively(Isern,J.etal.,sub-mitted),weresortedandculturedunderSchwanncelldifferentiationconditions,aspreviouslydescribed34,ormesenchymalcultureconditions,inMEMasupplementedwith10ngml21PDGF-AB,15?S.SchwanncellsandMSCswereco-culturedfor24hwithprimaryBMLSKcellsisolatedbyFACSfromcontrolormutantmice.SchwanncellsderivedfromsortedGFP1Pdgfra2BMprecursorswereadditionallyincubatedwithIL1ra(200ngml21)fortheco-cultureperiod.After24hofco-culture,haematopoieticcellswerewashedawayandtheremainingadherentSchwanncellswerefixedandblindlyanalysedforapoptosisbyTUNEL.

Histologicalanalyses,immunohistochemistryandimmunofluorescence.Haema-toxylinandeosinconventionalstainingwasperformedindeparaffinisedsectionsfollowedbyre-hydration.HarrishaematoxylinsolutionwasusedforstainingandeosinYsolutionwasusedforcounterstaining.Afterbrieflyrinsingtheslidesindistilledwater,dehydrationwasquicklyperformedin70%,95%andabsoluteethylalcohol.SectionswereclearedwithxyleneandmountedinDPX(Sigma).

ForMasson’strichromestainingofcollagen,sectionswerefixedagaininBouin’ssolutionfor1hat56uCandstainedwithWeigert’sironhaematoxylinworkingsolutionfor10min,followedbyBiebrichscarlet-acidfuchsinsolutionfor10–15min.Phosphomolybdic-phosphotungsticacidsolutionwasaddedfor10–15minoruntilcollagenlosttheredstaining.Sectionsweretransferreddirectlytofastgreensolutionandstainedfor1min,brieflyrinsedandincubatedwith1?eticacidsolutionfor2-5min.Afterabriefrinseindistilledwater,dehydrationwasquicklyperformedin70%,95%andabsoluteethylalcohol;sectionswereclearedwithxyleneandmountedwithDPX.

GordonandSweet’sstainingprotocolwasusedtovisualizereticulinfibres.Briefly,deparaffinisedsectionswereoxidisedin1?idifiedpotassiumpermanganatefor5min,followedby1%oxalicacidtodecolourise,andmordantin2.5%ironalumfor15min.Sectionswereimpregnatedinammoniacalsilversolutionfor2minandreducedwith10%aqueousformalinfor2min.Afterwardsthesectionswereincubated

?2014Macmillan Publishers Limited. All rights reserved

RESEARCHLETTER

ingoldchloridefor2minandfixedwith5%aqueoussodiumthiosulphate.A15sincubationinFastgreenwasusedforcounterstaining.Afterabriefrinseindistilledwater,dehydrationwasquicklyperformedin70%,95%andabsoluteethylalcohol;sectionswereclearedwithxyleneandmountedwithDPX.

ForVanGieson’sstainingofcollagen,nucleiwerestainedwithcelestinebluefor2min,brieflyrinsedindistilledwater,incubatedwithHarrishaematoxylinsolutionfor2minandwashedunderrunningtapwaterfor5min.Curtisstain(saturatedaqueouspicricacid,1%PonceauSandglacialaceticacidmix)wasperformedfor5min,untilcollagenwaspink.Afterabriefrinseindistilledwater,dehydrationwasquicklyperformedin70%,95%andabsoluteethylalcohol;sectionswereclearedwithxyleneandmountedwithDPX.

Immunofluorescencestainingofcryostatsectionswasperformedaspreviouslydescribed8.Forwholemountstainingofthecalvaria,allincubationtimeswereextended.Theantibodiesusedwereanti-TH(RabbitpAb,Millipore)andanti-GFAP(RabbitpAb,Dako).Confocalimageswereacquiredwithalaserscanningconfocal(ZeissLSM700).Atleast3differentsectionswereusedforquantificationusingImageJsoftware.

Fornestin/CD34immunohistochemistryofhumanBMsamples,12controlBMbiopsies(2fromhealthydonors,2frompatientswithreactiveperipheralleuko-cytosis,and8performedforlymphomastaging,butunaffectedbylymphoma)and28MPN(13ofthemwereJAK2(V617F)1)werestainedfornestinapplyingthemonoclonalantibody10C2fromAbDSerotec(OBT1610)atadilutionof1:50usinganautomatedimmunostainer(Benchmark,Ventana/Roche).Antigenretrievalwasachievedbycellconditioning(CC1fromVentana/Roche)treatmentfor60min.Incubationfor60min,signalamplificationandvisualization(amplifierandchromo-genultraviewuniversaldiaminobenzidinefromVentana/Roche)followed.Fornestin/CD34doublestainings,themonoclonalantibodyQBEnd/10(Ventana/Roche790-2927)wasusedafternestindetectionandvisualizedusinganalternativechromo-genicdetectionkit(basicaminoethylcarbazolefromVentana/Roche).Thenumberofnestin1perivascularniches(eithersinglecellsorclustersofupto3cells)wasblindlyscoredintheBMsamples.Anaverageareaof7.2mm2wasevaluatedforeachcase.ForTHimmunofluorescenceofhumanBMsamples,2controland16MPN(5essentialthrombocythaemia,4chronicmyeloidleukaemia,and7primarymyelofi-brosis)BMbiopsieswereused.SectionsweredeparaffinedandantigenretrievalwasperformedwithEDTApH9.After30minofpermeabilisationinmethanol,immu-nofluorescencestainingandquantificationwasperformedasdescribedabove.ELISA.Bio-PlexProMouseTh17cytokinePanelA6-plex(M60-00007NY,Bio-Rad)wasperformedfollowingthemanufacturer’sprotocol.Cxcl12proteinlevelsweremeasuredbyconventionalELISA.Briefly,96-wellplateswerecoatedovernightat4uCwith2mgml21ofmonoclonalanti-humanandmouseCXCL12/SDF-1antibody(MAB350,R&DSystems).Afterblocking,bonemarrowextracellularfluidswereincubatedfor2hatroomtemperature,followedbyadditionofbiotinylatedanti-humanandmouseCXCL12/SDF-1antibody(BAF310,RD).Streptavidin-horseradishperoxidaseconjugate(RPN1231V,Dako)wasusedforreportingsig-nal,andreactionwasstoppedwithhorseradishperoxidasesubstrate(TMB,ES001-500ML,Chemicon,Millipore).Stardardcurvewasperformedwithrecombinanthuman,feline,rhesusmacaqueSDF-1alpha(350-NS,R&D).

RNAisolationandqPCR.RNAisolationwasperformedusingtheDynabeadsmRNADIRECTMicroKit(Invitrogen).ReversetranscriptionwasperformedusingtheReverseTranscriptionSystem(Promega),followingthemanufacturer’srecommendations.Theexpressionlevelofeachgenewasdeterminedbyusingtherelativestandardcurvemethod.Briefly,astandardcurvewasperformedbydoingserialdilutionsofamouseorhumanreferencetotalRNA(Clontech).Theexpress-ionlevelofeachgenewascalculatedbyinterpolationfromthestandardcurve.AllvalueswerenormalizedwithGapdhasendogenouscontrol.Thefollowingprimerswereused:Human:

GFAP-Fw:CCGACAGCAGGTCCATGTGGFAP-Rv:GTTGCTGGACGCCTTGC

NESTIN-Fw:CAACAGCGACGGAGGTCTCNESTIN-Rv:GCCTCTACGCTCTCTTCTTTGAGAPDH-Fw:GCATGGCCTTCCGTGTTC

GAPDH-Rv:CCTGCTTCACCACCTTCTTGATMouse:

Adrb2-Fw:AGCAATAGCAACGGCAGAACAdrb2-Rv:TTCACAAAGCCTTCCATGCCAdrb3-Fw:AAACTGGTTGCGAACTGTGGAdrb3-Rv:TAACGCAAAGGGTTGGTGAC

Angpt1-Fw:CTCGTCAGACATTCATCATCCAGAngpt1-Rv:CACCTTCTTTAGTGCAAAGGCTCasp1-Fw:TTGGAGCTCAAGTTGACCTCAGCasp1-Rv:TGTCAGAAGTCTTGTGCTCTGG

Cxcl12-Fw:TGCATCAGTGACGGTAAACCACxcl12-Rv:TTCTTCAGCCGTGCAACAATCGapdh-Fw:GCATGGCCTTCCGTGTTC

Gapdh-Rv:CCTGCTTCACCACCTTCTTGATGfap-Fw:CGGAGACGCATCACCTCTGGfap-Rv:AGGGAGTGGAGGAGTCATTCGIl1b-Fw:GAAATGCCACCTTTTGACAGTGIl1b-Rv:TGGATGCTCTCATCAGGACAGIl1r-Fw:GTGCTACTGGGGCTCATTTGTIl1r-Rv:GGAGTAAGAGGACACTTGCGAATIl1rn-Fw:GAGAAACAACCAGCTCATTGCIl1rn-Rv:GGATGCCCAAGAACACACTATGKitl-Fw:CCCTGAAGACTCGGGCCTA

Kitl-Rv:CAATTACAAGCGAAATGAGAGCCLifr-Fw:TACGTCGGCAGACTCGATATTLifr-Rv:TGGGCGTATCTCTCTCTCCTTMbp-Fw:AATCGGCTCACAAGGGATTCAMbp-Rv:TCCTCCCAGCTTAAAGATTTTGGMobp-Fw:CCAGGCTCTCCAAGAACCAGMobp-Rv:GGTCCACGATCTCACGCTTNestin-Fw:CCCTGAAGTCGAGGAGCTGNestin-Rv:CTGCTGCACCTCTAAGCGAPkp4-Fw:GAACCTGTCATACCGGCTGGPkp4-Rv:TTCCGAGTCTTTGCTGGGAGAPlekhb1-Fw:CTGGAAGCGGAATTGGTTCGPlekhb1-Rv:TGCCGTCTCGTCATGGTAGTAPlp1-Fw:TGAGCGCAACGGTAACAGGPlp1-Rv:TTCCCAAACAATGACACACCCS100b-Fw:TGGTTGCCCTCATTGATGTCTS100b-Rv:CCCATCCCCATCTTCGTCC

Slit2-Fw:CCATGTAAAAATGATGGCACCTGSlit2-Rv:ATCACAGTCCTGACCCTTGAA

RNaseqandmicroarray.Fornext-generationsequencing,totalRNAwasisolatedusingtheArcturusPicopureRNAisolationkit(LifeTechnologies)fromsmallnumbersofFACSsortedCD452CD312Ter1192GFP1cells,obtainedfromtheBMofNes-gfp;Mx1-cre;JAK2(V617F)miceandcontrollittermates6weeksafterpIpCtreatment.Eachsamplewasapoolfrom3differentanimals.RNAwasamplifiedandpreparedforRNA-SequsingtheOvationRNA-SeqSystemv2(NuGEN)followingthemanufacturer’srecommendations.TheRNAsequencinglibrarywaspreparedwiththeTruSeqRNASamplePreparationv2Kit(Illumina,SanDiego,CA)toconstructindex-taggedcDNA.Thequality,quantityandthesizedistributionoftheIlluminalibrariesweredeterminedusingtheDNA-1000Kit(AgilentBioanalyzer).LibrariesweresequencedontheGenomeAnalyzerIIx(Illumina)followingthestandardRNAsequencingprotocolwiththeTruSeqSBSKitv5.Fastqfilescontainingreadsforeachlibrarywereextractedanddemulti-plexedusingCasavav1.8.2pipeline.Sequencingadaptorcontaminationswereremovedfromreadsusingcutadaptsoftwaretool(MIT)andtheresultingreadsweremappedandquantifiedonthetranscriptome(NCBIM37Ensemblgene-build65)usingRSEMv1.1734.Expressiondatawascomparedbetweenbothsam-plesbytheanalysisofindividualselectedgenesfordifferentialexpression,andthroughgene-setenrichmentanalyses(GSEA)todetectcoordinatedchangesinsetsofgenesrepresentingpathways,functionalsignaturesortranscriptionfactortargets.GSEAwereperformedasdescribed(http://www.broadinstitute.org/gsea/index.jsp),usingaweightedstatistic,fold-changeranking,1,000gene-setpermu-tationsandseveralgenesetdatabasesfoundintheMolecularSignaturesDatabase.Formicroarrayanalyses,totalRNAwasisolatedaspreviouslydescribedfromBMCD452CD312Ter1192Nes-GFP1cellsobtainedfromNes-gfpmice10weeksaftertransplantationwithMx1-cre;JAK2(V617F)(n53)orcontrolcells(n51).RNAwasamplifiedusingtheNuGenOvationsystemandhybridizedtotheAffymetrixMoGene1.0STarray.DatanormalizationwasperformedusingtheRobustMulti-arrayAverage(RMA)algorithm.Toperformprincipalcomponentanalysis(PCA)comparisonwithpreviouslypublisheddata,GEOdatasetsweredownloadedandpre-processedusingtheGEOqueryBioconductorpackage35.Normalizeddatasetswereadjustedtothesameintensityrange,andbatcheffectcorrectionwasper-formedusingComBat36.

Statisticalanalyses.StatisticalanalysesandgraphicswerecarriedoutwithGraph-PadPrism5softwareandMicrosoftExcel.Unlessspecified,datasetswerecom-paredbyunpairedtwo-tailedt-tests;Pvalueslessthan0.05wereconsideredstatisticallysignificant.

31.Frenette,P.S.,Subbarao,S.,Mazo,I.B.,vonAndrian,U.H.&Wagner,D.D.Endothelialselectinsandvascularcelladhesionmolecule-1promotehematopoieticprogenitorhomingtobonemarrow.Proc.NatlAcad.Sci.USA95,14423–14428(1998).?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

32.33.34.Stemple,D.L.&Anderson,D.J.Isolationofastemcellforneuronsandgliafromthemammalianneuralcrest.Cell71,973–985(1992).Pajtler,K.etal.Productionofchickembryoextractforthecultivationofmurineneuralcreststemcells.J.Vis.Exp.45,e2380(2010).Biernaskie,J.A.,McKenzie,I.A.,Toma,J.G.&Miller,F.D.Isolationofskin-derivedprecursors(SKPs)anddifferentiationandenrichmentoftheirSchwanncellprogeny.NatureProtocols1,2803–2812(2007).Davis,S.&Meltzer,P.S.GEOquery:abridgebetweentheGeneExpressionOmnibus(GEO)andBioConductor.Bioinformatics23,1846–1847(2007).Johnson,W.E.,Li,C.&Rabinovic,A.AdjustingbatcheffectsinmicroarrayexpressiondatausingempiricalBayesmethods.Biostatistics8,118–127(2007).35.36.?2014Macmillan Publishers Limited. All rights reservedRESEARCHLETTER

ExtendedDataFigure1|BMMSCreductionincompoundmutantmiceandrecipientsofmutanthaematopoieticcellsisnotduetocell

differentiation.a,b,BloodcountsofNes-gfp;Mx1-cre;JAK2(V617F)andcontrolmice4-8weeksafterpIpCtreatment(a)andNes-gfpmice10-12weeksaftertransplantationwithBMcellsfromMPNandcontrolmice(b).

Notethesimilarerythrocytosis,neutrophiliaandthrombocytosisincompoundmutantmiceandNes-gfpmicetransplantedwithmutantcells.Eachdot

representsamouse.c,GFPfluorescenceinskullBMofNes-gfpmice6–8weeksaftertransplantationwithMPNandcontrolBMcells(n58;scalebar,100mm).d,FrequencyofCD452CD312Ter1192Nes-GFP1BMcellsinMPN(n59)andcontrol(n57)mice6–8weeksafterpIpCinduction.e,Giemsa-stainedfibroblasticcolony-formingunits(CFU-F)fromimmunomagnetically-enrichedBMCD452Ter1192cells30weeksaftertransplantationwithMPN

andcontrolBMcells.f,FACSanalysesofBMCD452CD312Ter1192cellsfromMPNandcontrolmice8weeksafterpIpCtreatment(n53).ThespecifiedMSCmarkerswereused.Mean6s.e.m.g,BMreticulinstaining;MPNmiceshowedincipientfibrosis(arrow)6weeksafterpIpC

treatment(magnification,3200).h,i,Lineage-tracingstudiesofBMnestin1cellsinMPN.h,Experimentaldesign.Nes-creERT2;RCE:loxPmiceweretransplantedwithMPNandcontrolBMcellsandfedwithtamoxifendiet.Diseasedevelopmentwasmonitoredover28weeks.i,BloodcountsshowingprogressiveneutrophiliaandthrombocytosisinrecipientsofMPNBMcells(n53).j,FemoralhaematoxylinandeosinstainingsshowingabnormalboneformationinrecipientsofMPNBMcells(magnification,3100).*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

ExtendedDataFigure2|TreatmentwiththeJAK1/2inhibitorruxolitinibreduceshaematopoieticcellexpansioninMPNbutitdoesnotrescueBMMSCs.Nes-gfpmiceweretransplantedwithBMcellsfrompIpC-inducedMx1-cre;JAK2(V617F)miceandtreatedwithruxolitinib(n54)orvehicle(n56)over8weeks,starting2weeksaftertransplantation.a,Peripheralbloodcounts.b,Frequencyofimmunophenotypically-definedBM

CD452CD312Ter1192cells.Mean6s.e.m.*P,0.05(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reservedRESEARCHLETTER

ExtendedDataFigure3|Depletionofnestin1cellsortheirCxcl12

productionacceleratesMPN.a,Experimentaldesignforinvivonestin1celldepletion.Nes-creERT2;iDTAandcontroliDTAmiceweretransplantedwithBMcellsfrompIpC-treatedMx1-cre;JAK2(V617F)andcontrolmiceandtreatedwithtamoxifen(n53).b–e,UnlikeSchwanncells,MSCsarereducedinNes-creERT2;iDTABM.b,c,FrequencyofBMCD452CD312Ter1192CD901cells(b)andfibroblasticcolony-formingunits(CFU-F,c)inBM

CD452Ter1192cellsfromNes-creERT2;iDTAandcontrolmiceafter4-weektamoxifentreatment.d,Expressionofglialfibrillaryacidicprotein(Gfap)mRNA(normalizedtoGapdh)intheBMofmiceina.e,QuantificationofGfapimmunostainingoffemoralBMfrommiceinb,c.f–k,EarlymutantHSCexpansionandmobilisationcorrelateswithreducedCxcl12expressioninBMnestin1cells.f,g,BMlin2sca-11c-kit1(f)CD342Flt32long-term(LT)andCD341Flt32short-term(ST)HSCs,andCD342Flt32multipotentprogenitors

(MPP)and(g)frequencyofhaematopoieticcolony-formingunitsin

culture(CFU-C)fromBMnucleatedcells,spleenorperipheralbloodfromNes-Gfp;Mx1-cre;JAK2(V617F)(n56)andcontrol(n511)mice6weeksafterpIpCtreatment.h,Appearance(inset;scalebar,1cm)andhaematoxylinandeosinstainingsofspleens(magnification,3100).i–k,Cxcl12protein(i)andmRNAlevels(j,k)inBMextracellularfluid(i),nucleatedcells(j)andstromalNes-GFP1cells(k)ofMx1-cre;JAK2(V617F)andcontrolmice6weeksafterpIpCtreatment(i,j)orNes-gfpmice10weeksaftertransplantationwithmutantBM(n54)orcontrolcells(n57)(k).l,CirculatingplateletsinNes-creERT2;Cxcl12fl/fl(n511)andcontrolCxcl12fl/fllittermates(n514)18weeksaftertransplantationwithBMcellsfromMPNmiceand12weeksaftertamoxifentreatment.Mean6s.e.m.*P,0.05,**P,0.01(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reserved

LETTERRESEARCH

ExtendedDataFigure4|NeuroglialdamageinMPN-affectedBMstroma.a–d,ExpressionofneuroglialgenesetsinMPNBMNes-GFP1cells.a–c,Selectedtranscriptexpressioninfragmentsperkilobaseofexonpermillionfragmentsmapped(FPKM)ofmesenchymalgenes(a),HSC-nicherelatedgenes(b)andneural-relatedgenes(c).RNA-Seqin

CD452CD312Ter1192GFP1cellsisolatedfromcompoundMPNandcontrolmice6weeksafterpIpCtreatment(samplespooledfrom3mice).d,Genesetenrichmentanalyses(GSEA)ofRNA-Seqdata.e,f,Enrichmentplotofcoordinatedchangesinneuroactiveligand-receptorinteractions(e)andthetranscriptomeofneurons,astrocytesandoligodendrocytes(f).g,h,BMNes-GFP1cellsaredifferentfromsympatheticnervefibresandmatureSchwanncells.Immunostainingoftyrosinehydroxylase(TH)tovisualizesympatheticnervefibres(g)andglialfibrillaryacidicprotein(GFAP)formatureSchwanncells(h)inNes-gfpBM.NotethecloseassociationofNes-GFP1(green)cellstodistinctiveTH1orGFAP1(red)cells.i,BMbiopsiesfrom(leftandmiddle)controlor(right)MPNpatientsimmunostainedwith(left)secondaryAbasnegativecontrolor(middleandright)anti-THantibodies.Nucleiwere

counterstainedwithDAPI(blue).Scalebars,75mm(g),50mm(h),100mm(i).

?2014Macmillan Publishers Limited. All rights reservedRESEARCHLETTER

ExtendedDataFigure5|ContributionofIL-1btoearlyMPNpathogenesis.a,IncreasedBMIL-1blevelsearlyinMPN.MultiplexELISAanalysisofpro-inflammatory(IL-1b,TNF-a,IFN-c,IL-6),regulatory(IL-17)andanti-inflammatory(IL-10)mediatorsinplasma(control,n516;MPN,n513)andBMextracellularfluidsamples(control,n511;MPN,n56;onlyIL-1bandTNF-aweredetectable)fromcontrolandMx1-cre;JAK2(V617F)mice

4–8weeksafterpIpCtreatment.b–f,Nes-gfpmiceweretransplantedwithBMcellsfromMPN(n56)andcontrol(n53)miceandanalysed2weekslater.b–d,mRNAexpressionofIL-1b(b)anditsactivatingenzymecaspase-1(Casp1,c),andBMfrequenciesoflin2sca-11c-kit1haematopoieticprogenitorsandCD11b1Ly-6G(1A8)2monocytes(d).e,f,mRNA

expressionofIL-1receptor(IL1r)(e)anditsantagonist(IL1ra)(f)inBMCD452CD312Ter1192Nes-GFP1/2cells.g,NumberofcirculatingplateletsinWTmicetransplantedwithMPNandcontrolBMcellsandtreatedover18weekswithIL1ra,starting2weeksaftertransplant.h,FrequencyofBMCD452CD312Ter1192CD901cellsinmiceing.i,j,qPCRanalysesofIL-1b(i)andCasp1(j)mRNAexpressioninhaematopoieticprogenitorsand

monocytesisolatedfromtheBMofmiceing.Gapdhwasusedashousekeepinggene(n55).Mean6s.e.m.*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

ExtendedDataFigure6|Compensatorytreatmentwithb3-adrenergicagonistBRL37344blocksMPNprogression.a,Expressionofb2-andb3-adrenergicreceptorsinimmunomagnetically-enrichedCD451andCD452BMcellsfromcontrolandpIpC-inducedMx1-cre;JAK2(V617F)mice(n53).b-c,BloodcountsofWTmice(b)4weeksaftertransplantationwithMPNorcontrolBMcells,before(control,n511;MPN,n513)and(c)4-12weeksafterchronicBRL37344(2mgkg21)orvehicletreatment(i.p.,twiceaday;n55).CD11b1monocytesandgranulocytes,B2201BcellsandCD31Tcellsweredeterminedbyflowcytometry.d,BloodcountsofWTmicetransplantedwithcontrolBMcellsandtreatedwithBRL37344(n56)orvehicle(n54)asdescribedabove.Mean6s.e.m.*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reservedRESEARCHLETTER

ExtendedDataFigure7|b3-adrenergicagonistsorneuroprotectivedrugspreventneuroglialgeneinductioninnestin1cellsandindirectlyaffectmutanthaematopoieticcells.a,InhibitionofmutantHSCexpansionbyb3-adrenergicagonistsorneuroprotectivedrugsisHSC-nichedependent.Immunomagnetically-enrichedCD451haematopoieticcellswereobtainedfromtheBMofNes-gfp;Mx1-cre;JAK2(V617F)miceandcontrollittermates16weeksafterpIpCtreatment(n53).BRL37344(BRL)and4-methylcatechol(4MC)wereaddedinvitroattheindicatedconcentrationsandthefrequencyofcolony-formingunitsinculture(CFU-C)wasscoredafteroneweekinculture.b,Compensatorytreatmentwiththehumanb3-adrenergicagonistMirabegrondelaysMPNinmice.BloodcountsofWTmice8weeksaftertransplantation

withMPNorcontrol(n55)BMcells.Mirabegron(2mgkg21,n58)orvehicle(n57)treatment(i.p.,twiceaday)wasadministeredthelasttwoweeks.c,BMNes-GFP1cellsfromMPNmicetreatedwithBRL37344donotexpressneuroglialgenes;mRNAexpressionoftheSchwanncellmarkersmyelinbasicprotein(Mbp),myelin-associatedoligodendrocytebasicprotein(Mobp)andpleckstrinhomologydomaincontaining,familyB(evectins)member1

(Pleckhb1)inBMCD452CD312Ter1192GFP1cellssortedfromNes-gfpmicetreatedover8weekswithBRL37344orvehicle,starting4weeksaftertransplantationwithMPNorcontrolBMcells(n54).Mean6s.e.m.*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

ExtendedDataFigure8|IncipientMPNsignsareimprovedbyBRL37344or4-methylcatechol.a,b,HaematologicalparametersofWTmice8weeksaftertransplantationwithBMcellsfrompIpC-treatedcontroland

Mx1-cre;JAK2(V617F)mice.Thelatterreceivedtheneuroprotectivedrug4-methylcatechol(10mgkg21,oncedaily),BRL37344(2mgkg21)orvehicleinjections(twiceaday)forthelast4weeks(i.p.;n55).MPNprogressionwasmonitoredinperipheralbloodandmicewerekilledwhentheyshowedonly

incipientsymptoms.a,Bloodcounts,spleensize,weightandnucleatedcellnumber,BMnucleatedcells(limbsandsternum),CD11b1myeloid,B2201B-lymphoidandCD31Tcells.b,BMlin2sca-11c-kit1(LSK)haematopoieticprogenitors,Ter1191CD712matureerythroblastsandCD411/CD421megakaryocyteprogenitors.Mean6s.e.m.*P,0.05,**P,0.01,***P,0.001(unpairedtwo-tailedttest).

?2014Macmillan Publishers Limited. All rights reserved

RESEARCHLETTER

ExtendedDataFigure9|Long-termBRL37344treatmentinhibitsJAK2(V617F)1HSCexpansionbutitdoesnotcompromisenormal

haematopoiesis.a,b,HaematopoieticprogenitorsquantitatedbyFACS(a)orcolony-formingunits(CFU-C)(b)inBMandbloodofWTmice12-16weeksaftertransplantationwithMPN(a)andWTBMcells(b),16weeksafterBRL37344(2mgkg21)orvehicleinjections(twiceaday,i.p.;n53).

Mean6s.e.m.*P,0.05(unpairedtwo-tailedttest).c,BRL37344treatmentdoesnotchangenormalHSCBMnumber.CD45.2WTmiceweretransplantedwithBMcellsfromcongenicCD45.1WTmice,togetherwithlimitingnumbersofBMcellsfromCD45.2controlmicetreatedwithvehicle/BRL37344over4weeks,starting11weeksafterthetransplant(n55).Thefrequencyofmicethatfailedreconstitution16weeksaftertransplantationisplottedagainstthenumberoftestedcells.HSCfrequenciesareindicated;n.s.,non-significant;Pearsonchi-squaredttest.

?2014Macmillan Publishers Limited. All rights reservedLETTERRESEARCH

ExtendedDataFigure10|DifferentiallyexpressedgenesinBM

CD452CD312Ter1192GFP1cellsfromNes-gfp;Mx1-cre;JAK2(V617F)mice.a,b,Topmostupregulated(a)anddownregulated(b)genesinRNaseqanalysesofBMCD452CD312Ter1192GFP1cellsfrom

Nes-gfp;Mx1-cre;JAK2(V617F)mice(red)versuscontrolmice(blue).Plotsrepresentdifferentiallyexpressedgenesinfragmentsperkilobaseofexonpermillionfragmentsmapped(FPKM).

?2014Macmillan Publishers Limited. All rights reserved

本文来源：https://www.bwwdw.com/article/s3sg.html