1、See discussions,stats,and author profiles for this publication at:https:/ PTEN loss by exosomal microRNA primes brainmetastasis outgrowthArticleinNature October 2015DOI:10.1038/nature15376CITATIONS327READS54822 authors,including:Some of the authors of this publication are also working on these relat
2、ed projects:Cloud manufacturing View projectMicrofluidic and imaging View projectLin ZhangBeihang University(BUAA)200 PUBLICATIONS4,315 CITATIONSSEE PROFILESiyuan ZhangUniversity of Notre Dame46 PUBLICATIONS2,490 CITATIONSSEE PROFILEJun YaoUniversity of Texas MD Anderson Cancer Center84 PUBLICATIONS
3、11,746 CITATIONSSEE PROFILEwen-chien HuangMackay Memorial Hospital14 PUBLICATIONS937 CITATIONSSEE PROFILEAll content following this page was uploaded by Jodi M.Saunus on 21 January 2016.The user has requested enhancement of the downloaded file.LETTERdoi:10.1038/nature15376Microenvironment-induced PT
4、EN loss by exosomalmicroRNA primes brain metastasis outgrowthLinZhang1,2*,SiyuanZhang1,3*,JunYao1,FrankJ.Lowery1,2,QinglingZhang1,Wen-Chien Huang1,PingLi1,MinLi1,XiaoWang1,Chenyu Zhang1,Hai Wang1,Kenneth Ellis1,Mujeeburahiman Cheerathodi4,Joseph H.McCarty4,Diane Palmieri5,Jodi Saunus6,Sunil Lakhani6
5、,7,8,Suyun Huang4,Aysegul A.Sahin9,Kenneth D.Aldape9,Patricia S.Steeg5&Dihua Yu1,2,10The development of life-threatening cancer metastases at distantorgans requires disseminated tumour cells adaptation to,andco-evolution with,the drastically different microenvironments ofmetastatic sites1.Cancer cel
6、ls of common origin manifest distinctgene expression patterns after metastasizing to different organs2.Clearly,the dynamic interaction between metastatic tumour cellsand extrinsic signals at individual metastatic organ sites criticallyeffects the subsequent metastatic outgrowth3,4.Yet,it is unclearw
7、hen and how disseminated tumour cells acquire the essentialtraits from the microenvironment of metastatic organs that primetheir subsequent outgrowth.Here we show that both human andmouse tumour cells with normal expression of PTEN,an import-ant tumour suppressor,lose PTEN expression after dissemina
8、tiontothebrain,butnot toother organs.ThePTENlevelinPTEN-lossbrain metastatic tumour cells is restored after leaving the brainmicroenvironment.This brain microenvironment-dependent,reversible PTEN messenger RNA and protein downregulation isepigenetically regulated by microRNAs from brain astrocytes.M
9、echanistically,astrocyte-derived exosomes mediate an inter-cellular transfer of PTEN-targeting microRNAs to metastatictumour cells,while astrocyte-specific depletion of PTEN-targetingmicroRNAsorblockadeofastrocyteexosomesecretion rescuesthePTEN loss and suppresses brain metastasisin vivo.Furthermore
10、,thisadaptivePTENlossinbrainmetastatictumourcellsleadstoanincreased secretion of the chemokine CCL2,which recruits IBA1-expressing myeloid cells that reciprocally enhance the outgrowthof brain metastatic tumour cells via enhanced proliferation andreduced apoptosis.Our findings demonstrate a remarkab
11、le plas-ticity of PTEN expression in metastatic tumour cells in responseto different organ microenvironments,underpinning an essentialrole of co-evolution between the metastatic cells and their micro-environment during the adaptive metastatic outgrowth.Our find-ingssignify thedynamic and reciprocal
12、cross-talk between tumourcells and the metastatic niche;importantly,they provide newopportunities for effective anti-metastasis therapies,especially ofconsequence for brain metastasis patients.The remarkable phenotypic plasticity observed in metastasis isindicative of co-evolution occurring at speci
13、fic metastatic organmicroenvironments5,6.To obtain insights into how disseminatedtumour cells acquire essential traits from metastatic microenviron-ments for successful outgrowth,we analysed public gene expressionprofiles of clinical metastases from distinct organs as well as organ-specific metastas
14、es from mice injected with various cancer cells(Extended Data Fig.1ac).Notably,PTEN mRNA was markedlydownregulated in brain metastases compared to primary tumours orotherorganmetastases.Ourimmunohistochemistry(IHC)analysesofPTEN expression confirmed a significantly higher rate of PTENloss(defined by
15、 an immunoreactive score(IRS)of 03)7in brainmetastases(71%)than in unmatched primary breast cancers(30%)(Fig.1a).PTEN loss was also detected at a significantly higher fre-quency in brain metastases(71%)than in matched primary breastcancers(37%)of an independent patient cohort(Fig.1b).To test a possi
16、ble role for PTEN loss in brain metastasis8,9,weintracarotidly injected PTEN-knockdown tumour cells and assessedexperimental brain metastasis;unexpectedly,neither incidence norsize of brain metastases was increased(Fig.1c).Furthermore,patientswith PTEN-normal or PTEN-loss primary tumours had compara
17、blelevels of brain-metastasis-free survival,and patients with or withoutbrain metastases had similar PTEN levels in their primary tumours(Extended Data Fig.1d,e).Thus,the observed PTEN loss in brainmetastases was unlikely to be derived from PTEN-low primarytumours.To investigate whether PTEN loss in
18、 brain metastasis is asecondary non-geneticeventimposed bythe brainmicroenvironment,we injected five PTEN-normal breast cancer cell lines either intomammary fat pad(MFP)or intracarotidly to induce brain metastasis.Notably,thePTENlevelwassignificantlydecreasedinbrainmetastasescomparedtotherespectiveM
19、FPtumoursorlungmetastases(ExtendedDataFig.2a,b).Werepeatedtheinjectionswithcellsclonallyexpandedfrom single PTEN-normal tumour cells,and observed similar pheno-types(Fig.1d),suggesting that PTEN-loss brain metastases were notselected from pre-existing PTEN-low cells in the primary tumours.Surprising
20、ly,established sublines from PTEN-low brain metastases(primary Br cells)regained PTEN expression in culture comparableto parental cells(Fig.1e).Analogously,two in-vivo-selected brain-seeking sublines exhibited similar PTEN levels to their matched par-ental cells in vitro(Extended Data Fig.2c).Re-inj
21、ecting the culturedPTEN-normalprimarybrainsublinesconferredadistinctPTENlossinsecondary brain metastases,but not in secondary MFP tumours,andPTENlevelsinsecondarybrainsublinecellswerefullyrestoredagaininculture(Fig.1f,g and Extended Data Fig.2d),indicating a reversiblenon-geneticPTENlossinthebraintu
22、mourmicroenvironment(TME).ToexplorehowthebrainTMEregulatesPTENinmetastaticcells1012,we co-cultured tumour cells with primary glia(.90%astrocytes)13,cancer-associatedfibroblasts(CAFs),orNIH3T3fibroblasts.Co-culturewith glia led to a significant decrease of PTEN mRNA and PTENprotein(Fig.2a,b and Exten
23、ded Data Fig.2e,f)in all tumour cells,but did not affect PTEN promoter methylation or activity(ExtendedData Fig.2g,h).This prompted us to examine whether glia reducePTENmRNAstabilitythroughmicroRNAs(miRNAs).FivemiRNAs(miR-17,miR-19a,miR-19b,miR-20aandmiR-92)inthemiR-1792cluster were functionally dem
24、onstrated to target PTEN(refs 1417),and Mirc1tm1.1Tyj/J mice have a floxed miR-1792 allele18.We*These authors contributed equally to this work.1Department of Molecular and Cellular Oncology,The University of Texas M.D.Anderson Cancer Center,Houston,Texas 77030,USA.2Cancer Biology Program,Graduate Sc
25、hool of Biomedical Sciences,Houston,Texas77030,USA.3DepartmentofBiologicalSciences,UniversityofNotreDame,NotreDame,Indiana46556,USA.4DepartmentofNeurosurgery,TheUniversityofTexasM.D.AndersonCancerCenter,Houston,Texas 77030,USA.5Womans Malignancies Branch,National Cancer Institute,Bethesda,Maryland 2
26、0892,USA.6The University of Queensland Centre for Clinical Research,Brisbane,Queensland4029,Australia.7TheSchoolofMedicineandPathologyQueensland,Brisbane,Queensland4029,Australia.8TheRoyalBrisbaneandWomensHospital,Brisbane,Queensland4029,Australia.9Department of Pathology,The University of Texas M.D
27、.Anderson Cancer Center,Houston,Texas 77030,USA.10Center for Molecular Medicine,China Medical University,Taichung 40402,Taiwan.1 0 0|N A T U R E|V O L 5 2 7|5 N O V E M B E R2 0 1 5G2015 Macmillan Publishers Limited.All rights reservedknocked out the miR-1792 allele in situ in Mirc1tm1.1Tyj/J miceby
28、 intracranial injection of astrocyte-specific Cre adenovirus(Ad-GFAP-Cre),then intracarotidly injected syngeneic mouse melanomaB16BL6 cells to form brain metastases(Fig.2c).Astrocyte-specificdepletionofPTEN-targetingmiRNAsblockedPTENdownregulation(Fig.2d)in the brain metastasis tumour cells in vivo
29、without signifi-cantlyalteringotherpotentialmiRNAtargets(ExtendedDataFig.3a),andsignificantlysuppressedbrainmetastasisgrowthcomparedtothePTEN IHC IRS scorePercentage020406080Primary(n=139)Brain mets(n=131)034679 1012Chi-square P 0.001eabcCell line with heterogeneousPTEN levelsClonal expansionPTEN no
30、rmal cells from single clone50 m50 m50 m50 m25 m25 mBrainMetsCell pelletMFP tumourBrain metsHCC1954 clone 44T1 clone 7ICA injection for brain metsMFP injection forprimary tumourBrainMetsCell pelletMFP tumourBrain mets50 m50 m50 m50 m25 m25 mIn vivo tumourIn vitro tumour-derived cell lineP 0.001 1 br
31、ain cell Brain mets2 brain mets 2 MFP tumour 2 brain cell 2 MFP cell ICAMFP0.00.51.01.5PTEN mRNA(fold)fBrain metastasesPrimary breast tumourPTEN IHC200 m200 m20 m20 m2 MFPtumour2 brainmetsBrainBrainMets50 m50 m50 m50 mHCC1954BrMacrometsMicrometsMacrometsMicrometsMacrometsMicromets0501001502002503003
32、50Metastase countsControl shRNAPTEN shRNA_1PTEN shRNA_2Chi-squareP=0.1253PTEN-actinControl shRNAPTEN shRNA_2PTEN shRNA_1MDA-MB-231Br1.00 0.40 0.45PPTEN-actinBr P Br P BrMDA 2314T1HCC1954BT474P BrBrainmets1 brain cell In vitro cultured cellsPTEN mRNA(fold)00.51.01.5P BrP BrP BrP BrMDA 2314T1HCC1954BT
33、4741.0 1.2 1.0 1.11.01.11.00.9dgPrimary(n=35)Brain mets(n=35)PTEN IHC IRS scorePercentage0346791012Chi-square P=0.0211806040200 2 brain mets2 MFP tumour 2 brain cell2 MFP cellFigure 1|Brain microenvironment-dependent reversible PTEN down-regulationinbrainmetastases.a,RepresentativeIHCstainingandhist
34、ogramsof PTEN protein levels in primary breast tumours(n5139)and unmatchedbrain metastases(mets)(n5131)(Chi-square test,P,0.001).b,Histogramsof PTEN protein levels in primary breast tumours and matched brainmetastases from 35 patients(Chi-square test,P50.0211).c,PTEN westernblots(left)and brain meta
35、stasis counts 30 days after intracarotid injection(right)of MDA-MB-231Br cells transfected with control or PTEN shRNAs.Macromets:.50mm in diameter;micromets:#50mm(mean6s.e.m.,Chi-square test,P50.1253).d,PTEN IHC staining of tumours derived fromclonally expanded PTEN-normal sublines.ICA,intracarotid
36、artery;MFP,mammary fat pad.e,Western blot and quantitative reverse transcriptase PCR(qRTPCR)of PTEN expression in the indicated parental(P)and brain-seeking(Br)cells under culture(3 biological replicates,with 3 technicalreplicateseach).f,Schematicofinvivore-establishmentofsecondary(2u)brainmetastasi
37、s,MFP tumour,and their derived cell lines.g,PTEN qRTPCR(mean6s.e.m.,t-test,3 biological replicates,with 3 technical replicates each)and PTEN IHC in HCC1954Br secondary tumours and cultured cells.abd1.Intracranial injection:Ad-GLuc or Ad-GFAP-Cre Mirc1tm1.1Tyj/J mouse(floxed-miR-1792)2.Left ICA injec
38、tion of B16BL6 melanoma cells for brain metsAd-GLucAd-GFAP-Cre50 m50 mTumourTumour50 m50 mBrainBrainTumourTumourBrainBrain0.2 cm0.2 cm0.2 cm0.2 cmPTENIHCBrightfieldPrimary CAFVim10 mPrimary gliaGFAP10 m0.00.51.01.5Medium control+CAF+Glia+NIH3T3P=0.015 P 0.001P 0.001HCC1954MDA-MB-231B16BL6Relative PT
39、EN mRNA(fold change)P 0.0014T1PTEN+CAF+Glia-actinPTEN-actin-actin+NIH3T3ControlHCC1954 MDA-MB-231B16BL610.763 0.157 1.14010.918 0.533 0.89910.930 0.236 0.849PTEN-actin4T1cfePTEN 3-UTR(3,303 bp)miR-17miR-19a_1miR-19a_2miR-21miR-26Normalized luciferaseP 0.050.00.51.01.5Wild typemut-17mut-19a_1mut19a_2
40、mut-21mut-26+Astrocyte+CAFg0.00.51.01.52.0MDA-MB-231 co-cultured with astrocytes2.5P 0.001P 0.001Gene expression in tumour cells(fold)PTENmiR-19aPTENGAPDHMDA-MB-231 co-cultured with astrocytesAd-CreAd-CreAd-GLucAd-GLuc1.002.40hAd-GFAP-CreAd-GLuc0204060Tumour volume(mm3)P=0.0024PTEN10.746 0.447 0.757
41、Figure 2|Astrocyte-derived miRNAs silence PTEN in tumour cells.a,PTEN mRNA in the indicated tumour cells after 25 days co-culture withGFAP-positive primary glia or vimentin(vim)-positive primary CAFs orNIH3T3 fibroblasts(mean6s.e.m.,t-test,3 biological replicates,with 3technicalreplicateseach).b,Wes
42、ternblotofPTENproteinunderco-cultureasin a(3 biological replicates).c,Schematic of astrocyte-specific miR-1792deletion by GFAP-driven Cre adenovirus(Ad-GFAP-Cre)in Mirc1tm1.1Tyj/Jmice.d,Representative image of tumour sizes and PTEN IHC of brainmetastases.e,Quantification of brain metastases volume(m
43、ean6s.d.,t-test,P50.0024).f,PTEN39-UTRluciferaseactivityafterco-culture(mean6s.e.m.,t-test,3 biological replicates,with 3 technical replicates each).bp,base pairs.g,qRTPCR analyses of miR-19a and PTEN mRNA in MDA-MB-231 cellsafter 48h co-culture with primary astrocytes from Mirc1tm1.1Tyj/J mice pre-
44、infected(48h)by adenovirus(Ad-bGLuc or Ad-GFP-Cre)(mean6s.e.m.,t-test,P,0.001,3 biological replicates,with 3 technical replicates each).h,Western blot of PTEN protein in MDA-MB-231 cells,co-cultured as in g.5 N O V E M B E R 2 0 1 5|V O L5 2 7|N A T U R E|1 0 1LETTER RESEARCHG2015 Macmillan Publishe
45、rs Limited.All rights reservedcontrol group(Fig.2d,e),indicating a tumour cell non-autonomousPTEN downregulation by astrocyte-derived PTEN-targeting miRNAs.Astrocyte-specific depletion of PTEN-targeting miRNAs also sup-pressed intracranially injected tumour cell outgrowth(ExtendedData Fig.3bf).To ex
46、amine which PTEN-targeting miRNA primarilymediates the PTEN loss in tumour cells when co-cultured with astro-cytes,the luciferase activities of the wild-type and mutated PTEN39-untranslatedregion(UTR)(containingvariousmiRNAbindingsitemutations)intumourcellswereassessed(Fig.2f).ComparedwithCAFco-cult
47、ure,astrocyteco-cultureinhibitedluciferaseactivityofwild-typePTEN 39-UTR,which was rescued by the miR-19a binding site muta-tion(position 1),but not by other mutations,indicating the major roleof miR-19a in astrocyte-mediated PTEN mRNA downregulation intumour cells.Furthermore,PTEN mRNA(Fig.2g and E
48、xtendedData Fig.3g)and PTEN protein(Fig.2h and Extended Data Fig.3h)were not downregulated in tumour cells co-cultured with primaryastrocytes from Mirc1tm1.1Tyj/J mice in which PTEN-targetingmiRNAs were depleted(Extended Data Fig.3i).After co-culture with Cy3-labelled miR-19a-transfected primaryastr
49、ocytes,wedetectedsignificantlymoreCy31epithelialcelladhesionmolecule(EpCAM)-positive tumour cells over time than under CAFco-culture(Fig.3aandExtendedDataFig.4a),suggestingthatmiR-19ais intercellularly transferred from astrocytes to tumour cells.miRNAsare transferable between neighbouring cells thro
50、ugh gap junctions orsmall vesicles19,20.Treating tumour cells with a gap junction channelinhibitor,carbenoxolone disodium salt,had no significant effect onmiR-19a intercellular transfer(data not shown),while adding astro-cyte-conditioned media to tumour cells led to an increase in miR-19alevels and