Differentiation of mammary tumors and reduction in ...genesdev.cshlp.org/content/30/1/34.full.pdfGayatri Arun,1 Sarah Diermeier,1 Martin Akerman,1 Kung-Chi Chang,1,2 J. Erby Wilkinson,3

Differentiation of mammary tumorsand reduction in metastasis upon Malat1lncRNA lossGayatri Arun1 Sarah Diermeier1 Martin Akerman1 Kung-Chi Chang12 J Erby Wilkinson3

Stephen Hearn1 Youngsoo Kim4 A Robert MacLeod4 Adrian R Krainer12 Larry Norton5 Edi Brogi5

Mikala Egeblad1 and David L Spector12

1Cold Spring Harbor Laboratory Cold Spring Harbor New York 11724 USA 2Molecular and Cellular Biology Program StonyBrook University Stony Brook New York 11790 USA 3Department of Pathology University of Michigan Ann Arbor Michigan48109 USA 4Ionis Pharmaceuticals Inc Carlsbad California 92010 USA 5Memorial Sloan Kettering Cancer Center New YorkNew York 10065 USA

Genome-wide analyses have identified thousands of long noncoding RNAs (lncRNAs) Malat1 (metastasis-associated lung adenocarcinoma transcript 1) is among the most abundant lncRNAs whose expression is alteredin numerous cancers Here we report that genetic loss or systemic knockdown of Malat1 using antisenseoligonucleotides (ASOs) in the MMTV (mouse mammary tumor virus)-PyMT mouse mammary carcinomamodel results in slower tumor growth accompanied by significant differentiation into cystic tumors and a reductionin metastasis Furthermore Malat1 loss results in a reduction of branching morphogenesis in MMTV-PyMT-and Her2neu-amplified tumor organoids increased cell adhesion and loss of migration At the molecular levelMalat1 knockdown results in alterations in gene expression and changes in splicing patterns of genesinvolved in differentiation and protumorigenic signaling pathways Together these data demonstrate forthe first time a functional role of Malat1 in regulating critical processes in mammary cancer pathogenesis ThusMalat1 represents an exciting therapeutic target andMalat1ASOs represent a potential therapy for inhibiting breastcancer progression

[Keywords antisense therapy breast cancer Malat1 metastasis noncoding RNA tumor differentiation tumororganoids]

Supplemental material is available for this article

Received August 26 2015 revised version accepted November 24 2015

The majority of the mammalian genome encodes for non-coding RNAs (ncRNAs) (Derrien et al 2012 Djebali et al2012 The ENCODE Project Consortium 2012) LongncRNAs (lncRNAs) represent a class of ncRNAs that aregt200 nucleotides (nt) and have been shown to participatein diverse cellular functions (for review see Wilusz et al2009 Rinn and Chang 2012 Bergmann and Spector2014) Over 16000 lncRNAs have been annotated in thehuman genome and thus far 8000 have been annotatedin the mouse genome (Harrow et al 2012) SeverallncRNAs have been implicated in various types of cancer(for review see Costa 2007 Prasanth and Spector 2007Prensner and Chinnaiyan 2011) For example HOTAIRis overexpressed in metastatic breast tumors and inducesgenome-wide repositioning of gene silencing complexes(Gupta et al 2010) Multiple other lncRNAs such asDD3PCA3 PCAT-1 and SChLAP1 are overexpressed in

prostate tumors (Bussemakers et al 1999 Prensner et al2011 2013) while the lncRNA PVT1 was shown to regu-late Myc protein levels in tumors with a 8q24 gain (Tsenget al 2014)

MALAT1 (metastasis-associated lung adenocarcinomatranscript 1) is a highly conserved lncRNA that was firstidentified as being up-regulated in lung tumors that hada higher propensity to metastasize (Ji et al 2003) andwas subsequently shown to be up-regulated in a broadspectrum of tumor types (Yamada et al 2006 Lin et al2007 Guffanti et al 2009) MALAT1 is among the mosthighly abundant lncRNAs and interestingly its primarytranscript is processed (Wilusz et al 2008) into a long 67-kb transcript which localizes to nuclear speckles (Hutch-inson et al 2007 Clemson et al 2009 Bernard et al 2010)

Corresponding author spectorcshleduArticle published online ahead of print Article and publication date areonline at httpwwwgenesdevorgcgidoi101101gad270959115

copy 2016 Arun et al This article is distributed exclusively by Cold SpringHarbor Laboratory Press for the first six months after the full-issue publi-cation date (see httpgenesdevcshlporgsitemisctermsxhtml) Aftersix months it is available under a Creative Commons License (At-tribution-NonCommercial 40 International) as described at httpcreativecommonsorglicensesby-nc40

34 GENES amp DEVELOPMENT 3034ndash51 Published by Cold Spring Harbor Laboratory Press ISSN 0890-936916 wwwgenesdevorg

Cold Spring Harbor Laboratory Press on December 9 2020 - Published by genesdevcshlporgDownloaded from

and a 61-nt tRNA-like small RNA which localizes to thecytoplasm (Wilusz et al 2008) MALAT1 has been impli-cated in regulating alternative pre-mRNA splicing andits knockdown was shown to result in cell cycle arrest(Tripathi et al 2010) MALAT1 has also been shown tobe required to activate E2F target genes by repositioningthem from polycomb bodies to transcriptionally activenuclear sites in a serum-dependent manner (Yang et al2011) More recently two genome-wide studies have indi-cated that MALAT1 binds to the transcription start sites(TSSs) and gene bodies of actively transcribing genes to-gether with another lncRNA NEAT1 (West et al 2014)In addition a genome-wide RNAndashRNA binding studyhas shown that Malat1 can also bind to nascent pre-mRNAs indirectly via protein partners (Engreitz et al2014) Despite these functional insights derived from var-ious cell lines Malat1 knockout mice developed in ourlaboratory and others (Eissmann et al 2012 Nakagawaet al 2012 Zhang et al 2012) exhibit a normal phenotypeThe mice are born in expected Mendelian ratios and nomajor alterations have been observed in tissue histologygene expression or pre-mRNA splicingAlthough Malat1 knockout mice have no overt pheno-

type MALAT1 loss in a lung cancer xenograft model re-sulted in reduced homing of human lung cancer cells tothe lungs compared with control cells with wild-type lev-els ofMALAT1 (Gutschner et al 2013) and knockdown ofMALAT1 in lung cancer cells in vitro impaired cellularmotility (Tano et al 2010 Gutschner et al 2013) In addi-tion MALAT1 gene mutations have been found to fre-quently occur in luminal-type breast tumors (Ellis et al2012) Based on these findings we were interested in pur-suing whether Malat1 has a direct causative in vivo func-tion in breast cancer progressionHere we examined the role of Malat1 in mammary tu-

mor progression using the MMTV (mouse mammary tu-mor virus)-PyMT mouse mammary tumor model ofhuman luminal B breast cancer (Guy et al 1992 Linet al 2003 Herschkowitz et al 2007) We found thatgenetic knockout of Malat1 or knockdown upon subcu-taneous delivery of antisense oligonucleotides (ASOs)resulted in significant differentiation of the primary tu-mors and a reduction in lung metastases In additionknockdown ofMalat1 inMMTV-PyMT andHer2neu-am-plified mammary tumor organoids inhibited branchingmorphogenesis ex vivo Detailed molecular analyses byRNA sequencing (RNA-seq) revealed significant changesin gene expression and pre-mRNA splicing of protumori-genic and differentiation-related genes These findingsidentify Malat1 as a therapeutic target in breast cancerand the utility of Malat1 ASOs as a therapeutic for inhib-iting breast cancer progression

Results

Mammary tumor progression is impaired upongenetic loss of Malat1 lncRNA

Human luminal B breast cancer is one of the most preva-lent subtypes of breast cancer affecting nearly 15ndash20

of individuals diagnosed with this disease (for review seeSchnitt 2010 Ades et al 2014) In order to investigate thepotential in vivo role of Malat1 in mammary tumor pro-gression we used theMMTV-PyMTmouse mammary tu-mor model of human luminal B breast cancer (Guy et al1992 Lin et al 2003 Herschkowitz et al 2007) In thismodel the polyomamiddle-T antigen is under the controlof the MMTV promoter resulting in highly penetrantmammary tumors and lung metastasis (Lin et al 2003)Malat1 homozygous knockout (Malat1minusminus) femalemice

previously generated in our laboratory (Zhang et al 2012)were crossed to MMTV-PyMT males (C57Bl6) to obtainMMTV-PyMTMalat1++ PyMTMalat1+minus and PyMTMalat1minusminus mice Individuals of all three genotypes wereborn in expectedMendelian ratioswith a normal gestationperiod and postnatal development In accord with pub-lished data (Lin et al 2003) tumor onset in MMTV-PyMTmice began soon after puberty and palpable tumorsdeveloped within 12ndash16 wk of birth in gt90 of the miceTumor onset did not differ significantly in the presenceor absence ofMalat1 (Fig 1A) Prior to 5 mo of age nearly90of themice developed palpablemultifocal mammarygland lesions demonstrating that loss of Malat1 doesnot affect the initiation of PyMT-induced mammarygland tumors Although the tumors that developed inthe Malat1minusminus background initially exhibited a relativelyslower tumor growth rate compared with control mice(Fig 1B) this difference was not found to be significantespecially at the later time points Surprisingly theMMTV-PyMTMalat1minusminusmice developedhighly cystic liq-uid-filled tumors (Fig 1C right panel) in contrast to thecontrol mice which developed poorly differentiated solidcarcinomas (Fig 1C left panel) Hematoxylin and eosin(HampE) staining of tumor sections from MMTV-PyMTMalat1minusminus mice demonstrated that nearly 90 of the tu-mors thatweregt1cmindiameterexhibited acysticpheno-type with gt75 of the tumor area showing encapsulatedwell-differentiated cystic histopathology (Fig 1D bottompanels) in stark contrast to the poorly differentiated solidcarcinomas observed in the control tumors (Fig 1D toppanels)In the MMTV-PyMT model nearly 80 of the mice

develop lung micro- and macrometastases Thereforewe examined the impact of Malat1 loss on lung metasta-sesMMTV-PyMTMalat1minusminus mice exhibited a significantreduction in lung macrometastases where only three of12 mice had macro lesions as compared with seven ofeight mice in control MMTV-PyMTMalat1++ mice(Fig 1EF) The lungs were further examined for the inci-dence of micrometastases (Fig 1G) and the majority ofMMTV-PyMTMalat1minusminus animals did not showanymicro-metastasis (nine of 12) Next we calculated the total met-astatic burden as a percentage of lung volume (Fig 1H) andfound that the MMTV-PyMTMalat1minusminus animals showeda significantly decreased metastatic burden (3ndash5)as compared with the control MMTV-PyMTMalat1++

mice for which the metastatic burden reached up to40 of the total lung volume These results clearly dem-onstrate a role for the lncRNAMalat1 inmammary tumorprogression

Differentiation of mammary tumors upon Malat1 loss

GENES amp DEVELOPMENT 35


MALAT1 is up-regulated in human breastcancer metastases

To understand the significance of MALAT1 in humanbreast cancer we evaluated its expression in human

breast tumors by RNA-FISH on matched primary tumorandmetastases tissue sections from the same individualsMALAT1 was generally expressed at a low level in theprimary lesions (Fig 2A top panels) of breast tumors irre-spective of the ERPR status However very high levels of

Figure 1 Impaired tumor progression and metastasis in MMTV-PyMT mice lacking Malat1 (A) Kaplan-Meier curves for tumor-freesurvival in MMTV-PyMTMalat1++ (n = 18) and MMTV-PyMTMalat1minusminus (n = 24) mice (B) Total tumor burden in MMTV-PyMTMalat1++ (n = 8) andMMTV-PyMTMalat1minusminus (n = 12) mice from the detection of tumors (sim6 mm) over a period of 9 wk Error bars repre-sent SEM (C ) Surgically removed tumors fromMMTV-PyMTMalat1++ (left panel) andMMTV-PyMTMalat1minusminus (right panel) mice Bar 2mm (D) HampE-stained sections of primary tumors fromMMTV-PyMTMalat1++ andMMTV-PyMTMalat1minusminusmice The entire tumor sec-tion and a higher magnification of a small region are shown from both genotypes Bars 2 mm and 200 microm for the differentmagnificationsrespectively (E) Whole-lung images showing lung metastatic nodules in MMTV-PyMTMalat1++ (left panel) compared with MMTV-PyMTMalat1minusminus (right panel) Bar 5 mm (F ) Quantitation of number of macrometastatic nodules per lungs (total nodules2) Errorbars represent standard deviation (SD) (lowastlowastlowast) P lt 0001 by Wilcoxon signed rank test (G) HampE-stained lung sections showing regions ofmicrometastasis inMMTV-PyMTMalat1++ lungs (left panel) andMMTV-PyMTMalat1minusminus lungs (right panel) Bar 2 mm (H) Percentageof lungmetastatic burden over total lung area ofMMTV-PyMTMalat1++mice (n = 7) andMMTV-PyMTMalat1minusminus mice (n = 4) Error barsrepresent SD (lowastlowastlowast) P lt 0001 by Wilcoxon signed rank test

Arun et al

36 GENES amp DEVELOPMENT


MALAT1 were observed in the lung metastases from thesame individuals (Fig 2A bottom panels) Interestinglyhigher levels of MALAT1 expression were confined onlyto the tumor cells while the adjacent stromal cellsshowed a veryweak signal forMALAT1 in both the prima-ry tumor and the metastatic lesions (Supplemental Fig

S1A) Some primary tumors showed elevated MALAT1expression in only a small fraction of the cancer cells orcell clusters while neighboring cancer cells remainedvery low in MALAT1 expression suggesting that thesehighMALAT1-expressing cancer cells may escape the pri-mary tumors preferentially to establish metastasis or that

Figure 2 Efficient knockdown ofMalat1 in tumors using ASOs (A) RNA-FISH using aMALAT1-specific probe on matched human pri-mary tumor and lung metastases from breast cancer patients Bar 200 microm Magnification scale is 15times using Aperio Scanscope (B) Quan-titation ofMALAT1 FISH signal in primary tumor and lung metastasis tissue array samples from breast cancer patients (C ) Schematic ofthe ASO treatment protocol in MMTV-PyMT mice (D) RT-qPCR of Malat1 knockdown in tumors of mice treated with scrambled ASOcompared with Malat1 ASO1 or ASO2 Relative fold change calculated based on the geometric mean of Gapdh β-actin and Hprt RNAlevels Error bars represent SD (lowastlowast) P lt 001 by Studentrsquos t-test (E) RNA-FISH of tumor sections using aMalat1-specific probe Bar 100 micromMagnification scale is 20times using Aperio Scanscope (F ) Normalized tumor growth curve ofMMTV-PyMTmice treatedwith control scram-bledASO (ScASO) (n = 12)Malat1ASO1 (n = 11) orMalat1ASO2 (n = 7) Error bars represent SEM (lowast) P lt 005 (lowastlowast) P lt 001 byANOVA (G)Percentage of Ki-67-positive cells in ScASO-treated tumors orMalat1ASO1- or ASO2-treated tumors n = 3 tumors from each group withat least two sections from each tumor Error bars represent SD (lowast) P lt 005 (lowastlowast) P lt 001 by Studentrsquos t-test




potentially microenvironmental cues may influenceMALAT1 expression levels (Supplemental Fig S1BC)To determine the extent of occurrence of high MALAT1expression in human samples we obtained a breast cancertissue array and lungmetastasis tissue array and evaluatedMALAT1 expression level using RNA-FISH MALAT1transcript levels were increased between 15-fold andthreefold in gt60 of the lung metastases compared withthe primary tumor sections (Fig 2B) The lung metastasisarray was generated from primary breast cancer patientsWhile only some of the primary tumor samples andlung metastases were matched we could detect highMALAT1 expression in the majority of the lung metasta-sis samples confirming our matched sample data HighMALAT1 expression was also observed in sections of hu-man breast cancer brainmetastatic lesions (SupplementalFig S1D) Interestingly the MMTV-PyMT tumors alsodisplayed a similar expression pattern for Malat1 wherehigherMalat1 expression was confined to a small fractionof cells (Supplemental Fig S2AndashE) Furthermore morehigher Malat1-expressing cells were found in advancedcarcinomas comparedwith adenomas or hyperplasia (Sup-plemental Fig S2AndashE) This observation suggests thatMALAT1 likely plays a crucial role in human breast can-cer pathogenesis prompting us to further evaluate its po-tential as a therapeutic target

Knockdown of Malat1 results in mammarytumor differentiation

Next we were interested in determining whether wecould recapitulate the phenotype observed by geneticknockout upon systemically knocking down Malat1levels in established tumors in a more therapeutic sett-ing Since Malat1 is a nuclear retained lncRNA we usedASOs to achieve knockdown Malat1-specific ASOs(16mers) were synthesized using a next-generation designcomprised of phosphorothioate-modified short S-cEt(S-2prime-O-Et-2prime4prime-bridged nucleic acid) gapmer chemistry(Teplova et al 1999) We used two independent Malat1-specific ASOs targeting two different regions of theMalat1 transcript as well as a control scrambled ASO(ScASO)

Three-month-old to 4-mo-old mice were divided intothree cohorts (seven to 12 mice each) and randomizedbased on their age tumor size and number of tumorsEach mouse in the cohort received either ScASO orMalat1-specific ASO1 or ASO2 via subcutaneous injec-tions of 25 mgkg per day (125 mgkg per week) for 5 dwith a rest period of 2 d (Fig 2C) The injections were car-ried out for a period of 8 wk upon which at least one tu-mor from most of the control mice reached 2 cm in sizeDuring the course of the treatment tumors were mea-sured twice per week At the end of the treatment periodthe animals were euthanized and the primary tumors andlungs were collected An average Malat1 knockdown ofsim60 was achieved in the mammary tumor tissue inmice treated with either ASO1 or ASO2 as comparedwith ScASO-treated control mice (Fig 2D) Knockdownefficiencywas also confirmed byRNA-FISH on tumor sec-

tions (Fig 2E) Different regions of the treated tumorsshowed varying knockdown efficiencies ranging from20 to 80 knockdown (data not shown)

Interestingly when tumor growth was monitored overthe course of treatment the Malat1 ASO-treated groupshowed a 50 slower tumor growth rate than tumorstreated with ScASO (Fig 2F) The decrease in growthratewas observed as early asweek 3 into the treatment pe-riod and by the end of the treatment the tumor volumewas reduced by 50 in Malat1 ASO-treated mice com-pared with the ScASO-treated group Consistent withthe slower tumor growth rate Ki-67 labeling of the tumorsections showed a marked decrease in Ki-67-positive nu-clei The proliferative index of Malat1 ASO1- or ASO2-treated tumors calculated as a measure of Ki-67 positivi-ty was found to be 22 and 26 respectively comparedwith 51 in the control tumors treated with ScASO (Fig2G) Although this was in contrast to the knockout exper-iment where there was no significant difference in prima-ry tumor growth rate we reason that this was due to thefact that the Malat1 knockout tumors were extensivelycystic and when we measured the tumor volume it islikely that the cyst growth contributed significantly tothe total volume rather than actual tumor mass

The primary tumors from each group were analyzedfor histopathological changes Notably the Malat1 ASO-treated tumors were cystic liquid-filled encapsulatedwell-differentiated primary tumors closely resemblingthose from theMMTV-PyMTMalat1minusminus animals whereasthe tumors in the ScASO-treated control animals re-mained poorly differentiated as in tumors from MMTV-PyMTMalat1++mice (Fig 3A)Also similar to theknock-out phenotype a single layer of epithelial cells linedthe enlarged ducts in the Malat1 ASO-treated miceThere was a significant increase in cysticductular tu-mors (31 in ASO1-treated and 22 in ASO2-treated)(Fig 3B) whereas only 4 of the ScASO-treated micedeveloped cystic tumors similar in morphology to theMalat1 ASO-treated groups Furthermore there was atwofold decrease in the number of solid carcinomas char-acterized by poorly differentiated cells in Malat1 ASO-treated mice Quantitation of the cystic area of Malat1ASO-treated tumors showed that on average 40 of thetotal tumor area contained cysticductular histopathology(Fig 3C)

Given that Malat1 ASO-treated tumors display aremarkable histological differentiation phenotype we as-sayed for differentiation markers by immunofluorescenceon sections of primary tumors E-cadherin amajor cell ad-hesion protein present on the cell membrane of differenti-ated epithelial cells exhibited increased labeling on thecell membrane of Malat1 ASO-treated tumors (Fig 3Dbottom panels) whereas the control tumors showed lowerlevels to undetectable levels of E-cadherin in several re-gions within the tumors of the control mice (Fig 3Dtop panels) Surprisingly intense β-casein localizationwas observed in the enlarged ducts ofMalat1ASO-treatedtumors (Fig 3E bottompanels) consistentwith the differ-entiation phenotype In contrast β-casein was absent inthe control tumor sections (Fig 3E top panels)

Arun et al



Malat1 ASO treatment significantly reduceslung metastasis

Given the reduction of metastasis seen in the MMTV-PyMTMalat1minusminus animals (Fig 1EndashH) it was important

to determine whether Malat1 knockdown could also re-sult in a reduction in the incidence of lung metastasis inthis therapeutic setting Indeed Malat1 ASO treatmentresulted in a significant reduction in lung metastatic inci-dence (Fig 4A) A small fraction of Malat1 ASO-treated

Figure 3 Malat1 knockdown results in differentia-tion of MMTV-PyMT tumors (A) RepresentativeHampE-stained primary tumor sections from micetreated with ScASO (top panels) or Malat1 ASO1 orASO2 (bottom panels) Bar 100 microm (B) The frequencyof occurrence of different histologic grades in scram-bled versus Malat1 ASO-treated tumor sectionsstained with HampE nge 20 tumors from each group(lowast)P lt 005 by Fisherrsquos exact test (C ) Percentage of cys-tic area in the tumor sections from scrambled (n = 4)versusMalat1ASO1-treated (n = 6) and ASO2-treated(n = 4) mice Error bars represent SD (lowastlowast) P lt 001 byWilcoxon signed rank test (D) Representative E-cad-herin immunolabeling on cryosections of tumorsfrom ScASO- or Malat1 ASO2-treated mice Bars100 microm (E) Representative β-casein immunolabelingon cryosections of tumors from ScASO-treated miceor Malat1 ASO1-treated mice Bars 100 microm




animals (six of 18) developed macrometastatic nodulesas compared with the control mice (10 of 12) (Fig 4B)Furthermore the lungs were serial-sectioned to deter-mine the incidence of micrometastases and obtain thetotal metastatic burden for a given lung volume Con-sistent with data obtained from genetic loss of Malat1(Fig 1EndashH) the Malat1 ASO-treated group showeda significantly reduced metastatic burden (lt10) (Fig4CD) compared with the larger metastatic burden(gt30) observed in the ScASO-treated group Micrometa-stases were also not observed in lungs that were free ofmacrometastatic nodules in mice treated with Malat1ASOs The livers and brains from theMalat1ASO-treatedgroups and control groups were also analyzed for meta-static incidence and no evidence of macrometastatic le-sions was found in those organs (data not shown)Thus we conclude thatMalat1 loss impairs the metastat-ic progression of the disease As suchMalat1 ASOs repre-sent a potent therapeutic approach for metastatic breastcancer

Malat1 loss affects branching morphogenesisin tumor-derived organoids

Three-dimensional (3D) organotypic culture systems al-low studies of physiologically relevant cellular processesas well as disease states under defined culture conditions(Ewald et al 2008 for review see Sato and Clevers 2013Sachs and Clevers 2014) Mammary gland organoidspreserve the gland characteristics and undergo variousmorphogenetic changes upon growth factor stimulusthat recapitulate normal physiological processes such asbranching morphogenesis (Ewald et al 2008 Cheunget al 2013) To investigate the role ofMalat1 in these pro-cesses organoids were prepared from MMTV-PyMT tu-mors lt1 cm in size These tumor organoids were grown

on Matrigel for up to 6 d in the presence or absence ofMalat1 ASOs allowing us to track growth dynamicsand branching morphogenesis of the organoids (Fig 5A)

TheMalat1ASOswere taken up ldquofreelyrdquo by the organo-ids without the use of transfection reagents and an sim90knockdown of theMalat1 transcriptwas achievedwith ei-ther Malat1 ASO relative to the ScASO control (Fig 5B)After 6 d in culture mock-treated organoids or ScASO-treated organoids underwent extensive branching mor-phogenesis (Fig 5C [top panels] D) In contrast organoidstreated with Malat1 ASO1 or ASO2 failed to undergobranching morphogenesis (Fig 5C [bottom panels] D)and remained as spherical acini that were smaller thanthe branched control organoids The loss of branchingwas found to be specific for ASOs targeting Malat1 asknockdown of Neat1 another nuclear lncRNA encodedin the same chromosomal region as Malat1 resulted inorganoids with branching morphogenesis similar to con-trol PyMT tumor organoids (Supplemental Fig S3A) Fur-thermore detailed time-lapse differential interferencecontrast (DIC) imaging showed that control organoids un-derwent active remodeling throughout the imaging win-dow (4 d) displaying extensive movement within theMatrigel (Supplemental Movie S1) The mock-treated orScASO-treated organoids expanded rapidly forming lu-mens and as early as 72 h initiated the branching process(Supplemental Movie S1) Numerous cells migrated with-in the organoids and these cells could be followed usingorganoids developed from MMTV-PyMTCAGH2B-GFPbitransgenic mice treated with ScASO (SupplementalMovie S3) In contrast Malat1 ASO1- or ASO2-treatedorganoids remained in the same position over the 84-htime course of imaging and did not show collective cellmigration within the Matrigel (Supplemental Movie S2)Most importantly cells within individual Malat1 ASO-treated organoids also lacked cellular motility and the

Figure 4 Lung metastasis is impaired inMalat1 knockdown mice (A) Representa-tive lungs with metastatic nodules fromScASO-treated and Malat1 ASO1- orASO2-treated mice (B) Quantitation oflung metastatic nodules from ScASO-treat-ed and Malat1 ASO1- or ASO2-treatedmice Each dot represents one mouse (lowast)P lt 005 (lowastlowast) P lt 001 by Wilcoxon signedrank test (C ) Representative HampE-stainedlung sections from ScASO-treated andMalat1ASO1-treatedmice Arrowheads in-dicate representativemicrometastases Bar2mm (D) Quantitation of the percentage oflung metastatic burden from ScASO-treat-ed and Malat1 ASO1- or ASO2-treatedmice n = 5 lung sections from each groupError bars represent SD (lowast) P lt 005 by Wil-coxon signed rank test

Arun et al



Figure 5 Malat1 knockdown affects branching morphogenesis in the tumor-derived organoids (A) Schematic of preparation of tumor-derived organoids (B) qRTndashPCR of relative Malat1 RNA level in MMTV-PyMT tumor-derived organoids mock-treated or treated withScASO orMalat1ASO1 or ASO2 n = 4 independent experiments Error bars represent SD (C ) Differential interference contrast (DIC) im-ages of organoids that were mock treated or treated with ScASO or Malat1 ASO1 or ASO2 after 6 d of culturing Mock represents PBS-treated organoids Bar 125 microm (D) Percentage of branched organoids from PyMT tumors subjected to mock ScASO or Malat1 ASO1or ASO2 treatments n =ge 200 organoids from four biological replicates Error bars represent SD (E) Representative transmission electronmicroscopy (TEM) images of a ScASO-treated PyMT organoid showing loosely opposed tumor cells with prominent internal deposit ofMatrigel material (MG) and a Malat1 ASO1-treated PyMT organoid showing closely opposed polarized tumor cells with tight junctions(TJs) and evidence of secretory dense-core granules (DCGs) and lipid droplets (LDs) The insets are higher-magnification images of tightjunctions dense-core granules and lipid droplets Bars 1 microm (F ) DIC images of tumor organoids fromMMTV-PyMTMalat1++ (top panel)and MMTV-PyMTMalat1minusminus (bottom panel) Bar 100 microm (G) RT-qPCR of relative Malat1 RNA level in MMTV-CreFL-neo-NeuNT tu-mor-derived organoids n = 3 biological replicates Error bars represent SD (H) DIC image ofMMTV-CreFL-neo-NeuNT organoids treatedwithmock ScASO andMalat1ASO1 andASO2 after 6 d of culturing Bar 125 microm (I ) Percentage of branched organoids fromMMTV-CreFL-neo-NeuNT tumors subjected to treatment n =ge80 organoids from three biological replicates Error bars represent SD



organoids appeared highly compact (Supplemental MovieS4) Interestingly organoids treated with Malat1 ASOsshowed increased apoptotic cell death and a significantreduction in the number of dividing cells (SupplementalFig S3B)

To further characterize the distinct morphologicalchange induced upon Malat1 ASO treatment transmis-sion electron microscopic analysis was performed on theorganoids to evaluate ultrastructural differences (Fig 5ESupplemental Fig S3C) Consistent with the differentia-tion phenotype observed in the tumors treated withMalat1 ASOs the organoids also showed well-differenti-ated subcellular structures such as the presence of secre-tory lipid droplets and dense core granules that wereabsent in the control organoids treated with ScASO orPBS Importantly Malat1 ASO-treated organoids exhibit-ed tight cellndashcell contacts with little to no intercellularspaces An increase in cellndashcell junctional complexessuch as desmosomeswas observed inMalat1ASO-treatedorganoids compared with the control branched organoids(Supplemental Fig S3C) Consistent with this observa-tion desmosomes have previously been shown to nega-tively regulate branching morphogenesis (Basham et al2013) The cells within theMalat1ASO-treated organoidswere aligned adjacent to each other in a polarized man-ner with intercellular ducts that are present in normalmammary glands these ducts are one of the distinct fea-tures of differentiated epithelial cells (Zhan et al 2008Watson andCampbell 2009) The ScASO- ormock-treatedPyMT organoids lacked polarity and the vast majority ofthe organoids contained Matrix proteins within the orga-noids which reflected the active extracellular matrix(ECM) remodeling necessary for migration (SupplementalFig S3C) Last organoids derived from MMTV-PyMTMalat1minusminus tumors did not undergo branching morphogen-esis and remained as spherical tightly packed acini(Fig 5F) whereas the MMTV-PyMTMalat1++ organoidswere extensively branched The loss of branching pheno-type observed in the MMTV-PyMTMalat1minusminus organoidscould be rescued by full-length mouse Malat1 that wasnucleofected into those organoids (Supplemental FigS4AndashD) We observed sim30 branching in MMTV-PyMTMalat1minusminus organoids that were nucleofected with full-length mouse Malat1 In contrast 5 of organoids dis-played branching in mock transfected or GFP transfectedorganoids (Supplemental Fig S4D) The nucleofection ef-ficiency ranged between 35 and 45 (data not shown)The effect ofMalat1 knockdown was also studied in orga-noids derived from MMTV-creFlox-neo-neuNT mice amodel that phenocopies the HER2-amplified subtype ofhuman breast cancer Here the first exon of endogenousneu (mouse homolog of HER2ERBB2) has been replacedwithCre-inducible activated neu-cDNA resulting in con-stitutively active Neu protein transcribed from its endog-enous promoter in mammary epithelial cells (Andrecheket al 2000) Control organoids derived from thismodel un-derwent branching morphogenesis while the Malat1ASO-treated organoids failed to undergo branching mor-phogenesis resulting in the formation of tightly packedspherical acinar structures (Fig 5GndashI)

Malat1 knockdown in tumors disrupts protumorigenicsignaling pathways

Next we sought to investigate the gene expression chang-es that are triggered upon Malat1 knockdown in both tu-mor organoids and primary tumors byRNA-seq RNAwasisolated from day 6 organoids that were either untreatedmock control or treated with ScASO or Malat1 ASO1 orASO2 We used tumors from four MMTV-PyMT mice toprepare organoids that were subjected to different ASOintervention The differentially expressed genes inMalat1ASO1- or ASO2-treated organoids or tumor sections werecompared with ScASO- or mock-treated controls Whenwe subjected the entire organoid data set to pathway anal-ysis using the Kyoto Encyclopedia of Genes and Genomes(KEGG) several tumorigenic signaling pathways were en-riched Interestingly enrichment of genes belonging tothe ECM receptor focal adhesion and cell adhesion path-ways was observed (false discovery rate [FDR] lt 01) (Sup-plemental Table S1) These pathway changes are highlyconsistent with the phenotypic changes observed in theorganoids upon Malat1 ASO treatment ie loss of cellu-lar migration and increased adhesion observed by the lossof branching morphogenesis

RNA-seq analysis was also carried out from tumors de-rived fromMalat1ASO1- or ASO2-treatedmice that werecompared with tumors from ScASO-treated mice and 52genes were affected upon Malat1 ASO treatment with asignificance threshold of FDR lt 01 (Supplemental FigS5A) Additionally 478 genes were found to be differen-tially expressed these genes were filtered based on foldchange cutoff from both ASO1- and ASO2-treated tumorsand exhibited a significant change in one or the othertreated tumors (Supplemental Table S2) The sample cor-relation plot also reflected the heterogeneity between thetumors Nevertheless the top 10 up-regulated and down-regulated genes from the tumors were indeed very inter-esting candidates (Fig 6AB) Of the up-regulated genesthe casein gene family was significantly enriched Thisincluded Csn2 that encodes for the β-casein protein themajor milk protein and Csn1s2a a member of the α-casein family Additionally Wap (whey acidic protein)Lao and Cel which are also constituents of mammalianmilk were also up-regulated in many of the tumors Inlight of the significant differentiation phenotype thatwas observed in theMalat1ASO-treated tumors it was re-assuring to observe that these genes were also up-regulat-ed gt20-fold in these tumors Additionally up-regulation ofthe Pip gene is a confirmation of the cystic phenotype asPip (also known as gross mammary gland cystic fluid) is aprotein that is secreted in cystic mammary gland lesions(Collette et al 1986) Many of the lactation-related geneswere validated by RT-qPCR from the ASO-treated tumorsas well as Malat1 knockout tumors (Supplemental FigS5B) Interestingly none of the genes genomically adja-cent to Malat1 showed significant expression change ex-cept Cdc42ep2 which showed 13-fold up-regulation inthe ASO-treated tumors We subjected the entire dataset to gene set enrichment analysis (GSEA) (Subramanianet al 2005 Luo et al 2009) and interestingly significant

Arun et al



enrichment of several gene sets involved in mammary tu-mor progression was observed (Fig 6C) indicating thatloss of Malat1 inhibits several gene expression programsinvolved in tumor progression Detailed pathway analysisusing KEGG also resulted in enrichment of genes belong-ing to the ECM signaling focal adhesion and cell adhe-sion pathways (Fig 6D) Nearly 90 of the tumor RNA-seq enriched pathways (eight of nine) were also enriched

in the organoid RNA-seq data analysis One of the mostsignificantly affected pathways was integrin signalingwith many of the protumorigenic integrins (Itga2b andItga5b3) being down-regulated inMalat1-depleted tumors(Fig 6E) To identify the transcription factors potentiallycooperating with Malat1 we performed both known andde novo motif analysis of the promoter region (minus900+100 in relation to the TSS) of the differentially expressed

Figure 6 Knockdown of Malat1 affects tumor progression by altering many signaling pathways (A) Top 10 significantly up-regulatedgenes (B) Top 10 significantly down-regulated genes (C ) GSEA of tumor data sets showing enrichment of genes that affect tumor progres-sion uponMalat1 knockdown (D) List of pathways that were significantly affected in tumors afterMalat1ASO1 or ASO2 treatment com-pared with a ScASO-treated control Highlighted in gray are common pathways that are affected in both organoids and tumors (E)Representative example of the integrin pathway that was significantly altered in Malat1 ASO-treated tumors versus the control (F ) Denovo and known motif analyses of the promoters of differentially expressed genes




genes and they showed significant enrichment in theSox5- Tcfcp2l1- and E2A-bindingmotifs in their promot-er regions (Fig 6F) In addition we took a sequence-inde-pendent approach by searching for overlaps betweenMalat1 putative targets and gt359 ChIP-seq (chromatinimmunoprecipitation [ChIP] combined with deep se-quencing) and ChIPndashChIP experiments annotated in theChIP Enrichment Analysis (ChEA) database From a totalof 140 transcription factors Sox9 was identified as a topscorer (P = 00002) Sox9 is a HMG protein that sharesthe similar DNA-binding element that was predicted forSox5 in our motif analysis (data not shown) Interestinglywhen comparing differentially expressed genes uponMalat1 knockdown to CHART-seq (West et al 2014) per-formed in MCF7 cells we found that 11 of our differen-tially expressed genes possessed MALAT1 peaks(Supplemental Table S3) suggesting that Malat1 coulddirectly regulate gene expression by physically bindingto regulatory regions of those genes Interestingly one ofthe up-regulated genes Tnxb (Tenascin Xb) an ECM pro-tein that has been shown to have anti-metastatic proper-ties (Matsumoto et al 2001) had five Malat1-bindingpeaks with high peak scores (Supplemental Table S3)

Malat1 knockdown affects multiple splicing events

As Malat1 has been implicated in alternative pre-mRNAsplicing regulation in several cell lines (Tripathi et al2010) and in modulating the recruitment of splicing fac-tors to an actively transcribing transgene reporter (Bernardet al 2010) we investigated whether Malat1 knockdownalso resulted in changes in the in vivo splicing pattern ofgenes involved in tumorigenesis The SpliceTrapSplice-Duo pipeline (Wu et al 2011) was applied to the RNA-seq data sets generated from the tumors and organoidsSpliceTrap performs probabilistic inference of the extentto which an exon is included in the mature transcript isskipped or is subjected to size variations due to alterna-tive 3prime5prime splice sites or intron retention SpliceDuo imple-ments a noise-based methodology to identify statisticallysignificant alternative splicing events across two condi-tions (see Supplemental Fig S6A for splicing analysispipeline) Interestingly numerous alternative splicingchanges were observed in both the Malat1 ASO-treatedtumors and organoids when compared with the ScASO-treated samples (Fig 7AB Supplemental Table S4) Wedetected a total of 1351 splicing changes spanning thefour major classes of alternative splicing patterns in atleast two biological replicates resulting in a 60 overlapbetweenASO1 andASO2 treatment (Fig 7B) This findingsuggests an involvement ofMalat1 in alternative splicingofmany important genes Interestingly Esr1 coding for es-trogen receptor α (ERα) was found to be alternativelyspliced in its 5prime untranslated region (UTR) in Malat1ASO-treated samples compared with ScASO-treated tu-mors (Fig 7C Supplemental Table S4) When we lookedat ERα protein expression using immunoblot analysis(Supplemental Fig S6B) we observed that the control tu-mors showed lower ERα expression compared withMalat1 ASO-treated tumors It is known that the

MMTV-PyMT tumors lose ER expression in the advancedcarcinoma stage Since most of the Malat1 ASO-treatedtumors did not progress to advanced carcinomas it is pos-sible that they retained ERα expression It remains to befurther investigated whether the switch in the ER isoformconfers stability to the protein

Of particular note Sox5 was alternatively spliced giv-ing rise to an isoform that lacked the exon that encodesfor a loop domain of this HMG transcription factor (Sup-plemental Fig S6CD) This finding was exciting as motifanalysis of differentially regulated genes upon Malat1knockdown also showed enrichment of Sox5-bindingmotifs in their promoters (Fig 6F) Furthermore manymembers of ECM signaling were also alternativelyspliced and Itga2b (Fig 7D Supplemental Fig S6E) be-longing to the integrin signaling pathway showed alter-nate 3prime splice site utilization that resulted in retentionof part of an intron This transcript was shown to bedown-regulated in our gene expression analysis as wellby qRTndashPCR analysis

Overall Malat1 knockdown affects both gene expres-sion and alternative splicing of numerous protumorigenicsignaling molecules and differentiation-related genes re-sulting in a dramatic shift inmammary tumors to a highlydifferentiated less aggressive state

Discussion

Malat1 a nuclear-localized lncRNA is not required fornormal development (Eissmann et al 2012 Nakagawaet al 2012 Zhang et al 2012) Here using a mouse modeland 3D organotypic cultures we demonstrate for the firsttime a functional role for Malat1 in breast cancer pro-gression Malat1 loss results in a dramatic differentiationof the primary tumor and a significant reduction inmetas-tasis These findings demonstrate a critical role forMalat1loss in redirecting the program of a poorly differentiatedcarcinoma to that of a highly differentiated ldquotissuerdquo Im-portantly systemic knockdown of Malat1 using ASOsphenocopies genetic loss thereby providing an excitingfuture avenue for exploring the use of Malat1 ASOs in atherapeutic setting

Malat1 level increases upon cancer progression

Consistent with our findings that Malat1 plays a role intumor progression RNA-FISH analysis of paired humanbreast tumor samples andmetastases indicated that whilethe level of Malat1 is higher in the primary tumor thanin the surrounding stroma its level is further increasedin the lung metastases (Fig 7E) Of particular interest isthe fact that some cells in the primary tumor exhibit ahigher level ofMalat1 andwe suggest that itmay be thesecells that escape from the primary tumor and metastasizeto the lungs The up-regulation ofMalat1may be triggeredby physiological stress andor oncogene activation Inter-estingly several previous studies in other cancers havefound a correlation of Malat1 levels with EMT transcrip-tion factors such as Zeb1 and Snail Although we did not

Arun et al



observe changes in the expression of EMT transcriptionfactors in our data sets it is possible that Malat1 expres-sion could increase in cells that are undergoing EMTand that may correlate well with those cells preferentiallymetastasizing In this context it has been shown previ-ously that hypoxia-inducible factor HIF1A regulatesMALAT1 expression (Choudhry et al 2014 Michaliket al 2014)Hif1a is a major transcription factor that gov-erns oxidative stress and regulates tumor progression andhas also been shown to play a role in EMT (Semenza et al2003) In a number of cell culture studies Malat1 wasfound to be up-regulated upon stress such as serum starva-tion hypoxia and genotoxic stress all of which are linkedto tumor progression (Yang et al 2011 Mizutani et al

2012 Choudhry et al 2014) Consistent with this ourfindings demonstrated that Malat1 loss in dedifferentiat-ed primary tumors results in extensive alteration ingene expression activating pathways that are involvedin morphogenesis and differentiationMammary gland differentiation has been shown to be

regulated by GATA3 (Kouros-Mehr et al 2008) and Statsignaling (Liu et al 1997) Although we found that theJakndashStat signaling pathway is altered in tumor organoidsupon loss of Malat1 this change was not captured in theprimary tumors and Gata3 was not differentially ex-pressed in our data sets Thus additional factors likelyregulate mammary tumor differentiation upon Malat1loss Consistent with this possibility ECM proteins and

Figure 7 Malat1 knockdown impacts alternative pre-mRNA splicing of the pre-mRNAs (A) Venn diagram showing the total number ofsplicingchanges inASO1- andASO2-treated tumorsand tumororganoidswith respect to scrambled-treatedcontrolsNearly60of chang-es overlap between ASO1- and ASO2-treated groups in at least two biological replicates (B) Different types of splicing changes that are af-fected and the number of genes that are alternatively spliced uponMALAT1ASO treatment in each of them (C ) University of California atSantaCruz (UCSC) screenshot showing an example of an alternatively spliced exon in theEsr1 transcript (D) UCSCscreenshot showing anexample of alternatively spliced exons in the Itga2b transcript resulting in partial intron retention (E) Schematic of a working model ofMalat1 RNA levels Malat1 is expressed at a lower level in the differentiated mammary gland (normal) and early lesions (hyperplasia)Malat1 up-regulation inmammary tumors (carcinoma) is concomitant with dedifferentiation andmetastasis of the tumors (F ) Schematicof the proposed molecular mechanism The lncRNA Malat1 (green squiggles) shuttles between nuclear speckles and TSSs (arrowhead)where it can serve as a scaffold to regulate efficient transcription and alternative pre-mRNA splicing (red line)




integrin signaling are altered upon loss ofMalat1 and dif-ferentiation markers of the mammary gland such as ca-sein genes are up-regulated several-fold resulting insecretion of these proteinswithin the cysts and thereby re-storing a functional activity of the mammary gland Inter-estingly ECM proteins and integrin signaling have beenpreviously shown to play a crucial role in the regulationof casein genes (Streuli et al 1991 1995 Pontier andMul-ler 2009) and prior studies have indicated that lactationconfers cancer protection to the mammary gland due todifferentiation of epithelial cells (Kotsopoulos et al 2012)

Poorly differentiated primary tumors are often associat-ed with a poor outcome and higher incidence of metasta-sis (for review seeMedema 2013) Therefore the ability todifferentiate primary tumors by loss or knockdown ofMalat1 significantly reduces the possibility that theywill undergo EMT-like processes and become invasiveMultiple lines of evidence converge on this aspect ofMalat1 function as its loss has been previously shownto reduce motility in cell lines and the homing behaviorof lung cells in xenograft mouse models (Tano et al2010 Gutschner et al 2013) Our findings add signifi-cantly to these prior observations by showing that lossof Malat1 in tumor-bearing mice results in a 70 reduc-tion in lungmetastasis We also found thatMalat1 loss in-hibited branching morphogenesis in tumor organoidsconsistent with its requirement for migration and localECM remodeling of the organoids in Matrigel a processthat is similar to invasion and matrix degradation ob-served in vivo in malignant tumors (Ewald et al 2008Cheung et al 2013) In addition Malat1 loss in bothMMTV-PyMT and Her2neu-amplified MMTV-creFlox-neo-neuNT tumor organoids resulted in increased epithe-lial cell adhesion and loss ofmigration of cells resulting intightly packed spherical acinar structures Given that thebranching phenotype could be restored by addition ofmouse full-lengthMalat1Malat1 is essential for this pro-cess and the phenotype observed is specific for Malat1Our data add significant support to the concept that orga-noids can provide a rapid means of assessing therapeuticpotential in a patient-specific manner (Boj et al 2015)

Malat1 functions as a cotranscriptional splicing scaffold

Although Malat1 knockout mice exhibit no phenotype(Eissmann et al 2012 Nakagawa et al 2012 Zhanget al 2012) upon Malat1 knockout or knockdown in theMMTV-PyMT mammary cancer model we observedchanges in gene expression and pre-mRNA splicing of nu-merous cancer-relevant genes Lack of significant overlapof differentially expressed genes upon Malat1 loss amongvarious studies supports a role forMalat1 in a context-de-pendent manner Based on cross-linking studies it hasbeen shown that Malat1 is bound to either chromatin(West et al 2014) or pre-mRNAs (Engreitz et al 2014)Since it is well documented that transcription and pre-mRNA splicing are coupled processes (for review seeBentley 2014) our data suggest that Malat1 may act asa scaffold to coordinate these processes in a gene- andcontext-specific manner (Fig 7F) The presence ofMalat1

in nuclear speckles domains that are commonly foundnear actively transcribing genes and are also enriched insplicing and transcription factors (for review see Spectorand Lamond 2011) could facilitate the function ofMalat1at sites of transcription Interestingly based on compari-sons between a CHART-seq data set from MCF7 cellsthat identified directMALAT1-binding sites across the ge-nome and our gene expression data set we identifiedsim11 of our differentially expressed genes to be potentialdirect targets ofMalat1 which is consistent with the pos-sibility of Malat1 coordinating expression of these genesIn linewith this we previously showed that loss ofMalat1impairs the recruitment of SR splicing factors to a stablyintegrated transgene array (Bernard et al 2010) It hasalso been previously shown that numerous splicingchanges occur during tumor progression giving rise tonovel oncogenic splice variants (Venables et al 2008Eswaran et al 2013) In this regard SRSF1 a SR protein-splicing factor can act as an oncoprotein that promotesbreast cancer (Anczukow et al 2012) During tumor pro-gression Malat1 may facilitate efficient transcriptionand splicing of protumorigenic genes such as integrinsECM proteins and genes involved in migration and me-tastasis leading to a more favorable outcome for the tu-mor Thus loss of Malat1 results in altered transcriptionand pre-mRNA splicing events more consistent with thedifferentiated phenotype of the impacted tissue Giventhe importance of estrogen signaling in luminal breastcancer and our observation that the estrogen receptor(Esr1) transcript was alternatively spliced in its 5prime UTRuponMalat1 loss it will be important in follow-up studiesto determine whether Esr1 isoform switching plays a rolein the observed differentiation phenotype

It is presently unclear whether Malat1 binds to certainproteins or transcription factors to elicit its regulatoryrole Previously published CLIP-seq (cross-linking immu-noprecipitation [CLIP] combined with deep sequencing)and RNA pull-down studies from several groups haveshown that Malat1 binds to splicing factors such asSRSF1 TDP43 and PRC2 components including EZH2(Tripathi et al 2010 Tollervey et al 2011 Yang et al2011 Guil et al 2012) Detailed motif analysis of the pro-moters of the differentially expressed genes in our datasets show them to be enriched for TCFCP2L1 SOX5and E2A transcription factor-binding sites which have es-tablished roles in tissue differentiation and pluripotency(Chen et al 2008 Kamachi and Kondoh 2013 Ye et al2013) Therefore we suggest that the specificity of theMalat1-responsive genes may be context-dependent andarise from the transcription factors that primarily regulatethose genes For example TCFCP2L1 has been shown tocooperate with the estrogen receptor family of transcrip-tion factors such as ESRRB (Boroviak et al 2014) andhas been shown to be associatedwith breast cancermetas-tasis to the lungs (Landemaine et al 2008)

Malat1 ASOs are a promising therapeutic

Of potential clinical impact we employed antisense-mediated knockdown using Malat1-specific ASOs in a

Arun et al



mouse model of metastatic mammary cancer ASOs arecurrently in clinical trials for many diseases includingspinal muscular atrophy and Duchenne muscular dystro-phy (Aartsma-Rus et al 2004 Rigo et al 2012) The in vivoefficacies of Malat1 ASOs have been tested in primateswhere gt80 knockdown efficiency was achieved (Kolleret al 2011 Hung et al 2013) Here we show that Malat1knockdown results in slower tumor proliferation anddifferentiation and significant reduction in metastasisall recapitulating the phenotype of the MMTV-PyMTMalat1minusminus genetic model thus confirming the specificityof these ASOs Future studies using patient-derived xeno-graft models as well as patient-derived tumor organoidswill enable us to further study the efficacy of Malat1ASOs as a therapeutic to induce differentiation of primaryhuman breast tumors and significantly reduce tumorprogression

Materials and methods

Animals

Animal experiments were carried out in the Cold Spring HarborLaboratory Animal Shared Resource in accordance with Institu-tional Animal Care and Use Committee-approved proceduresMalat1 knockout (C57BL6) mice were generated as describedpreviously (Zhang et al 2012) and bred with MMTV-PyMT(C57BL6) mice (Guy et al 1992)CAGH2B-EGFPmice (Hadjan-tonakis and Papaioannou 2004) were obtained from Jackson Lab-oratory and bred with MMTV-PyMT mice MMTV-Cre and Fl-neo-neu-NT mice were obtained from Dr William Muller (Mc-Gill University Montreal) and were bred to generate MMTV-CreFl-neo-neu-NT

RNA in situ hybridization

RNA in situ hybridization was performed on formalinPFA-fixedparaffin-embedded tissue sections using specific probes againsthuman ormouseMalat1 purchased fromAffymetrix The humantissue sections were obtained fromDr Edi Brogi (Memorial SloanKetteringCancer Center) Bothmatched primary tumor sectionsmetastasis and unmatched array samples were scored forMALAT1 level The in situ hybridization protocol using View-RNA technology was performed according to the manufacturerrsquosinstructions (Affymetrix) Slides were scanned using an Aperioscanner and the images were analyzed using Leica Scanscopesoftware Alternatively nick-translated Malat1 probes labeledwith fluorescently conjugated dUTPs were used for RNA-FISHas described previously (Zhang et al 2012) Images were capturedusing a DeltaVision microscope (GE) For all of the quantitativeanalysis the fluorescent intensity profiles were obtained fromat least five different random fields When array spots were usedfor intensity measurement the total intensity of the spot withROImarkupwas used to estimate the signal intensity The inten-sity values were binned to obtain the range which was then usedfor plotting

Organoid culture

Organoids from MMTV-PyMT and MMTV-CreFlox neo-neu NTtumors were prepared and cultured as described previously(Ewald 2013) Organoids were mixed with Matrigel and platedin 24-well Mat-Tek dishes For ASO treatment 500 nM Malat1ASO or scASO was added to 1 mL of organoid culture medium

Images were acquired using a Zeiss Axio-Observer light micro-scope For live-cell imaging organoids from MMTV-PyMT orMMTV-PyMTCAGH2B-EGFP mice were treated with Malat1or ScASOs At least five organoids per well were imaged in threedimensions over a period of 4 d at 15- or 30-min intervals using aPerkin-Elmer spinning-disc confocal microscope Time-lapse im-ages were processed using Volocity (PerkinElmer) and ImageJ(National Institutes of Health) software

Electron microscopy

Organoids grown inMatrigel in 12-well culture dishes were fixedovernight with 2 glutaraldehyde and 2 paraformaldehyde inPBS While in the multiwell dish the samples were rinsed withdistilled water and then post-fixed with 1 osmium tetroxidein 15 potassium ferrocyanide for 1 h After a distilled waterrinse discs of Matrigel containing embedded organoids were lift-ed out of the culture dishes cut into 2-mmtimes 10-mmtimes 4-mmpiec-es and transferred into 20-mL glass scintillation vials Sampleswere dehydrated in a graded series of ethanol and then rinsed in100acetone Resin embeddingwas donewith continuous agita-tion in 50 epon-araldite resin in acetone for 1 h and overnightincubation in 100 epon-araldite with agitation Samples werethen placed into BEEM capsules (placed flat into the cap end ofa conical capsulemade by cutting off the conical end) and the res-in was polymerized overnight at 60degC in a vented oven Sampleswere sectioned at 100 nm collected on butvar-coated nickelslot grids (1-mmtimes 2-mmopening ElectronMicroscopy Sciences)and counterstained with lead citrate Sections were examinedin a Hitachi H-7000 transmission electron microscope operatedat 75 kV Representative images were recorded on Kodak 4489film Images were scanned at 2400 dpi with an Epson PerfectionVP750 Pro

ASO treatment and tumor studies

Themice selected for ASO treatment were grouped based on ageinitial tumor size and separation of siblings into three cohortsThe ASOs against Malat1 or ScASO were reconstituted at a con-centration of 5 mgmL in PBS and injected subcutaneously everyday at a concentration of 25mgkg per day in a volume ofsim125 μLfor 5 d with a rest period of 2 d Tumor growth was measuredtwice perweek using calipers in a nonblinded fashion Tumor vol-ume was calculated based on the formula volume =width2 timeslength2 Mice were sacrificed after 8 wk of treatment and tu-mors were removed and fixed in 4 PFA Upon sacrifice thelungswere also removed and themacronodules were counted us-ing a dissection microscope Other internal organs including thespleen liver kidney and brain were examined for metastasis orabnormalities using a dissection microscope

Histology

Primary tumors and lungswere fixed in 4PFA for up to 24 h andwashed in PBS Primary tumors were embedded in paraffin sec-tioned and stained with HampE Slides were scanned using anAperio slide scanner The pathologist performed blind analysesto assess the histologic grade The cystic area calculation wasdone using Aperio Scanscope software After fixation lungswere incubated in 20 sucrose overnight and embedded inOCT and the lungs were cross-sectioned 2 mm apart The lungsections were embedded horizontally to obtain serial sectionsof the entire lung and HampE-stained sections were scanned withan Aperio image scope The lungmetastatic burden was calculat-ed using Aperio Scanscope and ImageJ (National Institutes ofHealth) software




Immunofluorescence

Immunofluorescence on the tumor cryosections was performedas described previously Briefly sections were washed in PBSand permeabilized for 5 min in 005 Triton X The sectionswere blocked using 1 goat serum followed by incubation withprimary antibody (Ki-67 from Vector Laboratories E-cadherinfrom BD β-casein was a gift fromDr Mina Bissel) for 4 h to over-night at 4degC Subsequently the slides were washed with PBS +001 Triton X Finally the sections were incubated in Alexafluor-conjugated secondary antibodies (Jackson ImmunoRe-search) followed by washes with PBS The sections were stainedwith DAPI and mounted in 90 glycerol and 10 PBS plusPPD (paraphynelediamine) Images were acquired using a ZeissLSM 710 or Zeiss Axio-Observer microscope Ki-67 was quanti-fied by counting the number of positive nuclei for every 100 nu-clei from multiple fields on the slide to obtain the percentage ofpositive cells

RNA-seq and analysis

Organoids generated from fourMMTV-PyMTmicewere culturedin 3DMatrigel domes in the presence of FGF2 for 6 d The organo-ids were treated with 500 nM mock ScASO or Malat1 ASOsIn addition tumors were removed from MMTV-PyMT micethat had undergone injections with ScASO (threemice) orMalat1ASO1 (threemice) or ASO2 (fourmice) Total RNAwas extractedusing TRIzol (Life Technologies) fromboth organoids and homog-enized tumor tissue Libraries for polyA+ RNA-seq were preparedfrom 1 μg of RNA per sample using TruSeq chemistry (Illumina)multiplexed and sequenced to obtain paired-end 101-base-pair(bp) reads on an Illumina HiSeq 2000 platform resulting in 20million to 86 million reads per libraryThe quality of the raw data was evaluated using FastQC

(httpwwwbioinformaticsbabrahamacukprojectsfastqc)and reads were mapped to mm9 using STAR (Dobin et al2013) resulting in an overall mapping efficiency of gt90 TheGencode mV1 annotation set was used as a reference and thereads per gene record were counted using the HTSeq package(Anders et al 2015) Replicate analysis was carried out usingedgeR (Robinson et al 2010) Differential gene expression wasperformed with DESeq2 (Love et al 2014) Functional analysisof KEGG pathways and gene ontology (GO) terms was carriedout with the RBioconductor packages GAGE (Luo et al 2009)and Pathview (Luo and Brouwer 2013) Motif analyses were per-formed using Homer (httphomersalkeduhomer Heinz et al2010)

Splicing analysis

Alternative splicing including cassette exons alternative 3prime5prime

splice site and intron retention events were identified and quan-tified using the SpliceTrapSpliceDuo pipeline (Wu et al 2011)This tool combinesRNA-seq datawith prebuilt transcriptmodelsto quantify the level of inclusion of every exon in a transcript Thetranscript models are exon trios composed of alternative exoncandidates with their annotated flanking exons These were de-rived from the mm9 TXdb database which accounts for a totalof gt230000 exon trio models describing the alternative splicingof sim190000 mouse exons in sim44000 transcripts SpliceTrapuses the Bowtie read aligner to align reads against TXdb Biologi-cal replicates were analyzed independently and alternative splic-ing events that were reproduced in at least two samples wereselected (SpliceDuo FDR lt 01) Inconsistent splicing eventsshowing opposite splicing directions (skipping versus inclusion)were further removed

RT-qPCR

Onemicrogram of RNAwas used tomake cDNAs using oligo-dTprimers One-tenth the volume of the cDNA reaction mixturewas used in the qPCR reaction containing 1times SYBRGreenmastermix (Applied Biosystems) and specific forward and reverseprimers The relative fold change was calculated using the ΔΔCt

method The primers used were as follows Malat1 F (AACCAGTTTCCCCAGCTTTT) and Malat1 R (CTACATTCCCACCCAGCACT)

Accession number

RNA-seq data presented in this study have been submitted to theNCBI Gene Expression Omnibus (GEO http httpwwwncbinlmnihgovgeo) under accession number GSE67647

Acknowledgments

We thank members of the Spector laboratory for critical discus-sions throughout the course of this study and Mona Spector forcritically reviewing the manuscript We thank Dr William Mul-ler (McGill University Montreal) for providing MMTV-Creand Fl-Neo-NeuNT mice We also thank Dr Mina Bissel forthe β-casein antibody We thankDr Alain Bessis (Institut de Biol-ogie de lrsquoEacutecole Normale Supeacuterieure France) for the full-lengthmouse Malat1 cDNA We acknowledge the Cold Spring HarborLaboratory Cancer Center Animal Animal and Tissue Imagingand Microscopy Shared Resources (National Cancer Institute2P3OCA45508) We thank Cecilia Yan for assistance with theorganoid movies and Jacqueline Cappalleni for assistance withtheMMTV-PyMTmice This research was supported by Nation-al Cancer Institute 5P01CA013106-Project 3 (DLS) We also ac-knowledge support from the Babylon Breast Cancer Coalition DLS is a consultant to Ionis Pharmaceuticals

References

Aartsma-Rus A Janson AA Kaman WE Bremmer-Bout M vanOmmen GJ den Dunnen JT van Deutekom JC 2004 Anti-sense-induced multiexon skipping for Duchenne musculardystrophy makes more sense Am J Hum Genet 74 83ndash92

Ades F Zardavas D Bozovic-Spasojevic I Pugliano L FumagalliD de Azambuja E Viale G Sotiriou C Piccart M 2014Luminal B breast cancer molecular characterization clinicalmanagement and future perspectives J Clin Oncol 32 2794ndash2803

AnczukowORosenbergAZAkermanMDas SZhanLKarniRMuthuswamySKKrainerAR 2012The splicing factor SRSF1regulates apoptosis and proliferation to promotemammaryep-ithelial cell transformation Nat Struct Mol Biol 19 220ndash228

Anders S Pyl PT Huber W 2015 HTSeqmdasha Python frameworkto work with high-throughput sequencing data Bioinfor-matics 31 166ndash169

Andrechek ER HardyWR Siegel PM Rudnicki MA Cardiff RDMullerWJ 2000 Amplification of the neuerbB-2 oncogene ina mouse model of mammary tumorigenesis Proc Nat AcadSci 97 3444ndash3449

Basham KJ Kieffer C Shelton DN Leonard CJ Bhonde VR Van-kayalapati H Milash B Bearss DJ Looper RE Welm BE 2013Chemical genetic screen reveals a role for desmosomal adhe-sion in mammary branching morphogenesis J Biol Chem288 2261ndash2270

BentleyDL 2014 CouplingmRNAprocessingwith transcriptionin time and space Nat Rev Genet 15 163ndash175

Arun et al



Bergmann JH Spector DL 2014 Long non-coding RNAs modu-lators of nuclear structure and function Curr Opin Cell Biol26 10ndash18

BernardD PrasanthKVTripathiVColasse SNakamuraTXuanZ Zhang MQ Sedel F Jourdren L Coulpier F et al 2010 Along nuclear-retained non-coding RNA regulates synaptogen-esis by modulating gene expression EMBO J 29 3082ndash3093

Boj SF Hwang CI Baker LA Chio II Engle DD Corbo V JagerMPonz-Sarvise M Tiriac H Spector MS et al 2015 Organoidmodels of human and mouse ductal pancreatic cancer Cell160 324ndash338

Boroviak T Loos R Bertone P SmithANichols J 2014 The abil-ity of inner-cell-mass cells to self-renew as embryonic stemcells is acquired following epiblast specification Nat CellBiol 16 516ndash528

Bussemakers MJ van Bokhoven A Verhaegh GW Smit FP Kar-thaus HF Schalken JA Debruyne FM Ru N Isaacs WB1999 DD3 a new prostate-specific Gene highly overex-pressed in prostate cancer Cancer Res 59 5975ndash5979

Chen X Xu H Yuan P Fang F Huss M Vega VB Wong E OrlovYL ZhangW Jiang J et al 2008 Integration of external signal-ing pathways with the core transcriptional network in embry-onic stem cells Cell 133 1106ndash1117

Cheung KJ Gabrielson E Werb Z Ewald AJ 2013 Collective in-vasion in breast cancer requires a conserved basal epithelialprogram Cell 155 1639ndash1651

Choudhry H Schodel J Oikonomopoulos S Camps C Grampp SHarris AL Ratcliffe PJ Ragoussis J Mole DR 2014 Extensiveregulation of the non-coding transcriptome by hypoxia role ofHIF in releasing paused RNApol2 EMBO Rep 15 70ndash76

Clemson CM Hutchinson JN Sara SA Ensminger AW Fox AHChessA Lawrence JB 2009An architectural role for a nuclearnoncoding RNA NEAT1 RNA is essential for the structure ofparaspeckles Mol Cell 33 717ndash726

Collette J Hendrick JC Jaspar JM Franchimont P 1986 Presenceof α-lactalbumin epidermal growth factor epithelial mem-brane antigen and gross cystic disease fluid protein (15000daltons) in breast cyst fluid Cancer Res 46 3728ndash33

Costa FF 2007 Non-coding RNAs lost in translationGene 3861ndash10

Derrien T Johnson R Bussotti G Tanzer A Djebali S Tilgner HGuernecGMartinDMerkel A KnowlesDG et al 2012 TheGencode v7 catalog of human long noncoding RNAs analysisof their gene structure evolution and expression GenomeRes 22 1775ndash1789

Djebali S Davis CA Merkel A Dobin A Lassmann T MortazaviA Tanzer A Lagarde J LinW Schlesinger F et al 2012 Land-scape of transcription in human cells Nature 48 101ndash108

Dobin A Davis CA Schlesinger F Drenkow J Zaleski C Jha SBatut P ChaissonM Gingeras TR 2013 STAR ultrafast uni-versal RNA-seq aligner Bioinformatics 29 15ndash21

Eissmann M Gutschner T Hammerle M Gunther S Caudron-Herger M Gross M Schirmacher P Rippe K Braun T ZornigM et al 2012 Loss of the abundant nuclear non-coding RNAMALAT1 is compatible with life and development RNA Biol9 1076ndash1087

Ellis MJ Ding L Shen D Luo J Suman VJ Wallis JW Van TineBA Hoog J Goiffon RJ Goldstein TC et al 2012 Whole-ge-nome analysis informs breast cancer response to aromatase in-hibition Nature 486 353ndash360

The ENCODE Project Consortium 2012 An integrated encyclo-pedia of DNA elements in the human genome Nature 48957ndash74

Engreitz JM Sirokman K McDonel P Shishkin AA Surka CRussell P Grossman SR Chow AY Guttman M Lander ES

2014 RNAndashRNA interactions enable specific targeting ofnoncoding RNAs to nascent Pre-mRNAs and chromatin sitesCell 159 188ndash199

Eswaran J Horvath A Godbole S Reddy SD Mudvari P OhshiroK Cyanam D Nair S Fuqua SA Polyak K et al 2013 RNAsequencing of cancer reveals novel splicing alterations Scien-tific Rep 3 1689

Ewald AJ 2013 Isolation of mouse mammary organoids for long-term time-lapse imaging Cold Spring Harb Protoc 2013130ndash133

Ewald AJ Brenot A DuongM Chan BSWerb Z 2008 Collectiveepithelial migration and cell rearrangements drive mammarybranching morphogenesis Dev Cell 14 570ndash581

Guffanti A Iacono M Pelucchi P Kim N Solda G Croft LJ TaftRJ Rizzi E Askarian-Amiri M Bonnal RJ et al 2009 A tran-scriptional sketch of a primary human breast cancer by 454deep sequencing BMC Genomics 10 163

Guil S Soler M Portela A Carrere J Fonalleras E Gomez A Vil-lanueva A Esteller M 2012 Intronic RNAs mediate EZH2regulation of epigenetic targets Nat Struct Mol Biol 19664ndash670

GuptaRA ShahNWangKC Kim J HorlingsHMWongDJ TsaiMC Hung T Argani P Rinn JL et al 2010 Long non-codingRNAHOTAIR reprograms chromatin state to promote cancermetastasis Nature 464 1071ndash1076

Gutschner T Hammerle M EissmannM Hsu J Kim Y Hung GRevenko A Arun G Stentrup M Gross M et al 2013 Thenoncoding RNAMALAT1 is a critical regulator of the metas-tasis phenotype of lung cancer cells Cancer Res 73 1180ndash1189

GuyCT Cardiff RDMullerWJ 1992 Induction ofmammary tu-mors by expression of polyomavirus middle T oncogene atransgenic mouse model for metastatic disease Mol CellBiol 12 954ndash961

Hadjantonakis AK Papaioannou VE 2004 Dynamic in vivo im-aging and cell tracking using a histone fluorescent protein fu-sion in mice BMC Biotechnol 24 33

Harrow J Frankish A Gonzalez JM Tapanari E Diekhans MKokocinski F Aken BL Barrell D Zadissa A Searle S et al2012 Gencode the reference human genome annotation forThe ENCODE Project Genome Res 22 1760ndash1774

Heinz S Benner C Spann N Bertolino E Lin YC Laslo P ChengJXMurreC SinghHGlassCK 2010 Simple combinations oflineage-determining transcription factors prime cis-regulato-ry elements required for macrophage and B cell identitiesMol Cell 38 576ndash589

Herschkowitz JI Simin K Weigman VJ Mikaelian I Usary J HuZ Rasmussen KE Jones LP Assefnia S Chandrasekharan Set al 2007 Identification of conserved gene expression fea-tures between murine mammary carcinoma models and hu-man breast tumors Genome Biol 8 R76

Hung G Xiao X Peralta R Bhattacharjee G Murray S Norris DGuo S Monia BP 2013 Characterization of target mRNA re-duction through in situ RNA hybridization in multiple organsystems following systemic antisense treatment in animalsNucleic Acid Ther 23 369ndash378

Hutchinson JN Ensminger AW Clemson CM Lynch CR Law-rence JB Chess A 2007 A screen for nuclear transcripts iden-tifies two linked noncoding RNAs associated with SC35splicing domains BMC Genomics 8 39

Ji P Diederichs S Wang W Boing S Metzger R Schneider PMTidow N Brandt B Buerger H Bulk E et al 2003 MALAT-1 a novel noncoding RNA and thymosin β4 predict metasta-sis and survival in early-stage non-small cell lung cancerOncogene 22 8031ndash8041




Kamachi Y Kondoh H 2013 Sox proteins regulators of cell fatespecification and differentiation Development 140 4129ndash4144

Koller E Vincent TM Chappell A De S Manoharan M BennettCF 2011 Mechanisms of single-stranded phosphorothioatemodified antisense oligonucleotide accumulation in hepato-cytes Nucleic Acids Res 39 4795ndash4807

Kotsopoulos J Lubinski J Salmena L Lynch HT Kim-Sing CFoulkes WD Ghadirian P Neuhausen SL Demsky R TungN et al 2012 Breastfeeding and the risk of breast cancer inBRCA1 and BRCA2 mutation carriers Breast Cancer Res14 R42

Kouros-Mehr H Bechis SK Slorach EM Littlepage LE EgebladM Ewald AJ Pai SY Ho IC Werb Z 2008 GATA-3 links tu-mor differentiation and dissemination in a luminal breast can-cer model Cancer Cell 13 141ndash152

Landemaine T Jackson A Bellahcene A Rucci N Sin S AbadBM Sierra A Boudinet A Guinebretiere JM Ricevuto Eet al 2008 A six-gene signature predicting breast cancerlung metastasis Cancer Res 68 6092ndash6099

Lin EY Jones JG Li P Zhu LWhitney KDMullerWJ Pollard JW2003 Progression to malignancy in the polyoma middle Toncoprotein mouse breast cancer model provides a reliablemodel for human diseases Am J Pathol 163 2113ndash2126

Lin R Maeda S Liu C KarinM Edgington TS 2007 A large non-codingRNA is amarker formurine hepatocellular carcinomasand a spectrum of human carcinomasOncogene 26 851ndash858

Liu X Robinson GW Wagner KU Garrett L Wynshaw-Boris AHennighausen L 1997 Stat5a ismandatory for adultmamma-ry gland development and lactogenesis Genes Dev 11179ndash186

Love MI Huber W Anders S 2014 Moderated estimation of foldchange and dispersion for RNA-seq data with DESeq2 Ge-nome Biol 15 550

Luo W Brouwer C 2013 Pathview an RBioconductor packagefor pathway-based data integration and visualization Bioin-formatics 29 1830ndash1831

LuoW FriedmanMS SheddenKHankensonKDWoolf PJ 2009GAGE generally applicable gene set enrichment for pathwayanalysis BMC Bioinformatics 10 161

Matsumoto K Takayama N Ohnishi J Ohnishi E Shirayoshi YNakatsuji N Ariga H 2001 Tumour invasion and metastasisare promoted in mice deficient in tenascin-X Genes Cells 61101ndash1111

Medema JP 2013 Cancer stem cells the challenges ahead NatCell Biol 15 338ndash344

Michalik KM You X Manavski Y Doddaballapur A Zornig MBraun T John D Ponomareva Y Chen W Uchida S et al2014 Long noncoding RNA MALAT1 regulates endothelialcell function and vessel growth Circ Res 114 1389ndash1397

Mizutani R Wakamatsu A Tanaka N Yoshida H Tochigi NSuzuki Y Oonishi T Tani H Tano K Ijiri K et al 2012 Iden-tification and characterization of novel genotoxic stress-in-ducible nuclear long noncoding RNAs in mammalian cellsPLoS One 7 e34949

Nakagawa S Ip JY Shioi G Tripathi V Zong X Hirose T Pra-santh KV 2012 Malat1 is not an essential component of nu-clear speckles in mice RNA 18 1487ndash1499

Pontier SM Muller WJ 2009 Integrins in mammary-stem-cellbiology and breast-cancer progressionmdasha role in cancer stemcells J Cell Sci 122 207ndash214

Prasanth KV Spector DL 2007 Eukaryotic regulatory RNAs ananswer to the lsquogenome complexityrsquo conundrum Genes Dev21 11ndash42

Prensner JR Chinnaiyan AM 2011 The emergence of lncRNAsin cancer biology Cancer Discov 1 91ndash407

Prensner JR IyerMK BalbinOADhanasekaran SMCaoQ Bren-ner JC Laxman B Asangani IA Grasso CS Kominsky HDet al 2011 Transcriptome sequencing across a prostate cancercohort identifies PCAT-1 an unannotated lincRNA implicat-ed in disease progression Nat Biotechnol 29 742ndash749

Prensner JR Iyer MK Sahu A Asangani IA Cao Q Patel L Ver-gara IA Davicioni E Erho N Ghadessi M et al 2013 Thelong noncoding RNA SChLAP1 promotes aggressive prostatecancer and antagonizes the SWISNF complexNat Genet 451392ndash1398

Rigo F Hua Y Krainer AR Bennett CF 2012 Antisense-basedtherapy for the treatment of spinal muscular atrophy J CellBiol 199 21ndash25

Rinn JL Chang HY 2012 Genome regulation by long noncodingRNAs Ann Rev Biochem 81 145ndash166

Robinson MD McCarthy DJ Smyth GK 2010 edgeR a Biocon-ductor package for differential expression analysis of digitalgene expression data Bioinformatics 26 139ndash140

Sachs N Clevers H 2014 Organoid cultures for the analysis ofcancer phenotypes Curr Opin Genet Dev 24 68ndash73

Sato T Clevers H 2013 Growing self-organizing mini-guts froma single intestinal stem cell mechanism and applications Sci-ence 340 1190ndash1194

Schnitt SJ 2010 Classification and prognosis of invasive breastcancer from morphology to molecular taxonomy ModernPathol 23 S60ndashS64

Semenza GL 2003 Targeting HIF-1 for cancer therapy Nat RevCancer 3 721ndash732

Spector DL Lamond AI 2011 Nuclear speckles Cold SpringHarbor Persp Biol 3 a000646

Streuli CH Bailey N Bissell MJ 1991 Control of mammaryepithelial differentiation basement membrane induces tis-sue-specific gene expression in the absence of cellndashcellinteraction and morphological polarity J Cell Biol 1151383ndash1395

Streuli CH Edwards GM Delcommenne M Whitelaw CB Bur-donTG Schindler CWatsonCJ 1995 Stat5 as a target for reg-ulation by extracellular matrix J Biol Chem 270 21639ndash21644

Subramanian A Tamayo P Mootha VK Mukherjee S Ebert BLGillette MA Paulovich A Pomeroy SL Golub TR LanderES et al 2005 Gene set enrichment analysis a knowledge-based approach for interpreting genome-wide expression pro-files Proc Natl Acad Sci 102 15545ndash15550

Tano K Mizuno R Okada T Rakwal R Shibato J Masuo YIjiri K Akimitsu N 2010 MALAT-1 enhances cell motilityof lung adenocarcinoma cells by influencing the expressionof motility-related genes FEBS Lett 584 4575ndash4580

TeplovaMMinasov G Tereshko V Inamati GB Cook PD Man-oharan M Egli M 1999 Crystal structure and improved anti-sense properties of 2prime-O-(2-methoxyethyl)-RNA Nat StructBiol 6 535ndash539

Tollervey JR Curk T Rogelj B Briese M Cereda M KayikciM Konig J Hortobagyi T Nishimura AL Zupunski V et al2011 Characterizing the RNA targets and position-dependentsplicing regulation by TDP-43 Nat Neuro 14 452ndash458

Tripathi V Ellis JD Shen Z SongDY PanQWatt AT Freier SMBennett CF Sharma A Bubulya PA et al 2010 The nuclear-retained noncoding RNA MALAT1 regulates alternativesplicing by modulating SR splicing factor phosphorylationMol Cell 39 925ndash938

Tseng YY Moriarity BS Gong W Akiyama R Tiwari A Kawa-kami H Ronning P Reuland B Guenther K Beadnell TC

Arun et al



et al 2014 PVT1 dependence in cancer withMYC copy-num-ber increase Nature 512 82ndash86

Venables JP Klinck R Bramard A Inkel L Dufresne-Martin GKoh C Gervais-Bird J Lapointe E Froehlich U Durand Met al 2008 Identification of alternative splicing markers forbreast cancer Cancer Res 68 9525ndash9531

WatsonCJ Campbell JJ 2009 Three-dimensional culturemodelsof mammary gland Organogenesis 5 43ndash49

West JA Davis CP SunwooH SimonMD Sadreyev RI Wang PITolstorukov MY Kingston RE 2014 The long noncodingRNAs NEAT1 and MALAT1 bind active chromatin sitesMol Cell 55 791ndash802

Wilusz JE Freier SM Spector DL 2008 3prime end processing of a longnuclear-retained noncoding RNA yields a tRNA-like cyto-plasmic RNA Cell 135 919ndash932

Wilusz JE Sunwoo H Spector DL 2009 Long noncoding RNAsfunctional surprises from the RNA world Gene Dev 231494ndash1504

Wu J Akerman M Sun S McCombie WR Krainer AR ZhangMQ 2011 SpliceTrap a method to quantify alternative splic-

ing under single cellular conditions Bioinformatics 273010ndash3016

Yamada K Kano J TsunodaH Yoshikawa H Okubo C IshiyamaT Noguchi M 2006 Phenotypic characterization of endome-trial stromal sarcoma of the uterus Cancer Sci 97 106ndash12

Yang L Lin C LiuW Zhang J Ohgi KA Grinstein JD DorresteinPC Rosenfeld MG 2011 ncRNA- and Pc2 methylation-de-pendent gene relocation between nuclear structures mediatesgene activation programs Cell 147 773ndash788

Ye S Li P Tong C Ying QL 2013 Embryonic stem cell self-re-newal pathways converge on the transcription factor Tfcp2l1EMBO J 32 2548ndash2560

Zhan L RosenbergA BergamiKC YuM XuanZ Jaffe AB AllredC Muthuswamy SK 2008 Deregulation of scribble promotesmammary tumorigenesis and reveals a role for cell polarity incarcinoma Cell 135 865ndash878

Zhang B Arun G Mao YS Lazar Z Hung G Bhattacharjee GXiao X Booth CJ Wu J Zhang C et al 2012 The lncRNAMalat1 is dispensable for mouse development but its tran-scription plays a cis-regulatory role in the adult Cell Rep 2111ndash123




101101gad270959115Access the most recent version at doi originally published online December 23 2015302016 Genes Dev

Gayatri Arun Sarah Diermeier Martin Akerman et al

lncRNA lossMalat1Differentiation of mammary tumors and reduction in metastasis upon

Material

Supplemental

httpgenesdevcshlporgcontentsuppl20151223gad270959115DC1

References

httpgenesdevcshlporgcontent30134fullhtmlref-list-1

This article cites 88 articles 27 of which can be accessed free at

License

Commons Creative

httpcreativecommonsorglicensesby-nc40at Creative Commons License (Attribution-NonCommercial 40 International) as described

) After six months it is available under ahttpgenesdevcshlporgsitemisctermsxhtmlsix months after the full-issue publication date (see This article is distributed exclusively by Cold Spring Harbor Laboratory Press for the first

ServiceEmail Alerting

click hereright corner of the article or

Receive free email alerts when new articles cite this article - sign up in the box at the top

copy 2016 Arun et al Published by Cold Spring Harbor Laboratory Press


and a 61-nt tRNA-like small RNA which localizes to thecytoplasm (Wilusz et al 2008) MALAT1 has been impli-cated in regulating alternative pre-mRNA splicing andits knockdown was shown to result in cell cycle arrest(Tripathi et al 2010) MALAT1 has also been shown tobe required to activate E2F target genes by repositioningthem from polycomb bodies to transcriptionally activenuclear sites in a serum-dependent manner (Yang et al2011) More recently two genome-wide studies have indi-cated that MALAT1 binds to the transcription start sites(TSSs) and gene bodies of actively transcribing genes to-gether with another lncRNA NEAT1 (West et al 2014)In addition a genome-wide RNAndashRNA binding studyhas shown that Malat1 can also bind to nascent pre-mRNAs indirectly via protein partners (Engreitz et al2014) Despite these functional insights derived from var-ious cell lines Malat1 knockout mice developed in ourlaboratory and others (Eissmann et al 2012 Nakagawaet al 2012 Zhang et al 2012) exhibit a normal phenotypeThe mice are born in expected Mendelian ratios and nomajor alterations have been observed in tissue histologygene expression or pre-mRNA splicingAlthough Malat1 knockout mice have no overt pheno-

type MALAT1 loss in a lung cancer xenograft model re-sulted in reduced homing of human lung cancer cells tothe lungs compared with control cells with wild-type lev-els ofMALAT1 (Gutschner et al 2013) and knockdown ofMALAT1 in lung cancer cells in vitro impaired cellularmotility (Tano et al 2010 Gutschner et al 2013) In addi-tion MALAT1 gene mutations have been found to fre-quently occur in luminal-type breast tumors (Ellis et al2012) Based on these findings we were interested in pur-suing whether Malat1 has a direct causative in vivo func-tion in breast cancer progressionHere we examined the role of Malat1 in mammary tu-

mor progression using the MMTV (mouse mammary tu-mor virus)-PyMT mouse mammary tumor model ofhuman luminal B breast cancer (Guy et al 1992 Linet al 2003 Herschkowitz et al 2007) We found thatgenetic knockout of Malat1 or knockdown upon subcu-taneous delivery of antisense oligonucleotides (ASOs)resulted in significant differentiation of the primary tu-mors and a reduction in lung metastases In additionknockdown ofMalat1 inMMTV-PyMT andHer2neu-am-plified mammary tumor organoids inhibited branchingmorphogenesis ex vivo Detailed molecular analyses byRNA sequencing (RNA-seq) revealed significant changesin gene expression and pre-mRNA splicing of protumori-genic and differentiation-related genes These findingsidentify Malat1 as a therapeutic target in breast cancerand the utility of Malat1 ASOs as a therapeutic for inhib-iting breast cancer progression

Results

Mammary tumor progression is impaired upongenetic loss of Malat1 lncRNA

Human luminal B breast cancer is one of the most preva-lent subtypes of breast cancer affecting nearly 15ndash20

of individuals diagnosed with this disease (for review seeSchnitt 2010 Ades et al 2014) In order to investigate thepotential in vivo role of Malat1 in mammary tumor pro-gression we used theMMTV-PyMTmouse mammary tu-mor model of human luminal B breast cancer (Guy et al1992 Lin et al 2003 Herschkowitz et al 2007) In thismodel the polyomamiddle-T antigen is under the controlof the MMTV promoter resulting in highly penetrantmammary tumors and lung metastasis (Lin et al 2003)Malat1 homozygous knockout (Malat1minusminus) femalemice

previously generated in our laboratory (Zhang et al 2012)were crossed to MMTV-PyMT males (C57Bl6) to obtainMMTV-PyMTMalat1++ PyMTMalat1+minus and PyMTMalat1minusminus mice Individuals of all three genotypes wereborn in expectedMendelian ratioswith a normal gestationperiod and postnatal development In accord with pub-lished data (Lin et al 2003) tumor onset in MMTV-PyMTmice began soon after puberty and palpable tumorsdeveloped within 12ndash16 wk of birth in gt90 of the miceTumor onset did not differ significantly in the presenceor absence ofMalat1 (Fig 1A) Prior to 5 mo of age nearly90of themice developed palpablemultifocal mammarygland lesions demonstrating that loss of Malat1 doesnot affect the initiation of PyMT-induced mammarygland tumors Although the tumors that developed inthe Malat1minusminus background initially exhibited a relativelyslower tumor growth rate compared with control mice(Fig 1B) this difference was not found to be significantespecially at the later time points Surprisingly theMMTV-PyMTMalat1minusminusmice developedhighly cystic liq-uid-filled tumors (Fig 1C right panel) in contrast to thecontrol mice which developed poorly differentiated solidcarcinomas (Fig 1C left panel) Hematoxylin and eosin(HampE) staining of tumor sections from MMTV-PyMTMalat1minusminus mice demonstrated that nearly 90 of the tu-mors thatweregt1cmindiameterexhibited acysticpheno-type with gt75 of the tumor area showing encapsulatedwell-differentiated cystic histopathology (Fig 1D bottompanels) in stark contrast to the poorly differentiated solidcarcinomas observed in the control tumors (Fig 1D toppanels)In the MMTV-PyMT model nearly 80 of the mice

develop lung micro- and macrometastases Thereforewe examined the impact of Malat1 loss on lung metasta-sesMMTV-PyMTMalat1minusminus mice exhibited a significantreduction in lung macrometastases where only three of12 mice had macro lesions as compared with seven ofeight mice in control MMTV-PyMTMalat1++ mice(Fig 1EF) The lungs were further examined for the inci-dence of micrometastases (Fig 1G) and the majority ofMMTV-PyMTMalat1minusminus animals did not showanymicro-metastasis (nine of 12) Next we calculated the total met-astatic burden as a percentage of lung volume (Fig 1H) andfound that the MMTV-PyMTMalat1minusminus animals showeda significantly decreased metastatic burden (3ndash5)as compared with the control MMTV-PyMTMalat1++

mice for which the metastatic burden reached up to40 of the total lung volume These results clearly dem-onstrate a role for the lncRNAMalat1 inmammary tumorprogression








Arun et al

















Arun et al
















Arun et al











Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative












Arun et al

















Arun et al
















Arun et al











Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative






















Arun et al
















Arun et al











Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative
















Arun et al
















Arun et al











Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative





















Arun et al











Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative














Arun et al











Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative
















Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative













Arun et al















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative




















Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative














Discussion




Arun et al


















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative























Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative
















Arun et al





Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative










Animals




Organoid culture



Electron microscopy




Histology





Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative








Immunofluorescence





Splicing analysis



RT-qPCR



Accession number


Acknowledgments


References








Arun et al













































































Arun et al






















Material

Supplemental


References



License

Commons Creative


















































































Arun et al






















Material

Supplemental


References



License

Commons Creative













































Arun et al






















Material

Supplemental


References



License

Commons Creative



























Material

Supplemental


References



License

Commons Creative











Material

Supplemental


References



License

Commons Creative








Documents

Differentiation of mammary tumors and reduction in ...genesdev.cshlp.org/content/30/1/34.full.pdfGayatri Arun,1 Sarah Diermeier,1 Martin Akerman,1 Kung-Chi Chang,1,2 J. Erby Wilkinson,3