tmem2 is a sox4-regulated gene that mediates … · progression as a result of increased cell...

Molecular and Cellular Pathobiology

TMEM2 Is a SOX4-Regulated Gene That MediatesMetastaticMigrationand Invasion inBreastCancerHyeseung Lee, Hani Goodarzi, Sohail F. Tavazoie, and Claudio R. Alarc�on

Abstract

The developmental transcription factor SOX4 contributes to themetastatic spread ofmultiple solid cancer types, but its direct targetgenes thatmediate cancer progression arenotwell defined.Using asystematic molecular and genomic approach, we identified theTMEM2 transmembrane protein gene as a direct transcriptionaltarget of SOX4. TMEM2was transcriptionally activated by SOX4 inbreast cancer cellswhere, like SOX4, TMEM2was found tomediateproinvasive and promigratory effects. Similarly, TMEM2 was suf-

ficient to promote metastatic colonization of breast cancer cellsand its expression in primary breast tumors associated with ahigher likelihood of metastatic relapse. Given earlier evidence thatgenetic inactivation of SOX4 or TMEM2 yield similar defects incardiac development, our findings lead us to propose that TMEM2may not only mediate the pathologic effects of SOX4 on cancerprogression but also potentially its contributions to embryonicdevelopment. Cancer Res; 76(17); 4994–5005. �2016 AACR.

IntroductionBreast cancer is the most common cancer in women and the

second most common cancer in the United States (The NationalCancer Institute Surveillance, Epidemiology, and End Resultsdatabase of 2015). Most breast cancer patients die from metas-tasis, the spread of cancer cells from a primary tumor to distalorgans (1). To study the biology of cancer metastasis, we andothers have previously in vivo selected for breast cancer cells withhigh metastatic capability (2). This system has been used toidentify molecular factors and key regulatory pathways thatmediatemetastasis in various cancer types (3–8). In breast cancer,molecular analysis of independently derived in vivo–selected cellpopulations led to the identification of miR-335 as one of threeendogenous human miRNAs that suppress breast cancer metas-tasis (9). In that study, SOX4 was identified as a miR-335–regulated gene. Reduction of miR-335 levels in breast cancer cellsled to increased expression of SOX4, resulting in metastaticprogression as a result of increased cell invasion and migration.Subsequent to its discovery as a mediator of breast cancer metas-tasis, SOX4 expression has been shown to be also increased inother cancer types, including leukemia (10, 11), glioblastoma(12, 13), medulloblastoma (14, 15), hepatocellular carcinoma(HCC; ref. 16), and prostate cancer (17–19).Despite this evidencesupporting the tumorigenic properties of SOX4, comparative

analyses of the SOX4 transcriptional network revealed a highdegree of variation between the target genes in different cancers,which suggest the tumor-specific and context-dependentmechan-isms of target gene selection (20). Recent studies have implicatedSOX4 as a regulator of TGF-b–induced epithelial–mesenchymaltransition in breast cancer cells (21, 22). However, the broadertranscriptional targets of SOX4 in breast cancer are poorly char-acterized. SOX4 transcriptional networks have also been studiedin other cancer types, and these studies have revealed that SOX4targets are involved in various oncogenic pathways (10, 11, 18). Inthis study, we used an unbiased genomic approach to identifyTMEM2 as a metastasis-promoting gene that is transcriptionallyregulated by SOX4 in breast cancer cells and mediates SOX4-driven metastatic progression in breast cancer. Furthermore, wefind that TMEM2 expression in breast cancer positively correlateswith reduced overall survival of patients. Our work implicatesTMEM2 as a direct transcriptional target and a mediator of SOX4in breast cancer progression.

Materials and MethodsAnalysis of mRNA expression by quantitative real-time PCR

Total RNA fromcancer cellswas extracted andpurifiedusing themirVana (Applied Biosystems) or Total RNA Purification Kits(NorgenBiotek).mRNAexpressionquantificationwas performedas described previously (7). One microgram of total RNA wasreverse transcribed using the cDNA First-Strand Synthesis Kit (LifeTechnologies) and diluted 1:5 with water. Real-time PCR reactionmix consisted of 5 mL of the diluted cDNA, 2.5 mL of 1 mmol/Lprimer mix, 17.5 mL of water, and 25 mL of FAST SYBR GreenMaster mix (Life Technologies, 4385612). Each reaction wasperformed in quadruplicate andmRNAexpressionwas quantifiedby quantitative real-time PCR (qRT-PCR) using an ABI Prism7900HT Real-Time PCR System (Applied Biosystems). 18S orGAPDH was used as an endogenous control for normalization.Primers are listed in Supplementary Table S1.

Animal studiesAll animal work was conducted in accordance with a protocol

approved by the Institutional Animal Care and Use Committee(IACUC) at The Rockefeller University (New York, NY). For lung

Laboratory of Systems Cancer Biology, Rockefeller University, NewYork, New York.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Current address of H. Goodarzi: Department of Biochemistry & Biophysics,Department of Urology, Helen Diller Family Comprehensive Cancer Center,University of California, San Francisco, San Francisco, CA.

CorrespondingAuthors: Sohail F. Tavazoie, Laboratory of SystemsCancer Biology,Rockefeller University, Box 16, 1230 York Avenue, NewYork, NY 10065. Phone: 212-327-7208; Fax: 212-327-7209; E-mail: [email protected]; and ClaudioR. Alarc�on, [email protected]

doi: 10.1158/0008-5472.CAN-15-2322

�2016 American Association for Cancer Research.

CancerResearch

Cancer Res; 76(17) September 1, 20164994

on September 27, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 21, 2016; DOI: 10.1158/0008-5472.CAN-15-2322

http://cancerres.aacrjournals.org/

colonization assays, 4 � 104 MDA-LM2 cells or 5 � 104 CN34-LM1A cells were resuspendedwith 100mLof PBS and injected intothe lateral tail vein of NOD-SCID gamma female mice (TheJackson Laboratory) age-matched between 6 and 8 weeks. Lungcolonization was monitored once every week using biolumines-cence. For orthotopic lung metastasis assays, 5 � 105 MDA-LM2cells were resuspended in a 1:1mixture of PBS and growth factor–reduced Matrigel (BD Biosciences), and bilaterally injected intothe mammary fat pads of NOD-SCID gamma female mice in atotal volume of 100 mL at day 0. Primary tumor growth rates weremonitored from day 14 to day 21, and tumor volume wascalculated on the basis of the formula: pLW2/6 (L, tumor length;W, width). At day 21, primary tumors were resected, and orotho-pic lung metastatic progression was monitored using biolumi-nescence. At day 50, lungs were extracted and fixed in 4% para-formaldehyde in PBS overnight at 4�C, underwent serial washeswith PBS, 50% ethanol, and 70% ethanol, for 30 minutes each atroom temperature, and were hematoxylin and eosin stained.Quantification of in vivo bioluminescence (the total photon fluxin the region of interest) was done in Living Image software (L.P.Larson Corporation).

Cell cultureThe MDA-MB-231 cell line was derived from the pleural effu-

sion of a breast cancer patient who developed distant metastasisyears after primary tumor resection, and its metastatic derivativesLM2 cellswere selected from theMDA-MB-231 cell population fortheir ability to metastasize to lungs after two rounds of in vivoselection (3). CN34 parental cells were obtained from the pleuralfluid of another patient with metastatic breast cancer, and itsmetastatic derivatives LM1A cells were isolated from the lungmetastatic nodules originating from the CN34 parental cells (9).These cell lines (MDA-parental, MDA-LM2, CN34-parental, andCN34-LM1A) were authenticated by RNA-seq profiling fromGSE45162 (23). Breast cancer cell lines, HEK293T, and 293LTVcells were propagated in vitro with DMEM supplemented with10% FBS, 1% penicillin–streptomycin, 2 mmol/L L-glutamine, 1mmol/L sodium pyruvate, and 2.5 mg/mL fungizone. Mycoplas-ma contamination was monitored periodically.

Chromatin immunoprecipitation and qPCRChromatin immunoprecipitation (ChIP) was performed

using EZ-Magna ChIP A/G Chromatin ImmunoprecipitationKit (Millipore, 17-10086) according to the manufacturer'sprotocol. A total of 2 � 107 LM2 cells were used per immu-noprecipitation and the nuclear pellet was sonicated threetimes for 15 seconds at an amplitude of 60 with 3-minutelapse between each pulse (Branson sonifier). The lysate wasincubated at 4�C for 4 hours with 5 mg of anti-SOX4 antibody(Abcam, ab70598) or control IgG. Real-time PCR reaction mixconsisted of 1 mL of the ChIP DNA, 2.5 mL of 1 mmol/L primermix, 21.5 mL of water, and 25 mL of FAST SYBR Green Mastermix (Life Technologies, 4385612). The optimized primers usedfor ChIP-qPCR are listed in Supplementary Table S1.

Clinical correlation analysisExpression Project for Oncology data. Expression levels of SOX4and the putative target genes were analyzed from a total of 236breast cancer samples obtained from the Expression Project forOncology (expO) microarray database (GSE2109).

Gene expression–based outcome for breast cancer online data.Assessment of gene expression levels and association with overallsurvival outcome for SOX4 and TMEM2 in breast cancer sub-groups (stage 0, I, II, III, IV) were done by gene expression–basedoutcome for breast cancer online (GOBO; http://co.bmc.lu.se/gobo/gobo.pl; ref. 24).

Kaplan–Meier plotter for breast cancer data. Evaluation of TMEM2in predicting survival of intrinsic subtypes (HER2 positive andluminal B) of breast cancer patients was done as described inref. 25 (http://kmplot.com/analysis/).

Generation of overexpressing and knockdown cell linesFor SOX4 overexpression studies, human cDNA of SOX4

with N-terminal FLAG tag was cloned into the pBabe-puroretroviral expression vector. Ten micrograms of DNA weretransfected into the H29 packaging cell line in a 10-cm plateby Lipofectamine 2000 (ratio of mg DNA: mL Lipofectamine2000 is 1:4). Virus-containing supernatant at 48 and 72 hoursafter transfection were harvested, filtered through a 0.45-mmsyringe filter, and added to breast cancer cells with polybrene (8mg/mL). The cells were selected with puromycin (1 mg/mL) for2–3 days. For doxycycline-inducible SOX4 overexpression stud-ies, FLAG-SOX4 was subcloned into pEN_Tmcs Entry vector(Addgene plasmid #25751) and then digested with BamHI andEcoRV to be ligated with pSLIK-Neo lentiviral vector (Addgeneplasmid #25735) that was digested with AleI and XhoI. Lenti-virus was generated by cotransfecting 2.5 mg of appropriatelentiviral plasmid with 7.5 mg of lentiviral packaging plasmids(5 mg of gag/pol and 2.5 mg of env) with Lipofectamine 2000into 293LTV cells in a 10-cm plate.

For TMEM2 overexpression studies, human cDNA of TMEM2with C-terminal V5 tag was obtained from the CCSB LentiviralExpression Library (OHS6087; ref. 26). The C-terminal frameshiftmutation present in the ORFeome library was corrected andcloned into the pLX304 lentiviral expression vector by Gatewaycloning.

For knockdown studies, Sigma shRNA glycerol stockswere purchased [TRCN0000018216, TRCN0000018217,TRCN0000148114,TRCN0000147636, control shRNA(SHC002)].

Luciferase reporter assayThe SOX4-bound regions of TMEM2 obtained from ChIP-

qPCR were cloned into the pGL3-Promoter vector (Promega,E1761) that contains the Firefly luciferase gene downstream ofa multiple cloning site, into which the promoter of interest isinserted. The three putative SOX4-bindingmotifs in this constructwere mutated using the Multi Site-Directed Mutagenesis Kit(Agilent, 210515). These vectors (25 ng) were cotransfected withthe Renilla luciferase vector (2.5 ng) into HEK293T cells eithertransiently overexpressing FLAG-tagged SOX4 (22.5 ng) or con-trol emptyplasmid (22.5ng). After 30hours post-transfection, thecells were processed with the Dual-Luciferase Reporter AssaySystem (Promega, E1910) according to the manufacturer's pro-tocol. Briefly, cells were lysed with 100 mL of Passive Lysis bufferfor 20 minutes at room temperature. Thirty microliters of thelysate were used per luciferase reaction. The Firefly and Renillaluciferase activities were measured in a luminometer (Perkin-Elmer EnVision plate reader) and transfection efficiency wasnormalized by the Renilla luciferase activity.

TMEM2, a Breast Cancer Prometastatic Gene Regulated by SOX4

www.aacrjournals.org Cancer Res; 76(17) September 1, 2016 4995




siRNA-mediated mRNA knockdownA total of 1 � 106 cells were seeded in a 10-cm plate. Next

day, siRNAs targeting either SOX4 (Thermo Scientific oligo IDROSJN-00001, ROSJN-00003) or TMEM2 [Integrated DNA Tech-nologies (IDT) siRNA identity HSC.RNAI.N013390.12.10, HSC.RNAI.N013390.12.2] and control siRNA (Thermo Scientific,siGENOME control siRNA #1) were transfected with Lipofec-tamine reagent according to the manufacturer's protocol. Forty-eight hours after transfection, the cells were collected andapplied to either transwell invasion assay or transwell migra-tion assay.

Statistical analysisUnless otherwise noted, all data are represented asmean� SEM

with P values: ��, P < 0.0001; ��, P < 0.0005; ��, P < 0.001; �, P <0.05. One-tailed Student t test was used and P < 0.05 wasconsidered to be statistically significant. For Kaplan–Meier plotanalysis, two-tailed log-rank t test was used. Statistical analyseswere conducted using PRISM version 6 for Mac (GraphPadsoftware).

Transcriptomic profiling analysisFor poorly and highly metastatic cells, RNA-seq data from

GSE45162 (23) were used in this study. For SOX4 microarrayanalysis, 1 mg of total RNA was extracted using the MiRvana kit(Ambion) from two independent RNAi-mediated knockdowns ofSOX4 and two control cells. Each sample was labeled and hybrid-ized on Illumina BeadChip array (GPL10558_HumanHT-12_V4)by theGenomics core facility at Rockefeller University (New York,NY). Each sample wasmedian-normalized and the average of twocontrol cells and the average of two SOX4 siRNAs were used tocalculate the relative fold change upon SOX4 knockdown bydividing a value from control cells over a value from SOX4knockdown cells for each gene. The data from SOX4 knockdownmicroarrays were deposited in Gene Expression Omnibus underthe accession number GSE79202. Heatmaps depicting rela-tive expression levels of genes from the microarray data werecreated with matrix2png (27).

Transwell invasion assayThe transwell invasion assays were performed as described

previously (9). Briefly, cells were incubated overnight in theculture media with 0.2 % FBS. The next day, the cells weretrypsinized, resuspended in the low serum media, and seededat 5 � 104 for MDA or 1 � 105 for CN34 per well into thegrowth factor–reduced Matrigel invasion chambers (8-mm poresize, BD Biosciences). After 18–22 hours, the chambers werewashed with PBS twice and the cells on the apical side of eachinserts were scraped off. The invaded cells fixed with 4%paraformaldehyde were stained with DAPI using VectaShield(Vector Laboratories) and were counted in four fields per insertand then quantified with ImageJ (NIH, Bethesda, MD).

Transwell migration assayTranswell migration assays were performed as described

previously (9) with minor modifications: overnight serum-starved cells (1 � 105) were seeded per cell-culture insert madeof Track-etched polyethylene terephthalate (PET) membraneswith pores (3-mm pore size, BD Biosciences). After 12 hours, themigrated cells were processed as described above for the trans-well invasion assays.

Western blottingCells were lysed with ice-cold RIPA buffer containing pro-

tease inhibitor cocktail (Roche) and sonicated. The clearedlysate by centrifugation was reduced and run on Bis-TrisNuPAGE gels in MOPS buffer (Life Technologies). Proteinswere transferred to polyvinylidene difluoride membrane. Thefollowing primary antibodies were used: 1:100 for custom-made rabbit polyclonal anti-Sox4, Yenzyme, immunogenCGR SPA DHR GYA SLR; 1:1,000 for rabbit polyclonalanti-Tmem2 (Abcam, ab98348); 1:4,000 for rabbit polyclonalanti-Gapdh (Sigma, G9545).

ResultsSystematic analysis of SOX4-regulated direct target genes inbreast cancer

To identify potential SOX4 target genes that could mediatebreast cancer metastasis, we employed RNA-sequencing andtranscriptomic profiling approaches. Endogenous SOX4expression becomes upregulated in highly lung-metastaticbreast cancer subpopulations (MDA-LM2, CN34-LM1A) thatwere derived through in vivo selection from poorly metastaticparental cell lines (MDA-parental, CN34-parental, respective-ly; ref. 9). Given this, and knowing that SOX4 is an establishedtranscriptional activator (28–30), we searched for genes forwhich expression levels are increased in the metastatic cellsubpopulations and were also reduced upon RNA interfer-ence–mediated SOX4 depletion. RNA-sequencing of MDA-parental/MDA-LM2 and CN34-parental/CN34-LM1A revealed100 genes that were significantly upregulated in both highlymetastatic sublines. To determine whether any of these genesare transcriptional targets of SOX4, we performed transcrip-tomic profiling of MDA-LM2 cells transfected with two inde-pendent siRNAs targeting SOX4 as well as a control siRNA.Analysis of this data revealed that a total of 1,356 genes weresignificantly downregulated upon SOX4 depletion in MDA-LM2 cells. Of the 1,356 genes downregulated by SOX4 knock-down, 26 were also upregulated in the two highly metastaticcell lines based on the above mentioned RNA-sequencinganalysis, resulting in a list of candidate SOX4-regulated genesin breast cancer cells (Fig. 1 and Supplementary Tables S2–S4).

We next asked whether the expression levels of any of these26 SOX4-regulated genes positively correlated with SOX4expression in breast cancer. To do this, we examined geneexpression levels in publicly available transcriptomic profilingdatasets [Oncomine Cancer Microarray database (31) andExpression Project for Oncology microarray database (expO;GSE2109)]. This analysis showed that 8 of these 26 candidateSOX4-regulated genes were overexpressed in multiple datasetscomprised of noncancerous and cancerous breast tissues whenSOX4 was overexpressed (Fig. 2A). Of these 8 genes, we found apositive and significant correlation between MMP1, SEMA4B,and TMEM2 levels and SOX4 levels in the expO microarraydatabase (Fig. 2B). To find SOX4 targets that would mediatemetastatic progression of breast cancer, we next examined thetranscript levels of SOX4 and these three genes in patients withhigh-risk primary breast tumors to evaluate the prognosticvalue of these putative targets. Of the three genes, only TMEM2expression significantly correlates with reduced overall survival(OS) in high-risk grade 3 breast tumors as shown by theKaplan–Meier Plotter software (Fig. 3A). Moreover, combined

Lee et al.

Cancer Res; 76(17) September 1, 2016 Cancer Research4996




expression levels of SOX4 and TMEM2 performed as a superiorprognostic factor relative to TMEM2 alone (Fig. 3B). To furtherinvestigate whether TMEM2 might be associated with anyintrinsic subtype of breast cancer, we tested the prognosticvalues of either TMEM2 alone or TMEM2 with SOX4 in intrinsicsubtypes of breast cancer. TMEM2 expression associated withpoor prognosis in luminal B and HER2-positive subgroups. InHER2-positive subgroup, high expression of TMEM2 alone

most significantly correlated with poor distant metastasis–freesurvival (DMFS, Fig. 3C) and OS (Fig. 3D). In relapse-freesurvival (RFS) of HER2-positive group and OS of the luminalB group, expression of TMEM2 alone did not show a statisti-cally significant correlation; however, in both cases, the com-bination of TMEM2 with SOX4 significantly correlated withpoorer outcomes compared with low expression of both genes(Fig. 3E–H). These analyses reveal that TMEM2 is a SOX4-

MDA-p

aren

tal

MDA-L

M2

21Relativeexpression

CN34-p

aren

tal

CN34-L

M1A



siCon

trol(a

vg)

siSOX4(

avg)

SOX4

AKAP13

ALDH3A2

CADM1

DUSP4

EREG

FLRT2

FSCN1

FZD8

IGF2BP3

ITGB4

KLF9

MMP1

ODZ3

PCDH20

PHGDH

PHLDA2

PPP2R2C

SASH1

SEL1L3

SEMA4B

SERPINB5

SERPINE2

SLCO4A1

ST8SIA4

TMEM154

TMEM2

MDA-LM2BA C

Figure 1.

A systematic strategy for identification of SOX4 target genesthat regulate breast cancer metastasis. Heatmap depictingrelative transcript levels of the 26 putative SOX4 targetgenes that were identified by the overlap of genes whoseexpression was upregulated in MDA-LM2 relative toMDA-parental cells (A; q < 0.05), upregulated in CN34-LM1Arelative to CN34-parental cells (B; q < 0.05), and decreasedupon SOX4 depletion using two independent siRNAs relativeto control siRNA (C; P < 0.05).






regulated gene, for which expression positively associates withreduced overall survival in breast cancer and suggest thatTMEM2 may serve as prognostic factor for HER2-positive sub-types of breast cancer and potentially luminal B breast cancer.Importantly, using the Cancer Tissue Atlas, we observed thatTMEM2 is expressed at the protein level in breast cancer tissuesas well as in almost every cancer type analyzed; and multiplecancer types including hepatocellular carcinomas and a fewpancreatic, colorectal and gastric cancer samples showed strongpositive staining for the antibody targeting TMEM2 (Supple-mentary Fig. S1). Therefore, we sought to further investigate therole of SOX4 as a transcriptional activator of TMEM2.

SOX4 regulates TMEM2 transcriptionNext, using qRT-PCR, we validated the upregulation of SOX4

and TMEM2 transcript levels in highly metastatic MDA-LM2 cellsrelative to the poorly metastatic MDA-parental cells (Supplemen-tary Fig. S2A and S2B). The relative expression levels of TMEM2and SOX4 were also assessed by qRT-PCR in the CN34-parentalcell line, an independentmalignant breast cancer cell population,and its highlymetastatic lungderivative, theCN34-LM1Acell line.(Supplementary Fig. S2C and S2D). In both cell lines, SOX4 andTMEM2were upregulated in highlymetastatic sublines relative topoorly metastatic parental cell lines. Depletion of SOX4 by twoindependent SOX4-targeting siRNAs in MDA-LM2 cells also sig-nificantly reduced TMEM2 transcript levels (Supplementary Fig.S2E and S2F). In addition, SOX4 depletion in CN34-LM1A cells

by RNAi also significantly reduced TMEM2 transcript levels (Sup-plementary Fig. S2G and S2H). The efficiency of knockdown ofSOX4 by siRNAs was measured by Western blot analysis (Sup-plementary Fig. S2I).

We next questioned whether SOX4 overexpression is sufficientto increase TMEM2 transcript levels. MDA-parental cells over-expressing Flag-tagged SOX4 displayed increased TMEM2 tran-script levels relative to control cells (Fig. 4A and B). To determinethe temporal regulation of SOX4-mediated TMEM2 expression,we generated doxycycline-inducible Flag-tagged SOX4-expressingMDA-parental and CN34-parental cell lines. In both cell lines,TMEM2 mRNA levels were increased at least at 24 hours afterinducing SOX4overexpression (Fig. 4C–F). Thesefindings suggestthat TMEM2 transcript levels are regulated by SOX4 in breastcancer cells.

We next investigated whether TMEM2 is a direct transcriptionaltarget of SOX4 in breast cancer cells. To do this, we assessed SOX4binding to the TMEM2 promoter using ChIP-qPCR. We designedqPCR primers spanning from 4.8 kilobase (kb) pairs upstream ofthe TMEM2 transcription start site (TSS) to 2.2 kb pairs down-stream of the TSS. This experiment revealed that a genomic DNAregion proximal to the TMEM2 TSS and spanning base pairs 856–1178 upstream of the TSS, was highly enriched in SOX4-boundgenomic DNA relative to that bound by IgG control (Fig. 5A). Thefunctional consequence of SOX4 binding to this TMEM2 pro-moter region was further investigated using a luciferase promoterassay. To do this, we generated reporter constructs consisting of a

expO Data (243 tumors)

Correlationwith SOX4

Spearman RP

(two-tailed)

MMP1 0.2625 < 1E–04

SEMA4B 0.2331 2E–04

TMEM2 0.2126 9E–04

A

B

Tissue type

Gene Median rank P 1 2 3 4 5 6 7 8 9

SOX4 2109.0 3.59E–11

ST8SIA4 4981.0 2.18E–08

TMEM2 5637.0 3.47E–07

AKAP13 6635 1.46E–05

SEMA4B 4805.0 2.63E–04

MMP1 1199 1E–03

PPP2R2C 3434.0 6E–03

IGF2BP3 8523.0 1.6E–02

PHLDA2 4463.0 4.9E–02

151025%

Most expressed

1

2

3

4

5

6

7

8

Breast Adenocarcinoma Type: Ductal Breast Carcinoma

Intraductal Cribriform Breast Adenocarcinoma vs. Normal

Invasive Breast Carcinoma vs. Normal

Invasive Ductal and Lobular Carcinoma vs. Normal

Invasive Ductal Breast Carcinoma vs. Normal

Invasive Lobular Breast Carcinoma vs. Normal

Male Breast Carcinoma vs. Normal

Mixed Lobular and Ductal Breast Carcinoma vs. Normal

9 Mucinous Breast Carcinoma vs. Normal

251051

Least expressed

Figure 2.

Clinical relevance of putative SOX4target genes in breast tumors. A,mRNA expression levels of SOX4 andits target genes for which positivecorrelation is observed in breastcarcinoma and normal tissue. Thevalues were obtained from OncomineCancer Microarray database (www.oncomine.org). The color of each boxrepresents the expression level foreachgene asdepicted at the bottomofthe table. The median rank for eachgene across each of the analyses andtheir P value are shown in the table.B, the expression correlation of SOX4with its target genes MMP1, SEMA4B,and TMEM2 as analyzed using theexpO breast cancer database (www.intgen.org), which consists of a total of243 tumors from four patients instage 0, 38 patients in stage I, 129patients in stage II, 67 patients in stageIII, and 5 patients in stage IV.Bonferroni correction was used toadjust confidence intervals (P < 0.02).

Lee et al.





luciferase reporter cloned downstream of the indicated TMEM2promoter region (Fig. 5B). SOX4 overexpression increased lucif-erase activity relative to control, consistent with a SOX4-depen-dent increase in transcription from the TMEM2 promoter (Fig.5C). The SOX transcription factor family is known to preferen-tially bind to the hexameric core sequence WWCAAW, where W

indicates A or T (32–34). To refine the TMEM2 promoter regionsnecessary for the observed SOX4-dependent increase in transcrip-tion, we searched for and identified three instances of theWWCAAW motif in the TMEM2 promoter region. These core-binding motifs were then mutated, either individually or com-bined, and these constructs were tested in luciferase promoter

HG

BA

500 100 150 200

LowHigh

P = 1.4E–02

N = 111DM

FS

(%

)

Time (months)

TMEM2

HER2+

0

20

40

60

80

100

500 100 150 200

OS

(%

)

Time (months)

TMEM2

HER2+

0

20

40

60

80

100

P = 7.6E–03

N = 89LowHigh

500 100 150 200

RF

S (

%)

Time (months)

TMEM2

HER2+

0

20

40

60

80

100

P = 2E–01

N = 208LowHigh

500 100 150 200

RF

S (

%)

Time (months)

SOX4+TMEM2

HER2+

0

20

40

60

80

100

P = 1.8E–02

N = 208LowHigh

P = 6.1E–02

N = 320LowHigh

Time (months)

TMEM2

Luminal B

500 100 150 200

OS

(%

)

0

20

40

60

80

100

P = 3.3E–03

N = 320Low

High

Time (months)

SOX4+TMEM2

Luminal B

500 100 150 200

OS

(%

)

0

20

40

60

80

100

DC

FE

6420 108

0

20

40

60

80

100 TMEM2

Grade 3

P = 3.40E–02

N = 262

LowMediumHigh

6420 108

0

20

40

60

80

100 SOX4+TMEM2

Grade 3

P = 4.84E–03

N = 262

LowMediumHigh

OS

(%

)

Time (years)

OS

(%

)

Time (years)

Figure 3.

Prognostic value of TMEM2 on breast cancersurvival.A andB,Kaplan–Meier overall survival(OS) curves depicting the segregation ofpatients diagnosed with grade 3 tumors (n ¼262) expressing different levels of TMEM2 (A),and both SOX4 and TMEM2 (B). Data wereobtained from GOBO. Red line, patients withhigh gene expression; yellow line, patientswith intermediate; gray line, patients withlow gene expression. TMEM2_low (log2expression �2.531 to �0.243); medium (log2expression �0.24 to 0.245); high (log2expression 0.25 to 3.357); SOX4 þTMEM2_low (log2 expression �3.111to �0.2); medium (log2 expression �0.199 to0.215); high (log2 expression 0.22 to 2.196).C–H, Kaplan–Meier plots for DMFS (C), OS (D),RFS (E) for TMEM2, and both SOX4 andTMEM2 (F) in HER2þ subtype are shown.G–H, Kaplan–Meier OS plots for TMEM2 (G)and both SOX4 and TMEM2 (H) in luminal Bsubtype are shown.






reporter assays. Upon SOX4 overexpression, cells transfected witha reporter containing the wild-type TMEM2 promoter regionsexhibited increased expression, while reporters containing muta-tions in the SOX4-binding motif did not exhibit enhancement ofSOX4-dependent expression (Fig. 5C). The requirement for theregions of all three SOX4-binding sites suggests that cooperativemultisite binding by this transcription factor is required fortranscriptional activation. Given that pre-mRNAs are transientintermediates and their steady-state levels reflect endogenoustranscription prior to nuclear processing (35–39), we usedqRT-PCR to analyze the levels of intron-containing TMEM2pre-mRNA in control and SOX4-depleted cells. SOX4 knockdown

significantly reduced TMEM2 pre-mRNA levels, consistent witha transcriptional role for SOX4 in TMEM2 regulation (Fig. 5Dand E). Collectively, these findings support a model wherebySOX4 transcriptionally promotes TMEM2 expression throughdirect interaction with the promoter region of TMEM2.

TMEM2 promotes invasion, migration, and breast cancermetastasis

Highly metastatic MDA-LM2 cells express higher levels ofSOX4 relative to MDA-parental cells, and their ability to invadeand metastasize to the lungs of mice is reduced upon SOX4depletion (9). To determine whether an increase in the expres-sion of SOX4 is sufficient to enhance the invasive capacity ofbreast cancer cells, we performed Matrigel invasion assays inMDA-parental and CN34-parental breast cancer cell lines engi-neered to overexpress SOX4. SOX4 overexpression significantlyincreased the invasive capacity of breast cancer cells relative tocontrol cells (Fig. 6A and B). Next, to investigate whether theenhanced cancer cell invasion capacity observed upon SOX4overexpression was mediated by TMEM2, we used shRNAs toknockdown TMEM2 levels in SOX4-overexpressing cells. Twoindependent shRNAs targeting TMEM2 were validated by West-ern blot analysis (Supplementary Fig. S3). Depletion of TMEM2in SOX4-overexpressing cells significantly reduced the invasioncapability of these cells (Fig. 6C). Furthermore, transient over-expression of SOX4 in CN34-parental cells increased theirinvasion capacity, an effect that was abrogated upon TMEM2depletion (Fig. 6D–F). This suggests that the increased invasioncapacity conferred by SOX4 overexpression is dependent, atleast in part, on its downstream target gene, TMEM2. Similarly,significantly greater invasion and migration capacity wasobserved in MDA-parental cells overexpressing TMEM2 relativeto control cells (Fig. 6G and H). Consistent with this, areduction in cell invasion capacity was observed upon TMEM2depletion in MDA-LM2 cells (Fig. 6I and J). These findings areconsistent with a model where the proinvasive effects of SOX4are mediated, at least in part, through transcriptional inductionof TMEM2.

Next, we asked whether TMEM2 can promote in vivo breastcancer metastatic colonization in a similar manner as SOX4. Todo this, we assayed the metastatic lung colonization capacity ofMDA-LM2 cells with stable TMEM2 knockdown or withTMEM2 overexpression. Depletion of TMEM2 in highly meta-static MDA-LM2 cells significantly reduced metastatic coloni-zation (Fig. 7A and B). Consistently, knockdown of TMEM2 byan independent shRNA in CN34-LM1A cells also significantlydecreased the ability of cells to colonize the lungs (Supple-mentary Fig. S4A and S4B). Conversely, TMEM2 overexpressionsignificantly enhanced the metastatic capacity of MDA-LM2cells (Fig. 7C and D). To study the effect of TMEM2 in primarytumor growth as well as in metastatic potential, we conductedorthotopic injection of MDA-LM2 cells with stable TMEM2knockdown in mammary fat pads of the mice. In both in vitroproliferation assays (Fig. 7E) and in vivo primary tumor growthassays (Fig. 7F), depletion of TMEM2 did not have an effect onneither cell proliferation nor primary tumor growth relative tocontrol. However, TMEM2 depletion caused a decrease in lungmetastasis development after primary tumor resection (Fig. 7Gand H). Therefore, TMEM2 acts as a promoter of lung meta-static colonization in breast cancer cells and can promote spon-taneous breast cancer metastasis.

0

10

20

30

40

Rel

ativ

e ex

pres

sion

MDA-parentalSOX4

C

MDA-parentalSOX4

A

0

1

2

3

Rel

ativ

e ex

pres

sion

MDA-parentalTMEM2

D

MDA-parentalTMEM2

B

0

1

2

3

4

5

Rel

ativ

e ex

pres

sion

CN34-parentalTMEM2

Rel

ativ

e ex

pres

sion

CN34-parental FE

Contro

l

SOX4 OE

0

20

40

60

80

Rel

ativ

e ex

pres

sion

***

Contro

l

SOX4 OE

0.0

0.5

1.0

1.5

Rel

ativ

e ex

pres

sion

***

+ Dox 240 48 72 (h)

+ Dox 240 48 72 (h)36

+ Dox 240 48 72 (h)

0

10

20

30SOX4

+ Dox 240 48 72 (h)36

Figure 4.

SOX4 is necessary and sufficient for regulation of TMEM2. Induction of TMEM2transcript levels by SOX4 overexpression was measured by qRT-PCR. A and B,relative mRNA levels for SOX4 (A) and TMEM2 (B) in control or SOX4-overexpressing MDA-parental cells (n¼ 3). C and D, time course of doxycycline(1 mg/mL) treatment in doxycycline-inducible SOX4-overexpressingMDA-parental cells depicting relative mRNA levels for SOX4 (C, n ¼ 2),and TMEM2 (D, n¼ 2). E and F, same as C and D in CN34-parental cells (n ¼ 2).Error bars, SEM. �� , P < 0.0005.

Lee et al.





DiscussionSOX4 belongs to the HMG box superfamily of DNA-binding

proteins and contains a single HMG domain through which itbinds DNA in a sequence-specificmanner (34, 40–43). SOX4wasdiscovered as a transcriptional activator in B and T lymphocytes(40, 44) and early genetic studies revealed that SOX4 is requiredfor normal cardiac development (44–46).

SOX4 has also been found to be overexpressed in variouscancers (30, 47, 48) and amplified in lung cancer (49). Theidentification of miR-335 as a suppressor of breast cancer metas-tasis led to the identificationof a cancer progression role for SOX4,as a driver of cancer cell invasion andmigration. SOX4was shownto be downstream of the metastasis suppressor miRNA miR-335.As a result of miR-335 silencing in metastatic breast cancer, SOX4expression is enhanced (9). The enhancement of invasive andmigratory phenotypes causedby SOX4upregulation increased themetastatic lung colonization capacity of breast cancer cells. Otherstudies have shown a role for SOX4 in promoting cancer pro-

gression in a number of additional cancer types, including pros-tate cancer (17–19), hepatocellular carcinoma (16), glioblastoma(12, 13), and leukemia (10, 11).

Using a genomic and molecular approach, we have identifiedTMEM2 as a direct transcriptional target of SOX4 in breast cancercells. We find that SOX4 directly binds to the promoter region ofTMEM2 and transcriptionally activates it in a sequence-specificmanner. Molecular studies reveal that SOX4 also regulatesTMEM2 pre-mRNA levels, consistent with SOX4 transcriptionalactivation of this gene. TMEM2 promotes breast cancer cellmigration, invasion, and metastatic colonization, phenocopyingSOX4-dependent effects. Moreover, we find that expression ofTMEM2, a gene that has not been previously implicated in cancerprogression, correlates with breast cancer clinical outcome.

TMEM2 is a poorly described type II transmembrane proteinthat is composed of a N-terminal transmembrane domain and aC-terminal extracellular portion containing a G8 and a GGdomain. To date, no biochemical functions have been identifiedfor these domains (50, 51). Recent studies have revealed a role for

siCon

trol

si 1SOX4

si 2SOX4

0.0

0.5

1.0

0.0

0.5

1.0

Rel

ativ

e ex

pres

sion

SOX4

****

****

siCon

trol

si 1SOX4

si 2SOX4

Rel

ativ

e pr

e-m

RN

Aex

pres

sion

TMEM2

****

***

C

ED

B

A

–1391 –1087

WT

B3 B2 B1TMEM2

tttgtttacaaattcaaattcCCC tacCCC GGGgtt

B123M

B1M

B2M

B3MLuci

fera

se r

epor

ters

Contro

lW

T

B123M B1M B2M B3M

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e lu

cife

rase

act

ivity

FLA

G-S

OX

4/C

ontr

ol

******* *** **** ****

#1 #2 #3 #4 #5 #6 #7 #8 #9 #10

#11

#12

#13

#14

#15

#16

#17

#18

#19

#20

#21

0

100

200

300

400

500

Rel

ativ

e S

OX

4 IP

/IgG

ChIP-qPCR for TMEM2 (–4.8kb to +2.2kb)

*

–1010 bp

TSS

Figure 5.

SOX4 transcriptionally regulatesTMEM2. A, ChIP-qPCR assessment ofthe TMEM2 promoter regionimmunoprecipitated with an anti-SOX4antibody or control IgG. Pairs of primerswere designed to cover a region from�4.8 kb to þ2.2 kb from the TSS.Sequences upstream of the TSS aremarkedas (�)and thosedownstreamofthe TSS aremarked as (þ). Experimentswere performed on MDA-LM2 cells anddata from biological triplicates areshown. Data represented as foldenrichment over SOX4 relative to anegative control IgG. B, a schematicdiagram depicting promoter region ofTMEM2 amplified by qPCR primer #15used inA. TMEM2 promoter region from�1438 to �881 base pairs upstream ofthe TMEM2 TSS have three putativeSOX4-binding motifs marked by B1, B2,and B3. C, functional validation of theChIP-qPCR by luciferase reporter assay.TMEM2 promoter region containing allthree putative SOX4 motifs (WT) orsequences having mutations in all oreach of the three motifs shown in Bwere cloned upstream of a luciferasereporter. The relative luciferase activityof theWT sequence compared with thecontrol empty vector sequence wasmeasured in the context of transientFLAG-SOX4 overexpression inHEK293T cells. Experiments wereperformed in triplicates in two differentbiological replicates.DandE,pre-mRNAlevels of TMEM2 in the MDA-LM2 cellsexpressing either siSOX4 or siControlwere assessed by qRT-PCR, withprimers detecting intronic regions closeto 30UTR of TMEM2. Data fromquadruplicates in two differentbiological replicates are shown. Errorbars, SEM. � , P < 0.05; �� , P < 0.0005;�� , P < 0.0001.






Figure 6.

TMEM2 mediates the effects of SOX4 on cell invasion and migration. A and B, invasion assays by 5 � 104 MDA-parental cells expressing either a control vector orFLAG-SOX4 (A, n¼ 6) and by 1� 105 CN34-parental cells expressing either control or FLAG-SOX4 (B, n¼ 5). Images of DAPI-stained cells that invaded the Matrigelwere obtained at �10 magnifications and quantified by ImageJ. C, Matrigel invasion by MDA-parental cells after knockdown of TMEM2 in the context of constitutiveexpression of FLAG-SOX4 (n ¼ 4–5). D, Matrigel invasion by knockdown of TMEM2 in the context of doxycycline-inducible FLAG-SOX4 overexpression inCN34-parental cells (n ¼ 3–4). E and F, expression of SOX4 (E) and TMEM2 (F) in CN34-parental cells expressing doxycycline-inducible FLAG-SOX4 48 hoursafter doxycycline induction (n¼ 2).G,MDA-parental cells overexpressing TMEM2 showed enhancedMatrigel invasion comparedwith control cells (n¼ 4).H, increasedtranswell migration by MDA-parental cells overexpressing TMEM2 compared with the control cells (n ¼ 4). I and J, reduced Matrigel invasion (I) and transwellmigration (J) by knockdown of TMEM2 in MDA-LM2 cells (n ¼ 4). Error bars, SEM. � , P < 0.05; �� , P < 0.001; �� , P < 0.0005; �� , P < 0.0001.

Lee et al.





tmem2 in zebrafish heart development (52, 53). Of the twodifferent zebrafish mutants of tmem2 identified, one displayeda specific loss of cardiac looping at late developmental stages,

while the other exhibited aberrant atrioventricular canal differ-entiation. Both mutations result in a shortened Tmem2 extracel-lular domain, suggesting the missing extracellular domain is

100 20 301 100

1 101

1 102

1 103

1 104

Days post-injection

Nor

mal

ized

lung

pho

ton

flux

(Log

) shCTRL

100 20 301 101

1 102

1 103

1 104

1 105

Days post-injection

Nor

mal

ized

lung

pho

ton

flux

(Log

) ControlTMEM2 OE

Contro

l

TMEM

2 OE

0

1

2

3

4

Rel

ativ

e ex

pres

sion

MDA-LM2

shCTRL

0.0

0.5

1.0

Rel

ativ

e ex

pres

sion

MDA-LM2

shC

TR

Lsh

TM

EM

2T

ME

M2

OE

Con

trol

5.0

4.0

3.0

2.0

1.0

106X

1.0

0.8

0.6

0.4

0.2

108X

BA

DC

25 30 35 40 45 50 550

2 109

4 109

6 109

8 109

1 1010

Days post-injection

Lung

pho

ton

flux

shCTRL

sh1T

MEM

2

sh2T

MEM

20

50

100

150

Met

asta

tic n

odul

es

GE

12 14 16 18 20 220

100

200

300

400

Days post-injection

Tum

or v

olum

e (m

m3) shCTRL

sh1TMEM2

sh2TMEM2

0

2

4

6

8

Cel

l cou

nt (

×10

5)

HF

shCTRLsh1TMEM2

sh2TMEM2

shCTRL

sh1T

MEM

2

sh2T

MEM

2

shTMEM2

shTM

EM2

Figure 7.

TMEM2 promotes in vivometastatic lung colonization. A, bioluminescence imaging plot of lung metastasis by 4 � 104 MDA-LM2 cells expressing control shRNA orshRNA targeting TMEM2 (n¼ 5). NOD-SCID gamma female mice were imaged every seven days and bioluminescence signals were normalized to the signal on theday of tail-vein injection (day 0). B, knockdown efficiency of TMEM2 measured by qRT-PCR in MDA-LM2 cells injected into the tail-vein of mice. C and D,bioluminescence imaging plot of lung metastasis by 4 � 104 MDA-LM2 cells expressing control or TMEM2 (C, n ¼ 5) and mRNA levels of TMEM2 measured byqRT-PCR in MDA-LM2 cells injected into the tail-vein of NOD-SCID gamma female mice (D). Error bars, SEM. �� , P < 0.001; �� , P < 0.0005; �� , P < 0.0001.E, in vitro cell proliferation by 5 � 104 MDA-LM2 cells expressing either shControl or shTMEM2. Cells were seeded in triplicate at day 0. Dead cells wereexcluded by Trypan blue and live cells were counted at day 5. Data from biological triplicates are shown (n ¼ 3). F, in vivo primary tumor growth by 5 � 105

MDA-LM2 cells expressing either shControl or shTMEM2. Cells were bilaterally injected into the mammary fat pads of NOD-SCID gamma female mice atday 0. Primary tumor growth was monitored from day 14 to day 21 (control, n ¼ 5; TMEM2 KD #1, n ¼ 4; TMEM2 KD #2, n ¼ 4). G and H, at day 21, tumors wereresected and orthotopic metastasis was monitored. Lungs were harvested on day 50 (G), hematoxylin–eosin stained, and the number of nodules perlungs was counted (H). � , P < 0.05 obtained using two-tailed Mann–Whitney test. Error bars, SEM.






important for Tmem2 function during cardiac development. Thiszebrafish Tmem2-mutant phenotype could be rescued by knock-down of Bmp4, suggesting that Tmem2 regulates zebrafish heartdevelopment through inhibition of Bmp4. Interestingly, the heartdevelopmental phenotype observed in zebrafish Tmem2mutantsresembles theheart developmental phenotype, a commonarterialtrunk defect, that is observed in Sox4-null mice (44–46). In themyocardium, Tmem2 promotes themigration of myocardial andendocardial cells. It is interesting that Tmem2 functions in heartdevelopment in part by regulating cell migration, which resem-bles the role of SOX4 in both mouse heart development andbreast cancer metastasis. Because of these similarities, it would beinteresting to investigate whether TMEM2 regulates cell migrationand invasion by engaging BMP signaling pathways.

While the regulatory relationship between SOX4 and TMEM2was previously unknown, the regulation of TMEM2 by SOX4 inbreast cancer cells, the similarity of the SOX4 and TMEM2 phe-notypes in cancer, and the common phenotypes resulting fromthe genetic inactivation of these genes, suggest that SOX4 andTMEM2may form a evolutionarily conserved pathway that couldgovern normal cardiovascular development as well as cancerprogression. Further studies are needed to determine whetherSOX4 regulates TMEM2 expression in the developing heart and, ifso, whether enforced TMEM2 expression could rescue the SOX4loss-of-function phenotype.

While we have identified TMEM2 as one transcriptionaltarget of SOX4, our findings do not exclude the possibility ofadditional SOX4 transcriptional targets that mediate theSOX4-dependent proinvasive and promigratory phenotypes.In addition, it will be interesting to investigate the transcrip-tional and genomic landscape of SOX4 binding during devel-opment to establish the breadth of its developmental tran-scriptional program.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: H. Lee, S.F. Tavazoie, C.R. Alarc�onDevelopment of methodology: H. Lee, C.R. Alarc�onAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): H. LeeAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): H. Lee, H. Goodarzi, C.R. Alarc�onWriting, review, and/or revision of the manuscript: H. Lee, H. Goodarzi,S.F. Tavazoie, C.R. Alarc�onAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): H. Lee, S.F. TavazoieStudy supervision: S.F. Tavazoie, C.R. Alarc�on

AcknowledgmentsWe thankmembers of our laboratory; N. Halberg andH. Nguyen for helping

with tail-vein injections, J. Loo, L. Fish, and A. Nguyen for their helpfulcomments on this manuscript. We are grateful to S. Simon, D. Mucida, andS. Acharyya for valuable advice and intellectual suggestions. We thank C. Zhao,C. Lai, and N. Nnatubeugo of the Rockefeller Genomics Resource Center forassistance with transcriptomic profiling.

Grant SupportThis research was supported by Department of Defense Era of Hope Scholar

Award. C.R. Alarc�on was previously supported by the Anderson Cancer CenterFellowship and is currently supported by the Robert S. Bennett PostdoctoralFellowship. S.F. Tavazoie is a Department of Defense Breast Cancer Collabo-rative Scholars and Innovators Award recipient, and aBCRF andPershing SquareFoundation awardee.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 21, 2015; revised May 25, 2016; accepted June 14, 2016;published OnlineFirst June 21, 2016.

References1. Gupta GP, Massagu�e J. Cancer metastasis: building a framework. Cell

2006;127:679–95.2. Fidler IJ. Selection of successive tumour lines for metastasis. Nat New Biol

1973;242:148–9.3. Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, Giri DD, et al. Genes that

mediate breast cancer metastasis to lung. Nature 2005;436:518–24.4. Gupta GP,Minn AJ, Kang Y, Siegel PM, Serganova I, Cordon-Cardo C, et al.

Identifying site-specific metastasis genes and functions. Cold Spring HarbSymp Quant Biol 2005;70:149–58.

5. Acharyya S, Oskarsson T, Vanharanta S, Malladi S, Kim J,Morris PG, et al. ACXCL1 paracrine network links cancer chemoresistance and metastasis.Cell 2012;150:165–78.

6. Pencheva N, Tran H, Buss C, Huh D, Drobnjak M, Busam K, et al.Convergent multi-miRNA targeting of ApoE drives LRP1/LRP8-dependentmelanoma metastasis and angiogenesis. Cell 2012;151:1068–82.

7. Png KJ, Halberg N, Yoshida M, Tavazoie SF. A microRNA regulon thatmediates endothelial recruitment and metastasis by cancer cells. Nature2011;481:190–4.

8. Loo JM, Scherl A, NguyenA,Man FY,Weinberg E, Zeng Z, et al. Extracellularmetabolic energetics can promote cancer progression. Cell 2015;160:393–406.

9. Tavazoie SF, Alarc�on C, Oskarsson T, Padua D, Wang Q, Bos PD, et al.Endogenous human microRNAs that suppress breast cancer metastasis.Nature 2008;451:147–52.

10. ZhangH, Alberich-JordaM, AmabileG, YangH, Staber PB,Diruscio A, et al.Sox4 is a key oncogenic target in C/EBPalpha mutant acute myeloidleukemia. Cancer Cell 2013;24:575–88.

11. Ramezani-Rad P, GengH, Hurtz C, Chan LN, Chen Z, JumaaH, et al. SOX4enables oncogenic survival signals in acute lymphoblastic leukemia. Blood2013;121:148–55.

12. Ikushima H, Todo T, Ino Y, Takahashi M, Miyazawa K, Miyazono K.Autocrine TGF-beta signaling maintains tumorigenicity of glioma-initiat-ing cells through Sry-related HMG-box factors. Cell Stem Cell 2009;5:504–14.

13. Lin B, Madan A, Yoon J-G, Fang X, Yan X, Kim T-K, et al. Massivelyparallel signature sequencing and bioinformatics analysis identifies up-regulation of TGFBI and SOX4 in human glioblastoma. PLoS One2010;5:e10210.

14. Lee CJ, Appleby VJ, Orme AT, ChanWI, Scotting PJ. Differential expressionof SOX4 and SOX11 in medulloblastoma. J Neurooncol 2002;57:201–14.

15. de Bont JM, Kros JM, Passier MMCJ, Reddingius RE, Sillevis Smitt PAE,Luider TM, et al. Differential expression and prognostic significance of SOXgenes in pediatric medulloblastoma and ependymoma identified bymicroarray analysis. Neuro Oncol 2008;10:648–60.

16. Liao YL, SunYM,ChauGY,ChauYP, Lai TC,Wang JL, et al. Identification ofSOX4 target genes using phylogenetic footprinting-based prediction fromexpression microarrays suggests that overexpression of SOX4 potentiatesmetastasis in hepatocellular carcinoma. Oncogene 2008;27:5578–89.

17. Liu P, Ramachandran S, Ali Seyed M, Scharer CD, Laycock N, Dalton WB,et al. Sex-determining region Y box 4 is a transforming oncogene in humanprostate cancer cells. Cancer Res 2006;66:4011–9.

18. Scharer CD, McCabe CD, Ali Seyed M, Berger MF, Bulyk ML, Moreno CS.Genome-wide promoter analysis of the SOX4 transcriptional network inprostate cancer cells. Cancer Res 2009;69:709–17.


Lee et al.




19. Lai YH, Cheng J, ChengD, FeaselME, Beste KD, Peng J, et al. SOX4 interactswith plakoglobin in a Wnt3a-dependent manner in prostate cancer cells.BMC Cell Biol 2011;12:50.

20. Vervoort SJ, van Boxtel R, Coffer PJ. The role of SRY-related HMG boxtranscription factor 4 (SOX4) in tumorigenesis and metastasis: friend orfoe? Oncogene 2013;32:3397–409.

21. Zhang J, Liang Q, Lei Y, Yao M, Li L, Gao X, et al. SOX4 induces epithelial-mesenchymal transition and contributes to breast cancer progression.Cancer Res 2012;72:4597–608.

22. Tiwari N, Tiwari VK, Waldmeier L, Balwierz PJ, Arnold P, Pachkov M, et al.Sox4 is a master regulator of epithelial-mesenchymal transition by con-trolling Ezh2 expression and epigenetic reprogramming. Cancer Cell2013;23:768–83.

23. Furlow PW, Zhang S, Soong TD, Halberg N, Goodarzi H,MangrumC, et al.Mechanosensitive pannexin-1 channels mediate microvascular metastaticcell survival. Nat Cell Biol 2015;17:943–52.

24. Ringn�erM, FredlundE,H€akkinen J, BorgÅ, Staaf J.GOBO: gene expression-based outcome for breast cancer online. PLoS One 2011;6:e17911–1.

25. Gyorffy B, Lanczky A, Eklund AC, Denkert C, Budczies J, Li Q, et al. Anonline survival analysis tool to rapidly assess the effect of 22,277 genes onbreast cancer prognosis using microarray data of 1,809 patients. BreastCancer Res Treat 2010;123:725–31.

26. Yang X, Yang X, Boehm JS, Yang X, Salehi-Ashtiani K, Hao T, et al. A publicgenome-scale lentiviral expression library of human ORFs. Nat Methods2011;8:659–61.

27. Pavlidis P, Noble WS. Matrix2png: a utility for visualizing matrix data.Bioinformatics 2003;19:295–6.

28. Wotton D, Lake RA, Farr CJ, Owen MJ. The high mobility group transcrip-tion factor, SOX4, transactivates the human CD2 enhancer. J Biol Chem1995;270:7515–22.

29. Castillo SD, Matheu A, Mariani N, Carretero J, Lopez-Rios F, Lovell-BadgeR, et al. Novel transcriptional targets of the SRY-HMG box transcriptionfactor SOX4 link its expression to the development of small cell lungcancer. Cancer Res 2012;72:176–86.

30. Huang HY, Cheng YY, Liao WC, Tien YW, Yang CH, Hsu SM, et al. SOX4transcriptionally regulates multiple SEMA3/plexin family members andpromotes tumor growth in pancreatic cancer. PLoS One 2012;7:e48637.

31. Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, Ghosh D, et al.ONCOMINE: a cancer microarray database and integrated data-miningplatform. Neoplasia 2004;6:1–6.

32. Harley VR, Lovellbadge R,GoodfellowPN.Definitionof a consensusDNA-binding site for SRY. Nucleic Acids Res 1994;22:1500–1.

33. Mertin S, McDowall SG, Harley VR. The DNA-binding specificity of SOX9and other SOX proteins. Nucleic Acids Res 1999;27:1359–64.

34. Lefebvre V, Dumitriu B, Penzo-M�endez A, Han Y, Pallavi B. Control of cellfate and differentiation by Sry-related high-mobility-group box (Sox)transcription factors. Int J Biochem Cell Biol 2007;39:2195–214.

35. Lipson KE, Baserga R. Transcriptional activity of the human thymidinekinase gene determined by a method using the polymerase chain-reactionand an intron-specific probe. Proc Natl Acad Sci U S A 1989;86:9774–7.

36. Luco RF, Misteli T. More than a splicing code: integrating the role of RNA,chromatin and non-coding RNA in alternative splicing regulation. CurrOpin Genet Dev 2011;21:366–72.

37. de Almeida SF, Carmo-Fonseca M. Design principles of interconnectionsbetween chromatin and pre-mRNA splicing. Trends Biochem Sci 2012;37:248–53.

38. Montes M, Becerra S, S�anchez-�Alvarez M, Su~n�e C. Functional coupling oftranscription and splicing. Gene 2012;501:104–17.

39. Shukla S, Oberdoerffer S. Co-transcriptional regulation of alternative pre-mRNA splicing. Biochim Biophys Acta 2012;1819:673–83.

40. Van deWeteringM,Oosterwegel M, Vannorren K, Clevers H. Sox-4, an Sry-like HMGbox protein, is a transcriptional activator in lymphocytes. EMBOJ 1993;12:3847–54.

41. Bergsland M, Werme M, Malewicz M, Perlmann T, Muhr J. The establish-ment of neuronal properties is controlled by Sox4 and Sox11. Genes Dev2006;20:3475–86.

42. Dy P, Penzo-M�endez A, Wang H, Pedraza CE, Macklin WB, Lefebvre V. Thethree SoxC proteins–Sox4, Sox11 and Sox12–exhibit overlapping expres-sion patterns and molecular properties. Nucleic Acids Res 2008;36:3101–17.

43. Jolma A, Yan J, Whitington T, Toivonen J, Nitta KR, Rastas P, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327–39.

44. Schilham MW, Oosterwegel MA, Moerer P, Ya J, de Boer PA, van deWetering M, et al. Defects in cardiac outflow tract formation and pro-B-lymphocyte expansion in mice lacking Sox-4. Nature 1996;380:711–4.

45. Ya J, Schilham MW, de Boer PA, Moorman AF, Clevers H, Lamers WH.Sox4-deficiency syndrome in mice is an animal model for common trunk.Circ Res 1998;83:986–94.

46. Penzo-Mendez A,DyP, Pallavi B, Lefebvre V.Generation ofmice harboringa Sox4 conditional null allele. Genesis 2007;45:776–80.

47. Frierson HF, El-Naggar AK, Welsh JB, Sapinoso LM, Su AI, Cheng J,et al. Large scale molecular analysis identifies genes with alteredexpression in salivary adenoid cystic carcinoma. Am J Pathol 2002;161:1315–23.

48. Sinner D, Kordich JJ, Spence JR, Opoka R, Rankin S, Lin SC, et al. Sox17and Sox4 differentially regulate beta-catenin/T-cell factor activityand proliferation of colon carcinoma cells. Mol Cell Biol 2007;27:7802–15.

49. Medina PP, Castillo SD, Blanco S, Sanz-Garcia M, Largo C, Alvarez S, et al.The SRY-HMG box gene, SOX4, is a target of gene amplification atchromosome 6p in lung cancer. Hum Mol Genet 2009;18:1343–52.

50. He QY, Liu XH, Li Q, Studholme DJ, Li XW, Liang SP. G8: a novel domainassociated with polycystic kidney disease and non-syndromic hearing loss.Bioinformatics (Oxford, England) 2006;22:2189–91.

51. Guo J, Cheng H, Zhao S, Yu L. GG: a domain involved in phage LTFapparatus and implicated in humanMEB and non-syndromic hearing lossdiseases. FEBS Lett 2006;580:581–4.

52. Smith KA, Lagendijk AK, Courtney AD, Chen H, Paterson S, Hogan BM,et al. Transmembrane protein 2 (Tmem2) is required to regionally restrictatrioventricular canal boundary and endocardial cushion development.Development 2011;138:4193–8.

53. Totong R, Schell T, Lescroart F, Ryckebusch L, Lin YF, Zygmunt T, et al. Thenovel transmembrane protein Tmem2 is essential for coordination ofmyocardial and endocardial morphogenesis. Development 2011;138:4199–205.






2016;76:4994-5005. Published OnlineFirst June 21, 2016.Cancer Res Hyeseung Lee, Hani Goodarzi, Sohail F. Tavazoie, et al. Migration and Invasion in Breast CancerTMEM2 Is a SOX4-Regulated Gene That Mediates Metastatic

Updated version

10.1158/0008-5472.CAN-15-2322doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2016/06/21/0008-5472.CAN-15-2322.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/76/17/4994.full#ref-list-1

This article cites 53 articles, 13 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/76/17/4994.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/76/17/4994To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-15-2322

http://cancerres.aacrjournals.org/content/suppl/2016/06/21/0008-5472.CAN-15-2322.DC1

http://cancerres.aacrjournals.org/content/76/17/4994.full#ref-list-1

http://cancerres.aacrjournals.org/content/76/17/4994.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/76/17/4994


tmem2 is a sox4-regulated gene that mediates … · progression as a result of increased cell...

Documents