Downregulation of RBMS3 Is Associated with Poor Prognosis ... · Priority Report Downregulation of RBMS3 Is Associated with Poor Prognosis in Esophageal Squamous Cell Carcinoma Yan

Priority Report

Downregulation of RBMS3 Is Associated with PoorPrognosis in Esophageal Squamous Cell Carcinoma

Yan Li1, Leilei Chen2, Chang-jun Nie1, Ting-ting Zeng1, Haibo Liu1, Xueying Mao1, Yanru Qin3,Ying-Hui Zhu1, Li Fu2, and Xin-Yuan Guan1,2

AbstractDeletions on chromosome 3p occur often in many solid tumors, including esophageal squamous cell

carcinoma (ESCC), suggesting the existence at this location of one or more tumor suppressor genes (TSG).In this study, we characterized RBMS3 gene encoding an RNA-binding protein as a candidate TSG located at3p24. Downregulation of RBMS3mRNA and protein levels was documented in approximately 50% of the primaryESCCs examined. Clinical association studies determined that RBMS3 downregulation was associated withpoor clinical outcomes. RBMS3 expression effectively suppressed the tumorigenicity of ESCC cells in vitro andin vivo, including by inhibition of cell growth rate, foci formation, soft agar colony formation, and tumorformation in nude mice. Molecular analyses revealed that RBMS3 downregulated c-Myc and CDK4, leading tosubsequent inhibition of Rb phosphorylation. Together, our findings suggest a tumor suppression function forthe human RBMS3 gene in ESCC, acting through c-Myc downregulation, with genetic loss of this gene in ESCCcontributing to poor outcomes in this deadly disease. Cancer Res; 71(19); 6106–15. �2011 AACR.

Introduction

Esophageal cancer is one of the most common fatal cancersworldwide (1). It ranks as the ninth most frequent cancer inthe world, with a remarkable geographic distribution betweenand within countries. High-risk areas include northern Iran,South Africa, and northern China, where esophageal squa-mous cell carcinoma (ESCC) is still the main cancer burdenand the fourth most common cause of death (2). Althoughepidemiologic studies indicate that tobacco and alcohol con-sumption are the major risk factors for squamous esophagealcancer in the low-risk regions of Europe and North America,the etiologic agents in high-risk regions have yet to beconvincingly identified. Previous studies in high-risk regionsshowed a strong tendency toward familial aggregation (3, 4),suggesting that genetic susceptibility may play a role in thepathogenesis of ESCC.

Chromosomal deletion is a major cause of the inactivationof tumor suppressor genes (TSG). Loss of 3p is one of the mostfrequent genetic alterations in ESCC detected by comparative

genomic hybridization and LOH (5–8), suggesting the exis-tence of one or more TSGs at frequently deleted regions. In ourrecent study, single-nucleotide polymorphism (SNP) massarray was applied to investigate LOH on 3p in 100 primaryESCC cases (9). Four commonly deleted regions (CDR) at3p26.3, 3p22, 3p21.3, and 3p14.2 were identified. Two candi-date TSGs within these CDRs, including PLCD1 (10) and PCAF(11), have been investigated. In this study, another candidateTSG, RBMS3 at 3p24, was investigated. RBMS3 was firstidentified by screening a human fibroblast cDNA librarywith a labeled DNA fragment derived from a2(I) collagenpromoter Box 5A (12). The protein contains 2 RNA-bindingdomains and it can bind strongly to synthetic poly-U andpoly-A oligoribonucleotides, suggesting that it is an RNA-binding protein (12). RBMS3 belongs to the family of c-Mycgene single-strand binding proteins (MSSP), which has2 additional members, RBMS1 and RBMS2 (13, 14). RBMS1was reported to be involved in binding to c-Myc protein (14),but no such function has been described for RBMS3. Itwas reported that RBMS3 expressed in activated hepaticstellate cells and increased expression of transcription factorPrx1 (15). Nevertheless, there was no report investigatingthe functions of RBMS3 in tumor development and/orprogression.

In this study, expression of RBMS3 in primary ESCCs andESCC cell lines was investigated by real time-PCR and immu-nohistochemistry (IHC) staining. To examine the possibletumor-suppressive activity of RBMS3 in ESCC, RBMS3 wastransfected into 2 ESCC cell lines, KYSE30 and KYSE180. Thetumor-suppressive effects of RBMS3 were characterized byboth in vitro and in vivo assays. The tumor-suppressingmechanisms of RBMS3, as well as its clinical significance,were also investigated.

Authors' Affiliations: 1State Key Laboratory of Oncology in SouthernChina, Cancer Center, Sun Yat-sen University, Guangzhou; 2Departmentsof Clinical Oncology, The University of Hong Kong, Pokfulam, Hong Kong;and 3Department of Clinical Oncology, The First Affiliated Hospital,Zhengzhou University, Zhengzhou, China

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

Corresponding Author: Xin-Yuan Guan, State Key Laboratory of Oncol-ogy in Southern China, Cancer Center, Sun Yat-sen University, Room 605,651 E. Dongfeng Road, Guangzhou 510060, China. Phone: 86-20-87343166; Fax: 852-2816-9126; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-10-4291

�2011 American Association for Cancer Research.

CancerResearch

Cancer Res; 71(19) October 1, 20116106

on March 1, 2021. © 2011 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 15, 2011; DOI: 10.1158/0008-5472.CAN-10-4291

http://cancerres.aacrjournals.org/

Materials and Methods

Cell lines and primary tumor specimensThree esophageal cancer cell lines—KYSE30, KYSE140, and

KYSE180—were obtained from DSMZ, the German ResourceCenter for Biological Material (16, 17) in 2006. Chinese ESCCcell lines EC18, EC109, and HKESC1 were kindly provided byour colleagues at the University of Hong Kong (ProfessorsG. Srivastava and G.S. Tsao; ref. 18). The cells were confirmedby cytogenetics as being of human origin in 2009. Forty pairs ofprimary ESCCs and adjacent nontumorous tissues were col-lected immediately after surgical resection at Linzhou CancerHospital (Henan, China). Tissue samples used in the studywere approved by the Committees for Ethical Review ofResearch Involving Human Subjects at Zhengzhou University(Zhengzhou, China).

SNP detection in tissue samplesGenomic DNA was extracted from tumor samples using

TIANamp genomic DNA Kit (Tiangen) according to the man-ufacturer's instructions. SNP site (rs987693) within RBMS3gene was PCR amplified and sequencing analyzed withdesigned primers (Supplementary Table S1).

5-AZA-20-deoxycytidine treatmentTo study whether demethylation could restore RBMS3

expression in ESCC cell lines, 3 � 105 cells were treated withthe DNA demethylating agent 5-AZA (Sigma-Aldrich) at 0.1, 1,10, 50, and 100 mmol/L for 3 days. Drugs and culture mediumwere refreshed every day during the treatment.

Establishment of RBMS3-expressing cell linesOmicslink-RBMS3 was purchased from GeneCopoeia Inc.

and subcloned into pcDNA3.1þ (Invitrogen) according to themanufacturer's protocol. RBMS3 was then stably transfectedinto ESCC cell line KYSE30 (RBMS3-30) using Lipofectamine2000 (Invitrogen). Blank vector pcDNA3.1þ transfected cells(Vec-30) were set as controls. RBMS3 was also transientlytransfected into ESCC cell lines KYSE140, KYSE180, andHKESC1 cells.

Quantitative real-time PCRTotal RNA from cultured cells and tumor tissues was

extracted using Trizol (Invitrogen) according to the man-ufacturer's instructions. Reverse transcription of total RNAwas done using SuperScript Kit (Invitrogen), and cDNA wassubjected to quantitative reverse transcriptase PCR (qRT-PCR) for RBMS3 with 2 pairs of primers (SupplementaryTable S1). Glyceraldehyde-3-phosphate dehydrogenase(GAPDH) or 18s rRNA was used as internal control. qRT-PCR was done using the SYBR Green Supermix andABI7900HT Fast Real-Time PCR system (Applied Biosys-tems). The assays were done in triplicate and values werenormalized using the internal control. Quantitative datawere exported and RBMS3 expression was normalized byinternal control. PCR products were subjected to dissocia-tion curve analysis using the light cycler system to excludeamplification of nonspecific products. Quantitative RT-PCR

data were processed using the DCT method as describedpreviously (19).

Tissue microarray and immunohistochemistryArchives of paraffin blocks from 183 patients with ESCC

were obtained from Linzhou Cancer Hospital (Henan, China).None of the patients in this study had received follow-upradiation or chemotherapy. The age of these ESCC patientsranged from 40 to 80 years at the time of surgery (median age,59 years) and the male:female ratio was 1.3:1. The tissuemicroarray (TMA) blocks were constructed according to amethod described previously (20). Multiple sections (5-mmthick) were used for IHC. In brief, paraffin-embedded TMAsections were deparaffinized and blocked with goat serumfor 30 minutes at room temperature, followed by incubationwith Mouse anti-RBMS3 (1:40; Abnova) overnight at 4�C.After incubation with horseradish peroxidase–linked sec-ondary antibody (Real Envision Detection Kit; Gene Tech)for 30 minutes, TMA sections were counterstained withMayer's hematoxylin. The degree of immunostaining wasviewed and scored separately by 2 independent investigatorswithout any prior knowledge of clinicopathologic data.Expression of RBMS3 was scored as absent (total absenceof staining), very weak (faint staining in <25% of tumor cells),moderate (moderate staining in �25% to <75%, or strongstaining in <25% of tumor cells), and strong (moderatestaining in �75%, or strong staining in �25% of tumorcells). In this study, moderate/strong nuclear staining wasdefined as positive staining and absent/very weak staining asdownregulation.

Tumor-suppressive function of RBMS3To test the tumor-suppressive function of RBMS3, XTT

assay, foci formation assay, and colony formation in soft agarwere done on cells that overexpressed RBMS3 and vectorcontrol cells. Cell proliferation rates of RBMS3-overexpressingcell colonies and vector control cells were compared byXTT assay. Cells were seeded into 96-well plates at a densityof 2 � 103 cells per well. The cell growth rate was detectedusing CCK-8 Kit (Tojindo) according to the manufacturer'sinstructions. Three independent assays were conducted. Forfoci formation assay, 2 � 103 cells were plated into 6-wellplates. Two weeks later, the surviving colonies were fixed,stained with crystal violet, and counted. The result wasrepresented as means � SD of 3 independent assays. Colonyformation in soft agar was done by growing 2 � 103 cells in0.4% bacto-agar on a bottom layer of solidified 0.6% bacto-agarin 6-well plates. After 2 weeks, colonies consisting of morethan 50 cells were counted and the result was represented asmeans � SD of triplicate independent assays.

In vivo tumor formation assayThe in vivo tumor-suppressive effect of RBMS3 was in-

vestigated by xenograft mouse experiments as previouslydescribed (10). Animal experiments were done in compliancewith the guidelines for the Welfare of ExperimentalAnimals in Sun Yat-sen University. RBMS3-30 and Vec-30cells (2 � 106) were injected s.c. into left and right flanks of

RBMS3 in ESCC

www.aacrjournals.org Cancer Res; 71(19) October 1, 2011 6107




4-week-old male nude BALB/c mice (n¼ 8), respectively. Micewere then examined for tumor growth by measuring tumorsize with calipers every 3 days for 20 days. Tumor volume wascalculated (V ¼ 0.5 � L � W2). Following sacrifice, tumorswere excised, fixed in 10% formalin, and embedded in paraffinblock for IHC study.

Western blot analysisWestern blot analysis was done according to the standard

protocol with antibodies for RBMS3 (Abnova), GAPDH,

c-Myc, Phospho-Rb (Ser780, Ser795, and Ser807/811), CDK2,CDK4, cyclin D1, and cyclin E (Cell Signaling Technology).

Chromatin immunoprecipitation PCRThe chromatin immunoprecipitation (ChIP) experiment

was conducted using an EZ-Magna ChIP G Kit (UpstateBiotechnology) according to the manufacturer's instructions.RBMS3-binding DNA fragments were pulled down by an anti-RBMS3 antibody (Abnova) or pooled IgG from mouse (SantaCruz Biotechnology) as a negative control. The primers used

Figure 1. Downregulation ofRBMS3 in ESCCs. A, expressionof RBMS3 in 40 primary ESCCcaseswas compared by qRT-PCRbetween tumor tissues and theirpaired nontumorous tissues. Theaverage relative expression inthese tumor tissues (6.02 � 4.01)was significantly lower than that inpaired nontumorous tissues(24.33 � 10.85; P < 0.05). GAPDHwas set as internal control. B, theexpression of RBMS3 in 6 ESCCcell lines was compared withpooled samples from5 nontumorous esophagealtissues (normal) by RT-PCR (top).Expression level was quantitatedusing ImageJ software (WayneRashband; bottom). C,representative samples ofRBMS3 expression (brown colorstaining in nucleus) were detectedby IHC in a pair of ESCC tumortissues and paired nontumorousesophageal tissues (�200magnification). D, representativechromas of rs987693 in tumortissues and paired nontumortissues. The SNP site (rs987693)is indicated by red arrows.

Li et al.

Cancer Res; 71(19) October 1, 2011 Cancer Research6108




for PCR amplification of the precipitated DNA fragments arelisted in Supplementary Table S1.

Electrophoretic mobility shift assayBiotin end-labeled probe (Supplementary Table S1) was

synthesized by Invitrogen. Nuclear extract was extracted usingNucBuster Extraction Kit (Novagen). Electrophoretic mobilityshift assay (EMSA) was done as the procedure of LightShiftChemiluminescentEMSAKit (ThermoScientific). Briefly, biotin-labeled probe (100 fM or 200 fM) and 15-mg nuclear extract were

incubated for 20 minutes at room temperature. Free probe wasseparated fromDNA–protein complexes by electrophoresis on anative 6% polyacrilamide gel in �0.5 TBE buffer. After electro-phoresis, the DNA was transferred to positively charged nylonmembrane, cross-linked, and detected by chemiluminescence.

RNA interferenceRBMS3 expression was silenced by double-stranded siRNA

targeting RBMS3 (siRNA-RBMS3; Supplementary Table S1)and scramble siRNA, which were obtained from Ambion'spredesigned siRNA database (Ambion, Inc.). RBMS3-30 cellswere transfected with siRNA using Lipofectamine 2000 (Invi-trogen) according to the manufacturer's instructions. Genesilencing effect was evaluated by qRT-PCR 48 hours aftersiRNA treatment (21). Scramble siRNA was used as negativecontrol. The effect of RNA interference (RNAi) was evaluatedby foci formation assay 48 hours after siRNA treatment.

Statistical analysisStatistical calculations were carried out with the SPSS

statistical software package (Version 16.0; SPSS, Inc.). Alldata were expressed as means � SD. The Pearson c2 test wasused to analyze the relationship between RBMS3 expressionand clinicopathologic features. Survival curves were gener-ated according to the Kaplan–Meier method and the sta-tistical analysis was done by log-rank test. Student t test wasused to analyze data from cell growth, foci formation, softagar assays, and tumor formation in nude mice. The value ofP < 0.05 was considered statistically significant.

Results

RBMS3 is frequently downregulated in ESCCTo determine the expression pattern of RBMS3 in ESCC

tissues, qRT-PCR was done in 40 pairs of tumor tissues andtheir corresponding nontumorous esophageal tissues. RBMS3expression was normalized by internal control GAPDH. Theresult showed that downregulation of RBMS3 was detectedin 18 of 40 (45%) of ESCC tissues compared with their pairednontumorous tissues (Fig. 1A). The expression of RBMS3in 6 ESCC cell lines was also tested by RT-PCR, and the resultshowed that RBMS3 was downregulated in 3 cell lines(HKESC1, KYSE30, and KYSE140; Fig. 1B). Expression ofRBMS3 protein was also compared between tumor andpaired nontumor samples using a TMA containing 183 pairsof ESCCs by IHC. Informative TMA cases were observed in 161normal tissues and 119 tumor cases. The noninformativesamples included lost samples and samples with too few cellsor with inappropriate staining. Moderate and strong nuclearstaining of RBMS3 was detected in 115 of 161 (71.43%)informative normal tissues. Absent and very weak expressionof RBMS3 was detected in 63 of 119 (52.9%) tumor tissues(Fig. 1C).

Downregulation of RBMS3 is associated with DNA copynumber loss

Because no CpG island was found in the promoter region ofRBMS3, demethylation agent 5-Aza-20-deoxycytidine was used

Table 1. Association of RBMS3 downregulationwith clinicopathologic characteristics of ESCCpatients

Informativecases

RBMS3downregulation

Pa

Age 0.115�57 y 58 35 (60.3%)>57 y 61 28 (45.9%)

Gender 0.845Male 52 27 (51.9%)Female 67 36 (53.7%)

Differentiation 0.4951 17 7 (41.2%)2 78 44 (56.4%)3 24 12 (50%)

T status 0.435T1–2 55 27 (49.1%)T3–4 64 36 (56.2%)

N status 0.567N0 103 55 (53.4%)N1 16 8 (50%)

aPearson c2.

Figure 2. Kaplan–Meier analysis shows that expression of RBMS3 wassignificantly associated with better overall survival in 119 ESCC cases(P ¼ 0.034, log-rank test).

RBMS3 in ESCC





to treat 3 ESCC cells (HKESC1, KYSE30, and KYSE140) withdownregulated RBMS3 to test whether promoter methylationaffects the expression of RBMS3. The result showed thatdemethylation treatment could not restore the expressionof RBMS3 (data not shown). In our previous study, LOH atthe SNP site rs987693 within the RBMS3 gene was detected in31 of 51 (61%) of primary ESCCs (9). LOH at rs987693 site wastested in the cohort of 40 ESCC samples used in this study.LOH was detected in 11 of 17 (64.7%) of the informative cases(Fig. 1D). Downregulation of RBMS3 was observed in 8 of 11(72.7%) of the ESCCs with LOH at RBMS3 site, which wassignificantly higher than those cases without LOH (1/6, 16.7%,P ¼ 0.0498).

Clinical significance of RBMS3 downregulationTo examine the clinical significance of RBMS3 down-

regulation in ESCC, the correlation of RBMS3 downregula-tion with clinicopathologic features was investigated. Theassociation study showed that downregulation of RBMS3was not significantly associated with age, gender, tumordifferentiation, or TNM stage of the patient (Table 1). In-

terestingly, Kaplan–Meier analysis revealed that the down-regulation of RBMS3 was significantly (log-rank test, P ¼0.034) correlated with poorer outcome of patients with ESCC(Fig. 2).

RBMS3 has tumor-suppressive abilityTo investigate whether RBMS3 has tumor-suppressive

ability, RBMS3 was stably transfected into KYSE30 cells(RBMS3-30) and transiently transfected into KYSE180 cells(RBMS3-180). Empty vector–transfected cells were used ascontrol (Vec-30 or Vec-180). Expression of RBMS3 in RBMS3-30 and RBMS3-180 cells was confirmed by RT-PCR (Fig. 3Aand B). The tumor-suppressive function of RBMS3was testedby both in vitro and in vivo assays. Cell growth assay showedthat the growth rates were significantly decreased in RBMS3-30 colonies (P < 0.05, Student t test) and RBMS3-180 cells(P < 0.05, Student t test), respectively, compared with controlcells (Fig. 3C). Soft agar assay showed that RBMS3 couldsignificantly inhibit colony formation in soft agar in RBMS3-30 (P < 0.01, Student t test) and RBMS3-180 cells(P < 0.05, Student t test) compared with control cells

Figure 3. A, expression of RBMS3in RBMS3-transfected KYSE30cells was confirmed by RT-PCRand qRT-PCR. Empty vector–transfected cells were used ascontrol. B, expression of RBMS3in RBMS3-transfected KYSE180cells was confirmed by RT-PCR.Empty vector–transfected cellswere used as control. C, XTTassay was used to compare cellgrowth curves between Vec-30and RBMS3-30 cells (left) andVec-180 and RBMS3-180 cells(right). The cell growth rate wassignificantly decreased in RBMS3overexpressed cells(*, P < 0.05; **, P < 0.01). Values areexpressed as mean � SD of 3independent experiments. D, thefrequency of colony formation insoft agar (left, representative;right, summary) was significantlylower in RBMS3 overexpressedcells (RBMS3-30 and RBMS3-180) compared with vector controlcells (*, P < 0.05; **, P < 0.01). Theresults are expressed as mean �SD of 3 independent experiments.

Li et al.





(Fig. 3D). Similarly, RBMS3 could also significantly inhibitfoci formation ability in RBMS3-30 (P < 0.01, Student t test)and RBMS3-180 cells (P < 0.05, Student t test) comparedwith control cells (Fig. 4A).The tumor-suppressive potential of RBMS3 was also

evaluated by xenograft tumor formation in athymic nudemice. Subcutaneous visible tumor was observed in the rightflank (Vec-30) in all 8 tested animals on day 8 afterinjection. However, visible tumor in the left flank(RBMS3-30) was only observed in 4 nude mice on day 8.Xenograft tumor growth curve showed that tumor inducedby RBMS3-30 cells grew much more slowly than did the Vec-30 cells (P < 0.001; Fig. 4B). Twenty days after injection,tested mice were sacrificed and the tumors were excised forfurther analysis. The average volume of tumors induced byRBMS3-30 cells (91.9 � 68.7 mm3) was significantly de-creased compared with tumors induced by Vec-30 cells(651.97 � 120.65 mm3, P < 0.001; Fig. 4C). RBMS3 expressionin xenograft tumors was studied by IHC, and the resultshowed that RBMS3 expression was only detected in

tumors induced by RBMS3-30 cells but not in tumorsinduced by Vec-30 cells (Fig. 4D).

RBMS3 downregulates c-Myc and CDK4Because RBMS3 belongs to the family of MSSP, the effect

of RBMS3 on c-Myc expression was investigated by qRT-PCRand Western blot analysis. The results showed that theexpression of c-Myc was downregulated in RBMS3-30 cellscompared with that of Vec-30 cells (Fig. 5A). Expression ofother cell-cycle–related proteins, including CDK2, CDK4,cyclin D1, and cyclin E, was also studied by Western blotanalysis, and the results showed that only CDK4 was down-regulated (Fig. 5A). The downregulation effect of RBMS3 onc-Myc and CDK4 in KYSE30 cells was confirmed by qRT-PCR(Fig. 5B). The downregulation effect of RBMS3 on c-Myc andCDK4 was further confirmed in 2 other ESCC cell lines,KYSE140 and HKESC1, by qRT-PCR. The result showed thattransient transfection of RBMS3 into these 2 cell lines couldeffectively downregulate expression of c-Myc and CDK4(Fig. 5C).

Figure 4. A, foci formation assayshowed that the number of fociformation (left, representative;right, summary) was significantlydecreased in RBMS3overexpressed cells (RBMS3-30and RBMS3-180) compared withvector control cells (P < 0.01). Theresults are expressed as mean �SD of 3 independent experiments.B, summary of tumor growthcurves in nude mice induced byRBMS3-30 and Vec-30 cells. Theaverage tumor volume wasexpressed as mean � SD in 8inoculated sites for each group(**, P < 0.001). C, representativesamples of tumor formation innude mice. Tumors induced byRBMS3-30 (left) and Vec-30 cells(right) are indicated by red andblack arrows, respectively.Excised tumors are shown in thebottom insets. D, representativesamples of RBMS3 expressiondetected by IHC in xenografttumors induced by RBMS3-30(left) and Vec-30 cells (right).

RBMS3 in ESCC





RBMS3 binds to the putative DNA replication originof c-Myc

A sequence of 21 bp about 2 kb upstream of the humanc-Myc gene has been identified as putative DNA replicationorigin and a transcriptional enhancer, which can be bound byMSSPs (22). ChIP assay was done in EC18 as describedpreviously (23) to confirm the interaction of RBMS3 withthe fragments (F1: nt -2588/-2421; F2: nt -2313/-2085; andF3: nt -1202/-1020, F4: -717/-503) from the 50 flanking region ofc-Myc (Fig. 5D). The results showed that only DNA F2 contain-ing the putative DNA replication origin sequence of c-Myc, butnot F1, 3, and 4, could be detected in RBMS3-ChIPed DNAfragments by PCR (Fig. 5D).

To further confirm that RBMS3 could bind to the putativeDNA replication origin sequence of c-Myc, EMSAwas done usinga 60-bp probe containing the binding sequence. As shown inFig. 5E, the EC18 nuclear extract bound to the probe and formedDNA–protein complex. These results showed that RBMS3 is ableto bind directly to the putative DNA replication origin of c-Myc.

Silencing RBMS3 by RNAi inhibits its tumor-suppressive ability

To further show that the tumor-suppressing function ofRBMS3 was through downregulation of c-Myc expression,RNAi assay was applied to silence RBMS3 expression. qRT-PCR results showed that the expression of RBMS3 in RBMS3-30

Figure 5. RBMS3 downregulatesc-Myc and CDK4 expression. A,expression of RBMS3, c-Myc,CDK2, CDK4, cyclin D1, andcyclin E was compared betweenRBMS3-30 and control Vec-30cells by Western blot analysis.GAPDH was used as loadingcontrol. B, qRT-PCR analysisshowed that ectopic expression ofRBMS3 could downregulate c-Myc and CDK4 expression inKYSE30 cells. Results areexpressed as mean � SD of 3independent experiments(*, P < 0.05). C, when RBMS3was transiently transfected intoKYSE140 and HKESC1 cells,c-Myc and CDK4 expression wasdecreased significantly(*, P < 0.01). Results are expressedas mean � SD of 3 independentqRT-PCR experiments. D,schematic diagram representingthe distribution of RBMS3-bindingloci in the regulatory region ofc-Myc (top). Four individualfragments representing 50

upstream sequences of c-Myc(F1: nt -2588/-2421; F2: nt -2313/-2085; F3: nt -1202/-1020; and F4:nt -717/-503). PCR was used todetect the existence of fragments1, 2, 3, and 4 in DNA fragmentspulled down by an anti-RBMS3antibody. IgG was used as anegative control. E, EMSA wasused to detect the interactionbetween RBMS3 and c-Mycdouble-stranded DNA probes(100 fM or 200 fM). Supershiftband could be detected whennuclear extract (15 mg) fromEC18 cells was added.

Li et al.





cells could be effectively silenced by siRNA against RBMS3,which subsequently upregulated c-Myc and CDK4 expression,compared with RBMS3-30 cells treated with scramble siRNA(Fig. 6A). Foci formation assay showed that silencing RBMS3expression could increase foci formation ability comparedwith control cells (Fig. 6B).

RBMS3 reduces CDK4 and phosphorylated RbexpressionBecause CDK4 plays a critical role via the inactivation of Rb,

the level of inactive form of Rb (phosphorylated Rb) was thencompared between RBMS3-30 and Vec-30 cells by Westernblot analysis. The result showed that no obvious change wasdetected in total Rb protein between RBMS3-30 and Vec-30cells. However, inactive forms of Rb (phosphorylated atSer807/811 and Ser780) were reduced in RBMS3-30 cellscompared with those in Vec-30 cells (Fig. 6C). PhosphorylatedRb (ser780) in cells increased significantly when RBMS3 wassilenced by RNAi compared with scramble control.

Discussion

Deletion of 3p is one of the most frequent alterations inprimary ESCC. Here we report the characterization of one

candidate tumor suppressor gene, RBMS3, at 3p24. Down-regulation of RBMS3 in mRNA and protein level was detectedin 45% and 52.9% of primary ESCCs, respectively. LOH studyshowed that DNA copy number loss was detected in 64%(11/17) of ESCC cases. Interestingly, downregulation ofRBMS3 was significantly associated with poor outcome ofESCC patients (P ¼ 0.034), suggesting that RBMS3 might playan important role in ESCC development and progression. Thetumor-suppressive function of RBMS3 was characterized in 2ESCC cell lines by either stable transfection (KYSE30) ortransient transfection (KYSE180). Both in vitro and in vivoassays were used to investigate the tumor-suppressive poten-tial of RBMS3. In addition, silencing RBMS3 expression byRNAi could inhibit its tumor-suppressive ability. The resultsshowed that RBMS3 could effectively suppress cell growth,decrease foci formation and colony formation in anchorage-dependent and anchorage-independent assays, and inhibittumor formation in nude mice.

RBMS3 belongs to a family of MSSPs including 3 members.MSSPs are believed to regulate DNA replication, transcription,apoptosis, and cell-cycle progression by interacting with asequence of 21 bp about 2 kb upstream of the human c-Mycgene (22). Moreover, MSSP-1 and MSSP-2 can bind directly tothe C-terminal portion of c-Myc, which, along with Max, forms

Figure 6. A, siRNA againstRBMS3 (si-RBMS3) coulddownregulate RBMS3 expressionand upregulate c-Myc and CDK4expression in RBMS3-30 cells.Scramble siRNA was used asnegative control. Results areexpressed as mean � SD of 3independent qRT-PCRexperiments. B, the frequency offoci formation (left, representative;right, summary) was significantlyhigher in RBMS3-30 cells treatedwith si-RBMS3 than in cellstreated with scramble siRNA(P < 0.01). The results areexpressed as mean � SD of3 independent experiments.C, expression of phosphorylatedRb (ser807/811 and ser780) wascompared between RBMS3-30and Vec-30 cells by Western blotanalysis (left). siRNA againstRBMS3 could upregulatephosphorylated Rb (ser780) inRBMS3-30, EC109, and EC18cells, compared with scramblecontrol. GAPDH was used asloading control.

RBMS3 in ESCC





a ternary complex, which loses the binding activity to therecognition sequence of c-Myc/Max complex, thereby abro-gating the E-box–dependent transcription activity of c-Myc(14). In another study, MSSP has been reported to be releasedfrom a putative DNA replication origin of the c-Myc gene afterit complexed with a catalytic subunit of a DNA polymerasealpha (24). Although RBMS3 contains 2 consensus motifs of anRNA-binding protein, like other MSSP family members, whichcould artificially bind to DNA through chimeric GAL4 hybridsystem (12), its role in the regulation of c-Myc expression hasnot been shown.

In this study, we found that RBMS3 could directly bind to theputative DNA replication origin sequence about 2 kb upstreamof the c-Myc gene and, subsequently, downregulated c-Mycexpression in both mRNA and protein levels. This resultsuggests that RBMS3 could also abrogate the transcriptionactivity of c-Myc like MSSP-1 and MSSP-2. In addition, ectopicexpression of RBMS3 could downregulate CDK4 expression inESCC cells. It is not clear whether RBMS3 could directlydownregulate CDK4 expression. c-Myc is a proto-oncogenethat plays an important role in tumorigenesis in many malig-nances through stimulating expression of cyclin A, D, E, CDK2,and CDK4 (25, 26). CDK4 is a master integrator that initiatesthe phosphorylation of the central oncosuppressor Rb andinactivates its blocking activity of E2F transcription (27, 28).Therefore, we further studied the phosphorylation status of Rband the result showed that expression of RBMS3 could reduce

the phosphorylation form of Rb. These results suggest that thetumor-suppressive mechanism of RBMS3might be through theinhibition of c-Myc and CDK4 expression, and sequentiallydownregulates Rb phosphorylation and prevents E2F activity,which is required for cell-cycle progression, and further down-regulates c-Myc expression. Taken together, our data show thatRBMS3 is a novel tumor suppressor gene in ESCC, and itsdownregulation is associated with poor prognosis in ESCC.Further characterization of the tumor-suppressive function andmechanism of RBMS3 will not only greatly facilitate our un-derstanding of ESCC development and progression but will alsoprovide novel therapeutic targets in ESCC treatment.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Grant Support

This work was supported by Grants from the National Natural ScienceFoundation of China (30700820, 30772475, and 30971606), Research Fund for theDoctoral Program of Higher Education of China (20070558272) and Sun Yat-SenUniversity "Hundred Talents Program" (85000-3171311), National Key Sci-TechSpecial Project of China (2008ZX10002-022), Hong Kong Research Grant CouncilCentral Allocation (HKUST 2/06C).

Received November 26, 2010; revised August 4, 2011; accepted August 8,2011; published OnlineFirst August 15, 2011.

References1. Parkin DM, Bray FI, Devesa SS. Cancer burden in the year 2000. The

global picture. Eur J Cancer 2001;37 Suppl 8:S4–66.2. Wei WQ, Yang J, Zhang SW, Chen WQ, Qiao YL. [Analysis of the

esophageal cancer mortality in 2004–2005 and its trends during last30 years in China]. Zhonghua Yu Fang Yi Xue Za Zhi 2010;44:398–402.

3. Carter CL, Hu N,WuM, Lin PZ,Murigande C, Bonney GE. Segregationanalysis of esophageal cancer in 221 high-risk Chinese families. J NatlCancer Inst 1992;84:771–6.

4. Hu N, Dawsey SM, Wu M, Bonney GE, He LJ, Han XY, et al. Familialaggregation of oesophageal cancer in Yangcheng County, ShanxiProvince, China. Int J Epidemiol 1992;21:877–82.

5. Hu N, Wang C, Ng D, Clifford R, Yang HH, Tang ZZ, et al. Genomiccharacterization of esophageal squamous cell carcinoma from a high-risk population in China. Cancer Res 2009;69:5908–17.

6. Kwong D, Lam A, Guan X, Law S, Tai A, Wong J, et al. Chromosomalaberrations in esophageal squamous cell carcinoma among Chinese:gain of 12p predicts poor prognosis after surgery. Hum Pathol2004;35:309–16.

7. Ogasawara S, Maesawa C, Tamura G, Satodate R. Frequent micro-satellite alterations on chromosome 3p in esophageal squamous cellcarcinoma. Cancer Res 1995;55:891–4.

8. Yen CC, Chen YJ, Chen JT, Hsia JY, Chen PM, Liu JH, et al.Comparative genomic hybridization of esophageal squamous cellcarcinoma: correlations between chromosomal aberrations and dis-ease progression/prognosis. Cancer 2001;92:2769–77.

9. Qin YR, Fu L, Sham PC, Kwong DL, Zhu CL, Chu KK, et al. Single-nucleotide polymorphism-mass array reveals commonly deletedregions at 3p22 and 3p14.2 associate with poor clinical outcome inesophageal squamous cell carcinoma. Int J Cancer 2008;123:826–30.

10. Fu L, Qin YR, Xie D, Hu L, Kwong DL, Srivastava G, et al. Charac-terization of a novel tumor-suppressor gene PLC delta 1 at 3p22 inesophageal squamous cell carcinoma. Cancer Res 2007;67:10720–6.

11. Zhu C, Qin YR, Xie D, Chua DT, Fung JM, Chen L, et al. Character-ization of tumor suppressive function of P300/CBP-associated factor

at frequently deleted region 3p24 in esophageal squamous cellcarcinoma. Oncogene 2009;28:2821–8.

12. Penkov D, Ni R, Else C, Pinol-Roma S, Ramirez F, Tanaka S. Cloningof a human gene closely related to the genes coding for the c-mycsingle-strand binding proteins. Gene 2000;243:27–36.

13. Kimura K, Saga H, Hayashi K, Obata H, Chimori Y, Ariga H, et al. c-Myc gene single-strand binding protein-1, MSSP-1, suppresses tran-scription of alpha-smooth muscle actin gene in chicken visceralsmooth muscle cells. Nucleic Acids Res 1998;26:2420–5.

14. Niki T, Izumi S, Sa€egusa Y, Taira T, Takai T, Iguchi-Ariga SM, et al.MSSP promotes ras/myc cooperative cell transforming activity bybinding to c-Myc. Genes Cells 2000;5:127–41.

15. Fritz D, Stefanovic B. RNA-binding protein RBMS3 is expressed inactivated hepatic stellate cells and liver fibrosis and increasesexpression of transcription factor Prx1. J Mol Biol 2007;371:585–95.

16. Shimada Y, Imamura M. Prognostic significance of cell culture incarcinoma of the oesophagus. Br J Surg 1993;80:605–7.

17. Shimada Y, Imamura M, Wagata T, Yamaguchi N, Tobe T. Charac-terization of 21 newly established esophageal cancer cell lines.Cancer 1992;69:277–84.

18. Wong ML, Tao Q, Fu L, Wong KY, Qiu GH, Law FB, et al. Aberrantpromoter hypermethylation and silencing of the critical 3p21 tumoursuppressor gene, RASSF1A, in Chinese oesophageal squamous cellcarcinoma. Int J Oncol 2006;28:767–73.

19. Livak KJ, Schmittgen TD. Analysis of relative gene expression datausing real-time quantitative PCR and the 2(-Delta Delta C(T)) Method.Methods 2001;25:402–8.

20. Wang Y, Wu MC, Sham JS, Tai LS, Fang Y, Wu WQ, et al. Differentexpression of hepatitis B surface antigen between hepatocellularcarcinoma and its surrounding liver tissue, studied using a tissuemicroarray. J Pathol 2002;197:610–6.

21. Li Y, Nie CJ, Hu L, Qin Y, Liu HB, Zeng TT, et al. Characterization of anovel mechanism of genomic instability involving the SEI1/SET/

Li et al.





NM23H1 pathway in esophageal cancers. Cancer Res 2010;70:5695–705.

22. Negishi Y, Iguchi-Ariga SM, Ariga H. Protein complexes bearing myc-like antigenicity recognize two distinct DNA sequences. Oncogene1992;7:543–8.

23. Chen L, Chan TH, Yuan YF, Hu L, Huang J, Ma S, et al. CHD1Lpromotes hepatocellular carcinoma progression and metastasis inmice and is associated with these processes in human patients. J ClinInvest 2010;120:1178–91.

24. Niki T, Galli I, Ariga H, Iguchi-Ariga SM. MSSP, a protein binding to anorigin of replication in the c-myc gene, interacts with a catalyticsubunit of DNA polymerase alpha and stimulates its polymeraseactivity. FEBS Lett 2000;475:209–12.

25. Sherr CJ, Roberts JM. CDK inhibitors: positive and negativeregulators of G1-phase progression. Genes Dev 1999;13:1501–12.

26. Beier R, B€urgin A, Kiermaier A, Fero M, Karsunky H, Saffrich R, et al.Induction of cyclin E-cdk2 kinase activity, E2F-dependent transcrip-tion and cell growth by Myc are genetically separable events. EMBO J2000;19:5813–23.

27. Coqueret O. Linking cyclins to transcriptional control. Gene 2002;299:35–55.

28. Paternot S, Bockstaele L, Bisteau X, Kooken H, Coulonval K, RogerPP. Rb inactivation in cell cycle and cancer: the puzzle of highlyregulated activating phosphorylation of CDK4 versus constitutivelyactive CDK-activating kinase. Cell Cycle 2010;9:689–99.

RBMS3 in ESCC





2011;71:6106-6115. Published OnlineFirst August 15, 2011.Cancer Res Yan Li, Leilei Chen, Chang-jun Nie, et al. Esophageal Squamous Cell CarcinomaDownregulation of RBMS3 Is Associated with Poor Prognosis in

Updated version

10.1158/0008-5472.CAN-10-4291doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2011/08/15/0008-5472.CAN-10-4291.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/71/19/6106.full#ref-list-1

This article cites 28 articles, 6 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/71/19/6106.full#related-urls

This article has been cited by 8 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/71/19/6106To 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-10-4291

http://cancerres.aacrjournals.org/content/suppl/2011/08/15/0008-5472.CAN-10-4291.DC1

http://cancerres.aacrjournals.org/content/71/19/6106.full#ref-list-1

http://cancerres.aacrjournals.org/content/71/19/6106.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/71/19/6106


Documents

Downregulation of RBMS3 Is Associated with Poor Prognosis ... · Priority Report Downregulation of RBMS3 Is Associated with Poor Prognosis in Esophageal Squamous Cell Carcinoma Yan