Tumor Suppressor Functions of miR-133a in Colorectal …mcr.aacrjournals.org/content/molcanres/11/9/1051.full.pdf · Tumor Suppressor Functions of miR-133a in Colorectal Cancer Yujuan

Genomics

Tumor Suppressor Functions of miR-133a in ColorectalCancer

Yujuan Dong1,2, Junhong Zhao1, Chung-Wah Wu1, Lijing Zhang3, Xiaodong Liu4, Wei Kang5,Wing-Wah Leung2, Ning Zhang1, Francis K.L. Chan1, Joseph J.Y. Sung1, Simon S.M. Ng2, and Jun Yu1

AbstractDysregulated microRNA (miRNA) expression was profiled through a miRNA array comparison between

human colorectal cancer tumors and their adjacent normal tissues. Specifically, using laser capture micro-dissection, miR-133a was shown to be significantly downregulated in primary colorectal cancer specimenscompared with matched adjacent normal tissue. Ectopic expression of miR-133a significantly suppressedcolorectal cancer cell growth in vitro and in vivo. Cell-cycle analysis revealed that miR-133a induced a G0/G1-phase arrest, concomitant with the upregulation of the key G1-phase regulator p21

Cip1. We further revealedthat miR-133a markedly increased p53 protein and induced p21Cip1 transcription. Studies in silico revealedthat the 30UTR of the ring finger and FYVE-like domain containing E3-ubiquitin protein ligase (RFFL), whichregulates p53 protein, contains an evolutionarily conserved miR-133a binding site. miR-133a repressed RFFL-30UTR reporter activity and reduced RFFL protein levels, indicating that miR-133a directly bound to RFFLmRNA and inhibited RFFL translation. Moreover, miR-133a sensitized colon cancer cells to doxorubicin andoxaliplatin by enhancing apoptosis and inhibiting cell proliferation. These data add weight to the significanceof miR-133a in the development of CRC.

Implications: miR-133a serves as a potential tumor suppressor upstream of p53 in colorectal cancer and maysensitize cells to therapeutics. Mol Cancer Res; 11(9); 1051–60. �2013 AACR.

IntroductionColorectal cancer (CRC) is the thirdmost common cancer

and the third leading cause of cancer-related death in theworld, with an estimated incidence of 1.2 million new casesand a mortality of more than 600,000 deaths annually (1).While CRC carcinogenesis involving multistep progressionhas been extensively studied at the genetic level, recent datahighlight that microRNAs (miRNAs) are closely associatedwith CRC tumorigenesis. miRNAs belong to a class ofhighly conserved 22-nucleotide single-stranded RNAs thatsuppress target genes translation or induce messenger RNA

(mRNA) degradation through binding to the 30-untranslat-ed region (UTR; ref. 2). It is estimated that miRNAs canregulate up to 50% human genes translation (3). By target-ing multiple transcripts, miRNAs epigenetically regulatefundamental cellular processes such as cell proliferation,apoptosis, differentiation, and migration, which stronglyindicates that they may function as potential oncogenes ortumor suppressors in cancer development. Indeed, a globalimpairment of miRNA was described in various humancancers including CRCs (4). However, the precise mechan-isms of these aberrantly expressed miRNAs in CRC genesis,progression, and therapeutic response are still largelyunknown. An investigation into the mechanism of thedysregulated miRNAs in pathogenesis of CRCs can helpidentify novel molecular events and signal pathways for thedevelopment of anticancer therapy. We recently identifieddysregulated miRNAs in human CRC through comparisonbetween tumors and their adjacent normal tissues bymiRNA array (GSE45349). In this study, we aim to clarifymiR-133a biologic function, molecular basis, and targetgene in CRCs.

Materials and MethodsTissue samplesTwo cohorts of totally 95 patients with histologically

confirmed CRC from 1999 to 2009 who underwent surgeryat the Prince of Wales Hospital, Hong Kong, were included

Authors' Affiliations: 1Institute of Digestive Disease and Department ofMedicine and Therapeutics, Li Ka Shing Institute of Health Sciences,CUHK-Shenzhen Research Institute, and 2Department of Surgery, Princeof Wales Hospital, The Chinese University of Hong Kong, Hong Kong;3Department of Surgery, The First Affiliated Hospital of Hebei MedicalUniversity, Shijiazhuang, Hebei, China; 4Department of Anaesthesia andIntensive Care, Prince of Wales Hospital, and 5Department of Anatomicaland Cellular Pathology, The Chinese University of Hong Kong, Hong Kong

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Authors: Jun Yu, Department of Medicine and Therapeu-tics, Prince of Wales Hospital, The Chinese University of Hong Kong,Shatin, NT, Hong Kong. Phone: 852-3763-6099; Fax: 852-2144-5330;E-mail: [email protected]; and Simon S.M. Ng. Phone: 852-2632-1495; Fax: 852-2637-7974; E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-13-0061

�2013 American Association for Cancer Research.

MolecularCancer

Research

www.aacrjournals.org 1051

on June 4, 2018. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst May 30, 2013; DOI: 10.1158/1541-7786.MCR-13-0061

http://mcr.aacrjournals.org/

in this study, including 4 patients for building of themiRNAexpression profiles and 91 patients for miRNA validation(Supplementary Tables S1 and 2). CRC tissues andpaired adjacent normal tissues were obtained from theresected surgical specimens. The adjacent normal tissue iscomposed of normal colonic mucosa located approximately10 cm away from the cancer tissue. In addition, 9 biopsyspecimens of normal rectal mucosa from healthy controlsand 11 pairs of CRC paraffin-embedded tissue samples wereobtained at the Prince of Wales Hospital. All subjectsprovided informed consent before specimen collection. Thestudy protocol was approved by the Ethics Committee ofThe Chinese University of Hong Kong.

Cell cultureCell lines included in the present study (Caco2, colo205,

DLD1, HCT116, HT29, LoVo, LS180, SW480, SW620)were obtained from American Type Culture Collection andpropagated at 37�C in the presence of 5% CO2. LoVo wascultured in Dulbecco's Modified Eagle's Medium (GibcoBRL) supplemented with 10% (v/v) FBS (Hyclone, ThermoFisher Scientific), HCT116 and HT29 were cultured inMcCoy's 5Amedium (Sigma-Aldrich) containing 10% (v/v)FBS. When required, media were supplemented with 2 mg/mL puromycin (Invitrogen).

miRNA arrayReverse transcription for miRNA microarray was carried

out using Megaplex Primer pools, Human Pools A v2.1and B v3.0 (Applied Biosystems). The cDNA product wasused to conduct miRNA array. miRNA profiling of 754human miRNAs was conducted using TaqMan HumanMiRNA Array Set v3.0. Quantitative real-time PCR(qRT-PCR) was carried out using Applied Biosystems7900HT Real-Time PCR System (Applied Biosystems).Results were analyzed by the SDS RQ Manager 1.2software (Applied Biosystems).

miRNA and mRNA expression analysesTotal RNA was extracted from cell pellets or tissues using

Quizol reagent (Qiagen). TaqMan miRNA assays were usedto quantify levels of mature miR-133a and RNU6B (KIT002246 and 001093, Applied Biosystems). mRNA levelswere determined with qRT-PCR using SYBR Green mastermixture on HT7500 system (Applied Biosystems) with thehousekeeping gene ACTB as an internal control. Cyclethreshold (Ct) values >35 were set as the detection limit.Primers used in this study are listed in Supplementary TableS3.

Microdissection and RNA extraction from paraffinsectionsParaffin-embedded tissue samples used for laser capture

microdissection were randomly selected from validationcohort II. Samples were cut into 10 mm thick sections usinga microtome. Tissue slides were reviewed by a pathologist,and the muscularis mucosae in normal tissues were thenexcluded using laser capture microdissection under micro-

scope. Total RNA was extracted from the tissues usingRecoverAll Total Nucleic Acid Isolation Kit (Ambion).

miRNA transfection and RNA interferenceHuman miR-133a precursor (PM10413) was purchased

from Applied Biosystems. A control miRNA (miR-Ctrl;AM17110, Applied Biosystems) did not target any humangenes was used as a negative control. Cells were transfectedwithmiRNAs using Lipofectamine 2000 (Invitrogen). Shorthairpin RNAs (shRNA) against human RFFL (Origene)were delivered into cells using Lipofectamine 2000, andstable cell lines were created from puromycin-resistantcolonies. siRNAs against p53 (Santa Cruz Biotechnology)were delivered into cell using Lipofectamine 2000. Cellstransfected with miRNA or siRNA were harvested 12 to 48hours after transfection.

Colony formation and MTT cell viability assayAfter 24 hours of transfection, cells were collected and

seeded (500–1,000 per well) in a fresh 6-well plate for 7 to 9days. Colonies were counted after staining with crystal violetsolution. Cell viability was determined using the conversionof MTT to formazan via mitochondrial oxidation. MTTsolution was then added to each well at a final concentrationof 0.2mg/mL per well and the plates were incubated at 37�Cfor another 1 hour. After incubation, 200 mL of dimethylsulfoxide was added to each well to dissolve the formazanformed, and the absorbance was measured at 570 nm using aspectrophotometer. All the experiments were repeated in 3independent experiments in triplicate.

Flow cytometryAt 24 hours posttransfection, cells were treated with

nocodazole (25 ng/mL, Sigma-Aldrich) for an additional14 hours, or doxorubicin (100 nmol/L, Sigma-Aldrich), oroxaliplatin (1 mmol/L, Sigma-Aldrich) for additional 48hours. Cells were fixed and stained with propidium iodide(BD Biosciences). Cells undergoing apoptosis were detectedby Annexin V/7-aminoactinomycin D (AAD) staining (BDBiosciences) as per manufacturer's instructions. The cellswere then sorted by FACSCalibur (BD Biosciences). A totalof 30,000 events were counted for each sample and cell-cycleprofiles were analyzed by ModFit 3.0 software (BD Bios-ciences). All the experiments were repeated in 3 independentexperiments in triplicate.

Western blot analysisProtein concentration was measured by the Bradford DC

protein assay (Bio-Rad). About 5 to 20 mg of protein fromeach sample was separated on 12% Bis–Tris PAGE throughelectrophoresis and blotted onto nitrocellulose membranes(GE Healthcare). Blots were immunostained with primaryantibodies at 4�Covernight and secondary antibody at roomtemperature for 1 hour. Protein was visualized using ECLPlusWestern Blotting Detection Reagents (GEHealthcare).Antibodies against p53 (sc-126; Santa Cruz, dilution1:1,000), CDKN1A (p21) (sc-6246; Santa Cruz, dilution1:1,000), and RFFL (Rockland, dilution 1:2,000) were used

Dong et al.

Mol Cancer Res; 11(9) September 2013 Molecular Cancer Research1052




in Western blot analysis. Anti-GAPDH antibodies (sc-25778; Santa Cruz, dilution 1:2,000) were used to test forequal loading.

Plasmids and dual-luciferase reporter assayThe full-length open reading sequence of human RFFL

(NM_001017368.1) was PCR-amplified from HCT116cDNA. The PCR aliquots were subcloned into pcDNA3.1 (Invitrogen) to generated pRFFL plasmid and thenverified by DNA sequence. Human miR-133a expressionplasmid pmiR-133a (SC400158) and control vectorpCMVMIR were purchased from Origene. To generatethe reporter plasmid pMIR-RFFL-30UTR and pMIR-RFFLmut-30UTR, the annealed oligonucleotides contain-ing wild-type or mutated human RFFL 30-UTR were clonedinto pMIR-REPORT vector (Ambion) at the SacI and SpeIsites, respectively. Primer sequences are listed in Supple-mentary Table S3. pGL13 plasmid (containing 14 copies ofthewild-type p53 response elements that drives the luciferasereporter gene: 50-CCAGGCAAGTCCAGGCAGG-30) andp21-luc plasmid (containing 2.1-kb p21 promoter regionthat drives the luciferase reporter gene) were used to inves-tigate the p53 signaling pathways modulated by miR-133a.Cotransfection of reporter plasmids and miRNA in 24-wellplates were carried outwith Lipofectamine 2000 as describedby the manufacture. Per well, 195 ng luciferase reportplasmid, 5 ng pRL-cytomegalovirus (CMV) vector, and15 pmol miRNA were applied. Cells were harvested 48hours posttransfection and luciferase activities were analyzedby the dual-luciferase reporter assay system (Promega).

Tumor xenografts in nude mice modelIn the transient miR-133a expression model, approxi-

mately 2 � 106 HCT116 cells were subcutaneouslyimplanted into the right flank of the female BALB/c nudemice (4 weeks old, 4mice per group). miR-133a ormiR-Ctrlwere injected into the tumor using syringes and 30G needleswhen tumor size reached to 3 mm (length)� 3mm (width).miRNAs were prepared by preincubating miRNA (0.3nmol) with Lipofectamine 2000 (2.5 mL) (Invitrogen) for15 minutes, and injection was made in a final volume of 100mL inMcCoy's 5Amedium (Sigma-Aldrich) permouse. Theinjection was repeated every 3 days and consisted of 4consecutive injections as described before (5). In the stablemiRNA expression model, 2 � 106 HCT116 cells stablyexpressing miR-133a or control vector were injected sub-cutaneously into the left flank of the female BALB/c nudemice (4 weeks old, 4 mice per group), respectively. Tumorsizes were measured starting from the first injection andthe tumor volume was calculated as (length �width2)/2.The experiments were approved by Animal Experimenta-tion Ethics Committee, The Chinese University of HongKong.

Statistical analysisThe results were expressed as mean� SD. The difference

in miRNA levels between paired tissue samples was deter-mined by the Wilcoxon matched-pairs test. P < 0.05 was

taken as statistical significance. All the tests were conductedby GraphPad Prism 5.0.

ResultsDownregulation of miR-133a in primary human CRCand colon cancer cell linesTo identify candidate miRNAs associated with CRCs, we

compared miRNA expression profiles of 4 pairs of CRC andadjacent normal tissues using miRNA microarray covering754 miRNAs. We found that miR-133a was significantlydownregulated in CRCs based on the array data(GSE45349). We then verified the downregulation ofmiR-133a in 91 pairs of primary CRCs and paired normaltissues from 2 cohorts of patients with CRC by qRT-PCRanalysis. U6 small nuclear 2 (RNU6B) was used as aninternal control. Two pairs of tissues were excluded fromthis study due to undetectable miR-133a or RNU6B.Among the 89 pairs of CRC tissues specimens examined,80 (89.9%) tumor tissues showed significantly lower miR-133a expression when compared with matched adjacentnormal tissues (36 of 44, 81.8%, P < 0.001 for cohort I;and 44 of 45, 97.8%,P< 0.001 for cohort II; Fig. 1A1 and 2,Supplementary Table S4), with a median difference of 0.15-fold [interquartile range (IQR), 0.46–0.05] in cohort I and0.05-fold (IQR, 0.13–0.02) in cohort II, respectively. miR-133a was first identified as a muscle-specific miRNA (6).The presence of smooth muscle in the adjacent normaltissues might affect the results in our qRT-PCR assay. Toclarify the issue, we eliminated the muscularis mucosae withmicrodissection in randomly selected 11 pairs of CRCsand adjacent normal tissues. The level of miR-133a wassignificantly decreased in tumor tissues compared withnormal epithelial cells (Fig. 1A3, P < 0.01). These datafurther strengthened the significance of Fig. 1A1 and 2. Inaddition, we evaluated the expression of miR-133a in 9colon cancer cell lines and 9 normal rectal tissues fromhealthy subjects. miR-133a was dramatically underex-pressed in all 9 colon cancer cell lines (100%) examined,suggesting aberrant downregulated expression of miR-133a in CRCs (Fig. 1B).

Ectopic expression of miR-133a inhibits tumorigenicproperties of CRC cellsThe frequent silence of miR-133a in CRC primary

cancers and cell lines suggests that miR-133a might be aputative tumor suppressor. In this regard, we examined theeffect of ectopic expression of miR-133a on cell growth incolon cancer cell lines. We noticed that the miR-133a levelsinHCT116 and LoVowere about 104-fold less than those inthe normal tissues (Fig. 1B). To eliminate the nonspecificeffects generated by overtransfection, we adjusted the trans-fection amount of miRNA and restored miR-133a levelapproximately equivalent to the physiologic level (Fig. 2A).Ectopic expression of miR-133a significantly suppressedcolony formation with 20% reduction in HCT116 and26% reduction in LoVo (P < 0.01) when compared withmiR-Ctrl–transfected cells (Fig. 2B). This effect was furtherconfirmed by MTT assay. Overexpression of miR-133a

Role of miR-133a in Colorectal Cancer

www.aacrjournals.org Mol Cancer Res; 11(9) September 2013 1053




caused a significant decrease in cell viability in both p53wild-type cell lines HCT116 (P < 0.01) and LoVo (P < 0.05; Fig.2C andD). miR-133a also exerted inhibitory effects on p53-mutant cell lines DLD1 and HT29 (Supplementary Fig.S1). These results indicated that miR-133a exerts growth-inhibitory ability in colon cancer cells and functions as apotential tumor suppressor.

miR-133a inhibits tumor growth in nude miceBecause miR-133a can suppress CRC cell growth in vitro,

we next tested whether miR-133a could suppress tumori-genesis in vivo. We used 2 different miRNA expressionmodels to study themiR-133a function in vivo: (i)miR-133aprecursor transient expression model and (ii) vector-basedmiR-133a constitutive expression model. In the transientexpression model, the BALB/c nude mice intratumoralinjected with miR-133a precursor showed a reduced tumorgrowth rate compared with the miR-Ctrl–injected mice (P <0.05; Fig. 2E). In the constitutive expression model, wegenerated the miR-133a and control vector stable transfec-tants (designated as HCT116-miR-133a and HCT116-miR-Ctrl). HCT116-miR-133a exhibited markedlyreduced growth rate in nude mice compared with the

HCT116-miR-Ctrl cells (P < 0.01; Fig. 2F), supportingthat miR-133a suppressed the colon cancer growth in vivo.We also checked the miR-133a expression levels in bothxenograft models. In the transient expression model, wefound that the miR-133a levels showed a 105-fold increasecompared with those in the control group 24 hours after thefinal injection at day 13, and themiR-133a in the tumors stillshowed a 102-fold increase at the end of the experiment atday 36 (23 days following the final injection). In the stabletransfectants group, the amounts of miR-133a were approx-imately 102-fold higher in HCT116-miR-133a than thosein the control group at the beginning (day 4). No significantmiR-133a reduction was observed at the end of the exper-iment (day 34; Supplementary Fig. S2).

miR-133a induces G0–G1 phase arrestTo elucidate the mechanism of miR-133a function in the

control of CRC cell growth, we next examined the effects ofmiR-133a on cell-cycle regulation. Cells were synchronizedwith an antimitotic drug nocodazole (7). Transfectionof miR-133a significantly increased the fraction of cells inG0–G1 phase compared with the miR-Ctrl–transfected cellsin both HCT116 (P < 0.01) and LoVo (P < 0.01; Fig. 3A

Figure 1. miR-133a frequently downregulated in primary CRC tissues and colon cancer cell lines. A1 and A2, levels of miR-133a in cohort I (A1, n ¼ 44)and cohort II (A2, n ¼ 45) of colorectal tumors and adjacent normal tissues. The miR-133a level was normalized to RNU6B and P value indicatedsignificant difference in miR-133a level between paired samples determined by the Wilcoxon matched-pairs test. A3, miR-133a was frequentlydecreased in primary CRCs (paraffin-embedded tissues). Normal epithelia cells and tumor cells were isolated from paraffin blocks using themicrodissection.The miR-133a level was normalized to RNU6B. B, the expression profile of miR-133a in 9 colon cancer cell lines and normal rectal tissues (n ¼ 9) byqRT-PCR with RNU6B as an internal control.

Dong et al.





and B). The role of miR-133a in apoptosis was also evalu-ated, and no significant proapoptotic effect was observed.

miR-133a upregulates p53/p21 pathwayThe G0–G1 cell-cycle arrest phenomenon promoted us to

examine the regulation role of miR-133a on the CDKinhibitor CDKN1A (p21) and tumor suppressor p53.Ectopic expression of miR-133a dramatically increasedp21 mRNA and protein expressions (Fig. 3C and D). Theincreased p21 localized in the nucleus (Supplementary Fig.S3). We found that the p53 protein but not mRNA wassignificantly upregulated after treatment with miR-133a in

bothHCT116 and LoVo cells (Fig. 3C and E). On the otherhand, c-myc, an oncoprotein that could suppress p21expression, was unchanged after transfection with miR-133a in both cell lines (Fig. 3C), favoring the hypothesisthat miR-133a positively regulated p53/p21 pathway. Toconfirm the interaction between miR-133a and p53/p21pathway, we further conducted the promoter luciferaseactivity assays using pGL13 and p21-Luc luciferase reportersin HCT116 cells. As expected, cotransfection of miR-133aand pGL13 significantly increased p53 luciferase reporteractivities by 2.6-fold in HCT116 (Fig. 3F; P < 0.001) and1.4-fold in LoVo (Fig. 3G; P < 0.05). Cotransfection of

Figure 2. miR-133a inhibited coloncancer cell growth in vitro and invivo. A, ectopic expression of miR-133a in HCT116 and LoVo cellswere measured by qRT-PCR. B,miR-133a significantly inhibitedHCT116 and LoVo cells colonyformation. Top, representativeimages of the colony in HCT116and LoVo cells transfected withmiR-133a or miR-Ctrl. Quantitativeanalysis of colony numbers isshown in the bottom (n¼3,mean�SD). C and D, miR-133asignificantly inhibited HCT116 andLoVo cell viability. E, HCT116intratumoral injected with miR-133a precursor showed reducedtumor growth rate compares withmiR-Ctrl injection. Tumor volumesare shown in the left. Data aremean� SD (n ¼ 4 per group). Arrowsindicated the time of miRNAinjection. Images of the tumors areshown in the right. F, tumor growthcurve of HCT116-miR-133a cells innude mice compared withHCT116-miR-Ctrl cells. Data aremean � SD (n ¼ 4/group). Imagesof the tumors are shown in the right.






miR-133a and p21-Luc also increased p21 luciferase report-er activities by 1.9-fold inHCT116 (Fig. 3F; P < 0.001) and1.8-fold in LoVo (Fig. 3G; P< 0.05).Moreover, we depletedp53 by siRNA (si-p53) and found that the enhanced p21expression in miR-133a-transfected cells was completelyabrogated in both HCT116 and LoVo cells (Fig. 3H),confirming that p53 is required for the miR-133a–mediatedinduction of p21. In addition, knockdown of p53 alsoantagonized the growth-inhibitory effect of miR-133a (Fig.

3I and J). These data suggested that miR-133a suppressescell growth, at least in part, by upregulation of p53 and itsdownstream effector p21.

miR-133a inhibits RFFL by direct interaction with its30UTRWe then searched for the direct target of miR-133a based

on the following criteria: the target should show oncogenicproperty and regulate the p53/p21 signaling pathway.

Figure 3. miR-133a positively regulated p53/p21 pathway. A, representative images of the G0–G1 phase arrest by miR-133a overexpression. Cell-cycledistribution was analyzed by fluorescence-activated cell sorting after G2 trapping assay. Briefly, cells were transfected with miR-133a or miR-Ctrl.Nocodazole (25 ng/mL) was added 48 hours after transfection for another 14 hours. B, miR-133a caused accumulation of cells at G0–G1 phase. C, miR-133aincreased the protein level of p21 and p53 in HCT116 and LoVo cells. c-myc protein was unchanged after treated with miR-133a in both cell lines. D,ectopic expression of miR-133a induced p21 mRNA expression in HCT116 and LoVo. E, ectopic expression of miR-133a showed no significant effectson p53mRNA level in HCT116 and LoVo. F, ectopic expression of miR-133a activated p53/p21 pathway in HCT116 cells. G, ectopic expression of miR-133aactivated p53/p21 pathway in LoVo cells. p53-binding activities and p21 promoter transcription activities were determined by dual luciferase assaysystem at 48 hours after transfection (n¼ 3,mean�SD). H, si-p53 antagonizedmiR-133a–mediated p21 induction. HCT116 and LoVo cells were transfectedwith the miR-133a, siRNA against p53, or the corresponding RNA control in the indicated combination for 24 hours. The protein levels of the indicatedgene were determined by Western blotting. I and J, si-p53 antagonized miR-133a–mediated growth-inhibitory effect in MTT assay. HCT116 (I) and LoVo (J)cells were transfected with the miR-133a, siRNA against p53, or the corresponding RNA control in the indicated combination. NS, not significant.

Dong et al.





Among the top 200 targets of miR-133a predicted by onlinealgorithms (Targetscan, mirSVR), Ring finger, and FYVE-like domain-containing E3 ubiquitin protein ligase (RFFL)shows a high prediction score and also fit our hypothesis.RFFL contains an evolutionarily conserved binding site formiR-133a and was also reported to attenuate p53 signalingtransduction (8–10). To test whether RFFL is targetedby miR-133a in vitro, we attached the predicted RFFL30UTR-binding site, as well as its mutant form to fireflyluciferase reporter gene, respectively (Fig. 4A). HCT116cells transfected with miR-133a repressed wild-type RFFL-30UTR reporter activity for about 49% compared withthe cells cotransfected with empty vector (P < 0.01). Onthe other hand, miR-133a had no inhibition effect on themutant RFFL-30UTR reporter activity (Fig. 4B), indictingthe direct regulation of miR-133a in the 30UTR of RFFLmRNA.To determine whether these finding reflect the regula-

tion of endogenous RFFL by miR-133a, we transientlyreintroduced miR-133a precursor into 3 colon cancer celllines HCT116, HT29, and LoVo. Ectopic expression ofmiR-133a remarkably reduced RFFL protein expressionfor 35%, 55%, and 40% of the control group in the 3 celllines (Fig. 4C), suggesting that RFFL is a bona fide targetof miR-133a.

RFFL suppresses p53 and promotes cancer cell growthTo validate the effect of RFFL on p53 expression in colon

cancer, RFFL was transiently transfected inHCT116. Over-

expression of RFFL in HCT116 significantly suppressedp53 protein expression when compared with the controlvector (Fig. 5A). The effect of RFFL on p53 was furtherverified by RFFL knockdown experiment. Knockdown ofRFFL led to a 48% to 60% decrease of RFFL mRNAexpression in HCT116. Silencing of RFFL induced p53and its downstream target p21 protein expression and causedabout 5% increase in G0–G1 phase arrest (Fig. 5B and E).Moreover, RFFL-depleted cells exhibited impaired cellgrowth compared with control cells as revealed by MTTassay and colony formation assay (Fig. 5C). The coloniesformed by RFFL knockdown in HCT116 cells were fewerthan the control cells (P < 0.05; Fig. 5D). We also inves-tigated the level of RFFL protein in 18 randomly selectedpatients from the validation cohort. Our data showed thatRFFL protein was increased in 7 of 18 (38.9%) tumors.Exon sequencing of p53 gene on these 18 CRC tumorsamples was conducted, and the results showed that all 7RFFL-upregulated CRC tumor samples carrying wild-typep53 (data not shown).

miR-133a sensitizes CRC cells to chemotherapeuticdrugs treatmentp53 status was associated with the cell responses to

chemotherapeutic agents. It is thus possible that restorationof miR-133a would render cancer cell sensitivity to chemo-therapeutic agents. To test this hypothesis, we exposed cellsto doxorubicin (a commonly used DNA damage agent) oroxaliplatin (a first-line drug for advanced CRC treatment),

Figure 4. Analysis of RFFL asa miR-133a target gene. A,conservation of RFFL 30UTR-binding site for miR-133a amongdifferent species. Six of 8 basescorresponding to the miR-133aseed region were mutated in theassay. The conserved miR-133abinding sequence or its mutatedform was inserted into C-terminalof the luciferase gene to generatepMIR-RFFL-30UTR or pMIR-RFFLmut-30UTR, respectively. B,miR-133a targeted the wild-typebut not the mutant 30UTR of RFFL.miR-133a was concurrentlytransfected with pMIR-REPORT(the empty vector) or reportersconstructed containing either wild-type or mutant RFFL 30UTR intoHCT116. The luciferase activityassay was repeated 3 times intriplicate. C, miR-133a decreasedthe RFFL protein level in HCT116,HT29 and LoVo cells, asdetermined by Western blotting.The RFFL protein band intensitieswere quantified and normalizedto glyceraldehyde-3-phosphatedehydrogenase (GAPDH)intensities (top, n¼ 3, mean� SD).






respectively, for 48 hours after miR-133a transfection. Asexpected, cell viability was significantly reduced in miR-133a–transfected HCT116 upon exposure to doxorubicin(P < 0.001) or oxaliplatin (P < 0.05; Fig. 6A). Colonyformation was significantly suppressed when HCT116 cellstransfected with miR-133a after exposure to doxorubicin(P < 0.001) or oxaliplatin (P < 0.001), respectively (Fig. 6B).We further quantified the apoptotic cell by Annexin V/7-AAD double stain and revealed that overexpression of miR-133a sensitized HCT116 to chemotherapeutic drug–induced apoptosis, with about 23% increase in early apo-ptotic population for doxorubicin (40% vs. 63%, P < 0.01)and about 6% increase for oxaliplatin (17% vs. 23%, P <0.01), respectively (Fig. 6C). No significant differences werefound in the late apoptosis. These data suggested that theoverexpression of miR-133a had proapoptotic effects whichsensitized colon cancer cell to chemotherapeutic treatments.

DiscussionHuman miR-133a is encoded by 2 distinct genes, miR-

133a-1 and miR-133a-2, which are processed into an iden-ticalmature sequence (http://www.mirbase.org). miR-133a-1 and miR-133a-2 are embedded in the MIB1 gene onchromosome 18 andC20orf166 on chromosome 20, respec-tively. The sequence of miR-133b differs from miR-133a

with a single nucleotide on the 30 flank. However, the Ctvalues of miR-133b in normal tissues were significantlylower than that of miR-133a (data not shown), suggestingthat miR-133a is the dominant form of miR-133 in normalcolorectal tissue. Epigenetic silencing of miRNAs withtumor suppressor features by CpG island hypermethylationis a common hallmark of human tumors (11).Most recently,it was documented that hypermethylation of the CpGislands upstream of miR-1-133a might be involved in thedownregulation of miR-133a in CRCs (12). In the presentstudy, we showed that miR-133a was significantly down-regulated in primary CRC tissues using microdissectiontechnique.The effect of miR-133a in colon cancer was therefore

examined in vitro and in vivo. To avoid the artificialinterference, we restored miR-133a in those cell linescloser to the physiologic level. Ectopic expression ofmiR-133a in these silenced colon cancer cells significantlyinhibited cell viability and clonogenic survival. Increase inG0–G1 phase arrest was shown to contribute to the cellgrowth inhibition by miR-133a. To assure the physiologicrelevance, we used 2 xenograft models to study thefunction of miR-133a. Our study clearly showed thatintratumoral injection of either miR-133a or miR-133astable expression in HCT116 cells significantly inhibited

Figure 5. Effects of RFFL onp53/p21 regulation and tumorcell growth. A, ectopic expressionof RFFL mRNA (top) andprotein (bottom) in HCT116were evidenced by qRT-PCRand Western blot analysis.Overexpression of RFFLsignificantly reduced theendogenous level of p53 inHCT116. B, knockdown of RFFLincreases p53andp21protein levelin HCT116 cells. RFFL knockdownefficiency was determined by qRT-PCR (top). The levels of RFFL, p53,and its downstream target p21were assessed by immunoblottingafter silencing endogenous RFFL(bottom). C, knockdown of RFFLsignificantly inhibits cell viability inHCT116. shRFFL was a pool ofshRFFL1 and shRFFL2. D,knockdown of RFFL inhibitsHCT116 colony formation. Therepresentative images of colonyformation are shown. E,knockdown of RFFL inducedHCT116 cell-cycle arrest at G0–G1

phase. Cell-cycle distribution wasanalyzed by fluorescence-activated cell sorting after G2

trapping. Quantitative analysis ofG1 phase was shown in the right(n ¼ 3, mean � SD).

Dong et al.





CRC xenograft tumor growth. Collectively, these dataindicated for the first time that miR-133a function astumor suppressor in CRCs and mediated the growth-suppressing effects by negatively governing cell-cycle pro-gression. Therefore, miR-133a would contribute to thedepressed cell proliferation in colon cancer cells.G0–G1 arrest by miR-133a re-expression was shown to be

associated with the induction of p53 and its downstreamtarget p21. p53 is essential for the miR-133a–mediatedinduction of p21 based on the following evidences: p21was found to be markedly upregulated in HCT116 andLoVo at both protein and mRNA level by miR-133a.Moreover, we showed that the re-expression of miR-133astrongly increased p53 protein level, enhanced p53 responseelement binding activity, and the p21 promoter reporteractivity, which suggests that miR-133a positively modulatesthe p53/p21 signaling pathway. Finally, silencing of p53abrogated miR-133a–mediated induction of p21 andreversed the growth inhibition property ofmiR-133a. Takentogether, these findings indicate that miR-133a is a novelcomponent in the p53 tumor suppressor network andpositively regulates p53/p21 signaling pathway in coloncancer. Inhibition of miR-133a will impair the p53 safe-guard function and allow tumor growth in a wild-type p53background.

Having observed substantial suppression of colon cancercell growth by miR-133a through mediating p53/p21 sig-naling, we carried out experiments to test the potential targetfor miR-133a. Among the different putative targets for miR-133a predicted using miRNA target prediction algorithms,RFFL is recognized as the most promising target as othershave reported that p53 and phospho-p53 were the maintargets of RFFL (9). In the present work, we revealed a directinteraction that miR-133a negatively regulating RFFL pro-tein in colon cancer cell lines as evidenced by luciferaseactivity assay and by Western blotting. Re-expression ofmiR-133a in the miR-133a–low-expressed tumor cellsinduces activated p53/p21 signaling and G0–G1 phasearrest; the same phenotypes can also be exerted by knock-down of RFFL in the cells. Moreover, we showed thatsilencing of RFFL suppressed CRC cell growth. Thesefindings supported the idea that miR-133a suppressed CRCgrowth through directly targeting the potential oncoproteinRFFL and activating the p53/p21 pathway.Recent researches have shed light on the potential use of

miRNAs as noninvasive markers for CRC early diagnosis, aswell as their associationwithCRC clinical phenotypes. Chenand colleagues mentioned that patients with CRC withliver metastasis were associated with lower expression ofmiR-133a (12). Unfortunately, we did not observe such

Figure6. miR-133a sensitizedCRCcells tochemotherapeutic drug treatment. A, cell viability ofmiR-133a–transfectedHCT116wassignificantly reducedupontemporary exposure to doxorubicin and oxaliplatin (n ¼ 3, mean � SD). B, miR-133a significantly reduced colony formation ability when cells weretemporarily exposed to chemotherapeutic agents. Representative images of colony formation in HCT116 were shown. Quantitative analysis of colonynumbers was shown in the bottom (n ¼ 3, mean � SD). Briefly, doxorubicin or oxaliplatin was added 24 hours after transfection for another 48 hours,then the supernatant was replaced by fresh McCoy's 5A medium supplemented with 10% FBS cells for another 7 days. C1 and C2, miR-133asensitized HCT116 to doxorubicin and oxaliplatin by increasing early apoptosis. Left, representative fluorescence-activated cell-sorting analysis' imagesof HCT116 transfected with miRNA followed by doxorubicin or mock reagent treatment in the indicated combination. Right, the quantitative analysis of earlyand late apoptotic cells (n ¼ 3, mean � SD). DOXO, doxorubicin; Oxa, oxaliplatin.






correlation between miR-133a and patients with CRCoverall survival and disease-free survival in our cohort.Nonetheless, our finding that miR-133a significantly sen-sitized colon cancer cells to doxorubicin and oxaliplatincould have a beneficial impact of miR-133a on the clinicalpractice and in adjuvant chemotherapy of CRC after surgicalresection.In conclusion, we showed that miR-133a was frequently

downregulated in CRC tissues and colon cancer lines.Restoration of miR-133a in colon cancer cells resulted inthe G0–G1 cell-cycle arrest and therefore suppressed cancercell growth. The tumor growth inhibition by miR-133acould be attributed, at least in part, to miR-133a directlyreducing RFFL translation and activating p53/p21 path-way. Besides, miR-133a could sensitize cancer cell tochemotherapeutic agents, suggesting its potential clinicaltherapeutic implication.

Disclosure of Potential Conflicts of InterestF.K.L. Chan has honoraria from speakers' bureau fromEisai. No potential conflicts

of interest were disclosed by the other authors.

Authors' ContributionsConception and design: Y.J. Dong, C.W. Wu, F.K.L. Chan, S.S.M. Ng, J. YuDevelopment of methodology: J. Zhao, S.S.M. NgAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): Y.J. Dong, J. Zhao, X. Liu, W. Kang, S.S.M. Ng, J. YuAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis): Y.J. Dong, J. Zhao, L. Zhang, J.J.Y. Sung, S.S.M. NgWriting, review, and/or revision of the manuscript: Y.J. Dong, C.W. Wu, F.K.L.Chan, S.S.M. Ng, J. YuAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): J. Zhao, L. Zhang, W.W. LeungStudy supervision: F.K.L. Chan, J.J.Y. Sung, S.S.M. Ng, J. Yu

Grant SupportThis work was supported by the National High-tech R&D Program of China 863

program (2012AA02A203, 2012AA02A506), the Innovation and Technology FundHong Kong (ITS/276/11), Shenzhen Municipal Science and Technology R & Dfunding of basic research programs(JCYJ20120619152326450), Major projects ofNational Science and Technology (2011ZX09307-001-05), and The Chinese Uni-versity of Hong Kong direct grant (2010.1.085).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Received January 31, 2013; revised April 9, 2013; accepted May 1, 2013;published OnlineFirst May 30, 2013.

References1. JemalA,BrayF,CenterMM, Ferlay J,WardE, FormanD.Global cancer

statistics. CA Cancer J Clin 2011;61:69–90.2. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and func-

tion. Cell 2004;116:281–97.3. FriedmanRC, FarhKK, BurgeCB,Bartel DP.MostmammalianmRNAs

are conserved targets of microRNAs. Genome Res 2009;19:92–105.4. Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D, et al.

MicroRNA expression profiles classify human cancers. Nature 2005;435:834–8.

5. YangY, Zhou L, Lu L,WangL, Li X, JiangP, et al. A novelmiR-193a-5p-YY1-APC regulatory axis in human endometrioid endometrial adeno-carcinoma. Oncogene. 2012 Aug 20. [Epub ahead of print].

6. Chen WS, Leung CM, Pan HW, Hu LY, Li SC, Ho MR, et al. The role ofmicroRNA-1 and microRNA-133 in skeletal muscle proliferation anddifferentiation. Nat Genet 2006;38:228–33.

7. Georges SA, Biery MC, Kim SY, Schelter JM, Guo J, Chang AN, et al.Coordinated regulation of cell cycle transcripts by p53-induciblemicroRNAs, miR-192 and miR-215. Cancer Res 2008;68:10105–12.

8. McDonald ER III, El-Deiry WS. Suppression of caspase-8- and -10-associated RING proteins results in sensitization to death ligands andinhibition of tumor cell growth. Proc Natl Acad Sci U S A 2004;101:6170–5.

9. Yang W, Rozan LM, McDonald ER III, Navaraj A, Liu JJ, Matthew EM,et al. CARPs are ubiquitin ligases that promote MDM2-independentp53 and phospho-p53ser20 degradation. J Biol Chem 2007;282:3273–81.

10. YangW,DickerDT,Chen J, El-DeiryWS.CARPsenhancep53 turnoverby degrading 14-3-3sigma and stabilizing MDM2. Cell Cycle2008;7:670–82.

11. Lujambio A, Calin GA, Villanueva A, Ropero S, S�anchez-C�espedesM, Blanco D, et al. A microRNA DNA methylation signature forhuman cancer metastasis. Proc Natl Acad Sci U S A 2008;105:13556–61.

12. ChenWS, Leung CM, Pan HW, Hu LY, Li SC, HoMR, et al. Silencing ofmiR-1-1 andmiR-133a-2 cluster expression byDNAhypermethylationin colorectal cancer. Oncol Rep 2012;28:1069–76.

Dong et al.





2013;11:1051-1060. Published OnlineFirst May 30, 2013.Mol Cancer Res Yujuan Dong, Junhong Zhao, Chung-Wah Wu, et al. Tumor Suppressor Functions of miR-133a in Colorectal Cancer

Updated version

10.1158/1541-7786.MCR-13-0061doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2013/05/30/1541-7786.MCR-13-0061.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/11/9/1051.full#ref-list-1

This article cites 11 articles, 5 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/11/9/1051.full#related-urls

This article has been cited by 4 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://mcr.aacrjournals.org/content/11/9/1051To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-13-0061

http://mcr.aacrjournals.org/content/suppl/2013/05/30/1541-7786.MCR-13-0061.DC1

http://mcr.aacrjournals.org/content/11/9/1051.full#ref-list-1

http://mcr.aacrjournals.org/content/11/9/1051.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/11/9/1051


Documents

Tumor Suppressor Functions of miR-133a in Colorectal …mcr.aacrjournals.org/content/molcanres/11/9/1051.full.pdf · Tumor Suppressor Functions of miR-133a in Colorectal Cancer Yujuan