The tumor-suppressive function of UNC5D and its repressed expression in renal cell carcinoma

Human Cancer Biology

The Tumor-Suppressive Function of UNC5D and ItsRepressed Expression in Renal Cell Carcinoma

DanLu1, DongDong1, YuZhou1,Min Lu2, Xue-WenPang1, YanLi1, Xiao-Jun Tian3, YuZhang1, and JunZhang1

AbstractPurpose: As a newly added member of the UNC5H receptors, the function of UNC5D/H4 in tumor-

igenesis remains poorly defined. The aim of this study was to examine the expression ofUNC5D in primary

renal cell carcinomas (RCC), analyze themechanisms responsible for its downregulation in RCC, and assess

its functional relevance to tumor growth and migration.

Experimental Design: Forty-four paired primary RCCs and corresponding adjacent noncancerous

tissues were collected. The mRNA and protein expression level of UNC5D was assessed by reverse

transcriptase-PCR, real-time PCR, or immunohistochemistry. Epigenetic alterations in UNC5D promoter

and LOH in the UNC5D locus were also analyzed. Ectopic expression of UNC5D in renal cancer cells with

silenced expression of UNC5D was used for analysis of the biologic functions of UNC5D.

Results:UNC5D expression was attenuated in multiple carcinoma cell lines including renal cancer cells.

Similar reductionwas also observed inprimary RCC tissues as comparedwith paired adjacent noncancerous

tissues. Methylation-specific PCR showed hypermethylation in UNC5D promoter in a significant propor-

tion (18 of 44) of tumor tissue (40.9%). LOH of UNC5D was observed in 13 of 44 patients with RCCs

(29.5%). Restoration of UNC5D expression in renal cancer cells significantly inhibited cell proliferation,

anchorage-dependent and -independent growth, as well as migration and invasion, whereas knockdown of

UNC5D promoted cell growth. Furthermore, ectopic expression ofUNC5D induced G2–M cell-cycle arrest.

Conclusions:UNC5D is a functional tumor suppressor that is frequently downregulated in RCCs due to

promoter hypermethylation and LOH. Clin Cancer Res; 19(11); 2883–92. �2013 AACR.

IntroductionKidney cancer is among the 10 most common cancers

worldwide (1). Most (>90%) kidney cancers are renal cellcarcinoma (RCC) that originate from the renal parenchyma(2). Clear cell carcinoma, papillary carcinoma, and chro-mophobe carcinoma constitute the majority of RCCs (3).Patients with RCCs are usually asymptotic in early stages,and at the time of diagnosis, a quarter of patients with RCCshave locally advanced or metastatic disease. One third ofpatients who undergo resection for local disease have risk ofcancer relapse (4). To date, the mechanisms of RCC onco-

genesis remain elusive. The identification of novel genesfunctionally involved in renal cancer development andprogression may help to find potential diagnostic andtherapeutic targets.

Cancer is a disease initiated and driven by the clonalselection of cells with either inherited or acquired genetic orepigenetic alterations of key tumor suppressor genes (TSG)and oncogenes that confer growth advantage (5). It is well-recognized that loss of function of TSGs may lead toneoplastic changes. TSGs can be inactivated by both geneticand epigenetic mechanisms, which include LOH, pointmutation, homozygous deletion, and promoter hyper-methylation (6, 7). Until now, the most common geneticalteration in sporadic clear cell RCCs is inactivation of theTSGVonHippel-Lindau (VHL) by chromosome3pdeletion(8). Recently, increasing numbers of TSGs associated withepigenetic alterations have been identified in RCCs, whichimply that epigenetic regulation of TSGs may be an impor-tant mechanism for tumorigenesis in RCCs. Genes whichhave been found to be aberrantly methylated in RCCs atdifferent frequencies include VHL, RASSF1A, p16, Timp-3,E-cadherin, b-catenin, SFRP1, SFRP2, SFRP4, SFRP5, DAL-1,COLIA1, KRT19, UCHL1, and UNC5C (9–15).

Netrin-1 and its receptors are expressed extensively inmultiple tissues and participate in an array of cell processes(16–18). An important feature of Netrin-1 receptors is thatthey function as "dependence receptors." When unbound

Authors' Affiliations: 1Department of Immunology, Key Laboratory ofMedical Immunology (Ministry of Health), 2Department of Pathology, Schoolof Basic Medical Sciences, Peking University Health Science Center; and3Department of Urology, Peking University Third Hospital, Beijing, China

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

D. Lu and D. Dong contributed equally to this work.

Corresponding Authors: Jun Zhang, Department of Immunology, PekingUniversity Health Science Center, Beijing 100191, China. Phone: 86-10-82805649; Fax: 86-10-82801436; E-mail: [email protected]; or YuZhang, Department of Immunology, Peking University Health ScienceCenter, Beijing 100191, China. Phone: 86-10-82805055; Fax: 86-10-82801436; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-12-2978

�2013 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 2883

by the ligandNetrin-1, these receptors induce apoptosis butinhibit apoptosis and provide a survival signal in the pres-ence of the ligand. Through such a mechanism of action,these receptors may act as tumor suppressors (19). In linewith this speculation, the receptor DCC shows reduced orlost expression in a variety of cancers (20). Another group ofNetrin-1 receptors referred to as UNC5H (1-3/A-C) are alsodownregulated in a number of tumors. Overexpression ofthese receptors inhibits tumor cell anchorage-independentgrowth and invasion (21–23).

UNC5D/H4 is themost recently identifiedmember of theUNC5H receptors (24). One study reported that UNC5D isinduced during DNA damage–mediated apoptosis and is adirect transcriptional target of p53 (25). Several otherstudies showed that rearrangement of 8p with loss of distal8p, where UNC5D is located, is one of the most frequentgenomic events in common epithelial cancers (26, 27).Nevertheless, the function of UNC5D in tumorigenesisremains poorly defined.

The present study was focused on the potential role ofUNC5D in the development of RCCs. We first examined itsexpression in tumor versus adjacent noncancerous tissueand explored themechanisms underlying its suppression intumor tissue. Subsequently, the implication of UNC5Ddownregulation in RCC was analyzed by monitoringaltered cell behaviors following restoration of its expressionin otherwise silenced cells. Data thus acquired support atumor-suppressive function of UNC5D in RCCs.

Materials and MethodsCell lines, tumor specimens, and 5-aza-dC treatment

A series of cancer cell lines were used for this study,including 5 RCCs (786-O, A498, ACHN, Caki-1, and Os-RC-2), 2 bladder carcinoma, 5 hepatoma, 5 lung cancer, 4

gastrointestinal cancer, 5 leukemia, 1 prostate carcinoma, 1breast cancer, 1 cervical carcinoma, 1 ovarian cancer, 2melanoma, and 1 osteosarcoma cell line. Human immor-talized embryonic kidney cell lines HEK293 and HEK293Twere also used. Renal carcinoma and corresponding non-cancerous tissues (n ¼ 44) were obtained from the PekingUniversity Third Hospital (Beijing, China) with patient’sconsents and institutional ethics approval. All of the speci-mens were pathologically confirmed. For demethylation,cell lines were treated with 10 mmol/L of 5-aza-20-deoxycy-tidine (5-aza-dC, Sigma-Aldrich) for 3 days with exchangeof reagents and medium every 24 hours.

Reverse transcriptase PCR and real-time PCRA human normal tissue cDNA panel was purchased from

Clontech. Conventional PCR and quantitative real-timePCR were carried out as previously described (15). Allprimers used in this study are shown in SupplementaryTable S1.

Immunohistochemistry and scoringThe tissuemicroarray (ShanghaiOutdoBiotechCo., Ltd.)

of 30 paired tumor and noncancerous tissue was incubatedwith anti-UNC5D antibody (Santa Cruz) at 1:100 dilutionovernight at 4�C. The scoring method which combinedintensity and percentage of positivity was previously de-scribed (28). Slides were scored by 3 reviewers, and dis-crepancies were resolved by a urological pathologist. Extentand intensity measures for each core were combined asweak (intensity weak), moderate (intensity moderate), andstrong (intensity strong).

Methylation-specific PCR and bisulfate genomicsequencing

Methylation-specific PCR (MSP) and bisulfate genomicsequencing (BGS) analysis were conducted as describedpreviously (29, 30). Amplified BGS PCR products werecloned into the pGEM-T Easy vector (Promega), and 10random clones from each sample were sequenced.

LOH analysis with microsatellite markersA set of 3 polymorphicmicrosatellite markers (D8S1750,

D8S505, and D8S1803) were used to determine LOH atchromosome 8p12. gDNA from paired tissue samplesand cell lines was amplified by PCR using fluorescentlylabeled primers for the indicated polymorphic microsatel-lite markers.

Transfection, infection, and Western blottingThe expression vectors (pCCL-Flag-UNC5D-FL, pCCL-

Flag-UNC5D-DDD, pCCL-Flag-UNC5D-DZU5) with thefull-length or deletedmutants ofUNC5Dwere constructed.Adenoviruses were packaged by the Vector Gene Technol-ogy Company Limited. Lentiviral particles were producedby transfecting HEK293T cells with the pCCL vector and thepackaging vectors psPAX2andpMD2.G. The786-OorA498cells were infected with adenoviruses or lentiviruses. ForWestern blotting, primary antibodies used include anti-

Translational RelevanceNetrin-1 receptors—UNC5H family members play

important roles in tumorigenesis. In this study,we foundthat UNC5D/H4, the newly added member of theUNC5H family, widely expressed in normal tissues, wasfrequently absent or attenuated in cancer cell lines andprimary renal cell carcinoma (RCC). Promoter CpGmethylation was responsible for this. Pharmacologicdemethylation restored UNC5D expression along withconcomitant promoter demethylation.UNC5Dmethyl-ation was also frequently detected in primary RCCs.LOH also contributed to its downregulation in RCCs.Ectopic UNC5D expression in silenced renal cancer cellline dramatically inhibited the growth of renal cancercells through promoting G2–M cell-cycle arrest. EctopicUNC5D expression also inhibited the migration andinvasion of renal cancer cells. These findings raise thepossibility that UNC5D might be a novel tumor sup-pressor in RCCs and serve as a potential diagnostic andtherapeutic target for RCCs in the future.

Lu et al.

Clin Cancer Res; 19(11) June 1, 2013 Clinical Cancer Research2884

GAPDH (Proteintech Group), anti-Flag (Sigma-Aldrich),and anti-cyclin B1, cyclin A, cyclin-D1 (Santa Cruz).

Knockdown of UNC5DThe pGPU6-GFP-neo shRNA expression vectors (Gene-

Pharma Co., Ltd.) against UNC5D or a control vector wastransfected into PC3 cells by jetPRIME (Polyplus-Transfec-tion Inc.). Cells were selected in 800 ng/mL G418 (Sigma-Aldrich). The short hairpin RNA (shRNA) target sequencesused in this study are shown in Supplementary Table S1.

Cell proliferationCell proliferationwas analyzed using Cell Counting Kit-8

(CCK-8; Dojindo Laboratories), according to the manufac-turer’s instructions. Absorbance at 450nmwasmeasuredona microplate reader at the indicated time points.

Cell-cycle analysis786-O cells were infected with adenoviruses encoding

UNC5D or EGFP alone, and nocodazole (300 ng/mL,Sigma-Aldrich) was added to the culture 24 hours afterinfection. After incubation for another 16 hours, cellswere harvested and replated. At indicated time points,cells were collected and fixed with ice-cold ethanol. Afterwashing with PBS, cell pellet was resuspended in PBScontaining 10 mg/mL propidium iodide (PI; Sigma-Aldrich) and 500 mg/mL RNase A (Sigma-Aldrich) andincubated at 37�C for 30 minutes. Samples were thenanalyzed on a BD FACSCalibur.

Colony formation assayTwo thousand infected cells were plated in each well and

maintained in medium for 2 weeks. Colonies were fixedwith precooled methanol, and colonies were then stainedwith 0.5% (w/v) crystal violet and counted. For soft agarcolony formation assay, a total of 500 cells weremixed with0.35% agarose and plated in each well on top of a layer of0.5% agarose. After 2 weeks, colonies were stained withcrystal violet. Colonies larger than 20 mm in diameter werecounted.

Wound-healing assayCell motility were determined by measuring the move-

ment of cells to close an artificial wound. Cells werewounded with a 200 mL pipette tip, washed with PBS, andincubated in medium containing 2% FBS. The distancetraveled by cells was monitored by phase-contrast micros-copy (Olympus) at indicated time points.

Cell migration and invasion assayFor migration assay, infected cells were seeded into the

upper chamber of a Transwell with a fibronectin-coatedfilter (8-mm pore size, Corning Life Sciences). The bottomchamber contained medium supplemented with 10% FBS.After 14-hour incubation, cells adherent to the upper sur-faceof thefilterwere removedusing a cotton swaband thoseattached to the bottom of themembranes were stained withcrystal violet following fixation with methanol. Cell num-

ber was counted in 6 randomly chosen fields. Cell invasionassay was conducted essentially as the migration assay,except that the Transwell insert was coated with 30 mg ofMatrigel (BD Biosciences) and the invasion time wasextended to 24 hours.

Statistical analysisQuantitative data are presented as individual data plots

or as mean � SD. Statistical analyses were done with SPSS,version 13.0 (SPSS, Inc.). Comparisons of UNC5D geneexpression levels between paired tumor and adjacentnonmalignant tissue samples were conducted using theWilcoxon signed rank test. Differences between two inde-pendent groups were analyzed by the Student t test. Thepaired t test was done for comparison of differences withinpairs. The c2 test was used to calculate differences in thepatient’s age, gender, tumor stage, histologic grade, LOH,and methylation status between paired tumor and adjacentnon-malignant tissue samples. P < 0.05 was consideredsignificant.

ResultsReduced expression of UNC5D in tumor cell lines

To elucidate the potential role of UNC5D in tumorigen-esis, bioinformatics analysis of UNC5D expression inhuman cancers was first conducted using an ICGC data set(www.icgc.org; ref. 31). Compared with normal tissues,UNC5D was downregulated in a variety of cancers, includ-ing those derived from breast, colon, kidney, lung, andrectum (Fig. 1A). Such a pattern was subsequently verifiedusing semiquantitative reverse transcription PCR (RT-PCR).As shown in Fig. 1B, UNC5D was broadly expressed innormal human tissues, with relatively high levels in brain,kidney, prostate, testis, small intestine, and colon tissue. Incontrast, lack or weak ofUNC5D expressionwas detected ina number of tumor lines originated from kidney, lung,gastrointestinal, liver carcinoma, or leukemia (Fig. 1C).Notably, UNC5D expression was maintained in immortal-ized epithelial cell lines such as HEK293 (Fig. 1C). Takentogether, these data indicate a general reduction inUNC5Dexpression in association with tumorigenesis.

Frequent inactivation ofUNC5DbyCpGmethylation intumor cell lines

Bioinformatics analysis revealed that the UNC5D geneharbors a typical DNA sequence fulfilling the criteria for aCpG island in a region covering the promoter, exon 1, andthe beginning of intron 1 (Fig. 2A). Given the importance ofepigeneticmechanisms in the regulationof gene expression,we sought to determine whether DNA methylation con-tributed to the silenced expression of UNC5D in cancercells. Indeed, MSP analysis showed that the UNC5D pro-moter was methylated in 2 of 4 RCC cell lines, 2 of 2 coloncancer cell lines, 2 of 3 lung cancer cell lines, 1 of 1 cervicalcancer cell line, and 1 of 2 leukemia cell lines (Fig. 2B). Incontrast, no methylation of the UNC5D promoter wasfound in the immortalized epithelial cell line HEK293. Wealso confirmed UNC5D methylation by high-resolution

UNC5D with Tumor Suppressor Activities in RCC

www.aacrjournals.org Clin Cancer Res; 19(11) June 1, 2013 2885

BGS of CpG sites within the CpG island, including 43 CpGsites analyzed by MSP. Again, high levels of DNA methyl-ation were detected in 786-O, A498, SW480, SW620, andHeLa cells but not in HEK293 cells (Fig. 2B). To show thatCpG methylation is functionally associated with UNC5Dsilence, several carcinoma cell lines were treated with theDNA demethylation reagent 5-aza-dC. Such treatmentrestored UNC5D expression (Fig. 2C, left), which wasaccompanied by a decrease in methylated promoter allelesand an increase in unmethylated alleles. Furthermore, thedemethylation was confirmed by BGS analysis (Fig. 2C,right). These results support that DNA methylation consti-tutes one of the major mechanisms responsible for thedownregulation or inactivation of UNC5D.

Loss or reduced expression of UNC5D in primary RCCThe study was extended to primary tumors, with a focus

on RCCs. RT-PCR and quantitative real-time PCR analysisrevealed that UNC5D mRNA expression was markedlyreduced or inactivated in a large proportion of RCC

samples (Fig. 3A and B). Specifically, 32 of 44 tumortissues showed more than 2-fold reduced expression incomparison with paired noncancerous tissue. Amongthem, 18 showed more than 10-fold reduced expression.We further examined UNC5D protein expression in atotal of 30 paired tissue sections by immunohistochem-istry. Strong UNC5D staining was shown in noncancer-ous renal tissue, mainly in the epithelium. On the con-trary, UNC5D signal was absent or barely detectable incancer tissue (Fig. 3C). The immunohistochemical stain-ing was further scored by taking into consideration ofboth staining intensity and percentage of cells showingpositive staining. The percentage of each group withstrong, weak, or moderate expression level was calculatedof noncancerous versus paired cancer tissue. As shownin Fig. 3D, compared with noncancerous tissue, UNC5Dprotein was dramatically downregulated in RCC tissue.We also analyzed the association between clinicopatho-logic features and the expression of UNC5D and observedno correlation between loss or reduction of expression of

Figure 1. UND5D expression isdownregulated in multiplecarcinoma cell lines. A, ICGC dataanalysis of UNC5D expression inmultiple cancer tissues. x-axis,BIC, breast invasive carcinoma;CA, colon adenocarcinoma; GM,glioblastoma multiforme; KRCCC,renal clear cell carcinoma; LA, lungadenocarcinoma; LSCC, lungsquamous cell carcinoma; PC,pancreatic cancer; RA, rectaladenocarcinoma. y-axis, log2 scalenormalized data for mRNAexpression. B, UNC5D expressionprofile in human adult tissues. C,RT-PCR analysis of UNC5Dexpression pattern in multiplecancer cell lines.

Lu et al.


UNC5D and the parameters including age, gender, tumorstage, and histologic grade (data not shown).

Frequent UNC5D promoter methylation and LOH inprimary RCCWe next investigated whether promoter methylation also

contributed to the attenuated or lost expression of UNC5D

in primaryRCCs. Among the 44paired samples, 18 (40.9%)showed markedly increased methylation of the UNC5Dpromoter in tumor in contrast to that only 5 samplesshowed very weak methylation in the adjacent noncancer-ous tissue (Fig. 4A). BGS analysis confirmed that theUNC5D promoter was markedly methylated in primaryRCCs but not in paired noncancerous renal tissues (Fig. 4B).

Figure 2. Promoter methylation contributes to the silence of UNC5D in multiple cancer cell lines. A, a CpG island spans the promoter, exon 1, and intron 1 ofUNC5D. Horizontal bars, CpG sites; primers for methylation analysis, MSP primers and BGS primers are indicated. Curved arrow, transcription start site. B,UNC5Dpromoter is hypermethylated inmultiple cancer cell lines. Left,MSP results ofUNC5Dpromoter inmultiple cell lines; right, detailedBGSanalysis of theUNC5D promoter in multiple cell lines. Circles, CpG sites analyzed; row of circles, an individual promoter allele that was cloned, randomly selected, andsequenced; filled circle, methylated CpG site; open circle, unmethylated CpG site. C, pharmacologic demethylation with 5-aza-dC restored UNC5Dexpression in methylated and silenced carcinoma cell lines. Left, RT-PCR analysis of UNC5DmRNA expression; right, detailed BGS analysis of the UNC5Dpromoter after pharmacologic demethylation.

Figure 3. UNC5D is attenuated orsilenced in primary RCCs. A,UNC5DmRNA expression in 4representative pairs of RCC andadjacent noncancerous tissue asassessed by RT-PCR. B,comparison of the relativeexpression levels of UNC5D in 44paired renal carcinoma and adjacentnoncancerous tissue as measuredby real-timePCR.C, a representativeresult of immunohistochemicalstaining for UNC5D proteinexpression in paired renal carcinomaand adjacent noncancerous tissue.D, comparison of the relativeprotein levels of UNC5D in 30 pairedrenal carcinoma and adjacentnoncancerous tissue as measuredby IHC. ��, P < 0.001.

Ca.

UNC5D

GAPDH

Adj.

Ca. Adj.

A B

C D

×200 ×200

Ca. Adj. Ca. Adj. Ca. Adj.

Ca. Adj.

Ca. Adj.

StrongModerateWeak

80

60

40

20

0

100

80

60

40

20

0

mR

NA

Ex

pre

ssio

n l

evel

(qR

T-P

CR

, li

near)

Fre

qu

en

cy (

%)



As noted above, promoter methylation may be the maincontributor to the inactivation of UNC5D in RCCs,although possibly there are other reasons for this loss.Human UNC5D is located at chromosome 8p12. Throughthe analysis of chromosomal rearrangement, we found thatin breast cancer therewere 2 types of translocationoccurringin 8p12 where UNC5D is located, which included ampli-con-to-amplicon translocation and inverted orientation(data not shown). Therefore, to search for other mechan-isms that may explain the reduced expression ofUNC5D inRCCs, we determined the frequency of allelic losses in 3markers (D8S1083, D8S505, and D8S1750). Of the 44specimens investigated, 13 (29.5%) cases showed allelicimbalance in at least one of the markers (Fig. 4C and D).Statistical analysis showed no correlation of methylation orLOH with age, gender, tumor stage, or histologic grade.Although it was not statistically significant, a trend in whichmost methylation of UNC5D occurred at tumor stage pT1(Supplementary Table S2) was found. A relatively goodcorrelation between UNC5D expression level and its meth-ylation or LOH status was observed in several randomly

chosen samples (Supplementary Table S3). In addition, inUNC5D-negative cells without methylation modificationsuch asNCI-H460, K562, andOs-RC-2, allelic imbalance inat least one marker was also observed (data not shown).Therefore, LOH may be indeed implicated in the loss ofUNC5D expression in these cell lines.

UNC5D inhibits proliferation and clonogenicity ofRCC lines

The frequent inactivation of UNC5D in cancer cell linesand primary renal carcinoma tissue suggests a potentialrole of UNC5D in tumorigenesis. To test this possibility,786-O and A498, 2 RCC lines which showed no UNC5Dexpression with a heavily methylated UNC5D promoterwere infected with UNC5D expressing viruses and mon-itored for changes in cell behavior. Enforced expression ofUNC5D significantly inhibited the proliferation of both786-O (Fig. 5A) and A498 cells (data not shown). Colonyformation in monolayer culture or soft agar was found tobe dramatically reduced in 786-O cells, compared withmock cells (Fig. 5B and C). Cell-cycle analysis with 786-O

Ca. 1

(45.2%)

Ca. Ca. Ca.

Adj. Adj. Adj.

D8S1083

A

B

D

CD8S505 D8S1750

D8S1083

UNC5D

D8S505

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 41 42 43 4435 36 37 38 39 40

D8S1750

Adj. 1

(4.6%)

Ca. 2

(38.3%)

Adj. 2

(2.3%)

Figure 4. Promoter hypermethylation and LOH contribute to the downregulated expression of UNC5D in primary RCCs. A, UNC5D promotermethylation was assessed by MSP in primary RCCs and adjacent noncancerous tissues. Results from 6 representative pairs are shown. B, detailed BGSanalysis of the UNC5D promoter in 2 representative paired tissues. C, representative results of LOH status on UNC5D in RCC and adjacent noncanceroustissue. Top, the illustration of adjacent noncancerous tissues; bottom, the tumor tissues which present a deletion. D, summary of LOH analysis for 44noncancerous/tumor pairs in the UNC5D locus, showing noninformative (homozygous) markers (white squares), retained markers (shaded squares), andmarkers with LOH (black squares). Each vertical line indicates a single patient and cases having LOH at the UNC5D flanking markers are indicated by �.

Lu et al.


cells revealed a decrease in G0–G1 phase and a concurrentincrease in G2–M phase (Fig. 5D). To more closely exam-ine cell-cycle progression, 786-O cells were synchronizedat G2–M phase with nocodazole. Enforced expression ofUNC5D delayed the re-entry of G2–M cells into a newcycle following removal of nocodazole (Fig. 5D). Con-

sistent with G2–M cell-cycle arrest, cyclin B1 and Aexpression level increased in UNC5D overexpressed cells(Fig. 5E). In contrast, UNC5D-DDD lost the ability toinduce cell-cycle arrest (data not shown) and abolishedthe inhibition on cell proliferation and colony formation(Fig. 5F–H). These data support that the death domain of

Mock

ns

ns

0 24 48Time (h)

72 96

0 24 48Time (h)

72 96

0 24 3612 48 60Time (h)

72 84

UNC5D

Mock

A B

C D

E F

G H

I J

Co

lon

y n

um

ber

UNC5D

Mock

UNC5D

Mock

UNC5D

Cell count

FL2-A

Un G0–G1

G2–MS

0 h 4 h 10 h

Un0 h 4 h 8 h Un0 h 4 h 8 h

Per

cent

age

(%)

Un 0 h 4 h 10 h

Mock UNC5D

Moc

k

UNC5DM

ock

UNC5DM

ock

UNC5DM

ock

UNC5D

IB: ααUNC5D

UNC5DIB: α FLAG

IB: αFLAG

UNC5D-ΔDD

UNC5D-ΔDD

UNC5D-ΔZU5

UNC5D-FL

UNC5D-FL

UNC5D-ΔDD

UNC5D-ΔDD

UNC5D-ΔZU5UNC5D-FL

UNC5D-FL

shRNA-NC

shRNA-1

shRNA-2

shRNA-NCshRNA-1shRNA-2

ns

shRNA-NC shRNA-1 shRNA-2

shRNA-N

C

shRNA-1

shRNA-2

1

1

1

953

875

847

847–936540–645

Extracellular domain Cytoplasmic domain

GAPDHIB: α GAPDH

IB: αGAPDH

Cyclin B1IB: α Cyclin B1

Cyclin D1IB: α Cyclin D1

Cyclin AIB: α Cyclin A

IB: αGAPDH

IB: αUNC5D

IB: αGAPDH

Moc

kUNC5D

OD

.450 n

m

1.8

1.5

1.2

0.9

0.6

0.3

0.0

OD

.450 n

m

1.2

0.9

0.6

0.3

0.0

OD

.450 n

m

1.8

1.5

1.2

0.9

0.6

0.3

800

600

400

200

0

Co

lon

y n

um

ber

900

600

300

0

Co

lon

y n

um

ber

600

400

200

0

Co

lon

y n

um

ber

Mock UNC5D

Mock

Mock

Mock

Mock

UNC5D

-ΔDDUNC5D

-ΔZU5

UNC5D

-FL

Moc

k

UNC5D

80

60

40

20

0

120

100

80

60

40

20

0

Figure 5. UNC5D inhibits renal cancer cell growth. 786-O cells were infected with mock or UNC5D-expressing adenoviruses (A–E) or lentiviruses(F–H) and monitored for growth using various assays. Each assay was repeated at least 3 times. A, cell proliferation was analyzed using CCK-8. Results arepresented as absorbance (OD) at 450 nm for triplicate wells from 1 representative experiment. The inset shows protein expression of UNC5D and GAPDH.B and C, colony formation of 786-O cells inmonolayer culture (B) or in soft agar (C). D,UNC5D inducedG2–Marrest in 786-O cells. Cell cycle was analyzed atdifferent time points after removal of nocodazole in synchronized or unsynchronized 786-O cells by flow cytometry. Representative histograms (left)and the percentage of cells at different phases (right) are shown. E, Western blot analysis of cell cycle hallmarks in synchronized 786-O cells after nocodazoleremoval. F, schematic structure of the full-length and the deletion mutants of UNC5D protein. G&H, UNC5D-DDD abolished the inhibitory function oncell proliferation (G) and the colony formation of 786-O cells in monolayer culture (H). I and J, knockdown of UNC5D expression in PC3 cells increased cellproliferation (I) and colony formation (J). ��, P < 0.01; ��, P < 0.001.



UNC5D is essential for the tumor-suppressive function ofUNC5D. In addition, knockdown of UNC5D expressionin PC3 cells promoted cell proliferation and colonyformation (Fig. 5I and J).

UNC5D inhibits cell migration and invasionWe also explored the impact of UNC5D on another

important aspect of tumorigenesis, migration, and inva-sion. First, scratch wound-healing assay was engaged toassess themotility of 786-O cells.Wound closurewas foundto be retarded for UNC5D overexpressing 786-O cells(Fig. 6A).Next, cellmigrationwas tested using the Transwellassay. Enforced expression of UNC5D led to a 2- to 3-foldreduction in the number of 786-O cells crossing over thefilter (Fig. 6B). Similar inhibitory effect was observed on cellinvasion in a Matrigel assay (Fig. 6C). We also obtainedsimilar results with A498 cells (data not shown). Moreover,UNC5D-DDD abolished its effects on cancer cell migration(Fig. 6D). These data suggest a suppressive function ofUNC5D on tumor cell migration and invasion.

DiscussionUNC5D is a newmember of the UNC5H family. Limited

information is currently available concerning its biologicfunction. In this report, we show thatUNC5D expression isfrequently reduced or lost in renal cancer tissues and in awhole variety of tumor lines.Moreover, enforced expressionof UNC5D in expression-silenced cells inhibits cell growthand migration.

Both epigenetic and genetic mechanisms are implicatedin the downregulation of UNC5D. In cancer lines, UNC5Dsilencing correlated well with hypermethylation of a CpGisland in the promoter region. Moreover, pharmacologicdemethylation was able to restore UNC5D expression,supporting a direct link between promoter methylationand UNC5D downregulation. High levels of promotermethylation were also observed in about 41% of primaryrenal cancer tissues. Previous studies have indicated that,compared with other major malignancies, RCC shows avery low frequency of DNA methylation. It is true for mostof the TSGs such as VHL, p16/CDKN2A, p14/ARF, APC,and UNC5C (9–12). However, there are several tumorsuppressors, including SFRP1, SFRP2 and SEMA3B, whoseexpression is mainly regulated by epigenetic mechanisms(32–34). UNC5D apparently belongs to this category. As alarge proportion of RCC cancer tissues exhibited hyper-methylation in the UNC5D promoter, this epigeneticchange should constitute a major mechanism for the sup-pression or silence of UNC5D in RCCs. In addition, LOHoccurred at theUNC5D locus in approximately 30%of renalcancer tissues. Of note, UNC5D maps to 8p12. Loss ofchromosome arm 8p together with amplification of prox-imal 8p have been commonly identified in a variety ofepithelial cancers including breast, bladder, colon, pancre-atic cancer, and others (26, 27, 35, 36). As such, LOH mayrepresent another important mechanism for the downre-gulation of UNC5D. Most recently, a genome-wide associ-ated study has identified UNC5D as one of candidate genes

0 3 6Time (h)

9 12 15 18 21

Mock

A B

C D

Nu

mb

er

of

mig

rati

ng

cells/f

ield

UNC5D

Mock

UNC5D

Mock

Mock

UNC5D

Mock

UNC5D

0 h 18 h Mock UNC5D

Mock UNC5D UNC5D-ΔDD

UNC5D-ΔDD

UNC5D-FL

UNC5D-FL

Dis

tan

ce

(m

m)

500

400

300

200

100

0

200

150

100

50

0

Nu

mb

er

of

mig

rati

ng

cells/f

ield

250

200

150

100

50

0

Nu

mb

er

of

mig

rati

ng

cells/f

ield

500

400

300

200

100

0

Mock

Figure 6. UNC5D inhibits renal cancer cell migration and invasion. 786-O cells were infected with adenoviruses (A–C) or lentiviruses (D) and evaluatedfor the capacity of migration and invasion. Each assay was repeated at least 3 times with triplicate. A, representative images showing cell migration inscratch wound-healing assay (left) and wound closure over time as measured by the distance between the front edges (right). B, cell migration in Transwellanalysis. C, cell invasion across Matrigel. The number of migration or invasion cells was counted in 6 randomly chosen fields and averaged for each of thetriplicate wells. D, UNC5D-DDD abolished the inhibitory function on cell migration of 786-O cells. �, P < 0.05; ��, P < 0.01; ��, P < 0.001.

Lu et al.


associated with colon cancer predisposition (37). It will beinteresting to see whether gene polymorphism also affectsUNC5D expression.The reduced or lost expression ofUNC5D in cancer cells

suggests a tumor-suppressive activity for UNC5D. Weshowed that restoration of UNC5D expression in renalcancer cell lines suppressed cell proliferation and anchor-age-dependent and -independent growth. EnforcedUNC5D expression also displayed a profound inhibitoryeffect on cell mobility, leading to a reduced capacity ofmigration and invasion.RCC is notorious for its resistance to chemo- or radio-

therapy due to its insensitive to cell death induction (38).Similarly, although UNC5D indeed induced apoptosis insome cell lines (25 and our unpublished data), it did notshow this ability in renal cancer cells (data not shown).Herewe showed that UNC5D exerted the tumor-suppressivefunction mainly by cell-cycle arrest in renal carcinoma.UNC5D induced G2–M cell-cycle arrest with the upregula-tion of G2–Mhallmark cyclin B1. To date, there is no relatedreports about UNC5A-C. Further studies will extend to testwhether cell-cycle arrest is a common event induced byUNC5Hs or a unique feature for UNC5D. Moreover, thegrowth-inhibitory effect of UNC5D was mainly mediatedby the death domain. The involvement of death domaincontaining receptors or signal molecules in cell-cycle con-trol is extensively studied. For example, one Netrin-1 recep-tor DCC induced G2–M arrest by inhibition of Cdk1 (39).Dependence receptor p75NTR suppresses the proliferationof human gastric cancer cells by cell-cycle arrest (40). Thenuclear death domain protein p84N5 activates a G2–M cell-cycle checkpoint (41). FasL or TRAIL signaling by deathreceptors inhibited cell death due to a cell-cycle arrest in S–G2–M (42), etc. Death domain is well known as the plat-form for interaction with other death domain containingproteins or signaling proteins. Intriguingly, by Scansitemotif analysis, we found that Thr893 of UNC5D withintheDD is a potential ATMkinase phosphorylation site. ATMkinase is activated in DNA damage which phosphorylatesseveral key proteins capable of initiating activation of the

DNA damage checkpoint, leading to cell-cycle arrest, DNArepair, or apoptosis. A representative molecule is PIDD,which is phosphorylated by ATM kinase on Thr788 withinthe DD, facilitating the binding of RAIDD and caspase-2activation (43). In future, identification of signal pathwaysinitiated by UNC5D which lead to cell-cycle arrest will bevery meaningful.

In summary, our findings highlight the potential ofUNC5D as a TSG in RCCs, which may serve as a potentialdiagnostic and therapeutic target for RCC intervention. Italso extends the current understanding of the general rolesof UNC5H receptors in tumorigenesis.

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

Authors' ContributionsConception and design: Y. Zhang, J. ZhangDevelopment of methodology: D. Lu, D. DongAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): D. Lu, D. Dong, Y. ZhouAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis):D. Lu,D.Dong, Y. Zhou,M. Lu, Y. Zhang,J. ZhangWriting, review, and/or revision of the manuscript: D. Lu, D. Dong, Y.Zhang, J. ZhangAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases):M. Lu, X.-W. Pang, Y. Li, X.-J. Tian, J.ZhangStudy supervision: Y. Zhang, J. Zhang

AcknowledgmentsThe authors thank Dr. Mike McNutt for his careful proofreading of the

manuscript.

Grant SupportThis work received support from the National Basic Research Program of

China (2011CB946103), Beijing Municipal Natural Science Foundation(7122104), and the National Natural Science Foundation of China(81072395).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received September 17, 2012; revised March 18, 2013; accepted April 3,2013; published OnlineFirst April 15, 2013.

References1. Cohen HT, McGovern FJ. Renal-cell carcinoma. N Engl J Med 2005;

353:2477–90.2. Rini BI, Campbell SC, Escudier B. Renal cell carcinoma. Lancet 2009;

373:1119–32.3. George DJ, Kaelin WG Jr. The von Hippel-Lindau protein, vascular

endothelial growth factor, and kidney cancer. N Engl J Med 2003;349:419–21.

4. Eisen T, Sternberg CN, Robert C, Mulders P, Pyle L, Zbinden S, et al.Targeted therapies for renal cell carcinoma: review of adverse eventmanagement strategies. J Natl Cancer Inst 2012;104:93–113.

5. Luo J, Solimini NL, Elledge SJ. Principles of cancer therapy: oncogeneand non-oncogene addiction. Cell 2009;136:823–37.

6. Baylin SB, Jones PA. A decade of exploring the cancer epigenome -biological and translational implications. Nat Rev Cancer 2011;11:726–34.

7. Walker CL, Ho S. Developmental reprogramming of cancer suscep-tibility. Nat Rev Cancer 2012;12:479–86.

8. Gnarra JR, Tory K,WengY, Schmidt L,WeiMH, Li H, et al.Mutations ofthe VHL tumour suppressor gene in renal carcinoma. Nat Genet 1994;7:85–90.

9. Baldewijns MM, van Vlodrop IJ, Schouten LJ, Soetekouw PM, deBruine AP, van Engeland M. Genetics and epigenetics of renal cellcancer. Biochim Biophys Acta 2008;1785:133–55.

10. Morris MR, Gentle D, Abdulrahman M, Clarke N, Brown M, Kishida T,et al. Functional epigenomics approach to identify methylated candi-date tumour suppressor genes in renal cell carcinoma. Br J Cancer2008;98:496–501.

11. Hoffman AM, Cairns P. Epigenetics of kidney cancer and bladdercancer. Epigenomics 2011;3:19–34.

12. Morris MR, Ricketts C, Gentle D, Abdulrahman M, Clarke N, BrownM,et al. Identification of candidate tumour suppressor genes frequentlymethylated in renal cell carcinoma. Oncogene 2010;29:2104–17.

13. Kagara I, Enokida H, Kawakami K, Matsuda R, Toki K, Nishimura H,et al. CpG hypermethylation of the UCHL1 gene promoter is



associated with pathogenesis and poor prognosis in renal cell carci-noma. J Urol 2008;180:343–51.

14. Sogabe Y, Suzuki H, Toyota M, Ogi K, Imai T, Nojima M, et al.Epigenetic inactivation of SFRP genes in oral squamous cell carcino-ma. Int J Oncol 2008;32:1253–61.

15. Lv D, Zhao W, Dong D, Qian XP, Zhang Y, Tian XJ, et al. Genetic andepigenetic control of UNC5C expression in human renal cell carcino-ma. Eur J Cancer 2011;47:2068–76.

16. Delloye-Bourgeois C, Fitamant J, Paradisi A, Cappellen D, Douc-RasyS, Raquin MA, et al. Netrin-1 acts as a survival factor for aggressiveneuroblastoma. J Exp Med 2009;206:833–47.

17. Delloye-Bourgeois C, Brambilla E, Coissieux MM, Guenebeaud C,Pedeux R, Firlej V, et al. Interference with netrin-1 and tumor cell deathin non-small cell lung cancer. J Natl Cancer Inst 2009;101:237–47.

18. Paradisi A, Mehlen P. Netrin-1, a missing link between chronic inflam-mation and tumor progression. Cell Cycle 2010;9:1253–62.

19. Goldschneider D, Mehlen P. Dependence receptors: a new paradigmin cell signaling and cancer therapy. Oncogene 2010;29:1865–82.

20. Mehlen P, Goldschneider D. [Dependence receptors DCC andUNC5H: the role of apoptosis in the control of tumorigenesis]. J SocBiol 2005;199:211–8.

21. Thiebault K, Mazelin L, Pays L, Llambi F, Joly MO, Scoazec JY, et al.The netrin-1 receptors UNC5H are putative tumor suppressors con-trolling cell death commitment. Proc Natl Acad Sci U S A 2003;100:4173–8.

22. Bernet A, Mazelin L, Coissieux MM, Gadot N, Ackerman SL, ScoazecJY, et al. Inactivation of the UNC5C Netrin-1 receptor is associatedwith tumor progression in colorectal malignancies. Gastroenterology2007;133:1840–8.

23. Hibi K,MizukamiH, Shirahata A,Goto T, SakataM,SanadaY. Aberrantmethylation of the netrin-1 receptor genes UNC5C and DCC detectedin advanced colorectal cancer. World J Surg 2009;33:1053–7.

24. Engelkamp D. Cloning of three mouse Unc5 genes and their expres-sion patterns at mid-gestation. Mech Dev 2002;118:191–7.

25. Wang H, Ozaki T, Shamim Hossain M, Nakamura Y, Kamijo T, Xue X,et al. A newly identified dependence receptor UNC5H4 is inducedduring DNA damage-mediated apoptosis and transcriptional target oftumor suppressor p53. Biochem Biophys Res Commun 2008;370:594–8.

26. Gelsi-Boyer V, Orsetti B, Cervera N, Finetti P, Sircoulomb F, Rouge C,et al. Comprehensive profiling of 8p11-12 amplification in breastcancer. Mol Cancer Res 2005;3:655–67.

27. Melchor L, Garcia MJ, Honrado E, Pole JC, Alvarez S, Edwards PA,et al. Genomic analysis of the 8p11-12 amplicon in familial breastcancer. Int J Cancer 2007;120:714–7.

28. Goode EL, Chenevix-Trench G, Hartmann LC, Fridley BL, Kalli KR,Vierkant RA, et al. Assessment of hepatocyte growth factor in ovariancancer mortality. Cancer Epidemiol Biomarkers Prev 2011;20:1638–48.

29. Olek A, Oswald J, Walter J. A modified and improved method forbisulphite based cytosine methylation analysis. Nucleic Acids Res1996;24:5064–6.

30. LeeMN, TsengRC, HsuHS, Chen JY, TzaoC, HoWL, et al. Epigeneticinactivation of the chromosomal stability control genes BRCA1,BRCA2, and XRCC5 in non–small cell lung cancer. Clin Cancer Res2007;13:832–8.

31. Dancey JE, Bedard PL, Onetto N, Hudson TJ. The genetic basis forcancer treatment decisions. Cell 2012;148:409–20.

32. Loginov VI, Khodyrev DS, Pronina IV, Maliukova AV, Kazubskaia TP,Ermilova VD, et al. [Two CpG-islands of SEMA3B gene: methylation inclear cell renal cell carcinoma]. Mol Biol (Mosk) 2009;43:1088–92.

33. Atschekzei F, Hennenlotter J, Janisch S, Grosshennig A, Tran-kenschuh W, Waalkes S, et al. SFRP1 CpG island methylation locusis associated with renal cell cancer susceptibility and disease recur-rence. Epigenetics 2012;7:447–57.

34. Marsit CJ, KaragasMR, Andrew A, Liu M, Danaee H, Schned AR, et al.Epigenetic inactivation of SFRP genes and TP53 alteration act jointlyas markers of invasive bladder cancer. Cancer Res 2005;65:7081–5.

35. Williams SV, Platt FM, Hurst CD, Aveyard JS, Taylor CF, Pole J, et al.High-resolution analysis of genomic alteration on chromosome arm8pin urothelial carcinoma. Genes Chromosomes Cancer 2010;49:642–59.

36. Garcia MJ, Pole JCM, Chin SF, Teschendorff A, Naderi A, Ozdag H,et al. A 1 Mb minimal amplicon at 8p11–12 in breast cancer identifiesnew candidate oncogenes. Oncogene 2005;24:5235–45.

37. Liu P, Lu Y, Liu H, Wen W, Jia D, Wang Y, et al. Genome-wideassociation and fine mapping of genetic loci predisposing to coloncarcinogenesis in mice. Mol Cancer Res 2012;10:66–74.

38. Pantuck AJ, Zeng G, Belldegrun AS, Figlin RA. Pathobiology, prog-nosis, and targeted therapy for renal cell carcinoma: exploiting thehypoxia-induced pathway. Clin Cancer Res 2003;9:4641–52.

39. Chen YQ, Hsieh JT, Yao F, Fang B, Pong RC, Cipriano SC, et al.Induction of apoptosis and G2/M cell cycle arrest by DCC. Oncogene1999;18:2747–54.

40. Jin H, Pan Y, Zhao L, Zhai H, Li X, Sun L, et al. p75 neurotrophinreceptor suppresses the proliferation of human gastric cancer cells.Neoplasia 2007;9:471–8.

41. Doostzadeh-Cizeron J, Terry NH, Goodrich DW. The nuclear deathdomain protein p84N5 activates a G2/M cell cycle checkpoint prior tothe onset of apoptosis. J Biol Chem 2001;276:1127–32.

42. Bosque A, Aguilo JI, del Rey M, Paz-Artal E, Allende LM, Naval J, et al.Cell cycle regulation by FasL and Apo2L/TRAIL in human T-cell blasts.Implications for autoimmune lymphoproliferative syndromes. JLeukocBiol 2008;84:488–98.

43. Ando K, Kernan JL, Liu PH, Sanda T, Logette E, Tschopp J, et al. PIDDdeath-domain phosphorylation by ATM controls prodeath versusprosurvival PIDDosome signaling. Mol Cell 2012;47:681–93.

Lu et al.


The tumor-suppressive function of UNC5D and its repressed expression in renal cell carcinoma

Documents