Oncogene regulation of tumor suppressor genes in tumorigenesis

Jimmy Sung, Joel Turner, Susan McCarthy,Steve Enkemann, Chan Gong Li, Perally Yan,Timothy Huang and Timothy J.Yeatman�

Department of Surgery and Interdisciplinary Oncology, H.Lee MoffittCancer Center and Research Institute, University of South Florida, Tampa,FL 33612, USA and Department of Molecular Virology, ComprehensiveCancer Center, Ohio State University, Columbus, OH 43210, USA

�To whom correspondence should be addressedEmail: yeatman@moffitt.usf.edu

We attempted to demonstrate whether there is an epige-netic link between oncogenes and tumor suppression genesin tumorigenesis. We designed a high throughput model toidentify a candidate group of tumor suppressor genespotentially regulated by oncogenes. Gene expression profil-ing of mock-transfected versus v-src-transfected 3Y1 ratfibroblasts identified significant overexpression of DNAmethyltransferase 1, the enzyme responsible for aberrantgenome methylation, in v-src-transfected fibroblasts. Sec-ondary microarray analyses identified a number of candi-date tumor suppressor genes that were down-regulated byv-src but were also re-expressed following treatment with5-aza-20-deoxycytidine, a potent demethylating agent. Thiscandidate group included both tumor suppressor genesthat are known to be silenced by DNA hypermethylationand those that have not been previously identified withpromoter hypermethylation. To further validate ourmodel, we identified tsg, a tumor suppressor gene thatwas shown to be down-regulated by v-src and foundto harbor dense promoter hypermethylation. Our modeldemonstrates a cooperative relationship between onco-genes and tumor suppressor genes mediated throughpromoter hypermethylation.

Introduction

It is well known that cancer cells are often the product ofmultiple genetic alterations that cause cellular transformation.To date, numerous specific genetic alterations have been iden-tified that activate proto-oncogenes genes or inactivate tumorsuppressor genes. In fact, the sine qua non of a cancer gene isone that is affected by a mutational event with a significantprevalence. In addition, a ‘third’ pathway to tumorigenesishas been identified whereby the expression of key genesis regulated through promoter hypermethylation and silencing.Tumor suppressor genes, in particular, may be subject tothis mechanism of inactivation, in addition to mutationalevents (1).Oncogenes and tumor suppressor genes have classically been

assigned distinct, independent roles in cancer development and

progression. The interrelationships, structural and temporal,between these tumor-promoting processes, however, are stillpoorly understood. We have hypothesized that the inhibition oftumor suppressor genes that occurs via promoter hypermethy-lation may be initiated and regulated by the activation ofoncogenes, explaining in part how a single oncogene can resultin cellular transformation. To test this hypothesis we designeda high throughput model to identify tumor suppressor genespotentially regulated by oncogenes through methylationevents. We used the well-defined cellular model ofv-src-mediated cellular transformation to demonstrate theserelationships (2). We have previously examined the genesthat were principally up-regulated by v-src and, in fact, wewere able to develop a ‘Src fingerprint’ of genes commonlyregulated by Src that was detected in human colon cancerspecimens well known to harbor high levels of Src activity(3). In the current study we chose to examine genes down-regulated by v-src as a means of understanding the relationshipof tumor suppressor genes to the v-src oncogene. We demon-strated that a single oncogene, v-src, can inhibit the expressionof a large number of genes during the process of cellulartransformation. Using a gene expression profiling approach,a comparison of v-src-transfected with mock-transfected ratfibroblasts identified a number of candidate tumor suppressorgenes that are down-regulated by v-src and then re-expressedto baseline levels following treatment with 5-aza-20-deoxycy-tidine (DAC) (4--6). The regulation of one of these candidategenes, tsg, was the principal subject of this study in which wedemonstrated that a tumor suppressor gene is silenced by anoncogene through promoter hypermethylation.

Materials and methods

Cell culture

Pools of stably transfected 3Y1 rat fibroblasts were a kind gift from RichardJove (H.Lee Moffitt Cancer Center and Research Institute). Mock-transfectedand v-src-transfected 3Y1 cells were grown in Dulbecco’s modified Eagle’smedium (Life Technologies) with 10% fetal bovine serum at 37�C, 5% CO2.

Microarray analysis

The methods for preparation and subsequent handling of RNA up throughhybridization followed the recommendations of the Affymetrix GeneChip�

Expression analysis system. Rat U34 GeneChips, containing ~8000 elementsper array, were used in all experiments. Scanned output files were visuallyinspected for hybridization artifacts and then analyzed using AffymetrixMicroarray 5.1 software.

Real-time quantitative PCR (real-time qPCR)

A quantitative primer/probe set was designed to evaluate and to quantitative rattsg (accession no. S72637), rat vhl (accession no. U14746), rat dnmt1 (acces-sion no. NM_053354) and gapdh in 3Y1 v-src-transfected cell lines. Theprimers were as follows. Rat tsg: forward primer 50-CACGCAGATTGTGG-GCC-30; reverse primer 50-TCCTAGGCACGCCCTGTTC-30; probe 50-AGCG-GCTGTGAGGCCAAGTCTATCC-30. Rat dnmt1: forward primer 50-GGAAG-GTGAGCATCGACGAA-30; reverse primer 50-GATCATCCGGAATGACC-GAG-30; probe 50-ACTCTGGAGGTGGGCGACTG-CG-30. Rat vhl: forwardprimer 50-GCTGCCTTTGTGGCTCAAC-30; reverse primer 50-GGTGCCC-GGTGGTAAGGT-30; probe 50-TTGATGGTGAGCCTCAGCCCTACCC-30.

Abbreviations: DAC, 5-aza-20-deoxycytidine; DNMT1, DNA-cytosine 5-methyltransferase.

Carcinogenesis vol.26 no.2 # Oxford University Press 2005; all rights reserved. 487

Carcinogenesis vol.26 no.2 pp.487--494, 2005doi:10.1093/carcin/bgh318

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

gapdh: forward primer 50-TGAAGGTCGGAGTCAACGG-30; reverse primer50-AGAGTTAAAAGCAGCCCTGGTG-30; probe 50-TTTGGTCGTATTGG-GCGCCTGG-30. Total RNA was extracted by the guanidine isothiocyanate,phenol/chloroformmethod (Trizol; Gibco) with the addition of 20 mg glycogenas a carrier for the RNA. Reverse transcription of RNA was performed using15 U omniscript reverse transcriptase (Qiagen), 1� cDNA synthesis buffer,40 U RNase inhibitor (Gibco), 5 mM DTT, 1 mM dNTP mixture, 0.5 mgoligo(dT) primer and 2 mg RNA. Primers and probes for real-time PCR weredesigned using Primer Express software (Applied Biosystems).

Methylation-specific PCR (MS-PCR)

Genomic DNAwas isolated from cells by digestion with 100 g/ml proteinase Kfollowed by standard phenol/chloroform (1:1) extraction and ethanol precipi-tation. DNA was treated with sodium bisulfite as follows. A sample of 1 ggenomic DNA was denatured by incubation with 0.2 M NaOH for 10 min at37�C. Then 10 mM hydroquinone (Sigma Chemical Co., St Louis, MO) and3 M sodium bisulfite (pH 5.0) (Sigma Chemical Co.) were added and thesolution was incubated at 50�C for 16 h. Treated DNA was purified using theWizard DNA Purification System (Promega Corp., Madison, WI), desulfo-nated with 0.3 M NaOH, precipitated with ethanol and resuspended in water.Modified DNA was stored at �80�C until used. MS-PCR was performedwith primers specific for the methylation reaction. The primers were asfollows. Methylated tsg promoter: forward 50-TGGCCCCGGCACCCCTGC-TCGTCGCGCG-30; reverse 50-CGCGGCCGCCGCGCGCGGTAAGTCTC-30.Unmethylated tsg promoter: forward 50-TGGCCCTGGCACCCCTGCTTG-TTGTGTG 30; reverse 50 TGTGGCTGCTGTGTGTGGTAAGTCTC-30.Negative control samples without DNA were included for each set of PCR.PCR products were analyzed on 2% agarose gels containing ethidium bromide.

Bisulfite sequencing and combined bisulfite restriction analysis (COBRA)assay

Sodium bisulfite modification of genomic DNA was conducted using aCpGenome DNA Kit according to the manufacturer’s instructions (Intergen).Bisulfite-treated DNA (~1 ng) was used as a template for PCR with specificprimers flanking the BstUI sites located within the CpG island regionsof interest. Primers used for amplification were as follows: tsg_121U23,50-TTAGAATTTAATTTTTTGAGGGA-30; tsg_294L22, 50-TTACTCCRCC-TAACTTTTCTAA-30. After amplification, PCR products incorporating 32Pwere digested with BstUI (New England Biolabs), which recognizes sequencesunique to the methylated alleles. The undigested control and digested DNAsamples were separated in parallel on 8% polyacrylamide gels and subjected toautoradiography using a PhosphorImager (Amersham-Pharmacia). The regionof interest within the tsg promoter site was amplified from bisulfite-treatedgenomic DNA by PCR using the same primer pairs as described earlier.Amplified products were subcloned using the TOPO-TA Cloning System(Invitrogen). Plasmid DNA of 10--15 insert-positive clones were isolatedusing a QIAprep Spin Miniprep kit (Qiagen) and sequenced using the ABIsequencing system (Applied Biosystems).

Western analysis

DNA-cytosine 5-methyltransferase (DNMT1) protein levels were assessedusing goat anti-rat antibody purchased from Santa Cruz Inc. (Santa Cruz,CA) under the following conditions. The phosphotyrosine level inv-src-transfected 3Y1 cells was dectected using anti-phosphotyrosine-specficantibodyclone4G10(UpstateBiotechnology,LakePlacid,NY).Cellswerelysedin 50 mM Pipes--KOH (pH 6.5), 2 mM EDTA, 0.1% 3-((3-cholamidopropyl)-dimethylammonio)-1-propanesulfonic acid, 5 mM DTT, 20 g/ml leupeptin,10 g/ml pepstatin A, 10 g/ml apoprotinin and 2 mM phenylmethylsulfonylfluoride. Lysates were centrifuged at 4�C for 30 min at 20 000 g and thesupernatant fraction recovered. Protein extracts (100 g) were fractionatedthrough a 10% criterion precast gel (Bio-Rad, Hercules, CA) and blottedonto pure nitrocellulose membranes (Bio-Rad).

Final protein detection used horseradish peroxidase-conjugated mouseanti-goat IgG secondary antibodies purchased from Biosource (Camarillo,CA) and Santa Cruz Biotech (Santa Cruz, CA) and SuperSignal West PicoChemiluminescent Substrate (Pierce, Rockford, IL).

tsg transfection

Full-length rat tsg was cloned and inserted into the PCDNA3.1 expressionvector (Invitrogen) and transiently transfected (48 h) using 2 ml of SuperFect(Qiagen) into v-src-transformed 3Y1 rat fibroblasts to assess effects on tumor-igenicity. Photographs were obtained 3 days after transfection.

Antisense and 5-aza-20-deoxycytidine treatments of cells in culture

DAC (200 nm) was dissolved in fresh phosphate-buffered saline each day andadded to fresh medium for 6 days.

Results

Verification of v-src transfection

We first verified the quality of our 3Y1 v-src transfection bycomparing Src protein activity between the mock- and v-src-transfected 3Y1 cells with a phosphotyrosine western blot(Figure 1). These data demonstrate a high degree of Src-specific activity in v-src-transfected cells versus mock-transfected controls and is consistent with the transformationphenotype observed.

Identification of genes down-regulated by v-src

In order to identify candidate tumor suppressor genes regu-lated by v-src we performed a genome-wide microarray ana-lysis (in triplicate) of 3Y1 rat fibroblasts stably transfected(and transformed) with v-src and compared them with mock-transfected controls. After standard normalization and scalingprocedures, we compared the expression of these two celllines, gene by gene, to identify the genes that were signifi-cantly (P 5 0.001) down-regulated by v-src. A large numberof significantly down-regulated genes were identified.A representative sample of these genes is shown in Table I.A complete list of all down-regulated genes is shown inSupplementary material, Table SI.

Identification of v-src-regulated methyltransferases

Because we hypothesized that v-src might regulate a numberof candidate tumor suppressor genes though promoter hyper-methylation and silencing, we examined the expression of allevaluable methyltransferases on the Affymetrix GeneChip. Ofnumerous methyltransferases examined, DNMT1 was promi-nently overexpressed in v-src-transformed fibroblasts whencompared with mock controls (Figure 2). DNMT1 expressionwas confirmed at both the RNA level by real time qPCR(Figure 3A) and at the protein level by western analysis(Figure 3B), supporting its potential in the regulation ofpromoter activity.

Fig. 1. Phosphotyrosine western blot demonstrating high levels ofphosphotyrosine in v-src-transfected rat fibroblasts when compared withmock-transfected controls.

J.Sung et al.

488

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

Experimental screening for candidate tumor suppressorgenes regulated by promoter hypermethylation: identificationof tsg

In order to identify genes that were regulated by promoterhypermethylation we combined the gene expression profilingapproach with a demethylating treatment of the same cellsusing DAC to find genes that were down-regulated by v-srcbut then re-expressed following demethylation. By comparingthe gene expression of cells transformed by v-src with thosesame cells treated with DAC we were able to identify a numberof candidate tumor suppressor genes potentially regulated bypromoter hypermethylation. A heat map identifying these can-didate genes and demonstrating their expression consequent toDAC treatment is shown in Figure 4. Prominent among anumber of genes on this list was the von Hippel-Lindautumor suppressor gene (vhl), previously shown to be regulatedby promoter hypermethylation, validating our approach to thisscreening process (7). In addition, a new gene called tumorsuppressor gene (tsg) was identified. We chose tsg as the

principal focus of this study because it has been previouslyidentified as a strong tumor suppressor gene candidate with thepotential to reverse the transformation of neoplastic cells (8).However, tsg was not previously know to be regulated bymethylation. We have confirmed the previously reportedtumor suppressive properties of tsg by demonstrating thattransient transfection of tsg into v-src-transformed cells canreverse the transformation phenotype of v-src-transected cells(Figure 5).Expression of both vhl and tsg was subsequently confirmed

by real-time qPCR to be down-regulated by v-src transfectionand then re-expressed by demethylation following treatmentwith DAC (Figure 6).

Validation of tsg promoter methylation by bisulfite sequencing

While data demonstrating re-expression of down-regulatedgenes after DAC treatment is compelling, definitive evidenceof promoter hypermethylation is required since it has beendemonstrated that not all genes re-expressed following

Table I. Genes down-regulated by v-src

Affinity probe Reduction inv-src cells

Gene name Genesymbol

Accesssionno.

Locuslink

Unigeneno.

X63744_g_at 189.9 Solute carrier family 1, member 3 Slc1a3 NM_019225 29483 Rn.34134K01932_f_at 158.6 Glutathione S-transferase, a1 Gsta1 NM_031509 24421 Rn.10460L19998_at 127.9 Sulfotransferase family 1A,

phenol-preferring, member 1Sult1a1 NM_031834 83783 Rn.1507

X95466_at 109.2 CPG2 protein CPG2 NM_019355 54317 Rn.89150D85183_s_at 88.7 Protein tyrosine phosphatase,

non-receptor type substrate 1Ptpns1 NM_013016 25528 Rn.53971

Z22607_at 86.5 Bone morphogenetic protein 4 Bmp4 NM_012827 25296 Rn.10318M92059_s_at 79.4 Adipsin Adn M92059 54249 Rn.16172rc_AI102795_at 68.2 Pleiotrophin Ptn NM_017066 24924 Rn.1653D28560_g_at 64.4 Ectonucleotide pyrophosphatase/

phosphodiesterase 2Enpp2 NM_057104 84050 Rn.20403

M26125_at 54 Epoxide hydrolase 1 Ephx1 NM_012844 25315 Rn.3603rc_AI237731_s_at 42.9 Lipoprotein lipase Lpl NM_012598 24539 Rn.3834J03637_at 42.7 Aldehyde dehydrogenase family 3,

member A1Aldh3a1 NM_031972 25375 Rn.105627

AJ005396_at 39.5 Procollagen, type XI, a1 Col11a1 AJ005396 25654 Rn.260U65656_at 38.5 Matrix metalloproteinase 2 (72 kDa

type IV collagenase)Mmp2 NM_031054 81686 Rn.6422

X81449cds_g_at 38.2 Keratin complex 1, acidic, gene 19 Krt1-19 X81449 117046 Rn.9359rc_AI172247_at 38.2 Xanthine dehydrogenase Xdh NM_017154 29289 Rn.9938rc_AA894298_s_at 33.3 Membrane metalloendopeptidase Mme NM_012608 24590 Rn.33598rc_AA875025_at 31.9 Cellular retinoic acid-binding protein I Crabp1 25061 Rn.3207U23056_at 29.8 C-CAM4 protein LOC287009 NM_173339 287009 Rn.92160rc_AI170411_s_at 23.1 Dipeptidase 1 Dpep1 NM_053591 94199 Rn.6051U66470_at 21.9 Cell growth regulatory with EF-hand domain Cgr11 NM_139087 245918 Rn.31842D42148_at 21.8 Growth arrest-specific 6 Gas6 NM_057100 58935 Rn.52228rc_AI230247_s_at 21.6 Selenoprotein P, plasma, 1 Sepp1 NM_019192 29360 Rn.1451U07183_at 20.3 Solute carrier family 10, member 2 Slc10a2 NM_017222 29500 Rn.85891rc_AA892258_at 18.6 NADPH oxidase 4 Nox4 NM_053524 85431 Rn.14744U38812_s_at 17.9 Inositol 1,4,5-triphosphate receptor 1 Itpr1 U38812 25262 Rn.2135rc_AI232194_at 16.4 Chimerin (chimaerin) 2 Chn2 NM_032084 84031 Rn.10521AF036548_at 16.3 Rgc32 protein Rgc32 NM_054008 117183 Rn.3504U07181_g_at 15.8 Lactate dehydrogenase B Ldhb NM_012595 24534 Rn.1785S35751_f_at 15 3-a-Hydroxysteroid dehydrogenase LOC191574 NM_138547 191574 Rn.10021AF030358_g_at 14.4 Chemokine (C-X3-C motif) ligand 1 Cx3cl1 NM_134455 89808 Rn.107266rc_AI176504_at 13.9 Glutaminase Gls NM_012569 24398 Rn.5762rc_AA891194_s_at 13.1 Arg/Abl-interacting protein ArgBP2 Argbp2 NM_053770 114901 Rn.24612Z78279_at 12.8 Collagen, type 1, a1 Col1a1 29393 Rn.2953AF037272_at 12.4 Wap four-disulfide core domain 1 Wfdc1 NM_133581 171112 Rn.3193X70369_s_at 11.4 Collagen, type III, a1 Col3a1 84032 Rn.3247AB018049_s_at 10.8 CMP-NeuAc:lactosylceramide

a-2,3-sialyltransferase; GM3 synthaseSiat9 NM_031337 83505 Rn.22706

AF020046_s_at 10.5 Integrin aE1, epithelial-associated Itgae NM_031768 83577 Rn.29975U50147_at 10.4 Discs, large homolog 3 (Drosophila) Dlgh3 NM_031639 58948 Rn.10238

Oncogene regulation of tumor suppressor genes

489

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

DAC treatment are controlled via promoter hypermethylation(4). We first compared the tsg promoter site in mock- andv-src-transfected 3Y1 cells using MS-PCR and found evidenceof promoter hypermethylation (Figure 7A). To further confirmthis finding, we then selected a region of the tsg promoter site(region 1) for more definitive evaluation by combined bisulfiterestriction analysis (COBRA) assay and bisulfate sequencing.We found that v-src-transfected 3Y1 cells contained a specificregion of high density methylation (Figure 7B and C). Finally,we sequenced the remaining region of the promoter site(region 2) and observed generalized hypermethylation of thetsg promoter site in v-src-transfected 3Y1 cells (Figure 7D).

Using this approach we have demonstrated that the promoterregion of tsg contains CpG islands and, once transfected withv-src, these CpG islands increased in methylation densityand were noted to contain specific regions of high densitymethylation.

Discussion

To identify a potential relationship between oncogenes andtumor suppressor genes, the classic model of v-src fibroblasttransformation was selected. We and others have previouslydemonstrated that v-src is a potent inducer of transformation

Fig. 2. Gene expression profile of all methyltransferases on a U34A GeneChip shows prominent up-regulation of dnmt1 in v-src-transfected rat 3Y1 fibroblastscompared with mock-transfected controls.

Fig. 3. Confirmation of induction of dnmt1 expression in v-src-transformed fibroblasts. (A) Real-time quantitative PCR analysis identified a 5-fold increase inexpression when compared with mock-transfected cells. (B) Protein expression determined by western analysis confirms the 5-fold increase in expression.

J.Sung et al.

490

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

that is characterized by a signature of genes whose expressionis both up-regulated as well as down-regulated relative tomock-transfected controls (3). Much attention has beendevoted to the induced genes, however, those genes suppressedby v-src that are potential tumor suppressors have been less

well examined. First, we sought to identify a number of genesthat were down-regulated by v-src and then re-expressedfollowing demethylation. Based on gene expression profilesof v-src-transformed fibroblasts, we hypothesized that v-srcmight inhibit the expression and function of a number of tumor

Fig. 4. Gene expression profile of genes significantly down-regulated by v-Src and re-expressed on treatment with 5-azacytidine. Data identify tumor suppressorgene candidates regulated by promoter methylation. Prominent amongst these genes were tsg and vhl.

Oncogene regulation of tumor suppressor genes

491

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

suppressor genes through the process of promoter hypermethy-lation mediated by DNA methyltransferases, in particularDNMT1. Both mock-transfected and v-src-transformed 3Y1rat fibroblasts were then treated with DAC, a demethylatingagent (9). The Affymetrix U34 rat GeneChip was used toidentify genes that were down-regulated in v-src-transformedas compared with mock-transfected cells but re-expressed aftertreatment with DAC. A comparison of mock-transfected ver-sus v-src-transfected rat fibroblasts (3Y1) found that DNMT1was prominently overexpressed in v-src- compared withmock-transfected cells (Figure 2). Further analysis using

MAS 5.0 software (Affymetrix) identified 912 probe sets sig-nificantly down-regulated (P 5 0.001) by v-src (Table I). Ofthese probe sets, 141 were up-regulated (P 5 0.001) by treat-ment with DAC and, within these probe sets, at least six ofthese were annotated genes with a putative tumor suppressorfunction and were rich in mathematically defined CpG islands(vhl, tsg, cdkn1a, par-4, p53 and st13) (10).Among the genes meeting our criteria, we first selected vhl,

Von Hippel-Lindeu tumor-suppressor gene, for further valida-tion. We were encouraged by the identification of vhl withour model because previous studies have definitively

Fig. 5. Effect of tsg transfection on cellular transformation. (A) Mock-transfected rat 3Y1 fibroblasts show normal density inhibition of growth. (B) 3Y1fibroblasts transfected with v-src take on the transformed phenotype with evidence of focus formation and loss of intercellular adhesion. (C) Reversion of thetransformed phenotype by tsg transfection shows the potential tumor suppressive properties of tsg.

Fig. 6. Real-time qPCR analysis of gene expression. (A) tsg re-expression following demethylation with 5-aza-20-deoxycytidine. (B) vhl re-expression aftertreatment with 5-aza-20-deoxycytidine (DAC).

J.Sung et al.

492

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

demonstrated that vhl is silenced by promoter hypermethyla-tion in renal carcinoma and can be reactivated by DAC (7).However, the involvement of v-src in the down-regulation ofvhl has not been reported. On the other hand, our modelidentified a second gene that is a known v-src-regulated puta-tive tumor suppressor gene, the 3Y1 tumor suppressor gene.Unlike vhl, the relationship between tsg and promoter sitehypermethylation has not been reported. tsg, also known asN03, was first identified as a novel gene that is down-regulatedin v-src, v-mos and SV-40 transformed 3Y1 cells (11). It wasnamed the ‘‘3Y1 v-src tumor suppressor gene’’ because sub-sequent studies confirmed tsg’s role in the suppression ofcolony formation in soft agar and tumorigenicity in nudemice (8). In accordance with these previous findings we alsoobserved the reversal of phenotype by transfecting 3Y1 v-srctransformed fibroblasts with tsg (Figure 5)To see whether tsgwas also silenced by CpG island promoter

hypermethylation, as in the case of vhl, we first validated ourmicroarray findings with real-time quantitative polymerasechain reaction (qPCR). We demonstrated a significant inhibi-tion of both tsg and vhl and expression in v-src-transformedcells but a dramatic induction and re-expression of vhl and tsgfollowing treatment with DAC (Figure 6).The main epigenetic modification in human beings is methy-

lation of the cytosine bases that are located 50 to a guanosine ina CpG dinucleotide. Short regions of the genome that are richin CpG dinucleotides are called CpG ‘islands.’ Hypermethyla-tion of these CpG islands of a promoter leads to silencing ofthe gene (12). Using computational analysis to survey thepromoter region of tsg, we found the region to be rich inCpG islands. We first evaluated the tsg promoter site methy-lation status with methylation specific polymerase chain

reaction (MS-PCR) and found hypermethylation of the tsgpromoter site in 3Y1 v-src cells compared to 3Y1 mock(Figure 7A). This finding was confirmed by combined bisul-fate restriction analysis (COBRA) (Figure 7B) and bisulfitesequencing (Figure 7C) which showed a specific region of thetsg promoter site, designated as region 1, to be hypermethy-lated in all 7 clones of the v-src transformed cells examined,whereas no mock-transfected cells (5 clones) were methylated.Furthermore, a complete bisulfate sequencing of the remainingpromoter site, designated as region 2, showed no methylationin the mock transfected clone and hypermethylation in all ofthe v-src transfected clones (Figure 7D). The issue as to whatregion of methylation regulates suppression of gene expressionis controversial. That is, suppression of gene expression maybe the result of a total increase in methylation density or a resultof hypermethylation of a particular region of the promoter site(13--16). In our analysis, we found both hypermethylation of aparticular region (region 1) in the 50 end of the tsg promotersite and a generalized total increase of methylation density ofthe entire promoter site (regions 1 and 2). Most importantly,we found that methylation status of the promoter site corre-lated with tsg expression (Figure 6A).Currently, several DNA methyltransferase have been asso-

ciated with hypermethylation, including DNMT1, DNMT2,DNMT3a, DNMT3b and DNMT3b3 (17--21). Of these, therole of DNMT1 in hypermethlation of tumor suppressor genesis best established (22,23). Furthermore, previous studies haveshown that DNMT1 is necessary and sufficient to maintainglobal methylation and aberrant CpG island methylation inhuman cancer cells (24,25). Therefore, we attempted to linkv-src to hypermethylation of tumor suppressor genes toDNMT1 expression. Previous studies by others showed

Fig. 7. Analysis of the methylation state of tsg in v-src-transformed 3Y1 fibroblasts confirmed hypermethylation of the tsg promoter. (A) Methylation-specificPCR analysis of tsg promoter suggesting hypermethylation. (B) Methylation analysis by COBRA of the tsg promoter site. (C) Bisulfite sequencing confirmedhypermethylation of the tsg promoter. Seven clones of v-src-transformed and five clones of mock-transfected 3Y1 fibroblasts were analyzed for methylationstatus. Methylated and unmethylated CpG sites are represented by closed and open circles, respectively. (D) Bisulfite sequencing of the remaining portion of thetsg promoter. Methylated and unmethylated CpG sites are represented by closed and open squares, respectively.

Oncogene regulation of tumor suppressor genes

493

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022

up-regulation of dnmt1 by Fos and Ras (26,27). In addition,analysis of the dnmt1 promoter site linking DNMT1 to the c-Junpathway provided an explanation to the responsiveness ofDNMT1 to oncogenic signals (28). In our real-time PCRanalysis we found significant up-regulation of DNMT1 inv-src-transformed compared with mock-transfected 3Y1 cells(Figure 3A). Furthermore, our western blot analysis demon-strated that DNMT1 protein is indeed up-regulated in v-src-transformed 3Y1 cells (Figure 3B). These findings furthersupport our hypothesis that promoter methylation is amechanismlinking the activities of oncogenes and tumor suppressor genes.Both oncogenes and tumor suppressor genes have been impli-

cated in carcinogenesis and in cancer progression. Longregarded as independent, we hypothesized that these distinctgenetic elements might be related through epigenetic mechan-isms. Our model provides evidence that v-src, an oncogene,silenced the vhl and tsg tumor suppressor genes through promo-ter hypermethylation in transformed rat fibroblasts. Our datasuggest that this same approach can be used with other trans-formed cell models (i.e. with activated Ras andMyc transforma-tion models) to identify additional tumor suppressor genecandidates that are epigenetically controlled by oncogenes.Moreover, this process can be evaluated in different speciesand different tissues of origin. The presence of an epigeneticlinkage between oncogenes and tumor suppressor genes suggeststhat tumorigenesis occurs by cooperative mechanisms.

Supplementary material

Supplementary material is available online at: http://www.carcin.oupjournals.org.

Acknowledgements

The authors would like to thank Bernard W. Futscher, Ph.D. of the ArizonaCancer Center for his helpful discussions regarding the preparation of thismanuscript.

References

1.Robertson,K.D. and Jones,A.P. (2000) DNA methylation: past, present andfuture directions. Carcinogenesis, 21, 461--467.

2.Yeatman,T.J. (2004) A Renaissance for SRC. Nature Rev. Cancer, 4,470--480.

3.Malek,R.L., Irby,R.B., Guo,Q.M. et al. (2002) Identification of Srctransformation fingerprint in human colon cancer. Oncogene, 21,7256--7265.

4.Suzuki,H., Gabrielson,E., Chen,W., Anbazhagan,R., van Engeland,M.,Weijenberg,M., Herman,J.G. and Baylin,S.B. (2002) A genomic screen forgenes upregulated by demethylation and histone deacetylase inhibition inhuman colorectal cancer. Nature Genet., 31, 141--149.

5.Toyota,M., Ho,C., Ahuja,N., Jair,K.-W., Li,Q., Ohe-Toyota,M.,Baylin,S.B. and Issa,J.-P.J. (1999) Identification of differentially methy-lated sequences in colorectal cancer by methylated CpG island amplifica-tion. Cancer Res., 59, 2307--2312.

6.Cameron,E.E., Bachman,K.E., Myohanen,S., Herman,J.G. and Baylin,S.B.(1999) Synergy of demethylation and histone deacetylase inhibition in there-expression of genes silenced in cancer. Nature Genet., 21, 103--107.

7.Herman,J., Latif,F., Weng,Y. et al. (1994) Silencing of the VHL tumor-suppressor gene by DNA methylation in renal carcinoma. Proc. Natl Acad.Sci. USA, 91, 9700--9704.

8.Ozaki,T.S., Sakiyama,S. (1994) Tumor-suppressive activity of N03 geneproduct in v-src-transformed rat 3Y1 fibroblasts.Cancer Res., 54, 646--648.

9.Christman,J., Schneiderman,N. and Acs,G. (1985) Formation of highlystable complexes between 5-azacytosine-substituted DNA and specificnon-histone nuclear proteins. Implications for 5-azacytidine-mediatedeffects on DNA methylation and gene expression. J. Biol. Chem., 260,4059--4068.

10.Takai,D.J.P. (2003) The CpG island searcher: a new WWW resource.In Silico Biol., 3, 235--40.

11.Ozaki,T. and Sakiyama,S. (1993) Molecular cloning and characterizationof a cDNA showing negative regulation in v-Src-transformed 3Y1 ratfibroblasts. Proc. Natl Acad. Sci. USA, 90, 2593--2597.

12. Jones,P.A., Baylin,S.B. (2002) The fundamental role of epigenetic eventsin cancer. Nature Rev. Genet., 3, 415--428.

13.Gonzalgo,M.L., Mayashid,T., Bender,C.M., Pao,M.M., Tsai,Y.C.,Gonzales,F.A., Nguyen,H.D., Nguyen,T.T. and Jones,P.A. (1998) Therole of DNA methylation in expression of the p19/p16 locus in humanbladder cancer cell lines. Cancer Res., 6, 1245--1252.

14.Chen,C., Yang,M.C.K. and Yang,T.P. (2001) Evidence that silencing ofthe HPRT promoter by DNA methylation is mediated by critical CpGsites. J. Biol. Chem., 276, 320--328.

15.Yan,P.S., Shi,H., Rahmatpanah,F., Hsiau,T.H.-C., Hsiau,A.H.-A.,Leu,Y.-W., Liu,J.C. and Huang,T.H.-M. (2003) Differential distributionof DNA methylation within the RASSF1A CpG island in breast cancer.Cancer Res., 63, 6178--6186.

16.Graff,J.R., Herman,J.G., Myohanen,S., Baylin,S.B. and Vertino,P.M.(1997) Mapping patterns of CpG island methylation in normal andneoplastic cells implicates both upstream and downstream regions inde novo methylation. J. Biol. Chem., 272, 22322--22329.

17.Weisenberger,D.J., Velicescu,M., Cheng,J.C., Gonzales,F.A., Liang,G.and Jones,P.A. (2004) Role of the DNA methyltransferase variantDNMT3b3 in DNA methylation. Mol. Cancer Res., 2, 62--72.

18.Leu,Y.-W., Rahmatpanah,F., Shi,H., Wei,S.H., Liu,J.C., Yan,P.S. andHuang,T.H.-M. (2003) Double RNA interference of DNMT3b andDNMT1 enhances DNA demethylation and gene reactivation. CancerRes., 63, 6110--6115.

19.Rhee,I. et al. (2002) DNMT1 and DNMT3b cooperate to silence genes inhuman cancer cells. Nature, 416, 552--556.

20.Ling,Y., Sankpal,U.T., Robertson,A.K., McNally,J.G., Karpova,T. andRobertson,K.D. (2004) Modification of de novo DNA methyltransferase3a (Dnmt3a) by SUMO-1 modulates its interaction with histonedeacetylases (HDACs) and its capacity to repress transcription. NucleicAcids Res., 32, 598--610.

21.Liu,K., Wang,Y.F., Cantemir,C. and Muller,M.T. (2003) Endogenousassays of DNA methyltransferases: evidence for differential activities ofDNMT1, DNMT2 and DNMT3 in mammalian cells in vivo. Mol. Cell.Biol., 23, 2709--2719.

22.Etoh,T., Kanai,Y., Ushijima,S., Nakagawa,T., Nakanishi,Y., Sasako,M.,Kitano,S. and Hirohashi,S. (2004) Increased DNA methyltransferase 1(DNMT1) protein expression correlates significantly with poorer tumordifferentiation and frequent DNA hypermethylation of multiple CpGislands in gastric cancers. Am. J. Pathol., 164, 689--699.

23.Vertino,P., Yen,R., Gao,J. and Baylin,S. (1996) De novo methylation ofCpG island sequences in human fibroblasts overexpressing DNA(cytosine-5)-methyltransferase. Mol. Cell. Biol., 16, 4555--4565.

24.Robert,M.F., Morin,S., Beaulieu,N., Gauthier,F., Chute,I.C., Barsalou,A.and MacLeod,A.R. (2003) DNMT1 is required to maintain CpGmethylation and aberrant gene silencing in human cancer cells. NatureGenet., 33, 61--65.

25.Kawasaki,H.T.K. (2004) Induction of DNAmethylation and gene silencingby short interfering RNAs in human cells. Nature, 431, 211--217.

26.Bakin,A.V. and Curran,T. (1999) Role of DNA 5-methylcytosinetransferase in cell transformation by fos. Science, 283, 387--390.

27.Ordway,J.M., Williams,K. and Curran,T. (2004) Transcription repressionin oncogenic transformation: common targets of epigenetic repression incells transformed by Fos, Ras or Dnmt1. Oncogene, 23, 3737--3748.

28.Bigey,P., Ramchandani,S., Theberge,J., Araujo,F.D. and Szyf,M. (2000)Transcriptional regulation of the human DNA methyltransferase (dnmt1)gene. Gene, 242, 407--418.

Received August 18, 2004; revised October 8, 2004;accepted October 12, 2004

J.Sung et al.

494

Dow

nloaded from https://academ

ic.oup.com/carcin/article/26/2/487/2476059 by guest on 09 February 2022