Cancer Research Donn Young EMT and Cancer Progression and Treatment
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cheng, C. K.
Right arrow Articles by Chan, V.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cheng, C. K.
Right arrow Articles by Chan, V.
[Cancer Research 65, 8646-8654, October 1, 2005]
© 2005 American Association for Cancer Research

Molecular Biology, Pathobiology and Genetics

The Cell Cycle Checkpoint Gene Rad9 Is a Novel Oncogene Activated by 11q13 Amplification and DNA Methylation in Breast Cancer

Chi Keung Cheng1, Louis W.C. Chow2, Wings T.Y. Loo2, Tai Kwong Chan1 and Vivian Chan1

1 University Department of Medicine, Queen Mary Hospital and 2 Department of Surgery, The University of Hong Kong, Pokfulam, Hong Kong, China

Requests for reprints: Vivian Chan, University Department of Medicine, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong, China. Phone: 852-2855-4249; Fax: 852-2855-1143; E-mail: vnychana{at}hkucc.hku.hk.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Human Rad9 (hRad9), a structural homologue of yeast Schizosaccharomyces pombe rad9, is involved in cell cycle checkpoints and apoptosis. hRad9 can serve as a corepressor of androgen receptor in prostate cancer cells, but little is known about its role in the development of breast or other cancers. In the present study, semiquantitative reverse transcription-PCR showed that Rad9 mRNA levels were up-regulated in 52.1% (25 of 48) of breast tumors, and this up-regulation correlated with tumor size (P = 0.037) and local recurrence (P = 0.033). Overexpression of Rad9 mRNA was partly due to an increase in Rad9 gene number as measured by quantitative PCR. In other breast tumors with Rad9 mRNA overexpression but without increase in gene number, there was differential methylation of two putative Sp1/3 binding sites within the first and second introns of the Rad9 gene, which was similarly found in MCF-7 breast cancer cell line with increased Rad9 mRNA. Silencing Rad9 expression by RNA interference in MCF-7 cell line inhibited its proliferation in vitro. Promoter assays indicated that the Sp1/3 site in intron 2 may act as a silencer. In vivo binding of Sp3 to intron 2 was shown by chromatin immunoprecipitation assays. Treatment of MCF-7 cell line with 5'-aza-2'-deoxycytidine reduced Rad9 mRNA expression and also increased binding of Sp3 to the demethylated intron 2 region. Collectively, these findings suggest that Rad9 is a novel oncogene candidate activated by 11q13 amplification and DNA hypermethylation in breast cancer and may play a role in tumor proliferation and local invasion.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Human Rad9 (hRad9) was originally identified as a structural homologue of yeast Schizosaccharomyces pombe rad9, which can partially rescue the sensitivity of rad9 null yeast to hydroxyurea, radiation, and the associated checkpoint defects (1). It is thought to interact with human Rad1 (hRad1) and human Hus1 (hHus1) via its proliferating cell nuclear antigen–like region to form a heterotrimeric hRad9-hRad1-hHus1 complex, which is recruited onto DNA lesions to trigger checkpoint signaling pathways (2–4). Besides its cell cycle checkpoint function, hRad9 contains a Bcl-2 homology 3 (BH3)–like domain at its NH2 terminus that can bind the antiapoptotic proteins Bcl-2 and Bcl-xL, thereby promoting apoptosis (5). Furthermore, the COOH terminus of hRad9 can interact with androgen receptor and result in suppression of androgen receptor activation in prostate cancer cells, suggesting that hRad9 may be a tumor suppressor (6).

The human Rad9 gene has been mapped to chromosome 11q13.1-13.2 (1). Amplification of 11q13 occurs at high frequencies in certain human cancers such as those derived from the breast, head and neck region, and esophagus (7). For breast cancer, this genetic alteration occurs in about 13% of patients and is associated with reduced survival (8, 9). The gene-dense 11q13 region spans several megabases and is believed to contain four amplification units that can be amplified independently or together in different combinations (10, 11). Various evidence suggested the existence of at least four putative oncogenes that are activated by 11q13 amplification (11). However, to date, only two oncogenes [i.e., CCND1 (encoding cyclin D1) and EMS1 (encoding cortactin)] have been shown as favored candidates for two of the amplification cores (11, 12).

Although hRad9 can function as a transcriptional corepressor of androgen receptor in prostate cancer cells (6), its role in the development of breast or other cancers remains unknown. To address this issue, we examined Rad9 mRNA expression in matched tumor and normal breast tissues using semiquantitative reverse transcription-PCR (RT-PCR) and evaluated possible correlation between Rad9 mRNA levels and clinicopathologic variables. We further investigated genetic and epigenetic mechanisms that may lead to aberrant Rad9 mRNA expression in breast tumors.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Tissue Samples and Cell Culture
Forty-eight unselected pairs of breast tumors and adjacent normal breast tissues resected from 47 women and one man at the Queen Mary Hospital or Tung Wah Hospital between August 2003 and April 2004 were used in this study, with ethics approval from the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster. Upon surgical resection, tissue specimens were immersed in RNAlater (Ambion, Inc., Austin, TX) and incubated overnight at 4°C before long-term storage at –80°C. The clinicopathologic characteristics of the tumors, including patient's age, histologic type, tumor size, tumor grade, local recurrence, lymph node status, lymphatic vessel permeation, nuclear receptor immunoreactivity, and c-erbB2 (HER2) immunoreactivity, were obtained from medical records. These tumors consisted of 45 invasive ductal carcinomas (IDC), one invasive lobular carcinoma, and two ductal carcinomas in situ (DCIS). Tumors were graded (grades 1, 2, and 3) according to the Bloom and Richardson classification (13). The mean age of the patients was 58.5 years (range, 30-102 years). Patients were followed-up for 4 to 60 months, with a median follow-up time of 12 months.

Breast cancer MCF-7, MDA-MB-231, and BT-483 cells were cultured under conditions recommended by the American Type Culture Collection (Manassas, VA). Culture media were renewed every 3 days, and cells were passaged when they reached about 80% confluence using trypsin-EDTA solution (Invitrogen Corp., Carlsbad, CA).

Studies of Human Breast Tissues
RNA extraction and semiquantitative reverse transcription-PCR. Total RNA from primary breast tissues and breast cancer cell lines was extracted using the RNeasy Mini Kit (Qiagen, Inc., Valencia, CA). First-strand cDNA was prepared from total RNA (0.5 µg) and oligo(dT) using the Superscript First-Strand Synthesis System (Invitrogen). For semiquantitative RT-PCR, Rad9 cDNA was coamplified with ribosomal protein S14 cDNA in a total reaction volume of 50 µL, each containing GeneAmp 1 x PCR Buffer II (Applied Biosystems, Inc., Foster City, CA), 1.5 mmol/L MgCl2, 200 µmol/L each of deoxynucleotide triphosphate (dNTP), 300 nmol/L each of Rad9 primers, 37.5 nmol/L each of S14 primers, 1.25 units AmpliTaq Gold (Applied Biosystems), and 1 µL first-strand cDNA template. The Rad9 and S14 primer sequences are listed in Table 1. To avoid possible amplification of contaminating genomic DNA, the forward Rad9 primer was designed to span an exon/exon boundary of the gene. The cycling conditions for linear amplification of PCR products were hot start at 95°C for 10 minutes, then 31 cycles at 94°C for 30 seconds and 60°C for 45 seconds, and a 10-minute final extension at 72°C. The predicted size of the products was 261 bp for Rad9 cDNA and 143 bp for S14 cDNA, and their authenticity was confirmed by nucleotide sequencing. PCR products were separated by 6% PAGE, and the intensity of each band was determined using the Quantity One software (ref. 14; Bio-Rad Laboratories, Hercules, CA).


View this table:
[in this window]
[in a new window]

  		Table 1. Nucleotide sequences of primers used

 
Genomic DNA extraction and quantitative PCR. Genomic DNA was available from 21 (14 of them had matched normal tissues) and 10 (all had matched normal tissues) breast tumor samples with and without Rad9 mRNA overexpression, respectively. DNA was extracted with the DNeasy Tissue Kit (Qiagen). The copy number of the Rad9 gene was determined by a Taqman-based quantitative PCR method using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems) as described previously (15). The forward and reverse primers as well as the Taqman probe were designed using the Primer Express Software (Applied Biosystems; Table 1). The ß-actin gene was coamplified as an internal control using the Taqman Assay endogenous control kit (Applied Biosystems) to enable normalization of the amount of DNA added to each sample well. Each 50-µL reaction in 1x Taqman Buffer A (Applied Biosystems) contained 3.5 mmol/L MgCl2; 200 µmol/L each of dATP, dCTP, and dGTP; 400 µmol/L dUTP; 50 nmol/L forward primer; 300 nmol/L reverse primer; 150 nmol/L Taqman probe; 2.5 µL ß-actin control kit; 1.25 units AmpliTaq Gold; and 0.5 unit AmpErase uracil N-glycosylase (UNG), together with 100 ng of DNA template. Each sample was analyzed in triplicate. The cycling conditions consisted of an initial incubation at 50°C for 2 minutes to allow the UNG to act, hot start at 95°C for 10 minutes followed by 45 cycles at 95°C for 15 seconds and 60°C for 1 minute. The expected size of the Rad9 gene fragment was 74 bp.

To examine whether the increase in Rad9 gene number is related to CCND1 amplification, the copy number of the CCND1 gene in the same 14 tumor samples was determined by a similar quantitative PCR method except for using 100 nmol/L reverse primer and 200 nmol/L Taqman probe (Table 1). The expected size of the CCND1 gene fragment was 63 bp.

Sodium bisulfite modification and sequencing. The methylation status of the Rad9 CpG island was determined by the bisulfite sequencing method. Genomic DNA (1 µg) was subjected to bisulfite modification using the CpGenome DNA Modification Kit (Chemicon International, Inc., Temecula, CA). Because the Rad9 CpG island spans over 900 bp from the promoter region to intron 2 of the gene, it was amplified as two overlapping regions. The first region contained the proximal Rad9 promoter and part of exon 1 [nucleotide (nt) –421 to 13 (with the first nt of ATG assigned as +1)]. The second region laid between nt –10 and 557; from a portion of the proximal promoter to part of intron 2. These two regions were amplified by either seminested or nested PCR. The first round PCR was conducted in a total reaction volume of 50 µL, each containing GeneAmp 1x PCR Buffer II, 1.5 mmol/L MgCl2, 200 µmol/L each of dNTP, 300 nmol/L each of primers, 1.25 units AmpliTaq Gold, and 5 µL of bisulfite-modified DNA as template. Primer sequences are listed in Table 1. The PCR conditions were hot start at 95°C for 10 minutes; then 40 cycles at 94°C for 30 seconds, 55°C for 1 minute, and 72°C for 1 minutes; and a 10-minute final extension at 72°C. One microliter of the first round PCR product was used as template in the nested PCR reaction. The reaction mix and cycling conditions were similar to previous except for a reduction to 35 cycles. As a control experiment, we did bisulfite sequencing of the PAX5ß promoter as described by Palmisano et al. (16) with DNA from MCF-7 cells as the starting template. PCR products were separated on 1.2% agarose gels and purified by the High Pure PCR Product Purification Kit (Roche Applied Science, Indianapolis, IN). The purified fragments were then subjected to direct sequencing using the DYEnamic ET Terminator Cycle Sequencing Kit (Amersham Biosciences, Piscataway, NJ). The levels of cytosine methylation of each CpG dinucleotide were determined by comparing the peak height of the cytosine signal with that of the cytosine plus thymine signal according to Melki et al. (17).

To confirm the direct bisulfite sequencing results, PCR products of the second region spanning nt –10 to 557 from the four pairs of tumor and normal breast tissues that showed hypermethylation as well as the MCF-7 cell line were cloned into the pGEM-T Easy vector (Promega Corp., Madison, WI) and the constructs were transformed into Escherichia coli JM109–competent cells (Promega). Plasmid DNA from five PCR clones of each sample was prepared using the QIAprep Spin Miniprep Kit (Qiagen) and sequenced using the T7 promoter and M13 reverse primers.

Studies of Breast Cancer Cell Line
Small interfering RNA transfection and cell proliferation assay. The role of Rad9 in breast cancer cell proliferation was examined by RNA interference–mediated gene silencing. Transfection of small interfering RNA (siRNA) was carried out using siPORT Amine (Ambion) according to the manufacturer's instructions. MCF-7 cells (2.3 x 105) were transfected with 75 pmol of a Silencer-validated Rad9 siRNA (ID 5114, Ambion) or a negative control siRNA (Ambion) in a total volume of 2.5 mL in six-well tissue culture plates. The negative control siRNA is a scrambled sequence that has no significant homology to the human genome. Untransfected cells were included to monitor the cytotoxicity of the transfection reagents. Cells were harvested for RT-PCR at 48 and 72 hours, or trypsinized for counting with a hemacytometer at 48, 72, and 96 hours after siRNA treatment.

5'-aza-2'-deoxycytidine treatment. To examine the role of DNA methylation in Rad9 mRNA expression, MCF-7 cells were treated with the demethylating agent 5'-aza-2'-deoxycytidine (DAC) as described by Palmisano et al. (16). Briefly, cells were seed at a density of 1 x 105 cells in six-well tissue culture plates and allowed to attach overnight. Cells were then treated with 1 µmol/L of DAC (Sigma-Aldrich Corp., St. Louis, MO) or an equal concentration of DMSO (Sigma-Aldrich) for 3 days before RT-PCR. To determine whether the reduced Rad9 mRNA expression is associated with demethylation, PCR products of the second region (nt –10/557) from MCF-7 cells before and after 1 µmol/L DAC treatment were analyzed by bisulfite sequencing.

Plasmid construction and transient transfection. To determine the possible effect of the intron 2 Sp1/3 binding site on Rad9 gene transcription, various Rad9 promoter-luciferase constructs were prepared and analyzed in MCF-7 cells. A full-length human Rad9 gene fragment spanning from the proximal promoter to the second intron (nt –533/609) was amplified. The reaction mix was similar to that used for bisulfite PCR of the Rad9 CpG island described above, with the addition of 5% DMSO and 0.1 µg of genomic DNA prepared from a normal breast tissue sample as template. The cycling conditions were also similar to previous but with an annealing temperature of 64°C. The PCR product was digested with KpnI and HindIII (New England Biolabs, Beverly, MA), purified by the High Pure PCR Product Purification Kit, and then cloned into the firefly luciferase reporter gene vector pGL3-Basic (Promega). Deletion constructs spanning nt –533/436, nt –533/187, nt –533/116, nt –533/34, and nt –533/–1 were prepared by PCR of the corresponding regions of the Rad9 gene with the full-length construct (i.e., nt –533/609) as template. Primer sequences are given in Table 1. A KpnI or HindIII restriction site was introduced into each primer for directional cloning. Thirty amplification cycles were made. The amplified fragments were cloned into the pGL3-Basic vector as for the full-length fragment.

All Rad9 promoter-luciferase constructs were transformed into E. coli JM109 competent cells. Plasmid DNA for transient transfection was prepared using the Qiagen Plasmid Midi Kit (Qiagen).

Transient transfection of promoter constructs was carried out using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Briefly, MCF-7 cells were seed at a density of 1 x 105 cells in 24-well tissue culture plates 1 day before transfection. Under serum-free conditions, 1 µg of Rad9 promoter-luciferase constructs or pGL3-Basic (as a control) was cotransfected with 12.5 ng of pRL-CMV (Promega) into the cells to enable normalization of various transfection efficiencies. The pRL-CMV vector contains a cytomegalovirus promoter to provide high-level expression of Renilla luciferase. After 6 hours of transfection, the medium was replaced with fresh culture medium supplemented with 10% fetal bovine serum (Invitrogen). Cells were harvested 48 hours after transfection and luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega).

Chromatin immunoprecipitation assay. To detect any in vivo binding of transcription factors to the 3' portion of intron 2 of the Rad9 gene, a chromatin immunoprecipitation (ChIP) assay was done using the ChIP Assay Kit (Upstate Biotechnology, Lake Placid, NY). Briefly, MCF-7 cells (1 x 106) were cross-linked with 1% formaldehyde (Sigma-Aldrich) at 37°C for 10 minutes. Cells were collected, resuspended in SDS lysis buffer (200 µL), and then sonicated with the Branson Sonifier 450 (Fisher Scientific, Los Angeles, CA). After centrifugation, supernatants were diluted 10-fold with ChIP dilution buffer and an aliquot of the diluted supernatant (1%) was saved as a positive control (input) for PCR. Protein-DNA complexes were immunoprecipitated with 10 µg of Sp1 or Sp3 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C overnight with rotation. A no-antibody immunoprecipitation was included as a negative control. Immunoprecipitated complexes were washed, eluted, and the cross-linking was reversed by heating the samples at 65°C for 4 hours. Samples were digested with proteinase K (20 µg, Sigma-Aldrich) for 1 hour at 45°C. DNA was recovered by phenol/chloroform extraction and ethanol precipitation. One tenth of the resuspended DNA (5 µL) was used for PCR. The reaction mix and cycling conditions were similar to those used for plasmid construction but with an annealing temperature of 58°C. Primer sequences are given in Table 1. The expected product was 200 bp in length and spanned from nt 406 to 605. PCR products were analyzed on a 1.5% agarose gel, and their authenticity was confirmed by nucleotide sequencing.

To examine the effect of DNA methylation on Sp3 binding to the intron 2 region, MCF-7 cells (5 x 105) were treated with 1 µmol/L of DAC or DMSO as previously described (see above) before ChIP analysis. The concentration of DNA in each sample was determined by spectrophotometry to ensure that a similar amount of DNA was used for immunoprecipitation in each case.

Analysis of data. For semiquantitative RT-PCR, the mRNA level of Rad9 in each sample was normalized against that of ribosomal protein S14. Relative gene expression level was determined by comparing the normalized Rad9 mRNA level in each tumor to that in the corresponding normal tissue obtained from the same patient. Data were expressed as mean ± SE from three independent experiments. Overexpression was defined as a 2-fold (or more) increase in Rad9 mRNA level in tumors compared with the corresponding normal tissues. For quantitative PCR, data were analyzed using the Sequence Detection Software (Applied Biosystems), and the mean of triplicates was used in each calculation. The relative Rad9 and CCND1 gene copy numbers in matched tumor and normal breast specimens were determined by the comparative CT method (18). Gene amplification was arbitrarily defined as ≥1.47-fold increase in gene copy number in tumors compared with the corresponding normal tissues. Unpaired t test was used to compare the mean of relative Rad9 mRNA level or gene copy number between groups. Pearson correlation test was used to analyze the correlation between Rad9 and CCND1 gene numbers.

For cell proliferation studies, data were expressed as mean ± SE from three independent assays. In the transfection experiment, firefly luciferase activity of each sample was normalized against Renilla luciferase activity. Relative luciferase activity was determined by comparing the normalized firefly luciferase activity of each construct to that of the pGL3-Basic vector. Data were expressed as mean ± SE from three independent experiments, each done in triplicate. All data were analyzed by one-way ANOVA followed by Tukey's multiple comparison test using GraphPad Prism 4 (GraphPad Software, Inc., San Diego, CA).

Data were considered statistically significant when P < 0.05.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Rad9 mRNA levels were frequently up-regulated in breast tumors. The mRNA level of Rad9 gene in matched tumor and normal breast tissues determined by semiquantitative RT-PCR are shown in Fig. 1. In 48 pairs of samples analyzed, all (both tumor and normal tissues) had detectable Rad9 mRNA expression. In addition, Rad9 mRNA was expressed in all three breast cancer cell lines tested (MCF-7, MDA-MB-231, and BT-483). Twenty-five tumor samples (52.1%) exhibited an up-regulation of Rad9 mRNA level (>2-fold) when compared with the adjacent normal breast tissues. In contrast, none of the tumor samples showed a significant reduction (>0.5-fold) in Rad9 mRNA expression. The mean ± SE Rad9 mRNA level of the 48 tumor samples was 3.37 ± 0.39. The level of Rad9 mRNA correlated positively with tumor size (P = 0.037) and local recurrence (P = 0.033) but not with other clinicopathologic variables including age, tumor grade, lymph node status, lymphatic vessel permeation, nuclear receptor immunoreactivity, and HER2 immunoreactivity (Table 2).



View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Representative results of semiquantitative RT-PCR analysis of Rad9 and ribosomal protein S14 mRNA expression in 10 primary breast tumors with Rad9 mRNA overexpression (left), two breast tumors without Rad9 mRNA overexpression (middle), and three breast cancer cell lines MCF-7, MDA-MB-231, and BT-483 (right). Abbreviations: N, normal sample; T, tumor sample.

 

View this table:
[in this window]
[in a new window]

  		Table 2. Relationship of Rad9 mRNA expression with clinicopathologic variables

 
Rad9 gene was amplified in breast tumors. Quantitative PCR revealed an increase in Rad9 gene copy number (≥1.5-fold) in 5 of 14 tumors that have Rad9 mRNA overexpression (Table 3). In contrast, none of the tumor samples (n = 10) without Rad9 mRNA overexpression had a significant increase in gene number (Table 3). Overall, amplification of the Rad9 gene occurred in 20.8% (5 of 24) of breast tumors examined. The mean ± SE Rad9 gene copy number was 1.51 ± 0.22 and 0.89 ± 0.08 for tumors with and without Rad9 mRNA overexpression, respectively. Increase in gene copy number was found to correlate significantly with mRNA overexpression (t = 2.31, unpaired t test; P = 0.031).


View this table:
[in this window]
[in a new window]

  		Table 3. Relative copy number of the Rad9 and CCND1 genes in breast tumors with or without Rad9 mRNA overexpression

 
The CCND1 gene copy number in the same 14 breast tumors is shown in Table 3. The relative number of CCND1 gene correlated with that of Rad9 gene (r = 0.643, Pearson correlation test; P = 0.013). However, there were a few discordant samples (3 of 14). CCND1 gene numbers were increased but Rad9 gene numbers were normal in samples T3 and T4. Conversely, Rad9 but not CCND1 gene number was increased in T13.

Certain CpG dinucleotides in the first and second introns of the Rad9 gene were differently methylated in breast tumors. Sequence analysis using the default settings (%GC = 55%, ObsCpG/ExpCpG = 0.65, length = 500 bp) of the CpG Island Searcher (19) revealed that the human Rad9 gene contains a putative 900-bp CpG island that extends from the proximal promoter region to part of the intron 2 of the gene (Fig. 2A). Bisulfite sequencing of the entire Rad9 CpG island showed that all 47 CpG dinucleotides in the promoter region and all nine CpG dinucleotides in the first two exons were unmethylated both in breast cancer MCF-7 cells and in all the primary breast tissues examined (24 pairs of matched tumor and normal samples plus seven tumor samples [without matched normal tissues]; Fig. 2B). Conversely, as a control, the PAX5ß promoter (31 CpG dinucleotides) was completely methylated in MCF-7 cells as reported previously (Fig. 2B; ref. 16). Certain CpG dinucleotides in the first and second introns of the Rad9 gene were found to exhibit partial to almost complete methylation patterns in MCF-7 cells and in four primary tumors (T2, T3, T8, and T10) that had exhibited Rad9 mRNA overexpression without increased gene copy number (Fig. 2B). Overall, aberrant DNA methylation occurred in 12.9% (4 of 31) of breast tumors examined. These results obtained by direct bisulfite sequencing were confirmed by sequencing five clones each of second region PCR products (nt –10/557) from the four pairs of tumor and normal breast tissues (T2, T3, T8, and T10) and MCF-7 cells (Fig. 2C). The three tumors (T2, T8, and T10) showing aberrant methylation at intron 2 were IDCs, whereas the tumor (T3) with aberrant methylation at intron 1 was DCIS.



View larger version (41K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Methylation analysis of the Rad9 gene in primary breast tumors and breast cancer cells. A, diagrammatic representation of the 5' portion of the Rad9 gene. Location of primers used for amplification of the Rad9 CpG island. Black arrows, primers used for seminested PCR of the promoter region and a portion of exon 1; white arrows, primers for nested PCR of the rest of the CpG island. Grey arrow, an internal primer for sequencing. Relative positions of the CpG dinucleotides are indicated as the vertical lines below (not drawn to scale). B, results of direct bisulfite sequencing of the Rad9 CpG island in four pairs of matched tumor and normal breast samples (three IDCs and one DCIS) and the breast cancer cell line MCF-7. Each square represents one CpG dinucleotide in the 5’-flanking region, exons 1 and 2 (E1 and E2), and introns 1 and 2 (I1 and I2) of the Rad9 gene. Varying degrees of methylation are indicated by different shading within the squares. As a control, methylation pattern of the PAX5ß promoter in MCF-7 cells was analyzed and all 31 CpG dinucleotides were methylated. C, confirmation of direct bisulfite sequencing with sequencing of cloned PCR products from the same four pairs of matched tumor and normal breast samples as in Fig. 2B. Direct bisulfite sequencing results (I1, E2, and I2 regions) shown in Fig. 2B are indicated. Each row of circles represents one PCR clone and five clones from each sample were sequenced. Unmethylated ({circ}) and methylated ({bullet}) CpG dinucleotides. Methylation patterns of either intron 1 or 2 are shown. D, comparison of the consensus Sp1/3 binding site with sequences surrounding the methylated CpG dinucleotides in introns 1 and 2 of the Rad9 gene. CpG dinucleotides that are methylated in breast tumors (underlined) and nucleotides sharing identity with the consensus Sp1/3 motif (boxed).

 
Rad9 gene silencing inhibited the proliferation of breast cancer cell line in vitro. Rad9 mRNA level of MCF-7 cells was drastically reduced 48 hours after Rad9 siRNA treatment and was barely detectable at 72 hours (Fig. 3A). In contrast, no reduction in Rad9 expression (compared with untransfected cells) was observed when the cells were treated with the negative control siRNA. Neither Rad9 nor the negative control siRNA affected ribosomal protein S14 mRNA level (Fig. 3A). Compared with the untransfected control, Rad9 siRNA transfection significantly decreased MCF-7 cell number by 51% (6.06 ± 0.21 x 105 versus 3.00 ± 0.32 x 105; P < 0.01) at 72 hours and 56% (8.54 ± 0.54 x 105 versus 3.72 ± 0.36 x 105; P < 0.001) at 96 hours, respectively (Fig. 3B). Transfection with the negative control siRNA had little effect (P > 0.05) both at 72 hours (5.04 ± 0.43 x 105) and 96 hours (7.52 ± 0.45 x 105; Fig. 3B).



View larger version (30K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Inhibition of breast cancer cell proliferation by Rad9 gene silencing. A,Rad9 mRNA expression in MCF-7 cells at 48 and 72 hours after siRNA transfection. In each case, the ribosomal protein S14 cDNA was coamplified as an internal control. The negative control siRNA is a scrambled sequence that has no significant homology to the human genome. An untransfected control was included to monitor the cytotoxicity of the transfection reagents. B, MCF-7 cell number ± SE at different time points after siRNA transfection (solid columns). *, P < 0.01 and **, P < 0.001 versus untransfected control.

 
5'-Aza-2'-deoxycytidine treatment inhibited Rad9 mRNA expression in breast cancer cell line. The functional significance of intron 2 hypermethylation in Rad9 gene expression was confirmed by a reduction in Rad9 mRNA level in MCF-7 cells (that had shown aberrant methylation in the second intron), after treatment with DAC (1 µmol/L) for 3 days (Fig. 4A). In contrast, no significant change in ribosomal protein S14 mRNA level could be observed. Bisulfite sequencing showed that the reduced Rad9 mRNA expression in MCF-7 cells in response to DAC was associated with demethylation of the second intron of the gene (Fig. 4B).



View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Effects of DAC treatment on Rad9 mRNA expression and intron 2 methylation in breast cancer MCF-7 cells. A, inhibition of Rad9 mRNA expression in MCF-7 cells after treatment with 1 µmol/L of DAC for 3 days compared with cells treated with an equivalent concentration of DMSO. In each case, the ribosomal protein S14 cDNA was coamplified as an internal control. B, bisulfite sequencing of intron 2 of the Rad9 gene in MCF-7 cells before (top) and after (bottom) treatment with DAC. Only sequences between nt 429 and 478 are shown (reverse orientation). CpG dinucleotide that was demethylated by DAC (boxed).

 
The 3' portion of intron 2 exerted a silencing effect on Rad9 promoter activity and bound to Sp3 in vivo. The Rad9 promoter (nt –533/–1) generated a high level of luciferase activity [78.3-fold compared with pGL3-Basic (vector only), P < 0.001] in breast cancer MCF-7 cells (Fig. 5). Inclusion of exon 1 (nt –533/34), intron 1 (nt –533/116), exon 2 (nt –533/187), and part of the intron 2 (nt –533/436) had no significant effect on Rad9 promoter activity (Fig. 5). However, further addition of the 3' portion of intron 2 (nt –533/609), containing the Sp1/3 binding site (that exhibited aberrant methylation in certain breast tumors and MCF-7 cells), reduced the promoter activity from 63.5- to 29.4-fold (P < 0.001; Fig. 5).



View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5. Transcriptional activities of different regions of the human Rad9 gene in breast cancer MCF-7 cells. Left, diagrammatic representation of the 5'-flanking region, exons 1 and 2 (E1 and E2), and introns 1 and 2 (I1 and I2) of the Rad9 gene. The location of the intron 2 Sp1/3 site that exhibited aberrant methylation in certain breast tumors and MCF-7 cells is also indicated. Each arrow represents the size and direction of the Rad9 gene fragment cloned into the pGL3-Basic vector. Right, columns, relative luciferase activity of each construct compared with pGL3-Basic (solid columns); bars, ±SE. a, P < 0.001 versus pGL3-Basic; b, P < 0.001 versus nt –533/609.

 
ChIP assays showed that a 200-bp DNA fragment corresponding to the 3' portion of intron 2 of the Rad9 gene (nt 406/605) was amplified from MCF-7 cells when chromatin was immunoprecipitated with Sp3 antibody (Fig. 6A). The Rad9 gene fragment could also be amplified from the positive control (input), whereas no PCR product was obtained when Sp1 or no antibody (negative control) were used for immunoprecipitation (Fig. 6A). Compared with DMSO, treatment of MCF-7 cells with DAC (1 µmol/L) for 3 days decreased Rad9 mRNA (Fig. 4A), with an increase in Sp3 binding to the intron 2 region (Fig. 6B).



View larger version (36K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6. ChIP analysis of in vivo binding of transcription factors to the 3' portion of intron 2 of the Rad9 gene in breast cancer MCF-7 cells. A, gel electrophoresis of PCR product (a 200-bp Rad9 gene fragment) from purified input DNA (positive control) and chromatin immunoprecipitated with Sp3, Sp1, or no antibody (negative control). B, increased Sp3 binding to the intron 2 region in MCF-7 cells after treatment with 1 µmol/L of DAC for 3 days compared with cells treated with an equivalent concentration of DMSO. A similar amount of DNA from each sample was used for immunoprecipitation.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The chromosomal region 11q13 is frequently amplified in breast cancer and this amplification is associated with a poor outcome (8, 9). It is generally thought that 11q13 amplification results in overexpression of genes that are critical for tumor growth (20). One of these genes is CCND1, whose overexpression can shorten the G1-to-S phase interval and render cells less dependent on growth factors (21). Several studies have indicated that the 11q13 amplicon is unusually large and amplification patterns in this region are highly complex (10, 11). Thus far, two oncogenes, CCND1 and EMS1, have been identified (11, 12). In the present study, we showed that the mRNA level of another 11q13 gene, Rad9, was frequently up-regulated in breast tumors and that this up-regulation in some but not all tumors was due to an increase in gene copy number. Our data also showed that overexpression of Rad9 mRNA correlated with a larger tumor size, suggesting that the cell cycle checkpoint protein may promote breast cancer proliferation. Consistent with this, silencing Rad9 expression could inhibit the proliferation of breast cancer cells in vitro (Fig. 3). Therefore, it seems that Rad9 is another candidate gene activated by 11q13 amplification, conferring selective growth advantage to breast tumor cells.

The copy number of Rad9 gene correlated with that of CCND1 gene, although there were three discordant results out of 14 samples (Table 3). This coamplification of both genes in human breast cancer may be expected because of the known amplification of the 11q13 region (10, 11), where both genes reside.

Similar to CCND1 (22, 23), increase in Rad9 gene copy number was only found in a portion of the breast tumors with mRNA overexpression (Table 3), suggesting that mechanisms other than gene amplification may also be responsible. We showed that certain CpG dinucleotides at introns 1 and 2 of the Rad9 gene were hypermethylated in some breast tumors and MCF-7 cells and demethylation of intron 2 could reduce this mRNA expression (Fig. 2A-C and Fig. 4). Sequence analysis revealed that these CpG dinucleotides reside within two putative Sp1/3 binding sites, which share 78% sequence identity with the consensus Sp1/3 motif (Fig. 2D). DNA hypermethylation in the exons as well as introns of a gene can either silence (24, 25) or enhance (26) its expression. We have confirmed, through measurement of transcriptional activities of different regions of the Rad9 gene, that the 3' portion of intron 2 (inclusive of the Sp1/3 site) could suppress Rad9 promoter activity (silencer effect; Fig. 5). In addition, we have shown that Sp3, which contains an inhibitory domain to repress gene transcription (27), binds to this intron 2 region of the gene in vivo (Fig. 6A). This suggests that the Sp1/3 site may act as a silencer. Conversely, methylation of cytosine residues within the site (shown in some breast tumors) would interfere with Sp3 (or other transcription factors) binding, thereby conferring an enhancer effect on Rad9 gene transcription. This is further substantiated by an increased Sp3 binding to the intron 2 region after demethylation with DAC (Fig. 6B). The functional relevance of hypermethylation of the two CpG dinucleotides in the first intron of the Rad9 gene is unclear. Because this phenomenon was only observed in DCIS but not in IDCs, it may represent an early methylation event during breast cancer progression. Studies of more breast tumor samples are needed to justify this view.

It is important to note that some breast tumors with Rad9 mRNA overexpression exhibit neither gene amplification nor aberrant DNA methylation (Table 4), indicating that other as yet unidentified mechanisms may be involved. These mechanisms may include chromosomal translocation and steroid hormone regulation, which have been shown to activate the two 11q13 genes, CCND1 and THRSP (encoding thyroid hormone–responsive protein; refs. 28, 29).


View this table:
[in this window]
[in a new window]

  		Table 4. Summary of gene amplification and aberrant DNA methylation in breast tumors showing Rad9 mRNA overexpression

 
In a small study of three prostate tumors, a lower level of Rad9 mRNA expression was found in tumors compared with the corresponding normal tissues (6). hRad9 has been shown to interact with androgen receptor and suppress its activation in prostate cancer cells (6). The authors suggested that the reduction in Rad9 mRNA expression in prostate tumors would promote androgen-induced cell proliferation. However, in the same study, hRad9 was found to have no interaction with estrogen receptor (6). This may explain the discrepant role of hRad9 in prostate and breast tumor proliferation. It seems that hRad9 may act in concert with other protein factors to mediate tissue-specific responses. Identification of these factors will shed insights into the mechanism of hRad9 action in breast cancer development.

The significance of Rad9 on the survival of breast cancer patients has not been fully addressed in this study due to the short follow-up period for our patients. However, overexpression of its mRNA showed correlation with an increased risk of local recurrence. In addition, based on the proapoptotic activity of hRad9 (5), breast tumors expressing more Rad9 mRNA may exhibit a higher rate of apoptosis. Because increased apoptosis is associated with reduced survival (30), Rad9 mRNA overexpression may predict a worse outcome in breast cancer. Long-term follow-up studies will confirm the prognostic value of Rad9 in the disease.

In conclusion, our present findings showed that overexpression of Rad9 mRNA is common in breast cancer and this change has implications in tumor proliferation and local invasion. Gene amplification and aberrant DNA methylation are two mechanisms causing the overexpression. These observations suggest that Rad9 is a new oncogene candidate on 11q13 that may have a role in breast cancer progression.


   Acknowledgments
 
Grant support: Committee on Research and Conference (C.K. Cheng and V. Chan) and Vice Chancellor's Development Fund (V. Chan), University of Hong Kong.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 11/29/04. Revised 7/12/05. Accepted 7/26/05.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Lieberman HB, Hopkins KM, Nass M, Demetrick D, Davey S. A human homolog of the Schizosaccharomyces pombe rad9+ checkpoint control gene. Proc Natl Acad Sci U S A 1996;93:13890–5.[Abstract/Free Full Text]
  2. Volkmer E, Karnitz LM. Human homologs of Schizosaccharomyces pombe rad1, hus1, and rad9 form a DNA damage-responsive protein complex. J Biol Chem 1999;274:567–70.[Abstract/Free Full Text]
  3. Rauen M, Burtelow MA, Dufault VM, Karnitz LM. The human checkpoint protein hRad17 interacts with the PCNA-like proteins hRad1, hHus1, and hRad9. J Biol Chem 2000;275:29767–71.[Abstract/Free Full Text]
  4. Zou L, Cortez D, Elledge SJ. Regulation of ATR substrate selection by Rad17-dependent loading of Rad9 complexes onto chromatin. Genes Dev 2002;16:198–208.[Abstract/Free Full Text]
  5. Komatsu K, Miyashita T, Hang H, et al. Human homologue of S. pombe Rad9 interacts with Bcl-2/Bcl-xL and promotes apoptosis. Nat Cell Biol 2000;2:1–6.[CrossRef][Medline]
  6. Wang L, Hsu CL, Ni J, et al. Human checkpoint protein hRad9 functions as a negative coregulator to repress androgen receptor transactivation in prostate cancer cells. Mol Cell Biol 2004;24:2202–13.[Abstract/Free Full Text]
  7. Schwab M. Amplification of oncogenes in human cancer cells. Bioessays 1998;20:473–9.[CrossRef][Medline]
  8. Tsuda H, Hirohashi S, Shimosato Y, et al. Correlation between long-term survival in breast cancer patients and amplification of two putative oncogene-amplification units: hst-1/int-2 and c-erbB-2/ear-1. Cancer Res 1989;49:3104–8.[Abstract/Free Full Text]
  9. Borg A, Sigurdsson H, Clark GM, et al. Association of INT2/HST1 coamplification in primary breast cancer with hormone-dependent phenotype and poor prognosis. Br J Cancer 1991;63:136–42.[Medline]
  10. Karlseder J, Zeillinger R, Schneeberger C, et al. Patterns of DNA amplification at band q13 of chromosome 11 in human breast cancer. Genes Chromosomes Cancer 1994;9:42–8.[Medline]
  11. Ormandy CJ, Musgrove EA, Hui R, Daly RJ, Sutherland RL. Cyclin D1, EMS1 and 11q13 amplification in breast cancer. Breast Cancer Res Treat 2003;78:323–35.[CrossRef][Medline]
  12. Schuuring E, Verhoeven E, Mooi WJ, Michalides RJ. Identification and cloning of two overexpressed genes, U21B31/PRAD1 and EMS1, within the amplified chromosome 11q13 region in human carcinomas. Oncogene 1992;7:355–61.[Medline]
  13. Bloom HJ, Richardson WW. Histological grading and prognosis in breast cancer. Br J Cancer 1957;11:359–77.[Medline]
  14. Kurebayashi J, Otsuki T, Kunisue H, Tanaka K, Yamamoto S, Sonoo H. Expression levels of estrogen receptor-{alpha}, estrogen receptor-ß, coactivators, and corepressors in breast cancer. Clin Cancer Res 2000;6:512–8.[Abstract/Free Full Text]
  15. Chan V, Yip B, Tse HY, Wong HS, Chan TK. Quantitative polymerase chain reaction for the rapid prenatal diagnosis of homozygous {alpha}-thalassaemia (Hb Barts hydrops fetalis). Br J Haematol 2001;115:341–6.[CrossRef][Medline]
  16. Palmisano WA, Crume KP, Grimes MJ, et al. Aberrant promoter methylation of the transcription factor genes PAX5 {alpha} and ß in human cancers. Cancer Res 2003;63:4620–5.[Abstract/Free Full Text]
  17. Melki JR, Vincent PC, Clark SJ. Concurrent DNA methylation of multiple genes in acute myeloid leukemia. Cancer Res 1999;59:3730–40.[Abstract/Free Full Text]
  18. Zhu CQ, Blackhall FH, Pintilie M, et al. Skp2 gene copy number aberrations are common in non-small cell lung carcinoma, and its overexpression in tumors with ras mutation is a poor prognostic marker. Clin Cancer Res 2004;10:1984–91.[Abstract/Free Full Text]
  19. Takai D, Jones PA. Comprehensive analysis of CpG islands in human chromosomes 21 and 22. Proc Natl Acad Sci U S A 2002;99:3740–5.[Abstract/Free Full Text]
  20. Moncur JT, Park JP, Memoli VA, Mohandas TK, Kinlaw WB. The "Spot 14" gene resides on the telomeric end of the 11q13 amplicon and is overexpressed in lipogenic breast cancers: implications for control of tumor metabolism. Proc Natl Acad Sci U S A 1998;95:6989–94.[Abstract/Free Full Text]
  21. Musgrove EA, Lee CS, Buckley MF, Sutherland RL. Cyclin D1 induction in breast cancer cells shortens G1 and is sufficient for cells arrested in G1 to complete the cell cycle. Proc Natl Acad Sci U S A 1994;91:8022–6.[Abstract/Free Full Text]
  22. Buckley MF, Sweeney KJ, Hamilton JA, et al. Expression and amplification of cyclin genes in human breast cancer. Oncogene 1993;8:2127–33.[Medline]
  23. Gillett C, Fantl V, Smith R, et al. Amplification and overexpression of cyclin D1 in breast cancer detected by immunohistochemical staining. Cancer Res 1994;54:1812–7.[Abstract/Free Full Text]
  24. Nagai H, Li Y, Hatano S, et al. Mutations and aberrant DNA methylation of the PROX1 gene in hematologic malignancies. Genes Chromosomes Cancer 2003;38:13–21.[CrossRef][Medline]
  25. Douglas DB, Akiyama Y, Carraway H, et al. Hypermethylation of a small CpGuanine-rich region correlates with loss of activator protein-2{alpha} expression during progression of breast cancer. Cancer Res 2004;64:1611–20.[Abstract/Free Full Text]
  26. Unoki M, Nakamura Y. Methylation at CpG islands in intron 1 of EGR2 confers enhancer-like activity. FEBS Lett 2003;554:67–72.[CrossRef][Medline]
  27. Dennig J, Beato M, Suske G. An inhibitor domain in Sp3 regulates its glutamine-rich activation domains. EMBO J 1996;15:5659–67.[Medline]
  28. Rosenberg CL, Kim HG, Shows TB, Kronenberg HM, Arnold A. Rearrangement and overexpression of D11S287E, a candidate oncogene on chromosome 11q13 in benign parathyroid tumors. Oncogene 1991;6:449–53.[Medline]
  29. Heemers H, Vanderhoydonc F, Heyns W, Verhoeven G, Swinnen JV. Progestins and androgens increase expression of Spot 14 in T47-D breast tumor cells. Biochem Biophys Res Commun 2000;269:209–12.[CrossRef][Medline]
  30. Vakkala M, Lahteenmaki K, Raunio H, Paakko P, Soini Y. Apoptosis during breast cancer progression. Clin Cancer Res 1999;5:319–24.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Mol Biol EvolHome page
R. Farcas, E. Schneider, K. Frauenknecht, I. Kondova, R. Bontrop, J. Bohl, B. Navarro, M. Metzler, H. Zischler, U. Zechner, et al.
Differences in DNA Methylation Patterns and Expression of the CCRK Gene in Human and Nonhuman Primate Cortices
Mol. Biol. Evol., June 1, 2009; 26(6): 1379 - 1389.
[Abstract] [Full Text] [PDF]

Home page
 J. Virol.Home page
S. Kim, N. Kim, B. Dong, D. Boren, S. A. Lee, J. Das Gupta, C. Gaughan, E. A. Klein, C. Lee, R. H. Silverman, et al.
Integration Site Preference of Xenotropic Murine Leukemia Virus-Related Virus, a New Human Retrovirus Associated with Prostate Cancer
J. Virol., October 15, 2008; 82(20): 9964 - 9977.
[Abstract] [Full Text] [PDF]

Home page
 BloodHome page
C. K. Cheng, L. Li, S. H. Cheng, K. M. Lau, N. P. H. Chan, R. S. M. Wong, M. M. K. Shing, C. K. Li, and M. H. L. Ng
Transcriptional repression of the RUNX3/AML2 gene by the t(8;21) and inv(16) fusion proteins in acute myeloid leukemia
Blood, October 15, 2008; 112(8): 3391 - 3402.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
Z. Hu, Y. Liu, C. Zhang, Y. Zhao, W. He, L. Han, L. Yang, K. M. Hopkins, X. Yang, H. B. Lieberman, et al.
Targeted Deletion of Rad9 in Mouse Skin Keratinocytes Enhances Genotoxin-Induced Tumor Development
Cancer Res., July 15, 2008; 68(14): 5552 - 5561.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
A. Zhu, C. X. Zhang, and H. B. Lieberman
Rad9 Has a Functional Role in Human Prostate Carcinogenesis
Cancer Res., March 1, 2008; 68(5): 1267 - 1274.
[Abstract] [Full Text] [PDF]

Home page
 Genome ResHome page
N. Paz, E. Y. Levanon, N. Amariglio, A. B. Heimberger, Z. Ram, S. Constantini, Z. S. Barbash, K. Adamsky, M. Safran, A. Hirschberg, et al.
Altered adenosine-to-inosine RNA editing in human cancer
Genome Res., November 1, 2007; 17(11): 1586 - 1595.
[Abstract] [Full Text] [PDF]

Home page
 J. Nutr.Home page
S. Ollier, C. Robert-Granie, L. Bernard, Y. Chilliard, and C. Leroux
Mammary Transcriptome Analysis of Food-Deprived Lactating Goats Highlights Genes Involved in Milk Secretion and Programmed Cell Death
J. Nutr., March 1, 2007; 137(3): 560 - 567.
[Abstract] [Full Text] [PDF]

Home page
 Mol. Cell. Biol.Home page
R. K. Pandita, G. G. Sharma, A. Laszlo, K. M. Hopkins, S. Davey, M. Chakhparonian, A. Gupta, R. J. Wellinger, J. Zhang, S. N. Powell, et al.
Mammalian rad9 plays a role in telomere stability, s- and g2-phase-specific cell survival, and homologous recombinational repair.
Mol. Cell. Biol., March 1, 2006; 26(5): 1850 - 1864.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cheng, C. K.
Right arrow Articles by Chan, V.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cheng, C. K.
Right arrow Articles by Chan, V.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online