Cancer Research Annual Meeting 2010 Jordan
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 Chang, B.-l.
Right arrow Articles by Xu, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chang, B.-l.
Right arrow Articles by Xu, J.
[Cancer Research 64, 1997-1999, March 15, 2004]
© 2004 American Association for Cancer Research

Regular Articles

A Polymorphism in the CDKN1B Gene Is Associated with Increased Risk of Hereditary Prostate Cancer

Bao-li Chang1, Siqun L Zheng1, Sarah D. Isaacs2, Kathy E. Wiley2, Aubrey Turner1, Ge Li1, Patrick C. Walsh2, Deborah A. Meyers1, William B. Isaacs2 and Jianfeng Xu1

1 Center for Human Genomics, Wake Forest University School of Medicine, Winston-Salem, North Carolina, and 2 Department of Urology, Johns Hopkins Medical Institutions, Baltimore, Maryland


   ABSTRACT
Top
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
The loss of cell cycle control is believed to be an important mechanism in the promotion of carcinogenesis. CDKN1B (p27) belongs to the Cip/Kip family and functions as an important cell cycle gatekeeper. Several lines of evidence from clinical studies and laboratory experiments demonstrate that CDKN1B is an important tumor suppressor gene in prostate cancer etiology. In addition, a case-control study has shown that the 326T/G (V109G) polymorphism in CDKN1B is associated with advanced prostate cancer. In light of the evidence for linkage between the chromosomal location of the CDKN1B gene (12p13) and prostate cancer susceptibility in several hereditary prostate cancer (HPC) populations, we hypothesized that sequence variants of CDKN1B play a role in HPC. To test this hypothesis, we first resequenced this gene in 96 HPC probands to identify germ-line mutations and sequence variants. We then genotyped the identified sequence variants among all family members of 188 HPC families and tested for their cosegregation with prostate cancer. In total, 10 sequence variants were identified, including three nonsynonymous changes. A family-based test, which is free from the effects of population stratification, revealed a significant association between single nucleotide polymorphism (SNP) -79C/T and prostate cancer (with a nominal P of 0.0005). The C allele of -79C/T was overtransmitted from parents to their affected offspring. Evidence for this association was primarily contributed by affected offspring whose age at diagnosis was <65 years. Together with the previous association study in a sporadic prostate cancer population, our new findings additionally suggest that germ-line variants of this gene play a role in prostate cancer susceptibility.


   INTRODUCTION
Top
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
The loss of cell cycle control accompanied by replication of a defective genome facilitates the emergence of cancer cells. CDKN1B (p27) belongs to the cyclin-dependent kinase (CDK) inhibitor Cip/Kip family and functions as a cell cycle check point at G1-S by regulating G1 cyclin-CDK complexes (1) . In prostate tumors, frequent genetic aberrations were observed at the chromosomal region 12p12-13, harboring CDKN1B. Indeed, a homozygous deletion region was identified in a metastatic prostate cancer xenograft (2) . The critical role of CDKN1B in regulating cell proliferation and as a tumor suppressor was directly demonstrated in the CDKN1B knockout mouse, which exhibits gigantism (because of increased cell number) and increased tumorigenesis (3, 4, 5) .

Recently, in a study of 96 prostate cancer cases and 106 controls, Kibel et al. (6) reported an association between a nonsynonymous change of CDKN1B and the risk of advanced prostate cancer. The single nucleotide polymorphism (SNP) 326T/G changes a valine at codon 109 to glycine. Among men with the TT genotype, an association with increased risk of advanced prostate carcinoma was reported (odds ratio = 1.95; 95% confidence interval = 1.09–3.47), and the association was stronger in patients with androgen-independent disease (odds ratio = 2.88, 95% confidence interval = 1.19–6.97). Other results implicating CDKN1B and the risk for prostate cancer have come from linkage studies. The chromosomal region of CDKN1B (12p13) has been reported to be linked to prostate cancer susceptibility in several hereditary prostate cancer (HPC) family studies (7 , 8) . In our HPC study population, no evidence for linkage was observed at 12p13 in the complete set of the 188 HPC families based on the log of odds (LOD = 0); however, significant evidence for linkage at this region was found among the 34 families with the highest linkage score at 10q23 (the chromosomal region of PTEN), with a LOD score of 2.75 (P = 0.0002). This conditional linkage analysis was prompted by a mouse model, which demonstrated that the combined effects of both the Pten and Cdkn1b genes rather than a single gene cause prostate cancer in mice (9) . On the basis of the critical role CDKN1B plays in cell cycle control, the strong functional and clinical implications of this gene in prostate carcinogenesis, and the positive findings of association and linkage studies, we hypothesized that germ-line mutations or polymorphisms in CDKN1B may modify an individual’s risk to prostate cancer, and this may lead to familial aggregation of prostate cancer. To test this hypothesis, we first screened for germ-line mutations in 96 probands from HPC families, followed by genotyping among all 188 HPC families for the identified mutations and polymorphisms that were either rare or of potential functional importance. To evaluate the risk CDKN1B imparts for HPC, cosegregation of the mutations and polymorphisms in the HPC families was evaluated using family-based linkage and association tests. This study is unique in describing the CDKN1B mutation spectrum and its implications in HPC.


   MATERIALS AND METHODS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Subjects.
HPC families (n = 188) were recruited at the Brady Urology Institute at Johns Hopkins Hospital (Baltimore, MD) through referrals, review of medical records for patients seen at Johns Hopkins Hospital for treatment of prostate cancer, and respondents to various lay publications describing our studies. The eligibility criterion for HPC was at least three first-degree relatives affected with prostate cancer. The diagnosis of prostate cancer was verified by medical records for each affected male studied. Age of diagnosis of prostate cancer was confirmed either through medical records or from two other independent sources. The mean age at diagnosis was 64.4 years, with 91 families having mean age at diagnosis < 65 years. One hundred fifty-four families are non-Ashkenazi white, 17 families are Ashkenazi, and 15 families are black. The number of families with 3, 4, and >=5 affected individuals was 28, 47, and 113, respectively. All individuals that participated in this study gave full informed consent.

The family history for other types of cancers was also collected. For the diagnoses of other types of cancers, the majority was confirmed through interview with at least two relatives, and a small proportion (7%) was verified by medical records. Among the 188 HPC families, 68 (36.2%), 36 (19.1%), and 22 (11.7%) have at least one, two, and three or more other types of cancers in the family, respectively. Seventy-three of the prostate cancer cases were also diagnosed with another primary cancer. There were 148 other subjects diagnosed with other cancers. Colon cancer was the highest concurrent cancer (35 cases), followed by breast cancer (34 cases) and lung cancer (20 cases).

Sequencing and Genotyping Methods.
The PCR products of all three exons, exon-intron junctions, and ~800-bp promoter region of CDKN1B were directly sequenced in 96 HPC probands. The primers used for PCR and the components and cycling conditions for PCR are available upon request. All PCR products were purified using the QuickStep PCR purification Kit (Edge BioSystems, Gaithersburg, MD) to remove deoxynucleoside triphosphates and excess primers. All sequencing reactions were performed using dye-terminator chemistry (BigDye; ABI, Foster City, CA), and separated in an ABI 3700 DNA Analyzer. SNPs were identified using Sequencher software version 4.0.5 (Gene Codes Corporation).

SNPs were genotyped using the MassARRAY system (Sequenom, Inc., San Diego, CA). All reaction components and conditions were performed according to the manufacture’s instruction (Sequenom, Inc.) and are available upon request.

Statistical Methods.
Hardy-Weinberg equilibrium tests and pairwise linkage disequilibrium tests for all possible pairs of the variants were performed using the Fisher probability test statistic, as described by Weir (10) . For each test, 10,000 permutations were performed, and the test statistic of each replicate was calculated. Empirical Ps for each test were estimated as the proportion of replicates that are as probable as or less probable than the observed data, as implemented in the software package Genetic Data Analysis.

Family-based association tests were performed for the SNPs genotyped in the 188 HPC families, using the family-based association test software package (11) . Family-based association test uses data from nuclear families, sibships, or a combination of the two to test for linkage and linkage disequilibrium (association) between traits and genotypes. The test for linkage is valid when multiple affected members/pedigree are used, and the power to detect linkage is increased if there is an association. The test for association is valid if one affected member/pedigree is used (the genotypes of all of the affected members can be included) or if the empirical variance is used to account for correlation between transmissions in families when linkage is present.


   RESULTS AND DISCUSSION
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
A total of ten sequence variants were identified when we screened CDKN1B for germ-line mutations in the ~800 bp promoter region, all three exons, and the exon-intron junctions among 96 HPC probands (Table 1)Citation . Among these variants, three are in the promoter region, one is in the 5'-untranslated region, four are in the coding region of exon 1, one is in intron 2, and one is in the 3'-untranslated region. Three of the four SNPs in the coding sequence cause amino acid changes: SNP 258G/C changes a glutamic acid at codon 86 to aspartic acid (E86D); SNP 326T/G changes a valine at codon 109 to glycine (V109G); and SNP 356T/C changes an isoleucine at codon 119 to threonine (I119T). The two nonsynonymous changes 258G/C (E86D) and 356T/C (I119T) are rare, and each occurred only once among the probands. The nonsynonymous change 326T/G (V109G) is frequent and has been reported to be associated with increased sporadic prostate cancer risk (6) . Interestingly, only 4 of these 10 SNPs were previously reported in the public database.


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

  		Table 1 Sequence variants identified in CDKN1B among 96 Caucasian hereditary prostate cancer probands

 
To preliminarily evaluate the functional impact of these identified SNPs, we performed bioinformatic analyses using various computational methods. In a search of the TRANSFAC database, SNPs -1220T/G, -987C/T, and -838C/A were found to lie in potential binding sites for transcription factors Pbx-1a, SP1, and E2F, respectively. The SNP -79C/T was found to eliminate one CpG site in a CpG island. All three nonsynonymous changes were found to be located in the cyclin dependent kinase inhibitor (CDI) domain (pfam02234), which is the CDK-CDKN interaction motif. In addition, missense changes V109G and I119T might alter the interaction between CDKN1B and its negative regulator p38jab1 because they are located in the interaction surface that spans amino acid residues 97 to 151 (6 , 12) . When the exonic SNPs were evaluated for changes in exonic splice enhancer elements, SNPs -79C/T and 258G/C were found to be in these potentially important areas.

To assess whether the identified nonsynonymous changes and other sequence variants are associated with prostate cancer risk, we genotyped 9 of these 10 variants among all 188 HPC families. The SNP -987C/T was excluded because a reliable assay could not be developed. Hardy-Weinberg equilibrium and pairwise linkage disequilibrium tests were performed among unrelated probands of non-Ashkenazi whites. Eight of the 9 remaining SNPs were in Hardy-Weinberg equilibrium (P > 0.05). The SNP -838C/A significantly deviated from Hardy-Weinberg equilibrium (P < 0.005) and therefore was dropped from additional analysis. The family-based association test was used to test whether the sequence variants in CDKN1B are associated with prostate cancer risk. Three SNPs were rare (114C/T, 258G/C, and 356T/C) and were uninformative for this analysis. Among the five remaining SNPs, -79C/T in the 5'-untranslated region was significantly associated with prostate cancer, and the C allele was found to be overtransmitted from parents to affected offspring (nominal P = 0.0005; Table 2Citation ). Eighty HPC families were informative for the test of this SNP. We further explored this result by performing a stratified analysis based on age of diagnosis. Interestingly, the association was primarily observed among affected offspring with age at diagnosis < 65 years (P = 0.0015). There was no significant association among affected offspring with age at diagnosis >= 65 years (P = 0.34). It is worth noting that age of diagnosis is a crude indicator for age of onset and is not a good surrogate for clinically aggressive disease. Because the empirical variance, which accounts for correlation between transmission in families when linkage is present (13) , was used in the statistical test, this significant result reflects an association between SNP -79C/T and prostate cancer. The significant association remained when the analysis was limited to non-Ashkenazi white families only (P = 0.02). None of the other SNPs was significantly linked or associated with prostate cancer.


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

  		Table 2 Family based association test for CDKN1B sequence variants among 188 hereditary prostate cancer families

 
Because CDKN1B has been implicated in other types of cancers, we investigated this possibility in the HPC families by testing the hypothesis that CDKN1B functions as a general tumor suppressor gene for all types of cancers. A family-based association test was used to examine the presence of an association between the SNP -79C/T and the risk to all types of cancers, including prostate cancer. A significant association remained when including other cancers in the test (P = 0.0009), implicating the SNP -79C/T as contributing not only to the risk of prostate cancer but also to other cancers among HPC families. To explore the possibility that the association between SNP -79C/T and the risk to other cancers is independent of prostate cancer, we included all cancers except for prostate cancer in an additional analysis. After excluding prostate cancer cases, no association was observed between SNP -79C/T and the risk to other cancers (P = 0.61). These results suggest CDKN1B plays a role as a general tumor suppressor gene, and its tumor suppressor function is particularly profound for prostate cancer.

The SNP -79C/T is located in the 5'-regulatory CpG island. Previous studies indicate that methylation of the CDKN1B promoter, although occurring in some cancer cells (14) , is rare in prostate carcinomas (15 , 16) , and the methylation that does occur in prostate cancer typically does not affect this particular CpG dinucleotide. Although a primary focus of CDNK1B gene regulation studies has been the increase in protein degradation rates mediated by SKP2 ubiquitin ligase complex, ultimately decreasing CDKN1B protein in prostate cancer (17) , CDKN1B protein levels can also be affected by the rates of gene transcription (18) as well as mRNA translation rates. Interestingly, this latter mechanism has been shown to be affected by sequences in the 5'-untranslated region of the CDKN1B transcript (19, 20, 21) . Whether the basal transcription or transcript translation rates are affected by the -79C/T polymorphism is unknown at present.

The evidence for cosegregation of the two rare nonsynonymous mutations with prostate cancer in our HPC families is inconclusive. For the SNP 258G/C (E86D), the variant allele C was only identified in a proband but not in other family members available for study. For the SNP 356T/C (I119T), all four affected siblings were found to be carriers of the variant allele C. In addition, one of the sisters of the proband was diagnosed with lung cancer and found to be a C allele carrier. However, the unaffected brother and three unaffected nephews (ages 42, 60, and 65 years) of the proband also carried the C allele. Additional follow-up studies on these two rare variants may clarify their roles in HPC.

Recently, in a study by Kibel et al. (6) , which included 96 cases and 106 controls from a European-American population, the SNP 326T/G (V109G) was found to be associated with the risk of advanced prostate cancer, with the genotype TT significantly increasing the risk of advanced disease (odds ratio = 1.95; 95% confidence interval = 1.09–3.47). In their study, the variant T allele frequency was found to be 72.4% in controls and 82.3% in cases. We observed a similar frequency for the variant T allele in our collection of unrelated probands of non-Ashkenazi white families with HPC (81%; Table 1Citation ). Furthermore, the T allele at SNP 326T/G was found to be overtransmitted, although not significantly, to affected offspring (observed S statistics = 334, expected statistics = 331.9, P = 0.74) in our family based association test. It is also possible that the SNP 326T/G may play an important role in sporadic cases but less so in the etiology of high-risk hereditary cases, where -79C/T may be more relevant.

Given the biological relevance of cell cycle controls to cancer etiology, as well as the recent findings of association between prostate cancer and cell cycle regulation candidate genes (CHEK2 and CDKN1B), additional studies are warranted. Specifically, follow-up studies should either evaluate the existing results in other populations or expand the analyses to other candidate genes in the same pathway. For CDKN1B, additional studies that evaluate interaction effects with PTEN among human populations will be interesting because these may build upon the biological mechanism already described in prostate cancer mouse models.


   ACKNOWLEDGMENTS
 
We thank all of the study subjects who participated in this study.


   FOOTNOTES
 
Grant support: USPHS SPORE CA58236 (W. B. Isaacs) and a grant from the Department of Defense (J. Xu).

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.

Requests for reprints: Jianfeng Xu, Center for Human Genomics, Wake Forest University School of Medicine, Medical Center Boulevard, Winston-Salem, NC 27157. Phone: (336) 716-5700; Fax: (336) 716-7575; E-mail: jxu{at}wfubmc.edu

Received 7/29/03. Revised 12/18/03. Accepted 1/20/04.


   REFERENCES
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 

  1. Sherr CJ, Roberts JM. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev, 13: 1501-12, 1999.[Free Full Text]
  2. Kibel AS, Schutte M, Kern SE, Isaacs WB, Bova GS. Identification of 12p as a region of frequent deletion in advanced prostate cancer. Cancer Res, 58: 5652-5, 1998.[Abstract/Free Full Text]
  3. Fero ML, Randel E, Gurley KE, Roberts JM, Kemp CJ. The murine gene p27Kip1 is haplo-insufficient for tumour suppression. Nature (Lond.), 396: 177-80, 1998.[CrossRef][Medline]
  4. Nakayama K, Ishida N, Shirane M. Mice lacking p27Kip1 display increased body size, multiple organ hyperplasia, retinal dysplasia, and pituitary tumors. Cell, 85: 707-20, 1996.[CrossRef][Medline]
  5. Kiyokawa H, Kineman RD, Manova-Todorova KO. Enhanced growth of mice lacking the cyclin-dependent kinase inhibitor function of p27Kip1. Cell, 85: 721-32, 1996.[CrossRef][Medline]
  6. Kibel AS, Suarez BK, Belani J, Oh J, Webster R, Brophy-Ebbers M, Guo C, Catalona WJ, Picus J, Goodfellow PJ. CDKN, 1A, and CDKN1B polymorphisms and risk of advanced prostate carcinoma. Cancer Res, 63: 2033-6, 2003.[Abstract/Free Full Text]
  7. Suarez BK, Lin J, Burmester JK, Broman KW, Weber JL, Banerjee TK, Goddard KA, Witte JS, Elston RC, Catalona WJ. A genome screen of multiplex sibships with prostate cancer. Am J Hum Genet, 66: 933-44, 2000.[CrossRef][Medline]
  8. Hsieh CL, Oakley-Girvan I, Balise RR, Halpern J, Gallagher RP, Wu AH, Kolonel LN, O’Brien LE, Lin IG, Van Den Berg DJ, The CZ, West DW, Whittemore AS. A genome screen of families with multiple cases of prostate cancer: evidence of genetic heterogeneity. Am J Hum Genet, 69: 148-58, 2001.[CrossRef][Medline]
  9. Di Cristofano A, De Acetis M, Koff A, Cordon-Cardo C, Pandolfi PP. Pten and p27KIP1 cooperate in prostate cancer tumor suppression in the mouse. Nat Genet, 27: 222-4, 2001.[CrossRef][Medline]
  10. Weir BS. . Genetic data analysis II: methods for discrete population genetic data, Sinauer Association, Inc. Publishers Sunderland, MA 1996.
  11. Laird NM, Horvath S, Xu X. Implementing a unified approach to family based test of association. Genet Epidemiol, 19: 36-42, 2000.[CrossRef]
  12. Tomoda K, Kubota Y, Kato J. Degradation of the cyclin-dependent-kinase inhibitor p27Kip1 is instigated by Jab1. Nature (Lond.), 398: 160-5, 1999.[CrossRef][Medline]
  13. Lake S, Blacker D, Laird N. Family based tests in the presence of association. Am J Hum Genet, 67: 1515-25, 2001.
  14. Nakatsuka S, Liu A, Yao M, Takakuwa T, Tomita Y, Hoshida Y, Nishiu M, Aozasa K. Methylation of promoter region in p27 gene plays a role in the development of lymphoid malignancies. Int J Oncol, 22: 561-8, 2003.[Medline]
  15. Kibel AS, Christopher M, Faith DA, Bova GS, Goodfellow PJ, Isaacs WB. Methylation and mutational analysis of p27(kip1) in prostate carcinoma. Prostate, 48: 248-53, 2001.[CrossRef][Medline]
  16. Konishi N, Nakamura M, Kishi M, Nishimine M, Ishida E, Shimada K. DNA hypermethylation status of multiple genes in prostate adenocarcinomas. Jpn J Cancer Res, 93: 767-73, 2002.[CrossRef][Medline]
  17. Yang G, Ayala G, Marzo AD, Tian W, Frolov A, Wheeler TM, Thompson TC, Harper JW. Elevated Skp2 protein expression in human prostate cancer: association with loss of the cyclin-dependent kinase inhibitor p27 and PTEN and with reduced recurrence-free survival. Clin Cancer Res, 8: 3419-26, 2002.[Abstract/Free Full Text]
  18. Nakamura N, Ramaswamy S, Vazquez F, Signoretti S, Loda M, Sellers WR. Forkhead transcription factors are critical effectors of cell death and cell cycle arrest downstream of PTEN. Mol Cell Biol, 20: 8969-82, 2000.[Abstract/Free Full Text]
  19. Millard SS, Vidal A, Markus M, Koff A. A U-rich element in the 5'-untranslated region is necessary for the translation of p27 mRNA. Mol Cell Biol, 20: 5947-59, 2000.[Abstract/Free Full Text]
  20. Miskimins WK, Wang G, Hawkinson M, Miskimins R. Control of cyclin-dependent kinase inhibitor p27 expression by cap-independent translation. Mol Cell Biol, 21: 4960-7, 2001.[Abstract/Free Full Text]
  21. Kullmann M, Gopfert U, Siewe B, Hengst L. ELAV/Hu proteins inhibit p27 translation via an IRES element in the p27 5'-UTR. Genes Dev, 16: 3087-99, 2002.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
E. L. Goode, B. L. Fridley, R. A. Vierkant, J. M. Cunningham, C. M. Phelan, S. Anderson, D. N. Rider, K. L. White, V. S. Pankratz, H. Song, et al.
Candidate Gene Analysis Using Imputed Genotypes: Cell Cycle Single-Nucleotide Polymorphisms and Ovarian Cancer Risk
Cancer Epidemiol. Biomarkers Prev., March 1, 2009; 18(3): 935 - 944.
[Abstract] [Full Text] [PDF]

Home page
 BloodHome page
J. Healy, H. Belanger, P. Beaulieu, M. Lariviere, D. Labuda, and D. Sinnett
Promoter SNPs in G1/S checkpoint regulators and their impact on the susceptibility to childhood leukemia
Blood, January 15, 2007; 109(2): 683 - 692.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Pathol.Home page
C Hughes, A Murphy, C Martin, O Sheils, and J O'Leary
Molecular pathology of prostate cancer
J. Clin. Pathol., July 1, 2005; 58(7): 673 - 684.
[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 Chang, B.-l.
Right arrow Articles by Xu, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chang, B.-l.
Right arrow Articles by Xu, J.

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