Skip Navigation


Carcinogenesis Advance Access originally published online on August 29, 2007
Carcinogenesis 2007 28(10):2160-2165; doi:10.1093/carcin/bgm167
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
28/10/2160    most recent
bgm167v1
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 ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chen, M.
Right arrow Articles by Gu, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chen, M.
Right arrow Articles by Gu, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

High-order interactions among genetic polymorphisms in nucleotide excision repair pathway genes and smoking in modulating bladder cancer risk

Meng Chen, Ashish M. Kamat1, Maosheng Huang, H.B. Grossman1, Colin P. Dinney1, Seth P. Lerner2, Xifeng Wu and Jian Gu*

1 Department of Epidemiology and Urology, The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA
2 Department of Epidemiology and Scott Department of Urology, Baylor College of Medicine, Houston, TX, USA

* To whom correspondence should be addressed at Department of Epidemiology, Unit 1340, The University of Texas M.D. Anderson Cancer Center, 1155 Pressler Boulevard, Houston, TX 77030, USA. Tel: +1 713 792 8016; Fax: +1 713 792 4657;Email: jiangu{at}mdanderson.org


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Funding
 References
 
Polymorphisms in nucleotide excision repair (NER) genes may cause variations in DNA repair capacity and increase susceptibility to bladder cancer through complex gene–gene and gene–smoking interactions. We applied two data mining approaches to explore high-order gene–gene and gene–environment interactions among 13 polymorphisms in nine major NER genes in 696 bladder cancer patients and 629 controls. Individually, only the XPD D312N variant genotypes exhibited a slightly increased risk for bladder cancer. In classification and regression tree analysis, we observed gene–gene interactions among CCNH V270A, ERCC6 M1097V and RAD23B A249V in ever smokers: smokers with the variant alleles at these three loci had an almost 30-fold increased risk of bladder cancer [odds ratio (OR): 29.6, 95% confidence interval (CI): 9.3–93.7]. When evaluating combined effect of above four single nucleotide polymorphisms, we found a significant gene dosage effect for increased bladder cancer risk with increasing numbers of unfavorable genotypes. Compared with individuals with less than 2 unfavorable genotypes, those with 2 unfavorable genotypes and more than 2 unfavorable genotypes exhibited increased bladder cancer risk with ORs of 1.14 (95% CI: 0.87–1.51) and 2.15 (95% CI: 1.56–2.97), respectively (P < 0.001). The risks were more evident in ever smokers with ORs of 1.43 (95% CI: 1.02–2.01) and 3.40 (95% CI: 2.24–5.15), respectively (P < 0.001). In multifactor dimensionality reduction (MDR) analysis, the five-factor model including smoking, CCNH V270A, ERCC6 M1097V, RAD23B A249V and XPD D312N had the best ability to predict bladder cancer risk. The contributions of these polymorphisms may jointly affect bladder cancer risk through gene–gene and gene–smoking interactions.

Abbreviations: CART, classification and regression tree; CI, confidence interval; CVC, cross-validation consistency; NER, nucleotide excision repair; MDR, multifactor dimensionality reduction; OR, odds ratio; SNPs, single nucleotide polymorphisms


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Funding
 References
 
As a complex multifactorial disease, the etiology of bladder cancer is still not clearly understood. Cigarette smoking is the major risk factor for bladder cancer, accounting for about half of the cases in men and one third of the cases in women. Cigarette smoking contains a variety of chemical carcinogens, including polycyclic aromatic hydrocarbons, aromatic amines and N-nitrogen compounds which may cause DNA damage by forming DNA adducts. The formation of DNA adducts is associated with increased cancer risk confirmed by numerous basic and epidemiological studies. As a defense mechanism of cell, nucleotide excision repair (NER) is the key player in removing bulky DNA adducts and maintaining genome stability (1). There have been compelling evidence that deficient DNA repair capacity is a cancer predisposing factor (2).

The NER pathway has been extensively studied and the main component genes involved in human NER have been elucidated (3). NER consists of two distinct pathways, termed global genomic repair and transcription-coupled repair (Figure 1). At least four basic steps are involved in human NER pathway: (i) damage recognition by XPA, RPA and XPC-RAD23B complex in global genomic repair or by CSA and ERCC6 complex in transcription-coupled repair; (ii) unwinding of the DNA helix around the lesion by TFIIH complex that includes two helicases XPB and XPD; (iii) dual incision of damaged oligonucleotides by 5' (XPF-ERCC1 complex) and 3' (XPG) endonucleases and (iv) filling gap by DNA synthesis (4,5). Most of these critical NER genes are highly polymorphic and a number of single nucleotide polymorphisms (SNPs) in the functional regions of these genes have been implicated as responsible in altering the function of the respective genes, contributing to inter-individual variations of DNA repair capacity.


Figure 1
View larger version (26K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Simplified scheme of NER process in mammalian cells showing the involvement of evaluated genes in this study.

 
Numerous studies have investigated SNPs in NER genes and cancer risk. The results have been inconsistent since cancer is a multifactorial disease involving complex interactions between multiple genes and environmental risk factors (6,7). The low-penetrance feature of individual allele coupled with a failure to consider the complexity of gene–gene interactions may be partly responsible for the lack of consistency in cancer association studies of common SNPs (8–11). Recently, a pathway-based multigenic approach to assess the combined effects of a panel of polymorphisms that interact and function in the same pathway has been applied to bladder cancer association studies and revealed complex gene–gene and gene–smoking interactions in modulating bladder cancer risk (12–14). Two data mining approaches, classification and regression tree (CART) analysis and multifactor dimensionality reduction (MDR) analysis, have been applied to explore high-order gene–gene and gene–environment interactions in bladder cancer susceptibility (12–14).

In this large bladder cancer case–control study, we used a variety of analytical approaches to identify high-order interactions among genetic polymorphisms in NER pathway genes and smoking in modulating bladder cancer risk. Thirteen potential functional polymorphisms in nine major NER genes were investigated: ERCC1 3' UTR, XPD K751Q, XPD D312N, XPA 5'UTR, XPF S662P, XPG H1104D, XPC PAT, XPC K939Q, XPC A499V, RAD23B A249V, CCNH V270A, ERCC6 M1097V and ERCC6 R1230P.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Funding
 References
 
Study population
Beginning in July 1999, bladder cancer cases were recruited from the University of Texas M.D. Anderson Cancer Center and Baylor College of Medicine in Houston area through a daily review of computerized appointment schedules. Approximately 70% of the cases are from Texas. We performed an analysis comparing education and income between cases from Texas and cases from other states and did not find significant differences in education (P = 0.84) and in household income (P = 0.62). All 696 cases were newly diagnosed within one year of recruitment with histological confirmation of bladder cancer (urothelial carcinoma). Among them, 388 (55.75%) cases had superficial bladder cancer and 296 (42.53%) had invasive disease. Twelve cases had unconfirmed cancer type. Patients with a medical history of prior systemic chemotherapy or radiotherapy and recurrent bladder cancer were excluded from this study. There were no age, gender and cancer stage restrictions. The control subjects are healthy individuals with no prior cancer history (except non-melanoma skin cancer). Six hundred and twenty-nine controls were recruited from Kelsey–Seybold, the largest multi-specialty physician group in Houston metropolitan area. Controls were frequency matched to cases on the basis of age (±5 years), gender and ethnicity. The potential control subjects were first surveyed by using a short questionnaire to elicit willingness to participate in the study and to provide preliminary demographic data for matching. A Kelsey–Seybold staff member provided the questionnaire to each potential control subject during clinical registration. The potential control subjects were contacted by telephone at a later date to confirm their willingness to participate and to schedule an interview appointment at a Kelsey–Seybold clinic convenient to the participant. Informed consent was obtained from all study participants before the collection of epidemiological data and blood samples by trained M.D. Anderson staff interviewers. The response rates were 75% for the controls and 92% for the cases. After signing the informed consents, the subjects completed a standard questionnaire through personal interview. The information includes demographics, smoking status, alcohol consumption and medical history. We defined an individual who has never smoked or has less than 100 cigarettes in their lifetime as a non-smoker. Smokers include former smokers and current smokers. A former smoker was a person who had quit smoking at least 1 year prior to diagnosis (cases) or interview (controls). A current smoker was someone who was currently smoking or who had stopped <1 year. At the end of the interview, 40-ml blood sample was obtained from each subject. This procedure was approved by the Institutional Review Boards of M.D. Anderson Cancer Center, Baylor College of Medicine and Kelsey–Seybold clinic. All of the analysis in this study is restricted to Caucasian due to small numbers of minority populations.

Genotyping
Genomic DNA was isolated from peripheral blood lymphocytes using the QIAamp DNA Blood Maxi Kit (Qiagen, Valencia, CA). The specific polymorphic loci ERCC1 3'UTR (rs3212986), XPD 751 (rs13181), XPF 662 (rs2020955), XPG 1104 (rs17655), XPC 939 (rs2228001), XPC 499 (rs2228000), RAD23B 249 (rs1805329), CCNH 270 (rs2266690), ERCC6 1097 (rs2228526) and ERCC6 1230 (rs4253211) were genotyped by the Taqman real-time polymerase chain reaction method using 7900 HT sequence detector system (Applied Biosystems, Foster City, CA). The primer and probe on request are either obtained from database (http://snp500cancer.nci.nih.gov) or designed by Primer Expression Software. In probes, fluorescent dye FAM and VIC were labeled on the 5' end and a quencher was labeled on the 3' end. Typical amplification mixes (5 µl) consist of DNA sample (5 ng), 1x TaqMan Buffer A, 200 µm dNTPs, 5 mmol MgCl2, 0.65 units of AmpliTaq Gold, 900 nmol/l primer each and 200 nmol/L probe each. The thermal cycling conditions included 1 cycle for 10 min at 95°C, and 40 cycles for 15 s at 95°C and 1 min at 60°C. The end point fluorescence was analysed by SDS version 2.1 software (Applied Biosystems). Water control, internal controls and previously genotyped samples were added into each plate as the calibrator to ensure the accuracy and consistency of the genotyping. The polymorphic loci XPA 5'UTR (rs1800975), XPC PAT and XPD 312 (rs1799793) were examined by polymerase chain reaction–RFLP (15,16).

Statistical analysis
Using the Intercooled Stata 8.0 statistical software package (Stata Co., College Station, TX), the Pearson's {chi}2 test for categorical variables (sex and smoking status) and the Wilcoxon rank sum test for continuous variables (age and cigarette smoking pack years) were used to test the differences in distribution between cases and controls. To control for confounding factors, unconditional multivariate logistic regression analysis was conducted and adjusted odds ratios (ORs) along with 95% confidence intervals (CIs) were calculated. The combined effect of four minor alleles CCNH 270, ERCC6 1097, RAD23B 249 and XPD 312 were also analysed as a categorical variable by grouping the subjects according to the number of unfavorable genotypes. All statistical analyses were two sided.

CART approach.
For higher order gene–gene interactions, CART analysis was performed using the HelixTree Genetics Analysis Software (version 4.1.0, Golden Helix). CART is a binary recursive-partitioning method that produces a decision tree to identify subgroups of subjects at higher risk. Specifically, the recursive-partitioning algorithm in HelixTree starts at the first node (with the entire data set) and uses a statistical hypothesis-testing method—formal inference-based recursive modeling—to determine the first locally optimal split and each subsequent split of the data set, with multiplicity-adjusted P values to control tree growth (P < 0.05). The same process continues in the subsequent split until the terminal nodes that further split into subsequent statistically significant nodes or reach a pre-specified minimum size of subject (less than 10) in analysis.

MDR approach.
The non-parametric MDR approach was selected to complement logistic regression for the analysis of gene–gene and gene–environment interactions. The MDR method was first described by Moore et al. (13). The goal of MDR is to find the main factor and the combinations of multiple factors such as genotypes and discrete environmental factors that are more frequently associated with cases than with controls. To search for the best n-factor (N = 1, 2, 3, 4, 5) model, all the data were divided into a training set including 9/10 of the data and a testing set including 1/10 of the data. Training set is used to search for the best model associated with the high-risk pattern depending on the ratio of cases to controls and the testing set is used to control the goodness-of-fit of the model. As a result, MDR approach will give two scores for each n-factor best model, a mean prediction error percentage and the cross-validation consistency (CVC) frequency. Mean prediction error percentage is the proportion of subjects for whom an incorrect class prediction was made; thus, the best model will have the lowest mean prediction error. CVC presents the number of times a particular combination of factors is identified in each possible testing set; therefore, the best model will have the maximum CVC frequency. This complete test will repeat 10 times using different random seeds to reduce the probability of biased results due to the chance division of the data in training and testing sets. To evaluate the magnitude of the prediction error and the CVC frequency, we permutated the status of cases and controls in the data set and repeated the analysis 1000 times and obtained the prediction error and the CVC of each n-factor models under the null hypothesis of no association. By comparing those data, P values associated with each prediction error and CVC were obtained.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Funding
 References
 
Characteristics of subjects
A total of 696 confirmed Caucasian bladder cancer cases and 629 healthy Caucasian controls were analysed in this study. Men were overrepresented among the cases compared with the controls (78.45 versus 72.66%, P = 0.014). The cases were slightly older than the controls (mean age: 63.9 versus 62.8 years, P = 0.06). The cases had more ever smokers than the controls (73.6 versus 53.7%, P < 0.001) and the ever smokers in cases had a heavier smoking history than those in controls (mean pack years: 43.0 versus 28.3, P < 0.001).

Risk associated with the individual variant genotype
We evaluated the main effect of each individual polymorphism in NER pathway on bladder cancer susceptibility using the unconditional multivariate logistic regression (Table I). Only the XPD D312N variant genotype showed a significant elevated risk with an OR of 1.28 (95% CI: 1.01–1.62). When stratified by smoking status, the risk of XPD D312N was more evident in ever smoker (OR: 1.41, 95% CI: 1.06–1.88). In addition, there were significant associations for the variant alleles of CCNH V270A (OR: 1.61, 95% CI: 1.18-2.18) and ERCC6 M1097V (OR: 1.43, 95% CI: 1.06–1.93) in ever smokers only.


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

  		Table I. Overall main effects of genotypes on bladder cancer risk and effects stratified by smoking status

 
CART analysis
To explore high-order interactions, we included these polymorphisms and the smoking status variables in the CART analysis using the binary partitioning method (Figure 2). The first split on the decision tree is smoking status indicating smoking is the main risk factor for bladder cancer. Also, considering that the main function of NER pathway is more important in ever smokers to repair the DNA damage caused by smoking, we observed gene–gene interactions among CCNH V270A, XPD D312N, ERCC6 M1097V and RAD23B A249V in ever smokers. The smokers with the variant alleles at these three loci (CCNH V270A, ERCC6 M1097V and RAD23B A249V) (terminal node 6) had an almost 30-fold increased risk of bladder cancer (OR: 29.6, 95% CI: 9.3–93.7) compared with the group with the lowest risk (reference group, terminal node 1). The reference group (Node 1) was smokers with the genetic polymorphisms of CCNH V270A (wt), XPD D312N (wt) and RAD23B A249V variant genotypes. The ORs were calculated by comparing each of the other groups to the reference group. It is intriguing that the wild-type RAD23B A249 genotype appears to confer an increased risk when in combination with smoking and CCNH V270A (wt), XPD D312N (wt) (Node 2 versus Node 1). The functional significance of this SNP has not been studied yet. It is possible that the variant may be associated with a better DNA repair in the context of smoking, CCNH and XPD wild-type genotype. It is also likely that this result was chance finding due to the data-driven nature of CART analysis and the small sample size of the reference group.


Figure 2
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. CART analysis in the NER pathway. *W: the homozygous wild-type genotype; V: variant genotypes including the homozygous wild-type genotype and heterozygous variant genotype. § Genotypes of the SNPs. OR and 95% CI was adjusted by gender, age and pack years.

 
Risk associated with combined effects of unfavorable genotypes
We evaluated combined effect of four unfavorable genotypes in bladder cancer risk. Those unfavorable genotypes are variant genotypes of XPD D312N, ERCC6 M1097V, CCNH V270A and RAD23B A249V, all of which were predicted to attribute to higher bladder cancer risk in CART analysis (Figure 2). We found a significant gene dosage effect for increasing numbers of unfavorable genotypes (Table II). Compared with individuals with less than 2 unfavorable genotypes, those with 2 unfavorable genotypes and more than 2 unfavorable genotypes had increased bladder cancer risk with ORs of 1.14 (95% CI: 0.87–1.51) and 2.15 (95% CI: 1.56–2.97), respectively (P for trend <0.001). The risks were more evident in ever smokers with ORs of 1.43 (95% CI: 1.02–2.01) and 3.40 (95% CI: 2.24–5.15), respectively (P for trend <0.001).


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

  		Table II. Joint effects of unfavorable genotypes in NER pathway

 
MDR analysis
We then used another data mining tool, the MDR analysis, to identify which combination of factors is the best model to predict the high-risk subgroup of subjects. Table III summarizes the CVC and prediction error obtained from MDR for all number of factors evaluated (N = 1, 2, 3, 4, 5, 6). Smoking status was the best one-factor model with the highest CVC (100%), and the lowest prediction error (39.4%) among all 14 factors. The prediction error was statistically significant (P < 0.001). More interestingly, the five-factor model including XPD D312N, ERCC6 M1097V, CCNH V270A, RAD23B A249V and smoking had a minimum prediction error rate of 36% and a maximum CVC of 100/100 (P < 0.001), which showed a better prediction than smoking only.


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

  		Table III. MDR Analysis for the bladder cancer risk predication

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Funding
 References
 
Our results suggest the existence of gene–smoking and gene–gene interactions in NER pathway in modulating bladder cancer risk. Smokers carrying variant alleles XPD D312N, ERCC6 M1097V, CCNH V270A and RAD23B A249V are associated with higher risk of bladder cancer. Individual contributions of any of these genetic variations in NER genes to bladder cancer risk are modest. However, these polymorphisms may jointly affect bladder cancer risk through gene–gene and gene–smoking interactions.

Smoking is well established as the main risk factor to bladder cancer (17,18), which has been validated by the data mining approaches in our study: the first split in CART and the best one-factor model in MDR both picked out smoking as the predominant risk factor for bladder cancer. In our logistic regression model, we found only XPD D312N variant allele was associated with an increased risk for bladder cancer. When we stratified the data by smoking status, we found significant associations of XPD D312N, CCNH V270A and ERCC6 M1097V in ever smokers. This indicates gene–smoking interactions may modify the function of those NER genes. We also conducted a parallel CART analysis among non-smokers. However, we did not detect any high-order interaction among the NER polymorphisms (data not shown). This result is biologically plausible since NER is mainly involved in repairing bulky DNA adducts caused by smoking. In non-smokers, the major DNA damage probably is not bulky DNA adducts, but base damages and DNA breaks, which mainly involves base excision repair and double strand break repair.

XPD gene encodes a 5'–3' DNA helicase with single-strand DNA-dependent ATPase activity. It is a subunit of basal transcription factor (TFIIH) that has dual functions in transcription and DNA damage repair. XPD is the binding site of P53 to TFIIF, thus also plays a role in apoptosis (19,20). For carriers with the XPD D312N variant allele, we found a 1.28-fold increased risk overall and a 1.41-fold increased risk in ever smokers. For the XPD K751Q variant allele, we found a 1.14-fold increased risk overall and a 1.28-fold increased risk in ever smokers. Extensive studies, including those of chromosome aberrations, p53 mutations, changes in DNA repair capacity, formation of DNA adducts and apoptosis, have conducted to explore the functional significance of XPD D312N and K751Q in different cancer types. In breast tissues, XPD D312N and K751Q variant alleles were associated with higher levels of polycyclic aromatic hydrocarbons-DNA adducts (21). Changes in DNA repair capacity were also evaluated by cytogenetic observation. Au et al. (22) showed that the D312N variants were associated with increased UV-induced chromosome aberrations. However, no association was detected between the XPD D312N variants and X-ray-induced chromosome aberrations (23), which suggests that NER pathway plays a role in UV-, but not X-ray, induced DNA damage repair. Furthermore, carriers of the D312N variants were associated with a higher apoptotic rate in human lymphoblastoid cells (24). A significant association of the variant allele of the D312N with lung cancer risk has been observed in meta-analyses of lung cancer (25,26). Similar associations have been observed in skin cancer (23,27), prostate cancer (28) and basal cell carcinoma (29). In a recent large Spanish bladder cancer case–control study, the variant alleles of both SNPs were associated with a non-significant slight increase in bladder cancer risk (30).

The more significant finding of this study is the consistent complex gene–smoking and gene–gene interactions identified through different statistical approaches. In the logistic regression model, a dose–response was found for the increased bladder cancer risk with the increasing number of adverse genotypes of the XPD D312N, CCNH V270A, ERCC6 M1097V and RAD23B A249V. This dose–response is evident only in ever smokers. In CART analysis, subjects who are smokers and bearing the variant genotypes of the CCNH V270A, ERCC6 M1097V and RAD23B A249V had the highest bladder cancer risk. In MDR analysis, smokers with the variant genotypes of the above four SNPs are the subgroup at the highest risk. Therefore, we proposed that the five-factor model including smoking, the variant genotypes of the XPD D312N, CCNH V270A, ERCC6 M1097V and RAD23B A249V as the best model to predict bladder cancer risk from this study. The XPD D312N was not included in the highest risk subgroup in CART analysis. This may be due to the limitation of sample size and statistical power, which did not allow the further split of terminal node 6 (Figure 2). In the five-factor model in MDR analysis, the combination of XPD D312N, CCNH V270A, ERCC6 M1097V and RAD23B A249V in ever smokers is associated with the highest bladder cancer risk, although individually these variants are only associated with modest increased bladder cancer risk in ever smokers. It indicates a non-additive epistasis effect of these variants in modifying bladder cancer risk, which have been suggested previously in breast cancer (31), prostate cancer (32) and bladder cancer studies (13,30). The mechanisms underlying these high-order interactions among genetic polymorphisms in NER pathway genes and smoking in modulating bladder cancer risk remain to be elucidated. It is interesting to note that the RAD23B and ERCC6 are critical component in damage recognition, and CCNH and XPD are part of RNA polymerase complex that are involved in both damage recognition and unwinding step (Figure 1). It is tempting to speculate that the variant genotypes of these four genes may corporate to hinder damage recognition of smoking-induced DNA bulky adducts that leads to more DNA damage and higher bladder cancer risk.

We also found a dose–response between the number of the adverse genotypes of NER pathway and the bladder cancer risk. Moreover, the dose–response between the adverse genotypes and levels of DNA adducts in peripheral blood cells have been reported by Matullo (33). We have recently shown that higher risk subgroup identified by CART analysis exhibited higher chromosome breaks and DNA damage than lower risk subgroups (12). Thus, the high-order interactions among genetic polymorphisms in NER pathway genes and smoking may affect the DNA damage repair capacity and contributes to increased bladder cancer risk in high-risk subgroups.

There are some limitations for this study. First, this is a hospital-based case–control study and there may be selection bias. For example, the cases and controls were not geographically matched, which may introduce selection bias. However, the overall goal of this study is largely driven by genetic hypothesis versus an environmental hypothesis. Hence, a control population with representative environmental exposures to the cases is less of a concern. M.D. Anderson is a referral center and there is an overrepresentation of advanced stage cases in this study. But there is no evidence to suggest that superficial and invasive bladder cancer have different etiologies. Our previous extensive publications have shown that the frequencies of most SNPs and exposures in our population are comparable with other US and European bladder cancer case–control studies. Second, given the potential bias in our case–control design, the post hoc data-driven nature of the CART and MDR approach and the small sample sizes in some of the terminal nodes, the results should be interpreted with caution. Further validations from independent populations are necessary to confirm these results.

In conclusion, this large bladder cancer case–control study examined the complex gene–gene and gene–smoking interactions in NER genes in modulating bladder cancer risk. Our results provide further support for using a multi-genetic approach and taking consideration of complex high-order interactions in cancer association studies. Applying multiple data mining tools, such as CART and MDR, may facilitate the identification of gene–environment and gene–gene interactions.


   Funding
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Funding
References
 
National Cancer Institute (CA 74880, CA 91846).


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Funding
 References
 

  1. Hoeijmakers JH. Genome maintenance mechanisms for preventing cancer. Nature (2001) 411:366–374.[CrossRef][Medline]
  2. Berwick M, et al. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review. J. Natl Cancer Inst. (2000) 92:874–897.[Abstract/Free Full Text]
  3. Wood RD, et al. Human DNA repair genes. Science (2001) 291:1284–1289.[Abstract/Free Full Text]
  4. Christmann M, et al. Mechanisms of human DNA repair: an update. Toxicology (2003) 193:3–34.[CrossRef][ISI][Medline]
  5. Volker M, et al. Sequential assembly of the nucleotide excision repair factors in vivo. Mol. Cell (2001) 8:213–224.[CrossRef][ISI][Medline]
  6. Goode EL, et al. Polymorphisms in DNA repair genes and associations with cancer risk. Cancer Epidemiol Biomarkers Prev. (2002) 11:1513–1530.[Abstract/Free Full Text]
  7. Wu X, et al. Genetic polymorphism in bladder cancer. Front. Biosci. (2007) 12:192–213.[CrossRef][ISI][Medline]
  8. Pharoah PD, et al. Association studies for finding cancer-susceptibility genetic variants. Nat. Rev. Cancer (2004) 4:850–860.[CrossRef][ISI][Medline]
  9. Wu X, et al. Genetic susceptibility to tobacco-related cancer. Oncogene (2004) 23:6500–6523.[CrossRef][ISI][Medline]
  10. Erichsen HC, et al. SNPs in cancer research and treatment. Br. J. Cancer (2004) 90:747–751.[CrossRef][ISI][Medline]
  11. Friedberg EC. How nucleotide excision repair protects against cancer. Nat. Rev. Cancer (2001) 1:22–33.[CrossRef][Medline]
  12. Wu X, et al. Bladder cancer predisposition: a multigenic approach to DNA-repair and cell-cycle-control genes. Am. J. Hum. Genet. (2006) 78:464–479.[CrossRef][ISI][Medline]
  13. Andrew AS, et al. Concordance of multiple analytical approaches demonstrates a complex relationship between DNA repair gene SNPs, smoking and bladder cancer susceptibility. Carcinogenesis (2006) 27:1030–1037.[Abstract/Free Full Text]
  14. Huang M, et al. High-order interactions among genetic variants in DNA base excision repair pathway genes and smoking in bladder cancer susceptibility. Cancer Epidemiol. Biomarkers Prev. (2007) 16:84–91.[Abstract/Free Full Text]
  15. Wu X, et al. XPA polymorphism associated with reduced lung cancer risk and a modulating effect on nucleotide excision repair capacity. Carcinogenesis (2003) 24:505–509.[Abstract/Free Full Text]
  16. Marin MS, et al. Poly (AT) polymorphism in intron 11 of the XPC DNA repair gene enhances the risk of lung cancer. Cancer Epidemiol. Biomarkers Prev. (2004) 13:1788–1793.[Abstract/Free Full Text]
  17. Pashos CL, et al. Bladder cancer: epidemiology, diagnosis, and management. Cancer Pract. (2002) 10:311–322.[CrossRef][ISI][Medline]
  18. Ross RK, et al. Bladder cancer epidemiology and pathogenesis. Semin. Oncol. (1996) 23:536–545.[ISI][Medline]
  19. Hoeijmakers JH, et al. TFIIH: a key component in multiple DNA transactions. Curr. Opin. Genet. Dev. (1996) 6:26–33.[CrossRef][ISI][Medline]
  20. Drapkin R, et al. Dual role of TFIIH in DNA excision repair and in transcription by RNA polymerase II. Nature (1994) 368:769–772.[CrossRef][Medline]
  21. Tang D, et al. Polymorphisms in the DNA repair enzyme XPD are associated with increased levels of PAH-DNA adducts in a case-control study of breast cancer. Breast Cancer Res. Treat. (2002) 75:159–166.[CrossRef][ISI][Medline]
  22. Au WW, et al. Functional characterization of polymorphisms in DNA repair genes using cytogenetic challenge assays. Environ. Health Perspect. (2003) 111:1843–1850.[ISI][Medline]
  23. Lunn RM, et al. XPD polymorphisms: effects on DNA repair proficiency. Carcinogenesis (2000) 21:551–555.[Abstract/Free Full Text]
  24. Seker H, et al. Functional significance of XPD polymorphic variants: attenuated apoptosis in human lymphoblastoid cells with the XPD 312 Asp/Asp genotype. Cancer Res. (2001) 61:7430–7434.[Abstract/Free Full Text]
  25. Benhamou S, et al. ERCC2/XPD gene polymorphisms and lung cancer: a HuGE review. Am. J. Epidemiol. (2005) 161:1–14.[Abstract/Free Full Text]
  26. Hu Z, et al. DNA repair gene XPD polymorphism and lung cancer risk: a meta-analysis. Lung Cancer (2004) 46:1–10.[CrossRef][ISI][Medline]
  27. Tomescu D, et al. Nucleotide excision repair gene XPD polymorphisms and genetic predisposition to melanoma. Carcinogenesis (2001) 22:403–408.[Abstract/Free Full Text]
  28. Rybicki BA, et al. DNA repair gene XRCC1 and XPD polymorphisms and risk of prostate cancer. Cancer Epidemiol. Biomarkers Prev. (2004) 13:23–29.[Abstract/Free Full Text]
  29. Vogel U, et al. Polymorphisms of the DNA repair gene XPD: correlations with risk of basal cell carcinoma revisited. Carcinogenesis (2001) 22:899–904.[Abstract/Free Full Text]
  30. Garcia-Closas M, et al. Genetic variation in the nucleotide excision repair pathway and bladder cancer risk. Cancer Epidemiol. Biomarkers Prev. (2006) 15:536–542.[Abstract/Free Full Text]
  31. Ritchie MD, et al. Multifactor-dimensionality reduction reveals high-order interactions among estrogen-metabolism genes in sporadic breast cancer. Am. J. Hum. Genet. (2001) 69:138–147.[CrossRef][ISI][Medline]
  32. Xu J, et al. The interaction of four genes in the inflammation pathway significantly predicts prostate cancer risk. Cancer Epidemiol. Biomarkers Prev. (2005) 14:2563–2568.[Abstract/Free Full Text]
  33. Matullo G, et al. DNA repair gene polymorphisms, bulky DNA adducts in white blood cells and bladder cancer in a case-control study. Int. J. Cancer (2001) 92:562–567.[CrossRef][ISI][Medline]
Received April 2, 2007; revised July 2, 2007; accepted July 4, 2007.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 CarcinogenesisHome page
R. L. Milne, R. Fagerholm, H. Nevanlinna, and J. Benitez
The importance of replication in gene-gene interaction studies: multifactor dimensionality reduction applied to a two-stage breast cancer case-control study
Carcinogenesis, June 1, 2008; 29(6): 1215 - 1218.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
28/10/2160    most recent
bgm167v1
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 ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Chen, M.
Right arrow Articles by Gu, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chen, M.
Right arrow Articles by Gu, J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?