Figures
Abstract
Background
In order to elucidate a combination of genetic alterations that drive tobacco carcinogenesis we have explored a unique model system and analytical method for an unbiased qualitative and quantitative assessment of gene-gene and gene-environment interactions. The objective of this case control study was to assess genetic predisposition in a biologically enriched clinical model system of tobacco related cancers (TRC), occurring as Multiple Primary Neoplasms (MPN).
Methods
Genotyping of 21 candidate Single Nucleotide Polymorphisms (SNP) from major metabolic pathways was performed in a cohort of 151 MPN cases and 210 cancer-free controls. Statistical analysis using logistic regression and Multifactor Dimensionality Reduction (MDR) analysis was performed for studying higher order interactions among various SNPs and tobacco habit.
Results
Increased risk association was observed for patients with at least one TRC in the upper aero digestive tract (UADT) for variations in SULT1A1 Arg213His, mEH Tyr113His, hOGG1 Ser326Cys, XRCC1 Arg280His and BRCA2 Asn372His. Gene - environment interactions were assessed using MDR analysis. The overall best model by MDR was tobacco habit/p53(Arg/Arg)/XRCC1(Arg399His)/mEH(Tyr113His) that had highest Cross Validation Consistency (8.3) and test accuracy (0.69). This model also showed significant association using logistic regression analysis.
Citation: Kotnis A, Namkung J, Kannan S, Jayakrupakar N, Park T, Sarin R, et al. (2012) Multiple Pathway-Based Genetic Variations Associated with Tobacco Related Multiple Primary Neoplasms. PLoS ONE 7(1): e30013. https://doi.org/10.1371/journal.pone.0030013
Editor: Suzannah Rutherford, Fred Hutchinson Cancer Research Center, United States of America
Received: June 4, 2011; Accepted: December 11, 2011; Published: January 11, 2012
Copyright: © 2012 Kotnis et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was partially supported by the Department of Biotechnology (DBT), Government of India (No. BT/PR5588/Med/12/231/2004). Fellowship to AK from Council for Scientific and Industrial Research (CSIR), Government of India is acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Tobacco related cancers (TRC) which include carcinoma of lung, esophagus, head-neck, cervix, bladder, stomach, kidney, pancreas, liver and myeloid leukemia account for almost half the global burden of cancer [1], [2]. Tobacco contains a variety of chemical carcinogens which are activated for detoxification by xenobiotic metabolism enzymes (XME). Activated carcinogens can cause DNA damage by forming harmful DNA adducts. The damaged DNA is repaired by elaborate DNA repair machinery. Cells with extensive DNA damage usually undergo apoptosis. Compromise in any of these cellular pathways promotes survival and growth of mutated cells leading to oncogenesis [3].
Genetic susceptibility could be an important determinant in TRC etiology as suggested by familial occurrence of TRC [4]. Identification and characterization of susceptibility factors in common multifactor disorders such as TRC is challenging. This is due to stringent requirement of appropriate samples to analyze the complex gene-environment interactions involved, limitations of conventional statistical methods to reliably determine gene-gene and gene-environment interaction and tools to correlate the genotype to phenotype. Genes important in carcinogenesis are highly polymorphic and contribute to cancer susceptibility. There are numerous large studies associating single nucleotide polymorphisms (SNPs) in a single gene or multiple genes in a single pathway. Often such variants have limited use in assessment of disease risk, since most of the variants have low penetrance and confer a relatively small risk. Multipathway based association study should help to identify a cumulative effect of low penetrance alleles in predisposition [5]. Consensus has failed to emerge regarding the combination of genetic alterations that drives tobacco carcinogenesis. There is no report yet on a multipathway based approach to study genetic susceptibility to cancer in tobacco associated multiple primary neoplasms (MPN).
In the present case-control study, we have investigated the hypothesis that cumulative effect of low penetrance alleles predispose to tobacco induced MPN (Fig. 1). It is believed that patients with MPN provide a genetically enriched resource to study predisposition to cancer [6]. Association of 21 candidate SNPs in 18 genes from pathways of xenobiotic metabolism, DNA repair, cell cycle regulation and apoptosis implicated in tobacco carcinogenesis was studied in a unique group of individuals with tobacco related MPN. The risk association was analyzed using logistic regression and higher order genetic interactions were studied using multifactor dimensionality reduction (MDR) analysis.
Results
In the present study 151 cases with MPN and 210 cancer-free controls were analyzed. Using IARC definition of Tobacco Related Cancer (TRC) these 151 MPN patients were further sub classified as patients with at least one TRC in the UADT (n = 113), none of the TRC in the UADT (n = 17) and those with no TRC (n = 21; Table 1). In all the categories, majority of the cancers were metachronous, that is, the second primary cancer was diagnosed 6 months or more after diagnosis of first cancer.
Majority of the patients had both cancers in the UADT region. Patients and controls were mainly from North India, with tobacco habit. Among the tobacco users, 48.6% cases and 61.4% controls used smokeless tobacco which was in the form of application of roasted tobacco (masheri) over gums or chewing a mix of tobacco with one or more ingredients like lime, betel nut or betel leaf. The quanta of tobacco consumed were not available for many subjects as it was self reported information.
The genotype distribution of SNPs was compared in cases using dominant model (homozygous wild type versus homozygous variant+heterozygous) and extreme model (homozygous wild type versus homozygous variant). Selection of these models was done on the biological plausibility that homozygous and heterozygous variant conferred risk compared to wild type.
Amongst the 21 SNPs selected from genes in different pathways, univariate analysis showed risk association in a few genes as shown in Table 2 and Table S1. Crude OR and OR adjusted to age and gender were considered for statistically significant association. Increased risk association was observed for patients with at least one TRC in the UADT category for xenobiotic metabolizing genes, SULT1A1 Arg213His (Extreme model, OR = 6.6, 95% CI 1.47–29.34; Dominant model, OR = 1.74, 95% CI 1.07–2.82), meH Tyr113His (Extreme model, OR = 2.55, 95% CI 1.19–5.47), DNA repair genes, hOGG1 Ser326Cys (Dominant model, OR = 1.81, 95% CI 1.12–2.90; Heterozygous OR = 1.91, 95% CI 1.16–3.15), XRCC1 Arg280His (Dominant model, OR = 1.80, 95% CI 1.06–3.08), BRCA2 (Heterozygous OR = 1.93, 95% CI 1.11–3.35). In the TRC outside group significant risk association was observed for xenobiotic metabolizing genes MPO G −463 A (Extreme model, OR = 8.06, 95% CI 2.16–30.72; Dominant model, OR = 3.66, 95% CI 1.37–9.79), SULT1A1 Arg213His (Extreme model, OR = 6.92, 95% CI 2.27–20.96; Heterozygous, OR = 7.15, 95% CI 2.34–21.67), protective association was observed for the cell cycle regulating gene Cyclin D1 A870G (Extreme model OR = 0.27, 95%CI 0.09–0.723; Heterozygous, OR = 0.26, 95%CI 0.09–0.82).
We have conducted MDR analysis accounting for missing values [7] and the one-way to five-way interaction models were considered. As shown in Table 3, the overall best model across one to five-way interaction models was habit/p53(Arg72Arg)/XRCC1(Arg399His)/meH(Tyr113His) that had maximum CVC (Cross validation consistency) and maximum test accuracy (CVC = 8.3; test accuracy = 0.69). To obtain effect size of individual genotype combination, OR MDR analysis [8] was conducted for this four-way interaction model as shown in Table S2. Because the number of individuals for each combination of genotype and habit was relatively small, only one variable combination Habit = 1, p53 (Arg72Arg) = 0, XRCC1 (Arg399His) = 1, and meH (Tyr113His) = 1 had odds ratio with significant confidence intervals (OR = 3.217; 95% CI 1.201–10.177).
All the 21 SNPs were analyzed for HWE of which 12 SNPs were in HWE for the control group where as 9 SNP were not in HWE. However, all the 3 SNPs (p53 (Arg72Arg)/XRCC1 (Arg399His)/mEH (Tyr113His) which showed significant association together with tobacco habit in the MDR analysis were in HWE. SNPs with significant association also showed HWE in controls group. We perfomed linkage disequilibrium analysis for variants in XRCC1, NAT2 and BRCA2 and observed significant association between NAT2 Ile114Thr and NAT2 Arg197Gln, NAT2 Ile114Thr and NAT2 Gly286Glu, BRCA2 Asp991Asn and BRCA2 Asn372His as shown in Table S3.
Discussion
In the present case-control study we have examined a set of biologically plausible SNPs implicated in tobacco carcinogenesis (Fig. 1). Risk association of these SNPs for tobacco related cancers has been investigated by using conventional statistics and through MDR analysis. Though not a consistent finding, each of the SNP identified by us, and a large number of other SNPs have been shown to be associated with tobacco related cancers in previous case-control studies or their meta-analyses [1], [9], [10], [11], [12], [13]. The evidence for cumulative effect of various genetic alterations on metabolic and cellular pathways involved in tobacco carcinogenesis although compelling [5], is based on piecemeal evidence from heterogeneous studies of single or few related SNPs, in a background of large number of genetic and environmental risk modifiers. Only few genome-wide association studies (GWAS) have been conducted for tobacco related cancers so far and these are yet to provide major leads in tobacco carcinogenesis [14].
It is emerging that for an unbiased qualitative and quantitative assessment of gene-gene and gene-environment interactions, clinically relevant insight in tobacco carcinogenesis may not come from additional studies confirming or refuting risk association of known SNPs, but through exploration of alternative research strategies, model systems and analytical methods. Towards this goal we have incorporated three distinct elements in our study. Firstly, we have adopted a biologically holistic approach of examining SNPs in the key genes of major pathways in tobacco carcinogenesis. We have chosen a biologically enriched clinical model system of tobacco related multiple primary neoplasms (MPN-TRC). We had earlier hypothesized [5], [10], [15], [16] that individuals who develop tobacco related MPN, represent a cohort of individuals with enhanced gene-environment and gene-gene interaction. Only recently, the unique biological and statistical utility of MPN in molecular epidemiological studies has been highlighted by others [6]. These authors provide empirical evidence that for MPN of the same organ, the relative risk is approximately the square of the relative risk as found in the traditional case-control studies using single primary cancers. In our study, 75% of the TRC MPN were within the UADT, which is one continuous epithelial lining exposed to tobacco carcinogens.
The third aspect of our study is the statistical approaches used. Studying higher order gene interactions using logistic regression is laborious and has low statistical power due to very high degrees of freedom. Hence we used multifactor dimensionality reduction (MDR) as a complementary statistical approach for studying higher order interactions among various SNPs analyzed and tobacco habit. This combination of testing multiple SNPs using MDR showed 4 factor model of tobacco habit, p53 (Arg72Arg), XRCC1 (Arg399His) and mEH (Tyr113His) with an OR of 3.217 (95% CI: 1.2–10.18) and cross validation consistency of 8.3 as strongest risk predictor to MPN. Cross validation consistency refers to the number of times a particular interaction model is selected across 10 cross-validation datasets and the best test accuracy of 0.69. We have taken several approaches to control for false positive findings which may emerge due to multiple testing. We further tested the 2–4 MDR genetic models using logistic regression and found significant association of the models with tobacco related MPN. The SNPs with significant association in univariate analysis also showed significant interaction in MDR. The advantage of this observation is that MDR makes no assumption about the data distribution and does not require correction of multiple testing, which is helpful for studies with small sample size. The controls were matched for ethnicity and tobacco habit reducing the confounding risk due to ethnicity.
It is biologically plausible that decreased detoxification due to variant meH Tyr113His results in increased DNA damage which is inefficiently repaired by the base excision repair protein XRCC1 (Fig. 2). Presence of Arg399Gln SNP in an evolutionarily conserved region of XRCC1, expression of meH, XRCC1 in mucosa of the upper aerodigestive tract [17], [18], protein interaction of BRCA2 and p53 [19] strengthen this model.
Meta-analysis has shown varying effects of XRCC1 Arg280His in tobacco related cancers [1], [9]. Studies in Chinese lung cancer (n = 108) and Korean gastric cancer population (n = 172) showed significant risk association with Arg/His or His/His genotype [1], [20]. An Indian study [21] on oral cancer observed marginal risk conferred by His/His. Some studies reported no association of Arg280His with esophageal [22], bladder [23], gastric [24], [25] and lung cancer [1] although a few studies showed protection by His/His genotype in lung cancer [26], [27]. We observed stronger risk association of Arg/His in the patient population with at least one cancer in UADT-MPN group. Ethnic differences could be one of the determining factors in risk association, as mean frequency of His/His reported in Asian population is 13% (range 3–36%) where as in Caucasians it is 36% (range 2–47%) [10].
The variant meH Tyr113His which results in 30–50% decreased enzyme activity [28] has been significantly associated with cancers of larynx and lung [11], [12], [29]. Other studies observed no association with lung [30], head and neck [31] and laryngeal cancers [26]. Most of the studies show trend towards risk for this polymorphism and it is probable that the risk effect was enriched in combination with other genes in the MPN population.
Dominant and heterozygous models showed significant risk association of hOGG1 Ser326Cys SNP in UADT-TRC group. Studies have reported contrasting observation on risk association of this SNP in UADT cancers. Recently published Indian studies [27], [32] showed significant protective effect in head-neck cancer and no association of the hOGG1 Cys/Cys and Ser/Cys variant, whereas meta-analysis and another study [1] observed risk associaton of Cys/Cys genotype in lung cancer. Biochemical evidences support Cys/Cys genotype as risk conferring genotype due to lower protein activity compared to Ser/Ser variant, observed in head-neck [33] and lung cancer [34].
The BRCA2 Asn372His SNP showed significant risk association for dominant model for at least one TRC in UADT category and trend for risk in other models in at least one TRC in UADT and TRC outside UADT categories. The variation is located in the conserved region of the BRCA2 gene. Not much is known about the functional role of this SNP and its association in tobacco related cancers. The SNP has been associated with breast cancers [35] lymphoma [36], not associated with lung cancer [13], [37]. BRCA2 Asn372His showed higher sensitivity to gamma radiation along with other polymorphisms in the DNA repair pathway [13].
There are conflicting reports on the risk association of the p53 Arg72Pro SNP. While most studies report a weak protective association or no association of the p53 wild type Arg72Arg genotype for various cancers [38], [39] in our MDR model the Arg/Arg genotype in combination with the other two genotypes and tobacco, was associated with the risk of tobacco related MPNs. Several other studies have shown similar risk association between the wild type Arg/Arg genotype and breast [40], gastric [41], head and neck [42] and colorectal cancers [43], [44]. It may however be noted that studies which examined the gene-gene or gene environment interactions, the protective effect of the p53 Arg72Pro variant allele was seen in combination with other genotypes like the p53 intron 6 diplotype for head and neck cancers [42] and gastric cancer [41] or with the use of non steroidal anti-inflammatory drugs in colorectal cancer [44].
Despite the strengths and biological plausibility of the associations observed in our study, there are inherent limitations. Reliable estimation on the quanta of tobacco and alcohol consumed was not available as it was based on self-reported information. It is quite likely that several other important gene – gene and gene environment interactions exist that have not been evaluated in our study. It is also possible some of the SNPs studied and their interactions failed to emerge as significant risk association due to the limited sample size.
This is the first study to examine key SNPs in major metabolic and biological pathways implicated in tobacco carcinogenesis in the unique Indian MPN population. This study supports MPN to be an enriched model to predict cumulative genetic interactions. We anticipate the relevance of correlating the cumulative effect of variant genotypes to cellular phenotype in response to tobacco carcinogens. More importantly, for tobacco carcinogenesis it is difficult to quantify the redundancy of individual SNP, genes and pathways and this may vary in different geo-ethnic groups due to significant differences in the frequency of specific SNPs and or exposure to environmental, dietary co-carcinogens and protective agents. However, our approach to examine the multi-pathway tobacco carcinogenesis incorporates large body of research findings in a genetically enriched clinical model. Our approach could complement the GWAS approach by testing the leads provided by high quality GWAS studies.
Materials and Methods
Ethics Statement
Approval from Hospital Ethics Committee, Tata Memorial Centre, Mumbai was obtained before starting the study. Blood was collected after obtaining written informed consent from patients as well as healthy donors.
Study Population
Genotyping was carried out on 151 consecutive multiple primary neoplasm (MPN) patients. The cases were accrued from a registry of patients with MPN or familial cancers established at the Tata Memorial Hospital, Mumbai, in 1996 by one of the authors (RS). All the cases had histological or cytological confirmation of the primary cancer and each of the cancers was classified as TRC or non-TRC as per the IARC criteria [2]. There was no restriction for age at diagnosis, gender or carcinogen exposure. For defining two cancers as distinct multiple primaries, modified Hong's criteria [45] was used, which states that (a) there is >2 cm of normal intervening mucosa between two primaries in head and neck region; (b) lung as second primary if present, should be of different histology, or be solitary and with characteristic radiology of lung cancer; and (c) there is no evidence of haematogenous spread. Bilateral cancers in paired organs such as breast, ovaries or kidneys were not classified as MPN. Majority of the MPN cases in the registry were from the western and northern parts of India.
The cancer-free controls (n = 210) were volunteers who consented to donate blood or buccal washes for the study and were of similar geo-ethnic background as the cases. They were either visiting our hospital in the Preventive Oncology Department for cancer screening (n = 131) or visiting government dental college for various non-malignant, dental ailments (n = 73). A few controls were healthy, ethnically matched workers from Mumbai (n = 6). A majority of them were tobacco users (89%).
Detailed questionnaire including ethnicity and lifetime history of tobacco and alcohol use was obtained from all cases and controls. Family history of cancer was obtained from majority of MPN cases and cancer-free controls. After obtaining informed consent, 3–6 ml of peripheral blood was collected from each subject. Exfoliated buccal cells (mouthwash samples) were collected in sterile phosphate buffered saline from control individuals who were reluctant to give blood (n = 79). The study was approved by the Hospital Ethics Committee, Tata Memorial Centre, Mumbai.
DNA extraction and genotyping
Genomic DNA was extracted from peripheral blood/mouthwash samples using phenol chloroform method standardized in our laboratory [46]. Genotyping was done either by PCR-RFLP (Restriction fragment length polymorphism) or by SNaPshot method (ABI, USA). Primer sequences for PCRs were obtained from published literature and the conditions for PCR were standardized. The primer sequence, PCR conditions and restriction enzymes used for RFLP are available upon request. PCR was done in 96-well thermal cycler (ABI) in 25 µL volume containing PCR buffer (Invitrogen), 0.2 mmol/L deoxynucleotide triphosphates (Invitrogen), 0.5 mmol/L MgCl2 (Invitrogen), 0.25 unit Taq-Polymerase (Invitrogen), and 40 ng template DNA.
The authenticity of the PCR products was confirmed by sequencing at least five PCR products at random on an automated DNA sequencer (ABI Prism 3100 Avant) using the Big Dye terminator kit (ABI Prism, Foster City, CA, USA) as per the manufacturer's instructions.
Multiplex genotyping
Seven polymorphisms (NAT1, NAT2, BRCA1, BRCA2, GPX, meH and NAT3) were genotyped by multiplex PCR using SNaPshot. The assay was performed using SNaPshot ready reagent kit (ABI, USA). To the ready reagent, SBE primers (0.3 ρmol), EXO-SAP purified PCR products were added in a total reaction volume of 5 µl and incubated for 25 cycles of 96°C for 10 seconds, 50°C for 5 seconds and 60°C for 30 seconds. After the reaction, samples were purified by incubating with SAP (0.5 U) 37°C for 60 minutes followed by 75°C for 15 minutes. The purified products were run by capillary electrophoresis performed in 96 well plates in the ABI Prism™ 3100 genetic analyzer and analyzed using the Genemapper software (version 3.5). For SNP detection, the post purification products were denatured with de-ionized formamide and Genescan™ 120 Liz® size standard (Applied Biosystems) as per the manufacturers instructions at 95°C for 5 minutes followed by instant chilling on ice prior to loading on to the Sequencer. The electropherograms were depicted as two coloured peaks corresponding to two alleles for each heterozygous marker (SNP) or of one coloured peak for homozygous samples. As the fragments for each SNP are of varying sizes, the peaks did not overlap. To assure distinct recognition of closely lying peaks and avoid any chance of overlapping, the SNPs were grouped into two distinct panels based on the fragment size using the Primer Focus software (ABI). The software analyzed the genotypes according to the size and position of the alleles and accepted alleles that fall into the predetermined panel and represented genotypes of the entire sample set in a readily usable excel format.
Statistical Analysis
Univariate analysis.
Hardy Weinberg Equilibrium in the healthy controls was evaluated using χ2 test. Crude odds ratio and 95 percent confidence intervals were calculated for univariate analysis. For risk estimation the genotypes were a priori classified as homozygous low-risk or high-risk alleles based on their function in respective pathway. For each SNP, the Odds Ratio (OR) with its 95% Confidence Interval (CI) was estimated for the variant allele in its heterozygous and homozygous form taking the wild type homozygous allele as reference. Two sided p values were reported and considered significant if p<0.05.
Statistical analysis of gene-gene/gene-environment interactions.
In order to analyze interactions between SNPs and between SNPs and tobacco habits contributing to cancer risk, multifactor dimensionality reduction (MDR) approach was used. MDR is a non-parametric and genetic model free gene-gene interaction analysis method. This method had been proposed to overcome limitation of logistic regression in the analysis of high order interaction models where sparse data occur frequently [47], [48]. To account for individual data with missing values, ‘Available’ MDR approach was adopted in the analysis and analysis was performed using impute MDR in R packages [7]. Available MDR approach uses all the individuals who have complete data for a set of SNPs or habit variable that are included in a considered interaction model, thus it uses different number of individuals for each of possible interaction models. In the analysis of gene-gene or gene-environment interactions, individuals with more than 5 missing values were excluded. The analysis was repeated 10 times after shuffling the order of individuals and average of cross-validation consistency (CVC), training and test accuracies are presented. CVC is defined as the number of times a particular interaction model is selected across 10 cross-validation datasets. For the final selected model, we conducted odds ratio based MDR analysis (OR MDR) [8] to get the individual genotype effects.
Supporting Information
Table S1.
Univariate analysis of SNPs which do not show significant effects.
https://doi.org/10.1371/journal.pone.0030013.s001
(DOC)
Table S2.
The OR MDR analysis results for the final best model. The odds ratios having significant asymptotic confidence interval are in bold.
https://doi.org/10.1371/journal.pone.0030013.s002
(DOC)
Table S3.
Linkage Disequilibrium (LD) analysis.
https://doi.org/10.1371/journal.pone.0030013.s003
(DOC)
Acknowledgments
The authors wish to thank all patients and healthy volunteers who have given blood or mouth wash samples and the staff of Cancer Genetics Clinic, Tata Memorial Centre Memorial Hospital.
Author Contributions
Conceived and designed the experiments: AK RM. Performed the experiments: AK NJ. Analyzed the data: AK SK JN TP. Contributed reagents/materials/analysis tools: RM RS. Wrote the paper: AK RS RM. Patient accrual: RS.
References
- 1. Hung RJ, Hall J, Brennan P, Boffetta P (2005) Genetic polymorphisms in the base excision repair pathway and cancer risk: a HuGE review. Am J Epidemiol 162: 925–942.
- 2. (2004) Tobacco smoke and involuntary smoking. IARC Monogr Eval Carcinog Risks Hum 83: 1–1438.
- 3. Balmain A, Gray J, Ponder B (2003) The genetics and genomics of cancer. Nat Genet 33: Suppl238–244.
- 4. Jefferies S, Foulkes WD (2001) Genetic mechanisms in squamous cell carcinoma of the head and neck. Oral Oncol 37: 115–126.
- 5. Kotnis A, Sarin R, Mulherkar R (2005) Genotype, phenotype and cancer: role of low penetrance genes and environment in tumour susceptibility. J Biosci 30: 93–102.
- 6. Kuligina E, Reiner A, Imyanitov EN, Begg CB (2010) Evaluating cancer epidemiologic risk factors using multiple primary malignancies. Epidemiology 21: 366–372.
- 7. Namkung J, Elston RC, Yang JM, Park T (2009) Identification of gene-gene interactions in the presence of missing data using the multifactor dimensionality reduction method. Genet Epidemiol 33: 646–656.
- 8. Chung Y, Lee SY, Elston RC, Park T (2007) Odds ratio based multifactor-dimensionality reduction method for detecting gene-gene interactions. Bioinformatics 23: 71–76.
- 9. Zheng H, Wang Z, Shi X, Wang Z (2009) XRCC1 polymorphisms and lung cancer risk in Chinese populations: a meta-analysis. Lung Cancer 65: 268–273.
- 10. Kotnis A, Kannan S, Sarin R, Mulherkar R (2008) Case-control study and meta-analysis of SULT1A1 Arg213His polymorphism for gene, ethnicity and environment interaction for cancer risk. Br J Cancer 99: 1340–1347.
- 11. Jourenkova-Mironova N, Mitrunen K, Bouchardy C, Dayer P, Benhamou S, et al. (2000) High-activity microsomal epoxide hydrolase genotypes and the risk of oral, pharynx, and larynx cancers. Cancer Res 60: 534–536.
- 12. Lin TS, Huang HH, Fan YH, Chiou SH, Chow KC (2007) Genetic polymorphism and gene expression of microsomal epoxide hydrolase in non-small cell lung cancer. Oncol Rep 17: 565–572.
- 13. Zheng YL, Kosti O, Loffredo CA, Bowman E, Mechanic L, et al. (2010) Elevated lung cancer risk is associated with deficiencies in cell cycle checkpoints: genotype and phenotype analyses from a case-control study. Int J Cancer 126: 2199–2210.
- 14. Chung CC, Magalhaes WC, Gonzalez-Bosquet J, Chanock SJ (2010) Genome-wide association studies in cancer–current and future directions. Carcinogenesis 31: 111–120.
- 15. Jhavar SG, Sarin R, Chopra S, Kotnis A, Mulherkar R, et al. (2005) Females with paired occurrence of cancers in the UADT and genital region have a higher frequency of either Glutathione S-transferase M1/T1 null genotype. J Carcinog 4: 6.
- 16. Jhavar S, Sarin R, Mulherkar R, Benner A, Agarwal JP, et al. (2004) Glutathione S-transferase M1 or T1 null genotype as a risk factor for developing multiple primary neoplasms in the upper aero-digestive tract, in Indian males using tobacco. Oral Oncol 40: 84–91.
- 17. Angiero F, Berenzi A, Benetti A, Rossi E, Del Sordo R, et al. (2008) Expression of p16, p53 and Ki-67 proteins in the progression of epithelial dysplasia of the oral cavity. Anticancer Res 28: 2535–2539.
- 18. Blomquist T, Crawford EL, Mullins D, Yoon Y, Hernandez DA, et al. (2009) Pattern of antioxidant and DNA repair gene expression in normal airway epithelium associated with lung cancer diagnosis. Cancer Res 69: 8629–8635.
- 19. Marmorstein LY, Ouchi T, Aaronson SA (1998) The BRCA2 gene product functionally interacts with p53 and RAD51. Proc Natl Acad Sci U S A 95: 13869–13874.
- 20. Qu T, Morimoto K (2005) X-ray repair cross-complementing group 1 polymorphisms and cancer risks in Asian populations: a mini review. Cancer Detect Prev 29: 215–220.
- 21. Ramachandran S, Ramadas K, Hariharan R, Rejnish Kumar R, Radhakrishna Pillai M (2006) Single nucleotide polymorphisms of DNA repair genes XRCC1 and XPD and its molecular mapping in Indian oral cancer. Oral Oncol 42: 350–362.
- 22. Hao B, Wang H, Zhou K, Li Y, Chen X, et al. (2004) Identification of genetic variants in base excision repair pathway and their associations with risk of esophageal squamous cell carcinoma. Cancer Res 64: 4378–4384.
- 23. Stern MC, Umbach DM, van Gils CH, Lunn RM, Taylor JA (2001) DNA repair gene XRCC1 polymorphisms, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 10: 125–131.
- 24. Lee SG, Kim B, Choi J, Kim C, Lee I, et al. (2002) Genetic polymorphisms of XRCC1 and risk of gastric cancer. Cancer Lett 187: 53–60.
- 25. Lee JM, Lee YC, Yang SY, Yang PW, Luh SP, et al. (2001) Genetic polymorphisms of XRCC1 and risk of the esophageal cancer. Int J Cancer 95: 240–246.
- 26. To-Figueras J, Gene M, Gomez-Catalan J, Pique E, Borrego N, et al. (2002) Microsomal epoxide hydrolase and glutathione S-transferase polymorphisms in relation to laryngeal carcinoma risk. Cancer Lett 187: 95–101.
- 27. Mitra AK, Singh SV, Garg VK, Sharma M, Chaturvedi R, et al. (2011) Protective association exhibited by the single nucleotide polymorphism (SNP) rs1052133 in the gene human 8-oxoguanine DNA glycosylase (hOGG1) with the risk of squamous cell carcinomas of the head & neck (SCCHN) among north Indians. Indian J Med Res 133: 605–612.
- 28. Hassett C, Aicher L, Sidhu JS, Omiecinski CJ (1994) Human microsomal epoxide hydrolase: genetic polymorphism and functional expression in vitro of amino acid variants. Hum Mol Genet 3: 421–428.
- 29. Amador AG, Righi PD, Radpour S, Everett ET, Weisberger E, et al. (2002) Polymorphisms of xenobiotic metabolizing genes in oropharyngeal carcinoma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 93: 440–445.
- 30. Buch SC, Nazar-Stewart V, Weissfeld JL, Romkes M (2008) Case-control study of oral and oropharyngeal cancer in whites and genetic variation in eight metabolic enzymes. Head Neck 30: 1139–1147.
- 31. Lacko M, Roelofs HM, Te Morsche RH, Voogd AC, Oude Ophuis MB, et al. (2008) Microsomal epoxide hydrolase genotypes and the risk for head and neck cancer. Head Neck 30: 836–844.
- 32. Anantharaman D, Samant TA, Sen S, Mahimkar MB (2011) Polymorphisms in tobacco metabolism and DNA repair genes modulate oral precancer and cancer risk. Oral Oncol 47: 866–872.
- 33. Paz-Elizur T, Ben-Yosef R, Elinger D, Vexler A, Krupsky M, et al. (2006) Reduced repair of the oxidative 8-oxoguanine DNA damage and risk of head and neck cancer. Cancer Res 66: 11683–11689.
- 34. Paz-Elizur T, Krupsky M, Blumenstein S, Elinger D, Schechtman E, et al. (2003) DNA repair activity for oxidative damage and risk of lung cancer. J Natl Cancer Inst 95: 1312–1319.
- 35. Nathanson KL, Weber BL (2001) “Other” breast cancer susceptibility genes: searching for more holy grail. Hum Mol Genet 10: 715–720.
- 36. Salagovic J, Klimcakova L, Ilencikova D, Kafkova A (2011) Association of follicular lymphoma risk with BRCA2 N372H Polymorphism in Slovak population. Med Oncol.
- 37. Soumittra N, Meenakumari B, Parija T, Sridevi V, Nancy KN, et al. (2009) Molecular genetics analysis of hereditary breast and ovarian cancer patients in India. Hered Cancer Clin Pract 7: 13.
- 38. Economopoulos KP, Sergentanis TN, Zagouri F, Zografos GC (2010) Association between p53 Arg72Pro polymorphism and colorectal cancer risk: a meta-analysis. Onkologie 33: 666–674.
- 39. Jiang DK, Ren WH, Yao L, Wang WZ, Peng B, et al. (2010) Meta-analysis of association between TP53 Arg72Pro polymorphism and bladder cancer risk. Urology 76: 765 e761–767.
- 40. Bisof V, Salihovic MP, Narancic NS, Skaric-Juric T, Jakic-Razumovic J, et al. (2010) TP53 gene polymorphisms and breast cancer in Croatian women: a pilot study. Eur J Gynaecol Oncol 31: 539–544.
- 41. De Feo E, Persiani R, La Greca A, Amore R, Arzani D, et al. (2009) A case-control study on the effect of p53 and p73 gene polymorphisms on gastric cancer risk and progression. Mutat Res 675: 60–65.
- 42. Galli P, Cadoni G, Volante M, De Feo E, Amore R, et al. (2009) A case-control study on the combined effects of p53 and p73 polymorphisms on head and neck cancer risk in an Italian population. BMC Cancer 9: 137.
- 43. Dahabreh IJ, Linardou H, Bouzika P, Varvarigou V, Murray S (2010) TP53 Arg72Pro polymorphism and colorectal cancer risk: a systematic review and meta-analysis. Cancer Epidemiol Biomarkers Prev 19: 1840–1847.
- 44. Tan XL, Nieters A, Hoffmeister M, Beckmann L, Brenner H, et al. (2007) Genetic polymorphisms in TP53, nonsteroidal anti-inflammatory drugs and the risk of colorectal cancer: evidence for gene-environment interaction? Pharmacogenet Genomics 17: 639–645.
- 45. Hong WK, Lippman SM, Itri LM, Karp DD, Lee JS, et al. (1990) Prevention of second primary tumors with isotretinoin in squamous-cell carcinoma of the head and neck. N Engl J Med 323: 795–801.
- 46. Koppikar P, Mulherkar R (2006) A simple method for extraction of high molecular weight genomic DNA from buccal cells in mouthwash. Indian Journal of Biotechnology 5: 477–481.
- 47. Hahn LW, Ritchie MD, Moore JH (2003) Multifactor dimensionality reduction software for detecting gene-gene and gene-environment interactions. Bioinformatics 19: 376–382.
- 48. Ritchie MD, Hahn LW, Roodi N, Bailey LR, Dupont WD, et al. (2001) Multifactor-dimensionality reduction reveals high-order interactions among estrogen-metabolism genes in sporadic breast cancer. Am J Hum Genet 69: 138–147.