Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention

Transcription

1 Cancer Epidemiology 34 (2010) Contents lists available at ScienceDirect Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: Gene gene interactions of drug metabolizing enzymes and transporter protein in the risk of upper aerodigestive tract cancers among Indians Soya Sisy Sam a, *, Vinod Thomas a, Kanipakapatanam Sathyanarayana Reddy b, Gopalakrishnan Surianarayanan c, Adithan Chandrasekaran a a Department of Pharmacology, Jawaharlal Institute of Post-graduate Medical Education and Research, Pondicherry , India b Department of Radiotherapy, Jawaharlal Institute of Post-graduate Medical Education and Research, Pondicherry , India c Department of ENT, Jawaharlal Institute of Post-graduate Medical Education and Research, Pondicherry , India ARTICLE INFO ABSTRACT Article history: Accepted 15 May 2010 Available online 15 June 2010 Keywords: Genetic polymorphisms Cytochrome P-450 enzymes Glutathione-S-transferases ABCB1 drug transporter Gene gene interaction Upper aerodigestive tract cancers Indian population Background: The combined genetic effects of single nucleotide polymorphisms may additively or synergistically contribute to the increased cancer risk. The interactions associated with xenobiotic metabolizing enzymes and transporter protein involved in the biotransformation and transport of xenobiotics could determine the functional outcomes over the independent effects of a single susceptibility gene in the risk of upper aerodigestive tract cancers. Methods: The hospital-based case control study evaluated CYP1A1 (*2A and *2C), CYP2E1 (*1B, *5B, and *6), GST (M1, T1, and P1) and ABCB1 3435C>T polymorphisms among 408 histopathologically confirmed cases and 220 controls using polymerase chain reaction based methods in an Indian population. Results: The multivariate logistic regression analyses demonstrated potentially high risk gene gene interactions with the concurrent deletions of the GSTT1 and GSTM1 genes and GSTP1 variant genotypes (OR 5.81; 95% CI ), the deletions of GSTT1 and GSTM1 genotypes with variant genotypes of CYP1A1*2A (OR 8.21; 95% CI ), GSTT1 and GSTM1 deficient genotypes along with CYP2E1*1B variant genotypes (OR 6.73; 95% CI ), the polymorphic genotypes of ABCB1 and deficient GSTT1 (OR 6.08; 95% CI ) and an enhanced risk with the combined variant genotypes of CYP1A1*2A, GSTT1 and ABCB1 (OR 11.14; 95% CI ). Conclusion: The findings indicate that the interactions associated with various drug metabolizing enzymes and transporter protein exhibit high risk for UADT cancers than that ascribed to a single susceptible gene. This was particularly established among the polymorphic carriers of CYP1A1*2A, GSTT1 and ABCB1 genes in the population investigated. ß 2010 Elsevier Ltd. All rights reserved. 1. Introduction The cancers of upper aerodigestive tract (UADT) constitute a worldwide public health problem, despite advances in diagnosis and treatment protocols. These cancers are related group of malignancies that involve oral cavity, pharynx and larynx. It is among the sixth most common cancer worldwide [1]. In India, the most common UADT cancers are those of the oral cavity and pharynx [2]. These cancers are most commonly squamous cell carcinomas, originating from the squamous cells that line the upper aerodigestive tract [3]. The inherited differences in the capacity of xenobiotic metabolizing enzymes have been found to be an important factor that determines the genetic susceptibility to various malignancies [4]. Majority of the environmental procarcinogens require * Corresponding author. Tel.: ; fax: address: soyasoya@rediffmail.com (S.S. Sam). metabolic activation to get converted into their respective reactive electrophilic intermediates by the cytochrome P-450 superfamily of phase I enzymes. Subsequently, the phase II enzymes detoxify these intermediates by conjugation reactions. The metabolic activation and detoxification of xenobiotics express significant inter-individual differences in humans which are the key factors that regulate the genetic predisposition or host susceptibility to various carcinogens [5]. The cytochrome P-450 (CYP) dependent mono-oxygenases play an important role in the metabolism of environmental carcinogens. The metabolic activation of procarcinogens to reactive metabolites is mediated by these polymorphically distributed CYP enzymes in humans. Among these, CYP1A1 and CYP2E1 have been reported to be involved in the pathogenesis of various malignancies in diverse ethnic groups [6]. Cytochrome P-450 1A1 is a key enzyme in the phase I bioactivation of xenobiotics [5]. It contributes to the aryl hydrocarbon hydroxylase activity, catalyzing the first step in the metabolism of a number of polycyclic aromatic hydrocarbons. These include tobacco carcinogen, benzo (a) pyrene and several /$ see front matter ß 2010 Elsevier Ltd. All rights reserved. doi: /j.canep

2 S.S. Sam et al. / Cancer Epidemiology 34 (2010) other tobacco related procarcinogens such as nitrosamines and aromatic amines. They are metabolized to their ultimate DNAbinding forms. CYP1A1 gene is expressed in many epithelial tissues especially in buccal mucosa, which is responsible for the in situ activation of tobacco carcinogens [7]. CYP2E1 is responsible for the metabolic activation of procarcinogens such as N-nitrosamines and various other low-molecular-weight compounds into reactive intermediates that play an essential role in chemical carcinogenesis [8,9]. Thus, CYP2E1 activity is an important determinant of human susceptibility to toxicity and carcinogenicity of industrial and environmental chemicals. The activated carcinogens are detoxified by phase II enzymes such as glutathione-s-transferases (GST) to protect the body from the toxic effects of the carcinogens by catalyzing the conjugation of glutathione with electrophilic compounds [10]. The homozygous deletion of GSTM1 and GSTT1 genes results in complete lack of enzymatic activity [11]. Hence, carriers of null genotypes may be at increased risk in the development of UADT cancers because of the reduced ability to detoxify carcinogens. Another GST isoenzyme, GSTP1 is widely expressed in tumour cells and is responsible for the detoxification of benzo (a) pyrene diol epoxide and acrolein present in cigarette smoke [12]. Besides phase I and phase II enzymes, the genetic polymorphisms in drug transporters that are involved in the transmembrane efflux of xenobiotics may also result in susceptibility of normal tissues to neoplastic transformation by various carcinogens [13,14]. Besides phase I and phase II enzymes, the phase III transporters such as ABCB1 that belongs to Adenosine triphosphate binding cassette (ABC) superfamily encodes P-gp which is an energy dependent efflux pump that reduces the intracellular accumulation of a wide range of xenobiotics [15]. Human ABCB1 was identified due to its over expression in cultured cancer cells [16]. The ABCB1 gene exhibits polymorphisms, resulting in interindividual variations in the activity of P-gp [15]. Among the numerous polymorphisms in human ABCB1, the 3435C>T polymorphism in exon 26 has been associated with reduced mrna expression in the liver [17]. Several studies showing associations between genotypes and susceptibility are largely based on the effects of single genes. These studies consider the main rather than the interactive effects of a gene even though the main effect could be undetectable at times while an interactive effect could be large. We have conducted studies earlier to investigate the independent role as well as the gene environment interactions of certain candidate genes such as CYP1A1, CYP2E1, GST and ABCB1 in the risk of UADT cancers in Tamilian population of south India [18 21]. Our findings indicated that the genetic polymorphisms such as, CYP1A1*2A, GSTT1 null deletion and transporter protein, ABCB1 3435C>T were strong predisposing independent risk factors in the susceptibility to UADT cancers in the population. Further, we have also evaluated the gene environment interactions associated with the risk of UADT cancers for the genes under study among the tobacco users and alcoholics. These analyses revealed significant gene environment interactions with certain presumed genotypes eventhough some of them did not show an association when analyzed individually. In addition to studying the independent genetic effects as well as the gene environment interactions in the risk of UADT cancers in the population, we then proposed to investigate the gene gene interactions that might provide further information on the UADT cancer risk associated with the combinations of genes that are involved in the biotransformation and transport of xenobiotics. The gene gene interaction or epistasis is a unique component of the genetic architecture of common diseases, such as cancer [22]. Generally, the effect of a single nucleotide polymorphism (SNP) might be less compared to the genetic effect of combinations of functionally relevant SNPs that may additively or synergistically contribute to the increased cancer risk. These interactions might determine the functional outcomes over the independent effects of any single susceptibility gene or its genetic polymorphism. Generally, the inconsistencies among the SNP studies could be due to the fact that each individual SNP alters the function of only single gene of many, which are involved in carcinogenesis. The biological events associated with cancer risk that are modestly affected by a SNP may be more greatly affected by a SNP in combination with additional SNPs [23,24]. Hence, these studies could benefit from examining the interactions between polymorphisms in distinct genes that may contribute to the UADT cancer risk. There are a few studies on gene gene interactions describing inconsistent reports on the association between the genes belonging to phase I and phase II group and the risk of UADT cancers in various ethnic groups [25 28]. Apart from UADT cancers, a study was also conducted in Japanese to assess the interaction of these genes associated with gastric cancer [29]. Although the incidence of UADT cancers is one of the highest among the malignancies in India, there is only limited information available on the gene gene interactions and the risk of UADT cancers among Indians till date except for the two of the studies conducted in western and northern part of India recently [27,28]. In the face of this lacuna, the present study was planned to analyze the gene gene interactions for various genes that are involved in xenobiotic biotransformation and transport, with the aim of pinpointing the possible role of the combined at-risk genetic variants in the disease initiation and progression in the Tamilian population of south India. Tamilians are an ethnic group from south Asia and they are ethnically, linguistically and culturally related to the other Dravidian population. The oldest Tamil communities are present in southern India and northeastern Sri Lanka. There are an estimated 70 million Tamilians around the world [30]. 2. Subjects and methods 2.1. Subjects In the case control study, the case group consisted of histopathologically confirmed 408 patients (269 males and 139 females) with squamous cell carcinoma of UADT. They were diagnosed at the Departments of ENT and Radiotherapy, JIPMER hospital, Pondicherry, India. The control group included 220 subjects (148 males and 72 females) without present or previous history of malignancies. The age and sex matched controls were selected randomly from the same hospital and in the same period who came for treatment of various diseases other than malignancy. The male:female ratio in both the groups was 2:1. All the subjects were interviewed using a standardized questionnaire, regarding smoking habits, tobacco chewing habits and alcohol drinking history. The lifetime smoking consumption was expressed in pack-years. One pack-year corresponds to smoking one pack of cigarettes (10 cigarettes) per day for one year. Data on frequency of alcohol consumption and tobacco chewing were also estimated. The study was approved by the Institute Research Council and Institute Ethics Committee (human), JIPMER. After explaining the purpose and protocol of the study, the written informed consent was obtained from all the subjects Molecular analyses About five milliliters of venous blood was collected from each subject using ethylene diamine tetra acetic acid (EDTA) as anticoagulant. The genomic DNA was extracted from the peripheral blood leukocytes using standard phenol:chloroform

3 628 S.S. Sam et al. / Cancer Epidemiology 34 (2010) method. The CYP1A1, CYP2E1, GST and ABCB1 genotypes were determined by the polymerase chain reaction (PCR) based methods as described earlier [18 21]. Genotyping procedures were validated by sequencing of representative samples Statistical analyses Statistical analyses were done using the Statistical Package for Social Sciences statistical software (SPSS Windows version release 13). For analyzing the gene gene interactions, stratified variables (genotype genotype) were generated and included in the logistic model simultaneously with the appropriate indicator variables. Gene gene interaction analysis was conducted using the homozygous wild-type individuals for both the genes as the reference group. A multiplicative interaction was suggested when, Odds ratio (OR) +/+ > OR +/ OR /+, where OR+/+ = OR when both factors are present, OR+/ = OR when only factor 1 is present, OR /+ = OR when only factor 2 is present [31,32]. When the observed OR (OR+/ +) was more than the expected OR (OR+/ OR /+), then an interaction is said to be present. The observed genotype frequencies were compared with expected frequencies to check for the Hardy Weinberg equilibrium. 3. Results The mean (SEM) ages of cases and controls were 52.8 (0.49) and 52.3 (0.60) respectively. In the study, 87.2% of the patients with UADT cancers were above 40 years of age at the time of diagnosis. The number of smokers, alcohol consumers and tobacco chewers were more in cases than controls. An increase in the intensity of exposure and duration of the tobacco smoking, tobacco chewing and alcohol consumption showed a trend of dose response effect that was associated with an increased risk for UADT cancers compared to the controls. Of the 408 patients, 187 (45.8%) had oral cavity cancer, 141 (34.5%) had pharyngeal cancer and 80 (19.6%) had cancer in the laryngeal region (Table 1). The observed genotype frequencies for all the genes were consistent with Hardy Weinberg equilibrium. The combined analysis of genotypes, CYP1A1*2A and CYP1A1*2C did not show a significant interaction on the multiplicative scale. Also, no statistically significant interactions were observed between overall cancer risk and combinations of CYP2E1*1B and CYP2E1*6, CYP2E1*1B and CYP2E1*5B or CYP2E1*6 and CYP2E1*5B (data not shown). When the gene gene interactions associated with CYP1A1*2A and CYP2E1 genotypes were analyzed, statistically significant interactions were not observed between combinations of CYP1A1*2A and CYP2E1*1B, CYP1A1*2A and CYP2E1*6 or CYP1A1*2A and CYP2E1*5B and UADT cancer risk. Similarly, significant interactions were absent for the combinations of CYP1A1*2C and CYP2E1 (CYP2E*1B, CYP2E1*6 and CYP2E*5B) genotypes in the development of malignancies of UADT (data not shown). The risk associated with all the three high-risk GST genotypes (GSTM1 null, GSTT1 null and GSTP1 Ile/Val or Val/Val) compared to no risk genotypes (positive genotypes of GSTM1, GSTT1 and Ile/Ile genotype of GSTP1 designated as the reference group) were also investigated. The combined analyses of the GST genotypes showed significant interactions that increased the risk for UADT cancers. The individuals who presented both GSTM1 and GSTT1 deletion genotypes had nearly a 4-fold risk (OR 3.76; 95% CI ) of having the cancers than individuals without these characteristics which was comparable with the multiplicative OR of However, significant risk was absent in individuals with either of the null genotypes. Although there was a reduction in the risk for UADT cancers in individuals with GSTP1 Ile/Ile genotype and GSTT1 null genotype (OR 1.74; 95% CI ), the risk was found to be increased (OR 4.25; 95% CI ) in patients having GSTP1 Table 1 Demographic characteristics of the study subjects. Variables Cases (408) n (%) Age, yrs (mean SEM) Controls (220) n (%) OR (95% CI) Gender Male 269 (65.9) 148 (67.3) Female 139 (34.1) 72 (32.7) Smoking, (PY) a (36.7) 141 (64.1) (16.2) 34 (15.4) 2.13 ( ) * (14.5) 23 (10.5) 2.61 ( ) ** (15.0) 14 (6.4) 3.71 ( ) *** >40 72 (17.6) 8 (3.6) 7.90 ( ) **** Tobacco chewers b Never 134 (32.8) 160 (72.7) 1.0 Occasional 84 (20.6) 34 (15.5) 2.54 ( ) **** Regular 190 (46.6) 26 (11.8) ( ) **** Alcohol consumers c Never 181 (44.4) 146 (66.4) 1.0 Occasional 65 (15.9) 44 (20.0) 0.65 ( ) Regular 162 (39.7) 30 (13.6) 2.17 ( )* Site of carcinoma Oral cavity 187 (45.8) Pharynx 141 (34.5) Larynx 80 (19.6) PY: Pack-years. a Odds ratio adjusted for tobacco chewing and alcohol consumption. b Odds ratio adjusted for smoking and alcohol consumption. c Odds ratio adjusted for smoking and tobacco chewing. * P < ** P < *** P < **** P < polymorphic variants and GSTT1 null genotype which was higher than the expected OR of These two polymorphic genotypes in the presence of GSTM1 non-null gene have shown to reduce the cancer risk (OR 3.87; 95% CI ). Eventhough there was a reduced risk, the interaction observed might be due to the presence of the GSTM1 gene, where the observed OR was more than the expected OR of 2.36 which implicates a significant interaction. Further, an increased risk of 6-fold was observed in individuals carrying GSTM1 null, GSTT1 null genotypes and GSTP1 polymorphic variants (OR 5.81; 95% CI ). The observed OR of 5.81 was higher than the expected OR of 3.07 on the multiplicative scale (Table 2). The combinations of GSTM1 and CYP1A1*2A at-risk genotypes showed a 3-fold increased risk (OR 3.22; 95% CI ) when compared to individuals with polymorphic genotypes of either genes. The polymorphic CYP1A1*2A genotypes along with deficient GSTT1 resulted in more than 6.5-fold risk for the cancers (OR 6.68; 95% CI ). Furthermore, the individuals deficient for both GSTM1 and GSTT1 genotypes along with CYP1A1*2A mutant genotypes had an increased risk (OR 8.21; 95% CI ). The observed ORs of 3.22, 6.68 and 8.21 were found to be higher than the expected ORs of 2.83, 6.13 and 7.96 respectively on the multiplicative scale that emphasizes the importance of these gene gene interactions in the risk of UADT cancers (Table 3). However, no statistically significant interactions was observed with CYP1A1*2C genotypes and GSTM1 and GSTT1 genotype combinations (data not shown). The null genotype of GSTM1 and CYP2E1*1B mutant genotypes increased the risk to nearly a 4-fold (OR 3.90; 95% CI ) and GSTT1 null and polymorphic CYP2E1*1B resulted in more than a 4.5-fold increased risk (OR 4.64; 95% CI ) when

4 S.S. Sam et al. / Cancer Epidemiology 34 (2010) Table 2 Gene gene interactions of GSTM1, GSTT1 and GSTP1 genotypes in the study subjects. GSTM1 and GSTT1 M1 (+/+) and T1 (+/+) M1 ( / ) and T1 (+/+) ( ) M1 (+/+) and T1 ( / ) ( ) M1 ( / ) and T1 ( / ) ( ) b GSTM1 and GSTP1 M1 (+/+) and Ile/Ile M1 (+/+) and Ile/Val or Val/Val ( ) M1 ( / ) and Ile/Ile ( ) M1 ( / ) and Ile/Val or Val/Val ( ) GSTT1 and GSTP1 T1 (+/+) and Ile/Ile T1 (+/+) and Ile/Val or Val/Val ( ) T1 ( / ) and Ile/Ile ( ) T1 ( / ) and Ile/Val or Val/Val ( ) c GSTM1, GSTT1 and GSTP1 M1 (+/+), T1 (+/+) and Ile/Ile M1 (+/+), T1 (+/+) and Ile/Val or Val/Val ( ) M1 ( / ), T1 (+/+) and Ile/Ile ( ) M1 ( / ), T1 (+/+) and Ile/Val or Val/Val ( ) M1 (+/+), T1 ( / ) and Ile/Ile ( ) M1 (+/+), T1 ( / ) and Ile/Val or Val/Val ( ) d M1 ( / ), T1 ( / ) and Ile/Ile ( ) M1 ( / ), T1 ( / ) and Ile/Val or Val/Val ( ) e b Observed OR, 3.76 > Expected OR, c Observed OR, 4.25 > Expected OR, d Observed OR, 3.87 > Expected OR, e Observed OR, 5.81 > Expected OR, Table 3 Gene gene interactions of GSTM1, GSTT1 and CYP1A1*2A genotypes on UADT cancer risk. GSTM1 and CYP1A1*2A M1 (+/+) and *1A/*1A M1 (+/+) and *1A/*2A or *2A/*2A ( ) M1 ( / ) and *1A/*1A ( ) M1 ( / ) and *2A *1A/*2A or *2A/*2A ( ) b GSTT1 and CYP1A1*2A T1 (+/+) and *1A/*1A T1 (+/+) and *1A/*2A or *2A/*2A ( ) T1 ( / ) and *1A/*1A ( ) T1 ( / ) and *1A/*2A or *2A/*2A ( ) c GSTM1, GSTT1 and CYP1A1*2A M1(+/+), T1 (+/+) and *1A/*1A M1(+/+), T1 (+/+) and *1A/*2A or *2A/*2A ( ) M1 ( / ), T1 (+/+) and *1A/*1A ( ) M1 ( / ), T1 (+/+) and*1a/*2a or *2A/*2A ( ) M1 (+/+), T1 ( / ) and *1A/*1A ( ) M1 (+/+), T1 ( / ) and *1A/*2A or *2A/*2A ( ) M1 ( / ), T1 ( / ) and *1A/*1A ( ) M1 ( / ), T1 ( / ) and *1A/*2A or *2A/*2A ( ) d b Observed OR, 3.22 > Expected OR, c Observed OR, 6.68 > Expected OR, d Observed OR, 8.21 > Expected OR, compared to the carriers with either one of the polymorphic genotypes. Furthermore, the deficient genotypes of GSTM1 and GSTT1 along with CYP2E1*1B mutant genotypes were shown to have a significant interaction on the multiplicative scale (OR 6.73; 95% CI ). The observed ORs of 3.90, 4.64 and 6.73 were found to be higher than the expected ORs of 1.86, 4.02 and 5.22 respectively (Table 4). But, the combinations of either CYP2E1*6 or CYP2E1*5B and GSTT1 or GSTM1 deletion genotypes did not show any significant interactions on the multiplicative scale. Statistically significant interactions were not observed when mutant genotypes of GSTP1 along with various genotypes of CYP1A1 or CYP2E1 were analyzed separately (data not shown). The xenobiotic transporter gene, ABCB1 when combined with different genotypes of CYP1A1 and CYP2E1, no statistically significant interactions were observed (data not shown). The transporter gene along with the deficient GSTT1 showed a significant interaction on the multiplicative scale (OR 6.08; 95% CI ) which was higher than the expected OR of However, significant interactions were absent with GSTM1 deletion and GSTP1 mutant genotypes (Table 5).

5 630 S.S. Sam et al. / Cancer Epidemiology 34 (2010) Table 4 Gene gene interactions of GSTM1, GSTT1 and CYP2E1*1B genotypes on UADT cancer risk. GSTM1 and CYP2E1*1B M1 (+/+) and A2A M1 (+/+) and A2A1 or A1A ( ) M1 ( / ) and A2A ( ) M1 ( / ) and A2A1 or A1A ( ) b GSTT1 and CYP2E1*1B T1 (+/+) and A2A T1 (+/+) and A2A1 or A1A ( ) T1 ( / ) and A2A ( ) T1 ( / ) and A2A1 or A1A ( ) c GSTM1, GSTT1 and CYP2E1*1B M1 (+/+), T1 (+/+) and A2A M1 (+/+), T1 (+/+) and A2A1 or A1A ( ) M1 ( / ), T1 (+/+) and A2A ( ) M1 ( / ), T1 (+/+) and A2A1 or A1A ( ) M1 (+/+), T1 ( / ) and A2A ( ) M1 (+/+), T1 ( / ) and A2A1 or A1A ( ) M1 ( / ), T1 ( / ) and A2A ( ) M1 ( / ), T1 ( / ) and A2A1 or A1A ( ) d b Observed OR, 3.90 > Expected OR, c Observed OR, 4.64 > Expected OR, d Observed OR, 6.73 > Expected OR, Table 5 Gene gene interactions of ABCB1 and GST genotypes on UADT cancer risk. ABCB1 and GSTM1 CC and M1 (+/+) CC and M1 ( / ) ( ) CT or TT and M1 (+/+) ( ) CT or TT and M1 ( / ) ( ) ABCB1 and GSTT1 CC and T1 (+/+) CC and T1 ( / ) ( ) CT or TT and T1 (+/+) ( ) CT or TT and T1 ( / ) ( ) b ABCB1 and GSTP1 CC and Ile/Ile CC and Ile/Val or Val/Val ( ) CT or TT and Ile/Ile ( ) CT or TT and Ile/Val or Val/Val ( ) b Observed OR, 6.08 > Expected OR, Further, the gene gene interactions associated with the triple combinations of ABCB1, CYP1A1*2A and GSTs were analyzed by combining the various genotypes of these genes. Nearly a 5-fold increased risk was observed among the individuals polymorphic for ABCB1 3435C>T, CYP1A1*2A and GSTM1 null genotype (OR 4.82; 95% CI ); a further increase in risk was noted with the deficient genotype of GSTT1 instead of GSTM1 null genotype (OR 11.14; 95% CI ). The observed ORs of 4.82 and were higher than the expected ORs of 4.48 and 9.68 (Table 6). There was no significant gene gene interactions associated with the polymorphic genotypes of ABCB1, GSTM1, GSTT1, GSTP1 along with CYP2E1*1B, CYP2E1*6 and CYP2E1*5B in the population (data not shown). 4. Discussion In complex polygenic diseases such as UADT cancers, the genetic susceptibility may be dependent on the action of several gene polymorphisms operating in concert. Since, single gene polymorphism may contribute only to a small extent to the pathogenesis of UADT cancer, it is likely that the cumulative effect of many polymorphisms will be more important in its pathogenesis. Studies showing associations between genotypes and susceptibility are largely based on the effects of single genes. They consider therefore, main rather than interactive effects of a gene even though the main effect could be undetectable at times while an interactive effect could be large. The gene gene interactions apparent with the combinations of multilocus deficient GST genotypes modulated the risk associated with UADT cancer susceptibility than the risk related to a single GST deletion genotype. The risk associated with GSTT1 null genotype in the presence as well as in the absence of GSTM1 genotype depicts the independent involvement of GSTT1 gene deletion in the etiology of UADT cancers. The observation was consistent with a previously reported meta-analysis evaluating nine head and neck cancer studies for the combined effects of

6 S.S. Sam et al. / Cancer Epidemiology 34 (2010) Table 6 Gene gene interactions of ABCB1, GST and CYP1A1*2A genotypes on UADT cancer risk. ABCB1, GSTM1 and CYP1A1*2A CC, M1 (+/+) and *1A/*1A CC, M1 (+/+) and *1A/*2A or *2A/*2A ( ) CC, M1 ( / ) and *1A/*1A ( ) CC, M1 ( / ) and *1A/*2A or *2A/*2A ( ) CT or TT, M1 (+/+) and *1A/*1A ( ) CT or TT, M1 (+/+) and *1A/*2A or *2A/*2A ( ) CT or TT, M1 ( / ) and *1A/*1A ( ) CT or TT, M1 ( / ) and *1A/*2A or *2A/*2A ( ) b ABCB1, GSTT1 and CYP1A1*2A CC, T1 (+/+) and *1A/*1A CC, T1 (+/+) and *1A/*2A or *2A/*2A ( ) CC, T1 ( / ) and *1A/*1A ( ) CC, T1 ( / ) and *1A/*2A or *2A/*2A ( ) CT or TT, T1 (+/+) and *1A/*1A ( ) CT or TT, T1 (+/+) and *1A/*2A or *2A/*2A ( ) CT or TT, T1 ( / ) and *1A/*1A ( ) CT or TT, T1 ( / ) and *1A/*2A or *2A/*2A ( ) c ABCB1, GSTP1 and CYP1A1*2A CC, Ile/Ile and *1A/*1A CC, Ile/Ile and *1A/*2A or *2A/*2A ( ) CC, Ile/Val or Val/Val and *1A/*1A ( ) CC, Ile/Val or Val/Val and *1A/*2A or *2A/*2A ( ) CT or TT, Ile/Ile and *1A/*1A ( ) CT or TT, Ile/Ile and *1A/*2A or *2A/*2A ( ) CT or TT, Ile/Val or Val/Val and *1A/*1A ( ) CT or TT, Ile/Val or Val/Val and *1A/*2A or *2A/*2A ( ) b Observed OR, 4.82 > Expected OR, c Observed OR, > Expected OR, deficient GSTM1 and GSTT1 genotypes [33]. The decreased risk noticed when GSTP1 wild genotype combined with GSTT1 null genotype might be due to the fact that tobacco or alcohol derived carcinogens and toxins are multiple substrates for GSTP1. This was confirmed by the increased risk observed among carriers of GSTP1 polymorphic variants along with GSTT1 null genotype. In addition, a significant association was observed for concurrent deletion of the GSTM1, GSTT1 genes and mutant genotypes of GSTP1 indicating that individuals having a defective genotype for more than one of these genes would therefore be at greater risk. Several carcinogens, including polycyclic aromatic hydrocarbons (PAHs) present in the tobacco smoke are reported to be detoxified by GSTs [34]. Similar results were reported for various combinations of GST polymorphisms and susceptibility to oral leukoplakia [35], larynx cancers [36] and head and neck cancers [37]. An Indian study had reported 2-fold risk associated with head and neck cancer patients carrying the GSTM1 and GSTT1 null genotypes individually, though the risk was not significant with multivariate logistic regression model analysis. The wild type genotype of GSTP1 in combination with GSTM1 null or GSTT1 null genotype increased the susceptibility to 2.5- and 2.8-folds respectively for the cancers. The risk was further increased to nearly 4.5-fold, when the combinations of GSTM1 null, GSTT1 null and wild type genotype of GSTP1 was analyzed [38]. Identification of individuals with null genotypes of GSTs may eventually assist in the prevention of UADT cancers by allowing early detection of individuals with a high risk. Therefore, the deficiency of GST genes especially in combinations is important in public health issues related to UADT cancers. The interaction between phase II deficient enzymes and a phase I hyperactive enzyme (CYP1A1) is of interest as it may result in a larger amount of toxic compounds that might have a role in the initiation or progression of UADT cancers. In the present study, the wild type genotypes of CYP1A1*2A and GSTM1 taken together as baseline, the risk associated with the combinations of the mutant genotypes of CYP1A1*2A and deficient GSTM1 were determined. Individuals with combined mutant genotypes of CYP1A1*2A and GSTM1 null genotype had an increased risk for the cancers that was higher than the risk elevation observed for the susceptible genotypes of CYP1A1 or GSTM1 gene alone. The interaction between CYP1A1 and GSTM1 is of interest as GSTM1 null may be associated with CYP1A1 inducibility [39]. Thus, our findings were in accordance with an Indian study that reported a significant risk for oral cancer among those carriers of both CYP1A1*2A homozygous variant and GSTM1 null genotype [27]. The Japanese, Korean and Brazilian studies also support our findings which demonstrated the definite role of combined CYP1A1*2A homozygous variant and GSTM1 deficient genotypes in the risk of head and neck cancers [25,40 42]. Further, the combined effects of mutant variants of CYP1A1*2A and deletion genotype of GSTT1 resulted in high risk for the carcinogenesis in our population. A lung cancer study done in a south Indian population had reported high risk associated with combinations of variant or null genotypes of CYP1A1*2A, GSTM1 and GSTT1 that further supports our findings [43]. While the individuals deficient for both GST genes in presence of mutant genotypes of CYP1A1*2A were at remarkably high risk than individuals who had either of the deletion along with CYP1A1*2A mutant genotypes. These observations indicate that metabolic imbalance of activation and detoxification of CYP1A1*2A and GST deletion genotypes respectively worked synergistically for UADT cancer risk. The initial step in the metabolic activation of PAHs is mediated by cytochrome P-450 gene products, mainly the isozyme, CYP1A1. In addition to the metabolic activation of PAHs by CYP1A1 enzyme, it is also involved in the bioactivation of several other tobacco related procarcinogens such as nitrosamines and aromatic amines to carcinogenic products [44,45]. The carcinogenic metabolites

7 632 S.S. Sam et al. / Cancer Epidemiology 34 (2010) produced by the hyperactive CYP1A1 enzyme failed to be eliminated as a result of the lack of GST genes that are mainly involved in the detoxification of carcinogenic substances. Interindividual differences in the rate of metabolism (activation/ detoxification) have been thought to be due to genetic polymorphisms. Among the CYP2E1 genotypes, polymorphic genotypes of CYP2E1*1B have observed to produce significant interactions among the GST deficient individuals. Interestingly, eventhough, GSTM1 or CYP2E1*1B genotypes when analyzed alone did not show a significant association with UADT cancers, the combined analysis of these genotypes showed a significant interaction corresponding to nearly a 4-fold increased risk. Similarly, CYP2E1*1B along with GSTT1, further increased the risk to more than 4.5-fold while both the GST deletion genotypes and CYP2E1*1B had a remarkably high interaction that contributed to nearly 7-fold risk in the initiation of UADT cancers. The interactions noticed along with the deficient GST might be due to the enhanced activity of CYP2E1 enzyme associated with polymorphism in CYP2E1*1B [46]. There are no studies reporting on the interaction of CYP2E1*1B and other presumed genotypes which indeed to be confirmed in other ethnic groups would be of clinical significance. However, in the present study, no significant interactions were observed for CYP2E1*5B and CYP2E1*6 when combined with GST genotypes. A study conducted in northern part of India reported a significant increase in risk to head and neck cancers in patients carrying a combination of variant genotypes of CYP2E1 (*5B and *6) and GSTM1 null genotype which was in contrast to our finding [28]. A Japanese study reported a significant interaction between CYP2E1*5B and GSTM1 null genotypes and the risk for gastric cancer [29]. On the contrary, a Chinese study reported that the GSTM1 non-null genotype had a synergistic effect along with CYP2E1 c1/c1 genotype causing an 8.5-fold risk to esophageal cancer [47]. However, an Italian study did not find significant gene gene interactions when the genotypes of CYP1A1, CYP2E1, GSTM1 and GSTT1 were analyzed for the risk of head and neck cancers [26]. The main activity that occurs during the phase III elimination is the excretion of the parent drugs, metabolites, and xenobiotics from the cell via the transmembrane efflux pump, P-gp [15,14]. In the present study, the subjects carrying the transporter protein, ABCB1 C3435C>T and deficient GSTT1 had more than 6-fold increased risk which was higher than the risk observed for individual polymorphic genes. Thus, it explains the risk related in the presence of a single malfunctioning elimination system compared to the increased risk associated with multiple detoxification or transport pathways that are polymorphic in the accumulation of carcinogens in the body which would result in the transformation of normal cells to neoplastic cells. Furthermore, the combined analysis of the genes associated with metabolic activation and elimination pathways have shown an enhanced risk for polymorphic ABCB1, GSTM1 and CYP1A1*2A genes. A remarkable 11-fold higher risk was observed, when deficient GSTT1 was replaced with GSTM1 in the presence of ABCB1 and CYP1A1*2A polymorphic genes. The variant forms of the three genes (ABCB1, GSTT1 and CYP1A1*2A) have been shown to play a significant role in the etiology of UADT cancers when it was separately analyzed. Therefore, the individuals carrying these polymorphic genotypes that are mainly involved in metabolic activation as well as elimination are highly susceptible to develop cancers associated with UADT. Hence, our study indicates that individuals polymorphic for CYP1A1*2A, GSTT1 and ABCB1 genes have a genetically high risk for UADT cancers, and combined genotypes of these susceptible genes revealed higher risk than that ascribed to a single susceptible gene. The structure activity analyses of various transport substrates of P-gp have revealed that, it is associated with the elimination of hydroxylated carcinogenic metabolites [48]. However, we did not find any published work that demonstrated an interaction of phases I and II along with phase III presumed genotypes especially ABCB1 3435C>T which need to be confirmed in other ethnic groups. By evaluating the associations between genotypes and disease phenotype, it may reflect a link between at-risk genotypes, alone or in combination, to identify subjects who are poor metabolizers and consequently more likely to suffer formation of carcinogen-dna adducts and initiation of carcinogenesis. Cancer is a complex disease in which multiple loci are involved and only the combined analysis of several markers could provide new clues for prediction of its occurrence. These positive findings suggest that interindividual variation in the metabolic pathways involved in the functionalization and detoxification of specific xenobiotics is an important susceptibility factor for UADT cancers in Tamilian population. Conflict of interest We declare that we have no conflict of interest in relation to this article. Acknowledgements The study was funded by Indian Council of Medical Research, New Delhi (ICMR Ref. No. 3/2/2/31/2003-NCD-III dated 09/10/ 2003). We are thankful to Dr. Subramanian (Vector Control Research Centre, Pondicherry, India) for his assistance in the statistical analyses. References [1] Khandekar SP, Bagdey PS, Tiwari RR. Oral cancer and some epidemiological factors: a hospital based study. Indian J Commun Med 2006;31(3): [2] Petersen PE. Strengthening the prevention of oral cancer: the WHO perspective. Community Dent Oral Epidemiol 2005;33(6): [3] Sanderson RJ, Ironside JA. Squamous cell carcinomas of the head and neck. BMJ 2002;325(7368): [4] Hung RJ, van der HO, Tavtigian SV, Brennan P, Boffetta P, Hashibe M. Perspectives on the molecular epidemiology of aerodigestive tract cancers. Mutat Res 2005;592(1 2): [5] Nebert DW. Role of genetics and drug metabolism in human cancer risk. Mutat Res 1991;247(2): [6] Gajecka M, Rydzanicz M, Jaskula-Sztul R, Kujawski M, Szyfter W, Szyfter K. CYP1A1, CYP2D6, CYP2E1, NAT2, GSTM1 and GSTT1 polymorphisms or their combinations are associated with the increased risk of the laryngeal squamous cell carcinoma. Mutat Res 2005;574(1 2): [7] Vondracek M, Xi Z, Larsson P, Baker V, Mace K, Pfeifer A, et al. Cytochrome P450 expression and related metabolism in human buccal mucosa. Carcinogenesis 2001;22(3): [8] Nouso K, Thorgeirsson SS, Battula N. Stable expression of human cytochrome P450IIE1 in mammalian cells: metabolic activation of nitrosodimethylamine and formation of adducts with cellular DNA. Cancer Res 1992;52(7): [9] Koop DR. Oxidative and reductive metabolism by cytochrome P450 2E1. FASEB J 1992;6(2): [10] Taningher M, Malacarne D, Izzotti A, Ugolini D, Parodi S. Drug metabolism polymorphisms as modulators of cancer susceptibility. Mutat Res 1999; 436(3): [11] Board P, Coggan M, Johnston P, Ross V, Suzuki T, Webb G. Genetic heterogeneity of the human glutathione transferases: a complex of gene families. Pharmacol Ther 1990;48(3): [12] Matthias C, Bockmuhl U, Jahnke V, Harries LW, Wolf CR, Jones PW, et al. The glutathione S-transferase GSTP1 polymorphism: effects on susceptibility to oral/pharyngeal and laryngeal carcinomas. Pharmacogenetics 1998;8(1):1 6. [13] Liska DJ. The detoxification enzyme systems. Altern Med Rev 1998;3(3): [14] Xu C, Li CY, Kong AT. Induction of phase I, II and III drug metabolism/transport by xenobiotics. Arch Pharm Res 2005;28(3): [15] Marzolini C, Paus E, Buclin T, Kim RB. Polymorphisms in human MDR1 (Pglycoprotein): recent advances and clinical relevance. Clin Pharmacol Ther 2004;75(1): [16] Ishikawa T, Onishi Y, Hirano H, Oosumi K, Nagakura M, Tarui S. Pharmacogenomics of drug transporters: a new approach to functional analysis of the genetic polymorphisms of ABCB1 (P-glycoprotein/MDR1). Biol Pharm Bull 2004;27(7):

8 S.S. Sam et al. / Cancer Epidemiology 34 (2010) [17] Wang D, Johnson AD, Papp AC, Kroetz DL, Sadee W. Multidrug resistance polypeptide 1 (MDR1, ABCB1) variant 3435C>T affects mrna stability. Pharmacogenet Genom 2005;15(10): [18] Sam SS, Thomas V, Reddy KS, Surianarayanan G, Chandrasekaran A. CYP1A1 polymorphisms and the risk of upper aerodigestive tract cancers in an Indian population. Head Neck 2008;30(12): [19] Soya SS, Vinod T, Reddy KS, Gopalakrishnan S, Adithan C. CYP2E1 polymorphisms and gene environment interactions in the risk of upper aerodigestive tract cancers among Indians. Pharmacogenomics 2008;9(5): [20] Soya SS, Vinod T, Reddy KS, Gopalakrishnan S, Adithan C. Genetic polymorphism of glutathione-s-transferase genes (GSTM1, GSTT1 and GSTP1) and susceptibility to upper aerodigestive tract cancers in a south Indian population. Eur J Cancer 2007;43(18): [21] Sam SS, Thomas V, Sivagnanam K, Reddy KS, Surianarayanan G, Chandrasekaran A. ABCB1 genetic polymorphism and risk of upper aerodigestive tract cancers among smokers, tobacco chewers and alcoholics in an Indian population. Pharmacogenet Genomics 2007;17(10): [22] Nagel RL. Epistasis and the genetics of human diseases. C R Biol 2005;328(7): [23] Moore JH. The ubiquitous nature of epistasis in determining susceptibility to common human diseases. Hum Hered 2003;56(1 3): [24] Goodman JE, Mechanic LE, Luke BT, Ambs S, Chanock S, Harris CC. Exploring SNP SNP interactions and colon cancer risk using polymorphism interaction analysis. Int J Cancer 2006;118(7): [25] Gattas GJ, de Carvalho MB, Siraque MS, Curioni OA, Kohler P, Eluf-Neto J, et al. Genetic polymorphisms of CYP1A1, CYP2E1, GSTM1, and GSTT1 associated with head and neck cancer. Head Neck 2006;28(9): [26] Boccia S, Cadoni G, Sayed-Tabatabaei FA, Volante M, Arzani D, De Lauretis A, et al. CYP1A1, CYP2E1, GSTM1, GSTT1, EPHX1 exons 3 and 4, and NAT2 polymorphisms, smoking, consumption of alcohol and fruit and vegetables and risk of head and neck cancer. J Cancer Res Clin Oncol 2008;134(1): [27] Anantharaman D, Chaubal PM, Kannan S, Bhisey RA, Mahimkar MB. Susceptibility to oral cancer by genetic polymorphisms at CYP1A1, GSTM1 and GSTT1 loci among Indians: tobacco exposure as a risk modulator. Carcinogenesis 2007;28(7): [28] Ruwali M, Khan AJ, Shah PP, Singh AP, Pant MC, Parmar D. Cytochrome P450 2E1 and head and neck cancer: interaction with genetic and environmental risk factors. Environ Mol Mutagen 2009;50(6): [29] Suzuki S, Muroishi Y, Nakanishi I, Oda Y. Relationship between genetic polymorphisms of drug-metabolizing enzymes (CYP1A1, CYP2E1, GSTM1, and NAT2), drinking habits, histological subtypes, and p53 gene point mutations in Japanese patients with gastric cancer. J Gastroenterol 2004;39(3): [30] Tamil Nation [home page on the internet]: a trans state nation: Tamil Diaspora. Available from: URL: Accessed on March 24, [31] Brennan P. Gene environment interaction and aetiology of cancer: what does it mean and how can we measure it? Carcinogenesis 2002;23(3): [32] Boccia S, Sayed-Tabatabaei FA, Persiani R, Gianfagna F, Rausei S, Arzani D, et al. Polymorphisms in metabolic genes, their combination and interaction with tobacco smoke and alcohol consumption and risk of gastric cancer: a case control study in an Italian population. BMC Cancer 2007;7:206. [33] Ye Z, Song H, Guo Y. Glutathione S-transferase M1, T1 status and the risk of head and neck cancer: a meta-analysis. J Med Genet 2004;41(5): [34] Coles B, Ketterer B. The role of glutathione and glutathione transferases in chemical carcinogenesis. Crit Rev Biochem Mol Biol 1990;25(1): [35] Nair UJ, Nair J, Mathew B, Bartsch H. Glutathione S-transferase M1 and T1 null genotypes as risk factors for oral leukoplakia in ethnic Indian betel quid/ tobacco chewers. Carcinogenesis 1999;20(5): [36] Hong YJ, Lee JK, Lee GH, Hong SI. Influence of glutathione S-transferase M1 and T1 genotypes on larynx cancer risk among Korean smokers. Clin Chem Lab Med 2000;38(9): [37] Hashibe M, Brennan P, Strange RC, Bhisey R, Cascorbi I, Lazarus P, et al. Metaand pooled analyses of GSTM1, GSTT1, GSTP1, and CYP1A1 genotypes and risk of head and neck cancer. Cancer Epidemiol Biomarkers Prev 2003;12(12): [38] Singh M, Shah PP, Singh AP, Ruwali M, Mathur N, Pant MC, et al. Association of genetic polymorphisms in glutathione S-transferases and susceptibility to head and neck cancer. Mutat Res 2008;638(1 2): [39] Vaury C, Laine R, Noguiez P, de Coppet P, Jaulin C, Praz F, et al. Human glutathione S-transferase M1 null genotype is associated with a high inducibility of cytochrome P450 1A1 gene transcription. Cancer Res 1995;55(23): [40] Tanimoto K, Hayashi S, Yoshiga K, Ichikawa T. Polymorphisms of the CYP1A1 and GSTM1 gene involved in oral squamous cell carcinoma in association with a cigarette dose. Oral Oncol 1999;35(2): [41] Cha IH, Park JY, Chung WY, Choi MA, Kim HJ, Park KK. Polymorphisms of CYP1A1 and GSTM1 genes and susceptibility to oral cancer. Yonsei Med J 2007;48(2): [42] Sato M, Sato T, Izumo T, Amagasa T. Genetic polymorphism of drug-metabolizing enzymes and susceptibility to oral cancer. Carcinogenesis 1999;20(10): [43] Sreeja L, Syamala V, Hariharan S, Madhavan J, Devan SC, Ankathil R. Possible risk modification by CYP1A1, GSTM1 and GSTT1 gene polymorphisms in lung cancer susceptibility in a South Indian population. J Hum Genet 2005; 50(12): [44] Roy B, Sikdar N. Polymorphisms in drug-metabolizing genes and risk of head and neck squamous cell carcinoma. Int J Hum Genet 2003;3(2): [45] Thier R, Bruning T, Roos PH, Rihs H, Golka K, Ko Y, et al. Markers of genetic susceptibility in human environmental hygiene and toxicology: the role of selected CYP, NAT and GST genes. Int J Hyg Environ Health 2003;206(3): [46] Haufroid V, Buchet J, Gardinal S, Ghittori S, Imbriani M, Hirvonen A, et al. Importance of genetic polymorphisms of drugmetabolizing enzymes for the interpretation of biomarkers of exposure to styrene. Biomarkers 2001;6(3): [47] Tan W, Song N, Wang GQ, Liu Q, Tang HJ, Kadlubar FF, et al. Impact of genetic polymorphisms in cytochrome P450 2E1 and glutathione S-transferases M1, T1, and P1 on susceptibility to esophageal cancer among high-risk individuals in China. Cancer Epidemiol Biomarkers Prev 2000;9(6): [48] Bain LJ, McLachlan JB, LeBlanc GA. Structure activity relationships for xenobiotic transport substrates and inhibitory ligands of P-glycoprotein. Environ Health Perspect 1997;105(8):812 8.