Abstract
Individual susceptibility to the toxic effects of cigarette smoke may be modified by inherited variability in carcinogen metabolism. The purpose of the present study was to investigate pancreatic cancer risk associated with cigarette smoking and 33 variants within carcinogen metabolism genes and examine whether these variants modify the association between smoking and pancreatic cancer. A population-based study was conducted with 455 pancreatic cancer cases and 893 controls. Epidemiological and smoking data were collected from questionnaires and variants were genotyped by mass spectrometry. Age- and sex-adjusted odds ratio (ASOR) and multivariate-adjusted odds ratio (MVOR) estimates were obtained using multivariate logistic regression, and interactions between each variant and smoking were investigated. Current smoker status [MVOR = 2.29, 95% confidence interval (95% CI): 1.62, 3.22], 10–27 pack-years (MVOR = 1.57, 95% CI: 1.13, 2.18), >27 pack-years (MVOR = 1.77, 95% CI: 1.27, 2.46) and longer durations of smoking (19–32 years: MVOR = 1.46, 95% CI: 1.05, 2.05; >32 years: MVOR = 1.78, 95% CI: 1.30, 2.45) were associated with increased pancreatic cancer risk. CYP1B1-4390-GG (ASOR = 0.36, 95% CI: 0.15, 0.86) and Uridine 5'-diphospho glucuronosyltransferase 1 family, polypeptide A7-622-CT (ASOR = 0.77, 95% CI: 0.60, 0.99) were associated with reduced risk. N-acetyltransferase 1-640-GT/GG (ASOR = 1.75, 95% CI: 1.00, 3.05), GSTM1 (rs737497)-GG (ASOR = 1.41, 95% CI: 1.02, 1.95), GSTM1 gene deletion (ASOR = 4.89, 95% CI: 3.52, 6.79) and glutathione S-transferase theta-1 gene deletion (ASOR = 4.41, 95% CI: 2.67, 7.29) were associated with increased risk. Significant interactions were observed between pack-years and EPHX1-415 (P = 0.04) and smoking status and N-acetyltransferase 2-857 (P = 0.03). Variants of carcinogen metabolism genes are independently associated with pancreatic cancer risk and may modify the risk posed by smoking.
Introduction
In Canada, there were ∼4000 new cases and 3900 deaths from pancreatic cancer in 2010 (1) with an overall 5-year survival of 5% (2). Even among those who undergo surgical resection with curative intent, clinical outcomes are not optimal with the 5-year survival of resectable patients being ∼15–20% (3). Such poor prognosis emphasizes the importance of a better understanding of modifiable pancreatic cancer risk factors and improved primary prevention.
Numerous studies, such as those that have examined cohorts of men (4) and women (5), conducted in different continents, such as North America (6), Europe (7) and Asia (8) and using large multicenter designs (9) have all suggested that cigarette smoking is a risk factor for pancreatic cancer. Two recent meta-analyses (10,11) on smoking status and pancreatic cancer yielded pooled relative risk estimates between 1.20 and 2.10, depending on the studies included and the definition of smoking status used. In addition, dose-response relationships have been demonstrated with higher cigarette consumption and longer smoking duration with increased pancreatic cancer risk, which may persist even 10 years after quitting (10).
Cigarette smoke contains numerous known carcinogens, such as polycyclic aromatic hydrocarbons, N-nitrosamines, aromatic amines, 1,3-butadiene, benzene, aldehydes and ethylene oxide (12). Direct evidence of nitrosamines, heterocyclic amines, aromatic amines and reactive oxygen metabolite-mediated DNA damage has been demonstrated in pancreatic tissue in both human specimens and animal models (13–17). Since several carcinogens are metabolized by xenobiotic-metabolizing enzymes such as cytochrome P450 (CYP) enzymes (18), it is plausible that variants of CYP and other metabolic genes are associated with individual variability in carcinogen metabolism and subsequent risk for pancreatic cancer development.
Few epidemiological studies have investigated the hypothesis of genetic polymorphisms modifying the association between smoking and pancreatic cancer, yielding mixed results. Duell et al. (19) report an increased risk for pancreatic cancer among heavy smokers with glutathione S-transferase theta 1 (GSTT1)-null genotype, whereas three other studies (20–22) report no association between GSTT1 genotype and pancreatic cancer risk. Similarly, results of studies regarding N-acetyltransferase 1 (NAT1) and N-acetyltransferase 2 (NAT2) (23–27) and Uridine 5'-diphospho glucuronosyltransferase 1 family, polypeptide A7 (UGT1A7) (28–30) have been conflicting. These mixed results make it challenging to draw definite conclusions and furthermore, the majority of studies in this area has investigated only a small number of variants and is limited by small sample sizes. The present study is a large population-based case–control study, investigating the possibility of effect modification by functionally important carcinogen metabolism gene variants of the association between cigarette smoking and pancreatic cancer risk. Our specific aims were to: (i) estimate the risk of pancreatic cancer associated with smoking; (ii) evaluate the association between variants of carcinogen metabolism genes and pancreatic cancer risk and (iii) determine if the association between smoking and pancreatic cancer is modified by these carcinogen metabolism gene variants.
Materials and methods
Study design
Population-based pancreatic cancer cases and population-based controls were recruited within Ontario, Canada. Pancreatic cancer cases were participants of the Ontario Pancreas Cancer Study (OPCS) and controls were participants of the Ontario Familial Colorectal Cancer Registry (OFCCR). The OPCS is one of seven study sites of the Pancreatic Cancer Genetic Epidemiology Consortium (31), whereas the OFCCR is one of six sites of the Colorectal Cancer Family Registry (32). Methodologies of the OPCS (6,33) and the OFCCR (32,34) have previously been described but are briefly summarized below.
Participants
Cases had a primary pathology-confirmed adenocarcinoma of the pancreas or adenocarcinoma metastasis, confirmed as pancreatic cancer by treating physicians, diagnosed between April 2003 and July 2009 and initially identified through the population-based Ontario Cancer Registry. Patients with neuroendocrine tumors or cystadenocarcinoma of the pancreas are excluded from the OPCS and consequently excluded from our study. After consent to contact the patient was received from the treating physician, eligible cases were mailed three self-administered questionnaires—family history questionnaire, personal history (epidemiology) questionnaire, clinic patient questionnaire and a consent form. Blood samples, medical records and a tissue or tumor sample were requested and obtained from consenting participants. Initially 456 cases were eligible for analyses, however, one case had missing smoking data and was subsequently excluded from analyses. A small number of cases (8/455 cases, 1.76%) had a proxy respondent (for example, a spouse) complete the questionnaires on their behalf. The response rate for cases in the OPCS is estimated to be ∼30% (33), which is consistent with other population-based studies of pancreatic cancer [e.g. (35,36)] due to the rapid fatality of this disease.
Population-based controls were initially recruited as controls by the OFCCR by random digit dialing methods and the Ministry of Finance Property Assessment Database during 2002–03. All controls did not report a personal history of colorectal or pancreatic cancer. Controls were mailed three self-administered questionnaires [family history questionnaire, personal history (epidemiology) questionnaire and diet questionnaire]. Initially 903 controls were eligible for analyses, however 9 controls had missing smoking data and 1 control failed genotyping of all variants of interest and were subsequently excluded. The response rate for controls in the OFCCR is estimated to be ∼61% (34).
Data collection
Data on cigarette smoking and potential confounders were collected from the personal history (epidemiology) questionnaire and family history questionnaire. To account for potential changes in smoking behavior associated with cancer development and diagnosis, all smoking variables were calculated from a reference point of 2 years prior to pancreatic cancer diagnosis for cases and 2 years prior to questionnaire completion for controls. Three variables were derived to describe smoking behavior: (i) smoking status for participants was classified by never-smoker (never smoked or smoked less than a threshold of >100 lifetime cigarettes or >1 cigarette per day for ≥3 months), former smoker (smoked cigarettes >2 years prior to cancer diagnosis for cases and >2 years prior to questionnaire completion for controls) and current smoker (smoked cigarettes ≤2 years prior to cancer diagnosis for cases and ≤2 years prior to questionnaire completion for controls); (ii) smoking pack-years defined as the number of packs of cigarettes smoked per day multiplied by the number of smoking years reported at questionnaire completion and (iii) smoking duration defined as the number of years smoking cigarettes reported at questionnaire completion. To determine whether the risk of cancer among former smokers was associated with the timing of cessation, a variable to describe smoking cessation was also derived (categories: never-smoker, current smoker, quit <5 years ago, quit 5–10 years ago and quit >10 years ago), and its potential association with pancreatic cancer was investigated.
DNA collection and genotyping
Peripheral blood specimens were obtained, and DNA was extracted from lymphocytes using either phenol–cholorform extraction or spin columns (Qiagen, Valencia, CA) and stored at 4°C. Sequence variants in genes encoding phase I and phase II enzymes known to be involved in smoking carcinogen metabolism were identified through literature searches and the National Center for Biotechnology Information single nucleotide polymorphism database (37). Variants with a published minor allele frequency >5% were selected for analysis, with preference given to sequence variants with potential impact on enzyme activity. Genetic variants that were tested in the present study are listed in Table III. Genotyping of variants was performed in three multiplex PCR using the Sequenom iPLEX genotyping assay and MassARRAY MALDI-TOF Mass Spectrometry system (Sequenom, San Diego, CA). Samples were randomly aliquoted onto 384-well plates and blinded by disease status. All plates included positive and negative controls for each variant as well as a 10% repeat samples for quality control. Genotypes were analyzed using the MassARRAY Workstation 3.0 software and confirmed by visual assessment of the data.
Table III.
Distribution of pancreatic cancer cases, controls and ASOR estimates for selected genetic polymorphisms in carcinogen-metabolizing enzymes
| Genetic polymorphism | Cases (n = 455) |
Controls (n = 893) |
ASOR | 95% CI | ||
| No.a | % | No.a | % | |||
| AHR-1661G>A (rs2066853) | ||||||
| GG | 354 | 79 | 694 | 78 | 1.00 | |
| GA | 82 | 18 | 182 | 21 | 0.91 | 0.68, 1.22 |
| AA | 11 | 3 | 11 | 1 | 1.92 | 0.82, 4.52 |
| COMT-472G>A (rs4680) | ||||||
| AA | 118 | 26 | 234 | 26 | 1.00 | |
| GA | 217 | 48 | 435 | 49 | 0.95 | 0.72, 1.26 |
| GG | 113 | 25 | 219 | 25 | 1.01 | 0.73, 1.39 |
| CYP1A1-2455A>G (rs1048943) | ||||||
| AA | 407 | 90 | 808 | 92 | 1.00 | |
| GA | 42 | 9 | 72 | 8 | 1.08 | 0.71, 1.63 |
| GG | 2 | 1 | 3 | 1 | 1.42 | 0.24, 8.60 |
| CYP1A1-3801T>C (rs4646903) | ||||||
| TT | 341 | 76 | 683 | 77 | 1.00 | |
| TC | 98 | 22 | 181 | 21 | 1.05 | 0.79, 1.39 |
| CC | 10 | 2 | 18 | 2 | 1.03 | 0.45, 2.32 |
| CYP1A1-3801T>C (rs4646903) | ||||||
| TT | 341 | 76 | 683 | 77 | 1.00 | |
| TC + CC | 108 | 24 | 199 | 23 | 1.04 | 0.79, 1.38 |
| CYP1A1-2453C>A (rs1799814) | ||||||
| CC | 419 | 93 | 820 | 92 | 1.00 | |
| CA | 28 | 6 | 67 | 8 | 0.83 | 0.52, 1.32 |
| AA | 2 | 1 | 1 | 1 | 4.20 | 0.38, 46.73 |
| CYP1A2-162A>C (rs762551) | ||||||
| AA | 228 | 51 | 454 | 52 | 1.00 | |
| CA | 176 | 39 | 347 | 39 | 1.03 | 0.80, 1.31 |
| CC | 43 | 10 | 79 | 9 | 1.11 | 0.74, 1.68 |
| CYP1A2-162A>C (rs762551) | ||||||
| AA | 228 | 51 | 454 | 52 | 1.00 | |
| CA + CC | 219 | 49 | 426 | 48 | 1.04 | 0.83, 1.32 |
| CYP1A1 combined variant (derived)b | ||||||
| Wild-type | 340 | 76 | 681 | 78 | 1.00 | |
| Increased activity | 106 | 24 | 194 | 22 | 1.05 | 0.80, 1.39 |
| CYP1B1-142C>G (rs10012) | ||||||
| CC | 218 | 48 | 445 | 51 | 1.00 | |
| CG | 192 | 43 | 356 | 41 | 1.10 | 0.86, 1.40 |
| GG | 41 | 9 | 78 | 9 | 1.10 | 0.73, 1.67 |
| CYP1B1-4326G>C (rs1056836) | ||||||
| CC | 153 | 34 | 288 | 32 | 1.00 | |
| CG | 207 | 46 | 422 | 48 | 0.97 | 0.75, 1.27 |
| GG | 89 | 20 | 177 | 20 | 0.98 | 0.70, 1.36 |
| CYP1B1-4390A>G (rs1800440) | ||||||
| AA | 330 | 73 | 620 | 70 | 1.00 | |
| GA | 115 | 26 | 230 | 26 | 0.88 | 0.67, 1.15 |
| GG | 6 | 1 | 33 | 4 | 0.36 | 0.15, 0.86 |
| CYP1B1-4390A>G (rs1800440) | ||||||
| AA + GA | 445 | 99 | 850 | 96 | 1.00 | |
| GG | 6 | 1 | 33 | 4 | 0.37 | 0.15, 0.89 |
| CYP1B1 combined variant (derived)c | ||||||
| Wild-type | 138 | 31 | 258 | 29 | 1.00 | |
| Increased activity | 308 | 69 | 616 | 70 | 0.94 | 0.73, 1.20 |
| CYP2C9-1075A>C (rs1057910) | ||||||
| AA | 378 | 85 | 771 | 87 | 1.00 | |
| CA | 67 | 15 | 105 | 12 | 1.34 | 0.96, 1.87 |
| CC | 1 | 1 | 7 | 1 | 0.29 | 0.04, 2.41 |
| CYP2E1-1293G>C (rs3813867) | ||||||
| GG | 420 | 94 | 833 | 94 | 1.00 | |
| CG | 29 | 6 | 52 | 6 | 1.07 | 0.66, 1.73 |
| CC | 0 | 0 | 1 | 1 | — | — |
| CYP2E1-1293G>C (rs3813867) | ||||||
| GG | 420 | 94 | 833 | 94 | 1.00 | |
| CG + CC | 29 | 6 | 53 | 6 | 1.05 | 0.65, 1.70 |
| CYP2E1-7362T>A (rs6413432) | ||||||
| TT | 368 | 82 | 728 | 82 | 1.00 | |
| AT | 79 | 18 | 156 | 18 | 1.02 | 0.75, 1.38 |
| AA | 1 | 1 | 6 | 1 | 0.35 | 0.04, 2.91 |
| GSTM3-del-AGG (rs1799735) | ||||||
| AGG | 322 | 72 | 630 | 71 | 1.00 | |
| AGG/DEL | 120 | 27 | 235 | 27 | 0.95 | 0.73, 1.24 |
| DEL | 6 | 1 | 22 | 2 | 0.46 | 0.17, 1.22 |
| GSTP1-492A>G (rs1695) | ||||||
| AA | 194 | 43 | 375 | 42 | 1.00 | |
| AG | 210 | 47 | 403 | 45 | 0.99 | 0.77, 1.26 |
| GG | 45 | 10 | 109 | 12 | 0.80 | 0.54, 1.18 |
| GSTP1-370 C>T (rs1138272) | ||||||
| CC | 371 | 83 | 736 | 83 | 1.00 | |
| TC | 72 | 16 | 148 | 17 | 0.99 | 0.73, 1.36 |
| TT | 4 | 1 | 6 | 1 | 1.47 | 0.41, 5.27 |
| EPHX1-612T>C (rs1051740) | ||||||
| TT | 217 | 48 | 432 | 49 | 1.00 | |
| CT | 192 | 43 | 374 | 42 | 1.07 | 0.84, 1.36 |
| CC | 42 | 9 | 78 | 9 | 1.00 | 0.65, 1.53 |
| EPHX1-415A>G (rs2234922) | ||||||
| AA | 296 | 66 | 566 | 64 | 1.00 | |
| GA | 135 | 30 | 280 | 32 | 0.90 | 0.69, 1.16 |
| GG | 20 | 4 | 36 | 4 | 1.13 | 0.64, 1.99 |
| NAT1-445G>A (rs4987076) | ||||||
| GG | 420 | 94 | 853 | 96 | 1.00 | |
| AG | 27 | 6 | 36 | 4 | 1.50 | 0.88, 2.55 |
| AA | 1 | 1 | 0 | 0 | — | — |
| NAT1-445G>A (rs4987076) | ||||||
| GG | 420 | 94 | 853 | 96 | 1.00 | |
| AG + AA | 28 | 6 | 36 | 4 | 1.56 | 0.93, 2.64 |
| NAT1-459G>A (rs4986990) | ||||||
| GG | 410 | 94 | 847 | 96 | 1.00 | |
| GA | 26 | 6 | 36 | 4 | 1.47 | 0.86, 2.51 |
| AA | 1 | 1 | 0 | 0 | — | — |
| NAT1-459G>A (rs4986990) | ||||||
| GG | 410 | 94 | 847 | 96 | 1.00 | |
| GA + AA | 27 | 6 | 36 | 4 | 1.53 | 0.90, 2.60 |
| NAT1-640T>G (rs4986783) | ||||||
| TT | 342 | 93 | 841 | 96 | 1.00 | |
| GT | 23 | 6 | 34 | 4 | 1.66 | 0.94, 2.93 |
| GG | 1 | 1 | 0 | 0 | — | — |
| NAT1-640T>G (rs4986783) | ||||||
| TT | 342 | 93 | 841 | 96 | 1.00 | |
| GT + GG | 24 | 7 | 34 | 4 | 1.75 | 1.00, 3.05 |
| NAT2-341T>C (rs1801280) | ||||||
| TT | 279 | 97 | 836 | 836 | 1.00 | |
| TC | 6 | 2 | 16 | 2 | 1.22 | 0.47, 3.16 |
| CC | 2 | 1 | 9 | 1 | 0.75 | 0.16, 3.51 |
| NAT2-590G>A (rs1799930) | ||||||
| GG | 236 | 52 | 450 | 51 | 1.00 | |
| GA | 174 | 39 | 360 | 41 | 0.93 | 0.73, 1.19 |
| AA | 41 | 9 | 71 | 8 | 1.11 | 0.73, 1.70 |
| NAT2-803 A>G (rs1208) | ||||||
| AA | 166 | 37 | 296 | 34 | 1.00 | |
| AG | 198 | 44 | 413 | 47 | 0.84 | 0.65, 1.09 |
| GG | 87 | 19 | 172 | 20 | 0.94 | 0.68, 1.30 |
| NAT2-803 A>G (rs1208) | ||||||
| AG + GG | 285 | 63 | 585 | 66 | 1.00 | |
| AA | 166 | 37 | 296 | 34 | 1.15 | 0.90, 1.47 |
| NAT2-857G>A (rs1799931) | ||||||
| GG | 415 | 93 | 849 | 95 | 1.00 | |
| GA | 31 | 7 | 40 | 4 | 1.58 | 0.97, 2.58 |
| AA | 1 | 1 | 1 | 1 | 2.23 | 0.14, 36.23 |
| NAT2-857G>A (rs1799931) | ||||||
| GG | 415 | 93 | 849 | 95 | 1.00 | |
| GA + AA | 32 | 7 | 41 | 5 | 1.60 | 0.99, 2.59 |
| UGT1A7-387T>G (rs17868323) | ||||||
| GG | 166 | 37 | 349 | 40 | 1.00 | |
| GT | 207 | 47 | 400 | 46 | 1.07 | 0.83, 1.38 |
| TT | 72 | 16 | 122 | 14 | 1.24 | 0.87, 1.76 |
| UGT1A7-622 T>C (rs11692021) | ||||||
| TT | 187 | 42 | 321 | 36 | 1.00 | |
| CT | 203 | 45 | 428 | 49 | 0.77 | 0.60, 0.99 |
| CC | 57 | 13 | 133 | 15 | 0.73 | 0.51, 1.05 |
| UGT1A7-622 T>C (rs11692021) | ||||||
| TT | 187 | 42 | 321 | 36 | 1.00 | |
| CT + CC | 260 | 58 | 561 | 64 | 0.76 | 0.60, 0.97 |
| UGT1A7 combined variant (derived)d | ||||||
| Wild-type | 185 | 42 | 321 | 37 | 1.00 | |
| Slightly reduced activity | 201 | 45 | 419 | 48 | 0.79 | 0.62, 1.02 |
| Very reduced activity | 57 | 13 | 127 | 15 | 0.78 | 0.54, 1.12 |
| GSTT1 (rs4630) | ||||||
| TT | 305 | 86 | 681 | 87 | 1.00 | |
| TC | 21 | 6 | 49 | 6 | 0.97 | 0.57, 1.67 |
| CC | 29 | 8 | 54 | 7 | 1.23 | 0.76, 1.98 |
| GSTM1 (rs737497) | ||||||
| AA | 239 | 73 | 642 | 78 | 1.00 | |
| AG | 17 | 5 | 42 | 5 | 1.10 | 0.60, 1.99 |
| GG | 73 | 22 | 143 | 17 | 1.41 | 1.02, 1.95 |
| GSTM1 genee | ||||||
| Presence | 827 | 93 | 329 | 72 | 1.00 | |
| Deletion | 66 | 7 | 126 | 28 | 4.89 | 3.52, 6.79 |
| GSTT1 genef | ||||||
| Presence | 406 | 89 | 868 | 97 | 1.00 | |
| Deletion | 49 | 11 | 25 | 3 | 4.41 | 2.67, 7.29 |
Bold represents statistically significant findings.
Numbers may not add to total due to missing values.
CYP 1A1 wild-type: 2455AA/AG + 3801TT or 2455GG + 3801TC/CC; CYP 1A1 increased activity: 2455AA/AG + 3801TC/CC.
CYB 1B1 wild-type: 142CC+4326CC/CG+4390AA/AG or 142CG+4326CC+4390AA or 142GG+4326CC+4390AG; CYP1B1 increased activity: all other combinations.
UGT1A7 wild-type: 387TT + 622TT/TC or 387TG/GG + 622TT; UGT1A7 slightly reduced activity: 387TG/GG + 622TC; UGT1A7 very reduced activity: 387GG + 622CC.
Deletion: null genotype for rs737497; Presence: genotype for rs737497.
Deletion: null genotypes for 407 (rs2266636), 558 (rs2266637), rs2234735, rs4630; Presence: all other combinations.
Statistical analyses
Multivariate unconditional logistic regression analysis was conducted to obtain age- and sex-adjusted odds ratio (ASOR) and multivariate-adjusted odds ratio (MVOR) estimates for the association between cigarette smoking (defined as smoking status, pack-years and duration) and pancreatic cancer. To identify possible confounding variables that should be adjusted for in the multivariate models, seven risk factors for pancreatic cancer (ethnicity, body mass index, alcohol consumption, personal history of diabetes mellitus diagnosis, aspirin use, non-steroidal anti-inflammatory drug use and family history of pancreatic cancer) were evaluated as potential confounders separately for each of the three types of cigarette smoking exposure variables investigated and applying the 10% change in the odds ratio (OR) method (38). Only potential confounders that changed the ASOR by >10% were deemed to be confounders and were subsequently adjusted for in multivariate models in order to build the most parsimonious multivariate models.
The possibility of effect modification between sex and cigarette smoking and polymorphism of each genotype investigated and cigarette smoking were investigated by conducting stratified analyses and using the likelihood ratio statistic comparing models with and without the interaction term at a statistically significant level set at P <0.05.
Statistical analyses were conducted using SAS version 9.0 (SAS Institute, Cary, NC). Deviation from Hardy–Weinberg equilibrium (HWE) was tested for each variant in control subjects using HWE test in the genetics package (39) of R (40) with the statistical significance level set at P <0.05.
Ethics approval
Ethics approval for this project was obtained from the research ethics board of University Health Network and Mount Sinai Hospital, Toronto, Ontario, Canada.
Results
In total, 455 pancreatic cancer cases and 893 controls had complete epidemiologic and genotyping data. The median age of diagnosis for cases was 65 years (range: 20–89 years) and 54% were male. The median age of questionnaire completion (recruitment) for controls was 64 years (range: 29–79 years) and 53% were male.
Table I describes the distribution of cases, controls and ASOR estimates for subject characteristics and established pancreatic cancer risk factors. Non-Caucasian ethnicity, body mass index ≥30, personal history of diabetes mellitus, aspirin use, non-steroidal anti-inflammatory drug use and a first- or second-degree family history of pancreatic cancer were all significantly associated with increased pancreatic cancer risk. In contrast, alcohol consumption was associated with reduced pancreatic cancer risk. All variables were assessed for potential confounding of the smoking–pancreatic cancer association, and only ethnicity changed the ASORs >10%. As a result, ethnicity was added (to age and sex) to all multivariate models.
Table I.
Distribution of pancreatic cancer cases, controls and ASOR estimates for participant characteristics, medical history and established pancreatic cancer risk factors
| Variables | Cases (n = 455) |
Controls (n = 893) |
ASOR | 95% CI | ||
| No.a | % | No.a | % | |||
| Age (years)b | ||||||
| <50 | 37 | 8 | 58 | 6 | ||
| 50–54 | 51 | 11 | 107 | 12 | ||
| 55–59 | 65 | 14 | 121 | 14 | ||
| 60–64 | 70 | 15 | 161 | 18 | ||
| 65–69 | 93 | 20 | 203 | 23 | ||
| 70–74 | 63 | 14 | 145 | 16 | ||
| 75–79 | 53 | 12 | 98 | 11 | ||
| ≥80 | 23 | 5 | 0 | 0 | — | |
| Sex | ||||||
| Male | 245 | 54 | 470 | 53 | ||
| Female | 210 | 46 | 423 | 47 | — | |
| Ethnicity | ||||||
| Caucasian | 384 | 85 | 853 | 96 | 1.00 | |
| Non-Caucasian | 68 | 15 | 40 | 4 | 3.74 | 2.45, 5.66 |
| BMI (kg/m2)c | ||||||
| <25 | 166 | 37 | 346 | 41 | 1.00 | |
| 25–29.9 | 186 | 42 | 349 | 41 | 1.15 | 0.88, 1.50 |
| ≥30 | 96 | 21 | 157 | 18 | 1.39 | 1.01, 1.92 |
| Alcohol consumption (number of drinks per week)d | ||||||
| <1 | 224 | 49 | 336 | 39 | 1.00 | |
| 1–6 | 104 | 23 | 257 | 30 | 0.59 | 0.44, 0.79 |
| ≥7 | 125 | 28 | 270 | 31 | 0.67 | 0.50, 0.89 |
| Diabetes mellituse | ||||||
| No | 383 | 85 | 830 | 94 | 1.00 | |
| Yes | 69 | 15 | 53 | 6 | 2.80 | 1.90, 4.14 |
| Aspirin usef | ||||||
| No | 287 | 64 | 657 | 75 | 1.00 | |
| Yes | 162 | 36 | 221 | 25 | 1.72 | 1.32, 2.23 |
| Non-steroidal anti-inflammatory drug useg | ||||||
| No | 386 | 87 | 812 | 93 | 1.00 | |
| Yes | 59 | 13 | 64 | 7 | 2.05 | 1.40, 3.00 |
| First- or second-degree family history of pancreatic cancer | ||||||
| No | 393 | 87 | 864 | 97 | 1.00 | |
| Yes | 59 | 13 | 26 | 3 | 4.71 | 2.90, 7.66 |
Bold represents statistically significant findings. BMI, body mass index.
Numbers may not add to total due to missing values.
Age at diagnosis for cases and age at questionnaire completion for controls.
Calculated using usual weight and height for cases and weight and height 2 years prior to questionnaire completion for controls.
Total number of alcoholic beverages (beer, wine and liquor) consumed ∼1 year ago for cases and total number of alcoholic beverages (beer, wine and liquor) consumed in 20, 30/40 or 50 s (age periods closest to age of questionnaire completion) for controls.
Diagnosed with diabetes mellitus 2 years prior to pancreatic cancer diagnosis for cases and 2 years prior to questionnaire completion for controls.
Regular use (at least once per week for at least 1 year) of aspirin/acetylsalicylic acid 1 year prior to questionnaire completion for cases and regular use (at least twice a week for more than a month) of aspirin about 2 years prior to questionnaire completion for controls.
Regular use (at least once per week for at least 1 year) of Motrin, ibuprofen or other non-steroidal anti-inflammatory drugs 1 year prior to questionnaire completion for cases and regular use (at least twice a week for more than a month) of ibuprofen-based medications about 2 years prior to questionnaire completion for controls.
Table II shows the distribution of cases, controls, ASOR and MVOR estimates for cigarette smoking. Current smoking status was associated with significantly increased pancreatic cancer risk [MVOR = 2.29, 95% confidence interval (95% CI): 1.62, 3.22], as were 10–27 pack-years (MVOR = 1.57, 95% CI: 1.13, 2.18) and >27 pack-years (MVOR = 1.77, 95% CI: 1.27, 2.46). Longer durations of smoking were also associated with increased risk (19–32 years: MVOR = 1.46, 95% CI: 1.05, 2.05; >32 years: MVOR = 1.78, 95% CI: 1.30, 2.45). Among former smokers, no statistically significant trend was observed between strata of smoking cessation (quit <5 years ago, quit 5–10 years ago and quit >10 years ago) compared with either current or non-smokers and pancreatic cancer risk (results not shown). Sex stratified analyses and the likelihood ratio statistic also did not show a significant interaction between all three smoking variables and sex in the association between smoking and pancreatic cancer (results not shown).
Table II.
Distribution of pancreatic cancer cases, controls, ASOR and MVOR estimates for smoking characteristics
| Smoking variables | Cases (n = 455) |
Controls (n = 893) |
ASOR | 95% CI | MVORb | 95% CI | ||
| No.a | % | No.a | % | |||||
| Smoking statusc | ||||||||
| Never-smoker | 186 | 43 | 387 | 44 | 1.00 | 1.00 | ||
| Former smoker | 157 | 36 | 389 | 44 | 0.87 | 0.67, 1.14 | 1.02 | 0.77, 1.35 |
| Current smoker | 95 | 22 | 111 | 13 | 1.93 | 1.39, 2.69 | 2.29 | 1.62, 3.22 |
| Smoking pack-yearsd | ||||||||
| Never-smoker | 186 | 41 | 387 | 46 | 1.00 | 1.00 | ||
| ≤9 | 71 | 16 | 154 | 18 | 1.01 | 0.72, 1.41 | 1.14 | 0.81, 1.62 |
| 10–27 | 97 | 22 | 158 | 19 | 1.32 | 0.96, 1.82 | 1.57 | 1.13, 2.18 |
| >27 | 97 | 22 | 147 | 17 | 1.52 | 1.10, 2.09 | 1.77 | 1.27, 2.46 |
| Smoking duration (years)e | ||||||||
| Never-smoker | 186 | 41 | 387 | 45 | 1.00 | 1.00 | ||
| ≤18 | 71 | 16 | 170 | 20 | 0.89 | 0.64, 1.25 | 1.05 | 0.74, 1.49 |
| 19–32 | 90 | 20 | 154 | 18 | 1.23 | 0.89, 1.71 | 1.46 | 1.05, 2.05 |
| >32 | 107 | 24 | 157 | 18 | 1.57 | 1.15, 2.14 | 1.78 | 1.30, 2.45 |
Bold represents statistically significant findings.
Numbers may not add to total due to missing values.
Adjusted for age, sex and ethnicity.
Smoking status 2 years prior to pancreatic cancer diagnosis for cases and 2 years prior to questionnaire completion for controls.
Pack-years smoked up to questionnaire completion for cases and controls. Categories based on tertiles of controls who were ever-smokers.
Years smoked up to questionnaire completion for cases and controls. Categories based on tertiles of controls who were ever-smokers.
Analysis of genotypes in the control population demonstrated that NAT2 341 T>C (P = 5.61 × 10−13) deviated from HWE, whereas all other genetic variants described in this study were within HWE (P > 0.05) (results not shown).
Table III illustrates genotype frequencies for cases and controls, and ASOR estimates for the association between each genotype and pancreatic cancer risk. The CYP1B1-4390-GG genotype was associated with reduced risk for pancreatic cancer (ASOR = 0.36, 95% CI: 0.15, 0.86) as was the UGT1A7-622-CT (ASOR = 0.77, 95% CI: 0.60, 0.99) genotype. The NAT1-640-GT and GG genotypes (combined) were associated with increased risk for pancreatic cancer (ASOR = 1.75, 95% CI: 1.00, 3.05) as were the GSTM1-GG genotype (ASOR = 1.41, 95% CI: 1.02, 1.95), GSTM1 gene deletion (ASOR = 4.89, 95% CI: 3.52, 6.79) and GSTT1 gene deletion (ASOR = 4.41, 95% CI: 2.67, 7.29).
Potential interactions between smoking and all genetic variants listed in Table III were investigated, and only the statistically significant and marginally significant results are shown in Table IV. Significant interactions were observed between pack-years and EPHX1-415 (P = 0.04) and smoking status and NAT2-857 (P = 0.03). Compared with never-smokers, those with >27 pack-years and the EPHX1-415-AA genotype (MVOR = 2.07, 95% CI: 1.37, 3.13) and those with 10–27 pack-years and the EPHX1-145-GG genotype (MVOR = 17.39, 95% CI: 1.92, 157.35) were at significantly increased risk for pancreatic cancer. Compared with never-smokers, current smokers with the NAT2-857-GG genotype were at significantly increased risk for pancreatic cancer (MVOR = 2.27, 95% CI: 1.59, 3.24). Marginally significant interactions were observed between pack-years and CYP1B1-142 (P = 0.05) and smoking duration and GSTM3-del-AGG (P = 0.06). Compared with never-smokers, those with 10–27 pack-years and CYP1B1-142-CG genotype (MVOR = 3.07, 95% CI: 1.79, 5.26) as well as those with >27 pack-years and the CYP1B1-142-CG genotype (MVOR = 2.60, 95% CI: 1.55, 4.35) were observed to be at significantly increased risk. Compared with never-smokers, those with the GSTM3-del-AGG/AGG genotype and >32 years of smoking were at significantly increased risk for pancreatic cancer (MVOR = 2.33, 95% CI: 1.61, 3.39).
Table IV.
MVOR estimates and 95% CI for smoking and pancreatic cancer risk, stratified by carcinogen-metabolizing genotypes
| Smoking variable | Genotype |
df | PInteractionb | |||||
| MVORa | 95% CI | MVORa | 95% CI | MVORa | 95% CI | |||
| CYP1B1-142C>G (rs10012) | ||||||||
| CC (n = 663) |
CG (n = 548) |
GG (n = 119) |
||||||
| Smoking pack-years | ||||||||
| Never-smoker | 1.00 | 1.00 | 1.00 | |||||
| ≤9 | 1.06 | 0.63, 1.76 | 1.33 | 0.76, 2.35 | 0.98 | 0.34, 2.82 | ||
| 10–27 | 1.10 | 0.69, 1.76 | 3.07 | 1.79, 5.26 | 0.76 | 0.25, 2.35 | ||
| >27 | 1.49 | 0.93, 2.40 | 2.60 | 1.55, 4.35 | 0.80 | 0.21, 3.03 | 6 | 0.05 |
| GSTM3-del-AGG (rs1799735) | ||||||||
| AGG/AGG (n = 952) | AGG/DEL (n = 355) | DEL/DEL (n = 28) | ||||||
| Smoking duration (years) | ||||||||
| Never-smoker | 1.00 | 1.00 | 1.00 | |||||
| ≤18 | 1.19 | 0.79, 1.81 | 0.92 | 0.47, 1.82 | <0.01 | <0.01, >999.99 | ||
| 19–32 | 1.46 | 0.98, 2.19 | 1.59 | 0.82, 3.08 | <0.01 | <0.01, >999.99 | ||
| >32 | 2.33 | 1.61, 3.39 | 0.89 | 0.45, 1.78 | 0.54 | <0.01, >999.99 | 6 | 0.08 |
| AGG/AGG (n = 952) | AGG/DEL + DEL/DEL (n = 383) | |||||||
| Smoking duration (years) | ||||||||
| Never-smoker | 1.00 | 1.00 | ||||||
| ≤18 | 1.19 | 0.79, 1.81 | 0.83 | 0.43, 1.61 | ||||
| 19–32 | 1.46 | 0.98, 2.19 | 1.41 | 0.74, 2.69 | ||||
| >32 | 2.33 | 1.61, 3.39 | 0.84 | 0.43, 1.64 | 3 | 0.06 | ||
| EPHX1-415A>G (rs2234922) | ||||||||
| AA (n = 862) | GA (n = 415) | GG (n = 56) | ||||||
| Smoking pack-years | ||||||||
| Never-smoker | 1.00 | 1.00 | 1.00 | |||||
| ≤9 | 1.04 | 0.67, 1.62 | 1.60 | 0.82, 3.11 | 0.43 | 0.06, 3.07 | ||
| 10–27 | 1.42 | 0.93, 2.15 | 1.48 | 0.81, 2.68 | 17.39 | 1.92, 157.35 | ||
| >27 | 2.07 | 1.37, 3.13 | 1.28 | 0.69, 2.35 | 2.30 | 0.30, 17.96 | 6 | 0.04 |
| NAT2-857G>A (rs1799931) | ||||||||
| GG (n = 1264) | GA (n = 71) | AA (n = 2) | ||||||
| Smoking status | ||||||||
| Never-smoker | 1.00 | 1.00 | 1.00 | |||||
| Former smoker | 1.10 | 0.82, 1.47 | 0.33 | 0.06, 1.72 | — | — | ||
| Current smoker | 2.27 | 1.59, 3.24 | 4.08 | 0.65, 25.60 | — | — | 2 | 0.03 |
| GG (n = 1264) | GA + AA (n = 73) | |||||||
| Smoking status | ||||||||
| Never-smoker | 1.00 | 1.00 | ||||||
| Former smoker | 1.10 | 0.82, 1.47 | 0.33 | 0.07, 1.53 | ||||
| Current smoker | 2.27 | 1.59, 3.24 | 4.17 | 0.66, 26.40 | 2 | 0.05 | ||
Bold represents statistically significant findings.
df, degrees of freedom.
Adjusted for age, sex and ethnicity.
P-value for interaction.
Discussion
A limited number of studies have examined the potential interactions between genetic variants, cigarette smoking and pancreatic cancer risk (19,25–28), and as concluded by a recent review (41), the results of these studies have been difficult to interpret due to their small sample sizes. To address this limitation, we conducted a large population-based study with 1348 participants—to our knowledge, the largest population-based pancreatic cancer case–control study to date. We used three descriptors of smoking (smoking status, pack-years and smoking duration) in an attempt to be comprehensive in capturing the smoking behavior of participants.
Cigarette smoking and pancreatic cancer risk
Consistent with the literature, the present study confirms that cigarette smoking is an established risk factor for pancreatic cancer (10,11). Current smokers were at the highest risk for pancreatic cancer with an estimated doubling of risk. We did not observe an obvious cessation year cut-off that conferred significantly reduced risk. However, dose-response relationships were observed with increasing duration of smoking and pack-years conferring increased pancreatic cancer risk.
Association between selected genetic variants and pancreatic cancer risk—comparison to previous studies
Numerous studies have investigated the potential roles of polymorphisms of carcinogen metabolism genes in lung cancer risk, where smoking is the dominant etiologic risk factor, with mixed results. Meta-analyses have failed to demonstrate an association between the NAT2 slow acetylator genotype (42), GSTT1 deletion (43) and lung cancer, whereas increased cancer risk were observed with the GSTM1-null genotype (OR = 1.22, 95% CI: 1.14, 1.30) (44) and GSTP1-313-GG or AG genotypes (OR = 1.11, 95% CI: 1.03, 1.21) (45). Subgroup analysis among Caucasians demonstrated a potential interaction between the GSTP1-313 variant and pack-years history of smoking (45). From this data in lung and other cancers, we therefore hypothesized that variants in carcinogen-metabolizing genes may be associated with smoking and pancreatic cancer risk.
CYP1B1-4390-A>G leads to Asn453Ser substitution (46,47). We found a reduced risk for pancreatic cancer associated with the CYP1B1-4390-GG genotype (ASOR = 0.36, 95% CI: 0.15, 0.86). In contrast, Crous-Bou et al. (48) studied 129 pancreatic cancer cases and 19 controls and reported no difference in CYP1B1-4390 genotype frequencies between their cases and controls. Similarly, a recent study with 285 cases and 469 controls reported no association between polymorphisms in the CYP1B1 gene and pancreatic cancer (49). However, given the paucity of studies to investigate CYP1B1 and pancreatic cancer and the small sample sizes, it is premature to draw conclusions regarding the potential role of CYP1B1 in pancreatic cancer development.
Following the observation that individuals with 4-aminobiphenyl–DNA adduct in their pancreatic tissues have the slow NAT1-1095-CC genotype, it has been proposed that NAT1 may be involved in pancreatic cancer development through the detoxification pathway of 4-aminobiphenyl (14). We observed a marginally significant increased risk for pancreatic cancer among individuals with the G allele of NAT1-640 compared with those without (ASOR = 1.75, 95% CI: 1.00, 3.05). Previous studies looking at NAT1 polymorphisms have reported a trend (26) or statistically significant (25,27) increase in pancreatic cancer risk associated with the *10 (1088T>A, 1095C>A) and *11 (V149I, T153T, S214A) alleles which code for the rapid phenotype.
Those with the UGT1A7-622-CT and CC genotypes were at reduced risk for pancreatic cancer (ASOR = 0.77, 95% CI: 0.60, 0.99 and 0.73, 95% CI: 0.51, 1.05, respectively). These findings indicate a possible dominant mode of inheritance for this protective allele with a reduced risk observed for those with the C allele for UGT1A7-622 (CC/CT genotypes ASOR = 0.76, 95% CI: 0.60, 0.97). Note that this is in contrast to previous findings by Ockenga et al. (28), who report a non-significant increased risk associated with UGT1A7*4 allele (OR = 1.25, 95% CI: 0.98, 2.71), which corresponds to the UGT1A7-622 CC genotype (50) in our study. Since UGT1A7-622 CC [UGT1A7*4, encoding W208R (50)] has been associated with slower glucuronidation compared with wild-type (50), it is uncertain as to how the C allele could be protective against pancreatic cancer, although slower glucuronidation may lead to detoxification of active metabolites by other phase II enzymes. Previous studies investigating polymorphisms of the UGT1A7 gene and pancreatic cancer risk have generally been mixed with two previous studies reporting no significant association (29,30) and one study (28) reporting increased risk associated with the UGT1A7*3 allele, which results in N129K, R131K and W208R amino acid changes.
Contrary to previous reports that concluded that GSTM1 genotype is not associated with pancreatic cancer risk (19–21,23), we observed a statistically non-significant increased risk for pancreatic cancer among those with the GSTM1-AG genotype (ASOR = 1.10, 95% CI: 0.60, 1.99) and a significantly increased risk for those with the GSTM1-GG genotype (ASOR = 1.41, 95% CI: 1.02, 1.95) and GSTM1 gene deletion (ASOR = 4.89, 95% CI: 3.52, 6.79). There appears to be a relationship between gene copy number and enzyme activity where individuals with homozygous GSTM1 gene deletion have the lowest enzyme activity, followed by heterozygotes and homozygotes. Individuals with both copies of the gene have the highest enzyme activity and are described to have ‘ultrarapid’ activity (51–53). The marked increased pancreatic cancer risk associated with GSTM1 gene deletion that we report is plausible given that those with the null genotype have low enzymatic activity and thereby may have comparatively lower carcinogen metabolism (52,53).
Previous studies observed statistically non-significant increased pancreatic cancer risk associated with GSTT1 gene deletion (19–22), with OR estimates ranging from 1.01 (95% CI: 0.68, 1.49) (21) to 1.9 (95% CI: 0.59, 5.9) (19). We observed a significantly increased risk of pancreatic cancer associated with GSTT1 gene deletion (ASOR = 4.41, 95% CI: 2.67, 7.29). Given the smaller sample sizes of previous studies compared with the present study, it is possible that previous studies had insufficient statistical power to detect a statistically significant increased pancreatic cancer risk associated with GSTT1 gene deletion. Biologically, the marked increase in pancreatic cancer risk associated with GSTT1 gene deletion is plausible as there also appears to be a dose-dependent effect between genotype and enzyme activity for GSTT1 (52) as demonstrated by two studies investigating GSTT1 activity (54,55); namely GSTT1 gene deletion is associated with lowest GSTT1 activity, followed by GSTT1 heterozygotes and then wild-types having the highest activity.
Interaction between genetic variants and cigarette smoking with regards to pancreatic cancer risk
We observed a marginally significant interaction effect of CYP1B1-142 on the association between pack-years and pancreatic cancer risk (P = 0.05), with individuals with the CYP1B1-142-CG genotype having the highest pancreatic cancer risk. To our knowledge, this study is the first to investigate a potential interaction effect of CYP1B1-142 on smoking and pancreatic cancer. We also observed a marginally significant interaction of GSTM3 (P = 0.06), which was notable upon comparing individuals with a GSTM3-deletion allele compared with those without. Our results suggest that the GSTM3-deletion allele may be protective against pancreatic cancer for varying durations of smoking, however given the inconsistency of our findings across the categories of smoking duration, it is difficult to interpret our results. One previous study investigated the role of the EPHX1 gene on pancreatic cancer risk and observed no main effect and also did not observe differences in smoking status among pancreatic cancer cases with different EPHX1 genotypes (56). Although the present study similarly did not observe a main effect of EPHX1 on pancreatic cancer risk, we did observe a significant interaction effect of EPHX1-415 genotype on the association between pack-years and pancreatic cancer (P = 0.04). However due to the lack of a clear pattern in our results, the interpretation of the effect of EPHX1 is not straightforward. Contrary to previous reports (24), we did not observe a main effect of NAT2 genotype on pancreatic cancer, but we do report a significant interaction effect of NAT2-857 on the association between smoking status and pancreatic cancer risk (P = 0.03) that remained marginally significant when individuals with the NAT2-857 A allele were compared with those without (P = 0.05); this appears to be consistent with a previous report that found significant interactions between smoking and NAT2 for men and women (25). Finally, contrary to previous reports, we did not observe a significant interaction between CYP1A2 genotype and smoking (25,27); however, consistent with another study (19), we also did not observe an interaction between CYP1A1 genotype and smoking.
Cigarette smoking, genetic variants and pancreatic cancer mortality
As our study did not collect survivorship data, we were unable to investigate the association of carcinogen metabolism gene variants and cigarette smoking with pancreatic cancer survival. Currently, the literature has limited studies on this topic. Although some studies have reported that cigarette smoking is associated with increased mortality from pancreatic cancer (57,58), to our knowledge, only one study has investigated the association of carcinogen metabolism gene variants with pancreatic cancer mortality. Jiao et al. (21) report that in their case–control study, pancreatic cancer patients receiving 5-fluorouracil treatment and having the GSTP1 Val/Val genotype had significantly longer survival with a median survival time of almost 36 months compared to patients with other GSTP1 genotypes. No associations were observed between GSTM1 and GSTT1 genotypes and pancreatic cancer survival (21). The possible interaction between cigarette smoking and variants of carcinogen metabolism genes with pancreatic cancer mortality is an identifiable area in need of further research.
Strengths and limitations
Survival bias is a concern in all retrospective pancreatic cancer studies since the disease has such poor prognosis. Hence, differences between case participants and non-participants may limit the external validity of our findings. Since stage is not captured for most cases, we cannot assess response rate by stage in our dataset. Response bias is a possibility as our response rates are not ideal with estimates for cases and controls being ∼30% (33) and ∼61% (34), respectively. However, it should be noted that most studies of fatal cancers, such as pancreatic cancer, are plagued with poor response rates [e.g. (35,36)]. Recall bias is also a possibility for self-reported behaviors like smoking; yet, the direction of the bias is unclear. Similarly, smoking behavior may change after cancer diagnosis; to account for this possibility we defined smoking status to the point of 2 years prior to cancer diagnosis and questionnaire completion for cases and controls, respectively, for comparability. As our analyses were limited to testing multiplicative interactions between genetic polymorphisms and smoking, the possibility of additive interactions cannot be excluded. Finally, although HWE was tested for all genetic variants investigated in the present study, one polymorphism that we investigated deviated from HWE (NAT2-341 T>C), and furthermore, the possibility of genotyping errors cannot be excluded.
Despite limitations, our study also has notable strengths. First compared with most previous studies investigating effect modification by carcinogen metabolism polymorphisms, our study had a relatively larger sample size and consequently adequate statistical power to detect associations between genetic polymorphisms and pancreatic cancer that previously may have been undetected. However, it is noted that perhaps our sample size was still inadequate, particularly for analyses involving polymorphisms with low population prevalence as illustrated by the low frequencies observed for certain polymorphisms that we investigated. It is also noted that the investigations of interactions were largely hypothesis generating. Second as our study is population-based, our results may be more generalizable than previous studies with comparable samples sizes but used hospital-based designs (25). Finally, since extensive data on personal health history, family history and diet were collected from participants, we were able to investigate potential confounding by several known pancreatic cancer risk factors and appropriately adjust for those that confounded the association of interest.
Conclusion
We report several findings regarding the interaction between carcinogen metabolism gene variants, cigarette smoking and pancreatic cancer risk. CYP1B1-4390-GG genotype was associated with reduced pancreatic cancer risk. Similarly, the UGT1A7-622-CT genotype was associated with reduced risk, as was the C allele of UGT1A7-622. The NAT1-640 G allele was associated with increased pancreatic cancer risk, as were the GSTM1-GG genotype and the deletion of the GSTM1 gene and GSTT1 gene. Non-significant interactions were observed between CYP1B1-142 genotype and pack-years and GSTM3 deletion allele and smoking duration. Statistically significant interactions were observed between EPHX1-415 genotype and pack-years and NAT2-857 genotype and smoking status.
We conducted one of the largest and most comprehensive case–control studies to date on the main effects and interactions of carcinogen metabolism genes and smoking among 455 pancreatic cancer cases and 893 controls. Despite our modestly large sample size, we were admittedly underpowered for analyses of variants of lower population prevalence, indicating the need of studies with even greater sample sizes in the future. As some of our findings are among the first in the genetic epidemiology of pancreatic cancer, replications of our findings are necessary to confirm our results.
Funding
This work was supported by funding from the Grant Miller Cancer Research Award, Faculty of Medicine, University of Toronto [to S.P.C.]; the National Institutes of Health [grant number R01 CA97075, as part of the Pancreatic Cancer Genetic Epidemiology consortium to S.G.]; the National Cancer Institute, National Institutes of Health [grant number RFA # CA-95-011, the Ontario Registry for Studies of Familial Colorectal Cancer (U01 CA074783)], and through cooperative agreements with members of the Colon Cancer Family Registry and Principal Investigators and the National Cancer Institute of Canada [grant number 13304 to S.G.].
Acknowledgments
The authors would like to thank Cheryl Crozier and Xiangdong Liu at the Analytical Genetics Technology Centre for their assistance with the genotyping experiments; Darshana Daftary and Teresa Selander at the Ontario Familial Colorectal Cancer Registry for their assistance in data collection and sample extraction and Laura Anderson for her assistance in data management and interpretation. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the Colon Cancer Family Registries nor does mention of trade names, commercial products or organizations imply endorsement by the US Government or the Colon Cancer Family Registry.
Conflict of Interest Statement: None declared.
Glossary
Abbreviations
- ASOR
age- and sex-adjusted odds ratio
- CI
confidence interval
- GSTT1
glutathione S-transferase theta 1
- HWE
Hardy–Weinberg equilibrium
- MVOR
multivariate-adjusted odds ratio
- NAT1
N-acetyltransferase 1
- NAT2
N-acetyltransferase 2
- OFCCR
Ontario Familial Colorectal Cancer Registry
- OPCS
Ontario Pancreas Cancer Study
- OR
odds ratio
- UGT1A7
Uridine 5'-diphospho glucuronosyltransferase 1 family, polypeptide A7
References
- 1.Canadian Cancer Society’s Steering Committee. Canadian Cancer Statistics 2010. Toronto: Canadian Cancer Society; 2010. [Google Scholar]
- 2.Canadian Cancer Society’s Steering Committee. Canadian Cancer Statistics 2009. Toronto: Canadian Cancer Society; 2009. [Google Scholar]
- 3.Cleary SP, et al. Prognostic factors in resected pancreatic adenocarcinoma: analysis of actual 5-year survivors. J. Am. Coll. Surg. 2004;198:722–731. doi: 10.1016/j.jamcollsurg.2004.01.008. [DOI] [PubMed] [Google Scholar]
- 4.Olsen GW, et al. A case-control study of pancreatic cancer and cigarettes, alcohol, coffee and diet. Am. J. Public Health. 1989;79:1016–1019. doi: 10.2105/ajph.79.8.1016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Stevens RJ, et al. Factors associated with incident and fatal pancreatic cancer in a cohort of middle-aged women. Int. J. Cancer. 2009;124:2400–2405. doi: 10.1002/ijc.24196. [DOI] [PubMed] [Google Scholar]
- 6.Anderson LN, et al. Lifestyle, dietary, and medical history factors associated with pancreatic cancer risk in Ontario, Canada. Cancer Causes Control. 2009;20:825–834. doi: 10.1007/s10552-009-9303-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Heinen MM, et al. Active and passive smoking and the risk of pancreatic cancer in the Netherlands Cohort Study. Cancer Epidemiol. Biomarkers Prev. 2010;19:1612–1622. doi: 10.1158/1055-9965.EPI-10-0121. [DOI] [PubMed] [Google Scholar]
- 8.Ji BT, et al. Cigarette smoking and alcohol consumption and the risk of pancreatic cancer: a case-control study in Shanghai, China. Cancer Causes Control. 1995;6:369–376. doi: 10.1007/BF00051413. [DOI] [PubMed] [Google Scholar]
- 9.Lynch SM, et al. Cigarette smoking and pancreatic cancer: a pooled analysis from the pancreatic cancer cohort consortium. Am. J. Epidemiol. 2009;170:403–413. doi: 10.1093/aje/kwp134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Iodice S, et al. Tobacco and the risk of pancreatic cancer: a review and meta-analysis. Langenbecks Arch. Surg. 2008;393:535–545. doi: 10.1007/s00423-007-0266-2. [DOI] [PubMed] [Google Scholar]
- 11.La Torre G, et al. Does quality of observational studies affect the results of a meta-analysis?: the case of cigarette smoking and pancreatic cancer. Pancreas. 2009;38:241–247. doi: 10.1097/MPA.0b013e318190d795. [DOI] [PubMed] [Google Scholar]
- 12.Hecht SS. Cigarette smoking: cancer risks, carcinogens, and mechanisms. Langenbecks Arch. Surg. 2006;391:603–613. doi: 10.1007/s00423-006-0111-z. [DOI] [PubMed] [Google Scholar]
- 13.Anderson KE, et al. Metabolic activation of aromatic amines by human pancreas. Carcinogenesis. 1997;18:1085–1092. doi: 10.1093/carcin/18.5.1085. [DOI] [PubMed] [Google Scholar]
- 14.Thompson PA, et al. Comparison of DNA adduct levels associated with exogenous and endogenous exposures in human pancreas in relation to metabolic genotype. Mutat. Res. 1999;424:263–274. doi: 10.1016/s0027-5107(99)00024-x. [DOI] [PubMed] [Google Scholar]
- 15.Fretland AJ, et al. Metabolic activation of 2-hydroxyamino-1-methyl-6-phenylimidazo[4,5-b]pyridine in Syrian hamsters congenic at the N-acetyltransferase 2 (NAT2) locus. Toxicol. Sci. 2003;74:253–259. doi: 10.1093/toxsci/kfg133. [DOI] [PubMed] [Google Scholar]
- 16.Kadlubar FF, et al. Comparison of DNA adduct levels associated with oxidative stress in human pancreas. Mutat. Res. 1998;405:125–133. doi: 10.1016/s0027-5107(98)00129-8. [DOI] [PubMed] [Google Scholar]
- 17.Wetscher GJ, et al. Free radical production in nicotine treated pancreatic tissue. Free Radic. Biol. Med. 1995;18:877–882. doi: 10.1016/0891-5849(94)00221-5. [DOI] [PubMed] [Google Scholar]
- 18.Nebert DW, et al. The role of cytochrome P450 enzymes in endogenous signalling pathways and environmental carcinogenesis. Nat. Rev. Cancer. 2006;6:947–960. doi: 10.1038/nrc2015. [DOI] [PubMed] [Google Scholar]
- 19.Duell EJ, et al. A population-based, case-control study of polymorphisms in carcinogen-metabolizing genes, smoking, and pancreatic adenocarcinoma risk. J. Natl Cancer Inst. 2002;94:297–306. doi: 10.1093/jnci/94.4.297. [DOI] [PubMed] [Google Scholar]
- 20.Liu G, et al. Polymorphisms in GSTM1, GSTT1 and CYP1A1 and risk of pancreatic adenocarcinoma. Br. J. Cancer. 2000;82:1646–1649. doi: 10.1054/bjoc.2000.1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jiao L, et al. Glutathione S-transferase gene polymorphisms and risk and survival of pancreatic cancer. Cancer. 2007;109:840–848. doi: 10.1002/cncr.22468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Vrana D, et al. The association between glutathione S-transferase gene polymorphisms and pancreatic cancer in a central European Slavonic population. Mutat. Res. 2009;680:78–81. doi: 10.1016/j.mrgentox.2009.09.005. [DOI] [PubMed] [Google Scholar]
- 23.Bartsch H, et al. Genetic polymorphism of N-acetyltransferases, glutathione S-transferase M1 and NAD(P)H: quinone oxidoreductase in relation to malignant and benign pancreatic disease risk. Eur. J. Cancer Prev. 1998;7:215–223. doi: 10.1097/00008469-199806000-00006. [DOI] [PubMed] [Google Scholar]
- 24.Ayaz L, et al. The association between N-acetyltransferase 2 gene polymorphisms and pancreatic cancer. Cell Biochem. Funct. 2008;26:329–333. doi: 10.1002/cbf.1445. [DOI] [PubMed] [Google Scholar]
- 25.Suzuki H, et al. Interaction of cytochrome P4501A2, SULT1A1 and NAT gene polymorphisms with smoking and dietary mutagen intake in modification of the risk of pancreatic cancer. Carcinogenesis. 2008;29:1184–1191. doi: 10.1093/carcin/bgn085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Jiao L, et al. Haplotype of N-acetyltransferase 1 and 2 and risk of pancreatic cancer. Cancer Epidemiol. Biomarkers Prev. 2007;16:2379–2386. doi: 10.1158/1055-9965.EPI-06-0992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Li D, et al. Polymorphisms of cytochrome P4501A2 and N-acetyltransferase genes, smoking, and risk of pancreatic cancer. Carcinogenesis. 2006;27:103–111. doi: 10.1093/carcin/bgi171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ockenga J, et al. UDP glucuronosyltransferase (UGT1A7) gene polymorphisms increase the risk of chronic pancreatitis and pancreatic cancer. Gastroenterology. 2003;124:1802–1808. doi: 10.1016/s0016-5085(03)00294-4. [DOI] [PubMed] [Google Scholar]
- 29.Verlaan M, et al. Polymorphisms of UDP-glucuronosyltransferase 1A7 are not involved in pancreatic diseases. J. Med. Genet. 2005;42:e62. doi: 10.1136/jmg.2005.032599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Piepoli A, et al. Lack of association between UGT1A7, UGT1A9, ARP, SPINK1 and CFTR gene polymorphisms and pancreatic cancer in Italian patients. World J. Gastroenterol. 2006;12:6343–6348. doi: 10.3748/wjg.v12.i39.6343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Petersen GM, et al. Pancreatic cancer genetic epidemiology consortium. Cancer Epidemiol. Biomarkers Prev. 2006;15:704–710. doi: 10.1158/1055-9965.EPI-05-0734. [DOI] [PubMed] [Google Scholar]
- 32.Newcomb PA, et al. Colon Cancer Family Registry: an international resource for studies of the genetic epidemiology of colon cancer. Cancer Epidemiol. Biomarkers Prev. 2007;16:2331–2343. doi: 10.1158/1055-9965.EPI-07-0648. [DOI] [PubMed] [Google Scholar]
- 33.Eppel A, et al. Allergies are associated with reduced pancreas cancer risk: a population-based case-control study in Ontario, Canada. Int. J. Cancer. 2007;121:2241–2245. doi: 10.1002/ijc.22884. [DOI] [PubMed] [Google Scholar]
- 34.Cotterchio M, et al. Ontario familial colon cancer registry: methods and first-year response rates. Chronic Dis. Can. 2000;21:81–86. [PubMed] [Google Scholar]
- 35.Ghadirian P, et al. Food habits and pancreatic cancer: a case-control study of the Francophone community in Montreal, Canada. Cancer Epidemiol. Biomarkers Prev. 1995;4:895–899. [PubMed] [Google Scholar]
- 36.Hanley AJ, et al. Physical activity, anthropometric factors and risk of pancreatic cancer: results from the Canadian enhanced cancer surveillance system. Int. J. Cancer. 2001;94:140–147. doi: 10.1002/ijc.1446. [DOI] [PubMed] [Google Scholar]
- 37. National Center for Biotechnology Information Single Nucleotide Polymorphism. http://www.ncbi.nlm.nih.gov/projects/SNP (12 February 2012, date last accessed)
- 38.Greenland S. Modeling and variable selection in epidemiologic analysis. Am. J. Public Health. 1989;79:340–349. doi: 10.2105/ajph.79.3.340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Warnes G, et al. Genetics: Population Genetics. R Package Version 1.3.4. 2008. [Google Scholar]
- 40.R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2008. ISBN 3-900051-07-0 http://www.R-project.org (12 February 2012, date last accessed) [Google Scholar]
- 41.Lin Y, et al. An overview of genetic polymorphisms and pancreatic cancer risk in molecular epidemiologic studies. J. Epidemiol. 2011;21:2–12. doi: 10.2188/jea.JE20100090. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Borlak J, et al. N-acetyltransferase 2 (NAT2) gene polymorphisms in colon and lung cancer patients. BMC Med. Genet. 2006;7:58. doi: 10.1186/1471-2350-7-58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Raimondi S, et al. Meta- and pooled analysis of GSTT1 and lung cancer: a HuGE-GSEC review. Am. J. Epidemiol. 2006;164:1027–1042. doi: 10.1093/aje/kwj321. [DOI] [PubMed] [Google Scholar]
- 44.Carlsten C, et al. Glutathione S-transferase M1 (GSTM1) polymorphisms and lung cancer: a literature-based systematic HuGE review and meta-analysis. Am. J. Epidemiol. 2008;167:759–774. doi: 10.1093/aje/kwm383. [DOI] [PubMed] [Google Scholar]
- 45.Cote ML, et al. Meta- and pooled analysis of GSTP1 polymorphism and lung cancer: a HuGE-GSEC review. Am. J. Epidemiol. 2009;169:802–814. doi: 10.1093/aje/kwn417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Stoilov I, et al. Sequence analysis and homology modeling suggest that primary congenital glaucoma on 2p21 results from mutations disrupting either the hinge region or the conserved core structures of cytochrome P4501B1. Am. J. Hum. Genet. 1998;62:573–584. doi: 10.1086/301764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Bailey LR, et al. Association of cytochrome P450 1B1 (CYP1B1) polymorphism with steroid receptor status in breast cancer. Cancer Res. 1998;58:5038–5041. [PubMed] [Google Scholar]
- 48.Crous-Bou M, et al. CYP1B1 polymorphisms and k-ras mutations in patients with pancreatic ductal adenocarcinoma. Dig. Dis. Sci. 2008;53:1417–1421. doi: 10.1007/s10620-008-0235-9. [DOI] [PubMed] [Google Scholar]
- 49.Vrana D, et al. CYP1B1 gene polymorphism modifies pancreatic cancer risk but not survival. Neoplasma. 2010;57:15–19. doi: 10.4149/neo_2010_01_015. [DOI] [PubMed] [Google Scholar]
- 50.Guillemette C, et al. Structural heterogeneity at the UDP-glucuronosyltransferase 1 locus: functional consequences of three novel missense mutations in the human UGT1A7 gene. Pharmacogenetics. 2000;10:629–644. doi: 10.1097/00008571-200010000-00006. [DOI] [PubMed] [Google Scholar]
- 51.Seidegård J, et al. Hereditary differences in the expression of the human glutathione transferase active on trans-stilbene oxide are due to a gene deletion. Proc. Natl Acad. Sci. USA. 1988;85:7293–7297. doi: 10.1073/pnas.85.19.7293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Ginsberg G, et al. Genetic polymorphism in glutathione transferase (GST): population distribution of GSTM1, T1, and P1 conjugating activity. J. Toxicol. Environ. Health B Crit. Rev. 2009;12:389–439. doi: 10.1080/10937400903158375. [DOI] [PubMed] [Google Scholar]
- 53.McLelland RA, et al. Characterization of a human glutathione S-transfearse μ cluster containing a duplicated GSTM1 gene that causes ultrarapid enzyme activity. Mol. Pharmacol. 1997;52:958–965. doi: 10.1124/mol.52.6.958. [DOI] [PubMed] [Google Scholar]
- 54.Sprenger R, et al. Characterization of the glutathione S-transfaserase GSTT1 deletion: discrimination of all genotypes by polymerase chain reaction indicates a trimodular genotype-phenotype correlation. Pharmacogenetics. 2000;10:557–565. doi: 10.1097/00008571-200008000-00009. [DOI] [PubMed] [Google Scholar]
- 55.Their R, et al. Species differences in the glutathione transferase GSTT1-1 activity towards the model substrates methyl chloride and dichloromethane in liver and kidney. Arch. Toxicol. 1998;72:622–629. doi: 10.1007/s002040050552. [DOI] [PubMed] [Google Scholar]
- 56.Ockenga J, et al. The role of epoxide hydrolase Y113H gene variant in pancreatic diseases. Pancreas. 2009;38:e97–e101. doi: 10.1097/MPA.0b013e31819feeed. [DOI] [PubMed] [Google Scholar]
- 57.Ansary-Moghaddam A, et al. The effect of modifiable risk factors on pancreatic cancer mortality in populations of the Asia-Pacific region. Cancer Epidemiol. Biomarkers Prev. 2006;15:2435–2440. doi: 10.1158/1055-9965.EPI-06-0368. [DOI] [PubMed] [Google Scholar]
- 58.Nakamura K, et al. Cigarette smoking and other lifestyle factors in relation to the risk of pancreatic cancer death: a prospective cohort study in Japan. Jpn. J. Clin. Oncol. 2011;41:225–231. doi: 10.1093/jjco/hyq185. [DOI] [PubMed] [Google Scholar]