Microsomal epoxide hydrolase (mEH) metabolizes polycyclic aromatic hydrocarbons, carcinogens found in cigarette smoke and cooked meat. Polymorphisms in exon 3 and exon 4 of the mEH gene have been found to alter mEH activity. We investigated the association between these polymorphisms and colorectal polyps within the Minnesota Cancer Prevention Research Unit case-control study. Cases were diagnosed with colonoscopically confirmed adenomas (n = 530) or hyperplastic polyps (n = 202); controls (n = 649) were polyp-free at colonoscopy. Smoking history and meat consumption were obtained from self-administered questionnaires before colonoscopy. mEH genotypes were determined by PCR/RFLP or oligonucleotide ligation assay.
The overall risks associated with exon 3 or exon 4 polymorphisms for both adenomas and hyperplastic polyps were not statistically different from 1.0. Compared with exon 3 Tyr/Tyr, 0 pack-years, risk was highest among those with the exon 3 His/His genotype and >25 pack-years of smoking [adenoma, odds ratio (OR) = 4.9 (1.9–12.8); hyperplastic, OR = 7.7 (2.5–24.0)]. Risks were not elevated among exon 4 homozygous variants, even in the presence of heavy smoking. Fried, baked, or broiled meat intake of ≥two servings/week (high) compared with ≤one serving/week was associated with a 2-fold increase in risk of adenoma. The highest risks were seen for those with the exon 3 His/His genotype and high cooked meat intake [OR = 3.3 (1.4–7.9); reference group: Tyr/Tyr, ≤ 1 serving/week).
Although mEH polymorphisms are not associated with an increased risk of colorectal polyps overall, genotypes that produce a slow phenotype appear to be associated with an increased risk in the presence of smoking and high intakes of cooked meat.
PAHs2 are carcinogens that are metabolized via complex enzymatic mechanisms involving both activation and detoxification. mEH is an enzyme that hydrolyzes epoxides, yielding the corresponding trans-dihydrodiols (1) . Usually, this hydrolysis acts as a detoxifying step, although in some instances, trans-dihydrodiols generated from PAHs are highly toxic and mutagenic (1) . An example is (+)-anti-7,8,diol-9,10-epoxide derived from benzo[a]pyrene (2) . The mEH protein and nucleic acid sequences are highly conserved, and the protein appears universally expressed (3) . Interspecies comparisons suggest that epoxide hydrolysis is a particularly well-developed bio-inactivation pathway in humans (3) . Several-fold variation in mEH activity in humans has been reported (3 , 4) , and this variation can be substrate-dependent. Some of this variation is attributable to known genetic polymorphisms. Hassett et al. (5) have described two common mEH coding-region variants, as well as, more recently, polymorphisms in the 5′ flanking region (6) .
The coding-region variants result in amino acid substitutions at two positions (Tyr113His in exon 3 and His139Arg in exon 4). In vitro expression analyses of the corresponding proteins showed a 40% decrease and 25% increase, respectively, most probably as the result of altered protein stability (4 , 5) . Harrison et al. (7) have reported a >3-fold increased risk of colorectal cancer associated with the exon 3 His/His genotype in a study of 101 cancer cases and 203 blood donors.
Colorectal carcinogenesis is a complex process involving genetic as well as environmental factors. Adenomatous and possibly hyperplastic polyps are precursors of colorectal cancer and share many common risk factors (8) . Tobacco smoke and high intakes of cooked, broiled, or well-done meat have been implicated as risk factors (9, 10, 11) . It is likely that carcinogenic substances derived from these exposures, including PAHs, contribute toward this elevated risk. The goals of this study were to investigate the association between the exon 3 and exon 4 mEH polymorphisms with colorectal adenoma and hyperplastic polyps and to evaluate potential interaction with exposures to smoking and cooked-meat consumption.
Materials and Methods
Subject recruitment for this case-control study has been described previously (12) . Briefly, cases and controls were recruited through a large multiclinic private gastroenterology practice, DH in metropolitan Minneapolis. All patients (30 to 74 years of age) who were scheduled for colonoscopy at DH clinics between April 1991 and April 1994 were screened for eligibility (see criteria below) and recruited before colonoscopy. Recruitment at all 10 DH sites was initiated at the time of scheduling with the intention of recruiting subjects with both patient and recruiter blind to the final diagnosis. The study protocol was approved by the internal review boards of the University of Minnesota and each DH endoscopy site and included written informed consent from each study participant.
Eligibility criteria for both cases and controls included: resident of Twin cities metropolitan area, 30–74 years of age, English speaking, no known genetic syndrome associated with predisposition to colonic neoplasia, no individual history of cancer (except nonmelanoma skin cancer), no previous adenomatous polyps, and no history of inflammatory bowel disease. Indications for colonoscopy have been published previously (13) . The participation rate for all colonoscoped patients was 68%.
At the colonoscopy visit, questionnaires were collected, blood was drawn, and the colonoscopy findings were recorded on standardized forms. Only participants with a complete colonoscopy reaching the cecum were eligible. All of the polyps were removed and evaluated histologically by a single study pathologist based on the diagnostic criteria established for the National Polyp Study (14) . Participants with polyps showing invasive carcinoma were excluded.
On the basis of colonoscopy and pathology findings, study participants were assigned to one of three groups: adenomatous polyp cases (≥one adenomatous polyp), hyperplastic polyp cases (≥one hyperplastic polyp, and no adenomas). Controls were polyp-free at colonoscopy.
Information on (a) dietary intake, (b) physical activity, (c) smoking habits, (d) anthropometric measurements, (e) medical information, (f) demographic information, and (g) reproductive history (women) was collected. Study staff followed up by phone when data were incomplete. Smoking history included current and past smoking status, age when smoking was begun, and average number of cigarettes smoked/day. Years since quitting smoking were also recorded. Pack-years of smoking were calculated as years smoked multiplied by current (or past, for those who quit) cigarettes smoked/day divided by 20.
Data on meat consumption were obtained from specific questions assessing amount consumed and preferences for meat preparation. These questions assessed the preferred degree of cooking (“doneness”) of red meat and poultry; the frequency of cooking by frying, broiling, baking, or barbecuing of red meat, poultry, and fish; and the frequency of the consumption of drippings.
Determination of the mEH exon 3 and exon 4 polymorphisms was conducted at the Core Laboratory of the Public Health Sciences Division of the Fred Hutchinson Cancer Research Center (J. B.). Epoxide hydrolase exon 3 (Tyr113His) and exon 4 (His139Arg) genotyping was performed using an oligonucleotide ligation assay (15) or RFLP (5) , respectively. For the Tyr113His polymorphism a fragment containing the mutation was amplified using primers 5′-CTTGAGCTCTGTCCTTCCCATCCC-3′ and 5′-GACGGCCGTTCTCATGACATACATCC-3′ (5) . The PCR reaction contained 10 mm Tris (pH 8.3), 50 mm KCl, 2 mm MgCl2, (Perkin-Elmer, Foster City, CA), 50 μg/ml BSA, 0.2 μm primers, 0.2 mm deoxynucleotide triphosphates, 1 unit of AmpliTaq DNA polymerase (Perkin-Elmer), and 100 ng of genomic DNA. The cycling conditions were: 5 min at 93°C, followed by 35 cycles at 93°C for 1 min, 55°C for 1 min, 72°C for 1 min, and a final extension at 72°C for 5 min (5) . For the ligation, the PCR reaction was diluted with 1.5 volumes 0.1% Triton X-100. Digoxigenin-tailing of the common primer was performed as described (15) . The 20-μl ligation reactions consisted of 10 μl of diluted PCR product, 20 mm Tris-HCl (pH 8.0), 10 mm MgCl2, 12.5 mm KCl, 1 mm DTT, 1 mm NAD, 0.1% Triton X-100, 8 fmol/μl biotinylated wild-type (5′-GGTGGAGATTCTCAACAGAT-3′) or mutant (5′-GGTGGAGATTCTCAACAGAC-3′) primer, 8 fmol/μl digoxigenin-tailed common primer (5′-ACCCTCACTTCAAGACTAAG-3′), and 0.015 units of thermostable ligase (Epicentre Technologies, Madison, WI). The cycling conditions for the ligation for all of the mutations were: 15 cycles of 93°C for 30 s and 58°C for 2 min. The reaction was stopped with 10 μl of a buffer containing 0.1 m EDTA (pH 8.0) and 0.1% Triton X-100.
The ligation reactions were transferred into streptavidin-coated 96-well plates. After incubation for 60 min at room temperature, the plates were washed twice with 10 mm NaOH, 0.05% Tween 20, followed by two washes with 200 μl of 100 mm Tris-HCl (pH 7.5), 150 mm NaCl, and 0.05% Tween. The plates were then incubated with 40 μl of a 1000-fold dilution of anti-digoxigenin Fab fragment-alkaline phosphatase conjugate (0.75 units/μl; Boehringer Mannheim, Indianapolis, IN) for 30 min at room temperature. After four washes with 100 mm Tris-HCl (pH 7.5), 150 mm NaCl, and 0.05% Tween 20, the Life Technologies, Inc. ELISA amplification system was applied for the color reaction according to the manufacturer’s recommendations. The absorbance at 495 nm was recorded using a SpectraMax 250 plate reader (Molecular Devices, Sunnyvale, CA).
For the His139Arg polymorphism, primers 5′-GGGGTACCAGAGCCTGACCGT-3′ and 5′-AACACCGGGCCCACCCTTGGC-3′ were used to amplify the fragment containing the polymorphism. PCR conditions were the same as for the Tyr113His polymorphism with the exception of the annealing temperature, which was 60°C (5) . For the restriction digestion, 20 μl of amplified PCR fragment, 4 μl of 10 × reaction buffer, two units of RsaI (Amersham Pharmacia, Arlington Heights, IL), and 15.8 μl of distilled water were combined and incubated at 37°C for 60 min. The digestion products were separated on a 3% NuSieve gel. The expected fragment sizes are 295 bp and 62 bp for the wild-type allele and 174 bp, 121 bp, and 62 bp for the variant allele.
Statistical Data Analysis.
Standard techniques for case-control studies were used. Unconditional logistic regression models were used to obtain maximum likelihood estimates, ORs, and 95% CIs. Both age- and sex-adjusted and multivariate ORs are presented. Multivariate adjustment included age, sex, regular use of nonsteroidal anti-inflammatory drugs (y/n), dietary fiber intake (g), total energy intake (kcal), alcohol intake (g), body mass index (kg/m2), and, among women, ever use of hormone replacement therapy (y/n). Analyses for meat variables were, in addition, adjusted for smoking (pack-years). These variables had either shown previously (12) to be modulators of risk of colorectal polyps in this population or altered some risk estimates by 10%. In general, confounding effects were small and often only apparent in stratified analyses; however, for consistency, a full multivariate adjustment was maintained throughout all of the analyses presented.
Separate analyses were performed for adenoma cases and hyperplastic cases, each with colonoscopy-negative individuals as controls. Effect modification of the relation between smoking or meat intake and risk of polyps by genotype was evaluated by testing for different slopes with increasing exposures (smoking and meat) across genotypes. We performed smoking-related analyses both for groups based on pack-years (adjusted for current or former smoking status) and based on smoking status (“current,” “former,” and “never” smoking). The results were comparable, and we focus largely on data grouped by pack-years, which reflects an overall lifetime dose. For analyses with combined exon 3 and exon 4 genotypes, groups were combined only if sample sizes were insufficient for obtaining stable estimates; grouping was not based on a priori assumptions. Because of small cell sizes, the group with the exon 4 homozygous variant genotype could not be split further. In addition, we present data grouped by “imputed phenotype” as suggested by Smith and Harrison (16) . This stratification results in “fast” (wild-type exon 3, heterozygous or homozygous variant exon 4), “normal” (wild-type exon 3/wild-type exon 4 or heterozygous exon 3/heterozygous exon 4), “slow” (heterozygous exon 3/wild-type exon 4), and “very slow” (homozygous variant exon 3/wild-type exon 4) phenotypes. However, several genotype combinations are not captured by this method, and the in vitro data supporting this classification for different substrates are limited.
Our study population was 97% Caucasian. Consequently, restricting the population only to Caucasians yielded virtually identical results; thus, we chose to present the analyses reflecting the entire study population. All tests of statistical significance were two-sided. SAS, version 6.11 (SAS Institute Inc., Cary, NC) was used for all of the analyses.
The study population has been described previously (12 , 13 , 17) . Selected characteristics are presented in Table 1⇓ . Briefly, cases with adenomatous polyps were significantly older than those with hyperplastic polyps or polyp-free controls (P < 0.001). The population was almost entirely Caucasian, and men were at higher risk of colorectal polyps than women. The mEH polymorphisms in exon 3 and exon 4 were in Hardy-Weinberg equilibrium in all of the populations, with an allele frequency of 0.28 for the exon 3 His allele and 0.20 for the exon 4 Arg allele in the control population. The polymorphisms were not in linkage disequilibrium.
The relevant exposures for PAHs are smoking and intake of meat that has been fried, baked, broiled, or barbecued. Risks of colorectal polyps associated with these exposures are presented in Table 2⇓ . Smoking-related variables were consistently associated with an increased risk of polyps, to a larger extent for hyperplastic polyps than for adenomas. For meat intake, a higher intake of “fried, baked, or broiled meat” was associated with an increased risk of adenomatous polyps only. Other meat variables (preferred level of red meat doneness or consumption of drippings) were not associated with either adenomatous or hyperplastic polyps.
The main effects of the mEH polymorphisms are shown in Table 3⇓ . Risks of colorectal polyps did not differ substantially by different exon 3 or exon 4 genotypes. Similarly, no clear pattern emerged when different combinations of exon 3 and exon 4 genotypes or imputed phenotypes were considered. Although not statistically significant, the results for the exon 3 and exon 4 homozygous variants are consistent with an opposite effect on enzyme function, as observed in in vitro studies [for adenomatous polyps exon 3 His/His, OR = 1.3 (0.8–2.2); exon 4 Arg/Arg, OR = 0.8 (0.4–1.5)].
Table 4⇓ presents results on mEH polymorphisms, stratified by relevant exposure variables (pack-years of smoking and intake of fried, baked, or broiled meat). In consideration of the somewhat smaller sample sizes for analyses involving meat consumption, analyses were stratified into only two groups of intake. Consistently, the groups with highest exposure levels (smoking >25 pack-years, or individuals with fried, baked, or broiled meat at least two times/week) and the exon 3 homozygous variant genotype were at highest risk of colorectal adenoma [smoking >25 pack-years, OR = 4.9 (1.9–12.8); high meat intake, OR = 3.3 (1.4–7.9)], although no statistically significant gene-environment interaction was observed. Those who were homozygous variant for the exon 4 allele showed ORs that were lower than expected on a multiplicative scale, suggesting an inverse interaction; however, these results are based on small cell sizes.
We further analyzed the risk of colorectal adenoma associated with cumulative smoking dose (≤25 or >25 pack-years) and exon 3 genotype, stratified by current or former smoking status. Consistently, in both former and current smokers, a homozygous exon 3 variant genotype conferred an increased risk among those with higher exposure compared with those who were wildtype and had zero pack-years of smoking. (Table 5)⇓ .
Analyses based on different genotype combinations of the exon 3 and exon 4 polymorphism and for a categorization based on imputed phenotype were also undertaken. (Table 6)⇓ . The cell sizes were small, but the highest risks associated with smoking >25 pack-years were seen for those who were homozygous variant for the exon 3 and wildtype or heterozygous for the exon 4 genotype [OR = 4.5 (1.7–12.3), compared with individuals with the double wild-type genotype]. Furthermore, among those with a “very slow” imputed phenotype (equivalent to homozygous variant exon 3/wild-type exon 4), smoking was associated with a about an 8-fold increase in adenoma risk, compared with an about 2-fold increase in risk among those with an imputed “normal” phenotype (wildtype exon 3/wildtype exon 4 or double heterozygous). Consistently, among these individuals, increases in risk associated with high intakes of cooked meat were most pronounced.
For hyperplastic polyps (Table 7)⇓ , the number of cases was substantially lower than for adenomas, resulting in wider CIs, especially for the homozygous variant genotypes. Nevertheless, the overall patterns, when stratified by smoking, are consistent with the findings on adenomatous polyps. Smoking is a stronger risk factor for hyperplastic polyps than for adenomatous polyps, and the most pronounced increases in risk were observed among individuals with the “slow” exon 3 His/His genotype. Relatively smaller increases in risk associated with smoking were seen among individuals with the exon 4 “rapid” Arg/Arg genotype compared with “wildtype,” consistent with opposite effects of these two polymorphisms (see Fig. 1⇓ ). As noted in Table 1⇓ , meat intake was not associated with risk of hyperplastic polyps, so no stratified analyses are presented for this exposure.
In this case-control study, the exon 3 and exon 4 polymorphisms in the mEH gene were not related to risk of colorectal polyps, either adenomatous or hyperplastic, when exposure status was not considered. This is in contrast to the findings of Harrison et al. (7) who reported a significantly elevated risk of colorectal cancer (OR = 3.5) among individuals with the exon 3 (His/His) “slow phenotype,” as well as an elevated risk (not statistically significant) among individuals with the exon 4 “rapid phenotype” (OR = 2.6). Their study compared 101 colorectal cancer patients with 203 blood-donor controls, and it is unclear whether the reported risk estimates were adjusted for any covariates. The discrepancy in findings may be attributable to these differences in study design and statistical analysis or simply reflect a difference in outcome (colorectal polyps versus colorectal cancer).
There are reported associations between the exon 3 genotype and other outcomes, including ovarian cancer (18) , spontaneous abortion (19) , lung cancer (20 , 21) , and bladder cancer (22) , although sample sizes in these studies were small. Clearly, there is need for further research, with adequate sample sizes to investigate the relationship between polymorphisms in this metabolizing enzyme, and interactions with relevant exposure variables.
Individuals with the exon 3 homozygous variant genotype (“slow” phenotype) who had a high cumulative smoking dose, as measured by pack-years, were at substantially increased risk of colorectal adenoma (greater than 3-fold). A similar increase in risk with this “slow” phenotype was seen among subjects with a higher intake of fried, baked, broiled, or barbecued meat. For the imputed “very slow” phenotype, an even stronger increase in risk associated with smoking was observed (about 8-fold). Although not statistically significant, the association with the exon 4 polymorphism variant among current smokers is consistent with opposite effects of these two polymorphisms on enzyme activity. Overall, our results indicate that slow activity of mEH in combination with a relevant exposure (smoking or fried/broiled meat intake) is associated with an increased risk of colorectal polyps. Considering that benzo(a)pyrene, a major carcinogen in cigarette smoke, is activated into a more carcinogenic substance by mEH, these findings appear counterintuitive. Thus, our results may indicate that components of cigarette smoke and cooked meat other than benzo(a)pyrene are responsible for the elevated risk of colorectal polyps associated with smoking and that these substances are detoxified by mEH. Alternatively, “activation” of benzo(a)pyrene may be beneficial, because it permits conjugation and excretion of the product by phase II enzymes.
Overall, risk associated with smoking was more consistently associated with colorectal polyps than the risk associated with cooked meat consumption (measured by intake as well as preferred meat doneness). Although both environmental exposures are associated with an “internal exposure” to polycyclic aromatic hydrocarbons, there are some differences: (a) the PAH dose associated with cigarette smoking in this study population would have been higher and more continuous than the dose associated with the more sporadic consumption of cooked meat; and (b) specific carcinogens in cigarette smoke versus cooked meat may differ. Furthermore, the type of cooking recorded in this study may not account for all of the variations in the PAH levels from foods to which study participants were exposed.
Our study illustrates a challenge in molecular epidemiology, genes with multiple known polymorphisms and little information regarding consequences for phenotype, and even the possibility of differences in effects on substrate specificity. One option is to assign an “imputed phenotype” to different genotype combinations based on in vitro data or assumptions that effects are additive. However, this approach is limited in the absence of sound data on functional significance for different genotype combinations. To evaluate the interactions between genetic and environmental factors adequately and without preexisting data on specific phenotypes, even larger sample sizes are needed.
This case-control study had several limitations. The study population was not necessarily representative of the entire population, because only individuals who underwent colonoscopy were eligible. The major advantage of this clinic-based approach was that the presence of polyps was clearly established and that the control group was free of any polyps (studies that use a population-based control group will probably include a substantial proportion of individuals with undetected polyps, which can attenuate study findings). A potential disadvantage is that controls who undergo colonoscopy may be more like adenoma cases on a variety of lifestyle features than they are like members of the community from which they are derived. In this study, we recruited in addition community controls for exactly this reason (12) . The community controls and colonoscopy-negative controls were very similar in relation to education, obesity, energy, fat intake, fiber intake, fruit and vegetable intake, smoking, and alcohol consumption (23) , as well as physical activity.3 The colonoscopy-negative controls had a higher positive family history than the community controls,4 and among women, slightly higher prevalence of a history of use of oral contraceptives and hormone replacement therapy (12) .
A strength of this case-control study was the relatively large study size, in particular for adenoma, which allowed us to investigate gene-environment interactions. Nevertheless, the sample size was clearly limited for exploring combined effects of the two genotypes, which underscores the need for large epidemiological studies to investigate these associations.
In summary, these findings indicate that the epoxide hydrolase exon 3 and exon 4 polymorphisms themselves do not alter risk for colorectal polyps substantially. However, among individuals exposed to relevant carcinogens that are metabolized by mEH (e.g., smokers), slow enzyme activity (exon 3 His/His) appears to result in a more pronounced risk of both colorectal adenoma and hyperplastic polyps. Furthermore, the study results are consistent with opposite effects of the exon 3 and exon 4 polymorphisms on enzyme function.
We thank Angela Bush for technical assistance with the DNA genotyping and Mari Nakayoshi at the FHCRC with graphical and word processing assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 To whom requests for reprints should be addressed, at Cancer Prevention Research Program, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue N, MP-900, Seattle, WA 98109-1024. Phone: (206) 667-7617; Fax: (206) 667-7850; E-mail:
↵2 The abbreviations used are: PAH, polycyclic aromatic hydrocarbon; mEH, microsomal epoxide hydrolase; DH, Digestive Healthcare; OR, odds ratio; CI, confidence interval.
↵3 J. D. Potter, L. Fosdick, R. Bostick, T. A. Louis, P. Grambsch, unpublished data.
↵4 J. D. Potter, unpublished data.
- Received September 1, 2000.
- Revision received May 25, 2001.
- Accepted June 4, 2001.