• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of carcinLink to Publisher's site
Carcinogenesis. Jul 2012; 33(7): 1352–1359.
Published online May 18, 2012. doi:  10.1093/carcin/bgs175
PMCID: PMC3499053

Polymorphisms in carcinogen metabolism enzymes, fish intake, and risk of prostate cancer

Abstract

Cooking fish at high temperature can produce potent carcinogens such as heterocyclic amines and polycyclic aromatic hydrocarbons. The effects of these carcinogens may undergo modification by the enzymes responsible for their detoxification and/or activation. In this study, we investigated genetic polymorphisms in nine carcinogen metabolism enzymes and their modifying effects on the association between white or dark fish consumption and prostate cancer (PCA) risk. We genotyped 497 localized and 936 advanced PCA cases and 760 controls from the California Collaborative Case–Control Study of Prostate Cancer. Three polymorphisms, EPHX1 Tyr113His, CYP1B1 Leu432Val and GSTT1 null/present, were associated with localized PCA risk. The PTGS2 765 G/C polymorphism modified the association between white fish consumption and advanced PCA risk (interaction P 5 0.002), with high white fish consumption being positively associated with risk only among carriers of the C allele. This effect modification by PTGS2 genotype was stronger when restricted to consumption of well-done white fish (interaction P 5 0.021). These findings support the hypotheses that changes in white fish brought upon by high-temperature cooking methods, such as carcinogen accumulation and/or fatty acid composition changes, may contribute to prostate carcinogenesis. However, the gene–diet interactions should be interpreted with caution given the limited sample size. Thus, our findings require further validation with additional studies.

Abbreviations:

AA
African American;
BMI
body mass index;
CI
confidence interval;
CNV
copy number variant;
EPIC
European Prospective Investigation into Cancer and Nutrition;
HCA
heterocyclic amine;
HCFA
Health Care Financing Administration;
LAC
Los Angeles county;
MAF
minor allele frequency;
NHW
non-Hispanic White;
OR
odds ratio;
PAH
polycyclic aromatic hydrocarbon;
PCA
prostate cancer;
PTGS2
prostaglandin- endoperoxide synthase 2;
PUFA
polyunsaturated fatty acids;
RDD
random-digit dialing;
SEER
Surveillance, Epidemiology, and End Result;
SES
socio-economic status;
SFBA
San Francisco Bay Area;
SNP
single-nucleotide polymorphism

Introduction

Diets high in fish have been associated with decreased risk of advanced prostate cancer (PCA) and decreased PCA mortality (1–4).
It has been proposed that omega-3 polyunsaturated fatty acids (PUFA) found in fish, particularly dark and oily species, might explain this association (5). Other studies, however, have reported no association (6–8) or positive associations (9). Recently, we reported that diets high in white fish (flounder, halibut, snapper, bass, cod or sole) were associated with increased PCA risk, especially among men who consumed white fish cooked with high-temperature methods such as pan-frying, grilling, oven-broiling or barbequing (10). Cooking meats and fish using these high-temperature methods generates mutagens such as heterocyclic amines (HCAs) and polycyclic aromatic hydrocarbons (PAHs), which we hypothesize may be responsible for the increased PCA risk (11). HCAs present in cooked fish can increase both mutation frequency and prostate tumor incidence in animal models (12,13). Moreover, PAH–DNA adducts have also been detected in human prostate cells (14).

HCAs and PAHs present in cooked fish are absorbed in the intestine and transported via the hepatic portal system to the liver where they undergo metabolism (15). Native HCAs, PAHs and metabolites can then reach the prostate gland via blood circulation where they are metabolized by similar enzymes as those present in the liver. HCAs are activated by CYP1A2 into N-hydroxylamines, which can then undergo further metabolism by SULT1A1, NAT1, NAT2 and PTGS2 to generate reactive species that can damage DNA (16,17). The N-acetoxy-HCAs that result from NAT1 and NAT2 metabolism can be substrates for detoxification by GSTM1, GSTT1 and GSTP1 (18). PAHs are activated by CYP1A1 and CYP1B1, generating PAH epoxides, which can be further activated by EPHX1 to form ultimate carcinogens that can attack DNA, or they can be detoxified by Phase II enzymes (GSTM1, GSTT1, GSTP1, GSTA1, UGT1A1, UGT1A6 and UGT1A9 (19,20)). In addition, PTGS1 and PTGS2, which are expressed during prostatic inflammation, can further metabolize PAH-dihydrodiols to reactive PAH-dihydrodiol-epoxides, capable of attacking DNA (21–23).

Given the metabolic role of the above mentioned enzymes, along with our recent finding that consumption of fish cooked under conditions conducive to HCA formation is positively associated with PCA risk (10), we hypothesized that genetic variation in carcinogen metabolism enzymes may play a role in the etiology of PCA. Furthermore, we hypothesized that these genetic variants may modify the association between fish consumption and PCA risk, particularly when fish is cooked using high-temperature methods. Using data and samples from the California Collaborative Case–Control Study of Prostate Cancer, we investigated the role of genetic variation in nine metabolic enzymes relevant for HCA and PAH metabolism in combination with fish intake and cooking practices. We included the following polymorphisms which have been reported to impact enzyme function and/or have been associated with PCA risk: GSTP1 Ile105Val (rs1695 (24,25)), PTGS2-765 G/C (rs20417 (26)), CYP1A2-154 A/C (rs762551 (27)), EPHX1 Tyr113His (rs1051740 (28)), CYP1B1 Leu432Val (rs1056836 (29)), UGT1A6 Thr181Ala and Arg184Ser (rs1105879 and rs2070959 (30)) and NAT2 Ile114Thr, Arg197Gln, Gly286Glu and Arg64Gln (rs1799930, rs1799931, rs1801279 and rs180120 (31)), in addition to two genes that have copy number variants, GSTM1 and GSTT1 (24,25).

Study design and methods

Study population

The California Collaborative Case–Control Study of Prostate Cancer, described elsewhere (32,33), is a multiethnic, population-based case–control study that identified incident cases of PCA through two regional cancer registries in Los Angeles County (LAC) and the San Francisco Bay Area (SFBA). In both study sites, PCA was classified as advanced disease if the tumor extended beyond the prostatic capsule or into the adjacent tissue, involved regional lymph nodes or metastasized to distant locations (Surveillance, Epidemiology, and End Result 1995 clinical and pathologic extent of disease codes 41–85). Intra-capsular prostatic cancers were categorized as localized disease. In LAC, controls were identified through a standard neighborhood walk algorithm (34) and frequency-matched based on age (±5 years) and race/ethnicity. In SFBA, controls were primarily identified through random-digit dialing. In addition, a subset of controls aged ≥65 years were also identified through random selections from the rosters of beneficiaries of the Health Care Financing Administration. Controls were frequency-matched to cases with advanced disease based on 5-year age group and race/ethnicity. In LAC, study eligibility was restricted to men willing to both complete an in-person interview and provide a blood sample; 1232 cases [376 African American (AA), 355 Hispanics and 501 non-Hispanic Whites (NHW)], of 1870 eligible cases diagnosed from 1999 to 2003, and 594 controls (163 AA, 122 Hispanics and 309 NHW) participated in the study. In SFBA, in-person interviews were completed with 568 (118 AA and 450 NHW) advanced cases (of 788 eligible cases diagnosed from 1997 to 2000), and 545 (90 AA and 455 NHW) control men of 868 eligible controls. Blood or mouthwash samples were obtained for 533 advanced cases (107 AA and 426 NHW) and 525 controls (85 AA and 440 NHW). No biospecimen samples were collected for localized cases (32). At the time we performed the genotyping for this study, DNA from blood and data on risk factors were available for a total of 800 controls (545 from LAC and 255 from SFBA), 535 localized cases (LAC) and 988 advanced cases (597 from LAC and 391 from SFBA). Written informed consent was obtained from all study participants at the time of in-person interview.

Data collection

A common structured questionnaire was used at both study sites, including a 74-item food frequency questionnaire that was adapted from Block’s Health History and Habits Questionnaire (32). The food frequency questionnaire also included questions on cooking methods and degree of doneness and browning (35) that were adapted from a commonly used cooking module developed by Sinha et al. (36). An aggregate-level socio-economic status (SES) variable was derived from 2000 census data as described previously (33). Body mass index (BMI) was calculated using the reported weight in the reference year (defined as the calendar year before diagnosis for cases and the calendar year before selection into the study for controls) and measured height at the time of the interview. BMI was calculated as weight (kg) divided by height (m) squared and categorized as normal weight (BMI < 25), overweight (BMI 25–29.9) and obese (BMI ≥ 30). Underweight men (BMI < 18.5, n 5 15) were grouped with normal-weight men.

Exposure variables

The food frequency questionnaire assessed all food intake and cooking methods during the reference year. Subjects reported the usual portion size and frequency of consumption of tuna fish (including tuna salad, tuna casserole and tuna sandwiches), deep fried fish (including fish sticks and fish sandwiches), dark fish (salmon, mackerel, catfish, trout, herring and sardines) and white fish (flounder, halibut, snapper, bass, cod or sole). Information was obtained on usual method of preparation of dark fish and white fish (pan-frying, oven-broiling, grilling, baking/roasting, microwaving and other methods). The participants also reported their usual preference for the level of doneness of dark fish and white fish by choosing from a series of color photographs showing three levels of doneness and browning (1): just until done (not browned on the outside), (2) well done (browned on the outside) and (3) very well done (charred on the outside). There were only 74 men who usually consumed fish ‘very well done - charred on the outside’; therefore, men in this category were combined with men who usually ate fish well done. Cooking by pan-frying, oven-broiling or grilling was considered high-temperature cooking methods, and cooking by baking or other methods was considered to be low-temperature cooking methods.

Adjustment for total energy intake

In our study population, total caloric intake was associated with PCA risk
(P < 0.001). Residuals obtained from regression of fish intake variables on calories showed non-normality and heteroskedasticity (even after log transformation) because our fish exposure variables had few unique values. Hence, we used the multivariate nutrient density approach to adjust for energy intake (37). Nutrient densities of white fish and dark fish were created multiplying their reported daily intake with the reciprocal of the total calories consumed per day.

Nutrient density variables were categorized into three levels of consumption: never/rarely (quintile 1), low (quintiles 2–4) and high (quintile 5). These categories were chosen to circumvent spurious negative associations that are possibly driven by the influence of large variation in the denominator of the density variable (inverse calorie intake) compared with the variance of the nutrient of interest in the numerator. When tabulating the quintiles of nutrient density intake against the quintiles of total calorie intake, we observed that the middle three quintile categories were strongly affected by this phenomenon and the division among them could almost exclusively be explained by energy intake; thus, the middle three quintiles were collapsed.

Genotyping

We genotyped individuals for 12 single-nucleotide polymorphisms (SNPs) in 9 genes: GSTP1 Ile105Val (rs1695), PTGS2-765 G/C (rs20417), CYP1A2-154 A/C (rs762551), EPHX1 Tyr113His (rs1051740), CYP1B1 Leu432Val (rs1056836), UGT1A6 Thr181Ala and Arg184Ser (rs1105879 and rs2070959) and NAT2 Arg197Gln, Gly286Glu, Arg64Gln and Ile114Thr (rs1799930, rs1799931, rs1801279 and rs1801280), in addition to two genes that had copy number variants, GSTM1 and GSTT1. All genotypes were obtained using Taqman assays, available on request from Applied Biosystems (Foster City, CA), following manufacturer’s instructions. No differences were found between observed genotypic frequencies and those expected under the Hardy–Weinberg Equilibrium. Call rates were >97%. The four NAT2 SNPs genotyped infer the NAT2 phenotype with estimated 98.4% accuracy in populations with similar SNP frequencies as in this study (38). In agreement with the existing classification, carriers of at least one copy of the wildtype haplotype were classified as ‘fast’ and carriers of all other haplotypes as ‘slow’ phenotype (39).

Statistical analysis

Given differences in the distribution of race/ethnicity and SES between the two study sites, we created a variable that classified subjects according to study site (LAC or SFBA), SES (five-level variable from low to high, as described previously (33)), as well as race/ethnicity (AA, Hispanic, NHW) to group individuals in conditional logistic regression models. SES was collapsed into three categories (quintiles 1–2, 3 and 4–5) at the SFBA site and four categories (quintiles 1, 2, 3 and 4–5) at the LAC site, leaving 6 SES/race groups from SFBA and 12 from LAC. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated from conditional logistic regression, with separate analyses performed for localized and advanced cases. We estimated ORs and CIs for each genotype using dummy variables and per variant allele assuming a log-additive model. All OR estimates were adjusted for age (years, continuous) and family history of PCA in first-degree relatives (yes, no). Minor allele frequencies (MAF), ranging from 0.16 to 0.49, did not differ significantly from those reported for the general reference population (www.hapmap.org). Our study design and sample size can detect ORs as small as 1.25 (MAF 5 0.49) to 1.34 (MAF 5 0.16) among localized cases and ORs as small as 1.21 (MAF 5 0.49) to 1.29 (MAF 5 0.16) among advanced cases with 80% power when testing at the 5% significance level (all power calculations performed with Quanto (40)). When investigating main effects of fish, we also adjusted for potentially confounding lifestyle and dietary factors. These variables were total fat intake (g/day), dietary vitamin D intake (g/day), alcohol consumption (g/day), total dairy intake (g/day), cigarette smoking (pack-years), total fruit consumption (g/day), total vegetable consumption (g/day), red meat consumption (g/day), white meat consumption (g/day) and processed meat consumption (g/day) (Supplementary Table I, available at Carcinogenesis Online). We did not see any appreciable differences in estimates and overall trends; therefore, further gene–environment analyses were not adjusted for these additional variables.

Analyses of gene-by-diet interactions were done coding each SNP as log additive and fish consumption as a three-level variable (never, low and high), using the median level of exposure among controls at each of the levels. We conducted both 2-df interaction tests by treating the three-level fish intake exposure variable as categorical, and 1-df interaction tests by treating this variable as ordinal. Interaction models were adjusted for age (years, continuous), BMI (<25, 25.0–29.9 and ≥30), total calorie intake (kcal/day, continuous) and family history of PCA in first-degree relatives (yes, no). Analyses of white fish were adjusted for dark fish intake and vice versa. To analyze gene-by-fish interactions while taking into account cooking practices, we created variables that captured frequency of intake of either white or dark fish when cooked using high-temperature methods (pan-frying, grilling, oven-broiling and barbequing) or when cooked until well done. Models that tested for interaction using variables that captured fish intake and cooking practices were also adjusted for total well-done meat consumption (g/day) and consumption of total meat cooked at high temperature (g/day). Total meat variables included red meat, poultry and processed meats. Tests of trend across categories of fish consumption were done by coding each category by its median value and modeling the category variable as continuous. To account for multiple testing, we used a P-value cutoff of 0.0056 5 0.05/9 for each of the exposures tested, which is the Bonferroni-corrected P-value 5 0.05 for testing interactions between each given exposure and nine different genotypes or phenotypes. Further correction accounting for the two stages of disease set the P-value cutoff at 0.0028 5 0.05/18. Our study design and sample size can detect interaction ORs as small as 1.62 (MAF 5 0.49) to 1.84 (MAF 5 0.16) among localized cases and interaction ORs as small as 1.49 (MAF 5 0.49) to 1.69 (MAF 5 0.16) among advanced cases with 80% power when testing at the 5% significance level (all power calculations performed with Quanto (40)). All hypothesis tests were two-sided and all analyses were done using the statistical software Stata S/E 11.0 for Windows (STATA Corporation, College Station, TX). After excluding 124 individuals (87 cases and 37 controls) with dietary data considered unreliable (i.e. daily energy intake <600 kcal or >6000 kcal), all dietary analyses were based on 1420 cases and 746 controls.

Table I.
Socio-demographic and lifestyle characteristics of cases and controls with available DNA from blood, by stage of disease and study site

Results

In Table I, we summarize the demographic and other relevant characteristics of cases and controls included in the present analysis. Age was normally distributed among controls. Mean age was comparable between study sites and between advanced cases and controls; however, localized cases had a higher mean age than both controls and advanced cases. Mean caloric intake was higher among LAC relative to SFBA as well as cases in relation to controls, but was similar among advanced and localized cases.

Among the cases and controls included in this study, we observed a 30% reduction of both localized and advanced PCA risk in men with a low or high intake of dark fish compared with those who never/rarely consumed dark fish (OR localized disease 5 0.72; 95% CI 5 0.53–0.96; OR advanced disease 5 0.73; 95% CI 5 0.55–0.85) (Supplementary Table I, available at Carcinogenesis Online). Furthermore, PCA risk was reduced by 50% among men who usually used low-temperature cooking methods, although the interaction between cooking methods did not reach statistical significance (OR localized disease/high intake 5 0.52; 95% CI 5 0.26–1.02; OR advanced disease/high intake 5 0.53; 95% CI 5 0.31–0.89) (Supplementary Table II, available at Carcinogenesis Online). High white fish consumption was associated with a 50–60% increase in risk of both localized and advanced PCA among men who consumed well-done fish; however, the risk estimate only reached significance among advanced cases (OR localized disease 5 1.53; 95% CI 5 0.81–2.86; OR advanced disease 51.65; 95% CI 5 1.02–2.69) (Supplementary Table III, available at Carcinogenesis Online). These findings are similar to those reported previously for the full study population regardless of DNA availability (10).

Polymorphisms in carcinogen metabolism enzymes and PCA risk

Allelic frequencies did not differ between racial/ethnic groups or study sites for any of the SNPs or glutathione-S-transferase null polymorphisms genotyped (data not shown). We observed a 1.2- to 1.7-fold increased risk of localized PCA associated with three polymorphisms in three different genes: EPHX1 Tyr113His (per His allele OR 5 1.27; 95% CI 5 1.04–1.56), CYP1B1 Leu432Val (per Val allele OR 5 1.32; 95% CI 5 1.09–1.61) and GSTT1 null/present (null OR 5 1.68; 95% CI 5 1.19–2.38) (Table II). We observed no comparable associations for advanced PCA. A test of heterogeneity for the difference of these associations between advanced and localized disease did not reach statistical significance for any gene. Moreover, inheritance of the variant alleles did not yield differences in the risk estimates for localized/advanced PCA across the three racial/ethnic groups and any of the analyzed genes.

Fish intake, metabolism enzymes and PCA risk

We considered interactions between each of the eight SNP genotypes and NAT2 predicted phenotype and intake of white or dark fish. We observed evidence that the PTGS2 765 G/C polymorphism modified the association between high white fish intake and advanced PCA risk (crude 1-df P for interaction 5 0.002, Bonferroni-corrected 1-df P for interaction 5 0.018) (Table III). Specifically, the association between high white fish intake and advanced PCA risk was restricted to carriers of the C allele, reaching a more than 3-fold increased risk (for high versus no/rare white fish intake) among carriers of two C alleles (OR 5 3.56; 95% CI 5 1.61–7.88). There was no evidence of interaction with white fish intake for any of the other polymorphisms investigated for either localized or advanced disease.

When considering dark fish intake, there was an inverse association between dark fish intake and advanced PCA risk among men with the GG genotype, but not among carriers of the C allele (Table III). However, this interaction was not statistically significant after correcting for Bonferroni. We found no evidence of interactions between dark fish intake and the other polymorphisms.

Table II.
Polymorphisms in metabolic genes and PCA risk, by stage of disease

For both white and dark fish intake, similar trends were observed for localized and advanced diseases; however, the tests for interaction were not statistically significant for localized disease. Interactions between the PTGS2 variant and low/high intake of white/dark fish in relation to localized/advanced PCA did not vary across the three racial/ethnic groups (data not shown).

Fish intake, metabolic enzymes, cooking practices and PCA risk

We further assessed whether polymorphisms in metabolic enzymes might modify the association between localized and/or advanced PCA risk and fish intake when taking into account cooking methods and level of doneness. For this purpose, we investigated fish intake when cooked using high-temperature cooking methods only (pan-frying, grilling, oven-broiling and barbequing), and intake of well-done fish only. We observed that PTGS2 genotype modified the association between high intake of well-done white fish and advanced PCA similarly to all white fish, although the interaction term was no longer significant after adjusting for Bonferroni (crude P for interaction = 0.021) (Table IV). As before, this association was restricted to carriers of the C allele, with estimated 2- to 4-fold increased risks (for high versus no/rare white fish intake) for carriers of one or two C alleles (OR CG genotype 5 2.17; 95% CI 5 1.05–4.48; OR CC genotype 5 3.79; 95% CI 5 0.92–15.70). These estimates were slightly stronger than what we observed for white fish without considering level of doneness. We observed no evidence of interactions when considering intake of white or dark fish cooked using high-temperature methods (data not shown).

Discussion

In this study, we report that genetic variants in several metabolic enzymes, CYP1B1 Leu432Val, EPHX1 Tyr113His and GSTT1 present/null, are associated with risk of localized PCA. We also found evidence that the PTGS2 765 G/C polymorphism modifies the association between diets high in white or dark fish and advanced PCA risk. The modifying effect appears stronger for consumption of well-done white (lean) fish.

There exists substantial biological plausibility for a role of CYP1B1 in the development of PCA. PAHs that enter the prostate induce CYP1B1 expression. CYP1B1 is over-expressed in prostate carcinomas (41) and can activate PAHs into mutagenic metabolites that are capable of forming DNA adducts (42). The CYP1B1 protein coded by the codon 432 Valine allele is more active than the one coded by the Leucine allele (29); therefore, our finding of an association between the Valine (Val) allele and higher PCA risk is compatible with the known functional impact of this variant allele on the CYP1B1 protein. We distinguished between advanced and localized PCA and we found the CYP1B1 codon 432 Val allele to be associated with increased risk of localized disease only. Our findings are supported by two studies that found the CYP1B1 Leu432Val Val allele to be associated with an increased risk of PCA among Caucasian (29) and Japanese (43) men. These studies (29,43) did not distinguish between localized or advanced cases; therefore, it is difficult to fully compare those findings with ours. Another study reported the Val allele being associated with a decreased risk of advanced disease in Hispanic men and with decreased disease aggressiveness (localized versus advanced) in NHW men (44) which is consistent with our study finding the increased risk in localized disease only.

Table III.
Fish intake and PCA risk stratified by PTGS2 genotype, by stage of disease
Table IV.
Well-done white fish intake and risk of advanced PCA, stratified on PTGS2 genotype

EPHX1 codes for the epoxide hydrolase protein mEH, a Phase 1 enzyme that converts PAH epoxides to PAH-dihydrodiol, which can undergo further metabolic activation by CYP1A1 and CYP1B1, or detoxification by enzymes in the glutathione-S-transferase or UGT family (45). Therefore, this enzyme plays an important role in the generation of species that are at the crossroads between carcinogenic activation and detoxification. The EPHX1 codon 113 Histidine (His) allele codes for a protein with reduced activity compared with the one coded by the Tyrosine allele (28). Our finding of a positive association between the His allele and localized PCA risk suggests that the slow EPHX1 His allele may contribute toward increased metabolic activation, perhaps by reducing the bioavailability of substrates for PAH epoxide detoxification. Alternatively, reduced levels of EPHX1 may result in an increased bioavailability of pro-carcinogenic PAH epoxides. PAH epoxides can serve as direct substrates for AKR1C3 in the prostate where they can undergo conversion to catechols, which may lead to the accumulation of DNA-damaging PAH-O-quinones (46). One study reported a positive association between the codon 113 His allele and advanced PCA risk in an Indian population, and another reported an association between this allele and higher grade tumors in an Israeli population (47,48).

GSTT1 belongs to the glutathione-S-transferase superfamily and is responsible for detoxifying many carcinogens, including HCAs and PCAs. Approximately 20% of Caucasians have a complete deletion of this gene, which may predispose to cancer due to a decreased ability to detoxify these substrates (24). In recent years, GSTT1 status has been extensively studied as a PCA risk factor; however, the results are inconsistent. A 2009 meta-analysis concluded that there was no significant evidence that the GSTT1 deletion increased risk for PCA, regardless of racial group (49). Since then, six additional studies have investigated the association with mixed results. Three reported an approximately 2-fold increase in risk with GSTT1 deletion in Iranian, Tunisian and Danish populations (50–52), whereas Taioli et al. (53) reported an inverse association with GSTT1 deletion in populations of African descent. Two additional studies reported no associations in the German European Prospective Investigation into Cancer and Nutrition cohort (54) and a population of African descent (55). Few of these studies distinguished between localized and advanced cases; however, Safarinejad et al. (52) found the increase to be even more substantial in those with advanced PCA. In our study population, we found the GSTT1 deletion variant to be associated with increased risk of localized PCA, with a similar trend, albeit not statistically significant, for advanced PCA. Our results are compatible with the hypothesis that lack of GSTT1 detoxification may contribute to accumulation of activated HCA and PAH metabolites, which could contribute to prostate carcinogenesis.

It is interesting to note that for these three polymorphisms, statistically significant associations were only found for localized disease, but not advanced disease. As we reviewed above, there are previous studies that have reported on associations between the EPHX1 Tyr113His and the GSTT1 deletion and risk of advanced disease. Therefore, further studies with larger numbers of both localized and advanced cases are needed to clarify whether these enzymes play a differential role for early or late disease.

Although we had slightly less statistical power for detecting significant gene–diet interactions for localized disease than advanced disease, the fact that we found no strong evidence that any of these three polymorphisms modified the association between consumption of white fish cooked at high temperature and localized or advanced PCA risk may suggest that there are other relevant substrates for these enzymes. Larger studies are needed to confirm this. As we recently reported, the observed association between high intake of white fish cooked at high temperature (i.e. mostly pan-frying) and PCA risk could be driven by other changes that occur in the fish during the cooking process (10). Specifically, it is plausible that changes in white fish fatty acid composition may occur when white fish is cooked at high temperature, such as pan-frying, and that these changes are the ones driving the observed associations (10). If this hypothesis were true, it would explain why we failed to find gene-by-environment interactions for these three metabolism enzyme polymorphisms, in spite of their overall association with localized PCA. Interestingly, as we discuss below, we only observed an effect modification for PTGS2, which participates in both carcinogen metabolism and PUFA metabolism.

To our knowledge, the only metabolic enzyme reported to date as a potential modifier of the association of PCA with high fish intake is the PTGS2 enzyme (56,57). In our study, we investigated eight other biologically plausible enzymes jointly with fish intake and found that the PTGS2 765 G/C polymorphism was the only one that showed evidence of interaction with white and dark fish intake. PTGS2 is over-expressed in PCA cells and is induced by HCAs and PAHs which it can activate into carcinogens (16,21,22). PTGS2 is also involved in the metabolism of omega-6 PUFAs, such as arachidonic acid, into prostaglandins that can contribute to carcinogenesis. Moreover, omega-3 PUFAs found in fish, which are particularly abundant in dark fish, have been shown to block the activity of PTGS2 in hepatic and colon cells (58). Therefore, it is biologically plausible that genetic variation in PTGS2 might modify the association between fish intake and risk of PCA. The PTGS2 765 G/C polymorphism is located in a binding site upstream of the translation initiation site, its effect on PTGS2 gene expression in the prostate is still unclear. In vitro, the C allele had a lower activity than the G allele when transfected into human cervical epithelium cancer cells, but opposite results were obtained when transfection was conducted on human neural cells (26,59). Sanak et al. (2010) recently reported an increase in prostaglandin production associated with the C allele in heart disease patients; however, they simultaneously reported reduced PTGS2 activity with the C allele in both human endothelial and human leukemia cell lines (23,60). We found that diets high in white fish are associated with PCA among carriers of the C allele, and that diets high in dark fish are inversely associated with PCA risk only among carriers of the G allele. These data suggest that the C allele might be the more active form of PTGS2 in the prostate, and that under conditions of exposure to high levels of pro-carcinogens such as HCAs, PAHs and/or omega-6 PUFAs, all found in diets high in fish intake, it might be associated with risk of PCA. We found the PTGS2 interaction with white fish to be stronger when we restrict the analysis to men who consumed well-done white fish. This could lend support to the hypothesis that the interaction is due, at least in part, to the formation of HCAs and PAHs, as we would expect that fish cooked longer (i.e. until well done) would accumulate more of these carcinogens and therefore intake of well-done fish would increase risk to a greater extent than intake of lightly browned fish. However, we cannot disregard the hypothesis that when white fish is cooked for a long time, more fatty acid changes are accumulated that may also contribute to PCA risk, as we discussed previously (10). The fact that we failed to find interactions for any of the other metabolism enzymes investigated seems to offer more support for the latter rather than the former hypothesis. However, given that we did not include in our study all potential enzymes that participate in HCA and PAH metabolism, we cannot dismiss a role for these carcinogens in the observed association between fish and PCA risk.

Strengths of our study include the use of population-based cancer registries for the ascertainment of cases, the oversampling of advanced cases, the inclusion of three racial/ethnic groups, and the consideration of different types of fish (dark and white) and various cooking methods and doneness levels. Weaknesses of our study are the inclusion of a limited number of metabolic enzymes that do not comprehensively cover the HCA and PAH metabolism pathways and the inclusion of selected polymorphisms instead of a comprehensive tagSNP or sequencing approach. Moreover, we genotyped only 4 of the 30 SNPs that accurately classify NAT2 haplotype (39). These four SNPs have been shown to infer NAT2 phenotype with an estimated 98.4% accuracy (38); however, not genotyping every SNP may have misclassified some ‘slow’ phenotypes as ‘fast’ (wildtype), thereby biasing the results toward the null. For these reasons, we cannot exclude the possibility that some of the genes we studied, for which we found no association, may indeed play a role in PCA risk. Although the study had a large sample size, small numbers in some subgroup analyses may explain our failure to find interactions with the genes investigated. Also, dietary questionnaires have been reported to result in an underestimation of HCA intake (36), which could attenuate gene–diet interactions toward the null. Additionally, we did not have information on the use of marinades or use of cooking pre-treatments (microwaving, boiling), which can reduce the amount of HCA formation (61) and may further contribute to exposure misclassification. Our analyses included advanced cases sampled from both LAC and SFBA, but localized cases were from LAC only, as the SFBA center did not collect blood for localized cases. Finally, given that at the time of interview there was no widespread knowledge that consumption of fish or certain fish cooking methods may be associated with PCA risk, differential misclassification due to recall bias among cases is unlikely.

This study reports associations between functional polymorphisms in three enzymes involved in HCA and PAH metabolism and risk of localized PCA. Furthermore, we also report that the association between diets high in fish and advanced PCA may be modified by a polymorphism that may affect activity of the PTGS2 enzyme, which is known to be involved in HCA and PAH metabolism as well as metabolism of PUFA found in fish. Overall, our finding of an interaction with PTGS2 supports a role for HCAs and PAHs, which accumulate in well-done fish, and/or fatty acid changes that occur when fish is cooked at high temperature in PCA carcinogenesis.

Supplementary Material

Supplementary Data:

Supplementary material

Funding

Prostate Cancer Foundation (to M.C.S.); National Institute of Environmental Health Sciences (5P30 ES07048); California Cancer Research Program (864A-8702-S3514 and 99-00527V-10182 to E.M.J. and 99-00524V-10258 to S.A.I.), under Interagency Agreement #97-12013 (University of California contract #98-00924V) with the Department of Health Services Cancer Research Program; National Cancer Institute, National Institutes of Health (R01CA84979 to S.A.I.). Cancer incidence data used in this publication have been collected by the Greater Bay Area Cancer Registry, of the Cancer Prevention Institute of California (formerly the Northern California Cancer Center), under contract N01-PC-35136 with the National Cancer Institute, National Institutes of Health, and with support of the California Cancer Registry, a project of the Cancer Surveillance Section, California Department of Health Services, under subcontract 1006128 with the Public Health Institute and the Los Angeles Cancer Surveillance Program of the University of Southern California with Federal funds from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services, under Contract No. N01-PC-35139, and the California Department of Health Services as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; Centers for Disease Control and Prevention (1U58DP000807-3).

Mention of trade names, commercial products, specific equipment or organizations does not constitute endorsement, guarantee or warranty by the State of California Department of Health Services or the U.S. Government, nor does it imply approval to the exclusion of other products. The views expressed in this publication represent those of the authors and do not necessarily reflect the position or policies of the Cancer Prevention Institute of California, the California Public Health Institute, the State of California Department of Health Services, or the U.S. Department of Health and Human Services.

Acknowledgements

We are grateful to the men who participated in this study, without whom this research would not be possible. We thank Mr. Andre Kim, Mr. Chris Yoon and Mrs. Sangeetha Rao for assistance with genotyping, data cleaning and literature review, respectively.

Conflict of Interest Statement: None declared.

References

1. Augustsson K., et al. A prospective study of intake of fish and marine fatty acids and prostate cancer. Cancer Epidemiol. Biomarkers Prev. (2003);12:64–67. [PubMed]
2. Chavarro J.E., et al. A 22-y prospective study of fish intake in relation to prostate cancer incidence and mortality. Am. J. Clin. Nutr. (2008);88:1297–1303. [PMC free article] [PubMed]
3. Pham T.M., et al. Fish intake and the risk of fatal prostate cancer: findings from a cohort study in Japan. Public Health Nutr. (2009);12:609–613. [PubMed]
4. Terry P., et al. Fatty fish consumption and risk of prostate cancer. Lancet. (2001);357:1764–1766. [PubMed]
5. Wendel M., et al. Anticancer actions of omega-3 fatty acids--current state and future perspectives. Anticancer Agents Med. Chem. (2009);9:457–470. [PubMed]
6. Richman E.L., et al. Intakes of meat, fish, poultry, and eggs and risk of prostate cancer progression. Am. J. Clin. Nutr. (2010);91:712–721. [PMC free article] [PubMed]
7. Schuurman A.G., et al. Animal products, calcium and protein and prostate cancer risk in The Netherlands Cohort Study. Br. J. Cancer. (1999);80:1107–1113. [PMC free article] [PubMed]
8. Sonoda T., et al. A case-control study of diet and prostate cancer in Japan: possible protective effect of traditional Japanese diet. Cancer Sci. (2004);95:238–242. [PubMed]
9. Allen N.E., et al. A prospective study of diet and prostate cancer in Japanese men. Cancer Causes Control. (2004);15:911–920. [PubMed]
10. Joshi A.D., et al. Fish intake, cooking practices, and risk of prostate cancer: results from a multi-ethnic case-control study. Cancer Causes Control. (2012);23:405–420. [PubMed]
11. International Agency for Research on Cancer (IARC) (1993). Summaries and evaluations PhIP (2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine) (Group 2B), vol. 56
12. Shirai T., et al. The prostate: a target for carcinogenicity of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) derived from cooked foods. Cancer Res. (1997);57:195–198. [PubMed]
13. Stuart G.R., et al. Prostate mutations in rats induced by the suspected human carcinogen 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine. Cancer Res. (2000);60:266–268. [PubMed]
14. Rybicki B.A., et al. Polycyclic aromatic hydrocarbon-DNA adducts in prostate cancer. Cancer Res. (2004);64:8854–8859. [PubMed]
15. Aeschbacher H.U., et al. Effect of heterocyclic amines and beef extract on chromosome aberrations and sister chromatid exchanges in cultured human lymphocytes. Carcinogenesis. (1989);10:429–433. [PubMed]
16. Moonen H.J., et al. Generation of free radicals and induction of DNA adducts by activation of heterocyclic aromatic amines via different metabolic pathways in vitro. Mol. Carcinog. (2002);35:196–203. [PubMed]
17. Williams J.A., et al. Metabolic activation of carcinogens and expression of various cytochromes P450 in human prostate tissue. Carcinogenesis. (2000);21:1683–1689. [PubMed]
18. Di Paolo O.A., et al. Expression of cytochromes P450 and glutathione S-transferases in human prostate, and the potential for activation of heterocyclic amine carcinogens via acetyl-coA-, PAPS- and ATP-dependent pathways. Int. J. Cancer. (2005);117:8–13. [PubMed]
19. Shimada T., et al. Metabolic activation of polycyclic aromatic hydrocarbons to carcinogens by cytochromes P450 1A1 and 1B1. Cancer Sci. (2004);95:1–6. [PubMed]
20. Bock K.W., et al. Functions and transcriptional regulation of PAH-inducible human UDP-glucuronosyltransferases. Drug Metab. Rev. (1999);31:411–422. [PubMed]
21. Kirschenbaum A., et al. Expression of cyclooxygenase-1 and cyclooxygenase-2 in the human prostate. Urology. (2000);56:671–676. [PubMed]
22. Wang W., et al. Chronic inflammation in benign prostate hyperplasia is associated with focal upregulation of cyclooxygenase-2, Bcl-2, and cell proliferation in the glandular epithelium. Prostate. (2004);61:60–72. [PubMed]
23. Rybicki B.A., et al. Polycyclic aromatic hydrocarbon-DNA adduct formation in prostate carcinogenesis. Cancer Lett. (2006);239:157–167. [PMC free article] [PubMed]
24. Strange R.C., et al. The glutathione S-transferases: influence of polymorphism on cancer susceptibility. IARC Sci. Publ. (1999);148:231–249. [PubMed]
25. Srivastava D.S., et al. Genetic polymorphism of glutathione S-transferase genes (GSTM1, GSTT1 and GSTP1) and susceptibility to prostate cancer in Northern India. BJU Int. (2005);95:170–173. [PubMed]
26. Papafili A., et al. Common promoter variant in cyclooxygenase-2 represses gene expression: evidence of role in acute-phase inflammatory response. Arterioscler. Thromb. Vasc. Biol. (2002);22:1631–1636. [PubMed]
27. Sachse C., et al. Polymorphisms in the cytochrome P450 CYP1A2 gene (CYP1A2) in colorectal cancer patients and controls: allele frequencies, linkage disequilibrium and influence on caffeine metabolism. Br. J. Clin. Pharmacol. (2003);55:68–76. [PMC free article] [PubMed]
28. Hassett C., et al. Human hepatic microsomal epoxide hydrolase: comparative analysis of polymorphic expression. Arch. Biochem. Biophys. (1997);337:275–283. [PubMed]
29. Tang Y.M., et al. Human CYP1B1 Leu432Val gene polymorphism: ethnic distribution in African-Americans, Caucasians and Chinese; oestradiol hydroxylase activity; and distribution in prostate cancer cases and controls. Pharmacogenetics. (2000);10:761–766. [PubMed]
30. MacLeod S.L., et al. (2000). Genetic polymorphisms. In Nagao M., Sugimura H., editors. (eds) Food Borne Carcinogens: Heterocyclic Amines John Wiley & Sons; West Sussex, UK: pp. 112 130
31. Hirvonen A. Polymorphic NATs and cancer predisposition. IARC Sci. Publ. (1999):251–270. [PubMed]
32. John E.M., et al. Sun exposure, vitamin D receptor gene polymorphisms, and risk of advanced prostate cancer. Cancer Res. (2005);65:5470–5479. [PubMed]
33. Schwartz G.G., et al. Prostate cancer in African-American men and polymorphism in the calcium-sensing receptor. Cancer Biol. Ther. (2010);9:994–999. [PubMed]
34. Pike M.C., et al. Estrogen-progestin replacement therapy and endometrial cancer. J. Natl. Cancer Inst. (1997);89:1110–1116. [PubMed]
35. John E.M., et al. Meat consumption, cooking practices, meat mutagens, and risk of prostate cancer. Nutr. Cancer. (2011);63:525–537. [PMC free article] [PubMed]
36. Sinha R., et al. Development of a food frequency questionnaire module and databases for compounds in cooked and processed meats. Mol. Nutr. Food Res. (2005);49:648–655. [PubMed]
37. Willett W.C., et al. (1997). Adjustment for total energy intake in epidemiologic studies Am. J. Clin. Nutr. 65 1220S 1228S; discussion 1229S 1231S [PubMed]
38. Hein D.W., et al. Accuracy of various human NAT2 SNP genotyping panels to infer rapid, intermediate and slow acetylator phenotypes. Pharmacogenomics. (2012);13:31–41. [PMC free article] [PubMed]
39. Hein D.W., et al. Update on consensus arylamine N-acetyltransferase gene nomenclature. Pharmacogenetics. (2000);10:291–292. [PubMed]
40. Gauderman W.J. Sample size requirements for association studies of gene-gene interaction. Am. J. Epidemiol. (2002);155:478–484. [PubMed]
41. Carnell D.M., et al. Target validation of cytochrome P450 CYP1B1 in prostate carcinoma with protein expression in associated hyperplastic and premalignant tissue. Int. J. Radiat. Oncol. Biol. Phys. (2004);58:500–509. [PubMed]
42. Martin F.L., et al. Constitutive expression of bioactivating enzymes in normal human prostate suggests a capability to activate pro-carcinogens to DNA-damaging metabolites. Prostate. (2010);70:1586–1599. [PubMed]
43. Fukatsu T., et al. Genetic polymorphisms of hormone-related genes and prostate cancer risk in the Japanese population. Anticancer Res. (2004);24:2431–2437. [PubMed]
44. Beuten J., et al. CYP1B1 variants are associated with prostate cancer in non-Hispanic and Hispanic Caucasians. Carcinogenesis. (2008);29:1751–1757. [PMC free article] [PubMed]
45. Pastorelli R., et al. Impact of inherited polymorphisms in glutathione S-transferase M1, microsomal epoxide hydrolase, cytochrome P450 enzymes on DNA, and blood protein adducts of benzo(a)pyrene-diolepoxide. Cancer Epidemiol. Biomarkers Prev. (1998);7:703–709. [PubMed]
46. Penning T.M., et al. Dihydrodiol dehydrogenases and polycyclic aromatic hydrocarbon activation: generation of reactive and redox active o-quinones. Chem. Res. Toxicol. (1999);12:1–18. [PubMed]
47. Mittal R.D., et al. Cytochrome P4501A1 and microsomal epoxide hydrolase gene polymorphisms: gene-environment interaction and risk of prostate cancer. DNA Cell Biol. (2007);26:791–798. [PubMed]
48. Figer A., et al. Analysis of polymorphic patterns in candidate genes in Israeli patients with prostate cancer. Isr. Med. Assoc. J. (2003);5:741–745. [PubMed]
49. Mo Z., et al. An updating meta-analysis of the GSTM1, GSTT1, and GSTP1 polymorphisms and prostate cancer: a HuGE review. Prostate. (2009);69:662–688. [PubMed]
50. Norskov M.S., et al. Copy number variation in glutathione-S-transferase T1 and M1 predicts incidence and 5-year survival from prostate and bladder cancer, and incidence of corpus uteri cancer in the general population. Pharmacogenomics J. (2011);11:292–299. [PubMed]
51. Souiden Y., et al. Polymorphisms of glutathione-S-transferase M1 and T1 and prostate cancer risk in a Tunisian population. Cancer Epidemiol. (2010);34:598–603. [PubMed]
52. Safarinejad M.R., et al. Glutathione S-transferase gene polymorphisms (GSTM1, GSTT1, GSTP1) and prostate cancer: a case-control study in Tehran, Iran. Prostate Cancer Prostatic Dis. (2011);14:105–113. [PubMed]
53. Taioli E., et al. Multi-institutional prostate cancer study of genetic susceptibility in populations of African descent. Carcinogenesis. (2011);32:1361–1365. [PMC free article] [PubMed]
54. Steinbrecher A., et al. Dietary glucosinolate intake, polymorphisms in selected biotransformation enzymes, and risk of prostate cancer. Cancer Epidemiol. Biomarkers Prev. (2010);19:135–143. [PubMed]
55. Lavender N.A., et al. Examination of polymorphic glutathione S-transferase (GST) genes, tobacco smoking and prostate cancer risk among men of African descent: a case-control study. BMC Cancer. (2009);9:397. [PMC free article] [PubMed]
56. Fradet V., et al. Dietary omega-3 fatty acids, cyclooxygenase-2 genetic variation, and aggressive prostate cancer risk. Clin. Cancer Res. (2009);15:2559–2566. [PMC free article] [PubMed]
57. Hedelin M., et al. Association of frequent consumption of fatty fish with prostate cancer risk is modified by COX-2 polymorphism. Int. J. Cancer. (2007);120:398–405. [PubMed]
58. Lim K., et al. Omega-3 polyunsaturated fatty acids inhibit hepatocellular carcinoma cell growth through blocking beta-catenin and cyclooxygenase-2. Mol. Cancer Ther. (2009);8:3046–3055. [PMC free article] [PubMed]
59. Cui J.G., et al. Characterization of a cyclooxygenase-2-765G-->C promoter polymorphism in human neural cells. Neuroreport. (2005);16:575–579. [PubMed]
60. Sanak M., et al. Functional promoter polymorphism of cyclooxygenase-2 modulates the inflammatory response in stable coronary heart disease. Pol. Arch. Med. Wewn. (2010);120:82–88. [PubMed]
61. Salmon C.P., et al. Effects of marinating on heterocyclic amine carcinogen formation in grilled chicken. Food Chem. Toxicol. (1997);35:433–441. [PubMed]

Articles from Carcinogenesis are provided here courtesy of Oxford University Press
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...