Gene-nutrient interactions in cancer etiology

Relationships between dietary components and cancer risk are often unclear, and the results from epidemiologic studies are inconsistent. While some inconsistencies could be due to study design issues, we propose that genetic heterogeneity of study populations could mask associations. In this report, we review the literature regarding meat consumption and risk of colon, breast, and prostate cancers, particularly in relation to phenotypes and genotypes for enzymes that metabolize food-borne promutagens. The role of consumption of fruits and vegetables, as well as the role of genetic variants in oxidative stress genes, in the risk of breast cancer are also discussed.

Cruciferous Vegetables, Genetic Polymorphisms in Glutathione S-Transferases M1 and T1, and Prostate Cancer Risk

Cruciferous vegetables contain anticarcinogenic isothiocyanates (ITCs), particularly the potent sulforaphane, which may decrease risk of prostate cancer through induction of phase II enzymes, including glutathione S-transferases (GSTs). We evaluated this hypothesis in a population-based, case-control study of prostate cancer, including 428 men with incident prostate cancer and 537 community controls. An in-person interview included an extensive food-frequency questionnaire. Genotyping for deletions in GSTM1 and GSTT1 was performed in a subset of men who provided blood. Intakes of cruciferous vegetables and of broccoli, the greatest source of sulforaphane, were associated with decreased prostate cancer risk at all levels above the lowest consumers [adjusted 4th quartile odds ratio (OR)=0.58; 95% confidence interval (CI)=0.38, 0.89, and 0.72 (95% CI=0.49, 1.06)], respectively. In relation to genotypes, there was a nonsignificant increase in risk with the GSTT1 null genotype (OR=1.51; 95% CI=0.98, 2.31) but no effects of GSTM1 genotype. However, men with GSTM1-present genotype and high broccoli intake had the greatest reduction in risk (OR=0.49; 95% CI=0.27, 0.89). Our findings provide evidence that two or more servings per month of cruciferous vegetables may reduce risk of prostate cancer, especially among men with GSTM1-present alleles, and are consistent with a role of dietary ITCs as chemopreventive agents against prostate cancer.

Myeloperoxidase Genotype, Fruit and Vegetable Consumption, and Breast Cancer Risk

Myeloperoxidase (MPO), an antimicrobial enzyme in the breast, generates reactive oxygen species (ROS) endogenously. An MPO G463A polymorphism exists in the promoter region, with the variant A allele conferring lower transcription activity than the common G allele. Because oxidative stress may play a role in breast carcinogenesis, we evaluated MPO genotypes in relation to breast cancer risk among 1,011 cases and 1,067 controls from the Long Island Breast Cancer Study Project (1996-1997). We also assessed the potential modifying effects of dietary antioxidants and hormonally related risk factors on these relationships. Women over 20 years with incident breast cancer who were residents of Nassau and Suffolk Counties, NY, were identified as potential cases. Population-based controls were frequency matched by 5-year age groups. Genotyping was performed with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF) technology, and suspected breast cancer risk factors and usual dietary intake were assessed during an in-person interview. Unconditional logistic regression was used to estimate odds ratios and 95% confidence intervals. Having at least one A allele was associated with an overall 13% reduction in breast cancer risk. When consumption of fruits and vegetables and specific dietary antioxidants were dichotomized at the median, inverse associations with either GA or AA genotypes were most pronounced among women who consumed higher amounts of total fruits and vegetables (odds ratio, 0.75; 95% confidence interval, 0.58-0.97); this association was not noted among the low-consumption group (P for interaction = 0.04). Relationships were strongest among premenopausal women. Results from this first study of MPO genotypes and breast cancer risk indicate that MPO variants, related to reduced generation of ROS, are associated with decreased breast cancer risk, and emphasize the importance of fruit and vegetable consumption in reduction of breast cancer risk.

STK15 polymorphism and breast cancer risk in a population-based study

STK15 is considered a potential cancer susceptibility gene owing to its functions in normal cell mitosis. Two common coding region polymorphisms in the gene (F31I and V57I) may affect ubiquitin-dependent degradation and thus the half-life of the encoded protein. There are limited data on the relevance of these polymorphisms to population cancer rates. To examine whether functional variation in STK15 may affect breast cancer risk, we genotyped a large series of incident breast cancer cases (n = 941) and age-matched population controls (n = 830) for the F31I and V57I polymorphisms. Individually, neither the F31I polymorphism [odds ratio (OR) 1.54; 95% confidence interval (CI) 0.96-2.47, comparing 31I with 31F homozygotes] nor the V57I polymorphism (OR 0.92; 95% CI 0.50-1.71, comparing 57I with 57V homozygotes) was significantly associated with breast cancer risk. A relatively common genotype, combining the two polymorphisms (31I-57V/31I-57V, 3% of controls) was related to a significant 2-fold increase in the risk of post-menopausal breast cancer (OR 1.96; 95% CI 1.01-3.79). No interaction was detected between STK15 variants and estrogenic risk factors, although the power of these analyses was limited. These results suggest that STK15 may represent a low penetrance type breast cancer susceptibility gene.

Genetic variants in SOD2, MPO, and NQO1, and risk of ovarian cancer

OBJECTIVE

One way in which parity and use of oral contraceptives may protect against ovarian cancer is by preventing inflammation and oxidative stress associated with ovulation. Since the genes superoxide dismutase (SOD2), myeloperoxidase (MPO), and NAD(P)H:quinone oxidoreductase 1 (NQO1) are involved in inflammation and oxidative stress, we investigated whether variants of these genes are associated with risk of ovarian cancer.

METHODS

In a hospital-based case-control study, we compared 125 cases and 193 controls with respect to prevalence of (1) the T-->C (val-->ala) substitution at the -9 position in the signal sequence of SOD2; (2) the G-->A substitution at the -463 position in the promoter region of MPO; and (3) the C-->T (pro-->ser) change in exon 6 of NQO1. Genotyping was done using PCR and gel electrophoresis for MPO and NQO1 and using MALDI-TOF mass spectrometry for SOD2.

RESULTS

For SOD2, women with the TC (val/ala) or CC (ala/ala) genotypes were at increased risk [odds ratio (OR) 2.1, 95% confidence interval (CI) 1.1-4.0]. Results for MPO and NQO1 were in the hypothesized directions but were not statistically significant. For MPO, there was a small inverse association among women with GA or AA genotypes (OR = 0.72, 95% CI 0.43-1.2). For NQO1, the TT (ser/ser) genotype was associated with somewhat increased risk (OR = 2.3, 95% CI 0.69-7.6).

CONCLUSIONS

While these results need to be confirmed in other studies, they point to a possible role for genes involved in oxidative stress in the development of ovarian cancer.

Association of SULT1A1 phenotype and genotype with prostate cancer risk in African-Americans and Caucasians

Exposure to heterocyclic amines may increase prostate cancer risk. Human sulfotransferase 1A1 (SULT1A1) is involved in the bioactivation of some dietary procarcinogens, including the N-hydroxy metabolite of the food-borne heterocyclic amine, 2-amino-1-methyl-6-phenylimidazo(4,5-b) pyridine. This study compares a polymorphism in the SULT1A1 gene, SULT1A1 enzyme activity, meat consumption, and the risk of prostate cancer in a population based case-control study. Prostate cancer patients (n = 464) and control individuals (n = 459), frequency matched on age and ethnicity, provided informed consent, answered a survey, and provided a blood sample. Platelets were isolated for phenotype analysis, and DNA was isolated from lymphocytes for genotype determination. Meat consumption was assessed using a dietary questionnaire. Caucasians homozygous for the SULT1A1*1 high activity allele were at increased risk for prostate cancer [odds ratio (OR), 1.68; 95% confidence interval (CI), 1.05-2.68] compared with individuals homozygous for the low-activity allele. The association between SULT1A1 genotype and prostate cancer risk in African-Americans did not reach significance (OR, 1.60; 95% CI, 0.46-5.62). When SULT1A1 activity was considered, there was a strong association between increased SULT1A1 activity and prostate cancer risk in Caucasians (OR, 3.04; 95% CI, 1.8-5.1 and OR, 4.96; 95% CI, 3.0-8.3, for the second and third tertiles of SULT1A1 activity, respectively) compared with individuals in the low enzyme activity tertile. A similar association was also found in African-American patients, with ORs of 6.7 and 9.6 for the second and third tertiles of SULT1A1 activity (95% CI, 2.1-21.3 and 2.9-31.3, respectively). When consumption of well-done meat was considered, there was increased risk of prostate cancer (OR, 1.42; 95% CI, 1.01-1.99 and OR, 1.68; 95% CI, 1.20-2.36 for the second and third tertiles, respectively). When SULT1A1 activity was stratified by tertiles of meat consumption, there was greater risk of prostate cancer in the highest tertile of meat consumption. These results indicate that variations in SULT1A1 activity contributes to prostate cancer risk and the magnitude of the association may differ by ethnicity and be modified by meat consumption.

Breast cancer risk in premenopausal women is inversely associated with consumption of broccoli, a source of isothiocyanates, but is not modified by GST genotype

The role of vegetable consumption in relation to breast cancer risk is controversial. Anticarcinogenic compounds may be present only in specific vegetables, thereby attenuating findings for total vegetable intake. Cruciferous vegetables contain precursors of isothiocyanates (ITCs), which may be chemopreventive through potent inhibition of phase I, and induction of phase II enzymes, such as glutathione S-transferases (GSTs). We investigated associations between consumption of cruciferous vegetables, sources of ITCs, and breast cancer risk, and potential modification of relations by GSTM1 and GSTT1 genotypes. Cases (n = 740) were Caucasian women with incident breast cancer identified from all major hospitals in Erie and Niagara counties. Community controls (n = 810) were frequency matched to cases by age and county. An in-depth interview including a validated FFQ was administered in person. Odds ratios (ORs) and 95% CIs were used to estimate relative risks. Consumption of cruciferous vegetables, particularly broccoli, was marginally inversely associated with breast cancer risk in premenopausal women [4th quartile OR = 0.6, 95% CI (0.40-1.01), P = 0.058]. Associations were weaker or null among postmenopausal women. No significant effects of GST genotype on risk were observed in either menopausal group. These data indicate that cruciferous vegetables may play an important role in decreasing the risk of premenopausal breast cancer.

Reproductive factors in relation to breast cancer characterized by p53 protein expression (United States)

OBJECTIVE

To evaluate the potential etiologic heterogeneity of breast cancer by examining whether associations with reproductive and other personal characteristics differed by p53 protein expression status.

METHODS

Data from the Carolina Breast Cancer Study, a population-based, case-control study of 861 cases and 790 controls, were utilized. Immunohistochemical staining for the p53 protein was performed on 638 archived tumor specimens; 46% of cases were classified as p53+. Two separate unconditional logistic regression models were used to calculate odds ratios (OR) and 95% confidence intervals (CI) for p53+ and p53- breast cancer relative to controls for reproductive and other personal characteristics. Analyses were performed separately for younger (< or = 45 years) and older (>45 years) women.

RESULTS

Risk factor profiles largely overlapped for p53+ and p53- breast cancer, with the exception of oral contraceptive (OC) use among younger women and a family history of breast cancer. Prolonged OC use was more strongly associated with p53+ breast cancer [OR 3.1 (95% CI: 1.2-8.1) than p53- breast cancer (OR 1.3 (95% CI: 0.6-3.2)] among younger women only. A first-degree family history of breast cancer was associated with p53+ breast cancer among younger women [OR 1.5 (95% CI: 1.0-2.2)] and older women [OR 1.4 (95% CI: 0.9-2.3)], but not p53- breast cancer in either age-group.

CONCLUSIONS

These results provide little evidence of breast cancer heterogeneity as classified by p53 expression status. However, although not statistically significant, OC use among younger women and family history of breast cancer may operate through a pathway involving p53 alterations to increase risk of breast cancer.

Diet and alcohol consumption in relation to p53 mutations in breast tumors

There is evidence linking alcohol consumption to p53 mutations in tumors, considerable evidence linking alcohol consumption with risk of breast cancer and some evidence that alcohol and folate consumption interact to affect risk. Further, while there is some indication that oxidation may play a role in breast cancer etiology, there has been little examination of an association of oxidative stress with p53 mutations. We examined several dietary components related to one-carbon metabolism and antioxidants to determine if these factors were related to the prevalence of p53 mutations in breast tumors. We conducted a case-control study of primary, histologically confirmed breast cancer in western New York. Controls <65 were selected from drivers license lists; those > or =65 were selected from Health Care Finance Administration lists. p53 mutations in archived tumor blocks were identified in exons 2-11 and flanking intron sequences. Usual dietary intake was assessed by interview regarding intake in the previous 2 years; alcohol consumption was queried for 2, 10 and 20 years in the past. Our data were consistent with increased likelihood of tumors with p53 mutations for premenopausal breast cancer with increased alcohol intake 10 or 20 years previous; for intake of 16 or more drinks per month in the period 20 years before the interview compared with non-drinkers, the OR was 5.25, 95% CI 1.48-18.58. For postmenopausal women, there was increased likelihood of tumors with p53 mutations among women with higher folate. Antioxidant nutrients were not differentially related to p53 mutations. These results indicate that there may be heterogeneity in breast tumors, as indicated by differences in associations for those with or without p53 mutations, and that causal pathways for these nutrients may vary for pre- and postmenopausal women. For premenopausal women, alcohol consumption in the past was associated with p53 mutations.

MnSOD polymorphism and breast cancer in a population-based case-control study

A polymorphism in the signal sequence (Ala-9Val) of the gene encoding the free radical-quenching manganese superoxide dismutase (MnSOD) has been reported to alter the risk for breast cancer. We evaluated this relationship in a population-based case-control study (476 breast cancer cases and 502 controls). Overall, relative risks were not significantly elevated in women with one (RR: 1.27; 95% CI: 0.91-1.77) or two (RR: 1.18; 95% CI: 0.81-1.73) Ala alleles compared to those homozygous for the wild-type Val genotype. Results do not support any overall association of the Ala-9Val MnSOD polymorphism to the development of breast cancer.

Association between a glutathione S-transferase A1 promoter polymorphism and survival after breast cancer treatment

Glutathione S-transferase (GST) enzymes detoxify chemotherapeutic drugs, and several studies have reported differences in survival for cancer patients who have variant genotypes for GSTP1, GSTM1 or GSTT1 enzymes. A recently described polymorphism alters hepatic expression of GSTA1, a GST with high activity in glutathione conjugation of metabolites of cyclophosphamide (CP). To consider the possible influence of the reduced-expression GSTA1*B allele on cancer patient survival, we have conducted a pilot study of breast cancer patients treated with CP-containing combination chemotherapy. GSTA1 genotype was determined by polymerase chain reaction and restriction fragment length polymorphism. Kaplan-Meier methods and Cox proportional hazards models were used to evaluate survival in relation to genotype. Among 245 subjects, 35% were GSTA1*A/*A, 49% GSTA1*A/*B and 16% GSTA1*B/*B; the genotype distribution did not differ by ethnic group, age or stage at diagnosis. Among patients who had 0 or 1 GSTA1*B allele, the proportion surviving at 5 years was 0.66 (95% CI = 0.59-0.72), whereas for GSTA1*B/*B subjects the proportion was higher, 0.86 (95% CI = 0.67-0.95). Significantly reduced hazard of death was observed for GSTA1*B/*B subjects during the first 5 years after diagnosis, hazard ratio (HR) = 0.3, 95% CI = 0.1-0.8. The association varied with time, with no survival difference observed for subjects who survived beyond 5 years. These results, although based on a small study population, describe an apparent difference in survival after treatment for breast cancer according to GSTA1 genotype. Further studies should consider the possible association between the novel GSTA1*B variant and outcomes of cancer therapy.

CYP17 genetic polymorphism, breast cancer, and breast cancer risk factors

BACKGROUND

Findings from previous studies regarding the association between the CYP17 genotype and breast cancer are inconsistent. We investigated the role of the MspAI genetic polymorphism in the 5' region of CYP17 on risk of breast cancer and as a modifier of reproductive risk factors.

METHODS

Questionnaire and genotyping data were obtained from a population-based, case-control study of premenopausal (n = 182) and postmenopausal (n = 214) European-American Caucasian women in western New York. Cases and controls were frequency matched by age and by county of residence. Odds ratios and 95% confidence intervals were used to estimate relative risks.

RESULTS

The CYP17 genotype was not associated with breast cancer risk; however, controls with the A2/A2 genotype (associated with higher estrogens) had earlier menarche and earlier first full-term pregnancy. Premenopausal women with A1/A1 genotypes, but not with A2 alleles, were at significantly decreased risk with late age at menarche (odds ratio = 0.37, 95% confidence interval = 0.14-0.99), and at increased risk with late age at first full-term pregnancy (odds ratio = 4.30, 95% confidence interval = 1.46-12.67) and with use of oral contraceptives (odds ratio = 3.24, 95% confidence interval = 1.08-9.73). Associations were weaker among postmenopausal women.

CONCLUSION

These results suggest that the effects of factors that may alter breast cancer risk through a hormonal mechanism may be less important among premenopausal women with putative higher lifetime exposures to circulating estrogens related to the CYP17 A2 allele.

Association between sulfotransferase 1A1 genotype and survival of breast cancer patients receiving tamoxifen therapy

BACKGROUND

Human sulfotransferase 1A1 (SULT1A1) catalyzes the sulfation of a variety of phenolic and estrogenic compounds, including 4-hydroxytamoxifen (4-OH TAM), the active metabolite of tamoxifen. A functional polymorphism in exon 7 of the SULT1A1 gene (SULT1A1*2) has been described that generates an enzyme that has approximately twofold lower activity and is less thermostable than that of the common allele SULT1A1*1. We investigated the hypothesis that that high sulfation activity would increase the elimination of 4-OH TAM by examining whether the presence of this polymorphism affects the efficacy of tamoxifen therapy.

METHODS

We examined the relationship between the SULT1A1*2 allele and survival in a cohort of 337 women with breast cancer who received tamoxifen (n = 160) or who did not (n = 177). SULT1A1 genotype was determined by restriction fragment polymorphism analysis. Patient survival was evaluated according to SULT1A1 genotype using Kaplan-Meier survival functions. Hazard ratios (HRs) were calculated from adjusted Cox proportional hazards modeling. All statistical tests were two-sided.

RESULTS

Among tamoxifen-treated patients, those who were homozygous for the SULT1A1*2 low-activity allele had approximately three times the risk of death (HR = 2.9, 95% confidence interval [CI] = 1.1 to 7.6) as those who were homozygous for the common allele or those who were heterozygous (SULT1A1*1/*2). Among patients who did not receive tamoxifen, there was no association between survival and SULT1A1 genotype (HR = 0.7, 95% CI = 0.3 to 1.5).

CONCLUSIONS

Sulfation of 4-OH TAM provides a previously unanticipated benefit, possibly due to alterations in the bioavailability of the active metabolite or to undefined estrogen receptor-mediated events. These data alternatively suggest that variability in the metabolism of tamoxifen may affect its efficacy.

The risk of breast cancer associated with dietary lignans differs by CYP17 genotype in women

Lignans are plant compounds metabolized in the gut to produce the phytoestrogens enterolactone and enterodiol. Reduced breast cancer risks associated with higher urinary lignan excretion may be related to competitive inhibition of endogenous estrogens. Evidence exists that associations with reproductive risk factors for breast cancer differ according to cytochrome P450c17alpha (CYP17) genotype. Genetic variability in estrogen metabolism could affect lignan metabolism thereby modifying risk associations. We examined breast cancer risk, dietary lignans and CYP17 genotype among 207 women with primary, incident, histologically confirmed breast cancer and 188 controls frequency matched to cases by age and county of residence. Self-reported frequency of intake of 170 foods and beverages during the 2 y before the interview and other relevant data were collected by detailed in-person interviews. Dietary lignan intake was expressed as the sum of enterolactone and enterodiol production from foods. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by unconditional logistic regression, adjusting for age, education and other breast cancer risk factors. Women in the highest tertile of dietary lignans tended to have reduced breast cancer risk (OR 0.45, 95% CI 0.20-1.01 and OR 0.59, 95% CI 0.28-1.27, pre- and postmenopausal women, respectively). Substantially reduced risks in the highest tertile of lignans were observed for premenopausal women with at least one A2 allele (OR 0.12, 95% CI 0.03-0.50). Our results suggest that CYP17 genotype may be important in modifying the effect on breast cancer risk of exogenous estrogens, particularly for premenopausal women.

Analysis of total meat intake and exposure to individual heterocyclic amines in a case-control study of colorectal cancer: contribution of metabolic variation to risk

A case-control study of colorectal cancer, consisting of 157 cases and 380 controls matched by sex, ethnicity, decade of age and county of residence was performed to explore the associations between environmental exposure, metabolic polymorphisms and cancer risk. Participants were required to provide a blood sample, undergo caffeine phenotyping and complete an in-person interview that evaluated meat consumption, cooking methods and degree of doneness. A color atlas of foods cooked to different degrees of doneness was used to estimate food preparation techniques and food models were used to estimate serving portion sizes. Data was analyzed using a reference database of heterocyclic amine (HCA) exposure based on the food preferences chosen from the atlas. Data regarding individual food items cooked to different levels of doneness, as well as summary variables of foods and of food groups cooked to different degrees of doneness were also evaluated in a univariate analysis for association with colorectal cancer case status. Three measures of metabolic variation, hGSTA1 genotype, SULT1A1 genotype and the phenotype for CYP2A6 were also evaluated for possible association with colon cancer. While higher exposure to HCAs was strongly associated with colorectal cancer risk, increased consumption of five red meats cooked well done or very well done produced comparable odds ratios (OR) for colorectal cancer risk (OR=4.36, 95% CI 2.08-9.60) for the highest quartile of exposure. Similarly, individuals in the most rapid CYP2A6 phenotype quartile showed an odds ratio (OR = 4.18, 95% CI 2.03-8.90). The ORs for the low activity hGSTA1 and low activity SULT1A1 alleles were 2.0, 95% CI 1.0-3.7 and 0.6, 95% CI 0.3-1.1, respectively. Individual measures of specific HCAs provided little improvement in risk assessment over the measure of meat consumption, suggesting that exposure to other environmental or dietary carcinogens such as nitrosamines or undefined HCAs may contribute to colorectal cancer risk.

Environmental factors in relation to breast cancer characterized by p53 protein expression

Findings from studies of cigarette smoking and low-dose ionizing radiation exposure and breast cancer are unclear. Laboratory studies indicate that both exposures can cause DNA damage, potentially increasing cancer risk if such mutations occur in growth control genes, such as p53. We examined the potential etiologic heterogeneity of breast cancer by evaluating whether associations between cigarette smoking and low-dose ionizing radiation and breast cancer differed by p53 protein expression status. Data were obtained from the Carolina Breast Cancer Study, a population-based, case-control study conducted among African-American and white women ages 20-74 years. Questionnaire data were available from 861 women with incident, primary invasive breast cancer and 790 community-based controls. p53 immunostaining was performed on tissue from 683 women with breast cancer; 46% were classified as p53+. Two separate unconditional logistic regression models were used to calculate odds ratios (ORs) for p53+ and p53- breast cancer, as compared with controls, in relation to smoking and low-dose ionizing radiation exposure. Smoking was not differentially associated with p53+ or p53- breast cancer, even when duration, dose, and passive smoking status were considered. Exposure to individual sources of radiation did not differ for p53+ and p53- breast cancers. However, ORs for combined exposure to chest X-rays and occupational radiation were higher for p53+ [OR, 2.2; 95% confidence interval (CI), 1.0-5.3] than p53- breast cancer (OR, 1.2; 95% CI, 0.5-3.4). Combined exposure to radiation from other medical sources as well as occupational exposure was also higher for p53+ (OR, 3.7; 95% CI, 0.8-16.8) than for p53- breast cancer (OR, 1.7; 95% CI, 0.3-10.5). Although preliminary, our results suggest that exposure to multiple sources of low-dose ionizing radiation may contribute to the development of p53+ breast cancer.

Exposures associated with serum organochlorine levels among postmenopausal women from western New York State

BACKGROUND

Organochlorines, such as 2,2-bis (4-chlorophenyl)-1,1-dichloroethylene (DDE), polychlorinated biphenyls (PCBs), and hexachlorobenzene (HCB) are lipophilic compounds that are ubiquitous in the environment and may cause adverse health effects in humans.

METHODS

We examined the correlation between epidemiological exposure variables and serum DDE, PCB, and HCB levels in a sample of 192 healthy, female postmenopausal western New York residents; a subset of the control group from a case-control study on breast cancer risk. Usual diet, reproductive and medical histories, and other lifestyle information were obtained by an extensive in person interview. Serum levels (ng/g) of DDE, HCB, and 69 PCB congeners were determined by gas chromatography with electron capture detection. Statistical analyses included computations of crude and age and lipid adjusted correlation coefficients, as well as multiple linear regression analysis.

RESULTS

Results indicated that the strongest predictors for serum DDE levels were age, serum lipids, parity, and fruit and dairy consumption. Statistically significant predictors of PCB levels included age, serum lipids, parity, and fish consumption. Serum HCB levels were related to age, serum lipids, and fruit and red meat consumption.

CONCLUSIONS

Our findings are consistent with previous investigations that reported strong associations between organochlorine levels and age and serum lipids. The absence of other data showing an association between fruit intake and organochlorine levels make the importance of this finding unclear.

Evidence for the presence of mutagenic arylamines in human breast milk and DNA adducts in exfoliated breast ductal epithelial cells

Aromatic and heterocyclic amines are ubiquitous environmental mutagens present in combustion emissions, fried meats, and tobacco smoke, and are suspect human mammary carcinogens. To determine the presence of arylamines in breast tissue and fluid, we examined exfoliated breast ductal epithelial cells for DNA adducts and matched human milk samples for mutagenicity. Breast milk was obtained from 50 women who were 4-6 weeks postpartum, and exfoliated epithelial-cell DNA was evaluated for bulky, nonpolar DNA adducts by (32)P-postlabeling and thin-layer chromatography. Milk was processed by acid hydrolysis, and the extracted organics were examined in the standard plate-incorporation Ames Salmonella assay using primarily strain YG1024, which detects frameshift mutations and overexpresses aryl amine N-acetyltransferase. DNA adducts were identified in 66% of the specimens, and bulky adducts migrated in a pattern similar to that of 4-aminobiphenyl standards. The distribution of adducts did not vary by NAT2 genotype status. Of whole milk samples, 88% (22/25) had mutagenic activity. Among the samples for which we had both DNA adduct and mutagenicity data, 58% (14/19) of the samples with adducts were also mutagenic, and 85% (11/13) of the mutagenic samples had adducts. Quantitatively, no correlation was observed between the levels of adducts and the levels of mutagenicity. Separation of the milk showed that mutagenic activity was found in 69% of skimmed milk samples but in only 29% of the corresponding milk fat samples, suggesting that the breast milk mutagens were moderately polar molecules. Chemical fractionation showed that mutagenic activity was found in 67% (4/6) of the basic fractions but in only 33% (2/6) of acidic samples, indicating that the mutagens were primarily basic compounds, such as arylamines. Although pilot in nature, this study corroborates previous findings of significant levels of DNA adducts in breast tissue and mutagenicity in human breast milk and indicates that breast milk mutagens may be moderately polar basic compounds, such as arylamines.

Carcinogen-DNA adducts in human breast epithelial cells

Diet and environmental exposures are often regarded as significant etiologic factors in human breast cancer. Chemicals that may be involved in these exposures include heterocyclic amines, aromatic amines, and polycyclic aromatic hydrocarbons, which also serve as strong mammary carcinogens in different animal models. In this study, we chose to quantify the major DNA adducts derived from one member of each of these classes of carcinogens, that is, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), 4-aminobiphenyl (ABP), and benzo[a]pyrene (B[a]P), respectively, in DNA isolated from exfoliated ductal epithelial cells in human breast milk. Milk was collected from healthy, nonsmoking mothers. The isolated DNA was digested to 3' nucleotides and subjected to (32)P-postlabeling. Adduct enrichment was achieved using Oasis Sep-Paks and the analyses were conducted by HPLC using radiometric detection. Critical to the analyses were the syntheses of bis(phosphate) standards for the C8-dG adducts of PhIP and ABP, and the N(2)-dG adduct of B[a]P, which were added to each reaction as UV markers. Of the 64 samples analyzed, adducts were found in 31 samples. Thirty samples contained detectable levels of PhIP adducts, with a mean value of 4.7 adducts/10(7) nucleotides; 18 were positive for ABP adducts with a mean value of 4.7 adducts/10(7) nucleotides; and 13 were found to contain B[a]P adducts with a mean level of 1.9 adducts/10(7) nucleotides. These data indicate that women are exposed to several classes of dietary and environmental carcinogens and that these carcinogens react with DNA in breast ductal epithelial cells, the cells from which most breast cancers arise.

Molecular epidemiology, biomarkers and cancer prevention

Molecular epidemiological studies within the field of cancer research provide the potential for elucidating the carcinogenic cascade at the molecular level. Identification of susceptible subsets of the population, based on polymorphisms in genes involved in carcinogenesis, has the potential to delineate more clearly those factors that might increase cancer risk among some, but not all, individuals. Rapid advances in human genomics are making it possible to develop detailed profiles of susceptible subgroups based upon genetic variants in multiple pathways. Here we discuss examples of recent susceptibility studies involving genes, such as those involved in carcinogen metabolism, DNA repair, cell cycle and immune status, that hold the promise of increasing our understanding of cancer etiology and possible prevention strategies.