Director: Fred F. Kadlubar, Ph.D.

Introduction

The strategic goals of the division are: 1) the development, validation, and clinical application of molecular biomarkers of carcinogen exposure and of individual susceptibility in humans; 2) the extrapolation of results from animal mechanistic studies and animal bioassays to humans; and 3) the development and validation of "alternatives" to the standard rodent bioassay for assessment of human carcinogenicity. The intent is to better understand the mechanisms of human carcinogenesis, to provide an estimation of human exposure to genotoxic carcinogens, to assess the importance of inter-individual differences in bioactivation and detoxification toward a specific carcinogen or class of chemical carcinogens, and to suggest intervention strategies for human cancer prevention. Emphasis is on the food-borne heterocyclic amines, aromatic amines, and polycyclic aromatic hydrocarbons, and the widely used drugs, including selected benzodiazepines, antihistamines, chloral hydrate, methylphenidate, estrogens and endocrine disruptors, as well as on alcohol and tobacco usage. Projects on the etiology of human cancers of the colon/rectum, pancreas, larynx, breast, ovary, prostate, liver, lung, urinary bladder, and esophagus are either ongoing or planned for 1997.

The division's experimental approach and project areas are:

Studies to identify genetic polymorphisms that influence drug and carcinogen metabolism, individual cancer susceptibility, and therapeutic drug efficacy:

1. Food-borne heterocyclic amines and colo-rectal cancer.
   a) Chemoprevention.
   b) Metabolic polymorphisms, DNA repair, and dietary risk factors.
2. Food-borne heterocyclic amines, aromatic amines, polycyclic aromatic hydrocarbons, and pancreatic cancer.
3. Food-borne heterocyclic amines, exogenous hormone exposure and breast cancer in African-American women.
   a) Metabolic polymorphisms and dietary risk.
   b) Metabolic polymorphisms in steroid hormone metabolism.
4. Food-borne heterocyclic amines and prostate cancer in African-American men.

Human biomonitoring, DNA adduct detection, and post-market surveillance for chemical toxicants found in foods, drugs, cosmetics, and medical devices:

1. Etiology of human breast, ovarian, prostate, liver, lung, urinary bladder, and esophageal cancers.
2. Alcohol and cigarette smoking as risk factors for laryngeal cancer.

Extrapolation between animal studies and human populations:

1. Evaluation of the neonatal mouse bioassay as an alternative bioassay for selected benzodiazepines, antihistamines, chloral hydrate, methylphenidate, drugs inducing peroxisomal proliferation or oxidative stress, catechol estrogens, and endocrine disruptors including chlorinated hydrocarbon pesticides and dinitroaniline herbicides.
2. Assessment of methylphenidate hepatocarcinogenicity in non-human primates.
3. Human intervention studies with chloral hydrate and heterocyclic amines.

FY 97 Plans and Goals

During 1997, the division will continue with the projects described under the section, "1996 Accomplishments" but with increased emphasis on chemoprevention, based on recent data that the coffee constituents, kahweol and cafestol, not only induce GSTs but also appear to down-regulate NAT2. In animal studies that will require further validation, the division has effectively changed the phenotype of rats from rapid to slow acetylators. A pilot study is also being planned in humans, using coffee prepared with and without the use of filter paper, which quantitatively retains these terpenoids. These results could have a profound impact on our ability to predict adverse drug reactions and the impact of coffee consumption on colon and bladder cancer risk.

Similarly, further attention will be placed on the regulation of CYP1A2 gene expression, given the importance of this enzyme as a determinant of therapeutic drug efficacy and cancer susceptibility. The mechanisms that control gene expression will be explored, including analysis of genetic variants as well as changes in gene methylation, an epigenetic mechanism. The CYP1A2 genotype and the methylation profile of the CYP1A2 gene will be determined in human liver tissues grouped according to age, gender, and smoking status. Methyltransferase and SAM/SAH levels will also be determined. Because CYP1A2 is also a major metabolizing enzyme for aflatoxin B₁, these efforts will subsequently lead to NCTR participation in an ongoing case-control study (in collaboration with the NCI) being conducted on aflatoxin B₁-induced liver cancer patients including both hepatitis B-positive and -negative individuals. At the same time, these samples will be utilized to examine the biological significance of a genetic polymorphism in CYP2E1, by comparing genotype to expressed levels of the enzyme in human liver microsomes. Recently in collaboration with a visiting scientist and Chief of the Molecular Epidemiology Division at the National Cancer Institute in Beijing, division research has provided the first evidence that this polymorphism in CYP2E1 is a strong risk factor for esophageal cancer in China (Linxian County), where food-borne nitrosamines known to be bioactivated by CYP2E1 are strongly implicated in the etiology of this cancer.

The division will also focus its attention on analysis of molecular epidemiological data from its case/control study on pancreatic cancer which, like colo-rectal cancer, shares common risk factors that suggest the role of food-borne heterocyclic amines, including high meat and fat consumption and low intake of cruciferous vegetables and fruits. Furthermore, pancreatic cancer risk is further increased by cigarette smoking and certain occupations, which suggests the involvement of carcinogenic aromatic amines or polycyclic aromatic hydrocarbons (PAH's) as etiologic agents. Initial studies in this laboratory on the metabolic activation of aromatic/heterocyclic amines by human pancreas tissues indicate that both NAT1 and NAT2 are expressed, albeit at lower levels than in the colon: while CYP1A2 and other CYP's involved in carcinogen metabolism are not detectable in smokers or non-smokers. Preliminary experiments on the analyses of putative carcinogen-DNA adducts in human pancreas by ³²P­postlabelling have also suggested the presence of both aromatic amine and PAH adducts; but only the latter group were elevated in cigarette smokers. Efforts are currently underway to confirm the identity of carcinogen-DNA adducts present by immunochemical and S³⁵-postlabeling techniques [in collaboration with Columbia University and MIT]. This study is expected to provide an assessment of the relative roles of dietary and environmental carcinogens in human pancreatic cancer and to result in appropriate recommendations for protecting public health.

In preliminary studies, the division has obtained evidence that the heterocyclic amine, 2­amino- ­carboline (A C), which is found in cooked foods and is the predominant aromatic amine carcinogen in cigarette smoke, is present as a major DNA adduct in human larynx. During 1997, they plan to complete characterization of this DNA adduct and develop an LC/MS method [with in the Division of Chemistry at NCTR] to confirm adduct levels in human tissues. In addition, we have data that indicate that acetaldehyde-DNA adducts may be formed in vivo after treatment with [³H]ethanol. Recently, others have shown that acetaldehyde-modified DNA can be analyzed by ³²P­postlabelling (after reduction to ethylated adducts). Acetaldehyde has been shown to be carcinogenic to rodents after inhalation and it induces chromosomal damage but not mutations in a variety of in vitro tests. On the other hand, ethanol has not been shown to be carcinogenic or mutagenic in experimental systems; however, both acetaldehyde and ethanol have been reported to be co-carcinogens (e.g., with benzo[a]pyrene, diethylnitrosamine, or vinyl chloride) in different animal models. Human epidemiological studies are also consistent with ethanol as a co-carcinogen, particularly when combined with tobacco usage. Of these, the relative risk for cancers of the upper aerodigestive tract, especially the larynx, show the most consistent synergism between total alcohol intake and heavy cigarette smoking. Thus, the division proposes to examine the hypothesis that ethanol forms acetaldehyde-DNA adducts in human larynx and that these adducts may serve to enhance the relative persistence or mutagenic outcomes of A C and other smoking-related DNA adducts.

Another major emphasis in 1997 will be directed at the possible role of chemical carcinogens in breast cancer etiology, and the modification of risk associated with these exposures by polymorphisms in genes involved in carcinogen biotransformation. Because breast cancer most commonly arises from ductal epithelial cells, in 1996 the division began a study to examine those cells shed into human breast milk for carcinogen-DNA adducts. They have been obtaining specimens from nursing mothers, both smoking and non-smoking, and have developed methodology to separate exfoliated ductal epithelial cells from human breast milk. In the coming year, they will continue with extraction and postlabeling of DNA for the presence and characterization of carcinogen-DNA adducts in human breast epithelial cells. Variability in adduct levels related to both exposure and to genetic susceptibility based on variability in carcinogen metabolism is expected. A study of chemical carcinogenesis in human breast epithelial cell studies, and of the effects of environmental and drug exposures and genetic susceptibility on these processes, should have important implications for future regulatory decisions. By sorting out etiologic mechanisms and putative risk factors in breast carcinogenesis, subgroups of individuals susceptible to specific carcinogens, particularly heterocyclic and aromatic amines, may be identified.

The division is also beginning case-control molecular epidemiologic studies of breast and prostate cancer in African-American women and men. Because African-American men have the highest incidence of prostate cancer in the world, and African-American women have twice the risk of Caucasians of premenopausal breast cancer, the NCTR is interested in evaluating possible genetic and environmental factors that may account for these racial disparities. These include dietary factors, particularly consumption of dietary, heterocyclic amines, hormonal factors (oral contraceptives, hormone replacement therapy and reproductive factors), and genetic variability in the metabolism of heterocyclic amines and steroid hormones. These hypotheses will be applied to both breast and prostate cancer. Extensive questionnaire data, as well as blood specimens and urine for caffeine phenotyping will be collected. A nested case-series study is also planned to identify and characterize DNA adducts in prostate tissue from men who are participating in the study. As in the study of exfoliated ductal epithelial cells in human breast milk, the division will evaluate levels of adducts in relationship to environmental exposures, as well as to polymorphisms in genes involved in metabolism of dietary and environmental carcinogens and endogenous steroid hormones. They are also planning to evaluate loss of expression of kang ai 1, a putative anti-metastases protein, in relation to stage and grade at diagnosis, levels of DNA adducts and environmental exposures, in collaboration with the NIEHS.

Because of the racial disparities in breast and prostate cancer incidence, and the likelihood that diet may play a profound role in the etiology of both diseases, the division is beginning a survey of dietary habits of rural African-American men and women in the Mississippi River Delta region in eastern Arkansas. There is little information regarding eating habits of rural African-Americans in the southern United States, and it is questionable if existing food-frequency questionnaires are relevant for these populations. Using 24-hour diet recalls, researchers hope to obtain detailed dietary information, from which they can evaluate the utility of existing survey instruments. These data may be extremely important for future studies on the role of diet in disease etiology in African-Americans, and may have major implications for FDA monitoring of food-borne carcinogens.

The division is also beginning to focus on the role of metabolism of steroid hormones in carcinogenesis in hormonally responsive tissues. This includes a study of estrogen metabolism in human ovarian tissues and of the effects of hormonal regulation in combination with genetic variability on these processes. Identification and characterization of estrogen-specific biotransformation pathways in normal tissues will enable the dissection of events that participate in the generation of, and/or the protection from, the production of highly estrogenic or DNA-damaging metabolites. Specifically, the study will: 1) identify the major metabolic enzymes acting on estradiol in the human ovary as a function of hormonal exposure associated with ovulation; 2) characterize metabolic phenotypes for these major enzymatic activities (e.g., high, intermediate and low); 3) identify variability in metabolic activity as a function of hormonal exposure and/or individual variability; and 4) determine the molecular mechanisms that account for any observed phenotypic variability. It is predicted that these studies will identify estrogen-specific enzyme isoforms expressed in the ovary that demonstrate significant hormonal responsiveness and/or interindividual variability. These studies have the potential to aid in the definition of a group of individuals at greater risk from estrogen exposures for developing hormonally induced cancers.

Other areas of research that are still in the early stages involve similar molecular biomarkers studies, with special emphasis on the role of CYP1B1 in steroid hormone and in carcinogen/drug metabolism. The division has thus far prepared anti-peptide antibodies and have established functional assays for CYP1B1, and is now planning immunohistochemical studies on tissue localization in humans. They have also set up the host reactivation DNA repair assays, using aromatic and heterocyclic amine-adducted DNAs, that will allow examination of interindividual differences in DNA repair for carcinogen adducts that have been specifically implicated in human cancer etiology.

In addition to the continuation of its efforts on validation of the neonatal mouse bioassay, the division will propose, in collaboration with the Division of Neurotoxicology, an addendum to an existing three-year protocol that has been examining the effects of methylphenidate (Ritalin) on cognitive function in young Rhesus monkeys. The dose range and blood levels used are similar to that used to treat narcolepsy in adults and more widely, to treat attention-deficit disorders (hyperactivity) in children. A continuation of this dosing for an additional three years should allow an assessment of any neoplastic changes in the dosed animals. There is currently a high level of concern within CDER due to an NTP report that found hepatoblastomas in rats (5/50) given high doses of methylphenidate. Since this tumor is extremely rare in rats, these findings must be carefully considered, particularly since the most common childhood liver tumor, albeit rare, is also hepatoblastoma.

FY 96 Accomplishments

During 1996, division studies on genetic polymorphisms were focused on the bioactivation and detoxification of the food-borne heterocyclic amines, which have been of increasing public health concern to FDA. Using animal models, human tissues, and molecular biomarkers in epidemiological studies, the bioactivation of heterocyclic amines to colon carcinogens in humans was found to involve N-oxidation followed by O-acetylation to form the N-acetoxy arylamine that binds to DNA to form carcinogen-DNA adducts. These steps are catalyzed by the hepatic enzymes, cytochrome P4501A2 (CYP1A2) and acetyltransferase-2 (NAT2), respectively, which NCTR and others have shown to be expressed polymorphically in humans. The division has recently found four variant alleles in the CYP1A2 gene and now have evidence for a genetic polymorphism that is associated with CYP1A2 inducibility. It is believed that this finding will have a major impact not only on cancer susceptibility, but also on therapeutic drug efficacy and hormonal interactions, since CYP1A2 is also the major enzyme metabolizing many drugs and estrogens. Thus far, research has shown that, consistent with the proposed metabolic activation pathway for heterocyclic amine carcinogens, subjects at greatest risk for colo-rectal cancer or non­familial polyps are those who possess both the rapid NAT2 genotype/phenotype and the rapid CYP1A2 phenotype and who are exposed to high dietary levels of carcinogenic heterocyclic amines. Moreover, a logistic regression model that included both metabolic genotypes/phenotypes and consumption of well-done red meat suggests that, in terms of attributable risk, these susceptibility factors together with food-borne heterocyclic amine carcinogen intake may account for about half of the sporadic colo-rectal tumor incidence observed in the U.S. At the same time, research has found that glucuronyl transferases (UGTs) and the -class glutathione S­transferases (GST), human GST A1 and rat GST 1a,1b, can effectively detoxify the heterocyclic amines. This finding is of particular significance since UGTs and -class GSTs are known to be inducible in humans and experimental animals by consumption of cruciferous and yellow­green vegetables; and the latter has been shown by epidemiological studies to be one of the most consistent protective factors against human colo-rectal cancer.

Since colo-rectal cancer is the most prevalent cancer in non-smokers in the U.S., appropriate intervention strategies and concomitant public health recommendations are of paramount importance. To determine the potential for dietary intervention in modulating the carcinogenicity of the heterocyclic amines, the effect of a variety of treatment regimens on the bioactivation and detoxification pathways was examined. The paradigms used are representative of classes of dietary substances associated with chemoprevention in humans and have been shown to affect tumorigenesis in animal models. These included dietary tannic acid, Oltipraz, -angelicalactone, quercetin, ethoxyquin, indole-3-carbinol, diallyl sulfide, benzylisothiocyanate, kahweol:cafestol (1:1), black tea, green tea, high dietary fiber, and physical exercise. Of these, consumption of black tea and dietary kahweol:cafestol (the major terpenoids in coffee) had the strongest effect on inhibiting heterocyclic amine-DNA adduct formation in the rat colon model. The mechanisms of action of these agents differed and appear to involve potent inhibition of CYP1A2 by black tea and induction of -class GSTs by kahweol:cafestol, suggesting the potential for combined use in dietary intervention studies.

In the area of human biomonitoring and DNA adduct detection, similar progress has been made in understanding the etiology of urinary bladder cancer, where occupational exposures and cigarette smoking are regarded as major risk factors. NCTR research had previously provided strong evidence, based on metabolic polymorphisms and DNA adduct detection, that smoking-related aromatic amines are major risk factors. However, other epidemiological studies have suggested that exposures to polycyclic aromatic hydrocarbons (PAHs) from smoking, occupational or environmental sources may also increase bladder cancer risk and could account for a substantial proportion of bladder DNA adducts. Recently, the division has found that PAH-DNA adduct levels are correlated significantly with the ability of the bladder to bioactivate benzo[a]pyrene (BP) and have identified the activating enzyme as lipoxygenase. In addition, genotyping studies have suggested that glutathione S­transferase (GST) M1 may play an important role in the detoxification of PAH bladder carcinogens, since individuals with the GST M1 deletion are at higher risk for bladder cancer. Thus, we have also examined the relationship between adduct levels and GST M1 genotype and the expression of GST enzymes in the bladder epithelium. However, there was no effect of the GST M1 genotype on DNA adduct levels and the predominant GST (» 90%) in bladder cytosol was found to be GST P1. Interestingly, it was found that the levels of this enzyme in the bladder appear to be expressed in a polymorphic manner and are inversely related to PAH-DNA adduct levels in this tissue. Moreover, individuals who are both rapid metabolizers for BP and possess low GST P1 activity have four to six-fold fold higher bladder PAH-DNA adduct levels than poor BP metabolizers with the high GST P1 activity phenotype.

This effort subsequently led to a collaboration with the University of Dundee and resulted in the identification of a genetic polymorphism in GST P1, which involves an amino acid change at the active site of the protein and thus alters substrate activity. Since this GST is the major detoxifying enzyme in human extrahepatic tissues and is up-regulated in human tumor tissues, it is expected that this polymorphism will play a critical role in both cancer susceptibility and in chemotherapeutic drug efficacy. Accordingly, collaborative studies have begun to test these hypotheses with the National Institute of Public Health in Budapest, with the NCI, and with the UAMS.

Another project involving the use of molecular biomarkers in a pancreas cancer case-control study is nearly completed with some 100 cases and 300 controls entered into the study.

International collaborative efforts in the area of human DNA adduct biomonitoring have also been undertaken by the division, together with the NCTR Division of Biochemical Toxicology, the U.S. EPA, and the International Agency for Research on Cancer. This working group has organized and are participating in interlaboratory trials for the detection of carcinogen-DNA adducts in humans and its application to human risk assessment. This effort now involves some 30 laboratories world-wide and is expected to form the basis for the use of DNA adduct measurements in making regulatory decisions. As part of NCTR's commitment to research progress in this area, the Division served as a principal organizer and sponsor of the 6th International Conference on Carcinogenic and Mutagenic Aromatic and Heterocyclic Amines (Monterey, CA) and the International Conference on DNA Adducts and Mutations in Human Biomonitoring (Stockholm).

Projects involving the extrapolation between animal models and humans have thus far focused primarily on the validation of the neonatal mouse bioassay as an alternative model for identifying genotoxic carcinogens. The evaluation of several widely used benzodiazepine and antihistamine drugs, as well as chloral hydrate, methylphenidate, drugs inducing peroxisomal proliferation or oxidative stress, catechol estrogens, and endocrine disruptors, including chlorinated hydrocarbon pesticides and dinitroaniline herbicides, are ongoing in the neonatal mouse and are being compared to studies being conducted by the National Institute for Environmental Health Sciences (NIEHS) on other alternative rodent bioassays. The compounds selected represent major classes of drugs that are widely used in human populations. A common concern for many of these compounds arises from drug-related increases in the incidence of mouse liver tumors observed in standard two­year carcinogenicity studies. In this regard, the mechanism of tumor induction is unclear and both genotoxic and nongenotoxic processes have been proposed. However, in the neonatal mouse tumorigenicity bioassay, only two doses of the test compound, given to preweanling animals, are required to obtain positive results after 12 months; and, thus far, only genotoxic carcinogens have been shown to be active in this test system. Therefore, we believe that this bioassay, when combined with relevant mechanistic information in human cells and in human epidemiological studies, will provide a more definitive assessment of the significance of marginal findings in the standard rodent bioassay and will also become a useful supplemental or alternate carcinogenicity screening method for FDA-regulated drugs or drug products.

Significance to the FDA

These research projects are being carried out to identify human polymorphisms in carcinogen and drug metabolism and to provide direct evidence for human exposure to specific chemical carcinogens. Furthermore, correlational analyses between DNA adduct levels and carcinogen-metabolizing enzymes in the same individuals allows not only the identification of populations who may be at higher risk for chemically-induced cancers but also provides evidence for the role of different chemical classes in human cancer etiology. Together, these efforts are expected to result in better public health monitoring and regulatory risk assessment of food and drug carcinogens and in appropriate strategies for earlier disease diagnosis and cancer prevention.