Hormone-associated cancer: mechanistic similarities between human breast cancer and estrogen-induced kidney carcinogenesis in hamsters.

Estrogens are risk factors for human breast cancer and induce kidney tumors in Syrian hamsters. Mechanistic features of the estrogen-induced hamster kidney tumor model have been compared with corresponding aspects of human breast cancer to gain insight into the mechanism of human mammary oncogenesis. Shared characteristics point to a mechanism of metabolic activation of steroidal estrogens to 4-hydroxylated catechol metabolites that may undergo metabolic redox cycling, a mechanism of generation of reactive free radicals. Tumors may arise in cells genetically altered by various types of estrogen-induced DNA damage. At the same time, these altered cells may respond to estrogen receptor-mediated stimuli in support of cell transformation and growth.


Estrogens and Breast Cancer
Estrogen administration is accepted by most epidemiologists as a risk factor of human endometrial adenocarcinoma (1,2). Thus, estrogens unopposed by progestins increase the risk of uterine tumors. This risk increases with increasing doses of estrogen and with the length of treatment (3). Obesity also increases uterine tumor risk, most likely because the aromatase activity of adipose cells elevates circulating estrone levels (3,4).
Increasing evidence shows that elevated levels of estrogens in the body are also a risk factor in breast cancer. Exogenous estrogens, alone or in combination with jliehr@marlin.utmb.edu Abbreviations used: E2, 17J-estradiol; LHP, lipid hydroperoxides; 2-OH-E2, 2-hydroxyestradiol; 4-OH-E2, 4-hydroxyestradiol. progestin, elevate breast cancer risk (5)(6)(7). Progestin added to the estrogen medications does not inhibit mammary carcinogenesis (8) because the former hormone appears to be the primary mitogen of mammary ductal epithelial cells (9), whereas estrogen appears to function in this manner in the uterus. This role of estrogen in human breast carcinogenesis is supported by risk factors of breast cancer such as high serum or urine estrogen levels (10,11), the early onset of menstruation, or late menopause (12). These relationships are illustrated by the fact that the breast cancer risk of a woman undergoing menopause at 45 years of age is half that of a woman experiencing menopause at 55 years of age (5,6,12). These lines of evidence indicate that the total lifetime exposure of a woman to exogenous or endogenous estrogen is a determinant of her breast cancer risk.

Animal Models of Estrogeninduced Cancer
For the examination of the mechanism of estrogen-induced tumorigenesis, animal models are needed in which steroidal estrogens induce tumors with high incidence. There are very few reports in rodent models of estrogen-induced mammary tumorigenesis not mediated by hydrocarbon carcinogens, viruses, or other initiating agents. Estrogens induce or enhance tumor formation in organs of rodents such as the uterus (13,14), pituitary (15,16), prostate (17), uterine cervix (18), testes (19), and liver (20). In addition, the estrogen-induced kidney tumorigenesis in Syrian hamsters has been used extensively for mechanistic studies (21) because 100% tumor incidence is routinely achieved using the natural hormone 17f-estradiol (E2) or estrone without coadministration of any other carcinogens. This animal model is useful for mechanistic studies because the tumor is estrogen induced and estrogen dependent, because cell lines have been developed that share growth characteristics with the primary tumor (22,23), and because the tumor incidence may be manipulated by coadministration of other hormones such as progesterone or testosterone (21).
Several synthetic steroidal estrogens were only weakly carcinogenic in this model system despite their powerful hormonal potencies (21,24,25). This lack of correlation of estrogenic characteristics with carcinogenic activities of steroids led to the conclusion that hormonal potency of an estrogen was necessary but not sufficient for tumor induction (24,25). It was further concluded that metabolic activation of estrogens to reactive intermediates is a key event in the development of estrogeninduced tumors.

Mechanism of Estrogeninduced Carcinogenesis in Hamsters
The metabolic activation of estrogens to catechol metabolites and further to reactive semiquinone or quinone intermediates is a central feature of this mechanistic proposal and is analogous to the metabolic activation of many other carcinogens. The primary metabolic conversion of the estrogenic hormones estrone or E2 in liver of most mammalian species is 2-hydroxylation, as shown for hamsters in Table 1 (26,27). However, an elevated formation of 4-hydroxyestrogen metabolites by specific estrogen-4-hydroxylases has been detected in hamster kidney (28), mouse uterus (29), and rat pituitary (30)-all organs in which estrogens are known to induce tumors (13)(14)(15)(16)21). The elevated formation of 4-hydroxyestrogens in rodent organs in which estrogens induce tumors is significant because catecholestrogens, including 4 Figure 1. Formation and redox cycling of 4-OH-E2. E2 is converted by cytochrome P450 enzymes, including P4501 Bi, to 4-OH-E2, which may undergo metabolic redox cycling between hydroquinone and quinone forms via semiquinone intermediates. The oxidation of this cycle is catalyzed mainly by organic hydroperoxide-dependent cytochrome P4501A1 and the reduction by NADPH-dependent cytochrome P450 reductase. The semiquinone is a free radical that may react with molecular oxygen to form superoxide. Detoxification pathways are phase 11 conjugation reactions of the catechol, including methylation by catechol-0-methyltransferase (COMT). Details are outlined in a recent review by Yager and Liehr (54).
metabolic redox cycling between hydroquinone and quinone forms (Figure 1). The catecholestrogen quinone intermediates may form covalent DNA adducts (31,32). Moreover, metabolic redox cycling generates free radicals that are capable of direct or indirect covalent modification of DNA ( Figure 2) (33,34). In line with this hypothesis, hydroxy radical damage such as 8-hydroxyguanine bases of DNA or DNA single strand breaks have been detected in the kidney of hamsters treated with estradiol (35)(36)(37). In addition, intranuclear, site-specific redox cycling involving catecholestrogen metabolites and copper ions may induce DNA damage by reactive oxygen, which is not mediated by metabolizing enzymes (38).
Lipid hydroperoxides (LHP) play a central role in the DNA damage process by catechol estrogens (Figure 2). LHP are formed by free radicals generated by redox cycling. They, in turn, support redox cycling as cofactors in the cytochrome P4501A-mediated oxidation step of the redox cycle (33,39). Finally, LHP may decompose into reactive aldehydes such as malondialdehyde, which covalendy bind to DNA (40).
One or several of these classes of covalent DNA damage may initiate and/or promote the carcinogenic process ( Figure  3). Tumors may develop from cells genetically altered by the processes outlined above. At the same time, these altered cells may respond to a receptor-mediated proliferative stimulus either by estradiol or by 4-hydroxyestradiol, which is known to be a long-acting estrogen (41).

Mechanistic Analogies in Hamster and Human Breast
Despite the obvious differences arising from the differences of organ system, cell type, latency of tumor development, and other factors, there are several common features of estrogen-induced kidney tumors in hamsters and human breast cancer (  (3). The quinone may t by NADPH-dependent cytochrome P450 reductase (4) or by NADH-dependent quinone reductase (5) in a tron reduction bypassing the semiquinone. The semiquinone may react with molecular oxygen to form radicals, which are reduced to hydrogen peroxide by superoxide dismutase (6). H202 may be reduced by to hydroxy radicals, which initiate the formation of LHP. DNA may be covalently modified by aldehyde d tion products of LHP, by hydroxy radicals, or by covalent binding of quinone metabolites. Although 2 brackets), the most common metabolite formed by liver enzymes, is capable of undergoing metabolic red its contribution to free radical generation and DNA damage may be negligible because it is convertec gates by phase 11 metabolism enzymes much more efficiently than the 4-hydroxylated isomer. The p estrogen metabolites to damage cells is potentiated by redox cycling and by the initiation of free radical Details of this pathway are outlined in a recent review by Yager and Liehr (54).  (53). (10,11). However, not every woman with high estrogen levels will develop this disease. Thus, an estrogen metabolite rather than the parent hormone likely is the ultimate carcinogenic factor. This reasoning is supported by the elevated urinary catecholestrogen concentrations in women at risk of breast cancer compared to controls (11). Moreover, in microsomes of breast cancer cells (42) Figure 3. Proposed mechanism of estrogen-induced carcinogenesis. E2 is activated by hydroxylation to 4hydroxylated catecholestrogens (CE), their redox cycling, and free radical generation as outlined above. The various types of DNA damage generated by metabolic activation of estrogens may induce gene mutations. Tumors are proposed to arise in cells genetically DNA altered by this estrogen-induced genotoxicity. At adducts the same time, these altered cells may respond to ER-mediated cell transformation and growth.
Thus, in target cells of mammary oncogenesis, estradiol may be preferentially converted to 4-hydroxyestradiol, which in the hamster kidney tumor model is a carcinogenic metabolite of estradiol (46). the natural Consistent with the pathway of metabolic ninantly to activation and DNA damage outlined xidation to above, elevated levels of 8-hydroxyguanine be reduced bases of DNA (47), increased lipid peroxitwo-elecdation (48), and increased levels of LHPsuperoxide derived aldehyde DNA adducts (49,50) metal ions have been detected in mammary tissues of 'ecomposibreast cancer patients. The same indicators lox cycling of indirect, estrogen-induced cell damage d to conju-have been detected previously in hamsters otential of and are listed in Table 2. Consistent with a reactions. hypothesis of tumor initiation by estrogens, vitamin C decreased mammary tumor incidence in humans (51), as has been observed iesis.
previously in hamsters (52). Conclusion reference Complex interactions of metabolic, chemical, physiological, and endocrine effects of estrogenic hormones are involved in the induction of tumors in the breast. This ? 6,37) complexity requires a multidisciplinary 5) design of experiments because emphasizing (39) various aspects of carcinogenesis has not er (52) sufficed to explain the full sequence of events resulting in the appearance of mam-:F-7 breast mary tumors. The mechanisms outlined above are complex and difficult to understand, yet represent the best explanation to MCF-7 date of the experimental data. Additional human research into better animal models, into the nocarci-relation of estrogen metabolism to breast stradiol cancer, and into better assays of estrogen droxylametabolites and metabolizing enzymes will !strogens provide a more detailed understanding of enzyme mammary carcinogenesis. '44,45).