Logo of nihpaAbout Author manuscripts Submit a manuscript HHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Int J Cancer. Author manuscript; available in PMC 2016 Mar 2.
Published in final edited form as:
PMCID: PMC4775086
NIHMSID: NIHMS762539
PMID: 20104523

Effects of estrogen on breast cancer development: role of estrogen receptor independent mechanisms

Wei Yue,1 Ji-Ping Wang,1 Yuebai Li,1 Ping Fan,1 Guijian Liu,1 Nan Zhang,1 Mark Conaway,2 Hongkun Wang,2 Kenneth S. Korach,3 Wayne Bocchinfuso,3 and Richard Santen1
Author information Copyright and License information Disclaimer

Associated Data

Supplementary Materials

Abstract

Development of breast cancer involves genetic factors as well as lifetime exposure to estrogen. The precise molecular mechanisms whereby estrogens influence breast tumor formation are poorly understood. While estrogen receptor α (ERα) is certainly involved, nonreceptor mediated effects of estradiol (E2) may also play an important role in facilitating breast tumor development. A “reductionist” strategy allowed us to examine the role of ERα independent effects of E2 on mammary tumor development in ERα knockout (ERKO) mice bearing the Wnt-1 oncogene. Exogenous E2 “clamped” at early follicular and midluteal phase levels (i.e., 80 and 240 pg/ml) accelerated tumor formation in a dose-related fashion in ERKO/Wnt-1 animals (p = 0.0002). Reduction of endogenous E2 by oophorectomy (p < 0.001) or an aromatase inhibitor (AI) (p = 0.055) in intact ERKO/Wnt-1 animals delayed tumorigenesis as further evidence for an ER-independent effect. The effects of residual ERα or β were not involved since enhancement of tumor formation could not be blocked by the antiestrogen fulvestrant. 17α-OH-E2, a metabolizable but ER-impeded analogue of E2 stimulated tumor development without measurable uterine stimulatory effects. Taken together, our results suggest that ER-independent actions of E2 can influence breast tumor development in concert with ER dependent effects. These observations suggest 1 mechanism whereby AIs, which block E2 synthesis, would be more effective for breast cancer prevention than use of antiestrogens, which only block ER-mediated effects.

Keywords: breast cancer, Wnt-1, estrogen receptor α knockout

It was estimated that approximately 192,000 women would be diagnosed with breast cancer in the United States in 2009 with 40,000 resulting deaths.1 Improved diagnostic and treatment strategies have decreased breast cancer mortality by 25% over the past 2 decades,2–4 but the physical and psychological burdens of surgery, radiotherapy, hormonal- and chemotherapy are substantial. For this reason, breast cancer prevention represents a major focus of current research.5 A greater understanding of the molecular mechanisms of carcinogenesis will be required, however, before development of improved strategies for breast cancer prevention.

Both genetic and hormonal factors have been implicated in the genesis of breast cancer. Genetic factors involve significant mutations in BRCA 1 and 2, CHEK2, TP53, LKB-1 and PTEN in 5–10% of patients and lower risk mutations inferred by identical twin and genome wide association studies in others.6–8 Epidemiologic and experimental data implicate estradiol (E2) as another contributing factor. In various animal models, E2 administration causes and antiestrogens prevent breast cancer.9,10 In women, bilateral oophorectomy before age 35 reduces the lifetime incidence of breast cancer by 75%.11,12 Increased lifetime exposure to estrogens, conferred by early menarche, late menopause, long-term menopausal estrogen therapy, obesity and high circulating E2 levels in pre- and postmenopausal women, are associated with an enhanced incidence of breast cancer.13–16 Data from 2 large studies demonstrated that postmenopausal women in the highest quintile of plasma free E2 experienced at least a 2.58-fold (95% CI 1.76–3.78) higher rate of breast cancer over the ensuing 10 years than those in the lowest quintile.16,17 Blockade of estrogen action with tamoxifen or raloxifene reduces the incidence of breast cancer by 50–75% in high-risk women.5,18,19 Finally, inhibition of E2 synthesis with aromatase inhibitors (AIs) or abrogation of its action with antiestrogens prevents the development of contralateral breast cancer during adjuvant therapy.20,21 Taken together, these data provide compelling evidence that E2 plays a major etiologic role in breast cancer development.

The precise molecular mechanisms whereby E2 influences breast cancer development are not well understood. The most widely accepted theory, supported by extensive experimental evidence,22,23 holds that E2, acting through ERα, stimulates cell proliferation and initiates mutations that occur as a function of errors during DNA replication. The promotional effect of E2 then supports the growth of cells harboring mutations, which then accumulate until cancer ultimately results. Clinical and experimental data also suggest the possibility that receptor independent effects of E2 may be mechanistically involved. In a recent review, Yager and Davidson describe in detail how estrogen metabolites can exert genotoxic effects, which contribute to the development of breast cancer.24 Estrogens are converted to quinone metabolites, which directly bind to DNA and form adducts. Additionally, catechol estrogen metabolites undergo redox cycling with generation of oxygen free radicals, which damage DNA-bound guanine to form 8-OXO-guanine. The quinone-adducts and 8-OXO-guanine bases are unstable and are deleted from the affected DNA segments through a process called “depurination.”24 Error prone DNA repair then results in the formation of mutations at the depurinated sites. Accumulation of these mutations would then contribute to the development of breast cancer.25 As predicted from the “estrogen genotoxic metabolite” hypothesis, a predisposition to breast cancer would be expected in women with combinations of mutations of estrogen metabolizing enzymes, a finding reported by Park et al.26 and Ritchie et al.27 In support of the depurination mechanism, 2 recent reports indicate that women with breast cancer or at high risk for the disease have significantly higher levels of depurinating estrogen-DNA adducts in their urine than women at normal risk for breast cancer.28,29

Cell culture and animal data have provided biochemical and biologic evidence that ER independent DNA damage from E2 occurs.30–32 However, the causal relationship between E2 metabolism and breast cancer development has been a controversial issue. To date, no direct proof of an ER-independent effect of E2 on mammary tumor formation in an animal model has been reported. For this reason, we decided to provide proof of the principle that ER independent effects of estrogen could influence breast cancer development in an in vivo system. We chose the estrogen-receptor knockout (ERKO) mouse model system, which would allow assessment of the effects of E2 acting independently of ERα function. Reasoning that E2 exerts modulating effects on breast cancer incidence in women with genetic defects, we knocked in the Wnt-1 gene in the ERKO mouse to generate a double transgenic mouse model for our studies. This model mimics high-risk patients with genetic defects and provides a system with a sufficiently high frequency of tumor development to make the studies feasible.

We recognized that various factors inherent in our experimental design would confound interpretation of data and therefore attempted to utilize a “reductionist approach,” a term originally introduced by Bernard in 1864.33 Accordingly, to minimize confounding factors inherent in our model, we removed the ovaries in ERα knockout mice to eliminate potential confounding effects of ovarian steroid and peptide hormones and administered exogenous estradiol. We also administrated the pure antiestrogen (fulvestrant) to ensure complete blockade of any residual ERα and ERβ. Using this approach, our data strongly suggest that E2 can influence tumor formation through ER-independent effects.

Material and Methods

Animals

Wnt-1 transgenic animals were obtained from Dr. Harold Varmus and bred at the National Institute of Environmental Health Sciences (NIEHS).34–36 These animals were then crossbred with heterozygous ERα knockout mice to generate Wnt-1 transgenic mice, which could then be further bred to produce ERα knockout/Wnt-1 double transgenics (ERKO/Wnt-1). With establishment of collaboration among investigators, breeding pairs were sent to the University of Virginia (UVA) and a separate colony established. At both institutions, the mice were housed and treated in accordance with the NIH guide to Humane Use of Animals in Research. All surgical procedures were approved by the Animal Care and Use Committees at UVA and NIEHS. Genotyping was performed as previously described.34 Full characterization of the phenotypic, biologic and biochemical properties of these animals have been published.34 Results obtained from the collaborative studies between the 2 institutions and using identical protocols were pooled for statistical analysis presented in Figure 3a.

An external file that holds a picture, illustration, etc. Object name is nihms762539f3.jpg

(a) Kaplan–Meier curves of tumor formation in noncastrate ER+/+/Wnt-1 (n = 79) and ERKO/Wnt-1 animals (n = 120) and the effect of oophorectomy performed before 16 days of age on tumor formation in ERKO/Wnt-1 animals (n = 48). The vertical dotted lines represent the 50% incidence time point as described in the text. The curves were drawn with pooled data from NIEHS and the University of Virginia. The differences among 3 groups were statistically significant (p < 0.001). (b) Kaplan–Meier curves of tumor-free survival of intact ERKO/Wnt-1 animals treated with (n = 24) or without (n = 120) letrozole (20 µg/mouse/day). The difference between these curves was marginally significant (p = 0.055).

“E2 clamp” method and drug administration

Silastic tubes of 0.19 cm internal diameter were filled with E2/cholesterol mixtures at various ratios. The lengths of the filled part of Silastic tubes were 2.5, 5 or 7.5 mm, respectively. Our prior studies validated the ability to “clamp” plasma E2 at levels ranging from 20 to 800 pg/ml over a 2-month period37 and demonstrated linear dose responses in uterine weight in castrate mice.38 In our study, plasma E2 was “clamped” at levels representing postmenopausal (5 and 10 pg/ml), early follicular phase (80 pg/ml) and midluteal phase (240 pg/ml) levels in women. The implants used contained the following E2/cholesterol ratios and lengths: 1:39/2.5 mm (5 pg/ml), 1:19/2.5 mm (10 pg/ml), 1:3/2.5 mm (80 pg/ml) and 1:3/7.5 mm (240 pg/ml). Implants were inserted under the skin in the backs of the mice and changed every 2 months. Fulvestrant dissolved in sesame oil was administered by subcutaneous injection once per week at a dosage of 5 mg/mouse. Letrozole was suspended in 0.3% carboxymethyl cellulose solution in saline and administered by subcutaneous injection once a day at a dosage of 20 µg/mouse, 5 days a week. The complete blocking effects of fulvestrant were demonstrated by bioassay of uterine weight as described later.

Bioassay of ER-dependent actions of E2

Measurements of uterine wet weight were used as a bioassay for ER-dependent actions of E2. Ovariectomy was carried out at 15 days of age. Treatment with E2 via silastic implants at a dose of 240 pg/ml alone or in combination with fulvestrant at 5 mg/week subcutaneously started immediately after ovariectomy and continued for 2 months. At the end of treatment, animals were anesthetized and uteri were dissected and weighed after blotting of fluid. In experiments involving assessment of tumor formation, uteri were collected and weighed at the animal sacrifice when tumors were detected or at the end of the experiment without tumor formation. No increase in uterine weight occurred as a function of age in postpubertal animals (Supporting Information Figure 1).

Preparation of whole mounts

The whole mounts from excised mammary glands were fixed and stained as previously described.39 The inguinal mammary glands were excised, placed on glass slides and immerged in Carnoy’s fixative for 2–4 hr at room temperature. The glands were washed with 70% ethanol for 15 min, gradually hydrated and then stained overnight in carmine alum solution (1 g carmine natural red, 2.5 g aluminum potassium sulfate in 500 ml distilled water and a crystal of thymol). The glands were dehydrated progressively in 70, 95 and 100% ethanol for 15 min during each step. The mammary fat pads were cleared in xylene. The mammary whole mounts were photographed using Olympus SZX12 microscope.

Endogenous and exogenous gene expression assays

The ERE-TATA-luciferase reporter system was previously described in detail.40 Progesterone receptors A and B were detected on Western blots using a monoclonal antibody against the progesterone receptor (Cell Signaling Technology, Beverly, MA).40

Statistical methods

The Kaplan–Meier analyses were used to compare the tumor-free survival time between different treatment groups of mice in the study. The rate of tumor development was compared among the various treatment groups to determine whether they are statistically significantly different. The Student’s t-test was used to compare mean uterine weights between 2 groups and a significance level of 0.05 was considered to be statistically significant.

Results

E2 effects on mammary proliferation in ERKO mice

To demonstrate that E2 did not induce proliferative effects on breast tissue in the ERKO/Wnt-1 model, we initially conducted systematic examination of mammary whole mounts. Our prior studies had demonstrated that knockout of ERα36 allowed development of only rudimentary mammary ductal structures (Supporting Information Figure 2B) but that introduction of the Wnt-1 gene into ERKO animals caused proliferation of the existing mammary rudiment (Supporting Information Figures 2D and 2F). Our current studies demonstrated that E2 did not influence proliferation of breast tissue in the absence of ERα. As we showed before, the mammary gland was fully developed in mice bearing wild-type ERα. Wnt-1 expression caused mammary gland hyperplasia (Fig. 2, bottom panel: intact). Removal of ovaries reduced the size of lobules, which could be reversed by administration of estradiol. The morphology of glands from mice receiving E2 plus fulvestrant was the same as that in ovariectomized mice (Fig. 2, bottom panel: ovx + E2 + ICI). The results implicated an important role of E2 in stimulation of lobular proliferation even in the presence of Wnt-1. In striking contrast, no substantial change in mammary gland morphology occurred in ERKO/Wnt-1 mice when the ovaries were removed (Fig. 1, top panel: ovx). More importantly, supplementation with 240 pg/ml E2 for at least 2 months did not stimulate lobule proliferation (Fig. 1, top panel: ovx + E2). Administration of fulvestrant (Fig. 1, top panel: ovx + E2 + ICI) did not alter mammary gland morphology in ERKO/Wnt-1 mice. To further examine proliferation, proliferating cell nuclear antigen (PCNA) in mammary glands of mice expressing wild-type ERα or with ERα knocked out was analyzed by Western blot. The average level of PCNA in ERKO mammary gland was less than 20% of that in ER+ gland (data not shown) even though the circulating estradiol in ERKO mice is 10-fold higher than in ER+ mice.41 These data provided evidence that knockout of ERα prevented proliferative effects of E2 on mammary gland in this model.

An external file that holds a picture, illustration, etc. Object name is nihms762539f1.jpg

Whole mounts of the mammary gland in ERα knockout (ERKO) animals cotransfected with the Wnt-1 gene (ERKO/Wnt-1, top panels) and in ERα positive wild-type animals bearing the Wnt-1 gene (ER+/+/Wnt-1, bottom panels). The entire mammary fat pad is shown with the central lymph node (dense blue) shown on each whole mount for orientation. Three to six animals from each group were examined and morphology of the mammary gland within each group was highly consistent. The groups include: intact: animals in which the ovaries have not been removed surgically; ovx: animals in which the ovaries have been removed surgically before day 16 of life; ovx + E2: animals whose ovaries have been surgically removed but received E2 with silastic implants designed to produce plasma levels of 240 pg/ml; and ovx + E2 + ICI: animals having undergone surgical removal of the ovaries and receiving E2 implants (240 pg/ml) plus fulvestrant (ICI, 5 mg/mouse/week).

An external file that holds a picture, illustration, etc. Object name is nihms762539f2.jpg

Kaplan–Meier curves comparing the effect of estradiol on tumor formation in oophorectomized, ERKO/Wnt-1 animals treated either with cholesterol (vehicle, n = 20) or with silastic implants producing plasma midluteal phase levels of estradiol (240 pg/ml, n = 20) or early follicular phase levels (80 pg/ml, n = 19). The differences between the vehicle-treated animals and those with plasma E2 levels clamped at 240 pg/ml were statistically significant (p = 0.0002). The animals with levels clamped at 80 pg/ml exhibited tumor development curves intermediate between those in the vehicle treated and those receiving 240 pg/ml.

ERα-independent effects of E2 on tumor development

We administered E2 or vehicle over a 24-month period to castrate ERKO/Wnt-1 animals to examine the effects of E2 on tumor formation in animals lacking ERα. With the “E2 clamp” methodology,42 we maintained plasma E2 at levels reflecting postmenopausal (5 and 10 pg/ml), early-follicular (80 pg/ml) and midluteal (240 pg/ml) phase concentrations in women. Two endpoints were utilized to assess effects of E2 administration: (1) the percentage of animals developing tumors and (2) the time in months at which 50% of the animals developed tumors (50% incidence time). Regarding the first end point, 50% of castrate animals treated with vehicle developed tumors over 2 years. In marked contrast, 80% of animals receiving follicular phase levels of E2 developed tumors and nearly 100% in those with luteal phase levels. Tumor incidence in animals receiving postmenopausal levels of E2 did not differ from those given vehicle (data not shown). With respect to the second end point, the 50% incidence time in the vehicle and postmenopausal E2-treated animals was 23 months. A striking difference was observed in animals with E2 “clamped” at luteal phase levels of 240 pg/ml, whose 50% incidence time point was 10 months (p = 0.0002). Those with follicular phase levels had an intermediate 50% incidence time point of 14 months (Fig. 2).

Effect of castration on tumor development in ERKO/Wnt-1 animals

As exogenous E2 increased tumor incidence and reduced latency, we reasoned that castration, by lowering endogenous E2, should also reduce tumor incidence from levels observed in intact ERKO/Wnt-1 animals. For these experiments, we compared intact and castrate ERKO/Wnt-1 animals. Castration delayed tumor onset (50% incidence time) from 12 to 23 months and reduced tumor incidence from 80% to 50% (p < 0.001).34 Our working hypothesis is that both ERα dependent and ERα independent effects of estradiol are involved in carcinogenesis. This experiment also allowed verification of the expected ERα dependent effect on the process of carcinogenesis. Tumors developed sooner (50% tumor incidence time was 6 months) in the ER+/+/Wnt-1 animals than in those lacking ERα (Fig. 3a).

Since castration reduced tumor incidence in the intact ERKO animals, we reasoned that pharmacologic suppression of estrogen production should also exert similar effects. We had previously developed a regimen sufficient to block ovarian estrogen production to castrate levels in rodents with high-dose letrozole (AI) administration.43 Letrozole, given at a dosage of 20 lg/day for 5 days a week, increased 50% tumor incidence time from 12 to 18 months (p = 0.055) (Fig. 3b). The effect of letrozole was dose-dependent, since no difference in 50% incidence time point or overall incidence was observed in animals receiving a lower dose (10 µg/day) of AI versus vehicle (data not shown).

Complete elimination of effects of truncated ERα and ERβ

A factor confounding our experimental design is that the “Korach” ERKO animals retain 5–20% of the ERα activity present in their ER+ counterparts. To eliminate these effects, we blocked ERα (and ERβ) function with the “pure antiestrogen” fulvestrant and examined the effect of E2 under these conditions. Fulvestrant or vehicle was administered to castrate ERKO/Wnt-1 animals with E2 “clamped” at 240 pg/ml. As evidence of an ER independent effect, E2 induced tumor formation similarly in the E2/fulvestrant treated as in the E2/vehicle-treated animals (Fig. 4a). To provide further support that residual ERα did not explain our results, we administered 17α-OH-E2, an estrogen analogue lacking ERα-mediated activity but capable of metabolism to potentially genotoxic metabolites.44 Our data demonstrated that 17α-OH-E2 induced tumors in the castrate ERKO/Wnt-1 animals at a rate similar to that in animals with E2 maintained at the same plasma level (i.e., 240 pg/ml) (Fig. 4b).

An external file that holds a picture, illustration, etc. Object name is nihms762539f4.jpg

(a) Kaplan–Meier curves comparing the effect of 240 pg/ml of E2 with or without fulvestrant (ICI) in castrate ERKO/Wnt-1 animals (n = 21). These curves were not statistically significantly different (p = 0.56). (b) Kaplan–Meier curves showing the effect of 17α-OH-E2 on tumor formation in ERKO/Wnt-1 animals receiving 240 pg/ml 17α-OH-E2 (n = 18). The difference between 17α-OH-E2 and 17β-OH-E2 was not statistically significant (p = 0.34). Data on the 240 pg/ml E2 and vehicle doses are reproduced from Figure 2 and are statistically significant (p = 0.0002).

Bioassay of “clamped” E2 on uterine weight

Our various strategies to examine the ER-independent effects of E2 critically depended on complete blockade of any residual ER activity resulting from a truncated ERα or from low levels of ERβ. Measurement of uterine weight provided a robust bioassay of E2 to determine whether complete blockade was achieved. We measured uterine weight after at least 2 months of E2 exposure under each experimental condition (Fig. 5). In the ERKO castrate animals, E2 stimulated uterine weight to 18% of that observed in ER+/+/Wnt-1 animals, an effect resulting from a biologic effect of the truncated 56KD receptor (Fig. 5). Fulvestrant completely blocked the residual ER responsiveness in the ERKO/Wnt-1 animals. Uterine weight fell to 7 ± 1 mg in the animals receiving 240 pg/ml E2 plus fulvestrant, a uterine weight similar to that observed in castrate animals (Fig. 5). In aggregate, these data demonstrated that fulvestrant was capable of completely abrogating the effects of residual ER activity in ERKO animals.

An external file that holds a picture, illustration, etc. Object name is nihms762539f5.jpg

Uterine wet weights in ER+/Wnt-1 (ER+, white bars) and ERKO/Wnt-1 (ER−, black bars) animals. Shown are mean (±SE) weights of the uterus under various conditions. Data from ER-animals (ovx and E2 groups) were pooled from different experiments (n = 48). 17α-E2/ovx: castrate animals receiving 17α-OH-E2 to produce plasma levels of 240 pg/ml (n = 17). Statistical analysis: ER+ groups: (a) compared to intact (n = 6), (b) compared to ovx (n = 3), (c) compared to ovx + E2 (n = 5); ERKO groups: (d) compared to intact (n = 39), (e) compared to ovx (n = 13), (f) compared to ovx + E2 (n = 48) with all p-values less than 0.001.

We also wished to confirm by bioassay that 17α-OH-E2 did not exert ER-mediated effects as further evidence that its activity on tumor formation was ER-independent. At a level of 240 pg/ml, this compound caused no stimulation in uterine weight (4 ± 0.5 mg) indicating its lack of uterotropic activity (Fig. 5). Specifically, we observed no stimulation of uterine weight in ER+/+ mice by 4-day injections of 17α-OH-E2, whereas the same doses of 17β-E2 caused a 3-fold increase in uterine weight (Supporting Information Figure 3). As further proof of the minimal ER-mediated effects of 17α-OH-E2, we tested its ability to stimulate transcription of endogenous and exogenous estrogen responsive genes and MCF-7 cell growth in vitro. The growth of MCF-7 cells in response to 17β-E2 and 17α-OH-E2 was evaluated by cell counting after 5-day exposure to these steroids. The potency of 17α-OH-E2 on transcription of an ERE-luciferase construct (exogenous reporter gene), on progesterone receptor synthesis (endogenous genes) and on cell growth was 1% or less than that of E2 (Supporting Information Figures 4A–C).

Discussion

Basic, epidemiologic and clinical studies provide strong evidence of a role for estrogen in the genesis of breast cancer but the precise mechanistic actions on tumor formation are incompletely understood.9,11,13,24,45,46 While compelling biologic and clinical evidence support a key role for ERα mediated effects, receptor independent pathways involving estrogen metabolites may also contribute to breast cancer initiation.25 Our study utilized a “reductionist” approach to provide in vivo support for the principle that estrogens could influence mammary tumor formation in the absence of functioning estrogen receptors. Use of an ERKO/Wnt-1, double transgenic, castrate mouse model allowed specific assessment of the role of exogenous E2 on tumor formation while “reducing” the confounding effects of prolactin, progesterone and other ovarian factors. Inhibition of endogenous E2 by oophorectomy or an AI in ERKO mice provided another means of assessment. Using these various experimental approaches, we demonstrated statistically significant, dose-related effects of E2 on breast tumor onset and incidence in the absence of a functioning ERα.

While there was no functional ERα in our model, interpretation of our results could be confounded by the effects of residual ERα, ERβ and GPR30, a membrane associated protein that mediates nongenomic effects of E2. The “Korach” ERKO mice express an mRNA species that yields a 56 Kd truncated ERα message41 whose translated protein retains its DNA and ligand binding domains. The truncated ERα has less than 10% of the biological activity of the native ERα. To minimize the possible confounding effects, we administered the “pure” antiestrogen fulvestrant to block the activities of residual ERα. Complete blockade of ERα activity by fulvestrant was confirmed by using uterine weight as a bioassay. Another approach to demonstrate a non-ER mediated effect was the use of 17α-OH-E2, a metabolizable but ER impeded estrogen. 17α-OH-E2 significantly influenced tumor onset and incidence in ERKO animals but lacked uterotrophic activity.

Several studies suggest that ERβ exerts minimal effects on rodent mammary glands.47 Expression of ERβ mRNA could only be detected by PCR methodology in mammary tissue of ERKO mice.48 Nevertheless, if ERβ were present and important for tumor formation, fulvestrant would block the activity of this receptor. Accordingly, our data provide strong evidence that ERβ did not explain our results. The G-protein coupled receptor GPR30 was recently reported to mediate estrogen promoted proliferative signaling in an ER-negative breast cancer cell line and human endometrial cells. However, 2 recent studies using GPR30−/− mice indicated that GPR30 does not mediate estrogenic responses in the uterus and the mammary gland.49,50 In addition, in vitro studies showed that fulvestrant did not inhibit but paradoxically activated GPR30.51 The ER impeded analogue, 17α-OH-E2 could not activate GPR30.51 Our results with fulvestrant and 17α-OH-E2 provided indirect evidence that GPR30 did not play a role in mammary tumor initiation in our model. Taken together, these multiple experimental approaches provided strong evidence in support of the principle that E2 can influence breast tumor development independently of ER functionality.

The ER independent effects of E2 on tumor formation likely occur via estrogen metabolites. As reviewed by Yager and Davidson, estrogens are hydroxylated at the 2-, 4- and 16α-positions.24 The 2 and 4 catechol estrogen metabolites are involved in redox cycling with generation of oxygen free radicals, which can cause DNA damage. An additional genotoxic mechanism involves further oxidation to the 2,3- and 3,4-estradiol-quinones, which form covalent adenine and guanine DNA adducts. The adenine and guanine adducts of the 4-hydroxylated products are unstable and result in cleavage from DNA with formation of depurinated sites.25 Mutation at these DNA sites can occur through error prone DNA repair. Our prior measurements in rodent and human mammary tissue demonstrated that both benign and malignant breast tissue are able to metabolize E2 to E2-3,4-quinones that react with DNA to form depurinating N3Adenine and N7Guanine adducts.52

Estrogen hydroxylation at the 2 and 16 positions appears to be less important than 4-hydroxylation for genesis of tumors. Adducts arising from 2-hydroxylated-estrogen metabolites depurinate minimally and are found at very low levels in women and men with cancer.28,29,53,54 Animal studies46 demonstrate that 4-OH-E2 but not 2-OH-E2 causes kidney and uterine cancer. Although a series of prior studies suggested the importance of 16α hydroxylation in cancer formation,55 recent reviews question this conclusion based on the lack of specificity of earlier assays used to measure these compounds.24,46

Comprehensive data from several published in vitro studies provide additional experimental support for the “genotoxic E2 metabolism hypothesis.”24,25,31,32,46,56 Estradiol caused DNA point mutations in ER negative V-79 and Big Blue rat cell culture mutation assays.30,31 Benign, ER negative MCF-10 mammary epithelial cells, when exposed to physiologic concentrations of E2, underwent malignant transformation and form tumors in immunodeficient mice.25,57 At physiologic concentrations, E2 caused loss of heterozygosity in MCF-10 cells at “hot spots,” which are commonly observed in human mammary tumors.25,32 Hormonally active but non-metabolizable estrogens exhibited a reduced ability to cause cancer and inhibitors of estrogen metabolism reduce the incidence of estrogen induced kidney tumors in the Syrian hamster.46

Clinical and epidemiologic data also support the possible biologic relevance of ER independent effects of E2 on tumor formation. In patients carrying the BRCA1 gene mutations, bilateral oophorectomy reduces the risk of breast cancer by 53%.58 Since only 10–24% of breast tumors in BRCA1 mutation carriers are ER+,59,60 it has been suggested that the protective effects of oophorectomy might occur independently of ERα.61,62 Women at high risk of breast cancer excrete larger amounts of depurinated adenine and guanine-estradiol conjugates than women at low risk of breast cancer.28,29 In pre-menopausal patients with mutations of multiple estrogen metabolizing genes, the risk of breast cancer has been reported to be 4-fold higher than in controls lacking these mutations.26 However, it should be noted that studies of patients with mutations of only 1 of these genes have reported conflicting results regarding breast cancer risk.27

We acknowledge that several confounding factors could have influenced the interpretation and validity of our results. The Wnt-1 gene is expressed in mammary tissue because of the LTR of the mouse mammary tumor virus. Increased tumor incidence in E2-treated animals might be an indirect result of altered Wnt-1 expression. However, our prior studies carefully examined the expression of Wnt-1 in the ERKO and wild-type animals and showed no alteration of expression.41 We also examined Wnt-1 expression by Western analysis of mammary gland and tumor samples from ER+/+/Wnt-1 and ERKO/Wnt-1 mice in our study and found no change in Wnt-1 levels at different period of experiments or with E2 treatment (Supporting Information Figure 5). These results render unlikely an alteration of Wnt-1 expression as an explanation of our results. Epigenetic breast imprinting in utero caused by an absence of ERα functionality could have resulted in increased susceptibility to cancer in our animal model. However, breast cancer is influenced by factors occurring in utero in women,63,64 and thus our model could possibly reflect such effects.

Several experiments pooled animals from our 2 institutions (UVA and NIEHS) to enhance the sizes of experimental groups. Justification for pooling included the use of identical protocols and the comparability of mean tumor onsets among groups (Supporting Information Table 1). Our studies were conducted over a period of several years (a problem caused by the slow onset of tumors in this model) and involved several independent experiments. This likely explains the variability in time of onset of tumors among the various experiments. This variability clearly confounded the E2 alone versus E2 plus fulvestrant experiment. Nonetheless, the data clearly ruled out residual effects of ERα on tumor formation as assessed by the direct concomitant comparison of E2 alone versus E2 plus fulvestrant even though the 50% tumor incidence time in both groups was delayed from other experiments (i.e., 18–20 months). It should be noted that tumor incidence in the vehicle alone groups in other experiments ranged from 23 to 24 months. While this variability was clearly a limitation of the study, 2 other observations also argue against an ERα effect. Exogenous E2 did not alter breast morphology in ERKO animals (Fig. 1). And 17α-estra-diol enhanced tumor development but did not stimulate uterine weight.

In summary, our data provided strong evidence supporting the principle that breast cancer development can be influenced by E2 via ER independent mechanisms. A necessary component of the proof was to fulfill Koch’s third postulate which requires the disease to be produced by administration of the putative etiologic factor, in this case, E2.65 While other actions of tamoxifen such as epigenetic and genotoxic activities might be involved,66–68 our study provides important mechanistic evidence in support of the use of AIs in preference to the antiestrogens for prevention of breast cancer. As shown in the cartoon in Figure 6, antiestrogens primarily block receptor mediated pathways whereas the AIs block both receptor mediated and receptor independent effects of E2. Two current clinical trials are examining the AIs for prevention of breast cancer.69,70 Finally, a speculative consideration for the future is that blockade of estradiol metabolism with CYP1B1 and 1A1 inhibitors might be a means to reduce breast cancer incidence without blocking formation of E2 itself.

An external file that holds a picture, illustration, etc. Object name is nihms762539f6.jpg

Diagrammatic representation of effects of an antiestrogen and an aromatase inhibitor on prevention of breast cancer. This model postulates that antiestrogens only block ERα mediated effects on breast cancer whereas the aromatase inhibitors, by inhibiting estrogen synthesis, abrogate both ERα mediated as well as the genotoxic effects of estrogen.

Supplementary Material

Yue et al

Click here to view.(247K, mht)

Footnotes

Additional Supporting Information may be found in the online version of this article

References

1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin. 2009;59:225–49. [PubMed] [Google Scholar]
2. Wingo PA, Tong T, Bolden S. Cancer statistics, 1995. CA Cancer J Clin. 1995;45:8–30. [PubMed] [Google Scholar]
3. Boyle P, Ferlay J. Cancer incidence and mortality in Europe, 2004. Ann Oncol. 2005;16:481–8. [PubMed] [Google Scholar]
4. Elmore JG, Armstrong K, Lehman CD, Fletcher SW. Screening for breast cancer. JAMA. 2005;293:1245–56. [PMC free article] [PubMed] [Google Scholar]
5. Cuzick J, International-Breast-Cancer-Intervention-Study A brief review of the International Breast Cancer Intervention Study (IBIS), the other current breast cancer prevention trials, and proposals for future trials. Ann NY Acad Sci. 2001;949:123–33. [PubMed] [Google Scholar]
6. Martin AM, Weber BL. Genetic and hormonal risk factors in breast cancer. J Natl Cancer Inst. 2000;92:1126–35. [PubMed] [Google Scholar]
7. Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A, Hemminki K. Environmental and heritable factors in the causation of cancer-analyses of cohorts of twins from Sweden, Denmark, and Finland. New Engl J Med. 2000;343:78–85. [PubMed] [Google Scholar]
8. Thompson D, Easton D. The genetic epidemiology of breast cancer genes. J Mammary Gland Biol Neoplasia. 2004;9:221–36. [PubMed] [Google Scholar]
9. Zumoff B. Does postmenopausal estrogen administration increase the risk of breast cancer? Contributions of animal, biochemical, and clinical investigative studies to a resolution of the controversy. Proc Soc Exp Biol Med. 1998;217:30–7. [PubMed] [Google Scholar]
10. Hollingsworth AB, Lerner MR, Lightfoot SA, Wilkerson KB, Hanas JS, McCay PB, Brackett DJ. Prevention of DMBA-induced rat mammary carcinomas comparing leuprolide, oophorectomy, and tamoxifen. Breast Cancer Res Treat. 1998;47:63–70. [PubMed] [Google Scholar]
11. Feinleib M. Breast cancer and artificial menopause: a cohort study. J Natl Cancer Inst. 1968;41:315–29. [PubMed] [Google Scholar]
12. Trichopoulos D, MacMahon B, Cole P. Menopause and breast cancer risk. J Natl Cancer Inst. 1972;48:605–13. [PubMed] [Google Scholar]
13. Clemons M, Goss P. Estrogen and the risk of breast cancer. New Engl J Med. 2001;344:276–85. [PubMed] [Google Scholar]
14. Chen WY, Manson JE, Hankinson SE, Rosner B, Holmes MD, Willett WC, Colditz GA. Unopposed estrogen therapy and the risk of invasive breast cancer. Arch Intern Med. 2006;166:1027–32. [PubMed] [Google Scholar]
15. Hulka BS. Epidemiologic analysis of breast and gynecologic cancers. Prog Clin Biol Res. 1997;396:17–29. [PubMed] [Google Scholar]
16. Kaaks R, Rinaldi S, Key TJ, Berrino F, Peeters PHM, Biessy C, Dossus L, Lukanova A, Bingham S, Khaw K-T, Allen NE, Bueno-de-Mesquita HB, et al. Postmenopausal serum androgens, oestrogens and breast cancer risk: the European prospective investigation into cancer and nutrition. Endocr Relat Cancer. 2005;12:1071–82. [PubMed] [Google Scholar]
17. Key T, Appleby P, Barnes I, Reeves G, EHaBCC G Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies. J Natl Cancer Inst. 2002;94:606–16. [PubMed] [Google Scholar]
18. Cummings SR, Eckert S, Krueger KA, Grady D, Powles TJ, Cauley JA, Norton L, Nickelsen T, Bjarnason NH, Morrow M, Lippman ME, Black D, et al. The effect of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA. 1999;281:2189–97. [PubMed] [Google Scholar]
19. Martino S, Cauley JA, Barrett-Connor E, Powles TJ, Mershon J, Disch D, Secrest RJ, Cummings SR. CORE Investigators. Continuing outcomes relevant to Evista: breast cancer incidence in postmenopausal osteoporotic women in a randomized trial of raloxifene. J Natl Cancer Inst. 2004;96:1751–61. [PubMed] [Google Scholar]
20. Howell A, Cuzick J, Baum M, Buzdar A, Dowsett M, Forbes JF, Hoctin-Boes G, Houghton J, Locker GY, Tobias JS, Group AT. Results of the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial after completion of 5 years’ adjuvant treatment for breast cancer. Lancet. 2005;365:60–62. [PubMed] [Google Scholar]
21. Thürlimann B, Keshaviah A, Coates AS, Mouridsen H, Mauriac L, Forbes JF, Paridaens R, Castiglione-Gertsch M, Gelber RD, Rabaglio M, Smith I, et al. Breast International Group (BIG) 1–98 Collaborative Group. A comparison of letrozole and tamoxifen in postmenopausal women with early breast cancer. New Engl J Med. 2005;353:2747–57. [PubMed] [Google Scholar]
22. Preston-Martin S, Pike MC, Ross RK, Henderson BE. Epidemiologic evidence for the increased cell proliferation model of carcinogenesis. Environ Health Perspect Suppl. 1993;101:137–8. [PMC free article] [PubMed] [Google Scholar]
23. Preston-Martin S, Pike MC, Ross RK, Jones PA, Henderson BE. Increased cell division as a cause of human cancer. Cancer Res. 1990;50:7415–21. [PubMed] [Google Scholar]
24. Yager JD, Davidson NE. Estrogen carcinogenesis in breast cancer. New Engl J Med. 2006;354:270–82. [PubMed] [Google Scholar]
25. Cavalieri E, Chakravarti D, Guttenplan J, Jankowiak R, Muti P, Rogan E, Russo J, Santen RJ, Sutter T. Catechol estrogen quinones as initiators of breast and other human cancers: implications for biomarkers of susceptibility and cancer prevention. Biochim Biophys Acta. 2006;1766:63–78. [PubMed] [Google Scholar]
26. Park SK, Yim D-S, Yoon K-S, Choi I-M, Choi J-Y, Yoo K-Y, Noh D-Y, Choe K-J, Ahn S-H, Hirvonen A, Kang D. Combined effect of GSTM1, GSTT1, and COMT genotypes in individual breast cancer risk. Breast Cancer Res Treat. 2004;88:55–62. [PubMed] [Google Scholar]
27. Ritchie MD, Hahn LW, Roodi N, Railey LR, Dupont WD, Parl FF, Moore JH. Multifactor-dimensionality reduction reveals high-order interactions among estrogen-metabolism genes in sporadic breast cancer. Am J Hum Genet. 2001;69:138–47. [PMC free article] [PubMed] [Google Scholar]
28. Gaikwad NW, Yang L, Muti P, Meza JL, Pruthi S, Ingle JN, Rogan EG, Cavalieri EL. The molecular etiology of breast cancer: evidence from biomarkers of risk. Int J Cancer. 2008;122:1949–57. [PMC free article] [PubMed] [Google Scholar]
29. Gaikwad NW, Yang L, Pruthi S, Ingle JN, Sandhu N, Rogan EG, Cavalieri EL. Urine biomarkers of risk in the molecular etiology of breast cancer. Breast Cancer: Basic Clin Res. 2009;3:1–8. [PMC free article] [PubMed] [Google Scholar]
30. Kong LY, Szaniszlo P, Albrecht T, Liehr JG. Frequency and molecular analysis of hprt mutations induced by estradiol in Chinese hamster V79 cells. Int J Oncol. 2000;17:1141–9. [PubMed] [Google Scholar]
31. Zhao Z, Kosinska W, Khmelnitsky M, Cavalieri EL, Rogan EG, Chakravarti D, Sacks PG, Guttenplan JB. Mutagenic activity of 4-hydroxyestradiol, but not 2-hydroxyestradiol, in BB rat2 embryonic cells, and the mutational spectrum of 4-hydroxyestradiol. Chem Res Toxicol. 2006;19:475–9. [PubMed] [Google Scholar]
32. Fernandez SV, Russo IH, Russo J. Estradiol and its metabolites 4-hydroxyestradiol and 2-hydroxyestradiol induce mutations in human breast epithelial cells. Int J Cancer. 2006;118:1862–8. [PubMed] [Google Scholar]
33. Bernard C. An introduction to the study of experimental medicine. New York: The MacMillan Company; 1927. p. 72. [Google Scholar]
34. Bocchinfuso WP, Hively WP, Couse JF, Varmus HE, Korach KS. A mouse mammary tumor virus-Wnt-1 transgene induces mammary gland hyperplasia and tumorigenesis in mice lacking estrogen receptor-alpha. Cancer Res. 1999;59:1869–76. [PubMed] [Google Scholar]
35. Tsukamoto AS, Grosschedl R, Guzman RC, Parslow T, Varmus HE. Expression of the int-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice. Cell. 1988;55:619–25. [PubMed] [Google Scholar]
36. Korach KS, Couse JF, Curtis SW, Washburn TF, Lindzey J, Kimbro KS, Eddy EM, Migliaccio S, Snedeker SM, Lubahn DB, Schomberg DW, Smith EP. Estrogen receptor gene disruption: molecular characterization and experimental and clinical phenotypes. Recent Prog Horm Res. 1996;51:159–88. [PubMed] [Google Scholar]
37. Masamura S, Santner SJ, Heitjan DF, Santen RJ. Estrogen deprivation causes estradiol hypersensitivity in human breast cancer cells. J Clin Endocrinol Metab. 1995;80:2918–25. [PubMed] [Google Scholar]
38. Yue W, Wang J-P, Hamilton CJ, Demers LM, Santen RJ. In situ aromatization enhances breast tumor estradiol levels and cellular proliferation. Cancer Res. 1998;58:927–32. [PubMed] [Google Scholar]
39. Bocchinfuso WP, Lindzey JK, Hewitt SC, Clark JA, Myers PH, Cooper R, Korach KS. Induction of mammary gland development in estrogen receptor-alpha knockout mice. Endocrinology. 2000;141:2982–94. [PubMed] [Google Scholar]
40. Jeng M-H, Shupnik MA, Bender TP, Westin EH, Bandyopadhyay D, Kumar R, Masamura S, Santen RJ. Estrogen receptor expression and function in long-term estrogen-deprived human breast cancer cells. Endocrinology. 1998;139:4164–74. [PubMed] [Google Scholar]
41. Couse JF, Curtis SW, Washburn TF, Lindzey J, Golding TS, Lubahn DB, Smithies O, Korach KS. Analysis of transcription and estrogen insensitivity in the female mouse after targeted disruption of the estrogen receptor gene. Mol Endocrinol. 1995;9:1441–54. [PubMed] [Google Scholar]
42. Shim W-S, Conaway M, Shigeru M, Yue W, Wang J-P, Kumar R, Santen RJ. Estradiol hypersensitivity and mitogen-activated protein kinase expression in long-term estrogen deprived human breast cancer cells in vivo. Endocrinology. 2000;141:396–405. [PubMed] [Google Scholar]
43. Sinha S, Kaseta J, Santner SJ, Demers LM, Bremmer WJ, Santen RJ. Effect of CGS 20267 on ovarian aromatase and gonadotropin levels in the rat. Breast Cancer Res Treat. 1998;48:45–51. [PubMed] [Google Scholar]
44. Merriam GR, MacLusky NJ, Johnson LA, Naftolin F. 2-Hydroxyestradiol-17 alpha and 4-hydroxyestradiol-17 alpha, catechol estrogen analogs with reduced estrogen receptor affinity. Steroids. 1980;36:13–20. [PubMed] [Google Scholar]
45. Key T, Appleby P, Barnes I, Reeves G, Group. EHaBCC Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies. J Natl Cancer Inst. 2002;94:606–16. [PubMed] [Google Scholar]
46. Liehr JG. Is estradiol a genotoxic mutagenic carcinogen? Endocr Rev. 2000;21:40–54. [PubMed] [Google Scholar]
47. Deroo BJ, Korach KS. Estrogen receptors and human disease. J Clin Inv. 2006;116:561–70. [PMC free article] [PubMed] [Google Scholar]
48. Tekmal RR, Liu YG, Nair HB, Jones J, Perla RP, Lubahn DB, Korach KS, Kirma N. Estrogen receptor alpha is required for mammary development and the induction of mammary hyperplasia and epigenetic alterations in the aromatase transgenic mice. J Steroid Biochem Mol Biol. 2005;95:9–15. [PubMed] [Google Scholar]
49. Otto C, Fuchs I, Kauselmann G, Kern H, Zevnik B, Andreasen P, Schwarz G, Altmann H, Klewer M, Schoor M, Vonk R, Fritzemeier K-H. GPR30 does not mediate estrogenic responses in reproductive organs in mice. Biol Reprod. 2009;80:34–41. [PubMed] [Google Scholar]
50. Windahl SH, Andersson N, Chagin AS, Martensson UEA, Carlsten H, Olde B, Swanson C, Moverare-Skrtic S, Savendahl L, Lagerquist MK, Leeb-Lundberg LMF, Ohlsson C. The role of the G protein-coupled receptor GPR30 in the effects of estrogen in ovariectomized mice. Am J Physiol Endocrinol Metab. 2009;296:E490–E96. [PubMed] [Google Scholar]
51. Filardo EJ, Quinn JA, Bland KI, Jr, Frackelton AR. Estrogen-induced activation of Erk-1 and Erk-2 requires the G protein-coupled receptor homolog, GPR30, and occurs via trans-activation of the epidermal growth factor receptor through release of HB-EGF. Mol Endocrinol. 2000;14:1649–60. [PubMed] [Google Scholar]
52. Devanesan P, Santen RJ, Bocchinfuso WP, Korach KS, Rogan EG, Cavalieri E. Catechol estrogen metabolites and conjugates in mammary tumors and hyperplastic tissue from estrogen receptor-alpha knock-out (ERKO)/Wnt-1 mice: implications for initiation of mammary tumors. Carcinogenesis. 2001;22:1573–6. [PubMed] [Google Scholar]
53. Zahid M, Kohli E, Saeed M, Rogan E, Cavalieri E. The greater reactivity of estradiol-3,4-quinone vs. estradiol-2,3-quinone with DNA in the formation of depurinating adducts: implications for tumor-initiating activity. Chem Res Toxicol. 2006;19:164–72. [PubMed] [Google Scholar]
54. Yang L, Gaikwad N, Meza JL, Cavalieri E, Muti P, Trock BJ, Rogan E. Novel biomarkers for risk of prostate cancer. Results from a case-control study. Prostate. 2009;69:41–8. [PMC free article] [PubMed] [Google Scholar]
55. Bradlow HL, Hershcopf R, Martucci C, Fishman J. 16 Alpha-hydroxylation of estradiol: a possible risk marker for breast cancer. Ann NY Acad Sci. 1986;464:138–51. [PubMed] [Google Scholar]
56. Yager JD. Endogenous estrogens as carcinogens through metabolic activation. J Natl Cancer Inst Monogr. 2000;27:67–73. [PubMed] [Google Scholar]
57. Huang Y, Fernandez SV, Goodwin S, Russo PA, Russo IH, Sutter TR, Russo J. Epithelial to mesenchymal transition in human breast epithelial cells transformed by 17β-estradiol. Cancer Res. 2007;67:11147–57. [PubMed] [Google Scholar]
58. Rebbeck TR, Kauff ND, Domchek SM. Meta-analysis of risk reduction estimates associated with risk-reducing salpingo-oophorectomy in BRCA1 or BRCA2 mutation carriers. J Natl Cancer Inst. 2009;101:80–7. [PMC free article] [PubMed] [Google Scholar]
59. Lakhani SR, van de Vijver MJ, Jacquemier J, Anderson TJ, Osin PP, McGuffog L, Easton DF. The pathology of familial breast cancer: predictive value of immunohistochemical markers estrogen receptor, progesterone receptor, HER-2, and p53 in patients with mutations in BRCA1 and BRCA2. J Clin Oncol. 2002;20:2310–18. [PubMed] [Google Scholar]
60. Foulkes WD, Metcalfe K, Sun P, Hanna WM, Lynch HT, Ghadirian P, Tung N, Olopade OI, Weber BL, McLennan J, Olivotto IA, Bégin LR, et al. Estrogen receptor status in BRCA1- and BRCA2-related breast cancer: the influence of age, grade, and histological type. Clin Cancer Res. 2004;10:2029–34. [PubMed] [Google Scholar]
61. Eisen A, Lubinski J, Klijin J, Moller P, Lynch HT, Offit K, Weber BL, Rebbeck T, Neuhausen SL, Ghadirian P, Foulkes WD, Gershoni-Baruch R, et al. Brest cancer risk following bilateral oophorectomy in BRCA1 and BRCA2 mutation carriers: an international case-control study. J Clin Oncol. 2005;23:7491–6. [PubMed] [Google Scholar]
62. King MC, Wieand S, Hale K, Lee M, Walsh T, Owens K, Tait J, Ford L, Dunn BK, Costantino J, Wickerham L, Wolmark N, et al. Tamoxifen and breast cancer incidence among women with inherited mutations in BRCA1 and BRCA2: National Surgical Adjuvant Breast and Bowel Project (NSABP-P1) Breast Cancer Prevention Trial. JAMA. 2001;286:2251–6. [PubMed] [Google Scholar]
63. Michels KB, Trichopoulos D, Robins JM, Rosner BA, Manson JE, Hunter DJ, Colditz GA, Hankinson SE, Speizer FE, Willett WC. Birthweight as a risk factor for breast cancer. Lancet. 1996;348:1542–6. [PubMed] [Google Scholar]
64. Adami HO, Signorello LB, Trichopoulos D. Towards an understanding of breast cancer etiology. Semin Cancer Biol. 1998;8:255–62. [PubMed] [Google Scholar]
65. Koch R. Classics in infectious diseases. The etiology of tuberculosis: Robert Koch. Berlin, Germany, 1882. Rev Infect Dis. 1982;4:1270–4. [PubMed] [Google Scholar]
66. Badia E, Oliva J, Balaguer P, Cavaillès V. Tamoxifen resistance and epigenetic modifications in breast cancer cell lines. Curr Med Chem. 2007;14:3035–43. [PMC free article] [PubMed] [Google Scholar]
67. Wozniak K, Kolacinska A, Blasinska-Morawiec M, Morawiec-Bajda A, Morawiec Z, Zadrozny M, Blasiak J. The DNA-damaging potential of tamoxifen in breast cancer and normal cells. Arch Toxicol. 2007;81:519–27. [PubMed] [Google Scholar]
68. Liu X, Pisha E, Tonetti DA, Yao D, Li Y, Yao J, Burdette JE, Bolton JL. Antiestrogenic and DNA damaging effects induced by tamoxifen and toremifene metabolites. Chem Res Toxicol. 2003;16:832–7. [PubMed] [Google Scholar]
69. Cuzick J. Aromatase inhibitors in prevention–data from the ATAC (arimidex, tamoxifen alone or in combination) trial and the design of IBIS-II (the second International Breast Cancer Intervention Study) Recent Results Cancer Res. 2003;163:96–103. discussion 264–6. [PubMed] [Google Scholar]
70. Ingle JN. Endocrine therapy trials of aromatase inhibitors for breast cancer in the adjuvant and prevention settings. Clin Cancer Res. 2005;11:900s–5s. [PubMed] [Google Scholar]

Formats: