| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
The Stehlin Foundation for Cancer Research, Houston, Texas 77003
 |
Abstract |
|---|
 Top Abstract I. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
|
I. Carcinogenicity of E2 |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
A. Carcinogenicity of E2 in animals
The evidence for the carcinogenic activity of 17ß-estradiol
(E2) in animals has been deemed sufficient by the
IARC to consider this hormone a carcinogen (1, 2). This conclusion is
based on numerous tests of E2 administered to
rodents by oral or subcutaneous administration. For instance, the
administration of E2 to mice increased the
incidence of mammary, pituitary, uterine, cervical, vaginal,
testicular, lymphoid, and bone tumors (3, 4, 5, 6). In rats,
E2 or estrone (E1)
increased the incidence of mammary and/or pituitary tumors (7, 8, 9). In
hamsters, a high incidence of malignant kidney tumors occurred in
intact and castrated males (10, 11, 12, 13) and in ovariectomized females, but
not in intact females (10). In guinea pigs, diffuse fibromyomatous
uterine and abdominal lesions were observed (14).
E2 also induced tumors when administered orally
in the drinking water or in rodent chow (4, 5, 15, 16). All these tumor
models have been developed using pharmacological doses of
E2 with the aim of examining the tumorigenic
activity of this hormone in a relatively short period of time.
The purpose of all these studies was the development of useful and practical animal models for the investigation of mechanistic aspects of hormone-induced tumorigenesis. No animal models have been developed in which tumors are induced by very low doses of E2, presumably because of the cost of maintenance of a large number of animals for such a model and the difficulty of dosing in view of the varying levels of endogenous estrogen in cycling females. The same considerations, however, are also true for almost all other carcinogens known to man, which have been established as carcinogens at high doses in small groups of animals over a short period of time. Although the predictive value of carcinogenicity testing at high doses has been questioned (17, 18), estrogens are nevertheless considered to be carcinogens, based mainly on two types of evidence (1, 2): various tumor types are induced in animals in many organ sites under a variety of treatment conditions as discussed above. Moreover, a consensus is developing that estrogens impart a defined carcinogenic risk to human populations exposed to the low concentrations of estrogens used for medication purposes as discussed below.
B. Carcinogenicity of E2 in humans
Estrogen administration is accepted by most epidemiologists as a
risk factor of human endometrial adenocarcinoma (19, 20). 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 (21). Obesity also increases uterine tumor risk,
most likely because the aromatase activity of adipose cells elevates
tissue and circulating E1 levels (21, 22).
Increasing evidence shows that slightly elevated levels of circulating estrogens are also a risk factor for breast cancer (23, 24). This role of endogenous estrogen in human breast carcinogenesis is supported by risk factors of breast cancer such as high serum or urine estrogen levels (25, 26), the early onset of menstruation, or late menopause (27). While early cohort studies failed to identify an association between serum hormone levels and breast cancer (28, 29) (presumably due to shortcomings of the assay methods), more recent cohort studies have demonstrated strong relationships between endogenous estrogen levels and breast cancer risk (25, 30, 31, 32, 33). The role of endogenous E2 as a risk factor in human breast cancer is reviewed in more detail in the epidemiological literature (Refs. 23, 24, 27 and references cited therein).
Exogenous estrogens, alone or in combination with progestin, also elevate breast cancer risk (34, 35, 36). Progestin added to the estrogen medications does not inhibit mammary carcinogenesis (37) because the former hormone appears to be the primary mitogen of mammary ductal epithelial cells (38), whereas estrogen appears to function in this manner in the uterus. Pike et al. (39) summarized the population-based studies of oral contraceptive use and breast cancer among women under 45 yr of age that had been published through 1990 and derived a weighted average of approximately 3.1% increase in breast cancer risk per year of oral contraceptive use (relative risk estimate: 1.36). The weighted relative risk for young women who consumed oral contraceptives for 10 yr before their first full-term pregnancy was 1.45 compared with nonusers.
Pike et al. (39) also summarized the population-based epidemiological studies that had been published through 1990 and derived a weighted average of the relative breast cancer risk from use of hormone replacement therapy. Of the 10 studies reviewed, 9 showed a positive association and the results of 5 were statistically significant. Based on these studies, the average annual increase in breast cancer risk was 3.1% per year of estrogen replacement therapy use. For women with 10 yr of use, the risk of breast cancer was 1.36 times that of women who have never used these preparations. In a more recent meta-analysis of more than 50 studies, the relationship has been examined between breast cancer risk and estrogen replacement therapy during menopause (40). Although no randomized, controlled, double-blind studies have been conducted, the observational data available show an increased risk of breast cancer with the use of estrogen replacement therapy for more than 5–10 yr. The relative risk of breast cancer under these circumstances increases by about 30%. The absolute risk is small with about one additional breast cancer case/100 women of age 50 who have taken estrogen for at least 10 yr.
C. Conclusion: carcinogenicity of E2
These biological studies in animals and epidemiological studies in
humans all clearly identify E2 as a carcinogen.
Tumors are induced in small groups of animals with pharmacological
doses of E2 in a short period of time. In humans,
slight elevations of circulating estrogen levels caused either by
elevated endogenous production of hormone or by therapeutic doses of
estrogen medications also increase breast or uterine cancer risk (1, 2, 39, 40). This carcinogenic activity of steroidal estrogens is
recognized by the IARC, which classifies the evidence for the
carcinogenicity of steroid estrogens to humans as sufficient (1, 2).
The human epidemiological data point to E2 and other estrogens as only weak carcinogens. This conclusion is not contradicted by laboratory animal tests, which provide only qualitative results given the difficulties with appropriate dosing. Thus in animals, E2 may well be only a weak carcinogen compared with other laboratory carcinogens such as benzo[a]pyrene or 7,12-dimethylbenzanthracene. However, only a weak carcinogenic activity is to be expected because E1, E2, and other steroidal estrogens are endogenous hormones at low picomolar levels and because a strong carcinogenicity would have provided an evolutionary disadvantage to humans and many other species.
|
II. Hormonal Contributions of E2 to Carcinogenesis |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
|
III. E2 as Epigenetic Carcinogen |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
A. Uncontrolled cell proliferation by E2
Tumorigenesis by uncontrolled stimulation of mammary epithelial
cell proliferation has been proposed by Furth (57). A more recent
modification of this mechanistic proposal is the hormone-dependent
receptor-mediated proliferation of mammary epithelial cells carrying
spontaneous replication errors (23). The absence of estrogen receptors
in proliferating human mammary epithelial cells (58, 59) provides
evidence against this mechanistic pathway, at least in the form
proposed. It is possible that estrogens stimulate growth factors by
receptor-mediated pathways in neighboring cells, which in turn
stimulate mammary epithelial cell proliferation (44). However, the
development of synthetic estrogens such as 17
-ethinylestradiol or
2-fluoroestradiol with well maintained hormonal potency but
significantly reduced carcinogenic activity in animal models (10, 13, 60) indicates that the background of spontaneous replication errors of
normal cells may not be sufficient for tumors to develop solely in
response to a proliferative stimulus. More likely, tumors may arise by
hormone receptor-mediated proliferation of cells transformed by
specific genetic damage in addition to background lesions. This view is
consistent with the ability of estrogens to induce various genetic
lesions as described below.
B. Carcinogenesis by covalent modification of E2
receptors
Fishman, Bradlow, and co-workers (61, 62) proposed the induction
of breast cancer by a covalent modification of E2
receptors resulting in a permanent uncontrolled stimulation of mammary
epithelial cell proliferation by receptor-mediated processes. According
to this hypothesis, 16-hydroxyestrone, an E1
metabolite, covalently binds to amino groups of proteins, including the
estrogen receptor protein, and thus permanently stimulates the receptor
and induces hormone-responsive processes, including gene expression and
cell proliferation, in an uncontrolled manner (63). In support of this
mechanism, many studies have been conducted with the aim of correlating
16-hydroxylation of estrogens with tumor induction in laboratory
rodents (61), with incidence of breast cancer and other diseases in
humans (62, 63, 64, 65), and with other parameters of tumorigenesis, such as
induction of oncogene expression (66, 67, 68, 69, 70). In most of the early studies
of Bradlow, Fishman and associates (71, 72), 2- and
16-hydroxylation of E1 were assayed by tritium
release from [2-3H]- and
[16-3H]estrone as substrates, respectively
(71, 72). These assays have never been fully validated against
established product isolation assays of estrogen metabolism but have
been questioned because of spurious release of tritium from
3H-labeled E1 (73, 74, 75, 76, 77). In
addition, the positive correlation between elevated 16-hydroxylation
rates and breast cancer risk observed by Fishman and Bradlow and
associates (62) and Osborne et al. (78) could not be
validated in other laboratories by other researchers (26, 79, 80).
Because of this lack of validation of the assay of
E1 2- and 16-hydroxylation and because of
inadequate corroboration of the molecular epidemiology results by other
laboratories using validated product isolation assays, further work is
needed to determine the validity of the mechanistic hypothesis
of breast cancer induction as proposed by Fishman and
Bradlow.
C. Estrogen-induced chromosomal abnormalities
Barrett and co-workers (81, 82) have reported the neoplastic
transformation of Syrian hamster embryo cells by
E2 and by the synthetic estrogen
diethylstilbestrol without detectable concomitant gene mutations at the
ouabain resistance and 6-thioguanine resistance loci. In contrast,
there was a consistent correlation of cell transformation with
aneuploidy. Both chromosome losses and gains were observed, suggesting
a nondisjunctional mechanism (81, 82). The lack of detectable gene
mutations at defined loci by synthetic and natural estrogens and the
occurrence of aneuploidy concomitant with cell transformation led
Barrett and co-workers (83, 84, 85) to propose an epigenetic pathway of
estrogen-induced carcinogenesis with the following features:
synthetic or natural estrogens including
E2 may disrupt microtubule organization of cells,
resulting in anaphase abnormalities and nondisjunction. The resulting
chromosomal aneuploidy subsequently may induce cell transformation.
However, in a study of the genetic changes occurring during the rare
spontaneous progression of Syrian hamster embryo (SHE) cells from
normal to immortalized and further to neoplastic transformed cells,
Endo et al. (86) observed chromosomal abnormalities in cells
that were not capable of inducing tumors in nude mice. Thus, these
authors (86) concluded that other genetic changes (mutations) were
necessary in addition to chromosomal abnormalities for cells to acquire
tumorigenicity. This view is also consistent with the concept of
Lengauer et al. (87) that aneuploidy is a part of multiple
types of genetic alterations, including base substitutions, deletions,
insertions, gene amplifications, numerical chromosomal changes, and
chromosomal translocations that together make up the genetic
instability leading to human cancer.
D. Epigenotoxic mechanism of estrogen carcinogenesis
Li and co-workers proposed an "epigenotoxic," multistage
scheme for estrogen carcinogenesis in the hamster kidney (88, 89, 90, 91). They
defined an epigenotoxic carcinogen as "an agent that is not involved
in direct (covalent) or indirect interactions with genetic material
but, nevertheless, is able to elicit heritable changes by alternative
mechanisms" (91, 92). According to this hypothesis, which has been
developed mainly by studying the hamster kidney model, estrogen-induced
carcinogenesis involves estrogen-mediated cathepsin D and peroxidase
induction, reparative cell proliferation, aneuploidy and inappropriate
protooncogene and suppressor gene expression such as amplification of
c-myc (91, 92, 93, 94, 95). The sustained overexpression of early
estrogen response genes such as c-fos and c-myc
is thought to be related to estrogen-induced genomic instability as
manifested by amplification of c-myc (95), which is a
mechanism of activation of this gene to a transforming oncogene. Tumors
are thought to arise from the distinct growth advantage of cells
overexpressing c-fos, c-myc, and c-jun
and other early estrogen response genes.
Li et al. (94, 96, 97, 98) postulated the induction of genetic instability by mechanisms other than direct covalent or indirect interactions of estrogen metabolites with genetic material because they detected only very low rates of metabolic conversion of E2 to the catechol metabolites 2- and 4-hydroxyestradiol, the precursors of reactive semiquinone and quinone intermediates (as discussed below). Their hypothesis is also based on their inability to confirm the formation of estrogen-induced DNA adducts (99) by 32P-postlabeling assay as described earlier by Liehr and co-workers (100, 101). Unfortunately, rates of metabolic conversion of estrogens to catechol metabolites determined by Li and co-workers (94, 96, 97, 98, 102) were measured using an unvalidated, indirect radioenzymatic assay that converts the unstable catecholestrogens to more stable methoxyestrogens catalyzed by catechol-O-methyltransferase (103, 104). This assay has been shown to underestimate rates of catecholestrogen formation by 2 to 3 orders of magnitude (105). Specifically, 4-hydroxylation of estrogens cannot be detected by this radioenzymatic assay in microsomal preparations expressing both estrogen 2- and 4-hydroxylase activity (105), because 2-hydroxyestradiol inhibits the catechol-O-methyltransferase-mediated methylation of 4-hydroxyestradiol and thus inhibits formation of assayable product (106).
In contrast, much higher rates of catechol formation than those described by Li et al. (94, 96, 97, 98, 102) were obtained in target organs where estrogens induce tumors using product isolation assays fully validated and cross-checked in several laboratories (105, 107, 108, 109, 110, 111, 112). In these studies, assays were validated in the same (Liehr) laboratory using two different product isolation procedures, a gas chromatography-based, and a TLC-based method or in two different laboratories (Liehr and Weisz) using the same hamster microsomal preparations (105, 109, 111). Finally, the rates of 2- and 4-hydroxylation of E2 determined in these studies are consistent with rates published by other authors [as reviewed by Zhu and Conney (112)]. Moreover, the covalent binding of estrogens including catecholestrogen metabolites to DNA, initially published by Liehr and associates (100, 101), has now been confirmed by Cavalieri et al. (see discussion below) and by Hayashi et al. (113).
Li and associates (89, 90, 94) questioned the ability of estrogens and their metabolites to induce DNA damage in the carcinogenesis process on the grounds that insufficient concentrations of E2 are present in target tissues of hormonal cancer and that rates of its conversion to catecholestrogens are too low to result in significant amounts of genotoxic metabolites. This critique is based upon the measurement of plasma E2 levels (114) and the assumption of concordance between plasma and tissue E2 levels (115). However, this assumption is clearly not correct, since, for instance, in premenopausal women, the ratio of mammary tissue to plasma E2 levels approximates 1:1, whereas in postmenopausal women the ratio is 10–50:1 (116). Thus, local concentrations of E2 in human mammary tissue and in breast tumors depend more likely on the aromatase activity of individual mammary cells (autocrine or paracrine action) than on the ovarian hormone supply. Further evidence in support of a predominant local production of hormone is provided by the high aromatase activity of individual mammary cells (117, 118, 119, 120). The importance of mammary aromatase activity for local E2 concentrations has also been documented by studies in nude mice inoculated on one side with MCF-7 breast cancer cells stably transfected with aromatase and on the other side with sham-transfected cells (121). Administration of the aromatase substrate androstenedione stimulated the proliferation only of the aromatase-positive MCF-7 tumors. The relative importance of in situ production of E2 vs. uptake from plasma was examined by administering SILASTIC implants of this hormone (121). The E2 levels were more than 4-fold higher in aromatase-positive than -negative tumors. These experiments identify the local production of E2 in hormone-responsive tissue including mammary gland as a more important determinant of tissue E2 levels than the hormone supplied by circulation.
The metabolic conversion of E2 to catecholestrogen metabolites has been underestimated by Li et al. as discussed above. A specific conversion of E2 to the carcinogenic catechol metabolite 4-hydroxyestradiol by a specific cytochrome P450 has been detected in organs of rodents where estrogens induce tumors and in human breast and uterine tissue, as discussed below. This specific metabolic process may also result in elevated local concentrations of catecholestrogen metabolites. Additional research is needed to correlate local tissue and cellular estrogen and estrogen metabolite concentrations with tumorigenesis.
E. Conclusion: E2 as epigenetic carcinogen
In summary, the proposals of estrogen as an epigenetic
(epigenotoxic) carcinogen as discussed above all emphasize features
that most likely participate in, but may not be sufficient for, the
development of hormone-responsive cancers. There is widespread
agreement that the action of estrogens as hormones by receptor-mediated
processes is necessary for oncogenesis. Also, the induction by
estrogens including E2 of genetic lesions such as
c-myc gene amplification or aneuploidy is a part of genetic
changes necessary for the induction of carcinogenesis, as postulated by
Lengauer et al. (87) and discussed below. The early reports
of a lack of DNA reactivity and of mutational effects of estrogens or
their metabolites, which served as the basis for the epigenetic
mechanistic hypotheses outlined above, may have been based on
inadequate experimental design and/or insufficiently sensitive
detection technology. In more recent studies from various laboratories,
sufficient evidence has been obtained, which demonstrates the ability
of estrogens to undergo metabolic activation and to directly or
indirectly modify DNA as discussed below.
Several studies in support of epigenetic mechanistic hypotheses have
been carried out with poorly validated and inadequate assays. For
instance, values for 16-hydroxylation and catecholestrogen formation
by radiometric or radioenzymatic assays have been obtained with
unvalidated assays and have not been corroborated in other
laboratories. Moreover, the roles of local formation and local
concentrations of estrogens and their metabolites have not been fully
examined in relation to the carcinogenesis process. Finally, breast
cancer is a complex disease. It is more likely that estrogens act in a
dual function as hormones, as outlined above, and as carcinogens, as
outlined below, with both these characteristics necessary for
completion of tumor development.
|
IV. E2 as Genotoxic Carcinogen |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
A. Metabolic conversion of E2 to catecholestrogens
2-Hydroxylation of steroidal estrogens is the major metabolic
oxidation of estrogenic hormones in most mammalian species as
illustrated in Fig. 1
(112, 125, 126). In
human or hamster liver, this oxidation is catalyzed by cytochrome P450
3A enzymes, whereas cytochrome P450 1A enzymes are the predominant
estrogen 2-hydroxylases in extrahepatic tissues (127, 128, 129, 130). These
estrogen 2-hydroxylases convert E2 to
approximately 80–85% 2-hydroxyestradiol and, due to a lack of
specificity of the enzyme(s), to 15–20% 4-hydroxyestradiol (76, 109).
In contrast, specific estrogen 4-hydroxylase(s), which convert
E2 mainly to 4-hydroxyestradiol (131), have been
identified (107, 108, 109) in those organs of rodents in which chronic
estrogen exposure induces malignant or benign tumors: hamster kidney
(10), mouse uterus (124, 132), or rat pituitary (133). The specific and
local formation of 4-hydroxylated estrogens is important, because
4-hydroxyestradiol is as carcinogenic as E2 in
the hamster kidney tumor model (52, 122, 123), whereas in the mouse
uterus the 4-hydroxylated estrogen was 9 times more carcinogenic than
the parent hormone (124).
|
B. Metabolic activation of catecholestrogens
Catecholestrogens are capable of metabolic redox cycling as
illustrated for 4-hydroxyestradiol in Fig. 2. This process consists of the organic
hydroperoxide-dependent oxidation of the catecholestrogen (the
hydroquinone) to the quinone, and the NADPH-dependent cytochrome P450
reductase-catalyzed reduction of the quinone intermediate back to
the hydroquinone (140). The semiquinone free radical is an intermediate
in each of these metabolic conversions. The estrogen semiquinone is a
reactive species and may react with molecular oxygen and form quinone
and superoxide radicals (141). Alternatively, nonenzymatic redox
couples between copper ions and catecholestrogens also generate
reactive oxygen radicals (142, 143). Thus, metal ion-catalyzed or
enzyme-mediated redox cycling is a mechanism of metabolic activation
resulting in the continuous formation of free radicals from possibly
small amounts of catecholestrogen substrates that are reused in this
process. This cycling reaction may go on indefinitely, depending on the
availability of catechol substrate and organic hydroperoxide cofactor
or metal ion for the oxidation step of the cycle.
|
C. Free radical-mediated DNA damage induced by
estrogens
Several types of free radical-mediated DNA damage are induced by
estrogens and/or their metabolites and are listed in Table 1. For instance, DNA single-strand breaks
are induced in MCF-7 human breast cancer cells in culture by
3,4-estrone quinone (146, 147), formed by oxidative metabolism of
4-hydroxyestrone. This type of DNA damage is also induced in 
X-174
RFI plasmid DNA by 2-hydroxyestradiol and 10 µM
Cu(II)sulfate and in vivo in the kidney of Syrian hamsters
treated with either E2 or 4-hydroxyestradiol many
months before the development of neoplasms in this organ (142, 148). A
tissue-specific induction of DNA single-strand breaks was observed in
the dorsolateral prostates of Nobel rats treated with
E2 plus testosterone for 16 weeks before the
development of E2 + testosterone-related prostate cancer in this
tissue (149). In contrast, this lesion was not detected in
ventral prostate, where cancers do not develop under these conditions,
and was not induced in either tissue by androgen treatment alone.
|
D. Indirect DNA adduct formation induced by E2
In addition to the direct free radical-initiated DNA damage
described above, estrogen exposure also results in indirect DNA adduct
formation (158, 162, 163). Some of these adducts have been formed by
reactive aldehydes such as malondialdehyde, which are generated by
decomposition of lipid peroxides produced by estrogen treatment of the
animals. For instance, malondialdehyde-DNA adduct levels were increased
over control values in hamsters treated with E2
(158). Adducts of this type have also been identified in mammary DNA of
breast cancer patients (164).
E. Direct estrogen DNA adducts
In addition to indirect DNA adduct formation, estrogen metabolites
also are capable of direct covalent binding to DNA. As shown in Fig. 2
, catecholestrogens may be oxidized to quinone intermediates, which may
covalently bind to DNA in vitro (165, 166). The adducts of
estrone-3,4-quinone, formed by oxidation of 4-hydroxyestrone, are
unstable and decompose to form apurinic sites (166, 167, 168) consistent
with adduction characteristics of carcinogenic hydrocarbons (169, 170).
In contrast, the DNA adducts of estrone-2,3-quinone, formed by
oxidation of 2-hydroxyestrone, are chemically stable and do not
generate appreciable amounts of apurinic sites. The formation of the
mutagenic apurinic sites by the carcinogenic 4-hydroxyestrogen
metabolites and the generation of stable DNA adducts by the weakly or
noncarcinogenic 2-hydroxyestrogen metabolites is consistent with adduct
patterns of carcinogenic vs. weakly carcinogenic or
noncarcinogenic hydrocarbons, respectively (169, 170). This adduct
pattern has been taken as evidence for a mechanism of carcinogenesis by
unstable adduct formation of 4-hydroxylated estrogens, induction of
gene mutation, and subsequent tumor initiation (167, 169, 170). In
incubations of estrone-3,4-quinone with the COIII gene, the estrogen
metabolite was covalently bound mainly to guanine (171). Furthermore,
the in vitro replication of the COIII template containing
these adducts was obstructed, indicating an arrest of DNA polymerase by
these estrogen metabolite-guanine lesions. 4-Hydroxyequilenin, a
metabolite of the equine steroidal estrogen equilenin (172), which is a
component of the common estrogen replacement medication
Premarin (Ayerst Laboratories, New York, NY), forms
unusual cyclic adducts with DNA in vitro (173, 174). Taken
together, these data demonstrate that steroidal estrogens may be
metabolically activated and form estrogen-DNA adducts in
vitro (167, 168) and in vivo (175).
|
V. E2-Induced Chromosomal or Genetic Mutations |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
.
|
Synthetic and natural estrogens including E2 also
induce structural chromosomal aberrations in addition to the numerical
changes discussed above. For instance, perinatal exposure of rodents to
estrogen results in chromosomal aberrations in the same target tissues
in which tumors subsequently develop (181, 182). Treatment of Syrian
hamsters with E2 also leads to structural
chromosomal aberrations such as deletions, inversions, and
translocations in kidney cells long before tumors develop in this organ
(93, 177, 178). The lower frequency of chromosomal aberrations in the
hamster kidney cortex induced by 17-ethinylestradiol compared with
frequencies induced by E2 or diethylstilbestrol
(178) points to a role of catechol metabolites in the genesis of this
lesion, because the rate of conversion of this synthetic estrogen to 2-
and 4-hydroxylated metabolites by hamster kidney microsomes is one
third the rate observed with the natural hormone (110) and correlates
with the low carcinogenic activity compared with that of
E2 (10). In summary, E2
induces aneuploidy and structural chromosomal changes (81, 85, 93, 176, 177, 178, 179, 180, 181, 182, 183), which may be viewed as part of a larger pattern of various
types of covalent damage to genetic material at the DNA or chromosome
level occurring in vitro or in vivo. These types
of chromosomal aberrations by themselves may not be sufficient for
tumors to develop (86) but may contribute to tumorigenesis by
compromising the integrity of the genetic material (87).
B. E2-induced gene mutations
The mutagenic potential of estrogens including the natural hormone
E2 has been highly controversial. Early studies
of the mutagenic activity of estrogens were all negative,
i.e., neither E2 nor its catechol
metabolites induced point mutations in the Ames bacterial reversion
test (51, 52, 53), in Syrian hamster embryo cells (81, 82, 83, 84, 85), or in V79
Chinese hamster cells (53, 54, 55) in the concentration ranges tested.
Estrogens including E2 were classified as
nonmutagenic and nongenotoxic based on this failure to induce gene
mutations (49, 50, 83, 84). However, these results are not consistent
with the various types of DNA damage discussed above, which are known
to be potentially mutagenic.
More recent observations point to estrogen-induced gene mutations in
several test systems. For instance, diethylstilbestrol induces
mutations at the
Na+/K+-ATPase locus (184).
Moreover, either E2 or the synthetic estrogen
diethylstilbestrol are mutagenic and inactivate the gpt transgene of
the Chinese hamster G12 cell line (185, 186). Specifically, the
inactivation of the gpt transgene is caused by a pattern of mutations
unique for a given mutagen. Diethylstilbestrol induces approximately
37% deletion and 25% methylation silencing among independent
6-thioguanine-resistant clones, whereas E2
produced 53% deletions and only a few methylation-silenced mutants
(186, 187). 4-Hydroxyestrone and 16-hydroxyestrone both induce
methotrexate resistance in MCF-7 breast cancer cells with an
enhancement factor of 88 and 2-hydroxyestrone with an enhancement
factor of 33 (188). In contrast, the parent hormone
E2 showed only a slight effect with an
enhancement factor of 3.2. These data clearly implicate the metabolic
activation of parent estrogens to catecholestrogens in the induction of
this type of mutation. The induction of methotrexate resistance did not
correlate with receptor-mediated responses (188). Both
E2 and 16-hydroxyestrone stimulated expression
of the pS2 gene, whereas 2- and 4-hydroxyestrone did not do so. The
authors concluded that the development of methotrexate resistance was
possible in the absence of estrogen receptors (188).
The testing of E2 at various concentrations demonstrated a low frequency of mutations of the hprt gene by this hormone at the lowest dose assayed (10-10 M E2) in V79 Chinese hamster lung cells, whereas at higher doses this effect was not observed (55). This mutagenic activity of E2 at that low dose but not at elevated doses was independently confirmed (T. Albrecht and J.G. Liehr, unpublished). Moreover, Markides et al. (159) provided an explanation for this concentration dependence of the mutagenic activity of E2 by demonstrating that only the catecholestrogen metabolites 2- and 4-hydroxyestradiol exhibit prooxidant characteristics and only at low physiological concentrations and in the presence of metal ions. In contrast, at higher micromolar concentrations, all estrogens, including catecholestrogen metabolites, act as antioxidants. These data may provide an explanation for the failure of estrogens to induce mutations in previous studies, because only micromolar concentrations of E2 have been examined in these previous assays (54, 55, 81, 82, 83, 84, 85).
In other more recent studies, a 2.4- to 3.6-fold amplification of the c-myc gene was detected by Southern blot analysis in 67% of primary renal tumors induced by E2 or diethylstilbestrol treatment of Syrian hamsters (95). The c-myc gene was localized to hamster chromosome 6qb by fluorescence in situ hybridization. This chromosome 6 has a high frequency of trisomies and tetrasomies in the kidney of hamsters treated for at least 5 months and in renal tumors (95). Li et al. (95) concluded that estrogen-induced genomic instability, as demonstrated by c-myc gene amplification and concurrent chromosomal changes, was a key element in carcinogenic processes induced by estrogens (95). In the same animal model, E2 has been shown to alter tandem repeat sequences of DNA (microsatellite instability) in premalignant kidney of hamsters treated with this hormone for 3 and 4 months and subsequently in kidney tumors that had developed after 7 months (189). This type of mutation has been shown to be inducible by free radicals (190) and may have been generated by metabolic redox cycling of estrogen metabolites (140, 141). This type of mutation is important because microsatellite instability has been detected in 100% of genital tract tumors induced in the daughters of women treated with the transplacental carcinogen and synthetic estrogen, diethylstilbestrol (191).
Taken together, these data demonstrate that estrogens, including the natural hormone E2, induce multiple forms of genetic lesions including DNA microsatellite instability, DNA sequence deletions, gene amplification, chromosomal aberrations, and changes in the number of chromosomes. Such genetic alterations have recently been proposed by Lengauer, Kinzler, and Vogelstein (87) to be the basis of most human cancers. It is possible that estrogens may only be weak mutagens. However, only a low frequency of mutations is expected from natural circulating hormones. Thus, this area of research requires additional studies with more refined assay conditions designed to detect weak mutagens. Moreover, several types of mutations, such as DNA microsatellite instability or gene amplification, may have been missed by classical gene mutation assays because these tests are designed to detect only single-point mutations in only one specific gene.
|
VI. Indirect Evidence for the Genotoxic and Mutagenic Activity of E2 |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
1. There are several synthetic estrogens such as 2-fluoroestradiol and
17-ethinylestradiol, which exhibit comparable hormonal potency, yet
poor carcinogenicity compared with E2, which
induces a 100% tumor incidence in the Syrian hamster kidney model (10, 13, 60). These poorly carcinogenic estrogens, 2-fluoroestradiol and
17-ethinylestradiol, have a decreased catecholestrogen formation
compared to that of the parent estrogens (101, 110, 192). The
existence of such poorly carcinogenic, yet hormonally potent, synthetic
estrogens directly contradicts tumor incidence mediated solely by
hormone receptor pathways. Their altered metabolism implicates
catecholestrogen metabolites to play a crucial role in tumor
initiation.
2. The induction of kidney tumors in hamsters by
E2 may be completely prevented by
coadministration of -naphthoflavone, an inhibitor of cytochrome P450
1A-mediated catecholestrogen formation, or inhibited by ascorbic acid
(vitamin C), a free radical scavenger and reductant of the DNA-reactive
catecholestrogen quinone metabolites (160, 161, 193). This modulation
of E2-induced carcinogenesis by decreasing
concentrations of catecholestrogen or catecholestrogen quinone
metabolites further supports the concept of tumor initiation by
reactive metabolic intermediates of this hormone.
In this context, it is noteworthy that Ah receptor agonists such as 2,3,7,8-tetrachlorodibenzo-p-dioxin, which induce the metabolic conversion of E2 to 4-hydroxyestradiol (136), do not appear to induce mammary carcinogenesis. To the contrary, in rats exposed to this chemical, spontaneous mammary and uterine tumorigenesis is decreased over controls, and the sizes of chemically induced tumors are reduced (194, 195). In humans, short-term exposure to this organochlorine compound (e.g., after an explosion of a chemical manufacturing facility in Seveso, Italy) may provide protection from mammary cancer (196), whereas long-term occupational exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin slightly elevates the risk for breast cancer (197, 198, 199). These apparently conflicting results may be due to various biological effects of this organochlorine chemical on the carcinogenesis process. In addition to stimulation of estrogen hydroxylation via the Ah receptor, it may also act as an antiestrogen and inhibit a variety of hormone receptor-mediated responses [reviewed by Safe (200)]. Thus, it is possible that this compound may stimulate tumor initiation by inducing metabolic activation, but then may inhibit the completion of tumor development by its hormone antagonism. These data illustrate that chemical modulators of estrogen-induced carcinogenesis may be useful for the study of mechanistic aspects only if they alter narrowly defined biological parameters. Mechanistic conclusions cannot be drawn from studies of agents with multiple biochemical and endocrine effects.
3. Estrogen receptors in the human mammary epithelium are localized in cells distinct and different from cells expressing markers of cell proliferation (58, 59). Moreover, SHE cells, which have been used to study the mechanism of estrogen-induced cell transformation (81, 82, 83, 84, 85), do not express measurable levels of estrogen receptor and estrogen treatment is not mitogenic to these cells (201). In this cell line, either estrogens or the hormone antagonists tamoxifen or ICI 164,384 induce morphological transformation and aneuploidy (176). These data indicate that estrogen-induced cell transformation and aneuploidy arise in cells early in the carcinogenesis process and do not require estrogen receptors. Moreover, receptor-mediated processes may be linked indirectly rather than directly to mammary cell proliferation during mammary oncogenesis as discussed in Section III.A.
4. The strongest evidence for an additional (carcinogenic) role of
estrogens in hormone-induced oncogenesis is provided by experiments in
transgenic mice. Mice overexpressing the Wnt-1 gene produce elevated
amounts of a protein important in cell signaling during embryonal
development. These mice develop mammary tumors with high incidence
within a few months after birth (202). These transgenic mice have been
cross-bred with estrogen receptor- knockout (ERKO) mice to examine
the role of estrogen receptors in breast tumor incidence (203). The
incidence of mammary tumors was delayed, but not eliminated, in the
cross-bred animals (48 weeks) compared with mice only overexpressing
the Wnt-1 gene (24 weeks). When the Wnt-1 overexpressing/estrogen
receptor- knockout cross-bred animals were ovariectomized to reduce
their E2 production, the mammary tumor incidence
was significantly reduced (203). The authors concluded that ectopic
expression of the Wnt-1 protooncogene induces mammary tumors in
transgenic mice in the absence of estrogen receptors. Moreover,
decreases in circulating E2 concentrations
achieved by ovariectomy of these animals decrease this tumor incidence.
The data support a role of genotoxicity of E2 in
mammary carcinogenesis and contradict oncogenesis in this organ
mediated solely by hormone receptor pathways.
All these data are consistent with and support the conclusion that genotoxic processes and gene mutations participate and play a tumor-initiating role in the induction of mammary tumors by the natural hormone E2. These data are inconsistent with tumor induction solely based on hormonal receptor-mediated processes (as postulated previously (23, 49, 50).
Estrogen-induced carcinogenesis in the mammary gland and in other organ
sites likely is complex and requires both receptor-mediated
and genotoxic events for neoplastic development. Indirect evidence in
support of this dual role of estrogens as hormones and as
tumor-initiating chemicals is the inhibition of tumor incidence either
by: 1) hormone antagonists interfering with receptor-stimulated cell
proliferation (44, 47, 48); or 2) inhibitors of metabolic activation of
estrogens (160, 161, 193). An important aspect of this proposed action
of estrogens is that inhibition of either of these events will inhibit
oncogenesis, albeit at a different stage of neoplastic development. The
modulation of receptor-mediated tumor cell proliferation by hormone
antagonists thus may leave intact the accumulated genetic lesions
induced by estrogens and/or other carcinogens. This concept is
supported by the inhibition of estrogen-induced renal carcinogenesis in
Syrian hamsters by tamoxifen without concomitant decrease in
estrogen-induced DNA adduct levels (48). In contrast, inhibitors of
metabolic activation of estrogens are proposed to act by inhibiting the
accumulation of potentially mutagenic DNA alterations induced by
estrogens. This concept is supported by the inhibition of
estrogen-induced renal tumorigenesis in hamsters by -napthoflavone
or ascorbic acid (vitamin C) (160, 161, 193). It is also supported by
the action of poorly carcinogenic, yet hormonally potent, synthetic
estrogens 2-fluoroestradiol or 17-ethinylestradiol (10, 13, 60).
Inhibitors of estrogen metabolism have not yet been explored for the
prevention of breast and other hormone-associated cancer in humans and
may offer an attractive alternative to hormone antagonists, because
they may inhibit mammary tumorigenesis at an early stage.
|
VII. Summary and Conclusion |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
The multiple forms of DNA damage induced by catecholestrogen metabolites after metabolic activation to quinone-reactive intermediates provide strong support for the conclusion that the natural estrogenic hormone E2 exerts genotoxicity most likely via metabolic activation to catecholestrogens. The induction of gene mutation by estrogens outlined above also supports this conclusion but requires further work and experimental detail. We do not yet know which critical genes are mutated by estrogen or their metabolites in the oncogenesis process and the mechanism of induction of mutations. Much additional research is needed to sketch the mechanistic events resulting in hormone-associated cancer.
Despite these deficiencies in our knowledge of the mutagenic activity of E2, the human epidemiological studies point to estrogen as a weak carcinogen adding approximately 3% breast cancer risk/year of estrogen exposure (39, 40). These human data are in line with animal carcinogenicity and cell culture data. They are also in agreement with the more moderate levels of DNA modification by estrogen compared with the substantial genotoxicity of potent carcinogens such as benzo[a]pyrene or 7,12-dimethylbenzanthracene (169, 170). The weak mutagenic activity of E2 at the hprt locus of V79 cells also points to E2 as a weak mutagen/carcinogen (55). In a comparison of the induction of aneuploidy by E2 in human and hamster fibroblasts, Tsutsui et al. noted the much weaker induction of this genetic instability in the human compared with the rodent cells (183). This weak mutagenic activity of E2 explains the difficulties of previous workers to detect any mutational events and the underlying genotoxicity induced by E2 and makes understandable the resulting eagerness to classify estrogens as epigenetic, nonmutagenic carcinogens. However, this classification will have to be reconsidered in light of the more recent evidence cited above. The weak mutagenic activity of E2 is also understandable in view of the role of this endogenous hormone in many physiological processes. A high mutagenic and carcinogenic activity of E2 would not have permitted the existence of many higher life forms including that of the human species.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Studies in the author’s laboratory have been supported by grants from
the National Cancer Institute (NIH Grants CA-63129 and
CA-74971). 
|
References |
|---|
|
TopAbstractI. Carcinogenicity of E2II. Hormonal Contributions of...III. E2 as Epigenetic...IV. E2 as Genotoxic...V. E2-Induced Chromosomal or...VI. Indirect Evidence for...VII. Summary and ConclusionReferences |
|---|
-ergocryptine. Br J Cancer 35:322–328[Medline]
