2006 Study Abstract
The synthetic estrogen diethylstilbestrol (DES) is a potent perinatal endocrine disruptor. In humans and experimental animals, exposure to DES during critical periods of reproductive tract differentiation permanently alters estrogen target tissues and results in long-term abnormalities such as uterine neoplasia that are not manifested until later in life. Using the developmentally exposed DES mouse, multiple mechanisms have been identified that play a role in its carcinogenic and toxic effects. Analysis of the DES murine uterus reveals altered gene expression pathways that include an estrogen-regulated component. Thus, perinatal DES exposure, especially at low doses, offers the opportunity to study effects caused by weaker environmental estrogens and provides an example of the emerging scientific field termed the developmental origin of adult disease. As a model endocrine disruptor, it is of particular interest that even low doses of DES increase uterine tumor incidence. Additional studies have verified that DES is not unique; when other environmental estrogens are tested at equal estrogenic doses, developmental exposure results in increased incidence of uterine neoplasia similar to that caused by DES. Interestingly, our data suggest that this increased susceptibility for tumors is passed on from the maternal lineage to subsequent generations of male and female descendants; the mechanisms involved in these transgenerational events include genetic and epigenetic events. Together, our data point out the unique sensitivity of the developing organism to endocrine-disrupting chemicals, the occurrence of long-term effects after developmental exposure, and the possibility for adverse effects to be transmitted to subsequent generations.
OVER THE LAST decade, mounting evidence from wildlife, epidemiological, laboratory animal, and in vitro studies suggests that numerous environmental and dietary chemicals can interfere with an organism’s complex endocrine signaling mechanisms and result in adverse consequences, especially alterations in reproductive tract tissues and function (1, 2, 3, 4, 5, 6). Because reduced fertility and reproductive problems are well acknowledged consequences of estrogen exposure as demonstrated by the estrogen mimics o,p′-dichlorodiphenyltrichloroethane (DDT) (7), the phytoestrogens in clover (8), and the seminal observations of the adverse effects of diethylstilbestrol (DES) on human health (9), initial concern focused on chemicals with estrogenic activity; however, it has become increasingly evident that this is not the only endocrine mode of action that merits attention. Interest has broadened to include chemicals that mimic or interfere with the normal actions of all endocrine hormones including estrogens, androgens, progestins, and thyroid, hypothalamic, and pituitary hormones; these chemicals are now collectively referred to as endocrine disruptors. With over 80,000 chemicals in commercial use in the United States alone today, only a few have been tested for endocrine-disruptor activity. It is generally assumed that most of these chemicals are not likely to pose a significant health risk at the levels of environmental exposures that exist; however, in reality, the full extent of the health consequences of these chemicals is simply unknown. Furthermore, we are just beginning to understand how chemicals act as endocrine disruptors and to appreciate the complexities of endocrine signaling mechanisms.
Sensitivity of the Developing Organism and Developmental Basis of Adult Disease:
Although adult exposure to endocrine-disrupting chemicals is important, the focus on the fetus and/or neonate is of primary concern because developing organisms are extremely sensitive to perturbation by chemicals with hormone-like activity. For example, an adverse effect may be most pronounced in the developing organism and occur at concentrations of the chemical that are far below levels that would be considered harmful in the adult. The exquisite sensitivity of the developing fetus and neonate has been described by Howard Bern in a chapter titled The Fragile Fetus (10) and is suggested to be due to numerous factors including undeveloped DNA repair mechanisms, an immature immune system, lack of detoxifying enzymes, primitive liver metabolism, lack of the development of the blood/brain barrier, and an increased metabolic rate.
Unique problems may be encountered when studying chemical exposures of the fetus and neonate undergoing critical developmental windows of differentiation. 1) Extrapolation of risks may be difficult because effects may not follow a monotonic dose-response relationship typically seen in toxicity studies; for example, higher concentrations of chemicals may show no effect, whereas lower concentrations do exhibit an effect. 2) The test chemical may have an entirely different effect in the embryo, fetus, or perinatal organism compared with effects seen in adults. 3) Effects may be manifested in offspring but not in their exposed parent. 4) Timing of exposure in the developing organism is critical in determining the final outcome in the adult organism. 5) Although critical exposure occurs during embryonic/fetal/neonatal development, manifestation of the effects may not be seen until much later in life (the developmental basis of adult disease). Developmental exposure can thus lead to a number of adverse effects in adults, which may include tumors in endocrine target tissues and adverse reproductive effects in males and females.
The scientific hypothesis that adult health and disease have an etiology arising in fetal or early neonatal development is not unique to the field of endocrine disruption. In the late 1980s, reports gained prominence by suggesting that the fetal environment, as reflected by low birth size and poor nutrition, was related to increased risk of noncommunicable diseases later in adult life; associations with coronary heart disease quickly extended to include type 2 diabetes, osteoporosis, and metabolic dysfunction (11). These findings led to the development of the developmental origins of health and disease paradigm in which a substantial research effort focused on the perinatal influences on chronic disease (12, 13, 14). Perinatal effects are no longer viewed in terms of just teratogenic changes or acute birth injury such as thalidomide-induced limb malformations, but whether changes induced in early development (preimplantation through prepubertal stages) may lead to life-long anomalies. Certainly unique problems exist in studying chemical exposures during development and their relationship to adult disease, but despite these difficulties, research findings continue to support the idea that environmental chemicals, in particular those with estrogenic activity, can have endocrine-disrupting effects that result in long-term health consequences.
As an example, the profound effects of estrogens on the developing reproductive tract have been demonstrated by prenatal exposure to DES (for review, see Refs. 15, 16, 17). Although DES effects were well recognized and firmly documented long before the proposed developmental origins of health and disease paradigm, DES clearly demonstrates that chemical exposure, in addition to nutrition and other perinatal factors, can significantly alter the developing organism and cause long-term effects in the adult.
DES as a Model Estrogenic Endocrine Disruptor:
For almost 30 yr, clinicians prescribed DES to women with high-risk pregnancies to prevent miscarriages and other complications of pregnancy. In 1971, a clinical report associated DES with a rare form of reproductive tract cancer termed vaginal adenocarcinoma, which was detected in a small number (<0.1%) of adolescent daughters of women who had taken the drug while pregnant (9). Subsequently, DES was also linked to more frequent benign reproductive tract problems in an estimated 95% of the DES-exposed daughters; reproductive organ dysfunction, abnormal pregnancies, reduced fertility, and disorders of the immune system were reported. Similarly, DES-exposed male offspring demonstrated structural, functional, and cellular abnormalities after prenatal exposure including hypospadias, microphallus, retained testes, inflammation, and decreased fertility (for review, see Ref. 17). DES became the first example of an in utero estrogenic toxicant in humans; it was shown to cross the placenta and induce a direct effect on the developing fetus. Based on the medical catastrophe it caused, DES can be viewed as the original endocrine-disrupting chemical. DES is no longer used clinically to prevent miscarriage, but a major concern remains that, as DES-exposed women age and reach the time at which the incidence of reproductive organ cancers normally increase, they will show a much higher incidence of cancer than unexposed individuals. Furthermore, the possibility of second-generation effects has been reported (18, 19, 20, 21), which puts still another generation at risk for developing problems associated with DES treatment of their grandmothers. Unfortunately, the DES episode continues to have serious health consequences and serves as a reminder of the toxicities that can be caused by hormonally active chemicals.
To study the mechanisms involved in DES-induced teratogenesis and carcinogenesis, experimental animal models were developed to study the adverse effects of developmental exposure to environmental estrogens on reproductive tract development and differentiation; these models continue to be used to study effects of other endocrine-disrupting chemicals. The high doses that were used in many of the early animal studies were similar to doses administered to pregnant women and thus have clinical relevance; lower doses are informative when using DES as a model estrogenic endocrine disruptor. The prenatal DES-exposed mouse model has been particularly successful in replicating and predicting abnormalities reported in DES-exposed humans (22, 23, 24, 25, 26, 27). A comparison of DES abnormalities (Table 1⇓) in humans and mice demonstrates the usefulness of DES-exposed experimental animal models.
Although vaginal adenocarcinoma was the original lesion of clinical interest because of its rare occurrence in young DES-exposed women, we have focused much of our research on uterine adenocarcinoma because it provides the potential to study a lesion that increases with age and occurs at a high incidence. Realizing that reproductive tract differentiation continues into neonatal life for both humans and mice, outbred CD-1 mice were treated neonatally with DES (2 μg/pup·d) on d 1–5; a high incidence of uterine cancer (90–95%) was seen in mice at 18–24 months of age (28). These tumors rarely metastasized, but in aged animals (24 months of age or older), the lesions were observed to spread to para-aortic lymph nodes or directly extended to contiguous organs (28). Investigations with other species neonatally treated with DES including rats (29) and hamsters (30) also reported a high incidence of uterine tumors. Because other experimental rodent models have duplicated uterine tumors seen in the CD-1 mouse, this lesion may be predictive of the carcinogenic potential of environmental estrogens in women as they age. It is significant that these mouse tumors progressed through the same morphological and biological continuum of hyperplasia to atypical hyperplasia to neoplasia similar to that seen in women and that the tumors were histologically similar to the human pathologies (28).
To establish a dose-response curve for uterine adenocarcinoma useful for endocrine disruptor studies, neonatal mice were treated with varying doses of DES ranging from 0.0002–2 μg/pup·d. Few studies have covered these low doses (31, 32, 34). The results showed that DES caused tumors even at a dose of 0.0002 μg/pup·d (Table 2⇓). Unlike other estrogen-responsive end points published by our lab and others that showed an inverted U dose-response relationship (33), uterine cancer followed a linear dose-dependent response with increasing incidence of tumors after increasing dose.
Interestingly, tumor incidence could be predicted based on estrogenic potency seen during neonatal life. Figure 1A⇓ shows neonatal estrogenicity in response to varying doses of DES and its relationship to subsequent carcinogenicity (Fig. 1B⇓) observed in aged mice. Both graphs show a dose-dependent linear increase, unlike some other end points that have shown nonmonotonic responses. Furthermore, DES induced estrogen-regulated genes like uterine lactoferrin (LF) on neonatal d 5, and the induction followed a similar linear response as shown by Western blotting techniques (Fig. 2⇓); thus, LF may be useful to predict future carcinogenic effects. Whether LF is directly associated with carcinogenicity or is simply a marker remains to be determined, but it is a subject of study in our laboratory.
To determine whether DES was unique or whether other environmental estrogens could cause uterine lesions, neonatal mice were similarly treated on d 1–5 with 17β-estradiol, tamoxifen, hexestrol, tetrafluorodiethylstilbestrol, ethinyl estradiol, 2-hydroxyestradiol, 4-hydroxyestradiol, genistein, nonylphenol, bisphenol A, or methoxychlor. Most of the studies were conducted using 2 μg/pup·d based on previous studies using DES; however, the weaker environmental estrogens (genistein, nonylphenol, bisphenol A, and methoxychlor) were tested at a higher dose (200 μg/pup·d). All compounds except methoxychlor caused uterine lesions in aged mice; results are summarized in Table 3⇓. Methoxychlor (pure, not technical grade) was the only compound tested that did not cause uterine lesions. Because reports in the literature suggest that a metabolite of methoxychlor is estrogenic, not methoxychlor itself (35), and because the neonatal liver is not fully functional during the time of treatment (36), the most likely explanation for lack of tumors is that the compound was not metabolized to an active estrogen form. All other compounds were estrogenic, and all were associated with uterine tumors after developmental exposure.
Uterine carcinomas have not been observed in untreated control CD-1 mice at corresponding ages, at various stages of the estrous cycle, or after similar adult short-term exposure to estrogens. This suggests that the developmental stage of uterine differentiation and the time of estrogen exposure as well as the estrogenic potency are all important factors in the development of uterine lesions.
Numerous studies have demonstrated that developmental exposure to DES interferes with normal differentiation of the Müllerian duct and regression of the Wolffian duct. Although the mechanisms are not completely understood, a molecular component in the malformation of the tissues and perhaps in the cellular changes may be responsible. Developmental studies have reported that HOX genes are involved in the structural differentiation of the reproductive tract (37) and that prenatal DES delays the expression of these genes (38). Thus, this molecular misprogramming is apparently responsible for the structural alterations observed in the DES reproductive tract (39). Additional investigations with Wnt genes also suggest DES is working through multiple gene pathways to cause structural changes (40, 41, 42).
We have also described permanent abnormal gene imprinting, which may be involved in tumor induction and other cellular alterations in the reproductive tract; neonatal exposure to DES caused demethylation of the estrogen-responsive gene LF in the mouse uterus (43). Studies to determine altered methylation patterns in other estrogen-responsive genes continue.
Similar to tumors described in mice, the neonatal estrogen-exposed hamster developed uterine carcinoma at a high frequency after developmental exposure to DES (30). Molecular studies with the hamster concluded that imbalances in the estrogen-regulated uterine expression of c-jun, c-fos, c-myc, bax, bcl-2, and bcl-x protooncogenes probably played a role in the molecular mechanisms by which neonatal DES treatment ultimately induced epithelial neoplasia in the rodent uterus (44). Microarray studies with the murine uterus in our laboratory revealed similar altered gene expression pathways that included an estrogen-regulated component (Grissom, S., W. N. Jefferson, E. Padilla-Banks, E. Lobenhofer, and R. R. Newbold, submitted for publication).
The role of estrogen receptor (ER) in the induction of abnormalities and tumors after developmental DES exposure has been studied using transgenic mice that overexpressed ERα (MT-mER). MT-mER mice were treated with DES during neonatal life and followed as they aged. It was hypothesized that because of abnormal overexpression of ERα, reproductive tract tissues of the MT-mER mice might be more susceptible to tumors after neonatal DES treatment. This was indeed the case because mice overexpressing ERα were at a higher risk of developing abnormalities including uterine carcinoma in response to neonatal DES compared with DES-treated wild-type mice. At 8 months, 73% of the DES-treated MT-mER mice compared with 46% of the DES-treated wild-type mice had uterine tumors. Furthermore, these lesions occurred at an earlier age compared with wild-type DES mice (45). These transgenic mouse studies suggested that ERα levels present in a tissue may be a determining factor in the development of estrogen-related tumors. Additional transgenic mouse models that expressed variant forms of ERα, and DES-treated ER knockout mice that lacked uterine tumors, also suggested that ERα played a role in the development of reproductive tract lesions (40). ERβ was not identified in the murine uterus (46, 47); thus, its role in uterine abnormalities is unclear and requires additional study.
The role of metabolism in DES-induced lesions has long been an area of investigation. Using the DES mouse model, catechol estrogens, in particular 4-hydroxyestradiol, were very effective at uterine tumor induction (36). Although both 2- and 4-hydroxyestradiol were carcinogenic, the latter induced a 9-fold higher tumor incidence compared with the parent hormone estradiol (36). In addition to hormone-related cell proliferation that may be associated with DNA damage, 4-hydroxyestradiol can be further oxidized to a quinone reactive intermediate. Metabolic redox cycling between this quinone and the hydroquinone (4-hydroxyestradiol) may then produce mutagenic free radicals. Thus, estrogenic compounds may induce tumors in target tissues by inducing DNA damage and genetic lesions and by stimulating proliferation of cells damaged by such processes (36). Together, these data suggested that estrogens may be operating through multiple mechanisms to induce tumors.
Because mechanistic studies provided support that estrogens caused both genetic and epigenetic alterations in developing target tissues, the possibility was raised that abnormalities seen after prenatal or neonatal DES treatment could be transmitted to subsequent generations. In fact, studies from our laboratory showed that prenatal or neonatal treatment with DES led to tumors in the female and male genital tract, and in addition, the susceptibility for tumors was transmitted to the descendants through the maternal germ cell lineage (18, 19); transmission via the DES-exposed male was not studied. Mice were treated with DES prenatally (2.5, 5, or 10 μg/kg·d) on d 9–16 of gestation, or neonatally (0.002 μg/pup·d) on d 1–5, which were the highest doses that did not drastically interfere with fertility later in life. When female mice (F1) reached sexual maturity, they were bred to control untreated males. Female and male offspring (DES lineage or F2) from these matings were aged to 17–24 months and examined for genital tract abnormalities. An increased incidence of proliferative lesions of the rete testis (an estrogen target tissue in the male) and tumors of the reproductive tract was observed in DES-lineage males (18). Furthermore, in DES-lineage females, an increased incidence of uterine adenocarcinoma was seen (19). The incidence was lower in DES descendants than in their parents; uterine tumor incidence in DES F1 at 18 months was 31% at the neonatal dose of 0.002 (Table 2⇑), whereas it was 11% in their DES descendants (19). These data suggest that alterations occurred in germ cells and were passed to subsequent generations. Interestingly, multigenerational effects of DES have been reported by other laboratories, and some of these report transmission through the paternal lineage (21, 48).
The mechanisms involved in these transgenerational events are unknown, but altered methylation patterns can be transmitted to subsequent generations. We have shown altered methylation patterns in several uterine genes that are permanently dysregulated after developmental DES treatment (43, 49). The estrogen-responsive proteins LF and c-fos were permanently up-regulated in the uterus after developmental exposure to DES, and the promoter region of these genes was shown to be hypomethylated (43, 49). Although the consequences of these types of alterations are unclear, studies suggested that methylation patterns can be passed to subsequent generations (50). A recent report supports this theory because prenatal exposure to vinclozolin or methoxychlor caused adverse effects on testis morphology and male fertility, and these effects were transmitted to subsequent generations (51). In addition, this report showed that these two chemicals caused epigenetic alterations in the DNA, specifically hyper- and hypomethylation, and that these alterations were also observed in subsequent generations (51) (also see Anway and Skinner article in this series, Ref. 52). Because the response of estrogen-regulated genes is set during development, altered hormone response may be transmitted to subsequent generations.
Transgenerational effects may also be associated with alterations in specific estrogen-responsive genes. For example, LF induction in prepubertal females that were exposed neonatally to DES showed that this gene continued to be overexpressed even after treatment was completed (Fig. 3A⇓). Furthermore, this gene was also overexpressed in uterine tissues from DES-lineage females, although these mice never received DES (Fig. 3B⇓). Other estrogen-responsive genes are being similarly studied in DES-lineage mice, and thus far, these data suggest involvement of an altered gene expression pathway that includes an estrogen-regulated component. Future study into transgenerational effects of other environmental chemicals and the mechanisms that govern these effects is a newly emerging research focus that deserves serious attention.
Summary and Recommendations:
Sufficient evidence has been accumulated through the years in experimental animals and humans to show that the developing fetus and neonate are uniquely sensitive to exogenous estrogen exposure. If exposure occurs during critical periods of differentiation, permanent adverse effects are well documented to result. Some of these effects, such as reproductive tract abnormalities and uterine tumors, may not be observed until much later in life, long after exposure occurs. Most importantly, evidence with experimental animals suggests that adverse effects may be transmitted to subsequent generations; however, more studies are needed to determine whether this transmission of tumor potential occurs in humans. An important cohort to follow that would answer many of the unresolved questions for humans is the grandchildren of DES-exposed women. Furthermore, additional studies in both experimental animals and humans are needed to identify and understand the mechanisms involved in the transmission of disease and to detect early markers of subsequent disease.
Although animal studies must be considered carefully before extrapolation to humans follows, the DES-exposed mouse model has provided some interesting comparisons to similarly exposed humans. The model has duplicated and predicted many of the lesions observed in DES-exposed women. Although DES is a potent estrogen, it continues to provide markers of the adverse effects of exposure to estrogenic and other endocrine-disrupting substances during development, whether these exposures come from naturally occurring chemicals, from synthetic or environmental contaminants, or from pharmaceutical agents. Ongoing mechanistic studies will help identify other potential reproductive toxicants and will help better access the risks of exposure to other endocrine-disrupting chemicals in the environment if chemical exposures occur during critical stages of development.
Sources: Adverse Effects of the Model Environmental Estrogen Diethylstilbestrol Are Transmitted to Subsequent Generations
by Retha R. Newbold, Elizabeth Padilla-Banks and Wendy N. Jefferson
NCBI, June 2006 – full PDF
More DES DiEthylStilbestrol Resources
- DES studies on cancers and screening.
- DES studies on epigenetics and transgenerational effects.
- DES studies on fertility and pregnancy.
- DES studies on gender identity and psychological health.
- DES studies on in-utero exposure to DES and side-effects.
- DES studies on the genital tract.
- Papers on DES lawsuits.
- DES videos and posts tagged DES, the DES-exposed, DES victims.