Estrogen Imprinting : When Your Epigenetic Memories Come Back to Haunt You

Persistent hypomethylation in the promoter of nucleosomal binding protein 1 (Nsbp1) correlates with overexpression of Nsbp1 in mouse uteri neonatally exposed to diethylstilbestrol or genistein

2008 Study Abstract

The concept of developmental programming by steroid hormones has been around for decades, and may be best understood in the brain, where early-life estrogens and androgens feminize and masculinize behavior. Other end organs, such as the liver, are also programmed by steroids during early development, leading to important and necessary sexual dimorphisms in structure and function. In addition to physiological programming or imprinting by steroids, it is well established that the reproductive tract, as well as other structures, can be reprogrammed by abnormal or inadvertent steroid exposures during early life. This phenomenon is best understood by extensive studies on the synthetic estrogen, diethylstilbestrol (DES). Beginning in the early 1950s, DES was prescribed to pregnant women to prevent spontaneous abortions, and its use has been estimated at 8 million women worldwide. However, DES utilization during pregnancy was discontinued in the United States in 1972 and worldwide by 1979 when it was discovered that maternal DES usage was linked to an increased incidence of vaginal clear-cell adenocarcinoma in DES-exposed daughters. Since that time, intrauterine DES exposure has resulted in a multitude of complications in female and male offspring, including T-shaped uteri, uterine fibroids, hypospadias, subfertility and infertility, early menopause, immune system disorders, psychosexual disturbances, and increased breast and testicular cancer risk. Although some adverse effects appear relatively early, such as vaginal cancers in young women, others are emerging as the exposed population ages, suggesting lifelong memory of early estrogen levels by cells and organs. In some instances these delayed responses may be triggered by secondary hormonal events such as onset of puberty or changing hormone levels with aging, which greatly complicates the picture. Furthermore, there is new evidence for transgenerational DES effects transmitted to DES granddaughters, the third generation, which indicates that adverse sequelae may be with us for some time.

Concern regarding fetal and neonatal exposure to estrogenic agents also continues today due to a variety of real-life circumstances that include unintended continuation of estradiol-containing birth control pills during the first months of an undetected pregnancy (estimated at > 1 million annually), exposure (sometimes daily) to environmental estrogens (e.g. bisphenol A, methoxychlor, dichlorodiphenyltrichloroethane or DDT), and consumption of high levels of phytoestrogens during pregnancy or in infant formula (e.g. soy with the isoflavone genistein), all of which have shown a variety of adverse effects in animal models . Although there is compelling evidence that soy or genistein ingestion can be chemoprotective and beneficial, as with so many things in life, this appears to depend on timing. Prepubertal genistein intake reduces breast cancer risk in humans and rodent models, whereas fetal exposure, at least in rat models, can increase risk if not accompanied by lifetime genistein intake. This serves to emphasize the subtle influences of adult life experiences on disease manifestation from early-life exposures. Together, these and a multitude of other studies indicate that early-life estrogen exposure fits the expanding paradigm of the developmental origins of human adult disease.

Over the years, researchers have conducted extensive studies on developmental reprogramming, also referred to as developmental estrogenization or estrogen “imprinting,” using rodent models that largely recapitulate, and in many instances predict, the adverse effects seen in DES-exposed humans. A major research effort during this time has been to identify the molecular underpinning whereby cells and tissues “remember” early-life estrogen exposures long after hormone withdrawal. In the late 1990s, McLachlan and colleagues provided the first clue that the process may have an epigenetic basis when they described a permanent change in uterine lactoferrin gene methylation as a function of fetal DES exposure. They went on to propose that developmental reprogramming by early-life estrogenic exposures may be mediated by altered epigenetic memory. As in lower species such as frogs in which early estrogen treatment changes gene methylation and alters the response of the vitellogenin gene to hormones later in life, they also speculated that the first hormonal experience may epigenetically alter the set point for the later hormone response in mammals. A few studies have since substantiated that epigenetic mechanisms underlie estrogen imprinting by demonstrating that early-life estrogens permanently alter DNA methylation and gene expression of specific genes in the uterus and prostate gland, which are associated with altered susceptibility to adult pathology in those tissues.

In the present issue of Endocrinology, Tang et al.  identify additional layers of complexity to epigenetic reprogramming that have not been previously appreciated. Earlier work in the laboratory of Newbold et al. determined that developmental DES resulted in a high incidence of uterine adenocarcinoma in the murine model, which could be predictive of cancers not yet observed in DES daughters. In partnership with the laboratory of Newbold et al., Tang et al.  used an unbiased approach of methylation sensitive restriction fingerprinting to identify altered DNA methylation patterns in prepubertal, adult, and aged mouse uteri that had been exposed during early life to high-dose or low-dose DES, or to genistein at doses consistent with human dietary intake. They identified 14 genes whose methylation patterns shift as a result of neonatal DES or genistein treatments, several of which were previously unrecognized participants in developmental estrogenization of this tissue. Thus, the present data further substantiate the original hypothesis and findings of McLachlan and colleagues. Tang et al.  went on to thoroughly characterize the altered DNA methylation at specific CG sites in the promoter of one gene in particular, nucleosomal binding protein 1 (Nsbp1), and directly correlated the changes with altered gene expression. This gene is particularly significant because Nsbp1 is a nucleosome-core-particle binding protein that plays a role in chromatin remodeling, which itself underpins a higher order epigenetic process. Nsbp1 is structurally similar to the conserved functional domains of the high-mobility group proteins, HMG 14/17, that interact with nucleosomes, transiently destabilize chromatin, increase access to DNA, and enhance gene transcription at targeted sites. Thus, modifications in Nsbp1 methylation and expression by early estrogens have the potential to affect higher order chromatin structure that, in turn, may alter expression of many genes and possibly drive oncogenesis. Of course, this awaits confirmation by future experiments that directly address this possibility. Nonetheless, these novel findings provide important evidence that epigenetic reprogramming by early-life estrogens may involve multiple epigenetic pathways that combine to initiate disease later in life.

However, the real novelty and significant implications of the study by Tang et al. are not in the aforementioned results, as important as they may be, but in the demonstration that epigenetic reprogramming itself involves a two-step process. Original studies by Newbold et al. showed that the onset of uterine adenocarcinoma in neonatal DES-exposed mice required a secondary hormonal “push” by pubertal steroids because prepubertally ovariectomized mice did not develop cancers. The study by Tang et al.  now shows that this is a direct function of specific DNA methylation changes induced by the secondary pubertal hormones that occur only in mice that were first exposed neonatally to DES or genistein. The authors found that the Nsbp1 promoter is moderately methylated, and the gene is expressed in uteri during prepubertal development in control animals, and becomes hypermethylated and largely silenced by early adulthood. This suggests that the role for Nsbp1 in the uterus is normally restricted to development. When mice were given DES or genistein immediately after birth, the Nsbp1 promoter became hyper or hypomethylated in a treatment-specific manner that led to decreased or increased Nsbp1 expression before puberty. This can be considered the first epigenetic modification. If the mice were prepubertally ovariectomized, these divergent methylation patterns persisted through aging. However, if the mice were allowed to undergo puberty with its secondary estradiol exposure, the Nsbp1 promoter was radically shifted to a hypomethylated state in all neonatal estrogen-exposed uteri. Importantly, this secondary epigenetic modification resulted in overexpression of Nsbp1 throughout life, which correlates with the onset of uterine adenocarcinoma in this model . Thus, in addition to overt epigenetic memories determining our fate, it appears that repressed epigenetic memories are also looming in the background, only to be epigenetically triggered by future events. These new findings stress the importance of secondary events in life that are necessary to uncover the initial cryptic modifications.

A “second hit” model has long been applied to genetic changes and the ontogeny of cancer . It now appears that a parallel two-step concept may apply to epigenetic alterations as well. A somewhat similar scenario was previously shown in Drosophila. Flies with a Krüppel gene mutation (Krif−1) exhibited a specific phenotypical outcome (eye outgrowth) only if they were “epigenetically sensitized” by chromatin modification. The authors concluded that an otherwise cryptic genetic variation can be modified epigenetically to unmask a predisposed phenotype. Ruden et al.  went on to propose further that a similar mechanism may be responsible for the transmission of early-life DES effects on uterine development and cancer. The study by Tang et al.  for the first time identifies molecular evidence to support this proposal. Moreover, the aforementioned hypothesis can now be expanded to state that cryptic epigenetic marks may be further epigenetically modified to unmask a predisposed phenotype. This begs the question whether the complexity of lifetime epigenetic interactions might be the basis for individual variability to early-life exposures or even to differential responsiveness to therapeutic interventions.

There are many things that remain to be done, not the least of which is connecting the dots between the epigenetically altered genes and the disease at hand. This is true not only for the uterus, but for other cancers (breast, testicular, prostate) and disease states that result from estrogen imprinting. Might there be other secondary triggers for repressed epigenetic memories aside from just hormones? Can we formulate an “epigenetic fingerprint” consisting of multiple genes that may be similar or unique for the separate estrogenic compounds, end organs, or second hits? If so, perhaps these could be used in the future for early detection of altered disease susceptibility as a result of known or unknown early-life exposures. Perhaps they could even be used to devise novel interventions. Clearly, we have only just begun.

Sources and more information
  • Estrogen Imprinting: When Your Epigenetic Memories Come Back to Haunt You, Endocrinology, Endocrine Society, Oxford University Press Volume 149, Issue 12, Pages 5919–5921, 1 December 2008.
  • Estrogen Imprinting featured image credit dennis43.
DES DIETHYLSTILBESTROL RESOURCES

Have your say ! Share your views