Clearing Estrogen's Bad Name

In the six years since the Women's Health Initiative, we're learning estrogens might not be so bad after all.

By Phyllis Wise

The summer of 2002 was a tense time for us. One announcement and all of our knowledge and experience as basic researchers was put to question.

On July 10th, Jacques Rossouw, the director of the Women's Health Initiative (WHI) clinical trial, which tracked the long term effects of hormone therapy (HT) on women, announced the trial was being halted three years ahead of schedule. The results showed an unacceptable proportion of women were harmed by the therapy. The data implied that HT did not protect women against heart disease or protect against memory loss and other neurodegenerative conditions such as dementia, and in fact increased their risk of stroke and breast cancer. At the time, some 14 million US women were taking HT to...

On July 10th, Jacques Rossouw, the director of the Women's Health Initiative (WHI) clinical trial, which tracked the long term effects of hormone therapy (HT) on women, announced the trial was being halted three years ahead of schedule. The results showed an unacceptable proportion of women were harmed by the therapy. The data implied that HT did not protect women against heart disease or protect against memory loss and other neurodegenerative conditions such as dementia, and in fact increased their risk of stroke and breast cancer. At the time, some 14 million US women were taking HT to relieve postmenopausal symptoms or to lower their risk for osteoporosis. In the days and weeks that followed, women around the world would call their doctors in panic, and many would stop taking the medication altogether. But the news went against much of what we had been learning about estrogens in the laboratory.

Related Articles

1 It was clear that one of the reasons our results seemed to differ was because the WHI study had tracked women before stroke had occurred to capture their risk, but didn't follow women after their stroke to see if women taking estrogen had improved recoveries, as our models had suggested.

It is important to point out that men also produce estrogen, and are also protected after stroke when estrogens are present. However, it is not usually possible to treat men with estrogens because the hormone has feminizing effects that are not acceptable. This is why it is so important for us to investigate the mechanisms by which estrogens act. Our hope is that this understanding would lead toward the development of compounds that use the same protective mechanisms, but do not have the other effects of estrogens.

The mechanisms by which estrogens act is an area of incredibly intense study. A cytosolic estrogen receptor discovered in the 1960's was thought to act by transporting estrogens attached to the receptor to the nucleus where the hormone-receptor complex bound to DNA and acted as a transcription factor, altering gene expression (see graphic on below). For more than 30 years, this was the exclusive way that estrogens were thought to act: via a single receptor that acted by binding to the promoter region of DNA to increase or decrease transcription of estrogen responsive genes. In 1996 Jan Ake Gustafsson and Ken Korach discovered that there are actually two different estrogen receptor subtypes: one called estrogen receptor alpha (ERα) and a new form, ERβ.2 More recently, my lab realized that estrogens may act without a receptor or by binding to a receptor which can then activate second messenger systems without ever traveling to the nucleus and binding to DNA.

Our laboratory has followed these studies with great interest because we wondered whether the protective actions of estrogens in the brain depend upon the classical pathway or the more novel mechanisms of actions. We designed an experiment to use either ERα- or ERβ-knock out mice and found that ERα plays a pivotal role in protecting the brain against cell death. When this receptor is knocked out, estradiol does not protect the brain. ERβ proved not to play a role in protection; when we knocked out ERβ, mice continued to be protected just like wild-type mice.

It appears that low physiological levels of estradiol therapy act predominantly through the classical estrogen mechanisms of action. It is interesting to contrast this with findings from James Simpkins' lab that much higher concentrations of some estrogen analogs act through the more recently discovered mechanisms which bypass direct estrogen receptor, ERα or ERβ activation. This is important because this information may allow us to develop drugs that target a specific receptor to allow better outcomes from stroke.

What did these studies mean in terms of the results of the WHI? Two aspects of the clinical study design are worth considering in this context. First, only synthetic hormones were used. Premarin and Prempro are composed of substances that attempt to mimic the endogenous hormones synthesized by the ovary, but they are not the same structures. This might account for differential actions of these substances compared to endogenous hormones. Second, the average age of women when they started in the WHI trial was 63 years old and most of them had not had hormone therapy previous to the trial. Since women typically undergo the transition to the postmenopausal state around the age of 51, it means that most of them had been estrogen deficient for about 12 years.

We decided to investigate what would happen if we delayed the time of hormone treatment in mice to match that of the WHI. Two of my postdoctoral fellows, Shotaro Suzuki and Candice Brown, found that if we did not treat mice immediately after "menopause," but waited for several weeks (the equivalent of several human years), that hormone treatment was totally ineffective. What was even more exciting is that we found that one way estradiol acts is by suppressing inflammation, thereby preventing cell death. But if we waited a few weeks before starting to treat the mice, we were unable to suppress inflammation.3 The ability of estradiol to protect the brain by suppressing inflammation also depends upon the presence of ERα, since delay in the initiation of treatment no longer induced an increase in this receptor the way immediate therapy did. Just as our earlier studies had suggested, the timing of treatment is a critical factor to consider.

Many other investigators have asked whether all kinds of estrogens, synthetic or endogenous, act the same way. The answer is a definite no. Different kinds of estrogens and different concentrations of hormone have different effects. And the hormonal milieu makes a big difference in how estrogens act as they can have completely opposite effects depending upon what other hormones are present and whether they have been there before or after estrogen exposure (click here for graphic). While the estrogen research community had started to scratch the surface of the complex actions of this hormone before the WHI trial, our collective knowledge would not have predicted the outcome then. Today, with the insights into human physiology that the WHI trial brought, we have a much clearer picture of the nuances of this hormone.

One of the most exciting areas in neuroscience today is that of neurogenesis. The dogma had been that all neurons were born during embryonic and early postnatal development. Whatever neurons we were endowed with was what we had to work with. If someone suffered from a neurological disease or suffered a brain injury, the only way we could recover brain function was to encourage the remaining neurons to function more efficiently. About 20 years ago, several investigators began to question this assumption. They found that under some circumstances, new neurons were born even in adulthood. This set off a chase for finding compounds that would stimulate the birth of new neurons.

We had learned that estrogens protected the brain after stroke by limiting damage caused by inflammation. Until now, however, we hadn't considered whether part of the improved post-stroke recovery could be cause by the birth of new neurons. Researchers had noted that estrogen was important in the developing brain of an embryo. In the fetal and early postnatal brain, estrogens are factors that mediate neurogenesis, synapse formation, and glial differentiation, to name a few effects. More recently our studies have shown that estrogens are potent neurogenic factors after stroke injury in the adult brain. We showed that this effect depends on both ERα and ERβ.4 Not only can estradiol limit the effects of inflammatory damage, but apparently it also promotes repair and birth of new neurons. At the present time, we are uncertain about how estradiol works to enhance neurogenesis and are studying the cellular and molecular mechanisms that underlie these important actions. We are most excited about these new findings since they hold promise for the use of estrogens in the long-term repair and recovery of brain function.

For more than 30 years, researchers thought that (1) estradiol's only mode of action was through a single cytosolic receptor that transported the hormone to the nucleus where it acted as a transcription factor. In the last decade, researchers found a second estrogen receptor (ER ), designating the two ERα and ERβ, and started to uncover additional molecular pathways for the action of estrogens. Today we know that ER s may not exclusively act from the cytosol, but also appear to function as (2) membrane-bound receptors. Our research has recently suggested a third pathway (3) which may bypass the ER altogether, act on a variety of second messenger signaling cascades and ultimately influence gene expression.

Clinical trials such as the WHI are impressive for the number of women they study. But because of prohibitive costs of performing such a large clinical trial, only a limited number of questions can be addressed. In the year following the announcement that the WHI trial was halted, 40% of women stopped taking hormone replacement therapy. By following up on these paradoxical results, the field realized that results should have been interpreted with more caution and that conclusions were more limited than they seemed at first.5 This powerful hormone, with its many different analogs, different receptors, and the multiple mechanisms of action, must be used carefully. Its functions depend upon dose, preparation, method of administration, the recipient's age, genetics and previous exposure to the hormone. Although all of these caveats apply to all hormones, estrogens should be expected to be among the most complex ones to understand: they exhibit a diurnal rhythm, a monthly rhythm that is determined by the menstrual cycle, and they act differently depending upon whether other reproductive hormones are present in high or low concentrations.

It was a unique time. Although the trial created controversy in the community, between those who thought estrogen harmed and those who were convinced it would help, both sides were deeply motivated to understand the full story. We learned that estrogens are not the panacea people thought they were, but neither are they always harmful. It made us interact with one another with an openness I hadn't seen before. A basic researcher can live her entire professional life without ever having her assumptions tested against the ultimate system of human physiology. For me, it was a rare opportunity.

1. D.B. Dubal et al., "Differential modulation of estrogen receptors (ERs) in ischemic brain injury: a novel role for ER in estradiol-mediated protection against programmed cell death," Endocrinology, 147:3076-84, 2006.
2.G.G. Kuiper et al., "Cloning of a novel receptor expressed in rat prostate and ovary," Proc Natl Acad Sci 93:5925-30, 1996.
3.S. Suzuki et al., "Timing of estrogen therapy after ovariectomy dictates the efficacy of its neuroprotective and anti-inflammatory actions," Proc Natl Acad Sci, 104:6013-8, 2007.
4. S. Suzuki et al., "Estradiol enhances neurogenesis following ischemic stroke through estrogen receptors a and b," J Comp Neurol, 500:1064-75, 2007.
5. J.L. Turgeon et al., "Complex actions of sex steroids in adipose tissue, the cardiovascular system, and brain: Insights from basic science and clinical studies," Endocr Rev, 27:575-605. 2006.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?