Unraveling the Estrogen Paradox: When is Estrogen Good for the Brain?


by Brenda Patoine

May, 2008

An Interview with John H. Morrison, Ph.D.
Professor and Chair of Neuroscience
Professor, Geriatrics and Adult Development
Director, Neurobiology of Aging Laboratory
Mount Sinai School of Medicine

Q: You reported research in 2007 showing that estrogen protects primates’ brain cells from age-related deterioration. What brain changes did you observe, and what are the implications for therapeutic use of estrogen?

A: We have shown in aged female rhesus monkeys that a long-term cyclical regimen of estrogen treatment (i.e., estradiol every 21 days for 2–3 years) enhanced cognitive skills dependent on the prefrontal cortex. This is the cortical area that mediates the highest levels of cognitive function in primates, including humans. We have also shown that neurons in the prefrontal cortex of these same aged estrogen-treated monkeys display an increase in [dendritic] spines, the neuronal sites for synaptic plasticity, which are required for learning complex tasks. The new spines are primarily small and highly plastic, which has led to the hypothesis that they may be particularly important to neural processes underlying learning.

Interestingly, estrogen has the same effects on this group of neurons and their spines in young female monkeys, but young monkeys with and without estrogen performed equally well on cognitive tasks. In fact, our data showed that these prefrontal neurons in young monkeys have enough resilience to retain a reasonably youthful profile even in the absence of estrogen, whereas the aged monkeys show a particularly dramatic loss of spines in the absence of estrogen.

This notion of loss of resilience with aging is a critically important concept in the study of the aging brain. The fact that the prefrontal cortex is so highly responsive to estrogen is particularly relevant to humans, since this area mediates the highest levels of cognitive function. Unfortunately, it is also highly vulnerable in aging and Alzheimer’s disease.

Q: How do you explain estrogen’s capacity to sustain and/or recover youthful synaptic structure and cognitive function in monkeys and the findings of the Women’s Health Initiative-Memory Study (WHIMS), where there was no cognitive benefit to estrogen replacement in women?

A: I don’t think that this finding reflects a difference between women and female monkeys with respect to potential effects of estrogen. Rather, I think it is more reflective of key differences in the formulation, timing, and schedule of hormone treatment, all of which need to be very carefully considered with respect to estrogen’s effects on the brain, menopause, and cognitive aging.

In terms of formulation, we used pure estradiol in our experiments, the exact same hormone that monkeys and women are exposed to naturally. The WHIMS study used Premarin for unopposed estrogen treatment, and PremPro for mixed treatment employing estrogens and progestins. Premarin is a complex mix of over 20 estrogenic compounds with varying levels of activity in the brain; PremPro adds progestins to Premarin.

In addition, we recapitulated the cyclical schedule of natural estrogen exposure in our monkeys (i.e., a 21- to 28-day cycle of repetitive high levels followed by low levels), which is similar to the menstrual cycle in women as well. The WHIMS study employed the very common practice of achieving stable levels of hormones through the daily ingestion of pills. Finally, our monkeys were the equivalent of perimenopausal women in relative age and estrogen levels, and their estrogen replacement began within 6–12 months of ovariectomy. Most of the women in the WHIMS study were in their 70s, well beyond menopause, and had had no hormone treatments for at least 10 years.

For these reasons, rather than attributing these differences in our findings to species differences, we strongly believe that treatments for women have to better simulate normal physiology with respect to the estrogen formulation and schedule of delivery, and that treatments should start in close proximity to the loss of estrogen, not after a long exposure to very low estrogen levels.

Q: You’ve suggested there is a critical win­dow of opportunity during which estrogen therapy may be helpful to women. What is your current thinking on when that window exists and why?

A: This is a very important concept, and it is supported by the preclinical studies as well as many of the human studies. Our monkeys began their hormone treatments soon after ovariectomy, whereas most of the women in the WHIMS study had a very long delay between menopause and initiation of hormone treatments, and were in their 70s when treated. When the women in the WHIMS studies who were in their 50s are considered separately, a protective effect of estrogen does emerge. Other human studies have also shown that protection against cognitive decline by estrogen is most evident in women who begin treatment in close proximity to menopause, and animal studies that experimentally induce a long delay before initiation of estrogen show decreased effectiveness for therapy as well.

Why does this happen? We are not sure, but it may involve the molecules that respond to estrogen, the estrogen receptors. The distribution and abundance of these receptors is likely altered with long-term absence of estrogen, but this has not yet been adequately studied. Aged neurons may well respond to estrogen differently than young neurons, and may even respond negatively to estrogen if they are already displaying age-related pathology. Most of us are confident that a “window of opportunity” exists, but we need a great deal more research to understand the neurobiological, cellular, and molecular underpinnings of such a time-limited opportunity to benefit from estrogen.

Q: How has WHIMS impacted basic research on estrogen’s effects on the brain? Do you perceive there to be more cross-talk between basic and clinical researchers post-WHIMS?

A: In my work on aging and Alzheimer’s disease, I have worked across the clinical and basic science communities for 25 years, and I have rarely seen such a positive outcome from what appeared to be a large gulf between the preclinical and clinical data. Initially, it was difficult for the basic science community, with some people wondering why we were still interested in such issues now that it was settled that hormone treatment is bad for women.

Fortunately, many of the leaders in the basic and clinical research communities realized that there were now more questions than answers. So we got together to try to reconcile the discordant findings. The National Institute of Aging (NIA) played a critically important role in bringing these groups together. I found myself at meetings where I represented the neuron, and on my left was a gynecologist and on my right a cardiologist. It was fantastic!

The key investigators in the WHIMS studies were wonderful to work with at these meetings, and the basic scientists—myself included—were eager to refine our animal models and re-define the key issues in a more clinically relevant manner. We went back to our labs with a far more translational mission.

We have a long way to go to figure out the key endocrine/neural interactions that underlie this important aspect of aging, but I have been so impressed with the clinical and basic science investigators in this field and the manner in which they are working together to move the field forward.

Q: Is “normal” age-related memory loss merely an earlier or less severe manifestation of the same neuropathology occurring in Alzheimer’s disease? What’s different?

A: The short answer is no, age-related memory loss is not just a less severe form of the same neuropathology as Alzheimer’s disease (AD). The key difference is that age-related memory loss appears to result primarily from the loss of synapses with very minimal death of neurons, whereas in AD, extensive yet selective neuron death occurs, causing a massive disruption of the very circuits that mediate cognition. In “normal” aging, these circuits are not functioning optimally; in AD they no longer exist.

And it isn’t just synapse loss that leads to age-related memory decline. It is also loss of the synaptic plasticity that is required for learning and memory. This is an important distinction: bringing back destroyed circuits as complex as those that die in AD is far beyond our current abilities and may simply not be feasible, but we may be able to restore or protect synaptic plasticity. This is precisely what I think estrogen does.

One of the key questions remaining is: do the synaptic alterations of “normal” aging leave a neuron vulnerable to death, or is the neuron death seen in AD initiated by a completely independent mechanism? This is a critically important question for all of us, because if we can prevent the transition to AD-related neuron death by maintaining synaptic health, then we have good reason for optimism in our efforts to decrease the incidence of AD.