30 November 2011. Mounting evidence indicates that estrogens contribute to learning and memory, reported researchers at the Society for Neuroscience annual conference, held 12-16 November 2011 in Washington, DC. Estrogen signaling induces the rapid remodeling of spines and synapses, and amps up excitatory transmissions in the brain (see Smejkalova and Woolley, 2010), said Catherine Woolley at Northwestern University, Evanston, Illinois, in an SfN special lecture, although she added that researchers still need to clarify what physiological role these changes effect. In addition, estrogen acts through distinct mechanisms in male and female brains, Woolley noted, and these sex-specific differences deserve more study as they might shed light on gender differences in neurodegenerative and neuropsychiatric conditions. At an SfN mini-symposium, six scientists detailed some advances in this field, which may eventually yield therapeutic applications. As Woolley summed up, “We are at the beginning of a revolution in understanding how steroids act in the brain.”
Studies in rodents indicate estrogen plays a role in learning and memory. Mice that lack estrogen receptor (ER) β have a deficit in hippocampal long-term potentiation (LTP), a form of synaptic plasticity, while activation of ERβ enhances spatial working memory (see Liu et al., 2008). But how does estrogen produce these effects? In the classic steroid receptor model, estrogen binds nuclear ERα and ERβ, which act on DNA to alter transcription of target genes. The speakers noted that this model does not necessarily fit for all the cognitive actions of estrogen, however. Estrogen’s remodeling of spines and synapses is too rapid to be mediated by transcription, but instead seems to involve the activation of local signaling cascades. Supporting this idea, hippocampal estrogen receptors are mainly found in the plasma membrane around synapses, (see, e.g., Milner et al., 2001 and Mitterling et al., 2010). Estradiol, the main human estrogen, is also made locally at synapses by the enzyme aromatase (see e.g., Balthazart and Ball, 2006 and Saldanha et al., 2011), putting it in the right place to modulate neuronal circuitry. At SfN, speakers dug down into the detailed mechanisms of this rapid, synaptic estrogen action.
Adding estradiol to hippocampal slices from either male or female rodents amplifies excitatory synaptic transmission through the action of ERβ, Woolley said. Enikö Kramár at the University of California, Irvine, provided some clues as to how this works. In hippocampal tissue from adult male rats, she previously found that estradiol modulates the synaptic cytoskeleton by triggering actin polymerization through a second messenger pathway that involves the small GTPase RhoA and the actin-severing protein cofilin (see Kramár et al., 2009). Blocking actin assembly prevents the estrogen-induced increase in excitatory transmission, Kramár said. The question remained, however, as to how membrane-bound ERβ managed to activate intracellular RhoA. In new work, Kramár and colleagues report that ERβ can interact with and activate synaptic, membrane-bound integrin receptors of the β1 family, which can signal to RhoA. Neutralizing antibodies against β1 integrin blocked estrogen’s ability to pump up excitatory transmission, strengthening the case for this scenario.
By controlling the actin cytoskeleton, estrogen has the potential to remodel dendritic spines and to regulate neurotransmitter receptor trafficking. Deepak Srivastava at Northwestern described his research on cultured rat cortical neurons, in which estradiol treatment stimulates the growth of new dendritic spines within 30 minutes, and also leads to increased shuttling of AMPA glutamate receptors (see Srivastava et al., 2008). The new spines made contact with presynaptic terminals, suggesting they could form active synapses, but did not last, disappearing within 60 minutes after estradiol treatment. What is the purpose of such transient connections? Srivastava reported that when he followed estradiol treatment with a long-term potentiation protocol, the new spines stayed in place 24 hours later, and AMPA transmission was up as well. He suggested that estrogen acts to “prime” neuronal circuitry, creating increased connectivity that is then locked into place by brain activity and learning. He added that he does not know if this process is female-specific.
In her talk, Elizabeth Waters at Rockefeller University in New York City focused on estrogen’s effect in female rats. She noted that estrogen treatment increases spine number in young females, but not in aged ones. Searching for the reason, Waters found that estrogen treatment depletes synapses of ERα in young animals, but has no effect in old. By contrast, estrogen increases synaptic ERβ in both old and young females (see Waters et al., 2011). The changing ratio of synaptic ERα to ERβ in aged females may alter estrogen-induced plasticity, Waters suggested.
Paul Mermelstein at the University of Minnesota, Minneapolis, looked at sex-specific estrogen effects, using cultured female hippocampal neurons. He reported that within seconds, estradiol treatment causes a fourfold increase in CREB phosphorylation and CREB activity. CREB is a key protein implicated in learning and memory (see, e.g., ARF related news story). Dissecting the pathway, Mermelstein and colleagues found that ERα and ERβ interact with caveolin membrane proteins and metabotropic glutamate receptors (mGluRs) to activate downstream kinases and eventually CREB (see Boulware et al., 2005; Boulware et al., 2007; Meitzen et al., 2011). Depending on whether the estrogen receptor binds to caveolin 1 or 3, it activates distinct signaling pathways, Mermelstein added. He noted that this pathway is not the only mechanism of rapid estrogen action, but it seems to be a pervasive one, and is specific to female neurons. In addition, in different brain regions, estrogen receptors couple with distinct members of the mGluR family, allowing for a cell type-specific modulation of estrogen function, Mermelstein said. He pointed out that, depending on the particular mGluR activated, estrogen can play a role in behaviors such as sexual receptivity, drug-induced dopamine release, and pain perception, suggesting that this pathway may have broad-ranging effects.
Feng Liu at Pfizer Global Research and Development, Groton, Connecticut, also noted that estrogen is implicated in numerous conditions that show sex-specific differences. These include schizophrenia, which has delayed onset in women and symptoms that fluctuate across the menstrual cycle, and Alzheimer’s disease, where data suggests that lack of estrogens at menopause may increase the risk of cognitive decline (see Hughes et al., 2009). Estrogens may also play a role in depression, pain, and anxiety, Liu added. To try to target the central nervous system effects of estrogen without engaging its peripheral effects, Pfizer is focusing on ERβ, which is highly expressed in several brain structures. Liu said that Pfizer has developed a specific ERβ agonist, WAY200070, that replicates many of estrogen’s beneficial effects on memory, including increasing spine number, dendritic complexity, LTP, and learning (see ARF related news story). WAY200070 is an experimental compound that does not have good drug-like properties, Liu added. In addition to its other effects, the agonist quickly increases levels of critical synaptic proteins, which correlate with stronger LTP. Looking for the mechanism, Liu reported that ERβ activation decreases phosphorylation of the translation inhibitor EIF2α, thus allowing synaptic mRNAs to be quickly translated into proteins. Blocking EIF2α dephosphorylation abolished the effect of estrogen on synaptic protein levels, further supporting this mechanism.
Moving to a systems level, Tracey Shors at Rutgers University, Piscataway, New Jersey, provided a dramatic example of how male and female brains may be wired differently. Male and female rats respond in opposite ways to stress, Shors reported, which boosts spine density in the male hippocampus and depletes it in the female (see Shors et al., 2004). Stress also modulates subsequent learning in a sex-specific way. Shors subjected adult rats to a stressful experience, such as a 20 minute swim, allowed them to rest for a day, then trained them to blink in response to a tone. In unstressed rats, females learned this behavior more quickly than males. After a stress, however, females did not learn at all, while the performance of the male rats improved compared to their unstressed baseline. Impaired female learning depended on estrogen, as it did not happen in ovariectomized or prepubescent females.
Each sex uses different brain regions and circuitry for post-stress learning, Shors said. In males, the bed nucleus of the stria terminalis (BNST), a region that carries output from the amygdala and varies in males and females, mediates the enhanced learning after stress (see Bangasser et al., 2005). Lesions to this structure have no effect on female post-stress learning. Instead, females’ lack of learning after stress requires an active medial prefrontal cortex connected to the amygdala (see Maeng et al., 2010). Intriguingly, this wiring pattern is not set in stone, but changes across the lifespan and in response to experience, Shors said. Females who become mothers, or even mother another rat’s pups, no longer experience a drop in learning after stress, and this change seems to be permanent. Some of this work is slated for publication in Behavioral Neuroscience (see also Leuner and Shors, 2006). The experience of motherhood protects female rats from the negative effects of stress on learning, and implies that life experiences can affect brain circuitry, Shors concluded.
The data from this mini-symposium were also published in the November 9 Journal of Neuroscience. Commenting on the overall findings, Roberta Brinton at the University of Southern California, Los Angeles, noted that the data imply that estrogen is not needed for all types of learning, but is activated only in specific cells and pathways. Estrogen regulates synaptic plasticity for the most cognitively demanding tasks, Brinton suggested, such as delayed memory tests that require animals or people to encode information over time. Brinton pointed out that brain structures that develop hypometabolism after menopause, i.e., the prefrontal cortex and posterior cingulate, are also some of the first to exhibit deficits linked to mild cognitive impairment. Yet not all people develop age-related cognitive decline, and an estrogen-based therapy would probably not help everyone, Brinton said. In the future, scientists should focus on finding biomarkers that will identify those aging women who are candidates for hormone therapy, Brinton suggested. Other work presented at SfN found that post-menopausal women who complain of memory problems activate more of their brains during a memory task than those with no complaints, suggesting that this functional MRI signature could be such a biomarker (see ARF related news story).—Madolyn Bowman Rogers.
Srivastava DP, Waters EM, Mermelstein PG, Kramár EA, Shors TJ, Liu F. Rapid estrogen signaling in the brain: implications for the fine-tuning of neuronal circuitry. J Neurosci. 2011 Nov 9;31(45):16056-63. Abstract