Scientists contacted for this story were enthusiastic about the findings. “It’s a very exciting set of papers,” said Roberta Brinton at the University of Southern California, Los Angeles. She was not involved in either study. “They tell us a lot about the basic biology.”

Previous work has made the case that estrogen or other ER agonists can shield the brain against neurodegeneration and inflammation, provided the agonists are given at the right time and activate the right pathway (see ARF related news story; ARF related news story on Liu et al., 2008; Tiwari-Woodruff and Voskuhl, 2009). In particular, ER agonists can ameliorate the symptoms of experimental autoimmune encephalomyelitis (EAE), a mouse model of MS that has a particularly strong inflammatory component (see Gold and Voskuhl, 2009; Glass and Saijo, 2010; Crawford et al., 2010). Since neurons, astrocytes, and microglia all express both ERα and ERβ, many questions remain about what cell types mediate the effects of estrogen.

Glass and colleagues focused on ERβ, since ligands selective for this receptor isoform help dampen inflammation in the brains of EAE mice. They did not study ERα. For most experiments, first author Kaoru Saijo used primary cultures of both mouse and human microglia and astrocytes, which express high levels of ERβ. Saijo tested two synthetic ERβ-selective ligands—Indazole-Br and Indazole-Cl—and found that they repress a suite of inflammatory responses in both microglia and astrocytes stimulated by lipopolysaccharide (LPS). The indazoles blocked inflammatory responses due to LPS in vivo as well. Most tellingly, when Saijo and colleagues gave indazoles to EAE model mice before inducing the disease, the animals stayed healthy. When the authors provided indazoles late in the disease course, the animals’ symptoms improved, showing that treatment with ERβ ligands can also help resolve established inflammation.

Saijo and colleagues wondered if there are endogenous ERβ-selective ligands that act like the indazoles. A steroid screen turned up an endogenous neurosteroid—5-androstene-3β,17β-diol (ADIOL). The authors found that ADIOL had nearly identical effects to the indazoles in all of the assays listed above. Microglia are the main brain cells that make ADIOL, converting it from its precursor dehydroepiandrosterone (see Jellinck et al., 2007). Intriguingly, the gene that encodes the enzyme responsible for this conversion lies on human chromosome 19q13, an MS-susceptibility locus (see Bonetti et al., 2009), suggesting a connection among ADIOL, inflammation, and MS.

The authors went on to detail the mechanism by which ADIOL represses inflammatory genes. They found that, when the brain is inflamed, cytosolic ERβ in microglia translocates to the nucleus and becomes tethered to the transcription factor c-Fos. When ADIOL is also bound to the receptor, the complex recruits a repressor called C-terminal binding protein (CtBP) and represses pro-inflammatory genes. Intriguingly, if 17β-estradiol, the main form of estrogen, binds ERβ instead of ADIOL, CtBP does not join the complex, and inflammation remains high. The authors also note that 17β-estradiol binds to ERβ with greater affinity than ADIOL does, suggesting it can out-compete ADIOL.

In an accompanying editorial in Cell, Serge Rivest and David Gosselin at Laval University, Québec, Canada, write, “The net outcome of ERβ may rest upon an equilibrium that depends upon the bioavailability of its endogenous ligands.” Women have more circulating 17β-estradiol than men. This, they speculate, “suggests a fascinating mechanism to explain why relapsing-remitting MS affects females more than males.”

In the PNAS paper, Voskuhl and colleagues focused instead on ERα’s effects on EAE mice. These authors did not look at ERβ. Studies with ERα knockouts have shown that this receptor is necessary for some neuroprotective effects in the brain (see, e.g., Dubal et al., 2001), but it was not known in what cell type the receptor acted. To answer this question, first author Rory Spence bred two conditional knockout mice. In the first, he knocked out ERα only in neurons, in the second, only in astrocytes. He did not address microglia in this study. Spence and colleagues then administered an ERα ligand as a preventive strategy starting seven days before inducing EAE. The neuronal ERα knockouts had a normal response to ERα ligand, developing less severe disease than mice that did not get the ligand. Astrocyte ERα knockouts, on the other hand, received no benefit from ERα ligand treatment. The results showed that the neuroprotective effects of ERα agonists are mediated through astrocytes, not neurons. Astrocytes could be protecting the brain by making neurotrophic factors, by removing glutamate, or by modulating the immune response, Voskuhl told ARF. She plans to investigate each possibility.

Spence and colleagues also examined the spinal cord, which sustains damage in EAE as well as in human MS. In control mice and neuronal ERα knockouts, ERα ligands reduced macrophage and T cell entry into the cord, also axon loss and reactive astrogliosis. Astrocyte ERα knockouts got none of these benefits from ERα ligand treatment, again showing that ERα on astrocytes mediates the protective effects of these ligands. Voskuhl said that in future work, they will look in the hippocampus to see if astrocytic ERα dampens inflammation in this brain area as well. If so, this could have implications not only for cognitive problems in MS, but also for other neurodegenerative diseases such as Alzheimer’s disease. Voskuhl said her group will also look at behavior in these mice, and examine what role ERβ is playing in this system.

Together, the two papers indicate that “Estrogen agonism, both through ERα and ERβ, is beneficial in reducing neuroinflammation,” said Terrence Town at Cedars-Sinai Medical Center, Los Angeles, California. The next step, several commentators noted, would be to examine ERα and ERβ in the same assay and look for interactions between them. Rena Li at the Roskamp Institute, Sarasota, Florida, pointed out that ERα and ERβ can form heterodimers, and therefore may work together.

Many scientists interviewed for this article were excited by the therapeutic implications. Inflammation is a common factor in many neurodegenerative diseases (see Saijo et al., 2010; Glass et al., 2010), and estrogen has demonstrated neuroprotective effects in animal models of several neurological disorders. One possibility, Brinton suggested, is that therapies that target these ER pathways might help prevent inflammation and demyelination in people at risk for AD. Brinton notes that estrogen receptors are also abundant in men’s brains, so such treatment need not be limited to women. As the first step in extending these findings to other diseases, Glass’ group has begun a collaboration with Ron Ellis at UCSD as well as other labs that study HIV, Saijo said. They are looking at whether ADIOL can be used as a biomarker for HIV-dementia.

However, Town sounded a note of caution, warning that “EAE is not MS.” The pathology is different, he said, and more basic research will be needed before these results can be translated to the clinic. He suggested that researchers should do a study of multiple estrogen pathway agonists in several disease models to see which ones work best for which neurodegenerative disease. Other researchers agreed that more preclinical work is needed.

These studies highlight the importance of estrogen pathways and the need for continued study in this area, several scientists said. Interest in estrogen treatment dimmed after the large-scale Women’s Health Initiative showed that estrogen replacement therapy did not lower AD risk and instead increased the risk of breast cancer (see ARF related news story and ARF Live Discussion on estrogen). However, a consensus has emerged in the field that this study failed because the wrong formulation of estrogen was provided too late in life, Voskuhl said. She pointed out that a large body of literature in both humans and animal models shows that, once estrogen has been withdrawn for a period of time, the body can no longer respond to it. Brinton agreed, and noted that there is a great unmet need for safe alternatives to estrogen therapy that will deliver the neuroprotective benefits to women while not increasing their risk of cancer.

Brinton said that, for this reason, many researchers are interested in selectively targeting ERβ, which is better characterized in the brain than ERα. Brinton is intrigued by ADIOL’s promise in this regard. “It’s very interesting in that you can now approach this estrogen receptor-mediated effect through a neurosteroid back door. I think the therapeutic potential is really quite remarkable,” Brinton said. Voskuhl noted that the estrogen estriol, which is mostly produced during pregnancy and weakly targets both ERα and ERβ, is currently in clinical trials for MS. She speculated that the next generation of selective estrogen receptor modifiers will be safer and more effective.—Madolyn Bowman Rogers.

References:
Saijo K, Collier JG, Li AC, Katzenellenbogen JA, Glass CK. An ADIOL-ERbeta-CtBP transrepression pathway negatively regulates microglia-mediated inflammation. Cell. 2011 May 13;145(4):584-95. Abstract

Gosselin D, Rivest S. Estrogen receptor transrepresses brain inflammation. Cell. 2011 May 13;145(4):495-7. Abstract

Spence RD, Hamby ME, Umeda E, Itoh N, Du S, Wisdom AJ, Cao Y, Bondar G, Lam J, Ao Y, Sandoval F, Suriany S, Sofroniew MV, Voskuhl RR. Neuroprotection mediated through estrogen receptor {alpha} in astrocytes. Proc Natl Acad Sci USA. 2011 May 9. Abstract