This is Part 1 of a two-part meeting report from the 37th annual meeting of the Society for Neuroscience, held 3-7 November, in San Diego. See Part 2.
14 November 2007. A mini-symposium at this year’s annual meeting of the Society of Neuroscience devoted attention to the role of glia in brain injury and disease. Chaired by Clive Svendsen, University of Wisconsin, Madison, the symposium covered the gamut from glial cell signaling to therapeutic approaches to Parkinson disease (PD), ALS, and spinal cord injury.
The simple idea that glia are passive bystanders, offering help whenever neurons get into trouble, no longer passes muster. Scientists now know that these cells do a whole lot more. They encourage axon growth, prevent glutamate excitotoxicity, provide trophic and metabolic support, and even help maintain the blood-brain barrier. Glial function and dysfunction has also been implicated in neurologic diseases such as schizophrenia, and neurodegenerative disorders such as amyotrophic lateral sclerosis (see ARF related news story), multiple systems atrophy (see ARF related news story), and Alzheimer disease (see ARF Live Discussion). Glia buffs even go as far as assigning Albert Einstein’s intellectual prowess to the factoid that a major difference between his brain and that of everyone else is reportedly an increase in astrocytic processes (see Colombo et al., 2006). In short, glia are pretty hot these days, even making the cover of this month’s Nature Neuroscience, which devotes a special focus to the role of these cells in disease.
Michael Sofroniew, University of California at Los Angeles, started the symposium, entitled “The Role of Glial Cells in Brain Injury and Disease,” by detailing some of his lab’s recent studies on astrocyte signaling. Sofroniew’s lab is studying the role of astrocytes in response to injury, which is relevant to AD given that astrogliosis is a prominent feature of the disease. There is currently considerable debate about whether reactive astrocytes are beneficial or harmful. Reactive astrocytes form glial scars in response to brain injuries, such as contusions. Do these scars merely impede neuronal regeneration or do they have positive roles, as well? Sofroniew suggested that it would be highly unlikely that such a process would have arisen and been conserved through evolution if it was purely detrimental, and said that his experimental data seem to support that view, as well.
His lab uses transgenic models to study glial scarring. In normal scarring, reactive astrocytes line up in a palisade manner, surrounding a central necrotic lesion that is often packed with inflammatory cells. Sofroniew was interested in what might happen if he disrupted this pattern. Because proliferation seems to be an important step in the astrocyte response, his lab abolished astrocyte cell division. The scientists did this by generating transgenic mice harboring a thymidine kinase (TK) gene driven by the astrocyte-specific glial fibrillary acidic protein (GFAP) promoter. In the TK-positive cells, ganciclovir, a thymidine analog and antiviral agent, becomes phosphorylated when the TK gene gets turned on, and the analog then disrupts DNA replication. Astrocytes that are not activated by injury are unaffected.
When the researchers used these mice in a contusion model of brain injury, they found that the glia scar no longer formed properly, nor did the distinct glial boundary around necrotic lesions. The downstream consequences of this change were that the volume of the necrotic lesion increased about threefold and the inflammatory cells—no longer contained—spread much farther from the necrotic site. The astroglial quiescence also had functional repercussions. While normal animals recover locomotive function after a moderate contusion, TK transgenic animals treated with ganciclovir performed poorly in a rotarod test of balance and strength, Sofroniew reported.
What kind of cell signaling controls this glial response, and how might it impact other cells? Sofroniew speculated for the sake of discussion that some of the signals astrocytes send out to limit immune cell responses might also prevent axon regeneration in diseases such as multiple sclerosis. This means that understanding astrocytic signal transduction might offer new ways to stimulate neuronal regeneration, he suggested. Astrocyte signaling, however, is still a bit of a black box. Numerous signal transduction pathways have been suggested to be involved, including those linked to protein kinase C, Notch, Smad, TNFα, and G-protein coupled receptors. For his part, Sofroniew has begun studying the role of the Jak/Stat signaling pathway. His coworkers use the Cre/lox system under the control of the GFAP promoter to make conditional knockouts (cKOs) of the stat3 gene in mouse astrocytes. Data yet to be published show that stat3-negative astrocytes appear normal in healthy animals, but after crush injury their response is compromised. The usual hypertrophy does not set in, and neither does induction of GFAP and vimentin, genes that kick in in wild-type animals in response to injury. The palisade boundary between astrocytes and the necrotic lesion does not form in stat3 cKOs either, and immune cells, as detected by CD45 immunoreactivity, flood out from the lesion site, causing a marked increase in inflammation. These mice also don’t recover function as well as wild-type.
What relevance do astrocyte responses in general, and their proliferation in particular, have for human disease? To approach this, Sofroniew and colleagues turned to experimental autoimmune encephalomyelitis (EAE), a common model of multiple sclerosis. In a poster session, Scott Peterson from Sofroniew’s lab reported that limiting astrocyte proliferation exacerbated the EAE. When he induced EAE in TK transgenic animals, their clinical severity score rose to ~2.5 after 3 weeks, but in animals also treated with ganciclovir, scores rose to ~4.5. As in animals with contusion injury, the morphology of glial scars was compromised, and there was a statistically significant increase in the number of CD45 positive cells in gray and white matter. Axonal pathology also seemed exacerbated, because levels of neurofilament 200, which drop by about half in EAE, were even further depleted in the anterior and lateral funiculi. It is worth noting that Adriano Aguzzi and colleagues at University Hospital, Zurich, Switzerland, using the very same TK/ganciclovir methodology, found that suppression of microglia actually protected against EAE (see Heppner et al., 2005). Together, the findings indicate that different types of glia can have distinct and apparently contrary roles.
These findings suggest that glial scarring is protective in at least two forms of neuronal injury—trauma and autoimmune disease. But astroglia can clearly contribute to disease as well. Work from Don Cleveland’s group at University of California, San Francisco, has shown, for example, that mutant superoxide dismutase, which causes familial forms of ALS, can lead to motor neuron disease in mice when expressed only in glia (see ARF related news story). How, then, do glia propagate disease? This question captivates Jeffrey Rothstein of Johns Hopkins University, Baltimore, Maryland. Rothstein noted that a major function of astrocytes is to scavenge glutamate, which can otherwise accumulate to neurotoxic levels in synapses. Astrocytes are armed with an array of glutamate transporters, also called the excitatory amino acid transporters (EAAT). But not all astrocytes are equally well endowed, noted Rothstein. His lab found that astrocytes in the cortex and hippocampus have 10-fold higher levels of EAAT2 than do spinal cord astrocytes. Increased thalamic EAAT3 levels have also been linked to schizophrenia. And work from Eliezer Masliah’s lab at the University of San Diego, California, suggests that EAAT2 is downregulated in AD brain and in transgenic mice expressing human APP under control of the Thy1 promoter (see Li et al., 1997 and Masliah et al., 2000). ALS patients have a dramatic loss of EAAT2, said Rothstein. The astrocytes are there, but are not expressing the protein.
To investigate the relevance of EAAT2 changes, Rothstein used bacterial artificial chromosome (BAC) transgenic animals to measure the expression of glutamate transporters. In late-stage motor neuron disease in rodent models, his group found a focal loss of Glt-1, the rodent homolog of EAAT2. The loss seems due to decreased Glt-1 mRNA levels, and Rothstein speculated that some microRNA inhibitors may be at work.
And what about other astrocyte pathways? For example, glia provide neurons with lactate. To facilitate this transfer, both astrocytes and neurons have membrane-bound monocarboxylate transporters (MCTs). The cells have different isoforms, however, with neurons expressing MCT2, while astrocytes express MCT1 and MCT4. Rothstein has looked at MCT expression in rodent models of ALS. He reported that in the spinal cord there is almost a complete loss of MCT1 in diseased animals, while MCT2 and MCT4 remain unchanged. To test whether this is a cause of effect of disease, the scientists used antisense RNA to individually ablate MCT expression in spinal cord cultures, and they found that losing MCT1 leads spinal cord neurons to die. The scientists tested postmortem tissue from ALS patients and found that they, too, have a loss of MCT1. The findings suggest that damage to glia might compromise their crucial metabolic support role, leading to neuronal death.
Rothstein is currently exploring ways to exploit these discoveries for therapeutic purposes. A screen for compounds that elevate EAAT2 turned up mostly beta-lactam antibiotics, which were published to increase expression of Glt-1 and increase survival in ALS mouse models (see ARF related news story). However, Rothstein was challenged in question time by a researcher who noted that this result has been difficult to reproduce. Rothstein suggested that the effect is not robust and that he is currently trying to improve on the EC50 by making structural analogs. He also said that other groups using these antibiotics found improvement in different models of glutamate-related diseases, such as depression, stroke, and epilepsy. Clinical trials of a different antibiotic for ALS have been a disappointment. Rothstein also reminded the audience astrocytes can communicate via gap junctions across long distances. He showed a video of a calcium wave propagating through astrocytes for hundreds of microns. It is possible that one sick astrocyte could spell disaster for many of its neighbors.—Tom Fagan.