This paper by Zamanian and colleagues represents a technical tour de force aimed at determining if heterogeneity exists in reactive astrocytes. To that end, they have provoked reactive astrogliosis either by 1) administering the bacterial endotoxin LPS peripherally or 2) by inducing ischemic stroke. The authors then used a flow cytometric sorting method to isolate astroglia and subjected them to gene chip analysis. Interestingly, they noted an approximately 50 percent difference in the gene expression profiles between astroglia isolated after the two treatments. These results provide a new paradigm for understanding astrocytic activation—with some forms being beneficial (e.g., in the case of ischemic stroke), and other types, deleterious (e.g., after LPS challenge). This conceptual framework developed by Ben Barres and colleagues will clearly be of high importance to the field of neuroinflammation, both from basic science and translational biology perspectives.

I find it compelling that the present findings extend hypotheses to reactive astrocytes that my group and Carol...  Read more

This paper by Zamanian and colleagues represents a technical tour de force aimed at determining if heterogeneity exists in reactive astrocytes. To that end, they have provoked reactive astrogliosis either by 1) administering the bacterial endotoxin LPS peripherally or 2) by inducing ischemic stroke. The authors then used a flow cytometric sorting method to isolate astroglia and subjected them to gene chip analysis. Interestingly, they noted an approximately 50 percent difference in the gene expression profiles between astroglia isolated after the two treatments. These results provide a new paradigm for understanding astrocytic activation—with some forms being beneficial (e.g., in the case of ischemic stroke), and other types, deleterious (e.g., after LPS challenge). This conceptual framework developed by Ben Barres and colleagues will clearly be of high importance to the field of neuroinflammation, both from basic science and translational biology perspectives.

I find it compelling that the present findings extend hypotheses to reactive astrocytes that my group and Carol Colton’s offered in the mid-2000s to explain heterogeneity of microglial activation states (Town et al., 2005; Colton et al., 2006). Around that time, reactive microglia were considered to be a single phenotype, and we now know that heterogeneity exists in their activated states. For example, one of the lessons learned from the Aβ "immunotherapy" field is that microglia can be beneficially activated by Aβ-specific antibodies to phagocytose and clear cerebral amyloid (Schenk et al., 1999; Bard et al., 2000). On the other hand, the low-level, persistent microgliosis that is characteristic of Alzheimer’s disease seems to be deleterious, as blocking it restores cognitive function, mitigates cerebral amyloidosis, and inhibits plaque-associated astrogliosis (Tan et al., 1999; Tan et al., 2002).

When taken together, then, a broader understanding of neuroinflammation emerges that encompasses both reactive astroglia and microglia. From the translational science side, therefore, a key question emerges: How can we convert these potentially damaging inflammatory cells into targets for disease therapies?

References:
Bard F, Cannon C, Barbour R, Burke RL, Games D, Grajeda H, Guido T, Hu K, Huang J, Johnson-Wood K, Khan K, Kholodenko D, Lee M, Lieberburg I, Motter R, Nguyen M, Soriano F, Vasquez N, Weiss K, Welch B, Seubert P, Schenk D, Yednock T. Peripherally administered antibodies against amyloid β-peptide enter the central nervous system and reduce pathology in a mouse model of Alzheimer disease. Nat Med. 2000 Aug;6(8):916-9. Abstract

Colton CA, Mott RT, Sharpe H, Xu Q, Van Nostrand WE, Vitek MP. Expression profiles for macrophage alternative activation genes in AD and in mouse models of AD. J Neuroinflammation. 2006 Sep 27;3:27. Abstract

Schenk D, Barbour R, Dunn W, Gordon G, Grajeda H, Guido T, Hu K, Huang J, Johnson-Wood K, Khan K, Kholodenko D, Lee M, Liao Z, Lieberburg I, Motter R, Mutter L, Soriano F, Shopp G, Vasquez N, Vandevert C, Walker S, Wogulis M, Yednock T, Games D, Seubert P. Immunization with amyloid-β attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature. 1999 Jul 8;400(6740):173-7. Abstract

Tan J, Town T, Paris D, Mori T, Suo Z, Crawford F, Mattson MP, Flavell RA, Mullan M. Microglial activation resulting from CD40-CD40L interaction after β-amyloid stimulation. Science. 1999 Dec 17;286(5448):2352-5. Abstract

Tan J, Town T, Crawford F, Mori T, DelleDonne A, Crescentini R, Obregon D, Flavell RA, Mullan MJ. Role of CD40 ligand in amyloidosis in transgenic Alzheimer's mice. Nat Neurosci. 2002 Dec;5(12):1288-93. Abstract

Town T, Nikolic V, Tan J. The microglial "activation" continuum: from innate to adaptive responses. J Neuroinflammation. 2005 Oct 31;2:24. Abstract

View all comments by Terrence Town

I agree with Terrence Town; the recent publication from the Barres lab is truly a technical “tour de force” that puts a strong voice to the importance of comparing results between injury states. The beauty of this study is the careful and thorough analysis of gene expression at three different time points after injury, its focus on a single cell type (astrocytes), the immunohistochemical confirmation of specific gene markers in brain, and the analysis of the findings. The data discovered may seem obvious, but, in fact, the results provide useful information that pounds home a message to us all. To my mind, this message says that: 1) it is possible to find specific markers that do not change with disease, allowing firm identification of the number and type of cells involved; 2) specific markers do change as a function of the type of injury and the time point at which the study is done (as well as with gender and age, which need additional study); and 3) complexity exists, particularly with subtle changes in fold expression of common genes that should not be ignored....  Read more

I agree with Terrence Town; the recent publication from the Barres lab is truly a technical “tour de force” that puts a strong voice to the importance of comparing results between injury states. The beauty of this study is the careful and thorough analysis of gene expression at three different time points after injury, its focus on a single cell type (astrocytes), the immunohistochemical confirmation of specific gene markers in brain, and the analysis of the findings. The data discovered may seem obvious, but, in fact, the results provide useful information that pounds home a message to us all. To my mind, this message says that: 1) it is possible to find specific markers that do not change with disease, allowing firm identification of the number and type of cells involved; 2) specific markers do change as a function of the type of injury and the time point at which the study is done (as well as with gender and age, which need additional study); and 3) complexity exists, particularly with subtle changes in fold expression of common genes that should not be ignored.

The need for clear, disease-independent markers for the field is obvious, but what has not been well appreciated is the need to map out gene profiles at multiple time points. Immune-related events in response to injury or disease have a time course that is likely to be oscillatory in nature. If acute, then the onset is clear cut, while the resolution phase may have more complex time and gene profiles. Thus, a single time point for analysis may provide some information on the presence and level of a specific cytokine such as TNF-α, but does not place that information in the context of the action of that cytokine during the time course of the biological response. Like most, if not all, immune factors, TNF-α demonstrates a functional repertoire that can result in opposing outcomes. The actions of TNF-α during immune-repair phases may be different from its actions during immune-toxicity phases. A broader profile of other genes expressed at the same time point is essential to add “context” and to better understand the disease process and the role of TNF-α as the biological response develops over time. Furthermore, more subtle results. such as a fivefold versus twofold change in gene expression, may signal the involvement of different regulatory events. For immune regulatory molecules such as nitric oxide (NO), this type of change may indicate activation of either survival mechanisms or death mechanisms. Zamanian et al. clearly show that wide profiles over multiple time points studied yield a more complete picture of the disease process. Chronic disease has additional problems, since by definition, chronic diseases such as Alzheimer’s disease do not have resolution (the resolution is typically death). Thus, one must exercise caution when comparing immune activity using acutely supplied agents such as LPS to generate immune activity versus a chronic disease state which clearly has a different course of prolonged immune activity.

The final statements on chronic neurodegeneration are somewhat surprising and disappointing. After clearly showing that gene regulation is complex, and that middle cerebral artery occlusion and LPS initiate overlapping but strikingly different gene profiles, the assumption that diseases such as PD or AD would be necessarily similar to LPS-mediated acute inflammation is puzzling. This type of reductive approach can be beneficial under some circumstances. However, it may also bias or reduce the search for understanding of how chronic disease is different from acute disease. In fact, what this publication effectively teaches us is that it is critical to explore the wide scope and timing of the disease process using these and other technologies. What we may find in the AD field is that the differences between acute and chronic disease are astounding.

View all comments by Carol Colton

Thanks for the kind comments about our paper! Just to respond to Carol Colton's question about LPS: We found that ischemia and LPS induced different reactive glial phenotypes, and we speculated, based on the nature of these phenotypes, that one type of reactive astrocyte was good and the other was bad. We further hypothesized—and it was just a hypothesis, though I think it could turn out to be right—that LPS induced a type of reactive astrocyte that might resemble reactive glial cells in neurodegenerative disease. This hypothesis is based on other observations in the literature, especially that the classical complement cascade is highly activated in many neurodegenerative diseases, and we found that LPS-induced reactive astrocytes had highly upregulated many components of the classical complement cascade. Moreover, LPS has been shown in mice to be a strong sensitizer to neurodegenerative disease, and by itself is sufficient to cause a Parkinson's-like syndrome; furthermore, systemic infections (which LPS partly mimics) induce reactive gliosis and, as recent studies in humans...  Read more

Thanks for the kind comments about our paper! Just to respond to Carol Colton's question about LPS: We found that ischemia and LPS induced different reactive glial phenotypes, and we speculated, based on the nature of these phenotypes, that one type of reactive astrocyte was good and the other was bad. We further hypothesized—and it was just a hypothesis, though I think it could turn out to be right—that LPS induced a type of reactive astrocyte that might resemble reactive glial cells in neurodegenerative disease. This hypothesis is based on other observations in the literature, especially that the classical complement cascade is highly activated in many neurodegenerative diseases, and we found that LPS-induced reactive astrocytes had highly upregulated many components of the classical complement cascade. Moreover, LPS has been shown in mice to be a strong sensitizer to neurodegenerative disease, and by itself is sufficient to cause a Parkinson's-like syndrome; furthermore, systemic infections (which LPS partly mimics) induce reactive gliosis and, as recent studies in humans show, exacerbate Alzheimer's disease. I think that our hypothesis was, therefore, well justified. Again, it is only a hypothesis, but our paper now provides the tools with which to test it, as the markers that we identified as specific to this type of LPS-induced reactive astrocyte can now be assessed to see if they are expressed by reactive astrocytes in Alzheimer's disease. An alternative possibility is that there are many other types of reactive astrocytes, and that a completely different type will turn out to characterize neurodegenerative diseases.

View all comments by Ben Barres

The recent paper by Zamanian et al. from Ben Barres’ group continues a conceptual approach established in a previous influential paper by Cahoy et al., 2008, from the same group where they performed global characterization and comparison of genes expressed by the major types of CNS cells. The present study aimed to characterize and compare the gene expression profiles of isolated mouse reactive astrocytes induced by different types of injury models. Reactive astrogliosis in those models was induced by either peripheral neuroinflammation with bacterial endotoxin-LPS or by acute ischemic stroke. They found that, although reactive astrogliosis was characterized by a rapid induction and subsequent attenuation of transcription of core sets of genes, the global gene expression repertoire exhibits substantial heterogeneity of molecular phenotypes following the two different types of injury. These different molecular phenotypes suggest the induction of different functional states in astrocytes. The identified subsets of genes represent the most comprehensive characterization of altered...  Read more

The recent paper by Zamanian et al. from Ben Barres’ group continues a conceptual approach established in a previous influential paper by Cahoy et al., 2008, from the same group where they performed global characterization and comparison of genes expressed by the major types of CNS cells. The present study aimed to characterize and compare the gene expression profiles of isolated mouse reactive astrocytes induced by different types of injury models. Reactive astrogliosis in those models was induced by either peripheral neuroinflammation with bacterial endotoxin-LPS or by acute ischemic stroke. They found that, although reactive astrogliosis was characterized by a rapid induction and subsequent attenuation of transcription of core sets of genes, the global gene expression repertoire exhibits substantial heterogeneity of molecular phenotypes following the two different types of injury. These different molecular phenotypes suggest the induction of different functional states in astrocytes. The identified subsets of genes represent the most comprehensive characterization of altered state of astrocytes in response to injury to date. These findings will provide an indispensable resource for the neuroscience community in the form of new markers for detection of reactive astrocytes in human neurological diseases; they will help generate new, testable hypotheses of reactive astrocyte function; and they may even help in identifying new therapeutic targets for a variety of neurological disease states.

We found compelling the author’s prevalent hypothesis for understanding astroglial activation that some forms can be beneficial (in ischemic stroke injury) or detrimental (in peripheral inflammation). It will be crucial for future studies to perform similar studies (or targeted investigations based on the current study results) at different time points after each type of injury. Similarly, it will be of interest to investigate the consequences of prolonged inflammatory challenge to make the results even more relevant to conditions such as Alzheimer’s disease and aging.

References:
Cahoy JD, Emery B, Kaushal A, Foo LC, Zamanian JL, Christopherson KS, Xing Y, Lubischer JL, Krieg PA, Krupenko SA, Thompson WJ, Barres BA. A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J Neurosci. 2008 Jan 2;28(1):264-78. Abstract

View all comments by Vahram (Harry) Haroutunian
View all comments by Pavel Katsel

This is indeed a very interesting paper, as it shows how different astrocytes are compared to oligodendrocytes—the latter migrating far and compensating for any cells that are lost, while the former stay put both during development and in adulthood.

In regard to astrocyte-induced disease, these findings may be relevant, as death of astrocytes in a given domain cannot be compensated for. However, the ablation experiments occur very early and may be less relevant to age-related neurodegenerative diseases.

I think the biggest progress will be to understand the region-specific differences and specialization these cells have, and, hence, understand how they are specialized to support the neurons in their domain.

View all comments by Magdalena Goetz

This is one of the most interesting papers ever written on astrocytes. The implications are very important. Basically, it is showing that each domain of the brain has its own molecularly distinct type of astrocyte, and that these astrocytes respect their own unique boundaries. Most likely this is a very important design plan of the brain. It suggests distinct, domain-specific role(s) for astrocytes. Perhaps they are critical for specificity of axon guidance during development so appropriate neural circuit wiring occurs, as suggested by an earlier paper in Cell by David Anderson a few years ago (see Hochstim et al., 2008). Or perhaps they control domain-specific synapse formation, function, or plasticity. In addition, the paper also shows that killing of astrocytes in one domain results in a substantial decrease in excitatory synapse formation. A role for astrocytes in controlling synapse formation has so far mostly been shown in vitro, so it is very exciting to see evidence that astrocytes also have this role in vivo (see also...  Read more

This is one of the most interesting papers ever written on astrocytes. The implications are very important. Basically, it is showing that each domain of the brain has its own molecularly distinct type of astrocyte, and that these astrocytes respect their own unique boundaries. Most likely this is a very important design plan of the brain. It suggests distinct, domain-specific role(s) for astrocytes. Perhaps they are critical for specificity of axon guidance during development so appropriate neural circuit wiring occurs, as suggested by an earlier paper in Cell by David Anderson a few years ago (see Hochstim et al., 2008). Or perhaps they control domain-specific synapse formation, function, or plasticity. In addition, the paper also shows that killing of astrocytes in one domain results in a substantial decrease in excitatory synapse formation. A role for astrocytes in controlling synapse formation has so far mostly been shown in vitro, so it is very exciting to see evidence that astrocytes also have this role in vivo (see also previous work by Gabriel Corfas on the role of glia in synapse formation in the inner ear: also Rio et al., 2002).

This is a beautiful paper with many important implications, and it shows that we don't begin to understand the roles of astrocytes in normal development and function, and why so much glial heterogeneity is needed. There may be important implications for disease. Perhaps specific types of astrocytes malfunction or degenerate in disease situations. And, since astrocyte gene expression so closely resembles stem cells, perhaps domain-specific types of astrocytes can more easily be induced to become the type of neurons in that domain in order to repair/regenerate lost neuron types.

View all comments by Ben Barres