SEARCH

SEARCH BY CITATION

Keywords:

  • astrocytes;
  • intermediate filaments;
  • GFAP;
  • vimentin;
  • reactive gliosis;
  • severe mechanical stress;
  • CNS trauma;
  • CNS regeneration

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Astroglial cells are the most abundant cells in the mammalian central nervous system (CNS), yet our knowledge about their function in health and disease has been limited. This review focuses on the recent work addressing the function of intermediate filaments in astroglial cells under severe mechanical or osmotic stress, in hypoxia, and in brain and spinal cord injury. Recent data show that when astrocyte intermediate filaments are genetically ablated in mice, reactive gliosis is attenuated and the course of several CNS pathologies is altered, while the signs of CNS regeneration become more prominent. GFAP is the principal astrocyte intermediate filament protein and dominant mutations in the GFAP gene have been shown to lead to Alexander disease, a fatal neurodegenerative condition in humans. Copyright © 2004 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Intermediate filaments (IFs) are the most puzzling component of the cytoskeleton. We still do not understand why their composition is specific for individual cell types, developmental stages, and sometimes even for a cell carrying out a particular function, such as wound healing. In contrast to other components of the cytoskeleton, the cell-type-specific composition of IFs offers an interesting experimental possibility: by genetic manipulation of IF proteins—including null mutations, overexpression, or expression of mutated genes in transgenic mice—one can gain insights into the function of IFs in a particular type of cell, and also sometimes into the function of the cell itself, both in health and in disease. Studies of astrocyte IFs have been a prime example of this. During the last decade, several mouse models of genetic overexpression or deficiency of these proteins have allowed the expansion of the knowledge about the function of astrocytes in several CNS pathologies. In this review, we focus primarily on the findings and implications of studies of partial or complete IF deficiency in mouse loss-of-function genetic models in a number of disease paradigms. We also describe the newly established connection between mutations in an astrocyte IF protein and a fatal neurodegenerative disease.

Astrocytes, IFs, and reactive gliosis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Astrocytes are the most abundant cells in the CNS, and it appears that the ratio of astrocytes to neurons increases with the increasing complexity of the CNS. Astrocytes have been relatively difficult to study in vivo. Although connected by gap junctions into a huge syncytium (Figure 1), they do not form any defined structures amenable to experimental manipulation.

thumbnail image

Figure 1. Astrocytes in the brain as visualized by antibodies against GFAP

Download figure to PowerPoint

Astrocytes are suspected of being involved in a wide range of CNS pathologies, including trauma, ischaemia, and neurodegeneration. In such situations, the cells change both their morphology and expression of many genes—among them, genes encoding the IF proteins glial fibrillary acidic protein (GFAP), vimentin, and nestin 1. GFAP was first isolated from brain lesions of patients with multiple sclerosis 2. In the CNS it is specifically produced by astroglial cells, and is widely used as a marker of astrocytes and as a diagnostic marker for astrocyte-derived human neoplasms.

In response to essentially any CNS pathology, astrocytes undergo a characteristic change in appearance—the hypertrophy of their cellular processes, a phenomenon referred to as reactive gliosis (Figure 2). A well-known feature of reactive astrocytes is increased production of IFs, and the increased expression of GFAP, but also vimentin and nestin, two IF proteins that are abundantly expressed in immature astrocytes. Thus, the attempts to deplete astrocytes of GFAP and other IF proteins have been an obvious way to learn more about the physiological and pathological functions of these cells 3. The emerging picture is highly interesting and it suggests that IFs in astrocytes are structures of fundamental importance in the pathogenesis of various CNS pathologies.

thumbnail image

Figure 2. Reactive astrocytes around CNS injury visualized by antibodies against GFAP. In many types of CNS injury, the affected region is surrounded by reactive astrocytes whose cellular processes extend towards the injury. The function of this reactive gliosis remains incompletely understood

Download figure to PowerPoint

Reactive astrocytes in mice deficient in GFAP (GFAP−/−) or vimentin (Vim−/−) have a reduced amount of IFs, and astrocytes from mice deficient in both proteins (GFAP−/−Vim−/−) are completely devoid of IFs 4. In reactive astrocytes from wild-type mice, the IFs contain GFAP, vimentin, and nestin. However, in reactive astrocytes from GFAP−/− mice, the IFs contain vimentin and nestin, and those from Vim−/− mice contain GFAP only, since nestin can neither self-assemble nor co-assemble with GFAP 4 (Table 1). Experiments with unchallenged GFAP−/− or Vim−/− mice did not show major CNS phenotypes, but suggested that astrocytes influence neuronal physiology in the hippocampus 5 and in the cerebellum 6, 7. The absence of IF proteins in astrocytes appears to alter communication between Bergmann glia and Purkinje cells, resulting in impaired eye-blink conditioning and long-term depression in the cerebellum of GFAP−/− mice 6, and impaired motor coordination in Vim−/− mice 7.

Table 1. Composition of IFs in non-reactive and reactive astrocytes of wild-type mice and mice deficient in GFAP and/or vimentin
GenotypeComposition of IFsReactive astrocytes : IF amount/bundling
Non-reactive astrocytesReactive astrocytes
Wild-typeGFAP, vimentinGFAP, vimentin, nestinNormal/normal
GFAP−/−No proper IFs (non-filamentous vimentin)Vimentin, nestinDecreased/normal
Vim−/−GFAPGFAP (non-filamentous nestin)Decreased/tighter
GFAP−/−Vim−/−No IFsNo IFs (non-filamentous nestin)

Astrocyte IFs and resistance to mechanical stress

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Although the connection between keratin IFs and resistance of the epidermis to mechanical stress is well established (for a review see refs 8 and 9), the function of astrocyte IFs in maintaining the mechanical integrity of the CNS is unclear. In GFAP−/− mice, non-reactive astrocytes—which account for the overwhelming majority of astrocytes in a healthy brain—are essentially devoid of IFs 5, 10. Nevertheless, in three independent studies, GFAP−/− mice lived normal lives and, if not challenged, had normal CNS morphology 5, 10, 11. Another group, however, found white matter pathologies and dysmyelination in their unchallenged GFAP−/− mice 12. This discrepancy has not been resolved. Interestingly, the same group reported increased susceptibility of GFAP−/− mice to experimental autoimmune encephalomyelitis, a model of multiple sclerosis 13.

The CNS is mechanically protected by the skull, the vertebral column, meninges, and the cerebrospinal fluid. Thus, even if astrocyte IFs are important for stabilizing CNS tissue, their absence might not become manifest unless suitable experimental models are used. And indeed, studies with the head percussion model suggested that astrocyte IFs contribute to the resistance of CNS tissue to a particular type of severe mechanical stress 14. GFAP−/− mice were subjected to head injury from a dropped weight. When placed on a wooden board to prevent head movement at impact, GFAP−/− mice survived as well as wild-type controls. However, when placed on a foam bed that allowed head movement at impact, most of the GFAP−/− mice, but none of the wild-type controls, died after the injury. The GFAP−/− mice showed prominent subpial and white matter bleeding in the region of the cervical spinal cord, possibly resulting from a vein rupture (Figure 3).

thumbnail image

Figure 3. Spinal cord bleeding after percussion injury performed on a foam bed to allow head movement at impact is absent in wild-type mice (a) but visible in GFAP−/− mice (b). Subpial and white matter bleeding can also be seen on transverse sections through the cervical spinal cord in GFAP−/− mice (d), but not in wild-type mice (c). Reprinted from Neuroreport, 1998, vol 9, Nawashiro H, Messing A, Azzam N, Brenner M, Mice lacking GFAP are hypersensitive to traumatic cerebrospinal injury, pp 1691–1696, by permission of Dr Michael Brenner 14

Download figure to PowerPoint

We recently assessed the effect of the absence of IFs in astroglial cells on the mechanical stability of the retina under severe mechanical stress. Being an accessible part of the CNS, the retina is well suited for such experiments, in this case performed in mice seconds after death while the retinal tissue was still alive. Application of severe mechanical stress left the retinas of wild-type controls intact. However, in GFAP−/−Vim−/− mice and, to a lesser extent, in Vim−/− mice, the inner limiting membrane and adjacent tissue separated from the rest of the retina (Figure 4). Electron microscopy showed that this retinal ‘crack’ occurred within the end-feet of Müller cells, radial glia-like cells in the retina that normally contain IFs composed of GFAP and vimentin 15. Thus, at least in some regions of the CNS, astrocyte IFs seem to be important for resistance to severe mechanical stress.

thumbnail image

Figure 4. Severe mechanical stress on the retina leads to the complete and partial separation of the inner limiting membrane and adjacent tissue from the rest of the retina in GFAP−/−Vim−/− mice and partial separation in Vim−/− mice. The retinas of wild-type or GFAP−/− mice remain intact. GCL = ganglion cell layer; INL = inner nuclear layer; ONL = outer nuclear layer. Reprinted from J Cell Sci, 2004, vol 117, Ludkvist A, Reichenbach A, Betsholtz C, Carmeliet P, Wolburg H, Pekny M, Under stress, the absence of intermediate filaments from Müller cells in the retina has structural and functional consequences, pp 3481–3488, by copyright permission of The Company of Biologists Ltd 15

Download figure to PowerPoint

As demonstrated first for keratins in epithelial cells and subsequently for other IFs, these proteins might primarily serve to increase the mechanical resistance of cells and tissues 8, 16. In the absence of any challenge, the lack of IFs in astrocytes does not have any major consequences 5, 10, 11, perhaps because other components of the cytoskeleton, such as actin filaments, can compensate and provide the support necessary for the development and maintenance of cellular processes of non-reactive astrocytes. However, when the structural demands imposed on a cell are increased, as in reactive astrocytes that undergo hypertrophy of cellular processes, IFs become indispensable (see below and eg 17), and compensatory mechanisms are no longer effective. Thus, astrocyte IFs, and perhaps IFs in general, might be viewed as subcellular structures of fundamental importance in stress, particularly when a maximal cellular response is required to avert an imminent threat 18.

Astroglial cells without IFs—implications for pathological vascularization of the retina

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

To assess the pathophysiological implications of the ‘retinal crack’ phenotype found in GFAP−/−Vim−/− mice described above, we exposed GFAP−/−Vim−/− mice and single mutants to retinal hypoxia. This leads to oxygen-induced retinopathy, a widely used model of retinopathy of immaturity that also exhibits some features of diabetic retinopathy 19. On postnatal day 7, mice are placed into an environment with a decreased oxygen concentration, which delays the development of the vascular system. Five days later, the mice are transferred to a normo-oxygenic environment, which leads to massive neovascularization triggered by relative hypoxia. The vessels grow from the retina into the vitreous body (as they do in premature babies or patients with diabetes), and their presence there can easily be quantified 19. Hypoxia-induced vascularization was decreased substantially in GFAP−/−Vim−/− mice and partially in Vim−/− mice (Figure 5). Thus, the absence of IFs in Müller cells of the retina decreases the resistance of their end-feet and consequently of the corresponding layer of the retina to mechanical stress, and it also reduces the extent of ischaemia-triggered pathological vascularization 15.

thumbnail image

Figure 5. The number of neovascular nuclei in the vitreous body, a measure of the extent of hypoxia-induced pathological vascularization, is substantially decreased in GFAP−/−Vim−/− retinas and modestly decreased in Vim−/− retinas. No difference was found between GFAP−/− and wild-type retinas. In the absence of hypoxia, blood vessels do not enter the vitreous body, and the normal vascularization of the retina does not depend on the presence of GFAP and vimentin. Reprinted from J Cell Sci, 2004, vol 117, Ludkvist A, Reichenbach A, Betsholtz C, Carmeliet P, Wolburg H, Pekny M, Under stress, the absence of intermediate filaments from Müller cells in the retina has structural and functional consequences, pp 3481–3488, by copyright permission of The Company of Biologists Ltd 15

Download figure to PowerPoint

Astrocyte IFs and osmotic stress

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

In vitro, astrocytes respond to a hypo-osmotic environment by transient swelling and within minutes show a tendency to return to their original cell volume 20, 21. This phenomenon, known as regulatory volume decrease, involves an efflux of osmotically active molecules from astrocytes, such as the amino acid taurine 20, 22–24. It was proposed that regulatory volume decrease by astrocytes might be the key mechanism in counteracting the development of brain oedema in response to brain ischaemia or trauma, and that cytoskeleton-linked stretch-activated plasma membrane channels serve as cell-volume sensors 25–28.

Ding et al 29 subjected primary astrocyte cultures from wild-type, GFAP−/−, Vim−/−, and GFAP−/− Vim−/− mice to hypo-osmotic stress (corresponding to a 25 mM reduction in NaCl) in perfusion chambers and assessed the efflux of [3H]taurine. Taurine release was up to 50% lower in GFAP−/−Vim−/− than in wild-type astrocytes, but tended to be only slightly decreased in the single mutants. Anderova et al 30 perfused spinal slices with an iso-osmotic solution with an increased concentration of potassium (50 mM) or a hypo-osmotic solution with a reduced sodium concentration and found smaller increases in the potassium concentration around astrocytes in slices from GFAP−/− mice than in those from wild-type controls. Thus, genetic ablation of astrocytic IFs seems to compromise the ability of astrocytes to respond to hypo-osmotic stress.

Do these findings have any relevance for brain pathologies, in particular those connected with prominent osmotic stress, such as brain ischaemia? Nawashiro et al 14 exposed GFAP−/− and wild-type mice to brain ischaemia induced by middle cerebral artery occlusion for 2 days and reported comparable infarct volumes in the two groups. However, when middle cerebral artery occlusion was combined with transient occlusion of the carotid artery, GFAP−/− mice had larger infarcts than controls. This raises the interesting and unresolved question of whether reactive astrocytes protect ischaemically compromised brain tissue around the infarct in stroke patients. Studies of GFAP−/−Vim−/− mice, whose reactive astrocytes are devoid of IFs 4, in various brain ischaemia paradigms should shed more light on this issue.

IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

To assess the role of IF up-regulation in reactive astrocytes in CNS trauma, we and others applied several trauma models to mice deficient in GFAP and/or vimentin. A fine needle injury of the brain cortex and transection of the dorsal funiculus in the upper thoracic spinal cord were two of the models used. The responses of wild-type, GFAP−/−, and of Vim−/− mice were indistinguishable. In GFAP−/−Vim−/− mice, however, the post-traumatic glial scarring was considerably looser and less organized, suggesting that up-regulation of IFs is an important step in astrocyte activation 17 (Figure 6). Similarly, prolonged healing after CNS injury was reported in mice in which dividing astrocytes were ablated by GFAP-driven expression of herpes simplex virus thymidine kinase and administration of ganciclovir 31, 32.

thumbnail image

Figure 6. Wound healing after transection of the dorsal funiculus in the upper thoracic spinal cord is prolonged in GFAP−/−Vim−/− mice, and the resulting glial scarring is reduced. Reprinted from The Journal of Cell Biology, 1999, vol 145, Pekny M, Johansson CB, Eliasson C, Stakeberg J, Wallen A, Perlmann T, Lendahl U, Betsholtz C, Berthold CH, Frisen J, Abnormal reaction to central nervous system injury in mice lacking glial fibrillary acidic protein and vimentin, pp 503–514, by copyright permission of The Rockefeller University Press 17

Download figure to PowerPoint

In another study, hemisection of the spinal cord at T12 was associated with increased axonal sprouting and better functional recovery in GFAP−/−Vim−/− mice than in wild-type controls 33. Two groups have addressed the role of astrocyte IFs in neurite outgrowth in vitro 34–36. One reported that GFAP−/−Vim−/− and GFAP−/− astrocytes were a better substrate for neurite outgrowth in vitro than wild-type astrocytes 34, 36. The other found comparable neurite outgrowth when neurons were cultured on wild-type and GFAP−/− astrocytes 35. The latter finding is also compatible with the normal axonal sprouting and regeneration seen after dorsal hemisection of the spinal cord in GFAP−/− mice 37.

Another study, conducted in mice with entorhinal cortex lesions, showed that reactive astrocytes devoid of IFs (GFAP−/−Vim−/−) exhibited only limited hypertrophy of cell processes. Many processes of GFAP−/−Vim−/− astrocytes were shorter than those of wild-type astrocytes and were not straight, although the volume of brain tissue accessed by a single astrocyte was the same as in wild-type mice 18 (Figure 7). In GFAP−/−Vim−/− mice, loss of neuronal synapses in the projection area of the entorhinal cortex (dentate gyrus of the hippocampus) was prominent 4 days after lesioning, and there was remarkable synaptic regeneration 10 days later. GFAP−/−Vim−/− reactive astrocytes, in contrast to wild-type, did not up-regulate their expression of endothelin B receptors, suggesting that the up-regulation of this novel marker of reactive astrocytes 38–41 is IF-dependent 18.

thumbnail image

Figure 7. Three-dimensional reconstruction of reactive astrocytes after dye filling visualizes even fine cellular processes and shows the typical bushy appearance of astrocytes (a–c). Wild-type and IF-free GFAP−/−Vim−/− reactive astrocytes reach comparable volumes of brain tissue (d). However, GFAP−/−Vim−/− astrocytes have fewer long cellular processes (e, g) and their processes are less straight (f, g). Reprinted from The Journal of Neuroscience, 2004, vol 24, Wilhelmsson U, Li L, Pekna M, Berthold C-H, Blum S, Eliasson C, Renner O, Bushong E, Ellsiman M, Morgan T, Pekny M, Absence of glial fibrillary acidic protein and vimentin prevents hypertrophy of astrocytic processes and improves post-traumatic regeneration, pp 5016–5021, Copyright 2004, with permission from The Society for Neuroscience 18

Download figure to PowerPoint

These findings, along with in vitro data on the morphology of IF-depleted astrocytes 42, showed a novel role for IFs in determining astrocyte morphology and implied that the effect of reactive astrocytes after CNS trauma is two-fold: they play a beneficial role in the acute stage after CNS injury, but later on act as strong inhibitors of CNS regeneration. These experiments showed that studies of IF-null mutants can provide insights into how reactive astrocytes affect the clinical outcome of various CNS pathologies. It is tempting to speculate that by affecting the abundance or the composition of IFs, one might be able to control the state of cellular differentiation and hence many cellular functions, which ultimately allow control of complex processes such as CNS regeneration. Ultimately, such knowledge might open the way for modulation of astrocyte reactivity for the therapeutic benefit of patients.

Astrocyte IFs, cell motility, and integration of CNS transplants

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Since IFs are substantially less dynamic than actin filaments and microtubules, their involvement in cell motility has never been the major research focus, and in vitro studies focusing on the motility of Vim−/− fibroblasts 43 provided somewhat conflicting data. In monolayer wounding experiments, vimentin expression did not affect the mobility of polarized cells at the edge of the wound 44. In contrast, Eckes et al reported that Vim−/− fibroblasts had less resistance to mechanical stress than wild-type fibroblasts and reduced migration both in the scrape wound assay and in Boyden chambers 45.

Lepekhin et al 42 assessed the motility of primary cultured astrocytes from GFAP−/−, Vim−/−, and GFAP−/−Vim−/− mice. The fast-moving subpopulation was depleted partially among GFAP−/− and Vim−/− astrocytes and more severely among GFAP−/−Vim−/− astrocytes (Figure 8). The in vivo relevance and molecular mechanisms of this finding remain to be established. However, since astrocytes migrate over considerable distances to sites of injury 46, the slower migration of IF-deficient astrocytes could partially explain the more discrete development of post-traumatic glial scar seen in GFAP−/−Vim−/− mice 17.

thumbnail image

Figure 8. Compared to wild-type, the migration of GFAP−/−Vim−/− reactive astrocytes in vitro is reduced, with the single mutants showing a dose effect (A). Fast-moving subpopulations of GFAP−/−Vim−/− astrocytes are clearly smaller than in wild-type, with GFAP−/− astrocytes and Vim−/− astrocytes exhibiting a dose effect (B). Reprinted from Journal of Neurochemistry, 2001, vol 79, Lepekhin EA, Eliasson C, Berthold CH, Berezin V, Bock E, Pekny M, Intermediate filaments regulate astrocyte motility, pp 617–625, by copyright permission of Blackwell Publishing 42

Download figure to PowerPoint

Because of their morphology and abundance in the adult CNS, astrocytes inevitably come in direct physical contact with any cell that moves from one place to another. This might be of major importance in situations such as the migration of immature neurons born from endogenous neuronal stem cells or the migration of neuronal precursors from CNS transplants. The Chen and Pekny groups transplanted dissociated retinal cells from 0–3-week-old donor mice that ubiquitously express enhanced green fluorescent protein 47 into the retinas of adult wild-type and GFAP−/−Vim−/− recipients and compared the efficiency of integration 48. In wild-type hosts, few transplanted cells migrated from the transplantation site, and even fewer integrated into the retina. In GFAP−/−Vim−/− hosts, however, the transplanted cells effectively moved through the retina, differentiated into neurons, integrated into the ganglion cell layer, and some of them even extended neurites about 1 mm into the optic nerve (Figure 9). Six months after transplantation, the cells remained alive and well integrated 48.

thumbnail image

Figure 9. Retinal transplants from mice ubiquitously expressing enhanced green fluorescent protein integrated much better in GFAP−/−Vim−/− than in wild-type recipients (a, c). In GFAP−/−Vim−/− recipients, transplanted cells migrated more efficiently from the transplantation site and integrated into the ganglion cell layer (c), exhibiting typical morphology of ganglion cells (b). Some of these neurons extended axons into the optic nerve. Reprinted from Nat Neurosci, 2003, vol 6, Kinouchi R, Takeda M, Yang L, Wilhelmsson U, Ludkvist A, Pekny M, Chen DF, Robust neural integration from retinal transplants in mice deficient in GFAP and vimentin, pp 863–868, by copyright permission of Nature Publishing Group 48

Download figure to PowerPoint

Thus, the absence of IFs in astroglial cells of the retina (astrocytes and Müller cells) increases the permissivity of the retinal environment for integration of neural transplants. The extent to which this reflects increased permissiveness for the migration of transplanted cells remains to be established. However, it is tempting to speculate that IF depletion in astroglial cells alters their differentiation state, rendering them functionally similar to more immature astrocytes, which are also more supportive of CNS regeneration 49. These findings may open up interesting therapeutic possibilities 50, 51. It is fair to assume that therapies that would control the expression of IFs and consequently affect both cellular migration and cellular differentiation might also be applicable outside the CNS.

GFAP mutations and Alexander disease

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

To study the role of GFAP in astrocyte hypertrophy, Messing et al 52 generated mice overexpressing human GFAP. The astrocytes of these transgenic mice formed complex intracytoplasmic aggregates of GFAP and small stress proteins that were identical to Rosenthal fibres. Rosenthal fibres, which accompany chronic reactive astrogliosis, are eosinophilic, elongated structures that, when examined ultrastructurally, appear as electron-dense, amorphous masses surrounded by and merging with dense bundles of IFs.

Rosenthal fibres are also a hallmark of Alexander disease, a rare and fatal leukoencephalopathy that most commonly affects infants and young children, who typically present with feeding problems, paraparesis, macrencephaly, mental and physical retardation, and seizures. Juvenile forms of Alexander disease cause predominantly pseudobulbar and bulbar signs, while adult forms are variable and can resemble multiple sclerosis.

The GFAP-overexpressing mice died at an early age and although the cause of death remains unknown, these results pointed to GFAP as a candidate gene for Alexander disease. Subsequent investigations determined that the majority of cases of infantile Alexander disease, and at least some cases of the later-onset juvenile and adult forms, are due to heterozygous missense mutations in the GFAP gene. The heterozygosity of the mutations suggests that they are dominant. In the majority of cases, the mutations seemed to occur de novo and were not found in either parent 53, 54. However, familial adult cases have been described, raising the interesting issue of reduced penetrance or germline mosaicism 55, 56. Although these results identify mutated GFAP as at least one of the culprits responsible for a fatal neurological disorder in humans, the mechanisms are still unclear. It is unknown how the mutant GFAP protein causes brain damage, and what the role of Rosenthal fibres in this process is 57, 58.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Astrocyte IFs, at least at selected locations within the CNS, seem to be of importance in situations connected with severe mechanical stress. However, the function of IFs extends beyond the concept of ropes traversing the cell cytoplasm and rendering the cells tougher at mechanical impact. A number of recent studies imply the key role of astrocyte IFs in determining the functional state of astrocytes. The absence of this component of the cytoskeleton seems to make astrocytes less efficient in dealing with the acute stage of various CNS injuries. On the other hand, IF-free astrocytes support diverse aspects of CNS regeneration, such as collateral sprouting, synaptic recovery, or long-term integration of CNS transplants. Detailed molecular understanding of these processes should ultimately help to develop better therapeutic strategies for treating the CNS diseases and in limiting their consequences.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References

Milos Pekny's laboratory was funded by the Swedish Research Council, the Swedish Cancer Foundation, and King Gustav V Foundation. Marcela Pekna's laboratory was funded by the Swedish Research Council, King Gustav V Foundation, the Swedish Stroke Foundation, and the Swedish Heart–Lung Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Astrocytes, IFs, and reactive gliosis
  5. Astrocyte IFs and resistance to mechanical stress
  6. Astroglial cells without IFs—implications for pathological vascularization of the retina
  7. Astrocyte IFs and osmotic stress
  8. IF-free astrocytes, brain and spinal cord trauma, and post-traumatic regeneration
  9. Astrocyte IFs, cell motility, and integration of CNS transplants
  10. GFAP mutations and Alexander disease
  11. Conclusions
  12. Acknowledgements
  13. References