Molecular profile of reactive astrocytes--implications for their role in neurologic disease

Astrocyte-Schwann cell interactions in culture

After injury, either as a result of trauma or degenerating/demyelinating diseases, axons of the central nervous system (CNS) normally fail to regenerate. Transplantation of glial cells, particularly Schwann cells, into areas of injury or demyelination has been considered a promising approach to promote recovery. However, the extent of Schwann cell interaction with CNS axons is greatly influenced by the presence of astrocytes which redefine the CNS-PNS (peripheral nervous system) boundary in a lesioned CNS, thereby preventing invasion of Schwann cells. The molecular basis for this restrictive effect of astrocytes on Schwann cells is not known. In the present study, we have cocultured astrocytes and Schwann cells to develop an in vitro model to characterize this interaction. Astrocytes in contact with Schwann cells appeared hypertrophied and showed increased staining for glial fibrillary acidic protein (GFAP). In cocultures maintained for 2-3 weeks, segregation of the two cell types was observed, Schwann cells appeared in groups, and each group was surrounded and separated from one another by astrocytic processes. Since the behavior of these two cell types observed in culture is very similar to their interaction seen in vivo, this coculture model may be useful in further studying the relationship between astrocytes and Schwann cells.

Glial fibrillary acidic protein: regulation by hormones, cytokines, and growth factors

Levels of glial fibrillary acidic protein (GFAP), an astrocyte-specific intermediate filament protein, are altered during development and aging, GFAP also responds dynamically to neurodegenerative lesions. Changes in GFAP expression can occur at both transcriptional and translational levels. Modulators of GFAP expression include steroids, cytokines, and growth factors. GFAP expression also shows brain region-specific responses to sex steroids and of astrocyte-neuronal interactions. The 5'-upstream sequences of rat, mouse, and human are compared for the presence of response elements that are candidates for transcriptional regulation of GFAP. We propose that the regulation of the GFAP gene has evolved a system of controls that allow integrated responses to neuroendocrine and inflammatory modulators.

Differential expression of the astrocytic enzymes 3-hydroxyanthranilic acid oxygenase, kynurenine aminotransferase and glutamine synthetase in seizure-prone and non-epileptic mice

Previous investigations in seizure-prone mice have suggested that an abnormally elevated production of the astrocyte-derived neuroexcitant, quinolinic acid (QUIN), plays a role in seizure susceptibility. In order to evaluate further the role of QUIN metabolism in genetic murine seizure models, the activities of its biosynthetic enzyme 3-hydroxyanthranilic acid oxygenase (3HAO), and of two other astrocytic enzymes, kynurenine aminotransferase (KAT) and glutamine synthetase (GS), were measured in the brains of seizure-prone EL and DBA/2 mice and two non-epileptic strains (BALB/c and Swiss-Webster). 3HAO activity was found to be markedly higher in both EL and DBA/2 mice than in the non-epileptic strains in all brain regions examined. The activity of 3HAO was not modified by the tossing procedure employed to promote seizures in EL mice. While some strain differences were noted in the activities of KAT and GS, these enzymes did not distinguish seizure-prone from the non-epileptic mice. In order to delineate better the relationship between glial activation and 3HAO, KAT and GS, further studies were performed in the ibotenate-lesioned hippocampus. In mice (but not in rats), the activity of 3HAO was selectively increased in gliotic tissue. These data demonstrate substantial species and strain differences in astroglial enzymes and in their response to brain injury. The observation of widespread abnormally high 3HAO activity in two distinct seizure-prone mouse strains strengthens the hypothesis that enhanced production of QUIN contributes to seizure susceptibility in mice.

Target-deprived CNS neurons express the NGF gene while reactive glia around their axonal terminals contain low and high affinity NGF receptors

Reactive gliosis is part of the response of central nervous system to injury and neurodegeneration. Cellular components of the reactive gliosis have the capability to synthesize neurotrophic factors, and thus are capable of affecting the fate of neuronal populations in the injured tissue. In this study, we explored the putative involvement of reactive glia-derived neurotrophins in sustaining the axonal projections of target-deprived neurons. Neuronal targets of the dorsal column nuclei neurons were suppressed through excitotoxic lesion of the ventrobasal complex of the rat thalamus (VB). Despite the development of reactive gliosis, neither up-regulation of NGF, nor BDNF or NT3 mRNA could be detected by solution hybridization in the lesioned site at all times tested. In contrast, expression of the LNGFR gene increased progressively up to 90 days post-lesion. Immunocytochemical studies localized the LNGFR protein in a subset of small cells with ramified processes resembling microglia at 7 and 20 days post-lesion. At longer times, double immunolabelling studies revealed that a substantial part of LNGFR-immunoreactive cells filling the area of neuronal loss were neither microglial cells nor astrocytes although presence of LNGFR in a subset of microglial cells could not be excluded. Previous ultrastructural studies of the kainate-lesioned VB suggest that these LNGFR-immunoreactive cells correspond to oligodendrocytes and/or Schwann cells. At 2 months post-lesion, when LNGFR expression was maximal, increased levels of trkA mRNA were detected in the lesioned site. Immunocytochemical studies revealed the presence of numerous trkA-immunoreactive astrocytes. TrkB mRNA, encoding the full-length high-affinity receptor for BDNF, remained undetectable by non-isotopic in situ hybridization. In contrast to the lack of neurotrophin gene expression by glial components of the lesioned VB, dorsal column nuclei neurons contained NGF mRNA as revealed by in situ hybridization studies at 10 days--prior to enhanced LNGFR expression in the lesion--and 2 months post-lesion. In addition, the number and the staining intensity of NGF mRNA-positive neurons was increased in the target-deprived neurons, as compared with the contra-lateral nucleus projecting to intact targets. These results show that glial cells present in a reactive gliosis which develops in the kainic acid-lesioned thalamus, do not synthesize neurotrophins but instead produce high levels of both low- and high-affinity NGF receptors, LNGFR by Schwann cells/oligodendrocytes and possibly a subset of microglial cells, and trkA by reactive astrocytes.(ABSTRACT TRUNCATED AT 400 WORDS)

GFAP and astrogliosis

One of the most remarkable characteristics of astrocytes is their vigorous response to diverse neurologic insults, a feature that is well conserved across a variety of different species. The astroglial response occurs rapidly and can be detected within one hour of a focal mechanical trauma (Mucke et al., 1991). Prominent reactive astrogliosis is seen; in AIDS dementia; a variety of other viral infections; prion associated spongiform encephalopathies; inflammatory demyelinating diseases; acute traumatic brain injury; neurodegenerative diseases such as Alzheimer's disease. The prominence of astroglial reactions in various diseases, the rapidity of the astroglial response and the evolutionary conservation of reactive astrogliosis indicate that reactive astrocytes fulfill important functions of the central nervous system (CNS). Yet, the exact role reactive astrocytes play in the injured CNS has so far remained elusive. This chapter summaries the various experimental models and diseases that exhibit astrogliosis and increase in glial fibrillary acidic protein (GFAP). Recent in vitro studies to inhibit GFAP synthesis are also presented.

A transgenic mouse model to assess the interaction of cytotoxic T lymphocytes with virally infected, class I MHC-expressing astrocytes

Astrocytes provide crucial support for neurons and their impairment by viruses or their interactions with anti-viral or autoimmune responses could contribute to neurological disease. We have developed a transgenic mouse model to assess lymphocyte-astrocyte interactions. The major histocompatibility complex (MHC) class I molecule, Db, was expressed in astrocytes under the transcriptional control of regulatory sequences from the glial fibrillary acidic protein (GFAP) gene. Baseline cerebral MHC class I mRNA levels from transgenic mice were elevated over those of non-transgenic controls, and a prominent increase in cerebral MHC class I expression occurred following focal, injury-induced astroglial activation within transgenic brains but not in non-transgenic controls. FACS analysis of explant astrocyte cultures from established transgenic lines demonstrated astroglial expression of the GFAP-Db fusion gene at the protein level. Functional antigen-presenting capacity was conferred by the Db transgene, as virus-infected primary astrocytes obtained from transgenic BALB/c mice (KdIdDdLd) expressing the Db molecule were lysed by Db-restricted anti-viral CTL.

Age-related deposition of glia-associated fibrillar material in brains of C57BL/6 mice

With advancing age, clusters of unusual granules appear in the brains of C57BL/6 (B6) mice. At the light, confocal laser and electron microscopic levels, the granules represent aggregations of fibrillar material often associated with astrocytes. The fibrillar material is largely free of normal organelles and has been located within astrocytic somata and processes, although in many cases the material is found in the neuropil and is surrounded by a discontinuous membrane. The deposits occur predominantly in hippocampus, but also in piriform cortex, cerebellum and less frequently in some other brain regions. They become evident about six months of age and increase markedly in both number and size thereafter. Incidence of the deposits varies greatly among inbred mouse strains. At six to 12 months of age, granules are abundant in male and female B6, and are absent in BALB/c, CBA, DBA/2 and A mice. In hybrid strains with a B6 background the deposits are also present and thus appear to manifest dominant genetic heritability. Similar granular structures have been described in adult brains of the senescence accelerated mouse and have been noted, albeit very rarely, in aged mice from other strains. While immunostaining of the granules with several polyclonal antisera was found by preabsorption with antigens to be non-specific, immunolabeling with monoclonal antibodies to heparan sulfate proteoglycan core protein and to laminin suggest these or related molecules as components of the fibrillar material. The presence of glycosaminoglycans is supported by staining with periodic acid-Schiff and Gomori's methenamine silver methods. The functional significance of the murine deposits is not yet clear. The deposits do not represent senile plaques with beta-amyloid deposition, but they might mimic the deposition of extracellular matrix molecules that is hypothesized to be a precursor condition for plaque formation and cerebral amyloidosis. Furthermore, the genetic differences in the incidence of the fibrillar deposits has potential to model aspects of familial neurodegenerative diseases.

Pathophysiology of glial growth factor receptors

Activation and proliferation of glial cells are common events in the pathology of the nervous system. Although we are only beginning to understand the molecular signals leading to glial activation in vivo, there is increasing evidence that growth factors and their receptors may play an important part. In this paper we summarize the data on the pathophysiology of glial growth factor receptors and their ligands in the central and peripheral nervous systems.

Expression of small heat-shock protein hsp 27 in reactive gliosis in Alzheimer disease and other types of dementia

Immunohistochemical and immunoblotting analysis of brain tissue of Alzheimer's disease (AD) patients showed highly induced expression of the small heat-shock protein hsp 27 in affected cortex. Expression of hsp 27 was present in a large number of proliferating astrocytes. The highest expression was exhibited by degenerative astrocytes in the areas rich in senile plaques. Neurofibrillary tangles, Hirano bodies and some hippocampal neurons were also positive. Expression of hsp 27 increased with the severity of AD-specific morphological changes, and with the duration of dementia. In control brains immunoreaction was restricted to the vessels and to occasional astrocytes in the white matter. Similar patterns of immunoreactivity were present in cases without dementia (Parkinson disease, lacunar state, or focal ischemic necrosis). Patients suffering from other types of dementia (Parkinson/dementia complex, multi-infarct dementia, normal pressure hydrocephalus) showed less expression of hsp 27 in reactive astrocytes than AD, but more than controls. These results indicate that increased expression of hsp 27, especially in astrocytes showing klazmatodendrosis, is associated with AD pathology.

Kainic acid induced alterations in antibody recognition of connexin43 and loss of astrocytic gap junctions in rat brain

Intracerebral administration of kainic acid (KA) in rats was previously shown to abolish immunohistochemical labelling for the astrocytic gap junction protein connexin43 (Cx43) at sites depleted of neurons (Vukelic et al: Neurosci Lett 130:120-124, 1991). This response of Cx43 has now been further investigated with a number of different sequence-specific anti-Cx43 antibodies. At lesion sites in the thalamus, striatum, and hippocampus examined immunohistochemically with an antibody against amino acids (aa's) 346-363 in the Cx43 sequence, the antibody used in the earlier study, Cx43-immunoreactivity was increased 5 h after KA injections, absent by 24 h and for up to 2 weeks post-injection, and began to return to less than normal levels by 2 to 3 weeks post-injection. Analyses of KA lesion sites with antibodies against other sequences of Cx43 (amino acids 283-298, 253-270, 241-260, 113-123, and 49-61) revealed not only the presence but in some cases an increased density of Cx43 immunoreactivity after a survival time of 1 week. Immunolabelling patterns at these sites consisted of relatively large, coarse profiles rather the fine punctate labelling typically seen in sections of normal brain. In homogenates of KA-injected striatum analyzed by Western blots, Cx43 was detected at near normal or slightly increased levels at various survival times examined. The 43 kDa phosphorylated form of Cx43 and its faster migrating 41 kDa dephosphorylated form which is generated post-mortem by a brain phosphatase were both present after standard methods of tissue preparation for Western blot analysis, while only the 43 kDa form was present in normal and KA-injected striatum after inactivation of brain metabolism by focused cranial microwave irradiation. Ultrastructural investigations of lesions sites within the thalamus revealed a virtual absence of astrocytic gap junctions. These results demonstrate that Cx43 levels initially increase after intracerebral KA treatment, that its molecular organization in resident astrocytes is altered such that epitopes that are normally accessible to antibody are hidden while those that may be hidden or relatively inaccessible are exposed, and that this molecular alteration in Cx43 is associated with loss of astrocytic gap junctions.

Activation of cerebral cytokine gene expression and its correlation with onset of reactive astrocyte and acute-phase response gene expression in scrapie

The pathogenesis of scrapie, a transmissible subacute spongiform encephalopathy, is unclear. However, certain aspects of the known cellular and molecular neuropathology in scrapie led us to hypothesize that cytokines could mediate cerebral pathological changes in this neurodegenerative disease. Therefore, expression of multiple cytokine genes in the brain and peripheral organs of scrapie-infected mice was examined. Late in the course of scrapie, expression of tumor necrosis factor alpha (TNF-alpha), interleukin-1 alpha (IL-1 alpha), and IL-1 beta mRNA was markedly increased in the brain but not the spleen, kidneys, or liver. In time course studies, scrapie-infected mice exhibited increased cerebral expression of the TNF-alpha, IL-1 alpha, and IL-1 beta mRNAs by week 15 postinoculation--a time point that coincided with the onset of clinical symptoms. Thereafter, the levels of these cytokine transcripts increased progressively to the terminal stages of of the disease (week 25). To determine the relationship of the increased cerebral expression of the cytokine mRNAs to the development of pathological changes in scrapie, we examined the expression of the glial fibrillary acidic protein gene (a marker for astrocytosis) and the murine acute-phase response gene homologous to the alpha 1-antichymotrypsin gene (designated EB22/5.3). Markedly increased expression of both the glial fibrillary acidic protein and EB22/5.3 mRNAs was observed in the brain but not the peripheral organs of scrapie-infected mice. The increased expression of both these gene products also occurred at week 15 of infection and, thereafter, increased progressively to the terminal stages of the disease. Therefore, infection of mice with scrapie resulted in significant increases in the expression of the TNF-alpha, IL-1 alpha, and IL-1 beta gene products, whose pattern correlated with the onset and development of molecular and clinical pathological changes. Since scrapie is known not to evoke an immune response, the present findings strongly suggest the existence of a localized cerebral host response to the agent during which proinflammatory cytokines could be key pathogenic mediators.

Brain injury in a dish: a model for reactive gliosis

Reactive gliosis is a powerful response to brain injury and subsequent neuronal damage in vivo. Neuronal cell cultures are now well established as assays to study this process in vitro. However, equivalent studies of purified glial cell populations have only recently been achieved, following the realization that glial cells produce many of the neuropeptides, transmitters and growth factors that are produced also by neurons. There is now scope for studies in vitro that use mixed, identified populations of glial and neuronal cells to dissect the interactions between the two. Such cultures also lend themselves to assays for potential therapeutic strategies for brain injury that take account of all the different cell types found in the brain.

Central nervous system damage produced by expression of the HIV-1 coat protein gp120 in transgenic mice

Many people infected with human immunodeficiency virus type 1 (HIV-1) develop neurological complications that can culminate in dementia and paralysis. The discrepancy between the severity of impairment and the paucity of detectable HIV-1 within neurons has led to an intense search for diffusible virus- and host-derived factors that might be neurotoxic (see ref. 2 for review). The HIV-1 envelope glycoprotein gp120 is an extracellular protein that is shed from infected cells and so has the potential to diffuse and interact with distant uninfected brain cells. Studies on cultured immature cells suggest that gp120 induces neurotoxicity (reviewed in refs 2, 4), and systemic injection of gp120 in neonatal rats and intracerebroventricular injection in adult rats results in deleterious effects on the brain. To assess the pathogenic potential of gp120 in the intact brain, we have now produced gp120 in the brains of transgenic mice and found a spectrum of neuronal and glial changes resembling abnormalities in brains of HIV-1-infected humans. The severity of damage correlated positively with the brain level of gp120 expression. These results provide in vivo evidence that gp120 plays a key part in HIV-1-associated nervous system impairment. This model should facilitate the evaluation and development of therapeutic strategies aimed at HIV-brain interactions.

Growth factors in CNS repair and regeneration

Traumatic central nervous system (CNS) injury is a significant clinical problem in the developed world. After injuries that penetrate into either the mature brain or spinal cord, damaged neurons initially begin to regrow, but this regeneration is aborted as a fibrotic scar is laid down within the wound. Reconnection of several neuronal pathways does not occur. Functional recovery from such injuries is therefore poor and morbidity severe, particularly for those patients with spinal cord damage. Although palliative measures are available to improve the quality of life, there is no accepted treatment to restore impaired sensory or motor function, so patients remain significantly and permanently debilitated. However, the rapid recent advances that have been made in our understanding of the underlying cellular and trophic pathology of such injuries offer the potential for development of novel therapies to control scarring, enhance neuron survival and stimulate axon regeneration, thereby promoting functional recovery.

An aggregate brain cell culture model for studying neuronal degeneration and regeneration

Rotation-mediated aggregating cell cultures from fetal rat telencephalons containing glial and neuronal cells mature in a fashion comparable to that known to occur in brain in vivo. Large aggregates of 300-500 microM diameters can now reproducibly be cultivated and maintained for more than 40 days in a well defined serum free medium. Validity of the use of such cultures for in vitro studies of various physiological, pharmacological and toxicological phenomenon has already been demonstrated. In this communication some observations suggesting the usefulness of such cultures for pharmacological studies clarifying the possible effects of drugs and other agents on excitatory amino acid induced pathological processes will be presented. The advantages and limitations of the use of aggregated brain cell culture based models for the development of agents potentially useful for the treatment of aging and dementia will also be discussed.

Deafferentation induces early and delayed differential changes in the pattern of expression of the various guanine nucleotide binding protein mRNAs in rat striatum

A polymerase chain reaction-derived method was used to identify and quantitate the relative abundance of the different mRNAs encoding various isoforms of the guanine nucleotide regulatory protein Gs, Gi, and Go alpha subunits in the striata of rats unilaterally injected with 6-hydroxydopamine in the substantia nigra. Thirty days after the lesion, the mRNA levels of the G(o) and of the Gi 1 alpha subunits were increased by about 2-3 times, those of the Gi 3 decreased by about 60% and those of Gi 2 and Gs unmodified. The pattern of expression of the G(o) alpha subunits mRNA changed in a time-dependent fashion, being significant 20 days after the lesion. The decrease in Gi 3 alpha subunit mRNA levels was maximum 10 days after the lesion and tended to be reduced in magnitude with time while the changes in Gi 1 alpha subunit mRNA showed a byphasic behaviour being reduced at 10 days and increased at 30 days after the lesion. These data suggest that the expression of the various G protein alpha subunits in the striatum are continuously regulated by factors originating from afferent neurons and surrounding cells.

Neurologic disease induced in transgenic mice by cerebral overexpression of interleukin 6

Cytokines are thought to be important mediators in physiologic and pathophysiologic processes affecting the central nervous system (CNS). To explore this hypothesis, transgenic mice were generated in which the cytokine interleukin 6 (IL-6), under the regulatory control of the glial fibrillary acidic protein gene promoter, was overexpressed in the CNS. A number of transgenic founder mice and their offspring exhibited a neurologic syndrome the severity of which correlated with the levels of cerebral IL-6 expression. Transgenic mice with high levels of IL-6 expression developed severe neurologic disease characterized by runting, tremor, ataxia, and seizure. Neuropathologic manifestations included neuro-degeneration, astrocytosis, angiogenesis, and induction of acute-phase-protein production. These findings indicate that cytokines such as IL-6 can have a direct pathogenic role in inflammatory, infectious, and neurodegenerative CNS diseases.

Molecular mimicry accompanying HIV-1 infection: human monoclonal antibodies that bind to gp41 and to astrocytes

Monoclonal antibodies that bound to HIV gp41 and cross-reacted with astrocytes were recovered from the blood of three patients infected with HIV-1. Mapping of the specificity of these monoclonal antibodies, using synthetic gp41 peptides, located their epitope to amino acids 644-663 and established their conformation dependence. Six other human monoclonal anti-HIV antibodies were found to bind to HIV gp41 or gp120 but not to reactive astrocytes in brain tissue. Sharing of linear or conformational protein determinants between disparate viral and host proteins is termed molecular mimicry. The consequences of such mimicry by anti-viral antibodies interacting with astrocytes may play a role in the dementia of AIDS patients since a major function of astrocytes is to maintain the appropriate milieu for neuronal function. The finding of such cross-reactive antibodies adds to the evidence for a possible autoimmune pathogenesis in some of the disease manifestations accompanying HIV infection.

Expression of cathepsin G-like and alpha 1-antichymotrypsin-like proteins in reactive astrocytes

The central nervous system (CNS) of many different species responds to diverse neurologic injuries with an activation of astrocytes. Yet, the exact function of this reactive astrocytosis is unknown. In this report, mouse astrocytes were activated in vivo by focal penetrating brain injury. Reactive astrocytes were stained with antibodies raised against the serine protease cathepsin G (cat.G), the serine protease inhibitor alpha 1-antichymotrypsin (ACT), or the astrocytic marker glial fibrillary acidic protein (GFAP). Reactive astrocytes expressing both cat.G-like and ACT-like antigens were found around cerebral wound margins between 18 h and 13 days after neural lesions. The injury-induced immunostaining was unaltered by 900 rads of total body irradiation, suggesting that the astroglial reaction was relatively independent of bone marrow-derived cells. The in vivo immunostaining was complemented with biochemical assays on cultured primary astrocytes. A synthetic peptide was used as a substrate in combination with specific inhibitors to identify a proteolytic activity within astroglial lysates and culture supernatants that closely resembles cat.G. This activity increased substantially upon stimulation of astrocytes with dibutyryl cyclic AMP and was neutralized by antibodies raised against cat.G. In a separate report, it was shown that astrocytes also contain an ACT-like inhibitory activity. The production of ACT- and cat.G-like antigens and activities by activated astrocytes should allow these cells to participate in a number of important biologic processes. Many of these processes may benefit the CNS by assisting in early wound repair. However, astroglial proteases and their inhibitors could also contribute to the pathogenesis of certain neurologic diseases.

Astrocytes are the primary source of tissue factor in the murine central nervous system. A role for astrocytes in cerebral hemostasis

Hemostasis in the brain is of paramount importance because bleeding into the neural parenchyma can result in paralysis, coma, and death. Consistent with this sensitivity to hemorrhage, the brain contains large amounts of tissue factor (TF), the major cellular initiator of the coagulation protease cascades. However, to date, the cellular source for TF in the central nervous system has not been identified. In this study, analysis of murine brain sections by in situ hybridization demonstrated high levels of TF mRNA in cells that expressed glial fibrillary acidic protein, a specific marker for astrocytes. Furthermore, primary mouse astrocyte cultures and astrocyte cell lines from mouse, rat, and human constitutively expressed TF mRNA and functional protein. These data indicated that astrocytes are the primary source of TF in the central nervous system. We propose that astrocytes forming the glia limitans around the neural vasculature and deep to the meninges are intimately involved in controlling hemorrhage in the brain. Finally, we observed an increase in TF mRNA expression in the brains of scrapie-infected mice. This modulation of TF expression in the absence of hemorrhage suggested that TF may function in processes other than hemostasis by altering protease generation in normal and diseased brain.