GFAPδ immunostaining improves visualization of normal and pathologic astrocytic heterogeneity

Neuropathological analysis of cellular mechanisms underlying gliosis and brain tumors is slowed by the lack of markers allowing to distinguish glial subpopulations in normal or pathological human brains. We therefore evaluated GFAPδ immunostaining in a wide panel of astrogliosis and gliomas, and compared these with GFAP and vimentin. In normal tissue, gliosis and gliomas, GFAPδ immunostaining was observed in astrocytes with relatively high GFAP levels. GFAPδ immunostaining was most conspicuous in glia limitans astrocytes. In Chaslin's gliosis accompanying chronic epilepsy, GFAPδ immunostaining evidenced the glia limitans reactive astrocytes as the source of the dense fibrillary meshwork typical of Chaslin's gliosis. Interestingly GFAPδ and vimentin immunostainings coincided in normal tissues and gliosis, but not in gliomas. Altogether these results show that combined GFAP, GFAPδ and vimentin labelling reveals fine gliofilament regulation in normal and pathological brain.     

Co-expression of glial fibrillary acidic protein- and vimentin-type intermediate filaments in human astrocytomas

Summary The expression of intermediate filament (IF) proteins was studied in 71 cases of malignant human astrocytoma and in 17 cases of reactive gliosis, using immunocytochemical techniques with polyclonal and monoclonal antibodies to glial fibrillary acidic protein (GFAP) and vimentin. In all cases of astrocytoma, varying in degree of malignancy from grade I to grade IV, co-expression of GFAP and vimentin was found. No change in vimentin- or GFAP-IF expression with increasing anaplasia was seen. In addition astrocytic cells in reactive gliosis showed simultaneous expression of GFAP and vimentin. The intracellular distribution of these IF proteins differed. Vimentin was found to be located in a more juxta-nuclear position, whereas GFAP immunoreactivity showed a more intense staining of the cellular processes. Astrocytes in reactive gliosis behaved more or less like neoplastic cells. However, thin cell processes of reactive astrocytes in the cortex and superficial white matter only contained GFAP immunoreactivity. Simultaneous expression of GFAP and vimentin and their proportion in malignant and reactive glial cells are discussed in the light of earlier reports on the IF content of glial cells during development and maturation, in which vimentin precedes GFAP-expression. The existence of two separate (functional) IF systems in astroglia is suggested.     

Immunohistochemical studies on the proliferation of reactive astrocytes and the expression of cytoskeletal proteins following brain injury in rats

The appearance of reactive astrocytes following brain injury was investigated in 4-week-old rats with special reference to their proliferation and chronological changes in the cytoskeletal proteins. Two days after the injury, glial fibrillary acidic protein (GFAP)-positive cells had increased in number around the lesion and spread to the entire ipsilateral cortex by 3 days after the injury. To investigate the distribution of mitotic cells and its chronological change, immunohistochemical staining with monoclonal antibody to bromodeoxyuridine (BrdU) was performed. BrdU-positive cells began to appear around the lesion and spread to the entire ipsilateral cortex by 3 days and their distribution was the same as that of GFAP-positive cells. To investigate the association of GFAP-positive cells with cell division, double labeling experiments using [3H]thymidine autoradiography and immunohistochemical staining with antiserum to GFAP were performed. Cells doubly labeled with GFAP and [3H]thymidine were localized in the area adjacent to the lesion, in the molecular layer of the cortex and in the white matter. By contrast, none of the cells were doubly labeled in the IInd to VIth layers of the cortex. Furthermore, only astrocytes in the former areas expressed vimentin transiently from 2 to 10 days after the injury. In the rats administered vincristine, cells arrested during mitosis were found in the regions which express vimentin. From these results, it was suggested that astrocytes in the molecular layer of the cortex and the white matter adjacent to the lesion proliferated in response to the injury and expressed vimentin transiently, then acquired GFAP, and that astrocytes in the IInd to VIth layers of the cortex became reactive astrocytes without mitosis.     

Plectin immunopositivity appears in the astrocytes in the white matter but not in the gray matter after stab wounds

Plectin is associated with intermediate filaments, such as glial fibrillary acidic protein (GFAP) and vimentin, which have an important role in the glial reactions after lesions. The present study investigates the plectin immunoreactivity following stab wounds in the cortex and the underlying white matter of adult rats. Without lesion, plectin immunoreactive astrocytes were not found in the cortex and only scarcely in the white matter. Following lesion, a number of astrocytes became immunopositive in the white matter, while no immunoreactivity was found in the overlying cortex. GFAP and vimentin showed intense immunopositivity in both areas. Therefore, the participation of the plectin seems to distinguish the glial reactions developed in the white or gray matters.     

A comparison of intermediate filament markers for presumptive astroglia in the developing rat neocortex: immunostaining against nestin reveals more detail, than GFAP or vimentin

The present study compares the immunopositive elements in the developing rat cortex between the day of birth (P0) and the 18th postnatal day (P18), after immunostaining against nestin, vimentin and glial fibrillary acidic protein (GFAP). Nestin immunostaining revealed more structural details than either vimentin or GFAP, or they together. While vimentin immunostaining preferred radial glia and GFAP preferred astrocytes, nestin immunostaining detected both. Stellate-shaped astrocyte-like cells were already seen at P0 and cells of typical astrocytic morphology were numerous at P3, and were predominating elements from P7, whereas GFAP-immunopositive astrocytes were very scarce even at P7, and became numerous only by P11, when nestin immunopositivity started to disappear. Nestin immunostaining revealed such structures which were not seen in GFAP- or vimentin immunostained sections: cell body-like structures 'hanging' at the end the radial fibers, seeming to divide with their fibers, or having astrocyte-like processes. Nestin immunostaining is therefore highly recommended for studies of the glial architecture in the early post-natal brain development.     

Identification of a human glial fibrillary acidic protein cDNA: a tool for the molecular analysis of reactive gliosis in the mammalian central nervous system

Two clones encoding human glial fibrillary acidic protein (GFAP) were isolated from a human astrocytoma cDNA library. The clones pHGFAP1 and pHGFAP2 were selected by the combined use of differential colony hybridization and hybridization-selection technique with polyclonal anti GFAP antiserum. The longer one, pHGFAP1, encompasses 3.0 kb and includes the 1.8 kb long 3' untranslated region specific to the human mRNA. Sequence data disclosed an extensive homology within the coding region of human and mouse GFAP cDNAs even in the end domains. Blot hybridization analysis of RNAs from human, rat and mouse brain revealed a single GFAP mRNA species of 3.1, 2.8 and 2.7 kb respectively and Southern blot experiments indicated that this mRNA is most probably transcribed from a unique gene. In situ hybridization performed with biotinylated probes on cultured mouse brain cells suggests both the sorting and the transport of GFAP mRNA throughout the cytoplasm and processes of the astrocytes. As a model of reactive gliosis secondary to degenerative disorders, 6-hydroxydopamine (6-OHDA) lesion of the substantia nigra in the rat was performed. GFAP mRNA increased 1.4 fold in the ipsilateral striatum on day 10 after the lesion. It then declined to the control level 4 months later contrasting with the lower and more sustained increase in preproenkephalin (PPE) mRNA. The interspecies cross-reactivity of the HGFAP probes make them useful as a tool for the molecular analysis of reactive gliosis in various experimental models.     

Unusual topographical pattern of proximal astrogliosis around a cortical devascularizing lesion

Class II vessels were disrupted on the cortical surface of adult rats within a circular 5-mm-diameter area. This consistently resulted in the formation of a conical lesion by day 1, with a cystic cavity forming by day 21. Four markers were used to identify the glial response surrounding the lesion. The antibody used against S100beta marked the largest astrocytic pool in the gray matter of the cerebral cortex; only approximately 5% of astrocytes were glial fibrillary acidic protein (GFAP)(+) in control animals. GFAP served as a marker for distal reactive gliosis and vimentin (VIM) for proximal gliosis. Isolectin B4 was used as an additional marker to distinguish VIM(+) microglia from astrocytes inside the lesion area. Three immunohistochemically distinct areas of reactive astrocytes surrounding the lesion were found within 24 hr of injury and lasted through day 6. The first area, in contrast to focal traumatic injuries, consisted of a 196-microm-thick boundary layer of S100beta(+) cells immediately surrounding the lesion that never expressed GFAP or VIM by day 6. This boundary layer turns into a GFAP(+) glial limitans encasing the cystic cavity by day 21. A second unusual extended area around the base of the lesion reaching partly into the corpus callosum consisted of S100beta(+)/GFAP(+)/VIM(+) cells. This region appears to be compatible with the local or proximal gliotic response usually found completely surrounding other focal-type injuries. The proximal response at the base of the lesion developed over the first 3 days in the following sequence: S100beta(+)/GFAP(-)/VIM(-) to S100beta(+)/GFAP(+)/VIM(-) to S100beta(+)/GFAP(+)/VIM(+). Ninety percent of the astrocytes in this area express VIM. This is very high compared with findings in stab-wound preparations, where only 10% of astrocytes (surrounding entire lesion) are found to be VIM(+). A third region, consistent with a remote or distal reactive gliotic response, demonstrated staining for S100beta and had increased GFAP contents throughout the neocortical hemisphere. Cells in this region were never found to be VIM(+). Among S100beta(+) cells close to the boundary region, more than 80% expressed detectable GFAP by 2 days after lesioning. S100beta(+) cells 1 mm more laterally (distal to lesion) did not express GFAP to the same level until day 6. Thus, we find three immunohistochemically distinct populations of reactive astrocytes surrounding the focal ischemic lesion. In contrast to the case for stab-wound traumatic injury, the response closest to and surrounding the lesion did not up-regulate GFAP or VIM by day 6. The proximal response was, instead, more remote and only at the base of the lesion, extending partly into the corpus callosum.     

Post traumatic lesion absence of beta-dystroglycan-immunopositivity in brain vessels coincides with the glial reaction and the immunoreactivity of vascular laminin

Following brain lesions, the gliovascular basal lamina undergoes destruction and the gliovascular connections 'decouple'. Laminin receptors, as dystroglycan, are essential in these processes. The present study compares the immunoreactivities of beta-dystroglycan, glial fibrillary acidic protein (GFAP), and laminin following stab wounds in adult rats. In intact brain the vessels were immunopositive to beta-dystroglycan, whereas the laminin of their basal lamina proved to be unavailable to immunoreactions. Following stab wound, however, the adjacent vessels lost their immunopositvity to beta-dystroglycan, whereas immunopositivity to laminin became detectable in them. In an advanced stage of glial reaction the territory of GFAP immunopositive reactive astrocytes coincided with the area where vessels lost their immunopositivity to beta-dystroglycan. When glial reaction regressed, the beta-dystroglycan immunopositivity re-appeared, and laminin immunopositivity became undetectable again. Post-lesional disappearance of vascular beta-dystroglycan immunostaining was described earlier, and was attributed to the cleavage of beta-dystroglycan by matrix metalloproteinases as a mechanism of the decoupling of the gliovascular connections. Our results, which were obtained in a different type of lesion support that the loss of vascular beta-dystroglycan immunopositivity is a general phenomenon following cerebral lesions, and an indirect marker of gliovascular decoupling. For the first time coincidences were presented between vascular immunonegativity to beta-dystroglycan, glial reaction and detectability of laminin. Manifestation of laminin immunoreactivity also indicates gliovascular decoupling. Coincidence between glial reaction and lack of vascular beta-dystroglycan suggests mutual enhancement between them. The observations may have clinico-pathologic importance since similar investigations may help to follow the progression and regression of post-lesion processes.     

Co-expression of glial fibrillary acidic protein and vimentin in reactive astrocytes following brain injury in rats.

The immunohistochemical expression of glial fibrillary acidic protein (GFAP) and vimentin (VIM) was studied in reactive astrocytes of the rat cerebral cortex 5 days after a brain injury. Seriated Epon semithin sections were immunostained alternatively for GFAP or VIM. Thereafter, both antigens were detected in consecutive sections of the same cell. Bordering the wound, an inner reactive glial layer 300-350 microns thick, showed positive astrocytes with the two immunohistochemical techniques. In this layer, about 60% of the GFAP positive astrocytes were also positive for VIM. Outside the inner layer, only GFAP positive astrocytes could be found.     

The amniotic band syndrome as a cause of anencephaly

Summary Extremely severe gliosis develops at the end stage of infantile neuronal ceroid-lipofuscinosis (INCL), a fatal encephalopathy characterized by accumulation of autofluorescent storage material in the brain and other tissues followed by a terminal subtotal neuronal and myelin loss. A major fraction of highly enriched intermediate filaments was obtained with a density gradient centrifugation method from INCL brain tissue, whereas the storage material represented only a minor fraction. SDS-polyacrylamide gel electrophoresis of the filament fraction showed a major protein with molecular weight of 51 kD and three to four polypeptides of 40–48 kD identified as glial fibrillary acidic protein (GFAP) and its degradation products by the immunoblotting technique with monoclonal antibodies against GFAP. Immunization experiments with the isolated INCL glial filament fraction produced antibodies reacting only with GFAP but not with other types of intermediate filament proteins, furthermore indicating a high content of GFAP in the isolated fraction. No significant amounts of vimentin or other types of intermediate filament proteins could be detected. These results document the extremely high content of glial filaments at the terminal stage of INCL and suggest that INCL brain may serve as a good human model for studies on the composition of glial filaments in vivo and on the pathogenesis of gliosis.Supported by grants from the Research Department of Rinnekoti Foundation, Finska Läkaresällskapet, the Sigrid Juselius Foundation and the Finnish Medical Research Council     

Human influenza viral infection in utero alters glial fibrillary acidic protein immunoreactivity in the developing brains of neonatal mice

Epidemiological reports describe a strong association between prenatal human influenza viral infection and later development of schizophrenia. Postmodern human brain studies, however, indicate a lack of gliosis in schizophrenic brains presumably secondary to absence of glial cells during the second trimester viral infection in utero. We hypothesized that human influenza infection in day 9 pregnant mice would alter the expression of glial fibrillary acidic protein (GFAP, an important marker of gliosis, neuron migration, and reactive injury) in developing brains of postnatal days 0, 14 and 35 mice. Determination of cellular GFAP immunoreactivity (IR) expressed as cell density in cortex and hippocampus of control and experimental brains showed increases in GFAP-positive density in exposed cortical (P = 0.03 day 14 vs control) and hippocampal cells (P = 0.035 day 14, P = 0.034 day 35). Similarly, ependymal cell layer GFAP-IR cell counts showed increases with increasing brain age from day 0, to days 14 and 35 in infected groups (P = 0.037, day 14) vs controls. The GFAP-positive cells in prenatally exposed brains showed 'hypertrophy' and more stellate morphology. These results implicate a significant role of prenatal human influenza viral infection on subsequent gliosis, which persists throughout brain development in mice from birth to adolescence.     

Immunohistochemistry of glial reaction after injury in the rat: Double stainings and markers of cell proliferation

The astrocytic reaction in the rat after brain injury has been studied immunohistochemically for intermediate filaments (GFAP and vimentin), also with double staining procedures, and for markers of proliferation (BrdU and PCNA). GFAP-positive reactive astrocytes appeared around the lesion, where they were vimentin-positive and at a distance. BrdU and PCNA showed a high labelling index around the wound at day 2 and scattered positive nuclei were also found at a distance in the ipsilateral side. BrdU-positive astrocytes represented a minor fraction of GFAP- and vimentin-positive astrocytes. The expression of vimentin persisted at least 15 days after the lesion. Our results could suggest that distant reactive astrocytes originate through hypertrophy while those close to lesion arise by hyperplasia from mature or immature glial cells. The hypothesis is formulated that cells of the periventricular matrix contribute to the post-traumatic proliferative activity.     

The expression of TAPA (CD81) correlates with the reactive response of astrocytes in the developing rat CNS

During the development of the brain, astrocytes acquire the ability to become reactive and form a scar. This change in the astrocytes occurs at approximately the same time that there is a decrease in the regenerative capacity of the CNS. Previous work from our laboratory had revealed that TAPA (Target of Anti-Proliferative Antibody, also known as CD81) is associated with reactive gliosis and the glial scar. TAPA is a member of the tetraspan family of proteins that appears to be associated with the regulation of cellular behavior. In order to define the role of TAPA in relation to the developmentally regulated CNS response to injury, we examined the levels of TAPA and GFAP immunoreactivity in rat pups that received a penetrating cerebral cortical injury. All of the animals injured at postnatal day 9 (PND 9), PND 18, or as adults, exhibited reactive gliosis scar formation when they were sacrificed 10 days after the cortical injury. Of the nine animals injured at PND 2, only three displayed reactive gliosis and scar formation. The remaining six rat pups had either a modest gliotic response or no detectable gliosis. The level of TAPA at the site of injury mimicked the reactive gliosis as defined by GFAP immunoreactivity. In all of the rats with a glial scar, there was a dramatic upregulation of TAPA that is spatially restricted to the reactive astrocytes. These results suggest that the upregulation of TAPA is an integral component of glial scar formation.     

Reactive gliosis in the brains of Lewis rats with experimental allergic encephalomyelitis

This paper assesses reactive gliosis in the optic tracts and other regions of brain in Lewis rats with experimental autoimmune encephalomyelitis (EAE). Enhanced immunostaining for glial fibrillary acidic protein (GFAP) in brains from rats with EAE occurred primarily in the white-matter tracts and was not restricted to sites of inflammation. Immunocytochemical staining for other putative astrocytic antigens demonstrated glutathione-S-transferase (Yb isoenzyme) to be localized extensively in GFAP-positive cells and vimentin to be present both in inflammatory cells and in some GFAP-positive astroglial cells. Positive staining for carbonic anhydrase and glutamine synthetase was observed in oligodendrocytes. In the optic tracts glutamine synthetase, but not carbonic anhydrase, was also observed in some astrocytes.     

Upregulation of metabotropic glutamate receptor subtype mGluR3 and mGluR5 in reactive astrocytes in a rat model of mesial temporal lobe epilepsy

Reactive gliosis is a prominent morphological feature of mesial temporal lobe epilepsy. Because astrocytes express glutamate receptors, we examined changes in metabotropic glutamate receptor (mGluR) 2/3, mGluR5 and transforming growth factor (TGF)-beta in glial cells of the hippocampal regions in an experimental rat model of spontaneous seizures. Rats that exhibited behavioural status epilepticus (SE) directly after 1 h of electrical angular bundle stimulation, displayed chronic spontaneous seizures after a latent period of 1-2 weeks as observed using continuous electrographic monitoring. SE resulted in hypertrophy of astrocytes and microglia activation throughout the hippocampus as revealed by immunolabelling studies. A dramatic, seizure intensity-dependent increase in vimentin immunoreactivity (a marker for reactive astrocytes) was revealed in CA3 and hilar regions where prominent neuronal loss occurs. Increased vimentin labelling was first apparent 24 h after onset of SE and persisted up to 3 months. mGluR2/3 and mGluR5 protein expression increased markedly in glial cells of CA3 and hilus by 1 week after SE, and persisted up to 3 months after SE. Double immunolabelling of brain sections with vimentin confirmed co-localization with glial fibrillary acidic protein (GFAP), mGluR2/3 and mGluR5 in reactive astrocytes. TGF-beta, a cytokine implicated in mGluR3-mediated neuroprotection, was also upregulated during the first 3 weeks after SE throughout the hippocampus. This study demonstrates seizure-induced upregulation of two mGluR subtypes in reactive astrocytes, which - together with the increased production of TGF-beta - may represent a novel mechanism for modulation of glial function and for changes in glial-neuronal communication in the course of epileptogenesis.     

SC1/hevin and reactive gliosis after transient ischemic stroke in young and aged rats

SC1 is a member of the SPARC family of glycoproteins that regulate cell-matrix interactions in the developing brain. SC1 is expressed in astrocytes, but nothing is known about the expression in the aged or after stroke. We found that after focal striatal ischemic infarction in adult rats, SC1 increased in astrocytes surrounding the infarct and in the glial scar, but in aged rats, SC1 was lower at the lesion edge. Glial fibrillary acidic protein (GFAP) also increased, but it was less prominent in reactive astrocytes further from the lesion in the aged rats. On the basis of their differential expression of several molecules, 2 types of reactive astrocytes with differing spatiotemporal distributions were identified. On Days 3 and 7, SC1 was prevalent in cells expressing markers of classic reactive astrocytes (GFAP, vimentin, nestin, S100β), as well as apoliprotein E (ApoE), interleukin 1β, aggrecanase 1 (ADAMTS4), and heat shock protein 25 (Hsp25). Adjacent to the lesion on Days 1 and 3, astrocytes with low GFAP levels and a "starburst" SC1 pattern expressed S100β, ApoE, and Hsp32 but not vimentin, nestin, interleukin 1β, ADAMTS4, or Hsp25. Neither cell type was immunoreactive for NG2,CC-1, CD11b, or ionized calcium-binding adapter-1. Their differing expression of inflammation-related and putatively protective molecules suggests different roles for starburst and classic reactive astrocytes in the early glial responses to ischemia.     

Astrocyte activation and wound healing in intact‐skull mouse after focal brain injury

Localised brain tissue damage activates surrounding astrocytes, which significantly influences subsequent long‐term pathological processes. Most existing focal brain injury models in rodents employ craniotomy to localise mechanical insults. However, the craniotomy procedure itself induces gliosis. To investigate perilesional astrocyte activation under conditions in which the skull is intact, we created focal brain injuries using light exposure through a cranial window made by thinning the skull without inducing gliosis. The lesion size was maximal at ~ 12 h and showed substantial recovery over the subsequent 30 days. Two distinct types of perilesional reactive astrocyte, identified by GFAP upregulation and hypertrophy, were found. In proximal regions the reactive astrocytes proliferated and expressed nestin, whereas in regions distal to the injury core the astrocytes showed increased GFAP expression but did not proliferate, lacked nestin expression, and displayed different morphology. Simply making the window did not induce any of these changes. There were also significant numbers of neurons in the recovering cortical tissue. In the recovery region, reactive astrocytes radially extended processes which appeared to influence the shapes of neuronal nuclei. The proximal reactive astrocytes also formed a cell layer which appeared to serve as a protective barrier, blocking the spread of IgG deposition and migration of microglia from the lesion core to surrounding tissue. The recovery was preceded by perilesional accumulation of leukocytes expressing vascular endothelial growth factor. These results suggest that, under intact skull conditions, focal brain injury is followed by perilesional reactive astrocyte activities that foster cortical tissue protection and recovery.     

Reactive astroglia-neuron relationships in the human cerebellar cortex: A quantitative, morphological and immunocytochemical study in Creutzfeldt-Jakob disease

In order to investigate the role of neuron-glia interactions in the response of astroglial to a non-invasive cerebellar cortex injury, we have used two cases of the ataxic form of Creutzfeldt-Jakob disease (CJD) with distinct neuronal loss and diffuse astrogliosis. The quantitative study showed no changes in cell density of either Purkinje or Bergmann glial cells in CJ-1, whereas in the more affected CJ-2 a loss of Purkinje cells and an increase of Bergmann glial cells was found. The granular layer in both CJD cases showed a similar loss of granule cells (about 60% ) in parallel with the significant increase in GFAP+ reactive astrocytes. GFAP immunostaining revealed greater reactivity of Bergmann glia in CJ-2 than in CJ-1, as indicated by the thicker glial processes and the higher optical density. Granular layer reactive astrocytes were regularly spaced. In both CJD cases there was strict preservation of the spatial arrangement of all astroglial subtypes—Fañanas cells, Bergmann glia and granular layer astrocytes. Reactive Fañanas and Bergmann glial cells and microglia/macrophages expressed vimentin, while only a few vimentin+ reactive astrocytes were detected in the granular layer. Karyometric analysis showed that the increase in nuclear volume in reactive astrloglia was directly related with the level of glial hypertrophy. The number of nucleoli per nuclear section was constant in astroglial cells of human controls and CJD, suggesting an absence of polyploidy in reactive astroglia. Ultrastructural analysis revealed junctional complexes formed by the association of macula adherens and gap junctions. In the molecular layer numerous vacant dendritic spines were ensheathed by lamellar processes of reactive Bergmann glia. Our results suggest that quantitative (neuron/astroglia ratio) and qualitative changes in the interaction of neurons with their region-specific astroglial partners play a central role in the astroglial response pattern to the pathogenic agent of CJD.     

Glial domains and nerve fiber patterns in the fish retinotectal pathway

Optic nerve fibers run parallel from the retina as far as the optic tract in fish, then suddenly criss-cross into a new pattern matching the tectal map. This change coincides with a unique demarcation between two astroglial territories in the retinotectal pathway, located where the optic chiasm occurs in other vertebrates, which we defined using antibodies directed against intermediate filaments (IF). We found that astroglia in optic nerve territory express an Mr 56,000 IF polypeptide, band 3, which we identify as the fish equivalent of vimentin in mammals. These astrocytic cells lack glial fibrillary acidic protein (GFAP; cf. Dahl and Bignami, 1973). Conversely, glia in brain territory, that is, in the optic tract and elsewhere in the CNS, lack the fish vimentin, but express GFAP. By electron microscopy, we obtained evidence that new retinal axons extend swiftly through the growing optic nerve, where they are tightly shepherded into a narrow track by newly differentiating glial cells, positive for the fish vimentin. In the GFAP-positive glial territory of the optic tract, by contrast, growing axons are slowed down and probably branch. We suggest that this allows them to fasciculate accurately with older fibers and thereby propagate a tectotopic pattern established by pioneer axons in the embryo.     

Glial fibrillary acidic protein and vimentin in the experimental glial reaction of the rat brain

Glial reaction has been studied in the rat by the immunohistochemical demonstration of glial fibrillary acidic protein (GFAP) and vimentin (VIM) in two experimental conditions. The first was represented by a necrotic cerebral lesion obtained by laser irradiation and the second by the development of experimental tumors induced by transplacental ethylnitrosourea. Reactive astrocytes develop not only in the proximity of the lesion but also distant from it. The intensity of the glial response seems to depend upon the normal distribution of astrocytes and the perilesional edema. GFAP decorates all the reactive astrocytes, whereas VIM is positive only in those at the edges of the lesion. The significance of the different responses in the two models and between the two intermediate filaments is discussed.