Search

It is now well established that the glial fibrillary acidic protein (GFAP) is the principal 8-9 nm intermediate filament in mature astrocytes of the central nervous system (CNS). Over a decade ago, the value of GFAP as a prototype antigen in nervous tissue identification and as a standard marker for fundamental and applied research at an interdisciplinary level was recognized (Raine, 135). As a member of the cytoskeletal protein family, GFAP is thought to be important in modulating astrocyte motility and shape by providing structural stability to astrocytic processes. In the CNS of higher vertebrates, following injury, either as a result of trauma, disease, genetic disorders, or chemical insult, astrocytes become reactive and respond in a typical manner, termed astrogliosis. Astrogliosis is characterized by rapid synthesis of GFAP and is demonstrated by increase in protein content or by immunostaining with GFAP antibody. In addition to the major application of GFAP antisera for routine use in astrocyte identification in the CNS, the molecular cloning of the mouse gene in 1985 has opened a new and rich realm for GFAP studies. These include antisense, null mice, and numerous promoter studies. Studies showing that mice lacking GFAP are hypersensitive to cervical spinal cord injury caused by sudden acceleration of the head have provided more direct evidence for a structural role of GFAP. While the structural function of GFAP has become more acceptable, the use of GFAP antibodies and promoters continue to be valuable in studying CNS injury, disease, and development.

Experimental allergic encephalomyelitis (EAE) was induced in adult Lewis rats with purified guinea pig CNS myelin and Freund's adjuvant. As soon as the very earliest clinical signs appeared the animals were perfused with fixatives and the spinal cord analyzed by electron microscopy, silver methods, and immunocytochemistry. Our findings suggest that in the early stages of EAE a sequence of events can be traced, although these events frequently overlap. The earliest morphological change appears to be astrocytic edema in both the cell body and processes. Increased amounts of glycogen particles and dispersion of glial filaments are prominent. These changes seem to occur just prior to the time when inflammatory cells begin to penetrate the capillary walls. Invasion of the neuropil mainly by macrophages and lymphocytes closely follows. Both macrophages and microglia seem to participate in phagocytosis of oligodendrocytes and myelin. Demyelination, however, is not a prominent feature at this early stage.

The outpouching of CNS tissues into the entering spinal nerve roots was documented by light and electron microscopy of human and bovine tissues. Astrocytic processes containing large bundles of glial filaments were very prominent in the nerve entry zone and extended for short distances into the adjacent endoneurium of the spinal nerve roots. Antiserum raised to glial acidic fibrillary (GFA) protein stained these glial elements, thereby characterizing the dome-shaped evaginations of CNS tissues into the nerve root entry zones. Antisera to CNS basic protein showed enhanced staining in the nerve entry zone. Analyses of nerve proteins by SDS gel electrophoresis disclosed a prominent 49,000 MW protein in the bovine and human nerve root entry zone. This protein was also prominent in spinal cord white matter, but was not seen in nerve roots which were not admixed with glial tissues. This finding supported the view that a 49,000 MW protein is a glial filaments but is not a component of bovine or human neurofilaments.

Twenty-day-old mice received a single tail vein injection of [guanido-14C]arginine. The cytoskeleton was extracted from the spinal cords at varying lengths of time thereafter. Glial fibrillary acidic protein (GFAP) formed a distinct, broad band that was widely separated from other protein bands in one-dimensional polyacrylamide gels. The purity of the GFAP band was verified by Western blot analysis of one- and two-dimensional electrophoretic patterns. In addition, enzyme-linked immunosorbent assay and quantitative Western blot analysis indicated that 95% of the total spinal cord GFAP was extracted in the cytoskeletal preparation. The specific activity of GFAP was obtained by eluting the protein from the cytoskeletal GFAP band in preparative one-dimensional gels. Specific activity reached a peak 2 h after injection with [14C]arginine. Forty percent of the incorporated radioactivity was still present in cytoskeletal GFAP at 9 weeks, indicating that a significant proportion of glial filaments turns over relatively slowly in vivo.

Acute experimental allergic encephalomyelitis (EAE) in the Lewis rat is a cell-mediated autoimmune disease of central nervous system myelin. The lesion has been characterized by breakdown of the blood-brain barrier, edema, and periventricular infiltration of macrophages and lymphocytes. At the early stage of the disease, the astrocytes show a marked increase in immunostaining for glial fibrillary acidic protein (GFAP). A corresponding increase in GFAP content, however, cannot be demonstrated. Electron microscopic examination of the early lesion shows a typical reactive astrocytic response expressed by an enlarged watery cytoplasm, particularly at the level of the processes surrounding neurons and blood vessels and in the neuropil itself. The astroglial processes contain numerous glycogen particles (aggregates and single particles). Glial filaments are also conspicuous and are arranged in small bundles or loose thin filaments adjacent to the bundles. The glial filaments that normally appear as tight bundles have expanded and appear less dense. We suggest that the increase in GFAP immunostaining of the astrocytes in the early lesion is due in part to edema, which causes dissociation of the filaments and thereby exposes more antigenic sites to the antibodies.

Glial fibrillary acidic (GFA) protein was synthesized in vitro in a rabbit reticulocyte lysate system programmed with messenger RNA (mRNA) extracted from Jimpy mouse spinal cord. It was identical in molecular weight and charge to that synthesized from normal mouse mRNA and GFA protein extracted from normal mouse cord. These data suggest that the Jimpy mutation does not affect the primary phenotypic expression of GFA protein.

In situ and in vitro degradation of glial fibrillary acidic (GFA) protein in mouse spinal cord was examined with electroblots stained for GFA protein by the peroxidase anti-peroxidase method. Non-degraded, intact GFA protein had a molecular weight of 48 Kdaltons and isoelectric points ranging from pH 5.8 to 6.4. The molecular weights of immunoreactive degradation products ranged from 47 to 28 Kdaltons. All of the degradation products had acid shifted isoelectric points (pH 5.8-5.2). Degradation was prevented by chelating calcium with EGTA. In contrast to in situ degradation, degradation in vitro with 3 mM CaCl2 occurred at a faster rate. The effect of pH and temperature on the degradation process were determined by incubating homogenized spinal cords in 3 mM CaCl2 solutions varying in pH from 4 to 10 and at 4, 37, and 60 degrees C. The greatest number of immunoreactive bands with the lowest molecular weights occurred at pH 8 and 37 degrees C. The results suggest that turnover of glial filaments is in part controlled by a calcium dependent proteinase active near neutral pH similar to that postulated for neurofilament turnover.

In acute experimental allergic encephalomyelitis (EAE), astrocytes in spinal cord tissue hypertrophy and stain intensely with antibody to the glial fibrillary acidic protein (GFAP). We attempted to determine if this activation is a result solely of hypertrophy of existing astrocytes or if astrocyte division might also occur. Lewis rats in various stages of acute EAE were injected with [3H]thymidine, the spinal cord sections were prepared, immunostained for GFAP and processed for radioautography. In spinal cords from rats administered thymidine on days 11-15 after sensitization a large number of mononuclear cells showed radioactive label. Many of these labeled cells, most likely monocytes and lymphocytes, were associated with inflammatory lesions, but others were located in the CNS parenchyma at great distances from the lesions. Most cells staining for the GFAP were hypertrophied with greatly extended cell processes, and the nuclei of some of these cells identified as astrocytes were overlaid with silver grains, indicating uptake of [3H]thymidine. In addition a few ependymal cells appeared to be labeled. No GFAP-stained cells from the Freund's adjuvant controls contained radioactive label. Similar studies using SJL/J mice with chronic relapsing EAE yielded very few labeled inflammatory cells or astrocytes. This study indicates that division takes place in some astrocytes in acute EAE, but occurs much less frequently in chronic EAE. Probably most of the increase in GFAP-stained material is a result of hypertrophy of astrocytes rather than of massive cell division.

Glial fibrillary acidic (GFA) protein has been synthesized in an RNA-dependent cell-free system derived from rabbit reticulocytes. The cell-free synthesized product appears to have the same size as GFA protein isolated from bovine spinal cord, thus showing that GFA protein does not undergo detectable proteolytic processing.

Guinea pig antiserum specific for purified bovine choline acetyltransferase has been shown to cross-react with rabbit enzyme. We used the peroxidase-antiperoxidase immunohistochemical method to demonstrate the localization of choline acetyltransferase in formalin-fixed and paraffin-embedded sections of rabbit spinal cord and cerebellum. In the spinal cord, in agreement with our and others' previous results using immunofluorescent techniques, choline acetyltransferase was found in the cell bodies of the ventral horn motor neurons. In the cerebellum, choline acetyltransferase was localized exclusively in the mossy fibers and the glomeruli of the cerebellar folia. The immunohistochemical findings in the cerebellum reveal the morphological detail of cholinergic axons and their terminals. The results are consistent with published biochemical data on the cerebellar distribution of choline acetyltransferase.

Cells cultured from the caudalmost area of regenerating teleost spinal cord differ both morphologically and in terms of molecular architecture from those of more rostral (more fully differentiated) areas of the cord. The caudalmost regenerating cord consists of an ependymal tube. Cells from this region have flattened or partially flattened cell somas and short spike projections in vitro; they do not exhibit the rounded cell somas nor the long, thin, branching neurites typical of differentiated neurons. A series of cultures taken from different areas along the length of regenerating spinal cord were examined for molecular differentiation by staining with a monoclonal antibody against non-phosphorylated neurofilament protein (SMI 32). None of the cells from the caudalmost culture of the regenerating spinal cord stained with antibody SMI 32. In cultures of more rostral regenerated cord, the cells with neuronal morphology do stain positively with the anti-neurofilament antibody. This result suggests that the cells in the cultures of caudalmost cord represent relatively undifferentiated ependymal cells, or ependyma-derived cells, which later differentiate into neurons and glia in the regenerating spinal cord.

Post-embedding peroxidase-antiperoxidase methods to stain glial fibrillary acidic (GFA) protein in Araldite-embedded sections for light and electron microscopy were developed. The Jimpy mouse spinal cord was used because it is gliotic and contains abundant glial filaments and GFA protein. For light microscopy, specific staining was obtained in thick and in ultrathin sections mounted on glass following removal of the plastic with sodium ethoxide. Consistent specific staining for GFA protein in ultrathin sections mounted on nickel grids required partial removal of the plastic with 1% sodium ethoxide and further treatment with 2% sodium dodecyl sulfate (SDS).

General aspects of metabolic features of the most prominent CNS intermediate filament proteins, the 200,000 (200K), 150,000 (150K), and 70,000 (70K) dalton proteins of the neuron, and the glial fibrillary acidic protein (GFAP) have been explored using the incubated spinal cord slice from the rat. Measurement of short-term uptake of 3H-labeled amino acids into the individual proteins separated on polyacrylamide gels revealed that of the three neurofilament proteins, 200K was most metabolically active, 150K was less active, and 70K contained very little incorporated radioactivity. Glial fibrillary acidic protein based on Coomassie blue stain affinity showed less metabolic activity than any of the neurofilament proteins. Those relationships were constant at all ages, but the metabolic activity of all CNS intermediate filaments decreased with age. When Ca2+ was present in the medium of the incubated slices, the intermediate filaments were rapidly destroyed, but GFAP was more resistant to degradation than the neurofilament proteins. GFAP and probably the neurofilament proteins also were relatively resistant to Ca2+-activated degradative mechanisms in spinal cords of rats at younger ages (15 day) than in those of older animals (10-18 months). It is likely that the Ca2+ activated protease is less active in developing animals in which the nerve tracts are still elongating, than in adults. These results suggest that GFAP is less active metabolically and more resistant to degradation than the neurofilament proteins at all stages of maturation, but that metabolic activity of all CNS intermediate filaments decreases with age while the susceptibility to degradation increases.

Immunohistological and ultrastructural studies were carried out on normal and regenerating spinal cord of the gymnotid Sternarchus albifrons, and in the brain and spinal cord of the goldfish Carassius auratus, to examine the distribution of glial fibrillary acidic protein (GFAP) in these tissues. Sections of normal goldfish brain and spinal cord exhibited positive staining for GFAP. In normal Sternarchus spinal cord, electron microscopy has revealed filament-filled astrocytic processes; however, such astrocytic profiles were more numerous in regenerated cord. Likewise, while normal Sternarchus spinal cord showed only a small amount of GFAP staining, regenerated cords were strongly positive for GFAP. Positive staining with anti-GFAP was observed along the entire length of the regenerated cord in Sternarchus, and was especially strong in the transition zone between regenerated and unregenerated cord. Both regeneration of neurites and production of new neuronal cell bodies occur readily in such regenerating Sternarchus spinal cords (Anderson MJ, Waxman SG: J Hirnforsch 24: 371, 1983). These results demonstrate that the presence of GFAP and reactive astrocytes in Sternarchus spinal cord does not prevent neuronal regeneration in this species.

The anatomical distribution of P2 protein was studied in human autopsy tissue. Spinal cord (SC) and peripheral nerve (PN) were stained by the peroxidase-antiperoxidase method with antisera to bovine P2, glial fibrillary acidic protein, and myelin basic protein (BP). P2 antiserum did not stain all of the myelin in the PN. The staining was randomly distributed and discontinuous along a given myelinated axon. P2 antiserum also stained SC myelin in a pattern similar to the PN. Only a fraction of the sheaths stained, in contrast to BP antiserum that stained all myelin sheaths in both the SC and PN. P2-positive myelin was distributed throughout the SC white matter, including an occasional myelinated fiber in the SC grey matter. P2 and BP antisera did not stain regions of demyelination in a case of idiopathic polyneuritis, while adjacent myelinated PN stained normally. Absorption of the P2 antiserum with P2, bovine PN or bovine SC (carefully dissected to eliminate PN contamination) nullified the specific staining in both the PN and SC; however, absorption with BP or hemispheric myelin did not eliminate P2 staining. The P2 antiserum formed a single immunodiffusion line with pure P2 and acid extracts of bovine SC and PN myelin, but not with an acid extract of bovine hemispheric myelin. Electrophoresis of defatted bovine SC produced a distinct band corresponding to P2. Therefore, three lines of evidence, immunocytochemical, immunodiffusion and electrophoretic, suggest that P2 is present in PN and SC but not in hemispheric myelin.