Glial fibrillary acidic protein expression in a new human glioma cell line in culture before and after xenogenic transplantation into nude mice.

A human glioma cell line, SA146, was initiated on precoated extracellular matrix from a stereotactic biopsy of a glioblastoma. We report modulation in the expression of glial fibrillary acidic protein (GFAP) by SA146 passed in vitro before or after xenogenic transplantation into nude mice. Immunofluorescence data show a decrease in the percentage of GFAP-expressing cells with increasing in vitro passages but a full reexpression (100% of GFAP-positive cells among vimentin-positive cells) was observed in cultures just derived from the xenotransplanted tumor. These changes are correlated with the mRNA content (Northern blot probed with a cDNA for GFAP) and with the protein level (cytoskeletal fraction analyzed by two-dimensional gel electrophoresis and Western blots probed with a monoclonal antibody). At the optimal level of GFAP expression, a large range of micro-heterogeneity in GFAP isoforms is reached for which post-translational events are clearly involved since mRNA translation in cell free system would provide at best three isomers. We suggest that SA146 would be an appropriate model to study the regulation of GFAP expression in the context of human glial tumor biology.

Disrupted glial fibrillary acidic protein network in astrocytes from vimentin knockout mice.

Glial fibrillary acidic protein (GFAP) is an intermediate filament protein expressed predominantly in astrocytes. The study of its expression in the astrocyte lineage during development and in reactive astrocytes has revealed an intricate relationship with the expression of vimentin, another intermediate filament protein widely expressed in embryonic development. these findings suggested that vimentin could be implicated in the organization of the GFAP network. To address this question, we have examined GFAP expression and network formation in the recently generated vimentin knockout (Vim-) mice. We show that the GFAP network is disrupted in astrocytes that normally coexpress vimentin and GFAP, e.g., those of the corpus callosum or the Bergmann glia of cerebellum. Furthermore, Western blot analysis of GFAP protein content in the cerebellum suggests that posttranslational mechanisms are implicated in the disturbance of GFAP network formation. The role of vimentin in this process was further suggested by transfection of Vim-cultured astrocytes with a vimentin cDNA, which resulted in the normal assembly of the GFAP network. Finally, we examined GFAP expression after stab wound-induced astrogliosis. We demonstrate that in Vim- mice, reactive astrocytes that normally express both GFAP and vimentin do not exhibit GFAP immunoreactivity, whereas those that normally express GFAP only retain GFAP immunoreactivity. Taken together, these results show that in astrocytes, where vimentin is normally expressed with GFAP fails to assemble into a filamentous network in the absence of vimentin. In these cells, therefore, vimentin appears necessary to stabilize GFAP filaments and consequently the network formation.

Glial fibrillary acidic protein mRNA isotypes: expression in vitro and in vivo.

Glial fibrillary acidic protein (GFAP) and its mRNA, primarily expressed in astrocytes, are also expressed in peripheral nervous system Schwann cells as well as in certain non-neural tissues. Schwann cells express a GFAP mRNA (GFAP-beta) which differs from the CNS-type mRNA (GFAP-alpha) by the presence of an extended 5' untranslated region. We have developed a polymerase chain reaction assay which allows distinction of these two GFAP mRNAs, as well as quantitative analysis of their levels. In the cultured rat Schwannoma cell line RT4-D6, GFAP-beta was the major GFAP mRNA species, accounting for at least 75% of total GFAP (alpha + beta) mRNA. GFAP-beta was also detected in primary rat astrocyte cultures, where it constituted approximately 5% of the total GFAP mRNA, as well as in RNA samples prepared from normal rat cerebral cortex, and from hamster and human brain. In rat cortex, the temporal expression of GFAP-beta mRNA paralleled that of total GFAP mRNA, with plateau levels reached between postnatal days 15 and 20. In astrocyte cultures, the relative levels of GFAP-alpha and -beta mRNAs were differentially regulated by exposure to interferon-gamma (10 to 25 units/ml), which caused an increase in GFAP-beta levels while at the same time no change or a small decrease in total GFAP levels. In rat brain cortical slices, 4 hr exposure to 25 units/ml interferon-gamma decreased total GFAP mRNA levels over tenfold, while GFAP-beta levels were unaffected. These data indicate that a second form of the GFAP mRNA is expressed in astrocytes both in vivo and in vitro and provide evidence for independent regulation of these two GFAP mRNA species.

Normal and pathological expression of GFAP promoter elements in transgenic mice.

The expression of the glial fibrillary acidic protein (GFAP), a component of astroglial intermediate filaments, is regulated under developmental and pathological conditions. In order to characterize DNA sequences involved in such regulations, we produced transgenic mice bearing 2 kb of the 5' flanking region of the murine GFAP gene linked to the Escherichia coli beta-galactosidase (beta-gal) reporter gene. Seven transgenic lines were obtained. We observed that the regulatory elements present in the transgene GFAP-nls-LacZ direct an expression in the neural and non-neural tissue and target in vivo an unexpected subpopulation of astrocyte. In the developing brain, beta-gal activity and GFAP appeared simultaneously and in the same region, on embryonic day 18 (E18), suggesting that the 2 kb of the promoter contains the regulatory sequences responsible for the perinatal vimentin/GFAP switch. In addition, we demonstrated that the 2 kb sequence of the GFAP promoter used in the transgene possess elements which are activated after a surgical injury, thus permitting to study some aspects of reactive gliosis in these transgenic mice. These transgenic lines provide a useful tool by enabling further studies of astroglial and, probably, neuronal physiologies.

Antibodies raised against a peptide corresponding to a Gs alpha domain reveal a vimentin-associated protein.

In an attempt to localize the guanine nucleotide binding protein Gs alpha by immunocytochemistry, we synthetized peptides corresponding to several stretches of residues deduced from the published cDNA sequence of Gs alpha and raised antibodies against them. Among the peptides, one corresponding to residues 367-381 elicited antibodies that immunocytochemically detected, at the optical level, what appeared to be vimentin in several cells and tissues. Studies at the ultrastructural level confirmed this observation and also showed weak staining of some areas of plasma membranes of glial and nerve cells. Analysis by Western blots of rat brain subcellular fractions indicated that: (a) the protein stained by the anti-peptide antibodies was associated with the cytoskeleton; and (b) it was not vimentin but a protein of higher molecular weight, 65 KD. We accordingly suggest that the Gs alpha-derived peptide elicited two types of antibodies, one recognizing Gs alpha in fixed tissues, the other recognizing an epitope located in a vimentin-associated protein. This study emphasizes the caution that is needed when conclusions are drawn on the basis of immunocytochemical studies using antipeptide antibodies.

Regional and developmental variations of GFAP and actin mRNA levels in the CNS of jimpy and shiverer mutant mice.

Gliosis is a common reaction to brain damage. Glial fibrillary acidic protein (GFAP) is a classical astrocytic marker. We have undertaken to measure the level of GFAP-mRNA as an index of gliosis in the brain of jimpy (jp) and shiverer (shi) murine mutants, in which hypomyelination is either severe or moderate, respectively. This study was conducted in five different CNS regions and at different ages. In young jp mutant, the amount of GFAP-mRNA was either normal or lower than in control animals; but after 3 wk of age, the level of GFAP-transcript increased dramatically in all regions examined. A parallel increase in actin-mRNA was also observed, mostly in the diencephalon and to a lesser extent in cortex and spinal cord, but not in the cerebellum and brainstem. In the shi mutant, variations in the amount of GFAP-mRNA were less important than in the jp with two exceptions: In brainstem of 3-wk-old animals, a 2.5-fold increase was observed, and in all the regions but the spinal cord of 12-d-old shi, the levels of GFAP-transcript were 2-5 times lower than in controls. In this mutant, the levels of actin message were usually close to normal, or slightly lower than in controls.

Developmental expression of glial fibrillary acidic protein and actin-encoding messages in quaking and control mice.

Quaking is a neurological mutation leading to pleiotropic phenotypic expression, the most prominent being disturbed myelin formation in the central nervous system (CNS) with minor abnormalities in the peripheral nervous system. Previous immunochemical measurements of glial fibrillary acidic protein (GFAP) revealed a marked increase in the protein in several areas of the CNS. To further characterize the regulation parameters of GFAP synthesis, we analyzed the levels of GFAP mRNA in 5 regions of the CNS, some with elevated levels of GFAP and some without. This was compared to the developmental expression of GFAP transcripts in the same regions in normal mice. To establish the specificity of the variations observed with this astroglial specific message, we conducted a similar investigation with actin RNA which is expressed by several cell types in the CNS. Both the actin and the GFAP message were found to be increased in the adult mutant throughout the CNS. In 2-year-old normal mice the messengers for both cytoskeleton proteins were expressed in a higher amount than in young adults.

Recent evolutionary origin of the expression of the glial fibrillary acidic protein (GFAP) in lens epithelial cells. A molecular and genetic analysis of various mouse species.

We have investigated the phylogenetic distribution of the glial fibrillary acidic protein (GFAP) in lens epithelial cells (LEC) of various mouse species within the genus Mus. We have shown that lens GFAP is expressed in mice of the Mus musculus complex and in Mus spicilegus and Mus macedonicus species (L.GFAP(+) phenotype) while it is absent in Mus spretus, Mus caroli and Mus cooki species (L.GFAP(-) phenotype). Our results argue in favour of one of the phenograms illustrating the probable phylogenetic relationships between these species in the genus Mus. In animals where lens GFAP was immunodetected, Northern blots of lens RNA extracts hybridized with a mouse GFAP cDNA probe, revealed a single 2.7 kb band. Comparative Northern blot analysis of lens tissue from L.GFAP(+) mice or of brain tissue from L.GFAP(+) or L.GFAP(-) mice did not show any size heterogeneity of the GFAP mRNA. The pattern of the GFAP immunostaining of astroglial cells in brain was identical in both L.GFAP phenotypes. Analysis of interspecific crosses showed that the L.GFAP(+) character is transmitted in a dominant fashion and seems to be linked to the Mus musculus Gfap gene. In this study we have also confirmed the localization of the mouse Gfap gene on chromosome 11.

Expression of glial fibrillary acidic protein and vimentin in mouse lens epithelial cells during development in vivo and during proliferation and differentiation in vitro: comparison with the developmental appearance of GFAP in the mouse central nervous system.

Analysis of glial fibrillary acidic protein (GFAP) and vimentin in mouse lens epithelial cells (MLEC) during ontogenesis revealed a two-step developmental expression similar to that observed in astrocytes. Vimentin was first immunostained at E11 corresponding with the closure of the lens vesicle, whereas GFAP was detected only after a further 7 days (E18); this protein appeared simultaneously in the mouse lens and CNS. In the latter case, it was present in the hypothalamic tanycytes and spinal cord. This similarity in the timing of appearance of GFAP in the non-neural MLEC and in fetal astrocytes suggests a common mechanism for its expression in tissues of different embryological origin. However, it has previously been observed that, in contrast to the situation in astrocytes, GFAP disappears from differentiating MLEC in vivo. We have shown that in vitro this protein also disappears rapidly from MLEC in the presence of fetal calf serum (FCS). However, the use of mouse serum instead of FCS inhibited the migration of MLEC out of the explant, and in these cells GFAP persisted.

Expression of carbonic anhydrase II gene in early brain cells as revealed by in situ hybridization and immunohistochemistry.

A mouse carbonic anhydrase (CA II) complementary(c) DNA probe was used for in situ hybridization on mouse brain cultured cells in order to follow CA II gene expression during brain development. An improved method was established using biotinated probes that resulted in a high sensitivity and an absence of background; this method could be combined with immunohistochemistry. Hypothalamic cells of embryonic day (ED) 12-14 mice were cultured for various periods. Chronologic appearance of CA II messenger(m)RNA and protein was studied. The CA II gene transcripts are detectable as early as ED 12-13, although the protein they encode is not detectable until ED 17-18. Gene expression is restricted to 0.1% of the total population. Northern blot analysis confirmed the presence of CA II transcripts in embryonic hypothalamus. At postnatal stage, the majority of glial cells express both the CA II mRNA and the protein. Our results favour the early appearance of a glial lineage in a precise area of the developing CNS. The precocity of CA II gene transcription makes in situ hybridization an invaluable approach in defining the onset of nerve cell lineages during embryonic development.

Are close contacts between astrocytes and endothelial cells a prerequisite condition of a blood-brain barrier? The rat subfornical organ as an example.

The microvessels of the rat subfornical organ (SFO) are heterogeneous: those of the caudal part lack a blood-brain barrier (BBB) unlike those of the rostral part. The astroglial environment of these microvessels has been studied by combining an immunocytochemical technique employing an anti-GFAP (glial fibrillary acidic protein) antiserum with the morphological detection of a barrier to the protein-silver complex. All the SFO microvessels are surrounded by astrocytes characterized by a tumescent aspect; however, the relative proximity between the astrocytic feet and the endothelial cells varies considerably. The capillaries provided with a barrier (rostral SFO) are contiguous with the astrocytes from which they are only separated by a basement membrane. The capillaries devoid of BBB (caudal SFO) are surrounded by a pericapillary space that keeps the astrocytes at a short distance (capillaries with a very rich vesicular endothelium) or at a long distance (capillaries with a fenestrated endothelium). The astrocytes are absent in the choroid plexus where all microvessels are fenestrated and lack a barrier. These data suggest that the astrocytes release one or more signals which in their vicinity inhibit the expression of endothelial morphological characteristics (fenestrations, vesicles) responsible for the leakage of plasmatic proteins from the blood to the cerebral parenchyma of the circumventricular organs.

Palisading pattern of subpial astroglial processes in the adult rodent brain: relationship between the glial palisading pattern and the axonal and astroglial organization.

The architectural organization of the subpial astrocyte processes was examined near the brain surface by single immunostaining methods. The astroglial processes were stained on brain sections made parallel to the pial surface. The astroglial glial fibrillary acid protein (GFAP) antigen was used as a specific marker. We show that these subpial astrocyte processes present a well organized palisading pattern in the adult mouse and rat spinal cord, medulla and pons. This adult astrocyte palisading pattern is compared to the palisading radial glia organization we previously demonstrated in the fetal mouse brain. The observed analogies afford a new and strong argument in favor of a derivation of the subpial astrocytes from radial glia. Double immunostaining methods, using GFAP and neurofilament antigens as glial and neuronal markers respectively, show the close relationship existing between the trajectories of axonal and glial processes. Beside the colinearity already observed between the axon trajectories and the glial palisades we demonstrate a new kind of axon/glia relationship. Axons are closely intermingled, within the palisading glial tufts, with the peripheral processes of the subpial astrocytes progressing to the pial surface. The findings suggest that fetal radial glia organization has a direct and indirect influence on the adult astroglial and perhaps the axonal pattern.

Early expression of carbonic anhydrase II (CAII) in transitory glial cells of the developing murine nervous system.

Using immunocytochemical methods carbonic anhydrase II (CAII) has been detected in the embryonic mouse central nervous system as early as in stage E16. In the spinal cord and the brain stem, the enzyme first appeared in transitory cells probably derived from the radial glia. In the cerebrum such transitory cells were never stained with anti-CAII sera. The bodies of CAII-positive cells were never visualized in the ventricular layer. The controversial cellular specificity of CAII made it impossible to specify the glial lineage to which these transitory CAII-positive cells are committed.

Evidence that mouse astrocytes may be derived from the radial glia. An immunohistochemical study of the cerebellum in the normal and reeler mouse.

Astrocytic cells of unusual aspect can be detected in the cerebellum of normal mice during the first 4 weeks of life. They are visualized with anti-GFAP (glial fibrillary acidic protein), anti-S100 and anti-vimentin immune sera. Their perikaryons, located in the white matter or in the granular layer, extend long processes which are inserted onto the pial surface. These cells may be transitional forms between the radial glial cells and some of the differentiated astroglial elements. These unusual astrocytes are more numerous and heavily stained in the reeler mutant than in the normal mouse and it is suggested that our observations signify some degree of glial immaturity in the cerebellum of the mutant.

Immunohistochemical demonstration of an organized cytoarchitecture of the radial glia in the CNS of the embryonic mouse.

Immunocytochemical staining using antivimentin and anti-glial fibrillary acidic protein sera on nervous tissue sections shows a well-organized pattern of radial glial fibers as early as 15 days in the embryonic mouse. The glial fibers, isolated or in fascicles, form parallel palisades in rectilinear and longitudinal alignment in the spinal cord, medulla and pons. Certain areas show a double system of palisades, perpendicular to each other. The architecture is more complex at the cerebral level. There is a close relationship between the tracks followed by the axon bundles and the pattern of the glial palisades. The fibers have a helical structure, often with a very regular periodicity. The functional implications of the location and structure of the radial fibers during ontogenesis are discussed.

Waldenström's macroglobulinemia and peripheral neuropathy: a clinical and immunologic study of 25 patients.

We investigated, by indirect immunofluorescence, the binding of monoclonal IgM to human peripheral nerve in 25 patients with Waldenström's macroglobulinemia and peripheral neuropathy. In 10 cases (40%), an antibody activity against the myelin sheaths was demonstrated. The reactivity was mediated by the F(ab')2 fragments of the IgM. Prior delipidation of nervous tissue was needed to allow full expression of the target antigen(s). In nine cases, the IgM reacted with both peripheral and central myelin of primates, but not mouse or rabbit nervous tissue. In one case, the IgM reacted only with human peripheral nerve. The patients, whose IgM had an antibody activity to myelin antigen(s), had some distinct hematologic and neurologic features. The peripheral neuropathy always antedated the hematologic symptoms by several years. The serum level of the monoclonal IgM was low in all cases, and overt lymphoid malignancy was frequently absent. In other patients with neuropathy, the monoclonal IgM, which lacked antimyelin antibody activity, displayed either cross-idiotypic antigenic determinants or anti-intermediate filament antibody activity. These results taken together emphasize the heterogeneity of antibody activity of the monoclonal IgM from patients with Waldenström's macroglobulinemia and peripheral neuropathy.

Induction of oligodendrocyte-like properties in a primitive hypothalamic cell line by cholesterol, an eye derived growth factor and brain extract.

A serum-free medium has been devised which permits proliferation of the mouse primitive nervous cell line F7. When cholesterol, eye-derived growth factor and brain extract are added in this medium for 48 h, 80-90% of oligodendrocyte-like cells are generated. These cells have diminished substrate adhesion. They acquire the capacity to synthesize carbonic anhydrase II and myelin basic protein, two specific proteins of oligodendrocytes. These observations suggest that F7 clonal cell line, which has been previously shown to be a neurophysin cell precursor, is also a precursor for oligodendrocytes, and represents a bipotent stem cell line for both neuronal and glial cell lineages.

Immunohistochemical localization of galactosyl and sulfogalactosyl ceramide in the brain of the 30-day-old mouse.

We have used purified antibodies against galactosylceramide (galCb) and sulfogalactosylceramide (sulf) to study the topographical distribution of these two lipid haptens in the brain of the 30-day-old mouse. This study has been conducted, using the indirect immunofluorescence method, on cerebellum, brain stem and hemispherical tissue sections. Both haptens are present in the myelin sheaths and in the oligodendrocytes within the myelinated bundles. Cortical oligodendrocytes as well as some of the subependymal cells are also galCb-positive but sulf-negative. On the contrary, ciliated ependymal cells and subpial astrocytic processes (especially the Bergmann glia fibers in the cerebellum) are sulf-positive and galCb-negative. Astrocyte cell bodies and other astrocytic cell processes are devoid of both haptens. Lastly, some-sulf positive galCb-negative processes, as yet unidentified, were also found in the periaqueductal gray matter and in the nucleus interpeduncularis.

Presumptive common precursor for neuronal and glial cell lineages in mouse hypothalamus.

The cellular localization of a neuronal and a glial cell specific protein (14-3-2 and S-100, respectively) has been explored in mouse hypothalamus in order to trace cell lineages. This study was performed on fixed slices, at the light microscope level, by using either the indirect peroxidase-labeled immunoglobulin technique or immunofluorescence. In the adult, only S-100 immunoreactivity was found in the ependymal layer. In contrast, the magnocellular neurons of the preoptic area displayed strong 14-3-2 immunoreactivity. At neonatal stages (fetal day 17-postnatal day 3), both 14-3-2 and S-100 immunoreactivities developed simultaneously in the same cells lining the ventral part of the third ventricle. Transient detachment of some of these ventricular cells could be visualized before migration in the hypothalamus where they remained as bipotential cells up to postnatal day 10. Later in the development, they differentiated into separate cells, one type containing 14-3-2 and the other S-100, like neurons and glial cells. These results argue for a developmental stage during which cells lining the ventricle are bipotential and may thus be candidates for the role of stem cells for both neuronal and glial lineages.