Alteration of mouse cerebellar circuits following methylazoxymethanol treatment during development: immunohistochemistry of GABAergic elements and electron microscopic study.

Methylazoxymethanol (MAM) injected postnatally affects cerebellar development in mice. A single injection at the fifth postnatal day produces hypogranular cerebella whereas a single injection at birth produces, in addition, a disorderly cytoarchitecture of the folium and alteration of Purkinje cell positioning (Bejar et al.: Exp. Brain Res. 57:279-285, '85). In the present study we have used immunohistochemistry with anti-GABA immune serum and electron microscopy to further characterize these alterations. In addition to the already-described nonoccupied dendritic spines of Purkinje cells both in mice injected the day of birth and or at the fifth postnatal day, we have observed, in animals injected at birth, the absence of pericellular baskets around Purkinje cells and the presence of heterologous synapses between mossy fibres and Purkinje cell dendrites. These heterologous synapses apparently disappear after postnatal day 20. By using an appropriate timing of MAM injection, different types of hypogranular cerebella, phenocopies of different mutants, can be obtained in large enough number to carry out extensive biochemical studies at each developmental age.

Radial glia and astrocytes in developing and adult telencephalon of the lizard Gallotia galloti as revealed by immunohistochemistry with anti-GFAP and anti-vimentin antibodies.

The development of radial glia and astrocytes in the telencephalon of the lizard Gallotia galloti was studied by immunohistochemistry with anti-vimentin and anti-GFAP antibodies. Vimentin appears at embryonic stage 32 (E32) in the proliferative zone of the lateral ventricle and subpial end-feet in the marginal zone. At E34-35 the staining intensity for vimentin in all radial glia is maximal. It then decreases and disappears in most structures in adult animals. GFAP appears at E35 in the end-feet in the marginal zone and its intensity increases until adulthood, particularly in radial and sinuous fibers and in fibers that originate from the sulci and invade the ventral striatum and the septum. In contrast, the reaction is weak in the cortex, in the anterior dorso-ventricular ridge, and in the amygdala nuclei. Radial glia is still present in the adult, and the composition of its intermediate filaments changes during development from vimentin to GFAP. No GFA-positive cell bodies except those of ependymal glia were detected in telencephalon.

Glial fibrillary acidic protein and vimentin immunohistochemistry in the developing and adult midbrain of the lizard Gallotia galloti.

The distribution of glial fibrillary acidic protein (GFAP)- and vimentin-containing cells was studied by immunohistochemistry in the midbrain of the lizard Gallotia galloti. At embryonic stage 32 (E32), vimentin immunoreactivity appeared first in cell bodies located in the ventricular walls, in radial fibers, and subpial end-feet and increased in these structures until E34/E35. Faint GFAP immunoreactivity gradually appeared in the same structures between E34 and E37, and this increased until adulthood, whereas vimentin immunoreactivity decreased after E35, becoming limited to a few end-feet and fibers in the adult, mainly in the tegmentum. Thus, in developing Gallotia midbrain a shift from vimentin-containing to GFAP-containing intermediate filaments begins around E36 or E37. At E40, in addition to the cell bodies in the ependymal area, dispersed GFAP-positive cells, possibly immature astrocytes appeared. These cells showed the same shift. In the adult lizard, GFAP-positive radial glia are still present and coexist with GFAP-positive astrocytes, which are prefentially located in the marginal optic tract and the oculomotor nuclei, but are absent in the fasciculus longitudinalis medialis. Optic tectum, pretectum, tegmentum, and isthmic nuclei are the areas richest in GFAP-positive radial fibers: these were much less abundant in the deep mesencephalic nuclei. Thus, in this lizard, GFAP-positive astrocytes display a clear cut regional distribution: they are present in mesencephalon, whereas they are absent in telencephalon.

Immunoelectron microscopy of alpha 2-glycoprotein. An astrocyte-specific antigen.

The cellular and fine structural localization of the soluble brain-specific acidic alpha 2-glycoprotein was investigated using the indirect immunohistochemical method. The electron microscope was used to unambiguously identify cells containing the antigen. A single type of cell, the astrocyte, was found to be labelled with specific antisera directed against alpha 2-glycoprotein. Immunoperoxidase reaction product was found in astrocyte perikarya, their processes and perivascular end feet. It was found to be apparently associated with the cytoplasmic surface of mitochondria, reticular membranes and the plasma membrane. No specific labelling of neurones, oligodendrocytes, myelin or capillary endothelial cells was observed. The data is discussed in relation to the immunological properties of alpha 2-glycoprotein already reported.

A monoclonal antibody recognizing subpopulations of neurones in mouse brain.

A monoclonal antibody, designated C138, was raised against a particulate fraction from young postnatal mouse cerebellum. In sections from adult mouse brain and cerebellum, this antibody stained a sub-population of neurons. Cell bodies of large neurones and fibre tracts were negative in all brain regions examined, whereas neuropil was heavily labelled. Immune-electron microscopy confirmed the specificity of the antibody for certain types of neurones and indicated that the antigen may be associated with microtubular structures. Immunofluorescence studies on dissociated cerebellar cultures showed that the antigen was expressed inside all tetanus toxin-positive neurones but not by non-neuronal cells.

Autoradiographic characterization of [3H]L-glutamate binding sites in developing mouse cerebellar cortex.

Postnatal changes of [3H]L-glutamate binding sites in mouse cerebellum were studied by in vitro autoradiography. These sites were already present at birth, their density globally increased until postnatal day 25, and at all ages it was higher when Cl- and Ca2+ were present in the incubation buffer. At birth, these binding sites were diffused through the whole cerebellar mass, but became distinctly concentrated in the molecular and the internal granular layers by postnatal day 10. From this age on, binding site sensitivity to ions and glutamate analogues takes a different course in each layer. The external granular layer and the white matter never displayed significant amounts of binding. In the molecular layer the Cl-/Ca2+ effect increased during ontogeny until, in adults, the ion-dependent binding was threefold higher than the ion-independent binding. Quisqualate-sensitive sites accounted for 80% of the total binding sites already at postnatal day 15, while displacement by alpha-amino-3-hydroxy-methyl-4-isoxazolepropionic and ibotenic acids attained the maximum (68%) at postnatal day 60. N-Methyl-D-aspartate displaced glutamate binding (50%) only in the presence of Cl- and Ca2+. Starting from postnatal day 15, binding site density in the molecular layer of lobules VIb and VII of the vermis was lower than in other lobules. In the internal granular layer, the Cl-/Ca2+ effect observed in young animals decreased during development. These transient binding sites were sensitive to quisqualic and ibotenic acid. In adults, the majority of glutamate binding sites were ion-independent and mainly sensitive to D,L-amino-5-phospho-valeric acid and N-methyl-D-aspartate. Throughout development and in both layers, sites displaced by kainate were present at low density and sites displaced by D,L-2-amino-4-phosphonobutyric acid were not detected. The localized postnatal changes of the [3H]L-glutamate binding sites were correlated with the events occurring during growth and maturation of cerebellar structures. The increase of the Cl-/Ca(2+)-dependent binding in the molecular layer is simultaneous with the growth of Purkinje cell dendrites and of parallel fibres and with the formation of the synapses between them. This suggests that these binding sites are localized in these synapses. The changing pattern of sensitivity to different agonists during development might correspond to the maturation of these synapses. The low density of [3H]L-glutamate binding in the molecular layer of lobules VIb and VII probably indicates the presence of specific nerve projections to these areas.(ABSTRACT TRUNCATED AT 400 WORDS)

L-glutamate and L-glutamine uptake in adult rat cerebellum: an autoradiographic study.

The compartmentation of L-glutamate in the central nervous system has been extensively studied and L-glutamine is believed to be the precursor of the neuronal releasable pool of the L-glutamate. In order to localize the sites of uptake of both L-glutamate and L-glutamine, autoradiography was used in tissue slices of adult rat cerebellum, where granule cells are considered to be glutamatergic. Incubation of the tissue with low concentrations of [3H]L-glutamate or [3H]L-glutamine produces in both cases a heavy labelling of the molecular layer. [3H]L-glutamate uptake seems to be essentially glial (Golgi epithelial cells and Bergmann fibres) while [3H]L-glutamine is more diffusely distributed over the molecular layer. Although no conclusions can be drawn on the nature of L-glutamine uptake, these results are in agreement with the model which considers L-glutamate uptake by glial cells to be the inactivating process of glutamatergic synapses.

A developmentally modified neurone-specific marker in rodent cerebellum.

Out of several monoclonal antibodies secreted by hybridomas resulting from the fusion of a mouse myeloma cell line with spleen cells from mice immunized with cerebellar membranes from 12 day old rats, one, called 11.9, produced an unusual immunolabelling pattern when tested on sections of rat cerebellum. The cerebellar distribution of the antigenic sites recognized by this antibody using an immunoperoxidase technique at the optical and ultrastructural levels is described in detail in this report. The immunoreaction product was found in the adult rat to be associated with the microtubules and the zone immediately beneath the plasma membrane of parallel fibres. In young animals the density of immunostaining appears to be higher than in the adult, and the staining is detectable in addition in the perikaryal cytoplasm of granule cells. Biochemical studies using the Western immunoblot technique demonstrate that the antigens consist of two polypeptides of molecular weights 120 and 185 kD. The possible relation of the antigens to cytoskeletal structures is discussed and the labelling pattern is compared with that produced by other known monoclonal antibodies.

Cellular localization of the brain specific alpha 2-glycoprotein in rat cerebellum: an immunohistological study.

The cellular localization of the brain-specific, soluble, acidic alpha 2-glycoprotein was studied in rat cerebellum by using the immunoperoxidase technique at the light-and electron-microscopy levels with monospecific immune serum directed against this glycoprotein. Only astrocytes, their processes, and their end feet (subpial or perivascular) contained heavy immunoperoxidase reaction product. Cerebellar neurones, oligodendrocytes, myelin and blood vessel endothelia did not stain. Thus it appears that alpha 2-glycoprotein is an astrocyte marker.

A new brain cell surface glycoprotein identified by monoclonal antibody.

Of 207 monoclonal antibodies produced against cultured mouse cerebellar cells, 16 reacted with cerebellar cell surfaces and 4 reacted with glycoproteins. One of them, called an anti-BSP-3 (Brain cell Surface Protein-3) defines a 48,000 molecular weight protein which can be iodinated at the surface of cultured cerebellar cells. Lectin-binding and sugar incorporation studies established the glycoprotein nature of the antigen. Astroglia (glial fibrillary acidic protein-positive cells) in primary cerebellar cultures were labelled intensely for this antigen by the indirect immunofluorescence method while neuronal cells and their processes were more weakly labelled. Fibronectin-positive cells were negative for BSP-3. In cerebellar sections using the immunoperoxidase method at both the optical and electron microscope levels, the difference in staining intensity between astrocytes and neuronal cells was not significant: in Purkinje cells and in the large neurones present in the deep cerebellar nuclei the immunoperoxidase percipitate was confined to the plasma, membrane while in both astrocytes and granule cells cytoplasmic labelling was also observed. Oligodendrocytes do not appear to react with the anti-BSP-3 monoclonal antibody; neither do endothelial or leptomeningeal cells. The availability of a monoclonal antibody produced by a stable hybridoma line will be a powerful tool in attempts to purify the BSP-3 antigen and to elucidate its function.

Double labeling immunohistochemical technique provides evidence of the specificity of glial cell markers.

Sequential treatment of rat cerebellar slices with two antisera, each specific for a different glial cell marker, revealed, by indirect immunofluroescence combined with the immunoperoxidase method on the same tissue section, that carbonic anhydrase isoenzyme II (CA II) is exclusively localized in oligodendrocytes while the glial fibrillary acidic protein (GFA) is, as well established, a specific marker for astrocytes.

A surface marker for murine vascular endothelial cells defined by monoclonal antibody.

Spleen cells from a rat immunized with mouse cerebellar cells were fused with mouse myeloma cells. One of the hybridomas secreted a monoclonal antibody that reacts with a surface antigen on vascular endothelial cells. The antibody stained endothelial cells lining blood vessels in brain, heart, lung, kidney, and liver. It did not, however, stain endothelial cells lining hepatic sinusoids. Parenchymal cells were always negative. So far, an antigen of similar tissue distribution has not been described in the mouse and we have called it mouse endothelial surface antigen-1 (MESA-1). The antibody could be used as a highly specific usefulness for identifying endothelium-derived cells in culture has been demonstrated on cultures of dissociated mouse cerebellum, where it stained a subclass of fibronectin-expressing cells.

Autoradiographic localization of [3H]-L-glutamate binding sites in a model of cerebellar granule cell ectopia generated by methylazoxymethanol treatment.

The distribution of [3H]glutamate binding sites was studied in a model of altered cerebellar development obtained by injecting methylazoxymethanol (MAM) in 5-day-old mice. In these mice, at the 25th postnatal day, cerebella were smaller than normal, stratification was normal except for the presence in some lobes of a thin ectopic granule cell layer in the middle of the molecular layer, the proportion of the distribution of [3H]glutamate binding sites between molecular and internal granule cell layers was maintained but site density of both quisqualate- and NMDA-sensitive types was increased in the two layers. In the molecular layer, this increase was uniform in spite of the presence of the ectopic cell layer. In the internal granular layer, the increase of quisqualate-sensitive and NMDA-sensitive [3H]glutamate binding sites is topographically segregated and the first corresponds to areas of lesser cellular density. These results show that MAM treatment induces persistent alterations of the cerebellar glutamatergic system, which consist of receptor over-expression, possibly due to deficit of innervation, reactive gliosis and immaturity of surviving granule cells.

Changes of pharmacological properties of (1S,3R)-ACPD-sensitive glutamate binding sites in developing mouse cerebellum.

Autoradiography of [3H]glutamate binding to mouse cerebellar sections was used to study the distribution of (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid-((1S,3R)-ACPD) sensitive [3H]glutamate binding sites and the sensitivity of these sites to drugs preferentially acting on one or few types of the metabotropic receptor family. The inhibitory effect of (1S,3R)-ACPD on [3H]glutamate binding and its estimated inhibition constant showed the presence of a different global metabotropic receptor population according to the region considered. During ontogeny, the (1S,3R)-ACPD binding site density increased in the molecular layer (ML), in contrast it decreased in the internal granular layer (IGL) and the deep cerebellar nuclei (DCN). In addition, different sensitivities to (RS)-alpha-methyl-4-carboxyphenylglycine (MCPG), (S)-4-carboxyphenylglycine (4-CPG), (2S,3S,4S)-alpha-(carboxycyclopropyl)glycine (L-CCG-I) and L-2-amino-4-phosphonobutyric acid (L-AP4) were demonstrated according to the region and the age. In the DCN, the high (1S,3R)-ACPD binding site density at PND 10 seems to be also sensitive to L-CCG-I but not to MCPG, 4-CPG or L-AP4. In the ML, the MCPG-, the 4-CPG- and the L-AP4-sensitive [3H]glutamate binding sites appeared during ontogeny and the L-CCG-I-sensitive [3H]glutamate binding sites were already present at PND 10. In the IGL, L-CCG-I-sensitive binding sites disappeared in contrast to the L-AP4-sensitive binding sites which appeared during development even if the total (1S,3R)-ACPD binding site density was relatively weak in the adults. These results all reflected the multiplicity of the receptor subtypes included in the cerebellar metabotropic component.

Developmental changes of EAA metabotropic receptor activity in rat cerebellum.

The potency but not the efficacy of t-ACPD stimulation of phosphatydil inositide hydrolysis changes in developing rat cerebellum. This suggests that the excitatory amino-acid-stimulated metabotropic receptors and/or their coupling are ontogenically regulated. In this, cerebellum differs from other CNS regions where only efficacy changes were described. Differently from hippocampus, the t-ACPD effect, at all ages, is independent of the activation of the NMDA receptor.

Stoichiometry of muscimol and benzodiazepine binding sites in developing mouse cerebellum.

The ontogeny of high affinity GABAA and central benzodiazepine receptors in the mouse cerebellum was investigated by measuring [3H]muscimol and [3H]flunitrazepam binding to membrane preparations during postnatal development. In the P2 fraction, [3H]muscimol binding was much more abundant than [3H]flunitrazepam binding at all ages. [3H]muscimol Bmax exhibited a peak around postnatal day 25 while [3H]flunitrazepam binding did not follow a parallel course. These results can be explained by the preferential presence in cerebellum of certain variants of the different subunits of the GABAA receptor complex and with different topographical distributions of the different receptor subtypes. Development dependent changes of organelle distribution during subcellular fractionation also contributed to the described developmental pattern.

Ectopic granule cells in mouse cerebellum molecular layer express high affinity GABAA receptor.

The antimitotic/mutagenic agent methylazoxymethanol when injected in mice pups at postnatal days 3 and 4 produces hypogranular adult cerebella with subpial cells clusters and with a supplementary, ectopic granule cells layer in the molecular layer. The internal granular layer of these treated mice displayed a much lower density of [3H]muscimol binding sites than in controls. At the same time these binding sites are expressed in the ectopic granule cells in the middle of the molecular layer, in spite of the unusual localization of the cells and of the alteration of their nerve inputs. The subpial cells do not express these sites. Our autoradiographic data confirm the suggestion that during ontogeny the GABAA receptors in the granule cells appear when these cells leave the subpial region and indicate that the expressed receptor subtype is the same whether granule cells are in the internal granular layer or in the middle of the molecular layer.

An improved method for the preparation of rat cerebellar glomeruli.

Cerebellar glomeruli consist of large portions of the mossy fiber giant terminal, granule cell dendrites and Golgi neuron terminals. By modifying previously reported procedures we have developed a new method for bulk preparation of this polysynaptic complex from rat cerebellum. We obtained well preserved isolated glomeruli of satisfactory purity and homogeneity as indicated by electron microscopy and by determination of appropriate biochemical markers. The method is fast and simple, and it provides a glomerular fraction suitable for investigation of neurotransmitter receptors.

Synaptosomal plasma membrane glycoproteins: fractionation by affinity chromatography on concanavalin A.

Synaptosomal plasma membrane glycoproteins were solubilized in 0.08% sodium dodecyl sulfate (SDS) and separated by affinity chromatography on concanavalin A-Sepharose. Three fractions were obtained. Fraction CO (unadsorbed) contained 63% of the protein, but only 23% of the sugar and was rich in fucose, galactose and N-acetyl-neuraminic acid (NANA) relative to the other sugars. Many proteins were detected in this fraction by polyacrylamide gel electrophoresis, but only on band stained well for carbohydrate. Fraction CR (retarded) contained glycoproteins which reacted weakly with concanavalin A and stained poorly with periodic acid-Schiff reagent (PAS). There was no enrichment in total sugar/mg protein relative to the original fraction, but there was a marked enrichment in N-acetyl-glucosamine and NANA relative to the other sugars. The protein profile of this faction was complex, but only one major PAS-positive band was detected. However, most, if not all, the proteins seemed to be weakly PAS-positive. Fraction C1 (adsorbed) was markedly enriched (5-fold) in sugar/mg protein, particularly in mannose and N-acetyl-glucosamine. It had a relatively simple protein profile, and most of the protwin bands stained well with the PAS reaction. Glucose was detected in the initial fraction, and in all the subfractions, but it could not be shown definitely to be either a contaminant or an intrinsic constituent of synaptosomal plasma membrane (SPM) glycoproteins. Minor sugars, if present, could, at most, account for less than 0.25% of the carbohydrate.

Synaptosomal plasma membrane glycoproteins. II. Isolation of fucosyl-glycoproteins by affinity chromatography on the Ulex europeus lectin specific for L-fucose.

After solubilization in sodium dodecyl sulphate, almost 90% of synaptosomal plasma membrane glycoproteins were separated from the bulk of synaptosomal plasma membrane proteins by sequential affinity chromatography on two immobilized lectins: concanavalin A and the Ulex europeus lectin specific for L-fucose. Four fractions were obtained and their sugar composition and electrophoretic patterns were determined. Fucosyl-glycoproteins contain more than 26% of protein and 85% of the protein-bound sugar of synaptosomal plasma membrane; hence they constitute a major class of glycoproteins in these membranes. The presence of some glucose in glycoproteins fractions obtained after affinity chromatography on the two lectins suggests that this sugar could be a structural component of some brain glycoproteins. Polyacrylamide gel electrophoresis revealed at least 28 major bands in fucosylglycoprotein fractions, and 11 in other fractions. Several of these major bands appear to contain more than one glycoprotein each. This heterogeneity appears to be mostly the result of the heterogeneity of the neuronal population in the central nervous system. Microheterogeneity of glycoprotein sugar chains and possible contamination of synaptosomal plasma membranes play, in our opinion, only a minor role.