Extracellular matrix of the superior olivary nuclei in the dog

Characterization of the superior olivary complex of Canis lupus domesticus

The superior olivary complex (SOC) is a collection of brainstem auditory nuclei which play essential roles in the localization of sound sources, temporal coding of vocalizations and descending modulation of the cochlea. Notwithstanding, the SOC nuclei vary considerably between species in accordance with the auditory needs of the animal. The canine SOC was subjected to anatomical and physiological examination nearly 50 years ago and was then virtually forgotten. Herein, we aimed to characterize the nuclei of the canine SOC using quantitative morphometrics, estimation of neuronal number, histochemistry for perineuronal nets and immunofluorescence for the calcium binding proteins calbindin and calretinin. We found the principal nuclei to be extremely well developed: the lateral superior olive (LSO) contained over 20,000 neurons and the medial superior olive (MSO) contained over 15,000 neurons. In nearly all non-chiropterian terrestrial mammals, the MSO exists as a thin, vertical column of neurons. The canine MSO was folded into a U-shaped contour and had associated with the ventromedial tip a small, round collection of neurons we termed the tail nucleus of the MSO. Further, we found evidence within the LSO, MSO and medial nucleus of the trapezoid body (MNTB) for significant morphological variations along the mediolateral or rostrocaudal axes. Finally, the majority of MNTB neurons were calbindin-immunopositive and associated with calretinin-immunopositive calyceal terminals. Together, these observations suggest the canine SOC complies with the basic plan of the mammalian SOC but possesses a number of unique anatomical features.

Unique features of extracellular matrix in the mouse medial nucleus of trapezoid body--implications for physiological functions

The medial nucleus of the trapezoid body (MNTB) is a vital structure of sound localization circuits in the auditory brainstem. Each principal cell of MNTB is contacted by a very large presynaptic glutamatergic terminal, the calyx of Held. The MNTB principal cells themselves are surrounded by extracellular matrix components forming prominent perineuronal nets (PNs). Throughout the CNS, PNs, which form lattice-like structures around the somata and proximal dendrites, are associated with distinct types of neurons. PNs are highly enriched in hyaluronan and chondroitin sulfate proteoglycans therefore providing a charged surface structure surrounding the cell body and proximal neurites of these neurons. The localization and composition of PNs have lead investigators to a number of hypotheses about their functions including: creating a specific extracellular ionic milieu around these neurons, stabilizing synapses, and influencing the outgrowth of axons. However, presently the precise functions of PNs are still quite unclear primarily due to the lack of an ideal experimental model system that is highly enriched in PNs and in which the synaptic transmission properties can be precisely measured. The MNTB principal cells could offer such a model, since they have been extensively characterized electrophysiologically. However, extracellular matrix (ECM) in these neurons has not yet been precisely detailed. The present study gives a detailed examination of the ECM organization and structural differences in PNs of the mouse MNTB. The different PN components and their distribution pattern are scrutinized throughout the MNTB. The data are complemented by electron microscopic investigations of the unique ultrastructural localization of PN-components and their interrelation with distinct pre- and postsynaptic MNTB cell structures. Therefore, we believe this work identifies the MNTB as an ideal system for studying PN function.

Distribution of perineuronal nets in the human superior olivary complex

Perineuronal nets (PNNs) are specialized assemblies of chondroitin sulfate proteoglycans (CSPGs) in the central nervous system that form a lattice-like covering over the cell body, primary dendrites and initial axon segment of select neuronal populations. PNNs appear to play significant roles in development of the central nervous system, neuronal protection, synaptic plasticity and local ion homeostasis. In seven human brainstems (average age=81 years), we have utilized Wisteria floribunda (WFA) histochemistry and immunocytochemistry for CSPG to map the distribution of PNNs within the nuclei of the human superior olivary complex (SOC). Within the SOC, the majority of net-bearing neurons are situated in the most medially situated nuclei, especially the superior paraolivary nucleus and medial nucleus of the trapezoid body. Net-bearing neurons are consistently found in the ventral nucleus of the trapezoid body and posterior periolivary nucleus, but to a lesser extent in the lateral nucleus of the trapezoid body. Finally, perineuronal nets are typically absent from the lateral and medial superior olives.

Immunolocalization of laminin-alpha1-like antigens around synapses in mouse cerebellar perineuronal nets

The hypothesis that extracellular matrix components may be related to neuronal development in the mouse cerebellar cortex was verified with immunohistochemistry by using an antibody against laminin-alpha1, a major extracellular matrix protein in various tissues. A commercially available polyclonal antibody, raised against the carboxyl-terminal 20-amino acid peptide of laminin-alpha1 was used. Some positive immunoreaction products were localized around large GABAergic interneurons in granular layers and others were around neurons in deep cerebellar nuclei. At the electron microscope level, diaminobenzidine immunoreaction products were localized around presynaptic boutons and in intercellular matrices around interneurons. Such immunoreaction products could be detected at postnatal day 20, when most of cerebellar synapses are assumed to be established. It has been known that a special feature of extracellular matrix, termed perineuronal nets, exists around specific subpopulation of neurons. In the mouse cerebellum, the present findings suggest that laminin itself or laminin-like-antigens exists in the perineuronal nets in relation to inhibitory neuron synapses.

Perineuronal nets in the rat medial nucleus of the trapezoid body surround neurons immunoreactive for various amino acids, calcium-binding proteins and the potassium channel subunit Kv3.1b

Perineuronal nets (PNs) are known as chondroitin sulfate-rich, lattice-like coatings of the extracellular matrix ensheathing mainly GABAergic, parvalbumin-containing neurons especially in the cerebral cortex. PNs have also been detected around GABA-immunonegative cells which were shown to be not aminergic, cholinergic, nitrinergic or peptidergic in various brain regions of some mammalian species. To find out whether glycine and aspartate may occur in net-bearing neurons the present study was focused on the rat medial nucleus of the trapezoid body (MNTB) which contains a large portion of cells immunoreactive for these amino acids, but appears to be devoid of GABA-immunoreactive cell bodies. PNs were detected around many glycine- and aspartate-immunopositive neurons in the MNTB by carbocyanine double labeling and confocal laser scanning microscopy. An additional finding was that the lectin-cytochemically stained extracellular matrix surrounds the calretinin-immunoreactive calyces of Held known as giant glutamatergic endbulbs which cover glycinergic principal cells in the MNTB. As elucidated by triple fluorescence labeling, the vast majority of somata co-expressed the calcium-binding proteins parvalbumin and calbindin, but not calretinin. The observed co-localization of PNs and immunoreactivity for the voltage-dependent potassium channel Kv3.1b - as an established marker of fast-firing parvalbumin-containing neurons - supports the assumed function of PNs as a cation exchanger ensuring rapid ion transport as required by highly active nerve cells.

Identification of Gal(beta 1-3)GalNAc bearing glycoproteins in cerebrospinal fluid of amyotrophic lateral sclerosis (ALS) patients

Glycoproteins in cerebrospinal fluid of 55 patients with amyotrophic lateral sclerosis (ALS), six disease controls (multifocal motor neuropathy, sensorimotor neuropathy, Guillain-Barré syndrome, spinal muscular atrophy type II, motor neuropathy with monoclonal gammopathy) and 20 healthy controls were separated by PAGE electrophoresis and then detected immunochemically with peanut agglutinin (PNA). In 36 amyotrophic lateral sclerosis patients the 262 kDa glycoprotein was significantly increased (over the normal mean +/- SD x 2), which was associated with a decrease in the 114 kDa fraction. In the remaining patients, both fractions were either equal in concentration or the 114 kDa glycoprotein predominated. In normal cerebrospinal fluid, the 114 kDa glycoprotein predominated over the other glycoproteins. The total amount of separated glycoproteins was increased in 15 amyotrophic lateral sclerosis patients. In 12 of them it was followed by an increase in the percentage of the 262 kDa glycoprotein. There was no correlation between the content of the peanut agglutinin-labelled glycoproteins and the patients' age, duration and severity of the disease. There was a correlation between the 262 kDa glycoprotein being increased in cerebrospinal fluid and the electrophysiological pattern of denervation seen in electromyographic study. The glycoproteins change, similar to that occurring in amyotrophic lateral sclerosis patients, was also observed in one case of multifocal motor neuropathy (MMN). We suggest that in amyotrophic lateral sclerosis and multifocal motor neuropathy, the peanut agglutinin-labelled glycoproteins are released in excess from the nervous tissues into the cerebrospinal fluid as a result of neuronal degeneration. The question to be answered is, whether the released glycoproteins are becoming targets for auto-antibodies.

Perineuronal nets: past and present

Golgi ranked the peripheral reticulum--which adheres intimately to nerve cell surfaces--alongside the intracellular reticulum, or Golgi apparatus,which immortalized his name. At first dismissed as an artefact of capricious staining techniques, this peripheral reticulum, or perineuronal net, is now recognized as a genuine entity in neurocytology. It represents a complex of extracellular matrix molecules interposed between the meshwork of glial processes, from which they are indistinguishable, and nerve-cell surfaces. In no other branch of neuroscience has the waxing and waning of interest in any morphological entity been so pronounced as in the case of the perineuronal net. This review traces the history of this enigmatic structure from its conception to the present time, brings to light the keen observational powers of morphologists at the turn of the century and reveals how their sagacious forethought anticipated current thinking on the role of perineuronal nets.

Lectin staining in the basal nucleus (Meynert) and the hypothalamic tuberomamillary nucleus of the developing human prosencephalon

Previous studies have demonstrated that extracellular matrix glycoconjugates, shown by lectin-histochemistry with Vicia villosa agglutinin (VVA) and peanut agglutinin (PNA) as so-called perineuronal nets, play an important role in brain maturation. Concanavalin A (ConA) binding to neuronal surface glycoconjugates may be a marker of synaptic junctions. The present study was done to demonstrate the binding sites of these lectins in two functionally related nuclei of the prosencephalon, the basal nucleus (Meynert) and the hypothalamic tuberomamillary nucleus. Fetal brains of 16-36 weeks of gestation were examined by using VVA, PNA, and ConA to determine appearance and distribution patterns of specific lectin-binding sites on glycoconjugates during fetal brain development. The basal nucleus and the tuberomamillary nucleus showed a characteristic "cellular staining" that may have been due to cytoplasmatic labeling, surface labeling, or both. Lectin-staining occurred much earlier in the basal nucleus than in the tuberomamillary nucleus. Although all three lectins were bound to neurons of the basal nucleus, only ConA-positive neurons were observed in the tuberomamillary nucleus. In conclusion, lectin-labeled cells most probably represent projection neurons that are GABAergic (tuberomamillary nucleus) or cholinergic (basal nucleus). Labeling with the three lectins demonstrated nuclear-specific staining patterns that occur early in fetal development and gradually increase. Binding sites for lectins characterizing perineuronal nets (VVA, PNA) occurred only in the basal nucleus, whereas binding sites for ConA on neuronal-surface glycoconjugates, which seem to play a role in early synaptogenesis, were present in the basal and the tuberomamillary nucleus. The basal nucleus, however, expressed ConA binding sites distinctly earlier, probably indicating early arriving afferents.

Immunohistochemical localization of neurocan in the lower auditory nuclei of the dog

Chondroitin sulfate proteoglycans are present at high levels in the lower auditory system of mammals. Axon terminals on the principal neurons in the superior olivary nuclei contain chondroitin 4- and 6-sulfate, while the broad extracellular matrix around axon terminals contains chondroitin sulfate D, a highly sulfated chondroitin sulfate rich in the disaccharide unit of GlcA(2S)beta1 --> 3GalNAc(6S), in the dog. In the present study, we investigated the immunohistochemical staining of neurocan, a brain-specific proteoglycan, in the lower auditory tract of the dog, including an analysis by immunoelectron microscopy. Immunolocalization of neurocan was conspicuous in the medial and lateral superior olivary nuclei and much less intense immunostaining was seen in the cochlear nucleus and posterior colliculus. No immunoreactivity were found in other nuclei. The immunostaining in the medial and lateral superior olivary nuclei was observed as perineuronal nets around large principal neurons at the light-microscopic level, while no immunostaining was observed in the upper segment of the medial superior olivary nucleus and the medial segment of the lateral superior olivary nucleus, in which medium-sized and small neurons were located. Immunoelectron microscopy revealed the reaction products of immunostaining on cell membranes of the perikarya of principal neurons and on cell membranes of presynaptic terminals which made axo-somatic synapses on the principal cells. No immunoreactivity was detected at synaptic junctions, in the extracellular matrix or within axon terminals. In the cochlear nucleus, immunoreactive perineuronal nets were found around a small number of neurons and immunoreactive nerve fibers were scattered in the anterior ventral cochlear nucleus. In the posterior colliculus, perineuronal nets, which were weakly immunostained, were sparsely distributed in the central nucleus. These results suggest that different locations of chondroitin sulfate proteoglycans, including neurocan, may be associated with focal sites composed of neuronal surface, terminal boutons and extracellular matrix in the lower auditory tract of the adult dog.

The presence of chondroitin sulfate A and C within axon terminals in the superior olivary nuclei of the adult dog

Localization of chondroitin sulfate A and C [GlcA beta 1--> 3GalNAc(4S) and GlcA beta 1--> GalNAc(6S)] has been determined in the medial and lateral superior olivary nuclei of the adult dog by light and electron microscopic immunocytochemistry. Neuropil around large neurons were heavily immunostained, whereas neuropil around small neurons showed only weak or no immunoreactivity. Electron microscopy revealed that presence of chondroitin sulfate A and C proteoglycan in axon terminals around neuronal cell bodies.

SPARC, an extracellular matrix glycoprotein containing the follistatin module, is expressed by astrocytes in synaptic enriched regions of the adult brain

Although extracellular matrix (ECM) glycoproteins play important roles in neural development, their levels are generally believed to decrease in the adult brain. Immunohistochemical analysis indicates that the anti-adhesive ECM glycoprotein SPARC/osteonectin, which contains a follistatin 'module,' is expressed in the adult rabbit nervous system. In the cerebellum, SPARC is present in Bergmann glia, with a strong signal along their radial fibres. SPARC, while enriched in membrane fractions, is not a transmembrane protein. In the hippocampus, colocalization of SPARC is observed in cells which express the astrocytic marker GFAP. The expression of SPARC by a subset of astrocytes, particularly in synaptic enriched areas, suggests a continuing role for the ECM in the adult brain.

Identification of Gal(beta 1-3)GalNAc bearing glycoproteins at the nodes of Ranvier in peripheral nerve

A subset of human anti-GM1 ganglioside antibodies cross-reacts with Gal(beta 1-3)GalNAc bearing glycoproteins in peripheral nerve and spinal cord. The same oligosaccharide determinant is recognized by the lectin peanut agglutinin (PNA) which binds at the nodes of Ranvier in intact peripheral nerve. The Gal(beta 1-3)GalNAc bearing glycoproteins were isolated using PNA lectin affinity chromatography followed by separation on Western blot, and the proteins were subjected to partial amino acid sequence analysis. Two major PNA binding glycoproteins were identified in peripheral nerve and spinal cord; one had an approximate molecular weight of 120 kD and had sequence homology to the oligodendrocyte-myelin glycoprotein (OMgp). The other migrated between 70 and 80 kD and had sequence homology to the hyaluronate binding domain of versican, which has been reported to share sequence homology with the 70 kD proteins hyaluronectin and the glial hyaluronic acid binding protein (GHAP). By immunocytochemistry, OMgp was localized to the paranodal region of myelin, and the protein homologous to the hyaluronate binding domain of versican was localized to the nodal gap in peripheral nerve. These PNA binding glycoproteins might be target antigens for autoantibodies in peripheral nerve.

Perineuronal nets--a specialized form of extracellular matrix in the adult nervous system

One century ago, Camillo Golgi described 'perineuronal nets' enwrapping the cell bodies and proximal dendrites of certain neurons in the adult mammalian central nervous system and suggested that they represent a supportive and protective scaffolding. Although other neuroanatomists validated the existence of these nets on selected neurons in the adult brain, there was a lack of agreement on their origins, composition and function. The application of modern molecular and ultrastructural methods has brought new insights and a renewed interest in these classic observations. Recent data suggest that perineuronal nets result from the visualization of extracellular matrix molecules that are confined to the space interposed between glial processes and the nerve cells that they outline. The material confined to these spaces can be visualized selectively by antibodies directed to glycoproteins (e.g., tenascin and restrictin/janusin), proteoglycans (e.g., chondroitin sulfates), markers for hyaluronan as well as by lectins recognizing N-acetylgalactosamine and by monoclonal antibodies directed to epitopes on unknown molecules (e.g., HNK-1, VC1.1 and Cat 301). This review examines the emerging clarification of classical observations of perineuronal nets and the functional implications suggested by their molecular composition. Also discussed are studies that further extend observations on the time of development and of the specificity in the occurrence of perineuronal nets. In the adult brain the molecules constituting the 'perineuronal nets of matrix' could serve as recognition molecules between certain neurons and their surrounding cells and participate in the selection and consolidation of their relationship.

Dynamics of Golgi impregnation in neurons

This paper describes the early stages of impregnation by the Golgi rapid method in sections and blocks of brain tissue. Aldehyde-fixed and potassium dichromate-treated sections of cerebral cortex were placed on glass slides and coverslipped. The dichromate solution was then replaced by a silver nitrate solution, and events taking place in the section were monitored and time-lapse recorded until the impregnation was interrupted and the sections subsequently prepared for electron microscopy. The tissue blocks, fixed and chromated in the same way, were placed into a silver nitrate solution for 30 min to 24 h and the progress of impregnation compared with the results obtained in the sections on the glass slides. Two basic modes of impregnation were observed, apparently in direct relation to the process of crystallization of silver chromate: crystals of silver chromate growing directly from the surface of the tissue into the nerve cell via its transected plasma membranes, and microcrystalline precipitate of silver chromate spreading into the nerve cell from nucleation centres dispersed in the tissue. The precipitate grows inside the cell as in a preformed channel until the cell has been filled. If the nucleation begins extracellularly, the precipitate extends into the narrow intercellular gaps. Electron microscopy showed that the crystalline precipitate consisted of multilamellar formations containing dense coalesced granules that did not cross plasma or endocellular membrane boundaries.

Ultrastructural development of the medial superior olive (MSO) in the ferret

When ferrets are born, four weeks before the onset of hearing, few synapses are evident in the medial superior olive (MSO). The synapses present are immature and almost exclusively found in the neuropil. The MSO somata are virtually devoid of synaptic contacts but are contacted by fine glial processes that increasingly ensheathe the somata during the first postnatal week. By P12, somatic synaptogenesis in the MSO is evident. Initially the terminals contain vesicles of irregular shape, size, and distribution. The glial lamellae appear to withdraw as the synaptic contacts form but continue to cover the asynaptic portions of the cell surface. The lamellae frequently extend from ensheathing the soma to encapsulate the immature terminals. During the next two weeks, synaptic density and terminal encapsulation proceed until the somata is surrounded by encapsulated synaptic terminals as in the adult ferret MSO. While most immature terminals contain round vesicles, during the first postnatal week some terminals with nonround vesicles can be distinguished. The first distinction between types of nonround vesicle-containing terminals, i.e., pleiomorphic and ovoid, is in the second postnatal week. This distinction becomes increasingly clear and by the end of the first postnatal month, terminal types can be reliably categorized. These observations indicate that: (1) synapses are present in the MSO neuropil one month prior to the onset of hearing, (2) the major period of synaptogenesis begins approximately two weeks prior to the onset of hearing, and (3) glial lamellae ensheathe MSO somata prior to the onset of somatic synaptogenesis, withdraw as synapses form, and subsequently re-extend to encapsulate newly formed synapses.

Patterns of reactivity of human anti-GM1 antibodies with spinal cord and motor neurons

Human anti-GM1 antibodies from patients with lower motor neuron disease or predominantly motor neuropathy recognize carbohydrate determinants shared by GM1 and related glycolipids and glycoproteins, but the identity of the antigens to which they bind in tissue is unknown. We examined the binding of anti-GM1 antibodies with differing fine specificities to spinal cord, isolated motor neurons, and dorsal root ganglia neurons in order to characterize the tissue antigens. All anti-GM1 antibodies tested bound to the surface of bovine spinal motor neurons and immunostained the gray matter of unfixed sections of spinal cord. The staining was blocked by cholera toxin, which is specific for GM1, indicating that GM1 itself was the target antigen. Binding to white matter was more variable and depended on fixation and the fine specificities of the antibodies. The anti-GM1 antibodies did not bind to dorsal root ganglia neurons in tissue sections or in culture. These studies suggest that the autoantibodies might exert their effect, in part, by binding to GM1 on the surface of motor neurons, and that the absence of binding to dorsal root ganglia neurons might explain the lack of sensory abnormalities in affected patients.

Chondroitin sulfate in the extracellular matrix of the medial and lateral superior olivary nuclei in the dog

Chondroitin sulfate was examined in the extracellular matrix of the canine medial and lateral superior olivary nuclei by light and electron microscopic immunocytochemistry. The extracellular matrix around the large neurons was intensely stained with a monoclonal antibody recognizing D-glucuronic acid 2-sulfate----N-acetylgalactosamine 6-sulfate (D-unit) and this staining degree was remarkably reduced after chondroitinase ABC digestion. Neuronal cytoplasm, glial cells or capillaries in these nuclei were not stained with the monoclonal antibody. The results indicate the presence of disaccharide residue of D-glucuronic acid 2-sulfate----N-acetylgalactosamine 6-sulfate in the chondroitin sulfate proteoglycan of the extracellular matrix.

Identification of glycoconjugates which are targets for anti-Gal(beta 1-3)GalNAc autoantibodies in spinal motor neurons

Human IgM anti-Gal(beta 1-3)GalNAc antibodies which bind to GM1 and GD1b, are implicated in the pathogenesis of predominantly motor neuropathy or motor neuron disease. By immunofluorescence microscopy, the human antibodies immunostain the surface of motor neurons from bovine spinal cord. The motor neurons are also immunostained by cholera toxin (CT), which is specific for GM1. Glycolipid analysis using thin-layer chromatography (TLC) and immunostaining reveals that the relative concentration of GM1 and GD1b in motor neurons is greatly reduced in comparison to whole spinal cord, and that other motor neuron gangliosides are unreactive with the anti-Gal(beta 1-3)GalNAc antibodies. By Western blot analysis, the antibodies react with several protein bands in motor neuron extracts, and many of the same proteins are also recognized by PNA. These data suggest that both glycoproteins and glycolipids might be targets for anti-Gal(beta 1-3)GalNAc antibodies in spinal motor neurons.