Characterization of proteoglycan-containing perineuronal nets by enzymatic treatments of rat brain sections

The role of chondroitin sulfate proteoglycans in regeneration and plasticity in the central nervous system

Chondroitin sulfate proteoglycans (CSPGs) consist of a core protein and glycosaminoglycan (GAG) chains. There is enormous structural diversity among CSPGs due to variation in the core protein, the number of GAG chains and the extent and position of sulfation. Most CSPGs are secreted from cells and participate in the formation of the extracellular matrix (ECM). CSPGs are able to interact with various growth-active molecules and this may be important in their mechanism of action. In the normal central nervous system (CNS), CSPGs have a role in development and plasticity during postnatal development and in the adult. Plasticity is greatest in the young, especially during critical periods. CSPGs are crucial components of perineuronal nets (PNNs). PNNs have a role in closure of the critical period and digestion of PNNs allows their re-opening. In the adult, CSPGs play a part in learning and memory and the hypothalamo-neurohypophysial system. CSPGs have an important role in CNS injuries and diseases. After CNS injury, CSPGs are the major inhibitory component of the glial scar. Removal of CSPGs improves axonal regeneration and functional recovery. CSPGs may also be involved in the pathological processes in diseases such as epilepsy, stroke and Alzheimer's disease. Several possible methods of manipulating CSPGs in the CNS have recently been identified. The development of methods to remove CSPGs has considerable therapeutic potential in a number of CNS disorders.

Organization of brain extracellular matrix in the Chilean fat-tailed mouse opossum Thylamys elegans (Waterhouse, 1839)

We investigated the structural and molecular organization of the extracellular matrix in Thylamys elegans, a marsupial representative of the mammalian order Didelphimorphia. Perineuronal nets (PNs) associated with distinct types of neurons were visualized by detection of chondroitin sulfate proteoglycans and hyaluronan, and by labeling with Wisteria floribunda agglutinin (WFA), a marker for PNs in the mammalian brain. In the neocortex of Thylamys, these methods revealed PNs on pyramidal cells. In contrast, parvalbumin-immunoreactive interneurons in the neocortex and hippocampal formation (displaying robust, WFA-labeled PNs in placental mammals) were ensheathed only with a delicate rim of hyaluronan and proteoglycans not detectable with WFA. The absence of WFA staining was characteristic also of some subcortical regions which contained PNs intensely labeled for chondroitin sulfate proteoglycan and hyaluronan. However, corresponding to placental mammals, numerous subcortical nuclei showed clearly WFA-stained PNs. Similar as in placental mammals, cholinergic basal forebrain neurons and tyrosine hydroxylase-immunoreactive neurons of the substantia nigra and locus coeruleus were devoid of PNs. Together with our earlier study on Monodelphis, the present results reveal that South American opossums show either a particular "marsupial" or "Didelphid" type of extracellular matrix chemoarchitecture, supporting the view that these components may vary phylogenetically as integral parts of neuronal physiology at the systems and single cell level.

Distribution of pyramidal cells associated with perineuronal nets in the neocortex of rat

Perineuronal nets are lattice-like accumulations of extracellular matrix components around the cell body and perisomatic portion of certain neurons. Whereas interneurons associated to this specific neuron-associated sheath have been elaborately classified, less effort has been undertaken to describe the occurrence of perineuronal nets around pyramidal neurons. Our aim was to give a detailed and comparative description of the occurrence of net-associated pyramidal cells throughout the rat neocortex as well as to systematically and comparatively analyze the relation of main projection types of principal neurons to the presence of perineuronal nets. The present study revealed that perineuronal nets stained with WFA were associated rather rarely to pyramidal cells compared to interneurons in layers II/III and V/VI of rat neocortex. However, their frequency was considerably different between various cortical areas with a maximum in visual cortex and with a minimum in secondary motor cortices. Further analysis revealed that neuron-associated matrix sheaths around principal cells were more common in the primary than in the secondary fields of corresponding areas and they were more numerous in infra-than in supragranular layers in most regions. Subfields of cortical areas also differed regarding the occurrence of net-associated principal cells, and the subtlety of cortical representation seemed to correlate with the frequency of perineuronal nets around pyramidal neurons in the primary somatosensory cortex. It appears that net-associated pyramidal cells do not have a projection pattern restricted to distinct target regions. Rather a functional heterogeneity of the pyramidal cell population contributing to specific intra-or subcortical projections is suggested.

Axon initial segment ensheathed by extracellular matrix in perineuronal nets

Perineuronal nets of extracellular matrix are associated with distinct types of neurons in the cerebral cortex and many subcortical regions. Large complexes of aggregating proteoglycans form a chemically specified microenvironment around the somata, proximal dendrites and the axon initial segment, including the presynaptic boutons attached to these domains. The subcellular distribution and the temporal course of postnatal formation suggest that perineuronal nets may be involved in the regulation of synaptic plasticity. Here we investigate structural and cytochemical characteristics of the extracellular matrix around axon initial segments virtually devoid of synaptic contacts. Wisteria floribunda agglutinin staining, the immunocytochemical detection of aggrecan and tenascin-R, as well as affinity-labeling of hyaluronan were used to analyze perineuronal nets associated with large motoneurons in the mouse superior colliculus. The molecular composition of perineuronal nets was divergent between neurons but was identical around the different cellular domains of the individual neurons. The axon initial segments largely devoid of synapses were covered by a continuous matrix sheath infiltrating the adjacent neuropil. The periaxonal zone penetrated by matrix components often increased in diameter along the initial segment from the axon hillock toward the myelinated part of the axon. The axonal and somatodendritic domains of perineuronal nets were concomitantly formed during the first three weeks of postnatal development. The common molecular properties and major structural features of subcellular perineuronal net domains were retained in organotypic midbrain slice cultures. The results support the hypothesis that the aggrecan-related extracellular matrix of perineuronal nets provides a continuous micromilieu for different subcellular domains performing integration and generation of the electrical activity of neurons.

Genetic variability in the extracellular matrix protein as a determinant of risk for developing HTLV-I-associated neurological disease

Aggrecan, which is a well-known proteoglycan in joint cartilage, also exists in the spinal cord and plays an important role in maintaining water content in the extracellular matrix structure. In this study, we first examined the variable number of tandem repeat (VNTR) polymorphism of the aggrecan gene in 227 HTLV-I associated myelopathy/tropical spastic paraparesis (HAM/TSP) patients, in 217 HTLV-I-infected healthy carriers (HCs), and in 85 normal controls. The VNTR allele 28 (1,630 bp) was more frequently observed in HAM/TSP patients than in HCs (chi2=12.02, p=0.0005, odds ratio 1.79, 95% C.I. 1.29-2.50) and in controls (chi2=13.43, p=0.0002, odds ratio 2.54, 95% C.I. 1.52-4.25), although this allele was not related to disease progression or to HTLV-I provirus load. We also found that the aggrecan concentration in cerebrospinal fluid (CSF) from rapidly progressive HAM/TSP patients was significantly higher than in slowly progressive patients (corrected p=0.0145) but not in infected non-inflammatory neurological other disease controls (OND) (corrected p=0.078). We then analyzed this aggrecan VNTR polymorphism in the different set of patients with HAM/TSP (n=58) and healthy carriers (n=70). This analysis, again, revealed that allele 28 was detected more frequently in HAM/TSP group than in HCs (chi2=11.03, p=0.0009, odd ratio 3.04, 95% C.I. 1.55-5.97). The reproducibility of our study was regarded as a second- or third-class association by comparing combined p values and the Better Associations for Disease and GEnes (BADGE) system. Our results suggest that aggrecan polymorphism can be a novel genetic risk factor for developing HAM/TSP.

A disintegrin and metalloprotease 21 (ADAM21) is associated with neurogenesis and axonal growth in developing and adult rodent CNS

We have reported that alpha6beta1 integrin regulates the directed migration of neuroblasts from the adult rodent subventricular zone (SVZ) through the rostral migratory stream (RMS). ADAM (a disintegrin and metalloprotease) proteins bind integrins. Here, we show that ADAM21, but not ADAM2, -3, -9, -10, -12, -15, or -17, is expressed in adult rats and mice by ependyma and SVZ cells with long basal processes, and in radial glia at early postnatal times. ADAM21-positive processes projected into the RMS, contacted blood vessels, and were present within the RMS intermingled with neuroblasts up to where neuroblasts start their radial migration and differentiation in the olfactory bulb. Tissue inhibitors of metalloproteases (TIMPs) 1, 2, and 3 are present in the ependymal layer but not in the SVZ and RMS. Thus, ADAM21 could regulate neurogenesis and guide neuroblast migration through cleavage-dependent activation of proteins and integrin binding. ADAM21 is also present in growing axonal tracts during postnatal development and in growing primary olfactory axons in adults. In the olfactory nerve layer, ADAM21 often, but not always, colocalizes with OMP, a marker of mature olfactory neurons, but is not colocalized with the immature marker betaIII-tubulin. This suggests that ADAM21 is involved in the final axonal outgrowth phase and/or synapse formation. TIMP3 is present in periglomerular neurons, where it could restrict ADAM21-mediated axonal growth to the glomeruli. ADAM21's unique disintegrin and metalloprotease sequences and its restricted expression suggest that it might be a good target for influencing neurogenesis and neuronal plasticity.

Perineuronal nets potentially protect against oxidative stress

A specialized form of extracellular matrix (ECM) termed perineuronal nets (PNs) consisting of large aggregating chondroitin sulfate proteoglycans (CSPGs), with hyaluronan and tenascin as main components, surrounds subpopulations of neurons. The glycosaminoglycan components of perineuronal nets form highly charged structures in the direct microenvironment of neurons and thus might be involved in local ion homeostasis. The polyanionic character suggests that perineuronal nets also potentially contribute to reduce the local oxidative potential in the neuronal microenvironment by scavenging and binding redox-active iron, thus providing some neuroprotection to net-associated neurons. Here, we show that neurons ensheathed by a perineuronal net in the human cerebral cortex are less frequently affected by lipofuscin accumulation than neurons without a net both in normal-aged brain and Alzheimer's disease (AD). As lipofuscin is an intralysosomal pigment composed of cross-linked proteins and lipids generated by iron-catalyzed oxidative processes, the present results suggest a neuroprotective function of perineuronal nets against oxidative stress, potentially involved in neurodegeneration.

The brain angiotensin system and extracellular matrix molecules in neural plasticity, learning, and memory

The brain renin-angiotensin system (RAS) has long been known to regulate several classic physiologies including blood pressure, sodium and water balance, cyclicity of reproductive hormones and sexual behaviors, and pituitary gland hormones. These physiologies are thought to be under the control of the angiotensin II (AngII)/AT1 receptor subtype system. The AT2 receptor subtype is expressed during fetal development and is less abundant in the adult. This receptor appears to oppose growth responses facilitated by the AT1 receptor, as well as growth factor receptors. Recent evidence points to an important contribution by the brain RAS to non-classic physiologies mediated by the newly discovered angiotensin IV (AngIV)/AT4 receptor subtype system. These physiologies include the regulation of blood flow, modulation of exploratory behavior, and a facilitory role in learning and memory acquisition. This system appears to interact with brain matrix metalloproteinases in order to modify extracellular matrix molecules thus permitting the synaptic remodeling critical to the neural plasticity presumed to underlie memory consolidation, reconsolidation, and retrieval. There is support for an inhibitory influence by AngII activation of the AT1 subtype, and a facilitory role by AngIV activation of the AT4 subtype, on neuronal firing rate, long-term potentiation, associative and spatial learning. The discovery of the AT4 receptor subtype, and its facilitory influence upon learning and memory, suggest an important role for the brain RAS in normal cognitive processing and perhaps in the treatment of dysfunctional memory disease states.

Perisynaptic barrier of proteoglycans in the mature brain and spinal cord

Cell bodies and their dendrites of motor neurons, motor-related neurons, and certain other subsets of neurons such as GABAergic interneurons in the mature brain and spinal cord possess intensely negatively charged perineuronal or perisynaptic nets of proteoglycans which are linked to the nerve cell surface glycoproteins. These perineuronal nets of proteoglycans are digested by chondroitinase ABC, hyaluronidase, or collagenase, but not by endo-alpha-N-acetylgalactosaminidase, which is reactive to the nerve cell surface glycoproteins. Aggrecan, versican, neurocan, and brevican are members of a family of chondroitin sulfate proteoglycans that bind to hyaluronan. Neurocan- or brevican-deficient mice showed a regionally heterogeneous composition of proteoglycans in perineuronal nets. Aggrecan glycoforms contribute to the molecular heterogeneity of the perineuronal nets. Proteoglycans such as phosphacan are included in matrix-associated proteoglycans. The extracellular matrix glycoprotein tenascin-R is accumulated in the perineuronal nets. The perineuronal proteoglycans are produced by associated satellite astrocytes just before weaning, while the nerve cell surface glycoproteins are produced by the associated nerve cells at earlier stages after birth. The perineuronal proteoglycans may entrap the tissue fluid and form a perineuronal gel layer which protects the synapses as a "perisynaptic barrier". Degradation of the perineuronal proteoglycans or perisynaptic barrier by treatment with chondroitinase ABC or hyaluronidase reactivates the neuronal plasticity or promotes the functional recovery of a severed nervous system. Another set of perineuronal nets occurs, which are intensely positively charged and contain guanidino compounds. It is considered that these intensely positively charged nets are intermingled with the intensely negatively charged ones of proteoglycans.

Releasing the peri-neuronal net to patch-clamp neurons in adult CNS

The extracellular matrix of adult neural tissue contains chondroitin sulphated proteogylcans that form a dense peri-neuronal net surrounding the cell body and proximal dendrites of many neuronal classes. Development of the peri-neuronal net beyond approximately postnatal day 17 obscures visualization and often access by patch electrodes to neuronal membranes with the result that patch clamp recordings are most readily obtained from early postnatal animals. We describe a technique in which the surface tension of a sucrose-based medium promotes partial dissociation of thin tissue slices from adult tissue. Surface tension spreads the tissue and loosens the peri-neuronal net from neuronal membranes within minutes and in the absence of proteolytic enzymes. Furthermore, the extent of dissociation can be controlled so as to maintain the overall slice structure and allow identification of specific cell classes. Excellent structural preservation of neurons and dendrites can be obtained and full access by patch electrodes made possible for current- or voltage-clamp recordings in tissue well beyond the development of peri-neuronal nets. We demonstrate the feasibility of using this approach through patch recordings from neurons in the brainstem and cerebellum of adult gymnotiform fish and in deep cerebellar nuclei of rats as old as 6 months.

Chondroitin sulphate proteoglycans: preventing plasticity or protecting the CNS?

It is well established that axonal regeneration in the adult CNS is largely unsuccessful. Numerous axon-inhibitory molecules are now known to be present in the injured CNS, and various strategies for overcoming these obstacles and enhancing CNS regeneration have been experimentally developed. Recently, the use of chondroitinase-ABC to treat models of CNS injury in vivo has proven to be highly beneficial towards regenerating axons, by degrading the axon-inhibitory chondroitin sulphate glycosaminoglycan chains found on many proteoglycans in the astroglial scar. This enzyme has now been shown to restore synaptic plasticity in the visual cortex of adult rats by disrupting perineuronal nets, which contain high levels of chondroitin sulphate proteoglycans (CS-PGs) and are expressed postnatally around groups of certain neurons in the normal CNS. The findings suggest exciting prospects for enhancing growth and plasticity in the adult CNS; however, some protective roles of CS-PGs in the CNS have also been demonstrated. Clearly many questions concerning the mechanisms regulating expression of extracellular matrix molecules in CNS pathology remain to be answered.

Diffuse perineuronal nets and modified pyramidal cells immunoreactive for glutamate and the GABAA receptor α1 subunit form a unique entity in rat cerebral cortex

Perineuronal nets (PNs) consisting of polyanionic chondroitin sulfate proteoglycans (CSPG) and other extracellular matrix components create an exceptional microenvironment around certain types of neurons. In rat neocortex, three types of PNs can be distinguished after staining with Wisteria floribunda agglutinin (WFA) by their different morphological structure: lattice-like PNs associated with subpopulations of nonpyramidal neurons, weakly labeled PNs showing a pyramidal morphology, and diffuse PNs that possess a thick, strongly labeled matrix sheath located mainly in layer VIb above the white matter. The type of neuron surrounded by diffuse nets has not been described so far. This study is focused on the cytochemical and morphological characteristics of neurons associated with diffusely contoured PNs in rat parietal cortex using immunocytochemical staining, intracellular injection, and retrograde tracing methods. Cells surrounded by diffuse PNs were glutamate-immunoreactive in contrast to nonpyramidal, net-associated neurons that showed immunoreactivity for GABA, the calcium-binding protein parvalbumin and the potassium channel subunit Kv3.1b. Both groups of PN-ensheathed cells were mostly immunoreactive for the GABA(A) receptor alpha1 subunit. Lucifer Yellow-injected neurons surrounded by diffuse PNs displayed the morphological properties of modified pyramidal cells with intracortical main axons. Many neurons with diffuse PNs were retrogradely labeled over a long distance after Fluoro-Gold tracer injection in the parietal cortex, but remained unlabeled after intrathalamic injection. We conclude that neurons associated with diffuse PNs are a subpopulation of glutamatergic modified pyramidal cells that could act as excitatory long-range intracortically projecting neurons.

Region and lamina-specific distribution of extracellular matrix proteoglycans, hyaluronan and tenascin-R in the mouse hippocampal formation

The extracellular matrix is known to show region-specific characteristics in the adult brain. Our comparative cytochemical study is focused on the laminar organisation of major extracellular matrix constituents in the murine hippocampal formation, including the regions CA1, CA2 and CA3 of the hippocampus proper, the dentate gyrus, the subiculum and the presubiculum. Components related to chondroitin sulphate proteoglycans were detected by N-acetylgalactosamine-binding Wisteria floribunda agglutinin, colloidal iron staining, and antibodies to different proteoglycan domains, including the Cat-301 and Cat-315 epitopes of aggrecan, as well as neurocan, brevican and phosphacan. The distribution patterns of these components were correlated with the patterns revealed for hyaluronan and the brain-specific extracellular matrix glycoprotein, tenascin-R, known to be ligands of extracellular matrix proteoglycans. Lectin binding clearly labelled perineuronal nets of the extracellular matrix around interneurons, which were preferentially located within or near the principal cell layers in all regions. In the hippocampus proper, the CA2 subfield showed an intense labelling of the neuropil around pyramidal cell bodies and the neuropil zones in the strata oriens and radiatum. These patterns were also seen after immunoreaction for chondroitin proteoglycan domains, brevican and phosphacan, as well as after detection of hyaluronan and tenascin-R. Characteristic laminar and intralaminar patterns were additionally expressed in the neuropil in all regions. In the dentate gyrus, the staining intensity for brevican, phosphacan and tenascin-R was predominant in the middle molecular layer, and for Cat-315 in the inner molecular layer, whereas immunoreactivity for neurocan increased within the outer molecular layer towards the hippocampal fissure. Our findings indicate that proteoglycans, hyaluronan and tenascin-R show differential patterns of co-expression in the individual regions and laminae of the hippocampal formation. The inhomogeneous composition of these major components suggests that the extracellular matrix is specifically adapted to the functional domains of intrahippocampal connections and afferent fibre systems.

Histochemical study of perineuronal nets in the retrosplenial cortex of adult rats

The retrosplenic cortex of rats, similar to many cortical or subcortical regions, is provided with special subsets of neurons that exhibited a fenestrated or reticular coat of condensed extracellular matrix on their soma, initial dendrites and proximal axon segment. This pericellular coating, currently termed "Perineuronal Nets", was detected on the surfaces of some neurons distributing throughout the cortical layers II-V. They presented direct interconnections with each other, and appeared in close association to the astroglial processes. In addition to their collagenous ligands, the perineuronal nets (PNs) were enriched with proteoglycans (PGs, sulfated glycoconjugates) and/or glycoproteins (GPs, unsulfated glycoconjugates with terminal N-acetylgalactosamine). Accordingly, the PNs were differentially identified as belonging to three categories, depending upon their organic nature or chemical composition. First, coats exclusively formed of PGs (stained with iron colloid); second, coats formed of GPs (labeled with plant lectins binding to terminal N-acetylgalactosamine); and third, complex coats formed of PG networks intermingled with glycoprotein molecules (double stained with iron colloid and lectin). Since differential distribution of protein containing substances (GPs and/or PGs) in the extracellular matrix contributes to functional terms, we suggest that these biochemical or morphological differences in the microenvironment of some retrosplenial neurons might reflect certain functional aspects concerned with processing of navigation or episodic memory.

Perineuronal nets in the rhesus monkey and human basal forebrain including basal ganglia

Perineuronal nets of extracellular matrix have been shown to characterize the microenvironment of individual neurons and the chemoarchitecture of brain regions such as basal forebrain nuclei. Previous work has also demonstrated that neurons in the human cerebral cortex ensheathed by perineuronal nets rarely undergo cytoskeletal changes in Alzheimer's disease, suggesting a neuroprotective effect of extracellular matrix components. It is not known, however, whether or not perineuronal nets are absent in the microenvironment of the cholinergic basal forebrain neurons that are involved early in the cascade of neurodegeneration in humans. Therefore, the present study was undertaken to examine the distribution patterns of perineuronal nets in the basal forebrain of the higher primates, rhesus monkey and human. Cytochemical staining was performed with the lectin Wisteria floribunda agglutinin and a polyclonal antibody to core proteins of chondroitin sulfate proteoglycans in the perfusion-fixed tissue of rhesus monkeys. In human brains, perineuronal nets were only stained with the immunoreaction for chondroitin sulfate proteoglycans. The results showed similar characteristics in distribution patterns of perineuronal nets in the medial septum, the diagonal band of Broca, the basal nucleus of Meynert (Ch1-Ch4), the lateral septum, the caudate-putamen, and the globus pallidus in both species. Double-labelling revealed that the vast majority of cholinergic neurons, labelled either with antibodies to choline acetyltransferase or the low-affinity neurotrophin receptor p75(NTR), were not ensheathed by perineuronal nets. A small subpopulation of net-associated neurons in close proximity to or intermingled with cholinergic neurons of the Ch1-Ch4 cell groups was found to be immunoreactive for parvalbumin. In the caudate-putamen, a large number of the parvalbumin-positive neurons were surrounded by perineuronal nets, whereas in the external and internal segments of the globus pallidus the coincidence of both markers was nearly complete. The study demonstrates that perineuronal nets of extracellular matrix are associated with different types of non-cholinergic neurons in the primate basal forebrain. The absence of nets around cholinergic basal forebrain neurons may be related to their slow modulatory activity but may also contribute to their susceptibility to degeneration in Alzheimer's disease.

Extracellular matrix molecules, long-term potentiation, memory consolidation and the brain angiotensin system

Considerable evidence now suggests an interrelationship among long-term potentiation (LTP), extracellular matrix (ECM) reconfiguration, synaptogenesis, and memory consolidation within the mammalian central nervous system. Extracellular matrix molecules provide the scaffolding necessary to permit synaptic remodeling and contribute to the regulation of ionic and nutritional homeostasis of surrounding cells. These molecules also facilitate cellular proliferation, movement, differentiation, and apoptosis. The present review initially focuses on characterizing the ECM and the roles of cell adhesion molecules (CAMs), matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs), in the maintenance and degradation of the ECM. The induction and maintenance of LTP is described. Debate continues over whether LTP results in some form of synaptic strengthening and in turn promotes memory consolidation. Next, the contribution of CAMs and TIMPs to the facilitation of LTP and memory consolidation is discussed. Finally, possible roles for angiotensins, MMPs, and tissue plasminogen activators in the facilitation of LTP and memory consolidation are described. These enzymatic pathways appear to be very important to an understanding of dysfunctional memory diseases such as Alzheimer's disease, multiple sclerosis, brain tumors, and infections.

Modification of extracellular matrix by enzymatic removal of chondroitin sulfate and by lack of tenascin-R differentially affects several forms of synaptic plasticity in the hippocampus

The extracellular matrix is a complex network of macromolecules including glycoproteins, polysaccharides and proteoglycans. Tenascin-R and chondroitin sulfate proteoglycans are essential components of hippocampal extracellular matrix co-localised in perineuronal nets on interneurons. Mutant mice deficient in expression of tenascin-R showed a two-fold reduction of long-term potentiation induced by theta-burst stimulation of Schaffer collaterals in the stratum radiatum of the CA1 region of the hippocampus, as compared to wild-type mice. The same reduction in potentiation was observed in slices from wild-type mice pretreated for 2h with chondroitinase ABC that completely removed chondroitin sulfates from the extracellular matrix. Treatment of slices from tenascin-R deficient animals with the enzyme did not further reduce potentiation in comparison with untreated slices from these mice, showing an occlusion of effects produced by removal of tenascin-R and chondroitin sulfates. However, the level of potentiation recorded immediately after theta-burst stimulation was significantly higher in wild-type than in tenascin-R deficient mice, whereas chondroitinase ABC had no significant effect on this short-term form of plasticity. Enzymatic treatment also did not affect short-term depression evoked by low-frequency stimulation, whereas this form of synaptic plasticity was reduced in tenascin-R deficient mice. In contrast, long-term depression in CA1 was impaired by digestion of chondroitin sulfates but appeared normal in tenascin-R mutants. Our data demonstrate that tenascin-R and chondroitin sulfate proteoglycans differentially modulate several forms of synaptic plasticity, suggesting that different mechanisms are involved.

Hyaluronan-associated adhesive cues control fiber segregation in the hippocampus

In various brain regions, particularly in the hippocampus, afferent fiber projections terminate in specific layers. Little is known about the molecular cues governing this laminar specificity. To this end we have recently shown that the innervation pattern of entorhinal fibers to the hippocampus is mimicked by the lamina-specific adhesion of entorhinal cells on living hippocampal slices, suggesting a role of adhesion molecules in the positioning of entorhinal fibers. Here, we have analyzed the role of extracellular matrix components in mediating this lamina-specific adhesion. We show that hyaluronidase treatment of hippocampal slices abolishes lamina-specific adhesion as well as layer-specific growth of entorhinal fibers to the dentate outer molecular layer in organotypic slice cultures. We conclude that hyaluronan-associated molecules play a crucial role in the formation of the lamina-specific entorhinal projection to the hippocampus.

Neutralizing antibodies against neurite growth inhibitor NI-35/250 do not promote regeneration of sensory axons in the adult rat spinal cord

Neutralization of the myelin-associated neurite growth inhibitors NI-35 and NI-250 by IN-1 antibodies can promote axonal regeneration of several types of central nervous neurons. Here, we investigated in adult rats whether IN-1 can promote regeneration of ascending sensory axons across a peripheral nerve bridge back into the spinal cord. IN-1 was administered by hybridoma cells injected in the cerebral cortex or thoracic cord, its presence confirmed in tissue sections and cerebrospinal fluid, and its effectiveness demonstrated in co-cultures of oligodendrocytes and sensory neurons. With a two week infusion of control vehicle into the dorsal spinal cord 3 mm rostral to the nerve graft, only 3+/-2% of the anterogradely labeled sensory fibers present at the rostral end of the nerve graft had grown up to 0.5 mm, but not farther into the spinal cord. A similar limited extent of regeneration was seen with IN-1 or with infusion of Dantrolene, an inhibitor of NI-35/250 activity in vitro. With infusion of nerve growth factor rostral to the nerve graft, 40% of the fibers at the rostral end of the graft were found at 0.5 mm, 34% at 1 mm, 24% at 2 mm and 14% at 3 mm (the infusion site) into the spinal cord. Treatment with IN-l antibodies did not enhance the growth-promoting effects of nerve growth factor. We suggest that the neurite growth inhibitors NI-35 or NI-250 do not play a major inhibitory role in the regeneration of the ascending sensory fibers across a nerve bridge and back into the spinal cord of the adult rat.

Postnatal development of perineuronal nets in wild-type mice and in a mutant deficient in tenascin-R

The extracellular matrix glycoprotein tenascin-R (TN-R), colocalizing with hyaluronan, phosphacan, and aggregating chondroitin sulphate proteoglycans in the white and grey matter, is accumulated in perineuronal nets that surround different types of neurons in many brain regions. To characterize the role of TN-R in the formation of perineuronal nets, we studied their postnatal development in wild-type mice and in a TN-R knock-out mutant by using the lectin Wisteria floribunda agglutinin and an antibody to nonspecified chondroitin sulphate proteoglycans as established cytochemical markers. We detected the matrix components TN-R, hyaluronan, phosphacan, neurocan, and brevican in the perineuronal nets of cortical and subcortical regions. In wild-type mice, lectin-stained, immature perineuronal nets were first seen on postnatal day 4 in the brainstem and on day 14 in the cerebral cortex. The staining intensity of these nets for TN-R, hyaluronan, phosphacan, neurocan, and brevican was extremely weak or not distinguishable from that of the surrounding neuropil. However, all markers showed an increase in staining intensity of perineuronal nets reaching maximal levels between postnatal days 21 and 40. In TN-R-deficient animals, the perineuronal nets tended to show a granular component within their lattice-like structure at early stages of development. Additionally, the staining intensity in perineuronal nets was reduced for brevican, extremely low for hyaluronan and neurocan, and virtually no immunoreactivity was detectable for phosphacan. The granular configuration of perineuronal nets became more predominant with advancing age of the mutant animals, indicating the continued abnormal aggregation of chondroitin sulphate proteoglycans complexed with hyaluronan. As shown by electron microscopy in the cerebral cortex, the disruption of perineuronal nets was not accompanied by apparent changes in the synaptic structure on net-bearing neurons. The regional distribution patterns and the temporal course of development of perineuronal nets were not obviously changed in the mutant. We conclude that the lack of TN-R initially and continuously disturbs the molecular scaffolding of extracellular matrix components in perineuronal nets. This may interfere with the development of the specific micromilieu of the ensheathed neurons and adjacent glial cells and may also permanently change their functional properties.

Nervous system proteoglycans as modulators of neurite outgrowth

The proteoglycans are multifunctional macromolecules composed of a core polypeptide and a variable number of glycosaminoglycan chains. The structural diversity and complexities of proteoglycan expression in the developing and adult Nervous System underlies the variety of biological functions that these molecules fulfill. Thus, in the Nervous System, proteoglycans regulate the structural organisation of the extracellular matrix, modulate growth factor activities and cellular adhesive and motility events, such as cell migration and axon outgrowth. This review summarises the evidences indicating that proteoglycans have an important role as modulators of neurite outgrowth and neuronal polarity. Special emphasis will be placed on those studies that have shown that proteoglycans of certain subtypes inhibit neurite extension either during the development and/or the regeneration of the vertebrate Central Nervous System.

The extracellular matrix in the mouse brain: its reactions to endo-alpha-N-acetylgalactosaminidase and certain other enzymes

As our previous studies have indicated, the cingulate cortex of the adult mouse brain contains many neurons with rich cell surface glycoproteins which are linked by collagenous ligands to perineuronal proteoglycans. The present study demonstrated that exclusive incubation with endo-alpha-N-acetylgalactosaminidase abolished the lectin Vicia villosa or Wisteria floribunda agglutinin (VVA or WFA) labeling of the nerve cell surface glycoproteins, while it neither interfered with the cationic iron colloid or aldehyde fuchsin stainings of the perineuronal proteoglycans nor abolished the Gömöri's ammoniacal silver impregnation of the collagenous ligands. Double incubations with endo-alpha-N-acetylgalactosaminidase and collagenase did not eliminate the lectin VVA or WFA labeling of the nerve cell surface glycoproteins, though they did eliminate the cationic iron colloid and aldehyde fuchsin stainings of the perineuronal proteoglycans as well as the Gömöri's ammoniacal silver impregnation of the collagenous ligands. Triple incubations with endo-alpha-N-acetylgalactosaminidase, collagenase, and endo-alpha-N-acetylgalactosaminidase abolished the lectin VVA or WFA labeling of the nerve cell surface glycoproteins, and also eliminated the cationic iron colloid and aldehyde fuchsin stainings of the perineuronal proteoglycans and the Gömöri's ammoniacal silver impregnation of the collagenous ligands. These findings indicate that: the nerve cell surface glycoproteins or their terminal N-acetylgalactosamines are digested by endo-alpha-N-acetylgalactosaminidase; these galactosamines associated with the collagenous ligands or perineuronal proteoglycans are not digested by endo-alpha-N-acetylgalactosaminidase; and the terminal N-acetylgalactosamines newly exposed by collagenase incubation are digested by this galactosaminidase. It was further demonstrated that hyaluronidase incubation neither digests the collagenous ligands nor revives the lectin VVA or WFA labeling of the nerve cell surface proteoglycans.

Mice deficient for tenascin-R display alterations of the extracellular matrix and decreased axonal conduction velocities in the CNS

Tenascin-R (TN-R), an extracellular matrix glycoprotein of the CNS, localizes to nodes of Ranvier and perineuronal nets and interacts in vitro with other extracellular matrix components and recognition molecules of the immunoglobulin superfamily. To characterize the functional roles of TN-R in vivo, we have generated mice deficient for TN-R by homologous recombination using embryonic stem cells. TN-R-deficient mice are viable and fertile. The anatomy of all major brain areas and the formation and structure of myelin appear normal. However, immunostaining for the chondroitin sulfate proteoglycan phosphacan, a high-affinity ligand for TN-R, is weak and diffuse in the mutant when compared with wild-type mice. Compound action potential recordings from optic nerves of mutant mice show a significant decrease in conduction velocity as compared with controls. However, at nodes of Ranvier there is no apparent change in expression and distribution of Na+ channels, which are thought to bind to TN-R via their beta2 subunit. The distribution of carbohydrate epitopes of perineuronal nets recognized by the lectin Wisteria floribunda or antibodies to the HNK-1 carbohydrate on somata and dendrites of cortical and hippocampal interneurons is abnormal. These observations indicate an essential role for TN-R in the formation of perineuronal nets and in normal conduction velocity of optic nerve.

Perineuronal nets of proteoglycans in the adult mouse brain are digested by collagenase

As our previous study has indicated, perineuronal proteoglycans in the adult mouse brain are associated with some collagenous molecules which can be stained with Gömöri's ammoniacal silver and are resistant to hyaluronidase digestion. The present study demonstrated that these molecules are thoroughly digested with collagenase, and suggests that they represent a hyaluronic acid-binding domain of the ligand proteoglycans connecting the perineuronal proteoglycans and nerve cell surface glycoproteins.

Perineuronal nets of proteoglycans in the adult mouse brain, with special reference to their reactions to Gömöri's ammoniacal silver and Ehrlich's methylene blue

As our previous studies have indicated, many subsets of neurons in the vertebrate brain possess a sulfated proteoglycan surface coat which reacts to cationic iron colloid and aldehyde fuchsin. The present study demonstrated that this surface coat is supravitally stained with Ehrlich's methylene blue, and doubly with this blue and aldehyde fuchsin, a finding suggesting its being identical to Cajal's superficial reticulum (red superficial) and to Golgi's reticular coating (revetement reticulare). The perineuronal surface coat was further stained with Gömöri's ammoniacal silver, and doubly with this silver and cationic iron colloid. These neurons with such a proteoglycan surface coat usually expressed cell surface glycoproteins which were labeled with lectin Wisteria floribunda agglutinin. Hyaluronidase digestion did not interfere with this lectin labeling of the glycoproteins, methylene blue and Gömöri's ammoniacal silver staining of the surface coat, while it erased the cationic iron colloid and aldehyde fuchsin staining of the surface coat. These findings suggest that the perineuronal proteoglycan surface coat is associated with some additional molecules which are resistant to hyaluronidase digestion and stainable with methylene blue and Gömöri's ammoniacal silver. The possibility is suggested that these molecules might represent "ligand proteoglycans" connecting the perineuronal proteoglycans and cell surface glycoproteins.

Low expression of extracellular matrix components in rat brain stem regions containing modulatory aminergic neurons

Extracellular matrix proteoglycans, particularly those accumulated in perineuronal nets (PNs), have been shown to form characteristic distribution patterns in cortical and subcortical regions of adult mammals. Their involvement in sustaining mechanisms that are especially related to fast activities of neurons has been discussed as one of the possible functions. The present study deals with the spatial organization of extracellular matrix proteoglycans in brain stem regions that contain aminergic neurons, such as substantia nigra, ventral tegmental area (VTA), raphe nuclei and locus coeruleus (LC). As these nuclei are known to influence brain activity by modulatory functions exerting patterns of slow electric activity, it could be expected that PNs would be absent around aminergic cells. The staining of PNs with Wisteria floribunda agglutinin (WFA) was combined with the detection of catecholaminergic neurons by tyrosine hydroxylase immunoreactivity and of serotonergic neurons by tryptophan hydroxylase (TH) immunoreactivity using double fluorescence microscopy. It was found that the catecholaminergic and serotonergic neurons in the nuclear accumulations, as well as those scattered in adjacent regions, were not ensheathed by PNs. In contrast, several non-aminergic neurons intermingled with aminergic neurons in the raphe nuclei, in the substantia nigra pars compacta (SNC) and in the VTA, as well as many cells in the reticular part of the substantia nigra, were found to be surrounded by PNs. It can be concluded from these results that the absence of PNs around aminergic brain stem neurons, also previously shown for cholinergic basal forebrain neurons, appears as a characteristic feature common to cells that exert slow modulatory functions.

Occurrence of a N-terminal proteolytic fragment of neurocan, not a C-terminal half, in a perineuronal net in the adult rat cerebrum

Neurocan is a nervous tissue-unique chondroitin sulfate proteoglycan (CSPG) whose expression and proteolytic cleavage are developmentally regulated. In the adult rat brain, neurocan is completely cleaved into some proteoglycan fragments including the C-terminal half known as neurocan-C and a N-terminal fragment with a 130 kDa core glycoprotein (neurocan-130). We describe here the differential distribution of these two neurocan-derived CSPGs in the adult rat cerebrum and the occurrence of neurocan-130 as a new member of a perineuronal net-constituting molecule. At the light microscopic level, neurocan-130 exhibited pericellular localization around a subset of neurons in addition to diffuse distribution in the neuropil. In contrast, neurocan-C was distributed only diffusely in the neuropil. Double staining with anti-neurocan-130 and anti-synaptophysin antibodies suggested that neurocan-130 was localized in the vicinity of the synapses, but not at the synapses. Immunoelectron microscopy showed that neurocan-130 was mainly localized in the cytoplasm of glial cell processes, the so-called glial perineuronal net, encompassing the cell bodies of certain neurons. The presence of neurocan-130 in a limited number of glial cells may reflect some functional heterogeneity of the glia.