Maturation of astrocytes in vitro alters the extent and molecular basis of neurite outgrowth

Distribution and possible roles of the highly polysialylated neural cell adhesion molecule (NCAM-H) in the developing and adult central nervous system

The neural cell adhesion molecule (NCAM) is a cell surface glycoprotein which is thought to mediate cell adhesion and recognition. During developmental stages, NCAM is highly polysialylated (NCAM-H) by a unique alpha-2,8-linked polysialic acid chain (PSA), and this PSA portion of NCAM-H has been found to be closely associated with various developmental processes of the nervous system. Further, recent immunohistochemical investigations have revealed that even in the adult nervous system, a persistent PSA expression has been found confined to several regions: the olfactory bulb, the piriform cortex, the hippocampal dentate gyrus, the hypothalamus, some nuclei of the medulla and the dorsal horn of the spinal cord, which are related directly or indirectly to sensory systems. Moreover, in the dentate gyrus and olfactory bulb the expression is connected with adult neurogenesis that may add new neuronal circuits to the adult neural tissue. Therefore, the possible role of NCAM-H in the central nervous system may be associated not only with neural development, but also with adult functions, such as the processing system of sensory information and neuronal plasticity.

Establishment and neurite outgrowth properties of neonatal and adult rat olfactory bulb glial cell lines

Two glial cell types surround olfactory axons and glomeruli in the olfactory bulb (OB) and may influence synapse development and regeneration. OB astrocytes resemble type-1 astrocytes, and OB ensheathing cells resemble non-myelinating Schwann cells. We have produced clonal OB astrocyte and ensheathing cell lines from rat neonatal and adult OB cultures by SV40 large T antigen transduction. These cell lines have been characterized by morphology, growth characteristics, immunophenotype, and ability to promote neurite outgrowth in vitro. Neonatal and adult ensheathing cell lines were found to support higher neurite outgrowth than OB astrocyte lines. Neonatal OB astrocyte lines were of two types, high and low outgrowth support. The low support astrocyte lines express J1 and a chondroitin sulfate-containing proteoglycan as do astrocytes encircling the neonatal glomeruli in vivo. The adult OB astrocyte cell lines supported lower levels of outgrowth than adult ensheathing cell lines. These results are consistent with a positive role for ensheathing cells in OB synapse regeneration, in vivo. Further, based on our results, we hypothesize that ensheathing cells and high-outgrowth astrocytes facilitate axon growth in vivo, while low outgrowth astrocytes inhibit axon growth and may facilitate glomerulus formation.

Novel neurite growth-inhibitory properties of an astrocyte proteoglycan

Conditioned medium (CM) of primary cultures of GFAP-positive adherent astrocytes from neonatal rat neocortex contained a chondroitin sulphate/dermatan sulphate proteoglycan (CDSPG) that co-eluted with a heparan sulphate proteoglycan (HSPG) by ion-exchange chromatography. The CDSPG was resolved from the HSPG by molecular sieve chromatography, which indicated that the molecular mass of the HSPG was greater than 300 kDa, while that of the CDSPG was approximately 50 kDa. Specific lyase digestion and urea/polyacrylamide gel electrophoresis established the homogeneity of the CDSPG and suggested molecular masses of the core protein and glycosylated protein as 54 kDa and 58 kDa respectively. Virtually all of the poly-D-lysine substrate-bound proteoglycan-associated neurite growth-promoting activity of astrocyte CM was accounted for by the HSPG. On poly-D-lysine the immobilized CDSPG displayed little neurite growth-stimulatory activity relative to the HSPG. However, the CDSPG inhibited the potent growth-promoting activity of the HSPG by displacing it from the poly-D-lysine substrate. Differential cellular regulation of production of growth-modulatory proteins with different binding avidity for the substrate of growth may determine the success of a regenerative axonal response by fully competent neurons.

Analysis of a segment of the human glial fibrillary acidic protein gene that directs astrocyte-specific transcription

To understand astrocyte-specific transcription, we have been studying the human gfa gene. This gene encodes glial fibrillary acidic protein (GFAP), an intermediate filament protein expressed primarily in astrocytes. A survey of the gfa 5' flanking region showed it to contain several segments that contribute to expression of a chloramphenicol acetyltransferase reporter gene in transfected cells. The most active of these was the 124-bp B region, which spans bp -1612 to -1489. We have now used site-directed mutagenesis to analyze this region in greater detail, and show that the B region itself contains several important elements. The most crucial of these is a consensus AP-1 sequence, the binding site for the Fos and Jun families of transcription factors. The presence of members of both these families in the glial fibrillary acidic protein-expressing U251 cell line used for our transfection studies was verified by gel mobility-shift experiments. This is the first demonstration of the functioning of a specific transcription factor site for astrocytes, and provides a focus for future studies of glial fibrillary acidic protein regulation during development and reactive gliosis.

Embryonic optic nerve tissue fails to support neurite outgrowth by central and peripheral neurons in vitro

The failure of axon regeneration in the injured mammalian central nervous system has been ascribed, in part, to the inhibitory effects of myelin proteins. To investigate the influence of myelination on neurite growth and regeneration by both central nervous system and peripheral nervous system neurons, isolated rat neonatal retinal ganglion cells and adult and neonatal dorsal root ganglion neurons were cultured on cryostat sections of both immature unmyelinated and mature fully myelinated adult rat optic nerve. In agreement with earlier studies using neonatal peripheral neurons, the adult optic nerve failed to support neurite outgrowth from any of the neurons tested. A new finding was that tissue sections from unmyelinated optic nerve (aged embryonic days 18 and 20, and postnatal days 1-3), also failed to support the growth of neurites from neonatal retinal ganglion cells and both neonatal and adult dorsal root ganglion neurons. Neonatal retinal ganglion cells also failed to extend neurites on sections of pre-degenerated sciatic nerve, a tissue shown in our previous work to be a good substratum for supporting neurite growth for both neonatal and adult DRG neurons. These results suggest that cells in the immature optic nerve either express widely acting axon growth inhibitory molecules unrelated to previously described myelin proteins, or do not synthesize appropriate axon growth promoting molecules. They also reveal that, for axon regeneration, central nervous system and peripheral sensory neurons require distinct substratum interactions.

Cycloheximide downsynthesizes inhibitory molecules for neurite outgrowth in neural transplantation

Rat fetal substantia nigra treated with cycloheximide, a protein synthesis inhibitor, were implanted to adult rat striatum. After 4 weeks, tyrosine hydroxylase (TH)-like immunoreactive (-LI) fibers of host striatum penetrated grafts, and TH-LI neurites in the grafts elongated and mingled with the host striatal neurites. Astrocytes proliferated in the grafts without glial scar between the graft-host border. A few chondroitin sulfate- or tenascin-LI glial cells were found in the grafts, while in transplants without cycloheximide, glial scar expressed strong immunoreactivity for these molecules. Downsynthesis of these inhibitory molecules may alter the glial character and permit neurites traversing the border.

The upregulation of a glial cell surface antigen at the astrocytic scar in the rat

The upregulation of a 106-kDa glial protein was examined using indirect immunohistochemical methods. The protein is recognized by the monoclonal antibody AMP1. Previous studies demonstrated that the AMP1 antigen is found on the external surface of cultured astrocytes and is involved in stabilizing adhesive interactions between these cells. In sections of injured adult rat brain, the spatial distribution of the AMP1 antigen correlated with the region of reactive gliosis. The relative intensity of immunofluorescence indicates that the AMP1 antigen is dramatically upregulated at the gliotic scar. The correlation between AMP1 antigen upregulation and reactive gliosis suggests that this molecule is critical to the process of CNS scar formation.

A cell surface antigen expressed by astrocytes and their precursors

Increasing evidence suggests that astrocytes of the mammalian CNS are a heterogeneous population of cells that express a number of common characteristics. In most cases, astrocytes or their precursors contain a class of intermediate filaments, composed in large part of glial fibrillary acidic protein (GFAP). While the expression of GFAP immunoreactivity is a specific characteristic of astrocytes in the vertebrate CNS, not all astrocytes contain detectable levels of glial filaments, particularly during early development. We have isolated a monoclonal antibody termed 7B11 which binds to the surface of astrocytes and glial precursors, but not to other major types of neural cells. The 7B11 antigen is expressed by astrocytes in the adult CNS in vivo and in cultures of developing CNS tissue, but not on cells of the peripheral nervous system. During early development, 7B11 immunoreactivity appears prior to the expression of GFAP and is retained as punctate staining on the surface of most classes of astrocytes. During CNS maturation, however, 7B11 immunoreactivity is lost from the surface of Bergmann glia in the cerebellum, suggesting that differentiative events lead to functionally and antigenically distinct classes of CNS glial cells. In the adult spinal cord, biochemical analysis suggests that the epitope recognized by 7B11 is associated with a group of polypeptides of apparent molecular weights 200-160, 140, and 92 kD. The cellular distribution of 7B11 expression suggests that astrocytes and their precursors share a distinct cell surface antigenic property and that the expression of 7B11 immunoreactivity may be a useful operational marker for astrocytes in the absence of detectable GFAP expression.

Reactive astrocytes involved in the formation of lesional scars differ in the mediobasal hypothalamus and in other forebrain regions

The fine organization of lesional scars was studied in adult rats at the level of 2 types of surgical cuts aimed at deafferentating the dorsal hypothalamus from its neuropeptide-Y innervation. These included: (i) lesions located dorsolateral to the dorsal hypothalamus, which were shown to form a permanent obstacle to the regeneration of transected neuropeptide-Y-fibers, and (ii) lesions located in the ventromedial hypothalamus, where transected neuropeptide-Y-fibers were shown to penetrate and eventually cross the lesional area. Double labeling immunocytochemistry and conventional electron microscopy were used to identify various molecules produced by reactive astrocytes and to visualize their ultrastructural organization within the scars, respectively. In the different portions of the dorsolateral scars, the large majority of reactive astrocytes was characterized by a strong immunoreactivity to glial fibrillary acidic protein, vimentin, and embryonic (polysialylated) NCAM. Intense laminin-immunoreactivity was also observed over large patches included in the scar. Electron microscope observations further indicated that the matrix of the scar was mainly composed of tightly packed astrocytic perikarya and processes connected by extended gap junctions. All around the extracellular and perivascular spaces, these astrocyte profiles were bordered by a thick basal lamina. Only scarce axonal profiles were detected in the core of the scar, most of which exhibited degenerative features. In the ventromedial hypothalamic scars, reactive astrocytes were found to exhibit intense immunoreactivity to both glial fibrillary acidic protein and vimentin. On the other hand, only slight immunostaining to embryonic NCAM and laminin were associated with this type of lesional scar. At the ultrastructural level, the main differences with the dorsolateral scars concerned (i) the gap junctions, which were less frequent and involved shorter portions of adjacent membranes; (ii) the basal lamina, which was essentially localized to the perivascular spaces; and (iii) the axonal profiles, which were frequently observed throughout the scar matrix. These data indicate that reactive astrocytes that formed the glial scar differ in the mediobasal hypothalamus and in other forebrain regions. This provides strong support for the hypothesis that the regeneration of neuropeptide-Y axons through a mediobasal hypothalamic surgical cut depends mainly on the particular organization of the astroglial scar.

Neuropeptide Y-producing neurons of the arcuate nucleus regenerate axons after surgical deafferentation of the mediobasal hypothalamus

Dorsolateral and ventomedial surgical deafferentiation of the hypothalamus were used to study the capacity of different types of neuropeptide Y-containing axons afferent to the dorsal hypothalamus to regenerate through surgical lesions. The kinetics of the postlesional responses of transected neuropeptide Y-axons was studied on 30-40 microns thick vibratome sections, either (i) by light or electron microscopy after peroxidase immunostaining for neuropeptide Y or (ii) by confocal microscopy after double fluorescence immunostaining for neuropeptide Y and for glial fibrillary acidic protein. The dorsolateral cut was found to sever 2 main pathways containing neuropeptide Y axons located, respectively, below the bed nucleus of the stria terminalis and in the perifornical region. In both regions transected fibers were found to abut onto the surgical lesion, but even 45 days after the lesion, they were very rarely observed to penetrate into the astroglial scar forming along the lesion. The ventromedial cut was found to sever numerous neuropeptide axons that originate in the underlying arcuate nucleus. Seven to 15 days after the lesion neuropeptide Y fibers located below this type of cut presented a dramatic increase in both their numerical density and their immunostaining intensity. With increasing post-surgery times, an increased number of neuropeptide Y fibers was observed to penetrate and to cross the lesional scar formed by densely packed astrocytic processes. Electron microscope observations further demonstrated that 45 days after the lesion, numerous neuropeptide Y-immunoreactive axonal profiles were included in the scar matrix, which appeared to be mainly composed of closely interdigitating astrocytic processes containing dense bundles of filaments. These data indicate that, in contrast to other neuropeptide Y neurons innervating the dorsal hypothalamus, neuropeptide Y neurons of the arcuate nucleus regenerate axons through the astroglial scar produced by a surgical lesion placed in the ventromedial hypothalamus.

Modulation of adhesion molecule expression on rat cortical astrocytes during maturation

During development of the vertebrate CNS the functional properties of astrocytes change significantly. Many of these functional changes result from modifications in the expression of cell surface adhesion molecules on astrocytes that mediate the interactions of astrocytes with other astrocytes, neurons, and growing axons. In this study we have compared the cell surface expression of HNK-1, NCAM, and laminin on rat cortical type-I-like astrocytes during maturation in vitro and in vivo. Both the proportion of immunoreactive cells and the relative levels of expression of these antigens on different aged astrocyte populations were assayed by flow cytometry. At birth, most cortical type-I astrocytes express high levels of HNK-1 and NCAM, while approximately 50% of the cells express laminin. During maturation in vitro, the proportion of cortical astrocytes that expressed these surface molecules decreased over a period of 28 days, even though cell size and glial fibrillary acidic protein content increased. During maturation in vivo, a qualitatively and temporally similar decrease in antigen expression on astrocytes was observed. This reduction in the expression of specific cell surface molecules on maturing astrocytes results from maturation of a single population of astrocytes and not differential proliferation of a nonexpressing subpopulation of astrocytes, as shown by cell cycle analysis of both immunoreactive and nonimmunoreactive cell populations. These data indicate that during maturation of rat cortical type-I-like astrocytes, the expression of cell surface adhesion molecules is regulated. Furthermore, this regulation appears to be cell autonomous and not dependent on environmental factors. Such regulation of adhesion molecule expression may have profound consequences for the functional properties of astrocytes during CNS maturation.

Mechanism of action of melanocortin peptides. Possible role in astrocyte regulation

Melanocortin peptides exert pleiotropic effect in numerous cell types, controlling processes ranging from adrenal steroidogenesis and melanocyte pigmentation to lacrimation and nerve regeneration. The binding of melanocortins to specific cell surface receptors initiates cellular responses via GTP binding proteins (G-proteins). The affinity of these peptides to the receptor is modulated by extracellular Ca2+ ions, a property unique to melanocortin receptors. In astrocyte cultures derived from the rat brain, melanocortin stimulation elevates cAMP levels that appear to induce morphological changes. However, a transient proliferative response to melanocortins in these cells appears to be cAMP independent. The presence of melanocortin receptors in brain tissue and their unique Ca2+ dependence are discussed in relation to their putative role as regulators of astrocytes.

Immunocytochemical demonstration of early appearing astroglial structures that form boundaries and pathways along axon tracts in the fetal brain

During normal development of the mammalian forebrain, the paired cerebral hemispheres are initially separated midsagittally by the connective tissue-filled longitudinal fissure. During subsequent stages, the hemispheres fuse as basal lamina is remodeled and fibroblasts are eliminated from the fissure to create new central nervous system (CNS) territory in the midline. Two axon pathways, the corpus callosum and dorsal callosal stria, eventually use this region as part of their pathway. In order to assess the possible role of glial cells in the fusion process and in the guidance of axons in this and several other areas of the forebrain, we have analyzed the developing brain in timed cat and mouse embryos with immunohistochemical and morphological techniques. With the use of astroglial-specific antibodies and electron microscopy, we have visualized two distinct, primitive astroglial structures associated with the cerebral midline, and seven more associated with other specific brain regions. The way in which one of these structures moves as a column along the hemispheric midline in synchrony with seam formation suggests the possibility that during morphogenesis of the telencephalon, astrocytes may aid in the fusion process. In addition, the compact assemblage, early appearance and location of this and the other glial structures in relation to well defined neuroanatomical landmarks or axon pathways suggest that they may transiently compartmentalize relatively large regions of the CNS and organize certain developing fiber systems by acting as guides or barriers at critical stages of ontogeny.

Regenerative neurite growth modulation associated with astrocyte proteoglycans

Adherent GFAP-positive cells of neocortical origin in vitro produce and release members of three families of sulphated proteoglycans and a sulphated protein that copurifies with heparan sulphate proteoglycan (HSPG). Conditioned medium (CM) and the proteoglycans contained in the CM have neurite growth-promoting activity when immobilized on defined substrates of growth but not when in the nonimmobilized compartment. On a poly-D-lysine substrate, the rank ordering of specific neurite growth activity based on protein concentration was 330 kDa HSPG >> 100 kDa HSPG/chondroitin sulphate (CS) PG mixture or hybrid > 330 kDa CSPG > 50 kDa CSPG/dermatan sulphate (DS) PG mixture or hybrid and the 31 kDa sulphoprotein. Astrocyte CM lost its growth facilitatory activity when prepared and released by astrocytes in the presence of soluble mediators of inflammation. Loss of activity could not be explained by qualitative or quantitative alterations of released proteoglycans but appeared to be associated with the presence of an inhibitor. The sulphoprotein that copurified with HSPG was a potent inhibitor of HSPG-mediated neurite growth.

Regeneration of dorsal root axons is related to specific non-neuronal cells lining NGF-treated intraspinal nitrocellulose implants

The regeneration of sensory axons from severed dorsal roots can be enhanced by the presence of nerve growth factor (NGF)-treated nitrocellulose strips implanted into an intraspinal lesion cavity. Rather than being directly apposed to the transplant, most regenerating axons are separated from the nitrocellulose by several layers of non-neuronal cells, suggesting that these cells may have a role in the promotion of axonal regrowth. The cellular layers associated with untreated nitrocellulose strips or NGF-treated implants were examined in this study to determine if there were differences in their arrangement or orientation along the implant which might explain some of the possible effects of substrate-bound NGF on axonal regrowth. Into a hemisection lesion cavity created in the adult rat lumbar spinal cord NGF-treated or untreated strips of nitrocellulose were placed vertically, with intact pieces of fetal spinal cord (FSC) tissue transplanted along each side. The distal ends of cut dorsal rootlets were apposed to the fetal tissue. Immunocytochemical and electron microscopic examination 30-60 days post-transplantation revealed a distinct layering of cell types along the NGF-treated strips. Closest to the nitrocellulose was a single layer of macrophages, followed by a separate layer of fibroblasts with dense collagen bundles, then a layer of astroglial cells, before reaching the neuropil of the fetal spinal cord tissue. A thickened basal lamina formed between the fibroblast and astrocytic cell layers and bundles of regenerated sensory axons extended along the interface between these two layers. In contrast, non-neuronal cells along untreated nitrocellulose strips were not as well organized, with an intermixing of fibroblasts and astroglial cells and only scattered macrophage-like cells. Axons rarely were found in conjunction with this mixed population of cells and, overall, fewer regenerated axons extended into transplants with untreated nitrocellulose. The results demonstrate consistent differences in the composition and organization of non-neuronal cells adjacent to NGF-treated nitrocellulose implants, compared to untreated implants. This suggests that the presence of bound NGF influences the recruitment of various cells from the surrounding transplant tissue as well as from the previously injured dorsal rootlets. The capacity for NGF to promote the regeneration of sensory axons may be an indirect effect that is mediated or potentiated by the non-neuronal cell population that gathers in response to the presence of bound NGF.

Divergent effects of astroglial and microglial secretions on neuron growth and survival

Brain glia have a secretory capacity which can modulate neuronal function. Astrocytes release proteins which enhance neuronal survival and induce neuronal growth and differentiation. These effects can be blocked by antagonists of voltage-dependent calcium channels and may be partly mimicked by Bay K 8644, a calcium channel agonist. Two of these neurotrophic proteins appear, on the basis of their physical properties and effects on ciliary ganglion neurons, to be ciliary neurotrophic factor and basic fibroblast growth factor. Activated microglia release a heat- and protease-stable neurotoxin of low molecular weight. This neurotoxicity is blocked by NMDA receptor antagonists. Ciliary neurons exposed to the microglial neurotoxin exhibit an abnormal distribution of neurofilament immunoreactivity, which becomes concentrated in a perinuclear region, while the astroglial growth factors induce neurofilament organization into an extensive neuritic network. The astrocyte-released growth factors can counteract the effect of the microglial neurotoxin and lead to unimpaired neural differentiation in the presence of the neurotoxin.

Time course of dorsal root axon regeneration into transplants of fetal spinal cord: I. A light microscopic study

Cut dorsal root axons regenerate into intraspinal transplants of fetal spinal cord and establish synaptic connections there. The aims of the present study were to describe the progression of dorsal root growth within the transplants and the maturation of transplant morphology and to determine whether the regenerated dorsal root axons persist within the transplants or eventually withdraw. Embryonic (E) day 14 spinal cord was grafted into the lumbar enlargement of adult Sprague-Dawley rats, and the L4 or L5 dorsal root was cut and juxtaposed to the transplants. The morphology of the transplants was examined from 1 day to over 1 year after surgery, and the regenerated dorsal roots were labeled with immunohistochemical methods to study the subset that contains calcitonin gene-related peptide (CGRP). Embryonic spinal cord transplants survived and grew within the host spinal cord in over 90% of the animals. Transplant volume increased and the morphology of the transplants matured over the first 12 weeks and then did not change for 48-60 weeks. During the first week the transplants were composed of dissociated neurons, glia, and hematogenous cells with considerable extracellular space between them. Subsequently, the grafted neurons became densely aggregated, and non-neuronal elements such as inflammatory cells and myelin debris disappeared. CGRP-immunoreactive dorsal roots began to regenerate into the transplants within 24 hours, formed dense bundles by 4 days, and were still present at 60 weeks, the longest survival period examined. Myelination of axons within transplants began at 2 weeks. Quantitative analysis showed that the area of the transplants occupied by CGRP-labeled axons and the distribution area of the labeled axons within the transplants increased until 12 weeks and persisted unchanged for over 48 weeks. These results indicate that regenerated dorsal root axons are permanently maintained within transplants of embryonic spinal cord and suggest that the transplants can contribute to the permanent restoration of damaged intraspinal neural circuits.

Regeneration of axons in the optic nerve of the adult Browman-Wyse (BW) mutant rat

We have studied the regeneration of axons in the optic nerves of the BW rat in which both oligodendrocytes and CNS myelin are absent from a variable length of the proximal (retinal) end of the nerve. In the optic nerves of some of these animals, Schwann cells are present. Axons failed to regenerate in the exclusively astrocytic environment of the unmyelinated segment of BW optic nerves but readily regrew in the presence of Schwann cells even across the junctional zone and into the myelin debris filled distal segment. In the latter animals, the essential condition for regeneration was that the lesion was sited in a region of the nerve in which Schwann cells were resident. Regenerating fibres appeared to be sequestered within Schwann cell tubes although fibres traversed the neuropil intervening between the ends of discontinuous bundles of Schwann cell tubes, in both the proximal unmyelinated and myelin debris laden distal segments of the BW optic nerve. Regenerating axons never grew beyond the distal point of termination of the tubes. These observations demonstrate that central myelin is not an absolute requirement for regenerative failure, and that important contributing factors might include inhibition of astrocytes and/or absence of trophic factors. Regeneration presumably occurs in the BW optic nerve because trophic molecules are provided by resident Schwann cells, even in the presence of central myelin, oligodendrocytes and astrocytes. All the above experimental BW animals also have Schwann cells in their retinae which myelinate retinal ganglion cell axons in the fibre layer. Control animals comprised normal Long Evans Hooded rats, BW rats in which both retina and optic nerve were normal, and BW rats with Schwann cells in the retina but with normal, i.e. CNS myelinated, optic nerves. Regeneration was not observed in any of the control groups, demonstrating that, although the presence of Schwann cells in the retina may enhance the survival of retinal ganglion cells after crush, concomitant regrowth of axons cut in the optic nerve does not take place.

Melanocortins stimulate proliferation and induce morphological changes in cultured rat astrocytes by distinct transducing mechanisms

Melanocyte stimulating hormone (MSH), adrenocorticotropic hormone (ACTH), and several peptides derived from pro-opiomelanocortin, are present in the dorsolateral hypothalamus and arcuate nucleus of several vertebrate species. These peptides affect central nervous system (CNS) functions including behavior, memory, and foetal brain development. In this study we investigated the effects of ACTH1-24, ACTH1-17, ACTH4-10, alpha-MSH, beta-MSH, and a potent analog (Nle4,D-Phe7)-alpha-MSH (melanocortins) on immunocytochemically defined astroglial cells prepared from primary cultures of 1-2-day-old rat brains. A cyclic adenosine 3',5'-monophosphate (cAMP) response to the melanocortins was only detected in astrocytes and not in other cell types in the culture. The extent of the cAMP response was greatest on day 21, the latest time tested. On the other hand, (methyl3H)-thymidine incorporation in astrocytes was significantly stimulated (1.5-2-fold) by melanocortins only in 7 and not in 14 and 21 day cultures. This mitogenic activity of melanocortins was not mimicked by other agents such as forskolin or isoproterenol which efficiently stimulate cAMP production in astrocytes. ACTH1-17 as a melanocortin representative induced significant morphological changes in 7 and 14 day cultures which included rounding of the cell body and process extension. This response, however, resembled that induced by forskolin and hence appears to be cAMP mediated. These findings suggest that astrocytes in the CNS may serve as a target for melanocortins. These peptides appear to affect differentiation and proliferation of these cells during certain developmental periods. While the morphological effects of melanocortins seem to be cAMP mediated, induction of proliferation of the astrocytes by melanocortins appears to involve an alternative signal transduction pathway.

Astroglia in Alzheimer's disease

Amyloid deposits are characteristic of Alzheimer's Disease (AD) and there is growing evidence that amyloid may play an important role in the genesis of this neurodegenerative disease. This review discusses data which suggests that reactive astrocytes and microglia may be a necessary concomitant with amyloid to produce the neuropathology which manifests as AD. Several hypotheses and supporting data for mechanisms by which reactive astrocytes may mediate this neuropathology are presented. These include the possibility that amyloid induces excitotoxicity by interferring with astrocytic glutamate uptake, the possibility that amyloid has this effect via an action on a tachykinin-related receptor and the possibility that proteoglycans released by astrocytes may facilitate the deposition of amyloid plaques. Both symptomatic treatment to enhance cognitive function and treatment to stop the progression of AD are needed. It is hoped that answers to some of the unique questions raised here may provide new insight into the etiology and treatment of AD.

Selective inhibition of neurite outgrowth on mature astrocytes by Thy-1 glycoprotein

THY-1, the smallest member of the immunoglobulin superfamily, is a major cell-surface component expressed by several tissues. The protein, carbohydrate and gene structures of this molecule are known, yet its function is not. It is highly expressed in nervous tissue, where it appears on virtually all neurons after the cessation of axonal growth. Here we show that expression of Thy-1 by a neural cell line inhibits neurite outgrowth on mature astrocytes, but not on other cellular substrata which include Schwann cells and embryonic glia. This inhibition of neurite extension on astrocytes can be reversed by low concentrations (nanomolar) of soluble Thy-1. If a similar interaction between neuronal Thy-1 and astrocytes occurs in vivo, it could stabilize neuronal connections and suppress axonal regrowth after injury in the astrocyte-rich areas of adult central nervous system.

A CNS specific proteoglycan associated with astrocytes in rat optic nerve

A novel CNS specific keratan sulfate proteoglycan is identified by a monoclonal antibody (mAb) TED15. The antibody recognizes a carbohydrate epitope that is removed from the core protein following keratanase digestion. The relative molecular weight of the glycoprotein is not altered following digestion with heparinase, heparitinase or chondroitinase ABC. By immunoblot analysis, the TED15 antigen is differentially distributed in diencephalic derivatives, being present in the optic nerve but absent from the retina. Within the optic nerve, the TED15 antigen is associated with astrocytes, while in the retina no labeling of astrocytes of Müller cells is observed.

The structural integrity of glial scar tissue associated with a chronic spinal cord lesion can be altered by transplanted fetal spinal cord tissue

The potential for fetal spinal cord (FSC) tissue transplants to modify an established glial scar or to restrict the reformation of a scar following surgical manipulation of a chronic lesion site was studied in the injured rat spinal cord. Six to eight weeks after preparation of a hemisection lesion cavity, glial scar tissue was left intact in one group, whereas in a second group it was excised prior to transplantation of a suspension of FSC tissue. From the first group, examination of serial sections through the graft-host interface that had been immunoreacted for glial fibrillary acidic protein (GFAP) demonstrated that in many cases the glial scar no longer was a continuous wall separating the two tissues. Quantitation of the area occupied by these discrete gaps in the scar provided an Index of Fusion, indicating the extent of direct contact between the transplant and host spinal cord. In some animals this constituted as much as 60% of the interface, while in others there were no breaks in the scar (0% fusion). Reinjury of the spinal cord lead to a rapid astrocytic response culminating in the reestablishment of a dense matrix of glial cells and processes covered by a basal lamina. This reformed scar effectively isolated the spinal cord from the external environment of the cavity. When FSC tissue was transplanted after first removing scar tissue the continuity of reformed glial scarring at the graft-host interface was altered. Distinct gaps in the scar appeared randomly along the interface. The mean Index of Fusion for animals receiving a moderate reinjury (removal of scar tissue only) was not as high as for those animals in which a more severe reinjury (expansion of the cavity by 0.5 mm) was performed before transplantation. The extent of graft-host fusion was not significantly improved when scar tissue was removed prior to transplantation. These findings support the hypothesis that the presence of FSC tissue will have an effect on the persistence of glial scar tissue in a chronic lesion site as well as limit the extent to which a new scar is formed in response to a second injury to the spinal cord.

A role for glial cells in activity-dependent central nervous plasticity? Review and hypothesis

Activity-dependent plasticity relies on changes in neuronal transmission that are controlled by coincidence or noncoincidence of presynaptic and postsynaptic activity. These changes may rely on modulation of neural transmission or on structural changes in neuronal circuitry. The present overview summarizes experimental data that support the involvement of glial cells in central nervous activity-dependent plasticity. A role for glial cells in plastic changes of synaptic transmission may be based on modulation of transmitter uptake or on regulation of the extracellular ion composition. Both mechanisms can be initiated via neuronal-glial information transfer by potassium ions, transmitters, or other diffusible factor originating from active neurons. In addition, the importance of changes in neuronal circuitry in many model systems of activity-dependent plasticity is summarized. Structural changes in neuronal connectivity can be influenced or mediated by glial cells via release of growth or growth permissive factors on neuronal activation, and by active displacement and subsequent elimination of axonal boutons. A unifying hypothesis that integrates these possibilities into a model of activity-dependent plasticity is proposed. In this model glial cells interact with neurons to establish plastic changes; while glial cells have a global effect on plasticity, neuronal mechanisms underlie the induction and local specificity of the plastic change. The proposed hypothesis not only explains conventional findings on activity-dependent plastic changes, but offers an intriguing possibility to explain several paradoxical findings from studies on CNS plasticity that are not yet fully understood. Although the accumulated data seem to support the proposed role for glial cells in plasticity, it has to be emphasized that several steps in the proposed cascades of events require further detailed investigation, and several "missing links" have to be addressed by experimental work. Because of the increasing evidence for glial heterogeneity (for review see Wilkin et al., 1990) it seems to be of great importance to relate findings on glial populations to the developmental stage and topographical origin of the studied cells. The present overview is intended to serve as a guideline for future studies and to expand the view of "neuro" physiologists interested in activity-dependent plasticity. Key questions that have to be addressed relate to the mechanisms of release of growth and growth-permissive factors from glial cells and neuronal-glial information transfer. It is said that every complex problem has a simple, logical, wrong solution. Future studies will reveal the contribution of the proposed simple and logical solution to the understanding of central nervous plasticity.

Structure, expression, and function of Ng-CAM, a member of the immunoglobulin superfamily involved in neuron-neuron and neuron-glia adhesion

The neuron-glia cell adhesion molecule (Ng-CAM) mediates neuron-neuron adhesion by a homophilic mechanism and neuron-astrocyte adhesion by a heterophilic mechanism. The protein is expressed on neurons and Schwann cells but not on astrocytes. It is most prevalent during development on cell bodies of migrating neurons and on axons during formation of nerves. Ng-CAM expression is greatly increased following nerve injury. Anti-Ng-CAM antibodies inhibited migration of granule cells along Bergmann glia in cerebellar explants and fasciculation of neurites in outgrowths from explants of dorsal root ganglia. The combined results indicate that Ng-CAM on neurons binds to Ng-CAM on adjacent neurons and to as yet unidentified ligands on astrocytes. Ng-CAM is synthesized in chicken neurons from a 6 kb mRNA as Mr approximately 200,000 forms which are cleaved to yield two components of Mr 135,000 and 80,000. It is glycosylated and can be phosphorylated. Amino acid sequence analysis indicates that it contains six immunoglobulin domains, five fibronectin type III repeats, a transmembrane domain and a cytoplasmic region. Structural analyses indicate that Ng-CAM is most closely related to the mammalian glycoprotein L1 but significant differences between them strongly suggest that they are not equivalent molecules. The recent identification of another structurally related molecule in the chicken called Nr-CAM underscores the notion that these molecules are members of a subfamily of neural cell adhesion molecules within the immunoglobulin superfamily that have related or complementary functions in the nervous system.

Neurite growth modulation associated with astrocyte proteoglycans: influence of activators of inflammation

Of the three classes of sulphated proteoglycans produced by type 1 astrocytes in vitro and released into conditioned medium, only heparan sulphate (HS) was associated with enhanced neurite growth by sensory neurons following pretreatment of a laminin substratum. Astrocyte-conditioned medium (ACM) produced in the presence of certain inflammatory mediators had reduced titers of neurite-promoting activity. The low activity ACM contained inhibitors of neurite growth. Heparan sulphate proteoglycans may modulate neurite growth when complexed to constituents of the extracellular milieu either directly or by interacting with other growth-promoting or growth-inhibitory factors.

Initial stages of retinofugal axon development in the hamster: evidence for two distinct modes of growth

In order to characterize differences in growth patterns of axons as they elongate toward their targets and during the initial stages of terminal arbor formation within the targets, we examined the primary visual system of fetal and newborn hamsters using three morphological methods: the Cajal-deCastro reduced silver method, the rapid Golgi technique, and anterograde transport of HRP. Axons emerge from the retina between the 10th and 11th embryonic days (E10-E11). The front of retinal axons crosses the chiasm, extends over the primitive dorsal nucleus of the lateral geniculate body (LGBd) by E13, and advances to the back of the superior colliculus (SC) by E13.5-E14. The rate of axon growth during this advance is nearly 2 mm/day. Collateral sprouts appear on axons around E15.5. In the LGBd and SC, these sprouts arise from multiple sites along the parent axons. Only one or a few of the sprouts continue to grow and branch, while others are eliminated. The net rate of axon collateral advance in this second phase is an order of magnitude slower than during the stage of axon elongation. Thus, formation of CNS projections may involve two qualitatively distinct modes of axon growth. The arborization mode contrasts with the elongation mode by the presence of branching, a lack of fasciculation and a slower average rate of extension. The stereotypic direct advance of axons during elongation also differs from the remodelling which occurs during arborization. The delay between axon arrival at targets and onset of arborization could be a reflection of axons "waiting" for a maturational change to occur in the retina or in targets. Arborization in the LGBd and SC is initiated around the same time, implicating the former possibility. However, a slower differentiation of retinal arbors in the SC, in addition to morphological differences of arbors in the two structures, suggests that alterations in substrate factors also play a critical role in triggering the early stages of arbor formation.

Astrocyte-derived TGF-beta 2 and NGF differentially regulate neural recognition molecule expression by cultured astrocytes

Because of the importance of neural recognition molecules expressed by glial cells to mediate interactions with neurons, growth factors and cytokines known to be functional during morphogenesis and in diseases of the nervous system were studied for their effects on recognition molecule expression by cultured immature and mature astrocytes from several brain regions. In cultures of immature astrocytes, transforming growth factors-beta 1 (TGF-beta 1) and -beta 2 (TGF-beta 2) and nerve growth factor (NGF) increased expression of the neural adhesion molecule L1, leading to a glia-mediated L1-specific increase in neurite outgrowth of dorsal root ganglion neurons on the astrocyte substrate. L1 expression induced by TGF-beta was inhibited by addition of antibodies to NGF, suggesting that TGF-beta influences L1 expression by modulating production of NGF by astrocytes. TGF-beta 1 and -beta 2 decreased expression of N-CAM by immature astrocytes. Since N-CAM expression was not affected by NGF and antibodies to NGF did not abolish the TGF-beta-induced decrease in N-CAM expression, NGF did not appear to be the mediator for regulating expression of N-CAM. Expression of the adhesion molecule on glia (AMOG) was not affected by any factor. NGF and TGF-beta 2 in latent form, but not TGF-beta 1 were found in the culture supernatants. Addition of interferon-gamma (IFN-gamma), interleukin-1 beta (IL-1 beta), interleukin-6 (IL-6), platelet-derived growth factor (PDGF), or basic fibroblast growth factor (bFGF) to the cultures did not change recognition molecule expression. REcognition molecule expression by mature astrocytes was not found to be modified by any of the factors tested. In view of the observation that levels of L1 and N-CAM expression correlated with the presence of TGF-beta 2 and NGF in the culture supernatants of immature astrocytes, an autocrine regulatory mechanism for recognition molecule expression by these cells is suggested to play a crucial role in regulation of neuron-glia interactions.

Cultured epithelioid astrocytes migrate after transplantation into the adult rat brain

A highly purified population of dividing epithelioid astrocytes has been prepared from postnatal rat corpus callosum. These cells were labelled in culture by incorporation of either [3H]thymidine or fluorescent microspheres and transplanted in a fibrin clot into the hippocampi of adult syngeneic rats. Transplanted cells divided in vivo and progressively migrated into the host brain from the site of implantation up to distances of about 1 mm. After a 1-week survival, transplant cells stained strongly for glial fibrillary acidic protein and had the thick sinuous processes characteristic of stellate astrocytes. Artefactual transfer of radiolabel to host cells was ruled out by control experiments in which either the proportion of transplant cells that were radiolabelled was varied or radiolabelled transplant cells were killed prior to implantation. Astrocyte migration over the first days after implantation was determined to occur at a rate of approximately 100 microns per day. Transplant cells moved into both grey and white matter areas of the host brain and over the migratory period were commonly observed to be associated with blood vessels. Some transplant cells were directly juxtaposed against neuronal perikarya and dendrites. Many labelled astrocytes were located in areas that were apparently completely free of damage caused by implantation. These results define a class of mature astrocytic cells that have the ability to migrate through the adult brain. The existence of pathways for cell movement in the adult CNS has implications for the mechanisms of tissue remodelling after injury and transplantation, for regenerative repair of the CNS, and for the dynamics of cell-cell contacts in the normal adult mammalian brain.

Basic fibroblast growth factor regulates the ability of astrocytes to support hypothalamic neuronal survival in vitro

The putative neurotrophic effects of basic fibroblast growth factor (bFGF) were tested on embryonic hypothalamic neurons in dissociated cell culture. Basic FGF dramatically increased the survival of embryonic hypothalamic astrocytes plated on a poly-L-lysine (PLL) substrate. Basic FGF treatment also increased the number of hypothalamic neurons surviving in vitro; however, no neurotrophic effects were observed when astrocyte proliferation was prevented by using serum-free N2 medium or by using the mitotic inhibitor cytosine arabinoside. In contrast to effects when PLL was used as a substrate, bFGF reduced the survival of hypothalamic neurons plated on a confluent, contact-inhibited monolayer of astrocytes. This effect appears to be due to the direct actions of bFGF on astrocytes: treatment of confluent astrocytes with 5 ng/ml bFGF caused the protoplasmic astrocytes to develop a fibrillar morphology and reduced the ability of the astrocyte monolayer to promote neuronal survival after a further 24 hr in bFGF-free medium. It is concluded that in addition to its mitogenic effects, bFGF acts as a differentiation factor for protoplasmic astrocytes in vitro, and these morphological and functional changes may reflect the process of normal astrocytic development and response to brain injury in vivo.

Immature type-1 astrocytes suppress glial scar formation, are motile and interact with blood vessels

Previous studies have shown that immature but not mature astrocytes have the capacity to suppress glial scar formation and enhance axon outgrowth when transplanted into the adult mouse brain. We report here that glial scar formation is suppressed following transplantation of purified immature but not mature cultured type-1 rat cortical astrocytes into the adult rat brain. To examine the fate of transplanted cells, cultured astrocytes were labeled with either fluorescent beads or BSA-conjugated colloidal gold and traced after transplantation using both light and electron microscopy. While both immature and mature astrocytes survived transplantation, mature astrocytes appeared more susceptible to phagocytosis by cells of the immune system than immature astrocytes. Furthermore, while mature astrocytes were restricted to the region of the implant, immature astrocytes migrated into the surrounding CNS and became closely associated with host blood vessels. Such blood vessels were impermeable to the diffusion of systematically applied Evans blue dye. To determine whether immature astrocytes were intrinsically more motile than mature astrocytes, their rate of translocation was compared in vitro. Immature astrocytes translocated more than twice as fast as mature astrocytes. This ability of immature astrocytes to translocate throughout the host CNS and become associated with blood vessels may be a major factor in their ability to suppress glial scar formation in the adult animal.

Astroglia in CNS injury

The astroglial response to CNS injury is considered in the context of neuron-glial relationships. Although previous models suggested that astroglial cells present in "scars" impede axon regrowth owing to irreversible changes in the glial cell following injury, recent in vivo and in vitro studies indicate that astroglial cells exhibit considerable plasticity, elevating expression of the glial filament protein and altering expression of properties which support axons, including extracellular matrix components and cell surface adhesion systems. Both in vivo and in vitro studies on neuron-glia interactions in different brain regions suggest that glia express region-specific properties, including ion channels, neurotransmitter uptake and receptor systems, and cell surface adhesion systems. Together these findings suggest that a more detailed analysis of glial response to injury in different brain regions will lead to an appreciation of the diversity of the astroglial response to injury, and its regulation by neuron-glia relationships.

Extracellular matrix and neuronal movement

During brain development, both neuronal migration and axon guidance are influenced by extracellular matrix molecules present in the environment of the migrating neuronal cell bodies and nerve fibers. Glial laminin is an extracellular matrix protein which these early brain cells preferentially attach to. Extracellular glycosaminoglycans are suggested to function in restricting neuronal cell bodies and axons from certain brain areas. Since laminin is deposited along the radial glial fibers and along the developing nerve pathways in punctate form, the punctate assemblies may be one of the key factors in routing the developing neurons in vivo. This review discusses the role of laminin in neuronal movement given the present concept of the extracellular matrix molecules and their proposed interactions.

Astrocyte-polymer implants promote regeneration of dorsal root fibers into the adult mammalian spinal cord

To overcome obstacles to the regeneration of crushed dorsal root fibers at the dorsal root entry zone, we have employed specially designed Millipore implants coated with embryonic astrocytes to serve as a substrate for axonal growth. This strategy was successful in promoting the growth of crushed dorsal root axons into the grey matter of the adult mammalian spinal cord in a small number of animals. Fiber ingrowth into the spinal cord was closely associated with the surface of the polymer implant. In addition, unique terminal arbor malformations, not normally present, were seen in several animals. A consistent finding was the presence of a limited inflammatory response in regions immediately adjacent to the implant where axons penetrate the spinal cord. Our findings suggest that providing the dorsal root entry zone with an embryonic milieu can stimulate a limited amount of axonal regeneration into the adult mammalian spinal cord.