Lack of the protein tyrosine phosphatase SHP-1 results in decreased numbers of glia within the motheaten (me/me) mouse brain.

Mice that are homozygous for the autosomal recessive motheaten allele (me/me) lack the protein tyrosine phosphatase SHP-1. Loss of SHP-1 leads to many hematopoietic abnormalities, as well as defects such as infertility and low body weight. However, little is known regarding the role SHP-1 plays in the development of the central nervous system (CNS). To define the role of SHP-1 in CNS development and differentiation, we examined the brains of me/me mice at various times after birth for neuronal and glial abnormalities. Although the brains of me/me mice are slightly smaller than age-matched wild-type littermates, both me/me and wild-type brains are similar in weight, possess an intact blood-brain barrier, and have largely normal neuronal architecture. Significantly, the current study reveals that me/me brain shows decreases in the number of glial fibriallary acidic protein (GFAP)+ astrocytes and F480+ microglia compared with wild-type mice. In addition, decreased immunostaining for the myelin-synthesizing enzyme CNPase was observed in me/me mice, confirming the loss of myelin in these animals, as reported (Massa et al. [2000] Glia 29:376-385). It is particularly significant that there is a decreased number of immunolabeled glia of all subtypes and that this deficit in glial number is not restricted to a particular class of glia. This suggests that SHP-1 is necessary for the normal differentiation and distribution of astrocytes, microglia, and oligendrocytes within the murine CNS.

Neuronal death, not axonal degeneration, results in significant gliosis within the cochlear nucleus of adult chickens.

Injury to the central nervous system initiates a series of events that leads to neuronal cell death and glial activation. Astrocytes respond to damage and disease by becoming hyperplastic and hypertrophied. This 'reactive gliosis' is also accompanied by the upregulation of the intermediate filament protein glial fibrillary acidic protein, the release of growth factors and the formation of the glial scar. However, the signaling cascades which regulate these events, and the molecular mechanisms that give rise to this diverse response, have not been fully elucidated. For example, the role played by degenerating neurons vs. degenerating axons in the activation of astrocytes remains to be determined. To investigate the influence of neuronal cell death vs. axonal degeneration on gliosis, the current study examines the astrocyte response to cochlea removal in two different breeds of adult chickens, one of which exhibits neuronal cell death within the brainstem nucleus magnocellularis (NM) following the lesion and one which does not. Our results indicate that degeneration of NM neurons leads to large increases in both glial proliferation and hypertrophy, while eighth nerve degeneration without NM cell death results in very small increases in glial proliferation.

Tyrosine phosphatase SHP-1 immunoreactivity increases in a subset of astrocytes following deafferentation of the chicken auditory brainstem.

Proliferation of astrocytes is a dramatic response of the central nervous system (CNS) to injury and disease. Such proliferation results in the formation of the neural/glial scar and the reconstitution of the glial limitans. However, not all astrocytes enter the proliferative cycle following injury, and for those that do, the period of cell division is limited. Little attention has focused on the events that regulate the duration and extent of astrocyte proliferation following damage, but clearly control mechanisms are in place as CNS injury does not result in the continuous astrocyte proliferation seen in glial tumorigenesis. Protein tyrosine phosphorylation has been implicated in both astrocyte proliferation and differentiation and plays an important role in the regulation of the cell cycle in a number of different systems. We have found a small subset of astrocytes in the chick auditory brainstem that are immunopositive for the protein tyrosine phosphatase SHP-1. SHP-1 appears to negatively regulate cellular division in the hematopoietic system and is involved in the mitogenic response to various growth factors. Following cochlea removal, there is a marked increase within the auditory brainstem nucleus, nucleus magnocellularis (NM), in both in the number of SHP-1-positive astrocytes and the length of their immunopositive fibers. Significantly, those animals showing the greatest increases in SHP-1 immunoreactivity do not exhibit large amounts of astrocyte proliferation. We hypothesize that the expression of SHP-1 plays a role in negatively regulating the mitotic behavior of astrocytes following deafferentation.

Rapid regulation of cytoskeletal proteins and their mRNAs following afferent deprivation in the avian cochlear nucleus.

During development, removal of neuronal input can lead to profound changes in postsynaptic cells, including atrophy and cell death. In the chicken brainstem cochlear nucleus, the nucleus magnocellularis (NM), deprivation of auditory input via unilateral cochlea removal or silencing the eighth nerve with tetrodotoxin leads to a loss of 25-30% of the neurons and the atrophy of surviving neurons. One intracellular component that may be involved in both cell atrophy and cell death is the cytoskeleton. The degradation of the cytoskeleton following deafferentation could potentially lead to either atrophy or death of NM neurons. However, little is known regarding the role of neuronal input on the cytoskeletal structure of NM neurons and whether changes in the cytoskeleton are responsible for cell death following deafferentation. The present study examined whether changes in the cytoskeleton of NM neurons occurred following cochlea removal. Several components of the cytoskeleton were analyzed following unilateral afferent deprivation. Levels of immunostaining for tubulin, actin, and microtubule-associated protein 2 (MAP-2), and levels of beta-tubulin and beta-actin mRNAs were assessed in NM neurons following cochlea removal. Our results revealed that afferent deprivation results in a rapid decrease in immunostaining for all three cytoskeletal proteins examined. These decreases were observed as early as 3 hours after cochlea removal and persisted for up to 4 days. In addition, these changes occurred in all deafferented NM neurons at the early time points, indicating that both dying and surviving NM neurons undergo a similar change in their cytoskeletons. In contrast to the decreases in immunostaining, levels of beta-tubulin and beta-actin mRNAs were not noticeably altered by deafferentation. Our findings indicate that the cytoskeleton is altered or degraded following deafferentation but that this process is not regulated at the transcriptional level.

Development of Cat-301 immunoreactivity in auditory brainstem nuclei of the gerbil.

The developing brainstem auditory system has been studied in detail by using anatomical and physiological techniques. However, it is not known whether immature auditory neurons exhibit different molecular characteristics than those of physiologically mature neurons. To address this issue, we examined the distribution of Cat-301 immunoreactivity in the developing auditory brainstem of gerbils. Cat-301 is a monoclonal antibody that recognizes a 680-kD chondroitin sulfate proteoglycan similar to aggrecan, a high-molecular-weight chondroitin sulfate proteoglycan found in cartilage. In the central nervous system, Cat-301 immunoreactivity is localized to the extrasynaptic surface of neurons. It has been hypothesized by Hockfield and co-workers (Hockfield et al. [1990a]Cold Spring Harbor Symp. Quart. Biol. 55:504-514) that the Cat-301 proteoglycan is a molecular marker indicating that a neuron has acquired mature neuronal properties. In the current study, Cat-301 staining is first seen at 7 days after birth in the anterior ventral cochlear nucleus (AVCN), the posterior VCN (PVCN), and the medial nucleus of the trapezoid body (MNTB) shortly before the onset of sound-evoked activity. By 21 days after birth, neurons in the AVCN, the PVCN, and the lateral and medial superior olive have attained adult-like distributions of Cat-301 staining concomitant with the physiological maturation of these neurons. Neurons in MNTB attain adult-like distributions of Cat-301 immunoreactivity at 1 year. The maturation of Cat-301 immunoreactivity parallels the physiological maturation of gerbil auditory neurons, and the Cat-301 proteoglycan may play a role in the formation and/or stabilization of auditory synapses.

Neurofilament proteins in avian auditory hair cells.

The distribution of middle-weight neurofilament protein (NF-M), an intermediate filament of neurons, was examined in the developing and mature avian inner ear by using immunocytochemical techniques. NF-M was detected in auditory hair cells and VIIIth cranial nerve neurons. NF-M-positive hair cells are first detected at embryonic day 11 (E11) in superior hair cells in the mid-proximal (mid-frequency) region of the chicken basilar papilla. With time, increasing numbers of hair cells express NF-M. Two developmental gradients occur: 1) a radial gradient, in which superior hair cells are labeled first, and progressively more inferiorly located hair cells are labeled during ontogeny, and 2) a longitudinal gradient, in which hair cells in the mid-proximal region are labeled first, and then progressively more distal (low-frequency) hair cells are labeled. There is also a small proximally directed progression of NF-M expression. By E19, NF-M-positive hair cells are found throughout the distal and mid-proximal regions, and this expression is maintained through 3 weeks posthatching. By 22 weeks posthatching, NF-M staining in hair cells is markedly diminished; staining is seen in only a few tall hair cells in the distal one-fourth of the papilla and in short hair cells in the distal one-half of the papilla. NF-M is never expressed by hair cells at the proximal (high-frequency) end of the papilla at any time examined. These findings suggest that some cell types that have traditionally been classified as nonneural may express neurofilament and that the basilar papilla of the neonatal chicken is not morphologically mature.

Neurofilament spacing, phosphorylation, and axon diameter in regenerating and uninjured lamprey axons.

It has been postulated that phosphorylation of the carboxy terminus sidearms of neurofilaments (NFs) increases axon diameter through repulsive electrostatic forces that increase sidearm extension and interfilament spacing. To evaluate this hypothesis, the relationships among NF phosphorylation, NF spacing, and axon diameter were examined in uninjured and spinal cord-transected larval sea lampreys (Petromyzon marinus). In untransected animals, axon diameters in the spinal cord varied from 0.5 to 50 microns. Antibodies specific for highly phosphorylated NFs labeled only large axons (> 10 microns), whereas antibodies for lightly phosphorylated NFs labeled medium-sized and small axons more darkly than large axons. For most axons in untransected animals, diameter was inversely related to NF packing density, but the interfilament distances of the largest axons were only 1.5 times those of the smallest axons. In addition, the lightly phosphorylated NFs of the small axons in the dorsal columns were widely spaced, suggesting that phosphorylation of NFs does not rigidly determine their spacing and that NF spacing does not rigidly determine axon diameter. Regenerating neurites of giant reticulospinal axons (GRAs) have diameters only 5-10% of those of their parent axons. If axon caliber is controlled by NF phosphorylation via mutual electrostatic repulsion, then NFs in the slender regenerating neurites should be lightly phosphorylated and densely packed (similar to NFs in uninjured small caliber axons), whereas NFs in the parent GRAs should be highly phosphorylated and loosely packed. However, although linear density of NFs (the number of NFs per micrometer) in these slender regenerating neurites was twice that in their parent axons, they were highly phosphorylated. Following sectioning of these same axons close to the cell body, axon-like neurites regenerated ectopically from dendritic tips. These ectopically regenerating neurites had NF linear densities 2.5 times those of uncut GRAs but were also highly phosphorylated. Thus, in the lamprey, NF phosphorylation may not control axon diameter directly through electrorepulsive charges that increase NF sidearm extension and NF spacing. It is possible that phosphorylation of NFs normally influences axon diameter through indirect mechanisms, such as the slowing of NF transport and the formation of a stationary cytoskeletal lattice, as has been proposed by others. Such a mechanism could be overridden during regeneration, when a more compact, phosphorylated NF backbone might add mechanical stiffness that promotes the advance of the neurite tip within a restricted central nervous system environment.

Glial cells of the lamprey nervous system contain keratin-like proteins.

Lamprey axons regenerate following spinal cord transection despite the formation of a glial scar. As we were unable to detect a lamprey homologue of glial fibrillary acidic protein (GFAP), a major constituent of astrocytes, we studied the composition of intermediate filament (IF) proteins of lamprey glia. Monoclonal antibodies (mAbs) were raised to lamprey spinal cord cytoskeletal extracts and these mAbs were characterized by using Western blotting and immunocytochemistry. On two-dimensional (2-D) Western blots, five of the mAbs detected three major IF polypeptides in the molecular weight (MW) range of 45-56 kD. Further studies were conducted to determine the relationship between the lamprey glial-specific antigen and other mammalian IF proteins. Antikeratin 8 antibody recognized two of the three polypeptides. Several of the glial-specific mAbs reacted with human keratins 8 and 18 on Western blots. Keratin-like immunoreactivity was found in all parts of the central and peripheral nervous systems in both larval and adult lampreys. The immunocytochemical staining patterns of glial-specific mAbs were indistinguishable on lamprey spinal cord sections. However, on brain sections, two distinct patterns were observed. A subset of mAbs stained only a few glial fibers in the brain, whereas others stained many more brain glia, particularly the ependymal cells. The former group of mAbs recognized only the two lower MW polypeptides on 2-D Western blots, but the latter group of mAbs recognized all three major IF polypeptides. This correlation is supported by the observation that the highest MW IF polypeptide has an increased level of expression in the brain relative to the spinal cord. Thus, in the lamprey, the glial cells of both spinal cord and brain express molecules similar to simple epithelial cytokeratins, but their IFs may contain these keratins in different stoichiometric proportions. The widespread presence in the lamprey of primitive glial cells containing keratin-like intermediate filaments may have significance for the extraordinary ability of lamprey spinal axons to regenerate.

Identification and purification of a chicken brain neuroglia-associated protein.

The identification, purification, and biochemical characterization of specific markers for neuroglial cells in the central nervous system is an essential step toward a better understanding of the function of glial cells. This manuscript reports the identification and purification of a neuroglia-associated protein (NAP-185) with an apparent molecular mass of 185 kDa. While its expression is not restricted to the brain, it was first identified in a specific subpopulation of glial cells when chick brain stem sections were analyzed with an affinity-purified rabbit antiserum raised against the catalytic domain of the T-cell protein tyrosine phosphatase. This 185-kDa antigen was purified to apparent homogeneity and confirmed to be responsible for the neuroglial staining observed. In spite of its immunological relation to T-cell protein tyrosine phosphatase, purified NAP-185 failed to display tyrosine phosphatase activity. The primary sequence of five NAP-185-derived peptides shows that this protein has not yet been characterized and that it is possibly related to AP180, a clathrin-associated protein.

Astrocyte proliferation in the chick auditory brainstem following cochlea removal.

Astrocytes in the central nervous system (CNS) respond to injury and disease by proliferating and extending processes. The intermediate filament protein of astrocytes, glial fibrillary acidic protein (GFAP) also increases in astrocytes. These cells are called "reactive astrocytes" and are thought to play a role in CNS repair. We have previously demonstrated rapid increases (< 6 hours) in GFAP-immunoreactive and silver-impregnated glial processes in the chick cochlear nucleus, nucleus magnocellularis (NM), following cochlea removal or activity blockade of the eighth nerve. It was not known whether these changes were the result of glial proliferation, glial hypertrophy, or both. The present study examined the time course of astrocyte proliferation in NM following cochlea removal. Postnatal chicks received unilateral cochlea removal and survived for 6, 12, 18, 24, 36, 48, and 72 hours. Bromodeoxyuridine was used to label proliferating cells. The volume and number of labeled cells in NM was calculated for both the experimental and control sides of the brains for experimental animals was well as for unoperated control animals. A subset of astrocytes continuously divide in the normal posthatch chick brainstem. The percentage of labeled nuclei increases within NM 36 hours following cochlea removal and is robust by 48 hours. This increase is due to astrocyte proliferation within, rather than migration to, NM. These results indicate that rapid increases in GFAP following reduced activity are independent of cell proliferation. The time course of astrocyte proliferation suggests that cellular degeneration within the nucleus may play a role in upregulating astrocyte proliferation.

Structure of reticulospinal axon growth cones and their cellular environment during regeneration in the lamprey spinal cord.

The large larval sea lamprey is a primitive vertebrate that recovers coordinated swimming following complete spinal transection. An ultrastructural study was performed in order to determine whether morphologic features of regenerating axons and their cellular environment would provide clues to their successful regeneration compared to their mammalian counterparts. Three larval sea lampreys were studied at 3, 4 and 11 weeks following complete spinal transection and compared with an untransected control. Müller and Mauthner cells or their giant reticulospinal axons (GRAs) were impaled and injected with horseradish peroxidase (HRP). Alternating thick and thin sections were collected for light and electron microscopy. A total of 9 neurites were examined. At all times, growth cones of GRAs differed from those of cultured mammalian neurons in being packed with neurofilaments and in lacking long filopodia, suggesting possible differences in the mechanisms of axon outgrowth. Morphometric analysis suggested that GRA growth cones contact glial fibers disproportionately compared to the representation of glial surface membranes in the immediate environment of these growth cones. No differences were found between glial cells in regenerating spinal cords and those of untransected control animals with regard to the size of the cell body and nucleus and the packing density of their intermediate filaments. Glial fibers in control animals and glial fibers located far from a transection were oriented transversely. Glial cells adjacent to the transection site sent thickened, longitudinally oriented processes into the blood clot at the transection site. These longitudinal glial processes preceded the regenerating axons. Desmosomes were observed on glia adjacent to the lesion but were scarce in the lesion during the first four weeks post-transection. These findings suggest that longitudinally oriented glial fibers may serve as a bridge along which axons can regenerate across the lesion. The presence of desmosomes might prevent migration of astrocytes near the transection, thus stabilizing the glial bridge.

A biphasic change in ribosomal conformation during transneuronal degeneration is altered by inhibition of mitochondrial, but not cytoplasmic protein synthesis.

Following loss of eighth nerve input, 20-40% of neurons in the neonatal chick cochlear nucleus, nucleus magnocellularis (NM), undergo cell death. Intracellular changes that precede the death of NM neurons include increased oxidative metabolism and mitochondrial volume, decreased cytoplasmic protein synthesis, and destruction of ribosomes. Six hours following afferent deprivation, dying NM neurons demonstrate complete loss of ribosomes and cessation of protein synthesis, suggesting that the rapid destruction of ribosomes leads to neuronal death. Increased NM neuron death occurs when mitochondrial upregulation is prevented by chloramphenicol, a mitochondrial protein synthesis inhibitor. This finding suggests that increased oxidative capacity is required for neuronal survival following loss of afferent input. To study changes in the ribosomes of afferent-deprived NM neurons, we obtained a monoclonal antibody to ribosomal RNA. This monoclonal antibody, Y10B, labels ribosomes of all NM neurons receiving normal synaptic activity. Following removal of afferent input, NM neurons demonstrate a biphasic change in their pattern of Y10B label. During the initial phase, there is a uniform decrease in the density of Y10B label. In the second phase, some NM neurons recover the capacity to bind the Y10B antibody while others remain unlabeled. During this second phase, NM neurons putatively destined to die, based on their failure to synthesize protein, are unlabeled by the Y10B antibody. New gene expression is not necessary to initiate the change in ribosomal immunoreactivity that leads deafferented NM neurons toward cell death. Blocking cytoplasmic protein synthesis with cycloheximide had no effect on the biphasic change in Y10B labeling of afferent-deprived NM neurons. Treating chicks with chloramphenicol, however, prevented the recovery of Y10B immunoreactivity in NM neurons during the second phase of the response to afferent deprivation.

Preferential regeneration of spinal axons through the scar in hemisected lamprey spinal cord.

Axons of lamprey spinal cord can regenerate across a complete spinal transection. Thus, unlike the scar of injured mammalian spinal cords, the scar in the lamprey is not an absolute impediment to regeneration. However, it is still not known whether the scar is a relative impediment or whether it provides a favorable environment for regeneration compared to the spinal cord parenchyma. In order to answer this question, the cords of 12 large larval sea lampreys (4-5 years old) were hemisected at the level of the third gill and the animals allowed to recover for 10 weeks. The large reticulospinal neurons (Müller and Mauthner cells) or their giant axons were injected intracellularly with HRP and their regenerating neurites visualized in central nervous system (CNS) wholemounts. Forty-five of seventy-one regenerating neurites (64%) grew beyond the level of the hemisection. Of these, 36 (82%) regenerated through the scar and remained on the same side of the cord as their parent axons, while only 8 (18%) crossed the midline and grew around the scar. Thus, regenerating neurites of giant reticulospinal axons tended to grow through the hemisection scar rather than around it. Once they passed the level of injury, they continued to elongate in their appropriate paths. It is possible that this tendency for axons to regenerate through the scar reflects the greater amount of empty spaces on the hemisected side. In order to rule this out, 13 animals received contralateral simultaneous hemisections at the level of the 3rd and 7th gills. This procedure created large numbers of degenerating axons and potential empty spaces both rostral and caudal to the scars within both hemicords; 92 of 158 neurites (58%) regenerated beyond the level of their respective hemisections. All of these grew through the scar and none crossed to the contralateral side. Distal to either hemisection, neurites remained on their correct side regardless of whether the contralateral cord contained normal CNS parenchyma or axonal debris and empty spaces produced by Wallerian degeneration. Moreover, in hemisected and double hemisected animals, as well as in completely transected control animals, neurites regenerating in their correct direction grew further than those that were misrouted. Because lamprey spinal axons grow preferentially through a scar rather than around it, the scar may play a positive role in supporting axonal regeneration.

Axonal regeneration in the adult lamprey spinal cord.

Larval sea lampreys recover from complete spinal transection by a process involving directionally specific axonal regeneration. In order to determine whether this is also true of adults, 14 adult lampreys were transected at the level of the 5th gill and allowed to recover for 10 weeks. Müller and Mauthner cells and their giant reticulospinal axons (GRAs) were impaled with microelectrodes and injected with horseradish peroxidase (HRP). The tissue was processed for HRP histochemistry and wholemounts of brain and spinal cord were prepared. All animals recovered coordinated swimming; 61 of 121 (50%) neurites emanating from 30 axons regenerated caudal to the scar into the distal stump. Of the neurites which had grown beyond the scar, 92% were correctly oriented, i.e., caudalward and ipsilateral to the parent axon. Retransection in two additional animals eliminated the recovered swimming. Thus, behavioral recovery in adult sea lampreys is accompanied by directionally specific axonal regeneration.

The need for cellular elements during axonal regeneration in the sea lamprey spinal cord.

Spinal axons in the larval sea lamprey regenerate following a complete spinal transection. It is not known whether regenerating growth cones require contact with cellular elements or whether the basement membrane and collagenous meninx primitiva which surround the spinal cord are sufficient for neurite out-growth. To determine this, a freeze lesion was made which severed axons, destroyed neuronal perikarya, and greatly reduced the number of glial cells. After at least 10 weeks of recovery, 50 neurites from 31 Müller and Mauthner axons were labeled by intracellular injection of HRP. Eighty-six percent of these neurites did not regenerate into the lesion site. No neurites grew through the lesion. No animals recovered coordinated swimming. These results suggest that glial and/or neuronal surfaces are required for axonal regeneration. Moreover, a monolayer of glial cells appears to be suboptimal and a three-dimensional matrix of cells may be necessary to promote regeneration in the lamprey spinal cord.