Processes Of Reactive Astrocytes Research Articles

A facilitatory role of astrocytes in axonal regeneration after acute and chronic spinal cord injury

Neuroscience dogma avers that astrocytic “scars” inhibit axonal regeneration after spinal cord injury (SCI). A recent report suggested however that astrocytes form “borders” around lesions that are permissive rather than inhibitory to axonal growth. We now provide further evidence supporting a facilitatory role of astrocytes in axonal regeneration after SCI. First, even 6months after SCI, injured axons are retained within regions of densely reactive astrocytes, in direct contact with astrocyte processes without being repelled. Second, 6 month-delayed implants of neural stem cells extend axons into reactive astrocyte borders surrounding lesions, densely contacting astrocyte surfaces. Third, bioengineered hydrogels implanted into sites of SCI re-orient reactive astrocytic processes to align along the rostral-to-caudal spinal cord axis resulting in successful regeneration into the lesion/scaffold in close association with astrocytic processes. Fourth, corticospinal axons regenerate into neural progenitor cells implanted six months after injury in close association with host astrocytic processes. Thus, astrocytes do not appear to inhibit axonal regeneration, and the close association of newly growing axons with astrocytic processes suggests a facilitatory role in axonal regeneration.

Upregulation of Annexin A1 in Reactive Astrocytes and Its Subtle Induction in Microglia at the Boundaries of Human Brain Infarcts

Annexin A1 (ANXA1) has multiple functions, including anti-inflammatory effects, and is thought to be neuroprotective in various pathophysiologies of the central nervous system. The importance of ANXA1 in microglia and endothelial cells in ischemic environments in the brain has been recognized, but its detailed behavior in astrocytes in the ischemic brain remains unknown. Using immunohistochemistry, we therefore assessed the altered distribution of ANXA1 in human brain infarcts using 14 autopsied samples and 18 surgical samples. Elevated expression of ANXA1 was observed in reactive astrocytes in peri-infarct regions. ANXA1 accumulated at the cell periphery and in swollen cytoplasmic processes of reactive astrocytes, as well as at the rim of vacuoles at the boundary of necrosis, and colocalized with aberrantly distributed aquaporin 4 and excitatory amino acid transporter 1. Foamy macrophages in the necrotic core also expressed abundant ANXA1, whereas resident microglia at the boundary of necrosis rarely showed intrinsic expression of ANXA1. This characteristic distribution of ANXA1 in human brain infarcts may represent the good adaptability of reactive astrocytes to ischemic damage.

Increased Expression of Slit2 and its Robo Receptors During Astroglial Scar Formation After Transient Focal Cerebral Ischemia in Rats.

Slit2, a secreted glycoprotein, has recently been implicated in the post-ischemic astroglial reaction. The objective of this study was to investigate the temporal changes and cellular localization of Slit2 and its receptors, Robo1, Robo2, and Robo4, in a rat transient focal ischemia model induced by middle cerebral artery occlusion. We used double- and triple-immunolabeling to determine the cell-specific changes in Slit2 and its receptors during a 10-week post-ischemia period. The expression profiles of Slit2 and the Robo receptors shared overlapping expression patterns in sham-operated and ischemic striatum. Constitutive expression of Slit2 and Robo receptors was observed in striatal neurons with weak intensity, whereas in rats reperfused after ischemic insults, these immunoreactivities were increased in reactive astrocytes. Astroglial induction of Slit2 and Robo in the peri-infarct region was distinct on days 7-14 after reperfusion and thereafter increased progressively throughout the 10-week experimental period. Slit2 and Robo were prominently expressed in the perinuclear cytoplasm and main processes of reactive astrocytes forming the astroglial scar. This observation was confirmed by quantification of the mean fluorescence intensity of Slit2 and Robo receptors over reactive astrocytes localized at the edge of the infarct area. However, activated microglia/macrophages in the peri-infarct area were devoid of any specific labeling for Slit2 and Robo. Thus, our data revealed a selective and sustained induction of Slit2 and Robo in astrocytes localized throughout the astroglial scar after ischemic stroke, suggesting that Slit2/Robo signaling participates in glial scar formation and brain remodeling following ischemic injury.

‘Hit & Run’ Model of Closed-Skull Traumatic Brain Injury (TBI) Reveals Complex Patterns of Post-Traumatic AQP4 Dysregulation

Cerebral edema is a major contributor to morbidity associated with traumatic brain injury (TBI). The methods involved in most rodent models of TBI, including head fixation, opening of the skull, and prolonged anesthesia, likely alter TBI development and reduce secondary injury. We report the development of a closed-skull model of murine TBI, which minimizes time of anesthesia, allows the monitoring of intracranial pressure (ICP), and can be modulated to produce mild and moderate grade TBI. In this model, we characterized changes in aquaporin-4 (AQP4) expression and localization after mild and moderate TBI. We found that global AQP4 expression after TBI was generally increased; however, analysis of AQP4 localization revealed that the most prominent effect of TBI on AQP4 was the loss of polarized localization at endfoot processes of reactive astrocytes. This AQP4 dysregulation peaked at 7 days after injury and was largely indistinguishable between mild and moderate grade TBI for the first 2 weeks after injury. Within the same model, blood-brain barrieranalysis of variance permeability, cerebral edema, and ICP largely normalized within 7 days after moderate TBI. These findings suggest that changes in AQP4 expression and localization may not contribute to cerebral edema formation, but rather may represent a compensatory mechanism to facilitate its resolution.

Prenatal Hypoxic-Ischemic Insult Changes the Distribution and Number of NADPH-Diaphorase Cells in the Cerebellum

Astrogliosis, oligodendroglial death and motor deficits have been observed in the offspring of female rats that had their uterine arteries clamped at the 18th gestational day. Since nitric oxide has important roles in several inflammatory and developmental events, here we evaluated NADPH-diaphorase (NADPH-d) distribution in the cerebellum of rats submitted to this hypoxia-ischemia (HI) model. At postnatal (P) day 9, Purkinje cells of SHAM and non-manipulated (NM) animals showed NADPH-d+ labeling both in the cell body and dendritic arborization in folia 1 to 8, while HI animals presented a weaker labeling in both cellular structures. NADPH-d+ labeling in the molecular (ML), and in both the external and internal granular layer, was unaffected by HI at this age. At P23, labeling in Purkinje cells was absent in all three groups. Ectopic NADPH-d+ cells in the ML of folia 1 to 4 and folium 10 were present exclusively in HI animals. This labeling pattern was maintained up to P90 in folium 10. In the cerebellar white matter (WM), at P9 and P23, microglial (ED1+) NADPH-d+ cells, were observed in all groups. At P23, only HI animals presented NADPH-d labeling in the cell body and processes of reactive astrocytes (GFAP+). At P9 and P23, the number of NADPH-d+ cells in the WM was higher in HI animals than in SHAM and NM ones. At P45 and at P90 no NADPH-d+ cells were observed in the WM of the three groups. Our results indicate that HI insults lead to long-lasting alterations in nitric oxide synthase expression in the cerebellum. Such alterations in cerebellar differentiation might explain, at least in part, the motor deficits that are commonly observed in this model.

Intracerebroventricular Shiga toxin 2 increases the expression of its receptor globotriaosylceramide and causes dendritic abnormalities

Neurological damage caused by intoxication with Shiga toxin (Stx) from enterohemorrhagic Escherichia coli is the most unrepairable and untreatable outcome of Hemolytic Uremic Syndrome, and occurs in 30% of affected infants. In this work intracerebroventricular administration of Stx2 in rat brains significantly increased the expression of its receptor globotriaosylceramide (Gb 3) in neuronal populations from striatum, hippocampus and cortex. Stx2 was immunodetected in neurons that expressed Gb 3 after intracerebroventricular administration of the toxin. Confocal immunofluorescence of microtubule-associated protein 2 showed aberrant dendrites in neurons expressing increased Gb 3. The pro-apoptotic Bax protein was concomitantly immunodetected in neurons and other cell populations from the same described areas including the hypothalamus. Confocal immunofluorescence showed that Gb 3 colocalized also with glial fibrillary acidic protein only in reactive astrocytic processes, and not in vehicle-treated normal ones. Rats showed weight variation and motor deficits as compared to controls. We thus suggest that Stx2 induces the expression of Gb 3 in neurons and triggers neuronal dysfunctions.

Role of GFAP in morphological retention and distribution of reactive astrocytes induced by scrapie encephalopathy in mice

We have previously demonstrated that mutant mice bearing astrocytes deficient in glial fibrillary acidic protein (GFAP) exhibited typical spongiform degeneration and prion plaque deposition. However, it remains to be determined whether there are astrocyte-specific alterations in the reactive response of astrocytes. Herein, we analyzed morphological features of Gfap − / − reactive astrocytes. Light microscopic morphometry of mutant reactive astrocytes revealed reduced outlined cell area and shorter distances among expanded cell space but with larger nuclei. Electron microscopy revealed mutant cells containing very few and sparse glial filaments as well as abnormal cytoarchitecture of reactive astrocytic processes. Furthermore, paired cell formation appeared frequently. The results suggest that GFAP is necessary for morphological retention and distribution of reactive astrocytes during prion disease, and that there is a GFAP-dependent function of glial filaments in reactive astrocytes.

Perineuronal nets are largely unaffected in Alzheimer model Tg2576 mice

Changes in the molecular organization of the extracellular matrix are key factors in neuropathology. We investigated aggrecan-based perineuronal nets (PNs) in relation to neurodegeneration and activation of glial cells in a transgenic mouse (Tg2576) model of Alzheimer's disease. The formation of amyloid plaques in the cerebral cortex occurred independently of the area-specific distribution of PNs. Matrix components were only affected in the core of plaques in advanced stages of pathology. PNs remained unchanged in the large marginal zone occupied by reactive astrocytic processes. We conclude that the aggrecan-based extracellular matrix of PNs is not enzymatically altered in peripheral plaque territories and is only removed after neuronal death.

Reactive Astrocytes Expressing Intense Estrogen Receptor-alpha Immunoreactivities Have Much Elongated Cytoplasmic Processes: An Autopsy Case of Human Cerebellar Tissue with Multiple Genitourinary and Gastrointestinal Anomalies

We performed an immunohistochemical study on the estrogen receptor alpha (ER-α) distribution in the cerebellum of a human neonate with multiple congenital anomalies, that had been acquired during autopsy. Although the exact pathology in the brain was not clearly elucidated in this study, an unidentified stressful condition might have induced the astrocytes into reactive states. In this immunohistochemical study on the neonatal cerebellum with multiple congenital anomalies, intense ER-α immunoreactivities (IRs) were localized mainly within the white matter even though ER-α IRs were known to be mainly localized in neurons. Double immunohistochemical staining showed that ER-α IR cells were reactive astrocytes, but not neurons. Interestingly, there were differences in the process length among the reactive astrocytes showing ER-α IRs. Our quantitative data confirmed that among the glial fibrillary acidic protein (GFAP)-expressing reactive astrocytes, the cells exhibiting intense ER-α IRs have much longer cytoplasmic processes and relatively weaker GFAP IRs. Taken together, the elongated processes of reactive astrocytes might be due to decreased expression of GFAP, which might be induced by elevated expression of ER-α even though the elucidation of the exact mechanism needs further studies.

Allogeneic bone marrow stromal cells promote glial–axonal remodeling without immunologic sensitization after stroke in rats

We evaluated the effects of allogeneic bone marrow stromal cell treatment of stroke on functional outcome, glial–axonal architecture, and immune reaction. Female Wistar rats were subjected to 2 h of middle cerebral artery occlusion. Rats were injected intravenously with PBS, male allogeneic ACI – or syngeneic Wistar –bone marrow stromal cells at 24 h after ischemia and sacrificed at 28 days. Significant functional recovery was found in both cell-treated groups compared to stroke rats that did not receive BMSCs, but no difference was detected between allogeneic and syngeneic cell-treated rats. No evidence of T cell priming or humoral antibody production to marrow stromal cells was found in recipient rats after treatment with allogeneic cells. Similar numbers of Y-chromosome + cells were detected in the female rat brains in both groups. Significantly increased thickness of individual axons and myelin, and areas of the corpus callosum and the numbers of white matter bundles in the striatum were detected in the ischemic boundary zone of cell-treated rats compared to stroked rats. The areas of the contralateral corpus callosum significantly increased after cell treatment compared to normal rats. Processes of astrocytes remodeled from hypertrophic star-like to tadpole-like shape and oriented parallel to the ischemic regions after cell treatment. Axonal projections emanating from individual parenchymal neurons exhibited an overall orientation parallel to elongated radial processes of reactive astrocytes of the cell-treated rats. Allogeneic and syngeneic bone marrow stromal cell treatment after stroke in rats improved neurological recovery and enhanced reactive oligodendrocyte and astrocyte related axonal remodeling with no indication of immunologic sensitization in adult rat brain.

Alginate enhances elongation of early regenerating axons in spinal cord of young rats.

Freeze-dried alginate sponge cross-linked with covalent bonds has been demonstrated to enhance nerve regeneration in peripheral nerves and spinal cords. The present study examined, at early stages after surgery, the outgrowth of regenerating axons and reactions of astrocytes at the stump of transected spinal cord in young rats. Two segments (Th7-8) were resected, and alginate was implanted in the lesion. As controls, collagen gel was implanted in place of alginate or the lesion was left without implantation. Two and 4 weeks after surgery, nerve outgrowth and astrocyte reactions were examined. Many regenerating axons, some of which were accompanied by astrocytic processes, were found to extend from the stump into the alginate-implanted lesion. In the all nonimplanted animals, large cystic cavities were formed at both interfaces with no definite axonal outgrowth into the lesion. In collagen-implanted animals, cavity formation was found in some rats, and regenerating axons once formed at the stumps did not extend further into the lesion. Astrocytic processes extending into alginate-implanted lesion had no basal laminae, whereas those found in control experiments were covered by basal laminae. These findings suggest that alginate contributed to reducing the barrier composed of connective tissues and reactive astrocytic processes, and served as a scaffold for the outgrowth of regenerating axons and elongation of astrocytic processes.

Short-term ethanol exposure alters calbindin D28k and glial fibrillary acidic protein immunoreactivity in hippocampus of mice

The effects of a short-term ethanol treatment on hippocampus have been studied in mice exhibiting intoxication signs. The alterations of neurons and astrocytes as well as quantitative changes of calbindin D28k-immunoreactivity and glial fibrillary acidic protein-immunoreactivity (GFAP-IR) in selected regions of the dorsal hippocampus were examined using anti-calbindin and anti-GFAP monoclonal anti-body (mAb), respectively. The administration of 6% (v/v) ethanol during first week led to the neuronal death and decrease of the total number of calbindin-IR neurons in the examined brain regions. Moreover, the calbindin positive neurons were shown to have diminished processes following short-term ethanol exposure. These neuronal changes were associated with the increase of the GFAP-IR astrocytes. Hypertrophy of cell bodies and cytoplasmic processes of reactive astrocytes were also seen. In addition, dense, thick and highly-stained GFAP-IR cells with long processes in granular cell layer appeared entering into molecular layer of dentate gyrus. In agreement with the discrepancy percentage of neuronal cell loss and increase of reactive astrocytes detected by calbindin and GFAP-IR using image quantitative analysis, the regional differences in the vulnerability to the neurotoxic effects following short-term ethanol exposure were found: CA3>CA2>CA1>DG. These findings also illustrate the importance of correlation between calbindin and GFAP-IR when determining the morphological alteration of neuron and astroglial following short-term ethanol treatment. The disruption of calbindin and GFAP could affect neuronal-astroglial interaction, resulting in disturbance of behaviors dependent on hippocampus.

Astrocyte-apolipoprotein E associations in senile plaques in Alzheimer disease and vascular lesions: a regional immunohistochemical study.

While apolipoprotein E (ApoE) and beta-amyloid (A beta) co-localize in senile plaques in cortex and cerebellum in Alzheimer disease (AD), the A beta-positive, predominantly diffuse plaques in the striatum do not exhibit ApoE immunoreactivity regardless of disease duration. As astrocytes are a major source of ApoE in the brain, we investigated potential regional differences in the ability of astrocytes to produce ApoE that might affect A beta processing and the progression of AD pathology. Using antibodies to ApoE, glial fibrillary acidic protein (GFAP), and A beta, we compared the pattern of immunoreactivity in senile plaques in AD autopsy tissue with that of reactive astrocytes surrounding subacute and old infarcts in both AD and non-AD cases. We found GFAP and ApoE immunoreactivity, but not A beta label in cell bodies and processes of reactive astrocytes in zones of infarction within cerebral cortex, striatum, and cerebellum, indicating that astrocytes are capable of upregulating ApoE within these 3 regions. In contrast, while astrocytes surrounded many neocortical neuritic plaques in AD, these GFAP-positive cells failed to label with ApoE; instead. ApoE label within plaques paralleled that of A beta. As expected, neither the ApoE-negative diffuse plaques of the striatum nor the ApoE-immunopositive diffuse plaques of the cerebellum were clearly associated with GFAP-immunoreactive astrocytes. The apparent absence of ApoE label in cortical plaque-associated astrocytes may signify a regulatory mechanism affecting ApoE synthesis and secretion, influenced by binding of ApoE to fibrillar amyloid within the plaques, neuritic changes, or other factors.

Rubral astrocytic reactions to proximal and distal axotomy of rubrospinal neurons in the rat

Spinal tractotomy-induced perineuronal astrocytic reaction of the rat rubrospinal system was studied using an antiserum to the astrocyte-specific glial fibrillary acidic protein as a marker. The effect of the proximity of axonal cut to cell bodies was also studied by comparing astrocytic reactions elicited by upper cervical and lower thoracic tractotomy. Fast blue was used as a retrograde tracer to identify axotomized neurons, which were found to concentrate in the caudal part of the contralateral red nucleus. The length of reactive astrocytic processes in the dorsomedial and ventrolateral parts of the nucleus was quantified separately since neurons in these two parts project to cervical and lumbar spinal cord, respectively. Those of the ipsilateral nucleus were also quantified. Sham operation caused a transient increase in reactive astrocytic processes one day after surgery. An early and a late increase of reactive astrocytic processes was found 2–5 days and 2–8 weeks following both thoracic and cervical tractotomy. Cervical axotomy of lumbar-cord-projecting rubral neurons caused an increase of reactive astrocytic processes similar in magnitude to that generated by thoracic axotomy. Following thoracic axotomy, the uninjured dorsomedial area of the contralateral nucleus also displayed an increase concomitant with that which occurred within the neighboring, injured ventrolateral nuclear area suggesting the action of diffusible factor(s). Surprisingly, cervical and thoracic tractotomy also elicited a similar increase of reactive astrocytic processes in the ipsilateral nuclei, independent of the number of ipsilaterally projected neurons present in each nucleus. This may be attributed to the retrograde influence from the denervated spinal target sites which were carried by fibers of the intact rubrospinal tract known to terminate bilaterally. In the lesioned nucleus, reactive astrocytic processes were often located close to axotomized cell bodies as early as 3 days following upper cervical and also, to a lesser extent, lower thoracic tractotomy. However, reactive astrocytic processes in the ipsilateral nucleus usually remained in the neuropil. These results suggest that axotomy induces two levels of retrograde astrocytic reactions within the soma area of intrinsic central neurons. Reactive astrocytic processes located proximally to axotomized cell bodies may have a different functional role from those distributed in the neuropil.

Immunoreactivity of growth inhibitory factor in normal rat brain and after stab wounds — an immunocytochemical study using confocal laser scan microscope

The growth inhibitory factor (GIF) is a new member of the metallothionein family that is downregulated in Alzheimer's disease brain. Using a confocal laser scan microscope with polyclonal and monoclonal antibodies to GIF, and monoclonal antibodies to glial fibrillary acidic protein (GFAP) and MAP-2, we demonstrated that GIF immunoreactivity was expressed primarily in astrocytes and much less in neurons. In astrocytes of normal rat brain GIF immunoreactivity was detected mainly in the cell bodies, while GFAP immunoreactivity was detected mainly in the processes. GIF immunoreactivity was more strongly expressed in reactive astrocytes. These findings were confirmed with both polyclonal and monoclonal antibodies. Following stab wounds, a number of GIF-positive reactive astrocytes were detected around the wounds at 3 days postoperation. After 7 days GIF immunoreactivity was detected in cell bodies and processes of reactive astrocytes. The number of GIF-positive astrocytes and the intensity of the immunoreactivity remained elevated over the control levels at least through 28 days. These immunocytochemical findings correlated well with changes in GIF protein and mRNA levels. Not only changes in GIF protein and mRNA levels but also intracellular localization of GIF in normal rat brain and after stab wounds in rat brain were different from those of GFAP. These results support the concept that GIF plays an important role in the processing of reconstruction after brain damage.

β-Amyloid-induced changes in cultured astrocytes parallel reactive astrocytosis associated with senile plaques in Alzheimer's disease

One neuropathological characteristic of Alzheimer's disease is an abundance of reactive astrocytes, particularly in association with senile plaques. Neither the factor(s) responsible for initiating the reactive astrocytosis nor the effects of this event on disease progression are known. We investigated the possibility that β-amyloid protein, the primary constituent of plaques, contributes to reactive astrocytosis by comparing results derived from both culture studies and immunohistochemical analyses of Alzheimer brain tissue. We report that β-amyloid peptides, in an aggregation-dependent manner, rapidly induce a reactive phenotype in cultured rat astrocytes. Reactive morphological changes are accompanied by increased immunoreactivities for glial fibrillary acidic protein and basic fibroblast growth factor. Although toxic to other types of central nervous system cells, aggregated β -amyloid peptides do not significantly decrease astrocyte viability. Rather, the processes of cultured astrocytes envelop aggregated deposits of β-amyloid peptide. In Alzheimer brain, the processes of reactive astrocytes were also observed to engulf β-amyloid deposits. Similar to the in vitro findings, the astrocytic response was associated only with β-amyloid plaques exhibiting an aggregated structure. Further, the plaque-associated reactive astrocytes showed enhanced immunoreactivities for glial fibrillary acidic protein and basic fibroblast growth factor. These data suggest that β-amyloid which has assembled into β -sheet fibrils significantly contributes to the reactive astrocytosis characteristic of Alzheimer's disease. Thus, in addition to its hypothesized direct effects on neuronal viability, β-amyloid may also influence disease progression indirectly via reactive astrocytosis.

Reactive astrocytes in viral infections of the human brain express endothelin-like immunoreactivity

In order to investigate the expression of endothelin-like immunoreactivity in astrocytes of viral infections of the human brain the avidin-biotin peroxidase complex method and a polyclonal antiserum were used. Autopsy material was obtained from 5 cases of herpes simplex encephalitis, two of progressive multifocal leukoencephalopathy (PML) and two of subacute sclerosing panencephalitis (SSPE). All the 5 herpes simplex encephalitis cases presented groups of immunoreactive astrocytes around necrotic, inflammatory lesions. The PML cases exhibited a large number of immunoreactive astrocytes in and around lesions of the white matter. The cases of SSPE disclosed numerous, markedly stained fibrillary immunoreactive astrocytes; they were most abundant in degenerated regions of the white matter. The processes and peripheral cytoplasm of giant astrocytes in the PML cases contained immunoreactive material but the perinuclear region was devoid of such material. In the herpes simplex and the SSPE cases immunoreactivity was present throughout the cytoplasm and processes of reactive fibrillary astrocytes. Many nerve cells in the cerebral cortex, hippocampus, cerebellum and pons of control cases exhibited endothelin-like immunoreactivity but this occurred in only exceptional astrocytes of control cases. Endothelin-like immunoreactivity was not present in the oligodendrocytes and vascular endothelial cells of controls and cases of virus infection. The expression of endothelin-like immunoreactivity in astrocytes in human viral diseases reflects probably an increased intracellular content of endothelin. If this peptide is released from such astrocytes, it may act as a mitogen and by inducing constriction of arterioles it may influence the microcirculation.

Response of striatal astrocytes to neuronal deafferentation: An immunocytochemical and ultrastructural study

This ultrastructural and light microscopic immunocytochemical study describes the time course of anatomical changes that occur in striatal astrocytes in response to neuronal deafferentation in young adult rats and the coordinate distribution of two astrocytic proteins involved in reactive synaptogenesis, glial fibrillary acidic protein and clusterin. We found that following a unilateral lesion of the cerebral cortex, striatal astrocytes undergo a rapid ultrastructural transformation from a protoplasmic to a reactive type of astroglia and are the primary cells involved in the removal of degenerating axon terminals, but not axons of passage, from the neuropil. In addition, at 10 and 27 days postlesion, processes of reactive astrocytes are also seen to occupy vacant postsynaptic spines after degenerating presynaptic terminals are removed, suggesting that they may also participate in the reinnervation of the deafferented neurons. By immunocytochemistry, reactive astrocytes were characterized by a significant increase in the intensity of glial fibrillary acidic protein staining beginning at three days postlesion and lasting for at least 27 days postlesion. Reactive astrocytes were characterized by cellular hypertrophy and an increase in the density of immunoreactive processes distributed throughout the deafferented striatum. However, our analysis of astrocyte cell number found no evidence of astrocyte proliferation in response to the deafferentation lesion. Although previous in situ hybridization studies have reported elevated clusterin messenger RNA in reactive astrocytes after decortication, clusterin immunoreactivity was not seen in the cell soma of reactive astrocytes but was distributed as punctate deposits, ranging from 1 to 2μ m in diameter, within the neuropil of the deafferented striatum. At 10 days postlesion, the distribution of clusterin staining appeared as large aggregates of immunoreactive deposits adjacent to neurons. However, by 27 days postlesion, the aggregates of clusterin reaction product were replaced by a fine scattering of individual punctate deposits distributed evenly over the dorsal part of the deafferented striatum. These data support the notion that reactive astrocytes serve multiple, time-dependent roles in response to brain injury and are involved in both the removal of degenerative debris from the lesion site as well as in reforming the synaptic circuitry of the damaged brain. Our data suggest that, in response to decortication, reactive astrocytes are the primary cells responsible for removing degenerating axon terminals, but not axons of passage, from the deafferented striatum and that the coordinate increase in glial fibrillary acidic protein may serve to stabilize the extension of reactive astrocytic processes during phagocytosis. By comparison, clusterin is most likely an extracellular protein released by reactive astrocytes in response to brain damage. Possible roles for clusterin in reactive synaptogenesis include serving as a lipoprotein to facilitate the distribution of recycled lipid to actively sprouting axons and dendrites or regulation of complement-mediated responses.

Morphological and neurochemical features of cultured primary skin fibroblasts of Fischer 344 rats following striatal implantation.

In order to assess the feasibility of using primary skin fibroblasts as a donor population for genetic modification and subsequent intracerebral grafting, the present study examines the structural and neurochemical characteristics of intrastriatal grafts of isogeneic primary fibroblasts over a period of 6 months. In culture, primary skin fibroblasts obtained from a female Fischer 344 rat display robust growth, but once confluent these cells exhibit contact inhibition. Following the implantation of cultured primary cells within the striatum of other adult rats from the same inbred strain, isologous grafts stain immunohistochemically for fibronectin at 1 week, and this immunostaining persists up to 6 months. Immunoreactivity for laminin is intense within the grafts from 1 to 8 weeks, but decreases by 6 months. Astrocytes within the striatum respond dramatically to the implantation of primary fibroblasts, such that immunohistochemical staining for glial fibrillary acidic protein increases markedly from 1 to 8 weeks after implantation. Although the intensity of immunostaining for glial fibrillary acidic protein diminishes among striatal astrocytes between 8 weeks and 6 months, the astrocytic border between the grafts and striatal neuropil remains intensely immunoreactive. Capillaries within the grafts stain immunohistochemically for glucose transporter (a facilitated glucose uptake carrier) as early as 3 weeks after implantation. Following intravenous infusions of peroxidase, capillaries within fibroblast grafts do not permit the extravasation of this macromolecule at 8 weeks and 6 months. Thus, capillaries formed within intracerebral grafts of primary skin fibroblasts exhibit a functional impermeable barrier to macromolecules similar to those capillaries of the host striatum. At the ultrastructural level, grafts possess numerous fibroblasts and have an extracellular matrix filled with collagen. Reactive astrocytic processes filled with intermediate filaments are found throughout the grafts. Hypertrophied astrocytes and their processes also appear to form a continuous border between the grafts and striatal neuropil. Grafts of primary fibroblasts also possess an extensive vasculature that is composed of capillaries with nonfenestrated endothelial cells; the occurrence of reactive astrocytic processes closely associated with or enveloping capillaries is variable. These results provide direct morphological and neurochemical evidence for the long-term survival of isologous fibroblasts after implantation within the rat striatum. From these data, we propose that isologous skin fibroblasts can be considered as donor candidates for successful intracerebral grafting following gene transfer.

Reactive astrocytes are substrates for the growth of adult CNS axons in the presence of elevated levels of nerve growth factor

To assess the necessary parameters for the growth of axons within the adult rat CNS, we have used intracerebral grafts of primary fibroblasts genetically engineered to express nerve growth factor (NGF). Following the implantation of NGF-producing primary fibroblasts within the striatum, cholinergic axons arising from the nucleus basal is grow toward and penetrate these grafts between 1 and 8 weeks. Grafts of noninfected control cells do not elicit axon sprouting at any time. Unmyelinated axons grow into grafts of NGF-producing cells only on reactive astrocytic processes, which contribute to a surrounding glial border. From our data concerning axon growth within the adult rat CNS, we conclude that reactive astrocytes can act as conducive substrates for growing axons; and only in the presence of elevated levels of NGF will permissive substrates (e.g., astrocytes) support axon growth by NGF-sensitive neurons.