Cytotoxic Effect of Brain Macrophages on Developing Neurons

Neurotoxicity of soluble macrophage products in vitro--influence of dexamethasone

When macrophage conditioned medium is added to neurons in vitro, there is a loss of cell membrane integrity, a loss of cell processes, and a large increase in apoptotic neurons. We tested the influence of a potent anti-inflammatory steroid on the interaction between macrophages and neurons. Dexamethasone was applied to macrophages in culture for 24 h while the culture was stimulated with lipopolysaccharide and hypoxia. Conditioned medium was collected after dexamethasone was removed. The dexamethasone pretreated medium was not toxic to hippocampal neurons in contrast to medium from stimulated macrophages not treated with steroid. The dexamethasone effect was concentration dependent. Pretreatment of macrophages with indomethacin and transforming growth factor beta had similar but less impressive effects when compared to dexamethasone. The effect of dexamethasone may have been mediated by inhibiting the synthesis or release of neurotoxic macrophage protein(s), as a combination of medium from steroid pretreated macrophages with medium from nontreated macrophages was not neuroprotective. The toxin(s) did not appear to be tumor necrosis factor alpha or arginase. A role for most neutral proteases was also excluded. We also assessed the consequence of stressing neurons with a mild hypoxic exposure immediately prior to conditioned medium application. Medium from dexamethasone-treated macrophages did not exaggerate hypoxic neuronal injury, unlike medium from non-dexamethasone-treated macrophages. It did not, however, block the exaggerating effect when coapplied in equal volume with medium from nontreated macrophages. Dexamethasone at 100 nM had no impact when applied directly to neurons while they were being exposed to conditioned medium. This in vitro protection by dexamethasone may be relevant to the demonstrated benefit of glucocorticoids in selected brain and spinal cord conditions. Suspicion of a potential link between this in vitro finding and in vivo CNS injury justifies an assessment of more specific agents acting on macrophage protein synthesis or secretion.

Induction of tumor necrosis factor-alpha in the mouse hippocampus following transient forebrain ischemia

To assess the role of tumor necrosis factor-alpha (TNF-alpha) in modulating the process of cerebral ischemic injury, we identified TNF-alpha-producing cells and studied the time course of TNF-alpha expression. Immunoreactivity for TNF-alpha appeared in white matter of the mouse hippocampus as early as 1.5 h following a 30-min global ischemic insult. Double staining for TNF-alpha and glial fibrillary acidic protein (GFAP) suggested that the TNF-alpha-positive cells are most likely microglia, not astrocytes. TNF-alpha immunostaining decreased at 6 and 24 h but increased again at 3 days, when pyramidal neurons showed degeneration. Adjacent-section staining for microglia and double staining with GFAP suggested that TNF-alpha-positive cells in the pyramidal cell layer were microglia and those in the white matter were astrocytes. By 5 days TNF-alpha immunostaining disappeared from these glial cells, while a number of microglia were accumulated in the degenerated hippocampal pyramidal layer. Pyramidal neurons never expressed TNF-alpha immunoreactivity. Western blotting confirmed biphasic TNF-alpha expression. Our findings suggest that early production of TNF-alpha by microglia may activate a cytokine network in post-ischemic brain resulting in TNF-alpha synthesis by astrocytes.

Suppression of the oxidative burst in murine microglia by nitric oxide

The effect of nitric oxide (NO) on the oxidative burst was analyzed in purified murine microglial cells in vitro. The generation of reactive oxygen derivatives was monitored with the use of luminol-dependent chemiluminescence. After inducing the endogenous NO production with interleukin 1beta (IL-1beta) and interferon-gamma (IFN-gamma) the superoxide anion release was significantly reduced, which was reversed by the inhibition of the NO synthase. Additionally, chemical NO-releasing compounds reduced the generation of reactive oxygen derivatives rapidly and independently of the pathway used to trigger the oxidative burst. This effect of NO was not mediated via guanylyl cyclase and cGMP, or due to the scavenging of released superoxide anions. This attenuation of superoxide anion generation by NO may limit deleterious effects of the release of reactive oxygen derivatives in tissue inflammation or injury.

Increased expression of IL-1beta converting enzyme in hippocampus after ischemia: selective localization in microglia

Although the interleukin-1beta converting enzyme (ICE)/CED-3 family of proteases has been implicated recently in neuronal cell death in vitro and in ovo, the role of specific genes belonging to this family in cell death in the nervous system remains unknown. To address this question, we examined the in vivo expression of one of these genes, Ice, after global forebrain ischemia in gerbils. Using RT-PCR and Western immunoblot techniques, we detected an increase in the mRNA and protein expression of ICE in hippocampus during a period of 4 d after ischemia. Chromatin condensation was observed in CA1 neurons within 2 d after ischemia. Internucleosomal DNA fragmentation and apoptotic bodies were observed between 3 and 4 d after ischemia, a period during which CA1 neuronal death is maximal. In nonischemic brains, ICE-like immunoreactivity was relatively low in CA1 pyramidal neurons but high in scattered hippocampal interneurons. After ischemia, ICE-like immunoreactivity was not altered in these neurons. ICE-like immunoreactivity, however, was observed in microglial cells in the regions adjacent to the CA1 layer as early as 2 d after ischemic insult. The increase in ICE-like immunoreactivity was robust at 4 d after ischemia, a period that correlates with the DNA fragmentation observed in hippocampal homogenates of ischemic brains. These results provide the first evidence for the localization and induction of ICE expression in vivo after ischemia and suggest an indirect role for ICE in ischemic damage through mediation of an inflammatory response.

Specific domains of beta-amyloid from Alzheimer plaque elicit neuron killing in human microglia

Alzheimer's disease (AD) is found to have striking brain inflammation characterized by clusters of reactive microglia that surround senile plaques. A recent study has shown that microglia placed in contact with isolated plaque fragments release neurotoxins. To explore further this process of immunoactivation in AD, we fractionated plaque proteins and tested for the ability to stimulate microglia. Three plaque-derived fractions, each containing full-length native A beta 1-40 or A beta 1-42 peptides, elicited neurotoxin release from microglia. Screening of various synthetic peptides (A beta 1-16, A beta 1-28, A beta 12-28, A beta 25-35, A beta 17-43, A beta 1-40, and A beta 1-42) confirmed that microglia killed neurons only after exposure to nanomolar concentrations of human A beta 1-40 or human A beta 1-42, whereas the rodent A beta 1-40 (5Arg-->Gly, 10Tyr-->Phe 13His-->Arg) was not active. These findings suggested that specific portions of human A beta were necessary for microglia-plaque interactions. When coupled to microspheres, N-terminal portions of human A beta (A beta 1-16, A beta 1-28, A beta 12-28) provided anchoring sites for microglial adherence whereas C-terminal regions did not. Although itself not toxic, the 10-16 domain of human A beta was necessary for both microglial binding and activation. Peptide blockade of microglia-plaque interactions that occur in AD might prevent the immune-driven injury to neurons.

Excess glutamate in the cerebrospinal fluid in bacterial meningitis

We investigated possible neurotoxic components in the cerebrospinal fluid (CSF) of patients with bacterial meningitis. On murine cerebellar neuronal cell cultures, CSF exerted a dose-dependent toxic effect, which was attenuated by the NMDA receptor antagonist MK-801. Glutamate concentrations in the CSF of patients with bacterial meningitis were measured by an enzymatic assay and found to be significantly elevated (p < 0.001) as compared to viral meningitis and non-inflammatory neurological diseases. The concentration of glutamate in the CSF of patients with bacterial meningitis varied considerably and correlated with the severity of the disease as scored by the Glasgow Coma Scale. Cells in the CSF, mainly comprising polymorphonuclear granulocytes, did not release any glutamate into the culture medium, whereas blood monocytes produced remarkable amounts. These findings implicate an important role of monocytic inflammatory cells in bacterial meningitis by the release of glutamate, which may contribute to neuronal cell death.

Complement and clusterin in the injured nervous system

Peripheral nerve injury and neuronal degeneration resulting from toxic ricin induce activation of the classical pathway of complement close to the injured motorneuron perikarya or sensory terminals. In contrast, degeneration of central myelinated fibers is not accompanied by complement expression. The main source of complement in peripheral nerve injury and toxic ricin degeneration appears to be microglia. Brain contusion is associated with complement activation. Some of the complement in this situation may derive from plasma, because the blood-brain barrier is disrupted. Clusterin expression is increased in astrocytes along with their activation in the vicinity of lesioned neurons. In addition, axotomized motorneurons show a marked clusterin upregulation. A relationship between clusterin and cell death is suggested by the prominent aggregation of clusterin in neuronal perikarya destroyed by the effects of toxic ricin, as well as by the neosynthesis of clusterin in apparently degenerating nonneuronal cells, presumed to be oligodendrocytes. Our findings indicate that the expression of complement and clusterin are prominent features of neural degeneration and regeneration, as it is in Alzheimer's disease brains as well. The nerve injury conditions described, therefore, offer attractive experimental models to elucidate the roles of these molecular components in neurodegenerative disorders, thereby providing useful insights into potentially new therapeutic approaches in these conditions.

Activation of the complement cascade and increase of clusterin in the brain following a cortical contusion in the adult rat

The aim of the present study was to examine the glial cell response and the possible involvement of the complement cascade following a cerebral cortical contusion. The lesion was produced using a standardized weight-drop technique in adult rats. The blood-brain barrier was damaged, as demonstrated by a decrease of immunoreactivity for a tight junction protein normally expressed by endothelial cells of small vessels in the central nervous system. Increased immunoreactivity for microglial (OX42) and astroglial cells (glial fibrillary acidic protein), as well as macrophages expressing ED1-immunoreactivity (IR) were found in the vicinity of the lesion at all postoperative survival times (2-14 days). In the present study complement factor C3d- and C9-IR was found around the lesion, indicating that activation of the complement cascade had taken place. Furthermore, immunoreactivity for the putative complement inhibitor clusterin (sulfated glycoprotein-2) was found in some of the injured neurons. The contralateral hemisphere showed no evidence of the reaction found in the ipsilateral hemisphere. The balance between complement activation and complement inhibitors may have an impact on the degenerative components in the brain following traumatic injury and in particular on the events leading to nerve cell death.

Microglia: a sensor for pathological events in the CNS

The most characteristic feature of microglial cells is their rapid activation in response to even minor pathological changes in the CNS. Microglia activation is a key factor in the defence of the neural parenchyma against infectious diseases, inflammation, trauma, ischaemia, brain tumours and neurodegeneration. Microglia activation occurs as a graded response in vivo. The transformation of microglia into potentially cytotoxic cells is under strict control and occurs mainly in response to neuronal or terminal degeneration, or both. Activated microglia are mainly scavenger cells but also perform various other functions in tissue repair and neural regeneration. They form a network of immune alert resident macrophages with a capacity for immune surveillance and control. Activated microglia can destroy invading micro-organisms, remove potentially deleterious debris, promote tissue repair by secreting growth factors and thus facilitate the return to tissue homeostasis. An understanding of intercellular signalling pathways for microglia proliferation and activation could form a rational basis for targeted intervention on glial reactions to injuries in the CNS.

Inhibition of microglial superoxide anion production by isoproterenol and dexamethasone

Microglia, like other tissue macrophages, are a component of the hypothalamic-pituitary endocrine-immune axis and, as such, are responsive to both neural and endocrine factors. Using cultured neonatal hamster microglia, we have examined the effect of isoproterenol, a beta-adrenergic agonist, and dexamethasone, a synthetic glucocorticoid, on superoxide anion production. For these experiments, microglia were pretreated with isoproterenol or dexamethasone and then induced to produce superoxide anion by exposure of the cells to phorbol myristate acetate (PMA). Our study demonstrates that the PMA-stimulated production of superoxide anion was decreased by acute (30 min) and chronic (24 h) pretreatment of the microglia with isoproterenol and was blocked by the beta-adrenergic receptor antagonist, propranolol. Since a rise in intracellular cAMP may be a prime factor in the inhibition of superoxide anion production in isoproterenol-treated cells, we used forskolin, a known activator of the adenylate cyclase in place of isoproterenol and re-investigate superoxide anion production. Short term exposures to forskolin produced a lower amount of superoxide anion than PMA-stimulated alone and, thus, mimicked the effect of isoproterenol. However, treatment with the same concentration of forskolin for 24 h prior to the induction of the NADPH oxidase did not significantly change PMA-stimulated superoxide anion production from untreated values. Thus, chronic exposure to forskolin produced a different effect than chronic exposure to isoproterenol. Isoproterenol and forskolin both increased immunoreactivity for the protein products of the early response genes, c-fos and c-jun. Pretreatment with dexamethasone for 24 h also inhibited superoxide anion production and was blocked by the protein synthesis inhibitor, cycloheximide. The simultaneous addition of varying concentrations of dexamethasone and 5 microM isoproterenol did not produce a greater inhibition in superoxide anion production than either agent alone. The down-regulation of microglial function by adrenergic agonists and by glucocorticoids provides a way in which the cytotoxicity of these immune cells can be reduced and may be a factor in the paracrine regulation of microglia.

Study of receptor-mediated neurotoxins released by HIV-1-infected mononuclear phagocytes found in human brain

Although there is growing evidence that neurotoxic molecules produced by HIV-1-infected mononuclear phagocytes damage neurons, the precise mechanisms of neuronal attack remain uncertain. One class of cytotoxin involves neuronal injury mediated via the NMDA receptor. We examined blood monocytes and brain mononuclear cells isolated at autopsy from HIV-1-infected individuals for the ability to release NMDA-like neuron-killing factors. We found that a neurotoxic amine, NTox, was produced by blood monocytes and by brain mononuclear phagocytes infected with retrovirus. In vivo injections of minute quantities of NTox produced selective damage to hippocampal pyramidal neurons. NTox can be extracted directly from brain tissues infected with HIV-1 and showed structural features similar to wasp and spider venoms. In contrast to NTox, HIV-1 infection did not increase the release of the NMDA excitotoxin quinolinic acid (QUIN) from mononuclear cells. Although we found modest elevations of QUIN in the CSF of HIV-1-infected individuals, the increases were likely attributable to entry through damaged blood-brain barrier. Taken together, our data pinpoint NTox, rather than QUIN, as a major NMDA receptor-directed toxin associated with neuro-AIDS.

Neurocytopathic effects of beta-amyloid-stimulated monocytes: a potential mechanism for central nervous system damage in Alzheimer disease

Growing evidence indicates that cells of the mononuclear phagocyte lineage, which includes peripheral blood monocytes (PBM) and tissue macrophages, participate in a variety of neurodestructive events and may play a pivotal role in neurodegenerative conditions such as Alzheimer disease. The present study sought to determine whether exposure of PBM to beta-amyloid peptide (A beta), the major protein of the amyloid fibrils that accumulate in the brain in Alzheimer disease, could induce cytopathic activity in these cells upon their subsequent incubation with neural tissue. PBM were incubated with A beta for 3 days, centrifuged and washed to remove traces of cell-free A beta, and then applied to organotypic cultures of rat brain for varying periods of time. By using a cell-viability assay to quantitate neurocytopathic effect, an increase in the ratio of dead to live cells was detected in cultures containing A beta-stimulated PBM versus control PBM (stimulated with either bovine serum albumin or reverse A beta peptide) as early as 3 days after coculture. The ratio of dead to live cells increased further by 10 days of coculture. By 30 days of coculture, the dead to live cell ratio remained elevated, and the intensity of neurocytopathic effect was such that large areas of brain mass dissociated from the cultures. These results indicate that stimulation of PBM with A beta significantly heightens their neurocytopathic activity and highlight the possibility that inflammatory reactions in the brain play a role in the neurodegeneration that accompanies Alzheimer disease.

Protective effect of cyclosporin A on white matter changes in the rat brain after chronic cerebral hypoperfusion

BACKGROUND AND PURPOSE

Activation of glial cells and rarefaction of the white matter have been reported in rat brain after bilateral permanent occlusion of the common carotid arteries. Using this model, we investigated the effects of the immunosuppressant cyclosporin A on the activation of glial cells and the white matter rarefaction.

METHODS

Both common carotid arteries were ligated bilaterally in 40 male Wistar rats. Twenty-two of these rats received an intraperitoneal injection of cyclosporin A, and the remaining 18 received a vehicle-solution injection. Microglia/macrophages were investigated with immunohistochemistry for the major histocompatibility complex class I and II antigens as well as for leukocyte common antigen. Astroglia were examined with glial fibrillary acidic protein as a marker. Activation of glial cells and white matter rarefaction were then investigated from 7 to 30 days after the ligation.

RESULTS

In vehicle-treated animals, there was a persistent and extensive activation of both microglia/macrophages and astroglia in the white matter, including the optic nerve, optic tract, corpus callosum, internal capsule, and traversing fiber bundles of the caudoputamen. In cyclosporin A-treated rats, the number of activated microglia/macrophages was significantly reduced (P < .01) to approximately one fifth of that in vehicle-treated animals. Similarly, rarefaction of the white matter was much less intense in cyclosporin A-treated rats (P < .01).

CONCLUSIONS

Cyclosporin A suppressed both glial activation and white matter changes after chronic cerebral hypoperfusion. These results suggest that immunologic reaction may play a role in the pathogenesis of the white matter changes and that the present model may be useful in investigating the pathophysiology of white matter changes induced by chronic cerebral hypoperfusion.

Reactive microglia in cerebral ischaemia: an early mediator of tissue damage?

Microglial cell activation is a rapidly occurring cellular response to cerebral ischaemia. Microglia proliferate, are recruited to the site of lesion, upregulate the expression of several surface molecules including major histocompatibility complex class I and II antigens, complement receptor and the amyloid precursor protein (APP) as well as newly expressed cytokines, e.g. interleukin-1 and transforming growth factor beta 1. The ischaemia-induced production of APP may contribute to amyloid deposition in the aged brain under conditions of hypofusion. Ultrastructurally, microglia transform into phagocytes removing necrotic neurons but still respecting the integrity of eventually surviving neurons even in the close vicinity of necrotic neurons. Microglial activation starts within a few minutes after ischaemia and thus precedes the morphologically detectable neuronal damage. It additionally involves a transient generalized response within the first 24 hours post-ischaemia even at sites without eventual neuronal cell death. In functional terms, the microglial reaction appears to be a double-edged sword in ischaemia. Activated microglia may exert a cytotoxic effector function by releasing reactive oxygen species, nitric oxide, proteinases or inflammatory cytokines. All of these cytotoxic compounds may cause bystander damage following ischaemia. Pharmacological suppression of microglial activation after ischaemia has accordingly attenuated the extent of cell death and tissue damage. However, activated microglia support tissue repair by secreting factors such as transforming growth factor beta 1 which may limit tissue damage as well as suppress astroglial scar formation. In line with ultrastructural observations microglial activation in ischaemia is a strictly controlled event. By secreting cytokines and growth factors activated microglia most likely serve seemingly opposed functions in ischaemia, i.e. maintenance as well as removal of injured neurons. Post-ischaemic pharmacological modulation of microglial intervention in the cascade of events that lead to neuronal necrosis may help to improve the structural and functional outcome following CNS ischaemia.

Cultured rat striatal and cortical astrocytes protect mesencephalic dopaminergic neurons against hydrogen peroxide toxicity independent of their effect on neuronal development

Reactive oxygen species (ROS), including hydrogen peroxide, are supposed to be involved in the degeneration of dopaminergic neurons in Parkinson's disease. The potential role of astrocytes against neurotoxic effects of ROS was studied in cocultures of rat mesencephalic neurons and rat striatal or cortical astrocytes. Neuronal [3H]dopamine uptake, a marker of dopaminergic neuron integrity, was enhanced by striatal astrocytes, but not by cortical astrocytes, compared to uptake in mesencephalic neurons cultured alone. Whereas hydrogen peroxide at concentrations up to 100 microM reduced the [3H]dopamine uptake in neuronal cultures, no reduction of the uptake was observed in cocultures, regardless of the origin of the supporting astrocytes. These results suggest that astrocyte mediated protection of neurons against hydrogen peroxide induced toxicity is not directly related to a region-specific neurotrophic effect.

Protection from oxidation enhances the survival of cultured mesencephalic neurons

Oxidative stress has been linked to the destruction of dopaminergic neurons in the substantia nigra and may be a significant factor in both Parkinson's disease and MPTP toxicity. Using primary cultures of embryonic rat mesencephalon and standard immunocytochemical techniques, we have examined the survival of tyrosine hydroxylase-containing (TH+) neurons cultured in the presence of antioxidants and/or in an environment of low oxygen partial pressure. The number of TH+ neurons increased approximately twofold if superoxide dismutase, glutathione peroxidase (GP), or N-acetyl cysteine (NAC) were added to the culture media. Exposure of the neurons to a 5% oxygen environment (38 torr, i.e., 38 mm Hg) also increased the survival of TH+ neurons by about twofold. A dramatic enhancement of survival, however, was seen when NAC was used in combination with the 5% oxygen environment. In this case, the number of TH+ neurons increased fourfold from nontreated controls. Morphological changes were also noted. GP increased the average neurite length while NAC increased the average area of the cell body in the TH+ neuron. These results suggest that manipulation of oxidative conditions by changing the ambient O2 tension or the level of antioxidants promotes survival of TH+ neurons in culture and may have implications for transplantation therapies in Parkinson's disease.

Mechanisms of delayed cell death following hypoxic-ischemic injury in the immature rat: evidence for apoptosis during selective neuronal loss

The mechanisms leading to delayed cell death following hypoxic-ischemic injury in the developing brain are unclear. We examined the possible roles of apoptosis and microglial activation in the 21-day-old rat brain following either mild (15 min) or severe (60 min) unilateral hypoxic-ischemic injury. The temporal and spatial patterns of DNA degradation were assessed using gel-electrophoresis and in-situ DNA end-labelling. Microglial activation, mitochondrial failure and cell death were examined using lectin histochemistry, 2,3,5,triphenyl-H-tetrazolium chloride (TTC) staining and acid fuchsin staining, respectively. Selective neuronal death produced by the 15 min injury was associated with the development of apoptotic morphology, DNA laddering and acidophilia from 3 days post-hypoxia. The 60 min injury accelerated this process with some cells showing signs of DNA degradation at 10 h post-hypoxia. However, in the cortex, which developed infarction after the 60 min injury, a different pattern of cell loss occurred. The DNA and mitochondria remained intact, and cells basophilic, until after 10 h post-hypoxia, then widespread necrosis developed by 24 hr. In contrast to regions of selective neuronal loss, DNA degradation was initially random (at 24 hr), with 180bp DNA ladders not detected until 3 days post-hypoxia. There was no morphological evidence of apoptosis. Microglial activation coincided with the onset of DNA degradation in regions of selective neuronal loss but not infarction, suggesting a possible role in selective neuronal death. The results suggest that cortical infarction, which was delayed for at least 10 h, was necrotic, and occurred independently of microglial activation and apoptosis. In contrast, selective neuronal death was apoptotic.

Glial responses to hypoxic/ischemic encephalopathy in neonatal rat cerebrum

This study was undertaken to examine glial responses in the immature rat brain to hypoxic/ischemic injury. The results indicate that developing microglia are the first and principle glial element that responds to hypoxic/ischemic injury during the neonatal period. The astrocyte response occurs later and macrophage infiltration may be delayed or absent.

Expression of basic fibroblast growth factor and its receptors in human fetal microglia cells

The presence of basic fibroblast growth factor (bFGF) and FGF receptors was investigated in microglia cells derived from human fetal brain long-term cultures. Production of bFGF was suggested through the capability of microglial extracts to stimulate plasminogen activator (PA) synthesis in endothelial cells. The identity of PA-stimulating activity with bFGF was confirmed by its high affinity for heparin and its cross-reactivity with polyclonal antibodies to human recombinant bFGF. These antibodies recognized a cell-associated M(r) 18,000 protein as well as trace amounts of the M(r) 24,000 bFGF isoform in Western blot. All microglial cells showed bFGF immunoreactivity in the cytoplasm and, sometimes, in the nucleus. Scatchard plot analysis of 125I-bFGF binding data revealed the presence of low affinity heparansulphate proteoglycans (380,000 +/- 60,000 sites/cell; Kd = 730 +/- 200 nM) and of high affinity tyrosine-kinase receptors (10,300 + 2500 sites/cell; Kd = 30 +/- 9 pM). Immunocytochemistry confirmed the presence of FGF receptor (1/flg) on the cell surface of some, but not all microglial cells, with prevalent association to ameboid microglia. Transcripts for FGF receptors 1, 2, 3 and 4 were found in microglia by Northern blot analysis. Co-expression of bFGF and its receptors in human fetal microglia suggests an autocrine role of bFGF in these cells.

High dose methylprednisolone therapy reduces expression of JE/MCP-1 mRNA and macrophage accumulation in the ischemic rat brain

The effects of glucocorticoid (GC) on ischemic brain remain to be investigated. Since GC modulates immunological system, it also may inhibit macrophage accumulation in the ischemic brain. The GC effect, if any, on macrophages in ischemic brain, may be mediated through modulation of JE/MCP-1 gene, a strong monocyte attractant, which is expressed in the rat brain after ischemia. The purpose of the present study is to elucidate the effect of high dose methylprednisolone (MP) treatment on (1) macrophage infiltration, (2) histopathology of the ischemic lesion, and (3) expression of JE/MCP-1 mRNA, in a focal cerebral ischemia model of the rat. Thirty Wistar rats were used in this study. Focal cerebral ischemia was induced by advancing a nylon monofilament into the internal carotid artery until the origin of the middle cerebral artery (MCA) was occluded. For JE/MCP-1 mRNA study, animals (n = 9) were randomly injected with MP 75 mg/kg (x 3) (n = 3), 100 mg/kg (x 3) (n = 3), or same volume of saline (n = 3) and killed 24 h after onset of MCA occlusion. Three animals were used as a normal control, and a section of the liver from one rat was used as an internal control for JE/MCP-1 mRNA. Northern blot analysis was performed using murine JE c-DNA. For the histopathological study, animals (n = 17) were randomly divided into a MP group (MP 100 mg/kg x 3, n = 9) and a control group (saline treated, n = 8), and killed 72 h after onset of MCA occlusion.(ABSTRACT TRUNCATED AT 250 WORDS)

Evidence for activation of the terminal pathway of complement and upregulation of sulfated glycoprotein (SGP)-2 in the hypoglossal nucleus following peripheral nerve injury

In a previous study, we found immunoreactivity for complement factors C3, C3d, and C4d, as well as endogenous IgG in the hypoglossal nucleus following hypoglossal nerve transection, suggesting that activation of the complement cascade had taken place in the vicinity of the axotomized motorneurons. In the present study, we found increased immunoreactivity for complement factor C1 and C1q in reactive microglia, indicating an increased potential for initiation of the classical pathway by binding of IgG to C1q. Furthermore, we found immunoreactivity for C9, which contributes to the formation of C5b-9, the final lytic product of the complement cascade close to the axotomized neurons and perineuronal glia. In addition, immunoreactivity and mRNA labeling of sulfated glycoprotein (SGP-2), a putative complement inhibitor, was increased in a subpopulation of the axotomized motorneurons. SGP-2 immunoreactivity was also increased in astroglial cells ipsilateral to the nerve injury. The results lend further support to the hypothesis that the complement cascade is activated in the vicinity of axotomized neurons, which in turn may be protected by complement inhibitors. The balance between activation of complement and complement inhibitors might have an impact on the degenerative components of the axon reaction and, in particular, the events leading to nerve cell death.

Glutamate production by cultured microglia: differences between rat and mouse, enhancement by lipopolysaccharide and lack effect of HIV coat protein gp120 and depolarizing agents

Glutamate release from rat and mouse microglia subcultures grown in a serum-free medium was substantially greater in the presence than in the absence of a physiological concentration of glutamine (0.5 mM). Mouse microglia produced and released more glutamate than rat microglia. Glutamate accumulation in the medium increased with time and cell density, which is consistent with the virtual absence of glutamate reuptake. Lipopolysaccharide (LPS; 10-100 ng/ml), HIV coat protein gp120 (0.1-10 nM), high K+ (35 mM) or ATP (150 microM), did not affect glutamate release from cells maintained in serum-free medium. In the presence of 1% dialyzed serum, however, LPS induced a dose- and time-dependent increase in the accumulation of glutamate in the medium, suggesting that, as in other cell types, serum factors are required for LPS binding to its receptors.

Downregulation of in vitro neurotoxicity of brain macrophages by prostaglandin E2 and a beta-adrenergic agonist

Brain macrophages (BM), a subpopulation of microglia, have the ability to kill neurons by producing reactive oxygen intermediates. Cocultures of neurons and macrophages derived from the cerebral cortex of rat embryos were used to look for regulation of BM neurotoxicity. Isoproterenol (10(-7) M), a beta-adrenergic agonist, induced a significant inhibition of BM neurotoxicity and this effect was abolished in the presence of propranolol, a beta-adrenergic antagonist. BM neurotoxicity was also reduced in the presence of prostaglandin E2 (10(-8), 10(-6) M), a metabolite derived from arachidonic acid. These results suggest endogenous mechanisms of neuroprotection operating either during development or following lesions.

Brain macrophages stimulate neurite growth and regeneration by secreting thrombospondin

The presence of macrophages in the developing or lesioned central nervous system (CNS) led us to study the influence of these cells on neuronal growth. Macrophages were isolated from embryonic rat brain and we observed that factors released in vitro by these cells stimulate neurite growth and regeneration of cultured CNS neurons. This effect was inhibited by antibodies directed against thrombospondin, an extracellular matrix protein that we found to be synthesized and released by brain macrophages. Immunodetection of thrombospondin in the adult rat brain lesioned by kainic acid confirmed the production of this protein by brain macrophages and indicated an early intraparenchymal accumulation of thrombospondin following injury. These results suggest that brain macrophages contribute actively to neurite growth or regeneration during the development or in pathological contexts.

Secondary cell death and the inflammatory reaction after dorsal hemisection of the rat spinal cord

Local spinal cord lesions are often greatly enlarged by secondary damage, a process which leads to massive additional cell death. This process is poorly understood. In order to investigate which types of cells could play a role in increasing the size of the lesion, we have analysed the events occurring at rat spinal cord lesion sites from 1 h to 3 months after partial transection using cell type-specific markers. One hour after transection, the lesion site was small and corresponded to the zone of primary mechanical damage. Extravasation of blood and an opening of the blood-brain barrier occurred. Rapidly thereafter, at 3 and 6 h, an area of secondary cell death developed around the zone of the primary lesion. This secondary cell death, which was probably largely of the necrotic type, affected neurons, macroglia and microglial cells indiscriminately. It was virtually complete at 12 h. Recruitment of inflammatory cells followed a time course which lagged behind that of secondary cell death. Adhesion of neutrophils to the inside of blood vessels was observed at 3 h. They appeared in large numbers at 6 h at the site of the primary lesion, but not yet in the area of secondary cell death. They were numerous throughout the lesion site at 24 h and then disappeared rapidly. Proliferation and recruitment of macrophages and microglial cells became predominant 2 days after injury. Their density was highest within the lesion site between 4 and 8 days. Very few astrocytes were present in the lesion site during the first week.(ABSTRACT TRUNCATED AT 250 WORDS)

Mechanisms of sprouting in the adult central nervous system: cellular responses in areas of terminal degeneration and reinnervation in the rat hippocampus

Neurons of the adult mammalian central nervous system are limited in their ability to regenerate in response to injury. In certain circumstances, however, such neurons can exhibit morphological plasticity (e.g. sprouting). Unilateral transection of the perforant path in the adult rat induces terminal degeneration of entorhinal axons within the molecular layer of the ipsilateral hippocampal dentate gyrus. Cholinergic (and other) afferents subsequently sprout within the denervated zone. We show that despite the breach in the blood-brain barrier at the site of the aspirative lesion, the barrier remains intact in the areas of terminal degeneration (and reinnervation), and peripheral monocytic macrophages do not infiltrate this area to participate in the degenerative and/or regenerative events. Perforant path transection does not induce expression of major histocompatibility antigens on reactive cells within the denervated zone, nor are T lymphocytes recruited to this area. T lymphocyte-deficient Nude rats exhibit normal cholinergic sprouting. Perforant path transection does induce rapid and robust proliferation of microglia, and astrocytes to a lesser extent, in areas undergoing terminal degeneration. Histological evaluation after antimitotic administration shows that this glial proliferation is not required for the subsequent neuronal sprouting events. These results show that the reparative process in this model system involves interactions between cells endogenous to the brain in a non-immune context. Knowledge of these cellular responses provides a framework from which to further investigate putative molecular signals involved in initiating the neuronal sprouting events. Discovering the cellular and molecular interactions taking place under sprouting conditions is likely to be critical for understanding the mechanisms of reactive neuronal growth and, furthermore, may provide insights as to why regeneration is so limited in the central nervous system.

Tetrodotoxin blocks HIV coat protein (gp120) toxicity in primary neuronal cultures

HIV-1-associated cognitive/motor complex is a frequent neurological complication of the acquired immunodeficiency syndrome (AIDS). The pathogenesis of this syndrome implicates immunopathological and toxic events such as the production of cytokines. The HIV envelope glycoprotein gp120 seems also to play a major role in this process. Gp120 could produce a slow neuronal death probably via the release of neurotoxic factors by CNS macrophages/monocytes. NMDA antagonists and Ca2+ channel blockers in vitro have a powerful neuroprotective effect against gp120 neurotoxicity. The purpose of the present work is to determine whether gp120-induced neurotoxicity is associated with an abnormal neuronal depolarization induced by putative neurotoxins. We have compared in vitro the neuroprotective effects of Tetrodotoxin a Na+ channel blocker, the Ca2+ channel blocker nifedipine and the NMDA antagonist MK-801 in primary cortical neurons taken from embryonic rat and intoxicated with gp120. We observed comparable neuroprotective effects with the 3 precited compounds suggesting that gp120-induced neurotoxic factors act on Na+ channels, NMDA receptors and Ca2+ channels in a cascade of cellular events. We confirmed that the presence of macrophages is needed to trigger a marked gp120-induced neurotoxicity. These results underline the fact that depolarization is an important component of gp120 neurotoxicity in primary neuronal cultures.

Growth factors in CNS repair and regeneration

Traumatic central nervous system (CNS) injury is a significant clinical problem in the developed world. After injuries that penetrate into either the mature brain or spinal cord, damaged neurons initially begin to regrow, but this regeneration is aborted as a fibrotic scar is laid down within the wound. Reconnection of several neuronal pathways does not occur. Functional recovery from such injuries is therefore poor and morbidity severe, particularly for those patients with spinal cord damage. Although palliative measures are available to improve the quality of life, there is no accepted treatment to restore impaired sensory or motor function, so patients remain significantly and permanently debilitated. However, the rapid recent advances that have been made in our understanding of the underlying cellular and trophic pathology of such injuries offer the potential for development of novel therapies to control scarring, enhance neuron survival and stimulate axon regeneration, thereby promoting functional recovery.

Selective MAO-A and B inhibitors, radical scavengers and nitric oxide synthase inhibitors in Parkinson's disease

In the absence of identification of either an endogenously or an exogenously derived dopaminergic neurotoxin, the most valid hypothesis currently envisaged for etiopathology of Parkinson's disease (PD) is selective oxidative stress (OS) in substantia nigra (SN). Although OS is not proven, a significant body of evidence from studies on animal and Parkinsonian brain neurochemistry supports it. This hypothesis is based on excessive formation of reactive oxygen species (O2 and OH.) and demise of systems involved with scavenging or preventing the formation of such radicals from H2O2, generated as a consequence of dopamine oxidation (autoxidation and deamination). Since MAO (monoamine oxidase A and B are the major H2O2 generating enzymes in the SN much attention has been paid to their selective inhibitors as symptomatic and neuroprotective agents in PD. Attention should also be given to radical scavengers (e.g. iron chelators, lipid peroxidative inhibitors and Vitamin E derivatives) as therapeutic neuroprotective agents in PD. This is considered valid since a significant elevation of iron is known to occur selectively in SN zone compacta and within the remaining melanized dopamine neurons of Parkinsonian brains. Although all the mechanism of iron induced oxygen free radical formation is not fully known there is no doubt that it participates with H2O2 (Fenton chemistry) to generate cytotoxic hydroxyl radical (OH.) and induce tissue OS and neurodegeneration in 6-hydroxydopamine model of PD. The dramatic proliferation of reactive amoeboid macrophages and microglia seen in SN of PD brains together with OS is highly compatible with an inflammatory process, similar to what has been observed in Alzheimer's disease and multiple sclerosis brains. This has led us to examine the ability of reactive macrophages to produce oxygen free radicals in response to nitric oxide (NO) production. The latter radical has been implicated in the excitotoxicity of glutaminergic neurons innervating the striatum and SN. Indeed we have now observed that in reactive macrophages NO acts as a signal transducer of O2 production which can synergize with dopamine oxidation.

Reactive mononuclear phagocytes release neurotoxins after ischemic and traumatic injury to the central nervous system

Reactive microglia and invading macrophages, which appear in brain damaged by stroke or trauma, secrete neuron-killing factors. This release of cytotoxic substances is a delayed process and is not detected until inflammatory cells reach a peak of reactivity by the second day after injury. Proximity to the site of injury and density of mononuclear phagocytes determine in part the amount of neurotoxic activity released by injured tissues. Moreover, drugs that suppress the accumulation of reactive microglia and macrophages also reduce tissue production of neuron poisons. Neurotoxins released by brain inflammatory cells or extracted directly from inflamed tissues are heat-stable, protease-resistant molecules < 500 daltons with actions blocked by N-methyl-D-aspartate (NMDA) receptor antagonists. These molecules are distinguished from free radical intermediates, bind to cation exchange resins, lack carboxyl moieties, and are separated from excitatory amino acids including glutamate or aspartate and from the NMDA receptor-mediated toxin quinolinic acid by ion exchange and reverse phase chromatography. Our data suggest that an unrecognized class of neuron-killing molecules produced by inflammatory cells mediate the delayed neuronal loss associated with stroke and trauma.

Glutamate uptake by astrocytes is inhibited by reactive oxygen intermediates but not by other macrophage-derived molecules including cytokines, leukotrienes or platelet-activating factor

By their property to release glutamate and reactive oxygen intermediates, macrophages may play an important role in neurotoxicity. In the present study we have investigated whether macrophage-derived molecules also impair the detoxification of glutamate by astrocytes. Cytokines, including interleukin (IL)-1, 6 and 10, interferon (IFN)-alpha/beta, tumor necrosis factor (TNF)-alpha and transforming growth factor (TGF)-beta 1, as well as leukotriene (LT) B4 and C4, prostaglandin (PG) E2 and nitric oxide radicals had no effect on the uptake of [3H]glutamate by murine astrocytes in culture. In contrast, exposure of astrocytes to the enzyme glucose oxidase (100-200 mU ml-1), which maintains steady-state levels of hydrogen peroxide, reduced glutamate uptake by 30-50%. By their dual effect, comprising secretion of glutamate and inhibition of its detoxification by astrocytes, activated macrophages and microglial cells may contribute to exacerbate excitotoxic mechanisms in neurological diseases.

Distribution and axonal relations of macrophages in a neuroma

After axotomy in the peripheral nervous system, most axons regrow and re-establish contact with their targets. Depending on the type of lesion, a varying number of nerve fibers fail to regenerate and terminate far from the target, forming a neuroma. Sensory axons trapped in a neuroma show abnormal sensitivity to various stimuli, and often fire spontaneously. In this study we have examined the distribution and axonal relations of macrophages in rat sciatic neuromas three days to one year after cutting and ligating the nerve. ED1-immunoreactive macrophages migrated into the neuroma in large numbers within the two first weeks after the injury. Most cells were at that time located 0.5-1 mm proximal to the ligature. From three weeks on, a majority of the ED1-immunoreactive cells contained numerous large vacuoles filled with myelin fragments. At sites of focal demyelination, macrophages often had direct contact with axonal membranes. At later survival stages (three months to one year) ED1-immunoreactive cells were seen not only in the area just proximal to the ligature, but also several millimeters proximal to this. Macrophages persisted in considerable numbers in the neuroma for at least one year. These data suggest that neuroma macrophages may participate in the genesis of electrophysiological abnormalities thought to underly chronic pain after neuroma formation, possibly by creating demyelinated axonal regions susceptible to external stimuli from e.g. neighboring nerve fibers, by releasing substances which influence regeneration and remodelling of axonal growth cones, or by direct actions on the denuded axonal membranes.

Evidence for activation of astrocytes via reactive microglial cells following hypoglossal nerve transection

Following peripheral nerve injury, resident microglial cells proliferate and astrocytes undergo hypertrophy, as evidenced, e.g., by an increase in the levels of glial fibrillary acidic protein (GFAP). In a previous study we have shown that infusion of cytosine arabinoside (ARA-C) into the rat brain blocks the axotomy-induced proliferation of microglial cells. This experimental approach has been used in the present study in order to explore the issue of whether the reactive microglial cells are mediators of the increased GFAP expression in the hypoglossal nucleus of the rat following axotomy. Quantitative analysis of sections processed for immunocytochemistry or in situ hybridization demonstrated a marked increase in GFAP-like immunoreactivity and GFAP-mRNA, respectively, in the ipsilateral hypoglossal nucleus 4 and 7 days after axotomy in control experiments. These increases failed to occur in axotomized animals treated with ARA-C. Therefore, our data are compatible with the hypothesis that activation of astrocytes following axotomy as measured by increased expression of GFAP and its mRNA is induced secondarily to the microglial response.

The envelope glycoprotein of human immunodeficiency virus type 1 stimulates release of neurotoxins from monocytes

Mononuclear phagocytes infected with human immunodeficiency virus 1 (HIV-1) produce soluble factors that kill neurons in culture. To define the molecular events that lead to neuron killing, HIV-1 proteins were tested for the ability to trigger release of neurotoxins from human monocytes and lymphocytes. None of the recombinant-derived HIV-1 proteins examined (reverse transcriptase, protease, gag, nef, or gp120) were directly neurotoxic at concentrations from 100 pM to 10 nM. The envelope glycoprotein gp120 did, however, stimulate both isolated human blood monocytes and the monocytoid line THP-1 (but not lymphocytes or the lymphoid cell line H9) to discharge neurotoxic factors. These toxins consisted of heat-stable, protease-resistant molecules (< 500 Da) that copurified with neurotoxins from HIV-1-infected THP-1 cells and were blocked by antagonists to N-methyl-D-aspartate receptors. Release of neurotoxins through gp120 stimulation involved monocytoid CD4 receptors because toxin production could be inhibited either by a monoclonal antibody to the CD4-binding region of gp120 or by soluble CD4 receptors. Alternatively, production of neuron-killing factors could be induced with a peptide from the CD4-binding region of gp120. These data show that the HIV-1 envelope glycoprotein alone can stimulate neurotoxin release by binding to CD4 receptors of mononuclear phagocytes. Such neurotoxic factors may, in turn, contribute to the central nervous system dysfunction associated with HIV-1 by acting on neurons through N-methyl-D-aspartate receptors.

Influence of interleukin-1 and tumor necrosis factor alpha on the growth of microglial cells in primary cultures of mouse cerebral cortex: involvement of colony-stimulating factor 1

The influence of monokines and CNS-derived colony-stimulating factors (CSF) on the growth of microglia has been studied in mixed glial primary cultures stemming from mouse embryos. We observed that spontaneous growth of microglial cells in the presence of astrocytes is blocked by adding anti-colony-stimulating factor 1 (CSF-1) antibodies to the cultures. Both interleukin-1 (IL-1) and tumor necrosis factor-alpha(TNF alpha) strongly increased the number of microglial cells in mixed glial cultures and this effect was prevented by anti-CSF-1 antibodies. In contrast, anti-interleukin-3 (IL-3) or anti-granulocyte-macrophage colony-stimulating factor (GM-CSF) antibodies did not significantly affect the in vitro growth of microglia. These results provide functional significance to astrocytic productions of CSF-1 and their modulations by IL-1 or TNF alpha.

The possible contribution of microglia and macrophages to delayed neuronal death after ischemia

Macrophages have long been known to be involved in cytotoxic actions in many tissues in the body following infection. Knowledge of the post-injury actions of blood-borne macrophages in the brain, and their resident counterparts, the microglia, have been limited to the "mopping-up" of cellular debris. However, other functions are now coming to light and there is evidence that they contribute to both growth promotion and cytotoxicity following injury in the brain. This review raises the possibility that macrophages may contribute to delayed neuronal death following ischemia. Growth factors including certain cytokines produced by these cells protect against ischemia-induced neuronal death. In contrast, cytokines can also induce macrophages to synthesize nitric oxide synthase and indoleamine-2,3-dioxygenase which results in the production of the cytotoxins nitric oxide and quinolinic acid. It is hypothesized that viable cells produce or concentrate growth factors which prevent the induction of these enzymes, whereas damaged cells cannot.

Cytotoxicity of microglia

The most characteristic property of microglia is their swift activation in response to neuronal stress and their capacity for site-directed phagocytosis. The transformation of microglia into intrinsic brain macrophages appears to be under strict control and takes place if neuronal and/or terminal degeneration occurs in response to nerve lesion. The differentiation of microglia into brain macrophages is accompanied by the release of several secretory products, e.g., proteinases, cytokines, reactive oxygen intermediates, and reactive nitrogen intermediates. Interference with the microglial activation or the productions of cytotoxic metabolites by microglia may thus offer new therapeutic opportunities for the prevention of neuronal cell death in CNS disease.

The microglial reaction in the rat hippocampus following global ischemia: immuno-electron microscopy

Transient arrest of the cerebral circulation leads to neuronal cell death in selectively vulnerable regions of the central nervous system. It has recently been shown at the light microscopical level that neuronal necrosis is accompanied by a rapid microglial reaction in ischemia (Gehrmann et al. (1992) J. Cereb. Blood Flow Metab. 12:257-269). In the present study we have examined the postischemic microglial reaction in the dorsal rat hippocampus at the ultrastructural level using immuno-electron microscopy. Global ischemia was produced by 30 min of four-vessel occlusion and the microglial reaction then studied after 8, 24 and 72 h. In sham-operated controls microglial cells were not phagocytic; they were randomly distributed throughout the neuropil and occasionally made contacts with other structures such as dendrites in CA1. Ultrastructural signs of activation were observed from 1 day postlesion onward. Reactive microglial cells were consistently seen to phagocytose degenerating neurons particularly in the CA1 stratum pyramidale and in the CA4 sector. They were sometimes interposed between two morphologically distinct types of CA1 neurons, i.e., "dark" (degenerating) and "pale" (surviving) types of neurons. Phagocytic microglial cells also became positive for major histocompatibility complex (MHC) class II antigens at these locations from 1 day after ischemia onward. Furthermore, activated microglial cells were frequent along degenerating dendrites in the stratum radiatum of CA1. After survival times of up to 72 h microglial cells, but not astrocytes, were occasionally observed to undergo mitosis. In addition to their random distribution across the neuropil, microglial cells were frequently observed in a perivascular position under normal conditions.(ABSTRACT TRUNCATED AT 250 WORDS)

Evidence for activation of the complement cascade in the hypoglossal nucleus following peripheral nerve injury

Following hypoglossal nerve transection in adult rats, immunoreactivity for complement factor C3 and one of its degradation products C3d as well as C4d and immunoglobulin G (IgG) was observed in the ipsilateral hypoglossal nucleus. Double-labelling experiments indicated that these antigens were present in perineuronally located reactive microglial cells. In addition, increased levels of complement factor C3-mRNA was found in perineuronally located cells ipsilateral to nerve lesion. These results suggest that the complement cascade is locally activated in the vicinity of axotomized neuronal perikarya and that microglial cells have a key role in this process, alternatively that C3, C3d, C4d and IgG are involved in other so far unknown processes.

Interleukin 1 and tumor necrosis factor-alpha stimulate the production of colony-stimulating factor 1 by murine astrocytes

Astrocytes have the ability to secrete colony-stimulating factor 1 (CSF-1), a growth factor known to stimulate the proliferation of brain macrophages. We have studied the effect of cytokines such as interleukin 1 (IL-1), tumor necrosis factor-alpha (TNF alpha), and interleukin 6 (IL-6) on the production of CSF-1 by cultured primary astrocytes and an astrocytic cell line derived from embryonic mouse brain. We observed that both TNF alpha and IL-1 increased CSF-1 mRNA and protein levels in the astrocytic cultures. In contrast, IL-6 was ineffective. The CSF-1 mRNA levels were strongly reduced by incubating immortalized astrocytic cells with staurosporine, a protein kinase C inhibitor, both in the absence and in the presence of cytokines. Conversely, 12-O-tetradecanoylphorbol 13-acetate, a protein kinase C activator, increased CSF-1 mRNA levels. These results suggest a mechanism whereby mononuclear phagocytes could favor their own recruitment in the CNS by producing cytokines.

Macrophage-induced cytotoxicity of N-methyl-D-aspartate receptor positive neurons involves excitatory amino acids rather than reactive oxygen intermediates and cytokines

The co-localization of activated macrophages and damaged neurons observed in brain injury and degenerative brain diseases may hint to macrophage-induced neuronal cytotoxicity. Recently, macrophages have been found to secrete neurotoxic molecules such as radical oxygen intermediates and glutamate, the latter interacting with N-methyl-D-aspartate (NMDA) receptors. As shown in the present study, brain macrophages termed microglial cells co-cultured with differentiated cerebellar neurons excert potent neurotoxic effects. Neurotoxicity is unlikely to be due to cytokines since tumor necrosis factor (TNF)-alpha, interleukin (IL)-1 beta, IL-6 and interferon (IFN)-alpha/IFN-beta/IFN-gamma had no such effects. In contrast, when treating neurons with H2O2 or oxygen radical-generating systems cytotoxicity was induced. Furthermore, microglia were found to produce O2- and H2O2 when triggered with phorbol 12-myristate 13-acetate. However, in co-cultures of neurons and microglia, oxygen-radical scavengers catalase and superoxide dismutase, failed to protect neurons from microglia-induced killing. Moreover, when using undifferentiated neurons which are susceptible to H2O2 but not to NMDA receptor-dependent killing, microglia did not destroy the neurons. Thus, the amount of reactive oxygen intermediates produced by microglia in co-culture do not reach the critical concentrations required for neurotoxicity. As dibenzocyclohepteneimide, an antagonist to NMDA receptors neutralized neurotoxicity in microglia-neuronal co-cultures, excitatory amino acids released by microglia are suggested to compose the major determinant of neurotoxicity.