Search

Aquaporin-4 (AQP4) is located on astrocyte endfeet that face blood vessels in the brain and in the pia. It is thought to play a crucial role in the development of brain edema. To confirm the notion that sex steroids and dexamethasone influence brain edema through AQP4 regulation, we investigated the effects of 17β-estradiol, testosterone, and dexamethasone on the expression of AQP4 in cultured astrocytes. Testosterone significantly up-regulated AQP4 at the level of both protein and mRNA. At a concentration of 100 nM, testosterone significantly increased AQP4 protein levels and ameliorated the osmotic fragility of astrocytes from hypoosmotic stress, suggesting that the increased levels of AQP4 facilitated the testosterone function. Moreover, this effect was attenuated by the protein kinase C activator 12-O-tetradecanoylphorbol 13-acetate, which can rapidly decrease AQP4 mRNA expression, indicating that the response was specific. These results indicate that AQP4 can alter the osmotic fragility of astrocytes and that testosterone can influence brain edema through AQP4 regulation, whereas 17β-estradiol and dexamethasone cannot. © 2003 Wiley-Liss, Inc.

Microglia are well known to become activated during various kinds of neuropathological events. The factors that are responsible for the activation, however, are not fully determined. In the present study, L-Ser was shown to enhance production of nitric oxide (NO), interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF alpha) by lipopolysaccharide (LPS)-stimulated cultured rat microglial cells. L-Ser, however, did not enhance the expression of mRNAs encoding inducible NO synthase, IL-6 and TNF alpha. On the other hand, astrocytes did not depend on L-Ser for release of IL-6 and TNF alpha. The expression of an enzyme 3-phosphoglycerate dehydrogenase (3PGDH), which is essential for L-Ser biosynthesis from a glycolytic intermediate 3-phosphoglycerate, was investigated. As revealed by Western blotting and immunocytochemical staining, 3PGDH-protein expression in vitro was the highest in astrocytes, intermediate in neurons and the lowest in microglial cells. Semiquantitative RT-PCR showed that microglial cells expressed 3PGDH-mRNA at a lower level than astrocytes. In frozen sections from rat forebrain, only astrocytes were immunoreactive for 3PGDH. The present study suggested that L-Ser is able to modulate microglial function mainly at the translation level because microglial cells cannot synthesize sufficient amount of L-Ser due to the scarce expression of 3PGDH.

Cytochrome c was detected by immunoblotting in the cytosolic fraction 3 h after 5-min ischemia in the non-ischemia-tolerant CA1 region in which about 96% of neurons had developed delayed neuronal death, while less cytosolic cytochrome c was detected in the ischemia-tolerance-induced CA1 region where many more neurons survived. In the immunohistochemical study using anti-non-native cytochrome c monoclonal antibody, immunoreactivity was observed throughout the cytoplasm in the non-ischemia-tolerant CA1 neurons, but not in the normal and ischemia-tolerant CA1 neurons. Then we determined whether Bcl-2, Bax, Bcl-xL and Bcl-xS, which regulate the release of cytochrome c from mitochondria, were altered in the ischemia-tolerant CA1 region. Bcl-2 and Bax were up-regulated in the ischemia-tolerant group, but Bcl-xL and Bcl-xS showed no apparent difference in their expression. These results suggest that cytochrome c release is prevented in CA1 neurons in gerbils in which ischemia-tolerance had been induced and that the altered ratio of Bcl-2 to Bax may play a part in this mechanism.

Reactive oxygen and nitrogen species (RO/NS) such as nitric oxide (NO), hydroxyl radical (OH.), and superoxide anion (O(2)(-)) are generated in a variety of neuropathological processes and damage neurons. In the present study, we investigated the neuroprotective effects of rat astrocytes against RO/NS-induced damage using neuron-glia cocultures, and the effects were compared to those of microglial cells. Sodium nitroprusside (SNP), 3-morpholinosydnonimine (SIN-1), and FeSO(4) were used to generate NO, O(2)(-) and NO, and OH., respectively. Solely cultured neurons, which were transiently exposed to these agents, degenerated, possibly through apoptotic mechanisms as revealed by in situ detection of DNA fragmentation, whereas neurons cocultured with either astrocytes or microglial cells were viable even after exposure to RO/NS. In contrast, most neurons cocultured with meningeal fibroblasts degenerated. Astrocyte-conditioned medium partially attenuated RO/NS-induced neuronal damage. When neurons were cultured on astrocyte-derived extracellular matrix (AsECM), neuronal death induced by SNP and FeSO(4) was almost completely inhibited. AsECM contained significant amounts of laminin and fibronectin, and pure fibronectin and laminin also protected neurons against RO/NS-induced damage in the same manner as AsECM. These results suggest that astrocytes can protect neurons against RO/NS-induced damage by secreting soluble and insoluble factors.

Although microglial cells are well known to become activated in the pathological brain, mechanisms underlying the microglial activation are not fully understood. In the present study, with an aim to elucidate whether neurons are involved in the microglial activation, we compared the morphology and the superoxide anion (O2−)-generating activity of rat microglial cells in pure culture with those of cells cocultured with rat primary cortical neurons. Microglial cells in pure culture in serum-free Eagle's minimum essential medium on poly- l -lysine-coated coverslips displayed ramified morphology and suppressed activity of O2−generation. In contrast, microglial cells in neuron–microglia coculture under the same conditions as those for the pure culture displayed ameboid shape and upregulated activity of O2−generation. Electron microscopic observation revealed that microglial cells in coculture were more abundant in Golgi apparatus and secretory granules than those in pure culture and that some of microglial cells in the vicinity of neurites exhibited membrane specialization reminiscent of a junctional apparatus with high electron density between a microglial soma and a neurite. Microglial cells in coculture tended to tie neurites in bundles by extending processes. Medium conditioned by neurons significantly enhanced O2−generation by microglia, but microglial cells in contact with or in close apposition to cocultured neurons were much more intensely activated than those remote from the neurons. Furthermore, the membrane fraction of cortical neurons activated microglial cells, and this effect was abolished by treating the neuronal membrane with trypsin or neuraminidase. In conclusion, neuronal–microglial contact may be necessary to mediate microglial activation. The present findings suggest that the contact of microglia with damaged neurons in the brain is a plausible cause to activate microglia in the neuropathological pro cesses.

Cultured microglial cells usually exhibit ameboid morphology and peripheral macrophage-like properties, which are distinct from those observed in the normal mature brain. This might be caused by the inappropriate culture of microglial cells in high concentrations (approximately 200-400 microM) of Gly and Ser, although the concentrations of the amino acids in extracellular spaces of the brain parenchyma are quite low (approximately 5 microM). In the present study, we focused on the concentration-dependent effects of glycine (Gly) and serine (Ser) on microglial morphology and function. Under Gly/Ser-free and serum-free condition, the majority of rat microglial cells displayed round morphology, whereas in the presence of 5 microM Gly and 25 microM Ser, which correspond to the concentrations of Gly and Ser in the cerebrospinal fluid, they extended multiple branched processes and formed clusters of rough endoplasmic reticulum. On the other hand, Gly and Ser did not affect morphology of astrocytes. The viability of microglia was not affected by the changes in the concentrations of Gly and Ser. Metabolic activity, activities of acid phosphatase and inducible nitric oxide synthase, and superoxide anion (O2-) generation were all strongly suppressed in Gly/Ser-free medium or in medium containing physiological concentrations of both amino acids. Such activities were all enhanced in harmony with increases in the concentrations of Gly and Ser. Thus, microglial cells cultured in Gly/Ser-free medium, even though exhibiting ameboid morphology, appears to be in the functionally resting state. Furthermore, once the resting state was achieved, the microglial cells remained inactive even after the subsequent 24 h culture in serum-supplemented medium containing 400 microM of both amino acids. The medium conditioned by microglial cells that were cultured in the presence of 400 microM of Gly and Ser was toxic to cortical neurons, whereas the microglia-conditioned medium obtained in the absence of both amino acids facilitated the survival of cortical neurons. Therefore, microglial cells in the resting state, which was induced in the Gly/Ser-free condition, are likely to support neurons. Microglial cells could ramify on glass coverslips coated with astrocyte-derived extracellular matrix or on coverslips coated thinly with fibronectin and/or laminin even under the Gly/Ser-free condition. The ramified cells as induced in this way kept suppressed O2- generating activity. These findings suggest that resting ramified microglial cells with a neurotrophic activity can be induced with the combination of Gly/Ser-free medium and small amounts of extracellular matrix proteins, and that the resting state is rather stable.

Microglia transform from ameboid to ramified cells during development and display an ameboid appearance again under certain pathological conditions. Some cytokines produced by astrocytes may be responsible for the microglial transformation. In the present study, we compared the effects of cytokines, granulocyte/macrophage colony-stimulating factor (GM-CSF), macrophage colony-stimulating factor (M-CSF), and interleukin-3 (IL-3) on the morphology of rat cultured microglia. For quantitative evaluation, we employed "transformation index" as calculated by (perimeter of cell)2/4 pi (cell area). GM-CSF facilitated the ramification of cultured rat microglia, which was effectively induced in a serum-free medium. However, M-CSF and IL-3 did not induce the ramification. A certain serum adhesion protein (possibly vitronectin) as well as other high molecular weight substances in fetal calf serum inhibited the GM-CSF-induced microglial ramification. Among ordinary supplements for a chemically defined medium, progesterone, insulin, and a high concentration of glucose suppressed the ramification. These findings suggest that GM-CSF may be involved in microglial ramification and that many kinds of supplements that are added to culture media profoundly affect the morphology of microglial cells.

Because of variations in the morphology and function of microglial cells, it has often been claimed that microglial cells should be classified into two or more subtypes. However, such subtypes have not fully been characterized. In the present study, we isolated microglial cells expressing microglia-markers CD11b and CD68 from rat mixed glial cultures on the fifth and on the thirteenth days in vitro (DIV 5 and 13) and demonstrate that these two populations of microglial cells have distinct morphology and function. Microglial cells isolated on DIV 5, which we have termed immature cells, are characterized by the presence of large somata, large peroxidase- and alkaline phosphatase-positive granules, and high proliferative activity and suppressed responsiveness to lipopolysaccharide (LPS). In contrast, the microglial cells isolated on DIV 13, which we have termed mature cells, are devoid of granules, appear to be in a state of cell cycle arrest, and respond to LPS by the induction of inducible nitric oxide synthase (iNOS), tumor necrosis factor-α, and interleukin-6. Isolated immature cells maintained in pure culture failed to express iNOS in response to LPS. However, if these cells were cultured on astrocyte-derived extracellular matrix (AsECM) or pure laminin, the cells exhibited an induction of iNOS in response to LPS. AsECM and laminin were also able to induce a state of cell cycle arrest in cultured isolated immature cells. Thus, classification into two types of microglial cells is possible, but both types are in the same cell lineage, because the immature cells can differentiate into mature microglial cells in the presence of laminin or AsECM. Therefore, “microglioblasts” may be the appropriate term for the immature cells. © 2003 Wiley-Liss, Inc.

The non-essential amino acids l -serine (Ser) and glycine (Gly) have recently been shown to exhibit specific actions in the nervous system. In the present study, l -Ser and Gly promoted the survival of cultured rat cerebrocortical neurons in a concentration-dependent manner as revealed by Alamar blue assay and microtubule-associated protein-2 (MAP2) immunoblotting. The maximum effects of the amino acids were detected at the concentrations of 30–100 μM. l -Ser was more effective than Gly. d -Ser failed to promote neuronal survival. l -Ser and Gly upregulated expression of the anti-apoptotic gene product Bcl-w, while they did not affect the expression of Bcl-xL. The promotion of neuronal survival by l -Ser and Gly may be, at least in part, attributable to the upregulated Bcl-w.

We examined the cytochrome c oxidase (COX) activity in gerbil hippocampal CA1 neurons after 5-min ischemia by a histochemical method in the presence or absence of exogenous cytochrome c. In the CA1 neurons, COX activity without exogenous cytochrome c decreased from 1 h after ischemia, but was restored by the addition of exogenous cytochrome c in the following 6 h after ischemia. These results suggest that it is not COX activity but endogenous cytochrome c that is changed in the early phase after ischemia, and that COX activity begins to decrease 9 h after ischemia. We examined caspase-3 in the CA1 region by immunoblotting, as caspase-3 is known to take part in the cell-death cascade downstream from cytochrome c. Although pro-caspase-3 was strongly detected, active caspase-3 was not detected before and until 84 h after 5-min ischemia. Our data suggested that delayed neuronal death is likely to progress via cytochrome c-release but not via caspase-3 activation.

We examined cytosolic cytochrome c in gerbil hippocampal CA1 and CA3 regions after induction of 5-min ischemia by immunoblotting. In the CA1 region, cytochrome c was detected in the cytosolic fraction from 1 to 6 h after ischemia by Western blotting, while it was not detected in the CA3 region. Following intraventricular administration of cyclosporin A (CsA), detectable cytosolic cytochrome c was dramatically decreased, and about 80% of CA1 neurons survived after ischemia. The present studies demonstrate that cytochrome c is translocated from mitochondria to the cytosol in the early stage of delayed neuronal cell death, and suggest the involvement of the mitochondrial permeability transition.

It is believed that ramified resting microglial cells in the brain are differentiated from macrophage-like ameboid cells, although the mechanism for the differentiation is not fully understood. In the present study, we investigated whether the differentiation of microglial cells is observable in mixed brain cell culture prepared from newborn rat forebrains. In confluent mixed brain cell culture, both ramified and ameboid microglial cells were simultaneously present. The ramified cells were located in or under the astrocyte monolayer, while the ameboid cells were over the layer as revealed by confocal laser scan microscopy. The majority of ramified cells appeared after the astrocyte layer was completely formed and they downregulated the expression of the major histocompatibility complex antigen. Fibronectin was detected around ramified microglial cells, and laminin was also present in the astrocyte monolayer in mixed brain cell culture, while both proteins were not distributed near ameboid cells over the monolayer. When purified microglial cells were cultured on astrocyte-derived extracellular matrix in serum-free medium, they ramified. These results show that the differentiation of microglial cells is observable in culture and that astrocytes may play pivotal roles in the differentiation mainly by secreting insoluble factors.

Activated microglial cells and peripheral macrophages are hardly distinguishable from the viewpoints of morphology and function. There are various immunological markers common to both microglial cells and peripheral macrophages. In the present study, however, we found that microglial cells have distinct characters in terms of adhesion and morphology. By using a "rheoscope," that is an apparatus to rheologically measure the strength of cell adhesion to substrates, rat microglial cells were found to attach to polystyrene dishes much more weakly than alveolar and peritoneal macrophages. Interferon-gamma (IFNgamma) strengthened the adhesion of alveolar and peritoneal macrophages, whereas it weakened that of microglial cells. Morphological changes of microglial cells induced by IFNgamma were also different from those of peripheral macrophages. Furthermore, alveolar and peritoneal macrophages produced NO in response to IFNgamma, while microglial cells did not. When cultured on astrocyte-derived extracellular matrix (AsECM) in serum-free medium, only microglial cells extended multiple ramified processes. Conversely, alveolar and peritoneal macrophages on AsECM shrunk their ruffling membrane and rounded up. These distinctions between microglial cells and macrophages may reflect differences in cell lineages as well as environments in which individual cells reside.

Apoptotic neuronal death is known to occur in the developing brain and in the mature brain of patients with ischemic and degenerative disorders. Although microglial cells are known to become activated in specific conditions, it has not been elucidated whether they enhance or prevent neuronal apoptosis. The present study was intended to observe how microglial cells are involved in neuronal death. When rat primary cortical neurons were incubated with a nitric oxide (NO) donor sodium nitroprusside (SNP; 300 microM) for 10 min, neuronal death occurred 12-16 hr later. The NO-induced neuronal death was inhibited by cycloheximide, and the SNP-treated neurons were characterized by nuclear fragmentation and intact cell membrane under electron microscopy. Agarose gel electrophoresis demonstrated DNA fragmentation of the SNP-treated neurons. Thus, the NO-induced neuronal death appeared to be apoptosis. When neurons were cocultured with rat primary microglial cells, the SNP treatment failed to induce the neuronal death. Because microglia-conditioned medium also prevented apoptotic neuronal death, microglial cells were considered to secrete antiapoptotic factors. The microglia-conditioned medium rescued neurons even when they were added to neuronal cultures after the SNP treatment, implying that the factors acted on neurons in a manner other than scavenging NO. Interleukin-3, interleukin-6, macrophage colony-stimulating factor, and basic fibroblast growth factor, which are known to be secreted by microglial cells, were not effective in preventing NO-induced neuronal death. Among microglia-derived substances, tumor necrosis factor alpha and plasminogen, which are heat-labile proteins, inhibited neuronal apoptosis. The neuroprotective action of the microglia-conditioned medium, however, still remained, even after it was heated. These findings suggest that microglial cells protect neurons against NO-induced lethal damage by secreting heat-labile and heat-stable neuroprotective factors in vitro.

Effects of steroid hormones on the regulation of function and morphology of microglial cells were investigated using the cultured cells isolated from forebrain of newborn rats. Cortisol, corticosterone, and aldosterone at 100 nM caused a strong shrinkage of microglial cells cultured in a serum-supplemented medium. However, cholesterol, pregnenolone, testosterone, estradiol, and dehydroepiandrosterone did not exhibit any significant effects. The corticosteroids also inhibited the GM-CSF-mediated ramification of microglia in a serum-free medium. An anti-glucocorticoid agent RU38486 abolished the effects of corticosteroids on the microglial morphology, suggesting the presence of functional glucocorticoid receptor (GR) in microglial cells. The presence of GR was confirmed by immunoblotting with an antibody to the receptor. Cytokines GM-CSF and interleukin-3 altered the level of GR expression. Binding experiments with [3H]-corticosterone demonstrated the presence of not only GR but also mineralocorticoid receptor (MR): the dissociation constants (Kd) and the number of binding sites (Bmax) were 0.8 nM and 15 fmol/mg protein for MR and 5.0 nM and 73 fmol/mg protein for GR, respectively. The pure glucocorticoid RU28362 and dexamethasone at 20 nM (but not aldosterone and corticosterone at the same concentration) inhibited proliferation of microglial cells, as revealed by PCNA immunocytochemistry. RU28362 inhibited the activities of inducible nitric oxide synthase and acid phosphatase at concentrations higher than 1 nM. Aldosterone and corticosterone exhibited the similar inhibitory effect at 100 nM, and this inhibition was completely overcome by RU38486. On the other hand, corticosterone and aldosterone at concentrations lower than 1 nM enhanced the activities of both enzymes. The antimineralocorticoid agent spironolactone eliminated the stimulatory effects of corticosterone on the enzyme activities. In accordance with these biochemical results, electron microscopic observations revealed that glucocorticoids enhanced the formation of lysosomal vacuolation in microglial cells and aldosterone increased the number and size of lysosomes. In conclusion, it is suggested that GR and MR mediated the opposite effects of corticosterone on the functions of microglial cells; the hormone acted as an inhibitor through GR and as an stimulator through MR.