Immunohistochemical localization of macrophages and microglia in the adult and developing mouse brain

Proliferation of differentiated glial cells in the brain stem

Classical studies of macroglial proliferation in muride rodents have provided conflicting evidence concerning the proliferating capabilities of oligodendrocytes and microglia. Furthermore, little information has been obtained in other mammalian orders and very little is known about glial cell proliferation and differentiation in the subclass Metatheria although valuable knowledge may be obtained from the protracted period of central nervous system maturation in these forms. Thus, we have studied the proliferative capacity of phenotypically identified brain stem oligodendrocytes by tritiated thymidine radioautography and have compared it with known features of oligodendroglial differentiation as well as with proliferation of microglia in the opossum Didelphis marsupialis. We have detected a previously undescribed ephemeral, regionally heterogeneous proliferation of oligodendrocytes expressing the actin-binding, ensheathment-related protein 2'3'-cyclic nucleotide 3'-phosphodiesterase (CNPase), that is not necessarily related to the known regional and temporal heterogeneity of expression of CNPase in cell bodies. On the other hand, proliferation of microglia tagged by the binding of Griffonia simplicifolia B4 isolectin, which recognizes an alpha-D-galactosyl-bearing glycoprotein of the plasma membrane of macrophages/microglia, is known to be long lasting, showing no regional heterogeneity and being found amongst both ameboid and differentiated ramified cells, although at different rates. The functional significance of the proliferative behavior of these differentiated cells is unknown but may provide a low-grade cell renewal in the normal brain and may be augmented under pathological conditions.

Tangential migration of ameboid microglia in the developing quail retina: mechanism of migration and migratory behavior

Long distance migration of microglial precursors within the central nervous system is essential for microglial colonization of the nervous parenchyma. We studied morphological features of ameboid microglial cells migrating tangentially in the developing quail retina to shed light on the mechanism of migration and migratory behavior of microglial precursors. Many microglial precursors remained attached on retinal sheets containing the inner limiting membrane covered by a carpet of Müller cell endfeet. This demonstrates that most ameboid microglial cells migrate tangentially on Müller cell endfeet. Many of these cells showed a central-to-peripheral polarized morphology, with extensive lamellipodia spreading through grooves flanked by Müller cell radial processes, to which they were frequently anchored. Low protuberances from the vitreal face of microglial precursors were firmly attached to the subjacent basal lamina, which was accessible through gaps in the carpet of Müller cell endfeet. These results suggest a mechanism of migration involving polarized extension of lamellipodia at the leading edge of the cell, strong cell-to-substrate attachment, translocation of the cell body forward, and retraction of the rear of the cell. Other ameboid cells were multipolar, with lamellipodial projections radiating in all directions from the cell body, suggesting that microglial precursors explore the surrounding environment to orient their movement. Central-to-peripheral migration of microglial precursors in the retina does not follow a straight path; instead, these cells perform forward, backward, and sideways movements, as suggested by the occurrence of (a) V-shaped bipolar ameboid cells with their vertex pointing toward either the center or the periphery of the retina, and (b) threadlike processes projecting from either the periphery-facing edge or the center-facing edge of ameboid microglial cells.

Identification, isolation, and characterization of daintain (allograft inflammatory factor 1), a macrophage polypeptide with effects on insulin secretion and abundantly present in the pancreas of prediabetic BB rats

A bioactive macrophage factor, the polypeptide daintain/allograft inflammatory factor 1 (AIF1), has been isolated from porcine intestine. It was discovered when searching for intestinal peptides with effects on insulin release, and its purification was monitored by the influence of the peptide fractions on pancreatic glucose-induced insulin secretion. Daintain/AIF1 is a 146-aa residue polypeptide with a mass of 16,603 Da and an acetylated N terminus. An internal 44-residue segment with the sequence pattern -KR-KK-GKR- has a motif typical of peptide hormone precursors, i.e., dibasic sites for potential activation cleavages and at the sequentially last such site, the structure GKR. The latter is a signal for C-terminal amide formation in the processing of peptide hormones. Daintain/AIF1 is immunohistochemically localized to microglial cells in the central nervous system and to dendritic cells and macrophages in several organs. A particularly dense accumulation of daintain/AIF1-immunoreactive macrophages was observed in the insulitis affecting the pancreatic islets of prediabetic BB rats. When injected intravenously in mice, daintain/AIF1 at 75 pmol/kg inhibited glucose (1 g/kg)-stimulated insulin secretion, with a concomitant impairment of the glucose elimination, whereas at higher doses (7.5 and 75 nmol/kg), daintain/AIF1 potentiated glucose-stimulated insulin secretion and enhanced the glucose elimination. Its dual influence on insulin secretion in vivo at different peptide concentrations, and the abundance of macrophages expressing daintain/AIF1 in the pancreatic islets of prediabetic rats, suggest that daintain/AIF1 may have a role in connection with the pathogenesis of insulin-dependent diabetes mellitus.

Expression of LFA-1alpha and ICAM-1 in the developing rat brain: a potential mechanism for the recruitment of microglial cell precursors

Several studies agree that microglial cells derive from monocytes that infiltrate the central nervous system during development, but the precise mechanism by which these cells enter into the nervous tissue is still unknown. In this way, the aim of the present study was to analyze the expression of two cell adhesion molecules involved in the recruitment of blood leukocytes into tissues, the lymphocyte function-associated antigen-1alpha (LFA-1alpha) and the intercellular adhesion molecule-1 (ICAM-1) in the developing rat brain (from E16 to P18). By means of immunohistochemistry, our observations showed that LFA-1alpha and ICAM-1 were expressed in the developing rat brain with a definite distribution pattern and a characteristic time course of appearance. In the embryonic period, LFA-1alpha immunoreactivity was displayed not only by intravascular blood cells but also by intraparenchymal round cells with a horseshoe-shaped nucleus, showing the typical morphological features of monocytes. Monocyte-like cells present in the embryonic brain parenchyma often displayed mitotic profiles. LFA-1alpha immunohistochemistry also revealed the presence of some LFA-1alpha-positive cells belonging to the ameboid microglial population (mostly in the white matter from E18). In the postnatal period, LFA-1alpha immunoreactivity was displayed by some ameboid microglial cells (P0-P9) and also by some ramified microglia. LFA-1alpha immunoreactivity observed in ramified microglia was weaker when compared to LFA-1alpha stained ameboid microglia. In contrast, ICAM-1 immunolabeling during the embryonic period was mainly located in endothelial cells of parenchymal brain blood vessels (principally from day E18). Blood vessels in choroid plexus and meninges also expressed ICAM-1 during the embryonic time. In postnatal animals, ICAM-1 immunoreactivity was found in relation to endothelial cells of blood vessels, but the density of ICAM-1-positive blood vessels was lower than that during the embryonic period. The gradual regulation in the expression of LFA-1alpha by monocyte-like cells and cells of the microglial lineage, and the expression of ICAM-1 by the brain vasculature strongly suggest that the LFA-1/ICAM-1 system may be a mechanism involved in the entry of microglial cell precursors into the developing rat brain.

Immunological aspects of kaolin-induced hydrocephalus

Adult female Sprague Dawley rats were administrated 0.1 ml Kaolin (250 mg/ml) into cisterna magna. One, 4 and 8 weeks later, brains were analyzed using antibodies against MHC class I (OX18), MHC class II (OX6), CD4 (OX38), CD8 (OX8), OX42, ED1, NF, GFAP, AChE and TH. Remarkably high numbers of T lymphocytes, and OX42- and ED1-positive macrophages were found aggregated in subarachnoid spaces, and in the third and fourth ventricles. Marked aggregations of ED1-positive reactive microglial cells were also found in paraventricular structures, medial septum, retrosplenic cortex and commissural structures. However, no such cells were found in hippocampus. ED1-positive areas were also positive for round cells with a rim of MHC I fluorescent cytoplasm as well as for OX42-positive cells and MHC II positive microglial cells. At week 1, in ventro-frontal areas of cortex, CD8-positive cells and MHC I positive astroglial fibers were detected. At week 1, MHC I positive ramified microglial cells were also recognized in almost the entire brain. These positive cells gradually decreased with time and finally remained rounded with a rim of fluorescent cytoplasm. In addition, ED1 positive partly ramified microglial cells could be recognized in corpus callosum, probably representing cells in transition between ramified and reactive microglia. CD8+ cells entered ventral brain structures, and were found in the horizontal diagonal band at week 4, and had disappeared at week 8. Finally in cortex, ED1 positive microglial cells could be identified only in the retrosplenic cortex, and there were also "dark shrunken neurons" in light microscopic stainings. However, there was only a moderate GFAP positive gliosis. In conclusion, kaolin-induced hydrocephalus leads to immune reactions in several defined areas such as cholinergic systems, corpus callosum, circumventricular organs, pontine cerebellar peduncles and the vestibular nucleus.

Identification of transient microglial cell colonies in the forebrain white matter of developing rats

Herein, we describe the existence of distinct colonies of transient microglial cells that reside in well-defined zones of the forebrain white matter. Rats, aged at postnatal day (P) 0, P2, P5, P7, P10, P15 or adult, were anaesthetised with halothane gas, and various neural centres were injected unilaterally with the tracer biotinylated Dextran. The neural centres injected were cingulate or sensorimotor cortices, ventral nuclei of the dorsal thalamus, and the pontine reticular formation of the brainstem. Rats were allowed to survive to various stages, from 4 hours to 21 days, after the injection. They were then anaesthetised with sodium pentobarbitone, and their brains were aldehyde-fixed and processed by using standard methods. The following is a description of what is seen after injections at P0, P2, P5, P7, P10; we saw no labelled cells (described below) in the rats injected at P15 or adult. From 2 to 21 days after an injection of dextran into the above-mentioned centres, labelled microglial cell colonies, identified by using double-labelling with anti-OX-6 or Griffonia simplicifolia (Bandeiraea; isolectin B4), were seen in small isolated zones in the forebrain white matter. These colonies were in the corpus callosum, the dorsal and ventral regions of the external capsule, and the internal capsule. A striking feature of these labelled microglial cell colonies was that they were seen on both sides of the brain. Thus, regardless of the location of the injection site in either the cortex, thalamus, or brainstem, the same microglial cell colonies were labelled with dextran in the forebrain white matter. After injections of different coloured fluorescent dextrans into the cortex and into the brainstem of the same animal, many double-labelled cells in each of the colonies were seen. From our short-term survival cases (4 hours to 1 day), a rather strict sequence or progression of labelling of the colonies across the white matter from the injection site was seen; in general, the microglial cell colonies closest to the injection site became labelled well before (about a day) those further away. These results lead us to suggest that the microglial cells in each colony become labelled after a slow diffusion of the tracer through the extracellular space from the injection site.

Glial and endothelial cell response to a fetal transplant of purified neurons

Astrocytes, microglia and endothelial cells display very specific phenotypic characteristics in the intact adult CNS, which appear quite versatile when grown in culture without neurons. Indirect evidence from in vitro co-culture studies and analysis of the effects of specific neuronal removal in vivo, does accordingly favour a role of neurons for the phenotypic repression of these cells in the intact brain. In order to provide more direct evidence for such neuronal influence, we attempted to induce, in the rat brain, a reversal of the post-lesional activation of astrocytes, microglia and endothelial cells by transplantation of fetal neurons purified by immunopanning. Host microglial cells which have been activated by the lesion process, penetrated the neuronal graft during the few days after the transplantation. Reactive astrocytes began to appear in the lesioned parenchyma and gathered around the transplant. Thereafter they first sent their processes in the direction of the neuronal graft, before they migrated into the graft a few days later. At this time, which was at the end of the first week post-transplantation, the host endothelial cells sprouted "streamers" of basal lamina within the graft forming small capillaries. During the second week post-transplantation, numerous astrocytes and microglial cells, both displaying a reactive hypertrophied morphology, were observed throughout the grafts. Finally, by the end of the first month, the activated cells differentiated towards a quiescent, resting morphology. At this time the grafts contained a vascular network with morphological characteristics comparable to those observed in the intact brain parenchyma. The results indicate that the interaction of activated astroglia and microglia and endothelial cells with neurons causes the cells to re-differentiate and regain phenotypic features characteristic of intact brain parenchyma, strongly suggesting that neurons play an essential role in the phenotypic restriction of glial and endothelial cells in the adult central nervous system.

Microglia responses in the CNS following sciatic nerve transection in C57BL/Wld(s) and BALB/c mice

The present study employed C57BL/Wld(s) mice to investigate whether a delay in microglia reaction would occur similar to the delay that occurs in macrophage response after sciatic neurectomy. The results were compared with control BALB/c mice. The observations showed that in both strains of mice there was no delayed microglia response around lesioned motoneurons and around the central processes of the dorsal root ganglion cells after sciatic neurectomy in the adult. The increased Mac-1 staining appeared as early as 1 day postoperation (dpo). This indicates that microglial cells and macrophages respond to different signals generated by neurectomy. In both strains of mice, the number of microglia in the neonate was much less than that in the adult and the increase in Mac-1 staining was detectable only at 3 dpo in both strains of mice. A significant loss of motoneurons was detected after sciatic neurectomy in the neonate. However, there were no significant differences in the mean percentages of motoneuron loss between the two strains of mice at 5, 10, and 15 dpo. It is surmised that the lack of an adequate number of mature microglia in the neonates and their tardy expression of CR3 antigenicity may contribute to the motoneuron loss.

Early generation of glia in the intermediate zone of the developing cerebral cortex

Radial glia are present at the earliest stage of cerebral cortical development, and later they transform into astrocytes. Other glial cells including astrocytes and oligodendrocytes are thought to appear only after neuron generation is complete and the cortical layers are formed. Little is known of when and where microglia enter the central nervous system and proliferate. We addressed the question of the origin of these three glial cell types in the developing ferret cerebral cortex. We assessed the temporal pattern of glial cell division by administering [3H]thymidine to label cells in S phase, and by using survival periods of 1-2 h to label dividing cells in situ. Labeled cells were identified in the developing intermediate zone of the ferret cerebral wall. These cells were present at E28, and reached a maximum number at P1. Double labeling experiments identified these cells as astrocytes, oligodendrocytes or microglia. None of the dividing cells expressed neuronal markers. These data show that all three types of glia are generated in the developing subcortical white matter, and that glial progenitors are present in the intermediate zone as soon as it becomes a recognizable structure. These data also show that the period of glial generation overlaps extensively with the period of neuron generation, since neuron generation is not complete until the end of the second postnatal week in the ferret.

Rapid upregulation of the Pi isoform of glutathione-S-transferase in mouse brains after withdrawal of the neurotoxicant, cuprizone

Cuprizone intoxication has been used as a model for reversible demyelination in the CNS. During the course of cuprizone intoxication, the glutathione-S-transferase isoform, Pi, normally and oligodendrocytic marker, appears in reactive astrocytes (Cammer ad Zhang, 1993). The present experiments address the changes in expression of Pi after removal of cuprizone from the diet of the affected mice. In order to localize Pi message, a riboprobe was prepared and in situ hybridization (ISH) performed. Western blots and immunocytochemistry were used to examine Pi protein and other glial cell markers. The data indicated that Pi protein increased during the first 2 d after withdrawal of the toxicant, when the level of the myelin marker, 2',3'-cyclic nucleotide-3'phosphohydrolase, remained minimal. Results of ISH suggested that levels of Pi message in the corpus striatum decreased during cuprizone feeding and began to recover within 2d after withdrawal of the toxicant. Both microglia and astrocytes appeared during the first week of cuprizone administration and persisted during two to three additional weeks on cuprizone. Reactive astrocytes remained in the tissue for at least 6 wk after cuprizone was withdrawn, while microglia receded within days. The findings suggest that astrocytes continue to express Pi after withdrawal of cuprizone.

[Morphological study of Organum cavum pre-lamina terminalis (OCPLT), implicated in thirst in rats]

It has been shown that micro-injections of angiotensin II in the Organum cavum pre-lamina terminalis (OCPLT) of the rat brain trigger water intake. Our anatomical study shows that the anterior, vertical extension of this Organum corresponds to the anterior inter-hemispheric fissure, and its posterior, horizontal extension to the Cavum septi pellucidi (CSP). Contrary to classical descriptions, the CSP persists in the adult rat, though reduced to one third of its initial antero-posterior extension. It appears rostrally as an opening into the anterior inter-hemispheric fissure. The absence of ependymal cells and of communication with the lateral ventricles points to the fact that this diverticulum is not a median cerebral ventricle. Its pial sheathing and the presence of leptomenix in the cavity confirm its inter-hemispheric origin. The pinocytotic vesicles of the meningeal capillaries suggest an active transfer of substances. Circumstantial evidence suggests circulating angiotensin could be one of them.

Labeling of amoeboid microglial cells and intraventricular macrophages in fetal rats following a maternal injection of a fluorescent dye

Amoeboid microglial cells (AMC) in fetal brains were labeled by rhodamine B isothiocyanate (RhIc) when injected intravenously or intraperitoneally into mother rats at late state of pregnancy. The fluorescent cells were immunostained with antibodies OX-42 and OX-18 that recognize complement type 3 (CR3) receptors and major histocompatibility complex class I (MHC-I) surface antigen, respectively. RhIc-labeled AMC were first observed in the cavum septum pellucidum and subependymal cysts associated with the cerebral aqueduct as well as the fourth ventricle, and subsequently at other sites including the corpus callosum and other subcortical white matter. The fluorescence intensity increased with time after RhIc administration so that after 1 day the cells were brightly labeled. The majority of the labeled cells were round, with some elongated ones bearing two or three processes. Besides AMC, macrophages in the ventricular system were also labeled. All fluorescent cells were double labeled with OX-42 and OX-18 antibodies. Present results suggest that when introduced into the maternal circulation, RhIc could readily gain access into the fetal brain through the inefficient placental, blood-brain and blood-cerebrospinal-fluid (blood-CSF) barriers. The avid uptake of RhIc in circulation by brain macrophages indicates an active scavenging role of these cells in fetal brain. The labeling of cells by maternal route offers a rapid method for study of distribution of brain macrophages in fetuses.

Normal and reactive NG2+ glial cells are distinct from resting and activated microglia

We have previously used antibodies to the NG2 proteoglycan and the alpha receptor for platelet-derived growth factor (PDGF alpha receptor) to identify oligodendroglial progenitor cells in vivo and in vitro. It has recently become evident that the GD3 antigen, which has been widely used as a marker for oligodendrocyte progenitor cells, is also expressed by microglial cells. In this study we have examined the relationship between the NG2+/PDGF alpha receptor+ glial progenitor cells and microglial cells in normal developing and mature rat brain and in inflammatory lesions in mice with experimental autoimmune encephalomyelitis (EAE). Double-labeling of sections from normal rat brain using anti-NG2 antibodies and lectin from Griffonia simplicifolia (GSA I-B4) or monoclonal antibody 4H1 indicated that there is no overlap between NG2+ glial progenitor cells and microglia in the parenchyma of the central nervous system. In EAE lesions, both NG2+ cells and microglia, identified by antibodies to F4/80 and CD45, displayed reactive changes characterized by increased cell number and staining intensity and shortening and thickening of cell processes. Both cell types were found surrounding perivascular infiltrates of lymphocytes. Double-labeling EAE sections for NG2 and F4/80 or CD45 failed to reveal cells that co-expressed both antigens, suggesting that reactive NG2+ cells are distinct from activated microglia. However, a close spatial relationship between NG2+ cells and microglia was observed in the normal brain and to a greater extent in EAE, where processes of an activated microglial cell were sometimes seen to encircle an NG2+ cell. These observations are indicative of a functional interaction between microglia and the NG2+ glial cells.

Genetic influences on cellular reactions to brain injury: activation of microglia in denervated neuropil in mice carrying a mutation (Wld(S)) that causes delayed Wallerian degeneration

This study examines the relationship between the appearance of degenerative changes in synaptic terminals and axons and the activation of microglia in denervated neuropil regions of normal mice of the C57BL/6 strain and mutant mice (Wld(S)), in which Wallerian degeneration is substantially delayed. The time course of degenerative changes in synaptic terminals and axons was assessed using selective silver staining. Microglial cells were identified by immunostaining for Mac-1, a monoclonal antibody to the CR3 complement receptor, and by histochemical staining for nucleoside diphosphatase (NDPase). Increased argyrophilia, indicative of degenerative changes, was evident as early as 1 day postlesion in normal mice, but was not seen until 6-8 days in mice with the Wld(S) mutation. Microglial activation in normal C57BL/6 mice was evident by 24 hours postlesion, as evidenced by increased immunostaining for Mac-1, increased histochemical staining for NDPase, and morphological changes indicative of an activated phenotype (short, thick processes). Quantitative evaluation of immunostaining for Mac-1 revealed that peak activation occurred between 2 and 6 days postlesion with a return to a quiescent phenotype by 12 days. In contrast, the microglial response was significantly delayed and prolonged in mice bearing the Wld(S) mutation. Activated microglia were not seen within the deafferented area until 6 to 8 days postlesion and peak activation occurred between 12 and 20 days postlesion. These data suggest that the response of microglia in denervated neuropil zones is triggered by the same types of degenerative changes that cause increased argyrophilia as detected by selective silver staining methods.

Impaired neuroglial activation in interleukin-6 deficient mice

Astrocyte activation is a ubiquitous hallmark of the damaged brain and has been suggested to play an important regulatory role in the activation, survival, and regeneration of adjacent neurons, microglia, and oligodendrocytes. Little is known, however, about the endogenous signals that lead to this activation of astrocytes. Here we examined the regulation of interleukin 6 (IL6), a proinflammatory cytokine, its receptors, and the effects of IL6-deficiency in a model of traumatic central nervous system injury in the axotomized mouse facial motor nucleus. Facial nerve transection led to a massive but transient upregulation of IL6 mRNA in the disconnected motor nucleus, while IL6-receptor subunits were constitutively expressed on motoneurons and astrocytes. Absence of IL6 in genetically IL6-deficient mice led to massive reduction in the number of activated GFAP-positive astrocytes, a more moderate decrease in microglial activation and proliferation, and an increase in the late neuronal response to axotomy. These results emphasize the role of IL6 in the global regulation of neurons, astrocytes, and microglia and their activation in the injured nervous system.

Soman-induced seizures rapidly activate astrocytes and microglia in discrete brain regions

Neurons in the piriform cortex and the pontine nucleus locus coeruleus express elevated levels of the immediate early gene protein product, Fos, within 30-45 minutes of a seizurogenic dose of the anticholinesterase, soman (Zimmer et al., [1997] J. Comp. Neurol. 378:468-481). By 24 hours following soman injection, there is marked neuropathology in the piriform cortex. These findings suggest selective, regional vulnerability in response to the seizurogenic actions of soman. In the present study, we determined that soman-induced seizures also cause selective, rapid activation of astrocytes and microglia in the piriform cortex and other brain regions. Animals were killed at different intervals between 1 hour and 24 hours after a convulsive dose of soman. Brain sections were processed for immunocytochemical detection of astrocytes with antibodies against glial fibrillary acidic protein, and microglia and macrophages with antibodies against the complement receptor 3 protein, OX-42. The results demonstrate that following soman administration: (1) there is a rapid increase in glial fibrillary acidic protein staining in astrocytes of the piriform cortex (1 hour); (ii) reactive astrocytes are specifically restricted to layer II and the superficial boundaries of layer III of the piriform cortex. These are the same layers in which neurons express Fos within 30-45 minutes following soman administration; (3) between 1 and 4 hours, resting (ramified) microglia in the piriform cortex and the hippocampus alter their morphology to resemble active microglia. From 4-8 hours, active microglia undergo morphological changes characteristic of reactive microglia that resemble macrophages. Taken together, these observations indicate that astrocytes and microglia in brain regions susceptible to soman become rapidly "reactive" in response to seizures. The highly specific anatomical codistribution of reactive glia and Fos-expressing neurons suggests that intensely active neurons provide local signals that trigger reactive changes in neighboring glia.

Early activation of microglia in the pathogenesis of fatal murine cerebral malaria

Microglia are pluripotent members of the macrophage/monocyte lineage that can respond in several ways to pathological changes in the central nervous system. To determine their role in the pathogenesis of fatal murine cerebral malaria (FMCM) we have conducted a detailed study of the changes in morphology and distribution of retinal microglia during the progression of the disease. Adult CBA/T6 mice were inoculated with Plasmodium berghei ANKA. These mice died 7 days post inoculation (p.i.) with the parasite while exhibiting cerebral symptoms, increased permeability of the blood-brain barrier, and monocyte adherence to the vascular endothelium. Mice were injected i.v. with Monastral blue 2 h prior to sacrifice to identify "activated" monocytes, and their isolated retinae were incubated with the Griffonia simplicifolia (GS) lectin or reacted for the nucleoside diphosphatase enzyme to visualize microglia and the vasculature. Changes in microglial morphology were seen within 2-3 days p.i., that is, at least 3 days prior to the onset of cerebral symptoms and 4 days before death. Morphological changes included retraction of ramified processes, soma enlargement, an increasingly amoeboid appearance, and vacuolation. There was also increased staining intensity and redistribution of "activated" microglia toward retinal vessels, but no increase in density of NDPase-positive cells. The GS lectin only labeled a small population of microglia in the uninfected adult mouse retina. However, there was a striking increase in the focal density of GS-positive microglia during the progression of the disease. Extravasation of monocytes also was observed prior to the onset of cerebral symptoms. These results provide the first evidence that microglial activation is a critical component of the pathological process during FMCM.

Microglia from the developing rat medial septal area can affect cholinergic and GABAergic neuronal differentiation in vitro

The normal development of the central nervous system is regulated by glia. In this regard, we have reported that astrocytes, stimulated by epidermal growth factor or transforming growth factor alpha, suppress the biochemical differentiation of rat medial septal cholinergic neurons in vitro, as evidenced by a decrease in choline acetyltransferase activity. In this study, we found that, in contrast to astrocytes, microglia enhance rather than suppress this aspect of cholinergic cell expression. When in excess, microglia can revert the effects of epidermal growth factor on the septal cholinergic neurons without altering the astroglial proliferative response to this growth factor. In the absence of growth factors or other glial cell types, microglia increase choline acetyltransferase activity above control levels and thus, may be a source of cholinergic differentiating activity. The increase in enzyme activity induced by microglia is rapid in onset, detected as early as 2 h after their addition to the septal neurons and maintained up to six or seven days in vitro. Furthermore, in the absence or presence of other glial cell types, microglia also influence septal GABAergic neurons by significantly increasing glutamate decarboxylase activity. As microglia affect neither septal cholinergic nor GABAergic neuronal cell survival, they appear to enhance the biochemical differentiation of these two neuronal cell types. Specific immunoneutralizing antibodies were used to identify the microglia-derived factors affecting these two neuronal types. In this regard, we found that the microglia-derived cholinergic differentiating activity is significantly suppressed by antibodies raised against interleukin-3. Furthermore, interleukin-3 was detected in both conditioned media and cell homogenates from septal neuronal-microglial co-cultures by western blotting. Finally, although basic fibroblast growth factor and interleukin-3 significantly increase septal glutamate decarboxylase activity, neither appears to be implicated in the GABAergic cell response to the microglia. In conclusion, these results demonstrate that microglia can enhance the biochemical differentiation of developing cholinergic and GABAergic neurons in vitro.

Microglial cells internalize aggregates of the Alzheimer's disease amyloid beta-protein via a scavenger receptor

Microglia are immune system cells associated with Alzheimer's disease plaques containing beta-amyloid (A beta). Murine microglia internalize microaggregates of fluorescently labeled or radioiodinated A beta peptide 1-42. Uptake was confirmed using aggregates of unlabeled A beta detected by immunofluorescence. Uptake of A beta was reduced by coincubation with excess acetyl-low density lipoprotein (Ac-LDL) or other scavenger receptor (SR) ligands, and Dil-labeled Ac-LDL uptake by microglia was blocked by excess A beta. CHO cells transfected with class A or B SRs showed significantly enhanced uptake of A beta. These results show that microglia express SRs that may play a significant role in the clearance of A beta plaques. Binding to SRs could activate inflammation responses that contribute to the pathology of Alzheimer's disease.

Astrocyte reactivity in neonatal mice: apparent dependence on the presence of reactive microglia/macrophages

In neonatal mice, an acute injury produced by a stab wound to the cortex results in minimal astrocyte reactivity, as has been observed by others. However, if the source of the stab wound, a piece of nitrocellulose (NC) membrane, were now implanted in the cortex for a period of time (chronic NC implant injury), then extensive astroglial reactivity in the neonatal brain ensues. The astrogliosis is manifested by increased mRNA, protein content, and immunoreactivity for GFAP, and by ultrastructural changes. Given the previous reports that inflammatory cytokines are possible mediators of astrocyte reactivity (e.g., Balasingam et al: J Neurosci 14:846, 1994), we examined the brain parenchyma of neonatal mice following an NC stab or implant injury, with minimal or extensive astrogliosis, respectively, for a possible differential representation of inflammatory cells. A significant correlation (r = 0.87, P < 0.05) was observed between the occurrence of astrogliosis and the presence of reactive microglia/macrophages; no other inflammatory cell type was detected in the brain parenchyma of neonatal mice following NC implant injury. We suggest that reactive microglia/macrophages are required for the evolution of cells into reactive astrocytes following insults to the neonatal brain.

A spontaneously immortalized mouse microglial cell line expressing CD4

We have derived a microglial clone, named C8-B4, from the 8-day mouse cerebellum organ culture which gave rise to distinct astroglial cell lines as previously reported. Indeed, the C8-B4 clone expresses classical microglial markers (MAC1, F4/80, 2-4G2) and appears to be derived from a committed microglial precursor since it does not express differentiation antigens present during the early stage of the monocytic lineage. This microglial clone expresses two characteristics not previously reported for microglial cell lines: it synthesizes the CD4 molecule and produces and releases large amounts of glutamate.

Carbonic anhydrase II in microglia in forebrains of neonatal rats

In the brains of adult rodents carbonic anhydrase II (CA) immunoreactivity has been observed in the choroid plexus and in oligodendrocytes, astrocytes, and myelin. Localization and functions of CA in the neonatal brain, however, have been controversial. One issue is whether the CAII-immunopositive round and ameboid cells in the corpus callosum and cingulum in the rat CNS during the first postnatal week are oligodendrocytes or microglia. Colocalization of CAII with the microglial antigen, ED1, and the microglia-specific isolectin, BSI-B4, suggested that most (approx. 60%) of the CAII-positive round and ameboid cells in rat brain during the first postnatal week were, indeed, macrophages and microglia. During that initial week, some CAII-positive protoplasmic astrocytes (approx. 40%) were observed as well. At the end of the first postnatal week smooth-surfaced CAII-positive cells began to appear in the corpus callosum. Those cells also bound MAbO4, a marker for the oligodendrocyte cell line. We conclude that during the first postnatal week most of the CAII-positive cells are macrophages and microglia, and that some are protoplasmic astrocytes. During the second postnatal week CAII-positive cells in the oligodendrocyte lineage become apparent, and by the end of that week there are few CAII-positive microglia. Confocal microscopy suggests that in brains of three-day-old rats the ameboid microglia are associated with nerve fibers, where they may perform phagocytosis of axons, directional guidance of axons, or disinhibition of axonal growth.

Microglia in the nerve fiber layer of the cat retina: detection of postnatal changes by a new monoclonal antibody

This paper describes changes in the appearance and distribution of microglia in postnatal cat retina as demonstrated by a new antibody, H386F. This fractionated IgM antibody was created via an intrasplenic immunization of a single BALB/C mouse with about 2-3 x 10(5) large, whole cells isolated from 46 minced cat retinae. To confirm that the labeled cells are microglia, the staining properties of H386F were compared with those of four commercially available antibodies, OX-33, OX-41, OX-42, and ED-1, that have been used by others to distinguish between microglia and other cells in rat brain. These experiments show that H386F is the only antibody of the five to label only microglia in both the cat retina and hippocampus.

Differential immune responses to fetal intracameral spinal cord and cortex cerebri grafts

While the central nervous system (CNS) has been characterized as an immunologically privileged site, there are also several reports describing immunological reactions within the CNS. A certain degree of immunological privilege has also been ascribed to the anterior chamber of the eye. We have used the intraocular transplantation model to study immunological reactions in transplants of embryonic neural tissue. Outbred Sprague-Dawley rats and inbred Fisher rats were used. Pieces of rat parietal cortex or the cervical spinal cord were prepared from embryonic day 14 and implanted into the eye chambers of adult rats of the same strain. Following intraocular maturation, grafts were analysed using antibodies against: major histocompatibility complex (MHC) class I, MHC class II; rat antigens CD4, CD8, CD11b; T-cell receptor; rat antigen ED1; and glial fibrillary acidic protein. Using this set of markers for immunological reactions, transplants were scored on a blind basis. We found no significant differences in immunological scores between transplants obtained from different litters of fetuses of the outbred animals. Grafting in the outbred strain led to increased numbers of immunologically reactive cells in the grafts. This was not seen in grafts in the inbred strain. Spinal cord transplants led to a significantly higher degree of cytotoxic immunity-related cells expressing MHC class II as well as CD4-positive cells. There was a positive correlation between ED1 negativity and well-developed ramified microglia. From these results we conclude also that well-developed intraocular CNS tissue grafts do contain cellular evidence of immunological events and that different areas of the CNS may provoke different degrees of response. Reactive microglial proliferation appears to be one of the most sensitive ways to monitor the immunological condition of grafted CNS tissue.

Identification of surface glycoconjugates in the olfactory system of turtle

Lectin binding histochemistry was performed on the olfactory system of Pseudemys scripta to investigate the distribution and density of defined carbohydrate terminals on the cell surface glycoproteins of the olfactory receptors and their terminals in the olfactory bulbs. The lectin staining patterns indicate that the receptor cells of the olfactory mucosa are characterized by glycoconjugates containing alpha-D-galactose and N-acetyl-D-glucosamine terminal residues. The vomeronasal receptor cells contain instead alpha-N-acetyl-D-galactosamine, N-acetyl-D-glucosamine and alpha-D-galactose residues. The results demonstrate that the vomeronasal receptor cells contain high density of alpha-N-acetyl-D-galactosamine sugar residues that are not expressed by receptor cells of the olfactory mucosa. The presence of specific glycoproteins, whose terminal sugars are detected by lectin binding, might be related to the chemoreception and transduction of the odorous message into a nervous signal or in the histogenesis of the olfactory system. In fact, the olfactory receptors are the only known neurons in the vertebrate nervous system that undergo a continual cycle of proliferation not only in developing animals but also in mature ones. Moreover the results show that BSA-I-B4, an alpha-D-galactosyl-specific isolectin, targets the terminal sugar residues in the ramified microglial cells.

Glial cells in transected optic nerves of immature rats. I. An analysis of individual cells by intracellular dye-injection

The glial response to Wallerian degeneration was studied in optic nerves following unilateral enucleation in immature rats, aged 21 days old (P21). The three-dimensional morphology of dye-filled glia was determined in intact nerves, at post-enucleation day 21 in normal nerves from untreated P21 rats, by correlating laser scanning confocal microscopy and camera lucida drawings of single cells. In normal and transected nerves, the majority of dye-filled cells comprized astrocytes (54% and 65%, respectively). In normal P21 nerves, the predominant astrocyte form had a complex stellate morphology and had a centrally-located cell body from which branching processes extended randomly. Two other distinct forms were transverse and longitudinal astrocytes, which had a polarized process extension in a plane perpendicular or parallel to the long axis of the nerve, respectively. These forms were recognized in transected nerves also, but astrocytes in transected nerves had a simple morphology on the whole, and extended few, dense processes which branched infrequently. Quantitative analysis of astrocyte morphology confirmed that individual astrocytes underwent considerable remodelling in response to Wallerian degeneration. A prominent reaction was that astrocytes had withdrawn radial processes and extended a greater proportion of processes longitudinally, parallel to the long axis of the nerve and along the course of degenerated axons. A further, notable feature of transected nerves was the development of novel longitudinal forms and of hypertrophic astroglia. These results indicated that all astrocytes became reactive following enucleation and that glial scar formation was not the function of a single astrocyte subtype. Oligodendrocytes in transected nerves had lost their myelin sheaths and appeared as small cells with numerous bifurcating processes which extended radially, but a small number of oligodendrocytes were recognized which apparently supported myelin sheaths (9%, compared to 40% in normal nerves). In addition, there was a significant population of indeterminate cells in transected nerves (26%, compared to 6% in normal nerves) and, although some of these were identified as microglia/macrophages, it was concluded that many were likely to be dedifferentiated oligodendrocytes.

Regulation of thrombospondin in the regenerating mouse facial motor nucleus

Thrombospondin (TSP) is a multifunctional extracellular matrix protein that plays a role in neuronal migration and axonal outgrowth in the developing central nervous system. In the current study we have examined the localization and regulation of TSP immunoreactivity (TSP-IR) during neuronal regeneration in the axotomized facial motor nucleus using Western blotting and light and electron microscopy. Transection of the facial nerve led to a gradual increase in TSP-IR in the regenerating motoneurons, peaking 4-7 days after injury (DAI). In addition to regenerating neurons, axotomy also caused a rapid upregulation of TSP-IR on activated microglia throughout the facial nucleus, with a maximum of 2-3 DAI, and a second increase at 14-21 DAI on microglial aggregates surrounding degenerating motoneurons and in neuronophagic microglia. In summary, injury leads to the induction of thrombospondin on axotomized neurons and activated microglia, peaking at the times of maximal posttraumatic microglial proliferation and during neuronal phagocytosis. Since thrombospondin is a multimodal extracellular matrix protein with a variety of cell attachment sites, thrombospondin might serve to link microglia and injured neurons, followed by microglial proliferation and removal of the neuronal debris.

Interactions of neurotrophic factors GDNF and NT-3, but not BDNF, with the immune system following fetal spinal cord transplantation

Glial cell line-derived neurotrophic factor (GDNF) is known to stimulate survival of dopaminergic and spinal cord motor neurons. However, little is known of the possible immune sequelae of GDNF exposure, or that of other putative trophic factors. To address these questions we utilized in oculo grafts of spinal cord, wherein we could induce different levels of immune responses via allogeneic vs. syngeneic combinations. Adult female Sprague-Dawley and Fisher rats were used as hosts for allogeneic and syngeneic grafts, respectively. Embryonic age 14-15-day-old fetuses were taken from pregnant dams of each strain, and cervical spinal cords were removed and dissected. Pieces of the spinal cord were transplanted into the anterior chamber of the eye within each strain. At 5-day intervals, 0.5 microgram of GDNF, brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3) or cytochrome c (CC) was injected into the anterior chamber of the eye and the sizes of the transplants were measured for the Sprague-Dawley rats. The same injections and measurements, but only for GDNF and CC, were carried out using Fisher rats. As expected, GDNF increased transplant survival and growth in both the Sprague-Dawley and Fisher animals. At day 41-42, all rats were sacrificed. Cameral graft appearance was evaluated by cresyl violet and immunohistochemically using antibodies against neurofilament (NF), calcitonin gene-related peptide (CGRP) and glial fibrillary acidic protein (GFAP). To monitor immune responses, the following monoclonal antibodies were used: OX38 against CD4, OX18 against MHC class I (MHCI), OX8 against CD8, OX6 against MHC class II (MHCII), OX42 against CD11b, R73 against alpha and beta T cell receptor (TcR), and ED1. In the Sprague-Dawley grafts, significantly higher amounts of CD8+, T lymphocyte+, MHCI+ and MHCII+ antigen-presenting cells (APC) were observed in GDNF-treated transplants. These markers were also increased in NT-3-treated groups. There were two types of OX-42+ cells, one was the ordinary ramified microglial cell, the other appeared to be a phagocytic cell, looking like the interstitial proliferating variety. Interestingly, the phagocytic OX-42+ cells had the same distribution as ED1+ and MHCII+ cells. In contrast, there were few immunoreactive cells after GDNF treatment in the inbred Fisher animals, similar to the CC control group. These results suggest that GDNF and to some extent NT-3, can activate the immune system in allogeneic graft combinations, but that these trophic factors do not produce overt rejection, and do not per se induce immune responses.

Specific uptake of intracranial horseradish peroxidase (HRP) by microglial cells in the goldfish

Intracranial injections of horseradish peroxidase (HRP) in the goldfish labeled a population of cells with many similarities to microglia. The fact that neither macroglia nor neurons appeared to be labeled by these injections supports this identification. The labeled cells responded to Wallerian degeneration of the optic paths by accumulating in the optic terminal zones, a cellular behavior which strongly supports their identification as microglia.

Activation of microglia in haemorrhage microzones in human embryonic cortex. An ultrastructural description

We have noted the presence within parietal cortex of microzones of haemorrhage in two 14 week gestation fetuses. These were obtained from mothers with no clinical history indicative of infection or other pathological developmental data. The microzones of haemorrhage were complicated by degenerative changes in adjacent neurones and activation of putative microglial cells while other surrounding regions of developing cortex showed no signs of destructive alterations. This indicates that these zones are not simply the result of the termination procedure. Microglia associated with these haemorrhagic microzones showed increased vacuolization, phagocytic activity and abundant phagosomes. Although the cause of the microzones of haemorrhage is unknown, these observations are of interest since they are the first demonstration of possible activation of microglia in such an early stage of human development. This finding suggests that the presence of microhaemorrhage and associated neuronal death may provide an inducible stimulus of microglial cells during early brain development, even before programmed neuronal death occurs.

Gender influences the magnitude of the inflammatory response within embolic cerebral infarcts in young rats

BACKGROUND AND PURPOSE

The inflammatory response within cerebral infarcts may have an influence on tissue damage. Since old animals with an impaired immune response have decreased inflammation after experimental cerebral infarction, we postulated that female animals with an increased immune response will have an increased inflammatory response after cerebral infarction.

METHODS

Embolic cerebral infarcts were produced by photochemical irradiation of the right carotid artery in 12 female Fischer rats. The inflammatory response within 4-day-old infarcts was quantitated by histology with the use of computer-assisted image analysis and compared with that in 12 male rats from a previous series.

RESULTS

Severe infarcts had the most pronounced inflammatory response. Female rats had an increased inflammatory response in infarcts of all severity, which was statistically significant in severe cerebral infarcts even after adjustment for infarct size. Severe infarcts in males were significantly larger than those in females.

CONCLUSIONS

Gender influences the outcome of embolic cerebral infarcts after photochemical damage to the carotid artery, both in terms of the magnitude of the inflammatory response and infarct size. There are numerous gender-related differences in neurochemicals, cytokine production, and drug metabolism that may influence tissue damage after stroke and responsiveness to therapeutic intervention. The preponderance of male animals in stroke research may produce results not applicable to female stroke patients. The use of female animals will be required to provide adequate models for the study of stroke in women.

The role of macrophage subpopulations in autoimmune disease of the central nervous system

In this review the role of various subpopulations of macrophages in the pathogenesis of experimental autoimmune encephalomyetitis is discussed. Immunohistochemistry with macrophage markers shows that in this disease different populations of macrophages (i.e. perivascular cells, microglia and infiltrating blood-borne macrophages) are present in the central nervous system. These subpopulations partially overlap in some functional activity while other activities seem to be restricted to a distinct subpopulation, indicating that these subpopulations have different roles in the pathogenesis of encephalomyelitis. The studies discussed in this review reveal that immunocytochemical and morphological studies, combined with new techniques such as in situ nick translation and experimental approaches like the use of bone marrow chimeras and macrophage depletion techniques, give valuable information about the types and functions of cells involved in central nervous system inflammation. The review is divided in three parts. In the first part the experimental autoimmune encephalomyelitis model is introduced. The second part gives an overview of the origin, morphology and functions of the various subpopulations. In the third part the role of these subpopulations is discussed in relation to the various stages (i.e. preclinical, clinical and recovery) of the experimental disease.

Removal of cobalt-labeled neurons and nerve fibers by microglia from the frog's brain and spinal cord

We investigated the microglial reaction around cobalt-labeled degenerating neurons and nerve fibers in the frog central nervous system. The aim of these studies was to reveal the routes of migrating microglial cells during debris removal and the effect of seasonal changes on this process in a cold-blooded animal. Oculomotor and spinal motoneurons were filled with cobaltous-lysine complex through their axons. In the torus semicircularis and the isthmic nucleus, neurons were labeled with iontophoretically applied cobaltous-lysine complex through their injured dendrites and axons. The animals were left to survive for 1 to 50 days. During the summer, oculomotor neurons disintegrated by the seventh postoperative day. The debris from the neurons were phagocytosed by microglia-like cells identified by the presence of cobalt in their cytoplasm. Some of these cells were wedged between ependymoglial cells of the cerebral aqueduct, others appeared at the pial surface of the mesencephalon. The speed of this process was twice as fast during the summer as during the winter. Part of cobalt-labeled microglial cells in the torus semicircularis and the isthmic nucleus moved toward the ependyma of the optic ventricle and the cerebral aqueduct, respectively. Cobalt-loaded microglial cells did not move toward the surface in the spinal cord and the deep part of mesencephalic tegmentum, and left the brain probably via blood vessels. We conclude that microglial cells loaded with phagocytosed tissue debris may leave the brain tissue via three routes and their activity depends on the environmental temperature.

Molecular cloning of F4/80, a murine macrophage-restricted cell surface glycoprotein with homology to the G-protein-linked transmembrane 7 hormone receptor family

F4/80 is a monoclonal antibody that recognizes a murine macrophage-restricted cell surface glycoprotein and has been extensively used to characterize macrophage populations in a wide range of immunological studies. Apart from the tightly regulated pattern of expression of the F4/80 antigen, little is known about its possible role in macrophage differentiation and function. We have sought to characterize the molecule at the molecular level, through the isolation of cDNA clones, and now describe the sequence of the F4/80 protein. The primary amino acid sequence demonstrates homology to two protein superfamilies. The NH2-terminal region consists of seven epidermal growth factor-like domains, separated by approximately 300 amino acids from a COOH-terminal region that shows homology to members of the seven transmembrane-spanning family of hormone receptors. The potential role of these distinct domains is discussed with respect to the possible function of the F4/80 molecule.

Phagocytozing ameboid microglial cells studied in a mouse corpus callosum slice preparation

Highly motile brain macrophages/microglial cells were observed in the cingulum and supraventricular corpus callosum, an area termed by del Rio-Hortega the "fountain of microglia." This is the first study that uses time lapse video microscopy in acute cortical brain slices to analyze directly the motile and phagocytic behaviour of these cells. The cells migrated within minutes to the slice surface and actively screened their surrounding with velum-like processes. Dead/damaged cells on the slice surface were contacted by the processes and phagozytozed within minutes. A method to add red blood cells in a defined density was used to observe the phagocytosis.

The immunophenotype of perivascular cells in the human brain

The immunophenotype of perivascular cells (PC) in temporal lobe tissues obtained at autopsy in 48 patients (aged 41-88 years) was characterized using light and electron microscopic immunocytochemistry with a variety of antibodies. In all cases studied, PC bearing CD11c (Ki-M1P) and CD68 (KP1) were distributed throughout the temporal cortex. In addition to Ki-M1P and KP1, the monoclonal antibodies against major histocompatibility complex (MHC) class II antigen (Ag) (LN-3, CR3.43), anti-leucocyte common antigen (LCA), LN-5, Mac 387 were all found in PC with variable immunoreactivity. In contrast, LN-1 and OPD4 were not found in PC, although the former showed nearly constant staining of resting microglia. Semiquantitative analysis disclosed differences in the numbers of cells labeled with the markers in the 21 normal brains (Ki-M1P > KP1 >> LCA, LN-3, LN-5 >> Mac 387). Ultrastructurally, immunoreactivity for Ki-M1P, KP1, and LN-3 was observed in PC with cytoplasm containing dense lysosomal bodies. In brains from patients with Alzheimer's type dementia, PC were seen in the wall of beta-amyloid protein-positive small vessels. However, there was no definite alteration of antigenicity in PC from AD brains compared with those from normal brains. The immunophenotype of PC was similar to that of macrophages, which were observed in the perivascular spaces and the leptomeninges in some normal and diseased brains. In contrast with PC, however, macrophages showed high incidence of labeling for some macrophage markers LN-5 and Mac 387. These findings demonstrate that PC may be a normal constituent of the adult human brain with a variable expression of monocyte/macrophages markers and MHC class II Ag and that PC could be distinguished from resting microglia by their morphology and by their immunophenotype.

NADPH-diaphorase activity in brain macrophages during postnatal development in the rat

NADPH-diaphorase histochemistry, that allows the visualization of cells producing the gaseous intercellular messenger nitric oxide, was used in the study of the forebrain during the first three postnatal weeks in the rat. Subpopulations of NADPH-diaphorase positive neurons were observed at all ages studied. In addition, non-neuronal NADPH-diaphorase-stained cells were detected in the subcortical white matter, and were very numerous in the supraventricular portion of the corpus callosum, and in the internal and external capsules. These cells were present during the first two postnatal weeks, and were especially prominent at the end of the first postnatal week. They were round-shaped and morphologically similar to the brain macrophages, whose phagocytic activity has been shown in previous studies to play a role in naturally occurring cell death and elimination of exhuberant axons. Series of sections adjacent to those stained with NADPH-diaphorase were processed with immunohistochemistry, using two different antibodies (OX-42 and ED-1) that detect macrophagic and microglial markers, and antibodies that recognize the neuronal form of nitric oxide synthase. Furthermore, brain sections from rats at postnatal day 7 were sequentially processed for either OX-42 or nitric oxide synthase immunohistochemistry followed by NADPH-diaphorase histochemistry. The morphological features and distribution of the non-neuronal NADPH-diaphorase-positive cells were superimposable to those obtained with OX-42 and ED-1 immunohistochemistry. In addition, these cells did not display nitric oxide synthase immunoreactivity. Double-labelled NADPH-diaphorase-positive and OX-42-immunoreactive cells were detected at postnatal day 7. The present results show that brain macrophages express NADPH-diaphorase activity during the early stages of the normal postnatal maturation and suggest that nitric oxide produced by brain macrophages could be involved in the development reshaping of the central nervous system.