Role of microglia in multiple sclerosis

Microglia are the resident macrophages of the nervous system. They serve to protect and preserve neuronal cells from pathogens and facilitate recovery from metabolic insults. In addition, they appear to play a role in the neuropathology of noninfectious inflammatory disorders of the central nervous system, especially those that are autoimmune. Presentation of neural autoantigens to autoreactive T cells by microglia and the attendant secretion of proinflammatory cytokines are thought to facilitate the inflammatory process in diseases such as multiple sclerosis. They also serve as scavengers of damaged myelin following death of oligodendrocytes and the destruction of myelin and may, therefore, promote recovery of myelin damaged by the inflammatory insult. This review examines the current controversies on the pathology of multiple sclerosis and the role played by microglia in the development of central nervous system demyelination.     

Differential regulation of myelin phagocytosis by macrophages/microglia, involvement of target myelin, Fc receptors and activation by intravenous immunoglobulins.

Macrophages/microglia are the key effector cells in myelin removal. Differences exist in the amount and time course of myelin uptake in the central (CNS) and peripheral nervous system (PNS), the basis of this difference, however, is not yet clarified. In the present experiments we studied the phagocytosis rate of CNS or PNS myelin by macrophages and microglia in vitro. Additionally, the effects of intravenous immunoglobulins (IVIg) on this process were investigated. In the PNS experiments, sciatic nerves were cocultured with peritoneal macrophages. Optic nerve fragments were used to characterize the myelin-removing properties of microglia. Cocultures with peritoneal macrophages aimed at investigating the differences in phagocytosis between resident microglia and added macrophages. The myelin phagocytosis in sciatic nerve fragments was higher than in optic nerves, indicating differences in the myelin uptake rate between peripheral macrophages and microglia. IVIg increased the phagocytosis of PNS myelin by macrophages, but not by microglia in optic nerves. The addition of peritoneal macrophages to optic nerve fragments did not lead to an increase in the phagocytosis of CNS myelin either. The IVIg induced phagocytosis of PNS myelin by peripheral macrophages was associated with an increased expression of macrophage Fc receptors measured by FACS. Blocking of Fc receptors by anti-Fc receptor antibody reduced the IVIg induced PNS myelin phagocytosis to basic levels, indicating that the induced but not the basic myelin uptake by macrophages is Fc receptor dependent. In contrast to peripheral macrophages, IVIg did not increase Fc receptor density on microglia. These data indicate that phagocytosis of PNS and CNS myelin by macrophages or microglia is differentially regulated. Local factors within the CNS or PNS may affect this process by modulating the surface receptor profile and activation state of the phagocytic cell or the structure of the myelin sheath.     

Astrocytes modulate macrophage phagocytosis of myelin in vitro

Previous work from this laboratory has shown that both macrophages and microglia phagocytize relatively little myelin in vitro under basal conditions. In an effort to better simulate the conditions within the central nervous system (CNS), we have co-cultured these cells with astrocytes, the most numerous of the neural cells in the CNS, and have compared myelin phagocytosis in the co-cultures with that in cells cultured alone. Both macrophages and microglia in company with astrocytes phagocytized about three times as much myelin as controls, as measured by the formation of cholesterol ester, while astrocytes alone showed little evidence of myelin phagocytosis. Astrocyte-conditioned medium increased phagocytic activity in macrophages by 2.3-fold, and by 3.5-fold in microglia. A number of adhesion molecules and extracellular matrices were tested for their effects on myelin phagocytosis. Matrigel was most effective in activating the macrophages, and in the presence of conditioned medium, stimulated these cells to phagocytize as much myelin as when co-cultured with astrocytes. On the other hand, Matrigel inhibited myelin phagocytosis in microglia. These results indicate that activation of macrophages by astrocytes may be due to an adhesion component, as well as to soluble factors secreted by the astrocytes. While microglia were also stimulated by conditioned medium, adhesion to astrocytes or Matrigel induced a downregulation in phagocytic activity.     

Resident microglia and hematogenous macrophages as phagocytes in adoptively transferred experimental autoimmune transferred experimental autoimmune encephalomyelitis: An investigation using rat radiation bone marrow chimeras

Hematogenous macrophages are known to be involved in the induction of tissue damage in the central nervous system (CNS) as well as of clinical symptoms in experimental autoimmune encephalomyelitis (EAE). Although resident microglia can become phagocytic under certain circumstances, little is known about the role of these cells in brain inflammation in vivo. We thus studied EAE in the model of radiation bone marrow chimeras that allows us to distinguish donor-derived hematogenous cells from resident effector cells. Inflammation in the CNS was qualitatively and quantitatively similar in chimeras compared to fully histocompatible Lewis rats. Although activated resident microglial cells were outnumbered four-to sevenfold in EAE lesions by hematogenous macrophages, the number of resident microglia with ingested myelin was equal to that of macrophages containing myelin debris. Phagocytic resident microglia, expressing the macrophage activation marker ED1, showed ramified as well as amoeboid morphology. From our studies the following conclusions can be drawn. First, a considerable proportion of resident microglia upregulated ED1. Second, resident microglia provide a small but substantial source of brain macrophages in EAE as compared to the large influx of macrophages. Third, our results suggest that macroglia, due to their strategic position within the CNS, are more effective in removal of myelin debris compared to hematogenous macrophages. © 1995 Wiley-Liss, Inc.     

Microglia and macrophage activation and the regulation of complement-receptor-3 (CR3/MAC-1)-mediated myelin phagocytosis in injury and disease

Microglia and macrophages play critical roles in the response of the central and peripheral nervous systems (CNS and PNS, respectively) to injury and disease, one of which is the removal of degenerated myelin by phagocytosis. Myelin removal is efficient during Wallerian degeneration, which follows injury to PNS axons, and in CNS autoimmune demyelinating diseases (e.g., multiple sclerosis) but is inefficient after injury to CNS axons. We suggest that inefficient myelin removal results from deficient microglia activation, reflected by the failure to up-regulate Galectin-3/MAC-2 expression, which marks a state of activation correlated with efficient myelin phagocytosis. Surprisingly, whether or not executing myelin phagocytosis, CNS microglia express the alphaM/beta2 integrin complement receptor-3 (CR3/MAC-1), which has the potential of mediating efficient myelin phagocytosis. We hypothesize that CR3/MAC-1 might be present in distinct inactive and active states that determine, respectively, efficient and inefficient CR3/MAC-1-mediated myelin phagocytosis. We present evidence that CR3/MAC-1-mediated myelin phagocytosis is regulated in microglia and macrophages. First, CR3/MAC-1- mediated myelin phagocytosis has complement-dependent and -independent components. Second, an active complement system augments CR3/MAC-1-mediated myelin phagocytosis. Third, anti-alphaM monoclonal antibodies (MAbs) inhibit and anti-beta2 MAbs augment CR3/MAC-1-mediated myelin phagocytosis in the presence and absence of an active complement system. Fourth, an active complement system modulates MAb-induced regulation of CR3/MAC-1-mediated myelin phagocytosis. Overall, MAb-induced phagocytosis regulation might range three- to sevenfold from inefficient to efficient. We suggest that one of the mechanisms underlying MAb-induced phagocytosis regulation is the induction/stabilization of inactive and active conformational changes. Monoclonal antibody-induced phagocytosis regulation must reveal a mechanism by which native extracellular molecules bind to and regulate CR3/MAC-1-mediated myelin phagocytosis in microglia and macrophages.     

Phagocytosis of myelin by microglia in vitro

Previous experiments from this laboratory have shown that peritoneal macrophages in culture will phagocytize myelin. Myelin preopsonized with myelin antibodies is phagocytized to a much greater extent than untreated myelin, indicating that macrophages ingest myelin by an Fc receptor. The present work was undertaken to determine the characteristics of myelin phagocytosis by microglia, the resident macrophages of the central nervous system. Microglia isolated from 4–5 day primary cultures of newborn rat brains were shown to bind and phagocytize myelin labeled in the lipids by ¹⁴C-acetate. Both binding and phagocytosis as shown by the appearance of ¹⁴C-cholesterol ester were greatly increased if labeled myelin was preopsonized with antiserum to myelin basic protein or galactocerebroside. Both preopsonized and untreated myelin were phagocytized more actively by microglia than by peritoneal macrophages under the same culture conditions. Microglia cultured in the presence of GM-CSF showed slightly increased cholesterol ester production from opsonized myelin, but the effect of GM-CSF was significantly greater than myelin pretreated with control serum (34% increase) or untreated myelin (154% increase). There was no significant effect of GM-CSF on myelin phagocytosis by peritoneal macrophages. Cerebrospinal fluid containing immunoglobulin drawn from rabbits with acute EAE also opsonized myelin to increase phago cytosis by microglia, as has been previously shown with peritoneal macrophages. These results indicate that microglia may actively participate in myelin destruction in demyelinating diseases where myelin antibodies or a source of GM-CSF may be present. © 1993 Wiley-Liss, Inc.

1 This article is a US Government work and, as such, is in the public domain in the United States of America.

     

Differential effects of central and peripheral nerves on macrophages and microglia

The poor ability of injured central nervous system (CNS) axons to regenerate has been correlated, at least partially, with a limited and suppressed postinjury inflammatory response. A key cell type in the inflammatory process is the macrophage, which can respond in various ways, depending on the conditions of stimulation. The aim of this study is to compare the activities of macrophages or microglia when encountering CNS and peripheral nervous systems (PNS), on the assumption that nerve-related differences in the inflammatory response may have implications for tissue repair and thus for nerve regeneration. Phagocytic activity of macrophages or of isolated brain-derived microglia was enhanced upon their exposure to sciatic (PNS) nerve segments, but inhibited by exposure to optic (CNS) nerve segments. Similarly, nitric oxide production by macrophages or microglia was induced by sciatic nerve segments but not by optic nerve segments. The previously demonstrated presence of a resident inhibitory activity in CNS nerve, could account, at least in part, for the inhibited phagocytic activity of blood-borne macrophages in CNS nerve as well as of microglia resident in the brain. It seems that the CNS microglia are reversibly immunosuppressed by the CNS environment, at least with respect to the activities examined here. It also appears from this study that the weak induction of early healing-related activities of macrophages/microglia in the environment of CNS might explain the subsequent failure of this environment to acquire growth-supportive properties in temporal and spatial synchrony with the needs of regrowing axons. GLIA 23:181–190, 1998. © 1998 Wiley-Liss, Inc.     

On the role of peripheral macrophages during active experimental allergic encephalomyelitis (EAE)

Experimental allergic encephalitis (EAE) is an experimental autoimmune inflammatory condition of the central nervous system (CNS) that serves as a disease model for multiple sclerosis (MS). The primary effector mechanisms of the immune system leading to tissue destruction during EAE remain still controversial. T-cells, microglia, and macrophages infiltrating the brain parenchyma are suggested to be involved. To clarify the role of these cells during disease Lewis rats were immunised with different immunisation protocols: Immunisation with myelin basic protein (MBP) in complete Freunds adjuvant (CFA) containing high dose of mycobacterial components induced severe disease, whereas immunisation with low dose of mycobacterial components induced only mild disease. Severely and mildly diseased animals were analysed with respect to infiltration of T-cells, macrophages and upregulation of MHC class II molecules on microglia in the brain. All immunised rats showed high T-cell infiltration accompanied by microglia activation. The degree of disease and the infiltration of macrophages varied with dose of adjuvant. Lowering the dose of adjuvant prevented the development of disease but also the influx of peripheral macrophages into the brain without affecting the peripheral T-cell response to the autoantigen. Thus, appearance of (autoreactive) T-cells in the brain and microglia activation were probably not sufficient for development of disease. It can be concluded that peripheral macrophages play an essential or even key role in the pathogenesis of active EAE.     

Chronic stress promotes lymphocyte reduction through TLR2 mediated PI3K signaling in a β-arrestin 2 dependent manner

Reactive macrophages/microglia exert both protective or damaging effects in multiple sclerosis (MS), which contribute to the relapsing-remitting nature of MS. CD163 is considered a marker of M2 (alternatively activated) macrophages. In the MS brain, CD163(+) perivascular macrophages express molecules for antigen recognition and presentation. Here we further investigated the accumulation of CD163(+) macrophages/microglia in the parenchyma of MS brains. CD163 expression pattern was investigated in different lesions of brain tissue specimens from five MS brains and five neuropathologically unaffected controls by immunohistochemistry. In the parenchyma of normal brain samples, immunoreactivity (IR) of CD163 was absent. In acute active lesions and at the rim of chronic active lesions of MS, strong accumulation of CD163(+) macrophages/microglia was seen. In chronic inactive lesions and in the center of chronic active lesion, CD163(+) macrophages/microglia were rare. Further, double-labeling showed that parenchymal and perivascular CD163(+) macrophages/microglia were myelin basic protein positive and HLA-DR(+), suggesting that CD163(+) macrophages/microglia could ingest and present antigen. In addition, in vitro incubating macrophage RAW264.7 cells with myelin turned LPS-induced inflammatory macrophages into an anti-inflammatory phenotype, indicating that myelin basic protein positive, CD163(+) macrophages/microglia in MS might have anti-inflammatory effects. The parenchymal CD163(+) macrophages/microglia, which had the capacity for antigen ingestion and presentation, might contribute to the resolution of inflammation in MS.     

Galectin-3/MAC-2 in experimental allergic encephalomyelitis

The removal of degenerating myelin by phagocytosis is central to pathogenesis and repair in traumatized and diseased nervous system. Galectin-3/MAC-2 is a differentiation and activation marker of murine and human monocytes/macrophages/microglia. Galectin-3/MAC-2, along with MAC-1 that mediates myelin phagocytosis, marks an in vivo activation state in macrophages, which are involved in myelin degeneration and phagocytosis in injured mouse peripheral nerves. In contrast, high levels of MAC-1 but extremely low levels of Galectin-3/MAC-2 are expressed in vivo in injured CNS where myelin degeneration and phagocytosis progress extremely slowly. The present study was aimed at testing whether an activation state marked by Galectin-3/MAC-2 is present in vivo in the CNS of EAE mice concomitant with autoimmune induced myelin degeneration and phagocytosis. EAE was inflicted by mouse spinal cord homogenate. Demyelination was assessed by light microscopy and Galectin-3/MAC-2, MAC-1, and F4/80 expression by immunocytochemistry. We presently document that Galectin-3/MAC-2 expression is up regulated, along with MAC-1 and F4/80, in spinal cords and optic nerves of EAE mice in areas of demyelination and myelin degeneration, in myelin phagocytosing microglia and macrophages. Copolymer 1 (Glatiramer acetate) suppresses EAE, demyelination, and Galectin-3/MAC-2 expression. EAE pathogenesis thus involves a state of activation in microglia and macrophages characterized by the expression Galectin-3/MAC-2 along with MAC-1. Furthermore, the in vivo responses to injury and autoimmune challenge in the CNS differ in the activation pattern of microglia and macrophages with regard to Galectin-3/MAC-2 expression and the corresponding occurrence of myelin degeneration and phagocytosis.     

IL‐4 drives microglia and macrophages toward a phenotype conducive for tissue repair and functional recovery after spinal cord injury

Macrophages and microglia play a key role in the maintenance of nervous system homeostasis. However, upon different challenges, they can adopt several phenotypes, which may lead to divergent effects on tissue repair. After spinal cord injury (SCI), microglia and macrophages show predominantly pro‐inflammatory activation and contribute to tissue damage. However, the factors that hamper their conversion to an anti‐inflammatory state after SCI, or to other protective phenotypes, are poorly understood. Here, we show that IL‐4 protein levels are undetectable in the spinal cord after contusion injury, which likely favors microglia and macrophages to remain in a pro‐inflammatory state. We also demonstrate that a single delayed intraspinal injection of IL‐4, 48 hours after SCI, induces increased expression of M2 marker in microglia and macrophages. We also show that delayed injection of IL‐4 leads to the appearance of resolution‐phase macrophages, and that IL‐4 enhances resolution of inflammation after SCI. Interestingly, we provide clear evidence that delayed administration of IL‐4 markedly improves functional outcomes and reduces tissue damage after contusion injury. It is possible that these improvements are mediated by the presence of macrophages with M2 markers and resolution‐phase macrophages. These data suggest that therapies aimed at increasing IL‐4 levels could be valuable for the treatment of acute SCI, for which there are currently no effective treatments. GLIA 2016;64:2079–2092     

Microglial signal transduction as a target of alcohol action in the brain

Although alcohol is believed to exert deleterious effects on the nervous system in general, its specific effect on the brain's immune system remains poorly understood. In particular, the effects of alcohol consumption on the immune and inflammatory responses in the central nervous system (CNS) have not been extensively investigated. Here, reviewed is the recent progress on how ethanol influences the signal transduction pathways of the inflammatory activation of brain microglia, which are thought to function as the resident immune defense system of the brain. Microglia are the CNS representatives of macrophages, which have the ability to clean up cellular debris. Microglia participate in neuroinflammation in response to various intrinsic or extrinsic stimuli. It has been recently suggested that microglial signal transduction is one of the main targets of ethanol action in the brain: ethanol exposure selectively modulates the intracellular signal transductions of microglia, rather than globally inhibiting signaling pathways in a nonspecific manner. Deregulation by ethanol of the inflammatory activation signaling of microglia may contribute to the derangement of CNS immune and inflammatory responses.     

Selective and antigen-dependent effects of myelin degeneration on central nervous system inflammation

Damage to myelin sheath or oligodendrocytes may precede or even provoke inflammation of the central nervous system (CNS), but the extent to which these degenerative changes affect inflammation remains largely undefined. To study these processes in more detail, we used CNS antigen-specific T cells in the presence or absence of anti-myelin antibodies to induce experimental autoimmune encephalomyelitis (EAE) in transgenic Lewis rats with low-grade subclinical myelin degeneration and associated microglia cell activation, and in wild-type Lewis rats with an intact CNS. We found that myelin degeneration affects the localization of inflammatory lesions, the numbers of T cells recruited to these lesions, and the severity of the resulting clinical disease. In addition, myelin degeneration and associated microglia cell activation jointly enhance the susceptibility of the CNS to the action of anti-myelin antibodies. Our data show that even subtle alterations of myelin and oligodendrocytes may massively amplify the extent of demyelination and tissue damage, involving different immune effector mechanisms. A similar causal relationship might also operate in human patients with multiple sclerosis, where T cell-mediated inflammation and antibody-mediated demyelination have been documented, and where genetic factors might determine the susceptibility of the target tissue for immune-mediated injury.     

Deficient activation of microglia during optic nerve degeneration

Transection of an optic nerve (ON) is followed by slow removal of myelin. We studied microglia for the expression of molecules that characterize activated myelin phagocytosing macrophages: MAC-1, FcγII/III receptor (FcR), MAC-2, and F4/80. In-vitro, microglia expressed all molecules and phagocytosed myelin. In-vivo, intact ON displayed high levels of MAC-1, little FcR and F4/80, and no MAC-2. The expression of these molecules was upregulated differentially in in-vivo degenerating ON: MAC-1 uniformly, FcR and F4/80 variably, and MAC-2 sporadically. The distribution of MAC-2 expression correlated best with a pattern of sporadic structural degeneration. Thus in-vivo, ON injury is followed by deficient microglia activation, which we suggest contributes significantly to the slow clearance of myelin.     

Inflammation and the brain

Inflammation in the brain selectively damages the myelin sheath resulting in a variety of clinical syndromes of which the most common is multiple sclerosis. In these disorders, the areas of inflammation and demyelination can be identified in life by magnetic resonance imaging. Events occurring at the blood-brain barrier depend on T-cell activation, which increases immune surveillance within the central nervous system. T-cells activated against brain antigens persist to establish the conditions needed for inflammatory demyelination and this depends on local release of cytokines, culminating in removal of oligodendrocytes and their myelin lamellae by macrophages or microglia. These interactions involve binding between receptors present on microglia for the Fc portion of antibody and complement components to corresponding ligands on target cells. Taken together, the evidence from clinical and experimental studies provides a rationale for the issue of immunological treatments in patients with multiple sclerosis.     

Detection of lysosomal cysteine proteinases in microglia: Flow cytometric measurement and histochemical localization of cathepsin B and L

The activation and differentiation of microglia is a prominent pathophysiological process in numerous inflammatory and demyelinating diseases of the central nervous system, including Alzheimer's disease and the AIDS encephalopathy. The tissue damage during these diseases has partly been attributed to lipid peroxidating reactive oxygen intermediates for which activated microglia are a major source. The destruction of tissue may also involve the release of proteolytic enzymes, such as the lysosomal cysteine proteinases cathepsin B and L, which are present notably in phagocytic cells. The cathepsins B and L are endopeptidases with a substrate specificity including important proteins, like myelin basic protein, extracellular matrix components, or the class II major histocompatibility complex. Because of this pathophysiological relevance the cathepsins B and L were chosen for histochemical demonstration in isolated and cultured rat microglia and measurement by a new flow cytometric method. Cathepsin B/L activity was measured flow cytometrically in single viable cells by the intracellular cleavage of non-fluorescent (Z-Phe-Arg)₂-rhodamine 110 to the green fluorescent monoamide Z-Phe-Arg-rhodamine 110 and rhodamine 110. In microglia we measured a cathepsin B/L activity that was 2.5 times higher than in thioglycolate-elicited, i.e., inflammatory peritoneal rat macrophages. In elicited peritoneal macrophages the formation of fluorescent product was 6.2 times higher than in unstimulated resident peritoneal macrophages, demonstrating that the activation and differentiation of mononuclear phagocytes is accompanied by an increased cathepsin B/L enzyme activity. The subcellular localization of cathepsin B/L activity in plated viable microglia was demonstrated histochemically by the use of Z-Ala-Arg-Arg-4-methoxy-2-naphthylamide. Its blue fluorescent cleavage product 4-methoxy-2-naphthylamide was found in lysosomes. Our study shows that activated microglia are an important potential source of cathepsin B/L. This is particularly interesting as enzymatically active cathepsins have recently been found extracellularly at high levels in the senile plaques of Alzheimer's disease, which are known to contain many activated microglia. The release of proteinases by microglia may play a crucial role in the pathomechanism of tissue-destructing diseases in the brain.     

Interleukin-12 induces mild experimental allergic encephalomyelitis following local central nervous system injury in the Lewis rat

The major pathological feature in the central nervous system (CNS) following traumatic brain injury is activation of microglia both around and distant from the injury site. Intraperitoneal administration of interleukin-12 (IL-12) after brain injury resulted in a 7% weight loss, clinical signs of mild EAE and significant myelin basic protein (MBP)-specific splenic cell proliferation. The extent of pathology, in terms of the number of inflammatory perivascular cuffs and activation of microglia was greatest if IL-12 was administered immediately compared to a week following brain injury, whether at one or two sites. Specifically immunostaining for MHC class II and iNOS on macrophages and microglia, ICAM-1 on endothelial cells and macrophages was observed around the site of injury. A degree of myelin processing was apparent from immunostaining of MBP in inflammatory cells distant from the lesion. Inflammatory cuffs comprising macrophages, activated microglia, CD4⁺ T cells and iNOS⁺ cells were also detected distant to the injury site in the medulla and spinal cord of animals treated with IL-12. These results suggest that immune-mediated events in which IL-12 production is stimulated as for example viral infection, superimposed on a brain injury, could provide a trigger for a MS-like pathology.     

Modulation (inhibition and augmentation) of complement receptor-3-mediated myelin phagocytosis.

The removal of damaged myelin is central to repair after injury to axons and in autoimmune demyelinating diseases. Complement receptor 3 (CR3/MAC-1) plays a major role in mediating the phagocytosis of damaged myelin by macrophages and microglia. We studied the modulation (inhibition and augmentation) of CR3/MAC-1 mediated myelin phagocytosis by mAbs that bind to distinct epitopes of subunits alphaM and beta2 of CR3/MAC-1. mAb M1/70 anti-alpha(M) and mAb 5C6 anti-alpha(M) inhibited, whereas mAb M18/2 anti-beta2 augmented myelin phagocytosis. This mAb-induced modulation of myelin phagocytosis occurred in the presence and absence of active complement. Inhibition induced by M1/70 or 5C6 did not add when the two were combined. Combining M1/70 or 5C6 with M18/2 reduced the augmentation induced by M18/2 alone. CR3/MAC-1-mediated myelin phagocytosis may thus be subjected to modulation between efficient and inefficient functional/activation states. These observations and conclusions may offer an explanation for the observed discrepancy between efficient myelin phagocytosis in experimental allergic encephalomyelitis and inefficient myelin phagocytosis after injury to CNS axons, although in both instances macrophages/microglia express CR3/MAC-1.     

Intravenous lipopolysaccharide induces cyclooxygenase 2-like immunoreactivity in rat brain perivascular microglia and meningeal macrophages

Production of prostaglandins is a critical step in transducing immune stimuli into central nervous system (CNS) responses, but the cellular source of prostaglandins responsible for CNS signalling is unknown. Cyclooxygenase catalyzes the rate-limiting step in the synthesis of prostaglandins and exists in two isoforms. Regulation of the inducible isoform, cyclooxygenase 2, is thought to play a key role in the brain's response to acute inflammatory stimuli. In this paper, we report that intravenous lipopolysaccharide (LPS or endotoxin) induces cyclooxygenase 2-like immunoreactivity in cells closely associated with brain blood vessels and in cells in the meninges. Neuronal staining was not noticeably altered or induced in any brain region by endotoxin challenge. Furthermore, many of the cells also were stained with a perivascular microglial/macrophage-specific antibody, indicating that intravenous LPS induces cyclooxygenase in perivascular microglia along blood vessels and in meningeal macrophages at the edge of the brain. These findings suggest that perivascular microglia and meningeal macrophages throughout the brain may be the cellular source of prostaglandins following systemic immune challenge. We hypothesize that distinct components of the CNS response to immune system activation may be mediated by prostaglandins produced at specific intracranial sites such as the preoptic area (altered sleep and thermoregulation), medulla (adrenal corticosteroid response), and cerebral cortex (headache and encephalopathy). J. Comp. Neurol. 381:119-129, 1997. © 1997 Wiley-Liss, Inc.