Main

Two approaches have been used to visualize microglia and monocytes in brain and evaluate changes in their distribution and morphology after injury: histochemistry and immunocytochemistry. A variety of enzyme histochemical markers have recently been developed to distinguish the cell classes originally described by Rio-Hortega(16), including several lectin-based methods. The Griffonia simplicifolia B4-isolectin(17) selectively recognizesα-D-galactose residues contained in membrane glycoconjugates, and histochemical methods based on incubation of tissue sections with this lectin label all microglial cell forms, as well as other macrophages and microvasculature. Immunologic methods rely on antibodies directed against epitopes specific for the cell of interest. Among the most commonly used MAb for detecting monocytic cells is ED-1, which recognizes an uncharacterized cytoplasmic antigen expressed by all cells of the rat monocyte/macrophage lineage. In contrast with lectin histochemistry which labels both resting and activated microglia, ED-1 selectively labels only activated cells. Of note, ED-1 labels activated microglia in normal immature rat brain(15).

In this study, we used a well characterized perinatal rodent stroke model(18, 19) to evaluate the temporal and anatomic features of the microglial response to ischemic neuronal injury in the developing brain. We used both lectin histochemistry and complementary immunocytochemistry to delineate changes in the morphology and distribution of microglia and related cell types after focal ischemic injury, elicited by right carotid ligation and 3-h exposure to 8% oxygen in P7 rats. We found that hypoxic-ischemic forebrain injury rapidly induced microglial activation and subsequent accumulation of activated microglia in lesioned forebrain structures, peaking 2-4 d after injury.

METHODS

Animal lesioning. Sprague-Dawley rats were purchased from Charles Rivers Laboratories and maintained under a 12-h light/dark cycle with free access to food and water.

P7 rats were anesthetized with diethyl ether and underwent right carotid ligation, using previously reported methods(18, 19). After a 1-h recovery period, they were placed in plastic chambers, partially submerged in a warmed water bath (36.5°C), and exposed to 8% oxygen/92% nitrogen for 3 h. One-half hour after hypoxic exposure, they were returned to their dams. All surgical protocols were approved by the University of Michigan Committee on Use and Care of Animals in research.

Four groups of animals (n ≥ 5/group) were killed at 4, 12, 24, and 48 h posthypoxia-ischemia (i.e. after the completion of exposure to 8% oxygen); adjacent sections from these samples were subject to lectin histochemistry and ED-1 immunocytochemistry. In addition, because preliminary experiments suggested that lectin histochemistry was the most sensitive method for detection of the earliest changes in microglial morphology after injury, this assay was also performed in samples from three additional groups of lesioned animals (n ≥ 3/group) killed at 10 min, 2.5 h, and 8 h posthypoxia-ischemia. To assess the subsequent evolution of the microglial response, additional lesioned animals (n ≥ 5/group) were killed 72 and 120 h posthypoxia (P10 and P12), and the distribution of microglia was assayed by lectin histochemistry. In general, animals examined at each time interval were derived from at least two independent experiments. For each age group, two normal control animals were also evaluated.

Tissue preparation. Animals were deeply anesthetized with chloral hydrate and were perfused transcardially with either 4% paraformaldehyde or the PLP fixative(20). Brains were postfixed for at least 12 h in PLP or 4% paraformaldehyde at 4°C, then cryoprotected in graded sucrose solutions, and stored in 20% sucrose in PBS, pH 7.4(21). Preliminary experiments were performed to optimize tissue preservation of friable, acutely lesioned tissue from neonatal animals. The two perfusion solutions yielded identical results and were used interchangeably; most consistent results were obtained if, immediately before tissue sectioning, brains were immersed in a 2:1 solution of 20% sucrose and OCT (Miles, Elkhart, IN) for 45 min, and then were frozen, embedded in this solution, by immersion in isopentane (cooled to -40°C). Twenty-micrometer frozen, coronal sections were mounted on gelatincoated slides and stored overnight at -20°C. Sections were thawed at RT, postfixed for 30 min in 4% paraformaldehyde, encircled with rubber cement, and hydrated and rinsed in 0.1 M PBS, pH 7.4, containing 0.1% Triton X-100 for 15 min.

Histochemistry and immunocytochemistry assays. In each brain, the regional distribution of activated microglia was assessed by two independent methods, a histochemical assay using G. simplicifolia B4-isolectin (GSA I-B4-HRP, Sigma Chemical Co., St. Louis, MO) and an immunocytochemic assay using the mouse MAb ED-1 (Bioproducts for Science, Indianapolis, IN). At least 25 sections/brain were used for each assay. Immunocytochemistry assays were performed using a commercial kit (Vectastain ABC Elite Kit, Vector Laboratories, Burlingham, CA).

Sections used for the lectin histochemistry assay were immersed in 0.3% H2O2 in methanol for 10 min, washed in PBS/0.1% Triton X-100, and then incubated with 60 μL of the G. simplicifolia-B4-isolectin-horseradish peroxidase conjugate, 10 μg/mL, overnight at 4°C. After three washes with PBS/Triton, sections were incubated with 3,3′-diaminobenzidine-H2O2 for 5-8 min. Sections were then rinsed, counterstained, dehydrated, cleared with xylene, and coverslipped. Controls included omission of the lectin or incubation with 10 μg/mL lectin in 0.1 M melibiose (a competitive inhibitor for lectin binding). Concurrent vascular endothelial staining was invariably observed; it was markedly reduced by a 10-min preincubation with 0.3% H2O2 in methanol to block endogenous peroxidase activity.

Sections undergoing immunocytochemistry were preincubated for 20 min with normal serum (from the same species as the secondary antibody), and then incubated with 60 μL of ED-1 (1:200) overnight at 4°C. To prepare negative control samples, equivalent dilutions of mouse IgG1 were substituted for ED-1. Sections were washed three times with PBS/0.1% Triton-X, and then incubated for 30 min at RT with biotinylated rat-adsorbed, horse anti-mouse IgG (Vector). After three PBS/0.1% Triton-X washed, sections were immersed in 0.3% H2O2 in methanol for 10 min to block endogenous peroxidase activity. After three rinses, the sections were incubated with the avidin-biotin enzyme complex (Vector) for 30 min at RT, rinsed in PBS, and reacted with 3,3′- diaminobenzidine in H2O2 (8-12 min). Sections were then rinsed in water, lightly counterstained with cresyl violet, dehydrated through a graded series of alcohols, cleared with xylene, and coverslipped with Permount mounting medium. Nonspecific staining in lesioned tissue, which was noted in preliminary experiments, was eliminated by using a secondary antibody that was preadsorbed with rat tissue.

RESULTS

There is some controversy about the nomenclature of CNS phagocytic cells, arising from questions about their cellular origins(6, 7, 15, 22–27). We used morphologic and immunohistochemical criteria to distinguish cell populations. Resting microglia (also called ramified microglia) were identified by their small cell bodies (about 7 μm) and numerous long thin processes. These cells were labeled with lectin (Fig. 1G), but were nonreactive with the ED-1 antibody. Activated microglia, also termed ameboid microglia(22–25, 27) or brain macrophage(15), were identified by their relatively large size (12-16 μm), stout processes, and intense staining either with lectin histochemistry (Fig. 1E) or ED-1 immunocytochemistry (Fig. 1F). We also identified round cells that had a smooth surface, intermediate size (10-12 μm), and lacked processes, with both the lectin assay and the ED-1 antibody (not shown); although this cell type appeared morphologically distinguishable from the activated microglial cell, immunohistochemical differentiation was not possible with the methods used.

Figure 1
figure 1

This montage compares the distributions of resting and activated microglia in corpus callosum and cortex of normal P8 rat brain, visualized by lectin histochemistry (A and C) and, in adjacent sections, by immunohistochemistry with anti-ED-1 antibody(B and D). In the corpus callosum, lectinstaining labels both resting and activated microglia, as well as endothelial cells(A; area outlined is enlarged in E); in contrast, only activated microglia are ED-1 immunoreactive (B; area outlined is enlarged in F). In the cortex, lectin-staining labels ramified microglia and endothelial cells (C; area outlined is enlarged inG); there are no ED-1-immunoreactive cells (D).E illustrates the distinctive morphologic features of lectin-stained activated microglia (arrowheads), resting microglia(asterisks), and an endothelial cell (star); F demonstrates the uniform morphology of ED-1-immunoreactive cells in corpus callosum (arrowheads); G demonstrates the extensive distribution of lectinstained resting microglia (arrowheads) in cortex. Magnification: A-D, 10×; scale bar, 100 μm;E-G 40×, scale bar = 10 μm.

Full size image

In normal P7-12 rat brain, lectin and ED-1-reactive activated microglia were concentrated in the corpus callosum (Fig. 1, A and B); they were also readily detectable in internal and external capsules, choroid plexus, and ependyma. Occasional round cells (lectin- and ED-1-positive) were also noted in the white matter as well. Resting microglial cells (lectin-positive; ED-1-nonreactive) were distributed rather uniformly throughout the white and gray matter (Fig. 1, C and D).

In this perinatal stroke model, injury is typically confined to the hemisphere ipsilateral to the carotid artery ligation. Histopathologic outcome is somewhat variable; consistent features include evidence of irreversible neuronal injury and/or gross infarction in the hippocampus, cortex, striatum, and thalamus ipsilateral to ligation. A distinctive feature of hippocampal histopathology is the widespread distribution of injury, involving CA subfields CA1 and CA3, as well as the dentate gyrus(18, 19). In P7 rat brain, exposure to moderate hypoxia, alone, does not elicit evidence of tissue injury.

Lectin histochemistry. The lectin histochemistry assay revealed evidence of microglial activation and accumulation in the right hippocampus in the first 8 h after right carotid artery ligation and 3 h of 8% O2 exposure (Fig. 2). Subtle, yet consistent, changes in microglial morphology were evident as early as 10 min after hypoxia-ischemia(the earliest time evaluated); cell bodies were enlarged, and coincidentally their processes began to shorten and thicken (Fig. 2, D and F). These early changes were most prominent in the CA1 and CA3 subfields of the lesioned right hippocampus, especially within the lacunosum moleculare cell layer, in all five brains assayed. At 4 h after the hypoxic-ischemic insult (Fig. 2, G-I), increased numbers of activated microglia were consistently evident in CA1, and to a lesser extent CA3; they were also detected in the thalamus, cortex, and striatum (not shown). At 8-h posthypoxia-ischemia (Fig. 2, J-L) there was more extensive accumulation of activated microglia.

Figure 2
figure 2

This montage presents the earliest evidence of microglial activation and accumulation in the CA1 subfield of the lesioned hippocampus in the first 8 h after right carotid artery ligation and 3 h exposure to 8% O2 in P7 rats. Lectin histochemistry was performed in tissue sections from a normal P7 control, and three lesioned animals, killed 10 min, 4 h, or 8 h posthypoxia-ischemia. The pyramidal cell layer is identified by asterisks in A, D, G, and J; the outlined areas are enlarged in corresponding adjacent panels. In the normal control (A-C), resting microglia (arrowheads) and endothelial cells (star) are stained. In a lesioned animal, killed 10 min posthypoxia-ischemia (D-F), there is subtle evidence of microglial activation. E and F illustrate the two major morphologic features of microglial activation; in E microglial cell bodies are enlarged, in comparison with cells in B, and inF, shortening and thickening of the processes are readily apparent(arrowheads), in comparison with C. In G-I, from an animal killed 4 h posthypoxic-ischemic insult, the majority of microglia(arrowheads) have few identifiable processes. In J-L, from a lesioned animal killed at 8 h posthypoxia-ischemia, the microglial cell bodies are enlarged (arrowhead) and more intensely stained than at the earlier times, and their cell processes are markedly shortened and thickened. In C, F, I, and L, stars identify readily discernible lectin-stained endothelial cells. Magnification: A, D, G, J, 10×, scale bar = 100 μm; B, E, H, K, 20×, scale bar = 100 μm; C, F, I, L, 40×, scale bar = 10 μm.

Full size image

Lectin histochemistry demonstrated progressive increases in the microglial reaction at 24-48 h posthypoxia-ischemia (Fig. 3). At 24 h posthypoxia-ischemia (Fig. 3, A-D), activated microglia infiltrated the pyramidal cell layer and the dentate gyrus of the lesioned right hippocampus. At 48 h posthypoxia-ischemia, in a severely lesioned brain(Fig. 3, E-H), the right hippocampus anatomic landmarks could not be discerned, and activated microglia were densely and diffusely disseminated.

Figure 3
figure 3

These photomicrographs demonstrate progression of the microglial reaction, delineated by lectin histochemistry, in hippocampus, 24 h(A-D) and 48 h (E-H) after right carotid artery ligation and 3 h of exposure to 8% O2 in P7 rats. In the left hippocampus at 24 h posthypoxia (A) resting microglia are detected; both activated and resting microglia are concentrated in the overlying corpus callosum; note in the area of enlargement (C) that no activated microglia infiltrate the pyramidal cell layer (identified by asterisk). In contrast, in the right hippocampus (B) both intensity of staining and concentration of microglia are increased, and in the area of enlargement(D), activated microglia infiltrate the lesioned CA3 pyramidal cell layer (identified by asterisk). At 48 h, in this severely injured brain, increased numbers of microglial cells are detected bilaterally (E and F). Yet, in the left hippocampus(G) the cellular infiltrate includes predominantly resting microglia, that do not invade the pyramidal cell layer (asterisk); in contrast, in the right hippocampus (H), activated microglia predominate and they clearly infiltrate the pyramidal cell layer(asterisk). Sections were lightly counterstained with cresyl violet to facilitate identification of anatomic landmarks. Magnification: A, B, E, F, 4×; C, D, G, H 10×, scale bars = 100μm.

Full size image

The microglial response was similar in the cortex, increasing progressively in the first 48 h posthypoxic-ischemic insult. Figure 4 compares the distribution of microglia in left and right cortex, based on lectin histochemistry, in complementary sections from the same severely lesioned animal, that was presented in Figure 3, E-H. In the left cortex (Fig. 4A), the density of lectin-stained cells is somewhat increased (in comparison with normal P8 cortex,Fig. 1C); yet, higher magnification reveals only resting microglia (Fig. 4C). In contrast, only activated microglia are detected in the right cortex (Fig. 4D).

Figure 4
figure 4

These photomicrographs compare the distributions of lectin-stained cells in left (A and C) and right(B and D) cortex 48 h after right carotid artery ligation and 3 h exposure to 8% O2 in P7 rats; tissue sections were prepared from the same brain in which hippocampal histopathology was presented inFigure 3, E-H. In the left cortex (A), the density of lectin-stained cells is somewhat increased (in comparison with unlesioned P8 brain, see Fig. 1C); yet, higher magnification (C) reveals that the lectin-stained microglia invariably manifest the morphology of the resting cell type(arrowheads). In the lesioned right cortex (B), the lectin-stained cells are larger and more darkly stained; at higher magnification (D), only the activated microglial cell type is seen(arrowheads). In both regions, endothelial cells (stars) are also detectable. Magnification: A and B, 4×, scale bar = 100 μm; C and D, 40×, scale bar = 10μm.

Full size image

At all time intervals evaluated, ramified microglia were distributed throughout the left cerebral hemisphere; only rarely was this cell type identified in the lesioned right hemisphere at 24 h or subsequent time intervals after hypoxic-ischemic injury. At 24-72 h posthypoxia-ischemia, activated microglia and round cells were widely distributed in all lesioned fore-brain structures; in addition to the microglial reaction in cortex and hippocampus, findings were similar in thalamus, striatum, and habenula (not shown). Although lectin-staining of endothelium was not systematically evaluated, a trend toward increased endothelial staining was also noted in the lesioned right hemisphere at these times.

ED-1 immunocytochemistry. ED-1 immunocytochemistry did not reveal any corresponding focal increase in immunoreactivity at 10 min posthypoxia-ischemia. ED-1-immunoreactive cells were first detectable within the parenchyma of the right hemisphere fourth posthypoxia-ischemia, in 3/5 animals evaluated; of note, lectin histochemistry revealed increased accumulation of activated microglia in corresponding sections from all five samples evaluated at this early time. Increased numbers of ED-1 immunoreactive cells were consistently detected in the adjacent ependyma and choroid plexus at this time interval and subsequently. Figure 5 demonstrates the absence of ED-1 immunoreactive microglia in normal P7 hippocampus (Fig. 5A), and progressive increases in the accumulation of ED-1-immunoreactive microglia in the right hippocampus and adjacent structures 4-48 h posthypoxia-ischemia. At 24 h, ED-1-reactive cells infiltrate the lesioned pyramidal cell layer (Fig. 5C), and at 48 h, the widespread distribution of ED-1 immunoreactive cells in lesioned right hippocampus and adjacent regions closely paralleled the distribution of activated microglia, identified by lectin histochemistry(compare Fig. 3F). Figure 6 illustrates the periventricular distribution of ED-1-immunoreactive cells in P7 brain (Fig. 6A), and the prominent increases in their accumulation in periventricular white matter, choroid plexus, ependyma, as well as in adjacent striatum acutely, 12 h after hypoxic-ischemic lesioning.

Figure 5
figure 5

This montage demonstrates features of the acute microglial response in lesioned hippocampus that are identifiable by immunocytochemistry using the ED-1 antibody. Photomicrographs of ED-1 immunocytochemistry in the right hippocampus of four animals are compared[A, normal control P8; B-D, animals that underwent right carotid artery ligation, followed by 3 h of 8% O2 exposure and were killed 4 (B), 24 (C), or 48 h (D) later]. In the normal P8 hippocampus (A), there are ED-1 immunoreactive cells in the corpus callosum (compare Fig. 1). At 4 h posthypoxia-ischemia (B), the ependyma and choroid plexus, adjacent to the lesioned hippocampus, demonstrate increased accumulation of ED-1-immunoreactive cells. No reactive cells are detected within the pyramidal cell layer. At 24 h postinjury, ED-1-immunoreactive cells infiltrate the pyramidal cell layer (in C, most evident in CA1). At 48 h, there is a marked diffuse infiltration of ED-1-immunoreactive cells throughout the hippocampus. All sections were lightly counterstained with cresyl violet to facilitate identification of anatomic landmarks. CC, corpus callosum. Magnification: 4×, scale bar = 100 μm.

Full size image
Figure 6
figure 6

These photomicrographs demonstrate that the acute microglial response elicited by hypoxic-ischemic injury, identified immunocytochemically with the ED-1 antibody, is also readily evident in periventricular ependyma at the level of the striatum (compareFig. 5B). This montage compares the distributions of ED-1-immunoreactive cells in white matter, choroid plexus, ependyma, and parenchyma bilaterally in a P7 animal that underwent right carotid artery ligation, followed by 3 h of 8% O2 exposure and was killed 12 h later. In the left hemisphere (A), the distribution of ED-1-immunoreactive cells is similar to that in normal P7 brain; ED-1-immunoreactive cells are detected in white matter and choroid plexus. In the hemisphere ipsilateral to ligation (B), the density of immunoreactive cells is increased in white matter, choroid plexus, and ependyma, and in addition there is infiltration of adjacent striatum. Magnification: 4×, scale bar = 100μm.

Full size image

Based on results of both lectin histochemistry and ED-1 immunocytochemistry, the density of activated microglia in the lesioned right hemisphere peaked between 2 and 4 d after lesioning. At P12, the latest time interval postinjury that was evaluated, the microglial reaction in the lesioned right hemisphere was consistently less intense than at the preceding times evaluated (not shown); in 8/8 animals, an attenuated microglial response was still detected in cortex, striatum, and hippocampus of the right hemisphere.

DISCUSSION

These data provide histochemical and immunophenotypic analysis of the microglial reaction elicited by hypoxic-ischemic brain injury in immature rats. Lectin histochemistry demonstrated morphologic evidence of microglial activation in the first 4 h after hypoxia-ischemia in the right hippocampus. The increase in ED-1 immunoreactivity noted at 4 h posthypoxia-ischemia, and likely reflecting de novo synthesis of this macrophage specific antigen(7), provided important confirmatory evidence of the acute onset of microglial activation.

Within the first 24 h after injury, increased numbers of activated microglia accumulated in all lesioned forebrain structures. In general, the regional distribution of the microglial response corresponded with the distribution of histopathology in this perinatal stroke model. In the hippocampus, in contrast with the selective vulnerability of CA1 in adult cerebral ischemia models, CA1 and CA3 subfields are both vulnerable to irreversible injury(18); microglia infiltrated the pyramidal cell layer to a similar extent in both regions. The temporal evolution of the microglial response differed considerably from findings in adult brain; infiltration of activated microglia in lesioned forebrain peaked at about 48 h posthypoxia, plateaued at this level from 48-96 h posthypoxia, and then began to wane. In contrast, previous reports in adult cerebral ischemia models reported initial evidence of microglial proliferation 2-3 d after injury, which was then sustained for up to 4 wk(5, 8, 28). Increased numbers of morphologically distinct round cells were also noted in lesioned tissue. Regional accumulation of round cells has been reported previously in adult animals after injury(6, 29, 30); they may be peripherally derived monocytes. Whether these cells have functional properties distinct from those of activated microglia is unknown.

A distinctive feature of the acute microglial response was the early evidence of an ipsilateral periventricular microglial reaction. Of interest, in this perinatal stroke model, there is also early evidence of disruption of cytochrome oxidase activity, an indicator of neuronal metabolic integrity, in periventricular zones(31). Although a periventricular distribution might suggest exposure to a soluble factor disseminated in cerebrospinal fluid, both the acute microglial response and loss of cytochrome oxidase activity are limited to the ischemic hemisphere, and the factors that account for this anatomic distribution are uncertain.

Whether microglia play an important role in the progression of ischemic injury through the release of cytokines or other potential injury mediators remains uncertain(32, 33). There is currently noin vivo evidence that suppression of microglial activity is neuroprotective in brain; yet, there is an interesting report that treatment with chloroquine, which presumably acts by inhibition of the cytotoxic and/or phagocytic, microglial activity. improves neurologic function in rabbits subjected to spinal cord ischemia(34). Activated microglia likely also contribute to the removal of cell death-associated debris induced by ischemic injury and, as in the normal developing brain, may be active in tissue remodeling and/or repair(35). Thus, the function of activated microglia after ischemia may be bidirectional, as mediators of further injury, as well as serving as cleansing scavengers and agents of repair.

Considerable controversy(27, 36–40) has arisen about the cellular origin of microglia and brain macrophages. Brain macrophages and microglia may constitute two distinct phagocytic cell populations in the developing brain, with the microglia remaining as the resident cell group in the adult brain(15), or alternatively, blood-derived macrophages may differentiate into the resting microglial cell in brain(39, 41, 42). This study did not attempt to address this complex issue. Our findings do suggest a morphologic transition from ramified microglia to activated phagocytic cells in lesioned tissue. When the microglial reaction peaked 2-4 d after injury, the number of resting microglia in the lesioned hemisphere decreased markedly, and these cells were virtually nondetectable in severely injured regions.

Activated microglia may release soluble factor(s) that initiate astrocyte hypertrophy and/or proliferation(43). Yet, in this perinatal brain injury model, increased accumulation of glial fibrillary acidic protein (mRNA and protein) in the lesioned forebrain persists for at least 2 wk after injury(44), whereas the microglial reaction waned abruptly after P11. This disparate temporal pattern suggests that, although microglia may well contribute to the initial stimulation of astrogliosis, they are unlikely to provide the molecular signals that contribute to sustained gliosis.

Our data suggest that, in the developing brain, injury-induced microglial activation and accumulation proceeds more rapidly than in the injured adult brain. This may reflect distinct developmental stage-specific properties of microglia and the distinctive roles of activated microglia in normal CNS development, for example, in the remodeling of tissue cytoarchitecture that occurs in early postnatal life(45, 46). It is also possible that the adverse impact of inappropriate microglial activation is greater in the immature nervous system. Because macrophages may secrete a variety of factors that stimulate axonal growth, glial proliferation(47, 48), and angiogenesis(12, 49), these cells may also influence not only physiologic maturation of both gliogenesis and angiogenesis, but may also contribute to their pathologic disruption after perinatal brain injury. Similarly, activated macrophages are a potent source of plasminogen activator(40, 50); because the plasminogen activator-plasmin system is involved in neuronal migration(51), it is also conceivable that pathologic microglial activation can also lead to indirect injury-induced disruption of neuronal migration.

A particularly important issue in understanding the actions of microglia after injury is the relationship of morphologic activation with biochemical activation; whether morphologic activation inevitably precedes or accompanies functional activation is difficult to assess directly in vivo. Similarly, it is unknown if the same molecular signals initiate and regulate morphologic and functional changes. The chemotactic signals that enable phagocytic cells to target degenerating elements in the CNS are currently unknown(42, 45); our preliminary studies indicate that injury-induced expression of the potent chemokine monocyte chemoattractant protein-1, which is detectable at 4 h posthypoxia in the lesioned hemisphere, may be one of the molecular signals that initiates microglial accumulation(52). In addition, the functional roles of the ramified microglia(23), which exhibit a high degree of pinocytotic activity and motility(53), after injury are uncertain. Of greatest interest is the question of whether attenuation of microglial activation would modulate outcome after acute brain injury. In the neonatal brain, where microglial activation may play important physiologic roles, development of therapeutic interventions targeted to limit pathologic microglial activation may be particularly complex.