Radioautographic investigation of gliogenesis in the corpus callosum of young rats. II. Origin of microglial cells

Inosine diphosphatase as a histochemical marker of retinal microvasculature, with special reference to transformation of microglia

Nucleoside diphosphatase (IDPase), localized using inosine diphosphate as substrate, allows the selective staining of blood vessels and cells of vascular origin, such as macrophages and microglia, whereas the neuroglial, the neuronal and the pigment epithelial cells remain unstained. The staining pattern observed in the retina of mouse, rat, cat and monkey are similar; some apparent quantitative differences reflect species differences in the distribution of retinal microvasculature. At the electron-microscopic level, most of the enzyme activity in the blood vessels appears to be located along the outer wall. The cell membrane, parts of the smooth endoplasmic reticulum and the nuclear membrane in the microglial perikarya appear positive; profiles of microglial processes are intensely stained. In the developing eyes of rats and mice, the blood vessels are stainable from the earliest stage of their appearance. An array of amoeboid cells precede the growing blood vessels and spread out over the future vascularized part of the retina. These cells eventually develop characteristic microglial features, and extend many elongated and branched processes between the neuroepithelial cells while remaining in contact with, or in close proximity to, the blood vessels. Intense IDPase activity in the microglial cells, in contrast to the absence of the enzyme in the neuroglial Müller cells, suggests that microglia are involved in phosphate metabolism and indicates functional compartmentalization within the glial tissue lying between the blood retinal barrier and the retinal neurons.

Transient populations of presumptive macrophages in the brain of the developing hamster, as indicated by endocytosis of blood-borne horseradish peroxidase

During postnatal development, clusters of cells associated with the mononuclear phagocytic system appear within the white matter of rodents and cats. We studied the distribution and morphology of these cells in the hamster's brain during the first 2 weeks after birth. In animals of different ages, horseradish peroxidase was injected into the heart. After 3-6 h survival, the animals were perfused with aldehydes and had their brains removed, cut and reacted. In another series, fixed brain sections from horseradish peroxidase-injected and non-injected animals were reacted for a non-specific esterase expressed by monocytes and macrophages. The horseradish peroxidase reaction-product was seen throughout the nervous tissue at the first postnatal day, appearing more concentrated in certain brain sectors from postnatal day 3 through 10, to finally become restricted to a few regions at postnatal day 16. Horseradish peroxidase-labeled cells appeared in increasing numbers from postnatal day 1 to 8, decreasing thereafter to disappear completely at postnatal day 16. Some labeled cells were roundish or elliptical with few, if any, processes; others had several clearly detectable processes. Horseradish peroxidase-labelled cells formed clusters within the dorsal subventricular zone, dorsal cortical white matter, corpus callosum and several other prosencephalic fiber tracts. The morphology of esterase-reactive cells was less clearly outlined but their distribution and relative density correlated with those of horseradish peroxidase-labeled cells. Also, many horseradish peroxidase-labeled cells were esterase-positive in most clusters. We conclude that (1) some cells in the developing brain selectively endocytose and accumulate blood-borne horseradish peroxidase in their cytoplasm, (2) these cells do not appear to be neurons but a particular cell type associated to the mononuclear phagocytic system and (3) they cluster transiently in particular sectors of the cortical and subcortical white matter during the first 2 weeks after birth.

Ultrastructural characterization of glial cells in the rat pineal gland with special reference to the pineal stalk

In the present study the "interstitial" cells of the superficial pineal gland and the nonparenchymal cells of the pineal stalk in Sprague-Dawley rats were examined ultrastructurally with the aim of defining the cells more closely. The "interstitial" cells of the superficial pineal gland do not represent a homogeneous cell population. The most abundant cell type is the mononuclear phagocyte, most easily recognized by its dark appearance and its content of primary and conspicuous secondary lysosomes. Astrocytes can be distinguished by the typical appearance of their nuclei (i.e., a thin continuous rim of heterochromatin adjacent to the nuclear membrane), identical to that of astrocytes in the CNS. Depending on the absence or presence of glial filaments and their amount, a spectrum of astrocytic cells is present. Mature astrocytes with filaments throughout their cytoplasm are rare. Immature glial cells with few or no filaments predominate. In the vicinity of blood vessels pericytes are present. In view of the fact that the "interstitial" cells could generally be identified it is suggested to abandon the term interstitial for the cells in question. In the pineal stalk mature astrocytes predominate; they have some features in common with pinealocytes, i.e., the presence of intergrade endoplasmic reticulum and grumose bodies (lysosomes). Other unusual features are a relative abundance of coated pits and vesicles. Oligodendrocytes are restricted to the proximal part of the stalk, near the deep pineal, where myelinated axons are abundant. More distally a few Schwann cells were seen.

Genesis of resting microglia in the gray matter of mouse hippocampus

The genesis of resting microglia in the gray matter of mouse hippocampus was studied by 3H-thymidine autoradiography in combination with electron microscopy. Newborn mice were injected with 3H-thymidine singly or repeatedly at different postnatal stages, and killed shortly after the injection or after various intervals. Tissue specimens of the hippocampus at CA1 and CA2 were processed for light and electron microscopic autoradiography. The results showed that at least 91% of glial cells in the stratum radiatum of the hippocampus are produced after birth. About three-fourths of astroglia in this area are produced before the sixth postnatal day, and a larger part of resting microglia are formed after the ninth postnatal day. Morphological transition can be traced from either proliferating cells in the stratum radiatum at late postnatal days to resting microglia, or from those in early postnatal days to astroglia. A continuous morphological transition was observed between the proliferating cells at the late postnatal days (microglial production period) and those at the early postnatal days (astroglial production period). The latter retain some fine structural characteristics similar to small glioblasts in the subependymal layer. These findings strongly suggest that resting microglia, as well as astroglia, are derived from glioblasts, and are of neurectodermal origin.

Immunohistochemical studies of blood monocytes infiltrating into the neonatal rat brain

Brains of normal rats ranging in age from newborn to adult were observed with immunofluorescence technique using anti- granulomonocyte antiserum. For the first 10 days after birth, many cells with positive fluorescence were found in the white matter, the subependyma , the extra-parenchymal spaces, and the leptomeninx , but very few in the gray matter. They were mononuclear, rich in cytoplasm, and globular or irregular in shape. After about day 10 p.n., the positive cells decreased in number and became slender. However, there was no change in the distribution pattern. After about 3 weeks of age, no positive cells were detected in the brain parenchyma, except for very rare necrobiotic ones. It was suggested that blood monocytes infiltrate into the brain parenchyma of normal neonatal rat, but only for a while in the limited areas (white matter and subependyma ). They have the morphology and distribution of the "ameboid microglia" of neonatal brain. These monocytes disappear from the brain finally by the end of month 1 p.n.

Transitory macrophages in the white matter of the developing visual cortex. II. Development and relations with axonal pathways

Clusters of 'gitter cells' develop in the white matter of the occipital cortex of the cat at the end of the first postnatal week. These clusters, and others already present at birth, disappear by the end of the first postnatal month. The life span of the clusters in the occipital white matter corresponds to the period when transitory callosal axons are eliminated. The clusters have close contact with callosal axons and can be labeled by HRP injected in the contralateral hemisphere and transported through the corpus callosum. One of the clusters clearly forms in a part of the white matter crossed by transitory callosal axons. The 'gitter cells' might be involved in the elimination of these axons. Consistent with this hypothesis, ultrastructural observations show groups of axons completely surrounded by 'gitter cell' cytoplasm as if they were being phagocytosed.

Transitory macrophages in the white matter of the developing visual cortex. I. Light and electron microscopic characteristics and distribution

Injections of horseradish peroxidase (HRP) into the occipital cortex of the kitten and diffusing to the white matter label a widely distributed microglial population and in addition, cells with light and electron microscopic features of 'gitter cells'. The latter are concentrated in a complex and highly consistent system of interconnected clusters in the white matter of the lateral, postlateral, middle suprasylvian and posterior ectosylvian gyri, as well as on the roof of the lateral ventricle. The 'gitter cells' have the ultrastructural and (as described by others) chemical characteristics of macrophages, and may be involved in the elimination of transitory axons.

Immunohistochemical localization of a macrophage-specific antigen in developing mouse retina: phagocytosis of dying neurons and differentiation of microglial cells to form a regular array in the plexiform layers

In the developing mouse retina degenerating neurons can be observed initially in the ganglion cell layer followed by a phase of cell death in the inner nuclear layer. Using an immunohistochemical method to localize the mouse macrophage specific antigen F4/80, we show that macrophages migrate from the vascular supply overlying the developing retina and phagocytose the degenerating neurons. The macrophages subsequently differentiate to become the microglia of the retina and form a regularly spaced distribution across the retina in the inner and outer plexiform layers. These experiments provide strong evidence for the mesodermal origin of central nervous system microglia.

A fine structural study of the cells that proliferate in the partially denervated dentate gyrus of the rat

Tritiated thymidine autoradiography has established that after interrupting the commissural afferents to the dentate gyrus a number of non-neuronal cells proliferate in the molecular layer. In the present study the fine structure of the proliferating cells was analyzed by reembedding the 2-microns thick plastic sections of the dentate gyrus which had been previously coated with a nuclear emulsion and processed for light microscopic autoradiography. The location of the labeled cells was plotted with a camera lucida and a few ultrathin sections were taken from the re-embedded sections. In these the labeled cells were re-identified and photographed in an electron microscope. Most of the identified proliferating cells exhibited the following morphological features: The nuclei were irregularly oval, sometimes with deep indentations and contained dense clumps of chromatin; their diameters ranged between 4.5 and 6.5 microns. The cytoplasm was generally disposed to one side of the nucleus and often extended into a few broad processes. The Golgi apparatus was well developed. Many rosettes of free ribosomes were scattered throughout the cytoplasm, and the rough endoplasmic reticulum usually consisted of a few short cisternae. Small multilamellated bodies were common, but dense inclusion bodies were infrequent. The observations reported in this paper suggest: 1. that the nonneuronal cells which proliferate in a neuropil undergoing a mild denervation are morphologically closely related to microglia; 2. that in young adult animals these cells do not seem to have been previously involved in intense phagocytic activity; and 3. that the proliferating cells are present in the neuropil at the time of the denervation.

Life history of decidual cells: a review

Decidual cells are a distinctive cell population observed in the mammalian endometrium during pregnancy. Their appearance can also be induced with appropriate stimuli in the hormone-primed pseudopregnant uterus. This review deals with their life history, including the dynamic morphological events during the process of decidualization, cytochemical markers for decidual cell reaction, the surface markers and functions of decidual cells, and, finally, the origin and fate of this cell class. The recent discovery of the bone marrow origin of decidual cell precursors adds a new dimension to decidual cell biology. Future studies of steps in the differentiation of the decidual cell lineage await the identification of unique lineage-specific cell--surface of cytoplasmic marker(s); a precise knowledge of decidual cell functions can only be obtained from a discriminating analysis using purified cell populations.

Morphological studies on neuroglia. VI. Postnatal development of microglial cells

The postnatal development of microglial cells was investigated in the neonatal rat brain by use of light- and electron microscopy, including enzyme-histochemical techniques. Microglial cells were selectively stained by demonstration of their nucleoside diphosphatase (NDPase) activity and classified into three types: 1) In the early postnatal period "primitive microglial cells" showing scantily ramified processes were found in the cerebral cortex, the hippocampal formation, and the hypothalamus. During the course of the first postnatal week the processes of this cell type developed gradually and the cells were transformed into typical ramified microglial cells, called "resting microglial cells". 2) "Amoeboid microglial cells "showing typical features of macrophages were characteristic of the cerebral white matter. 3) "Round microglial cells" possessing a round soma and few pseudopodia but no characteristic processes occurred in large numbers in the subventricular zone of the lateral ventricle and as single elements in the vicinity of blood vessels. Histochemically, thiamine pyrophosphatase (TPPase) was demonstrated only in the fully developed, ramified microglial cells ("resting microglial cells"), which could be readily observed in the central nervous tissue from the age of 14 day. "Round and amoeboid microglial cells" did not show TPPase activity and disappeared after 14 days of postnatal life. By use of electron microscopy, in neonatal rats NDPase activity was apparent in the plasma membrane of the three types of microglial cells ("primitive, round, and amoeboid" types). They showed basically similar submicroscopic characteristics, i.e., well-developed Golgi apparatus, long strands of rough-surfaced endoplasmic reticulum, single dense bodies and vacuoles, and numerous ribosomes. "Amoeboid microglial cells" were characterized by their well-developed cytoplasmic vacuoles and phagocytic inclusion bodies. The present study strongly suggests a mesodermal origin for these microglial elements.

The early formation of the corpus callosum: a light and electron microscopic study in foetal and neonatal rats

A light and electron microscopic study of the developing corpus callosum was carried out in foetal and neonatal rats in order to determine the mode of growth of the earliest callosal axons across the midline and to investigate the potential role played by non-neuronal cells during the formation of the tract. The axons of the corpus callosum first cross the midline between the 18th and 19th days of gestation by traversing the anterodorsal aspect of the pre-existing hippocampal commissure. Prior to the appearance of the callosal axons at the midline, there is an aggregation of astrocyte processes anterior and dorsal to the hippocampal commissure. Careful examination of these processes in different planes of section shows that they are not organized in any obvious way that would provide a clearly defined path for the growing axons; nor are there any preferentially oriented extracellular spaces at the midline. No specialized membrane contacts could be seen between non-neuronal cell processes and the early callosal axons. Thus, there is no overt morphological evidence for an active role of non-neuronal cells in axon guidance in the initial formation of the corpus callosum. The development of the corpus callosum is accompanied by the formation of a temporary cavum septi pellucidi, which is always closed to the subarachnoid space. The cavum persists during the first postnatal week, after which time it becomes populated by cells of the lateral septal nuclei. Macrophages are present within the cavum and may play a role in its formation. Macrophages are also found within parts of the corpus callosum. No obvious degeneration of axons is seen in the corpus callosum during its early development.

Perisomatic changes in the maturing hypoglossal nucleus after axon injury

The perisomatic retrograde reactions to peripheral nerve injury during postnatal maturation were investigated by quantitative light and electron microscopy. The hypoglossal nerve was crushed, ligated or transected in 10 and 21 day postnatal (dpn) rats. Only crush injury was made in 7 dpn rats. Survival periods ranged from 3 to 40 days postoperative (dpo). Normal and sham operated animals of corresponding ages served as controls. A remarkable, transient response of neuronal somata in the young (7-10 dpn) rats, following all three types of axon injury, was the formation of excrescences which engulfed neuronal processes: boutons, dendrites, small neurites and a few myelinated axons in adjacent neuropil. The somal engulfment was rarely evident after nerve injury to 21 dpn rats. Bouton displacement from the somal surface, accompanied by glial incursion, followed each type of nerve injury but was less extensive and occurred later in the young rats. There seemed to be no association in the amount of boutons displaced from neuronal somata with the type of nerve injury for any of the three experimental age groups. However, the rate and intensity of the perineuronal glial reaction were related to the severity of the nerve injury in the older (21 dpn) but not in the younger (7-10 dpn) rats. Substantial loss of neurons occurred in the affected nucleus after each type of trauma to the young neurons. The degree of neuronal loss was related to the age and the volume of axoplasm disrupted. Only nerve transection in 21 dpn rats resulted in appreciable neuronal loss. The responses of axotomized neurons, their afferent axon terminals and the surrounding glia are quantitatively and qualitatively different in the hypoglossal nucleus of rats of increasing postnatal ages.

Hematogenous origin of the inflammatory response in acute poliomyelitis

To determine the origin of the inflammatory response, and in particular the microglial rod cell response, in acute viral encephalitis, 4-week-old Swiss mice were injected with tritiated thymidine to label actively dividing cells prior to infection with the Lansing type 2 strain of poliovirus. As expected, the majority of polymorphonuclear and mononuclear leukocytes within the central nervous system perivascular infiltrates were shown to be hematogenous in origin. As early as 24 hours after infection, isotope-labeled cells having light histological and ultrastructural features consistent with microglia and microglial rod cells were identified within brain parenchyma and were shown to participate in neuronophagia and formation of glial nodules. Supraependymal and suprachoroidal cells were also shown to contain the label. However, neither endothelial cells nor pericytes contained label as determined by electron microscopy. These studies support a hematogenous origin for all cellular elements of the classic inflammatory response in viral infections of brain.

Affinity of chick microglia in vitro for ricin 120

The distribution of binding sites for ricin 120 in cell cultures of spinal cord from the chick embryo was examined. A characteristic pattern was observed, which remained similar in cultures fixed by a variety of methods. Light microscopy demonstrated that the most prominent staining was of small rounded or amoeboid cells. Electron microscopy showed that ricin was bound over their entire surfaces. The ultrastructural characteristics of these cells suggest their identification as microglia.

Morphological studies on neuroglia. II. Response of glial cells to kainic acid-induced lesions

The cellular response of non-neuronal elements of the pyramidal cell layer of the rat hippocampus, especially the area CA3, was observed electron microscopically following destruction of this formation by means of intraventricular administration of kainic acid (KA). The neuroglial cell types responding to the KA-induced lesion included astrocytes and the "microglia-like reactive cells". In addition, numerous brain macrophages appeared in the damaged area CA3. Oligodendrocytes and pericytes revealed no morphological changes. Swollen astrocytes were seen in the KA-induced lesion during the early stage. Glial filaments gradually developed in the soma and cell processes of these cells. Brain macrophages were seen in the KA-induced lesion during the early stage; they gradually decreased in number with time. Numerous small cells displaying a dark nucleus appeared in the damaged area CA3 during the first two days after the KA-administration, and gradually increased in number. During the later stage this cell type could hardly be distinguished from the intrinsic microglial cells. It is open to discussion whether this cell type originates from the intrinsic microglial cells or from the hematogenic monocytes; therefore it is designated as "microglia-like reactive cell" in the present study.

Hematogenous cells in the central nervous system of developing human embryos and fetuses

Examination of large blocks of Epon-embedded, 1.0-micrometer sections of human embryos and fetuses reveal the presence of hematogenous cells in various stages of differentiation in neural tissue. In every embryo and fetus of 10 weeks ovulation age and younger, hematogenous cells are found randomly scattered in the cerebrum, cerebellum, and spinal cord. Many of these cells appear to undergo spontaneous degeneration in neural tissue and become rarer in older fetuses. Also identified in the neuropil of normal embryos and fetuses are cells with the typical morphological appearance of macrophage containing numerous inclusions of various kinds, both inside and outside the blood vessels. In addition, scattered in the subpial, perivascular, and perineuronal regions of the neural parenchyma are small cells with fusiform nuclei and a small amount of cytoplasm as well as cells with a moderate amount of elongated cytoplasm containing various inclusions and oblong nuclei. All of these cells have clumped heterochromatin along the nuclear membrane which differs from other neuroectodermal cells of the developing human CNS. Although there is no direct evidence to indicate transformation of macrophages to "microglia" or vice versa, the presence of cells having similar nuclear morphology and chromatin pattern while appearing to be transitional forms of macrophage, varying from undifferentiated to fully developed, suggest a common lineage of these latter types. It is concluded that migration of hematogenous cells into neural tissue is a ubiquitous developmental phenomenon in young human embryos and fetuses.

Non-specific esterase activity in reactive cells in injured nervous tissue labeled with 3H-thymidine or 125iododeoxyuridine injected before injury

Tritiated thymidine (3H-TdR) injected before a stab wound of the spinal cord or transection of the hypoglossal nerve has resulted in many labeled reactive cells in the CNS after injury, most of which have the ultrastructural features of microglia. To test for the possible origin of these labeled cells from monocytes, we examined them for the presence of sodium fluoride- (NaF) sensitive non-specific esterase (NSE), an enzyme characteristic of monocytes. Some of the labeled cells in stab wounds had NaF-sensitive NSE, but no such cells were found in the nucleus of the injured hypoglossal nerve. To test for the possibility that the NSE-negative labeled cells had been labeled by reutilization of 3H-TdR, we used 125I-5-iodo-2'-deoxyuridine (125I-UdR), a thymidine analogue with a much lower rate of reutilization, to label blood mononuclear cells prior to either a spinal cord stab wound or hypoglossal axotomy. The number of labeled cells was decreased in the spinal cord wound, but more than half were NSE-negative. No labeled blood mononuclear cells were found in the hypoglossal nucleus, although there was no decrease in the hyperplasia of unlabeled non-neuronal cells. When 125I-UdR was injected on the fourth day after hypoglossal axotomy, or when both 3H-TdR and 125I-UdR were injected simultaneously before hypoglossal axotomy, many labeled cells were found in the hypoglossal nucleus, indicating that 125I-UdR can be used by the reactive cells and that it did not inhibit their proliferation. Therefore, the microglial cells that proliferate in response to peripheral nerve injury are not recently derived from any type of circulating large blood mononuclear cell. The most likely explanation for the presence of the 3H-TdR-labeled cells in the nucleus of the injured hypoglossal nerve is that they were proliferating intrinsic cells labeled by reutilization of 3H-TdR.

Use of carbon labeling to demonstrate the role of blood monocytes as precursors of the 'ameboid cells' present in the corpus callosum of postnatal rats

Cells with features suggestive of ameboid motion and phagocytic properties are observed in the rat corpus callosum during the first few days of life. These cells, hereafter referred to as 'ameboid cells', have been investigated in several ways. An electron microscopic study of the corpus callosum in 5- to 7-day-old rats indicated that most 'ameboid cells' were typical macrophages, but some displayed features of monocytes, while others appeared to be transitional between the two types. These observations raised the possibility that blood monocytes were the precursors of 'ameboid cells'. This possibility was tested by injecting a suspension of carbon particles into the circulation of rats of various ages to label and trace monocytes. Within 15 minutes after injection, carbon particles were seen between cells in blood smears as well as in the lumen of capillaries, but not between cells and axons in corpus callosum. By a half hour, a few of the circulationg monocytes, and with time, up to half of them, contained carbon particles. Five days after injection, carbon particles were observed in cells of the corpus callosum identified as 'ameboid cells' of the monocytic and macrophagic type. Such carbon-containing cells were seen in many of the animals injected at the age of 0-1 day, in few of those injected at 3-5 days, and in none of the older animals. Since free carbon had not been observed in corpus callosum spaces, it was concluded that 'ameboid cells' did not pick up carbon locally. The alternative was that blood monocytes, after ingesting carbon particles in the circulation, migrated to the corpus callosum and settled as 'ameboid cells'. In the hope of obtaining a direct confirmation of this conclusion, blood cells obtained from carbon-injected Lewis rats were centrifuged in a Percoll gradient to obtain a fraction which contained 70-80% monocytes, less than 2% granulocytes, and 20-30% lymphocytes. Carbon was present in up to half of the monocytes and 1% of the granulocytes, but not in the lymphocytes; and it was calculated that over 99% of the carbon-labeled cells were monocytes. The cell fraction was then introduced into the blood circulation of 2- to 3-day-old syngeneic Lewis rats, and the animals were sacrificed 5 days later. Occasional carbon-labeled cells appeared not only in liver, spleen and connective tissue, but also in the corpus callosum, where they were identified as 'ameboid cells' of the monocytic and macrophagic type. Even though such cells were infrequent, their presence conclusively demonstrated that at least some 'ameboid cells' of the corpus callosum were derived from circulating blood monocytes.

Gliogenesis of astrocytes and oligodendrocytes in the neocortical grey and white matter of the adult rat: electron microscopic analysis of light radioautographs

The identification of newly formed glial cells in the normal adult cerebral cortex is unresolved, since the identification of cells incorporating [H3] thymidine has not been demonstrated in the adult by electron microscopy. In the present study, this problem has been studied by combining the resolution of the electron microscope with radioautography of 1-micrometer sections. Four normal male rats were injected at 90 days of age with [H3] thymidine and allowed to survive for 30 days. Labeled cells were found in 1-micrometer sections of the visual cortex of these adult rats, and electron micrographs of selected cells from these same sections demonstrated clearly two types of cells labeled, astrocytes and oligodendrocytes, in both grey and white matter. The few cells that were tentatively identified as labeled microglia in the light microscope proved to resemble oligodendrocytes when examined in the electron microscope. In 1-micrometer sections of the cortical grey matter, heavily labeled astrocytes (13 or more silver grains over the nucleus) represent about 0.08% of the total astrocytic population, and heavily labeled oligodendrocytes also were about 0.08% of their population. In the cortical white matter, about 0.03% heavily labeled astrocytes were observed, compared to about 0.07% heavily labeled oligodendrocytes. For all neuroglial cells in both white and grey matter, the average percent heavily labeled cells was 0.066%, a value large enough to suggest a slow turnover of neuroglial cells during the lifespan of the rats.

Endocytic activity of subependymal microglial cells in the toad brain: a cytochemical study of peroxidase uptake

A population of microglial cells that rapidly incorporate extracellular material introduced into the ventricular system has been identified just beneath the ependyma of all four cerebral ventricles in the toad (Bufo marinus). In untreated tissue these cells appear to be scattered, possess few processes and have an elongate shape with their long axes lying parallel to the ventricular surface. Their most distinctive ultrastructural features are nuclei containing clumps of chromatin, cytoplasmic dense bodies and single strands of granular endoplasmic reticulum. When horseradish peroxidase (HRP) is perfused through the ventricular system and the tissue processed using the DAB cytochemical method, the cells change shape and incorporate HRP into cytoplasmic structures. Even after very short perfusion periods (2-5 minutes) cells become rounded, the surface is ruffled and pseudopodia develop that contain characteristic flocculent material. Reaction product for HRP is contained in plain and coated vesicles, tubules, vacuoles and long structures composed of two closely apposed membranes. At these early times, relatively few multivesicular bodies and dense bodies contain reaction product, but when the cells are viewed at longer time periods after the ventricular perfusion of HRP an increasing proportion of the multivesicular bodies and dense bodies contain reaction product. By 320 minutes reaction product is found almost exclusively in these two organelles. In addition, many pseudopodia containing dense bodies with peroxidase activity are found in the neurophile; some, but not all, can be traced from the subependymal microglial cells. The cell bodies have resumed their flattened shape. When compared to the subependymal microglial cells, other brain cells--oligodendrocytes, astrocytes, ependymal cells and neurons--contain relatively little reaction product at short time intervals; only by 320 minutes are moderate amounts of HRP present. Because of the position of the microglial cells and their ingestive capacity, it is suggested that they function to protect the brain from foreign substances entering from the CSF.

Radioautographic investigation of gliogenesis in the corpus callosum of young rats. I. Sequential changes in oligodendrocytes

The corpus callosum of young rats was examined to clarify the behavior of the three subtypes of oligodendrocytes (the large organelle-rich "light oligodendrocytes," the smaller and more densely stained cells referred to as "medium oligodendrocytes," and the even smaller and denser "dark oligodendrocytes"). It was hoped to find out whether cells of the three subtypes undergo division and how they are related to one another. 3H-thymidine was given intraperitoneally as single or three shortly spaced injections to a first group of 19- to 20-day old rats weighing about 40 g, and to a second group of 25-day old rats weighing about 80 g. The animals were sacrificed at various time intervals from 2 hours to 35 days after 3H-thymidine administration. Pieces of corpus callosum were taken near the superior lateral angle of the lateral ventricles; and semithin sections were radioautographed and stained with toluidine blue. Two hours after 3H-thymidine injection, label is virtually absent from light, medium and dark oligodendrocytes, from microglia, and probably from astrocytes, but is present in about 10% of the immature glial cells, which include the poorly differentiated glioblasts and the partially differentiated oligodendroblasts and astroblasts. Hence, the cells undergoing DNA synthesis and mitosis in the corpus callosum are these three types of immature cells. During the week that follow the administration of 3H-thymidine, label appears in oligodendrocytes and astrocytes, which presumably have arisen from the initially labeled immature cells. The oligodendrocytes acquire label in a sequential manner: the light cells show label first and their labeling index reaches a peak at the seven-day interval; the medium oligodendrocytes become labeled next with a labeling peak toward the 14- and 21-day intervals and, finally, the dark oligodendrocytes with a peak around the 28-day interval. Analysis by the method of Zilversmit et al. ('42-'43) provides precise details on the sequence: immature cells presumed to be oligodendroblasts give rise to light oligodendrocytes which, after four to seven days, transform into medium oligodendrocytes which, after another 11 to 18 days, transform into dark oligodendrocytes. The dark cells may persist indefinitely or turn over at a very slow rate. It is concluded that oligodendrocytes arise from the last division of oligodendroblasts and develop in three main periods: a light stage lasting less than a week, a medium stage lasting about two weeks, and a very long lasting dark stage.