The shape and distribution of astrocytes in the retina of the adult rabbit

Effect of epiretinal electrical stimulation on the glial cells in a rabbit retinal eyecup model

Introduction

We examined how pulse train electrical stimulation of the inner surface of the rabbit retina effected the resident glial cells. We used a rabbit retinal eyecup preparation model, transparent stimulus electrodes, and optical coherence tomography (OCT). The endfeet of Müller glia processes line the inner limiting membrane (ILM).

Methods

To examine how epiretinal electrode stimulation affected the Müller glia, we labeled them post stimulation using antibodies against soluble glutamine synthetase (GS). After 5 min 50 Hz pulse train stimulation 30 μm from the surface, the retina was fixed, immunostained for Müller glia, and examined using confocal microscopic reconstruction. Stimulus pulse charge densities between 133-749 μC/cm2/ph were examined.

Results

High charge density stimulation (442-749 μC/cm2/ph) caused significant losses in the GS immunofluorescence of the Müller glia endfeet under the electrode. This loss of immunofluorescence was correlated with stimuli causing ILM detachment when measured using OCT. Müller cells show potassium conductances at rest that are blocked by barium ions. Using 30 msec 20 μA stimulus current pulses across the eyecup, the change in transretinal resistance was examined by adding barium to the Ringer. Barium caused little change in the transretinal resistance, suggesting under low charge density stimulus pulse conditions, the Müller cell radial conductance pathway for these stimulus currents was small. To examine how epiretinal electrode stimulation affected the microglia, we used lectin staining 0-4 h post stimulation. After stimulation at high charge densities 749 μC/cm2/ph, the microglia under the electrode appeared rounded, while the local microglia outside the electrode responded to the stimulated retina by process orientation inwards in a ring by 30 min post stimulation.

Discussion

Our study of glial cells in a rabbit eyecup model using transparent electrode imaging suggests that epiretinal electrical stimulation at high pulse charge densities, can injure the Müller and microglia cells lining the inner retinal surface in addition to ganglion cells.

Retinal toxicity of intravitreal rituximab in albino rabbits

PURPOSE

To evaluate retinal toxicity of intravitreal rituximab.

METHODS

Twelve albino rabbits were included in the study. 1 mg/0.1 mL of rituximab was injected to the right (experimental) eye of each rabbit, and 0.1 mL of saline was injected into the left (control) eye. Electroretinogram was recorded before injection and 3 hours, 4 days, 1 week, 2 weeks and 4 weeks after injection (12 rabbits), and visual evoked potential was recorded before injection and 4 weeks after injection (10 rabbits). Histology and glial fibrillary acidic protein immunocytochemistry (12 rabbits) were performed at 4 weeks after injection. Clinical examination was conducted at all time points in all rabbits.

RESULTS

The average dark-adapted electroretinogram b-wave Vmax ratios, and the average light-adapted b-wave amplitude ratios were approximately 1, and the average log σ difference was around zero throughout the follow-up period. The average visual evoked potential amplitude ratio and the average visual evoked potential implicit time difference were approximately 1 and 0, respectively. No histologic damage was seen, but glial fibrillary acidic protein was mildly expressed in 6 of 12 experimental eyes. Results of clinical examination were normal in all the eyes.

CONCLUSION

Intravitreal injection of 1 mg rituximab does not cause functional or histologic signs of retinal toxicity in albino rabbits. Mild glial fibrillary acidic protein expression in Müller cells probably indicates a mild degree of retinal stress.

Spatial patterning of cholinergic amacrine cells in the mouse retina

The two populations of cholinergic amacrine cells in the inner nuclear layer (INL) and the ganglion cell layer (GCL) differ in their spatial organization in the mouse retina, but the basis for this difference is not understood. The present investigation examined this issue in six strains of mice that differ in their number of cholinergic cells, addressing how the regularity, packing, and spacing of these cells varies as a function of strain, layer, and density. The number of cholinergic cells was lower in the GCL than in the INL in all six strains. The nearest neighbor and Voronoi domain regularity indexes as well as the packing factor were each consistently lower for the GCL. While these regularity indexes and the packing factor were largely stable across variation in density, the effective radius was inversely related to density for both the GCL and INL, being smaller and more variable in the GCL. Consequently, despite the lower densities in the GCL, neighboring cells were more likely to be positioned closer to one another than in the higher-density INL, thereby reducing regularity and packing. This difference in the spatial organization of cholinergic cells may be due to the cells in the GCL having been passively displaced by fascicles of optic axons and an expanding retinal vasculature during development. In support of this interpretation, we show such displacement of cholinergic somata relative to their dendritic stalks and a decline in packing efficiency and regularity during postnatal development that is more severe for the GCL.

A direct contact between astrocyte and vitreous body is possible in the rabbit eye due to discontinuities in the basement membrane of the retinal inner limiting membrane

Different from most mammalian species, the optic nerve of the rabbit eye is initially formed inside the retina where myelination of the axons of the ganglion cells starts and vascularization occurs. Astrocytes are confined to these regions. The aforementioned nerve fibers known as medullated nerve fibers form two bundles that may be identified with the naked eye. The blood vessels run on the inner surface of these nerve fiber bundles (epivascularization) and, accordingly, the accompanying astrocytes lie mostly facing the vitreous body from which they are separated only by the inner limiting membrane of the retina. The arrangement of the astrocytes around blood vessels leads to the formation of structures known as glial tufts. Fragments (N = 3) or whole pieces (N = 3) of the medullated nerve fiber region of three-month-old male rabbits (Orictolagus cuniculus) were fixed in glutaraldehyde followed by osmium tetroxide, and their thin sections were examined with a transmission electron microscope. Randomly located discontinuities (up to a few micrometers long) of the basement membrane of the inner limiting membrane of the retina were observed in the glial tufts. As a consequence, a direct contact between the astrocyte plasma membrane and vitreous elements was demonstrated, making possible functional interactions such as macromolecular exchanges between this glial cell type and the components of the vitreous body.

Differentiation and migration of astrocyte precursor cells and astrocytes in human fetal retina: relevance to optic nerve coloboma

The presence of astrocyte precursor cells (APCs) and time course and topography of astrocyte differentiation during development were investigated by triple-label immunohistochemistry with intact fetal and adult human retinas. Throughout retinal development and adulthood, expression of Pax2 was restricted to cells of the astrocytic lineage. Three distinct stages of astrocytic differentiation were identified during development: i) Pax2+/vimentin+/GFAP- APCs; ii) Pax2+/vimentin+/GFAP+ immature perinatal astrocytes; and iii) Pax2+/vimentin-/GFAP+ mature perinatal astrocytes. In adult, cells with the antigenic phenotype of mature perinatal astrocytes were restricted to a region surrounding the optic nerve head (ONH), whereas cells at a fourth stage of differentiation, adult astrocytes (Pax2-/vimentin-/GFAP+), were apparent throughout the vascularized retina. APC appearance was centered around the ONH and preceded the appearance of perinatal astrocytes. A cluster of Pax2+ somas was also present in a small region surrounding the ONH at the ventricular surface of the developing retina, which suggests the existence of two distinct sites of astrocytic differentiation. The coincidence in the location of APCs and perinatal astrocytes at the ventricular zone with that of optic nerve colobomas, together with the association of Pax2 gene mutations with this condition, suggests that coloboma formation may result from impaired astrocyte differentiation during development.

Immunohistochemistry in association with scanning electron microscopy for the morphological characterization and location of astrocytes of the rabbit retina

The purpose of the present investigation was to establish a method for the morphological characterization and location of the several types of astrocytes in the rabbit retina. Whole retinas were incubated with unlabeled antibody to glial fibrillary acidic protein (GFAP) and, afterwards, treated with secondary antibody labeled according to the requirements for the visualization of the antigen-antibody reaction either with the confocal or transmission electron microscope. Specimens treated similarly to the latter were osmium enhanced and analyzed with scanning electron microscopy (SEM). The different immunohistochemical approaches led to the conclusion that the cells selectively visualized with the SEM are astrocytes. The higher resolution and depth of focus of this instrument allowed a better morphological characterization and a more precise location of the astrocytes in the several levels of the inner portion of the rabbit retina. The method described herein, in which pre-embedding immunohistochemistry for GFAP on rabbit retinas was associated with osmium enhancement and examination with SEM, proved to be reliable and efficient for the morphological characterization and location of astrocytes.

Confocal microscopic study of glial-vascular relationships in the retinas of pigmented rats

Astroglia are interposed between the cerebral vasculature and neurons, where they may mediate the transfer of substances from the circulation to neurons and couple changes in neuronal activity to changes in cerebral blood flow. The retina is a particularly advantageous model system for studying glial-vascular interactions in situ. Confocal microscopy and three-dimensional image reconstruction were used to study the anatomical relationships between glia and the surface vasculature in retinas acutely isolated from adult pigmented rats. Retinas were immunostained using antibodies directed against the basal lamina surrounding the vasculature as well as antibodies directed against glial fibrillary acidic protein. Surface vessels of all calibers were contacted by the processes of astrocytes. The vitreal surfaces of the large retinal vessels were covered by a meshwork of immunoreactive astrocyte processes of a variety of shapes, whereas the scleral surfaces of the vessels were supported by thick bundles of astrocyte processes. In addition, glial cells were filled intracellularly with the gap junction-permeable tracers Lucifer yellow and Neurobiotin. Intracellular fills clearly demonstrated the presence of astrocytes with somata that were closely apposed to the large retinal vessels. Tracer-filled astrocytes displayed a variety and complexity of shapes that was not apparent in immunostained material. Gap junctional coupling was stronger between astrocytes adjacent to the same artery than between periarterial astrocytes and astrocytes located away from arteries. Significantly fewer Müller cells were labeled when Neurobiotin was injected into astrocytes associated with arteries than when Neurobiotin was injected into astrocytes that were distant from arteries.

Developmental changes in astrocyte density in the macaque perifoveal region

We studied astrocyte density both in the perifoveal region and in extrafoveal regions within the same distance of the optic disc (OD) over a time period from foveal pit formation (embryonic day E112) until 2 months after birth. The study was prompted by earlier observations that the adult macaque displays an almost astrocyte-free region around the fovea which, however, at birth is occupied by astrocytes. Thus, we wanted to determine if the perifoveal region is invaded by astrocytes during early development to the same degree as other regions in the central retina, and how the reduction in density can be explained. From the earliest age we studied (embryonic day 112), less astrocytes were found in the perifovea than in other regions equidistant from the OD. In addition, the number of astrocytes steadily declined both in the perifovea and outside until birth. During the first week after birth, there was a further dramatic decline in perifoveal astrocyte density. Double-labelling with glial fibrillary acidic protein (GFAP) immunocytochemistry and the TUNEL method showed that during the whole observation period astrocytes undergo DNA fragmentation and presumably die. However, the rate of TUNEL-positive astrocytes did not significantly differ between perifovea and other regions equidistant to the OD, and at no time did we find a significant peak of apoptosis rate. Thus, the reduction in perifoveal astrocyte density cannot be explained by missing invasion or by selectively elevated apoptosis rates in the foveal and perifoveal regions. Alternative hypotheses are discussed.

Human retinal astroglia. A comparative study of adult and the 18 month postnatal developmental stage

The immunohistochemical location of glial fibrillary acidic protein (GFAP) was used to study the state of maturation of retinal astrocytes from an 18-mo-old infant and to compare it with the situation in the adult. Infant astrocytes showed intense GFAP immunoreactivity in the perikarya and possessed spindle-like enlargements in their processes, while in the adult immunoreactivity in the perikarya was scarce and the spindle-like enlargements were not evident. Two types of astrocyte were observed in adult and child retinas: elongated and star-shaped. In the adult, the star-shaped type tend to be more stylised and to have longer processes than in the infant. In the infant, numerous astrocyte cell bodies were observed over vessels, while in the adult these were scarce. In the infant, the star-shaped astrocytes made up a honeycomb plexus, but this was not fully developed. These results suggest that at 18 mo of postnatal development the retinal astrocytes are still increasing and growing into the astroglial structure found in adults.

Contribution of O4+ oligodendrocyte precursors and astrocytes to the glial ensheathment of vessels in the rabbit myelinated streak

The barrier properties and glial ensheathment of blood vessels in the retinal myelinated streak of adult New Zealand White rabbits were characterized at the ultrastructural level by intravascular injection of horseradish peroxidase (HRP) and immuno-electron microscopy with monoclonal antibody O4 and antibodies to glial fibrillary acidic protein (GFAP). Vessels within the myelinated streak did not leak HRP, and they exhibited tight junctions between adjacent endothelial cells. However, unlike their adult counterparts, the retinal blood vessels at postnatal day 18 exhibited substantial endocytotic activity. Both GFAP+ astrocytes and O4+ cells were evident surrounding the preretinal blood vessels of the myelinated streak. Furthermore, O4+ cells exhibited features indicative of high synthetic activity, including a large proportion of extended chromatin and prominent nucleoli within the nucleus, as well as a well-developed Golgi apparatus and numerous mitochondria in the cytoplasm. O4+ cells also exhibited variable quantities of heterochromatin, indicative of early stages of cellular differentiation. These observations are consistent with previous data showing that O4+ cells in the myelinated streak include oligodendrocyte precursor cells, pre-oligodendrocytes, and immature oligodendrocytes (Morcos Y, Chan-Ling T. Glia 21:163-182, 1997). The present data indicate that the preretinal vessels of the myelinated streak possess barrier properties typical of microvasculature in the central nervous system, and that both O4+ cells and astrocytes contribute to the glial ensheathment of these vessels. These vessels thus differ markedly from the leaky preretinal vessels associated with pathological conditions such as retinopathy of prematurity.

Comparative study of astrocytes in human and rabbit retinae

Immunohistochemical location of Glial Fibrillary Acidic Protein (GFAP) was used to compare the morphology of astrocytes in vascularized and partially vascularized retinae (human and rabbit, respectively). Astrocytes in human and rabbit retinae were found in the same regions as the blood vessels. These cells in partially vascularized retinae differed from those in vascularized retina in several respects. Firstly, there were six morphological types in rabbit retina and only two in human retina. Secondly, in rabbit retinae there were astrocytes only related to blood vessels called "perivascular astrocytes" which were absent in human retinae. Thirdly, the astrocytes were located in the nerve fiber layer and ganglion cell layer in both types of retinae, but in human retinae these cells could also be seen in the internal nuclear layer. These observations demonstrate that there are many differences between astrocytes in human and rabbit retina, suggesting that rabbit retina could be used with caution as an experimental model in comparative studies with human retina.

Microglia-derived cytotoxic factors. Part I: Inhibition of tumor cell growth in vitro

The rarity of neoplasms in the adult mammalian retina has led us to hypothesize the presence or increased expression of 'tumor-inhibitory molecules' in the mature differentiated retina. We have begun to investigate the source(s) of these molecules, and the following study describes the inhibitory activity of a soluble microglia-derived cytotoxic factor on the proliferation of C6 cells, a glial tumor cell line. C6 cells were treated for 24, 48, or 72 h with basal medium or basal medium conditioned by retina-derived Muller cells (MCCM) or microglial cells (MGCM) and assayed for cell proliferation and/or cell death using various techniques involving fluorescent probes, lactate dehydrogenase release, or bromodeoxyuridine uptake. C6 cells increased in number from 24 to 72 h following incubation in basal medium or MCCM, but not in MGCM, where the cells rounded up and retracted their processes. The number of dead cells appeared to be the same in all groups at each time point. Similar findings were observed in the presence of 1-10% serum. About 25% of cells treated with basal medium for 72 h were positive for bromodeoxyuridine as compared with < 1% in MGCM-treated cultures. Our studies suggest that retina-derived microglial cells secrete soluble product(s) that inhibit the growth of C6 cells in culture. These molecules may provide protection for the mature retina against the invasion of tumor cells and may prove useful in the treatment of cancer.

Structural specializations of human retinal glial cells

Electron microscopy and immunohistochemistry have been used to study the structural specializations of astrocyte and Müller glia cells in human retinas. The astrocytes and Müller cells contribute to the formation of the internal limiting membrane, the retina, the blood vessel glial limiting membranes and the glial sheaths of the ganglion cells. Two types of junctions were observed among retinal glial cells. Adherent junctions were found between astrocytes and Müller cells, and between adjacent astrocytes. Gap junctions were only observed between astrocyte processes. These similarities suggest that astrocytes and Müller cells can perform the same functions in human retinas. Finally, the "perivascular astrocyte of Wolter", related only to the blood vessels, was not found. All the retinal astrocytes have the same ultrastructural characteristics, confirming the absence of these astroglial cells in human retinas observed by immunohistochemical techniques.

Heterogeneous morphology and tracer coupling patterns of retinal oligodendrocytes

The present study characterizes the morphology and tracer coupling patterns of oligodendrocytes in the myelinated band of the rabbit retina, as revealed by intracellular injection of biocytin or Lucifer yellow in an isolated superfused preparation. Based on the observed heterogeneity in morphology, we have grouped the presumptive oligodendrocytes into three categories termed 'parallel', 'stratified' and 'radial'. Most parallel oligodendrocytes were tracer coupled to nearby oligodendrocytes and astrocytes, whereas the stratified and radial oligodendrocytes rarely showed coupling. We conclude that the different categories of oligodendrocytes may be stages in a developmental series, with radial oligodendrocytes being premyelinating cells, parallel oligodendrocytes being mature myelinating cells and the stratified cells representing a transition between these categories.

Immunohistochemical study of human retinal astroglia

Immunocytochemical localization of glial fibrillary acidic protein (GFAP) has been used to study astrocyte distribution and morphology in whole mounted human retinas and vertical sections. Two types of astrocytes can be distinguished: elongated astrocytes are located in the nerve fibre layer (NFL); and star-shaped astrocytes are found in the ganglion cell layer (GCL). Astroglial processes join to form bundles. The bundles formed by the elongated astrocytes lie along and separate the nerve fibre bundles. Processes from star-shaped astrocytes reach towards other star-shaped astrocytes and towards the vessels to form a morphologically honeycombed plexus. These astrocytes also send other processes towards the internal nuclear layer (INL), forming an irregular plexus which accompanies the GCL capillaries that extend into the INL. Often, the cell bodies of the star-shaped GCL astrocytes lie over vessels and form cell clusters. Finally, none of the retinas examined for this study evidenced the "perivascular astrocytes" described by Wolter in the human retina.

Immunocytochemical demonstration of astrocytes and microglia in the whale brain

Whale brains have attracted the attention of neuroscientists but there are only sparse studies on whale glial cells. Here we report on immunolabeling of astrocytes by antibodies to glial fibrillary acidic protein (GFAP) or protein S-100 beta (both by the streptavidin/biotin technique), and labeling of microglial cells by Griffonia simplicifolia agglutinin (GSA I-B4, coupled to horseradish peroxidase), in the neocortex of a harbour porpoise Phocoena phocoena L. Many subpial and perivascular astrocytes were stained; they differed greatly in thickness and length of their processes. Subpial astrocytes were coarse with a few stout stem processes, whereas perivascular astrocytes deeper in the brain had many long and slender processes. Additionally, some long radial astrocytes were observed. Microglia were labeled throughout the brain, and showed similar features as 'resting' (ramified) microglia in the brain of other mammals.

In vivo staining of oligodendroglia in the rabbit retina

We have discovered that a strongly fluorescent dye, sulforhodamine 101, when injected intravitreally in vivo, very effectively stains a class of star-shaped cells in the innermost layers of the rabbit retina. The cells were strictly confined to the region containing medullated fibers and emitted dichotomously branching processes that ended up running some distance along the myelinated fibers. In favorable cases they could be seen to ensheath the fibers in a tube-like fashion. No other retinal cells were stained. Shortly (hours) after the injection, the stain appeared in the cell cytoplasm, but it later became progressively more localized to intracellular granules. Most of the dye had disappeared after 2 days. Oligodendrocytes and astrocytes are the only cells known to be confined to the region of the medullated fibers in the rabbit retina, and hence the sulforhodamine 101-stained cells should be one of these two types. Sulforhodamine 101-stained cells were indistinguishable from oligodendrocytes identified by 2',3'-cyclic nucleotide phosphodiesterase (CNP) immunohistochemistry, and sequential staining showed them to be the same. Sulforhodamine 101-stained cells were microinjected with lucifer yellow after lightly fixing rabbit retinas with formaldehyde and were found to be indistinguishable from oligodendrocytes. Glial fibrillary acidic protein staining for astrocytes showed fiber bundles that to some extent were similar to the bundles stained by sulforhodamine 101, but at the level of individual fibers, it was impossible to establish any concordance. Sulforhodamine 101 thus appears to stain oligodendrocytes rather than astrocytes in the rabbit retina. A related dye, rhodamine 123, also stained rabbit oligodendrocytes, but with poor contrast because many other cells and structures were also stained.(ABSTRACT TRUNCATED AT 250 WORDS)

A quantitative study of the lateral spread of Müller cell responses to retinal lesions in the rabbit

A wide variety of retinal pathology is associated with an increase in Müller glial cell expression of glial fibrillary acidic protein (GFAP). In this study the time course and spatial spread of the Müller cell GFAP response following argon laser photocoagulation lesions was examined in wholemounted rabbit retina. At 24 hours single focal lesions were surrounded by GFAP positive Müller cell end feet which declined in density with distance but extended as far as 2-3 mm from the lesion. The Müller cell reaction reached a maximal spread of 4-5 mm at 14 to 21 days and had started to contract by 30 days, leaving a core of GFAP positive processes immediately around the lesion site at 60 days. This zone of spread was much larger than the area of disrupted pigment epithelium. Isodensity plots did not reveal any correlation with the trajectory of retinal ganglion cell axons. The spread of reaction was more confined for lesions within the visual streak than in the dorsal or ventral retinal periphery. Multiple lesions within a focal region of retina resulted in a greater density of GFAP reactive end feet with a corresponding greater spread. However, when five to ten lesions were made in a horizontal row, the Müller cells over the entire retina became GFAP immunoreactive. This pan-retinal reaction took several days to spread, peaked at 7-14 days, and contracted back to the primary lesion sites by 2 months. This spread of Müller cell reactivity may be triggered by the diffusion of substances released by injury or it may be due to direct cellular communication. The extensive indirect effect on Müller cells of laser irradiation might be an important component of the clinical effect of laser photocoagulation and indicates a long distance communication mechanism between retinal glia which is poorly understood. This study also shows the importance of the time at which the Müller cell response is assessed.

Glia cells of the monkey retina. I. Astrocytes

Morphology and distribution of retinal astrocytes have been studied in macaque monkeys by immunocytochemical localization of glial fibrillary acidic protein (GFAP). With the exception of the fovea and the far periphery, astrocytes are ubiquitous in the nerve fiber layer (NFL) and the ganglion cell layer (GCL) of the monkey retina. The morphology of NFL astrocytes changes gradually, from star-shaped in the periphery to bipolar close to the optic disc. By contrast, GCL astrocytes maintain their star-shaped appearance throughout the retina. Astrocytes are unevenly distributed in the monkey retina, showing the highest concentration around the optic disc, and particularly low densities in the perifoveal region and the far periphery. The fovea proper is devoid of astrocytes. Employing high-resolution confocal microscopy, we could demonstrate that astrocytes form manifold contacts to blood vessels. In addition, bundles of NFL astrocyte processes are co-localized with axon bundles, individual astrocytes forming contacts to several axon bundles. In contrast, a similar affinity of astrocytes to ganglion cell somata was never observed. Thus, our data confirm and extend the current knowledge of morphology and putative function of astrocytes in mammalian and especially the primate retina.

The role of Müller cells in the formation of the blood-retinal barrier

We have compared the ability of Müller cells and astrocytes to induce the formation of barrier properties in blood vessels. Müller cells cultured from the rabbit retina, and astrocytes and meningeal cells cultured from the rat cerebral cortex, were injected into the anterior chamber of the rat eye, where they formed aggregates on the iris. We have examined the barrier properties of the vessels in those aggregates and, for comparison, the barrier properties of vessels in the retina, ciliary processes and iris. Two tracers were perfused intravascularly to test barrier properties. The movement of Evans Blue was assessed by light microscopy, and the movement of horseradish peroxidase by light and electron microscopy. Our results indicate that Müller cells share the ability of astrocytes to induce the formation of barrier properties by vascular endothelial cells, and we suggest that Müller cells play a major role in the formation of barrier properties in retinal vessels.

Retinal perivascular astroglia: an immunoperoxidase study

Different morphological types of retinal perivascular astrocytes have been studied in wholemount preparations of rabbit retina. Astrocytes were immunohistochemically demonstrated using glial fibrillary acidic protein monoclonal antibodies (GFAP clone GA-5). Three types of perivascular astrocytes were found. Type I has an ovoid perikaryon which gives rise to numerous hair-like processes, and it shows strong GFAP reactivity; these cells are associated with medium-size vessels and capillaries. Type II is star-shaped and its spherical perikaryon has a basal cone with four to ten small, protruding, radial processes; these astrocytes show high GFAP reactivity and are located on larger and medium-size vessels. Type III astrocytes show the classic, star-shaped morphology in which processes emerge directly from the perikaryon which lacks a basal cone. However, different from Types I and II, Types III astrocytes show low GFAP reactivity and are positioned between capillaries. These cells are the only ones that can contact other astrocytes of the same type to form a network.

Effect of elevated ambient glucose upon polyphosphoinositide turnover in bovine retinal endothelial cells and rat astrocytes

We investigated the turnover of polyphosphoinositides in bovine retinal microvascular endothelial cells and rat astrocytes cultured in the presence of high ambient concentrations of glucose in order to study the possible involvement of this pathway in the pathogenesis of diabetic retinopathy. a 35-45% decrease in the amount of 32P incorporated into phosphatidylinositol(4)phosphate (PIP) and phosphatidyl-inositol(4,5)biphosphate (PIP2) occurred in rat astrocytes but not bovine retinal endothelial cells grown for 14 +/- 3 days in a medium with an elevated (28 mM) glucose concentration. Incorporation of 32P into phosphatidylinositol, phosphatidylcholine, phosphatidylserine and phosphatidylethanolamine was not altered by these conditions. A 39-45% decrease in 32P incorporated into PIP/PIP2 was also found in rat astrocytes grown in 28 mM glucose which were detergent solubilized and incubated with [32P]ATP. Exposure to elevated concentrations of glucose decreased the amount of PIP/PIP2 cleaved by ionomycin or fluoroaluminate treatment, but did not disturb phospholipase C activity. Thus, the lower level of PIP/PIP2, induced by exposure to elevated concentrations of glucose, appears due to changes in phospholipid substrate levels, or polyphosphoinositide kinase activity, rather than a decrease in ATP levels or phospholipase C activity. These results suggest that high ambient glucose levels alter second-messenger generation by astrocytes. In turn, cellular interactions dependent upon these second messengers and important for maintenance of normal microvessel function in the retina may be disrupted.

Intraorbital transection of the rabbit optic nerve: consequences for ganglion cells and neuroglia in the retina

Rabbit retinae were stained with antibodies to glial fibrillary acidic protein (GFAP) at various times up to 5 months after complete unilateral intraorbital optic nerve transection, which is known to induce degeneration of ganglion cell axons and perikarya in the retina. A transient immunoreactivity for GFAP was observed in Müller glial cells that normally lack this marker. Müller-cell GFAP immunoreactivity became detectable 4 days after the lesion, but Müller cells were no longer labeled 3 months later. GFAP-labeled astrocytes located in the nerve fiber layer showed no change in immunoreactivity at any stage after transection. Application of horseradish peroxidase to the left and right superior colliculus of a rabbit whose optic nerve had been transected unilaterally 2 years before confirmed the completeness of the transection. Yet electron microscopy showed the presence of some healthy-looking ganglion cell axons in the lesioned retina, although these cells were deprived of their target. Labeling retinal wholemounts with neurofilament antibodies confirmed the presence of some ganglion cell axons and perikarya in the retina more than 2 years after transection. The course of these axons suggested that they were remnants of axons. Using antibodies to galactocerebroside (GC) we found that, as in the normal rabbit, these persisting ganglion cell axons were myelinated in the medullary rays. Although many ganglion cell axons had disappeared after 2 years, the number of neuroglial cells (including astrocytes and oligodendrocytes) present in the medullary ray region was not altered. The cell bodies of some oligodendrocytes were covered with a myelin sheath, an aberrant feature not seen normally.

Interferon-gamma and tumour necrosis factor induce expression of major histocompatibility complex antigen on rat retinal astrocytes

Cultured rat retinal astrocytes were tested by indirect immunofluorescence staining for their ability to express class I and II major histocompatibility complex (MHC) antigens under basal culture conditions and after three days of stimulation with two recombinant cytokines, rat interferon-gamma (IFN-gamma) and human tumour necrosis factor alpha (TNF alpha). Under basal culture conditions low levels of class I antigens were detected on a small percentage of cells, but there was no visible class II. IFN-gamma and TNF alpha stimulation enhanced class I expression. TNF alpha had no effect on class II expression, whereas IFN-gamma induced the expression of class II in a dose dependent manner. These findings suggest that retinal astrocytes might play a part in immunological events occurring in the retina.

Factors determining the migration of astrocytes into the developing retina: migration does not depend on intact axons or patent vessels

Astrocytes migrate into the cat retina from the optic nerve, beginning from embryonic day (E) 52. Once they have entered the retina they concentrate along major axon bundles and fail to enter regions of the retina with high densities of neurones, in particular the area centralis region of the ganglion cell layer. These nonuniformities appear as the astrocytes spread over the retina during development, and in this study we have examined factors that might control their spread. First we examined astrocytes in a retina in which the axon bundles had degenerated following an optic nerve lesion at birth. The area over which astrocytes had spread was normal, suggesting that their spread does not depend on the presence of intact axons. Second, we noted that, despite the degeneration of all ganglion cells following the nerve lesion, astrocytes still did not spread over the area centralis. Their spread is apparently not inhibited by concentrations of neurones. Third, we examined astrocytes in retinas of animals raised in an atmosphere containing 70-80% oxygen, which prevents the formation of retinal vessels. Again, the area over which the astrocytes had spread was normal, suggesting that their spread does not depend on the presence of patent blood vessels. These negative findings led us to compare the distribution of spindle cells (precursors of retinal vasculature) and astrocytes in the cat during development. The close correspondence in their topographical distribution and the earlier spread of the spindle cells lead us to suggest that spindle cells provide a basal lamina component that may guide the migration of astrocytes.

Factors determining the morphology and distribution of astrocytes in the cat retina: a 'contact-spacing' model of astrocyte interaction

The retina provides a valuable opportunity to examine the interaction of astrocytes with neurones and vasculature, in adult tissue and in vivo. We have studied astrocytes in cat retina to delineate the interactions that determine their morphology and distribution. Their morphology varied with their interaction with surrounding cells, from a classic stellate shape to an elongated bipolar form associated with axon bundles. Evidence is presented that the distribution of astrocytes across the retina is determined by their morphology and by a previously unrecognised interaction between astrocytes, which we term 'contact-spacing,' in which astrocytes maintain contact with their neighbours through their processes, but keep their somas apart. Evidence is also presented that astrocytes are not influenced in their distribution by surrounding neurones, and the influence of developmental mechanisms is identified. These observations are summarised in a contact-spacing model of astrocyte distribution, and four predictions of the model are tested. The concentration of astrocytes along axon bundles dispersed when the axons degenerate but not when vessels were prevented from forming. Further, when both axons and vessels were eliminated, the concentrations of astrocytes dispersed and they became stellate in form. Finally, in the retina of the rat, in which astrocytes show no affinity for axons, the distribution of astrocytes is essentially uniform. We suggest that the contact-spacing interaction among astrocytes provides the anatomical basis of a functional glial network extending across the retina and throughout the central nervous system.

Glial fibrillary acidic protein (GFAP) immunoreactivity in rabbit retina: effect of fixation

The binding of monoclonal and polyclonal antibodies to glial fibrillary acidic protein (GFAP) antigenic sites in the rabbit retina was shown to be sensitive to aldehyde fixation. In chemically unfixed retina, the polyclonal anti-GFAP labeled Müller cells, astrocytes, and unidentified profiles in the outer plexiform layer; the monoclonal anti-GFAP labeled Müller cell endfeet and astrocytes only. The outer plexiform layer label with the polyclonal antibody was lost after fixation for 1 hr in 1% paraformaldehyde; elsewhere, the label was reduced. Fixation also reduced labeling by the monoclonal antibody. Such fixation sensitivity may underlie the different patterns reported for retinal GFAP immunoreactivity in the literature.

Postnatal gliogenesis in the nerve fiber layer of the rabbit retina: an autoradiographic study

Cell proliferation was studied in the retina of rabbits at various postnatal stages. Autoradiography was performed with animals that received a single injection of 3H-thymidine and were sacrificed 1 hour later. This short survival time allowed the determination of the position of a cell undergoing DNA synthesis at that moment. Between birth and day 6, cells engaged in DNA synthesis were seen in the inner nuclear layer (INL) of the entire retina. Cytogenesis ceases in this layer after the first postnatal week. Few labeled cells were detectable in the INL at day 9; these were found close to the ora serrata. Thus neurogenesis, which is known to occur in this layer of the retina, ceases by that time. In the nerve fiber layer (NFL), labeled cells were found at all ages between birth and day 27, which was the oldest stage examined in this study. By using horizontal sections through the NFL of entire retinae, it was observed that almost all labeled cells were confined to the medullary ray region, which is the neuroglia (astrocyte and oligodendrocyte)-bearing part of the NFL. Microglial cells, the only cellular elements present in the NFL outside the medullary ray region, were rarely labeled, and thus do not play a major role in gliogenesis occurring in the NFL. In addition to neuroglia, some endothelial cells were labeled after day 9. It is concluded that gliogenesis taking place in the NFL persists after the cessation of neurogenesis, suggesting that both processes occur independently.

Müller cells in adult rabbit retinae: morphology, distribution and implications for function and development

We describe the morphology and distribution of Müller cells in wholemounts of rabbit retinae labelled with either monoclonal antibodies (anti-Vimentin, 3H3, 4D6, and 4H11), or intracellular horseradish peroxidase. Several new features of Müller cell organization are noted. First, Müller cells appear to compose a single morphological class and their morphology varies systematically with retinal thickness. Second, in contrast to other retinal glia, Müller cells have a neuronlike distribution, with a peak density of 10,700-15,000 cells per mm2 at the visual streak and a minimum density of 4,400-6,000 per mm2 at both the superior and inferior retinal edges. There are 4.2 +/- 0.5 x 10(6) Müller cells per retina. Third, unlike in other species, rabbit Müller cells do not contact blood vessels, suggesting that they do not participate in the transfer of metabolites or in the blood:retinal barrier. Fourth, each Müller cell has a vitread endfoot about 20-40 microns in diameter composed of numerous fimbriae. The fimbriae from a single Müller cell generally contact several axon fascicles in the nerve fibre layer, and at each point along its length each fascicle is enclosed by the overlapping fimbriae from several Müller cells. Fifth, in the inner and outer plexiform layers, numerous filamentous branchlets extend 20 microns or more from the radial trunk, interweaving with branchlets from nearby Müller cells to form dense and continuous strata. In the ganglion cell layer and outer nuclear layer, Müller cell processes completely wrap neuronal somata, whereas in the inner nuclear layer they partially wrap somata. We discuss the functional and developmental implications of these observations.

Evidence for three morphological classes of astrocyte in the adult rabbit retina: functional and developmental implications

Commercially available antibodies to glial fibrillary acidic protein (GFAP) and vimentin were used in conjunction with our own antibodies (3F11, MA1 and 4D6) to label astrocytes in the adult albino rabbit retina. Anti-GFAP labels the entire astrocyte population and shows them to be morphologically diverse. By contrast, the remaining four antibodies label different subpopulations of astrocytes. Comparisons of the shapes and distributions of cells composing these subpopulations led us to distinguish three morphological classes: Class A are predominantly perivascular astrocytes, Class B are territorial and rarely contact vessels or axons, Class C (which seems to contain four subclasses) are astrocytes that are predominantly associated with axons.

Enzyme-histochemical demonstration of microglial cells in the adult and postnatal rabbit retina

Enzyme-histochemical methods for thiamine pyrophosphatase (TPPase) and nucleoside diphosphatase (NDPase) were applied to wholemounted rabbit retinae to demonstrate the shape and distribution of microglial cells in early postnatal and adult animals. At birth, microglial cells were already present in the entire retina. They acquired their adult "resting shape" during the first 3 postnatal weeks. Early postnatally labeled microglial cells were scattered throughout the nerve fiber layer, the inner plexiform layer, and the outer plexiform layer (OPL); at adulthood, they were not detected in the OPL. Nissl-stained retinae revealed that the number of microglial cells continuously increased during postnatal development. The same Nissl-stained preparations were used to evaluate the topography of degenerating cells in the developing postnatal retina of the rabbit. Large numbers of degenerating pyknotic cells were observed throughout the entire retinal ganglion cell layer during the first postnatal week. Later their number decreased, and from the third postnatal week onward degenerating cells were rare. Also discussed is that the emergence of microglial cells during development may be related to cell death, whereas at adulthood the function(s) of microglial cells remains obscure. Evidence for the blood-derived origin of microglia was not obtained in this study. It is argued here that if this mode of development, which has been demonstrated for other species, is also applied to the rabbit retina, then microglia would have to migrate over considerable distances, since, postnatally, the rabbit retina is avascular for more than 1 week.

The rabbit retina: a suitable mammalian tissue for obtaining astroglia-free Müller cell cultures

Monolayer cultures were prepared from two distinct parts of early postnatal rabbit retinae. Cell suspensions obtained from the developing medullary ray (MR) region contained neurons, Müller (glial) cells, and astrocytes, cells obtained from the remainder (peripheral) part of the retina contained neurons and Müller cells, but no astrocytes. Müller cells lack glial fibrillary acidic protein (GFAP) immunolabeling in situ but some of them acquire faint GFAP labeling in both types of cultures. Strongly GFAP-labeled cells, most likely astrocytes, were seen in MR cultures only. We propose that the periphery of the rabbit retina is ideal for obtaining astroglia-free Müller cell cultures to study their functional properties in vitro.

The development of astrocytes in the cat retina: evidence of migration from the optic nerve

To test recent ideas of the origin of retinal astrocytes we have studied the distribution of astrocytes, identified by anti-GFAP antibodies, in the developing retina of the cat. GFAP+ cells first appeared at the optic disc at E53 (embryonic day 53). At subsequent ages, GFAP+ cells covered successively larger regions surrounding the optic disc, and were found at the edge of the retina by P35 (postnatal day 35). During development, the GFAP+ cells near the optic disc were strongly related to blood vessels and axon bundles; in a more peripheral zone they were closely associated with the immature capillary net; while the most peripheral GFAP+ cells appeared to extend exploratory processes towards the margin of the retina. The velocity at which the 'front' of GFAP+ cells spread over the retina was estimated at 170-240 microns/day. At no time during development were GFAP+ cells observed in the area centralis. Except at the area centralis, the spread of GFAP+ cells preceded the formation of capillaries, by a small but distinct margin. GFAP+ cells also extended for a short distance from the optic disc along the proximal part of the hyaloid artery. These results support the view that retinal astrocytes migrate into the retina from the optic disc, in close association with the formation of retinal vasculature.

Distribution of Müller cells in the turtle retina: an immunocytochemical study

Müller cells are the major type of glial cell in the vertebrate retina, and appear to participate in important structural and metabolic functions. Although the morphological features of Müller cells have been extensively studied, their topographic distribution across the retina has not been previously reported. We have used a Müller cell-specific monoclonal antibody, 19-33, to study the distribution of Müller cells in turtle retina. The antibody was obtained during a search for cell type-specific monoclonal antibodies in the rat retina. Immunoblotting studies show that 19-33 reacts with a 58 KDa protein that is present in Müller cells. Immunocytochemical studies with en face sections of turtle retina show that the density of Müller cells is fairly uniform across the retina although there are small regional differences. We estimate that the mean Müller cell density is about 1600 cells mm-2 of turtle retina and that each turtle retina contains about 54,000 Müller cells.

The development of astrocytes and blood vessels in the postnatal rabbit retina

Antibodies to glial fibrillary acidic protein (GFAP) have been used to study the shape and location of astrocytes in whole mount preparations of developing postnatal rabbit retina. At all developmental stages GFAP-positive astrocytes were detectable. At birth, they were few in number and only weakly labelled. With further development, their number as well as their labelling intensity increased. Following Nissl counterstaining it was observed that GFAP-positive astrocytes, always situated in the nerve fibre layer, are capable of cell division during about the first 4 postnatal weeks. GFAP-positive astrocytes were always confined to a wing-shaped area extending horizontally from both sides of the optic nerve head. It is suggested that astrocytes are not generated in the entire rabbit retina, which is in clear contrast to the second glial cell type of the rabbit retina, the Müller cell; and it has been concluded that the confinement of astrocytes to the medullary rays region in the adult rabbit is established during ontogenesis, and is not due to a secondary restriction of astrocytes to this region. Horizontal sections cut through entire rabbit retinae at various postnatal stages revealed that the first intraretinal blood vessels are not found before postnatal day 9. This is more than 1 week later than the first astrocytes are detectable. It is suggested that, at least in the rabbit, retinal astrocytes do not co-migrate with blood vessel endothelial cells from the optic disc into the retina, a hypothesis considered recently for the cat retina. It was, however, not possible to decide unequivocally if, in this material, astroglial progenitors are derived from retinal neuroepithelial cells or invade the retina from the optic nerve head.

The development of astrocytes in the albino rabbit retina and their relationship to retinal vasculature

We have examined the development of astrocytes in the albino rabbit retina, using antibodies to glial fibrillary acidic protein (GFAP) and vimentin. Vimentin immunoreactive (vimentin+) astrocyte-like cells first appear at the 24th postconceptional day (24 PCD), in a pattern similar to that of the adult. GFAP immunoreactivity was first detected in astrocytes at the 29 PCD, in a similar pattern. Vessels enter the retina from 29 PCD. The presence of astrocytes in a mature distribution prior to the ingrowth of vessels indicates that astrocytes are not dependent on the vessels for their early positioning and differentiation. In contrast with the rat and cat, we found no evidence of migration of astrocytes into the rabbit retina from the optic nerve.

Distribution of orthogonal arrays of particles in the Müller cell membrane of the mouse retina

In the present study we investigated the Müller cell membrane of the mouse retina by freeze-fracturing. The mouse retina is vascularized and the vessels running outside the nerve fiber layer are completely encased by Müller cell endfeet. Orthogonal arrays of particles (OAP) reside in all membrane areas of the Müller cells. The paravitreous as well as the pericapillary endfeet reveal a considerably higher density of OAP than the nonendfoot membranes including the perikaryal ones. This is in contrast to the Müller cell membrane of the rabbit retina studied previously (Wolburg and Berg: Neurosci, Lett., 82:273-277, 1987). There we found a completely different distribution of OAP; practically all OAP reside in the endfoot membrane facing the vitreous body. The nonendfoot and perikaryal membranes were devoid of OAP. The OAP distribution in both species corresponds roughly to the distribution of the K+ conductances measured by Newman (J. Neurosci., 7:2423-2432, 1987). The putative relationship between OAP and K+ channels, including functional aspects, is discussed.

Astrocytes in the guinea pig, horse, and monkey retina: their occurrence coincides with the presence of blood vessels

In the present study the distribution of astrocytes in the nerve fiber layer (NFL) has been studied in the sparsely vascularized retinae of the guinea pig and horse and in the richly vascularized retina of the Old World monkey (Cercopithecus aethiops) using immunocytochemical methods. In the guinea pig retina glial fibrillary acidic protein (GFAP)-positive astrocytes could not be detected. They were found, however, in the myelinated region of the optic nerve. The optic nerve head and a small retinal region immediately adjacent to it contained few vimentin-positive astrocytes. Histological sections confirmed the restriction of astrocytes to a small retinal region and showed that this is also the only retinal area that is vascularized. Astrocytes showing GFAP and vimentin immunoreactivity were absent from most of the horse retina. They were found only in a narrow zone close to the optic disc, which is also the only region of the horse retina that is vascularized. Thus, as in the rabbit retina (Schnitzer: J. Comp. Neurol. 240:128-142, 1985), in the guinea pig and horse retina astrocytes are not present ubiquitously in the NFL but coexist with blood vessels. In the monkey retina, GFAP-positive astrocytes were found ubiquitously in the NFL. Astrocytes were absent from the avascular foveal region only. It is suggested that the concurrence of retinal astrocytes and intraretinal vascularization may be a feature common to many, if not all, mammalian species.

Retinal astrocytes: their restriction to vascularized parts of the mammalian retina

The distribution of astrocytes has been studied in whole-mounted horse and monkey retinae by the immunocytochemical localization of glial fibrillary acidic protein (GFAP). In the horse, astrocytes were found to be restricted to a narrow zone close to the optic nerve head. This is also the only region of the horse retina that is vascularized. In the monkey, astrocytes were found ubiquitously in the nerve fiber layer of the retina, apart from the avascular fovea centralis which lacked astrocytes. These observations strongly suggest that retinal astrocytes co-occur with blood vessels, a feature which may be common among mammals.

Immunocytochemical localization of S-100 protein in astrocytes and Müller cells in the rabbit retina

The localization of S-100 protein was studied in histological sections of retinae from adult rabbits. By use of double-immunolabeling techniques it was shown that most but not all radially oriented vimentin-positive Müller cells were co-labeled by an antiserum to S-100 protein. Glial fibrillary acidic protein-positive astrocytes, which in the rabbit retina are restricted to the medullary rays formed by myelinated optic nerve fibers, consistently showed S-100 protein immunoreactivity. The present report shows that, with respect to S-100 protein staining, Müller cells represent a heterogeneous population of glial elements.