Transferrin binding protein is expressed by oligodendrocytes in the avian retina.

It has been documented that some axons of ganglion cells in the nerve fiber layer of avian retina are wrapped in a myelin sheath. However, the identity of myelin-forming cells has not been established. In this study we demonstrated immunohistochemical evidence for the existence of a large population of oligodendrocytes in avian retina, using an antiserum against transferrin binding protein (TfBP), the avian homologue of the mammalian GRP 94 family of stress-regulated proteins. TfBP+ cells were mostly confined to the ganglion cell and optic nerve fiber layers of the retina, in which they were closely associated with the soma and axons of ganglion cells. The double-labeling experiments clearly show that TfBP is specific to oligodendrocytes. The morphology, distribution, and antigenic properties indicated by our findings suggest that TfBP+ cells are retinal oligodendrocytes that may be responsible for the myelination of ganglion cell axons in avian retina. A putative tropic role of TfBP+ oligodendrocytes to the ganglion cells is also discussed in conjunction with the physical properties of TfBP and avascular retinae of birds.

Origins of the extracellular glutamate released during total metabolic blockade in the immature retina.

Previous studies have shown that complete blockade of metabolism in embryonic chick retina causes a time-dependent increase in the release of glutamate into the extracellular space. The present study examined the cellular source of this glutamate, i.e., neuronal and/or glial. Pure cultures of retinal neurons or glia were labeled for 10 min at 37 degrees C with [3H]acetate. Retinal glia, but not retinal neurons, were found to selectively and preferentially metabolize acetate, thus producing 3H-labeled amino acids in the glial compartment. This finding provides direct evidence to substantiate findings from several other laboratories that have indirectly determined the preferential metabolism of acetate by glia by using mixed neuronal/glial populations. To study the cellular source of glutamate released during total metabolic blockade, whole retina were prelabeled with [3H]acetate plus [U-14C]glucose (to label the neuronal compartment). Total metabolic blockade was instituted with a combination of iodoacetate (IOA) plus KCN, and the release of glutamate into the medium was followed at 5, 15, and 30 min. During total energy blockade, net extracellular glutamate was not elevated at 5 min [0.17 +/- 0.02 vs. 0.12 +/- 0.01 microM for treated vs. control retina (means +/- SEM), respectively], but was increased significantly at 15 (1.2 +/- 0.26 microM) and 30 min (2.6 +/- 0.22 microM). Total [3H]glutamate in the medium during IOA/KCN treatment was unchanged at 5 min, but was increased 1.5- and threefold above basal levels at 15 and 30 min, respectively. During the time when extracellular glutamate increased, the specific activity of [3H]glutamate remained fairly constant, 731 +/- 134 and 517 +/- 82 dpm/nmol (means +/- SEM) at 15 and 30 min, respectively. In contrast, 14C-labeled glutamate in the medium did not increase during IOA/KCN treatment and paralleled basal levels. Thus, the specific activity of 14C-labeled extracellular glutamate decreased from 309 +/- 87 dpm/nmol at 15 min to 42 +/- 8 dpm/nmol at 30 min. Prior loading of the tissue with 0.5 mM trans-pyrrolidine-2,4-dicarboxylate (t-PDC), a glutamate transport inhibitor, blocked 57% of the glutamate released at 30 min of IOA/KCN exposure, suggesting that reversal of an Na+-dependent glutamate transporter was a key contributor to the appearance of extracellular glutamate during energy deprivation. The increase in extracellular [3H]glutamate, constancy of the specific activity of extracellular [3H]glutamate, decrease in the specific activity of extracellular [14C]glutamate, and attenuation of release by prior loading with t-PDC indicate that glial pools of glutamate released via reversal of the transporter contribute significantly to the rise in extracellular glutamate after metabolic inhibition in this preparation.

Identification of a population of amacrine cells rich in insulin receptors.

Insulin receptor immunoreactivity in the developing chick retina was examined by immunocytochemistry. Insulin receptor immunoreactivity could be detected throughout the retina at all stages studied. Beginning at E12, a limited number of amacrine cells located in the inner level of the inner nuclear layer were strongly immunoreactive. By E19 there was a decrease in immunoreactivity throughout the retina, with the exception of the ganglion cell layer and a few amacrine cells and their process; this distribution was present in 3-day-old posthatched chick retina. The pattern of immunoreactivity of insulin receptors may indicate a unique role for insulin in the development and physiology of some amacrine cells.

High affinity binding of transferrin in cultures of embryonic neurons from the chick retina.

Immunohistochemical and autoradiographic analysis of neuronal cultures from embryonic day 8 (E8) and day 11 (E11) chick retina indicate that transferrin receptors and binding sites are present on soma and neurites. Cultures maintained in the presence of transferrin expressed elevated transferrin binding due to an increase in the number of transferrin receptors. Cultures from E11 neural retina exhibited a decrease in transferrin binding when compared to E8 cultures. This appears to be due to a decrease in the number of binding sites. Neurons maintained in a transferrin-free medium supplemented with 0.4 microM of iron sulfate generally expressed slight increases in transferrin binding.

Transferrin can alter physiological properties of retinal neurons.

The role of transferrin as a possible neurotransmitter was examined in cultured chick retinal cells. Brief exposure to transferrin caused a dramatic and transient increase in intracellular calcium levels in approximately 20% of the total population of cultured retinal neurons. The increase in intracellular calcium was observed in cell bodies and neuronal processes. Electrophysiological analysis of a subset of the population, bipolar-like neurons, demonstrated that more than half of these cells responded to the application of transferrin with a transient membrane depolarization. Under voltage clamp conditions, the currents evoked by transferrin were similar to glutamate in that they both displayed non-linear voltage dependence. Furthermore, acute transferrin exposure resulted in a 200% increase in the amount of Na+ independent [3H]glutamate binding observed in these cultures. These results suggest that transferrin may function as a neurotransmitter or neuromodulator in the developing vertebrate nervous system.

The ontogeny of transferrin receptors in the embryonic chick retina: an immunohistochemical study.

Transferrin receptor (TfR) immunoreactivity in the developing chick retina was examined. Immunoreactivity was detectable in the ganglion cells of embryonic day (E) 4 retina. At E9, diffuse TfR immunoreactivity appeared in the outer portion of the inner nuclear layer. Amacrine cells were the most intensely TfR-positive cells in the inner nuclear layer. At E11, the inner segment of photoreceptor cells showed moderate immunoreactivity. With the appearance of the outer segments, positive immunoreactivity was observed in these structures. TfR's developmental distribution in the retina may reflect the developmental and physiological role of transferrin.

Excitatory amino acid-induced toxicity in chick retina: amino acid release, histology, and effects of chloride channel blockers.

Acute excitotoxicity in embryonic chick retina and the ability of Cl- channel blockers to prevent toxicity were evaluated by measurement of endogenous amino acid release and histology. Treatment of retina with kainate, quisqualate, or N-methyl-D-aspartate resulted in a large dose-dependent release of gamma-aminobutyric acid and taurine, moderate release of glutamine and alanine, and no measurable release of glutamate or aspartate. Concentrations inducing maximal gamma-aminobutyric acid release were 50 microM quisquaalate, 100 microM kainate, and 100 microM N-methyl-D-aspartate. Treatment with 1 mM glutamate resulted in significant gamma-aminobutyric acid release, as well as an elevation in medium aspartate levels. Typical excitotoxic retinal lesions were produced by the agonists and, at the lower concentrations tested, revealed a regional sensitivity. There was a positive correlation between the amount of gamma-aminobutyric acid release and the extent of tissue swelling, suggesting that release may be secondary to toxic cellular events. Omission of Cl- completely blocked cytotoxic effects due to kainate or glutamate. Likewise, addition of the Cl-/bicarbonate anion channel blocker 4,4'-diisothiocyanatostilbene-2,2'-disulfonate at 600 microM protected retina from cytotoxic damage from all excitotoxic analogs and restored amino acid levels to baseline values. Furosemide, which blocks Na+/K+/2Cl- cotransport, was only minimally effective in reducing amino acid release induced by the agonists. Consistent with the latter, histological examination showed the continued presence of the lesion but with general reduction of cellular edema. These results indicate that although influx of Cl- is a central component of the acute excitotoxic phenomenon, mechanisms other than passive Cl- flux may be involved.

Transferrin concentration and location during formation of chick retina: developmental correlates.

The amount of transferrin in chick retina was measured during development and compared to transferrin location seen immunocytochemically. Between embryonic day 6 (E6), and 5 days post hatching, two periods occur in which transferrin concentrations rise sharply and decline. During the first, transferrin concentration rises 5-fold between E6 and 10, then rapidly declines by E14. A second increase begins on E17 and peaks by E19-20. Immunocytochemical findings demonstrate that during the first rise in concentration, transferrin is located primarily in neuritic layers. Later in development, when levels again increase, newly forming photoreceptor outer segments are strongly transferrin positive. These findings are discussed in light of developmental events occurring during retinal maturation.

Transferrin and iron in cultured chick embryonic neurons: a comparison between human and chick transferrins.

Transferrin was not required for the short-term survival of cultured chick retinal neurons. Both human and chick transferrin failed to enhance the in vitro survival of 8- or 11-day embryonic chick retinal neurons when cultured in a defined medium. Furthermore, maintenance of neurons in the presence of chick transferrin antibody did not alter in vitro survival. Retinal neurons, however, could bind and internalize human or chick transferrin when assayed for by fluorescence immunohistochemical techniques. Binding and internalization of chick transferrin appeared to be greater than human transferrin. Iron uptake was measured in cultures maintained in the absence of transferrin. After incubation with 59FeCl3, iron uptake was 3.5 +/- 1.1 fmoles/cell. The presence of chick transferrin antibody did not significantly alter the amount of iron uptake occurring in this assay. In a comparison of human and chick transferrin mediated iron uptake, chick transferrin was 50% more effective than human transferrin in transporting iron. This study demonstrates that cultured embryonic retinal neurons are not dependent on transferrin for survival or iron uptake, although they actively bind and internalize transferrin. Results also demonstrate that whereas cultured chick retinal neurons can bind and utilize human transferrin, they do so with less efficiency than chick transferrin.

Transferrin in chick retina: distribution and location during development.

Chick retinas from embryonic day 6 (E6) to 3 weeks post-hatching were examined for the presence and location of endogenous transferrin. Immunocytochemistry revealed that transferrin was differentially distributed in retinal layers. Furthermore, the pattern of transferrin distribution changed with developmental age. At day E6, transferrin was found in 2 distinct bands which were located in the area of the Müller cell end-feet. By day E9, additional regions of transferrin immunoreactivity could be found in the inner and outer plexiform layers (IPL, OPL) and the nerve fiber layer (NFL). These latter 3 bands (IPL, OPL and NFL) became more prominent from E9 until E17 as the synaptic layers and nerve fiber layer increased in density and maturation. Perikarya in the nuclear layers size, density and maturation. Perikarya in the nuclear layers were negative. At day E17 and later, the newly forming outer segments of photoreceptor cells were strongly reactive for transferrin while the somas of the photoreceptor cells, in the ONL, were negative. Retinas from chicks 1 day to 3 weeks post-hatching retained strong immunoreactivity for transferrin in the photoreceptor cell outer segments and OPL, lessened immunoreactivity in the IPL and loss of immunoreactivity in the NFL. Iron distribution in the retina for all ages examined showed only 2 bands that locally corresponded to the Müller cell end-feet. Iron stores were not found in the synaptic layers or photoreceptor cell outer segments. These studies suggest an iron storage function for retinal glia and a role for transferrin in neuronal development and differentiation.

Neurons and glia in purified retinal cultures identified by monoclonal antibodies to intermediate filaments.

Two monoclonal antibodies known to bind intermediate filaments were used to identify neurons and glia from embryonic chick neural retina. Neurofilament specific antibody RT-97-F1, bound neuroepithelial cells, migrating neurons, as well as the photoreceptor layer, plexiform layers and optic fiber layer. The other, 3A7, bound intermediate filaments of Müller cells. In purified neuronal cultures, RT-97-F1 bound 75, 83 and 98% of the cells present at 24, 48 and 72 h respectively, while 3A7 bound 26, 15 and 0.3% for the same times in vitro. In purified glial cultures, RT-97-F1 showed a weak perinuclear binding, while 3A7 strongly bound intermediate filaments of nearly 100% of the cells. These antibodies confirmed and quantitated the high purity of our cultures.

Glutamate and kainate are not directly toxic to developing amacrine cells: analysis in a Lucifer Yellow-labeled population.

Lucifer Yellow (LY)-labeled retinal amacrine cells were examined for toxin sensitivity to glutamate (Glu) or kainate (KA) and for high-affinity uptake (HAU) of Glu. In vitro, a 24-h exposure to either toxin caused a 50% loss of neurons with no loss of LY neurons. A 1-h exposure of intact retina to the toxins resulted in a two-fold increase in non-viable cells with no change in LY neurons. While histological damage within the retina was seen following toxin exposure, the LY population appeared unaffected. HAU in vitro was found in 5% of neurons but never colocalized with LY neurons. These studies show that Glu and KA do not directly affect the LY-labeled amacrine cells and support the hypothesis that their eventual loss may be an indirect consequence of toxin exposure.

Lucifer Yellow labeling of embryonic chick retinal amacrine cells in vitro.

Lucifer Yellow (LY) selectively labels embryonic chick amacrine cells from days 11 until 14 in vivo. Its usefulness as an in vitro marker was investigated. In vivo labeling and subsequent culturing was not possible due to dye leakage. Neurons, however, could be labeled at various times in vitro. The number of neurons labeled with LY in vitro was consistent with the number of neurons expected to be labeled and increased when selected areas of the retina known to be rich in LY-labeled neurons were used in culturing. Neurons plated at times when labeling was not found in vivo (Embryonic day 8, E8) began to label only at times that were equivalent to times when labeling was found in vivo (E10-E11). This suggests that the selectivity of labeling is preserved in vitro and that LY can be used as an in vitro marker for retinal amacrine cells.

The effects of glutamate and kainate on cell proliferation in retinal cultures.

Two different monolayer culture preparations were used to investigate the effects of glutamate and kainate on retinal cell proliferation. Glial monolayer cultures were produced from 8-day-old chick embryo neural retina. Glial monolayers were exposed to glutamate (5 mM) or kainate (0.4 mM) at either 3 or 5 days in vitro. It was found that exposure to glutamate at 3 days in vitro significantly reduced the number of glia present at 6 days, but a day 5 exposure or exposure to kainate did not significantly affect the number of glia. This glutamate-induced decrease in cell number was seen even though thymidine kinase (EC 2.7.75) and thymidylate synthetase (EC 2.1.1-) activities were higher in treated cultures. Neuroblast monolayer cultures were produced from 6-day-old chick embryo neural retina. These cultures also were exposed to glutamate (5 mM) or kainate (0.4 mM) Glutamate, but not kainate treated cultures, showed a decrease in neuroblast survival and proliferation compared with controls. Thus, at the concentrations tested, glutamate appears to be a general retinal toxin, while the survival of immature neurons or glia is not affected by kainate.

A diffusible molecule may mediate glutamate neurotoxicity.

Purified neuronal cultures from 8-day chick embryo neural retinas were incubated in either serum-free N1 or N1 supplemented with heart extract. The addition of glutamate decreased neuronal survival only in the presence of heart extract; however, if heart extract was dialyzed, neuronal survival remained at control values. This suggests a diffusible molecule may mediate glutamate neurotoxicity Kainic acid decreased neuronal survival in all conditions tested.

GABA uptake and release in purified neuronal and nonneuronal cultures from chick embryo retina.

Uptake and release of gamma-aminobutyric acid (GABA) have been studied using glia-free, purified neuronal cultures from 8-day chick embryo retina. At 3 days in vitro 65% of the neurons showed high-affinity GABA uptake. These neurons appeared heavily labeled after incubation in 5 X 10(-8) M [3H]GABA, but no labeling was detected when the incubation was carried out at 4 degrees C, or in the absence of Na+ ions. Diaminobutyric acid (DABA) also blocked completely the neuronal uptake of GABA, while beta -alanine was ineffective at similar concentrations. At 6 days in vitro Na+- and temperature-dependent GABA uptake was present in 50% of the neurons. In addition, in 80% of those neurons the uptake was insensitive to DABA or beta -alanine, whereas in the remaining 20% it was blocked by DABA but not by beta -alanine. Important developmental changes were also found in the capacity of the neurons to release GABA into the medium. Spontaneous GABA release (i.e. that taking place in regular medium, containing 5 mM K+) was higher at 3 than at 6 days in vitro. However, increasing the K+ concentration to 56 mM had minimal effects at 3 days in vitro, but induced a 2 to 3-fold increase in GABA release at 6 days in vitro. This K+-induced release appeared to be Ca2+-dependent, since it was substantially reduced the presence of 10 mM Co2+. Cultures containing a confluent monolayer of nonneuronal flat cells were generated by seeding retinal cell suspensions on poorly adhesive substrata. Retina nonneuronal cells showed, during the first 10 days in vitro, a high-affinity mechanism for GABA uptake which was Na+- and temperature-dependent, and was reduced by 85% by DABA but was practically unaffected by beta-alanine. This uptake mechanism seemed to be lost towards the end of the second week in vitro, and could not be detected after 21 days culture.