Immunocytochemical localization of the amino acid neurotransmitters in the chicken retina

Postembedding immunocytochemistry was used to determine the cellular localization of the amino acid neurotransmitters glutamate, aspartate, gamma-aminobutyric acid (GABA), and glycine in the avian retina. The through retinal pathway was glutamatergic, with all photoreceptors, bipolar cells, and ganglion cells being immunoreactive for glutamate. Bipolar cells displayed the highest level of glutamate immunoreactivity, with the cell bodies terminating just below the middle of the inner nuclear layer. All lateral elements, horizontal cells, amacrine cells, and interplexiform cells were immunoreactive for glycine or GABA. The GABAergic neurons consisted of two classes of horizontal cells and amacrine cells located in the lower part of the inner nuclear layer. GABA was also localized in displaced amacrine cells in the ganglion cell layer, and a population of ganglion cells that co-localize glutamate and GABA. Both the horizontal cells and GABAergic amacrine cells had high levels of glutamate immunoreactivity, which probably reflects a metabolic pool. At least two types of horizontal cells in the avian retina could be discriminated on the basis of the presence of aspartate immunoreactivity in the H2 horizontal cells. Glycine was contained in a subclass of amacrine cells, with their cell bodies located between the bipolar cells and GABAergic amacrine cells, two subclasses of bipolar cells, displaced amacrine cells in the ganglion cell layer, and ganglion cells that colocalize glutamate and glycine. Glycinergic amacrine cells had low levels of glutamate. We have also identified a new class of glycinergic interplexiform cell, with its stellate cell body located in the middle of the inner nuclear layer among the cell bodies of bipolar cells. Neurochemical signatures obtained by analyzing data from serial sections allowed the classification of subclasses of horizontal cells, bipolar cells, amacrine cells, and ganglion cells. © 1993 Wiley-Liss, Inc.     

Localization of two calcium binding proteins, calbindin (28 kD) and parvalbumin (12 kD), in the vertebrate retina

We used immunocytochemistry to locate two calcium binding proteins, calbindin (CaB) and parvalbumin (PV), in the retina of goldfish, frog, chick, rat, guinea pig, dog, and man. The location of CaB depended on the type of dominant photoreceptor cells in birds and mammals. In cone-dominant retinas such as those of the chick, CaB-like immunoreactivity was found in the cones, cone bipolars, and ganglion cells. Amacrine cells 5-12 microns across were also labeled. In rod-dominant retinas, such as those of the rat, guinea pig, and dog, horizontal cells, small amacrine cells (about 6 microns across), and cells in the ganglion cell layer were labeled. In the human retina, which has both cones and rods in abundance, cones, cone bipolars, ganglion cells, horizontal cells, and small and large amacrine cells were labeled. In the frog and goldfish, the level of CaB-like immunoreactivity was low. In the frog, a few cones, amacrine cells, and cells in the ganglion cell layer were labeled. No immunoreactive structures were seen in the goldfish retina. PV-like immunoreactivity was found in chicks, rats, and dogs. No such immunoreactive structures were seen in the other species. In the chick, only amacrine cells were labeled. In the rat, amacrine cells and several displaced amacrine cells were labeled. In the dog, in addition to amacrine cells and displaced amacrine cells, horizontal cells were strongly labeled. Thus, PV-like immunoreactivity was found in those elements relating to the modulation of the main pathway of the visual transmission system.     

Aspartate aminotransferase-like immunoreactivity in the guinea pig and monkey retinas

The excitatory amino acids, aspartate and glutamate, have been proposed as retinal neurotransmitters. Aspartate aminotransferase (AAT) is an enzyme which is involved in the routine metabolism of these amino acids and may be involved in the specific synthesis of glutamate and/or aspartate for use as a neurotransmitter. On the basis of the hypothesis that increased levels of aspartate aminotransferase may reflect a transmitter role for aspartate and/or glutamate, we have localized aspartate aminotransferase in the guinea pig and cynamolgus monkey retinas with light and electron microscopic immunohistochemical techniques. AAT-like immunoreactivity is localized to the cones of guinea pig retina and to monkey rods. Both species contain a subpopulation of immunoreactive amacrine cells as well as a subpopulation of immunoreactive cells in the ganglion cell layer. Immunostaining is seen in bipolar cells and terminals in the monkey but not in the guinea pig retina. We have performed quantitative analysis of the immunoreactive staining in the outer plexiform layer and described the synaptic organization of immunoreactive processes in the inner plexiform layer (IPL). Labeled amacrine processes in both species form synaptic contacts predominantly to and from bipolar terminals in the inner third of the IPL and to and from other amacrine and small unidentified processes in the outer portion of the IPL. The majority of labeled bipolar terminals in the monkey retina are seen in the inner third of the IPL where they synapse exclusively onto amacrine processes. Labeled bipolar terminals in the outer third of the IPL occasionally synapse onto ganglion processes.     

Immunocytochemical localization of the glutamate transporter GLT-1 in goldfish (Carassius auratus) retina

Glutamate is the major excitatory neurotransmitter in the retina of vertebrates. Electrophysiological experiments in goldfish and salamander have shown that neuronal glutamate transporters play an important role in the clearance of glutamate from cone synaptic clefts. In this study, the localization of the glutamate transporter GLT-1 has been investigated immunocytochemically at the light and electron microscopical levels in the goldfish retina using a GLT-1-specific antibody. GLT immunoreactivity (IR) was observed at the light microscopical level in Müller cells, bipolar cells, the outer plexiform layer (OPL), and the inner plexiform layer (IPL). At the electron microscopical level, membrane-bound and cytoplasmic GLT-IR in the OPL was located in finger-like protrusions of the cone terminal located near the invaginating postsynaptic processes of bipolar and horizontal cells. GLT-IR was not observed in the vicinity of synaptic ribbons. This location of GLT-1 allows modulation of the glutamate concentration in the synaptic cleft, thereby shaping the dynamics of synaptic transmission between cones and second-order neurons. In the inner IPL, GLT-IR was observed in the cytoplasm and was membrane bound in mixed rod/cone bipolar cell terminals and cone bipolar cell terminals. The membrane-bound GLT-1 was generally observed at some distance from the synaptic ribbon. The morphology of the bipolar cell terminal together with the localization of GLT-1 suggests that at least these glutamate transporters are not primarily involved in rapid uptake of glutamate release by the bipolar cells. The GLT-IR in the cytoplasm of Müller cells was located throughout the entire goldfish retina from the outer limiting membrane to the inner limiting membrane. The location of GLT-1 in Müller cells is consistent with the role of Müller cells in converting glutamate to glutamine.     

Amino acid signatures in the developing mouse retina

This study characterizes the developmental patterns of seven key amino acids: glutamate, γ-amino-butyric acid (GABA), glycine, glutamine, aspartate, alanine and taurine in the mouse retina. We analyze amino acids in specific bipolar, amacrine and ganglion cell sub-populations (i.e. GABAergic vs. glycinergic amacrine cells) and anatomically distinct regions of photoreceptors and Müller cells (i.e. cell bodies vs. endfeet) by extracting data from previously described pattern recognition analysis. Pattern recognition statistically classifies all cells in the retina based on their neurochemical profile and surpasses the previous limitations of anatomical and morphological identification of cells in the immature retina. We found that the GABA and glycine cellular content reached adult-like levels in most neurons before glutamate. The metabolic amino acids glutamine, aspartate and alanine also reached maturity in most retinal cells before eye opening. When the overall amino acid profiles were considered for each cell group, ganglion cells and GABAergic amacrine cells matured first, followed by glycinergic amacrine cells and finally bipolar cells. Photoreceptor cell bodies reached adult-like amino acid profiles at P7 whilst Müller cells acquired typical amino acid profiles in their cell bodies at P7 and in their endfeet by P14. We further compared the amino acid profiles of the C57Bl/6J mouse with the transgenic X-inactivation mouse carrying the lacZ gene on the X chromosome and validated this animal model for the study of normal retinal development. This study provides valuable insight into normal retinal neurochemical maturation and metabolism and benchmark amino acid values for comparison with retinal disease, particularly those which occur during development.     

The effects of intraocular injection of kainic acid on the synaptic organization of the goldfish retina

Kainic acid (KA), an extended analog of L-glutamate, was injected into the eyes of living goldfish. After survival times ranging from 15 min to 6 days, retinae were inspected for KA-induced degeneration at both the LM and EM levels. KA had little effect on photoreceptors, mixed rod-cone bipolar cells, Müller cells, at least two types of amacrine cells and the optic nerve. Reversible edema was seen in both rod and cone horizontal cells. Pure cone bipolar cells and the majority of amacrine cells appeared to be destroyed by KA. The effect of KA is selective not only on the cell types involved, but also in the location of KA-induced edema on the affected cells, i.e., soma and proximal portions of dendrites of cone horizontal cells as opposed to the distal ends of dendrites of rod horizontal cells. Implications of these data are discussed in regard to the use of KA as a probe for glutamatergic pathways in the retina. One hypothesis suggested by our results is that rods use glutamate whereas cones use aspartate as their neurotransmitter.     

Shifting of parvalbumin expression in the rat retina in experimentally induced diabetes

The AII amacrine cell, a unique rod signal integrator passing through the cone bipolar cell to ganglion cells, uses parvalbumin as a transducer of cytosolic calcium ion signals in the mammalian retina. For clarification of whether AII amacrine cell network contributes to the early neuropathogenesis of diabetic retinopathy, this study first analyzed alteration of parvalbumin expression in experimental diabetic retinas using immunohistochemical methods. Parvalbumin immunoreactivity was found in AII amacrine cells, some amacrine cells of a wide-field type, and displaced amacrine cells of the normal rat retina. During diabetes, cell density of each parvalbumin immunoreactive amacrine cell type showed no large changes despite decrease in immunoreactivity especially in AII amacrine cells. In addition to these parvalbumin immunoreactive amacrine cell types, a type of cone bipolar cells co-expressing glutamate transporter 1b and connecting electrically with AII amacrine cells appeared clearly by 4 weeks of diabetes, and thereafter sharply increased in number to that of AII amacrine cells. Protein levels of parvalbumin throughout the diabetic retinas also showed no large changes, except a transitional slight increase at 4 weeks of diabetes. These results suggest that the parvalbumin expression propagates from AII amacrine cells to a type of cone bipolar cell through electrical synapses due to dysfunction of biased mechanism in calcium ion buffering, caused by diabetic injury, and thus AII amacrine cells are closely involved in neuropathogenesis of ongoing diabetic retinopathy.     

Retinal GABA neurons: localization in vertebrate species using an antiserum to rabbit brain glutamate decarboxylase

Retinal gamma-aminobutyric acid (GABA) neurons have been localized immunocytochemically using a new antiserum against rabbit brain glutamate decarboxylase (GAD). The animals examined were: dogfish, ratfish, goldfish, catfish, turtle, chick, mouse, rat, pig, rabbit, cat and New World monkey. GAD-containing processes, observed as punctate deposits of immunochemical reaction product, formed discrete bands within the inner plexiform layers of all retinas examined. Immunoreactive, and therefore presumably GABAergic, amacrine cells were observed in all species. Displaced GABAergic amacrine cells were observed in the retinas of goldfish, catfish, turtle and chick, and sparsely in the rabbit as well. GABAergic horizontal cells were detected in catfish, goldfish, chick and turtle. Interplexiform cells in the cat and the rat were clearly immunoreactive for glutamate decarboxylase; this is the first report of GABAergic interplexiform cells in the rat.     

Spatial density and immunoreactivity of bipolar cells in the macaque monkey retina.

The anatomical substrates of spatial and color vision in the primate retina are investigated by measuring the immunoreactivity and spatial density of bipolar, amacrine and horizontal cells in the inner nuclear layer of the macaque monkey retina. Bipolar cells can be distinguished from amacrine and horizontal cells by their differential immunoreactivity to antisera against glutamate, glycine, GABA, parvalbumin, calbindin (CaBP D-28K), and the L7 protein from mouse cerebellum. The spatial density of bipolar cells is compared to the densities of photoreceptors and ganglion cells at different retinal eccentricities. In the centralmost 2 mm, cone bipolar cells outnumber ganglion cells by about 1.4:1. The density of cone bipolar cells is thus high enough to allow for input to different (parasol and midget) ganglion cell classes by different (diffuse and midget) bipolar cell classes. The density gradient of cone bipolar cells follows closely that of ganglion cells in central retina but falls less steeply in peripheral retina. This suggests that the convergence of cone signals to the receptive fields of ganglion cells in the peripheral retina occurs in the inner plexiform layer. The density of cone bipolar cells is 2.5-4 times that of cones at all eccentricities studied, implying that cone connectivity to bipolar cells remains constant throughout the retina. Different subgroups of bipolar cells are distinguished by their relative immunoreactivity to the different antisera. All rod and cone bipolar cells show moderate to strong glutamate-like immunoreactivity. The bipolar cells that show weak to moderate GABA-like immunoreactivity are also labeled with the antiserum to the L7 protein and are thus identified as rod bipolar cells. Nearly half of all cone bipolar cells showed glycine-like immunoreactivity. The results suggest that the inhibitory neurotransmitter candidates GABA and glycine are segregated respectively in rod and cone bipolar cell pathways. A diffuse, cone bipolar cell type can be identified by the anti-parvalbumin and the anti-calbindin antisera. All horizontal cells show parvalbumin-like immunoreactivity. Nearly all amacrine cells show GABA-like or glycine-like immunoreactivity; a variety of subpopulations also show immunoreactivity to one or more of the other markers used.     

Retinal GABA neurons: Localization in vertebrate species using an antiserum to rabbit brain glutamate decar☐ylase

Retinal γ-aminobutyric acid (GABA) neurons have been localized immunocytochemically using a new antiserum against rabbit brain glutamate decar☐ylase (GAD). The animals examined were: dogfish, ratfish, goldfish, catfish, turtle, chick, mouse, rat, pig, rabbit, cat and New World monkey. GAD-containing processes, observed as punctate deposits of immunochemical reaction product, formed discrete bands within the inner plexiform layers of all retinas examined. Immunoreactive, and therefore presumably GABAergic, amacrine cells were observed in all species. Displaced GABAergic amacrine cells were observed in the retinas of goldfish, catfish, turtle and chick, and sparsely in the rabbit as well. GABAergic horizontal cells were detected in catfish, goldfish, chick and turtle. Interplexiform cells in the cat and the rat were clearly immunoreactive for glutamate decar☐ylase; this is the first report of GABAergic interplexiform cells in the rat.     

Glutamate, GABA, and glycine in the human retina: an immunocytochemical investigation.

The distribution of the neuroactive amino acids glutamate, GABA, and glycine in the human retina was examined in consecutive semithin sections treated with antisera specific for fixed glutamate, GABA, and glycine, respectively. Glutamate immunoreactivity was conspicuous in all photoreceptor cells (rods more strongly labelled than cones), and in a majority (85-89%) of the cells in the inner nuclear layer (INL). Rod spherules and cone pedicles showed a greater enrichment of glutamate immunoreactivity than the parent cell bodies and inner segments. Also, structures of the inner plexiform layer (IPL) were labelled. A large majority (83-91%) of cells in the ganglion cell layer (GCL) was strongly stained, as were most axons in the nerve fibre layer. Müller cell processes appeared unstained. GABA immunoreactivity was present in presumed amacrine but not in bipolar-like cells. The stained cells were restricted to the inner 1/3 of the INL and were more frequent in central than in peripheral retina (40% and 26% of all cells in the inner 1/2 of INL, respectively). GABA positive cell processes, probably originating from interplexiform cells, appeared to traverse the INL and end in the outer plexiform layer. Dense immunolabeling was found in the IPL. GABA immunoreactive cells (some also stained for glutamate) comprised 23% of all GCL cells in the peripheral retina, but only 5% in the central retina. Most of them were localized adjacent to the IPL. A few GABA positive (possibly ganglion) cells extended a single fibre toward the nerve fibre layer. Solitary GABA positive fibres were seen in this layer and in the optic nerve. Glycine immunoreactivity was observed in cells with the location typical of amacrine and bipolar (peripheral retina) cells, as well as in punctate structures of the IPL. In contrast to the GABA positive cells, the glycine positive cells were more frequent in the peripheral than in the central retina (42% and 23% of all cells in inner 1/3 of INL, respectively). A few cells in the GCL (0.5-1.5%) were glycine positive. Glutamate colocalized with GABA or glycine in a majority of the cells stained for either of these inhibitory transmitters (90-95% of the GABA positive cells, and 80-86% of the glycine positive cells, in the INL). Some bipolar cells were stained for both glutamate and glycine. Colocalization of GABA and glycine occurred in a subpopulation (3-4%) of presumed amacrine cells, about half of which was also glutamate positive.     

Differential expression of phospholipase D1 in the developing retina

Expression patterns of phospholipase D1 (PLD1) in the developing rat retina were investigated using immunocytochemistry and Western blot analysis and compared with the expression patterns of glutamine synthetase. PLD1 immunoreactivity appeared first in a few neuroblasts in the middle of the mantle zone of the primitive retina by embryonic (E) day 13. PLD1-immunoreactive primitive ganglion cells were characterized in the ganglion cell layer by E17. Faint immunoreactivity at E17 profiled radially orientated cells and this pattern appeared up to postnatal (P) day 7. In the ganglion cell layer at P3, displaced amacrine cells and ganglion cells were classified. At P5, presumptive horizontal cells and amacrine cells were identified. By P7, a thin outermost layer of newly formed segments of the photoreceptor cells was also PLD1 immunoreactive. PLD1 immunoreactivity at P8 was limited to radial Müller cells and the outer segment layer of the photoreceptor cells, and the expression pattern was conserved to adulthood. Western blot analysis showed relatively high amounts of PLD1 protein at E17 and P3, a decrease at P7, and moderate amounts from P8 onward. Co-expression of PLD1 with glutamine synthetase in the retina appeared first after birth in differentiating neurons and in Müller cells by P8; thereafter the pattern was maintained. The expression pattern of the PLD1 during development of the retina suggests that PLD1 plays important roles in glutamate-associated differentiation of both specific neurons and radial glial cells, and in glutamate-mediated cellular signalling in Müller cells.     

Dopamine D1-receptor immunolocalization in goldfish retina.

Dopamine, a neuromodulator in the vertebrate retina, is involved in numerous functions related to light adaptation. However, unlike in mammals, localization of retinal D1-dopamine receptors in nonmammalian vertebrates has been hampered due to a lack of antisera. To address this problem, an antiserum against the 18 C-terminal amino acids of the goldfish D1 receptor (gfD1r) was generated in chicken eggs and tested in retinae of goldfish and rat, and rat caudate putamen, by using immunoblots and light microscopic immunocytochemistry. No labeling was observed in any tissue or immunoblots with preabsorbed gfD1r antiserum. Immunoblot analysis of goldfish retina revealed a single band at about 101 kDa. The patterns of gfD1r immunoreactivity (gfD1r-IR), found in rat caudate putamen and rat retina were virtually identical to that previously reported with other D1-receptor ligands and antisera. In goldfish retina, gfD1r-IR was most intense over cell bodies in the ganglion cell layer, amacrine cells in the proximal inner nuclear layer (INL), and bipolar cells in the distal INL. Weaker gfD1r-IR was observed over horizontal cell bodies and both plexiform layers. Müller cells and axons of cone photoreceptors were labeled as well. Double labeling showed that all protein kinase C-immunoreactive bipolar cells (ON type) were gfD1r-IR on the soma, axon terminal, and dendrites. All glutamate decarboxylase-immunoreactive (i.e., gamma-aminobutyric acid utilizing) amacrine cells and horizontal cells were gfD1r-IR. Retinal D1r distribution is more extensive than dopamine neuron innervation, but is consistent with physiologic estimates of dopamine function, suggestive of both wiring and volume transmission of dopamine in the retina. The gfD1r antiserum displays cross-reactivity to dopamine receptors in a mammal and a nonmammal and should prove useful in future studies of dopaminergic systems.     

Immunocytochemical localization of kainate-selective glutamate receptor subunits GluR5, GluR6, and GluR7 in the cat retina

Localizations of the kainate-selective glutamate receptor subunits GluR5, 6, and 7 were studied in the cat retina by light and electron microscopic immunocytochemistry. GluR5 immunoreactivity was observed in the cell bodies and dendrites of numerous cone bipolar cells and ganglion cells. The labeled cone bipolar cells make basal or flat contacts with cone pedicles in the outer plexiform layer, leading to their identification as OFF-center bipolar cells. Reaction product within the inner plexiform layer was observed in processes of ganglion cells at their sites of input from cone bipolar cells. Staining for GluR6 was localized to A- and B-type horizontal cells, numerous amacrine cells, and an occasional cone bipolar cell. The larger ganglion cells were also immunoreactive. As with other GluR molecules, labeling was usually confined to one of the two postsynaptic elements at a cone bipolar dyad contact. Immunoreactivity for GluR7 was very limited and was seen only in a few amacrine and displaced amacrine cells. Findings of this study are consistent with a major role for kainate receptors in mediating OFF pathways in the outer retina with participation in both OFF and ON pathways in the inner retina.     

Distribution of muscarinic acetylcholine receptors on processes of isolated retinal cells

Binding of propylbenzilylcholine mustard, a muscarinic acetylcholine receptor antagonist, to isolated retinal cells was examined with light microscopic autoradiography. Dissociation of the adult tiger salamander retina yielded identifiable rod, cone, horizontal, bipolar, amacrine/ganglion, and Müller cells. Preservation of fine structure was assessed with conventional electron microscopy. For all cell types, the plasmalemma was intact and free of adhering debris; in addition, presynaptic ribbon complexes were present in photoreceptor and bipolar axon terminals indicating that synaptic structures were retained. Specific binding to cell bodies and processes was analyzed separately by using morphometric and statistical techniques. The highest grain densities occurred on processes of amacrine/ganglion cells and axons and 2 degrees and 3 degrees dendrites of bipolar neurons. Bipolar cells, however, seemed to be a heterogeneous population because there was great variation in the density of binding sites on both their axons and distal dendrites. Intermediate levels of binding were found on bipolar 1 degree dendrites and horizontal cells. No specific binding was detected on Müller cells and most parts of photoreceptors. Comparisons between cells showed that grain densities were similar for bipolar axons and amacrine/ganglion cell processes but bipolar dendrites were richer in binding sites than horizontal cell dendrites. Thus, muscarinic receptors in the salamander retina are located on amacrine/ganglion, bipolar, and horizontal cells and primarily confined to the processes which compose the two synaptic layers. In the inner plexiform layer, muscarinic receptors reside on processes from all three inner retinal neurons: in the outer synaptic layer, receptors are only on second-order cells and are more numerous along bipolar than horizontal cell dendrites.     

Identification and characterization of an aquaporin 1 immunoreactive amacrine-type cell of the mouse retina

Using immunocytochemistry, a type of amacrine cell that is immunoreactive for aquaporin 1 was identified in the mouse retina. AQP1 immunoreactivity was found in photoreceptor cells of the outer nuclear layer (ONL) and in a distinct type of amacrine cells of the inner nuclear layer (INL). AQP1-immunoreactive (IR) amacrine cell somata were located in the INL and their processes extended through strata 3 and 4 of the inner plexiform layer (IPL) with thin varicosities. The density of the AQP1-IR amacrine cells increased from 100/mm(2) in the peripheral retina to 350/mm(2) in the central retina. The AQP1-IR amacrine cells comprise 0.5% of the total amacrine cells. The AQP1-IR amacrine cell bodies formed a regular mosaic, which suggested that they represent a single type of amacrine cell. Double labeling with AQP1 and glycine, gamma-aminobutyric acid (GABA) or GAD(65) antiserum demonstrated that the AQP1-IR amacrine cells expressed GABA or GAD(65) but not glycine. Their synaptic input was primarily from other amacrine cell processes. They also received synaptic inputs from a few cone bipolar cells. The primary synaptic targets were ganglion cells, followed by other amacrine cells and cone bipolar cells. In addition, gap junctions between an AQP1-IR amacrine process and another amacrine process were rarely observed. In summary, a GABAergic amacrine cell type labeled by an antibody against AQP1 was identified in the mouse retina and was found to play a possible role in transferring a certain type of visual information from other amacrine or a few cone bipolar cells primarily to ganglion cells.     

Calbindin-D28k and calretinin as markers of retinal neurons in the anuran amphibian Rana perezi

In the present study we have analyzed the distribution of the calcium binding proteins calbindin-D28k (CB) and calretinin (CR) immunoreactive cells in the retina of the anuran Rana perezi using poly- and mono-clonal antibodies that were proven to be specific in the amphibian brain, without cross-reactivity. Double immunohistofluorescence techniques were used to demonstrate colocalization of both proteins in the same retinal cells. In addition, retrograde tracing experiments from the optic nerve were conducted to labeled ganglion cells and these were observed in combination with CB and/or CR immunohistochemistry. Cells containing CB were identified as all cones, scattered bipolar and amacrine cells together with cells in the ganglion cell layer. The pattern of CR immunoreactivity was strikingly different. Abundant cells contained CR in the inner retinal layers including horizontal, bipolar and amacrine cells, and cells in the ganglion cell layer. By means of double immunohistochemistry it was found that only subpopulations of amacrine cells and cells in the ganglion cell layer contained both CB and CR. Tracing from the optic nerve revealed retrogradely labeled ganglion cells with different morphologies and most of them contained CB and/or CR. All these data taken together suggest that in amphibians CB and CR are distinctly and precisely distributed in retinal neurons showing, however, peculiar features not observed previously in other vertebrates.     

Immunocytochemical localization of GABAA receptors in goldfish and chicken retinas

A monoclonal antibody (mAb 62-3G1) to the GABAA receptor/benzodiazepine receptor/Cl- channel complex from bovine brain was used with light and electron microscopy in goldfish retina and light microscopy in chicken retina to localize GABAA receptor immunoreactivity (GABAr-IR). GABAr-IR was found in the outer plexiform layer (OPL) in both species, in three broad bands in the inner plexiform layer (IPL) of goldfish, and in seven major bands of the chicken IPL. A small percentage of amacrine cell bodies (composing at least three types) were stained in chicken. In goldfish OPL, GABAr-IR was localized intracellularly and along the plasma membrane of cone pedicles, whereas rod spherules were lightly stained, but always only intracellularly. In chicken, all three sublayers of the OPL were GABAr-IR. The presence of GABAr-IR on photoreceptor terminals is consistent with data indicating feedback from GABAergic horizontal cells to cones. In the goldfish IPL, GABAr-IR was localized to postsynaptic sites of amacrine cell synapses; intracellular staining of processes in the IPL also was observed in presumed "GABAergic" targets. A comparison of GABAr-IR with the distributions of 3H-muscimol uptake/binding, glutamate decarboxylase-IR, GABA-IR, and 3H-GABA uptake in the IPL showed either a reasonable correspondence or mismatch, depending on the marker, species, and lamina within the IPL. The distribution of GABAr-IR in the retina corresponded better with the 3H-muscimol than with 3H-benzodiazepine binding patterns yet overall was in excellent agreement with many other physiological and anatomical indicators of GABAergic function. We suggest that intracellular GABAr-IR represents the biosynthetic and/or degradative pathway of the receptor and we conclude that mAb 62-3G1 is a valid marker of GABAA receptors in these retinas and will serve as a useful probe with which to address the issue of mismatches between the localization of GABAA receptors and indicators of presynaptic GABAergic terminals.     

Synaptic contacts of somatostatin-immunoreactive amacrine cells in goldfish retina

Retinal amacrine cells containing somatostatinlike immunoreactivity (SLI) were labeled by the peroxidase-antiperoxidase technique and their connections were analyzed by light and electron microscopy. The labeled processes were found in two distinct plexuses--one near the most proximal border of the inner plexiform layer and the other near the most distal border. They received most (89%) of their input from amacrine cells and the remainder from bipolar cells. A majority (56%) of their output synapses go to processes of amacrine cells, a substantial proportion (38%) go to ganglion cell dendrites, and the remainder go to bipolar cell axon terminals. The relative frequencies of each of the types of contacts were nearly identical in the distal and proximal plexuses. The cells containing SLI are different in their morphology and synaptic connections from any goldfish amacrine cells containing conventional neurotransmitters, but one type of amacrine cell containing SLI resembles certain other peptidergic amacrine cells in the goldfish retina.     

Suppression by zinc of transient OFF responses of carp amacrine cells to red light is mediated by GABA(A) receptors

Modulation by Zn(2+) of ON and OFF responses of transient amacrine cells driven by red- and green-sensitive cones was investigated in isolated, superfused carp retina, using intracellular recording techniques. Zn(2+) selectively abolished the OFF response to red flash of the transient amacrine cells. This Zn(2+) effect was mimicked by GABA application and was blocked by bicuculline, indicating the involvement of GABA(A) receptors. Such differential modulation was observable neither in bipolar cells nor in sustained OFF amacrine cells. It is suggested that the Zn(2+) effect reported here might be due to a direct action of Zn(2+) on GABA(A) receptors of the transient amacrine cells.