GABA and GAD-like immunoreactivity in the primate retina

GABA immunoreactivity was studied and compared with GAD immunoreactivity in the retinae of baboon, cynomolgus monkey and man. The central and peripheral parts of the retinae were investigated separately in cynomolgus monkey and in man. The same kinds of structures were stained with both antisera. Cells with a position corresponding to amacrine cells were stained, as well as processes in the inner plexiform layer and some cells in the ganglion cell layer. The outer plexiform layer and some cells with the position and configuration of horizontal cells also appeared immunoreactive. Staining was also observed in bipolar-like cells, in man most clearly when using the GABA antiserum in sections from the central parts of the retina. It is possible that horizontal cells, as well as bipolar-like cells, may play a previously unsuspected role in GABAergic transmission in the primate retina.     

GABA neurones in retinas of different species and their postnatal development in situ and in culture in the rabbit retina

Summary The localisation of GABA immunoreactive neurones in retinas of a variety of animals was examined. Immunoreactivity was associated with specific populations of amacrine neurones in all species examined, viz. rat, rabbit, goldfish, frog, pigeon and guinea-pig. All species, with the exception of the frog, possessed immunoreactive perikarya in their retinal ganglion cell layers. These perikarya are probably displaced amacrine cells because GABA immunoreactivity was absent from the optic nerves and destruction of the rat optic nerve did not result in degeneration of these cells. GABA immunoreactivity was also associated with the outer plexiform layers of all the retinas studied; these processes are derived from GABA-positive horizontal cells in rat, rabbit, frog, pigeon and goldfish retinas, from bipolar-like cells in the frog, and probably from interplexiform cells in the guinea-pig retina.The development of GABA-positive neurones in the rabbit retina was also analysed. Immunoreactivity was clearly associated with subpopulations of amacrine and horizontal cells on the second postnatal day. The immunoreactivity at this stage is strong, and fairly well developed processes are apparent. The intensity of the immunoreactivity increases with development in the case of the amacrine cells. The immunoreactive neurones appear fully developed at about the 8th postnatal day, although the immunoreactivity in the inner plexiform layer becomes more dispersed as development proceeds. The immunoreactive horizontal cells become less apparent as development proceeds, but they can still be seen in the adult retina.The GABA immunoreactive cells in rabbit retinas can be maintained in culture. Cultures of retinal cells derived from 2-day-old animals can be maintained for up to 20 days and show the presence of GABA-positive cells at all stages. In one-day-old cultures the GABA immunoreactive cells lacked processes but within three days had clearly defined processes. After maintenance for 10 days a meshwork of GABA-positive fibres could also be seen in the cultures.     

Somatostatin and VIP neurons in the retina of different species

Neurons displaying somatostatin or vasoactive intestinal polypeptide (VIP) immunoreactivity were detected among the amacrine cells in the retina of baboon, cynomolgus monkey, squirrel monkey, cow, pig, cat, rabbit, guinea-pig, rat, mouse, frog and goldfish. Generally, immunoreactive cell bodies were located in the inner nuclear layer with processes ramifying in three more or less well-defined sublayers in the inner plexiform layer. The density of the sublayers and their location varied with the peptide and species investigated. In most cases there was a sublayer in the outermost part (Ramon y Cajal's sublamina 1) of the inner plexiform layer and this sublayer was usually the best developed. In some species a few somatostatin fibres were also detected in the outer plexiform layer, suggesting that some interplexiform cells contain somatostatin. In the baboon VIP was found exclusively in interstitial amacrine cells which have their cell bodies and processes entirely within the inner plexiform layer.     

Somatostatin and VIP neurons in the retina of different species

Summary Neurons displaying somatostatin or vasoactive intestinal polypeptide (VIP) immunoreactivity were detected among the amacrine cells in the retina of baboon, cynomolgus monkey, squirrel monkey, cow, pig, cat, rabbit, guinea-pig, rat, mouse, frog and goldfish. Generally, immunoreactive cell bodies were located in the inner nuclear layer with processes ramifying in three more or less well-defined sublayers in the inner plexiform layer. The density of the sublayers and their location varied with the peptide and species investigated. In most cases there was a sublayer in the outermost part (Ramon y Cajal's sublamina 1) of the inner plexiform layer and this sublayer was usually the best developed. In some species a few somatostatin fibres were also detected in the outer plexiform layer, suggesting that some interplexiform cells contain somatostatin. In the baboon VIP was found exclusively in interstitial amacrine cells which have their cell bodies and processes entirely within the inner plexiform layer.     

Neuropeptide Y (NPY) immunoreactive neurons in the retina of different species

Neurons displaying Neuropeptide Y (NPY) immunoreactivity were found among amacrine cells in the retina of baboon, pig, cat, pigeon, chicken, frog, trout, carp and goldfish. The immunoreactive cell bodies were located in the middle and the innermost cell rows of the inner nuclear layer with processes forming one, two or three more or less well-defined sublayers in the inner plexiform layer. The location and the density of the sublayers varied with the species investigated. In the frog retina, bipolar-like cell bodies were found in the middle of the inner nuclear layer as well as sparsely occurring ovoid cell bodies in the ganglion cell layer. Like the amacrine cells, these cells emitted processes ramifying in three sublayers in the inner plexiform layer.     

Some horizontal cells of the bovine retina receive input synapses in the inner plexiform layer

Bovine retinae were stained immunocytochemically with antibodies against the calcium-binding protein, calbindin. Horizontal cells in the outer plexiform layer were heavily labelled. The processes of most horizontal cells were confined to the level of the outer plexiform layer, and the tips of their dendrites were positioned as the lateral elements of the cone triads, viz. the usual mammalian arrangement. However, some of the horizontal cells had additional thick processes descending to branch within the inner plexiform layer, where they were postsynaptic at bipolar cell dyads and where they also received input from amacrine cells. No output synapses of horizontal cells were observed in the inner plexiform layer.     

IMMUNOHISTOCHEMICAL LOCALIZATION OF GABA IN CHAMELEON RETINA (CHAMALEO CHAMALEO)

We used a policlonal antiserum against GABA and demonstated GABA-immunoreactivity (GABA-IR) in several populations of amacrine cells in the inner nuclear layer (INL), and other cells in the inner plexiform layer (IPL) of the central and peripheral retina of the chameleon. Horizontal cells do not contain GABA-IR and the chameleon retina is therefore an exception among non-mammals. GABA-IR was not seen in cell bodies in the position of photoreceptor, bipolar and interplexiform cells suggesting that GABA is not involved in synaptic transmission in the outer plexiform layer of chameleon retina.     

Cellular and subcellular localization of heme oxygenase-2 in monkey retina

Carbon monoxide (CO), an activator of soluble guanylate cyclase (SGC) and generated enzymatically by heme oxygenases (HO), is considered to function as an intra- and intercellular neuromodulator or neurotransmitter in the central and peripheral nervous systems. HO-2 is the constitutive isoform of HO and is more prevalent in nervous tissues than in the other peripheral tissues. Because previous studies have demonstrated different distributions of HO-2 in the retina depending on the species of animals, the aim of this study was to identify which cell types of the monkey retina express HO-2. The expression of HO-2 protein was examined in monkey retina by Western blot analysis. Immunoblottings from monkey homogenates revealed a single clear protein band with a molecular mass of 36 kDa that is corresponding to rat HO-2. Immunoreactivity of HO-2 was found in the perikarya of ganglion cells. Density of immunoreactive ganglion cells was higher in the central area of retina than in the peripheral retina, and somata of larger ganglion cells were stained more densely than smaller ones. In electron microscopy, immunoreactivity of HO-2 was localized on the membrane of the endoplasmic reticulum and the nuclear outer membrane of the ganglion cells. By contrast, inner plexiform layer, inner nuclear layer and outer nuclear layer were devoid of HO-2 immunoreactivity. cGMP were strongly localized in all of ganglion cells. Some cells contributed to the relatively faint cGMP staining were seen in the inner nuclear layer. In combination of HO-2 and cGMP immunocytochemistry, the overlap of co-localization of HO-2 and cGMP would suggest that HO-2 in the ganglion cells would serve as a source for CO generation and CO could serve as a gaseous signaling molecule modulator of neural activity in the retina of monkey.     

Somatostatin-like immunoreactivity in the retinae of adult and embryonic chickens

Somatostatin, a tetradecapeptide that inhibits growth hormone release, has a widespread distribution in the central and peripheral nervous systems and other cell types. In the present investigation, the chicken neural retina was studied for the presence of structures exhibiting somatostatin-like immunoreactivity by utilizing an indirect immunofluorescence technique. Controls for specificity of staining were performed on alternate sections. Several types of distinctly labeled neurons and their processes were evident in sections of adult and late embryonic retinae. Cresyl violet staining showed that these neurons, which were scattered peripherally and more numerous centrally, occupied several strata within the inner nuclear, inner plexiform, and ganglion cell layers. Labeled neurites of immunoreactive perikarya coursed within these layers as well, often approaching other immunoreactive cells and fibers. The morphology and position of the somatostatin-containing neurons indicated that these neurons were amacrine, horizontal, or ganglion associational cells. These findings indicate that somatostatin is first detectable in the retina during the late embryonic stages of the chicken.     

Differential expression of syntaxin-1 and synaptophysin in the developing and adult human retina

Synaptophysin and syntaxin-1 are membrane proteins that associate with synaptic vesicles and presynaptic active zones at nerve endings, respectively. The former is known to be a good marker of synaptogenesis; this aspect, however, is not clear with syntaxin-1. In this study, the expression of both proteins was examined in the developing human retina and compared with their distribution in postnatal to adult retinas, by immunohistochemistry. In the inner plexiform layer, both were expressed simultaneously at 11–12 weeks of gestation, when synaptogenesis reportedly begins in the central retina. In the outer plexiform layer, however, the immunoreactivities were prominent by 16 weeks of gestation. Their expression in both plexiform layers followed a centre-to-periphery gradient. The immunoreactivities for both proteins were found in the immature photoreceptor, amacrine and ganglion cells; however, synaptophysin was differentially localized in bipolar cells and their axons, and syntaxin was present in some horizontal cells. In postnatal-to-adult retinas, synaptophysin immunoreactivity was prominent in photoreceptor terminals lying in the outer plexiform layer; on the contrary, syntaxin-1 was present in a thin immunoreactive band in this layer. In the inner plexiform layer, however, both were homogeneously distributed. Our study suggests that (i) syntaxin-1 appears in parallel with synapse formation; (ii) synaptogenesis in the human retina might follow a centre-to-periphery gradient; (iii) syntaxin-1 is likely to be absent from ribbon synapses of the outer plexiform layer, but may occur at presynaptic terminals of photoreceptor and horizontal cells, as is apparent from its localization in these cells, which is hitherto unreported for any vertebrate retina.     

Local Differences in GABA Release Induced by Excitatory Amino Acids During Retina Development: Selective Activation of NMDA Receptors by Aspartate in the Inner Retina

Glutamate and GABA are the major excitatory and inhibitory neurotransmitters in the CNS. In the retina, it has been shown that glutamate and aspartate and their agonists kainate and NMDA promote the release of GABA. In the chick retina, at embryonic day 14 (E14), glutamate and kainate were able to induce the release of GABA from amacrine and horizontal cells as detected by GABA-immunoreactivity. NMDA also induced GABA release restricted to amacrine cell population and its projections to the inner plexiform layer (E14 and E18). Although aspartate reduced GABA immunoreactivity, specifically in amacrine cells of E18 retinas, it was not efficient to promote GABA release from retinas at E14. As observed in differentiated retinas, dopamine inhibited the GABA release promoted by NMDA and aspartate but not by kainate. Our data show that different retinal sites respond to distinct EAAs via different receptor systems.     

Immunohistochemical and quantitative analysis of 5-hydroxytryptamine in the retina of some vertebrates

5-Hydroxytryptamine immunoreactive neurons were found in retinae from chicken, pigeon, frog and goldfish. They were localized among the amacrine cells with a distribution of cell bodies and nerve fibres that varied with the species. In chicken and pigeon, bipolar-like cell bodies were also found in the middle of the inner nuclear layer, sending processes inwards to the inner plexiform layer and outwards to the horizontal cells. The signalling direction of these cells is doubtful. No 5-hydroxytryptamine immunoreactivity was found in retinae from cow, pig, cat, rabbit, guinea-pig, rat or mouse.Quantitative analyses were performed with HPLC on extracts from chicken, pigeon, frog and goldfish retinae. High concentrations were found in goldfish and frog whereas less, about 100 ng/g, was observed in chicken and pigeon.The results suggest that 5-hydroxytryptamine is the transmitter of a set of amacrine cells in cold-blooded vertebrates and perhaps also in birds. The transmitter of the indoleamine accumulating neurons of mammals remains to be further elucidated.     

The structure and development of the rat retina: an immunofluorescence microscopical study using antibodies specific for intermediate filament proteins

Rat retina structure was studied between embryonic day 14 and adult with antibodies specific for vimentin, glial fibrillary acidic protein (GFA) and the proteins of the neurofilament triplet. Vimentin could be detected in radial processes throughout the retina at all stages studied. These processes are believed to correspond, in the developing retina, to ventriculocytes, and in the mature retina to Müller cells. They could not normally be stained with any of the other intermediate filament antibodies employed here. We did find, however, that some older albino rats possessed GFA staining in addition to vimentin in these processes. Since we never saw such staining in the retinae of mature non-albino rats, and the retinae of older albino rats often showed signs of degeneration, we concluded that such GFA expression was most likely pathological. Neurofilament protein-positive processes were first detectable at embryonic day 15 1/2 in the inner regions of the retina, and corresponded to the axons of retinal ganglion cells. Such processes were equivalently displayed with antibodies to 68 K and 145 K protein, but were negative with 200 K protein. Some 68 K and 145 K positive fibers could also be decorated with vimentin antibody at this stage, though at later stages this was not the case. At later development stages more 68 K and 145 K neurofilament positive processes appeared, and after the first post-natal week progressively more of such processes became in addition 200 K positive, so that almost all neurofilament positive fibers in the adult stained for all three proteins. Such fibers, in the mature retina corresponded to 68 K and 145 K positive optic nerve fibers, and the processes of neurones in the inner plexiform layer. All fibers in the mature optic nerve fiber layer, but not all of those in the inner plexiform layer were stainable with 200 K antibodies. At 4 days post-natal we were able to detect 68 K and 145 K protein positive profiles in the outer regions of the developing retina, the prospective outer plexiform layer. Such profiles were always in addition vimentin positive, but negative for 200 K protein. During further development such profiles became ordered into a well defined layer and from about post-natal day 13 all of them began to acquire 200 K protein. They could be identified as the processes of horizontal cells. They continued to express vimentin in addition to the three triplet proteins in the adult, a so far unprecedented situation. We were able to detect neurofilament staining in the mature retina only in the above described regions, the inner and outer nuclear layer and the photoreceptor processes being completely free of staining. GFA was first detected in short processes adjacent to the inner limiting membrane which penetrated the optic nerve fiber layer. Such profiles were first detectable in the eye of the newborn animal, and were invariably identically stainable with vimentin at this age. These profiles could be stained with both vimentin and GFA at all later stages examined, although GFA staining became very much stronger than vimentin staining in some profiles in the adult. The results presented here are discussed in terms of development of the different retinal cell types.     

Neuron-specific enolase-like immunoreactivity in the vertebrate retina: selective labelling of Müller cells in Anura.

Neuron-specific enolase (NSE) immunocytochemistry was carried out in retinae of goldfish, axolotl, clawed frog, cane toad, lizard, chick, guinea-pig, rabbit, rat, cat and human. With the exception of Anura, strong immunoreactivity was seen in the large ganglion, amacrine cells and horizontal cells of the retina in all of the other species. Photoreceptors were found to be labelled in the rat and human retina and only one cone type in rabbit. Photoreceptor pedicles and ellipsoids were stained in the goldfish and the somata and inner segments of some photoreceptors in axolotl. In the axolotl retina, besides neurons, Müller cells (MCs) were also immunolabelled. In the retina of the cane toad and the clawed frog MCs were the only stained elements. Similarly in other parts of the central nervous system of the cane toad, glial elements of the optic tectum and spinal cord were immunoreactive. In contrast, in the peripheral nervous system, neurons of the 1st sympathetic ganglion and the 2nd dorsal root ganglion were labelled. In double-labelling experiments, glial fibrillary acidic protein and NSE showed colocalisation both in the glial elements of the optic tectum and spinal cord and in MCs of the retina of the cane toad.     

GABA release induced by aspartate-mediated activation of NMDA receptors is modulated by dopamine in a selective subpopulation of amacrine cells

Glutamate and GABA are the major excitatory and inhibitory neurotransmitters in the CNS, including the retina. In the chick retina, GABA is located in horizontal and amacrine cells and in some cells in the ganglion cell layer. It has been shown that glutamate and its agonists, NMDA, kainate, and aspartate, promote the release of GABA from isolated retina and from cultured retinal cells. Dopamine, the major catecholamine in the retina, inhibits the induction of GABA release by NMDA. Two to seven-day-old intact chicken retinas were stimulated with different glutamatergic agonists and the GABA remaining in the tissue was detected by immunohistochemical procedures. The exposure of retinas to 100 μ M NMDA for 30 minutes resulted in 50% reduction in the number of GABA-immunoreactive amacrine cells. Aspartate (100 μ M) treatment also resulted in 60% decrease in the number of GABA-immunoreactive amacrine cells. The number of GABA-immunoreactive horizontal cells was not affected by either NMDA or aspartate. In addition, dopamine reversed by 50% the reduction of the number of GABA-immunoreactive amacrine cells exposed to NMDA or aspartate. Kainate stimulation promoted a 50% reduction in the number of both GABA-immunoreactive amacrine and horizontal cells. Dopamine did not interfere with the kainate effect. While in control and in non-stimulated retinas a continuous and homogeneous immunolabeling was observed throughout the inner plexiform layer, retinas exposed to NMDA, kainate and aspartate displayed only a faint punctate labeling in the inner plexiform layer. It is concluded that, under our experimental conditions, both NMDA and aspartate induce the release of GABA exclusively from amacrine cells, and that the release is modulated by dopamine. On the other hand, kainate stimulates GABA release from both amacrine and horizontal cells with no interference of dopamine.     

Target cells of vitamin D in the vertebrate retina

Using PAP technique, cellular localization of vitamin D-dependent calcium-binding protein (D-CaBP) was investigated in vertebrate retina with monospecific antisera against chick duodenal D-CaBP. In the chick retina, the receptor cells were positive. In the inner nuclear layer, horizontal cells and some bipolar cells were also positive. Some amacrine cells as well as different levels of the inner plexiform layer were also positive for D-CaBP. A few interspersed ganglion cells were positive but their axons forming the optic tract were negative. Müller's cells were negative. In 1-day-old chicks and 4-week-old rachitic chicks there was paucity and absence, respectively, of D-CaBP staining in horizontal cells. In the mouse, rat, and rabbit the receptors had only trace amounts of reaction product in their outer segment and pedicle. Horizontal cells were densely positive throughout their cellular body and processes. Some amacrine cells in the inner nuclear layer were positive. In the mouse and rat three horizontal levels of the outer plexiform layer were very prominent because of their dense staining for D-CaBP. Many ganglion cells were also positive along with their axons forming the optic nerve. In the rabbit, no positive layers were seen in the inner plexiform layer, and ganglion cells with their fibers were negative. In the frog retina there were smaller amounts of D-CaBP in the receptor cells and horizontal cells than that of the chick retina. Also, the fibers of the ganglionic cells were positive for D-CaBP. In all species studied, some amacrine cells were stained for D-CaBP. Because of its possible roles in membrane calcium transport and intracellular Ca++ regulation, it has perhaps similar functions in these positive cells. The synthesis of D-CaBP is dependent upon vitamin D. These positive cells are thus target cells of vitamin D.     

Double-labeling techniques demonstrate that rod bipolar cells are under GABAergic control in the inner plexiform layer of the rat retina

The synaptic connectivity between rod bipolar cells and GABAergic neurons in the inner plexiform layer (IPL) of the rat retina was studied using two immunocytochemical markers. Rod bipolar cells were stained with an antibody specific for protein kinase C (PKC, α isoenzyme), and GABAergic neurons were stained with an antiserum specific for glutamic-acid decarboxylase (GAD). Some amacrine cells were also labeled with the anti-PKC antiserum. All PKC-labeled amacrine cells examined showed GABA immunoreactivity, indicating that PKC-labeled amacrine cells constitute a subpopulation of GABAergic amacrine cells in the rat retina. A total of 150 ribbon synapses established by rod bipolar cells were observed in the IPL. One member of the postsynaptic dyads was always an unlabeled AII amacrine cell process, and the other belonged to an amacrine-cell process showing GAD immunoreactivity. The majority (n=92) (61.3%) of these processes made reciprocal synapses back to the axon terminals of rod bipolar cells. In addition, 78 conventional synapses onto rod bipolar axons were observed, and among them 52 (66.7%) were GAD-immunoreactive. Thus GABA provides the major inhibitory input to rod bipolar cells.     

Neuron-specific enolase-like immunoreactivity in the vertebrate retina: selective labelling of Müller cells in Anura

Summary Neuron-specific enolase (NSE) immunocytochemistry was carried out in retinae of goldfish, axolotl, clawed frog, cane toad, lizard, chick, guinea-pig, rabbit, rat, cat and human. With the exception of Anura, strong immunoreactivity was seen in the large ganglion, amacrine cells and horizontal cells of the retina in all of the other species. Photoreceptors were found to be labelled in the rat and human retina and only one cone type in rabbit. Photoreceptor pedicles and ellipsoids were stained in the goldfish and the somata and inner segments of some photoreceptors in axolotl. In the axolotl retina, besides neurons, Müller cells (MCs) were also immunolabelled. In the retina of the cane toad and the clawed frog MCs were the only stained elements. Similarly in other parts of the central nervous system of the cane toad, glial elements of the optic tectum and spinal cord were immunoreactive. In contrast, in the peripheral nervous system, neurons of the 1st sympathetic ganglion and the 2nd dorsal root ganglion were labelled. In double-labelling experiments, glial fibrillary acidic protein and NSE showed colocalisation both in the glial elements of the optic tectum and spinal cord and in MCs of the retina of the cane toad.On leave of absence from Department of Zoology, Attila József University, Szeged, Hungary     

Immunocytochemical and biochemical localization of parvalbumin in the retina

Summary The cellular distribution of parvalbumin-like immunoreactivity (PA-LI) in normal retina of rat, monkey, and human was investigated by immunohistochemical peroxidase antiperoxidase methods, and the levels of PA-LI in normal rat retina and brain were measured by radioimmunoassay. The antibody, raised in rabbits using rat skeletal muscle parvalbumin, did not cross-react with other Ca²⁺-binding proteins such as calmodulin or S-100 proteins. In rat retina, PA-LI-containing cells are located in the proximal inner nuclear layer and send processes to the external half of the internal plexiform layer, suggesting that they are amacrine cells. In monkey and human retina, PA-LI positive cells exist in the outermost sublayer of inner nuclear layer, and PA-LI-containing fibers that extend horizontally are found in the internal zone of outer plexiform layer. The radioimmunoassay revealed that the rat retina contained 1710±91 ng PA-LI/mg protein, the levels of which were higher than that of brain (881±165 ng PA-LI/mg protein). These results show an additional location for PA-LI outside skeletal muscle and brain, and also provide information on the function of interneurons of retina, which are still poorly understood.     

Electron microscopic observation of pituitary adenylate cyclase-activating polypeptide (PACAP)-containing neurons in the rat retina

The distribution and localization of pituitary adenylate cyclase-activating polypeptide (PACAP) in the rat retina were studied by immunocytochemistry with both light and electron microscopy. PACAP-like immunoreactivity (PACAP-LI) was detected in the amacrine and horizontal cells as well as in the inner plexiform layer, the ganglion cell layer and the nerve fiber layer. PACAP-LI seemed to be concentrated predominantly in the neuronal perikarya and their processes, but not in other cells in the retina. At the ultrastructural level, PACAP-LI was visible in the plasma membranes, rough endoplasmic reticulum, and cytoplasmic matrix in the PACAP-positive neurons in the inner nuclear layer. In the inner plexiform layer, PACAP-positive amacrine cell processes made synaptic contact with immunonegative amacrine cell processes, bipolar cell processes, and ganglion cell terminals. These findings suggest that PACAP may function as a neurotransmitter and/or neuromodulator.