Subgroup of alpha ganglion cells in the adult cat retina is immunoreactive for somatostatin.

We have previously shown that two types of cells in the ganglion cell layer of the adult cat retina are immunoreactive for somatostatin (White et al., '90). One of the types was identified by morphological criteria as a wide-field amacrine cell. The other cell type had a large, angular soma that resembled the alpha ganglion cell, but evidence was not available to identify it definitively as a ganglion cell. Both cell types were distributed preferentially in the inferior retina. In this report, we demonstrate that the two types of cell are, indeed, displaced amacrine cells and alpha ganglion cells. First, when retrograde tracers were injected into central visual targets, the immunoreactive large cells but not the displaced amacrine cells were found to be labeled. Second, after unilateral section of the optic nerve, the immunoreactive large cells disappeared from the retina on the lesioned side, but the displaced amacrine cells occurred in the same numbers in both retinae. In the periphery, the large cells ranged in diameter from 33 to 47 microns, comparable only to alpha ganglion cells (Boycott and W?ssle, '74). An antiserum to parvalbumin was used to visualize the dendrites (R?hrenbeck and W?ssle, '88) of somatostatin-immunoreactive large cells. Based on dendritic stratification within the inner plexiform layer (Famiglietti and Kolb, '76), the somatostatin-immunoreactive large cells were found to include both on-center cells and off-center cells, but were predominantly of the off-center type. Within a local region, they were found to be arrayed with greater regularity than the overall population of alpha ganglion cells. These results indicate that alpha ganglion cells of the cat retina can be subdivided on the basis of their immunoreactive staining for somatostatin and suggest that the diversity of ganglion cells in the cat retina may be greater than has been recognized on the basis of morphological criteria alone.     

Morphology and distribution of neurons in the retina of the American garter snake Thamnophis sirtalis

It is confirmed that cone photoreceptors observed in flatmounts of the American garter snake Thamnophis sirtalis, retina correspond to the retinal mosaic viewed in the living eye (Land and Snyder, Vision Res. 11:105-114, '85). The garter snake has three major morphological types of cones; large single cones, small single cones, and double cones. The brightly reflecting components seen in the living eye are large single cones and principle cones of double cones, whereas irregularly spaced dark regions within this mosaiac mark the positions of small single cones. The "sparkle" of the retinal mosaic originates from the ellipsoid region of the cones where microdroplets of high refractive index are densely packed. Unlike conventional oil droplets, these microdroplets reside adjacent to mitochondrial cristae within the ellipsoid. However, the microdroplets may function collectively as a single large oil droplet to increase the angular sensitivity of the inner segments, thus reducing a potentially large Stiles-Crawford effect predicted for this geometrically small eye. The ganglion cell layer of the garter snake comprises two morphologically distinct populations of presumed neurons; classical neurons and microneurons. Density distribution maps for neurons in the ganglion cell layer and the photoreceptor layer reveal the presence of a putative area centralis and a horizontal visual streak. The topography of large cones parallels that of classical neurons. Small single cones have a more circular distribution, but also peak in density at the area centralis. The convergence of cones to classical neurons is lowest at the area centralis, 2.5:1, and highest, 4:1, at the retinal edge. With its interesting structural features, the garter snake retina provides helpful insight into different strategies in eye design.     

NADPH-diaphorase neurons in the retina of the hamster

NADPH-diaphorase-positive neurons have been demonstrated in the inner nuclear layer and ganglion cell layer of the retina of different mammalian species, but so far no experiments have been conducted to identify whether these cells are amacrine cells and/or retinal ganglion cells. We attempted to solve this problem by studying the NADPH-diaphorase-positive neurons in the hamster retina. From the NADPH-diaphorase histochemical reaction, two distinct types of neurons in the hamster retina were identified. They were named ND(g) and ND(i) cells. The ND(g) cells were cells with larger somata, ranging from 10 to 21 μm in diameter with a mean of 15.58 μm (S.D.= 2.59). They were found in the ganglion cell layer only. The ND(i) cells were smaller, with the somata ranging from 7 to 11 μm and having the mean diameter of 8.77 μm (S.D. = 1.24). Most of the ND(i) cells were found in the inner nuclear layer, and only very few could be observed in the inner plexiform layer. On average, there were 8,033 ND(g) and 5,051 ND(i) cells in the ganglion cell layer and inner nuclear layer, respectively. Two experiments were performed to clarify whether any of the NADPH-diaphorase neurons were retinal ganglion cells. Following unilateral optic nerve section, which leads to the retrograde degeneration of retinal ganglion cells, the numbers of both ND(g) and ND(i) cells did not change significantly for up to 4 months. In addition, when retinal ganglion cells were prelabeled retrogradely (horseradish peroxidase of flurescent microspheres) and retinas were then stained for NADPH diaphorase, no double-labeled neurons were detected. These results indicated that the NADPH-diaphorase neurons in the hamster retina were the amacrine cells in the inner nuclear layer and displaced amacrine cells in the ganglion cell layer. Dendrites of the ND(g) and ND(i) cells were found to stratify in sublaminae 1, 3, and 5 of the inner plexiform layer, with a prominent staining in the sublamina 5. The possible importance of this arrangement in the rod pathway is also discussed. © 1994 Wiley-Liss, Inc.     

Presynaptic and postsynaptic localization of GABAB receptors in neurons of the rat retina

The recently cloned GABA_B receptors were localized in rat retina using specific antisera. Immunolabelling was detected in the inner and outer plexiform layers (IPL, OPL), and in a number of cells in the inner nuclear layer and the ganglion cell layer. Double-labelling experiments for GABA (γ-aminobutyric acid) and GABA_B receptors, respectively, demonstrated a co-localization in horizontal cells and amacrine cells. Electron microscopy showed that GABA_B receptors of the OPL were localized presynaptically in horizontal cell processes invaginating into photoreceptor terminals. In the IPL, GABA_B receptors were present presynaptically in amacrine cells, as well as postsynaptically in amacrine and ganglion cells. The postnatal development of GABA_B receptors was also studied, and immunoreactivity was observed well before morphological and synaptic differentiation of retinal neurons. The present results suggest a presynaptic (autoreceptor) as well as postsynaptic role for GABA_B receptors. In addition, the extrasynaptic localization of GABA_B receptors could indicate a paracrine function of GABA in the retina.     

Ontogeny of catecholaminergic and cholinergic cell distributions in the cat's retina

The development of catecholaminergic and cholinergic neurones in the cat's retina has been examined with antibodies against their respective rate-limiting enzymes, tyrosine hydroxylase (TH) and choline acetyl transferase (ChAT). ChAT-immunoreactive (IR) cells were first detected at E (embryonic day) 56 with somata in the ganglion cell layer (GCL) or in the inner cytoblast layer (CBL). At P (postnatal day) 1, two faint bands of ChAT-IR fibres were evident in an inner and outer strata of the inner plexiform layer (IPL) and by P26, the bands were similar to those in the adult. TH immunoreactivity was first detected at E59 in either darkly labelled somata in the inner CBL with processes extending toward the IPL or in lightly labelled somata also located in CBL but with no processes. At P1, most TH-IR cells had prominently labelled dendrites and, by P8, most of the features of the adult cells were evident. Soma size gradients among TH-IR cells were first detected at P8, with cells in temporal retina being larger than those in nasal retina or at the area centralis. The smaller sizes of cells at the area centralis emerged after P26. The smaller sizes of ChAT-IR somata at the area centralis, by contrast, emerged between P8 and P26. The number of both TH-IR and ChAT-IR cells declined from the time they first appeared till adulthood. The decline was smaller among ChAT-IR cells (24%) than among TH-IR cells (68%). In distribution, the differential expansion of the retina appeared to be largely responsible for generating the final adult distribution of ChAT-IR cells. However, during late postnatal development (P26 to adulthood), the density of ChAT-IR cells in the periphery declined more than that of the ganglion cells, suggesting that some ChAT-IR cells may die in the periphery during this time. Prior to P26, the changes in the distribution of TH-IR cells were inconsistent with the pattern of retinal expansion. It is suggested that during this period, regional cell loss and cell addition may account for the changes in distribution of TH-IR cells. Later in development (P26 to adulthood), the changes in the density of TH-IR cells closely conformed to the differential expansion of the retina.     

Differential distribution of voltage-gated calcium channels in dopaminergic neurons of the rat retina

We studied by immunocytochemistry and Western blots the identity and cellular distribution of voltage-gated calcium channels within dopaminergic neurons of the rat retina. The aim was to associate particular calcium channel subtypes with known activities of the neuron (e.g., transmitter release from axon terminals). Five voltage-gated calcium channels were identified: alpha1A, alpha1B, alpha1E, alpha1F, and alpha1H. All of these, except the alpha1B subtype, were found within dopaminergic perikarya. The alpha1B channels were concentrated at axon terminal rings, together with alpha1A calcium channels. In contrast, alpha1H calcium channels were most abundant in the dendrites, and alpha1F calcium channels were restricted to the perikaryon. The alpha1E calcium channel was present at such a low density that its cellular distribution beyond the perikaryon could not be determined. Our findings are consistent with the available pharmacological data indicating that alpha1A and alpha1B calcium channels control the major fraction of dopamine release in the rat retina.     

Substance P-immunoreactive neurons in the human retina

Substance P (SP) is a neuropeptide that acts as a neurotransmitter or a neuromodulator in the retina. The aim of this study was to identify the type(s) and the distribution of the SP-immunoreactive (SP-IR) cells in the human retina. We have used an antiserum to SP to immunostain neurons in postmortem human retinae. Immunostained retinae were processedwith the avidin-biotin complex (ABC) to visualize the cells either whole mounted in glycerol or embedded in plastic. Some retinae were also sectioned at 20 μm in order to obtain radial views of stained cells. SP-IR amacrine cells stain intensely and appear to be of a single type in the human retina. They are large-field cells with large cell bodies (16 μm diameter) lying in normal or displaced positions on either side of the inner plexiform layer (IPL). Their sturdy, spiny, and appendage-bearing dendrites stratify in stratum 3 (S3) of the IPL, where many overlapping, fine dendrites intermingle to form a plexus of stained processes. Either cell bodies or primary dendrites emit an “axon-like” process that, typically, divides into two long, fine processes, which run in opposite directions for hundreds of micrometers in S5 and S3 before disappearing as distinct entities in the stained plexus in S3. Long, fine dendrites also pass from the dendritic plexus to run in S5 and down to the nerve fiber layer to end as large varicosities at blood vessel walls. In addition, fine processes are emitted from the dendritic plexus that runs in S1, and some pass up to the outer plexiform layer (OPL) to run therein for short distances. The SP-IR amacrine cell has many similaritiesto the thorny, type 2 amacrine cells described from Golgi studies. In addition to the SP-IR amacrine cells, a presumed ganglion cell type is faintly immunoreactive. Its 20–22 μpm cell body gives rise to a radiate, sparsely branched, widespreading dendritic tree running in S3. Its dendrites and cell body become enveloped by the more intensely SP-IR processes and boutons from the SP-IR amacrine cell type. The SP-IR ganglion cell type most resembles G21 from a Golgi study. © 1995 Wiley-Liss, Inc.     

Serotonin-like immunoreactivity in the adult and developing retina of the leopard frog Rana pipiens

Recent work in nonmammalian vertebrate retinas has suggested that other cell types besides the generally accepted amacrine cells may contain serotonin. We have used immunocytochemical methods to study serotonin-like immunoreactivity (5-HTLI) in the retina of the developing and mature frog Rana pipiens. In the adult, two types of serotonin immunoreactive (5-HT-ir) cells were found in the inner nuclear layer (INL) of the retina. Additionally, a large population of cells in the retinal ganglion cell layer (RGCL) had 5-HTLI. These cells were grouped into three types based on their soma size and their primary dendritic branching pattern. The optic nerve fiber layer was also intensely stained with serotonin antisera although staining intensity decreased progressively as the fibers approached the optic nerve head. Severing the optic nerve resulted in 5-HT-ir elements that extended up the optic nerve shaft from the lesion site toward the retina. Both regional and temporal changes in the pattern of 5-HTLI were seen. In middle regions of retina, approximately 30% of the cells in the RGCL were 5-HT-ir. Nasal and temporal regions of central retina had significantly fewer 5-HT-ir cells. Early in development only scattered cells in the RGCL were 5-HT-ir. As the animals matured there was an increase in both the proportion and the staining intensity of these cells. Our results suggest that in studying the function and development of the visual system in this animal, the role of serotonin must be examined. © 1993 Wiley-Liss, Inc.     

Spatial density and distribution of choline acetyltransferase immunoreactive cells in human, macaque, and baboon retinas.

Whole-mounted human, macaque, and baboon retinas were labelled with an antiserum to human choline acetyltransferase (ChAT), by the immunoperoxidase technique. Previous work in nonprimate species has shown that these cells correspond to the starburst amacrine cells. Labelled somata were disposed on either side of the inner plexiform layer, and their processes formed two narrow zones within it. In human retinas, the ratio of labelled somata in the ganglion cell layer (GCL) to those in the inner nuclear layer (nominal Sb/Sa ratio) was about 60/40 at all locations, similar to that found in nonprimate mammalian species. The density of labelled cells in the human GCL ranged from 1,000 to 1,150 mm-2 near the fovea to 300 to 400 mm-2 in the periphery. Labelling tended to be more erratic in macaque retinas. Nevertheless the Sb/Sa ratio was as high as 70/30 and spatial densities were similar to those of humans. The overlap factor in macaque retinas outside the nasal quadrant was about 10 at all retinal eccentricities, based upon dendritic-field sizes from a Golgi study. About each labelled soma there was a region 20 to 120 microns in diameter in which the probability of the occurrence of other labelled somata was lower than elsewhere. No such nonrandomness was found between labeled cells in the GCL and those in the amacrine cell layer. The packing factor was about 0.3 in well-labelled regions, independent of retinal position or spatial density. Published data on ChAT-labelled cells in rabbit and rat show a similar value. This invariance is consistent with the hypothesis that this nonrandomness is a residual consequence of somal contiguity at an early developmental stage.     

Ontogeny of somatostatin immunoreactivity in the cat retina.

In the ganglion cell layer of the adult cat retina, subgroups of displaced amacrine cells and alpha ganglion cells are immunoreactive for somatostatin or a somatostatinlike substance. Both types of immunoreactive cells are found preferentially in inferior retina. We studied the development of somatostatin immunoreactivity in the prenatal and postnatal cat retina to determine how such unusual distributions of immunoreactive cells arise. Somatostatin-immunoreactive profiles were first observed at embryonic day (E) 30, within the inner retina in a central region that included the optic disk and the area centralis. By E36, immunoreactivity had virtually disappeared from the central retina but was present throughout the periphery. The immunoreactive profiles could not be classified morphologically because of their immaturity but were most likely retinal ganglion cells, the earliest born cells of the inner retina. Of the two types of immunoreactive cell observed in the adult, the first to be recognized morphologically was the displaced amacrine cell, at E45. These cells were virtually adultlike in morphology and number by E51, two weeks before birth. In contrast, immunoreactive alpha ganglion cells were not apparent until five days after birth and did not achieve their mature numbers and immunoreactive staining characteristics until more than a month later. From the time they could initially be recognized, both immunoreactive displaced amacrine cells and alpha cells were distributed mainly in the inferior retina. A third type of somatostatin-immunoreactive cell was transiently observed in the superior and inferior retina during prenatal and early postnatal development. These cells were characterized by granular staining in irregular shapes and few, if any, faintly stained processes. Injections of retrograde tracers into retinorecipient targets revealed that many of these cells were retinal ganglion cells. They disappeared by postnatal day 38. Our results indicate that somatostatin immunoreactivity initially follows a central-to-peripheral pattern of development, as is typical of other developmental events in the mammalian retina. They also indicate that the two types of somatostatin-immunoreactive neurons present in the adult cat retina (displaced amacrine and alpha ganglion cells) attain their mature immunocytochemical properties with very different timecourses. Finally, the observation that somatostatin immunoreactivity appears transiently in the granular-staining ganglion cells, distributed throughout the superior and inferior retina, suggests that the peptide may play a regulatory role in the development of the retina and/or retinofugal pathways.     

GABA-like immunoreactive cells containing nicotinic acetylcholine receptors in the chick retina.

The possibility that GABA-like immunoreactive cells of the chick retina also contain neuronal nicotinic acetylcholine receptors was investigated by means of immunohistochemical techniques. Double-labeled cell bodies containing GABA-like immunoreactivity and nicotinic receptor-like immunoreactivity were seen in the inner third of the inner nuclear layer and were presumably amacrine cells. Approximately 29-36% of the GABA-positive cells in the inner nuclear layer contained nicotinic receptor immunoreactivity. Their soma sizes ranged from 5-12 microns. Some double-labeled cells ranging from 7-21 microns were observed in the ganglion cell layer as well. Between 9-37% of the GABA-positive cells in this layer contained nicotinic receptor-like immunoreactivity. Following injection of a retrograde tracer into the optic tectum, some of the retrogradely labeled cells were also double labeled with antibodies against GABA and nicotinic receptors. This indicates that at least some of the GABA-positive cells containing nicotinic acetylcholine receptors in the ganglion cell layer are indeed ganglion cells. The present data appear to represent the first demonstration of the presence of acetylcholine receptors in GABA-containing cells in the retina, thus providing a basis for a possible influence of acetylcholine upon those presumptive GABAergic cells.     

Pathway-dependent modulation by P2-purinoceptors in the mouse retina

Adenosine trisphosphate (ATP) activates purinoceptors and acts as a neurotransmitter in the nervous system. In the retina, we previously reported that the immunohistochemical distribution of the subset of P2-purinoceptors differs between the ON and OFF pathways. Here, we investigated whether ATP activates P2-purinoceptors and modulates the physiological function of the mouse retina. We also examined if signal processing by P2-purinoceptors is pathway specific. Results showed that ATP activated both ON- and OFF-cholinergic amacrine cells. However, responses in OFF-cholinergic amacrine cells were greater than those in ON-cholinergic amacrine cells. Pharmacological studies in OFF-cholinergic amacrine cells showed that the response of OFF-cholinergic amacrine cells is mediated P2X₂-purinoceptors. Further, ATP increased γ-aminobutyric acid (GABA)ergic inhibitory postsynaptic currents (IPSCs) in OFF- but not ON-cholinergic amacrine cells. The increase in GABAergic IPSCs was mediated by P2-purinoceptors. P2-purinoceptor-mediated signals suppressed OFF ganglion cells but activated ON ganglion cells. Our findings indicate that ATP physiologically modulates signal processing of the ON and OFF pathways in a pathway-specific manner through P2-purinoceptors.     

Colocalization of immunoreactive substance P and neurotensin in amacrine cells of the goldfish retina

Using an immunohistochemical double-label technique, neurotensin and substance P immunoreactivity were localized to amacrine cells in the goldfish retina. Both peptides were found in a single population of unistratified amacrine cells branching in sublamina 3 of the inner plexiform layer. A monostratified amacrine cell branching in sublamina 1 contained only substance P immunoreactivity and a bistratified cell branching in sublaminae 1 and 3 contained only neurotensin immunoreactivity.     

Synaptic input to the on-off directionally selective ganglion cell in the rabbit retina

A physiologically identified on-off directionally selective (DS) ganglion cell with its preferred-null axis defined was stained with horseradish peroxidase (HRP) and prepared for electron microscopy. A continuous series of thin sections were used to examine the cell's synaptology. Although the DS cell dendrite received the majority of its synaptic input from a heterogeneous population of amacrine cell processes, a frequently observed synaptic profile consisted of a DS cell dendrite receiving synapses from a cluster of several amacrine cell processes. These clusters of processes were assumed to be from a fascicle of amacrine cells, most of which probably belonged to several different cholinergic starburst amacrine cells. The most frequently observed presynaptic profile within the clusters consisted of a synaptic couplet in which two processes synapsed with each other before one of them finally synapsed with the DS ganglion cell dendrite; occasionally, a chain of three serial synapses was seen. In addition, a specific microcircuit that has the potential to exert lateral feedforward inhibition was also observed. This microcircuit consisted of two cone bipolar cell terminal dyad synapses where one dyad contained an amacrine cell process making a reciprocal synapse and a DS ganglion cell dendrite receiving direct excitation; the other dyad synapse, found lateral to the first dyad, contained two amacrine cell processes that both made reciprocal synapses, but one fed forward to make a putative inhibitory synapse with the DS cell dendrite.     

Morphology and mosaics of VIP-like immunoreactive neurons in the retina of the rhesus monkey.

Vasoactive intestinal peptide (VIP) is a 28-amino acid peptide that has been demonstrated to reside in cells ( = VIP+ cells) of the retinae of various vertebrate species. In an attempt to study the morphology and distribution of VIP+ cells in the retina of the rhesus monkey in more detail, we subjected VIP+ cells observed in cryostat sections or wholemounts rhesus monkey retinae to a quantitative analysis. VIP+ cells were found to reside in the innermost row of the inner nuclear layer (INL) and in the ganglion cell layer (GCL) in similar numbers (estimate: 50 cells/mm2 at 6-10 mm eccentricity each) and only on rare occasions (12% of all VIP+ cells) in varying positions within the inner plexiform layer (IPL). Somata of VIP+ cells were circular and had a mean diameter of 9.1 microns. They gave rise to 1-3 main dendrites, which were usually oriented toward the IPL. Main dendrites ramified widely into thin fibers (dendritic field diameter less than = 1 mm), carrying varicose swellings. The fibers that contributed to one and the same plexus of VIP+ fibers preferred the middle third of the IPL, independent of the positions of the parent somata. A quantitative analysis of nearest-neighbour distances in the retinal wholemount preparation suggested that VIP+ cells in the GCL and in the INL might be distributed according to 2 independent mosaics. A comparison with Golgi-stained material leads to the tentative equation of VIP+ cells with the "spiny" A12 amacrine cell of Mariani ('90). Whereas the low density and large dendritic field size of VIP+ cells might suggest a more widespread function, the varicose dendritic morphology seems to be more compatible with functionally independent dendritic subunits mediating localized effects.     

Distribution of immunoreactive substance P neurons in the hypothalamus and pituitary of the rhesus monkey

The distribution of immunoreactive substance P (IR-SP) neurons was examined in the hypothalamus and pituitary gland of the rhesus monkey by using the peroxidase-antiperoxidase immunocytochemical technique. Immunoreactive SP cell bodies were observed in the arcuate nucleus, in the region lateral to the arcuate nucleus, and in the median eminence (ME). Immunoreactive SP cells were also seen in the periventricular area of the dorsal tuberal region. A rich network of SP fibers was concentrated in the arcuate region, and the fiber stain was particularly dense in the external zone of the median eminence and in the external layer of the infundibular stalk. Also, substance P fibers were seen in the internal layer of the pituitary stalk and in the neural lobe of the pituitary gland. Outside the hypothalamus a dense network of IR-SP fibers was observed in the globus pallidus.     

Expression and function of the neuronal gap junction protein connexin 36 in developing mammalian retina

With the advent of transgenic mice, much has been learned about the expression and function of gap junctions. Previously, we reported that retinal ganglion cells in mice lacking the neuronal gap junction protein connexin 36 (Cx36) have nearly normal firing patterns at postnatal day 4 (P4) but many more asynchronous action potentials than wild-type mice at P10 (Torborg et al. [2005] Nat. Neurosci. 8:72-78). With the goal of understanding the origin of this increased activity in Cx36-/- mice, we used a transgenic mouse (Deans et al. [2001] Neuron 31:477-485) to characterize the developmental expression of a Cx36 reporter in the retina. We found that Cx36 was first detected weakly at P2 and gradually increased in expression until it reached an adult pattern at P14. Although the onset of expression varied by cell type, we identified Cx36 in the glycinergic AII amacrine cell, glutamatergic cone bipolar cell, and retinal ganglion cells (RGCs). In addition, we used calcium imaging and multielectrode array recording to characterize further the firing patterns in Cx36-/- mice. Both correlated and asynchronous action potentials in P10 Cx36-/- RGCs were significantly inhibited by bath application of an ionotropic glutamate receptor antagonist, indicating that the increase in activity was synaptically mediated. Hence, both the expression patterns and the physiology suggest an increasing role for Cx36-containing gap junctions in suppressing RGC firing between waves during postnatal retinal development.