Rod vision: pathways and processing in the mammalian retina.

Bipolar cells in the mammalian retina are postsynaptic to either rod or cone photoreceptors, thereby segregating their respective signals into parallel vertical streams. In contrast to the cone pathways, only one type of rod bipolar cell exists, apparently limiting the routes available for the propagation of rod signals. However, due to numerous interactions between the rod and cone circuitry, there is now strong evidence for the existence of up to three different pathways for the transmission of scotopic visual information. Here we survey work over the last decade or so that have defined the structure and function of the interneurons subserving the rod pathways in the mammalian retina. We have focused on: (1) the synaptic ultrastructure of the interneurons; (2) their light-evoked physiologies; (3) localization of specific transmitter receptor subtypes; (4) plasticity of gap junctions related to changes in adaptational state; and (5) the functional implications of the existence of multiple rod pathways. Special emphasis has been placed on defining the circuits underlying the different response components of the AII amacrine cell, a central element in the transmission of scotopic signals.

Rabbit cone bipolar cells: correlation of their morphologies with whole-cell recordings.

The superfused retinal slice preparation was used to examine the morphology and glutamate-activated whole-cell currents of rabbit bipolar cells. There were six morphologically distinct types of cone bipolar cells and a rod bipolar cell that had axon terminals stratifying in stratum 3 to 5 of sublamina-b. All of these bipolar cell types exhibited an outward current in response to the application of the metabotropic glutamate receptor, mGluR6, agonist AP-4 (APB), and had I/V curves indicative of membrane channel closure. Conversely, there were no currents activated during the application of kainate, the AMPA/kainate receptor agonist. These data demonstrate they were on-bipolar cells. In addition, there were six morphologically distinct cone bipolar cells that stratified in sublamina-a. Every cell with axonal arborizations in stratum 1 and 2 exhibited an inward current when the ionotropic glutamate receptor agonist kainate was applied. This current was blocked by application of the AMPA/kainate receptor antagonist CNQX. These cells also decreased their membrane resistance in response to kainate, a characteristic of the opening of channels within the plasma membrane. Without exception, no cells stratifying in sublamina-a responded to the mGluR6 agonist AP-4, further identifying them as off-bipolar cells.

A flattened retina-eyecup preparation suitable for electrophysiological studies of neurons visualized with trans-scleral infrared illumination.

We present an in vitro flattened retinal-scleral preparation suitable for electrophysiological studies from visually targeted amacrine and ganglion cells of the rabbit retina. In a newly designed superfusion chamber, the retinal-scleral tissue is stained with Azure B allowing for imaging of neurons in the ganglion cell layer with an infrared (IR)-sensitive CCD camera via trans-scleral IR illumination. Neurons can be visually identified and targeted for both extracellular and intracellular recordings made singly or in simultaneous pairs. The quality and stability of the recordings are excellent and the tissue remains viable for up to 10 h. This relatively simple preparation avoids the extensive surgical manipulations inherent to those based on isolated retinas or retinal slices. Moreover, the use of trans-scleral IR illumination rather than fluorescent dyes to visualize and target neurons allows for electrophysiological studies of the retina under controlled adaptational states including dark-adapted conditions.

GABA responses of rod bipolar cells in rabbit retinal slices.

GABAergic responses of rabbit rod bipolar cells were reexamined by using whole-cell recordings in the superfused slice preparation to determine if there is GABA(C) receptor input to their axon terminal and to characterize the contribution that GABA(A) and GABA(C) receptors make to the total GABA current on the axon terminals of these cells. Pharmacological agents specifically blocking GABA(A) and GABA(C) receptor currents demonstrated that 37% of the GABA-activated current was blocked by either the GABA(A) antagonists bicuculline or SR-95531, whereas the remaining 63% of the GABA current was blocked by a mixture of bicuculline and the GABA(C) antagonist TPMPA. This indicated that GABA(C) receptors were present on the axon terminal of the rabbit rod bipolar cell and that they were responsible for mediating the bicuculline insensitive GABA current.

The shapes and numbers of amacrine cells: matching of photofilled with Golgi-stained cells in the rabbit retina and comparison with other mammalian species.

Amacrine cells of the rabbit retina were studied by "photofilling" a photochemical method in which a fluorescent product is created within an individual cell by focal irradiation of the nucleus; and by Golgi impregnation. The photofilling method is quantitative, allowing an estimate of the frequency of the cells. The Golgi method shows their morphology in better detail. The photofilled sample consisted of 261 cells that were imaged digitally in through-focus series from a previous study (MacNeil and Masland [1998] Neuron 20:971-982). The Golgi material consisted of 49 retinas that were stained as wholemounts. Eleven of these subsequently were cut in vertical section. Of the many hundreds of cells stained, digital through-focus series were recorded for 208 of the Golgi-impregnated cells. The two methods were found to confirm one another: Most cells revealed by photofilling were recognized easily by Golgi staining, and vice versa. The greater resolution of the Golgi method allowed a more precise description of the cells and several types of amacrine cell were redefined. Two new types were identified. The two methods, taken together, provide an essentially complete accounting of the populations of amacrine cells present in the rabbit retina. Many of them correspond to amacrine cells that have been described in other mammalian species, and these homologies are reviewed.

Rabbit retinal Müller cells undergo antigenic changes in response to experimentally induced proliferative vitreoretinopathy.

Experimental proliferative vitreoretinopathy (PVR) was induced in the rabbit eye by injecting mitotically active Müller cells into the vitreal chamber. Two weeks after the initiation of PVR, the retina and the epiretinal membrane that formed were examined to ascertain the antigenic expression of Müller cells in the retina and in the epiretinal membrane. Examination of various regions of the retina from the experimental PVR eye demonstrated that vimentin, glial fibrillary acidic protein (GFAP), cellular retinaldehyde binding protein (CRALBP), and beta-amyloid precursor protein (beta-APP), which were present in the Müller cells of the retina from the control eye, increased their expression, while the antigenicity of glutamine synthetase (GS), did not change; these proteins were also present in the cells contained within the experimentally induced epiretinal membrane. Alpha smooth muscle actin (alpha-SMA), a cytoskeletal protein that is associated with migration and tractional forces in many cell types, was not only present in the cells embedded within the epiretinal membrane, but was also present in the Müller cells underlying the epiretinal membrane. However, Müller cells that were in the inferior portion of the retina, where epiretinal membrane pathology was absent, did not express alpha-SMA. Although this protein is not normally found in Müller cells, they do express it de novo when they are maintained in culture. This suggests that a localized mechanism associated with epiretinal membrane formation induces the expression of alpha-SMA in Müller cells while the increased expression of GFAP, beta-APP, vimentin, and CRALBP are probably regulated via a more general mechanism.

Antigenic changes of rabbit retinal Müller cells in culture.

PURPOSE: To determine whether dissociated and cultured Müller cells from the avascular rabbit retina undergo the same phenotypic changes as Müller cells that are dissociated and cultured from a vascular retina. METHODS: Müller cells were dissociated from adult rabbit retinas by using an enzymatic digestion-mechanical trituration technique and a cell attachment method that provided Müller cell- enriched cell cultures. Indirect immunofluorescence localization of vimentin, glial fibrillary acidic protein (GFAP), glutamine synthetase (GS), beta-amyloid precursor protein (beta-APP), and (alpha-smooth muscle actin (alpha-SMA) was carried out on Müller cells that were freshly dissociated, on those that had been in culture 2 and 6 days, and on confluent primary cultures and late-passage cultures. The specificity of the antibodies and changes in protein expression were examined by western blot analysis. RESULTS: The expression of vimentin, GFAP, GS, and beta-APP was present 2 days after dissociation and was retained through 6 days in culture, at which time alpha-SMA began to be expressed in a small number of cells. The confluent, primary cultures no longer expressed GS, but vimentin and beta-APP were still expressed, and the expression of alpha-SMA was increased. During the late-passage stage, the morphologic appearance of the Müller cell cultures was large and amorphous, with additional changes in antigenicity. Although there was loss of expression of the intermediate filament proteins GFAP and vimentin, the expression of beta-APP was maintained, whereas alpha-SMA was increased and appeared to be a major cytoskeletal protein. CONCLUSIONS: Dissociated Müller cells that were maintained in culture underwent phenotypic changes that included a large, amorphous appearance; the loss of detectable vimentin, GFAP, and GS expression; the persistent presence of beta-APP; and the de novo appearance of alpha-SMA. The phenotypic and antigenic changes that occurred in cultured Müller cells from an avascular retina were similar but not identical to the changes observed in cultured Müller cells from a vascular retina.

Alpha ganglion cells of the rabbit retina lose antagonistic surround responses under dark adaptation.

Alpha ganglion cells from the midperiphery of the rabbit retina were recorded intracellularly under visual control, in a superfused everted eyecup, and labeled with HRP. Their physiology and large somata with broad dendritic arbors identified them as uniform populations of ON- and OFF-center alpha ganglion cells, which typically displayed transient/sustained light-evoked responses. When dark adapted, the light-evoked responses from both ON- and OFF-center alpha ganglion cells were more sustained than those generally seen under light-adapted conditions. During dark-adapted (scotopic) conditions, stimulation with dim full-field illumination and small spots, either positioned over the soma or displaced 450 microns from the soma, all elicited pure center responses. After light adaptation (photopic conditions), the displaced small spots that previously evoked center responses elicited antagonistic surround responses from both ON- and OFF-center cells. Thus, as originally described in cat retina (Barlow et al., 1957), the receptive-field organization of ganglion cells changed between dark and light adaptation, and an absence or presence of surround antagonism was indicative of scotopic versus photopic states.

Protein kinase modulation of GABAA currents in rabbit retinal rod bipolar cells.

1. Protein kinase modulation of gamma-aminobutyric acid-A (GABAA)- and glycine-activated Cl- currents in freshly dissociated, morphologically identified rabbit retinal rod bipolar cells was studied under voltage clamp with the use of the whole cell patch-clamp technique. Responses to pulses of GABA and glycine were monitored before, during, and after application of adenosine 3',5'-cyclic monophosphate (cAMP)-dependent protein kinase (PKA) and protein kinase C (PKC) activators, inactive analogues, and inhibitors. 2. Bath perfusion with either forskolin, an adenylate cyclase activator, or its inactive analogue, 1,9 dideoxyforskolin, reduced the GABA-activated Cl- currents by 30-50%; coapplication of N-[2-(Methylamino)ethyl]-5-isoquinolinesulfonamide hydrochloride (H-8), a PKA inhibitor, did not prevent the forskolin effects. The membrane-permeable cAMP analogues, 8-bromo-cAMP and 8-(4-Chlorophenylthio)-cAMP, and intracellularly dialyzed cAMP, did not modulate either the GABA- or glycine-activated Cl- current. Perfusion of the phosphodiesterase inhibitor 3-isobutyl-1-methylxantine (IBMX) had no direct effect on the GABA-activated current and did not alter the results with cAMP or its membrane-permeable analogues. Collectively, these results make it very unlikely that PKA represents an important mechanism of either GABAA or glycine channel modulation in the rabbit rod bipolar cell. 3. Although the isoquinoline sulfonamide protein kinase inhibitor H-8 was without discernible effect, the related compounds 1-(5-Isoquinolinesulfonyl)-2-methylpiperazine dihydrochlorine (H-7) and N-(2-Aminoethyl)-5-isoquinolinesulfonamide dihydrochloride (H-9) both dramatically reduced the GABA response. H-7 also strongly reduced the response to glycine, whereas H-8 had no effect and H-9 had an intermediate effect. Because only certain members of this inhibitor class of agents proved effective, and their effectiveness appeared unrelated to the established activity profiles, these agents probably inhibit the Cl- currents in a phosphorylation-independent manner. Direct interaction of these inhibitors with binding sites in the GABAA receptor-channel complex has been previously reported in some other preparations. 4. The phorbol ester and PKC activator phorbol 12,13 dibutyrate (PDB) led to a 35-55% reduction in the GABA-activated Cl- current of the rod bipolar cell. The broad-spectrum kinase inhibitor staurosporine, and the more PKC-specific inhibitor calphostin C, had no direct effect on GABA responses, but prevented Cl- current reduction when coapplied with PDB. Phorbol 12-myristate 13-acetate (PMA) reduced the GABA-activated current in a fashion very similar to PDB. Staurosporine and calphostin C blocked the PMA effect. No reduction of Cl- current was seen with the inactive analogue, 4-alpha-PMA, used as a control for PKC-independent phorbol ester effects. 5. PDB effectively reduced the GABA-activated Cl- current of the rod bipolar cell at low concentrations, whereas PMA had a diminished effect at low concentrations. This is consistent with the reported concentration-dependent abilities of these agents to promote translocation of PKC-alpha immunoreactivity from the membrane to the cytosolic compartment in the rabbit retinal rod bipolar cell. Collectively, the data from phorbol esters, inactive analogues, and kinase inhibitors support the existence of a PKC-mediated mechanism for GABA-activated Cl- current reduction in these cells. 6. The naphthalenesulfonamide PKC activator N-(n-Heptyl)-5-chloro-1-naphthalenesulfonamide (SC-10) also potently and reversibly reduced the GABA-activated current. Staurosporine and calphostin C eliminated this effect. When the nonhydrolyzable guanosine 5'-triphosphate (GTP) analogue guanosine 5'-O-(3-thiotriphosphate) tetralithium salt (GTP-gamma-S) replaced GTP in the recording pipette, the SC-10-mediated GABA current reduction became irreversible.(ABSTRACT TRUNCATED)

Connections of two types of flat cone bipolars in the rabbit retina.

The synaptic connections of two types of cone bipolar cells in the rabbit retina were studied with the electron microscope after labeling in vitro with 4',6-diamidino-2-phenylindole (DAPI), intracellular injection with Lucifer Yellow, and photooxidation (Mills and Massey [1992] J. Comp. Neurol. 321:133). Both types of bipolars belong to the flat variety, because they make basal junctions with a group of four to ten neighboring cone pedicles. One cell type has an axonal arborization that occupies strata 1 through 3 of the inner plexiform layer (IPL). At ribbon synaptic junctions, it is presynaptic to ganglion cell dendrites and to reciprocal dendrites belonging to narrow-field bistratified (AII) amacrine cells. In addition, it contacts and is contacted by other amacrine cell processes of unknown origin. The other cell type has an axonal arborization entirely confined to stratum 2 of the IPL; it is pre- or postsynaptic to a pleomorphic population of amacrine cell processes, and, in particular, it receives input from the lobular appendages of AII. Thus, these two bipolar types probably belong to the off-variety because they make basal junctions with cone photoreceptors and send their axon to sublamina a of the IPL, which is occupied by the dendrites of off-ganglion cells. They are also part of the rod pathway because they receive input from AII amacrine cells.

Light responses from one type of ON-OFF amacrine cells in the rabbit retina.

1. The light responses from one type of ON-OFF amacrine cell were recorded intracellularly in the superfused rabbit retina under various conditions of light adaptation. These recordings were obtained from cells located in a central area. 5-7 mm inferior and directly below the optic nerve head. 2. ON-OFF amacrine cells responded to the initiation and termination of light stimuli with transient depolarizations. Their receptive fields were approximately 0.8-1 mm diam and did not exhibit antagonistic center-and-surround organization. 3. The cells received rod input because they responded to very dim scotopic stimuli. With prolonged dark adaptation, the cells became more sensitive to the initiation than termination of the stimulus, because the ON component of the light response had a lower threshold than the OFF component. 4. The cells continued to respond to test flashes when the retina was adapted to a background illumination of rod-saturating intensity. Thus ON-OFF amacrine cells also receive cone input. Under these photopic conditions, a secondary afterpotential was observed following the OFF component. Its characteristics were different from those of the rod aftereffect reported in other retinal cells of the rabbit because its latency and amplitude changed with increasing stimulus intensity. 5. Intracellular injections of horseradish peroxidase showed that the recordings were obtained from a class of ON-OFF amacrine cells whose wide-field, unistratified dendrites were rigorously confined to the middle of the inner plexiform layer or stratum 3. 6. The conspicuous rod and cone inputs into a class of amacrine cells that are connected neither to rod bipolars nor to All amacrine cells strongly support the idea that in the rabbit the rod pathway uses cone bipolars as interneurons to distribute scotopic signals to ganglion and cone-driven amacrine cells.

GABA- and glycine-activated currents in the rod bipolar cell of the rabbit retina.

1. Voltage- and ligand-gated currents were recorded from solitary rabbit rod bipolar cells using the whole cell patch-clamp technique. The rod bipolar cell forms a single, stereotypical physiological and morphological class of cells that was easily identified from other neurons and support cells after enzymatic and mechanical dissociation from isolated retina. Protein kinase C immunoreactivity confirmed the validity of using a purely morphological identification of this cell type. 2. Voltage steps in 15-mV increments from a holding potential of -45 mV elicited a large outward current activated near -30 mV. These voltage-gated currents were eliminated by using equimolar substitutions of Cs+ and tetraethylammonium+ for K+ in the pipette, indicating that they represent a mixture of K+ currents. 3. The putative inhibitory neurotransmitters gamma-aminobutyric acid (GABA) and glycine activated inward Cl- currents when pressure-applied from pipettes placed near the axon terminals of rod bipolar cells, which were voltage-clamped at -45 mV. With changes in intracellular or extracellular Cl- concentration, the reversal potential of these ligand-gated currents changed as predicted by the Nernst equation for Cl- activity. The dose-response curves for GABA and glycine were sigmoidal with saturating concentrations of 100 and 300 microM, respectively. 4. GABA-activated currents were 1) reversibly reduced by the allosteric inhibitor picrotoxin and the competitive antagonist bicuculline; 2) potentiated by the benzodiazepine diazepam and the barbiturate barbital sodium; and 3) indistinguishable from muscimol-activated currents. There was no response to the GABAB agonist baclofen. Collectively, these data strongly suggest that the GABA-activated currents in rabbit rod bipolar cells are mediated by the GABAA receptor. This is similar to the GABA-activated currents in other mammalian rod bipolar cells. 5. Application of the conformationally restricted GABA analogue cis-4-aminocrotonic acid (CACA) failed to elicit a response, whereas the conformationally extended GABA analogue trans-4-aminocrotonic acid (TACA) elicited a response similar to that of GABA. Although bicuculline appeared to suppress the GABA-activated current slightly more than the TACA-activated current (not significant using Student's t-distribution), GABA- and TACA-activated currents were equally suppressed by picrotoxin and equally enhanced by diazepam and barbital sodium. These data, coupled with the inefficacy of CACA, argue against the existence of a GABAC-type channel in the rod bipolar cell of the rabbit and suggest that GABA and TACA were activating the same GABAA receptor-channel complex.(ABSTRACT TRUNCATED AT 400 WORDS)

Cone bipolar cells as interneurons in the rod pathway of the rabbit retina.

In the mammalian retina, rod signals are transmitted by rod bipolars to the narrow-field, bistratified (AII) amacrine cell. This neuron, in turn, makes gap junctions with the axonal arborization of cone bipolar cells that reside in the vitreal half (sublamina b) of the inner plexiform layer (IPL). After examining rod bipolars and AII amacrines in the rabbit retina, we have now reconstructed from electron micrographs of continuous series of thin sections the synaptic connections of the axonal arborizations of cone bipolar cells that make the highest number of gap junctions with AII amacrines. These axonal arborizations were narrowly confined to stratum 4 (S4) of the IPL and made ribbon synapses to dyads of postsynaptic dendrites that belonged to either ganglion or amacrine cells. In the population of postsynaptic processes, 30% were ganglion cell dendrites. These dendrites were probably originating, at least in part, from on-center ganglion cells because their course was confined to sublamina b of the IPL. Of the remaining postsynaptic processes, 51.7% belonged to amacrine cells and 18.3% were not identified. Among the postsynaptic amacrine cell processes, 33.3% returned a reciprocal synapse onto the cone bipolar endings. These reciprocal synapses represented 21.3% of the total input onto the axonal arborizations, the remaining fraction (78.7%) arising from a heterogeneous population of amacrine dendrites that were purely presynaptic to the cone bipolars endings. Pre- and postsynaptic amacrines were part of several distinct microcircuits which suggest complex local processing of both rod and cone signals. Thus, the cone bipolars that make gap junctions with AII amacrines in sublamina b of the rabbit IPL exhibit a substantial output onto ganglion cells. This fact, in conjunction with our previous observations that in this sublamina ganglion cells receive negligible input from rod bipolars and AII amacrines, demonstrates that in the rabbit cone bipolars represent a necessary link in the pathway followed by rod signals to enter on-center ganglion cells. Thus, rod and cone signals ultimately share the same integrating mechanisms and converge onto the same set of ganglion cells.

Synaptic connections of the narrow-field, bistratified rod amacrine cell (AII) in the rabbit retina.

The synaptic connections of the narrow-field, bistratified rod amacrine cell (AII) in the inner plexiform layer (IPL) of the rabbit retina were reconstructed from electron micrographs of continuous series of thin sections. The AII amacrine cell receives a large synaptic input from the axonal endings of rod bipolar cells in the most vitreal region of the IPL (sublamina b, S5) and a smaller input from axonal endings of cone bipolar cells in the scleral region of the IPL (sublamina a, S1-S2). Amacrine input, localized at multiple levels in the IPL, equals the total number of synapses received from bipolar cells. The axonal endings of cone bipolar cells represent the major target for the chemical output of the AII amacrine cell: these synapses are established by the lobular appendages in sublamina a (S1-S2). Ganglion cell dendrites represent only 4% of the output of the AII amacrine and most of them are also postsynaptic to the cone bipolars which receive AII input. The AII amacrine is not presynaptic to other amacrine cells. Finally, the AII amacrine makes gap junctions with the axonal arborizations of cone bipolars that stratify in sublamina b (S3-S4) as well as with other AII amacrine cells in S5. Therefore, in the rabbit retina 1) the rod pathway consists of five neurons arranged in series: rod-->rod bipolar-->AII amacrine-->cone bipolar-->ganglion cell; 2) it seems unlikely that a class of ganglion cells exists that is exclusively devoted to scotopic functions. In ventral, midperipheral retina, about nine rod bipolar cells converge onto a single AII amacrine, but one of them establishes a much higher proportion of synaptic contacts than the rest. Conversely, each rod bipolar cell diverges onto four AII amacrine cells, but one of them receives the largest fraction of synapses. Thus, within the pattern of convergence and divergence suggested by population studies, preferential synaptic pathways are established.

Axonless horizontal cells of the rabbit retina: synaptic connections and origin of the rod aftereffect.

An axonless horizontal cell (AHC) of the rabbit retina was penetrated with a microelectrode and stained with horseradish peroxidase after recording its light responses. The cell was then serially sectioned and its connections examined with the electron microscope. Physiologically, the cell exhibited cone-dominated responses and a minor rod influence known as rod aftereffect. Electron microscopy showed that this AHC was only connected to cones. Therefore, the rod aftereffect could only invade the cell through the gap junctions between the synaptic endings of rod and cone photoreceptors. In the synaptic invaginations of the cone pedicles contacting the cell, only one of the lateral elements was stained. This suggests that the two lateral elements of each cone-invaginating synapse belong to two different horizontal cells. By staining intracellularly adjacent AHCs, we showed that the two lateral processes may originate from two horizontal cells belonging to the same morphological type.

Synaptic connections of rod bipolar cells in the inner plexiform layer of the rabbit retina.

We have reconstructed from electron micrographs of a continuous series of thin sections the synaptic connections of the axonal arborizations of all the rod bipolar cells contained in a small region of the retina of the rabbit. We observed that all rod bipolars share the same pattern of connectivity and are probably functionally equivalent. As a rule, they do not contact ganglion cells. Their prevalent synaptic output is on narrow-field, bistratified, and indoleamine-accumulating amacrine cells. Their dominant inputs are the reciprocal synapses from the indoleamine-accumulating amacrines, but they also receive a sizable number of synaptic contacts from other, non-reciprocal, amacrine cells. The lateral spread of scotopic signals at the synapse between rod bipolars and narrow-field, bistratified amacrines is small. Finally, in the rabbit, as in the cat, a narrow-field, bistratified amacrine is inserted in series along the rod pathway.

Connections of indoleamine-accumulating cells in the rabbit retina.

To study the connections of the neurons of the rabbit retina that accumulate indoleamines, we injected 5,7-dihydroxytryptamine into the vitreous body. It accumulated within a subset of amacrine cells and could be visualized there by aldehyde-induced fluorescence. The fluorescent labeling was photo-converted to an insoluble, osmiophilic product by irradiation in the presence of diaminobenzidine, and the tissue was examined by electron microscopy. Preservation of the structure of the tissue after photoconversion was satisfactory and the dendrites of the indoleamine-accumulating cells could easily be identified. They form a dense plexus near the junction of the inner plexiform and ganglion cell layers, where they exhibit large synaptic endings that occupy a substantial fraction of the surface of rod bipolar terminals. The dendrites of the indoleamine-accumulating cells receive input from rod bipolars at dyad synapses, where the other postsynaptic partner is a dendrite of a narrow-field, bistratified amacrine cell; in addition, they receive amacrine cell input throughout the inner plexiform layer. The only outputs we observed are reciprocal synapses onto the rod bipolar endings. Thus, these amacrine cells appear to exert an important effect on the transmission of scotopic information through the retina.

Rod photoreceptors dissociated from the adult rabbit retina.

Rod photoreceptors have been isolated from the adult rabbit retina using enzymatic and mechanical dissociation procedures; their fine structure, synaptic activity, and long-term viability were examined using conventional electron-microscopic, quick-freezing, and cell culture techniques. Freshly dissociated photoreceptors were well-preserved compared to their counterparts in the intact retina. About half of the cells, however, exhibited broad continuity between inner and outer segments. Quick-frozen, freeze-substituted rods differed from chemically fixed cells in 3 respects: (1) there was an increased amount of granular matrix in the cytoplasm, mitochondria, and rough endoplasmic reticulum; (2) branching and anastomosing profiles of smooth endoplasmic reticulum had disappeared from the inner segment; and (3) the number of synaptic vesicles within the spherule was highly variable, in some cases leaving synaptic ribbons completely denuded of their halo of vesicles. Light-adapted, solitary rod cells continued to be synaptically active: their endings were capable of endocytosis when placed in the dark in the presence of extracellular ferritin and tracer was incorporated into vesicles and vacuoles; this uptake was much reduced when the cells were incubated with the tracer in the light. Thus, synaptic vesicle regeneration was stimulated in the dark, suggesting that vesicles underwent exocytosis in the dark. Isolated rod cells adhered poorly to most standard substrates; without proper adhesion, cells deteriorated in 2-4 hr. However, photoreceptors did adhere to glutaraldehyde-fixed Vitrogen gels and could be maintained for over 48 hr on this substrate if kept in a complete medium at 22 degrees C. In contrast, Müller cells adhered quickly to a laminin substrate with their endfoot processes. The differential adhesion properties of Müller and photoreceptor cells may be useful in obtaining pure populations of glial cells or neurons from the adult mammalian retina.