1. Intracellular and extracellular recordings were obtained from ganglion cells in the rabbit retina. The effect of N-methyl-DL-aspartate (NMDLA) and N-methyl-D-aspartate (NMDA) antagonists were studied with the use of a perfusion method for drug application. 2. NMDLA excited all ganglion cell types and caused a characteristic burst firing pattern, which is not typical of physiological responses in the retina. When synaptic transmission was blocked with cobalt, NMDLA still excited ganglion cells, indicating a direct action. 3. A comparison of DL-2-amino-5-phosphonopentanoate (DL-AP-5) and DL-2-amino-7-phosphonoheptanoate (DL-AP-7) revealed that DL-AP-7 was a more specific NMDA antagonist. DL-AP-5 partially blocked the b-wave of the electroretinogram (ERG), an action typical of L-2-amino-4-phosphonobutyrate (L-APB), which specifically blocks on channels in the retina. 4. DL-AP-7 reversibly blocked the action of NMDLA on all ganglion cell types, but the effects of kainate (KA) and carbachol were unchanged. AP-7 was stereospecific and pharmacologically specific, with action typical of a competitive NMDA antagonist in the rabbit retina. 5. DL-AP-7 did not block light responses driven by center or surround stimulation for ON or OFF ganglion cells. Directional selectively was unchanged by DL-AP-7. However, most ganglion cells showed a reduction, typically 20-30%, in the number of action potentials produced by light stimulation. 6. In contrast to a previous report, we found no evidence that DL-AP-7 specifically inhibited sustained ON ganglion cells. The inhibition of sustained ON responses by DL-AP-5, previously attributed to NMDA antagonism, is probably because of the weak APB activity of L-AP-5. 7. We conclude that NMDA receptors do not mediate the major light-driven input to ganglion cells in the rabbit retina. By exclusion, transmission from bipolar cells to ganglion cells appears to be carried mostly by KA or quisqualate (QQ) receptors. However, because NMDA antagonists reduced the number of action potentials produced by light stimulation, it is likely that NMDA receptors carry a portion of the signal transmission to ganglion cells. The presence of NMDA receptors on third-order neurons is consistent with the release of glutamate from presynaptic neurons such as bipolar cells.

A variety of glutamate and aspartate analogues were used to characterize the excitatory amino acid receptors in the mudpuppy retina. This approach revealed two general classes of receptors which were represented by the agonists kainic acid and N-methyl aspartic acid. Kainic acid was found to be a potent photoreceptor transmitter agonist on all three types of second-order neurons, and it was a powerful excitant of amacrine and ganglion cells. N-Methyl aspartate had little effect in the outer retina, but it had potent stimulatory effects on inner retinal neurons. N-Methyl aspartate antagonists selectively blocked light responses in some sustained OFF ganglion cells. These results suggest that both photoreceptors and bipolar neurons may use glutamate Or an analogue, whereas aspartate may be utilized by a class of sustained ON amacrine cells.

Intracellular electrophysiological recordings were obtained from amacrine and ganglion cells in an isolated, superfused retina-eyecup preparation of the rabbit. Cells were characterized physiologically, after which cell-staining was accomplished by intracellular iontophoresis of HRP. A computer-assisted image-processing system was used to study the dendritic stratification pattern of HRP-labeled neurons within the inner plexiform layer (IPL). Our results support the concept that the IPL is functionally divided into a distal OFF region and proximal ON layer. ON and OFF ganglion and amacrine cells show dendritic arborizations consistent with this division and ON-OFF ganglion cells have processes in both portions of the IPL. It appears that these functional subdivisions of the IPL reflect excitatory, but not necessarily inhibitory, inputs. Thus, the pattern of dendritic arborization of a cell appears to predict its physiological response polarity, regardless of the type of inhibition it receives.

SUMMARY1. Intracellular and extracellular recordings were obtained from ganglion cells in the rabbit retina. The effects of glutamate analogues and antagonists were studied using a perfusion method for drug application.2. Kainate (KA) excited all ganglion cells directly and caused a large increase in firing rate. N-Methyl-DL-aspartate (NMDLA) also excited ganglion cells but it was less potent and caused burst firing.3. Quisqualate (QQ) and (RS)-2-amino-3-hydroxy-5-methyl-isoxazole-4-propionic acid (AMPA) excited many ganglion cells and were approximately as potent as KA. Less frequently, QQ and AMPA had inhibitory effects possibly due to polysynaptic action.4. General glutamate antagonists such as cis-2,3-piperidine dicarboxylic acid (PDA) and kynurenic acid blocked the light input to all ganglion cells. PDA and kynurenic acid blocked the effects of KA and NMDLA, but not carbachol, indicating that they act as glutamate antagonists in the rabbit retina. Kynurenic acid did not block the excitatory action of QQ, even though light responses were abolished.5. Amacrine cells were depolarized by KA or QQ and less potently by NMDLA. Their light-evoked responses were blocked by PDA.6. We conclude that the light input to ganglion cells in the rabbit retina is predominantly mediated by KA receptors. This is consistent with the idea that 'on' and 'off' bipolar cells are excitatory and release glutamate.

The neuropeptide somatostatin (SS) has been localized to neurons of the rabbit retina by immunochemical and biochemical methods (Sagar et al., 1982, 1986; Marshak and Yamada 1984). We examined the effects of bath-applied SS on neurons of the rabbit retina, using intra- and extracellular electrophysiological techniques in an in vitro retina eyecup preparation. All commonly encountered ganglion cell receptive field types were affected by SS, and the effects were of 3 kinds: The first was a general excitation, occurring with a threshold concentration of about 100 nM; the onset of the excitation was too slow (seconds) for SS to participate in any rapid light-evoked responses. The second SS effect was an increase in the "signal-to-noise ratio," defined here as the ratio of light-evoked to spontaneous spiking, which resulted from a decrease in spontaneous activity and, usually, a concomitant increase in light-evoked spiking. The third effect was a shift in center-surround balance towards a more dominant center. The signal-to-noise and center-surround effects were evident at concentrations as low as 0.5 nM; both were slow onset (tens of seconds) and long lasting (tens of minutes). SS acted at multiple levels within the retinal circuitry to produce the observed changes in ganglion cell output. These effects included direct actions on ganglion and amacrine cells, and a decrease in the efficiency with which horizontal cells could drive the retinal network. At least part of these SS actions on third-order neurons resulted from a decrease in conductance to ions with an equilibrium potential more positive than dark membrane potential. The degradation-resistant SS agonist SMS201-995 had effects qualitatively and quantitatively similar to those of SS, suggesting that SS may be degraded slowly enough to act at a distance from its sites of release. While no adequate SS antagonist is available, the greater sensitivity to exogenous SS, in retinas depleted of their SS content (with cysteamine), suggests a role for endogenous SS. The potency of SS also reinforces this view. The results of this study suggest that SS may be a neuromodulator in the rabbit retina, producing long-lasting changes in the "signal-to-noise" discharge pattern and center-surround balance of ganglion cells.