Wide-field amacrine cells show ON-OFF physiology. A : Intracellular recording of the response of a wide-field amacrine cell to full-field illumination consisting of large synaptic potentials at light onset and offset with transient bursts of spiking. Trace below the recording indicates onset and offset of the light stimulus. B : Application of 50 m M L-AP4 completely abolished the ON response component, whereas the OFF response remained. This result indicates that the OFF response component is generated via the OFF retinal pathway. 

Context in source publication

Context 1

... corresponding to eccentricities of 0.6–2.5 mm from the medullary rays. The amacrine cells in this study displayed dendritic arbors extending Ͼ 1.5 mm and, as detailed later, shared a large number of morphological and physiological features in com- mon. Because their extremely large size was the characterizing feature of these cells, we use the generic term “wide-field cells” to identify them. However, up to 10 different subtypes of wide-field amacrine cells have been described in the rabbit retina ~ MacNeil & Masland, 1998; MacNeil et al., 1999 ! and so it needs to be emphasized that the cells in this study represent only a small subset of the wide-field category and, as detailed later, most likely can be classified as a single amacrine cell subtype. The wide-field amacrine cells possessed round somata, 8–10 m m in diameter, which were exclusively situated in the proximal region of the INL. They showed three and four primary dendrites ~ Fig. 1 ! , each of which divided up to five times within the first 75 m m of their course, giving rise to long, terminal dendritic branches that showed few crossings ~ Figs. 1A, 2A ! . The circular, radially symmetric dendritic arbors were very extensive, with diameters ranging from 1500 to 2500 m m. The terminal branches, which were 600–1300 m m in length, appeared uniform in caliber except for the presence of numerous varicosities irregularly distributed throughout their length ~ Figs. 1B to 1D ! . Curiously, most cells also possessed one to two terminal branches that were relatively short, extending only 50–100 m m. These shorter branches were well labeled and showed clear termina- tions, indicating that they did not result from an artifact because of poor intracellular diffusion of Neurobiotin. Despite the large size of the wide-field cells, the overall appearance of their dendritic arbors was one of a homogeneous structure. That is, there was no indication of a separation of the arbor into dendritic and axonal systems as displayed by the polyaxonal amacrine cells ~ Völgyi et al., 2001 ! . Computer reconstructions of nine cells indicated that their dendrites were unistratified in the middle region of the IPL. All cells stratified within stratum 3, with some cells showing processes that straddled the borders with the adjacent strata 2 and 4 ~ Fig. 1B ! . As described in Methods, we found that the error in the measure of stratification in the IPL using the computer reconstruction system was 2 m m and so the variability in the stratification level across the cell population likely reflects experimental error. The wide-field amacrine cells showed very extensive tracer coupling. This included homologous coupling to neighboring wide- field cells as well as heterologous coupling to other amacrine cell subtypes ~ Figs. 1 and 2 ! . Overall, these coupled amacrine cells had somata in the INL that varied in size and labeling intensity, suggesting coupling to at least two other amacrine cell subtypes ~ Fig. 3 ! . Interestingly, for every wide-field cell injection we found that the array of coupled amacrine cell somata occupied an area that matched that of the dendritic arbor of the injected cell. This circumscribed coupling pattern is difficult to explain considering the large dendritic arbors of the wide-field cell, but it suggests that distal dendrites of one amacrine cell do not form gap junctions with the distal dendrites of adjacent cells. On three occasions, we also visualized a few tracer-coupled ganglion cell bodies in the GCL, which were identified by their axons that projected to the optic disk ~ not shown ! . However, in these instances, the injections resulted in deposits of Neurobiotin in the extracellular space around the injection site. Thus, it is likely that this rare amacrine-to-ganglion cell coupling reflected artifac- tual labeling by tracer uptake from the extracellular space. The wide-field amacrine cells responded to full field light stimulation with ON-OFF responses consisting of transient trains of spikes riding atop of large excitatory synaptic potentials at light onset and offset ~ Fig. 4A ! . The synaptic potential latencies were similar for the ON and OFF responses ~ 100.8 6 6.3 vs . 96.4 6 5.2 ms ! suggesting that they were both center-mediated and did not reflect a combination of a faster center-mediated response and a slower surround-mediated response ~ Werblin & Dowling, 1969 ! . Application of the mGluR6 agonist, L-AP4, reversibly blocked the ON response, but it had no effect on the OFF response ~ Fig. 4B ! . This result provides further support for the idea that the ON and OFF responses reflect inputs from the ON and OFF retinal channels. Application of L-AP4 on the cell illustrated in Fig. 4 revealed a small, sustained hyperpolarization during light, possibly reflect- ing a sustained input from OFF bipolar cells, but this effect was not seen for most cells. To determine the size of wide-field cell receptive fields we first calculated their area summations using concentric spots of light of increasing size. Although this measure was somewhat approximate because of the large intervals between spot sizes, it was evident that the area summations of the ON and OFF slow potential components were different. For example, the cell responses illustrated in Fig. 5 show a summated response for the OFF slow potentials to relatively small spots of light, between 750–1250 m m in diameter, whereas the ON response showed an area summation between 1750–2500 m m. This difference in area summation between the ON and OFF response components was consistent across the population of wide-field amacrine cells tested ~ n ϭ 9 ! . In addition, the area summation profiles in Fig. 5B demonstrate that neither the ON nor the OFF responses showed evidence of attenuation during presentation of relatively large spots of light ~ Ͼ 2500 m m ! . This indicates that surround-mediated inhibition does not play a significant role in shaping the light evoked activity of these wide-field amacrine cells. We also computed the Gaussian diameters of the wide-field cells as a second method to measure receptive field size. As detailed in Methods, a narrow slit of light was moved across the retina in discrete steps from which the relationship between slow potential amplitude and slit distance was used to compute the Gaussian diameters. It was evident from the raw recordings that the OFF response component was lost first as the slit was moved away from the center position. For the cell illustrated in Fig. 6A, the OFF responses were completely lost when the slit was posi- tioned only 350– 400 m m from the center position. In contrast, the ON slow response disappeared only when the slit was displaced beyond 1 mm. This dramatic difference was evident when the Gaussian diameter was computed for the cell; 1747 m m for the ON component and 372 m m for the OFF component ~ Fig. 6C ! . The large difference in the Gaussian diameter of the ON and OFF responses was seen across the entire population of wide-field cells studied ~ 1821 6 82 m m for ON responses vs . 358 6 19 m m for OFF responses; n ϭ 12 ! . Interestingly, we found that wide-field amacrine cells with larger dendritic fields displayed ON response components with larger Gaussian diameters ~ r 2 ϭ 0.87 on ...