A17: a broad-field amacrine cell in the rod system of the cat retina.

Abstract

A17 amacrine cells of the cat retina have been penetrated with horseradish peroxidase (HRP)-filled microelectrodes and their light responses recorded. These cells depolarize in sustained fashion to steps of light. Viewed in retinal wholemounts, HRP-injected cells have a spokelike radiating splay of very fine dendrites (0.1 micron diam) passing diffusely through all strata of the inner plexiform layer (IPL) to run primarily in strata 4 and 5. There are as many as 1,000 large, regularly spaced beads borne on the 500- to 1,200-micron diameter dendritic field. Cell body sizes range from 9 to 13 micron. In the electron microscope, the dendritic beads in sublamina b of the IPL are seen to synapse reciprocally with rod bipolar axon terminals. Dendritic beads in sublamina a rarely make synapses, but between the beads in this layer, input from at least three distinctive amacrine profiles occurs. Though diffuse at the light microscopic level, A17 thus appears to be structurally bistratified, with amacrine input in sublamina a and bipolar input in sublamina b. It is likely that A17 can be identified with AI. A17 signals are driven almost exclusively by rods. The spectral sensitivity peaks at 507 nm, identical with that of pigment epithelial cells. Light adaptation abolishes all but a small hyperpolarizing component of the signal. The overall intensity-response range is similar to that of AII amacrine cells. When receptive fields of A17 cells are mapped with slit stimuli, a broad, single-component curve is measured approximately covering the dendritic field. The receptive field is well described by a linear electrical model with a mean space constant of 259 +/- 97 micron (SD). On the other hand, responses to centered slit stimuli of varying width yielded space constants of only 38 +/- 29 micron. A17 amacrines are thus broad-field components of the cat's rod system but with very little capacity for spatial integration. Receptive-field measurements are not supportive of the notion of isolated dendritic regions.