Transmitter interactions in the rabbit retina.

surfaces (Fig. 1). Most cells were not labelled and could be compared with controls. The morphology of the ThyI-labelled cells in culture was quite distinctive; they are the largest, with an eccentrically placed nucleus. Furthermore these cells do not accumulate I'H Iglycine, I 3H IGABA* or D-l )H laspartate as do other cells in the culture (see Beale & Osborne 1983). We therefore conclude that Thy-I antigen is restricted to rat ganglion cells in culture as it appears to be in intact tissue. Thy-I antibody was subsequently used as a probe for obtaining 'enriched' populations of ganglion cells. Cells dissociated from neonatal and adult rat retinas were separated by densitygradient centrifugation. The proportion of ThyI -positive neonatal cells was increased from about 0.4% in the initial dissociates to about 8% in the most enriched fraction of a Percoll step gradient. Among adult cells initial 0.7% ThyI-positive cells were increased to roughly 4% in the best fraction of a metrizamide step gradient. The presence of relatively large numbers of ThyI-positive cells in other fractions suggests that it would be difficult to increase the proportion of rat ganglion cells (either neonatal or adult) by methods based on their sedimentation properties. We have, in fact, used a variety of other gradients in an attempt to increase yields, but to no avail. The degree of enrichment resulting could, nevertheless, be a useful preliminary step to other immunoselection procedures such as fluorescence-activated cell sorting, panning (Wyscocki & Sati, 1978) or cell selection using antibody-coated magnetic beads (Meier ef al., 1982), all of which also depend on celltype-specific antibodies. One of the main arguments against 5-hydroxytryptamine being a transmitter in the mammalian retina is the fact that present procedures are unable to show that the endogenous amine is restricted to specific retinal neurons (Osborne, 1982; Ehinger, 1982). When rabbit retinas are incubated with exogenous 5-hydroxytryptamine (0.1 p ~ ) at 37°C and then processed by immunohistochemistry to localize 5-hydroxytryptamine, specific amacrine neurons can be seen to have taken up the exogenous amine (Fig. 2). Clearly, if it is possible to isolate these '5-hydroxytryptamine-accumulating cells', then it would be possible to design experiments to establish whether these cells contain endogenous 5-hydroxytryptamine. Specific neurons from neonatal rabbit retinal cultures also take up exogenous 5-1 'H lhydroxytryptamine (Fig. 2) and these cells can be distinguished from other neurons, which have the capacity to take up I 'H Iglycine, I'H Idopamine. I 'H IGABA or I3H laspartate. An enriched population (3-fold increase) of the 5-1 )H Ihydroxytryptamine-accumulating cells from cultures using a Percoll step gradient (see Schaeffer, 1982) has been produced. It is still important, however, to improve on this enrichment before the cells can be analysed for meaningful * Abbreviation: GABA, y-aminobutyrate. data on their endogenous 5-hydroxytryptamine content. We propose to use I "C ldansyl(5-dimethylaminonaphthalene1sulphony1)chloride (see Osborne 1974) to determine simultaneously the total and -'H-labelled amount of 5-hydroxytryptamine in the different cell populations of a cell culture fractionated on a Percoll gradient. It may then be possible to establish whether mammalian retinal neurons that take up exogenous 5-hydroxytryptamine also contain low amounts of the amine. It has been suggested that neurons in the mammalian retina which take-up exogenous 5-hydroxytryptamine actually do not contain the amine but utilize another chemical as a neurotransmitter (Ehinger & Floren, 1980: Floren & Hansson, 1980). In summary, histological and electrophysiological studies have clearly confirmed the heterogeneity of neuron types in the mammalian retina. Although immunohistological studies have facilitated the chemical characterization of certain retinal neurons in intact tissues, the data are limited and restricted primarily to the amacrine cells. Attempts have been made to isolate specific groups of cells, e.g. ganglion cells, for biochemical analysis. Unfortunately, the populations of neurons so far produced are not of a purity sufficient for chemical analysis.