Retinal ganglion cells (RGCs) process visual information in the eye and transmit it to the brain. All RGCs share key features, but distinct subtypes have been identified that differ in morphology, biophysical properties, visual responsiveness, synaptic inputs and synaptic targets. These RGC subtypes have been studied intensively with anatomical and physiological methods, both because they are critical determinants of visual perception and because they provide an accessible system for addressing general issues of neuronal diversification and categorization. Unfortunately, however, few if any molecular markers are known that distinguish functionally relevant mammalian RGC subclasses, so it has been impossible to answer basic questions about how many subtypes really exist, when they arise, how they acquire their definitive identities, or what molecules underlie their subtype-specific properties. Here, we propose to identify molecular markers for two pairs of RGC subsets that have already been defined structurally and physiologically, and are of clear functional importance: alpha- vs. beta-like (which differ in soma size, dendritic spread and briskness of response) and ON vs. OFF (which differ in inputs, outputs, dendritic stratification, and responsiveness to onset vs. termination of light). To this end, we have already generated and characterized transgenic mice in which the Green Fluorescent Protein (GFP) labels RGCs of multiple types in their entirely, allowing us to identify and isolate them. We will prepare and characterize cDNA from individual transgene-positive RGCs, using methods we have recently used successfully to characterize chick RGCs. The probes will then be hybridized to commercially available microarrays (Affymetrix GeneChips), allowing us to assess expression of about 27,000 genes in cells of each of four classes (alpha-ON, alpha-OFF, beta-like-ON, beta-like-OFF). The output will be analyzed to identify genes expressed by defined RGC subsets or that define novel subsets. Finally, candidate markers will be validated by in situ hybridization. We believe these experiments will (a) provide markers for developmental studies, (b) uncover genes that themselves play roles in subtype diversification or function, and (c) contribute to development of generally applicable methods for gene expression profiling from single cells.
