COMPREHENSIVE EXPRESSION PROFILES OF DEVELOPING RETINAL NEURONS

Jeff Trimarchi (Postdoctoral Fellow)

Tim Cherry (Graduate Student)

Doug Kim (Former Postdoctoral Fellow)

As described in the section, “Retinal progenitor cell heterogeneity”, we have been profiling single developing retinal cells on Affymetrix chips. To date, almost 170 cells have been profiled. Comparison of these profiles between cell types, and among cells of a single type, has revealed a large amount of heterogeneity, as well as provided markers for retinal cell types. For example, hundreds of genes have been shown to be expressed in subsets of amacrine and ganglion cells. These data can be mined for the identification of genes to be used as classical markers, as well as for candidate genes that might drive the formation of diversity, or provide starting points for studies of transcription factor networks. For example, an examination of newborn ganglion cells shows many transcription factors, as well as differences in transcription factor expression among the single cells. Differences in genes regulating neurotransmitter metabolism, such as receptors and synthesis genes, and molecules that might differentially direct axonal targeting can also be seen. See Figures 1 and 2 below, as well as Trimarchi et al. 2007.

Figure 1. Single ganglion cells were profiled on Affymetrix chips. The signal from ganglion cells as well as from other retinal neurons, including adult rods, are shown in the columns. The signals>10,000 are shown as bright red, those <1,000 as black, and others are graded. Each row is the signal from the gene listed. A subset of transcription factors are shown. From Trimarchi et al. 2007.

Figure 2. The signals from single ganglion cells are shown for genes that might be involed in the targetting of ganglion axons are shown.

Amacrine cells, a diverse class of inhibitory interneurons, are thought to mediate the majority of the processing of the visual signal that occurs within the retina. Although morphological characterization of amacrine cells has demonstrated extensive diversity among neurons of this class (Figure 3), the molecular nature of this diversity is largely unknown. Furthermore, it is not known how this diversity arises during development. We have performed single cell profiling of 32 single amacrine cells during development (e.g. see Figure 4). The profiles of these cells have been compared to each other, and to other retinal neurons and progenitor cells. These analyses have revealed the molecular diversity of this cell class. In addition to the molecular profiles, we have also been labeling amacrine cells in development using plasmid electroporation (Figure 4). These methods allow genetic access to these cells and reveal details about their morphology during development, as well as allow us to examine promoter elements that drive expression in different amacrine cell types (e.g. see Matsuda and Cepko, 2004).

Figure 3. Amacrine cells have been classified by morphology, revealing great complexity in their arbors. A summary from the Macneil and Masland is shown. Immunohistochemistry against two proteins that characterize paritcular subtypes are also shown in sections.

Figure 4. Single retinal cells from mouse from postnatal days 4 and 5 were profiled on Affymetrix chips.The signals from a selection of single amacrine cells are shown in columns, with each row representing the indicated gene. The amacrine cells all expressed Pax6, as expected. Several were GABAergic, as indicated by expression of GAD1, while others were glycinergic, as seen by expression of GlyT1. In addition, calretinin can be seen in 2 cells, and the marker of AII cells, Dab1, can be seen in 2 cells. Additional anaylses has revealed many diverse markers of amacrine cells. Confirmation of these expression patterns is then carried out by in situ hybridization. From Cherry et al, submitted.

Bipolar cells are another diverse class of interneurons. As with amacrine cells, they have been classified by the arbors, as shown in teh summary in Figure 5. Bipolar cells also have been profiled (Kim et al. 2008). Many new markers have been defined on the basis of the comparisons among bipolar cells, and by comparisons between bipolar cells and other retinal cells. In situ hybridization for some of the bipolar subtype markers are shown (Figure 6). Quantification of the overlap of marker expression using dissociated in situ hybridization have been carried out. A summary of one such analysis for cells that express Bip3, an ion channel, is shown. (From Kim et al. 2008).

Figure 5. Bipolar cells have been classified on the basis of the projections of their dendrites to the inner plexiform layer, their presynatpic partner (rods or cones), as well as their cell body location. A summary from Euler and Wassle (1995) for rat is shown. Several divisions can be made, such as those that project to the ON sublaminae, those that project to the OFF sublaminae, or those that connect to rods or cones.

Figure 6. In situ hybridization on retinal sections from mouse postnatal day 21 with probes for genes defined as being expressed in all bipolar or in subsets of retinal bipolar cells. Chx10 is expressed in all bipolar cells, prkca and pcp2 in rod bipolar cells (as shown previously), and Bip1, 2, 3 are novel markers expressed in subsets of bipolar cells. From Kim et al. 2008.

Figure 7. The expression of bipolar markers was assessed by performing in situ hybridization on dissociated retinal cells. The overlap of gene expression was assessed by in situ hybridization with two probes on the same cell preparation. By performing in situ with several overlapping pairs of probes, the set of cells expressing Bip3, a marker newly disovered by the microarray analysis, can be defined into subsets. A summary of these data are shown. From Kim et al. 2008.