PHOTORECEPTOR DEVELOPMENT

Takahiko Matsuda (Postdoctoral Fellow)

Mark Emerson (Postdoctoral Fellow)

Karolina Mizeracka (Graduate Student)

Joe Corbo (Former Postdoctoral Fellow, now at Washington University)

Photoreceptors mediate the first step in vision, capturing light and carrying out phototransduction, ultimately resulting in a neural signal to the brain. Rod photoreceptors are active in dim light, while cone photoreceptors are active in bright light and initiate color vision. We are interested in how these cells are produced from multipotent retinal progenitor cells, and how rods and cones diversify relative to each other. In addition, we are interested in how they are patterned across the retina (see section on Retinal Patterning). Our interest in these questions is in the basic mechanisms, as well as in the potential therapeutic applications. The generation of photoreceptor cells from stem cells is being carried out in several laboratories for possible future therapies in diseases where photoreceptors malfunction and/or die. Knowledge of the steps required to make rods and cones is required for successful genesis of such cells for engraftment.

Our lineage studies showed that rods and cones are made by multipotent progenitor cells. Even in a terminal division, when a progenitor cell divides to make 2 postmitotic daughters, one cell might be a rod and one might be a bipolar cell. This observation suggests that the fate decision to be a rod is made either in the progenitor cell, and then handed down asymmetrically to the daughter cells, or it is made by newly postmitotic daughter cells. To better understand which of these possibilities pertains, we have been studying both the progenitor cells and the newly postmitotic cells. Single cell profiling of both types of cells has been carried out (see section on Progenitor Cell Heterogeneity). In addition, perturbation of genes expressed in progenitor cells, and/or in postmitotic daughter cells, is being pursued. Figure 1 shows our current understanding of some of the genes that are crucial in the decision to be a photoreceptor cell, and the decision to be a rod and a cone. The earliest step is in the mitotic multipotent cell, where the Notch receptor is active. This receptor represses formation of photoreceptors, as shown by its removal in a conditional knock-out mouse (Jadhav et al. 2006). When removed early, primarily cone photoreceptors were made. When it was removed later in development, when rods and not cones are normally made, only rod photoreceptors were made. Also in mitotic progenitor cells, in G2 of the cell cycle in a progenitor that is about to make a postmitotic daughter, the Otx2 transcription factor is made (Trimarchi et al. 2008). The Furukawa lab showed that loss of this gene led to a transfating of photoreceptor cells to amacrine interneurons (Nishida et al. 2003). Crx, a close relative of Otx2, is also crucial, but for differentiation, not determination, as loss of this gene led to a lack of differentiation of determined rods and cones (Furukawa et al. 1999). The laboratory of Anand Swaroop showed that Nrl, a transcription factor expressed only in rods, was required for rods to develop, as loss of Nrl led to the transfating of rods to cones (Mears et al. 2001). In part, Nrl’s action is carried out by NR2E3, another rod-specific transcription factor. We found that loss of NR2E3 led to the derepression of some cone genes in rods (Corbo and Cepko, 2005). Finally, the bHLH transcription factor, NeuroD, is required for rod survival subsequent to differentiation (Morrow et al. 1999). (Other genes are required as well, such as Rb, but are not included here for simplicity.)

Figure 1. Transcription factors important in photoreceptor development and differentiation, and the point in development where their action is required, as shown by analysis of knock-out mice.

We carried out microarray and SAGE analysis of retinas from mice in which one of more of these transcription factors have been inactivated (Livesey et al. 2000; Blackshaw et al. 2004; Corbo et al. 2005, 2007; Hsiau e tal 2007). These data were analyzed for changes in rod and cone genes. Genes whose expression levels changed were also analyzed for conserved sequences that might be responsible for cis regulation. Sequences that are activated by Crx and Nrl were recognized near the start site for many of the misregulated genes. Some of the misregulated genes are transcription factors (Figure 2). Many were photoreceptor-specific or enriched genes. From these data, it appears that Crx and Nrl directly regulate many of the photoreceptor differentiation genes, as well as a set of transcription factors that likely also regulate some of the photoreceptor transcription factors. The network can be portrayed in its simplest form in Figure 3.

Figure 2. Microarray analysis of knock-out mice for Crx and Nrl showed that they are at the top of a hierarchy of transcription factors expressed in photoreceptors during development (See Corbo and Cepko, 2005 and Hsiau et al. 2007.)

Figure 3. The microarray analysis conducted on the Crx, NR2E3 and Nrl knock-out mice showed up and down-regulation of photoreceptor-enriched and specific genes in rods, as indicated.

Another approach to understanding the transcription factor network that regulates photoreceptor specific genes, and to gain a deeper appreciation for rod vs. cone genesis, we have directly analyzed the cis regulatory sequences for rod and cone-specific genes. To identify enhancer elements, we first locate phylogenetically conserved noncoding sequences located near known early photoreceptor expressed genes (e.g. Figure 4). These sequences are then cloned into a reporter plasmid with a minimal promoter upstream of EGFP ires AP and tested in vivo or in vitro for specific expression of one of these two reporters in photoreceptors. The DNA is delivered via electroporation and cell type specificity scored by cell morphology, position, and/or immunohistochemistry using markers for particular cell types (e.g. Figure 5). Positive enhancers can then be serially mutagenized to identify the precise sequences that are required for proper expression. Using this approach, we hope to functionally analyze enough enhancer elements to distill from this collection the precise DNA parameters for photoreceptor specific expression. In addition, as we are targeting some of the earliest genes expressed specifically in rods and/or cones, we hope to learn how rods and cone cell fate determination occurs.

Figure 4. Regulatory sequences are often phylogenetically conserved. The upstream regions of phylogenetic conservation are shown here for the photoreceptor gene, IRBP (RBP3).

Figure 5. Plasmids encoding reporters were electroporated at embryonic day 5 into the developing chick retina. One plasmid used photoreceptor-specific regulatory elements to drive expression of GFP, while a control plasmid used a broadly active regulatory region and promoter to drive expression of lacZ, to mark cells that took up the plasmids. The tissue was cultured for 2 days and then stained with anti-visinin, an antibody for photoreceptor cells.