Understanding gene regulatory networks (GRNs) that control neuronal differentiation will provide systems-level perspectives on neurogenesis. We have previously constructed a model for a GRN in retinal ganglion cell (RGC) differentiation in which four hierarchical tiers of transcription factors ultimately control the expression of downstream terminal genes. Math5 occupies a central node in the hierarchy because it is essential for the formation of RGCs and the expression of the immediate downstream factor Pou4f2. Based on its expression, we also proposed that Isl1, a LIM-homeodomain factor, functions in parallel with Pou4f2 and downstream of Math5 in the RGC GRN. To determine whether this was the case, a conditional Isl1 allele was generated and deleted specifically in the developing retina. Although RGCs formed in Isl1-deleted retinas, most underwent apoptosis, and few remained at later stages. By microarray analysis, we identified a distinct set of genes whose expression depended on Isl1. These genes are all downstream of Math5, and some of them, but not all, also depend on Pou4f2. Additionally, Isl1 was required for the sustained expression of Pou4f2, suggesting that Isl1 positively regulates Pou4f2 after Math5 levels are diminished. The results demonstrate an essential role for Isl1 in RGC development and reveal two distinct but intersecting branches of the RGC GRN downstream of Math5, one directed by Pou4f2 and the other by Isl1. They also reveal that identical RGC expression patterns are achieved by different combinations of divergent inputs from upstream transcription factors.cell differentiation ͉ gene regulatory network ͉ transcription factors ͉ retinogenesis R etinal development is especially attractive for gene regulatory network (GRN) analysis because it represents a relatively simple, highly amenable sensory tissue with a small number of neuronal cell types connected to visual centers in the brain by the optic nerve (1, 2). A canonical GRN has been proposed for retinal determination genes in insects and mammals that is composed of seven to eight highly conserved transcription factors (3). In the mouse, one of the retinal determination factors, the pair-rule homeobox factor Pax6, is essential for the formation of all retinal cell types with the exception of amacrine cells (4). Pax6 is required in retinal progenitor cells (RPCs) for the expression of several critical transcription factors that then control the development of individual retinal cell types. One of these factors is Math5, a proneural basic helix-loop-helix factor that is essential for specifying retinal ganglion cells (RGCs) (5, 6), the first cells to differentiate in the developing retina. Pax6-Math5-expressing RPCs define a competence field that permits the subsequent steps of RGC formation to proceed (7,8). Based on these facts, we have constructed a GRN model for RGC development that features a downward cascade of transcription factors occupying distinct hierarchical tiers (8).In the RGC GRN, Math5 activates genes that encode immediat...

Math5, a mouse homolog of the Drosophila proneural bHLH transcription factor Atonal, is essential in the developing retina to establish retinal progenitor cell competence for a ganglion cell fate. Elucidating the mechanisms by which Math5 influences progenitor cell competence is crucial for understanding how specification of neuronal cell fate occurs in the retina and it requires knowledge of the downstream target genes that depend on Math5 for their expression. To date, only a handful of genes downstream of Math5 have been identified. To better define the gene network operating downstream of Math5, we used custom-designed microarrays to examine the changes in embryonic retinal gene expression caused by deletion of math5. We identified 270 Math5-dependent genes, including those that were expressed specifically either in progenitor cells or differentiated ganglion cells. The ganglion cell-specific genes included both Brn3b-dependent and Brn3b-independent genes, indicating that Math5 regulates distinct branches of the gene network responsible for retinal ganglion cell differentiation. In math5-null progenitor cells, there was an up-regulation of the proneural genes math3, neuroD, and ngn2, indicating that Math5 suppresses the production of other cell types in addition to promoting retinal ganglion cell formation. The promoter regions of many Math5-dependent genes contained binding sites for REST/NRSF, suggesting that release from general repression in retinal progenitor cells is required for ganglion cell-specific gene activation. The identification of multiple roles for Math5 provides new insights into the gene network that defines progenitor cell competence in the embryonic retina.

During early pregnancy in ruminants, progesterone (P4) from the corpus luteum and interferon tau (IFNT) from the conceptus act on the endometrium to regulate genes important for uterine receptivity and conceptus growth. The use of the uterine gland knockout (UGKO) ewe has demonstrated the critical role of epithelial secretions in regulation of conceptus survival and growth. A custom ovine cDNA array was used to identify alterations in gene expression of endometria from Day 14 cyclic, pregnant, and UGKO ewes (study 1) and from cyclic ewes treated with P4 or P4 with ZK 136,317 antiprogestin and control proteins or IFNT (study 2). In study 1, expression of 47 genes was more than 2-fold different between Day 14 pregnant and cyclic endometria, whereas 23 genes was different between Day 14 cyclic and UGKO endometria. In study 2, 70 genes were different due to P4 alone, 74 genes were affected by IFNT in a P4-dependent manner, and 180 genes were regulated by IFNT in a P4-independent manner. In each study, an approximately equal number of genes were found to be activated or repressed in each group. Endometrial genes increased by pregnancy and P4 and/or IFNT include B2M, CTSL, CXCL10, G1P3, GRP, IFI27, IFIT1, IFITM3, LGALS15, MX1, POSTN, RSAD2, and STAT5A. Transcripts decreased by pregnancy and P4 and/or IFNT include COL3A1, LUM, PTMA, PUM1, RPL9, SPARC, and VIM. Identification and analysis of these hormonally responsive genes will help define endometrial pathways critical for uterine support of peri-implantation conceptus survival, growth, and implantation.

Brn3b/Brn-3.2/POU4f2 is a POU domain transcription factor that is essential for retinal ganglion cell (RGC) differentiation, axonal outgrowth and survival. Our goal was to establish a link between Brn3b and the downstream events leading to RGC differentiation. We sought to determine both the number and types of genes that depend on Brn3b for their expression. RNA probes from wild-type and Brn3b-/- E14.5, E16.5 and E18.5 mouse retinas were hybridized to a microarray containing 18,816 retina-expressed cDNAs. At E14.5, we identified 87 genes whose expression was significantly altered in the absence of Brn3b and verified the results by real-time PCR and in situ hybridization. These genes fell into discrete sets that encoded transcription factors, proteins associated with neuron integrity and function,and secreted signaling molecules. We found that Brn3b influenced gene expression in non RGCs of the retina by controlling the expression of secreted signaling molecules such as sonic hedgehog and myostatin/Gdf8. At later developmental stages, additional alterations in gene expression were secondary consequences of aberrant RGC differentiation caused by the absence of Brn3b. Our results demonstrate that a small but crucial fraction of the RGC transcriptome is dependent on Brn3b. The Brn3b-dependent gene sets therefore provide a unique molecular signature for the developing retina.

No abstract