I. Clinical Studies of Uveal Coloboma Since initiating this research, I have recruited and examined 90 famlies where at least one member is affected by uveal coloboma. All probands and their first degree relatives (when available) have complete ophthalmic exams. General physical examinations and targeted systemic testing (e.g., renal ultrasounds, echocardiograms) were performed on probands, as needed. Lymphoblastoid cell lines were established on all participants for candidate gene analysis. Preliminary analysis of these patients has resulted in four important clinical observations. First, patients with coloboma often have thickened corneas as part of their developmental defect. This affects the measurement of their eye pressures. Second, I have identified a new syndrome in which missing thoracic and/or lumbar vertebrae co-segregates with coloboma. This was seen in two, unrelated autosomal dominant pedigrees. In collaboration with Dr. Fielding Hejmantcik, we have performed initial genetic linkage analysis in these families and isolated a candidate locus on chromosome 1. Third, I have determined the yield of systemic testing on patients with apparent isolated coloboma. Fourth, I have defined microforms of coloboma that aid in genetic counseling of families. II. Laboratory Studies of Uveal Coloboma A. Mouse Models of Coloboma. 1. A novel Pax2 mutant mouse model of coloboma. My lab identified and characterized a mouse model of autosomal dominant congenital optic nerve excavation. Linkage analysis suggested Pax2 as a candidate gene and sequencing studies revealed a missense mutation predicted to change a highly-conserved threonine to alanine in the paired domain of the Pax2 gene. Pax2 is dynamically expressed at the closing edges of the optic fissure and homozygous mutation results in uveal coloboma. This mutation is particularly interesting in that it is in one of the few residues where missense mutations have been reported in patients with papillorenal syndrome. We have characterized the effect this and other missense mutations have on Pax2 function, including transactivation, subcellular localization, and protein expression studies. A manuscript has been submitted to PLoS Genetics detailing this work, with favorable reviews. We nearly complete revisions and anticipate publication in the near future. 2. The RICO Mouse Model of Coloboma The RICO mouse arose from the random insertion of a transgene (NSE-VEGF) on chromosome 13 of C57BL/6 mice. Although the transgene is not expressed in the adult or embryonic mouse, its insertion led to a classic coloboma phenotype. The region of insertion did not suggest any previously-identified coloboma genes. This phenotype is semi-dominant and does not significantly affect the viability or fertility of mice. Tiling arrays of chromosome 13 show no evidence of major deletion. Pax2 expression is not affected in RICO embryos, indicating that the responsible gene is either independent or downstream of Pax2. We have narrowed the region of insertion to approximately 15MB and have also constructed a BAC library created from RICO mice and isolated two positive clones containing the transgene. As an ongoing project, we are sequencing the smaller of these two clones to determine the precise site of integration. B. Identification of coloboma candidate genes by molecular characterization of gene expression during optic fissure closure. 1. Zfp503 and Zfp703 We have performed laser capture microdissection of the tissue surrounding the optic fissure in wild-type mice at three time points corresponding to before, during and after closure. After amplification of the resulting RNA, microarray expression analysis revealed 168 annoted genes and 54 predicted/hypothetical genes, several of which have been verified with RT-PCR and in situ hybridization. Analysis of these data with Ingenuity Pathways software identified 10 high priority targets for further investigation. Two such genes were the zinc-finger proteins, Zfp503 and Zfp703. Relatively little is known about these genes, other than they appear to be important in early development of the zebrafish midbrain and may act as repressors of transcription. When we knocked-down expression using a morpholino strategy in zebrafish, we were able to induce coloboma (Figure 2). Histology reveals retinal lamination defects &persistence of fetal vasculatureboth phenotypes observed in humans with coloboma. While knockdown of Zfp703 results in reduced Pax2 expression, Zfp503 knockdown increases Pax2 expression. We have shown that this effect has some specificity, in the other markers of ventral eye development (e.g., Vax1 and Vax2) are not affected. Furthermore, we have shown, using chromatin immunoprecipitation, that both Zfp703 and Zfp503 are capable of binding at a DNA segment approximately 2kb upstream of the Pax2 initiation site. We have shown in vitro that Zfp703 and Zfp503 are able to regulate baseline Pax2 transcription. Because this site is adjacent to putative binding sites for transcription factors known to be important in ventral eye development (e.g., a Smad binding site for the bone morphogenetic protein BMPpathway and a retinoic acid receptor binding site), we have been investigating whether Zfp703 and Zfp503 modulation of Pax2 transcription is, in fact, affected by signaling through the BMP and retinoic acid pathways. Lastly, we have screened our patient DNA samples for mutations in ZFP703 and ZFP503. We have identified a c.C919T sequence change in ZFP703 that is predicted to change a well-conserved proline residue to a serine. Molecular modeling in collaboration with Dr. Yuri Sergeev shows that this change occurs in a proline-rich region of the protein, and is expected to disrupt the tight packing a alpha helices. Furthermore, preliminary in vitro assays of mutant protein function indicate that its regulation of Pax2 transcription differs from that of the wild-type protein. Whereas wild-type zfp703 mRNA partially rescues the morpholino phenotype in our zebrafish model, mutant mRNA does not. All these data point to a functional consequence of this mutation and that this allele is a risk factor for coloboma in patients. A manuscript describing this work is in preparation. 2. FAT1 and FAT4 Another candidate gene that seemed promising from our expression array analysis was FAT4, a protocadherin cell adhesion molecule important in the development of cell polarity. We have characterized the expression of Fat4 and the three other Fat molecules (Fat1, Fat2, and Fat3) in mouse eye development. Both Fat1 and Fat4 mRNA are expressed in a polarized fashion in primary lens fiber cells and in the neuroepithelium at the presumptive retina/retinal pigment epithelium junction. Specifically, we see notable expression at the closing edges of the optic fissure. We have obtained knock-out mice for both of these mice and are currently characterizing their ocular development.
Showing the most recent 10 out of 17 publications
