Mapping of transcription start sites of human retina expressed genes.

BACKGROUND: The proper assembly of the transcriptional initiation machinery is a key regulatory step in the execution of the correct program of mRNA synthesis. The use of alternative transcription start sites (TSSs) provides a mechanism for cell and tissue specific gene regulation. Our knowledge of transcriptional initiation sequences in the human genome is limited despite the availability of the complete genome sequence. While genome wide experimental and bioinformatic approaches are improving our knowledge of TSSs, they lack information concerning genes expressed in a restricted manner or at very low levels, such as tissue specific genes. RESULTS: In this study we describe the mapping of TSSs of genes expressed in human retina. Genes have been selected on the basis of their physiological or developmental role in this tissue. Our work combines in silico analysis of ESTs and known algorithm predictions together with their experimental validation via Cap-finder RACE. We report here the TSSs mapping of 54 retina expressed genes: we retrieved new sequences for 41 genes, some of which contain un-annotated exons. Results can be grouped into five categories, compared to the RefSeq; (i) TSS located in new first exons, (ii) splicing variation of the second exon, (iii) extension of the annotated first exon, (iv) shortening of the annotated first exon, (v) confirmation of previously annotated TSS. CONCLUSION: In silico and experimental analysis of the transcripts proved to be essential for the ultimate mapping of TSSs. Our results highlight the necessity of a tissue specific approach to complete the existing gene annotation. The new TSSs and transcribed sequences are essential for further exploration of the promoter and other cis-regulatory sequences at the 5'end of genes.

Identification of genes selectively regulated by IFNs in endothelial cells.

IFNs are highly pleiotropic cytokines also endowed with marked antiangiogenic activity. In this study, the mRNA expression profiles of endothelial cells (EC) exposed in vitro to IFN-alpha, IFN-beta, or IFN-gamma were determined. We found that in HUVEC as well as in other EC types 175 genes were up-regulated (>2-fold increase) by IFNs, including genes involved in the host response to RNA viruses, inflammation, and apoptosis. Interestingly, 41 genes showed a >5-fold higher induction by IFN-alpha in EC compared with human fibroblasts; among them, the gene encoding the angiostatic chemokine CXCL11 was selectively induced by IFN-alpha in EC along with other genes associated with angiogenesis regulation, including CXCL10, TRAIL, and guanylate-binding protein 1. These transcriptional changes were confirmed and extended by quantitative PCR analysis and ELISA; whereas IFN-alpha and IFN-beta exerted virtually identical effects on transcriptome modulation, a differential gene regulation by type I and type II IFN emerged, especially as far as quantitative aspects were concerned. In vivo, IFN-alpha-producing tumors overexpressed murine CXCL10 and CXCL11, guanylate-binding protein 1, and TRAIL, with evidence of CXCL11 production by tumor-associated EC. Overall, these findings improve our understanding of the antiangiogenic effects of IFNs by showing that these cytokines trigger an antiangiogenic transcriptional program in EC. Moreover, we suggest that quantitative differences in the magnitude of the transcriptional activation of IFN-responsive genes could form the basis for cell-specific transcriptional signatures.

Recruitment of human umbilical vein endothelial cells and human primary fibroblasts into experimental tumors growing in SCID mice.

Generation of a vascular network is a hallmark of solid tumor growth, and attempts to switch off the tumor angiogenic phenotype are promising. However, this angiogenic potential might also be exploited to obtain incorporation into tumor vessels of genetically modified third-party cells, which could behave as targets of immunologic or pharmacologic attack. With this in mind, we addressed the efficiency and selectivity of third-party cell recruitment into experimental tumors generated in severe combined immunodeficiency mice. The animals were inoculated intraperitoneally with human ovarian carcinoma cell lines and with beta-galactosidase (beta-gal)-transduced human umbilical vein endothelial cell (HUVEC) or human fibroblasts. Transgenic HUVEC were scattered in tumors, but not in normal mouse tissues; immunohistochemical analysis revealed their selective homing to tumor vascular structures, over 50% of which contained beta-gal(+) cells. Injection of beta-gal-transduced human fibroblasts was also associated with transgenic cell incorporation into tumor masses; however, beta-gal(+) fibroblasts did not home to tumor blood vessels and were only localized within the tumor stroma. These findings show that the recruitment of primary third-party cells into the different compartments of experimentally induced tumors is an efficient and selective phenomenon and indicate possible alternative ways of confronting the tumor angiogenic potential in cancer therapy.

Expression from cell type-specific enhancer-modified retroviral vectors after transduction: influence of marker gene stability.

The enhanced green fluorescent protein (EGFP) is increasingly used as a reporter gene in viral vectors for a number of applications. To establish a system to study the activity of cis-acting cellular regulatory sequences, we deleted the viral enhancer in EGFP-carrying retroviral vectors and replaced it with cell type-specific elements. In this study, we use this system to demonstrate the activity of the human CD2 lymphoid-specific and the Tie2 endothelial cell type-specific enhancers in cell lines and in primary cells transduced by retroviral vectors. Furthermore, we compare findings obtained with EGFP as the reporter gene to those obtained replacing EGFP with d2EGFP, an unstable variant of EGFP characterized by a much shorter half-life compared to EGFP, and by reduced accumulation in the cells. d2EGFP-carrying vectors were generated at titers which were not different from those generated by the corresponding vectors carrying EGFP. Moreover, the activity of a Moloney murine leukemia virus enhancer could be readily detected following transduction of target cells with either EGFP- or d2EGFP-carrying vectors. However, the activity of the relatively weak CD2 and Tie2 enhancers was exclusively detected using EGFP as the reporter gene. These findings indicate that enhancer replacement is a feasible and promising approach to address the function of cell type-specific regulatory elements in retroviral vectors carrying the EGFP gene.