Bioinformatic analysis ofp53 response in vivo

p53 is a tumor suppressor protein that transmits stress response stimuli to the cell fate control mechanisms. It is a transcription factor that acts primarily through activation or repression of its target genes.; In order to discover novel functions of p53, we identified by microarray hybridization the p53-regulated genes in human and mouse cells of different p53 status and tissue origin.; Several of the new p53-activated genes encode known secreted growth inhibitory factors. This suggests that p53 can cause “bystander effect” by inducing export of growth suppressive stimuli from damaged cells to neighboring cells. Consistently, after gamma irradiation a p53-dependent accumulation of growth inhibitory factors was observed in the cell culture media and in the urine of irradiated mice. This indicates a previously unknown role for normal cells in chemo- and radiation therapy of cancer.; Among p53-repressible genes in prostate cell line LNCaP we found prostate-specific antigen (PSA) that is broadly used for early diagnostics of prostate cancer. wt p53 strongly repressed, while dominant negative p53 mutants stimulated PSA promoter driven transcription and secretion of PSA. These findings helped to explain the upregulation of PSA in prostate tumors and lead us to an idea to look for new “PSA-like” tumor markers among the genes that are downregulated by tumor suppressors.; We designed an extensive program of identification of such marker genes. Firstly, mice deficient in tumor suppressor (p53, PTEN pRB) function in prostate were analyzed by microarray hybridization. Tumor suppressor-repressed genes were identified, and their human homologs were selected as potential markers. Secondly, EST database mining was used to select genes whose expression pattern satisfied reliability and specificity criteria required of diagnostic markers. cDNA microarray carrying the candidate genes was screened with probes derived from cells of normal prostate, prostate tumor and other origin. Genes with prostate tumor-specific expression were further validated by hybridization with tissue microarrays containing the specimens of normal prostate and prostate cancer at various stages of progression. EP14, a novel transcript, was identified by our approach and found to have a prostate tumor-specific expression pattern similar to such of PSA, but with lower expression in normal prostate.