Transcriptional profiling develops molecular signatures for ovarian tumors

Of the cancers unique to women, ovarian cancer has the highest mortality rate. Over 26,000 women are diagnosed with this disease in the U.S. annually, and 60% of those diagnosed will die of the disease. One of the greatest problems with this disease is the lack of strong early warning signs or symptoms resulting in advanced stage at presentation in most women, followed by the outgrowth of chemotherapy-resistant disease in the majority of patients. The 5-year survival for patients with early stage disease ranges from 50-90%, but it is less than 25% for advanced-stage disease. In collaboration with researchers at Millennium Predictive Medicine (Cambridge, MA), the Ovarian Cancer Program of the Mayo Clinic Cancer Center analyzed gene expression in over 50 primary ovarian tumors, as compared with normal ovarian epithelial cells. The technologies utilized included microarray analysis with nitrocellulose filters containing 25,000 arrayed human cDNAs, as well as the construction of subtraction suppression hybridization cDNA libraries and their subsequent sequencing. Our specific focus has been on genes that are underexpressed during the development of ovarian cancer, although this analysis has revealed a large number of consistently up- and down-regulated genes. There were more down-regulated genes in ovarian tumors than up-regulated genes. In addition, the number of genes that had altered expression levels was quite large. For example, we found 409 genes down-regulated at least 5-fold, and 72 genes up-regulated at least 5-fold in 33% of the tumors analyzed. We also observed that most of the expression alterations observed in later stage (Stages III/IV) tumors were also observed in early-stage tumors (Stages I/II). This was corroborated using comparative genomic hybridization analysis on the same tumors that were expression profiled. This analysis revealed that the late-stage tumors had more gene amplification than early-stage tumors, but most regions of change (either increases or decreases) were in common between different stage tumors. We also have verified the altered expression levels of several of these genes using several complementary strategies. Finally, we are taking top candidate genes that are consistently under-expressed in ovarian tumors and attempting to determine their functional role in the development of ovarian cancer.