| The majority of cancer related deaths are due to metastasis, highlighting the importance of early detection and intervention. However, in melanoma and other cancers, early stage tumors are clinically heterogeneous, where subsets of patients with no evidence of metastatic spread eventually recur and succumb to the disease. Identifying molecular biomarkers to identify high risk patients at the time of diagnosis is in critical need. To that end, in this dissertation I develop an in situ hybridization method (ISH) to detect and quantify microRNA (miRNA) expression in tissue microarrays to evaluate the hypothesis that miRNAs, in regulating target protein expression, may be a valuable class of biomarkers. I also assess the human relevance of recently identified prometastasis oncogenes and their prognostic value in melanoma.;Robust miRNA detection has been difficult with most methods requiring total RNA extracts, which in the case of tissue samples is often variably contaminated with normal and stromal cells. ISH however, allows the direct assessment of expression in individual cells and importantly allows one to determine the contribution of malignant as well as normal cells to observed changes in expression. In quantifying miRNA expression, the ISH is multiplexed with DAPI and protein immunofluorescence for a tissue specific marker of tumor cells such as cytokeratin or S100. This co-localization based approach uses the AQUA (Automated Quantitative Analysis) technology to establish subcellular compartments within tumor cells resulting in a score directly proportional to the number of molecules per unit area. The miRNA ISH assay is vigorously validated using genetically modified mouse tissue, transfected cell lines, blocking oligos, and assessment of reproducibility. In testing miR-21, miR-205, miR-34a, miR-92a, miR-221, and let-7a, miR-221 is found to have prognostic value in breast cancer independent of age, tumor size, node status, hormone status, and Her2 status. In addition, miR-205 is found to be prognostic in two independent melanoma cohorts supporting its role as a tumor suppressing miRNA in cancer. In assessing the functional implications of miRNA expression we evaluated the relationship between miRNAs and respective protein targets on a large population based cohort analysis. Interestingly the presumed inverse relationship was not discerned for any miRNA-protein pair suggesting this relationship to be more complicated than in vitro cell line models suggest. Nevertheless these studies provide proof of concept for the use of miRNAs as prognostic biomarkers in cancer as detected and quantified using in situ hybridization.;Recently identified prometastasis proteins and the Met oncogene were validated for their human relevance by demonstrating differential expression associated with transformation or progression of melanoma. Furthermore, ACP5 is shown to be another promising prognostic biomarker in melanoma. Finally, a literature review of the histopathology and function of the metastasis promoting c-Met protein in cancer is presented. |
