Ovarian cancer results from the dysregulation, in normal ovarian epithelial cells, of genes responsible for the control of critical biological processes. Since their discovery 20 years ago, microRNAs have increasingly been implicated in that dysregulation due to their role mediating gene expression; changes in microRNA expression levels in cancer have been linked with tumor growth, proliferation and metastasis. Their imputed involvement in cancer has led to the possibility of their use as biomarkers and to their potential clinical use.
Using mRNA and microRNA microarray analysis to compare human gene expression in normal ovarian surface epithelial (OSE) cells and epithelial ovarian cancer (EOC) cells, we explored the interactions between microRNAs and genes. First, we validated in silico predictions of microRNA targets by comparing them with in vitro evidence after exogenous microRNA transfection. We found that pairs of microRNAs with identical 7-nt (nucleotide) seed regions shared 88% of their predicted targets and 55% of their in vitro targets, confirming the importance of the seed as a targeting mechanism. But more importantly, we found that even a single nucleotide change in the seed region can result in a significant shift in the set of targeted genes, implying strong functional conservation of the seeds and their corresponding binding sites.
Next, we discovered a 3-element network motif which explains the upregulation of nearly 800 genes in ovarian cancer which, as predicted microRNA targets, might be expected to be down- regulated. This model shows that, under certain circumstances, repressor genes which are down- regulated in cancer can apparently override the repressive effects of microRNAs, resulting in the upregulation of predicted microRNA targets.
Finally, we developed a phenomenological network model, based on the Pearson correlation of microarray gene expression data, to identify subnetworks dysregulated in cell cycle and apoptosis. While our methodology reported many genes previously associated with ovarian cancer, it significantly suggested potentially oncogenic genes for further investigation. This network model can easily be extended to identify dysregulated genes in other cancers.
| Identifer | oai:union.ndltd.org:GATECH/oai:smartech.gatech.edu:1853/53086 |
| Date | 12 January 2015 |
| Creators | Hill, Christopher G. |
| Contributors | McDonald, John |
| Publisher | Georgia Institute of Technology |
| Source Sets | Georgia Tech Electronic Thesis and Dissertation Archive |
| Language | en_US |
| Detected Language | English |
| Type | Dissertation |
| Format | application/pdf |
Page generated in 0.002 seconds