Characterizing Molecular Biomarkers in Cancer as a Tool for Prognostication and for Identification of Pathways that are Amenable to Personalized Therapy

Identifying and characterizing molecular targets in cancer that have the potential to aid in clinical prognosis and selecting targeted drug therapies were examined in this thesis. The first part of the thesis involved in silico discovery and molecular validation of putative urinary biomarkers that are able to predict survival outcomes in patients with urothelial carcinoma (UC) after neoadjuvant chemotherapy. Using this approach, gammaglutamyl hydrolase (GGH) and diazepam-binding inhibitor (DBI) were identified as being present in urine and highly expressed in UC. Using disease-free survival as a marker for clinical outcome, the combination of GGH and DBI was able to significantly stratify disease free survival (P=0.024) after neoadjuvant chemotherapy suggesting that this two-gene signature may putatively be utilized clinically to assess prognostic outcomes in UC patients. The second part of the thesis examined the relationship between c-Src and the cyclin-dependent kinase (CDK) inhibitor p21. Increased expression and/or activation of the tyrosine kinase c-Src has classically been associated with increased aggressiveness in cancer cells. Interestingly, it was demonstrated that c-Src expression in the UMUC3 cell iii line had a direct correlation with p21 expression suggesting that c-Src could act as a suppressor of proliferation in UC via increased p21 expression. Importantly, these data also suggested that c-Src inhibitors would be contraindicated in UC patients with high cSrc expression. However, additional studies demonstrated no significant growth phenotypes through modulation of c-Src levels. Thus, while c-Src is likely important in regulating p21 expression in UC, the phenotypic impact of this relationship is likely subtle. The final part of the thesis involved utilizing a genome wide functional short hairpin RNA (shRNA) screen aimed at identifying suppressors of in vivo tumor growth. This screen revealed amylo-alpha-1-6-glucosidase-4-alpha-glucanotransferase (AGL), a glycogen debranching enzyme, as a novel tumor suppressor. A reduction in AGL levels concomitantly increased cellular glucose uptake, enhanced lactate production and increased lipid biosynthesis. These findings demonstrate for the first time that disruption of glycogen metabolism results in aggressive tumor phenotypes via dependence on lipid biosynthesis/catabolism and extracellular glucose uptake. Hence, these pathways represent new potential therapeutic targets in UC patients with low expression of AGL.

Note: Abstract extracted from PDF text