Thesis (Ph. D.)--University of Washington, 2004. / Vita. Includes bibliographical references.

No description available.

published_or_final_version / Physiology / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / toc / Zoology / Master / Master of Philosophy

published_or_final_version / abstract / Paediatrics and Adolescent Medicine / Doctoral / Doctor of Philosophy

published_or_final_version / Medicine / Master / Master of Philosophy

Ovarian cancer is a highly metastatic disease having a poor prognosis (<25%). The factors and underlying mechanisms that regulate ovarian cancer metastasis, however, are still incompletely understood. p70 S6 kinase (p70S6K), a serine/threonine kinase, is frequently activated in high-grade malignant human ovarian cancer. The aim of this study is to investigate the molecular mechanisms by which p70S6K may promote ovarian cancer metastasis. The results show that p70S6K is a critical regulator of the actin cytoskeleton, peritoneal adhesion and dissemination, and multicellular aggregates/spheroids formation in the acquisition of the metastatic phenotype. The regulation of p70S6K on the actin cytoskeleton is through two important functions: as an actin cross-linking protein and as a Rho family GTPase-activating protein. Ectopic expression of constitutively active p70S6K induced a marked reorganization of the actin cytoskeleton and directional migration of ovarian cancer cells. Actin binding and immunofluorescence studies showed that p70S6K had a direct interaction with the actin filaments with no other proteins involved. This interaction did not affect actin polymerization kinetics but cross-linked the actin filaments to inhibit cofilin-induced actin depolymerization. In addition, p70S6K mediated the activation of Rac1 and Cdc42 GTPases and their downstream effector p21-activated kinase 1, but not RhoA. Peritoneal adhesion and dissemination is regulated by p70S6K through integrin expression. Expression of p70S6K siRNA efficiently inhibited ovarian cancer cell adhesion to fibronectin and laminin among different peritoneal extracellular matrix components, as well as to human primary peritoneal mesothelial cells. These effects were associated with the expression of alpha5 and beta1 integrin. Studies into the mechanisms suggest that p70S6K may upregulate alpha5 integrin by a transcriptional mechanism whereas beta1 integrin is regulated at a post-transcriptional level. Enhanced expression of alpha5 and beta1 integrin by active p70S6K mediated the subsequent peritoneal adhesion. In ovarian cancer xenografts, p70S6K and beta1 integrin interference significantly inhibited peritoneal dissemination through reduction in the number and weight of tumors. Multicellular spheroids are present in the malignant ascites of ovarian cancer patients. Using a 3-dimensional culture system, expression of p70S6K siRNA resulted in inhibition of multicellular spheroid formation, which was mediated by N-cadherin but not E- or P-cadherin. In addition to spheroid formation, inhibition of p70S6K was associated with reduced growth of spheroids and disaggregation capabilities on different extracellular matrix components. Taken together, these findings indicate that p70S6K plays an important role in the biology of ovarian cancer metastasis through regulation of several critical steps in dissemination and migration, suggesting that p70S6K could be explored as a potential therapeutic target in ovarian cancer. / published_or_final_version / Biological Sciences / Doctoral / Doctor of Philosophy

Protein kinases constitute essential biological and target class owing to the vital function of reversible phosphorylation catalyzed by these enzymes. With more than 500 kinases in the human genome, containing conserved structure and overlapping function, pose challenging targets for inhibition. Alternative methods for targeting protein kinases remain warranted as the traditional methods are biased toward ATP-competitive compounds. These methods have yielded successful therapeutics, however toxicity due to nonselectivity and limited development potential due to intense drug discovery efforts renders alternative modes of action attractive as new goals for protein kinase inhibition.Herein is presented a bivalent methodology for targeting protein kinases comprising staurosporine tethered a phage display cyclic peptide library such that the cyclic peptide is directed to areas on the kinase surface distinct from the ATP-site where staurosporine is bound. Presented in detail is this strategy as it was successfully applied to Protein Kinase A and the subsequent analysis of bivalent ligands. Since this initial study several kinases have been targeted with this methodology and Application to Aurora Kinase A will be explored in detail. An essential analysis of results to date is included as it applies to the redesign, construction, and application of new cyclic phage libraries. Finally, to complete the first successful application against Protein Kinase A, we explore kinase expression for structural studies.

Protein kinases function as key regulators in a variety of signaling pathways by executing the phosphorylation of a variety of protein substrates. Perturbation in the activity of numerous proteins kinases has been implicated in a large number of diseases including cancer, diabetes, inflammation and neurological disorders. Therefore, selective modulation of kinase activity is highly desirable for the dissection of complex signaling pathways and substantiating therapeutic targets. To develop potent and selective inhibitors for an array of kinases, our group has developed a fragment based bivalent methodology utilizing phage display. The strategy involves an ATP active site targeted small molecule which directs the selection of cyclic peptides, from a phage displayed library, on the target kinase surface through coiled coil interactions. The selected cyclic peptides can be conjugated to the ATP mimetic to generate bivalent inhibitors. In this thesis, I have expanded the scope of the bivalent phage-display selection approach. To interrogate the generality of this approach, we targeted several kinases from different groups within the human kinome using the staurosporine warhead. Fyn and PDGFRβ represented the tyrosine kinase group and CLK2 and Pim-1 kinases represented the CMGC and CaMK groups respectively. The selections against these four kinases did not result in potent inhibitors though they provided an avenue for the refinement of the bivalent phage-display approach as well as method development. Application of this methodology to AKT2 in the AGC family resulted in bivalent inhibitors which were interrogated for their selectivity and mode of action. The bivalent strategy was further explored for its utility to target inactive kinases, and success was achieved against AKT1. Finally, we demonstrated the modularity of ATP site targeted ligand by carrying out a selection against STK33 kinase using a new small molecule warhead, sunitinib. This resulted in potent and selective bivalent inhibitors for STK33. The use of different ATP site targeting molecules potentially increases the number of targetable kinases with our strategy. In all the selections, the identified cyclic peptides inhibited the kinase and showed a non-competitive mode of inhibition with respect to the kinase substrate. This suggests that the selected peptides do not target the substrate site and possibly bind to unidentified pockets on the kinase surface, which potentially provides new methods to target kinases outside the traditional ATP binding cleft. The strategy may prove to be a robust method to discover new allosteric sites on kinases as well as other proteins. The potent and selective bivalent inhibitors obtained by our strategy have the potential to provide insight towards the design of new non-ATP targeted approaches for inhibiting protein kinases and elucidating their specific functions.

No description available.