Cancer signaling networks and their implications for personalized medicine

Abstract

Amongst the unique features of cancer cells perhaps the most crucial one is the change in the cellular decision-making process. While both non-cancer and cancer cells are constantly integrating different external cues that reach them and computing cellular decisions (e.g. proliferation or apoptosis) based on the integration of these cues; this integration and consequently the cellular decisions taken by cancer cells are arguably very distinct from the decisions that would be expected from non-cancer cells. Since cellular signaling networks and its different states are the computational circuits that determine cellular outcome, it is clear to many that these networks will be highly dysregulated in cancer cells. Thus, developing and applying methods that will be capable of mapping and predicting how cancer mutations translate into signaling network perturbations, which could explain cancer development as well as cancer resistance to treatment, represent not only a huge challenge, but also one with potentially extreme benefit for our understanding of the disease and for patients. This thesis summarizes my efforts during the last years in contributing positively to overcome this challenge. This thesis is divided into six parts. Starting with a brief introduction to the history and some basic concepts of cancer, signaling networks and human protein kinases (part I), we quickly move on to describing existing methods to analyze cancer signaling networks, including methods proposed by us, as well as three of the articles that are part of this PhD thesis (part II). In part III, we illustrate with an article that has been submitted recently, how next-generation sequencing data and mass spectrometry data can be combined to uncover genome-specific signaling networks. In part IV, I describe the two computational methods that I have developed and how they can be integrated with the aim of predicting how signaling networks will be dysregulated in cancer. As a matter of fact, the following part (part V) proves the usefulness of the method by identifying a functional mutation in a group of ovarian clear cell carcinoma cell lines that could cause their resistance to cisplatin treatment. Part VI closes the thesis by summarizing its main points and proposing some future perspectives for the work presented here. All in all, this work establishes a new framework for the prediction of mechanisms underlying cancer development and evolution which, one would hope, should help close the gap between cancer genotype and phenotype.