Systematic proteomic strategies to map the human signaling landscape

Abstract

Signal transduction is the process by which cells continuously sense and integrate environmental cues in order to make real-time decisions, e.g. how to direct metabolic flux or whether to enter mitosis. Constructing detailed maps of cellular signaling networks has proved a valuable way to summarize knowledge, formulate new hypotheses, and devise pharmacological control strategies for a myriad of human diseases. However, despite remarkable progress in this area, only recently have we begun to appreciate the true vastness and diversity of signaling landscapes. The capability to measure molecular systems at near genome scale represents a major paradigm shift for biology, and mass spectrometry (MS) approaches can now provide multiplexed quantitative measurements of thousands of cellular proteins and phosphorylation events in a single sample. In this dissertation, I use MS-proteomics to study protein networks in cultured human cell lines. First, I describe an integrative proteomic analysis of twenty breast cancer cell lines. In this work I show that protein expression varies dramatically across cells derived from the same tissue type and across several canonical signaling pathways. I identify distinct patterns of protein expression that are found in triple negative breast cancer cells compared to luminal breast cancer. Further, I suggest that genetic aberrations and protein expression are interconnected, and together they affect the responsiveness of cells to cancer therapeutics. Next, I used mass spectrometry to investigate phosphorylation-dependent signaling networks. I first performed a deep characterization of protein phosphorylation events in HeLa cells exposed to sixteen different stimuli (e.g. epidermal growth factor, tumor necrosis factor-alpha, osmotic stress), quantifying more than one hundred thousand phosphorylation sites. In this study I provide a detailed view of the vast signaling landscape present within an individual cell type and reveal the extent of regulatory cross-talk, whereby the same phosphorylation sites and proteins are regulated by multiple stimuli. To further dissect the topology, dynamics, and cross-talk of pathways frequently mutated in cancer I performed additional experiments using time-courses and systematic kinase inhibition against several key signaling nodes in the presence of growth factors and cellular stress in HeLa and MCF7 cells. Finally, I use the data collected during these experiments to evaluate and address several technical limitations of mass spectrometry-based phosphoproteomics analysis. I discuss a method to mine these large-scale data to rapidly generate targeted MS assays which now enable versatile, high-throughput, sensitive, and reproducible analysis of cellular signal transduction networks.