Optimal control models of epithelial-mesenchymal transition for the design of pancreas cancer combination therapy

点击翻译按钮获取中文摘要

PROJECT SUMMARY Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal and common cancer, with an overall five-year survival rate of 6%. Among the factors contributing to this dismal statistic is the observation that epithelial- derived PDAC cells, sometimes in direct response to therapy, can de-differentiate to a mesenchymal state in which they are more chemoresistant. This observation prompts the question: should epithelial-mesenchymal transition (EMT) be targeted to promote therapeutic response and increase patient survival? The main barrier to exploring this idea is that we do not know how to target EMT precisely, especially in light of the complex multivariate cell signaling dynamics that drive EMT and maintain it as a feedback response to chemotherapy. We recently undertook a preliminary study to identify a group of druggable cell signaling pathways that may cooperatively drive the mesenchymal state in PDAC. However, the translational potential of our current analysis is limited in that it merely identified potential targets; it does not provide any systematic actionable understanding, nor testable predictions, of how best to schedule combinations of drugs in time to maximize therapeutic efficacy and minimize unintended toxicity. Consequently, we now seek to extend our preliminary studies to develop a systems biology platform for the systematic determination of scheduled combination therapy approaches for PDAC designed to maximally suppress EMT during treatment. In Aim 1, we will make dynamic measurements of signaling pathway activity and cell phenotypes in PDAC cells treated with drivers of EMT, antagonists of EMT, and chemotherapeutics. Our measurements will cover those pathways already identified in our preliminary work as the most likely druggable regulators of EMT, and will include the effects of hypoxia and cancer-associated fibroblasts, elements of the tumor microenvironment that may impact EMT regulation. The goal is to obtain an information-rich data set to be used subsequently for model identification and control computations. In Aim 2, we will use the dynamic data to develop the computational platform for determining optimal changes to the drivers and antagonists required to achieve maximal suppression of EMT, to be implemented as scheduled combination therapies for PDAC. This will be accomplished through: (i) identification of a dynamic model for epithelial or mesenchymal cell state determination in response to phosphoprotein perturbations (i.e., quantitative characterization, in the form of a computational model, the EMT response to changes in its drivers and antagonists) and (ii) deploying the model “in reverse” to determine, via optimal control principles, how best to combine and schedule drugs for optimal maintenance of the epithelial phenotype. In Aim 3, we will test the model-based schedules for combination therapy in a sequence of in vitro and in vivo experiments. Ultimately, these studies will provide pre-clinical validation for a new strategy to develop therapeutic regimens that target a pathological process in PDAC that limits therapeutic response. New approaches are urgently needed, as PDAC survival rates have not changed in nearly 40 years.