Role of systems biology in chemoprevention drug identification and implementation

ED07-01

Systems biology is the study of the emergence of functional properties that are present in a biological system but that are not obvious from a study of its individual components. Systems biology represents a paradigm shift from an in-depth study of one molecule or a few molecules at a time to a study of a system of integrated, interacting molecules. Systems biology attempts to understand how a process, a cell, a group of cells, or an organism works at a global level and how different components of the process cooperate to attain the "correct" functional outcome. It is now recognized that component-by-component analysis is not sufficient for the study of signal transduction, gene regulatory and biochemical networks, oncogenic transformation, and other processes in which many genes and proteins interact. Understanding the dynamics of such systems, both qualitatively and quantitatively, and constructing mathematical models with robust predictive capabilities will result in powerful new tools for biomedical research. The systems approach will be directly applicable to patient care as we implement targeted therapeutics and molecular makers leading to personalized molecular medicine. At the lowest level, systems biology will develop models that can improve our ability to select patients likely to respond to particular therapies and also greatly improve our ability to develop rational drug combinations. In its maturity, systems biology will be able to predict unintuitive consequences of targeted therapeutics. An outstanding example of this is the recent termination of a clinical trial of an mTOR inhibitor due to apparent increased tumor growth. Our preliminary systems biology-based mathematical and experimental models of the phosphatidylinositol 3-kinase/mTOR signaling network accurately predict these consequences as well as the biochemical processes involved. Further, the models suggest combinations of targeted therapeutics likely to reverse the negative effects of the mTOR inhibitors converting the outcome from negative to positive in terms of tumor growth.
 Many of the concepts of systems biology are not new. Biologists and biochemists have long known that the detailed (reductionist) study of individual proteins is just the first step toward an understanding of the overall (integrated) life process. Current advances in biology are a direct result of the success of the reductionist approach. The available experimental procedures necessarily forced a "one protein at a time" analysis during the middle of the 20th century. Advances in experimental methodology (high-throughput screening and functional genomics technologies) have made the "global" view accessible for the first time, finally allowing scientific research at the overall level of the cell or the organism possible.
 Recent developments in computational systems analysis are expanding the traditional hypothesis driven model in an intriguing new direction. By combining comprehensive representations of the data (with explicit ontologies) with machine learning tools, it becomes possible to add an automated capability for hypothesis generation. Critically, this integration between computational modeling and interventional laboratory bench studies provides an iterative approach to validate and constrain the computational models. New approaches are being developed by quantitative scientists, such as computational biologists, statisticians, mathematicians, computer scientists, engineers, and physicists, to improve our ability to make these high-throughput measurements and create, refine, and retest the models until the predicted behavior accurately reflects the phenotype observed. However, the successful implementation of this approach requires close collaboration between quantitative scientists and interventional biologists, who provide the original experimental data to develop the models and, more importantly, to test the hypotheses developed from the models. The importance of developing a cadre of scientists comfortable with both interventional bench studies and computational modeling is evidenced by a number of Roadmap, NCI and NIH training grants and RFAs related to this area.
 It is now clear that systems biology is a data-driven process. Although our analytical technologies have also improved over the last several years, most of them can generate only high-throughput data of low quality (ie microarrays), or low-throughput data of high quality (ie enzymatic assays). We need both high-throughput and high-quality data to create accurate cell response models. We also need technologies that simultaneously measure multiple parameters at the level of individual cells. Technological advances are beginning to provide the comprehensive databases at the DNA, RNA, and protein level necessary to integrate systems biology with cancer biology. Combining these patient and model-based databases with the ability to interrogate functional networks by a systematic analysis using siRNA libraries and chemical genomics provides an ability to link in silico modeling, computational biology, and interventional approaches to develop robust predictive models applicable to patient management.