Reaction Flux and Information Flow in the COX-2 Allosteric Regulation Network
March 7, 2018
Foege N130A, Wallace H. Coulter Seminar Room
Cells exist in a state of constant flux, subject to intrinsic and extrinsic perturbations that directly affect the dynamics of biological network processes. These perturbations can disrupt network signal execution in ways that are not easily discernible from experimental approaches, which typically measure network end-points in signal execution. This limited understanding of signaling network mechanisms often leads to an incomplete picture of cellular process execution. This gap in our knowledge is not due to a lack of experimental or modeling efforts, but rather to the complexity associated with signaling flux and information flow in biochemical reaction networks. Here we present a novel approach that combines model exploration, Bayesian calibration, kinetic analysis, and Information Theory approaches to study how information flows in complex biochemical networks. We study the reaction kinetics allosteric modulation of the Cycloxygenase-2 (COX-2) catalytic network, a protein central in inflammation and the target of most NSAIDs to date. Our COX-2 Reaction Model (CORM) describes the dynamics of COX-2 processing arachidonic acid and 2-arachidonyl glycerol into prostaglandin (PG) and prostaglandin-glycerol (PGG). The network was calibrated to experimental data using a Bayesian inference method to determine converged parameter distributions. We analyze the network with a non-equilibrium flux-analysis method to characterize multiple pathways in the network cooperate toward signal processing outcomes that lead to PG or PGG production under different substrate concentrations. Global sensitivity analysis with the products as outputs aligns well with known COX-2 drug efficiencies. However, global sensitivity analysis performed in different regions of the substrate space indicates that the sensitivity of PG and PGG outputs to different perturbations is dependent on initial substrate concentrations. We apply information theory methods to understand the link between biochemical flux and the maximum channel capacity through a given path from reactants to products. Our initial analysis indicates that channel capacity is highly variable within the reaction network and exhibits dependence on the correlation across inputs. Our observations have direct implications for pharmaceutical targeting of COX-2 in vivo as well as for other enzyme-substrate systems that exhibit multiple heterogeneous substrate inputs.
Carlos Lopez, Ph.D., is an Assistant Professor at Vanderbilt University School of Medicine in Nashville, TN, with appointments in the Department of Cancer Biology and Biomedical Engineering and member of the Vanderbilt-Ingram Cancer Center the Center for Quantitative Sciences. Prior to his current position, Carlos completed his doctoral degree in Physical Chemistry (2004) and his bachelor of science in Chemistry/Biochemistry (1998) at the University of Miami. The overarching goal of his research is to develop and apply numerical, modeling, and statistical methods to understand cellular processes and their dysregulation. In particular, his work develops and applies novel computational modeling tools and leverages strong experimental collaborations to describe intracellular biochemical signaling networks using a wide range of mathematical methods and further our understanding of cellular decision-making processes. This work could lead to a better understanding of dynamic cellular systems, how they are regulated in health, dysregulated in diseases such as cancer, and use this knowledge to guide experiments toward novel therapies.