Abstract
Matrix metalloproteinases (MMPs) are a family of proteases that drive degradation of extracellular matrix (ECM) across many tissues. MMP activity is antagonized by tissue inhibitors of metalloproteinases (TIMPs), resulting in a complex multivariate system with many MMP isoforms and TIMP isoforms interacting across a network of biochemical reactions - each with their own distinct kinetic rates. This system complexity makes it very difficult to identify which specific molecules are most responsible for driving ECM turnover in vivo and therefore the most promising therapeutic targets. To help elucidate the specific roles of various MMP and TIMP isoforms, we present a computational systems biology model of collagen turnover capturing all possible interactions between type I collagen, four different MMP isoforms (MMP-1, -2, -8, and -9), and three different TIMP isoforms (TIMP-1, -2, and -4). We used dye-quenched fluorescent collagen to monitor the degradation of collagen in the presence of various MMP+TIMP cocktails, and we then used these experimental data to fit hypothetical reaction system topologies in order to investigate their respective accuracies. We determined kinetic rate constants for this system and used post-myocardial infarct time courses of collagen, MMP, and TIMP levels to perform a parameter sensitivity analysis across the model reaction rates and predict which molecules and interactions are the important regulators of ECM in the infarcted heart. Notably, the model suggested that MMP degradation and inactivation terms were more important for driving collagen levels than TIMP interaction terms. In sum, this work highlights the need for systems-level analyses to distinguish the roles of various biomolecules operating with a complex system, prioritizes therapeutic targets for post-infarct cardiac remodeling, and presents a computational framework that can be applied to many other collagen-rich tissues.