Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)



Committee Chair/Advisor

Dr. William Richardson

Committee Member

Dr. Bruce Z. Gao

Committee Member

Dr. Robert Latour

Committee Member

Dr. Yongren Wu


Heart failure (HF) is a chronic, progressive condition defined as an abnormality of cardiac function with the inability of the heart muscle to pump enough blood to meet the body’s requirements for metabolism. HF has various contributing pathologies, including hypertension (86 million Americans), myocardial infarction (MI, 800,000 Americans per year and 300,000 recurrent infarctions each year), both of which promote fibrosis. Myocardial fibrosis contributes to left ventricular (LV) dysfunction and is histologically defined by excessive deposition of fibrous tissue relative to the mass of cardiomyocytes within the myocardial tissue. Quantitatively, myocardial fibrosis is characterized by increased collagen volume fraction (CVF) or percentage of myocardial tissue with collagen fibers. Currently, there are no prescribed therapeutics for preventing cardiac fibrosis, and clinicians are unable to predict which patients at what time and to what extent are more likely to develop fibrosis. Collagen accumulation contributes to increased stiffness and loss of function in failing hearts, and cardiac fibrosis remains a significant barrier to the treatment and prevention of HF. Collagen remodeling is regulated by a complex network of extracellular interactions, including: (1) collagen secretion, (2) protease secretion, activation, and degradation of collagen (namely Matrix Metalloproteinase (MMP) and Cathepsins), and (3) tissue inhibitors of metalloproteinases (TIMP) secretion and inhibition of MMPs. Importantly, this network is sensitive to mechanical tension. Fibroblast expression of collagen, MMPs, and TIMPs all depend on tension, and it is known that an excessive amount of tension can damage matrix fibers. There is also evidence that protease degradation of collagen can depend on fiber tension. However, it is unknown how tension affects collagen degradation by different proteases and protease mixes. The overarching objective of this dissertation is to develop a computational model of collagen turnover under combinatory chemo-mechano-conditions as a predictive tool for stratifying fibrotic risk for HF patients. Firstly, we tested the effect of tensile loading on collagenous tissue degradation by proteases. We picked four proteases and quantified the role of mechanical loading on the degradation of collagenous tissue by each protease. As matrix degradation leads to decaying force levels, sample degradation rate was quantified for different strain levels for each protease. Secondly, we developed a detailed biochemical network computational model of collagen I proteolysis capturing all interactions of type I collagen, four MMPs, and three TIMPs in a cell-free, well-stirred environment. We monitored the proteolytic activity of MMPs and inhibitory activity of TIMPs and then used the results from experimental data to fit five different hypothetical reaction topologies and determined kinetic rate constants for collagen degradation by MMPs, MMP inhibition by TIMPs, MMP and TIMP inactivation, MMP cannibalism, and MMP and TIMP distraction. We also used post-MI time courses of collagen, MMP, and TIMP levels in animal experiments from the literature to perform a parameter sensitivity analysis across the model reaction rates to identify which molecules or interactions are the essential regulators of ECM post-MI for both early and late time-periods. Lastly, we developed an ensemble classification algorithm for diagnosing HF patients with preserved ejection fraction (HFpEF) within a population of 459 individuals, including HFpEF patients and referent control patients. We concluded that machine learning algorithms could substantially improve the predictive value of circulating plasma biomarkers. Additionally, we built a mechanistic model to predict ECM component degradation using a genetic algorithm to connect ECM remodeling to the plasma biomarkers to help us with HFpEF patients’ classification. Our findings demonstrate that machine learning-based classification algorithms show promise as a non-invasive diagnostic tool for HFpEF patients’ classification while also suggesting priority biomarkers for future mechanistic studies to elucidate more specific regulatory roles. Our work suggests that computational modeling can serve as a beneficial tool for HF prognosis and potentially developing novel therapeutics.

Author ORCID Identifier




To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.