Towards automated assessment of drug efficacy on cardiomyopathic human heart function.

Abstract

Structural and functional abnormalities of ventricular myocardium related to Hypertrophic and Dilated Cardiomyopathies (HCM and DCM) are commonly caused by inherited mutations in sarcomeric proteins. Automating predictions of mutation and drug effects on human cardiac muscle function could enhance early diagnosis and evaluation of response to treatments for cardiomyopathies. However, the development of automation tools that integrate results from multiple experiments, conducted under various experimental conditions at different time and spatial scales from multiple species, is a significant challenge. Such automation tools usually require large sets of training data, which are extremely limited or even nonexistent from humans, thus any information generated could be significantly inaccurate or even misleading. Recent advances in multiscale computational tools, coupled with multiple experiments, for the translation of data concerning sarcomere level molecular interactions, including the effect of drugs, to the level of cardiac muscle fibers and tissues, might provide the most relevant approach for generating databases for automation algorithms. Using the MUSICO platform for precise modeling of protein-protein interactions and Ca2+ regulation in cardiac muscle cells, coupled with the finite element (FE) solver Mexie via Mijailovich-Prodanovic surrogate model, we can simulate experimental data from various sources. These tools collectively enable quantitative assessment of the changes in human cardiac muscle contractile responses caused by HCM and DCM extending from the molecular level to whole organ function, as well as prediction of improvements of cardiac muscle function by drugs. Our simulations successfully predicted pressure and volume changes due to troponin-C mutations and the effect of drugs, disopyramide and digoxin, in improving heart function in HCM and DCM, respectively, during multiple heartbeats in human heart. These simulations can potentially provide sufficient information for physicians to evaluate the progress of disease and the efficacy of administrated therapeutics.