Goal: Accelerate and reduce the cost of drug development.
Problem: Developing new drugs is a long and expensive process with a lot of trial end error which can take more than 15 years. Before a drug can reach a patient, it must go through several stages in the development process and rigorous testing to determine whether it is safe, and effective at treating the condition it was developed for and to ascertain the correct dosage and appropriate administration route.
How do we do it? We are focusing on early-stage and preclinical trials of the drug discovery process by eliminating inappropriate compounds and therefore reducing the cost and time several folds. During our research studies, we have identified three major pathways defined by the principal action of specific drugs: (1) modulation of Calcium transient, (2) changes in kinetic parameters, and (3) changes in macroscopic parameters.
Results: Using our MUSICO platform we were able to quantitatively assess the effects of several drugs (mavacamten, disopyramide, digoxin, 2-deoxy adenosine triphosphate (dATP)…) on human muscle behavior across multiple length and time scales as well as the effect of the dosage. Moreover, by coupling MUSICO with FE solvers, we have been able to predict the effects of drugs on whole organs.
Goal: Gain quantitative insight into cardiomyopathies pathways.
Problem: Cardiomyopathies are one of the most common causes of death among young adults. At the moment there is no cure and diagnosis, and prediction of disease progression remains difficult. Yet one in 500 people carries a mutation associated with the disease. About 40% of these mutations occur in myosin and another 40% in myosin-binding protein C (MyBP-C) and the rest are mostly in thin filament proteins troponin and tropomyosin.
How do we do it? In order to understand the effects of mutations and gain quantitative insights into disease pathways we have been conducting and analyzing experiments on different temporal and spatial scales, starting from the experiments in solution and motility assays all the way to cell and tissue mechanics and whole organ assessments. The experimental observations were then used as the input data with the appropriate tools in MUSICO platform which enabled us to predict muscle behavior under influence of genetic mutations at multiple levels.
Results: During our research in SilicoFCM project we have established a workflow for discovering and simulating the effects of genetic mutations on cardiac muscles. Currently, we have over 3000 simulations from mice, rats, and humans in our database, covering about 30% of known human cardiac mutations.
Goal: Building the framework for simulating human heart behavior at multiple temporal and spatial scales.
Problem: Rapid development of information technologies, simulation software packages, and medical devices in recent years provides an opportunity for collecting a large amount of clinical information. Creating comprehensive and detailed computational tools became essential to process specific information from an abundance of available data. From the point of view of physicians, it becomes of paramount importance to distinguish “normal” phenotypes from the appearance of the phenotype in a specific patient in order to estimate its disease progression, therapeutic responses, and future risks. On the other hand, patient-specific modeling presents many new challenges, including (1) the lack of details regarding the physical and biological properties of the human heart; (2) the need for subject-specific estimation of parameters from limited, noisy data, typically obtained using non-invasive measurements; (3) the need to perform numerous large-scale computations in a clinically useful time-frame; and (4) the need to store and share model metadata that can be re-used without compromising patient confidentiality.
How do we do it? Despite these difficulties, multiscale models of the heart can include a level of detail sufficient to achieve predictions that closely follow observed transient responses, providing solid evidence for prospective clinical applications. Modeling provides a route to join up multiple studies including underlying genetics, protein, structural and kinetics alternations, observed cell and tissue changes in disease to clinical outcomes. FilamenTech’s new integrative approach consists of three major components: (1) Multiscale Computational Models (MCM) that integrate experimental data at the level of single molecular interactions between contractile proteins and small molecules (e.g., calcium), and mechanical measurements at the level of muscle fibers, fibrils, and cardiac tissues, (2) Surrogate Models (SM) that mimic simulations from MCM for computationally effective calculations with small memory requirements and fast execution time and (3) Finite Element (FE) program packages for simulations at the level of whole heart or ventricles. These systems are coupled and provide the translation of the effects of mutations and other abnormalities in contractile proteins to functional changes in cardiac behavior in patients with myopathies.
Results: Developed over 18 years and critically acclaimed in over 60 peer-reviewed journals, our MUSICO Platform consists of multiple submodules (MUSICO SL, MUSICO MA, MUSICO Fiber, MUSICO X, and MP surrogate) for simulating a wide variety of muscle experiments, from molecular kinetics affected by genetic mutations and protein-drug interactions to functional behavior of muscle cells and tissues. Moreover, with our research partners from Seven Bridges Genomics, the University of Kragujevac, and Barcelona Supercomputing Center we have developed workflows for integrating patient-specific genetics and structural data in order to translate the effects of mutations and drugs to whole organ behavior.
Goal: Enhance understanding of contractile dysfunctions mechanisms in skeletal muscle tissues.
Problem: Continuous degeneration of muscle tissue, inflammatory processes, and fibrosis characterized by a loss of muscle mass, formation of micro-scars, adipose tissue in the muscles and eventual muscle punctures are often signs of muscular dystrophies (dystrophinopathies). These neuro-muscular diseases result from genetic mutations of a structural protein called dystrophin. The absence of functional dystrophin leads to the most common and severe form of muscular dystrophy, Duchenne muscular dystrophy (DMD). The severity of dystrophic diseases can be assessed from muscle fiber and fiber bundle mechanical experiments. However, these observations may not be sufficient to quantitatively evaluate the main underlying cause and the progress of the disease. Moreover, it has been hypothesized that nebulin, a filamentous protein extending along the length of the actin filament, maintains a constant length of thin filaments, but in disease, e.g., human nemaline myopathy, mutations in, or lack of, nebulin can result in much shorter randomly distributed thin filament lengths which is associated with impairment of myosin binding to actin.
How do we do it? Typically, within one muscle bundle, there are so-called fast and slow muscle fibers that shorten and lengthen at different speeds during muscle contraction. The use of computational multiscale models, such as our MUSICO platform, may inform the understanding of how normal or pathological variations in microstructure are related to the observed macroscopic behavior of dystrophic muscle. We evaluated how the presence or absence of dystrophin affects the connective tissue deformation between fast and slow muscle fibers. By adjusting the elasticity of the extracellular matrix layer, we estimated the magnitude of the shear strain under unloaded and lightly loaded fiber contractions caused by differences in shortening velocities between fast and slow fibers. On the other hand, we used MUSICO to quantitatively separate the effects of structural changes, the kinetics of cross-bridge cycling, and calcium sensitivity of the thin filaments. This enabled us to examine whether the force deficit in nebulin-based nemaline myopathy can be explained by (a) shorter and variable thin filament lengths, (b) altered cross-bridge cycling kinetics, and (c) reduced myofilament calcium sensitivity.
Results: The simulations for DMD showed that without dystrophin large shear strains (up to six times larger in the DMD than in the healthy muscles) are generated causing local microinjury and inflammation leading to further muscle degeneration. The main cause of these large strains is a difference in contractile characteristics between fast and slow muscles, i.e., between shortening velocities and rate of force development after the onset of activation. On the other hand, MUSICO simulations for nebulin deficiency showed that variation in thin filament length cannot by itself account for experimental observations of the contractility in nebulin-deficient muscle, but instead must be accompanied by a decreased myosin binding rate. Additionally, to match the observed calcium sensitivity, the rate of troponin-I detachment from actin needed to be increased.
Goal: Build computational tools to determine the structure of the muscle under physiological conditions and to efficiently extract information from experimental X-ray diffraction patterns.
Problem: Synchrotron small angle X-ray fiber diffraction is the method of choice for obtaining structural and physiological information in the same experiment from active muscle. Experimental questions addressed range from basic biophysical questions regarding mechanisms of force production and regulation to increasingly pre-clinical questions relating structure to functional phenotype in animal models for cardiomyopathies and skeletal muscle disease as well as human muscle biopsies. Critical barriers to progress, however, has been the lack of robust, user-friendly tools for data reduction and computational tools for modeling diffraction patterns to interpret the data.
How we do it? We are developing a new tool, MUSICO X for predicting two-dimensional X-ray diffraction patterns from striated muscle. MUSICO X is a new extension module for the multiscale simulation platform MUSICO that predicts small-angle X-ray fiber diffraction patterns simultaneously with the physiological data as a novel “forward problem” approach to extracting maximal information from static and dynamic time resolved X-ray fiber diffraction experiments on striated muscle. This new module assigns electron densities to sarcomeric components using predicted molecular positions from MUSICO and predicts simulated diffraction patterns that are tested and refined against representative data sets.
Results: Using our MUSICO X tool integrated within MUSICO platform enabled us to predict the structure of skeletal and cardiac muscles under physiological conditions as well as distinguish individual components that contribute to X-ray observations. Moreover, we have been able to estimate forces in individual actin filaments and crossbridges and predict changes in monomer spacings during fiber contraction by comparing experimental X-ray diffraction profiles to MUSICO X simulations.