Collaborating Investigator(s): Rai Winslow, PhD
Institution: Johns Hopkins
Dr. Winslow is the Raj and Neera Singh Professor of the Department of Biomedical Engineering and the director of the Institute for Computational Medicine and the Center for Cardiovascular Bioinformatics and Modeling at the John Hopkins University School of Medicine & Whiting School of Engineering. Dr. Winslow's research interests focus on computational modeling of intracellular signaling, metabolism, and electrical excitability in cardiac myocytes, integrative modeling of cardiac function in health and disease, biomedical data representation, and database design and Grid-computing and data-sharing. Dr. Winslow is a specialty editor-in-chief for Frontiers in Computational Physiology and Medicine and has been on numerous advisory committees over the last 10 years.
One of the goals of Dr. Winslow's research is to advance understanding of the molecular basis of cardiac arrhythmias, particularly those related to intracellular Ca2+ dynamics in the setting of heart failure. To do this, his group has developed computational models of normal and failing cardiac myocytes based on extensive experimental data. Using these models, they have shown that disruptions of intracellular Ca2+ handling play a key role in arrhythmogenesis. Motivated by these findings, they are now developing multiscale models of Ca2+ signaling in order to understand how signaling events within the cell contribute to risk of arrhythmia.
This Driving Biomedical Project involves all of the TR&Ds. Currently, Dr. Winslow can only qualitatively verify the existence of what are known as "metabolic sinks" and the extent to which they recover on re-perfusion by examining reduced resolution cut planes through the image data. Tools that enable quantitative data analysis and interactive navigation into these very large and complex 3D datasets are required. These two goals are enabled by the innovation in the Image and Geometric Analysis TR&D and Visualization TR&D. In the Image and Geometric Analysis TR&D, we apply methods for user-assisted segmentation to images of cardiac tissue, followed by visualization of combined segmentations and fluorescent channel data using FluoRender. Finally, in order to properly characterize the nature of these metabolic sinks, we use the models we generate from segmentations in the Simulation & Estimation TR&D as the basis for a simulations to identify active electrophysiological tissue and associate cardiac behavior with those regions.
The Center is helping to advance the study of metabolic sinks while also advancing the technical contributions for several cross-cutting TR&Ds. Dr. Winslow is advancing the understanding of metabolic sinks and their relationship to arrhythmias by identifying the mechanisms by which reactive oxygen species (ROS) are produced under conditions of ischemia and in the setting of heart failure. Demonstrating the existence of metabolic sinks will provide definitive evidence supporting a novel hypothesis regarding mechanisms of arrhythmia. While helping to solve these problems, we are pulling technologies from all three TR&Ds to improve the segmentation and building of models for simulations while also advancing the visualizations used to display these large datasets.
The methods used in this DBP to help understand the molecular basis of cardiac arrhythmias require input from all of the TR&Ds in the image modeling pipeline. Beginning with the Image and Geometric Analysis TR&D, we use semiautomated segmentation algorithms to isolate nuclei and mitochondria that are corrupted with noise from the fluorescent stain. We can use these segmentations to quantify the distribution of myocytes in the tissue and calculate diameter, length, and volume statistics. Also of importance are the physioelectric simulations we perform to identify cardiac arrhythmia behavior. Following these results, we use FluoRender's advanced rendering techniques for multi-channel confocal data to interactively position the heart and select cutting planes to visualize and understand both the properties of metabolic sinks and sheet structure using original image data, segmented nuclei and mitochondria, and simulation outputs. Combined, these three TR&Ds provide a pipeline for statistical image analysis, simulation, and visualization of results necessary for understanding cardiac metabolic injury.
