Abstract 10881: Graph-Based Analysis of Cardiomyocyte Network Connectivity

Abstract

Introduction: Cardiomyocytes branch and interconnect with one another, allowing for mechanical and electrochemical coupling. Despite being an essential structural feature enabling coordinated contraction, analysis of this branching structure has been limited. Herein, we use graph-based methods to analyze the cardiomyocyte network structure. Hypothesis: Cardiomyocytes are fully-connected, rather than being bundled into discrete, tract-bundles. We therefore expect that the size of connected graph components will increase with graph size, rather than tending towards a constant tract-bundle size. Methods: Rodent myocardium (n = 4) was stained for collagen and imaged using extended volume confocal microscopy. A sheetlet portion was segmented (~150 cells). The centerline of each cardiomyocyte was tracked. Vertices were identified where the cell bodies of adjacent cardiomyocytes merged. From the connected network (Fig 1A), an adjacency matrix was determined and the shortest-paths between all vertices calculated. The largest component was recorded as a function of graph size. Simulated fully-connected and tract-bundle networks (Fig 1B) were then compared to cardiomyocyte networks (Fig 1C). Results: For the simulated fully-connected network, the largest component increased with graph size, while the tract-bundle network levelled-off based on tract size (Fig 1B). In the cardiomyocyte networks (Fig 1C, red squares), the largest component continued to increase with graph size. Myocardial connectivity appears to lie between a tract-bundle and fully-connected simulated networks (Fig 1C). This suggests that the myocardium tends towards being a fully-connected network, although confirmation requires a larger image volume. Conclusion: Graph-based analysis of cardiomyocyte network structure distinguished between simulated fully-connected and tract-bundle networks. Cardiomyocyte networks are more similar to the fully-connected networks.