Automated Sarcomere Structure Analysis for Studying Cardiotoxicity in Human Pluripotent Stem Cell-Derived Cardiomyocytes

Abstract Drug-induced cardiotoxicity is one of the main causes of heart failure (HF), a worldwide major and growing public health issue. Extensive research on cardiomyocytes has shown that two crucial features of the mechanisms involved in HF are the disruption of striated sarcomeric organization and myofibril deterioration. However, most studies that worked on extracting these sarcomere features have only focused on animal models rather than the more representative human pluripotent stem cells (hPSCs). Currently, there are limited established image analysis systems to specifically assess and quantify the sarcomeric organization of hPSC-derived cardiomyocytes (hPSC-CMs). Here, we report a fully automated and robust image analysis pipeline to detect z-lines and myofibrils from hPSC-CMs with a high-throughput live-imaging setup. Phenotype measurements were further quantified to evaluate the cardiotoxic effect of the anticancer drug Doxorubicin. Our findings show that this pipeline is able to capture z-lines and myofibrils. The pipeline filters out disrupted sarcomere structures and irrelevant noisy signals, which allows us to perform automated high-throughput imaging for accurate quantification of cardiomyocyte injury.