A simplified co-culture reveals altered cardiotoxic responses to doxorubicin in hPSC-derived cardiomyocytes in the presence of endothelial cells

Abstract Cardiotoxicity is a significant challenge in cancer therapies, particularly with doxorubicin, a widely used anthracycline known for its broad anti-cancer spectrum but life-threatening cardiac side effects. There is a critical need for more predictive in vitro models to understand doxorubicin-induced cardiotoxicity and patient-specific drug responses. In this study, we used human pluripotent stem cell (hPSC)-derived cardiomyocytes (hPSC-CMs), cardiac fibroblasts (hPSC-cFBs) and endothelial cells (hPSC-ECs) to investigate the cardiotoxic effects of doxorubicin in two-dimensional mono-and multi-cell type cultures. By mimicking the cumulative dose effect seen in patients through repeated doxorubicin treatments and using a machine learning-based in silico image analysis tool, we could precisely quantify caspase 3/7 activity as an early toxicity marker and identify hPSC-CMs in multi-cell type cultures. This innovative approach allowed continuous monitoring of apoptosis from phase-contrast images, revealing that hPSC-ECs showed higher sensitivity to doxorubicin than isogenic hPSC-CMs or hPSC-cFBs and significantly enhanced cardiomyocyte toxicity in co-culture. In contrast, dermal fibroblasts differentiated from the same hPSC line showed no toxic response under the same treatment regimen. These results challenge the conventional focus on cardiomyocytes as the target of drug-induced cardiac damage. Our findings not only highlight the complex interplay among different cardiac cell types in mediating the toxic effects of doxorubicin, but also demonstrate the potential of AI-enabled tools to advance personalized drug screening and safety assessments.