Human induced pluripotent stem cells (iPSCs) hold great promise for reducing the mortality of cardiovascular disease by cellular replacement of infarcted cardiomyocytes (CMs). CM differentiation via iPSCs is a lengthy multiweek process and is highly subject to batch-to-batch variability, presenting challenges in current cell manufacturing contexts. Real-time, label-free control quality attributes (CQAs) are required to ensure efficient iPSC-derived CM manufacturing. In this work, we report that live oxygen consumption rate measurements are highly predictive CQAs of CM differentiation outcome as early as the first 72 h of the differentiation protocol with an accuracy of 93%. Oxygen probes are already incorporated in commercial bioreactors, thus methods presented in this work are easily translatable to the manufacturing setting. Detecting deviations in the CM differentiation trajectory early in the protocol will save time and money for both manufacturers and patients, bringing iPSC-derived CM one step closer to clinical use.
Keywords: cell manufacturing; critical quality attributes; induced pluripotent stem cells; machine learning; oxygen consumption rate; time-series analysis.
© 2023 Wiley Periodicals LLC.