Scalable process technologies for human pluripotent stem cells (hPSCs) expansion and differentiation

Human pluripotent stem cells (hPSCs) can serve as an unlimited, renewable "raw material" for essentially any therapeutic cell product, placing them at an exciting intersection of biology, manufacturing and the emerging field of regenerative medicine. In the last decade, the field has progressed to the initiation of >100 clinical trials applying hPSC-derivatives for a spectrum of diseases. However, for enabling the envisioned routine hPS cell therapies in future, novel industry-scale manufacturing processes compliant with clinical and regulatory needs are required. Recently, the bioprocessing field has shifted from proof-of-concept studies demonstrating that pluripotent cells and their derivatives can be grown in suspension culture in dynamic systems, toward advanced processes that can be standardized, intensified, and successfully transferred between production sides. In this article, we provide a brief historical background on hPS cell manufacturing but will mainly focus on recent developments in the field. This includes suspension culture (3D)-based bioprocessing strategies for both the cultivation/ expansion of hPSCs at the pluripotent state and their lineage-directed differentiation as well, such as clinically relevant examples of functional cell types representative of all three germ layers. Since industrial-scale bioprocessing is more than just cell yields, aspects of volumetric productivity, production time, challenges for process reproducibility and others will be outlined. Additional aspects will include the optimization of upstream expansion and its integration with directed differentiation protocols as well as downstream handling into coherent, GMP-ready, clinically compliant production processes. Finally, an outlook is provided, indicating that bioreactor-based hPSC manufacturing in future will be defined by more controlled, data-rich, and economically defensible production processes feedback-controlled by artificial intelligence and machine learning. This will promote moving from retrospective process analytics to validated in process decision support for reducing production batch failure across individual hPSC lines, tightening comparability, and enabling more consistent product definition across production sites.