Xie Awarded $2 Million NIIMBL Stem Cell Research Grant
Wei Xie, assistant professor of mechanical and industrial engineering, has received a $2 million award from NIIMBL, the National Institute for Innovation in Manufacturing Biopharmaceuticals, for the project “Advanced Bioprocessing Sensor and Analytical Technologies for Induced Pluripotent Stem Cell Culture Online Monitoring and Automation.” Project collaborators include co-PI Jared Auclair, the director of the Biopharmaceutical Analysis Training Lab (BATL) at Northeastern, along with multiple universities and industry partners.
NIIMBL supports innovation in biopharmaceutical manufacturing technology and workforce development. The organization recently announced $15.8 million in funding for 14 projects in these areas, including Xie’s.
Large-scale manufacturing of induced pluripotent stem cells (iPSCs) is essential for cell therapies and regenerative medicines. Yet iPSC productivity and pluripotency are highly sensitive to culture conditions; subtle changes in those conditions can lead to stress that adversely affects cell differentiation. To help meet emerging therapeutic needs, Xie’s research focuses on improving the monitoring and control technology to accelerate intensified, reliable, and scalable iPSC production in bioreactors and support Good Manufacturing Practice development via quality-by-design and automation.
Improving Online Sensor Technology to Facilitate Real-Time Release
The two-photon excitation (TPE) optical sensor has been used in clinical studies to detect cell differentiation. This device has the advantage of being able to monitor intracellular metabolites and detect the metabolic state shift of each individual cell.
“Optical sensors will definitely play a critical role to support bio-drugs, like monoclonal antibodies, cell and gene therapies, and vaccines, as well as manufacturing online monitoring and real-time release,” Xie affirms.
However, she goes on to explain that bioreactors have much more complex and dynamic environments than clinical test applications, often employing agitation, which makes TPE online monitoring difficult. A major goal of her project is to refine the TPE sensor so that it can detect spatial heterogeneity and support large-scale cell processing. This real-time monitoring will allow technicians to react quickly to changes in cell metabolic and functional behaviors that might alter the manufacturing process in undesired ways.
Interpretable AI/ML for Cell Culture Process Mechanism Learning and Control
Hand-in-hand with this effort is Xie’s second major goal: the improvement of interpretable AI and machine-learning algorithms for bioprocessing control and automation. TPE is just one of several sensor technologies used to monitor iPSC cultures; combining various online/offline measurements together immensely increases the volume and complexity of the data they gather.
Due to strong cell-cell interactions, extracellular matrix secretion, and iPSC self-aggregation, compact cell spheres will be naturally formed. The large aggregates can cause hypoxia and low nutrient concentration for the cells located at the core, which can result in decreased viability, loss of pluripotency, and cell differentiation. This issue can be more serious for large-scale bioreactors.
“Each cell is a system. Cells aggregated together are a system. The many aggregations in the bioreactor are a complex system-of-systems,” Xie says, illustrating the need for a multi-level bioprocess hybrid model characterizing biological, physical, and chemical interactions at molecular, cellular, and system levels, as well as facilitating mechanism learning from heterogeneous measurements and advanced sensors.
To this end, Xie’s team is building a knowledge graph hybrid model-based machine learning (KG-ML) framework. Ideally, it will advance risk- and science-based understanding of cellular metabolism by correlating online sensing with the intracellular metabolic state of the culture process. This KG hybrid model can integrate heterogeneous online/offline measurements, improve the prediction of cell response to micro-environmental perturbation, and support the control of cell-to-cell variation.
“Ultimately, the KG-ML, coupled with the TPE sensor, will enable reliable online monitoring and ensure robust control to enhance productivity and quality for the iPSC culture process,” Xie says. “Once you have information on the individual cell level that’s complex and dynamic, a big innovation is needed to leverage that information to support optimal and flexible process control of stem cell cultures.”
Xie believes that the demand for iPSC-based technology will grow as populations continue to age and experience increasing rates of neurologic, skeletal, muscular, and other disorders.
“Cell therapies and regenerative medicines will be a huge market,” she says, “but technology definitely needs to further improve.”
This project was developed with an award from the National Institute for Innovation in Manufacturing Biopharmaceuticals (NIIMBL) and financial assistance from the U.S. Department of Commerce, National Institute of Standards and Technology (70NANB21H086).