Model-based estimation of myeloid hematopoietic progenitor cells in ex vivo cultures for cell and gene therapies

Abstract

Ex vivo production of hematopoietic progenitor cells has potential applications for cell therapy to alleviate cytopenias associated with chemotherapy and for gene therapy. In both therapies, progenitor and stem cells are considered crucial factors for therapeutic success. Assays for progenitor cells, however, take 2 weeks to complete, which is similar to the length of a typical culture. Therefore, a real-time estimation of the percentage or number of progenitor cells, based on rapid measurements, would be useful for optimization of feeding and harvest decisions. In this study, metabolic activity assays and flow cytometric analysis were used to estimate the content of progenitor cells. The measured metabolic activities are a collective contribution from all types of cells. Cells in granulomonocytic cultures have been lumped into six cell types and metabolic rates have been modeled as a linear function of cell composition and growth rate and as a nonlinear function of cell density. Data from 24 experiments were utilized to determine the model parameters in a calibration step. These data include flow cytometric analysis of more mature hematopoietic cells, progenitor cell colony assays, total cell content, and metabolite concentrations, and cover a wide range of cell composition, cell density, and growth rate. After calibration, the model is able to deliver good predictions of progenitor cell content for cultures with higher percentages of progenitor cells, as well as the peak progenitor cell content, based only on parameters that can be rapidly measured. With the aid of those predictions a harvest strategy was developed that will allow optimizing the harvest time based on the culture kinetics of each patient or donor inoculum, rather than using retrospective analysis to determine a uniform harvest time.