Zhao Abstract

Hematopoietic stem cells (HSCs) are an invaluable clinical resource for the treatment of a variety of inherited and acquired blood and immune disorders by hematopoietic cell transplantation (HCT). The ability to predict the long-term engraftment potential of donor HCT products, particularly after ex vivo manipulation such as genome editing as a part of gene therapy or immunotherapy, represents an unmet clinical need. However, current technologies such as high-dimensional flow cytometry and single cell transcriptomics lack precision in predicting long-term HSC fate at single cell resolution and remain too resource intensive and costly for routine screening applications. To overcome these limitations, we propose to leverage on the Fred Hutch Cancer Center (FHCC)’s

CCEH HCPPS core stem cell products and services to develop a method of using live imaging and machine learning (ML) to identify HSCs with long-term engraftment potential. To achieve this, we will take advantage of our established ex vivo HSC culture systems (endothelial co-culture environment and engineered niche cytokine environment) for live imaging of single index-sorted HSCs from adult bone marrow or mobilized peripheral blood (provided by FHCC CCEH HCPPS core), which will be coupled by downstream functional assessment of long-term in-culture

maintenance and engraftment in immunodeficient mice. Recorded HSC live imaging series with corresponding long-term maintenance and engraftment outcome will be used in ML for classification of HSCs. Success in these efforts would provide novel insights into the behavior of HSCs and their niche interactions at single cell resolution and advance translational applications to optimize the long-term engraftment potential of HSCs manipulated ex vivo or de novo HSCs engineered from pluripotent stem cells.