Sifting through the haystack: Optimizing hematopoietic stem cell purification for gene therapy

From the Kiem Lab, Clinical Research Division

November 16, 2020 • By L Schmidt

Gene therapy, involving directed ‘corrections’ or modifications of the genetic material in afflicted tissues, holds promise for the treatment of many human diseases. The lasting success of gene therapy relies on efficient genetic modification and engraftment of cells that can a) reconstitute disease-relevant cell populations and b) persist within the body for long periods of time. Hematopoietic stem cells (HSCs), long-lived, multipotent adult stem cells that maintain and generate billions of mature blood cells every day, are targets for some of the most common gene therapy applications against hematologic disorders. Current HSC gene therapy paradigms involve the isolation of HSCs from the body, genetic modification of these HSCs in vitro, and infusion of the modified HSCs into patients. Dr. Stefan Radtke of the Kiem lab (Clinical research Division) and co-authors performed a side-by-side comparison of currently proposed HSC isolation techniques to determine the optimal strategy for gene therapy applications. Their findings, which established an improved two-marker strategy for enriching HSC populations, were recently published in Molecular Therapy, Methods & Clinical Development.

HSC-Targeted Gene Therapy Schematic — Image provided by Dr. Radtke

The current gold standard HSC isolation strategy involves single-marker purification of CD34⁺ cells. CD34⁺ cell isolation yields a population of cells that is enriched for HSC properties compared to un-purified populations. However, it is known that the CD34⁺ population remains highly heterogenous, consisting mostly of less potent progenitor cells committed to limited blood lineages, and relatively few true HSCs with the capability of long-term reconstitution of all hematopoietic cell types. Thus, many CD34⁺ cells must be isolated, modified, and infused into patients to increase the chances of successful editing and engraftment of the relatively rare HSCs present in this population. One way to improve the efficiency and cost-effectiveness of HSC gene therapy is to refine the target population for gene modification. Several strategies have been proposed for enhanced purification of HSCs, including enrichments of CD133⁺, CD38^low, or CD90⁺ subsets from the greater CD34⁺ population. Radtke and others in the Kiem lab previously established that long-term multi-lineage engraftment is mediated by CD34⁺CD90⁺ HSCs in a non-human primate transplantation model. Here, in a side-by-side comparison, the Kiem Lab confirmed that human CD34⁺CD90⁺ cell purification yielded superior outcomes in terms of enrichment of HSC potential and multi-lineage reconstitution by gene-modified cells compared to current and other proposed isolation methods.

First, the authors assessed the frequencies of, and defined the hierarchical relationships between, each of the CD34⁺ subsets by performing flow cytometry on bone marrow-derived cells from 9 healthy donors. CD133⁺ cells made up the largest fraction of the CD34⁺ population, followed by the CD38^low subset, and finally the CD90⁺ subset, representing 2.4, 5.8, and 12.5-fold enrichments of target cells over the standard CD34⁺ purification strategy, respectively. By fluorescence-activated cell sorting (FACS),. each population was purified for further phenotypic characterization. Interestingly, they found that most of the CD38^low subset expressed the CD133 marker, and that the CD90⁺ subset was almost entirely CD133⁺ and CD38^low, indicating a hierarchical relationship between these subsets. In other words, the CD90⁺ cells represented a subset of the CD38^low population, which formed a subset of the CD133⁺ population, all of which constituted a fraction of the bulk CD34⁺ population. Thus, the CD90⁺ fraction represented the most refined subset of CD34⁺ cells.

Single-cell RNA sequencing (scRNA-seq) was used to assess the transcriptional heterogeneity and similarity to known progenitor signatures amongst each subset. First, a reference map using principal component analysis (PCA) of bulk CD34⁺ cells from 2 healthy donors, defining seven distinct clusters of cells was generated. Cluster 1 represented the population most enriched for known progenitor-like HSC signatures, and the other 6 clusters included populations enriched for genes associated with lymphoid, myeloid, and erythroid differentiation. Next, the authors used scRNA-seq to overlay the sorted CD133⁺, CD38^low, and CD90⁺ subsets onto their map. CD133⁺ cells mapped both to the progenitor-like HSC cluster as well as to lympho-myeloid clusters. The CD38^low and CD90⁺ subsets each mapped more exclusively to the progenitor cluster, but were essentially indistinguishable from one another in this setting. To profile these populations more deeply, they performed bulk RNA-seq on multi-parameter sorted populations. Using this technique confirmed enriched expression of early progenitor-like genes in both CD90⁺ and CD38^low subsets, but showed drift toward erythroid-like signatures and more donor-to-donor variability in CD90^-CD38^low cells. Importantly, the progenitor phenotype was stable in the CD90⁺ subset after G-CSF-mobilization, a procedure commonly performed prior to harvesting HSCs, confirming that the CD90⁺ population is enriched for HSC signatures in clinic-like settings.

Next, the Kiem Lab tested the in vivo reconstitution potential of the CD90⁺ subset compared to CD90^- and bulk CD34⁺ populations. To assess this, G-CSF-mobilized cells were sorted into populations and injected into into sub-lethally irradiated immunocompromised (NSG) mice. CD90⁺ cells showed superior engraftment in most tissues, including blood, bone marrow, and thymus, compared to the other populations. Importantly, only populations containing CD90⁺ HSCs were capable of repopulating the CD90⁺ HSC compartment in the bone marrow, and these populations had superior performance in ex vivo CD34⁺ cell differentiation assays. Finally, the efficiency of gene therapy in the HSC-enriched CD90⁺ population was tested and compared to bulk CD34⁺ cells. Importantly, due to the reduced number of cells in the refined target population, they were able to achieve higher transduction efficiencies in the CD90⁺ subset using clinically-approved gene therapy protocols. Upon transfer into immunocompromised mice, the Kiem Lab. observed similar engraftment efficiencies between the CD90⁺ subset and bulk CD34⁺ cells. However, gene-modified cells were more prevalent in the tissues of CD90⁺-reconstituted mice, indicating more efficient gene therapy outcomes in this setting.

The Kiem Lab and co-authors have established a method for purifying a refined CD34⁺ HSC subset with enriched primitive HSC expression signatures, better engraftment potential compared to other CD34⁺ fractions, and superior enrichment of gene-modified tissues in vivo compared to the current gold standard. “Simplicity of this HSC-enriched phenotype and the ease of isolation will hopefully facilitate the implementation of this HSC purification step into existing as well as future gene therapy protocols for pre-clinical and clinical applications”, commented the lead author. These findings will allow for more efficient and cost-effective manufacturing of HSC gene therapy products for patients in the clinic.

This work was supported by the National Institutes of Health, the National Cancer Institute, the Evergreen Fund, the Immunotherapy Integrated Research Center, the Cuyamaca Foundation, the Markey Molecular Medicine Investigator award, the José Carreras/E. Donnall Thomas Endowed Chair for Cancer Research and the Fred Hutch Endowed Chair for Cell and Gene Therapy awards, and the Fleischauer Family Endowed Chair in Gene Therapy Translation award.

UW/Fred Hutch Cancer Consortium members Jennifer Adair (Fred Hutch), and Hans-Peter Kiem (Fred Hutch) contributed to this work.

Radtke S, Pande D, Cui M, Perez AM, Chan YY, Enstrom M, Schmuck S, Berger A, Eunson T, Adair JE, Kiem HP. Purification of Human CD34+CD90+ HSCs Reduces Target Cell Population and Improves Lentiviral Transduction for Gene Therapy. Mol Ther Methods Clin Dev. 2020 Jul 15;18:679-691. doi: 10.1016/j.omtm.2020.07.010.