Preservation of Gene Edited Hematopoietic Stem Cells By Transient Overexpression of BCL-2 mRNA

Introduction: Site-specific gene correction of the point mutation causing sickle cell disease (SCD) in hematopoietic stem cells (HSCs) constitutes a precise strategy to generate a life-long source of gene-corrected erythrocytes that do not sickle. However, low efficiency of homology-directed repair (HDR) in primitive reconstituting HSCs is currently a limit to the use of therapeutic genome editing for treatment of severe genetic blood disorders. To identify the mechanism(s) that underlie decreased HDR efficacy in primitive HSCs relative to that in more mature progenitor populations, we assessed: efficiency of gene delivery and expression after electroporation of in vitro transcribed mRNA; functional ZFN-mediated endonuclease activity; cell cycle status; gene expression of key HDR genes; and cytotoxic responses; in the following immunophenotypically-defined human cell populations: HSCs (CD34⁺/CD38^-/CD90⁺CD45RA^-); multipotent progenitors (MPPs) (CD34⁺/CD38^-/CD45RA^-/CD90^-); and progenitor cells (CD34⁺/CD38⁺).

Methods: CD34⁺ cells were enriched from human G-CSF-mobilized peripheral blood and cultured for 1-3 days prior to electroporation of in vitro transcribed mRNA encoding GFP or a pair of zinc finger nucleases (ZFN). The ZFNs, designed to target the sickle mutation in exon 1 of the human beta-globin gene, were co-delivered with one of the homologous donor templates containing the corrective base (A/T): an integrase-deficient lentiviral vector (IDLV) or a 101bp single-stranded oligodeoxynucleotide (oligo). Percentages of alleles containing insertions/deletions (indels) and/or HDR-mediated gene correction were analyzed by high throughput sequencing (HTS). Acute cytotoxicity was determined by flow cytometry, identifying viable cells as 7AAD/AnnexinV neg. cells. To assess HDR-mediated gene correction in vivo after three months, gene-edited cells were transplanted (>1E6 viable CD34⁺ cells/mouse, I.V.) one day after electroporation into irradiated (250cGy) NOD/SCID/IL2R gamma^-/- (NSG) mice.

Results: In HSCs, MPPs and progenitor populations, no differences were observed in delivery and expression from electroporated GFP mRNA [%GFP(+) and MFI]. To assess the activity of ZFN mRNA in the stem and progenitor populations, ZFNs were delivered to CD34⁺ cells through electroporation of in vitrotranscribed mRNA. The CD34+ cells were then FACS-sorted into the respective populations and HTS was used to determine the percentage of alleles containing indels; the frequencies of indels were equivalent among the populations indicating equivalent ZFN mRNA activity. To evaluate the efficacy of site-specific HDR in HSCs and progenitor cells, ZFN mRNA was co-delivered with either an IDLV or an oligodeoxynucleotide donor template to modify the single base-pair involved in SCD. We observed lower percentage of HDR-mediated gene modification in the HSC population compared to progenitors with all donor templates.

Due to the cell cycle phase restriction of HDR, we pre-stimulated CD34⁺ cells for 1-3 days prior to electroporation of ZFN mRNA and the oligo donor, and analyzed the cell cycle phases at the time of electroporation, and the frequencies of HDR and NHEJ produced by HTS. Only a small percentage of the immunophenotypic HSCs were in S/G2 phase after 24 hours of pre-stimulation; no HDR modification was observed in these cells. After 2-3 days of pre-stimulation, the HDR levels increased as the percentage of HSCs in S/G2 phase reached 20%. Importantly, assessment of relative cytotoxicity of the genome editing procedure (electroporation of ZFN mRNA and oligo donor) revealed a heightened sensitivity of HSCs/MPPs compared to progenitors, resulting in ~80% cell death in HSC vs. ~30% in progenitors under the conditions we are using. Transient expression of BCL-2 mRNA, co-electroporated with the genome editing reagents, improved HSC survival and significantly increased the numbers of HDR gene-corrected HSCs both in vitro and in vivo.

Conclusions: These data indicate an elevated sensitivity to cytotoxicity from the gene editing process for HSCs compared to the mature progenitor cells under our conditions, which may explain the lower levels of gene modification seen using in vivo compared to in vitro assays. Transient overexpression of BCL-2 mRNA preserves HSC survival after HDR-based gene editing, increasing the frequency of gene-corrected HSCs.