Therapeutic Base Editing of Human Hematopoietic Stem Cells

Base editing with fusions of RNA-guided DNA-binding proteins and nucleotide deaminases represents a promising approach to permanently remedy genetic blood disorders without obligatory double strand breaks, however its application in engrafting hematopoietic stem cells (HSCs) remains unexplored. Here we purified A3A(N57Q)-BE3 protein for RNP electroporation of human peripheral blood (PB) mobilized CD34+ hematopoietic stem and progenitor cells (HSPCs). We found that sgRNAs targeting for cytidine base editing the core GATA1 binding motif of the BCL11A +58 erythroid enhancer resulted in efficient on-target base edits (81% allele frequency) with low indels. There was similar HbF induction in erythroid progeny as compared to Cas9:sgRNA RNP nuclease mediated modification of the same target sequence (36% median HbF in base edited cells with 81% allele modifications, 39% HbF in 3xNLS-SpCas9:sgRNA#1617 nuclease edited cells with 99% indels, and 5%HbF in unedited cells). A single therapeutic base edit of the BCL11A enhancer was sufficient to ameliorate the pathobiology of both sickle cell disease (SCD) and beta-thalassemia. Base editing of CD34+ HSPCs from plerixafor mobilized PB SCD patient donors resulted in potent HbF induction (24-31% HbF level in bulk cells with 87-90% on-target allele modifications) and reduced sodium metabisulfite induced sickling (from 84% to 29% of erythrocytes). Base editing of non-mobilized PB CD34+ HSPCs from 3 beta-thalassemia patient donors potently induced gamma-globin and improved ineffective erythropoiesis in vitro, with 59-75% on-target allele modifications. Moreover highly efficient multiplex editing could be achieved in HSPCs. Simultaneous disruption of the BCL11A +58 erythroid enhancer core GATA1 motif and correction of the HBB -28A>G promoter mutation in HSPCs resulted in production of at least one therapeutic allele in 35 of 35 erythroid colonies analyzed, with each colony achieving improved alpha-to-non-alpha globin chain balance. Finally we found that base editing could be efficiently produced in engrafting HSCs. Using 3 donors, we observed 62-64% base edits at the BCL11A +58 enhancer GATA1 motif in input HSPCs with one round of electroporation and 73-85% base edits with two rounds of electroporation (separated by 24 hours), albeit with modest reduction of cell viability with tandem electroporation. We observed 70-94% human chimerism and 35-82% base edits in engrafting human cells in primary recipients and 11-61% base edits in secondary recipients as measured 16 weeks following cell infusion. Similar edit frequencies were observed across all assayed hematopoietic lineages including engrafting B-lymphocytes, erythroid precursors, and HSPCs. We observed that compared to non-engrafting progenitors, long-term engrafting HSCs favored C to T (as compared to C to G/A) editing. There was a strong correlation (Spearman r 0.86, P<0.0001) between base edit frequency and HbF level of engrafting erythroid cells, with in vivo HbF level induced from median 1.8% in unedited recipients to 22% in recipients of tandem base edited cells. Together these results demonstrate, to our knowledge for the first time, the potential of RNP base editing of human HSPCs as an alternative to nuclease editing for HSC-targeted therapeutic genome modification.