Base Editing Eliminates the Sickle Cell Mutation and Pathology in Hematopoietic Stem Cells Derived Erythroid Cells

Sickle cell disease (SCD) is a chronic, life-altering multisystem disorder that affects millions of individuals worldwide. Strategies for genetic therapy of autologous SCD hematopoietic stem cells (HSCs) include lentiviral vector (LV) delivery of an anti-sickling β-like globin gene and genome editing, either to revert the SCD mutation by homology-directed repair (HDR) or to induce the expression of fetal hemoglobin (HbF, α2γ2) in red blood cell (RBC) by non-homologous end-joining repair (NHEJ). Base editing is a newer technology that offers the potential for facile generation of more precise genetic alterations with improved safety features. Adenosine base editors (ABEs) convert A-T base pairs to G-C pairs at loci targeted by a single guide (sg) RNA and Cas9 nickase. In contrast to LV vectors and standard Cas9 genome editing, ABEs function through mechanisms that are independent of double-stranded DNA breaks (DSBs), which can cause large deletions, structural DNA rearrangements and TP53-mediated DNA damage responses leading to cell death or malignant transformation.

Base editors cannot generate the T-to-A transversion required to revert the mutant SCD codon (Val, GTG) to wild-type (Glu, GAG). However, ABEs can convert the Val codon to Ala (GCG) to generate the naturally occurring, non-sickling variant hemoglobin "Makassar" (HbG). Hemoglobin Makassar heterozygotes and 1 reported homozygote exhibit normal RBC indices, indicating that the variant is benign. We used protein directed evolution to generate a new ABE (ABE8e-NRCH) that accesses a nearby CACC PAM to convert HbS alleles to HbG efficiently in heterologous cells. To demonstrate therapeutic proof of concept, we electroporated ABE8e-NRCH mRNA and targeting sgRNA or ABE8e-NRCH/sgRNA ribonucleoprotein (RNP) complex into 3 different SCD donor CD34+ hematopoietic stem and progenitor cells (HSPCs). After 48 hours, conversion of the HbS allele to HbG was 58±5% with ABE8e-NRCH mRNA/sgRNA and 34±5% with RNP (n=3). On target editing was maximal at 144 hours: 80±2% with ABE8e-NRCH mRNA/sgRNA and 44±7% with ABE8e-NRCH protein (n=3). The indel rate resulting from inadvertent DSBs was <1%, and missense bystander mutations were not detected. In vitro differentiation of base-edited CD34+ cells generated late stage erythroid cells with 76% HbG/14% HbS protein after editing with ABE8e-NRCH mRNA/sgRNA, and 52% HbG/37% HbS after editing with RNP. The sickling rate of in vitro-generated reticulocytes exposed to hypoxia (2% O₂) was 52% in controls derived from unedited CD34+ cells, 16% after editing with ABE8e-NRCH mRNA/sgRNA and 27% after editing with RNP.

To assess base editing in repopulating HSCs, we transplanted ABE8e-NRCH-edited SCD CD34+ cells into NBSGW mice. The editing frequency before transplantation was ~64% for ABE8e-NRCH mRNA/sgRNA and 40% for RNP at 48 hours after electroporation. Mice were euthanized after 16 weeks and base editing was analyzed in human donor-derived cells. Overall editing rates were preserved in repopulating cells: 68%±3% for ABE8e-NRCH mRNA/sgRNA and 36%±3% for RNP (n=4). The editing frequencies were similar in donor cell-derived myeloid, erythroid, and B cell-lineages, indicating that base editing did not alter hematopoietic development. Bone marrow erythroblasts derived from base-edited and control CD34+ HSPCs exhibited similar maturation profiles and enucleation. Erythroblasts generated in vivo from HSPCs from SCD patient exhibited potentially therapeutic levels of HbG protein: 58±3% with ABE8e-NRCH mRNA/sgRNA and 27%±3% with RNP (n=4). Adenine base editor conversion of the HbS allele to the Makassar variant in autologous HSCs represents a new therapeutic approach for SCD.