Comparative Studies Reveal Robust HbF Induction By Editing of HBG1/2 Promoters or BCL11A Erythroid-Enhancer in Human CD34+ Cells but That BCL11A Erythroid-Enhancer Editing Is Associated with Selective Reduction in Erythroid Lineage Reconstitution in a Xenotransplantation Model

Beta hemoglobinopathies resulting from dysfunctional or deficient adult beta-globin expression are some of the most prevalent inherited blood disorders in the world. Upregulation of developmentally-silenced fetal gamma-globin would replace adult beta-globin to ameliorate disease symptoms. One of the approaches to reactivate fetal globin expression in erythroid cells is through gene editing by zinc finger or CRISPR-Cas9 nucleases to disrupt the expression of a transcription factor BCL11A, which mediates fetal globin silencing. As BCL11A-deficiency leads to hematopoietic stem cells (HSCs) defects, the current editing approaches target the BCL11A erythroid-enhancer region located in intron 2 of the BCL11A gene to selectively reduce BCL11A expression in erythroid cells.

Instead of targeting BCL11A, we sought to identify novel cis-regulatory elements at the beta-globin locus for targeted gene editing to achieve fetal globin reactivation. From a lenti-CRISPR mediated saturated mutagenesis screen covering the beta-globin locus using Human Umbilical Cord Blood-Derived Erythroid Progenitor (HUDEP)-2 cells, multiple fetal hemoglobin (HbF)-inducing genomic domains were identified. Most of the hits were concentrated at the gamma-globin (HBG1/2) promoters, clustered at known hereditary persistence of fetal hemoglobin (HPFH) mutation hotspots. In-depth genotype to phenotype analysis further defined the indels responsible for HbF induction in these subdomains. We interrogated multiple families of nucleases and guide RNA (gRNA) combinations with or without single-stranded oligodeoxynucleotides (ssODN) to guide editing outcome. gRNAs were selected based on their HbF induction potential (up to 40%) when introduced into mobilized peripheral blood (mPB) CD34+ hematopoietic stem and progenitor cells (HSPCs) as ribonucleoprotein (RNP) complexes. HSPCs transfected with RNPs targeting either the BCL11A erythroid-enhancer or the HBG1/2 proximal regions were then injected into NBSGW mice to study the editing in SCID-repopulating cells (SRC) and their multilineage reconstitution potential.

All groups achieved high levels of human chimerism (>70% hCD45+/hCD45+mCD45) and comparable monocytes, granulocytes, B lymphocytes, and hCD34+ HSPCs reconstitution. However, BCL11A-edited cells showed selective reduction in erythroid lineage (CD235a+) output, up to 4-fold lower than untreated or HBG1/2 promoter-edited HSPCs. Sequencing analysis from lineage-specific sorted cells further revealed reduced editing levels at BCL11A erythroid-specific enhancer in the erythroid compartment compared to unfractionated bone marrow (BM) or other human lineages (70% in erythroid vs. 90% in all other lineages). Furthermore, the nonproductive fraction of indels that did not disrupt the BCL11A GATAA motif was significantly enriched in erythroid cells (22% in erythroid vs. 8% in other lineages). Ex vivo erythroid cultures suggests BCL11A erythroid-enhancer editing may lead to slightly increased apoptosis during erythroid differentiation. In contrast, HBG1/2 promoter-edited cells had similar editing levels and indel patterns across all lineages with no significant lineage skewing. When chimeric BM from HBG1/2 promoter-edited groups were cultured in erythroid conditions, ex vivo-derived erythroid cells had significantly elevated levels of HbF compared to controls. When CD235a+ cells were sorted without further culture from chimeric BM of mice engrafted with HBG1/2 promoter-edited cells, significantly increased levels of HbF were detected by UPLC (up to 30%) compared to controls (~6%). Thus, long-term HSCs have been edited productively at the HBG1/2 promoters. These cells were able to generate erythroid progenitors that harbor HbF-inducing indels, which in turn, gave rise to erythroid cells in vivo with a clinically-relevant levels of HbF in a xenotransplantation model.

Together, our data suggest that BCL11A-edited cells have an erythroid differentiation defect or survival disadvantage in NBSGW mouse model that warrants further investigation. In contrast, editing of the HBG1/2 promoters in mPB CD34+ cells achieved sustained HbF expression in erythroid lineage while maintaining multilineage differentiation potential. Targeting of the HBG1/2 promoters in HSPCs may be an attractive strategy for the development of potential gene editing medicines for beta hemoglobinopathies.