Search

Induction of fetal hemoglobin (HbF) via clustered regularly interspaced short palindromic repeats/Cas9–mediated disruption of DNA regulatory elements that repress γ-globin gene (HBG1 and HBG2) expression is a promising therapeutic strategy for sickle cell disease (SCD) and β-thalassemia, although the optimal technical approaches and limiting toxicities are not yet fully defined. We disrupted an HBG1/HBG2 gene promoter motif that is bound by the transcriptional repressor BCL11A. Electroporation of Cas9 single guide RNA ribonucleoprotein complex into normal and SCD donor CD34(+) hematopoietic stem and progenitor cells resulted in high frequencies of on-target mutations and the induction of HbF to potentially therapeutic levels in erythroid progeny generated in vitro and in vivo after transplantation of hematopoietic stem and progenitor cells into nonobese diabetic/severe combined immunodeficiency/Il2rγ(−/−)/Kit(W41/W41) immunodeficient mice. On-target editing did not impair CD34(+) cell regeneration or differentiation into erythroid, T, B, or myeloid cell lineages at 16 to 17 weeks after xenotransplantation. No off-target mutations were detected by targeted sequencing of candidate sites identified by circularization for in vitro reporting of cleavage effects by sequencing (CIRCLE-seq), an in vitro genome-scale method for detecting Cas9 activity. Engineered Cas9 containing 3 nuclear localization sequences edited human hematopoietic stem and progenitor cells more efficiently and consistently than conventional Cas9 with 2 nuclear localization sequences. Our studies provide novel and essential preclinical evidence supporting the safety, feasibility, and efficacy of a mechanism-based approach to induce HbF for treating hemoglobinopathies.

Targeting the BCL11A erythroid enhancer by gene editing is a promising approach to fetal hemoglobin induction for beta-hemoglobinopathies. HbF levels vary widely among individuals, suggesting potential heterogeneity in HbF responses after therapeutic intervention. We hypothesize that maximizing both gene edit frequency and HbF induction potential could promote consistently favorable clinical outcomes. Here we compared CRISPR-Cas9 endonuclease editing of the BCL11A +58 enhancer with alternative gene modification approaches, including +55 erythroid enhancer editing alone or in combination with the +58 enhancer, as well as editing targeting the HBG1/2 promoter -115 BCL11A binding site and transduction by an shRNA knocking down the BCL11A transcript in erythroid precursors. We found that combined targeting of the BCL11A +58 and +55 enhancers with 3xNLS-SpCas9 and two sgRNAs resulted in the most potent HbF induction (52.4%±6.3%) of tested approaches (BCL11A +58 editing alone, 29.1%±3.9%; BCL11A +55 editing alone, 34.8±5.1%; HBG1/2 promoter editing, 34.1% ±5.4%; shmiR-BCL11A, 32.2%±4.4%; mock, 7.6%±3.4%). Based on assays in bulk and single cell derived erythroid cultures and xenografted immunodeficient mice, we found that disruption of core half E-box/GATA motifs at both the +58 and +55 enhancers was associated with greatest HbF induction, whether by small indels, interstitial 3.1 kb deletion, or 3.1 kb inversion. Rare gene edited clones with alleles that only partially disrupted these motifs were associated with intermediate HbF induction phenotypes. Combined editing of BCL11A +58 and +55 enhancers was compatible with HSC self-renewal in primary and secondary xenotransplant, with intact lymphoid, myeloid and erythroid repopulation. We conducted gene-edited cell product manufacturing process development and developed conditions using a MaxCyte electroporation instrument achieving mean 97.3±1.8% gene edits and 88.9%±6.4% viability 24 hours after electroporation in 3 engineering runs at clinical scale. We obtained similar results at small-scale with plerixafor-mobilized HSPCs from sickle cell disease (SCD) donors or G-CSF mobilized PBMCs from transfusion-dependent beta-thalassemia (TDT) donors, including 94.2%±4.4%, 99.5%±0.3% and 91.8%±6.3% of gene edits in engrafting cells from NBSGW 16 week mouse bone marrow of healthy, SCD and TDT donors respectively. Off-target analyses by pooled amplicon sequencing of 601 candidate off-target sites for the +58 and +55 targeting sgRNAs, nominated by a range of computational (CRISPRme) and experimental (GUIDE-seq and ONE-seq) methods, did not identify reference genome off-target edits at a sensitivity of 0.1% allele frequency. We evaluated +58/+55 enhancer combined targeting in nonhuman primates by performing ribonucleoprotein (RNP) electroporation in rhesus macaque mobilized peripheral blood CD34+ HSPCs with autologous re-infusion following busulfan myeloablation. We observed highly efficient gene edit frequency (85.2%, 88.8% and 84.9%) and durable HbF induction (26.4%, 57.5%, and 45.9% F-cells and 12.7%, 41.9%, and 28% gamma-globin) in the peripheral blood in 3 animals at most recent recorded time point post infusion (127, 78, and 54 weeks respectively). Single colony analyses and bulk ddPCR and unidirectional sequencing demonstrated that the long-term engrafting cells displayed a similar distribution of indels, 3.1 kb deletions, and 3.1 kb inversions as the input cell products. Erythroid stress due to hydroxyurea treatment, with or without phlebotomy, was associated with substantially augmented HbF responses (to 75.9%, 88.2%, and 57.8% F-cells and 47.9%, 68%, and 35.7% gamma-globin). No hematologic or other toxicities attributable to gene editing were observed. Together these results suggest that combined BCL11A +58 and +55 erythroid enhancer editing produces highly efficient on-target allelic disruption, erythroid-specific BCL11A downregulation, heightened HbF induction capacity compared to alternative approaches, preserved long-term multilineage engraftment potential by both human xenotransplant and rhesus autotransplant assays, and absence of evident genotoxicity, under clinically relevant SpCas9 RNP electroporation conditions. Disclosures No relevant conflicts of interest to declare.

CRISPR gene editing holds great promise to modify somatic genomes to ameliorate disease. In silico prediction of homologous sites coupled with biochemical evaluation of possible genomic off-targets may predict genotoxicity risk of individual gene editing reagents. However, standard computational and biochemical methods focus on reference genomes and do not consider the impact of genetic diversity on off-target potential. Here we developed a web application called CRISPRme that explicitly and efficiently integrates human genetic variant datasets with orthogonal genomic annotations to predict and prioritize off-target sites at scale. The method considers both single-nucleotide variants (SNVs) and indels, accounts for bona fide haplotypes, accepts spacer:protospacer mismatches and bulges, and is suitable for personal genome analyses. We tested the tool with a guide RNA (gRNA) targeting the BCL11A erythroid enhancer that has shown therapeutic promise in clinical trials for sickle cell disease (SCD) and β-thalassemia (Frangoul et al. NEJM 2021). We find that the top predicted off-target site is produced by a non-reference allele common in African-ancestry populations (rs114518452, minor allele frequency (MAF) = 4.5%) that introduces a protospacer adjacent motif (PAM) for SpCas9. We validate that SpCas9 generates indels (~9.6% frequency) and chr2 pericentric inversions in a strictly allele-specific manner in edited CD34+ hematopoietic stem/progenitor cells (HSPCs), although a high-fidelity Cas9 variant mitigates this off-target. This report illustrates how population and private genetic variants should be considered as modifiers of genome editing outcomes. We expect that variant-aware off-target assessment will be required for therapeutic genome editing efforts going forward, including both ongoing and future clinical trials, and we provide a powerful approach for comprehensive off-target prediction. CRISPRme is available at crisprme.di.univr.it. Disclosures No relevant conflicts of interest to declare.

Patients with Familial Platelet disorder (FPD) have a germline RUNX1 mutation and are at high risk to developing hematologic malignancies (HM), primarily myelodysplastic syndrome and acute myeloid leukemia (lifetime risk~40%). To understand how germline RUNX1 mutations predispose to HM in vivo, we developed a Runx1 R188Q/+ mouse strain , mimicking the FPD-associated R201Q missense mutation. Analysis of the bone marrow cells in Runx1 R188Q/+ mice revealed a significant increase in the total number of bone marrow cells. Immunophenotypic analysis using Sca-1 and Cd86 markers revealed a significant increase in Sca-1 expression in hematopoietic stem and multi-potential progenitor cells, indicating a systemic inflammation in the bone marrow. In addition, the frequency of common-myeloid, granulocytic-monocytic and granulocytic progenitor cells were found significantly increased in the Runx1 R188Q/+ bone marrow. Accordingly, their colony-forming unit capacity was increased when compared to wildtype controls (wt/Runx1 R188Q/+ CFU average = 45/85), indicating a myeloid bias. The number and size of platelets were not altered in Runx1 R188Q/+ mice. However, platelet function was significantly reduced. The activation of the Cd41/Cd61 fibrinogen receptor complex in membrane after thrombin treatment was reduced in Runx1 R188Q/+ platelets. Similarly, the translocation of P-selectin by alpha granules and the secretion of serotonin by the dense granules were also reduced. Hematopoietic progenitor cells isolated from Runx1 R188Q/+ mice revealed a significant reduction in DNA-damage repair response in vitro. Quantitative analysis of nuclei with 53bp1-positive foci in response to ionizing radiation showed a marked increase in 53bp1-positive foci in Runx1 R188Q/+ nuclei, suggesting that Runx1 R188Q/+ cells have a defective repair of double strand DNA breaks. Furthermore, expression of DNA-damage repair pathway-associated Pmaip1 (Noxa) was significantly reduced in irradiated Runx1 R188Q/+ hematopoietic progenitor cells. To understand underlying mechanism responsible for the observed myeloid bias in Runx1 R188Q/+ cells, transcription profiling analysis was performed in myeloid progenitors from wildtype and Runx1 R188Q/+ mice, utilizing RNA-sequencing. A total of 39 genes were significantly deregulated (> 1.5 FC; FDR To study the FPD-associated pre-leukemic process in vivo, wildtype and Runx1 R188Q/+ mice were monitored for 20 months. Although Runx1 R188Q/+ mice remained healthy for 18 months, somatic mutations in their leukocytes were evident at 12 months. Targeted sequencing of 578 cancer genes (mIMPACT panel) in leukocyte DNA of two Runx1 R188Q/+ mice identified somatic mutations in Kdm6a, Setd1b, Amer1, and Esco1 (variant allele frequencies between 0.5% and 2.8%). These mutations were confirmed at stable frequency for eight following months. Since loss of the second Runx1 allele is a frequent somatic event in progression to FPD/HM, we evaluated the predisposition to HM in Mx1Cre-Runx1 R188Q/fl mice over time. Unlike Runx1 R188Q/+ mice, Runx1 R188Q/Δ mice succumbed to myeloid leukemia with a median latency of 37.5 weeks and full penetrance. In addition, the expression of oncogenic Nras-G12D, in Runx1 R188Q/Δ mice reduced the median latency to 14.7 weeks. These studies demonstrate that FPD-associated Runx1 germline mutations induce inflammation in hematopoietic stem cells, induce myeloid expansion with defective DNA-damage response and predispose to HM over time. These studies suggest that anti-inflammatory therapies in pre-symptomatic FPD patients may reduce clonal expansion and predisposition to HM. Disclosures Ebert: Exo Therapeutics: Membership on an entity's Board of Directors or advisory committees; Skyhawk Therapeutics: Membership on an entity's Board of Directors or advisory committees; Celgene: Research Funding; Deerfield: Research Funding; GRAIL: Consultancy. Levine: Isoplexis: Membership on an entity's Board of Directors or advisory committees; Auron: Membership on an entity's Board of Directors or advisory committees; C4 Therapeutics: Membership on an entity's Board of Directors or advisory committees; Zentalis: Membership on an entity's Board of Directors or advisory committees; Celgene: Research Funding; QIAGEN: Membership on an entity's Board of Directors or advisory committees; Ajax: Membership on an entity's Board of Directors or advisory committees; Imago: Membership on an entity's Board of Directors or advisory committees; Mission Bio: Membership on an entity's Board of Directors or advisory committees; Gilead: Honoraria; Prelude: Membership on an entity's Board of Directors or advisory committees; Janssen: Consultancy; Lilly: Honoraria; Morphosys: Consultancy; Roche: Honoraria, Research Funding; Incyte: Consultancy; Astellas: Consultancy; Amgen: Honoraria.

Signal transduction pathways regulate the proliferation and viability of acute myeloid leukemia (AML) blasts. The regulation in the expression of cytokine receptors in AML is not well understood. In this study, we investigated how the CBFβ-SMMHC fusion protein regulates expression of cytokine receptors in inv(16) AML, with focus on the co-receptor Neuropilin-1 (NRP1). Knock-down of CBFβ-SMMHC expression, utilizing shRNA transduction, induced G1 phase of cell cycle arrest and reduced the viability of inv(16) ME-1 cells in culture. Expression profile analysis of CBFβ-SMMHC knock-down cells revealed a significant repression of genes associated with transmembrane receptor protein kinase pathways, including NRP1 (-5 fold), FGFR1 (-4.2 fold) , FLT3 (-2 fold) and TGFBR2 (-1.2 fold). The expression of NRP1 was significantly upregulated in inv(16) AML cases when compared to other AML sub-types and to hematopoietic stem and progenitor cells. Functionally, NRP1 knock-down reduced the viability of ME-1 cells with a similar dynamics as when using CBFβ-SMMHC shRNAs. In addition, the proliferation of inv(16) AML cells was reduced 4.1-fold when treated with a function-blocking antibody for the FV/VIII extracellular NRP1 domain, while having no effect in non-inv(16) AML cells or when using blocking antibody for the CUB extracellular domain. Furthermore, deletion of Nrp1 by gene editing reduced the colony-forming unit capacity of primary mouse Cbfb +/MYH11 leukemic cells and extended the median leukemia latency in vivo. To identify the genes regulated by NRP1 in inv(16) AML, we analysed the transcription profile of NRP1 knock-down in ME-1 cells. Gene Set Enrichment and Pathway Analysis revealed a repression in STAT5 pathway, and in signalling receptor activity, including FLT3 (-1.8 fold) and TGFBR2 (-1.8 fold) expression, indicating that NRP1 mediates transcriptional regulation of FLT3 and TGFBR2 expression in inv(16) AML. Furthermore, the regulation of FLT3 and TGFB2 expression by CBFβ-SMMHC and by NRP1 was validated by gene editing in inv(16) AML blasts. Accordingly, NRP1 knock-down in AML cells reduced SMAD2/3 phosphorylation. The repression of RUNX1/CBFβ function, using small molecule inhibitors, in inv(16) AML cells with CBFβ-SMMHC knockdown restored NRP1 expression, suggesting that RUNX1 may repress NRP1 expression in AML cells. To evaluate if RUNX1 directly regulates NRP1 expression, we tested RUNX1 binding in the NRP1 locus of AML cells with CBFβ-SMMHC knockdown. RUNX1 binding at one of six sites with RUNX1 occupancy identified by chromatin immunoprecipitation followed by sequencing (RE5, regulatory element 5) was increased in the CBFβ-SMMHC knock-down cells. The RE5 is located 178kb upstream of the NRP1 transcription start site and it is evolutionarily conserved in vertebrates. The deletion of RE5 by gene editing (~50% editing efficiency) increased NRP1 expression 1.8-fold, suggesting that RUNX1 may repress NRP1 expression at by binding to the RE5 enhancer. Taken together, these studies demonstrate that CBFβ-SMMHC regulates expression of cytokine receptors in inv(16) AML. Specifically, it directly regulates expression of the co-receptor NRP1, which is essential for AML survival, acting (at least in part) by regulating FLT3 and TGFB pathways. Disclosures Guzman: SeqRx: Consultancy; BridgeMedicines: Consultancy; Cellectis: Membership on an entity's Board of Directors or advisory committees; Daiichi Sankyo: Membership on an entity's Board of Directors or advisory committees.

Elevated fetal hemoglobin (HbF, α2γ2) levels are clinically beneficial for patients with β-hemoglobinopathies. Editing of the erythroid-specific BCL11A enhancer induces HbF, inhibiting sickling and restoring globin chain balance in erythroid cells derived from hematopoietic stem and progenitor cells (HSPCs) from SCD and β-thalassemia patients respectively, without detectable genotoxicity or adverse effects on hematopoietic stem cell (HSC) function (Wu Y, Nat Med, 2019). Here, we sought to evaluate engraftment and HbF induction potential of erythroid-specific BCL11A enhancer edited CD34+ HSPCs in a non-human primate transplantation model in which hemoglobin switching is conserved. We targeted the erythroid-specific +58 DNAse I hypersensitive site of BCL11A, which has identical human and rhesus sequences at the spacer and protospacer adjacent motif (PAM) of the potent #1617 sgRNA. Ribonucleoprotein complex (RNP) composed of 3x-NLS SpCas9 protein and either BCL11A enhancer targeting (#1617) or AAVS1 targeting sgRNA was electroporated into rhesus CD34+ HSPCs (n=3). Following erythroid differentiation, substantial γ-globin expression (54-77%, p We tested BCL11A enhancer editing with autologous HSC transplant in two cohorts, with two macaques per cohort. For cohort 1, we performed competitive engraftment of BCL11A enhancer and AAVS1 edited HSPCs to test long-term reconstitution. For cohort 2, we evaluated BCL11A enhancer editing alone to evaluate HbF induction and hematopoietic reconstitution. For each cohort, purified CD34+ HSPCs were electroporated with RNP one day after G-CSF and plerixafor mobilization and cultured for two days prior to cryopreservation. HSPCs were thawed and infused following 2×5 Gy total body irradiation. For cohort 1 (n=2, ZL25 and ZL22, 1.34-1.39×106 CD34+ HSPCs/kg), we observed reduced indel frequencies (8-41%) at early post-infusion time points compared to cell products (18-49%), suggesting indels in unfractionated HSPCs may overestimate those in engrafting cells and/or hematopoietic ablation was incomplete. From weeks 6 to 83, stable indel frequencies were detected in both BCL11A (~3-18%) and AAVS1 (~10-45%), suggesting no selective advantage for BCL11A enhancer edited, AAVS1 edited, or non-edited HSCs. For cohort 2 (BCL11A enhancer editing alone (n=2, ZM17 and ZM26, 1.78-6.06×106 CD34+ cells/kg), cell products showed improved editing with ~95% indels and ~65-78% γ-globin protein after in vitro erythroid culture. Animals engrafted with typical kinetics and displayed stable indel ratios up to 28 weeks post-transplantation. A significant correlation was detected between γ-globin level and indel frequency comparing all 4 transplanted animals and unedited controls (R2=0.76, p In summary, we evaluated the clinical potential of autologous BCL11A erythroid enhancer editing in rhesus macaques. BCL11A enhancer edited HSCs can persist for at least 83 weeks post-transplant and provide therapeutic levels of HbF in peripheral RBCs without anemia or other apparent hematologic toxicity. Furthermore, these results emphasize input CD34+ HSPC dose and conditioning intensity as critical variables that influence gene editing following autologous HSCT. Overall, these findings support BCL11A erythroid enhancer genome editing as a promising strategy for therapeutic HbF induction. Disclosures Weiss: GlaxoSmithKline: Consultancy; Cellarity INC: Consultancy; Esperian: Consultancy; Beam Therapeutics: Consultancy; Rubius INC: Consultancy.

Severe congenital neutropenia (SCN) is a life-threatening disorder of insufficient granulocytes. Lifelong granulocyte colony-stimulating factor (G-CSF) injections are the mainstay of treatment, yet there remains a high risk of myelodysplastic syndrome and acute myeloid leukemia. The most common etiology of SCN is germline ELANE mutation. These dominantly acting mutations preserve expression but alter the structure of the neutrophil elastase protein product resulting in altered protein folding and/or trafficking with excess cell death at the promyelocyte/myelocyte stage of maturation. Recent advances in gene editing technologies have enabled targeted genetic modification of hematopoietic stem cells (HSCs); nonetheless genetic repair of specific disease-associated mutations remains challenging. We hypothesized that introduction of premature termination codons (PTCs) by nuclease-mediated frameshift mutations within early exons of ELANE could constitute a universal, highly efficient, simple therapeutic approach for ELANE-associated SCN. We predicted that the PTCs would trigger nonsense mediated decay (NMD) of the mutant transcript resulting in its loss of expression and thus bypassing neutrophil precursor cell death and consequent neutropenia. The mild phenotype observed in the Papillon-Lefevre syndrome, characterized by combined serine protease deficiency, suggests that isolated neutrophil elastase deficiency would not result in clinically significant immunodeficiency. We delivered 3xNLS-SpCas9 and ELANE targeting sgRNA as ribonucleoprotein (RNP) complexes to primary human CD34+ hematopoietic stem and progenitor cells (HSPCs) and conducted in vitro neutrophil maturation culture. Introducing indels at exon 2 of ELANE efficiently triggered NMD. Edited cells were fully competent for neutrophil maturation similar to neutral locus targeted control cells. Using three human donors, we found that ELANE exon 2 edited HSPCs produced similar human bone marrow (BM) chimerism as unedited cells in NBSGW recipient mice 16 weeks following infusion. We found similar lymphoid, erythroid, and myeloid engraftment including similar fraction of human neutrophils (13.4% of total human BM cells in unedited and 13.7% in ELANE exon 2 edited, despite 97.3% on-target indel frequency and 84.3% reduction in ELANE expression in the latter). Using CD34+ HSPCs from four ELANE mutant SCN patient donors, we demonstrated that exon 2 targeting RNPs achieve highly efficient editing exceeding 95% indel frequency, trigger ELANE transcript decay, and rescue promyelocyte stage maturation arrest. In contrast to these ameliorating early exon frameshifts, naturally occurring SCN-associated frameshifts affect late exons of ELANE, suggesting that these mutations might escape NMD. Indeed we found that targeting ELANE exon 5 in HSPCs resulted in robust indels (93.5%), preserving ELANE expression but resulting in cell death at the promyelocyte/myelocyte stages of development, recapitulating an SCN phenotype. To our surprise, we found that only -1 frameshifts and not -2 frameshifts induced by gene editing with NHEJ repair led to the SCN-like phenotype, although we noted that all 23 reported naturally occurring SCN-associated ELANE frameshift mutations result from -1 but not -2 bp frameshifts. Using xenograft of NBSGW recipients, we found that an RNP complex leading to efficient -1 frame indels in ELANE exon 5 produced profound neutrophil maturation block, with reduction from 13.4% neutrophils in controls to 0.5% neutrophils in ELANE exon 5 targeted recipients, with otherwise indistinguishable human monocyte, lymphoid, and erythroid reconstitution as compared to controls. This dramatic phenotype contrasts with mice engineered to express SCN-associated Elane mutations that do not exhibit neutropenia, indicating species differences in granulopoiesis. Together these results support the development of ELANE early exon targeting as a highly efficient universal therapy for ELANE mutant SCN, feasible with existing gene editing technology. Moreover, by late exon ELANE gene editing we have developed a robust new model of SCN using primary human HSPCs that recapitulates neutropenia in vivo following xenotransplant, refines the molecular genetics of mutant ELANE induced neutrophil maturation arrest, and offers opportunities to explore novel therapeutic approaches. Disclosures Newburger: TransCytos LLC: Consultancy; X4 Pharmaceuticals: Consultancy, Honoraria.

Induction of fetal hemoglobin (HbF, α2γ2) via genome editing-mediated disruption of DNA regulatory elements that repress expression of γ-globin genes (HBG1 and HBG2) is a promising therapeutic strategy for b-hemoglobinopathies including sickle cell disease (SCD) and β-thalassemia. Optimal technical approaches and safety profiles are yet to be fully defined. We used CRISPR/Cas9 to target a DNA repressor element near the distal CCAAT box of the HBG1/HBG2 promoters. This region contains a "TGACC" motif recognized by BCL11A, a transcriptional repressor protein that regulates γ-to-β globin switch after birth. Rare germline variants at or near this motif are associated with hereditary persistence of fetal hemoglobin, a benign genetic condition that alleviates the clinical manifestations of co-inherited b-hemoglobinopathies. Previously, we showed that transduction of human CD34+ cells with lentiviral vector encoding Cas9 and guideRNA (gRNA) targeting the HBG1/HBG2 promoter caused induction of HbF in red blood cell (RBC) progeny generated in vitro (Traxler et. al, Nature Medicine v22,2016). Here we present a clinically tractable approach for disrupting the HBG1/HBG2 BCL11A binding site in human hematopoietic stem cells (HSCs). Electroporation of Cas9:gRNA ribonucleoprotein (RNP) complex into healthy or SCD donor CD34+ cells resulted in up to 80% on-target insertion-deletion (indel) mutations and 35% HbF in erythroid progeny generated in vitro. Sixteen to 17 weeks after transplantation of gene edited CD34+ cells into immunodeficient NBSGW mice, up to 75% of donor CD34+ cells in recipient bone marrow contained on-target indels, demonstrating efficient modification of repopulating human HSCs. No differences in CD34+ cell regeneration or differentiation into erythroid, T, B, or myeloid cell lineages were observed between edited and control cells. Moreover, up to 78% of gene edited erythroid cells stained with anti-HbF antibody ("F-cells") compared to 15% in control erythroid cells, suggesting a "pan-cellular" pattern of HbF expression after editing. Strikingly, human donor-derived erythroid cells in recipient bone marrow expressed up to 40% HbF compared to 3% HbF in controls. Although the editing frequencies of HBG1 and HBG2 promoters varied between different donor CD34+ cells, an engineered variant of Cas9 containing 3 nuclear localization sequences (Wu et. al,Nature Medicine v25, 2019) edited repopulating HSCs more efficiently and consistently than conventional Cas9 with two nuclear localization signals. Simultaneous on-target RNP-induced DSBs at both HBG1 and HBG2 can result in the deletion of the intervening 4.9-kb region, leaving a single hybrid gene with HBG2 promoter sequences fused to the downstream HBG1 gene. We detected this deletion in approximately 30% of edited cells, with no associated decline in HbF expression determined by clonal analysis of erythroid colonies. No off-target mutations were detected by targeted sequencing of the 26 top candidate sites identified by CIRCLE-seq, an in vitro genome-scale method for detecting Cas9 activity. Analysis of gene edited human donor cells purified from mouse bone marrow showed no chromosomal rearrangements by G-banding (n=20) or fluorescence in situ hybridization with a probe located distal to the HBG1/HBG2 loci (n=225). Taken together, our studies provide novel and essential preclinical evidence supporting the safety, feasibility, and efficacy of a CRISPR-Cas9 genome editing approach to induce HbF for treating hemoglobinopathies. Figure. Gene editing of the HBG1/HBG2 promoters in HSCs and HbF induction of erythroid progeny in vivo. Plerixafor-mobilized CD34+ cells from an individual with SCD were edited with RNP and transplanted into NBSGW mice, which were analyzed 16-17 weeks later. A. On-target indel frequency before (Pre) and after bone marrow transplantation (BM). The black bars represent a 13-nucleotide deletion associated with human hereditary persistence of fetal hemoglobin. B. Human erythroblasts and reticulocytes derived from RNP-edited and non-edited Control CD34+ donor cells. Scale bar = 10 mm. C. HbF immunostaining control and RNP edited erythroid cells in recipient bone marrow assessed by flow cytometry. D. %HbF protein in hemolysates of control (C) and RNP edited erythroid cells assessed by ion-exchange HPLC. n= 3 biological replicates. **** P < 0.0001. Figure Disclosures Metais: MBIO: Other: St. Jude Children's Research Hospital has an existing exclusive license and ongoing partnership with Mustang Bio for the further clinical development and commercialization of this XSCID gene therapy. Sharma:Doris Duke Foundation: Research Funding; Vertex Pharmaceuticals: Other: Study PI. Weiss:Beam Therapeutics: Consultancy; Rubius INC: Consultancy; GlaxoSmithKline: Consultancy; Cellarity INC: Consultancy; Esperian: Consultancy.

Although therapeutic genome editing of autologous hematopoietic stem cells (HSCs) in principle could cure β-hemoglobinopathies, CRISPR-Cas9 mediated gene modification has demonstrated variable efficiency, specificity, and persistence in HSCs. Here we demonstrate selection-free on-target editing of the BCL11A erythroid enhancer by Cas9:sgRNA ribonucleoprotein in patient-derived HSCs as a nearly complete reaction lacking detectable genotoxicity or deleterious impact on stem cell function. First we screened a set of 20 guide RNAs targeting the functional core of the +58 BCL11A enhancer for maximal HbF induction by RNP delivery. We used SpCas9 protein with additional NLS sequences, synthetic modified sgRNA, and optimized electroporation buffer to produce >95% on-target indels disrupting a critical GATA1 binding site within the +58 BCL11A enhancer in CD34+ HSPCs. Clonal analysis showed that even 1 bp indels around the cleavage site were sufficient for HbF reactivation. Specificity was evaluated by CIRCLE-seq, a method to define genome-wide target sequences susceptible to RNP cleavage in vitro. Amplicon deep sequencing of 24 possible off-target sites from edited CD34+ cells did not reveal any off-target editing with limit of detection 0.1% allele frequency. Despite transient induction of a p53 transcriptional response peaking at 4-8 hours after RNP electroporation, we found no evidence of selection for TP53 or 94 other hematologic malignancy associated mutations by targeted deep sequencing. Edited CD34+ HSPCs from healthy donors contributed to multilineage engraftment of primary and secondary immunodeficient mouse recipients at similar levels as unedited control cells. Likewise we found that edited CD34+ HSPCs from a plerixafor-mobilized SCD donor contributed to marrow engraftment and multilineage hematopoiesis in immunodeficient NBSGW mice after 16 weeks at similar levels as unedited cells, with ~95% indel frequency for engrafting healthy and SCD donor cells. Edited engrafting SCD cells were similarly competent for secondary transplantation as unedited controls. Erythroid progeny of edited engrafting sickle cell disease HSCs expressed therapeutic levels of fetal hemoglobin (HbF) and resisted sickling. Erythroid progeny of edited CD34+ HSPCs from 7 transfusion-dependent ß-thalassemia donors showed restored globin chain balance and amelioration of microcytosis and poikilocytosis. We found that compared to the bulk CD34+ HSPC pool, HSCs preferentially underwent nonhomologous as compared to microhomology mediated end-joining repair based on three assays: sorting of CD34+ CD38- CD90+ CD45RA- cells; isolation of cells in G0, G1, S, and G2/M cell cycle phases; and evaluation of long-term engrafting cells in immunodeficient mice. Together these data show that NHEJ-based BCL11A enhancer editing approaching complete allelic disruption is a practicable therapeutic strategy to produce durable HbF induction in SCD and ß-thalassemia. Disclosures Esrick: Bluebird Bio: Honoraria. Williams:Bluebird Bio: Research Funding.

Chromatin complexes regulate gene expression in normal and malignant hematopoiesis. The significance of transcription factor deregulation on the control of epigenetic dynamics in leukemia is poorly understood. The leukemia fusion CBFβ-SMMHC is expressed in inv(16) acute myeloid leukemia (AML), and deregulates the activity of the transcription factor RUNX1. This fusion protein blocks myeloid differentiation, expands pre-leukemic myeloid progenitor cells, and drives AML development. The CBFβ-SMMHC inhibitor, AI-10-49, specifically disrupts its binding to RUNX1 resulting in an acute RUNX1 release, inducing apoptosis in inv(16) AML cells and delaying leukemia latency in mice. However, the mechanism by which AI-10-49 induces apoptosis is unknown. In this study, we utilize pharmacologic, genomic and genetic approaches to demonstrate a RUNX-dependent mechanism by which CBFβ-SMMHC maintains leukemia survival, and that this mechanism can be pharmacologically targeted in the treatment of inv(16) AML. Gene Set Enrichment Analysis of RNA-seq data from inv(16) AML cells treated with AI-10-49 identified the deregulation of a MYC signature, including cell cycle, ribosome biogenesis and metabolism, and the reduction of MYC transcript levels (10-fold). MYC shRNA knockdown induced apoptosis of inv(16) AML cells, and MYC expression rescued AI-10-49 induced apoptosis. Furthermore, mouse leukemia cells transduced with Myc shRNAs showed significant delay in leukemic latency upon transplantation, validating the requirement of MYC in inv(16) AML maintenance in vivo. Similarly, pharmacologic inhibition of MYC activity, using a combined treatment with AI-10-49 and the BET-family bromodomain inhibitor JQ1, revealed a strong synergy in inv(16) AML cells and significant delay in leukemia latency in mice. Analysis of chromatin immunoprecipitation followed by deep-sequencing (ChIP-seq) revealed that AI-10-49 treatment increased RUNX1 occupancy at three MYC distal enhancer elements (including two new enhancers: +0.18 Mb and +0.5 Mb, and the E3 enhancer of the BDME superenhancer at +1.7 Mb) downstream from MYC TSS but not at the MYC TSS or the T-cell leukemia associated N-ME enhancer. The open-chromatin and active nature of these regions was further validated by Assay for Transposase-Accessible Chromatin with high-throughput sequencing (ATAC-seq) and H3K27Ac ChIP-seq. Since the MYC locus is occupied by a number of tumor-specific enhancers, we analyzed the enhancer interactions in a 4 Mb region (-1Mb to +3 Mb of MYC TSS) by chromosome conformation capture carbon copy (5C) in inv(16) AML cells. The 5C analysis revealed that the three enhancers are physically connected with each other and with the MYC promoter. Again, the T-cell ALL associated N-ME enhancer was not active in these cells. Furthermore, the six-hour AI-10-49 treatment had not altered these interactions, indicating that MYC repression was probably not due to changes in open chromatin but to alterations in regulatory factors. Therefore, we analyzed the occupancy of chromatin complex components at the three enhancers, N-ME and MYC TSS regions, utilizing ChIP-quantitative PCR. We demonstrate that AI-10-49 treatment results in the repression of the SWI/SNF complex component BRG1 and recruitment of polycomb-repressive complex 1 (PRC1) component RING1B at the three MYC enhancers. This switch was associated with a reduction of active enhancer mark H3K4me1 and an increase of the repressive mark H3K27me3 at these sites. Finally, deletion of a small region in each of these enhancers surrounding the RUNX1 binding site (41 bp, 67 bp and 295 bp) by genome editing (CRISPR/Cas9) reduced MYC transcript levels by 50% and the viability of inv(16) AML cells in vitro, indicating that each one of these enhancers plays a critical role in regulating MYC levels and sustaining the survival of inv(16) AML cells. Our results demonstrate that CBFβ-SMMHC regulates MYC levels by repression of RUNX1 activity at three distal MYC enhancers in inv(16) AML cells. AI-10-49 treatment induces an acute release of RUNX1, increases RUNX1 occupancy at the distal enhancers, and disrupts enhancer chromatin dynamics which in turn induces apoptosis by repressing MYC. Furthermore, this study suggests that combined treatment of inv(16) AML with AI-10-49 and BET-family inhibitors may represent a promising targeted therapy. Disclosures No relevant conflicts of interest to declare.