Knock out of CD123 or CLL-1 By CRISPR-Cas9 Editing from Human Hematopoietic Stem Cell Transplants Provide New Possibilities for Increasing Therapeutic Index and Safety for AML Treatment

Acute myeloid leukemia (AML) is a clonal disorder of hematopoiesis and the most common form of acute leukemia in adults. Most patients with AML relapse despite intensive chemotherapy. Allogeneic hematopoietic cell transplant (HCT) has become the standard of care for patients with intermediate or adverse genetics, with >3,500 transplants performed annually in the US. However, leukemia relapse post-HCT occurs in ~40% of these patients with a 2-year survival rate at <20%, necessitating new approaches to reduce relapse and improve overall outcomes.

Antigen-specific immunotherapies require cell surface antigens to be uniquely expressed on cancer cells to minimize "on-target, off-tumor" toxicity. CD123 and CLL-1 are highly expressed on normal myeloid cells, thus impeding the use of therapies targeting these antigens due to myelotoxicity. To circumvent such toxicity, we aim to create CD123 or CLL-1 negative human hematopoietic stem and progenitor cells (hHSPCs) for HCT such that this can be combined with subsequent use of targeted therapy against these antigens to prevent post-HCT relapse. Here, we present the pre-clinical evaluation of engineered hHSPCs, derived from mobilized peripheral blood of healthy donors, where CD123 or CLL-1 proteins were ablated by CRISPR/Cas9 gene editing. We have identified highly efficient guide RNAs that result in >80% on-target editing and can achieve greater than 90% biallelic gene knockout (KO). Deep sequencing followed by hybrid capture of up to 1000 potential off-target sites predicted to contain a maximum of 5 mismatches revealed minimal to no detectable off-target editing events. CD123 or CLL-1 KO hHSPCs showed >85% cell viability post-editing. Loss of CD123 or CLL-1 protein did not impact the differentiation of hHSPCs into granulocytic, monocytic, or erythroid lineages in vitro. Additionally, myeloid cells derived from CD123 or CLL-1 KO hHSPCs also retained their function, demonstrating similar phagocytotic capacity and the production of inflammatory cytokines in response to TLR agonists compared to unedited control cells. Importantly, CD123 or CLL-1 KO hHSPCs xenotransplanted into NSG or NBSGW mice showed no defect in long-term engraftment (human chimerism in NBSGW mice 16 weeks post transplant: 92.3±5.1% for CD123 KO, 91.2±4.1% for CLL-1 KO, and 93.4±3.1% unedited hHSPC engrafted mice, n=15 each). This in vivo model confirmed no observable impact on multilineage differentiation after CD123 or CLL-1 deletion either, supported by equal distribution of 10 hematopoietic lineages between groups. Bone marrow analyses revealed persistence of high gene editing frequency after 16 weeks (71±10% from CD123 KO and 86±6% from CLL-1 KO engrafted mice, n=15 each, compared to original 83% input CD123 KO and 86% input CLL-1 hHSPCs). Sorted subpopulations, including granulocytes, monocytes, dendritic cells, and mast cells, from the bone marrow similarly retained high levels of editing. Together these data suggest that there was no counter-selection against CD123 or CLL-1 KO cells, no outgrowth of any cells with particular edits or hematopoietic lineages to warrant tumorigenic concerns. Most notably, proof-of-concept experiments showed that cells depleted of CD123 or CLL-1 protein are indeed protected from the cytotoxicity of CD123 or CLL-1 targeted CAR-T cells.

In conclusion, we demonstrate that CD123 or CLL-1 negative human HSPCs can successfully carry out functional hematopoiesis that is resistant to CD123 or CLL-1 targeted therapies. Our findings provide a next-generation HCT strategy that supports the safe and effective use of antigen-directed immunotherapy treatments for patients with AML.