Background: Acute myeloid leukemia (AML) is a heterogeneous disease characterized by abnormal clonal expansion and is the most common form of adult acute leukemia. Though hematopoietic stem cell transplant is the standard of care for high-risk AML patients, relapse post-transplantation occurs in 40% of these patients, highlighting the need for new therapeutic approaches such as immunotherapy.
Flow cytometric profiling of 26 AML patient samples demonstrated that CD33 is expressed on >94% of blasts/LSCs (leukemic stem cells) while CLL-1 is expressed on >85% of blasts/LSCs at density levels targetable by immunotherapies (1283-2260 molecules per blast), suggesting that immuno-targeting both CD33 and CLL-1 can address AML heterogeneity and reduce chances of tumor resistance. Targeting these antigens, however, can lead to cytopenia due to shared expression on normal hematopoietic cells.
Aims: We propose to delete both CD33 and CLL-1 from hematopoietic stem cell grafts, thereby restricting these antigens to AML cells only to enable subsequent immunotherapy without risk of on-target off-tumor toxicities.
Methods: Mobilized peripheral blood CD34+ human hematopoietic stem and progenitor cells (hHSPCs) were gene edited using CRISPR/Cas9 with gRNAs against CD33 and CLL-1, resulting in biallelic deletion of both genes in >80% cells.
Results: The viability of multiplex-edited hHSPCs was >90% with no impact on the distribution of hematopoietic stem cells, multi-potent progenitors, myeloid or lymphoid progenitors. The editing frequencies were maintained in each of these HSPC compartments. Loss of CD33 and CLL-1 proteins did not impact myeloid and erythroid potential or differentiation into granulocytic or monocytic lineages in vitro. Additionally, myeloid cells derived from multiplex-edited hHSPCs also retained their function, demonstrating similar phagocytotic capacity and cytokine secretion compared to unedited control cells. Multiplex-edited hHSPCs xenotransplanted into NOD-scid IL2Rgammanull (NSG) mice showed no defect in long-term engraftment. No impact was observed in lymphoid and myeloid lineage reconstitution (10 hematopoietic lineages analyzed) while CD33 and CLL-1 antigens were simultaneously absent from >95% of cells due to genome editing (CD33+/CLL-1+ monocytes: 0.9±0.4% for multiplex-edited and 71.5±1.4% unedited arms; CD33-/CLL-1- monocytes: 92.2±2.5% for multiplex edited and 1.1±0.3% unedited arms). This data highlights the long-term persistence of high levels of biallelic editing at both genes with no impact to myeloid differentiation in vivo. Quantification of on-target editing by NGS amplicon sequencing revealed no loss of total editing frequencies after engraftment and most importantly, no detectable translocations due to our temporally staggered editing strategies to reduce simultaneous double stranded breaks at both loci. Together, these pre-clinical data indicate that gene modifications in dual engineered cells can persist long-term post-engraftment and that there were no counterselection for these cells. Lastly, CD33 and CLL-1 dual edited cells also showed significant protection from CD33 and/or CLL-1 CAR-T cells.
Summary/Conclusion: Here we show that CD33 and CLL-1 multiplex edited hHSPCs maintain robust hematopoiesis with high levels of editing at both loci. Pairing multiplex-edited hHSPCs with subsequent multi-specific immunotherapy can obviate concerns around tumor heterogeneity and escape mechanisms related to single antigen downregulation, transforming the current treatment approach for AML.