273. Genome Editing Using CRISPR-Cas9 to Increase the Therapeutic Index of Antigen-Specific Immunotherapy in Acute Myeloid Leukemia

Abstract

Antigen-specific immunotherapy using chimeric antigen receptor (CAR) T cells or bispecific T cell engaging (BITE) antibodies directed against CD19 leads to elimination of malignant and normal B cells. These approaches have a broad therapeutic index in acute lymphoid leukemia (ALL) due to a combination of high anti-tumor activity and tolerability of prolonged B cell depletion. In contrast, in acute myeloid leukemia (AML) the absence of surface antigens specific to leukemia implies that potent myeloid-directed immunotherapy will eradicate normal as well as malignant cells, leading to protracted myeloablation and bone marrow failure, as has been shown in several preclinical studies of CD33 or CD123 directed CART cell therapy. We developed a novel approach to circumvent this problem by genetically modifying normal hematopoietic stem cells (HSCs) so that they do not express the antigen recognized by CART cells, thus generating normal myeloid progeny that is resistant to CART cell therapy.CD33 is expressed on the majority of AML as well on normal myeloid progenitors and its function is poorly characterized. We hypothesized that CD33 knockout HSCs and their myeloid progeny would develop normally yet be resistant to treatment with anti-CD33 CART cells (CART33). Electroporation of human CD34+ cells with Cas9 protein and in-vitro transcribed sgRNA generated deleterious mutations in the CD33 gene with 70-80% efficacy as measured by flow cytometry and DNA sequencing. Addition of a single-stranded oligonucleotide template with a short insertional mutation at the sgRNA cut site increased the knockout efficacy up to 90%. In vitro cytotoxicity assays showed CD33 KO HSC progeny are resistant to killing by CART33 compared to wild-type HSC controls (% live cells after CART33 co-culture: 60% vs. 18%, p=0.005). CD33 KO HSC progeny showed a similar growth and differentiation profile in Methocult semi-solid media compared to control cells that were electroporated with an unrelated sgRNA. The differentiated cells had normal neutrophil and macrophage morphology and immunophenotype. We also found that phagocytosis and cytokine secretion capabilities of the CD33 KO progeny were identical to WT control. NOD-SCID-g-/- (NSG) mice injected with CD33 KO HSCs showed normal engraftment with differentiation of both myeloid and lymphoid lineages. As expected, the percentage of CD33+ myeloid cells was lower in CD33 KO recipients than WT control (24% vs 63%, p<0.0001), yet the percentage of CD14+CD11b+ monocytes (45% vs 47%, p=0.77) and CD14-CD11b+ neutrophils (34% vs. 33%, p=0.87) were unaltered.This approach yields a potential strategy to treat AML with potent CD33-specific immunotherapy, followed by infusion of gene-modified CD33 KO HSCs, which will allow persistent antigen-specific immune attack along with reconstitution of effective hematopoiesis.