Anti-Human CD117 CAR T-Cells Efficiently Eliminate Hematopoietic Stem and CD117-Positive AML Cells

Introduction: Acute Myeloid Leukemia (AML) is a clonal disease of the hematopoietic system that originates from immature hematopoietic stem and progenitor cells (HSPC). Because some AML-initiating cells are comparatively resistant to conventional cytotoxic agents, disease relapses are common with current treatment approaches. As an alternative, immunological eradication of leukemic cells by adoptively transferred chimeric-antigen receptor T-cells (CAR T-cells) might be considerably more efficient. To date, however, the search for AML-specific surface antigens has remained largely elusive. To circumvent this problem, we propose to target the stem cell antigen c-Kit (CD117) that is expressed by physiological HSPC as wells as by leukemic blasts in >90% of AML patients. For translation into a clinical setting, CAR T cell treatment must then be followed by depletion of CAR T-cells as well subsequent healthy/allogeneic HSC transplantation.

Methods: A lentiviral vector was generated which incorporates the CAR (scFv linked to intracellular CD3ζ and 4-1BB signaling domains via stalk and transmembrane regions derived from CD8), followed by a T2A ribosomal skip sequence and RQR8 as selection marker and depletion gene (surface expression of CD34 and CD20 epitopes). The scFv was extracted from a previously published bivalent anti-CD117 antibody (clone 79D) that was derived from an artificial human phage library (Reshetnyak et al., PNAS, 2013). 79D exhibits high binding affinity to an epitope in the membrane-proximal domain of human CD117. Human CD117 was cloned in human CD117 negative HL-60 AML cells and cell lines with stable expression of CD117 at various levels were derived from these.

Results: T-cells were isolated from healthy donors or AML patients in complete remission and both healthy donor and AML pateint derived T-cells exhibited sustained growth after activation with recombinant human IL-2 and CD3/CD28 beads. Lentiviral transduction yielded consistently high transduction rates, ranging from 55 - 75% as determined by staining for RQR8 and the scFv.

In co-culture assays, CAR T-cells eliminated more than 90% of CD117^high leukemia cell lines within 24 hours at effector-to target ratios (E:T) of 4:1 and 1:1 and more than 50% at E:T of 1:4. CAR-mediated cytotoxicity correlated with levels of CD117 surface expression as the elimination of CD117^low target cells was less efficient compared to CD117^high and CD117^intermediate cells. In long-term cytotoxicity assays (45d), only CD117^low cells were able to escape CAR-mediated killing. In the setting of primary cells, anti-CD117 CAR T-cells effectively depleted >90% of lin^-CD117⁺CD34⁺CD38⁺ and >70% of lin^-CD117⁺CD34⁺CD38^- cells from healthy bone marrow in vitro within 48 hours. Similarly, >70% of patient derived leukemic blasts were eliminated by autologous anti-CD117 CAR T-cells within 48 hours (1:1 ratio of CAR T cells:blasts). In a long-term assay, no outgrowth of leukemic blasts was observed in the presence of autologous CAR T-cells over 3 weeks.

To determine effectivity of CAR T-cells in vivo, humanized mice (NSG & MTRG-SKI) were engrafted with umbilical cord blood derived CD34⁺ cells. A single injection of 2x10⁶ anti-CD117 CAR T-cells resulted in >90% depletion of CD117⁺ cells in the bone marrow within 6 days. Finally, humanized mice transplanted with bone marrow from AML patients expressing CD117 were treated with patient-derived autologous CAR T-cells. At 6 weeks after injection of CAR T-cells, >98% of hu-CD45 CD117⁺ cells were depleted in the bone marrow while control human T-cell treated mice showed full-blown CD117 positive AML.

Conclusions: We provide proof of concept for the generation of highly-potent CAR T-cells re-directed against CD117 from healthy human donors and AML patients. Anti-CD117 CAR T-cells exhibit high cytotoxic activity against CD117⁺ cell lines as well as primary healthy HSPC and patient AML cells in vitro and in vivo in murine xenograft models. Strategies for the complete elimination of CAR T-cells (immunologic or small molecule based) are required before translation of this approach to the clinical setting.