Therapies for acute myeloid leukemia (AML) have not improved patient long-term survival for decades. Novel treatment options, such as chimeric antigen receptor (CAR) T cells, are needed for patients with this disease. However, AML cells lack ideal targeting antigens that are safe to target with CAR T cells. CD33, a commonly targeted antigen, is expressed in about 85-90% of AML cases but is also present on normal myeloid progenitors and myelocytes. Our aim was to engineer and validate CD33-directed CAR T cells, with the intention to open a phase I clinical trial in patients with relapsed AML. Given the potential toxicity associated with targeting CD33 in patients, an elimination gene was included in the construct design to allow CAR T cell clearance after disease eradication.

We constructed a CAR specific for CD33 by utilizing the variable heavy and light chains of the humanized M-195 antibody (HuM-195), CD28 and zeta signaling domains, and the IL-12 gene. To facilitate CAR T cell elimination if necessary, a truncated epidermal growth factor receptor (EGFRt) gene was also inserted into the retroviral vector to create the EGFRt/HuM195-28z/IL-12 CAR construct. Retroviral transduction with this tri-cistronic construct resulted in high transduction efficiency of human T cells, as determined by detection of EGFRt with fluorescently-labeled cetuximab or CAR with fluorescently-labeled CD33 molecule by flow cytometry. T cells transduced with the EGFRt/HuM195-28z/IL-12 CAR secreted functional IL-12, as assessed by culturing CAR cell supernatant with peripheral blood mononuclear cells and detecting IFN-g produced in response to IL-12. When stimulated through the CAR by co-culturing with the CD33+ AML cell line Molm-13, EGFRt/HuM195-28z/IL-12 CAR T cells proliferated and produced significant levels of the pro-inflammatory cytokines IFN-g and IL-2. In addition, EGFRt/HuM195-28z/IL-12 CAR T cells mediated significantly cytotoxicity against Molm-13 at a range of effector-to-target ratios in standard 51Cr-release assays, as compared to control CAR T cells.

The in vivo anti-tumor efficacy of EGFRt/HuM195-28z/IL-12 CAR T cells was tested in two preclinical mouse models of AML. First, SCID/Beige mice were xenografted with Molm-13 cells. Mice subsequently treated with EGFRt/HuM195-28z/IL-12 CAR T cells exhibited significant long-term survival, as compared to untreated or control CAR T cell-treated mice (p = <0.0001). In addition, NSG mice were engrafted with patient-derived CD33+ AML (PDX) cells. PDX mice treated with EGFRt/HuM195-28z/IL-12 CAR T cells exhibited reduction of peripheral CD33+ disease, as compared to untreated or control CAR T cell-treated mice. These studies demonstrate the capacity of the EGFRt/HuM195-28z/IL-12 CAR to redirect the specific anti-tumor function of T cells to CD33+ AML tumor cells. Ongoing studies aim to validate elimination of CAR T cells via EGFRt in vitro in antibody-dependent cellular cytotoxicity and complement-dependent cytotoxicity assays, as well as in mice are currently underway. These data support the utilization of CD33-specific CAR T cells in the clinic for patients with relapsed AML as a means to decrease disease burden prior to consolidative therapies such as allogeneic transplantation.


Brentjens:Juno Therapeutics: Consultancy, Membership on an entity's Board of Directors or advisory committees, Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.