Chimeric antigen receptor (CAR) T-cells targeting CD19 has become a promising treatment option for relapsed/refractory B cell acute lymphoblastic leukemia and diffuse large B Cell lymphoma. For acute myeloid leukemia (AML), CAR T-cells targeting molecules such as CD33 and CD123 are under clinical evaluation. Regardless of target diseases or antigens, it is essential to understand mechanisms underlying both on- and off-target effects of CAR T-cells such as cytokine release syndrome.

In the study, we aimed to develop a CAR T-cell treatment for poor prognosis AML. To this end, we analyzed gene expression of patient-derived AML-initiating cells with demonstrated capacity for in vivo AML development in a NOD/SCID/Il2rgKO (NSG) xenogeneic transplantation assay. CD25 (IL-2 receptor alpha chain), previously reported as a marker for poor prognosis in AML, was over-represented in AML-initiating cells as compared with normal CD34+CD38- hematopoietic stem/progenitor cells (HSPCs) (Saito et al., Science Translational Medicine 2010). In addition, the antigen is expressed in other hematologic malignancies such as CML, adult T cell leukemia/lymphoma, and Hodgkin's lymphoma.

We therefore engineered lentiviral vector containing TCR sequence and Fab antigen recognition sites for human CD25 antigen. Following transduction of CD25-CAR lentiviral particles into cord blood-derived human T cells, we achieved in vitro CD25-CAR T-cell expansion to more than 2x107 cells.

In vivo treatment of human AML patient-derived xenotransplantation (PDX) mice with 5x106 CD25-CAR (25CAR) T-cells (patient n=3, PDX n=3 for each patient) resulted in reduction of patient-derived leukemic cells in the peripheral blood (PB) of PDX mice, but abundant leukemic cells remained in the bone marrow (BM). To improve homing and targeting of AML cells in the BM, we engineered CAR construct with mouse CXCR4 expression (CXCR4-25CAR). Injection of 5x106 CXCR4-25CAR T-cells resulted in complete elimination of human AML cells in PB of PDX mice (0.0+/-0.0 hCD33+ cells/ml, n=3), while patient-derived AML cells remained in peripheral blood of PDX treated with 5x106 non-CXCR4-expressing 25CAR T-cells (1416.2+/-661.0 hCD33+ cells/ml PB, n=3) (Untreated PDX: 15677 hCD33+ cells/ml PB, n=1). At 4 weeks post-CAR T-cell injection, we found complete eradication of hCD33+ AML cells only in mice treated with CXCR4-25CAR T (CXCR4-25CAR T-cell treated: BM 0.0+/-0.0 cells, spleen 0.0+/-0.0 cells spleen, n=3; 25CAR T-cell treated: BM 1.9+/-0.4x107 cells, spleen 6.2+/-2.8x107 cells, n=3). In addition, histopathological examination demonstrated no xenogeneic GVHD in liver, lung, and intestine of the CXCR4 25 CAR-treated mice. In one CXCR4-25CAR T-cell-treated PDX mouse with longer-term observation, the number of CXCR4-25 CAR T-cells decreased and murine CD45+ hematopoietic cells increased in PB without evidence of AML relapse after 3 months (2 weeks post-injection: hCD33+AML 1040.3 +/-159.7 cells/ml, hCD3+T cells 474.7 +/-85.8 cells/ml, mouse CD45+ cells 751 +/-31.0 cells/ml, n=3; 3 months post-injection: hCD33+AML 0.0+/-0.0 cells/ml, hCD3+T cells 136.3 +/-77.6 cells/ml, mCD45+ cells 1864 +/-428 cells/ml, n=3). These findings indicate that elimination of human AML cells is mediated by specific targeting of CD25 by CAR T-cells.

Injection of CXCR4-expressing CD25-CAR T-cells in NSG mice engrafted with normal CB CD34+CD38- HPSCs resulted in no change in the percentage of human CD3+Foxp3+ cells among PB CD4+ T cells (Pre-injection 10.5+/-1.4%, n=2; 4 weeks post-injection: 14.8+/-0.9% n=2). Additionally, CXCR4-25CAR did not affect to cell numbers of CD3+CD4+FoxP3+ cells, total CD3+ cells, CD19+ cells, CD33+ cells, and CD56+ cells in PB, BM, and spleen.

CXCR4-expressing CD25-CAR T-cells is a promising treatment strategy for poor prognosis AML.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.