Allogeneic TCRαβ deficient CAR T-cells targeting CD123 in acute myeloid leukemia

Abstract

Acute myeloid leukemia (AML) is a disease with high incidence of relapse that is originated and maintained from leukemia stem cells (LSCs). Hematopoietic stem cells can be distinguished from LSCs by an array of cell surface antigens such as CD123, thus a candidate to eliminate LSCs using a variety of approaches, including CAR T cells. Here, we evaluate the potential of allogeneic gene-edited CAR T cells targeting CD123 to eliminate LSCs (UCART123). UCART123 cells are TCRαβneg T cells generated from healthy donors using TALEN® gene-editing technology, decreasing the likelihood of graft vs host disease. As safety feature, cells express RQR8 to allow elimination with Rituximab. UCART123 effectively eliminates AML cells in vitro and in vivo with significant benefits in overall survival of AML-patient derived xenograft mice. Furthermore, UCART123 preferentially target AML over normal cells with modest toxicity to normal hematopoietic stem/progenitor cells. Together these results suggest that UCART123 represents an off-the shelf therapeutic approach for AML.

Fig. 1. Evaluation of cytotoxicity by UCART123 against MOLM13 (AML cell line) expressing CD123 in vitro and in vivo.

a CAR construct containing a scFv derived from an anti-CD123 murine antibody linked to a CD8-derived hinge/transmembrane domain, as well as 4-1BB and CD3zeta signaling domains. The CAR is co-expressed with RQR8, an artificial cell surface protein containing two CD20 epitopes recognized by Rituximab. b CD123-negative cell line, Jurkat, and CD123-positive AML cell line, MOLM13, were co-cultured with UCART123 with indicated various E:T ratios for 24 h. UCART123 induced significant cell death in MOLM13 at all E:T ratios, but not in Jurkat cells. Each symbol represents independent experiments (n = 23) and bar represents the average with the SD. ****p < 0.0001 c IFN-γ levels in supernatants were measured after 24 h when UCART123 was co-cultured with MOLM13 and Jurkat cells. Each symbol represents independent experiments (n = 19) and bar represents the average with the SD. ****p < 0.0001. d NSG mice were injected with MOLM13 cells expressing luciferase (MOLM13) and were treated with UCART123 at different cell doses on day 7 post injection of MOLM13. Leukemia burden was monitored weekly with bioluminescence imaging (BLI) and survival was evaluated. Diagram created with biorender.com e Weekly tracking of leukemia burden by BLI. UCART123-treated mice efficiently eliminated leukemia (except one mouse in the cohort treated with the lowest dose of UCART123), n = 6. f UCART123-treated mice achieved significant longer survival compared to control (p = 0.0027 for each of the UCART123-treated cohorts compared to control, log-rank test). Each symbol represents one mouse. g MOLM13 engrafted mice were treated with PBS, UCART123, or UCART123 followed by rituximab (RTX). Representative BLI images at pre-treatment, day 28 and day 42 measured are shown for each group. h top, Frequencies of UCART123 (%) in mononuclear cells in peripheral blood on day 23 post UCART123 treatment were measured with flow cytometry. Each symbol represents one mouse and bar represents the average with the SD. ****p < 0.0001, one-way ANOVA. bottom, Average radiance measured with BLI on day 28 and day 42 are shown, Each symbol represents a mouse in the cohort (UCART123/vehicle n = 10; UCART123/RTX n = 8) and the bar represents the average with the SD. ***p < 0.001, Mann–Whitney test. Source data are provided as a Source Data file.

Fig. 2. UCART123 inhibited CD123 + primary AML remarkably and selectively with minimum toxicity to normal hematopoietic cells in vitro.

Primary AML cells and cord blood (CB) cells were cultured alone, or co-cultured with TCRαβ KO T cells or UCART123 cells at different effector to target ratios (E:T) in vitro. Cell viabilities were assessed by flow cytometry at all E:T ratios and cytokines released in supernatants were measured after 24 h co-culture at 1:1 E:T ratio. a Percent of dead cells relative to untreated control at 24 h after co-culture with TCRαβ KO (white columns) or UCART123 (gray columns) under indicated E:T ratios are shown here. Each symbol represents the mean of replicates of a primary AML sample, n = 5 (5:1 ratio), n = 10 (1:1 ratio), n = 11 (0.5:1 ratio). Samples tested independently more than one time are represented in colors, AML37(red), AML17(blue), and AML95(black). Bar represents the average with the SD. ** p = 0.0013, ****p < 0.0001 two-tailed t-test. b Cytokine levels in supernatants after 24 h co-culture at all E:T ratios were measured. The average concentrations (pg/mL) of each cytokine are presented in the heatmap, ranging 0–30,000. c Percent of dead cells of cord blood (CB, n = 6) cells relative to untreated control at 24 h after co-culture with UCART123 under indicated E:T ratios. Each symbol indicates one well of replicates. Bar presents the mean with the SD. d Percent of dead cells of CB or primary AML at 24 h when mixed cells of CB (n = 3) and a single primary AML (AML37) were co-cultured with UCART123 at 5:1, 1:1, and 0.5:1 E:T ratios to assess selectivity of UCART123. UCART123 induced significant less cell death in CB cells compared to cell death in AML37 at all E:T ratios. Each symbol indicates the mean of one CB or AML in replicates. ***p = 0.0001, ****p < 0.0001. e Percent colony-forming units (CFU) relative to control in colony-forming assays of primary AML (n = 9) cells plated after 4 h co-culturing with TCRαβ KO or UCART123 at 1:1 E:T ratio. Each symbol represents one well in replicates. Bar represents the mean with the SD. ***p = 0.0002 paired two-tailed t-test f, Percent CFU relative to control in colony-forming assays of CB (n = 10) cells plated after 4 h co-culturing with UCART123 or TCRαβ KO (see Supplementary fig. 2a) at 5:1, 1:1, and 0.5:1 E:T ratios. Myeloid (gray) and erythroid (red) colonies were evaluated separately. Each symbol represents one well in replicates and bar represents the mean with the SD. ****p < 0.0001, ***p < 0.001, **p < 0.01, one-way ANOVA. Source data are provided as a Source Data file.

Fig. 3. UCART123 targets AML cells in vivo and results in improved overall survival in patient-derived xenografts (PDX) model.

a Schematic representation of the experimental procedure for the AML- PDX models (AML2 and AML37), showing that animals were treated with PBS, Ara-C, TCRαβ KO T cells or UCART123. Leukemia and T cells in peripheral blood (PB) or bone marrow (BM) were monitored starting on day 2 and later every 2–3 weeks and overall survival was evaluated. b, c Survival curves of PDX-AML2 cohorts treated with 1 × 10⁶ UCART123 (n = 7), 2.5 × 10⁶ UCART123 (n = 15) TCRαβ KO (n = 15), Ara-C (n = 15) or PBS (control; n = 14), and of PDX-AML37 cohorts treated with 2.5 × 10⁶ UCART123 (n = 7) TCRαβ KO (n = 6), Ara-C (n = 5), or PBS (control; n = 5). Treatment with UCART123 significantly prolonged overall survival in PDX-AML2 (b) and PDX-AML37(c) compared to those in control, Ara-C, or TCRαβ KO. UCART123 1 × 10⁶ or 2.5 × 10⁶ vs Ara-C; **** p < 0.0001, UCART123 2.5 × 10⁶ vs TCRαβ KO; ***p = 0.0002, UCART123 2.5 × 10⁶ vs Ara-C; * p = 0.0171, log-rank test. d NPM1 mutant in AML cells and UCART123 transcripts in PB were monitored by ddPCR on day 2 and every 2–3 weeks until mice are dead or killed due to sickness. The ddPCR plot shows simultaneous detection of human ABL1, mouse ABL1, NPM1 mutant, and UCART123 transcripts (left column). Representative examples are shown for a mouse that relapse (top) and one mouse that remained disease free (bottom). Evaluation of leukemia blasts and UCART123 cells in PB and BM at end of study on day 221 by flow cytometry demonstrated one mouse with progression of disease due to loss of UCART123 (top right) and another mouse in remission with sustaining UCART123 (bottom right). Source data are provided as a Source Data file.

Fig. 4. UCART123 killed primary AML selectively in co-engrafted model with human normal hematopoiesis and human AML.

a Schematic representation for the competitive BM and AML model. Normal CD34⁺ bone marrow cells (nBM) (6 × 10⁶ cells, HLA-A2-) were co-injected with T-cell depleted human AML cells (2.5 × 10⁵ cells, HLA-A2⁺) into sub-lethally irradiated NSG mice. After 7 weeks, mice were treated with PBS, 1 × 10⁶ UCART123 or 1 × 10⁶ TCRαβ KO T-cells. Normal hematopoietic cells and leukemia cells in PB and BM were monitored with flow cytometry starting on day 2 and later every 1–2 weeks. b Human cells gated with mouse CD45 versus human CD45 (left) were evaluated for CD123 expression level after normal cells (HLA-A2⁻, blue rectangle) and leukemic cells (HLA-A2⁺ , red rectangle) were distinguished (right). Leukemic cells express higher CD123 levels than normal cells and were preferentially eliminated by UCART123 (right bottom). c Normal cells and leukemic cells were tracked in PB by flow cytometry starting on day 2 and at indicated timepoints. Leukemic cells outcompeted normal cells over time in PBS and TCRαβ KO cohorts; cohort size: n = 5 PBS, n = 8 TCRαβ KO, n = 10 UCART123. d Normal cells and leukemic cells were evaluated in BM on day 36 post treatment. PBS (n = 5) and TCRαβ KO (n = 8) cohorts showed progression of leukemia, while UCART123 (n = 10) treated mice did not (top). Evaluation of subsets in BM showed 2-fold decrease in CD33⁺ myeloid cells in UCART123-treated mice, while lymphoid lineages were not impacted. Each symbol represents a mouse and bar represents the mean with the SD. ***p < 0.001, *p < 0.05, one-way ANOVA. Source data are provided as a Source Data file.