C-JUN overexpressing CAR-T cells in acute myeloid leukemia: preclinical characterization and phase I trial

Abstract

Chimeric antigen receptor (CAR) T cells show suboptimal efficacy in acute myeloid leukemia (AML). We find that CAR T cells exposed to myeloid leukemia show impaired activation and cytolytic function, accompanied by impaired antigen receptor downstream calcium, ZAP70, ERK, and C-JUN signaling, compared to those exposed to B-cell leukemia. These defects are caused in part by the high expression of CD155 by AML. Overexpressing C-JUN, but not other antigen receptor downstream components, maximally restores anti-tumor function. C-JUN overexpression increases costimulatory molecules and cytokines through reinvigoration of ERK or transcriptional activation, independent of anti-exhaustion. We conduct an open-label, non-randomized, single-arm, phase I trial of C-JUN-overexpressing CAR-T in AML (NCT04835519) with safety and efficacy as primary and secondary endpoints, respectively. Of the four patients treated, one has grade 4 (dose-limiting toxicity) and three have grade 1-2 cytokine release syndrome. Two patients have no detectable bone marrow blasts and one patient has blast reduction after treatment. Thus, overexpressing C-JUN endows CAR-T efficacy in AML.

Fig. 1. CAR T cells are less effective at killing myeloid leukemia than B-lineage leukemia.

A Representative histograms showing CD33 expression. B Cytolytic activity of CD33 CAR T cells in vitro. n = 3 biological replicates with T cells from different donors per point. C Cytolytic activity of CLL1 and CD123 CAR T cells in vitro, n = 3. D Percentage of CD25 and CD69, IFN-γ, Granzyme B (GZMB), and IL-2 in CAR T cells, n = 3. E Cytolytic activity of CD33 CAR T cells against different cell lines in vitro, n = 3. F Representative histograms showing CD38 expression on sorted primary samples and CD38 CAR T cytolytic activity in vitro, n = 4 in ALL group, n = 5 in AML group. G Schematic of mouse model. NSG mice were intravenously injected with tumor cells, followed by 1 × 10⁶ CAR or PCDH T cells 5 days later. Peripheral blood (PB), spleen (SP), and bone marrow (BM) cells were collected from mice euthanized on days 3–5. H CAR T-cell counts in SP, BM, and PB from (G), n = 4 in SP and BM, n = 5 in PB. I Percentage of Annexin V⁺ of CAR T cells in spleen from (G), n = 3. J Percentage of CD25, GZMB, IFN-γ, and IL-2 of CAR T cells in spleen from (G), n = 3 in IL-2 expression, n = 4 otherwise. K Total Naïve (N, CD45RA⁺CD62L⁺), central memory (CM, CD45RA⁻CD62L⁺), effector memory (EM, CD45RA⁻CD62L⁻) and effector (E, CD45RA⁺CD62L⁻) CAR T cells were assigned to CD4⁺ and CD8⁺ CAR T cells in spleen, n = 3. L Percentage of PD-1, TIM-3, and TOX in CAR T cells in spleen, n = 5 in TOX, n = 3 otherwise. For all bar plots, data are shown as mean ± SD. Assays were performed on day 10 after T-cell initial activation. Two-sided unpaired t-tests or multiple two-sided unpaired t tests were used to assess significance in (B, C, D, F, H–L). One-way ANOVA was used in (E). All numbers defined by “n” indicate the number of biological replicates with different human donors or mice. Data are representative of two independent experiments. NS not significant. Source data are provided in the Source Data file.

Fig. 2. CAR T had an impaired effector program and defective antigen receptor signaling in AML.

A, B GSEA results from running RNAseq data of U937^CD33 co-incubated- versus Nalm6^CD33 co-incubated-CD33 CAR T cells (A). Nominal P values, FDR q values, and normalized enrichment score (NES) were calculated using GSEA software (Broad Institute). Heat maps (B) indicating the expression of genes enriched in GSEA from (A) and the known related genes not included in the GSEA gene set. The genes shown in heatmaps meet the parameters: fold change ≥ 1.5-fold in each of the two biological replicates. Each RNA sample was pooled from three technical replicates with T cells from one donor, and we conducted experiment with two different donors, n = 2. C Intensity of intracellular calcium in CAR T cells co-incubated with tumor cells, n = 3. Percentage of phosphorylated ZAP70 (D), phosphorylated ERK1/2 (E), phosphorylated C-JUN (F), and phosphorylated JNK (G) in CD33 CAR T cells pre-incubated with tumor cells and re-stimulated with U937 cells, n = 3. For all bar plots, data are shown as mean ± SD. Assays were performed on day 10 after T-cell initial activation. Two-sided unpaired t-tests were used to assess significance in (C–G). All numbers defined by “n” indicate the number of biological replicates with different human donors. Data are representative of two independent experiments. NS not significant. Source data are provided in the Source Data file.

Fig. 3. CD155 mediates myeloid tumor evasion to CAR-T killing.

A Heat maps showing inhibitory genes that are up- or down-regulated more than two-fold. Primary samples each had four biological replicates. B Histograms showing CD155 expression of in tumor cell lines and primary samples, and the MFI of CD155 in primary samples, n = 3. C Cytolytic activity of CD33 CAR T cells against U937 cells in the presence of 500 ng/ml CD155 blocking antibody, n = 3. D Flow cytometric plots showing knockout efficiency of CD155 on U937 cells. E Cytolytic activity of CD33 CAR T cells against U937^{CD155 WT} and U937^{CD155 KO} cells, n = 3. F Percentage of p-ERK1/2, p-C-JUN, and p-ZAP70 in CAR T cells pre-incubated with U937^{CD155 WT} and U937^{CD155 KO} cells and re-stimulated with U937 cells, n = 3. G Expression of CD155 in control and CD155-overexpressing (CD155 OE) Nalm6 cells, and CD19 CAR T cell cytolytic activity against them, n = 3. H Histograms showing CD112 expression in AML and ALL cell lines. I Cytolytic activity of CD33 CAR T cells against U937 cells with 10 µg/ml CD112 antibody, n = 3. J Cytolytic activity of CD33 CAR T cells against U937 cells with 10 µg/ml CD226, 50 µg/ml TIGIT, or 20 µg/ml CD96 antibody, n = 4 in CD226 antibody group, n = 5 otherwise. K Cytolytic activity of scramble-, CD96 knockout (CD96 KO), and TIGIT knockout (TIGIT KO)-CD33 CAR T cells against U937 cells, n = 3. The scrambled CD33 CAR T were electroporated with non-targeting sgRNAs. L Cytolytic activity of CD33 CAR T cells against U937 cells with both 50 µg/ml TIGIT and 20 µg/ml CD96 antibody, n = 3. For all bar plots, data are shown as mean ± SD. Assays were performed on day 10 after T-cell initial activation. Two-sided unpaired t-tests were used to assess significance in (B, C, E–G, I, J, L). One-way ANOVA was used in (K). All numbers defined by “n” indicate the number of biological replicates with different human donors. Data are representative of two independent experiments. NS not significant. Source data are provided in the Source Data file.

Fig. 4. Overexpression of C-JUN restored anti-AML activity of CAR T cells.

A Schematic of CAR constructs (gemtuzumab ozogamicin, GO). B Schematic of mouse model. NSG mice received U937 cells intravenously followed by CAR T cells or control PCDH T cells 5 days later. C Representative bioluminescence imaging of mouse model after CAR T treatment, n = 2 in IL15-CAR T group, n = 3 in other groups. Survival curve (D) and quantification of tumor burden (E) as indicated by average radiance (p/sec/cm²/sr) of (C), n = 2 in IL15-CAR T group, n = 3 in other groups. F CAR T-cell counts in PB, data are summarized from two independent experiments, n = 3 in PCDH and IL15-CAR T group, n = 4 in LCK-CAR T and ZAP70-CAR T group, n = 5 in Control CAR T and C-JUN-CAR T group. G, J, M Representative bioluminescence imaging of mouse model after CAR T treatment, n = 3. H, K, N Quantification of (G, J, M) showing the tumor burden as indicated by average radiance (p/sec/cm²/sr), n = 3. (I, L, O) Survival curve of NSG mice in (G, J, M), n = 3. P Cytokine levels in tail blood collected on days 7, 14, and 21 post-infusion, n = 6. Q Pathological analysis of the heart, liver, spleen, lung, and kidney of the representative two mice in the control and C-JUN CAR T-treatment group at their terminal stage by using HE staining. Magnification, 200×. R Cytolytic activity of control and C-JUN CAR T cells against U937 cells in vitro, CAR T cells were sorted from the mice spleen, n = 3. For all bar plots, data are shown as mean ± SD. Assays were performed on day 10 after T-cell initial activation. Two-sided unpaired t-tests or multiple two-sided unpaired t tests were used to assess significance in (P, R). Survival curves were compared using the log-rank Mantel-Cox test in (D, I, L, O). All numbers defined by “n” indicate the number of biological replicates with different mice. Data are representative of two independent experiments. NS not significant. Source data are provided in the Source Data file.

Fig. 5. The effect of C-JUN overexpression in the phenotype of CAR T cells.

A CAR T cell counts in SP, BM, and PB collected from day 10 to 20 post-infusion, n = 6. B Percentage of EDU⁺ cells in CD8⁺ CAR T cells, n = 3. C Percentage of CD25⁺ cells in CAR T cells, n = 5. D Percentage of IL-2 and IFN-γ in CAR T cells, n = 7 in CD4⁺ CAR T cells in the control CAR T group, n = 5 otherwise. E Percentage of naïve (N, CD45RA⁺CD62L⁺), central memory (CM, CD45RA⁻CD62L⁺), effector memory (EM, CD45RA⁻CD62L⁻), and effector (E, CD45RA⁺CD62L⁻) cells in CAR T cells, n = 5. F Percentage of PD-1, TIM-3, LAG-3, TOX, and TIGIT in CAR T cells, n = 3 in TOX, n = 6 in TIGIT, n = 5 otherwise. G Cytolytic activity of control and C-JUN CAR T cells against U937 cells in vitro, CAR T cells were cocultured with U937 cells at the indicated E:T ratios, n = 3 replicates per point, representative of three donors. H Percentage of IL-2 and IFN-γ in CAR T cells cocultured with U937 cells in vitro, n = 3. I Percentage of Annexin V⁺ in control and C-JUN CAR T cells cultured alone, n = 3. J Percentage of PD-1, TIM-3, and LAG-3 in control and C-JUN CAR T cells cultured alone, n = 3. K Expression of phosphorylated CD3ζ in control CAR-T, C-JUN CAR-T, and PCDH (non-CAR transduction) T cells cultured alone, n = 3. For all bar plots, data are shown as mean ± SD. Assays were performed on day 10 after T-cell initial activation. Two-sided unpaired t-tests or multiple two-sided unpaired t tests were used to assess significance in (A–J). One-way ANOVA was used in (K). All numbers defined by “n” indicate the number of biological replicates with different human donors or mice. Data are representative of two independent experiments. NS not significant. Source data are provided in the Source Data file.

Fig. 6. C-JUN re-activates ERK and upregulates co-stimulatory molecules and cytokines.

A, B GSEA results from RNAseq of U937 co-incubated- control versus C-JUN CAR T (A). Nominal P values, FDR q values, and NES were calculated using GSEA software (Broad Institute). Heat maps (B) indicating the expression of genes enriched in GSEA from (A) and the known related genes not in the GSEA gene set. The genes in heatmaps meet: fold change ≥ 1.5-fold in each of the two biological replicates. Each RNA sample was pooled from three technical replicates with T cells from one donor, and we conducted experiment with two different donors, n = 2. C Differentially accessible regions of indicated genes from ATAC-seq analysis. D Intensity of intracellular calcium in CAR T cells, n = 3. Expression of p-ZAP70 (E), p-ERK1/2 (F), p-C-JUN (G), and p-JNK (H) in CAR T cells pre-incubated with U937 and re-stimulated with U937, n = 3. I Percentage of 4-1BB, CD28, CD86, and IL-21 in CAR T cells cocultured with U937, n = 3. J Cytolytic activity of C-JUN CAR T cells against U937 with 50 µM U0126 or 50 µM SP600125, n = 3. Percentage of 4-1BB, CD28, CD86 (K), and IL-2, IL-21, IFN-γ (L) in T cells from (J), n = 3. (M) Percentage of p-ERK1/2 in U937 pre-incubated-control CAR T, C-JUN CAR T, and C-JUN CAR T with 1, 5, 10 µM U0126, n = 3. Percentage of 4-1BB, CD28, and CD86 (N) and IL-2, IL-21, and IFN-γ (O) in U937-co-incubated control CAR T, C-JUN CAR T, and C-JUN CAR T with 5 µM U0126, n = 4 in CD28 and CD86 expression, n = 3 otherwise. For all bar plots, data are shown as mean ± SD. Assays were performed on day 10 after T-cell initial activation. Two-sided unpaired t-tests or multiple two-sided unpaired t tests were used to assess significance in (D–I). One-way ANOVA was used in (J–O). All numbers defined by “n” indicate the number of biological replicates with different human donors. Data are representative of two independent experiments. NS not significant. Source data are provided in the Source Data file.

Fig. 7. Safety of C-JUN-overexpressing CD33 CAR T cells in patients with r/r AML.

A CONSORT diagram of the clinical trial. B Swimmer plot (n = 4) demonstrating the occurrence of CRS and ICANS after the infusion of CAR T cells. Each bar represents an individual patient. The severity is indicated by different colors. C CRS and ICANS management with etanercept in patient 2, n = 1. D The peak numbers of CAR T cells in PB of patients according to the grade of CRS, GVHD, and ICANS. Each dot represents one patient, n = 4. E Kinetics of serum cytokines for all patients in the first 30 days after CAR T-cell infusion. Each line represents one patient, n = 4. F Kinetics of monocyte, neutrophil, and thrombocyte counts in PB at different time points post-infusion. Each line represents one patient, n = 4. CRS cytokine release syndrome, ICANS immune effector cell-associated neurotoxicity syndrome, GVHD Graft-versus-Host Disease. Source data are provided in the Source Data file.

Fig. 8. Activity of C-JUN-overexpressing CD33 CAR T cells in patients with r/r AML.

A Swimmer plot (n = 4) showing patient responses. Each bar represents an individual patient. Responses were determined on day 15 and day 30 and were indicated by different colors. Bars with solid arrows represent patients in an ongoing follow-up. B Dot plots indicating blasts in the BM samples before CAR T-cell infusion (patient 1 at 21 days, patient 2 at 72 days, patient 3 at 12 days, patient 4 at 12 days) and on day 30 post-infusion (except patient 1 who was analyzed on day 15 post the second CAR T-cell infusion), as determined by flow cytometry. Kinetics of CAR vector transgene copies (C), CAR T cell counts (D), percentage of CAR T cells on lymphocytes (E), and percentage of CD33⁺ cells in PB (F) of individual patients at different time points post-infusion, as determined by quantitative PCR and flow cytometry. Each line represents one patient, n = 4. CR complete remission, CRi complete remission with incomplete hematologic recovery, DLT dose-limiting toxicity, MRD minimal residual disease, NR no response, PR partial remission, Pt patient, SCT stem cell transplantation. Source data are provided in the Source Data file.