Efficacy analysis of CAR-T cells against CD123-positive malignant tumors and exploration of the feasibility of combination therapy with venetoclax Authors: Meng Zhang, Xiao Chai, Xiaomei Zhang, Wenyi Lu, Hairong Lyu, Yedi Pu, Juanxia Meng, Cuicui Lyu, Mingfeng Zhao Published: 2026 ABSTRACT AbstractChimeric antigen receptor T cells (CAR-T cells) have been widely applied in hematologic malignancies. It has shown limited efficacy in the treatment of myeloid neoplasms. We retrospectively analyzed the safety and efficacy of CD123 CAR-T cells in 15 patients with myeloid neoplasms (ChiCTR2000041054). We investigated the effects of Venetoclax (Ven) on the proliferation, subset distribution, and cytotoxic activity of CD123 CAR-T cells, and explored the underlying mechanisms through multi-omics sequencing. Clinical trials indicated that the efficacy was poor. 4 patients achieved MRD-CR and 1 patient attained MRD + CR. Ven at 10nM significantly reduced the viability of HL-60, Molm13, and THP1 cells, but had no significant effect on CAR-T cells. Within a certain concentration range, as the concentration of Ven increased, the cytotoxic effect of CAR-T cells gradually intensified. In the group of 100 nM, the killing rate (p < 0.000) and cytokine secretion (p < 0.000, 0.001, and 0.028) were significantly higher than those in the group of 0nM. Further research demonstrated that pretreatment of CAR-T cells with Ven (to generate V-CD123 CAR-T cells) significantly enhanced their cytotoxic activity at a low effector-to-target ratios (1:10). Multi-omics analysis combining transcriptomic and metabolomic sequencing identified distinct alterations in key metabolic pathways between CD123 CAR-T cells and V-CD123 CAR-T cells. We propose that low-dose venetoclax combined with CD123 CAR-T cells enhances cytotoxicity, which may be attributed to metabolic optimization in CAR-T cells.Keywords: CD123 CAR-T, Efficacy, Acute myeloid leukemia, Venetoclax, Metabolism Introduction Chimeric antigen receptor (CAR) T cell therapy has achieved remarkable efficacy against hematologic tumors to date. Notably, CD19 CAR-T cell therapy has achieved an impressive complete response (CR) rate of 70% − 90% in patients with relapsed/refractory (R/R) B-cell acute lymphoblastic leukemia (B-ALL) [1]. However CAR-T cells represent a promising therapeutic strategy for acute myeloid leukemia (AML). CD123 is highly expressed on the surface of most AML stem and progenitor cells, with only low-level expression detected in normal hematopoietic stem/progenitor cells. Studies have shown that CD123 CAR-T cells can efficiently eliminate tumor cells in mouse models, highlighting CD123 as a promising target for CAR-T cell-based therapy against AML [2, 3]. To date, clinical trials of CD123 CAR-T cells have been conducted in only a small patient cohort. Specifically, seven AML patients received CD123 CAR-T cell therapy. Among them, 3 patients achieved CR, while another 3 developed PD. Importantly, no severe off-target toxicities were observed [4]. Nevertheless, subsequent successful clinical cases remain scarce. The suboptimal therapeutic efficacy of CAR-T cells arises from multiple underlying mechanisms, including tumor antigen escape, limited persistence of CAR-T cells in vivo, and the presence of an immunosuppressive tumor microenvironment (TME) [5]. To overcome these challenges and enhance clinical outcomes, recent research has increasingly focused on combinatorial strategies that pair CAR-T cell therapy with venetoclax (a targeted therapeutic agent) or other targeted therapeutics. Prominent examples include hypomethylating agents (e.g. azacitidine), Bruton’s tyrosine kinase inhibitors (e.g. ibrutinib), and immune checkpoint inhibitors (e.g. PD-1 monoclonal antibodies) [6–8]. Venetoclax (Ven), an orally bioavailable, highly selective BCL-2 inhibitor, exerts dual mechanisms to potentiate the cytotoxic activity of CD19-directed CAR-T cells. Mechanistically, pretreatment of conventional T cells with Ven significantly enhances their antitumor activity via ROS-dependent amplification of oxidative stress [9]. Concurrently, Ven-conditioned CD19 CAR-T cells exhibit enhanced functionality, which is attributed to the upregulation of caspase-3/7-mediated apoptotic signaling enhancement [10]. Notably, pre-treatment of tumor cells with Ven induces upregulation of CD19 antigen on the cell surface, thereby rendering malignant cells more susceptible to CAR-T cell-mediated lysis [11]. Despite these promising findings in CD19 CAR-T models, the modulatory effects of Ven on CD123-targeted CAR-T cells remain poorly defined. This study aimed to evaluate the efficacy and safety of CD123 CAR-T cells in eliminating CD123-positive tumor cells, and to preliminarily explore whether venetoclax can enhance their cytotoxic activity. Materials and methods Patients and methods 15 myeloid tumors patients at the Department of Hematology, Tianjin First Central Hospital were enrolled in CAR-T cell clinical trials ChiCTR2000041054. The main inclusion criteria were as follows: (1) patients who were diagnosed with R/R myeloid tumors; (2) CD123 was expressed in tumor cells; (3) without active bacterial, fungal, or viral infection; (4) the ability to understand and the willingness to provide written informed consent. The main exclusion criteria were a history of malignancy and severe renal, hepatic, and cardiac dysfunction. All patients received a lymphodepleting chemotherapy with cyclophosphamide (500mg/m2) and fludarabine (30mg/m2) 3 days before CAR-T cell infusion. On Day 14 after CAR-T cells therapy, efficacy was assessed via bone marrow morphology and flow cytometry. During the first 3 months, monthly efficacy evaluations are required, along with peripheral blood and biochemical assessments at least twice weekly. Then evaluations are conducted every 3 months. All efficacy assessments adhere to the NCCN guidelines. CRS is graded in accordance with the criteria established by Lee et al. Construction of CAR The CAR plasmid is constructed by sequentially linking an anti-CD123 single-chain variable fragment (scFV), the hinge region derived from CD8, the transmembrane domain of CD28, the 4-1BB co-stimulatory signaling domain, and the intracellular signaling domain of CD3ζ. Subsequently, the CAR sequence is cloned into the pCDH-MND-MCS-T2A-Puro backbone vector, with the incorporation of an RQR8 epitope tag that concurrently functions as a suicide switch. Preparation of CAR-T cells After obtaining peripheral blood specimens, mononuclear cells are extracted using lymphocyte separation medium. Subsequently, enrich CD3 + T cells using the immunomagnetic bead method. The obtained cells are then cultured in a specialized T-cell culture medium containing IL-2 and glutamine, and CD3/CD28 magnetic beads are added to activate the T cells. After 3 days of culture, the T cells are infected with lentivirus to prepare CD123 CAR-T cells. Cell proliferation The detection was carried out using a Cell Proliferation and Cytotoxicity Assay Kit. CAR-T cells and tumor cell lines in their logarithmic growth phase were selected. The cell density was adjusted to 5 × 10⁴ cells per well, with 90 µL added to each well of a 96-well plate. Three replicate wells were set up for each group. Ven was prepared at various concentrations, and 10 µL of each concentration was added to the 96-well plate. The cells were then incubated in an incubator for 24, 48, and 72 h. The microplate reader was calibrated, and absorbance was measured at a wavelength of 450 nm. The cell proliferation viability score was determined using the formula: (experimental group / control group) × 100%. Specific killing of CAR-T cells THP1 tumor cells labeled with green fluorescent protein (GFP) were co-incubated with effector CD123 CAR-T cells at varying effector-to-target (E: T) ratios for 24, 48, and 72 h. The control group consisted of untransfected T cells. Subsequently, an appropriate volume of the cell mixture from the aforementioned culture systems was collected. The residual number of tumor cells was determined by flow cytometry, and the killing efficiency of CAR-T cells was calculated using the formula: CAR-T cell killing rate = (number of tumor cells in the control well - number of tumor cells in the experimental well) / number of tumor cells in the control well. Additionally, the concentrations of IL-6, TNF-α, and IFN-γ in the supernatant were measured using the Luminex assay. Statistical analysis Experimental results were analyzed using GraphPad Prism software. Given the limited sample size in clinical case-related data, most analyses were of a descriptive nature. For normally distributed measurement data, results were presented as mean ± standard deviation (X±s) and analyzed using the t-test. For experiments conducted with three independent replicates, data were expressed as mean ± standard deviation. p< 0.05 was considered statistically significant. Results Patients We retrospectively analyzed 15 patients, which included 3 cases of blastic plasmacytoid dendritic cell neoplasm (BPDCN) and 12 cases of AML(ChiCTR2000041054). Baseline characteristics are detailed in Table 1, with 9 males and 6 females (median age:31 years; range:10–66 years). Patients received a median of 7 prior lines of therapy (range:2–12). Among them, four patients had extramedullary disease, two presented with complex karyotypes, two harbored FLT3-ITD mutations, one was AML1-ETO-positive, and one carried the BCR-ABL fusion gene. The median pre-infusion bone marrow blast percentage, serving as a surrogate marker of tumor burden, was 20% (range, 0.07%–59.97%). Table 1Characteristics of patientsIDAge/SexDiagnosisFusion gene/ Gene mutationKaryotypePrior lines of treatmentExtramedullary invasionPre CAR-T disease burden (%)CAR-T cells /kgEfficacy(D14)CRS140/MBPDCNN46,XY[7]12central nervous system1.422.6 × 106MRD-CR2210/MBPDCN//8testicle0.079.3 × 106MRD-CR1317/FBPDCNN46,XX,21p+9N0.336.8 × 106MRD-CR1424/FAML//2N2.542.2 × 107MRD-CR2529/FAMLWT-1, FLT3-ITD/7N27.914.2 × 106PR3627/FAMLAML1-ETO, WT-1, RUNX1-RUNXIT146,XX, t(8;21)(q22;q22)[20]6N202.8 × 106MRD + CR2718/MAMLWT-16N307 × 105NR1831/MAMLNPM1, FLT3/6liver/8.8 × 106NR1963/MAMLTP53complex karyotype7N39.291.3 × 107NR11066/MAML//10/59.971.4 × 107NR21139/MAMLWT146, XY[3]7axillary lymph nodes/7.8 × 106NR11236/FAMLTLS/ERG, NRAScomplex karyotype5N41.379 × 107NR21342/FAMLBCR-ABLp210, JAK2, ASXL1, SF3B146,XX, t(9;22) (q34,q11)3N8.035 × 106NR11414/MAMLFLT3-ITD/7N575.7 × 106NR11540/MAML//4N12.446.6 × 106NR2BPDCN blastic plasmacytoid dendritic cell neoplasm, CAR-T chimeric antigen receptor T-cell, CRS cytokine release syndrome Characteristics of patients BPDCN blastic plasmacytoid dendritic cell neoplasm, CAR-T chimeric antigen receptor T-cell, CRS cytokine release syndrome Response and toxicity The median cell dose of CD123 CAR-T cells was 6.8 × 10⁶/kg (range 7 × 10⁵/kg − 9 × 10⁷/kg). Four patients achieved MRD-CR and 1 patient attained MRD + CR. Encouragingly, one patient achieved partial remission (PR), with the percentage of blasts decreasing from 27.91% to 5.95%. Unfortunately, the remaining 9 patients achieved no remission (Fig. 1A). Only 1 patient developed grade 3 cytokine release syndrome (CRS), which was effectively managed with corticosteroids. Importantly, none of the patients exhibited signs of Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS). Responding group had notably higher CAR-T expansion (18% vs. 3%, p<0.000) and IL-6 secretion than non-responding group (1000pg/ml vs. 246 pg/ml, p = 0.003) (Fig. 1B and C). Clinical outcome following CD123 CAR-T cells. (A) The therapeutic efficacy of CD123 CAR-T cell therapy for myeloid neoplasms. (B) The peak expansion level of CAR-T cells in the responding group is significantly higher than that in the non-responding group (p < 0.000). (C) The secretion of IL-6 in the responding group is significantly higher than that in the non-responding group (p = 0.003) The impact of Ven on the proliferation of AML cells and CAR-T cells We exposed AML tumor cell lines (THP1, Molm13, HL-60) and CD123 CAR-T cells to different Ven concentrations. Compared with controls, Ven at 10nM significantly reduced the viability of HL-60, Molm13, and THP1 cells, but had no significant effect on CAR-T cell viability (Figs. 2A and B, 4). CD123 CAR-T cells were mixed with THP1 cells at a 1:1 ratio to form a co-culture system, to which different concentrations of Ven were added. The killing rate showed a progressive increase with rising Ven concentration (Fig. 2C). In the group of 100 nM, the killing rate (p < 0.000) and cytokine secretion (p < 0.000, 0.001, and 0.028) were significantly higher than those in the group of 0nM (Fig. 2D). The effect of Ven on cell proliferation activity. (A) The changes in cell proliferation activity at different time points after treating THP-1, Molm-13, and HL-60 cells with Ven at different concentrations. (B) The changes in cell proliferation activity at different time points after treating CD123 CAR-T cells with Ven at different concentrations. (C) The cytotoxic effect of different concentrations of Ven in combination with CD123 CAR-T cells on THP-1 cells. (D) The secretion of IL-6, TNF-α, and IFN-γ during the cytotoxic process of THP-1 cells induced by different concentrations of Ven and CD123 CAR-T cells (A) Molecular structure diagram of anti-CD123 CAR; (B) Transfection efficiency of CD123 CAR-T cells; (C) CD123 surface expression in different AML cell lines Cytotoxicity of Ven pre-treatment of CAR-T cells on tumor cells CD123 CAR-T cells were pre-treated with 100nM Ven for 24 h to generate V-CD123 CAR-T cells, followed by an assessment of their cytotoxic activity. When the effector-to-target (E: T) ratio was 1:1 or 1:2, there was no statistically significant difference in the killing rates (p > 0.05) (Fig. 3A, B). In contrast, when the E: T was 1:10, V-CD123 CAR-T cells showed strong cytotoxicity at 48 and 72 h (p < 0.000, p < 0.000) (Fig. 3D). They also secreted more TNF-α and IFN-γ (p = 0.035, p < 0.000) (Fig. 3E). The subsets of CAR-T cells are closely related to their killing function and durability. After CD123 CAR-T cells were treated with different concentrations of Ven, there were no marked changes in the proportions of Tn (naive T cells), Tcm (central memory T cells), Teff (effector T cells), and Tem (effector memory T cells) (p > 0.05) (Fig. 3F). The results indicate that Ven may enhance the function of CAR-T cells through direct activation. (A-D) The differences in cytotoxic activity exhibited by CD123 CAR-T cells and V-CD123 CAR-T cells at different effector-to-target ratios. (E) The secretion of IL-6, TNF-α, and IFN-γ during THP-1 killing by CD123 CAR-T cells or V-CD123 CAR-T cells. (F) The impact of Ven on CD123 CAR-T cells subsets Transcriptomic profiling revealed significant differences (p < 0.05) in core metabolic pathways between CD123 CAR-T and V-CD123 CAR-T cells, including amino acid biosynthesis, one-carbon metabolism, and oxidative phosphorylation (Figs. 4 and 5). Metabolomic profiling further demonstrated an upregulation of multiple amino acids (e.g., pentoproline, proline, and proline derivatives) alongside a concurrent downregulation of glycerophospholipid metabolites (e.g., phosphatidylethanolamine and acylcholine) (Fig. 6). These findings collectively indicate that Ven pretreatment activates diverse intracellular metabolic pathways in CAR-T cells, which may enhance their therapeutic efficacy. Bar graph and bubble chart of KEGG enrichment analysis. (A) Bar graph of KEGG enrichment analysis for CD123 CAR-T cells and V-CD123 CAR-T cells. The x-axis shows the enriched KEGG pathways, and the y-axis indicates the degree of significant differences. (B) Bubble chart for KEGG enrichment analysis of CD123 CAR-T cells and V-CD123 CAR-T cells. The X-axis shows the proportion of differentially expressed genes on a KEGG pathway out of all differentially expressed genes, and the Y-axis lists the KEGG pathways Matchstick diagram and clustering heatmap. We took the Fold Change values of the differential metabolites, did a base-2 logarithmic conversion on them, and then ranked the metabolites. We picked the top 20 up-regulated and top 20 down-regulated ones to show in a matchstick plot. (A) Substances with significant fold changes in metabolites in negative ion mode. (B) Heatmap of clustering analysis for metabolites in negative ion mode. (C) Substances with significant fold changes in metabolites in positive ion mode. (D) Heatmap of clustering analysis for metabolites in positive ion mode Discussion Our clinical trial involving 15 patients with R/R myeloid neoplasms demonstrated that CD123 CAR-T cell therapy exhibits a manageable safety profile, while its therapeutic efficacy remains limited [4]. As a distinct class of “living therapeutics,” the functionality of CAR-T cells is modulated by multiple factors. Multiple strategies can be employed to enhance the therapeutic efficacy of CAR-T cells, including: (1) structural optimization of the CAR molecule, (2) development of dual-antigen targeting CAR constructs, (3) combinatorial regimens with chemotherapy or targeted agents, and (4) tumor microenvironment remodeling [12–15]. We observed that low-dose Ven inhibits tumor cells proliferation without compromising CAR-T cells expansion. This phenomenon may be attributed to the acquisition of resistance to BCL-2/MCL-1 inhibitors by T cells during CAR-T cell manufacturing, which involves stimulation with IL-2 and CD3/CD28 magnetic beads [16]. These findings establish the preclinical rationale for combining Ven with CD123-targeted CAR-T cells in the treatment of AML. Combination treatment with graded venetoclax doses and CD123 CAR-T cells demonstrated potential synergistic effects at optimal concentrations. Notably, venetoclax exerted no marked influence on the subpopulation composition of CAR-T cells [10]. Despite the typical in vivo predominance of tumor cells over CAR-T cells, our data demonstrated a significantly enhanced cytotoxic activity of V-CD123 CAR-T cells at low E: T ratios. These findings suggest that low-dose Ven can potentiate CAR-T cell functionality. Metabolomic profiling showed that venetoclax pretreatment increased amino acid metabolites including proline and methionine in CD123 CAR-T cells [17]. PRODH2-mediated proline metabolic reprogramming has been reported to boost CAR-T cell function [18]. Furthermore, Ven also increased nucleotide metabolites in CAR-T cells. There are relatively few reports on the influence of regulating nucleotide metabolism on CAR-T cell function. Overexpression of adenosine deaminase (ADA) can enhance the function of CAR-T cells [19]. In conclusion, our findings indicate that CD123 CAR-T cell therapy exhibits suboptimal efficacy in the treatment of myeloid neoplasms, whereas the combination of Ven with CD123 CAR-T cells represents a promising potential therapeutic strategy. Multi-omics analyses suggest that Ven may enhance the cytotoxic function of CD123 CAR-T cells by improving their metabolic profiles, though the specific underlying mechanisms warrant further validation. Acknowledgements We are grateful to the patients and their families who participated in this study. We also thank all the physicians, nurses and other patient caregivers who participated in the study. Abbreviations Chimeric antigen receptor T cell Complete response Relapsed/Refractory B-cell acute lymphoblastic leukemia Acute myeloid leukemia Immunosuppressive tumor microenvironment Venetoclax Blastic plasmacytoid dendritic cell neoplasm Patient achieved partial remission Cytokine release syndrome Immune Effector Cell-Associated Neurotoxicity Syndrome Naive T cells Central memory T cells effector T cells Effector memory T cells Author contributions MZ was a major contributor in writing the manuscript. XC have made substantial contributions to the conception. MZ, XMZ have drafted the work. HRL and WYL have substantively revised it. JXM,CCL,XMZ contributes to experimental design. YDP have collected data. MFZ and HRL reviewed the draft. All authors reviewed the manuscript. Funding This work is supported by Beijing Natural Science Foundation (7254351). Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate The study was approved by the Ethics Committee of Tianjin First Central Hospital. The need for informed consent was waived due to the retrospective nature of the study and because the data were anonymously analyzed. Written informed consent was obtained from the individual for the publication of any potentially identifiable images or data included in this article. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Meng Zhang and Xiao Chai contributed equally to this work. Contributor Information Meng Zhang, Email: zhangys218@sina.com. Mingfeng Zhao, Email: mingfengzhao@sina.com. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.