Effiefficacy and safety of allogeneic CD123 CAR-NK cell therapy in Relapsed/Refractory Acute Myeloid Leukemia

Abstract Background: While CAR-T cell therapy has revolutionized the treatment of certain hematologic malignancies, including B-cell malignancies and multiple myeloma, its application in R/R AML presents distinct and formidable challenges, including target antigen heterogeneity, immunosuppressive microenvironments, and treatment-related toxicities such as CRS and neurotoxicity. NK cells, with their inherent tumor-killing mechanisms, favorable safety profile, and potential for 'off-the-shelf’ allogeneic use, offer a compelling alternative. The high and consistent expression of CD123 on AML blast cells and leukemia stem cells further positions CD123-directed CAR-NK therapy as a promising novel approach. In this study, we developed a novel CD123 CAR-NK product using NK cells instead of T cells, aiming to evaluate its safety, antitumor activity, and optimal dosing in patients with R/R AML, and to explore strategies to further enhance its therapeutic efficacy. Patients and methods: A fully human-derived CD123-targeting CAR-NK cell therapy was developed and validated preclinically. CD123 CAR-NK cells were generated via lentiviral transduction using PBMCs from a single donor and cryopreserved for clinical use. Five eligible adult patients (aged 18–75 years) with R/R AML were enrolled and treated with escalating doses of CD123 CAR-NK cells. All patients had undergone extensive prior therapies. Two patients had previously failed allogeneic HSCT. Prior to CAR-NK cell infusion, patients received preconditioning regimen consisted of fludarabine (30 mg/m²/day), cyclophosphamide (300-500 mg/m²/day), and decitabine (15 mg/m²/day) for 3-5 days, followed by multiple administrations of CD123 CAR-NK cells. Clinical outcomes, including adverse events, treatment response, in vivo CAR-NK cell expansion, and persistence, were systematically assessed before and after infusion. To investigate the robust CAR-NK cell expansion observed in patient 3, comprehensive multi-omics analyses were conducted, encompassing whole-exome sequencing, CAR integration site profiling, single-cell RNA sequencing, and KIR-HLA typing. Results: Five patients received CD123 CAR-NK cell therapy and completed treatment evaluations. The median age was 45 years (range: 39-68), with a median tumor burden of 50.18% (range: 1.12%-69.6%). The median number of prior treatment lines was 3 (range: 2-6). On Day 28, 60% of patients achieved CR or CRi, and 20% achieved PR, resulting in an ORR of 80%. With respect to safety, no cases of ICANS were observed throughout the treatment course. Three patients experienced grade 1 CRS, while one patient developed grade 3 CRS, and another patient experienced grade 4 CRS. Pharmacokinetic analysis showed that CAR-NK cells in patients reached peak concentration between days 7 and 14 post-infusion. Patient 3 exhibited particularly notable pharmacodynamics, with robust CD123 CAR-NK cell expansion that corresponded with the development of grade 4 cytokine release syndrome. In-depth analysis of patient 3 revealed that the abnormally expanded NK cells in vivo were predominantly derived from the infused product. Interestingly, the CAR-positive cell population did not demonstrate a significant in vivo expansion advantage. Instead, the CD56dimCD16high subset within the infused product exhibited remarkable expansion in patient. This expanded population displayed enhanced functional characteristics, including robust cytotoxicity, immune regulatory capacity, and improved cell adhesion and migration properties. Corresponding with this cellular expansion, post-infusion plasma analysis demonstrated significant upregulation of multiple inflammatory and immune-activating factors, including TNF-α, IL-2RA, Granzyme B, IL-15, GM-CSF, and IL-8. Further immunogenetic analysis identified a KIR-HLA mismatch between the donor-derived infused product and the recipient. Conclusions: These findings demonstrate the clinical efficacy of CD123 CAR-NK cell therapy in heavily pretreated relapsed/refractory AML patients. However, the observed instances of excessive CAR-NK cell proliferation raise significant safety concerns that require thorough investigation and careful clinical surveillance.