Safety and efficacy of CD33-targeted CAR-NK cell therapy for relapsed/refractory AML: preclinical evaluation and phase I trial

Abstract Background Due to the lack of effective treatment options, the prognosis of patients with relapsed/refractory acute myeloid leukemia (R/R AML) remains poor. Although chimeric antigen receptor (CAR)-T-cell therapy has shown promising effects in acute lymphoblastic leukemia (ALL) and lymphoma, its application in R/R AML is limited by “off-target” effects, which lead to severe bone marrow suppression and limit its clinical application. CAR-natural killer (NK) cells not only exhibit antitumor effects but also demonstrate increased safety and universality. We have developed a new CAR construct that targets CD33 and modified NK cells, specifically eliminating AML cells while reducing severe side effects on stem cells. Methods The CD33-targeting domain was selected by CAR-T cells, and this optimized CAR construct was subsequently transduced into umbilical cord-derived NK cells via a retroviral vector. Preclinical efficacy and safety studies were conducted both in vitro and in vivo. Ten eligible patients with R/R AML aged 18–65 years who received one or more infusions of anti-CD33 CAR-NK cells following the preconditioning regimen were enrolled. We assessed the response rates and treatment-related side effects post-infusion, while also documenting the long-term efficacy of the therapy. Results The CD33 sequence was selected on the basis of its antitumor efficacy and safety in CAR-T-cell studies conducted both in vitro and in vivo. CD33 CAR-NK cells demonstrated efficacy comparable to that of CD33 CAR-T cells but showed limited toxicity to hematopoietic stem cells (HSCs). Ten patients, with a median of five prior lines of treatment, completed the efficacy evaluation (range, 3–8). No grade 3–4 adverse events were observed, except bone marrow suppression, which was relieved within one month. No cases of immune effector cell–associated neurotoxicity syndrome (ICANS) or graft-versus-host disease (GVHD) were reported following CAR-NK cell infusion. Only one patient experienced grade 2 cytokine release syndrome (CRS) and presented with persistent fever. By day 28, six of ten patients had achieved minimal residual disease (MRD)-negative complete remission. Conclusions Our preclinical and clinical data demonstrated the primary efficacy and safety of CD33 CAR-NK cells for patients with R/R AML. Expanded samples and longer follow-up periods are needed to provide further efficacy data. Trial registration NCT05008575 ( https://clinicaltrials.gov/study/NCT05008575 ).