Engineering CLL-1 CAR-NK cells via mRNA–LNP for potent antitumor activity and reversal of HLA-E–mediated resistance in acute myeloid leukemia

BACKGROUND: Acute myeloid leukemia (AML) remains a highly lethal malignancy, with relapse primarily driven by resistance to chemotherapy or targeted therapies. Existing chimeric antigen receptor T cell (CAR-T) strategies are limited by toxicity and complex manufacturing, underscoring the need for transient, controllable, and safe CAR-engineering platforms that can selectively target multidrug-resistant (MDR) AML cells. METHODS: We developed a lipid nanoparticle (LNP)–based mRNA delivery platform for scalable generation of C-type lectin-like molecule-1 (CLL-1) CAR-NK cells. NK phenotype, cytotoxicity, cytokine secretion, and safety were evaluated ex vivo against AML cell lines and patient-derived blasts, with in vivo efficacy tested in xenograft NSG mouse models. Mechanisms of adaptive resistance were investigated through transcriptomic profiling, modulation of the NKG2A/HLA-E axis, and functional interrogation of the JAK2–STAT1 signaling pathway. RESULTS: Drug-response profiling across AML cohorts identified a multidrug-resistant subgroup marked by a distinct transcriptomic program in which CLL-1 was the only validated CAR target upregulated. mRNA–LNP transfection enabled efficient generation of primary CLL-1 CAR-NK cells with preserved phenotype and potent, antigen-specific cytotoxicity against AML cells, while sparing normal hematopoietic progenitors. In vivo, repeated CAR-NK infusions markedly suppressed leukemia progression and prolonged survival. Transcriptomic analyses of tumor cells surviving CAR-NK exposure revealed inflammatory activation with progressive HLA-E upregulation, which impaired CAR-NK function. NKG2A blockade restored cytotoxicity ex vivo and enhanced leukemia clearance and survival in vivo. Mechanistically, prolonged CAR-NK engagement or IFN-γ stimulation activated a JAK2–STAT1 axis that drove sustained HLA-E induction. JAK2 knockdown reduced HLA-E expression and sensitized AML cells to CAR-NK–mediated killing, whereas pharmacologic JAK2 inhibition also decreased HLA-E expression but concurrently impaired NK-cell activation, thereby limiting the overall therapeutic benefit. CONCLUSIONS: Transient, non-integrating mRNA–LNP–transfected CLL-1 CAR-NK cells provide a safe and effective strategy for MDR AML. Repeated dosing enables robust antitumor activity, while adaptive resistance via NKG2A/HLA-E axis can be mitigated through checkpoint blockade. The JAK2–STAT1 pathway represents a potential upstream modulator, providing opportunities for rational combinatorial approaches to optimize CAR-NK therapy.