Background The use of tyrosine kinase inhibitors (TKIs) has dramatically modified the therapy of chronic myeloid leukemia (CML), generating durable remissions and prolonging survival in TKI-responders. However, progression to blast crisis (BC) still occurs especially in TKI-resistant patients and represents a clinical challenge. We and others have identified IL2RA/CD25 as a typical cell surface marker of BC-CML and reported that its overexpression is correlated with the progression of CML from CP-CML to BC-CML (Imeri et al, Cells 2023). Here we show the experimental development of a third-generation CAR-NK therapy strategy against the CD25 based on the scFV of the clinically approved monoclonal humanized antibody, Basiliximab. Methods: As NK cell model, we have used the NK92 cell line which has a well-established and clinically demonstrated NK cell activity. We have lentivirally transduced NK92 cells with the CAR construct containing a selectable gene (GFP). After FACS-sorting of GFP-positive cells, phenotypical characterization was performed by FACS. The expression of the CAR-CD25 at the surface of the cells was demonstrated an anti-Fab antibody and double-positive (Fab/GFP) cells were further purified. The functionality of the cells was evaluated using CD107a degranulation assay after 3h co-culture and ELISA for IFN-gamma release. We have in parallel engineered K562 cells expressing CD25 by lentiviral transduction (K562-CD25) as well as also a second target cell line (RAJI) using the same strategy. Annexin V staining of target K562 cells was used for in vitro cytotoxicity assessments. For in vivo assays, NSG mice were intraperitoneally (IP) injected with K562-CD25 cells expressing Luciferase at Day-3 (3.10 6 cells/mouse, n = 13). At Days 0, 3, and 7, mice were treated by IP injection of either irradiated CD25 CAR-NK92 cells (10 .10 6/mouse n=6) or irradiated Wild-type (WT)-NK92 cells (n=5). The clinical evolution of mice transplanted mice was followed weekly by luminescence (IVIS 200). Results: After cell sorting, we obtained more than 90% of double-positive NK92 CAR+/GFP+ cells. Lentiviral transduction did not affect the activatory or inhibitory signals of NK92 cells. No statistical differences were observed between CD25 CAR-NK92 and WT- NK92 cells for the expression of NKp30, NKp46, KirDl2-3, TIGIT, and DNAM. However, we observed a strong increase in the Granzyme B and Perforin in CD25 CAR-NK92 cells after co-culture with K562-CD25 as compared to WT NK92 cells (p<0.001). Importantly, we have found increased levels of degranulation after co-culture of target K562-CD25 with CD25-CAR-NK92 cells (40%) as compared to cells co-cultured with WT NK92 alone (20% ) suggesting strongly the occurrence of an additional specific effect due to CAR-CD25. IFN-gamma levels after co-culture of CAR CD25 NK92 cells were also found to be significantly increased ( 400 pg/ml) in as compared to co-cultures of target cells with WT-NK92 (200 pg/ml)(p<0.0001). Similarly,in vitro cytotoxicity assays showed induction of higher levels of apoptosis in target cells (K562-CD25 and Raji-CD25) when co-cultured with CD25 CAR-NK92 as compared to NK92 WT (p<0.0001). In in vivo experiments, we have analyzed K562-CD25 leukemia-bearing mice treated with CAR-NK92 cells (n=6) or WT-NK92 cells (n=5). These experiments analyzed at D+30 post-transplant showed stronger anti-leukemia effect of CAR-NK therapy by IVIS imaging with a survival rate of 84% for mice treated with CD25 CAR-NK92 versus 40% for those treated with WT-NK92. All control mice transplanted with K562-CD25 cells and left untreated died by D+20. Conclusion: We show here for the first time the potential use of an NK cell-mediated CAR therapy strategy targeting CD25 which has been shown to be upregulated in CML blast crisis. The experimental data show a significantly increased and selective in vitro and in vivo cytotoxicity of CD25 CAR-NK92 cells against CD25-expressing leukemia cells as compared to WT-NK92 cells. These results suggest that targeting CD25 by a CD25 CAR based on Basilixiamb's scFV might be an interesting tool in BC-CML and in all acute leukemias overexpressing CD25. In order to translate these findings to NK cells derived from induced pluripotent stem cells (iPSCs), we have produced iPSCs expressing CAR-CD25 constructs and experiments are underway to evaluate the therapeutic potential of iPSC-derived CD25 CAR-NK cells in CML blast crisis or AML models.