Enhancing CD33.CAR NK cell therapy for acute myeloid leukemia through TGFβ imprinting and NKTR-255

Abstract Background: Acute myeloid leukemia (AML) accounts for 25% of acute leukemias in children and 33% in adults. CD33 is overexpressed in AML blasts, and natural killer (NK) cells engineered with anti-CD33 chimeric antigen receptor (CD33.CAR NK) using a Cas9/AAV6 system exhibit significantly enhanced cytotoxicity against AML (Naeimi Kararoudi, M./Lee D., et al. Cell Rep Method, 2022). Separately, chronic TGFβ stimulation during NK expansion (TGFβ imprinting or TGFβi) induces NK cells to secrete high levels of IFNγ and TNF and confer resistance to TGFβ-induced suppression in tumor microenvironment (Foltz J./Lee D. et.al. Cancers, 2018). The safety and feasibility of infusing ex-vivo expanded universal donor TGFβi NK cells have been demonstrated in a preliminary report from a Phase I/II studies in children with relapsed/refractory neuroblastoma (Ranalli M/Lee D, et al. ASCO. 2025). Interleukin-15 (IL-15) is a key homeostatic cytokine for NK cell survival, activation, and proliferation. NKTR-255, a polymer-conjugated IL-15 receptor agonist, has been shown promises in enhancing NK-based immunotherapies (Luo W/Cairo MS, et al. Mol Ther Oncol, 2024) The goal of this study is to determine whether TGFβi would further enhance anti-AML efficacy of CD33.CAR NK cells and whether NKTR-255 further augments these effects. Methods: Unmodified NK, NK TGFβi, CD33.CAR NK, and CD33.CAR NK TGFβi were evaluated with and without NKTR-255 (40ng/ml, generously provided by Nektar) in long-term in vitro cytotoxicity assays against GFP-luciferase labeled Molm13 AML cell line. NK cells were co-cultured at 2:1 effector-to-target ratio, and fresh tumor cells were added every 24 hours. Cytotoxicity and IFNγ secretion were assessed 24 hours after each cycle as previously described (Chu, Y./Cairo, MS et al. Mol Ther Oncol, 2024). Immunophenotypes of NK were evaluated by flow cytometry. In vivo anti-leukemic activity and persistence were assessed in xenograft models using immunodeficient mice engrafted with Molm13 or MV4-11 AML cells and treated with NK cell products ± NKTR-255. Results: CD33.CAR NK cells exhibited superior cytotoxicity and IFNγ secretion compared to unmodified NK throughout the assay (p0.01) but their potency declined with repeated tumor challenges. For instance, the cytotoxicity of CD33.CAR NK cells decreased from 97.9 ± 1.8% after cycle 1 to 59.4 ± 4.2% after cycle 5 (p0.01). TGFβ imprinting significantly preserved both cytotoxicity and IFNγ secretion in CD33.CAR NK cells across repeated tumor challenges (84.9 ± 3.5% tumor killing at cycle 5). NKTR-255 significantly enhanced cytotoxicity and IFNγ levels in all NK groups. The combination of TGFβ imprinting, CD33.CAR expression and NKTR-255 yielded the most sustained anti-leukemic activity: cytotoxicity of CD33.CAR NK TGFβi + NKTR-255 remained 98.7 ± 0.3% after 5 cycles of tumor cell challenges. Notably, NKTR-255 also rescued the declining cytotoxicity of NK TGFβi alone, achieving a cytotoxicity of 83.8 ± 0.9% at cycle 4, versus 58.2 ± 7.2% for NK + NKTR-255, and 40.7 ± 4.7% for NK TGFβi (p0.01), suggesting a synergistic effect between TGFβ imprinting and NKTR-255. Flow cytometry analysis revealed significantly improved NK persistence in TGFβi and NKTR-255 modified conditions. More than 80% of viable CD56+ CD33.CAR NK TGFβi cells in the presence of NKTR-255 was detected by the end of cycle 5, while there was less than 5% of viable NK cells under unmodified conditions (p0.001). TGFβi and NKTR-255 each independently and additively enhanced phosphorylation of STAT5 and expression of CD25 in NK cells (p0.05). NKTR-255 alone also significantly upregulated TRAIL expression (44.52 ± 0.27% without NKTR-255 vs 79.29 ± 5.02% with NKTR-255; p 0.001). Consistent with in vitro studies, the combination of TGFβ imprinted CD33.CAR NK cells and NKTR-255 achieved the greatest survival benefits in experimental mice xenografted with either Molm13 or MV4-11 AML cells. The median life span of MV4-11 xenografted mice was improved from 25 days in the vehicle group to 34 days with CD33.CAR NK cell treatment (p 0.05) and 41 days post CD33.CAR NK TGFβi (p 0.01) and further to 60 days post CD33.CAR NK TGFβi + NKTR255 (p 0.01). Conclusions: TGFβ imprinting and NKTR-255 independently and additively enhanced and sustained anti-AML activity of NK and CD33.CAR NK cells, supporting a novel strategy to improve NK-based immunotherapy for AML.