Cytokine-induced Expandable Memory NK Cells with Significant Metabolic, Epigenetic Remodeling, and Persistence Properties

Cytokine-induced Expandable Memory NK Cells with Significant Metabolic, Epigenetic Remodeling, and Persistence Properties | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Cytokine-induced Expandable Memory NK Cells with Significant Metabolic, Epigenetic Remodeling, and Persistence Properties Zhiying He, Hongye Wang, Jingwei Tu, Yi Luo, Xinxin Yuan, Xin Fang, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8091256/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Recent investigations have indicated that NK cells induced overnight with IL-12/15/18 cytokines possess memory-like characteristics (refers as CIML NK, cytokine induced memory-like NK cells), demonstrating enhanced effector function and prolonged persistence. Preclinical studies and very early clinical trials revealed the promising clinical activity of CIML NK cells in the treatment of hematological malignancies, especially acute myelocytic leukemia (AML). However, the current manufacturing method for CIML NK cells do not endow these cells the ability to expand robustly ex vivo . The low production efficiency limits their clinical application. In this study, we established a novel expansion platform for these memory-like NK cells. This method can effectively and stably expand healthy donor-derived Peripheral NK (PBNK) cells by more than 200-fold within two weeks, with a purity exceeding 95%. Importantly, these ex vivo expanded NK (ExNK) cells exhibited hypomethylated state in the CNS-1 region (Conserved Noncoding Sequence 1) of IFNG gene, stronger metabolic ability, and enhanced effector function when compared to CIML NK cells. Based on these features, we named these expanded NK cells but still possess memory-like characteristics as mExNK (memory-like expanded NK cells). Bulk mRNA-seq further uncovered a high similarity between mExNK and CIML NK cells, with both highly expressing genes being related to proliferation, metabolism, and memory. Mouse tumor models proved that the infused mExNK exhibited longer persistence in vivo and stronger effector functions. Taken together, the method described herein showed an enhanced expansion efficiency for human memory NK cells and would facilitate their clinic applications in cancer therapy. Biological sciences/Immunology/Translational immunology Biological sciences/Immunology/Adaptive immunity/Cellular immunity/Immunological memory memory-like NK cells ex vivo expansion metabolism reprograming epigenetic remodeling persistence Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Natural killer (NK) cells are commonly described as a type of innate lymphocytes with natural cytotoxicity. This functional definition means that NK cells can exhibit a rapid cytolytic activity against malignant tumor cells and virally infected cells without requiring prior antigen sensitization ( 1 , 2 ). NK cells express a variety of germline-encoded, stochastically expressed activating and inhibitory receptors that control their functional responses to the target cells ( 3 ). Activated NK cells can eliminate tumor cells through multiple killing mechanisms: they can release perforin and granzymes to act directly on target cells; and they also enable to mediate antibody-dependent cellular cytotoxicity (ADCC) via the Fc gamma receptor (FcγRIIIa / CD16a) expressed on their surface. Therefore, NK cell-based adoptive cell therapy (ACT) has been investigated for decades. Early clinical studies indicated that NK cell infusion, even with those originating from allogeneic sources, has a higher safety profile ( 4 , 5 ). While this NK cell-mediated protective immunity holds great promise in cancer therapy, challenges remain, including their absence of immune memory so that NK cells frequently display a limited persistence and functionality in vivo , which would compromise their clinical effectiveness ( 6 ). In the last two decades, however, accumulating evidence have revealed that the innate immune system, similar to the adaptive immune system, is also capable of exhibiting immune memory, a phenomenon observed in cells such as macrophages, dendritic cells, and NK cells, known as "trained immunity" ( 7 – 10 ). Currently, the NK cell memory has been studied in three main scenarios: hapten specific, virus specific and cytokine induced ( 11 – 14 ). While in human, the cytokine induced memory-like NK (CIML NK) cells represent one of the two major types of memory NK cells ( 15 ). CIML NK cells are induced by a triple cytokine cocktail consisting of IL-12, IL-15 and IL-18 ( 11 , 16 ). Recent studies indicated that a brief exposure to this cocktail initiates a transient proliferation of NKG2A + CD57- NK cell subsets that enable to simultaneously express certain enzyme markers such as EZH2 (enhancer of zeste homolog 2, a histone methyltransferase) and CD39 (the ectonucleoside triphosphate diphosphohydrolase-1) ( 17 , 18 ). These proliferating NK cells immediately exhibit a shift in the glycose metabolism from oxidative phosphorylation to aerobic glycolysis and the inheritable epigenetic changes such as the DNA methylation state in the CNS-1 region (conserved noncoding sequence 1) of IFN-γ encoding gene IFNG ( 19 – 21 ). The synergistic coordination between the metabolic rewiring and the epigenetic reprogramming described above, and among the others, composed the molecular basis of NK cell memory ( 15 ). Upon re-encountering tumor targets or cytokines, CIML NK cells demonstrate markedly enhanced IFN-γ production and effector functions ( 22 , 23 ). They can persist for several months in both clinical patients and mouse models, significantly longer than conventional NK cells ( 24 – 26 ). Multiple clinical trials (NCT02782546, NCT03068819, NCT04024761, NCT04290546, NCT04354025, and NCT04634435, from clinicaltrials.gov ) have been initiated with CIML NK cells, demonstrating notable safety and efficacy in treating acute myeloid leukemia (AML) patients. However, the current preparation methods for CIML NK cells used in these trials lack effective cell expansion, limiting their broad clinical application. The growing clinical demand for large quantities of CIML NK cells urgently requires the development of more efficient and convenient methods for their ex vivo expansion. In this context, the robust expansion of CIML NK cells was proved achievable when using a combination of CD16a antibody and IL-15/IL-18 cytokines in the early activation phase to induce the development of immune memory, followed by the use of a cytokine cocktail containing IL-2, IL-15, and IL-21 in the later phase to ensure the maintenance of memory properties during the high-efficiency expansion stage. Here, we developed a novel feeder-free platform for expanding memory NK cells. By using of this method, we achieved more than 200-fold expansion of NK cells after 14 days in culture. We then systematically and rigorously validated the immune memory attributes of these expanded NK cells compared with CIML NK and PBNK cells across five dimensions: phenotype, metabolism, epigenetics, function, and persistence. Thus, we named these named these expanded but still possess memory-like characteristics NK cells as mExNK (memory-like expanded NK cells). The scalable expansion of mExNK cells would pave a new avenue for their clinical applications. Results Effective expansion of human NK cells with the feeder-free, cytokine induced system While unstimulated NK cells exhibit limited antitumor efficacy in adoptive cell therapy, multiple activation and expansion protocols have been developed to enhance their cytotoxic function ( 27 – 29 ). To prevent T cell overgrowth in final products, we first depleted CD3 + T cells from PBMCs using immunomagnetic beads. Building on evidence from cytokine-induced memory-like (CIML) NK studies, combinations of CD16a antibody with cytokines such as IL-12, IL-15, and IL-18 can promote memory-like properties ( 11 , 30 ). According to studies, NK cell activation upregulates the metalloprotease ADAM17, which cleaves surface receptors including CD16a ( 31 , 32 ). Therefore, we stimulated NK cells with CD16a antibody plus IL-15/IL-18 during the first two days to induce memory-like characteristics (Fig. 1A) . The ADAM17 inhibitor INCB3619 was added concurrently to limit receptor shedding and preserve CD16a expression. CD137 antibody and IL-21, recognized as strong proliferative signals for NK cells ( 33 – 35 ), were subsequently introduced to drive expansion (Fig. 1A) . From day 4 onward, IL-2, IL-15, and IL-21 were supplemented to sustain long-term culture. This feeder-free system consistently yielded expanded NK (ExNK) cells with high viability (> 80%) and purity (≥ 95%) (Fig. 1B-D) . Across 12 healthy donors, the mean expansion at two weeks was ~ 230-fold (range from 110 to 543), with some donors sustaining growth for up to four weeks (Fig. 1E-F) . Flow cytometry revealed that ExNK cells expressed high levels of activating receptors (CD16, NKG2D, NKp30, NKp44, NKp46) and inhibitory receptors NKG2A and CD158b, whereas CD158a and CD158e1 were detected on fewer than 20% of cells (Fig. 1G) . Real-time cytotoxicity assays (Incucyte) against K562 targets showed that at 4 hours, ExNK cells mediated ∼20% killing at a 1:1 E:T ratio, increasing to ∼80% at 10:1 (Fig. 1H-I) . In summary, we established a feeder-free expansion system that generates highly pure (≥ 95%), functionally active NK cells expressing multiple activating receptors and exhibiting potent tumor-killing capacity. Although early activation included CD16 antibody and IL-15/IL-18 to promote memory-like features, whether such properties persist after two weeks of expansion requires further investigation. ExNK cells showed apparent memory-like properties as determined by phenotype, metabolism and epigenetic remodeling Prior studies have established that CIML NK cells exhibit distinct features—including a specific surface phenotype, metabolic remodeling, and epigenetic reprogramming—which collectively contribute to their elevated cytotoxicity (Summarized in Fig. 2A) ( 11 , 15 , 36 ). Here, we systematically evaluated our expanded NK (ExNK) cells against these established CIML benchmarks. Firstly, we isolated and purified peripheral blood NK cells (PBNKs) from healthy donors and cultured them in low-dose IL-15 (1ng/ml) for 3–5 days to serve as our negative control. PBNKs from the same donor underwent IL-12/15/18 pre-activation for 16–24 hours, followed by resting in low-dose IL-15 (1 ng/ml) for 3–5 days, to serve as the positive control cells (CIML NK). Morphologically, both CIML NK and ExNK cells formed prominent clusters, whereas PBNKs did not (Fig. 2B) . IL-12/15/18 pre-activation strongly upregulates CD25 (IL-2Rα), enabling formation of high-affinity IL-2 receptors and supporting robust proliferation and cytotoxicity even under low IL-2 conditions ( 37 ). Flow cytometry revealed that ExNK and CIML NK cells similarly upregulated CD25 and the early activation marker CD69 compared to PBNKs, while expression of CD16, NKp46, and NKG2D remained largely unchanged (Fig. 2C and Sup Fig. 1A) . Metabolically, resting NK cells exhibit low activity, but activation induces a pronounced glycolytic shift, accompanied by increased expression of nutrient transporters such as GLUT1 (SLC2A1), amino acid transporters (SLC1A5, SLC7A5, SLC3A2/CD98), and CD71 (transferrin receptor) ( 38 , 39 ). Consistent with a CIML-like profile, ExNK cells showed elevated levels of these transporters, comparable to CIML NK cells (Fig. 2D-E) . Functional metabolic assays confirmed that both ExNK and CIML NK cells exhibited enhanced glycolysis, glycolytic capacity, and glycolytic reserve relative to PBNKs (Fig. 2F-G) , though oxidative phosphorylation (OXPHOS) parameters—including basal respiration, spare respiratory capacity, and ATP production—remained unchanged across all groups (Sup Fig. 1B-C) . The ECAR:OCR ratio, reflecting a shift toward glycolytic metabolism ( 40 ), was similarly reduced in both ExNK and CIML NK cells (Fig. 2H) . A hallmark of CIML NK cells is their capacity for enhanced IFN-γ production upon restimulation, which is epigenetically regulated by demethylation of CpG sites in the conserved noncoding sequence 1 (CNS‑1) of the IFNG locus ( 20 , 41 – 43 ). We assessed the methylation status of six key CpG sites within the IFNG CNS‑1 region as an indicator of memory-like reprogramming. ExNK cells exhibited profound hypomethylation at these sites, with an average methylation level of only 6%, compared to 54.8% in CIML NK and 76.7% in PBNK cells (Fig. 2I-J and Sup Fig. 1D-E) , indicating that ExNK cells undergo extensive epigenetic remodeling. In summary, ExNK cells closely resemble CIML NK cells in their immunophenotype, metabolic profile, and epigenetic state of the IFNG CNS‑1 region—distinguishing them clearly from PBNKs. Based on these findings, we designate these expanded cells as mExNK (memory-like expanded NK cells). mExNK had potent effector function and persistence ability in vitro Building on our demonstration that mExNK cells possess immune memory, we postulated that they would also exhibit enhanced effector functions, akin to CIML NK cells. To test this, we first evaluated cytokine expression following a 6-hour co-culture with K562 target cells (E:T ratio = 1:1). Flow cytometric analysis revealed that both mExNK and CIML NK cells produced significantly higher levels (measured by both percentage and MFI) of IFN-γ, TNF-α, and Granzyme B than PBNKs cultured in low-dose IL-15, with mExNK cells consistently showing the most robust enhancement (Fig. 3A, 3B, 3D) . Perforin expression was also elevated in mExNK compared to PBNK, though not significantly different from CIML NK (Fig. 3C) . These results were corroborated by ELISA after 24-hour co-culture, confirming that mExNK and CIML NK cells secreted significantly more IFN-γ, Granzyme B, and perforin than PBNKs (Fig. 3F-H) . Traditional methods for detecting NK cell effector function in vitro primarily focus on measuring cytokine secretion, and killing of target cells. However, due to the short co-culture times and high E:T ratios typically used, these methods often fail to accurately and objectively reflect the true anti-tumor capacity of immune cells in vivo ( 44 , 45 ). Tumor rechallenge assay can be coupled with profiling immune cell activation, exhaustion and persistence phenotypes, provide a more accurate reflection of the in vivo anti-tumor capacity of immune cells and are widely used, particularly in the in vitro functional evaluation of CAR-T cells ( 46 ). Given evidence that CIML NK cells display prolonged persistence ( 22 , 24 , 25 ), we developed a rechallenge model to assess serial killing capacity under high tumor burden. PBNK, CIML NK, and mExNK cells were co-cultured with three tumor lines (K562, Raji, THP-1) at E:T ratios of 1:1 or 1:2. Fresh, fluorescently labeled target cells were replenished every two days, and killing was monitored over three rounds (6 days) using Incucyte live-cell imaging. mExNK cells consistently exhibited stronger and more sustained tumor killing than both CIML NK and PBNK cells, with differences emerging as early as 6–8 hours and maintained throughout the assay (Fig. 3H-G and Sup Fig. 2A-I) . Together, these data demonstrate that mExNK cells possess not only enhanced effector function relative to CIML NK cells, but also superior persistence in serial killing assays. The mExNK cells display an advanced effector and activation gene expression, but similar memory-like profiles with CIML NK We next performed bulk mRNA-sequencing to define the transcriptional landscapes of mExNK and CIML NK cells relative to PBNK cells. The analysis yielded two central findings. First, mExNK and CIML NK cells are highly similar: the number of DEGs between each of them and PBNK was far greater than the number of DEGs between them ( Fig. 4A and Sup Fig. 3A) , and their sample correlation was markedly higher than with PBNK (Fig. 4B) . This similarity was quantified by the substantial overlap in their individual DEG sets compared to PBNK (~ 71–72% of up- and down-regulated genes; Sup Fig. 3A-C ) and confirmed by GO analysis, which showed shared enrichment in pathways like DNA replication, oxidative phosphorylation and cellular amino acid metabolic process (Sup Fig. 3F) . The second key finding was the specific signature of mExNK cells. Direct comparison with CIML NK cells revealed a unique enrichment in pathways governing T cell activation, cell-cell adhesion, and cytokine regulation (Sup Fig. 3D) . Importantly, this transcriptional prediction of enhanced adhesive and effector capacity was functionally validated by the superior performance of mExNK cells in the tumor rechallenge assay (Sup Fig. 3E) . Gene Set Enrichment Analysis (GSEA) using MSigDB Hallmark gene sets further highlighted strong similarities between mExNK and CIML NK cells (Sup Fig. 3G) . Both populations showed significant upregulation of gene sets involved in E2F targets, G2M checkpoint, mitotic spindle, IL2‑STAT5 signaling, and mTORC1 signaling compared to PBNK (Sup Fig. 3H-I) . Given the established roles of IL2‑STAT5 and mTORC1 pathways in NK cell survival, metabolism, and function ( 47 , 48 ), we focused subsequent validation on these, along with oxidative phosphorylation (OXPHOS). GSEA confirmed that mExNK and CIML NK cells exhibited significant upregulation of all three pathways relative to PBNK (Fig. 4C-E) . Notably, even in a resting state, mExNK cells maintained higher expression of key effector genes, including TNFSF10 (TRAIL), FASLG (FASL), GZMA , GZMB , and IFNG (Fig. 4F) . Both mExNK and CIML NK cells highly expressed KLRC1 (NKG2A), IL2RA (CD25), and genes encoding nutrient transporters such as TFRC (CD71), SLC1A5 , and SLC7A5 (Fig. 4G) , consistent with our phenotypic and metabolic data (Fig. 2C-E) . Transcriptomic profiling also revealed elevated expression of EZH2 and MKI67 , alongside low expression of B3GAT1 (CD57), in both mExNK and CIML NK cells (Fig. 4G) . This profile matches with a previously reported CD57-NKG2A + EZH2 + MKI67 + CIML NK subset associated with rapid cycling and enhanced trained immunity ( 18 ). Furthermore, we observed upregulation of activation-related genes (e.g., IRF4 , TNFRSF9 , MYC ) and downregulation of maturation- and adhesion-related genes (e.g., FGFBP2 , KLF2 , CX3CR1 ), mirroring recently reported memory-like NK transcriptional signatures ( 17 ) (Fig. 4H) . In summary, mExNK cells share a core memory-like transcriptional program with CIML NK cells but also display a distinct gene expression profile linked to adhesion and effector function—consistent with their enhanced and sustained anti-tumor activity. mExNK cells showed efficacy against THP‑1 leukemia tumors and persist long‑term in vivo Next, we were interested to investigate the effector function and persistence of our mExNK cells in vivo during targeting the tumor cells. We employed an NSG mouse model engrafted with THP-1-Luc human leukemic monocytic cells to evaluate the anti-tumor activity of mExNK cells ( 49 ). THP-1-Luc⁺ tumor cells were administered intravenously via the tail vein at a dose of 2×10⁵ per mouse. Three days later, mExNK cells were infused intravenously at a dose of 1×10⁷ per mouse. Tumor-bearing mice without transferred NK cells served as the control group. After NK cell transfer, each mouse received intraperitoneal injections of cytokines (50,000 U IL-2 and 1 ng IL-15 per injection) twice to three times per week for three weeks to support NK cell survival (Fig. 5A) . Tumor growth and mouse body weight were monitored weekly until control mice largely died around 7 weeks. Compared to the control group, mExNK cell treatment significantly reduced tumor burden and prolonged survival (Fig. 5B-D) , indicating potent in vivo anti-tumor capability of mExNK cells. Body weight monitoring revealed no significant difference between mExNK-treated and control mice, suggesting no apparent toxicity associated with mExNK cells (Fig. 5E) . To test our hypothesis that mExNK cells can persist longer in vivo , we collected blood from the orbital venous bleeds of mice at 1-, 2-, and 4-weeks post-infusion (under tumor-bearing conditions) and measured the proportion of hCD45⁺&hCD56⁺ NK cells. Although the number of mExNK cells decreased weekly, they were still detectable even four weeks after infusion (Fig. 5F) . To further benchmark the efficacy of mExNK cells, we included NK cells expanded using a Miltenyi Biotec commercial kit (hereafter termed cExNK, conventional expanded NK) as a control. While cExNK cells met baseline quality criteria with viability and purity both exceeding 80%, their average expansion fold within 2 weeks was substantially lower (~ 40-fold) than that achieved with our system (Sup Fig. 4A-D) . Phenotypically, cExNK cells showed minimal CD25 expression, indicating a lack of canonical memory-like feature (Sup Fig. 4E) . Functionally, they also exhibited significantly reduced expression of key nutrient transporters and secreted lower levels of the effector molecules IFN-γ, Granzyme B, and Perforin compared to mExNK cells (Sup Fig. 4F-I) . Transcriptomically, bulk mRNA-seq revealed a strikingly greater number of differentially expressed genes (DEGs) between cExNK and mExNK (4204) than between CIML NK and mExNK (1149) (Sup Fig. 4J) . GO analysis indicated that cExNK cells were significantly deficient in pathways related to gene expression, immune response, and cell surface receptor signaling (Sup Fig. 4K) . This was further evidenced by the marked downregulation of critical effector molecules (e.g., IFNG , GZMB ), metabolic transporters (e.g., SLC7A5 , SLC1A5 ), and crucial transcription factors (e.g., STAT1 , STAT3 , RUNX1 ) in cExNK cells (Sup Fig. 4L) . Collectively, these multi-faceted results demonstrate that mExNK cells possess superior properties not only relative to CIML NK cells but also, and more profoundly, to conventionally expanded NK cells. IL-21 acts in synergy with IL-2 to promote memory NK cell expansion ex vivo Finally, we sought to define the key mechanisms and culture factors that enable the generation and expansion of memory-like NK cells. Transcriptomic analysis revealed that both CIML NK and mExNK cells upregulate genes involved in mitotic cell cycle and DNA replication (Sup Fig. 3F) , indicating their intrinsic proliferative potential. However, conventional IL-12/15/18 stimulation alone does not lead to substantial expansion of CIML NK cells, highlighting a major challenge in scaling memory-like NK cells for clinical use. To identify pathways critical for expansion, we compared mRNA-seq profiles of mExNK and CIML NK cells. Over-representation analysis of genes upregulated in mExNK versus CIML NK cells revealed consistent and significant enrichment of the JAK-STAT signaling pathway across multiple databases (KEGG, GO-BP, Reactome, Wikipathways) (Fig. 6A) . This aligns with the known roles of cytokines in our culture system: IL-15 supports NK cell survival ( 50 ), IL-21 activates STAT3 and upregulates MYC to regulate proliferation and metabolism ( 51 ), and IL-2 potently drives NK cell expansion and cytotoxicity ( 52 , 53 ). Transcription factor activity analysis further indicated differential activation of STAT1 , STAT3 , STAT5A , STAT5B , NFKB1 , and STAT6 in mExNK cells (Fig. 6B) , and key JAK-STAT pathway genes (including BCL2 , CCND3 , IL21R , JAK1 , JAK3 , and STAT1/3 ) were more highly expressed in mExNK than in CIML NK cells (Fig. 6C) . To functionally validate these findings, we tested different cytokine combinations in a 10-day expansion assay. While IL-21 increased IL-21R expression non-significantly (Fig. 6D) , it clearly enhanced phosphorylation of STAT1 and STAT3 (Fig. 6E–F) . Co-stimulation with IL-2 and IL-21 further elevated p-STAT5 levels, indicating synergistic JAK-STAT activation (Fig. 6G) . Expansion folds were 17-, 52-, 45-, and 128-fold for the IL-2-/IL-21-, IL-2+/IL-21-, IL-2-/IL-21+, and IL-2+/IL-21 + groups, respectively (Fig. 6H) . Together, these results underscore that IL-2 and IL-21 are essential for activating the JAK-STAT pathway and driving NK cell expansion, with their combination exerting a potent synergistic effect. Discussion Over the past two decades, immunotherapy with functionally improved T and NK cells has emerged as a novel pillar of cancer treatment ( 6 , 54 ). Memory NK cells, especially CIML NK cells, represent an attractive cell type for developing this therapeutic strategy, mainly due to their enhanced effector function, enhanced ability to persist and proliferate in vivo ( 22 , 36 ). In this context, the robust expansion ex vivo of these cells is essential in order to meet clinical demands. Currently, few established methods exist for expanding memory-like NK cells ( 55 , 56 ). One study showed that pre-activating NK cells with cytokines (IL-12, IL-15, and IL-18) together with stimulation by PM21 particles—derived from K562-41BBL-mbIL21 feeder cells—resulted in an 8200-fold expansion of NK cells within two weeks ( 56 ). However, the absence of proper CIML NK controls and in vivo tumor-killing assays made the reported expansion of memory-like NK cells less convincing. Moreover, the therapeutic use of such tumor feeder cell-derived particles presents considerable safety concerns, in addition to challenges related to GMP-grade storage and handling. In the same year, Shrestha et al. reported a feeder-free protocol based on a human tissue factor (TF) fusion protein incorporating IL-12/15/18 cytokines (HCW9206), along with an anti-TF-TF IgG1 antibody (HCW9101) that may deliver a stimulating signal via CD16a receptor ( 55 ). This combination of signals induced approximately a 250-fold expansion of memory NK cells over two weeks. Although such cytokine fusion proteins enable efficient ex vivo expansion of memory-like NK cells, their adoption has been limited by the specialized preparation required, making it challenging for other research groups to reproduce the methodology. In the present study, we developed a novel feeder-free system for the ex vivo expansion of memory like NK cells. All necessary materials are commercially available reagents, and the culture procedure is relatively simple, which should facilitate broader translational applications in the future. Based on insights from memory-like NK cell studies, combinations of CD16a antibody with cytokines such as IL-12, IL-15, and IL-18 led to enhanced IFN-γ production upon restimulation with tumor cells or cytokines, underscoring the importance of CD16a engagement in conferring memory-like properties ( 11 , 30 ). However, activation with cytokines or exposure to target cells were found to substantially reduce CD16a expression. Previous studies have shown that NK cells express ADAM17, and selective inhibition of this metalloprotease prevented proteolytic shedding of CD16a, thereby augmenting interferon-γ production—especially when activation was mediated via CD16a ( 57 ). Thus, we selected a combination of CD16 antibody, IL-15, IL-18, and the ADAM17 inhibitor INCB3619 as the initial signaling cocktail to induce immune memory in NK cells. Activation of the CD137 signaling axis is essential for achieving optimal NK cell expansion. Previous studies indicate that the synergistic interaction among CD137L, IL-15, and IL-21 critically influences the extent of NK cell proliferation ( 58 ). The anti-CD137 agonist urelumab has been shown to sustain expression of key activating receptors such as NKG2D and effector molecules including granzyme B and IFN-γ, thereby counteracting TGFβ-mediated suppression of human NK cell proliferation and antitumor function ( 59 ). Moreover, combined stimulation with IL-2 and IL-21 co-stimulation has been identified as the most effective combination for enhancing NK cell cytotoxicity, sustaining proliferative capacity, and enhancing IFN-γ secretion ( 60 ). Further evidence confirms that IL-21 signaling activates multiple pathways—including Jak-STAT, MAPK, and PI3K—to collectively support NK cell expansion ( 61 ). Based on the aforementioned research findings, we supplemented the dT-PBMC culture with CD137 antibody (Urelumab), IL-2, and IL-21 at 48 hours after pre-activation, followed by regular additions of IL-2/15/21 starting from day 4 to optimize NK cell expansion outcomes (Fig. 1) . The efficient expansion of memory NK cells offers an excellent opportunity for us to understand their biological attributes. Our data comprehensively demonstrate that mExNK cells possess classic memory-like attributes across multiple layers. The mExNK cells exhibited a definitive memory phenotype, characterized by high CD25 expression similar to CIML NK cells (Fig. 2C, 4G) , much higher than PBNK and cExNK (Fig. 2C and Sup Fig. 4E, 4L) . They also underwent significant metabolic remodeling, evidenced by increased nutrient transporter expression, enhanced glycolysis, and transcriptional upregulation of OXPHOS and mTORC1 signaling pathways, which collectively support their enhanced effector functions (Fig. 2D-H, 4D-E; Sup Fig. 3F and 4E-F) . At the epigenetic level, mExNK cells showed much lower methylation in the IFNG CNS-1 region than CIML NK cells (Fig. 2I-J; Sup Fig. 1D-E) . Supporting this permissive epigenetic state, mExNK cells produced more IFN-γ upon stimulation (Fig. 3A, 3F) , demonstrating a direct link between their epigenetic and functional enhancements. Concurrently, mExNK cells expressed higher levels of cytotoxic molecules, including Perforin and Granzyme B (Fig. 3A-G, Fig. 4F and Sup Fig. 4G-I) , and exhibited superior in vitro tumor-killing and persistence capacity (Fig. 3H-J and Sup Fig. 2A-I) compared to CIML NK cells. More importantly, mExNK cells exhibited superior in vivo tumor-killing and persistence capacity, significantly reduced tumor burden and prolonged mice survival (Fig. 5B-F) . Foltz et al. originally identified CD117⁻CD57⁻NKG2A⁺CD39⁺ enriched memory-like (eML) NK cells as the subpopulation possessing genuine immune memory ( 17 ). Our mRNA-seq analysis revealed a high degree of similarity between the transcriptomes of mExNK cells and these eML NK cells, including concordant changes in key genes (Fig. 4H) . Specifically, compared to CIML NK cells, mExNK cells expressed lower levels of KIT (encoding CD117) and B3GAT1 (encoding CD57), and higher levels of ENTPD1 (encoding CD39) (Fig. 4H, Sup Fig. 1A) , suggesting a greater enrichment of the bona fide eML subset. Integrating these phenotypic, metabolic, epigenetic, transcriptomic, and functional analyses, we conclude that mExNK cells likely possess more robust immune memory characteristics than CIML NK cells. Therefore, despite undergoing expansion, mExNK cells retain strong immune memory properties. The expansion of mExNK cells significantly enhances the production efficiency and clinical therapeutic potential of traditional CIML NK cells, while simultaneously improving their functionality, offering greater clinical applicability. However, how to further improve the expansion efficiency of memory-like NK cells remains a challenge to be overcome. Materials and methods Mice Housing and all experimental protocols for mice used in this study were performed in accordance with the guidelines established by the Institutional Animal Care and Use Committee in the East Hospital of Tongji University. NCG mice (NOD/ShiLtJGpt-Prkdc em26Cd52 Il2rg em26Cd22 /Gpt) were purchased from The GemPharmatech. If not stated differently, 6-8-week-old male mice were used for all experiments. Cell Culture Cell lines THP-1-Luciferase cell line was bought from Xiamen Immocell Biotechnology Co.,Ltd., K562 and Raji tumor cell lines were obtained from Cell Bank of the Committee on Type Culture Collection, Chinese Academy of Sciences‌. They were all cultured in RPMI 1640 complete media (Gibco) supplemented with 10% FBS (Gibco). The information of cell lines are listed in Table S3 . PBNK cells Freshly isolated human peripheral blood NK (PBNK) cells were purified from PBMCs using the NK Cell Isolation Kit (Miltenyi) according to the manufacturer's instructions. Cell purity, defined as the percentage of CD56⁺CD3⁻ cells, was determined by flow cytometry. Only preparations with a purity exceeding 90% were used for subsequent cultivation or pre-activation. Then, PBNK cells were rested in HIPP-T009 medium (Bioengine) supplemented with 5% UltraGRO™-Advanced Serum Replacement (Helios) and low-dose rhIL-15 (1 ng/ml) for about 3–5 days to generate negative control PBNK cells. CIML NK cells Freshly purified PBNK were pre-activated with rhIL-12 (10ng/mL), rhIL-15 (100ng/mL), and rhIL-18 (50ng/mL) for 16–24 hours to generate cytokine induced memory-like NK cells (CIML NK) as positive control NK cells, then harvested, washed, and allowed to differentiate for about 3 days supported with low-dose rhIL-15 (1 ng/ml) as described ( 11 , 22 , 62 ). PBNK and CIML NK cells were used as negative and positive controls in the FACS test, ELISA assay, metabolism assessment, conventional bisulfite sequencing, cytokine production test and tumor rechallenge assay. cExNK cells Freshly purified PBNK cells were activated and expanded with NK cell Activation/Expansion Kit (Miltenyi) by using the NKp46/CD2 coupling microbeads as described per manufacturer’s instructions. The expansion medium was HIPP-T009 medium (Bioengine) supplemented with 5% UltraGRO™-Advanced Serum Replacement (Helios), 500U/mL rhIL-2 (PeproTech). After 14 days, this conventional expanded NK cells (cExNK) were used in the FACS tests and ELISA assay. Isolation and expansion of memory-like peripheral blood NK cells Healthy anonymous human PBMC were obtained by Ficoll (Stem Cell Technology) centrifugation of cells from leukoreduction filters following platelet apheresis. The CD3-depleted PBMCs (with a residual T cell percentage ≤ 1%) were isolated using Dynabeads™ CD3 (Thermo) as described per manufacturer’s instructions. Then dT-PBMCs were placed in anti-CD16a antibody (T&L) immobilization plate containing HIPP-T009 medium (Bioengine) supplemented with 5% UltraGRO™-Advanced Serum Replacement (Helios), 500U/mL rhIL-2 (PeproTech), 50ng/mL rhIL-15 (T&L), 50ng/mL rhIL-18 (T&L) and 4µM ADAM inhibitor (MCE) at 37℃ in a humidified atmosphere containing 5% CO2. After 2 days, fresh medium with 500U/mL rhIL-2 (PeproTech), 50ng/mL rhIL-15 (T&L) and 1µg/mL Urebumab (MCE) were added. From day 4, medium containing 500IU/mL rhIL-2 (PeproTech), 50ng/mL rhIL-15 (T&L) and 50ng/mL rhIL-21 (T&L) was replenished as necessary during the NK expansion process. Flow Cytometry Analysis For analysis of surface markers, cells were stained in PBS (Gibco) containing 1% FBS (Gibco). Surface proteins were stained at 4℃ for 30min. Intracellular staining of SLC1A5, SLC7A5, GLUT1 and cytokines was performed with Cytofix Fixation Buffer (BD) and Perm/Wash Buffer I (BD), according to the manufacturer’s instructions. Data were obtained using a CytoFLEX flow cytometer (Beckman Coulter) and analyzed using FlowJo Software. The following antibodies were used: APC anti-human IFN-γ antibody, PE anti-human TNF-α antibody, PE anti-human CD107a (LAMP-1) antibody, FITC anti-human Perforin, APC anti-human/mouse Granzyme B Recombinant antibody, PE anti-human CD56 (NCAM) antibody, FITC anti-human CD56 (NCAM) antibody, APC anti-human CD56 (NCAM) antibody, PE anti-human CD45 antibody, PE anti-human CD3 antibody, PE anti-human CD25 antibody, PE anti-human CD69 antibody, PE anti-human CD314 (NKG2D) antibody, APC anti-human CD159a (NKG2A) antibody, PE anti-human CD57 antibody, PE anti-human CD335 (NKp46) antibody, APC anti-human CD337 (NKp30) antibody, PE anti-human CD336 (NKp44) antibody, APC anti-human CD16 antibody, PE anti-human CD158 (KIR2DL1/S1/S3/S5) antibody, FITC anti-human CD158b/j (KIR2DL2/L3/S2) antibody, FITC anti-human CD158e1 (KIR3DL1) antibody, APC anti-human CD71 antibody, PE anti-human CD360 (IL-21R) Antibody, FITC Mouse IgG1, κ Isotype Ctrl (FC) antibody, PE Mouse IgM, κ Isotype Ctrl antibody, PE Mouse IgG1, κ Isotype Ctrl antibody, FITC Mouse IgG2b, κ Isotype Ctrl antibody, APC Mouse IgG1, κ Isotype Ctrl antibody (all from Biolegend), Alexa Fluor™ 488 Rabbit Anti-Human GLUT1 (BD), PE Mouse Anti-Human CD98 (BD), Alexa Fluor™ 488 Rabbit IgG Isotype Control (BD), PE Mouse IgG1, κ Isotype Control (BD), Anti-SLC7A5/LAT1 antibody [EPR26260-66] (Abcam), Anti-SLC1A5/ASCT2 antibody[CAL33] (Abcam), Goat Anti-Rabbit IgG H&L / FITC secondary antibody (HZbscience), Phospho-Stat1 (Tyr701) (D4A7) Rabbit mAb (CST), Phospho-Stat3 (Tyr705) (D3A7) XP® Rabbit mAb (CST), Phospho-Stat5 (Tyr694) (D47E7) XP® Rabbit mAb (CST). All antibodies and their working conditions are listed in Table S1 . Incucyte-Based NK cytotoxicity Assays K562 target cells were labeled with Incucyte® Cytolight Rapid Green Dye (Sartorius, Germany) for 20 min at 37℃ in the dark and washed twice with complete media to quench the labeling reaction. For kinetic analysis of tumor cell killing, Green dye labelled K562 cells were plated at a concentration of 1e4 per well in 96-well flat bottom plate. Expaned NK cell were added at E:T ratios ranging from 1:1 to 40:1. All co-cultures were performed in HIPP culture media. The number of viable target cells was monitored by hourly fluorescence imaging over 24 hours using an IncuCyte S3 live cell imaging and analysis system (Sartorius, Germany). Live cell numbers were quantified by IncuCyte S3 software and normalized to the number of live cells remaing in the target cell-only control group. Degranulation Assay and Cytokine Production PBNK, CIML NK and ExNK cells were co-cultured with K562 tumor cells for 6 hours at an effector/target ratio of 1:1 in HIPP growth medium with 1 ng/ml IL-15. For degranulation assays, anti-CD107a mAb was added at the start of the functional assay in 1:100 dilution ratio. To assay cytokine production, Brefeldin A and monensin (eBioscience) were added within the first hour of co-culture with K562. After 5 hours, cells were washed and stained for cell surface marker hCD56 (1:200 dilution), fixed/permeabilized and intracellularly stained for IFN-γ, TNF-α, Perforin, Granzyme B (1:100 dilution). ELISAs PBNK, CIML NK and ExNK cells were co-cultured with K562 tumor cells at an effector/target ratio of 1:1 in HIPP growth medium without exogenous cytokines. The supernatant was collected after 24h, and Granzyme B, Perforin, IFN‑γ secretion was measured using the Granzyme B Valukine ELISA Kit (R&D Systems), ELISA Flex Human Perforin (HRP) (MABTECH) and Human IFN-γ Broad range ELISA Kit (STARTER, China) individually according to the manufacturer’s instructions. Metabolism Assessment Seahosre assays were performed according to the manufacturer’s instructions with modifications to simultaneously analyze glycolysis and oxidative mitochondrial metabolism. Briefly, 1.5e5 cells were plated per well in triplicate into Poly-D-Lysine coated Seahorse XF24 plates and analyzed with a Seahorse XF24 Analyzer (Agilent Technologies). For ECAR assays, the assay medium was prepared by Seahorse XF RPMI Medium (pH 7.4) contained 2mM L-glutamine. Glucose, oligomycin and 2-deoxyglucose (2-DG) at a final concentration of 10mM, 2.5µm and 50mM individually, were serially injected. For OCR assays, the assay medium was prepared by Seahorse XF RPMI Medium contained 10mM Glucose, 1mM Pyruvate, and 2mM L-glutamine, pH 7.4. Mitochondrial respiration was monitored at basal state and after sequential injection of the 2.5µM oligomycin, 1µM FCCP, 2.5µM antimycin A and 2.5µM rotenone. Glycolysis was calculated as average post-glucose ECAR values minus average basal ECAR values. Glycolytic reserve was calculated as average maximal ECAR values minus post-glucose ECAR values. Glycolytic capacity was calculated as average post-oligomycin ECAR values minus average basal ECAR values. Basal respiration was determined by monitoring OCR in the absence of any inhibitors. SRC (Spare Respiration Capacity) measurements were calculated as average maximal OCR values minus average basal OCR values. ATP respiration was calculated as average basal OCR values minus average post-oligomycin values. Conventional Bisulfite Sequencing of IFNG-CNS1 region Genomic DNA from PBNK, CIML NK and ExNK cells was extracted with FastPure Blood/Cell/Tissue/Bacteria DNA Isolation Mini Kit (Vazyme, DC112-01) and treated with EZ DNA Methylation-Direct Kit according to the manufacturer’s instructions. Bisulfite-treated DNA was subjected to PCR amplification and cloned into pGEM-T vector (Promega). Individual clones were sequenced by standard Sanger sequencing. Data were analyzed by BISMA ( http://services.ibc.uni-stuttgart.de/BDPC/BISMA/ ) ( 63 ). Primers sequences for the bisulfate PCR and sequencing were listed in Table S4 . Repeated Tumor Rechallenge Assays Persistence of NK cells in vitro were evaluated in a serial tumor challenge assay as described before( 46 ). In brief, 2.5e4 or 1.25e4 NK cells and 2.5e4 target cells (K562, Raji or THP-1) labelled with cytolight red (Sartorius) were co-cultured in a poly-D-lysine pre-coated 96-well plate using HIPP fresh growth medium with 1ng/ml IL-15 to support NK survival. Maintained the plate in a 37℃, 5% CO2 incubator for 6 days to perform 3 repeated tumor rechallenge, additional cytolight red labelled tumor targets in fresh medium administered to the co-culture every other day. The number of viable target cells was monitored by every 2 hours fluorescence imaging over 140 hours using the IncuCyte S3 live cell imaging and analysis system (Sartorius). Live cell numbers were quantified as red area and normalized to the beginning time (0d0h0m). In vivo tumor xenograft models For in vivo experiments testing ExNK cell function again THP-1 tumor cells in a single infusion model, 6- to 8-week-old male NSG mice were injected intravenously in tail-vein with 2 or 5e6 luciferase expressing THP-1 tumor cells. There days later mice were distributed into 2–3 groups randomly, each group with at least 8 mice. Memory-like expanded NK cells (mExNK) were resuspended in 1e7/0.2mL PBS buffer and injected intravenously in tail-vein of tumor bearing mice. Transferred NK cells were supported with rhIL-2 (50000U/mice) and rhIL-15 (10 ng/mice) by intraperitoneal (i.p.) injection at least 2 times every week for up to 3 weeks. BLI was performed weekly to monitor tumor progression, at the same time monitor the mice body weigh weekly. THP-1-Luciferase tumor-bearing mice without NK cell transfer (PBS) or with conventional expanded NK cells (cExNK) served as controls. Animal technicians were blinded to expected outcomes. The experiments were performed in the East China Normal University. BLI was conducted using an IVIS Spectrum, and images were analyzed using Live Imaging Software (Perkin Elmer). Survival was graphed as Kaplan-Meier curves, and the log-rank (Mantel-Cox) test was used to determine statistical significance. Persistence of adoptively transferred human NK cells in NSG mice The experiment was approved by the Institutional Animal Care and Use Committee in the East China Normal University. After 1 week of acclimation, a single dose of 1e7 memory-like and conventional expanded NK cells (mExNK and cExNK) was injected intravenously through the tail vein into male NSG mice (6- to 8-week-old; GemPharmatech) with rhIL-2 (50000U/mice) and rhIL-15 (10 ng/mice) (i.p. injection, 2 times per week for 3 weeks) support. For blood collection, about 100µl blood obtained from orbital venous bleeds using EDTA anticoagulant at 1, 2, 4 weeks post-infusion for immediate flow cytometry. Blood sample were incubated with the appropriate antibody fluorophore conjugates (mouse CD16/32 Fc block (Biolegend, 1:100), viability Live/Dead (Thermo, 1:1000), PE-hCD45 and APC-hCD56 (Biolegend, 1:50)) for 30mins on ice in staining buffer (DPBS + 2%FBS + 2mM EDTA). Then stained blood samples were incubated with ACK buffer (Gibco) for 5mins on ice to lyse red blood cells. After lysis, samples were washed with cold staining buffer once, then resuspended in 90µl staining buffer plus 10µl CountBright absolute counting bbeads (Thermo), and data was acquired in CytoFLEX flow cytometer (Beckman Coulter). Population densities were quantified to hCD45+&Hcd56 + cells/total cells using in sample CountBright beads. RNA-Seq Sample Preparation, Sequencing, and Data Analysis Total RNA was extracted by TRIzol Reagent (Thermo) to constructed library using using VAHTS Universal V10 RNA-seq Library Prep Kit for MGI (Vazyme, NRM606). The libraries were size-selected for cDNA fragments of 200–300 bp on 2% Low Range Ultra Agarose and then subjected to PCR amplification using Phusion DNA polymerase (NEB) for 15 PCR cycles. Following quantification using TBS380, the paired-end libraries were sequenced using the DNBSEQ-T7 platform (Shanghai BIOZERON Biotech. Co., Ltd). The initial paired-end reads underwent trimming and quality control using Trimmomatic, applying the parameters (SLIDINGWINDOW:4:15 MINLEN:75). Subsequently the resulting clean reads were individually aligned to the reference genome in orientation mode using hisat2 software with default settings. The quality of these data was assessed using qualimap_v2.21. Clean reads were aligned to the ribosome database of human genome (hg38) using the alignment tool bowtie2 to evaluate the proportion of rRNA in the featurecount to count each gene reads. We analyzed differentially expressed genes (DEGs) using the R package edgeR for all pairwise comparisons among PBNK, CIML NK, and mExNK samples, and for the comparison between cExNK and mExNK. Genes with an adjusted p-value < 0.05 (using a significance threshold of p-adjust < 0.05) were considered statistically significant DEGs. Statistical analysis Figures were prepared using GraphPad Prism8. All data are presented as mean ± s.e.m, excepting informed specially. Statistical significance excepting the survival curve of mice and the ELISA result of IFN-γ was calculated using an unpaired, two-tailed student’s t-test with 95% confident intervals. ns, not significant (p > 0.05). * p < 0.05, ** p < 0.01, *** p < 0.001. The statistical significance of mice survival curve was calculated using Log-rank (Mantel-Cox) test. The statistical significance of IFN-γ ELISA result was calculated using F-test. Abbreviations ACT, doptive cell therapy; ADCC, Antibody-Dependent Cellular Cytotoxicity; AML , Acute Myelocytic Leukemia; CIFF, Cytokine-induced and Feeder-free; CIML NK, Cytokine Induced Memory-like NK cells; CNS-1, Conserved Noncoding Sequence 1; ECAR, Extracellular Acidification Rate; ELISA , Enzyme-Linked Immunosorbent Assay; cExNK , Convential Expanded NK cells; mExNK , Memory-like Expanded NK cells; EZH2, Enhancer of Zeste Homolog 2; GSEA, Gene Set Enrichment Analysis; IL , Interleukin, MDS, Myelodysplastic Syndromes; MFI , Mean Fluorescent Intensity; eML NK, Enriched Memory-Like NK cells; MM , Multiple Myeloma; NK , Natural killer; OCR , Oxygen Consumption Rate; ORA, Over Representation Analysis; PBNK , Peripheral NK cell; SEM, Standard Error of the Mean; TFs, Transcriptional Factors; VIPER, Visualization Pipeline for RNA-seq analysis; Declarations Funding This work was funded by the National Natural Science Foundation of China (82471592, 82270638, 82300718, 82301904), Shanghai Clinical Research Center for Cell Therapy (23J41900100), Shanghai Engineering Research Center of Stem Cells Translational Medicine (20DZ2255100), Peak Disciplines (Type IV) of Institutions of Higher Learning in Shanghai. The authors declared that these funders had no role in the study design, data collection, analysis, interpretation, manuscript writing, or the decision to submit the article for publication. Authors’ Contributions Q.D., Z.H. and H.W. performed study concept and design; W.Z., Q.D., Z.H. and H.W. performed development of methodology and writing, review and revision of the paper; H.W. proved acquisition, analysis and interpretation of data, and statistical analysis; H.W. performed the vast majority assays, excepting the ELISA test and development of expansion system. Q.D., J.T. and X.Y. developed the expansion system of memory-like NK cells. J.T., X.Y., F.X. and Z.L. performed the ELISA assay and some phenotype tests. H.W. and Y.L. performed the mRNA-seq and analysis. L.H. and Y.X. provided technique support of mice assay. All authors read and approved the final paper. Ethics statement The studies involving human peripheral blood and PBMC cells from healthy volunteers were approved by Shanghai Hycells Biotechnology Co., Ltd (Ethics Approval Number: HYS-LQ-001 and EC-SMP-2024017). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was required from the participants. The animal study was approved by the Institutional Animal Care and Use Committee in the East Hospital of Tongji University (Ethics Approval Number: Animal Research Review No. (273), 2024). Declaration of interests The authors declare that they have no conflict of interest. References Liu S, Galat V, Galat Y, Lee YKA, Wainwright D, Wu J. NK cell-based cancer immunotherapy: from basic biology to clinical development. Journal of hematology & oncology. 2021;14(1):7. Vivier E, Raulet DH, Moretta A, Caligiuri MA, Zitvogel L, Lanier LL, et al. Innate or adaptive immunity? The example of natural killer cells. Science (New York, NY). 2011;331(6013):44-9. Martinet L, Smyth MJ. Balancing natural killer cell activation through paired receptors. Nature reviews Immunology. 2015;15(4):243-54. Iliopoulou EG, Kountourakis P, Karamouzis MV, Doufexis D, Ardavanis A, Baxevanis CN, et al. A phase I trial of adoptive transfer of allogeneic natural killer cells in patients with advanced non-small cell lung cancer. Cancer immunology, immunotherapy : CII. 2010;59(12):1781-9. Sakamoto N, Ishikawa T, Kokura S, Okayama T, Oka K, Ideno M, et al. Phase I clinical trial of autologous NK cell therapy using novel expansion method in patients with advanced digestive cancer. Journal of translational medicine. 2015;13:277. Berrien-Elliott MM, Jacobs MT, Fehniger TA. Allogeneic natural killer cell therapy. Blood. 2023;141(8):856-68. Bowdish DM, Loffredo MS, Mukhopadhyay S, Mantovani A, Gordon S. Macrophage receptors implicated in the "adaptive" form of innate immunity. Microbes and infection. 2007;9(14-15):1680-7. Netea MG, Quintin J, van der Meer JW. Trained immunity: a memory for innate host defense. Cell host & microbe. 2011;9(5):355-61. Kleinnijenhuis J, Quintin J, Preijers F, Joosten LA, Ifrim DC, Saeed S, et al. Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of monocytes. Proceedings of the National Academy of Sciences of the United States of America. 2012;109(43):17537-42. Ochando J, Mulder WJM, Madsen JC, Netea MG, Duivenvoorden R. Trained immunity - basic concepts and contributions to immunopathology. Nature reviews Nephrology. 2023;19(1):23-37. Romee R, Schneider SE, Leong JW, Chase JM, Keppel CR, Sullivan RP, et al. Cytokine activation induces human memory-like NK cells. Blood. 2012;120(24):4751-60. Kleinnijenhuis J, Quintin J, Preijers F, Joosten LA, Jacobs C, Xavier RJ, et al. BCG-induced trained immunity in NK cells: Role for non-specific protection to infection. Clinical immunology (Orlando, Fla). 2014;155(2):213-9. Schlums H, Cichocki F, Tesi B, Theorell J, Beziat V, Holmes TD, et al. Cytomegalovirus infection drives adaptive epigenetic diversification of NK cells with altered signaling and effector function. Immunity. 2015;42(3):443-56. Peng H, Tian Z. Natural Killer Cell Memory: Progress and Implications. Frontiers in immunology. 2017;8:1143. Tarannum M, Romee R. Cytokine-induced memory-like natural killer cells for cancer immunotherapy. Stem cell research & therapy. 2021;12(1):592. Cooper MA, Elliott JM, Keyel PA, Yang L, Carrero JA, Yokoyama WM. Cytokine-induced memory-like natural killer cells. Proceedings of the National Academy of Sciences of the United States of America. 2009;106(6):1915-9. Foltz JA, Tran J, Wong P, Fan C, Schmidt E, Fisk B, et al. Cytokines drive the formation of memory-like NK cell subsets via epigenetic rewiring and transcriptional regulation. Science immunology. 2024;9(96):eadk4893. Zhang C, Yin J, Zheng J, Xiao J, Hu J, Su Y, et al. EZH2 identifies the precursors of human natural killer cells with trained immunity. Cancer biology & medicine. 2021;18(4):1021-39. Terrén I, Orrantia A, Mosteiro A, Vitallé J, Zenarruzabeitia O, Borrego F. Metabolic changes of Interleukin-12/15/18-stimulated human NK cells. Scientific reports. 2021;11(1):6472. Ni J, Hölsken O, Miller M, Hammer Q, Luetke-Eversloh M, Romagnani C, et al. Adoptively transferred natural killer cells maintain long-term antitumor activity by epigenetic imprinting and CD4(+) T cell help. Oncoimmunology. 2016;5(9):e1219009. Becker-Hapak MK, Shrestha N, McClain E, Dee MJ, Chaturvedi P, Leclerc GM, et al. A Fusion Protein Complex that Combines IL-12, IL-15, and IL-18 Signaling to Induce Memory-Like NK Cells for Cancer Immunotherapy. Cancer immunology research. 2021;9(9):1071-87. Berrien-Elliott MM, Foltz JA, Russler-Germain DA, Neal CC, Tran J, Gang M, et al. Hematopoietic cell transplantation donor-derived memory-like NK cells functionally persist after transfer into patients with leukemia. Science translational medicine. 2022;14(633):eabm1375. Keppel MP, Yang L, Cooper MA. Murine NK cell intrinsic cytokine-induced memory-like responses are maintained following homeostatic proliferation. Journal of immunology (Baltimore, Md : 1950). 2013;190(9):4754-62. Bednarski JJ, Zimmerman C, Berrien-Elliott MM, Foltz JA, Becker-Hapak M, Neal CC, et al. Donor memory-like NK cells persist and induce remissions in pediatric patients with relapsed AML after transplant. Blood. 2022;139(11):1670-83. Shapiro RM, Birch GC, Hu G, Vergara Cadavid J, Nikiforow S, Baginska J, et al. Expansion, persistence, and efficacy of donor memory-like NK cells infused for posttransplant relapse. The Journal of clinical investigation. 2022;132(11). Berrien-Elliott MM, Cashen AF, Cubitt CC, Neal CC, Wong P, Wagner JA, et al. Multidimensional Analyses of Donor Memory-Like NK Cells Reveal New Associations with Response after Adoptive Immunotherapy for Leukemia. Cancer discovery. 2020;10(12):1854-71. Deng X, Terunuma H, Nieda M, Xiao W, Nicol A. Synergistic cytotoxicity of ex vivo expanded natural killer cells in combination with monoclonal antibody drugs against cancer cells. International immunopharmacology. 2012;14(4):593-605. Huang RS, Lai MC, Shih HA, Lin S. A robust platform for expansion and genome editing of primary human natural killer cells. The Journal of experimental medicine. 2021;218(3). Li X, He C, Liu C, Ma J, Ma P, Cui H, et al. Expansion of NK cells from PBMCs using immobilized 4-1BBL and interleukin-21. International journal of oncology. 2015;47(1):335-42. Pahl JHW, Koch J, Götz JJ, Arnold A, Reusch U, Gantke T, et al. CD16A Activation of NK Cells Promotes NK Cell Proliferation and Memory-Like Cytotoxicity against Cancer Cells. Cancer immunology research. 2018;6(5):517-27. Mishra HK, Dixon KJ, Pore N, Felices M, Miller JS, Walcheck B. Activation of ADAM17 by IL-15 Limits Human NK Cell Proliferation. Frontiers in immunology. 2021;12:711621. Pomeroy EJ, Hunzeker JT, Kluesner MG, Lahr WS, Smeester BA, Crosby MR, et al. A Genetically Engineered Primary Human Natural Killer Cell Platform for Cancer Immunotherapy. Molecular therapy : the journal of the American Society of Gene Therapy. 2020;28(1):52-63. Parrish-Novak J, Dillon SR, Nelson A, Hammond A, Sprecher C, Gross JA, et al. Interleukin 21 and its receptor are involved in NK cell expansion and regulation of lymphocyte function. Nature. 2000;408(6808):57-63. Shman TV, Vashkevich KP, Migas AA, Matveyenka MA, Lasiukov YA, Mukhametshyna NS, et al. Phenotypic and functional characterisation of locally produced natural killer cells ex vivo expanded with the K562-41BBL-mbIL21 cell line. Clinical and experimental medicine. 2023;23(6):2551-60. Vidard L. 4-1BB and cytokines trigger human NK, γδ T, and CD8(+) T cell proliferation and activation, but are not required for their effector functions. Immunity, inflammation and disease. 2023;11(1):e749. Bakhtiyaridovvombaygi M, Yazdanparast S, Mikanik F, Izadpanah A, Parkhideh S, Shahbaz Ghasabeh A, et al. Cytokine-Induced Memory-Like NK Cells: Emerging strategy for AML immunotherapy. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie. 2023;168:115718. Leong JW, Chase JM, Romee R, Schneider SE, Sullivan RP, Cooper MA, et al. Preactivation with IL-12, IL-15, and IL-18 induces CD25 and a functional high-affinity IL-2 receptor on human cytokine-induced memory-like natural killer cells. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation. 2014;20(4):463-73. O'Brien KL, Finlay DK. Immunometabolism and natural killer cell responses. Nature reviews Immunology. 2019;19(5):282-90. Keating SE, Zaiatz-Bittencourt V, Loftus RM, Keane C, Brennan K, Finlay DK, et al. Metabolic Reprogramming Supports IFN-γ Production by CD56bright NK Cells. Journal of immunology (Baltimore, Md : 1950). 2016;196(6):2552-60. Nicholas D, Proctor EA, Raval FM, Ip BC, Habib C, Ritou E, et al. Advances in the quantification of mitochondrial function in primary human immune cells through extracellular flux analysis. PloS one. 2017;12(2):e0170975. Tato CM, Martins GA, High FA, DiCioccio CB, Reiner SL, Hunter CA. Cutting Edge: Innate production of IFN-gamma by NK cells is independent of epigenetic modification of the IFN-gamma promoter. Journal of immunology (Baltimore, Md : 1950). 2004;173(3):1514-7. Balasubramani A, Mukasa R, Hatton RD, Weaver CT. Regulation of the Ifng locus in the context of T-lineage specification and plasticity. Immunological reviews. 2010;238(1):216-32. Wilson CB, Rowell E, Sekimata M. Epigenetic control of T-helper-cell differentiation. Nature reviews Immunology. 2009;9(2):91-105. Long AH, Haso WM, Shern JF, Wanhainen KM, Murgai M, Ingaramo M, et al. 4-1BB costimulation ameliorates T cell exhaustion induced by tonic signaling of chimeric antigen receptors. Nature medicine. 2015;21(6):581-90. Gattinoni L, Klebanoff CA, Palmer DC, Wrzesinski C, Kerstann K, Yu Z, et al. Acquisition of full effector function in vitro paradoxically impairs the in vivo antitumor efficacy of adoptively transferred CD8+ T cells. The Journal of clinical investigation. 2005;115(6):1616-26. Wang D, Starr R, Alizadeh D, Yang X, Forman SJ, Brown CE. In Vitro Tumor Cell Rechallenge For Predictive Evaluation of Chimeric Antigen Receptor T Cell Antitumor Function. Journal of visualized experiments : JoVE. 2019(144). Gotthardt D, Trifinopoulos J, Sexl V, Putz EM. JAK/STAT Cytokine Signaling at the Crossroad of NK Cell Development and Maturation. Frontiers in immunology. 2019;10:2590. Panwar V, Singh A, Bhatt M, Tonk RK, Azizov S, Raza AS, et al. Multifaceted role of mTOR (mammalian target of rapamycin) signaling pathway in human health and disease. Signal transduction and targeted therapy. 2023;8(1):375. Caruso S, De Angelis B, Del Bufalo F, Ciccone R, Donsante S, Volpe G, et al. Safe and effective off-the-shelf immunotherapy based on CAR.CD123-NK cells for the treatment of acute myeloid leukaemia. Journal of hematology & oncology. 2022;15(1):163. Vahidi S, Zabeti Touchaei A, Samadani AA. IL-15 as a key regulator in NK cell-mediated immunotherapy for cancer: From bench to bedside. International immunopharmacology. 2024;133:112156. Ojo EO, Sharma AA, Liu R, Moreton S, Checkley-Luttge MA, Gupta K, et al. Membrane bound IL-21 based NK cell feeder cells drive robust expansion and metabolic activation of NK cells. Scientific reports. 2019;9(1):14916. Sharma R, Das A. IL-2 mediates NK cell proliferation but not hyperactivity. Immunologic research. 2018;66(1):151-7. Wu Y, Tian Z, Wei H. Developmental and Functional Control of Natural Killer Cells by Cytokines. Frontiers in immunology. 2017;8:930. Sun D, Shi X, Li S, Wang X, Yang X, Wan M. CAR‑T cell therapy: A breakthrough in traditional cancer treatment strategies (Review). Molecular medicine reports. 2024;29(3). Shrestha N, Dee MJ, Chaturvedi P, Leclerc GM, Mathyer M, Dufour C, et al. A "Prime and Expand" strategy using the multifunctional fusion proteins to generate memory-like NK cells for cell therapy. Cancer immunology, immunotherapy : CII. 2024;73(9):179. Oyer JL, Croom-Perez TJ, Hasan MF, Rivera-Huertas JA, Gitto SB, Mucha JM, et al. PM21-particle stimulation augmented with cytokines enhances NK cell expansion and confers memory-like characteristics with enhanced survival. Frontiers in immunology. 2024;15:1383281. Romee R, Foley B, Lenvik T, Wang Y, Zhang B, Ankarlo D, et al. NK cell CD16 surface expression and function is regulated by a disintegrin and metalloprotease-17 (ADAM17). Blood. 2013;121(18):3599-608. Vidard L, Dureuil C, Baudhuin J, Vescovi L, Durand L, Sierra V, et al. CD137 (4-1BB) Engagement Fine-Tunes Synergistic IL-15- and IL-21-Driven NK Cell Proliferation. Journal of immunology (Baltimore, Md : 1950). 2019;203(3):676-85. Cabo M, Santana-Hernández S, Costa-Garcia M, Rea A, Lozano-Rodríguez R, Ataya M, et al. CD137 Costimulation Counteracts TGFβ Inhibition of NK-cell Antitumor Function. Cancer immunology research. 2021;9(12):1476-90. Hwang I, Lee J, Kim SW, Sung YC. Dual IL-2/IL-21 priming augments the anti-tumor function of CD16V-CAR/TRAIL-engineered NK cells in combination with monoclonal antibodies. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie. 2025;189:118289. Zeng R, Spolski R, Casas E, Zhu W, Levy DE, Leonard WJ. The molecular basis of IL-21-mediated proliferation. Blood. 2007;109(10):4135-42. Romee R, Rosario M, Berrien-Elliott MM, Wagner JA, Jewell BA, Schappe T, et al. Cytokine-induced memory-like natural killer cells exhibit enhanced responses against myeloid leukemia. Science translational medicine. 2016;8(357):357ra123. Rohde C, Zhang Y, Reinhardt R, Jeltsch A. BISMA--fast and accurate bisulfite sequencing data analysis of individual clones from unique and repetitive sequences. BMC bioinformatics. 2010;11:230. Additional Declarations (Not answered) Supplementary Files Supplementaltables.xlsx Supplemental tables SupFig1.jpg Supplemental Figure 1 SupFig2.jpg Supplemental Figure 2 Supfig3.jpg Supplemental Figure 3 supfig4.jpg Supplemental Figure 4 Cite Share Download PDF Status: Under Review Version 1 posted Unknown event 17 Feb, 2026 Editorial decision: Reject after peer review 10 Feb, 2026 Reviewers invited by journal 25 Nov, 2025 Submission checks completed at journal 13 Nov, 2025 First submitted to journal 12 Nov, 2025 Unknown event 12 Nov, 2025 Editor assigned by journal 11 Nov, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8091256","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":550416867,"identity":"f95dd840-c66d-411c-948e-887f4f5c8220","order_by":0,"name":"Zhiying He","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2UlEQVRIiWNgGAWjYFACHgaGD2DGwQYGBjYitTDOSCBVCzNPAoxDjBb5GbkHH9v+sMmTdzzcwPCh7DAD/+wG/FoYZ+QlG+ckpBUbHjjYwDjj3GEGiTsH8Gthlsgxk85JOJy4seFgAzNv22EGA4kE/FrYJHLMf1vAtPwlRgsP0BZmBqCW+cAQY2YkRosEzxtjyZ60tMQNQC0He86l80jcIKBFvj3H8MMPG5vE+TOOP3zwo8xajn8GAS0MAlAFBjcOMBxgACcGQoD/ANS6/gbCikfBKBgFo2BkAgAebkauouzIBAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-8298-7835","institution":"Shanghai East Hospital","correspondingAuthor":true,"prefix":"","firstName":"Zhiying","middleName":"","lastName":"He","suffix":""},{"id":550416868,"identity":"d8706c45-08f2-42b2-87cb-9476d6a4fa2e","order_by":1,"name":"Hongye Wang","email":"","orcid":"","institution":"Shanghai East Hospital, School of Life Sciences and Technology, Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Hongye","middleName":"","lastName":"Wang","suffix":""},{"id":550416869,"identity":"1f2aaf29-cc82-4977-b6f0-8e7e0d76399d","order_by":2,"name":"Jingwei Tu","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Jingwei","middleName":"","lastName":"Tu","suffix":""},{"id":550416870,"identity":"1a5abc4f-d68d-4ba6-9c9b-868422943d88","order_by":3,"name":"Yi Luo","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Luo","suffix":""},{"id":550416871,"identity":"38b2b45b-c7a3-4e19-8bfe-76e52d25d712","order_by":4,"name":"Xinxin Yuan","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Xinxin","middleName":"","lastName":"Yuan","suffix":""},{"id":550416872,"identity":"6128e379-825c-44e2-8985-8ada2faaa4bb","order_by":5,"name":"Xin Fang","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Fang","suffix":""},{"id":550416873,"identity":"c32cba41-a5e0-4aef-b4e8-fe236e28d136","order_by":6,"name":"Zhenyi Lu","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Zhenyi","middleName":"","lastName":"Lu","suffix":""},{"id":550416874,"identity":"2487a2fd-4f61-4b27-a582-cb39465882b9","order_by":7,"name":"Li Han","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Han","suffix":""},{"id":550416875,"identity":"5e085b64-6d39-42a8-9a66-a5006b7f323e","order_by":8,"name":"Yongjia Xue","email":"","orcid":"","institution":"Shanghai East Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yongjia","middleName":"","lastName":"Xue","suffix":""},{"id":550416876,"identity":"75dae8ec-aa59-4c22-a962-223a5f1d01bf","order_by":9,"name":"Wencheng Zhang","email":"","orcid":"https://orcid.org/0000-0003-2848-1611","institution":"Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Wencheng","middleName":"","lastName":"Zhang","suffix":""},{"id":550416877,"identity":"8cf4ad1f-ccfb-4081-9f5e-fe66993b107e","order_by":10,"name":"Qianggang Dong","email":"","orcid":"","institution":"Shanghai iCELL Biotechnology Co., Ltd","correspondingAuthor":false,"prefix":"","firstName":"Qianggang","middleName":"","lastName":"Dong","suffix":""}],"badges":[],"createdAt":"2025-11-12 02:40:59","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8091256/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8091256/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97007902,"identity":"45bd9e45-886c-4151-aaf2-1a4f89270f44","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"jpg","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3237984,"visible":true,"origin":"","legend":"","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/0f966681ac0e4d565f8d09cb.jpg"},{"id":97007894,"identity":"ab2074a5-5704-48dd-a4e9-13d42eabcf29","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174327,"visible":true,"origin":"","legend":"","description":"","filename":"MainDocument.docx","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/caf1c0aaa69ef4bd863834e3.docx"},{"id":97139807,"identity":"30fc957c-6b90-4319-a39e-4fc8052f9364","added_by":"auto","created_at":"2025-12-01 10:02:32","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6098879,"visible":true,"origin":"","legend":"","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/4f8fcee6cfdb82431c91d404.jpg"},{"id":97007901,"identity":"70549425-d0e8-4cf8-beb8-d5fb3b8f7f8b","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1734263,"visible":true,"origin":"","legend":"","description":"","filename":"fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/8fad852cd88367d8331c7c64.jpg"},{"id":97140390,"identity":"7911dfec-91e8-4576-bf94-1ac69025f32b","added_by":"auto","created_at":"2025-12-01 10:04:48","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5599702,"visible":true,"origin":"","legend":"","description":"","filename":"fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/d553c254982d835a37998e57.jpg"},{"id":97007908,"identity":"f871bb03-9f12-4b91-a280-d216551315d2","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4112242,"visible":true,"origin":"","legend":"","description":"","filename":"fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/ed82ca43b89dd663fd8ca8ee.jpg"},{"id":97007915,"identity":"3b370af5-bb74-4a21-920b-449bd48c35d8","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5413968,"visible":true,"origin":"","legend":"","description":"","filename":"fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/f31c2338e0113ad25adc667c.jpg"},{"id":97140105,"identity":"1522e1a2-9399-4fe2-b23d-2c516fc7bbac","added_by":"auto","created_at":"2025-12-01 10:03:49","extension":"json","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11960,"visible":true,"origin":"","legend":"","description":"","filename":"CDDIS257230.json","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/8b9a527f03df0c43d5d0263a.json"},{"id":97007919,"identity":"baeff9e5-b0cf-42cc-8455-323c0d51f454","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7150538,"visible":true,"origin":"","legend":"","description":"","filename":"SupFig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/a9c0856416a1a1ac34e45509.jpg"},{"id":97007926,"identity":"2d2a2277-606a-4f30-9847-875bc00c5033","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3773541,"visible":true,"origin":"","legend":"","description":"","filename":"SupFig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/b7a4d553c31621f47ad56709.jpg"},{"id":97007923,"identity":"109a0854-f3fe-48bd-997d-8393f01c8bc6","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5268133,"visible":true,"origin":"","legend":"","description":"","filename":"Supfig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/77e1f1436f8df6109317d918.jpg"},{"id":97139462,"identity":"50dc2faa-c151-413e-9f7e-adf2a4a33cdc","added_by":"auto","created_at":"2025-12-01 10:00:26","extension":"xlsx","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23116,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaltables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/af12621e27377faf730183c5.xlsx"},{"id":97007911,"identity":"78f755fe-0687-4d1c-b42f-52e3371a024d","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4418563,"visible":true,"origin":"","legend":"","description":"","filename":"supfig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/4279f80d0f0dc84ec9f55bb7.jpg"},{"id":97007913,"identity":"3f088a20-4ed0-4e8e-9480-131046db377a","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"xml","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":167017,"visible":true,"origin":"","legend":"","description":"","filename":"CDDIS2572300enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/30b9c5a43f9c08ff04f5d549.xml"},{"id":97007930,"identity":"0435c6d7-3d9c-47be-815f-0a2e0972719c","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3237984,"visible":true,"origin":"","legend":"","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/f8699340a667f620a828e883.jpg"},{"id":97138378,"identity":"a16ed060-6c02-4ee4-b841-bdae6396a33a","added_by":"auto","created_at":"2025-12-01 09:58:49","extension":"jpg","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6098879,"visible":true,"origin":"","legend":"","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/e3d739d54755bacbfd893974.jpg"},{"id":97007928,"identity":"9cd0c560-36c1-4519-ab77-6cbd55a908d9","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1734263,"visible":true,"origin":"","legend":"","description":"","filename":"fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/442dfedbbca7b47dd36851ca.jpg"},{"id":97007921,"identity":"e90265df-1a66-40e2-bc45-b3473955651c","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5599702,"visible":true,"origin":"","legend":"","description":"","filename":"fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/ff48e6ac49c4a94dbe1e80ad.jpg"},{"id":97140127,"identity":"a00281e5-4df2-4841-9cd7-cba1dac5fd37","added_by":"auto","created_at":"2025-12-01 10:03:52","extension":"jpg","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4112242,"visible":true,"origin":"","legend":"","description":"","filename":"fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/30581fbdd63143610f6faf8d.jpg"},{"id":97007927,"identity":"628e5500-ffc3-4c9b-9f38-45b9549dd267","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5413968,"visible":true,"origin":"","legend":"","description":"","filename":"fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/3d896b73dc26d1e97e1552a7.jpg"},{"id":97007916,"identity":"29d34ac9-739d-4177-aec5-c5883bcb1bd9","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":264781,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig1.png","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/7ff6ef5478854db0e3734433.png"},{"id":97139460,"identity":"443f6ea5-78ea-4f29-911f-0f3626fa6e89","added_by":"auto","created_at":"2025-12-01 10:00:26","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":507502,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig2.png","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/bb0f7bc96eecd7d57d634cef.png"},{"id":97137914,"identity":"3ae30850-8484-4466-8d36-ce0b90f72df8","added_by":"auto","created_at":"2025-12-01 09:58:18","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174483,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig3.png","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/cbd149c31e0330f1f00641e1.png"},{"id":97007918,"identity":"9db73deb-2d37-49ea-b1b0-3749ab494e0d","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":369722,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig4.png","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/5f3a872bc0ac8ea3937370b4.png"},{"id":97007922,"identity":"673b1f22-5937-43cf-bd76-4f3e3de2fb19","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":394485,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig5.png","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/c845c3640900039dc99660c9.png"},{"id":97007924,"identity":"8011de4d-bf77-4b94-89d7-12940e893e0f","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":338565,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig6.png","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/12617d07affdcbb71ef8e013.png"},{"id":97139319,"identity":"e727903a-f4c4-4857-b2ed-e8f1ae67deb0","added_by":"auto","created_at":"2025-12-01 10:00:01","extension":"xml","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":163629,"visible":true,"origin":"","legend":"","description":"","filename":"CDDIS2572300structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/5f3bfdc9f3111f92a5180a80.xml"},{"id":97007925,"identity":"312ff0e8-9d4f-4aa7-839e-bc969b3780b7","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"html","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":181173,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/d79899dc66fc6f252f51d72b.html"},{"id":97138478,"identity":"ee1a6029-4ba4-4008-a5cd-f8441640d32a","added_by":"auto","created_at":"2025-12-01 09:58:56","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3237984,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u0026nbsp;\u003c/p\u003e","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/b29eab54e44de8f3b947c458.jpg"},{"id":97007895,"identity":"41ac6c5a-3ba9-4b67-809e-51699c8556cb","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":6098879,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/6495fac1167a533c05edfbc5.jpg"},{"id":97007896,"identity":"cffae5d7-d8dd-4ec9-88e6-7697578f7155","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1734263,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/692f434d81fab9e74f73bf41.jpg"},{"id":97007904,"identity":"8f65bc16-5179-4a75-8e34-e82479a82c21","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":5599702,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u0026nbsp;\u003c/p\u003e","description":"","filename":"fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/0f799379ca99854199c4d1b3.jpg"},{"id":97007903,"identity":"c5c64d67-2195-4388-9e73-4ba209a55513","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":4112242,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u0026nbsp;\u003c/p\u003e","description":"","filename":"fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/53adbd3dfc902af76e73e887.jpg"},{"id":97007910,"identity":"d5222ba0-f291-47d0-bcf5-91885f785fb4","added_by":"auto","created_at":"2025-11-28 15:02:27","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":5413968,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u0026nbsp;\u003c/p\u003e","description":"","filename":"fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/83f88f1a49b84005f3edcccc.jpg"},{"id":102342927,"identity":"55763bfb-ef11-4b4e-8b8c-c832069d8ef7","added_by":"auto","created_at":"2026-02-10 17:01:51","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":15671714,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/286a2053-536a-43a8-b22b-d9b986f6deff.pdf"},{"id":97007893,"identity":"bc671572-2ea0-4896-aebf-d86fb323654c","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":23116,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental tables\u003c/p\u003e","description":"","filename":"Supplementaltables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/640dba4f847109f38aad347a.xlsx"},{"id":97007898,"identity":"fe84dd82-821c-4772-b24e-ffd07e160ce0","added_by":"auto","created_at":"2025-11-28 15:02:26","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":7150538,"visible":true,"origin":"","legend":"Supplemental Figure 1","description":"","filename":"SupFig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/bfe59d230d74c448b600d0f6.jpg"},{"id":97139187,"identity":"f01e0624-7204-4600-a0c1-921cf7f8fc98","added_by":"auto","created_at":"2025-12-01 09:59:44","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":3773541,"visible":true,"origin":"","legend":"Supplemental Figure 2","description":"","filename":"SupFig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/c860e87f1e2a38c8d621dc2c.jpg"},{"id":97137811,"identity":"34a39906-e805-48f6-8b27-d6e7c0e4b66d","added_by":"auto","created_at":"2025-12-01 09:58:11","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":5268133,"visible":true,"origin":"","legend":"Supplemental Figure 3","description":"","filename":"Supfig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/30dd9d2b32e4d7cafcc57b13.jpg"},{"id":97138928,"identity":"22e7943c-7e93-4608-b253-b9c67ba9218d","added_by":"auto","created_at":"2025-12-01 09:59:26","extension":"jpg","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":4418563,"visible":true,"origin":"","legend":"Supplemental Figure 4","description":"","filename":"supfig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8091256/v1/09584edc51ade63a90ef173b.jpg"}],"financialInterests":"(Not answered)","formattedTitle":"Cytokine-induced Expandable Memory NK Cells with Significant Metabolic, Epigenetic Remodeling, and Persistence Properties","fulltext":[{"header":"Introduction","content":"\u003cp\u003eNatural killer (NK) cells are commonly described as a type of innate lymphocytes with natural cytotoxicity. This functional definition means that NK cells can exhibit a rapid cytolytic activity against malignant tumor cells and virally infected cells without requiring prior antigen sensitization (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). NK cells express a variety of germline-encoded, stochastically expressed activating and inhibitory receptors that control their functional responses to the target cells (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). Activated NK cells can eliminate tumor cells through multiple killing mechanisms: they can release perforin and granzymes to act directly on target cells; and they also enable to mediate antibody-dependent cellular cytotoxicity (ADCC) via the Fc gamma receptor (FcγRIIIa / CD16a) expressed on their surface. Therefore, NK cell-based adoptive cell therapy (ACT) has been investigated for decades. Early clinical studies indicated that NK cell infusion, even with those originating from allogeneic sources, has a higher safety profile (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). While this NK cell-mediated protective immunity holds great promise in cancer therapy, challenges remain, including their absence of immune memory so that NK cells frequently display a limited persistence and functionality \u003cem\u003ein vivo\u003c/em\u003e, which would compromise their clinical effectiveness (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn the last two decades, however, accumulating evidence have revealed that the innate immune system, similar to the adaptive immune system, is also capable of exhibiting immune memory, a phenomenon observed in cells such as macrophages, dendritic cells, and NK cells, known as \"trained immunity\" (\u003cspan additionalcitationids=\"CR8 CR9\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Currently, the NK cell memory has been studied in three main scenarios: hapten specific, virus specific and cytokine induced (\u003cspan additionalcitationids=\"CR12 CR13\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). While in human, the cytokine induced memory-like NK (CIML NK) cells represent one of the two major types of memory NK cells (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). CIML NK cells are induced by a triple cytokine cocktail consisting of IL-12, IL-15 and IL-18 (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). Recent studies indicated that a brief exposure to this cocktail initiates a transient proliferation of NKG2A\u0026thinsp;+\u0026thinsp;CD57- NK cell subsets that enable to simultaneously express certain enzyme markers such as EZH2 (enhancer of zeste homolog 2, a histone methyltransferase) and CD39 (the ectonucleoside triphosphate diphosphohydrolase-1) (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). These proliferating NK cells immediately exhibit a shift in the glycose metabolism from oxidative phosphorylation to aerobic glycolysis and the inheritable epigenetic changes such as the DNA methylation state in the CNS-1 region (conserved noncoding sequence 1) of IFN-γ encoding gene IFNG (\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). The synergistic coordination between the metabolic rewiring and the epigenetic reprogramming described above, and among the others, composed the molecular basis of NK cell memory (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). Upon re-encountering tumor targets or cytokines, CIML NK cells demonstrate markedly enhanced IFN-γ production and effector functions (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). They can persist for several months in both clinical patients and mouse models, significantly longer than conventional NK cells (\u003cspan additionalcitationids=\"CR25\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Multiple clinical trials (NCT02782546, NCT03068819, NCT04024761, NCT04290546, NCT04354025, and NCT04634435, from \u003cem\u003eclinicaltrials.gov\u003c/em\u003e) have been initiated with CIML NK cells, demonstrating notable safety and efficacy in treating acute myeloid leukemia (AML) patients.\u003c/p\u003e\u003cp\u003eHowever, the current preparation methods for CIML NK cells used in these trials lack effective cell expansion, limiting their broad clinical application. The growing clinical demand for large quantities of CIML NK cells urgently requires the development of more efficient and convenient methods for their \u003cem\u003eex vivo\u003c/em\u003e expansion.\u003c/p\u003e\u003cp\u003eIn this context, the robust expansion of CIML NK cells was proved achievable when using a combination of CD16a antibody and IL-15/IL-18 cytokines in the early activation phase to induce the development of immune memory, followed by the use of a cytokine cocktail containing IL-2, IL-15, and IL-21 in the later phase to ensure the maintenance of memory properties during the high-efficiency expansion stage. Here, we developed a novel feeder-free platform for expanding memory NK cells. By using of this method, we achieved more than 200-fold expansion of NK cells after 14 days in culture. We then systematically and rigorously validated the immune memory attributes of these expanded NK cells compared with CIML NK and PBNK cells across five dimensions: phenotype, metabolism, epigenetics, function, and persistence. Thus, we named these named these expanded but still possess memory-like characteristics NK cells as mExNK (memory-like expanded NK cells). The scalable expansion of mExNK cells would pave a new avenue for their clinical applications.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eEffective expansion of human NK cells with the feeder-free, cytokine induced system\u003c/h2\u003e\u003cp\u003eWhile unstimulated NK cells exhibit limited antitumor efficacy in adoptive cell therapy, multiple activation and expansion protocols have been developed to enhance their cytotoxic function (\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). To prevent T cell overgrowth in final products, we first depleted CD3\u003csup\u003e+\u003c/sup\u003e T cells from PBMCs using immunomagnetic beads. Building on evidence from cytokine-induced memory-like (CIML) NK studies, combinations of CD16a antibody with cytokines such as IL-12, IL-15, and IL-18 can promote memory-like properties (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAccording to studies, NK cell activation upregulates the metalloprotease ADAM17, which cleaves surface receptors including CD16a (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Therefore, we stimulated NK cells with CD16a antibody plus IL-15/IL-18 during the first two days to induce memory-like characteristics \u003cb\u003e(Fig.\u0026nbsp;1A)\u003c/b\u003e. The ADAM17 inhibitor INCB3619 was added concurrently to limit receptor shedding and preserve CD16a expression. CD137 antibody and IL-21, recognized as strong proliferative signals for NK cells (\u003cspan additionalcitationids=\"CR34\" citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e), were subsequently introduced to drive expansion \u003cb\u003e(Fig.\u0026nbsp;1A)\u003c/b\u003e. From day 4 onward, IL-2, IL-15, and IL-21 were supplemented to sustain long-term culture. This feeder-free system consistently yielded expanded NK (ExNK) cells with high viability (\u0026gt;\u0026thinsp;80%) and purity (\u0026ge;\u0026thinsp;95%) \u003cb\u003e(Fig.\u0026nbsp;1B-D)\u003c/b\u003e. Across 12 healthy donors, the mean expansion at two weeks was ~\u0026thinsp;230-fold (range from 110 to 543), with some donors sustaining growth for up to four weeks \u003cb\u003e(Fig.\u0026nbsp;1E-F)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eFlow cytometry revealed that ExNK cells expressed high levels of activating receptors (CD16, NKG2D, NKp30, NKp44, NKp46) and inhibitory receptors NKG2A and CD158b, whereas CD158a and CD158e1 were detected on fewer than 20% of cells \u003cb\u003e(Fig.\u0026nbsp;1G)\u003c/b\u003e. Real-time cytotoxicity assays (Incucyte) against K562 targets showed that at 4 hours, ExNK cells mediated \u0026sim;20% killing at a 1:1 E:T ratio, increasing to \u0026sim;80% at 10:1 \u003cb\u003e(Fig.\u0026nbsp;1H-I)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eIn summary, we established a feeder-free expansion system that generates highly pure (\u0026ge;\u0026thinsp;95%), functionally active NK cells expressing multiple activating receptors and exhibiting potent tumor-killing capacity. Although early activation included CD16 antibody and IL-15/IL-18 to promote memory-like features, whether such properties persist after two weeks of expansion requires further investigation.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eExNK cells showed apparent memory-like properties as determined by phenotype, metabolism and epigenetic remodeling\u003c/h3\u003e\n\u003cp\u003ePrior studies have established that CIML NK cells exhibit distinct features\u0026mdash;including a specific surface phenotype, metabolic remodeling, and epigenetic reprogramming\u0026mdash;which collectively contribute to their elevated cytotoxicity \u003cb\u003e(Summarized in Fig.\u0026nbsp;2A)\u003c/b\u003e (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Here, we systematically evaluated our expanded NK (ExNK) cells against these established CIML benchmarks.\u003c/p\u003e\u003cp\u003eFirstly, we isolated and purified peripheral blood NK cells (PBNKs) from healthy donors and cultured them in low-dose IL-15 (1ng/ml) for 3\u0026ndash;5 days to serve as our negative control. PBNKs from the same donor underwent IL-12/15/18 pre-activation for 16\u0026ndash;24 hours, followed by resting in low-dose IL-15 (1 ng/ml) for 3\u0026ndash;5 days, to serve as the positive control cells (CIML NK). Morphologically, both CIML NK and ExNK cells formed prominent clusters, whereas PBNKs did not \u003cb\u003e(Fig.\u0026nbsp;2B)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eIL-12/15/18 pre-activation strongly upregulates CD25 (IL-2Rα), enabling formation of high-affinity IL-2 receptors and supporting robust proliferation and cytotoxicity even under low IL-2 conditions (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e). Flow cytometry revealed that ExNK and CIML NK cells similarly upregulated CD25 and the early activation marker CD69 compared to PBNKs, while expression of CD16, NKp46, and NKG2D remained largely unchanged \u003cb\u003e(Fig.\u0026nbsp;2C and Sup Fig.\u0026nbsp;1A)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eMetabolically, resting NK cells exhibit low activity, but activation induces a pronounced glycolytic shift, accompanied by increased expression of nutrient transporters such as GLUT1 (SLC2A1), amino acid transporters (SLC1A5, SLC7A5, SLC3A2/CD98), and CD71 (transferrin receptor) (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e). Consistent with a CIML-like profile, ExNK cells showed elevated levels of these transporters, comparable to CIML NK cells \u003cb\u003e(Fig.\u0026nbsp;2D-E)\u003c/b\u003e. Functional metabolic assays confirmed that both ExNK and CIML NK cells exhibited enhanced glycolysis, glycolytic capacity, and glycolytic reserve relative to PBNKs \u003cb\u003e(Fig.\u0026nbsp;2F-G)\u003c/b\u003e, though oxidative phosphorylation (OXPHOS) parameters\u0026mdash;including basal respiration, spare respiratory capacity, and ATP production\u0026mdash;remained unchanged across all groups \u003cb\u003e(Sup Fig.\u0026nbsp;1B-C)\u003c/b\u003e. The ECAR:OCR ratio, reflecting a shift toward glycolytic metabolism (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e), was similarly reduced in both ExNK and CIML NK cells \u003cb\u003e(Fig.\u0026nbsp;2H)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eA hallmark of CIML NK cells is their capacity for enhanced IFN-γ production upon restimulation, which is epigenetically regulated by demethylation of CpG sites in the conserved noncoding sequence 1 (CNS‑1) of the \u003cem\u003eIFNG\u003c/em\u003e locus (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan additionalcitationids=\"CR42\" citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). We assessed the methylation status of six key CpG sites within the \u003cem\u003eIFNG\u003c/em\u003e CNS‑1 region as an indicator of memory-like reprogramming. ExNK cells exhibited profound hypomethylation at these sites, with an average methylation level of only 6%, compared to 54.8% in CIML NK and 76.7% in PBNK cells \u003cb\u003e(Fig.\u0026nbsp;2I-J and Sup Fig.\u0026nbsp;1D-E)\u003c/b\u003e, indicating that ExNK cells undergo extensive epigenetic remodeling.\u003c/p\u003e\u003cp\u003eIn summary, ExNK cells closely resemble CIML NK cells in their immunophenotype, metabolic profile, and epigenetic state of the \u003cem\u003eIFNG\u003c/em\u003e CNS‑1 region\u0026mdash;distinguishing them clearly from PBNKs. Based on these findings, we designate these expanded cells as mExNK (memory-like expanded NK cells).\u003c/p\u003e\u003cp\u003e\u003cb\u003emExNK had potent effector function and persistence ability\u003c/b\u003e \u003cb\u003ein vitro\u003c/b\u003e\u003c/p\u003e\u003cp\u003eBuilding on our demonstration that mExNK cells possess immune memory, we postulated that they would also exhibit enhanced effector functions, akin to CIML NK cells. To test this, we first evaluated cytokine expression following a 6-hour co-culture with K562 target cells (E:T ratio\u0026thinsp;=\u0026thinsp;1:1). Flow cytometric analysis revealed that both mExNK and CIML NK cells produced significantly higher levels (measured by both percentage and MFI) of IFN-γ, TNF-α, and Granzyme B than PBNKs cultured in low-dose IL-15, with mExNK cells consistently showing the most robust enhancement \u003cb\u003e(Fig.\u0026nbsp;3A, 3B, 3D)\u003c/b\u003e. Perforin expression was also elevated in mExNK compared to PBNK, though not significantly different from CIML NK \u003cb\u003e(Fig.\u0026nbsp;3C)\u003c/b\u003e. These results were corroborated by ELISA after 24-hour co-culture, confirming that mExNK and CIML NK cells secreted significantly more IFN-γ, Granzyme B, and perforin than PBNKs \u003cb\u003e(Fig.\u0026nbsp;3F-H)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTraditional methods for detecting NK cell effector function \u003cem\u003ein vitro\u003c/em\u003e primarily focus on measuring cytokine secretion, and killing of target cells. However, due to the short co-culture times and high E:T ratios typically used, these methods often fail to accurately and objectively reflect the true anti-tumor capacity of immune cells \u003cem\u003ein vivo\u003c/em\u003e (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Tumor rechallenge assay can be coupled with profiling immune cell activation, exhaustion and persistence phenotypes, provide a more accurate reflection of the \u003cem\u003ein vivo\u003c/em\u003e anti-tumor capacity of immune cells and are widely used, particularly in the \u003cem\u003ein vitro\u003c/em\u003e functional evaluation of CAR-T cells (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). Given evidence that CIML NK cells display prolonged persistence (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e), we developed a rechallenge model to assess serial killing capacity under high tumor burden. PBNK, CIML NK, and mExNK cells were co-cultured with three tumor lines (K562, Raji, THP-1) at E:T ratios of 1:1 or 1:2. Fresh, fluorescently labeled target cells were replenished every two days, and killing was monitored over three rounds (6 days) using Incucyte live-cell imaging. mExNK cells consistently exhibited stronger and more sustained tumor killing than both CIML NK and PBNK cells, with differences emerging as early as 6\u0026ndash;8 hours and maintained throughout the assay \u003cb\u003e(Fig.\u0026nbsp;3H-G and Sup Fig.\u0026nbsp;2A-I)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTogether, these data demonstrate that mExNK cells possess not only enhanced effector function relative to CIML NK cells, but also superior persistence in serial killing assays.\u003c/p\u003e\u003cp\u003e\u003cb\u003eThe mExNK cells display an advanced effector and activation gene expression, but similar memory-like profiles with CIML NK\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe next performed bulk mRNA-sequencing to define the transcriptional landscapes of mExNK and CIML NK cells relative to PBNK cells. The analysis yielded two central findings. First, mExNK and CIML NK cells are highly similar: the number of DEGs between each of them and PBNK was far greater than the number of DEGs between them (\u003cb\u003eFig.\u0026nbsp;4A and Sup Fig.\u0026nbsp;3A)\u003c/b\u003e, and their sample correlation was markedly higher than with PBNK \u003cb\u003e(Fig.\u0026nbsp;4B)\u003c/b\u003e. This similarity was quantified by the substantial overlap in their individual DEG sets compared to PBNK (~\u0026thinsp;71\u0026ndash;72% of up- and down-regulated genes; \u003cb\u003eSup Fig.\u0026nbsp;3A-C\u003c/b\u003e) and confirmed by GO analysis, which showed shared enrichment in pathways like DNA replication, oxidative phosphorylation and cellular amino acid metabolic process \u003cb\u003e(Sup Fig.\u0026nbsp;3F)\u003c/b\u003e. The second key finding was the specific signature of mExNK cells. Direct comparison with CIML NK cells revealed a unique enrichment in pathways governing T cell activation, cell-cell adhesion, and cytokine regulation \u003cb\u003e(Sup Fig.\u0026nbsp;3D)\u003c/b\u003e. Importantly, this transcriptional prediction of enhanced adhesive and effector capacity was functionally validated by the superior performance of mExNK cells in the tumor rechallenge assay \u003cb\u003e(Sup Fig.\u0026nbsp;3E)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eGene Set Enrichment Analysis (GSEA) using MSigDB Hallmark gene sets further highlighted strong similarities between mExNK and CIML NK cells \u003cb\u003e(Sup Fig.\u0026nbsp;3G)\u003c/b\u003e. Both populations showed significant upregulation of gene sets involved in E2F targets, G2M checkpoint, mitotic spindle, IL2‑STAT5 signaling, and mTORC1 signaling compared to PBNK \u003cb\u003e(Sup Fig.\u0026nbsp;3H-I)\u003c/b\u003e. Given the established roles of IL2‑STAT5 and mTORC1 pathways in NK cell survival, metabolism, and function (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e), we focused subsequent validation on these, along with oxidative phosphorylation (OXPHOS). GSEA confirmed that mExNK and CIML NK cells exhibited significant upregulation of all three pathways relative to PBNK \u003cb\u003e(Fig.\u0026nbsp;4C-E)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eNotably, even in a resting state, mExNK cells maintained higher expression of key effector genes, including \u003cem\u003eTNFSF10\u003c/em\u003e (TRAIL), \u003cem\u003eFASLG\u003c/em\u003e (FASL), \u003cem\u003eGZMA\u003c/em\u003e, \u003cem\u003eGZMB\u003c/em\u003e, and \u003cem\u003eIFNG\u003c/em\u003e \u003cb\u003e(Fig.\u0026nbsp;4F)\u003c/b\u003e. Both mExNK and CIML NK cells highly expressed \u003cem\u003eKLRC1\u003c/em\u003e (NKG2A), \u003cem\u003eIL2RA\u003c/em\u003e (CD25), and genes encoding nutrient transporters such as \u003cem\u003eTFRC\u003c/em\u003e (CD71), \u003cem\u003eSLC1A5\u003c/em\u003e, and \u003cem\u003eSLC7A5\u003c/em\u003e \u003cb\u003e(Fig.\u0026nbsp;4G)\u003c/b\u003e, consistent with our phenotypic and metabolic data \u003cb\u003e(Fig.\u0026nbsp;2C-E)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTranscriptomic profiling also revealed elevated expression of \u003cem\u003eEZH2\u003c/em\u003e and \u003cem\u003eMKI67\u003c/em\u003e, alongside low expression of \u003cem\u003eB3GAT1\u003c/em\u003e (CD57), in both mExNK and CIML NK cells \u003cb\u003e(Fig.\u0026nbsp;4G)\u003c/b\u003e. This profile matches with a previously reported CD57-NKG2A\u0026thinsp;+\u0026thinsp;EZH2\u0026thinsp;+\u0026thinsp;MKI67\u0026thinsp;+\u0026thinsp;CIML NK subset associated with rapid cycling and enhanced trained immunity (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). Furthermore, we observed upregulation of activation-related genes (e.g., \u003cem\u003eIRF4\u003c/em\u003e, \u003cem\u003eTNFRSF9\u003c/em\u003e, \u003cem\u003eMYC\u003c/em\u003e) and downregulation of maturation- and adhesion-related genes (e.g., \u003cem\u003eFGFBP2\u003c/em\u003e, \u003cem\u003eKLF2\u003c/em\u003e, \u003cem\u003eCX3CR1\u003c/em\u003e), mirroring recently reported memory-like NK transcriptional signatures (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e) \u003cb\u003e(Fig.\u0026nbsp;4H)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eIn summary, mExNK cells share a core memory-like transcriptional program with CIML NK cells but also display a distinct gene expression profile linked to adhesion and effector function\u0026mdash;consistent with their enhanced and sustained anti-tumor activity.\u003c/p\u003e\u003cp\u003e\u003cb\u003emExNK cells showed efficacy against THP‑1 leukemia tumors and persist long‑term\u003c/b\u003e \u003cb\u003ein vivo\u003c/b\u003e\u003c/p\u003e\u003cp\u003eNext, we were interested to investigate the effector function and persistence of our mExNK cells \u003cem\u003ein vivo\u003c/em\u003e during targeting the tumor cells. We employed an NSG mouse model engrafted with THP-1-Luc human leukemic monocytic cells to evaluate the anti-tumor activity of mExNK cells (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). THP-1-Luc⁺ tumor cells were administered intravenously via the tail vein at a dose of 2\u0026times;10⁵ per mouse. Three days later, mExNK cells were infused intravenously at a dose of 1\u0026times;10⁷ per mouse. Tumor-bearing mice without transferred NK cells served as the control group. After NK cell transfer, each mouse received intraperitoneal injections of cytokines (50,000 U IL-2 and 1 ng IL-15 per injection) twice to three times per week for three weeks to support NK cell survival \u003cb\u003e(Fig.\u0026nbsp;5A)\u003c/b\u003e. Tumor growth and mouse body weight were monitored weekly until control mice largely died around 7 weeks. Compared to the control group, mExNK cell treatment significantly reduced tumor burden and prolonged survival \u003cb\u003e(Fig.\u0026nbsp;5B-D)\u003c/b\u003e, indicating potent \u003cem\u003ein vivo\u003c/em\u003e anti-tumor capability of mExNK cells. Body weight monitoring revealed no significant difference between mExNK-treated and control mice, suggesting no apparent toxicity associated with mExNK cells \u003cb\u003e(Fig.\u0026nbsp;5E)\u003c/b\u003e. To test our hypothesis that mExNK cells can persist longer \u003cem\u003ein vivo\u003c/em\u003e, we collected blood from the orbital venous bleeds of mice at 1-, 2-, and 4-weeks post-infusion (under tumor-bearing conditions) and measured the proportion of hCD45⁺\u0026amp;hCD56⁺ NK cells. Although the number of mExNK cells decreased weekly, they were still detectable even four weeks after infusion \u003cb\u003e(Fig.\u0026nbsp;5F)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTo further benchmark the efficacy of mExNK cells, we included NK cells expanded using a Miltenyi Biotec commercial kit (hereafter termed cExNK, conventional expanded NK) as a control. While cExNK cells met baseline quality criteria with viability and purity both exceeding 80%, their average expansion fold within 2 weeks was substantially lower (~\u0026thinsp;40-fold) than that achieved with our system \u003cb\u003e(Sup Fig.\u0026nbsp;4A-D)\u003c/b\u003e. Phenotypically, cExNK cells showed minimal CD25 expression, indicating a lack of canonical memory-like feature \u003cb\u003e(Sup Fig.\u0026nbsp;4E)\u003c/b\u003e. Functionally, they also exhibited significantly reduced expression of key nutrient transporters and secreted lower levels of the effector molecules IFN-γ, Granzyme B, and Perforin compared to mExNK cells \u003cb\u003e(Sup Fig.\u0026nbsp;4F-I)\u003c/b\u003e. Transcriptomically, bulk mRNA-seq revealed a strikingly greater number of differentially expressed genes (DEGs) between cExNK and mExNK (4204) than between CIML NK and mExNK (1149) \u003cb\u003e(Sup Fig.\u0026nbsp;4J)\u003c/b\u003e. GO analysis indicated that cExNK cells were significantly deficient in pathways related to gene expression, immune response, and cell surface receptor signaling \u003cb\u003e(Sup Fig.\u0026nbsp;4K)\u003c/b\u003e. This was further evidenced by the marked downregulation of critical effector molecules (e.g., \u003cem\u003eIFNG\u003c/em\u003e, \u003cem\u003eGZMB\u003c/em\u003e), metabolic transporters (e.g., \u003cem\u003eSLC7A5\u003c/em\u003e, \u003cem\u003eSLC1A5\u003c/em\u003e), and crucial transcription factors (e.g., \u003cem\u003eSTAT1\u003c/em\u003e, \u003cem\u003eSTAT3\u003c/em\u003e, \u003cem\u003eRUNX1\u003c/em\u003e) in cExNK cells \u003cb\u003e(Sup Fig.\u0026nbsp;4L)\u003c/b\u003e. Collectively, these multi-faceted results demonstrate that mExNK cells possess superior properties not only relative to CIML NK cells but also, and more profoundly, to conventionally expanded NK cells.\u003c/p\u003e\u003cp\u003e\u003cb\u003eIL-21 acts in synergy with IL-2 to promote memory NK cell expansion\u003c/b\u003e \u003cb\u003eex vivo\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFinally, we sought to define the key mechanisms and culture factors that enable the generation and expansion of memory-like NK cells. Transcriptomic analysis revealed that both CIML NK and mExNK cells upregulate genes involved in mitotic cell cycle and DNA replication \u003cb\u003e(Sup Fig.\u0026nbsp;3F)\u003c/b\u003e, indicating their intrinsic proliferative potential. However, conventional IL-12/15/18 stimulation alone does not lead to substantial expansion of CIML NK cells, highlighting a major challenge in scaling memory-like NK cells for clinical use. To identify pathways critical for expansion, we compared mRNA-seq profiles of mExNK and CIML NK cells.\u003c/p\u003e\u003cp\u003eOver-representation analysis of genes upregulated in mExNK versus CIML NK cells revealed consistent and significant enrichment of the JAK-STAT signaling pathway across multiple databases (KEGG, GO-BP, Reactome, Wikipathways) \u003cb\u003e(Fig.\u0026nbsp;6A)\u003c/b\u003e. This aligns with the known roles of cytokines in our culture system: IL-15 supports NK cell survival (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e), IL-21 activates \u003cem\u003eSTAT3\u003c/em\u003e and upregulates \u003cem\u003eMYC\u003c/em\u003e to regulate proliferation and metabolism (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e), and IL-2 potently drives NK cell expansion and cytotoxicity (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e). Transcription factor activity analysis further indicated differential activation of \u003cem\u003eSTAT1\u003c/em\u003e, \u003cem\u003eSTAT3\u003c/em\u003e, \u003cem\u003eSTAT5A\u003c/em\u003e, \u003cem\u003eSTAT5B\u003c/em\u003e, \u003cem\u003eNFKB1\u003c/em\u003e, and \u003cem\u003eSTAT6\u003c/em\u003e in mExNK cells \u003cb\u003e(Fig.\u0026nbsp;6B)\u003c/b\u003e, and key JAK-STAT pathway genes (including \u003cem\u003eBCL2\u003c/em\u003e, \u003cem\u003eCCND3\u003c/em\u003e, \u003cem\u003eIL21R\u003c/em\u003e, \u003cem\u003eJAK1\u003c/em\u003e, \u003cem\u003eJAK3\u003c/em\u003e, and \u003cem\u003eSTAT1/3\u003c/em\u003e) were more highly expressed in mExNK than in CIML NK cells \u003cb\u003e(Fig.\u0026nbsp;6C)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTo functionally validate these findings, we tested different cytokine combinations in a 10-day expansion assay. While IL-21 increased IL-21R expression non-significantly \u003cb\u003e(Fig.\u0026nbsp;6D)\u003c/b\u003e, it clearly enhanced phosphorylation of STAT1 and STAT3 \u003cb\u003e(Fig.\u0026nbsp;6E\u0026ndash;F)\u003c/b\u003e. Co-stimulation with IL-2 and IL-21 further elevated p-STAT5 levels, indicating synergistic JAK-STAT activation \u003cb\u003e(Fig.\u0026nbsp;6G)\u003c/b\u003e. Expansion folds were 17-, 52-, 45-, and 128-fold for the IL-2-/IL-21-, IL-2+/IL-21-, IL-2-/IL-21+, and IL-2+/IL-21\u0026thinsp;+\u0026thinsp;groups, respectively \u003cb\u003e(Fig.\u0026nbsp;6H)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTogether, these results underscore that IL-2 and IL-21 are essential for activating the JAK-STAT pathway and driving NK cell expansion, with their combination exerting a potent synergistic effect.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOver the past two decades, immunotherapy with functionally improved T and NK cells has emerged as a novel pillar of cancer treatment (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). Memory NK cells, especially CIML NK cells, represent an attractive cell type for developing this therapeutic strategy, mainly due to their enhanced effector function, enhanced ability to persist and proliferate \u003cem\u003ein vivo\u003c/em\u003e (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). In this context, the robust expansion \u003cem\u003eex vivo\u003c/em\u003e of these cells is essential in order to meet clinical demands. Currently, few established methods exist for expanding memory-like NK cells (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eOne study showed that pre-activating NK cells with cytokines (IL-12, IL-15, and IL-18) together with stimulation by PM21 particles\u0026mdash;derived from K562-41BBL-mbIL21 feeder cells\u0026mdash;resulted in an 8200-fold expansion of NK cells within two weeks (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). However, the absence of proper CIML NK controls and \u003cem\u003ein vivo\u003c/em\u003e tumor-killing assays made the reported expansion of memory-like NK cells less convincing. Moreover, the therapeutic use of such tumor feeder cell-derived particles presents considerable safety concerns, in addition to challenges related to GMP-grade storage and handling. In the same year, Shrestha et al. reported a feeder-free protocol based on a human tissue factor (TF) fusion protein incorporating IL-12/15/18 cytokines (HCW9206), along with an anti-TF-TF IgG1 antibody (HCW9101) that may deliver a stimulating signal via CD16a receptor (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). This combination of signals induced approximately a 250-fold expansion of memory NK cells over two weeks. Although such cytokine fusion proteins enable efficient \u003cem\u003eex vivo\u003c/em\u003e expansion of memory-like NK cells, their adoption has been limited by the specialized preparation required, making it challenging for other research groups to reproduce the methodology. In the present study, we developed a novel feeder-free system for the \u003cem\u003eex vivo\u003c/em\u003e expansion of memory like NK cells. All necessary materials are commercially available reagents, and the culture procedure is relatively simple, which should facilitate broader translational applications in the future.\u003c/p\u003e\u003cp\u003eBased on insights from memory-like NK cell studies, combinations of CD16a antibody with cytokines such as IL-12, IL-15, and IL-18 led to enhanced IFN-γ production upon restimulation with tumor cells or cytokines, underscoring the importance of CD16a engagement in conferring memory-like properties (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). However, activation with cytokines or exposure to target cells were found to substantially reduce CD16a expression. Previous studies have shown that NK cells express ADAM17, and selective inhibition of this metalloprotease prevented proteolytic shedding of CD16a, thereby augmenting interferon-γ production\u0026mdash;especially when activation was mediated via CD16a (\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). Thus, we selected a combination of CD16 antibody, IL-15, IL-18, and the ADAM17 inhibitor INCB3619 as the initial signaling cocktail to induce immune memory in NK cells.\u003c/p\u003e\u003cp\u003eActivation of the CD137 signaling axis is essential for achieving optimal NK cell expansion. Previous studies indicate that the synergistic interaction among CD137L, IL-15, and IL-21 critically influences the extent of NK cell proliferation (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e). The anti-CD137 agonist urelumab has been shown to sustain expression of key activating receptors such as NKG2D and effector molecules including granzyme B and IFN-γ, thereby counteracting TGFβ-mediated suppression of human NK cell proliferation and antitumor function (\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e). Moreover, combined stimulation with IL-2 and IL-21 co-stimulation has been identified as the most effective combination for enhancing NK cell cytotoxicity, sustaining proliferative capacity, and enhancing IFN-γ secretion (\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e). Further evidence confirms that IL-21 signaling activates multiple pathways\u0026mdash;including Jak-STAT, MAPK, and PI3K\u0026mdash;to collectively support NK cell expansion (\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). Based on the aforementioned research findings, we supplemented the dT-PBMC culture with CD137 antibody (Urelumab), IL-2, and IL-21 at 48 hours after pre-activation, followed by regular additions of IL-2/15/21 starting from day 4 to optimize NK cell expansion outcomes \u003cb\u003e(Fig.\u0026nbsp;1)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eThe efficient expansion of memory NK cells offers an excellent opportunity for us to understand their biological attributes. Our data comprehensively demonstrate that mExNK cells possess classic memory-like attributes across multiple layers. The mExNK cells exhibited a definitive memory phenotype, characterized by high CD25 expression similar to CIML NK cells \u003cb\u003e(Fig.\u0026nbsp;2C, 4G)\u003c/b\u003e, much higher than PBNK and cExNK \u003cb\u003e(Fig.\u0026nbsp;2C and Sup Fig.\u0026nbsp;4E, 4L)\u003c/b\u003e. They also underwent significant metabolic remodeling, evidenced by increased nutrient transporter expression, enhanced glycolysis, and transcriptional upregulation of OXPHOS and mTORC1 signaling pathways, which collectively support their enhanced effector functions \u003cb\u003e(Fig.\u0026nbsp;2D-H, 4D-E; Sup Fig.\u0026nbsp;3F and 4E-F)\u003c/b\u003e. At the epigenetic level, mExNK cells showed much lower methylation in the IFNG CNS-1 region than CIML NK cells \u003cb\u003e(Fig.\u0026nbsp;2I-J; Sup Fig.\u0026nbsp;1D-E)\u003c/b\u003e. Supporting this permissive epigenetic state, mExNK cells produced more IFN-γ upon stimulation \u003cb\u003e(Fig.\u0026nbsp;3A, 3F)\u003c/b\u003e, demonstrating a direct link between their epigenetic and functional enhancements. Concurrently, mExNK cells expressed higher levels of cytotoxic molecules, including Perforin and Granzyme B \u003cb\u003e(Fig.\u0026nbsp;3A-G, Fig.\u0026nbsp;4F and Sup Fig.\u0026nbsp;4G-I)\u003c/b\u003e, and exhibited superior \u003cem\u003ein vitro\u003c/em\u003e tumor-killing and persistence capacity \u003cb\u003e(Fig.\u0026nbsp;3H-J and Sup Fig.\u0026nbsp;2A-I)\u003c/b\u003e compared to CIML NK cells. More importantly, mExNK cells exhibited superior \u003cem\u003ein vivo\u003c/em\u003e tumor-killing and persistence capacity, significantly reduced tumor burden and prolonged mice survival \u003cb\u003e(Fig.\u0026nbsp;5B-F)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eFoltz et al. originally identified CD117⁻CD57⁻NKG2A⁺CD39⁺ enriched memory-like (eML) NK cells as the subpopulation possessing genuine immune memory (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). Our mRNA-seq analysis revealed a high degree of similarity between the transcriptomes of mExNK cells and these eML NK cells, including concordant changes in key genes \u003cb\u003e(Fig.\u0026nbsp;4H)\u003c/b\u003e. Specifically, compared to CIML NK cells, mExNK cells expressed lower levels of KIT (encoding CD117) and B3GAT1 (encoding CD57), and higher levels of ENTPD1 (encoding CD39) \u003cb\u003e(Fig.\u0026nbsp;4H, Sup Fig.\u0026nbsp;1A)\u003c/b\u003e, suggesting a greater enrichment of the bona fide eML subset. Integrating these phenotypic, metabolic, epigenetic, transcriptomic, and functional analyses, we conclude that mExNK cells likely possess more robust immune memory characteristics than CIML NK cells.\u003c/p\u003e\u003cp\u003eTherefore, despite undergoing expansion, mExNK cells retain strong immune memory properties. The expansion of mExNK cells significantly enhances the production efficiency and clinical therapeutic potential of traditional CIML NK cells, while simultaneously improving their functionality, offering greater clinical applicability. However, how to further improve the expansion efficiency of memory-like NK cells remains a challenge to be overcome.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003eMice\u003c/h2\u003e\u003cp\u003e Housing and all experimental protocols for mice used in this study were performed in accordance with the guidelines established by the Institutional Animal Care and Use Committee in the East Hospital of Tongji University. NCG mice (NOD/ShiLtJGpt-Prkdc\u003csup\u003eem26Cd52\u003c/sup\u003eIl2rg\u003csup\u003eem26Cd22\u003c/sup\u003e/Gpt) were purchased from The GemPharmatech. If not stated differently, 6-8-week-old male mice were used for all experiments.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eCell Culture\u003c/h2\u003e\u003cp\u003e\u003cstrong\u003eCell lines\u003c/strong\u003e\u003cp\u003e THP-1-Luciferase cell line was bought from Xiamen Immocell Biotechnology Co.,Ltd., K562 and Raji tumor cell lines were obtained from Cell Bank of the Committee on Type Culture Collection, Chinese Academy of Sciences\u0026zwnj;. They were all cultured in RPMI 1640 complete media (Gibco) supplemented with 10% FBS (Gibco). The information of cell lines are listed in \u003cb\u003eTable \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e\u003c/b\u003e.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003ePBNK cells\u003c/strong\u003e\u003cp\u003eFreshly isolated human peripheral blood NK (PBNK) cells were purified from PBMCs using the NK Cell Isolation Kit (Miltenyi) according to the manufacturer's instructions. Cell purity, defined as the percentage of CD56⁺CD3⁻ cells, was determined by flow cytometry. Only preparations with a purity exceeding 90% were used for subsequent cultivation or pre-activation. Then, PBNK cells were rested in HIPP-T009 medium (Bioengine) supplemented with 5% UltraGRO\u0026trade;-Advanced Serum Replacement (Helios) and low-dose rhIL-15 (1 ng/ml) for about 3\u0026ndash;5 days to generate negative control PBNK cells.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eCIML NK cells\u003c/strong\u003e\u003cp\u003eFreshly purified PBNK were pre-activated with rhIL-12 (10ng/mL), rhIL-15 (100ng/mL), and rhIL-18 (50ng/mL) for 16\u0026ndash;24 hours to generate cytokine induced memory-like NK cells (CIML NK) as positive control NK cells, then harvested, washed, and allowed to differentiate for about 3 days supported with low-dose rhIL-15 (1 ng/ml) as described (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e). PBNK and CIML NK cells were used as negative and positive controls in the FACS test, ELISA assay, metabolism assessment, conventional bisulfite sequencing, cytokine production test and tumor rechallenge assay.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003ecExNK cells\u003c/strong\u003e\u003cp\u003eFreshly purified PBNK cells were activated and expanded with NK cell Activation/Expansion Kit (Miltenyi) by using the NKp46/CD2 coupling microbeads as described per manufacturer\u0026rsquo;s instructions. The expansion medium was HIPP-T009 medium (Bioengine) supplemented with 5% UltraGRO\u0026trade;-Advanced Serum Replacement (Helios), 500U/mL rhIL-2 (PeproTech). After 14 days, this conventional expanded NK cells (cExNK) were used in the FACS tests and ELISA assay.\u003c/p\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eIsolation and expansion of memory-like peripheral blood NK cells\u003c/h3\u003e\n\u003cp\u003eHealthy anonymous human PBMC were obtained by Ficoll (Stem Cell Technology) centrifugation of cells from leukoreduction filters following platelet apheresis. The CD3-depleted PBMCs (with a residual T cell percentage\u0026thinsp;\u0026le;\u0026thinsp;1%) were isolated using Dynabeads\u0026trade; CD3 (Thermo) as described per manufacturer\u0026rsquo;s instructions.\u003c/p\u003e\u003cp\u003eThen dT-PBMCs were placed in anti-CD16a antibody (T\u0026amp;L) immobilization plate containing HIPP-T009 medium (Bioengine) supplemented with 5% UltraGRO\u0026trade;-Advanced Serum Replacement (Helios), 500U/mL rhIL-2 (PeproTech), 50ng/mL rhIL-15 (T\u0026amp;L), 50ng/mL rhIL-18 (T\u0026amp;L) and 4\u0026micro;M ADAM inhibitor (MCE) at 37℃ in a humidified atmosphere containing 5% CO2. After 2 days, fresh medium with 500U/mL rhIL-2 (PeproTech), 50ng/mL rhIL-15 (T\u0026amp;L) and 1\u0026micro;g/mL Urebumab (MCE) were added. From day 4, medium containing 500IU/mL rhIL-2 (PeproTech), 50ng/mL rhIL-15 (T\u0026amp;L) and 50ng/mL rhIL-21 (T\u0026amp;L) was replenished as necessary during the NK expansion process.\u003c/p\u003e\n\u003ch3\u003eFlow Cytometry Analysis\u003c/h3\u003e\n\u003cp\u003eFor analysis of surface markers, cells were stained in PBS (Gibco) containing 1% FBS (Gibco). Surface proteins were stained at 4℃ for 30min. Intracellular staining of SLC1A5, SLC7A5, GLUT1 and cytokines was performed with Cytofix Fixation Buffer (BD) and Perm/Wash Buffer I (BD), according to the manufacturer\u0026rsquo;s instructions. Data were obtained using a CytoFLEX flow cytometer (Beckman Coulter) and analyzed using FlowJo Software. The following antibodies were used: APC anti-human IFN-γ antibody, PE anti-human TNF-α antibody, PE anti-human CD107a (LAMP-1) antibody, FITC anti-human Perforin, APC anti-human/mouse Granzyme B Recombinant antibody, PE anti-human CD56 (NCAM) antibody, FITC anti-human CD56 (NCAM) antibody, APC anti-human CD56 (NCAM) antibody, PE anti-human CD45 antibody, PE anti-human CD3 antibody, PE anti-human CD25 antibody, PE anti-human CD69 antibody, PE anti-human CD314 (NKG2D) antibody, APC anti-human CD159a (NKG2A) antibody, PE anti-human CD57 antibody, PE anti-human CD335 (NKp46) antibody, APC anti-human CD337 (NKp30) antibody, PE anti-human CD336 (NKp44) antibody, APC anti-human CD16 antibody, PE anti-human CD158 (KIR2DL1/S1/S3/S5) antibody, FITC anti-human CD158b/j (KIR2DL2/L3/S2) antibody, FITC anti-human CD158e1 (KIR3DL1) antibody, APC anti-human CD71 antibody, PE anti-human CD360 (IL-21R) Antibody, FITC Mouse IgG1, κ Isotype Ctrl (FC) antibody, PE Mouse IgM, κ Isotype Ctrl antibody, PE Mouse IgG1, κ Isotype Ctrl antibody, FITC Mouse IgG2b, κ Isotype Ctrl antibody, APC Mouse IgG1, κ Isotype Ctrl antibody (all from Biolegend), Alexa Fluor\u0026trade; 488 Rabbit Anti-Human GLUT1 (BD), PE Mouse Anti-Human CD98 (BD), Alexa Fluor\u0026trade; 488 Rabbit IgG Isotype Control (BD), PE Mouse IgG1, κ Isotype Control (BD), Anti-SLC7A5/LAT1 antibody [EPR26260-66] (Abcam), Anti-SLC1A5/ASCT2 antibody[CAL33] (Abcam), Goat Anti-Rabbit IgG H\u0026amp;L / FITC secondary antibody (HZbscience), Phospho-Stat1 (Tyr701) (D4A7) Rabbit mAb (CST), Phospho-Stat3 (Tyr705) (D3A7) XP\u0026reg; Rabbit mAb (CST), Phospho-Stat5 (Tyr694) (D47E7) XP\u0026reg; Rabbit mAb (CST). All antibodies and their working conditions are listed in \u003cb\u003eTable \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eIncucyte-Based NK cytotoxicity Assays\u003c/h2\u003e\u003cp\u003eK562 target cells were labeled with Incucyte\u0026reg; Cytolight Rapid Green Dye (Sartorius, Germany) for 20 min at 37℃ in the dark and washed twice with complete media to quench the labeling reaction. For kinetic analysis of tumor cell killing, Green dye labelled K562 cells were plated at a concentration of 1e4 per well in 96-well flat bottom plate. Expaned NK cell were added at E:T ratios ranging from 1:1 to 40:1. All co-cultures were performed in HIPP culture media. The number of viable target cells was monitored by hourly fluorescence imaging over 24 hours using an IncuCyte S3 live cell imaging and analysis system (Sartorius, Germany). Live cell numbers were quantified by IncuCyte S3 software and normalized to the number of live cells remaing in the target cell-only control group.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eDegranulation Assay and Cytokine Production\u003c/h2\u003e\u003cp\u003ePBNK, CIML NK and ExNK cells were co-cultured with K562 tumor cells for 6 hours at an effector/target ratio of 1:1 in HIPP growth medium with 1 ng/ml IL-15. For degranulation assays, anti-CD107a mAb was added at the start of the functional assay in 1:100 dilution ratio. To assay cytokine production, Brefeldin A and monensin (eBioscience) were added within the first hour of co-culture with K562. After 5 hours, cells were washed and stained for cell surface marker hCD56 (1:200 dilution), fixed/permeabilized and intracellularly stained for IFN-γ, TNF-α, Perforin, Granzyme B (1:100 dilution).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eELISAs\u003c/h2\u003e\u003cp\u003ePBNK, CIML NK and ExNK cells were co-cultured with K562 tumor cells at an effector/target ratio of 1:1 in HIPP growth medium without exogenous cytokines. The supernatant was collected after 24h, and Granzyme B, Perforin, IFN‑γ secretion was measured using the Granzyme B Valukine ELISA Kit (R\u0026amp;D Systems), ELISA Flex Human Perforin (HRP) (MABTECH) and Human IFN-γ Broad range ELISA Kit (STARTER, China) individually according to the manufacturer\u0026rsquo;s instructions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eMetabolism Assessment\u003c/h2\u003e\u003cp\u003e Seahosre assays were performed according to the manufacturer\u0026rsquo;s instructions with modifications to simultaneously analyze glycolysis and oxidative mitochondrial metabolism. Briefly, 1.5e5 cells were plated per well in triplicate into Poly-D-Lysine coated Seahorse XF24 plates and analyzed with a Seahorse XF24 Analyzer (Agilent Technologies). For ECAR assays, the assay medium was prepared by Seahorse XF RPMI Medium (pH 7.4) contained 2mM L-glutamine. Glucose, oligomycin and 2-deoxyglucose (2-DG) at a final concentration of 10mM, 2.5\u0026micro;m and 50mM individually, were serially injected. For OCR assays, the assay medium was prepared by Seahorse XF RPMI Medium contained 10mM Glucose, 1mM Pyruvate, and 2mM L-glutamine, pH 7.4. Mitochondrial respiration was monitored at basal state and after sequential injection of the 2.5\u0026micro;M oligomycin, 1\u0026micro;M FCCP, 2.5\u0026micro;M antimycin A and 2.5\u0026micro;M rotenone.\u003c/p\u003e\u003cp\u003eGlycolysis was calculated as average post-glucose ECAR values minus average basal ECAR values. Glycolytic reserve was calculated as average maximal ECAR values minus post-glucose ECAR values. Glycolytic capacity was calculated as average post-oligomycin ECAR values minus average basal ECAR values. Basal respiration was determined by monitoring OCR in the absence of any inhibitors. SRC (Spare Respiration Capacity) measurements were calculated as average maximal OCR values minus average basal OCR values. ATP respiration was calculated as average basal OCR values minus average post-oligomycin values.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eConventional Bisulfite Sequencing of IFNG-CNS1 region\u003c/h2\u003e\u003cp\u003eGenomic DNA from PBNK, CIML NK and ExNK cells was extracted with FastPure Blood/Cell/Tissue/Bacteria DNA Isolation Mini Kit (Vazyme, DC112-01) and treated with EZ DNA Methylation-Direct Kit according to the manufacturer\u0026rsquo;s instructions. Bisulfite-treated DNA was subjected to PCR amplification and cloned into pGEM-T vector (Promega). Individual clones were sequenced by standard Sanger sequencing. Data were analyzed by BISMA (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://services.ibc.uni-stuttgart.de/BDPC/BISMA/\u003c/span\u003e\u003cspan address=\"http://services.ibc.uni-stuttgart.de/BDPC/BISMA/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) (\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e).\u003c/p\u003e\u003cp\u003ePrimers sequences for the bisulfate PCR and sequencing were listed in \u003cb\u003eTable \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e\u003c/b\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eRepeated Tumor Rechallenge Assays\u003c/h2\u003e\u003cp\u003ePersistence of NK cells in vitro were evaluated in a serial tumor challenge assay as described before(\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). In brief, 2.5e4 or 1.25e4 NK cells and 2.5e4 target cells (K562, Raji or THP-1) labelled with cytolight red (Sartorius) were co-cultured in a poly-D-lysine pre-coated 96-well plate using HIPP fresh growth medium with 1ng/ml IL-15 to support NK survival. Maintained the plate in a 37℃, 5% CO2 incubator for 6 days to perform 3 repeated tumor rechallenge, additional cytolight red labelled tumor targets in fresh medium administered to the co-culture every other day. The number of viable target cells was monitored by every 2 hours fluorescence imaging over 140 hours using the IncuCyte S3 live cell imaging and analysis system (Sartorius). Live cell numbers were quantified as red area and normalized to the beginning time (0d0h0m).\u003c/p\u003e\u003cp\u003e\u003cb\u003eIn vivo\u003c/b\u003e \u003cb\u003etumor xenograft models\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFor in vivo experiments testing ExNK cell function again THP-1 tumor cells in a single infusion model, 6- to 8-week-old male NSG mice were injected intravenously in tail-vein with 2 or 5e6 luciferase expressing THP-1 tumor cells. There days later mice were distributed into 2\u0026ndash;3 groups randomly, each group with at least 8 mice. Memory-like expanded NK cells (mExNK) were resuspended in 1e7/0.2mL PBS buffer and injected intravenously in tail-vein of tumor bearing mice. Transferred NK cells were supported with rhIL-2 (50000U/mice) and rhIL-15 (10 ng/mice) by intraperitoneal (i.p.) injection at least 2 times every week for up to 3 weeks. BLI was performed weekly to monitor tumor progression, at the same time monitor the mice body weigh weekly. THP-1-Luciferase tumor-bearing mice without NK cell transfer (PBS) or with conventional expanded NK cells (cExNK) served as controls. Animal technicians were blinded to expected outcomes. The experiments were performed in the East China Normal University.\u003c/p\u003e\u003cp\u003eBLI was conducted using an IVIS Spectrum, and images were analyzed using Live Imaging Software (Perkin Elmer). Survival was graphed as Kaplan-Meier curves, and the log-rank (Mantel-Cox) test was used to determine statistical significance.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003ePersistence of adoptively transferred human NK cells in NSG mice\u003c/h2\u003e\u003cp\u003e The experiment was approved by the Institutional Animal Care and Use Committee in the East China Normal University. After 1 week of acclimation, a single dose of 1e7 memory-like and conventional expanded NK cells (mExNK and cExNK) was injected intravenously through the tail vein into male NSG mice (6- to 8-week-old; GemPharmatech) with rhIL-2 (50000U/mice) and rhIL-15 (10 ng/mice) (i.p. injection, 2 times per week for 3 weeks) support. For blood collection, about 100\u0026micro;l blood obtained from orbital venous bleeds using EDTA anticoagulant at 1, 2, 4 weeks post-infusion for immediate flow cytometry.\u003c/p\u003e\u003cp\u003eBlood sample were incubated with the appropriate antibody fluorophore conjugates (mouse CD16/32 Fc block (Biolegend, 1:100), viability Live/Dead (Thermo, 1:1000), PE-hCD45 and APC-hCD56 (Biolegend, 1:50)) for 30mins on ice in staining buffer (DPBS\u0026thinsp;+\u0026thinsp;2%FBS\u0026thinsp;+\u0026thinsp;2mM EDTA). Then stained blood samples were incubated with ACK buffer (Gibco) for 5mins on ice to lyse red blood cells. After lysis, samples were washed with cold staining buffer once, then resuspended in 90\u0026micro;l staining buffer plus 10\u0026micro;l CountBright absolute counting bbeads (Thermo), and data was acquired in CytoFLEX flow cytometer (Beckman Coulter). Population densities were quantified to hCD45+\u0026amp;Hcd56\u0026thinsp;+\u0026thinsp;cells/total cells using in sample CountBright beads.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003eRNA-Seq Sample Preparation, Sequencing, and Data Analysis\u003c/h2\u003e\u003cp\u003eTotal RNA was extracted by TRIzol Reagent (Thermo) to constructed library using using VAHTS Universal V10 RNA-seq Library Prep Kit for MGI (Vazyme, NRM606). The libraries were size-selected for cDNA fragments of 200\u0026ndash;300 bp on 2% Low Range Ultra Agarose and then subjected to PCR amplification using Phusion DNA polymerase (NEB) for 15 PCR cycles. Following quantification using TBS380, the paired-end libraries were sequenced using the DNBSEQ-T7 platform (Shanghai BIOZERON Biotech. Co., Ltd).\u003c/p\u003e\u003cp\u003eThe initial paired-end reads underwent trimming and quality control using Trimmomatic, applying the parameters (SLIDINGWINDOW:4:15 MINLEN:75). Subsequently the resulting clean reads were individually aligned to the reference genome in orientation mode using hisat2 software with default settings. The quality of these data was assessed using qualimap_v2.21. Clean reads were aligned to the ribosome database of human genome (hg38) using the alignment tool bowtie2 to evaluate the proportion of rRNA in the featurecount to count each gene reads. We analyzed differentially expressed genes (DEGs) using the R package edgeR for all pairwise comparisons among PBNK, CIML NK, and mExNK samples, and for the comparison between cExNK and mExNK. Genes with an adjusted p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 (using a significance threshold of p-adjust\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were considered statistically significant DEGs.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eFigures were prepared using GraphPad Prism8. All data are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;s.e.m, excepting informed specially. Statistical significance excepting the survival curve of mice and the ELISA result of IFN-γ was calculated using an unpaired, two-tailed student\u0026rsquo;s t-test with 95% confident intervals. ns, not significant (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). * p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, ** p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, *** p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. The statistical significance of mice survival curve was calculated using Log-rank (Mantel-Cox) test. The statistical significance of IFN-γ ELISA result was calculated using F-test.\u003c/p\u003e\u003c/div\u003e"},{"header":"Abbreviations","content":"\u003cp\u003e\u003cstrong\u003eACT,\u0026nbsp;\u003c/strong\u003edoptive cell therapy;\u003cstrong\u003e\u0026nbsp;ADCC,\u0026nbsp;\u003c/strong\u003eAntibody-Dependent Cellular Cytotoxicity;\u003cstrong\u003e\u0026nbsp;AML\u003c/strong\u003e, Acute Myelocytic Leukemia;\u0026nbsp;\u003cstrong\u003eCIFF,\u003c/strong\u003e Cytokine-induced and Feeder-free; \u003cstrong\u003eCIML NK,\u0026nbsp;\u003c/strong\u003eCytokine Induced Memory-like NK cells; \u003cstrong\u003eCNS-1,\u003c/strong\u003e Conserved Noncoding Sequence 1;\u0026nbsp;\u003cstrong\u003eECAR,\u003c/strong\u003e Extracellular Acidification Rate;\u0026nbsp;\u003cstrong\u003eELISA\u003c/strong\u003e, Enzyme-Linked Immunosorbent Assay; \u003cstrong\u003ecExNK\u003c/strong\u003e, Convential Expanded NK cells; \u003cstrong\u003emExNK\u003c/strong\u003e, Memory-like Expanded NK cells; \u003cstrong\u003eEZH2,\u003c/strong\u003e Enhancer of Zeste Homolog 2; \u003cstrong\u003eGSEA,\u003c/strong\u003e Gene Set Enrichment Analysis; \u003cstrong\u003eIL\u003c/strong\u003e, Interleukin, \u003cstrong\u003eMDS,\u003c/strong\u003e Myelodysplastic Syndromes;\u0026nbsp;\u003cstrong\u003eMFI\u003c/strong\u003e, Mean Fluorescent Intensity;\u0026nbsp;\u003cstrong\u003eeML NK,\u003c/strong\u003e Enriched Memory-Like NK cells; \u003cstrong\u003eMM\u003c/strong\u003e, Multiple Myeloma; \u003cstrong\u003eNK\u003c/strong\u003e, Natural killer;\u0026nbsp;\u003cstrong\u003eOCR\u003c/strong\u003e,\u0026nbsp;Oxygen Consumption Rate; \u003cstrong\u003eORA,\u0026nbsp;\u003c/strong\u003eOver Representation Analysis; \u003cstrong\u003ePBNK\u003c/strong\u003e, Peripheral NK cell; \u003cstrong\u003eSEM,\u0026nbsp;\u003c/strong\u003eStandard Error of the Mean; \u003cstrong\u003eTFs,\u003c/strong\u003e Transcriptional Factors; \u003cstrong\u003eVIPER,\u003c/strong\u003e Visualization Pipeline for RNA-seq analysis;\u0026nbsp;\u003c/p\u003e\n"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was funded by the National Natural Science Foundation of China (82471592, 82270638, 82300718, 82301904), Shanghai Clinical Research Center for Cell Therapy (23J41900100), Shanghai Engineering Research Center of Stem Cells Translational Medicine (20DZ2255100), Peak Disciplines (Type IV) of Institutions of Higher Learning in Shanghai.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe authors declared that these funders had no role in the study design, data collection, analysis, interpretation, manuscript writing, or the decision to submit the article for publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eQ.D., Z.H. and H.W. performed study concept and design; W.Z., Q.D., Z.H. and H.W. performed development of methodology and writing, review and revision of the paper; H.W. proved acquisition, analysis and interpretation of data, and statistical analysis; H.W. performed the vast majority assays, excepting the ELISA test and development of expansion system. Q.D., J.T. and X.Y. developed the expansion system of memory-like NK cells. J.T., X.Y., F.X. and Z.L. performed the ELISA assay and some phenotype tests. H.W. and Y.L. performed the mRNA-seq and analysis. L.H. and Y.X. provided technique support of mice assay. All authors read and approved the final paper.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe studies involving human peripheral blood and PBMC cells from healthy volunteers were approved by Shanghai Hycells Biotechnology Co., Ltd (Ethics Approval Number: HYS-LQ-001 and EC-SMP-2024017). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was required from the participants. The animal study was approved by the Institutional Animal Care and Use Committee in the East Hospital of Tongji University (Ethics Approval Number: Animal Research Review No. (273), 2024).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eDeclaration of interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eLiu S, Galat V, Galat Y, Lee YKA, Wainwright D, Wu J. NK cell-based cancer immunotherapy: from basic biology to clinical development. Journal of hematology \u0026amp; oncology. 2021;14(1):7.\u003c/li\u003e\n\u003cli\u003eVivier E, Raulet DH, Moretta A, Caligiuri MA, Zitvogel L, Lanier LL, et al. Innate or adaptive immunity? The example of natural killer cells. Science (New York, NY). 2011;331(6013):44-9.\u003c/li\u003e\n\u003cli\u003eMartinet L, Smyth MJ. Balancing natural killer cell activation through paired receptors. Nature reviews Immunology. 2015;15(4):243-54.\u003c/li\u003e\n\u003cli\u003eIliopoulou EG, Kountourakis P, Karamouzis MV, Doufexis D, Ardavanis A, Baxevanis CN, et al. A phase I trial of adoptive transfer of allogeneic natural killer cells in patients with advanced non-small cell lung cancer. Cancer immunology, immunotherapy : CII. 2010;59(12):1781-9.\u003c/li\u003e\n\u003cli\u003eSakamoto N, Ishikawa T, Kokura S, Okayama T, Oka K, Ideno M, et al. Phase I clinical trial of autologous NK cell therapy using novel expansion method in patients with advanced digestive cancer. Journal of translational medicine. 2015;13:277.\u003c/li\u003e\n\u003cli\u003eBerrien-Elliott MM, Jacobs MT, Fehniger TA. Allogeneic natural killer cell therapy. Blood. 2023;141(8):856-68.\u003c/li\u003e\n\u003cli\u003eBowdish DM, Loffredo MS, Mukhopadhyay S, Mantovani A, Gordon S. Macrophage receptors implicated in the \u0026quot;adaptive\u0026quot; form of innate immunity. Microbes and infection. 2007;9(14-15):1680-7.\u003c/li\u003e\n\u003cli\u003eNetea MG, Quintin J, van der Meer JW. Trained immunity: a memory for innate host defense. Cell host \u0026amp; microbe. 2011;9(5):355-61.\u003c/li\u003e\n\u003cli\u003eKleinnijenhuis J, Quintin J, Preijers F, Joosten LA, Ifrim DC, Saeed S, et al. Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of monocytes. Proceedings of the National Academy of Sciences of the United States of America. 2012;109(43):17537-42.\u003c/li\u003e\n\u003cli\u003eOchando J, Mulder WJM, Madsen JC, Netea MG, Duivenvoorden R. Trained immunity - basic concepts and contributions to immunopathology. Nature reviews Nephrology. 2023;19(1):23-37.\u003c/li\u003e\n\u003cli\u003eRomee R, Schneider SE, Leong JW, Chase JM, Keppel CR, Sullivan RP, et al. Cytokine activation induces human memory-like NK cells. Blood. 2012;120(24):4751-60.\u003c/li\u003e\n\u003cli\u003eKleinnijenhuis J, Quintin J, Preijers F, Joosten LA, Jacobs C, Xavier RJ, et al. BCG-induced trained immunity in NK cells: Role for non-specific protection to infection. Clinical immunology (Orlando, Fla). 2014;155(2):213-9.\u003c/li\u003e\n\u003cli\u003eSchlums H, Cichocki F, Tesi B, Theorell J, Beziat V, Holmes TD, et al. Cytomegalovirus infection drives adaptive epigenetic diversification of NK cells with altered signaling and effector function. Immunity. 2015;42(3):443-56.\u003c/li\u003e\n\u003cli\u003ePeng H, Tian Z. Natural Killer Cell Memory: Progress and Implications. Frontiers in immunology. 2017;8:1143.\u003c/li\u003e\n\u003cli\u003eTarannum M, Romee R. Cytokine-induced memory-like natural killer cells for cancer immunotherapy. Stem cell research \u0026amp; therapy. 2021;12(1):592.\u003c/li\u003e\n\u003cli\u003eCooper MA, Elliott JM, Keyel PA, Yang L, Carrero JA, Yokoyama WM. Cytokine-induced memory-like natural killer cells. Proceedings of the National Academy of Sciences of the United States of America. 2009;106(6):1915-9.\u003c/li\u003e\n\u003cli\u003eFoltz JA, Tran J, Wong P, Fan C, Schmidt E, Fisk B, et al. Cytokines drive the formation of memory-like NK cell subsets via epigenetic rewiring and transcriptional regulation. Science immunology. 2024;9(96):eadk4893.\u003c/li\u003e\n\u003cli\u003eZhang C, Yin J, Zheng J, Xiao J, Hu J, Su Y, et al. EZH2 identifies the precursors of human natural killer cells with trained immunity. Cancer biology \u0026amp; medicine. 2021;18(4):1021-39.\u003c/li\u003e\n\u003cli\u003eTerr\u0026eacute;n I, Orrantia A, Mosteiro A, Vitall\u0026eacute; J, Zenarruzabeitia O, Borrego F. Metabolic changes of Interleukin-12/15/18-stimulated human NK cells. Scientific reports. 2021;11(1):6472.\u003c/li\u003e\n\u003cli\u003eNi J, H\u0026ouml;lsken O, Miller M, Hammer Q, Luetke-Eversloh M, Romagnani C, et al. Adoptively transferred natural killer cells maintain long-term antitumor activity by epigenetic imprinting and CD4(+) T cell help. Oncoimmunology. 2016;5(9):e1219009.\u003c/li\u003e\n\u003cli\u003eBecker-Hapak MK, Shrestha N, McClain E, Dee MJ, Chaturvedi P, Leclerc GM, et al. A Fusion Protein Complex that Combines IL-12, IL-15, and IL-18 Signaling to Induce Memory-Like NK Cells for Cancer Immunotherapy. Cancer immunology research. 2021;9(9):1071-87.\u003c/li\u003e\n\u003cli\u003eBerrien-Elliott MM, Foltz JA, Russler-Germain DA, Neal CC, Tran J, Gang M, et al. Hematopoietic cell transplantation donor-derived memory-like NK cells functionally persist after transfer into patients with leukemia. Science translational medicine. 2022;14(633):eabm1375.\u003c/li\u003e\n\u003cli\u003eKeppel MP, Yang L, Cooper MA. Murine NK cell intrinsic cytokine-induced memory-like responses are maintained following homeostatic proliferation. Journal of immunology (Baltimore, Md : 1950). 2013;190(9):4754-62.\u003c/li\u003e\n\u003cli\u003eBednarski JJ, Zimmerman C, Berrien-Elliott MM, Foltz JA, Becker-Hapak M, Neal CC, et al. Donor memory-like NK cells persist and induce remissions in pediatric patients with relapsed AML after transplant. Blood. 2022;139(11):1670-83.\u003c/li\u003e\n\u003cli\u003eShapiro RM, Birch GC, Hu G, Vergara Cadavid J, Nikiforow S, Baginska J, et al. Expansion, persistence, and efficacy of donor memory-like NK cells infused for posttransplant relapse. The Journal of clinical investigation. 2022;132(11).\u003c/li\u003e\n\u003cli\u003eBerrien-Elliott MM, Cashen AF, Cubitt CC, Neal CC, Wong P, Wagner JA, et al. Multidimensional Analyses of Donor Memory-Like NK Cells Reveal New Associations with Response after Adoptive Immunotherapy for Leukemia. Cancer discovery. 2020;10(12):1854-71.\u003c/li\u003e\n\u003cli\u003eDeng X, Terunuma H, Nieda M, Xiao W, Nicol A. Synergistic cytotoxicity of ex vivo expanded natural killer cells in combination with monoclonal antibody drugs against cancer cells. International immunopharmacology. 2012;14(4):593-605.\u003c/li\u003e\n\u003cli\u003eHuang RS, Lai MC, Shih HA, Lin S. A robust platform for expansion and genome editing of primary human natural killer cells. The Journal of experimental medicine. 2021;218(3).\u003c/li\u003e\n\u003cli\u003eLi X, He C, Liu C, Ma J, Ma P, Cui H, et al. Expansion of NK cells from PBMCs using immobilized 4-1BBL and interleukin-21. International journal of oncology. 2015;47(1):335-42.\u003c/li\u003e\n\u003cli\u003ePahl JHW, Koch J, G\u0026ouml;tz JJ, Arnold A, Reusch U, Gantke T, et al. CD16A Activation of NK Cells Promotes NK Cell Proliferation and Memory-Like Cytotoxicity against Cancer Cells. Cancer immunology research. 2018;6(5):517-27.\u003c/li\u003e\n\u003cli\u003eMishra HK, Dixon KJ, Pore N, Felices M, Miller JS, Walcheck B. Activation of ADAM17 by IL-15 Limits Human NK Cell Proliferation. Frontiers in immunology. 2021;12:711621.\u003c/li\u003e\n\u003cli\u003ePomeroy EJ, Hunzeker JT, Kluesner MG, Lahr WS, Smeester BA, Crosby MR, et al. A Genetically Engineered Primary Human Natural Killer Cell Platform for Cancer Immunotherapy. Molecular therapy : the journal of the American Society of Gene Therapy. 2020;28(1):52-63.\u003c/li\u003e\n\u003cli\u003eParrish-Novak J, Dillon SR, Nelson A, Hammond A, Sprecher C, Gross JA, et al. Interleukin 21 and its receptor are involved in NK cell expansion and regulation of lymphocyte function. Nature. 2000;408(6808):57-63.\u003c/li\u003e\n\u003cli\u003eShman TV, Vashkevich KP, Migas AA, Matveyenka MA, Lasiukov YA, Mukhametshyna NS, et al. Phenotypic and functional characterisation of locally produced natural killer cells ex vivo expanded with the K562-41BBL-mbIL21 cell line. Clinical and experimental medicine. 2023;23(6):2551-60.\u003c/li\u003e\n\u003cli\u003eVidard L. 4-1BB and cytokines trigger human NK, \u0026gamma;\u0026delta; T, and CD8(+) T cell proliferation and activation, but are not required for their effector functions. Immunity, inflammation and disease. 2023;11(1):e749.\u003c/li\u003e\n\u003cli\u003eBakhtiyaridovvombaygi M, Yazdanparast S, Mikanik F, Izadpanah A, Parkhideh S, Shahbaz Ghasabeh A, et al. Cytokine-Induced Memory-Like NK Cells: Emerging strategy for AML immunotherapy. Biomedicine \u0026amp; pharmacotherapy = Biomedecine \u0026amp; pharmacotherapie. 2023;168:115718.\u003c/li\u003e\n\u003cli\u003eLeong JW, Chase JM, Romee R, Schneider SE, Sullivan RP, Cooper MA, et al. Preactivation with IL-12, IL-15, and IL-18 induces CD25 and a functional high-affinity IL-2 receptor on human cytokine-induced memory-like natural killer cells. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation. 2014;20(4):463-73.\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Brien KL, Finlay DK. Immunometabolism and natural killer cell responses. Nature reviews Immunology. 2019;19(5):282-90.\u003c/li\u003e\n\u003cli\u003eKeating SE, Zaiatz-Bittencourt V, Loftus RM, Keane C, Brennan K, Finlay DK, et al. Metabolic Reprogramming Supports IFN-\u0026gamma; Production by CD56bright NK Cells. Journal of immunology (Baltimore, Md : 1950). 2016;196(6):2552-60.\u003c/li\u003e\n\u003cli\u003eNicholas D, Proctor EA, Raval FM, Ip BC, Habib C, Ritou E, et al. Advances in the quantification of mitochondrial function in primary human immune cells through extracellular flux analysis. PloS one. 2017;12(2):e0170975.\u003c/li\u003e\n\u003cli\u003eTato CM, Martins GA, High FA, DiCioccio CB, Reiner SL, Hunter CA. Cutting Edge: Innate production of IFN-gamma by NK cells is independent of epigenetic modification of the IFN-gamma promoter. Journal of immunology (Baltimore, Md : 1950). 2004;173(3):1514-7.\u003c/li\u003e\n\u003cli\u003eBalasubramani A, Mukasa R, Hatton RD, Weaver CT. Regulation of the Ifng locus in the context of T-lineage specification and plasticity. Immunological reviews. 2010;238(1):216-32.\u003c/li\u003e\n\u003cli\u003eWilson CB, Rowell E, Sekimata M. Epigenetic control of T-helper-cell differentiation. Nature reviews Immunology. 2009;9(2):91-105.\u003c/li\u003e\n\u003cli\u003eLong AH, Haso WM, Shern JF, Wanhainen KM, Murgai M, Ingaramo M, et al. 4-1BB costimulation ameliorates T cell exhaustion induced by tonic signaling of chimeric antigen receptors. Nature medicine. 2015;21(6):581-90.\u003c/li\u003e\n\u003cli\u003eGattinoni L, Klebanoff CA, Palmer DC, Wrzesinski C, Kerstann K, Yu Z, et al. Acquisition of full effector function in vitro paradoxically impairs the in vivo antitumor efficacy of adoptively transferred CD8+ T cells. The Journal of clinical investigation. 2005;115(6):1616-26.\u003c/li\u003e\n\u003cli\u003eWang D, Starr R, Alizadeh D, Yang X, Forman SJ, Brown CE. In Vitro Tumor Cell Rechallenge For Predictive Evaluation of Chimeric Antigen Receptor T Cell Antitumor Function. Journal of visualized experiments : JoVE. 2019(144).\u003c/li\u003e\n\u003cli\u003eGotthardt D, Trifinopoulos J, Sexl V, Putz EM. JAK/STAT Cytokine Signaling at the Crossroad of NK Cell Development and Maturation. Frontiers in immunology. 2019;10:2590.\u003c/li\u003e\n\u003cli\u003ePanwar V, Singh A, Bhatt M, Tonk RK, Azizov S, Raza AS, et al. Multifaceted role of mTOR (mammalian target of rapamycin) signaling pathway in human health and disease. Signal transduction and targeted therapy. 2023;8(1):375.\u003c/li\u003e\n\u003cli\u003eCaruso S, De Angelis B, Del Bufalo F, Ciccone R, Donsante S, Volpe G, et al. Safe and effective off-the-shelf immunotherapy based on CAR.CD123-NK cells for the treatment of acute myeloid leukaemia. Journal of hematology \u0026amp; oncology. 2022;15(1):163.\u003c/li\u003e\n\u003cli\u003eVahidi S, Zabeti Touchaei A, Samadani AA. IL-15 as a key regulator in NK cell-mediated immunotherapy for cancer: From bench to bedside. International immunopharmacology. 2024;133:112156.\u003c/li\u003e\n\u003cli\u003eOjo EO, Sharma AA, Liu R, Moreton S, Checkley-Luttge MA, Gupta K, et al. Membrane bound IL-21 based NK cell feeder cells drive robust expansion and metabolic activation of NK cells. Scientific reports. 2019;9(1):14916.\u003c/li\u003e\n\u003cli\u003eSharma R, Das A. IL-2 mediates NK cell proliferation but not hyperactivity. Immunologic research. 2018;66(1):151-7.\u003c/li\u003e\n\u003cli\u003eWu Y, Tian Z, Wei H. Developmental and Functional Control of Natural Killer Cells by Cytokines. Frontiers in immunology. 2017;8:930.\u003c/li\u003e\n\u003cli\u003eSun D, Shi X, Li S, Wang X, Yang X, Wan M. CAR‑T cell therapy: A breakthrough in traditional cancer treatment strategies (Review). Molecular medicine reports. 2024;29(3).\u003c/li\u003e\n\u003cli\u003eShrestha N, Dee MJ, Chaturvedi P, Leclerc GM, Mathyer M, Dufour C, et al. A \u0026quot;Prime and Expand\u0026quot; strategy using the multifunctional fusion proteins to generate memory-like NK cells for cell therapy. Cancer immunology, immunotherapy : CII. 2024;73(9):179.\u003c/li\u003e\n\u003cli\u003eOyer JL, Croom-Perez TJ, Hasan MF, Rivera-Huertas JA, Gitto SB, Mucha JM, et al. PM21-particle stimulation augmented with cytokines enhances NK cell expansion and confers memory-like characteristics with enhanced survival. Frontiers in immunology. 2024;15:1383281.\u003c/li\u003e\n\u003cli\u003eRomee R, Foley B, Lenvik T, Wang Y, Zhang B, Ankarlo D, et al. NK cell CD16 surface expression and function is regulated by a disintegrin and metalloprotease-17 (ADAM17). Blood. 2013;121(18):3599-608.\u003c/li\u003e\n\u003cli\u003eVidard L, Dureuil C, Baudhuin J, Vescovi L, Durand L, Sierra V, et al. CD137 (4-1BB) Engagement Fine-Tunes Synergistic IL-15- and IL-21-Driven NK Cell Proliferation. Journal of immunology (Baltimore, Md : 1950). 2019;203(3):676-85.\u003c/li\u003e\n\u003cli\u003eCabo M, Santana-Hern\u0026aacute;ndez S, Costa-Garcia M, Rea A, Lozano-Rodr\u0026iacute;guez R, Ataya M, et al. CD137 Costimulation Counteracts TGF\u0026beta; Inhibition of NK-cell Antitumor Function. Cancer immunology research. 2021;9(12):1476-90.\u003c/li\u003e\n\u003cli\u003eHwang I, Lee J, Kim SW, Sung YC. Dual IL-2/IL-21 priming augments the anti-tumor function of CD16V-CAR/TRAIL-engineered NK cells in combination with monoclonal antibodies. Biomedicine \u0026amp; pharmacotherapy = Biomedecine \u0026amp; pharmacotherapie. 2025;189:118289.\u003c/li\u003e\n\u003cli\u003eZeng R, Spolski R, Casas E, Zhu W, Levy DE, Leonard WJ. The molecular basis of IL-21-mediated proliferation. Blood. 2007;109(10):4135-42.\u003c/li\u003e\n\u003cli\u003eRomee R, Rosario M, Berrien-Elliott MM, Wagner JA, Jewell BA, Schappe T, et al. Cytokine-induced memory-like natural killer cells exhibit enhanced responses against myeloid leukemia. Science translational medicine. 2016;8(357):357ra123.\u003c/li\u003e\n\u003cli\u003eRohde C, Zhang Y, Reinhardt R, Jeltsch A. BISMA--fast and accurate bisulfite sequencing data analysis of individual clones from unique and repetitive sequences. BMC bioinformatics. 2010;11:230.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"cell-death-discovery","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"cddiscovery","sideBox":"Learn more about [Cell Death Discovery](http://www.nature.com/cddiscovery/)","snPcode":"41420","submissionUrl":"https://mts-cddiscovery.nature.com/","title":"Cell Death Discovery","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"memory-like NK cells, ex vivo expansion, metabolism reprograming, epigenetic remodeling, persistence","lastPublishedDoi":"10.21203/rs.3.rs-8091256/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8091256/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eRecent investigations have indicated that NK cells induced overnight with IL-12/15/18 cytokines possess memory-like characteristics (refers as CIML NK, cytokine induced memory-like NK cells), demonstrating enhanced effector function and prolonged persistence. Preclinical studies and very early clinical trials revealed the promising clinical activity of CIML NK cells in the treatment of hematological malignancies, especially acute myelocytic leukemia (AML). However, the current manufacturing method for CIML NK cells do not endow these cells the ability to expand robustly \u003cem\u003eex vivo\u003c/em\u003e. The low production efficiency limits their clinical application. In this study, we established a novel expansion platform for these memory-like NK cells. This method can effectively and stably expand healthy donor-derived Peripheral NK (PBNK) cells by more than 200-fold within two weeks, with a purity exceeding 95%. Importantly, these \u003cem\u003eex vivo\u003c/em\u003e expanded NK (ExNK) cells exhibited hypomethylated state in the CNS-1 region (Conserved Noncoding Sequence 1) of IFNG gene, stronger metabolic ability, and enhanced effector function when compared to CIML NK cells. Based on these features, we named these expanded NK cells but still possess memory-like characteristics as mExNK (memory-like expanded NK cells). Bulk mRNA-seq further uncovered a high similarity between mExNK and CIML NK cells, with both highly expressing genes being related to proliferation, metabolism, and memory. Mouse tumor models proved that the infused mExNK exhibited longer persistence \u003cem\u003ein vivo\u003c/em\u003e and stronger effector functions. Taken together, the method described herein showed an enhanced expansion efficiency for human memory NK cells and would facilitate their clinic applications in cancer therapy.\u003c/p\u003e","manuscriptTitle":"Cytokine-induced Expandable Memory NK Cells with Significant Metabolic, Epigenetic Remodeling, and Persistence Properties","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-28 15:02:21","doi":"10.21203/rs.3.rs-8091256/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"transferred","content":"Cell Death Discovery","date":"2026-02-18T04:10:22+00:00","index":"","fulltext":""},{"type":"decision","content":"Reject after peer review","date":"2026-02-10T16:54:08+00:00","index":"","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-25T08:15:42+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-13T15:26:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"Cell Death \u0026 Disease","date":"2025-11-13T01:04:59+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2025-11-12T14:54:53+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-12T02:39:56+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"cell-death-discovery","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"cddiscovery","sideBox":"Learn more about [Cell Death Discovery](http://www.nature.com/cddiscovery/)","snPcode":"41420","submissionUrl":"https://mts-cddiscovery.nature.com/","title":"Cell Death Discovery","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"af1adc66-bd20-46a3-a467-d3112683792e","owner":[],"postedDate":"November 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":58564211,"name":"Biological sciences/Immunology/Translational immunology"},{"id":58564212,"name":"Biological sciences/Immunology/Adaptive immunity/Cellular immunity/Immunological memory"}],"tags":[],"updatedAt":"2026-03-03T08:00:53+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-28 15:02:21","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8091256","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8091256","identity":"rs-8091256","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}